摘要:

Chemical Society Reviews Vol 3 No 2 1974 Page The Photochemistry of the Uranyl Ion 139By H. D. Burrows and T. J. Kemp Non-conventional Electrophilic Aromatic Substitutions and Related Reactions By S. R. Hartshorn 167 Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes By D. V. Fenby and L. G. Hepler 193 Chemistry of the Production of Organic Isocyanates By H. J. Twitchett 209 Metalloboranes and Metal-Boron Bonding By Norman N. Greenwood and Ian M. Ward 23 1 The Chemical Society London Chemical Society Reviews Chemical Society Reviews appears quarterly and comprises approximately 25 articles (ca. 600 pp) per annum. It is intended that each review article shall be of interest to chemists in general, and not merely to those with a specialist interest in the subject under review. The articles range over the whole of chemistry and its interfaces with other disciplines.Although the majority of articles are specially commissioned, the Society is always prepared to consider oEers of articles for publication. In such cases a short synopsis, rather than the completed article, should be submitted to The Editor, Reports and Reviews Section, The Chemical Society, Burlington House, Piccadilly, London, W1V OBN. Members of The Chemical Society may subscribe to Chemical Society Reviews at €3-00 per mum; they should place their orders on their Annual Subscription renewal forms in the usual way. Non-members may order Chemical Soc;iety Reviews for f10.00 per mum (remittance with order) from: The Publications Sales Officer, The Chemical Society, Blackhorse Road, Letchworth, Herts., SG6 lHN, England. 0Copyright reserved by The Chemical Society 1974 Published by The Chemical Society, Burlington House, London, W1V OBN Printed in England by Eyre & Spottiswoode Ltd, Thanet Press,Margate

