|
1. |
Proceedings of the Chemical Society. August 1961 |
|
Proceedings of the Chemical Society ,
Volume 1,
Issue August,
1961,
Page 273-320
Preview
|
PDF (5163KB)
|
|
摘要:
PROCEEDINGS OF THE CHEMICAL SOCIETY AUGUST 1961 TILDEN LECTURE* Electron Configuration and Structure of Transition-metalComplexes By R. S. NYHOLM THEaim of this Lecture is to bring together some recent experimental advances in the sti-uctural chem- istry of the later transition elements and to correlate these with modern views on atomic structure. Whilst a considerable number of problems have been solved during the past few years many intriguing ones still await solution. Attention is directed towards certain of these since a lecture of this type is more likely to be of some value if it indicates the avenues along which research might proceed rather than if it provides merely a summary of past achievements. It being necessary to limit our objectives we shall be concerned mainly with the structure of the later transition elements is.,those containing 5-10 non-bonding d-electrons in the oxidation state being dis- cussed.Some reference however will be made to compounds containing fewer than five d-electrons owing to the marked similarities which arise from related electron configurations (e.g. d3 with d8,d4 with d9). Instead of discussing the complex chemistry in terms of horizontal and vertical sequences of ele- ments we shall find it more convenient to approach it in terms of the non-bonding configuration of the metal atom because so many of the structural similarities are dependent upon isoelectronic con- figurations. The survey is concerned almost entirely with a-bonded complexes it excludes .n-donors of the ethylene benzene etc.type except in so far as generalisation based on o-bonded complexes are applicable to these also. Also the structure of ionic compounds which have been dealt with in a recent review,l will not be discussed here. Taking a wide view of structure one is interested in the value of and the correlation between the following (a) oxidation state; (b) co-ordination number; (c) stereochemistry of the metal atom (and much less frequently molecular conformation) ; (d)nature of the bonds. The structure of transition-metal complexes may be discussed by starting from one of two extremes using purely electrostatic (crystal-field) or covalent- bond2 (valence-bond or the molecular-orbital) models.In fact these are complementary rather than contradictory and a satisfactory compromise is found in the so-called ligand-field theory which combines covalent binding with the crystal-field theory. It is convenient to start from the crystal-field model when the bonds are fairly polar (e.g. as in [FeF6I3-) and a valence-bond model when the bonds are essentially covalent [e.g. as in Ni(CO),]. It is trite but true to say that whereas the formation of an organic com- pound is usually kinetically controlled the product formed in an inorganic reaction is usually decided by * Delivered before The Society on November loth 1960 at Burlington House London on December 8th 1960 at Aberystwyth and on April 20th 1961 at St. Salvator’s College St. Andrews.Orgel and Dunitz in “Advances in Inorganic and Nuclear Chemistry,” ed. Emelkus and Sharpe Academic Press, New York 1960 Vol. 11 p. 1 ;see also Nature 1957 179,462. Pauling “The Nature of the Chemical Bond,” Cornell Univ. Press New York 1960 3rd edn. 273 the laws of thermodynamics. Hence the energetics of bond formation are of major interest. Unfortunately our knowledge of the relationships between atomic levels and bond energies leaves much to be desired; however for comparative purposes promotion energies are of considerable value. Even these are often unavailable and considerable use will be made here of ionisation potentials and electron affinities; we prefer to use these experimental data rather than derived electronegativity values which vary with the method used to compute them.An excellent survey of these is available.” Many of the anomalies between different sets of electro- negative values have been explained. Two main questions arise at the outset first why do the transition elements differ in their behaviour from the non-transition elements ? and secondly why do the three transition series show certain marked differences from one another? The usual answer given to the first question is simply that the transition elements can have an unfilled d-shell; without further elaboration this is not an answer to but a rephrasing of the original question. The answer to both questions lies in the relative shielding powers of filled (and unfilled) s- p- and d-shells and this together with the equally important and inter- dependent atomic radius will be our main concern.If the metal-ligand bond is essentially covalent then the strength of attachment of a ligand must depend mainly on the attractive force exerted upon its bond- ing pair of electrons by the metal atom. To obtain a relative estimate of this we evaluate certain simple concepts as a guide. In the absence of accurate quantum-mechanical calculations as yet on com- plexes of the transition metals one must fall back on the simplest experimental parameters available to us. As a background to the discussion of the various configurations it is convenient first to examine ionisation potential electron affinity and screening effects for the dlO-configuration.A comparison of these for the dlOsl-triad (Cu,Ag and Au) and the alkali metals provides a useful way of comparing the non-transition and the transition elements. Each of these atoms has in the ground state a single s-electron outside a filled s2- s2p6-,or dlO-sub-shell. The first ionisation potentials (I.P.) are given in Table 1 and Figure 1.3 The graph emphasises the differences between the non-transition and the transition elements and also the big increase in I.P. in passing from copper and silver to gold. To interpret these figures and to apply PROCEEDINGS them to structural chemistry it is desirable to have some idea of the relative importance of the effective nuclear charge Z* felt by the ns-electron and the radius of the atom.The value of the effective nuclear charge for a polyelectronic atom increases as the electron approaches closer to the nucleus. For example it is unity at a long distance from a Cuf ion but increases rapidly towards +29 as we move in- wards through the electronic shells. In order to obtain a comparative estimate based on experimental data of the value of Z* at radii of the order observed in complexes it is convenient to use the radius and the ionisation potential of the free atom. Although as discussed by Slater! Z* still varies with the radius when the latter is rMo,at this distance the rate of change of Z* with radius is relatively small. For a hydrogen like atom the radius of the nth orbit is n2h2/4wZme2and the energy of the electron is given by -2r?Z2eQm/n2h2.Thus En = Ze2/2r = 7.20Z/rif E is in ev r is in A and Z in nuclear units (H = 1).Using this expression we have calculated relative Z* values for the various sl-atoms as shown in Table 1. It is fully recognised that this is at best an approximate procedure. The chief difficulty lies in estimating the radius (rMo)of the metal atom in A. We have chosen the Pauling2 metallic radii in metallic crystals where the effective co-ordination number is twelve. This is deliberately an arbitrary procedure and the estimated radii are undoubtedly less than ideal radii owing to partial ionisation of the atoms in the metallic crysta1.t A further complication arises from the fact that the percentage ionisation varies from element to element.This is expected to be less serious and to work in the sense of under- rather than over-estimating the differences between the Z* values because the percentage ionisation will we assume decrease as the I.P. increases. Hence the errors in the estimates of values are probably r,O greater for the alkali metals than for the copper- silver-gold triad. This error will increase all Z* values particularly those of the alkali metals. It is in any case difficult to define precisely what really is the radius of the s-electron in these atoms; pre- sumably it is the most probable distance of the electron from the nucleus i.e. similar to the circular Bohr orbit. The justification for using these metallic radii is simply that they are the most convenient experimental data suitable for comparative purposes; however it is considered unwise to place any reliance upon absolute values of Z* or on small differences between these.t The effect of covalent metal-metal bonding is ignored in this approximate treatment. Allred and Hensley J. Znorg. Nuclear Chem. 1961 17,43 and refs. therein. Moore “Atomic Energy Levels,” Nat. Bur. Standards (U.S.) Circular No. 467 1952 (Vol. 11); 1958 (Vol. 110; see also Finkelnburg and Humbach Nafurwrss. 1955 42 36; Lakatos Bohus and Medgyesi Acta Chem. (Hungary). 1959,21,203. Slater “Quantum Theory of Atomic Structure,” McGraw Hill New York 1960 p. 227. AUGUST1961 275 TABLE 1. Properties of elements of s1 type. Ionisation Atomic rd rw potential Experimental Z* z*/(r,o)B Element number Radius Radius calc.from I.P.x Calc. from of ion of atom' I.P. = (ev) experimental (A) (4 7.20/rM0 I.P. (ev) H 1 -0-529 13-59 13.59 1.00 3.60 Li 3 0.60 1.55 4-67 5.39 1-16 048 Na 11 0.98 1-90 3.80 5-14 1.35 0.374 K 19 1.33 2.35 3.00 4.34 1-41 0.255 Rb 37 1*48 2-48 2.90 418 1.44 0,234 cs 55 1*69 2.67 2.7 1 390 1.44 0.202 cu 29 0.96 1.28 5.65 7.72 1.37 0.837 Ag 47 1*26 1.44 5.01 7.57 1.51 0.725 AU 79 1-37 1-44 5.01 9.22 1*84 0.855 Z* = r, x I.P. (exp.)/7-20. * Ref. 2. Ref. 3. 14 H 12 AU cu Ga In TI Q I I I I J 20 40 60 80 Atomic no. FIG. 1. Ionisation potentials (full lines) and s-p promotion energies (broken lines). Since various sets of Z* values which differ from the above in absolute value although not in relative order have been put forward it is desirable to emphasise the difference between Z* values which are mathematically convenient and those which are physically realistic.Thus the energy separation between successive s-orbitals in the sodium atom (e.g. 3s -+ 4s) can be represented by a simple expression 1 E= -1 (4-fig-("t -4' in which n and nB are the values of successive principal quantum numbers. In this expression all of the deviation from the ideal hydrogen-atom be haviour is embodied in the quantum defect d and the value of Z* is taken as formally as unity. However Slaterg suggests that the effective 2for energy can be obtained from the equation E =Z2eff/n2.With H = 1-00this leads to the values Li 1.26 Na 1.84 K 2-22 Rb 2.77 Cs 3.21 Cu 3.02 Ag 3.73 and Au 4.94. All of these are considerably greater than the values we use above but in this instance all of the deviation from the ideal hydrogen atom behaviour is em- bodied in the value of Zeff,the value of n being taken as formally equal to the principal quantum number. These Zeffvalues however lead to what appear to be unduly large atomic radii (e.g.,Cu 2-14 Ag 3.54 Au 3-86A; cf. Table 1). We suggest that both Zeff and neff differ from ideal values and consider that a value for neff and Zeffbased upon experimental values of atomic radii represents more realistically the actual conditions obtaining in these atoms. Fortunately these Z* values fall naturally into three classes (i) the lithium atom where the 2s- electron is so well shielded by the Is2-pair that Z* is not much greater than for the hydrogen atom; (ii) the remaining alkali metals in which Z* is roughly 1.4 (in each of these elements the outer electron is shielded by the less effective s2p6-octet); (iii) the transition elements.Although Z* for gold is very high the values for copper and silver are similar to 0:- ~_.__I 60 40--___ 80 Atomic no. FIG. 2. Efective nuclear charge (Z*) and Z*lr2M. values for various 9-elements. PROCEEDINGS one another and to those of the alkali metals. The observed order of the I.P. for the elements copper and silver and the values of these compared with those for the alkali metals arise more from their difference in radii; we show in Fig.2 values of Z*/rMo2, a measure of the attractive force acting upon the outer electron and also of the polarising power of the metal ion at the radius rMo. The values for copper and silver are now in opposite order to their Z* values. The manifestation of poor shielding by a reduction in radius is well known and its effect with unfilled d-shells is illustrated by the reduction in the radii of the transition-metal ions as the number of d-electrons increases. For example the radius of Ti2+ (d2),0.90A decreases to 0.72 A for Ni2+ (ds).6Finally we have the very large value of Z* for gold arising partly from the inefficient shielding of the 5d1°-shell but in addition from the effect of the 4f14-shell (the lan- thanide contraction) which is still evident at the end of the third transition series.It is of interest that the poor shielding effect of a filled (n -l)dlo-shell is almost entirely neutralised if an ns2-shell is added. Thus the first ionisation potentials for the Group IIIB elements are given in Fig. 1 ;it is obvious that after boron the values are almost constant at 6 ev. As a result univalent gallium indium and thallium are much more similar to rubidium czesium and francium in their behaviour than to univalent copper silver and gold. It is of interest that in their first ionisation potentials both francium and thallium show the lanthanide effect even though shielded by the s-electron pair.The larger effective nuclear charge in the transition series is also reflected in the behaviour of the elements after copper silver and gold. Thus the variation in the energy required to break the ns2-pair and to effect s -+p or s -+d promotion causes effects which are evident as far along as the halogens; much of the unexpected behaviour of the higher oxidation states of these non-transition elements can be understood if this be taken into account. For example each of the elements nitrogen phosphorus antimony and bismuth forms a trichloride but only phosphorus and antimony give rise to pentachlorides. The failure to obtain a pentachloride of nitrogen is obviously due to the absence of 2d-orbitals but in the cases of arsenic and bismuth the difficulty arises from the large energy required to uncouple the s-pair.This is conveniently demonstrated as shown in Fig. 3 by comparing the increase in energy (d) required to remove an electron from the bipositive (s2p)ion with that required for the terpositive (s2) ion. Fig. 3 also gives the promotion energies for the s2p1 +-s2d1 li George and McClure in “Progress in Inorganic Chemistry,” ed. F. A. Cotton Interscience Publ. Inc, New York 1959 Vol. I p. 381. AUGUST1961 As4 4. A 2+ P As2 -t 1 I 0‘ ’ I I FIG.3. Promotion energies for Group V elements. Top sl+dl. Middle I.P.,-I.P., i.e. ($+$)-(s2ps2).Bottom s2p1+s2d1. transition for bipositive ions and the s14dl transi-tion for tetrapositive ions of the Group V elements; the effect of removing the shielding effect of the s2-pair for arsenic and bismuth is strikingly demon- strated.Further examples of this effect from other groups include HClO and HIO are known but HBrO has never been prepared; H,SeO is a power- ful oxidising agent but H2S04 and H2Te04 are not. Some caution is needed in using ionisation potentials as a guide to promotion energies because the order of these can change as the charge on the ion increases. For ions of the same charge however this is unusual. The ns -+np promotion energies for &elements are given in Fig. 1. The close parallel between these and the first ionisation potentials lends support to the use of the latter as a guide to promo- tion energies in covalent compounds when promo- tion energies are not available.Although attention has been directed mainly to ionisation potentials and promotion energies of the dl0-and dlOsl-metals their electron affinities are also important and these are much larger than one normally expects for a metal. The estimation of electron affinities is difficult? the usual method in- volves extrapolation of ionisation potentials for an isodectronic series. Unfortunately this depends on differences between large numbers some of doubtful reliability. A better method is to plot the square root of the ionisation potential but this means that nega- tive electron affinities are excluded; it is however the simplest comparative method when one is con-fident that electron affinities are positive.Thia method has been used to obtain the values in Table 2 which are at least comparative. The large values for the dlOsl-elements can be correlated with the fact that they are one electron short of a closed sub-shell arrangement. The high values for copper silver and gold are not so unusual if these metals be regarded as pseudo-halogens as one can picture hydrogen. In support of this view the bond-energy data in Table 2 for related diatomic molecules are relevant. The bond strengths of Cu, Ag, and Au exceed that of I2and those of Cu2and Au2exceed Br2. The Au, Ag, and Cu,molecules are of course unstable other than at high temperatures in the gas phase owing to the fact that the heat of atomisation of the metals is so high. Thus AH for the reaction Au(gas)+ 2Aupetal)is7 about 116 kcal.mole-l. It is of interest that metal-metal bonding in compounds containing atoms with an sl-confisura- tions is very common. For example Hg2C12 is the second member of a dlosl isoelectronic sequence. Metal-metal bonding as a general phenomenon occurring in compounds in which the configuration is one electron less than the s2 dlO,d*,and d6 spin-paired configurations will be discussed in the section on d9-complexes. Ionisation potentials and electron affinities of atoms may also be used to obtain some idea of the relative ease with which metal atoms will form dative double bonds with suitable ligands (see p. 283). Oxidation State.-The formal oxidation state of a metal atom is conveniently defined as the charge left on the atom when the attached ligands are removed in their closed shell configurations e.g.ammonia as NH,and chlorine as Cl-. Certain atoms or groups notably hydrogen and nitric oxide give rise to difficulty because they can be attached in more than one way in a compound and some knowledge of structure is needed before an oxidation state can be assigned. Thus if the H atom in PtBr.H.(Et,P) is regarded as a replacement for a Br atom in PtBr,(Et,P), its removal as H-gives the platinum an oxidation state of +2 consistent with the square coplanar shape of the complex. However if the hydrogen were removed as H-from the compound HCO(CO)~ the cobalt atom would have an oxidation state of + 1 whereas in the salt Na+[Co(CO),]- it would be -1.This inconsistency arises because the hydrogen atom in HCo(CO) exists essentially as a proton imbedded in a region of high electron density. In examples such as this the proton needs to be Pritchard Chem. Rev. 1953,33 529. ’ Cottrell “The Strength of Chemical Bonds,” Butterworths London 1959,2nd edn. * Chatt Duncanson and Shaw Proc. Chem. Soc. 1957 343; see also Owston Partridge and Rowe Actu Cryst., 1960,13,246. PROCEEDINGS TABLE 2. Electron afiities of atoms andbondenergies for diatomic molecules. Bond Atom Electron (A) cu affinity (ev)2*04 1st Ionisation potential' 7-72 sum of 1.P. + E.A. 9.7 energy of A8 mole-cule (ev)' 2-05 AgAu 2.2&e 2-7" 7.57 9-22 9.8 11.9 1 069 2.26 Li Na 1.lace 0.6~ee 5.39 5-14 6.5 5.7 1-09 0.75 H 0.7 13.59 14.3 4.48 F 3-48h 17.42 20.9 1-52 c1 Br 3*6gh 3*45h 13.01 11.84 16.7 15.3 2.48 1 -95 1 3~15~ 10.45 13.6 1 *55 Zn 0.7" 10.435 11.1 0.26 Cd 0.6' 8-99 9.6 0.10 Hs 1 -00 9.39 10.4 0.14 bNi bPd bPt al.2(0*8)dal*3( l.O)d a2-4(1-6)* 5.81 8.33 8-20 7.0 9.6 10.6 -- Estimated by using a plot of atomic number against dp.P.).For tPoconfiguration. For these elements Dyatkina and Syrkin "The Struc-ture of Molecules," Butterworths London 1950 p. 40, give the following estimates Na 0.3 Li 05 Au 1.0 Ag 1-35 AU 2.4. d I am lndebted to Dr. H. A. Skinner for his valuable advice in connection with the estimation of electron affinities. He believes that owing to a "kink" in the extra- polation from the zerovalent atom to the uninegative atom my estimates for Ni Pd and Pt are probably about 0.5 v too large and suggeststhe values in parenthesis.This does not affect however either the relative order of the estimates or the fact that these elements are still highly dtctron-seeking even when thiscorrection is made. Since this correction has not been made for the-other transition dements I have left the original figures intact to enable comparisons to be effected more readrly. a Edlh (J. Chem.Phys. 1960,33,98) estima? Li 0.82 and Na 0.47. For these elements the extrapolauon corn-tion is less than the value -0.5 v proposed by Dr. Skinner for Ni Pd and Pt. f Ref. 3. 8 Ref. 6. * Edlkn loc. cit. and Cubiciotti J. Chem. Phys. 1959 31,1646. removed as H+.The presence of a typical covalent metal-H bond is usually indicated by a strong stretching frequency near 2000 cm.-l. Similarly nitric oxide can be attached in various ways as is reflected in the N-O stretching frequency. In metal carbonyls and metal nitrosyls 2NO replace 3CO groups or 1NO is equivalent to a metal-metal @ Lewis Sci. Progr. 1959 47 509. loPauling J. 1948 1461. bond + ICO. Thus Fe(C0)6 + Fe(CO),(NO),, whilst Mn2(CO), -+ Mn(CO),NO. If we remove NO in its closed-shell configuration as NO-this gives the metal atoms oxidation states of +2 in Fe(CO),(NO) and +1in Mn(CO)*NO. This would however make the oxidation state of the iron atom in K,Fe(CN),NO +4. The NO group is therefore usually regarded as NO+,giving the iron an oxidation state of +2.In these compounds nitric oxide shares three of its electrons with the metal; a m-bond is formed by nitric oxide to the metal in addition to the a-bond; hence the NO+ donates a welectron pair and receives a a-pair from the metal. However whether one regards the metal atom as oxidised or reduced by the NO group in complex formation is purely a matter of convenience the main feature being that the metal atom normally reaches a closed -shell (inert-gas) configurati~n.~ The variable valency shown by the transition elements is connected with the fact that the energy to remove successive d-electrons is comparable with the extra energy obtained by the formation of extra bonds. Of course the metal does not acquire a posi-tive charge corresponding to the oxidation state but one may conveniently explain the relative stabilities of different oxidation states for a vertical sequence of elements by comparing the total ionisation energies of the metal atoms.Thus the increasing stability of higher oxidation states as we pass down a vertical group is connected with the steadily de- creasing total ionisation energies. As an example one might compare the stability of the oxidation states of nickel with those of platinum. It is known that NiCl is more stable than PtCl, but whereas K2PtCl is a very stable compound the correspond- kg quadrivalent nickel derivative is unknown. Factors such as heats of atomisation and lattice energies come into such comparisons but for com-paring higher oxidation states the total ionisation potential is probably the most important item.The difference between nickel and platinum is due to the fact that the sum of the first two ionisation potentials for Ni2+ (25.78 ev) is less than the sum for Ptz+ (27.50 ev) but the converse is true for the quadri- valent state viz. Ni*+ (1 17.0 ev) > R4+(97.1 ev). Co-ordination Number.