ISSN:0308-6003    

DOI:10.1039/PR97774FX001

出版商:RSC    

年代:1977

数据来源: RSC

ISSN:0308-6003    

DOI:10.1039/PR97774BX003

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

2 Electrolyte Solutions By A. K. COVINGTON Department of Physical Chemistry University of Newcastle Newcastle-upon- Tyne NE17RU and A. 0. PETHYBRIDGE Department of Chemistry University of Reading Reading RG6 2AD 1 Introduction In September 1977 the Faraday Division of the Chemical Society held a General Discussion on ‘Ion-Ion and Ion-Solvent Interactions’.’ This took place in Oxford almost exactly 50 years after the celebrated 1927 General Discussion of the Faraday Society there on the ‘Theory of Strong Electrolytes’2 and 20 years after the 1957 General Discussion entitled ‘Interaction in Ionic solution^'.^ The Senior Reporter suggested that for this Annual Report we should while reviewing the 1977 literature consider assessing the progress that has been made in understand- ing the behaviour of electrolyte solutions over the past 50 years.We have also referred to the appropriate Annual Reports for 1926,4 1927,5 and 19586 in order to place the General Discussion materisri in perspective. The selection of material of course reflects the interests of the Reporters. Last year’s Annual Reports contained an article on solvation’ and some overlap of coverage with this is inevitable as it is impossible to discuss ionic solutions without reference to the important problem of the solvent molecules closest to an ion and their interaction with it. Views on the nature and extent of solvation effects at increasing distance from the ion remain controversial. It is here that spec- troscopic methods of study have made greatest impact.Unfortunately the 1977 Oxford meeting revealed a wide gulf between pro- ponents of spectroscopic methods and those who have never made such measure- ments themselves and retain a profound distrust of the conclusions reached from them. The interpretation of spectral information is only unequivocal when entirely new spectral features appear and variation of other parameters such as concen- tration temperature etc. permits unambiguous identification of the species responsible. Thus fine structural detail can be obtained which is not furnished by classical methods. Otherwise spectral information consists of band shifts or similar ’ Faraday Discuss. Chem. SOC.,1977,No.64. ’ Trans. Faraday SOC.,1927,23 333-554. Discuss.Faraday SOC.,1957,No.24.J. E.Coates and J. A. V. Butler Ann. Reports 1926 23 11. H. Hunter Ann. Reports 1927,24,11. J. E.Prue Ann. Reports 1958,55 14. ’M. C. R. Symons Ann. Reports (A) 1976,73,91. 5 A. K. Covington and A. D. Pethybridge changes in spectral parameters the interpretation of which is by inference in the same way as for example the existence of ion-pair formation is inferred from abnormal thermodynamic behaviour. The interpretation must have regard to the appropriate spectroscopic rules and it is when entirely independent interpretations of a given system’s behaviour coincide that one can have confidence in the conclusions. Too often spectroscopic results have been evaluated with reference to the models and figurative picturesque descriptions introduced in the 1950’s of for example structure breaking and making flickering clusters ‘ice likeness’,’ and hydrophobic interactions which although harmless in themselves have been widely adopted and misused by biologists and others.The recent Oxford meeting revealed considerable differences even between some of the principal workers in the field in the meaning of some of the commoner terms used in describing ionic solution behaviour. There remains far too much recourse to such qualitative devices and other intuitive considerations. It is therefore no wonder that those who seek to understand interactions in complicated biological systems involving ion transport and equilibria in and across membranes tend to misuse such terms.Therein lies one of the principal problems of the electrolyte solution field. The area of interest is vast but the number of workers who would call themselves ‘solution chemists’ is small and even then the range of interest covered is enormous. Another problem lies in the sophistication of much of the experimental apparatus and data-handling methods now available. It is very difficult for anyone who has not actually had direct experience of a particular technique to be able to appreciate its contribution and assess results from it critically; thus implicit assumptions in data treatment and inherent theoretical assumptions tend to be overlooked and results and conclusions are regarded as more definitive than can be justified. Economic stringencies coupled with increased instrumental sophistication have led to the multi-technique researcher who was such an encouraging phenomenon of the 1960’s becoming much less common.2 Conductance Conductance measurements although not requiring the on-line linking to computers now regarded as almost essential for many spectroscopic measurements nevertheless depend heavily on computer data analysis for the testing of improved theories. In fact the experimental technique has had to be pushed to *0.005 S cm2 mol-1 in order to discriminate between rival theories. By some quirk of fate or masterstroke of timing each of the three Faraday discussion^'-^ has coincided with important advances in the theoretical back- ground to conductance studies. The topic was widely discussed at the 1927 Dis-cussion the first half being devoted to the mobilities of ions with Onsager’ giving a summary of results obtained with his classical extension” of the earlier work of Debye and Huckel.It is clear that many of the topics discussed in 1927 are still of interest today although the advent of more accurate measurements and analysis and the arrival of computers have greatly changed the extent of the information that can be extracted from conductance measurements. Ref. 3 p. 222. L. Onsager Ref. 2 p. 341. lo L. Onsager Physik. Z. 1927,28 277. Electrolyte Solutions In the 1957 Discussion3 no papers specifically on conductance were presented but many mentions of the topic were made in the Discussion and the paper of Fuoss and Onsager," published in the same year and described as a 'most notable recent advance',12 served to focus attention once again on the theoretical aspects of the subject.(The equally important work of Pitts13 only received recognition at a later date.) Thereafter a steady stream of modifications and extensions to the theory flowed from the Fuoss school culminating in 1975 in the de~cription'~ of a new model to replace the primitive model of rigid non-polarizable spheres non- conducting only when in contact on which all previous conductance equations had been based. Fuoss based his model on the concept of the Gurney cosphere and ionic association was taken into account from the start rather than simply being used to account for deviations from the behaviour expected on the assumption of complete dissociation.The distance of closest approach of free ions R (or d),was not set equal to the contact distance of the two ions but was defined as the distance from the reference ion beyond which the solvent can be treated as a continuum and within which ions are treated as paired provided that unique partners can be statistically defined. The years 1976-78 have seen a flurry of contributions to the development of conductance theory with several new names emerging. Carman has reported" several theoretical studies of the electrophoretic and relaxation terms and the latter work has been summarized in a more accessible journal.'6 Most previous workers in the field have expanded appropriate functions and have expressed the conduc- tance equation in the form A = Ao-Sc* + (E~A~-E~)c In c+Jlc +J2c3 (1) where J1and J2are functions of the distance parameter and El and E2are not.Opinion is divided as to whether the same distance parameter should be used in both J1and J2 or whether because any error due to neglect of higher terms will be included in the J2term a selected fixed distance should be used in J1and the distance parameter in J2alone adjusted. Several papers in the 1977 Discussion' comment on this problem. By contrast Carman16 expresses the relative pertur- bation of the electric field MIX in terms of eleven dimensionless ratios which have been calculated and tabulated as functions of KR.Unfortunately there are relatively few points tabulated for each of these functions and consequently it is not possible to determine adequate interpolation functions with any precision and the equation can only be adapted to full computer analysis with some inaccuracy.Carman expresses his equation in the form A(~+Bc)=~[A~(~+AX/X)+A,/(~+KR)] (2) with Aeo the limiting Debye-Huckel-Onsager electrophoretic term B the Jones- Dole viscosity coefficient and y the fraction of free ions which is related to the association constant by $ (1-Y)/Y~CY= KA (3) l1 R. M. Fuoss and L. Onsager J. Phys. Chem. 1957,61,668. l2 J. E.Prue ref. 6 p. 18. l3 E.Pitts Roc. Roy. SOC.,1953 A217,43. l4 R.M.Fuoss J. Phys. Chem. 1975,79,525 1983. l5 P. C. Carman J. S. African Chem. Inst. 1975,28,80. l6 P.C. Carman and D. P. Laurie J. Solution Chem.1976,5,457. A. K. Covington and A. D. Pethybridge Carman uses the Debye-Huckel limiting law for y* (cf. ref. 17) whereas most workers include a factor (1 + KR)in the denominator of the expression for In y*. It has been shown'* that the omission of B from equation (2) has only a minor effect on the values of the best-fit parameters for the dilute aqueous solutions of MgS04 analysed by Carman,16 but changing the expression used for y+ has a much greater effect on R and KA. Moreover a slight extension of the concentration range analysed (from 1-10 to 1-40 X lop4mol dm-3) causes a very sharp decrease in the accuracy of the fit particularly with the limiting-law expression for y*.18 Carman has also published a detailed studylg of the application of his equation to existing data for a large number of 1 1 electrolytes in aqueous solution.He has concluded that for many salts there is reasonable agreement between the best-fit distance parameter obtained from his analysis and the estimated sum of the ionic radii. However for a large number of equally precise sets of data there is no such correlation. In 1975 Fuoss did not expand his new equation F75 in the form of equation (l) but expressed the relaxation and electrophoretic terms in a series of interpolating polynomials in KRwhich reproduce the explicit functions to within 0.01%. Pre-viously it had been shown that different best-fit parameters are often obtained with full and expanded [equation (l)] versions of the same conductance equation but Fuoss's approach eliminates this problem and also greatly shortens the computer time needed for analysis.After considerable correspondence in the literature Fuoss reported2' that Carman and Laurie16 were correct in their deduction that one of the boundary conditions he had imposed in deriving F75 should have given rise to a series expansion in KRrather than to the constant term he had used. Very little work had been done on comparing the F75 equation with earlier equations based on the primitive model but Mattina and FUOSS~' showed that it was as good as any for fitting some new LiI results and Pethybridge and Soltani Taba18 showed that it gave a superior fit of precise data for dilute aqueous solutions of MgS04 to that given by earlier equations despite being based on a false premise.Fuoss then changed2' the offending boundary condition from lim (fil/n2)z CO c +o to fil(R)= o where fi1 is the probability of simultaneously finding an ion i in an element of volume d Vl and an ion j in d Vzand R is the distance within which ions are treated as paired. Fuoss has now derived22 a new conductance function (which we can call F78) for the conductance of the free ions which he has expressed in the form of complex polynomials in KRas used in F75. However this change in boundary condition constitutes a slight but significant change of model and although the A. values are virtually unaltered the best-fit KAand R values are changed. Moreover for MgS04 the minimum standard deviation is greater than it was with the erroneous F75 equation but it is still lower than for equations based on the primitive model.Simultaneously with this change Fuoss has modified2' his model by dividing the l7 C. W. Davies ref. 3 p. 117. A. D. Pethybridge and S. Soltani Taba ref. 1 p. 274. l9 P. C. Carman J. Solution Chem. 1977,6 609. 2o R. M. Fuoss J. Solution Chem. 1977,6,331 (note added at end). 21 C. F. Mattina and R. M. Fuoss J. Phys. Chem. 1975 79 1604. '* R. M. Fuoss Proc. Nut. Acad. Sci. U.S.A.,1978 75 16. Electrolyte Solutions ion pairs for which a =s r sR into two classes solvent-separated or ‘diffusion’ pairs and contact pairs in accordance with the following scheme A++B- A+. . . . B-+ A+B- Stage 1 C 0 0 Stage 2 YC (1-Y)C 0 Stage 3 YC (1-a)(l- y)c a(1-y)c Long-range electrostatic interactions (for both activity coefficients and conduc- tivity) are governed by the concentration of free ions yc but Fuoss claims that some of the diffusion ion pairs can conduct a current in the field direction whereas contact ion pairs behave like dipoles and can only contribute to the charging current not to net charge transport.Fuoss then derives” his final equation by adding the contri- butions of all ions in the first two categories {total concentration [l-a(1-y)]c} and assuming that the conductances of both free ions and diffusion ions pairs are governed by the F78 expression for long-range electrostatic effects and obtains the equation of the form A = [l-a(l-y)][Ao(l+ AX/X)+(EL)] (4) in place of the expression A=(l-~)[Ao(l +AX/X)+(EL)] AX/X and (EL)are the relaxation and electrophoretic terms and are calculated in terms of the free-ion concentration yc.To the Reporters it seems highly unlikely that the mechanism and magnitude of the conductance of these diffusion ion pairs are such that it can be treated in exactly the same way as the conductance of free ions separated by distances greater than the sum of their Gurney cospheres. Fuoss has used22 equation (4) to analyse data for aqueous solutions of alkali-metal halides (except fluorides) finding that R values are almost independent of the anion but decrease with increasing atomic number of the cation which is consistent with the expected effects of these ions on water structure. The application of the Bjerrum concept of association to conductance measure- ments has been the subject of much discussion during this decade.Recent papers by Justice and Justicez3 have reviewed its application to activity-coefficient and conductance equations derived from the Debye-Hiickel and Meeron distribution functions. They have also shown that their expression can be generalized to more realistic types of interaction by introducing a term for the short-range potential between two ions such as the charged square-well model of Rasaiah and Friedman. Justice and Justicez4 show that there is quite a good correlation for alkali-metal halides (except fluorides) between the height of the square mound d+-as calculated from osmotic coefficients and from conductance measurements.Ebeling2’ emphasized the same point. Apart from a pioneering attempt by Murphy and Cohen,26all theoretical exten- sions have until recently been restricted to single symmetrical electrolytes 23 M. C. Justice and J. C. Justice Colloques internat. C.N.R.S. 1975,246,241;J. Solution Chem. 1976,5 543. 24 M. C. Justice and J. C. Justice ref. 1,p. 265. ” W.Ebeling ref. 1 p. 322. 26 T.J. Murphy and E. G. D. Cohen J. Chem. Phys. 1970,53,2173. 10 A. K. Covington and A. D. Pethybridge although they have often been used to analyse results for more complicated systems. In 1975 Quint and Viallard27 extended the Fuoss-Onsager equation to cover the problem of mixed electrolytes and recently Lee and Wheaton28a*b have shown that this approach can also be used for the analysis of conductance data for unsymmetrical electrolytes which give rise to associated but conducting ion pairs.The analysis of data for such systems is much more time-consuming and the equation is a great deal more complex than for single symmetrical electrolytes but there now exists a method for dealing rigorously with systems closer to the needs of other chemists although still only for dilute solutions. Lee and Wheaton”‘ have now embarked on the derivation of the conductance equation for the full general case where any number of ionic species of any valency may be present and paired in solution but using the Fuoss 1975 model as well as the primitive model. They have then obtained the simplified equations for the ‘special case’ of a single symmetrical electrolyte by reduction of these general equations.Unpublished preliminary calculation^^^ suggest that their latest equa- tion (with the Fuoss 1975 model) is capable of fitting precise data for aqueous solutions of such electrolytes as well as or better than any previous conductance equation most notably in the testing case of aqueous MgS04 solutions where the irritating systematic deviation between the calculated and observed conductance values at the lowest concentrations18 is totally eliminated. In one of three papers submitted for the Faraday Discussion just before his death Onsager returned to the topic of conductance. In this paper now published elsewhere Chen and Onsager3’ developed an expression for the coefficient of the c In c term in equation (1) for mixtures of strong electrolytes containing any number of ionic species of any valence type.Unfortunately they do not test the fit of their equation to any experiment$ results nor have they extended their treat- ment to the higher terms in c and c2 so their equation cannot be compared with those of Quint and Viallard27 or Lee and Wheaton.28 Chen and 0nsager3O show that for a single symmetrical electrolyte the coefficient of the c In c term should be (Elno+2E2) as earlier shown by Carman. Very few authors have paid much attention to the possible distortion introduced into their derived parameters by the method of fitting and the limits of approxima- tion and cut-off adopted therein despite the fact that the o, against R plot is frequently very shallow.Therefore a most welcome development is the recent paper by Schollmeyer and Seide131 who describe a computer program which tests these effects in a proper statistical manner even although the authors have only used a conductance function in the form of equation (1). Despite these theoretical advances it is a sad fact that a high proportion of the precise conductance results published recently are still only analysed in terms of the Fuoss-Onsager equations derived from the primitive model and if the raw data are not given it is not possible to reanalyse them. The goodness of fit may not be a great deal worse but the derived best-fit parameters have less value. 27 J. Quint and A. Viallard J. Chim. phys. 1975 72 335.28 (a)W. H. Lee and R. J. Wheaton 1.Chim. phys. 1977,74,689;(6)J. Phys. Chem. 1978,82,605.(c) J.C.S. Faraday II,in press. 29 A. D. Pethybridge unpublished work 1977. 30 M.-S. Chen and L. Onsager J. Phys. Chem. 1977,81,2017. 31 E.Schollmeyer and W. Seidell Z. phys. Chem. (Leipzig). 1976,257 1103. Electrolyte Solutions 11 An interesting paper by Janz and torn kin^^^ presents a critical assessment of current practice in conductance-cell calibration. It is clear from their analysis of methods used by most of the major groups working in the field that calibration techniques based on the interpolation functions for the conductance of KCl due to Fuoss (which in turn are based on Jones and Bradshaw standards) are superseding calibrations based solely on the Jones and Bradshaw recommended standards.Other interpolation functions derived by Justice and the Rostock school are shown to be compatible with the Fuoss functions but the latter extend up to 1mol dm-3 and can be used for cells with larger cell constants. It is also clear that calibration problems for molten-salt cells and for cells used at high temperature and pressure are greater and the precision of the standards is not nearly as good. On the experimental side most interest has been as usual in non-aqueous and mixed-solvent systems although it is perhaps surprising that despite the amount of work done over many years there should be a few alkali-metal halides for which precise conductimetric results for aqueous solutions at 25 “C have only been obtained recently uiz.NaI,33 LiI,21*33 and the alkali-metal The behaviour of sparingly soluble LiF is quite untypical the other alkali-metal halides having very small association constants whereas LiF has an association constant of about 2 dm3 mol-’ (depending on the precise equation and R value used). There is now general ag~eernent’~*’~”~ that conductance measurements should be made over as wide a range of concentration as permitted by solubility considerations and by the upper limit of KRfor the appropriate theory. It is also important that points at low concentrations ca. 1x mol dm-3 are included since A. values obtained by fitting techniques are dependent upon points at the low-concentration end. However it is comforting to know that such fitting techniques can accurately determine both A.and KA from suitable data even for solutes as weak as benzoic Particularly impressive from the experimental standpoint has been the work of Barthel and co-workers with non-aqueous systems measured over a range of temperatures. Barthel et aZ.36 analysed their results for various tetra-alkylam- monium iodides and alkali-metal salts in propanol acetonitrile and propylene carbonate with R fixed at the Bjerrum critical distance 4 = z2e2/skT,a tempera- ture-dependent quantity. It is of course arguable that a constant temperature- independent value of R or the best-fit values should have been used; the latter approach would probably lead to the same conclusions but with a higher degree of uncertainty.They found that the KA values for tetra-alkylammonium iodides showed a minimum with increasing T whereas the alkali-metal salts showed a steady increase. They suggest that the free-energy change on formation of an ion pair from separate ions can be divided into a coulombic and a residual term uiz. -RT In (KA/Kz)=AGZ = AG + AGZ (6) After calculating values of AGE from the appropriate theory they obtained values of AGk for the residual interaction term. Plots of a quantity related to AGZ ’’ G. J. Janz and R. P. T. Tomkins J. ElectrochemSOC.,1977,124,SSC. ” A. D. Pethybridge and D. J. Spiers J.C.S. Faraday I 1977,73 768. 34 J. Mallanoo and R. H. Stokes Austral. J. Chem. 1977,30 1375. ’’ L.A.Strong private communication 1977. 36 J. Barthel R. Wachter and H.-J.Goves ref. 1 p. 285. 12 A. K. Covington and A. D. Pethybridge against 1/T showed that for tetra-alkylammonium iodides AH was negative and independent of temperature whereas for the alkali-metal salts it was positive. This difference is attributed to the different character of the interactions between ions ion pairs and the surrounding solvent. Kata~ama~~ has made extensive measurements on dilute aqueous solutions of cobalt calcium cadmium and zinc sulphates in the range 045“C. He analysed these only in terms of the 1957 Fuoss-Onsager theory which must cast doubt on the precise values of the parameters obtained but quoted values of AGO AH0 and AS0 for the simple ion association process M2++ MSO,. Kata~ama~~ assumed a linear variation of log K with 1/T but closer inspection of his plots reveals that they are slightly curved.His best-fit values of the distance parameter generally increase slowly with T,in line with the variation of the Bjerrum distance. Various groups of workers have reported conductance measurements with aqueous solutions of salts containing complex ions. Pethybridge and Spiers38 report work with cis-and trans-[C~(en)~(NO~)~]’ and [Co(edta)]- salts; Masterton and Bierly3’ have published further measurements on [CO(NH~)~NO~]SO~ in water and water-dioxan mixtures and the effect of high pressure on the conductance of aqueous solutions of this salt has also been studied.,’ Unsymmetrical complex-ion conductance studies reported recently include those on hexacyanoferrate(~~),~~ hexacyanoruthenate(~~),~’hexacyanoferrate(~~~),~~ and octacyanotungstate(~)~~ and on complexes of the type [{NH2(CH2),NH2}Co](C104)3.43 The dramatic effect of the addition of small quantities of water on the A.and KA values for solutions of hydrochloric acid in hydroxylic solvents has again been demonstrated by a recent study4 with 2-methoxyethanol as solvent. A compre- hensive study of alkali-metal halides and tetra-alkylammonium bromides across the full range of solvent compositions for ethanol and water has also been rep~rted.~’ An interesting observation has been of the ability of certain hexaoxa ‘crown’ ethers to act as effective proton binding agents in aprotic solvents of low dielectric constant. Thus highly conducting solutions of toluene-p-sulphonic acid and picric acid in 1,2-dichloroethane were obtained on the addition of 18-cr0wn-6.~~ With trifluoroacetic acid in 1,2-dichloroethane triple-ion formation is postulated to account for the variation of conductance with Addition of water or lower alcohols to this system lowers the conductivity (probably owing to forma- tion of a hydrate which prevents the subsequent formation of a triple ion) whereas the conductance increases in the absence of the crown ether.Erdey-Gruz and Lengyel have reviewed the subject of proton transfer in transport processes in 37 S. Katayama J. Solution Chem. 1976,5 241. A. D. Pethybridge and D. J. Spiers J.C.S.Faraday I 1976,72 64 73. ”W. L. Masterton and T. Bierly J. Solution Chem. 1976 5,721.40 M. Ueno K. Shimizu and J. Osugi Rev. Phys. Chem. Japan 1973,43 33. 41 A. Fidler and J. Celada Coll. Czech. Chem. Comm. 1977,42 773. 42 R. J. Lemire and M. W. Lister J. Solution Chem. 1976 5 171. 43 H. Tanaka and K. Harada Electrochim. Acta 1976 21 615; ibid. 1977 22 815. 44 G. V. Merken H. P. Thun and F. Verbeek Electrochim. Acta 1976,21 11. 45 R. L. Kay and T. L. Broadwater J. Solution Chem. 1976,5 57. 46 N. Nae and J. Jagur-Grodzinski,(a)J. Amer. Chem. SOC. 1977,99 498; (b) J.C.S. Faraday I 1977 73 1951. 47 T. Erdey-Gruz and S. Lengyel in ‘Modern Aspects of Electrochemistry’ ed. J. OM. Bockris and B. E. Conway Vol. 12 1977 pp. 140. Electrolyte Solutions 13 The preoccupation of workers in this field with accounting for the concentration variation of molar conductance to the exclusion of the interpretation of limiting ionic conductances was criticized by In 1927 Bury49 expressed ‘doubt whether Stokes’s Law was applicable to bodies as small as ions’.Notwithstanding this doubt periodically repeated over 50 years no major improvement has been suggested. It is to be hoped that this problem will be tackled in the next decade. 3 Activity Hunter” in his opening sentence to a section on ‘Strong Electrolytes’ in Annual Reports for 1927 said ‘Debye-Hiickel theory continues to attract considerable attention’ but he devoted his space to results which were discrepant with the theory and concluded by remarking that the theoretical formulae were being applied sometimes with more enthusiasm than discretion.This stricture was still true in 1957. Then Bells1 noted that it was necessary to take into account solvent struc- ture i.e. to treat inner-shell water molecules individually and the rest as a continu- ous dielectric. This remains a fundamental weakness of most approaches for the dielectric constant is a bulk property with no molecular meaning. Pit~er~~ in a masterly survey has reviewed the improvements in electrolyte theory since Debye and Huckel under the headings of the selection of the model yielding the interionic potentials of mean force calculation of the radial distribution functions and from these the calculation of thermodynamic functions. He points out that speculation about net inaccuracies in the Debye-Huckel approach are best answered by comparison with calculations based on more exact theories.The Debye-Hiickel calculation of interionic distribution functions is surprisingly accurate for aqueous 1 :1electrolytes with ionic diameters of about 4 A. fried mar^^^ concluded that the peak in interest in the Debye-Huckel primitive model has been passed. New attempts to improve on the treatment by modifying the model have been reported54 and summarized by Spitzer and Bennetto.” Doubts remain whether this type of model can ever be free of worrying inconsistencies. Pitzer in an important series of has established a consistent system of equations for the thermodynamic properties of electrolytes both singly and in mixtures. Essentially the approach is an extension of the Scatchard-Harned-Robinson-Guggenheim treatment of including short-range interaction parameters.Parameters derived from data-fitting for single electrolytes can be used to predict the properties of mixed solutions of strong electrolytes. The application of the equations to evaluation of the properties of phosphoric acid5* and more importantly to sulphuric acid5’ solutions is a 48 R. H. Stokes ref. 1 p. 354. 49 C. R. Bury ref. 2 p. 330. 50 H. Hunter ref. 5 p. 22. 51 R. P. Bell in ref. 3 p. 19. ’* K. S. Pitzer Accounts Chem. Res. 1977,10 371. ” H. L. Friedman ref. 1 p.7. ” H. P. Bennetto and J. J. Spitzer J.C.S. Faraday I 1977 73 1066. 55 J. J. Spitzer and H. P. Bennetto in ‘Thermodynamic Behavior of Electrolytes in Mixed Solvents’ ed. W. F. Furter American Chemical Society Advances in Chemistry Series No.155 1976 p. 197. 56 K. S. Pitzer J. Phys. Chem. 1973,77 268. ’’ K. S. Pitzer and G. Mayorga J. Phys. Chem. 1973,77,2300. K. S. Pitzer and L. F. Silvester J. Solution Chem. 1976 5 269. 59 K. S. Pitzer R. N. Roy and L. F. Silvester J. Amer. Chem. SOC.,1977 aS,4930. 14 A. K.Covington and A. D. Pethybridge considerable advance. Iterative data-fitting techniques enable cell e.m.f. and osmotic coefficient data to be analysed and ionic interaction parameters not otherwise determinable to be obtained. A tabular survey of best values for various properties of aqueous sulphuric acid from to 6 mol kg-’ is given.” Another but more extensive analysis has been prepared by Staples.60 In a recent paper Silvester and Pitzer6’ extend their ideas to aqueous NaCl over the temperature range 0-300°C.They conclude that even at 300°C sodium chloride is still a strong electrolyte with little tendency to form ion pairs. In the 1927 Discussion Brransted,62 in an introductory paper to the second section called ‘Activity’ reviewed his 1922-23 ‘Principle of Specific Interaction of Ions’ formulated to explain his solubility measurements in mixed electrolytes. This Principle was adopted by G~ggenheim~~ in his 1935 approach to the problem of mixed electrolyte solutions and reassessed by Guggenheim and Turgeon- in 1955 on the basis of data published in the intervening years. Panckhurst and Macaskil16’ have re-examined Guggenheim’s approach recognizing that ion-solvent inter-actions must be included and have made comparisons with the equation of Harned and of Pitzer and with Stokes and Robinson’s hydration theory treatment.The Panckh~rst-Macaskill~~ treatment allows equations to be written for single-ion activity coefficients and indicates that these will not be identical even for sym- metrical electrolytes. This conclusion suggests that some conventions adopted for single-ion activity coefficients are unsuitable. This s~bject,~~,‘~ like the related problem of single electrode potentials,68 is perennial. Spedding6’ and co-workers have reviewed the osmotic coefficient data for aqueous CaCI2 often used as an isopiestic standard. This has also been done by Staples and N~ttall.~’ A bibliography of sources of experimental activity and osmotic coefficient data for multivalent electrolyte^^^ is available.Spedding and co-worker~~~*~~ have reported data on rare-earth salts. 4 Theoretical Studies Besides the application of computers for data handling and analysis already mentioned their use in theoretical studies allied to solution chemistry is a most important development since 1957. Adams and Ra~aiah~~ have made Monte Carlo calculations for ion-solvent forces and energies for a single model of a pair of 6o B. R. Staples ‘Activity and Osmotic Coefficients of Aqueous Sulfuric Acid’ N.B.S. Special Pub- lication National Bureau of Standards Washington 1978. “ L. F.Silvester and K. S. Pitzer J. Phys. Chem. 1977,81 1822. 62 J. N. Br~nsted ref. 2 p. 416. E. A.Guggenheim Phil. Mag. 1935,19,588. E. A. Guggenheim and J. C. Turgeon Trans. Faraday SOC.,1955,51,747. 6s M. H. Panckhurst and J. B. Macaskill J. Solution Chem. 1976,5,469. 66 M.Bonciocat Electrochim.Acta 1977 22 1047. ” B.D. Struck Electrochim.Acta 1977,22 1057. 68 R. Gomer and G. Tryson J. Chem. Phys. 1977,66,4413. 69 J. A. Rard A. Habenschuss and F. H. Spedding J. Chem. and Eng. Data 1977,22,180. ’O B. R. Staples and R. L. Nuttall J. Phys. Chem. Ref. Data. 1977,6 385. ” R. N. Goldberg B. R. Staples R. L. Nuttall and R. Arbuckle ‘A Bibliography of Sources of Experimental Data leading to Activity or Osmotic coefficients for Polyvalent Electrolytes in Aqueous Solution’ N.B.S. Special Publication No. 485 National Bureau of Standards Washington 1977. 72 J.A. Rard H. 0.Weber and F. H. Spedding J. Chem. and Eng. Data 1977,22 187. 73 J. A. Rard L. E. Shiers D. J. Heiser and F. H. Spedding J. Chem. and Eng. Data 1977,22 337. 74 D.J. Adams and J. C. Rasaiah ref. 1 p. 22. Electrolyte Solutions 15 oppositely charged ions in a polar solvent. The contour maps obtained show that as with an isolated ion there is only a single well defined oriented layer of solvent molecules about the pair of ions with only a small disturbance of solvent structure beyond this first layer. Heinzinger and his group in the latest of a series of papers,7s have done a molecular dynamics simulation on a system containing 200 water molecules with eight Na+ and eight Cl- ions and compared the results obtained for the radial pair-correlation function with that from X-ray scattering.Whereas the average ion-water and water-water nearest-neighbour distances coincide there are some differences between the results. A second important attempt to correlate information from two sources is that of Friedman Zebolsky and Kalman76 who have compared pair correlation functions derivable from X-ray-scattering experiments on aqueous tetraphenylarsonium chloride solutions with those calculable from models which fit osmotic coefficient data for this salt. One of the first papers on the application of X-ray scattering to ionic solutions was presented at the 1957 General Discu~sion.~ Both this technique and neutron scattering are developing in importance although neither is without some contro- versial interpretative points or definitive without the other or in combination with a good theory for obtaining partial structure functions.77 A paper on aqueous NiC12 solutions from the Sardinian group was presented at the 1977 Discussion78 and confirms that the co-ordination geometry is octahedral with the C1- ions sur- rounded by their own hydration shells.No evidence was found for quasi-lattice structure an interpretation favoured by some other workers in this field. A very recent paper7' reports a theoretical treatment of concentrated solutions on the basis of further reassessment of the lattice-like model. The capacity and speed of modern high-powered computers now permits the application of ab initio molecular orbital calculation methods to molecules and ions of interest to the solution chemist.Radom" has calculated the equilibrium geometry of many polynuclear anions such as BF4- NO3- NO2- C032- and acetate. It is still necessary to contract the basis orbital set to make the problem tractable. Schuster et d8'have surveyed progress by this method in calculating the equilibrium geometry between water molecules and simple ions. It is now possible to confirm by calculation that water is oppositely oriented about a cation and an anion by virtue of its dipole a fact that was pointed out by LowryS2 as a consequence of Fowler's picture of water at the 1927 Discussion. Sadlej and Sadleja3 have taken such calculations a step further by using ab initio SCF LCAO molecular orbital approaches to calculate the force constants and polarizabilities and hence the change in frequencies and intensities to be expected in the infrared and Raman spectra of the symmetrical OH-stretch and LiOH-bending vibrations '' G.Palinkas W. 0.Reide and K. Heinzinger,Z. Naturforsch. 1977 32n. 1137. 76 H.L.Friedman D. M. Zebolsky and E. Kalman J. Solution Chem. 1976,5 853. 77 R. Triolo 'Recent Applications of X-ray and Neutron Diffraction to Electrolyte Solutions' Abstract of paper presented at One-day Symposium University of Reading Sept. 12th. 1977 to be published in J. Solution Chern. 78 R. Caminiti G. Licheri G. Piccaluga and G. Pinna. ref. 1 p. 62. 79 I. Ruff J.C.S. Faraday II 1977 73 1858. L. Radom Austral. J. Chem. 1976,29 1635. P. Schuster W.Jakubetz and W. Marius Topics Current Chem.1975,60 1. 82 T. M. Lowry ref. 2 p. 541. 83 J. Sadlej and A. J. Sadlej. ref. 1 p. 112. 16 A. K. Covington and A. D. Pethybridge due to interaction of Li' with a single water molecule. It will be necessary to extend such calculations to take into account hydrogen-bonding to other water molecules and perhaps the effect of negative ions before comparison with observed spectra becomes realistic. Nevertheless this is an important advance which highlights the inadequacy of the simple H\O. -M' interaction picture which is much favoured H' by spectroscopists. It should be possible to extend such calculations to the models used to interpret the temperature dependence of the 'H n.m.r. chemical shifts in water induced by various ions.84 One awaits with some interest the analysis of the interaction of water molecules with the polynuclear anions considered by Radom but such studies will be very expensive of computer time.5 Water In the 1927 meeting Harried" reported on some cell e.m.f. studies of the ionic product of water (Kw). A recent careful studys6 over the temperature range 273-328K had as its aim investigation of the so far unexplainable discrepancy between the enthalpy of ionization of water AH* derived from cells and the directly determined calorimetric values. Extrapolation to zero ionic strength for cells containing various salts was achieved using Pitzer's equation^.^^*^^ The dis- crepancy in AH* now stands at 90 cal mol-1 and pKw(298)= 14.004. A second study reportss7 values for the ionic product of DzO over the extended temperature range 298-523 K from cells with liquid junction.The results are in close agreement with a previous modern study which reached only to 323K. Born-theory predictions of very high solvation energies for ions at 573K have been verified by Cobble and Murray" from heat of solution measurements in a pres- surized calorimeter. The papers presented at a symposium on high-temperature and -pressure studies held in Guildford in 1973 have belatedly been p~blished.'~ Three i.r. studies of water modes deserve mention. Gigukre and Turrell" have obtained improved spectra in the region 8004000 cm-' for the aqueous hydro- halide acids up to saturated solutions. These confirm their previous assignments which had been challenged.They conclude that there is no evidence for species such as Hs02+ and that H30+ has a much longer lifetime in strong acids than in water. Paquette and Jolicoeurgl have obtained differential i.r. spectra of the effect of various ions and non-electrolytes on the 1200 and 1450 cm-' bands in H20 and HOD-D,O respectively. The spectra expressed as structural temperature shifts for each ion can be interpreted in terms of changes in equilibrium populations of two categories of OH oscillators which differ by an enthalpy of 2-3 kcal mol-l. Schioberg and discuss in terms of a mixture model for water the i.r. 84 J. W. Akitt ref. 1 p..102. " H. S. Harned ref. 2 p. 462. 86 A. K. Covington M. I. A. Ferra and R. A. Robinson J.C.S. Faraday I 1977,73 1721.D. W. Shoesmith and W. Lee Canad. J. Chem. 1976,74,3553. J. W. Cobble and R. C. Murray ref. 1 p. 144. 89 High Temperature High Pressure Electrochemistry in Aqueous Solutions ed. R. W. Staehle D. de G. Jones and J. E. Slater International Corrosion Conference Series Vol. 4 National Association of Corrosion Engineers Washington D.C. 1976. 90 P. A. Gigutre and S. Turrell Canad. J. Chem. 1976 54 3477. 91 J. Paquette and (3. Jolicoeur J. Sohtion Chem. 1977,6 403. 92 D. Schioberg and W. Luck Spectroscopy Letters 1977,10,613. Electrolyte Solutions 17 spectra in the OH-stretching (-3700 cm-') and bending (-1600 cm-') regions for 1:1 complexes of water with bases of increasing acceptor strength (acetonitrile dimethyl sulphoxide and triethylamine) decoupled in carbon tetrachloride.6 Solvation and Related Topics Last year's Annual Reports contained a section devoted to spectroscopic studies of solvation.' We shall therefore be content with mentioning some of the more important studies that have appeared over the past year. 133Cs and 39K n.m.r. shifts have been measured for various salts in different solvents and Popov and co- worker~~~ report that infinite-dilution shifts correlate with Gutmann's donor numbers (DN) for many solvents with the exception in the case of '33Cs,of water methanol acetonitrile and dimethyl sulphoxide. All these solvents are known to form complexes with water which may be taken as a measure of their peculiarity. The extent of ion-pair formation derived for some caesium salts was noted to depend not only on the dielectric constant but also on DN.The same has interpreted the downfield shift in the 133Cs resonance which occurs when 18- crown-6 is added to solutions of caesium tetraphenylborate in pyridine changing to an upfield shift after pairing 1:1 in Cs :crown as due to the formation of a 1:1 complex followed by a 2 :1 sandwich Cs :crown complex with K1=lo6 and KZ = 44 f2 at 298 K. Kesler" and co-workers provide additional evidence that donor- acceptor interaction makes an important contribution to the mechanism of quadrupole relaxation of ionic nuclei which was challenged by Hertz.96 The spin-lattice correlation time (l/Tl) for Cs' (as for Li' and Na') was found to correlate with DN. Hertz and his group have continued their studies of first-hydration-sphere inter- actions by proton9' and other nuclei9' relaxation times.Hertz9' clearly does not believe in the usually accepted Frank and Wen model for the first sphere. Doubts about the appropriateness of this model for anions have also been expressed by Symons and co-workerslOO from low-temperature i.r. studies. Secondary solvation is viewed not as an anion-induced phenomenon but as a natural result of the desire of the solvent to form hydrogen bonds. Lowered temperatures allow exchange to be slowed and new features to be observed in 'H n.m.r. spectra"' and the same is true to a lesser extent in i.r. spectra. In anhydrous methanol in CH30D at 148 K features due to cation-solvent anion-solvent and solvent-separated ion pairs can be discerned"' for lithium iodide in the OH-stretching region.Generally perchlorate ions interact only weakly with most cations which makes study difficult but a of 35Cl longitudinal and transverse relaxation time 93 (a)W. J. Witte L. Liu E. Mei J. L. Dye and A. I. Popov J. Solution Chem. 1977,6 337; (b)J. S. Shih and A. I. Popov Inorg. Nuclear Chem. Letters 1977 13 105. 94 E. Mei J. L. Dye and A. I. Popov J. Amer. Chem. Soc. 1977,99 5308. 9s Yu. M. Kessler A. 1. Mishustin and A. I. Podkovyrin J. Solution Chem. 1977,6 111. 96 H. G. Hertz J. Solution Chem. 1975,4 790. 97 H. Langer and H. G. Hertz Ber. Bunsengesellschaft phys. Chem. 1977,81,468. 98 M. Contreras and H. G. Hertz ref. 1 p. 33. 99 H. G. Hertz ref.1 p.77. 100 S. E. Jackson I. M. Strauss and M. C. R. Symons J.C.S. Chem. Comm. 1977 174. lo' Y. Ruben and J. Reuben J. Phys. Chem. 1976,80,2394. lo' I. M. Strauss and M. C. R. Symons J.C.S. Faraday Z 1977,73 1796. lo3 P. Reimarsson and B. Lindman Znorg. Nuclear Chem. Letters 1977 13 149. 18 A. K. Covington and A. D. Pethybridge measurements suggests that this is a promising method. Ions affect Tl in the order Fe3' >Ba" >Ca" >Mg2+>Li' >Na+>NH,+ >H+ but a full analysis of the effect of cation concentration appears to be a daunting prospect. A new non-spectroscopic method for determining ionic solvation numbers has been ~uggested"~ based on gas solubility. The method involves hexamethyl- phosphotriamide (HMPT) which has special cation-solvating properties.HMPT bonds to the acidic proton of methylacetylene (MA) through its oxygen but this oxygen is used to co-ordinate a cation. Thus it is suggested that there should be a relationship between the Henry's Law constant for MA in the absence (h) and presence (h') of a cation with a solvation number n. If the concentration of M is small then -=I-n-[M+l h h' [HMPT] (7) For NaClO, n = 3 from a straight-line plot of equation (7). For the other alkali-metal perchlorates curves are found attributed to ion-pair formation but limiting slopes suggest n =4 for Li' and K+ and n = 5 for Cs'. Ion-pairing in KClO with an association constant of 0.12 mol-' dm-3 which would be too small to be derivable from conductance studies was confirmed by an e.s.r.technique. The determination of an association constant for CuSO ion pairs from U.V. spectroscopy was one of the now classical papers'os of the 1957Discussion. A new paperlo6 reports the characterization of inner- and outer-sphere complex formation in the first-row transition-metal sulphates by absorption spectra and ther-modynamic measurements. Most are predominantly outer-sphere except FeS04'. Fox and Hayon rep~rt~~~*'~* far-u.v. spectra for I- Br- and OH-. The analysis depends on computerized curve resolution techniques which is also true of the interpretation of the Raman spectra of alkali-metal salts in liquid ammonia.'o9 Following pioneering work in the late 1940's on nitrates and sulphates now made easier by laser excitation Raman spectroscopy continues to be a powerful and popular method.Irish and Jarv"' report on the diffuse low-frequency lines in the 350-500 cm-' region assigned to symmetric stretching vibrations of aquated cations. James and Frost'" have analysed polarized Raman spectra of aqueous metal nitrate solutions to yield vibrational relaxational and rotational reorien- tational times for the nitrate ion. Their interpretation appeared to generate some controversy. In mixed solvents the ions may be preferentially solvated by one or other of the solvent components. This may have a profound effect on the properties of the solution and has important implications for certain technological processes such as salt effects on liquid-vapour equilibrium,"' electrolysis and chemical kinetics.The suggestion that curvature of the iodide c.t.t.s. spectral shifts in a binary mixed solvent could be used to evaluate the extent of preferential solvation was made by 104 W. Martir A. E. Alegria and G. R. Stevenson J. Amer. Chem. SOC.,1976,98 7955. W. G. Davies R. J. Otter and J. E. Prue ref. 3 p. 103. '06 K. G. Ashurst and R. D. Handck J.C.S. Dalton 1977 1701. lo' M. Fox and E. Hayon J.C.S. Faraday I 1977,73 1003,872. lo8 M. Fox R. McIntyre and E. Hayon ref. 1 p. 167. lo' P. Gans and J. B. Gill ref. 1 p. 150. 'lo D. E. Irish and T. Jarv ref. 1 p. 95. D. James and R. L. Frost ref. 1 p. 48. '12 W. F. Furter Canad.J. Chem. Eng.,1977,55 229. Electrolyte Solutions 19 Smith and Symon~~~~ in the 1957 Discussion.This and its implications for kinetics were followed up by Langford who has reviewed his recent work.'14 The inter- pretation of U.V. spectral shifts in this way is on less sound theoretical grounds than that of n.m.r. chemical shifts for solute ions. Covington and Newman115 have reviewed and extended their treatment achieving an important link with cell- e.m.f .-derived free energy of transfer measurements. They found methanol-water one of the hardest solvent systems to understand but concluded that Na+ Rb+ Cs' F- and C1- were all preferentially hydrated. Hertz and co-worker~~~~*"' have studied n.m.r. relaxation rates for protons and for 35Cl 81Br 12'1 23Na and 87Rb ions in methanol-water and agree that values extrapolated to zero salt concentration support the conclusion from shift studies that the cations are preferentially hydra- ted.However they conclude"' that C1- and Br- are preferentially solvated by methanol and that I-shows no preference. It is important that this discrepancy should be cleared up. The quantitative treatment of n.m.r. shifts in mixed solvents has close parallels with the mathematical treatment of the formation of weak molecular complexes by the Benesi-Hildebrand equation. 0. B. Nagy and co- workers118*'l9 have described a competitive preferential solvation treatment of molecular association which has algebraic similarities to Covington's treatment of n.m.r. shift data.115 N.m.r. chemical shifts have often been employed to evaluate pair association in non-electrolyte mixtures.Hertz's group120*121 have also applied n.m.r. relaxation time (Tl) measurements to two such systems NN-dimethyl- formamide-benzene and acetone-chloroform. Again there is not complete agreement with conclusions from chemical shift measurements. Studies of non- electrolyte mixtures have great relevance to understanding the more difficult pro- blem of salt solutions in such mixtures as mixed solvents. It is a noticeable trend that some groups are making separate studies of such solvent systems,122 sometimes even abandoning electrolyte studies. 123 Parker has reviewed'24 his work on the solvation of ions as studied by enthalpies entropies and free energies of transfer and has applied12' his ideas to problems of electrowinning and refining.Organic nitriles stabilize Cu' with the result that the electrode potentials E(Cu' I Cu) and E(Cu2' 1 Cu') for aqueous solutions differ from those for aqueous solutions to which some acetonitrile or 3-hydroxypropanonitrile has been added owing to preferential solvation of Cu' by nitrile and Cu2' by water. Electrolysis in a mixed solvent in an inert atmosphere thus provides an alternative possibility for copper winning and refining. '13 M. Smith and M. C. R. Symons ref. 3 p. 206. 11' C. H. Langford and J. P. K. Tong Pure Appl. Chem. 1977 49 93 [also available in "on-Aqueous Solutions-5 (Leeds 1976)' ed. J. B. Gill Pergamon Oxford 1977 p. 931. 11' A. K. Covington andK. E. Newman in ref. 55 p. 153. 11' D. S. Gill H. G. Hertz and R. Tutsch J.C.S. Faraday I 1976,72 1559.11' M. Holz H. Weingartner and H. G. Hertz J.C.S. Faraday I 1977,73 71. M. wa Muanda J. B. Nagy and 0.B. Nagy Tetrahedron Letters 1974 3421. 119 0.B. Nagy and J. B. Nagy in 'Environmental Effects on Molecular Structure and Properties' ed. B. Pullman Reidel Dordrecht 1976 p. 179. ''O H. G. Hertz B. Kwatra and R. Tutsch Z. phys. Chem. (Frankfurt) 1976 103 259. A. L. Capparelli H. G. Hertz B. Kwatra and R. Tutsch 2.phys. Chem. (Frankfurt) 1976,103 279. 122 C. de Visser G. Perron J. E. Desnoyers W. M. Heuvelsland and G. Somsen J. Chem. and Eng. Data 1977,22 74. lZ3 R. H. Stokes J.C.S.Furuday I 1977,73 1140. lZ4 A. J. Parker Electrochim. Acta 1976 21 671. ''' I. D. Macleod D. M. Muir A. J. Parker and P. Singh Austral. J. Chem. 1977 30 1423.20 A. K. Covington and A. D. Pethybridge The development of precision calorimeters (such as LKB) microflow calori- meters (Picker) and densitometers (Anton Paar Sodev) have greatly facilitated the making of enthalpy heat capacity and density measurements and a great deal of useful data is now being collected and correlated. 126~127 Desrosiers and Desnoyers12* have applied scaled particle theory (SPT) (for a recent review see Pierotti12') to their results for the transfer of tetrabutylammonium ion from water to aqueous mixed solvents involving urea or t-butyl alcohol. They conclude that rather surprisingly using only cavity terms SPT can be used to predict signs magnitudes and trends in the results. Treiner13' has used SPT to examine the validity of the large ion assumption.He concludes that the free energy of transfer to water of the tetraphenylborate ion can be predicted to within 20% from methanol ethanol acetonitrile and formamide. It was unsuccessful for propylene carbonate dimethyl sulphoxide and sulpholan where strong solute-solvent inter- actions occur. The theory of hydrophobic interaction originates from the anomalously large loss of entropy which accompanies dissolution of polar substances in water. This leads to the conclusion that partitioning phenomena which are so important in biological systems result almost entirely from repulsive interactions between water and non-polar substances. Cramer131 has drawn attention to discrepancies between the existing theory of hydrophobic interaction and experimental data.By comparing the free-energy changes for the processes (a) aqueous solution + vapour (b) octanol solution + vapour and (c) aqueous solution + octanol solution for various solutes and variation with their molar volumes he concludes that the simple features exhibited by (c) must arise by compensation of some complex interaction of the solute with water and octanol. He suggests that SPT would be a better explanation of the effects. Heuvelsland and S~msen'~~*'~~ have selected tetra-n- butylammonium bromide in NN-dimethylformamide-water to investigate the hydrophobic hydration effect from measurements of volume enthalpy and heat capacity changes on solution. From a co-operative hydration model the enthalpy changes were interpreted in terms of two parameters viz.the number of water molecules surrounding an alkyl group and the enthalpic effect due to hydrophobic hydration in pure water. A minimum in the apparent molar volume at-infinite dilution in water-rich solvents was attributed to hydrophobic interactions. Considerable understanding of the complex interactions which take place between solvent molecules and between solvent molecules and ions in solution may arise from the excellent gas-phase studies of Kebarle.'34.'35 lZ6 Y.-S.Choi and C. M. Criss ref. 1 p. 204. 12' C. de Visser G. Perron and J. E. Desnoyers J. Amer. Chem. SOC. 1977,99,5894. lZ8N. Desrosiers and J. E. Desnoyers Canad. J. Chem. 1976,54 3800. lZ9 R. A. Pierotti Chem. Rev. 1976,76 717. C.Treiner J.C.S. Faraday I 1976,72,2007. lJ1 R. D. Cramer J. Amer. Chem. Soc. 1977,99,5408. 132 W. J. M. Heuvelsland and G. Somsen J. Chem. Thermodynamics 1976,8 873. lJ3 W. J. M. Heuvelsland and G. Somsen J. Chem. Thermodynamics 1977 9 231. lJ4 P. Kebarle Ann. Rev. Phys. Chem. 1977 28 445. 135 P. Kebarle W. R. Davidson M. French J. B. Cumming and T. B. McMahon ref. 1 p. 220. Electrolyte Solutions 7 Concluding Remarks It remains to attempt to consider the question whether much progress has been made since 1927 and 1957. The answer must surely be that ‘rapid progress continues in this field’,’36 particularly in the application of new techniques n.m.r. probably being the most important and widely ap~licab1e.l~~ Anyone outside the field who maintains that solution chemistry remains in the Debye-Huckel era is uninformed.Yet he is not to be blamed since that excellent best-selling textbook by Robinson and Stokes13* remains essentially unrevised since its second edition in 1965 and a more recent two-volume text on electro~hemistry’~~ contains only five pages devoted to a review of spectroscopic methods and with no examples. Of recent books only that of Gordon,14* aimed at physical organic chemists attempts an up-dated picture. Firstly we consider the solution chemist has a responsibility to tidy-up the sloppiness of expression that still exists and secondly there is a need for more communication between technique-oriented specialists for it is through the cor- relation of results from the application of different approaches that advances will come.We cannot agree with last year’s Reporter7 that attempts to reproduce spectroscopic data quantitatively are premature for surely the subject is entering an exciting phase. Onward then to 1987 when there will be a need for the fourth Faraday Discussion! 13‘ J. E. Coates and J. A. V. Butler ref. 4 p. 21. 13’ A. K. Covington and K. E. Newman in ref. 47 p. 41. 13* R. A. Robinson and R. H. Stokes ‘Electrolyte Solutions’ 2nd. Edn. Butterworths London,1965. 139 J. O’M. Bockris and A. K. N. Reddy ‘Modern Electrochemistry’ Vol. 1 Macdonald London 1970. 140 J. E. Gordon ‘The Organic Chemistry of Electrolyte Solutions’ Wiley-Interscience New York 1975.

ISSN:0308-6003    

DOI:10.1039/PR9777400005

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

3 Solid Electrolytes By M. D. INGRAM Department of Chemistry University of Aberdeen Aberdeen Scotland and C. A. VINCENT Department of Chemistry University of St. Andrews St. Andrews Scotland. 1 Introduction 'Solid electrolytes' are phases which have an electric conductance wholly or partly due to ionic motion within a solid lattice. While such substances have been known for over a century there has been a great renewal of interest in them recently. Some ten years ago on reviewing the field of solid state electrochemistry Raleigh' pointed out that as far as practical applications were concerned 'the main drawback is the high internal resistance associated with the low specific conductivity of these materials under many conditions of interest'. Within a year the classic papers of Bradley and Greene' and of Owens and Argue3 had described solid electrolytes of the form MAg& with conductances at ambient temperatures of >10 S m-' the same value as the conductance of a molar KCl aqueous solution.That same year Yao and Kummer4 showed that a single crystal of p -alumina exhibited rapid diffusion of univalent cations in the plane perpendicular to the c-axis. They reported the diffusion coefficient of sodium ions to be 1x m2s-' at 300 "C a value comparable with that of the same ions in molten NaN03. A number of structural features have been found to characterize the solids which have rapid ionic transport and to distinguish them from the more usual ionic crystals. Typically their structures are not close-packed but contain networks of 'ion-sized' passageways consisting of interconnected polyhedra of the fixed ions through which selected mobile ions may move.In general the number of sites available for the mobile ions is much larger than the number of mobile ions themselves hence the solid has a highly disordered structure. The high conductivity is brought about by a combination of (i) a high concentration of mobile particles (ii) a low enthalpy of activation for ion motion from site to site. Many authors have referred to such highly conducting solids as superionic conductors. We regard this terminology as inappropriate since a connection is implied with electronic superconductors (where none exists) and in addition ionic conductivity in solids can vary over a very wide range of values.In this Report for ' D. 0.Raleigh Progr. Solid-state Chem. 1966,3 83. ' J. N.Bradley and P. D. Greene Trans. Faraday SOC.,1967,63 2516. B. B. Owens and G. R. Argue Science 1967,157,308. Y.-F. Yao and J. T. Kummer J. Inorg. Nuclear Chem. 1967,29,2453. 23 M. D. Ingram and C. A. Vincent simplicity we shall refer to solids which exhibit high ionic conductance as solid electrolytes or optimized ionic conductors using the terminology of van Gool.’ During the past decade there has been a rapid development in the field of solid electrolytes spurred on on the one hand by the aim of producing better batteries and fuel cells and on the other by attempts to understand the mechanism of ‘fast’ ionic motion in the solid state.Reports of NATO Advanced Study Institutes held in 1972 and 1975 gave an impetus to the study of transport’ and electrode processes6 and a number of reviews’-’’ and two books 12*13have outlined developments up to 1976. This Report concentrates on a number of the more interesting advances in the period 1976-77. We have excluded from our review the defect stabilized ceramic oxide ion conductors which are electrochemically important only at rather higher temperatures and have instead concentrated on those materials which are highly conducting in a more accessible temperature range. 2 Theory Two principal questions in the field of solid electrolytes are ‘why are certain materials such good ionic conductors?’ and ‘what is the detailed mechanism of ion transport?’.Holzapfel and Rickertg have pointed out that the common property of all the best solid electrolytes is that they are optimized both with respect to ion mobility (i.e.they are indeed fast ion conductors) and to the density of mobile ions. This latter requirement is generally more restrictive and the main problem is to identify crystal structures where the mobile ion forms one of the major constituents of the material. Structure.-The first necessary condition for generalized ionic motion is that there must be more available sites (of almost the same energy) than there are mobile ions to fill them. One of the main advances in understanding has therefore stemmed from crystallographic studies notably by Geller and co-workers (see for example refs.5 12 and 13) which have convincingly demonstrated the importance of a detailed knowledge of structure. Such work has not only indicated the types and numbers of sites available to the diffusing ions but has emphasised the significance of the spatial relationship between sites. Ionic mobility is enhanced when the ion sites form a network of channels through which the ions can move and the ‘simplicity’ of these passageways has an important influence on the value of the conductivity. Another useful parameter made available by crystallographic studies is the mean distribution of mobile ions over sets of non-equivalent lattice sites. Recent structural investigations are referred to in the sections on particular elec- trolyte types given below.‘Fast ion transport in solids’ ed. W. van Gool North-Holland/American Elsevier Amsterdam 1973. ‘Electrode processes in solid state ionics’ ed. M. Kleitz and J. Dupuy D. Reidel/Dordrecht-Holland Boston 1976. R. A. Huggins in ‘Diffusion in solids’ ed. A. S. Nowick and J. J. Burton Academic Press New York 1975 p. 445. K. Funke Progr. Solid-state Chem. 1976 11 345. G. HoMpfel and H. Rickert Natunuiss. 1977 64 53. lo P. McGeehin and A. Hooper J. Materials Sci. 1977 12 1. M. Lazzari and B. Scrosati J. Power Sources 1977,1 333. l2 ‘Superionic conductors’ ed. G. D. Mahan and W. L. Roth Plenum Press New York 1976. l3 ‘Topics in Applied Physics’ ed. S. Geller Springer-Verlag Heidelberg/New York 1977 vol. 21. Solid Electrolytes Models of Ionic Motion.-The second necessary condition for a material to exhibit high ionic conductivity is that the mobile ion has a high diffusion coefficient.Development of models of ionic motion may be considered to have taken place in three stages. Of the early transport mechanisms suggested (apart from the free-ion model of Rice and Roth) most tried to extend the simple hopping model from its successful application in defect conductivity mechanisms to systems where the mobile ion concentration was high. In the second group of theories recognition was given to what O’Keeffe’ has termed co-operative motion where the influence of mobile ions on one another was taken into consideration. Finally for systems in which spectroscopic studies had shown that the ion flight time between sites was of the same order of magnitude as the dwell time motion was characterized in terms of quasi-liquid sublattice behaviour.Simple Hopping Motion.-In this representation the solid is considered as a simple vibrational system and an Eyring approach to the rate of motion of say univalent cations between sites leads to an expression for the jump frequency v+ of the form (e.g.ref. 5) U. v+ = ni { *}YO exp {-EO-$h[vO-C (ut -ui)]}/kT vi 1 where yo is the vibrational frequency of the jumping particle in the direction of the barrier vi and v are other vibrational frequencies (t refers to those of the transition state) and Eo is the height of the energy barrier. Depending on the exact model used i.e. on the partition functions chosen to describe the initial and transition states such equations can be expressed in the form v+ = vo exp (-got/kT) For random motion in an isotropic crystal and ignoring correlation factors the diffusion coefficient is given by d+=kva,* where a+ is the cation mean jump distance.Finally from the Nernst relationship the cation conductivity u+is given by u+(oru+T)=pN+ exp (-got/kT) where p contains a number of constants (different for u+and u+T)and N+ is the concentration of mobile ions. The height of the activation potential barrier is recognised to be the major factor influencing the rate of movement of an ion. However the gradient of an Arrhenius type plot will only give a mean value of this parameter if the ions require no additional energy to excite them from a ‘normal’ lattice site to a ‘defect’ or high energy site in a diffusion pathway.Indeed while the majority of experimental results give linear Arrhenius plots unique and meaningful interpretation of the slopes is very dependent on the precise model used-for instance whether there are one two or a continuous distribution of possible energy states. Flygare and Huggins developed an explicitly atomistic model for the trans- port of ions through crystallographic channels using a-AgI as an example. Recently a similar theory was applied14 to the motion of interstitial alkali metal ions within a rutile structure. The basis of this calculation was the determination of an energy profile for a single interstitial ion as a function of position in the host lattice.This l4 0.B. Ajayi L. E. Nagel J. D. Raistrick and R. A. Huggins J. Phys. Chem. Solids 1976 37,167. M. D. Ingram and C. A. Vincent was done by calculating the total electrostatic interaction energy of the mobile ion and the surrounding lattice ions as a sum of the Coulomb dipole polarization monopole-permanent dipole and overlap repulsion forces. The authors point out that motion of ions between inequivalent positions within a particular group of sites occurs with much smaller activation energy than is the case for normal site-to-site jumping. However the model does not take into consideration important factors such as relaxation of the fixed lattice around a mobile ion and interactions between mobile species. Nevertheless a realistic estimate for the activation energy of Li’ ion motion in Ti02 was made (just over a factor of two too large).An attempt to extend the hopping model to take into account ion-lattice interactions was made by Mahan and Pardee who applied small polaron theory to the ionic diffusion. However there is some disagreement on the choice of coupling constant suggested.” Returning to the simple hopping model another criticism is that the pre- exponential factors evaluated for highly conducting systems are usually difficult to reconcile with theory. Phillips” has suggested that since local potentials are rather flat along certain lines which connect neighbouring sites normal energy barriers as generally used in this model may not be appropriate. Co-operative Motion Models.-In general it may be concluded that the simple hopping model is not a particularly good description of solid electrolytes which have a high density of mobile ions.In such conditions it is necessary to recognise that situations must arise where mobile ions will interact influencing each other’s motion and where the movement of an ion into a particular site will depend on the occupation of neighbouring sites. Thus diffusion of ions in systems such as the AgI-based electrolytes and p -alumina must be regarded as highly correlated pro- cesses. Here changes in local potentials due to the motion of an ion are comparable in magnitude to the energy needed for a diffusion step. In the extreme case a form of ‘caterpillar motion’ may occur where an ion jumps into an occupied site thereby inducing the occupying ion to perform a jump in the same direction.* Information on the conductivity correlation factor fI is given by measurements of the Haven ratio D*/Dmwhich is the ratio between the tracer diffusion coefficient and that based on conductance measurements (by way of the Nernst relationship).While this parameter may be more correctly equated to h./fI,where fi is the diffusion correlation factor the latter term is not likely to differ much from unity in systems such as MAg& (where there are many vacant sites) and hence D*/D = l/fI is often taken to be a reasonable approximation. Sat0 and Kikuchi proposed a path probability method to deal with the problem of co-operative motion and have recently discussed this f~rther.’~.~~ The theory is essentially a modification of the random walk approach in which the influence of mobile particles on one another is introduced and the effect of this on the cor- relation factor is determined.It is shown that provided long range ordering is not caused repulsive interactions enhance diffusion while attractive interactions suppress it. The method is specifically designed to describe highly co-operative processes occurring in a lattice structure but further developments are required to include correlation of atom positions past the nearest-neighbour type before all types of co-operative forces can be included. The theory has been applied with ’’ J. C. Phillips J. Electrochem.Soc. 1976,123,934. Solid Electrolytes some success to p-alumina systems where a correlated hopping mechanism seems appropriate.However as will be discussed below there is experimental evidence to show that such behaviour which implies a long dwell time compared with the time of flight is not characteristic of conductors based on silver iodide. The domain theory considers that mobile ions occupy specific sets of related sites within the excess number of sites available. The basic idea is that local situations involved in the diffusion mechanism interact with one another-i.e. the model assumes that the diffusion mechanism is spatially co-operative but not necessarily co-operative in time.5 Local configurations would be likely to differ from the time- and space-averaged structure derived from X-ray diffraction studies.The question of the probability of domain formation is still unresolved as it would appear that the establishment of a charged domain wall would require a considerable amount of energy and the associated enthalpy increase and entropy decrease would have to be compensated in some way. Until recently there was no direct evidence for domain structure in solid electrolytes. Thermal evidence for microdomain forma- tion in a-Ag2S has been discussed by Phillips16 who also proposed that proof of the existence of such features might be obtainable from electron diffraction studies. However confirmation has been obtained in the case of p -eucryptite17 where diffuse X-ray scattering has shown that large one-dimensional ordered regions are conserved in the lithium channels (see below) up to temperatures of 600 "C.Quasi-liquid Sublattice Models.-In 1972 O'Keeff e5 discussed the likelihood that certain diffusion rates might be directly determined by phonon lifetimes. The question then arises of how liquid-like is the behaviour of the mobile ions. In free diffusion (e.g. in gases) the majority of time is spent in flight with 'instantaneous' collisions producing random changes in the velocity vectors. Diffusion in many liquid systems is intermediate between this and hopping motion. An itinerant oscillator model applicable to liquids has been applied with some success to a-AgI at higher temperatures.' The main requirement of experimental methods capable of resolving such questions is that they should yield information on the same time scale as that of the basic steps of the ionic motion.The application of such techniques which include microwave far-i.r. and Raman spectroscopy and the quasielastic scattering of cold neutrons by himself and other workers to AgI-based solid electrolytes has been reviewed by Funke.' The conclusion of such studies is that the silver ion motion may be best characterized by a superimposition of jump diffusion on a local motion of large amplitude and strong damping. The mean time of flight appears to be around ten times larger than might be expected if the ions moved directly from one point to another. Thus local motion of the silver ion occurs not only with the crystal voids (such as the ellipsoidal regions centred at the tetrahedral sites of a-AgI) but also during flight from one void to another as interactions with its moving neighbours cause fluctuations on a time scale of s (see Figure 1).Quantitative agreement has been shown by Funke and co-workers between such a model and the response of a number of AgI-based electrolytes in the microwave region and the model is also consistent with results of quasielastic neutron scattering experiments.l6 J. C. Phillips Electrochim.Actu 1977 22 709. l7 (a) U. von Alpen H. Schulz G. H. Talat and H. Bohm Solid Srate Comm. 1977,23 911; (6)U. von Alpen E. Schonherr H. Schulz and G. H. Talat Electrochim Acru 1977,22 805. M. D. Ingram and C. A. Vincent Simple diffusion Ag'in solid electrolyte Jump diffusion < -0.5nm > < -0.5 rm > Residence time =O Time of flight t residence Time of flight (8 residence time time Figure 1 Diffusion pathways in different systems (after ref.8). Briiesch Pietronero Strassler and Zeller18 have developed a mathematical model for the complex conductivity of ionic conductors of this type and have compared the theoretical behaviour with experimental far-i.r. data. The model is based on a Langevin equation with the potential given by a memory function of the form exp [ -(t -?')/TI where T is the transition time from oscillatory to diffusive behaviour. Good agreement was obtained for far4.r. response but Funke has pointed out that the theory cannot easily predict the characteristic dispersion of such systems in the microwave range.Huberman and Martinlg have shown that an analysis of light scattering due to low frequency fluctuations in highly conducting solid electrolytes would reveal specific information on local site symmetries and flight times. Strong damping due to coupling between the acoustic phonons of the fixed lattice and the diffusive excitations of the mobile silver ions were reported for RbAg415 and other systems. 3 Experimental Measurement of Ionic Conductivity The measurement of conductivity in solid electrolytes presents special problems partly because these materials may exhibit both ionic and electronic conductivity and also because in polycrystalline materials both inter- and intra-crystalline effects can be important. In the interpretation of experimental data especially with variable frequency a.c.it is important to appreciate the differing characteristics of blocking and non-blocking electrodes and also to have some idea of possible conductivity dispersions which may arise from the nature of the conduction mechanism or the presence of certain features in the sample microstructure. Thus if solid electrolytes are to be used in battery applications it is important to establish that electronic conductivity is negligible so as to avoid internal short circuits. This is not a straightforward procedure and although various semi-quan- titative tests can be applied the problem is by no means completely solved. In an l8 P. Briiesch L. Pietronero S. Strassler and H. R. Zeller Electrochim. Actu 1977,22 717.'' B. A. Huberman and R. M. Martin Phys. Rev. 1976 B13,498. Solid Electrolytes important note2’ Kennedy has recalled the exact significance of the electronic conductance as measured by the Wagner blocking technique. He points out that the Faradaic efficiency of a battery is not a simple function of this quantity since hole and electron carrier concentrations are dependent on electrode potential and thus on conditions of ohmic polarization when current is flowing in the cell. He suggests that in addition to quoting say a value of u+’the hole conductivity an electronic decomposition potential should also be reported in order to define practical limits for cells using the electrolyte. The further problem of studying electronic conduc- tivity in ternary systems namely the requirement of fixing the electrochemical potential of two constituents of the electrolyte is considered by Matsui and Wagner.21 Many solids do exhibit predominantly ionic conductivity but nevertheless conductivity measurements may not be straightforward because of grain-boundary effects.It is frequently of great importance to detect the presence of these inter- particle impedances which can have an adverse effect on the properties af poly-crystalline electrolytes e.g. causing deterioration in the selectivity of 02-conduc-tors,” and reduced lifetime of p-alumina ceramic^.'^ The use of complex plane diagrams for separating bulk grain-boundary and electrode effects was first described by Ba~erle.~~ He was able to extract resistance values from semicircles corresponding to series RC elements from the appropriate intercepts on the real axis of the complex udmittunce (Y*) diagram.Later workers see for example refs. 22 and 23 have generally preferred to use the impedunce formalism where Z* = (Y*)-’. A full discussion of the use of complex impedance diagrams has been given by Armstrong et with reference to an equivalent circuit Figure 2 containing a series array of parallel RC elements. Blocking electrodes are represented by a single capacitance and separate RC elements are used to distinguish different physical features such as grain boundaries and different types of crystal. Indeed it is fair to say that the standard procedure for determining electrolyte conductivities (aside from d.c.measurements with rever- sible electrodes) involves the use of blocking electrodes and complex impedance diagrams. It is sometimes difficult to assess the virtues of different kinds of data Figure 2 Equivalent circuit of an ‘ideal’ solid electrolyte *’ J. H. Kennedy J. Electrochem. SOC.,1977,124 865. 21 T.Matsui and J. B. Wagner jun. J. Electrochem SOC.,1977 124 300 610 937 941. ” N. M. Beekmans and L. Heyne Electrochim. Acta 1976,21,303. 23 R. W. Powers and S. W. Mitofi J. Electrochem SOC. 1975,122,226. 24 J. E. Bauerle J. Phys. Chem. Solids 1969,30,2657. 25 R. D. Armstrong T. Dickinson and P. M. Willis J. Electroanalyt. Chem. 1974 53,389. 30 M. D. Ingram and C. A. Vincent analysis. Recently Macdonald and Garber26 have compared the use of admittance and impedance formalisms.They showed that essentially the same results can be obtained from either method when the data are treated by complex nonlinear least squares analysis. Measurement of impedance in the frequency range lo2-lo7 Hz by conventional a.c. bridges is straightforward but time-consuming. The use of phase-sensitive detectors and more recently automatic frequency response analysers has made possible rapid measurements over the whole range down to 104Hz and the impedance data can be processed by an on-line computer.27 Using four-terminal a.c. measurements Powers and Mitoff 23 have shown that the electrode dispersion can be effectively eliminated. Farrington” has also described a method for looking at inter-grain and intra-grain resistivities in polycrystalline p -alumina using square waves with a rise time of 7 X lO-’s.Data obtained in the effective frequency range 10-4-5 x lo7Hz agree with those obtained by conventional methods. Despite these successes impedance measurements have provided little informa- tion concerning the intracrystalline effects arising for example from the preferred orientation of anisotropic crystals. These essentially bulk phenomena might give rise to conductance dispersions or dielectric losses closely related to the Maxwell- Wagner effect which for a highly conducting solid would normally occur at frequencies in the range 109-10’2Hz i.e. well outside the range of normal ax. equipment. Furthermore the impedance formalism highlights the larger resis- tances in the series equivalent circuit and hence information from the more conductive circuit elements (e.g.favourably oriented crystals) is effectively suppressed. These factors have been discussed by Hodge Ingram and West,29 who have pointed out the advantages of low-temperature measurements (bringing the conductivities below S m-’ and the conductance dispersions into a more convenient frequency range) and the use of the complex inverse permittivity or electric modulus formalism. Briefly M* = (E*)-’ = joCoZ* where j = a, o is the angular frequency and Co is the vacuum capacitance of the cell. In the complex M* plane semicircles corresponding to the higher frequency relaxations are given full weight and it has been possible to distinguish the contributions to electrolyte resistivity from the different crystal orientations in polycrystalline /3 -alumina,29 and from highly conducting ‘Suzuki phases’ in single-crystal sodium ~hloride.~’ The equivalent circuit in Figure 2 contains an array of RC elements and if the individual time constants are sufficiently well resolved semicircular arcs appear in the appropriate complex-plane diagrams.Alternatively if the data are plotted with the imaginary component versus the logarithm of the frequency characteristic Debye peaks are obtained. This may be regarded as the ‘ideal’ solid electrolyte behaviour; for many practical purposes this approach is justified. However it is 26 (a)J. R. Macdonald and J. A. Garber J. Electrochem. SOC.1977 124 1022; (b)J. R. Macdonald J. Chem. Phys. 1974,61,3977. 27 R. D. Armstrong M. F. Bell and A. A. Metcalfe J. Electroanalyt. Chem. 1977 77 287. 28 G. C. Farrington J. Electrochem.SOL 1976,123 1213. 29 (a)I. M. Hodge M. D. Ingram and A. R. West J. Electroanalyt. Chem. 1976,74 125; (b)R. J. Grant M. D. Ingram and A. R. West Electrochim. Acra 1977,22,729; (c)R. J. Grant M. D. Ingram and A. R. West Nature 1977 266 42. 30 N. Bonanos and E. Lilley personal communication. Solid Electrolytes 31 worth remembering that Figure 2 is essentially only a model of electrolyte behaviour and that the shapes of the complex plane diagrams (admittance impedance and modulus) may be quite different. There is growing evidence that many aspects of solid electrolyte behaviour are fundamentally ‘non-Debye’ in nature.Thus Jonscher3‘ has argued that in any system containing interacting mobile charges (ions electrons or rotating dipoles) interionic forces will lead to an expected or ‘universal’ frequency response which will reflect a fixed proportion of energy stored and dissipated per cycle. The complex admittance for example will then be given by an equation of the form Y* = Ao + jBw ” where A,B and n are constants. Despite a number of recent investigations using single-crystal electrolytes opinion is still divided regarding the origin of such additional dispersions. Raistrick et al.32have recently reported an impedance dispersion with blocking electrodes (sputtered Au or Pt) on crystals of P-PbF2 which they attributed to interionic .~ forces in the disordered inter-phase region.However Armstrong et ~1attributed~ similar effects observed with P-alumina to the existence of surface roughness and have shown that ‘ideal blocking behaviour’ can be restored by electrode polishing (see also ref. 26). It has been that the intracrystalline response however is governed mainly by the concentration of mobile ions. Thus materials having relatively ordered structures and small concentrations of mobile defects such as P-PbFz32 and AgC1,34 approximate to the ideal behaviour and the impedance diagrams are almost perfect semicircles. However with single-crystal P-alumina there is a pronounced conductivity dispersion which can be interpreted” as evidence either for strong interactions among the mobile Na+ ions or for the existence of a wide distribution of ion jump frequencies which reflects the disorder in the conduction plane.The equivalent circuit given in Figure 2 implies that the conductivity can either remain constant or else increase with increasing frequency and it might be expec- ted that this behaviour will continue into the microwave (GHz) region. However Armstrong and Taylor,35 and also Funke’ have found with electrolytes of very high conductivity (the optimized conductors such as a-CuI or RbAg&) that in this region the conductivity decreases with increasing frequency and negative capi- tances can be measured. The a(w)spectrum is very much like that predicted by the Drude model and Funke’ gives a detailed theory in terms of the acceleration and .retardation of ions in the oscillating electric field. However Grant et ~1have ~~ pointed out that similar effects can be observed in the radiofrequency region with P-alumina electrolytes if the circuit contains stray inductance. Ultimately the conductance spectrum terminates in the far-i.r. where resonance absorption spectra can be assigned to the collective and localized ionic motions.18 A 31 A. K. Jonscher Phys. Stat. Sol. 1975,32 665. 32 I. D. Raistrick C. Ho Y. W.Hu and R. A. Huggins J. Electroanalyt. Chem. 1977 77 319. 33 R. D. Armstrong and R. A. Burnham J. Electroanalyt. Chem. 1976,72,257. 34 R. B. Buck D. E. Mathis and R. K. Rhodes J. Electroanalyt. Chem. 1977 80 245. 35 R. D. Armstrong and K.Taylor J. Electroanalyt. Chem. 1975,66 258. ’‘ R.J. Grant M. D. Ingram and A. R. West J. Electroanalyt. Chem. 1977,80 239. 32 M. D. Ingram and C A. Vincent wide range of spectroscopic techniques including i.r. absorption Raman and quasi-elastic neutron scattering are now providing a detailed knowledge of the conduction mechanism especially in AgI-type electrolytes. For further details reference should be made to the review by Funke.8 Rapid Screening Methods.-It has been known for many years that the motional narrowing of n.m.r. lines can be used as a guide for the existence of translational motions in solids. This method is still valuable and in a recent paper Shannon et aZ.37reported 'Li n.m.r. measurements as an initial screen to determine the poten- tial for Li+-ion conductivity in a large number of samples.Motional narrowing of the dipolar-broadened 7Li resonance can be due either to long-range motions or to short-range motions not directly related to the ionic conductivity. The authors concluded that the absence of motional narrowing could be taken as a fairly sure guide to the absence of ionic conductivity but that the motional narrowing (when observed) was often dominated by short-range motions with lower activation energies than the corresponding d.c. conductivities. Shriver et aL3' also describe a method which they claim may be useful for screening potential ionic conductors. They examined the variation with tempera- ture of the Raman bandwidths of Ag2Hg14 and Cu2Hg14. The order-disorder (/3 $a) transitions at 50 and 70 "C were clearly indicated by band broadening at the Hg142- stretching frequency but in "12Hg14 (which is a poor ionic conductor) such effects were absent.4 Silver Ion Conductors Above 149"C the unexceptional P-AgI undergoes a phase change into the a-form. This transformation in which the structure becomes highly disordered with silver ions distributed over a large number of equivalent sites is accompanied by an increase in silver ion conductivity of almost four orders of magnitude. It was initially considered that the prime role of the large foreign ions such as Rb' in the MAg& class of solid electrolytes was to 'stabilize the a-structure' down to room temperature. X-Ray studies of the systems however revealed a more complex situation with unique structures not directly related to a-AgI.Over fifty silver ion conductors based on AgI are now known." These may be divided into two general classes (i) conductors formed by a formal partial substitution of the cation generally by K' or Rb' or by one of a large variety of organic substituted ammonium ions and (ii) conductors formed by a formal partial substitution of the anion mainly by oxoanions. An important feature of this class is that many of the best conductors are metastable phases or glasses. In reviewing the crystal structures of the cation-substituted conductors RbAg& (Me4N)2Ag13115 and (C5H5NH)Ag516 Geller' showed that the common structural element was a system of multiple face-shared iodide ion polyhedra some of which were occupied by silver ions.The interconnected polyhedra provided a system of 'passageways' through which silver ions might migrate. At that time Armstrong 37 R. D. Shannon B. E. Taylor A. D. English and T. Berzins Electrochim.Actu. 1977,22 783. 38 D.F.Shriver G. Joy and D. Greig J. Electrochem. Soc. 1976,123 588. Solid Electrolytes 33 Bulmer and Dickinson3’ suggested that the energetics of cation migration through a tetrahedral face would be more favourable if the cation was relatively stable in both three- and four-co-ordinate sites -a property common to both Ag’ and Cu’ ions. More recently Chan and Geller4’ have reported the structure of the crystalline electrolyte based on silver iodide/silver tungstate which has a formula Ag26118W4016.Here the conduction pathways involve ninety face-sharing iodide polyhedra and in addition fifty-six mixed iodide+xide polyhedra per unit cell. A high silver ion occupancy of the mixed polyhedra implied that mobility in these was lower than in the pure iodide structures. A number of silver ions were also resident in polyhedra which were not part of the conduction pathways. Other structures have been determined by Coetzer and Tha~keray.~~ The importance of face- sharing polyhedra in silver ion electrolytes has been discussed by Raleigh4* who considers that the principal function of the added foreign ions is to force the overall structure into a greater degree of face sharing. Addition of a cation that CO-ordinates six iodide ions or that requires an octahedral iodide ion hole fulfils this purpose.Further it appears that introduction of such foreign ions permits a reduction in the distortion (and hence distortion energy) which must accompany an attempt to form a space-filling structure using only regular tetrahedra. Figure 3 shows a number of interlinked tetrahedra forming say an iodide ion lattice. It can be seen that a silver ion traversing a pathway through this structure will pass through a trigonal co-ordination configuration on its way between neighbouring tetrahedral sites. Phillips” has discussed the formation (mainly) of cation substituted conductors in terms of a quantitative theory of chemical bonding. He proposed that for the polarizable bonds characteristic of ‘soft’ structures like silver iodide there is a balance between covalent and ionic forces which determines phase transitions between competing structures.A particular phase may then be stabilized by a minority cation. The theory goes on to postulate the existence of scaffolded struc- tures where iodide complexes of the minority cation provide a skeletal framework. For the large organic based cations there is evidence from X-ray diffraction studies for the determinative structural role suggested by Phillips. However it is very difficult to perceive the structure of the anion-substituted silver iodide Ag26118W401640 as ‘scaffolded’. A number of years ago Kunze’ pointed out that in the AgI/Ag2Se04 system in addition to a conducting crystalline phase still better conducting but metastable or glass-like phases could be obtained.He pointed out that glass-like structures might show particular advantages over crystalline systems because of the possibility of wide ranges of composition and the use of glass technology for the production of practical devices. Since that time a large number of glass-forming and non-glass forming anion-substituted AgI electrolytes have been reported. During the past two years Lazzari Scrosati Vincent and co-worker~~~ have discussed electrolytes 39 R. D. Armstrong R. S. Bulmer and T. Dickinson J. Solid State Chem. 1973,8 219. 40 L. Y. Y. Chan and S. Geller J. Solid State Chem. 1977 21 331. 41 J. Coetzer and M. M. Thackeray Electrochim. Acta 1976,21 37 Acta Cryst.1976 B32 1248 2966. 42 D. 0.Raleigh J. Electrochem. SOC.,1977,124 1157. 43 (a)M. Lazzari B. Scrosati and C. A. Vincent Electrochim. Acta 1977,22,51;(b)B. Scrosati A. Ricci and M. Lazzari J. Applied Electrochem.. 1976 6 237; (c) F. Bonino M. Lazzari A. Lonardi B. Rivolta and B. Scrosati J. Solid State Chem. 1977 20 315. M. D. Ingram and C. A. Vincent Figure 3 Interconnected tetrahedra providing a passageway for ion migration mobile ions must pass through a three-co-ordinate configuration on their way between tetrahedral sites .~~ containing Crz07’- and Mo207’- Schiraldi et ~1 Te042- Cr207’- WO:- Moo4’- and Cr04’- Kuwano and Kat~~~ Cr2072- and P20$- and ~ Minami et ~1.~ and P2074-. The work of Chan and Geller4’ on the tungstate system has been referred to above.An important feature of a number of these systems is that unlike RbAg415 and KA&15 they are stable in moisture. Further unlike most cation-substituted electrolytes they do not react with uncomplexed iodine. On the other hand with the exception of the tungstate electrolyte all the oxoanion substituted systems have a limited temperature range over which they are stable decomposing before their melting points. The morphology of the conducting phases is one of the main points of interest in this class of compound. With the possible exception of the AgI/Ag2Moz07 system all the electrolytes can be produced in glass-like form by a rapid quenching of the melt in liquid nitrogen or by splat-forming techniques. It is considered that the high viscosity of the melt prevents the rearrangement of ions into a crystal structure when the cooling is rapid.The best electrical performance is generally found with 44 A. Magistris A. Schiraldi and G. Chiodelli J. Applied Electrochem. 1976 6 25 1 ;Electrochim. Acta 1977 22 689. ” J. Kuwano and M. Kato Denki Kagaku 1975,43 734; ibid. 1977,45 104. 46 (a)T. Minami H. Nambu and M. Tanaka J. Amer. Ceram. SOC.,1977,60,283,467; (b)T. Minami Y. Takuma and M. Tanaka J. Electrochem. SOC. 1977,124,1659. Solid Electrolytes this fast-quenched material. As the temperature of the quenched electrolyte is raised devitrification processes occur which may be accompanied by decom- position into poorly conducting phases. Thus the conducting vitreous phases may be regarded as thermodynamically unstable at room temperature.On the other hand the kinetics of devitrification or decomposition become appreciable only at higher temperatures thus giving a number of these systems an acceptable stability at room temperature for practical purposes. It has been found that a prolonged annealing time is necessary in order to complete the transformation of a glassy electrolyte into the thermodynamically stable phases.43 The silver iodide/silver metaphosphate system has been studied by Malugani et al.47-it is not made clear whether trimetaphosphate or a mixed metaphosphate was used. A maximum conductance of 1.5 S m-' at 25 "C was noted in glasses with 57.5 m/o (mole per cent) AgI this composition appeared to represent the limit of solubility of AgI in AgPO,.In all respects this electrolyte seemed similar to those discussed above. In a second paper the same authors report ionically conducting glasses in the system lead iodide/silver metaphosphate. The maximum conductance (0.7 S m-' at 25 "C) and minimum activation energy (21 kJ mol-') were obtained at a composition of 19m/o PbIz. However it seems probable that the conductance is due to the formation of a AgI-AgP03 electrolyte together with non-conducting Pb(P03X. Silver ion conductors of a completely different type have been discussed by Tell and ~o-workers.~~ AgCrSz and AgCrSez are members of a large class of compounds of formula AMX2 which have structures containing two-dimensional layers of A and M.In general such compounds are electronic conductors and it was therefore necessary to block the electrons with a layer of RbA&I in order to determine the ionic component of the conductivity. For AgCrSz this was 0.03 S m-' at room temperature with an activation energy of 38.6 kJ mol-' below 150 "C. A phase change at 170"C was accompanied by a rapid rise in the ionic conductance. For AgCrSez the silver ion conductance at room temperature was comparable to that of RbAhI, although the electronic component was several times larger than the ionic value. A phase change was noted at 138 "C for AgCrSez. The authors point out that the latent heat of such transformations was much smaller than that associated with the more significant structural changes associated with the a! $p transformation of AgI or AgzS.On the other hand they had magnitudes similar to that of the second-order phase transformation of RbAgJ,. In a further paper ionic conduction was reported for two classes of ternary chalcogenide. The first group included AgGa5S8 CuGa& and CuGaSSes and is closely related td the well known 'cation deficient' ionic conductors a! -Ag2Hg14 and a! -Cu2Hg14 which have a simple zinc blende structure but with only three of the four cation sites filled in the unit cell. The second group the chalcopyrite semiconductors (e.g. AgGaS2) also show some silver ion conductivity at higher temperatures. However neither of these two latter groups of compound show appreciable ionic conductivity under ambient conditions. 47 J.P.Malugani A. Wasniewski M. Doreau and G. Robert Compt. rend. 1976,283,C 111 and 1977 284,c,99. 48 (a)B.Tell S. Wagner and H. M.Kasper J. Electrochem. SOC. 1977,124,536;(6)D. W.Murphy H. S. Chen and B. Tell J. Electrochem. SOC. 1977,124 1268. 36 M. D. Ingram and C. A. Vincent 5 Copper Ion Conductors The synthesis of stable copper ion electrolytes with good electrical properties is of some technological importance owing to the relatively low cost of copper compared with silver. Since 1973 a number of conductors based on the introduction of organic substituted ammonium ions into copper(1) halides have been reported by Taka- hashi and co-workers and by Sammels et af. (see for instance Takahashi"). In contrast to the silver ion conductors copper solid electrolytes have been obtained not only with cuprous iodide but also with cuprous bromide and chloride.Here Phillips' the~ry'~ on bond ionicity and interatomic forces (originally proposed before the synthesis of such copper 'double salt' electrolytes) has proved notably successful. Conductivity trends etc. with different halides have been found to be consistent with his predictions. Since energy differences between different coor- dination configurations are larger for the cuprous halides than for silver iodide a larger minority cation is required for the stabilization of disordered cuprous struc- tures. Recently Takahashi and co-worker~~~ have extended the range of organic ammonium halides used to include substituted and unsubstituted piperidinium morpholinium and sulphonium halides.It was concluded that the formation of conducting systems was related not only to the size of the substituted ammonium ion but also to its flexibility. For instance 1-methylpyridinium iodide did not form an ionic conductor when combined with CuI whereas the similarly sized 1-methylpiperidinium iodide did. Of the sulphonium halide systems studied 1-methyl-1-thionacyclohexaneiodide/CuI (83.4 m/o) and 4-methyl-1-oxa-4-thio- nacyclohexane iodide/CuI (83.4 m/o) gave the highest ionic conductivities of 0.14 S m-' and 0.072 S m-' respectively at room temperature with an activation energy of 22 kJ mol-'. A galvanic cell using the N-methylhexamethylene bromide/CuBr electrolyte copper anode and chalcogen cathode has been developed successfully by Takahashi and Yamam~to,~' which was able to achieve current densities of 50-100 pA cm-* at room temperature.Cells using Se or Te could be operated up to temperatures of 150 "C and were found to show long term stability. However investigations by different workers have shown that certain organic-substituted copper ion conductors are unstable over long periods and are incompatible with copper and with halogen electrodes.*' An intermediate compound in the KI/CuI system KCu415 stable between 257 "Cand 332 "C was reported by Bradley and Greene.2 This compound has been re-examined by Bonino and LazzariSo and later by Matsui and Wagner.21 Both groups found the conductivity of this compound to be essentially ionic throughout its stability range.The conductance values reported from the two laboratories varied by a factor of ten and the activation energies by a factor of two this is likely to be a function of the different heat treatments used. The investigation of KCu4I5 by Matsui and Wagner formed part of a broad study aimed at the preparation of copper solid eleccrolytes without organic additives. These workers examined a number of systems (i) CuI doped with an aliovalent cation (uia CdIz) in order to 49 (a)T. Takahashi N. Wakabayashi and 0.Yamamoto J. Electrochem. SOC. 1976,123,129; J. Applied Electrochem. 1977,7 253; (6)T. Takahashi and 0.Yamamoto J. Applied Electrochem. 1977,7 37. F. Bonino and M. Lazzari J. Power Sources 1976,1 103. Solid Electrolytes 37 change the concentration of lattice defects without affecting the charge neutrality.Mobilities and energies of formation and of migration of copper vacancies and interstitials were derived from the temperature dependence of the conductivity. Studies were also made at high (20 m/o and 33 m/o) Cd12 concentrations which showed that the ionic conductance of CuI was enhanced by about a factor of ten at temperatures above 200°C. (ii) CuCl doped by Cu2S. Here the mobility charac- teristics of copper interstitials and the concentrations and energies of formation of Frenkel defects were found. (iii) The compound Cu4CdCls which showed slightly higher conductance above 130"C than pure CuCl. (iv) The KI/CuI system (refer- red to above) the RbI/CuI system and the RbCl/CuCl system.In this last system ionic conductivity at room temperature was found over a wide range of composi- tion. The highest conductance for a composition equivalent to RbCu3C14 was 0.225 S m-' at 25 "C. Below about 80 "C the activation energy was 18 kJ mol-'; above 110"C it was 13.5 kJ mol-' and it seemed that a second-order phase tran- sition occurred near 100 "C. X-Ray powder diffraction and thermal analysis stu- dies showed that RbCu3C14 was a new compound and not a mixture or solid solution. Takahashi and co-workersSf have also studied the CuI/Pb12 and CuBr/PbBr2 systems. No intermediate compounds were found in the former but a compound of formula CuPbsBr7 was shown to be stable between its incongruent melting point at 300 "C and 160 "C.At 200 "C it had a conductance of 3 S m-' with an activation energy of 22 kJ mol-' -similar to that of many 'average structure' compounds. Transport number measurements confirmed that copper ions were effectively the only charge carrier. A further investigation in the same laboratory found that the a-phase of the mixed conductor Cu2-8Se had a high ionic conductance again associated with an average structure. By blocking the electrons with an organic- based copper electrolyte a room temperature ionic conductance of 3 S m-' was determined. It was noted that the a p transition temperature fell as S was raised so that for 6 20.15 the a-phase was stable at room temperature. An interesting new group of ternary copper compounds studied by von Alpen Fenner Marcoll and Rabena~,~' has the general formula CuTeX where X represents chloride bromide or iodide.Conductivity polarization and trans- ference studies have shown that at 100 "C the copper ion conductance ranges from around 2 X S m-' for CuTeI to 0.1 S m-' for CuTeBr; at 250 "C the ionic conductance is as high as 1S m-' for the latter compound. In all three electrolytes the electronic conductivity is 1x S m-l at room temperature and increases to 5 X S m-' for the bromide and chloride and to 5 x S m-' for the iodide at 300 "C. The crystal structure of these compounds is of interest four types of Cu' site are found and these are populated statistically. A marked difference is noted in the CuTeCl occupation probabilities compared with the other two compounds in which the probability of occupation of all sites is nearly equal.Conductance-temperature plots for a representative selection of silver and copper electrolytes are given in Figure 4. 51 (a) T. Takahashi 0. Yamamoto and H. Takahashi I. Solid State Chem. 1977 21 37; (6) T. Takahashi 0.Yamamoto F. Matsuyama and Y. Noda J. Solid State Chem. 1976,16,35. 52 U. von Alpen J. Fenner J. D. Marcoll and A. Rabenau Electrochim.Acta. 1977,22,801. M. D. Ingramand C. A. Vincent T I "C 300 200 100 50 0 -50 I 1 I I I I +2 +1 h r E I/) -b Y 0, -00 -1 1.5 2.5 3.5 4.5 Figure 4 Conductance-temperature behaviour of some silver and copper solid electrolytes 6 Indium(1) and Thallium(@ Conductors The conductivity and Raman spectra of the ionic conductors In4Cd16 InzZn14 and f12Zn14 and of the related compounds T14Cd16 were discussed by Ammlung eta^'^ The three conductors underwent order-disorder phase transitions at temperatures above 200 "C which were detected using Raman spectroscopy by the broadening of the Rayleigh scattering line.Below the transition temperatures the conductivity of In4Cd16 and InzZn14 was almost totally electronic (as is the case in the related AgzHg4). Above the transition temperature the ionic component was some fifty 53 R L. Ammlung D. F. Shriver M. Kamimoto and D. H. Whitmore J. Solid State Chern.,1977,21,185. Solid Electrolytes 39 times greater than the electronic component. For TI2ZnI4 the electronic conduc- tance was negligible above the transition temperature.The conductivity was noted to be frequency dependent but no explanation for this effect has been found. Experiments with blocking and non-blocking electrodes confirmed the motion of TI’ and In’ ions. 7 B-AIumina and Related Electrolytes /3 -Alumina.-The principal motive for studying 6-alumina remains the develop- ment of electrolytes for sodium-sulphur cells and following the classic paper of Yao and Kummer4 this is still probably the most extensively studied solid elec- trolyte system. 6-Alumina itself has the ‘ideal stoicheiometry’ Na20 1 1AI2O3 but it always contains between 15 and 30% excess Na20. A crystal structure refinement of 6-alumina was reported in 1971 by Peters et dS4 This confirmed the general picture of a structure containing alternating spinel blocks and conduction planes oriented perpendicular to the c-axis and showed that both the Beevers- Ross (BR) and mid-oxygen (m0)sites were partially occupied by Na’ ions.However the anti Beevers-Ross sites (aBR) were found to be empty (cf. the structure of ion-exchanged Ag 6-alumina).” A long-standing problem concerns the method of charge compensation for the ‘excess’ Na+ ions in the apparently nonstoicheiometric 6-alumina. Thus Peters et al. thought that the excess positive charge was balanced by A13+ vacancies in the spinel blocks but subsequently Roth has favoured the presence of extra (inter- stitial) oxygens located in the conduction plane. This problem has recently been discussed by Roth Reidinger and La Placa12 from the standpoint of releasing the stresses within the spinel structure.In @-alumina stress is relieved by the formation of ‘compound defects’ with interstitial 02-ions held in position by A13’ interstitials (Frenkel defects) located immediately above and below the conduction planes. However in the 6”-aluminas (related polytypes of typical stoicheiometry Na20,Mg0,5AI2O3) it is released by selective replacement of A13+ by Mg2+ or Li’ ions on tetrahedral sites within the spinel blocks. Roth Chung and Storys6 have described the characterization of additive-modified 6-alumina ceramics by 23Na n.m.r. The additives were found to fall into three groups. First there are those like Y203 which are insoluble in 6-alumina and have no influence on sodium motion.Secondly there are those like LizO MgO and NiO which enter the spinel blocks and generally cause an enhancement of Na’ mobility. Finally there are those like CaO where Ca2’ ions replace Na’ ions in the conduction planes and there is a decrease in conductivity. These findings are in broad agreement with the results of Boilot et ul.,” who have also studied the influence of foreign ions on the relative stability and electrical conductivity of p and p” phases. They found that the incorporation of certain ions (e.g. Mg2+ and Ni2+) with ionic radius less than 97 pm favours the formation of the p”-phase by stabilizing the spinel structure but also causes a corresponding ” C. R. Peters M. Bettman J. W. Moore and M. D. Glick Acta Cryst. 1971 B27 1826.” W. L. Roth J. Solid State Chem. 1972,4 60. s6 (a) W. L. Roth J. Chung and H. S. Story J. Amer Ceram. SOC.,1977 60 311; (b)W. Bailey S. Glownikowsky H. Story and W. L. Roth J. Chem. Phys. 1976,64,4126. 57 J P. Boilot A. Kahn J. Thkry R. Collongues J. Antoine D. Vivien C. Chevrette and D. Courier Electrochim. Acta 1977 22 741. 40 M. D. Ingram and C. A. Vincent increase in the conductivity of @-alumina. It was argued that in these doped crystals the smaller concentration of interstitial 02-ions allows more room in the conduction planes for Na' ion motion and hence the conductivity is increased. Thus the conductivity of polycrystalline material depends on the concentration of dopants rather than simply on the proportion of @-and @"-phases.In a somewhat different approach Sat0 and Hirotsu5* have argued that @-alumina is in fact a stoicheiometric compound but that its stoicheiometry is far from the idealized formula Na20,1 lA1203. The existence of the apparent 'solu-bility range' for Na20 is due to the ability of the structure to accommodate mixed periods rather than to the presence of excess Na' ions in the conduction planes. In 1971,Whittingham and H~ggins~~ reported on the electrical conductivity of a range of ion-exchanged (Li+ K' Ag' or Tl+)single crystals of @-alumina. They evaluated the Haven ratios and proposed at least for Nap-alumina that conduc- tion involves an interstitial type mechanism. This interpretation is consistent with a more detailed theory by Sat0 and Kikuchi discussed earlier,'2*'3 where the cor- related motions of large numbers of ions are considered.There is now theoretical and experimental evidence that long range conductive and diffusive processes do involve the collective motions of groups of ions. Thus Wang et aL60 have cab culated potential energy curves for cationic motion within a two-dimensional hexagonal pathway. They show that the energy of the system is minimized if the 'excess' Na20 leads to a local rearrangement of mobile ions and to the formation of 'interstitial pairs' (two Na' ions on mO sites next to a vacant BR site). Theoretical activation energies calculated from the in-phase migration of these interstitial pairs (a modified interstitialcy mechanism) are in agreement with experimental values but the energy of simply moving a single Na' ion from a BR to a neighbouring aBR site is approximately an order of magnitude larger.The conductivity of @-alumina in the frequency range 107-10'3Hz has been investigated by Strom et d6' They conclude that the quadratic temperature dependence of conductivity at a fixed frequency of lO'OHz and the large width (ca. 23 cm-') of the 65 cm-' phonon mode are consistent with a conduction mechanism where co-operative effects similar to those proposed by Wang give rise to a spectrum of excitation energies (cf. the conductance dispersions found at lower temperatures and in the frequency range 102-107Hz discussed above.) In this context it is interesting to note that Bailey et aLS6have described an n.m.r.study of Na' ion motion measuring the splitting and line-width of the 23Na n.m.r. line as a function of temperature for several crystal orientations. At low temperatures the Na' ions occupy several independent sites in the conduction plane as indicated by the nuclear quadrupolar splitting. However above -160 "C,this structure gives way to a single symmetric resonance line because of rapid jumping of Na+ ions among the inequivalent sites. The presence at a single temperature of both the 'static' spectrum and the motionally narrowed line suggested to the authors the existence of a wide distribution of ion jump frequencies -a conclusion however not inconsis- tent with the idea of co-operative ionic motions. H. Sato and Y. Hirotsu Materials Res. Bull. 1976 11 1307.59 M. S. Whittingham and R. A. Huggins NBS Special Publication 1972,364 139. 6o J. C. Wang M. Gaffari and S. Choi J. Chem. Phys. 1975,63 772. U.Strom P. C. Taylor S. G. Bishop T. L. Reinecke and K. L. Ngai Phys. Rev. 1976 B13 3329. Solid Electrolytes 41 Polycrystalline @ -alumina is used in electrochemical devices because of the anisotropic properties of single crystals and also because of the greater ease of fabrication of the polycrystalline material in the shapes required. The use of polycrystalline material however introduces two features not encountered with single crystals namely tortuosity and grain boundaries. These aspects have been extensively studied by Powers and Mitoff and have been reviewed by them in a number of publication~.'~*~~~~~ For example Wynn Jones and Miles had earlier noted that the activation energy for conductivity typically approaches the single- crystal value at higher temperatures but increases at low temperatures to much higher values.By using four-terminal a.c. methods and a simplified equivalent circuit to simulate the electrolyte response Powers and Mitoff have been able in favourable circumstances to measure the intra- and inter-crystalline contributions to the resistivity of polycrystalline materials. The intercrystalline resistivity has a higher activation energy (28 as compared with 18 kJ mol-') and below about 250 "C the major part of the electrolyte resistance may be located within the grain boundaries which thus determine the overall activation energy.Depending on the microstructure the activation energy can vary from sample to sample. When current flows through @-alumina a large part of the voltage drop will therefore occur across the grain boundaries giving rise to intense local electric fields. It has been suggested that these intense fields may cause accelerated degradation of the p-alumina ceramic during operation in sodium sulphur cells. Mitoff6* has also discussed the 'tortuosity factor' which is the ratio of the resistivities of the intracrystalline to the single-crystal material. Using a Monte- Carlo statistical approach he estimated that this ratio can be as low as 1.34 and concluded therefore that at least from this point of view there is little advantage to be gained from the use of single-crystal electrolytes.However Grant et al.29 have studied the anisotropic conductivity of single crystal 6-alumina. They show that crystal anisotropy does influence polycrystalline behaviour at very low tempera- tures (ca. -13O"C) and that the impedance contains contributions from both favourably and unfavourably oriented crystals. The increasing activation energy (and decreasing conductivity) of polycrystalline p-alumina compared with the single-crystal electrolytes as the temperature falls below 0 "C can thus depend not only on the presence of grain-boundary impedances but also on an increase in the tortuosity factor. Finally it must be remembered that 6-alumina is not a single substance but that these ceramic electrolytes differ considerably in composition and in microstructure.Thus the polycrystalline material may contain varying amounts of p-and @"-alumina depending on the method of preparation. For example using transmission electron microscopy De J~nghe~~ has demonstrated that p-alumina contains both stacking sequence faults and 0'' intergrowths which are introduced in the early stages of sintering. Related Electrolytes.-Both -and p"-alumina can be modified considerably by ion-exchange while still retaining their essential structural properties and thus can be regarded as 'parents' of a wide range of electrolytes. These new electrolytes may 62 R.w.Powers and S. P. Mitoff 'Technical Information Series' General Electric Company Repdrt No. 77CRD055 1977. 63 L. C.De Jonghe J. Materials Sci. 1977 12 497. 42 M. D. Ingram and C A. Vincent be important in their own right or simply as model systems for further investigation of @-alumina. Thus Dudley and Steele64 have studied the potassium ferrites of general stoicheiometry K1+xFe11017 where A13' ions are replaced by Fe3' ions and the 'excess' K' ions are compensated by the presence of Fe2' ions. These materials are mixed ionic/electronic conductors and are potentially important as reversible K' electrodes. Using a special four-terminal technique with voltage probes reversible to either potassium ions or electrons they measured both ionic and electronic components of the conductivity and showed that the ionic conduc- tivity varied with excess K' content in a manner at least qualitatively consistent with the theoretical predictions of Wang et aL6' A number of authors have reported on the properties of sodium and potassium 0-gallates and also mixed aluminate/gallate/ferrite ~ystems.~'*~~ Of particular significance is the investigation by Chandrashekhar and Foster66 of the conductivity of single crystals of the gallium analogues of sodium 0-and 0"-alumina.They showed that whereas the electronic conductivity is insignificant in these materials the ionic conductivity is extremely high. Indeed Na p"-gallate has the highest Na' conductivity (ca. 10's m-' at 300 "C) of any solid electrolyte. Yao and Kummer4 had shown earlier that p-alumina has markedly different affinities for univalent cations (Li' Na' etc.)substituting in the conduction planes.They found for example that it is difficult to replace more than 50% of the Na' ions by Li'. Farrington and R~th~~ have recently discussed this phenomenon in the light of ionic interactions and the likely existence of inequivalent sites within the conduction planes. They have designated the existence of a new class of co-ionic conductors as typified by 1:1Li/Na p-alumina. This latter material is remarkable in that the Li' ion conductivity at 25 "C (0.5 S m-') is appreciably better than that of 'pure' Li p-alumina (ca. 0.01 S m-') and indeed higher than that of any other Li+-ion conducting solid at room temperature. 8 Other Alkali Metal Ion Conductors At present one of the most successful applications of solid electrolytes is the use of Li/12 cells in heart pacemaker batteries despite the low conductivity of LiI at room temperature (ca.lo-' S m-'). Developments in this area continue and Liang and Barnette6* have recently described a high-energy density long-life battery system where the electrolyte is a mixture of LiI and aAl,O of somewhat higher conduc- tivity. This type of cell is however limited in its application because of the low power output and in the last few years a great many workers have made strenuous efforts to discover more highly conducting solid electrolytes. Since in theory the highest energy densities could be obtained with Li cells particular attention has been given to the search for new Li' ion conductors. It has long been known that Li2S04 undergoes a reversible p$a transition rather like AgI and that the highly conducting (a)form is stable above 585 "C.It has nevertheless proved difficult to 6.i G. J. Dudley and B. C. Steele J. Solid Stare Chem. 1977,21,1. 65 K.Kuwabara and T. Takahashi J.Applied Electrochem. 1977,7 339. 66 G.V.Chandrashekhar and L. M. Foster J. Electrochem SOC.,1977,124,329. " G.C.Farrington and W. L. Roth Electrochim. Acta 1977 22,767. C. C. Liang and H. Barnette J. Electrochem.SOC.,1976,123,453. Solid Electrolytes 43 find appropriate additives which will extend the stability of highly conducting phases into a more convenient range (200-300 "C). However Heed et ~1.~'have recently described a range of sulphate-based mixtures which it is claimed can be used as electrolytes in batteries right down to room temperatures.H~ggins,~'and Shannon et al.37 have both described investigations of Li' conductivity in a wide range of solids. In a number of cases data were obtained for identical systems and were in very good agreement. The unit cell of Li4Si04 is believed to contain two Si04,- tetrahedra linked by eight Li' ions which are distributed over eighteen possible sites and promising conductivities at high temperatures (>350 "C) were earlier reported by West. Huggins and Shannon both investigated solid solutions of Li3 PO and Li4Si04. They showed that at 100"C an enhancement in the conductivity of Li4Si04of about five orders of magnitude can be obtained by adding 40 m/o Li3P04. Huggins7' also observed relatively high conductivity in Li3N (a very open struc- ture with intersecting tunnels in two dimensions) in LiAlCL below its melting point at 146"C and in a third group of materials based on the antifluorite structure including Li2S Li20 and related compounds with partially occupied cation sublat- tices such as Li3A104 Li,GaO, and Li6Zn04.The ionic conductivity of these latter materials rose very rapidly at about 380"C reaching high values (ca. 30 S m-') at 450 "C. However Johnson et al'l have since shown that this sudden rise in conductivity occurs only in moist environments and apparently depends on the presence of LiOH within the material. .~~ Shannon et ~ 1 looked at Li' conduction mainly in oxide ceramics such as boracite Li4B170, (Cl Br) (see also ref.72) and LiM,P,O, (M = Zr or Hf) which have framework structures and those such as Li3.7sSio.7sPo.2304 (see above) Li3.4Sio.7So.304, and Li2,,Co.,,B0.,,O3 which contain isolated polyhedra (IP) in a network of edge-linked Li polyhedra. At 300 "C. the conductivities of these last three electrolytes are all approximately 1S m-' where the activation energy EAis -55 kJ mol-'. Shannon et al. suggest that a common conductivity mechanism is operative in the IP structures and that other IP structures will have similar conductivities. They conclude that future searches for good Li conductors should probably be concentrated on compounds with framework structures. It has long been recognized that framework structures with proper channel sizes should be good candidates for high ionic conductivity and in this context the search for Li' conductors can be seen as one aspect of conduction by alkali metal ions has generally.Thus H~ng~~ prepared crystals of general formula Nal+ Zr2Si,P3- 012, which contain a rigid three dimensional network of PO4 and SiO tetrahedra sharing corners with Zr06 octahedra. He claims to have 'designed out' any 'bottlenecks' in the structure which might inhibit Na' motion. Certainly the conductivity of these materials is high and in this respect Na3Zr2PSi2012 ceramic at 300 "C is comparable with sintered 6"-alumina. It is interesting to note that the isotypic compound Lio.8Zrl.8Tao.2P3012 to was found by Shannon et ~21.~~ 69 B. Heed A. LundCn and K. Schroeder Electrochim.Act4 1977,22 705. 70 R. A. Huggins Electrochim. Act4 1977 22 773. 71 R. T. Johnson R. M. Biefeld and J. D. Keck Materials Res. Bull. 1977,12,577. 72 J. M. RCau G. Maginiez B. Calks C. Fouassier and P. Hagenrnuller Materials Res. Bull. 1976,11 1087. 73 (a)H. Y.-P.Hong Materials Res. Bull. 1976,11 173; (b)J. B. Goodenough H. P.-P.Hong and J. A. Kafalas ibid; 203. 44 M. D. Ingram and C A. Vincent have the lowest activation energy (41 kJ mol-') and the highest room temperature conductivity (ca. 5 X lo- S m-') of the Li electrolytes which they had studied. More recently Shannon et aL7 have reported on the conductivity of Na5GdSi401Z which appears to be very similar to that of Hong's electrolyte Na,Zr2Si,PO12. Also Delmas et aZ.75 describe K' conduction in new lamellar structures of general formula KL02-M02 where L = In or Sc and M = Zr Sn or Pb.Above 250 "C the conductivity of K0.721n0.72Hf0.2802 for example exceeds that of sintered p-alumina although the activation energy is greater and the conductivity decreases at lower temperatures. A great many cationic conductors have now been discovered which show good to average conductivity but are otherwise not particularly noteworthy. However von Alpen et a1." have shown that P -eucryptite LiAlSiO, exhibits several interesting features. First this is a one-dimensional Li+-ion conductor and measurements with single-crystals show that conductivity along channels parallel to the c-axis is about three decades higher than the conductivity in the perpendicular direction.The activation energy (in both orientations) is 71 kJ mol-' and the conductivity along the c-axis which is not remarkable reaches 10 S m-' at 600 " C. However there is evidence from diffuse X-ray scattering of large one-dimensional ordered regions (domains) surviving until temperatures around 600 "C and it is argued that Li-ion diffusion is a co-operative process involving strongly correlated Li+-ion motions. They suggest that P-eucryptite could be a model system for the study of predominantly one-dimensional motion in crystals. Bottelberghs et aZ.76have also reported on the properties of sodium tungstate. The conductivity of the intermediate P-phase (cr =1S m-') stable in the very limited temperature range 587-589"C is higher than that of both the low- temperature (y) and high temperature (a)phases.Isovalent doping by anions such as and Cr042- causes a metastable extension of the P-phase region by about 40 K but cationic doping for example by Ag' or K' ions causes the /3-phase to disappear completely. It has been established that conduction in the poorly conducting y-phase depends on extrinsic defect concentrations and that the rise in conductivity occurring around 550 "C involves a change to an intrinsic conduction mechanism. Matsui and Wagner77 have reported mixed ionic/electronic conduction in the lithium ferrite spinel LiFe508 where Li' ions occupy a portion (1/8) of the octahedral sites in the inverse spinel structure. Under an atmosphere of oxygen most of the conductivity is by Li' ions and reaches around 1 S m-' at 300 "C (similar to the materials examined by Huggins and Shannon).However under an atmosphere of argon the electron transport dominates because of the presence of Fe2' as well as Fe3' in the crystal. Future interest in this material may depend on its use as an electrode material reversible to Li' ions (cf.the behaviour of potassium ferrites discussed earlier). Conductivity data for some of these new alkali ion conductors are compared with @-alumina in Figure 5. It is apparent that there is still a shortage of good Li'-ion conductors. 74 R. D. Shannon H.-Y. Chen and T. Berzins Materials Res. Bull. 1977 12 969. 75 C. Delmas C. Fouassier J.-M. Rbau and P. Hagenmuller Materials. Res. Bull. 1976 11 1081.76 (a)P. H. Bottelberghs and G. H. J. Broers Electrochim. Acta 2976 21 719; (b)H. T. Cahen P. H. Bottelberghs and G. H. J. Broers Materiufs Res. Bull. 1977 12 693. 77 M. Matsui and J. B. Wagner J. Electrochem. SOC. 1977 124 1141. Solid Electrolytes TI"C 600 300 200 100 50 0 +1 0 CI 7 -1 E m \ b -2 3 a -0 -3 -4 1.o 2.o 3.0 4.O lo3K T Figure 5 Conductance-temperature behaviour of some alkali metal ion conductors 9 Proton Conductors The subject of proton conduction in the solid state was reviewed by Glasser'* in 1975. However despite their potential as fuel cell electrolytes few of such materi- als have been found which exhibit the high ionic conductance of the other elec- trolytes considered in this Report.Two classes of proton conductor may be identified (i) those in which there are no hydrogen-bonded chains and where proton conduction is by an interstitial motion between suitable co-ordination sites and (ii) those which have chains of linked hydrogen-bonds in which co-operative proton transfers can occur. No material of the first group has been found which possesses high ionic conductance. Of the second group the best conductors usually have a structure which permits molecular rotation and hence reorientation of Grotthuss-type chains. Indeed many materials only become proton conductors at higher temperatures above a transition to a phase where rotation can occur. Thus L. Glasser Chem. Reo. 1975,75 21. M. D. Ingram and C.A. Vincent one of the best proton conductors H3O+C1O4- has a structure which (above -30 "C) has the H30+ disordered between two different orientations with a barrier to rotation of only 7.5 kJ mol-I. The proton conduction is 0.03 S m-' at 25 "C but as the electrolyte decomposes at 50 "C and is very hygroscopic its practical appli- cations are limited. Takahashi et aZ.79 have recently reported protonic conductivity in tri-ethylenediamine sulphate and hexamethylenetetramine sulphate. The latter compound had a conductance of 0.003 S m-' at 100"C and 0.1 S m-l at 200 "C at which temperature it was quite stable the e.m.f. behaviour of a hydrogen concen- tration cell using this electrolyte was Nernstian. High protonic conductivity was found by Shilton and Howe8' in the insoluble layered hydrate HU02P04,4H20.A sudden fall in d.c. conductivity from a value of about 0.25 S rn-l at 0°C was attributed to a disruption of the conductance pathways by freezing of the semi-free water in the grain boundaries. The hydrate of phosphotungstic acid H3(PW,,0so) 29H,O has the highest protonic conductance in the solid state yet reported," with a value of 4 S m-' at room temperature; this material has been discussed by Takahashi. l2 Results of recent conductance studies on optimized proton conductors are shown in Figure 6. T /"C 300 200 100 50 0 -50 +1 0 -4 -5 1.5 2.5 3.5 4.5 103~ T Figure 6 Conductance-temperature behaviour of some solid state proton conductors 79 T. Takahashi S. Tanase 0.Yamamoto and S.Yamanchi,J. Solid State Chem. 1976,17,353. Solid Electrolytes 47 10 Fluoride Ion Conductors Much of the current interest in this field is concerned with the detailed understand- ing of the concentrations and mobilities of various types of defect in pure or lightly doped materials which do not show high ionic conductance. However in recent years a number of optimized fluoride conductors have been discovered which have properties at moderate temperatures which match the best solid electrolytes. Fluoride ion conductors may be grouped into two classes based on fluorite and tysonite structures respectively. It is well known that oxide ion conductors with fluorite lattice structures such as the stabilized zirconias thorias and cerias have very low conductance values at temperatures below 500-600°C.On the other hand the fluoride ion in similar structures is considerably more mobile. The fluorite structure may be regarded as a face-centred cubic array of cations with anion sites in the tetrahedral interstices. Fluoride ion motion is due mainly to anion vacancy movement and addition of dopants which increase the anion vacancy concentration greatly enhances the ionic conductivity. The second class of fluoride ion conductor is based on a lamellar configuration as in the tysonite structure (LaF3). In this compound the lanthanum ions and one third of the fluoride ions form a series of hexagonally based layers while the remaining fluoride ions lie between these and are relatively mobile. Fluorite Structures.-A number of solid solutions based on the fluorite structure (with high concentrations of foreign ions) have been studied and shown to have optimized conduction properties.Similarly ternary systems with formulae (LF3)1-x(MF2)x (where L is one of the lighter lanthanides and M is an alkaline earth metal) and tysonite structure have been reported which also have high conduc- tance. An interesting feature common to both these classes of fluoride conductor is that they show a gradual disordering with increasing temperature. Thus the conductivity passes smoothly from values typical of ‘normal’ ionic solids to values typical of fused salts. This type of transition has been termed the ‘Faraday tran- sition’ by O’Keeffe12*82 and is in effect a gradual change from vibrational to translational degrees of freedom.Kennedy and Miless3 used aliovalent doping to determine the nature of the charge carriers in P-PbF2 at low temperatures. While addition of NaF or KF increased the conductance (KF addition raised it from 5 x lo-’ S m-’ to 0.1 S m-’ at room temperature) BiF3 appeared to decrease it. This finding is consistent with the view that ionic current at temperatures below 180“Cwas carried solely by mobile fluoride ion vacancies. However earlier work with different tervalent ion-doping had shown an increase in conductance. This anomaly was pointed out by Bonne and SchoonmanS4 and the possibility of interfacial polarization or reaction between the electrode and Bi3+-doped electrolyte was suggested. The latter authors pro- vided additional data on pure and LaF3- and AgF-doped p-PbF2.From the temperature dependence of the conductance the concentrations and mobilities of M. G. Shilton and A. T. Howe Materials Res. Bull. 1977 12 701. 0. Nakamura T. Kodama I. Ogingo and Y. Miyake Japan Kokai 76 106 683 (1976). M. O’Keeffe and B. G. Hyde Phil. Mag. 1976,33,219. 83 J. H. Kennedy and R. C. Miles J. Electrochem. SOC. 1976,123,47. 84 R. W. Bonne and J. Schoonman J. Electrochem. SOC. 1977,124,28. 48 M. D. Ingram and C A. Vincent both anion vacancies and interstitial anions were found. A kink in the Arrhenius plot was attributed to a change from extrinsic to intrinsic conduction by fluoride vacancies. Support for the idea of interstitial anion mobility in @-PbF2 was given by a study of thin films using a-particle retrodiffusion measurements by PistrC et aLg5 The conductivity characteristics of PbF containing large quantities (5-25 m/o) of Bi3+ or A13+ were investigated by Lucat et a1.86and Joshi and Liang,87 respec- tively.The former workers studied Pbl-,Bi,F2+ for (0 S x S 0.5). The lattice parameters of the @-PbF2 were found to decrease linearly with increasing BiF3 indicating the formation of a solid solution. The maximum conductance was reported for the system with x = 0.25 which gave a value of 0.04 S m-' at 60 "C and 50 Sm-' at 350 "C (similar to that of @-alumina). Joshi and Liang discovered that inclusion of large concentrations of AlF3 stabilized the p -form which unlike pure @-PbF2 did not revert to the a-form when subjected to pressure or mechani- cal shock.PbF2 containing 12 m/o AIF3 had a room temperature conductance of 0.011S m-' an activation energy of 34 kJ mol-1 over the range -50-70 "C and an electronic conductance of S m-'. While this electrolyte was completely stable at room temperature its conductance decreased substantially on annealing at 100"C or above. As with Pbl-,Bi,F2+ the increase in conductivity with increasing x revealed a complex mechanism involving vacancies interstitials and possibly defect clusters. Lucat and co-workers also studied the similar system Gal-,Y,F2+ for (0 s x s 0.38),88 and were able to interpret successfully their measurements of conductance and activation energy as a function of x in terms of the interstitial- vacancy cluster model previously suggested for this system as a result of neutron diffraction measurements.The same group of workers also showed that KBiF4 RbBiF, and the high temperature or p-form of TlBiF had fluorite structures and conductance behaviour similar to but slightly better than Pb0.75Bi0.25F2.25. Tysonite Structures.-There have been comparatively few studies on tysonite type materials except for an investigation of solid solutions based on CeF3 with the aliovalent dopants CaF2 SrF, BaF, and ThF by Takahashi and ~o-workers.*~ In general a tysonite phase was maintained only for x <0.1. The highest conductance was found for (CeF3)o.g5(CaF2)o.05 which had a value of 1S m-l at 200 "C.This value is higher than for the equivalent LaF3 system and similar to that of KF doped PbF,.Since ThF decreased the conductivity it was concluded that ionic current was carried by mobile fluoride ion vacancies the intrinsic defect concentration in pure CeF3 was decreased by the addition of ThF,. Electronic conductance was again low -about 1x lO-'S m-' at 300 "C. A preliminary report has also been madeg0 on the tysonite related BiO,F3_,(x = 0.09 0.10). This was found to have a conductance intermediate between that of Sno.05Lao.05F2.95 and Ca0.27Yo.73P2.73. Conductance data for a number of recently studied fluoride ion conductors are compared with P-PbF and CeF3 in Figure 7. 85 J. D. Pistre J. Saiardene and P. Smutek Electrochim. Acta 1977. 22,693. 86 (a)C. Lucat G. Compet J.Claverie J. Portier J.-M. Riau and P. Hagenmuller Materials Res. Bull. 1976 11 167; (b) J.-M. RCau C. Lucat G. Campet J. Claverie and J. Portier Electrochim. Acta 1977,22,761. 87 A. V. Joshi and C. C. Liang J. Electrochem. SOC. 1977,124 1253. 88 (a) C. Lucat P. Sorbe J. Portier J.-M. RCau P. Hagenmuller and J. Gramec Materiuls Res. Bull. 1977,12 145; (b)J.-M. RBau C. Lucat G. Campet J. Portier and A. Hammou J. Solid State. Chem. 1976,17 123. 89 T. Takahashi H. Iwahara and T. Ishikawa J. Electrochem. SOC. 1977,124,280. 90 J. Schoonman G. J. Dirkson and R. W. Bonne Solid State Comm. 1976,19 783. Solid Electrolytes T 1°C 300 200 100 50 0 -50 +2 I I I I I 1 v -3 --4- I I 1.5 2.5 3.5 4.5 103K T Figure 7 Conductance-temperature behaviour of some fluoride solid electrolytes 11 Co-ionic Conductivity Sodium /?-alumina in which more than half of the sodium ions have been replaced by lithium ions has already been referred to (see Section 7).In this material the lithium transport number is close to unity and the Li' ions can move through the lattice without altering the Na:Li ratio. Farrington and Roth have termed this phenomenon co-ionic conductivity 12,67 Similar behaviour has been observed by Bazin and Schmidt'' for the motion of Cu' in a-AgI and by Lazzari and Scrosatig2 for Cu' in RbAg& Me4NI,6AgI Me3N(CH2)6NMe312 12AgI and possibly for Cu' in Ag3SBr. There is as yet no detailed understanding of the phenomenon but it is likely that this will become one of the more important fields'for future research.91 J. C. Bazln and J. A. Schmidt J. Applied Electrochem. 1976,6,411. 92 M.Lazzari and B. Scrosati Electrochim. Acta 1978,23 75.