ISSN:0306-0012    

DOI:10.1039/CS97403FP003

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS97403FX005

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS97403BX007

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

The Photochemistry of the Uranyl Ion By H. D. Burrows THE CHEMICAL LABORATORY, UNIVERSITY OF COIMBRA, PORTUGAL T. J. Kemp DEPARTMENT OF MOLECULAR SCIENCES, UNIVERSITY OF WARWICK, COVENTRY, CV4 7AL, WARWICKSHIRE 1 Introduction Whilst historically predating organic photochemistry, only in the last decade has inorganic photochemistry received comparable study of the nature of the primary photochemical and photophysical processes involved.1 The important photo- induced reactions in inorganic systems (particularly those involving metal ions) are redox reactions, ligand substitutions, and molecular rearrangemenkl The uranyl ion presents a model species for photochemical study as, in addition to the high photosensitivity of its many complexes,2~3 the excited uranyl ion luminesces, thus providing an additional tool for the study of the excited-state processes involved.Nevertheless, the two most recent reviews on the photo- chemistry of this species2J both indicate that many of the results are intricate and confusing, and that our understanding of the primary processes involved is far from satisfactory. In large part, these problems stem from both a lack of detailed understanding of the spectroscopy of the uranyl ion, and a lack of application of fast-reaction techniques to the study of mechanisms. We feel now, however, that this situation has been remedied sufficiently to enable a more detailed but consistent picture of the photochemistry of uranium(v1) to be presented. Photochemical studies on uranium(w) date back to the beginning of the nineteenth century when uranyl oxalate was found to decompose under the influence of light.4 Detailed investigation of the photochemistry received con- siderable impetus both from the discovery that uranyl oxalate provides a suitable actinometer5 and, much later, from the Manhattan project and later related See, for example, A.W. Adamson, W. L. Waltz, E. Zinato, D. W. Watts, P. D. Fieischauer, and R. D. Lindholm, Chem. Rev., 1968, 68, 541; W. L. Waltz and R. G. Sutherland, Chem. SOC.Rev., 1972,1,241; or ref. 3. * E. Rabinowitch and R. L. Belford, ‘Spectroscopy and Photochemistry of Uranyl Com- V. Balzani and V. Carassiti, ‘Photochemistry of Co-ordination Compounds’, Academic pounds’, Pergamon Press, Oxford, 1964.Press, London, 1970. A. Buchholz, Ann. Chirn. Phys., 1805, 56, 142, quoted in ref. 3. See, for example, C. R. Masson, V. Bockelheide, and W. A. Noyes, jun., ‘Photochemical Reactions’ in ‘Techniques of Organic Chemistry, Vol. II’, ed. A. Weissberger, Interscience, New York, 2nd edn., 1956, p. 294, The Photochemistry of the Uranyl Ion studies, where one of the objectives was to determine methods of separating pure uranium and obtain isotopic enrichment. Here we will concentrate on those aspects of the subject which have developed since the reviews of Balzani and Carassiti,3 and Rabinowitch and Belford,2 especially those which have led to a clarification of the primary processes in- volved, and of the mechanistic pathways leading to stable final products.How-ever, where appropriate, references are included from these earlier reviews. 2 Structure and Bonding The dominating structural factor in uranyl compounds is the 0-U-O system, which is generally thought to be linear both in the solid statea and in s~lution.~ Around this are co-ordinated a further 4, 5, or 6 ligands; e.g. in the crystal structure of uranyl nitrate dihydrate8 (Figure 1) the metal atom exhibits an 0' U--05=1.770 U-O'= 1.749 05U06=179.1' Figure 1 Structure of UOz(N03)2,2Hz0from ref. 8 (05 and 06, the atoms in the U02 entity, are directly above and below the U atom and all other atoms are approximately coplanar) (Reproduced by permission from Acta Cryst., 1965, 19, 536) overall co-ordination number of 8, with two water molecules and two nitrate ligands co-ordinated in the plane perpendicular to the 0-U-0 axis.Further, the U-0 bonds to the ligands are always longer than those in the U02 moiety, e.g. in the uranyl nitrate complex, U-0 (nitrate) = 2.504, 2.547 A, U-O (water) = 2.397 A, U-0 (uranyl) = 1.749, 1.770 A. The metal-ligand bonds are thought For example, J. E. Fleming and H. Lynton, Chem. and Ind., 1960,1415. S. P. McGlynn and J. K. Smith, J. MoZ. Spectroscopy, 1961, 6, 164 (see also ref. 15).* J. C. Taylor and M. H. Mueller, Acru Cryst., 1965,19,536; N. K. Dalley, M. H. Mueller, and S. H. Simonsen, Inorg. Chem., 1971,10,323. Burrows and Kemp to be covalent in character rather than purely electrostatic,g so the short U-O bond distances in the uranyl species provide good evidence for involvement of n-bonding in this group.A number of theoretical treatment~~JO-1~ have been performed on U0z2+ which support the existence of wbonding. The electronic configuration7 of U is 5f36d17s2. In U022+ the 6d-orbitals of uranium are thought to be involved both in 0-and n-bonding with the 2p-orbitals of the 0 atoms.7912-15 Evidence for .f-orbital involvement is less clear. Calculations show only limited 5f-orbital involvement12-14 but Coulson and Lesterll have demonstrated that 6f-orbitals may be involved, at least in the bonding to the other ligands. Although the ac- tual ordering of the molecular orbitals has not been settled,7J3J4 theoretical calcu- lations on uranyl, assuming Dmh symmetry, suggest that the ground state is a singlet state ('Ig+) with electronic configuration7 (lau+)2 (la$)2 (l7~~)4(17~~)~.Magnetic measurements confirm that uranyl complexes are predominantly diamagnetic, with a weak temperature-independent paramagnetism.lO The magnetic properties are discussed in more detail by Eisenstein and PrycelO and by McGlynn and Smith.7 Whilst spectroscopic studiesl6J7 of uranyl complexes indicate that the symmetry is often different from Dmh, this does not significantly affect the above picture. Further, whilst uranyl forms complexes with a wide range of organic and inorganic ligands, and also hydrolyses in a slow, pH-dependent reaction to give polymeric species,ls the photochemical properties in both cases can be con- sidered to be predominantly those of the UOz2+ moiety.In the hydrolysed species, the U02 system probably remains intact, with bonding between adjacent units via hydroxy bridges.19 Fuller details on the structure of uranyl complexes are given elsewhere.20p21 3 Spectroscopy and Photophysics A. Absorption Spectrum.-The electronic absorption spectrum of the uranyl ion is probably one of the most extensively studied regions of molecular spectroscopy, yet our understanding of it remains far from complete. The spectrum shows a For a discussion, see, for example, N. W. Alcock, J.C.S. Dalton, 1973, 1616. lo J. C. Eisenstein and M. H. L. Pryce, Proc. Roy. SOC.,1955, A229,20. l1 C. A. Coulson and G.R. Lester, J. Chem. SOC.,1956, 3650. la R. L. Belford, J. Chem. Phys., 1961, 34, 318. l3 R. L. Belford and G. Belford, J. Chem. Phys., 1961, 34, 1330. l4J. B. Newman, J. Chem. Phys., 1965, 43, 1691. Is S. P. McGlynn, J. K. Smith, and W. C. Neely, J. Chem. Phys., 1961,35, 105. l6 C. Gorller-Walrand and S. De Jaegere, J. Chim. Phys., 1972, 69, 726. l7 C. Gorller-Walrand and S. De Jaegere, Spectrochim. Acta, 1972, A28,257. See, for example, H. S. Dunsmore, S. Hietanen, and L. G. Sillen, Acta Chem. Scand., 1963, 17,2644 and other papers in this series. lrnJ. T. Bell and R. E. Biggers, J. Mol. Spectroscopy,1967, 22,262. ao A. E. Comyns, Chem. Rev., 1960,60,115. I1 I. I. Chernayev, 'Complex Compounds of Uranium', Israel Program for Scientific Trans- lations, Jerusalem, 1966.141 The Photochemistry of the Uranyl Ion remarkable sensitivity to its environment and to temperat~re.2~-~4However, the main spectral features are typified by uranyl perchlorate in wate1-2~ (Figure 2) where the absorbing species is probably [UO2(H2O)sl2+. These consist of an initial series of weak bands between 500 and 360 nm, with progressively stronger bands towards the U.V. 19.0 L 15.2 -II) d) 11.4 -2 -2 &nj-2 7.6 --3.8 --L340 380 420 460 500 Wavelength Alnm Figure 2 Absorption spectrum of uranyl perchlorate in aqueous HC104. Filled circles-experimental spectrum ofuO2(c104)2 (9.15 mM, [HC104] = 1.4 mM); full line joining circles-computer simulated spectrum ; separate curves-com-ponents of simulated spectrum (Numbering scheme due to Bell and Biggers, ref.24) (Reproduced by permission from J. Mol. Spectroscopy, 1965, 18, 247) aa G. N. Dieke and A. B. F. Duncan, ‘Spectroscopic Properties of Uranium Compounds’ McGraw-Hill, New York, 1949. 23 B. Jezowska-Trzebiatowska and A. Bartecki, Spectrochirn. Ada, 1962, 18, 799 and other papers in this series. I4 J. T. Bell and R. E. Biggers, J. Mol. Spectroscopy, 1965, 18, 247. Burrows and Kemp Jorgensen25 suggested that the weak visible bands are probably Laporte- forbidden, and arise from a charge-transfer transition of an electron from a ligand orbital to an empty 5f-orbital on the central U atom. However, most other workers interpret the transition as being completely within the U0z2+ gr0uping.~J6 Following the theoretical demonstration of the importance of n--bonding in uranyl, McGlynn and Smith7 assigned the lowest-energy transition to excitation of an electron from the highest filled morbital to a non-bonding orbital on the uranium, and this interpretation is generally accepted.Such an excited state may be expected to possess certain free-radical character, and in fact hydrogen atom abstraction by the excited uranyl ion is observed exper- imentally (see Section 4B). Assignment of the electronic states involved in the absorption spectrum is, however, rather more controversial. McGlynn and Smith7 suggested that the lowest-energy absorption between 20 500 and 30 000 cm-1 results from a triplet -+ singlet transition to either 3nUor 34, states (with a preference for the former).Because of the presence of significant spin-orbit coupling, this transition is reasonably intense (emaxN 101mol-1 cm-1) even though it is spin-forbidden. They argued that these bands could be divided into three groups with origins at 22 050,24 125, and 27 OOO cm-1 representing the three components of the triplet state which is split by the local field. Further vibronic fine structure is super- imposed on these absorptions. McGlynn and Smith’s assignment of the lowest excited state to a triplet state stems primarily from theoretical arguments. Sup- port for this assignment has been proposed on the basis of magnetic circular dichroism (m.c.d.) studies on (Bu4N)U02(N03)3 in a polymer matrix at 10 K,27 from studies of the effect of complexation on the absorption spectrum of uranyl sol~tions,~~~~~-~~ In a particularly and from further theoretical analyse~.~~#31-33 detailed study, Bell and Bigger~19~~4~3~~32 showed that the complicated uranyl absorption could be resolved mathematically using a non-linear least-squares computer program into a series of 24 Gaussian bands in 7 groups (Figure 2).Following McGlynn and Smith’s model,7 they assigned the first 13 bands to the three components of a triplet state 3nuat 21 329, 24 107, and 27 731 cm-1. Brint and McCafferyz7 have argued that at least in the case of (Bu4N)U02(N03)3 the lowest-energy transition, which is parity-forbidden, is to two close-lying triplet states (3Elg and 3E~gin their nomenclature).Alternative assignments of the low-energy absorptions have, however, been gb C. K. Jargensen, Acta Chem. Scand., 1957, 11, 165. See, for example, Z. Lib&, J. Inorg. Nuclear Chem., 1962, 24,619. !2’ P. Brint and A. J. McCaffery, Mol. Phys., 1973, 25, 31 1. M. M. Fayt, E. M. Vandeput, and S. De Jaegere, Nature, 1963,200, 355. S. De Jaegere and T. Govers, Nature, 1965, 205, 900. 30 V. Baran, Coll. Czech. Chem. Comm., 1966, 31, 2093. a1 J. T. Bell and R. E. Biggers, J, Mol. Spectroscopy, 1968,25,312. 32 J. T. Bell, J. Inorg. Nuclear Chem., 1969, 31, 703. 3s Y. J. Israeli, Bull. SOC.chim. France, 1965, 196. The Photochemistry of the UranylIon propo~ed.~~-~~ have investigated uranyl solutions Volod’ko and co-~orkers~~ and crystals at low temperatures and concluded that the absorption spectrum has four series of lines in the region 330-500 nm with substantially different characteristics which arise from transitions to four separate singlet levels, and have given an energy state diagram based on this model.Similarly, Gorller- Walrand and Vanqui~kenborne~~9~3 favour interpretation of the low-energy region in terms of singlet -singlet transitions because of the effect of ligands on the uranyl spectrum. They find that the spectra of a large number of uranyl complexes are virtually independent of the chemical nature of the ligand, but are strongly dependent on the total symmetry of the complex. Siddall and Prohaska35 have also questioned the assignment of the low-energy bands to a triplet +-singlet transition, following a study of solvent effects on spectra of organophosphorus and organonitrogen adducts of uranyl.The evidence is not strong enough to allow an unambiguous assignment of the lowest excited state but we favour the triplet assignment. In support, McGlynn and Smith7 noted that a weak paramagnetism (different from the permanent, temperature-independent paramagnetism) could be detected in an excited uranyl complex. Unfortunately, no further details on this have been forthcoming. Similar uncertainty exists in the assignment of the upper excited states. McGlynn and Smith7 assigned the excited states at 34 000 and 48 OOO cm-1 (2nd and 3rd excited states in their classification) to the lowest excited singlet states.Israeli33 assigned states at 29 000, 34 500, 42 600, and 53 000 cm-l to the singlet states Idu,Inu, I&+, and ln,. Volod’ko and co-workers3* have indicated the difficulty of assigning states higher than 31 OOO cm-l in uranyl salt solutions where the counter-ion is nitrate, sulphate, or any other anion possessing absorptions in this region. They assign the excited states between 20 000 and 31 OOO cm-1 to four singlet states. Brint and M~Caffery,~~ however, on the basis of low- temperature absorption and m.c.d. studies assign the absorption between 29 OOO and 36 000 cm-1 to the parity-allowed T +-S transition 3Elu +-lAlg. Bell and Biggers24JlJZ overcame the problem of overlapping counter-ion absorption by using uranyl perchlorate, and produced possibly the most comprehensive analysis of the spectrum.They assigned the state at 31 367 cm-l to the second triplet 34 A. N. Sevchenko, Analele Stiint. Univ., ‘A. I. Cuza’, Iasi Sect. 1, 1961,7, 121 ;Chem. Abs., l963,59,2300e. 36 T. H. Siddall and C. A. Prohaska, Nature, 1963, 202, 1088. 38 A. N. Sevchenko, Izvest. Fiz. Inst. ANEB, Bulg. Akad. Nauk, 1968, 17, 69; Chem. Abs., 1968, 69, 81974~. 37 L. V. Volod’ko, A. N. Sevchenko, and D. S. Umreiko, Doklady Akad. Nauk S.S.S.R., 1967,172,1303;Chem. Abs., 1967,66,109 903k. 3a L. V. Volod’ko, A. I. Komyak, and L. E. Sleptsov, Optics and Spectroscopy, 1967,23, 397. sD D. S. Umreiko, A. N. Sevchenko, and G.G. Novitskii, Doklady Akad. Nauk Beloruss. S.S.R., 1968,12, 884; Chem. Abs., 1969, 70’42 522x. 40 A. I. Komyak and L. V. Volod’ko, Vesti Akad. Navuk Belorusf. S.S.R.Ser. Fiz-Mat. Navuk, 1969, 115; Chem. Abs., 1969, 71, 17311r. *l D. N. Sanwal and D. D. Pant, Proc. Indian Acad. Sci., Sect. A, 1969, 69,324; Chem.A&., 1970, 72 7520q. C. Gorller-Walrand and L. G. Vanquickenborne, J. Chem. Phys., 1971,54,4178. 4a C. Gorller-Walrand and L. G. Vanquickenbome, J. Chem. Phys., 1972,57! 1436. Burrows and Kemp state, and resolved five further excited states which, whilst possessing steadily increasing oscillator strengths, were separated by an approximately constant energy gap of 6015 k 39 cm-l. Two groups have observed absorptions in the uranyl spectrum at energies less than 20 OOO cm-1.Be1144 reported very weak bands at 507.7 and 531 .O nm, which he assigned to hot bands involving transitions from the first vibrational level of the ground state to the lowest excited state. De Jaegere and co-workersZ8 found a band at 555 nm in the absorption spectrum of uranyl nitrate solution under 110 kG cm-2 of oxygen. This, however, is believed to result from a charge-transfer transition in a uranyl-02 complex. In the presence of strongly complexing species the uranyl spectrum is signi- ficantly modified.l6117,24,42,43145However, the changes are predominantly in the intensities and bandwidths of the absorptions, and only slightly, if at all, in the actual positions of the bands, Such a change has been interpreted in terms of a variation of the uranyl symmetry,16J7 and a good correlation has been observed between the vibronic fine structure and the geometry of the UO2 co-ordination.In addition to these transitions within the U0z2+ grouping (probably involving transitions of n-electrons to progressively higher orbitals on the U atom) there is some evidence for charge-transfer-to-metal (CTTM) transitions of fairly low energy from the surrounding ligands to the central metal atom.46147 In particular, a study by Barnes and Day47 showed that with U022+(H~O)n-1X complexes (X = Br-, NCS-, S2032-, SO$-, H20) there are bands between ca. 30 000 and 50 OOO cm-1 which, from a comparison of similar bands in cobalt complexes, are almost certainly of a charge-transfer character, As commented by Balzani and Carassiti,3 similar CTTM transitions should be observed in complexes with organic ligands.Indeed, such transitions appear in the uranyl chloride-tributyl phosphate complex,l7 although it is not clear whether the charge transfer is from the chloride or organic part. Similarly, Heidt and colleagues have suggested that CTTM bands are present in oxalatouranium(v1) ~pecies.~8 B. Luminescence.-Uranyl salts exhibit a highly characteristic luminescence both in the solid state and in solution. The luminescence, which is assigned to emission from the lowest excited 3nustate only24 (and is therefore strictly an example of phosphorescence, although the majority of workers adhere to the description ‘fluorescence’), displays considerable vibrational structure.Bell and BiggersZ4 resolved six bands in the luminescence spectrum of uranyl perchlorate in water (see Figure 4). As with the absorption spectrum, the shape and intensity of the emission spectrum are very sensitive to the particular environment of the UO$+ species. Volod’ko and co-workers found that the luminescence decay of uranyl salts in various organic solvents at low temperatures could be divided into three 44 J. T. Bell, J. Mol. Spectroscopy, 1972, 41, 409. 46 R. L. Belford, A. G. Martell, and M. Calvin, J. Inorg. Nuclear Chem., 1960, 14, 169. 46 J. L. Ryan and C. K. Jerrgensen, Mol. Phys., 1963, 7, 17; J. H. Miles, J. Znorg. Nuclear Chem., 1965,27, 1595. O7 J. C.Barnes and P. Day, J. Chem. Soc., 1964, 3886, *.3 L. J. Heidt, G. W. Tregay, and F. A. Middleton, jun., J. Phys. Chem., 1970, 74, 1876. Tlie Photochemistry of the Uranyl Ion first-order components of species with different emission spectra, assigned to different uranyl complexes.49 In suitable cases the luminescence quantum yield (#F) is unity,2$50 but in solution #F decreases markedly, possibly being influenced both by physical effects (as evinced by the decrease in +F with decreasing viscosity) and by chemical reactions of the excited state. The luminescence of uranyl salts, either in solution or in the solid state, is affected both in intensity and in lifetime by deuteriation of the ligands; for ex- ample, the replacement of H2O by D2O in crystals of UOzS04,3H20 increases the luminescence lifetime,50 although with KdJOz(S04)2,2H20, deuteriation has no effect (however,+p is unity in this case). Similarly, the luminescence yield and lifetime of U022+in D2O solution is 1.7-2 times greater than in H20.51152The reason for this isotope effect is not at present clear.Three possi- bilities exist: (i) Intramolecular radiationless deactivation of the excited state via the O-H stretch modes of co-ordinated water molecules. The lower frequency of the 0-D stretch vibration compared with the O-H stretch allows less efficient deactivation. Similar isotope effects have been observed in radiationless processes in organic molecules.53 (ii) Intermolecular energy transfer from excited uranyl ion to vibrational modes of solvent water molecules.Thus, with the lowest state of uranyl being at 21 000 cm-l, this could energy transfer to the sixth overtone of the O-H stretch, but only to the eighth or ninth overtone of the 0-D stretch vibration. Similar intermolecular energy transfer to solvent vibra- tions has been proposed to account for solvent effects on the fluorescence lifetimes of excited rare-earth i0ns,~4 and singlet 0xygen.~5 However, the energy of the excited state involved is considerably lower in these cases, so that the process would be expected to be more significant than with U022+ ion. (iii) Chemical deactivation of excited uranyl by hydrogen abstraction from water, i.e. equation (1): [a similar oxidation has been suggested in the photochemistry of aqueous 49 L.V. Volod’ko and E. A. Turetskaya, Zhur. priklad. Spectroskopii, 1965, 3, 248; Chem. Abs., 1966, 64, 7548c. D. D. Pant, D. N. Pande, and H. C. Pant, Indian J. Pure Appl. Phys., 1966,4,289; Chem. Abs., 1966, 65, 179268; see also D. D. Pant and H. B. Tripathi, Indian J. Pure Appl. Phys., 1969, 7, 140; Chem. Abs., 1969, 70, 92050n; and D. D. Pant and H. C. Pant, Indian J. Pure Appl. Phys., 1968,6,219; Chem. Abs., l968,69,63314a. 61 J. L. Kropp, J. Chem. Phys., 1967,46,843. 62 R. J. Hill, T. J. Kemp, and (in part) D. M. Allen and A. Cox, J.C.S. Faraday I, 1974, 70, 847. 6s See, for example, J. B. Birks, ‘Photophysics of Aromatic Molecules’, Wiley, London, 1970. 64 See, for example, Y.Haas, and G.Stein, J. Phys. Chem., 1971,75, 3668. 66 P. B. Merkel and D. R. Kearns,J. Amer. Chem. Soc., 1972,94,7244. Burrows and Kemp cerium(rv) solutions].56 One would expect a primary kinetic isotope effect on equation (1). No e.s.r. signal attributed to OH is observed on pho- tolysis of uranyl perchlorate in an aqueous glass. However, under the influence of light, uranyl ion has been shown to exchange its oxygen atoms57 with H20,1* probably via production of the catalytic UV. It is not at present possible to say definitely which of these explanations is responsible for the observed isotope effect on IF1.Detailed consideration of (i) and (ii) must await theoretical study on radiationless processes in inorganic complexes. It seems likely that, at least in the solid state, the intramolecular deactivation process (i) is involved.In solution, however, a chemical process of type (1) could easily occur. Similar hydrogen abstraction from water by the triplet states of benz~phenone~~ has recently been reported and may and q~inones~~ be important in photosynthesis. The failure to observe OH radicals may merely imply rapid back reaction between 6H and the UV species in the solvent cage; indeed, radiation chemical studies indicate that the reaction between UV and 6H is rapid.60 Whereas in water and certain other solvents the decay of the uranyl lumi- nescence follows good first-order kinetics,2 in certain cases, particularly in crystals under high light intensities, non-exponential decay of the luminescence is observed.61-63 This has been attributed to formation of colour centres in the crystals.61 Recent flash photolysis experiments in gla~ses~~~~ and in solu-have shown that the lowest excited state of U022+possesses absorp- tions in the visible and near4.r.regions. For example, flash photolysis of a 2 x M solution of uranyl perchlorate in water yields a transient absorption 66 J. J. Weiss and D. Porret, Nature, 1937, 139, 1019; in this case, however, the photo- oxidation is suggested to proceed via electron transfer from water co-ordinated to CeIV. 67 G. Gordon and H. Taube, J. Znorg. Nuclear Chem., 1961,16,272. M. B. Ledger and G. Porter, J.C.S. Faraday Z, 1972,68, 539. K. C. Kurien and P. A. Robins, J. Chem. SOC.B, 1970, 855 and references therein. aoV. G. Firsov, Doklady Akad. Nauk S.S.S.R., 1961, 138, 1155; Chem. Abs., 1962, 56, 12459h. 61 N. A. Tolstoi, A. P. Abramov, and I. N. Abramova, Optics and Spectroscopy, 1966, 20, 415. A. M. Bonch-Bruevich, E. N. Kaliteevskaya, G. 0.Karapetyan, V. P. Kolobkov, P. I. Kudryashov, T. K. Razumova, and A. L. Reishakhrit, Optics and Spectroscopy, 1969,27, 433. O3 N. A. Tolstoi, A. P. Abramov, and I. N. Abramova, Proc. Znt. Conf. Lumin., 1969 (Publ. 1970), 330, ed. F. Williams, North-Holland, Amsterdam; Chem. Abs., 1971, 74, 17705. 64 C. C. Robinson, J. Opt. SOC.Amer., 1967, 57, 4; U.S.Dept. Comm. A.D. 624361 (1965) (Chem. Abs., 1967, 67, 16387a); L. A. Cross and L. G. Cross, Proc. Z.E.E.E., 1966, 54, 1460.N. A. Tolstoi, I. N. Abramova, and A. P. Abramov, Optics and Spectroscopy, 1969, 26, 314; I. N. Abramova, A. P. Abramov, and N. A. Tolstoi, ibid., 1969, 27, 293; V. P. Gapontsev, E. I. Galant, M. E. Zhabotinskii, Y. P. Rudnitskii and E. I. Sverchkov, Int. Conf. Microwaves Opt. Generation Amplification, [Proc.], Eighth Meeting, 1970 (publ. 1971), 13/22; Chem. Abs., l973,78,36076d. 66 L. N. Rygalov, A. K. Chibisov, A. V. Karyakin, E. V. Beuogova, B. F. Myasoedev, and A. A. Nemodruk, Teor. ieksp. Khim., 1972, 8, 484; Chem. Abs., 1973, 78, 130520r; G. Sergeeva, A. Chibisov, L. Levshin, and A. Karyakin, J.C.S. Chem. Comm., 1974, 159. 67 D. M. Allen, H. D. Burrows,A. Cox, R.J. Hill, T. J. Kemp, and T. J. Stone, J.C.S. Chem. Comm., 1973, 59. The Photochemistry of the Uranyl Ion (Figure 3) with Amax 590 nm, which is completely formed within the pulse N of 50 ns from a frequency-doubled ruby la~er.~~,~~ The decay of this absorption i 3-23[2-1-400 500 hjnm Figure 3 Flash photolysis spectra of aqueous solutions of uranyl salts.Full line- spectrographic recording of absorption from 2 x lop2 M uranyl perchlorate following ps flash; X-photoelectrically recorded absorptions on laser flashing 2 x 10-2 M uranyl perchlorate, -k -analogous experiments with uranyl nitrate (from ref. 67). (kl = 8.02 x lo5s-l) parallels the decay of the luminescence as determined by single-photon counting (kl = 3.85 x lo5 ~-9~~and similar parallels are found for the analogous absorptions and emissions in HzS04 and aqueous H3P0452p67 and rigid silicate glass solutions.62 Vibronic structure is perceptible in most of these absorptions, e.g. with average band separations of 580 cm-1 in fluid s0lution5~ and 700 k 30 cm-l in UV1-doped glasses.62 The identification of the electronic level from which the 590 nm absorption takes place with that from which emission occurs is confirmed by the identical quenching action of methanol and [2H3]methanol upon the kinetics of both processes (Table l).S2 This implies a triplet energy level scheme of the type shown in Figure 4.Reabsorption of uranyl luminescence by the 317, state under high-intensity excitation could account for both the observed non-exponential decay and the failure to observe significant laser action in uranyl glasses.68 In addition, this excited-state absorp- tion may act as an internal Q-switch in uranyl-Nd3+ glasses,69 although the occurrence of uranium(v) impurity centres cannot be ruled 0~t.70 A curious and, as yet, unexplained fact is that the uranyl excited state can be fie0. Risgin, U.S.Dept.Comm. A.D. 470360 (1965); Chem. Abs., 1967, 67, 16650f; H. Hartmann, G. Nitschmann, and G. Scharmann, Z. angew. Phys., 1968,24,69.(* N. T.Melamed, C. Hirayama, and P. W. French, Appl. Phys. Letters, 1965,6, 43. 'O C.B.Greenberg and P. W. French, J. Opt. SOC.Amer., 1968,58,472. 148 Burrows and Kemp Table 1 Absolute rate constants and isotope efects for the interaction between (U022+)*and alcohols measured by laser flash photolysis (ref.52) Alcohol Medium 10-6k2/l mol-l s-l ~C-HIC-D ~O-HIO-D CH30H H2O 6.40 k 0.08 (k~) ~I/~II= 2.76 ~I/~III= 0.98 CD30H H2O 2.32 f 0.10 (~II) CH30D D2O 6.50 k 0.19 (~III) k111/k1v= 2.39 k11/km= 0.85 CD30D D2O 2.72 k 0.04 (~Iv) (CH3)sCHOH H2O 85.3 k 0.5 2.41 (CH&CDOH H2O 35.4 k 0.9 CyClO-C6H11OH H2O 294 k 9 2.31 CyClO-C6D11OH H2O 127 k 3 excited by electrolysis71 or radiolysis72 of uranyl solutions, as indicated by the resultant characteristic luminescence. C. Energy Transfer.-A number of studies have been performed on energy transfer involving the U022+ i0n,~3-8~ with particular relevance to the use of 71 V. P. Kazakov, Zhur.fiz. Khim., 1965,39,2936; Chem. Abs., l966,64,12195e.72 C. Gopinathan, G. Stevens, and E. J. Hart, J. Phys. Chem., 1972, 76, 3698. 73 G. I. Kobyshev, G. N. Lyalin, and A. N. Terenin, Doklady Akad. Nauk S.S.S.R., 1963, 148, 1294; Chem. Abs., 1963,59, 127d. 74 G. N. Lyalin and G. 1. Kobyshev, Optics and Spectroscopy, 1963, 15, 135. 76 H. W. Gandy, R. J. Ginther, and J. F. Weller, Appl. Phys. Letters, 1964,4, 188; NASA Accession No. N64-27523, Rept. No. AD 441832, 1964; Chem. Abs., 1965, 62, 3777h; Appl. Phys. Letters, 1965,6,46; Rep. NRL Progr., 1966, 1;Chem. Abs., 1967, 66, 100095r; L. G. DeShazer and A. Y. Cabezas, Proc. I.E.E.E., 1964, 52, 1355; A. Kitamura, J. Phys. SOC.Japan, 1965,20,1283; M. E. Zhabotinskii, Y. P. Rudnitskii, V. V. Tsapkin, and G. V. Ellert, Zhur. exp. teor.Fiz., 1965, 49, 1689; Chem. Abs., 1966, 64, 12058h; M. J. Vogel,U.S. P., 3 265 628/1966; Chem. Abh., 1966, 65, 13000b; Y. I. Krasilov, Y. A. Polyakov, and Y. P. Rudnitskii, Izvest. Akad. Nauk S.S.S.R. Neorg. Materialy, 1956, 2, 2186; Chem. Abs., 1967, 66, 89910e; M. T. Artamonova, C. M. Briskina, and V. F. Zolin, Zhur. priklad. Spektroskopii, 1967, 6, 112; Chem. Abs., 1967, 67, 59134. 76 A. Y. Cabezas and L. G. DeShazer, U.S. P. 3 417 345/1968; Chem. Abs., 1969, 70, 42797r. 77 M. V. Artamonova, C. M. Briskina, V. F. Zolin, and N. M. Noginova, ‘Proceedings of International Conference on Luminescence’, 1966(publ. 1968),Vol. 2, p. 1699,ed. G. Szigeti, Akad. Kiado, Budapest; Chem. Abs., 1969, 70, 52786a; N. S. Belokrinitskii, M. E. Zhabotinskii, A.D. Manulskii, Y. P. Rudnitskii, M. S. Soskin, V. V. Tspakin, and G. V. Ellert, Doklady Akad. Nauk S.S.S.R., 1969, 185, 557; Chem. Abs., 1969, 70, 119857~; G. M. Gaevoi, M. E. Zhabotinskii, Y. I. Krasilov, Y. P. Rudnitskii, and G. V. Ellert, Izveht. Akad. Nauk Arm. S.S.R., Fiz., 1968, 3, 431; Chem. Abs., 1969, 71, 34378k; N. E. Alekseev, I. M. Buzhinskii, V. P. Gapontsev, M. E. Zhabotinskii, Y. P. Rudnitskii, V. V. Tspakin, and G. V. Ellert, Izvest. Akad. Nauk S.S.S.R., Neorg. Materialy, 1969, 5, 1042; Chem. Abs., l969,71,54715w. 78 G. Blasse,J. Electrochem. Soc., 1968, 115, 738. 7B M. V. Hoffman, J. Electrochem. Soc., 1970, 117, 227. F. J. Avella, U.S. P. 3 586 634/1971; Chem. Abs., 1971, 75, 6945%. N. M. Pavlushkin, M. V. Artamonova, C. M. Briskina, and G.V. Leuta, Steklo, 1971, 1, 32; Chem. Abs., l971,75,11443Ou. 84 B. M. Antipenko and V. L. Ermolaev, Izvest. Akad. Nauk S.S.S.R., Ser. Jiz , 1968, 32, 1504; Chem. Abs., 1969,70, 15730g. (refs. 83-85 overleaf) 10--60 (60930)-(54950)--50 (48930)-(4381?)-.Band -40 I d 19 Q -30 -20 c Band Ga (SingIet?)b Band F (Singlet ?) Band E (Singlet?) D (Singlet?) (3687,3?7 Band CC-t T,? !70)r'(21,329) 1 Vibration level spacing 5315430 744& 92h 715f18' L Fluorescence 855; 201* Figure 4 Energy level diagram fvom U022+ion (based on the data and assignments of Bell and Biggers, refs. 24 and 31) a Nomenclature of U.V. bands of Bell and Biggers (ref. 31). Multiplicity suggested by Bell and Biggers.Band C displays no vibrational structure in absorption and is regarded by Bell and Biggers as a member of the (singlet) series C to G which are separated by a constant spacing of 6015 f 39 cm-'. This level may be degenerate with a triplet level (denoted T,) which does display vibronic structure following excitation from the ground triplet state 317uwith Amax 574 nm (17420 cm-l); the latter absorption indicates T3 to be positioned at 38 750 cm-l, i.e. very near to band C at 36 873 cm-'. That T8 is triplet in character is suggested by the strong extinction of the 574 nm absorption which is clearly highly allowed. The transition depicted as I (of 10 038 cm-') may be the origin of the near4.r. absorption band found in the flash photolysis spectra of uranyl-doped silicate glasses (ref.62). ~2.eThe groups of levels centred around 31 367 and 24 107 cm-' are regarded as triplet states 3Au and 31721 respectively by Bell and Biggers (ref. 31) and McGlynn and Smith (ref. 7). f Assignments due to Bell and Biggers (ref. 31); emission from the 21 270 cm-l level is only 4.66% of the total. Data from ps flash photolysis absorption (ref. 52). hd Fine structure due to symmetric stretching frequency in the excited state (ref. 31). Fine structure due to symmetric stretching frequency in the ground state (ref. 31). 83 L. 1. Kononenko, N. S. Poluektov, and L. M. Burtnenko, Optics and Spectroscopy, 1967, 23, 321; Chem. Abs., 1968, 68, 55191~. L. M. Burtnenko, L. I. Kononenko, and N. S. Poluektov, Monokrist, Stsintill Org.Lyuminofory, 1968.4,246; Chem. Abs., I970,74,117990g. 8s R. Mabushima, Chem.Letters, 1973, 11 5. Burrows and Kemp U022+ in laser systems. For example, intramolecular energy transfer has been observed from organic ligands to the uranyl i0n,83,84 whilst the latter has been shown to transfer energy to tervalent lanthanide ions in both the solid ~tate~~-~l and solution.51~82 Energy transfer to phthalocyanine has also been 0bserved.~3*~~ UOz2+ has recently been shown to photosensitize the aquation of hexacyano-cobaltate(II1) ion in aqueous solution,s5 presumably via energy transfer to the lowest ligand field state of the CoIII complex. 