-The preferred co-ordina- tion number of a metal atom in a complex compound depends on several factors the most important of which appear to be (i) the oxidation state of the metal; (ii) the nature of the ligand; (iii) the particular metal atom chosen; (iv) the type of bond (especially if double bonding occurs); (v) steric effects. As originally proposed by Pauling,'* it is con-sidered that the electroneutrality principle is of AUGUST1961 279 major importance in deciding the co-ordination number.In regard to the first of the above we would expect an increase in co-ordination number as the oxidation state increases; thus in the isoelectronic seriesAg+,Cd2+,andIn3+, the complex ions[AgCl,r [CdCI4I2- and [InCl6I3- are the preferred species with c1-as ligand. Next we consider the ligand; for a metal ion Mn+it is to be expected that the negative charge transferred to the metal ion from the ligand in bond formation could be obtained either by taking a lot of charge from a small number of easily polarised ligands or by taking a little charge from each of a greater number of less easily polarised ligands.As an example we find that Zn- has a preferred co-ordination number of six towards water but of only four for the more easily polarised halide ions. In using this approach one must bear in mind that the co-ordination number might also be de-creased by steric effects of bulky ligands donor atoms but we believe that these effects are much less important for simple donor atoms than has often py 4 pxorbitals Diagonal spx hybr I ds d-orbitals contract. Highly electronegative groups are required to achieve this and H20+is expected to be much more effective than the Cl atom for this purpose; in other words the more electronegative groups favour higher co-ordination numbers. It should be mentioned that certain readily polarisable groups such as CN-ion are exceptions to this generalisation.Thus whereas Cl- favours four- co-ordination for FeIII and Fe" CN- forms six- co-ordinate complexes. In this case however d,-p bond formation using d,-electrons of the metal atom is believed to occur and this by the removal of an excess of negative charge from the iron atom. allows it to exhibit a higher co-ordination number. We pass now to the third factor mentioned above. The fore- going picture of polarisation of the ligand implies some degree of covalency and the greater the percentage covalency the more important will it be for suitable metal orbitals to be energetically avail- able for complex formation. The main factor which decides the ease with which these may be hybridised ,---pa orbitaI / Trigona I sp.py hybrids TWO-Gold Three -co-wdt nat ion at rn ceordI noC ion FIG.4.Hybridisation scheme for univalent gold. been supposed. This explanation of the variation of the preferred co-ordination number with the electro- negativity of the ligand can be explained more elegantly by using Craig and Magnussen's orbital- contraction theory." To compare the co-ordination of Zn2+with 6H20 and 6Cl-to form [Zn(H20)6]2+ and [ZnCI,]*- respectively a suitable model is ob- tained by transferring one electron from each ligand to the metal ion yielding in each case Znk with a configuration 4s24p4.For bond formation to occur this ion must now be excited to the 4s4p34d2-con- figuration and the energy resulting from bond fonna- tion by the (half-filled) 4s4ps4d2-orbitals with the six ligands (i.e.6H20+and 6Cl) must fuUy compensate for this. In fact however the 4d-orbitals are too diffuse to overlap effectively at reasonable bond lengths so as to give strong bonds unless the to form equivalent bonds is the energy separation between them. Consider as an example the atoms Cuo,Ago and Auo for which the s-p separations are respectively 3.79 3-66 and 4-43 ev.9 As shown on the left of Fig 4 this separation canbe used to explain why for the same ligand the formation of sp*-tetrahedral hybrid bonds is more difficult for AuI than it is for CuI and AgI. The splitting shown on the left of Fig. 4 leads to linear two-co-ordinate Complexes and that on the right to threeco-ordinate derivatives.In spite of the larger s-p separationin ALP,forma-tion of the four-co-ordinate derivatives is still possible if one uses a chelate group which forces the atom to use all four sPg-orbitals. Chelate groups such as the ditertiary mine are able to effect thisincrease in preferred co-ordination number. Craig and Magnussen Discuss. Faraday SOC.,1958,26 116; J. 1956,4895. 280 An alternative crystal-field approach to the prob- lem of two-co-ordination in Au' Hgrr and TP has been advanced by Orgel.lSl2 It is suggested that since the non-bonding configuration 5dg6s is not far removed from the 5d1° configuration (see Table 6) the formation of linear complexes could arise because the metal atom is distorted spontaneously from a spherical shape by mixing of nd2-and (n + 1)s-orbitals.By putting two electrons from the d2-orbital into the 4k(d,2 -s)-orbi tal strong elec- tronegativity will develop along the fz and -z axes which favours linear two-co-ordination. In Fig. 5 the energy separation d10 -+ d9s is given; it +20- t 16-n 25t12-Y) V .-01-L w t8- C .-c, E 2 t4-a 0- -4' FIG. 5. Promotion energies full lines dlO+dss; broken lines dl0s-tdl0p. will be observed that ease of promotion d103d9s is relatively easier for the third transition series when the positive charge is high i.e. T13+< In3+< Ga3+ and Hg2+ < Zn2+< Cd2+ whereas this order is not preserved when the charge is low.Thus for univalent ions Au+ < Cu+< Ag+ and for zerovalent ions the order becomes Nio < Pto < PdO a reversal of the previous trend. In view of the earlier estimates of ionisation potentials and electron affinities it is reasonable to assume that the charge on these metal ions is much less than that given by the formal oxidation state. If this charge is between 0 and + 1 as seems most likely then this theory would favour two-co-ordination in the first rather than the third transition series. Thus Niomight be expected to form PROCEEDINGS two-co-ordinate complexes readily but this has not been observed. Similarly the d10 -+d9s promotion energy for Cul+ is closer to that of Aul+ than it is to Agl+ and this would tend to make Cul+ more like Ad+ than Agl+ which does not seem to be borne out by structural studies on these elements.It may be seen from Fig. 5 however that for ions with low positive charge there is still a large dlos -+ dl0p separation in the third series which favours the sp-hybrid theory. In short if the residual charge on the metal atom is snialZ an explanation of two-co-ordination based upon the larger s-p separation in the third series seems reasonable; if however the positive charge on the metal ion is appreciable then Orgel's crystal-field picture based upon the smaIZer d'O -+ d9sseparation is probably more satisfactory. It is of interest that the above two approaches are not perhaps as different as they might appear at first sight and the following model shows that they are complementary.Let us assume that as a result of a large s-p separation caused by a large effective nuclear charge (as in Au+) two linear sp-hybrid bonds are formed. This places two electron pairs on opposite sides of the spherical dlO-atom and this in turn is expected to cause polarisation of the dlO-shell leading to the kind of (n-1)d ns mixing which favours the formation of two linear bonds as predicted by crystal field theory. In connection with the number of bond mbitals available reference should be made to the way in which the inert-gas rule can be widened to include paramagnetic Complexes by regarding them as special cases of a nine-orbital rule-five (n- l)d one ns and three np. The magnetism and co-ordination number of a large number of transition-metal cyanide and substituted phosphine and arsine com- plexes and of substituted carbonyls can be under-stood by assuming that all nine orbitals are used either (a) by bonding pairs or (b) by a non-bonding electron or electron pair.The extent to which spin- pairing occurs will depend upon the relative energies of electron-pairing and the ligand-field splitting. A closer examination reveals that there are deviations from this generalisation but these occur almost exclusively at the beginning and the end of each transi- tion series. In general having been given the oxida- tion state of the metal atom in a complex in which all or most of the ligands are like CO (e.g. CN- R3P R3As RNC etc.) there is usually no doubt as to the co-ordination number.As an example the diamagnetism of the complex Rev complex cyanide points strongly to the eight-co-ordinate ion [Re(CN),l2-.I3 It has been suggested l4 that the best explanation of the inert-gas nile for carbonyls is the l3Orgel J. 1958,4186; see also Jorgensen Thesis Copenhagen 1957. l* Colton Peacock and Wilkipon Nature 1958 182 393. 1960.communicahon,personalCraig 1' AUGUST1961 281 simplest one namely that just as the preferred stable covalent bonding; (iv) n-double bonding. Spin-orbit atomic electron configuration is that of the nearest coupling becomes progressively more important as inert gas so an atom in a molecule or complex ion we pass from left to right in a given series and as the strives to reach the state of maximum electrical sym-atomic number increases in a vertical column.When metry. The deviations from the inert-gks rule which covalent binding becomes important an approach occur at the beginning and the end of the transition using bond orbitals becomes more convenient. Thus series can be understood if one bears in mind the fact for largely ionic bonds electron-pair repulsions be-that at the beginning of the series the (n-I)d-np tween ligands are less important than ligand-field TABLE 3. Stereochemistry expected from simple crystal-field theory.* Number of Unpaired non-bonding electrons Four-co-ordinate Six-co-ordinate d-electrons Spin-free 0,10 ,or 5 0 or 5 Regular tetrahedral Regular octahedral 9 or 4 1 Square planar Tet ragonal 8 or 3 2 Distorted tetrahedral Regular octahedral 7 or 2 3 Regular tetrahedral Almost regular octahedral 6 or 1 4 Almost regular tetrahedral Almost regular octahedral Spin-paired I 1 or 2 3 1 Almost regular tetrahedral -4 0 Regular tetrahedral Almost regular octahedral 5 1 Distorted tetrahedral Almost regular octahedral 6 0 Distorted tetrahedral Regular octahedral 7 1 Square planar Tetragonal 8 0 Square planar Te tragonal * Spin-orbit coupling lattice forces covalent bonding,etc.are ignored. separation is considerable whereas indeed near the effects; hence in this instance the shape is expected end of a series the use of nd-and (n+l)s-orbitals to agree more or less with that shown in Table 3. becomes energetically feasible and these rather than However as the degree of covalency increases the (tz-l)d- ns- and np-orbitals are more easily repulsions between bond pairs become more im- hybridised to reach maximum symmetry.portant and as in the cases16 of [CuC14]’-and As we shall see it is possible for certain d8-con- [CuBr,I2- are large enough to cause such a distor- figurations e.g. [Au Diarsine2Cl,]+ to employ two tion that the shape of the [CuX,I2-ion is half-way orbitals more than are available on the inert-gas rule. between a square and a regular tetrahedron. Finally Similarly one orbital less can be used as for example the use of the term “almost regular” in Table 3 is with many NiII complexes these are discussed below. trivial since effects arising from the asymmetry of the Stereochemistry.-For a given co-ordination num- non-bonding d-shell are almost certainly smaller than ber the stereochemistry is usually approached in the those due to spin-orbit coupling and covalent first instance by using a crystal-field model.lJ5 The bonding.shape expected on this model for four- and six-Nature of the Bond.-Some discussion of the co-ordination (spin-orbit coupling being ignored) is nature of the metal-ligand bond has been implicit in given in Table 3. In practice the deviations arising the foregoing but it is desirable to sum up at this from the following are by no means negligible point and to make some remarks about &-bonding (i) spin-orbit interaction leading to non-simple in metal complexes. Taking the view that the distinc- ground states (e.g.as though one mixed d5 d3, with tion between “ionic” and “covalent” bonds is a d6 d2? for octahedral NP); (ii) lattice forces; (iii) 0-matter of degree rather than of kind we sum up the l5 (a)Griffith and Orgel Quart. Rev. 1957,11 381; Orgel J. Chem. Phys. 1955,23 1819; (6) Gillespieand Nyholm Quart. Rev. 1957 11 339. l6 Helmholtz and Kruh J. Amer. Chern. SOC.,1952 74 1176; see also Felsenfeld Proc. Roy. SOC.,1956 A 236 506; Morosen and Lingafelter Acta Cryst. 1960,13 807. PROCEEDINGS variation from so-called “purely electrostatic” to for normal lone-pair acceptors (e.g. H+ or BF3) “covalent” in Table 4. forms stable complexes with zerovalent transition The use of the nephelauxetic series as a guide to metals.Without proposing some charge-transfer the covalent character of a bond in the transition back from the metal to the ligand one expects to find series has proved of considerable value.l& an improbably large negative charge on the metal Most c covalent .1 J. (1) (2) Covalent Covalent complex complex with with all non-equivalent o-bonds a-bonds. equivalent. Eg.,Fe(CO),. E.g. Cr(CO),. Three trigonal Six d2sp3- sp2-and two hybrid bond linear d2p- orbitals. bonds.* TABLE 4. Metal ion + ligand I J. (3) Mixed covalent and semi-imic. E.g. HgC12,2Py. Two linear covalent Cl-Hg bonds and four longer bonds which are largely ionic. .1 (4) Semi-ionic E.g.9 [Fe(H,O) In+. Bonds essentially ionic but some covalent character using outer orbitals.-+ Most ionic .1 (5) Ionic crystal. E.g. NafC1-. Stereochemist ry decided by sizes and ratio of number of ions. No discrete molecules or complex ions. * This division is oversimplified for convenience. Some s-character is possible in all bonds but the sp2 and dp descrip-tion represents an extreme model. It is customary to postulate d,-bonding between metals and certain ligands to explain many of the problems of complex formation. It is surprising however what few unambiguous experimental results are available to support this hypothesis. We shall develop the view that whilst d,-bonding is important in complexes with incomplete non-bonding d-shells it is much less common for the dlO-configuration than is often supposed.For double bonding involving donation of d-electrons of the metal to the ligand the following conditions must be satisfied (a) The metal atom must have available electrons in suitable d-orbitals. Hence for the non- bonding configuration dn the number of possible double bonds decreases as n approaches zero (the non-existence of carbonyls of the non-transition elements can thus be explained). (b)The ligand must have empty d-(as in R3P) or p-orbitals which can be made available (as in CO) to receive these d-electrons. (c) The sizes of the orbitals on the metal and the ligand must be such as to ensure effective overlap. It is in connection with the last that most difficulties arise.Some years ago it was shown17 that in certain model compounds some overlap is feasible but some way of assessing its comparative importance for different oxidation states electron configurations and different metals is needed. We confine our atten- tion here to carbonyls for convenience. Double bonding is invoked in these complexes largely in order to explain why a ligand which shows no affinity atom. Experimental support for double bonding based on bond-length data is unconvincing partly owing to their paucity but also because it is difficult to know just what should be the metal-CO bond distance if no double bonding occurred. Electric di- pole moments are useful,ls but stretching frequencies of the C-0 bond are probably the most compelling data available.The use of stretching frequencies is based on the premise that if carbon monoxide has thetriple bond structure C 0 then the fom-ation of a d,-bond involving a contribution from the structure M=C=O must necessarily reduce the C-0 stretching frequency. It is difficult to predict the effect of changing the metal because one is changing electronegativity size and perhaps the number of available d-electrons but one can compare other factors more readily. Thus in the isoelectronic dl0 carbonyls Ni(CO), [Co(CO),]- and [Fe(CO)J2-the negative charge increases along the series and one would expect &-bond formation by the metal to be facilitated. This is at least borne out by the figures in Table 5; the decrease in the C-0 stretching fre- quency and the bond order; suggest that the metal-C bond order has increased.The attachment of halogen to the metal atom is expected to increase its positive charge making it a poorer &-donor; this is reflected in the increase in C-0 stretching frequencies for the [Pt(CO)X,] and [Fe(CO),X,]o compounds as compared with the zerovajent carbonyls. Finally if one replaces some of the carbonyl groups with (pre- Schaffer and Jorgensen Rome conference on Co-ordination Compounds J. Inorg. Nuclear Chem. Special Suppl. 1958 143. l7 Craig Maccoll Nyholm,Orgel and Sutton J. 1954 332. Chatt and Hart J. 1958 1474. AUGUST1961 283 sumably) poorer d,-acceptors allowing more observed (cf. diarsine complexes of iron and double bonding in the remaining M-CO bonds then molybdenum).the expected decrease in C-0 stretching frequency is In order to obtain a priori some criterion of the TABLE 5. Eflecf of various factors on the C-0 stretching frequency of metal carbonyls.* Variable Compound and stretching frequency of C-0 bond (force constant lo5 dyne/cm.) Cr(CO)," Fe(CO),a Ni(CO)," (0 (0 @lo) ~~ -~ MO(CO),~ MoDiars(CO)," Mo(Diars),(CO) 2a Replacement 1984 R2022 R2119 2026 1938 1923 1914 1887 1828 of CO by fa = 16.6 fco = 15.4 and 15.3 f, = 13.9 weaker d,-accep t or Fe(C0) FeDiars(CO),b ligand 2019 1998 1996 1916 1887 f, = 16-1and 17-1 fco = 14.8 and 15.2 i Electro-t [Pt(CO)Cl2]2 1 [Pt(CO)Br,l [Pt(CO)I,]OC 2112 negative I 21 52 I __ __ 21 30 __ -attached I-__-Fe(CO),Cl is I-Fe(CO),Br," Fe(CO),I,a groups unstable I I -_____-Charge on com-I [Fe(CO),IZ-plex carbonyl 1788 -1 iond \ ___--__~-fm in i CO, 15-5 CO 19.0 i CH,=C=O 15.5 I * See also Abel Bennett and Wilkinson (J.1959 2324) for data on similar series. R = Raman. a In cyclohexane (Banaclough Thesis London 1960; see also Chatt and Hart J. 1960 1378 2047). In CHC13 (idem,loc. cit.). Irving and Magnussen J. 1956 1860. d Stammreich Sala Tavares Krumholz and Behmoiras J. Chern. Phys. 1960 32 1482. TABLE 6. Promotiorr and ionisation energies for d'O atoms and ions (e~).~ Atom Ground Promo tion Promotion Ionisation Electron or state energy energy potential affinity ion dl0-+dgsl d'*+d9pl dl0+d9 dlO-+dlOsl N io dss2+ -1-80t 1 -72 5.83 -1.2 PdO d10 0.8 1 4.23 8-33 -1.3 PtO d 9s1* -0.76 3.28 8-20 -2.4 cu+ d lo 2-72 8.25 20.29 -7.72 Ag+ 4.87 9.94 21.48 -7.57 Au+ 1.87 7-83 20-45 -9.22 Zn2+ 9.65 17.1 39.7 -17-96 Cd2+ 10.0 16.6 44.5 -16.90 Hg2+ 5.31 12.8 34.2 -18.75 Ga3+ 18.7 27.8 64-2 -30.7 1n3+ 16.0 23.1 58.0 -28.8 ~13+ 9.31 18.3 50-7 -29.8 * Energy of dlO-configuration above ground state is 1.80 ev for Ni and 0 76 ev for PtO.Negative sign Indicates energy evolved in process. ease of d,,-bonding we shall consider first d10 car-bonyls. For this purpose the data in Table 6 are useful. We assume that the relative ease with which non-bonding d-electrons will overlap empty orbitals on the ligand will be related to the ionisation potential of the metal atom in the spin-paired state (dlO) (the use of d10 -+ d9s or d10 -+ d9p promotion energies might be preferred instead; as may be seen from Table 6 they are roughly parallel with ionisation potential data).The figures suggest that the abilities to form d,-bonds are (a) Nio 9 PtO PdO; (6) Au+ > Cu+ > As+; (c) zerovalent metals > bivalent > tervalent. In practice the positive charge on the metal will not reach the values given in the Table but it is to be expected that there will be an increase as we pass from NiO -+ Cul 3 Zn*I -f GaIII for example. These data can be correlated satisfactorily with what is known of the stability of d10 carbonyls. The non-existence of carbonyls of bivalent zinc cadmium and mercury can be understood in terms of the difficulty of promoting d-electrons for n-bond formation.Univalent gold forms the solid carbonyl halide AuCOCl but the corresponding CuI deriva- tive is much less stable. The CO stretching fre- quency19 in AuCOCl (2152 cm.-l) in any case sug- gests that little double-bonding occurs. Indeed it is possible that the CO group is attached to the gold atom via the oxygen. Finally we note that neither Pt(CO) nor Pd(C0)4 has yet been prepared. E~COCEEDINGS This could perhaps be due solely to the heats of atomisation plus promotion energies to the spin-paired state but as may be seen from Table 7 the energies required for the reactions Mcryst. -+ M,pgasare 6.2 3.8 and 6.4 ev for nickel palladium and platinum respectively.On this basis alone the preparation of Pd(CO) at least should provide no difficulty. However the relative ease with which an electron can be removed from the dlO-non-bonding shell (Table 7) clearly favours Ni(CO) over Pd(CO) and Pt(CO),. Further evidence in support of the hypothesis is available from other dlO-complexes. First one can prepare substituted carbonyls of the type Pt(CO)2(R,P)2;20 in these compounds the limited donor capacity of the platinum atom is apparently increased because the metal becomes more negatively charged as the result of the donor bonds from the two PR ligands-the lone pair on which there is a much better donor than the lone pair in carbon monoxide. Also in Pt(CO),(R,P) there are fewer CO groups competing for double bonding than in pt(CO)4 and this is expected to lead to greater stability of the R-CO bond.Finally we refer to the relative ease of preparing carbonyls with less than ten non-bonding d-electrons. The fact that the d,- and do-orbitals are now in the Same quantum shell is expected to facilitate d,-bonding. The energy required to convert the metal TABLE 7. Energy (ev) for promotion and ionisation of certain transition elements (related to carbonyl formation). Metal atom Ground state and (unpaired electrons) Ionisation potential of ground statel Spin-paired state M* and (unpaired electrons) Energy required to effect spin-pairing2 Ionisation potential of spin- paired statez Heat of atomisa-tion Heat of Heat of atomisation reaction+ spin-M(CO),+ pairing M* + nCO energy4 Cr d5s1 6.76 dc6 7.33 4-01 3-67 11.0 3.77 x 6 Fe (6)d6s2 (4) (d5s1+d5) 7.90 (d6s2+d%) 7-63 (0) d2dy2 (0) d10 6-62 1.82 3.90 (d,6d,2+ d,6d l) (d,6+d,") 5.81 4.16 4.38 10.8 6-2 3.86 x 5 = 21.6 = 19.3 3-34 X 4 8.33 (dSs2-td9) d10 (0) 0.0 8.33 (dlO+d9) 3.845 3.8 ? = 13.4 8.96(d1°+-d9) dl0(0) 0-76 (dl0+ds) 8-20 5.605 6.4 ? (dgsl+d (0) (d1°+d9) M* indicates metal atom in gas phase in spin-paired state.Ref. 3. Skinner and Sumner J. Inorg. Nuclear Chem. 1957 4 245. Cottrell ref. 7; also Quill "Chemistry and Metallurgy of Miscellaneous Materials. Thermodynamics," McGraw-Hill New York 1960. * Heat required for the transition M (cryst.) -f M* atom in the spin-paired state. More recent values (Dreyer and Margrave J. Phys. Chern. 1960 64 1323) are Pd 3.94 Pt 5.85 ev.l9 Nyholm unpublished work 1960. 