ISSN:0308-6003    

DOI:10.1039/PR9777400023

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

4 Luminescence of Organic Solids By J. 0. WILLIAMS Edward Davies Chemical Laboratories University College of Wales Aberystwyth Dyfed 1 Introduction The luminescence of organic solids deals with light emission from the excited states of molecules in crystalline environments. Even in the organic solid state lumines- cence is a vast subject and one which is developing rapidly. Accordingly we have much information about the photophysics and photochemistry of a variety of molecular systems. Studies of luminescence provide a wealth of information about the nature of the excited molecules their structures sizes and shapes their orien- tation lifetime and indirectly their fate when not undergoing radiative decays. Studies of localized and delocalized crystalline excited states (excitons) have gained in importance over recent years and there has been a significant development in studies having direct relevance to the photochemistry of the organic solid state.In this article attention is focused on those luminescence and related phenomena which have made a direct contribution to the understanding of the photochemical properties of molecules in crystalline environments. Attention is given mainly to the behaviour of single crystalline solids but studies of molecules in various other crystalline environments e.g. in glasses at low temperatures and in well defined molecular arrays e.g. monolayer assemblies are considered. Naturally some emphasis is placed on photophysical phenomena insofar as they are essential to an understanding of the chemical processes e.g.determination of energy levels and lifetimes of excited states. It is well accepted that photoreactivity is in direct competition with luminescence from both singlet and triplet excited states and together with other non-radiative decay processes and energy transfer accounts for the disappearance of the excited molecules. For the chemist to understand fully the photochemistry of organic solids information must be sought on the relationships between luminescence and crys- talline structure and whether reactivity occurs at the site of photon absorption or at a distant point in the crystal following energy transfer. To understand the energy transfer properties attention must be given to the generation and transport of excitons and the role of impurities and structural defects in terminating the exci- ton’s migration i.e.exciton localization or trapping. Furthermore molecules in crystals cannot exist or react in isolation and the role of diffusional reorien- tational and translational motions in governing the disposition of a molecule relative to its nearest neighbours is crucial. Often such effects have important repercussions in the luminescence properties and may thus be conveniently investigated using steady-state time-dependent and time-resolved techniques. It is 51 52 J. 0.Williams therefore the intention of this Report to review recent advances in the field of luminescence which assist in the understanding of photoreactivity in crystalline environments.This necessarily leads to shortcomings in some areas and overem- phasis in others. This is accepted as inevitable; indeed in part it is deliberate. There is no attempt to discuss experimental techniques in detail except when these are specific to the study of organic solids or provide a significant advance or have been inadequately treated in earlier reviews on luminescence. Clearly the importance of the development and applications of laser sources to steady state and transient luminescence studies cannot be over-emphasised and this together with improvements in detection and recording systems have opened up exciting possi- bilities e.g. two-photon and picosecond spectroscopy in luminescence studies. Indeed it is significant that the majority of studies on unimolecular systems over the past year utilize the narrow linewidths ca.1cm-’ and short-time resolution s5 ns currently available with such equipment. At the outset and for an essential background the reader is referred to the many recent books and reviews dealing with luminescence in general,’-9 excitonic behaviour’@-20 and the more photophysical aspect^.'^-^' The series of bibliog-R. B. Cundall and T. F. Palmer Ann. Reports (A),1973 70 31. J. Bourdon and B. Schnwiger in ‘Physics and Chemistry of the Organic Solid State’ eds. D. Fox M. M. Labes and A. Weissberger Wiley New York 1967 Vol 3. ‘Proceedings of the Michael Kasha Symposium’ J. Phys. Chem. 1976,80,2143. M. W. Windsor in ‘Physics and Chemistry of the Organic Solid State’ eds.D. Fox M. M. Labes and A. Weissberger Wiley New York 1965 Vol. 2. J. Jortner S. A. Rice and R. M. Hochstrasser Adv. Photochem. 1969,7 149. ‘J. G. Calvert and J. N. Pitts jun. ‘Photochemistry’ Wiley New York 1966. ’ ‘The Triplet State’ ed. A. B. Zahlan Cambridge University Press 1967. B. Stevens Adv. Photochem. 1971,8 161. D. J. Trecker in ‘Organic Photochemistry’ ed. 0.L. Chapman Marcel Dekker New York 1969 Vol. 2 p. 63. lo A. S. Davydov ‘Theory of Molecular Excitons’ Plenum Press New York 1971. G. Fischer in ‘Transfer and Storage of Energy by Molecules’ eds. G. M. Burnett A. M. North and J. N. Sherwood Wiley London 1974 Vol. 4 p. 1. l2 Proceedings of the International Conference of Luminescence’ Tokyo Japan 1975 J. Luminescence 1976 14. l3 D.P. Craig and S. H. Walmsley ‘Excitons in Molecular Crystals’ Benjamin London 1968. l4 R. S. Knox ‘Theory of Excitons’ Solid State Phys. Supp. 1963 5. l5 S. A. Rice and J. Jortner ‘Physics and Chemistry of the Organic Solid State’ eds. D. Fox M. M. Labes and A. Weissberger Wiley New York 1967 Vol. 3 p. 199. ’‘ G. W. Robinson Ann. Rev. Pbys. Chem. 1970,21,429. l7 M. R. Philpott Adv. Chem. Phys. 1973,23 227. l8 R.A. Silbey Ann. Rev. Phys. Chem. 1976 27 203. l9 R.M. Hochstrasser Ann. Rev. Phys. Chem. 1966 17 457. *O C. E. Swenberg and N. E. Geacintov ‘Organic Molecular Photophysics’ ed. J. B. Birks Wiley New York 1973 Vol. 1 p. 489. 21 J. B. Birks ‘Photophysics of Aromatic Molecules’ Wiley London 1970. 22 ‘Organic Molecular Photophysics’ ed. J. B.Birks Wiley London 1973 Vol. 1. 23 ‘Organic Molecular Photophysics’ ed. J. B. Birks Wiley London 1975 Vol. 2. 24 J. B. Birks Reports Progr. F’hys. 1975 38 903. 2s ‘Creation and Detection of the Excited States’ ed. A. A. Lamola Marcel Dekker New York Vol. 1 Parts A and B. 26 K. F. Freed Topics in Current Chem. 1972,31 105. ’’ S. P. McGlynn T. Azumi and M. Kinoshita ‘Molecular Spectroscopy of the Triplet State’ Prentice Hall New Jersey 1969. *’ S. Kimel and S. Speiser Chem. Rev. 1977 77 437. 29 See articles by D. Phillips and K. Salisbury in ‘Photochemistry’ ed. D. Bryce-Smith (Specialist Periodical Reports) The Chemical Society London 1977 Vol. 7. 30 P. M. Rentzepis Adv. Chem. Phys. 1973 23 189. Luminescence of Organic Solids 53 raphies on the spectroscopy of molecular crystals covers the literature up to 1975.31-33 2 Anthracene and Other Polynuclear Aromatic Hydrocarbons.Steady State Studies.-Over the past five years photophysical studies on anthracene single crystals have concentrated on an understanding of the various excitonic properties of this material whilst chemists have been occupied in unravel- ling the details of the various photochemical processes that may occur following irradiation by U.V. light e.g. photo-oxidation and photodimerization. It now tran- spires that the chemist interested in photoreactivity and the impurity and defect properties of anthracene finds himself in a much stronger position than the physicist who is primarily concerned with properties that are characteristic of the pure perfect material-a situation still not attainable with this or any other organic molecular solid.Indeed the unification of photophysical and photochemical prop- erties may be achieved by recognising the undisputed role of structural defects and chemical impurities in controlling such properties. Because of anthracene’s sensi- tivity to light and the generation of photoproducts it is extremely difficult to obtain high purity single crystals. Accordingly much of the luminescence behaviour of single crystals may be interpreted in terms of emission from impurity molecules and anthracene molecules situated at or near structural defects and surrounding impurities. Whilst the importance of structural defects and impurities was recog- nised in the early studies of Helfrich and Lip~ett~~ it is only in and Wolf et ul.,35736 recent years that these effects have been quantified.The papers by Lyons and Warren,37 and Bridge and Vincent3* are important contributions and the role of defects in the luminescence of organic solids has been reviewed by Williams and Thomas39 (up to 1972). Subsequent work has clearly identified the emission characteristics of anthracene crystal? containing various impurities and separated the effects of such impurities from those of structural defects. The incorporation of ‘guest’ chemical impurity molecules into ‘host’ anthracene crystals to form mixed crystals has been studied extensively by means of spec-troscopic and crystallographic techniques. The determining factors in controlling the nature of the resulting mixed crystal are the relative molecular geometry symmetry and sizes of the host and guest species on the one hand together with their relevant electronic properties on the other.These in turn affect both the structure and behaviour of the mixed crystals. Indeed it is known for many such systems that incorporation of the guest into the host occurs by substitution at sites in the ordered lattice up to a certain solubility limit. Above such a limit guest molecules enter by formation of clusters or they concentrate at internal surfaces in 31 E. F. Sheka V. S. Makarova and E. D. Simonovskaya Mol. Cryst. Liq. Cryst. 1975 30 239. 32 E. F. Sheka V. S. Makarova and E. D. Simonovskaya Mol. Cryst. Liq.Cryst.1976 33 261. 33 E. F. Sheka V. S. Makarova and E. D. Simonovskaya Mol. Cryst. Liq. Cryst. 1977,39,259. 34 W. Helfrich and F. R. Lipsett J. Chem. Phys. 1965 43,4368. 35 E. Glockner and H. C. Wolf Z. Naturforsch 1969 24a 943. 36 K. W. Benz W. Hacker and H. C. Wolf 2.Naturforsch. 1970 25a 657. 37 L. E. Lyons and L. J. Warren Austral J. Chem. 1972,25 141 1-1427. 38 N. J. Bridge and D. Vincent J.C.S. Faraday 11 1972 68 1522. 39 J. 0.Williams and J. M. Thomas ‘Surface and Defect Properties of Solids’ eds. M. W. Roberts and J. M. Thomas (Specialist Periodical Reports) The Chemical Society London 1972 Vol. 2 p. 229. J. 0.Williams the host crystalsprovided by such structural imperfections as dislocations and grain boundaries. Two extreme cases of mixed crystals exist so far as prompt fluores- cence studies are concerned depending on the relative magnitude of the excited singlet state energies (&).When the Sl level of the guest is lower than the bottom of the host exciton band e.g. tetracene and pentacene4' as guest molecules in anthracene there is usually efficient energy transfer (see later) to the guest result- ing in an emission spectrum that is characteristic of the guest. Such a situation leads to the chemical traps (see Table 1) for the singlet exciton. In the other case e.g. naphthalene and carbazole in anthracene where the S energy of the guest is higher than the top of the host exciton band trapping is most efficient at host molecules perturbed by the guest (X-traps). The systems have been studied extensively but recently an intermediate situation has also been shown to exist in the cases of perdeuterio anthracenef' and a~ridine~~ in anthracene.Here the S of the guest molecules falls within the exciton band of the host resulting in an amalgamated exciton band. The present situation for various impurities in anthracene is sum- marized in Table I which includes information about the energy states of various molecules in relation to both lowest singlet and triplet states. Two obervations deserve comment. The first is that for the 1-amino 2-amino and 2-hydroxy derivatives there are two values of the trap depth below the singlet level arising from the two symmetry-related orientations of these molecules in the anthracene latti~e.~~'~~ The second observation is that 2-hydroxyanthracene and Table 1 Energy states" for impurity molecules below the lowest singlet and triplet states of anthracene Impurity Energy Below S1at 25 095 an-' Energy Below Tl at 14 738 cm-' /cm-' /cm-' 1-aminoanthracene 1A 3927 2-aminoan thracene (1 NH2A) 13 2A 3771 3296 2- hydroxyanthracene (2NH2A) 2B 2A 3017 938 31 1 (20HA) 2-methylanthracene 2B 751 181 73 (X-trap) 132 (X-trap) (2MeA) 9-hydroxyanthracene 9-methylanthracene 1970 672 tetracene 4849 perylene 2840 mronene 3046 pyrene 59 pentacene 8698 See refs.37 38,40,42 43. * These energy values correspond to origins in the molecular emission spectra of the impurities. The difference between host origin and impurity origin corresponds to the trap depth.Wavenumbers are referred to vacuum. 40 A. Brillante and D. P. Craig J.C.S. Furuduy ZI 1975 71 1457. 41 E. Glockner and H. C. Wolf Chern. Phys. 1975,10,479. 42 J. 0.Williams and B. P. Clarke J.C.S. Furaduy ZI 1977,73 1371. 43 N. J. Bridge and L. P. Gianneschi J.C.S. Furuduy ZZ 1976,72 1622. Luminescence of Organic Solids 2-methylanthracene induce chemical traps in the singlet manifold but yield X-traps for the triplets. The situation for trapping centres introduced as a result of the presence of structural defects is summarized in Table 2. This information has been derived from the work of Craig and Rajikan,44 Lisovenko Shpak and Antoni~k,~~ Williams et uZ.,~~and Williams and Clarke.42 The situation here is not quite so clear as it is for the chemical traps since the introduction of a particular type of dislocation or a specific displacement of molecules in a unique fashion is difficult for molecular crystals.Invariably the most careful deformation induces more than one slip system and therefore several families of dislocations and indeed as has been shown recently deformation introduces a new polymorphic modification into the anthracene Furthermore since such defects are sensitive to crystalline condition and environment the method of crystal growth appears to have a crucial effect on the type of emission behaviour observed e.g. the emission with AEs ca. 275 cm-' appears to be in a single maximum4* in melt grown crystals whereas in thin vapour grown crystals two sharp maxima with AEs of 235 cm-' and 275 cm-' appear.48 It is also significant that similar traps are observed for molecules at dislocations and molecules surrounding impurities e.g.at 275 cm-' for singlets and 130cm-' for triplets. During the past two years studies on triplet exciton trapping in molecular crystals have been few. However Williams and Zboinski have re-evaluated the delayed fluorescence behaviour of anthracene They again find structural imperfections to exert a controlling influence and conclude that even in the most pure anthracene crystals available (impurity and defect content 2:lo-'' M M-') the triplet exciton is not free at room temperature. Table 2 Trapping centres introduced into the host anthracene structure as a result of the presence of structural defects and impurities Type of Defect Depth Below (AEs) Depth Below (AET) /cm-/cm-Single anthracene molecular 73 (20HA) displaced from regular 247 235 129 lattice sites by chemical 261 275 132 (2MeA) impurity at a structural 277 defect uiz.dislocations 320 and polymorphic modifi- 1450 cation. 1565 Pairs of anthracene mole-quasi-continuum extending cules in various relative ca. 4040 ca. 4895 orientations producing an the above two levels are prob- from ca. 4700. excimeric type emission ably part of a quasi continu- um extending downwards from ca. 7000 cm-' 44 D. P. Craig and J. Rajikan J.C.S.Faraday ZZ 1978 74 292. 45 V. A. Lisovenko M. T. Shpak and V. G. Antoniuk Chem. Phys. Letters 1976,42 339. 46 J.0.Williams B. P. Clarke J. M. Thomas and M.J. Shaw Chem. Phys. Letters 1976 38,41. 47 W. L. Rees M. J. Goringe G. M. Parkinson J. 0.Williams and J. M. Thomas Zmt. Phys. Conf. Ser. No. 36 1977 Ch. 10. 48 J. 0.Williams and Z. Zboinski J.C.S. Furaduy IZ 1978,74,618. 56 J. 0.Williams The routes controlling triplet exciton decay are extrinsic via trapping sites. Their observations are interpreted in terms of perturbed anthracene molecules both as separate entities and in pairs associated with dislocations and a polymorphic modification that yields a quasi-continuum of triplet trapping levels extending downwards from ca. 4700 cm-’. This situation is similiar to that observed for the singlet exciton (see Table 2). Together with advances in the experimental studies of singlet and triplet exciton trapping centres there has been an encouraging increase of interest in the relevant theoretical aspects.The effect of strain on the exciton properties of molecular crystals has been examined by S~hipper.~’ It was found that such exciton parameters as the site shift and transfer terms are sensitive to changes in the lattice parameters of the crystal. Schipper also developed a theoretical approach to the treatment of surface states of molecular crystal~.~’ A surface state lying below the crystal band is a potential exciton trap with repercussions in enhanced reactivity in surface regions or in fluorescence from the surface. Craig Dissado and Walmsley” have also treated exciton trapping and self-trapping in molecular crystals.The traps arise from chemical traps formed by foreign molecules physical traps formed by structural disturbance of the host crystal and self-traps. The self-traps occur when the resonance transfer is slow compared with local lattice relaxation. The model developed which applies in varying degree to all three types is an initial excitation to a delocalized exciton state followed by degeneration into a trap state at a rate governed by lattice forces and dependent on exciton-phonon coupling. Klafter and J~rtner~~ have investigated the effects of structural disorder on the absorption line shapes of the lowest exciton states in molecular crystals at zero temperature. The disorder considered is present in crystalline organic solids in the form of vacancies structural defects and strain inhomogeneities.When the lowest exciton state is perturbed by disorder scattering the following features of the optical absorption line shape are exhibited at T = 0 K (i) finite linewidth (ii) asymmetric line broadening of the absorption line shape (iii) the high energy portion of the absorption line is quasi-Lorenzian and (iv) the low energy edge is sharp. Dimen- sionality effects are also considered. To supplement the existing data on the experimental determination of the nature of various structural defects in molecular crystal~~~~~~ several attempts have been made to theoretically compute the defective structures. By using the atom-atom intermolecular potential approach based on Buckingham type potentials which has been developed by Kitaigorodskii and Williams for the determination of the stable crystal structures of molecular crystals Craig Ogilvie and Reynolds” have shown that there are at least three crystal structures of anthracene close in energy to the observed P2,/a phase.Bulk concentrations of misoriented molecules up to 10-2-10-3 mole per mole at room 49 P. E. Schipper Mol. Cryst. Liq. Cryst. 1974 28,401. 50 P. E. Schipper Mol. Phys. 1975 29 501. D. P. Craig L. A. Dissado and S. H. Walmsley Chem. Phys. Letters 1977,46 191. 52 J. Klafter and J. Jortner Chem. Phys. Letters 1977,50 202. s3 J. M. Thomas and J. 0.Williams in ‘Surface and Defect Properties of Solids’ ed. M. W. Roberts and J. M. Thomas (Specialist Periodical Reports) The Chemical Society London 1972 Vol.1 p. 129. 54 J. 0.Williams Sci. Progr. 1977 64 247. 55 D. P. Craig J. P. Ogilvie and P. A. Reynolds J.CS. Faraday Zl 1976 72 1603. Luminescence of Organic Solids 57 temperature are found by crude calculations and are associated with vacancies and dislocations. The misoriented molecules form strings of 10-100 units along the [Ol 01 direction. Similarly Mirsky and Cohens6 have calculated the energies involved with the creation of several dislocation families in anthracene. They confirm that dislocations of the type (201)3[1021 (001)$[ 1001 (001)$[ 1lo] and (001%[IT01 lead to minimum energy configurations. Recently Ramdas et af.57 and .~~ Parkinson et ~1have combined the use of transmission electron microscopy the theoretical method outlined above and fluorescence techniques to investigate the formation of a distinct polymorph of anthracene under mechanical stress.The new phase produced by compression along a direction approximately perpendicular to the basal (001) faces is triclinic and belongs to the Pi space group with unit cell dimensions a =8.6& b =6.0& c = 11.2 A a = 123” p = 113.5” and ?=82.5”. This new modification yields excimeric emission of long lifetime and has important repurcussions in the photodimerization of anthracene crystals (see later). Such a crystalline modification introduced by compression and where the molecules are in a different mutual orientation with respect to the normal P2Ja structure has been but has implicated in studies on anthracene under press~re,~~’~~ never been conclusively identified.Time-dependent Studies.-Time-dependent prompt and delayed fluorescence measurements yield information about energy transfer involving singlet and triplet states respectively. Much work was done in this area in the late 60’s and early 70’s and with respect to singlet excitons in organic solids this was reviewed by Powell and Soos in 1975.61 In their review article Powell and Soos consider the usefulness of different types of experimental techniques in elucidating the characteristics of both exciton migration (diffusion) and trapping (mainly by impurities i.e. chemical traps). The effects of radiative reabsorption on the results obtained by different types of measurements are also discussed.Such considerations are crucial to any consideration of the chemical reactivity of organic solids. Energy transfer in the triplet state has received less attention lately but particular aspects of the early work have been reviewed by Avakian and Merrifield,62 Ern,63 Wolf,64 and Williams and Thomas.39 Indeed singlet-singlet exciton transfer is more relevant to an understanding of the photochemistry of anthracene than the corresponding triplet- triplet transfer and we shall therefore concentrate on the former and summarize the present position. Host-sensitized energy transfer consists of two distinct physical mechanisms migration (or diffusion) through the host and trapping at some specific site. Thus two primary experimentally measured parameters the energy transfer rate k (t) and the trapping time (T),are necessary to characterize this process and these may be 56 K.Mirsky and M. D. Cohen J.C.S. Furuduy ZI 1976,72,2155. 57 S. Ramdas J. 0.Williams J. M. Thomas G. M. Parkinson and M. J. Goringe ‘Eighth Molecular Crystal Symposium’ Santa Barbara U.S.A. 1977. 58 G. M. Parkinson M. J. Goringe S. Ramdas J. M. Thomas and J. 0.Williams J.C.S. Chem. Comm. 1978,134. 59 M.Nicol M. Vernon and J. T. Woo J. Chem. Phys. 1975,63 1942. 6o M.Nicol J. Phys. Chem. 1976,80,2200. 61 R.C. Powell and Z. G. Soos J. Luminescence 1975,11 1. 62 P. Avakian and R. E. Merrified Mol. Cryst. Liq. Cryst. 1968 5 37. 63 V. Ern Mol. Cryst. Liq. Cryst. 1972 18 1. H. C. Wolf in ‘Advances in Atomic and Molecular Physics’ Academic Press New York 1967 Vol.3. 58 J. 0.Williams determined under conditions that eliminate reabsorption by recording the time- dependent emission of both host and impurity. This includes analysis of both the rise and decay of the luminescence in response to a short-pulse excitation (S1ns). Such values of the energy transfer rate must agree with values of the diffusion constant measured by independent means and corrected for reabsorption. Energy transfer can occur either radiatively or radiationlessly. The former process consists of the emission of a photon and its subsequent reabsorption by another molecule whereas the non-radiative process between two impurity molecules in a crystal can be described as a quantum mechanical ‘resonance’ interaction between the elec- tromagnetic multipole radiation fields of the two molecules which results in the exchange of a virtual photon.For singlet-singlet transitions the most important type of interaction mechanism is dipole-dipole and this type of energy transfer is referred to by several different names including quantum mechanical resonance inductive resonance Forster-Dexter transfer and long range resonance transfer (LRRT). Exciton migration can be described as a random walk and in the limit of many steps this can equivalently be described by a diffusion formalism. The form of the diffusion coefficient depends on the size of the mean free path. If the mean free path’ is of the order of one lattice spacing the exciton hops incoherently through the lattice and is scattered at each molecule.The diffusion coefficient is then expressed in terms of the lattice spacing and hopping time. If the mean free path is greater than the lattice spacing the exciton moves coherently over several lattice spacings before being scattered and it is more appropriate to use the band model. The diffusion coefficient can then be expressed in terms of the mean free path and the exciton velocity. For singlet exciton migration in typical organic solids such as anthracene at room temperature the incoherent hopping model is generally thought to be appropriate. In ultra-pure crystals at low temperatures where neither impurities nor phonons are effective in restricting the exciton’s motion it has been possible to detect some coherent exciton motion.Interest in problems of coherence in molecular crystals is increasing rapidly (see Section 8 p.72). The rise times of activator (impurity) fluorescences reported in the literature (see data in ref. 61) are not consistent with a simple diffusion transfer model but as pointed out by Powell and Soos require a faster initial transfer rate as also found in the time dependent studies. The problems involved in singlet energy transfer in molecular crystals have also been considered in some detail by Birk~.~’ For tetracene- and perylene-doped anthracene crystals it had been shown earlier by and Gammill and that the rise time of the activator fluorescence is faster than the decay time of the host fluorescence. The time-dependent host- guest energy transfer rate thus led to a generalized random-walk which is identical with a time-dependent diffusion controlled process described by the Smoluchowski relation.For tetracene-doped anthracene crystals such a model yields values of 0,the diffusion coefficient of 1.6-9.6 X lo-’ cm2s-’ which are one or two orders of magnitude less than previously accepted values and values of R the energy transfer distance of 60 A-100 &much greater than the previously 65 J. B. Birks Mol. Cryst. Liq. Cryst. 1974 28 117. 66 R.C. Powell Phys. Rev. 1970 B2 1159. 67 L.S. Gammill and R. C. Powell Mof. Cryst. Liq. Cryst. 1974 25 123. Z. G. Soos and R. C. Powell Phys. Rev. 1972 B6 4065. Luminescence of Organic Solids assumed collisional distance of about lOA.Thus our ideas of mixed organic crystals were inadequate and the addition of impurities like tetracene introduces extended regions of guest-induced defects (see earlier) in the vicinity of each impurity molecule. This situation has now been verified for many impure organic molecular crystals as has been discussed earlier (see for example refs. 40 and 42) and it may be accepted that the addition of impurities to an anthracene crystal in excess of a very small concentration (~1 p.p.m.) introduces so much structural disorder that there appears to be little relation between the photophysical prop- erties of undoped and tetracene-doped anthracene crystals. Thus Birks suggested three methods of obtaining the information about ‘pure’ crystals sought from mixed crystal studies.These were (i) to study energy transfer from the host crystal to its ‘natural’ guest impurities e.g. 2-OHA and 2-MeA at low temperatures (ii) to use isomorphic guest molecules e.g. deuteriated species (iii) to study exciton-exciton interactions in ‘pure’ crystals using laser excitation to generate a high initial exciton concentration. Since it is likely that in a ‘pure’ anthracene crystal the host-guest interaction is probably collisional with R =6-10 A and the relatively low concentration of exciton traps increases D to ca. 10-4cm2s-1 time resolution of the order of picoseconds (10-50 ps) is required to resolve the energy-transfer problems involving singlet excitons in anthracene and many other organic molecular crystals.It is thus not surprising that such studies have not been carried out during the past five years -only with the further development of picosecond spectroscopy and its application to organic solids will further progress be made. Nevertheless time dependent prompt fluorescence studies have been performed on anthracene crystals with a time resolution of ca. 1ns and some useful informa- tion has been obtained. Galanin et d9using a single-photon counting technique have measured fluorescence decay times of anthracene crystals at room tempera- ture and 4.2 K. Different decay times are found corresponding to different emis- sion lines at the lower temperature. It is concluded that the intrinsic excitonic luminescence occurs in a two-step way. Excited and thermalized excitons have a lifetime of 1ns.The emission from these excitonic states probably free excitons takes place together with energy transfer to the other crystal excited states (local- ized excitons or self-trapped excitons) with a lifetime of about 3.5 ns. The energy transfer to impurities probably takes place from both states and the resulting emission lifetime is longer ca. 5-7 ns. Similar measurements have been per- formed by Bale Bridge and Smith” who conclude that the radiative lifetime of free excitons in anthracene is less than 1.911s at 5.4K.This is attributed to the super-radiant behaviour predicted for states at the bottom of the exciton band. Guest origin bands corresponding to 20 HA have decay times of ca. 8.0 ns and the broad emission with a maximum at ca.22 830 cm-I attributable to structural defects and a decay time of ca. 6 ns. Williams et ~l.,~~ who introduced dislocations into anthracene crystals thus increasing the intensity of the defect origins at 24 770 cm-’ and 23 528 cm-I recorded decay times of ca. 7 ns for both emissions. 69 M. D. Galanin Sh. D. Khan-Magometova Z. A. Chizhikova M. I. Demchuk and A. F. Chernyavskii J. Luminescence 1975,9 459. ’O A. G. Bale N. J. Bridge and D. B. Smith Chem. Phys. Letters 1976 42 166. 60 J. 0. Williams The importance of performing such studies with picosecond time resolution may be considered by reference to recent work on tetracene crystals doped with penta~ene.~~ In this system the rise-time of the guest fluorescence is the same as the decay time of the host emission.For low pentacene concentrations the exponential decay of the host at 170K is consistent with a diffusion model for singlet migration from host to guest. At high intensities no evidence was found for guest saturation because of the dominating effect of bimolecular exciton anni- hilation in the host. Rise and decay times were ca. 260 ps at 170 K. Previously the emission in ‘impure’ tetracene crystals with a dual exponential decay uiz. 12 ns and 0.8ns had been attributed by Smith and Weiss7* to different impurity species. Experimentally naphthalene crystals are easier to investigate than anthracene and tetracene since the prompt fluorescence lifetime lies in the range 20 ns to 150 ns. Recently the temperature dependence of the fluorescence decay times of pure and anthracene-doped naphthalene single crystals has been reported by Kohler Schmid and Wolf.73 Above 60 K all decay curves were single exponentials and the time dependence was analysed in terms of a simple hopping model.Below 60 K a biexponential decay was observed and this is attributed to additional shallow traps effective at these low temperatures. The results do not require the postulation of a long range resonance transfer of energy. A discussion of the binary solid solutions of naphthalene in perdeuterio-naphthalene and arithracene in perdeuterio-anthracene is given by Wolf and and the same authors employ p-halo- genated naphthalene molecules in naphthalene crystals to study energy transfer processes and the geometrical and electronic structure of Direct trap to trap energy transfer is observed at low concentration of the dopant and it has been possible using e.s.r.to measure the misorientation of the naphthalene X-traps introduced. The fluorescence of anthracene crystals has been studied over a wide range of nitrogen laser optical pumping by Avanesjan et af.76Lattice defects are found to play a crucial role in the process of exciton-exciton annihilation. In relatively perfect crystals exciton concentrations of -2 x 10’’ cm-3 may be achieved with pumping levels of -2 x cm-2 s-l. Energy transfer due to triplet excitons has also been investigated by monitoring the sensitized delayed fluorescence from ‘real’ crystals following the earlier studies outlined by Wolf and Ben~.~~ Naphthalene crystals doped with anthracene have been studied7’ and work on anthracene doped with tetracene has been reported by Zschokke-Granacher and co-w~rkers.~~ The problem of whether the triplet exci- ton should be treated as a collective state of the crystal (coherent description) or rather as an incoherent hopping of molecular excitations still remains unclear for ” A.J. Campillo S. L. Shapiro and C. E. Swenberg Chem Phys. Leffers 1977 52 11. 72 A. W. Smith and C. Weiss Chem. Phys. Letters 1972 14 507. 73 M. Kohler D. Schmid and H. C. Wolf J. Luminescence 1976,14,41. 74 H. C. Wolf and H. Port ‘Molecular Spectroscopy of Dense Phases -Proceedings of the 12th European Congress on Molecular Spectroscopy’ France 1975 Elsevier Amsterdam 1976.” H. C. Wolf and H. Port J. Luminescence 1976,12/13 33. 76 0. S. Avanesjan V. A. Benderskii V. Kh. Brikenstein V. L. Broude L. J. Korshunov A. G. Lavrushko and I. I. Tartakovskii Mof. Cryst. Liq. Cryst. 1974,29 165. 77 H. C. Wolf and K. W. Benz Pure and Appl. Chem. 1971,27,439. 78 H. Port M. Bader G. Weber and H. C. Wolf Z. Naturforsch 1975 30a 277. 79 J. Funfshilling and I. Zschokke Granacher Mol. Cryst. Liq. Cryst. 1974 28 383. Luminescence of Organic Solids 61 several systems. At room temperature it is fairly well agreed that the motion is incoherent but at liquid helium temperatures the situation is often complicated by the presence of traps. As mentioned earlier it is profitable in such cases to study the incorporation of isotopically labelled molecules.In the ['Hs]naphthalene in [*H,]naphthalene system it is possible to focus on excitonic problems and a further attraction is that the triplet state of two adjacent translationally inequivalent ['H'lnaphthalene guest molecules may be studied by optical spectroscopy as well as e.s.r. Davydov components of the pair have been observed in the phosphorescence spectrum and Botter et aZ.,80 on the basis of electron spin echo and optical detection of magnetic resonance techniques conclude that the pair may be considered as a 'mini-exciton' possessing a degree of coherence at low temperatures. In fact this study demonstrates that coherence experiments together with optically detected magnetic resonance prove to be very successful in unravelling the nature of energy delocalization in the naphthalene resonance pair and also make possible the determination of the lifetime and zero field splittings of the two Davydov states (see Section 8 p.72). Singlet exciton states of molecular crystals such as anthracene corresponding to moderately intense transitions require submicron crystals for optical transmission experiments and they typically display broad spectral features which are greatly influenced by defects and surface effects. An improved understanding of excitons has come from reflection studies" (which lie outside the scope of this Report) and recentlys2 parity forbidden two-photon fluorescence excitation spectra of the naphthalene 3200 band were studied in large crystals following earlier The use of non-linear spectroscopy using photons in the 'transparent' regions of large crystals is an exciting development in the probing of the character of bulk states.Experiments so far conducted on naphthalene and anthracene single crys- tal~'~ yield new insights regarding polaritons in these systems. The necessity to treat some of the spectroscopic properties of anthracene within a polariton frame- work has been discussed by Ferguson.'' (A polariton is a quasi-particle resulting from the interaction of photon and exciton.) Photochemistry.-The photochemistry of anthracene in the solid state in particular the photodimerization has been extensively studied for several years. Anthracene may be considered as one of those solids which violate the topochemi- cal principle of Cohen and Schmidts6 in that it yields a stable photodimer upon irradiation with U.V.light in crystals whose predominant structure is not conducive to photodimerization. It is now well accepted that certain types of structural offer easy pathways to the photodimer but it is only recently that the details of the process have been elucidated. Thus over the past few years there has been an accelerated effort to understand the precise nature of structural defects in anthracene on the one hand and to unravel the prompt fluorescence behaviour that 80 B. J. Botter C. J. Nonhoff,J. Schmidt and J. H. van der Waals Chem. Phys. Letters 1976,43 210. 81 see e.g. M. R. Philpott and J. M. Turlet J. Chem. Phys.1976,64 3852. 82 R. M. Hochstrasser and H. N. Sung J. Chem. Phys. 1977,66 3276. 83 R. M. Hochstrasser H. N. Sung and J. E. Wessel Chem. Phys. Letters 1974 24 168. 84 R. M. Hochstrasser and G. R. Meredith J. Chem. Phys. 1977,67 1273. 85 J. Ferguson Chem. Phys. Letters 1975 36 316. 86 M. D. Cohen and G. M. J. Schmidt J. Chem. SOC.,1964 1966. J. 0. Williams serves as a sensitive probe for the photochemistry on the other. The correctness of the assumption that the path of the photodimerization of anthracene in the solid state passes through an excimeric state was established in 1974 by the elegant work of Ferguson and Mau.” It was shown that for a preformed anthracene pair in a crystal of dimer the photodimerization and excimer emission processes are in direct competition.The quantum yield of the emission is unity at low temperatures whereas at high temperature (> 160K) the yield of photodimerization is unity and that of emission essentially zero. A similar situation has been shown to exist for the photodimerization in fluid solution by Cohen Ludmer and Yakhot.88 In their study Ferguson and Maus’ identified two excimeric species (‘red’ and ‘green’ emitting) varying in terms of different mutual orientations of the two molecules. In a dianthracene single crystal host the ‘red’ fluorescent excimer yielded upon thermal activation the stable photodimer whereas the green emitting species was unreactive. Studies of the photochemistry and photophysics of 1,2-di(9-anthry1)ethane (AEA) have allowed Ferguson Morita and Puzas9 to make deduc- tions about the most likely conformations responsible for the different spectral properties of the excimers.Accordingly the species which is responsible for the red emission is compound (l) whilst a likely agent for the green emission is compound (2). n U Later studies by Ferguson Morita and Puzag0 on bridged anthracenes of the type [2n](9,10)-anthracenophanes (n= 2 3 or 4) 1,3-di(9-anthryl)propane and 1,4-di(9-anthryl)butane and by Hayashi et aL9’ on a range of anthracenophanes bring into focus the great variability in excimeric emissions dependent upon only slight modifications in the mutual orientations of the molecular pairs. In the six bridged anthracenes produced by photocleavage of the corresponding photoisomer crystals studied by Ferguson Morita and Puzago at least one non-fluorescent conformation has been found which is transferred into another intermediate which can partition either way to form both ground state isomers.Although focussing on excimeric emission properties and stressing the importance of the surrounding medium (solvent cage) in controlling mutual orientations of molecular pairs a transfer of the above information directly to the photodimerization of anthracene crystals (and the subsequent cleavage of the photodimer) is hindered by the fact that the bridges associated with the intramolecular excimers in the above molecules ’’ J. Ferguson and A. W. H. Mau Mol. Phys. 1974,27 77. ’’ M. D. Cohen Z. Ludmer and V. Yakhot Chem. Phys.Letters 1976 38 398. 89 J. Ferguson M. Morita and M. Puza Chem. Phys. Letters 1976 42 288. 90 J. Ferguson M. Morita and M. Puza Chem. Phys. Letters 1977,49 265. 91 T. Hayashi N. Mataga Y. Sakata S. Misumi M. Morita and J. Tanaka J. Amer. Chem. Soc. 1976 98 5910. Luminescence of Organic Solids 63 restrict the pathway of conformational relaxation following photochemical cleavage of the photoisomer. Thus to understand fully the solid state photodimerization of anthracene experiments have to be performed on single crystals. Such studies have been reported recently by Williams Donati and Thomas,92 Williams and Clarke,42 and Craig and In their study Williams Donati and Thomas92 investigate the fluorescence spectra and lifetimes of anthracene crystals subjected to photodimerization and photo-oxidation.Whereas photo-oxidation does not appreciably change the spectral characteristics of the anthracene crystals even at 4 K photodimerization significantly modifies both the spectrum and the fluores- cence lifetimes of the crystals. The emissions are attributed to anthracene 'incipient' dimers and to single displaced anthracene molecules in close proximity to the stable photodimer. They are respectively long and short lived and may be classed as deep (ca.4000 cm-') and shallow (ca. 300 cm-') traps. The shallow trap is similar to the one previously identified in deformed anthracene in melt grown anthracene crystals doped with carbaz~le,~~ and in most other pure melt grown The deep excimeric emitting trap is similar to the 'green' exci- meric species identified by Ferguson and Mau.*' In their study Craig and Rajikan44 concentrate on the effect of photodimerization on the shallow defect-induced traps revealed in the low-temperature fluorescence spectrum.Thus they are probably monitoring the effect of photodimerization on the surrounding lattice rather than investigating the species that are directly involved in the reaction. Most of their X-trap emissions e.g. at 24 07 1 cm-' (AE,) = 22 cm-') 24 848 cm-' (A& = 247 cm-') 24 818 cm-' (A& = 277 cm-') 24 581 cm-' (AEs= 514 cm-') and 24 426 cm-' (A& = 669 cm-') disappear and new fluorescence emission lines appear at 25 064 cm-' (A& =31 cm-') -24 989 cm-' (A& = 106 cm-') 24 984 cm-' (A& = 1 I1 cm-') 24 972 cm-' (AEs= 123 cm-') and 24 914 cm-' (AE = 181 cm-').Significantly in crystals exhibiting emission at 24 834 cm-' (A&= 261 cm-') irradiation does not cause new emission lines to appear; the changes then are a broadening of the bands and an increase in the emission from the trap at AEs =261 cm-'. This accounts for the major difference between the response of thin vapour and thick melt grown crystals to irradiation and to a large extent accounts for the minor differences reported by the two research groups. Despite the lack of a definite assignment to many of the various trapping centres and the precise molecular orientation involved there is overall agreement as to the mechanism of anthracene photodimerization particularly when the role of the newly discovered polymorphic modification of anthra~ene~~~~~ is considered.This triclinic P1 modification facilitates photodimerization by providing pairs of neigh- bouring anthracene molecules both within its own structure and at the interfaces with the normal P2Ja structure that are conducive to photodimerization. The emission characteristics of anthracene crystals containing a large proportion (ca. 10%) of this new modification exhibit a range of excimeric emissions in the wavelength range 450-550 nm with lifetimes ranging from ca. 85 ns-175 ns. Steady state and temporal dependences of delayed fluorescence as a function of temperature for a range of high purity anthracene crystals before and after pho- todimerization confirm that deep triplet traps extending downwards from ca.92 J. 0.Williams D. Donati and J. M. Thomas J.CS. Furuduy ZI 1977 73 1169. 93 D. P. Craig and J. Rajikan Chem. Phys. Letters 1977,47,20. 64 J. 0. Williams 4700 cm-' are removed upon prolonged U.V. irradiation whereupon other shal- lower traps appear.48 The mechanism of photodimerization therefore involves the following. Initiation occurs at structural defects with an induction period in which the total intensity of prompt fluorescence remains practically constant. With further irradiation photodimerization leads to a decrease in the fluorescence intensity. A disappearance of the deep singlet traps provided by incipient dimers occurs during the induction period and later a generation of additional deep traps to propagate the reaction (which is known to be autocatalytic) occurs.Shallower traps are continually being formed and consumed during the reaction following molecular motion and/or diffusion at room temperature. In impure anthracene crystals the formation of the hetero-photodimer formed when a molecule of the host combines with an impurity competes with the generation of the homo-photodimer. Such processes have been monitored by Craig and Rajikan44 and Williams and Clark.42 Craig has used the term 'site selective photochemistry' to describe the preferential reactivity at one of the two sites occupied by the 2-OHA impurity. 3 9-Cyanoanthracene Pyrene and Related Excimeric Emitting Systems 9-Cyanoanthracene (9-CNA) is one of several substituted anthracenes which if perfect have quasi one-dimensional stacks of molecules packed face-to-face with a mirror symmetric (cis) orientation with respect to the CN group.In 9-CNA the intermolecular spacing along the stack is 3.93 8 and therefore according to the topochemical preformation theory photodimerization should occur to give the cis-photodimer. However the fact that in the solid state (and in solution) reaction produces the centrosymmetric photodimer in quantitative yield led to the proposals that the reaction is initiated at certain types of dislocations.94995 The situation regarding the structure sensitivity of this and many other related solid state photo- chemical reactions has been reviewed In this review we shall concentrate on the luminescence properties of such systems and how these relate to photochemical reactivity.The cis-geometry of molecular stacks as in the 9-CNA structure should favour excimeric emission from the association of pairs of such molecules. Several spectroscopic studies have been made of the excimeric emission and of the emission originating from photochemical products -the earliest being those of Stevens and Dickin~on~~ who showed that at room temperature a broad excimer-like emission could be converted by U.V. irradiation into a blue somewhat structured emission. This blue emission was due to 9-CNA molecules emitting in a matrix of photodimer~.~~ Subsequently Sarti-Fantoni and Teronilm performed similar experiments and Cohen et a1.'" repeated these measurements and found the green emission to be excimer-like and to shift substantially to lower energy y4 M.D. Cohen Z. Ludmer J. M. Thomas and J. 0.Williams Chem. Comm. 1969 1172. 95 M. D. Cohen Z. Ludmer J. M. Thomas and J. 0.Williams Roc. Roy. SOC.,1971 A324,459. 96 J. M. Thomas S. E. Morsi and J. P. Desvergne Adv. in Phys. Org. Chem. 1977 15,63. 97 M. D. Cohen and B. E. Green Chem. in Britain 1973,9 490. '* B. Stevens and T. Dickinson Speccrochim. Acta 1963 19 1865. 99 B. Stevens T. Dickinson and R. R. Sharpe Nature 1964 204 876. loo P. Sarti-Fantoni and R.Teroni Mol. Cryst. Liq. Cryst. 1970,6 431. M. D. Cohen Z. Ludmer and V. Yakhot Phys. Stahts Solidi 1975 b67 51. Luminescence of Organic Solids upon cooling from room temperature to 90K.Recently the polarized emission spectra of melt and vapour grown single crystals of 9-CNA have been reported."* In crystals previously unirradiated at room temperature the emission is excimer-like with a double peak (550 nm E Ic) and (525 nm Ell c) at 2 K and a single peak at 300 K (488 nm E Iand IIc). Exposure to U.V. irradiation induces fluorescence associated with the production of photodimers and at 2 K this emission is highly structured and co-exists with excimer emission. A poorly structured band appear- ing at 490nm is attributed to emission from isolated 9-CNA molecules i.e. X-traps. Recently Ludmerlo3 has reported a new metastable crystal form of 9-CNA showing a red excimeric emission at ca.580 nm. The structure of this excimer is believed to be centrosymmetric.As in the case of the photodimerization of anthracene crystals much work remains to be done on 9-CNA. In particular recent publications have demonstrated that a given compound can give rise to more than one excimer e.g. 9-~hloroanthracene,'~' 2,4-dichloro-3'-methyl-trans-stilbene,'04 a~ridine,~'~~~~~ and that in excimers with infinite stack structures the emission is shifted by change in temperature much more strongly than in structures based on pair-wise packing,"' e.g. pyrene. Accepting that the excimeric state is a precursor to the 9-CNA photodimer the precise geometry and mutual orientation of the molecular pairs involved needs to be known. Both steady state and time-depen- dent luminescence studies will prove invaluable in this respect.As we have seen for anthracene itself (see Section 1 p. 51) studies of the products of photochemical cleavage of dimers and related compounds in rigid glass matrices or in single crystals of the photodimer yield valuable indirect information about the mechanism of both the photodimerization and photodissociation in single crystals. The most extensive studies relating to anthracene derivatives and mixed dimers have been carried out by Ferguson Mau and following the earlier work of Chandross and Fergu~~n'~~~''~ and Chandross Ferguson and McRae.'" The most significant contribution in this area over the past two years is that of Ferguson and Miller112 on a number of 9-substituted anthracenes. Photochemical and photo- physical studies on 9-substituted anthracene sandwich pairs in their corresponding photodimer crystal matrices and in methyl-cyclohexane at 6 K were performed.Photodimerization of the 9-methyl 9-chloro and 9-cyano derivatives in the cor- responding photodimer matrix occurs at 6K with unit quantum yield. The presence of excimer fluorescence from sandwich pairs indicates the lack of a perfect topochemical orientation while activation processes leading to photodimerization involve molecular re-orientation from more stable ground state configurations which are achieved within the constraints imposed by the solvent or crystalline cage. Future studies of the photodimerization in 9-CNA and other substituted lo' R. M. Macfarlane and M. R. Philpott Chem. Phys. Letters 1976 41 33.'03 Z. Ludmer Phys. Status Solidi 1977 a43 695. R. Cohen Z. Ludmer and V. Yakhot Chem. Phys. Letters 1975,34 271. Io5 B. P. Clarke J. M. Thomas and J. 0.Williams Chem. Phys. Letters 1975,35 251. lo' J. 0.Williams B. P. Clarke and M. J. Shaw Chem. Phys. Letters 1976 39 142. lo' J. Ferguson A. W. H. Mau and J. M. Morris Austral. J. Chem. 1973,26 91. lo' J. Ferguson A. W. H. Mau and J. M. Morris Austral. J. Chem. 1973 26 103. E. A. Chandross and J. Ferguson J. Chem. Phys. 1966,45,397. 'lo E. A. Chandross and J. Ferguson J. Chem. Phys. 1966,45,3554. E. A. Chandross J. Ferguson and E. G. McRae J. Chem. Phys. 1966,453546. '12 J. Ferguson and S. E. H. Miller Chem. Phys. Letters 1975,36 635. J. 0. Williams anthracenes should focus on the role of impurities in the energy transfer process the determination of the precise geometry of the molecular pairs responsible for the various excimeric emissions and the identification of the precursor to the photodimer.A thorough understanding of the excimeric state is essential to the understanding of the photochemistry of anthracenic and other related systems. In this respect pyrene has been the material upon which attention has been concentrated both from the experimental and theoretical Hitherto it has been assumed that pyrene crystals contain but one type of molecular pair and such experimental parameters as the shift in excimeric emission with temperature together with corresponding changes in the emission half-width have been accounted for on this basis.The importance of purity and crystalline perfection of the normal monoclinic form has never been in doubt since such parameters are known to reduce drastically the decay time of the excimeric emission. Thus a recent report on a new crystalline modification of pyrene114 confirming earlier work’” may lead to a reappraisal of our general understanding of this material particularly since there appear to be changes in the geometry of molecular pairs in the new form. Indeed the shape and position of fluorescence bands observed in pyrene crystals and pyrenophanes have been interpreted in terms of structural differences in the geometry of the excited state.”6 Port and Mistelberger’” in their study of the temperature dependence of the singlet-triplet absorption in pyrene crystals were able to freeze-in the high temperature modification by fast cooling through the transition temperature.This allowed them to obtain information about energy transfer and exciton-phonon coupling in the triplet state of pyrene and they demonstrated the applicability of an exciton diffusion approach. Little experimental work has been reported on the anthracene derivatives 1,8-dichloro-9-methyl anthracene (1,8-C12-9-MeA) and 1,8-dichloro- 10-methyl anthracene (1 ,8-C12- 10-MeA) since it was demonstrated that dislocations control their phot~dimerization.”~.”~ The main problem involved here concerns the availability of highly purified single crystals. However the results of a compu- tational approach to the study of extended defects in 1,8-C12-9-MeA have been reported by Ramdas Thomas and Goringe.”’ Using the pairwise evaluation of non-bonded interactions approach the extent of structural changes that occur at an extended planar fault such as a stacking fault or antiphase boundary of a pres- cribed type has been estimated.The way in which the fault energy is diminished as a result of allowing the molecules in the vicinity of the fault to relax and the lattice to expand is demonstrated. In particular there are indications that for a (100) fault plane the lowest energy is achieved by incorporating a translation vector of & [250] and a small degree of folding of the constituent molecules in and adjacent to the plane of the fault. Such structural relaxation gives rise to incipient dimer pairs in ‘I3 MID.Cohen E. Klein Z. Ludmer and V. Yakhot Chem Phys. 1974,5 15. W. Jones and M. D. Cohen Mol. Cryst. Liq. Cryst. Letters 1978 41 103. R. Zallen C. H. Griffiths M. L. Slade M. Hayek and 0.Brafman Chem. Phys. Letters 1976,39,85. M. E. Michele-Beyerle and V. Yakhot Chem. Phys. Letters 1977 49,463. ‘I7 H. Port and K. Mistelberger J. Luminescence 1976,12/13 351. J. P. Desvergne J. M. Thomas J. 0.Williams and H. Bouas-Laurent J.C.S. Perkin ZZ 1974 363. J. M. Thomas J. 0.Williams J. P. Desvergne G. Guarini and H. Bouas Laurent J.CS. Perkin IZ 1975 1974. S. Ramdas J. M. Thomas and M. J. Goringe J.C.S. Faraday IZ 1977 73 551. Luminescence of Organic Solids which contiguous molecules oriented in a trans registry and with the G-Clo distance in the range 4.54.7& are 'preformed'.The calculations are in agreement with the earlier experimental observations. Exciting developments in the field of photodimerization of organic solids include the use of such reactions as a method of enantiomeric purification e.g. of 1-aryl ethanols121 and the photochemical reactions in inclusion molecular complexes. 122 4 Stilbenes Styrenes and Polyphenyls Cis-trans isomerizations have figured prominently in attempts to develop a mechanistic theory for many photochemical reactions. From the theoretical point of view these reactions which involve a rotation about a double bond have the advantage of possessing a reaction co-ordinate that can be easily separated from the other molecular co-ordinates.In 1975 it was shown by Orlandi and Siebrand"' that the conventional description of the cis-trans isomerization of stilbene was based on an incomplete potential energy diagram. An alternative description was proposed based on an additional low-energy singlet state (labelled S2).To under-stand the relation of S to the ground state So and the lowest excited singlet state in the planar configuration S1 it was found instructive to describe stilbene in terms of two interacting benzyl radicals in the .rr-electron approximation. Thus instead of the direct and sensitized cis-trans photoisomerization proceeding via a TIstate it was proposed that the S level can account for all known observations. Birch and Birk~'~~ subsequently measured the fluorescence lifetime of trans-stilbene in dilute methylcyclohexane/iso-hexane solution and determined the mean S radiative (kF) radiationless (k,) and cis-trans isomerization (k,) rate parameters from -90-60 "C.S1 consists of a fluorescent trans ('B:) state (kg = 6.0X lo8s-') which undergoes reversible thermal-activated rotational internal conversion (AH= 7.42 kJ mol-' AS = 44.3 J K-' mol-') to a non-fluorescent perp ('A:) state. The p('A8) state lies 610 cm-' above t('B,*) with ail intermediate S potential maxi- mum and undergoes internal conversion (k,= 5.8 X lo8s-') to p('A,) leading to cis-isomerization; this is the main isomerization channel over the whole tempera- ture range. Such a mechanism has been confirmed recently by Teschke Ippen and H~ltom'~~ in their picosecond study of the dynamics of the singlet excited state of trans- and cis-stilbene.In both trans- and cis-stilbene in hexane excited-state absorption is observed to decrease rapidly as the initially excited singlet state relaxes to its conformational equilibrium. Even though an improved structure of trans-stilbene in the solid state has been presented126 there has been little work on the solid state luminescence and isomerization behaviour. For styrenes photo- chemical and photophysical investigations have been performed in the gas phase,'" in solution and in low temperature matrices128 with information lacking for cor- responding crystalline studies. M. Lahav F. Laub E. Gati L. Leiserowitz and Z. Ludmer J. Amer. Gem. Soc.,1976,98,620. M. Lahav L.Leiserowitz L. Roitman and C.-P. Tang J.CS. Chem. Cbmm. 1977,928. ''' G. Orlandi and W. Siebrand Chem. Phys. Letters 1975,30 352. lZ4 D. J. S. Birch and J. B. Birks Chem. Phys. Letters 1976,38 432. 0. Teschke E. P. Ippen and G R. Holtom Chem. Phys. Letters 1977,52,233. C. J. Finder M. G. Newton and N. L. Allinger Acta Cryst. 1974 B30,411. ''' R. P. Steer M. D. Swords P. M. Crosby D. Phillips and K. Salisbury Chem. Phys. Letters 1976,43 461. ''13 P. M. Crosby and K. Salisbury J.C.S. Chem. Comm 1975,477. 68 J. 0. Williams The situation is a little improved for certain polyphenyls particularly biphenyl and p-terphenyl. The influence of molecular geometry on the fluorescence spectra of biphenyl and p-terphenyl in rigid matrices was studied by Naqvi Donatsh and Wild.'" Their spectra are interpreted in terms of a model in which the equilibrium configuration of the fluorescent state is planar but that of the ground state deviates from planarity.The fluorescence emission of p-terphenyl (PT) molecules in solu- tion and in the crystalline state at various temperatures in the range 3004 K has been compared by Williams.'30 The appearance of new bands at higher energies with decreasing temperature corresponds with a phase transition in the crystalline material. Since the structure of this low temperature phase of PT has been identified by X-ray diff ra~fion,'~' electron diff ~action'~' and computed by the pair-wise summation of non-bonded the emission spectra may be correlated with the precise molecular geometry above and below the phase tran- sition temperature.It is concluded that the outer and central rings of the molecule have similar orientations in the low temperature modification and in solution. Following some earlier work Hochstrasser and Sung134 report on the two-photon excitation spectra for ['H,,]biphenyl and [2Hlo]biphenyl neat and mixed crystals at 77 K and 1.8 K. Their results confirm the existence of two electronic states of biphenyl in the 3000A region but may require re-examination in the light of the recent suggestion that at low temperatures (<75 K) the biphenyl molecules are non-planar in the crystalline Indeed a combined study of the optical and magnetic resonance properties of the triplet state of ['Hlo]biphenyl in [2H,,]biphenyl and a neutron diffraction investigation of ['H,,]biphenyl at 4.2 K by Hochstrasser Scott Zewail and Fu~ss'~~ does indicate the low temperature struc- ture of biphenyl to be disordered but the exact nature of the disorder was not determined.5 Heterocyclic and Hetero-aromatic Systems Nitrogen Heterocyclics.-In this class of molecule attention has been focused mainly on acridines and tetrazines. It has been shown by Hochstrasser and King'37 that s-tetrazine in mixed crystal systems undergoes isotopically selective photo- chemical decomposition from both the lowest n-n* singlet and triplet states to yield with near unit quantum efficiency stoicheiometric quantities of nitrogen and hydrogen cyanide. No intermediate species were observed in this photolysis even when performed in organic crystals at 1.6 K.Furthermore by reference to 1,4-s-['5N,]tetra~ine,138 low temperature high resolution mixed crystal absorption spec- troscopy is demonstrated to be a novel non-destructive method for isotopic analysis. The high resolution 1B3u(n7r*)-1A,absorption single isotopic fluores- cence and single isotopic fluorescence excitation spectra of normal s-tetrazine and K. R. Naqvi J. Donatsch and U. P. Wild Chem. Phys. Letters 1975,34 285. 130 J. 0.Williams Chem. Phys. Letters 1976 42 171. 13' J. L.Badour Y. Delugeard and H. Cailleau Acta Cryst. 1976 B32 150. 13* W.Jones J. M. Thomas and J. 0.Williams Muter. Res. Bull. 1975 10 1031. 133 S.Ramdas and J. M. Thomas J.C.S. Furuduy 11 1976,72 1251.134 R.M.Hochstrasser and H. N. Sung J. Chem. Phys. 1977,66 3265. 135 G-P. Charbouneau and Y.Delugeard Acta Cryst. 1976 B32 1420. R. M.Hochstrasser G. W. Scott A. H. Zewail and H. Fuess Chem. Phys. 1975,11 273. 13' R. M.Hochstrasser and D. A. King J. Amer. Chem. Soc. 1975,97 4760. D. S. King C. T. Denny R. M. Hochstrasser and A. B. Smith J. Amer. Chem. Soc. 1977,9!3, 271. 13' Luminescence of Organic Solids 69 the 15N 13C,and *Hmonosubstituted s-tetrazines in natural abundance in benzene at 1.6 K; the photophysics and photochemistry of s-tetrazine in solution (300 K) and solid (4.2-1.6K); and the laser isotope enrichment of s-tetrazine in mixed crystals (1.6 K) are reported and also discussed by Hochstrasser and King.139 Similar studies on dimethyl-s-tetrazine and phenyl-s-tetrazine in neat and mixed crystals have also been and the subnanosecond dynamics of the fluorescence and singlet absorption of s-tetrazine reveal a new S spectrum peaking at 473 nm corresponding to an allowed transition with an S lifetime from absorp- tion of 424 f75 ps and a spontaneous emission lifetime of 450* 55 ps.Such a low lifetime of the S state is consistent with the high quantum yield of the photoreac- tion in this material. Fluorescence emission from azines and N-heterocyclics is invariably weak and normally involves low lying nv* states. This is indeed so for s-tetrazine and a recent study on phenazine,141 following singlet excitation with a picosecond pulse and observation of the build-up of triplet-triplet absorption yields a rate constafit for the build-up of Tl ca.6 x 10” s-’ consistent with a short fluorescence lifetime and a fast intersystem crossing in this material. Studies on acridine by Hirata and Tanaka14* and Sundstrom Rentzepis and Lim143 indicate similar values for the rate of build-up of the triplet-triplet absorption. Thus it is concluded that both phenazine and acridine in n-hexane and other hydrocarbon solvents have a lowest singlet excited state (S,) that is nr* in character. In acridine the fast intersystem crossing is due to the Sl(nr*)+ T3(vv*)process whereas in phenazine the T3(rr*)lies above the Sl(nn*) and the corresponding process is energetically difficult. Thus in phenazine the mechanism and dynamics of the triplet state formation are different from that of acridine and may be a result of the SI(nv*)+ Tl(mr*) transition.The temperature dependence of the emission and lifetimes of several crystallographic modifications of acridine has also been rep~rted,’~”~~ following earlier observations of broad excimeric emission in single crystals of acridine I1 and acridine III.’45 The properties of acridine V are characteristic of a ‘monomeric’ structure whereas acridine 111 known to possess discrete antiparallel molecular pairs exhibits excimeric emission. Two different acridine molecular pairs comprising the acridine I1 structure also display excimeric emission with their respective maxima appreciably red shifted from the molecular emission level and exhibiting different polarization behaviour.Triplet states are believed to offer easy non-radiative pathways following thermal activation from both ‘monomeric’ and ‘excimeric’ crystailine states. Aromatic Carbonyl Derivatives and Related Systems.-As with nitrogen-contain- ing heteroatomic molecules the emission from the lowest singlet states of carbonyl derivatives of aromatic molecules is particularly weak and intersystem crossing followed by triplet emission is the dominant process. Accordingly most studies are related to the properties of the triplet states of such systems. Benzophenone and its 139 R. M. Hochstrasser and D. S. King J. Amer. Chem. SOC.,1976,98 5443. I4O R. M. Hochstrasser D. S. King and A. B. Smith J. Amer. Chem. Soc. 1977,99 3923. 14’ Y. Hirata and I. Tanaka Chem.Phys. Letters 1976,43,568. 14* Y. Hirata and I. Tanaka Chem Phys. Letters 1976 41 336. 143 V. Sundstrom P. M. Rentzepis and E. C. Lim J. Chem. Phys. 1977,66,4287. 144 J. 0.Williams and B. P. Clarke J.C.S. Faraday I 1977,73 514. 145 B. P. Clarke J. M. Thomas and J. 0.Williams Chem. Phys. Letters 1975 35 251. J. 0.Williams derivatives have featured prominently in luminescence studies over the years and reference should be made in particular to the work of Hochstrasser and co-wor- kerS146-148 on excitons and traps involving optical magnetic resonance and ENDOR studies in neat and mixed crystals and to the work of Batley and Bram- ley'49 who assigned the 3n7r* or 37m*of the lowest triplet states of several derivatives. More recently Hirayama and K~bayashi''~ have reported a pico- second-laser study of triplet state population in a range of carbonyl derivatives of anthracene and a similar study OR anthrone and fluorenone has been described by Kobayashi and Nagakura."' In the solid state studies have been few in recent years but Anderson Hochstrasser and P0wna11'~~ have reported on the second excited state of xanthione; Anderson et ~1.'~~ report on direct measurements of internal conversion between excited electronic states of 4-(1-naphthylmethyl) benzophenone and Miyagi Koyanagi and Kanda'54 describe Stark effects on the phosphorescence of p-benzoquinone which indicate the existence of a double minimum potential in the TIstate.6 Polymeric Systems The polymer to receive the greatest attention in the years between 1965 and 1975 was poly(viny1 carbazole) (PVK).As far back as 1969 the energy transfer in amorphous PVK was studied by fluorescence quenching experiments using pery- lene trinitrofluorenone and hexachloro-p-xylene as guest rn01ecules.'~~ It was demonstrated that PVK exhibits excimeric emission owing to a favourable intramolecular pairing of the carbazole molecular units. The results are discussed in terms of exciton migration competitive exciton trapping at excimer and guest sites and Coulombic transfer from excimers to guests. In the pure polymer the singlet exciton visits about lo3 monomer units during its lifetime. Later it was that monomeric carbazole derivatives show phosphorescence and that PVK also emits phosphorescence in dilute frozen solutions together with delayed fluorescence owing to triplet-triplet annihilation.The phosphorescence of PVK is governed by two traps -one of which is the excimer-forming site. Interest in the fluorescence properties of polymers has been at a low ebb over the past two or three years but recently there has been a resurgence of interest with the reports on the luminescence emission from single crystalline bis(to1uene-p-sulphonate) dia~etylene'~~ and from 1,6-di-(N-carbazolyl)-2,4hexadiyne (DCH).15' It is true 14' R. M. Hochstrasser and T. S. Lin J. Chem Phys. 1968 49,4929. 147 S. Dym R. M. Hochstrasser and M. Schafer J. Chem. Phys. 1968,48 646. 14' R. M. Hochstrasser G. W. Scott and A. H. Zewail J. Gem Phys. 1973,58 393.149 M. Batley and R. Brarnley Chem. Phys. Letters 1972 15 337. 150 S. Hirayama and T. Kobayashi Chem. Phys. Letters 1977,52 55. T. Kobayashi and S. Nagakura Chem. Phys. Letters 1976,43 429. Is* R. W. Anderson jun. R. M. Hochstrasser and J. H. Pownall Chem. Phys. Letters 1976,42 224. 153 R. W. Anderson jun. R. M. Hochstrasser H. Lutz and G. W. Scott Chem. Phys. Letters 1975 32 204. 154 Y. Miyagi M. Koyanagi and Y. Kanda Chem. Phys. Letters 1976,40,98. W. Klopffer J. Chem. Phys. 1969,50 2337. Is' W. Klopffer and D. Fischer J. Poly. Sci.Symp. No. 40 1973,43. 15' D. Bloor D. N. Batchelder and F. H. Preston Phys. Status Solidi 1977 a40 279. V. Enkelrnann G. Schleier G. Wegner H. Eichele and M. Schwoerer personal communication (work submitted for publication).Luminescence of Organic Solids 71 to say that perfect fully polymerized single crystals of the bis(to1uene-p-sulphonate) ester of 2,4-hexadiyne- 1,6-diol (TSHD) and the related poly-diacetyl- enes do not normally display luminescence emission but defective partially ply-merized crystds of several diacetylene monomers behave differently. In the study of Bloor Batchelder and Pre~ton'~' the fluorescence is attributed to polymer chain ends trapped in the vicinity of dislocations. Interestingly paramagnetic triplet species had been observed earlier during the thermal polymerization of a diacetyl- ene by Stevens and Bl00r."~ In DCH the monomer crystals show strong phos- phorescence from four traps the depths of which are 23 71 224 and 369cm-' respectively together with a broad band fluorescence which may be excimeric in origin.The sensitive monitoring of the polymerization by changes in the lumines- cence intensity is an exciting possibility in these systems. A novel method for the determination of polymer conformation based on the magnetic modulation of delayed fluorescence has been reported by Avakian et aZ.l6' for the system poly(2- vinylnaphthalene). The observations are interpreted in terms of a local ordering of the naphthalene chromophores with their molecular planes parallel. 7 Other Systems Exciton transitions in 9,lO-dichloroanthracene (DCA) 1,4-dibromonaphthalene (DBN) and 1,2,4,5-tetrachlorobenzene(TCB) are of interest because the stacking of molecules in the crystalline state limits significant exciton transfer interactions to one dimension.The work of Burland Konzelmann and Macfarlane16' has unequivocally established that the triplet exciton absorption linewidth for both crystal sites (site I at 20 192 cm-' and site I1 at 20 245 cm-') in DBN broaden in a non-exponential manner from 440 K and that both transitions are frequency shifted to lower energy with increasing temperature. In addition they have deter- mined that the lowest optic phonon has an energy of 17 cm-' and that the phonon spectrum extends to ca. 96 cm-' with states at 25 cm-' and 39 cm-'. The lineshape function for the So+ Tl absorption has been analysed in terms of exchange theory by Harris.162 It is shown that some exciton coherence could persist over most of the higher temperature range studied i.e.up to 40 K because the phonon absorption rate for optical dephasing or line broadening is significantly slower than the exciton exchange time p-' which gives rise to the band width of 4p in such one- dimensional systems. There has also been a report on the dynamics of triplet exoiton trapping in TCB crystals by Guttler von Schutz and Wolf.'63 The results indicate that exciton trapping is not limited by the motion of the excitons towards the traps but by the trapping step on the trap molecule (rate ca. 3x lo6s-I). The absorption fluorescence and reflectivity spectra of the charge transfer system anthracene-PMDA (pyromelletic acid dianhydride) as single crystals have been investigated at 2 K by ha are^.'^^ From the intensity of the zero-phonon transition and from the width of the phonon side bands information is obtained about the G.C. Stevens and D. Bloor Chem. Phys. Letters 1976,40 37. '60 P. Avakian R. P. Groff A. Suna and H. N. Cripps Chem Phys. Letters 1975,32 466. 161 D. Burland N. Konzelmann and R. M. Macfarlane personal communication (to be published) see also D. Burland and R. M. Macfarlane J. Luminescence 1976,12,123. C. B. Harris Gem. Phys. Letters 1977,52,5. W. Guttler J. U. von Schutz and H. C. Wolf Chem Phys. 1977 24 159. '6-1D. Haarer J. Chem. Phys. 1977,67,4076. 72 J. 0. Williams strength of the electron-phonon coupling. A model calculation on a localized charge transfer complex assuming that the excited state relaxation arises primarily from the Coulombic attraction of the ionic donor-acceptor states supports the rather strong electron-phonon coupling.Photochemical reactions in organized monolayer assemblies have been investi- gated by Whitten and co-worker~.'~~-'~~ Such studies are an exciting extension of the earlier preparation of monolayers and associated energy transfer studies by Kuhn and his co-workers.16* It is found that for derivatives of 4-stilbazole and l-phenyl-4-(4-pyridyl)buta-1,3-diene in the solid state excimer fluorescence and dimerization are the dominant processes. The same processes occur in condensed monolayer assemblies; in addition cis- to trans-isomerization of the surfactant stilbazole occurs but not the reverse. In contrast for both micelles and solutions only cis-trans isomerization and monomer fluorescence are observed and although there is some evidence for photocyclization no dimers or excimers have been detected.The differences in reactivity are attributed to the influence of packing phenomena in both single crystals and in the condensed monolayer assemblies. 8 Application of Novel Techniques Double Resonance and Related Techniques.-Double resonance methods involv- ing the lowest excited triplet states of molecules in organic crystals were performed some years ago and the various experimental techniques have since been reviewed by E1-Sa~ed.l~~ In most if not all the experiments the population difference in the magnetic sublevels of the triplet state is achieved by an optical pumping scheme involving absorption from the ground state to the higher excited singlet state which is followed by a rapid non-radiative process to the lowest excited singlet state (S,)and then to the lowest triplet state (Tl).The non-radiative S1-TIprocess or the T,-+ So process in which the spin quantum number changes could be aniso- tropic and could thus lead to a difference in the population of the three spin levels of the lowest triplet. Experiments are performed at low temperatures to slow down the spin lattice relaxation process and thus preserve the population difference created by the pumping process. Thus depending on the mode of detection or the mode of changing the population of the zero-field levels several different types of double resonance methods have been developed in the study of the triplet state at low temperatures.The most commonly used technique is that in which phos- phorescence is used for detection and is thus named phosphorescence microwave double resonance (PMDR). A related technique is that of microwave-induced delayed phosphorescence (MIDP). Indeed PMDR has been a powerful tool in determining the structure of triplet energy traps in terms of the precise molecular geometry of impurity molecules in host cry~tals,"~ and in elucidating the mechanism of the S1-T non-radiative proce~ses.'~~ Recently the zero field 16' F. H. Quina and D. G. Whitten J. Amer. Chem. Soc.,1975 97 1602. 166 G. Sprintschink H. W. Sprintschink P. P. Kirsch and D. G. Whitten J. Amer. Chem. Soc. 1976,98 2337.167 F. H. Quina and D. G. Whitten J. Amer. Chem. SOC.,1977,99 877. '" see e.g. H. Kuhn D. Mobius and A. Bucher in 'Physical Methods of Chemistry' Vol. 1 Part 3B ed. A. Weissberger and B. Rossiter Wiley New York 1972 p. 588. 169 M. A. El-Sayed Ann. Rev. Phys. Chem. 1975,26 235. M. A. El-Sayed and C. T. Lin in 'Sixth Molecular Crystals Symposium' Schloss Elmau 1973. A. Campion and M. A. El-Sayed J. Phys. Chem 1976,80,2201. Luminescence of Organic Solids 73 splitting parameters of benzene and deuteriated benzene (C&) in a borazole crystal host have been measured by Vergragt and van der Waal~.~~~ Vergragt Kooter and van der Waal~’~~ employ a combination of e.s.r. and MIDP to investi- gate the triplet state of p-xylene in an isotopically mixed crystal and several studies combining e.~.r.l~~ and microwave induced changes in have been performed on various porphins in n-octane crystals.Other examples of magnetic resonance techniques are the spin echo methods that have been developed and used recently by S~hmidt”~ and van der Waals and co-workers.’77 Such methods allow fast changes within a few microseconds in the triplet sublevel populations to be monitored. A particularly interesting develop- ment has been the use of an electron spin echo technique to follow a photochemical reaction in the solid ~tate.”~ This reaction involves the formation of diphenyl-methylene (DPM) by laser photolysis of diphenyldiazomethane (DPDAM) contained in a single crystalline host at 1.2 K.The laser pulse is followed by a pair of resonant microwave pulses and the subsequent spin echo. The echo intensity is a measure of the population difference between two of the DPM triplet sublevels provided the time scale is less than the spin-lattice relaxation times (ca. 3 ms in this system). After the DPM spins have reached thermal equilibrium a reference experiment is performed without the laser flash. The results are interpreted in terms of the preferential population of one of the DPM triplet sublevels during the photochemical reaction which is in turn a result of a spin-orbit coupling due to the central carbon atom of the molecule and is insensitive to molecular geometry. It is known that the coupling of a resonant microwave field with a pair of spin levels of a photo-excited triplet state in zero-field creates a superposition of the two spin states which corresponds to a magnetic dipole moment oscillating at a frequency wo equal to the splitting of the two levels.When applying pulsed microwave fields the decay of the free induction and the formation of e!ectron spin echoes may be observed by means of a microwave receiver tuned to the resonance frequency of the transition. The detection of such coherence as ‘quantum beats’ of frequency wo in the phosphorescence from the triplet state of tetra-methyl[2Hl,]pyrazine in a durene host has been reported by Schadee et all7’ When after a few microseconds the phase coherence is lost it can be restored by a second microwave pulse and the beats then appear as an echo signal (cf.ref. 178). A dephasing time T2of 1.2 ps was obtained. In the visible (or u.v.) region of the spectrum the coherent transients will manifest the dynamics of two different elec- tronic manifolds that are coupled by a weak excitation from a CW laser. Photon echoes (the optical analogues of the spin echo) in mixed molecular crystals have been investigated by Aartsma and Wiersma,’80 and Aartsma Morsink and Wier- sma,18’ and Zewail and Orlowski’82 have recently reported on the phenomenon of 17’ P. J. Vergragt and J. H. van der Waals Chem. Phys. Letters 1976,42 193. 173 P. J. Vergragt J. A. Kooter and J. H. Van der Waals Molecular Physics 1977,33 1523. 174 J. A. Kooter G. W. Canters and J. H. Van der Waals Molecular Physics 1977,33 1545.175 G. Jansen and J. H. van der Waals Chem Phys. Letters 1976,43 413. 176 J. Schmidt Chem. Phys. Letters 1972 14,411. 177 B. J. Potter D. C. Doetschman J. Schmidt and J. H. Van der Waals Mol. Phys. 1975,30 609. 178 D. C. Doetschman J. Phys. Chem. 1976,80 2167. 179 R. A. Schadee C. J. Nonhof J. Schmidt and J. H. Van der Waals Mol. Phys. 1977 34 171. *” T. J. Aartsma and D. A. Wiersma Chem. Phys. Letters 1976,42 520. T. J. Aartsma J. Morsink and D. A. Wiersma Chem. Phys. Letters 1977,47 425. A. H. Zewail and T. E. Orlowski Chem. Phys. Letters 1977,45 399. J. 0. Williams ‘optical ringing’ in solids. In the experiments of these two groups coherent and incoherent transients in crystals of pentacene in p-terphenyl were measured at different temperatures.In the ‘optical ringing’ experiment a single site of the solid was excited by a single mode CW dye laser that is capable of forming a coherent superposition of the ground and the electronic excited state. The laser was then switched electro-optically allowing the detection of the super-radiant state. Thus the phase memory of the solid and the nature of the photon-crystal interactions were examined. The results show that localized states of pentacene in p-terphenyl at 1.7 K have a definite phase memory that can be destroyed only above 3 K owing to the increased phonon population. In the experiments of Aartsma and Wiersma180 and Aartsma Morsink and Wiersma18’ two low intensity laser pulses having the same divergence were passed through delay lines before impinging on the solid with a time separation of ca.20 ns. The time dependence of the photon echo intensity yields a dephasing time and its temperature dependence yields in turn an energy identified with the resonant phonon frequency in the ground state. Such experiments in addition to providing information on homogeneous linewidths of electronic transitions scattering phenomena and molecular parameters may prove useful in probing phase transitions and structural defects in a very sensitive way. Optical ‘Hole-burning’.-It is well known that at low temperatures broad-band absorption and emission spectra of many large molecules in solutions are broadened mainly inhomogeneously . The elimination of this inhomogeneous broadening reveals the fine structure inherent in the spectra of separate molecules.Thus the development of methods for the elimination of inhomogeneous broaden- ing in various kinds of spectra is of great current interest. Inhomogeneous broadening in fluorescence spectra may be eliminated by selective laser excitation near the 0-0 transition frequency (see e.g. refs. 183 and 184). In this respect absorption spectra are also of interest and recently there have been reports of ‘hole-burning’ spectra as a new method for obtaining fine structure in the spectra of organic m01ecules.’~~”~~ Of particular interest to chemists is the recent work of V~elker’~’ on photoisomerization and photochemical hole-burning of free-base porphyrin in n-octane at 4.2 K. When a porphyrin molecule is trapped in one of the two possible sites in an n-octane crystal it may occur in two distinct tautomeric forms.’88 The mechanism of the photo-induced tautomerism has been studied by single level excitation with a pulsed dye laser of molecules in a particular By studying the photoisomerization ‘hole-burning’ in the inhomogeneously-broadened zero-phonon lines of the S +Sotransition has been ob~erved,’~~,’~~ the width of the holes being determined by the bandwidth of the pulsed laser used (S1 cm-l).Such investigations reflect the differences between porphyrin molecules in the two sites and indicate the nature of the broadening mechanism. E. I. Al’shits R. I. Personov A. M. Pyndyk and V. I. Stogov Opt. i. Spektroskopiya 1975 39,274. K. Cunningham J. M. Morris J.Funfschilling and D. F. Williams Chem. Phys. Letters 1975,32,581. B. M. Kharlamov L. A. Bykovskaya and R. I. Personov Gem. Phys. Letters 1977,50 407. M. R. Topp and H.-B. Lin Chem. Phys. Letters 1977,50,412. I” S. Voelker in ‘Eigth Molecular Crystals Symposium’ Santa Barbara U.S.A. 1977. W. G. Van Dorp M. Soma J. A. Kooter and J. H. Van der Waals Mol. Phys. 1974 28 1551. S. Voelker and J. H. Van der Waals Mol. Phys. 1976 32 1703. 19’ H. DeVries and D. Wiersma Phys. Rev. Lefters 1976 36 91. 19’ A. Gorokhovski R. Kaarli and L. A. Rebane Opt. Comm. 1976 16 282. Luminescence of Organic Solids 75 Circularly Polarized Luminescence Spectroscopy.-Recently there has been considerable interest in using the emission analogue of circular dichroism (CD) to investigate the structural features of the emitting states in chiral luminescent molecular systems.This phenomenon has been referred to as circularly polarized emission (CPE) or circularly polarized luminescence (CPL) in which a differential (spontaneous) emission of left and right circularly polarized light occurs from chiral luminescent systems. The subject has been reviewed re~ently.~~~*~~~ Just as all materials when placed in a magnetic field exhibit circular dichroism in their absorption region -the phenomenon of magnetic circular dichroism (MCD)-so should all luminescent materials exhibit CPL in their emission spectra when a magnetic field is applied to them along the direction of emission detection. This emission analogue of MCD has been observed and has been referred to as MCPL (magnetic circularly polarized luminescence).Both CPL and MCPL may have appreciable relevance to organic solids but hitherto studies have been mainly concentrated on obtaining structural (configurational and conformational) information on molecular systems (including biomolecules) in solution. 19* I. Steinberg in ‘Biochemical Fluorescence Concepts’ Vol. 1 ed. R. F. Chen and H. Edelhoch Marcel Dekker New York 1975. 193 F. S. Richardson and J. P. Riehl Chem Rev. 1977,6 773.