4 Solution Photochemistry of U022+ U022+ possesses a rather low redox potential (UOz2+ + e-= UOz+;Eo = +0.05 V)s6 and thermal oxidations are normally comparatively The important processes in the photochemistry of uranyl systems are (i) oxidation of ligands or other species present in solution (including the solvent), usually, but not in- variably, with accompanying net reduction of the uranium to UIVand (ii) in the presence of oxygen, sensitized oxidations2 which depend on regeneration of UVI by autoxidation of reduced states of uranium.Both of these processes are thought to proceed via the intermediacy of UV species. Rabinowitch and Belford2 classified the oxidations into two types, i.e. (i) those proceeding via excitation of cluster complexes between uranyl and the oxidizing species, and (ii) those involving a kinetic encounter between (U022+)* and the substrate.We have found it convenient to use a slightly different division of the reaction types, set out below as A (excitation of complexes), B (intermolecular abstraction of a hydrogen atom), C (intermolecular electron transfer), and D (energy transfer). A. Excitation of Complexes.-U022+ ion readily forms complexes with a variety of ligands, including organic acids and inorganic ions. When a variety of carboxylate ions are added to solutions of uranyl ion, there is a modification in the shape and intensity of the spectrum.17 As with inorganic ligands, this may well result from a charge-transfer transition from the ligand to the metal ion (CTTM). In all cases photolysis destroys the ligand, but there are differences from acid to acid, and we shall present the results seriatim. (i) Oxalic acid.There have been a very large number of individual studies of this system,2,5,48@~89particularly because of its former near-universal use as a chemical actinometer, although most contemporary groups prefer that based on potassium ferrioxalate.90 The uranyl oxalate system is normally based on measurement of oxalate destroyed, but an ingenious development depends on the sensitive g.1.c. determination of CO releasedgl (after conversion into CH4.92 W. M. Latimer, ‘Oxidation Potentials’, 2nd edn., Prentice-Hall Inc., New York, 1952, Chapter 21. 87 T. W. Newton and F. B. Baker, Adv. Chem. Ser., No. 71, 1967, p. 268. 88 D. H. Volman and J. R. Seed, J.Amer. Chem. SOC.,1964,86,5095. C. A. Discher, P. F. Smith, I. Lippman, and R. Turse, J. Phys. Chem., 1963, 67, 2501. C. G. Hatchard and C. A. Parker, Proc. Roy. SOC.,1956, A235,518. s1 K. Porter and D. H. Volman, J. Amer. Chem. SOL,1962, 84,2011. O1 K. Porter and D. H. Volman, Analyt. Chem., 1962,34,748. 151 The Photochemistry of the Uranyl Ion Studies by a number of techniques suggest that the main complexes present in aqueous solution in the pH range 1-7 are the mono- and bis-oxalato-species.48 Unlike many other uranyl complexes, photodecomposition of the oxalato- uranium(v1) complex involves little net reduction of UVI (unlike the ferrioxalate systemg0 where$[FeII] is high). The chief pathways are (2) and (3), hu U022+ + HOzCCOzH -+ U022+ + C02 + CO + H2O (2) U022+ + HOzCCOzH -+ U022+ + C02 + HCOzH hu (3) and the redox breakdown (4) is very minor hu UOz2++ HOzCCOzH + U02H+ (i.e.UIV) + H+ + 2C02 (4) The ratio C#Z:$~ is determined by pH:48 whereas (2) accounts for ca. 96 % of all photodecomposition of oxalate at pH 0, at pH 3 and 7 this figure has fallen to 25 % and zero respectively. C02 production is pH-invariant andC#[UIV] increases slowly from pH 2 to pH 6, but is always low (i.e. between 0.0023 and 0.0063) but depends somewhat on [U1VJ4s993 Disregarding reaction (4) means that we are dealing with an apparent ‘sensitization’ reaction2 whereby although Uvl is instrumental in absorbing the radiation, probably via a CTTM transition to give a UV species (I) in reaction (3,the UV is reconverted into Uvl rather than under- going disproportionation to UVI + UIV.(u02’+* CZO~~-) [uv ’ COZ-] + c02 (5)-+ (1) The ‘complexed radical’ (I) acts as the source of formate at high pH by a reversal of the electron transfer (6), [UV* COz-1 + H2O -+ [UVI * HC02-1 + OH-(6) At low pH, the replacement of HC02- by CO as coprincipal product with C02 must lie in an alternative acid-catalysed breakdown (7): [UV * C02-1 + H30+ -+ UOzVIOH++ HCO+ (7) (the formatouranyl complex itself is not decomposed at these acidities to give CO). The quantum yield for oxalate disappearance shows a small but definite sensitivity to the wavelength of the irradiation? for example $(365.5 nm) = 0.492 whereas$(435.8 nm) = 0.573 (both k 0.002).Two pieces of physical evidence provide support for a long-lived intermediate in this photoreaction: Parker and Hatchard have detected a long-lived but reversible optical absorption on flash photolysis of a UVI-oxalate solution con- taining excess oxalic a~id.9~ Photolysis at 77 K of concentrated aqueous oxalic 93 J. J. McBrady and R. Livingston, J. Phys. Chem., 1946, 50, 176. 94 C. A. Parker and C. G. Hatchard, J. Phys. Chem., 1959, 63,22. Burrows and Kemp acid containing uranyl perchlorate95 yielded an e.s.r. signal displaying no hyper- fine splitting but g-factor anisotropy with the average g-factor of 2.0044, which is rather high for free C02- radicals. However, in the neighbourhood of a metal atom, the value typical of free o-radicals is likely to be increased to perturbation effects,96 consistent with an assignment to (I). (ii) Benzilic acid.The only product which could be identified by thin-layer chromatography was benzophenone.97 In addition some gas (probably C02) is evolved. Further information on the mechanism in this case has come from e.s.r. and flash photolysis studies.98 Upon intense photolysis of cold ethanolic benzilic acid and uranyl perchlorate in the cavity of an e.s.r. spectrometer, the e.s.r. spectrum of the diphenylketyl radical (Ph2COH) is observed. Similarly, upon ps flash photolysis of either aqueous or ethanolic solutions of uranyl ion and benzilic acid, the very well-known spectrum of the ketyl radical is observed immediately after the flash (Figure 5).This can readily be interpreted in terms of I 3 400 500 A/nm Figure 5 Flash photolysis spectrum obtained from a solution in ethanol of 2 mM uranylperchlorate and 2 mM benzilic acid using h > 300 nm and attributed to the radical Ph&(OH) (from ref. 98) (Reproduced by permission from J. Amer. Chern. SOC.,1971,93,2539) a CTTM transition leading to oxidation of the acid. The resulting species is expected to decarboxylate very rapidly (kl = 109-1010 to give the observed ketyl radical (8) and (9) (the two steps may possibly be concerted as the carboxyl radical is likely to be formed in an energy-rich state) : s6 D. Greatorex, R. J. Hill, T. J. Kemp, and T. J. Stone, J.C.S. Faraday I, 1972, 68,2059. P. W. Atkins, N.Keen, and M. C. R. Symons, J. Chem. SOC.,1962,2873. H. D. Burrows and T. J. Kemp, unpublished observations. s8 H. D. Burrows, D. Greatorex, and T. J. Kemp, J. Amer. Chem. SOC.,1971,93,2539. s9 L. Herk, M.Feld, and M. Szwarc, J. Amer. Chem. SOC.,1961, 83, 2998. The Photochemistry of the Uranyl Ion hv PhzC U022+ -+ U02+ + H+ + Ph2C(OH)C02* (8) \/c=o I OH Ph2C(OH)COi +Ph2COH + CO2 (9) Under high metal-ion concentrations, the ketyl radical is then oxidized to the observed product, benzophenone (10): PhzCOH + U0z2f-+PhzC0 + U02H2+ (10) This is similar to the photo-oxidations both of benzilic acid by ‘paraquat’loo and of 2-hydroxyisobutyric acid by Uvl to acetone and UIV.lO1 (iii) Lactic acid. Here the extent to which oxidative decarboxylation occurs depends on the medium; thus photolysis of solutions of uranyl nitrate and lactic acid in aqueous solution at high pH levels leads to the formation of acetaldehyde and carbon dioxide (1 l), whereas at low pH the organic product is pyruvic acid (12);lo2 in both pathways UIV is produced in stoicheiometric proportions.pH > 3.5 U022+ + MeCH(0H)COzH __+ MeCHO + UN species (11) Yo U0z2++ MeCH(0H)COzH --+ MeC + Urn species (12) ‘COzH This result is readily interpreted in terms of the species present. At high pH, the spectra indicate that lactate anion is complexed to UVI and decarboxylation can occur to yield initially the MeCHOH radical which is further oxidized to acetaldehyde. However, at low pH the extent of complexation is much less, and the photochemical reaction is intermolecular hydrogen atom abstraction by excited uranyl to yield the MeC(0H)COzH radical, which yields pyruvic acid on further oxidation.In support of this view, e.s.r. studies on frozen lactic acid solutions of UVI at 77 K confirm the presence of these radicals at pH 7 and pH 1 respectively (aIthough at pH 7 a small quantity of MeC(0H)COzH radical is also observed).95 Kemp et a1.67 report a preliminary figure for the rate constant of UVI* attack upon lactic acid of 9.8 x lo61mol-l s-l at pH 1.5 using laser flash pho tolysis. (iv) Acetic acid and its homologues and dicarboxylic acids. Heckler et aI.lQs have looJ. R. Barnett, A. S. Hopkins, and A. Ledwith, J.C.S. Perkin IZ, 1973, 80.lol K. Pan and W.-K. Wong, J. Chinese Chem. SOC.(Taiwan), Ser. 11, 1961, 8, 1;Chem. Abs., 1962,57, 3003g. loas.Sakuraba and R. Matsushima, Bull. Chem. Soc. Japan, 1970,43, 1950. lPa G. E. Heckler, A. E. Taylor, C. Jensen, D. Percival, R. Jensen, and P. Fung,J. Phys.Chem., 1963, 67, 1. Burrows and Kemp carefully determined the gaseous products of anaerobic photolysis of a number of carboxylic acids and di-acids in the presence of U022+ ion at pH values of 0.6-1.8, confirming the conclusion of earlier workers that the principal reaction is decarboxylation, viz. hv RCHzCOzH +RCH3 + C02 (13)UO,'+ Several dicarboxylic acids were selected for detailed micromanometric study (C02 being then the sole gaseous product) and it was found (i) that there is a general increase in the rate of C02 evolution with increase of pH, and (ii) that this rate is dependent on the concentrations of mono-ionized acid for succinic and glutaric acids and of di-ionized acid for malonic acid.These observations imply that UVLanion complexes are the photoreactive species: at high ( N 10 mM) concentrations of acid anion (or dianion) the rate reaches a plateau value corresponding to essentially complete complexation of UOz2+. Absorbence measurements of UVI-substrate solutions indicate the stoicheiometric compositions of the complexes to be UV1.(HO2CCH2CHzCO2-)2 and UV1.(02CCH2C02-). Unfortunately these authors record no yields for Uvl, but the net reaction (13) as written involves no consumption of oxidant.Conse- quently, if the first step is one of oxidative decarboxylation of the type effected by CelV and PbIV under irradiation, then the radical resulting from reaction (14) UV1-RCH2C02H -+ UV.RCH2 + C02 + H+ (14) might back-react with UV (as in the case of oxalic acid) to give RCH2- and thence RCH3 (the course with Ce'V and Pbm is that RCH; is oxidized by a second equivalent of oxidant to yield ReH2 and thence RCH20H). Left to itself RCHi can either dimerize to RCHzCHzR (the dimeric product is not normally found, although there exists an early report of C2H6 as a major product in the IF-sensitized decomposition of acetic acid which has not been subsequently con- firmed), or possibly attack further acid, especially at high substrate concentra- tions: RCH, + RCHzCOzH --t RCH3 + RCHC02H (1 5) The product from dimerization of RcHC02H is not, however, recorded in investigations that we are aware of.E.s.r. work indicates that reaction (14) does not appear to operate significantly for 'neat' solutions at 77 K;95 photolysis of frozen solutions of UVI in neat substrate yields e.s.r. spectra of the radical corresponding to abstraction of a hydrogen atom from carbon adjacent to the CO2H group for the following acids : acetic, propionic, isobutyric, cyclobutanecarboxylic, and cyclohexanecarboxylic. (The alkyl radical is obtained during photo-oxidation by CelV under these re- action conditions.)lW However, the alkyl radical expected from step (14) is found lo4D. Greatorex and T.J. Kemp, Trans. Faraday SOC.,1971,67, 1576. The Photochemistry of the UranylIon with succinic and trifluoroacetic acids, and in the case of acetic acid both CH3. and CH2C02H are found at 77 K, although only the latter radical survives on warming to 140 K. Irradiation of solutions of WI in neat acetic and propionic acids at 293 and 255 K respectivelylo5 gave well-resolved signals of CH2C02H and CH3cHC02H in conformity with the low-temperature data. It is possible that the abstraction of an a-hydrogen atom is predominant in highly concentrated solutions whereas decarboxylation is more important in dilute aqueous solution- such a complete change in mechanism with solvent composition has been estab- lished for the UVI*-methanol system (see below).Another possibility is that the entity W.RcH2 in (14) is too labile even at 77 K for e.s.r. detection and that the e.s.r. signals detected refer to a relatively minor reaction pathway. With certain other systems containing UVI, an intramolecular CTTM mechan- ism is also likely. Thus on photolysis of uranyl iodide complexes2J iodine formed may well arise from process (16), which is followed by steps (17) and (18): u021+--+ u02++ 1-(16) I*+ I---f 12-(17) 212---f I2 + 21-It appears that a photostationary equilibrium is attained in this system2J between the reactants and UIV and 12. Wan and col1eagueslo6 find that irradiation of aqueous tetrabutylammonium iodide yields a black precipitate of tributyliodo-ammonium iodide (C~HQ)~IN+I- without loss of UVI, and a sensitization mechan- ism is invoked.Recent studies featuring spin-trapping of radicals by benzylidene nitrone and t-nitrosobutane have shown that on photolysis of uranyl nitrate in methanol solution the immediate product radical is methoxyl (CH30-).107 This can then abstract a hydrogen from solvent methanol to yield CH20H, the radical observed upon photolysis in a glass at 77 K. The reaction can most readily be interpreted in terms of a CTTM transition from co-ordinated methanol to the U atom. The reaction, however, shows a marked solvent sensitivity. Upon photolysis of uranyl nitrate in 2:l (v/v) water-methanol, the only radical detect- able by spin-trapping is CHzOH.107 Possibly in this case only water is co- ordinated to the uranyl, such that an intermolecular mechanism must be involved (see later).B. Intermolecular Abstraction of a Hydrogen Atom by UVI*.-The quenching of the UVI emission by a number of alcohols, as measured by the Stern-Volmer quenching constant K~v~08-l~~correlates very well with the corresponding lo6 D. Greatorex, R. J. Hill, T. J. Kemp, and T. J. Stone, J.C.S. Faraday I, 1974,70, 216. lo*J. K. S. Wan, E. A. Schuck, J K Foote, and J. N. Pitts, jun., Canad. J. Chem., 1964,42, 2029. lo' A. Ledwith, R. J. Russell, and L. H. Sutcliffe,Proc. Roy. SOC.,1973, A332, 151. lo*S. Sakuraba and R. Matsushima, Bull. Chem. SOC.Japan, 1970,43,2359. logS. Sakuraba and R. Matsushima, Bull. Chem. Suc. Japan, 1971,44,2915. R. Matsushima and S.Sakuraba, J. Amer. Chem. SOC.,1971, 93, 5421. ll1 R. Matsushima, J. Amer. Chem. SOC.,1972,94,6010. Burrows and Kemp quantum yield108JlOJ11 for the photo-oxidations determined by c$(UIV). Again, the absolute rate constants for the quenching of UVI emission by CH30H and CD30H and other similar pairs of alcohols agree with those for the rates of decay of the 590 nm absorption of uVI* in the presence of these added substrates52 (Table 1). These observations imply that the quenching of UVI luminescence is both chemical and collisional in nature, viz. hv u022+__+ (u022+)* knr (U022+)**u022+ kr (U022+)*--+ u022++ hv kQ(U022+)*+ RCHzOH +U022+ + RcHOH + HL,, (22) where the subscripts refer to non-radiative, fluorescent, and quenching pathways respectively.kq implies a 'kinetic encounter' between (U022+)* and the alcohol, of the type proposed by Rabinowitch and Belford.2 That the quenching act involves inter- molecular hydrogen abstraction primarily from a C-H bond of the hydroxylated carbon atom is strongly suggested by the following evidence. (i) Luminescence quenching. For the series of alcohols methanol, ethanol, propan-2-01 and t-butyl alcohol, the quenching increases rapidly with the first three (with KSVvalues of 12,60, and 11 3 1mol-l respectivelyllO) as anticipated from the progressive reduction of the relevant C-H bond energies, but falls off sharply with t-butyl alcohol with KSV = 5 1 mol-l (all KSVmeasured at 293 K). Again substitution at the weakest bond by deuterium reduces both the value of KSVand the photoreactivity of the alcohol as determined by UIV production (Table 2).(ii) Quenching of emission Igetime. The emission of U0z2+ in water (kl = 3.85 x 105 s-l) is systematically reduced on addition of either CH30H or CD30H,52 yielding second-order rate constants for the quenching process of k(CH30H) = 4.48 k 0.08) x 1061mol-1 s-1 and k(CD30H) = (1.73 +_ 0.13) x 106 1 mol-1 s-l withk~lk~= 2.59 k 0.24. (iii) Quenching of the excited-state absorption of UVI. The characterization of the absorption spectrum of excited U022+ ion (see above) affords a convenient direct method of determination of rate constants of its reactions utilizing laser flash photolysis. Data exist for the photo-oxidation of three alcohols, including some in D2O for methanol52 (Table 1).(iv) E.s.r. spectra of irradiated glassy and liquid alcoholic solutions of WI. Irradiated glassy solutions of Uvl in alcohols at 77 K yield e.s.r. spectra95 of RcHOH in the case of primary alcohols except for n-propanol (see later), but with secondary alcohols C-C cleavage occurs: propan-2-01 gives CHj and butan-2-01 and pentan-3-01 both give 'C2H5. 2-phenylethanol gives the spectrum 157 The Photochemistry of the UranylIon Table 2 Isotope eflects on the Stern-Volmer quenching constants and quantum yields for UIV production in the interaction between (U022+)*and alcohols in aqueous solution Alcohol Ksva KHsv/KDsva C$H~IV/C$D~I~ CH30H 5.26 2.76 - CD30H 1.90 CH3CH20H 60 2.06 2.21 CD3CD20H 29 (CH3)2CHOH 113 2.76 1.94 (CH3) 2CD OH 41 CyClO-C6H11OH 369 2.30 C~CIO-C~DIIOH 160 a Data from Ref.52; data from Ref. 110. of CsH5CH2.; in this connection it is interesting to note that Ledwith et aZ.lo7 spin-trapped both benzyl radical and 2-phenylethanoxyl radical PhCH2CH20. during peroxydisulphate oxidation of 2-phenylethanol. The latter radical is considered to fragment, viz. PhCHzCH20. ---t PhCH2. + CH2=O (23) C-C cleavage is, as expected, the principal route with tertiary alcohols at 77 K : Substrate Principal Radical (CH3)sCOH CH3 (CH3)zCHC(CH&OH CH( CH 3) 2 CH3CH2C(CH3)20H CH2CHs CH3CH2CH2C(CH3)(C2H5)0H CH2CH3 CH3CH2CH(CH3)C(CH&OH CH3 + ? Liquid-phase e.s.r.studies98J05 were restricted to a narrower range of substrates: methanol, ethanol, propan-1-01, and butan-1-01 yielded RCHOH radicals (Figure 6), but (CH&CHCH20H gave (CH3)2CCHzOH, 2-phenylethanol gave benzyl radical, and propan-2-01 gave (CH&COH. The result with CH30H is interesting in connection with the related spin- trapping work of Ledwith et aI.lo7 mentioned in the previous section. In liquid methanol, the principal radical spin-trapped is CH30. (and not CH20H) although the latter is predominant in aqueous methanol. Clearly direct examina- Burrows and Kemp 2.5 mT H Figure 6 Electron spin resonance spectra of irradiatedfluid solutions in (a) methanol and (b) ethanol, of uranyl perchlorate (from R. J.Hill, Ph.D. thesis, University of Warwick, 1974) tion of the UVI-methanol system in the solid state or liquid yields a secondary radical formed by attack of methoxyl upon solvent : CH30. + CH30H -+ CH20H + CH30H (24) With all other alcohols, the spin-trapped radical is ReHOH107and it seems that methanol is quite atypical in its behaviour with IF*. (v) Photolysis products. That from ethanol is acetaldehyde in conformity with fast, secondary thermal oxidation of CH3eHOH.lO8 (vi) Substituent eflect on photolysis rate. The rate of photoreduction of UVI by various aliphatic alcohols correlates well with the Taft (T*va1ues110-112 (Figure 7). These studies also indicate that the polarity of the substituent is a more important factor than any steric effect on the rate of photolysis.It is evident from these results that (UO$+)* shares with triplet-state benzo- phenone, hydroxyl, and alkoxyl radicals a strong propensity to function oxida- tively by abstracting hydrogen atoms from activated C-H bonds, and it seems probable that these reagents share the property of a high spin density at the 11* R.Venkatraman and S. B. Rao, Indian J. Chern., 1971, 9, 500. The Photochemistry of the Uranyl Ion 0.0 -0.5 n B.-s-8 -1.0 12 13 --1.5 0.0 0.5 1.o. 1.5 &* Figure 7 Plot of log (#+o) as a function of Ca* where 40is the quantum yield for U" with Ca* = 0 (propan-2-01) and4 is the formation quantum yield divided by the number of a-hydrogen atoms (1) s-butyl alcohol, (2) propan-2-01, (3) i-butyl alcohol, (4) n-butanol, (5) n-propanol, (6)ethanol, (7) 3-chloropropan-l-ol,(8) methanol, (9) ethylene glycol, (10) benzyl alcohol, (1 1)2-phenoxyethanol, (12) 2-bromoethanol, (1 3) 2-chloroethanol.(Reproduced by permission from J. Amer. Chem Soc., 1972, 94, 6010) oxygen atom. E.s.r. investigation of UVI-doped glasses and crystals of organic materials following irradiation at 77 K indicates that hydrogen-atom abstraction is predominant with certain acids, esters, amides, aldehydes, ketones, and phos- phorus(rr1) corn pound^.^^ However, in many cases there is a competition between hydrogen abstraction and C-C bond cleavage: for example, in photolysis of uranyl perchlorate in an acetaldehyde glass, the spectra both of CH3.(indicating C-C cleavage) and of CHZCHO(dueto hydrogen abstraction) are observed. Pro-ducts studied on the photolysis of uranyl nitrate and acetaldehyde in dilute sulphuric acid solution indicate the formation of biacetyl,ll3 suggesting yet a further reaction, i.e. hydrogen-atom abstraction from the carbonyl carbon, but n8S. Sakuraba and R. Matsushima, Chem.Letters, 1972, 911. Burrows and Kemp no acetyl radical was found in the solid-state e.s.r. spectrum, in contrast to the result with photolysis of CelV solutions in a~eta1dehyde.l~~ In cases where carbon-carbon bond cleavage is observed, these radicals presumably arise either from a CTTM transition between complexed substrate and UVI or from intermolecular electron transfer from substrate to the excited uranyl.For example, photolysis of U022+ in a t-butyl alcohol glass gives the e.s.r. spectrum of the methyl radical. This is the product expected from decompo- sition of the t-butoxy radical,115 (25), (26): (CH&COH + U022+-+ U02+ + (CH3)3CO. + H+ (25) Similarly, as indicated above, solid-state e.~.r.~5 studies show that on photo- oxidation with U022+, 2-phenylethanol yields benzyl radical, via the inter- mediacy of the alkoxyl radical (PhCH2CH20.) [cf. equation (23)]. In all cases, the radicals observed upon C-C bond cleavage are those anti- cipated from the mass spectral breakdown pattern of the substrate. Thus, the mass spectrum of t-amyl alcohol116 displays its strongest peak at m/e 59, attri-+ buted to (CH&COH, which suggests loss of C2H5.; similarly, on photolysis of uranyl perchlorate in t-amyl alcohol glass at 77 K, the radical observed initially is C2Hs*.95 Further, the mass spectrum of ethyl methyl ketone indicates cleavage of the ethyl-carbonyl bond,l17 whereas on photolysis of a uranyl perchlorate- ethyl methyl ketone glass, the ethyl radical is observed by e.s.r.95 In the case of n-propanol, the mass spectrum116 shows its strongest peak at mle = 31, implying loss of C2H5' and the e.s.r.spectrum of glassy n-propanol irradiated in the presence of UVI shows CZH5' to be the principal radical118 (ethanol gives no C-C fission product). Whilst the radicals observed in cases where there is C-C bond cleavage are most readily explained in terms of an electron transfer to the uranyl group, it is not at present clear whether this takes place via an inter- molecular route, or via a CTTM transition in a uranyl-substrate complex. C.Intermolecular Electron Transfer to Excited Uranyl Ion.-While there is no spectroscopic evidence for ground-state interaction between U022+ and aromatic hydrocarbons, the latter efficiently quench uranyl fluore~cence~~~J~~ to give good, linear Stern-Volmer plots. These afford a correlation between the quenching constant KSVand the polarity of the aromatic compound as measured by the Yukawa-Tsuno Q constant for the substituent. There is no net photochemical reaction, nor is there a deuterium isotope effect on KSV, and the process has been interpreted in terms of an intermediate exciplex in which charge is transferred 114 D.Greatorex and T. J. Kemp, J.C.S. Faraduy I, 1972, 68, 121. 115 P. Gray and A. Williams, Chem. Rev., 1959, 59, 239. 116 R. A. Friedel, J. L. Schultz, and A. G. Sharkey, jun., Analyt. Chem., 1956,28,926. A. G. Sharkey, jun., J. L. Schultz, and R. A. Friedel, Analyt. Chem., 1956, 28,934. 11* R. J. Hill, Ph.D. thesis, University of Warwick, 1974. 119 S. Sakuraba and R. Matsushima, J. Amer. Chem. SOC.,1971,93, 7143. me Photochemistry of the Uranyl Ion from the aromatic wcloud to UVI. This is followed either by the reverse process or by radiationless deactivation of the complex. A number of metal ions have also been found120 to quench UVI*; where the quencher ion has a low-lying excited state then the quenching may simply result from electronic energy transfer.Some preliminary work indicates that the quenching in other cases may best be interpreted in terms of an intermolecular electron-transfer process.121 D. Energy Transfer.-One of the early mechanisms suggested for the photo- oxidation of oxalate by UOz2+involved energy transfer from UYI to the ligand122 but this was later disc0~nted.l~~ Energy transfer has since been observed between (U022+)* and lanthanide ions51~82 and also cO(cN)6.3-.85 Energy transfer from (U022+)*to acrylamide (to give a triplet state) has been invoked to explain the photosensitization by UOz2+ of the polymerization of acrylamide,l24 but it has been found that UV1 photo-oxidizes a number of vinyl monomers95 to give free radicals detected by e.s.r.and a freeradical mechanism may be responsible for the observed polymerization. 5 Reactions of Intermediate Species Following photo-oxidation of the substrate (either via electron-or atom-transfer) the immediate inorganic product is a uraniurn(v) species (probably U02+ or, in solutions of high uranyl concentration,lz5 U2ch3+, a species which has been identified kineticallylz6). UV is known to disproportionate rapidly in a process which depends on pH, anion concentration, and UOz2+ c~ncentrationl~~J~~ to yield UVI and UIV, and normally UIV is the observed end product of photo-oxidations by UVI. UV has also been shown to be oxidized back to UVI by inorganic ions such as copper(II)lO* and mercury(II).12s The Uv may also be oxidized back to UVI by complexed organic radicals (see above).The substrate radicals may undergo the following processes : (i) Further oxidation either by UOz2+or U02+, e.g. PhzCOH + UOz2+-+ PhzCO + UOa+ + H+ (27) 120 M. NovAk, Jadernd Energie, 1957,3, 44 (Chem. Abs., 1960, 54, 10505g); G. Alberti and A. Saini, Analyt. Chim. Acra, 1963, 28, 536; I. K. Shutov and D. S. Umreiko, Izvest. Akad. Nauk. S.S.S.R., Ser. fir., 1970,34, 1349 (Chem. Abs., 1970, 73, 71776a); Y.Kato and H. Fukutomi, Bull. Tokyo Inst. Technol., 1971,102,55(Chem. Abs., 1972,76,28227~);I. A. Taha and H. Morawetz, J. Amer. Chem. SOC., 1971,93, 829 and references cited therein. 181 H. D. Burrows, S. J. Formosinho, G.Miguel, and M. C. S. Viais, unpublished observation. W. West, R. H. Miiller, and E. Jette, Proc. Roy. SOC.,1928, A121, 294. A. H.Carter and J. J. Weiss, Proc. Roy. SOC.,1940, A174, 351. 1-24 K. Venkatarao and M. Santappa, J. Polymer Sci., Part A-I, Polymer Chem., 1970,8, 1785, 3429. 1-25 T.W. Newton and F. B. Baker, Inorg. Chem., 1965,4, 1166. 188 R. Bressat, B. Claudel, M. Fbve, and G. Giorgio, Compt. rend., 1968, 267, C, 707; G. Giorgio, Ann. Chim. (France), 1971,6, 53. la' J. Selbiu and J. D. Ortego, Chem. Rev., 1969,69,657. 118 S. Sakuraba and R. Matsushima, J. Amer. Chem. SOC.,1972, 94,2622. 162 Burrows and Kemp This reaction mode has been confirmed by flash photolysis.129 Similar oxidation of aliphatic radicals by metal ions has been observed.For example, Cohen and Meyersteinl30 studied the oxidation of several aliphatic radicals in aqueous solution by [C0(NH3)6l3+ and [Ru(NH3)3]3+ using the pulse radiolysis technique. (ii) Disproportionation, e.g. 2CH3CHOH -+ CH3CHO + CH3CH20H (28) This reaction mode has been identified by radiation chemical studieslzl and in many cases kinetic data have been obtained from pulse radi0lysis.l3~ While Matsushhna and Sakuraba108J28 favour this as the main route for the formation of the organic products in photo-oxidations of alcohols by UVI, based on the absence of any effect of other metal ions on4 (organic product) it is probable that in most cases both reactions of type (27) and (28) occur. Both yield the same final product, and the predominance of one process will be favoured by variations in the metal ion and radical concentrations.(iii) Dimerization. The latter is normally important only at low metal-ion concentrations, but is most probably responsible for the formation of biacetyl in the photo-oxidation of acetaldehyde by uranyl perchlorate in dilute sulphuric acid .l13 The processes of dhnerization and disproportionation with organic radicals have recently been reviewed.133 (iv) Decomposition. This is most probably the fate of the radicals produced prior to C-C bond cleavage, e.g. the formation of the benzyl radical from the 2-phenylethanoxy radical [equation (23)]. (v) Hydrogen abstraction. This has been observed both by e.s.r. studies on glasses at 77 K105 and, in the case of the methanoxy radical, by spin-trapping studies in solution at room tem~erature.10~.(vi) Polymerization. Photolysis of vinyl monomers in the presence of U0z2+ readily effects polymerization.3J24J34 6 Applications of the Photochemistry of Uranyl Compounds A number of applications of the photochemical reactivity of UVI have been suggested, of which a few representatives are as follows. The photoreduction of UVI to UIV by, for example, ethanol, and the subsequent precipitation of the UIV as the oxysulphate, has been used as a method for separating uranium from other la@H. D. Burrows and T. J. Kemp, to be published. lSo H. Cohen and D. Meyerstein,J. Amer. Chem. SOC.,1972,94,6944. lS1 J. T. Allan, E. M. Hayon, and J.J. Weiss, J. Chem. SOC.,1959, 3913. 13a M. Simic, P. Neta, and E. Hayon, J.Phys. Chem., 1969,73,3794. lS3M. J. Gibian and R.C. Corley, Chem. Rev., 1973,73,441. lS4 S. Okamura and S. Tazuke, Jap. P. 7034392/1970 (Chem. Abs., 1971, 74, 64633~); M. Amagasa, S. Kinumaki, Y.Kasuga, H. Chiba, K. Konno, H. Ueda, and H. Mukunaski, Kobunshi Kagaku, 1970,27,718 (Chem. Abs., l971,75,6401b). The Photochemistry of the Uranyl Ion metal ions.135 Further, the photoreduction, and subsequent analysis of UIV, has been used as a rapid method for determining UVI, particularly at trace levels.136 Uranyl-ion photosensitized polymerization has been mentioned earlier. This can be used as a method for coating metal, paper, wood, or glass surfaces.137 A silver-free photographic process has been proposed, based on the reduction of HgII by a photolysed uranyl-organic system in a viscous polyvinyl alcohol system.126 Finally, mention should be made of a study of the uranyl acetate-photo- sensitized killing of Escherichiu coli K-12.Neely and co-workersl38 found that under the conditions they used, the uranyl-photosensitized killing occurred only after 150 minutes. Binding of U0z2+to the cell was shown to be fairly rapid (complete within 30 minutes) and the observation was attributed to the necessity for penetration of U0z2+ into the cell's interior to effect photosensitization.The photochemistry of IF1may thus provide a probe into the structure of the interior of cells. Perhaps surprisingly, few other reports exist on uranyl photoreactions in biological systems, although Wacker and c0lleagues~3~ have noted that uV1 can photochemically cleave thymine dimers to give thymine, and another unidentified product.7 Some Comparisons with Other Metal Ions Whilst UOz2+ shares with several lanthanide ions the comparatively unusual phenomenon of 1umine~cence~~J~~J~~ and with a number of one-equivalent oxidants such as CeIV,98J04J14 TP, PbW,142 FeIII,143 and CuII143 the ability to photo-oxidize a variety of substrates, it is the best-known exponent of each of these properties and the only member common to both groups. The luminescence of Eu3+ is mure intense in methanol than in water, i.e. photochemical attack on C-H bonds by (Eu3+)* is imperce~tible.5~J~~ In the group of photo-oxidizing metal ions Uvl is the weakest as regards the thermodynamic energy released on the reduction process and this is reflected in its tendency to oxidize molecules like t-butyl alcohol by C-H cleavage to -CH2C(CH3)20H rather than by the C-C cleavage (to give CH3.) exhibited by CeIV on photo-oxidation.Again, UVI photo-oxidation of CH3CHO gives CH2CHO whereas the CelV reaction yields the acyl radical CH3C0*.l14 Photochemical studies of PbIV* have been confined to carboxylic acid complexes142 and the data with FelI1 and CuII are compara- tively few,3J43 making comparison with IF1difficult at this stage. 136 C. R. Das and D. Patnaik, Chem.and Znd.,1973,699 and references given therein. 136 W. M. Riggs, Analyt. Chem., 1972,44,390.13' K. Juna, H. Nakayama, and K. Asada, Ger. Offen, 2 010 867/1970; Chem. Abs., 1970,73, 121659n. 13* W. C. Neely, S.P. Ellis, and R. M. Cody, Photochem. and Photobiol., 1971, 13, 503. 139 A. Wacker, H. Dellweg, L. Trager, A. Kornhauser, E. Lodemann, G. Turck, R. Selzer, P. Chandra, and M. Ishimoto, Photochem. and Photobiol., 1964, 3, 369. 140 Y. Haas and G. Stein, Chem. Phys. Letters, 1972, 15, 12. lol Y. Haas and G. Stein, J. Phys. Chem., 1972,76, 1093. 142 K. Heusler and H. Loeliger, Helv. Chim. Acta, 1969, 52, 1495; H. Loeliger, Helv. Chim. Acta, 1969,52, 1516; P. B. Ayscough, J. Machovi, and K. Mach, J.C.S. Faraday IZ, 1973, 69, 750. 143 R. J. Hill and T. J. Kemp, unpublished results. 164 Burrows and Kemp As regards U.V. spectra, a comparison between U022+ and other actinide 0x0-cations has been made32 and the results have been accommodated by a single molecular orbital model. One of us (H.D.B.) is grateful to the Instituto de Alta Cultura for financial support.

