摘要:

Quarterly Reviews No 4 Vol21 1967 Page Organic Reactions involving ElectrophiIic Oxygen By J. B. Lee and B. C. Uff 429 Chemical Applications of Oxygen-17 Nuclear and Electron Spin Resonance By Brian L. Silver and Zeev Luz 458 The Elementary Particles By B. H. Bransden 474 The Carbanion Mechanism of Olefin-forming Elimination By D. J. McLennan 490 Electronic Properties of Binary Compounds of the First-row Transition Metals By A. T. Howe and P. J. Fensham 507 The Chemistry of the C,,-Diterpene Auraloids By S. W. Pelletier 525 The Chemical Society London Quarterly Reviews contains articles by recognised authorities on selected topics from general physical inorganic and organic chemistry. The Journal and Annual Reports interest primarily the research worker Quarterly Reviews is designed for a wider range of readers. It is intended that each review article shall be of interest to chemists generally and not only to workers in the particular field being reviewed. The submission of reviews for publication is welcomed but intending authors are advised to write in the first place to the Editor The Chemical Society Burlington House Piccadilly London W. 1. Such preliminary communications should be accompanied by an outline of the ground to be covered (about two quarto pages) rather than by the completed manuscript. Price to non-fellows ;E4 10s. Od. per annum @ Copyright reserved by The Chemical Society 1967 Published by The Chemical Society Burlinglon House London. Printed in England by The Thanet Press Margate.

ISSN:0009-2681    

DOI:10.1039/QR96721FP007

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Organic Reactions involving Electrophilic Oxygen By J. B. Lee and B. C. Uff UNIVERSITY OF TECHNOLOGY LOUGHBOROUGH LEICESTERSHIRE 1 Introduction In many of the more familiar reactions of organic compounds in which an oxygen atom is directly involved at the reaction centre the oxygen atom behaves as a nucleophile. In such reactions as the solvolysis of alkyl halides epoxides esters and amides by hydroxylic solvents,l and in the protonation of epoxides aldehydes etc.,lP2 in cases of neighbouring-group participation by oxygenated groups? the oxygen atom clearly forms bonds to other atoms with a pair of its own electrons. In some other cases as for example in the breakdown of (postu- lated) chromate or vanadyl esters mechanisms may be drawn formally involving either nucleophilic (1) or electrophilic behaviour (2) (or even radical formation) (2) and considerable experimental work may be needed to establish the correct mechanism.At the other end of this ‘spectrum’ of behaviour are a number of reactions in which oxygen clearly behaves electrophilically the second molecule involved (the nucleophile) donating a pair of electrons to the (at least incipiently) electron-deficient oxygen. Thus the Baeyer-Villiger conversion of ketones into esters5 provides an early example of this type of behaviour and it is now appre- ciated that electrophilic behaviour by an oxygen atom occurs commonly in a wide range of reactions. One major group of reactions where such electrophilic behaviour is shown involves rearrangements of peroxy-compounds where a group (alkyl or aryl) migrates from carbon to oxygen (3) (X = OR).Apart from other heterolytic E.g. see C. A. Bunton ‘Nucleophilic Substitution at a Saturated Carbon Atom’ ed. E. D. Hughes Elsevier London 1963. E.g. J. N. Bronsted M. Kilpatrick and M. Kilpatrick J. Amer. Chem. SOC. 1929 51,428; H. J. Lichtenstein and G. H. Twigg Trans. Faraday Soc. 1948,44,905. E.g. S. Winstein and R. Buckles J. Amer. Chem. Soc. 1942 64 2780. E.g. J. S. Littler and W. A. Waters J . Chem. SOC. 1959 4046; W. A. Waters Quart. Rev. 1958 12 277; Helv. Chirn. Acta 1962 45 2554; L. Kaplan J. Amer. Chem. SOC. 1955 77 5469. A. von Baeyer and V. Villiger Ber. 1899 32,3625; 1900,33 858. 429 Quarterly Reviews q - 7 f A R,c-0-x - R&b + x- (3) peroxide reactions the rearrangement of ozonides? the breakdown of certain oxye~ters,~ and certain reactions of amine oxides and sulphoxides for example are among reactions of this type.Consideration of the rearrangements occurring in peroxy-compounds has drawn attention* to features which these reactions and certain other rearrange- ments have in common. Compare for example the simplified equation* (4) for the Beckmann reaction9 with the generalised equation* (3) whilst these are (4) essentially intramolecular reactions there are certain unifying aspects mechanis- tically in a series of both inter- and intra-moIecular reactions which we hope to underline in this Review which revolve essentially around the generation of an electron-poor oxygen atom in the presence of some suitable nucleophilic group. These reactions often take place in such a way as to avoid the generation of an atom (usually 0 C N) having only six outer valency electrons.It is not always possible to ascertain whether such an electron-deficient atom is ever formed completely or whether concerted migration prevents its formation. In some cases intermediate discrete ionic species are formed but this does not occur invariably. In intermolecular reactions by the term ‘electrophilic oxygen’ we imply a process which may be generalised as (5) where Y is a group capable of forming a stable ion (e.g. acyl) and Z usually is hydrogen. Loss of a proton from (A) gives the stable neutral product the conjugate base of (A). Before discussing particular examples it is worth noting the following charac- teristics one might expect most of these reactions to exhibit.1° (a) The reaction will be of the second order; first order with respect to X first order with respect * The representation of charged intermediates in these equations is not to be taken to imply complete ion formation necessarily.J. E. Leffler Chem. Rev. 1949 45 385. M. Anbar I. Dostrovsky D. Samuel and A. D. Yoffe J. Chem. SOC. 1954 3603. R. Criegee Annalen 1948,560 127; G. Robertson and W. A. Waters J. Chem. Soc. 1948 1574; see also W. A. Waters ‘Mechanisms of Oxidation of Organic Compounds’ Methuen London 1964 ch. 3. @ W. Z . Heldt and L. G. Donaruma Org. Reactions 1960,11 ch. 1. lo J. 0. Edwards ‘Peroxide Reaction Mechanisms’ Interscience Landon 1962 p. 69. 430 Lee and Uf to YOZ. (b) Increased electron availability at the reacting centre of X will accelerate rate of reaction; decreased electron availability will decelerate it.