2o Malatesta and Carriello J. Inorg. Nuclear Chem. 1958 8 561; J. 1958 2323. AUGUST1961 285 atom from the ground-state configuration to the dn support the view that chromium and iron tend to spin-paired valency state has been estimated by form stronger r-bonds. The overall heats of forma- Skinner and Surnner,,l as shown in Table 7. Skinner22 tion of course are not necessarily the same since has pointed out that CrO and Feo in their spin-paired lattice energies and promotion energies are also in- dn-states have ionisation potentials about the same volved. It is proposed therefore that as a rough as those of the alkali metals. Thus these configura- guide to the &bonding capacity of a metal one tions will tend to lose electrons to the ligand fairly might use the ionisation potential of the metal atom readily.The Table shows also that in terms of latent in its actual (valency) state in the compound i.e. the heat of vaporisation plus energy of spin pairing only spin-paired state as in carbonyl compounds. This the carbonyls of chromium and iron should be much might be carried further in order to understand the more difficult to prepare than Ni(CO),. However if behaviour of elements in higher oxidation states double bonding is important it is clear that it will where ionisation potentials for spin-paired states are take place much more readily for chromium and iron available. than for nickel. In short although more energy is The Configurationsd'o to d5.-The dlO-configuraiion.required to reach the spin-paired state for chromium Table 8 summarises the elements and the oxidation and iron these atoms should then be capable of states which give rise to well-established dlO-com- forming very good d,-bonds-the property con-plexes. The spherically symmetrical dlO-arrangement sidered essential to enable carbonyls to be stable. is expected to show simple stereochemistry and have estimated indeed for 2- 3- 4-,5- and 6-ca-ordination one finds Cotton Fischer and Wilkin~on~~ that for the reactions iM(CO) (gas) -+ M*(gas) + examples respectively of linear { [Ag(NHJ,]+} tri- CO where M = Cr Fe or Ni n = 6 5 and 4 gonal {polymeric [Cu(CN),],-} tetrahedral respectively and M*(gas) represents the metal atom { [ZnCl4I2-) pentagonal bipyramidal (ZnCl,,Ter- in the spin-paired state the AH values are respec- pyridyl) and octahedral { [Zn(H,O) J2+) complexes.tively 87 kcal. molee1 for Cr(CO), 89 for Fe(CO), Certain unusual features however call for comment. and 77 for Ni(CO)4. These figures give an idea of the (a) Co-ordination number. As discussed above energy required to break the M-CO bond to form the preferred co-ordination number falls as the the metal atom in the spin-pared valency state and oxidation state decreases and also as the separation TABLE 8. The dlO-conJiguration. Co-ordn. co-' Nio cu' no. 2 3 -4 Ni(CO) -5 -6 PdO 2 3 4 5 6 AuI 2 CAu(CN21-3 AuT(p-Diars) 4 [A~(Diars),]~f -5 6 Note Since many apparent low oxidation states have been shown to involve a metal-hydrogen bond caution should be exercised in the interpretation of these.a Terpyridyl. p-Diars = p-phenylenebisdimethylarsine.References to each of the compounds quoted as examples in Tables 8-16 are not given except when the compound is discussed in the text. For other substances see recent Annual Reports of the Chemical Society. 21 Skinner and Sumner J. Inorg. Nuclear Chem. 1957 4 245. 22 Skinner personal communication 1960. 23 Cotton Fischer and Wilkinson J. Amer. CJzem. Soc. 1959 81 800. between the energy levels of the orbitals required for bond formation increases. The apparent exception in Ni(CO)B arises because of the charge delocalisation via dn-pn bonding (see below).Several well-estab- lished cases of three co-ordination are now available. In addition to the polymeric [Cu(CN),-] manion upon which X-ray data are available 24 the following are believed to contain three-co-ordinated metal atoms PROCEEDINGS M-M bond is expected to favour disproportionation. (c) Stereochemistry of Hg”. A noteworthy feature of the stereochemistry of the chloro-complexes of mercury is the tendency for the linear Cl-Hg-Cl unit to be retained. Addition of further ligands then tends to occur giving compounds the stereochemistry of which appear to be decided by simple electrostatics. AgSCN(Cyclohexy1,P) (b) Metal-metal bonding. The ease with which dlOsl-metals can form metal-metal bonds in Cu, Ag, and Au2 has already been referred to.The non- bonding shell is in fact dlO and such compounds are conveniently discussed here. The oxidation states involving a dlO-configuration wherein one might expect metal-metal bonds include (Ni,Pd,Pt)l- (Cu,Ag,Au)O (Zn,Cd,Hg)l+ (Ga,In,T1)2+ and (Ge,Sn,Pb),+. Examples of many of these are avail- able as may be seen in Table 10. Well-established metal-metal bonding cccurs in dimeric Ga IISsZs and in the Group IV compounds of the type R3MIII- MIIIR3,29 but the compound H2[Ni:(C0)6]2- is still little studied.30 No definite examples of compounds with TlII-TlII bonds appear to be available. The non- existence of Zn-Zn or Cd-Cd bonds similar to those in [Hg2I2+ is of interest; this seems to be connected with the greater electron affinity and/or ionisation potential in the third row.Thus the isoelectronic gold compound (Au,) is more stable than the lighter cU2 or Ag molecules. The energy involved in reactions involving disproportionation of HgI -+ Hg” + Hg are in any case small (e.g. for Hg,Cl -+HgC1 4 Hg AH = + 8 kcal. mole-l) and dG = + 6.3 kcal. m01e-l)~~~~ and any significant weakening of the Thus two more ligands give a square and six a (Et,P,Agc~cPh),,2~ CUT,(M~,PC,H,NM~,),~~(distorted) octahedron. These extra bonds are largely ionic in character. Examples of these complexes are discussed by Orgel and Dunitz.’ The d9-configuration.-Two aspects are of special interest. First as discussed earlier the paramag- netic dg-configuration is expected from crystal-field theory to favour the square four-co-ordinate or the tctragonal six-co-ordinate arrangemenkl However as more structural data become available many of the compounds in which the Cull atom was believed to be square co-ordinated are found to involve either a six-co-ordinate tetragonal or a four-co-ordinate tetrahedral metal atom.The latter is particularly true when the attached ligands form bonds which have a good deal of covalent character. The second feature is the comparative rarity of FIG.6. Structure of [Fe-1,(C0),]2-. TABLE 9. The dg-confguration. Co-ordn. no. Fe-I coo NiI CU’IU 4 -[Ni(CN) 13-a [Cu(NH3),l2+,[CUCI,]~-“i2(CN),l4-I 5 [Fe,(CO),I‘-[cO,(co),1° CuCl ,,Terp y 6 --[Cu(NH3)612+ 4 5 6 a Paramagnetism (1 unpaired electron) is displayed by most Cull and AgII complexes and by K,Ni(CN),.Similar compounds are formed by Rh and Ir. 24 Crorner J. Phys. Chem. 1957 61 1388; see also Lindquist Acta Cryst. 1957 10 29. a5 Blake Calvin and Coates Proc. Chem. SOC.,1959 396. 26 Cass Coates and Hayter Cfiem.and Ind. 1954 1485; J. 1955 1007. 27 Turco Panottoni and Frasson Ricerca xi. 1959 29 540. 28 Hahn and Frank Z. anorg. Chem. 1955 278 340. 29 Coates “Organ0 Metallic Compounds,” Methuen London 1960 2nd edn. 30 Hieber Kroder and Zahn Z Naturforsch. 1960 15b 325. 31 Quill “Chemistry and Metallurgy of Miscellaneous Materials. Thermodynamics,” McGraw-Hill New York 1960. AUGUST1961 287 TABLE10. Metal-metal bonds. Oxidation state Co-ordn. Electron number I‘ransi- -1 0 config.(including tion M-M bond) series 1 1 2 3 or 4 2 3 1 5 or 4 23 1 6 2 3 1 7 2 3 C.P. = cyclopentodienyl ion CIH5-. a The oxidation state 43 occurs in Group IV e.g.,Ph,Pb-PbPh,. $1 +2” Cf. (CP,Ni2(C0)2)0. FIG.7. Structure of Co,(CO),. paramagnetism except with Cu” or less commonly Ag”. Oxidation states with a formal ds-configuration are common (see Table 9) but in most cases the spare electron is paired off to achieve in effect a d8-non-bonding arrangement. Alternatively the con- figuration is so unstable that the atom passes to the next higher oxidation state (dlO) either by reducing the ligand or by disproportionation. We summarise the possible methods avaifable to the metal atom giving examples.(i) Metal-metal bonding (cf. Table 10). (a) Normal 0-bonding. This occurs in the [Fe2(C0),l2- ion (Fig. 6); each Fe-* atom is at the centre of a trigonal bipyramid with a metal-metal bond along the axial direction.32 This bond is also believed to occur in the diamagnetic anion of K4[Ni2(CN)B],33 the nickel becoming effectively bivalent with square dsp2-co-ordination. (b) Bent a-bonding. Bonds of this type have been postulated34 in various compounds and there are reasons for believing that a type of “banana” or bent structure can be visualised as two octahedra sharing a triangular face; at two corners of this triangle are bridging CO groups and at the third an overlap utilising the singly occupied orbitals of each CO atom ;in effect the complex uses 3d24s4p3-octahedral bonds.(c) &Bond. The third manner in which electron pairing can occur is by the formation of a &bond o-bond occurs in dimeric cobalt ~arbonyl.~~~~~ The with face-to-face overlap of all four lobes of each of probable structure of this is shown in Fig. 7. The two d,2-y2-orbitals as shown in Fig. 8. In essence 32 Mills. Dersonal communication. 1960. 33 Gnt€ith and Wilkinson J. Znorg. Nuclear Cheni. 1958,7 235; cf. Pfab and Nast Z. Krist. 1959 111,4. Coulson “Valence,” Oxford Univ. Press 1952 p. 205. 35 Mills and Robinson Pfoc. CJ7em. Soc. 1959 156. PROCEEDINGS "'p H2O FIG.8. Structure of [Cu(Acetate),H,O J2. both Cul* atoms are square pyramidal the two copper atoms interacting weakly.Compounds in which this type of bonding is postulated include cupric acetate monohydrate and the cupric salts of many other long chain carboxylic acids.% In these compounds the magnetic moment is less than the usual figure of 2.1 B.M. owing to the spin-spin inter- action and may be as low as zero e.g. as in cupric thi~acetate.~~,~ The moment of cupric acetate mono- hydrate decreases as the temperature falls; in those cases where the curve fits a Boltzmann distribution it is reasonable to postulate a temperature-dependent relationship between the diamagnetic dl0-and the paramagnetic ds-configuration. The Cu-Cu bond energy is clearly of the order of kT (-200 cm.-l = 1 kcal. mole-'). the &bond is clearly much weaker than the 0-and n-bonds and is really of quite a different kind.However since it can lead to diamag- netism its inclusion here is convenient. It is of interest that the paramagnetic Cull and AgII complexes generally occur with highly electronegative ligands such as water. Greater metal-metal inter-action with a decrease in magnetic moment occurs with the more easily polarised ligands as with thio- acetic a~id.3~ It is tempting to suggest that this arises because in these cases there is greater transfer of electron density to the copper atom with consequent increase in size of the dZ2-,2-orbitals and presum- ably an increase in overlap between the orbitals of the two different atoms. On the other hand the con- verse process of forming a paramagnetic dg-complex from a normally diamagnetic substance (e.g.Co2(CO),) is expected to occur most readily if one replaces some of the CO groups by more electro- negative donor atoms. Evidence for the formation of paramagnetic derivatives when one replaces some of the CO groups with diarsine has been obtained but the products are very unstable.% It is possible that the mechanism of certain catalysts such as Co,(CO) in the 0x0-reaction and of the Cull amines in the aerial oxidation of alcoh0l,3~ may be connected with this para-diamagnetic equilibrium. (ii) An alternative method for pairing of the single electron is by disproportionation i.e. 2d9-+d10 + d8.This provides an explanation for the instability of the AulI complexes; thus the reaction 2Au2+ -+Au3+ + Au+(AH + 10-4ev) is less difficult to bring about than the corresponding reactions for Cu:+ (+16.7 ev) and Ag2+ (+17.7 ev).(iii) Finally chemical reactions involving reduction of the anion must be considered. The best known examples of this are the spontaneous decompositions of Cur1 complexes with covalent ligands such as I-and CN-. Also Morris40 has shown that whilst AgCl is stable thermodynamically with respect to silver and chlorine it is unstable with respect to AgCl and chlorine; this means that the ion Ag2+ has such a strong tendency to gain an electron to reach the d10 configuration that it will oxidise chlorine in order to do so. The da-confgurutiun.-This electronic arrangement is together with the d10 and d6 the most widely occurring in the transition series.It extends from bi- negative Cr Mo and W to terpositive Cu,Ag and Au as shown in Table 11. Only in the cases of NiIr and CuI11 do we observe the existence of spin-free complexes (containing two unpaired electrons) and it will be convenient to discuss features of interest concerning these first. (a) Spin-free complexes. Only the co-ordination numbers four and six need be seriously considered. Six-co-ordination occurs widely in Nil1 complexes but in the case of CuIJ1 it is known only in the com- plex fluoride K,CuF,. These octahedral nickel com- plexes are formed by a large number of electronega-tive ligands such as water and ammonia. For this configuration one expects a magnetic moment slight- ly larger than the spin-only value of 2.83 B.M.because some mixing-in of upper states via spin-orbit coupling occurs. Thus the magnetic moment of [Ni(H20),I2+(3.2 B.M.) taken together with the known value of A the ligand-field splitting obtained Niekerk and Schoening Actu Crysf. 1953 6 227; see also Martin and Watennan J. 1959 1359 2960 and references therein; Ross Trans. Furuduy Soc. 1959 55 1057. 37 Nortia Suomen Kem. 1960,33 161. Nyholm and Vallarino unpublished work 1960. Brackman Chem. SOC. Special Publ. 1959 No. 13 p. 128. 40 Morris J. Phys. and Chem. Solids 1958 7 214. AUGUST1961 from spectroscopic data enables one to estimate that the effective spin-orbit coupling constant has been reduced from 335 to about 200 ~m.-l.~~ From this we conclude that there has been a considerable amount of electron-transfer from the ligand to the metal indicating an appreciable amount of covalent binding.Probably the most important development in the chemistry of paramagnetic ds-complexes during the past decade has been the clarification of the problem field theory to inorganic chemistry in 1952,therefore occurred at a time when no definite examples of tetrahedral Nil1 complexes were known. When argu- ments were based upon crystal-field stabilisation energies even the possibility of the existence of tetra- hedral NiII complexes was seriously questioned. It can be shown that given a spin-free d8-configuration the crystal-field stabilisation energies favour an octa- hedral over a tetrahedral complex by some 15-20 kcal.mole-' depending upon the ligand involved. TABLE 11. The da-spin-paired configuration. Mn-I CO' NilI -Au"' [Au(CN) I-[Auwiars),I 12+ [Au(Diars),I I+ TAS = Triarsine = Me,As~[CH,~,-AsMe~[CH,I,~AsMe,-. * This formula is still doubtful. concerning the existence and structure of paramag- netic four-co-ordinate Nil1 complexes. In his classic work on the correlation of magnetic data with structure Pauling employed Ni" as a particular example. Using a valence-bond approach he showed that for a four-co-ordinate NiII complex paramag- netism is expected to be associated with the tetra- hedral and diamagnetism with a square-planar arrangement. Crystal-field theory has provided further understanding of this but its over-enthusiastic application led on occasions to some erroneous con- clusions.In the earlier thirties the only two likely tetrahedral complexes were Ni(N03)2,2Et,P and Ni(acetylacetone),. A crystal-structure determina- tion on the latter was started some years and this led quickly to the conclusion that the compound was a trimer almost certainly involving octahedrally co-ordinated NiII. The structure has now been fully determined43 and is shown in Fig. 9 confirming that the nickel atom is octahedral. Now although Ni(N03),,2Et,P is a monomer in benzene it is at least feasible for the nitrate to be attached as a bi- dentate group in such a way as to give the Nil1atom a (distorted) octahedral arrangement.The wide- spread application of the well-established crystal- 41 Owen Proc. Roy. SOC.,1955 A 227,183. 42 Bullen Nature 1956 177,537. 43 Bullen Mason and Pauling Nature 1960 189,291. However this energy is but a small part of the total energy involvedin the formation ofa complex ion (per- haps 5-10%); furthermore it assumes that in com- parison of the two arrangements (a) all bonds are ionic i.e. that covalent bonding is negligible (6)all ligands are the same in both types of complex (c)all FIG.9. Structure of bis(acety1acetone)nickel. bond lengths are equal and (d)steric effects areneglig- ible. There is no question of the fact that if we com- pare the properties of a succession of ions of the same transition series with the same charge and with the same ligand (e.g.complex ions of the type [M(H2O),I2+},then crystal-field stabilisat ion energies -small as they are-have a clearIy discernable effect upun the relative values of heats of hydration and similar q~antities.~~~~ However their effects are much less significant when we are dealing with two com- plexes of the same ion with different ligands e.g. [Ni(H20)6]2' compared with [NiCl,]2-. We take the view that tetrahedral nickel complexes arise largely as a result of covalent binding between the Nil* atom and a weak-field ligand. Suitable ligands include the halide ions and mixtures of these with triphenylphos- phines. It has now been established that both [NiBr2,2Ph3P]0 and [P~,M~AS]~[N~C~,]~- 45 con-tain a tetrahedrally co-ordinated Wi" ion the distor- tion being large in the former but very small in the latter.The distortion in [NiCl4I2- involves lengthen- ing of one Ni-CI bond by no more than 0.05 A. In sharp contrast with regular octahedral com- plexes one expects to observe for tetrahedral Nil1 compounds a magnetic moment considerably in excess of the spin-only value. The room-temperature moment of miCl,]" (3.89 B.M.) certainly agrees with this prediction but the values for [NiBr,I2- and [NiI,I2- are much less (3.80 and 3.50 B.M. respec- ti~ely);,~ furthermore if one determines pelf over a range of temperature one finds that even lower values are obtained. Thus at -100" the moments46 of the tetrachloride tetrabromide and tetraiodide are res- pectively 3.23,3.53 and 3-21 B.M.It is apparent that the small ligand-field splitting spin-orbit coupling and deviations from perfect tetrahedral symmetry all assist to reduce the moment. A deviation from the regular tetrahedral arrangement may be geometrical- ly negligible but at low temperatures it has a very marked effect upon the separation of degenerate levels with a disproportionately large effect upon the magnetic moment. It is therefore. important to emphasise that whilst a magnetic moment consider- ably in excess of 2.83 B.M. strongly suggests a regular tetrahedral arrangement a moment Iess than this by no means excludes tetrahedral co-ordination. A final word is appropriate concerning the rarity until recently of tetrahedral complexes.Most in- organic reactions are carried out in water. In this solvent the added ligand is always competing with the very strong co-ordinating agent H,Q and the failure to isolate [NiC14j2- complexes from aqueous solution is primarily due to the fact that the Nix1 bond strong as it may be is ruptured by the solvent. Just as one cannot prepare tin tetrachloride in the presence of water-in spite of the strength of the Sn-Cl bond-similarly one cannot conclude that the [NiC1,I2- ion is unstable because it too is decom- 44 Powell and Venanzi. Proc. Chem. SOC..1956. 6. PROCEEDINGS posed by water. The use of alcohol and even more weakly co-ordinating solvents has widened the pos- sibilities in this field; thus the use of methyl cyanide for example as a solvent has enabled to prepare the [VCl,]-ion.Also by leaving out the solvent altogether and using fused reactants complex ions such as [TiCI4l2- can be prepared and isolated. MakiP has discussed recently the possibility of the occurrence of square-co-ordinated paramagnetic NiII complexes. Cotton49 has suggested that the com- plex [NiII(Ph,PO),][ClO,] may be of this type; however in the solid form one oxygen of each CD,-ion could occupy the fifth and sixth positions of what is in effect an octahedral complex. Therefore until experimental structural studies are forthcoming the existence of square paramagnetic Nil1 complexes should be regarded as somewhat speculative. (b) Spin-paired complexes. The study of diamag- netic ds-complexes was until recently largely the investigation of squareplanar four-co-ordinate com- plexes.The noteworthy exceptions to this generalisa- tion were the pentacarbonyls of iron ruthenium and osmium. Whether one takes a ligand-field or valence- bond approach the diamagnetism of a four-co-ordinate ds-complex indicates a square-planar arrangement of four ligands about the metal atom. We shall adopt the valence-bond approach in the first instance since the complexes to be discussed are formed mainly with ligands forming essentially co-valent bonds e.g. diarsine. The electron-configura- tion of square four-co-ordinate complexes is nd (n+ 0s (n+ 1lP t ++--f Non-bonding Bonding Special interest attaches to the relative ease with which the vacant p-orbital can be used for bond formation to yield a five-co-ordinate complex.Recent work has shown that five-co-ordinate d8-complexes are much more common than formerly supposed. Their preparation however requires the use of an appropriate ligand. As may be seen from Table 1 1 practically all of the known five-co-ordinate complexes of metals in the oxidation -11 -I and 0 are formed with carbon monoxide as the ligand. However for the @-oxidation states +I +II and +111we can use diarsine the cyanide ion and similar 45 Gill Pauling and Nyholm Nature 1958 182 168; Gill and Nyholm J. 1959 3997. 46 Mabbs Thesis London 1950. 47 Scaife Chem. SOC.Special Publ. 1959 No. 13 p. 152. 48 Makii J. Chem. Phys. 1958 28 162; 1959 29 1129.49 Cotton and Holm J. Chem. Phys. 1959 31 788; see also Bannister and Cotton J. 1960 1873 1878. AUGUST 1961 ligands as well as mixtures of these with halide ions. Cur11 and Ag"1 appear to give rise to one diamagnetic complex only namely KCuOz which presumably has a polymeric anion with oxygen bridging. Usually CulI1 and AgllI form spin-free complexes such as K,CuF, as mentioned earlier. Two features concerning the occurrence of dSfive-co-ordinate complexes are noteworthy. First the tendency towards five-co-ordination in a particular transition series decreases as the valency of the metal rises. Secondly it decreases as we pass from the first to the second and the third transition series. For example the red [Ni(CN),I3- ion is formed5* readily in aqueous solution by the reaction [Ni(CN),l2- + CN-+ [Ni(CN),'J3- whereas the reaction between the CN-ion and the corresponding [Pd(CN),I2- and [Pt(CN),I2- ions is negligible.51 Similarly five-co- ordinate metal complexes of the type [MII(Di-arsine),Hal]ClO are formed by Ni" Pd" and Pt"; although these are uni-univalent electrolytes in sol-vents such as nitromethane or nitrobenzene in water the Pdfl and Pt" complexes are largely dissociated to form the four-co-ordinate [MI1(Diarsine),l2+ Also the tritertiary arsine group Me,As.