ISSN:0306-0012    

DOI:10.1039/CS9740300139

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Non-conventional Electrophilic Aromatic Substitutions and Related Reactions By S. R. Hartshorn DEPARTMENT OF CHEMISTRY , UNIVERSITY OF DURHAM, SCIENCE LABORATORIES, SOUTH ROAD, DURHAM, DH1 3LE 1 Introduction Electrophilic aromatic substitutions have been studied extensively.1 The general features of this class of reaction are now well understood and can be explained in terms of the two-step sE2 mechanism which involves a relatively unstable a-complex (Wheland intermediate) :2 X ArX + E+ +Ar+/ +ArE + X+ \ E There have been attempts to modify this mechanism by introducing additional intermediates (wcomplexes),3 but in general such extensions are thought to be unnecessary.4 Until recently most attention was given to those nuclear substitutions in which X = H (or one of its isotopes), i.e.conventional electrophilic substitutions, but there is now a growing interest in other types of electrophilic reaction. In particular, there are those substitutions in which X is a group other than hydrogen,5 and there are those reactions which appear to be initiated by the attack of an electrophile but which give products other than those expected for a conventional substitution.6 The different types of behaviour may be classified as follows: (i) Nuclear substitution. The electrofugal leaving group may or may not be hydrogen. (ii) Side-chain substitution. This usually involves an alkyl side-chain. R. Taylor in ‘Comprehensive Chemical Kinetics’, ed. C. H. Bamford and C. F. H. Tipper, Elsevier, Amsterdam, 1972, vol.13. (a) H. Zollinger, Adv. Phys. Org. Chem., 1964, 2, 163; (b) C. K. Ingold, ‘Structure and Mechanism in Organic Chemistry’. 2nd edn., Bell, London, 1970; (c) P. Rys, P. Skrabal, and H. Zollinger, Angew. Chem. Internat. Edn., 1972, 11,874. * G. A. Olah, Accounts Chem. Res., 1971, 4, 240. 4 D. V. Banthorpe, Chem. Rev., 1970,70,295. 6 (a)C. L. Perrin, J. Org. Chem., 1971, 36,420; (6)C. L. Perrin and G. A. Skinner, J. Amer. Chem. SOC.,1971, 93, 3389; (c) P. B. Fischer and H. Zollinger, Helv. Chim.Acra, 1972,55, 21 39. E. Baciocchi and G. Illuminati,Progr. Phys. Org. Chem., 1967,5, 1. 167 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions (iii) Rearrangement. Initiated by an electrophilic reagent.(iv) Elec tr ophilic addition. (v) Other reactions. Examples include the formation of dienones, quinones, and products of coupling. These reactions all appear to be closely related mechanistically and the pathway followed in a particular reaction may depend upon the reaction conditions, the structure of the substrate, and the nature of the leaving group. The reactions of class (ii) have been called non-conventional electrophilic substitutions ;6 the term non-conventional will be applied in this review to all of the above processes other than nuclear substitutions in which hydrogen is the leaving group. It should be emphasized that a reaction need not be confined to a single path- way, and indeed rearrangements and additions frequently accompany substitu- tion, sometimes with evidence for a common reaction intermediate. Because of the number of pathways available, complicated mixtures of products are often obtained and it is perhaps for this reason that non-conventional reactions have been overlooked in the past, since minor components of reaction products have not always been characterized.2 Electrofugal Leaving Groups other than Hydrogen Many electrophilic substitutions in which the leaving group is other than hydrogen have been reviewed.l The different types of reaction are well estab- lished and interest now lies in finding an order for the leaving abilities of the different electrofugal group^.^ One recent attempt to establish such an order, based on the magnitudes of hydrogen isotope effects, has led to the following sequence for increasing leaving ability:5a C1+ z NO2+ z R+ < Br+ < D+ z ArNz+ z SO3 z RCO+ < NO+ z H+ z I+ < Hg2+ < Me&+.It has been suggestedsa that the compound (1; X = Cl) should be a good model for a o-complex. The reactions of (1) should therefore give some in- X NO, -ass formation about the behaviour of the Wheland intermediates involved in non- conventional electrophilic substitutions. In mixtures of acetic acid and acetic anhydride (1) decomposes by loss and migration of NOz+ rather than C1+, suggesting that NO2+ is a better leaving group than Clf; with the corresponding bromo-compound, however, Br+ rather than NOz* is lost. Similarly, the re- Hartshorn arrangement of 1-chloro-1 -methyl-2-keto-l,2-dihydronaphthalenein acetic acid7 indicates that C1+ is a better leaving group than Me+; the sequence of leaving abilities Me+ < C1+ < N02+ < Br+ is suggested by these results.In the diazodehalogenation of (2) the ease of displacement of halogen is found to be C1+ < Br+ < I+ < H+.5c It was pointed out that this represents only an apparent order of leaving ability, because in the case X = Br the reaction is catalysed by thiosulphate ions, but such catalysis is not observed in the other cases. The rates of bromination of certain phenols provide another means for estimating relative leaving abilities.* The general agreement among the various methods is encouraging, but several difficulties remain. One problem, illustrated by the results for diazodehalogena- tion, is that the loss of the electrofugal group may involve interaction with a nucleophile. Relative leaving abilities might therefore show some variation with the reaction conditions depending upon the mechanism by which the electrofugal group is removed from the a-complex.Such a consideration implies that a single order of leaving abilities may be inappropriate. Another problem concerns the identity of the actual electrophilic reagent, which appears to be in question for certain reactions. For example, in the nitrodeiodination of 4-iodoanisole the preliminary step is thought to involve nitro~odeiodination.~ Substituent effects on the reactivity of an aromatic compound towards conventional electrophilic substitution are expressed quantitatively in terms of partial rate factors,2b which give the reactivity of a ring position, ortho, meta, or para to the substituent, relative to a single position in benzene.For the types of reaction discussed in this section it is necessary, in addition, to have a measure of the reactivity of the ring position at which the substituent is attached; the name @so has been suggested for this position, and an ips0 partial rate factor (ifR) has been defined for this purpose:5 ~k~ x r% attack at ips0 position in ArR ifR = kArH x % attack at corresponding position in ArH ~RIn this expression ~A represents the rate constant for total reaction of ArR, the system of interest, and kArH that for the corresponding system in which the substituent R is replaced by hydrogen.A limited number of @so partial rate factors have been determined and these are collected in Table 1. The choice of name for the term ifR is unfortunate since partial rate factors are generally understood to be measures of reactivity relative to a position in benzene. The definition of ifR, however, means that these factors do not express reactivities relative to benzene, or indeed, to any single reference compound. 3 Electrophilic Side-chain Halogenation The side-chain halogenations of polymethylbenzenes under certain conditions K. Fries and K. Schimmelschmidt, Annalen, 1930,484,245.* L. G. Cannell, J. Amer. Chem. Soc., 1957, 79,2932.* A. R. Butler and A. P. Sanderson, J.C.S. Perkh ZZ, 1972, 989. * Apparent ipso-factor, see text and ref. f. 169 2 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions Table 1ips0 Rate factors (iP)for some electrophilic substitutions Reaction R if R Ref. Pro todesilylation Me& ca. lo5 a Bromodesilylation Me3Si ca. 108 a Diazodesilylation Me3Si <1 a Protodeboronation B(OH)2 ca. 4 b Bromodemet hylat ion Me 0.29 C Bromode-t-butylation But 0.25 C Bromine exchange Br < 2 x 10-7 C Protodecarboxylation CO2H 2 x 10-3 d I 0.18 e Nitrodehalogenation Br c1 0.079 0.061 e e I 0.149 f Diazodehalogenation Br 0.0089* f c1 0.0070 f a C.Eaborn, ‘Organosilicon Compounds’, Butterworths, London, 1960; b H. G. Kuivila and K. V. Nahabedian, J. Amer. Chem. SOC.,1961, 83,2159; C E. Baciocchi and G. Illuminati, J. Amer. Chem. SOC.,1967,89,4017; J. L. Longridge and F. A. Long, J. Amer. Chem. SOC., 1968,90,3092; C. L. Perrin and G. A. Skinner, J. Amer. Chem. SOC.,1971, 93, 3389; f P. B. Fischer and H. Zollinger, Helv. Chim. Acta, 1972,55,2139. appear to be electrophilic reactions.6JO Thus hexamethylbenzene and molecular chlorine react together rapidly in acetic acid, in the absence of light and catalyst, to give mainly chloromethylpentamethylbenzene (3);loU9 l1 it is known that reaction by a free-radical mechanism under similar conditions occurs only slowly. The kinetic orders and relative reactivities observed with other hexa- substituted benzenes (CCMe5X) suggest that the slow step in side-chain chlorina- tion is electrophilic in character.The relative rates given in Table 2 show the similarities between side-chain reactions and conventional electrophilic substitu- tions. 10 (a) E. Baciocchi and G. Illuminati, Tetrahedron Letters, 1962, 637; (b) G. Illuminati and F. Stegel, Ricerca scient., 1964, 34,458; (c) E. Baciocchi and G. Illuminati, Ricercu scient., 1964, 34, 462. 11 E. Baciocchi, A. Ciana, G. Illuminati, and C. Pasini, J. Amer. Chem. Soc., 1965, 87, 3953. Hartshorn Table 2 Relative rates for some diflerent chlorination reaction@ Reaction kMe/kcl kMe/kCN Substitution (side-chain), CsMesX ca.3 x 103 ca. 1.4 x lo6 Electrophilic substitution (nuclear), CsHsX 0.8 x lo3 12.8 x lo6 Radical substitution (side-chain), CsMesX 2.2 4.3 a E. Baciocchi, A. Ciana, G. Illuminati, and C. Pasini, J. Amer. Chem. SOC.,1965,87, 3953. The kinetics of the reactions of hexamethylbenzene with bromine in acetic acidlOCJ2 and with iodine monochloride in carbon tetrachloridel3 are also of the forms expected for electrophilic processes. The similarities between side-chain and conventional electrophilic substitutions have led to the suggestion that the slow step in the former type of reaction is the formation of a o-complex. It is Scheme 1 assumed that an electrophile may initially attack any activated position of the aromatic system (Scheme l), regardless of whether a substituent is present and regardless of whether a nuclear displacement will subsequently occur.Little is known about the steps following the slow step, although several mechanisms have been suggested for the formation of side-chain products.6JocJ1 Tn one view the electrophile migrates to the methyl group attached to the position initially attacked (Scheme 2); a process analogous to the quinobenzilic rearrangement observed with many dienones.14 Another suggestion (Scheme 3) is that the electrophile migrates to an ortho-methyl group. Since the o-complex y t c )$.. ... Scheme 2 l* E. Baciocchi, M. Casula, G. Illuminati, and L. Mandolini, Tetrahedron Letters, 1969,1275. Is R.M. Keefer and L.J. Andrews, J.Org. Chem., 1966,31,541 l4 V.V.Ershov, A, A, Volod'kin, and G. N. Bogdanov, Rum. Chem. Rev., 1963,32,75. Non-conventional Electrophilic Aromatic Substitutions and Related Reactions is an acidic species,15 it first loses a proton to give (4) which then rearranges to give side-chain product (5). Both mechanisms assume that the rearrange- Scheme 3 ment is intramolecular in order to explain the high retention of halogen that is observed. The chlorination of hexamethylbenzene in acetic acid gives mainly chloromethylpentamethylbenzeneand only a small amount (ca. 5 %) of solvo- lysis product, apparently formed by solvent capture of a reaction intennediate.11 Bromina t ion under the same conditions gives bromome thylpent ame thy1 benzene with no evidence of accompanying solvolysis.lOc It has been pointed out11 that the greater tendency for retention in the case of bromine is in the same order as the ability of the halogens to form bridged intermediates.The small amount of solvolysis accompanying chlorination, however, indicates that reaction with an external reagent can compete with the rearrangement. Proton loss from the a-complex can lead to (6),1OC but (4) is thought to be the (6) W. von E. Doering, M. Saunders, H. G. Boyston, H. W. Erhart, E. F. Wadely, W. R. Edwards, and G. Laber, Tetrahedron,1958,4,178. Hartshorn more probable intermediate, presumably because rearrangement of (6) is expected to be an intermolecular process and so does not explain the high re- tention of halogen.However, under the usual reaction conditions it is probable that the ionization of (6), or (4), does not lead to kinetically free (i.e. fully dissociated) ions;l6 rearrangement may then occur at an ion-pair stage with no appreciable intervention by external nucleophiles (Scheme 4), although such inter- vention (by anion exchange) need not be totally excluded. CH,E Scheme 4 Some support for ionic intermediates in the rearrangement comes from the observation that in the chlorination of he~aethylbenzenel~ the amount of solvolysis accompanying side-chain substitution (ca. 15%) is significantly larger than the 5% found in the case of hexamethylbenzene. This difference can be explained in terms of the greater stability of the secondary carbonium ion in- volved in the former rearrangement, which provides more opportunity for interference by external nucleophiles.There is no direct evidence at present about which reaction pathway is important in side-chain halogenation; kinetic arguments are inapplicable, and product studies are ambiguous. Thus in the chlorination of isodurene, substitu- tion occurs almost exclusively in the methyl group at the 5-position (7).l* Of the a-complexes possibly involved, (8) and (9), in which the positions attacked are both activated by two ortho-methyl groups and onepara-methyl group, appear to be the most likely. Such a consideration seems to eliminate the mechanism l6 S. Winstein, B. Appel, R. Baker, and A. Diaz, Chem. SOC.Special Publ., No.19, 1965, p. 109. l7 G. Illuminati, L. Mandolini, E. M. Amett, and R. Smoyer, J. Chem. SOC.(B), 1971, 2206. G. Illuminati, L. Mandolini, and A. Patara, Tetrahedron Letters, 1972, 4161. Non-conventional Electrophilic Aromatic Substitutions and Related Reactions involving rearrangement into the gem-methyl group, but does not distinguish between the other possibilities. 4 Electrophilic Addition Accompanying Halogenation Addition is known to accompany the chlorination and bromination of many polycyclic aromatic hydrocarbons.19 Most examples refer to chlorination by molecular chlorine in acetic acid,20 although other solvents have been used,zoc and other chlorinating agents, such as sulphuryl chloridezl and hypochlorous acid,22 have been investigated.Fewer results are available for bromination, but the conditions studied include molecular bromine in methan01,~s in aqueous acetic and in carbon tetrachl~ride.~S A general difficulty in studying these additions arises because the initial product often undergoes further reactions, leading to complicated products and product mixtures.lg This difficulty may be minimized by a suitable choice of substrate, such that the structure limits the nature of the products obtained. The chlorination of phenanthrene by molecular chlorine in acetic acid shows second- order kinetics and gives the following mixture of products:20c(12), 35 %; (13), 38%; (14), 10%; (15), 5%; (16), 12%. The reaction shown in Scheme 5 has been pr0posed.1~ The addition of lithium chloride increases the amount of trans-dichloride (14) but does not affect appreciably the amount of cis-dichloride (13).The addition of sodium acetate increases the amount of acetoxy-chloride products and also increases the ratio of cis- to trans-dichloride. These effects are readily explained if two intermediates are assumed to be involved in the reaction. Capture of (1 1) l9P. B. D. de la Mare and R. Bolton, ‘Electrophilic Addition to Unsaturated Systems’, Elsevier, Amsterdam, 1966. 80 (a) P. B. D. de la Mare, M. D. Johnson, J. S. Lomas, and V. Sanchez del Olmo, J. Chem. SOC.(B), 1966, 827; (b) P. B. D. de la Mare, G. Cum, and M. D. Johnson, J. Chem. SOC. (C), 1967, 1590; (c) P. B. D. de la Mare, A. Singh, E. A. Johnson, R.Koenigsberger, J. S. Lomas, V. Sanchez del Olmo, and A. Sexton, J. Chem. SOC.(B), 1969, 717. 81 (a) R. Bolton, P. B. D. de la Mare, and H. Suzuki, Rec. Trav. chim., 1966, 85, 1206; (b)P. B. D. de la Mare and H. Suzuki, J. Chem. Soc. (C), 1967, 1586. P* P. B D. de la Mare and L. Main, J. Chem. SOC.(B), 1971,90. a3 J. van der Linde and E. Havinga, Rec. Trav. chim., 1965, 84, 1047. 2* L. Altshuler and E. Berliner, J. Amer. Chem. Suc., 1966, 88, 5837. F. R. Mayo and W. B. Hardy, J. Amer. Chem. SOC.,1952,74,911. 174 Hartshorn -\U4 / / // / / '\\ \ I\\ \ \L \O \ n n b,0, vI W \ \543\ \ I\\\ Non-conventional Electrophilic Aromatic Substitutions and Related Reactions by added nucleophiles gives predominantly, though not necessarily entirely, trans-addition products.The insensitivity of the yield of cis-dichloride (13) to the presence of added salts indicates that the major part of this product is produced from another intermediate (10) without the intervention of external reagents. Although cis-addition could be a one-step reaction involving a cyclic transition state (17), the dependence of the rate of addition on the polarity of the solvent indicates that a transition state with ionic character is involved.20c This is represented by (lo), but the ion pair (18) might be an alternative representation; internal return16 from (18) could give cis-dichloride, whereas further dissocia- tion would give (11). C' ...........C' y............i/ 7 -c v (1 7) Similar considerations can be used to explain the chlorination of naphthalene, although in this case there is the additional complication of the further reaction of the initial Addition accompanies the chlorination of 1-methyl-naphthalene, and in this case a small amount of side-chain substitution (ca.6 %) also occurs.20* 5 Other Non-conventional Processes in Halogenation The bromination of 1,3,5-tri-t-butylbenzenein acetic acid is accompanied by acetoxylation, but the latter reaction does not occur in the absence of bromina- tion. The results suggest that bromination and acetoxylation involve common reaction intermediates, and a mechanism (Scheme 6) including addition-elimina- tion pathways has been proposed.26 The yield of aryl acetates is increased by the addition of sodium acetate, which suggests that acetoxylation depends on the ability of acetate ions to capture these intermediates [(19) and (20)].In the case of the o-complex (19) capture by acetate ion competes with proton loss, which implies a relatively slow proton transfer in the pathway leading to (21), a cir- cumstance confirmed by the observation of a primary hydrogen isotope effect.26~~~ The formation of products (22) and (23) shows that the t-butyl group is a good electrofugal leaving group under these conditions. This example illustrates how conventional and non-conventional processes can be closely related. P. C. Myhre, G. S. Owen. and L. L. James, J. Amer. Chem.Soc., 1968, 90, 2115. I7P. C. Myhre, Acta Chem. Scand., 1960,14,219. Hartshorn + + Scheme 6 6 Non-conventional Processes in Nitration:Side-chain Substitution Reports of anomalous nitrations are to be found in the early literat~re.~~~~~ Recently, interest has revived in certain of these reactions, particularly in substitu- tion occurring in an aromatic side-chain. The results of product and kinetic studies indicate that these reactions, like the side-chain halogenations, involve ionic rather than radical processes. Thus the nitration of substituted penta- methylbenzenes (24) by fuming nitric acid in dichloromethane at low tempera- tures exhibits high positional selectivity.30 Substitution (to give side-chain nitrates) takes place almost exclusively at the ortho-position (25) with the strongly electron-withdrawing substituents X = NOZ, C02H, or C0~Me.3~With the substituents X = C1, Br, or I some substitution occurs at the rneta-position, in D.V. Nightingale, Chem. Rev., 1947, 40, 117. *9 A. V. Topchiev, ‘Nitration of Hydrocarbons and other Organic Compounds’, Pergamon,London, 1959. *O (a) H. Suzuki, BUN. Chem SOC.Japan, 1970, 43, 481; (b) H. Suzuki and K. Nakamura, ibid., 1971, 44,227. 177 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions addition to predominant ortho-subs ti tu t ion, whereas with the elec tron-releasing groups X = OH, OMe, or NHCOMe substitution occurs mainly at the meta-position (although complications arise owing to reactions involving the sub- sti tuents).30 b The kinetic data for the substituted pentamethylbenzenes also support the idea of an ionic (electrophilic) reaction pathway and are given in Table 3.The Table 3 Relative rates of nitration for some side-chain and nuclear substitutionsa Compound x , 1 Me H Br NOz CaMesX 1 2 x 10-2 4 x 10-4 3 x 10-6 CsHsX 1 io-1-10-2 10-3 10-5-10-* 0 K. Nakamura, Bull. Chem. SOC.Japan, 1971,44, 133. similarities between the substituent effects in the series CsMe5X and CeHsX indicate that side-chain substitution and conventional nuclear substitution are mechanistically similar and suggest that both processes have similar rate- controlling steps, presumably the formation of a o-complex (cf.side-chain halogenation). The stabilities of the a-complexes might account for the posi- tional selectivity observed,31 although this has also been discussed in terms of the slow step in proton removal from the side-chain.32 One interesting feature of side-chain nitrations is that the products obtained often depend upon the reaction conditions. The nitration of hexamethylbenzene by fuming nitric acid in dichloromethane at 5 "Cgives a variety of products which mostly appear to be derived from the initial product (26).33 Nitration by (26) (27) (28) 81 S. R. Hartshorn and K. Schofield, Progr. Org. Chem., 1973,8, 278.*' K. Nakamura, Bull. Chem. Soc. Japan, 1971,44, 133. ss H. Suzuki, Bull. Chem. SOC.Japan, 1970,43, 879. Hartshorn an equimolar amount of nitric acid in acetic anhydride at 0 "C,however, gives (27) as the major product, together with smaller amounts of the acetate (28) and the nitrate (26).34 It has been suggested that the nitrate is formed by an intramolecular cyclic rearrangement (Scheme 7), in which the initially formed nitrite ester is readily oxidized to nitrate under the reaction conditions, and that the nitro-compound 0 is formed by some other mechanism.% Presumably the importance of the different reaction pathways depends upon the substrate and the reaction con- ditions.The formation of acetate (28) indicates that side-chain substitution can involve an external reagent, as for example the capture of the cyclohexadiene intermediate (29) by an electrophile (30) or nucleophile (3 1). Cyclohexadiene intermediates with the ortho-quinoid structure have been considered, although it has been pointed out that those with the para-quinoid structure (32) might be expected to be more likely.35 34 R.Astolfi, E. Baciocchi, and G. Illuminati, Chimica e Industria,1971, 53, 1153. 36 E. Hunziker, J. R. Penton, and H. Zollinger, Helv. Chim. Ada, 1971,54,2043. Non-conventional Electrophilic Aromatic Substitutions and Related Reactions There is some evidence to suggest that the nitro-compound is only formed in the presence of a nitrating agent (see next section) and thus involves electro- philic attack; however, a nucleophilic pathway cannot be excluded on the present evidence. One such mechanism involves rearrangement within an ion pair (Scheme 8);3236 since the nitrite anion is an arnbident nucleophile37 both side- or fHzONO Scheme 8 chain nitrates and nitro-compounds can be produced. Although there has been much speculation, little is actually known about the reaction pathways leading to side-chain products.7 Non-conventional Processes in Nitration :Addition-Elimination The moderately stable cis-and trans-acetoxynitrodimethylcyclohexadienes(33) and (34) have been isolated during the nitration of o-xylene by nitric acid in 36 H. Suzuki, K. Nakamura, and M. Takeshima, Bull. Chem. SOC.Japan, 1971,44,2248. 37 D. H. Rosenblatt, W. H. Dennis,jun., and R.D. Goodin, J. Amer. Chem. SOC.,1973, 95, 2133. Hartshorn acetic anhydride,38 The structures of these adducts have been determined by spectroscopic methods and it is found that the characteristic absorptions ob-served in the 1H n.m.r.spectrum in the region 3.54.67 provide a particularly convenient method for detecting the formation of adducts in nitrating solutions. The adducts (33) and (34) readily decompose, either on heating or in aqueous acidic media, with loss of nitrous acid to give the aryl acetate (35). It therefore appears that the acetoxylation that accompanies the nitration of many methyl- benzenes in acetic anhydride occurs by an addition-elimination path~ay.3~3~~ In the case of p-xylene, the adduct* (36) isolated from the nitration mixture is found to decompose, by an apparently intramolecular mechanism, to give acetate (37) by a 1,2-migration of the acetoxy-gr0up.3~ OAC A number of adducts have now been isolated (Table 4) and still others have been detected as reaction intermediates.Two features appear to be common to all of the adducts: first, the nitro-group is always attached to a ring position substituted by a methyl group or other alkyl side-chain; secondly, only 1,4-adducts are formed and these occur as pairs of cis-and trans-isomers. This last observation suggests that a two-step mechanism is involved in adduct formation. The extent of acetoxylation accompanying nitration depends upon the structure of the aromatic substrate.39 Appreciable amounts of aryl acetates are produced with substrates such as o-xylene, p-xylene, and hemimellitene, but not with those such as toluene and rn-xylene. There is no reason to doubt that, in general, ring nitration occurs by conventional electrophilic substitution, and acetoxylation probably by addition-elimination. The proportion of acetoxylation to nitration can therefore be explained in terms of the extent to which the forma- tion of a a-complex by @so-attack competes with the formation of conventional o-complexes (Scheme 9).31p39Apparently the former process only provides an important reaction pathway if the @so-position concerned is activated by at least one ortho-or para-methyl group or an equivalent activating substituent.* In the following it is implicit that all adducts occur as cisltrans-isomers. sB D. J. Blackstock, J.R. Cretney, A. Fischer, M. P. Hartshorn, K. E. Richards, J. Vaughan,and G. J. Wright, Tetrahedron Letters, 1970, 2793. 89 D. J. Blackstock, A. Fischer, K. E. Richards, and G. J. Wright, Austral. J. Chem., 1973, 26,775. 181 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions Table 4 1,4-Adducts isolated from nitrations in acetic anhydride H OAc OAc H OAc H OAc Ref, a,b b b$02NP 02N@ @It II II II OAc H OAc AcO H 0 Me0 OAc Ref. d e f g Ref. h h i a D. J. Blackstock, A. Fischer, K. E. Richards, J. Vaughan, and G. J. Wright, Chem. Comm., 1970,641 ;b D. J. Blackstock, J. R. Cretney, A. Fischer, M. P. Hartshorn, K. E. Richards, J. Vaughan, and G. J. Wright, Tetrahedron Letters, 1970, 2793; A.Fischer and J. N.C Ramsay, J.C.S. Perkin ZI, 1973, 237; A. Fischer and A. L. Wilkinson, Canad. J. Chem., 1972, 50, 3988; e A. Fischer, C. C. Greig, A. L. Wilkinson, and D. R. A. Leonard, Canad. J. Chem., 1972, 50,221 1 ;f A. Fischer and D. R. A. Leonard, Canad. J. Chem., 1972, 50, 3367 ;g A. Fischer and D. R. A. Leonard, J.C.S. Chem. Comm., 1973, 300; h A. Fischer and C. C. Greig, J.C.S. Chem. Comm., 1973, 396; R. C. Hahn and M. B. Groen, J. Amer. Chem. SOC.,1973, 95, 6128. Addition-elimination is also thought to provide an important pathway for side-chain substitution. The nitration of 1,4-dimethylnaphthalene by nitric acid in acetic anhydride gives (38). If the reaction is carried out at -40 "Cand the reaction mixture quenched at that temperature with ammonia, the main product, isolated by low-temperature chromatography over alumina, is the adduct (39).40 This decomposes in acetic acid at 30 "Cto give 1,4-dimethyl-2-naphthyl acetate 40 A.Fischer and A. L. Wilkinson, Canad. J. Chem., 1972, 50, 3988. Hartshorn @-ring aatoxylation H OAcI __f ring nitration Scheme 9 (cf. behaviour of p-xylene) together with small amounts of the nitromethyl compound (38). In contrast, the adduct decomposes in acetic anhydride con- taining nitric acid to give (38) as the main product. The mechanism shown in Scheme 10 has been proposed for the side-chain nitration. Scheme 10 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions When the nitration is performed at -60 "Can adduct is obtained whose n.m.r.spectrum is similar to, but not identical with, that given by (39); the structure (41) has been s~ggested.~O As the temperature increases the adducts (39) and (41) are found to interconvert. It is relevant that nitration by nitric acid in deuteriated dichloromethane, in the absence of acetic anhydride, also gives the nitromethyl product (38), and the adduct involved is presumed to be (41). On the basis of these observations it is not certain which adduct, (39) or (41), is a necessary intermediate. Some uncertainty also surrounds the supposed intermediate (40). Contrary to earlier suggestions,4l there now appears to be no evidence for the formation of (40).40 This observation does not necessarily exclude such an intermediate, but it does imply that if involved it must be more reactive than other intermediates in the reaction pathway.The nitration of 1,4-dimethyInaphthalene in acetic anhydride,40~4~ in nitro- and in dichloromethane43 gives the side-chain nitro-compound. The nitration of hexamethylbenzene in acetic anhydride also gives mainly the nitro- compound, but in dichloromethane the nitrate is formed. As already mentioned, the different products might be produced by different reaction pathways, the nitrate by an intramolecular mechanism and the nitro-compound by an inter- molecular mechanism, e.g. the addition-elimination pathway, although other possibilities have been considered. The difference in behaviour between hexa- methylbenzene and 1,4-dimethylnaphthaIene then presumably arises from a structural effect on the partitioning of some reaction intermediate between the different pathways.Some clarification of the processes involved is required. 8 The Formation of Dienones and Quinones Nitration of the 4-X-o-xylenes and of the 5-X-hemimellitenes (42; X = Br, OAc, or OMe, R = H or Me) with nitric acid in acetic anhydride gives the corresponding nitrocyclohexadienones (43) in addition to conventional substitu- tion products.@ In the case of 3,4-dimethylanisole, the adduct (44) has been isolated and shown to decompose readily to 3,4-dimethyl-4-nitrocyclohexa-2,5-41 R. Robinson, J. Chem. SOC.(B), 1970, 1289. 4a R. Robinson and H. W. Thompson, J. Chem. SOC.,1932,2015.43 H. Suzuki and K. Nakamura, Bull. Chem. SOL.Japan, 1971,44,303. 44 D. J. Blackstock, M. P. Hartshorn, A. J. Lewis, K. E. Richards, J. Vaughan, and G. J. Wright, J. Chem. SOC.(B), 1971, 1212. Hartshorn dien-l-one.45 In the absence of the substituent X the only major products, apart from those of conventional substitution, are aryl acetates.39 Apparently the mode R RQ X X X OAc \ (44 (44) OAc of decomposition of the adduct (44)depends upon the nature of the substituent X. The small amounts of nitrodienones and cyclohexenones sometimes obtained as by-products in other nitrations46 can probably be accounted for in similar terms, although in some cases the adduct thought to be involved has a structure analogous to (41) rather than to (44).46a+d Dienones are frequently obtained as the products of electrophilic reactions involving phenols and their derivatives.14 The bromination of 2,4,6-trialkyl-phenols by molecular bromine in aqueous acetic acid, for example, gives high yields (70-90 %) of the corresponding bromocyclohexadienones (Scheme 1 l).47 The yield and the stability of the dienone both increase with the size of the R R Br R Br Scheme 11 ‘E.A.Fischer and D. R. A. Leonard, J.C.S. Chem. Comm., 1973, 300. 46 (a) H. Suzuki, M. Sawaki, and R. Sakimoto, Bull. Chem. SOC.Japan, 1972, 45, 1834; (b) H. Suzuki and K. Nakamura, J.C.S. Chem. Comm., 1972, 340; (c) H. Suzuki, M. Sawaki, and R. Sakimoto, Buff. Chem. SOC.Japan, 1972, 45, 1515; (d) H.Suzuki and K. Nakamura, Bull. Chem. SOC.Japan, 1972,45,1270; (e)A. J. M. Reuvers, F. F. van Leeuwen, and A. Sinnema, J.C.S. Chem. Comm., 1972,828. 47 A. A. Volod‘kin and V. V. Ershov, Bull. Acad. Sci.,U.S.S.R.,1962, 1039. Non-conventional Electrophilic Aromatic Substitutions and Related Reactions groups R; with bulky groups (e.g. t-butyl or t-pentyl) dienones may be isolated when only the 2-and 6-positions are substituted.48 Similar results have been obtained for nitration, and nitrocyclohexadienones are readily formed under various reaction condition^.^^ In many reactions dienones, although not actually isolated, are thought to be reaction intermediates.14950 Several modes of decomposition seem probable, including nuclear substitution, substitution with rearrangement, and side-chain substitution (Scheme 12).The last-mentioned reaction is referred to as the R$5JR Me E J R*R E RQR OHI Me E R9RCH& Scheme 12 quinobenzilic rearrangement and is usually considered to be intramolecular, but some evidence points to an intermolecular pathway under certain condition~.~~~~~1 Quinones are by-products in many electrophilic reactions,52 particularly in the nitration of polyphenolic compounds, a circumstance which may be connected with the fact that the reaction conditions employed are favourable for oxidative side reactions.14 However, the formation of quinones may be an indication that non-conventional electrophilic pathways are involved. Thus the major product from the nitration of 5-X-1,2,3-trimethoxybenzene(X = OMe or OAc) by nitric acid in acetic anhydride is the quinone (45)produced by decomposition of the 46 A.A. Volod'kin and V. V. Ershov, Bull. Acad. Sci., U.S.S.R.,1962, 1931. 4s V. V. Ershov and G. A. Zlobina, Bull. Acad. Sci., U.S.S.R.,1963, 1524. 6o (a) V. V. Ershov and A. A. Volod'kin, Bull. Acad. Sci., U.S.S.R.,1962, 1925; (b) V. V. Ershov and A. A. Volod'kin, ibid., p. 1935. 61 G. M. Coppinger and T. W. Campbell, J. Amer. Chem. SOC.,1953, 75, 734. 62 (a) L. I. Smith and F. J. Dobrovolny, J. Amer. Chem. SOC.,1926, 48, 1693; (b) H. H. Hodgson and J. Nixon, J. Chem. SOC.,1930, 1085; (c) C. C. Price and C. Weaver, J. Amer. Chem. SOC.,1939,61,3360; (d)K. Ley and E. Miiller, Chem. Ber., 1956,89,1402; (e)W.F. Gum, M. R. W. Levy, and M. M. Joullie, J. Chem. SOC.,1965, 2282. Hartshorn dienone (46).53 A likely mechanism for the formation of (45) therefore involves a non-conventional pathway in which the initial ITpso-attack of the nitrating agent occurs at a nuclear carbon atom bearing a methoxy-substituent (Scheme 13). eO eMeOQ OMe Me M Me0 NO, v M Me0 oNO,p e ___, II__3. #' X X X OAc J Scheme 13 9 Coupling Accompanying Electrophilic Substitution Substituted diphenylmethanes are produced in the nitration of pentamethyl-benzene by fuming nitric acid either in chloroform or in dichl~romethane,~~ and in the nitration of durene by nitronium hexafluorophosphate in nitr~methane.~S The formation of such compounds may be rationalized in terms of the same types of intermediate used to explain side-chain substitution.A speculative scheme, using pentamethylbenzene as an example, is outlined in Scheme 14; it is assumed that the intermediate involved in side-chain substitution may also react with substrate to give the substituted diphenylmethane. The positive hydroxylation of polymethylbenzenes by mixtures of peroxytri-fluoroacetic acid and boron trifluoride is believed to be an electrophilic process.56 In addition to the expected phenols, products of rearrangement, disproportiona- tion, and coupling are also frequently obtained. The formation of such products 63 B. A. Collins, K. E. Richards, and G. J. Wright, J.C.S. Chem. Comm., 1972, 1216. 64 H.Suzuki and K. Nakamura, Bull. Chem. SOC.Japan, 1970,43,473. 66 S. B. Hanna, E. Hunziker, T. Saito, and H. Zollinger, Helv. Chim. Acra, 1969, 52, 1537. 68 (a) C. A. Buehler and H. Hart, J. Amer. Chem. SOC.,1963, 85,2177; (b) H. Hart and C. A. Buehler, J. Org. Chem., 1964, 29, 2397; (c) H. Hart, C. A. Buehler, A. J. Waring, and S. Meyerson, J. Org. Chem., 1965, 30, 331. 187 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions can again be accounted for by assuming that non-conventional pathways are important. ozNoHzcv Scheme 14 Substituted diphenylmethanes have also been reported in the sulphonation of several polysubstituted benzenes under conditions favouring the Jacobsen re- arrangement.57 Thus treat men t of 1-chlor o-3,4,5,6-te trame t h ylbenzene with sulphuric acid at room temperature gives chloropentamethylbenzene (by dis- proportionation) and 3',4-dichloro-2,2',3,4',5,5',6-heptamethyldiphenylmethane. The formation of diphenylmethanes in this and other electrophilic reactions is of interest in connection with the suggestion that similar compounds may be inter- mediates in the disproportionation of primary alkylbenzenes under Friedel- Crafts conditions.58 Diarylalkanes have also been identified among the products of reaction of various monoalkylbenzenes with aluminium chloride59 and with antimony pentachloride .60 Substituted biphenyls are formed as by-products in certain nitrations of alkylbenzenes.61 The conditions necessary for the optimum yield of biphenyls appear to be the addition of 90 % nitric acid to the aromatic at low temperatures (ca.-25 "C).61b Under such conditions a small amount of 2-nitro-3',4,4',5-tetramethylbiphenyl is formed during the nitration of o-xylene, in addition to the expected nitroxylenes. The formation of these products of nuclear coupling may be an indication that non-conventional pathways are involved. 10 Rearrangements Accompanying Electrophilic Substitutions Rearrangements are of two general types, those involving a change in the structure of the aromatic substrate and those involving the electrophile and 67 H. Suzuki and Y. Tamwa, Chem. Comm., 1969,244. 68 A. Streitwieser and L. Reif, J. Amer. Chem. Soc., 1964, 86, 1988. R. M. Roberts, A.A. Khalaf, and R. N. Greene, J. Amer. Chem. SOC.,1964, 86, 2846. 6o P. Kovacic and A. K. Sparks, J. Org. Chem., 1963, 28, 972. I1I. Puskas and E. K. Fields, (a)J. Org. Chem., 1966, 31, 4204; (b) ibid., 1967, 32, 589. 188 Hartshorn occurring during the substitution. Among the first type are the well-known migrations of alkyl substituents that occur in Friedel-Crafts and Jacobsen reactions.62 Examples of methyl-group migrations occurring in other electrophilic reactions have been reported more recently in connection with non-conventional pathways. The action of cold fuming nitric acid on 1,2,3,4-tetramethyl-5,6-di-nitrobenzene gives a high yield of 2,3,6,6-tetramethyl-2,3,4,5-tetranitrocyclohex-4-enone.460 The mechanism proposed for this change (Scheme 15) assumes that the methyl migration occurs after the formation of a non-conventional a-complex. Such a pathway may also account for the 1,2-methyl migrations observed in the Scheme 15 positive hydroxylation of several polymethylbenzenes.63 Thus the dienone (47) is formed during the hydroxylation of 1,2,3,4-tetramethylbenzeneand (48) is formed from hexamethylbenzene.A more unusual type of behaviour is indicated by the isolation of the products (49) and (50) from the nitration of 2,4,6-tri-t-b~tyl-l-nitrobenzene.~The rearrangement-fragmentations(Scheme 16) are assumed to involve the a-complex (51) and to be competitive with conventional ring nitration. This mechanism implies that proton loss from (51) be relatively slow, for the non-conventional pathways to compete with the conventional substitution, and indeed the nitration shows a primary hydrogen isotope effect.65 H.J. Shine, ‘Aromatic Rearrangements’, Elsevier, Amsterdam, 1967. 69 H. Hart, P. M. Collins, and A. J. Waring, J. Amer. Chem. SOC.,1966, 88, 1005. a4 P. C. Myhre, M. Beug, K. S. Brown, and B. Ostman, J. Amer. Chem. SOC.,1971,93, 3452. 66 P. C. Myhre, M. Beug, and L. L. James, J. Amer. Chem. SOC.,1968,90, 2105. 189 Non-conventional Electrophilic Aromatic Substitutions and Related Reactions Scheme 16 The second general type of rearrangement that can occur involves the migra- tion of the electrophile from the nuclear position initially attacked either to another nuclear position or to a side-chain; the latter type of migration has already been discussed.The existence of a pathway involving rearrangement of the electrophile to another ring position may be overlooked, particularly if the final product can also be formed directly by a conventional substitution. How- ever, it is essential to estimate the contribution made by such pathways if reliable information about substituent effects and relative reactivities is to be obtained. The nitration of o-xylene by nitric acid in aqueous sulphuric acid appears to be a normal substitution, although the large change in the isomer proportions as the acidity of the sulphuric acid increases from 50% to 70% is difficult to explain.66 It has been shown that this change can be accounted for in terms of a contribution from the non-conventional pathway involving the intermediate (52) (Scheme 17).67 This intermediate can either be captured by the solvent to give phenolic products, or undergo an intramolecular 1,Zmigration of the nitro- group to give 3-nitro-o-xylene.The former pathway apparently predominates at the lower acidities, whereas rearrangement becomes dominant at higher acidities. The 4-nitro-o-xylene is produced only by conventional substitution. The ob- served change in isomer proportions (3-nitro : 4-nitro) is therefore to be associated with a solvent effect on the partitioning of (52) between solvent capture and rearrangement. It has also been suggested that some of the product (4-chloro-2-nitroanisole) from the nitration of p-chloroanisole is formed by a rearrangement involving a 1,3-migration of the nitro-group.5b Similar migrations also seem to be involved 66 R.G. Coombes and L. W. Russell, J. Chem. SOC.(B), 1971, 2443. 67 P. C. Myhre, J. Amer. Chem. SOC.,1972,94,7921. 190 Hartshorn &NO2 -> phenolic products '.-& A Scheme 17 in the nitration of (53).68 Little is known about the mechanism of these migra- tions except that they appear to be intramolecular,69~70 and possibly involve successive 1,2-nitro-shXts.70 11 summary Electrophilic attack on an aromatic system need not necessarily lead only to the products expected for a conventional substitution. Various non-conventional pathways are available and the examples discussed in this review show that for suitable substrates and reaction conditions products of a surprising variety may be formed.68 R. C. Hahn and M. B. Groen, J. Amer. Chem. SOC.,1973,95,6128. 69 A.Fischer and C. C. Grieg, J.C.S. Chem. Comm., 1973, 396. 'O G.A.Olah, H. C. Lin, and Y.K. Mo, J. Amer. Chern. SOC.,1972,94,3667. Non-conventional Electrophilic Aromatic Substitutions and Related Reactions The reactions considered can all be rationalized in terms of the initial forma- tion of a a-complex which can then react by any of a number of available path- ways. In outline these seem to be fairly well established, although many details require further clarification. The establishment of non-conventional reactions has important implications concerning quantitative studies of aromatic reactivity.The demonstration that products of conventional ring substitution can be formed by addition4imina- tion mechanisms and by a mechanism involving rearrangement of the electro- phile suggests that care should always be taken when interpreting results, even for seemingly conventional substitutions. The recognition that an electrophile can attack a substituted ring position also emphasizes the need to consider substituent effects at the @so-position.