(c) Increased electron attraction by Y will accelerate the rate of reaction and vice-versa. ( d ) Where the atom in the radical Y next to the oxygen is capable of protonation general or specific acid catalysis may be observed. (e) A salt effect may be observed. cf) Increased solvent polarity may increase the reaction rate. ( g ) The entropy of activation should be negative if an activated complex having a definite orientation is involved. (h) Energies of activation will be low (e.g. 12-1 6 kcal./mole). High values would suggest homolytic bond cleavage leading to a radical-chain or other free-radical mechanism.ll The precise mechanism of the reaction will determine to what extent each characteristic is found. In this Review reactions are considered under three headings ; intermolecular intramolecular and other reactions.2 Intermolecular Reactions Involving Electrophilic Oxygen These reactions can be further classified by the nature of the nucleophile. A. Olefins and Related Systems.-(i) The Prileschajew reaction.12 Peroxy-acids react with olefinic double bonds to give epoxides.13 The reaction involves an electrophilic attack by the peracid on the C=C bond. This is borne out by the observation that substituents on the bond which enhance its nucleophilic character (e.g. alkyl) increase the reaction rate whereas electron-withdrawing groups (e.g. -CO.X) considerably decrease reaction rate.14 Further the stronger the peracid the greater the rate; tduoroperacetic acid is particularly effective.16 The reaction which is of the second order (first order with respect to olefin and to peracid) is most easily effected in a non-ionising solvent such as benzene and is not subject to a salt effect.ls This suggests a non-ionic transition state HOCOR of the type (I)17 where proton transfer occurs by a concerted intramolecular process.An alternative polar stepwise addition involving charge-separated intermediates (II III) seems unlikely. l1 K. M. Ibne-Rasa and J. 0. Edwards J. Amer. Chem. SOC. 1962 84,767. la N. Prileschajew Ber. 1909 42 4811; J. Russ. Phys. Chem. SOC. 1910 42 1387; 1911 43 609; 1912,44,613. l4 D. Swern J. Amer. Chem. Soc. 1947 69 1692. l6 W. D. Emmons A. S. Pagano and J. P. Freeman J. Amer. Chem. SOC. 1954,76 3472; W. D. Emmons and A. S. Pagano ibid. 1955 77 89. l6 B. M. Lynch and K. H. Pausacker J. Chem. SOC. 1955 1525.l7 P. 0. Bartlett Rec. Chem. Prog. 1950 11 47. (a) D. Swern Chem. Rev. 1949,45 1 ; (b) D. Swern Org. Reactions 1953 7 378. 43 1 Quarterly Re views i The stereochemistry of olefin epoxidation has recently been reviewed.18 An interesting alternative mechanism1g involving a 1,3-dipolar addition as the rate determining step has recently been proposed by Kwart et aZ.lga The mechanism \ / n has received some criticism.lgb Thus for example dipolar addition of nitrones is strongly accelerated by electron withdrawing groups attached to the double bond whereas such groups drastically inhibit epoxidation.lgc Further sub- stituent effects on the rate of epoxidation e.g. the additive effect of alkyl groups on either side of the double bond suggest a fairly symmetrical transition state.In contrast the dipolar addition mechanism would predict a heavily unsym- metrical transition state.lgc (ii) Acetylenes. Treatment of the acetylenic triple bond with peroxyacids gives a mixture of products.20,21 Addition of peracid by a concerted mechanism (similar to the olefin case) has been postulated.20,21 This produces the oxiren (IV) which can be further converted into the 1,3-dioxobicyclo [l ,l,O]butane (V) both intermediates undergoing further reaction. 0 R-C=C-R - [ R-C-C-R ;”; ] - k-C53-R] 0 I- Products L Products (iii) ap- Unsaturated ketones with hydrogen peroxide in alkali. ai3-Unsaturated ketones form epoxides on treatment with alkaline hydrogen peroxide.22 The mechanism23 does not involve nucleophilic attack on the peroxide. The carbonyl group confers electrophilic character on the C=C bond which is attacked by the -0OH anion (VI).l8 H. B. Henbest in Chem. SOC. Special Publ. No. 19 1965 p. 83; H. C. Stevens and A. J. Kaman J. Amer. Chem. SOC. 1965 87 734. l9 (a) H. Kwart and D. M. Hoffman J. Org. Chem. 1966,31,419; H. Kwart P. S. Starcher and S. W. Tinsley Chem. Comm. 1967 335; (6) K. D. Bingham G. D. Meakins and G. H. Whitham Chem. Comm. 1966 445; (c) G. H. Whitham private communication. 2o R. N. McDonald and P. A. Schwab J. Amer. Chem. SOC. 1964 86,4866. 21 J. K. Stille and D. D. Whitehurst J. Amer. Chem. SOC. 1964 86 4871. 22 Reviewed by C. H. Hassall Org. Reactions 1957 9 73. 23 C. A. Bunton and G. 0. Minkoff J. Chem. SOC. 1949 665. 432 The peracid epoxidation of olefins is known to proceed stereospecifically; the configuration of the resulting epoxide is usually identical with that of the original 0 1 e f i n .~ ~ ~ ~ ~ ~ ~ - ~ ~ Under similar reaction conditions ap-unsaturated aldehydes2' and alkylidene- malonic esters28 can be epoxidised. (iv) Nitriles and ap-unsaturated nitriles. Nitriles are hydrolysed to amides by alkaline hydrogen peroxide.29 The peroxycarboximidic acid MI) formed initially decomposes via a concerted reaction which involves a further molecule of hydrogen peroxide.3O In the case of an @-unsaturated nitrile self-epoxidation occurs31 (VlII) to give the ep~xy-amide~l,~~ (IX). Addition of a competing olefin cyclohexene halves the yield of glycidamide (IX) with concomitant production of acrylamide (X) which is also obtained but in very low yield in the absence of cy~lohexene.~~ Good use of the intermediate (VIII) has been made in enabling added olefins to be oxidised under mildly alkaline conditions.33,34 Keto-olefins have thus been epoxidised without formation of Baeyer-Villiger products since the latter reaction is usually acid-~atalysed.~~ (v) a@ Unsaturated ketones with organic peracids.Organic peracids react with 24 S. Winstein and R. B. Henderson R. C. .Elderfeld 'Heterocyclic Compounds' vol. 1 John Wiley New York 1950 p. 1. 26 H. 0. House and R. S. Ro J. Amer. Chem. SOC. 1958 80,2428. 26 H. E. Zimmerman L. Singer and B. S. Thyagarajan J. Amer. Chem. SOC. 1959 81 108. 27 G. B. Payne J. Org. Chem. 1960 25 275; J. Amer. Chem. SOC. 1959 81 4901. 2Q B. Radziszewski Ber. 1884 17 1289. 30 K. B. Wiberg J. Amer. Chem. SOC. 1953 75 3961 ; 1955 77 2519. 31 G. B.Payne and P. H. Williams J. Org. Chem. 1961 26 651. sa G. B. Payne P. H. Deming and P. H. Williams J. Org. Chem. 1961,26,659; G. B. Payne ibid. 1961 26 663 668; Tetrahedron 1962 18 763; Y. Ogata and Y. Sawaki ibid. 1964 20 2065. 33 G. B. Payne P. H. Deming and P. H. Williams J. Org. Chem. 1961 26 659. 34 Y. Ogata and Y . Sawaki Tetrahedron 1964,20,2065. 35 C. H. Hassall Org. Reactions 1957 9 74. G. B. Payne J. Org. Chem. 1959,24,2048. 