ISSN:0308-6003    

DOI:10.1039/PR9777400051

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

5 Studies of Molecular Motion in Liquids and Solids using Low Frequency Dielectric Relaxation and related Techniques By G. WILLIAMS* Edward Davies Chemical Laboratories University College of Wales Aberystwyth Dyfed and J. CROSSLEY Chemistry Department Lakehead University Thunder Bay Ontario Canada 1 Introduction The translational and reorientational motions of whole molecules or parts of molecules may be studied by spectroscopic relaxation and scattering techniques which cover in total the frequency range 10-4-1013 Hz. Dependent upon molecu- lar structure and the phase (liquid supercooled liquid glass crystal or liquid crystal) and temperature and applied pressure molecular motions may be observed anywhere in this large frequency range. For the purposes of the present account the range may be arbitrarily divided into three parts (i) lo1’ Hz< v <1013Hz; (ii) 10’ Hz< Y < lo1’ Hz; (iii) Hz< Y < 10’ Hz.For range (i) the fast motions of molecules are observed e.g. for small molecules in the liquid state or rotator-phase crystalline state. Such motions may be studied using near i.r. and Raman vibration- rotation spectroscopy,’*2 far i.r. ~pectroscopy,~ inelastic light-scattering spectros- copy4 (polarized and depolarized Rayleigh line-broadening) and neutron scattering ~pectroscopy,~ microwave dielectric relaxation techniques,6 and pico-second Kerr- effect relaxation technique^.^ A recent text which gives the results of extensive studies of mainly non-viscous liquids is that by Lascombe.’ The theoretical approach to molecular motion involves time-correlation functions and advanced4.’-l2 and more elementary review^'^-'^ of these quantities are available R.G. Gordon Adv. Magn. Resonance 1968,3 1. ’W. G. Rothschild G. J. Rosasco and R. C. Livingston J. Chem. Phys. 1975,62 1253. M. Evans J.C.S. Faraday II,1975,71 2051. B. J. Berne and R. Pecora ‘Dynamic Light Scattering’ Wiley-Interscience New York 1976. J. W. White in ‘Molecular Spectroscopy’ ed. P. Hepple Institute of Petroleum 1972. N. E. Hill W. Vaughan A. H. Price and M. Davies ‘Dielectric Properties and Molecular Behaviour’ Van Nostrand New York 1969. P. P. Ho W. Yu and R. R. Alfano Chem. Phys. Letters 1976,37 91. * ‘Molecular Motions in Liquids’ ed. J. Lascombe Reidel Dordrecht 1974.B. J. Berne in ‘Physical Chemistry An Advanced Treatise Vol. VIIIB The Liquid State’ ed. H. E. Eyring D. Henderson and W. Jost Academic Press New York 1974 pp. 540-713. lo W. A Steele Adv. Chem. Phys. 1976 34 21. C. Brot in ‘Les Houches 1973’ Gordon and Breach 1976. M. W. Evans Adv. Mol. Relax. Proc. 1977 10 203. l3 R. Zwanzig Ann. Rev. Phys. Chem. 1965 16,67. l4 G. Williams Chem. Rev. 1972,72 55. G. Williams Chem. SOC. Rev. 1978 7 89. 77 78 G. Williams and J. Crossley which describe the relationship between time-correlation functions and experi- mentally observed quantities (e.g. line shapes of absorption spectra scattered spectra). The experimental data obtained for liquids and solids in range (i) may either be fitted using assumed forms for certain time-correlation functions or be transformed numerically to yield experimental time-correlation functions.For fast motions inertial factors may make important contributions to the decay of the correlation functions and this has been extensively discussed.’ Also computer simulations of the dynamics of assemblies of molecules using the technique of ‘molecular dynamics’ have been used successfully to test the validity of model theories for fast motions and to simulate experimentally observed behaviour for liquids (see for example refs. 9 11 16 and 17). Aspects of such work have been discussed by Rowlinson and Evans18 in the Annual Report for 1975. For range (ii) one is concerned with the motions of fairly large molecules in solution (e.g.tri-n-butyl ammonium picrate in benzene) of macromolecules in solution or bulk (at higher temperatures) of moderately sized molecules in their rotator-phase crystalline state (e.g. substituted benzenes) or of liquid crystalline systems above the clearing temperature. Such motions are conveniently studied by dielectric,6 light-~cattering,~”’ n.m.r. (see for example ref. 20) and time-resolved fluorescence depolarization21 techniques. For range (iii) motion is observed for molecules in their supercooled and highly viscous states or in the liquid crystal or rotator-phase crystal states or for macromolecules in their bulk and solution states. Suitable techniques include dielectric,6 n.m.r.,20 dynamic Kerr-effect (see for example refs.22 and 23) and light-~cattering~.~~ (photon correlation) techniques. The present Report is mainly concerned with work published since 1975 for liquids rotator-phase solids liquid crystals and polymers and will primarily refer to studies made in range (iii) (10-4Hz-108Hz). The major proportion of this work is concerned with liquid crystals and polymers owing to the continuing commercial interest in such systems. The Report is arranged to consider liquids liquid crystals rotator-phase crystals and polymers -each in turn. For each class of systems we first give a brief resumC of the recent work and then give more detailed accounts of selected publications. The account will emphasize dielectric studies but will include studies using proton and I3C n.m.r.light-scattering Kerr- effect relaxation fluorescence depolarization piezo-electrical relaxation thermally stimulated current and related techniques. In order to keep the Report to a reasonable length we exclude discussions of mechanical relaxation and rheo-optical properties of polymers or of mechanical and dielectric relaxation of inorganic l6 B. Quentrec and C. Brot Phys. Rev. A 1975,12 272. l7 A. Rahman and F. H. Stillinger J. Chem. Phys. 1971,55 3336. J. S. Rowlinson and M. W. Evans Ann.Reports (A) 1975,72,5. l9 G. R. Alms D. R. Bauer J. I. Braumann and R. Pecora J. Chem. Phys. 1973 58 5570; 1973 59 5310; 1973,59 5321; 1974,61 2255; 1975,63,53. ’O W. P. Slichter and T. M. Connor in ‘NMR Basic Principles and Progress’ Springer Verlag 1970 Vol.4 p. 209 246. B. Valeur and L. Monnerie J. Polymer Sci. (Polymer Phys.) 1976 14 11 29. *’ E. Fredericq and C. Houssier ‘Electric Dichroism and Electrical Birefringence’ Oxford University Press 1974. 23 M. S. Beevers J. Crossley D. C. Garrington and G. Williams J.C.S. Furuduy IZ 1976,72 1482. 24 H. Z. Cummins and E. R. Pike ‘Photon Correlation and Light Beating Spectroscopy’ Plenum Press New York 1974. Studies of Molecular Motion in Liquids and Solids 79 crystals (vacancy and ion-hopping processes). No attempt is made to make the Report comprehensive but rather it reflects the interests of the authors. In anticipation of the following account we note that most recent work involves the documentation of the motional behaviour using a single experimental technique followed by a detailed interpretation using assumed models for motion.However ‘slow’ reorientational and translational motions are statistical processes governed by the time-dependent fluctuations in local thermodynamic q~antities~’~’~’ and hence give rise to time-correlation functions of the motion which resemble exponential or weighted sums of exponential functions of time. The resultant experimentally observed quantities (e.g. permittivity intensity correlation function for scattered light24) are broad and featureless. It is all too easy to fit such data from a single experimental technique to a variety of models for motion each with its adjustable parameters. Such a situation does not arise for high-resolution quantum spectroscopy where the data from a single technique may suffice to determine the required molecular quantities.Given the fact that motional pro- cesses lead to broad spectral features an improvement in our understanding of the molecular processes will not be obtained by making more precise measurements for a given technique but rather by comparing data obtained using different but complementary techniques and then seeking a unified interpretation. The basic problem may be illustrated by the example of the reorientational motion of a molecule possessing axial symmetry and which has associated with its z axis a dipole moment p = pu and a molecular polarizability ellipsoid of components a,,a,,, and a = a,,.If f(Q,t) dfl is the conditional probability that the unit vector ur lies in the element of solid angle da around Ln at time t given that it was along the +t laboratory axis at t = 0 then for the special case of axially symmetric reorien- tation of the vector with respect to its initial direction which may be a reasonable approximation in the liquid state f(n,t) may be written as a series expansion involving Legendre polynomials Pn(cos 0) and orientational (auto) time-correlation functions 4 (t) as follo~~.~*~~*~~ rL,(t)= (pn [cos @(t)J) = f(JI,tPn(cos 6)dR (2) 8 is the angle between u and z &(t)= 1.All the information on the reorientation process is contained in f(n,t). Experimentally dielectric studies may Jll(t) whilst dynamic Kerr-effect studie~,~~*~~ depolarized light-scattering and time-dependent fluorescence depolarization studies” may in suitable cases yield 492(t).Note that n.m.r. studies conducted at a single resonance frequency may in suitable cases26 yield a correlation time T~ = j&(t) dt for the reorientation of an inter-nuclear vector but cannot yield &(t) itself. Clearly dielectric studies on their own yielding cLl(t) or each of the above mentioned techniques on their own yielding $z(t) are insufficient to specify f(n,t) and therefore the mechanism for motion. It is essential that the results of different but complementary studies be compared e.g. t,bl(t) and t,b2(t),in order to rule out certain mechanisms and favour others. 25 L. D. Landau and E. M. Lifshitz ‘Statistical Physics’ Addison-Wesley Reading Mass.1958. 26 A. Abragam ‘Principles of Nuclear Magnetism’ Oxford University Press 1961. 80 G. Williams and J. Crossley This simple example suffices to illustrate the basic problem with studies of molecular motion. More generally the situation is much more complicated since one may be concerned with anisotropic molecular reorientation which involves of time-correlation function~~’~ the elements of Wigner rotation matrices DL,K(a,p y). In addition angular correlations between molecules (e.g. alcohols liquid crystals) or between groups within a molecule (e.g. flexible-chain polymers) lead both to auto- and cross-correlation function^^*^*^^'^^ which contribute to the observed processes. As we shall see only a few recent studies have attempted to compare and contrast dynamical data from different experimental techniques and the lack of effort in this direction severely limits our understanding of how mole- cules move.However many studies do not seek the mechanism for motion the aim being to know what is moving and how fast it is moving at a given temperature pressure and composition condition. 2 Glass-forming Molecular Liquids and Glasses This section is concerned with studies of organic glass-forming liquids. Reviews of studies made before 1976 have been given by Williams2’ (dielectric studies) Lamb2’ (viscoelastic studies) and by Wong and Ange11.29 More recently accounts of the nature of the primary and secondary relaxations in glass-forming systems have been given by Johari3’ and others3’ and by G~ldstein.~~-~~ For convenience our discussion will be in two parts namely for T > Tgand for T < T, where Tgis the apparent glass-transition temperature.For T > TBtwo motional processes may be observed these being the primary (a)process due to gross Brownian motions of molecules and the secondary (p)process due to local motions or partial reorien- tations of molecules. Usually only the a process is studied but it is established that for glass-forming and for amorphous solid both a and p processes coexist for a limited temperature range for T> T and coalesce at high temperatures. For T < T the p process is observed. Ranger >T,.-Dielectric studies have been made for polar solutelo -terphenyl solution~,~’-~~ tri-tolyl phosphate,39 n-propyl benzene4’ and for mono- and di- hydroxy alcohol^.^^.^^ Dynamic Kerr-effect experiments have also been made on 27 G.Williams in ‘Dielectric and Related Molecular Processes’ ed. M. Davies (Specialist Periodical Reports) The Chemical Society London 1975 p. 151. 28 J. Lamb in ref. 8 p. 29. 29 J. Wong and C. A. Angell ‘Glass Structure and Spectroscopy’ Marcel Dekker Inc. New York 1976. 30 G. P. Johari Ann. New York Acud. Sci. 1976 279 117. See Discussion p. 141-149 in ref. 30. 32 M. Goldstein. J. Phys. Paris Colloq. 1975 C2 C2. 33 L. Hayler and M. Goldstein J. Chem. Phys. 1977,66 4736. 34 M. Goldstein J. Chem. Phys. 1977,67 2246. 35 G. P. Johari J. Chem. Phys. 1973,58 1766. 36 G. P. Johari and G. Williams Faruduy Symposia Chem. SOC.1972 No.6,42. 37 M. Nakamura H. Takahashi and K.Higasi Bull. Chem. SOC.(Japan) 47 1593. M. S. Beevers J. Crossley D. C. Garrington and G. Williams Furuduy Symposia Chem. SOC.1976 No.‘ll 38. 39 M. S. Beevers J. Crossley D. C. Garrington and G. Williams J.C.S. Faruday 11 1977,73 458. 40 T. G. Copeland and D. J. Denney J. Chem. Phys. 1976,80,210. 41 J. Crossley and G. Williams J.C.S. Furuduy 11 1977,73 1651. 42 J. Crossley and G. Williams J.C.S. Furuduy 11 1977 73 1906. Studies of Molecular Motion in Liquids and Solids 81 certain of these ~y~tem~,~~*~~*~~*~~ whereas viscoelastic relaxation studies have been made for he~ane-triol~~ and eugen01~~. Inelastic light-scattering studies have been made for ethyl benzoate,4547 benzyl benzoate,4547 for butane- 1,3-di01,~~ for viscous gly~erol,~~,~~ dimethyl sulfoxide/water mixtures,51 and for Na20,K20,6Si02.52 In addition the motion of ions in concentrated aqueous acids has been studied with electrical conductivity relaxation53 and the Mossbauer tech- nique has been useds4 to probe motion in glycerol.The general form of the a process also known as the structural relaxation process is well established from dielectric viscoelastic and now from recent studies using the dynamic Kerr-effe~t’~*’~ and light-scattering technique^.^^-^^ The different techniques probe different aspects of the motion since they relate to different time-correlation functions of the m~tion.~*~~*’~ However the form of the relaxation function for non-associated viscous liquids (and for glycerol) as obtained by the various techniques all resemble that of a Cole-Davidsons6 or Williams-Watt~~~ function while the average relaxation (or ‘correlation’) time T is strongly dependent upon temperature near Tg, having a form T = T~ exp [B/(T-To)].The overall similarity in behaviour for liquids of very different chemical structure is striking and unexpected and implies that the mechanism for motion involves the co-operatiue motions of large assemblies of molecules (see for example refs. 28-30). Williams and co-worker~~~*~~ have used dielectric and Kerr-eff ect techniques for solutes in o-ter~henyl~~ and for pure liquid tri-tolyl phosphate and its mixtures with o-terpheny13’ in an attempt to define the mechanism for the a relaxation in these and other glass-forming systems.For the fluorenonelo -terphenyl system these techniques probe the auto-correlation functions for angular motion of the dipole as Ijll(t) and &(f) as indicated in the introduction. Different models for motion predict different relationships between t+bl(t)and t+b2(t).For reorientation by small- angle steps (small-angle rotational diffusion) 1 ‘a f(a,t)=-2 (2n+ l)P,(cos 6)exp [-n(n + l)Drt] (31 4~ n=O SO Ijll(t) = exp (-2Drt) and t+b2(t)= exp (-6D,t) where D is the rotational diffusion coefficient. Thus for this mechanism both dielectric and Kerr-eff ect processes ~~ would be characterized by single relaxation time functions with TK (decay ~ transient)= ($)T~~~~~~~~~~, and the Kerr-effect rise transient would be far slower than O3 R.Kono and H. Yoshizaki J. Appl. Phys. 1976,47 1359. O4 M. G. Kim J.C.S. Faraday II 1975,71,415. ” P. Bezot G. M. Searby and P. Sixou J. Chem. Phys. 1975,62 3813. O6 G. Searby P. Bezot and P. Sixou Faraday Symposia Chem. SOC.,1976 No. 11,63. O7 P. Bezot G. M. Searby and P. Sixou Opt. Comm. 1976,16 276. O8 P. W. Drake J. F. Dill C. J. Montrose and R. Meister J. Chem. Phys. 1977,67 1969. O9 C. Demoulin C. J. Montrose and N. Ostrowsky Phys. Rev. A 1974.9 1740. ’O C. Demoulin P. Lallemand and N. Ostrowsky Mol. Phys. 1976,31 581. ” P. W. Drake and R. Meister J. Phys. Che~., 1976,80 2780. ’* C. C. Lai P. B. Macedo and C. J. Montrose J. Amer. Ceram. SOC. 1975,58 120. 53 I. M. Hodge and C. A. Angell J.Chem. Phys. 1977,67,1647. M. Soltwisch M. Elwenspoek and D. Quitmann Mol. Phys. 1977 34 33. 55 C. J. Montrose and T. A. Litovitz J. Acoust. SOC. of America 1970,47 1250. 56 D. W. Davidson Canad. J. Chem. 1961 39,571. ” G. Williams D. C. Watts S. B. Dev and A. M. North Trans. Faraday SOC. 1971,67 1323. G. Williamsand J. Crossley the decay transient the rise transient being a mixture of &(t) and &(t) For reorientation by jumps through angles of arbitrary size (fluctuation-relaxation so q91(t)= 4b2(t)=((t) and hence the dielectric and Kerr-effect relaxation would be characterized by the relaxation function ((t) which is not necessarily an exponen- tial function of time or a weighted sum of such exponential functions. This would ~~~.give TKerr (decay transient) = P~~~~ (rise-transient) = P~ Experimentally it was found38 for the fluorenonelo -terphenyl system that the rise and decay Kerr-eff ect transients were equivalent at a given temperature being characterized by the same broad relaxation function and the same relaxation time as that obtained from the dielectric relaxation studies and thus for the first time experimentally ruling out rotational diffusion as the mechanism for motion and favouring a mechanism based on fluctuations leading to relaxation. Interestingly of the large ion-pair solute tri-n-butyl ammonium picrate in o-terphenyl by dielectric and Kerr-eff ect techniques gave behaviour of the form observed for fluorenonelo -terphenyl for T close to T,but for higher temperatures a systematic change to behaviour charac- teristic of small-angle rotational diffusion of the solute was observed.Such information could not be obtained from either technique on its own. Thus comparative studies using the dielettric and Kerr-eff ect techniques on model systems suggest that molecules may reorientate in the structural relaxation process uia a ‘fluctuation-relaxation’ mechanism in which ((t) is a function which charac- terizes the temporal evolution of the fluctuating ‘local’ thermodynamic quantities. The fact that ((t) =exp-(t/~)’.’ for a variety of non-associated viscous liqUids23,28,38.39,41 implies that the fluctuations have the same time dependent form independent of chemical structure* and bear out the suggestions of the earlier work of Montrose and Litovitz’’ and Lamb2* and the earlier discussions of Johari and Goldstein and of Williams and co-workers (see for example refs.27 and 30). Of special interest regarding the mechanism of structural relaxation are the recent studies using the dynamic light-scattering techniques. Excellent reviews are given by Berne and Pecora4 and by Searby and co-~orkers~~ of the theory and practice of obtaining polarized and depolarized dynamic scattering data both in the frequency- and time-domains. In contrast to dielectric relaxation where the molecular probe of motion is the dipole moment or the dynamic Kerr-effect where the molecular probe is the combination of the molecular dipole moment and the molecular polarizability ellipsoid light scattering may arise from a variety of processes some purely macroscopic and others which may be related to microscopic (molecular) quantities.The scattering of a linearly polarized light beam arises basically from non-uniformity of the dielectric permittivity caused by the continual random (Brownian) motions of the molecules. Light scattered with the same polarization at the incident radiation contains information about isotropic fluctua- tions whereas the depolarized scattered light gives information about anisotropic fluctuations -and hence reorientational motions of molecules. As explained by * Associated viscous liquids may behave in a different manner see for example refs. 41 and 42. Studies of Molecular Motion in Liquids and Solids 83 Berne and Pe~ora,~ by Dill and co-w~rkers,~~ Searby and co-w~rkers,~~ and by Demoulin and co-w~rkers,~~ the important parameters characterizing the spectrum of the VV scattered light from a viscoelastic liquid are the frequency shift of the Brillouin lines oB,the stress relaxation time T~, and the thermal diffusion rate Ak2/(pC,) where A p and C are thermal conductivity density and constant pressure specific heat capacity respectively and k is the scattering wave vector.Three ranges are to be di~tinguished.~~ >> OB >> Ak2/(pCp) Range (i) (1/~) Here the Rayleigh line is due to the isobaric portion of the density fluctuations and is associated with thermal diff rision giving a Lorentzian line-shape of half-width Ak2/(pC,),and a correlation function of the form exp -[(Ak*t/(pC,)].Range (ii) WB >> (1/~) >> Ak2/(pCp) In this case the non-propagating fluctuations in the local structure lead to a component known as the ‘Mountain line’ which superposes but is broader than the normal Rayleigh central component due to thermal diffusivity. Range (iii) wB>>Ak2/(pC,)>>(1/~) The central scattered line again consists of two components but now the Mountain line is narrower than the normal Rayleigh component. The relaxational behaviour responsible for the shape of the Mountain line is isothermal rather than adiabatic. For a viscoelastic liquid Ranges (i) (ii) and (iii) occur for liquid viscosities less than 0.1 N s m-2 (1 poise) 0.1 < r) < 1O3 N s m-2 and q > 1O3 N s m-2 respectively.Structural relaxation may be studied using the Mountain line in polarized light- scattering a Fabry-Perot instrument being appropriate to Range (ii) and a photon- correlation spectrometer being appropriate to Range (iii). Demoulin and co-w~rkers~~ studied glycerol in Range (iii) in the temperature range 224-1 92 K by photon-correlation spectroscopy and observed the cor-relation function in the time domain. It was found to resemble the Davidson-Cole function with distribution parameter in the range 0.4-0.5. Such data may a!so be fitted to the empirical function of Williams and Watts i.e. @(t)= exp -(t/~)’,with p -0.7-0.8. This is the correlation function for the fluctuations arising from structural relaxation in glycerol and Demoulin and co-w~rkers~~ found that the average structural relaxation times obtained by the light-scattering technique are very similar to values extrapglated from low-frequency ultrasonic structural relax- ation studiess9 and also that the Davidson-Cole function with relaxation parameter of 0.44 was common to the data of both experimental techniques.The similarity of the form of structural relaxations for a given liquid as determined by different techniques has been previously n~ted.~~*~~*~~ Models for structural relaxation have been proposed involving fluctuations in local orders5 and defect-diff usion60-62 and these lead to broad relaxations of the Davidson-Cole form. However B~rdewijk~~ 58 J. F. Dill P. W. Drake and T. A. Litovitz ASLE Trans. 1975,18 202.59 R. Piccirelli and T. A. Litovitz J. Acoust. SOC.Amer. 1957 29 1009. 6o S. H. Glarum J. Chem. Phys. 1960,33,639. 61 M. C. Phillips A. J. Barlow and J. Lamb Proc. Roy. SOC.,1972 A329 193. 62 M. F. Shears G. Williams A. J. Barlow and J. Lamb J.C.S. Faraday II 1974 70 1783. P. Bordewijk Chem. Phys. Letters 1975 32 592. 84 G. IVilh’ams and J. Crossley has re-examined the defect-diffusion models and has demonstrated that whereas one-dimensional diffusion leads to non-exponential relaxation functions three dimensional diffusion leads to single relaxation-time behaviour. Clearly there is a need for a theory of structural relaxation which naturally leads to a prediction that common relaxation times and common (non-exponential) correlation functions will be obtained from different experimental techniques.Recently Demoulin and co- workersSo have given a microscopic theory of light-scattering for a pure fluid and different predictions are made dependent upon thermodynamic parameters and the type of relaxation. Their theory is applied to structural relaxation in viscous glycerol.50 As for polarized scattering studies of depolarized scattering for normal and highly viscous liquids are very recent and the pattern of behaviour is only now being established for such liquids as ethyl benzoate benzyl benzoate tolane triphenyl phosphate and benzyl alcohol. Searby and co-w~rkers~~~~ have given critical accounts of such studies. They show that the hydrodynamic theories* of Keyes and Kivel~on~~~ adequately describe the line-shape of and of Anderson and Pec~ra~~~ the depolarized spectrum at high temperatures but at lower temperatures (viscous liquid above T,)a weak depolarized doublet is seen corresponding to propagating transverse modes and is not accommodated by theory.We note that Pecora and co-worker~~~ have observed the fast reorientational motions (T -lO-’Os) for a number of non-viscous liquids via the linewidth of the/ depolarized Rayleigh line. Quite recently Crossley and William~~’.~’ made comparative dielectric and dynamic Kerr-eff ect studies of relaxation for 2-methylpentane-2,4-diol (l) dipro- pylene glycol (2) pentane-2,4-diol (3) 2-ethylhexane-l,3-diol (4) and 6-methyl- heptan-3-01 (5). For compounds (1)and (2) structural relaxation was observed similar to that dicussed above for fluorenonelo-terphenyl giving Tkmr= T~,~ = TD and common relaxation functions.For the mono-alcohol (5) single relaxation-time behaviour is maintained for both dielectric and Kerr-effect studies with TD = 7k.d =3 (+)T~,,being indicative of the motion of OH groups through larger angles (‘coarser’ motion) than is found for liquid H20 (see ref. 17). The results for compounds (3) and (4) show T~,~ and TD but their to be substantially greater than both T~,~ interpretation is not simple in this case owing to the combined effects of inter- and intra-molecular hydrogen bonding. Range T < T,.-It is well established that the formation of a glass is experimentally a wholly kinetic process. The structure and formation of certain glasses has been investigated by computer ~irnulation,~’ where the glass was prepared by rapid quenching and the thermodynamic aspects have been discussed.66 In the simula- tion by Rahman and ~o-workers~~ rapid cooling of Lennard-Jones hard spheres gave a ‘vitrification’ at about 0.7Tmelting with the relaxation time T at Tgbeing66 -3.5 x 10-los i.e.very much smaller than that in normal experiments where T = 10’s at T,. Gibbs and co-workers66 have given evidence to suggest that for 64 (a) T. Keyes and D. Kivelson J. Chem. Phys. 1972 56 1876; (6) H. Anderson and R. Pecora ibid. 1971,54 2584. 65 A. Rahman M. J. Mandell and J. P. McTague J. Chem. Phys. 1976,64,1564. ‘‘J. M. Gordon J. H. Gibbs and P. D. Fleming J.Chem. Phys. 1976,65 2771. *Involving k2r/7/(p) where r/ is the shear viscosity and 7 is a correlation time and R a coupling parameter. Studies of Molecular Motion in Liquids and Solids 85 hard spheres ‘glass-formation’ is a purely kinetic phenomenon with no underlying thermodynamic second order transition temperature T2( T2< T,) being anticipated in the cooling behaviour -in contrast to that which has been suggested for polymers by Gibbs and DiMar~io.~~.~~ The molecular motion present in the glassy state for a variety of simple rigid- molecule molecular glasses (and some fused salts) has been studied by Hayler and G~ldstein.~~ It is found that the temperature of maximum loss for the p process at 1kHz is -0.75 T in many cases although not all the glasses exhibit a /3 process.This work and earlier work by G~ldstein,~~ Johari (see refs. 30 and 31) and by Johari and Goldstein (see refs. 27 and 30) show that the p process in such systems has an extremely broad contour (as loss-factor us frequency) with a half-width of up to five decades of frequency; thus there is unlikely to be a simple molecular mechanism for the process. The fact that the p process may be observed above T both for small molecule glass-forming ~ystems~~’~~ and for amorphous solid poly- mers (see for example ref. 36) implies that it is not a special property of the glassy state. Williams and co-w~rkers~~.~~ have given general relations which accom- modate the observations that (i) the p process is observed for T> Tg and for T < T, (ii) that the a and p processes coalesce above a given temperature and (iii) that the relative magnitudes of a and p processes are coupled despite the fact that the mechanisms for a and p processes are independent of each other.The general relations arise for a model of partial reorientation (p process) followed by total reorientation (aprocess) of a molecular group and take the form27*36*69 for permit- tivity E* so and are the low and high frequency permittivities respectively 9indicates a one-sided Fourier transformation 4u(t) is the relaxation function for the micro- Brownian motions of the dipole &,(t) is that for the local motions of the dipole in environment r Opr is the probability of obtaining environment r for the dipole 4ur= (1-qor)= [(pr)I2/p2 where (p)is the mean moment residing in environment r as observed on a time-scale long compared with that for the Pr process.In equation (5) it is implied that the a process is the structural relaxation process discussed above for T> Tg whereas the mechanism for the p process may generally include intra- and inter-molecular motions. The use of equation (5) has been extensively dis~ussed~~.~~,~~ in relation to molecular glass-forming liquids and for amorphous solid polymers. 3 Other Low Frequency Relaxation of Liquids The low frequency motions of ions ion-pairs and their aggregates have been described by Lestrade and co-workers in a recent review.70 Recently there has been an interest in studying molecular motion using non-linear dielectric relaxation 67 J.H. Gibbs and E. A. DiMarzio J. Chem. Phys. 1958,28,373. E. A. DiMarzio and J. H. Gibbs J. Chem. Phys. 1958 28 807. 69 G. Williams and D. C. Watts in ‘NMR Basic Principles and Progress’ Springer Verlag 1970 Vol. 4 271. ’O J. C. Lestrade J. P. Badiali and H. Cachet in ref. 27 p. 106. G. Williamsand J. Crossley effects and this has been reviewed by Jones.71 Non-linear relaxation techniques also afford information on the dynamics of chemical equilibria in suitable systems. Following the earlier work of Bergmann Eigen and DeMae~er,~’ Hellemans and DeMae~er~~ have developed a new experimental technique for measuring the real and imaginary parts of the electric field-induced increment in permittivity for a medium in the range 20-100MHz.The technique was applied to the system E -caprolactam/cyclohexane and the dispersion and absorption curves were well- fitted by a single relaxation time expression. For a situation where a cyclic dimer D2 is formed from an open dimer D1 by successive hydrogen-bonding between the monomers M i.e. M+M$D1 +D2 which was used to interpret thermodynamic studies of this system then the relaxation time T would be given by (1/T)2= klCo+kg (6) where Co is the concentration of lactam and kl and kZ are the rate coefficients. However the data indicate a linear relationshipbetween (1/~) and Co,which is not the case for equation (6). Thus the proposed dimerization mechanism is inconsis- tent with the observed non-linear dielectric permittivity behaviour.This and the closely related studies of H~pmann~~ are discussed by Hellemans and DeMaeyer. Nauwelaers and co-w~rkers~~ have studied the dynamics of ion-pair formation and dissociation for tetra-n-butyl ammonium picrate in diphenyl ether using a non-ohmic relaxation technique at high fields in the range 60 Hz to 10 kHz. The field dissociation effect (the ‘second’ Wien effect) was observed to follow the equation of a single relaxation time T& which for the simple dissociation kd A+B- s A++B-kr (7) corresponds to Tch = kd + k(CA -k CB) (8) where CAand CBare the concentrations of the ions at equilibrium. For only slight dissociation of ion-pairs equation (8) becomes where KD is the ionic dissociation constant and Cois the total concentration of the solute (tetra-butyl ammonium picrate).Nauwelaers and co-~orkers~~ found that 7,h was proportional to ck for all concentrations studied allowing determination of k, kD,and KD.All relevant parameters calculated from the dynamic data and from thermodynamic data are discussed. The modern experimental techniques for the measurement of the non-linear dielectric propertie~~~.~~.~~ appear to be a valuable addition to available techniques for measuring the individual rate coefficients for chemical equilibria. 71 G. P. Jones in ref. 27,p. 198. 72 K. Bergmann M. Eigen and L. DeMaeyer Ber. Bunsengesellschaft Phys. Chem. 1963,67 819. 73 L.Hellemans and L. DeMaeyer J. Chem. Phys. 1975,63 3490. 74 R.Hopmann Ber.Bunsengesellschaft Phys. Chem. 1970,74,935; 1973,77,52; J. Phys. Chem. 1974 78 2341. ’’ F. Nauwelaers L. Hellemans and A. Persons J. Phys. Chem. 1976,80 767. Studies of Molecular Motion in Liquids and Solids 4 Liquid Crystals Introduction.-Owing to their remarkable physical properties and their practical importance many studies have recently been made for liquid crystals in their different phases. In view of the large number of publications we can only cover a selection for the purpose of the present account. Useful introductory reviews of liquid crystal behaviour and structure have been given by Sa~pe,~~ Margerum and Miller,7g and by Also texts have been published by deGennesgO and Gray and Winsor.gl Luckhurst and Faberg’ gave a recent Annual Report of progress in liquid crystal structure and dynamics and further reviews of structure and dynamics have been given by Ohtsukag3 (cholesterics) and by de Jeug4 (dielectric properties).Accounts of dynamics are also given by deGenne~,~~*~’.~~ Hirotakes7 (n.m.r.) Freedgg (n.m.r.) Frankling9 (n.m.r.) Agostini and co-worker~~~ (magnetic and dielectric relaxation) and by North and Pethrick” (acoustic relaxation). For convenience in this Report we first indicate references to work on liquid crystals using a particular experimen- tal technique each taken in turn and then discuss selected papers for a given technique in greater detail. Dielectric Relaxation.-Dielectric studies of molecular rotation have been carried out for a wide variety of liquid crystals with nematic N-(p-methoxy benzy1idene)- p-butyl aniline (MBBA) the azo and azoxy benzenes and substituted biphenyls and their closely related derivatives receiving most attention.We may quote the work of Agarwal and Price92 and Cummins and co-~orkers~~ as illustrating the behaviour of disoriented and oriented MBBA whereas more recent studies of MBBA are due to Moscicki and co-worker~,~~,~~ Rondelez Ikeno and co-w~rkers,~~ and Mircea-Ro~ssel,~~ Yano and co-worker~,~~ Agarwal and co-worker~,~~ and by 76 G. H. Brown J. Coll. Interference Sci. 1977 58 534. 77 A. Saupe J. Coll. Interference Sci. 1977 58,549. J. D. Margerum and L. J. Miller J. Coll. Interference Sci. 1977 58 559. 79 D. M. Small J. Coll. Interference Sci. 1977 58 581.P. G. deGennes ‘The Physics of Liquid Crystals’ Oxford University Press 1974. 81 ‘Liquid Crystals and Plastic Crystals’ eds. G. W. Gray and P. A. Winsor Vol. 1 and 2 Ellis Horwood Ltd. Chichester 1974. G. R. Luckhurst and T. E. Faber Ann. Reports (A) 1975 72,31. 83 T. Ohtsuka Oyo Butsuri 1976,45,498. 84 W. H. deJeu to be published in ‘Advances in Liquid Crystal Research’ ed. Orsay Liquid Crystal Group as a supplement in the ‘Solid State Physics’ series 1978 Vol. 34 Ch. 4. P. G. deGennes Phys. Letters 1969,30A 454. 86 see ref. 81 pp. 92-8 (Vol. 1). K. Hirotake Nippon Butsuri Gakkaishi 1975,30 195. J. H. Freed J. Chem. Phys. 1977,66,4183. 89 W. Franklin Mol. Cryst. Liq. Cryst. 1977 40 91. 9o G. Agostini P. L. Nordio G. Rigatti and U. Segre Att.Accad. Naz. Lincei Mem. Cl. Sci. Fis. Mat. Nut. Sez. 1975,13,20pp. 91 A. M. North and R. A. Pethrick in ‘Transfer and Storage of Energy in Molecules’ ed. G. M. Burnett A. M. North and J. N. Sherwood Wiley New York 1974 Vol. 4,441. 92 V. K.Agarwal and A. H. Price J.C.S. Faraday ZZ 1974,70,188. 93 P. G. Cummins D. A. Dunmur and N. E. Jessup Liq. Cryst. Ordered Fluids 1973,2,341. 94 J. K. Moscicki Solid State Comm. 1976 20,481. ” J. K. Moscicki X. P. Nguyen S. Urban S.Wrobel M. Rachwalska and J. A. Janik Mol. Cryst. Liq. Cryst. 1977 40 177. 96 S. Ikeno M. Yokayama and H. Mikawa Mol. Cryst. Liq. Cryst. 1976 36 89. 97 F. Rondelez and A. Mircea-Roussel Mol. Cryst. Liq. Cryst. 1974 28 173. 98 V. K. Agarwal V. P. Arora and A. Mansingh J. Chem. Phys. 1977,66,2817.99 S. Yano Y. Hayashi M. Kuwahara and K. Aoki Japanese J. Appl. Phys. 1977,16 649. G. Williams and J. Crossley Prost and Pershan.'" The work of Davies and co-workers,'ol Cummins and co-workers,102 and of Ratna and Shashidharlo3 illustrate the behaviour of p-substituted cyano-biphenyls. Quite recently Schadt'04 has reported the dielectric and Kerr-electro-optical relaxations of heterocyclic analogues of the p-cyano- biphenyls which contain nitrogen atoms in the aromatic rings. Azoxybenzenes and their related structures have been studied by Moscicki and co-workersg5 Mircea- Roussel and R~ndelez,~" Wrobel and co-workers,'06 Obayashi,lo7 and by Bata and co-workers,'08 while certain phenyl benzoates have been studied by deJeu and Latheuwer~,'~~ Kresse and co-workers,'10 and by Bata and Molnar."' In addition the conductivity relaxation of liquid crystal mixtures containing Schiff's base ester or biphenyl components with a positive dielectric anisotropy have been studied by Schadt and von Planta.l12 The dielectric behaviour of two nematic phenyl pyri- midines was studied by Kresse and co-worker~~~~ while Schadt and M~ller"~ studied two nematogens containing Schiff's base and cyano components and investigated the effect of organic dopants on the static and dynamic electro-optical properties of the resultant twisted nematic displays.Examples of recent work on the dynamics of the cholesteric phase as studied by dielectric relaxation are the papers of Kashnow and co-~orkers~~~ and Evans and co-workers."6 Agarwal and Price9* studied MBBA in its isotropic and nematic states over the range lo3Hz to 1.2 x lo8Hz.For the ordered nematic phase (saturation achieved by a magnetic field B > 0.25T) one relaxation was observed with relaxation times 711 in the range 1.8~ 10-8s to 1.01X10-7s for T=313-300K. Here 711 is the relaxation time for motions of dipoles parallel to the axis of the director. In contrast the relaxation time in the isotropic phase was 138 ps at 323 K being much shorter than the 711 value under equivalent conditions. Rondelez and Mircea- Ro~ssel~~ followed up earlier work on MBBA,"' where they had shown that in the range lo5-lo7 Hz relaxation was observed for EII but not for E~ of the dielectric tensor. Here 11 and Iare with respect to the direction of the nematic alignment.One process was observed for 811 with an apparent activation energy of 0.65 eV and which peaks at lo6Hz at 295 K. No detailed interpretation of the process was made. Ikeno and co-w~rkers~~ studied the dielectric and optical transmission properties of MBBA and the ethoxy analogue EBBA. For a sample of MBBA loo J. Prost and P. S. Pershan J. Appl. Phys. 1976 47 2298. M. Davies R. Moutran A. H. Price M. S. Beevers and G. Williams J.C.S. Faraday ZZ 1976 72 1447. lo2 P. G. Cummins D. A. Dunmur and D. A. Laidler Mol. Cryst. Liq. Cryst. 1975 30,109. '03 B. R. Ratna and R. Shashidhar Pramana 1976,6 278. '04 M. Schadt J. Chem. Phys. 1977,67 210. '05 A. M. Mircea-Roussel and F. Rondelez J.Chem. Phys. 1975 63 2311. Io6 S. Wrobel J. A. Janik J. Moscicki and S. Urban. Acta. Phys. Polon. 1975 A48,215. lo' C. Obayashi J. Phys. SOC.(Japan) 1975 38,1787. lo' L. Bata A. Buka and G. Molnar Mol. Cryst. Liq. Cryst. 1977 38,155. W.H.deJeu and W. Th. Latheuwers Mol. Cryst. Liq. Cryst. 1974 26 225. 'lo H. Kresse D. Demus and C. Krinzner 2.phys. Chem. (Leipzig) 1975,256 7. ''I L. Bata and G. Molna Chem. Phys. Letters 1975 33,535. '12 M.Schadt and C. von Planta J. Chem. Phys. 1975,63,4379. 'I3 H. Kresse P. Schmidt and D. Demus Phys. Status Solidi A. 1975 32,315. M. Schadt and F. Miiller J. Chem. Phys. 1976,65 2224. 'I5 R. A. Kashnow J. Bigelow H. S. Cole and C. R. Stein Liq. Cryst. Ordered Huids 1973 2,483. M. Evans R. Moutran and A. H. Price J.C.S.Faraday ZZ 1975,71 1854. F. Rondelez D. Diguet and G. Durand Mol. Cryst. Liq. Cryst. 1971 15 183. Studies of Molecular Motion in Liquids and Solids 120 p thick and at 297 K the permittivity first increases with field strength E (up to 800 V cm-') then rapidly decreases as E increases to 2 kV cm-' and gradually increases at still higher values of E. Similar complex behaviour was observed at different temperatures in the range 297-316 K for MBBA and also for the ethoxy compound. The region of decreasing permittivity also corresponds to a rapid decrease in light transmittance through specimens suggesting that this region arises due to hydrodynamic flow in the nematic and a change of molecular alignment. No detailed analysis is given for this behaviour.Studies of the liquid crystal and solid phases of MBBA are described by Moscicki and co-~orkers.~~ Below 294K MBBA has stable and metastable solid states-as may be demonstrated calorimetrically.95 Studies in the range 180-280 K for frequencies in the range 0.3-300 kHz showed that there is dipole motion in the metastable phase but none in the stable phase for this range of frequency. They suggest that the process is due mainly to the motion of methoxy groups. Similar results and conclusions are reached by Agarwal and co-worker~.~~ Moscicki and co-workers also studied the microwave dielectric properties of isotropic and nematic MBBA finding general agreement with the observations of Price92s118 except that their activation energies for the motion are lower than that of Price and co-workers.We note that Prost and Pershamloo observed flexoelectrical effects for p-butoxybenzal-p-(P-methyl butyl-aniline (BBMBA) and p-cyanobenzylidene-p -octyloxyaniline (CBOOA) in both their nematic and smectic-A phases. Relaxation times for splay-like nematic fluctuations were obtained. MBBA azo-benzenes and azoxybenzenes have a staggered molecular arrangement of aromatic rings thus their 'long-molecular axes' are not well- defined. In contrast the p-cyanobiphenyls which form nematic liquid crystals have a well-defined long axis which also is the director axis in the liquid crystal phase. Dielectric relaxation for 4,4'-n-pentyl cyanobiphenyl (PCB) and for the similar 4,4'-n-heptyl cyanobiphenyl (HCB) have been reported by Cummins and co-workers102 and by Davies and co-workers,lO1 respectively.In general terms the equilibrium permittivities and el of the magnetically or electrically aligned nematic phase are adequately described by the Maier and Meier or Bordewijk theories involving the order parameter S and angular correlations factors with indications"* that S is dependent on the strength of the directing field. ELIshows a large low frequency dispersion for both systems101*102 having a nearly single- relaxation-time contour and with relaxation times in the range 10-7-10-9 s. Davies and co-workerslO' also studied the E* dispersion whose magnitude was found to be -($) of that for with 71<~11,and which is rather broad with a Fuoss-Kirkwood distribution parameter -0.65.Using the theory of Martin et al. which relates TI(and T~ to retardation factors" yli and y1 and to an 'order-free' relaxation time TO,where 70 may be estimated from the extrapolated relaxation times of the isotropic liquid phase. With 4= 7.7 kJ mol-' very good agreement was obtained between the observed and calculated 711and T* values for a range of temperature. This implies that the simple extension of the Debye theory by Martin et al. to include the nematic ordering potential provides at least a semi- *''P. Maurel and A. H. Price Trans. Faraday SOC.,1973,69,1486. * y11and yL are functions of a nematic potential parameter q and of temperature. G. Williams and J. Crossley quantitative explanation of the relaxations parallel and perpendicular to the nema- tic direction.In physical terms the molecule behaves as a rotational diffuser with small angle steps whose molecular axis is on average preferentially aligned along the director axis. As the nematic potential q is increased 711 increases and T* decreases (711increases more rapidly than T~ decreases) and in addition AEI(/AE,. increases. The success of the simple theory for cyanobiphenyls is gratifying since in this case the analysis is uncomplicated by the presence of transverse dipole-moment components which occur in MBBA and most other liquid-crystal-forming molecules. Dielectric behaviour similar to that found by Davies and co-workers for a cyanobiphenyl has also been observed by Schadtlo4 for two alkylcyanophenyl- pyrimidines and their mixtures.This work will be discussed below in relation to parallel Kerr-electro-optical relaxation studies of these systems by Schadt. lo4 We note that Evans and Evans'" studied the far i.r. spectrum of 4,4-n-pentyl cyanobiphenyl. Using a memory function approach (involving parameters Kl(0) and rl) they have found that the low frequency dielectric relaxation behaviour"' of the nematic predicts a sharp and distinct far i.r. absorption which peaks at 137 cm-' in agreement with the observed far i.r. spectrum. Kerr-eff ect Relaxation.-The electro-optical properties of liquid-crystalline phases have been extensively studied. However information about the behaviour of the molecules is not easily obtained from such studies owing to the complexity of the mesophase at the local and macroscopic levels (see for example ref.114). For the present Report we restrict discussion to studies of the pre-transition region of the isotropic liquid and to studies in non-polar solvents. Such studies have been made by Coles and Jennings for MBBA'" and for PCB,"l by Davies and co-workers'" for HCB by Schadtlo4 for two alkylcyanophenyl pyrimidines and their mixtures and by Wong and Shen122'123 for MBBA and EBBA. The pre-transitional behaviour of the static Kerr-constant is remarkable in that it follows a Curie-Weiss type relation with temperature dependence of (T-T*)a,and where T* is a little below the clearing temperature T,. This large variation in magnitude of the Kerr-constant is due to the onset of extensive angular between the polarizable dipolar molecules in the pre-transition region and is formally represen- ted by expressions given by Kielich.12s Such angular correlations involve the molecular axes and the molecular dipole directions thus are a mixture of Pl(cos 6,) and Pz(cos 6,) terms.Since dielectric relaxation involves Pl(cos 6,) only and since the Kerr-constants are dominated by the cross-correlation terms between mole- cules for T =r T",then it follows that the dielectric relaxation times and Kerr-effect relaxation times may be quite different in the pre-transition region. This is implied from a comparison of dielectric" and Kerr-effect'" relaxation data for MBBA. As 'I9 G. J. Evans and M. Evans J.C.S.,Faraday ZZ,1977 73,285. H. J. Coles and B.R. Jennings Mol. Phys. 1976 31,571. H. J. Coles and B. R. Jennings Mol. Phys. 1976 31,1225. G. K. L. Wong and Y.R. Shen Phys. Rev. Letters 1973 30,895. G. K. L. Wong and Y. R. Shen Phys. Rev. (A),1974,10,1277. lZ4 M. S. Beevers and G.Williams J.C.S. Faraday IZ 1976 72,2171. S. Kielich in 'Dielectric and Related Molecular Processes' ed. M. Davies (Specialist Periodical Reports) The Chemical Society London 1972 Vol. 1 p. 338. Studies of Molecular Motion in Liquids and Solids 91 the temperature is decreased towards T* the Kerr-effect relaxation time TKerr increases rapidly whereas T~~~~shows only a normal small increase. TKerr >> TDiel indicating that the angular correlations between the molecules (which dominate the strength-factor of the Kerr-constant) are more slowly lost in time than the vector- correlations for the reorientation of single molecules (flip-flop motion of the molecules with respect to the local director).Such behaviour is substantiated by the work of Schadtlo4 for alkylcyanophenyl pyrimidines above the nematic-isotropic transition temperature. It was shownlo4 that the plot (1/~~~~~) against (T- T*) was ~~~ was linear for (T-T*) from 0 to 2 K. TKranged from 10-8-10-4 s while T~i~l -5 x lop9s in this range. Schadt analysed the Kerr-effect data using the Peterlin- Stuart theory which includes contributions from permanent and induced dipole moments. The analysis indicated that the Kerr-constant and thus the Kerr-eff ect relaxation was primarily due to the induced moment contribution where we would say that this was dominated by cross-correlation terms between molecules.Dynamic Light Scattering.-As with electro-optical effects many studies have recently been made of the dynamic light-scattering from liquid crystals. As exam- ples we quote work of Zulauf and co-workers 126 Kirov and co-worker~,'~' Mada and Kobayashi,128 and of Jakeman and co-worker~.'~~-'~~ The interpretation of the photon-correlation function for scattered laser light is complicated for mesophases and the problems (especially multiple scattering problems) are discussed by Zulauf and co-workers.'26 Molecular interpretations of the observations of the dynamics of the mesophase must be regarded as tenuous and such data may be more generally interpreted macroscopically using the random phase-screen model of Jakeman and co-workers (see refs.126 129-131 and refs. therein). Perhaps the most elegant use of the light-scattering technique for liquid-crystal systems is that described by Gierke and F1~gare.l~~ They studied the intensity and linewidth of the depolarized Rayleigh light scattered from MBBA in the pretransition region and observed premonitory behaviour analogous to that observed by Kerr-eff ect tech- nique~~~~ as discussed previously. The pair angular correlation function g2 was shown to vary from near unity up to about 50 and the correlation time for motion increased remarkably as T was approached. Thus the equilibrium angular cor- relations between molecules and the time-scale for the reorientation of the 'direc- tors' associated with the clusters of molecules may be readily determined by the light-scattering method.Magnetic Resonance Studies.-Studies of motion in liquid crystals using n.m.r. and e.s.r. techniques are numerous and a detailed report cannot be given here. We note that studies of the n.m.r. Tl T2,and Tl behaviour may yield information on (Pzcos O(t)) and thus the motions of the director in a mesophase. Accounts have 126 M. Zulauf M. Bertolotti and F. Scudieri J. Appl. Phys. 1975,46 5152. 127 N. Kirov P. Simova and M. Sabeva Mol. Cryst. Liq. Cryst. 1976 33 189. H. Mada and S. Kobayashi J. Appl. Phys. 1977,41,2898. 12' E.Jakeman in ref. 24 p. 75. 130 E. Jakeman P.N. Pusey and J. M. Vaughan Opt. Comm. 1976,17 305.13' P. N.Pusey and E. Jakeman J. Phys. (A) 1975,8 392. 132 T. D. Gierke and W. H. Flygare J. Chem. Phys. 1974,61. 2231. 92 G. Williams and J. Crossley been given recently for a variety of nematic~'~~-'~~ and in and of smecti~s'~'-'~~ lyotropi~s.'~~ Studies of molecular dynamics of liquid crystals by n.m.r. and e.s.r. techniques are described by several authors in the report of the 19th AmpCre Congress 1976 (see ref. 151) and will not be detailed here. 5 Molecular Crystals Whilst few recent studies of low-frequency motions in molecular crystals have been reported excellent reviews up to 1976 have been given by Sixou and Dansas,15* Brot and Lassier Gover~,'~~ The review by Sixou and Dansas15' and by P0w1es.l~~ is concerned with motion in clathrates as detected by a wide variety of techniques including those of dielectric relaxation i.r.Raman n.m.r. n.q.r. e.s.r. and neu- tron scattering. That by Brot and Lassier Gover~'~~ emphasizes mechanisms for short and long time motions and outlines the Group-Theory methods for deducing time-correlation functions for motion in classical barrier systems. Dielectric studies of low-frequency motions of HCN and MeCN in quinol clathrates are discussed by Dansas and Sixou.155*156 studied the dielectric properties Cook and co-worker~~~~ of clathrates of ethanol heptanol and chloroform with Dianin's compound. Mul- tiple relaxations were observed e.g. at 104.5,and 106.5 Hz at 92 K for the ethanol clathrate. These data were analysed in terms of the sites available in the cavities and the internal flexibility of the guest molecule.Johari and co-worker^^^*-^^^ have given accounts of the dielectric behaviour of H20 and D20 ices 158,160.161 and H20 ice (6) at low temperature^.'^^ The study161 of poly-133 P. L. Nordio and U. Segre Mol. Cryst. Liq. Cryst. 1976,36 255. 134 C. E. Tarr F. Vosman and L. R. Whalley J. Chem. Phys. 1977,67 868. 135 P. Ukleja J. Pirs and J. W. Doane Phys. Rev. (A),1976 14,414. 136 R. Y. Dong E. Tomchuk J. J. Visintainer and E. Bock Canad. J. Phys. 1976 54 1600. 137 A. Loesche S. G. Grande P. M. Borodin and Yu. V. Molchanov Kristallografiya 1976 21 856. 13' I. Zupancic V. Zagar M. Rozmarin I. Levstick F. Kogovsek and R. Blinc Solid State Comm. 1976 18 1591. 13' W.Woelfel F. Noak and M. Stohrer Z. Naturforsch 1975 30a 437. 140 R. T. Thompson D. W. Kydon and M. M. Pintar Chem. Phys. Letters 1976,42,586. 141 R. T. Thompson D. W. Kydon and M. M. Pintar J. Chem. Phys. 1974,61,4646. 142 K. Hayamizu and 0.Yamamoto J. Chem. Phys. 1977,66 1720. 143 B. M. Fung C. G. Wade and R. W. Orwoll J. Chem. Phys. 1976,64 148. 144 J. J. Visintainer R. Y. Dong E. Bock E. Tomchuk D. B. Dewey A. Li-Kuo and C. G. Wade J. Chem. Phys. 1977,66 3343. 145 I. Zupancic M. Vilfan M. Sentjure M. Schara F. Pusnik J. Pirs and R. Blinc Liq. Cryst. Ordered Fluids 1973,2 525. C. E. Tarr M. E. Field and L. R. Whalley Mol. Cryst. Liq. Cryst. 1976 35 225. 147 R. Y. Dong M. Wiszniewska E. Tomchuk and E. Bock Canad. J. Phys. 1975,53 1646. 14' F. Poddy M.Dvolaitzky and C. Taupin Chem. Phys. Letters 1976,42 449. 149 G. J. Kruger H. Spiesecke and R. Van Steelwinkel J. Phys. (Paris) Colloq. 1976 3 123. 150 M. Bloom E. E. Burnell S. B. W. Roeder and M. I. Valic J. Chem. Phys. 1977,66,3012. 15' 'Magnetic Resonance and Related Phenomena Proceedings of the 19th AmpBre Congress' ed. H. Brunner K. H. Hauner and D. Schweitzer 1976. 15' P. Sixou and P. Dansas Ber. Bunsengesellschaft Phys. Chem. 1976,80 363. C. Brot and B. Lassier Govers Ber. Bunsengesellschaft Phys. Chem. 1976 80 31. lS4 J. G. Powles Ber. Bunsengesellschaft Phys. Chem. 1976,80,259. 155 P. Dansas and P. Sixou Mol. Phys. 1976,31 1319. lS6 P. Dansas and P. Sixou Mol. Phys. 1976 31 1297. lS7 J. S. Cook R. G. Heydon and H. K. Welsh J.C.S. Faraday ZZ 1974,70 1591.G. P. Johari and S. J. Jones J. Chem. Phys. 1975,62,4213. 159 G. P. Johari and E. Whalley J. Chem. Phys. 1976 64 4484. G. P. Johari J. Chem. Phys. 1976,64 3998. 16' G. P. Johari and S. J. Jones Proc. Roy. SOC., 1976 A349,467. 14' Studies of Molecular Motion in Liquids and Solids 93 crystalline D20 ice is particularly comprehensive and reviews earlier work. A special feature is observed16' in that the plot of log (relaxation time) vs K-l is found to have a high-temperature region of constant activation energy Q but at lower temperatures the activation energy first decreases then increases with decreasing temperature. Explanations are considered,16' one of which involves change of relaxation mechanism (one dominated by the diffusion of orientational defects the other by the diffusion of intrinsic ionic defects) but it is suggested'61 that the decrease in Q near 250K is likely to be associated with impurity-generated orientational defects.It is suggested that the increase in Q at the lowest tempera- tures is due to the co-operative rearrangement of molecules being similar to that observed for amorphous materials just above the glass transition temperature (see Section 2 p. 80). This is the first crystalline solid for which such non-Arrhenius behaviour has been observed. 6 Polymers in Solution Introduction.-In view of the numerous publications concerned with molecular motion of polymers in solution we shall quote only a limited number of pub- lications and discuss some of these in greater detail.After a brief account of theoretical work we consider each experimental technique in turn making cross- reference where appropriate. Theoretical Studies.-The dynamics of polymer chains are considered in recent review^'^*.'^^ and paper^.'^^-'^* The majority of these are concerned with the generalized equations of motion of flexible chains where interactions occur within ~~~~ chains (including excluded volume eff e ~ )t and with solvent (including ~ ~ special hydrodynamic interactions"' and motions constrained to a It is often difficult to relate the results of such theories to experimental observations of molecular motion. Theories of Valeur and co-~orkers,'~~*'~~ Gotlib and co- and of Jones and stock ma ye^,'^^ which consider the motion of a flexible chain on a tetrahedral lattice are specifically aimed at an understanding of n.m.r.162 C. F. Curtiss R. B. Bird and 0.Hassager Adu. Chem. Phys. 1976,35 31. 163 W. H. Stockmayer in 'Fluides Moleculaires' ed. Balian and Weil Gordon and Breach New York 1976 p. 107. 164 P. G. deGennes (a) Macromolecules 1976,9 587; (b)J. Chem. Phys. 1977,66,4736. 16' F. Brochard and P. G. deGennes J. Chem. Phys. 1977,67,52. 166 D. A. MacInnes (a)J. Polymer Sci. (Polymer Phys.) 1977,15,657; (b)J. Polymer Sci. (Polymer Phys.) 1977,15,465. 167 W. C. Forsman J. Chem. Phys. 1976,65 4111. 16' M. Bixon J. Chem. Phys. 1977,66 5500. 169 D. E. Kranbuehl and P. H. Verdier J. Chem. Phys. 1977,67 361. 170 R. Kapral D. Ng and S. G. Whittington J.Chem. Phys. 1975,64,539. 171 B. Valeur L. Monnerie and J. P. Jarry J. Polymer Sci. (Polymer Phys.) 1975,13 675. 17' B. Valeur J. P. Jarry F. Geny and L. Monnerie J. Polymer Sci. (Polymer Phys.) 1975 13 667. 173 A. A. Jones and W. H. Stockmayer J. Polymer Sci. (Polymer Phys.) 1977,15 847. Y. Y. Gotlib A. A. Darinski and I. M. Neyelov Vysokornol. Soed 1976 Al8 1528. 17' B. K. Oh M. M. Labes and R. E. Salomon J. Chem. Phys. 1976,64,3375. 176 K. F. Freed J. Chem. Phys. 1976,64 5126. 177 K. F. Freed S. F. Edwards and M. Warner J. Chem. Phys. 1976,64 5132. 17' P. G. Wolynes and J. M. Deutch J. Chem. Phys. 1977,67 733. 179 H. J. Hilhorst and J. M. Deutch J. Chem. Phys. 1975,63 5153. P. G. Wolynes and J. M. Deutch J. Chem. Phys. 1976,65 2030. 18' M.Schwarz and D. Poland J. Chem. Phys. 1976,65 2620. 18' M. Fixman and G. T. Evans J. Chem. Phys. 1976 64 3474. G. Williams and J. Crossley dielectric and fluorescence-depolarization studies of the motion of groups within polymer chains. Valeur and co-~orkers'~~*'~~ have extended the earlier work of Dubois-Violette and co-worker~'~~ for the jump-motions of chains on the tetra- hedral lattice by allowing the chains to move uia three-bond motions and have deduced the autocorrelation functions (Pl(u, t)) and (P2(u,t)) where u = cos 8 for the motion of a unit vector which has the same direction as a representative bond-vector of the chain. It is shown that (P2(u,t)) has the form (P~(u, t))= exp (t/c)erfc -(10) (Y where erfc indicates the complementary error function and u is a characteristic relaxation time for the jump process.Thus a 'non-exponential' correlation function naturally emerges from the theory and appears to be in semi-quantitative accord with experiment (fluorescence-depolarization studies of anthracene labelled poly- styrene solutions2') if the above correlation function is multiplied by the factor exp -(t/0) which allows for stochastic motions of the lattice governed by a cor- relation time 0.Geny and Monnerie18* have demonstrated that for their model of motions on a tetrahedral lattice (Pl(u,t)) = (Pz(u,t)). They have shown that the dielectric loss-curves have a shape dependent on the ratio (@/CT). For (O/o) = 0 the curve is that for a single relaxation time process characterized by 0,for @+a the loss curve is that of a Cole-Cole function with spread parameter equal to (i) while for (0/(+)= 1 the curve is that for a Davidson-Cole function with spread parameter of (4).Experimental dielectric data for poly-p -chlorostyrene in solution are well fitted with (0/u)= 2 indicating that the motion of these chains is consis- tent with. a model in which the local units move by local conformational rear- rangements (involving u)and at the same time are being randomized by an overall rotation process (motion of the reference tetrahedral lattice involving the relax- ation time 0).Jones and stock ma ye^^^^ have noted that equation (10) is of the same form as that obtained by Glarum and by Hunt and Powles for defect-diffusion models (and applied to viscous liquids see Section 2 p.80). They'73 derived equations using a different procedure from Valeur and co-workers in which the effects of strict lattice chain directional correlations on the rearrangement kinetics are cut-off sharply rather than gradually as a function of distance from the central link of a ~rankshaft.'~~ The results are comparable numerically with those of Valeur and co-workers. Jones and Stockmayer have predicted I3C n.m.r. relax- ation behaviour using this model predicting Tl and the nuclear Overhauser enhancement behaviour for 13C-2H dipole-dipole relaxation. This theory remains to be fully tested with experimental data for polymers in solution. Dielectric Relaxation.-Studies have been made for polyalkyl isocyanate~,'~~-~~~ poly~tyrene,"~poly-p-~hlorostyrene,~~' polyvinyl acetate,192 methyl methacryl- lE3 E.Dubois-Violette F. Geny L. Monnerie and 0.Parodi J. chim. Phys 1969,66 1865. ''* F. Geny and L. Monnerie J. Polymer Sci. (Polymer Phys.) 1977 15 1. '" R. F. Boyer Rubber Chem. Technol. 1963,34 1303. 186 J. S. Anderson and W. E. Vaughan Macromolecules 1975,8 455. ''' H. J. Coles A. K. Gupta and E. Marchal Macromolecules 1977 10 182. '*' M. S.Beevers D. C. Garrington and G. Williams Polymer 1977 18 540. 189 B. L. Brown and G. Parry-Jones J. Polymer Sci. (Polymer Phys.) 1975 13 599. ''O K. Adachi I. Fujihara and Y. Ishida J. Polymer Sci. (Polymer Phys.) 1975,13 2155. 19' S. Mashimo Macromolecules 1976 9 91. ''* K. Adachi M. Hattori and Y. Ishida J.Polymer Sci. (Polymer Phys.) 1977 15 693. Studies of Molecular Motion in Liquids and Solids 95 ate/styrene co-polymer~,'~~ polypropylene poly-s -~aprolactone,~~~ ploy(&-carbobenzoxy-~-lysine),'~~ and polyvinyl chloride. 197 It is well known that polyalkyl isocyanates resemble rigid rods for molecular weights less than lo4 and tend to coil-up at higher molecular weights. Anderson and Vaughan'86 studied poly-n-hexyl isocyanate (mol. wt. = 1.0X lo5) in carbon tetrachloride n-heptane and cyclohexane over a range of temperature. They found that the variation of relaxation time with temperature and solvent was not accurately in accord with the predictions of the Kirkwood-Auer-Riseman theory for the intrinsic viscosity and dielectric relaxation of rods and rigid coils in solution and thus suggested that the molecules (i) do not behave as rigid entities and (ii) that the pitch of the helix decreases with increasing temperature.Similar conclusions were reached by Beevers and co-workersls8 from comparative dielectric and Kerr- effect studies (see below) for poly-n-butyl and poly-n-octyl isocyanates. Coles and co-workers also carried out such comparative studies for poly-n-hexyl isocyanate and this work will also be discussed below. Ishida and CO-WO~~~~S~~~*~~~~~~~ have made comprehensive dielectric studies of the systems poly~tyrene/toluene,~~~ and polyvinyl chl~ride/tetrahydrofuran,~~~ polyvinyl a~etate/toluene.'~~ The composition of a given system was varied from dilute polymer solution to bulk polymer.In addition n.m.r. linewidth and differen- tial thermal analysis (d.t.a.) measurements were made in order to determine motional narrowing behaviour and the apparent glass-transition temperatures respectively. Multiple dielectric relaxations were observed. For example for a solution of 44% polystyrene in toluene,lgO peaks were observed at 100 160 180 and 220K for 10 kHz being attributable respectively to the local motions of polystyrene and toluene the rotation of toluene phenyl group rotation and segmental motion of polystyrene. The combination of the evidence from dielectric n.m.r. and d.t.a. studies allowed Ishida and co-workers to assign the origin of each process and also to indicate the mechanism for the motion (e.g. phenyl libration in polystyrene).Kerr-effect Relaxation.-Studies have been made for the polyalkyl iso-cyanate~,~~~'~~~ and of cellulose derivative~.~~~-~" polybenzyl-~-glutamate,~~~ Coles and co-~orkers~~~ and Beevers and co-workers'88 made comparative dielectric and Kerr-eff ect relaxation studies on poly-n-hexyl isocyanate and both poly-n-butyl and poly-n-octyl isocyanates respectively. Evidence was given for association of several poly-n-hexyl isocyanates (mol. wt. from 6.2 x lo4 to 2.9 x lo5) in toluene. They18' noted that the Kerr-rise transients were slower than the corresponding decay transients being indicative that the contribution to the Kerr- effect from permanent dipole orientation far exceeded that from the induced dipole contribution. Such behaviour is also indicative of a rotational diffusion mechanism 193 Y.Iwasa and A. Chiba J. Polymer Sci. (Polymer Phys.) 1977 15 881. 194 S. Yano R. R. Rahalker S. P. Hunter C. H. Wang and R. H. Boyd J. Polymer Sci. (Polymer Phys.) 1976,14 1877. 19' A. A. Jones W. H. Stockmayer and R. Molinari J. Polymer Sci. (Polymer Symposia) 1976 54 227. 196 I. Omura A. Teramoto and H. Fujita Macromolecules 1975,8 284. K. Adachi and Y. Ishida J. Polymer Sci. (Polymer Phys.) 1976 14 2219. 19* K. Tsuji and H. Watanabe J. Chem. Phys. 1977 66 1343. 199 A. R. Foweraker and B. R. Jennings Polymer 1975 16 720. *O0 M. Isles and B. R. Jennings Brit.Polymer J. 1976 No. 3 34. 201 A. R. Foweraker and B. R. Jennings Polymer 1976 17 508. 19' 96 G.Williams and J.Crossley for the reorientation of the rod-like macromolecules. However they found that both the dielectric relaxation and Kerr-effect decay transient curves were broader than a single relaxation time process thus making difficult any comparison between the corresponding relaxation times. Beevers and co-~orkers'~~ observed similar features for poly-n-butyl and poly-n-octyl isocyanates in carbon tetrachloride. They showed that the ratio (al/a2)of permanent to induced dipole moment contributions to the Kerr-constant was -lo5. They found that al,a2,and the chain dipole moment all decreased with increasing temperature being indicative of a conformation change and/or increased flexibility of the chain (see p. 93 and ref. 186). Whilst the Kerr-effect rise-function was far slower than the decay function for both polymers and the observed TKerr,decay =(f)~~i~l, where the relax- ation time of the Kerr-effect is defined as time taken for the birefringence to decay to (l/e) of its initial value the area analysis of rise and decay transients was not in agreement with the model of small-angle rotational diffusion.An analysis of the area below the Kerr-effect decay transient defines an average relaxation time (~~,d) dt which is determined by p2(M),Ag(M) and p(M)where p Ag, =j(j/k,d(t) and p are dipole moment optical polarizability anisotropy and molecular weight (M)distribution function respectively. The average dielectric relaxation time (T~) obtained from the condition of maximum loss 27rf,,,~,,= 1 is a different average over k2(M)and p(M).Beevers and co-workers188 showed that ((TD)/(T~,~)) = AD/AK for rigid-rods and Kratky-Porod stiff-coils. Here ADand AK are defined by the relations TD= ADMI,T~.~ =AKMd,where d is a parameter. For rotational diffusion (AD/AK)= 3 so the fact that the experimental values for this ratio are near unity for poly-n-butyl and poly-n-octyl isocyanate1ss indicates that the mole- cules do not move by rotational diffusion. This evidence taken with other factors cited above and with the evidence from the works of Coles and co-worker~,~~~ Anderson and Vaughan lS6 and of Bur and Roberts202 suggests that rigid rod-like behaviour is obtained up to M -lo5 but at higher molecular weights the chains coil-up and have internal flexibility.However detailed interpretations require a knowledge of (i) the form of the molecular weight distribution and (ii) the depen- dence of p and Ag upon molecular weight and (iii) the relative importance of internal and overall rotational modes of motion. Light Scattering Studies.-Photon-correlation spectroscopy is now a well-established technique for studying the translational and reorientational motions of polymers in solution. Recent studies (up to mid-1976) have been reviewed by Cummins and PuseyZo3 in an article containing 193 references. Further studies are given in refs. 204L210 and include depolarized Rayleigh scattering studies by Wang and co-worker~.~~~-~~~ No attempt can be made to review such studies as a *02 A. J. Bur and D. E. Roberts J.Chem. Phys. 1969 51 406. 2c3 H. Z. Cummins and P. N. Pusey in 'Photon Correlation Spectroscopy and Velocimetry' eds. H. Z. Cummins and E. R. Pike Plenum Press New York 1977 pp. 164-199. 204 J. Hendrix B. Saleh K. Gnadig and L. de Mayer Polymer 1977 18 10. '05 E. Geissler and A. M. Hect J. Chem. Phys. 1976,65 103. '06 G. D. J. Phillips G. B. Benedek and N. A. Mazer J. Chem. Phys. 1976,65 1883. '07 A. M. Hecht and E. Geissler J. Chem. Phys. 1977,66 1416. '08 F. C. Chen A Yeh and B. Chu J. Chem. Phys. 1977,66 1290. 209 Y. H. Lin and C. H. Wang J. Chem. Phys. 1977,66 5578. 'lo D. R. Jones and C. H. Wang J. Chem. Phys. 1976,65 1835. 97 Studies ofMolecular Motion in Liquids and Solids whole and we shall only discuss the work of Brown and co-workers211 on ordered macromolecule solutions and of Jones and Wang2" for polypropylene glycol in solution.Brown and co-workers211 have studied aqueous dispersions of charged poly- styrene spheres (mean radius -25 nm) at very low ionic strengths. The con- ventisnal light scattering data show an angular dependence of the scattered light consistent with a structured solution in which the particles appeared to be regularly spaced with mean interparticle spacings approaching lo3nm; i.e. a distance of about twenty particle diameters. The photon-correlation data however show that the system is in a dynamic state with the polymer particles free to move in single particle Brownian diffusional motion at 'short' times and collectively at longer times.This is apparent from the dependence of the form of the intensity-intensity correlation function on the scattering angle. At a given scattering angle this correlation function decays in two parts (i) a 'short-time' free particle region and (ii) a 'longer-time' collective motion region. For large scattering angles (i.e.large values of the scattering vector k) the correlation function is dominated by the free-particle motions. Thus the free particle motions and collective motions both translational in origin may be deduced for this rather special system. It seems highly desirable to study concentrated solutions of rod-like macromolecules using the same technique where it is anticipated that the free Brownian translational and rotational diffusion and collective motions could be extracted separately.Jones and Wang2l0 studied the depolarized Rayleigh-line spectrum for poly- propylene glycols of molecular weight of 425 1025 and 2025 both for the pure liquids and for their solutions in cyclohexane. Orientational relaxation times TR in the range 0.2-2.6 ns were obtained. Comparisons were made with the dielectric data available for polypropylene glycol solutions and it appears that TR is a factor of ten larger than that for dielectric relaxation. Since the dielectric process is due to the reorientation of the ether groups it might have been hoped that TR==T~i~l. Clearly a systematic comparison of correlation times obtained from Rayleigh line-broadening studies dielectric studies and n.m.r. studies of polypropylene glycol and similar polymers (i.e.polymers not possessing rotatable side-groups) is highly desirable and will be required in order to understand fully the motions being observed in any given experiment.Fluorescence Depolarization Studies.-It has been indicated (see p. 93) that several workers have deduced the time correlations for the reorientational motions of a model chain on a tetrahedral lattice. Valeur and Monnerie2' have carried out studies of the time-dependent fluorescence depolarization of solutions of poly-styrene chains containing anthracene or 9-styryl- 10-phenyl anthracene in the main chain in order to see if such models represent the chain dynamics. The emission anisotropy factor r(t)=& (P2(u,t)) for the motion of the fluorescent group.It was found that equation (10) with the added factor [exp- (t/O)]gave a good represen- tation of the data and that (O/u)>1 so both conformational and overall rotational modes contribute to r(t). Also (O/u) decreased with increasing solvent viscosity (the solvent was ethyl acetate/tripropionin). This indicates that the relative contributions of internal and overall motions to the correlation function (and hence 21 1 J. c.Brown P. N. Pusey J. W. Goodwin and R. H. Ottewill J. Phys. (A),1975,8 664. G. Williams and J. Crossley its shape) depend upon solvent viscosity. Clearly further evidence is required from related experimental techniques in order to be more certain of the relative contri- butions being made by the competing mechanisms for motion.N.M.R.-Reviews are available212 of studies of molecular motion using n.m.r. techniques covering the period June 1975 to May 1977. The advent of 13C n.m.r. techniques has opened up new and powerful methods for studying the motions of specific carbon atoms for polymers in solution. As examples of these studies we may refer to the recent work of Bovey and co-~orkers~’~-~~~ ‘-* and of Heatley and co-wor k ers . The paper by Cais and Bovey213 on the 13C n.m.r. behaviour of polystyrene peroxide in solution serves as an excellent review and source of reference to earlier 13C n.m.r. studies of structure and dynamics. Tl values were given for all five resolvable carbons in the equimolar polystyrene peroxide and polystyrene copolymer together with the T1values for the corresponding carbons in polystyrene.The solvents were chloroform deuteriochloroform and benzene. The 2 :1 ratio of Tl values for the C and CB carbons confirms the dipole-dipole nature of the relaxation mechanism. A comparison of the Tlvalues for the main chain carbons of the polyperoxide with those for polystyrene show that the poly- peroxide backbone reorientates with about one half of the correlation time of that for polystyrene. For polystyrene the correlation times of the phenyl C2 C3 and C4 carbons are comparable with those for the C carbon indicating limited librational motion of the phenyl ring. For the polyperoxide however the correlation times for the C2 and C3 carbons are about 0.6 of those for the C and C4 carbons showing that the phenyl groups here enjoy a measure of additional rotational freedom by comparison with those in polystyrene.This is physically reasonable owing to the fact that the additional oxygen in the chain is less sterically restrictive than a 0-methylene group. Thus the 13C n.m.r. technique gives information on the motions of specific carbons in the chain. Cais and Bo~ey*~” studied polybut-1-ene sulphone (7) and polystyrene sulphone (8) in chloroform solution. For compound (7) the values of Tl for backbone CH and CH2 units and the ethyl branch CH2 and CH units were found to be the same for molecular weights of 48 600 182 000 and 272 500. Also the nuclear Overhauser enhancement factor was independent of molecular weight. These results indicate that the polymer units move by segmental motions and that these occur at a rate which is independent of molecular weight.In contrast earlier dielectric studies2lg had shown that T~ was a function of molecular weight. This paradox was discussed by Cais and Bovey who conclude that the segmental motions in (7) are not as severely hindered as had been They suggest a possible mechanism for motion which allows the CH groups to reorientate but which does not reorientate the sulphone dipole. The mechanism is based on the concerted motion of a C-S-C-C-S-C sequence according to a ‘second ’’’ (a) F. Heatley in ‘Nuclear Magnetic Resonance’ ed. R. J. Abraham (Specialist Periodical Reports) The Chemical Society London 1976 Vol. 6 p. 141 (6) ibid. 1977 Vol. 7 in the press.’13 R. E. Cais and F. A. Bovey Macromolecules 1977 10 169. ’14 F. A. Bovey F. C. Schilling T. K. Kwei and H. L. Frisch Macromolecules 1977 10 559. ’”(a) R. E. Cais and F. A. Bovey Macromolecules 1977 10 752; (6) 1977 10,757. ’I6 F. Heatley and A. Begum Polymer 1976.17 399. ’I7 F. Heatley and M. K. Cox Polymer 1977,18 225. ‘18 F. Heatley A. Begum and M. K. Cox Polymer 1977,18 637. ’19 T. W. Bates K. J. Ivin and G. Williams Trans. Faraday Soc. 1967,63 1964. Studies of Molecular Motion in Liquids and Solids 99 type’ mechanism classified by Helfand.220 Concurrent with the publication by Cais and Bo~ey,~~’* studies of the 13C n.m.r. relaxations of poly-o-methyl pentene sulphones were published by Stockmayer and co-workers221a and for fifteen poly- alkyl sulphones by Fawcett and co-workers;222 both studies show segmental motion as the origin of the relaxations in accord with Cais and Bo~ey.~~~’ Stockmayer and co-workers221a have tentatively rationalized the anomaly between the 13Cn.m.r.data and dielectric relaxation data219 in terms of a mechanism for four-bond rearrangements which reorientate CH bonds but not sulphone dipole groups. It should be apparent from these examples and the recent work of Stockmayer and co-workers221b that 13Cn.m.r. relaxation especially when considered together with dielectric Kerr-eff ect and fluorescence depolarization results represents a most promising technique for the study of the motions of specific groups in chains. Its only drawback is that the correlation times T, deduced from TIand nuclear Overhauser enhancement factors are integrals over correlation functions in time (i.e.they are transport coefficients). It is therefore not easy to distinguish between models for motion based on a single exponential or a sum of exponential cor- relation functions in time. However this is not of great importance in distinguish- ing between the rates of motion for individual groups and is of no importance in distinguishing between segmental and overall motions -as was achieved for the polyolefin sulphones .215*221a*222 We note that in addition to the 13C n.m.r. and fluorescence-depolarization techniques for following the motions of specific groups in polymer chains such motions may also be studied through the motions of nitroxide spin-labels covalently attached to the chains.223 Veksli and Miller223 have given rotational correlation times (10-8-10-10 s) obtained from e.s.r.spectra for three different nitroxide labels attached to polystyrene. 7 Polymers in Bulk Introduction.-Reviews of molecular motion for solid polymers have been given by Hed~ig,~~~ Boyd,226 Wada,227 and by Allen and Watts.228 Only a selec- B~yer,~~’ tion of recent papers will be referenced in this Report. Whilst we shall not discuss the mechanical relaxation behaviour we note that B~yer~~’ has given an appraisal of recent mechanical relaxation studies and has suggested topics for future research. We first indicate some recent theoretical studies and then consider experimental results taking each technique in turn and cross-referencing where appropriate.220 E. Helfand J. Chem. Phys. 1971,54,4651. 221 (a) W. H. Stockmayer A. A. Jones and T. L. Treadwell Macromolecules 1977 10 762; (b) K. Matsuo K. F. Kuhlmann H. W. H. Yang F. Geny and W. H. Stockmayer J. Polymer Sci. (Polymer Phys.) 1977 15 1347. 222 A. H. Fawcett F. Heatley K. J. Ivin C. D. Stewart and P. Watt Macromolecules 1977,10 765. 223 Z. Veksli and W. G. Miller Macromolecules 1975 8 249. 224 P. Hedvig ‘Dielectric Spectroscopy of Polymers’ Adam Hilger Ltd. Bristol 1977. 225 R. F. Boyer Polymer 1976,17,996. 226 R. H. Boyd in ‘Molecular Basis of Transitions and Relaxations’ Midland Macromolecular Institute Monograph ed. D. J. Meier 1976 No. 4. ’”Y. Wada in ‘Dielectric and Related Molecular Processes’ ed.M. Davies (Specialist Periodical Reports) The Chemical Society London 1977 Vol. 3 p. 143. ”* G. Allen and D. C. Watts in ‘International Review of Science Phys. Chem. Series 2 Vol. 2 Mol. Struct. and Props. ed. A. D. Buckingham Butterworths 1975. G. Williams and J. Crossley Theoretical Studies.-Shore and Zwanzig2*’ considered relaxation for a one-dimensional Ising-lattice of coupled rotators. The resulting auto-correlation function.for dipole motion was non-exponential in time and resembled that for the Williams-Watts empirical relation @(t) = exp -(t/~)’ with y = 0.5. This gives a reasonable representation of the dielectric a relaxation for amorphous solid polymers. Jons~her~~~~~~ noted that the plots of log x” us.log(frequency) for published data for a wide variety of amorphous polymers were linear for w7~1 and for WT >> 1. He proposed the empirical relation (5) =(;Jm+(E)l-where x” is the imaginary part of the complex dielectric susceptibility and wl,02,n and rn are constants to be determined from experiment. He suggested that a general feature of polymers would be that the ratio of the real to imaginary parts of ,y would be a constant independent of frequency for frequencies greater than that for maximum loss. Given this phenomenological description of the dielectric relaxations (for both a and p processes) Jonscher has sought an interpretation in terms of a model of the ‘screened’ hopping of charges or dipoles. The rate equations and hence the form of the relaxation function for such motions have not yet been formulated thus it remains to be seen if the model can predict the shape of the observed relaxations.Brereton and Da~ies~~~ have proposed an explanation of the universal occur- rence of two relaxations (aand p)for simple glass-forming systems. Their basic model is a two-site model for which the dipole reorientates between the sites according to classical first-order rate equations. Correlation of motions between units is explicitly introduced by allowing the energy difference W between the sites to be determined in part by the relative occupation of the two sites according to the relation nl and n2 are the average occupation numbers of the sites Wois a constant and T is a critical temperature.They find that the model predicts one process which takes on different forms for T< T and for T>T,.For T>T the process is charac- terized by a single relaxation time and has a temperature dependent activation energy Q which increases with decreasing temperature. At T=T,there is a discontinuity in the relaxation time and for T<T the relaxation time first decreases then increases with decreasing temperature. For T > T,the static permittivity so decreases with increasing temperature and for T < T, E~ decreases with decreasing temperature. Clearly the qualitative features of the model for T> T,and T< T,are in accord with observed behaviour with T,resembling Tg. However only one process the site-hopping process is involved at all temperatures.We have seen (Section 2) that for certain viscous glass-forming liquids and for amorphous solid polymers containing flexible side-chains both a 229 J. E. Shore and R. Zwanzig J. Chem. Phys. 1975,63,5445. 230 A.K.Jonscher Kature 1975 253 717. *’’ A.K.Jonscher Koll. Z. 1975 253 231. 232 A. K.Jonscher Nature 1977,267,673. 233 M.G.Brereton and G. R. Davies Polymer 1977,18 1764. Studies of Molecular Motion in Liquids and Solids 101 and p processes are observed for T > Tgwhere the p process is a continuation of the process observed in the glass. Equation (5) has been used by Williams and to rationalize these and other observations of multiple relaxations in glass-forming systems. It remains to be seen if the model of Brereton and Da~ies~~~ gives a more adequate explanation of observed behaviour than that suggested earlier.Solunov and P~nevsky*~~-~~~ have given interesting accounts of the deter- mination of the temperature dependence of relaxation times from limited exoeri- mental data,234 for the evaluation of apparent activation energies from data at two frequencie~~~’ and for the evaluation of relaxation parameters from thermally stimulated current data.236 In all cases the time-temperature superposition prin- ciple is assumed. Dielectric Relaxation.-Studies have been reported for oxidized and chlorinated p~lyethylene,~~’ alternating ethylene/carbon monoxide poly-ethylene,239 ethylene/vinyl acetate chlorinated polyethylene vul- ~anizate,~~l*~~~ polytrifl~oroethylene,~~~ poly~hloroprene,~~~ polyalkyl methacryl- ateS,245-248 p~lyalkylacrylates,~~~ polystyrene/plasticizer styrene/N-substituted maleimide copolymers,251 styrene-acrylonitrile copolymer^,^'^ poly-p~lycarbonate,~’~-~’~ ethylene te~ephthalate,~~~-~~~ polyester^,*^^-^^' polyvinyl polyvinylidene fl~oride,’~~-~~~ chloride,262 polychlor~trifluoroethylene,~~~ poly-234 Ch.Ponevsky and Ch. Solunov J. Polymer Sci. (Polymer Phys.) 1975,13 1467. 235 Ch. Solunov and Ch. Ponevsky J. Polymer Sci (Polymer Phys.) 1976,14 1801. 236 Ch. Solunov and Ch. Ponevsky J. Polymer Sci. (Polymer Phys.) 1977,15,969. 237 C. R. Ashcraft and R. H. Boyd J. Polymer Sci. (Polymer Phys.) 1976,14 2153. 238 H. Starkweather J. Polymer Sci. (Polymer Phys.) 1977 15,247. 239 P.Fischer and P. Rohl J. Polymer Sci. (Polymer Phys.) 1976,14 531 543. 240 M. E. Baird and E. Houston Polymer 1975,16 308. 241 M. Naoki and T. Nose J. Polymer Sci. (Polymer Phys.) 1975,13 1747. 242 M. Naoki M. Motomura T. Nose and T. Hata J. Polymer Sci (Polymer Phys.) 1975,13 1737. 243 V. P. Petrosyan U. Berner and Sh. T. Yegurtdzhyan Vysokomol. Soedin.,1976 A18 1376. 244 C. L. Choy Y. K. Tse S. M. Tsui and B. S. Hsu Polymer 1975,16,501. 245 (a)K. Shimizu 0.Yano and Y. Wada Rep. Progr. Polym. Phys. (Japan) 1975,18,387;(b) J. Polymer Sci. (Polymer Phys.) 1975,13 1959. 246 A. M. North R. A. Pethrick and D. W. Phillips Polymer 1977 18 324. 247 Id Kryszewski M. Zielinski and S. Sapieha Polymer 1976 17 213. 248 T. Tetsutani M. Kakizaki and T. Hideshima Rep.Progr. Polym. Phys. (Japan) 1976,14 307. 249 A. Tanaka and Y. Ishida J. Polymer Sci. (Polymer Phys.) 1975,13,436. P. J. Hains and G. Williams Polymer 1975 16,725. ’” H. Block P. W. Lord and S. M. Walker Polymer 1975,16 739. 252 M. Cook T. T. Jones and G. Williams Polymer 1975,16 835. 253 K. Sawada and Y. Ishida J. Polymer Sci. (Polymer Phys.) 1975,13 2247. 254 B. S. Hsu and S. H. Kwan J. Polymer Sci. (Polymer Phys.) 1976 14 1591. 255 B. S. Hsu S. H. Kwan and L. W. Wong J. Polymer Sci. (Polymer Phys.) 1975,13 2079. 256 E. Sacher J. Macromol. Sci. 1975 Bll 403. 257 (a)Y. Aoki and J. 0. Brittain J. Appl. Polymer Sci. 1976 20 2879; (b) J. Polymer Sci. (Polymer Phys.) 1976,14 1297; (c)J. Polymer Sci. (Polymer Phys.) 1977,15 199. E. Ito K. Sawamura and S.Saito Koll. Z. 1975,253,480. 259 C. Lacabanne and D. Chatain J. Polymer Sci. (Polymer Phys.) 1975,79 283. 260 C. Lacabanne D. Chatain J. Guillett G. Seytre and J. F. May J. Polymer Sci. (Polymer Phys.) 1975 13,445. ’“ M. Ito S. Nakatani A. Gokan and K. Tanaka J. Polymer Sci. (Polymer Phys.) 1977 15 605. 262 Y. Kihara K. Matsusaka and I. Murakami Polymer 1975 16,265. 263 G. Samara and I. J. Fritz J. Polymer Sci. (Polymer Letters) 1975,13 93. 264 M. G. Brereton G. R. Davies A. Rushworth and J. Spence J. Polymer Sci. (Polymer Phys.) 1977,15 583. 265 M. Murayama and H. Hashizume J. Polymer Sci. (Polymer Phys.) 1976,14 989. 266 H. Ohigashi J. Appl. Phys. 1976 47 949. 102 G. Williamsand J. Crossley diethyl ~iloxane,~~~ and polyvinyl polypropylene glycols,268 polyvinyl carba~oles,~~~ an thracenes 269 hexafluoroisobutylene/ 1,l difluoroethylene alternating and solid solutions of n-butyl-4,5,7 trinitrofluorenone-2-carboxylate with p~lycarbonate.~~~ The large number of studies reflect the continuing interest in polymers whose properties are generally well-established (e.g.polyethylene and polyalkyl methacrylates) and the interest in polymers and polymer systems which may have great practical application in the future (e.g. polyvinyl carbazole). Much of the work in refs. 237-271 involves a careful documentation of the multiple dielectric relaxation processes (a,0 y etc.) over a wide range of frequency temperature and in some instances of applied For amorphous solid polymers the a and p relaxations have common features with the cor- responding relaxations observed for small-molecule glass-forming systems and since their mechanisms have been discussed frequently in recent (see pp.80 and loo) no further discussion is necessary in this Section. For crystalline polymers there continues to be difficulty in assigning the origin of individual processes. Over the past twenty years there has been considerable effort directed towards understanding the origin of the a 0 and y relaxations of different polyethylenes. In a particularly valuable paper Ashcraft and have given dielectric data over the range lo-lo4 Hz for low-density and high-density poly- ethylenes which were (i) oxidized to give 0.5-1.7 carbonyl groups per 1000 CH2 units and (ii) chlorinated to give 14-22 chloride groups per 1000 CH units.The dielectric evidence suggested that the dipolar groups were both in the crystalline and non-crystalline regions of the materials but with selective partitioning favour- ing their presence in the non-crystalline regions. From a critical analysis of their data taken together with the many earlier studies they concluded (i) that the a process is due to the motions of non-rigid (twisting) chains in the crystalline regions -in general agreement with an earlier analysis272; (ii) that the p process in oxidized low-density polyethylene had relaxation parameters similar to those for a dynamic glass rubber transition; and (iii) that the y process observed in all cases was due to motions in the ‘amorphous’ regions and that these motions occurred by local conformational changes in which the units move in a barrier system deter- mined by internal chain energetics and by environmental packing energetics.The latter factor is thought to give rise to a range of barriers since the chains may find themselves in a variety of environments. This leads to a distribution of relaxation times and hence very broad loss-curves (of half-width -4 decades of frequency). This model for the y process in polyethylenes appears to be similar to that proposed for the small broad p process of amorphous solid polymers and other glass-forming systems (see Section 2 p. 80). N.M.R. Relaxation.-Broad-line (and hence T2)proton n.m.r. studies have been made for branched and linear polyethylenes273 for isotropic and drawn high-density 267 J.M. Pochan C. L. Beatty D. F. Hinman and F. E. Karasz J. Polymer Sci.(PolymerPhys.) 1975,13 977. 268 T. Alper A. J. Barlow and R. W. Gray Polymer 1976,17 665. 269 J. M. Pochan and D. F. Hinman J. Polymer Sci. (Polymer Phys.) 1976 14 2285. 270 J. M. Pochan D. F. Hinman M. F. Froix and T. Davidson Macromolecules 1977,10 113. 271 J. M. Pochan D. F. Hinman and S. R. Turner J. Appl. Phys. 1976,47,4245. 272 J. D. Hoffman G. Williams and E. Passaglia J. Polymer Sci.,C 1976 14 173. 273 V. J. McBrierty and I. R. McDonald Polymer 1975.16 125. Studies of Molecular Motion in Liquids and Solids 103 polyethylene^,^^^ for oriented polyethylene tere~hthalate,~~~ for several aromatic and polychlorotrifluoroethylene Proton n.m.r.TI and T2 studies have been made for p~lystyrene,~~' styrene/N-substituted maleimide~,*~~ polyhexamethylene sebacate2" and a macroscopic 'crystal' of styrene/butadiene co-po1ymer.281 Such studies are concerned with a careful documentation of proton n.m.r. relaxation data for the multiple relaxations which occur and the origin of these processes and their mechanisms may be established. For example McBrierty and McDonald273 have assigned mechanisms for the CY p and y processes in linear (high-density) and branched (low-density) polyethylenes. These are in general agreement with the mechanisms suggested by Ashcraft and Boyd237 and have been discussed previously (see p. 101). The possibility of the anisotropic motion of A continuing difficulty with such n.m.r.chain-units can be e~amined.~'~-~'~ experiments is that measurements are commonly and conveniently carried out at a single resonance frequency using temperature as the means to vary TIand T2 in terms of correlation times and their attendant distributions. Kimmich and S~hmauder~'~*~*~ have discussed the TI relaxation times of polypropylene glycols of molecular weights 1000,4000 10000 and 27 000 which had been determined earlier by Preising and Noak286 over the range 104-108.5 Hz at a temperature of 70°C. Given such comprehensive data in the frequency domain Kimmich and S~hmauder~'~ were able to resolve three distinct motional processes. These were assigned to (i) rapid anisotropic segmental motions (ii) longitudinal chain-diff usion in a 'tube' ('reptation') and (iii) configurational fluctuations of the 'tube'.The 13Cn.m.r. techniques for solid polymers have now been de~eloped~'~-~'~ to the stage that high resolution spectra may be obtained which are comparable with those obtained from solution measurements. Given such resolution Schaefer and co-worker~~'~ were able to measure TI Tlp,nuclear Overhauser enhancement factors and cross polarization relaxation quantities for a number of glassy solid polymers including polymethyl methacrylate polycarbonate polyphenylene oxide polystyrene polysulphone polyether sulphone and polyvinyl chloride. This work represents a remarkable advance in experimental technique. The detailed inter- pretations of motion resulting from those experiments are not complete since further measurements are required over a wide range of temperature but it is clear that when the results of such studies are available and are compared with those for dielectric and mechanical relaxation studies our understanding of the detailed 274 J.B. Smith A J. Manuel and I. M. Ward Polymer 1975 16,57. 27s A.Cunningham A. J. Manuel and I. M. Ward Polymer 1976 17,125. 276 V. Frosini G. Levita J. Landis and A. E. Woodward J. Polymer Sci. (Polymer Phys.) 1977 15 239. 277 D. C. Douglas V.J. McBrierty and T. A. Weber J. Chem. Phys. 1976,64 1533. 278 M. F. Froix D. J. Williams and A. 0.Goedde Macromolecules 1976 9 354. 279 H. Block D. Evans and S. M. Walker Polymer 1977 18,786. 280 M. F. Froix and A. 0.Goedde J.Macromol. Sci. B 1975 11,345. G. E. Wardell D. C. Douglas and V. J. McBrierty Polymer 1976 17,41. 282 D. C. Douglas V. J. McBrierty and T. A. Weber Macromolecules 1977,10 178. "' C. R. Dybowski and R. W. Vaughan Macromolecules 1975,8 50. 284 R. Kimmich Polymer 1977 18,233. 28s R.Kirnmich and Kh. Schmauder Polymer 1977,18 239. G. Preising and F. Noak Progr. Coil. Polymer Sci. 1975 57,216. 287 J. Schaefer E. 0.Stejskal and R. Buchdahl Macromolecules 1975 6 291. 288 J. Schaefer and E. 0.Stejskal J. Amer. Chem. SOC. 1976 98 1031. 289 J. Schaefer E. 0.Stejskal and R. Buchadahl Macromolecules 1977,10 384. 104 G. Williams and J. Crossley motions of individual groups will be greatly enhanced. Komoroski and co-worker~~~~*~~~ have reported 13Cn.m.r.spin-relaxation parameters for amorphous and partially crystalline ~is-polyisoprene~~~ and for linear For cis-polyisoprene these data at 67.9 MHz were compared with earlier data for TI at 22.6 MHz obtained by S~haefer.~~~ These data290-292 give valuable new informa- tion on chain mobility but a fuller understanding of the nature and extent of the motions requires data for a wide range of temperature and additional information to be provided by other techniques. E.S.R. Spin-probe Studies.-The motions of spin-probes (e.g. nitroxides) incorporated into a solid polymer may be studied by e.s.r. techniques and provide a measure of molecular mobility in the polymer. Recent studies using this technique have been made for a p~lyamide,~~~ acrylonitrile/butadiene and for polysil~xanes,~~~ polyvinyl chloride,295 poly~tyrene,~~’ p~lybutadiene,~~’ poly-carbonate,295 p~lyethylene,~~’ polyvinyl fluoride and polyvinylidene polyethylene oxide and poly~xymethylene,~~~ ethylene/propylene co-p01yrners,~~’ ethylene/vinyl acetate ~o-polymers,~~~ In and styrene/butadiene co-polyrner~.~~~ the comprehensive study by Kumler and B~yer~~’ the average e.s.r.line-narrowing temperature ‘T50G’was found to be well-correlated with the apparent glass-tran- sition temperature T obtained by conventional means with ‘T50G’and Tg cor-responding to frequencies of observation of about lo7Hz and 1Hz respectively. Piezo-electric Relaxation and Related Phenomena.-Molecular motions in solid polymers give rise to both mechanical and dielectric relaxation processes.As a result the electro-mechanical properties of solid polymers exhibit complicated behaviour as a function of frequency and temperature which find interpretations in terms of the multiple relaxations observed by the normal mechanical and dielectric techniques. Reviews of this important subject which includes piezo-electrical and electrostrictional relaxation are Of the many papers which have appeared in the past three years those by Koga and co-w~rkers~~’ for poly-y- methyl glutamate and by Furukawa and F~kada~~* for poly-y-benzyl glutamate may be taken as representative of the piezo-electric behaviour of films formed from rod-like macromolecules. 8 Conclusions We have seen that studies of molecular motions of a wide variety of systems have been actively pursued using a wide range of experimental techniques.The primary objectives of identifying the motional units and specifying their rate of motion as a function of temperature pressure and composition have been achieved for many systems even for those of considerable chemical and physical complexity. It is 290 R. A. Komoroski J. Maxfield and L. Mandelkern Macromolecules 1977 10 545. 291 R.A.Kornoroski J. Maxfield F. Sakaguchi and L.Mandelkern Macromolecules 1977,10 550. 292 J. Schaefer Macromolecules 1972 5,427. 293 P.Tormola KoN. Z. 1977,255 209. 294 N.Kusumoto S. Sano N. Zaitsa and Y. Motazato Polymer 1976 17 448. 295 P.L.Kumler and R. F. Boyer Macromolecules 1976,9,903. 296 R.Hayakawa and Y.Wada Adu. Polymer Sci. 1973,11,1. 297 K. Koga T. Kajiyama and M. Takayanagi J. Polymer Sci. (Polymer Phys.) 1976,14 401. 298 T. Furukawa and E. Fukada J. Polymer Sci (Polymer Phys.) 1976 14 1979. Studies of Molecular Motion in Liquids and Solids apparent however that comparisons between the results from different experi- ments for a given system have not been frequently made. It is hoped that such comparisons will be made for future studies in order that a unified interpretation can be made in terms of detailed mechanisms for motion. This is particularly important in relation to the newly-developed 13Cn.m.r. techniques which are able to follow the motions of specific CH bonds in a molecule but owing to their apparent limitation to only a limited number of frequencies require additional information from other experiments in order to specify relaxation mechanisms.The authors acknowledge helpful correspondence from Drs. F. Heatley and P. N. Pusey regarding n.m.r. studies and inelastic light-scattering studies respectively.