[CH,],.AsMe. [CH,],.AsMe (TAW gives rise to derivatives of the type IWBr,,TAS with Nil* PdlI and Pt*I;53 the nickel complex is a non- electrolyte in nitrobenzene but the PdI1 and Pt" salts are ionised much more than the Nil1 compound.Powell and his collaboratorsa have recently deter- mined the crystal structure of certain of these com- plexes and find that in the solid state whereas NiBr,,Triarsine contains a truly five-co-ordinate Nil1 atom (see Fig. lo) the PdII and PtX1 structures do not. 6r2 Br ' FIG.10. Structure of NiBr,Triarsine. Triarsine = Me,As. [CH,],-AsMe. [CH,],-AsMe,. 291 In the Nil* complex the second bromine atom com-pletes a square pyramid even though this Ni-Br bond is unusually long but for the other two com- plexes the second bromine atom is present as a Br-ion. A similar trend is apparent with the less investi- gated sequence CoI Rh' and IrI; whilst Cox forms five-co-ordinate isocyanide derivatives e.g.[Co(RNC),]ClO, RhI gives a four-co-ordinate complex.55 The difference between the first and the later two transition series shows up most readily with ligands other than carbon monoxide i.e. those whose greater electronegativity and lower v-bonding capacity are believed to lead to a greater positive charge on the metal atom than occurs with carbon monoxide. We consider that this factor explains the observed trend. It seems reasonable to postulate that the ease of formation of five-co-ordinate complexes depends upon the relative ease with which the third np-orbital can be hybridised with the (n-l)dnsnp2 orbitals; this in turn depends upon the energy separation. This will increase as the positive charge on the metal atom rises.The residual positive charge on the metal atom will increase as we increase the (positive) oxidation state; this explains why the tendency towards five-co-ordination decreases as we pass along the sequence Cr-11 Mn-1 Feo &+I Ni+I1. It is also expected to increas in any vertical column as the effective nuclear charge rises i.e. NirI < Pdrr < PtI1. There are experimental data to support this view. Near the end of a transition series the energy separation between (n-1)d-,ns- and np-orbitals is very sensitive to the effective charge on the metal atom. Thus as shown in Table 6 the promotion energies(n-l)d1° -+ (n-l)d9 ns are Nio 1-71 < Pto 3.28 < PdO 4.23 ev; however this order changes markedly for the bipositive Zn2+ Cd2+ and Hg2+ ions for which the sequence is Hg2+ 12.8 < Cd2+ 16.6 < Zn2+17-1 ev.We consider that a high posi- tive charge on the ds-metal atoms is most unlikely and suggest that it lies between 0 and +1 at the outside. A guide to the (n-l)d -np separation for &corn- plexes is obtained by comparing the promotion * energies for the transition (n-l)d9 + (n-l)d*np * transition where ds indicates the sgin-paired con- figuration such as occurs in diamagnetic ds-com- plexes. These data for atoms with charges 0 + 1 and +2 are given in Table 12. The most striking feature is that when the charge is zero the order for the 50 Morris and Nyholm quoted in Chem. Rev. 1953 53 263; McCullough Jones and Penneman 3. Znorg. Nuclear Chem. 1960 13 286.51 Reddy Thesis London 1961. 52 Hans Nyholm and Phillips J. 1960 4379. 53 Barclay and Nyholm Chem. and Znd. 1953 378; Parish Thesis London 1959. ti4 Mair Powell and Henn Proc. Chem. SOC.,1960,415. 55 Malatesta ref. 5 p. 283; see also Vallarmo J. 1956 1867; 1957 2287 2473. d -+ p separation is 1st row < 2nd row < 3rd row. However for a positive charge of f2 units this order is completely reversed. TABLE12. Ionisation and promotion energies of dQ-atomsand ions (ev). Element coo Ni+ cu2+ d9+d8Pa 0.85 2-90 6-01 I.P. 6-67 18-15 37.08 Element Rho Pd+ Ag2* d9+d8p a 1-60 3-39 5-76 I.P. 7.3 1 19-42 34.82 Element Iro Pt+ Au2+ d94d8g 2.4 3~05~ 5*2b I.P. 7.95 18.56 30.5 a This refers to the dS(2D)-+@('D)P(~F) transition the dS('D)state being spin-paired."ese are estimated values based on ionisation potentials and other promotion energies. They are believed correct to within f0.1 ev. One needs to mention also the role of the ligand in deciding whether five- or four-co-ordination is favoured. As the ligand becomes more electro-negative the metal atom becomes more positive and this will favour four-co-ordination. This is illustrated by four-co-ordination in complexes such as [NiCI2,2Et,P]O. Finally reference should be made to the bisdi- methylglyoximato-metal11 complexes of NF Pd" and Ptr*. These all have an unusual layer structure involving the stacking of molecules one on top of the other so that the metal atoms are ~o-linear.~~ Studies of the pleochroism support the suggestion that metal-metal interaction occurs;57 in the case of the Nix[ compound the Ni-Ni bond has an energy of the order of 12 kcal.mole-l. The extreme structure pro- posed for the Nirr complex utilises the &-pair of electrons on one metal atom to form a bond with the vacant p-orbital on the next (Fig. 11). In a sense one can regard the insolubility of bisdimethylglyoximato- nickel(Ii) and its consequent value in analytical chemistry as one result of the strong tendency on the part of NF to utilise the vacant 4p-orbital. The stereochemistry of these five-co-ordinate complexes is of interest. On the basis of infrared and Raman spectra the structures of Fe(CO) is now fairly well established as a trigonal bipyramid;= dipole-moment studies on Fe(CO),(PhNC) support PROCEEDINGS this conclusion.59 Surprisingly however the com- pound NiBr,Triarsine is square pyramidal albeit with slight distortionM (see Fig.10). A lot more t A .' w I N I FIG.11. Structureof bisdimethylglyoximatonickel(n). structural work is needed to enable us to understand why d8-complexes are sometimes trigonal bi-pyramidal and sometimes square pyramidal. There has been much speculation whether a dx2-y2- or a dz2-orbital would be hybridised with the s-and three p-orbitals in these five-co-ordinated complexes ; it appears that this depends upon the ligands used and the ease with which d- s- and p-orbitals can be hybridised. We are tempted towards the view that with ligands which are more covalent in type such as carbon monoxide one approaches a symmetry of quasi-inert-gas type.If this occurs then the stereo- chemistry with maximum symmetry namely the tri- gonal bipyramid is likely. However with the more polar ligands greater d-s-p separation is likely and as a result the structure can be pictured as resulting from the use of square dsp2-bond orbitals with an extra p-orbital less effectively mixed in. This would naturally favour the square-pyramidal arrangement. Finally we refer to the six-co-ordinate complexes. The total number of electrons in the valency shell of the metal atom in this instance exceeds 18 and hence 10 orbitals are needed to accommodate them. It is noteworthy that six-co-ordination appears to arise only where the charge on the metal atom is high jaGodycki and Rundle Acta Cryst.1953 6 478; Rundle J. Phys. Chem. 1957 61 45. 57 Yamada and Tsuchida Bull. Chem. SOC.Japan 1954 27 156; Yamada Nakamura and Tsuchida ibid. 1957, 30,953. j8 Sheline and Pitzer J. Amer. Chem. SOC.,1950 72 1107; King and Lippincourt ibid. 1956,78,4192; cf. O'Dwyer, J. Mol. Spectroscopy 1958 2 144. 5g Cotton and Parish .I., 1960 1440. AUGUST 1961 before the extra two ligands are attached to the square four-co-ordinate structure e.g. two for Cotton and Holm62 recently extended earlier worke3 on the correlation of magnetic data with stereochem- [Pd(Diarsine),I2+ and three for [A~(Diarshe)~]~+. istry and it is now apparent that tetrahedral com-The attachment of the two extra ligands is probably due to the greater charge on the cation two of the bonds being largely ionic.Crystal-field theory pre- dicts a tetragonal arrangement; this may be inter- preted in valence-bond language by assuming that there are two separate kinds of hybridisation at the metal atom-four square (n -l)dnsnp2-bonds and two longer and largely ionic npnd- or perhaps plexes can have magnetic moments as high as 4.9 B.M. (as in the [CoI,I2- ion). This is a large deviation from the spin-only moment of 3-88 B.M. but this is a value calculated for three unpaired elec- trons in a dy4d5-configuration without spin-orbit coupling. By allowing for this and for the small ligand-field splitting (-3000 cm.-l) the value of 4-90 B.M.can be. understood. Since I-probably TABLE13. The d 7-spin-paired configuration. MnO FeI ~I a I NjlII a Co-ordn. no. CI-' 4 --[Co(Diars),I2+ -5 -[Mn(CO),Ph,P]* [FeDiars(CO),I]* [Co(PhNC),I2+ NiBr,(Et,P) 6 [cr2(co),,l2-Mn,(CO),o -[WNO2),l4-[Ni(Diars),CI ]* * These elements also Rive rise to suin-free comdexes (seebelow). Similar complex anims are formed by Mo and W. Reo and Tc"fom-similar compound. - np(n+ 1)s-bonds normal to the square. Few struc- tural results are yet available on complexes of this cc-ordination number but X-ray studies on the com- pound PdI,Diarsine indicate the expected tetragonal arrangement with four coplanar square Pd-As bonds of the usual length and two unusually long Pd-I bonds normal to this plane.60 Of the various con- figurations discussed in this Lecture few offer as much scope for preparative structural and theoreti- cal studies as does the subject of five- and six-co- ordination in d8-complexes.The d7-configuration.-(a) Complexes with the three unpaired electrons required for the spin-free configuration occur widely in the case of bivalent cobalt ;a few paramagnetic NiIrI and PdIII derivatives are known but these all contain only one unpaired electron. They are discussed below. The correlation between the co-ordination number stereochemistry spectra and magnetism of bivalent cobalt complexes has been widely investigated. The crystal-field stabilisation energy of the dY4d,3-configuration is favourable for a tetrahedral arrangement and this is observed in the complex halides and in many com- pounds of the type MX,(Hal) where X can be for example pyridinen61 In general the ligands of high electro-negativity such as water favour a co-ordination number of six whereas those which are inore easily polarised such as the halide ions give rise to four-co-ordinate tetrahedral complexes.forms the weakest ligand field of all the value of 4.90 B.M. is near the upper limit for tetrahedral complexes. X-Ray structural studies on six-co-ordinate COII complexes show that although the Co" atom is octahedral there is in all cases some deviation from the regular arrangement.64 (b) Spin-paired complexes. In discussing the spin- paired d 7-configuration one must first distinguish those compounds in which the expected unpaired electron is observed from those in which it is paired to form a metal-metal bond.First as to the para- magnetic derivatives these are restricted to the ele- ments of the first transition series. Four-co-ordina- tion is observed in a few CoII complexes usually in forced configurations such as the pthalocyanine and porphyrin complexes. As with the d8-configuration five-co-ordination is quite common examples of this being quoted in Table 13. In all cases the nine orbitals are being used one containing the unpaired electron. Unfortunately the stereochemistry of none of these compounds has been determined. Arguing from dipole-moment measurements Jensen and Nygaa~d~~ proposed that the compound NiBr3,2Et,P is square pyramidal but no X-ray study has been carried out and it would indeed be difficult as the complex decomposes fairly rapidly.However the compounds Mn(CO),Ph3P,66 Fe(CO),DiarsineI,B7 and [CO(RNC),]X~~ are all stable and warrant detailed investigation. Of special interest is the para- 6o Harris Nyholm and Stephenson Nafure 1956 177 1127; Harris and Nyholm J. 1950 2061; see also Harris and Reece Nature 1958 182 1665. 61 Gill and h'yholm J. Inorg. Nuclear Cltern. 1961 18 88. 62 Cotton and Holm J. Chem. Phys. 1960,32 1168 2979. 63 Figgis and Nyholm J. 1954 12; 1959 338. G4 See for example Bullen Acta Cryst. 1959 12 703. 65 Jensen and Nygaard Acta Chem. Scand. 1949 3,474. 66 Hieber and Fryer Gem. Ber. 1959 92 1765.67 Nigam Nyholm and Rao J. 1959 1397. magnetic compound [Mn(CO),Ph,PIo. If one uses certain other phosphines or the diarsine as ligand it is possible to obtain derivatives in which the metal-metal bond is retained; in these the "d7"-con- figuration is only formal for if we regard the Mn-Mn bond as a true covalent bond then Mn,(CO), is really an octahedral d,6-complex. This diamagnetic six-co-ordination also occurs in Re2(CO)1,. Finally a few six-co-ordinae complexes of Coil and of NiIII e.g. [CO(NO)~]~-, [Co(Di-arsine),I2+ [Ni(Diarsine),Cl,]+ are known.15b On the basis of crystal-field theory it seems likely that these will be found to have a tetragonal structure but again no X-ray work is available. The d6-confjguration (see Table 14).-This is of PROCEEDINGS one might expect for a D state even after allowance for the fact that spin-orbit coupling should increase bfffrom the spin-only value of 4-9 B.M.to about 5-2B.M. (d:d arrangement). Further work on the correlation of magnetism spectra and stereochem- istry of these complexes is needed. Four-co-ordinate presumably tetrahedral complexes with R3P and R,As have also been described by Naldi~~i;~~ the spin-paired (d:) arrangement is specially favourable for the formation of octahedral complexes for which the crystal-field stabilisation energy is 24 Dq (-150 kcal. mole-l) for CO(H~O),~+. Even the complex oxide LiCoO is spin-pa~red,~~ the F-ion alone giving rise to the spin-free state. The d:-configuration arising in zero- uni- and ptrv 6 W(c0)(3 Re(C0)61 z(Di ars) ,lo [Ir(CN)6 ]3-[PtCl612-a These elements also give rise to spin-free complexes (see above).Presumably seven-covalent complexes of the type IrC13 (chelate)* are also known. TABLE15. The d5-spinpaired configuration. Co-ordn. vo Crl MnIra FeIrIU CO'V no. 6 V(CO),? [Cr(RNC)6i+ [Mn(CN)f3]4-[Fe(CN)f3]3-[CoF6i" cf. [V(DiPY),l0 NbO Mol Tcrl Ru"' RhIv 6 -[TcCI&( Diars)23' [RU( CN)6l5- [RhF6I2-Tao W' Re" 0s"' IrIV 6 -[ReCl,(Diars),]O [os(cr\T)6]3-[IrC16]2-a These elements also give rise to spin-free complexes (see below). TABLE14. The d6-spin-paired configuration. Co-ordn. no. CrO Mnl Fe1Ia Coma 6 cr(cO)6 Mn(CO)S1 IFe(cN)6 [CdCN), Moo Tcl RU" RhIII 6 MdCO) -[Rua2(DiaE)2l0 [Rh(cN>6]3- W0 Re1 0s" IrIII b NiIv "iF6 1,-PdIV [PdF6l2-very common occurrence especially in the spin- paired state and calls for little comment.Only in the cases of ColI1,e.g. K3COF6 and Fe" are spin-free complexes known. In addition to its common octa- hedral compounds bivalent iron is now known to give rise to tetrahedral complexes of the type 2A-[FeX4I2- where A may be a large organic cation and X = C1 Br or I.68 The complex chloride [Ph3MeAs],FeCl4 is isomorphous with the corres- ponding tetrahedral nickel complex. The magnetic moments (5-43, 5.36,and 5-44 B.M. for the chloride bromide and iodide respectively) are higher than 68 Gill unpublished work 1960; see also ref. 45. bi-valent states is stabilised readily by ligands of n-acceptor type.Only in the case of TcI are well- defined compounds lacking-a deficiency which one may confidently expect to be made good in the near future.70a* The low molecular conductivity in nitrobenzene of certain complexes of the general type IrC13,2chelate is interpreted by Livingstone as evidence for seven- co-ordinate Ir(m). The d5-conJiguration.-The spin-free arrangement with five unpaired electrons occurs only for bivalent manganese and tervalent iron (see Table 15). Al-69 Naldini Gazzetta 1960 90 391. 70 Cossee Rome Conference on Co-ordination Compounds Special Suppl. to J. itwrg. Nuclear Cfienz.,1958 p. 483; see also Bongers Thesis Leiden 1957. 7mChiswell and Livingstone J. 1960 3181. * Professor H.Kaesz of the University of California has now reported the preparation of Tc,(CO),, Tc(CO),I and [Tc(CO),I] (personal communication). AUGUST1961 295 though the ligand is fluorine the complex ion [CO~~,F,J~-has a spin-paired arrangement. Another case where the spin-free arrangement might arise is Tc". Simple and complex halides of the type TcX or [TcX,]~ are unknown but the compound [T~(Diarsine)X,]~ is definitely spin-paired (see The spin-free configuration dsis spherically symmetrical and gives rise to regular tetrahedral e.g. [Mn1C1,l2- (ref. 45) and [FelI[,Cl,]- and regular octahedral e.g. [FeF,]& complexes. There had been some doubt about the first of these because the pyridinium cation used originally could theoretically give rise either to a complex acid of Mn" i.e.H2[MnC1,,2Py] or a tetrahedral compound [PyH],[MnCl,]. The use of cations such as [Et,N]+ and [Ph,MeAs]+ together with spectral and X-ray studies have established that the green [MnC1,I2- is definitely tetrahedra1.45s72 The spin-paired compounds show the usual behaviour the higher oxidation states e.g. Cow being stabilised by highly electronegative halogens whilst the lowest oxidation state (VO) requires CO. Considerable interest attaches to the carbonyl V(CO)6 for several reasons. Its preparation was announced about the same time by Pruett and Wyn~an~~ and by Calderazzo Cini and ErcoIi.' However whilst the former considered it was dia- magnetic having the formula V2(C0),, the latter claim that it is a paramagnetic with one unpaired electron.Molecular-weight data presented by the latter are difficult to interpret but there seems little doubt that the complex is paramagnetic. On the basis of the inert-gas rule one would expect the dimer to be the more stable form but against this one needs to invoke a co-ordination number of seven for purely a-bonded compounds in the first transition series. However if one takes the view that the cyclopenta- dienyl ion forms three bonds then vanadium is at least formally seven-co-ordinate in C,H,V(CO) and titanium is eight-co-ordinate in (C,H,),Ti(CO),. Univalent chromium has been stabilised with bi- pyridyl and with isocyanides but no carbonyl halides are known. Until recently the bivalent state for technetium was unknown and rhenium complexes were ill-defined.Well-characterised six-co-ordinate complexes with diarsine have been prepared recently. These have the general formula [MrIX,(Diars),]* where M = Tc or Re and X = C1 Br or I. Their 71 Fergusson and Nyholm Chem. and Id. 1960 347. 73 Buffagni and Dunn Nature 1960 188 937. 73 Pruett and Wyman Chem. and Ind. 1960 119. 74 Calderazzo Cini and Ercoli Chenz. and Ind. 1960 934. physical properties include magnetic and molecular- weight data and X-ray studies support the proposed structure. They are strong reducing agents being easily oxidised to Tc111 and ReI* derivati~es.~lJ~ Some apprently four-co-ordinate diamagnetic Re11 complexes of the type ReC12,2Py have been The unusual magnetic behaviour and the lack of physical data concerning these compounds makes speculation as to their structure unprofitable at this juncture.The configurations do to d*.-The remarks below are confined to compounds displaying unusual co-ordination numbers especially seven or eight several examples of which have been reported recently. Some of the oldest examplesf5 of seven- and eight-co-ordination occur in the complex fluorides of Zr Hf Nb and Ta e.g. [ZrF,]% and [TaF,I3-. Although these compounds contain dismzte complex anions the polar nature of the metal-fluorine bond makes it probable that interionic interactions are at least comparable with interelectronic (bond-pair) repulsions. Also the co-ordination number is un-doubtedly decided here more by size than by the orbitals available.By using a-type ligands however which favour spin-pairing and the use of 211 nine orbitals it is possible to prepare many examples of stable seven-and eight-co-ordinate compounds whose physical properties lend themselves to investi- gation by many techniques less applicable to fluorides. Suitable ligands include CN- as in [Mo(CN),I3-and [Mo(CN),]~-,~'and mixtures of CO and Diarsine with halide ions e.g. [MoIIDi-ar~ine(CO),I,]~.~* In Table 16 we summarise the possible co-ordination numbers for various electron configurations with examples of each. The field has been relatively little investigated and poses many interesting problems. Thus if we consider the iso- electronic sequence MotV ReV and Osv* well-defined eight-co-ordinate complexes of the first two are known (see Table) but as yet no similar deriva- tives of Osvl.It is possible that the unusual carbonyl fluoride of platinum Pt(CO),F can be explained in terms of this sequence.79 Finally brief reference should be made to the frequency with which metal-metal bonding occurs in this region of the Periodic Table. A feasible mechanism for the diamagnetism of MoCl (d4)in-volving metal-metal bonds in the [Mo61ClgI4+ unit 75 Curtiss Fergusson and Nyholm Chem. and Ind. 1958 625; Fergusson and Nyholm ibid. p. 1555. 76 Tronev and Bondin Doklady Akad. Nauk S.S.S.R. 1952 86 87. 77 Hoard and Nordsieck J. Atner. Chern. SOC. 1939 61 2853. Nigam Nyholm and Stiddard J.1960 1806. 7y Sharpe Proc. Chern. SOC.,1960 317 PROCEEDINGS TABLE 16. Co-ordination number and do to d4electron configurations. No. of non- Unpaired Bond bonding d- electrons orbitals Examples of cJ-bonded electrons available* complexes 0 0 9 1 1 8 2 0 8 2 7 3 1 7 3 6 4 0 7 2 6 4 5 * Co-ordinationnumber. has been discussed by Sheldon.,O A similar type of bond is likely in chromous acetate in the d3-ion [W2Cl,]3- and in several diamagnetic d2-complexes to mention but a few examples. Summary and Outlook It seems clear that the structure of complexes of metals with isoelectronic configurations are closely related provided that the differences in oxidation state are not too large. Structural variations especially in the co-ordination number do arise when ligands of very different electronegativity are used and when the separation between orbitals re- quired for bond formation becomes appreciable; in general the effect of an increased separation between orbitals is to decrease the preferred co-ordination number.In conclusion it is appropriate to draw attention to some problems of special relevance to the above discussion which call for more intensive investiga- tion. (1) First a better understanding of the nature of the metal-ligand bond in terms of experimentally available parameters is needed. In particular the rela- tion between bond strength and properties of the metal ion such as charge size and electronegativity and the size and polarisability of the ligand calls for more theoretical work.Our interpretation of even apparently simple reactions such as the replacement of fluoride ions by iodide ions and vice versa would be helped considerably if we knew more concerning the covalent character of these bonds and the charge transferred to the metal atom in the metal-ligand bond. Secondly the nature of n-bonding in complexes calls for more detailed study. A better picture of this bond and a critical evaluation of the circumstances where it is feasible are needed. (2) Thermodynamic data on metal complexes which provide actual heats of reaction for metal- Sheldon Nature 1959 184 1210. -[MoV(CN)Sl3- [ReV'(CN),]'- [Mo1V(CN),IP [ReVCl,@iars),]+ -[Mo"'(CN),]4-[Crrrr( CN) [MorW1 ,(Diars) ,] +-[Mol*I2Diars(CO),]0 [Crrl(CN)J4-?[MnrW,]2-[MnCI,Diars,H,O ]+ ligand bond formation rather than free energies of replacement of one ligand by another would assist materially in providing a better understanding of the strengths of bonds in complexes.