433 Quarter Zy Reviews ab-unsaturated ketones to give enol esters (XlII) or lactones. The mechanism3s is a specific example of the Baeyer-Villiger oxidation. ? '+ F f+c+a-O-C-Me - R-CH-qk-C-Me 0 &Ill> (XI 0 The reaction is c ~ n s i d e r e d ~ ~ ~ ~ to be acid-catalysed. The route via the carbonium-ion intermediate (XIII) is additionally favoured by a stabilising aryl group for R.Benzylideneacetophenone for instance gives exclusive formation of enol ester on oxidation with peracetic acid.39 PhCH=CH-COMe - Ph CH=CH-O-COMe Similarly the benzylidene cycloallcanones give the lactones (Mv) which can be further oxidised to (Xv).40 O\ Ph-CH=$-(p _t Ph-u-I=y-O-yO - P h ~ H - ~ - O - ~ O [CH;;I,CH2 (XIV) P%-cH2 (xv) P-41,CH2 Where acid catalysts are absent and benzylidene-ketones are not involved the use of organic peracids usually,22 but not i n ~ a r i a b l y 3 ~ ~ ~ ~ ~ leads to enol esters. 5,6-Dibromo-3-hydroxy-16-pregnen-2O-one acetate (XVI) gives with per- benzoic acid not enol ester but the 16,17-epoxy-derivative (XVII);41 since addi- & __c Ace@ AcO Br (XVI) Br (XVII) tion to carbonyl is often the rate-determining stage this suggests that attack at the carbonyl is more difficult in this case than at the 16,17-double bond.Parallel s' E. Wenkert and M. Rubin Nature 1952 170 708. P. Wieland and K. Miescher Helv. Chim. Acta 1949 32 1768. s8 L. H. Sarett J. Amer. Chem. Soc. 1947 69 2899. s9 J. Boeseken and A. L. Soesman Rec. Trav. chim. 1933 52,874. 40 H. M. Walton J. Org. Chem. 1957 22 1161; C. R. Zanesco Helv. Chim. Acta 1966 49 1002. 4a G. B. Payne and P. H. Williams J. Org. Chem. 1959,24,284. P. L. Julian E. W. Meyer and I. Rydon J. Amer. Chem. Soc. 1950 72 367. 434 Lee and Uf difficulty in carbonyl addition in saturated analogues has been noted.43 However with strong peracids carbonyl addition occurs,44 thus providing a useful route for conversion of 17-acyl-steroids into the 17-hydroxy-analogues.OCOMe (7 W{fi - {& B. Peracid Oxidation of C=N and N-N Bonds.-(i) Schif bases. Peracid oxidation of the C = N bond has been demonstrated with Schiff bases-* and benzoylimine~~~ (the nitrogen analogues of ap-unsaturated ketones). Schiff-base oxidation yields oxazirans (XVIII). E m r n o n ~ ~ ~ ~ has demonstrated clearly that the isomeric nitrone (XIX) is not formed. Irradiation converts (XlX) into (XVIII) ;48,50 thermal isomerisation regenerates nitrone from ~ x a z i r a n . ~ ~ ~ ~ The available evidence suggests initial nucleophilic attack by peroxyacid on. the positive carbon followed by nucleophilic displacement of an acyl ion by the electron-rich nitrogen (XX). RYR2 N R3’ This is supported by in certain condition^.^^ A similar mechanism the isolation of the intermediate hydroperoxide (XXI) ~ _ t oxazirane O.OH (XXI) cthv can be suggested for the related oxidations of C- and N-benzoylimines [Section B.(ii)]; contrast the mechanism with the two other possibilities (6) and (7) both involving electrophizic attack by peracid initially. p3 J. L. Mateos J. Org. Chem. 1959,24,2034; R. Cetina and J. L. Mateos ibid. 1960,25,704. 44 R. E. Marker J. Biol. Chem. 1940 62 650. 45 H. Krimm Chem. Ber. 1958,91 1057. 46 W. D. Emmons J. Amer. Chem. SOC. (a) 1956 78 6208; (b) 1957 79 5739. 47 L. Horner and E. Jiirgens Chem. Ber. 1957,90,2184. 48 E. Schmitz ‘Advances in Heterocyclic Chemistry’ ed. A. R. Katritsky Academic Press New York 1963 vol. 2 p. 83. 49 A. Padwa J. Amer. Chem. SOC. 1965 87,4365. 50 J. S. Splitter and M. Calvin J. Org. Chern. 1958 23 651.51 E. Hoft and A. Rieche Angew. Chem. Internat. Edn. 1965 4 525. 435 Quarterly Reviews (7) The analogous mechanism (6) to the Prileschajew reaction would in any case seem less favourable than direct attack by the more available nitrogen electron pair { cf. the peracid oxidation of azo-compounds and phenylhydrazones [Section B. (iii)] } (7) but the greater thermodynamic stability of n i t r o n e ~ ~ ~ p ~ ~ eliminates this mechanism as nitrones are not isolated. (ii) C- and N-Benzoylimines. The action of organic peracids on C-benzoy- limines produces oxa~irans.~~ Monocyclohexyliminobenzil (XXII) yields with m-chloroperbenzoic acid 3-benzoyl-2-cyclohexyl-3-phenyl-oxaziran (XXIII) which rearranges to NN-dibenzoylcyclohexylamine (XXIV) on heating. /O\ Ph (XXII) Ph (XXIII) WV.l PhCO*FN-C,H, -+- Ph C0.~-N-C6H, - (PhCO)~-C,H, Peracid oxidation of the related N-benzoylimines (XXV) proceeds by a Baeyer-Villiger type of rearrangement of the peracid adduct (XXVI) first formed.=NCOPh "'0 0' (XXV) C=NCOPh - The product (XXVII) is usually hydrolysed in the work-up. The observed migratory aptitudes49 in the rearrangement step (XXVI) -f (XXVII) are con- sistent with migration to an electron-deficient atom.52 An alternative rearrange- ment of an oxaziran intermediate% seems less plausible unless one postulates relief of steric strain as the driving force. The case differs from Schiff-base oxidation via intermediate where the Schiff-base nitrogen is considerably more nucleophilic. (iii) Phenylhydrazones and azo-compounds. Phenylhydra~ones~-~~ and azo- 52 P.A. S. Smith 'Molecular Rearrangements' ed. P. de Mayo John Wiley New York 1963 part 1 p. 585. 53 Footnote 29 in ref. 49. 54 K. H. Pausaker J. Chem. SUC. 1950,3478; B. M. Lynch and K. H. Pausaker ibid. 1953 2517; 1954 3340. 55 B. M. Lynch and K. H. Pausaker J. Chem. SOC. 1954 1131. 56 B. Witkop and H. M. Kissman J. Amer. Chem. SOC. 1953,75 1975. 57 J. N. Brough B. Lythgoe and P. Waterhouse J. Chem. SOC. 1954 4069. 436 Lee and Uff compounds5* both give azoxy-compounds on oxidation with peracids although the products differ ~tructurally.~~ Ar ’ CH==N- NHA? 0 7 / (XXIX) (XXX) (XXVI I I) A? CH,-N=&A~~ or A r‘cH2-N=NAr2 A ~ H 2- N= N-A? In the case of the aromatic aldehyde phenylhydrazones (XXVIII) the dispute concerning alternative structures (XXrX) and (XXX)54-57959 for the products has resolved in favour of the cis-form of (XXIX).57 The reaction has been extended to the aromatic aldehyde alkylhydrazones.gO The isomeric azo-compounds give both (XXIX) and (XXX) the former pre- dominatingF7 both however in the trans configuration.The hydrazone-peracid reaction is of the second order and has identical entropy of activation in polar and non-polar solvents.55 Electrophilic molecular (non-ionic) attack by the peracid followed by rapid rearrangement of (XXXI) seems likely. N:-f &C-Me ct R C ~ N ~ N - RCH=N\ HO- R/ ‘H q/ R’ %H (xxx I) Similar attack by the nitrogen lone pair (contrast then-bond attack of olefins) can occur with azo-compounds. C. Amines.-(i) Oxidation with peracids. Tertiary amines yield amine oxides with organic peracids.61 Slower reaction occurs with hydrogen p e r o ~ i d e .