ISSN:0306-0012    

DOI:10.1039/CS9740300167

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes By D. V. Fenby and L. G. Hepler* DEPARTMENT OF CHEMISTRY, UNIVERSITY OF OTAGO, DUNEDIN, NEW ZEALAND 1 Introduction On the basis of formidable experimental evidence it is believed that complex formation between like and/or unlike molecules takes place in many systems. This intermolecular complex formation is significant in many areas of chemistry and biochemistry and consequently has attracted considerable attention. Complexes are usually classified according to the intermolecular interaction primarily responsible for their formation. Hydrogen bond complex formation, first recognized explicitly by Latimer and Rodebushl in 1920, has been discussed comprehensively in a number of monograph^.^ -4 The theory of the hydrogen bond has been reviewed recently.5 Current interest in charge transfer complex formation began with the spectroscopic studies of Benesi and Hadebrand6 in 1948.The vast amount of research since this pioneering work has been discussed el~ewhere.~-12 The thermodynamic properties associated with complex formation are of fundamental importance. Reliable values are necessary for many chemical and biochemical applications, and also for testing such extrathermodynamic relationships as the Badger-Bauer correlation13 of hydrogen bond enthalpy *Mellor Visiting Professor, on leave from Department of Chemistry, University of Lethbridge. Alberta, Canada. 1 W. M. Latimer and W. H. Rodebush, J. Amer. Chem.Suc., 1920,42,1419.* ‘Hydrogen Bonding’, ed. D. Hadzi and H. W. Thompson, Pergamon Press, Oxford, 1959. G. C. Pimentel and A. L. McClellan, ‘The Hydrogen Bond‘, W. H. Freeman, San Francisco, 1960. 4 S. N. Vinogradov and R. H. Linnell, ‘Hydrogen Bonding’, Van Nostrand Reinhold Co., New York, 1971. 5 P. A. Kollman and L. C. Allen, Chem. Rev., 1972,72,283. 6 H. A. Benesi and J. H. Hildebrand, J.Amer. Chem. SOC.,1948,70,3978; ibid., 1949’71,2703. 7 R. S. Mulliken and W. B. Person, ‘Molecular Complexes’, Wiley-Interscience, New York, 1969.* R. Foster, ‘Organic Charge-Transfer Complexes’, Academic Press, London, 1969. M. A. Slifkin, ‘Charge Transfer Interactions of Biomolecules’, Academic Press, London, 1971. G. Briegleb, ‘Elektronen-Donator-Acceptor-Komplexe’,Springer-Verlag, Berlin, 1961.L. J. Andrews and R. M. Keefer, ‘Molecular Complexes in Organic Chemistry’, Holden- Day, San Francisco, 1964. l2 J. Rose, ‘Molecular Complexes,’ Pergamon Press, Oxford, 1967. l3 R. M. Badger and S. H. Bauer,J. Chern.Phys., 1937,5,839. Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes with i.r. shift and the correlation14 of charge transfer bond energy with the enhancement of dipole moment. A wide range of techniques, discussed in the books referred to above, has been applied to investigations of the thermodynamics of complex formation. Following a review of the data available in 1959, Pimentel and McClellan3 commented that ‘. . . the most outstanding conclusion is that additional and systematic studies are needed.’ With the substantial number of investigations since then and the ever greater appreciation of some of the complicating factors, the situation has improved.But even in 1969 Fosters wrote that ‘none of the methods currently used yields results in which, for all systems, high confidence can be placed.’ Given this situation and the importance of the thermodynamics of complex formation, all new methods for their investigation deserve scrutiny. In this paper we discuss a number of calorimetric methods developed in recent years. As these possess certain advantages over traditional methods, they represent a significant contribution to the investigation of complex formation in solution, Because they have been introduced only in the last few years, calorimetric methods have received brief mention only in the more recent4ssJ of the general references given above.2 Thermodynamic Properties of Complex Formation At first we consider the formation of only one complex of 1 :1 stoicheiometry between unlike molecules as represented by A+B+AB (1) It is customary (and reas~nable~~~~J~) to take all activity coefficients to be unity so that the equilibrium constant for this reaction can be expressed in terms of the equilibrium composition of the solution. For example, we commonly express the equilibrium composition in terms of concentrations (molarities) and thence obtain an equilibrium constant that we represent by Kc.Equilibrium compositions may also be expressed in terms of mole fractions and molalities, thus leading to equilibrium constants represented by Kzand Km,respectively.It should be emphasized that the assumption of all activity coefficients equal to unity (or the nearly equivalent assumption of a ratio of activity coefficients equal to unity) is common to most methods used in investigation of complex formation in non-electrolyte systems. In particular, we emphasize that similar assumptions about activity coefficients are used in the various spectroscopic methods and also the calorimetric methods discussed here. For a dilute solution of A and B in some solvent S, the most common situation experimentally, the K values and dGovalues are related as follows : 14 H.Ratajczak and W. J. Orville-Thomas,J. Chem. Phys., 1973,58,911. 15 K. Denbigh, ‘The Principks of Chemical Equilibrium’, 2nd edo., Cambridge University Press, London, 1966. Fenby and Hepler where Ms is the molar mass of solvent S and Y,,S is the molar volume of S. The standard enthalpy of complex formation at infinite dilution (AH") is given by the van't Hoff equation, AH" = -R(-)alnKm al/T P It follows from equation (2) that AH" is related to the temperature coefficients of K,, Km,and Kc as follows : where as is the thermal expansivity of the solvent. The term RT2~~16J7 in equation (6)is mistakenly omitted from many calculations; with as M K-1 for 'typical' solvents at 300 K, RT2as M 1 kJ mol-l (i.e., often about 10%of AH").The standard entropies of complex formation are related by 3 Determination of Thermodynamic Properties The thermodynamic properties of complex formation have usually been obtained from measurements of some equilibrium constant and its variation with tem- perature. The methods used have been revie~ed.~ -497-12 The equilibrium constant has most frequently been obtained from spectro- scopic measurements (i.r., u.v., visible, n.m.r.) carried out over a range of com- positions of one or both of the reactants. Thermodynamic procedures such as distribution between immiscible liquids and measurement of various colligative properties, dielectric studies, polarography, etc., have been used to a much lesser extent. In all of these methods it is usual to incorporate the assumption that activity coefficients may be approximated to unity.As most measurements have involved dilute solutions of A and B in some 'inert' solvent, the molarity equilibrium constant Kchas been most widely used. (This is often experimentally more convenient than Km or Kzand is entirely 'proper' when AH" is calculated correctly as noted above.) Perusal of published equilibrium constants suggests that in many cases little confidence can be placed in the results. Reported equilibrium constants for a l6 E. J. King, 'Acid-Base Equilibria', Pergamon Press, Oxford, 1965. l7 E. M. Woolley, J. G. Travers, B. P. Erno, and L. G. Hepler, J. Phys. Chem., 1971,75, 3591. 195 Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes particular complex frequently differ by substantially more than the assigned uncertainties; for example, in Table 1 we list equilibrium constants for the Table 1 Equilibrium constants for the formation of the 1,3,5-trinitrobentene- hexamethylbenzene complex at 293 K Solvent KJdm3 mol--1 Reference cyclohexane 17.50 i-0.20 18 13.5 k 0.4 19 carbon tetrachloride 4.87 k 0.08 18 5.7 & 0.3 19 7.10 k 0.05 20 chloroform 0.92 k 0.10 18 0.76 k 0.05 19 0.81 21 1,3,5-trinitrobenzene-hexamethylbenzenecomplex at 293 K obtained from U.V.studies. The situation has been improved by critical analyses, such as that of Person,22of the various spectroscopic procedures. With careful experimental work and data analysis, reliable equilibrium constants can be obtained.Having measured equilibrium constants over a temperature range, the value of AH" can be obtained by using equations (4)-(6). (The values of AH" com-monly lie between -10 and -30 kJ mol-l.) Since differentiation procedures always result in a loss of precision, AH" (and AS,") will be less precisely known than InK,. King16 has described a thorough and rigorous analysis of the relationship between random errors in equilibrium constants and the thermo- dynamic properties derived by differentiation. Even with equilibrium constants obtained from careful spectroscopic studies, the propagation of errors can lead to standard deviations in AH" of 1-2 kJ mol-l. As King's statistical analysis excludes systematic errors, the standard deviations obtained represent lower limits for the total uncertainties.In view of the above, it is not surprising that the discrepancies between published results for AH" are often large, sometimes exceeding 100%. [Extensive comparisons of K and AH" values obtained in different ways are given in refs. 27, 28, 32, 35, 36, 39,44, and 64, discussed later in this review.] A number of methods have been advanced recently for evaluation of AH" (and usually K) from results of calorimetric measurements. With most of these procedures, AH" is obtained from calorimetric results at a single temperature and without recourse to the van't Hoff equation. It thus appears that these calorimetric methods offer significant advantages over other approaches which C.C. Thompson and P. A. D. de Maine, J. Phys. Chem., 1965,69,2766. l9 R.Foster, J. Chem. Soc., 1960, 1075. 2o G. Briegleb and J. Czekalla, Z. Elektrochem., 1955, 59, 184. 21 C. E. Castro, L. J. Andrews, and R. M. Keefer, J. Amer. Chem. SOC.,1958,80,2322. W. B. Person, J. Amer. Chem. SOC.,1965,87,187. 196 Fenby and Hepler rely on differentiation procedures. Lambert~~~ has provided a general review of calorimetric investigations of molecular complexes in solution while Chris- tensen et a1.24-26have reviewed applications of titration calorimetry to a variety of reactions in solution. Our discussion is intended to supplement these earlier reviews. It is convenient to divide the calorimetric methods into those in which complex formation takes place in dilute solution in some 'inert' solvent and those in- volving no solvent. 4 Complex Formation in Inert Solvents In this section we consider the complex formation represented by the equation A(S) + B(S) + AB(S) (8) in which (S) indicates that the preceding species is in dilute solution in the solvent S.Following usual practice, we will use the molarity equilibrium constant Kc. With the notation given in Table 2, Kc is given by Table 2 Summary of symbols used Stoicheiometric number of moles Stoicheiometric mole fractions Stoicheiometric molarities Equilibrium numbers of moles Equilibrium mole fractions Equilibrium molarities This equation can be rearranged to C' -(CA + CB + Kc-l)~+ CACB = 0 (10) Calorimetric methods involve measurement of the enthalpy change Q associated with formation of n mol of complex AB according to the reaction represented by equation (8).This measured Q is related to the standard enthalpy of complex formation at infinite dilution, AH",by Q = AHo = cVAH" (1 1) in which Vrepresents the total volume of the solution. The enthalpy change of 23 L. Lamberts, Ind. chim. belge, 1971, 36, 347. 24 J. J. Christensen, J. Ruckman, D. J. Eatough, and R. M. Izatt, Thermochim.Ada, 1972,3, 203. 25 D. J. Eatough, J. J. Christensen, and R. M. Izatt, Thermochim.Acra, 1972,3,219. 26 D. J. Eatough, R. M. Izatt, and J. J. Christensen, Thermochim.Am, 1972,3,233. Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes interest is obtained experimentally as a difference between a principal experi- mental enthalpy change and effects extraneous to the complex formation, such as enthalpies of solution or dilution.A widely used procedure is illustrated by the following cycle: 0 S AB (S,dilute).1 *II A (S,dilute) + B (S,dilute) -/ A (pure) represents either pure solid A or pure liquid A. The desired enthalpy change Q for reaction (8) is obtained as the difference between the two experi- mental enthalpy changes QI and QII: Q = QI -QII In some procedures, A(pure) is replaced by A(S, concentrated). All of the techniques to be discussed in this section are based on a combination of equations (9) or (10) withequation (1 1).In eachcase we derive the fundamental equation and indicate the method used in its application. No attempt has been made to list all of the systems to which these methods have been applied, but some key references are given. Equation (10) is a quadratic equation in c and can be solved to give [The other root of equation (10) is unacceptable.] Using a Kc value determined independently (for example, from spectroscopic measurements), equation (13) permits calculation of c values for a series of stoicheiometric molarities CA and CB. These values of c, together with the cor- responding Q values, can be used in equation (ll), according to which a plot of Q vs. cV will be linear with slope AH" (and zero intercept).This procedure was introduced by Arnett and co-worker~.~~ The results obtained have been com- paredZ7s28with those obtained using another calorimetric method (to be dis- cussed in the next section) and with spectroscopic results. While dependent upon other studies for values of Kc, this method does not involve trial-and-error calculations as do most of the techniques capable of yielding both Kc and AH" from calorimetric measurements alone. The AH" value obtained is very sensitive to variations in Kc when the latter is small. Methods advanced to obtain simultaneously both Kc and AH" from equations *' E. M. Arnett, T. S. S. R. Murty, P. von R. Schleyer, and L. Joris, J. Amer. Chem. SOC., 1967, 89, 5955. 28 E. M. Arnett, L. Joris, E. Mitchell, T.S. S. R. Murty, T. M. Gorie, and P. von R. Schleyer, J. Amer. Chew, SQC.,1970,92,2355, Fenby and Hepler (10) and (11) have involved trial-and-error methods of calculation. The limits of such methods arising from the propagation of errors have been discussed by Cabani and Gian~~i,~~ who have shown that a good fit between calculated and experimental data can be compatible with fictitious values of Kc and AH" if the experiments are conducted with a constant concentration of one of the reactants. Procedures to reveal systematic errors are listed and a weighting procedure, which takes into account the different propagation of errors for different experimental conditions, is proposed.29 Equations (11) and (13) can be combined to give Using this equation, Lamberts and Zeegers-Huy~kens~~ and others31 have obtained both Kc and AH" by way of trial-and-error calculations.For each set of experimental results (CA, CB, V, Q), AH"values are calculated from equation (14) for a range of values of Kc.From plots of AH"vs. CA or CB for each of these Kc values, one selects as the 'best' value of Kc that which gives a constant value of AH" (ix.,a line of zero slope). A method introduced by Bolles and drag^^^ and extensively used by Drago and ~o-workers~~9~~ involves the following combination of equations (10) and (11): Again both Kc and AH" are obtained by trial-and-error calculations. For each set of experimental results (CA, CB, V, Q), AH"values are calculated for a range of values of Kc,and Kc-l is plotted against AH".The curves thus obtained, one for each set of experimental results, should ideally intersect at a single point, this being the unique solution of equation (15).A statistical procedure for this method of data analysis35 has been applied. Comparisons between calorimetric and spectroscopic results have been given.32*35 Another approach, making use of calorimetric measurements at two tempera- tures, has been proposed by Neerinck et 01.~~Ignoring RT2as in equation (6) and assuming AH" to be temperature independent over the temperature range of interest, one obtains 29 S. Cabani and P. Gianni,J. Chem. SOC.(A), 1968,547. 30 L. Lamberts and F. Zeegers-Huyskens,J. Chim.phys., 1963,435. 31 L. Abello and G.Pannetier, Bull. SOC.chim. France, 1967, 3752. 32 T. F. Bolles and R. S. Drago, J. Amer. Chem. SOC.,1965,87, 5015. 33 R. M. Guidry and R. S. Drago,J. Amer. Chem. SOC., 1973,95,759. 34 F. L. Slejko and R. S. Drago, Inorg. Chem., 1973,12, 176. T. D. Epley and R. S. Drago, J. Amer. Chem. SOC.,1967,89,5770. 36 D. Neerinck, A. VanAudenhaege,L. Lamberts, and P. Huyskens,Nature, 1968,218,46 I. 199 Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes where Kc(T)is the molarity equilibrium constant at temperature T. Substituting equation (9) into equation (16) and rearranging gives on the basis of experiments arranged so that the stoicheiometric molarities CA and CB are the same at temperatures 7'1 and Tz.In this equation c1 and c2 re-present the molarities of the complex at temperatures TI and T2, respectively.Substituting c = Q/VAH" from equation (11) into equation (17) gives an equation from which AH" can be obtained by trial and error. Having obtained AH", equation (9) leads to Kc. Results from this approach were compared36 with those obtained by the calorimetric method of Bolles and DragoS2 discussed above; the AH" values agree well but there are considerable discrepancies between the Kcvalues. All of the above methods capable of yielding both Kc and AH" involve trial- and-error methods of calculation. A new and potentially better procedure, which we propose here, avoids the necessity for trial-and-error calculations. When the stoicheiometric molarities c A and CB (at constant temperature) are chosen so that their sum is always the same we define a constant x by Combining equations (15) and (18) and rearranging gives VCACB ---.-1Q +-x Q (AH)2 V AH" according to which a plot of VCACBIQvs.Q/V should be linear with slope -~/(LW")~and intercept x/dH". From such a plot we can therefore obtain AH" and, if CA and CB are chosen so that (CA + CB)K,is not considerably greater than unity, Kc can also be evaluated. We believe that this approach could prove to be preferable to those dependent upon trial-and-error methods of calculation. A thorough test of this method, which appears to be especially suited to flow calorimetry, should involve evaluation of K, and AH" from measurements at different values of (CA + CB).5 Complex Formation Studies without Inert Solvents In this section we will be concerned with the mixing of pure liquid A and pure liquidB to give a liquid mixture, all at constant temperature and pressure: The influence of complex formation reactions, such as that represented by equation (l), on the thermodynamic properties of liquid mixtures has long been Fenby and Hepler recognized but remains difficult to assess. At the beginning of this century D0lezalek3~ and his followers tried to account for all deviations from ideal solution behaviour in terms of chemical equilibria between the like and/or unlike molecules. This approach was soon shown to be chemically unrealistic and numerically inadequate for many systems.But it has also been shown that complex formation as represented by chemical reaction equations such as (1) is the principal interaction in many solutions. Properties of these solutions may therefore be realistically and usefully interpreted in terms of the thermodynamics of complex formation as described below. In this section it is convenient to use the mole fraction equilibrium constant Kz (again assuming the activity coefficient ratio to be unity): Combining this assumption with the Gibbs-Duhem equationl5 leads to The mixture of the three chemical species (A, B, AB) therefore behaves as an ideal solution and all deviations from ideality observed experimentally for a mixture of NA (stoicheiometric) mol A with NB (stoicheiometric) mol B can be attributed to the formation of the complex AB.In the present situation, the common assumption of activity coefficient ratio equal to unity is therefore equivalent to assuming that the system is an ‘ideal associated solution’.38 The enthalpy change associated with the mixing process represented by equation (20) is supposed to arise entirely from the formation of the complex. An alternative formulation of the discussion above is useful in that it leads directly to a relationship between the enthalpy of mixing and the thermodynamics of complex formation.39 In this formulation we discuss the thermodynamics of mixing in terms of ‘physical’ (non-complexing) and ‘chemical’ contributions, with the latter due to complex f~rmation.~~-~~ According to this model, the molar enthalpy of mixing, AH,, can be expressed as in which AHmphys and AHmchemrepresent the molar ‘physical’ and ‘chemical’ contributions.For systems with large exothermic enthalpies of mixing (-AHm greater than ca. 1.5 kJ mol-1 at x = 0.5), it is usually reasonable to neglect 37 F. Dolezalek,Z. phys. Chem., 1908,64, 727. 38 I. Prigogine and R. Defay, ‘Chemical Thermodynamics’, translated by D. H. Everett, Longmans Green and Co., London, 1950. 39 L. G. Hepler and D. V. Fenby, J. Chem. Thermodynamics, 1973,5,471. 40 R. Anderson, R. Cambio, and J. M. Prausnitz, J. Amer. Inst. Chem. Eng., 1962,8,66. 41 I. D. Watson and A. G. Williamson,J. Sci. Id. Res., India, 1965,24,615. 4* R. L. Scott, ‘Weak Complexes: Molecular Models and Thermodynamic Evidence’, paper read before the Chemical Society, Exeter, 1967.201 3 Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes AHmphYsin the above expression. On this basis, we proceed to relate molar enthalpies of mixing to the equilibrium constant Kzand the enthalpy of complex formation AH".To do so we use n =Q/AHOfrom equation (1 1) and which relates the measured enthalpy Q and stoicheiometric amounts of A and B to the molar enthalpy of mixing that is usually reported. Using the notation summarized in Table 2 with NS =0, equation (21) can be written as n(NA +NB -n)K, = (NA -n)(NB -n) Substitutingn =Q/AH"from equation (1 1) in (25) and rearranging yields NANB Kz +1 Kz +1 Q =-Kz(AHo)2 Q +(NA +NB)-KzAH" Further substitution of Q =AHm(NA +NB)from (24) in (26) and use of the definition of stoicheiometric mole fraction leads to According to this equation, a plot of XAXB/AHm vs.AHm should be linear; Kzand AH" can be evaluated from the slope and intercept. We have applied equation (27) to investigation of complex formation between chloroform and triethylamine39 and between chloroform and various ethers.*3 It is also possible to relate differential or partial molar enthalpies of solution to Kz and AHo in ways that permit useful evaluation of these latter quantities from results of calorimetric measurement^.^^ We begin by identifying ZA and ZBas partial molar enthalpies of solution of A and B, respectively. A formal definition is given by First, we will be concerned with partial molar enthalpies of solution at infinite dilution (ZAO and ZB'). These limiting partial molar enthalpies of solution may be obtained45 almost directly from 'small increment' solution calorimetry or 43 N.F. Pasco, D. V. Fenby, and L. G. Hepler, Canad. J. Chem., in the press. 44 T. Matsui, L. G. Hepler, D. V. Fenby, J. Phys. Chem., 1973,77,2397. H. C. Van Ness, 'Classical Thermodynamics of Non-Electrolyte Solutions', Pergamon Press, New York, 1964. Fenby and Hepler (less accurately) from 'integral' enthalpies of mixing such as those already represented by AH,. We have recently compared z" values obtained by way of these two appro ache^.^^ In the limit NA -+ 0 it follows from equation (25) that Kz+ n/(NA -n).Consequently, in this limit, the fraction of A that is complexed is given by As all enthalpy changes have been attributed to the formation of the complex, or AH" = (1 + Kz)1~"/Kz (31) Because the system A + B is symmetrical when AB is the only complex, we also have Equation (26), a quadratic in Q, can be solved and the result differentiated with respect to NAto give a general expression for ZAas previously described.44 In the limit NA-+0 this general expression reduces to equation (30), which has already been derived in a simpler way. In the case NA = NB, the general ex- pression reduces to44 in which 20.5represents the partial molar enthalpy of solution of either A or B in solution with XA = XB = 0.5.We have applied equations (31)--(33) to investigation of complex formation in the chloroform-triethylamine system in two ways.44 First, calorimetrically determined zvalues have been combined with previously reported Kzvalues to yield the desired AH". Second, these simultaneous equations have been solved (either graphically or analytically) to yield values of both AH"and Kz. For those cases in which Kz% 1, equations (29), (31), and (32) reduce to and LAo z AH" (35) In such cases it follows from equation (34) that (almost) all of the A molecules (in infinitely dilute solution in B) are involved in complex formation, The application of equation (35) to evaluation of AH", while appealingly simple, Calorimetric Investigations of Hydrogen Bond and Chrge Transfer Complexes requires caution as it is only applicable if the equilibrium constant is sufficiently large.Arnett et aZ.27928have applied equation (35) to a large number of systems for which it appears to be a valid approximation. In this work pure A was a solid, making it necessary to obtain ZA' indirectly. This was done by introducing an 'inert' solvent S and a 'model compound' M, a compound similar to A but unable to undergo complex formation with B (e.g. A = phenol, M = anisole or fluorobenzene). The value of ZA' was obtained by using the expression in which ZX'(~)(Y) represents the partial molar enthalpy of solution of solid X in infinitely dilute solution in the liquid Y.Although the results reported by Arnett et al.27128 are in good agreement with the results obtained using their method discussed earlier in this review, [equations (11) and (13)], a critical analysisg0 has indicated that the AH" values obtained are strongly dependent on the choice of both the 'inert' solvent and the 'model compound'.Duer and Bertrandg6 have modified equation (36) to determine the difference in the standard enthalpies of formation of two complexes, AB and AC (both in the same solvent): This procedure eliminates the effect of the 'inert' solvent. The AH' differences obtained were found to be nearly independent of the choice of 'model compound' M; hence this appears to be a useful procedure for obtaining relativeAH' values. The approximation represented by equations (34) and (35) has also been implied in other investigation^^^ -4Q in which its validity is questionable.6 Further Complex Formation In many systems complex formation in addition to that represented by equation (1) takes place. Some of the methods discussed in the last two sections can be extended to these more complicated situations. Systems in which AB and AB2 complexes are present appear to be quite common. McGlashan and Rast~gi,~~ assuming that such systems behave as ideal associated solutions, have proposed a method whereby the two equilibrium constants can be obtained from vapour pressure measurements. These equili- brium constants, or those obtained in any other way, can then be combined with results of calorimetric measurements to yield the standard enthalpies of complex 46 W.C. Duer and G. L. Bertrand,J. Amer. Chem. Soc., 1970,92,2587. 47 S. Murakami, K. Amaya, and R. Fujishiro, Bull. Chem. SOC.Japan, 1964,37, 1776. 48 S. Murakami, M. Koyama, and R. Fujishiro, Bull. Chem. SOC.Japan, 1968,41,1540. 49 T. J. V. Findlay, J. S. Keniry, A. D. Kidman, and V.A. Pickles, Trans. Furaday Soc., 1967, 63,846. 50 M. L.McGlashan and R.P.Rastogi, Trans. Fafaday SOC.,1958,54,496. 204 Fenby and Hepler formation.50-63 The method based on partial molar enthalpies of solution that was discussed in the last section has been extended44 to allow the standard enthalpies of formation of both the AB and AB2 complex to be evaluated from previously reported equilibrium constants and calorimetric ZA" and ZB'values.Calorimetric results have also been used in investigations of various self- association reactions that can be represented by nA(S) + An(S) (38) in which (S) indicates that monomeric A and associated species An are in solution in some solvent S. We first call attention to the pioneering investigation of S~hellman.~4 He used osmotic coefficients for evaluation of the equilibrium constant for dimerization of urea in aqueous solution, and then combined the equilibrium constant with enthalpies of dilution for evaluation of AH" of dimerization. Subsequent investigations by Stokes have provided further experimental results and con- siderably more extensive theoretical analysis of self-association of urea in aqueous solutions5 and in various other solvents.56 We also call attention to the work of Gill et aZ.,57-60 who have derived useful relationships between enthalpies of dilution and thermodynamic functions for self-association reactions such as represented by equation (38).They have investigated several aqueous systems of biochemical interest. Many investigations2 -* have provided convincing evidence that carboxylic acids form hydrogen-bonded dimers in more or less 'inert' solvents. On the assumption that only monomer-dimer equilibria need be considered, i.r. and other non-calorimetric measurements have led to many equilibrium constants for dimerization reactions in dilute solutions. Although some AH" values have been calculated by way of the van't Hoff equations (4)-(6), uncertainties appear to be large and there are substantial disagreements between results of different investigators.Woolley et al.61,s2have measured enthalpies of dilution of acetic acid in carbon tetrachloride and benzene and of some chloro-substituted acetic and propionic acids in carbon tetrachloride. They have calculated Kc and AHo values for dimerization reactions, with Kc values in reasonable agreement with those derived from earlier non-calorimetric investigations. The AH" values from these calorimetric investigations6lPB2 are probably more accurate than those derived from equilibrium constants at several temperatures. 61 E. R. Kearns, J. Phys. Chem., 1961,65,314. 6p K. W. Morcom and D. N. Travers, Trans. Farday SOL,1965,61,230.53 D. V. Fenby and L. G. Hepler, J. Chem. Thermodynamics, 1974, 6, 185. J. A. Schellman, Compt. rend. Trav Lab. Carlsberg, 1955,29,223. bb R. H. Stokes, Austral.J. Chem., 1967,20,2087. 6o D. Hamilton and R. H. Stokes, J. Solution Chem., 1972,1,223. 67 S. J. Gill, M Downing,and G F. Sheats, Biochemistry, 1967,6,272. 58 R. Stoesser and S. J. Gill. J. Phys. Chem., 1967,71, 564. 59 S. J. Gill and E. L. Farquhar, J. Amer. Chem. SOC.,1968,90,3089. 6o E. L. Farquhar, M. Downing,and S. J. Gill, Biochemistry, 1968,7, 1224. N.S. Zaugg, S. P. Steed, and E. M. Woolley, Thermochim.Acta, 19723. 349, 62 N. S. Zaugg, L. E. Trejo,and E. M. Woolley, Thermochim. Acra, 1973,6,293. Calorimetric Investigations of Hydrogen Bond and Charge Transfer Complexes Although it appears well established that monomers and dimers are the principal carboxylic acid species in dilute solutions, recent i.r.measurements and ‘factor analysis’ of the results by Bulmer and ShurvelP provide evidence that larger associated species are also present in all but the most dilute solutions. At present we have no quantitative information about these larger species, nor do we know how much error in thermodynamic properties of dimerization reactions results from failure to consider trimers, etc. Factor analysis of results from spectroscopic investigations of self-association of ROH compounds (discussed below) is certainly desirable, and it may be that this important data treatment technique can be extended to apply to enthalpies of dilution.Numerous investigations have provided convincing evidence for self-associ- ation of ROH compounds in various solvents, but there is widespread dis- agreement as to identification of predominant associated species, and few reliable data for well-defined association reactions are available. We now discuss phenol as typical of this class of compounds. Many, but not all, of the i.r. and other spectroscopic investigations (see refs. 17 and 64 for a review) of phenol in various ‘inert’ solvents indicate almost conclusively that self-associated species larger than dimers are important even in dilute solutions. This conclusion is supported by data for the distribution of phenol between water and carbon tetrach10ride.l~ Finally, enthalpies of dilu- ti0n1~~6~also show that there must be self-associated species larger than dimers.Calculationsl7~64 based on these calorimetric results indicate that trimers may be the predominant self-associated species in most solutions [n = 3 in the reaction represented by (38)], a result that is in agreement with some spectro- scopic investigations, But a recent thorough i.r. investigation by Whetsel and Lady65 suggests that dimerization and stepwise further self-association to various n-mers is the most realistic representation. Although there seems to be no reason to question the reliability of either the calorimetric or spectroscopic experimental results, it is obvious that one or both of the conflicting inter- pretationsl7~64.65 must be in error.Possible means of resolving present un- certainties have been s~ggested.~~ 7 Conclusions Investigations based on methods discussed in this paper have shown that it is possible to obtain AH” or in some cases both K and AH” for reactions of type represented by equations (1) and (8) from results of calorimetric measurements at a single temperature. The most important general advantage of the calori- metric methods isthat they permit evaluation of AH” without recourse to the van’t Hoff equation and the uncertainties derived from differentiation of experi- mental results. It is probable that the ‘best’ thermodynamic functions for 69 J. T.Bulmer and H. F. Shurvell,J. Phys. Chem., 1973,77,256. 64 E.M.Woolley and L.G. Hepler, J. Phys. Chem., 1972,76,3058. 65 K.B.Whetsel and J. H. Lady, ‘Spectrometry of Fuels’, Plenum Press, New York, 1970, pp. 259-279. 206 Fenby and Hepler reactions of type (1) and (8) can be obtained as a result of combination of equilibrium constants determined by spectroscopic (or other non-calorimetric) means with calorimetric enthalpies. For more complicated reaction schemes such as those which involve complexes represented by AB2 or An,it is generally impracticable to obtain all of the desired thermodynamic functions from only one kind of measurement. In these cases it is especially advantageous to combine calorimetric results with those derived from spectroscopic measurements, vapour pressures, or other non-calorimetric measurements. We are pleased to acknowledge the assistance of the Donors of the Petroleum Research Fund, administered by the American Chemical Society, the National Research Council of Canada, and the Research Committee of the New Zealand Grants Committee.