ISSN:0308-6003    

DOI:10.1039/PR9777400077

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

PART II INORGANIC CHEMISTRY 6 Introduction By M. F. LAPPERT School of Molecular Sciences University of Sussex Brighton BN 1 9QJ The general form of the 1977 Annual Reports (Vol 74) on Inorganic Chemistry follows the pattern set last year and is written by much the same group of people except that Chapter 7 Part 11 Group 3 now has Prof. G. E. Toogood as the sole author whereas Chapter 9 while still in the hands of Drs. D. J. Cardin and K. R. Dixon also has Drs. C. J. Cardin and R. J. Norton as participants. The aims and limitations of the coverage have been amply discussed in the Introductions to the two previous Reports (Vols. 72 and 73). We therefore merely emphasise that there is a considerable element of selection in the material discussed but that hopefully over the three year period a reasonable balance has been provided.At the outset (Vol. 72) this group of reporters was faced with a new problem in that Specialist Periodical Reports had become available over a wide front of Inorganic Chemistry as a result of which it clearly became desirable that our treatment of the literature be different and distinct. Now that we are coming to the end of our tenure we note with disquiet that the Chemical Society has discontinued a number of the S.P.R.’s most notably as far as Inorganic Chemistry is concerned those dealing with respectively main-group element and transition-metal chemistry. There is little doubt that the Annual Reports of the future will have to respond appropriately to this new challenge.Over the three years we have gradually tightened up standards with respect to the listing of numerical results in SI units and the adoption of stricter rules on the presentation of formulae and in systematic nomenclature. We trust the same pattern will be followed by our successors. During the year an important advanced text book’ has appeared. There is little doubt in my mind that it will become one of the more influential of the advanced single volume treatises. The approach is distinct from its competitors. It is not based primarily on the Periodic Table as a foundation and descriptive chemistry starts rather late (Chapter 13 out of a total of 19 chapters corresponding to 694 pp. out of a total of 11 16 pp.) with co-ordination chemistry through organometallic chemistry to molecular polyhedra (boron hydrides and metal clusters being treated together) and ends with biochemical applications.In the area of applied inorganic chemistry a new edition of Kirk Othmer is timely. Of more immediate use to students and teachers however will be a Chemical Society special publication2 dealing with the inorganic chemicals industry; it is written by industrial chemists. ’ K. F. Purcell and J. C. Kotz ‘Inorganic Chemistry’ W. B. Saunders Company Philadelphia 1977. ’ The Modern Inorganic Chemicals Industry’ ed. R. Thompson Chern. SOC.Special Publ. No.31 1977. 109 110 M. F. Lappert The subject of inorganic rings and chains is covered in ref. 3 and ref. 4 provides a much needed compilation of computer-analysed thermochemical data.The elements phosphorus,’ gold,6 and mercury7 are well served by the appearance of useful monographs; one of these is a multi-author Finally attention is drawn to the continuing and increasing activity by the Gmelin Institute.8 Recent additions have included texts dealing with compounds of boron,8a tin,86 iron,“ sulphur-nitrogen,8d and tellurium.8 I should like to end on some personal notes first in recording my thanks to the contributors and to the editorial staff (Books Section) of the Chemical Society. The authors deserve credit for producing their interesting manuscripts on time and I regard as a notable tribute to Philip Gardam and his CS colleagues the fact that only about six months elapsed between his receipt of typewritten manuscripts and the appearance of well produced and bound volumes.Finally I am pleased to note that the Chemical Society has so far resisted and hopefully will continue so to do the temptation to produce Annual Reports by direct photographic reproduction of typescripts. Homoatomic Rings Chains and Macromolecules of Main Group Elements’ ed. A. L. Rheingold Elsevier Amsterdam 1977. J. B. Pedley and J. Rylance ‘Computer analysed thermochemical data organic and organometallic compounds’ Sussex University Press 1977. D. E. C. Corbridge ‘Phosphorus an Outline of its Chemistry Biochemistry and Technology’ Elsevier Amsterdam 1978. R. J. Puddephatt ‘The Chemistry of Gold’ Elsevier Amsterdam 1978. ’‘The Chemistry of Mercury’ ed. C. A. McAuliffe Macmillan London 1977 * Gmelin L. ‘Handbuch der Anorganischen Chemie’ New Supplement to 8th Edition Springer Berlin (a)Vol. 37 Part 10; Vol. 42 Part 11; Vol. 43 Part 12; and Vol. 44 Part 13; (b) Vol. 35 Part 4; (c)Vol. 36 Part 13; (d) Vol. 31 Vol. 32; (e) 8th Edition Main Series Supplement Vol. B.