~ ~ ~ Caro’s acid has been ~ ~ e d . ~ ~ ~ ~ R3N:~@jOS0,H - R3N+0 + H,S04 Ogata and Tabushis4 observed second-order rate constants with Caro’s acid and p-substituted NN-dimethylaniline; substituent effects were consistent with the amine’s behaving nucleophilically. R o d 5 had earlier found second-order kinetics in the hydrogen peroxide-triethylamine r e a ~ t i o n . ~ ~ ~ ~ In contrast primary and secondary amines give mixed products on oxidation; initially a 58 D. Swern Chem. Rev. 1949 45 1. 59 M. Bergmann R. Ulpto and C. Witte Ber. 1923 56 679. 6o B. T. Gillis and K. F. Schimmel J. Org. Chem. 1962 27,413. 61 D. Swern Chem. Rev. 1949,454 6. 62 W. R. Dunstan and E. Goulding J. Chem. SOC. 1899,75 1004. 63 A. C. Cope and E. R. Trumbull Org. Reactions 1960 11,317.64 Y. Ogata and I. Tabushi Bull. Chem. SOC. Japan 1958,31,969. 65 S. D. Ross J. Amer. Chem. Soc. 1946 68 1484. 66 A. A. Oswald and D. L. Quertin J. Org. Chem. 1963,28 651. 67 H. Wieland Ber. 1921 54 2353. 437 Quarterly Reviews hydroxylamine forms ; nitrones (from secondary amines) or nitroso- and nitro- compounds (from primary amines) are later products. Condensation reactions give rise to b y - p r o d ~ c t s . ~ ~ ~ ~ ~ ~ Detailed s t ~ d i e s ~ ~ ~ ~ of the oxidation of anilines with peracetic acid show slow formation of phenylhydroxylamines preceding compounds. 71 Kinetic results are consist en t with by amine on the peroxide oxygen. Overall oxidation proceeds according to (8). rapid conversion into nitroso- bimolecular nucleophilic attack - RNO[-RNOJ (8) Edwards et al.72 showed that the oxidation of nitroso- to nitro-benzene gave similar kinetic results indicating a similar nucleophilic attack by the nitrogen on peroxy-oxygen.This is in conflict with Bunton’s suggestion73 that the reactants behave in a reverse manner i.e. that the peroxyacid behaves nucleophilically. Direct oxidation of primary arylamines to nitro-compounds is most easily achieved by use of trifluoroperacetic acid.74 (ii) Oxidation with acyl peroxides. The reaction of tertiary amines with benzoyl peroxide initially gives a quaternary derivative -1) by nucleophilic attack on the per~xide.~~-~l Evidence includes substituent effects ;75 electron-donating f R3N + (PhCO)202 - [lNOCOPh]PhCOp - products yoocl 0 groups accelerate electron-withdrawing retard reaction when substituted in the a ~ n i n e ~ ~ but act conversely when substituted in the peroxide.76 It is sug- gestedE2 that decomposition of (XXXII) is a radical process.Dimethylaniline N-oxide also gives an intermediate similar to =I) with acetic anhydride. 83 Secondary amines with benzoyl peroxide give 0-benzoylhydroxylarnines or 68 E. Baumberger and T. Scheutz Bet. 1901,34,2262. 69 J. D ’ h s and A. Kneip Ber. 1915 48 1136. 71 K. M. Ibne-Rasa and J. 0. Edwards J. Amer. Chem. SOC. 1962 04 763. 72 K. M. Ibne-Rasa C. G. Lauro and J. 0. Edwards J. Amer. Chem. Sac. 1963 85 1165. 7a C. A. Bunton ‘Peroxide Reaction Mechanisms’ ed. J. 0. Edwards Interscience New York 1962 p. 21. 74 W. D. Emmons J. Amer. Chem. SOC. 1954,76 3470; 1957,79 5528. 76 L. Horner and K. Scherf Annalen 1951,573 35.76 L. Horner and W. Kirmse Annalen 1955,597,48. 77 L. Horner and H. Junkermann Annalen 1955 591 53. 78 M. Imoto and S. Choe J. Polymer Sci. 1955 15 485. 78 C. Walling and N. Indictor J. Amer. Chem. SOC. 1958 80 5814. 82 For a summary see C. Walling ‘Free radicals in solution’ Wiley New York 1957 p. 590. 83 V. Boekelheide and D. C. Harrington Chem. and Ind. 1955 1423. I. P. Gragerov and A. F. Levit Zhur. obshchei Khim 1960 30 372. D. Buckley S. Dunstan and H. B. Henbest J Chem. Suc. 1957 4901. W. B. Geiger J. Org. Chem. 1958 23 298. 438 Lee and U? their rearrangement products. There is evidence for participation of free radicals in some instance^^,^^ but a non-radical mechanism has also been s u g g e ~ t e d . ~ ~ ~ ~ Using benzoyl peroxide labelled with oxygen-1 8 in the carbonyl positions (XXXIV) Denney and Denneys6 showed that dibenzylamine (XXXIII) and diphenylamine react by nucleophilic displacement on an link to give o.oxygen of the peroxide t PhC02H The suggestione2sBs that the transition state (XXXVI) decomposes via ions with subsequent proton transfer is not supported by the observatione7 that solvent polarity has little effect on reaction rate. More likely is a concerted proton transfer through a cyclic intermediate (XXXVII). D. Peroxidation of Sulphur Compounds.-The peroxidation of sulphidess8 involves two steps. A sulphoxide is first formed and further converted at a slower rate into a sulphone The mechanism has been shown clearly to involve electrophilic peroxide attack on sulphur. Thus when bis-(p-chlorobenzyl) sulphide is oxidised by substituted benzoic acidss9 the reaction is accelerated by electronattracting substituents and vice versa with a Hammett reaction constant of p = + 1-05 The rate which is of the first order in peroxide and sulphide (or sulphone) is 84 Kh.S. Bagdasaryan and R. I. Milyutinskaya Zhur. fiz. Khim. 1953,27,420. 86 D. A. Chaltykyan E. N. Anatasyan N. M. Beileryan and G. A. Marmaryan Russ. J. Phys. Chem. 1959,33,36. 86 D. B. Denney and D. Z. Denney J. Amer. Chem. SOC. 1960 82 1389. *'Ya. K. Syrkin and I. I. Moiseev Russ. Chem. Revs. 1960 29 193. 88 D. Barnard L. Bateman and J. I. Cunneen 'Organic Sulphur Compounds' ed. N. Kharasch Pergamon London 1961 p. 229. 89 C. G. Overburger and R. W. Cummins J. Amer. Chem. Soc. 1953,75,4250. 439 Quarterly Reviews not subject to a salt effect and is faster in toluene than isopropyl alcohol s~ggesting~~ an intramolecular hydrogen-transfer (XXXVIII-XI,).The increased rate in toluene which accompanies lower energy and entropy of activation could however be due to the strong hydrogen bonding possible in isopropyl alcohol hindering proton transfer. Peracid oxidation of sulphoxide has similarly been shown to involve electro- philic attack by the p e r a ~ i d . ~ ~ ~ ~ When hydrogen p e r o ~ i d e ~ ~ ~ ~ or hydro peroxide^^^-^^ are the oxidants the medium can have marked effects on the kinetics and solvent may participate in the transition state Addition of acetic acid to the ethanethiol-t-butyl hydroperoxide reaction system produces a sharp change in rate when amounts added are large the acid thus behaving more like a solvent.Acid catalysis seems to be specific rather than general,96 which is also the case for the oxidation by hydrogen peroxide of other system^.^^-^^ General acid catalysis has been observed in the oxidation of sulphite ion by hydrogen peroxide.100 E. Peroxidation of Phosphorus Compounds.-(i) With diaroyl and diacyl peroxide Laible et al.lol report that dibenzoyl peroxide reacts with trialkyl phosphites to give benzoic anhydride and phosphate ester. The reaction with phosphines has been studied in greater detail and is probably ionic.lo2 (ii) With hydrogen peroxide and hydroperoxides. Hydroperoxides react readily with both tertiary phosphines and tertiary phosphites,lo3 and since analogous products are obtained it seems reasonable to assume that analogous mechanisms are involved. Primary phosphines also react vigorously to give initially primary G.Modena Gazzetta 1959 89 843; 1960 90 3 11. s1 G. Kreeze W. Schramm and G. Gleve Chem. Ber. 1961,94,2060. 