(3) The spectra and magnetic properties of com- plexes and the correlation of these with structure especially for the heavier transition elements offer an important area for research. In particular the stage has now been reached where the subtle effects of small deviations from regular stereochemical ar- rangements and the changes in spin-orbit coupling consequent upon the redistribution of electrons between metal and ligand can be profitably investigated.(4) %Ray studies of more complex compounds are urgently needed. The purposeful study of com- plexes in isoelectronic series is of considerable interest and the changes in stereochemistry with changes in the electronegativity of the ligand will add much to our understanding of the factors influencing the structure of metal complexes. Finally we have now reached a point where the detailed study of accurate bond lengths in suitable series of complexes e.g. the C-O bond in Mo(CO), MoDiarsine(CO), and Mo(Diarsine),(CO), is feasible and should be as valuable in inorganic chemistry as have been the corresponding studies of polycyclic hydrocarbons in organic chemistry. Important as have been the results from X-ray crystallography in the past its purpose- ful application in the future is even more necessary and will undoubtedly yield exciting results.(5) Finally it is hoped that in the near future many more data on atomic energy levels particularly for the heavier atcms will become available. Reliable ionisation potentials for these heavy elements are few and rarely extend beyond the first or second stage of ionisation. Perhaps one of the main features which emerge from this survey is the great value of such information in correlating the experhental facts of inorganic chemistry. AUGUST 1961 297 COMMUNICATIONS The Solution Spectra and Crystal Structure of Lifschitz Nickel(n) Complexes By S. C. NYBURG and J. S. WOOD DEPARTMENT COLLEGE STAFFORDSHIRE, (CHEMISTRY UNIVEFWTY OF NORTH KEELE) and W.C. E. HIGGINSON (CHEMISTRY UNIVERSITY DEPARTMENT OF MANCHESTER) SINCELifschitz Bos and Dijkemal first prepared obtained from ethanol were of variable stability. We complexes of nickel(@ with meso-and active have not repeated these experiments with absolute 1,2-diphenylethylenediamine(“stilbenediamine” = ethanol nor have we attempted to prepare yellow stien) and with phenylethylenediamine there have (11) crystals but we have found from X-ray measure been several attempts to correlate the colours and ments that there are in fact at least three blue magnetic properties of these complexes with their crystals (I) (11) and (111). Blue (11) crystals are assumed structures. Lifschitz and his co-workers obtained from dilute 95 % ethanolic solution and fall found that the complexes of Ni(rneso-stien),(RCO& to a yellow-green powder when kept in air.However where R = H Me CH,CI CHCl, or CC1 can be they can be kept indefinitely in a gelatine capsule and prepared in both yellow (diamagnetic) and blue we have measured the unit cell dimensions by this 1 ,-CO Me solution Dilutn. Yellow-green Unstable in Blue (11) f +-powder 95% (paramagnteic) forms. The dichloroacetate (dca) complexes are of particular interest since they can be prepared as blue and as yellow crystals simul- taneously from 95 % ethanolic solution. Kein and Muller2 found that from absolute ethanol only blue crystals of the dichloroacetate complexes could be prepared but “otherwise” (which we take to mean in ,undried ethanol) they obtained both blue and yellow crystals.They called the yellow crystals “yellow (I)”; from the blue crystals which we call “blue (I),” they obtained what they took to be a different yellow form “(II),” by moistening the crystals with water and then drying them over phosphorus pentoxide (see Chart). The blue crystals air Ethanolic solution means. We are not certain what conditions are responsible for the preferential formation of blue (I) or blue (11). Yellow (I) crystals appear to be those in equilibrium with saturated 95 % ethanolic solution somewhat above room temperature whereas at room temperature blue (I) crystals are stable. The trans- formation from yellow (I) to blue (I) requires some days in saturated solution (dry crystals are stable) so that after cooling of a reasonably concentrated solution both forms are found side by side.In acetone dried over potassium carbonate and redistilled all samples of the complex give the same absorption spectrum (Fig. 1A). On addition of water the spectrum changes as shown and the colour from Lifschitz Bos and Dijkema 2. anorg. Chem. 1939,242,97. -a Hein and Miiller 2.anorg. Chem. 1956 283,172. PROCEEDINGS ~~ ~ blue to yellow. It is tempting to ascribe this to the equilibrium (1) Ni(stien),(dca),,xCOMe + aq + Ni(stien),’+(aq) + 2dca-(aq) + xCOMe,(aq) ...(1) a view apparently supported by the fact that addition of sodium dichloroacetate to the aqueous acetone solution reduces the prominent 22,000 cm.-l peak and enhances that at 16,500 cm.-l which is more prominent in anhydrous acetone (Fig.2). This inter- pretation however is not easily reconcilable with our complete three-dimensional crystal-structure analysis of the blue (I) crystals [1460 independent F(hkl) values; R = 0.2151. The structure given in Fig. 3 has two stien molecules centrosymmetrically co-ordinated to the nickel atom but the axial posi- 50t 401 I\ I I 30t ia I I I 1 I I 10,OOO 20,OOO 30,OOO t5.000 25,000 Wavelength ern.-' FIG.1. Spectra of the Ni(meso-stien),(dca) complex in acetone containing varying amounts of water. H,O:A 0; By6; C 14; D 24; E 36 mole l.-l. FIG.2. Spectra of a solution of the Ni(meso-stien),(dca) complex in 55 :45 vlv acetone-water (G) and (F) plus a small quantity of 2~sodium dichloroacetate in ethanol.tions are occupied not by dichloroacetate ions as eyn. (1) might lead us to suspect but by two water molecules. These water molecules are hydrogen-bonded to neighbouring dichloroacetate ions. The nature of the organic component in the solvent does not seem to affect the spectra markedly; we find that 95 % ethanolic solutions can be matched very closely with a corresponding water-acetone solution. A more realistic interpretation of the equilibrium in organic solvent-water solutions therefore seems to be N i(stien) ,( H ,O),(dca) ,.x(solv) + aq + Ni(~tien),(H,O),~+(aq) + 2dca-(aq) + xsolv(aq) ... (2) where solv = the organic solvent.We believe these results can be reconciled with recent theoretical work on nickel(1r) complexes. For many years it was assumed that the yellow forms had planar and the blue forms tetrahedral or “octa- hedral” (i.e. for bisbidentate cases tetragonal) con- figurations the type of configuration being governed by solvent temperature etc. Maki3 threw serious doubt on this interpretation as did Ballhausen and Liehr shortly after.4 The notion of a configurational change was rejected tetragonal co-ordination being assumed throughout. Changes in spectra and mag- netic susceptibility were attributed to changes in the effective ligand field at the axial positions. In a sufficiently strong in-plane ligand field an increase in axial field will destabilise the lAlg ground state and stabilise the 3B,g state.5 In sufficiently strong axial ligand fields the triplet (blue paramagnetic) 0 -6 FIG.3.Crystal structure of the blue (I) Ni(meso-stien),(H,O),(dca) complex. state becomes the ground state. It seems that the stien ligands in these complexes exert a field strong enough to bring the nickel ion to a critical condition in the presence of axially disposed water molecules. On this view all modifications of the complex should have square in-plane co-ordination of stien groups but differing axial environments. If the interpretation suminarised in eqn. (2) is correct the change in electronic state of nickel(I1) is caused at least in part by the presence of dichloro- acetate anions which are not in the first co-ordination sphere.We shall be in a better position to examine this problem more critically when we have deter- mined the X-ray crystal structure of the yellow (I) form which is under examination. We thank the English Electric Co. Data Process- ing Division for generous help with crystal-structure computations and the British Ceramic Research Association for financial assistance to one of us (J.S.W.). (Received May 31st 1961.) 3 Maki J. Chem. Phys. 1958 28 651. Ballhausen and Liehr J. Amer. Chem. SOC.,1959,81 538. Ref. 3 Fig. 6. AUGUST1961 299 Determination of tbe Geometry of the B/C Fusion of Rotenoids by Means of Long-range Asymmetric Magnetic Shielding of the Carbony1 Group By L. CROMBIE and J.W. Low (DEPARTMENTS UNIVERSITY STRAND,W.C.2, OF CHEMISTRY OF LONDON,KING’SCOLLEGE SOUTHKENSINGTON, and IMPERIALCOLLEGE S.W.7) VARIATION in the line position assigned to the be negatively shielded. The corresponding A com-1-aromatic proton in the nuclear magnetic resonance pounds1 have the l-hydrogen line near T 3.3 this spectra of rotenoids and related ketones provides a indicates that it lies in a region of low positive or convenient criterion for deciding the geometry of the negative shielding and agrees with the cis-B/c con- T Values for rotenoid derivatives. H(1) of ketol H(1) of 12a-Me ether (&)-Isorotenolone A (Ic; R = OH R’ = H) ?7 B (IIc; R = OH R’ = H) Rotenolone A(Ia; R = OH R’ = H)t 3-37 (3-43*) 2.13 3.36* 3-23 1*99 3.51 99(k)-Tephrosin B (IIa; R = OH R’ = H)$(Ie; R = OH R’ = H) 2-18*3.34 1 *96 - * All values in CDC1 except these three which are in dioxan.t Unseparated 6aj3,12aj3,5’B-and 6aa,12a a,5’j34astereoisomers. # Unseparated 6a/3,12a 4’P-and 6aa,12aj3,5’j3-diastereoisomers. Blc-ring fusion. The 1-hydrogen atoms in 6a712a- formation (VII; R = OH or OMe) where the dehydroisorotenone (IVc) (7 1-67) and 6a,12a-hydrogen atom is near the presumed nodal surface dehydrodeguelin (We) (T 1.60) are approximately of carbonyl magnetic shielding (cf. VI).2*3 Such coplanar with the 12-carbonyl group and show the stereochemical assignments agree with our earlier expected negative shieldings [cf. 12-deoxy-6’,7’-proposals,l and in cis-rotenolone A and iso-dihydrorotenone (IIIb) (T3-34)].Rotenolone B iso-rotenolone A conformation (VII) rather than (VIII) is further supported by study of hydrogen bonding.’ Shielding by aryl-ring D has been estimatedq and does not influence the interpretation. rotenolone B (12a-hydroxy-derivatives),l and their According to the nuclear magnetic resonance methyl ethers have T values of about 2.0 for their criterion natural optically active rotenone (Ia; R = I-hydrogen atoms (Table 1) and this is consistent R’ = H) (T 3.32) sumatrol (Ia; R = H R’= OH) with the rigid and fairly flat trans-B/c fusion (V; R (T 3-23) elliptone (Id; R = R’ = H) (T 3-21) and = OH or OMe) in which the l-hydrogen atoms must a-toxicarol (Ie; R = H R’= OH) (T 3.16) all have Crombie and Godin Proc. Chem. SOC., 1960,276; J. 1961,2861.Jackman “Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry,” Pergarnon London 1959. Pople Proc. Roy. Soc. 1957 A 239 550. * Johnson and Bovey J. Chem. Phys. 1958 29 1012. the cis-B/c-fusion. This applies also to (&)-p-toxicarol (linear DIE arrangement) (7 3-16) iso-rotenone (Ic R = R' = H) (r 3-21),and a number of other rotenoid derivatives with the thermodynami- cally stable B/c-fusion. By using the earlier establish- ment of the complete absolute configuration of roten~ne,~ similar conclusions can be reached for natural rotenoids by comparison of optical rotatory Buchi Kaltenbronn Crombie Godin and Whiting Kaltenbronn Siddalingaiah and Whiting J. 1961 2843. Djerassi Ollis and Russell J. 1961 1448.PROCEEDINGS dispersion curves.6 The latter method compares absolute configurations but the present method is applicable to both optically active and (f)-6a,12a-compounds and it is not necessary to draw upon a previously worked out standard of comparison. We thank Dr. L. M. Jackman for interesting discussions. (Received May 1 1 th 1961.) Proc. Chem. SOC.,1960 274; Buchi Crombie Godin The Application of the Weermann Reaction to Polysaccharides By P. S. O'COLLA,J. J. O'DONNELL, and J. A. MULLOY (DEPARTMENT UNIVERSITY GALWAY, OF CHEMISTRY COLLEGE IRELAND) AN important development in the application of periodate oxidation to elucidation of polysaccharide structure has been the discovery of two methods for the elimination of oxidised fragments without cleavage of the linkages between sugar units in the non-oxidised portion of the molecule namely the use of phenylhydrazinel and of sodium boro-hydride.We have been examining other reactions for this purpose and we have found that the Weer- mann reaction3 may be used to decompose 0x0-starch and to bring about stepwise degradation of snail galactan. Periodiate-oxidised starch was further oxidised by bromine water in the presence of strontium car- bonate. The water-insoluble acid starch { ash 0-3%; neutralisation equiv. 95 ;[a] of Na salt + 10"(c. 0-85 in H,O) was esterified with propylene oxide in aqueous suspension.* The hydroxypropyl ester [cyID + 4"(c. 1 *7 in pyridine) was suspended in methanol and treated with ammonia to yield a water-soluble compound [a] + 4" (c.1.5 in H,O) whose analyses indicated that it was a polyimide (Found N 8-3. Calc. for [CGH,O,N], N 8.1 %). The same non-dialysable polyimide was obtained by evaporat- ing to dryness a solution of the arnnionium salt of the acid starch. The starch polyimide was cleaved under Weermann conditions (alkaline hypochlorite) to yield a complex mixture of dialysable substances of low molecular weight. The degradation was carried out also with hypochlorite at pH 6 and at pH 4. Previous ~tudiesl.~ have proved that snail galactan has a highly branched structure in which half the galactose units are present as non-reducing end- groups and the remainder are triply linked (at posi- tions 1 3 and 6) and non-vulnerable to periodate.Hence this galactan is a suitable material for testing reactions which eliminate oxidised fragments from polysaccharides. Samples of periodate-oxidised galactan were de- graded once by the established techniques of Barry' and Smith.2 In each case a degraded galactan was recovered in 50% yield ([a]?+ 35" (c. 0.5 in H,O)}. Another sample of oxidised galactan was further oxidised with bromine water in the presence of calcium carbonate. The free acid was formed by deionisation neutralised with ammonia and dried to yield a water-soluble compound whose analyses indicated a polyimide [a] + 22" (c. 0.6 in H20) [Found N 4.8.Calc. for (C&f@,*C&O,N), N 4~6x1.The polyimide was treated under Weermann conditions and a degraded galactan was obtained in 50% yield very similar to those from the other degradations ([a] + 35" in H,O).The three degraded galactans were oxidised with periodate. Each consumed one mol. of periodate and produced 0.5 mcl. of formic acid per 162 g. The oxo- galactans were degraded by Barry Smith and Weermann techniques ; again degraded galactans were obtained in 50% yield. These experiments con-firm results obtained by us1 by application of the Barry technique to the stepwise degradation of snail galac tan. We thank the Chemical Society for a grant from the Research Fund. (Received May 25th 1961.) Barry Nature 1943,152,537; OColla Proc. Roy. Irish Acad. 1953,55 B7,165; Dillon O'Colla and O'Ceallachain ibid. B9 331 ;Barry and Mitchell J.1954 4020. Smith and Montgomery "The Chemistry of Plant Gums and Mucilages," Chapman & Hall London 1959. Weermann Rec. Trav. chim. 1917 37 16. Deuel Heh. Chim. Acta 1947 30 1523; Bell and Baldwin J. 1938 1463. AUGUST1961 301 Chain-inhibition of a Non-chain Reaction By SAUL G. OHEN LAUFER STANLEYORMAN and DANIEL OF CKEMISTRY,BRANDEIS WALTHAM (DEPARTMENT UNIVERSITY 54 MASS.,U.S.A.) THE long-known photochemical reduction of benzo-dicating that the inhibition was not due to quenching phenone to benzpinacol by isopropyl alcohol' by the mere presence of a compound containing appears to be a radical reaction2 of quantum yield sulphur. Nickel chloride that quenches6 excited 0.93 proceeding through an excited triplet state3 of triplet states also had no effect.When solutions of benzophenone and the free radicals (I) and (TI). the disulphides in isopropyl alcohol were used as external filters little or no diminution in the yield hv of benzpinacol resulted thereby indicating that only (1) Ph,CO + Ph,CO* a small proportion of the effective radiation would (2) Ph,CO* + Me,CH.OH -f be absorbed by the additive. Ph,C.OH (1) + Me,C.OH (11) Reactions 1-4 do not constitute a chain mechan- ism and may not be subject to the usual type of (3) Ph2C0+ (11) 3 (I) + Me,CO inhibition. Nevertheless mesityl disulphide by its (4) 2Ph2C*OH+ [Ph,C(OH)] chemical reactions prevented conversion of a 50-fold excess of benzophenone which was recovered un- It was of interest to examine the effect of thiols changed.Diphenylmethanol and acetone were not and disulphides on this reaction since these com- detected while both thiol and disulphide were pounds enter into related photochemical oxidation- quickly converted into a mixture of the two sulphur- reduction reactions4 and further thiols have been containing species during the inhibition period. The shown to act as hydrogen-transfer reagents or sulphur compounds appear to be effective as thiol catalysts in a number of free radical proce~ses.~ and as thiyl radical or disulphide converting Degassed solutions of benzophenone in isopropyl radicals (I) and (11) (formation equation 2) largely alcohol were irradiated with a mercury-arc lamp and into benzophenone and isopropyl alcohol respec- benzpinacol was isolated after measured times.The tively. Each sulphur species is converted thereby into reactions in both the absence and the presence of the alternate reactive valency state providing a chain Eflects of compounds of sulphur on the benzophenone-isopropyl alcohol reaction. [Ph,CO] 0.500 mole/l. Additive Jrradn. Benzpinacol Rel. Compound mole/l. fir.) (%) rate None -8 95 1-00 Mesityl disulphide 0-0025 8 23 0.24 Mesitylenethiol 0-0025 8 38 0.40 Mesityl disulphide 0.0075 20 4 002 Mesityl disulphide 0.0100 8 0 0.00 Phenyl disulphide 04050 3 5 0.14 Thiophenol 0-0050 6 16 0.22 Benzyl disulphide 0.0100 8 43 0.46 Diphenyl sulphide 0.0240 7 88 1.00 Nickel(r1) chloride 0.0250 3 36 1-00 additives were of approximately zero order.The mechanism for the inhibition of the non-chain results are summarised in the Table. process Thiols and disulphides in low concentration led to retardation and inhibition of formation of benz-(5) (I) + RS. -+ Ph,CO + RSH pinacol the effect depending on their structure and (R2S2) (RSH + RS-) concentration. Diphenyl sulphide had no effect in- (6) (11) + RSH -f Me,CH-OH + RS-Ciamician and Silber Ber. 1900 33 2911; 1901 34 1541. a Pitts Letsinger Taylor Patterson Recktemodd and Martin 1.Amer. Chern. SOC.,1959 81 1068. Hammond and Moore J. Amer. Chem. SOC.,1959,81 6334. Wang and Cohen J. Amer. Chem. SOC.,1959,81 3005. Bickel and Kooijman Nature 1952 170 211; Harris and Waters ibid. p. 212; Wang and Cohen J. Amer. Chenz. SOC.,1957 79 1924.Linschitz and Pekkarinen J. Amer. Chern. SOC., 1960 82 2411. We also find that formation of benzopinacol by e°Co y-irradiation of benzophenone in isopropyl alcohol7 is similarly inhibited by mesityl disulphide. This work has been supported by the U.S. Atomic Burr and Strong J. Phys. Chem. 1959,63 873. PROCEEDINGS Energy Commission. Preliminary experiments were carried out while S. G. Cohen was Fulbright Senior Research Scholar at King’s College London. (Received,June 5th 1961#) ‘C’ phosphoranes (I;R1= H R2= II OEt Me or Ph) gave no acetyl- C enes although the phosphorane 0’ ‘p (I; R1= R2 = Ph) at 300” gave (9 almost quantitative yields of di-phenylacetylene and triphenyl-phosphine oxide. More recent work has shown that when R1and R2are both alkyl the product contains a few percent of the desired dialkylacetylene as shown by gas-liquid chromatography.However when R1or R2 is phenyl or carbonyl or its equi- valent we now find that pyrolysis at 280”/10m.of the phosphoranes (1) constitutes a new and con- venient synthesis of disubstituted acetylenes (see A New Synthesisof Acetylenes By S. T. D. GOUGH and S. TRIPPETT OF ORGANIC ~DS, (DEPARTMENT CHEMLFTRY,THE UNIVERSITY 2) + WEpreviously found’ that pyro- We find that the stable ester (11) and nitrile (EI) R1 lysis of the /%ketoalkyltriphenyl- phosphoranes can be utilised in this way to give Ph3P phosphoranes which on pyrolysis lead to 4-acetylenic esters and nitriles respectively e.g. Ph,P:CHCO,Et + CHMe:CH-COCI -+ (19 Ph,P:C(CO,Et).COCH:CHMe Ph,P:CH.CN + Ph.COCI -+ Ph,P:C(CN).COPh 57% (111) 68% The synthesis of ap-acetylenic ketones and of more highly conjugated acetylenes by this method is being investigated.Hendrickson5 recently assigned structure (IV)to the compound obtained by reaction of phenacyl- idenetriphenylphosphoranewith dimethyl acetylene- dicarboxylate. Pyrolysis of this compound in vacm gave the acetylene (V) in high yield. This reaction bears a striking resemblance to the general synthesis of acetylenes described above and suggests that Hendrickson’s compound may perhaps have the phosphorane structure (VI). This point is also being investigated. One of us (S.T.D.G.) thanks the Department of Scientific and Industrial Research for a maintenance grant.(Received June 23rd 1961.) Table). (1) R’ R2 Ph Me Me Ph C02Et Me CO2Et C3H7 C0,Et Ph Acetylene Yield ( %) PhC =C-Me 59 PhCrC. Me 55 CH3C=C*CO&t 91 C,H,-CrC-CO&t 85 PhCrCC0,Et 9 1 C0,Et CH:CH.Me Me-CH:CH-C~C-CO,Et 87 CN Ph PhCrCCN 85 Suitable p-ketoalkylphosphoranes (I) may be prepared (a) from a-halogeno-ketones Ph,P OH-PhCHBr-COMe-+ Ph.CHAc.PPh,+ Br-+ Ph,P:CPh*Ac (b) from a-keto-esters, e.g. Ph,PCI CH,Ac*CO,Et -+ Ph,P:CAc.CO,Et -Et,N (c) by the action of acyl halides on suitable phos- phoranes2p4eg. Ph,P:CHMe + PhCOCI -+ Ph,P:CMeCOPh Trippett and Walker J. 1959 3874. Trippett and Walker J. 1961 1266. Homer and Oediger Chem. Ber.1958,91,437. Bestmann Tetrahedron Letters 1960 No. 4 7. Hendrickson J. Amer. Chem. Sac. 1961,83,2018. AUGUST 1961 303 The Measurement of Chemical Shifts by Nuclear Magnetic Double Resonance By J. D. BALDESCHWIELER and E. W.RANDALL (DEPARTMENT HARVARD CAMBRIDGE OF CHEMISTRY UNIVERSTTY 38 MASS.,U.S.A.) NUCLEARmagnetic double resonance has been suggested as a technique for measuring the chemical shifts of nitrogen-14 nuclei bonded to protons? The method involves irradiation of a sample with a vari- able radiofrequency field at about the 14N resonance frequency while the proton magnetic resonance spectrum is observed. The double resonance spectrum depends in detail on the frequency w, and amplitude of the second radiofrequency field H2.2For a sufficiently large amplitude of H,; and when w2 exactly equals the nitrogen resonance frequency the double resonance spectrum is particularly simple.Under these conditions the 14N nuclei are essentially broadened by coupling with the nitrogen. Double- resonance effects can be detected only by the sharpen- ing of the a-proton signal (seeFig. 1). The accuracy of the measurement is therefore poor-about -+ 30 cycleslsec. The effect of double resonance on the amhe-proton of the pyridinium ion in trifluoro- acetic acid is shown in Fig. 2. The third component “decoupled” from the other spins in the m~lecule.