ISSN:0308-6003    

DOI:10.1039/PR9777400107

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

7 The Typical Elements ByR. H. CRAGG Department of Chemistry University of Kent Canterbury CT2 7NH J. D. SMITH School of Molecular Sciences University of Sussex Brighton BN 1 9QJ G. E. TOOGOOD Department of Chemistry University of Waterloo Waterloo Ontario Canada N2L 3G 1 PART 1 Groups I and I1 By R. H. Cragg 1 Group1 The quantitative determination of 6Li and 'Li in Rhine water and the isotopic distribution in a number of commercial lithium salts have been evaluated by the application of FD mass spectrometry in combination with a multichannel analyser. 6Li and 'Li were determined quantitatively (with a total error of 5%) by the isotopic dilution method using 6Li-enriched lithium fluoride as the standard. The results obtained showed that the isotopic distribution in lithium salts deviated from the mean natural abundance in contrast to the natural distribution found in the case of samples of Rhine water.The results of a number of theoretical studies on lithium compounds have been reported. For example the information obtained from accurate ab initio configuration interaction calculations has been used to obtain the potential energy curves of the ground state of LiH- and LiH.2 For LiH the calculated values of the bond length dissociation energies and spectroscopic constants obtained from the potential energy curve were within 1% of the experimental values. An SCF correlation diagram has been reported for low-lying linear and bent states of Li2H and SCF potential curves have been computed for the collinear reactions H +Liz and Li +LiH.3 The results indicated that Li2H at the present an unknown molecule may be obtained from the interaction of Li and LiH.Pressure-composition-temperature data have been collected for the Li-LiD system over the ranges 0-750 Torr 1-99 mol YO LID and 705-871 "C by measur- ing the equilibrium deuterium pressures over encapsulated Li-LiD mixture^.^ The information obtained enabled a series of five Pb,us. NLiDisotherms to be con- ' W. D. Lehmann and H. R. Schulten Angew. Chem. Internat. Edn.,1977,16 852. 'B. Liu K. 0.Ohata and K. K. Docken J. Chem. Phys. 1977,67 1850. W. B. England N. H. Sabelli A. C. Wahl and A. Karo J. Phys. Chem. 1977,81 772. E. Veleckis J. Phys. Chem. 1977 81,526. 111 R. H. Cragg J.D. Smith and G. E. Toogood structed whose shapes suggest the existence of two homogeneous terminal solutions which are separated by a wide miscibility gap. Two synthetic methods for the formation of Li2Sn in certain aprotic media have been rep~rted.~ These involve either the direct electrochemical reduction of S8 or alternatively a direct in situ reaction of Sg with Li or Li2S. The phosphorescence spectra of a series of group I and group 11 metal (M) acetylacetonates (M = Li Na K Rb Cs Mg Ca Sr or Ba) have been measured in solid ethanol glass solutions at 77 K and their spectral profiles have been observed to be similar to that of [ Al(a~ac)~]. A detailed description of the measurement conditions and data treatment used to obtain accurate rate constants and activation energies from 23Na pulsed Fourier- transform data has been rep~rted.~ The technique has been applied to a study of the interaction of Na' in the free state and its complex with 2,2,2-cryptand in four solvents (edta py THF and H20).The results indicate that the exchange process proceeds uia dissociation of the complex. The rewlts of a study* by 39K n.m.r. for a series of potassium salts in a range of solvents in the 0.07-1.0 moll-' concentration range demonstrate that there is a reasonable linear relationship as has been observed from 23Na n.m.r. studies on sodium salts between the magnitude of the chemical shift in a given solvent and the donicity of the latter expressed by the Gutmann donor numbers. The combination of e.s.r.and e.s.r.-optical techniques has been applied to a study of M-THF (M = K Rb or Cs) solutions in the presence of dicyclohexyl-18- crown-6 (CR).9 The recombination process of the photoelectrons in the M-THF systems followed pseudo-first-order kinetics and was dependent upon the CR concentration. The results were interpreted in terms of two competitive reactions namely e+M"+M* and e+MCR'+e MCR?. The activation energy for the recombination process is in the range 8.4-33.5 kJ mol-' for low and high CR concentrations respectively. There has been considerable interest in the nature of ion aquation with parti- cular concern for the relative importance of cation-H20 and anion-H20 inter- actions.'' The application of i.r. spectroscopy to this problem has produced inter- esting results.The effect on the H20 stretching-mode absorptions produced by changing the salt oxyanions while holding the cation Na' fixed or alternatively by retaining the same anion ClO.,- and varying the cation has been measured; the results are shown in Table 1. The novel cyclic polyether-ester compounds 2,6-dioxo- 18-crown-6 and 2,4- dioxo-19-crown-6 (4) have been prepared and their binding with Na' K' or Ba2' has been compared with results for 18-crown-6 (1) or valinomycin (2)." The data (Table 2) show significant differences between cation selectivity of (3) or (4)and (1) as indicated by the log K values. R. D. Rauh F. S. Shuker J. M. Marston and S. B. Brummer J. Inorg. Nuclear Chem. 1977,39 1761. ' J. A. Kemlo J. D. Nielson and T.M. Shepherd J. Znorg. Nuclear Chem. 1977 39 1945. ' J. M. Ceraso P. B. Smith J. S. Landers and J. L. Dye J. Phys. Chem. 1977 81 760. J. S. Shih and A. I. Popov Inorg. Nuclear Chem. Letters 1977 13 105. A. Friedenberg and H. Levanon J. Phys. Chem. 1977,81,766. lo G. Ritzhaupt and J. P. Devlin Inorg. Chem. 1977 16 486. R. M. Izatt J. D. Lamb G. E. Maas R. E. Assay J. S. Bradshaw and J. J. Christensen J. Amer. Chem. SOC.,1977 99 2365. The Typical Elements Table 1 Vibrational frequencies (cm-l) for H20 molecules isolated in glassy salt matrices at 80 K Assignment Li[C104] Na[C104] K[C104] Na[C103] Na[N03] v, inner shell 3550 3570 3590 3450 3360 vSyminner shell 3490 3510 3540 -v2 inner shell 1630 1630 1630 1640 1650 Y outer shell 3250 3230 3240 3250 3240 Table 2 log K AH and TAS values of the binding at 25 "C in CH30H of Na+ K' or Ba" with the macrocycles (1)-(4) Compound M"' log K AHlkJ mol-' TAS/kJ mol-' (1) Na+ 4.36 -35.2 -10.0 K+ 6.05 -56.0 -21.8 Ba2+ 7.0 -42.80 2.9 (2) Na+ 0.67 -K+ 4.9 -19.00 9.00 Ba2+ 3.34 -(31 Naf 2.5kO.l -9.50 *0.75 4.6 K+ 2.79 f 0.02 -24.56* 0.04 -8.62 Ba2+ 3.1 k0.2 -1.92 *0.46 15.9 (4) Naf 1.8*0.18 -4.6k0.8 5.9 K+ 2.55 f0.3 -33.10* 0.25 -18.54 Ba2' 1.41k0.11 -20.42 *0.54 -12.6 The structure and energies of small compounds containing lithium or beryllium with ionic multicentre or co-ordinate bonding have been investigated by MO techniques at a uniform level of approximation.l2 The significant conclusions are as follows (a) The single M-M bonds are long and relatively weak and bonds to lone-pair atoms are strong and short owing to n-bonding.In order to maximize n-bonding NH2 adopts a planar configuration in LiNH2 and HBeNH2 whereas LiOH and HBeOH are linear. (b) All systems containing a formal Be=X double bond with the exception of Be2 are bound in singlet and triplet states. (c) Doubly and triply bridged dimers of metal hydrides are generally more favourable than singly bridged forms. (d) Co-ordinate bonds require polarization functions for proper description and may require methods better than STO-3G. 2 Group11 The hydrolysis of Mg2+ has been studied in 1.0mol dm-3 (Na or H) NO3 aqueous solution at 25 "Cby means of e.m.f. measurements of a cell containing suitable test solutions.l3 The results indicate that Mg2+ hydrolyses to form polynuclear complexes [Mg2(0H)2]2' and [Mg2(0H),I2'.Monometallic complexes of Mg2+ and Ca2' with adenosine 5'-triphosphate (ATP) have been studied by 'H n.m.r. and the results show that a bis-M2'-ATP complex is formed.14 '* J. D. Dill P. von R. Schleyer J. S. Brinkley and J. A. Pople J. Amer. Chem. SOC.,1977,99 6159. l3 H. Einaga J.C.S. Dalton 1977 912. l4 J. Granot and D. Fiat J. Amer. Chem. Soc. 1977 99 70. 114 R. H. Cragg J. D. Smith and G. E. Toogood A number of studies relate to a search for a suitable reagent for the spec- trophotometric determination of calcium and magnesium tons. The evaluation of a series of dyes for the spectrophotometric measurement of magnesium ions showed that 2-(2‘-carboxy- 1’-benzeneaz0)- 1,8-dihydroxynaphthalene-3,6-disulphonic acid has constant absorbance in the pH range 7-10 and the capability of distinguishing between Ca” and Mg2+ binding.15 However the free ligand absorbs significantly at the same wavelength as the complex and the formation constant for the calcium complex although smaller than that for magnesium is large enough to cause some interference.A thermogravimetric study of the dimesoperiodates M21209 nH20 (M = Ca n = 8; M =Sr or Ba n = 3) at temperatures of 220-240 “C for 2 h indicated that the iodates M[IO,] are formed as intermediates.16 On increasing the temperature these decompose to the orthoperiodates M5(106) possibly via M21207-type compounds. The synthesis of well characterized complexes with benzo- 15-crown-5 has been reported,” including a series of Ca2’ complexes of this ligand in which the counter- anion is organic such as 2,4-dinitrophenolate 2-hydroxybenzoate 4-hydro- xybenzoate 3,5-dinitrobenzoate or 2,4,6-trinitrophenolate.l8 It was observed that the Ca-crown interaction was sensitive to the nature of the anion and depended upon its basicity and ion-pairing ability.PART 11 Group 111 By G. E. Toogood 1 Boron Boranes Substituted Boranes and Borane Anions.-Theoretical Studies. The pub- lication of Professor Lipscomb’s Nobel Prize lecture’ gives an excellent perspective to this area of study and several significant papers this year demonstrate that he is not ‘resting on his laurels’. The effects of orbital vacancies in boron compounds have been investigated* and an extended set of styx topological rules derived to include structures with up to one vacant orbital per boron atom.3 The extended rules based upon earlier ideas,4 have been used to suggest vacant orbital structures which successfully predict observed distortions from idealized topological struc- tures for the known species B4HI0 B5Hll and B3Hs-.Using PRDDO methods the transient boranes B2H4 B3H7 and B4Hs are calculated to have preferred structures involving one or more vacant orbitals while B3H9 and B4Hl2 have not. Energies of these species all of which are of special interest as likely reaction D. G. McMinn and D. Kratochvil Canad. J. Chem. 1977 55 3909. l6 G. S. Sanyal and K. Nag J.Inorg. Nuclear Chem. 1977,39 1127. 17 N.S. Poonia B. P. Yadov V. W. Bhagiwat V. Naik and H. Manohar Inorg. Nuclear Chem. Letters 1977,13 119. N. S. Poonia V. W. Bhagwat and S. K. Sarad Inorg. Nuclear Chem. Letters 1977 13 119. ’ W. N. Lipscomb Science 1977,196 1047. W. N. Lipscomb Pure & Applied Chem. 1977 49,701. I. M. Pepperberg T. A. Halgren and W. N. Lipscomb Inorg. Chem. 1977,16 363. J. A. Dupont and R. Schaeffer J. Inorg. Nuclear Chem. 1960 15 310. The Typical Elements intermediates in the pyrolysis of diborane have been evaluated using minimal-basis STO-3G and extended-basis 4-31G calculations. The modified styx rules for a structure containing (u) vacant orbitals are s +x =4,s+t =p -v t +y =p -4/2 where the other symbols have their usual meanings and only the middle equation is different from the unmodified rules.Application of the new rules to B2H4,shows three possibilities 0012 (v =2) 1011 (v = l) and 2010 (v =0) but the latter two are topologically forbidden because they involve both 2-and 3-centre bonds connecting adjacent borons. Two conformers are possible for the allowed 0012 (v =2) structure. The DZdform not surprisingly in the absence of r-bonding possibilities is calculated to be more stable than the eclipsed D2h structure. Similarly for B3H9only one admissible topology 3003 (t’ =0) exists and this is calculated to have D3, symmetry. By constrast B4Hs under the extended formal- ism produces at least 14 structures; however B4HI2produces just the 4004 (v =0) structure with Czosymmetry.A simple relationship was found between the ‘Valency,’ V (as defined by other^),^ styx and p. Specifically equation (1) V = 2.55s +2.70t +2.20~+2.00(p +x) (1) Valencies calculated from this equation fitted remarkably well the ‘actual’ PRDDO valencies for numerous boranes. The relationship between ‘valency’ topology and energy was investigated. It is apparent that the borane structures are the result of competing tendencies to maximize ‘valency’ and to minimize the strain energy associated with the bent B-H-B bonds. By incorporating equation (1) into the previously reported relationship between styx and p and the atomization energy A [namely A = 90(p +x)+ 107s+93t +goy in kcal mol-’1 and equating the two units of valency for each of the (p+x)B-H bonds with the bond energy of 90 kcal mol-’ the relationship A =45 V -8s -28t -19y is obtained.The difference in energy between isomeric structures is then equation (2) AE = -45AV +8As +9At (kcal mol-‘) (2) An analogous equation (3) AE = -45AV+ 17As’+ 28At’ (kcal mol-’) (31 enables the calculation of the energies involved in processes of interest in the pyrolysis scheme for diborane for example 2BH3(0002)-+B2H6(2002). From Equation (l) AV is 1.1 also As’= +2 and At’ = 0 (by inspection of the styx numbers)so AE is [from equation (3)] -15.5 kcal mol-’ in excellent agreement with the -12.2 kcal mol-’ value from 4-31G calculations. [It is noteworthy that the first unambiguous proof of the formation of BH3during B2H6pyrolysis has been obtained using molecular beam velocity analysis spectrometry a new technique which enables neutral species to be differentiated before they undergo ionization and frag- mentation in the mass spectrometer i~nizer.~] A theoretical study of the NH3 HzO (CH3),0 and CO Lewis-base adducts of triborane(7) has been performed using PRDDO STO-3G and STO-4-31G methods.’ In all cases the preferred geometry resembles the (1 104) crystal struc- ’D.R. Armstrong P. G. Perkins and J. P. Stewart J.C.S. Dalton. 1973 838. B. Askins and C. Riley Inorg. Chem. 1977 16 481. ’L. D. Brown and W. N. Lipscomb ibid. 1. R. H. Cragg J. D. Smith and G. E. Toogood ture of B3H7.C0 with one BHB bond and C,symmetry. However the preference for this structure decreases with increasing strength of the Lewis-base.This trend ultimately results in a preference for the alternative (2013) structure for B3H8- which may be regarded as a B3H7 adduct of the strong base H-. Structures have been proposed for the hypothetical closo-borane anions of formula B,Hn2- for n = 13 to 24 and various synthetic routes for these compounds suggested.8 The most intriguing is the joining together of the two halves of a metal sandwich compound by the removal of the metal and an extension of this route to include triple-decker sandwich compounds as precursors. The former has already been utilized by Grimes and co-workers (see ref. 44). The structures proposed for the BnHn2- ions resemble the solutions to the well-known problem of arranging objects which repel one another on the surface of a sphere.By coincidence this topic has been revisited for n values 5 to 16.” The new calculations are not always in agreement with earlier work’” nor with the proposed structures for the borane anions.8 However there are factors at work in the boranes and related compounds which are certainly not taken into account in the calculations so the differences between the chemical and mathematical proposals are only to be expected. Prejudice against semi-empirical methods especially those like MIND0/3 and its relatives should be diminished by a recent series of papers by Dewar et al. which clearly indicate the merits of such procedures. Despite the unsatisfactory results of MIND0/3 for boron-containing compounds a newer method MNDO has proved very successful in studying the chemical behaviour of boron hydrides and their anions” (as well as numerous other boron compounds).” The basic principles of the method are given e1~ewhere.l~ It is sufficient to note here that the calculations are parametrized to reproduce experimental data e.g.Afi for a carefully chosen ‘basis set’ of molecules and it is herein that lies its major weakness at present -namely that there is a dearth of accurate thermochemical data for boron compounds particularly bond strengths and heats of formation. Semi-empirical methods parametrized to reproduce the results of ab initio calculations are exem- plified by the PRDDO method of Lipscomb and co-workers and it is interesting that both PRDD0’4 and MNDO” methods have been applied this year to the BnHn2- ions so that their results can be compared.There is general qualitative agreement with respect to the calculated boron charges but the HOMO energies are considerably different. The structural predictions of MNDO are in good agreement with experiment except for B9H9’-. Like B5H9 (where the MNDO predictions also break down) B9Hg2- contains an abnormally large number of three centre bonds whose energies MNDO tends to underestimate. The PRDDO method essentially assumes the structures of the species concerned and hence its structural ‘predictions’ cannot be compared with MNDO. Extended-Hiickel MO calculations have been reported for the icosahedral platinaboranes and carbaboranes [Ft(FI-13)2B1lH1 [Pt(PH3)2CB10H11]- and ’ L.D. Brown and W. N. Lipscomb ibid. 2989. T. W. Melnyk 0.Knop and W. R. Smith Cunud. J. Chem. 1977,55 1745. lo L. Foppl and J. Reine Angew Math. 1912 141 251. M. J. S. Dewar and M. L. McKee (in press). ’’ M. J. S. Dewar and M. L. McKee J. Amer. Chem. SOC.,1977 99 5231. l3 M. J. S. Dewar and W. Thiel ibid. 4899. D. A. Dixon D. A. Kleier T. A. Halgren J. H. Hall and W. N. Lipscomb ibid. 6226. The Typical Elements 117 [Pt(PH3)2C2B9H1 and the failure of polyhedral skeletal electon-counting rules in the case of the platinacarbaboranes attributed to the unequal bonding capabilities of the platinum 5dx and 5d, orbitals in the Pt(PH3) fragment.15"6 The conformations of these compounds are rationalized on the basis of the symmetry characteristics of the LUMOs of the carbaborane and the HOMOS of the Pt(PH3) moiety.Analogous dsmetal compounds are predicted to have complementary conformations and to be more thermodynamically stable. It is salutary to include at this point a warning that electron-counting rules may not always be violated even when they appear to be! The use of X-ray structural data alone to characterize metallocarbaboranes can often result in an inaccurate hydrogen atom count. For example '[Pt(PPh3)2SB8Hs]' has been found on closer investigation using 'H and "B n.m.r. and mass spectrometry to be [Pt(PPh3)2SB8Hlo] whose nidu-structure is thus perfectly understandable. l7 Distortions due to the second-order Jahn-Teller effect have been used to rationalize the effects of additional electron pairs upon the structures of boranes and related compounds.'* Treatment was primarily restricted to systems contain- ing 5 6 or 7 cage atoms.Structural changes such as D3h(trigonal bipyramid) to C4"(square based pyramid) upon addition of an electron pair (e.g. 1,2-C2B3H5to 1,2-C2B3H7) and from C4"to C (e.g. B5H9 to B5HIl) with one more pair are readily accounted for. Interestingly no first order distortion of an original Cho-Oh structure (14 electrons) can produce any of the other six-vertex polyhedra upon addition of electron pairs. Also planar geometries such as DSh(five atoms) D6h (six atoms) and D7h(seven atoms) cannot be obtained by Jahn-Teller distortions of three-dimensional structures. A graph-theoretical interpretation has been given to the bonding topology in polyhedral boranes carbaboranes and metal cluster^.'^ It provides a theoretical basis for the stability of closed deltahedral systems with n vertices and 2n +2 skeletal electrons and can also be applied to open structures.The application to these three-dimensional systems is analogous to that in delocalized two-dimen- sional systems like benzene Transmission of Electronic Effects in Boranes.-"B n.m.r. chemical shifts at cage atoms antipodal to the point of substitction (by halogen SH or ethoxide) in boranes or heteroboranes whose molecular structure is based on the icosahedron have been correlated with the 13C chemical shifts of the para-position in monosub- stituted benzenes. A parallelism between antipodal (A) and mesomeric (M)affects has been suggested.20 The existence of such shifts has been reported previo~sly~'-~~ but their nature has not been elucidated.The A-effect increases generally in the l5 D. M. P. Mingos M. I. Forsyth and A. J. Welch J.C.S. Chem. Comm. 1977,605. l6 D. M. P. Mingos J.C.S. Dalton 1977 602. 17 D. A. Thompson T. K. Hilty and R. W. Rudolph J. Amer. Chem. SOC.,1977,99,6774. l8 C. Gildewell J. Organometallic Chem. 1977 128 13. l9 R. B. King and D. H. Rouvray J. Amer. Chem. SOC. 1977. 99 7834. S. Heimanek J. PleSek V. Gregor and B. Stibr J.C.S. Chem. Comm. 1977 561. 21 S. Heiminek V. Gregor B. Stibr J. PleSek Z. JanouSek and V. A. Antonovich Coll. Czech. Chem. Comm. 1976,41 1492. 22 S. Heimanek J. PleSek and B. Stibr 3rd Internat.Conference Boron Chem. 1976 Ettal Munich Germany Abstract 52. 23 R. Weiss and R. N. Grimes J. Organometallic Chem. 1976 113 29. 118 R. H. Cragg J. D. Smith and G. E. Toogood order I <Br <C1< SH <F <OR in accord with increasing electron donation rather than with increasing electronegativity. Clearly the M-and A-effects are not identical because the aromaticity of benzene-like systems although analogous with," is not the same as the 'pseudoaromaticity' of boranes which is produced by the delocalization of electrons in an electron-deficient skeleton. Nevertheless the transfer of various effects through a system possessing a high degree of electron delocalization is qualitativeIy similar in the two systems. The transmission of electronic-eff ects via 'apical-apical conjugation' (an equivalent term to 'antipodal') has been suggested for BloHIo2- derivative^^^ and recently to explain certain features of the structure of the dinitrogen-bridged complex (H)2(Ph3P)3Ru(N2- BIoH8SMe2),3C6H6.The electrons provided by the SMe2 group antipodal to the N2 help the Blo cage to donate electrons to the N2 thus maintaining the integrity of the N-N bond which is otherwise likely to be weakened by aN2+Ru and Ru +(77*)N2synergic bonding. A high GNN and short N-N bond are accounted for by this effect. N.M.R. Spectroscopy* and Cage Boron Compounds.-The use of n.m.r. spec- troscopy in the area of borane and carbaborane chemistry continues to expand and several highly significant papers have appeared this year.Proton-proton spin coupling in metallocarbaboranes has been investigated using a triple resonance technique in which IIB and 'H decouplings are conducted simultaneously.26 The technique is frequently capable of removing uncertainties in the interpretation of conventional n.m.r. spectra and allowing unequivocal assignments of resonances. Specifically three-bond proton-proton spin coupling in 2,4-C2B5H7 and a number of metallocarbaboranes was investigated and coupling constants of the order of 10-20Hz for H-C-C-H 1-10Hz for H-C-B-H and 0-4Hz for H-B-B-H were found. Four-bond coupling was rarely observed except for protons in trans locations. Most of the species considered were 7-vertex pentagonal bipyramidal structures and the coupling between equatorial protons was stronger than between equatorial and apical protons.H-B and H-C resonances are found to be inherently sharp with natural line widths -1-2 Hz and the broad peaks observed in undecoupled spectra as previously ~uggested,~~ are due primarily to unresolved coupling which can be removed with "B and 'H decoupling. It is clear that although "B quadrupole relaxation contributes broad- ness to "B n.m.r. spectral lines it is not a significant factor in 'H n.m.r. line broadening.28 Hence when appropriate decoupling is performed the 'H spectra of many boron compounds can be analysed in detail in much the same manner as those of organic compounds. As an illustrative example the "B-decoupled proton resonances of 2,4-CzB5H7are shown in the upper line of Figure 1.In the remaining 24 W. H. Knoth J. Amer. Chem. SOC.,1972 94 104. 2s K. D. Schramm and J. A. Ibers Inorg. Chem. 1977,16 3287. 26 V. R. Miller and R. N. Grimes Inorg. Chem. 1977,16 15. 27 T. Onak and E. Wan J.C.S. Dalton 1974 665. R. Weiss and R. N. Grimes J. Amer. Chem. Soc. 1977.99 1036. * It has been decided informally among workers in the subject that in future publications (1977 onwards) "B downfield shifts will be given a positive sign (the reverse of the previous convention) to bring consistency to the signs among the commonly employed nuclei such as 'H and 13C.BF3,0Et2will remain as the zero standard. See for example J. Organometallic Chem. 1977 131. C43. The Typical Elements spectra the boron nuclei and selected protons (as listed at the left of the figure) are decoupled simultaneously.It is clear from the H-C (2’4) proton decoupling that both H3and H5,6are coupled to H-C whereas HI,,decoupling has no effect on the HZe4or H5,6resonances although it does remove the secondary ‘triplet-of-triplets’ H24 DKOUPLED I-$ DECOUPLEO HwDKOUQLED A OECOUPLEO A Figure 1 120 R. H. Cragg J. D. Smith and G. E. Toogood fine structure in H3 thus demonstrating slight coupling between the apical protons and H3 and none with H2,4.5,6. The technique can also be used to distinguish between positional isomers. For example although both 2,4- and 2,3-C2B5H7 have C2"symmetry and cannot be differentiated on the basis of "B or 'H n.m.r. peak areas and chemical shifts alone their triple resonance spectra will clearly be different.In particular a characteristically large H-C-C-H interaction will occur in the 2,3-isomer. The correlation of "B and 'H spin-lattice relaxation times with molecular structure has been studied for a number of carbaboranes and metallocar-baboranes.28 From the "B work TI values for the carbaboranes are found to be longer than for the analogous mono-metallocarbaboranes which in turn have longer Tl values than the corresponding dimetallocarbaboranes. For example in 2,4-C2B5H7,T (in ms) is 38.6 for B(e) (using the same numbering scheme as in Figure 1) whereas in [~-CO-~~~-C~H~)-~,~-C~B~H~] for the equivalent boron it is 9.7 and in [~,~-(CO-~~-C~H~)~-~,~-C~B~H~] it is 2.86. Also as reported earlier for boranes and some carbaborane~,~~~~~ T1values of apical borons are generally longer than those of non-apical boron nuclei.Thus B1, in 2,4-C2B5H7 have T1= 68.3 whereas B5,6 have T1= 47.7 and B3 has T = 38.6 ms. This is probably due to the more symmetric environment of the apex boron which in turn produces a smaller field gradient and a longer T1.29The symmetry-field gradient- T relationship has been expanded28 to include comparisons between non-apical borons in the same molecule. Estimates of the degree of symmetry involve the consideration of the "B nucleus in question as the origin of a co-ordinate system from which vectors are drawn to all bonded neighbouring nuclei. The directions of the vectors are obtained from the appropriate crystal structures and the magnitudes taken as the atomic numbers of the bonded atoms.These vectors are then added and the mag- nitude of the resultant taken as a measure of the field gradient of the "B nucleus in question. For 2,4-C2B5H7 the resultants for B5,6 are 12.57 and for B3 13.13 -consistent with the trend of the Tlvalues reported above. It should be stressed that these symmetry arguments break down when comparing two atoms of different co-ordination number or two atoms in different molecules. Another interesting aspect of the proton relaxation study is the fact that Tl values for protons bound to cage carbon atoms were found to be strongly affected by decoupling of "B nuclei. In each instance a substantial lengthening of T was noted upon decoupling.The change in T for a C-H proton ATl induced by saturation of a given "B nucleus should be indicative of JBHand in a general sense of distance. In the case of 2,4-C2B5H7 saturation of each equatorial "B resonance produced comparable AT,'s in accordance with the fact that all equatorial BH groups are adjacent to CH. However decoupling the apical borons which are farther away produced a much smaller ATl (5.3 s compared to about 12 s). The application of double resonance Tl (and ATl)measurements in structure determination is a promising addition to the n.m.r. techniques currently employed but its full implications for the development of n.m.r. theory of polyhedral cluster molecules remain to be deter- mined. 29 A. Allerhand J. D. Odom and R.E. Moll J. Chem. Phys, 1969 50 5037. 30 R. R. Rietz and R. Schaeffer J. Amer. Chem. SOC.,1973,95,4580. The Typical Elements 121 The calculation of charges on boron and carbon atoms in cage boranes carbaboranes and related compounds was discussed in Annual Reports 1975 72A 96. A recent paper3’ has provided a potential experimental approach to the determination of the relative charges within a given cage by measuring the charges of the hydrogens attached to the cage atoms. Since the hydrogens are a mirror of the attached borons or carbons the trends in the charges of the latter will be reflected in the trends of the hydrogen charges. The approach is based upon the aromatic solvent induced ‘H n.m.r. shift (ASIS) phenomenon. ASIS effects have often been observed in organic chemistry32 and occasionally for b~ranes.~”’~ However their quantitative correlation with hydrogen charges has not previously been undertaken.The proton shifts for boranes and carbaboranes dissolved in benzene or hexafluorobenzene (compared to data obtained in a non-aromatic solvent e.g. cyclohexane) follow the order H-C >H,-B >H >B with the shifts being upfield for benzene and downfield for C6F6. The correlation of the shifts AT with the hydrogen charges Q which were evaluated by ab initio minimum basis set MO calculations was found to be good -so much so that the reverse procedure of evaluatingQ (ASIS) from Ar(AS1S) seems a viable method especially when nearest neighbour cage hydrogen effects are considered.31 Metalloboranes Carbaboranes and Related Compounds.-Metalloboranes.The reactions of B5Hs- with FeC12 and NaC5H5 produce predominently [2-(Fe-q ‘-C5H5)B5Hlo]35 in sharp contrast to the analogous reaction involving CoC1 which generates a host of three- and four-boron metalloboranes but no detectable five- boron The iron compound undergoes thermal rearrangement in which the metal migrates from a basal to an apical location to produce the ferrocene-like isomer containing a cyclic B5 ligand (Figure 2). In the cobalt system the main primary product is the (2n +4)electron [2-(Co-q5-C5H5)B4H8] which upon pyrolysis goes chiefly to the 1-isomer with the cobalt in an apical position an analogous compound to that with iron except that there is a B4 cyclic ligand (Figure 3).Other products of the pyrolysis include the (2n + 2) electron systems [1,2-(Co-and [1,~,~-(CO-~‘-C~H~)~B~H~], q5-C5H5)2B4H6] to both of which were assigned octahedral geometries subsequently confirmed by X-ray studies for the latter and the (2n) electron systems [(CO-~’-C~H~)~B~H~] and [(Co-q’- C5H5),B4H4]. The former has a capped octahedral structure37 unique among boron cages but analogous to [Os,(CO),,] and [Rh7(CO):6] which are also electron- hyperdeficient. In contrast the tetra metal compound appears to have the DZd dodecahedra1 structure expected for the corresponding 2n +2 system e.g. [(C0-q5- C5H5)2C2B4H6], rather than the capped closo-structure predicted for the electron- hyperdeficient structure. Trans-cage metal-metal bonds may be present and make 31 T.Onak W. Inman H. Rosendo E. W. Distefano and J. Nurse ibid.,1977 99 6488. 32 See for example R. S. Armstrong M. J. Aroney R. K. Duffin H. J. Stootman and R. J. W. Lefevre J.C.S. Perkin 1973 1272 and 1362. 33 P. M. Tucker T. Onak and J. B. Leach Znorg. Chem. 1970.9 1430. 34 G. M. Bodner and L. G. Sneddon ibid. 1421. jS R. Weiss and R. N. Grimes J. Amer. Chem. SOC.,1977 99 8087. 36 V. R. Miller R. Weiss and R. N. Grimes ibid.,5646. 37 J. R. Pipal and R. N. Grimes Znorg. Chem. 1977,16 3255. R.H. Cragg J. D. Smith and G.E. Toogood up for the deficiencies of two skeletal electrons and stabilize the system. Such a mechanism would be less likely in the Co3system which is perhaps why that adopts the capped closo-configuration.OH Figure 2 Figure 3 The Typical Elements An interesting practical consequence of the replacement of borons by cobalts in these clusters is the greatly increased hydrolytic stability of at least some of the cobaltaboranes compared to their borane analogues. Especially noteworthy is the difference in behaviour of [(Co-q5-C5H5)B4H8] compared to B5H9. As is well known the latter is highly pyrophoric whereas the metalloborane is resistant to attack by air and water. [Recently38 the hypo-B5H9,L (L is a diphosphine) adducts have also been shown to be stable to hydrolysis]. Cobaltaboranes which do contain five borons have been made from metal atom attack on ~entaborane(9).~’ [(Co-q5-C5H5)3B5H5] has been made by this route and it will be interesting to see whether this (2n)system adopts a capped closo-structure analogous to the ‘C03B4’ species as suggested above.A cobaltaborane analogue of B10H14 has been characterized. The compound [5-(Co-q’-(C5H5)B9H13] has a structure derived from a Bl0HI4- like cage in which the 5-BH unit has been replaced by [Co(q5-C5H5)] (Figure 4). Figure 4 3” N. W. Alcock H. M. Colquhoun G. Haran J. F. Sawyer and M. G. H. Wallbridge J.C.S. Chem. Comm. 1977,368. 39 L. W. Hall G. J. Zimmerman and L. G. Sneddon ibid. 45. 4” J. R. Pipal and R. N. Grimes Inorg. Chem. 1977 16,3251. 124 R.H. Cragg J. D. Smith and G. E. Toogood Alternatively it may be viewed as a pseudo-ferrocene type complex with the B4 face of the borane ligand nearly parallel to the q5-CsHs reminiscent of [2-(Co-q5-C5H5)B4H8] referred to previously.Tetracarbon Metallocarbaboranes and Carbaboranes. It is clear from recent k~~*~~ ~ ~ rthat the chemistry of tetracarbon metallocarbaboranes briefly mentioned in last years Report is not merely a routine extension of the well-studied dicarbon carbaborane area. An extensive series of 12- 13- and 14-vertex systems contain- ing iron cobalt or nickel has been characterized by X-ray diffraction and n.m.r. spectroscopic methods. These compounds were made from Me&&& by reduc- tion to the dianion followed by metal ion insertion. The iron system is particularly interesting in that four isomers of the 14-vertex [(Fe-r) 5-C5H5)2Me4C4B8H8] and one of the 12-vertex [(Fe-r) 5-C5H5)Me4C4B,H8] were identified among the primary reaction products.Subsequently the thermal rearrangement of the 14-vertex system was investigated and several new isomers ~haracterized.~~ As Figure 5 shows although the structure of the final isomer VIII is consistent with trends previously observed in metallocarbaborane chemistry -notably the preference of metal atoms for high-co-ordinate vertices and the tendency of carbon atoms to migrate ultimately to non-vicinal positions -the stereochemistry of these polyhedra is novel in several respects. Most striking is the existence of three kinds of polyhedra within the five isomers structurally characterized -a situation almost unprecedented in carbaborane and metallocarbaborane chemistry. Normally iso- mers adopt the same basic framework -the only previous exception being the different (but closely related) gross structures of nido-and closo-[Pt(R3P),(Me2C,B6H6)].It is believed that kinetic effects related to the extreme size of the polyhedra and the mode of insertion of the metal ions into the [Me4C4B8H82-] dianion are responsible for the initial production of geometries other than the thermodynamically preferred closo -polyhedron (species VIII in Figure 5). In I 11 and V two framework atoms (1 boron 1 carbon) are bound to both metals whereas VII and VIII have no such feature. In order to achieve the closo-isomer structures at least two Fe-B/C links in V must be broken -evidently something which only occurs at an appreciable rate at elevated temperatures.The persistence of the nido-geometry below 300 "C suggest that steric effects of this type are likely to be encountered in other 14-vertex (or larger) systems. These 2n +2 systems which are effectively frozen into nido-configurations for steric reasons constitute a new class of polyhedra which have been dubbed 'pseudo- nido' to distinguish them from the true (2n+4) nido-species. It is conceivable that these pseudo-nido-systems will provide routes to the synthesis of larger (1 5-vertex) polyhedra by insertion of a two-electron donor metal group into their open faces. One other aspect of the structures of these di-iron species is noteworthy. It is apparent that the tendency of carbons to occupy low co-ordinate vertices is less important than their propensity to occupy non-adjacent positions.This is especi- ally obvious in the rearrangement-product V which has two less low-co-ordinate carbons than the isomers I and 11. However even the separation of the carbon atoms is secondary to the achievement of a closo-structure. 41 W. M. Maxwell R. F. Bryan and R. N. Grimes J. Amer. Chem. Soc. 1977,99,4008. 42 W.M. Maxwell R. Weiss E. Sinn and R. N. Grimes ibid.,4016. The Typical Elements b700 1140' ? 1 150' TI P 300" Figure 5 R. H. Cragg J. D. Smith and G. E. Toogood [(Fe-7 5-C5H5)Me4C4B7H8] and its cage-isoelectronic relatives [l-(Co-$-C5H5)-12-(CZH50)-2,3,7,8-(MeC)4B,H6] (Figure 6) and Me4C4B8H4 have been structurally ~haracterized.~~’~~’~~ Since all are 2n + 4 systems nido-structures were CH cc i13 0BH B Figure 6 expected.However the structures are not identical and although the former is clearly nido with an obvious six-membered open face (with a B-H-B bridge on the edge) the latter two are best viewed as distorted icosahedra as shown [(6) and (7)]. In the carbaborane (Figure 7) two carbon-carbon interactions C2-C7 C3-C8 have been stretched to the point of breakage to accommodate the extra two electrons in the system. Such distortions have increased in the cobalt compound to the point where three carbon-carbon interactions are non-bonding (uiz. C2-C7 C3-C7 C3-C8) (Figure 6). The absence of bridging hydrogens or other features which might interfere with polyhedral closure is considered to favour the ‘nearly’ isosahedral structure of Me4C4BsH4 while the presence of such features is respon- W.M. Maxwell K.4. Wong and R. N. Grimes Inorg. Chem. 1977 16 3094. 44 D. P. Freyberg R. Weiss E. Sinn and R. N. Grimes ibid. 1847. The Typical Elements Figure 7 sible for the obvious nido-form of the iron complex. The cobalt system is an intermediate situation. The cobalt tetracarbaborane was prepared from the reaction of bis-(dicarborane) complex [Fe(H)2{2,3-(MeC)2B4H4}2] with the precursors of (Co-7 5-C5H5).43 This iron complex is a remarkable species even by the standards of metallocarbaborane chemistry being able to undergo at least five different types of reaction namely (i) oxidative metal elimination and fusion of the C2B4 ligand groups to form e.g.Me4C4B8H4; (ii) reversible metal deprotonation by NaH; (iii) substitution of one carbaborane ligand by CO groups; (iv) oxidation to iron(II1) metallocarbaboranes; and (v) insertion of transition- and main-group metals into the framework. The latter reaction was reported last year for the preparation of the electron-hyperdeficient [(Co-77 5-C5H5)FeMe4C4B8H8] and has been used recently43 to pro- duce germanium and tin compounds [MFeMe4C4B,H8] with similar but not iden- tical structures (Figure 8). Presumably the difference in the positions of the ‘wedging’ atoms is due to steric effects -the bare tin being able to occupy a position which is unavailable to the 75-C5H5-capped cobalt. Tetracarbon metallocarbaboranes containing molybdenum or tungsten have also been prepared.[M(C0)2Me4C4B8H8] 13-vertex systems were made by direct reaction of [M(CO),] with the neutral Me4C4B4H8.41 These are expected to be closo-systems and are isoelectronic with the previously reported [MO(CO)~(C~B 10H12)]2-diani~n.~~ 45 D. F. Dustin G. B. Dunks and M. F.Hawthorne J. Amer. Chem. SOC.,1973.95 1109. R. H. Cragg J. D. Smith,and G. E. Toogood b Figure 8 Me4C4B4H4 has recently been synthesized by a novel and potentially very inter- esting route namely the photochemical addition of 2-butyne to [Fe(C0)3B4H8].46 This is an analogous reaction to those involving [Fe(C0)3C4H4]47 and constitutes a rational route for the insertion of carbon atoms into the borane cage. Metullodicarbuborunes.The formal electronic configuration of the metal has been shown to have a pronounced effect on the degree of cage opening in a set of isoelectronic metallodicarbaboranes. Thus in the series of 18-electron compounds [Re(C0)3C2B9H1 (Re' d6) [Au(Et2NCS2)C2B9Hl 1] 1] (Au" d') [Hg(PPh3)C2B9H11] (Hg" d lo) and [(T1C2B9Hll)-] (Tl* dlos2),there is a pro-gressive cage opening culminating in the thallium complex where the bonding between metal and cage is considered to be an ion-pair intera~tion.~' This is consistent with the ready loss of thallium from the [T1(CH3)C2B9H11] cluster in the mass spe~trometer.~~ Last year's Report dealt with the catalytic potential of hydrido-transition metal dicarbaboranes. An interesting extension of that work is the preparation of a polymer-bound metallocarbaborane catalyst [3,3-(Ph3P),-3-H-4-(poly-styrylmethy1)-3,l ,2-RhC2B9Hlo] an analogue of the homogeneous hydrogenation catalyst [3,3-(Ph3P)2-3-H-3,1,2-RhC2B9Hll].The new species was found to be an efficient catalyst for the reduction of alkenes to alkanes." The homogeneous catalyst mentioned above showed little or no stereoselectivity in the process 46 T. P. Fehlner ibid.,1977 99 8355. 47 J. S. Ward and R. Pettit ibid.,1971 93 262. 48 H. M. Colquhoun T. J. Greenough and M. G. H. Wallbridge J.C.S. Chem. Comm. 1977 775. 49 J. Smith G. Allender and H. D. Smith jun. Inorg. Chem. 1977 16 1814. 50 B. A. Sosinsky W. C. Kalb R. A. Grey V. A. Uski and M. F. Hawthorne J. Amer. Chem. Soc. 1977 99 6768. The Typical Elements 129 whereby terminal hydrogens were replaced by deuterium in a variety of boranes and related compounds in marked contrast to [RU(H)(C~)(PP~,)~].” The first small metallodicarbaboranes containing tin and lead have been prepared recently.52 [SnC2B4H6] [SnMe2C2B4H4] and the corresponding lead compounds were produced by reaction of MX (M = Sn or Pb X = C1 or Br) with [C2B4H,]-or [C,C’-Me2C2B4HS]-.All were characterized as closo -pentagonal bipyramidal systems with the metal atom occupying an apical position. Interes-tingly no metallo products were observed when [C2B5H2]*- was employed. Reac- tion of [SnMe2C2B4H4] with [Co(q5-C5H5)(C0),] gave [ 1-(Co-qs-C5Hs)-2,3-(MeC)2B4H4] and [~-(CO-~~-C~H~)-~-S~-~,~-(M~C)~B~H~] -a cage system containing tin and cobalt in adjacent positions of the same polyhedron (Figure 9).In all cases the group IV metallic elements were ‘bare’ outside the cage with their unshared electron pairs presumably directed away from the polyhedral surface. H3C OR CH Figure 9 Boranes Carbaboranes and Related Species containing Sulphur Selenium or Tellurium. Reaction of atomic (‘D)sulphur with carbaboranes has yielded a variety of B-mercapto~arbaboranes.~~ In the case of 2,4-C2BSH7 all three isomers namely 5-SH- 3-SH- and 1-SH-2,4-C2BsH6 were produced. Photolytic decomposition of these mercapto species resulted in the formation of several isomers containing E. L. Hoel M. Talebinasab-Savari and M. F. Hawthorne ibid.,4356. 52 K.-S. Wong and R. N. Grimes Inorg.Chem. 1977 16,2053. 53 J. S. Plotkin and L. G. Sneddon J. Amer. Chem. Soc. 1977 99 3011 R. H. Cragg J. D. Smith and G. E. Toogood x Figure 10 boron-bonded bridging disulphide groups [5,5’-S2-(2,4-C2B~H6)~], [3,3’-S2-(2,4-C2B5H,),] and [ 1 1’-Sz-(2,4-CzB5H6)2] (Figure lo) and perhaps others. These disulphides are especially interesting since they appear to be both hydrolytically and thermally stable. Furthermore they have ‘open’ carbon positions for substitu- tion or polymerization reactions. A bridging disulphide group has also been found in the [B12HllSSB12H11]4- structure. This species was made by an oxidative coup- ling reaction of [Bl2HllSHI2- with iodosoben~oate.~~ Upon addition of the disulphide to acidified solvents there is rapid formation of an exceptionally stable free radical -presumably [Bl2HllSI2- or [BI2HlISH]T with a th of about 8 days in solution.Oxidation of the disulphide produces [B12HI ,SOSBl2H1 1]4- which is surprisingly reluctant to undergo further oxidation. In contrast [ 10,lO’-{B10H8(SMe2)2}2] produced by electrochemical oxidation of [Bl,H9(SMe2)]- like [Bz0Hl8l2- consists of two Blo species linked by a pair of adjacent axial and equatorial borons from each cage forming two localized three-centre B -B-B bonds. The dimethyl sulphide groups are bonded to the other axial borons and are not involved in the cage-to-cage link.55 54 G. R. Wellum E. I. Tolpin A. H. Soloway and A. Kaczmarczyk Inorg. Chem. 1977 16,2120. ’’ 0.P. Anderson and A. P. Schmitt ibid.,1630.The Typical Elements 131 The first selena- and tellura-boranes have been prepared by reaction of Na2Se4 and Na,Te in aqueous ammonia with de~aborane(l4).~~ The initial products are the [BloHI1E-] species (E = Se or Te) which upon acidification yield [BloH12E]. Treatment of these neutral compounds with KOH and CoCl or FeC1 gives the expected [M1r(BloHloE)2]2-. [(M-~'-C5H5)BloHloE] compounds can also be made. A by-product of the polyselenide/decaborane reaction was the novel electron-rich [B9H9Se2] which is postulated to have an 11-atom nido-cage with adjacent Se atoms in the open face. The crystal structure of 1-NN-dimethylthiocarbamoyl-1,2-dicarba-closo-dede-caborane(l2) suggests that the energetically favoured orientation of the thio- carbamoyl group with respect to the carbaborane cage results from a directional 7-bond involving the carbon atoms of the cage and the CS group all of which are coplanar.'' Metal Tetrahydroborates.The chemistry of covalently bonded tetrahydroborates of the transition elements including the lanthanoids and actinoids has been compre- hensively reviewed up to the middle of 1976.5sInterest in the structures of these compounds continues to focus on the various modes of co-ordination of the [BH4]- groups despite (or perhaps because of) the well-documented problems associated with the X-ray diffraction procedure and with the n.m.r. studies for these systems. The variation in the nature of the bridging involved in the attachment of [BH,]-to a central metal atom in an anionic species compared to that in the neutral tetrahydroborate has been investigated using Raman spectroscopy.Specifically the anions [AI(BH,),]- [Be(BH4),]- and [U(BH,),]'- have been compared to their neutral parent^.'^ The spectrum of [Be(BH4)3)- indicates three bidentate ligands and a monomeric structure analogous to the isoelectronic [Al(BH,),] but in contrast to the helical polymeric structure of solid [Be(BH4),] which involves bridging and terminal ligands. [Al(BH,),Ip apparently involves three bidentate and one tridentate ligand giving the aluminium a formal co-ordination number of 9. The relationship between the neutral and anionic U4+species parallels that of beryllium in that there is a change from a polymeric neutral compound to a monomeric anion in which the co-ordination about the metal is only slightly modified.In both cases four bidentate and two tridentate ligands provide a co- ordination number of 14 for uranium. The much-debated structure of gaseous [Be(BH,),] has been investigated by "B and 'H n.m.r. and the results show conclusively that none of the previously postulated triangular B-Be-B structures is correct. Instead a linear configuration is proposed.60.h1 Noteworthy in view of previous contentions to the contrary,62 is the fact that the n.m.r. study shows only one species is present in gaseous [Be(BH4),]. A single M-H bridged tetrahydroborate of copper Cu(MePh2P),(BH4) has been independently reported by two groups. The X-ray 56 J. Little G. D. Friesen and L.J. Todd ibid.,869. 57 H. Beall Inorg. Nuclear Chem. Letters 1977 13,111. " T. J. Marks and J. R. Kolb Chem. Rev. 1977,77 263. 59 A. C. Bond and F. L. Himpsl jun.. J. Amer. Chem. Soc. 1977 99 6906. 6o D. F. Gaines J. L. Walsh and D. F. Hillenbrand J.C.S. Chem. Comm. 1977 224. D. F. Gaines J. L. Walsh J. H. Morris and D. F. Hillenbrand (in press). 62 A. Alrnenningen G. Gundersen and D. P. Novak. Acta. Chem. Scand. (in press quoted in ref. 58). 132 R. H. Cragg J. D. Smith and G.E. Toogood structure63 confirms the previously inferred monodentate character of the bonding.64 The fluxional nature of solid [M(BH4),] (M = Zr or Hf) has been studied by ‘H n.m.r. spectr~scopy.~~ Two motional processes have been observed. One is ascribed to a rapid bridge-terminal hydrogen permutation and showed an activa- tion energy E, of about 22 kJ mol-’ for Zr and about 35 kJ mol-’ for Hf.The second is considered to involve rotation about the M-B-H axis. For this process E,(Zr) = 22.6 and E,(Hf) = 19.2 kJ mol-’. Bidentate [BH4]- has been suggested as acting as a 4-electron donor (analogous to the q3-allyl ligand)66 and this has been nicely confirmed in an elegant study of [Mo(CO),BH,]-ion the first reported example of a stable transition metal [BH4]- complex in which the formal oxidation state of the metal is zero and in which the only other ligand type is C0.67 The similarity between [BH4]- and q3-allyl is strikingly revealed when the appropriate dimensions are compared for the ‘C,MO(CO)~’ and ‘BH,Mo(CO),’ moieties in [Mo(CO)~{~ and 3-C3H4P(C6Hs)3)] [Mo(CO),(BH,)]-.The ligand bite angles are almost identical (60.5”,59.5”) as are the analogous Mo-C lengths. Mention should also be made of the first triple ‘metal’ hydride AlH4MgBH4 produced from Na[AlH4] and MgC1(BH4).68 The kinetics of the hydrolysis of [BH,]- in acetonitrile have been studied. Solutions up to 0.6M-H20 do not hydrolyse [BH,]-measurably over a period of days. However 10-3M-CF3S03H in CH3CN completely hydrolyses [BH4]- in less than 10 s and buffered or unbuffered solutions of acetic acid give measurable rates with tj typically around 30s. Support is provided for the postulate that BHS is an intermediate in aqueous solution hydroly~is.~~ Octahydrotriborates. Electrolysis of MeCN or CH2C12 solutions of [B3H8]- in the presence of PPh3 has proved a useful route to metalloboranes the nature of which depends upon the anode material.70 Copper or silver dissolved to produce [M(B3H8)(PPh3)2]; the silver compound is new.Zinc or cadmium anodes also resulted in metal dissolution but the only identifiable products were Ph3PB3H7 and its cleavage products; these oxidation products were also obtained when platinum electrodes were employed. However anode dissolution was not observed. Oxida- tion of [B,H,]- by HgC12 in the presence of phosphine ligands also produced LB3H7 and its cleavage and chlorinated products.” The static structure of [Be(B3H&] has been confirmed by a low temperature X-ray diffraction It consists of two bidentate [B3H8]- ligands bound to a central Be by two Be-H-B bridge hydrogen bonds from adjacent borons in each ligand.[B3H8]- derivatives of beryllium have been found to be 63 J. L. Atwood R. D. Rogers C. Kutal and P. -4.Grfitsch J.C.C.Chem. Comm. 1977 593. 64 J. C. Bommer and K. W. Morse ibid. 137. 65 I.-S. Chuang T. J. Marks W. J. Kennelly and J. R. Kolb J. Amer. Chem. SOC.,1977,99 7539. 66 B. D. James and M.G. H. Wallbridge Prog. Inorg. Chem. 1970 11 99. 67 S. W. Kirtley M. A. Andrews R. Bau G. W. Grynkewich T. J. Marks D. L. Tipton and B. R. Whittlesey J. Amer. Chem. SOC.,1977 99 7154. 68 E. C. Ashby and A. B. Goel Inorg. Chem. 1977,16,2082. 69 R. F. Modler and M. M. Kreevoy J. Amer. Chem. SOC.,1977 99 2271. 70 B. G. Cooksey J. D. Gorham J. H. Morris and L.Kane J.C.S. Dalton 1978 141. 71 A. Drummond and J. H. Morris Inorg. Chim. Acta 1977 24 191. 72 D. F. Gaines and J. H. Morris J.C.S. Chem. Comm. 1975 626. 73 J. C. Calabrese D. F. Gaines S. J. Hildebrandt and J. H. Morris J. Amer. Chem. SOC., 1976,98,5489. The Typical Elements 133 fluxional based upon a variable temperature "B and 'H n.m.r. study of [Be(B3Hs)2] [Be(q'-C5H5)(B3Hs)] and [Be(Me)(B3Hs)2].74 The latter like [AlMe2(B3Hs)] shows non-equivalent methyl groups at low temperatures due to the orientation of the [B3HS]- ligand. Reactions of [Be(B3HS)2] and [Be(BH4)2] have also been studied. Of particular interest is the novel reaction of [Be(BH4)2] with 1-C1B5Hs to produce B5HloBeBH4.74 The crystal structure of [B3Hs]- is known to involve the 2013 (styx) bonding pattern while in solution the hydrogens and borons are equivalent on the n.m.r.time scale suggesting a difference of 35 kJ mol-' or less between the 1104 and 2013 structures. Previous calculations75 had indicated the 1104 structure to be the more stable but recent results from SCF-CI and STO-4-3 1G calculations show the 2013 is very slightly (-5 kJ mol-') more Metal Borides.-The long standing problem of the nature of the y-A1Bl2 structure has now been solved.77 It is based upon a framework of boron icosahedra intimately related to the frameworks recently found in cu-AlB1278 and and all three are derivable from that of @-rhombohedra1 boron." The basic layer structure in these compounds consists of interconnected 48-boron units designated as B4'-(Td) formed from four icosahedra whose centres lie at the corners of a tetrahedron.Two distinct methods of stacking these 'layers' will preserve the preferred pentagonal pyramidal bonding of boron noted previously in elemental boron and many of its compounds.'' One arrangement produces the three-layer framework structure (Fd3m) found in a -A1BI2 and p-rhombohedra1 boron while the other produces the y-AlB, structure (P63/mmc). Further insight into the relationship among these structures is achieved by focusing attention upon yet another structural unit in which twelve interbonded icosahedra define the vertices of a regular truncated tetrahedron. Both framework types can be generated by appropriately sharing hexagonal faces of this B144-(Td) unit.Each of the large holes thus defined at the centre of the 'tetrahedron' has twelve internally directed unsatisfied bonds that sharply limit the size geometry and symmetry of the structural units which can be contained therein. In y-A1B12 the units found in these holes are B20species formed by condensing two icosahedra in either of two ways one of which has C2 symmetry and the other C (Figure 11). The two forms alternate throughout the structure. In a-A1BI2 only the Cz symmetry species is found. Finally in p-rhombohedra1 boron the unit in the hole is formed by conden- sation of three icosahedra to give a B28 species of C3"symmetry (Figure 12). These smaller units fully satisfy the internal bonding within the B144-(Td) clusters and interlink in a subsidiary framework through the open hexagonal 74 D.F. Gaines J. L. Walsh J. H. Morris and D. F. Hillenbrand (in press). 7s P. E. Stevensen J. Amer. Chem. SOC.,1973,95 54. 76 Cited as 'work in progress' in ref. 7. 77 R. E. Hughes M. E. Leonowicz J. T. Lemley and L.-T. Tai J. Amer. Chem. SOC.,1977,99,5507. 78 I. Higashi T. Sakurai and T. Atoda J. Solid State Chem. 1977 20 67. 79 Quoted in ref. 77. R. E. Hughes C. H. L. Kennard D. B. Sullenger H. A. Weakliem D. E. Sands and J. L. Hoard J. Amer. Chem. SOC.,1963,85 361. J. L. Hoard and R. E. Hughes in 'The Chemistry of Boron and its Compounds' ed. E. L. Muetterties Wiley New York 1967. R. H. Cragg J. D. Smith,and G.E. Toogood 0 b Figure 11 Diboron Halides.-Although the structures of crystalline B2F4s2and B2C14,s3 which involve eclipsed BX2 groups (molecular symmetry D2,,),and gaseous B2C1484 with staggered BC12 groups (molecular symmetry D2d),have been reported previously results from both spectroscopic experiments and theoretical calculations have hitherto faiied to agree on the molecular structure of gaseous B2F4.However an electron diffraction investigations5 has now revealed that the molecule is planar (D2h), in contrast to gaseous B2C14. Considerable double bond character is implied for the B-F link since the boron-fluorine distance (131.4 pm) is 5-6 pm shorter than the sum of the covalent single bond radii. Resonance structures such as I and F F+ +F F+ \-/ \-B-B -/ +F/B-B\F F/ \F I I1 B2 L. Trefonas and W.N. Lipscomb J. Chem. Phys. 1958,28 54. 83 M. Atoji P. J. Wheatley and W. N. Liuscomb ibid. 1957 27 196. 84 R. R. Ryan and K. Hedberg ibid. 1969 50 4986. 85 D. D. Danielson J. V. Patton and K. Hedberg J. Amer. Chem. Soc. 1977 99 6484. The Typical Elements Figure 12 I1 clearly contribute to the overall structure and the implied charge distribution on the boron atoms is also in accord with the long B-B bond (171.9pm). [It is noteworthy that the isoelectronic N204 molecule is also planar and nas a long N-N and short N-0 The difference between the structures of gaseous B2F4 and B2C14 is seen as a result of delicate balancing between the effects of conjugation which favours the planar form and steric effects which clearly favour the staggered configuration.Conjugation is more important in B2F4 than in B2C14 where the B-Cl bond (175 pm) is longer than the sum of the covalent single bond. radii implying negligible B-Cl double bond character in contrast to the situation in the monoboron halide^.^' [The preference of boron for 2p,-2p bonding with a first row element compared to 2p,-3p bonding with a second row element shows up in the structure of bis(diethy1amine) dithiaboretane which as predicted earlier,88 has been found to include a 4-membered planar ring with alternate B and S atoms.89 The B-S bonds are undoubtedly single whereas the B-N linkage shows considerable double bond character.] On the other hand steric repuIsion between two eclipsed BF2 groups should be less than that between two BC12 groups in a hypothetical planar B2C14.It is interesting that the B-B distance is shorter by 1.8 pm in staggered B2Cl than in the eclipsed B2F4. The fact that B2C14 is planar in the solid state is presumably due to packing forces within the crystal. 86 B. McClelland G. Gundersen and K. Hedberg 1.Chem. Phys. 1972 56 4541. 87 See for example Annual Reports 1975 72A,11 1 and refs. therein. J. A. Forstner and E. L. Muetterties Znorg. Chem. 1966 5 164. 89 G. Bushnell and G. A. Rivett Canad.J. Chem. 1977,55 3294. 136 R. H. Cragg J. D. Smith and G. E. Toogood Boron-Nitrogen Compounds.-Rotations about B -N Bonds. The stereodynamical behaviour of aminoboranes is of longstanding interest and several papers have appeared this Barriers to B-N bond rotation are typically 70-100kJ mol-' for monoaminoboranes and 40-50 kJ mol-' for bis(amino) boranes and this has been rationalized by postulating that the ?r-bond order of an individual B-N bond would be less in a bis(amino)borane than in a similar mono(amin0)- borane because in the former two nitrogen lone pairs would compete for overlap with the empty boron orbital.Obviously if such an argument is valid tris(amin0)- boranes should have still lower rotational barriers -but this has not proved to be the case when one of the amino groups is bulky. In this circumstances the bulky amino group may be forced to assume a position in which its nitrogen is unable to participate in B-N ?r-bonding and the rotational barrier corresponds to that of a bi~(amino)borane.~~ Of course the presence of bulky groups of any kind (whether or not they contain nitrogen) would be likely to increase the rotational barrier and this has been proposed as the chief reason for the high B-N rotational barriers in N-trimethyl-silyl -germyl or -stannyl derivatives of amin~boranes.~~ The rota- tional dynamics of tris(amino)boranes is also of interest because they are stereo- chemically correspondent with triaryl compounds of general formula Ar3X which have been extensively investigated.Both are known to be helically chiral with propellor-:ike ground states. Using 'H and 13Cvariable temperature n.m.r. spec- troscopy the steroc hemicall y rigid structures of bis (di- isoprop ylam ino)aI kylamino- boranes at low temperatures have been found to isomerize at higher temperatures by means of concerted rotations of the amino group^.^^,^^ Their low energy mode of isomerization involves a one-amine flip as opposed to the two-flip mechanism of isomerization of the triaryl species.The former seems to result from both steric and electronic effects whereas the latter is primarily sterically-decided. The factors at work in the aminoborane case seem to include the fact that the transition state for the one-amine flip strikes the best balance between the conflicting ?r-bonding and steric requirements. New Boron-Nitrogen Compounds. Among the interesting B-N compounds reported this year are mixed amino-hydrazino-boranes of high chemical and thermal stability,94 and diazadiboretidines (containing four-membered B2N2 rings).95 The latter have been synthesized in good yield by reaction of diazadistan- netidines with R'BX2 (R' = CH3or Cl) according to Scheme 1.Diazaborolines have CR3 CR3 I I N N RBX /\ R2Sn/ \SnR2 d2R'-B B-R' \N/ \N/ I I CR3 CR3 Scheme 1 90 R. H. Neilson and R. L. Wells Inorg. Chem. 1977 16 7. 91 K. K. Curry and J. W. Gilje J. Amer. Chem. Soc. 1977 99 8262. 92 K. K. Curry and J. W. Gilje ibid. 1978 100 1442. 93 M. J. '3. Dewar and P. Rona ibid. 1969 91 2259. 94 B. K. Christmas and K. Niedenzu 2.Naturforsch. 1977,326 157. 95 W. Storch W. Jacktiess H. Noth and G. Winter Angew. Chem. Internat. Edn. 1977 16,478. The Typical Elements 137 been complexed with a metal for the first time and a (CO)3Cr-one-faced sandwich Boron-nitrogen-phosphorus rings have not been derivative ~haracterized.~~ synthesized previously but one has recently been reported (l).97 (1) Boron-Sulphur Compounds.-B2S3 has been found98 to adopt a layered structure with only van der Waals linkages between the layers in marked contrast to crystalline B203 which has a three-dimensionally linked structure involving BO tetrahedra.Within each B2S3 layer there are zig-zag chains of planar six-membered B3S3 rings each of which is bonded via a planar four-membered B2S2 ring to a six-membered ring of the adjacent chain. The average B-S bond length in the structure is 180.8 pm but somewhat longer (182.3 pm) in the strained smaller ring. These values may be compared to the 184.4pm found in bis(diethy1- amin0)dithiaboretane which contains a similar four-membered ring.89 It is possible that the length of the B-S bonds in the latter compound is a consequence of the use of boron 2p orbitals to form .rr-bonds to the nitrogen atoms outside the ring (as mentioned in the boron halide section of the Report).The S-B-S and B-S-B bond angles are about the same (104-105" and 76" respectively) for these compounds. B& five-membered rings with S-S bonds (Vs Raman 440 cm-') are considered to be the building block of the polymeric material of composition H2S,xBS2 formed by thermal decomposition of meta -thioboric acid (HBS2)3 or from the previously unknown H2B& which was synthesized from H2S and 12B2S3 in benzene Considerable coverage was given in last year's Report to metal complexes of diborolines and thiadiborolines and it is only necessary here to mention the recent production of a triple-decker sandwich containing three thiadiboroline rings and two cobalt atoms.'" 2 Aluminium Gallium Indium and Thallium Aluminium.-Theoretical Considerations of Aquated A13+.The hydration number and energy of A13+ have been discussed on the basis of LCAO-MO-SCF ab initio calculations.101 A model of the hydrated cation which is limited to the first hydration shell does not account for the experimental value of 6 reported for the co-ordination number. Furthermore the computed hydration energy is well below the experimental enthalpy of hydration. An improved model where each water of q6 G. Schmid and J.Schulze ibid. 249. " R. T. Paine J. Amer. Chem. SOC.,1977,99,3884. 98 H. Diercks and B. Krebs Angew. Chem. Internat. Edn. 1977 16,313. 99 A.S. Gates and J. G. Edwards Inorg. Chem. 1977 16,2248. loo W.Siebert and W. Rothermel Angew Chem. Internat. Edn. 1977 16,333. lo' H.Veillard J. Amer. Chem. SOC.,1977,99 7194. 138 R. H. Cragg J. D. Smith and G. E. Toogood the first hydration shell is hydrogen-bonded to two water molecules in the second hydration shell yields a hydration energy which is close to the experimental value and also accounts for the co-ordination number. Up to six water molecules prefer to bond to the A13+ (in the primary shell); after this it is energetically favourable for the additional water molecules to occupy the second shell and to form hydrogen bonds to the first shell molecules rather than to insert in the first shell.[A similar analysis has been given for the (Li+ 5H20) cluster where one water molecule prefers to be in the second co-ordination ~phere]."~ The importance of strong secondary solvation has recently been invoked to account for the effects of temperature and added aprotic solvents on the 'H n.m.r. spectra of aqueous AI3+,lo3 and hydrogen-bonded interactions between the water molecules of the first shell and chloride anions in the second co-ordination sphere have been included in the interpretation of the X-ray diffraction data of aqueous solutions of lanthanoid M3' ions.'04 The S 1 mechanism proposed on the basis of n.m.r. studies,Io5 for the exchange reaction of water molecules between the aquo complex and non-co-ordinated water molecules is also substantiated in the theoretical study lo' pro-vided the exchange is considered to take place between the first and second hydration shells.The calculated activation energy 88 kJ mol-' is in good agreement with the experimental value of 112 kJ mol-'. It would be interesting to repeat the calculations for Ga3' to see whether the SN2mechanism proposed for its analogous exchange reaction'" is confirmed. Aluminium Trihalide-THF Interactions. A combined i.r. Raman and 27A1 n.m.r. spectroscopic investigation of aluminium trihalide-tetrahydrofuran interactions in solution and in the solid state has been carried OU~.~~~*~~~ The vibrational spectra of two general types of solid complex A1X3 THF and AlX3,2THF have been totally assigned.A molecular structure is indicated for the 1 :1 adduct and an ionic one [AIX2(THF),]'[AIX4]- for the 1:2 adduct (X = C1 or Br). Interestingly the ionic species is converted into the molecular one even at room temperature in Lhe solid state (under vacuum). The solution results were obtained in THF and in a mixed THF/dichloromethane solvent -the latter being needed to keep the AIX3,THF complexes in solution at high concentration. At room temperature the main species present were the molecular 1 1 or 1:2 complexes with the latter dissociat- ing into the ionic form noted above when X=C1 but not when X=Br. The equilibrium constants and activation energies for the reactions listed below were obtained $A12X6 +THF + AlX3,THF AIX,,THF+THF + AIX3,2THF 2[A1C13,2THF] + [A1CI2(THF),]'[AICl4]-ci~-AlC13,2THF$ ~xz~.s-AIC~~,~THF lo* P.A. Kollman and I. D. Kuntz ibid. 1974 96 4766. lo3 M. C. R. Symons Spectrochim. Actu 1975 A31 1105. lo4 M. L. Steele and D. L. Wertz J. Amer. Chem. SOC.,1976 98 4424. lo5 D. Fiat and R. E. Connick ibid. 1968 90 608. 106 J. Derouaul and M. T. Forel Znorg. Chem. 1977 16 3207. lo' J. Derouault P. Granger and M. T. Forel ibid. 3214. The Typical Elements Gaseous Complexes of Aluminium Gallium and Indium Halides with Transition Metal Halides. The volatility enhancement of certain transition metal halides in the presence of various 'acidic' A& gases (A=Al Ga or In; X=C1 or Br) has attracted a good deal of attention and the nature of the complexed species present in the gas phase has been the subject of several investigations.'08-''o It is apparent that the most stable species are of general formula MA2& (or MX2.2AX3) (M= Fe Co Ni or CU).'~~.'~~ However the 1 1 adducts are also ob~erved.'~' The extension of this work to the lanthanoids has been initiated with a view to the production of gas lasers involving AlX3/LnX3 gas phases."' Gallium.-Gallium Citrate Complexes.Gallium(II1) ions exhibit considerable anti- cancer activity and this fact has focused attention on the aqueous co-ordination chemistry of gallium the literature on which is very limited and often con-tradictory. Gallium radioisotopes are clinically administered as citrate buffered solutions and the extent of gallium localization in tumours depends on the citrate concentration.Citrate also seems to play a significant role in serum transportation of this metal. For this reason interest has concentrated on the nature of gallium citrate in aqueous ~olution,'~~~''~ but a general study has also been carried out to measure the stability constants of Ga'I' complexes with multidentate ligands most of which were carboxylic or aminocarboxylic acids.'I4 The Ga*I'-citrate system was investigated by 'H 71Ga,'12 and I3C'l3 n.m.r. spectroscopy. In moderately acidic solutions with [citrate] = [Ga3+],or in neutral solutions with [Ga3']/[citrate] >0.5 gallium citrate (1 1) polymers are formed. At high acidities these polymers break down to lower molecular weight species with 1 1metal :citrate ratios.In neutral or very slightly basic solutions the polymers break down into GaCit complexes. The only gallium species present in strongly basic solution (pH 2 12) is [Ga(OH4)4]-. The form of these polymers and complexes is reminiscent of the behaviour of Fe"' and presumably the biological interactions of Ga"' are influenced by their forma- tion in an analogous fashion to the interactions involved in the metabolism and transport of Fe"'.' 15.' l6 Indium.-Cloro -and Aquochloro -indate (111) Complexes. With the recent deter- mination of the crystal structure of trimethylammonium hexachloroindate(III),' l7 data are now available on the In-C1 bond lengths in all the [InCl,+,]"- species (x = 1 2 or 3) as well as for the aquochloro species [InC16-n(H20)n]m (n = 1,2 or 3; m = -2 -1 or 0).Two features consistent with the highly ionic character of the In-Cl bond previously discussed,' " and more recently confirmed by X-ray are clearly shown. Firstly within the non-aquated ions increasing co- F. P. Emmenegger Inorg. Chem. 1977 16 343. F. P. Emmenegger and F. Dienstbach ibid. 2957. 'lo G. N. Papatheodorov and G. H. Kucera ibid. 1006. G. N. Papatheodorov and G. H. Kucera A.C.S. Fall Meeting 1977 Abstract 135. J. D. Glickson T. P. Pitner J. Webb and R. A. Gams J. Amer. Chem. SOC.,1975 97 1679. C. H. F. Chang T. P. Pitner R. E. Lenkinski and J. D. Glickson ibid. 1977 99 5858. W. R. Harris and A. E. Martell Inorg. Chem. 1976 15 713. T. G. Spiro L.Pape and P. Saltman J. Amer. Chem. Soc. 1967,89 5555. T. G. Spiro G. Bates and P. Saltman ibid. 5559. J. G. Contreras F. W. B. Einstein M. M. Gilbert and D. G. Tuck Actu Cryst. 1977 B33 1648. 'I8 J. G. Contreras and D. G.Tuck Inorg. Chem.,1972 11 2967. B. H. Freeland J. J. Habeed and D. G. Tuck Canud. J. Chem. 1977,595 1527. R. H. Cragg J. D. Smith and G. E. Toogood ordination number is accompanied by monotonic lengthening of the bond. Secondly replacement of C1- ligands by water leads to a shortening of the remain- ing In-Cl bonds especially those which are trans to a water molecule. Consistent with this latter observation are the results of a recent Raman spectroscopic study of the aquochloro complexes present in aqueous solutions of indium(II1) chloride.120 The V (In-C1 stretch) bands in the four species characterized were respectively 279 289 300 and 31 1cm-' for [In(H20),C14]- [In(H20),C1,] [In(H20)4C12]' and [In(H20)5C1]2+. Electrochemical Behaviour of Indium Metal. The passivity of indium metal to the solution of indium ions has been shown to be virtually absent in contrast to the situation with aluminium. The probable cause is the lack of oxide coating on the indium metal although other (electronic) factors may be important. Overvoltage to H2evolution was quite high but almost zero for indium metal deposition. 