92 C. G. Overburger and R. W. Cummins J. Amer. Chem. Soc. 1953,75,4783. 93 S. D. Ross J Amer. Chem. Soc. 1946 68 1484. Q4 L. Bateman and K. R. Hargrave Proc. Roy. SOC. 1954 A 224,389. L. Bateman and K. R. Hargrave Proc. Roy. Soc. 1954 A 224 399. Q6 J. 0. Edwards and D. H. Fortnum J. Org. Chem. 1962 27 407. s7 I. R. Wilson and G. M. Harris J. Amer. Chem. SOC. 1960 82 4515; 1961 83 286. Q* H. A. Liebhafsky and A. Mohammed J. Amer. Chem. SOC. 1933,553977; 1934,56,1680. sQ E. Abel Monatsh. 1907 28 1239; K. Sandred and J. B. Hotte Chem. Abs. 1939,33,4856. loo P. M. Mader J. Amer. Chem. SOC. 1958 80 2634. lol R. C. Laible R.M. Esteve and J. D. Morgerum J. Appl. Polymer Sci. 1959 1 376; Chem. Abs. 1960 54 10827. loZL. Horner and W. J. Jurgeleit Ann. Chem. 1955 591 138; M. A. Greenbaum D. B. Denney and A. K. Hoffman J. Amer. Chem. SOC. 1956,742563; D. B. Denney and M. A. Greenbaum ibid. 1957 79 979. lo3 (a) M. A. Greenbaum D. B. Denney and B. Goldstein J. Amer. Chem. SOC. 1960 82 1396; (b) C. Walling and R. Rabinowitz ibid. 1957 79 5326; (c) I. S. Bengelsdorf Ph.D. Thesis Univ. of Chicago 195 1. 440 Lee and U . phosphine oxideslo4 which rapidly decompose at room temperature. Bengels- dorf's original suggestionslo5 were confirmed by later workerslo6 who suggested that an intermediate quasi-phosphonium compound was produced (XLI). RO,H + P(OEt)3 - [ROkOEt)3*0H] - ROH + OP(OEt) WI) However in an examinationlo7 of the reaction of triphenylphosphine with trans-9-decalyl hydroperoxide retention of configuration and a lack of exchange of labelled oxygen with either product when labelled water was added was found.In formation and breakdown of an intermediate of type (XLL) one would most reasonably suggest either inversion or racemisation of the group R as likely to occur and lack of exchange of labelled oxygen would require (XLI) to decompose much more quickly than equilibration of OH- and water. A simpler explanation is to assume a mechanism in common with other heterolytic peroxide oxidations. The reaction should show similar characteristics to other reactions of this type e.g. protonation of 0 should accelerate reaction. However although the reaction has been applied synthetically,lo8 no kinetic study has been made.Denney et al.lo9 examined the oxidation of optically active phosphines with various oxidising agents and assumed that t-butyl hydroperoxide reacted with- out inversion but no direct evidence is available on this point. (iii) With dialkyl peroxides. Walling and Rabin~witzlO~~ considered that the reaction of di-t-butyl peroxide with phosphites was a radical process (this may be a steric effect; see footnote 11 in ref. 110). The interesting observationlll that peroxides but not hydroperoxides reduce pyridine N-oxide in the presence of triethyl phosphite led Rees and Emerson to postulate a radical process. The reaction has certain analogies with the reduction of N-oxides by H202 in acetic acid,ll0 which would imply nucleophilic attack by the N-oxide upon peroxide.Io4 S. A. Buckler and M. Epstein Tetrahedron 1962 18 1221. Io5 I. S. Bengelsdorf Ph.D. Thesis Univ. of Chicago 1951. lo' D. B. Denney W. F. Goodyear and B. Goldstein J. Amer. Chem. SOC. 1960,82 1393. 108 T. Tanaka Yakuguku Zasshi 1960 80,439; (Chem. Abs. 1960,54 19470). lo9 D. B. Denney and J. W. Hanifin Tetrahedron Letters 1963 No. 30,2177; cf. D. B. Denney etal. J. Arner. Chem. SOC. 1960,82,1393; 1961,83 1726; 1962,84,4737; 1956,78,2563. 110 I. J. Pachter and M. C. Kloetzel J Amer. Chern. SOC. 1951 73 4958. D. B. Denney M. A. Greenbaum and B. Goldstein J. Amer. Chem. SOC. 1960 82 1396. T. R. Emerson and C. W. Rees Pruc. Chem. SOC. 1960,418. 441 QuarterZy Reviews These authors foundlll that an electron-donating (e.g. p-OMe) substituent accelerated reaction while an electron-attracting substituent slowed reaction which would fit a heterolytic mechanism but the evidence is too slight to permit a clear decision.(iv) With peresters. In a detailed study112 of the oxidation of tertiary phosphines with peresters isotopic and stereochemical results indicated clearly that a quinquecovalent phosphorus compound resulted from attack by phosphorus on peroxy-oxygen atoms. Since only slight rate enhancement occurs with increased solvent polarity (contrast the reaction of triphenylphosphine with sulphur,l13 for example) a relatively small separation of charge in the transition state is indicated. The entropy of activation observed114 is in line with this. Tertiary phosphites behave ~imilar1y.l~~ (v) Oxygen other than peroxy-oxygen. Apart from peroxide systems attack by phosphorus on oxygen is claimed to occur in a number of cases.For example triethyl phosphite with benzoquinones reacts mainly to give the products of 1,6-addition o(LIV).1169117 Consideration of other r e ~ u l t s ~ ~ ~ J ~ ~ suggests an alternative mechanism. 112 D. B. Denney W. F. Goodyear and B. Goldstein J. Amer. Chem. SOC. 1961,83 1726. 113 P. D. Bartlett and G. Meguerian J. Amer. Chem. SOC. 1956 78 3710. 114 A. A. Frost and R. G. Pearson ‘Kinetics and Mechanism’ Wiley New York 1953 ch. 7. 115 J. B. Lee unpublished observations. 116F. Ramirez E. H. Chen and S. Derschowitz J. Amer. Chem. SOC. 1959 81 4338; F. Ramirez and S. Derschowitz ibid. 1959,81 587; J. Org. Chem. 1957,22,856; 1958 23,778. 117 V. A. Kukhtin and K. M. Orekhova Proc. Acad. Sci. U.S.S.R.Chem. Sect. 1959 73; (Chem. Abs. 1961,55,1567); V. A. Kukhtin K. M. Orekhova and N. S. Garifyanov J. Gem Chem. U.S.S.R. 1961,31 1070. 118 T. Reetz U.S. Patent 2,935,518/1960; (Chem. Abs. 1960 54 19598); Abs. of papers of the 134th meeting A.C.S. Chicago Ill. 1958 p. 86-P. 119 R. G. Harvey E. G. de Sombre and E. V. Jensen Abs. of papers of the 135th meeting A.C.S. Boston Mass. 1959 p. 69-0; R. G. Harvey and E. V. Jensen Abs. of papers of the 144th meeting A.C.S. Los Angeles Calif. 1961 p. 44-0; G. Kamai and V. A. Kukhtin J. Gen. Chem. U.S.S.R. 1958 28 913; (Chem. Abs. 1958 52 17162); Proc. Akad. Sci. U.S.S.R. Chem. Sect. 1957 129; (Chem. A h . 1957,51 13742); Zhur. obshchei Khim. 1957 27,2376; (Chem. Abs. 1958,52,7127). 