ISSN:0306-0012    

DOI:10.1039/CS9740300193

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Chemistry of the Production of Organic Isocyanates By H. J. Twitchett RESEARCH DEPARTMENT, ICI LTD., ORGANICS DIVISION, BLACKLEY, MANCHESTER 1 Introduction The manufacture of organic isocyanates has grown at a rapid rate in the past twenty years and world productive capacity now stands at over 1OOO OOO tons per annum. The reactivity of isocyanates and the versatility of derived polymers ensures increasing demand for them and new plants continue to be erected. The complexity of the plants and the hazards associated with the raw materials and the products themselves have restricted production to the larger chemical companies whose resources are adequate to deal with many of the chemical and engineering problems arising in manufacture. Much of the information relating to isocyanate production is described only in patent specifications. The large number of patents issued and the breadth of the claims contained in them makes it difficult for scientists not directly involved in manufacture to get a clear idea how processes are operated.At least 90% of total world production is accounted for by two products, tolylene di-isocyanate (TDI) and diphenylmethane di-isocyanate (MDI). Small amounts of other di-isocyanates are made, including 1,6-hexamethylene di- isocyanate, 1,haphthalene di-isocyanate, xylylene di-isocyanate, 4,4’-dicyclo- hexylmethane di-isocyanate and 3-isocyanatomethyl-3,5,5-trimethylcyclohexyl isocyanate. Also manufactured are various monoisocyanates, including methyl, n-propyl, n-butyl, cyclohexyl, phenyl, and 4-chloro- and 3,ddichloro-phenyl isocyanates. The di-isocyanates are used to make urethane polymers such as foams, elastomers, surface coatings and fibres.The monoisocyanates are used for substituted ureas and carbamates important as herbicides and crop protection agents. Some of the basic di-isocyanates are converted on a commercial scale into derived products containing residual isocyanate groups. Thus TDI is trimerized and sold as a solvent solution of its trifunctional isocyanurate (I), and hexa- methylene di-isocyanate is converted into the substituted biuret (2). Di-iso-cyanates may also be treated with deficiencies of diols (HOXOH) to give non- volatile bis-carbamates (3) containing terminal isocyanate groups. The manufacture of these latter products is not discussed in this article, which is restricted to describing the chemistry of the processes used for the main commercial di-isocyanates.2 Preparation of Isocyanates Isocyanates can be made in many ways; the classical laboratory routes, the Chemistry of the Production of Organic Isocyanates OCN NCO Me-&&&Me oc, ,co OCN(CH&NCONH(CH2)'eNCOI(LoMe CONH(CHa)eNCO (1) (2) OCN *R*NHCO*O*X RvNCO*OCONH* (3) Curtius, Hoffman, and Lossen rearrangements (reactions 1, 2, 3), which may involve a nitrene as an intermediate, are not at all satisfactory for large-scale operation, NaN, -N* RCOCl__+ RCON3 +RCON -+ RNCO (1) NaOBr -HBr RCONHz -+ RCONHBr +RCON --+ RNCO (2) NH,OH -Ha0 R1COOR2+R20H + RTONHOH _jRlCON ---+ RlNCO (3) The use of azides in the Curtius reaction is hazardous and the utility of the Hoffmann and Lossen rearrangements is limited to preparation of aliphatic isocyanates as, in the aqueous media employed, aromatic isocyanates react readily with water to form substituted ureas.t-Butyl hypochlorite can be used for non-aqueous Hoffmann rearrangements but is too costly for technical applications. In practice, only phosgenation of a primary amine is important (reaction 4). This route, first used by Hentschell in 1884, enables a wide range of aliphatic and aromatic isocyanates to be obtained from the corresponding amines. The laboratory preparation of numerous examples by this method has been described by Siefken.2 COCI, -HCl RNHz __+ RNHCOCl +RNCO (4) W.Hentschel, Ber., 1884, 17,1284. W. Siefken, Annalen, 1949, 562, 75. 210 Twitchett 3 The Phosgenation of Primary Amines Relatively little detailed attention has been given to phosgenation processes other than those for tolylene di-isocyanate and diphenylmethane di-isocyanate, but it is generally considered that the mechanism of the phosgenation reaction and the pattern of impurities resulting from side reactions is similar for many amines, both aromatic and aliphatic in nature. Phosgenation of a primary amine is usually carried out by dissolving it in an inert solvent before treating with phosgene. As an alternative the amine may first be converted into a suspension of a salt by passing dry hydrogen chloride or carbon dioxide into the solution before contacting with phosgene.Improved results are sometimes claimed for this procedure although, as a rule, phosgena- tion of pre-formed salts is slower than direct reaction of the amine. Relatively volatile monoamines, e.g. cyclohexylamine or aniline, can be phosgenated quite satisfactorily using a vapour-phase method; an inert gas carrier is sometimes used. Typical processes have been patented by ICI and Ba~er.~ Another variation is sometimes referred to as 'interfacial' phosgenation. An aqueous solution of a primary amine and an acid acceptor, such as an inorganic alkali or a tertiary amine, is added to an agitated solution of phosgene in a water- immiscible organic solvent. The aqueous phase is discarded and the isocyanate recovered from the organic solvent.The method is unsuitable for aromatic isocyanates because of the readiness with which they react with water, but it is useful for aliphatic isocyanates which are hydrolysed only slowly. The procedure is particularly useful for phosgenating primary amines containing unsaturated bonds sensitive to the action of hydrogen chloride present in solvent or vapour- phase phosgenation? The solvent used as reaction medium must be inert to phosgene, hydrogen chloride, and the isocyanate being prepared. Chloro- and dichloro-benzene are frequently used as they are sufficiently high boiling to allow suitably high reaction temperatures to be attained.Overall reaction rates and yields can be increased by using more polar solvents such as ketones and esters, although neither type is satisfactory for large-scale use; hydrogen chloride is highly soluble in them and mild or stainless steel plant is rapidly corroded when they are used. Isocyanates take part in very many reactions5 and in consequence they are difficult to prepare in high yield and purity. The methods used and the chemistry involved in the successful large-scale production of isocyanates are described below. Tolylene Di-isocyanate.-The isocyanate at present manufactured in largest quantity is tolylene di-isocyanate. The commercial material, a colourless liquid boiling at 251 "C,is a mixture of the 2,4- and 2,6-isomers, the product containing 80% of the former being in greatest demand.Toluene is treated with a mixture Bayer, U.S.P. 2 823 221 ;ICI, B.P. 1 165 831. ICI, B.P. 1 152 877; B.P. 1 208 862. J. H. Saunders and R. J. Slocombe, Chem. Rev., 1948,43,204; R.G. Arnold, J. A. Nelson, and J. J. Verbanc, ibid., 1957, 57, 47; S. Ozaki, ibid., 1972, 72,457. 211 Chemistry of the Production of Organic Isocyanates of nitric and sulphuric acids to give dinitrotoluene, usually in a continuous operation; the product is separated from the 'spent' acid, neutralized and washed thoroughly. Small amounts of nitrocresols can be produced during nitration and conditions are selected to minimize their formation as they can act as catalyst poisons in the subsequent hydrogenation stage.The product comprises approxi- mately 76%2,4-, 19%2,6-, 2.5% 3,4-, 1.5% 2,3- and 1.0%3,6-dinitrotoluene. The mixture of isomeric dinitrotoluenes was originally reduced to tolylene diamine with iron and aqueous acid but catalytic hydrogenation is now used. This is usually performed under pressure in an alcohol solvent with either a nickel or a palladium catalyst.6 A high yield is obtained and after filtration the solution is concentrated; the diamine may be distilled to remove some tar. Deamination to 4-toluidine occurs to a small extent. The boiling point of the mixed diamines is ca. 178 "Cat 30 mmHg. As explained later, the presence of 2,3- and 3,4-diamines is disadvantageous as they are responsible for loss of yield and quality defects in the finished pro- ducts.The iron reduction process eliminated ortho-diamines by forming involatile iron co-ordination complexes. Catalytic hydrogenation, however, gives higher yields of the meta-diamines and is easier to operate. Manufacturers therefore minimize formation of ortho-compounds or remove them at some stage. The process used by a particular company is usually not disclosed but many methods for reducing the content of ortho-diamines have been patented. Purification by crystallization is too costly for large-scale operation but separation of isomers by fractional distillation has been claimed as satisfactory in a patent.7 It is also possible to eliminate ortho-isomers by adding compounds which react pref- erentially with ortho-amino-groups to give derivatives which are stable but involatile under the distillation conditions for the meta-diamines.Typical additives include boric acid,* oxalic acid and its ester^,^ urea,1° carbon di- sulphide,ll and tolylene di-isocyanate residues;l2 e.g. the reaction with urea results in the formation of a high melting methylbenzimidazolone (4). Scheme 1 Bayer, B.P. 821 220; TCT,B.P. 907 154.'Allied Chemical Corpn., B.P. 1 133 668. 8 rcr,R.P.966 812. ICI, B.P. 982 483. lo Mobay, B.P. 1 145 431. l1 F.M.C. Corpn., U.S.P. 3 134 813. l2 F.M.C. Corpn., U.S.P. 3 246 035. Twitchet t Phosgenationof Tolylene Diamine. Tolylene diamine was originally phosgenated 'batchwise' by adding a solution of it, e.g. in a halogenated aromatic solvent, to a stirred reactor containing a solution of phosgene in the same solvent.After completion of addition the reactor was heated to 150-200 "C with gaseous phosgene passing until no more hydrogen chloride was evolved. Better yields are now obtained using continuous processes with a series of stirred vessels. Phosgenation in this way was first described by Bayerl3 and modifications of the basic process are still in use today. U Reactors No.1 No.2 Nd.3 Figure 1 Continuous phosgenation of a primary amine In the continuous process (Figure 1) the tolylene diamine mixed isomers are dissolved in an inert solvent such as chloro- or dichloro-benzene, and added continuously together with excess liquid phosgene to a well-agitated first-stage reactor.Conditions of temperature and pressure and reactor design vary from one manufacturer to another. The reactor may be fitted with cooling jackets or internal coils to control the highly exothermic initial reaction (-37.5 kcal mol-l diamine). Alternatively, a very large excess of phosgene (b.p. 8 "C)may be used and allowed to reflw to achieve the same purpose. No satisfactory study of the kinetics of the very rapid reaction of tolylene diamine with phosgene has been published. Model systems using monoamines have been examined but these are considered to have little relevance to the technical systems involving phosgena- tion of polyamines. The reaction mixture flows continuously from the first reactor to a second, third, or even fourth-stage reactor.Residual phosgene and hydrogen chloride are removed from the solution of crude isocyanate which is then concentrated and finally distilled to give a technically pure product. Reactions taking place in the first-stage reactor. The reactions taking place in the first-stage reactor may be summarized as follows: l3 Bayer, Ger. P. 844 896. 213 Chemistry of the Production of Organic Isocyanates -NH2 + COClz +-NHCOCI + HCI (exothermic) (5) -NH2 + HCl +--NH2HCl (exothermic) (6) -NH2 +-NHCOCI +NCO +-NH2,HCI (7) -NHCOCl +-NCO + HCI (labile equilibrium) (8) -NH2 +-NCO -t-NHCO.NH--(exothermic) (9) The initial interaction of phosgene and the mixture of tolylene diamine isomers therefore results in a complex mixture.Dihydrochlorides of the diamine isomers, e.g. (9,are precipitated as also are monohydrochlorides containing isocyanate (6) and carbamoyl chloride groups (7). Substituted ureas (8) are formed in NHa,HCl NH,,HClI I Me+ MeNH-CONH 0 consequence of reaction (9),an undesirable one referred to later. Ureas con- stitute loss of yield and their formation is minimized by working in dilute solu-tion using a large excess of phosgene, and having efficient agitation in the reactor. Reactions taking place in the second-stage reactor. An excess of phosgene is used in the first-stage reactor in order to minimize the formation of substituted ureas (reaction 9). The temperature attained in this reactor is therefore limited, unless it is operated at super-atmospheric pressure, The precipitated suspension obtained in the first-stage reactor flows continuously to a second-stage reactor maintained at a higher temperature (80-120 "C).The bulk of the excess phos- gene here evaporates and is evolved together with hydrogen chloride which results from dissociation of carbamoyl chlorides (reaction 7). The amine hydro- chlorides precipitated at the first stage react in the following manner: -NH2HCI + COCl2 + -NHCOCl + HCl (10) Twitchett This reaction, which has not been closely studied, probably requires initial dissociation of the salt to free amine and hydrogen chloride. The overall reaction involves a heterogenous phase, the precipitated mixed amine hydrochlorides, and it is important to obtain this material in a finely divided form in order that it may react rapidly in the second reactor.If this does not happen third and fourth hot stage reactors will be required to complete the reaction. It is more difficult to obtain the mixed hydrochlorides in a dispersed form in the process in which the diamine is contacted first with hydrogen chloride to form the salt. This makes the overall process more prolonged than in direct phosgenation. Much attention has been given to the study of mixing conditions in the first-stage reactor in order to obtain finely divided precipitates as well as to minimize formation of sub- stituted urea. Many patents claiming improved procedures have been filed and reference is made to typical processes.14 The concentration of phosgene in the hot stage reactors is relatively low, while the concentration of isocyanate is high because of the dissociation of carbamoyl chlorides; it is considered that additional formation of substituted ureas occurs as a result of the following overall reaction which competes with reaction (10).-NH2,HCI + -NCO --* -NH*CO*NH-+ HCI (1 1) The formation and further reactions of substituted ureas. In addition to the substituted linear ureas of type (8) formed in consequence of reactions (9) and (1 1) a cyclic type of urea, methylbenzimidazolone (4), is also produced at the initial mixing stage when the tolylene 2,3-and 3,4-diamine isomers react with phosgene: (4) Scheme 2 All ureas, whether substituted linear ones or cyclic methylbenzimidazolones, are undesirable because their presence results in loss of yield and leads to impurities in the final product.In addition to the direct loss of yield resulting from reaction of an amine group with an isocyanate group to form a urea, further loss occurs because the hydrogen atoms in both linear and cyclic ureas are reactive to isocyanate and combination to form substituted biurets (9), triurets (lo), and higher polyurets (reaction 13) can take place. The presence of small amounts of compounds containing reactive -NH-groups can therefore lead to disproportionate losses of yield. l4 Bayer, B.P. 761 590; ICI, B.P. 1 034 285; Bayer, Ger. P. DT 2 153 268. 215 Chemistry of the Production of Organic Isocyanates -NH -EH ;co -NH (9) Scheme 3 Compounds containing the stable isocyanurate ring (ll), cf.the trimer of TDI (l), are present in the reactor mixture. Increased amounts form during concentration and distillation. It is considered that these may result from ring closure of the higher polyurets. R-NH-CO-NH-R + R R Scheme 4 This type of mechanism would explain the known observation that heating a monoisocyanate with a small amount of a compound containing active hydrogen may polymerize virtually all the isocyanate present. Substituted aromatic ureas react readily with phosgene. By analogy with the aliphatic series,15 it is considered that thermally unstable substituted allophanoyl chlorides (12) and substituted chloroformamidines (13) are first formed.The former can eliminate hydrogen chloride to form isocyanate, although available evidence indicates that this reaction takes place only to a limited extent during the production of TDI. The chloroformamidines appear to react with additional phosgene to give a substituted chloroformamidine-N-carbonylchloride (14) the presence of which can be demonstratedl6 in the phosgenation products of the linear urea (8). It has also been shown16 that, under the catalytic influence of hydrogen chloride, the chloroformamidine-N-carbonylchlorides decompose by two routes to yield substituted carbodi-imides (15) and substituted isocyanide l6 H. Ulrich, J. N. Tilley, and A. A. R.Sayigh, J. Org. Chem., 1964, 29, 2401.l6 R. P. Redman, private communication. Twitchett dichlorides (16). The latter have boiling points close to the analogous isocyanates and are not readily separated by distillation. These compounds, although present in small amounts, therefore contribute to persistent chlorine-containing im-purities in the final product. Mebd-1 \ NH-CO-NH \ Me NCO1 Y1 COC1aI NCO NCO Meb1 NH-Co-N ,& NCO OH COCltI NCO J toa Iheat OCOCl -COa*I -HC1 NCOI NCO NCO NCOI - HCI + NCO Meb1 \ NCO NCOI lHC' NCOI Scheme 5 Chemistry of the Production of Organic Isocyanates The cyclopolymerization of 2,3-and 3,4-tolylene di-isocyanates. Although ortho- diamines react with phosgene mainly to form cyclic ureas, it has been found that the 2,3- and 3,4-isomers of tolylene diamine can react with phosgene to give low yields of the corresponding ortho-di-isocyanates (17).The formation of these compounds has been studied by Schnabel and Kober17 who have suggested the mechanism shown. gCo-wc,-HCI \ ‘ NCO 7 Scheme 6 The presence of the ortho-di-isocyanates as impurities in the mixed 2,4- and 2,6-tolylene di-isocyanates used commercially can give rise to undesirable turbidity. This is the result of spontaneous cyclopolymerization to (18), which is particularly insoluble. [--Nq--CO-N$-]co n Manufacturers take steps to limit the presence of ortho-di-isocyanates in their products. The best solution is to remove ortho-isomers from the tolylene diamine l7 W.J. Schnabel and E.Kober, J. Org. Chem., 1969,34, 1162. Twitchett before phosgenation. The formation of cyclopolymers, including those from ortho-di-isocyanates, has recently been reviewed.18 Distillation of tolylene di-isocyanate. After phosgenation, tolylene di-isocyanate must be recovered in a pure form from the dilute crude reactor material. A complicated work-up procedure is necessary because of the reactivity of the di- isocyanates with their own impurities, the linear and cyclic substituted ureas and because of side-reactions which may introduce persistent chlorine-containing impurities. Concentration and distillation by ‘batchwise’ methods originally used gave poor yields and these operations are now carried out continuously in specialized equipment.Procedures vary according to the solvent employed. Climbing-film and thin-film evaporators19 are used to separate the di-isocyanate from the involatile but reactive by-products without delay. The residence time of material in evaporators of this nature is of the order of seconds and further reaction of TDI with its impurities is minimized. The distillate from the evaporators is virtually free from impurities containing active hydrogen atoms and can be refined further using conventional fractionating columns. The commercial product is of high purity; the ratio of 2,4- to 2,6-isomer is strictly controlled and manufacturers usually specify the content of isocyanate and the levels of acidic and chlorine-containing impurities present.TDI is largely used for making flexible foams by reaction with mixtures of water and polyether polyols. Catalysts are used in the foam-making process, and as impurities present in TDI have a deleterious influence on catalytic activity they must be kept at low and consistent levels. Strict precautions must be taken when handling tolylene di-isocyanate ; its vapour irritates the respiratory system and may cause bronchial asthma. After exposure to the vapour individuals may become sensitized and develop acute symptoms on subsequent exposure to minute amounts of the di-isocyanate. Pure 2,4-tolylene di-isocyanate. Pure 2,4-tolylene di-isocyanate is used industrially in small quantities. It can be obtained from the more important mixture con- taining 80% 2,4- and 20% 2,6-isomers by freezing, when the pure material, m.p. 19.5-21.5 “C separates.Conditions can be regulated to leave a mixture containing 65 % 2,4- and 35% 2,6-isomers, which is used commercially, and can also be made by nitrating toluene to give a mixture of 2-and 4-nitrotoluenes, separating the isomers by distillation, and nitrating the 2-nitrotoluene further to give a ca. 65:35 mixture of 2,4- and 2,6-dinitrotoluene. After reduction the mixed diamines are phosgenated to give the corresponding di-isocyanates. Diphenylmethane Di-isocyanate (MDI).-Diphenylmethane di-isocyanate (19) is the second most important polyisocyanate in Iarge-scale production. The pure 4,4’-isomer is made as a distilled product but the principal commercial product l8 G.C. Corlield, Chern. SOC.Rev., 1972, 1, 523. l9 ‘Encyclopedia of Chemical Process Equipment’, ed. W. J. Mead. Reinhold, New York, 1964, pp. 357, 968. Chemistry of the Production of Organic Isocyanates is an undistilled mixture known as ‘polymeric MDI’ containing about 50% 4,4‘-isomer together with higher methylene-bridged polyphenylene polyiso- cyanates, e.g. (21) and (22). A small amount (5-10 %) of the 2,4’-isomer (20) is usually present in marketed polymeric MDI, but some brands recently introduced contain considerably more together with significant amounts of the 2,2’-isomer. OCNeH2+C0 -(1 9)-Nco Other polyisocyanates resulting from side-reactions are also present and are referred to below.The mean functionality of the commercial product is about 2.8, i.e. on average each molecule contains 2.8 -NCO groups. Although the material is a complex mixture, every effort is made during manufacture to maintain consistent composition to ensure reproducibility in the various poly- urethane systems in which it is used. The product is analysed for its main iso-cyanate components as well as for acidic, chlorine-containing, and other impurities which could upset the polymer-forming reactions if present in excessive amounts. Twitchett The 4,4'-isomer, a colourless solid of low melting point, is treated with polyols to make elastomers, thermoplastic materials, shoe-soling compositions, and elastomeric fibres of the 'Spandex' type.Polymeric MDI, a viscous liquid, is used in large amounts to make rigid polyurethane foam required for heat and sound insulation. A particularly valuable feature of pure and polymeric MDI's is low vapour pressure which makes them less toxic and safer to handle than tolylene di-isocyanate. Pure 4,4'-MDI is normally isolated from the polymeric MDI mixture of polyisocyanates. The starting material in the manufacture is aniline, which reacts with formaldehyde to give a mixture of isomeric diaminodiphenyl- methanes and higher methylene-bridged polyphenylene polyamines. Two distinct processes are used industrially. In one, a large excess of aniline is heated with formaldehyde under pressure in the presence of a catalyst which may be an acid, a salt, or an acid-activated clay.In the second and more commonly em- ployed process, aniline is treated with defined amounts of formaldehyde and hydrochloric acid in an agitated reactor at relatively low temperature. The hydrochloric acid addition may precede or follow the formaldehyde addition. The reaction mixture is heated to 100 "Cto convert intermediate products into primary amines. The reaction of aniline with formaldehyde is exothermic, (-11 kcal mol-l aniline reacted), and a complicated sequence of concurrent and consecutive reactions takes place in both types of process. Many patents describe both catalytic and hydrochloric acid processes and reference is made to typical examples.20 Mechanism of the formation ofpolyamine. The reaction of aniline with formalde- hyde was reviewed by Sprung in 1940,2l and little fundamental work has been published since.22 In general, investigation has been confined to industrial laboratories.It is now considered23 that the mechanism of polyamine formation is as outlined below. When formaldehyde is added to aniline in the presence of acid, the electrophile (23) is produced with the elimination of water: scheme 7 *O ICI, B.P.1 038 266; B.P. 1 229 695; Mobay, B.P. 1 154 980. M.M.Sprung, Chem. Rev., 1940,26,297. *2 J. F. Walker, 'Formaldehyde', Reinhold, New York, 1964, p. 369. 23 R. C. Smith, private communication. Chemistry of the Production of Organic Isocyanates This species then substitutes into an aromatic nucleus to give p-aminobenzyl- aniline (24)by reaction with aniline, or higher polymeric species (25)by reaction with either secondary amine, e.g.(24)itself or higher analogues already produced. A polymeric secondary amine of type (25) is referred to as poly(aminobenzy1- aniline) or poly(anhydroaminobenzy1 alcohol). The second stage of the reaction, known as the ‘isomerization’ stage, converts secondary amine into primary amines. The reaction is, in fact, an intermolecular reaction between a protonated secondary amine and a non-protonated primary amine, so is extremely pH sensitive, and is usually accelerated by heating at 95-100 “Cafter formaldehyde addition. The ‘isomerization’ is illustrated here by the reaction between a protonated polymeric secondary mine and aniline to give 4,4‘-diaminodiphenylmethane(26)and a lower polymer of the secondary amine.J 2,4’-Diaminodiphenylmethane (27) is thought to arise from species (23) substituting into an aromatic nucleus ortho to the NHz group. Triamines, e.g. (28), arise from the diamines (26)or (27) taking part in the initial reaction with formaldehyde to give analogues of (23). Higher methylene-bridged polyphenylene polyamines are formed in a similar manner. Seven isomeric triamines and fifteen isomeric tetra-amines are possible, although a smaller number predominate because of stereoselectivity at the reaction stage. As the reactions taking place during addition of formaldehyde are exothermic it is important to control the temperature and maintain good mixing in the Twitchett reactor to avoid undesirable side-reactions occurring as a result of local excesses of formaldehyde or overheating. After addition of formaldehyde is complete the reaction mixture is heated until ‘isomerization’ of secondary to primary amines is virtually complete. It will be noted that ‘isomerization’ depends on the presence of both free aromatic amine and a protonated species.The rate of ‘isomerization’ is controlled therefore by the ratio of aniline to acid initially used as well as by dilution and temperature. As might be anticipated, the rate of ‘isomerization’ decreases when the molar ratio of acid to aniline exceeds unity because of the decrease in the concentration of free amine. After the ‘isomerization’ stage, the reaction mixture is neutralized with caustic soda when the polyamine separates as an oil which is washed free from salt.This operation is usually done continuously in appropriate equipment at a sufficiently high temperature (90-100 “C)to keep the polyamine liquid. The polyamine is next processed to remove water and residual aniline. It is of interest that a considerable amount of aniline does not react even when less than two mol of aniline per mol of formaldehyde is employed. This is the result of the speed with which formaldehyde reacts with diamines to form tri- amines and so on. Methanol, frequently used to stabilize aqueous formaldehyde against polymerization to paraformaldehyde, is recovered from the saline wash- water as a marketable by-product.Recovered aniline is re-used in the process and it is of interest that it contains 2-and 4-toluidine, as well as N-methylaniline, formed as impurities during the addition of formaldehyde. Composition of the poZyamine.The composition of the polyamine can be varied widely according to the conditions used, the principal factors being ratio of aniline to formaldehyde, ratio of aniline to hydrochloric acid, dilution, tempera- ture, and rate of addition of formaldehyde. Thus large excesses of aniline favour Chemistry of the Production of Organic Isocyanates formation of the 4,4’-diamine (26), whereas small amounts of hydrochloric acid result in higher amounts of the 2,4’-diamine (27) together with larger proportions of the higher met h ylene- brid ged pol yamines .The cat a1 ysed high- temperature pressure process gives a polyamine containing less 4,4’-diamine and much more of the 2,4‘-and 2,2’-isomers than does the aqueous hydrochIoric acid route, owing to reduced stereoselectivity at the higher reaction temperatures. Some typical analyses of polyamines are given in Table 1. Table 1 Composition of polyamine Process Diamines Higher Aqueous HCl 60% 4,4’- 5.0% 2,4’- trace 2,2’- 35% polyamines Catalytic high temperature 18.3% 18.6% 4.7% 58% Small amounts of 3-methyldiphenyImethane-4,4‘-diamine(29) are often found in the polyamine as well as a little un-isomerized secondary amine and 3-phenyl- 3,4-dihydroquinazoline(30). The presence of secondary amine and the quinazol- ine is undesirable as both give rise to acidic chlorine containing impurities during phosgenation.Phosgenation of the polyamine. The polyamine can be distilled under reduced pressure to give fractions containing the high-boiling diamine and triamine isomers, but normally it is phosgenated directly to the corresponding mixture of polyisocyanates. The equipment and conditions used are similar to those for the continuous phosgenation of tolylene diamine. The same inert solvents, e.g. chloro- or dichloro-benzene, are used and the product is obtained as a dilute (5-25 %) solution. Some care is necessary in selecting materials of construction for the phosgenation reactors because, although isocyanate solutions containing phosgene and hydrogen chloride are not particularly corrosive to mild or stain- less steel, some of the by-products of side-reactions can be.Mixtures of poly- arnine-derived ureas and phosgene or hydrogen chloride corrode steel quite Twitchett readily, although the iron co-ordination compounds formed have not been studied in any detail. Special grades of stainless steel are normally used to minimize corrosion. The use of solvent containing small amounts of water or inadequately dried polyamine also leads to corrosion. As the bulk of the product is used as an un- distilled product, corrosion products remain in it. High levels of iron and other transition metals in ‘polymeric’ MDI are deleterious as the metals have consider- able catalytic activity in the reactions of the polyisocyanate in many polyurethane applications.Variable contents in particular are undesirable as they lead to inconsistency in the production of derived urethane polymers. Reactions taking place during phosgenation. The reactions taking place during the phosgenation of the polyamine are even more difficult to study than those occurring with tolylene diamine because of the greater complexity of the amino- compounds involved. In general, the main reactions appear to follow the pattern of reactions (5)--(11) described above for tolylene diamine. The reaction taking place in the first-stage reactor has a similar order of exotherm and the suspension which flows from this reactor appears to be highly solvated with phosgene and less crystalline than in the case of tolylene diamine.It reacts more readily at the second, hot stage. This is attributed to the greater complexity of the polyamine mixture giving a precipitate of smaller particle size. Linear ureas (reaction 9) appear to be formed and there is evidence that these react with isocyanate groups to form biurets and polyurets (see reaction 13) and with phosgene to form N-carbonyl chlorides [cf. (13)]. A secondary amine, e.g. (24), reacts with phos- gene to give a secondary carbamoyl chloride (31) which cannot dissociate in the manner of a primary carbamoyl chloride and remains as a chlorine-containing impurity. 3-Phenyl-3,4-dihydroquinazoline(30) reacts with hydrogen chloride to give a stable hydrochloride which also remains in the product. In some respects, control of side reactions is of greater importance than in production of tolylene di-isocyanate as polymeric MDI is not distilled and impurities remain as contaminants.Concentrationof dilute polymeric MDI. As in the case of tolylene di-isocyanate, the solvent is removed from dilute polymeric MDI without delay. Any reaction of isocyanate groups with impurities such as substituted weas containing active hydrogen atoms is minimized by using climbing-film and thin-film evaporators Chemistry of the Production of Organic Isocyanates having low hold-up times, for continuous concentration. The polymeric MDI produced contains appreciable amounts of chlorine-containing impurities, and it is necessary to eliminate these to some extent to make the material usable for polyurethane applications.This is done by a short treatment at high temperature under specialized condition^.^^ Phosgene and hydrogen chloride are evolved during the treatment and it is considered that one of the reactions involves de- composition of a chloroformamidine-N-carbonylchloride(32), formed as a result of phosgenation of urea groups, to a substituted carbodi-imide (33). On cooling, carbodi-imide groups reversibly combine with isocyanate groups to form substituted uretonimines (34), the presence of which in polymeric MDI at low temperature results in a viscosity increase in consequence of the increase in molecular complexity. INI1 cII0 Scheme 9 Yet another reaction occurring as the result of the heat treatment process concerns the formation of an insoluble sediment arising from the dimerization of isocyanate groups.Dimerization of isocyanates. Many aromatic isocyanates can be dimerized25 to substituted uretediones (35); the reaction is reversible and in fact it is the dimer which is thermodynamically stable at room temperature, although its rate of formation at this temperature in the absence of suitable catalysts (e.g. trialkyl-phosphines or 4dimethylaminopyridine) is very slow. The sediment which forms 24 ICI, B.P. 1 080 717; du Pont, B.P. 1 015 977. 25 R.G.Arnold, J. A. Nelson, and J. J. Verbanc, Chem. Rev., 1957, 57, 54. Twitchett CO 2RNCO 9R--N' 'N--R'd Scheme 10 both in polymeric MDI and the pure 4,4'-MDI comprises mainly the very insoluble compound (36).At 200 "C, dissociation of the uretedione to 4,4'-MDI is virtually complete; at 120 "C the equilibrium mixture contains 6.3 % dimer. If, after concentration and heat treatment, polymeric MDI is allowed to cool slowly as it may do in bulk, appreciable amounts of uretedione form and are precipitated as sediment. It is therefore necessary to heat polymeric MDI to around 200 "C and then cool rapidly to freeze the equilibrium and so prevent deposition of sediment. Dimer forms slowly in polymeric MDI on prolonged storage and also in pure 4,4'-MDI in the crystalline state. For this reason the latter is sometimes kept in refrigerated storage.In spite of its sterically unhindered 4'4socyanate group, 2,4'-MDI does not dimerize readily. Pure 4,4'-diphenylmethane di-isocyanate. Technically pure 4,4'-diphenylmethane di-isocyanate is obtained by distillation or crystallization of polymeric MDI. Many variations of these basic processes have been patented. As the presence of quite small amounts of the 2,4'-isomer can have a detrimental effect on derived elastomers, a polymeric MDI precursor is made by phosgenating a polyamine containing a low content of 2,4'-diamine and a high proportion of the 4,4'- isomer. This type of polyamine can be obtained by treating a large excess of aniline with formaldehyde in the presence of approximately equimolecular proportions of hydrochloric acid and aniline.After phosgenation, polymeric MDI containing corresponding proportions of 2,4'- and 4,4'-MDI is obtained. A fraction rich in 4,4'-MDI, obtained by distillation or crystallization, is usually purified further to remove chlorine-containing impurities and residual 2,4'-isomer by re-distillation or re-crystallization or by a combination of these methods. The boiling point of 4,4'-MDI is high (156-158 "Cat 0.1 mmHg) and, like tolylene di-isocyanate, it must be distilled in specialized thin-film evaporators Chemistry of the Production of Organic Isocyanates capable of operating at low pressures and having short hold-up times if undue degradation is to be avoided. To achieve a currently viable process it is essential to use the undistilled or uncrystallized residues from the polymeric MDI precursor.The composition of the original polyamine, and hence the polymeric MDI precursor and the pro- portion of 4,4'-MDI removed, are therefore arranged so that the composition of the residue is similar to the product normally used for making polyurethane foams. Typical analyses of the precursor, technically pure 4,4'-MDI, and poly- meric MDI residue are given in Table 2. Table 2 Composition of pure and polymeric diphenylmethane di-isocyanate Material Diphenylmethane Tri-Higher di-isocyanate isocyanates poly- 4,4'-2,4'-isocyanates Precursor 65 % 4.0 % 20% 11% Pure MDI 98% 2.0 % nil nil Polymeric MDI residue 56 % 4.4 % 25 % 13.7% Technically pure 4,4'-diphenylmethane di-isocyanate is a solid, m.p.38-39 "C. A Iiquefied product, also used commercially, is obtained by converting a small proportion of the isocyanate groups into carbodi-imide. This can be done by heating alone,26 but is usually achieved by heating in the presence of a catalytic phosphorus compound such as a trialkyl ph~sphate.~' The most effective catalysts known are the phosp holine oxides:* e.g.the 3-methyl-1-e t hy 1or 3-methyl-1-p hen y1 derivatives (37), which convert isocyanates into the corresponding carbodi- (37) R = Et or Ph hides at room temperature. A mechanism for the catalytic action of these oxides has been proposed.29 The use of titanium and other metal alkoxides as catalysts for carbodi-imide formation has been patented.30 26 Bayer, B.P.899 036. 27 Upjohn, B.P. 1 069 858. 2a du Pont, U.S.P. 2 853 473. eg J. J. Monagle and J. V Mengenhauser, J Org. Chem., 1966,31,2321. 30 du Pont, U.S.P. 3 426 025. Twitchett On cooling, the carbodi-imides associate with isocyanate groups to form substituted uretonimines [see (34)]. 4 The Cyanation and Carbonylation Processes for Making Xsocyanates The large number of patents issued shows that several companies have closely examined the possibilities of the cyanation and carbonylation processes as alternatives to phosgenation of primary amines for manufacturing isocyanates. The cyanation process, in which an alkyl or aralkyl chloride, sulphate, or phosphate is treated with a metal cyanate was used by Wurtz31 to synthesize the first organic isocyanate: RCl + NaCNO -+ RNCO + NaCl Numerous patents have been granted claiming the making of alkyl and aralkyl mono- and poly-isocyanates.Particular attention has been given to the treatment of bis(chloromethy1) derivatives of benzene and xylene with sodium cyanate to obtain polyisocyanates comparable in cost with tolylene di-isocyanate but capable of giving urethane polymers less subject to discoloration. The hetero- geneous reaction with finely ground sodium cyanate is not easy to control, how- ever, and considerable amounts of substituted isocyanurates (39) are usually formed as by-products, as well as the di-isocyanate (38). A typical process38 has been patented by the Marathon Oil Co. Scheme 11 Bayer33 claims a similar process for making methoxymethyl isocyanate (40).MeOCHzCl + NaCNO --f MeOCHzNCO + NaCl (40) Sodium cyanate is much dearer to produce than phosgene and adds to the expense of the process, while unchanged chloro-compounds are sometimes 31 A. Wurtz, Compr. rend., 1848, 27,241; Annalen, 1849,71,326. sa Marathon Oil Co.,U.S.P.3 458 448. 33 Bayer, B.P. 1 104 917. Chemistry of the Production of Organic Isocyanates difficult to separate from derived isocyanates because of the closeness of their boiling points. Many of the reactive chlorine compounds which have been used for cyanation are most readily obtained by chloromethylation of aromatic compounds. A hitherto unsuspected hazard concerning this reaction has recently been reported.% Bis(chloromethyl)ether,a by-product and/or an intermediate in chloromethyla- tion has been found to be a potent carcinogen as well as an acute lung irritant and it is recommended that this reaction be avoided.Cyanation is not seen at present as a feasible route for manufacturing other than very specialized isocyanates. In the carbonylation process an aromatic nitro-compound is converted directly into an isocyanate on treatment with carbon monoxide. The process is potentially attractive for making tolylene di-isocyanate, as the intermediate stage of reducing dinitrotoluene to the diamine is eliminated. The direct use of carbon monoxide also avoids initial treatment with chlorine to obtain phosgene and the subsequent excess production of hydrogen chloride.The process has been examined by several manufacturers and numerous patents have been published by American Cyanamid, Shell, ICI, and Olin Mathieson. The first direct synthesis of an isocyanate by carbonylation was claimed by Hardy and Bennett35 of American Cyanamid, who obtained a 35% yield of phenyl isocyanate by treating nitrobenzene with carbon monoxide and a palladium or rhodium on alumina catalyst in the presence of a Lewis acid. A 57% yield of tolylene di-isocyanate from dinitrotoluene is claimed in a patent by Olin Mathie~on~~ using a mixture of palladium, molybdenum, iron, and chromium catalysts, while Shell claim 70-76% yields using a mixed palladium, iron, and molybdenum catalyst in the presence of pyridine.37 Expensive rare-metal catalysts seem to be essential for the carbonylation synthesis of isocyanates.As far as is known, no large-scale plant operating this type of process has been commissioned although patents continue to appear. The author would like to thank his colleagues, in particular Drs W. Costain, R. P. Redman, and R. C. Smith for the help they have given him in preparing this review. 34 ‘Organic Reactions’, Wiley, New York, 1972, Vo.119, p. 422. 35 W. B. Hardy and R. P. Bennett, Tetrahedron Letters, 1967, 961. 36 Olin Mathieson, B.P. 1 252 517. Shell, U.S.P. 3 719 699.