12' Thallium.-Thallium Complexes. (i) T1I". Interest in the complexes of Tl"' chlorides continues with several new reports during the year. 122-124 Many of these materials have been formulated as ionic complexes of the type exemplified by [T1C12(terpy)]'[T1C14]-, on the basis of i.r.and conductivity data. Some doubt has been cast on this interpretation by the discovery that their X-ray p.e.s. show no evidence for more than one T1 site within each c~mpound.'~~ On the other hand the differences in the spectra of different compounds are also not very large and it may well be that the technique is not suited for such distinctions. The first tertiary phosphine complex of T1Cl3 has been reported. '22 Diethylphenylphosphine forms T1C13L2,5 when the ligand and [TlC13(py)2]n are mixed in acetone. Other phos- phines however reduce the Tl"' to Tl'. (ii) Tl'. The 'inert' pair on T1' seems to be stereochemically inactive when that ion is complexed with macro-cyclic crown ethers.The products of reaction between various T1' salts and 15-crown-5 and dibromo-benzo- 15-crown-5 respec- tively have analogous sandwich structures to those formed by potassium with two molecules of ligand to one T1' ion.126 The inactivity of the lone pair in these compounds contrasts sharply with earlier studies on other T1' complexes. '*' PART 111 Groups 1V and V By J.D. Smith 1 =-Bonding between First and Second Row elements It is 20 years since the publication of the Valence Shell Electron Pair Repulsion (VSEPR) Theory. This has been remarkably successful in interpreting shapes of simple molecules 2nd cases where its predictions are incorrect have aroused much T. Jaarv J. T. Bulnier and D.E. Irish J. Phys. Chem. 1977,81,649. I*' A. N. Campbell Canad. J. Chem. 1977 55 1710. 12* V. Ozimec T. J. Smith and R. A. Walton J. Znorg. Nuclear Chern. 1977 39 362. K. C. Malhotra and Balkrishnan ibid. 387. K. C. Malhotra and Balkrishnan ibid. 1523. R. A. Walton J. Znorg. Nuclear Chem. 1977 39 549. M. E. Farago Inorg. Chim. Acta 1977 23 211. See for example Annual Reports 1975 72A,p. 118 refs. 22a-c. The Typical Elements interest. Some of the problems in making deductions about electronic effects from the shapes of molecules have been illustrated by structural determinations in 1977. The compound N(SCF,),(l) is like N(SiH3)3 practically non-basic and the NS3-skeleton has been shown by electron diffraction to be almost planar.' It is CF3 / CF'3 (11 however too simple to deduce from bond angles alone that p+d 7-bonding between nitrogen and sulphur is established since the S .. . S distance is close to the sum of the 'one-angle' radii,2 i.e. the distance below which there is serious repul- sion between two atoms attached to a third. More reliable evidence for rr-bonding comes from the N-S bond length [ 17032) pm]. Although the C-S bond in Me2S is 2.1 pm longer than the C-C1 bond in MeC1 the N-S bond in N(SCF3)3 is 5.4pm shorter than the N-Cl bond in NCl,. 7-Bonding may be expected to influence the barrier to rotation about the N-S bond. An unexpected feature of the structure of N(SCF3) is that two trifluoromethyl groups lie above the NS3 plane and one lies below. At 20°C the CF3-groups appear to be equivalent on the 19F n.m.r.timescale but two environments are distinguished at -96 "C. The energy barrier to rotation about the N-S bond found from these n.m.r. spectra is 25 kJ mol-' but semi-empirical molecular orbital calculations suggest the p -+ d rr-bonding though important is not crucial. It may actually be greater in the conformation with the carbon atom in the NS plane than in the conformation adopted. If p +d 7-bonding is important in Si-N bonds bond lengths are expected to be shorter in species with negative charges on nitrogen than in neutral species. The Si-N distance in K[N(SiMe3)2]2THF (164pm) is indeed shorter than those in NMe2(SiH3) (171.6 pm) NH(SiH3)* (172.5 pm) and N(SiH3)3 (173.4pm); all are shorter than the sum of the covalent radii (187 pm).N-Sodiohexamethyldisilazane is dimeric in hydrocarbon solution but an X-ray structure determination3 shows that in the solid Na' cations and [N(SiMe3)2]- anions (2) are linked into chains. N-SiMe3 The Si-N bond distance is 169(2)pm compared with 168-175 pm in other compounds M[N(SiMe,),] (M = Eu Cr Ni Co Cr Be Sc Fe H or Al; n = 1-3. The Si-N-Si angle 125.6(1)" is in a range shown by other bis(trimethylsily1) amido-derivatives but there is no simple correlation between Si-N bond lengths and Si-N-Si bond angles. ' C. J. Marsden and L. S. Bartell J.C.S. Dalton 1977 1582. C. Glidewell Znorg. Chem. Am 1975 12 219. R. Griining and J. L. Atwood J. Otganometal. Chem. 1977,137 101. R. H. Cragg J. D.Smith and G. E. Toogood Table 1 Bond lengths and angles in compounds (R3Si)20 Pauling R electronegativity Si-O(pm) Si-0-Si(O) F c1 4.0 3.2 158.0(25) 159.2(10) 155.7(20) 146(4) Ph H 3.0 2.2 161.6(1) 163.4(2) 180.0 144.1(9) A similar picture emerges from disiloxanes with the Si-0-Si skeleton (Table 1). An X-ray determination4 of the crystal structure of the triphenyl compound (Ph3Si),0 shows that the molecule is centrosymmetrical so the Si-0-Si group is linear. Although electronegative substituents on silicon are associated with short Si-0 bonds there is again no clear correlation between bond lengths and Si-0-Si bond angles. The neutral compound (Ph3Si)20 is isoelectronic with the ion (Ph3P),N’ which is linear in crystalline [(Ph,P),N][V(CO),] [P-N = 153.9(2)pm] but more usually bent with an average P-N distance taken over several compounds of 157.5(2)~m.~ The P-N-P angles are in the range 134.6- 141.8’.These results provide further evidence that energy differences between linear and bent forms are small. The double ylide Ph3P=C=PPh3 is another member of the same isoelectronic series. In the solid there are two crystallo- graphically independent molecules with P=C=P angles 143.8(6) and 130.1(6)”. Since the part played by intermolecular forces in the solid is not clear the related molecule Me3P=C=PMe3 has been examined by gas phase electron diffraction.6 A different complication now arises since the observed structure may deviate from that of the ground state because of excitation of a low frequency bending vibration.The best interpretation of the electron diffraction data gives molecular parameters P=C = 159.4(3); P-C = 181.4(3) pm; P=C=P = 147.6(5) C=P-C = 116.7(4)” H I C Ph3P//yL PPh3 (3) (41 (51 and is based on a model allowing free rotation about P=C bonds. Free rotation is consistent with a linear P=C=P skeleton since in contrast to allenes the orbitals available for r-bonding (phosphorus d and carbon p) are all members of degenerate pairs. The P=C=P angle found by electron diffraction must then be explained by assuming a low frequency bending vibration of about 80 cm-’ similar to that in carbon suboxide (63 cm-I). The related molecule Me3P=CH2 has also been examined;’ the C3P=C skeleton has very similar parameters to those in Me3P=C=PMe3.The P=C bond [164.0(6)pm] is the shortest yet reported for compounds with 3-co-ordinate carbon; the slightly shorter bond in C. Glidewell and D. C. Liles J.C.S. Chem. Comm. 1977 632. R. D. Wilson and R. Bau J. Amer. Chem. Soc. 1974 96 7601. E. A. V. Ebsworth T. E. Fraser D. W. H. Rankin 0.Gasser and H. Schmidbaur Chem. Ber. 1977 110,3508. E. A. V. Ebsworth T. E. Fraser and D. W. H. Rankin Chem. Ber. 1977 110 3494. The Typical Elements 143 Me3P=C=PMe3 is consistent with a change in carbon hybridization from sp’ to sp. This still leaves unexplained the P=C=P angle in the crystalline hexaphenyl compound Ph3P=C=PPh3. The related ion (3) has been isolated as bromide and shown by X-ray diffraction to have P-C 170.3pm and P-C-P 128.2(3)”.By comparison the compound Ph2P(:Se)CH2P(:Se)Ph2 (4) has P-C 184.3(14) pm and P-C-P 117.9(6)”. The angles at carbon appear to be insensitive to changes in bond lengths.8 Structural studies on silyl isocyanate (m.p. 183 K) raise similar problems. Vibra- tional and rotational spectra show that the molecule is a symmetric top but electron diffraction gives a Si-N-C angle of 152”. Again the discrepancy has been explained by a low-frequency bending vibration. An X-ray determination’ at 140 K has shown that the molecule (5) is planar except for H’ with Si-N = 172.3(4) pm Si-N-C angle of 158.2(3)” and N-C-0 angle of 176.2(5)”. The linear ground state for the molecule in the vapour does not persist in the crystal where the low frequency vibration is apparently constrained by packing forces.2 B-Group Metal Clusters The use of alkali metal-crown ether derivatives to characterize polyatomic anions Sng4- Pb5’- and Sb73- was described in the Annual Report for 1975. Several further examples of this procedure have now been published. The intermetallic compounds K5Bi4 and K3Bi2 react with solutions of the crown ether N(CH2CH20CH2CH20CH2CH2)3N (crypt) in ethylenediamine (en) to yield deep green-red dichroic solutions from which crystals of [crypt KI2+[Bi4l2- precipitate.” The Bi4’- ion is square planar like the isoelectronic Te,”.” and the mean Bi-Bi distance (293.9 pm) indicates a high bond order [c.f. 307 pm (~3) and 353 pm (~3) in the metal]. Similarly the compound [crypt KI2+[Te3l2- en is formed from K2Te; the Te32- ion is isoelectronic with I,+ and has Te-Te 269.2(5) and 272.0(4)pm and a Te-Te-Te angle of 113.1(2)”.The slight deviation from CZvsymmetry appears to arise from weak hydrogen bonding. Molecular parameters for the trigonal bipyramidal ion SnS2- isostructural with Pbs2- described earlier have also been found by X-ray diffraction. ” The characterization of these polyatomic anions is useful in understanding the structures of compounds where the bonding is intermediate between ionic and metallic. Thus Bi groups are observed in crystals of Cal Bile.l7 The AS^^-cluster (6) is found in the compound Ba3As14 obtained by fusing the elements at 1000-1 100 K.’ Related cationic clusters Te2Se4” (7) and Te,Se,” have been isolated as hexafluoro-arsenates(V) and -antimonates(~).’~ If one Se” is added between Se’ and Se‘ in (7) the resulting ion SeSTe24+ becomes isostructural with AS^^-(6) and a-As4Se3(8).l6 In the grey air-sensitive Zintl phase Cs2As, P.J. Carroll and D. D. Titus J.C.S.Dalton 1977 824. ’M. J. Barrow S. Cradock E. A. V. Ebsworth and M. M. Harding J.C.S. Chem. Comm. 1977 744. I” A. Cisar and J. D. Corbett Znorg. Chem. 1977 16 632 2482. J. Barr R. J. Gillespie G. P. PPZ,P. K. Urnrnat and 0.C. Vaidya Znorg. Chem. 1971 10 362. P. A. Edwards and J. D. Corbett Inorg. Chem. 1977,16 903; J. Amer. Chem. Soc. 1977,99 3313. l3 K. Deller and B. Eisenrnann. 2.Narurforsch.. 1976 31b 29 1023 1146. W. Schrnettow and H. G. von Schnering Angew Chem. Internat. Edn. 1977 16 857. R.J. Gillespie W. Luk E. Maharajh and D. R. Slim Inorg. Chem. 1977 16 892; R. J. Gillespie W. Luk and D. R. Slim J.C.S. Chem. Comm. 1976 791. I‘ T. J. Bastow and H. J. Whitfield J.C.S. Dalton 1977 959. R. H. Cragg J. D. Smith,and G. E. Toogood 2+ however arsenic atoms are linked into puckered linear chains with As-As 252 pm (c.f. 250.4 pm in rhombohedra1 arsenic). If the compound is considered to be ionic the anions AS^^-and Ass*'-are isoelectronic with Sen2- (n = 4 or 8) found in polyselenides. In SrSb2 the antimony atoms form zig-zag chains Sb,"- which are isoelectronic and isostructural with the Ten chains in tellurium metal. The shortest Sb-Sb distances (290-292 pm) are again comparable with those in the element (290.8 pm)." The deep red compound [crypt K]6+Ge92-Ge94- 2.5 en contains two separate polyatomic anions." Assignment of charges to these from bond lengths alone is not straightforward.One anion is only slightly distorted from the closo D3,,struc-ture (9a) based on a square-capped trigonal prism shown by other clusters e.g. B9H92- with 10 skeletal electron pairs. This is assigned as Ge92-. The addition of two electrons to a closo structure based on a polyhedron of n corners is predicted18 to lead to a nido structure in which one corner is missing from a polyhedron of n + 1 corners. The second polyatomic anion like Sn94- described previously,12 has a structure (9b) based on a capped square antiprism as expected for a cluster with 11 skeletal electron pairs. Transformations from D3,,(9a) to C4"(9b) require only small atomic displacements and little energy.An exactly similar argument relates Te64+ and Te3Se3''. Te64+ with 10 pairs of electrons has the expected square-capped trigonal prismatic structure (9a) with 3 corners (shown by daggers) missing. Te3Se3*+ (7) with 11 pairs has the capped antiprismatic structure (9b); the three missing corners are again shown by daggers. The structure of the other €3 Group cluster compounds may similarly be rational- ized by applying the rules for counting electrons worked out for boron hydride carbaborane and transition-metal clusters." Thus Bi42- and Te4*+ with 7 skeletal electron pairs adopt an octahedral structure with two corners missing; Sns2- and PbS2-(6 pairs) form a closo-trigonal bipyramid; (12 pairs) can be considered as having the structure (10)of an eleven-cornered polyhedron from which four corners are missing.It would be interesting to find further examples. For cationic species it appears to be possible to accommodate two electrons in almost non-bonding orbitals thus Bi9'+ isoelectronic with the yet uncharacterized Pb94- has the D3,,(9a) instead of the expected C4"structure (9b). Polyatomic anions have also been generatedlg from reactions of sodium-tin -germanium and -antimony alloys with ethylenedlamine in the absence of crown ether but the solids isolated are not so l7 C. H. E. Belin J. D. Corbett and A. Cisar J. Amer. Chem. Soc. 1977 99,7143. l8 K. Wade Chem. Brit. 1975 11 177; Adv. Inorg. Chem. Radiochem. 1976 18 1. l9 L.Diehl K. Khodadadeh D. Kummer and J. Strahle Chem. Ber. 1976 109 3404. 145 The Typical Elements well-characterized the ether seems more effective than ethylenediamine in prevent- ing return of charge to the alkali metal ions. Measurements of conductivities have shown that the clusters persist in ethylenediamine solution. 3 Oxyanions One of the most striking contrasts between the chemistry of first and second row elements is in the stoicheiometry and structures of oxyanions. Carbon and nitrogen give planar ions C032-and NO3- whereas the second row elements are tetra- hedrally co-ordinated in silicates and phosphates. There have been reports in 1977 of both ‘orthonitrate’ N043- and ‘metaphosphate’ PO3- counterparts to the long- established species.The reaction between sodium nitrate and disodium oxide at 300°C yields a compound Na3N04.20 It is very reactive towards moisture and carbon dioxide and quickly decomposes to carbonate and hydroxide on exposure to air. The Raman spectrum is different from that expected for nitrate or nitrite; instead it shows four peaks characteristic of a tetrahedral species. It seems reasonable to attribute these to N043- since the Raman shifts are very similar to those of the isoelectronic NF4’. In a similar experiment starting from sodium nitrite and disodium oxide the product was Na3NO3; this however was shown by X-ray studies to have an antiperovskite structure represented by [Na’],[02-]. [NO2-] i.e. without discrete ‘orthonitrite’ anions N033-. The monomeric metaphosphate ion has not been isolated in a crystalline compound but it has been suggested as an intermediate in the hydrolysis of aryl phosphates21 [equation (l)].At pH 7-43 measurements on [‘80]-2,4-dinitrophenyl phosphate (11) show a kinetic isotope effect of 2% compared with the 160-compound which suggests that the slow step is breaking of the P-0 bond. O,N In the presence of pyridine or dioxan the reaction appears to be bimolecular suggesting that the PO3-ion reacts rapidly with bases. A much lower isotoFz effect is observed in the buffer-catalysed hydrolysis of the triester (PhCH20)2P(0)OC6H3(N02)2-2,4;in this case therefore a 5-co-ordinated 2o M. Jansen Angew. Chem. Internat. Edn. 1977 16 534; 1976 15 376. D. G. Gorenstein Y.-G.Lee and D.Kar J. Arner. Chem. Soc. 1977,99 2264. R. H. Cragg J. D. Smith,and G.E. Toogood transition state seems more likely. A new bond to the attacking nucleophile is formed before the P-OAr bond is broken. The main features of silicate chemistry have been established largely by X-ray methods. In many silicate minerals Si044- tetrahedra are linked into chains as in pyroxenes (12) double chains (13) as in amphiboles or sheets as in micas. A hitherto puzzling aspect has been the lack of structures with several pyroxene chains linked into bands e.g. (14) intermediate between the amphiboles and sheet silicates. (12) (13) (14a) (14b) Structures with 3 4 or 6 chains of SiO tetrahedra linked together have now been identified in jade22 using high resolution electron microsc~py.~~ This tech- nique applied to very thin specimens a few nm thick may be made to generate a ‘lattice image’ which is related to the charge density within the specimen.23 The lattice image may be compared with images calculated for various structures.Under favourable circumstances the comparison may give information not obtainable by classical X-ray methods about planar defects and coherent growth between related structures. The jade specimens examined show wide variations in internal structure. Examples have been found of isolated single 3- 4- and 6-fold chains of linked Si04 tetrahedra- in matrices of amiphibole structure (13). Triple chain structures (14a) have also been observed as ordered sequences about 40 nm wide in surrounding amphibole structures (13) and isolated double chains (13) have been observed in surrounding triple chain structure (14a).New and unsuspected linear structural faults have also been revealed as interconversions of various multiple chains. High resolution electron microscopy has also shown a similar rich variety of linear and planar defects in the chain silicate wollastonite CaSi03 which has SiO sequence (15). The chains pack together in several ways in samples from various geological sources.24 It is a little surprising that in spite of widespread studies on solid silicates simple species in silicate solutions are not well characterized. Equilibria are often approached slowly potentiometric measurements are hampered by interaction between the species under study and buffer solutions and the concentration of the neutral Si(OH) appears to be too low for many structural measurements though SiO(OH)3- and Si02(OH)22- have been detected by Raman spectroscopy.A potentiometric method has now been used to study the ionization of silicic acid and polysilicate formation in 1M-NaCl between 60 and 290°C.25 At the lowest Si‘” concentrations (0.005 M) only mononuclear species occur over wide ranges of 22 L. G. Mallinson J. L. Hutchison D. A. Jefferson and J. M. Thomas J.C.S.Chem. Comm. 1977 910. 23 J. S. Anderson Chem. Brit. 1977 13,182. 24 J. M. Palomino D. A. Jefferson J. L. Hutchison and J. M. Thomas J.C.S. Dalton 1977 1834. ’’ R. H. Busey R. E. Mesmer Znorg. Chem. 1977 16 2444. The Typical Elements temperature and pH but polysilicates are detected at higher concentrations.Pre- cise thermodynamic data have been obtained for the reaction (2). Si(OH) +OH-S SiO(OH)3-+H20 (2) The equilibrium is only weakly dependent on the sodium ion concentration. Careful thermodynamic measurements have also been made on aqueous sodium phosphate solutions (Na20:P205 = 1.8-10.0) at 250-300 "C. Detailed under- standing of phase equilibria is important because sodium phosphate is widely used for treatment of water for power station boilers. More alkaline saturated solutions are in equilibrium with a complex orthophosphate 11Na20.4P205,3H20but less alkaline solutions are in equilibrium with Na2[HP0,] at 250°C and Na4[P07] at 300 "C. Various hydrates at Na2[HP0,] may be crystallized at lower temperatures and above 300 "C there appears to be a region of liquid-liquid immiscibility.26 4 Ylides For many applications in organic syntheses the most useful ylides are alkylidenetriarylphosphoranes Ar3P=CR'R2.Reactions of alkylidenetrialkyl-phosphoranes are more complicated and more subtle and continue to provide a focus for much research.27 In 1977 interest has been concentrated on factors affecting positions for proton abstraction intra- or inter-molecular attack of the ylide carbon on electrophilic centres and on complexes with transition metal compounds. The structures of the ylides Me3P=CH2 and Me3P=C=PMe3 are referred to in Section 1 above. Problems relating to positions for proton abstraction are illustrated in the reaction sequences (3)-(5).28 Although the reaction between the triphenylphosphonium salt (16) and base involves abstraction of two a-protons from the side chain to give the ylide in which the phosphorus is exocyclic [equation (3)] the trimethyl compound reacts with abstraction of protons from separate methyl groups to give the compound (17) in which the phosphorus is in the ring.The bromide (18) however does not 26 D. Broadbent G. G. Lewis and E. A. M. Wetton J.C.S. Dalton 1977,464. 27 H. Schmidbaur Accounts Chem. Res. 1975,8 62; Adv. Organometallic Chem. 1976 14 205. 28 H. Schmidbaur and H. P. Scherm Ckem. Ber. 1977 110 1576. R. H. Cragg J. D. Smith,and G. E. Toogood [Ph3P(CH2)4Br]Br b Ph3P=CH(CH2)3BrNaNH2 -+ Ph,P=C 3 (31 (16) NaNH2[Me3P(CH2)4Br]Br +Me2P(CH2)4Br+ I1 Br--+ (4) react in the same way to give the ylide (19):instead initial proton abstraction seems to be from the 3-bromopropyl side chain with formation of the cyclopropylphosphorus (19) ylide [equation (5)].Products in reactions of this kind are clearly difficult to predict. The ylides Me,Bu:-,P=CH2 may be made at low temperatures from tetra-Me3p4X" [Me P (CH 2)3B r]B r NaNH2+ Me3P=CH(CH2)2Br -P Br-+ Me\ /P+CH2 (18) Me (5) alkylphosphonium salts and sodamide in tetrahydr~furan.~~ The tri-t-butyl-deriva- tive Bu:P=CH2 decomposes at 20 "C with elimination of but-1-ene. The di-t-butyl derivative reacts similarly at 80 "C. The elimination is attributable to the crowding in the molecule shown by 'H and 13 C n.m.r.spectra which indicate restricted rotation about P-C bonds. The usual reaction between an ylide and a chloroalkane gives a phosphonium salt [equation (7)]. R:P=CH2 + R2Cl + [R:PCH2R2]CI (7) Although several examples are known in which the alkyl group R2 bears a silicon- containing substituent reactions between trimethyl-or triethyl-methyl-idenephosphorane and chloromethylsilane give the product (20; R = Me or Et).30 It appears therefore that the attack of the ylide carbon is on silicon rather than carbon probably reflecting the decreased polarity of the C-Cl bond in SiH3CH2Cl compared with CH3CH2Cl. The subsequent 1,2-hydride shift is confirmed by the reaction between D3SiCH2C1 and Me3P=CH2 which gives Me3P=C(SiD2CH2D) as product.An intramolecular attack of an ylide carbon on 29 H. Schmidbaur G. Blaschke and F. H. Kohler Z. Nuturforsch 1977 32b 757. 3" H. Schmidbaur and B. Zimrner-Gasser Angew. Chern. Internut. Edn. 1977 16 639. The Typical Elements 149 R3P=CH2 R3P=CH2 +H3SiCH2CI -+ [R3PCH2SiH2CH3]Cl -[R3PMelC D R3P=CHSiH2CH3 H3SiCH2C' b R3P=C(SiH2Me)2 (20) a silicon atom is suggested by the observation31 that the silylaminophosphonium compound (21) reacts with n-butyl-lithium even at O'C to give the phosphine imine (22) rather than the ylide (23). Me3Si CH2 CH3SiMe3 \ I1 I N-P-Me -+ Me3Si-N=P-Me (8) [(Me3Si)2NPMe3]'I-Bu"LI I I Me&/ Me Me (21) (23) (22) Double ylides R2MeP=C=PR2Me (R=Me or Ph) are easily formed by de- protonation of the phosphonium ions [R2MePCH2PR2Me]2'.Although proton shifts in such systems occur extremely easily there is no evidence for non-cumulene isomers e.g. R3P=CH-PR2=CH2. In attempts to force deprotonation at the terminal methyl groups phosphonium salts [Ph2MePCMe2PPh2]Cl and [Ph2MePCMe2PPh2Me]12 were made.32 The ylide products from these were (24) and (25). Me Me CMe CH -H+ +/\ / [Ph2MePCMe2PPh2]+-Ph,P \-PPh + Ph,P \c \/ ;; CH (9) (24) [PhzMePCMe2PPh2Me]" -2H1 b Ph2MeP=C=PPh2CHMe2 (10) (25) Again the key step seems to be nucleophilic attack by the ylide carbanion this time on an adjacent phosphorus. Further progress has been made in isolating chlorine-substituted phosphorus ylides. Dichloromethylidinetriphenylphosphorane Ph3P=CC12 may be isolated by the reaction between [Ph3PCCl3]C1 and P(NMe2)3 but the same preparative route is unsuccessful for Ph3P=CHC1 since the strong base P(NMe2)3 promotes dehy- drochlorination as well as dechlorination.The chloromethylene compound is however neatly made by transylidation. Ph3PzCH2 + [Ph3PCH*Cl]Cl + [Ph3PCH3]Cl+ Ph,P=CHCl (1 1) This exploits the increasing basicity in the series Ph3P=CC12 <Ph,P=CHCl< Ph3P=CH2 as electron withdrawing CI-substituents on carbon are replaced by hydrogen.33 31 J. C. Wilburn and R. H. Neilson J.C.S. Chem. Comm. 1977 308. 32 A. Wohlleben and H. Schmidbaur Angew. Chem. Internal. Edn. 1977 16 417; H. Schmidbaur 0. Gasser and M. S. Hussain Chem. Ber. 1977 110 3501. 33 R. Appel and W. Morbach Angew.Chem. Internat. Edn. 1977.16 180. R. H. Cragg,J. D. Smith and G. E. Toogood It is now clear that phosphorus and arsenic ylides show an extremely rich chemistry involving complexes with metal centre^.^' The ylide groups may be of several general types (26)-(30) (E = P or As). The arrangement (26; R = H) in which the ylide is a two-electron donor is found in the compound [(Ph3AsCH2)2Cu]C1 made from copper(1) chloride and the ylide Ph3A~CH2.34 Related phosphorus compounds were described in 1974. Neutral compounds with sequence (26) may be obtained from diarylcarbene complexe~.~~ [(CO)5W=CAr2] Me3P + [(CO)5W{CAr2PMe3}] (Ar = Ph 2-C4H3S or 2-C4H30) Type (27) is illustrated by the compound [(q-CSHs)2Ti(CH2)2PMe2]36 obtained from the reaction between bis(cyclopentadieny1)titanium dichloride and the lithi- ated ylide Me2P(CH2)2Li.Type (28) is found in the complexes (31) derived from ylide (17) and in the group of compounds (32) (M = Cu Ag or Au; E = P 0.1-As). Several new examples of this kind have been de~cribed.~’ The Au-Au distance [302.3(1) pm] in the compound (32; M = Au; E = P; R = Et) 14 pm longer than that in gold metal shows that the metal-metal interaction is weaker than in the compound (33) (Au-Au 259.7 pm). Complexes with bivalent metals (34; M = Mg Cd or Zn; X = CH2,R = Me) are insoluble in organic solvents38 and are presumed to have polymeric structures like those of the isoelectronic phosphinates (34; X = 0). Few complexes with the grouping (29) are known but two compounds (35) and (36) have been described,39 in which both chelate systems (27) and (29) are found.A series of bis(dipheny1phosphino)methane complexes analogous to (32) and (33) have 34 W. Richter Y. Yamamoto and H. Schmidbaur Chem. Ber. 1977 110 1312. 35 F. R. Kreissl and W. Held Chem. Ber. 1977 110 799. 36 H. Schmidbaur W. Scharf and H.-J. Fuller 2.Nufurforsch,1977 32b,858. ” H. Schmidbaur J. R. Mandl W. Richter V. Bejenke A. Frank and G. Huttner Chem. Ber. 1977,110 2236; 1976,109,466. 38 H. Schmidbaur and J. Eberlein Z. anorg. allg. Chem. 1977 434 145. 39 H. Schmidbaur and J. R. Mandl Angew. Chem. Infernat. Edn. 1977 16 640. The Typical Elements 151 CI I Et CH2-Au-CH2 Et \/ \/ P P /\ i /\ Et CH,-Au-CH Et I Cl RR (33) (34) also been made.40 These have the system (37)with an extra hydrogen compared with (29).Complexes with the ring system (30;X = N or CH)41 were described in last year’s Report. CH3 / Et2P PPh H-C \. Pd \ /PEt2 [(Me3P)2PdC121 LiCH(PPh2)z ,[Pd{CH(PPh2)2)21 \\CH2 4 \ / FH,\ PPh CH (35) The reactions between dimethylaminotitanium chlorides (Me2N)3TiCl and (Me2N)2TiC12 and the ylide Me3P=CH2 give the product (38) in which titanium atoms are bridged by a single ylide carbon;36 the structure is reminiscent of the p -trimethylsilylmethylidene compounds (39; M = Nb or Ta).42 Another surprising product from the reaction between an ylide and a transition metal compound is the iron compound (40) obtained from the tin ylide (Me3Sn)2C=PPh3 and tri-irondode~acarbonyl.~~The two Fe(C0)3 groups are bridged by a Ph PMe3 IIC (MC~~N)~T~’‘Ti(NMe2) SiMe3 I C (Me3SiCH2)2M\.-,yM(CH2SiMe3)2/--1 II PMe3 ‘c’ SiMe3 C I 40 H.Schmidbaur A. Wohlleben U. Schubert A. Frank and G. Huttner Chem. Ber. 1977 110,2751. 41 H. Schmidbaur 0. Gasser C. Kruger and J. C. Sekutowski Chem. Ber. 1977 110,3157; H. Schmidbaur H. J. Fuller V. Bejenke A. Frank and G. Huttner Chem. Ber. 1977 110,3528 3536. 42 W. Mowat and G. Wilkinson J.C.S. Dalton 1973 1120. 43 M. R. Churchill F. J. Rotella E. W. Abel and S. A. Mucklejohn J. Amer. Chem. SOC.,1977.99 5820. R. H. Cragg J. D. Smith,and G. E. Toogood -C(CHo)PPhzc6H4-ligand derived from the original ylide and a carbonyl group from the Fe3(C0)1z. The positive charge on the phosphorus appears to be balanced by negative charge delocalized over the C6H4 ring.In a further series of complexes the ylide behaves as a three-electron donor. Thus the carbyne complexes (41; M = Cr R = C6H6 1,4-Me2C6H4 or 1,3,5-Me3C6H3; M = Mn or Re; R = C5H5) react with trimethylphosphine to give the compounds (42) [Equation (13)].44 At -78"C a second mole of trimethylphosphine is taken up by the rhenium compound (42; R = C5H5 M = Re).45 Ph PMe3 I [q-C5H5)(C0)2Re=C/ ][BC14]+Me3P + [(q-C5H5)(C0)2Re-CPh ][BC14] (14) I 'PMe3 PMe3 (43) The product (43) is easily dissociated at room temperature to trimethylphosphine and the complex (42). In the yellow diamagnetic compound (44) the ylide group acts as a 3-electron bridging ligand.46 Ph PMe3 \C/ Me3P /\ [(C0)5Re(CO)4W~CPh]+(C0)4Re-W(C0)4 \C/ II 0 (44) 5 Phosphorus-Nitrogen Compounds Last year's Report was concerned mainly with compounds having four-membered P2N2 rings; this year's reverts to discussion of the widely studied cyclic phosphazene trimers tetramers and polymer^.^' An important objective in phosphazene chemistry is the synthesis of high poly- mers with only alkyl or aryl substituents.Such polymers are expected to have good thermal stability and to be less susceptible to depolymerization to cyclic oligomers than polyphosphazenes with bulky or easily ionized substituents. Such high mole- cular weight polymers may in principle be obtained by polymerization of the well established oligomers (R2PN)3or or by alkylation of polyhalogenophosphazenes.44 F. R. Kreissl P. Stiickler and E. W. Meineke Chem. Ber. 1977 110,3040. 45 F. R. Kreissl K. Eberl and P. Stuckler Angew. Chem. Internat. Edn. 1977 16,654. 46 F. R. Kreissi P. Friedrich T. L. Lindner and G. Huttner Angew. Chem. Internat. Edn. 1977 16,314. 4' H. R. Allcock Angew. Chem. Internat. Edn. 1977 16 147. The Typical Elements 153 Both possibilities have been examined. The compounds (Me2PN) and (Me2PN)4 may be interconverted between 200 and 350°C but they do not p~lymerize.~~ The reaction is inhibited by base (e.g. NaNH2 or NaOMe) and accelerated by acid suggesting that the isomerization involves protonation of a ring nitrogen. So far dimethylphosphazene polymers have not been obtained either by heating diaminodimethylphosphonium chloride or from the oligomers (Me2PN)3 or4.Since the compounds (X2PN)3,,4 (X =F C1 or Br) and (MeZSi0)30r4 can polymerize it is difficult to see why the dimethylphosphazenes cannot; it may be that the tetramer (Me2PN)4 is an ‘energy trap’ and that formation of high molecular weight polymer requires generation of a species that can bypass the tetramer in a rapid chain propagation ’process. Even at 350°C only 2% of the dimethylphosphazene is decomposed in a week this suggests that polymeric compound would if formed show good stability at this temperature. High molecular weight poly(difluorophosphazene) was chosen for alkylation studies49 since the chloro-compounds are known to be degraded with cleavage of P-N bonds by Grignard or organolithium reagents.Carefully controlled poly- merization of (F2PN) at 350 “C in an autoclave yields a material that is soluble in solvents such as perfluoro-2-butyltetrahydrofuran or perfluorodecalin with a little diethyl ether. This polymer reacts with phenyl-lithium or diethylmagnesium to give the phosphazenes (45) and the remaining fluorine substituents may be replaced in the usual way by treatment with CF3CH20Na [equation (16)]. F F R F R OCH2CF3 [-N=P-\/In +[-N=P-\/In +[-N=P-\/ In (16) (45;R =Ph or Et) A polymer in which 62% of the side groups were Ph and 38% OCH2CF3 had a molecular weight of ca. lo6 and appeared to be more stable at 300°C than did [(CF3CH20)2PN]n. Complete replacement of the halogens in (45) by phenyl could be achieved; but the molecular weight of the resulting polymer was reduced to ca.2500. Phosphazene polymers with OCH2CF3 groups are water-and oil-repellant. Polymers with amino groups however are soluble in water and therefore possible complexing agents for square-planar platinum compounds used as anticancer drugs.” Complexing prevents rapid excretion and mitigates side effects such as kidney damage observed with uncomplexed C~S-[P~C~,(NH~)~]. The polymer [(MeNH),PN] (n =ca. 15 000) reacts with potassium chloroplatinate(I1) in chloro- form in the presence of 18-crown-6-ether to form a material from which platinum is not removed by dialysis in water; preliminary screening tests show that this is particularly active as an antitumor agent.In order to discover whether the pla- tinum is bound to chain or side-group nitrogen atoms the compound (46) was isolated from the tetramer [(MeNH),PN],. An X-ray study shows that the square- planar platinum is bound to 2,6-skeletal nitrogen atoms and it is reasonable to assume that the platinum is similarly bound in the polymer. In acid solution the 48 H. R. Allcock and D. B. Patterson Inorg. Chem. 1977 16 197. 49 H. R. Allcock D. B. Patterson and T. L. Evans J. Amer. Chem. Soc. 1977 99 6095. 50 H. R. Allcock R. W. Allen and J. P. O’Brien J. Amer. Chem. Soc. 1977.99 3984. R. H. Cragg J. D. Smith and G. E. Toogood (MeNH),P-N=P(NHMe), IlCV' I N-Pt -N I II (MeNH),P =N-P(NHMe) (46) tetramers [(MeNH),PN] and [Me2PNI4 react with potassium chloroplatinate(I1) to give the salts [(MeNH)8P4N4H2][PtC14] and [Me8P4N4H2][PtC14] both of which show an ti tumor activity .There is continued progress in elucidating details of substitution reactions of cyclic trimers and tetramers. For example the phosphazenes [Ar06-x(CF3CH20),P3N3] (Ar = 0-or p-NO&& p-C1C6H4 or Ph) may be obtained by the reaction between the trimers [(ArO),PN] and CF3CH20Na." The non-geminal compounds [(Ar0)3(CF3CH20)3P3N3] are formed in largest amounts. At 65 "C in dimethylformamide or tetrahydrofuran carbon-oxygen bond cleavage with formation of ethers ArOCH2CF3 competes with the nucleophilic substitution. Similar reactions leading to tetramers [(RNH) C18-,P4N4] R = Et or But have also been described.For R = Et seven compounds with n = 1,2 (two isomers) 3 4 (two isomers) and 8 have been isolated and characterized by 'H and 31 P n.m.r. spectroscopy. Non-geminal isomers are formed until n = 4. For R = But the products are with n = 1 2 (two isomers) 3 and 8; again non-geminal compounds are formed even though the trimer (Cl,PN) reacts with t-butylamine to give only geminal products. The relative proportions of various products formed in these substitution reactions depend on solvent temperature and precise reaction conditions in a way that is still incompletely under~tood.'~ The other series of compounds to be studied in detail has the formulae [(Me,N>,Cl,-,P,N,]. With n = 4 three of the four possible non-geminal isomers have been separated and the crystal structures of two the cis cis trans trans- and the cis cis,cis trans- isomers (47) have been determined.The most remarkable feature of compound (47) is the Me2N pMe2 ,P-N'=P CI" I1 I -*-a NN Me2N I II /c1 P=N -P.. a" ~ "Me2 (47) inequality of the two N'-P bond distances (162 and 148pm) even though the neighbouring phosphorus atoms carry the same sub~tituents.~~ Although a number of arsenic compounds (R2AsN) n = 3 or 4 have been made there has so far been little work on mixed compounds e.g. [R6AsnP3-,N3]. The reaction between the arsenic(v) chlorides R1R2AsCl3 and the linear phosphazene [(H,N)Ph,P=NPPh,(NH,)]CI at 20 "C,in the presence of triethylamine gives the H. R. Allcock and L. A. Smeltz J. Amer. Chem. SOC., 1976,98 4143.52 S. S. Krishnamurthy A. C. Sau A. R. Vasudeva Murthy R. Keat R. A. Shaw and M. Woods J.C.S. Dalton 1976 1405; 1977 1980. 53 M. J. Begley T. J. King and D. B. Sowerby J.C.S. Dalton 1977 149. The Typical Elements compounds (48; R'R2 = Ph2 PhCl or Me2) and the chloride [(Ph2P),N3AsPhC1] may be treated with sodium methoxide or dimethykmine to yield the derivatives (48; R'=Ph; R2=OMe or NMe2). The compounds have been characterized (48) spectroscopicallys4 and further structural studies should yield information about the effects of larger size and lower .rr-bonding potential of arsenic compared with phosphorus. 6 Difluorophosphino Derivatives Several compounds with the F2P-group bound to Main Group elements have been characterized by a wide variety of spectroscopic methods.The ether (F2P)20 has been known for about 12 years and a convenient route has now been found to the sulphur and selenium analogue^^^^^^ [equation (17)]. E(SnBu3)2+2PBrF2 -+ E(PF2)2+2SnBrBu3 (17) E = S or Se (49) Similar reactions using E(SiH3) in place of E(S~BL~)~ were slower. With protonic species the compounds (49) react with P-E bond cleavage [equation (IS)]. E(PF2)2+HX -+ PF2HE +PF2X(X = C1 Br NH2 OMe or SMe) (18) (49) When X = SH or SeH the phosphorus(II1) product is not observed presumably because it undergoes further reaction. The reaction between E(PF2)2 and chlorine is more complicated; both fragments are converted into phosphorus(v) compounds PC13F2 PC1F2E and further rearrangements may follow.The compounds (49) also behave as donors e.g. in the complexes (F2P)2EBH3 and [(F2P),EMo(CO),]. The corresponding nitrogen derivatives N(PF2) H3- were made earlier by the gas phase reactions between ammonia and difluorohalogenophosphines. The compound F2PNH2 was obtained in good yield but considerable experimental skill was required to isolate pure N(PF2)2H and N(PF2)3.57 The primary and secondary difluorophosphinoamines have been used to make a series of germylamines which are more stable thermally than trigermylamine.58 The compound N(GeH3)(PF2)2 [equation (19)] is unchanged in the gas phase after 10 h at 32 "C; NH(GeH3)PF2 made similarly is decomposed in ca. 10 min and NH(PF2)2+GeH31+NMe3 -+ N(GeH3)(PF2)2+[NHMe3]I (19) 54 D. B.Sowerby and R. J. Tillott J.C.S.Dalton 1977 455. 55 G.N. Bockerman and R. W. Parry J. Fluorine Chern. 1976 7 1. 56 D. E. J. Arnold E. R. Cromie and D. W. H. Rankin J.C.S. Dalron 1977 1999. '' D. E. J. Arnold and D. W. H. Rankin J.C.S. Dalton 1976 1130; 1975 889. 5R E. A. V. Ebsworth D. W. H. Rankin and J. G. Wright J.C.S. Dalton 1977 2348. R. H. Cragg J. D. Smith,and G. E. Toogood N(GeH3)2PF2 has intermediate stability. In contrast undiluted N(GeH3)3 decom- poses within a few minutes with formation of GeH2 and so cannot be condensed or revapourized. The difluorophosphinogermylamines may be stabilized by intramolecular H-F interactions and this suggestion is supported by an analysis of n.m.r. parameters. Pseudohalides F2PX (X = CN NCO NCS or NCSe) have also been made and details of low-frequency vibrations have been deduced from i.r.spectra. As with silyl and germyl pseudohalides H3EX (E=Si or Ge) these are important in the interpretation of electron diffraction data (c.f. Section 1 above).s9 The reaction between aminodifluorophosphine and ammonia [equation (20)] is formally an oxidative addition. The product diaminodifluorophosphorane (50) has been characterized spectro- scopicallys7 and by electron diff raction.60 As expected the molecule is trigonal bipyramidal with axial fluorine atoms [P-N 164.0(5)pm]. The configuration of the P(NH2) groups cannot be found from the electron diffraction data alone but CNDO/2 calculations suggest that the P(NH2) system is planar and perpendicular to the equatorial plane of the trigonal bipyramid (50).Interactions between fluorine and hydrogen account for the observation that rotation of the NH2 groups is slowed on the n.m.r. timescale at -40 "C. Me I 7 Monomeric Compounds of Gel' and Sn" Most simple derivatives of germanium(I1) or tin(11) are oligomeric e.g. [Sn(NMe2)2]2or [SnF2I4,or form structures with extended frameworks e.g. Ge12 or SnC12. Monomeric species are rare. It is found however that molecular compounds EL2 (E = Ge or Sn) may be obtained e-ither when L is a chelate ligand or when L is very large. A variety of chelate structures have been reported but these may be illustrated by the compound Ge(acac) obtained from Ge12 and Na(acac) (acac =pent-2,4-dionato) in refluxing tetrahydrofuran,61 and [Sn(S2COMe)2](51)made from SnC1 and potassium 0-methyl dithiocarbonate.62 59 S.Cradock E. A. V. Ebsworth M. L. McConnell D. W. H. Rankin and M. R. Todd J.C.S. Dalton 1977,1925. " D. E. J. Arnold D. W. H. Rankin and G. Robinet J.C.S. Dalton 1977 585. 61 A. Rodgers and S. R. Stobart J.C.S. Chem. Comm. 1976 52. 62 P. F. R. Ewings P. G. Harrison and T. J. King J.C.S. Dalton 1976 1399. The Typical Elements 157 An X-ray study of compound (51) shows that the co-ordination of tin is trigonal bipyramidal with the lone pair occupying an equatorial position. The amides M(NR'R2)2 (R' =SiMe2 R2=But or R' =R2 =SiMe2 M =Ge Sn or Pb; R' = R2=GeMe3 SiEt, or GePh, :M=Ge or Sn) have been made from metal(I1) halides and lithium amide~.~~ The compounds are diamagnetic monomers soluble in hydrocarbons and a wide range of physical evidence suggests that the structures are bent (52),with singlet ground states.The amides are pale yellow to red; the R 'R*N \ /Sn: R'R~N (52) colour becomes deeper on heating and almost disappears at -196"C and the moderate extinction coefficients for the bands in the visible suggest the presence of thermally accessible singlet states. When the amides are irradiated in the cavity of an e.s.r. spectrometer signals characteristic of E"' radicals are observed except for the most bulky amides when the aminyl radical is observed. Half lives range from more than 5 months for Ge{N(SiMe3)2}3 to about 5 min for S~{NBU'(S~M~~)}~. 8 Trifluoromethyl Derivatives of Nitrogen and Phosphorus The chemistry of the trifluoromethyl compounds of phosphorus and arsenic was first explored about 25 years ago.The nitrogen derivatives have been studied more recently.64 Two routes have now been found to the very simple compound trifluoromethylamine (53).65 One involves elimination of chlorine from the +2CIF +3HCI/-10OoC CF3-N=SF2 dCF3NC12-2~12 CF3NH2*HClB2 CF3NH2 (21) -SF4 (54) (53) (B =quinoline pyridine Et,NH or Me3N) dichloro compound (54). This reaction which is similar to that used for the preparation of trifluoromethanol (from CF30Cl+HC1 at -120 "C) involves combination of positively and negatively charged chlorine atoms. The second reaction [equation (22)] involves careful cleavage of t-butyl trifluoromethyl- carbamate (55).CF3NC0+HOBU' + CF3NHC02BuL HC1/-25 "C CF3NH2'HCl (22) -c02 -B"'C Trifluoromethylamine is a thermally unstable colourless solid which decomposes easily with elimination of hydrogen fluoride. It may be sublimed with only partial decomposition and is less volatile (extrapolated b.p. 50 "C)than methylamine; in 61 M. J. S. Gynane D. H. Harris M. F. Lappert P. P. Power P. Rivikre and M. Riviere-Baudet J.C.S. Dalton 1977 2004. 64 H. G. Ang and Y. C. Syn Ado. Inorg. Chem. Radiochem. 1974,16 1. 65 G. Kloter W. Lutz K. Seppeit and W. Sundermeyer Angew. Chem. Internat. Edn. 1977 16 322 707. R. H. Cragg J. D. Smith,and G. E. Toogood contrast CF30H is more volatile than CH30H. Spectroscopic evidence appears to exclude intermolecular hydrogen bonding.As expected CF3NH2 is a weaker base than methylamine. Bis(trifluoromethy1)amine is available from the reaction sequence of Equation (23) 1 CF3NO+C2F4+ CF3NOCF2CF24CF3N:CF2 HF (CF,),NH (23) Salts containing the (CF3)2N- anion may be made by treatment of (CF3)2NH with triethylamine or from the reaction between CF3N:CF2 and ceasium fluoride. The presence of the anion is shown by reactions with ethyl bromoacetate or 3-bromo- propene.66 ,=b (CF3)2NCH2C02Et CF,N=CF2 CsF Cs[(CF3)2N] I 1 The molecular structure of tris(trifluoromethy1)phosphine has been redetermined by electron diffraction and earlier results have been ~onfirmed.~’ The P-C bond length [190.4(7)pm] is slightly longer than in trimethylphosphine [184.6(3)pm] suggesting that the C-F bond may be strengthened at the expense of the adjacent P-C bond.The C-P-C angle in (CF3)3P is 97.2(7)0 compared with 98.6(3)0 in Me3P. The silicon analogue (SiF3),P has been made in 84% yield by the mercury- sensitized cophotolysis of Si2F6 with PF3 [equation (25)].68 3Si2F6+PF3 + 3SiF4+(SiF3)3P (25) The colourless air-sensitive product is much more reactive than (CF3)3P. Thus it is hydrolysed by water vapour to PH3 Si20F6 and SiF whereas (CF3)3P is not attacked by water. Bromine readily cleaves the Si-P bonds at room temperature to give PBr5 and SiF3Br; C-P bonds in (CF3)3P are cleaved by bromine only on heating. The methylation of the readily available trifluoromethyl-substituted chloro-phosporanes by methyl-lithium is complicated by ylide formation or fluoride abs- traction.The less reactive reagents Me,Sn or Me4Pb give much cleaner products and a number of phosphoranes Me(CF3)3PX (X = F C1 OMe SMe or NMe2) have been ~btained.~~ The stereochemistry and fluxional behaviour are established by ‘H 19F and 31Pn.m.r. spectroscopy. When X = F or C1 the halogen atoms occupy axial positions in the trigonal bipyramidal array round phosphorus and the methyl groups occupy equatorial positions. There are thus one axial and two equatorial trifluoromethyl groups. When X = OMe SMe or NMe2 methyl and X substituents occupy equatorial places in the trigonal bipyramid so that one CF group is 66 A. F. Gontar 8. G. Bykhovskaya and I. L. Knunyants Izuesr Akad Narc&,S.S.S.R.,Ser.Khim 1975 24,2279; 1976,25,212. 67 C. J. Marsden and L. S. Bartell Inorg. Chem. 1976 15 2713. K. G. Sharp J.C.S. Chem. Comm. 1977 564. 6q K. I. The and R. G. Cavell Inorg. Chem. 1977,16 2887 1463. The Typical Elements 159 equatorial and the other two axial. Except for the dimethylamino derivative Me(CF3),PNMe2 the CF,-environments are averaged on the n.m.r. timescale at 33 "C but distinguished on cooling coalescence temperatures decrease in the series NMe2> SMe > OMe >> F> C1. The hydrolyses follow equations (26) in neutral solution and (27) in alkali. Me(CF3),PX + 2H20 -D 2CF3H+ Me(CF3)P02-+ H++ HX (26) Me(CF3),PX + H20+ 20H-D 3CF3H+ MeP032-+ HX Methylbis(trifluoromethy1)phosphoranes Me(CF3),PXY may be made by methyl- ation of (CF3)2PF3 or from the dichlorophosphorane Me(CF3)2PC12 itself made by chlorination of Me(CF3)2P.Where X = Y = halogen the low temperature struc- tures determined by n.m.r. indicate that halogen atoms occupy axial positions but in the compounds CH3(CF3)2P(F)Y (Y = OMe or NMe2) one trifluoromethyl group occupies an axial and the other an equatorial position. All these structures may be rationalized on the usual basis that in trigonal bipyramidal structures electro- negative groups preferentially occupy axial positions. Part IV Groups VI-VIII By R. H. Cragg 1 GroupVI Two areas still continue to dominate (i) the synthesis of crown ethers and their complexes and (ii) the chemistry and physical properties of S-N systems. The synthesis of 2,3,11,12-tetraphenyl-1,4,7,10,13,16-hexaoxa-cyc1o-octadeca-2,ll-diene the first macrocyclic crown ether containing a double bond in the ring starting from benzoin has been reported.' On refluxing (1) and (2) in boiling benzene with 50% aqueous NaOH as base and Bun4NBr as a phase-transfer catalyst the crown ether (3) is obtained.Ph I 0 Ph Ph 07 CH-OH I + c=o 'Ph -+ Ph'O,)) I Ph A study of the binding of ammonia and amines with crown-ether phenols has shown that the phenolic group is strongly influenced by the encircling oxygen atoms of the crown and the binding of ammonia and amines.' Complexes have been prepared of 18-crown-6 with hydrated halides of Mn" Co" and Ni" from acetone solution^.^ In the case of both Mn and Ni hydrated complexes were obtained ' A.Men Angew. Chem. Internat. Edn. 1977 16 467. M. A. McKervey and D. L. Mulhollan J.C.S. Chem. Comm. 1977,438. M. E. Farago Inorg. Chim. Acra 1977 25 71. R. H. Cragg J. D. Smith and G.E. Toogood which appear to undergo a rearrangement from square planar or octahedral to tetrahedral on heating. Benzo- 15-crown-5 and dibromobenzo- 15-crown-5 form complexes with CoC12 and MnBr2 in which the co-ordinated metal appears to be in a tetrahedral environment. Thallium(1) complexes with macrocyclic crown ethers have also been rep~rted.~ Complexes of T1' salts and benzo-15-crown-5 or 43-dibromobenzo- 15-crown-5 appear to have a sandwich structure where two mole- cules of the ligand are co-ordinated to one thallium ion. In contrast the complexes with 18-crown-6 ether have a 1 1crown to metal ratio.Measurements of the enthalpy of transfer of cryptand-2,2,2 ([222]) and of the Na' and K+ cryptates have been combined with heats of complex formation in water to yield by a thermochemical cycle the heats of complex formation in methanol.' The values obtained for the enthalpies of transfer from water to methanol (kJ mol-' at 298 K) are 58.2 (cryptand-2,2,2) 23.4 (Na' [222k]) and 18.0 (K',[222]) respectively. Ab initio MO calculations using an STO-3G basis set have been carried out for cyclohexasulphur.6 The results obtained are consistent with the molecular mechanics calculations in indicating that the chair and twist forms are two stable conformations with the chair form more stable by ca.63 kJmol-' while the boat (C2") form is a twist-twist rotational transition state. The relative stability of ten possible electronic and conformational states of S4 has been studied by ab initio STO-3G and 44-31G calculation^.^ Molecular mechanics calculations suggest that cyclotetrasulphur is unlikely to exist because it has a strain energy almost ten times that for cyclohexasulphur and an S-S bond energy ca. 50 kJ mol-' smaller than that for cyclohexasulphur. The S4 molecule is known to exist in both the gas and liquid phases and several possible structures have been investigated. In the gas phase the triplet helix biradical (4) is favoured over the other forms and the bent S4 is the next most stable state. 203.2 pm \<104" / q5 = 90" 208.4 pm (4) Red monoclinic prisms shown by X-ray structure analysis to be an allotrope of cyclo-octaselenium but different from LY and /3 monoclinic selenium crystallized from a solution of dipiperidinotetraselane in CS2.8 The dipiperidinotetraselane (the X-ray structure is also reported) was obtained from the reaction of black selenium powder and the amir~e.~ The electronic structures of SF2 SF4 and SF6 have been investigated from the standpoint of ab initio generalized valence bond (GVB) calculations.lo Analysis of the GVB orbitals in these molecules and in the model reaction SF4-+ SF2+ 2F shows that the stability of the SF and SF6systems is largely due to the incorpora- M. E. Farago Znorg. Chim. Acta 1977,23 2 11. M. H. Abraham A. F. D. Namor and W.H. Lee J.C.S. Chem. Comm. 1977 893. J. Kao and N. L. Allinger Znorg. Chem. 1977 16 35. ' J. Kao Znorg. Chem. 1977 16 2085. 0. Foss and V. Janickis J.C.S. Chem. Comm. 1977 834. 0. Foss and V. Janickis J.C.S. Chem. Comm. 1977 833. lo P. J. Hey J. Amer. Chem. SOC.,1977.99 1003. The Typical Elements tion of charge-transfer configurations 3d functions on the sulphur playing a lesser role. The excited electronic states of SF6 and the lowest states of SF6' have been investigated by ab initio calculations in an extended Gaussian basis." Two new i.r. absorption bands at 811.6 and 552.1 cm-' have been observed in the spectra of the products of vacuum-u.v. photolysis at 104.8 and 106.7 nm of SF6 and its derivatives SF,X (X = C1 Br or SF,) in dilute argon matrices at 8 K and are assigned to the SF radical a new species of C4vsymmetry.I2 A new absorption at 594 cm-' generated by charge-transfer processes involving the photoionization of NO Na and K in the presence of matrix-isolated SF6 has been tentatively assigned to the SF6- ion.' The same absorption was also obser- ved to result from the photolysis of SF6 with the 104.8 and 106.6nm resonance lines of argon.The failure of SF6 to react with nucleophilic reagents in contrast to the high reactivity of SF has been suggested to be due to kinetic rather than to ther- modynamic factors. It is therefore of considerable interest to note that pentafluorosulphur halides SFSX (X = C1 or Br) react with Me3SiNMe2 at -78 "C to form Me2NSF4X that CF,SF,Cl undergoes defluorination and reduction to form CF3SC1(NMe2)2 at 25 or -78"C,14 and that in both reactions the S-X bond is retained.The solvated entites Se," and Seg2' have been identified in reaction mixtures of dilute SeC1 and elemental selenium by a potentiometric method supplemented by a spectrophotometric method,15 the solvent being low-melting NaC1-AlCl3 (37 :65 mole%) at 150"C. The results obtained also indicate the possible ions Se22' Se1221 and SeIG2' in solution. It has been observed that SeC1 and SeBr4 form Sex3+ in contrast to Se2C12 or Se2Br2 which form Se,*' and Sex,' when dissolved in disulphuric acid.I6 In acid containing chlorine and bromine selenium forms Sex3+ and Se2C12 forms Se2C13' and Se2C12Br' but Se2Br2 forms Se2Br3+ with Se2X3 disproportionating to give Sex3' and Se.The results of a I2'Te Mossbauer investigation of frozen solutions of Te in HS0,F and in oleum have been reported together with data for solid samples which may have contained Te42C Te,"' and Te62' cations." Quadrupole splitting reflects the sequential oxidation of Te through the stages Te + Te4,' + Te,"' -+ Te" + Te'" + Te02 Ammonium tetrasulphurpentanitride,[NH4][S4N5],has been obtained from the reaction of S2C12 SCl, SCl,(ClSN), and S4N4 with NH3.I8 In order to provide some understanding of the initial stage of the solid-state polymerization of (SN), semi-empirical INDO-type ASMO-SCF calculations have been carried out for the precursor S2N2 several deformed structures of (SN), the dimeric unit (SN), and the trimeric unit (SN)6.19 The results suggest that the triplet l1 P.J. Hey J. Amer. Chem. Soc. 1977 99 1013. R. R. Smardzewski and W. B. Fox J. Chem. Phys. 1977.67 2309. l3 J. E. Barfield and W. A. Guillory J. Phys. Chem. 1977 81 634. l4 T. Kitazume and J. M. Shreeve J. Amer. Chem. SOC. 1977,99 3690. lS R. Fehrmann and N. J. Bjerrum Inorg. Chem. 1977,16,2089. l6 A. Bali and K. C. Malhotra J. Inorg. Nuclear Chem. 1977 39 957. l7 C. H. W. Jones Canad. J. Chem. 1977,55 3076. 0.S. Scherer and G. Wolmershauser Chem. Ber. 1977 110 3241. l9 T. Yamabe K. Tanaka K. Fukui and H. Kato J. Phys. Chem. 1977,81,727. 162 R. H. Cragg J. D. Smith and G. E. Toogood biradical nature is favoured in appropriately deformed structures of (SN),.This is in agreement with the experimental findings wherein the paramagnetism (g = 2.005) is observed at the initial stage of polymerization.! In contrast (SN) and (SN) show no triplet nature probably corresponding to the fact that the system gradually becomes diamagnetic as the polymerization proceeds. In an examination of the electronic states of the intermediates involved in the polymerization of (SN), the deformed (SN), it was shown that the triplet biradical nature begins to emerge as one of the S-N bonds opens; this corresponds to the interaction of the opened S-N 0-bond with the adjacent S-N p,-bond in the molecular plane. Mass spectrometric techniques have been used to study the vapour-phase species obtained by heating (SN) at 100-160 OC.,' Field ionization and field desorption mass spectra of (SN) show that (SN) vaporizes as a (SN) species having a structure different from the cradle-like structure of S4N4,and this may be linear.The heat of sublimation of (SN) is 121.3~t2.1kJ mol-'. Crystals of the anisotropic metallic conductor polymeric (SN) employed as electrodes with and without depolarizers present yield well defined cyclic voltammograms.21 They are found to be similar to those obtained with other solid electrodes but have unusual electrode process characteristics which depend upon the crystal orientation. A new phase of (SN), with the same chain geometry as the monoclinic form but having ortho- rhombic symmetry and very different interchain interactions has been produced by a martensitic-like phase transformation.22 The phase transformation results from mechanical shear on planes which are parallel to the chain axis direction which can be very easily produced by grinding phase I polymer in a mortar and pestle.There has been considerable interest in improving the metallic characteristics of (SN) polymers. For example the electrical properties of (SN) have been modified by reaction with Br, ICl or On bromination blue-black fibrous crystals of composition (SNBro,,) are obtained and the crystals supercooled with T ca. 0.31 K. This suggests that the (SN) chain structure remains substantially unchanged after bromination. Similar products which are highly conducting solids have also been obtained from the interaction of Br, ICl IBr or I2 with S4N4.24 For example the reaction with bromine yields a product (SNBro.,) This compound which is prepared by heating the reagents at 80 "C for 4 h undergoes a change on heating to give a metallic-like compound of composition (SNBT~.,~),.~' Crystal and films of (SNBr',,), prepared by the method outlined in the previous paper have been characterized.26 The visible and i.r.reflectance spectra and room-tempera- ture conductivity were found to be comistent with metallic behaviour. An X-ray crystallographic study has shown that the new cyclic cation [S4N4I2' has different structures in the compounds [S4N4][SbC16]2 and [S4N4][SbF6]* 'O R. D. Smith J. R. Wyatt J. J. De Corpo F. E. Saalfeld M. J. Moran and A. G. MacDiarmid J. Amer. Chern. SOC.,1977,99 1726.'' R. J. Nowak H. B. Mark A. G. MacDiarmid and D. Weber J.C.S. Chern. Cornrn. 1977,9. 22 R. H. Baughman P. A. Apgar R. R. Chance A. D. MacDiarmid and A. F. Garito J.C.S. Chem. Cornm. 1977,49. 23 G. B. Street W. D. Gill R. H. Geiss R. L. Green and J. J. Mayerle J.C.S. Chem. Comm. 1977 407. 24 G. B. Street R. L. Bingham J. I. Crowley and J. Kuyper J.C.S. Chem. Comm. 1977,464. 25 M. Akhtar C. J. Chiang A. J. Heeger and A G. MacDiarmid J.C.S. Chem. Cornm. 1977 846. 26 M. Akhtar J. Kleppinger A. G. MacDiarrnid J. Milliken M. J. Moran C. K. Chiang M. J. Cohen H. Heeger and D. L. Peebles J.C.S. Chem. Cornrn. 1977,473. The Typical Elements [Sb3F14].27 The complexes were readily obtained from the reactions of S4N4 or S3N,Cl with SbCls and S4N4 with SbF6 S4N4+3SbC15 % [S4N4][SbC16]2 +SbC13 The mass and i.r.spectra have been investigated for the complexes formed between S4N4 and MC12 (M=Ni Co or Pd) in methaol solution.3* The mass spectra indicate that the stoicheiometry of the complexes is M(S2N2H)* and the i.r. spectra confirm the presence of NH groups and suggest a cis-planar configuration for the S2N2H chelate groups (5). The reaction of caesium or tetra-alkylammonium azides with S4N4 in ethanol produced salts of the S3N3- anion which are explosive under the influence of heat or pressure.29 On the basis of vibrational spectra the S3N3- anion was assigned a cyclic structure. The crystal structure of Ph3As,S3N4 has been solved from diffractometer X-ray single-crystal data.30 Three new S4N4 complexes of Ag’ namely Ag[C104],S4N4H4 (6) 2Ag[C104],S4N4H4(7) and Ag[C104],2S4N4H4 (8) have been prepared from acetone solutions.31 The mass spectra of (6) and (8) indicate metal-ion-S4N4H4 ring complexation and i.r.data provide evidence for the existence of bidentate perchlorate bridging -0C1020- between the solid state as indicated in (9). The interaction of N2S4 and dicyclopentadiene has been reported to result in the formation of a 1 1 complex.32 On the basis of i.r. spectra structure (10) has been proposed. H H /N-s-N / s\ S N-S-N O\ HI /0 0-Cl-0-0/ Ag -0-CI-0\0 s / N-S-N H H S (9) (10) 27 R. J. Gillespie D. R. Slim and J. D. Tyrer J.C.S. Chem. Comm. 1977 253. 28 I. S. Butler and T. Sawai Cunad. J. Chem.1977 22 3828. 29 J. Bojes and T. Chivers J.C.S. Chem. Comm. 1977 453. 30 E. M. Holt S. L. Holt and K. J. Watson J.C.S. Dalton 1977 514. 31 S. N. Nabi J.C.S. Dalton 1977 1152. ” R. R. Adkins and A. G. Turner Inorg. Chim. Acta. 1977 25 233. R. H. Cragg J. D. Smith and G.E. Toogood The photoelectron spectrum of S2N2 has been determined and tentatively assign- ed.33 Raman i.r. and far-i.r. spectra of solid and dissolved (in CS,) S7NH S7ND S715NH and S715NDhave been All fundamental frequencies for S7NH were observed and assigned in accordance with C molecular symmetry. The crystal and molecular structure of cycloheptas~lphur~~ and S7036have also been reported. The first optically active sulphurane (+)-1-chloro-3,3-dimethyl-1-phenyl-[3H-2,1-benzoxathiole](11) has been synthesized with known absolute configuration in 95% optical purity.37 Reaction with water in the presence of NN-dimethylaniline is rapid and leads to retention of configuration about sulphur.Me Me (11) CS has been shown to undergo direct fluorination at low temperature the product of the reaction being characterized by mass and 19F n.m.r. ~pectra.~' F2-He (60%) cs2 -120°C + F3SCF2SF3 Another area which continues to be of interest is that of organic metals. The synthesis and properties of p-phenylenebistetrathiafulvalane (12) in which the (12) molecule contains two donor units has been described.39 It is of interest as a source of charge-transfer complexes with a band structure higher than the quasi- one-dimensional structure characteristic of all presently known electrically conducting C.T.complexes. For example with TCNQ (7,7,8,8-tetra-cyanoquinodimethane) a 1 1 complex is obtained whose room-temperature resis- tivity is ca. 10 times that of TCNQ-TTF. Monosubstituted TTF (tetra-thiafulvalene) and their selenium analogues [e.g.(13)] have been prepared by the formation of tetrathiafulvalenyl-lithium and subsequent reaction with C02 or tri- alkyloxonium This is in contrast to direct substitution methods such as chlorination which have yielded only the radical-cation salts or dication salts. The 33 D. C. Frost M. R. Le Geyt N. L. Paddock and N. P. C. Westwood J.C.S. Chem. Comm. 1977,217. 34 R. Steudel J. Phys. Chem. 1977,81 343. 35 R. Steudel R. Reinhardt and F. Schuster Angew.Chem. Internat. Edn. 1977 16 715. 36 R. Steudel K. Reinhardt and T. Sandow Angew. Chem. Internat. Edn. 1977 16 716. 37 J. C. Martin and T. M. Balthazor J. Amer. Chem. SOC.,1977,99 153. 38 L. A. Shimp R. J. Lagon Znorg. Chem. 1977 11,2974. 39 M. L. Kaplan R. C. Hadden and F. Weidl J.C.S. Chem. Comm. 1977 388. 40 D. C. Green J.C.S. Chem. Comm. 1977 161. The Typical Elements 'S S CO,H (andOEt) (13) (also the Se compounds) presence of one carboxylate function appears not to detract from its electron-donor properties and may allow it to be attached to other molecular systems in order to impart its unique donating properties to such systems. The replacement of the sulphur atoms in the charge-transfer salt TTF-2,5- diethyltetracyano-p-quinodimethanewith selenium yielded the isostructural and better conducting analogue in which the Peierls transition temperature is decreased.41 Finally the new organic r-donor (14)has been prepared and shown to react with TCNQ to give two semiconducting a new general synthesis of substituted (14) diselenadithiafulvalanes and tetraselenafulvalene (TSeF) from readily available starting materials has been developed,43 and TTF (15) has been prepared directly from cs2.44 2 GroupVII The products of the reaction of alkali fluoride or chloride molecules with F2 ClF or C12 have been investigated in argon matrices at 15 K,45 and assignments to MX3 species (Table 1) have been reported.A strong i.r. band at 550 cm-' and a Raman band at 461 cm-' observed in the spectra of the products of the reaction of CsF and RbF with F2 were assigned to v3 and v1 respectively for the F3-anion in MfF3-.The formation of K'F3-and the failure to isolate Na'F3- suggest that the F3- species is relatively unstable and requires a large counter-cation for stabiliza- tion in an argon matrix. 41 J. R. Anderson R. A. Craven J. E. Weidenborner and E. M. Engler J.C.S. Chem. Comm.,1977,526. 42 D. J. Sandman A. P. Fisher T. J. Holmes and A. J. Epstein J.C.S. Chem. Comm. 1977,687. 43 P. Shu A. N. Block T. F. Carruthers and D. 0.Cowan J.C.S. Chem. Comm. 1977,505. 44 W. P. King A. N. Block and D. 0.Cowan J.C.S. Chem. Comm. 1977,660. 45 B. S. Ault and L. Andrews Znorg. Chem. 1977 16 2024. R. H. Cragg J. D. Smith and G.E. Toogood Table 1 I.r. absorption wavenumbers (cm-') assigned to M'X3-species derived from F2 ClF or C12 in solid argon at 15 K Cs' Rb+ K+ Na' F3- 550 550 549 - FClF- 566 565 571 589 ClFF 365 371 - - ClFCl 474 480 486 511 ClClF 412 409 - - High yields of alkali fluoride or chloride monomers and dimers alkali trihalide species and in the case of the heavy alkali metals a stable M'ClF- species were obtained from the alkali-metal ClF matrix reaction,46 the species being trapped and identified by i.r. Raman and U.V. spectroscopy. The strong bands observed in both i.r. and Raman near 340cm-' were assigned to the interatomic (Cl-F)- mode and strong metal-dependent i.r. absorptions were attributed to an interionic M't (C1F)- mode. The intense U.V.bands near 290 and 250nm were assigned to M'ClF-and MfC12F-. The reaction between Cs[IC12] with aqueous Ag[N03] saturated with Ag[I03] is both rapid and q~antitative.~~ Based upon enthalpies of reaction with aqueous ApTN03] the standard enthalpies of formation of crystalline Cs[ICl,] and Rb[IC12] ha:.. 5een determined as -5 11.1 *4.1 and -491 k4.1 kJ mol-' respectively. As part of a MO study of electron donor-acceptor complexes the energy and charge distribution decomposition analysis have been carried out for a series of donor-acceptor complexes involving halogen^.^' Based on energy components for binding the following classifications were proposed H3N-F2 H3N-C12 and H2CO-F2 form weak electrostatic charge-transfer complexes HF-C1 is a weak electrostatic complex and Fz-F2 is a very weak dispersion-charge-transfer complex.A single-crystal X-ray study on IF302 shows that the compound exists as a dimer in the crystalline state and is centrosymmetric with oxygen bridges and trans- equatorial oxygen^.^^ The same compound has been studied in the vapour phase using the combined techniques of electric deflection and mass spectrometry of a modulated molecular beam." Flight-time distributions show that the molecule is essentially dimeric at room temperature and using flight-time distributions the decomposition of the dimer to monomer has been followed. The decomposition becomes significant at 100"C and is almost complete at 185 "C. Electric deflection measurements show that the monomer has a polar structure whereas the dimer is non-polar and therefore must be symmetric.The 19Fn.m.r. and Raman spectra of adducts of IO2F3 with AsF, SbFS NbFS or TaF and with IFS or IOF3 have been measured." The spectra demonstrate that the adducts are oxygen-bridged polymers of the types (I02F4MF4) and (I02F410F2),. The adducts of I02F3 and Lewis acid pentafluorides have polymeric 46 E. S. Prochaska B. S. Ault and L. Andrews Znorg. Chem. 1977,16 2021. 47 A. Finch P. N. Gates and S. J. Peake J.C.S. Dalton 1977 397. 48 H. Umeyama K. Morokuma and S. Yamabe J. Amer. Chem. SOC.,1977,99 330. 49 L. E. Smart J.C.S. Chem. Comm. 1977 519. '' M. J. Vasile W. F. Falconer F. A. Stevie and I. R. Beattie J.C.S. Dalton 1977 1233. 51 R. J. Gillespie and J. P. Krasznai Znorg.Chem. 1977 16 1384. The Typical Elements structures containing I02F4groups alternating with 02MF4 groups. On the basis of n.m.r. studies in the molten and solution state it is suggested that both O2MF4 and I02F4 units exhibit cis-trans isomerism with the cis-isomers favoured. Ab initio LCAO-MO-SCF calculations indicate that the hydrogen bond between the fluoride anion and formic acid is stronger than that in the bifluoride anion previously believed to represent the upper limit of hydrogen-bond strength (calc. 250 kJ mol-'; cf. 220 kJ m~l-l),~~ and the calculated hydration energies of halide (F C1 or Br) ions for symmetrical hydration structures containing 1-6 waters of hydration have been rep~rted.'~ The reaction of dimethylchloramine with an excess of iodomethane at room temperature resulted in the formation of a crystalline compound which was shown by low-temperature (-60 "C) X-ray analysis to contain the [Me4N]' cation and the novel [Me2N(IC1),]- anion (16) which is analogous to the [I5]- ion with NMe2 as a pseudohalide.The complex decomposes ina few hours at 20 TS4 Me Me 3 Group VIII Xenon tetrafluoride has been observed to form an intercalate with graphite on interaction over a period of three weeks." The reaction was generally completed after 10 days and the compound could be easily handled outside a vacuum line although it partially hydrolyses giving hydrogen fluoride. The residue was surpris- ingly stable neither being shock-sensitive nor exhibiting appreciable oxidizing power.In preliminary studies to evaluate its properties the compound was found to be a moderate fluorinating agent for the synthesis of organic fluorides. For example on reaction with benzene C6H5F (26%) and p-C6H4F2 (11%) were obtained. This is in contrast to the fact that XeF2 fluorinates benzene only in the presence of HF as catalyst. Krypton difluoride also reacts directly with graphite to form intercalates which have been studied by 19Fwide-line n.m.r. and are of variable composition with the general formula C,KrF (n>2).56' It appears that the reaction is a balance between oxidative fluorination and intercalation with fluorination most likely dominant at edges and internal defects. Xenon difluoride dissolved in anhydrous HF is a moderately strong oxidant which can be used selectively in many situations to prepare a wide range of fluorine-containing compound^.^' Binary and complex 52 J.Emsley 0.P. A. Hoyte and R. E. Overill J.C.S. Chem. Comm. 1977 225. 53 K. J. Spears J. Phys. Chern. 1977 81 186. N. W. Alcock S. Esperas J. F. Sawyer N. D. Cowan C. J. Ludman and T. C. Waddington J.C.S. Chem. Comm. 1977,403. 55 H. Salig M. Rabinovitz I. Agranat C. H. Lin and L. Ebert J. Amer. Chem. Soc. 1977,99 953. 56 H. Selig P. K. Gallagher and L. B. Ebert Inorg. Nuclear Chem. Letters 1977,13 427. " R. C. Burns I. D MacLeod T. A. O'Donnell T. E. Peel K. A. Phillips and A. B. Waugh J. Inorg. Nuclear Chem. 1977,39 1737. R. H. Cragg J. D. Smith and G. E. Toogood fluorides oxide fluorides and carbonyl fluorides have been isolated from the reactions of XeF2 with elements lower fluorides other halides or halogeno-complexes oxides or carbonyls.Some indication of the range of reactions can be seen from the examples given in Table 2. Table 2 Products of reaction with XeF2 in HF Reactants Products Reactants Products Elements s8 Mo SF6 MOF6 Hg0s Hg2F2 +HgF2OsF5 and OsF6 -+ OsF5 adducts Lower fluorides TlF TIF3 COF~ CoF3 AsF~ AsF~ MoF5 MoF~ Halides and halo-complexes MoBr4 K[WF~IOxides Re02 K2[RhF61RuF~+RuF~-B RuF~ UF4 and Uo2F2 Ru02 MOF6 Carbon yls [Ru(C0)3C121 The preparations of XeOF2 Cs[XeOF3] and Cs[Xe02F3] have been reported and information concerning their structures has been deduced from their Raman spectra.58 XeOF2 was obtained from the interaction of XeF and water in HF solvent and the structure proposed (17) is consistent with its Raman spectrum.On addition of CsF at low temperatures and slowly warming to 0°C while pumping Cs[XeOF3] was obtained. By altering the conditions Cs[Xe02F,] was also isolated. Their proposed structures (18) and (19),were based upon Raman studies. .F F FFO -\I i' I i' \xk // Xe F-Xe=O \.. //I OF *.' // OF 1 58 R. J. Gillespie and G. J. Schrobilgen J.C.S. Chem. Comrn. 1977 595.