442 Lee and Uf CI CL CL CL Cl CL 0- 0- &(OR& C l \ C l CL /CL Q 0- (XLVII) Although electrophilic attack by phosphite esters upon episulphide sulphur is easy reaction with epoxides involves attack upon carbon.120 Similarly attack on nitrogen seems more likely than direct attack on oxygen in the deoxygenation (with phosphines and/or phosphites) of pyridine N-oxides,121 and some related reactions.1224 Attack upon carbonyl carbon is the usual mode of reaction of phosphites with aldehydes and anhydrides but in the unique case of phthalic anhydride in reaction with trialkyl phosphites attack upon oxygen to give phosphate and a carbene has been re~0rted.l~~ F.Peroxidation of Boron Compounds.-The recent extensive use of hydrobora- tion with ensuing oxidation to hydrate double bonds involves hydrogen peroxide oxidation of intermediate boranes probably involving attack on electron-deficient oxygen.Hydrogen peroxide oxidation of phenyl boronic acid was suggested to involve attack upon the boron by hydroperoxide followed by migration of the (phenyl) group in the hydroperoxide and expulsion of a hydroxyl ion with production of phenyl borate. C. B. Scott J. Org. Chem. 1957,71 578; N. P. Neureister and F. G. Bordwell J. Amer. Chem. SOC. 1959,81 578. lal F. Ramirez and A. Aguiar Abs. of papers of 134th A.C.S. meeting Chicago Ill. 1958 4 2 - ~ . M. Haman J. Pharm. SOC. Japan 1955,75 139; T. Emerson and C. Rees Proc. Chern. SOC. 1960 418. laa P. Bunyan and J. I. G. Cadogan Proc. Chem. SOC. 1962,78; T. Mukaiyama H. Nambu and M. Okamato J. Org. Chem. 1962,27,3651. la3 J. I. G. Cadogan and M. Cameron-Wood Proc. Chem. SOC.1962 361. lZ4T. Mukaiyama H. Nambu and M. Okamato J. Org. Chem. 1962 27 3651. laS P. Bunyan and J. I. G. Cadogan Proc. Chem. SOC. 1962 78; cf. T. T. Mukaiyama H. Nambu. and M. Okamato J. Org. Chem. 1962 27 3651; J. H. Boyer and S. Ellzey ibid. 1961,26,4684. ltrH. G. Kuivila J. Amer. Chem. SOC. 1954 76 870; 1955 77 4014 4830; 1957 79 5659. 443 Quarterly Reviews PhB(CH) + -0OH - Ph-B(OH) - PhOB(OH) + OH- [ k a - H l So far no detailed kinetic study of oxidation of trialkylboranes has been reported. The first significant observation was probably that of Johnson and Van Campen,12' who demonstrated the rapid quantitative low-temperature reaction of tri-n-butylborane with perbenzoic acid. Their observation on the similar action of alkaline hydrogen peroxide was later developed into an analy- tical technique.128 Other workers129 found the reaction proceeded rapidly at low temperature.The absence of free radicals is indicated130 and the lack of any rearrangement of the migrating group is well estab1ished,l3O the migrating group retaining its optical structure unchanged.l15J31 G. Peroxidation of Silicon Compounds.-Triorganosilyl peroxides (XLVII) and silanes (XLIX) have been prepared by a variety of method~,1~~-~ most involving nucleophilic attack by peroxy-oxygen. R3Si -0-0-Si R R3Si -O-O-CR (XLVII I) (XLIX) Breakdown of the peroxides (XLVIII) on distillation possibly involves a rearrangement (XLVIII - L). _ct R,S,i-OMe (L) OSi R The occurrence of rearrangement in the peroxysilanes (XLIX) is not well established and attempts to prepare triorganosilyl peroxyesters (LI) led directly to rearranged products but under similar conditions trialkylperoxysilanes did not rearrange.133 Rearrangement of the silyl perbenzoates seems to occur much more readily than for the analogous carbon compounds.115 Since silicon shows 12' J.R. Johnson and M. G. Van Campen J. Amer. Chem. SOC. 1938 60 121. 12* R. Belcher Mikrochim. Acta 1952 40 76. lZ9 H. C. Brown and B. C . Subba Rao J. Amer. Chem. SOC. 1956 78 5694; A. A. Carotti and P. F. Winternitz ibid. 1960 82 2430. 130 Inter al. H. C. Brown and G. Zweifel J. Amer. Chem. SOC. 1959 81,247; W. J. Wechter Chem. and hd. 1959 294; R. Koster G. Griansnow W. Larbig and P. Binger Annalen 1964 672 1; H. C. Brown N. R. Ayyanger and G. Zweifel J. Amer. Chem. SOC. 1964,86 397; D. K. Shumway and J. D. Barnhurst J. Org.Chem. 1964 29 2320; M. Nussin Y . Mamr and F. Sondheimer ibid. 1964 29 1120 1131. 131 H. C. Brown and G. Zweifel J. Amer. Chem. SOC. 1964 86,393; F. R. Jensen ibid. 1960 82 148 246. 132 W. Hahn and L. Metzinger Makromol. Chem. 1956 21 113. 133 E. Buncel and A. G. Davies Chem. and Ind. 1956 1052; J. Chem. SOC. 1958 1550. 134 R. A. Pike and L. A. Shaffer Chem. and Ind. 1957 1294; Chem. Abs. 1955,49,13290. 444 Lee and Uf a great ease of front-side attack in displacement mechanism of the type 0.1 -+ LII) should proceed rapidly. a synchronous Further evidence of this process is desirable. Since phenyl migrates in pre- ference to methyl and acid catalysis is shown one may assume that reaction occurs as in the carbon case.115 A comprehensive review of the chemistry of organometalloid and organo- metallic peroxides is recently a~ai1able.l~~~ 3 Intramolecular Rearrangements Involving Electrophilic Oxygen A.Peroxyesters.-In 1944 Criegee noted136 that acetate and benzoate esters of trans-9-decalyl hydroperoxide (LIII) rearranged on standing to give isomeric esters of type (LIV) ; subsequent hydrolysis gave 6-hydroxycyclodecanone (LV) (R= Me or Ph). 0-0-COR (LIV) OH (LV) 03 (LIII) Kinetic studies showed that the rate increased with more polar solvent and also with increased electron-withdrawing character of R (in LIII LIV).13' Criegee and K a ~ p e r l ~ ~ pointed out the analogy in this respect with carbonium- ion rearrangements of the Wagner-Meerwein pinacol-pinacolone and Beck- mann types. They suggested that heterolytic fission of the peroxide-bond occurred but that the benzoate ion was never free i.e.that an ion-pair mechanism was involved. Further study showed that the reaction was of the fist order,138,139 subject to slight acid catalysis and that the rates of rearrangement of para-substituted benzoate esters fitted the Hammett equation (p = +1-34).138 Furthermore in 135 (a) L. H. Sommer and 0. F. Bennett J. Amer. Chem. SOC. 1957,79 3295; 1959 81 251 ; (b) G. Sosnovsky and J. H. Brown Chem. Rev. 1966 66 529. 136 R. Criegee Ber. 1944 77 722. 13' R. Criegee and R. Kaspar Annalen 1948 560 127. 138 P. D. Bartlett and J. L. Kice J. Amer. Chem. SOC. 1953 75 5591. ls9 H. L. Goering and A. C. Olsen J. Amer. Chem. Suc. 1953 75 5853. 445 QuarterZy Reviews the presence of added para-substituted benzoate ions there is no interchange between these and the benzoate ions of the decalyl perbenz~ate.l~~J~~ A number of mechanisms have been pr~posedl~~-l~l and it seems likely that an intermediate of the type (LVIII) or (LIX) is i n ~ o l v e d .~ ~ ~ ~ ~ ~ The Criegee rearrangement has been further demonstrated14* in hydrindanyl peroxyesters (La leading to the cyclononanone system (LMI). B. Hydroperoxides.