ISSN:0306-0012    

DOI:10.1039/CS9740300209

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Metalloboranes and Metal-Boron Bonding By Norman N. Greenwood and Ian M. Ward DEPARTMENT OF INORGANIC AND STRUCTURAL CHEMISTRY, THE UNIVERSITY OF LEEDS, LEEDS LS2 9JT, ENGLAND Introduction Compounds in which a metal atom is directly bonded to a boron hydride group have been increasingly studied since the first examples were synthesized some ten years ago, and it is now clear that the field of metalloborane chemistry is potentially as large and diverse as the older field of organometallic chemistry. Boron is the element immediately preceding carbon in the Periodic Table and it therefore has one less electron available for bonding but the same number of orbitals (the 2s and the three 2p orbitals). The covalent radius, electronegativity, and ionization energies of the two atoms are very similar and this, coupled with the propensity of boron to form polyatomic clusters both with itself and with other elements, permits an extensive covalent chemistry to be deve1oped.l Just as the numerous and novel types of structure and bonding that can be adopted by compounds between metals and hydrocarbons or their derivatives have been elucidated and exploited during the past 25 years, so it now appears that a similarly rich variety of structural and bonding types can be generated by com- pounds between metals and boron hydrides or their derivatives.Metalloborane compounds can be divided into four broad categories : (i) Ionic hydroborates, of which the simplest are the tetrahydroborates such as NaBH4 and Ba(BH4)2; these have recently been reviewed by James and Wallbridge.2 Also in this category are the ionic octahydrotriborates MB3H8 and a variety of nido-and arachno-polyhydroborates such as MBsHs,3 MBeH9,3 CsB 9H14,~ and NaBloH13.5 Ionic closo-polyhydroborates of general formula M2BnHn(n = 6-12) are likewise well known and adequately documented,lp6 though the special case of Cu2BloHio will be considered further (on p.255). (ii) Metal hydroborates containing hydrogen-bridge bonds M-H-B, for example Al(BH4)3,7 (Ph3P)zCuBH4,8 and (Ph3P)KuB3Hs.g N. N. Greenwood, ‘Boron’, Chapter 11 in ‘Comprehensive Inorganic Chemistry’, ed. J. C. Bailar, H. J. Emelkus, R. S. Nyholm, and A. Trotman-Dickenson, Pergamon Press, Oxford, 1973. B. D. James and M.G. H. Wallbridge, Progr. Inorg. Chem., 1970,11,99. H. D. Johnson, R. A. Geanangel, and S. G. Shore, Znorg. Chem., 1970,9,908.* N. N. Greenwood, J. A. McGinnety, and J. D. Owen, J.C.S. Dalton, 1972, 986; N. N. Greenwood, H. J. Gysling, J. A. McGinnety, and J. D. Owen, Chem. Comm., 1970, 505. L. G. Sneddon, J. C. Huffman, R. 0. Schaeffer, and W. E. Streib, J.C.S. Chem. Comm., 1972, 474. W. H. Knoth and E. L. Muetterties, ‘Polyhedral Boranes’, M. Dekker Inc., New York, 1968.’A. Almenningen, G. Gunderson, and A. Haaland, Actu Chem. Scand., 1968, 22, 328. S. J. Lippard and K. M. Melmed, Znorg. Chem., 1967, 6, 2223.* S. J. Lippard and K. M. Melmed, Znorg. Chem., 1969, 8, 2755. 231 Metalloboranes and Metal-Boron Bonding (iii) Metal-carbaborane complexes including wbonded ‘sandwich’ com-pounds; this is probably the most extensive class of metal-boron compound to date and several reviews are available.10-13 (iv) Compounds containing direct metal-boron bonds except those in the carbaborane category (iii) above.It is these compounds which form the subject of the present review. The sequence of presentation will be according to the various synthetic routes that have been developed for these compounds. The chemical reactions, physical properties, structure, and bonding will be incor- porated within this framework, and a concluding section will summarize the types of bonding so far observed in these diverse structures. Seven broadly distinguishable preparative routes have been devised ; they overlap to some extent but form, nevertheless, a convenient framework for discussion.The routes are:* 1 borane plus metal hydride compound; 2 borane plus metal alkyl or aryl; 3 borane plus transition-metal complex; 4 borane anion plus transition-metal complex; 5 metathesis (i.e. mutual interchange of metal atoms); 6 borane anion plus metal halide (or alkylmetal halide); 7 borane (or borane anion) plus transition-metal carbonyl. 1 Borane plus Metal Hydride Compound One of the first syntheses of a covalently bonded metal derivative of a polyhedral borane cluster was reported in 1965 by Greenwood and McGinnety.l4 Deca- borane(l4) was allowed to react in ethereal solution at room temperature with trimethylamine-alane to give a quantitative yield of the highly reactive colourless complex [Me3NH]+[B1oH12A1H2]-,nEt2Oywhere n= 1.5-2.0: Et,O &OH14 + Me3NAlH3 +[M~~NH~+[B~OH~~AIH~]-,~E~~O+ H2 (1) The structure of the metalloborane anion has not been established with certainty *Note added in proofi An important additional route has recently been published by 0.Hollander and S. G. Shore, 167th Natl. Meeting Amer. Chem. SOC., Los Angeles, April 1974, Abstract INORG 184; this involves skeletal build-up by insertion of BH, into a pre- formed metalloborane complex [see equation (62)]: p -B,H,,Fe(CO), + KH _.+ K+[tc-B,H,Fe(CO)J-+ H, K+b-B,H,Fe(CO)J-+ *B,H, __+ K+[B,H,,Fe(CO) J-K+[B,H,,Fe(CO)J-+ HC14 [B,H,,Fe(CO),] + KC1 + H2 The product is stable at -78 OC, decomposes over several hours at 0 OC, and is the first example of a previously unknown neutral borane (B,H1,) to be stabilized by co-ordination to an electron-withdrawing group.lo R.Snaith and K. Wade, ‘International Review of Science, Inorganic Chemistry, Series 1’,M.T.P.,Butterworths, 1972, Vol. 1, Ch. 4. l1 M.F.Hawthorne, Endeavour, 1966,25, 146 I* M. F. Hawthorne, Pure Appl. Chem., 1968,17, 195. Is M. F.Hawthorne, Pure Appl. Chem., 1972,29,547. N. N.Greenwood and J. A.McGinnety, Chem. Comm., 1965,331. Greenwood and Ward but its reactions and i.r. spectrum are consistent with its formulation as the 6,9-dihapto-derivative of B10H142-shown in Figure 1. Th~s,1~J~ Figure 1 PossibZe structure and topology of [BloH12AIHz]-&OH14 + Me3NAlHzCl + nEtzO -+ [Me3NH]+ [BI oH12AlHz]-,nEtzO + HCl (2) &OH14 + Me3NAID3 + nEtzO -+ [Me3NH]+[B10HlzAlDz]-,nEt20+ HD (3) [Me3NH]+[B1oH1zAlHz]-,nEtzO+ 3Hz0 3 + Al(OH)3 + H2 + nEtzO (4)[M~~NH]+[B~oH~~]-l5 N.N. Greenwood and J. A. McGinnety, J. Gem. SOC.(A), 1966, 1090. Metalloboranes and Metal-Boron Bonding In these reactions it is noteworthy that the HCI liberated in reaction (2) does not react with the co-ordinated AlH2 moiety, and that in reaction (3) almost pure HD and co-ordinated AID2 are produced. When ,u*-bridge-deuteriated decaborane was used, a partially deuteriated product was obtained plus one molar equivalent of composition H2 26%, HD 74%, D2 < 0.1 %. This indicates that it is the bridge hydrogens which are involved in the reaction.The hydrolysis of the complex with water or alcohols was explosively violent unless moderated by a very large excess of ether; under these conditions the aluminium was quantitatively eliminated but only one mole of hydrogen was evolved, due to the formation of B10H15- ion. The i.r. spectrum of the original complex indicates that the ether molecules are co-ordinated (presumably to the aluminium) rather than free; for this reason the dihapto-structure shown in Figure 1 is preferred rather than the 6,7,8,9-tetrahapto-structure,though this structure, which is known in other metallo-boranes (see later), is not rigorously excluded. The reaction of decaborane with trimethylaminegallane in ether at room temperature proceeds rather differently, according to equation (5).Et,O &OH14 + Me3NGaH3 __+ [Me3NH]+[BloGaH16]-(5) No hydrogen is evolved and the product is a fine, white, solvent-free powder, stable under dry nitrogen, and soluble as a 1:1 electrolyte in ethanol.14 In view of the violent reactivity of [BloH12AlH2]-,nEtzO to water it is remarkable that the anion [B10GaHls]- can be recovered unchanged even from 1 M-hydrochloric acid; it is, in fact, the only known gallium hydride derivative which is hydro- lytically stable. Few other direct reactions of boranes with metal hydride compounds have been exploited to yield metalloboranes but one example is the reaction of diborane with germylpotassium in 1 ,Zdimethoxyethane at low temperatures to give the monohapto anionic derivative potassium germyltrihydroborate(1- ) :16 (MeOCH,),4B2H6 + KGeH3 __I_, K+[HsGeBH3]-(6) This anion can be regarded either as a germyl-substituted BH4- ion or as an adduct between the germyl anion GeH3- and BH3.The salt KH3GeBH3 melts at 99 "C and decomposes at 200 "C to germanium, germanium hydrides, hydrogen, and potassium tetrahydroborate(1 -). It is fairly stable in alkaline aqueous solutions but hydrolyses completely in acids to germane, hydrogen, and boric acid. 2 Borane plus Metal Alkyl or Aryl A. Group 11Metals.-The first reaction of this type to be reported concerned decaborane and diethylcadmium;17J8 these reacted slowly in equimolar propor- l6 D. S. Rustad and W. J. Jolly, Inorg. Chem., 1968, 7, 213.l7 N. N. Greenwood and N. F. Travers, Inorg. Nuclear Chem. Letters, 1966, 2, 169. la N. N. Greenwood and N. F. Travers, J. Chem. Soc. (A), 1967, 880. Greenwood and Ward tions in diethyl ether at room temperature to give a white, crystalline product which was subsequently shown to be dimeric9 Eta0 2B10H14 +2EtaCd +[(EtaO)aCdBloH~a]z+2EtH (7) The reaction is more rapid in THF and the corresponding product is pale yellow. No reaction occurred in benzene solution unless a co-ordinating ether was added. Organozinc reagents, R2Zn (R = Me, Et, or Ph), reacted similarly to give air-sensitive, yellow crystalline solids in yields which depended both on the nature of R and the solvent S.17820 ethereal solvent 2B10H14 +2R2Zn -[SsZnBloH12]2 +2RH (8) s A stoicheiometric complex could be obtained in the presence of a stronger ligand; e.g.decaborane reacted with the bipyridyl adduct of diethylzinc in diethyl ether solution at room temperature according to the stoicheiometry indicated in equation (9).21 Eta0 B10H14 +(bipy)ZnEtz _j(bi~y)ZnBioHi2+2EtH (9) Diethylmagnesium reacted rapidly with an ethereal solution of decaborane even at -78 "C to give high yields of the cream-coloured complex (EtzO)zMgBioH~a.~~ By contrast, organomercury compounds did not react even at elevated tempera- tures, thereby indicating the influence of metal-carbon bond polarity on reactivity. The crystal and molecular structure of [(Et20)2CdB10H12]2 has been deter- mined by single-crystal X-ray diffraction techniq~es.1~ The compound comprises two B10H122- icosahedral fragments each acting as a bis-dihapto-ligand to the two cadmium atoms (Figure 2).Each cadmium atom is bonded by three-centre bonds so as to bridge the two borane clusters and is also co-ordinated by two oxygen atoms from the ether molecules to give an overall distorted tetrahedral bonding environment. The hydrogen atoms were not located but the structure clearly indicates that two bridge protons in each of the decaborane clusters have been replaced by the cadmium atoms to give the topology shown in Figure 3. This structure immediately rationalizes the chemical properties of the com- pound.l* Thus, it dissolves as a non-electrolyte in methylene chloride and reacts smoothly and quantitatively with dry hydrogen chloride in ether to regenerate decaborane: loN. N.Greenwood, J. A. McGinnety, and J. D. Owen, J.C.S. Dalton, 1972, 989. *ON. N. Greenwood and N. F. Travers, J. Chem. SOC.(A), 1968, 15. z1 D. W. Waite, Ph.D. thesis, 1973, Univ. Newcastle upon Tyne, England. 235 Metalloboranes and Metal-Boron Bonding Figure 2 Structure of [(EtzO)zCd(BloHlz)]zshowing positions of boron, cadmium, and oxygen atoms. The shortest Cd-B distances are in the range 237-246 pm and theCd-Odistancesare 224-238 pm. The non-bonding Cd-Cd distance is 486 pm and the angle OCdO is 87" Figure 3 Topology of the BioHi2 unit showing the three-centre bonding to the two bridging cadmium atoms In solvents such as water, ethanol, and methyl cyanide the compound (and the bis-tetrahydrofuranate) dissolved to give a 2: 2 electrolyte as shown in equation (11).This proved to be a general route to the novel series of metalloborane [(EtzO)zCdBloHnlz+Cd2++ [Cd(BioHi2)2l2-+ 4Etz0 (11) complexes [M(BioH12)2I2- (M = Mg, Zn, or Cd), which could be obtained as yellow or orange crystalline solids when precipitated from solution by large Greenwood and Ward cations such as Me4N+ and Ph3MePt.18920J2 The isostructural mercury com- plexes were obtained by the alternative routes discussed in Section 6. The structure of [P~~M~P]+~[Z~(BIOHI~)~]~-was determined by single- crystal X-ray diffraction analysis, and the positions of all 121 atoms in the compound were located.23 The anionic complex, shown in Figure 4, can be 2-2' Figure 4 Structure of the anion [Zn(BloHl2)2l2-showing the positions of the zinc and boron atoms only; each boron has one hydrogen atom attached and each cluster has two hydrogen atoms bridging the 6,7 and 8,9 positions considered to comprise two bidentate B10H12~- ligands co-ordinated tetra- hedrally to zinc, as indicated in Figure 5.The distances between zinc and B(5), B(10), B(53, and B(10') fall in the range 220-221 pm, and the distances between zinc and B(6), B(9), B(6'), and B(9') fall in the range 243-247 pm. The average B-H (terminal) distance is 105 pm and the average B--H (bridge) is 131 pm, which are within the normal range. Most of the boron-boron distances within pp N. N. Greenwood and N.F. Travers, J. Chem. SOC.(A), 1971, 3257. p3 N. N. Greenwood, J. A. McGinnety, and J. D. Owen, J. Chem. Suc. (A), 1971, 809. Metalloboranes and Metal-Boron Bonding %"I2 Figure 5 A three-dimensional representation of the 2640 topology suggested for the bidentate ligand B10Hr2~-the clusters are similar to those in decaborane, the greatest deviations (5-1 3 pm) occurring, as expected, for those boron atoms closest to the zinc atom. The angles subtended at zinc by the mid points of B(5)B(6) and B(9)B(10) are 77.4" and 77.2" for the two cages, implying substantial deviation from the ideal tetrahedral angle of 109". The ix., u.v., and n.m.r. spectra of these compounds, including the high- resolution 1lB n.m.r. spectra at 80.53 MHz, have been reported.1*J0122 The reaction of other boranes with Group I1 metal alkyls has been much less extensively studied but it is known that hexaborane(l0) reacts with the di- methyls of magnesium, zinc, and cadmium (but not mercury) to form somewhat unstable compounds of formula MeMBsH9,nTHF :2* THF B6HlO +MezM _jMeMBsH9,nTHF +Me In the case of the magnesium compound the product is still sufficiently reactive to form a bishexaboranyl derivative [reaction (13)], which is stable for weeks at room temperature.THF B6HlO +MeMgBsH9,nTHF +Mg(BgHg)z,xTHF +MeH No structural data were published but the 1lB n.m.r. spectra indicated that the B6 cluster remained intact and that there was an equivalent interaction (at least on the n.m.r.time-scale) between all the basal boron atoms and the metal atom. B6HlO was also shown to react with MeMgBr to give BeHgMgBr and methane.24 24 D. L. Denton and S. G. Shore, Abstracts of 162nd meeting of A.C.S., Washington D.C., 1971, INORG, 4. Greenwood and Ward B. Group III Metals.-Trimethylthallium reacts with decaborane in diethyl ether to give two dimethylthallium derivatives by competing reactions:25 &OH14 + Me3Tl-Eta0 [Me2Tl]+[BioHi3]-+ MeH (14) Eta0 B10H14 + 2Me3Tl _j [Me2Tl]+[BioHizTlMe2]-(1 5) The yellow crystalline product of reaction (14) is simply the dimethylthallium salt of the tridecahydrodecaborate(1 -) anion but the product of reaction (15) is a true metalloborane complex. Thus the 1Hn.m.r. spectrum of the compound in hexadeuterioacetone showed three distinct sets of methyl protons in the ratio 2: 1 :1, the more intense of which corresponded in chemical shift and Tl-(C)-H coupling constant to the known Me2Tl+ cation. The other two non-equivalent sets of methyl protons were assigned to the 9,10,5,6-tetrahapto-complexshown in Figure 6.This would be analogous to the known chelating tendency of Figure 6 Proposed heavy-atom structure of the anion [BioH12TlMe2]-; numbering as in decaborane(l4). Each boron atom carries one terminal hydrogen atom and there are bridging hydrogen atoms between B(6)-B(7) and B(8)-B(9) B10H122- in [M(B1oH12)22-] (see Figure 4). However, it would imply a high-resolution 1lB n.m.r. spectrum in which there were six doublets (in the absence of accidental overlap): four with relative intensity 2 and two with intensity 1.Unpublished work at 80.53 MHz21 suggests that there may be a resolved triplet of relative total intensity 1(implying the presence of a BH2 group) together with three partially resolved doublets of intensity 1 and three doublets of intensity 2. This could be explained if the Me2Tl group were bonded to the decaborane cluster by a three-centre, two-electron bond involving B(5) and B(6), whilst the m N. N. Greenwood, N. F. Travers, and D. W. Waite, Chem. Comm., 1971, 1027. Metalloboranes and Metal-Boron Bonding B(9) atom carried two terminal hydrogen atoms -this would destroy the CzU symmetry of the borane cluster and render each boron atom structurally distinct, though it would be expected that several pairs of boron atoms would be sufficiently similar to give apparently coincident doublets of relative intensity 2, as observed.It is also possible that the structure in solution differs from that in the solid. The reaction of trimethylindium with decaborane in benzene solution26 is analogous to that of the corresponding thallium compound. Methane is evolved and, depending on the stoicheiometry of the reagents, two products are pro- duced, [MezIn]+ [MezInBl oHiz]-, which crystallizes from solution during the course of the reaction, and MezInB10H13, which remains in solution. [Me2In]+[MezInBloHlz]- is a pale yellow crystalline solid, the formulation of which is confirmed by analytical, i.r., n.m.r., and conductivity data.The pro- posed structure of the [MezInB10H12]- anion is analogous to that of the thallium derivative (see Figure 6). MezInBloH13 is a yellow solid which is formulated as a simple derivative of B10H14. 3 Borane plus Transition-metal Complex The first report of the preparation of a metalloborane by this route involved the direct reaction of (cleaved) diborane in ethereal solution with carbonyl deriva- tives of manganese and rhenium:27 THF,BH3 + NaMn(C0)s --+ Na+[(CO)sMn(BH3)]-+ THF (16) THF,BH3 + Na+ [(Ph3P)Mn(C0)4]- 3Na+ [(Ph3P)Mn(C0)4(BH3)]- + THF (17) In a reaction analogous to (16) the rhenium complex NaRe(C0)a absorbed either one or two moles of borane from solution in bis-(2-methoxyethyl) ether: diglymeQB2H6 + NaRe(C0)a --+ Na+[(C0)5Re(BH3)]-(18) diglyme *B2H6 + Na+[(C0)5Re@H3)1--Na+[(C0)5Re(BH3)2]-(19) These sodium salts of pentacarbonyl metal boranes crystallize as ethereal solvates but the non-solvated Et4N+ and Bu4Pf salts were isolated by metathesis in THF.The manganese analogue was less stable and lost diborane on standing for several days at -25 "C: > -25 "C [(C0)5Mn(BH3)]-[Mn(C0)5]-+ THF,BH3 (20)-78 "C However, the phosphine derivative [reaction (17)] was stable at room tempera- ture. The i.r., electronic, and n.m.r. spectra led to the conclusion that the metal was acting as an electron-pair donor to the borane, i.e. [LaM+BH3]-where 26 N. N. Greenwood and B. S.Thomas, unpublished observations. 27 G. W Parshall, J.Amer. Chem. SOC.,1964, 86, 361. Greenwood and Ward L = Ph3P and M = Mn or Re. The reaction of decaborane with a hydrido-platinum chloride complex affords another example of this class of reaction:28 BioHi4 + 2(EtsP)2Pt(H)Cl--+ [Pt(BloHrz) (EtsP)a] + (Et3P)aPtCla + 2H2 (21) This is a variant of a more general reaction to be considered in the next section. Another synthesis of a metalloborane by the reaction of a borane with a transition-metal complex involved the novel use of decaborane as an oxidizing agent.29 Thus decaborane reacts with (py)4FeIJBrz in benzene according to equation (22) to give the FelI1 compound (A) in 94 % yield and this, on treatment with THF, was cleaved to give an insoluble complex (py)FeBrf~BloH11(py)2 and a compound that was soluble in THF and which, on recrystallization, gave an almost quantitative yield of the brown, tetrahedral, high-spin iron(n1) complex (py)FeBraB10H13(py) (B).Complex (B) was monomeric and non-conducting in methyl cyanide and had a magnetic moment perf = 5.60 BM. Infrared spectro- scopy confirmed the presence of pyridine co-ordinated both to iron and to the boron cluster respectively, thus leading to the formulation of the complex as a monohapto-metalloborane (Figure 7). This structure can be thought of as being derived from BlOH15- in which the H- ligand has been replaced by py and one hydrogen atom substituted by the iron complex. Analogous but more com- plicated oxidative reactions occurred between decaborane and complexes of cobalt(rr) and tin@).More recently,3O hexaborane(l0) has been shown to react with K[(q-C2H4)PtC13] at low temperatures to give the compound (BgH10)2PtC12, which is unstable at room temperature but forms yellow crystals below -20 "C: 2BeHlo + [(q-C2H4)PtC13]---t (BgH10)2PtC12 + C2H4 + C1-(23) The complex has been shown by single-crystal X-ray diffraction analysis to have the trans-square-plar,ar configuration about the platinum atom, each B6 unit being bridge-bonded via the unique B(4)B(5) basal boron atoms (see Figure 8). The most noteworthy feature of the icosahedral Be fragments is the lengthening of the unusually short B(4)--B(5) distance (160 pm) found in uncomplexed F. Klanberg, P. A. Wegner, G. W. Parshall, and E. L. Muetterties, Inorg.Chem., 1968,7, 2072. ts N. N. Greenwood and H. Schick, Chem. Comm., 1969,935. 30 J. P. Brennan, R. Schaeffer, A. Davison, and S. S. Wreford, J.C.S. Chem. Comm., 1973, 354. 241 5 Metalloboranes and Metal-Boron Bonding PY Br Br Figure 7 Proposed structure of the iron(Ir1) complex (py)FeBr2(B10Hl3py);each of the nine boron atoms B(l)-B(5) andB(7M10) carriesa terminal hydrogen atom, shown on& for B(9), in addition to the four bridge hydrogens indicated B6HlO to a value typical of B-B distances in triangulated boranes (N 182 pm). Each borane cluster is thus acting as a neutral dihapto-ligand and the three- centre two-electron B(4)B(5)Pt bonds are nearly perpendicular to the Cl-Pt-Cl axis. Figure 8 Molecular structure of (B6Hlo)~PtClz;typical interatomic distances are Pt-Cl 231.3 pm, Pt-B 227 pm, all 33-B 182 & 5 pm; angles about boron 60 2 2", internal angles of basal boron atoms 108 t 2" 4. Borane Anion plus Transition-Metal Complex This route to metalloboranes was first exploited in an extensive series of reactions by the du Pont group28 and has since been employed to obtain a wide variety of metalloboranes.The results will be reviewed in order of decreasing size of the co-ordinating borane moiety. Greenwood and Ward A. Higher Borane Anions.-A series of transition-metal complexes of n-B18Hm2-and i-B1sHzo2- have been prepared by direct reaction with phosphine- and carbonyl-substituted complexes of Co, Rh, Ni, Pd, and Pt.31 The compounds have the general formula (B~Hzo)M(ligand)~ and typical reactions are shown in equations (24)--(28).THP Na+2[n-BlsHzo]* + (PhsP)aNiClz __+ (n-Bi~Hzo)Ni(PhsP)z(red) (24) THF Na+2 [n-B~sHzo]~- + (diphos)MClz 4 (n-B~sHzo)M(diphos)(purple, Ni ;yellow, Pd) (25) {diphos = PhzPCHzCHzPPh2; M = Ni or Pd} THF Na+z [i-B1 sH2oI2- + (diphos)MClz (Call (i-B~eHzo)M(diphos) (red, Ni; yellow, Pd and Pt) (26) (M = Ni, Pd, or Pt) (i) THF Na+z[n-B18H20]~-+ ~coz(c0)s+ (ii) Me,NCI [Me4N]+[n-B1 sH~o)CO(CO)Q]- (orange-red) (27) MeCN [Me4N]+z[n-B1sHzo]~-+ (Ph3P)2Rh(CO)Cl + [Me4N]+ [(n-Bl sHzo)Rh(CO)Ph3P] (red) (28) The i.r., u.v.-visible, and 1lB n.m.r. spectra of several of these complexes were studied and it is suggested that the metal atom is bonded to one of the open faces of the B18H2O2- ion as a tetrahapto bidentate chelating ligand, as illustrated in Figure 9.Figure 9 Proposed structure of (n-B1 sHao)Ni(PhaP)a R.L.Sneatb and L.J. Todd,Inorg. Chem., 1973,12,44. Metalloboranes and Metal-Boron Bonding It is convenient to consider at this point some sandwich compounds formed by the reaction of thiaundecaborate anions with transition metals. During an extensive study of thiaboranes and their derivatives Muetterties and his co- workers showed that the reaction of the BroHloS2- dianion with a variety of transition-metal halides gave a series of icosahedral metallothiaboranes such as [(B~oHIoS)ZCO]-.~~Thus decaborane was found to react quantitatively with aqueous ammonium polysulphide to give BgHrzS-, which could then be con- verted into BloH11S- by pyrolysis at 200 "C.Subsequent treatment with a strong base such as butyl-lithium gave the required intermediate BloHloS2-.In THF solution anhydrous FeCI2 gave the complex [(B~oHIoS)~F~]~-, isolated as the pink bis(tetramethy1ammonium) salt or the maroon di-caesium monohydrate salt. These compounds could be polarographically oxidized to [(BI oHloS)zFe]- and partially halogenated with chlorine or bromine. A similar sequence of reactions using anhydrous CoClz gave the brown crystalline cobalt(1rr) salt [M~~N]+[(B~oH~oS)~CO]-;the yellow-orange complex Cs+[(B1 oHloS)zCo]- and the orange crystalline compound [(C~H~)~CO]+[(B~OH~OS)~CO]-were also characterized.The spectroscopic properties of these metallothiaborane anions encourage their formulation as sandwich compounds structurally similar to the carbollyl complexes of B gCzH112- (see Figure 10). Other sandwich compounds were prepared by ligand-replacement reactions, e.g. the yellow complex [(BloH10S)Pt(Et3P)z] and the colourless anion [(BIOHIOS)R~(CO)~]-.~~ d Figure 10 Proposed structure of the anion [(BioHloS)zCo]-$a W. R. Hertler, F. Klanberg, and E. L. Muetterties, Znorg. Chern., 1967,6,1696. Greenwood and Ward B. Deeaborate Anions.-The B10H13- anion reacts with a wide variety of transition-metal halide complexes to give three types of metal-decaborane complexes, [M(B1oHl2)2l2-, [M(B~oH~~)L~]-, in which the and [M(B~oH~~)L~], BloH12~- dianion can be considered to be acting as a tetrahapto bidentate chelating ligand.28 Typical reactions are as follows : 4B10H13-+ (Ph3P)2MC12 4[M(BioHi2)2I2-+ 2B10H14 + 2Ph3P + ZCl-(29)M = Co, Ni, Pd, or Pt 2BloH13-+ (PhsP)2M(CO)CI__+ [M(BioHi2)(CO)(Ph3P)21-+ BioHi4 + C1-M = Co, Rh,or Ir (30) 2BlOH13-+ (Et3P)2PtC12 --t [Pt(BioHlz)(EtsP)z] + B1oHi4 + 2C1- (31) This last reaction type also occurs with bis(triarylphosphine)dichloropalladium, but not when the bis(tria1kylphosphine)dichloropalladium analogues are used.In general, good yields were obtained from reactions (29)--(31) by using NaBl0H13 in diethyl ether or THF and then precipitating the anionic complexes with large cations such as Me4Nf or Bu3PHf. Extensive i.r.and n.m.r. data were reported on these compounds but the structure-motif was established by the single-crystal X-ray analysis of [Me4N]+2[Ni(B10H12)2]2-.33 The structure of the anion is shown in Figure 11; 2-Figure 11 The structure of the anion [Ni(BloH12)2]~-aa L.J. Guggenberger,J. Amer. Chem. SOC.,1972,94, 114. Metalloboranes and Metal-Boron Bonding it comprises two bidentate B10H1z2- units chelating on to the central nickel ion in such a way as to preserve planar bonding geometry. The relation to the structure of the [Zn(B10H12)2]~- ion in Figure 4 is obvious, the only difference being the relative orientation of the chelating borane clusters; thus, rotation of one of the borane clusters through 90"about the z-axis transforms one structure into the other.The shortest nickel-boron distances involved positions 5, 6, 9, and 10, with the Ni-B(5) and Ni-B(10) distances (211 and 218 pm) being signi- ficantly shorter than those of Ni-B(6) and Ni-B(9) (224 and 222 pm). Boron- boron distances were similar to those in decaborane, as expected, with an average B-B distance of 179 pm. The average B-H terminal distance of 122 pm and B-H bridge distance of 134 pm are also within the expected range. The angle subtended at the nickel atom from the midpoints of the B(5)-B(6) and R(9)-B(10) bonds is 80.8'. The caesium salts of [M(BloH~a)z]~- (M = Co, Ni, Pd, or Pt) were soluble in water and could be purified by crystallization as the tetramethylammonium salt from this solvent.The Me4N+ salts were moderately soluble in ethanol and very soluble in methyl cyanide, acetone, and DMF. They behave as normal 2:l electrolytes in methyl cyanide. Salts of the cobalt complex are burgundy red, the nickel complex is orange, and the palladium and platinum are yellow. There is also a significant difference in thermal stability, the cobalt and nickel complexes decomposing below 180 'C in nitrogen, the palladium complex at about 240 "C, and the platinum complex above 300 "C. A qualitative ordering of the hydrolytic stability of the dianion series is Pt > Pd > Ni S Coy the complexes being moderately stable at pH < 7 but rapidly degrading to BQH14- and metal hydrox- ides in basic solutions. The neutral complexes M(B~oH~~)(R~P)~ behave similarly. The dianion B10H142- can also be used to prepare complexes of B10H12~-, the overall reaction being:% 2B10H14~-+ M2+---t [M(BioHia)ala-+ 2H2 Thus CszB10H14 reacted with NiCl2 and CoCl2 in aqueous acetone to give the complex anions in 55 % yield, the products being isolated as the insoluble tetra- methylammonium saIts.Likewise, (cyclo-octa-1 ,5-diene)di-iodoplatium reacted to give [Me4N]+2[Pt(B10Hl2)2]2- in 30 % yield, and the analogous palladium complex was obtained in 25 % yield by a similar route. Despite these reactions, PtC12 itself gave only a 7% yield of [Pt(B1oH12)2I2- with B10H142- and PdClz was completely reduced ;zinc chloride was unreactive, as were (Ph3P)2Ir(CO)Cl and (PhsP)zPdClz. In a parallel study it was shown that the progress of reactions involving the B1oHls- anion is critically dependent on the counter-cation used.Thus, although reactions (29)-(31) proceed smoothly with NaBloHls in ethereal solution, the reaction of this salt with mercury(11) halides yields unstable products; however, Me4NBloHls reacts smoothly with HgCl2 in THF to give the desired product I4 A. R.Side and T.A. Hill, J. Inorg. Nuclear Chem., 1969.31.3874. Greenwood and Ward [M~~N]+~[H~(BIoHI~)~]~-.~Likewise NaBl0H13 could not be made to react with (py)2CoC12 or (py)2CoBr2, but when either Me4NBloH13 or E~~NHBIoHI~ was used, reaction (33) proceeded to give an 80% yield of the complex anions [(py)CoXz(BioHiipy)I-:% THF Me4NBloH13 + (py)~CoXz-+ [Me4Nl+[(py)CoX2(B10Hlipy)l-+ €32 (33) The brown crystalline products were insoluble in ethers, slightly soluble in methyl cyanide, and more soluble in nitromethane, in which they were shown to be 1:1 electrolytes.The effective magnetic moments at room temperature fell within the range 3.34.0 BM, consistent with the presence of CoIL,d7 with 3 unpaired electrons. The i.r. spectra show the presence of two types of pyridine, bound to the metal and to the borane cluster, respectively. These various pro- perties, in conjunction with the structures determined by X-ray diffraction already shown in Figures 4 and 11, strongly suggest that the structure of the anionic complex is as shown in Figure 12. In this structure the (BloHllpy-) moiety can be considered as a bidentate chelating ligand that is isoelectronic with B10H12~- in which one H-has been replaced by py as a ligand. Figure 12 Proposed structure of the anion [(py)CoCla(B~oHr~py)]-;the B(9) atom carries apendent pyridine Iigand andeach of the other nine boron atoms is covalently bonded to a terminal hydrogen atom (not shown) C.Smaller Borane Anions.-Three new classes of nido-metalloboranes were reported by the du Pont group in 1970.s6 The thiaborane anion BgHlzS- reacted with trans-[(Et3P)sPt(H)Cl] with evolution of hydrogen as indicated in equation (34): 36 N. N. Greenwood and D. N. Sharrocks,J. Chem. SOC.(A), 1969,2334.A.R Kane, L. J. Guggenberger, and E. L. Muetterties,J. Amer. Chem. SOC.,1970,92,2571. 247 Metalloboranes and Metal-Boron Bonding CsBeHmS + trans-[(Et3P)2Pt(H)CI] -+ cis-[(Et3P)zPt(H)BgH10S] + H2 (34) The neutral platinum-thiaborane complex was monomolecular in solution and has a Pt-H stretching vibration at 2214 cm-1.lH and 31P n.m.r. spectra indicate non-equivalent phosphine ligands, and the definitive structure of the complex was established by single-crystal X-ray analysis (see Figure 13). It can be seen that both platinum and sulphur are incorporated within the 1 1-atom icosahedral fragment. The Pt-S distance is 243 pm and the three close Pt-B contacts are in the range 220-225 pm. The BgHloS- ligand is isoelectronic with B10H1z2- considered in the preceding subsection, thereby implying that the platinum co-ordination polyhedron is a quasi-five-co-ordinate distorted square pyramid.n U Figure 13 Structure of (Et3P)2Pt(H)BsHloS; each of the nine boron atoms is bonded to a terminal hydrogen atom (not shown) and a further borane hydrogen atom (also not depicted) is probably associated with one or more of the atoms in the open face of the icosahedral fragment The other new types of metalloborane were (E~~P)zP~B~HIILand (EtsP)zPtBsHiz. Thus the B9H12-anionreacts with trans-[(Et3P)2Pt(H)CI] in the presence of donor molecules, liberating one mole of hydrogen to give the PtB9 complexes, and these can be degraded in alcohol to the PtBs species:36 CsBsHlz + trans-[(Et3P)2Pt(H)CI] + L + (Et3P)zPtBgHiiL + CSCI + Hz (35) (Et3P)2PtBgHiiL + 3EtOH +-(EtsP)zPtBsHi2 + B(OEt)3 + L + H2 (36) Organic ligands based on amines, nitriles, phosphines, and sulphides function Greenwood and Ward effectively as L in reaction (35).1H and 31P n.m.r. spectra do not distinguish between the two Et3P ligands bonded to platinum, and the most reasonable structure is that shown in Figure 14, in which the BgHllL2- ligand is effectively Figure 14 Proposed structure of (Et3P)2Pt(BgHiiL) trihapto-bidentate, with the square-planar platinum orbitals bisecting the B(2)-B(5) and B(2)-B(7) edges. The PtB8 spies could then plausibly be derived from this by elision of BH(L)2+ and the addition of two bridge protons to give the structure shown in Figure 15. Figure 15 Proposed structure of (Et3P)2Pt(BaHiz) Metalloboranes and Metal-Boron Bonding The metalloboranes (Ph3P)zCuBsHs and (PhsP)2CuBsHs have been syn-thesized by reaction of the corresponding potassium boranates with tris(tri- phenylphosphine)copper(r) chloride in the mixed solvent methylene chloride- THF at -78 KB5Hs + (PhsP)aCuCl-+ (PhsP)aCuBsHs + KCl + Ph3P (37) KB~HQ+ (Ph3P)3CuCl 4(PhsP)&uBaHg + KCl + PhsP (38) Both complexes are white orystalline solids and are the first air-stable derivatives of B5H9 and B6H10, respectively.The parent boranes can be regenerated in 75 % yield by treating the complexes with ethereal hydrogen chloride. The structure of the complexes has not been elucidated but may consist of a bridging (PhsP)zCu group spanning two of the basal boron atoms in each case. In a short communication it has been reported that NaBsHs reacts with an excess of CoC12 and NaCsHs in THF to give a wide range of The new compounds isolated so far are shown in Figure 16, and include dark-red needles x, 0 0 Fig 16 (a and b) ( b) V.T. Brice and S. G. Shore, Chem. Comm., 1970, 1312. V. R. Miller and R. N. Grimes, J. Amer. Chem. SOC.,1973,95,5078. Greenwood and Ward 0 3 8 0 Figure 16 Proposed structures for cobalt-borane complexes (see text) ;the terminal hydrogen atoms in (f 1 have been omitted for clarity 25 1 Metalloboranes and Metal-Boron Bonding of (~T-C~H~)CO(B~H~) (a), which are converted into the pale-yellow crystalline isomer (b) by gas-phase pyrolysis at 200 "C for thirty minutes; (7T-C5H5)2- Co2(1,2-B4Hs) (c), which is a violet, crystalline solid isoelectronic with C2B4H6 and BsHs2-; (~T-C~H~)~CO~(~-C~CIO-C~H~-~,2-B4H5) (d), which contains equi- valent (7-CgH5) groups; (~T-C~H~)~CO~(~-C~C~U-C~HQ-~,2-B&) (e), which con- tains non-equivalent (7r-CsH5) groups; and (T-C~H~)CO(S-B~H~~) (f), which is a red solid, isoelectronic with decaborane(l4).The proposed structures in Figure 16(a)--(f) have been formulated on the basis of lH and 1lB n.m.r. spectra and mass spectrometric evidence. An intriguing series of .rr-borallyl(7r-BsH72-) complexes with nickel, palladium, and platinum have recently been synthesized and the stabilities have been found to increase in the order Ni < Pd < Pt.39p40 The~BsH7~- dianion is isoelectronic with the wallyl ion 77-C3W5-.The platinum complexes were synthesized by allowing caesium octahydrotriborate(1- ) to react with cis-[(RsP)2PtC12] in the mixed solvent methyl cyanide-triethylamine : CSBsH8 + ci~-[(RsP)2PtC12]+ EtsN --t (RsP)2PtB3H7 + CSCl + EtsNHCl (39) The complexes were white, or off-white, and typical phosphine ligands were EtsP, EtzPhP, PhsP, MezPhP, (p-tOl)sP, and Ph2PCH2CH2PPh2. The structure determined by X-ray diffraction of the bis(dimethy1phenylphosphine) complex is shown in two projections in Figure 17. The two phosphine ligands and the Figure 17(a) 38L.J. Guggenberger, A. R. Kane, and E. L. Muetterties, J. Amer. Chem. SOC.,1972, 94, 5665. 