ISSN:0308-6003    

DOI:10.1039/PR9777400111

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

8 Chemistry of the d-and f-Block Metals By J. R. DILWORTH G. J. LEIGH and R. L. RICHARDS ARC Unit of Nitrogen Fixation University of Sussex Brighton BN 1 9QJ K. W. BAGWALL Department of Chemistry University of Manchester Manchester M13 9PL PART I The Transition Metals By J. R. Dilworth G. J. Leigh and R. L. Richards 1 Introduction This year we have again concentrated on a selection of topics which we have judged to be particularly active and interesting. Inevitably some topics which were dis- cussed last year are again included and in particular sulphur ligands receive more detailed coverage. 2 Macrocyclic Ligands Within this wide area we have chosen to discuss in some detail metal porphyrin complexes complexes of dinucleating ligands and the influence of ring size on the properties of complex compounds of macrocyclic ligands.Metal-Porphyrin Complexes.-These compounds merit considerable attention because of their relevance to natural species such as chlorophyll haemoglobin myoglobin catalase peroxidase and the cytochromes. We will not cover the dioxygen-carrier properties of porphyrin complexes in detail since it was discussed last year but rather will concentrate on their general chemical and structural features. Structural aspects of metal-porphyrin complexes have been reviewed this year,’ and a recent book gives a more general coverage of their chemistry.’ We have used the following abbreviations PH =general porphyrin; TPP(R)H2 = apy&tetraphenylporphyrin with substituent R on phenyl ring or rings; OEPH2 = octaethylporphyrin; PP(IX)(DME)H2= protoporphyrin(XX)(dimethyl ester); TNMPP(3- or 4-)H2= tetra(3- or 4-N methylpyridy1)porphyrin; TPrPH = tetra-n-propylporphyrin; DP(DME)H2 = deuteroporphyrin(dimethy1 ester).Less com-monly used porphyrins will be named in the text as appropriate. Titanium. The complexes [TiX2(TPP)] (X = F C1 or Br) prepared by treatment of [TiO(TPP)] with the gaseous halogen acids in dichlorom,:thane easily revert to ’ W. R. Scheidt Accounts Chem. Res. 1977 10 339. ‘The Porphyrins’ ed. K. M. Smith Elsevier Amsterdam 1975. 169 J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall their oxide precursor on hydrolysis.3 Reduction of [TiF,(TPP)] at a rotating pla- tinum electrode in CH2C12-[Bu4N][PF6] gives a purple product formulated as [TiF(TPP)] which reacts readily with dioxygen to give a mixture of [Ti(O,)(TPP)] and [TiF2(TPP)].3 The series [TiO(TPPR)] (R=C1 OMe Me NEt, CF3 Pr’ or OH as para-substituent of one TPP phenyl group) has been prepared and the rate of rotation of the substituted phenyl group determined from its ‘H n.m.r.spectrum. The rate is greater when R is electron-donating than when it is electron-withdraw- ing but is not proportional to the upvalues of R. Relative rotation rates for these complexes and analogues with other metals are in the order [TiP(TPPR)]> [InCl(TPPR)]>[RU(TPPR)(CO>(BU‘C,H~N)].~ Niobium and Molybdenum. The complexes [M203(TPP)2] (M = Nb or Mo),’ although having the same empirical formula have been shown by X-ray crystallo- graphy to have quite different structures.The niobium complex (1) has a triple oxygen bridge between metal atoms whilst the molybdenum complex (2) has a 0 N-N-N-N ‘h6’ /\\ I 000 0 I \\ 1 Nb 0 (1) (2) linear O(Mo)O(Mo)O group.6 This structural difference is thought to reflect the different affinities of the metals for the porphyrin core. Thus the more strongly binding molybdenum atom is displaced by only 90pm from the mean nitrogen atom plane towards terminal oxide. This requires an energetically demanding radial expansion of the porphinato-core and a close intramolecular porphyrin interplanar spacing. In the niobium complex the much larger displacement of the more weakly held metal atom (101pm) towards oxygen allows a decreased metal- porphyrin interaction and less radial strain within the porphyrin ring.6 A similar large displacement of niobium has also been observed’ in [Nb(O)(CO,Me)(TPP)].It has been suggested6 that the complexes [M203(TPP)2] (M =W or Re)5 have the same structure as (2). Treatment of [MoOP] (P = TPP or OEP) with anhydrous HCI in benzene gives the complexes [MC12P]. Thz X-ray structure of [MoCl,(TPP)] (p = 2.9 B.M.) shows that although the molybdenum atom is in the plane of the porphyrin nitrogen atoms the MoCl distances differ (234.7 and 227.6 pm) thus far inexplicably.8 M. Nakajima J.-M. Latour and J.-C. Marchon J.C.S. Chem. Comm. 1977 763. S. S. Eaton and G. R. Eaton J. Amer. Chem. SOC.,1977,99 6594. B. Fleisher and T. Srivastava Inorg.Chim. Acta 1967 5 151; J. W. Buchler L. Puppe K. Rohbock and H. H. Schneehage Chem. Ber. 1973,106 2710. J. F. Johnson and W. R. Scheidt J. Amer. Chem. SOC.,1977,99 294. ’C. Lecomte J. Protas R. Juilard B. Fliniaux and P. Fournari J.C.S.Chem. Comm. 1976 434. T. Diebold B. Chevrier and R. Weiss Angew. Chem. Internat. Edn. 1977 16 788. Chemistry of the d-and f-Block Metals Manganese and Rhenium. An X-ray study of [Mn(TPP)] which binds dioxygen at low temperatures showed minimal structural differences at -175 "C compared to 20 "C. The manganese(I1) d5-ion is expected to be too large to fit into the plane of the porphyrin ring and the metal atom has a high degree of thermal motion in which it probably alternates between positions above and below the porphyrin plane.' On the basis of ab initio calculations it has been suggested that dioxygen should bind to this complex in a terminal-bent rather than sideways fashion and that the valency formalism [Mn"'-O2-] rather than [Mn'V-022-] (as was previously suggested; see last year's Report) should describe the metal-dioxygen interaction." Displacement of the large high-spin manganese(I1) ion out of the plane of the porphyrin nitrogen atoms (by 56 pm) occurs in the square-pyramidal complexes [Mn(TPP)(l-methylimidazole)] A smaller displacement (27 pm) of the smaller manganese(Ir1) ion occurs in [MnCl(TPP)] which also has a (slightly distorted) square-pyramidal structure.l2 Dechelation of manganese(I1) from the complex [Mn{TNMPP(-3)}] occurs under acidic conditions at a rate more than lo6 times faster than for analogous complexes of manganese(111).~~ Oxidation of [{Re(C0)3}2(TPP)](3) with SbC15 gives the products shown in reaction (1).The oxidation state of rhenium in (4)is formally 1.5 and its Re-Re distance (295 pm) which is shorter than that of (3) (312 pm) indicates that there is a metal-metal interaction. The TPP ligand of (4)is distorted; the two pyrrole rings which co-ordinate both rhenium atoms are coplanar with the mean plane of the macrocycle whereas the other two rings are tilted towards the metal ions to which they are co-ordinated [see structure (4)].The structure of (5) which formally contains Re"' is considered to have a similar arrangement of metal atoms.14 Iron. Of all porphyrin complexes those of iron receive the most attention because of their obvious relevance to biological systems.Those aspects examined this year include the synthesis of dihaem- and polymer-bound haem-complexes the binding of various ligands at iron centres oxidation and reduction of iron porphyrin complexes and n.m.r. properties. The dihaem complex (6) has been synthesized and the ligation of the iron centres by dioxygen and carbon monoxide studied. Like monohaem compounds (6) binds dioxygen reversibly but unlike the monomeric complexes it binds two molecules of CO with different rate constants. It has been suggested that the first fast rate J. F. Kirner C. A. Reed and W. R. Scheidt J. Amer. Chem. SOC. 1977,99 1093. lo A. Dedieu and M. M. Rohmer J.Amer. Chem. SOC. 1977,99 8050. J. F. Kirner C. A. Reed and W. R. Scheidt J.Amer. Chem. SOC. 1977,99 2557. A. Tulinsky and B. M. L. Chen J. Amer. Chem. SOC. 1977,99 3647. l3 P. Hambright Znorg. Nuclear Chem. Letters 1977 8 403. S. Katon M. Tsutsui D. L. Cullen and E. F. Meyer jun. J. Amer. Chem. SOC. 1977 99 620. J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall 0 1 (4) (Ph groups omitted) X I Me Et 7 Me 0 MeO,CCH,CH Me I X (6) X (when present) =O2or CO corresponds to binding at one iron which is partially in a strained four-co-ordinate form. After binding CO this iron centre closes to a six-co-ordinate form making available a less strained conformer about the central C-C bond (A) so that the second iron atom binds CO by a slower associative me~hanism.'~ The stimulus for the above work is the occurrence of four integrated haem units in haemoglobin and another advance in the synthesis of analogues of natural carriers of dioxygen has been the linkage of an [Fe(PPIX)] core to a polymer surface cia condensation of a l5 T.G. Taylor Y. Tatsuno D. W. Powell and J. B. Cannon J. C. S. Chem. Comm. 1977,732. Chemistry of the d-and f-Block Metals 173 peripheral carboxylate group. The linking group and a porphyrin side-chain carry imidazole groups which can axially ligate the iron centre. The dithionite-reduced haem polymer is very soluble in water and can take up dioxygen reversibly through several cycles before being irreversibly oxidized.Reduction with dithionite restores dioxygen-carrier activity.16 Interaction of dioxygen with iron(I1)-porphyrin complexes generally leads to irreversible oxidation unless measures are taken to prevent it (see above and last year's Report). Generally the oxidation process is considered to involve a dioxygen-bridged intermediate. This year such an intermediate [{TPP(p-Me)}FeO,Fe{TPP(p-Me)}] (petf=2.2 B.M. at -83 "C) which is involved in the interaction of [Fe{TPP(p-Me)}] with dioxygen in toluene at -80 OC has been detected by its 'H n.m.r. spectrum. It changes into [{TPP(p-Me)}FeOFe{TPP(p-Me)}] at -30 OC." Axial inner-sphere electron-transfer mechanisms are thought to operate in the oxidation of high-spin Fell-porphyrin complexes by quinones (to give hydroquinones) or aromatic nitro-compounds.l8 Low-spin axially ligated five- or six-co-ordinated porphyrin complexes are much more difficult to oxidize; they generally require the presence of an acid and use an outer-sphere mechanism.'* The complex [FeCl(OEP)] catalyses the epoxidation of cyclohexene in nitrobenzene solution. The mechanism of the reaction is not clear but it does not appear to involve a direct iron-dioxygen interaction since the cobalt(II1) analogue which is inert to dioxygen is also an active catalyst." Iron-porphyrin complexes with axial mercaptide ligands reproduce the charac- teristic U.V. absorptions at about 450 nm and about 360 nm (generally obscured) observed for cytochrome P-450 in the presence of CO. It has been shown that this characteristic spectral pattern arises by mixing of a charge-transfer transition from a lone-pair sulphur orbital of the mercaptide to a porphyrin ring orbital [e.g.(T*)] with a transition of the porphyrin ring that has the same symmetry [al,(T) a2,(7r)-* e,(~)].~' Generally the P-450-type spectra are shown by iron(I1) complexes such as [Fe(SR)(TPP)(CO)] but the low-spin iron(II1) complexes formulated as [Fe(SR)2{PP(IX)DME}]- and [Fe(SR)(PP(IX)DME}(PEt,Ph)] (R = Bun Ph C6H4CH2 or p-NO2C6H4) have now been shown also to have these spectral features. Evidently the combinations of iron(II1) with either two SR groups or one SR plus one PEt2Ph group are equivalent to iron(I1) with ligating SR plus axial CO in producing the right conditions for suitable mixing of charge-transfer transitions.21 The complex [Fe(C,H,s)(TPP)(PhSH)] which also has a P-450-type spectrum changes from a high- to a low-spin configuration as its temperature is lowered.A multiple-temperature X-ray study of this process has revealed that at 115 K the complex forms a 1:2 disordered mixture of five-co-ordinate high-spin [Fe(PhS)(TPP)] and six-co-ordinate low-spin [Fe(PhS)(TPP)(PhSH)]. At 4.2 K [Fe(PhS)(TPP)(PhSH)] is virtually exclusively formed and thus the change to the l6 E. Bayer and G. Holzbach Angew. Chem. Internat. Edn. 1977,16 117. l7 D.-H. Chin J. D. Gandio G. N. LaMar and A. L. Balch J. Amer. Chem. SOC.,1977 99 5486. C. E. Castro G. M. Hathaway and R. Havlin J. Amer. Chem. Soc. 1977 99 8032; J. H. Ong and C. E. Castro ibid. p. 6740.l9 M. Baccouche J. Ernst J.-H. Fuhrhop R. Schlozer and H. Arzoumanian J.C.S.Chem. Comm. 1977 821. 20 L. K. Hanson W. A. Eaton S. G. Sligar I. C. Gunsalus M. Gouterrnan and C. R. Connell J. Amer. Chem. SOC.,1976,98,2676. H. H. Ruf and P. Wende J. Amer. Chem. SOC.,1977.99 5499. 174 J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall high-spin form is associated with the loss of a thiol ligand.22 Further refinement of this work will establish more clearly its detailed mechanism and its relation to the change from a low-spin to a high-spin configuration that is undergone by cyto- chrome P-450 on binding a substrate. The binding of various other neutral and anionic ligands to iron-porphyrin centres has been studied. The 'H n.m.r.linewidths of the complexes [FeCl(P)B,] (P = TPP OEP or TPrP; B = substituted imidazole or pyridine) have shown that the lability of the axial ligands B is increased by electron-donating substituents on the porphyrin electron-withdrawing groups on the imidazole or greater steric bulk of the substituents of B. Pyridines are more labile than imidazole~.~~ A kinetic study of the displacement of Me2S0 from the complex [Fe(TPP)(Me,SO)]' by imidazole (ImH) has shown that the first step is addition of ImH followed by the rate-determining displacement of Me,SO by ImH to give [Fe(TPP)(ImH)2]+.24 Deprotonation (by such bases as hydroxide or Bu'O-) of ImH ligating this latter complex [equation (2)] has been postulated to occur but attempts to isolate (7) and (8).gave only {Fe(TPP)(Im)}n.2s -H+ -H+ [Fe(TPP)(ImH)2]+ [Fe(TPP)(Im)(ImH)] + [Fe(TPP)(Im)J (2) +H+ +H+ (7) (8) Substituent effects on the ligation of pyridine in [FeCI{TPP(m-or p-R)}-(py),(DMFX-,] and [Fe{TPP(rn-or p-R)}( py),(DMF)2-,] (R = various substi- tuents n = 1 or 2 py = pyridine DMF = dimethylformamide) have been examined electrochemically.Values of the Hammett reaction parameter p derived from equilibrium constants for binding of pyridine were independent of the degree of both axial ligation and of the meta- or para-substituent for the iron(1r) series. For the iron(1II) complexes however whereas the value for [FeCl{TPP(m -R)}-(DMF)( py)] (-0.123) differs from that of its para-substituted analogue (-0.454) only a single value is obtained for [FeCl{TPP(m -or p-R)}( p~)~] (-0.433).26 Adducts of iron(I1) porphyrins with nitroso-alkanes e.g.[Fe(TPP)(RNO)B] (R = Me Pr' or PhCH2CH2; B = various N-donor bases),27 and with dichlorocarbene e.g. [Fe(TPP)(CC1,)],28 have been reported. The X-ray structures of two different crystalline forms of the nitrosyl adduct [Fe(TPP)(NO)(p-CH,C,H,N)] have been determined. These structures differ in their Fe-N (N = axial pyridine nitrogen atom) distances (232.8 and 246.3 pm) and corresponding Fe-N-0 angles (138.5' and 143.7" respectively). 'The iron atom is also displaced out of the porphyrin-N plane towards NO by 9 and 11 pm respectively. The reasons for the occurrence of two crystalline forms are not clear but a linear correlation between v(N0) and Fe-N for these and related nitrosyl adducts was demonstrated and was consid- ered to reflect the lengthening of the Fe-N bond as a result of varying electron 22 J.P. Collman T. N. Sorrell K. 0.Hodgson A. K. Kulrestha and C. E. Strouse J. Amer. Chem. SOC. 1977,99,5180. 23 J. D. Saterlee G. N. LaMar andT. J. Bold J. Amer. Chem. SOC.,1977 99 1088. 24 R. F. Pasternack and J. R. Stahlbush J.C.S.Chem. Comm. 1977 106. 25 M. Nappa J. S. Valentine and P. Snyder J. Amer. Chem. SOC.,1977 99 5799. 26 K. M. Kadish and L. A. Bottomley J. Amer. Chem. SOC.,1977,99 2380. 27 D. Mansuy P. Battini J. C. Chottard and M. Lange J. Amer. Chem. SOC.,1977,99 6441. 28 D. Mansuy M. Lange J. C. Chottard P. Gueri P. Maliere D. Brault and M. Rougee J.C.S. Chem. Comm. 1977 648. Chemistry of the d- and f-Block Metals 175 release from the trans-NO ligand to the iron d,.orbital. The more nearly linear is the NO ligand the greater is its electron release.29 The demetallation of the complexes [FeCl(P)] (P = TPP or DPME) by HCl only proceeds via formation of iron(I1tporphyrin intermediates in their solutions pro- duced by the addition of iron(^^).^' The aquation of [Fe(TNPP-4)IS' over a pH and ionic strength range has been found to follow equilibria (3),(4) and (5). At pH 1-3 the solution contains the five-co-ordinate high-spin complex but at pH 10-12 the low-spin six-co-ordinate p-0x0-dimer is the major species [FeP(H2O)I5'+ H20 2 [FeP(OH)(H20)l4+ (31 [FeP(OH)(H20)l4++ [FeP(OH>2l3' (4) 2[FeP(OH)*j3+ S [{FeP(OH)}20]"+ (P = TNPP 4-) Other n.m.r.and related examinations of iron porphyrin complexes have been made. The 'H n.m.r. data yield a highly anisotropic magnetic moment (pi= 4.9B.M. and pll= 2.2 B.M.) for [Fe(TPP)] and the observed contact shifts confirm the intermediate spin ground-state configuration [S = 1;(dxy)2 for (dz2)2,(d,, ~d,,)~] the complex.32 The porphyrin 'H n.m.r. shifts of a variety of high-spin five-co- ordinate axial adducts of [Fe(TPP)] and related complexes with pyridine or methyl-substituted imidazoles are relatively insensitive to the axial base and therefore unlikely to be useful probes of biological The axial imidazole shifts however are consistent with primarily o-spin transfer to the metal and may provide a probe for this interaction in other specie^.'^ The zero-field splitting parameter for high-spin [FeCl(TPP)] (D= 5.9* 0.1 cm-') obtained from its single- crystal anisotropy is less than earlier values but close to the value (6.95 cm-') for chlorohaem and [FeCl{PP(IX)DME}].34 Cobalt and Rhodium.Dinuclear cobalt complexes of cofacial porphyrin ligands have been synthesized in two separate laboratories. Kang35 has synthesized the ligand (9) and prepared a dicobalt(I1) adduct. In toluene-dichloromethane solution kc / LN?YN 0'1 Y H"-/ \ -I R (9) R = n-hexyl 29 W. R. Scheidt A. C. Brinegar E. B. Ferro and J. F. Kirner J. Amer. Chem. SOC.,1977,99 7315. 3" J. H. Espenson and R. J. Christensen Znorg. Chem. 1977 16,2561. '' R. F. Pasternack H. Lee P. Malek and C. Spencer J.Znorg. Nuclear Chem.. 1977 39 1865. '' H. Goff,G. M. LaMar and C. A. Reed J. Amer. Chem. SOC.,1977,99 3641. 33 H. Goff and G. N. LaMar J. Amer. Chem. Soc. 1977,99,6599. 34 D. V. Behere V. R. Marathe and S. Mitra J. Amer. Chem. Soc. 1977 99 4149. 35 C. K. Kang J.C.S. Chem. Comm. 1977,800. 176 J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall the dinuclear cobalt complex reacts with air and 1-triphenylmethylimidazole (Ph,CIm) to give a diamagnetic pperoxo-complex which can be converted into a p-superoxo-complex by oxidation with di-iodine. The e.p.r. spectrum of the latter paramagnetic complex is consistent with the unpaired electron being mainly on oxygen i.e. the bonding description [(Ph~CIm)Co"1P-02-Co'1'P(ImCPh3)].35 Related dinuclear porphyrin ligands have been prepared by Collman and co- worker~.~~ Their bis-Co" derivative (with 1-methylimidazole as the axial base) has on the basis of e.p.r.data 'face-to-face' interacting metal atoms separated by about 650-680 pm. A copper derivative has also been prepared (see beIo~).~~ Rever-sible binding of dioxygen to the complexes [Co(TPPR)] where R=H O(CH2)4C(0)NHC5H4Nor O(CH2)3C5H4N occurs at low temperature and its dependence upon the nitrogen-base substituents has been determined by U.V. spectroscopy. The pyridine-analogue substituents at the phenyl ring of TPP are able to bind axially to the metal but have little effect on the strength of binding of dioxygen.37 The X-ray structure of [CoCl(TPPNMe)] (TPPNMe = N-methyl apy8-tetra- phenylporphyrin) reveals that the complex is a distorted square pyramid with an axial chloride.The methylated nitrogen is a greater distance (238.1 pm) from cobalt than are the other nitrogen atoms (201.6 pm) and its associated pyrrole ring is distorted out of the porphyrin plane so as to block the sixth co-ordination position of the metal. The high-spin cobalt atom is displaced towards the ~hloride.~~ Free-energy data for the substitution of water in [Co(TNMPP4-) (H20)2]5+ by anionic ligands such as NCS- to give [CO(TNMPP~-)(H,O)(CNS)]~+ have been determined.39 The first hydride complex of a rhodium porphyrin derivative [RhH(OEP)] has been prepared by treatment of [RhCl(OEP)] with dihydrogen in methan01.~' In benzene solution it loses dihydrogen to give the violet diamagnetic dimer [{Rh(OEP)},] (see Chapter 9).Copper. Dinuclear cofacial copper(I1) analogues of the above cobalt complexes prepared by Collman also show a metal-metal intera~tion.~~ Copper-substituted cytochrome c has been synthesized by dialysis of freshly prepared solutions of cyctochrome c against copper(I1) acetate at 4 0C.41It has the same electrophoretic and ion-exchange behaviour as the native enzyme and its e.p.r. and U.V. properties at pH 4-1 1show it to contain six-co-ordinate copper. This behaviour is unique to the natural protein because copper porphyrins normally do not take on two axial ligands. The new technique of X-a multiple scattering has been used to derive the electronic structure and spectral parameters of square-planar copper porphyrins.The values obtained are reasonably consistent with those derived from more established methods except for the values concerned with electronic excited 36 J. P. Collman C. M. Elliot T. R. Halbert and B. S. Tovrog Proc. Nu?. Acad. Sci. U.S.A.,1977,74 18. 37 F. S. Molinaro R. G. Little and J. A. Ibers J. Amer. Chem. Soc. 1977 99 5628. '* 0.P. Anderson and D. K. Lavallee J. Amer. Chem. Soc. 1977.99 1404. 39 K. R. Ashley J. Inorg. Nuclear Chem. 1977 39 357. 40 H. Ogoshi J. Setsune and Z. Yoshida J. Amer. Chem. SOC.,1977 99 3869. M. C. Findlay L. C. Dickinson and J. C. W. Chien J. Amer. Chem. Soc. 1977 99 5168. 42 D. A. Case and M. Karplus J. Amer. Chem. SOC.,1977.99 6182. 177 Chemistry of the d-and f-Block Metals Complexes of Dinucleating Ligands.-The complexing of two metal ions by the same ligand has intrigued chemists for a long time because of the possibility of studying metal-metal interactions and of using a combination of catalytic prop- erties to effect complicated transformations of substrates.Ligands which can combine two metal ions are termed 'dinucleating' and transition-metal complexes with dinucleating ligands have been reviewed Complexes of dinucleating ligands are different from dinuclear complexes although the differences are often rather small. Thus in a recent example taken at random,44 in complex (10) the square-pyramidal copper atoms are in a dinuclear 2+ (10) complex. Nevertheless this complex is of interest since it is diamagnetic.The spin-pairing is indicative of a strong metal-metal interaction which is not always observed in such systems. On the other hand the complex (11; R = H) has a genuine dinucleating ligand,4s prepared from 5-methyl-2-hydroxy-isophthalaldehyde and 1,3-diaminopropane in the presence of CU(C~O~)~,~H~O. It can be reduced electrochemically to a CU'~-CU~ species which (as deduced from the four-line e.p.r. spectrum at liquid-nitrogen temperature) contains distinct Cur' and Cu'. This spectrum for unknown reasons is temperature-dependent. The reduced homologue (11;R =Me) also contains Cut' and Cur but the e.p.r. spectrum is not dependent on temperature. These species may contain five-co-ordinate Cu'. M e RwR \ (11) However there is also an intermediate stage between dinuclear and dinucleate represented by (12) in which one of the metal ions is not completely enclosed.Complex (12) is diamagnetic and therefore may contain Vv+Cu' rather than 43 V. Casellato M. Vidali and P. A. Vigato Coordination Chem. Rev. 1977 23 31. 44 J. S. DeCourcy T. N. Waters and N. F. Curtis J.C.S. Chem. Comm. 1977 572. 45 R. R. GagnC C. A. Koval and T. J. Smith J. Amer. Chem. Soc. 1977,99 8368. J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall r R' (12) X=C1 or Br V'" +Cur' which might be expected to yield overall pararnagneti~m.~~ This kind of interaction may have very significant consequences. For instance the complex [Eu(fod),] (fod = heptafluorodimethyloctanedionate) is a widely used n.m.r.shift reagent and has now been shown to be efficacious with complexes with which it is proposed to interact in the manner shown (13).47 Generally when two complexing sites are available in a dinucleating ligand then both may be occupied by metal atoms of the same kind. Examples include the ligand (14) (LH,). LH3 forms a (13) (14) complex containing Fe [Fe2C12(0H2)L] in which the co-ordination number of one iron atom is made up by the water A related system (without the central phenolic OH) binds Ni and Co using subsidiary ligands to make up co-ordination numbers.49 The ligand (15) with two similar sites binds 2 atoms of types such as Cu Ni and CO.~'However when the two sites in a dinucleating ligand are not identical then the possibility arises that a given metal atom might prefer one site to the other.Thus o -acetoacetylphenol condenses with ethylenediamine in EtOH to yield (16).51 Reaction with nickel(1r) acetate in dichloromethane-ethanol results in the nickel ion being complexed by the four oxygen atoms. Subsequent treatment with 46 K. Okawa and S. Kida Inorg. Chim. Acta 1977 23 253. 47 L. F. Lindoy and W. E. Moody J. Amer. Chem. SOC.,1977.99 5863. 4R N. A. Bailey E. D. McKenzie J. M. Worthington M. McPartlin and P. A. Tasker Inorg. Chim. Acta 1977.25 L137. 49 R. Robson and D. G. Vince Inorg. Chim. Actu 1977 25 191. D. E. Fenton and S. E. Gayda J.C.S. Dalton 1977 2095. D. E. Fenton S. E. Gayda U. Casellato M. Vidali and P. A. Vigato Inorg. Chim. Actu 1977 21 L29. Chemistry of the d-and f-Block Metals M c m M e 0 .o..H" ."'"'1 NH 0 HN H... ..H /0' 0 N NH 0 HN M e w M e (15) (16) uranyl(v1) acetate produces a new complex in which U02 is bound by the N202 donor set the nickel still being bound by the O4donor set. An extensive series of researches has been reported concerning the ligand (17).52 For R= (CH& and R'=R2=Me the reaction with one equivalent of copper(I1) acetate results in the N202donor set being occupied by Cu; two equivalents give a mixture of 3 complexes one with a copper atom bound by both N202and O4donor sets and the other two in which each set is taken up individually. Uranyl and vanadyl prefer the O4set whereas nickel prefers N202.52This is however subject to change if minor variations are made in the ligand.Thus copper prefers the N202 set in the ligand (17; R' = R2= alkyl). In (17; R' = alkyl R2=phenyl) copper apparently prefers the O4set. In (17; R1 = R2=Me) either set can be R'yyy fNH OH O R 'NH 0 OH R ,UR2 (17) By suitable choice of ligand and conditions a variety of mixed complexes has been prepared Ni[ N202] /Cu[ 04]; Ni [N202] /VO[04] ; Cu[N202]/U02[04]; and Cu[N202]/Zn[04]. One awaits with interest the reports of the mutual influence of one of these metal ions on the properties and reactivity of the other. Aspects of the Chemistry of Complexes with Macrocyclic Ligands.-Macrocyclic ligands have received a lot of attention lately often because of some supposed value for the understanding of biological systems.The properties of the complex metal ions are affected by factors such as the size of the macrocyclic ring ring unsaturation and ring substitution. A considerable amount of effort has been put into determining the influence of ring size and the study of saturated systems has been rewarding. These ligands are usually denoted by a trivial nomenclature '* D. E. Fenton and S. E. Gayda J.C.S. Dalton 1977 2101. J3 D. E. Fenton and S. E. Gayda J.C.S. Dalton 1977 2109. 180 J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall [ 13]aneN4 is a thirteen-membered saturated ring containing 4 nitrogen atoms generally symmetrically arranged. A thermodynamic study has been made of the complexing of Cu2' with [9]aneN3 and with H2N(CH2)2NH(CH2)2NH2.54 The respective equilibrium constants are very similar but the [9]aneN3 presumably takes up the constrained facial positions so that its enthalpy contribution is less than for the open-chain ligand.The entropy contribution almost exactly compensates for the deficiency in the enthalpy contri- bution. By contrast [14]aneN4 forms much more stable complexes with Cu2+ than does the open-chain analogue.55 A direct comparison has been made of the compounds [12]- [13]- [14]- and [15]-aneN as complexing agents for Cu". AHe is always greater than for the corresponding open-chain ligand and is a maximum at [14]aneN4. The enthalpy contribution to the extra macrocyclic stability is in fact less for Cu2' than for Ni2' but is still significant even for [12]aneN4.56 However it has also been suggested that the factor having the major influence on the value of the Dq's produced by macrocyclic ligands is the distance between metal and donor atoms whether the ligand is open-chain or cyclic being in~idental.~' For complexes of [ 14]aneN4 and of the open-chain tetramines H2N(CH2)2- NH(CH2)3NH(CH2)2NH2 (1 9) with (18) and H2N(CH2)3NH(CH2)2NH(CH2)3NH2 Ni2+ the equilibrium (6) has been observed (L is one of the three ligands just [NiL(H20)2]2+$ [NiL]*++ 2H20 (6) blue yellow cited).The respective equilibrium constants are in the order [14]aneN4 > (18)> (19) but it is claimed that there are no special cyclic factors involved and that the origin of these different equilibrium constants lies in steric repulsions.It has been suggested on the basis of spectral data that the [12]aneN4 forms stronger bonds to Ni2+ than either [14]- or [15]-aneN4 and that this is primarily a size That being so then a different ion of different size should fit another macrocycle better than it fits [12]aneN4. A series of complexes [C~L(anion)~]' has been prepared (L= [13]aneN4-[16]aneN4 inclu~ive).~~ Values of Dq due to the macrocycles are in the order [14] > [13]> [15]> [16]. It has been suggested that Co3' fits [14]aneN4 best and that [13]aneN4 is too small whereas [15]- and [16]-aneN4 are too big. This is also reflected in the chemistry. Thus for [CoLCl,]' (L=[13]aneN4-[16]aneN4 inclusive) the rate constants for the aquation of the first chloride in 0.1MHN03 vary from 2.6-1.1~10-~s-~ in the order [16]>[15]>[13]>[14].60 It has been found that for the unusual ion Mn3+ [14]aneN4 is a better ligand than [ 15]aneN4.61 " L.Fabbrizzi and L. J. Stompa Znorg. Nuclear Chem. Letters 1977,13 287. '5 M. Kodama and E. Kimura J.C.S.Dalton 1977 1473. 56 A. Anichini L. Fabbrizzi P. Paoletti and R. M. Clay Znorg. Chim. Actu 1977 22 L25. '' D. Gattesschi and A. Scozzafava Znorg. Chim. Acta 1977 21 223. '* A. Anichini L. Fabbrizzi P. Paoletti and R. M. Clay Znorg. Chim. Acru 1977 24 L21; L. Fabbrizzi Znorg. Chem. 1977 16 2667. 59 Y. Hung L. Y. Martin S. C. Jackels A. M. Tait and D. H. Busch J. Amer. Chem. SOC.,1977 99 4029. 6o Y. Hung and D. H. Busch J. Amer. Chem. SOC.,1977,99,4977. 6' P. S. Bryan and J. M. Calvert Znorg.Nuclear Chem. Letters. 1977 12 615. Chemistry of the d- and f-Block Metals Extension to substituted ligands62 introduces possibilities of isomer formation which are not entirely absent with the unsubstituted rings.60 If the ring is unsaturated that introduces further complications. A comparison has been made of the ligands [14]aneN4 (L') Me6[14]-4,11-dieneN4 (L') [141-4,7,11,14-tetraeneN (L3) and [ 141-4,7-dieneN4 with C02+.63 Complexes isolated include [CoL'I2+ [COL~(H~O)~]~+ All three have low-spin cobalt and [COL~(H~O)~]~+. but [COL'(H~O)~]~+ could not be synthesized. Evidently axial interactions increase with the degree of unsaturation of the ring. Hydrolysis of rneso-[CoLC12] (L = Me6[ 141-4,ll-dieneN,) by base has been shown to proceed in two steps both dissociative in character The replacement of the first chloride is nine times as fast for the complex of 5,12-Me2[14]-4,ll-dieneN,and between lo2 and lo3 times as fast as for the complex of the saturated ligand.Evidently ring unsaturation tends to stabilize the intermediate of low co-ordination number (in this case 5).64 The degree of unsaturation has been shown to affect the specific rates of hydrolysis reactions in similar A very extensive study of Co complexes with various [14]N4 rings suggests that the Co2'/Co' redox couple is much affected by ring unsaturation; the greater the degree of unsaturation the more stable is the lower oxidation state.66 This is also so in more complex systems. Thus the quinqueden- tate macrocyclic ligand (20) and Mn2' form pentagonal-bipyramidal complexes [Mn(20)(NCS)2] for 1 = rn = 3 and n = 2 and anion = NCS,67" with 5 nitrogen atoms in the pentagonal plane.Such a structure also seems to hold for 1= rn = n = 2 and for 1 = rn = 2 and n = 3. Such complexes can be oxidized to Mn3' species but oxidation does not occur so easily if the ligand is uncharged or the ring is unsaturated. Oxidation to Mn3' does not occur if the macrocyclic ligand (21) fN (14-membered ring) is in the complex. Oxidation of the complex of (20; 1= rn = n = 2; 15-membered ring) is easier than for the complexes of (20; 1 = rn = 2 n = 3; 16-membered ring) and (20; 1= rn = 3 n = 2; 17-membered ring).676 Perhaps not 62 R. W. Hay and D. P. Piplani J.C.S. Dalton 1977 1956; R.W. Hay D. P. Piplani and B. Jeragh J.C.S. Dalton; 1977 1951; T. J. Lotz and T. A. Kaden J.C.S. Chem. Comm. 1977 15. 63 J. F. Endicott J. Lilie J. M. Kuszaj B. S. Ramaswamy W. G. Schmonsees M. G. Simic M. D. Glick and D. P. Rillema J. Amer. Chem. Soc. 1977,99,429. 64 P. L. Kendall and G. A. Lawrence Austral. J. Chem. 1977,30 1841. 65 C.-K. Poon and C.-L. Wong J.C.S. Dalton 1977 523. 66 A. M. Tait F. V. Lovecchio and D. H. Busch Inorg. Chem. 1977,16,2206. 67 (a)M. G. B. Drew A. J. bin Othman S. G. McFall P. D. A. McIlroy and S. M. Nelson J.C.S.Dalton 1977,438; (b)J. C. Dabrowiak L. A. Nafie P. S. Bryan and A. T. Torkelson Inorg. Chem. 1977,16 540. 182 J. R.Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall surprisingly redox potentials also vary between diastereoisomers,68 but a considerable amount of work remains to be done before all the influences on the redox potential and mechanism69 can be disentangled and understood.3 Small Multiply- bonding Nitrogen Ligands The chemistry of the nitrosyl ligand has continued to be a major source of interest. This stems in part from the bent-straight (1-electron-donor-3-electron-donor) problem and in part from problems associated with the reactivity of NO and with its occurrence in exhaust fumes. There have also been several ingenious attempts to overcome the problems associated with the oxidizing power of NO when it encounters complexes that contain transition metals in low oxidation states. The bent-straight nitrosyl transition has been discussed in terms of the stereo- chemical control of valence.This is the phenomenon whereby a closed-shell complex can take on an extra ligand the ligand electrons being accommodated by passing an electron pair from the metal to a three-electron donor such as NO which consequently becomes a one-electron donor. The oxidation state of the metal changes accordingly. The complex [Fe(NO)(Me2AsC6H4-2-AsMe2)2]2+ has a square-pyramidal configuration with a straight NO ligand at the apex LFeNO = 172.8(7)" N-0 = 141.1(27) pm. Its reaction with NCS-produces [Fe(NCS)- (NO)(Me2AsC6H4-2-AsMe2)]+, in which LFeNO is now 158.6(9)0 and N-0 = 109.6(10)pm.70 In the former case the oxidation state of the iron is taken to be (I) and in the latter (111).Another series of iron nitrosyl complexes has been reported namely ~~~~s-[F~(NO)(S~CNM~~)~M~CN]' and cis- and trans-[Fe(NO)- (S2CNMe2)2X] (X=Br NO C1 Br I efc.). The first compound is made by the action of NO[BF,] on [Fe(S2CNMe2)2].71 The compound [Fe(NO)-(S2CNMe2)2(N02)] has a linear iron-nitrosyl system [ 174.9(5)"] with N-0 = 113.6(6) pm.71 There is however a problem of definition of oxidation state particularly where the configuration of the NO is not known. I.U.P.A.C. n~menclature'~ avoids this by the device of regarding NO as a neutral ligand in which no attention is paid to formalisms such as NO' or NO-. Thus the complexes [Mo(NO)ClSI2- and [Mo(NO)C14]- which by I.U.P.A.C. nomenclature would be designated as containing Mo'" have ;(NO) in the range 1675-1695cm-' and are claimed to contain M011,73a whereas the remarkable water-soluble species [Mo(NO)CI,(H,O)]~- [F(NO) = 1624 cm-'1 is said to contain Mo' predicated on NO+.736 However the NO is formulated the stability to water is unexpected.The iron nitrosyl species [Fe(CO),NO]- has been reportedly characterized by its i.r. spectrum and shown to have C3"symmetry. It now appears74 that in THF solution the i.r. spectrum is essentially doubled and this has been attributed to a J. Hanzlik A. Puxeddu and G. Costa J.C.S.Dalton 1977 542. 69 N. Al-Shatti M. G. Segal and A. G. Sykes J.C.S.Dalton 1977 1766. 70 J. H. Enemark R. D. Feltham €3. T. Hine P. L. Johnson and K. B. Swedo J. Amer. Chem. Soc. 1977 99 3285. " 0.A. Ileperuma and R. D.Feltham Inorg. Chem. 1977.16 1876. 72 'Nomenclature of Inorganic Chemistry Second Edition Definitive Rules 1970' I.U.P.A.C. Butter- worths London 1971. 73 S. Sarkar and A. Miiller (a)Angew. Chem. Internat. Edn. 1977 16,183; (6) ibid. p. 468. 74 K. H. Pannell Y.-S. Chen and K. L. Belknap. J.C.S.Chem. Comm. 1977 362. Chemistry of the d- and f-Block Metals 183 dissociation equilibrium involving a tight ion-pair based on a cation-nitrosyl inter-action the first of its kind to be recognized. Resolution of another spectroscopic conundrum has been achieved. The complex [Mn(C0)4(NO)] has been shown by X-ray analysis to have a trigonal- bipyramidal structure with NO in an equatorial position whereas the solution i.r. spectrum has been interpreted in terms of a trigonal bipyramid with axial NO.The spectrum of the complex in a matrix of N2 has now been shown to be in accord with the X-ray structure.75 Matrix (argon and methane) i.r. studies of [Fe(CO)2(NO)2] have demonstrated its dissociation to [Fe(CO)(N0)2]; in a dinitrogen matrix (at 20 K) N2 can be incorporated yielding the novel species [Fe(CO)(N,)(NO),] and [Fe(N2)2(N0)2].76 Similarly [CO(CO)~(NO)] yields [CO(CO)~(NO)] upon irradia- tion at 20 K and [Co(CO),(N,)(NO)] and possibly [Co(CO)(N,),(NO)] can be formed.77 NO is not labile under these conditions. The reduction of NO complexes if it is sufficiently extreme can lead to hyponi- trite (02N22-) complexes. The reduction of [Mn(CN)S(NO)]3- by sodium in liquid ammonia yields [Mn(CN)=,(NO)]"-; that of [Fe(CN),(NO)]-yields [Fe(CN)3(NO)]5- and [CI-(CN)~(NO)]~- yields in turn [Cr(CN)5(N0)]4- [ti(N0) = 1510cm-'1 [Cr2(CN)7(N0)2]7- [:(NO) = 1490 cm-'1 and [Cr2(CN)6(NO)2]'-[:(NO) = 1460 ~m-'].~' The structures of these remarkable species may contain bridging NO rather than hyponitrite but this has not been properly established.However [Pt(PPh,),(NO),] which reacts with carbon monoxide to yield CO + N20 has been shown to be a square-planar Pt" complex (22) with a planar five-membered ring.79 (22) Distances/pm Other bis(nitrogen) ligands have also been mentioned during the year. In the past for example the trioxodinitrate(2 -) ion has been characterized as its sodium salt and a few complexes have been identified in solution. It has now been shown that the reaction of [CO(NH,)~]~+ with MC12,6H20 (M=Mn Fe Co or Ni) in the presence of sodium trioxodinitrate yields crystalline materials [CO(NH,)~]~[M(N~~~)~],,~H~O.'~ These complexes have a characteristic i.r.spec- trum with bands at 1370-95 1240-60 1050-80 and 940-60 cm-' assigned to ti(N=N) and Y(N-0). The compound with M = Co rearranges in water to give [co(NH3)6]2[co2(N203)s], of unknown structure and all the compounds decom- pose slowly when heated yielding nitrosyl species. It has been suggested that the N20 in [Ru(NH,)~(N~O)]*+ is N-bonded to the ruthenium despite an earlier analysis of the i.r. spectrum which had been taken to indicate O-bonding.81 75 A. J. Rest and D. J. Taylor J.C.S. Chem. Comm. 1977 717. 76 0.Crichton and A.J. Rest J.C.S. Dalton 1977 656. 77 0.Crichton and A. J. Rest J.C.S. Dalton 1977 536. 78 J. Schmidt 2.anorg. Chem. 1977 431 284. '' S. Bhaduri B. F. G. Johnson A. Pickard P. R. Raithby G. M. Sheldrick and C. I. Zuccaro J.C.S. Chem. Comm. 1977 354. C. A. Lutz A Lomax and L. Toh J.C.S. Chem. Comm. 1977 247. F. Bottomley and W. V. F. Brooks Znorg. Chem. 1977,16,501. 184 J. R. Dilworth G.J.Leigh R. L. Richards and K. W.Bagnall The catalysis of the reaction (7) does not necessarily require the formation of a dinitrosyl or a hyponitrite as an intermediate though they may be involved. Thus C0+2NO -+ CO,+N,O (7) [RhC12(CO)J gives rise to a catalyst in ethanol solution but only under a mixture of CO and NO. Once generated the catalyst regenerates the starting material if exposed to CO alon'e and under NO alone it forms a red nitrosyl probably [RhClz(NO)z]-.The catalyst solution shows F(C0) at 2095 and F(N0) at 1715 and 1780 cm-' and a formulation for the catalytic species that has been suggested is the five-co-ordinate [RhClz(CO)(NO)2]- which contains Rh"' if one allows the formu- lation to contain NO'.82 A related series of complexes [M(NO)z(PPh3)z]' (M = Co Rh or Ir) has been shown to undergo exchange of phosphine on the basis of a 31P n.m.r. study but the mechanism is dissociative for M=Co and associative for M=Rh or Ir.83 The complexes [M(N0)2(PPh3)2] (M = Ru or 0s) react with CO to form N20 and CO and tricarbonyl species. It has been proposed that the key to this is the trans- formation @) with NO' and NO- then coupling to give N20.The Co and Fe analogues more prone to dissociative mechanisms do not produce N,0.83 [M"'(N0')2] 2 [M"2''(NO')(NO-)L] (8) Other interesting rhodium species mentioned during the year include [Rh(NO)- (PPh3),S02] which has a five-co-ordinate structure that is not easily described in ideal terms. The NO is 'bent'; LRhNO = 140.6(6)0 and N-0 = 19537) pm.84 The complex [Rh(cycl~-octadiene)~]' reacts with nitrosonium salts in MeCN to generate [Rh(MeCN)4(N0)]2'.85 This can be converted into [Rh(NO)-(S2CNMe2)3]+ which is fluxional down to -95 "C in solution has a low i(NO) and must be either octahedral with a straight NO or seven-co-ordinate with a bent NO.s5 A complex related to the precursor [Rh(MeCN)3(NO)(PPh3)2Jz+, has a bent NO; LRhNO = 118.4(6)" N-0 = 115.9(10) A rather unusual rhodium nitrosyl [Rh(NO)(N03)(PPh3)2] [;(NO) = 1655 cm-'1 has been obtained by con- verting [Rh(CO)CI(PPh,),] into [Rh(CO)(NO)(PPh,),Cl,] using sodium nitrite and hydrogen chloride and treating the latter [?(NO) = 1630 cm-'1 with silver nitrate." Finally the compounds [M(NO)(MeCN),(PPh,),]' (M = Rh or Ir) [?(NO) = ca.1540cm-'] react with catechol to form compounds (23) with a bent NO [i=ca. 1590 cm-'1 which in another manifestation of the stereochemical control of valence can lose a molecule of phosphine to give (24) in which NO is straight [Y(N0)=1850~m-'].~~ Bridging nitrosyls are not yet very common. New examples during the past year include [cL-(C~)-~-(N~)-{C~(C~HS))~I and [(ML),(NO)I' [{CL-(NO>CO(C,H,)}~I,~~ where M = Co or Fe and L = S(CH2)2NMe(CH2)2NMe(CH2)zS.90 Occasionally '* D.E. Hendrikson C. D. Meyer and R. Eisenberg Znorg. Chem. 1977.16 970. 83 S. Bhaduri K. Grundy and B. F. G. Johnson J.C.S.Dalton 1977 2085. 84 D. C. Moody and R. R. Ryan Inorg. Chem. 1977,16,2473. N. G. Connelly P. T. Draggett M. Green and T. A. Kuc J.C.S. Dalton 1977 70. 86 B. A. Kelly A. J. Welch and P. Woodward J.C.S. Dalton 1977 2237. A. Dowera and M. M. Singh Transition Metal Chem. 1977 2 74. M. Ghedini G. Dolcetti B. Giovanitti and G. Denti Inorg. Chem. 1977 16 1725. 89 W. A. Herrmann and I. Bernal Agnew. Chem. Internat. Edn. 1977,16 172. 90 H. N. Rabinowitz K. D. Karlin and S. J. Lippard J. Amer. Chem. SOC., 1977.99 1420.Chemistry of the d- and f-Block Metals (23) (24) cluster nitrosyls have been reported. Thus [Os3(CO),2] reacts with NO in octane at 126 "C to yield [OS,(CO),(NO)~] (two 3-electron donors replacing three 2-electron donors) which can react with CO to yield [OS,(CO)~,(NO)~] and with P(OMe)3 to give [Os3(CO),(NO)2{P(OMe)3}].91 The complex [Fe(CO)(N0)2(PMe2H)] reacts with [CO(C,H~)(CO)~] to produce (25) as one product of several.,' L Me2 J (25) Nickel nitrosyls are of interest because of the way they can change co-ordination number and stereochemistry. Thus [Ni(C0)2L2] reacts with NO[PFs] to form [Ni(NO)L,]' (n =2 or 3 for L=PPh3; n = 3 for L=PMe2Ph)., The compound [Ni(NO)(PPh,),]'[ G(N0) = 1755 cm-'1 has a C3"stereochemistry and dissociates in solution to yield [Ni(NO)(PPh,),]'.At -85 "C in solution it slowly reaches equilibrium with a square-planar form. It reacts with NaS2CNet2 to yield [Ni(NO)(PPh,),(S,CNEt,),] which is a single species and apparently fl~xional.,~ Cobalt nitrosyls of the type [CO(NO)~L~]Y = RCN ROH Me2C0 etc.; Y = (L PF6 BF, etc.) are readily available from the reaction of [{CO(NO)~C~},] with silver ion in the presence of L. They have ;(NO) at ca. 1800 1900 cm-.' with two linear NO The complexes [Co(WS,),]'- and [Fe(MS4)2]2- (M = Mo or W) react with NO to form 1 1adducts in which NO is bound to Co or Fe and has a stretching frequency of ca. 1700 ~m-'.,~ Ruthenium nitrosyls have long attracted attention and this is not changing. The complex [Ru(NO)(P~~PCH~CH~CH~PP~~)~]+ [N-0 = 120(1)pm _LRuNO = ca.174"] is very similar to its Ph2PCH2CH2PPh2 analogue in although its 1-electron-reduction product disproportionates more readily than that of the ethylene derivative. The novel nitrosyls [Ru(LL)~(NO)X] [fi(NO) = ca. 1910 cm-'1 (LL = violurate) have been rep~rted.~' The complexes [Ru(NO)(bipy),X]"+ (bipy = bipyridyl; X = CI N3 or NOz n = 2; X =NH3 py or MeCN n = 3) (py = 91 S. Bhaduri B. F. G. Johnson J. Lewis D. J. Watson and C. Zuccharo J.C.S. Chem. Comm. 1977 477. 92 E. Keller and H. Vahrenkamp Angew Chem. Internat. Edn. 1977 16 541. 93 S. Bhaduri B. F. G. Johnson and T. W. Matheson J.C.S. Dalton 1977 561. 94 D. Ballivet and I. Tkatchenko Inorg. Chem. 1977 16,945. 95 A. Miiller and S.Sarkar Angew. Chem. Internat. Edn. 1977 16 705. 96 G. Bombieri E. Forsellini R. Graziani and G. Zotti Transition Metal Chem. 1977 2 264. 97 C. Bremard M. Muller G. Nowogrocki and S. Sueur J.C.S. Dalton 1977 2307. 186 J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall pyridine) undergo reversible 1-electron reduction and [Ru(NO)(bipy),Cl]' has been isolated. It was found that P(N0) correlates linearly with Ei and shows a drop of ca. 300 cm-' upon redu~tion.~~ For this and other reasons it is believed that the site of reduction is the NO (characterized as strongly 'NO+') rather than the ruthenium. The complex [R~(bipy)~(NO)Cl]~' reacts with aryldiazonium ions ArN2' to yield [R~(bipy)~(N~Ar)Cl]" which have high fi(N=N) (>1980 cm-') and hence probably contain 'ArN2+'.These compounds undergo irreversible 1-elec-tron reduction losing N2.99 However parallels between N2Ar and NO can be misleading if only because the chemical reactions open to N2Ar are so much greater. Thus [Ir(CO)Cl(PPh3)2] reacts with diazonium salts to give varieties of products depending upon the conditions. Amongst those described are a dinuclear bis(aryldiazenidoj-complex,'oo ortho-metallated species,1oo diazene and hydrazine and compounds [Ir(CO)C1(N2Ar)(PPh3)2X](X = anionic ligand) which possess doubly bent N2Ar groups ('N2Ar-').'02 The nitrosyls of Group VI elements have also elicited considerable interest. New synthetic routes to molybdenum nitrosyls have been described.lo3 The novel [Cr(N02),(NO)( py)],py has octahedral co-ordination LCrNO = 180.0" N-O(nitrosy1) = 115.0(1) pm and 0-bonded nitrito-gro~ps.~~~ Parenthetically a further mode of binding for NO2 has been suggested in a reformulation of the structure of Vkzes Red Salt as (26).lo5 Other chromium complexes that have been reported include [Cr(CN),-x(H20)x(N0)]'x-3'-(x = 3,4 or 5) which have been studied electrochemically,'06 and (27) which has a linear NO system in which N-0 = 119.3(6) pm.'" 2 Finally molybdenum and tungsten complexes of the type [M(CO),(LL)] (LL = bidentate neutral ligandj have been used to synthesize new nitrosyls and aryldiazenido-complexes.Thus [W(C0),(Me2PCH2CH2PMe2)] reacts with NO[PF6] to give fUC(?)-[W(CO)3(NO)(LL)][PF6]and with an excess of NO[PF,] 98 R.W. Callahan and T. J. Meyer Inorg. Chem. 1977,16,574. 99 W. L. Bowden G. M. Brown E. M. Gupton W. F. Little and T. J. Meyer Inorg. Chem. 1977 16 213. ''' N. Farrell and D. Sutton J.C.S. Dalton 1977 2124. lo' A. B. Gilchrist and D. Sutton J.C.S. Dalton 1977 677. R. E. Cobbledick F. W. B. Einstein N. Farrell A. B. Gilchrist and D. Sutton J.C.S. Dalton 1977 373. F. King and G. J. Leigh J.C.S. Dalton 1977 423. C. M. Lukehart and J. M. Troup Inorg. Chim. Acta 1977 22 81. lo5 A. E. Underhill and D. M. Watkins J.C.S. Dalton 1977 5. J. MocBk D. Bustin and M. Ziakova. Inorg. Chim. Acta 1977 22 185. lo' D.Webster R. C. Edwards and D. H. Busch Inorg. Chem. 1977 16 1055. 187 Chemistry of the d-and f-Block Metals to produce a further complex postulated to be [(CO)3(NO)W(LL)W(CO)3-(N0)I2'.The bipyridyl complex [M~(CO)~(bipy)] reacts with Z[PF6] (Z = NO or ArN2) to produce fa~-[Mo(CO)~(bipy)Z]' and oxidation of the diazenido-complex with bromine yields {M~Br~(bipy)(N~Ph))~ which has V(N=N) = 1390 cm-'. Now [(MO(CO),(N,P~)}~] has V(N=N)= 1479 cm-' so the value of 1390 cm-' is taken to indicate either doubly bent or doubly bridging N2Ph.'09 Nitrogen fixation has continued to receive a great deal of attention. Amongst new dinitrogen complexes reported are [M(N2)6] (M=Ti V or Cr). These were generally made by reactions of the metal atoms in a nitrogen matrix at 10-15 ~.110.111The titanium complex has V(NEN) at 2131 2100 and 2095 cm".llo From an analysis of the crystal-field and charge-transfer spectra of these complexes and of the analogous [M(CO),] it was concluded (again) that N2 is a poorer 0-donor and n-acceptor than CO."' The complexes [HgX2N2] (X = C1 Br or I) have been observed in Ar matrices."2 The first stable chromium derivative containing a tertiary phosphine ci~-[Cr(N,),(PMe,)~l has been reported.It has V(N=N) = 1990 1918cm-' and decomposes at room temperat~re.''~ The osmium com- plex [OS(M~~ASC~H~-~-ASM~~)~C~(N~)]+, which has v(NZN) = ca. 2080 cm-' depending on the counter anion,'' has been prepared from its nitrosyl analogue [V(N-0) also high at ca. 1860 cm-'1 by its reaction first with hydrazine and then with acid. The X.p.e. spectra of [OS(NH~)~(N~)~]CI~ and [OsC1(NH3)4(N2)OsCl(NH3)4]C13 have been measured. The latter contains both 0s" and Osrrl,and not 'averaged' o~miums.''~ However the most unusual dini- trogen complex reported during the year is [RhCl(N2)(PPri3)2]."6 This has been shown by X-ray analysis to possess N2 that is sideways bonded to a single rhodium V(NGN) = 2100 cm-' N-N = 83(2) pm the first of its kind.This astoundingly short N-N distance is paralleled by similar short distances in the O2[103(l)pm] and ethylene [131.0(4) pm] analogues. This is the shortest C-C separation yet observed for complexed ethylene. However these extra-short separations may be an artefact of crystal disorder. These results are even more unexpected when it is realized that the rhodium N2-complex [RhH(N2)(PBut2Ph)2] has normal end-on co-ordination of N2. The reactivity of co-ordinated N2 has received considerable attention both theoretically and empirically.The bonding of diatomic molecules in general to transition metals has been discussed. It has again been suggested that CO may be a better .rr-acceptor than N2.' '' A MO description of diazenido- dinitrogen- and related complexes has been used to discuss the successive protonations of co-ordinated Nz to produce NH3.'18 The analysis of alkylation of co-ordinated N2 J. A. Connor P.1. Riley and C. J. Rix J.C.S. Dalton 1977 1317. '09 D. Condon M. E. Deane F. J. Lalor N. G. Connelly and A. C. Lewis J.C.S.Dalton 1977 925. R. Busby W. Klotzbiicher and G. A. Ozin Inorg. Chem. 1977,16,822. "' A. B. P. Lever and G. A. Ozin Inorg. Chem. 1977,16 2012. ''* D. Trevault D. P.Strommen and K. Nakamoto J. Amer. Chern. SOC.,1977 99 2997. H. H. Karsch Angew. Chem. Internat. Edn. 1977 16 56. F. Bottomley and E. M. R. Kiremire J.C.S. Dalton 1977 1125. C. Battistoni C. Furlani G. Mattogno and G. Tom Inorg. Chim. Acta 1977 21 L25. 'I6 C. Busetto A. D'Alfonso F. Maspero G. Perego and A. Zazzetta J.C.S. Dalton 1977 1828. 'I7 R. Hofmann M. M.-L. Chem and D. L. Thorn Inorg. Chem. 1977,16 503. 'I8 D. L. Dubois and R. Hofmann Nouueau J. de Chimie 1977 1,479. J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall suffers from being based upon a hypothetical reaction path which is different from that which has now been shown to be foll~wed."~ The general reactions of the complexes [M(N2)2(dppe)2] [M =Mo or W; dppe = 1,2-bis(diphenylpho~phino)ethane]with alkyl acyl or aroyl halides RX (X = C1 Br or I) to form [M(N2R)X(dppe)2] have been described in some detail.12' These reactions have been shown to involve a predissociation of the bis(dinitrogen) complex and the generation of radicals from the halides as shown in Scheme 1.In Scheme 1 inert solvents R attacks the remaining N2 to form [M(N2R)X(dppe),]."' Where then after protonation a diazobutanol derivative e.g. [MoBr{N2CH(CH2)30H}-formed and a solvent radical produced; in the case cited this is the b(CH2)sCH. radical. This then attacks the dinitrogen yielding first a diazenido-complex and then after protonation a diazobutanol derivative e.g. [MOB~(N~CH(CH~)~OH} (d~pe)~]+.l~l The diazenido-complexes can give rise to amines but only after rather severe treatment.'22 A very unusual reaction of co-ordinated dinitrogen in [Mn(C,HS)(CO),(N2)] has been With phenyl-lithium it generates [Mn(CsHs)(C02{N(Ph)=N}]- which with acid yields [Mn(CSHS)(CO),-{N(Ph)=NH}].These products are not very stable and the suggested structures are somewhat speculative. The conversion of N2 into ammonia has been observed with novel Haber cata- lysts (for example K-lamellar graphite-Fe -Ru or -0s) as well as with N2 complexes. In the graphitic catalyst cited the reactivity for ammonia production correlates with the rate of isotope scrambling in a mixture of 14N2 and 15N2.124 Presumably dissociation of dinitrogen is rate-determining. A much more unusual system with far-reaching implications uses H2 that has been produced by the photolysis of water chemisorbed on iron-doped Ti02.125 The photolysis of water by iron-doped TiO under Ar produces 2 moles of H2 and 1mole of 0,.The same photolysis in the presence of N2 at atmospheric pressure produces O2 in the same amount but the evolution of H2 is completely inhibited. Traces of NH3 and N2H4 were detected and the relevance of these observations to the reaction of N2 has been confirmed by using I5N2. Sunlight was found to be a sufficient activator. Yields of ammonia are still very small (0.2 g of TiO doped with 0.2% Fe,O, when irradiated with mercury light and equilibrated with water vapour produced 0.2 pmol of H2 1.05 pmol of 02,1.39pmol of NH3 and 0.15 pmol of N2H4 after 2 h) but the consumption of HZ by N2 is quantitative.It is assumed that N2H2 J. Chatt R. A. Head G. J. Leigh and C. J. Pickett J.C.S. Chem. Comm. 1977 299. J. Chatt A. A. Diamantis G. A. Heath N. E. Hooper and G. J. Leigh J.C.S. Dalton 1977 688. ''I P. C. Bevan J. Chatt A. A. Diamantis R. A. Head G. A. Heath and G. J. Leigh J.C.S. Dalton 1977 1711. I" P. C. Bevan J. Chatt G. J. Leigh and E. G. Leelamani J. Organometallic Chem. 1977,139 C59. 123 D. Sellmann and W. Weiss Angew. Chem. Internat. Edn. 1977 16,880. lZ4 M. E. Volpin Yu. N. Novikov V. A. Postnikov V. B. Shur B. Bayerl L. Kaden M. Wahren L. M. Dmitrienko R. A. Stukan and A. V. Nefedev 2.anorg. Chem. 1977 428 231. 12' G. N. Schrauzer and D. T. Guth J. Amer. Chem. Soc. 1977 99 7189. Chemistry of the d- and f-BlockMetals 189 is an intermediate in the reduction but direct evidence is as yet lacking.More conventional aqueous fixing systems that have been described during the year include molybdate-cysteine plus reductant'26 and [MoOC~(CNCH,)~]' plus reduc- tant.12' In both these cases the yields of ammonia are small but N2H2 and sideways-bound NZ are claimed to be involved and some supposed parallels with the rather poorly understood enzyme system nitrogenase are drawn. The degradation of complexed N2 to give ammonia has received considerable attention. The conversion of [Mo(N~)~(PM~~P~)~] into [MOC~~(N~H~)(PM~~P~)~] and thence by the action of sulphuric acid into NH3 (0.68 mol per atom of Mo) is believed to involve step-wise protonation of co-ordinated N2 and possibly MoEN or Mo=NH groups as intermediate^.'^^"^^ The compounds [W(N2)2(PMe2Ph)4] and [M(N2)2(PMePh2)4] (M=Mo or W)129 can also be protonated in methanol by sulphuric acid to yield ammonia.For M = W the yields of ammonia approach 2 mol per atom of meta1,13' and methanol is itself a sufficiently strong acid to produce NH3 from N2. The postulated reaction sequence supported in part by the inter- mediates actually isolated is shown in Scheme 2. The implications of this kind of H+ H+ H+ MO-NGN + M-N=NH -+ M=N-NH2 M-NHNH2 H+ H+ MV'+NH3 + M-NH2 + M=NH+NH Scheme 2 process for the functioning of nitrogenase have been presented."' An attempt to adapt this scheme to an electrochemical reduction of N2 by constructing an elec- trode that bears groupings such as (28) failed because direct discharge of protons from the solution is preferred to reduction of N2.I3l n 3 Ety Ph,P ,PPh ,Sn-0-Si-( CH 2)3C N MoN2 E~)J Ph2P' 'PPh, u Alkylimido-complexes are the nitrogen analogues of carbene complexes but are not generally graced with the sophisticated name 'nitrene complexes'.The reaction of Os04with Bu'N=PPh3 gives rise to [Os(NBu)03] and the poly(nitrene) species [Os(NBu),02] and [Os(NBu),O]. These have strong bands in the i.r. spectrum in the range 1160-1200cm-' assigned to i;(Os=N) at a higher frequency than 126 P. R. Robinson E. L. Moorhead B. J. Weathers E. A. Ufkes T. M. Vickrey and G.N. Schrauzer J. Amer. Chem. SOC.,1977,99 3657. 127 E. L. Moorehead B. J. Weathers E. A.Ufkes P. R. Robinson and G. N. Schrauzer J. Amer. Chem. SOC.,1977,99 6089. 128 J. Chatt and R. L. Richards J. Less-Common Metals 1977 54 477. 129 J. Chatt A. J. Pearman and R. L. Richards J.C.S. Dalton 1977 2139. 130 J. Chatt A. J. Pearman and R. L. Richards J.C.S. Dalton 1977 1852. 131 G.J. Leigh and C. J. Pickett J.C.S.Dalton 1977 1797. J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall bands assigned in other compounds to fi(0s~N). A similar situation has been noted in rhenium cherni~try.'~' Nitrene-(alkylimido-)complexes also arise from reactions of WC16 and organo-nitriles RCN. On the basis of i.r. and 'Hn.m.r. spectra these are believed to have the structure (29).133 They react with donors L such as Et20 THF and SEt2 to yield complexes [WCI,(NCCI,R)L] which have fi(W=N) at ca.1280 4 Dithiocarbamato- and Related Complexes As promised in last year's Report a more detailed account of the recent literature on dithiocarbamate and related 1,l-dithioacid ligands follows. The last compre- hensive account of the topic appeared in 1970,13' but a more recent review details the preparation and properties of dithiocarbamato-complexes with metals in unusual oxidation The iron(II1) dithiocarbamato-(hereafter abbreviated to dtc) complexes [Fe(dtc),] continue to be studied primarily because they exhibit spin-state cross- over between 6A (high-spin) and 'T2 (low-spin) states. When crystallized from benzene or nitrobenzene [Fe(S2CNB~n2)3] occludes solvent and the magnetic moment at room temperature (3.6B.M.) is markedly lower than in the unsolvated complex (5.3 B.M.).The solvent molecules are too well separated from each other and the dtc ligands for interaction but they do affect the average Fe-S distances. These variations of bond length appear to correlate well with the observed moments the highest moments corresponding to the longest Fe-S distances. 13' The analogous Se2CNR2 complexes also show spin-state equilibria which lie more towards the 2T2state. It had previously been reported that the diselenocarbamato- complexes were exclusively low-spin but the observations that led to this conclusion are now attributed to contamination with [Fe(Se2CNR2)2].138 The preparation of the mixed dtc complexes [Fe(S2CNR2)(S2CNR'2)2] has enabled more reliable calculations of ligand-field parameters to be made on the basis of the observed spectroscopic and magnetic properties.139 The iron(II1) dtc complexes [FeX(dt~)~] also have unusual magnetic properties in possessing an S =$ ground state and this has prompted the preparation of new complexes with X = NCO- NCS- NCSe- from [Fe(dtc),] and the appropriate silver ~a1t.l~' The complex with X=CF3C02- was obtained by dissolution of [Fe(dtc),] in trifluoroacetic acid. The complexes with X = C1 can also be prepared 13* A. 0.Chang K. Oshima and K. B. Sharpless J. Amer. Chem. Soc. 1977.99 3420. 133 G. W. A. Fowles D. A. Rice and K. J. Shanton J.C.S. Dalton 1977 1212. 134 G. W. A. Fowles D. A. Rice and K. J. Shanton J.C.S. Dalton 1977 2129.13' D. Coucouvanis Progr. Inorg. Chem. 1970 11,233. J. Willemse J. A. Cras J. J. Steggerda and C. P. Keijpers Structure and Bonding,1976 28 84. 13' E. J. Bukouskas B. S. Deaver and E. Sinn Inorg. Nuclear Chem. Letters 1977,13 282. 13' D. de Fillipo P. Depalano A. Diaz S. Steffe and E. F. Trogin J.C.S.Dalton 1977 1566. 139 C. A. Tsipis C. C. Hadjikostas and G. E. Manoussakis Inorg. Chim. Acta 1977 23 163. 140 E. A. Pasek and D. K. Straub Inorg. Chim. Acta 1977 21,29. Chemistry of the d- and f-Block Metals 191 by irradiation of [Fe(dtc),] in a halogenated solvent. This reaction is free-radical in character and has been studied in detail for dtc = S2CN(CH2Ph)2. It is believed to proceed via excited-state weakening of the bonding of one dtc ligand followed by interaction with the halogenated s01vent.l~~ Analogous irradiation of the ruthenium complex [Ru(S2CNEt2),] in CHCI or CH2C12 at 265 nm produces both [RuC1(S2CNEt2),] and [Ru~(S~CNE~~)~]C~.142 The latter complex can also be prepared by oxidation of [Ru(S2CNEt2),] with dichlorine and its structure was included in last year's Report. If di-iodine is employed as oxidant for [Ru(S2CNMe2),] a complex of stoicheiometry [Ru13(S2CNMe2),] (30) results and an X-ray structure revealed pentagonal-bi- (30) pyramidal [RuI(S~CNM~~)~] units linked into infinite chains by I2 molecules. The central 1-1 distance of 283.4(5)pm is some 13pm longer than in 12 and indi- cative of donor-acceptor interaction with the ligating iodide. 143 The reactions of [Fe(S2CNR2),] with di-iodine give complex mixtures of products in-cluding [Fe12(S2CNR2)2] [Fe(S2CNR2),]13 [R2 =Me2 Et2 or (CH2),] and for R2= Et2 [Fe13(S2CNR2)2].The structure of the [Fe12(S2CNR2)2] complexes may well be analogous to (30) with I2 molecules linking [FeI(S,CNR,),] units.'44 If higher oxidation states of the metal are not readily accessible treatment with dihalogen results in oxidation of the dithiocarbamate ligands. This is illus- trated by the oxidation of [Zn(S2CNBun2),] with di-iodine to give [Zn12(Bun2NCSS2CSNBun2)], containing ligating tetra-n-butylthiuram disulphide. A stopped-flow study of the reaction kinetics suggests that the reaction involves initial formation of the adduct [Zn(S2CNBu"2)2]12; the I2 interacts with a sulphur atom of the dithiocarbamate ligand.',' One of the dtc ligands of [Fe(S2CNEt2),] can also be displaced by neutral molecules L to give [Fe(S2CNEt2)2L2]+ (L = p-ClC6H4NC or iPh2PCH2CH2PPh2).Cyclic voltammetric studies at a platinum electrode show that these can be reversibly reduced to the neutral species [Fe(S2CNEt2)2L2]. The complex [Fe(S2CNEt2)2(p-C1C6H4NC)2] reacts with zinc iodide and the structure of the product which is shown in Figure 1,is such that zinc is bonded to the sulphur atoms 14' P.-H. Lin and J. I. Zink J. Amer. Chem. Soc. 1977 99 2155. 14* K. W. Given B. M. Mattson M. F. McGuiggan G. L. Miessler and L. H. Pignolet J. Amer. Chem. SOC.,1977,99,4855. 143 B. M. Mattson and L. H. Pignolet Inorg. Chem. 1977 16 488. 144 E.A. Pasek and D. K. Straub Znorg. Chim. Actu 1977 21 23. 145 H. Kita K. Tanaka and T. Tanaka Znorg. Chim. Actu 1977 21,229. J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall Figure 1 The strucfure of [Fe{CN( p-CIC6H4)}~(S2CNEt2)2Zn12] (Reproduced from J.C.S. Dalton 1977 359) of the dithiocarbamate ligand.'46 The complex [Cr(S2CNEt2)3] is isostructural with its iron and ruthenium analogues (discussed above) and it has an average Cr-S distance of 239(6) pm.147 An interesting related complex is prepared by the reac- tion of K2[Cr207] with sodium diethyldithiocarbamate in aqueous solution. The structure shown in Figure 2 was confirmed by X-ray crystallography and provides the first example of insertion of oxygen into a dtc ligand 148 although analogous insertion of sulphur was reported some years A systematic study of the mass spectra of dtc complexes of Cr Fe Co Ru and Rh showed molecular ions for all the complexes,15o but at widely different relative abundances.The major decombosition pathways involve loss of the dithiocarbamate ligand radical and loss of S Sz and SCNRz groups. For the dithiolen-substituted complex [Fe{SzCz(CF3)2}(S2CNRz)z], the dithiolen ligand is lost preferentially to the dtc ligands. A caveat for the mass spectral analysis of dtc complexes is provided by the presence of peaks due to [FeH(dtc)]' and [Fe(dtc),]' in the spectra of even carefully purified Na(dtc). The iron is evidently being chelated from the spectrometer by the dtc ligand. Another physical technique applied to dtc complexes has been U.V.He(1) photoelectron spectroscopy. The tris(dtc) complexes of Cr Fe and Co show distinct bands corresponding to d-orbital ionization and bands in the region 7.5-9.5 eV are believed to arise from orbitals principally derived from sulphur 3p ~rbitals.'~~.~~~ The first reported examples of osmium dtc complexes without tertiary phosphine ligands are provided by the preparation of [OS(S~CNE~~)~] (31)from [NH4]2[OsC16] 146 J. A. McCleverty S. McLuckie N. J. Morrison N. A. Bailey and N. W. Walker J.C.S. Dalron 1977 359. 147 C. L. Raston and A. H. White Ausrral. J. Chem. 1977 30 2091. 148 J. M. Hope R. L. Martin D. Taylor and A. H. White J.C.S. Chem. Comm. 1977 99. 149 D. Coucouvanis and J. P. Fackler J.Amer. Chem. SOC.,1967,89 1346. 150 K. W. Given B. M. Mattson G. L. Miessler and L. H. Pignolet J. Inorg. Nuclear Chem. 1977 39 1309. 151 C. Cauletti and C. Furlani J.C.S. Dalton 1977 1068. 152 C. Cauletti and C. Furlani Inorg. Chim. Am 1977 23 181. Chemistry of the d- and f-Block Metals Figure 2 The molecular structure of [Cr(S2CNEt2)2(02SCNEt2)], showing the ellipsoids of 50% probability (Reproduced from J.C.S. Chem. Comm. 1977 99) and excess Na(dtc). The i.r. spectrum of (31)suggests the presence of a unidentate dtc ligand and on heating under reflux in THF [OS(S~CNE~~)~] (32) is formed. The cyclic voltammetric behaviour of (31) at a platinum electrode is very complex whereas that of (32) is more straightforward showing two oxidation waves and one reduction wave all being reversible.153 Surprisingly the reaction of OsC13,xH20 with Na[S2CNR2] (R=Me or Et) in acetonitrile under reflux gives the p-nitrido- complexes [Os2N(S2CNR2),] in 70% yield. An X-ray crystal structure of the complex (33) showed the presence of both symmetrically bridging nitride and dtc Me Me \I N (33) 153 A. H. Dix J. W. Diesveld and J. G. M. van der Linden Znorg. Chirn. Actu 1977 24 L51. J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall ligand~.’~~ The bridging dtc ligand imposes a distortion on the 0s-N-0s system to an 0s-N-0s angle of 165(2)”. The 0s-0s separation of 349 pm provides the longest reported span for 1’1-dithiolate ligands. The origin of the nitride nitrogen has not yet been determined but it may arise from the excess dtc present in the preparative reaction.The chemistry of complexes of molybdenum with dtc continues to be an active area the emphasis being on dimeric complexes with a variety of bridging groups. The reaction of [Mo,O,(S,CNEt,)J with thiphenol produces a red-orange diamagnetic complex [Mo,O~(SP~)~(S,CNE~~)~], which has been shown by X-ray crystallography to have structure (34). One SPh group is trans to a terminal 0x0-groups whereas the other is cis and the two Mo-S bridging distances differ by about 20~m.l~~ Trace amounts of water cause (34) to revert to the starting I / Et Et 0x0-complex and halogenated solvents react to give [Mo2(p2-C1)(p2-SPh)2(S2CNEt2)][MoOC14(H20)], the structure of which was discussed in last year’s Report.The asymmetric dinuclear complex shown in Figure 3 is formed on reaction of [MoO~(S,CNE~~)~] with the hydrazines PhCYNHNH (Y = 0 or S) in methanol. The structure shows one molybdenum atom to have distorted square-pyramidal co-ordination with an apical oxygen whereas the geometry about the other is approximately trigonal-prismatic. The dimers undergo a reversible one-electron reduction at a potential about 0.3 V more positive for the derivative with Y = S than that with Y = O.Is6The synthesis of [Mo(S,CNR,)~(CO)~(PP~,),] was report- ed sometime ago’” and now the tungsten analogues have also been ~repared.’~’ Both are derived from the reaction of [MC12(C0)2(PPh3b](M= Mo or W) with excess Na(dtc) an improved synthesis of the tungsten dichlorobis(carbony1) complex also being reported.The ability of dithiocarbamate ligands to stabilize high oxidation states has been attributed to their ability to delocalize positive charge oia the resonance structure Is4 K. W. Given and L. H. Pignolet Znorg. Chem. 1977,16 2982. K. Yamanouchi J. H. Enemark J. W. McDonald and W. E. Newton J. Amer. Chem. Soc. 1977.99 3529. lS6 M. W. Bishop J. Chatt J. R. Dilworth G. Kaufman S. Kim and J. Zubieta J.C.S. Chem. Comm. 1977 70. 15’ R Colton and B. Tomkins Austral. J. Chem. 1970 23 lill. lS8 C. J. Chen R. 0.Yelton and J. W. McDonald Inorg. Chim. Ada 1977 22 249. Chemistry of the d-and f-BlockMetals Figure 3 The structure of [(MO(S~CNE~~)(P~CON~)}~O], which is obtuinedfrorn the reuction of [MoO~(S~CNE~~)~] with PhCONHNH2 (Reproduced from J.C.S.Chern.Cornm. 1977 70) (35). The presence of i.r. bands in the 1500-1600 cm-' region suggests that such structures are important for most dtc complexes. Two apparently independent studie~~~~.'~' of the pyrrolyldithiocarbamate ligand (36) indicate that there is very little contribution of resonance forms such as (35) with C-N multiple bonding R S-\+ / N=C / \s-R because this would require disruption of the aromatic character of the pyrrole ring. Cyclic voltammetric studies of [Fe(S2CNC4H4)3] show that it is easier to reduce and harder to oxidize than its counterparts with other dtc's indicating its enhanced stabilization of lower oxidation states.159 An e.p.r. study of [Cu(S2CNC4H4)2] produced spectral parameters consistent with a very covalent u and w metal-sulphur bond. 160 As the copper atom in copper dithiocarbamate complexes can have oxidation states of 1 2 or 3 and polynuclear derivatives with mixed oxidation states are also known,136 the chemistry in this area is extensive and continues to be studied. can The complex [{CUCI(S~CNE~~))~] be prepared either by the reaction or from of [Cu(S2CNEt2)2] with CUC~~,~H~O'~~anhydrous CuC12 and PhCOS2CNEt2.162 The benzoylated dtc represents a convenient source of anhy-drous dithiocarbamate because the sodium dtc salts are extremely difficult to A. G. El A'rnrna and R. S. Drago Inorg. Chem. 1977,16 2975. I6O R. D. Berernan and D.Nalewajek Inorg. Chem. 1977 16 2687. '" A. R. Hendrickson R. L. Martin and D. Taylor J.C.S. Chem. Comm. 1975 843. ''* C. G. R. Nair and K. K. M. Yusuff J. Znorg. Nuclear Chem. 1977,39 281. 196 J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall dehydrate. An X-ray structure161 of the dimer shows that in the solid state weak association into tetramers uia intermolecular Cu-S and Cu-Cl interactions occurs. The magnetic properties have been studied from 4.2 to 300 K and inter- preted in terms of an isotropic exchange interaction between each of the four copper An e.p.r. study of the redistribution reaction between and shows [CU(S~CNR~~)~][CU(S~CNR~~)~] that a statistical distribution of products results. However the analogous equilibrium of [Cu(S2CNR2),] and [CU{S~C~(CF~)~}~]~-gives exclusively the mixed product [Cu(S2CNR2)- {S2C2(CF3)2)1-.164 The phosphorodithioate (R2PS2-; R = alkyl aryl or alkoxy) xanthate (ROCS2-) or thioxanthate (Etrithiocarbonate) (RSCS2-) ligands are analogous to dithiocar- bamate.However as in the case of pyrrolyl-dtc they cannot delocalize positive charge by forming exocyclic multiple bonds and they tend to be stronger reducing agents than dithiocarbamates and to stabilize lower oxidation states. 31P n.m.r. spectroscopy has been applied to the determination of the bonding modes (37) (38) (39) and (40) of phosphorodithioate ligands for a range of substituents on phosphorus and for several The observed "P chemical-shift (relative to H3PO4) regions are >107p.p.m.for (39) <82p.p.m. for (40) and 87 and 101 p.p.m. for (39) and (40) which cannot be distinguished. The phosphorus atom is also a potential ligating atom and a structure of (Ni(S,P(chex),)J S S S-M \pH \P' \M \P/ / \s/ /\S-M / \S-M (37) (38) (39) (40) (chex = cyclohexyl) shows P,S-bonded phosphorodithioate ligands. The complex is prepared by direct reaction between the metal halide CS2 [PH(chex),] and base.166 Chelated bis( 1,l-dithioacid) ligands if bound to platinum or palladium are relatively labile and they can be forced into unidentate bonding or displaced altogether by the reaction of the complex with neutral donors such as tertiary phosphines. Prolonged reaction of [Pd(S2PMe2)2]with excess PPh20R (R = Me or Et) in benzene gives [Pd(S2PMe2)(PPh20)(PPh20H)](41) which has been shown by X-ray structure analysis to contain the symmetric hydrogen-bonded Ph2P0-* * H * * * OPPh2 ligand.An analogous complex with dtc can be prepared Ph2 Ph2 (41) 163 P. D. W. Boyd and R. L. Martin J.C.S. Dalton 1977 105. 164 W. Dietzsch J. Reinhold K. Kirmse E. Hoyer and I. N. Marov J. Znorg. Nuclear Chem. 1977 39 1377. C. Glidewell Znorg. Chim. Actu 1977 25 159. 166 E. G. Moers D. H. M. W. Thewissen and J. J. Steggerda J. Znorg. Nuclear Chem. 1977 39 1321. Chemistry of the d- and f-Block Metals and the mechanism is believed to involve attack by displaced 1,l-dithioate ligand on the alkoxy-group of a co-ordinated diphenylphosphinite ligand with the formation of a dithioacid The OR group of co-ordinated xanthate can undergo similar nucleophilic attack as in the reaction of [Pt(S2COR)2] with K[S2COR].Here the product is dependent on R; when R=Me or CHZPh [Pt(S2COR),(S2CO>]- and [Pt(S2CO)2]2- are formed whereas with R =Pr’ or Et [Pt(S2COR),] is the main species generated. The last complex has a structure consisting of two unidentate and one bidentate xanthate ligands as shown in Figure 4,168 whereas the analogous nickel complex on the basis of spectroscopic analysis is believed to contain three bidentate dithioacid ligand~.’~~ Figure 4 The structure of the anion [Pt(S2COEth]-(Reproduced from J.C.S. Dalton 1977,496) 5 Iron-Sulphur Cluster Complexes Much of the research in this area has originated from the laboratory of R.H. Holm and his recent review”’ makes a detailed comparison of the synthetic iron-sulphur clusters with those extant in non-haem iron proteins. This topic was last reported in 1975 and the following describes developments that have been made since then. The iron(II1) complex [Fe(S2-o-xyl)2]- (S2-0-xyl =o-xylyl-a,a-dithiolate) is a viable model for the oxidized rubredoxin (Rd,,) isolated from C. pasteurianum at 167 M. C. Cornock R. 0.Gould C. L. Jones andT. A. Stephenson J.C.S. Dalton 1977,1307. 168 M. C. Cornock R. 0.Gould C. L. Jones J. D. Owen D. F. Steele and T. A. Stephenson J.C.S. Dalton 1977,496. 169 D. Coucouvanis and J. P. Fackler Znorg. Chem. 1967,16,2047. 170 R. H. Holm in ‘Biological Aspects of Inorganic Chemistry’ ed. A. W.Addison W. R. Cullen D. Dolphin and B. R. James John Wiley and Sons New York London Sydney and Ontario 1977 p. 71. 198 J. R. Dilworth G.J. Leigh R. L. Richards and K. W. Bagnall least in terms of spectroscopic and magnetic properties. Although the X-ray structure shows that the iron sites in the synthetic and biological systems are related the Rd, site (42) appears to be more distorted with a very short Fe-S bond. However the structure of the protein is not yet highly refined making precise comparison difficult. The structure of the iron(”) complex [Fe(S2-o-xyl)2]2- has also been determined and shows that the average Fe-S distance increases by about 10 pm on reduction.I7’ Ab initio MO calculations on the hypothetical [Fe(SH)4]- produce the S = and S = 2 ground states found by magnetic measure- ments for the oxidized and reduced forms respectively of r~bredoxin.’~~ The structure of the biological counterpart of (Fe2S2(S2-o-~y1)2]2- has still not been determined but the close resemblance of magnetic and spectroscopic properties indicates that the synthetic cluster is a good model for the two-iron ferredoxins [2Fe-Fd,,] (43).The conditions required for interconversions of FeS, Fe2S2 and Fe4S4 clusters have now been delineated and are summarized in Scheme 3 which illustrates the S2-0-xyl 02 [FeCl4I2-(Fe(S2-o-xyl)21* -[Fe(S2-o-xyl)2]--1 02 v Rdred 3 NaHS. Rdox NaOMe [Fe2S2(S2-o-~y1)2]4-___L [Fe2S2(S2-o-xyl)2]3-[F~~S~(S~-O-XYI)~]~-173 v -1 49v 2 Fe -Fdred 2Fe- Fd, (1)PhSH (II)~MSO-H~O I [Fe4S,(SPh),l4‘-W [Fe4S4(SPh),13-[Fe4S,(SPh),l2--1 75v -1 04 v 4Fe-Fdred 4 Fe -Fd, Scheme 3 systematic build-up of Fe4S4 clusters from tetrachloroferrate(~~).” The equivalent biological systems where they exist are indicated.Mononuclear [Fe(S2-o-xyl)2]- is smoothly converted (in yield) into the dinuclear derivative by treatment with NaHS and NaOMe in methanol. Ligand exchange to the thiophenolate derivative [Fe2S2(SPh),I2-’ is necessary prior to dimerization to the [Fe4S4(SPh),I2- cluster in aqueous DMS0.’73 The (S2-o-xyl)-ligated dimer cannot be converted into the tetramer because the bidentate ligands cannot span the faces of the Fe4S4 cubane structure. The ready interconversion of the three Fe-S cores of the ferredoxins (42) (43) and (44) is illustrated by the reaction of iron(II1) chloride and base in DMSO with the tetracysteinyl peptide AcGlyCys(Gly2Cys),G1y2NH2.The spectrum of the 171 R. W. Lane J. A. hers R. B. Frankel G. C. Papaefthymiou and R. H. Holm. J. Amer. Chem. Soc.. 1977.99.84. R. A. Blair and W. A. Goddard. J. Amer. Chem. SOC.. 1977 99. 3505. ”’ J. Cambray. R. W. Lane A. G. Wedd R. W. Johnson. and R. H. Holm Inorg. Chem. 1977,16,2585. 199 Chemistry of the d -and f -Block Metals initial red-violet solution is very similar to that of [Fe(Sz-o-xyl)zJ- and Rd,,. The colour fades rapidly and the spectrum after 5 min is similar to that of [Ee2S2(S2-o- ~yl),]~-and 2Fe-Fd0,. Subsequent addition of sodium sulphide gives a solution with a spectrum characteristic of the [Fe,S,(SR),]’- clusters and the addition of excess thiophenol gives [Fe4S4(SPh),]’- in a yield corresponding to 48% conversion of the original iron(II1) ~hloride.”~ This represents the first direct synthesis of peptide analogues of the ferredoxins as these were previously prepared by exchange reactions of the pre-formed clusters.175 One of the most useful applications of the work on synthetic clusters is the core-extrusion technique for the identification of iron sites within proteins by treatment of the protein with excess thiophenol in a 4:1 HMPA-HzO medium. Any possibility of dinuclear-tetranuclear conversion during extrusion is suppressed by using a 500 molar excess of thiophenol and an aqueous component of pH b8. The use of this quantitative technique for the hydrogenase from C.pasteurianum suggests the presence of three Fe4S4-type iron sites in the ~r0tein.l~~ The presence of interfering chromophores in proteins such as succinate dehydrogenase will require the development of alternatives to thiophenol as the extrusion agent. In non-aqueous media the redox potentials of the analogue Fe4S4 clusters are substantially more negative than the equivalent ferredoxins. However if the [Fe,S,(SR),]*-clusters are rendered water-soluble by introducing the R groups CH2C00-,‘77 CHzCHzOH or Cys(Ac)NHMe the potentials become very similar to those of the ferredoxins. The effect of solvent medium on the redox properties of the water-soluble clusters has now been studied over the range 80% DMSO-HzO to pure Between these limits the El values (relative to the SCE) for the reduction of [Fe4S4(SR),IZ- to [Fe4S4(SR),I3- range from -1.05 to -0.75 V (R = CHzCHzOH) and -0.91 to -0.73 V [R =Cys(Ac)NHMe].The corresponding couple for the ferredoxin from C.pasteurianum decreases from -0.93 to -0.70 V down to 40% DMSO content whence it remains invariant. At this stage of 40% DMSO the protein becomes re-folded and the Fe sites are shielded from solvent effects. The variations of potential in both are attributed to progressive solvation by water as the content of water is As well as undergoing easy exchange with other thiols the RS-ligands of the Fez& and Fe4S4 cores can be replaced by chloride by reaction with benzoyl ~hloride.~’ The X-ray structures of the products [Fe2SZCl4] (45) and [Fe4S4CI4] G.Christou B. Ridge and H. N. Rydon,J.C.S. Chem. Cornm. 1977 908. 175 B. V. de Pamphilis B. A. Averill T. Herskovitz L. Que and R. H. Holm J. Arner. Chem. Soc. 1974 96 6042. 176 W. 0.Gillum L. E. Mortensen J.-S. Chen and R. H. Holm J. Amer. Chem. Sac. 1977 99 584. 177 H. L. Carrell J. P. Glusker R. Job andT. C. Bruice J. Amer. Chem. Soc. 1’977,99 3683. 178 C. L. Hill J. Renaud R. H. Holm and L. E. Mortensen J. Amer. Chern. Soc. 1977,99 2549. 200 J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall have been determined and the unequal Fe-Cl distances observed in (45) were attributed to Van der Waals interaction^."^ The first example of an iron-sulphur cluster functioning as a catalyst is in the reaction of EtSH with isocyanides mediated by [Fe,S4(SEt),l4- to give ethylthio-formimidates in essentially quan- titative yields.The postulated mechanism involves the attack of isocyanide at iron followed by migration of SEt from iron to carbon. Reaction with further EtSH liberates the organic product.'" An X-ray structure of [(q5-CsH5),Fe4S4]Br shows that oxidation from the neu- tral species is accompanied by a shortening of two of the four long non-bonding Fe-Fe distances within the cluster from 336 to 319pm the others remaining approximately constant.lgl The effect is attributed to removal of an electron from a cluster antibonding orbital with a subsequent orthorhombic Jahn-Teller dis-tortion. The structure of the dication'" shows that further oxidation produces a geometry with four short Fe-Fe distances of 283.4(3) pm and two of 330.4(5) pm.These structural results have been used as the basis of a qualitative molecular cluster model which rationalizes the observed geometries and predicts a further shortening of the four short Fe-Fe distances of the dication on oxidation to the as yet unisolated trication. Because to date only dianionic [Fe4S4(SR)4]Z-type clus- ters have been structurally characterized it remains to be seen if analogous geometric changes occur on reduction to [Fe,S,(SR),]'- clusters which can be isolated after reduction of the [Fe,S4(SR),JZ- clusters with sodium a~enaphthide.'~' 6 Models for Copper Proteins Among copper-containing biological systems the best characterized are the so-called blue copper which manifest typical intense u.~.bands at about 600 nm (E =5000) and their e.p.r. spectra have small copper hyperfine splitting constants all of 30-100 G. Such copper sites occur in electron-transfer proteins such as plastocyanins and azurin and in conjunction with other copper sites in the blue oxidases such as laccase. Detailed spectroscopic studies of the proteins themselves and investigations of a number of model cornple~es'~~ have led to a wide variety of proposals for the structure of the copper site. There appears to be a consensus of opinion on the presence of RS- originating from a cysteinyl group but other proposed co-ligands include inter alia imidazole methionine sulphur and phenolate. Structure (46) was proposed on the basis of spectroscopic studies of an extensive series of copper complexes of different geometries with various combinations of nitrogen sulphur and oxygen donor ligands.Ig5 It was not reported if the e.p.r.spectra of these models showed the same small hyperfine splittings as the protein. Several model complexes have been able to reproduce U.V. spectral characteristics of the proteins 179 M. 0.Bobrik K. 0.Hodgson and R. H. Holm Inorg. Chem. 1977.16 1851. A. Schwartz and E. E. van Tarnelen J. Amer. Chem. Soc. 1977,99 3189. T. Toan W. P. Fehlhammer and L. F. Dahl J. Amer. Chem. SOC.,1977,99 402. 18* T. Toan B. K. Teo J. A. Ferguson T. J. Meyer and L. F. Dahl J. Amer. Chem. Soc. 1977 99 408. 183 R. W. Lane A. G. Wedd W. 0. Gillum E.J. Loskowski R. H. Holm R. B. Frankel and G. C. Papaefthymiou J. Amer. Chem. SOC., 1977,99 2350. 184 H. Beinert Co-ordination Chem. Rev. 1977 23 119. A. R. Amudsen J. Whelan and B. Bosnich J. Amer. Chem. SOC.,1977 99 6730 and references therein. Chemistry of the d-and f-Block Metals but as in the case of [CU(HB~Z,}(SC~H,~-N~,)~'~~ (pz =pyrazole) they fail simul- taneously to give appropriate e.p.r. parameters. However complex (47),prepared R S/ H L m (47) L=H200r NvNH on the premise that the blue copper site contains imidazole and cysteinyl sulphur shows both an intense U.V. band at about 600 nm and a small all hyperfine splitting constant of 93 G.lS7 The analysis of the amino-acid sequences of azurin and plastocyanin reveals comparable spacings between cysteine histidine and methionine and on this basis a tetrahedral copper ligated by those three amino- acids has been proposed.188 However the absence of any methionine in the sequence of the blue copper protein stellacyanin would appear to rule it out of the copper site.7 Polynuclear Metal-Sulphur Cluster Complexes Copper(1) forms a prolific number of cluster complexes with sulphur ligands and recent studies have confirmed the following stoicheiometries and geometries Cussl2 (Cu cube) cuss6 (Cu trigonal bipyramid) Cu4Ss and cu4s6 (Cu tetra- hedron). The structure of [CU,(SP~)~]~-, shown in Figure 5 is quite different from the cu& type and can be visualized as arising by removal of an SPh moiety from a tetrahedral Cu4(SPh) unit and replacement by a linear (p 2-SPh)-Cu( p2-SPh) unit.Further examples of Cussl2 clusters were reported this year in [C~~(dts)~]~- where dts = (48),and in [Cu8(ded),]"- in which ded = (49).19'Their structures are analogous to that of [C~~(mnt)~]"-(mnt = [S2CNC(CN)2]2-) and comparison with the c&&cluster structure suggests that the Cu-Cu distances in the cubic array are optimal for maximum attractive and minimal repulsive interaction. When co- ordinated the ded type of ligand can undergo protonation at the dithiolate carbon and addition of four equivalents of acid to [C~~(Bu'ded)~]~- where Bu'ded = (49; But in place of Et gives [C~~~(Bu'ded)~(Bu'dedH)~], the first example of an aggregate that contains 10 copper atoms.191 The reaction of CuC12 with tetra- phenyldithioimidodiphosphinate [L (50)] gives the purple mixed-valence complex Cu3L4 which dissociates in solution to Cu13L3 and is oxidized by halogenated lR6 J.S. Thompson T. J. Marks and J. A. Ibers Proc. Nut. Acad. Sci. U.S.A.,1977,74 3114. Y. Sugiura and Y. Hirayama J. Amer. Chem. SOC.,1977 99 1581. G. McLandon and A. E. Martell J. Inorg. Nuclear Chem. 1977.39 191. lS9 I. G.Dance J.C.S. Chem. Comm. 1976 103. 190 F. J. Hollander and D. Coucouvanis J. Amer. Chem. Soc. 1977 99 6268. D. Coucouvanis D. Swenson N. C. Baenziger R. Redelty and M. L. Caffery J. Amer. Chem. SOC. 1977,99,8096. J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall n Figure 5 The structure of [Cu,(SPh),]’-(Reproduced from J.C.S.Chem. Comm. 1976 103) -S 0 \/ S COZEt c-c \/ II I c=c -S /c-c\o S / ‘CO,Et dts ded solvents toCur4L3. The structure of the last complex was shown by an X-ray structural analysis to belong to the Cu4S6 category and provides the first example of such a structure with a bidentate ligand.19’ The 1,8-dithiolate hexachloronaphtho[ 1,8-cd]- 1,2-dithiole reacts with [Ni(cyclo- octadiene),] in the presence of triphenylphosphine to give complex (5 l) which is a ,Ni -S-1 Ph,P \J 0.Siiman C. P. Huber and M.L. Post Inorg. Chim. Acta 1977 25 Lll. Chemistry of the d -and f-Block Metals rare example of a trinuclear nickel-sulphur ~1uster.l~~ There are few rhenium- sulphur clusters and only the second example of a Re,S cluster is provided by [Re4S4(CN),,l4- prepared by the reaction of [Re(CN),]-with molten KSCN.An X-ray structure of its tetraphenylphosphonium salt showed a cubane-like arrange- ment of the Re,S4 core with three CN groups bound to each Re. The selenium analogue is prepared using a KSeCN melt and is structurally isomorphous. The average Re-Re distances of about 280 pm indicate that there is a high degree of Re-Re bonding.’94 In an attempt to prepare the manganese complex [Mn2( ,u2-s)2(co)8], [Mn2(p-SSnMe3)2(C0)8](52) was oxidized with di-iodine. The analysis and the mass spectrum of the red crystalline product suggested the formula [Mn4S4(CO)15] and the X-ray structure revealed an unexpected linking of three manganese atoms by disulphide as shown in Figure 6.The initial product of oxidation of (52) is MnCO) MnKO) Mn(C0)S Figure 6 The sfrucfure of [Mn4S4(C0)15) (Reproduced from J.C.S. Chem. Comm. 1977,782) probably (53) which relieves the steric strain imposed by its unusual co-ordination at sulphur by dimerization with simultaneous transfer and elimination of co.lg5 193 W. P. Bosman and H. G. M. van der Linden J.C.S. Chem. Comm. 1977 714. M. Laing P. M. Kiernan and W. P. Griffith J.C.S. Chem. Comm. 1977 221. ”’ V. Kiillmer E. Rottinger and H. Vahrenkamp J.C.S. Chem. Comm. 1977 782. J. R. Dilworth G. J. Leigh R. L. Richards and K. W. Bagnall 8 Complexes of Sulphur Dioxide and Related Ligands Sulphur dioxide can either co-ordinate as a Lewis acid in which case the geometry of the M-S02 system is pyramidal or as a cr-donor and r-acceptor the M-SO geometry being planar.Recent structurally confirmed examples of these two co-ordination modes are given by [Pt(S0,),(PPh3)2]'96 (54)(both S02's pyramidal) 19' Rh(q5-C5H5)(q '-C2H4)(S02)] [ calculations on nitrosyl and sulphur dioxide complexes produce similar diagrams for the two formally analogous ligands and permit predictions of the geometry adopted by ligating SO,. Comparisons of tetrahedral SO and NO complexes suggest that SO has a greater tendency to bend owing to the smaller energy separation between the cr* and T* 0rbita1s.l~~ Complex (54) loses SO in toluene solution to form [Pt3(S02)3(PPh3)3] (56) which contains three bridging SO2group the sulphur and phosphorus atoms being nearly coplanar with the triangle of Pt atoms.199 That there is a third mode of bonding for SO2,via sulphur and oxygen has been established with the determination of the structure of [Rh(NO)(SOZ)(PPh3)2] (57).84 Its reactions with 1802 to give a sulphato-complex have been studied and 0 YPh3 0 Ph,P \ /oI Rh Ph3P ,Pt- \s/pt\PPh Ph,P /IN '\o I 0/\ 0 0 (56) (57) the distribution of products that has been established by i.r.spectroscopy could only be rationalized in terms of the unusual square-pyramidal intermediate (58).84 This is in contrast to the oxygenation of [RuC1(NO)(S0,)(PPh3h] with 1802(S-bonded SO,) and the reaction of SO with [IrC1('802)(PPh3)2] where the labelling distribution is consistent with the peroxysulphite intermediate (59).and (55) (SOz planar). Extended Huckel MO 196 D. C. Moody and R. R. Ryan Inorg. Chem. 1976,15 1823. 19' 19' R. R. Ryan P. G. Eller and G. J. Kubas Inorg. Chem. 1976 15,797. R. R. Ryan and P. G. Eller Znorg. Chem. 1976 15 494. 199 D. C. Moody and R. R. Ryan Znorg. Chem. 1977.16 1052. 205 Chemistry of the d- and f-Block Metals 0 \I \ Ru-S / +O 4 "'0-0 ,s-0 /I \o (59) The dependence of the stability and reactivity of co-ordinated SO on other ligands present is illustrated by the series of complexes [Rh(X)(ttp)(SO,)] where X = C1- N3- or CN- and ttp = MeC(CH,PPh,), and [Rh(ttp)L(S02)]' (L = CO PR3 or MeCN). In all the complexes the i.r. spectra suggest pyramidal SOzligands but there is a marked variation in reactivity towards 02,[RhCl(ttp)(S02)] not reacting in solution at 25 "C whereas [Rh(ttp)(MeCN)(SO,)] [AsF,] reacts almost instantaneously."' Co-ordinated SO can undergo nucleophilic attack at sulphur as in the complexes [ML(S020Et)]' [M=Co or Ni L=N(CH2PPh2)3 or P(CH2PPh2)3] prepared by treatment of the hydrated metal salts with L and SO in ethanol.An X-ray structure of the anion (60) of [Ni{N(CH2PPh2)3}(S0,0Et)] [BF,] showed fairly regular trigonal-bipyramidal geometry for the nickel with the S03Et ligand in an apical site.201 When functioning as a Lewis acid SO can attack at sites other than a metal as in the complexes [CU~I~(PP~~M~)~(SO~)]~'~ and [Cu(PMe2Ph),(SPh)(SOz)],zo3where X-ray structures showed the SOz to be attached to bridging iodide and mercaptide sulphur respectively.Both 0-and S-bonding of the thiosulphate ligand in [CO(S~O~)(NH~)~]' have been proposed but an X-ray structure has now shown it to be S-b~nded.~' Reduction of the complex by Cr" is then envisaged as proceeding via attack of Cr" at oxygen. Another reported reaction involving thiosulphate is with [C~(bipy)~]Cl~ which unexpectedly produces a trithionato-complex. An X-ray structure deter- mination revealed an infinite structure (Figure 7) with S3062-ions linking [Cu(bipy),]'+ units.205 An oxosulphur anion is also formed in the reaction of [Pt(O2)(PPh3),] with PhNSO in the presence of 02,PPh3 and moisture. The reaction yields [Pt2(0H)(PPh3),][S,O8] and aniline sulphate the persulphate anion being identified by its characteristic i.r.bands. The dimer is degraded to [Pt(S04)- (PPh3)*] in polar solvents,206 but the mechanisms operative in these and the pre- ceding reaction are far from clear. 2oo P. C. Blum and D. W. Meek fnorg. Chim. Acta 1977,24 675. 201 R. J. Restivo G.Ferguson and R. J. Balahura fnorg. Chem. 1977 16,167. 202 P.G. Eller G. J. Kubas and R. R. Ryan fnorg.Chem. 1977 16,2454. 203 P. G. Eller and G.J. Kubas J. Amer. Chem. Soc. 1977 99 4346. 204 R. J. Restivo G. Ferguson and R. J. Balahura fnorg. Chem. 1977 16,167. *OS M.B. Ferrari G. G. Fava and C. Pelizzi J.C.S. Chem. Comm. 1977 8. *06 C. La Monica and S. Cenini fnorg. Chim. Acta 1977 24 L17. J. R. Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall Figure 7 The structure of [C~(bipy)~(S~O~)~].projected on to the (010)plane. (Reproduced from J.C.S. Chem. Comm. 1977,8) Chemistry of the d-and f-BlockMetals 9 Hydrosulphido-complexes Hydrosulphido-(SH) complexes are few and far between; attempts to synthesize them have frequently led to polymeric sulphide-bridged materials or to binary metal sulphides. One of the earliest prepared was [Cr(SH)(H20)5][S04],'07 and its formulation as a monomer has now been confirmed by solution Raman studies and by the determination of the oxidation state of the metal.2os Oxidation with [Fe(H20)6]3' gives both [(H20)5CrS2Fe(H20)5]3'(61) and [(H20)5CrS2Cr(H20)5]4+ (62) the relative amounts depending on the pH. At acid concentrations >0.1 mol l-' [(H20)5CrS2HFe(H20)514f (63) is the major prod- uct.The further reaction of (61) with [Fe(H20),I2' to give (63) is also found and is surprising in view of the usual resistance of Cr"' to substitution.209 The synthesis of both mono- and bis-hydrosulphido-complexes of rhodium has now been described.210 The macrocyclic rhodium complex (64) undergoes an BF 0' '0 y;,i\J \ /Rh\ / u (641 oxidative addition reaction with one equivalent of H2S to give [RhH(SH)L] (L =macrocyclic ligand). The Rh-H stretching frequency appears in the i.r. as a strong band at 1910cm-' but v(S-H) is too weak to detect. The bis(hydro- su1phido)-complex [Rh(SH),L] (65) was formed by the reaction of [RhC12L] with Na[SGeEt3] in acetonitrile. The presumed intermediate [Rh(SGeEt,),L] under-goes hydrolysis to the final product.210 Complex (65) is surprisingly resistant to aerial oxidation both in the solid state and in solution.These Rh hydrosulphido- complexes may well find use as synthetic intermediates for a range of rhodium- sulphur derivatives. The stability of the above rhodium SH-complexes is attributable to the stability and steric requirements of the macrocyclic ligand and the full utilization of the metal orbitals in bonding preventing degradation by ligand dissociation and the formation of additional metal-sulphur bonds. The quadridentate ligands N(CH2CH2PPh2),(=np,) and P(CH2PPh2) (=pp3) function similarly and permit the preparation of ['M(SH)(pp3)][BPh4] (66) (M = Fe Co or Ni) and [M(SH) (pp3)] (67) (M = Co or Ni) by the reaction of H2Swith hydrated metal salts in DMF in the presence of np or pp3."' Attempts to prepare [Fe(SH)(np,)][BPh,] only gave iron sulphide.In the absence of H2S [M(H2O)LI2' are formed initially and depro- tonated by excess ligand to [M(OH)L]'. The SH complexes are probably formed analogously via [M(H2S)LI2' followed by deprotonation. Complexes (66) and (67) 207 M. Ardon and H. Taube J. Amer. Chem. SOC.,1967 89 3661. 208 T. Ramasami and A. G. Sykes Znorg. Chem. 1976 15 1010. 209 T. Ramasami R. S. Taylor and A. G. Sykes Znorg. Chem. 1977,16 1931. 'lo J. P. Collrnan R. K. Rothrock and R. A. Stark Znorg. Chem. 1977 16 437. 211 M. Di Voira. S. Midollini and L. Sacconi Znorg. Chem. 1977 16 1519. J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall showed no i.r.bands due to V(S-H) but an X-ray structure of [Ni(SH)(pp3)] [BPh4] confirmed the presence of the SH group.211 The co-ordination of nickel is approximately trigonal bipyramidal with the metal displaced 19.6 pm below the equatorial plane towards the sulphur atom. 10 Photolytic Properties of [Ru(2,2'-bipyridyl),l2' and Related Complexes The photoredox properties of [Ru(bipy),I2' are of high current interest because of the ability of its lowest energy transfer exited state [Ru(bipy),12+* to act as a donor or acceptor of electrons in redox reactions. It is hoped to combine this property with the oxidizing ability of [R~(bipy)~]~+ effect the solar conversion of to water into dihydrogen and dioxygen. Last year [see Annual Reports (A),Vol. 73 19761 monolayer assemblies of the related complexes [Ru(bipy),{4,4'-bis(C02Cl,H,,)-2,2-bipyridyl),] deposited on glass slides were reported to catalyse the photodecomposition of water.212 Unfortunately it appears that the 4',4'- dicarboxy ligand used contained quantities of impurities their concentration and nature not being reproducible but these were essential for the photolytic reaction.Attempts to repeat this reaction in the original213 and other214 laboratories with pure materials were unsuccessful. Nevertheless examination of the quenching of the [Ru(bipy),I2+* complex with a variety of substrates which are themselves excited reduced or oxidized has been actively pursued with the aim of eventually producing a storage system for solar energy that is based on the decomposition of water.Thus [Ru(bipy),12+* is quenched by dioxygen which is converted into its singlet excited The complexes [N4Co( p-02)(p-NH2)CoN4I4+ (N = NH3 or iH2NCH2CH2NH2) quench [Ru(bipy),I2'* and are reduced by one electron which is transferred to the superoxide bridging group rather than to the metal centre.'16 Quenching is rapid by [Fe(H20)6I3+ to give217 [Fe(H2o)6l2+ and by low concentrations of cu" ions in the presence of poly(viny1 sulphate).218 Reductive quenching of [Ru(bipy),I2+* can also occur to give [Ru(bipyX]' which is a reducing agent capable of releasing dihydrogen from water. Thus the photolysis of hydrophobic analogues of [Ru(bipy),I2' in dry acetonitrile has allowed the isolation of [Ru(bipyR),]' (R = various carboxylate s~bstituents).~~~ The complex [Ru(bipy),]+ has been independently prepared and characterized by electrolytic reduction of [Ru(bipy),I2+ in acetonitrile.It was then observed to be generated during the photolysis of [Ru(bipy),I2+in acetonitrile and to produce 02-by interaction with dioxygen in the presence of dimethylaniline.220 Thus in prin-'I2 G. Sprintschnik H. W. Sprintschnik P. P. Kirsh and D. G. Whitten J. Amer. Chem. Soc. 1976 98 2337. '13 G. Sprintschnik H. W. Sprintschnik P. P. Kirsch and D. G. Whitten J. Amer. Chem. Soc. 1977.99 4947. 'I4 S. Valenty and G. L. Gaines J. Amer. Chem. Soc. 1977,99 1285; A. Harrirnan J.C.S. Chem. Comm. 1977 777. *15 J. N. Demas E. W. Harris and R. P. McBride J. Amer. Chem. SOC.,1977,99 3547. 'I6 K. Chandrasekaran and P.Natarajan J.C.S. Chem. Comm. 1977 774. *" R. C. Young R. F. Kasne and T. J. Meyer J. Amer. Chem. SOC.,1977 99 2468. '"D. Meisel and D. S. Matheson J. Amer. Chem. SOC.,1977,99 6577. 'I9 P. J. Delaive J. T. Lee H. W. Sprintschnik H. Abruna T. J. Meyer and D. G. Whitten J. Amer. Chem. SOC.,1977.99 7094. "" C. P. Anderson D. J. Salmon T. J. Meyer and R. C. Young J. Amer. Chem. SOC.,1977,99 1180. Chemistry of the d-and f-Block Metals 209 ciple the component parts of a ruthenium-catalysed photolytic decomposition of water appear to be available but their combination in the required manner has not yet been realized. PART 11 Scandium Yttrium the Lanthanides and the Actinides By K. W. Bagnall A further volume of the Gmelin series covering scandium yttrium lanthanum and the lanthanides has appeared;’ this volume deals with the separation of the ele- ments from each other the preparation of the mctals and their applications as well as the toxicology of the elements.An extensive review of recent advances in the chemistry of the lanthanides in their less common oxidation states has also appeared .2 1 Scandium The structure of ScCl prepared by reaction of metallic scandium with the ScCll.’ phase .or better with ScC13 at 800”C has been determined. This is a sheet structure consisting of close-packed homoatomic layers (Cl-Sc-Sc-Cl along [OOl]) with antiprismatic co-ordination of the metal atoms; the short Sc-Sc distances indicate strong Sc-Sc bonding between the layers and less strong Sc-Sc bonding within the layer.3 The chloride of composition Sc7ClIo is transported to the hot zone when metallic scandium is heated with ScC13 under a temperature gradient (800-900°C); the infinite-chain structure is made up of two parallel chains of scandium octahedra which share a common edge and chlorine atoms cap all outward-facing metal triangular faces bridging to and between isolated Sc”’ ions; the compound can then be described as [(SCC~~~~C~~~~),(SC~C~~C~~~~)~] .In the structure’ of the THF (tetrahydrofuran) solvate of the tetrahydroborate SC(BH,)~,~THF the metal atom co-ordination is trigonal bipyramidal with the three BH groups in the equatorial plane and the two bonded THF molecules occupying axial positions. A structural study6 of the hydroxoacetate SC(HOCH~CO~)~,~H~O has shown that the compound is more correctly formulated as [SC(HOCH~CO~)~(OH~)~]+ [Sc(HOCH,CO),),]-; in both ions the metal atom is eight-co-ordinate in a distorted dodecahedra1 arrangement.2 Yttrium and the Lanthanides Hydrothermal ageing of the gel obtained by treating a lanthanide nitrate with aqueous ammonia (mole ratio 1 2) at pH 7.0 in the mother liquor yields crystalline products; in the erbium compound of composition Er402(OH)8,HN03 the unit ’ ‘Gmelin Handbuch der Anorganischen Chemie’ System-Nr. 39 ‘Seltenerdelemente’ Teil B ‘Die Elemente’ Lief. 2 Springer Verlag Berlin 1976. ’ D. A. Johnson Adv. Inorg. Chem. Radiochem. 1977,20 1. K. P. Poeppelmeier and J. D. Corbett Inorg. Chem. 1977 16 294. K.P. Poeppelmeier and J. D. Corbett ibid. 1977 16 1107. E. B. Lobkovskii S. E. Kravchenko and N. N. Semenenko Zhur. srrukt. Khim. 1977,18,389. A. S. Antsyshkina L. M. Dikareva and M. A. Porai-Koshits Tezisy Doklady Vses. Chugaevskoe Soveshch. Khim. Kompleksn. Soedinenii 12th. 1975 Vol. 2 241 (Chem. Abs. 1977,86 10 871). J. R. Dilworth G.J. Leigh R. L. Richards and K. W.Bagnall cell contains eight erbium atoms (two formula units) two of which are each bonded to nine OH groups in a tricapped trigonal-prismatic arrangement with the metal atom approximately central and the others bonded to five OH groups and two oxygen atoms in a singly capped trigonal-prismatic array with the metal atom displaced towards the cap. The Y Dy Ho Tm Yb and Lu analogues are i~omorphous.~ The previously described HoC12 14 phase has now been structurally identified' as Ho5Clll isostructural with Dy5CI11.The arrangement is best understood as a one-dimensional superstructure of the fluorite type built from five basic fluorite units with four additional anions per unit cell. In order to accommodate the latter half of the primitive cubic anion packing is transformed into closest packing.' In the enneahydrates [M(H20)9X3] [M =Pr or Yb X = Br03or EtSO,] the metal atoms are at the centre of a tricapped trigonal-prismatic arrangement of water 0 atoms with D3,,symmetry in the case of the bromates and C3hsymmetry in the ethyl ~ulphates.~ A number of hydrated lanthanide perrhenates M(Ke04)3,4H20 (M=Ho Er Tm Yb Lu or Y) have been reported and in the ytterbium compound the metal atom is eight-co-ordinate in a distorted tetragonal antiprism bridged by Re04 tetrahedra; the compounds are therefore more correctly written as [Yb2(Re04)6(H20)6],,2nH20.10 In the structure of the oxodiacetato-complex Na3[Ce{O(CH2C02)2}3],2NaC104,-6H20the metal atom in the cerate(Ir1) anion is surrounded by nine carboxylate and ether 0 atoms in a slightly distorted tricapped trigonal-prismatic array," an arrangement found also in the erbate(II1) anion in Er[Er{O(CH2C02)2}3],6H20 ['Er2{0(CH2C02)2}3,6H20']; the other erbium atom is eight-co-ordinate sur-rounded by a distorted square antiprism of 0 atoms made up by four atoms from the outer carboxylate groups of the four nearest erbate(II1) anions together with the 0 atoms of four water molecules.'2 The hydrated NN'-dimethylurea complex [Er{oC(NHMe)2}6(oH2)](C104)3 provides an example of seven-co-ordination for erbium(II1); this is apparently only the second example to be reported the geometry being a deformed pentagonal bipyramid.13 In the 1,%naphthyridine [napy (l)] complex [Pr(napy)6](C104)3 all six ligands are bidentate the twelve atoms forming a distorted icosahedron about the metal atom.14 Complexes with 2,7-dimethyl-1,8-naphthyridine(dmnapy) of composition [M(pd),(dmnapy)] (M = Pr-Yb pd = pentane-2,4-dionate) have been ' H.A. Wolcott W. 0.Milligan and G. W. Beall J. Inorg. Nuclear Chem. 1977,39,59. U. Lochner H. Barnighausen and J. D. Corbett Inorg. Chem. 1977.16 2134. J. Albertsson and I.Elding Acta Crysr. 1977 B33 1460. lo E. D. Bakhareva M. B. Varfolomeev V. P. Mashonkin and V. V. Ilyukhin Koord. Khim. 1976,2 1135. J. Albertsson and I. Elding Acta Chem. Scand. 1977 A31 21. I. Elding Acfa Chem. Scand. 1977 A31 75. M. C.Mattos E. Surcouf and J.-P. Mornon Acta Cryst.,1977 B33 1855. l4 A. Clearfield R. Gopal and R. W. Olsen Inorg. Chem. 1977 16 911. Chemistry of the d- and f-Block Metals 211 reported and dodecahedra1 co-ordination geometry has been suggested for them." A study of the thermal decomposition of the hydrazine complexes of the (n lanthanide oxalates M2(C204)3,4N2H4,nH20 = 2-6 M =Tb Dy Ho or Y) and M2(C204)3,3N2H4,nH20 (n= 2-5 M = Yb or Lu) has appeared.16 The results of this work weie taken to indicate that in the first group the NZH4 molecules occupied seven co-ordination sites whereas in the second group they occupied four sites;16 it would certainly be interesting to have crystallographic confirmation of these conclusions.Structural information for a number of sulphur donor complexes of the lanthanides is now available; in the complex anion of the salt [PPh,] [PI-(S~PM~~)~] the metal atom is co-ordinated to eight S atoms in a distorted tetragonal-antipris- matic arrangement;17 in the complexes [La{SzP(OEt),}3(PPh30)2] and [Sm{SzP(OEt)2}3(PPh30)3] the metal atom in the La complex is eight co-ordinate with square-antiprismatic geometry whereas the samarium complex is cationic [Sm{S2P(OEt)2}2(PPh30)3]+ [SzP(OEt)z]- and the samarium atom is seven-co-ordinate (four S and three 0 atoms) with pentagonal-bipyramidal geometry (four S and one 0 atom in the pentagonal plane)." In the tris-complexes [M{(C6H11)2PS2}3] (M = Pr or Sm) the metal atoms are co-ordinated to six S atoms at the corners of a distorted trigonal prism.'' The electronic spectra of a series of 00'-diethyldithiophosphato (ddtp) complexes [NEt4] [M(ddtp),] (M = La-Eu Tb or Ho) have also been reported.*' 3 The Actinides A spectrophotometric study of the hydrolytic behaviour of Npv" over the pH range 1-14 has shown that between pH 1 and pH 3-4 the neptunium is present as a cation (NpO,') while at pH 4-5 the hydroxide is formed and above pH 5 further hydrolysis to an anionic species occurs.21 It has also been established that oxidation of Npv' to Npv" by ozonized air occurs at pH 8.2 or above.22 Solutions containing s 8 g Np"" 1-' can be obtained by ozone oxidation of Npv' in 2-3 M-KOH and even higher concentrations of Npv" can be achieved by oxidation in 2-3 M-LiOH.23 The oxidation is a diffusion-controlled zero-order reaction and neptunium (also plutonium) is oxidized by the decomposition products of ozone most probably O-.23 The preparation of AmrV,CmrV,and CfIV in solution by oxidation of the tri- positive ions with K2[S208] in the presence of a tungstophosphate (K1~P2W17061) has been reported;24 AmIV was also prepared by electrochemical oxidation in this medium and the Am'V/Aml*' potential at 25 "C in this solution is reported to be l5 M.Ng See H. W. Latz and D. G. Hendricker J.Inorg. Nuclear Chem. 1977 39 71. 16 V. A. Sharov G. V. Bezdenezhnykh E. A. Nikonenko and E. I. Krylov Russ. J. Inorg. Chem. 1977 22 356. " A. A. Pinkerton and D. Schwarzenbach J.C.S. Dalton 1976 2464. l8 A. A. Pinkerton and D. Schwarzenbach J.C.S. Dalton 1976 2466. '9 Y. Meseri A. A. Pinkerton and G. Chapuis J.C.S. Dalton 1977 725. 2o M. Ciampolini and N. Nardi J.C.S. Dalton 1977 2121. 21 V. P. Shilov E. S. Stepanov and N. N. Krot Soviet Radiochem. 1977 19 59. 22 V. P. Shilov E. S. Stepanov and N. N. Krot Souier Radiochern. 1977,19 64. 23 M. P. Mefod'eva N. N. Krot and T. V. Afanas'eva Radiokhimiya 1977 19 245. 24 V. N. Kosyakov G. A. Timofeev E. A. Erin V. I. Andreev V. V. Kopytov and G. A. Simakin Radiokhirniya 1977 19 511. 212 J.R.Dilworth G. J. Leigh R. L. Richards and K. W.Bagnall 1.52k0.01 V. Reduction of AmrV to Am'" in this medium occurs only under radiolysis but CmIV and CfIV are reduced by water.24 The redox behaviour of Am Cm Bk Cf Es and Fm in aqueous media has been investigated by radiopolaro- graphic and radiocoulometric techniques. Electrochemical reduction at a mercury cathode appears to involve Fm" [Fm& +2e-+Hg S Fm(Hg)] whereas for the other actinides it is the tripositive ion which is reduced to the element.*' The Frn:,',)/Fm& potential has been found to be very close to that of the Yb&/Yb& system by means of a cocrystallization technique [254Fm& with SrC12 in the presence of Yb& (not Y2+ as stated)].26 The uranium atom in &UF5 is not seven-co-ordinate as previously believed but eight-co-ordinate the geometry being intermediate between dodecahedra1 and square antiprismatic; the terminal U-F(1) bond length [196(2) pm] is the shortest so far reported for a uranium(V) c~mplex.~' The co-ordination geometry about the uranium atom in U02Br is pentagonal bipyramidal (three 0 and two Br atoms in the equatorial plane) and the structure consists of (U02/203/3Br2/2) layers perpen- dicular to [OlO].The compound is prepared by the thermal decomposition of anhydrous U02Br2 at 650°C in uucuo in a sealed tube.28 The plutonium(v1) oxofluoride PuOF4 has been obtained by the controlled hydrolysis of PuF6 in anhydrous HF both by water and by using the calculated quantity of quartz wool to generate water.Although it is stable at room temperature it disproportionates in anhydrous HF to yield a mixture of PuF6 and Pu02F2 in contrast to UOF and NPOF which are stable in this respect. The compound is isostructural with a-UOF4 and NPOF,.~~ The uranium(v) fluorosulphonate UF2(S03F)3 is obtained by reaction of UF6 with SO in the gas phase or in CFCl3 solution a reaction in which S206F2 is also formed. The compound is stable to 120 "C and its vibrational spectrum indicates that there are two different types of S03F group; two of the groups are probably bridging bidentate and the other may be non-bridging bidentate or unidentate. The two fluorine atoms are terminal.30 Uranium(II1) complexes are uncommon because of the ease of oxidation to uranium(IV) but a number of complexes with A""'-tetramethyldicarboxylic acid amides of the type [UL4][BPh4I3 have now been obtained as precipitates by the addition of an ethanolic solution of Na[BPh,] to a solution of NH4UC14,5H20 and the amide in the same solvent.31 A theoretical treatment of the magnetic behaviour of a variety of six-co-ordinate uranium(1V) complexes of the type [UX4L2](X = C1 or Br L = PR30 or AsR30) of D4, or CZ0symmetry has been published; the calculations fit the experimental results over the temperature range 70-320 K.32 The u.v.-visible spectra of the hexakis(arsine oxide) complex UCl4,6AsMe3O indicate that it is of the form [UC1L6I3' 3C1- like UC14,6PMe30 *' F.David K. Samhoun and R. Guillaumont Rev. Chim. minkrule 1977 14 199. 26 N.B. Mikheev V. I. Spitsyn A. N. Kamenskaya N. A. Konovalova I. A. Rumer L. N. Aueman and A. M. Podorozhnyi Inorg. Nuclear Chem. Letters 1977.13 651. *' R. R. Ryan R. A. Penneman L. B. Asprey and R. T. Paine Acta Cryst. 1976 B32 3311. 28 J.-C. Levet M. Potel and J.-Y. le Marouille Actu Cryst. 1977 B33 2542. 29 R. C. Burns and T. A. O'Donnell Inorg. Nuclear Chem. Letters 1977,13 657. 30 W. W. Wilson C. Naulin and R. Bougon Inorg. Chem. 1977 16 2252. 31 J. I. Bullock and A. E. Storey J.C.S. Chem. Comm. 1977 507. 32 J. W. Gonsalves P. J. Steenkamp and J. G. H. du Preez Inorg. Chim. Acta 1977 21 167. Chemistry of the d-and f-BlockMetals 213 reported last year whereas the corresponding phosphine and arsine oxide XR30 complexes of the bromide and iodide are of the form [UL6I4"4Y- (X=P or As R = Me or Et).The pale (off-white to very pale green) colours of these last are also indicative of a centrosymmetric environment of the uranium atom.33 Three poly- morphs of [UO2(hfpd),{0P(OMe),}] (hfpd = 1,1,1,5,5,5-hexafluoropropane-2,4-dionate) have been detected in the course of thermal studies with this complex34 and the structures of the and p-f~rms~~ have been determined. In both of them the uranium atom adopts seven-co-ordinate pentagonal-bipyramidal geometry but whereas the (hfpd) planes are tilted by 22.5" to the plane of the pentagon in a boat formation in the a-form there is only a slight tilt in the 6-form. A number of papers describing crown ether solvates of uranium compounds have appeared.In crystals of the composition {[U(NCS)4(H20)4],(18-crown-6)1.5,3H20,MeCOBui},neither the crown ether nor the ketone is bonded to the uranium atom which is co-ordinated to four NCS groups via the N atom and to four 0 atoms of water molecules in a square antipri~m.~' Similarly 'Hn.m.r. evidence shows that in the hydrated U02(N03)2 and Th(N03)4 crown ether solvates (15-crown-5 benzo-15-crown-5 18-crown-6 dicyclohexyl-18-crown-6 dibenzo-24-crown-8) the crown ether is outer-~phere.~~ A report of the structure of [(12-~rown-4)UO~(OH~)~](NO~)~, in which the UO group is said to be sur- rounded equatorially by a near-planar hexagon consisting of four 0atoms from the ether and two from the water seems unlikely to be correct. Of the other crown ether species the high melting point 169-172 "C (decomp.) of the 15-crown-5 solvate of UO,(NO,) may indicate a possible bonding intera~tion;~" although the i.r.spectra of a number of hydrated 1 1solvates of U02C12 with various crown ethers are said to be consistent with co-ordination of the ether to the the shift in the C-0-C stretching frequency couId equally well be due to other causes. In the structure of NaU02(HC02),,H20 the metal atom is surrounded by a pentagonal bipyramid of oxygen atoms42 whereas in the neptunium(v) complex BaNp02(MeC02)3 the co-ordination geometry is hexagonal bi~yramidal.~~ Further examples of heteronuclear Schiff-base complexes have been reported in which the d transition-metal ion (Ni" or Cu2') is held in the (N2,02)site and the UOf' ion is held in the (02,02) ~ite.~~.~~ A number of uranium(1V) amido-complex structures have been reported; [U(NPh2)4] is a rare example of four-co-ordinate uranium(IV) the co-ordination 33 J.G. H. du Preez B. J. Gellatly and M. L. Gibson J.C.S. Dalton 1977 1062. 34 J. H. Levy and A. B. Waugh J.C.S. Dalton 1977 1628. 35 J. C. Taylor and A. B. Waugh J.C.S. Dalton 1977 1630. 36 J. C. Taylor and A. B. Waugh J.C.S. Dalton 1977 1636. 37 P. Charpin R. M. Costes G. Folcher P. Plurien A. Navaza and C. de Rango Inorg. Nuclear Chem. Letters 1977 13,341. 38 J. Klimes A. Knochel and G. Rudolph Inorg. Nuclear Chem. Letters 1977. 13,45. 39 N. Armggan Acta Cryst. 1977 B33 2281. 40 D. L. Williams and L. E. Deacon J. Inorg. Nuclear Chem. 1977 39 1079.41 D. L. Tomaja Inorg. Chim. Actu 1977 21 L31. 42 B. F. Mentzen Acta Cryst. 1977 B33 2546. 43 J. H. Burns and C. Musikas Inorg. Chem. 1977,16 1619. 44 D. E. Fenton S. E. Gayda U. Casellato M. Vidali and P. A. Vigato Inorg. Chim. Acta 1977,21 L29. 45 M. Vidali U. Casellato P. A. Vigato L. Doretti and F. Madalosso J. Inorg. Nuclear Chem. 1977 39 1985. J. R. Dilworth G.J. Leigh R. L. Richards and K. W. Bagnall geometry being a highly distorted tetrahedron whereas in [UO{NPhz)3Li(OEt2)]z the uranium atoms in the oxygen-bridged dimer are five-co-ordinate in an approx- imately trigonal-bipyramidal arrangement (three N two 0 atoms). Both compounds are very sensitive to oxygen; the first is prepared by transamination of U(NEtZ) with PhNH2 or by reaction of UCl with LiNPhz in ether while the second is obtained from the filtrate from the latter and evidently results from reaction with air.46 Transamination of the diethylamide with NN’-dimethyldi- aminoethane in pentane yields the trimeric molecule [U3(MeNCHzCH2NMe)6] in which the three U atoms form a linear chain; the central one which is on a centre of symmetry is linked by a triple nitrogen bridge to each of the terminal U atoms and is in an octahedral environment of N atoms.The terminal U atoms are at the centres of distorted trigonal prisms of N atoms the whole presenting a very unusual A closed tetramer [U4(MeNCH2CEizNMe)8] is formed as a minor product of the above reaction; in this the four U atoms are 360pm apart at the corners of a twisted square and each U atom is bonded to six N atoms at the corners of a highly distorted trigonal prism.48 A preliminary communication reports the formation of a black penta-aza-complex UOzL{[2,6-diacetylpyridine-bis-(2’-pyridylhydrazonato)-~~~~~]-dioxouranium(vI) (2)obtained by deprotonating the product of the reaction in ethyl acetate of U02(N03)2,6H20with 2,6-diacetylpyridine-bis-(2’-pyridylhydrazone) by means of 1,8-bis(dimethylarnino)naphthalene (proton sponge).This reacts with anhydrous methanol to yield a polymeric species of composition [H,L(Uo,)2(oMe),(MeoH)I~.49 46 J. G. Reynolds A. Zalkin D. H. Templeton and N. M. Edelstein Znorg. Chem. 1977 16 1090. 47 J. G. Reynolds A. Zalkin D. H. Templeton and N. M. Edelstein Znorg. Chem.1977 16 599. 48 J. G. Reynolds A. Zalkin D. H. Templeton and N. M. Edelstein Znorg. Chem. 1977 16 1858. 49 G. Paolucci and G. Marangoni Znorg. Chim. Acta 1977,24 L5.