~ When the amplitude of H is reduced or w2 is not equal to the nitrogen resonance frequency a com- plex intermediate case is obtained.Even when the amplitude of H is too low to give complete de coupling the nitrogen resonance frequency can be obtained from the double-resonance spectrum without resorting to detailed ca1culation.l The accuracy of the double-resonance method for the measurement of chemical shifts depends upon how well the nitrogen-14 spin-spin interaction with the protons is resolved. This depends on such factors as the 14N spin-lattice relaxation time and the rate of proton-exchange processes. For the tetrahedral NH4+ ion in acid solution where the nuclear quadrupole moment does not contribute to the 14N relaxation and where the exchange of protons is slow the proton spectrum is a sharp well-resolved triplet. In this case the 14N resonance frequency can be determined to about f 0-2 cycle/sec.For the pyridinium ion in acid solution the amine-hydrogen resonance is a triplet broadened by both 14N quadrupole relaxation and proton exchange. Jn this case the accuracy of the 14N chemical-shift deter- mination is about & 5 cycles/sec. In N-methyl-formamide the 14N relaxation time is so short that the amine-proton resonance is a single but broadened peak and the accuracy of the 14N shift measurement is only about i-10 cycles/sec. In the limit of very short 14N relaxation times double resonance does not change the proton spectrum and the 14Nshift cannot be measured by this technique. The double-resonance method has been applied to measure the 14N chemical shifts in pyridine and in pyridinium ion.In pyridine the a-protons are Baldeschwieler J. Chern. Phys. in the press. a Baldeschwieler. J. Chern. Phvs.. 1961. 34. 718. H FIG.1. a-Proton resonance of pyridine (pure liquid) at 40 Mc./sec. without and with double resonance. H FIG. 2. Amine-proton resonance of pyridiniuin ion in trifiuoroacefic acid at 40 Mc./sec. without and with double resonance. of the amine-proton resonance is overlapped by the intense signal of the carboxyl-proton of the acid. The 14N shift measurements (in p.p.m.) are com-pared below with results for ammonia and the am-monium ion NH3 0.0 C5H5N -302 f10 NH4+ -10.0 f0.5 CsHaN+ -179 f1 For NH the shift is in the direction to be expected from electron-density arguments. It is suggested that the low field value of the shift for pyridine relative to the pyridinium ion is due to a larger paramagnetic contribution to the pyridine shift arising because the energy required to promote an electron to the first Shoolery Discuss.Faraday So& 1955,19 215; Ogg and Ray ibid.,p. 239; Piette Ray and Ogg J. Mol. Spectro-scopy 1958 2 66. excited state is presumably much less in pyridine than in any of the other molecules considered. The importance of paramagnetic shifts for 14N resonances has been shown independently by usual nuclear magnetic resonance methods The paramagnetic effect may also be present at the &-protons of the ring. All the ring-proton resonances of pyridine are shifted to low field upon protonation.6 The shift for the a-protons however is much smaller than for the p-protons or for the y-protons which are shifted most of all.This relatively small shift of the a-protons probably arises because the a-proton resonance in pyridine is already shifted to Phillips personal communication. Smith and Schneider Canad. J. Chem. 1961,39 1158. Katritzky and Lagowski,J. 1961,43. PROCEEDINGS low field by the paramagnetic effect. The paramag- netic effect can also be invoked to explain the anomalous shifts of the protons at the 2-and the 3-position of 4-substituted pyridine 1-oxides. Kat- ritzky and Lagowskie have noted that in these mole- cules the proton chemical shifts are not consistent with calculated values of the charge distribution. The financial support of the National Science Foundation and of the Office of Naval Research is gratefully acknowledged.(Received June 12th 1961.) Structures Involving Unshed Electron Pair:Pyramidal Configuration of Trichlorostannite Ion By D. GRDENIC and B. KAMENAR OF STRUCTURAL CHEMISTRY INSTITUTE, (DEPARTMENT AND INORGANIC RUDERBOSKOVIC ZAGREB, YUGOSLAVIA) STANNOUS chloride dihydrate crystallisesl as a dichloroaquotin(n) hydrate containing the pyramidal complex Sn(H,O)Cl, in accordance with Sidgwick and Powell's rule2* for the tetrahedral arrangement of three bonding and one non-bonding pair of electrons. We have therefore redetermined the crystal structure of the compound formulated as K,SnCl,,H,O which-on the basis of an incomplete OCl Osn OHP OK The structure of potassium chloride trichlorostannite hydrate KCl,KSnCI,,H,O viewed at an angle of about 6" to the c-axis as to avoid overlapping of chlorine atoms.analysis4-was supposed to embody infinite chains of SnCl octahedra sharing two opposite edges. Our crystallographic data which are in good agree- ment with the earlier work are as follows Ortho- rhombic a = 8.24 b = 12-05,c = 9.14 A 2 = 4 space group Pbnm (No. 62). However our analysis based on a-and c-axial projections and refined to an R-value of 0-12 in each case indicates a different structural arrangement. The compound should be formulated as KCI,KSnCl,,H,O. The Sn atoms occupy the four-fold positions (c) of "International Tables"; the pyramidal SnC1,- lies across a rpirror- plane with two Sn-Cl bonds of length 2-54 8 and one of 2-63 A and with Cl-Sn-Cl angles of 87.6" and 90.7".The potassium ion is seven-co-ordinated with two isolated C1- at 3-15 and 3.24 A four Cl atoms of two different chlorostannite anions at average distances of 3.25 A and a water molecule at 2-81 A. The Figure shows the structure as it appears when viewed at an angle of about 6" to the c-axis. In its original formulation this compound appeared to resemble K,HgCl,,H,O in which mer- cury is octahedrally co-~rdinated,~ and the re-semblance was reinforced by similar (orthorhombic) cell dimensions. In fact as is implied by the difference between the space groups the structures are wholly dissimilar. A fuller account of this work will be published elsewhere.(Received May 24th 1961.) GrdeniC and Kamenar Proc. Chem. SOC.,1960 312. * Sidgwick and Powell Proc. Roy. SOC.,1940 A 176 153. * Gillespie and Nyholm Quart. Rev. 1957 11 339. Brasseur and Rassenfcsse 2. Krist. 1939 101 389. MacGillavry Wilde and Bijvoet 2. Krist. 1938 100 212. AUGUST 1961 305 The Relative Thermodynamic Affinities of Hydroxide and Thiophenoxide Ions for Carbon By J. F. BUNNETT F. HAUSER, CHARLES and K. V. NAHABEDIAN OF CHEMISTRY PROVIDENCE, (DEPARTMENT BROWNUNIVERSITY -ODE ISLAND U.S.A.) IN organic chemistry reactions of basic (nucleo- philic) reagents with carbon and with hydrogen atoms are of great importance. Each type of reaction has a kinetic and a thermodynamic aspect Thermodynamic (equilibrium phenomena) (1) Affinity for H OMe- or OH-> PhS- (2) Affinity for C this work.Kinetic (rate phenomena) (3) Reactivity towards H RS-> RO- (4) Reactivity towards C PhS-> OMe- or OH- It is instructive to compare mercaptide with alkoxide reagents with respect to these four qualities. Mercaptides are as is well known usually stronger nucleophiles towards carbon (quality 4)' though weaker bases towards hydrogen (quality 1). Also they are kinetically more reactive towards hydrogen (quality 3) in effecting E2 elimination from t-butyl chloride (PhS- > Et0-)2 and from aa-dimethyl- phenethyl chloride (EtS- > MeO-).3 Thus the two kinetic qualities seem to run parallel.We now have evidence that hydroxide ion has greater thermodynamic affinity for carbon (quality 2) than has thiophenoxide ion. We have determined equilibrium constants for dissociation of 9-Y-10- methyl-9-phenylacridans (I) into Y- and 10-methyl- 9-phenylacridinium ion (11) where Y-is OH-or PhS- in 37.6 %:62.4 % (w/w) acetone-water at room temperature (about 25 "). The concentration of ion (11) was measured photometrically at 425 mp in PhSH-PhS-buffers and in borate buffers the latter supplying OH- at several buffer ratios. The dis- sociation constants are 1-5 x and 1.1 x mole/l. for Y = PhS- and OH- respectively. The latter value is slightly uncertain because it is based on the assumption that the ionic dissociation con- stants for water and boric acid change in parallel fashion in going from water to 37.6% acetone.Nevertheless it is clear that OH-combines with ion (11) thermodynamically about a thousand times more tightly than does PhS-. The spectra of analytically pure adduct (I; Y = PhS) and its m.p. (126-1 3 1 ") suggest contamination with the isomeric adduct (111). The spectra of the isolated alcohol (I; Y = OH) m.p. 138-140" sup-port fully the single structure (I).* We found nothing in these systems to suggest that the effects observed are due to the formation of charge-transfer ion pairs5 rather than covalent bonds. The possibility that the high dissociation constant for compound (I; Y = SPh) arises from internal crowding not present in the alcohol (I; Y = OH) is not supported by examina- tion of molecular models.These indicate little inter- action between the phenylthio-group and the rest of the molecule. To the extent that generalisation from a single comparison is warranted these results imply that thermodynamic affinity for carbon parallels that for hydrogen. This thermodynamic factor can affect rates of nucleophilic substitution at unsaturated carbon reactions which usually occur by the intermediate complex mechanism I kl I k2 I Y-+ =c-x + -c-x + =c-Y + x-k-1 I Y iw The overall second-order rate coefficient is when k- % k, virtually equal to (kl/k-l)kz;it is thus the product of an equilibrium times a rate constant. It is possible for overall rates of displacement of a con- stant substituent X by various nucleophiles to depend principally on their thermodynamic affinities for carbon for example if k is relatively insensitive to the identity of Y in (IV).A probable example is that fluoride ion alone among halogen anions displaces p-nitrophenoxide ion from p-nitrophenyl acetate.' This work was supported by the National Science Foundation. Preliminary experiments by Dr. W. D. Merritt jr- provoked our interest in this subject. (Received May 18th 1961.) Bunnett and Merritt J. Amer. Chem. SOC., 1957 79 5967. de la Mare and Vernon J. 1956,41. Bunnett Davis and Tanida unpublished work. Decker J. prakt. Chem. 1892 45 161. Kosower J. Amer. Chem. SOC.,1958 80 3253. Bunnett "Theoretical Organic Chemistry Proceedings of the KekulC Symposium," Butterworths London 1959 p.144. Jencks and Carriuolo J. Amer. Chem. SOC.,1960 82 1778. PROCEEDINGS The Stereochemistry of (-)-2-Bromo-2-nitrocamphane T. A. HAMOR ROBERTSON, By D. A. BRUECKNER J. MONTEATH and G. A. SIM DEPARTMENT GLASGOW, (CHEMISTRY THEUNIVERSITY W.2) WE are investigating by the X-ray method the creasing numbers of atoms included in the phasing stereochemistry of a number of 2,2-disubstituted calculations as they became clearly defined on the camphanes (I) which are of interest in connection maps. Subsequent refinement by the least-squares with studies of optical rotatory dispersi0n.l For ( +)-10-bromo-2-chloro-2-nitros~camphane~ we have sho1m3that the stereochemistry is as in (I; R = Br R’ = C1 R” = NO) with the halogen atom at position 2 cis to the CMe bridge.A detailed crystal-structure determination of (-)-2-brom0-2-nitrocamphane~~~ has now been method has reduced to 12% the average discrepancy carried out and we find that in this molecule also the between measured and calculated structure ampli- halogen atom is cis to the CMe bridge (see I; R = tudes as measured by the R value. The refined H R’ = Br R“ = NO2). co-ordinates yield satisfactory bond lengths and (-)-2-Bromo-2-nitrocamphane crystallises in the valency angles. orthorhombic system space group P2,2,2, with We thank Dr. Stotherd Mitchell for a supply of four molecules in a unit cell of dimensions a = 11~07 (-)-2-bromo-2-nitrocamphane. The awards of an b = 13-46 c = 7-44 A.Three-dimensional X-ray I.C.I. Research Fellowship (to T.A.H.) and a diffraction data (825 structure amplitudes) were em- Fulbright Scholarship (to D.A.B.) are gratefully ployed and the crystal structure was elucidated by acknowledged. computing successive Fourier syntheses with in-(Received June 19th 1961.) Mitchell. Watson. and Dunloo. J.. 1950. 3440. * Davidson Ph.D. Thesis Uni;.‘Glasgow; 1958. Ferguson Fritchie Robertson and Sim J. 1961 1976. Ginnings and Noyes J. Amer. Chem. SOC.,1922 44 2567. Transition-metal Nitrites By C. C. ADDISON and A. WOJCICKI B. F. G. JOHNSON,N. LOGAN (DEPARTMENT THE UNIVERSITY OF CHEMISTRY NOTTINGHAM) STUDYof the literature reveals the surprising fact are mixed at room temperature; a smoke is formed that (with the single exception of silver) not one immediately and settles to a pale green deposit of simple anhydrous nitrite of any transition metal has pure nickel nitrite [Found Ni 38.5 N 18-6,NO2- been definitely characterised.The decreased stability 61.1. Ni(N02) requires Ni 38.9 N 18.6 NO,- of the alkaline-earth nitrites compared with those of 61.1 %I.* The ultraviolet spectrum of the aqueous the alkali metals has led to a general belief that solution shows A,,,. 357 mp (E 51.1). It is of interest although the NO group can exist as a ligand in that the nitrate (the formation of which is character- transition-metal complexes which already contain istic of reactions in liquid dinitrogen tetroxide) is not other strong ligands e.g. (R,P),Pt(N02)2 or in produced so that gas-phase reaction of dinitrogen anion complexes e.g.[CO(NO,),]~- the simple tetroxide with volatile transition-metal compounds nitrites are probably of low stability. becomes a general route to the preparation of their This is not the case. We have now observed that nitrites-certain transition-metal nitrites can be obtained by Nickel nitrite is stable up to 260” in an argon direct reaction of a metal carbonyl with dinitrogen atmosphere and decomposes at 220” in a vacuum. tetroxide in the gas phase for example The infrared spectrum of the solid shows NO2-bands Ni(CO) + N,O,+ Ni(NO,) + 4CO at 1388 1333 1240 and 830 cm.-l. In addition there are strong bands at 1575 cm.-l (as observed with All the carbonyl groups are displaced when the gases organic nitro-compounds R-NO,l) and at 1080 cm.? * Fraser and Trout (J.Amer. Chem. SOC.,1936,58,2201) obtained uncharacterised mixtures of products from this L. J. Bellamy “The Tnfrared Spectra of Complex Molecules,” Methuen London 1958 p. 298. react ion. AUGUST1961 (observed only in the case of nitrito-metal bonding M-0-N =02).Some covalent bonding of the nitrite groups is therefore involved and nickel nitrite exhibits slight volatility. The bands at 1575 and 1080cm.? disappear when the compound is exposed to atmosphere. The covalent bonding also modifies chemical reactivity in non-aqueous solvents. Thus nickel Penland Lane and Quagliano J. Amer. Cliem. SOC. 481 7. nitrite undergoes ready oxidation to the nitrate in liquid dinitrogen tetroxide whereas the ionic (e.g.alkali-metal) nitrites are unaffected. We thank the D.S.I.R. for a maintenance grant (to B.F.G.J.) and the National Science Foundation for a Postdoctoral Fellowship (to A-W-). (Received June 19th 1961.) 1956 78 587;Nakaaoto Fujita and Murata ibid. 1953 Reactions of Cupric Halides with Aromatic Hydrocarbons By D. C. NONHEBEL OF CHEMISTRY COLLEGE AND TECHNOLOGY, (DEPARTMENT ROYAL OF SCIENCE GLASGOW, C.1) KOCHI~ has shown that cupric chloride dissociates in aqueous acetone with the formation of cuprous chloride chloroacetone and hydrogen chloride. More recently Fort2 extended this work to cupric bromide and applied it to the bromination of several compounds containing active hydrogen atoms in methanol or aqueous methanol.The present work involves a study of the dissociation of anhydrous cupric halides under heterogeneous conditions generally in benzene or carbon tetrachloride. A stirred suspension of cupric bromide in refluxing toluene was quantitatively converted into cuprous bromide in three days. Hydrogen bromide was evolved and bibenzyl and traces of benzyl bromide were isolated. The formation of bibenzyl indicates a free-radical mechanism with the participation of benzyl radicals. The following reaction scheme is proposed CuBr $ CuBr + Br. Br. + PhCH -+ Ph-CH,. + HBr 2Ph.CH2* 4 Ph*CH2*CH2Ph 9-Bromoanthracene was obtained on reaction of cupric bromide with anthracene which is known to be reactive towards free radicals in 60% yield in benzene and in 99% yield in carbon tetrachloride.Pyrene in carbon tetrachloride gave a 94% yield of 1-bromopyrene the reaction being somewhat slower as expected because pyrene is less reactive towards free radical^.^ Cupric chloride behaves similarly though reacting Kochi. J. Amer. Chem. SOC..1955. 77. 5274. somewhat less rapidly. Thus in toluene bibenzyl was obtained. Anthracene and pyrene in carbon tetra- chloride were converted into 9-chloroanthracene and 1-chloropyrene in 80% and 90 % yield respectively. Phenanthrene which is relatively unreactive towards free radical^,^ was recovered in > 90% yield after six days with cupric chloride in refluxing carbon tetrachloride. These results are consistent with a reversible dis- sociation of the cupric halide to cuprous halide and halogen atoms.The halogen atoms can then either recombine with cuprous bromide or react with the organic substrate. Thus a halogen atom adds to anthracene giving an intermediate 1 0-halogeno-9,lO-dihydro-Panthryl which radical reacts with a further halogen atom giving the 9-halogenoanthracene and hydrogen halide. In all experiments hydrogen halide was evolved in large quantities and two moles of cupric halide were required per mole of anthracene consistent with the above mechanism. These reac- tions occur at temperatures considerably below those at which cupric chloride* and cupric bromide" normally dissociate to the cuprous halide. They indicate that cupric halides have potential uses as halogenating agents giving in the examples quoted better yields of the 9-halogenoanthracenes and 1-halogenopyrenes than previous methods.The author thanks Mr. M. G. Murray for technical assistance. (Received June 8th 1961.) Fort J. Org. Chem. 1961 26 765. Levy and Szwarc J. Amer. Chem. SOC.,1955 77 1949. Shchukareq and Oranskaya J. Gen. Chem. (U.S.S.R.) 1954,24,1889. Barret and Guenebaut-Thevenot. Bull. SOC.chim.France 1957 409. PROCEEDINGS The Constitution of Zierone By D. H. R. BARTONand G. S. GUPTA (IMPERIAL LONDON,S.W.7) COLLEGE THE sesquiterpenoid ketone zierone,l C15H220,2 has been investigated by Birch and his collaborator^^^^ as well as by Hildebrand and S~therland.~ It has been established that zierone is a bicyclic doubly un-saturated ketone with one of the ethylenic linkages present as an isopropylidene group.Reduction with lithium aluminium hydride followed by dehydro- genation afforded zierazulene (I) the constitution of which was rigidly established by synthesis by Collins.6 Based on these and other observations the constitution (11) or (III) has been proposed5 for zierone. We now report new evidence which shows that zierone has the constitution (IV). A The nuclear magnetic resonance spectra of zierone and its semicarbazone show the presence of one >CHMe grouping and of three vinyl Me groups. No vinyl hydrogen is present so that formulae such as (IT) and (III) are excluded.In agreement with Hildebrand and Sutherland's report5 ozonolysis of zierone semicarbazone affords acetone. The latter is also formed when zierone is heated with aqueous- ethanolic potassium hydroxide. These experiments show that an isopropylidene group must be present in the c&or equivalent)-relation to the ketone group. Reduction of zierone with sodium and ethanol gave dihydrozierol (V; R = H) purified though its dinitrobenzoate (V ;R = dinitrobenzoyl) m.p. (from ether-methanol) 103-106". Oxidation of dihydro-zierol (V; R = H) with chromic acid in acetone' at 10-15" afforded dihydrozierone (VI) Am,,. 206 and 294 mp (E 5500 and 330respectively) )cahoulder 227 mp (E 2200) vmax. 1685 characterised as its 2,4-dinitro- phenylhydrazone m.p.(from ethanol) 129-1 31 O Amax. (in CHCI,) 368 mp (E 25,800). The ultraviolet spectrum of dihydrozierone is that of a &un-saturated ketone.* Dihydrozierol 3,5-dinitrobenzoate had a nuclear magnetic resonance spectrum showing the presence of one >CHMe (see above) one vinyl Me and one CHMe grouping. The reduction of one ethylenic linkage by sodium and alcohol indicates that it must (x> be conjugated with the ketone group. That this ethylenic iinkage is that of the isopropylidene group (see above) is proved by this spectrum. Ozonolysis of dihydrozierol 3,5-dinitrobenzoate in methylene dichloride at -20" gave a diketo-ester (VTT; R = dinitrobenzoyl) m.p. (from ethanol) 129-131" [aID+ 5" (c 0.95) which contained a methyl ketone grouping (nuclear magnetic resonance and iodofom test).Similar ozonolysis of dihydro- zierol (V; R = H) furnished the analogous hydroxy- diketone (VII; R = H) vmax. 1720 (Ac) and 1702 (cycloheptanone) cm.-l which on 3,5-dinitro-benzoylation gave the dinitrobenzoate (VII; R = dinitrobenzoyl) already described. Reduction of zierone with lithium aluminium hydride gave zierol (VIII; R = H) Amax. 207 mp Penfold J. Rqv. SOC.New South Wales 1926 60 104. Bradfield Penfold and Simonsen J. Roy. SOC.New South WuIes 1933 67 200. Birch Collins and Penfold Chem. and Ind. 1955 1773. * Birch Collins Hildebrand Penfold and Sutherland Cheni. Soc.Terpene Symposium Glasgow July 1957. Hildebrand and Sutherland Austral. J. Chem. 1959 12 436. Collins J. 1959 531.'Bowden Heilbron Jones and Weedon J. 1946 39. Cookson and Wariyar J. 1956 2302. AUGUST 1961 309 (E 9500) hshoulder 226 and 233 mp (E 5000 and 4350 lished (see above) proves that zierone can only be respectively). This abnormal diene spectrum is com- represented as in (IV). The rearrangement of the parable with that of photosantonic acid (IX)9p10 methyl group from to C(3)(see IV) during where a similar structural peculiarity (overcrowding) dehydrogenation has many precedents.ll Biogenetic- pertains. Zierol was characterised as its 3,5-dinitro- ally zierone may plausibly be derived from a com- benzoate(VUU;R = dinitrobenzoyl) m.p. 108-1 10" pound of normal skeleton [(XI) or equivalent] by a [manipulated at room temperature and crystallised 1,2-shift (XI; see arrows).from light petroleum (b.p. 40-60")1. Unless specified to the contrary ultraviolet spectra Ozonolysis of zierone in methylene dichloride at refer to ethanol infrared and nuclear magnetic -20° and treatment with sodium hydrogen car- resonance spectra and [a] to chloroform solutions. bonate and hydrogen peroxide and then with diazo- We thank Drs. L. M. Jackman and J. W. Lown methane gave a keto-diester (X) which showed a for the determination and interpretation of the strong iodoform test [2,4-dinitrophenylhydrazone nuclear magnetic resonance spectra. m.p. (from ethanol) 111-112" h,, (in CHCl,) 365 mp (E 19,500)]. Earlier work3s6 has proved that the ketone group One of us (G.S.G.) thanks the Colombo Plan for of zierone is in the 7(or 9)-position of the nucleus Fellowship support and the University of Aligarh (see IV).The composition of the ester (X) coupled (India) for study leave. with this evidence and with other facts now estab- (Received June 14th 1961.) van Tamelen Levin Brenner Wolinsky and Aldrich J. Amer. Chem. Sac. 1958,80 501. lo Barton de Mayo and Shafiq J. 1958 3314. l1 See especially Buchi Chow Matsura Popper Rennhard and Wittenau Tetrahedron Letters 1959 No. 6 14. The Decomposition of Quinone Azine A New Carbene By CHI-HUA WANG OF CHEMISTRY UNIVERSITY MASS.,U.S.A.) (DEPARTMENT BRANDEIS WALTHAM THEpyrolysis of benzaldehyde azines to stilbenes and decomposes this azine in solution at room tempera- nitrogen is well knownf and reported by Curtius and ture. However the light-induced decomposition Jay and subsequently by Meisenheimer Pascal furnished mainly a high-melting solid m.p.300" Howard and Hilbert and has been shown by which possesses the characteristics of phenols and kinetic studies to be of the first order in the vapour only a trace of biphenyl-4-01.' This high-melting solid phase2 and in solution3 at about 300". Zimmerman probably consists of p-hydroxypolyphenylene ethers and Somasekhara4 suggested the decomposition involves an ionic chain mechanism in which aryldiazomethane is the chain transfer species. However it is also believed to follow at least to (A) (B) (C! some extent a free-radical process a carbene formed by rearrangement of the intermediate carbene presumably being produced as shown by catalysis to a biradical (B) or a dipolar structure (C).As of ethylene p~lymerisation.