40 E. L. Muetterties, Pure Appl. Chem., 1972, 29, 585. Greenwood and Ward Figure 17 (a) The molecular configuration of (Me2PhP)2Pt(wB3H7), all hydrogen atoms being omitted for clarity; (b) a side view of the molecule, illustrating the angle between the n-borallyl plane and the P(l)PtPf2)plane.(The two phosphorus atoms overlap in this view) asymmetrically bonded n-BsH 72-ligand complete an essentially square-planar co-ordination environment around the platinum. The PPtP bond angle is 96" and the Pt-P distances are 230.1 and 231.1 pm. The Pt-B distances are 238 pm to B(l) in the PPtP plane and 218, 213 pm to B(2) and B(3), whose mid point lies in this plane. The interboron distances are 186 k 5 pm for B(l)-B(2) and 192 2 4 pm for B(2)-B(3), and the dihedral angle between the B3 plane and the PtP2 planeis 117".TheB(l)-B(3) distance is 31 5 pm, i.e. non-bonding. Although the hydrogen atoms in the n-BsH72- ligand were not specifically located because of disorder, they are considered to be similar to those in B3H8- [in which B(l) and B(3) are directly bonded]. The bonding is perhaps best considered in mol- ecular orbital terms as an essentially =-bonded B3H72- group but in the valence- bond 0-n formalism the dsp2 hybrid orbitals on the platinum would be directed to B(3) and to the mid-point of B(l)-B(2), as indicated by the resonance form in Figure 18. The i.r., X-ray, photoelectron, and n.m.r. spectra of the complexes were also studied, and the lH n.m.r. spectra in particular showed that the n-borallyl complexes were stereochemically more rigid than the (hydrogen-bridged) cFB3H8- metalloborane complexes.An interesting additional reaction of the n-borallylplatinum complex was the facile replacement of the n-B3H72- ion by Metalloboranes and Metal-Boron Bonding H Figure 18 A valence-bond description of the bonding in the n--borallyl complex shown in Figure 17 trialkylphosphines to produce the first examples of tetrakis(trialky1phosphine)-platinum(0). For example, a solution of (Et3P)2PtB3H7 when heated at 125 "C for 10min in Et3P gave a dark orange-red solution from which off-white crystals of (Et3P)4Pt were obtained on cooling to -40 "C. 5 Metathesis (Interchange of Metal Atoms) Several metalloboranes of types similar to those discussed in preceding sections can be prepared by metathetical reactions. Thus, the red complex [M~~N]+~[CO(B~OH&]~-can readily be prepared by treating a solution of [(THF)2Zn(BloH12)]2 in acetone with a solution of anhydrous CoCl2 in the same solvent followed by precipitation with aqueous Me4NCl. In view of the known behaviour of the dimer in solution [see equation (1l)] it seems likely that the reaction proceeds as follows : MeSCO [(THF)2Zn(BloH12)]2 __+ Zn2+ [Zn(B1oH12)2I2- 3-4THF Me,CO[Z~(BIOHIZ)Z]~-+ CoC12 __+ [Co(B1oH12)2l2-+ ZnCl2 (40) Another typical metathesis occurs when [Me2TI]+[BioH12TlMe2]- reacts with MeHgCl in THF; an orange crystalline product was obtained as shown in equation (41).21 THF MeHgCl + [Me~Tl]+[B10Hr2TlMe21-+ [Me2Tl]+[B1 oHlzHgMe,THF]- + MezTlCl (41) Greenwood and Ward The product was then allowed to react with PhsMePIBr in dichloromethane to give crystals of unsolvated [PhsMeP]+ [BI ~HlzHgMel-.Spectroscopic proper- ties suggest that the anion has the chelated structure shown in Figure 19, though a bridging p-HgMe structure is not ruled out. Figure 19 Proposed structure of [B1oHl2HgMe]-;the terminal hydrogen atoms attached to the boron atoms and the methyl carbon atom are not shown The related reaction type of reductive metathesis was used in a very early study to prepare the copper(x) metalloborane CU~BIOH~O from an aqueous solution of the ionic closo-borane Na2B10Hlo and an excess of the copper(r1) salt Cu(C104)~.*~ The detailed course of the reaction was not investigated but the colourless crystalline product was subjected to a detailed X-ray structural analysis.This established that the B1oH102- unit was indeed a bicapped square Archimedian antiprism, as had been postulated on the basis of n.m.r. data, but the most interesting structural feature was the close approach of the copper atoms to the centre of diagonally opposite edges of the borane cage as indicated in Figure 20(a). The actual Cu-B distances were in the range 214-233 pm, with a mean value of 220 pm, close to the estimated Cu-B distance of 213 pm for a covalent two-electron three-centre BCuB bond. Each copper@ atom can be considered to be sp-hybridized and is equidistant from two such edges, so that the bonding environment of copper is as shown in Figure 2qb).The angle subtended by the two edge-boron atoms at the copper atom is only 46", thus making a four-co-ordinate bonding model for copper less attractive. In the extreme view there- fore the cZoso-BloHd- anion can be considered as a tetrakis(dihapt0) ligand donating one pair of electrons to each of four Cu+ cations; at the opposite extreme the structure can be considered as a cross-linked covalent polymer [Cud31 oHi 01~. 41 R.D.Dobrott and W.N.Lipscomb,J. Chem.Phys., 1962,37,1779. 255 Metalioboranes and Metal-Boron Bonding cu 2Q4pm205p7 \/* cu cu ( a) Figure 20 (a) The structure of the ten-atom boron cluster in Cu2BloH10 showing the close approaches of four copper(1) ions to opposite edges of the polyhedron; (b) the digonal dative-bonding environment of the copper 6 Borane Anion plus Metal Halide (or Alkylmetal Halide) It was mentioned in Section 2 that, unlike the dialkyls and diaryls of zinc and cadmium, organomercury compounds did not react with decaborane to give (R20)2Hg(BloH12), and that alternative routes to the derived bischelated anion [Hg(BloH12)2]2- had to be devised.22 It was also found that NaBloH13 did not give a satisfactory reaction with methylmercury(r1) halides.42 However, deca- boranylmagnesium iodide reacted with methylmercury(I1) halides in ethereal solution to give air-sensitive, orange crystalline solids of general formula M~~[H~(BIOHI~)~]X~O,~E~~O.The rates of reaction were in the order Cl > Br B-I, as expected from the polarities of the respective Hg-X bonds, and indicate that the borane is acting as a pseudoanion BIOHI~~--M~I~+.The compounds dissolved in water or ethanol, from which the orange tetramethylammonium salt [Me4N]+2 [Hg(BloH12)2]~- could be precipitated. The original products should thus be formulated as the mixed salts Mg[Hg(BioH12)2],5MgX2,nEtsO.The [Hg(B10H12)2]2- anion was found to be isostructural with the zinc and cadmium analogues (see Figure 4) and the high-resolution llB n.m.r. spectra of all three species have been studied in detai1.22 The i.r. spectra of the three tetramethyl- ra N. N. Greenwood and N. F. Travers, Chem. Comm., 1967, 216. Greenwood and Ward ammonium salts were identical except for a single band, which can therefore be assigned to the asymmetrical metal-boron stretching vibration of the tetra- hedral MB4 unit (see Figure 5).This occurs at 278, 235, and 225 cm-l for the zinc, cadmium, and mercury complexes, respectively. An alternative route to [Hg(BloHi2)zl2-is by allowing two moles of MeeNBloHls to react with mercury(??) chloride in THF:35 2Me4NBloH13 + HgClz -+ [Me4N]+z[Hg(Bi0Hiz)z]~-+ 2HC1 (42) An unidentified green powder was also formed in addition to the orange crystal- line mercury complex. The reaction is particularly sensitive to conditions and could not be run successfully with other salts of B10H13-, other mercury halides, or other solvents. The decaborane dianion BloH122- has also been used successfully to synthesize metalloborane complexes from metal halides and alkylmetal halides.Thus, (Ph4As)zB1oH12 reacted with anhydrous SnCl2 in two ways in a mixed THF- methylene chloride solvent as indicated in equations (43) and (44).43 (Ph4As)zB10H12 + SnC12 -[Ph4As]+z[BroH10]~-+ Sn + 2HCl (43) (P~~As)~BIoHIz (44)+ SnCh -+[P~~ASI+~[(B~OH~~)S~C~~I~-In reaction (43) the chlorine is acting as a proton abstractor which transforms the nido-B1oH1z2- anion into the cZoso-BloH102- anion with precipitation of elemental tin. In reaction (44)the B10H1z2- anion acts as a bidentate chelating ligand to give the desired stannaborane complex. The llgSn Mossbauer spectrum of [(BI 0Hlz)SnClz]2- at liquid-nitrogen temperature established the continuing presence of SnII rather than SnIV, the chemical isomer shift with respect to BaSnO3 being 8 = 3.17 mm s-l and the quadrupole splitting d = 1.26 mm s-1 The basis for this conclusion rests on the observation that the chemical isomer shift for tin(I1) compounds is always greater than 2.9 mrn s-l and that for tin(1v) compounds is always less than 2.0 mm s-l with respect to BaSn03.4Q The struc- ture of the complex anion is considered to involve a tetrahapto-borane group as shown in Figure 21.A related series of reactions of (P~~As)~BIoHI~ with dimethyltin(1v) chloride also involved both disproportionation and co-ordination, as set out in equations (45) and (46).43 (Ph4As)zBloH12 + MezSnClz --+ [P~~As]+~[B~oH~o]~-+ 2H+ + Me2SnCh2-(45) (P~~As)zB~oH~z (46)+ MezSnCle -+[P~~As]+~[(BIoH~~)S~(M~~)CI~]~-The structure of the metalloborane anion is considered to involve a tetrahapto- bidentate chelating B10H12~-unit as in Figure 17 but with the bonding environ- 4R N.N. Greenwood and B. Youll, to be published; B. Youll, Ph.D. thesis, 1970, Univ. Newcastle upon Tyne, England. 44 N. N. Greenwood and T. C. Gibb, ‘MiJssbauer Spectroscopy’, Chapman and Hall, London, 1971, p. 375. 257 Metalloboranes and Metal-Boron Bonding 2-WL Figure 21 Proposed structure of the anion [(B1oH12)SnC12l2-ment about the tin atom expanded from four-co-ordinate to six-co-ordinate. The presence of Snm was confirmed by the Mossbauer parameters: chemical isomer shift 8 = 1.75 mm s-1 with respect to BaSnOs, quadrupole splitting d = 3.13 mm s-1. The corresponding diethyl derivative was also prepared but the diphenyl analogue was less stable.An alternative route to stannaboranes utilized the reaction of NaBloH13 with MesSnCl in ether at 75 "C; MesGeBr reacted similarly.4B NaBloHls + MesSnCl --+ MezSnBloH12 + NaCl + MeH (47) NaBloH13 + MesGeBr --+ MezGeBloHi2 + NaBr + MeH (48) The air-stable products melted at 123 and 82 "C,respectively, and their structure is considered to be as in Figure 21 but with Me2Sn2+ or Me2Ge2+ groups replacing the SnC12. The +4 oxidation state in the neutral complex was confirmed by reaction with HCI to yield MezSnCla and regenerate BlOH14 in >95% yield. The analogous reaction with DCl gave the bis bridge-deuteriated borane P-BI 0H12D2.The reaction of BloH13- with the corresponding silane MesSiCl is strongly dependent on the alkali-metal cation used, the ethereal solvent chosen, and the 46 R. E. Loffredo and A. D. Norman, J. Amer. Chem. Soc., 1971,93,5587. 258 Greenwood and Ward temperature of reaction [see comments above on reaction (42)]. Thus, solvated LiBl0H13 prepared in diethyl ether at 35 "Cor in THF at 100 "C gave no reaction but LiBloH13,THF prepared at 50 "C reacted smoothly with an excess of MesSiCl over a period of 1 day at 120-140 0C:46 LiBloHls,THF + MesSiCl Med3iBloH13,THF + LiCl (49) The structure of the red, hygroscopic oily product was not determined. Likewise, various solvates of NaB10H13 and KBioHls were either unreactive or reacted in only small yields whereas unsolvated ~1OH13, formed by the reaction of BioH14 and an excess of KOH in THF at -30 "C, reacted well.The product presumably comprises a decaborane skeleton in which either a bridge or a terminal proton has been replaced by the SiH3 group, as established in the analogous derivatives of pentaborane(9) to be discussed in the next paragraph. Lithium octahydropentaborate(1- ) reacts with a variety of trialkyl Group IV metal halides in diethyl ether at -30 "C to give base-bridged products of general formula p-RsM-BsHs :47s48 LiBsHs + R3MX 3 pR3M-BsHs + LiX (50) M = Si or Ge; X = C1, Br, or I; R = H, Me, or Et; M = Sn or Pb; X = C1, Br, or I; R = Me; The structure of these dihapto-metalloboranes was first deduced from their llB and lH n.m.r.spectra and later confirmed by an X-ray single-crystal structure determination on l-Br-p-Me3Si-BsH 7 (see Figure 22).4Q This compound was chosen because p-H3Si-B5H8 is a liquid at room temperature and p-Me3Si-BsHs melts at 17 "C; the bromo-derivative was prepared by direct bromination of p-Me3Si-BsHs using bromine at room temperature. The structure is closely related to that of pentaborane(9), the p-Me3Si-group replacing one of the bridging protons, and the bonding is best considered as involving a two-electron, three-centre bond between B(2)SiB(3). The distance of the Si atom from the mid-point of B(2)B(3) is 216 pm. The compound p-Me3Si-BsHs was shown to be monomeric by mass spectro- metry. It was hydrolysed quantitatively by aqueous hydrochloric acid according to equation (51): The bridged Me& and MesGe-derivatives of pentaborane isomer& quanti- tatively at room temperature to the corresponding 2-substituted derivatives when in the presence of weak Lewis bases such as Me20 or Et20, and the progress of this isomerization could be followed readily by n.m.r.~pectroscopy.~7 The corresponding compounds p-Me3Sn-BsHs and p-Me3Pb-BsHs were too unstable to allow the isomerization to be studied in this way. 48 E. Amberger and P. Leidl, J. Organometallic Chem., 1969, 18, 345. 47 D. F. Gaines and T. V. Iorns,J. Amer. Chem. SOC.,1968,90,6617. 4a D. F. Gaines and T. V. Iorns, J. Amer. Chem. SOC.,1967,89,4249. 49 J. C. Calabrese and L. F.Dahl, J. Amer. Chem. SOC.,1971, 93, 6042. 259 Metalloboranesand Metal-Boron Bonding Figure 22 Structure of 1-Br-p-Me3Si-B5H7 and main interatomic distances. The dihedral angle between B(l)B(2)B(3) and B(3)B(2)Si is 182", between B(l)B(3)B(4) and B(4)B(3)H(3,4) is 194", and between B(l)B(4)B(5) and B(5)B(4)H(4,5) is 204" A1 t hoiigh metallocarbaborane complexes are not being considered specifically in this review, a number of R3M derivatives of the nido-carbaboranes C2B4Hs and (MeC)2B4Hs in which the metal atom replaces a bridging hydrogen atom between two basal boron atoms will be discussed since their structures are very similar to those of the ~-R~M-B~HE species just considered, both sets of compound involving a three-centre two-electron BMB bond.Thus NaC2B4H7 reacted with R3MCI (R = H or Me; M = Si or Ge) in THF at 0 "C to give good yields of p-H3M-C2B4H7:50 NaC2B4H7 + R3MC1 --t pL-R3M-C2B4H7+ NaCl (52) An alternative route using the lithium salts gave good yields only in the case of p-H3Si-C2B4H7. The compounds were structurally characterized by their lH and 1lB n.m.r. spectra and their mass spectra, and in each case were shown to have the p4,5-bridged-2,3-dicarba-nido-hexaborane(8)structure illustrated in Figure 23. The compounds were colourless liquids with low vapour pressures in the range 14.1 mmHg at room temperature. When heated to temperatures in N the range 80-200 "C (depending on the compound) they readily isomerized quantitatively to the terminal 4-R3M-derivatives. Heating 4-Me3SiC2B4H 7 at 220-400 "C yields the carbon-substituted derivative 2-MesSiCzB4H7, and M.L. Thompson and R.N. Grimes, Inorg. Chern., 1972, 11, 1925. Greenwood and Ward further heating converts the product into derivatives of the closo-carbaboranes CaBsH5 and C2B5H 7, presumably by disproportionation of the closo-CzB4 cage. Figure 23 Proposed structure ofp-RsM-C2B4H7 (M = Si or Ge; R = H or Me) Subsequent work showed that NaGB4H 7 reacted smoothly with Me3SnBr and MesPbCl in THF at room temperature to yield the corresponding p-tin and -lead derivatives as involatile, colourless liquids (p < 0.1 mmHg at 25 "C).51 The structures were again established by n.m.r. techniques. By contrast to the silicon and germanium compounds, however, the tin and lead derivatives showed no tendency to isomerize from bridge- to terminal-substitution: the tin compound was stable indefinitely at 120 "C but decomposed rapidly at 220 "C to give MesSnH and the parent carbaborane C2B4H8; the lead derivative was less stable and decomposed to the corresponding products in 2 days at room temperature and more rapidly at 120 "C.Both compounds reacted quantitatively with HCl and DCl to give the parent C2B4Hs and p-DC2B4H7 respectively. In an extension of these studies a bis(trimethylsily1) derivative was obtained as indicated in the Scheme.51 The product contained both a bridging B(4)B(5) -and Scheme O1 A. Tabereaux and R.N. Grimes, Inorg. Chem., 1973,12,792. Metalloboranes and Metal-Boron Bonding a terminal B(6)-trimethylsilyl group.The related p,p‘-bis(carbaborany1)silane was obtained as a white solid by reaction (53), and two possible structural isomers are illustrated in Figure 24. OBH .CH OH Figure 24 Proposed structures of the two possible geometrical isomers of p,p’-(CzB4H7)2SiH2 THF 2NaGB4H7 + SiHzCl2 -+ p,p’-(CaB4H7)2SiH2 + 2NaCl (53) A further variation was obtained by allowing the nido-carbaborane anion (MeC)gB4H5- to react with methylhalogeno-silanes and -germanes in ether, and an unstable product containing an added bridging boron atom was likewise synthesized using MezBCl :62 Na(MeC)2B4H5 + MesSiCl3 2,3,pMe3Si(MeC)zB4Hs + NaCl (54) Na(MeC)2B4H5 + MezSiClz --+ 2,3,pL-Me2Si(C1)(MeC)zB4H5+ NaCl (55) Na(MeC)2B4H5 + MesGeBr -+ 2,3,pMesGe(MeC)zB4Hs + NaBr (56) Na(MeC)2B4H5 + MezBClj 2,3,pMezB(MeC)2B4H5 + NaCl (57) Sodium can be replaced by lithium in these reactions. Of the four compounds so prepared, only the MesGe derivative isomerized to the terminal isomer.7 Borane (or Borane Anion) plus Transition-metal Carbonyl CompIex The reaction of cleaved diborane in THF solution with the carbonyl anions of manganese and rhenium has already been mentioned in Section 3. A variety of other reaction types have also been reported. sa C.G. Savory and M.G.H.Wallbridge, J.C.S. Dalton, 1972,918. Greenwood and Ward A unique carbonyl insertion into the BlOHl3- anion has been effected by irradiating a THF solution of NaBloH13 and Group VI metal hexacarbonyls at 350 nm in the absence of air:= hv,THFBioHis-+ M(CO)e __+ [(BioHioCOH)M(C0)4]-+ H2 + CO (58) In reaction (58) M = Cr, Mo, or W, and the metallocarbonylborane was isolated in 50% yield as the tetramethylammonium salt.Treatment of these products with sodium hydride in THF gave quantitative yields of the yellow tetrabutylammonium salts of the dianion [BIOHIOCO&CO(CO)~]~-: THF [(BioHioCOH)M(C0)4]-+ H-__+ [BIoH~oCO&CO(CO)~]~-+ H2(59) The structure of the molybdenum compound was established by single-crystal X-ray analysis and found to contain a novel ester link between the oxygen atom of the carbonyl group inserted into the BIOcluster and a carbon atom of an adjacent metal-bonded carbonyl group, as shown in Figure 25.It is also clear fromthe Figure that the dianion comprises a closo-icosahedral BloCMo cluster in which the metal and carbon atoms are linked via a cyclic MC(0)OC grouping. The monoanion [equation (59)] can be regenerated from the dianion (Figure 25) by direct protonation with aqueous acid under anaerobic conditions; this 0C OBH Figure 25 Structure of the dianion [BloHloCOMoCO(CO)3]2-with terminal -hydrogen atoms omitted for clarity; typical interatomic distanceslnm are shown 53 P. A. Wegner, L. J. Guggenberger, and E. L. Muetterties, J. Amer. Chem. SOC.,1970, 92, 3413. Metalloboranes and Metal-Boron Bonding cleaves the ether linkage to give a simple tetracarbonyl of the BloCM icosahedron in which there is no bonding between the carbonyl groups and the cage. All spectral data are consistent with this formulation.63 The mono- and di-anions of these metalloboranes of chromium, molybdenum, and tungsten undergo a further unusual reaction with base; this partially degrades the icosahedral cage with removal of the carbon atom, not a boron atom as normally occurs with carbaboranes.The resulting dianions [B10H12M(C0)4]2- are very air-sensitive, especially in acid solutions, but there is no hydrolysis in the absence of oxygen; their structures are presumably analogous to those of the B10H122- chelating complexes considered in Section 4(b). A phosphaundecaborane-metal carbonyl complex has been made by first deprotonating decaborane with sodium hydride in ethereal solution and then adding an ethereal solution of PhPClz to yield the nido-phosphaundecaborane PhPBloH12 in moderate yields; this was then allowed to react with two equi- valents of sodium hydride and one equivalent of (C0)sMnBr in THF to give the yellow 'metal1ocene'-type ion [(PhPB1oHlo)Mn(CO)3]-, which was isolated as the tetramethylammonium salt 9 Et,O NaBloHl3 + PhPC12 + NaH---+PhPBloHla + 2NaCl + Ha (60) THF PhPBloH12 + 2NaH + (C0)sMnBr __+ Me4NC1 [Me4N]+[(PhPBloHlo)Mn(C0)3]-+ NaCl + NaBr + 2H2 + 2CO (61) U.V.,ix., and n.m.r.data were reported, and the presumed 12-atom cluso-structure is shown in Figure 26. Very recently it has been shown that the B9H14'- anion reacts with (C0)sMnBr in THF to produce several metalloboranes, from which the anion [(BgH13)Mn(C0)3]-and the solvate [(BgH12)Mn(C0)3,THF] have been iso- lated.55 The air-stable reddish-orange crystalline salt [Me4N]+ [BgH13Mn(C0)3]-, m.p.200 "C (decomp.), was obtained in 40% yield and shown to be a 1:1 electrolyte in methyl cyanide. The neutral solvate is also an air-stable reddish- orange crystalline material, m.p. 189 "C(decomp.), that can be sublimed in high vacuum at 50-70 "C. Both compounds have unusually high thermal, hydrolytic, and oxidative stability compared to that of other metalloboranes. The single- crystal X-ray structure of the neutral solvate is illustrated in Figure 27 and shows several unique features. The Mn(C0)3 moiety can be considered to be substitut- ing the B(6) position in the decaborane cluster, being bound to the B9 ligand by two three-centre two-electron hydrogen-bridge bonds, BHMn, and a single boron-metal bond, B(2)-+Mn, that contributes two electrons to the metal.The l39 ligand is thus formally a terdentate five-electron donor, replacing two CO and one Br to preserve the 18-electron environment about manganese. The other unusual feature of the structure is the directly bonded THF donor which is tib J. L. Little and A. C. Wong, J. Amer. Chem. Soc., 1971, 93, 522. 66 J. W. Lott, and D. F. Gaines, H. Shenhav, and R. Schaeffer, J. Amer. CRem. SOC., 1973, 95, 3042. Greenwood and H7ard Figure 26 Proposed structure of the anion [(P~PBIoH~o)M~(CO)~]-Figure 27 fie structure of [(B sH12THF)Mn(C0)3] omitting terminal hydrogen atoms on tetrahydrofuran and the eight boron atoms 1,3-5, and 7-10 for clarity Metalloboranes and Metal-Boron Bonding attached insteadof ahydrogenatom to B(2).Typical bond distances are Mn-CO 179.3pm, Mn-H 175 pm, Mn-B(2)219.6 pm, Mn-B(5) 223.7 pm, and0-B(2) 152.6 pm. The bond angles subtended mutually by the three carbonyl groups and the two bridging hydrogen atoms at the manganese atom are all close to 90" but the angle B(2)MnH(bridge) is only 77.5", and the angle B(2)MnCO(trans) is 161.4'. The carbonyl ligands in both the anion and the neutral complex can be replaced successively by triphenylphosphine ligands under selected photolytic conditions, and even the complexed THF in Figure 27 can be replaced. Hexaborane(l0) can act as a dihapto-two-electron donor in the volatile, yellow, crystalline complex p-(BsH1o)Fe(CO)c, m.p.38 "C (decornp.). This is prepared by direct reaction of hexaborane(l0) with Fez(CO)g, as shown in reaction (62).56 On the basis of ir., n.m.r., and mass spectrometric data the 2B6Hl0 + Fe2(C0)9 ---f 2&-@6Hlo)Fe(C0)4] + CO (62) complex is formulated as a derivative of Fe(CO)5 in which one of the equatorial carbonyl ligands has been replaced by the unique B(4)-B(5) bond in the basal plane of B6H10, thereby forming a three-centre twoelectron BFeB bond as indicated schematically in Figure 28. The complex can be handled for brief periods in air but is thermally and photolytically unstable. (See also footnote on p. 232). co NCOFe 'co co Figure 28 Schematic representation of the structure of k-(BeHlo)Fe(C0)4] Direct (monohapto-) o-bonds between a borane and a metal carbonyl can be synthesized by the reaction of chloro- and bromo-derivatives of B5Hp with NaMn(C0)s and NaRe(C0)a baction (63)].Only basal-halogenated penta- boranes 2-XBsH8 would react, and in no case were apex-halogenated penta- boranes 1-XB5H8 successfully used.57 EtsO 2-XBsH8 + NaM(C0)5 I_+2- [M(C0)5]B5Hs (63)X = C1 or Br, M = Mn or Re 66 A. Davison, D. D. Traficante, and S. S. Wreford, J.C.S. Chem. Cornm., 1972, 1155. M D. F. GainesandT. V. Ioms,Inorg. Chem., 1968,7,1041. Greenwood and Ward The rhenium compound melts at 11 "Cand is stable at room temperature for several hours but the manganese derivative, upon melting at -10 "C, develops a yellow colour which becomes more pronounced as the sample is warmed to room temperature for several days. Spectroscopic properties indicate the struc- ture shown in Figure 29.co Figure 29 Schematic representation of the structure of 2-[M(C0)5]BsHa (M = Mn or Re) The reaction takes a dramatically different course when lithium octahydro- pentaborate(1-) is treated with pentacarbonylmanganese hydride:40 LiB5Hs + HMn(C0)5 -+ Li+[(B5H5)Mn(C0)3]-+ 2Hz + 2CO (64) Little information on the compound has yet been published but it is considered to comprise a B5Hs2-tetrahap to-ligand, electronically analogous to cyclo-butadiene, in which the square open face of the BS moiety is n-bonded to the metal atom as indicated in Figure 30.. co Figure 30Proposed structure of the anion [(Bas)Mn(CO)sr Metalloboranes and Metal-Boron Bonding 8 Concluding Summary of Structures and Bond-types Despite the considerable diversity of structures and bond-types adopted by metalloboranes, a general pattern is now beginning to emerge which enables the borane ligands to be classified according to their connexity (polyhapto-nature)58 and the number of electrons formally donated from the borane to the metal. It is a measure of the changed emphasis brought about by the synthesis of metallo-boranes that it is the electron-donor capacity of the boranes rather than their electron-acceptor capacity which is being exploited. Indeed, of all the metallo- boranes mentioned in this review, the only ones in which the borane unam- biguously acts as an electron-acceptor are the monohaptocomplexes of man-ganese and rhenium, [(CO)5M-+BH3]- and [(Ph3P)(C0)4Mn+BH3]-.The germanylborane anion [H3GeBH3]- could likewise be regarded as an adduct of the hypothetical anion GeH3-, in which BH3 is acting as an electron-pair acceptor [H3Ge-+BH3]-, although it is also possible to formdate this adduct as a germyl-substituted BH4- anion in which both H3Ge and BH3- contribute one electron to the covalent bond.16 The number of boron atoms involved in bonding to the metal, and the number of electrons formally donated from the borane ligand to the metal acceptor, are summarized in the Table. The classic two-electron a-donor-acceptor bond is exemplified by the monohapto-ligand pyBloH13-, which donates a pair of Table Classification of borane donors in metalloboranes(n = number of electrons donated by the borane moiety) Ligand n Examples Fig.Ref. monohapto aBH3-1 [H3GeBH3]--16 'B5H8 1 2-Me3M-B5Hs M = Si or Ge -47 1 2-[M(C0)5]B5H8 M = Mn or Re 29 57 pyB1 oHi3- 2 [(PYB~0Hl3)Fe(py)Brz1 7 29 dihapto B5H8-2 pu-R3M-B5H8 M = Si, Ge, Sn, or Pb 22 47,48 B~H~BT 2 p-MesSi-BsH7Br 22 49 B6H10 2 p-B 6H10Fe(C0)4 28 56 2 (B6Hio)zPtCh 8 30 7CzB4H7-2 P-R~M-C~B~H M = Si or Ge 23 51 .BioHiz. 2 [(B~oH~z)AIHz]-(chelate) 1 14 trihapto 7-T-B~H 3 (RsP)zPt(r-BsH7) (also Ni, Pd) 18 39,40 (or 77-B3H72-) 4 (R~P)~P~(T-B~H~)(also Ni, Pd) 17 39, 40 B4H8 2 ~~-CSH~)CO(B~HS)(donor atoms BH2) 16a 38 B8H1Z2-4 (Et3P)d'tBaHia 15 36 BgHiiL2-4 (Et3P)2€'t(B~HiiL) 14 36 CzB4Hs 3 6-Me3Si-pL-Me3Si-C2B4Hs -51 (i.e.mono-plus di-hapto) &* F. A. Cotton,J. Amer. Chem. SOC.,1968,90,6230. Greenwood and Ward Ligund n Examples Fig. Ref. tetrahapto B4H8 2 (~T-C~H~)CO(B~H~)(donor atoms B4) 16b 38 B5H52-4 [(B5HdMn(C0)31-30 40 B9H13-2 (T-C~H~)CO(~-B~H~~)(donor atoms B3H) 16f 38 B10H12~-4 [M(B1oH12)zl2-M = Mg, Zn, Cd, or Hg 4 23 4 [M(BloHi2)2l2-M = Co, Ni, Pd, or Pt 11 4 [Pt(Bi oHi z)(Et 3P)2 3 -28 4 [M(BioHiz)(CO)(Ph3P)z]- M = CO, Rh, Ir -28 4 [(B~oHia)HgMe]-19 21 4 [(BI oHlz)TlMez]- 6 25 4 [(BioHi2)SnC1~]~-,[(BioHiz)Sn(Me2)Cl~]~-21 43 4 [(BloHiz)GeMez], [(B10Hiz)SnMe2] 21 45 4 [(BIoHI~)M(CO)~]~-M = Cr, Mo, or W -53 BioHiipy-4 [(BioHllpy)Co(py)X21- X = C1 or Br 12 35 B 9Hi 0s-4 [(Et3P)zPt(H)B 9Hi OSI 13 36 n-B1 8HZ02- 4 [(~-BI~H~O)MLZ]M = Ni, Pd, or Rh 9 31 4 [(n-B18Hz0)Co(C0)3]--31 i-B18H202-4 [(i-BlsH20)M(diphos)] M = Ni, Pd, or Pt 31 bis-dihapto B1 oH12'- 4 [(EtzO)zCd(B10H12)]~(also Zn) 19 pentahap to Bi oHioCOH- 4 [(Bi oHioCOH)M(C0)41 M = Cr, Mo, or W 53 Bi oHi oC02- 4 [(BioHioCO)M~O(C0)~12~ M = Cr, Mo, or W 53 B9H13-5 [(B 9H13)Mn(C0)3 1-55 B gHizTHF 5 [(B9HdHF)Mn(CO)d 55 Bi oHioPPh- 5 [(B10HioPPh)Mn(C0)31-54 Bi oHioS2-6 [(B~OH~OS)~CO]-,[(BI OHI oS)2FeI2- 32 6 [(Bi oHi oS)Pt(EtsP)z 1 32 6 [(BI OHI 0S)Re(C0)31- 32 hexahapto no examples (yet !) heptahapto no examples (yet !) octahapto (i.e.tetrakis-dihapto) 41 electrons into the tetrahedral iron(m) atom as shown in Figure 7. By contrast the monohapto-ligand 'B5H8 donates only one electron to the B-M bond in 2-Me3MB5Hs and 2-[M(C0)5]B5Hs, thereby implying that a normal a-two- electron two-centre terminal B-H bondinB5Hg hasbeen replaced bya B-MMe3 or %--M(C0)5 bond (see Figure 29). The B5Hs unit can act as a dihapto- ('n-bonded') ligand in which the borane moiety donates both electrons to form Metalloboranes and Metal-Boron Bonding a three-centre two-electron bond, as in p-RsM-BsH8 and p-MeaSi-B~H7Br (see Figure 22). Similar bonding is shown by the B8HlO complexes illustrated in Figures 8 and 28. Dihapticity can also be generated by a bis-monohapto chelating Iigand, and a possible example of this is the [(BloHr2)AlH2]- ion shown in Figure 1.trihapto-Borane ligands can donate 2, 3, or 4 electrons to a metal atom. An example of a two-electron trihapto-ligand is B4H8 in (~-C~H~)CO(B~H~), the connecting atoms being the apex boron and two bridging hydrogen atoms [Figure 16(a)]. This complex can also exist as a structural isomer [Figure 16(b)] in which the Bas group acts as a two-electron tetrahapto-ligand, with the four basal boron atoms now being equally bonded to the cobalt atom. Three-electron trihapto-ligands are exemplified by the ‘mono-plus-dihapto’carbaboraneC2B4Hs (connected to the metal atom by three boron atoms) and the v-borallyl complex (R~P)~P~(?T-B~H~).If the bonding in this latter complex is considered to be a-plus*-bonding from BsH7-, as illustrated in Figure 18, then it is a three-electron donor, involving PtI; alternatively, as suggested by Figure 17, it may be considered as a true r-borallyl four-electron donor involving BsH~~- and PtII.This is an ambiguity which finds many counterparts in organometallic chemistry. Figures 14 and 15 illustrate borane clusters acting as four-electron trihapto- ligands, the B9HrlL2- and B8H122- groups acting as bidentate chelates, with pairs of electrons from B(2)B(5) and B(2)B(7) completing the square-planar bonding environment about the platinum atom in each case. tetrahapto-Borane ligands comprise the most numerous class to date. In many of the examples listed in the Table the borane acts as a bidentate chelate, donating four electrons to the metal.The bonding has already been fully discussed in connection with tetrahedral [Zn(Bl0Hl2)2]~- (Figures 4 and 5) and square-planar wi(B10H12)2l2- (Figure 11). Further examples are illustrated in Figures 6, 9, 12, 13, 19, and 21. In addition, BloH12~- can generate tetrahapticity by acting as a bis-dihapto-ligand in the dimeric complex [(Et20)2Cd(B10H12)]2, as discussed in connection with Figure 2. Three other modes of tetrahapto interaction are also known. The two-electron donation in (~-C~H~)CO(B~H~) has already been alluded to on p. 252. A similar geometrical arrangement occurs in the four- electron donor BsH~~-, which is an analogue of cyclobutadiene (see Figure 30). Finally, B9H13- can act as tetrahapto two-electron ligand by bonding to cobalt via three boron atoms and one bridging hydrogen atom as illustrated in Figure 16(f) [though by analogy with the isostructural cluster in decaborane(l4) itself the cobalt atom occupying the B(5) position could equally be thought of as being connected to B(1)B(2)B(6)B(1O)Hy thus making the ligand formally pentahapto].pentahapto-Borane ligands are known which donate 4, 5, or 6 electrons. It is interesting to note, as indicated in the Table, that all such ligands that have so far been prepared have at least one heteroatom involved in the bonding to the metal acceptor, and that when this heteroatom is C, P, or S,the formal number of electrons donated by the ligand is 4, 5, and 6. Illustrations are in Figures 25, 26, and 10, respectively.Bridge-bonding via hydrogen also occurs, as exemplified by the Bs-manganese complex shownin Figure 27, Greenwood and Ward No examples of hexahapto- and heptahapto-borane ligands have yet been synthesized though, of course, in the carbaborane-metal complexes a connexity of higher than five occasionally occurs, particularly if the carbaborane skeleton is joined to more than one metal atom. One example of an octahapto-ligand is known, namely CuzBloHlo; this high connexity is achieved by having four pairs of boron atoms bonding to four different copper atoms. The details of the bonding in this compound have not been fully elucidated, though it is certain that the picture given in Figure 20 of electron-pair donation into three-centre BCuB orbitals is an oversimplification, particularly in view of the known electron delocalization in the parent closo-borane anion B1oH1o2-.In conclusion, it can be stated that the new and rapidly expanding field of metalloboranes has now reached a stage where preliminary systematization is possible. A wide variety of structural and bonding types is apparent and these give promise of still further rapid advances in the near future.

ISSN:0306-0012    

DOI:10.1039/CS9740300231

出版商:RSC    

年代:1974

数据来源: RSC