ISSN:0308-6003    

DOI:10.1039/PR9777400169

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

9 Organometallic Compounds By C. J. CARDIN D. J. CARDIN and R. J. NORTON Chemistry Department Trinity College University of Dublin Dublin 2 Ireland and K. R. DIXON Department of Chemistry University of Victoria Victoria 8.C.,Canada V8W 2Y2 1 Introduction Literature coverage for this year has been for the main-group and early transition elements (Sections 1-19) by C.J.C. D.J.C. and R.J.N. with metal carbonyls fluxional processes and related topics (mainly of the later transition metals Sections 20-25) by K.R.D. The coverage is essentially for 1977 in the first part and November 1976-7 in the second. Further developments in carbene- and carbyne-complex chemistry have been such that a section has been devoted to them for the third successive year while the only other two topics dealt with previously i.e.syntheses involving metal atoms and ylides have not been covered for two years; there have been significant advances in these areas.2 Groups IA and IIA The methyl derivatives of rubidium and caesium have been obtained from the reaction of methyl-lithium with rubidium t-butoxide and caesium-2-methyl-2- pentoxide (chosen for its solubility) respectively in ether. X-Ray powder data support a hexagonal structure of the NiAs type with isolated anions and cations resembling methylpotassium but unlike the electron-deficient tetramers of methyl-lithium.'" The reaction between dilute solutions of butyl-lithium in benzene and 7bH-indeno[ 1,2,3-jk]fluorene affords species (l),in which the aromatic anions are coplanar and the Li atoms lie symmetrically between pairs of external six- membered rings.Sandwich structures of molecular species with hexahupto-ligands were previously unknown for lithium and considerations of charge density on the ring (greatest for the five-membered rings) suggest a significant covalent interaction between metal and carbanions." A useful method for chain extension of alkyl- lithiums by two carbon atoms is shown in equation (1). A related reaction in which R-Li (i) CH2=CHAsPh2 R(CH&X -+ R(CH2)2Li (ii) H20 (R = alkyl; X = C1 or Br) (a)E. Weiss and H. Koester Chem. Ber. 1977 110 717; (b)D. Bladanski H. Dietrich H. J. Hecht and D. Rewicki Angew Chem. Internal. Edn. 1977 16 474; (c) T. Kauffmann H. Ahlers H. J. Tilhard and A.Woltermann ibid. p. 710. 215 C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Li + Li + the arsenic reagent is replaced by phenyl vinyl selenide afforded much poorer yields in both stages leading to the halogeno-alkane." The structure and bonding in dicyclopentadienylberyllium continue to excite attention. The dipole moment of the compound coupled with electron-diff raction data suggested the unsymmetrical structure (2) rather than the symmetrical alter- native (3) (either D5,,or D5d).Indeed no symmetrical metallocenes are known for I Be I Be the elements of the first short period. From ab initio M.O. calculations Collins and Schleyer have concluded that the HOMOS(which occur in degenerate pairs) for the DShand D5dstructures have nodes passing through the central atom and that a distortion from these structures would be expected.'" Furthermore they make the interesting suggestion that molecules with 8 or fewer interstitial electrons (for rings of 4 or more atoms) would be expected to be stable with respect to 0-bonding isomers.Some possible first-row symmetrical sandwiches are shown in (4)-(6). By Li + Be '2" I ' (a)J. B. Collins and P. von. R. Schleyer Znorg. Chem. 1977 16 152; (6)D. S. Marynick J. Amer. Chem. SOC.,1977,99 1436. Organometallic Compounds contrast Marynick has concluded that the Cs,structure is not an energy minimum lying very close in the potential-energy surface to the DSd configuration.2* However the suggested structure (7) has one ring 0-bonded to the metal.Be Although believed to be intermediates in silyl-transfer reactions involving magnesium silylmagnesium species have not previously been isolated. Bis(tri-methylsily1)magnesium has been isolated both as its dimethoxyethane (DME) complex [equation (2)] and as the bis(tetrahydr0furan) adduct after undergoing Hg(SiMe& +Mg D% (Me3Si)2Mg(DME) (2) reaction at room temperature for 4-5 The crystal structure of the DME solvate shows l(Mg-Si) = 263 pm.36 The THF species loses the ether in solution; it is suggested that the resulting dimer may have bridging trimethylsilyl groups. Magnesium dialkyls of the bulky ligands neopentyl trimethylsilymethyl and 2-methyl-2-phenylpropyl have been prepared by displacement of the Schlenk equilibrium with dio~an.~~ The neopentyl species is unusual in being freely soluble in hydrocarbons (as a trimer) and subliming at 110 "C.Unfortunately exchange between bridging and terminal alkyl groups is rapid on an n.m.r. time-scale (both 'H and "C) so that a distinction between a cyclic and linear structure for the trimer could not be made. The other new compounds are believed to be polymeric. 3 Organo-actinides Two groups have reported the synthesis of the first indenyl-uranium and -thorium alkyls. Compounds of the type [MR(q5-C9H7)3] (M=Th or U; R=alkyl or alk-OXY)~" and [U(q5-C9H7)2Rz] (R = alkyl) were prepared.46 Few actinide alkyls other than of the type [MCp3R] were known previously. Indeed the electrical neutrality of the UR4 compounds reported by has been questioned by Sigurdson and Wilkinson who have reported the synthesis of the first compounds (homoleptic alkyl-actinide anions) containing six and eight c+-bonds between uranium and carbon (Scheme l).4d The lithium hexa-alkyluranate(1v) compounds are thermally unstable whilst the lithium octa-alkyluranate(v) compounds are fairly stable.(a) L. Rosch Angew. Chem. Internat. Edn. 1977 16 247; (6) A. R. Claggett W. H. Ilsley T. J. Anderson M. D. Glick and J. P. Oliver J. Amer. Chem. Soc. 1977,99,1797; (c) R. A. Andersen and G. Wilkinson J.C.S. Dalton 1977 809; M. R. Collier M. F. Lappert and R. Pearce ibid. 1973 445. (a)J. Goffart B. Gilbert and G. Duyckaerts Inorg. Nuclear Chem. Letters 1977 13 189; (6) A. M. Seyam and G. A. Eddein ibid. p. 115; (c)T.J. Marks and A. M. Seyam J. Organometallic Chem. 1974 67 61; (d)E. R. Sigurdson and G. Wilkinson J.C.S.Dalton 1977 812. C. J. Cardin D. J. Cardin,R. J. Norton and K. R. Dixon UC14+ RLi (excess) Et2O b Li2[UR,)8Et20 R = Me CH2SiMe3 Ph or o-Me2NCH2C6H4 [U(OEt)S]+ RLi (excess) Li3[UR8],3dioxan R = Me CH2CMe3 or CH2SiMe3 Scheme 1 4 Groups IVA VA and VIA The two bulky ligands -CHPh (benzhydryl) and -CH(SiMe3)2 have been used to prepare a series of metallocene derivatives of the early transition metals.sd It was not possible to obtain homoleptic alkyls with these ligands apart from those previously reported" {[VR,] and [CrR3]; R = CH(SiMe3),} and the species [ZrC1R3] and [HfCIR3]. However in the reaction of [Ti(C5H5)2C12] with 2 mol of either organolithium reagent the reaction appears to proceed by initial alkylation followed by homolytic cleavage to give the titanocene(II1) chloride which is alkyl-ated by a second mole of organolithium reagent to give [Ti(CsH5)2R].This reaction pathway is also observed with [V(CSHs),CI2] under the same conditions but in this case reductive alkylation would be expected. The greater steric demand of -CH(SiMe3)2 is consistent with the alkylations of [Zr(C5H5)2C12] and [Hf(CsHs)2C12] giving disubstitution when R = CHPh but only monosubstitution when R = CH(SiMe3)2 even under relatively forcing conditions. X-Ray crystal structure determinations have been reported for [M(CsHs)2(CHPh2)2] (M = Zr or Hf) and here the steric demand of -CHPh is shown by an exceptionally long average Zr-C bond length of 238.8(1.2)pm 13 pm longer than that previously observed for Zr-Me in [Zr(C9H7)2Me2].In spite of the very similar atomic radii of ZrIV and HfIV the Hf-C bond length appears normal. The oxidation of Ti'" to TiTV is a step in the synthesis of tetramesityl-titanium(~v)." In this case the homoleptic tetramesityltitanium(II1) is oxidized by molecular oxygen in clear parallel with previous Cr"' 3 CrIV oxidationssd [see equation (3)]. 0 TiCI3+ 4(mesityl)Li -+ Li[Ti(rne~ityl)~]2,[Ti(rne~ityl)~] (3) Titanium(II1) aryls [Ti(CsH5),(Ar)] have been shown to react with aryl cyanides (Ar'CN) at -10 "C to give a series of compounds [Ti(CsHs)2(Ar)(Ar'CN)].5' Some of these undergo coupling reactions on warming; the reactions involve the cyanide ligands and produce dimeric products (not yet fully characterized).Thermal decomposition of early transition-metal organometallics is an area of interest again this year. In contrast to earlier work on the titanium analogues the thermolysis of [Zr(CsHs)2R2] (R = Me or Ph) in CCl to give [Zr(CSHs)2C12] RH (a)J. L. Atwood G. K. Barker J. Holton W. E. Hunter M. F. Lappert and R. Pearce J. Arner. Chem. Soc. 1977,99,6645; (b)G. K. Barker and M. F. Lappert J. Organometallic Chem. 1974.76 C45; (c) W. Seidel and I. Buerger 2.Chem. 1977 17,105; (d)J. Muller and W. Holzinger Angew. Chem. Internat. Edn. 1975 14 760; (e)E. J. M. De Boer and J. H. Teuben J. Organometallic Chem. 1977 140,4 1. Organ ometa llic Compounds 219 and RCl is complex and involves the cyclopentadienyl group.6a The thermal and photochemical decomposition of substituted zirconium aryls [Zr(C5H5)2Ar2] (Ar = 0-m- or p-tolyl) in benzene gives a variety of products.6b These include presumed intermediates resembling ortho-metallated compounds (as shown in Scheme 2).+phxH 1 +'ZhH Cp2Zr$3 Cp2Zr& Cp2ZrPh2 -PhH @ Cp = q5-C5H5 cp2zD Scheme 2 The cyclopentadienyl group is also involved in the thermolysis of the 16-electron [V(C5H5),Ar] compounds (Ar =Ph o -tolyl m-tolyl p-tolyl or 2,6-~ylyl).~' Thermal stability is distinctly increased by substitution in the ortho-position of the aryl group probably due to shielding of vacant co-ordination sites in the metal by the ortho-methyl group.The principal products of thermal decomposition are ArH [V(C5H5),] and vanadocenes in which one or both C5H5 rings are substituted by migrating aryl groups. From a study of the mass spectra of the decomposition products of selectively deuteriated analogues a decomposition mechanism has been proposed in which the aryl group is transferred from the vanadium atom of the one molecule by interaction with the aryl group of another. The proposed '(a) G. A. Razuvaev V. N. Latyaeva L. I. Vishenskaya G. A. Vasil'eva V. I. Khruleva and M. I. Smirnova Doklady Akad. Nauk. S.S.S.R.,1976 231 114; (b) G. Erker J. Organometallic Chem. 1977,134 189; (c)C. P. Boekel A. Jelsma J. H. Teuben and H. J. de Liefde Meijer ibid. 1977,136 21 1; (d)C. P. Boekel J. H. Teuben and H.J. de Liefde Meijer ibid. 1977 128 375. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon mechanism is consistent with the observed (approximate) stoichiometry shown in equation (4). + [V(CSH~)~I ~[V(C~H~)ZA~] i-ArH+ [V(CSH~)(C~H~A~)I (41 The same authors have also studied the thermal decomposition of [Nb(C5H5)2Ph2] in which quantitative formation of benzene occurs again by intramolecular hydrogen abstraction from a C5H5 ring.6d The reaction is first- order with an activation energy of about 146 kJ mol-' in both the solid and in toluene solution rising to about 163 kJ mol-' on deuteriation of the Ph ring but remaining the same on deuteriation of the C5H5 ring. Last year's Report covered several papers describing Group VI alkyls containing multiple metal-metal bonds.Interest continues in this area and the work has now been extended to other elements. Using the magnesium dialkyls MgR (where R = Me CH2SiMe3 or CH2CMe3) as alkylating agents a series of dinuclear tran- sition-metal acetates [M(OAc),] (M = Cr Mo Re Ru or Rh) has been found to give alkyls containing metal-metal multiple bonds,'" e.g. equation (5). A crystal [Cr(O2CMe>,] +2MgR +2PMe3 -+ [C~ZR~(PM~~)~] (R = Me3SiCH2) (5) structure determination on this chromium(I1) alkyl reveals a molecule of approxi-mate C2symmetry in which the two chromium atoms are bridged by two of the alkyl groups. The Cr-Cr distance of 210.0 pm is indicative of strong metal-metal bonding and there seems also to be an interaction between each chromium atom and one of the a-hydrogens of the bridging alkyl groups Cr-H being about 230 pm long (see Figure 1).\ CH Si Me -230 ,,/Li--h H Si Me Figure 1 Bond lengrhslpm of [Cr2(CH2SiMe3)4(PMe3)2] ' (a)R. A. Andersen R. A. Jones G. Wilkinson M. B. Hursthouse and K. M. A. Malik J.C.S. Chem. Comm. 1977 283; (b)D. M. Collins F. A. Cotton S. Koch M. Millar and C. A. Murillo J. Amer. Chem. Soc. 1977 99 1259; (c) M. H. Chisholm F. A. Cotton M. Extine and B. R. Stults Inorg. Chem. 1976,15 2252; (d)F. A. Cotton S. Koch and M. Millar J. Amer. Chem. SOC.,1977.99 7372; (e)F. A. Cotton S. Koch K. Mertis M. Millar and G. Wilkinson ibid. p. 4989; (f) A. P. Sattelberger and J. P. Fackler ibid. p. 1258; (g) M. H. Chisholm F. A. Cotton M. W. Extine M.Millar and B. R. Stults Inorg. Chem. 1977 16 320. Organometallic Compounds 22 1 The first direct proof of the existence of the W-W quadruple bond has been obtained by the determination of the crystal structure of the disordered compound Li4[W2(Me)8-,C1,].7b This mixture of species was obtained from the attempted preparation of the [W2Mes14- ion by adding LiMe to WCl at -78 "C using a mole ratio LiMe WCl of 6 1. The red crystalline product obtained is thermally unstable above -20 "C and the X-ray analysis at -55 "C showed molecules of approximate D4,, symmetry in which C1 and Me groups are best considered to be internally disordered within each [W2(Me)8-,C1,]"- unit and present in approxi- mately equal amounts. It is however curious that the W-W bond length of 226.3 pm is slightly longer than the W-W triple bond of 225.5 pm recorded for [W2(CH2SiMe3)6],7C presumably indicating that bond length is almost insensitive to bond order when such high bond orders are involved.However since the stoi- cheiometry of the crystals is non-integral (approximating to [W2C14.7&k3.28]) the observed bond length is clearly averaged over a number of chemical species. Cotton has recently introduced a simple but useful method for comparing metal- metal bond lengths in relation to data on bonds between different metals by ratioing the absolute lengths to the approximate Pauling metallic radii.7d The quantity for comparison the 'formal shortness ratio' is given by the absolute distance observed divided by twice the Pauling metallic radius used.On this basis both [W2Me8l2- and [W2(CH2SiMe3)6] have formal shortness ratios in the region appropriate to triple bonds rather than to quadruple bonds. The ion [W2Me8l4- has been prepared from both WCl and WC15 by using higher molar ratios of LiMe but so far has not been obtained in crystals suitable for X-ray analy~is.~' In all these syntheses methyl-lithium serves both as reducing [to W"] and methylating agent and the optimum molar ratio for the preparation is that which gives a LiMe W ratio in the range 5-6 1 after allowance for reduction. The electronic spectrum of [W2Me814- at 77 K has been reported and in the visible region it is homologous to that obtained for Cr and Mo analogues. The position of the peak maximum is consistent with the previously established correlation between 8-8" transition energy and M-M di~tance.~' Amongst W-W triply bonded species a stereospecific reaction between anti-[W2C12(NEt2),] and [LiCH2SiMe3] has been recorded7' [equation (6)].The anti-unti-[W2Cl2(NEt2),]+ 2LiCHzSiMe3 -P ~nti-[W~(CH2SiMe~)~(NEt2)~] (6) product subsequently isomerizes to an equilibrium mixture of anti-and gauche-rotamers (K = 4). A variable-temperature n.m.r. study has given the activation parameters for the isomerization as AG' = 101* 1.0 kJ mol-'. The value for AG' is about 12.5 kJ mol-' higher than that previously determined for anti-[W2Me2(NEt2),] reflecting presumably the greater steric demand of the -CH2SiMe3 ligand although the mechanism of this isomerization is not yet known.5 GroupIB Amongst a series of copper alkyls stabilized by tertiary phosphine are the unusual dinuclear derivatives [ (C~R)~(dppe)~]. These (in a similar fashion to alkyl-coppers with unidentate phosphines) were obtained by the reaction of alkylaluminium C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon ethoxides with the phosphine and copper(I1) acetylacetonate and they are believed to have a structure with a bridging diphosphine ligand. Thermolqsis of these alkyls [see equation (7); R =Me Et Pr" or Bu'] affords alkane Ph2PCH=CH2 and the copper species shown in a remarkable reaction that is thought to involve the C-H C-P and C-Cu cleavages indicated.8" (8) Although both diphosphinomethanide and phosphoniumbis(methy1ide) groups are known as independent ligands for transition metals it appears that when both are in a complex they confer a greater degree of stability (including thermal stability) than might have been expected.86 The gold derivatives (like a related palladium compound) can be obtained from the known bis(methanide) by reaction with the 'mixed' ylide [equation (S)].(See also Section 14). Ph 2 P-Au-[AuCH(PPh2),] +Et2P(:CH2)Me +(2'PEt2 +CHZ(Ph2P)z (8) P-Au Ph2 Details are now available on the cluster compound [{CUR},] (R =Me3SiCH2).8' Both 0-and C-bonded acetylacetonates of metals have been recognized for several years and it is noteworthy that a reversible isomerism of the two forms is induced by phosphines in the Au"' system shown in equation (9).The rearrange- Me Me PR (9) Me ''0 +PR Me \/ ment is promoted by several phosphines the efficacy falling in the series PPhMe2 > PPh2Me>PPh3. With PPh3 the thermodynamic characteristics [AH = -45.98 kJ mol-'; AS =-146.3 kJ mol-'1 are consistent with an associative mechanism as expected.8d Although pure phenylsilver had been reported by van der Kerk and co-workers in 1972 using tin or lead derivatives as arylating agent diphenylzinc has been found to be more satisfactory in that it is a stronger organylating reagent and also lacks the reducing tendencies of aryl-lithium or -magnesium species.'" The reac- (a) A. Miyashita and A. Yamamoto Bull. Chem. SOC.Japan 1977 50 1102; (6) H. Schrnidbaur and J. R. Mandl Angew.Chem. Internat. Edn. 1977 16 640; (c)J. A. Jarvis R. Pearce and M. F. Lappert J.C.S. Cdton 1977,999; (d)S. Komiya and J. K. Kochi J. Amer. Chem. SOC.,1977,99,3695. (a)J. Boersma F. J. A. Des Tornbe F. Weijers and G. J. M. Van der Kerk J. Organometallic Chem. 1977 124 229; (6) S. Komiya J. C. Huffman and J. K. Kochi Znorg. Chem. 1977 16 1253; (c)ibid. p. 2138. Organometallic Compounds tion shown in equation (10) gives the silver compound in almost quantitative yield and in a more stable form than hitherto. Et2O PhzZn+ 2Ag[N03] o" 2[AgPhl+ Zn"O31~ In view of the supposed instability of monoalkyl-gold(II1) species the product of equation (11)is of note. However X-ray data on the product reveal that both + 2Br2 -+ [{AuB~~R}~] [{M~,AUB~}~] + 2MeBr (11) methyl groups are bonded to one of the gold atoms and it is not therefore surprising that thermal decomposition affords ethane by reductive elimination together with [AU~(~-B~,)AU~~~B~~].dimethylgold derivative, Another [AuMe,OSO,CF,] obtained from silver triflate and [AuIMe2] is like [AuMe,(aq)]' apparently stable in water.The hydrate [AuMe2(HZO)OSO2CF3 J has a square-planar arrangement about the metal atom.9b Despite the stability in aqueous solution a solution in benzene smoothly undergoes reductive elimina- tion affording ethane metallic gold and (presumably via AuOSO,CF,) [AU(OSO~CF~)~].~" 6 GroupIIB The bis(bis-ylide) derivative shown in equation (12) is the first tetra-alkyl derivative of mercury to have been reported.The mercury compound is a pale yellow M + 2[Me3P-CH-PMe3]C1 crystalline solid sensitive to air which shows an intense molecular ion in the mass spectrometer. N.m.r. studies [a 'H-I9'Hg experiment] carried out over the range -60 to +35"C show that the molecule is non-fluxional the four ligands being chemically (and magnetically) equivalent."" For some time it has been thought that o-phenylenemercury exists as a hexamer. Despite attempts to prepare a hexameric species (12 plausible syntheses have been examined) only trimers of formula (C6H4Hg) were obtained. Among the evi- dence for the trimeric formulation is an intense peak in the mass spectrum cor- responding to [CsH4Hg]3+ which must almost certainly correspond to the molecu- lar ion for a volatile species.1o6 7 Group IIIB As noted in last year's Report thiadiborolenes are known as q5-ligands in both double- and triple-decker sandwiches.An interesting recent development [equa- tion (13)] affords an 18-electron nickel complex (formally a replacement of two lo (a)H. Schmidbaur 0.Gasser T. E. Fraser and E. A. V. Ebsworth J.C.S. Chern. Comm. 1977 334; (b) S. B. Awad D. S. Brown S. C. Cohen R. E. Humphries and A. G. Massey J. Organometallic Chern. 1977,127,127. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon R HR CH units of the ring of nickelocene by BR groups) which as expected is diamag- netic though air-sensitive in solution."" The first triple-decker complexes containing only h'-thiadiborolens as ligands have been prepared with cobalt."b Previous complexes of this structure all had two (CsHs) ligands together with a single heterocyclic ligand.1,3-Di-t-butyl-2-methyl-A4-1,3,2-diazaboroline appears to be rather strongly bonded to the metal in (9). The complex is stable in Me (9) air and decomposes at 170"C. Spectroscopic evidence (from "B and 14N n.m.r. u.v. and p.e. spectra) suggests that the boroline behaves as a delocalized 677-electron ligand."' h 5-Complexes of pentaphenylborole have been obtained by direct interaction with enneacarbonyldi-iron(o) and with [Ni(CO4].'ld More remarkably the tendency of the 1-phenyl-4,5-dehydroborepinto undergo ring contraction under forcing conditions is exemplified by the reaction with [Fe(CO)5] in boiling mesitylene affording the h 5-(2-ethyl- 1-phenylborole) complex of Fe(C0)3.' As with magnesium (see above) aluminum had not previously afforded isolable compounds containing Al-Si bonds although such species have been postulated as reaction intermediates.The species [Al(SiMe3)3(THF)] has been obtained from aluminum and bis(trimethylsily1)mercury in pentane-THF but the base-free molecule has not been isolated.12 The adduct decomposes at about 50"C and inflames in air. 8 GroupIVB The preservation of mercury-carbon and silicon-carbon bonds during direct fluorination of organometallics was reported last year.130 Liu and Lagow have now reported the synthesis of tetrakis(trifluoromethy1)germanium from tetramethyl- germanium in more than 60% yield by direct fl~orination.'~' This method may become the one of choice for synthesis of some perfluorinated organometallic compounds.'I (a)W. Siebert and M. Boschmann Angew. Chem. Internat. Edn.,1977,16,468; (b)W. Siebert and W. Rothermel ibid.,p. 333; (c) G.Schmid and J. Schulze ibid. p. 249; (d)G.E. Herberich J. Hengesbach U. Kolle and W. Oschmann ibid.,p. 42. L. Rosch Angew. Chem. Internat. Edn. 1977,16 480. l3 (a) E. K. Liu and R. J. Lagow J. Amer. Chem. Soc. 1976 98 8270; (b)J.C.S. Chem. Cornm. 1977 450. Organometa1lic Compounds On the basis of experiments with optically active s-butyltrineopentyltin cleavage of organotin compounds with dibromine has generally been assumed to cause inversion of configuration at ~arb0n.l~~ Retention of configuration was found previously only in the case of optically active cyclopropyltin compounds and this was readily raticnaiized in terms of the known resistance of cyclopropyl derivatives to react with inversion.Retention of configuration has now been reported to be the predominant mode of reaction for s-butyltrialkyltin compounds and by extension for other tetra-alkyltin compounds. 14' The steric requirements of the neopentyl group are believed to be responsible for the special behaviour of s-butyltrineopen- tyltin. The silicon-silicon metathesis (Scheme 3) has been reported to proceed in refluxing benzene with palladium catalysts. 15n The maximum yields obtained were 33% (R = ethynyl) when using tetrakis(triphenylphosphine)palladium(o) as cata- lyst and 13% (R = vinyl) with bis(triphenylphosphine)(maleic anhy-dride)palladium(o).The vinyl and ethynyl groups appear to be activating and necessary for metathesis under less forcing conditions than previous silicon meta- these~."~ -+ Me3Si-SiMe2(CH2)3SiMe2-SiMe2R SiMez SiMe3 (R = vinyl or ethynyl) Scheme 3 A convenient new synthesis of dodecamethylcyclohexasilane from the readily available sym-dimethoxytetramethyldisilaneaffords 60% crude product under mild conditions [equation (14)].15' 6(MeO)Me2SiSiMe2(0Me)Na[OMe]-+ (SiMe2)6+6SiMe2(0Me)2 (1 4) Details of the oxidative addition of an alkyl or aryl halide to SnR2 [R= (Me2Si)2CH] have been publi~hed;'~" the compound SnCIR3 is a discrete monomer in the solid state (X-ray) unlike less bulky analogues such as SnC1Me3 (which has five-co-ordinate tin).9 GroupVB Synthetic routes to E(CF3)3compounds (E= N P As or Sb) fail to produce the analogous bismuth compound. The previously unknown Bi(CF,) has now been prepared by the reaction of Bi13 with cF3radicals produced by the glow discharge of C2F6.16a The compound is thermally unstable which accounts for previous failures to isolate it by standard methods. l4 (a)F. R. Jensen and D. D. Davis J. Amer. Chem. SOC.,1971 93,4048; (6) A. Rahm and M. Pereyre ibid. 1977 99,1672. l5 (a) H. Sakurai Y. Kamiyarna and Y. Nakadaira J. Organomeralfic Chem. 1977 131,147; (6) H. Watanabe K. Higushi M. Kobayashi T. Kitahara and Y. Nagai J.C.S. Chem. Comm. 1977 704; (c) M. J. S. Gynane M. F. Lappert S.J. Miles A. J. Carty and N. J. Taylor J.C.S. Dalton 1977 2009. l6 (a)J. A. Morrison and R. J. Lagow Inorg. Chem. 1977 16,1823; (b) H.Schmidbaur P. Holl and F. H. Kohler Angew. Chem. Internat. Edn. 1977 16 722; (c) R. Appel F. Knoll and H. D. Wihler ibid. p. 402; (d)A. J. Ashe and H. S. Friedman Tetrahedron Letters 1977 1283; (e)C. Jongsrna J. J. De Kok R. J. M. Wenstink M. Van der Ley J. Bulthuis and F. Bickelhaupt Tetrahedron 1977 33,205. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Phosphorane PHS is as yet unknown and hence there is interest in preparing simple organophosphorus(v) compounds. The recently reported 5-methyl-SA '-phosphaspiro[4,4]nonane (10) represents the simplest penta-alkylphosphorane yet synthesized (see Scheme 4).Ylide formation was suppressed by careful choice of reaction conditions. 16b Reagents i LiMe -60 "C;ii LiMe 20 "C Scheme 4 The synthesis of a novel phosphorus-carbon ring system (1 1) is shown in Scheme 5.16' The ring is apparently stabilized by withdrawal of rr-electron density into the external phosphorus-carbon bonds. Ph3P+CC14 + [Ph3P-CC13]+ C1-[Ph3P-C(CI)-PPh2(CI)]+ C1-I ii Reagents i 2Ph2PCl; ii P(NR2)3; iii A; iv Br2; V 0 Scheme 5 The synthetic route to arsabenzenes shown in equation (15) precludes the possi- bility of 3-substitution. 3-Substituted arsabenzenes have now been prepared in good yield by thermal cleavage reactions of the Diels-Alder 1,4-adducts of arsabenzene [equation (16)].16d(Note the first report of the trapping of a silaben- ~ene.~'~) R A novel synthesis (Scheme 6) of arsenic-substituted triptycene (12; X = As) also affords the new compound (12; X=Sb).The reaction appears to proceed via a Organometallic Compounds base + Scheme 6 benzyne intermediate since the use of a weak base leads to decomposition of the more rapidly formed carbanion along other reaction pathways. 16e Although earlier attempts to synthesize a tetraphenylphosphaferrocene failed,17a the synthesis of phosphaferrocenes [see equation (17)] has now been reported. The R R [{Fe(C5H5)(CO),},]+ 0150°C P R RpTFe-o Ph Ph R=HorMe authors suggest that the electron-withdrawing ability of the phenyl groups was responsible for the earlier failure. 17' The phosphaferrocenes undergo selective acetylation of the phospholyl moiety.X-Ray structural analysis shows that the molecule adopts the eclipsed conformation. The planarity of the cyclopentadienyl group and of the four carbons of the phospholyl group indicates that there is considerable aromaticity in the ligands. 10 Metallocyclo-alkanes and -alkenes For the purposes of this Report metallocyclic systems are subdivided into types A and B of which the former are dealt with under this heading and the latter in Section 11. M"> M<:) 'c (A) (B1 Metallocyclopentanes have been of interest for some time as potential inter- mediates in a number of reactions between transition metals and olefins and in last year's Report the work of Whitesides et al. on the decomposition of titanocyclo- pentane was described.18" A series of nickelocyclopentanes has now been synthesized by the 1,4-dilithium route [equation (18)].18b These complexes are extremely air-sensitive solids and in solution they appear to dissociate with loss of l7 (a)E.H. Braye and K. K. Joshi Bull. SOC. chim. belges 1971,80,651; (b)F. Mathey A. Mitschler and R. Weiss J. Amer. Chem. SOC. 1977 99 3537. (a) J. X. McDermott M. E. Wilson and G. M. Whitesides J. Amer. Chem. SOC. 1976,98,6529; (b)R. H. Grubbs A. Miyashita M. I. M. Liu and P. L. Burk ibid. 1977 99 3863; (c)R. H. Grubbs and A. Miyashita J.C.S. Chem. Comm. 1977 864. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon [NiCl,(P),] +LiCH2CH2CH2CH2Li-%[(P,?NiD] +2LiCI (P)2 = (Ph3P)2 [(cyclohexyl)3Pl2 or Ph2PCH2CH2PPh2 phosphine but they are also capable of co-ordinating added phosphine.When phosphine is added to give an overall stoicheiometry of [Ni(C4H,)P3] the major product of thermal decomposition is ethylene (P-carbon-carbon bond cleavage) but this depends on the ratio of added phosphine. The authors have proposed Scheme 7 to account for these observations. The same workers subsequently PNi(C4Hs) 5 P2Ni(C4H8) 5 P3Ni(C4Hs) 1 1 1 major product of ] but-1-ene cyclobutane ethylene [ thermolysis (@-hydrogen (reductive (@-carbon-transfer) elimination) carbon bond cleavage) Scheme 7 re-examined the thermal decomposition of [Ti(C4H8)(C5H5)2]'sc and in agreement with Whitesides et a1.,18afound ethylene to be the major product suggesting that the metallocycle was in equilibrium with a bis-olefin complex.In further support of this hypothesis an olefin-exchange reaction was found to give new metallocycles with both [Ni(C,H,)(Ph,P)] and [Ti(C,H,)(C,H,),] characterized by protonolysis e.g. as shown in equation (19). In both cases n-hexane was the major product CH3CH=CH2 +L,M(C4H8) + n-hexane +2-methylpentane +2,3-dimethylbutane (19) '/o (M=Ni) 53 3 6 '/o (M=Ti) 49 11 6 suggesting that the principal metallocycle formed in each case was (13) but the nickel metallocycle was only stable below -20°C and the titanium complex decomposed even below -50 "C. However the metallocycle from the reaction of [Ti(C4H8)(C5H5)2]and octa-l,7-diene has been isolated and characterized as (14) (C,H dzTi Me (13) (14) and the addition of carbon monoxide produced a mixture of cis- (17%) and trans-hydrindanone (68%).The insertion of CO into a metallocyclic ring to give a cyclopentanone now appears to be characteristic of metallocyclopentanes and therefore this is a suitable trapping reaction for thermally unstable species of this type. Clearly much remains to be done in this area. Organometallic Compounds A nickelacyclopentane in which a rigid configuration is enforced by the choice of substituents has been prepared by an oxidative addition/insertion in which 2 mol of 3,3-dimethylcyclopropenewas treated with 1mol of (a,a'-bipyridy1)cyclo-octa-1,5-diene)ni~kel(O).'~ The structure of the product (15) has been confirmed by X-ray diffraction.The thermal stability and decomposition of rhodacyloalkanes has been studied as a function of ring size.20" The reaction of [Rh12(PPh3)(CsMeS)] with the di-Grig- nard reagents BrMg(CH,),MgBr (n = 4,5,or 6) gave for n = 5 and 6 the expected six- and seven-membered rhodacycloalkanes but for n = 4 the pure rhodacyclo- pentane complex [Rh(C4H8)(CsMes)(PPh3)] could not be isolated -it was always contaminated with [Rh(CsMes)(PPh3)(CH2=CH2)].This is strong circumstantial evidence for decomposition of the rhodacyclopentane via 0-carbon-carbon bond cleavage and the first time a presumed olefin decomposition product has been isolated. The same products were isolated after using the cyclic dialkylmagnesium Mg(C4H8) as the alkylating agent.The dependence of the nature of the products of thermal decomposition on ring size was shown for the two larger ring compounds (n = 5 and 6) by heating these complexes at 160 "C under argon; the only volatile products were pentenes and hexenes respectively. Surprisingly the reaction of BrMg[CH,],MgBr with both [Co12(PPh3)(CsHs)] and [IrCl2(PPh3)(C5Mes)] leads straightforwardly to the expected metallocyclic products."* X-Ray crystallographic studies on both resulting complexes showed essentially identical features with the characteristic 'envelope' conformation of cyclopentane that is shown in Figure 2. The thermal decomposition of both these metallocycles in mesitylene at 110 "C leads only to C4 hydrocarbons; this is similar to the original observations on the decomposition of platinacyclopentane.20c Novel and surprising reactions occurred when Group VIA q 3-aIIyl complexes were treated with Na[BH4]; metallacyclobutanes (16) were produced [see equation (20)].20d Since the 'H n.m.r.spectrum for each of these compounds shows that the [M(CsHs),(C,H,)][PF6] +Na[BH4I + [M(C~H~)(CSHS)~I (20) (M = Mo or W) (16) two cyclopentadienyl groups are equivalent the metallacyclobutane ring must either be planar or more plausibly bent there being rapid interconversion of the 19 P. Binger M. J. Doyle J. McMeeking C. Krueger and Y.H. Tsay J. Organometallic Chem. 1977,135 405. 20 (a) P. Diversi G. Ingrosso and A. Lucherini J.C.S. Chem. Comm. 1977 52; (6) P. Diversi G. Ingrosso A. Lucherini W.Porzio and M. Zocchi ibid.,p. 811; (c)J. X. McDermott J. F. White and G. M. Whitesides J. Amer. Chem. SOC. 1976 98 6521; (d)M. Ephritikhine B. R. Francis M. L. H. Green R. E. Mackenzie and M. J. Smith J.C.S. Dalton 1977 1131. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Figure 2 The structure of [Co(C4H,)(PPh3)(C5H5)],omitting the phenyl groups. Unlettered atoms are carbon (Reproduced from J.C.S. Chem. Comm. 1977,811) two conformers. The exclusive formation of these ring compounds shows that the site of nucleophilic attack by H-on the co-ordinated allyl ligand is only the central carbon atom. This observation has led the authors to propose that in general where two or more sites are available for such attack by nucleophiles on an organometallic cation the preferred site will be determined by the electron- richness or -poorness of the metal centre.Thus an electron-rich centre would be attacked such that the increase in electron density at the metal centre in the product was minimized whereas an electron-poor cation would be attacked where electron density would be maximally increased. The reaction reported here is that expected for an electron-rich metal centre and it is noteworthy that attack by H-at C-1 or C-3 of the co-ordinated allyl ligand which does not occur is in fact a real alternative since the expected product i.e. [W(CsH5)(r12-MeCH=CH2)],is known and is thermally stable. Syntheses of metallocyclopentadienes from acetylenes are now well known but a series of cobaltocyclopentadienes have been reported in which the products are thermally stable and stable in air [see equation (21)] so that even solutions can be handled in air.21a As expected unsymmetrical acetylenes gave mixtures of isomers and the sequential addition of different acetylenes gives mixed rings via the initial acetylene complex.[Co(CSH,)(PPh3),] +2PhrCPh + [CO(CSH~)(PP~~)(C~P~~)] (21) A ring incorporating platinum and silicon has been obtained by treating a series of platinum-disilicon compounds with acetylenes [equation (22)].’lb 21 (a) H. Yamazaki and Y. Wakatsuki J. Organometallic Chem. 1977 139 157; (b) C. Eaborn T. N. Metham and A. Pidcock ibid. 1977 131 377. Organometallic Compounds 231 Rings that contain main-group and transition metals are also involved in the preparation and reactions of aluminacyclopentadiene.22".b The preparation"" is accomplished by the dilithium route giving the ether adduct of [PhAI(C4Ph4)].This compound acts as a cyclic diene and forms olefin complexes with transition metals. However when 2 mol of the compound are added to 1mol of [Ni(cod),] the aluminum is lost and the novel structure (17) is obtained. [See also Section 25.1 Ph \ Ph Ph (17) The molecule contains one nickel atom 7-bonded to a cyclobutadiene ring and a second nickel atom .rr-bonded to a cyclopentadiene ring with an ally1 group bridging the two metal atoms. The fragmentation of the aluminacyclopentadiene which must have occurred to produce this molecule remains to be explored.The same workers have prepared a 1,4-dialuminacyclohexadieneand confirmed the structure of this compound by X-ray analysisz2' The compound reacts with lithium in ether on addition of TMEDA to give tetra-ortho-metallation of the four phenyl substituents in the six-membered ring. A silicon analogue of the same type of ring system has been prepared by a photochemical dimerization of the silicon-substituted acetylene [PhC-CSiMe2SiMe3] catalysed by [PdC12(PEt3)2].23" This reaction is the first example of the formation of a 1,4-disilacyclohexadiene by dimerization of a silacyclopropene intermediate [equation (23)]. In a related reaction silacyclo- propene in the presence of phenylacetylene and [PdC12(PPh3)2] has been found to give a silacy~lopentadiene.~~~ Disilacyclobutanes have now been prepared by 22 (a) H.Hoberg and R. Krause-Going J. Organometallic Chem. 1977 127 C29; (b) H. Hoberg R. Krause-Going C. Kriiger and J. C. Sekutowski Angew. Chem. Internat. Edn. 1977 16 183; (c) H. Hoberg V. Gotor A. Milchereit C. Kriiger and J. C. Sekutowski ibid. p. 529. 23 (a)M. Ishikawa T. Fuchikarni and M. Kurnada J.C.S. Chem. Comm. 1977,352; (b)D. Seyferth D. P. Duncan and S. C. Vick J. Organometallic Chem. 1977,125 C5; (c)L. F. Cason and H. G. Brooks J. Amer. Chem. SOC.,1952 74 4582; (d) P. R. Jones and T. F. 0. Lim ibid. 1977 99 2013; (e)M. Mickiewicz and S. B. Wild J.C.S. Dalton 1977 704. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon dimerization of vinylsilanes and apart from the original this reaction appears to be the only case of a nucleophilic addition to the double bond of a vinylsilane when the silane also contained substituents susceptible to nucleophilic attack.23d The dimerization occurs on treating Me2Si(CH=CH2)Cl with an equimolar quantity of LiBu' at room temperature in hexane but the presumed intermediate [Bu'CH2CHLi-SiC1Me2] could not be trapped or identified spectroscopically so a silaethene intermediate cannot be ruled out.The first four-membered arsenic heterocycles have now been synthesized by reductive cyclization using sodium [equation (24)].23e As(CH2CH2CH2Cl)PhI+2Na THF H,C-CH, I I H,C-AsPh 11 Cyclometallation Reactions The cyclometallation reaction (principally that of ortho-metallation) has again attracted considerable attention this year and valuable reviews of the field have The first ortho-metallation in organotin chemistry [equation (25)] (18) R = Me X = C1 (19) R=H X=C1 (20) R = C1 X = C1 (21) R=H X=Br has been The presence of five-co-ordinate tin is indicated by the appearance of the appropriate quadrupolar splittings in the Mossbauer spectrum and the X-ray structure determination of compound (19) shows the co-ordination about tin to be distorted trigonal bipyramidal with nitrogen and one of the chlorine atoms axial.Since the ease of ortho-metallation increases in the sequence (20)C (19)<(18) and the presence of Lewis bases inhibits the rearrangement a co- ordinatively unsaturated and electrophilic metal centre is required.Another example of N-aryl ortho-metallation is shown in the formation of the compound [~UC~(PE~~)~{~N(A~)(CH~)~N(&H~M~-~)}] (whose structure has been verified by X-ray analysis) from the reaction of [RUCI~(PP~~)~] with +CN(Ar)(CH2)&Ar]2 (Ar = C6H4Me-4) and subsequent treatment of the inter- mediate with PEt3.256 This is a new type of reaction of an electron-rich olefin. 24 (a)M. I. Bruce Angew. Chem. Internat. Edn. 1977 16 86; (b) H.P. Abicht and K. Issleib 2.Chem. 1977,17,1. 25 (a)B. Fitsimmons D. G. Othen H. M. M. Shearer and K. Wade J.C.S. Chem. Comm. 1977,215;(b) P.B.Hitchcock M. F. Lappert and P. L. Pye ibid. p. 196. Organometallic Compounds A variety of new cyclometallations have been reported in transition-metal systems.Among the most novel must be the di-ortho-metallation observed when the preparation of an iron-ylide complex from [(Me3Sn)2C=PPh3] and [Fe3(C0)12] was attempted.26" The species isolated has the formula [Fe2(Co)6{C(CHo)P(Ph2c6H4)}], as elucidated by X-ray crystallography in which the ligand fC(CHO)P(Ph2C6H4)] is formally derived by attack of a carbonyl fragment at the ylide carbon atom and transfer of an ortho-hydrogen from the PPh3 group. The ortho-dimetallated phenyl ring approximates to that of the Meisen- heimer intermediate found in nucleophilic aromatic substitution reactions. The structure of one of the forms of the molecule that is present in the crystal is shown in Figure 3. Figure 3 Stereochemistry of one of the formsof the molecule [Fe2(Co),{C(CHo)P(Ph2C6H4)}] that is found in the crystal state.The 30% probability ellipsoids for all atoms other than hydrogen are shown in this ORTEP diagram (Reproduced by permission from J. Amer. Chem. Soc. 1977,99 5820) A three-membered ring formed by ortho-metallation has been reported for a platinum complex also containing a carbaborane ligand. Here the methyl group of PPh2Me is metallated [see equation (26)].26b An X-ray diffraction study ~is-[Pt(PPh~Me)~C12] + 1-LiR % /Pt J,PPh R R = 2-phenyl-1,2-dicarbadodecaborane(10) 26 (a)M. R. Churchill F. J. Rotella E. W. Abel and S. A. Muckeljohn J. Amer. Chem. SOC., 1977 99 5820; (b)S. Bresadola B. Longato and F. Morandini J. Orgunometullic Chem. 1977 128 C5; (c) R. J. McKinney C.B. Knobler B. T. Huie md H. D. Kaesz J. Amer. Chem. SOC., 1977 99 2988. C.J. Cardin D. J. Cardin R. J. -Norton and K. R. Dixon of a fpur-membfred-ring ortho-metallated manganese compound [(MeC6H4)2PC6H3MeMn(CO)4], has shown that there is a planar ring system.26c The geometry of the ring is sjmilar ,to that previously reported for the cyclo- metallated compounds [Ph2PC6H41r(PPh3)L2] [L = C2H4 or CO)27a and [IrH{PPh2(C6H4)}2(PPh3)],27b which also contain a four-membered ring. In the manganese compound LMnPC is 85.2(3)" whereas LPMnC is only 67.5(3)" readily accounting for the ease of ring expansion noted in the presence of donor ligands. Internal metallation of chelating diphosphine ligands can take place in several ways. An internal oxidative addition of a benzylic C-H bond in an Ir' complex to give the Ir"' hydride (22) takes place according to equation (27).28a The reaction (Z=Cl n =O; Z=CO n = 1; E=Por As) between IrC13 and the chelating alkyl-diphosphine Bu'~P(CH~)~PBU'~ takes place over 3 days in refluxing isopropyl alcohol to give a mixture of species.These are the internally metallated Irrrrhydride (23) a bimetallic hydride (24) and a third unidentified species.28b When the hydride (23) is sublimed at about 170°C and 15 Torr an almost black sublimate is obtained which results from reductive elimination of H2 to give a complex (25) that can be formulated either as an iridium(II1) ylide or as an iridium(1) carbene. The complex is converted back into the original iridium(II1) hydride in solution under H2 For an Ir"' ylide the ylide carbon atom (C-1) is expected to deviate by about 55 pm from the Ir-C-2-C-3 plane and the X-ray crystal structure shows an actual deviation of 30pm.However the Ir-C-1 bond length of 200.6 pm is in the range expected for trigonal carbon bonded to Ir'rl and the 13C n.m.r. peak for C-1 at 66.6 pm (relative to TMS) is much more highly shielded than a typical carbene carbon. A substantial contri- bution to the bonding from the ylide form is favoured. 27 (a) G.Perego G.Del Piero M. Cesari M. G.Clerici and E. Perrotti,J. OrganometallicChem. 1973,54 C51; (6) G. Del Piero G. Perego A. Zazzetta and M. Cesari Cryst. Struct. Comm. 1974 3 725. ** (a)M. A. Bennett R. N. Johnson and J. B. Tomkins J. Organometallic Chem.1977 128 73; (b) H. D. EmpsaI1 E. M. Hyde R. Markham W. S. McDonald M. C. Norton B. L. Shaw and B. Weeks J.C.S. Chem. Comm.. 1977 589. Organometallic Compounds 235 Metallation of purely alkyl-phosphines has been reported for the two phosphines PPr"Bui and PPr; when co-ordinated to Ir'.29a Four-and five-membered rings only are formed without isolation of the Ir' bis(phosphine) complex e.g. equation (28). When an alkenyl-phosphine is used in ortho-metallation a metallocyclic ring containing the alkenyl-iridium linkage is formed [e.g.equation (29)].29b (R = But or cyclohexyl) 12 Carbene and Carbyne Complexes The use of carbene-metal complexes in organic synthesis has been Details have been published of carbene complexes of Cro Moo Wo Cr" Mo" W" Mn' Feo Fe" Ruo Co' Nio and Ni" derived from an electron-rich olefin including (NN"-olefinbmetal complexes which may be precursors in some cases e.g.in the +CN(Me)(CH2)2NMe]2-[Cr(CO)6] The relative ability of carbene ligands to function as 0-donors and ?r-acceptors has been evaluated on the basis of approximate M.O. cal~ulations.~~" Methoxy-carbenes were found to be better .rr-acceptors than amino-carbenes but all are poorer than the carbonyl ligand in agreement with experimental data. However the positive charge on C& in [Cr(C0)5{Ccarb(X)(Y)}]appears to be less than that on C of co-ordinated carbo- nyl groups which is at variance with the usual interpretation of e.g. 13C Ccarb n.m.r. shift data. However both the n.m.r.shifts and the chemical reactivity of carbene complexes may be interpreted in a wide variety of ways without clear implication as to the charge density on Ccarb.The bonding of complexes thus 29 (a) S. Hietkamp D. J. Stufkens and K. Vrieze J. Organometullic Chem. 1977 139 189; (6) ibid. 1977 134 95. 30 (a)C. P. Casey Org. Chem. 1976 1,189; (b)M. F. Lappert and P. L. Pye J. Less-Common Metals 1977 54 191; M. F. Lappert P. L. Pye and G. M. McLaughlin J.C.S. Dalton 1977 1272; M. F. Lappert and P. L. Pye ibid.,p. 1283; P. B. Hitchcock M. F. Lappert and P. L. Pye ibid.,p. 2160; M. F. Lappert and P. L. Pye ibid.,p. 2172. 31 (a)T. F. Block and R. F. Fenske J. Organometallic Chem. 1977 139 235; (b)F. R. Kreisel and P. Friedrich Angew. Chem. Internat. Edn.1977 16,543; (c) P. Friedrich G. Besl E. 0.Fischer and G. Huttner J. Organometallic Chem. 1977,139,C68. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon normally involves insignificant multiple-bond character in M-Ccarb and it is therefore unexpected that one should find a significantly shortened Re-C bond (I = 197 pm) in [Re(Cp)(CO),C(PMe,)Ph][BCI,] which is therefore best regarded as an ylide316 [see Section 141. It is equally surprising that the metal-Cca, bond length in [Ru-CMe,(C,H,)(CO>,] (I = 187.2 186.4 pm) is not very much less than that in the corresponding methoxymethylcarbene complex [I(M-Ccarb) = 189 pm] where electron demand at Ccarb can be satisfied by forming a multiple bond with the oxygen atom.31c In contrast to the behaviour of tr~ns-[WrCPh(Br)(CO)~] with unidentate donors where carbonyl substitution occurs the reaction shown in equation (30) cH2 Ph As /\ AsPh oc I /x 1 ,x [W~cMe(x)(Co)~] +Ph2AsCH2AsPh2 -+ ,W ~e W ,co (30) oc I\ /I c Ill c Me 0 gives coupling of the carbyne ligands affording a co-ordinated butyne molecule.32" That there should be such a coupling implies that carbyne complexes might be functioning as intermediates in dismutation reactions of acetylenes just as carbene complexes do in olefin metatheses.The conversion of a co-ordinated alkyne into a carbyne complex has also been demonstrated (Scheme 8) but this is not simply the Q I lii Reagents i Na[BH4]-THF-P(OMe)3 -78 "C; ii benzene solution or in solid phase; iii H[BF4]-Ac20 -78 "C Scheme 8 reverse of the dimerization [equation (30)] and it involves a most unusual proton migration.32b This migration of hydrogen is hindered if free trimethyl phosphite is 32 (a)E.0.Fischer H. Ruhs P. Friedrich and G. Huttner Angew. Chem. Internat. Edn. 1977 16 465; (b)M. Bottrill and M. Green J. Amer. Chem. Soc. 1977,99,5795;(c) M. Brookhart and G. 0.Nelson ibid.,p. 6099. Organometa1lic Compounds 237 present suggesting that a vacant co-ordination site on Mo is required if the hydrogen transfer is to occur. Treatment of [FeCH,OMe(C,H,)(CO),] with HS03F-S02C1F at low tempera- tures led to a mixture of the cations [Fe(C,H,)(C,H,)(CO),]' and [Fe(CSHS) (CO),]' possibly by disproportionation of the intermediate methylene complex [FeCH2(CsHs)(CO),]'.A secondary cationic carbene complex of iron without stabilizing heteroatom groups on the sp2 carbene was however isolated after protonic demethoxylation of the phenyl analogue [Fe-CHPhOMe(C5H5)(CO)2] at temperaturbs below -100C.32c rneso-Tetraphenylporphinatoiron(I1) [(TPP)Fe"] reacts with carbon tetrachloride in benzene under reducing conditions in a remarkable reaction affording the dichlorocarbene complex [Fe"(TPP)CCl,]. This complex the first dichlorocarbene and metalloporphyrin carbene complex shows a molecular ion in the mass spectrometer and a peak for Ccarb in the 13C n.m.r. spectrum at S = 224.7 ~.p.m.~~" hexaco-ordinate chloro(ch1oro-A carbyne)porphinatoiron formulation is however still just compatible with the data available and the results of an X-ray crystallographic study are awaited with interest.It is noteworthy that related porphyrin complexes of cobalt having the ligand [(CO,Et)CHpy]' have also been reported.336 A novel source of co-ordinated carbenes reported this year is the carbodi-imides [e.g. equation (31)]. The intermediates which can be isolated appear to have the C-metallated formamidine structure.34 HCI-THF RN=C=NR+ [L,M]"-+ [L,M{C(NR)z}]"-[L,M{C(NHR),}] (31) [L,M = Cr(CO) or Fe(CSHS)(C0)2;R = Ph or C6Hll] 13 Carbene Analogues Two papers report the generation of univalent organo-boron species (borylenes or borynes RB:) and their reactions. The reaction of MeBBr with either NaK alloy or CsK and cyclohexene afforded a mixture of products amongst which (26)-(28) were identified by their molecular ions in the mass spectrum.A species with molecular weight twice that of (26) was believed to be a dimer with B-CH3-B bridge^.^'" Photolysis of tri-l-naphthylboron at wavelengths 2350 nm (the charge-transfer maximum) gives a variety of products some of which are shown in Scheme 9. The generation of the boryne together with the normal B-C cleavage process is evidenced by the carbene-like reactions of insertion into CH and CCl bonds and by addition by which the unusual boracyclopropane ring is formed.35b 33 (a) D. Mansuy M. Lange J. C. Chottard P. Guerin P. Maliere D. Brault and M. Rougee J.C.S. Chem. Comm. 1977 648; (b)A. W. Johnson and D. Ward J.C.S. Perkin I 1977 720. 34 W. P. Fehlharnmer A.Mayr and M. Ritter Angew.Chern. Internat. Edn. 1977,16,641. 35 (a)S. M. van der Kerk J. Boersma and G. J. M. van der Kerk Tetrahedron Letfers 1976 4765; (b)B. G. Ramsey and D. M. Aujo J. Amer. Chem. SOC.,1977,99 3182. C. J. Cardin D. J. Cardin R.J. Norton and K. R.Dixon oBHIC'aH-l Reagents i hv 350 nrn; ii hv; iii hv CCI,; iv hv 350 nm,cyclohexene; v hu,350 nm cyclohexane Scheme 9 Reactions of silylenes continue to provide routes to novel systems. Thus the insertion of dimethylsilylene into a silacyclopropene [equation (32)] affords a rare example of a l,2-disilacyclobutene.36" (See also Section 10.) Me Me \/ Me,Si Me,Si SiMe, wSiMe3 + :SiMe + -+ Me\w,Me (32) Si SI-Si \ Me/\Me Me' Me The first base-stabilized silylene complex of a transition metal has been obtained via a silyliron hydride which rearranged spontaneously [equation (33)].The H [Fe(C0)~1 + HSiMe2(NEt2) !% (OC)&/ NMe2+ CO \si' silylene adduct like the related (OC),FeSi(Cl>NPhCH,NPhH can only be kept if stored at temperatures below -20 0C.36b The co-ordination chemistry of the heavier Group IV metal donors has been extended to amides of Pb and Ge dialkyls. The new adducts include [Fe(CsHs)(CO)2(GeR2Cl)] and trans-[W(CO),(GeR,),] where R = CH(SiMe3)2 as well as [Pd(q-C3Hs)Cl(PbR2)] and [W(CO)S(GeR2)I in which R = N(SiMe3)2.36CA notable feature of the Pd-Pb compound is its thermal 36 (a)D. Seyferth and S. C. Vick J. Organometallic Chem. 1977 125 C11; (b)G. Schrnid and E. We@ Angew. Chem. Internat.Edn. 1977 16 785; (c)M. F. Lappert S. J. Miles P. P. Power A. J. Carty and N. J. Taylor J.C.S. Chem. Comm. 1977 458; (d)W. W. du Mont and H. J. Kroth Angew. Chem. Internat. Edn. 1977 16 792; (e) G. Dousse and J. SatgC Helv. Chim. Acta 1977 60 1381. Organometallic Compounds stability when compared with the free ligand PbR,; another feature of the amides in general is their donor strength despite their higher first ionization potentials (by about 1eV) with respect to isoelectronic alkyls. The first bis(phosphido)tin(II) derivative has been prepared containing the bulky di-t-butylphosphido ligand [equation (34)].36e 2KPBu +Et3PSnC12 -+ 2KC1+ Et,P + (BU;P)~S~ (34) The pure crystalline Snrr derivative was obtained in about 40% yield and it is dimeric in benzene.N.m.r. data indicate a symmetrical structure with three-co- ordinate tin and bridging PBu groups. Synthetic approaches and the charac- terization of bifunctional germylenes have been briefly reviewed.36e 14 Transition-metal Ylide Complexes This section heading excludes ylide chemistry of elements of the main groups a topic we hope to cover next year. The ability of the ylide ligand to stabilize unusual stereochemistry has been demonstrated by the preparation of the tetrahedral mercury(I1) species [Hg(CH2PMe2CHPMe2CH2),] (see Section 6). Less well studied is the behaviour of transition-metal ylide complexes as cata- lysts. The complex (29) prepared from [{Rh(cod)Cl},] and excess Me3P=CH2 has CH2 /\ [(cod)Rh PMe21 \/ CH2 been shown to catalyse the complete hydrogenation of hex-1-ene in lO-’M solution in benzene at 30°C with no detectable isomerization to internal olefins (at 16% ~ompletion).~’ Replacement of (cod) by CO generated the complexes [Rh(CO)2{(CH2)2PMe2}] and [(CO)2Rh(p-CH2PMe2)2Rh(C0)2], which as a mix- ture in solution have been found to catalyse the conversion of methyl iodide into methyl acetate although they are completely inactive in the hydroformylation of olefins.The exact nature of the catalytic species has not been established. High thermal stability is a feature of complexes containing both ylide and diphos- phomethanide ligands. Thus the palladium species [Pd{HC(PPh2)2} {(CH2)2PEt2}] is obtained by treating [PdCl,(PMe,),] sequentially with LiCH(PPh2) and Et2P(Me)=CH2.*’ Cyclic ylide complexes of Ti and Zr have been prepared the ring system that was obtained depending on the choice of starting A bridged dimeric system is obtained from [TiC12(Me2N),] and Me,P=CH [equation (35)] whereas [Ti(CSH5)2Cl2] and [{Ti(CsHs)2Cl}2] give the same complex [Ti(C,H,),{(CH2)2PMe2}1.37 R. A. Grey and L. R. Anderson Inorg. Chem. 1977 16,3187. ’* (a)H. Schmidbaur W. Scharf and H Fueller Z. Naturforsch. 1977 32b,858; (6) F. R. Kreissl K. Eberl and P. Stueckler Angew. Chem. Internat. Edn. 1977,16,654; (c) F. R. Kreissl P. Stueckler and E. W. Meinecke Chem. Ber. 1977 110 3040; (d)F. R. Kreissl and W. Held ibid. p. 799; (e)W. C. Kaska R. F. Reicheldorfer and L. Prizant J. Organometallic Chem. 1977 128 97. C.J. Cardin D. J. Cardin R.J. Norton and K. R. Dixon PMe3 I1 C / \ [TiCI2(NMe2)2]+ Me3P=CH2 -+ (Me2N);Ti Ti(NMe& (35) \C/ I1 PMe3 A 'semi-ylidic' ligand system [of the type M-C(PR:),R'] has been described for the first time38b [see equation (36)]. This new type of ligand contains two PMe3 + I [BC14]-+PMe 2 (C5H5)(C0)2Re-C-Ph [BCLI-(C5H5)(C0)2Re=C'PMeT I \Ph PMe3 ! 1 magnetically equivalent PMe3 groups. However loss of PMe from the complex to give the starting ylide complex takes place readily at room temperature. The same authors also report the synthesis of cationic ylide complexes by addition to the triple bond of cationic carbynes [equation (37)].38' Addition of PMe3 to a tung- M = Mn or Re; L = (CsHs)-sten-carbene complex has also been reported [equation (38)].38dA 'homoleptic' ylide complex of tungsten is obtained by displacement of methyl vinyl ketone (OC)5W=C +PMe3 -+ (OC),W-L-R' (38) [ :::I [ Le3] R',R2 = Ph 2-thienyl or 2-furoyl ligands by Ph3P=CH2 [equation (39)].38' This complex does not form adducts with CO and is cleaved by HCl to give Ph,PMeCI.The apparent co-ordinative unsaturation in this complex suggests there may be an interaction between the phenyl rings and the metal. [W(CH2=CHCOMe)3]+ 3Ph3P=CH2 %&$$ [W(CH2-PPh3)3] + 3CH2=CHCOMe (39) Finally sulphonium ylides have been shown to give a variety of products on reaction with trimethylgold(II1) complexes [equations (40) and (41)].39 [AuMe3(PPh3)]+ CH2=S(0)Me2 THF [Au{CH2S(0)Me2)Me3] (40) [AuMe3(PPh3)]+ NaH + Me3SCI -+ [Au{CH2SMe2}Me3] (41) '' J.P. Fackler and C. Paparzos J. Amer. Chem. SOC.,1977,99 2363. Organometallic Compounds 24 1 15 -Metal-Carbon Multiple Bonds The last two Reports have contained accounts of new chemistry of niobium and tantalum primary alkylidene systems [L,MCHR] and such systems are still rare for other transition metals (see Section 12). The reaction of neopentylidene complexes with olefins has now been studied (Scheme The ligating neopentylidene Ph H \/ CH CMe M = Ta(CSHS)C12 isolable metallocycle Reagents i HCI -78 "C; ii CzH4-pentane 25 "C; iii PhCH=CH;?; iv MeCH=CH2 0 "C. Proposed metallocyclobutane intermediates are shown in brackets. Scheme 10 carbon atom behaves as a nucleophile reacting with HCl at -78 "C in toluene to give [Ta(C5HS)C13(CH2CMe3)].With olefins the reaction is believed to occur uia metallocyclobutane intermediates to give the products shown in Scheme 10 with only one direction of addition of unsymmetrical olefins. This selectivity bears a direct relationship to that observed for addition of an olefin to a postulated tungsten alkylidene complex in an olefin metathesis A further striking abservation is the isolation in 95% yield of the metallocycle I [TaCH2CH2CH2CH2(C5H5)C12] from the reaction with ethylene. There is a clear parallel here with the demonstration that a titanocycle and a nickelocycle are in equilibrium with the diolefin complexes (Section 10 ref. 18c). The tantalocycle appears to be indefinitely stable at -30°C is cleaved by Br2 (giving mainly 1,.l-dibromobutane) and with CO undergoes the insertion characteristic of metal- locycles to give cyclopentanone.The metallocycle-forming reaction seems to be a general one and will undoubtedly be studied in greater detail. A formally related metal-carbon multiply bonded system is found in the rhenium complex [(C,H5)(CO)2Re(=C=CPh-CPh=CH2)Re(CO)2(C5Hs)], in which the 4" (a)S. J. McLain C. D. Wood and R. R. Schrock J. Arner. Chern. Soc. 1977,99 3519; (b) C. P. Casey H. E. Tuinstra and M. C. Saernan ibid.,1976 98 608; and references therein. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon ligand system is stabilized by phenyl sub~titution.~~ X-Ray structure determination has revealed that the second Re atom is 77-co-ordinated to the terminal double bond of the ligand as shown in (31).Ph In 1976 several groups reported new routes to silaethenes (see last year's Report). This year the addition of LiBu' to SiMezC1(CH=CHz) has been shown to give products which may well involve an intermediate silaethene and a related reaction appears to involve silabenzene intermediate^.^^" (See also Section 9). In this case the precursor is a silacyclohexadiene prepared by copyrolysis of cyclo- pentadiene and (ClzSiMe)z (a source of silylene) at -600 "C. The base-promoted elimination of HCl was achieved using N-lithiodisilazane and the silabenzene produced was trapped with hexafluorobut-2-yne. (See Scheme 11.) This is the first Me' N(SiMe Me Reagents i (Me3Si)2NLi-Et20; ii CF~CGCCF~ Scheme 11 example of a silabenzene intermediate.The synthetic utility of silaethenes has been increased by the establishment of a new route to these systems requiring very mild reaction conditions.426 This consists in thermolysis of Me2Si(OR)-C(Li)(SiMe3)2 which has been shown to undergo a variety of [4+ 21 and [2 + 21 cycloadditions. In Group V the first acyclic compounds containing an isolated As=C double bond have been synthesized [equation (42)].42' The rearrangement reaction to give the As=C linkage is only complete after 12 h at 50 "C. 41 N. E. Kolobova A. B. Antonova 0. M. Khitkova M. Yu. Antipin and Yu. T. Struchkov J. Organometallic Chem. 1977 137 69. 42 (a)T. J. Barton and D. S. Banasiak J.Amer. Chem. Soc. 1977,99,5199; (b)N. Wiberg and G. Preiner Angew. Chem. Internat. Edn. 1977 16 328; (c) G. Becker and G. Gutekunst ibid. p. 463; (d)M. J. Hopkinson H. W. Kroto J. F. Nixon and N. P. C. Simmons J.C.S.Chem. Comm. 1976 513; (e) C. Thomson ibid. 1977 32-2. Organometallic Compounds SiMe3 OSiMe3 The recent synthesis of the phospha-alkenes H2C =PH F2C=PH and H2C=PCl has shown that pT-pT bonding can sometimes occur for second-row elements.42dAb initio calculations of bond lengths total energies and bond angles in these molecules are now complete and the highest occupied orbital (HOMO) in these molecules is a .Ir-orbital whereas for H2C=NH it is a a-~rbital.~~' 16 Transition-metal Alkenyls The first homoleptic alkenyls of the transition metals [MR,] (M =Ti Zr or Hf) and related cyclopentadienyl-metal alkenyls have been prepared from the appropriate halide and organolithium reagent [equation (43)].MC14+4LiCPh=CMe2 + [M(CPh=CMe2)4] +4LiCI (43) The alkenyls are thermally stable (if M = Zr or Hf; that of Ti is less so) they are readily decomposed by water and other protic reagents and undergo insertion reactions into the M-C(sp2) bonds e.g. with methyl i~ocyanide.~~ In contrast to these P-hydrogen-free derivatives of the early transition elements stable ethenyls of platinum (and trifluoroethenyls) have been obtained from the appropriate halide cis-[PtClzL2] and the trimethyltin alkenyl. Reactions of one complex are shown in Scheme 12. Related reactions have also been demonstrated to occur for [ClPtP2C-CH=CH2] [Pt{C(NHBu")NHC6H40Me-p}-1 P = PEt2Ph [Pt (CH= CH2)Me12P2] Reagents i SiEt3H reflux; ii HCI-OEt2 1 equivalent; iii Cl2 (1 equivalent) in benzene; iv NaI- MeZCO; v MeI 4 weeks; vi MeO2CCrCCOzMe; vii pMeOC6H4NC followed by reflux in benzene; viii Ag[PF6]-MeOH; ix pMeOC6H4NC; x NBu"H2 reflux.Scheme 12 43 C. J. Cardin D. J. Cardin J. M. Kelly R. J. Norton and A. Roy J. Orgunomefullic Chem. 1977 132 C23. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon platinum(I1)phenylalkynyl species obtained via Me,snc~XR.~~" An accurate single-crystal X-ray structure determination of the complex trans-[Pt(CH=CHz)Cl(EtzPhP),l reveals that there is little if any multiple-bond character in the Pt-C(sp') bond.44* The length of the bond is 203(2) pm in good agreement with data on the complex ~~U~~-[P~B~(CH=CHP~)(PP~~)~].~~~ In the reaction of isopropenylacetylene with cis-[PtC12(PPh3),] in ethanol in the presence of hydrazine hydrate the complexes (32) (33) and (34) are formed.(32) (33) (34) Complex (32) is presumably formed by insertion of the acetylene into the platinum hydride that is formed by reduction of the cis-dichloride with hydrazine while the formation of (33) is believed to involve the reaction of a second mole of the acetylene together with hydrazine; the latter functions as a base to facilitate elimination of HCl. Single-crystal X-ray data on complexes (32) and (33) show 1[Pt-C(sp2)] of 207(1) and 209(2) pm respectively again suggesting the absence of back-bonding to the 0-unsaturated hydrocarbyl ligand; l[Pt-C(sp2)] in (33) is 199(1)pm.44d Related insertions of acetylenes into M-C bonds have provided a route to other alkenyl complexes (see also Section 18) particularly of the Group VIII metals.Examples with nickel and ruthenium are shown in equations (44) and (45). Curiously the metal complex fails to insert hex-3-yne under similar reaction The crystal structure of the ruthenium product [equation (45)] shows MeCECMe+trans-[Ni(PPh3)2BrPh] -* tran~-[Ni(PPh~)~BrCMe=CMePh](44) (c5H5)\y5H [Ru(PP~~)~M~(C~H~)] + HCECCF~+ CF3 (45) c4c H-C4 I CF the novel cumulene ligand to be almost planar. The unusual configuration of the oligomerized acetylene is believed to arise by metal co-ordination (an q '-adduct) 44 (a)C.J. Cardin D. J. Cardin and M. F. Lappert J.C.S.Dalton 1977,767; (6)C. J. Cardin and K. W. Muir ibid. p. 1563; (c) J. Rajaram R. G. Pearson and J. A. Ibers J. Amer. Chem. SOC. 1974 96 2103; (d) A.Furlani M. V. Russo A. C. Villa A. G.Manfredotti and C. Guastini J.C.S. Dalton 1977 2154. 45 (a)$. J. Tremont and R. G. Bergman J. Organometallic Chem. 1977,140 C12; (b)M. I. Bruce R. C. F. Gardner J. A. K. Howard F. G. A. Stone M. Welling and P. Woodward J. C.S.Dalton 1977,621; (c)P. M. Maitlis Accounts Chem. Res. 1976 9 93; (d)E. A. Kelly P. M. Bailey and P. M. Maitlis J.C.S. Chem. Comm. 1977 289; (e)I. G. Dinulescu S. Staicu E. Avram F. Chiralu and M. Avram J. Organometallic Chem. 1977 136 C15. Organometallic Compounds followed by elimination of methane and subsequent coupling under the influence of further liganding acetylenes; see Scheme 13.45b Me H Me $p' / \/ -MeH Ru + Ru I I I Ru €4 H m CF c*ccF3 I CF3 Reagent i HCECCF~.Scheme 13 a-Butadienyl complexes have been postulated in another acetylene-coupling reaction; that by palladium species which subsequently affords cyclobutadiene One such complex has now been isolated as a bipyridyl adduct obtained by treatment of [PdCl,(PhCN),] with 2 mol of Bu'CCECMe followed by bipyridyl. A crystal-structure determination of an isomeric species obtained by heating the initial (kinetically preferred) product shows a 0-4-chloro- 1,4-di-t- butyl-2,3-dimethylbutadienylligand co-ordinated to an essentially square-planar palladium atom bearing C1 and 2 N (of bipy) as the ligating The iso- merism arises because of restricted rotation about the ligand 2,3-bond owing to non-bonding interactions of the substituents.The intermediacy of such ligand types in the route to cyclobutadienyl complexes is demonstrated by the conversion of the initial acetylene products (i.e. without addition of bipy) into the cyclic complexes the structure of one of which was also confirmed by an X-ray A similar a-butadienylpalladium complex formed from another acetylene bearing bulky substituents [equation (46)] undergoes ring closure to produce the cor- responding benzotropone in high yield on treatment with chromium(v1) oxide in ~yridine.~~' 2,4,6-Me3C6H2CZCPh+ [PdCl,(PhCN),)] -+ Me C.J:Cardin D. J. Cardin R. J. Norton and K. R. Dixon A remarkable di-rhodium alkenyl has been obtained from insertion of the acetylene into the metal-metal bond of a new Rh" octaethylporphyrin dimer (Scheme 14).46" The rhodium hydride gives the dimeric Rh" porphyrin complex [Rh"'(OEP)Cl] I [Rh"'(OEP)H] iii 1ii /liv [Rh'(OEP)]-[Rh1I(OEP)]3+H2 I. OEPH2= octaethylporphyrin [Rh(OEP){CH=CPhRh(OEP)}] Reagents i H2-MeOH ii NaBH4; iii acid; iv benzene solution; v PhCzCH Scheme 14 which is diamagnetic and presumably contains a Rh-Rh bond into which the acetylene inserts itself. A new synthesis of alkenyl-cobalt porphyrins (octaethyl- porphyrin or meso-tetraphenylporphyrin) involves the reaction of the cobalt(m) bromide derivative with a variety of dia~oalkanes.~~~ Equivalents of 2-5 diazoalkane are required for the reaction which often proceeds in high yield in a few minutes in dichloromethane at 25 "C.The reaction is believed to occur via an intermediate (35) that has the organic adduct bridging Co and one of the porphyrin N atoms (an established structure in related 17 Paramagnetic Species The first stable paramagnetic carbene complexes have been prepared47 (see Scheme 15). The dinuclear dpe-bridged dication is remarkable in having two low-spin d7 Fe atoms without apparent Fe-Fe interaction. The compounds vary in hydrolytic and thermal stability from species that are observable only at tempera- tures below -20 "C (e.g. [Fe(C0)4LMe]') to the thermally and hydrolytically robust [Fe(C0)2LMe(PPh3)2]C.As noted in last year's Report interest is growing in paramagnetic organo- metallic compounds of the early transition elements. The new complexes [NbCp2Me2] and [Ta(MeCphMe2] have been isolated as red crystalline solids from the reaction of the corresponding halide (C1 and Br respectively) with LiMe. The 46 (a)H. Ogoshi J. Setsune and Z. Yoshida J. Amer. Chem. SOC.,1977,99,3869; (b) H. J. Callot and E. Schaeffer Tetrahedron Letters 1977 239. 47 M. F. Lappert J. J. MacQuitty and P. L. Pye J.C.S. Chem. Comm. 1977 411. Organometallic Compounds 247 [Fe(CO)3(LMe)Lf] / [Fe(CO)f(LMe)Lf]+[BF4]- [Fe(CO)2(LMe)Li]'[ BF41- [Fe (COML~")I LMe= CNMeCH2CH2&Me;L' = LMe PPh3 or PEt3 Reagents i Ag[BF4]-THF 20 "C; ii THF or L'-THF; iii dpe hv PhMe 25 "C;iv Ag[BFd]-THF 20 "C Scheme 15 niobium complex on thermolysis at 130°C affords a gas consisting of 96-8% methane 1-2% ethylene but no ethane in contrast to platinum(1v) methyls which afford that gas as the only hydrocarbon.Whether the H-abstraction occurs from a methyl group or from a cyclopentadiene has not been e~tablished.~~" Synthesis of paramagnetic organometallics with the chelating o-(dimethylamino- methy1)phenyl and related ligands has been extended to novel complexes of Tirrr CrIII VIII Mn" and CrII.48b.c.d P-Diketonate derivatives of Ti"' have provided particularly convenient starting materials for the preparation in high yields [e.g. 96% for [Ti(dpm){C6H4CH2NMe2-o}] from [Ti(dpm)C12(THF)2] (dpmH = dipivaloylmethane) of such paramagnetic The reaction of [LiC6H4CH2NMe2-o] with Mn12 or CrC12 in THF initially affords adducts [Li2(THF)2MX2(CH2C6H4NMe2-o)2] (M = Mn or Cr) from which solvated LiX can be removed by treatment with benzene.The resulting paramagnetic Cr" and Mn" products are very sensitive to air and are apparently associated to some extent (70-80%) in solution. The structure of the manganese derivative based on a single-crystal X-ray study is shown in (36). One of the ortho-NMe2 groups remains unco-ordinated probably as a result of steric constraints. The Mn-Mn distance is 281.O pm.48d 48 (a) L. E. Manzer Znorg. Chem. 1977 16 525; (b) J. Amer. Chem. SOC. 1977 99 276; (c) J. Organometallic Chem. 1977,135 C6; (d)L. E. Manzer and L.J. Suggenberger ibid. 1977,139 C34; (e)F. W. Van der Weij H. Scholtens and J. H. Teuben ibid. 1977 127 299. 248 C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon The titanium(II1) neopentyl [see equation (47)] reacts with nitrogen gas at -90 "C giving [Ti(C5H5)2(CH2CMe3)2N2]; this is a result of great interest in view of the known ability of related [Ti(C5H5),R] species (R = e.g. CH2SiMe3 or 2,6-Me2C6H3) to reduce dinitrogen in the presence of reducing agents such as sodio- na~hthalene.~'" [{Ti(C5H5)2Cl}2]+Me3CCH2MgCI+ [Ti(C5H5)2(CH2CMe3)] (47) 18 Photochemistry The photochemical reaction of M(C5H5)2Me2 compounds (M = Ti Zr or Hf) with acetylenes to yield metallocycles has been known for some time.49" It has now been shown that under similar conditions the titanium compound also affords the mono-insertion product (37) in 18% yield496 [equation (48)].When Ph Me / [Ti(C5H5)2Me2]+PhCrCPh % (C,H,),Ti 3;; + (C,H,)*Ti (48) )=.(Me Ph Ph Ph (37) bis(pentafluoropheny1)acetylene is employed the sole organometallic compound isolated is the mono-insertion product. The photorearrangement of thionicotinic S-aryl esters to give azathiaxanthones finds a novel parallel in the rearrangement of the pseudo-isoelectronic (2-chlorot selenonicotinic acid ester shown in equation (49).50a 0 hv (49) ___) + -HCI MeeSe-Sea'Me ' The rather curious reaction of (a$,y,S-tetraphenylporphinat9)ethylaluminium with COz induced by visible light in the presence of 1-methylimidazole has been reported to afford low yields of tetraphenylporphinatoaluminium pr~pionate.~'~ The photolysis of ethylenebis(triphenylphosphine)platinum(O) in different solvents and at different wavelengths has been studied5" (Scheme 16).The pho- tolysis product (39) is presumably formed by an initial orthometallation followed by insertion of C2H4 into the resulting platinum-hydrogen bond. The structure of (38) has not been determined though an orthometallation followed by elimination of benzene is thought to occur. 49 (a)H. Alt and M. D. Rausch J. Amer. Chem. SOC.,1974,96,5936;(b)W. H. Boon and M. D. Rausch J. C.S. Chem. Comm. 1977,387. (a)B.Pakzad K. Praefcke and H. Simon Angew. Chem. Internat. Edn. 1977 16 319; (6)S. Inoue and N. Takeda Bull.Chem. SOC. Japan 1977,50 984; (c)S. Sostero 0.Traverso M. Lenarda and H. Graziani J. Organometallic Chem. 1977,134 259. 249 Organometallic Compounds PPh, (39) Reagents i 280 nm EtOH; ii 254nm CH2C12 Scheme 16 Thermolysis of [Th(C5H5)3Pri] is known to cause intramolecular abstraction of hydrogen from a cyclopentadienyl ring and to result in the quantitative formation of propane [equation (50)].5'" Photolysis of the same compound has been reported 2[Th(CsH5)3Pri1 [{T~(C~HS)~(CSH~))~I (50) +2C3Hs to afford [Th(C5H5),] in high yield with roughly equal quantities of propane (55%) and propene (47'/0).~*' The fact that comparable quantities of propane and pro- pene are produced combined with other evidence leads to the conclusion that &elimination occurs according to Scheme 17.Thus the photolysis of [Th(C5H5)3Pri] not only provides a convenient new route to low-valent organo- thorium compounds but also appears to be the first example of photochemically induced &elimination of an organometallic compound which decomposes ther- mally by a different pathway. Investigations into thermal and photochemical reactions between compounds of the type [M(C5H5)(CO)3R'] (M = Mo or W; R' = Me CF3 or PhCH,) and acetyl- enes R2CCR2 (R*=Me or H) ~ontinue.~~"-~ The two types of reaction which occur are the formation of 16-electron 7r-complexes [M(C,H,)(CO)(R'CCR')R'] and the formation of 0-bonded compounds [M(C5H5)(CO)2(CR2CR2COR')]. [Mo(C5H,)(CO),(CF3)] forms a 7r-complex on photolysis in the presence of MeC-CMe."" In contrast [Mo(C5H5)(C0),(Me)] both thermally52" (with MeCrCMe) and photo~hemically~~~ (with HCrCH) yields the 0-bonded compounds.[Mo(C,H,)(CO)~(CH~P~)] forms the 0-bonded compound thermally with HCGCH.~~" [W(C,H5)(CO)3(Me)] has been reported to form a-bonded compounds with both MeCrCMe"" and HCGCH~~~ photochemically although Alt bas since reported the simultaneous production of a ?r-complex under these conditions.52c '' (a)T. J. Marks and W. A. Wachter J. Amer. Chem. Soc. 1976,98 703; (b)D.G.Kalina T. J. Marks and W. A. Wachter ibid. 1977,99,3877. 52 (a)F.G. A. Stone A. J. Welch and P. Woodward J.C.S. Chem. Comm. 1976 714; (6)H. G.Alt,and W. Stadler Z. Naturforsch.,1977,32b 144; (c)H. G.Alt J. Organometallic Chem.1977,127,349; (d) Angew. Chem. Internat. Edn. 1976,15 759. 250 C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon 19 Syntheses involving Metal Atoms The synthesis of organometallic compounds uia the direct reaction of metal atoms with appropriate ligands continues to receive attention and has been reviewed.53" A mechanistic investigation of the oxidative addition of alkyl halides to palladium vapour has been carried Studies involving trapping experiments free- radical scavenging and examination of the decomposition products revealed that insertion of a metal atom into the carbon-halogen bond occurs via a caged radical pair. Co-condensation of ArCH2Cl species with palladium vapour has been found to afford [{Pd(q 3-ArCH2)Cl}2] which oligomerize slo~ly.~~~.~ Treatment of the dimer with a phosphine causes a T to CT rearrangement of the organic group followed by cleavage of the chloride bridges in contrast with the behaviour of allylpalladium halide dimers in which bridge cleavage generally occurs first (Scheme 18).Et,P ArCH2Cl \/ Pl Pd (vapour) + Io"p I ArCH2-Pd-CI I PEt3 Reagents i PEt3 (2 equivalents) Scheme 18 The preparation and use of highly reactive metal powders has been reviewed by Rieke.54" The powders were produced by reduction of a metal salt with an alkali metal (usually potassium) in an ethereal or hydrocarbon solvent [equation (5 l)]. MX,+nK + M+nKX (51) Activated magnesium formed by this method reacts quantitatively with bromo- benzene within 5 min at -78°C.Addition of KI to the magnesium salt before reduction results in the generation of a more reactive powder. The greater reac- tivity of the powder is apparently due to the fact that the particles are smaller. Reactions of 'activated' and 'KI-activated' magnesium powders with alkyl or aryl halides have been and include syntheses of previously inaccessible '' (a)P. L. Timms and T. W. Turney Adv. Organometallic Chem. 1977,15,53; (b)K. J. Klabunde and J. S. Roberts J. Organometallic Chem. 1977 137 113; (c) J. S. Roberts and K. J. Klabunde ibid. 1975 85 C13; (d) J. Amer. Chem. SOC.,1977,99,2509. 54 (a)R. D. Rieke Accounts Chem. Res. 1977,10,301; (b)R. D. Rieke and S. E. Bales J. Amer. Chem. Soc. 1974,96 1775; (c) R. D. Rieke and S. J. Uhm Synthesis 1975,452; (d) L.Chao and R. D. Rieke J. Organometallic Chem. 1974 67 C64; (e) Synth. React. Znorg. Metal-Org. Chem. 1975 5 165; (f) R. D. Rieke W. J. Wolf N Kujundzic and A.N. Kavalinnas J. Amer. Chem. Soc. 1977,99,4159; (g) R. D Rieke K. Ofele and E. 0.Fischer J. Organometallic Chem. 1974 76 C19; (h) G. Page Ph.D. Thesis University of N. Carolina 1974; (i) R. D. Rieke and L. Chao Synth. React. Inorg. Metal-Org. Chem. 1974 4 101; (j) J. E. McMurtry and M. P. Fleming J. Org. Chem. 1976 41 896. Organometallic Compounds 251 Grignard reagents such as MeC6H4MgF. The high reactivity of the magnesium powders allows Grignard reactions to be carried out at -78 “C and permits exten- sion of the Grignard reaction to thermally unstable compounds. Similar methods have been used to generate reactive zinc powders which react rapidly wih alkyl bromides in refluxing THF giving quantitative yields of dialkyl-zinc compounds.The zinc powders have been employed to produce a great increase in the yield from the Reformatsky rea~tion.’~‘ Activated indium reacts with alky154d or phenylS4‘ iodides to yield the corresponding R21nI compounds in high yields. InMe3 was prepared in good yield from the metal powder and dimethylrnerc~ry.’~‘ The powders obtained on reduction of nickel palladium or platinum salts proved to be rather unreactive towards oxidative addition to carbon-halogen Reduc-tion of phosphine complexes of the salts or reduction of the salts in the presence of phosphines leads to reactive slurries which undergo oxidative addition with pentafluorobromobenzene in good yield [equation (52)].The procedure has been PEt3 I Ni + 2Et3P+ C6F5Br -+ C6FS-Ni-Br I PEt3 extended to unreactive halides. For example the palladium slurry reacts with chlorobenzene to yield chlorophenylbis(triethylphosphine)palladium(~~) in 54% yield. Highly reactive chromium powders have been generated by the reduction pro- c~ss.’~~ The compound [Cr(CO),] previously unobtainable by the direct reaction of CO with the metal was obtained in 50% yield by the reaction of a slurry of chromium powder in THF with CO in an autoclave. Other reactive metal powders that have been generated include those of thallium,54h aluminium,54i and titani~m.’~’ 20 Carbonyl Complexes Interest in metal carbonyl chemistry continues to centre around those areas repor- ted by us in 1975 and 1976.Thus the present section presents a concise selective overview and the next deals with the especially rapidly developing field of fluxional processes. Synthesis.-Matrix-isolation techniques continue to be important and the reaction of titanium atoms with CO at 10-15 K has been shown to give [Ti(CO),] as the product having highest ~toicheiometry.’~~ This preparation completes the series of binary carbonyls for the first transition series Ti to Cu. 1.r. studies suggest that it has a slightly Jahn-Teller-distorted octahedral structure as expected for a low-spin d4 c~rnplex.’’~ [Ti(N2),] may be obtained under similar conditions and the elec- tronic spectra of both molecules and of their vanadium and chromium analogues have been assigned.556 At the other end of the transition series matrix isolation of [Au(CO),] (n = 1 or 2) completes the observation of binary carbonyls for the Cu s5 (a)R.Busby W. Klotzbucher and G. A. Ozin Inorg. Chem. 1977,16,822; (6)A. B. P. Lever and G. A. Ozin ibid.,p. 2012; (c)D. McIntosh and G. A. Ozin ibid.,p. 51; (d)G. A. Ozin Accounts Chem. Res. 1977,lO. 21. 252 C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Ag Au group; co-condensation of Au atoms with a mixture of CO and O2at 10 K yields a single compound which can be formulated as a peroxyformate [(CO)Au{C(0)-O-~}].”c This complex decomposes to CO at 30-40 K and might therefore be considered a model for heterogeneous catalysis of the oxidation of CO at a gold surface.This approach may be more generally applied; for example a number of matrix-isolated M(CZH4) complexes have been considered as models for chemisorbed ethylene and the subject has been reviewed this year.55d Metal carbonylate anions have long occupied a central position in the synthesis of carbonyl derivatives and several groups this year have been interested in improved methods of preparation of these important intermediate^.^^ [Co(CO),]-may be simply obtained by the halide-ion-catalysed disproportionation of [Co,(CO),] or [co4(co)12],”aor alternatively by phase-transfer catalysis using [co2(co)g] in a benzene/aqueous NaOH system benzyltrierhylammonium chloride being the More generally Li[BEt3H] may be employed as a convenient reducing agent to generate [Co(CO),]- [Mn(CO)5]- [Mo(~’-C,H,)(CO),]- or [Fe(q5-C5H5)(C0),]- from the corresponding neutral dimeric carbonyl complexes.The method is more convenient than the traditional reductions with alkali metal and it produces only volatile by-products (BEt3 and H2)? Preliminary work on the novel trinegative ions [M(C0)4]3- (M=Mn or Re) was reported last year. Full details have now been and the work has been extended to the or reduction of N~[CO(CO)~] [M4(C0)12] (M=Rh or Ir) by sodium in liquid ammonia at -78 “C. The extremely air- and moisture-sensitive products have not been fully characterized but reactions with [EPh3Cl] (E=Ge Sn or Pb) give [Co(CO),(EPh,),]-anions which can be isolated as air-stable [NEt,]’ salts.These results suggest that the initial reduction products contain [M(C0),I3- ions.56d The addition of carbonylate anions to small clusters is a potentially interesting approach to designed syntheses of larger clusters. Thus the reaction of [Fe(C0)4]2- with followed by acidification yields [F~RU~(CO)~,H~] [Ru~(CO)~~] in good yield. A similar reaction using the mixed clusters [OsRu2(CO),2] and [OS~RU(CO)~I (obtained from copyrolysis of [Ru~(CO)~~] gives with [OS~(CO)~~]) [F~RU~OS(CO)~~H~] These hydrides are the first exam- and [F~RUOS~(CO)~~H~]. ples of clusters that contain three different transition Di-iron enneacarbonyl was prepared and correctly formulated in 1905. However its ruthenium and osmium analogues were unknown until recently when a preliminary report of the U.V.irradiation of [OS(CO)~] at -40°C showed that [Os2(CO),] was formed. The full paper published this year confirms this result but the analogous reaction of [Ru(CO)~] gives a very unstable product which is only partially characterized as [RU~(CO)~]. On the basis of i.r. measurements [OS,(CO)~] has been assigned a C20structure with two Os(CO) groups bound by a metal-metal bond bridged by only one CO group. It reacts with hydrogen to give the known [OS~(CO)~H~] and this compound is also formed by a hydrogen-abstraction process in reactions of [Os2(CO)9] with PEt3 or PPh3.” s6 (a)P. S. Braterman B. S. Walker andT. H. Robertson,J.C.S. Chem. Comm.,1977,651; (b)H. Alper H. Des Abbayes and D. Des Roches J.OrganomefallicChem. 1976,121 C31; (c) J. A. Gladysz G. M. Williams W. Tam and D. K. Johnson ibid. 1977 140; C1; (d)J. E. Ellis and R. A. Faltynek J. Amer. Chem. SOC.,1977 99 1801; J. E. Ellis P. T. Barger and M. L. Winzenburg J.C.S. Chem. Comm. 1977,686; (e)G. L. Geoffroy and W. L. Gladfelter J. Amer. Chem. SOC.,1977,99 304. 57 J. R. Moss and W. A. G. Graham J.C.S.Dalton 1977 95 and references therein. Organometallic Compounds 253 The alkylidynetricobalt clusters [CO,(CO)~(CR)] are very well known and have been extensively st~died.'~" The system is unique being the only organometallic cluster for which a substantial organic chemistry is accessible. An example is the ability of the cluster to form highly stabilized carbonium ions [Co3(C0),(CCHR)]' by the reaction of the alcohols [Co3(CO)9{CCH(OH)R}] with acids and it has now been shown that protonation of the vinyl derivatives [Co3(C0)9(CCR=CH2)] to yield [Co3(C0),(CCRCH3)]' is an alternative and probably more general route to these useful intermediate^.'^^ One unusual example of these carbonium ions is [Co,(CO),(CCO)]' originally prepared by the reaction of [CO~(CO)~(CCO,R)] complexes with acetic anhydride and acid and a useful intermediate in the synthesis of various [Co3(CO),{C(0)Y}] derivatives (e.g.Y = OR NR2 SR R or H) which can be produced by the addition of nucleophiles. The ion [Co,(CO),(CCO)]' has now been prepared by a novel reaction between [Co,(CO),(CCl)] and aluminium trichloride. The cation is obtained in high yield (83%) as its [A1C14,A1C13]- salt and the yield is not increased by the addition of CO.Thus the reaction involves a remarkable and very efficient transfer of CO on to the carbon atom in the cluster and also transfer of CO between ~l~~ter~.'~~ The reaction is an interesting contrast with the familiar insertion of CO into metal- carbon bonds. In view of the remarkable stability and extensive chemistry of the alkylidynetricobalt clusters it is somewhat surprising that analogues involving other metals have not been found. The hydrides [M3(CMe)(CO),H3] (M = Ru or 0s) are well established; they have a similar M3Ccore except that each M-M bond is bridged by a h~dride.'~~ The only other analogue appears to be [Ni3(CPh)(qs- C5H5)3] claimed to exist on the basis of 'H n.m.r.and mass spectroscopic evi- den~e.'~~ Thus the isolation of (40) as a product from the reaction of methyl- Et 'Fe' acetylene with [Fe3(C0)12] is especially interesting. The structure has been established by X-ray diffraction and the Fe3C core is structurally similar to those in the Ru Os and Co compounds.5sd (a)B. R. Penfold and B. H. Robinson Accounts Chem. Res. 1973,6 74; D. Seyferth Ado. Organo-metallic Chem. 1976 14 97; (b) D. Seyferth C. S. Eshbach G. H. Williams and P. L. K. Hung J. Organometallic Chem. 1977,134,67,and references therein; (c)D. Seyferth G. H. Williams and C. L. Nivert Inorg. Chem. 1977 16 758; (d) S. Aime,'L. Milone E. Sappa and A. Tiripicchio J.C.S. Dalton 1977 227 and references therein; (e) T. I. Voyevodskaya I.M. Pribytokova and Yu. A. Ustynyuk J. Organometallic Chem. 1972,37 187. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Radical Species and Mechanisms.-The early development of carbonyl chemistry rapidly established that the stable complexes are almost universally diamagnetic. In recent years investigations have begun into the chemistry of paramagnetic species which are mostly rather short-lived and may be regarded as metal-centred radicals. One approach involves electrochemical methods such as cyclic voitammetry and controlled-potential electrolysis and for example both oxidized and reduced derivatives of simple carbonyls {e.g. [Cr(CO)=,I7 [Cr(CO)6]t [Fe(CO)=,]t [Mn(q 5-C5H5)(CO)3]t and [MnBr(CO)5]t} have half lives of several Metal clusters are of especial interest in that they might be expected to have more potential as 'electron reservoirs'.Frequently however irreversible disintegration of the cluster occurs. Thus complexes [Fe3(C0)12- (PR3),] (n = 0-3) undergo reversible one-electron reduction to radical anions but two-electron reduction results in irreversible breakdown of the cluster.59b The alkylidynetricobalt clusters undergo one-electron reduction to radical anions [CO~(CO)~(CY)]~ (R = alkyl or halogen) which are unusually stable having lifetimes of several This is an interesting addition to the known ability of these CO~C clusters to stabilize positive charge in carbonium ions (see previous section). Even more remarkable is the very stable radical anion (42) which results from electrochemical reduction of (41) or reduction with sodium amalgam in solution in THF.Products of the reduction are N~[CO(CO)~] and the pyrophoric green sodium salt of (42). An X-ray diffraction study of the air-stable [N(PPh3),]' salt of (42) confirms the structure shown and gives a Co-Co bond length of 236 pm consistent with a bond order of -1.5. The radical anion is readily oxidized either chemically or electrochemically to (43) which presumably has a full Co-Co double bond.59d -r 7' 0 The recognition of the significance of radical intermediates has also been an important recent development in mechanistic studies of carbonyl complexes. Definitive evidence is not always easy to obtain. For example one notes the continuing debate on the photolysis products of [Mn2(CO)lo].It now seems clear that the initial product of this reaction is the [Mn(CO)=,] It is very (a)C. J. Pickett and D. Pletcher J.C.S. Dalton 1976 636 749; (b)A. M. Bond P. A. Dawson B. M. Peake B. H. Robinson and J. Simpson Inorg. Chem. 1977,16,2199; (c)B. M. Peake B. H. Robinson J. Simpson and D. J. Watson ibid. p. 405; A. M. Bond B. M. Peake B. H. Robinson J. Simpson and D. J. Watson ibid. p. 410; (d)N. E. Schore C. S. Ilenda and R. G. Bergman J. Amer. Chew. SOC. 1977,99,1781. (a)A. S. Huffadine B. M. Peake B. H. Robinson J. Simpson and P. A. Dawson J. Organometaflic Chem. 1976,121,391 and references therein; A. Hudson M. F. Lappert and B. K. Nicholson J.C.S. Dalton 1977 551 and references therein; (b) J. L.Hughey C. P. Anderson and T. J. Meyer J. Organometallic Chem. 1977 125 C49; (c)J. P. Fawcett and A. Poe J.C.S. Dalton 1977 1302 and references therein; (d)L. S. Benner and A. L. Balch J. Organometallic Chew. 1977,134 121; (e)P. .I. Krusic P. J. Fagan and J. S. Filippo jun. J. Amer. Chem. SOC.,1977,99 252; (f) M. Absi-Halavi and T. L. Brown ibid. p. 2982; (g) B. H. Byers and T. L. Brown ibid. p. 2527; (h)J. Organometallic Chem. 1977,127 181. Organometallic Compounds short-lived recombining at almost diff usion-controlled rates,606 and in basic solvents (S) it undergoes disproportionation reactions to yield [MII(S)~] [Mn(C0)5]2.60" Somewhat more surprising are observations that facile homolytic bond fission may also be important in thermal reactions.Thus kinetic studies of the thermal decomposition of [Mn2(CO)lo] and of its reactions with CO 02,or PPh3 have all shown initial reversible homolytic fission of the Mn-Mn bond as the major reaction path.60' Substitution of the decacarbonyl by tertiary phosphine or phosphite ligands labilizes the bond even further and [MII~(CO)~L~] species react with the spin trap nitroso-t-butane to give radicals6'" derived from [*Mn(CO),L] even in the dark at 23 oC.60d Other reactions in which radical processes have been identified include (a)the reaction of alkyl halides with Na[Fe(CO),(q'-C,H,)] in which alkyl radicals have been detected by e.s.r.;60e (6) the formation of [CO(CO)~(PP~~),][S~CI~] from [Co(CO),(SnCl,)] and PPh3 via homolytic cleavage of the Co-Sn bond to give the 17-electron [Co(CO),] species;60f and (c)tertiary phosphine substitution of [Re(CO),H] by a radical chain process initiated by abstraction of H atoms.60g This last reaction offers an interesting contrast to the corresponding reaction of [Mn(CO)'H].This proceeds by an unusual hydride- migration mechanism to give an intermediate formyl complex [Mn(CHO)-(CO),(PR3)] which undergoes rapid decarbonylation to [Mn(C0)5H(PR3)].60h Structure and Bonding.-The structure of [Mn2(CO),(dpm)2] (dpm = Ph2PCH2PPh2) provided the first example of a carbonyl group bridging a metal- metal bond by 0-bonding to one metal atom and r-bonding to the other (i.e.acting as a four-electron donor).61" The [Fe,(CO),,H]- ion studied by X-ray diffraction of the [NMe3(CH2Ph)]+ salt has a similar four-electron bridging CO but in this structure (shown in Figure 4)the wbonding is to an Fes triangle.Each Fe carries 3 terminal CO groups and the hydride bridges the nuclei Fe-2 and Fe-3. Important bond distances are Fe-4-C-13 181pm; Fe-2-C-13 210 pm; Fe-3-C-13,210 pm; Figure 4 The structure of [Fe4(C0)13H]-ion,as determined by X-ray diffraction (Reproduced from J.C.S. Chem. Comm. 1976,919) (a) R. Colton C. J. Commons and B. J. Hoskins J.C.S. Chem. Comm. 1975 363; (6) M. Manassero M. Sansoni and G. Longoni ibid. 1976 919. 256 C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Fe-1-C-13 217 pm; and Fe-1-0-13 200 pm.61b We reported last year a proposal that binary carbonyl structures {e.g. [Fe3(C0)12]us.[OS~(CO)~~]} can be rational- ized by considering the size of the cavity available to the metal atoms in various possible close-packed polyhedra of CO groups.62a Subsequent calculations62b of the non-bonded interactions in such polyhedra show that for idealized arrange-ments icosahedral packing {the [Fe3(C0)12] structure} should be favoured for both Fe and 0s dodecacarbonyls. Thus the results do not support the idea that non- bonded interactions are a singly important factor and bonded interactions must also change significantly with structure. The authors note however that the observed general tendency for larger clusters to adopt more regular carbonyl arrangements does suggest that ligand-ligand interactions are an important factor.62b Some extended Hiickel calculations62c on model M6H6clusters (M =Co or Ir) with octahedral bicapped tetrahedral and trigonal-prismatic geometries are also an important development of an area reported last year.The results show that the extension of the electron-counting schemes originally developed for borane complexes to metal clusters has a reasonable basis in M.O. theory. In particular the ‘capping principle’ that a capped polyhedral cluster has the same number of bonding skeletal molecular orbitals as the parent uncapped polyhedron is shown to have general quantum-mechanical validity.62c The synergic CT-T model of bonding between CO and transition metals has been accepted for a long time and many theoretical and spectroscopic studies have been carried out.Thus an experimental measurement of the actual charge distribution in [Cr(CO),] by a combination of X-ray and neutron diffraction techniques at liquid- nitrogen temperature is of obvious The authors have concluded that net charge transfer is approximately zero both forward and back donations being -0.3 electrons per C0.63a Interestingly a new M.O. treatment of the same molecule using SCF-X,-MSW calculations has led to the conclusion that the M-C interaction is due mainly to a-bonding there being only a relatively weak 7r-back- bonding contribution. The latter is however important in determining the bond order of CO. The authors point out that the charge transfer in the two components of the bond may still be roughly equal even though their contributions to the bond strength are different.63b 21 Fluxional Processes A recent provides much valuable background to this section.The majority of the following results are from 13C n.m.r. studies. Carbony1s.-Fluxional rearrangement in [Fe(CO)5] is very fast even at 100K and a polarized visible photolysis experiment shows that the extreme conditions of 20 K in an argon matrix are necessary if one is to observe the stereochemically rigid 62 (a)B. F. G. Johnson J.C.S. Chem. Comm. 1976 211; (b)P. B. Hitchcock R. Mason and M. Textor ibid. p. 1047; (c)D. M. P. Mingos and M. 1. Forsyth J.C.S. Dalton 1977 610. (a) B. Rees and A. Mitschler J. Amer. Chem. Soc. 1976 98 7918; (b) J. B. Johnson and W. G. Klemperer ibid. 1977,99 7132. 64 ‘Dynamic Nuclear Magnetic Resonance Spectroscopy’ ed.L. M. Jackman and F A. Cotton Academic Press New York 1975. Organometallic Compounds 257 m01ecule.~~" In contrast non-dissociative fluxional rearrangements of 6-co-ordinate species are expected to have much higher activation energies and were previously known only in tris-chelate complexes and [MH2L4] (M=Fe or Ru L = PR3). A series of recent papers656 has established the first examples of such rearrangements in simple 6-co-ordinate carbonyls namely [Fe(C0)4H2] and the complexes [M(C0)4(EMe3)2] (M = Fe Ru or 0s; E = Si Ge Sn,or Pb). In complexes involving more than one metal atom pairwise exchange of CO between two metal centres is well documented [see Annuul Reports (A),1975 Vol. 721 and a process equivalent to rotation of M(CO) units has also been established.For example [os6(co),8] has a bicapped tetrahedral arrangement of Os(CO) groups with all carbonyls terminal and although localized scrambling begins within two individual Os(CO) groups at -50 "C and is complete for all six groups at +lOO"C there is no inter-osmium scrambling even at +100"C.66 Definitive evidence on mechanisms of CO scrambling is often difficult to obtain and one interesting approach involves restricting the possible movements of the CO groups by attaching (CO),M-M'(CO) moieties to polyene or polyenyl groups. Inter-nuclear exchange is usually inhibited in these molecules but has been obser- ved for the first time in the azulene complexes (44).67a It appears that internuclear scrambling occurs only when rn # n and when the .rr-electron density of the polyene is capable of redistribution to allow transfer of carbonyls between metal atoms.67b Thus the intermediate from (45) is (46).67c (44)M =Fe or Ru (45) (46) A different type of process involving the cycling of coplanar carbonyl groups around a set of metal atoms has been reported by several groups this year.68 For example (47) where N-N = pyridazine is stereochemically rigid at -156 "C,but the equivalencing of the 6 equatorial CO groups is complete at -90°C.The CO groups are considered to cycle around the triangle via non-bridged intermediates but always preserving their initial cyclic Such a process would be (a)J. K. Burdett J. M. Grzybowski M. Poliakoff and J. J. Turner J. Amer. Chem.SOC. 1976 98 5728; (b)L. Vancea and W. A. G. Graham J. Organometallic Chem. 1977,134,219; R. K. Pomeroy L. Vancea H. P. Calhoun and W. A. G. Graham Znorg. Chem. 1977.16 1508 and references therein. C. R. Eady W. G. Jackson B. F. G. Johnson J. Lewis and T. W. Matheson J.C.S. Chem. Comm. 1975 958. (a)F. A. Cotton B. E. Hanson J. R. Kolb and P. Lahuerta Znorg. Chem. 1977,16,89 and references therein; (6) F. A. Cotton and B. E. Hanson ibid. p. 1861 and references therein; (c)F. A. Cotton B. E. Hanson J. R. Kolb P. Lahuerta G. G. Stanley B. R. Stults and A. J. White J. Amer. Chem. SOC. 1977,99,3673. (a)F. A. Cotton B. E. Hanson and J. D. Jamerson J. Amer. Chem. SOC. 1977 99 6588; (6)F. A. Cotton B. E. Hanson J. D. Jamerson and B. R. Stults ibid.,p. 3293; (c)B.F. G. Johnson J. Lewis B. E. Reichert and K. T. Schorpp J.C.S. Dalton 1976 1403; (d)G. L. Geoffroy and W. L. Gladfelter J. Amer. Chem. SOC. 1977 99 6775; (e) R. J. Lawson and J. R. Shapley ibid. 1976 98 7433; (f)M. Tachikawa S. I. Richter and J. R. Shapley J. Organometallic Chem. 1977 128 C9. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon unobservable in the parent carbonyls [M3(C0)12] (M=Fe Ru or Os) for which there has been extensive discussion of the mechanism of carbonyl scrambling. A ‘truly concerted’ process involving the reorientation of the Fes triangle within an icosahedron of CO groups has been for [Fe3(C0)12 1 but Cotton et a1.68n have pointed out that this mechanism cannot account for the results they obtained for (47).Other examples of similar processes include [Fe2(C0)7(pyridazine)]68b and [OS~(CO)~~(PE~~)~]~~~ in which the cycling process in each case occurs around two metal atoms and [F~RU~(CO)~~H~],~~~ [Rh3(C0),(q 5-C5H5)3],68e and [Os3(CO)lo(norbornadiene)],68f where metal triangles are involved. Possibly the most unusual fluxional process of all occurs in the [Ptg(C0)18]2- anion.7oa The remarkable series of anions [Pt3(CO)&- (n = -10,6,5,4,3,2 or 1) has been known for several years although complete details of the method of their synthesis by reductive carbonylation of [PtC16I2- have only recently been pub- li~hed.~” Some corresponding nickel complexes (n = 2 or 3) are also The structures of these anions are based on the stacking of triangular [M3(p- CO),(CO),] units which are approximately prismatic when M = Pt and antipris- matic when M =Ni (n = 2).Evidently the two structures are very close in energy and lg5Pt n.m.r. have now shown that in [Pt,(CO)18]2- there is rapid rotation of the outer Pt triangles with respect to the middle Pt triangle both at 25 “C and at -85 “C. Pt-Pt coupling is retained and so the process is intramolecu- lar. If [Pt12(C0)24]2- is also present there is exchange of [Pt3(C0)3(p-C0)3] units between the n = 3 and n =4 species at 25 “C. Organometallic Complexes.-The 1,2-shift has been generally accepted as the principal mechanism of fluxional processes in polyene-metal complexes.64 Although [Fe(q4-C8H8)(C0)3] was one of the first systems to be studied it has a very low activation energy (the ‘ring-whizzing’ persisting even in the solid state7Ia) and the validity of the 1,2-shift mechanism for this molecule was established only last year.71b A study of [Fe(allene)(q5-C5H5)]+ cations has confirmed that the mechanism is a 1,2-shift in these molecules also.The process is ‘concerted’ and not stepwise and must go via an (77 ‘-2-ally1)iron intermediate since chiral complexes are not racemized during the fluxional process.71c The 1,3-shift has been generally assumed to be a process that requires higher energy and is essentially 69 B. F. G. Johnson J.C.S. Chem. Comm. 1976 703. ’O (a) C. Brown B. T. Heaton P. Chini A. Fumagalli and G. Longoni J.C.S.Chem. Comm. 1977 309; (6) G. Longoni and P. Chini J. Amer. Chem. SOC.,1976 98 7225; (c) G.Longoni P. Chini and A. Cavalieri fnorg. Chem. 1976.15 3025; G. Longoni and P. Chini ibid.,p. 3029. ’‘ (a)A.J. Campbell C. E. Cottrell C. A. Fyfe and K. R. Jeffrey Inorg. Chem. 1976,15 1321 1326; (b) F. A. Cotton and D. L. Hunter J. Amer. Chem. SOC., 1976 98 1413; (c) B. Foxman D. Marten A. Rosan S. Raghu and M. Rosenblum ibid. 1977.99 2160. Organometallic Compounds forbidden. For example (48; R=H) is stereochemically rigid up to 373 K. However it now appears that comparatively subtle changes are sufficient to induce 1,3-shifts since the complexes (48; R = SiMe3) and (48; R = GeMe3) are fluxional FdCO) (48) the oscillatory motion of the Fe(C0)3 groups between the two possible q4-positions on the triene having the effect of a 1,3-shift.Coalescence temperatures are 388 K (R = SiMe3) and 373 K (R = GeMe3).72a Because 1,3-shifts were thought unlikely the fluxional processes in [M(q 6-CsHs)(CO)3] were originally considered to occur via a symmetrical q8-intermediate (i.e. 'random shifts'). 1,3-Shifts were favoured by other authors and the application of the ForsCn-Hoffman spin-saturation method to 13C n.m.r. spectra of [Cr(q6-CsHs)(CO)3] has now proved that the dominant rearrangement process is a 1,3-shift. 1,2-Shifts also occur as a competing process.72b Presumably the 1,3-shift proceeds via a 16-electron intermediate [M(q4-C8HS)(CO),] and a mechanism which is essentially the inverse of this has been demonstrated for a variety of [Pd(C7H9)L2]+cations (49).72c The fluxional process in the 16-electron complex (49) proceeds uia the 18-electron intermediates (50) to (51) i.e.overall a 1,3-shift or a q3-q5-q3interchange in contrast to the more usual q 3-q '-q3 mechanism in ally1 complexes. This process is very closely related to the 'twitching' process for complex (52) that was originally proposed by (52) Cotton et al. for the fluxion of Fe2(C0)6 groups bound to cy~lo-octatrienes~~" and extended and confirmed by other authors this year working on q6-(bicyclo-[6,1,0]nona-2,4,6-triene)hexacarbonyldi-iron72d and q6-(bicyclo[6,2,0]deca-2,4,6-triene)hexacarbonyldir~thenium.~~' Metal Hydrides.-The anions [Rh3(C0)24H5-n]n- (n = 2 or 3) are the only known clusters in which the arrangement of metal atoms approximates that found in a close-packed metal and in which a central metal atom is completely encapsulated 72 (a)L.K. K. Li Shing Man and J. Takats J. Organometallic Chem. 1976 117 C104; (b)B. E. Mann J.C.S. Chem. Comm. 1977 626 and references therein; (c)B. E. Mann and P. M. Maitlis ibid. 1976 1058; (d)G. Deganello L. K. K. Li Shing Man and J. Takats J. Orgunometullic Chem. 1977,132,265 and references therein; (e) G. Deganelio J. Lewis D. G. Parker and P. L. Sandrini Znorg. Chim. Acta 1977 24 165. C. J. Cardin D. J. Cardin R. J. Norton and K.R. Dixon D= CO-bridge Figure 5 A schematic representation of [Rh13(C0)24H5-n]n-, in which A represents the central rhodium atom and B and C are rhodium atoms to which carbon yl groups are attached (Reproduced from J.C.S.Chem. Comm. 1977 39) (see Figure 5). Thus the observation of rapid hydrogen migration around the inside of the cluster is especially interesting for its relevance to catalysis by clusters and to diffusion of hydrogen in metals.73" At ambient temperatures 'H n.m.r. spectra show a doublet of septets when either the Rh(B) or Rh(C) atoms are specifically spin-decoupled. The magnitude of Rh(AtH coupling (-23 Hz) suggests a one- bond coupling and this considered in conjunction with the high chemical shifts [T 36.7 (n = 2) and 39.3 (n = 3)] indicates that the hydrogen atoms are inside the cluster. The hydrogen migration is accompanied by CO-scrambling Another hydride-migration process of interest is the edge-terminal-edge exchange observed in [Ru~(CO)~~H~(P~~PCH~CH~PP~~)~.~~~ The hydrides in this type of tetrahedral cluster are edge-bridging {e.g.[Ru4(C0),2H4]} or face-bridging {e.g [Re4(C0),2H4]} and terminal hydrides are generally rare in cluster chemistry. Thus the migration process was expected to be based on the interconversidn of edge- and face-bridging 22 Isocyanide Complexes Although isocyanide ligands may be regarded as formally similar to carbonyls fairly extensive studies have not hitherto yielded the enormous variety of complexes found in carbonyl chemistry. This situation may be about to change. Generally the isocyanides tend to stabilize higher oxidation states than does CO and many of the known complexes are cationic. Thus isocyanide analogues of simple zerovalent binary carbonyls were previously known only in the Cr and Ni The complex [CO,(CNR)~] (R = 2,6-xylyl) has now been prepared by 73 (a)S.Martinengo B. T. Heaton R. J. Goodfellow and P. Chini J.C.S. Chem. Comm. 1977,39; (b)J. R. Shapley S. I. Richter M. R. Churchill and R. A. Lashewycz J. Amer. Chem. Soc. 1977,99,7384. 74 (a)G. K. Barker A. M. R. Galas M. Green J. A. K. Howard F. G. A. Stone T. W. Turney A. J. Welch and P. Woodward J.C.S. Chem. Comm. 1977 256 and references therein; (6) Y. Yamamoto and H. Yamazaki J. Organometallic Chem. 1977 137 C31. Organometallic Compounds 26 1 the reaction of [co2(c0)8] with the isocyanide at 80-90°C746 and the t-butyl analogue by reduction of [Co(CNBu'),][PF6] with potassium amalgam at -78 0C.74a Structures (53) analogous to the carbonyl-bridged [(CO)3Co(p-CO)2Co(CO)3] NR II C /\ (RNC)~CO-\/Co(CNR)3 C I1 NR form of [co2(co)8] have been proposed by both sets of author~'~ and confirmed by the X-ray analysis of [CO~(CNBU')~] There are also indications from variable- .74a temperature n.m.r.studies that other isomers exist in solution just as they do in the case of [co2(co)8]. Complexes [M(CNBU')~] (M=Fe or Ru) result from the reduction of M" complexes in the presence of excess isocyanide and a derivative [Ru(CNBU')~(PP~~)] The structure has been characterized by X-ray diffra~tion.~~" is notable in that two of the isocyanide ligands are bent (130") at the nitrogen. The only previous example of such bending in a terminal isocyanide i.e. trans- [Mo(CNM~)~(P~~PCH~CH~PP~~)~], was much less marked (156°).74" Preliminary accounts of the remarkable clusters obtained by the reaction of [M(cod),] (M =Ni or Pt; cod =cyclo-octa-1,5-diene) with t-butyl isocyanide had appeared previously and full papers this year.75 The molecule of [Pt3(CNBu')6] is a symmetrical triangle of Pt atoms with terminal isocyanides at each platinum and bridging each edge.75" [Ni4(CNBu')7] is based on an Ni4 tetrahedron with one terminal isocyanide at each nickel atom.The other three isocyanides act as four- electron-donor unsymmetrical bridges on each of the basal edges their attachment being somewhat similar to that of the bridging CO in [Mn2(CO)5(dpm)2] (see p. 255).6*" The existence of a palladium analogue of [Pt3(CNRI6] has not been conclusively demonstrated although it has been claimed that the product of reac- tion of [Pd(q3-C3H,)($-C,H5)] with Bu'NC is a t~imer.~,~ However an unprece- dented linear trinuclear complex has been obtained7" from the reaction of 2 mol of [Pd(CNMe),] with 1 mol of [Pd(CNMe)J2+.The product ion [Pd3(CNMe)8]2+ can be isolated as a crystalline [PF6-] salt but a derivative [Pd3(CNMe)6(PPh3)2] [PF6], was chosen for X-ray study because of its greater stability. The cation shown in Figure 6 has a linear Pd chain with very short bonds (259.2 ~m),~" The preparation of [Pd3(CNMe)8]2' is especially interesting in that the same reaction but with 1 :1 stoicheiometry had previously given good yields of dinuclear species [Pd2(CNMe)6]2+ and similar reactions yielded [PdPt(CNMe)6]2+ and [Pt2(CNMe)6]2+.These molecules also have metal-metal bonds with no bridging is~cyanides.~~~ 75 (a)M. Green J. A. K. Howard M. Murray J. L. Spencer and F. G. A. Stone J.C.S. Dalton 1977 1509; J. ForniCs M. Green A. Laguna M. Murray J. L. Spencer and F. G. A. Stone ibid.,p. 1515; (b) M. G. Thomas W. R. Pretzer B. F. Beier F. J. Hirsekorn and E. L. Muetterties J. Amer. Chem. SOC. 1977,99,743;(c)A. L. Balch J. R. Boehm H. Hope and M. M. Olmstead ibid. 1976,98,7431; (d)J. R. Boehm and A. L. Balch Znorg. Chem. 1977 16 778 and references therein. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Figure 6 An ORTEP drawing of the structure of [Pd3(CNMe)6(PPh3)]2+ (Reproduced by permission from J Amer. Chem.Soc. 1976,98 7431) 23 Catalytic and Related Reactions of Complexes containing Metal-Metal Bonds Interest in this area centres around the thesis that complexes containing metal- metal bonds especially clusters may serve as models for catalysis at metal surfaces. This relationship has been Although clusters certainly permit a great variety of interactions with organic species [e.g. see Annual Reports (A),1976 Vol. 731 well-studied examples are still relatively rare. One possible difficulty is that the co-ordinative unsaturation usually required for catalysis is rarely present in metal- metal-bonded complexes. [Ni2(cod)2(~-C2Ph2)] (cod = cyclo-octa-l,5-diene) which is the product of reaction of diphenylacetylene with [Ni(cod),] has the acetylene forming a four-electron bridge between two Ni(cod) groups.766 Thus a double M=M bond would be required to achieve electron counts of 18 at each Ni atom.The Ni-Ni bond length is 262 pm consistent with the presence of a single bond and the complex may therefore be considered to be co-ordinatively unsaturated although in a somewhat different sense to [0s2(CO),,H2] which has an Os=Os double bond. The Ni complex is a selective catalyst for hydrogenation of acetylenes to cis-olefins whereas co-ordinately saturated analogues such as [Co,(CO),(C,R,)] and [Ni2(q’-C5H,)2(C2R2)] are inactive.766 Reduction of [TiC12(q s-CsH5)2] with potassio-naphthalene at -80 “C gives (54).77aThis complex is interesting as an example of the rather rare bridging (54) 76 (a)H. F. Schaefer,Accounts Chem.Res. 1977,10,287;(b)V.W. Day S. S. Abdel-Meguid S. Dakestani M. G. Thomas W. R. Pretzer and E. L. Muetterties J. Amer. Chem. SOC., 1976 98 8289. ” (a) G. P. Pez J.C.S.Chem. Comm. 1977 560 and references therein; (b) N. E. Schore C. Ilenda and R. G. Bergrnan J. Amer. Chem. SOC.,1976,98 7436; (c) J. P. Collman R. K. Rothrock R. G. Finke and F. Rose-Munch ibid.,1977 99 7380. Organometallic Compounds 263 (ql,qS-CsH4) group and is able to bind hydrogen reversibly and to catalyse hydrogenation of olefins. It is also a catalyst for a novel conversion of ethylene into ethane and buta-1,3-diene and the proposed mechanism (see Scheme 19) +\ P Ti-Ti Ti-Ti double &hydrogen abstraction H H\ / Ti -Ti Ti -Ti C2H4 ++ C2H6 L-2 Scheme 19 emphasizes the unique importance of the presence of two metal (See also Section 10.) Two stoicheiometric reactions reported this year also emphasize this type of co-operativity (a) thermal decomposition of the metal-metal bonded dimer [CO(CH~)(~~-C~H~)(~-CO)]~ is remarkable in that it yields acetone in 85% yield and if CO is added acetone and [CO(~~-C~H~)(CO)~] are formed in 100% yield;776 and (6) the very reactive dianion [(CO)3Fe(p-PPh2)2Fe(C0)3]2-undergoes alkylation to yield an acyl derivative [(CO)2(COR)Fe(p-PPh2)2Fe(CO)3]-,rather than the expected iron alkyl.The alkyl-acyl rearrangement normally requires an external nucleophile and the authors have suggested that in the present case the formation of an Fe-Fe bond formally absent in the dianion fulfils this The catalysis of the water-gas shift reaction by [RU~(CO)~~] in alkaline solu- ti~n,~~~ of Fischer-Tropsch synthesis by [Ir4(CO)12] in NaCl.2AlCl3 and of hydroformylation by [Co3(p3-CC6H5)(C0)9] or [CO,(CO)~(~~-CO)~(~~-PCsHg)2]78c have all been reported this year.Another important reaction which is not generally accomplished by mononuclear catalysts is the reduction of triple bonds. [Ni4(CNBu‘)7] (see p. 261)756 is an active catalyst for a variety of processes including the conversion of acetylenes into benzenes and of butadiene into cyclo- octa- 1,5-diene the polymerization of allene and the selective hydrogenation of acetylenes to cis-olefin~.~’~ Recently it has also been shown to hydrogenate isocyanides and nit rile^.'^^ However the most interesting insight into these reduc- tions of triple bonds comes from a very elegant study of a stoicheiometric (not catalytic) reaction.79 The complete sequence shown in Scheme 20 has been established by the isolation and characterization of all the intermediates.It represents a unique stepwise reduction of CH3CN on the face of the Fe3cluster starting at either (55) or (56) and culminating in (57).79 ’13 (a)R. M. Laine R. G. Rinker and P. C. Ford J. Amer. Chem. Soc. 1977,99,252;(b)G. C. Demitras and E. L. Muetterties ibid. p. 2796; (c) R. C. Ryan C. U. Pittman jun. and J. P. O’Connor ibid. p. 1986; (d) E. Band W. R. Pretzer M. G. Thomas and E. L. Muetterties ibid. p. 7380. 79 M. A. Andrews and H. D. Kaesz J.Amer. Chem. SOC.,1977,99,6760. C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon 0 II \ /c*. / -Fe 'Fe-i, i/ \ -Ha/1 \ Me Me H 1 \/ C C=N Et I I (57) Reagents i MeCGN A; ii H+; iii 02; iv H2 80 "C; v hexane 65 "C;vi Hz 200 psi 25 "C in hexane; vii EtN02-THF A Scheme 20 24 Theoretical Studies of Organometallic Complexes The Chatt-Dewar-Duncanson approach to bonding in polyene-metal systems has played a central role in the development of organometallic chemistry. The original scheme is most suitable for relatively weak interactions and a topological Huckel model which is more appropriate for strongly interacting systems has now been proposed.'' The basis-set orbitals are chosen as the p,-orbitals of the olefin and two metal orbitals with T symmetry with respect to the M-olefin bond e.g.as shown in Figure 7. This assumption is roughly equivalent to suggesting that the M+L back-bonding component dominates in strongly bound complexes. The model successfully accounts for bond lengths in a wide range of q4-diene and polyene systems including those with phenyl rings fused to the co-ordinated diene. Using simple symmetry arguments this was previously possible only for butadiene itself.''" Development of the approach shows that q4-diene complexes may be represented as even-alternant bonding networks (a) and q3-allyl complexes as odd-alternant networks (b) as shown in Figure 8. Application of the pairing theorem to these networks shows that (a) has three bonding M.O.'s and requires 2 (a)D.M. P. Mingos J.C.S.Dalton 1977 20; (b) ibid. p. 26; (c) ibid. p. 31. Organometallic Compounds 0 Figure 7 The basis-set orbitals used in a topological Hiickel model for the bonding in a strongly interacting system of butadiene and a metal atom to which carbonyl groups are attached (Reproduced from J.C.S. Dalton 1977 20) 1 * c’ CZgel -P c3 4 Figure 8 Topological representations of bonding networks for (a) even-alternant and (b) odd- alternant systems. The valence orbitals are represented by vertices (closed circles for ligand orbitals and open ones for metal orbitals) and the connections represent orbital interactions (Reproduced from J.C.S. Dalton 1977 26) electrons from the metal for a stable closed-shell configuration whereas (b) has two bonding and one non-bonding M.O.and requires 3 electrons from the metal. These requirements are met by Fe(C0)3 and CO(CO)~ respectively* and hence we have an indication of the electronic basis of the 18-electron rule. The non-bonding orbital in the allyl case suggests the possibility of stable 16-electron complexes and [Fe(v3-allyl)(CO)3]’ cations are known examples.80” Alternatively a more sophis- ticated analysis of the 18-electron rule is possible in terms of a perturbation M.O. method within the topological Hiickel approach. Thus by considering the hypo- thetical union of an allyl complex with a methyl radical it is possible to predict whether ~2 or v4-co-ordination of the resulting butadiene is preferred.Union with an allyl radical permits distinctions between v4-,q5-,and v6-modes to be made for the resulting hexatriene and the approach may be generalized to a statement of the 18-electron rule.8oc Another interesting extension of basic theory is the development of a frontier- orbital approach to metal carbonyl and organometallic fragments. Potentially this * Six metal electrons (dr2 dx*-y2 and dxy)of an M(C0)3 fragment are essentially non-bonding with respect to the metal-olefin n-system. 266 C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon method can explain many features of structure and reactivity. For example a recent paper'la gives good references to basic theory and uses extended Huckel calculations to describe frontier orbitals for M(CO)3 and M(CH2) fragments (n = 3-8).The results show that M(q6-C6H6) or M(qS-CsHs) should be more effective in 0-bonding to another metal whereas M(CO)3 is better at 7r-bonding. Comparison of bond lengths in known structures tends to confirm these conclusions (e.g. the metal-metal bond lengths are 252 and 237 pm respectively in [co2(co)8] and [Ni2(q'-CsHs)2(CO)2]).81a The construction of frontier orbitals for M(q5-CsHs) and M(CO)3 fragments also leads to a M.O. treatment of the triple-decker sandwich compounds [(qs-C5H5)M(q5-CsH5)M(q5-C5H5)] and [(C0)3M(qS-CSH5)M(C0)3].816Two series of stable compounds are predicted that have 34 and 30 valence electrons respectively. Examples of such complexes are rare but those which are known seem to confirm this prediction.Thus [Ni2(q5-CsH5)3]+ has 34 valence electrons and [1,7,2,3-Co2(C2B2H3)(qs-C,H,),] and its 1,7,2,4-isomer both have 30 valence electrons.'lc The newly discovered complexes of 3,4-di- ethyl-2,5-dimethyl-l,2,5-thiadiborolen(Y) also have 30 valence electrons if the ligand is regarded as an aromatic dianion analogous to thiophen; i.e. [(CO)3Mn(Y)Mn(CO)3],'1c [(q'-CSHS)Fe(Y)Fe(q 5-C5HS)J,81d and [Co2(Y),] .'Ic A study by the frontier-orbital method of bent M(qS-CSH5)2 fragments has been reported.'" The results enable rationalization of both structural and reactivity data and the problems considered include (a)the geometry of [M(q5-CsHS)2L,] complexes such as [Re(q5-CsHs)2H] [V(qs-CsH5)2C12] and [Nb(q 5-C5H5)2H3]; (b) insertion reactions of olefins into Zr-H or Ti-R bonds; (c) insertion of CO into Zr-R bonds; and (d) oxidative coupling of diphenylacetylene by [Ti(q'-C5H,)2(CO)2] to yield the metallocycle [!ffCPh)].'lf 25 Synthesis and Structure of Organometallic tr -Complexes Because of the emphasis on complexes of large rings in last year's Report this section deals with q2-q5 complexes taking as its main themes the oligomerization of alkynes novel unsaturated ligands and fulvalene complexes.Several shorter items of general interest are also included. Bonding of alkynes to groups of metal atoms is becoming increasingly common and several new examples have been reported this year. [Rh3(q 5-CsH5)3(CO)- (C2Ph2)] and its C6F5 analogue have Rh triangles with cyclopentadienyls on each Rh and the alkyne essentially 0-bonded to two Rh atoms (causing an alkene-like geometry) and 7r-bonded to the third.Interestingly the CO is opposite the alkyne and it co-ordinates to all three rhodiums in the C2Ph2 complex whereas in the C2(C6F5)2 complex the CO bridges a Rh-Rh edge on the same side of the metal triangle as the alkyne.82" A somewhat different type of bonding occurs in [RU~(CO)~~(C~P~~)], where the four Ru(CO)~ groups and the two carbon atoms of (a)M. Elian M. M. L. Chen D. M. P. Mingos and R. Hoffman Inorg. Chem. 1976,15,1148 (b)J.W. Lauher M. Elian R. H. Summerville and R. Hoffman J. Amer. Chem. SOC.,1976 98 3219; (c) H. Werner Angew. Chem. Internat. Edn. 1977 16 1 and references therein; (d) W. Siebert T. Renk K. Kinberger M.Bochmann and C. Kruger ibid. 1976,15,779; (e)W. Siebert and W. Rothermal ibid. 1977,16,333; (f) J. W. Lauher and R.Hoffman J. Amer. Chem. SOC.,1976,98 1729. 82 (a)T. Toan R. W. Broach S. A. Gardner M. D. Rausch and L. F. Dahl fnorg. Chem. 1977,16,279; (b) B. F. G. Johnson J. Lewis B. E. Reichert K. T. Schorpp and G. M. Sheldrick J. C.S.Dalton 1977 1417. Organometallic Compounds the alkyne form a distorted closo octahedron. The arrangement is best described as a 'butterfly' Ru4 group capped by the alkyne.82b The oligomerization of alkynes by palladium(I1) complexes has attracted much recent effort and the main features of the reaction are now ~nderstood.~~" If the substituents on the alkyne are relatively small the principal reaction is cyclo- trimerization to substituted benzenes but bulky substituents may lead to substi- tuted cyclobutadiene or even mono-alkyne complexes.The mechanism is a sequence of insertions (similar to those shown in Scheme 21) to give a u-hexatrienyl Ph OMe I 'C I CII Ph 'Pd(0Ach' 4lll+Pd- % /\ Ph Ph li Ph OMe ring c-e- closure Pd \ Ph Ph Ph Reagents i PhC=CPh; ii MeOH Scheme 21 derivative which then undergoes internal insertion to a substituted palladia- methylcyclopentadiene complex e.g. (58). The final step in the cyclotrimerization is the formation of a substituted benzene from the substituted palladiamethyl- cyclopentadiene complex and the mechanism of this reaction is not yet clear.83a Scheme 21 represents an interesting variation of this process leading to the isolation of (59) from the reaction of palladium acetate with diphenylacetylene in methanol.The overall reaction involves cleavage of a C=C triple bond.836 In view of the novelty of this process it is interesting that thejeverse reaction has also been reported this year. [See Section 12 and equation (30).] 83 (a) P. M. Maitlis Accounrs Chem. Res. 1976 9 93; (b)T. R. Jack C. J. May and J. Powell J. Amer. Chem. SOC.,1977,99,4707;(c) H. Suzuki K. Itoh Y. Ishii K. Simon and J. A. Ibers J. Amer. Chem. SOC.,1976 98 8494; (d) P. Caddy M. Green E. O'Brien L. E. Smart and P. Woodward Angew. Chem. Internat. Edn. 1977,16,648. 268 C.J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon Cyclotrimerization of alkynes using Pdo complexes apparently proceeds by a different mechanism involving metallocyclic complexes (60) as intermediates.If an alkene is also present then co-cyclization to cyclohexa- 1,3-diene derivatives can occur in competition with benzene formation. Specific catalysis of cyclohexadiene formation requires that the alkene should bond much more strongly to (60) than does the alkyne suggesting that an electron-withdrawing alkyne in combination with an electron-rich alkene is required. In agreement with this suggestion (60; 2= C0,Me) is an efficient catalyst for the stereoselective cyclotrimerization of norbornene with dimethyl acetylenedicarboxylate which gives (61) in 94% yield.*% (60) (61) In contrast results on co-cyclization occurring in the reactions of alkynes with the indenyl complexes [Rh(alkene)2(q5-C9H7)] suggest that a metallocyclopent-2-ene complex is the intermediate rather than a metallocyclopentadiene as found in the Pdo case.The metallocyclopentene is formed by the union of alkene and alkyne on the metal and the insertion of a further molecule of alkyne followed by reductive cyclization gives complexes such as (62). This product was derived from [Rh(q2- CH,CHCN),(q'-C~H,)] and BU'C~H.'~~ Mechanisms of oligomerization of alkynes with some metals may .involve u-acetylides. In this context it is interesting that the u-T acetylide [(CO)3Fe(p- C=CBU')F~(CO)~]reacts with activated alkynes to form C-C bonds at the a-carbon of the acetylide. Typical products are (63).84 BU' (62) (63) R' =But; R2=R3= C02Et,CF3 Et or Ph An unusual q2 3-electron ligand results from the reaction of the car-byne complex [W(CC6H4CH3)(q'-C5H5)(C0)2]with PMe3.The product [W{C(0)CC6H4Me}(q'-C5H5)(CO)(PMe3)], contains a p-tolylketenyl ligand bound to tungsten approximately symmetrically uia the C-C bond (132pm) both the oxygen and the p-tolyl group being bent away from the 84 W. F. Smith N. J. Taylor and A. J. Carty J.C.S. Chem. Comm. 1976 896. 85 F.R.Kreissl P. Friedrich and G. Huttner Angew. Chem. Infernat. Edn. 1977,16 102. Organometallic Compounds Cyclopropenium ligands are known to form q3-complexes and also to undergo ring-opening reactions to form metallocycles. Reports from three different groups this year show that an entirely novel ligand can result from a combination of these processes in which a ring-opened cyclopropenium group acts as a 5-electron donor bridging a metal-metal bond.86 Dibenzylideneacetone (dba) complexes of Pdo and Pto react with triarylcyclopropenium salts followed by thallium acetyl- acetonate according to Scheme 22; complex (64; M,M' =Pd R' = p-C6&0Me R' R2 cf (64) a; R' =p-MeOC6H4 R2 = Ph b; R' = R2=p-MeOC6H4 Scheme 22 R2= Ph) has been characterized by X-ray diffraction.The q3-C3RlzR2 ligands bridge normal Pd-Pd bonds (266.3 pm) so that the C3 plane intersects the Pd-Pd axis. The apical carbons are slightly closer to the central Pd and the basal carbons closer to the terminal Pd (M'-CB = 231 M-CB = 255 M'-C = 213 M-C = 201 pm). The C,-CB distances are normal (142 pm) but there is essentially no bond between C and C (C,-C = 218 pm LC,C,C = 100o).86a A closely related complex (17) results from the reaction of NiBrz with a 1 1 mixture of (65) Ph PhG*, -Ph Ph Ph (65) with its dilithium salt.The authors have described the bridging ligand as being bound uia two u-bonds to the Ni(q5-C5Ph5) group and as a q3-allyl group to the Ni(q4-C4Ph4) group.22b However the overall geometry seems quite similar to that in (64) except that the basal carbons are more symmetrically placed relative to the metals. Thus in (17) Ni-Ni = 246 (C4Ph4)Ni-CB = 21 1 (C5Ph5)Ni-CB = 244 (C4Ph4)Ni-C,, = 195 (C5Phs)Ni-CC,, =202 pm.22* Tetrachlorocyclopropene reacts with nickel tetracarbonyl to yield [Ni4C12(C3C13)2(CO)4] whose structure is shown in Figure 9.866 The planar C3C13 ligands bridge Ni-Ni bonds and are generally similar to the previous examples (C-1-C-2 = 139.6 C-1-C-3 = 213.2 pm (a)A.Keasey P. M. Bailey and P. M. Maitlis J.C.S. Chem. Comm. 1977 178; (b)R.G.Posey G. P. Khare and P. D. Frisch J. Amer. Chem. Soc. 1977,99,4863. C. J. Cardin D. J. Cardin,R. J. Norton and K. R. Dixon 3 $"colc Figure 9 The structure of [Ni4C12(C3C13)2(C0)4] (Reproduced by permission from J. Amer. Chem. SOC.,1977,99,4863) LC-1-C-2-C-3 = 99.7'). They are however more symmetrically placed being orthogonal to the Ni4CI2 plane at the midpoints of the Ni-Ni bonds. Thus the Ni2(p2-C3C13) fragment constitutes a completely delocalized bonding arrange- ment.86b The stable cation derived by removal of P-hydride from @-(propane- 1,3-diyl)- bis(q 5-cyclopentadienyldicarbonyliron)has been postulated to be fluxional.The change of configuration occurs in two stages; these are very fast processes involving the two possible forms of the y-.rr-bridged structure [(q'-C5H5)-(CObFe(p -77 :q 2-CH2CH=CH2)Fe(C0)2($-C5H5)]+ and a slower process involving rotation about a bond.87" However an X-ray study of the [PF,]-salt shows structure (66) which is a stabilized 'non-classically bonded carbonium ion and the authors have suggested that this species is responsible for the low- temperature n.m.r. spectrum. There are fairly normal single bonds to the CH2 groups (Fe-CH2 = 212 pm) and longer-range interactions with the CH (Fe-CH = 265 pm average).The structure is interesting since weak 'stabilizing' interactions of this type have frequently been postulated to exist during reactions.87b '' (a)R. C. Kerber W. P. Giering T. Bauch P. Waterman and E.-H. Chou J. Organometullic Chem. 1976,120 C31; (b)M. Laing J. R. Moss,and J. Johnson J.C.S. Chem. Comm. 1977,656. Organometallic Compounds 27 1 The MzM triple-bonded complex [Mo~($-C,H,)~(CO),] is known to add alk- ynes to form the well-known four-electron bridging mode of alkyne bonding across a metal-metal bond e.g. [MO~(~~-C,H,)~(CO),(~-E~C~E~)]. The unsaturated complex also reacts with allene to provide the first example of an allene acting as a four-electron donor across a metal-metal bond. The Mo-Mo distance in [Mo2(q5- C,H,),(CO),(p-C,H,)] is 318pm slightly longer than in the above Et2C2 complex (298 pm) so as to accommodate the C3 chain.The allene is bent (146") so that the two virtually orthogonal ethylenic portions can each interact with one Mo atom." The stabilization of the trimethylenemethane radical in the complex tri-methylenemethaneiron tricarbonyl is unique and the bonding is of considerable interest. An ESCA study gives C 1s binding energies that are consistent with the presence of a high positive charge on the central carbon atom in the tri-methylenemethane group. Thus a crude representation of the bonding is (67) in agreement with a previous theoretical study which indicated that there is a very weak bond to the central carbon structure. The diagram (67) does not attempt to represent back-bonding.This will of course be important for the CO groups but is considered to be less significant for the trimeth~lenemethane.'~" An unusual molecule (68) which may be regarded as a hetero-analogue of tri-methylenemethaneiron tricarbonyl has been prepared by the reaction of [Mn(C=CPh)($-C5H5)(CO)2] with [Fe2(C0)9] and its structure established by X-ray diff ra~tion.~~' (67) (68) Compounds such as (69; L = PR3) in which an organo ligand is symmetrically v-bonding on a Pd-Pd bond have been known for several years [e.g. see ref. 90a and Annual Reports (A),1975 Vol. 721. The organo ligand may be allyl cyclo- pentadienyl or benzene. Reactions such as equation (53) had been postulated to I /\ Pd + PdL --) L-Pd-Pd-L (53) \c/ v R R 88 M.H. Chisholm L. A. Rankel W. I. Bailey jun. F. A. Cotton and C. A. Murillo J. Amer. Chem. SOC. 1977,99,1261. 89 (a)J. W. Koepke W. L. Jolly G. M. Bancroft P. A. Malmquist and K. Siegbahn Znorg. Chem. 1977 16 2659 and reference therein; (b) V. G. Andrianov Yu. T. Struchkov N. E. Kolabova A. B. Antonova and N. S. Obezyuk J. Organometallic Chem. 1976 122 C33. C. J. Cardin D. J. Cardin R. J. Norton and K. R. Dixon occur during their synthesis and it has now been shown that they represent a general synthetic route to (69) and related complexes.90b The reaction may also be applied to [Pd(q3-methallyl)2] (to yield two allyl bridges on the Pd-Pd bond) and to [Pd2(p-C1~(q3-methallyl)2] (to give one allyl and one halide bridge).A similar reaction of [Pt($-CsH5)(q3-methallyl)] with [Pt(PPr;),] gives the platinum ana- logue of (69). This is the first example of a platinum complex of this type.90b The stable end product of many reductions of biscyclopentadienyltitanium(rv) complexes has the formula [{Ti(C10H10)}2] ('titanocene'). Since the formation of titanocene is accompanied by loss of activity of the reduced species as a catalyst for olefin hydrogenation its structure is important and has occasioned much dis- cussion. Structure (70; X = H) is now generally accepted but there has been no X-ray proof since suitable crystals have not been obtained. However X-ray structural studies of two derivatives (70; X = OH)910 and (70; X = Cl),916 Qrepared by reactions of titanocene with H20 or HCl respectively have now confirmed structure (70).The Ti-Ti distances are 319.5 and 363.8 pm respectively and although some metal-metal interaction may exist in the hydroxy-derivative it seems certain that there is no M-M bond in the chloride. The observed diamag- netism may be explained by electronic coupling through the fulvalene bridge.91b Previous syntheses of fulvalene-bridged complexes of type (7 1)have used Ullmann I I M M coupling of iodocyclopentadienyl complexes. A new procedure based on the formation of the fulvalene dianion from sodium cyclopentadienide di-iodine and butyl-lithium is more versatile and has been applied to the synthesis of a (q5!q5-fulva1ene)hexacarbonyldimolybdenum dianionglc and (71; M =Fegld or Ni9le).The last two derivatives are both readily oxidized to +1 and +2 The reaction of the tetramer [{Mo(~'-C~H~)HL~}~] with nitrous oxide gives (72) and a cis-isomer thereof. At 50 "C this mixture is converted into (73) and further reaction with H[PF6] gives (74) whose structure has been determined by an X-ray diffraction study of the [PF6]- The other structures have been established by spectroscopic evidence and it is of interest that (72) is an analogue of the known niobium complex [(Nb($-C,H,)(p -q :q 5-C5H4)H}2].918 The ready conversion of 90 (a)H. Felkin and G. K. Turner J. Organometallic Chem. 1977,129,429;(b)H.Werner and A. Kiihn Angew. Chem. Internat. Edn. 1977 16 412. 91 (a) L.J. Guggenberger and F. N. Tebbe J. Amer. Chem. SOC.,1976 98 4137; (6) G.J. Olthof J. Organometallic Chem. 1977,128,367;(c)J. C. Smart and C. J. Curtis Inorg. Chem. 1977,16 1788; (d)C. LeVanda K. Bechgaard D. 0.Cowan U. T. Mueller-Westerhoff P. Eilbrecht G. A. Candela and R. L. Collins J. Amer. Chem. SOC.,1976,98,3181; (e)J. C. Smart and B. L. Pinsky J. Amer. Chem. SOC.,1977,99,956;(nN. J. Cooper M. L. H. Green C. Couldwell and K. Prout J.C.S.Chem. Comm. 1977 145; K. Prout and M. C. Couldwell Acra Cryst.,1977 B33.2146; (g)L.J. Guggenberger Znorg. Chem. 1973,12,294. Organometallic Compounds (72) into (73) provides some insight into the mechanism of formation of fulvalenes in titanocene and related complexes. Finally we note the isolation of another very unusual cyclopentadienyl derivative of titanium.During homogeneous hydrogenation of CO to CH4 using [Ti(q5- C5H5)2(CO)2]in toluene a royal-blue complex [Ti6(q 5-C5H5)608] is slowly formed. X-Ray diffraction shows an octahedron of Ti atoms with oxygen atoms triply bridging each face and cyclopentadienyl groups capping each Ti vertex. Unfortunately the cluster is catalytically inactive!92 92 J. C. Huffman J. G. Stone W. C. Krusell and K. G. Coulton J. Amer. Chem. SOC.,1977,99 5829.

ISSN:0308-6003    

DOI:10.1039/PR9777400215

出版商:RSC    

年代:1977

数据来源: RSC