-Industrially important preparations of phenols have been based upon the rearrangement of certain hydroperoxides. Thus condensation of propene and benzene followed by oxidation and rearrangement yields phenol and acetone.la Probably the ‘prototype’ of this reaction was the p-cymene hydroperoxide rearrangement.l& In these cases the migration of an aryl group occurs to the exclusion of the alternative alkyl migration.One of the earliest examples14*Ja7 established the ionic nature of the acid- catalysed rearrangement of triphenylmet hyl hydroperoxide (R = Ph). H+ R,C-O-O-H - R2C0 4- ROH Later work148 developed a onestage conversion of triaryhethanols into ketone and phenol. Initial o b ~ e r v a t i o n s ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ established the relative ‘migratory aptitude’ of the aryl groups an electron-releasing substituent increases an 140 D. B. Denney J. Amer. Chem. SOC. 1955 77 1706. 141 D. B. Denney and D. G. Denney J. Amer. Chem. SOC. 1957,79,4806. 142 S. Winstein and G. C. Robinson J. Amer. Chem. SOC. 1958 80 169. 143 R. Criegee and H. Zogel Chem. Ber. 1951 84 215. 14*E.g. P. W. Sherwood Ind. Chemist 1954 30 25 71. 146 H. Hock and S. Lang Ber. 1944 77 257. 148 Cf. L. H. Cone and M. Gomberg Ber.1904,37 3538. 14’ H. Wieland and J. Maier Ber. 1931 64 1205. 148 W. Dilthey F. Quint and H. Dierichs J. prukt. Chem. 1938 151 25. ld9 P. D. Bartlett and J. D. Cotman J. Amer. Chem. SOC. 1950 72 3095; cf. however Y. ISM A. Furuno and S. Sumi Kogyo Kaguku Zasshi 1961,64,472. 446 Lee and U ! electron-attracting decreases the tendency to migrate. Study of competitive migrations established150 that the migratory aptitude increased as the ability to stabilise a positive centre increased; both protonic and Lewis acids strongly accelerated the reaction. In secondary benzylic hydroperoxides aryl groups migrate almost to the complete exclusion of alkyl groups except vinylic groups which are similar in mobility to phenyl. A similar series of migratory aptitude is shown for example in the pinacol-pinacolone rearrangement.151 While the latter reaction shows a very marked steric effect with orthu substitution this effect is much less markedlSa in the peroxide rearrangement reflecting lower steric crowding at the reaction terminus in this case.(There will also be some effect here from interference in the orientation of the benzene ring but this is obviously small.) For alkyl groups the order tertiary > secondary > primary > methyl holds. The position of hydrogen in the series is more d o ~ b t f u l . ~ ~ ~ ~ ~ Kinetic s t ~ d i e s ~ ~ ~ J ~ ” ~ ~ ~ with and other evidence160 give strong support for a rate-determining concerted rearrangement of a protonated form of the hydroperoxide (LXIII). Reaction is of the second order overall; fist order in peroxide first order in acid pseudo-first order at a given acid concentration.Specific rather than 150 M. S. Kharasch A. Fono and W. Nudenberg J. Org. Chem. 1950 15 748; M. S. Kharasch A. Fono W. Nudenberg and A. C. Pushkus ibid. 1950 15 775; M. S. Kharasch and J. G. Burt ibid. 1951,16 150; M. S. Kharasch A. Fono W. Nudenberg and B. Bischof ibid. 1952 17 207. 151 C. K. Ingold ‘Structure and Mechanism in Organic Chemistry’ Bell London 1953 p. 477. 152 A. W. de Ruyter van Steveninck and E. C. Kooyman Rec. Trav. chim. 1960 79 413. 153 E.g. W. Pritzkow and K. A. Miiller Ber. 1956 89 2312. 154 E g . Cf. M. S. Kharasch and J. G. Burt J. Org. Chem. 1951 16 13. 128; N. Kornblum and H. E. De La Mare J. Amer. Chem. SOC. 1951 73 851; A. M. White Thesis Univ. of London 1953; W. J. Farrisey J.Org. Chem. 1962 27 3065; J. Amer. Chem. SOC. 1962 84 1002; H. B. van Leeuwen J. P. Wibaut I. F. Bickel and E. C. Kooyman Rec. Trav. chim. 1959 78 667; A. D. Boggs Thesis Ohio State Univ. 1954. 155 F. H. Seubold and W. E. Vaughan J. Amer. Chem. SOC. 1953,75 3790. 156 0. Wichterle and P. Cefelin CoZZ. Czech. Chem. Comm. 1957 22 1083. 16’ A. W. de Ruyter van Steveninck J. Chem. SOC. 1958 2066. 158 Yu A. Shiyapnikov Kinetika i Kataliz 1960 1,365; V. A. Shushunov and Yu. A. Shiyap- nikov ‘Works on Chemistry and Chemical Technology’ Gorkii 1958 p. 50 1959 p. 102; V. A. Shushanov and Yu. A. Shiyapnikov Doklady Akad. Nauk S.S.S.R. 1959,128 341. 159 M. Bussey C. A. Bunton A. G. Davies T. A. Lewis and D. R. Llewellyn J. Chem. SOC. 1955,2471. le0 A. G. Davies and R. Feld J. Chem. Soc. 1958 4637.447 Quarterly Reviews + f R2C0 3. AOH R-7-OA + H,O R general acid catalysis is inv01ved.l~~ Activation energies suggest an ionic rather than a free-radical r e a ~ t i o n . l ~ ~ J ~ ~ The effects of para-substituents on aryl migra- tion fitf52 the Hammett cr+ (rather than (7) relationship with a large reaction constant of - 4-57 supporting a concerted rearrangement of the protonated material (LXIII). l*O-Labelled methanols gave hydroperoxide free from labe1.lS9 R3C"0H 9 R$O-OH 4- H,'*O Since some rearrangement occurs in these conditions there cannot have been an equilibrium involving R-O+ ; rearrangement is either concerted therefore (via LXII) or occurs immediately upon formation of R-O+ (i.e. k3 k,) thus e + ? R2C-0 etc. e + R,C-0-OH & k H,O + R,C-O+ A The easy decomposition of I-phenylethylhydroperoxide as compared with t-butyl hydroperoxide suggests a synartetic process.The mechanism varies with structure.lsO The overall reaction is somewhat more complex than this.lsl Protonation to form two distinct conjugate acids is possible (LXIII LXIV) and whilst re- arrangement of (LXIV) will not occur a reversible breakdown to hydrogen peroxide and a carbonium ion is to be expected particularly with triarylmethyl hydroperoxides. +'+ R3C-Y-OH H G=== R3E i- H,O R3C-O-oH -9 (LXIV) In the presence of a second methanol equilibration between water methanol hydroperoxide and hydrogen peroxide occurs at a rate comparable with the rate of rearrangement.lel Information on hydroperoxide rearrangement in groups other than aralkyl hydroperoxides is relatively sparse.In nitrogen heterocycles the -C=N.R group migrates to the exclusion of D. E. Blissing C . A. Mutuszak and W. E. McEwen Tetrahedron Letters 1962 No. 17 763 ; J . Amer Chem. SOC. 1964 86 3824. 448 Lee and U$ phenyl alkyl or -N(Aryl) m i g r a t i ~ n l ~ ~ - ~ ~ ~ and one may assume that the oxidative breakdown of tryptophan to forrnylkyn~renine~~~~~~~ can be explained by a similar migration in a hydroperoxide intermediate. The rearrangement of metal alkyls to alkoxides is likely to involve an ionic rearrangement via hydroperoxide.ls8 In certain other cases where isomerisation of hydroperoxides occurs no evidence is available to permit distinction between electrophilic or nucleophilic 0 ~ y g e n . l ~ ~ The electrophilic attack by olefin upon hydroperoxide has been dis- cussed above.170 A number of reviews have appeared on hydroper~xides.l~l-l~~ C.The Baeyer-VWger Reaction.