~ generally believed a bivalent carbon species We now report the decomposition of quinone generated photochemically may not be electronically monoazine6 in solution at comparatively lower the same as that formed in the thermal process. It is temperatures follows essentially the free-radical also interesting that neither thermal- nor light-mechanism with formation of a carbene (A). induced decomposition furnished bip hen y1-4,4'-quinone. When the -OaOlM-solution of this azine in In toluene the product of decomposition was a benzene was refluxed the red colour changed mixture including p-benzylphenol. In hydrolytic gradually to pale yellow with almost quantitative solvents such as boiling ethanol this azine was evolution of nitrogen and the product was quanti- reduced quantitatively to 4,4'-dihydroxyazobenzene.tatively biphenyl-4-01. Light (G. E. sun lamp 60 We thank the National Science Foundation for cycle A.C. 110-125 v as the light source) also supporting this work. (Received June 8th 1961 .) Curtius and Jay J.prakt. Chem. 1889,39,45; Meisenheimer and Helm Annalen 355 274; Pascal and Normand Bull. SOC.chim. France 1911 9 1029 1059; 1912 11 21; Howard and Hilbert J. Amzr. Chem. Soc. 1932 54 3628. Williams and Lawrence Proc. Roy. Soc. 1936 156 A 444. Overberger and Chien J. Amer. Chem. Soc. 1960 82 5874. Zimmerman and Somasekhara J. Amer. Chem. Soc. 1960,82 .5865. Roedel U.S.P. 2,439,528. Willstatter and Berg Ber. 1904 37 4744; 1906 39 3488.' Friebel and Rassow J. prakt. Chern. 1901 63,455. 3 10 PROCEEDINGS Dehydrobufotenin By B. ROBINSON and G. F. SMITH (DEPARTMENT OF CHEMISTRY THEUNIVERSITY MANCHESTER) A. H. JACKSONand D. SHAW (DEPARTMENT OF CHEMISTRY THEUNIVERSITY LIVERPOOL) B. FRYDMAN and V. DEULOFEU (DEPARTMENT AIRES) OF SCIENCE THEUNIVERSITY BUENOS BUFOTHIONIN, one of the components of the skin of The many reactions of these two compounds Bufo bufo bufo and its hydrolysis product dehydro- reported by Wieland et a1.lP2 can be satisfactorily bufotenin were investigated by H. Wieland and his interpreted in terms of the new structures. Thus the co-workers,lP2 who proposed structures (I; R = resistance of dehydrobufotenin to acid or alkaline SO,H and H respectively).We now propose struc- hydrolysis which was one of the main weaknesses of ture (11; R = 0.S03-)for bufothionin and (11; R = Wieland’s structure is readily understood as is the 0-)for dehydrobufotenin. fact that bufothionin crystallises unchanged from RO~JJ-=-cH=cH*NM~~ “WH 3.077 PH 342t H ct-The nuclear magnetic resonance spectrum of N-aqueous sodium hydroxide,l which points to a dehydrobufotenin “hydrochloride” (111) in D,O is in betaine structure. accord with the proposed structure. The observed Witkop4 has already concluded that structure (I) resonances are assigned as shown on formula (111). could not be correct from his observation that the The two benzene proton frequencies are split into ultraviolet spectrum of dehydrobufotenin hydro- two doublets (J = 9 cycles/sec.) one centred at 2.96 chloride corresponds to that of a simple 5-hydroxy- and the other at 3.427 as expected for adjacent indole and that the compound is not reduced by protons.It is noteworthy that this absorption is borohydride. incompatible with the presence of a proton in posi- Wieland and Widand’s bromination of bufo-tion 4 of the indole nucleus such a system with a thionin,2 with the subsequent hydrolyses can be re- proton pair in a 1,3-relation would be expected to interpreted as shown in the annexed sequence and show fine splitting (J = 2.5 cycles/sec.) as is ob-thus runs parallel to the bromination of lysergic served for 5-hydroxytryptamine oxalate. The third acid5 and of benzoyltryptophan.6 aromatic proton is at 3.077 which corresponds to an The two methylenes appear as The reaction which led Wieland to propose indole cu-pr~ton.~ asymmetric triplets the one adjacent to the indole structure (I) the platinum-catalysed hydrogenation ring is centred at 7-07~?and that adjacent to the of the salt (111) in dilute hydrochloric acid to bufo- positive nitrogen is centred at 6.537.The methyl tenin (5-hydroxy-NN-dimethyltryptamine),must be protons give a symmetrical sharp peak at 6.767 the looked upon as hydrogenolysis of the arylam-area covered by this peak corresponds to about monium C-N bond. 5.8 protons. The quaternary nature of the nitrogen The quaternary iodide C,,H,,N201 obtained by was confirmed by the fact that addition of NaOD to Wieland by exhaustive methylation of dehydro-the solution of the salt (111) in D20 caused no bufotenin in alcoholic alkoxide should now be T-appreciable shift of the methyl and methylene proton looked upon as the methiodide of base (IV; R resonances.Me) and the Hofmann degradation product H. Wieland and Vocke Annalen 1930,481 230. H. Wieland and T. Wieland. Annalen. 1937. 528. 234. Cohen Daly Kny and Witkop J. Amer. Chem.‘ SOC.,1960 82 2184. Witkop J. Amer. Chem. SOC. 1956 78 2878. Troxler and Hofmann Helv. Chim. Acta 1959 42 793. Patchornik Lawson and Witkop J. Amer. Chem. Sor. 1958 80 4748. AUGUST1961 311 (IV) (V) (VI) x-Reagents 1 Br,-H,O at 0". 2 H,O at 100". 3 Aq. HBr at 100". 4 Hot aq. Ba(OH),. 5 Hot aq. HBr. Cl3HISN20 as (IV; R = Me) although the the corresponding tetra-0-acetyl tertiary base.' analytical data2 fit Cl4HI80N2better which would The biogenesis of dehydrobufotenin may well correspond to structure (V).The product of vigorous parallel the oxidation of laudanosoline by chloranil acetylation of dehydrobufotenin would now be to the compound (VI),' proceeding by dehydro- expected to have structure (IV; R = Ac) this involves genation of bufotenin to a para-quinonoid structure the loss of one N-methyl group a reaction which is which cyclises by nucleophilic addition of the NMe paralleled by the conversion of a compound (VI) into group. (Received June 16th 1 96 1.) Robinson and Sugasawa J. 1932 789. Chain-transfer by Anisole in the Cationic Polymerisation of Isobutene By J. PENFOLD and P.H. PLESCH (UNIVERSITY COLLEGE OF NORTH STAFFS.) STAFFORDSHIRE KEELE SINCEthe discovery of chain transfer by toluene in but unknown level by evacuating the reaction the cationic polymerisation of styrene1 the pheno- vessel according to a strict routine. The concentra- menon has been investigated in detail for a wide tion of titanium tetrachloride was of the order of range of aromatic s~bstances,~~~ but only three other 3 mmole/l. that of isobutene 0.1 mole/l. Anisole monomers have been studied and that but cursorily dried over calcium hydride was distilled in vucuo cis-and trans-stilbene with t~luene,~ and 4-methoxy- from a burette into the reaction vessel. The reactions styrene with p-methylani~ole,~ were stopped at 10-20 % conversion by breaking We now report results on chain transfer by anisole the methanol phial.The polymers were isolated by in the polymerisation of isobutene by titanium tetra- steam-distillation and the degree of polymerisation chloride-water in methylene dichloride at -9" (DP) was determined viscometrically. -29" -44" -61" and -89". The apparatus and Infrared and ultraviolet spectra showed the technique have been described.6 The apparatus was polymers to contain (presumably terminal) charged with a phial containing a solution of titan- p-methoxyphenyl groups. The Mayo plots of I/DP ium tetrachloride in methylene dichloride (catalyst against [anisoleJ/ [i-C,HJ obtained at -89" and solution) and one containing methanol. The con- -61" were straight lines; those obtained at -#" centration of water was controlled at a reproducible -29" and -9" were straight up to [anisole]/[i-C,H,] Plesch J.1953 1659. Overberger et al.,J. Amer. Chem. Soc. 1953,75,6349 and subsequent papers ;Sakurada Higashimura and Okamura J. Polymer Sci.,1958 33,496; Higashimura and Okamura J. High-Polymer Chem. (Japan) 1956,13 397. Haas Kamath and Schuler J. Polymer Sci. 1957 24 85. Brackman and Plesch J. 1958 3563. Kamath and Haas J. Polymer Sci. 1957 24 143. Biddulph and Plesch Chem. andInd. 1959 1482; J. 1960 3913. about 0.2 1 and beyond that became concave to the 1/DP axis. From the straight portions of the plots values of the relative transfer coefficient kr/kpwere obtained in the usual manner. Within the experimental un- certainty (probably about f 15%) kr/kp did not vary significantly with temperature the mean value being 5.3 x This can be compared with 1.62 for styrene2 (SnC1,-H,O-Ph -N02-CC1 at 0").The main reason for the difference of three powers of ten between kr/kp for anisole with styrene and with isobutene is probably the difference between the k,'s. According to Pepper's' and Longworth and Plesch'ss results kp for styrene is of the order of 10 1.mole-l sec.-l whereas for isobutene recent esti- matesg indicate that it is of the order of lo4Lmole-l sec.-l. For both monomers k varies only slightly with temperature. Our results explain the observation3 that whereas styrene polymerising cationically in the presence of pre-formed poly-(4-methoxystyrene) will form grafts on this isobutene will not.The transfer reaction can be represented PROCEEDINGS kr RCH,.CMe,+ + PhOMe -+p-R.C,H,+.OMe -1C4H* R.C6H4*O+Me R.C,H,.OMe + C4H,+ The side-reaction forming the oxonium ion is sug- gested to account for the low yields found in the presence of anisole in reactions not stopped by methanol; the high stability of the oxonium ion would make its formation a kinetic termination step. The probable reason why low yields were not re- ported for the styrene experiments is that the con- centration of anisole was very much smaller and that of growing chains probably greater than in our experiments. The reason for the curvature of the Mayo plots obtained at the higher temperatures and the effects of anisole on the rate and yield of the polymerisa- tions are under investigation.We acknowledge the award of a U.S. Rubber Company Research Studentship (to J.P.) and a grant for apparatus from Polymer Corporation of Canada. (Received June 23rd 1961.) Pepper Symposium on Macromolecules Wiesbaden 1959,Paper 111 A.9. Longworth and Plesch Zuc. cit. Paper 111 A. 11. Plesch and Rutherford unpublished work. Dews Structures as Intermediates in the Synthesis of Benzene Derivatives By C. E. BERKOFF J. HUDEC, R. C. COOKSON and R. 0. WILLIAMS (THEUNIVERSITY, SOUTHAMPTON) ONEof the classical molecules of organic chemistry that has never been synthesised is Dewar's structure for benzene. As a possible route to such a ring- system,* the properties of which would be of great theoretical interest we considered the addition of a cyclobutadiene to an acetylene.By reduction of 3,4-dichloro- 1,2,3,4tetrarnethyl- cyclobutene (I) with lithium amalgam Criegee et aZ.l made the cis-dimer (111) of tetramethylcyclobutadi- ene and they later succeeded in catching the monomer (11) itself as the dichloronickel complex.2 When the dichloride (I) was reduced with lithium amalgam in but-2-yne only the cis-dimer (111) was formed. With activated zinc dust however hexa- methylbenzene was produced in 20% yield under conditions that gave less than 1 % in control experi- ments (butyne and zinc with zinc chloride equivalent doubt that four atoms of the benzene ring came from those of the cyclobutene (I) was dispelled when )Q:: +QR / R (1) w> -/+JqJ--&R I\ n\ (mi yq7[;-1'"'' to that produced from the butene).Any lingering (IV,' (V) * We refer to bicyclo[2,2,0]hexa-2,5-diene,non-planar and with a normal if strained central bond not to the fictitious excited state of benzene planar and with a long p-bond across the ring which can be considered a contributor to the valence-bond description of benzene. Criegee and Louis Chem. Ber. 1957,90,417;Criegee and Moschel ibid. 1959,92,2181. Criegee and Schroder Annalen 1959,623,1. AUGUST1961 reduction of the dichloride (I) with zinc in dimethyl acetylenedicarboxylate similarly produced in 18% yield dimethyl tetramethylphthalate (V; R = CO,Me) m.p. 128-129". The cis-dimer (111) added dimethyl acetylene&- carboxylate in boiling n-butyl acetate to give the Diels-Alder adduct (VI) m.p.97-98 ",probably formed through the valency-bond tautomer of (111) octamethyl[4,2,0]bicyclo-octatriene. At 330-350" the adduct decomposed to the same phthalic ester (V; R C0,Me). An authentic sample of the ester -L (V; R = C0,Me) was made from 3,4-dimethylhexa- 2,4-diene (separated from the mixture of hydro- carbons3 got by dehydration of 3 &dimethylhexane- 3,4-diol) by addition of maleic anhydride (adduct m.p. 95-96") esterification (dimethyl ester m.p. 51-52") and dehydrogenation with palladium-charcoal. It seems most likely that the benzene derivatives (V) arise by addition of the cyclobutadiene (11) to the acetylene and rapid rearrangement of the result- ing Dewar structure (IV).No isomer could be detected that might conceivably have come from an intermediate with Ladenburg's prism formula or from an adduct (VII) generated from a cyclobuta- diene with a diagonal bond.4 We acknowledge an I.C.I. Fellowship (to J.H.) and support by the U.S. Air Force (Geophysics Research Directorate A.R.D.C.) through its European Office. We also thank Dr. I. D. R. Stevens for helpful discussions. (Received,June 8th 1961.) McCullum and Whitby Trans. Roy. SOC.Canada 1928 22 39. Lipscomb Tetrahedron Letters 1959 No. 18 20. The Catalytic Hydrogenation of Succinic Anhydrides By R. MCCRINDLE and R. A. RAPHAEL K. H. OVERTON OF CHEMISTRY GLASGOW, (DEPARTMENT THEUNIVERSITY W.2) THEreduction of substituted succinic anhydrides by a variety of reducing agents both chemical and catalytic has been reported previously,l the product normally obtained being the lactone (-CO.OCO- -+ -CH,*OCO-).We have found that hydrogenation of substituted succinic anhydrides with a platinum catalyst can lead to hydroxy-lactones [-CO.OCO- + -CIQ(OH).OCO-] lactones and most sur-prisingly methyl acids (-CO.O-CO--+ -CH H,OC-) as major products the course of reaction depending on the compound solvent and time of reduction. The Diels-Alder adduct (I; R = 8,but double bond at position 4,5),2when hydrogenated at 1 atm. with Adams catalyst in ethyl acetate formed the hydroxy-lactone (I;R = H,OH) (> 90%) and the lactone (I; R = H,) (< 2 %).The constitution of the hydroxy-lactone followed from its stability to boiling methanol the disappearance of anhydride bands (1865 1780 cm.-l) and appearance of a y-lactone band (1770cm.-l) in the infrared spectrum and its smooth reconversion into the anhydride (I; R = 0) with one equivalent of chromic acid. In acetic acid further hydrogenation led to the lactone (I; R = H,) and the methyl acid (11) in comparable amounts. The lactone was recovered unchanged under the same conditions of hydrogenation but the hydroxy-lactone gave a mixture (2:l)of lactone and methyl acid showing that all three products arise from reduction of the same anhydride carbonyl group. That this was the group remote from the acetoxyl followed from basic hydrolysis of either the lactone or the methyl acid which with concomitant decarboxylation afforded the bridge-opened hydroxy-ketones (I11; R = OH or H respectively).Reduction of the other structurally related anhydrides (IV,)3 (V; R = H),4 (V; R = Me) and (VI)5 showed the above results to be general within these limits although the pro- portions of the three possible products varied from case to case. Succinic anhydride gave butyrolactone and butyric acid (4 1) even in ethyl acetate while cyclobutane-l,2-dicarboxylicanhydride6 was unique Fichter and Herbrand Ber. 1896,29 1193; Eijkman Chem. Zentr. 1907 I 1616; Sicher SipoS and Jon& CoZZ. Czech. Chem. Comrn. 1960,26,262; Granger and Techer Compt. rend. 1960,250 142. * Clauson-Kaas and Elming Acta Chem.Scand. 1952,6 560. Diels and Alder Ber. 1929 62 554. Diels and Alder Annalen 1928 460 98. Jenkins and Costello J. Amer. Chem. SOC.,1946 68 2733. Buchman Reims Skei and Schlatter J. Amer. Chem. SOC.,1942 64 2696. in giving the methyl acid as the only isolable product. Our results afford a new and simple route to 1,2-aldehydo-carboxylic acids (in the form of their cyclic tautomers) from the corresponding diacids. The observed formation under mild conditions of methyl acids from succinic (as distinct from phthalic) PROCEEDINGS anhydrides is of mechanistic interest and is being further studied. We thank the D.S.I.R. for a maintenance grant (to R.McC.). (Received June 13th 1961.) Orientation in Some Elimination Reactions in Cyclic Systems By T.H. BROWNLEE Jr. and W. H. SAUNDERS (CHEMISTRY THE UNIVERSITY ROCHE~TER U.S.A.) DEPARTMENT OF ROCHESTER 20 NEWYORK EARLEER reports that heating trimethylneomenthyl- ammonium hydroxide gives mainly menth-3-ene1p2 (the Saytzeff- rather than the Hofmann-rule product) have recently been confirmed by Hughes and Wilby.3 The main reaction is indeed of E2 type but sur- prisingly it displays under certain conditions a first- order component.5 The unusual nature of elimina- tion in cyclic systems is confirmed by our results on orientation in eliminations from trimethyl-2-methyl- cyclohexylammonium salts and the corresponding toluene-p-sulphonic esters. The percentages of 1-methylcyclohexene in the olefin fractions are recorded in the Table.For the cis-isomers elimination toward the methyl group is strongly favoured with the ammonium salts to an extent unchanged by variations in base and herence to the Saytzeff rule than does the toluenep- sulphonate normally a Saytzeff-rule leaving group. The trans-isomers show less stereo-specificity than expected for E2 reactions. Under normal Hofmann conditions the ammonium hydroxide does give only 3-methylcyclohexene in a trans-elimination but in dilute aqueous solution in a sealed tube appreciable amounts of 1-methylcyclohexene result. Analyses are also more scattered than with the cis-isomers average deviations of 3-5 % being usual. If both El and E2 processes contribute variations in concen-trations from one run to another could produce this effect.Indeed the reaction of the ammonium salt with pyridine appears to be entirely of El type, since the two isomers give the same product mixture. Further evidence that our reactions are on or near Orientation in elimination reactions of 2-methylcyclohexy i derivatives. 1-Methyl-1-Methyl-cyclohexened cy clohexened Reaction" (%> cis-R-NMe with HO-(conc.) in H20b 95 cis-R-NMe with EtO- in EtOH 94 cis-R-NMe with BuQ- in BuQH 97 cis-R-NMe in pyridine 96 trans-R-NMe with HO- (conc.) in H20b 0 trans-R-NMe with HO-(dil.) in H,OC 5 trans-RsNMe with EtO- in EtOH 1 trans-R-NMe with BuQ- in ButOH 8 trans-R-NMe in pyridine 94 Reaction (%I cis-ROTS with EtO- in EtOH 76 cis-ROTS with ButO- in BuQH 40 cis-ROTS in pyridine 93 trans-ROTS with EtO- in EtOH 8 trans-ROTS with ButO- in BufOH 4 trans-ROTS in pyridine 52 a Solutions were 0-024.20~ in alkoxide.Reaction mixtures were usually heated under in substrate and 0.20460~ reflux. An initially 0.02M-SOlUtiOn was concentrated by distillation until reaction occurred. The O-O2~-solution was heated at 180-220" in a sealed tube. As determined by gas chromatography on a 15 ft. "Carbowax" column. Ts = p-CaH,MeSO,. solvent. Average deviations in repeated runs were 2% or less except as mentioned below and in selected cases the products were found to be stable to the reaction conditions. Trimethylammonium a Hofmann-rule leaving group shows firmer ad-Hiickel Tappe and Legutke Annalen 1940,543 191.Read and Henry J. 1952 153 Hughes and Wilby J. 1960,4094. the E1-E2 borderline is provided by results for the 2-methylcyclopentyl system. Here the Hofmann reaction of the trans-ammonium hydroxide and the action of ethoxide ion and of pyridine on the trans- toluene-p-sulphonate give mixtures markedly less AUGUST1961 stereospecific than their counterparts in the 2-methyl- cyclohexyl series. The cis-isomers again give reactions obeying the Saytzeff rule. These results accord with the well-known greater reactivity of cyclopentyl derivatives in solvolytic reactions.* Since tetra-alkylammonium salts are not normally prone to El reaction a steric effect such as that sug-gested by Hughes and Wilby3 probably intervenes in cyclic systems.The strain produced by forcing the trimethylammonium group into the axial conforma- tion seems to loosen the C-N bond to such an extent that the system approaches and in some cases even attains the El transition state. Toluene-p-sulphonyl- oxy though normally a better leaving group is less strained in an axial conformation and hence shows less extreme orientation effects. Other explanations of course can account for at least some of our results. Among these are isomerisation prior to elimination the Wittig cyclic mechanism and the Winstein merged mechanism.6 We hope to obtain more definitive evidence by examining kinetics and isotope effects in selected cases. This work was supported in part by the National Science Foundation Washington D.C.(Received July 3rd 1961.) Brown Fletcher and Johannesen J. Amer. Chem. SOC.,1951,73,212; Roberts and Chambers ibid. p. 5034. Wittig and Polster Annalen 1956 599 13. a Winstein Darwish and Holness J. Amer. Chem. SOC.,1956,78 2915. The Preparation of Cohulupone from Colupulone by D. WRIGHT (BREWING RESEARCH NUTFIELD, INDUSTRY FOUNDATION SURREY) THEformulae of the hulupones,l a new group of bitter substances derived from hops have been established by a comparison of the spectral pro- perties of synthetic tetrahydrocohulupone2 with those of the natural hulupones. Tetrahydrocohulupone has been synthesis4 from hexahydrocol~pulone~ but the route employed is not easily applicable to the preparation of cohulupone itself (11) from co-l~puIone*~~ (I).When a solution of colupulone in cyclohexane is shaken in oxygen considerably more gas is absorbed (1.6 mol.) than by hexahydrocolupulone (1 mol.) and this may be due to participation of the olefinic bonds as the iodine value of the resinous product is substantially reduced. In these investigations sodium sulphite was used to reduce hydroperoxy-groups before the treatment with iodine monochIoride solu- tion but the resin then obtained did not have the ultraviolet absorption spectrum of the expected tetrasubstituted acylphloroglucinol. While sodium sulphite in methanol was found to Spetsig and Steninger J. Inst. Brewing 1960 65,413. Stevens and Wright Proc. Chem. Soc. 1960,417. Riedl and Nickl. Chern. Ber..1956. 89. 1863. Howard Pollock and Tatchell J.,' 1955 174. have no effect upon colupulone itself yet in the presence of oxygen rapid uptake (1 mol.) occurred and led to a mixture of resins from which cohulupone was isolated in about 30% yield. Synthetic cohulupone (XI) had b.p. 100" (bath)/lW4 mm. and the infrared absorption spectrum of natural cohu1upone;l iodine uptake indicated the presence of 2 double bonds and reaction with o-phenylene- diamine and 2,3-naphthylenediamine yielded crystal- line quinoxalines. The structure (11) was confirmed by converting cohulupone into the known tetrahydrocohulupone;2 hydrogenation of its non-crystalline sodium salt gave a product that had ultraviolet light absorption like that of oxyhumulinic acid2ys and was oxidised by bismuth trioxide in acetic acid to tetrahydro-cohulupone.The mechanism of this conversion of colupulone into cohulupone has not been established but oxidation with use of sodium sulphite is applicable to certain other a-alkyl-/3-diketones. Thus dihydro-1,1,4tri- methylresor~inol~ by this method gave rise to 1,l-dirnethylcy~lopentane-3,4dione;~ the other on hand the parent dimedone was unaffected. The interest of Dr. A. H. Cook F.R.S. is gratefully acknowledged. (Received Ma) 17th 1961.) Riedl Annalen 1954 585 38. Cook,Howard and Slater,J. Inst. Brewing 1955 61 321. Desai J. 1932 1079. Kon J. 1922 513. 316 PROCEED INGS Gas-chromatographic Study of the Rate of DiebAlder Addition By EMANUEL GXL-Avand YONAHERZBERG-MINZLY (THEDANIEL THE WEIZMANNINSTITUTE OF SCIENCE, SIEFF RESEARCH INSTITUTE REHOVOTH ISRAEL) BASSETTand HABGOOD~have shown that catalytic conversion occurring under conditions of chromato- graphic elution can be used for the determination of kinetic data.Continuing earlier gas-chromato-graphic work2 we have measured the rate of reaction of aliphatic dienes with chloromaleic anhydride as column liquid. The diene plug passing through the column (internal diameter 4 mm.; length 2m.) reacts with the dienophilic stationary phase (coated on 50-80 mesh Johns-Manville C-22 fire brick; 1 :2) to form a non-volatile addition product which is held back by the column. The residual amount of re actant emerging from the column is measured on the chromatogram and the conversion can be followed as a function of contact time by varying the rate of flow.The reaction is bimolecular but should become of pseudo-first-order in the chromatographic condi- tions owing to the large excess of the dienophil. If further diffusion is rapid in relation to addition the latter is rate-determining. The rate of addition at every section of the column through which the plug of diene is passing will be given by dx/dt = -kHx where x is the total amount of diene in the plug at contact time t and His a factor by which x has to be multiplied to give the amount in the liquid phase i.e. the part in contact with the dienophil. Since here the coefficient of partition of the solute between the gas and the liquid phase is almost independent of the concentration H is the same at every point of the column and a plot of log xo/x against contact time should give a straight line.The Figure shows this to be true for all five cases measured. The linearity of the plot for widely varying flow rates of carrier gas (4-1 10 ml./min.) proves that as assumed above the rate of addition is not diffusion-controlled. In practice instead of dctermining the absolute values of xo and x the change of the ratio of the reacting diene with respect to an inert reference substance was measured. Butane or heptane added to the diene and in some cases the unreactive cis-isomer accom- panying the trans-diene investigated served as the inert component.The following pseudo-first-order constants (sec.-l at 40") were calculated from the slopes (kH) of the curves hexa-1,3-dien-5-yne 1.42 x lo4 (H 0.98) butadiene 1-43 x (HO.67),isoprene 1-70 x (H 0.90),trans-penta-l,3-diene 6.30 x lo3 (H O-W) trans,trans-hexa-2,4-diene 2.12 x (H 0.94). These values show that electron-donating groups (methyl) increase the rate of addition whereas the electron-withdrawing ethynyl group decreases it in apreement with the literature? 00 Contact time (sec) Plot of log (x,/x) against contact time at 40". (A) trans,trans-hexa-2,4-diene;(B) trans-penta-l,3-diene; (C) isoprene; (D) butadiene; (E) trans-hexa-1,3-dien-5-yne. * Experimental data for higher contact times faIl outside the graph. Some Arrhenius frequency factors and the energy of activation for the addition reaction were cal-culated from the rate constants at 25" and 40"as follows trans-penta-1,3-diene log A = 7-1 sec.-l mole-l l.E = 17.2 kcal./mole; isoprene log A = 6-2sec.-l mole-l l. E = 15 kcal./mole. These values are in agreement with the literat~re.~ The procedure is rapid and requires only a few mg. of substance per experiment and impurities are removed while the reactant moves along the column. The method is obviously applicable to many other cases. The authors thank Dr. E. Fischer and Dr. F. S. Klein for valuable discussions. (Received,June 19th 1961.) Bassett and Habgood J. Phys. Chem. 1960,64 769. Gil-Av Herling and Shabtai J. Chromatog. 1958 1 508; Shabtai Herling and Gil-Av ibid.1959,2,406. * Ingold "Structure and Mechanism in Organic Chemistry," Cornell Univ. Press Ithaca 1953 p. 718. Ref. 3 p. 715 Table 45-1. AUGUST1961 NEWS AND ANNOUNCEMENTS Chemical Society Liaisoh Offiwrs.-The following Eellows have agreed to act as Liaison Officers Arthur D. Little Research Insti- tute Inveresk .. .. Dr. W.Banks College of Advanced Tech-nology Birmingham .. Dr. A. Holt Ciba (A.R.L.) Ltd. Duxford.. *Dr. G. Winfield Heriot-Watt College Edinburgh Professor F. Bell College of Technology Liver- pool .. .. .. .. Mr. C. B. F. Rice College of Technology Ports- mouth .. .. .. *Mr. J. W. Griffin * In place of Dr. R. Webb and Dr. J. L. Latham, respectively who have resigned as Liaison Officers.Corday-Morgan Medal and Prize for 1960.-This award consisting of a silver medal and a monetary prize of 400 guineas is made annually to the chemist of either sex and of British nationality who in the judgment of the Council of the Chemical Society has published during the year in question the most meritorious contribution to experimental chemistry and who has not at the date of publication attained the age of thirty-six years. Copies of the rules governing the award can be obtained from the General Secretary of the Society. Applications or recommendations in respect of the award for the year 1960 must be received not later than December 31st 1961 and applications for the award for 1961 are due before the end of 1962. Royal Society Awards.-Awards under the Royal Society and Nuffield Foundation Commonwealth Bursaries Scheme include the following Dr.J. N. Phillips of the C.S.I.R.O. division of plant industry Canberra to enable him to extend his physicochemical studies on plant and animal pig- ments at Imperial College London for about eleven months from December 1961. Dr. Muriel L. Tomlinson University Demonstrator in organic chemistry Oxford to enable her to study polydentate compounds at Sydney University for about three months early in 1962. Harrison Howe Award.-The Rochester Section of the American Chemical Society is to present the Harrison Howe Award for 1961 to Professor M. J. S. Dewar of the University of Chicago. The award is presented annually by the Rochester Section to an outstanding Chemist.Beilby Medal and Prize 1961.-The Admini-strators of the Sir George Beilby Memorial Fund representing the Royal Institute of Chemistry the Society of Chemical Industry and the Institute of Metals have decided to make awards from the Fund in 1961-each consisting of the newly instituted gold medal with a prize of la0 guineas-to the following To Constantin Edeleanu M.A. Ph.D. in recogni- tion of his work on the corrosion of metals and alloys with special reference to the development of the potentiostat technique and its applications to the study of practical problems and on the character- istics of corrosion reactions in fused salts. To Professor Jack Nutting M.A. Ph.D. in recognition of his work in physical metallurgy especially in the application of the electron micro- scope to the study of the relationship between micro- structure and mechanical properties of metals and alloys and to the investigation of phase changes and dislocation interactions.In accordance with the revised conditions the Beilby Medal and Prize will henceforth be offered at intervals of two years but more than one award may be made on the same occasion as in this year if there are several candidates of sufficiently outstand- ing merit. No further award will therefore be made until 1963. War Office Appointment.-The Secretary of State for War Mr. John Profumo has appointed Mr. E. E. Haddon Director of the Chemical Defence Experi- mental Establishment Porton in succession to Dr.E. A. Perren who is retiring. Mr. Haddon is at present Director of Chemical Defence Research and Development at the War Office. On taking up his new appointment on July 24th he was promoted to Chief Scientific Officer. Dr. Perren will continue to serve in the research laboratory at C.D.E.E. Chemistry in the Service of Medicine.-The Faculty of the History of Medicine and Pharmacy of the Society of Apothecaries of London has chosen “Chemistry in the Service of Medicine” as the theme of its second annual congress which is to be held in London on September 28th and 29th 1961. The Chemical Society is co-operating in this meeting and the last of the four sessions will be held in the Chemical Society’s rooms in Burlington House.Application forms for membership of the Congress and copies of the programme may be obtained from the Honorary Secretary Dr. F. N. L. Poynter The Wellcome Historical Medical Library Euston Road London N.W.l. International Congresses etc.-An International Conference on Magnetism and Crystallography including a symposium on Electron and Neutron Diffraction and sponsored by the International Union of Pure and Applied Physics and the Inter- national Union of Crystallography will be held in Kyoto on September 25-30th 1961. Further en- quiries should be addressed to Mr. T. Nagamiya. PROCEEDINGS Science Council of Japan Ueno Park Tokyo Japan. An International Heat Transfer Conference will be held in London on January 8-12th 1962.Further enquiries should be addressed to the Secretary Institution of Mechanical Engineers 1 Birdcage Walk Westminster London S.W. 1. A Symposium on Non-Stoicheiometric Com-pounds arranged by the Inorganic Chemistry Section of the American Chemical Society will be held in Washington D.C. on March 20-29th 1962. Enquiries should be addressed to Mr. Roland Ward Chairman of the Symposium University of Con- necticut Storrs Connecticut U.S.A. An international Conference on the Physics of Semiconductors arranged by the Institute of Physics and the Physical Society on behalf of the Inter- national Union of Pure and Applied Physics and the British National Committee for Physics will be held at the University of Exeter on July 16-2Oth 1962.Enquiries should be addressed to Administration Assistant The Institute of Physics and the Physical Society 47 Belgrave Square London S.W.l. The Ninth International Symposium on Com-bustion arranged by the Combustion Institute will be held at Cornell University on August 27th- September lst 1962. Enquiries should be addressed to Combustion Symposium office Upsom Hall Sibley School of Mechanical Engineering Cornell University Ithaca New York U.S.A. Election of New Fellows.-79 Candidates whose names were published in Proceedings for June have been elected to the Fellowship. Deaths.-We regret to announce the deaths of the following Sir Kariamanikkam Krishnan (4.6.61) Director of the National Physical Laboratory New Delhi and Dr.J. Murray (24.6.61) of the University of otago. Personal.-Professor D. H. R. Barton was the guest at the 14th Convention of the South African Chemical Institute in Durban in July and visited the Universities of Salisbury Pretoria Pietermaritzburg and Cape Town. Mr. A. I. Biggs formerly Director of Chemistry Federation of Malaya has recently taken up an appointment with the Federation of British In-dustries London. Dr. R. Bonnet? Assistant Professor in the Chem- istry Department of the University of British Columbia has been appointed to a Lectureship in organic chemistry at Queen Mary College Univer- sity of London. Dr. F. Brasco has been appointed Temporary Lecturer in Chemistry at the University of Sheffield. Dr. I. G. M.Campbell Senior Lecturer in the Department of Chemistry of the University of Southampton has been appointed Reader. Dr. T.A. Crabbhas been appointed I.C.I. Research Fellow in the Department of Chemistry of the University of Southampton. Professor A. G. Evans Dr. J. E. Garside and Mr. G. H. Moore have been appointed by the Minister of Education and the Secretary of State for Scotland to serve on the Standing Advisory Com- mittee on Grants to Students under the Chairman- ship of Sir Francis Hill. Dr. F. A. Filby Lecturer in Chemistry at the South-East Essex Technical College has been appointed Senior Lecturer from September Ist 1961. Dr. M. Gordon Senior Lecturer in Physical Chemistry at the Imperial College of Science and Technology has been appointed to the University Readership in Physical Chemistry.Sir Harold Hartley has been appointed an Honorary Member of the Institute of Fuel. The title of Reader has been conferred upon Dr. P. F. Holt Lecturer in Organic Chemistry University of Reading. Mr. H. Humphreys Jones principal from 1908-50 of the former Liverpool School of Pharmacy has received the Honorary Degree of M.A. of the University of Liverpool. Mr. K. A. R. Julian formerly Chief Project Engineer with Cremer and Warner has taken up an appointment with R. and J. Dempster Ltd. Dr. J. A. Kitchener Reader in Physical Chemistry at the Imperial College of Science and Technology has been appointed to the University Readership in the Science of Mineral Processing.Mr. H. E. Krumnz has been appointed Works Manager of the African Metals Corporation Cape Province. Mr. J. W.Ladbury who was awarded the 1960-61 Silver Medal of the Plastics Institute has recently been seconded to British Visqueen Limited from Imperial Chemical Industries Limited (Plastics Division). Mr. P.J. March has been appointed a Director of Shell Chemical Company. Dr. R. G. Neville has been appointed member of the Technical Staff at Aerospace Corporation Los Angeles California U.S.A. Dr. H. E. Nursten of the University of Leeds has been granted leave of absence to become a research associate in the Department of Nutrition Food Science and Technology Massachusetts Institute of Technology for one year from July 1st. He will also visit leading American Institutes dealing with higher education and/or research for the food and leather industries.Dr. D. C. Pepper Fellow of Trinity College Dublin and Reader in Physical Chemistry has been appointed to the newly established Chair of Physical Chemistry in Trinity College University of Dublin. AUGUST1961 Sir Rudolph Peters has received the degree of Honorary D.Sc. from the Australian National University at Canberra. Dr. H. J.Phelps has been appointed consultant to Maxam Power Limited and Goodyear Pumps Limited but will continue as an independent con-sultant to the De La Rue Company and as a director of Ets. Wyns-Bristol S.A. Belgium. Dr. Eva M. Philbin Senior Lecturer in Organic Chemistry has been appointed to the Chair of Organic Chemistry tenable in University College Dublin.Dr. F. A. Robinson at present Director of Research of Allen & Hanburys Ltd. will take over the Chairmanship of the Board of Directors of the Crookes Laboratories from September 1 st. The present chairman is Mr. N. B. Smiley of Arthur Guinness Son & Co. which acquired the Crookes Laboratories jointly with Philips Industries last November. 3 19 Dr. J. B. Stenlake Senior Lecturer in Pharma- ceutical Chemistry at the Royal College of Science and Technology Glasgow has been appointed to the Chair of Pharmacy at the College. Dv. F. G. A. Stone Assistant Professor in the Chemistry Department of Harvard University has been appointed to a Readership in Inorganic Chem- istry at Queen Mary College University of London.Dr. L. E. Sutton Demonstrator in Physical Chemistry University of Oxford has been awarded a Leverhulme Fellowship for studies on the sig- nificance of molecular shapes and sizes. Dr. G. Swann Assistant Managing Director of Beck Holler and Company (England) Limited has been appointed a Director of the parent company Reichhold Chemicals Limited. Dv. G. P. Wannigama of the University of Ceylon is to spend a year at the Laboratory of the Chemistry of Natural Products National Heart Institute Bethesda 14 Maryland U.S.A. FORTHCOMING SCIENTIFIC MEETINGS Hong Kong Wednesday September 13th 1961 at 7.30 p.m. Lecture “Polyacetylenes from Micro-organisms,” by Professor E.R. H. Jones D.Sc. F.R.S. in the Department of Chemistry University of Hong Kong. The meeting is being held in connection with the Golden Jubilee Congress of the University which includes a Symposium on Phytochemistry of which Professor Jones is Chairman. New ZeaIand Lecture “Stereospecificity of Enzyme Action,” by Professor N. L. Edson Ph.D. F.N.Z.I.C. F.R.S.N.Z. to be given as follows Wednesday September 13th at 8 p.m. in the Easterfield Lecture Theatre Victoria University of Wellington. Thursday September 14th at 8 p.ni. at the Massey Agricultural College Palmerston North. APPLICATlONS FOR FELLOWSHIP (Fellows wishing to lodge objections to the election of these candidates should comiiiunicate with the Honorary Secretaries within ten days of the publication of this issue of Proceedings.Such objections will be treated as confidential. The forms of application are available in the Rooms of the Society for inspection by Fellows.) Abbott David B.A. 28 High Street Eton Windsor Berkshire. Acheson Albert James Alexander B.Sc. Northern Bank House Aughnacloy Co. Tyrone Northern Ireland. Adams Margaret Joan. Somerville College Oxford. Amonoo-Neizer Eugene Hammond B.Sc. 35/9 Queens Gardens London W.2. Appleby Ian Colin B.Sc. 18 Hurstwood Avenue Albany Park Bexley Kent. Arrowsmith John B.A. 6 Old Lane Eccleston Park Prescot Lancs. Austin Peter William B.Sc. 8 Beeches Close Saffron Walden Essex. Betts James B.Sc. Ph.D. 7 Radford Close Rosslare Road Roseworth Stockton-on-Tees Co.Durham. Holton Peter Dennis B.Sc.. Ph.D. 18 Percy Circus London W.C.l. Boocock Geoffrey B.Sc. Ph.D. 1 Ripon Grove Sale Cheshire. Cabani Sergio. Istituto di Chimica Fisica. Via Risorgi- mento 19 bis Pisa Italy. Canipion Thomas Howard. 6 Badger Paddock Brock- field Park York. Carnighan Robert A.B. 469 Noyes Laboratory. Univer- sity of Illinois Urbana Illinois U.S.A. Chambers Derek William Sinclair M.Sc. 146 Middle Renwick Road Blenheim Marlborough New Zealand. Clarke Brian B.Sc. 48 Woodhead Road Abbey Hulton Stoke-on-Trent . Coburn John Francis B.S. 158 Foster Street New Haven Connecticut,U.S.A. Crudace Derek William B.Sc. 33 Shakespeare Avenue West Hartlepool Co. Durham. Daniels Peter John Lovell M-Sc. Ph.D.3 1 Albert Road Heaton Moor Stockport Cheshire. Downie Ian Morrison M.Sc. 9 Glenister Park Road London S.W.16. Evelegh Nigel Markham Aldridge B.Sc. 13A Merton Road Southsea Hampshire. Fleming John Scott B.S. Kedzie Chemical Laboratory Michigan State University East Lansing Michigan U.S.A. Fretwell Terence Stanley. 78 Chadwick Road Notting- ham. Gabali Edmond Nagib M.Pharm. Ph.D. Street 16, House 2 Meadi Egypt. Garnett John Lyndon M.Sc. Ph.D. 29 Arabella Street Longueville New South Wales Australia. Gartod Geoffrey M.A. D.Phi1. Clarendon Laboratory Parks Road Oxford. Gibbms William A. B.Sc. Department of chemistry University of Alberta Edmonton Alberta Canada. Gorton Earl Mahaffey B.S. M.S. School of Chemistry, Georgia Institute of Technology Atlanta 13 Georgia, U.S.A.Graham Garry George B.Sc. 120 Woniora Road South Hurstville New South Wales Australia. Grice Roy B.Sc. A.R.C.S. 36C Lewis Estate Warner Road London S.E.5. Habicht Ernst R. A.B. Department of Chemistry, Stanford University Stanford California U.S.A. Heathcock Clayton B.Sc. 2224 Arapahoe Boulder Colorado U.S.A. Henderson Ian. Wootton Hall Fallowfield Manchester 14. Hewins Michael. 44 Richards Road Sheffie!d 2. Hill Ralph M. Ph.D. Esso Research Ltd. 50 Stratton Street London W.1. Hopkins Clarence Yardley M.A. Ph.D. 180 Carleton Road Rockcliffe Park Ottawa Canada. Horobin Richard William B.Sc. 2 Nicholls Street West Bromwich Staffordshire. Howard Bryan Charles BSc. 32 St. Johns Road Lowestoft Suffolk.Jones Neville M.Sc. A.R.I.C. 94 The Chesils Styve- chale Coventry. Katsanos Nicholas B.Sc. 57 Benibridge Road Leicester. Kennedy Anna B.Sc. 16 Dale Green Road London N.ll. Khalique Abdul M.Sc. Ph.D. East Regional Labora- tories P.C.S.I.R. Dacca Pakistan. Khwaja Tasneem Afzal M.Sc. University Chemical Laboratories Lensfield Road Cambridge. Kidwai Akhlaq R. M.S. Ph.D. Departnicnt of Chern-istry Aligarh Muslim University Aligarh India. Kumar Manniohan. 142 The Hides Harlow Essex. Lawrie Gordon William B.A. I The Cottages Hall Lane Wrightington nr. Wigan. Lea Leonard John. 29 Melford Avenue Barking Essex. Levy Edmond B.Sc. Weizmann Institute of Science Rehovoth Israel. McNae Colin James B.Sc. 6 Pagoda Avenue Richmond Surrey.Mann Ronald B.A 21 Berkley Road Gravesend Kent. Martin Patricia Ann B.Sc. 103 Barnsole Road Gilling- ham Kent. Middleton Eric John B.Sc. Chemistry Department University of Western Australia Nedlands West Australia . Miller Kenneth James M.Sc. 125 Osborne Road Jesmond Newcastle-upon-Tyne 2. Moore William Archibald John. 9 Morten Close Clarence Avenue London S.W.4. Naeser Charles Rudolph M.S. Ph.D. Department of Chemistry The George Washington University, Washington 6 D.C. U.S.A. Northmore Jack McKeer 76 Courtlands Avenue, Eltharn Road London S.E.12. Parr Alan Rodney. 28 Higher Green Ewell Epsoni Surrey. Perciaccante Vincent Anthony B.S. M.S. 31-36-41 Street Long Island City 3 New York City N.Y. U.S.A. Pitchfork Ernest David.30 Shadwell Gardens Watney Street London E.1. Podimuang Verapong B.Sc. Department of Organic Chemistry The University Liverpool. Potty Dasan N. M.Sc. Navaratna Pharmaceutical Laboratories P.B. No. 13 Cochin-2 India. Quinn Alan George. 7 Ripley Gardens London S.W.14. Ramachandran S. B.Sc. Ph.D. Department of Chem- istry Ohio State University Columbus 10 Ohio, U.S.A. Randall William John B.S. SO94 East John Street Champaign Illinois U.S.A. Reid William Kenneth B.Sc. 32 Walter Street Glasgow E.l. Rosich Ronald Steven B.Sc. 43 Great Eastern Highway Rivervale West Australia. Sahasrabudhey Ramachandra Hari Ph.D. D.Sc., F.R.I.C. Banares Hindu University Varanasi 5 India. Serafin Frank George B.A. 67 Hall Avenue Sonierville Massachusetts U.S.A.Shaw Stanley James B.A. Chemistry Department Trinity College Dublin. Taylor Max Ronald M.Sc. School of Chemistry Uni- versity of Sydney New South Wales Australia. Thomson Thomas Jardine B.Sc. 26 Jedburgh Road Dundee Scotland. Tress Roy. B.Sc. 40 Rosslyn Crescent Rentley, Doncas ter. Wallace Robert B.Sc. c/o Cochran 16 McKenzie Strcet Paisley Renfrewshire Scotland. Walton David Ralph h4iles M.Sc. A.R.I.C. 15 Port-land Road Clarendon Park Leicester. Willett John Edward B.Sc. 14 Mont le Grand Excter Devon. Williamson Robert Humphrey B.Sc. A.R.I.C. 59 Eastwick Park Avenue Great Bookham Surrey. Winfield Graham Ph.D. Prestcotian Rectory Estate Harlton Nr. Cambridge. Yeadon Alan BSc. 3 Lands Lane Guiseley Leeds.
ISSN:0369-8718
DOI:10.1039/PS9610000273
出版商:RSC
年代:1961
数据来源: RSC
|
|