-The oxidation of ketones by peracids proceeds via a peroxy-intermediate with subsequent carbon to oxygen migration leading to an ester or lactone. Since its discovery174 the reaction has been very widely a~p1ied.l’~ d * CO R2 d.CO*OR2 + R30H Use of acids other than Caro’s acid174 has led to impro~ernents.l~~J~~ Mechanistic studies177 led to the formulation of the reaction as initial nucleo- philic addition of peracid to the protonated carbonyl group followed by migra- tion of an electron-pair bearing group in the hydroxyperoxide produced to an electrophilic oxygen. A summary of this evidence is as follows. Rearrangement of [180]benzo- phenone gives phenyl benzoate with the same carbonyl oxygen isotope content 162 C.W. Bird J. Chem. SOC. 1965 3490. 163 D. M. White J. Amer. Chem. Soc. 1964 86 5685. 164 E. H. White and M. J. C. Harding J. Amer. Chem. SOC. 1964 86 5687. 165 L. A. Cohen and B. Witkop J. Amer. Chem. Sac. 1955,77 6595. 166 -4. Ek H. M. Kissman J. B. Patrick and B. Witkop Experientia 1952 8 38. 167 A. A. Haakim and K. A. Thiele Biochem. Biophys. Res. Comm. 1960,2,242. 168 E.g. H. Hock H. Kropf and F. Ernst Angew. Chem. 1959 71 541; cf. A. G. Davies ‘Organic Peroxides’ Buttenvorths London 1961 pp. 120-126 155-1 60. 169 E.g. G. 0. Schenk 0. A. Neumuller and W. Eisfed Angew. Chem. 1960 70 595. 170 W. F. Brill and N. Indictor J. Org. Chem. 1965,30,2074; 1964 29 710; J. Amer. Chem. Soc. 1963 85 141. 171 P. D. Bartlett Rec. Chem. Progr. 1950 11 47.172 J. E. Leffler Chem. Rev. 1949 45 385. 173 E. Testa ‘Oxydationen durch Wasserstoff peroxyd. und persauren die zur spaltung von C-C Bindungen fuhren’ Juris Verlag Zurich 1950. 174 A. von Baeyer and V. Villiger Ber. 1899 32 3625. 175 E.g. L. Ruzicka and E. Stoll Helv. Chim. Acta 1928 11 1159; V. Burkhardt and T. Reichstein ibid. 1942 25 821 1435; H. Heuser A. Segre and A. Plattner ibid. 1948,31 1183; V. Prelog L. Ruzicka P. Meister and P. Wieland ibid. 1945 28 618 1651 ; R. E. Marker J. Biol. Chem. 1940 62 650. 176 J. E. Leffler Chem. Rev. 1949 45 385; C. H. Hassall Org. Reactions 1957 9 74. 17’ R. Criegee Annalen 1948 560 127; S. L. Friess and N. Farnham J. Amer. Chem. SOC. 1950 72 5518; W. E. Doering and L. Speers ibid. 1950 72 5515; W. E. Doering and E. Dorfmann ibid. 1953,75,5595; J.L. Mateos and H. Mechata J. Org. Chem. 1964,29,2026. See ref. 19a in contrast. 449 Quarterly Reviews !+ d- C- R‘ U V ) RL? - 5-R‘ + -OR2 + H+ 0 (LXVI) (LXVII) as the initial ketone.178 The most likely intermediate is (LXVTII) or one of its O-protonated conjugate acids.179 (LXVII I ) PH R,C-O-OA The actual rearrangement is clearly intramolecular; several groups of w o r k e r ~ l ~ ~ ~ ~ ~ have reported complete retention of configuration in migrating centres. The relative ‘migratory aptitude’ of groups in unsymmetrical ketones has been extensively e ~ a m i n e d . ~ ~ ~ J ~ ~ para-Substitution in a phenyl group influences migratory aptitude in proportion to electron-releasing effectla8 suggesting a concerted rearrangement. ortho-Substitution reduces mobility vis-2-vis para.lS6 In general the group best able to sustain a positive charge migrates most readily but this electronic effect is not the sole factor.Sauersla7 confirmed the observation17* that the only lactone isolable from the action of Caro’s acid on camphor was the ‘abnormal’ a-campholide but showed in agreement with other that the ‘normal’ lactone (that is the product of migration of a tertiary group) was further decomposed in these conditions. Further the ‘normal’ lactone only was obtained on oxidation of camphor with buffered peracetic acid.la7 The expected bridgehead migration occurs with norcamphor.lsg Epicamphor and l-methylnorcamphorlgo differ sharply with exclusive methylene migration in the former case while fenchone gives mixed products. 178 W. E. Doering and E. Dorfmann J.Amer. Chem. SOC. 1953,75,5595; cf. C. A. Bunton Chem. and Ind. 1954 19 1. 179 Originally suggested by R. Criegee and R. Kaspar Annulen 1948 560 127. 180 T. F. Gallagher and T. H. Kritchevsky J. Amer. Chem. SOC. 1950,72,882; R. B. Turner ibid. 1950 72 878; L. H. Sarrett ibid. 1957 69 2899; R. E. Marker ibid. 1940 62 2543; V. Burkhardt and T. Reichstein Helv. Chim. Acta 1942 25 1434. 181 K. Mislow and J. Bremner J. Amer. Chem. Soc. 1953 75 2319. 182 J. W. Wilt and A. Danielzadeh J. Org. Chem. 1958 23,920; J. A. Berson and S. Suzuki J . Amer. Chem SOC. 1959 81 4088. 183 R. B. Turner J. Amer. Chem. SOC. 1950 72 879. lE4 M. F. Hawthorne W. D. Emmons and K. S. McCallum J. Amer. Chem. SOC. 1958 80 6393; R. R. Sauers and R. W. Ubersax J. Org. Chem. 1965 30 3939; J. R. Owen and W. H. Saunders J .Amer. Chem. SOC. 1966 88 5809 5816. 185 Y . Yukawa and T. Yokoyama Mem. Inst. Sci. Ind. Res. Osaka Univ. 1956 13 171; (Chem. Abs. 1957,51,2633); S . L. Friess and R. Pinson J. Amer. Chem. SOC. 1952,74 1302. 186 (a) W. H. Saunders J. Amer. Chem. SOC. 1955 77,4679; (b) W. E. DoeringandL. Speers ibid. 1950 72 55 15. 187 R. R. Sauers J. Amer. Chem. SOC. 1959 81 925. lBB J. D. Connolly and K. H. Overton Proc. Chem. SOC. 1959 188. lBo R. R. Sauers and G. P. Ahearn J. Amer. Chem. SOC. 1961 83 2759; cf. J. T. Edward and P. F. Morand Cunad. J . Chem. 1960 38 1325. J. Meinwald and E. Frauenglass J. Amer. Chem. SOC. 1960 82 5235. 450 Lee and U. A steric effect of addition to the carbonyl was suggested,lgl but since the relative migration tendency for the methylene and bridgehead carbons in these cases is somewhat dependent upon the reagent used we suggest that an additional factor may be the nature of the leaving group.Thus in the case of camphor two alternatives are either (i) that the addition reaction is rate-limiting with one reagent (Car03 acid) but rearrangement is rate-limiting with another or (ii) that the relative energies in the transition states in the rearrangement are greatly different. The sulphate ion is a better leaving group than acetate which should accelerate the rearrangement step and give less s e l e c t i ~ i t y ; ~ ~ ~ - ~ ~ ~ similar differ- ences have been reported with other peracids.lg2 Likewise the change in size of the acid group should alter the relative stabilities of the various transition states leading directly to altered product ratios.A complicating factor may be the rapid interconversionlg3 of epimeric adducts (LXIX) via a cyclic intermediate (LXX) although this seems at variance with the results of the isotopic labelling experiments. (LXIXb) Some workerslg4 suggest that a further factor is relief of Pitzer strain since where electronic factors are approximately equal (e.g. fenchone) oxidation proceeds mainly as forecast for maximum strain release. While saturated groups migrate without isomerisation may occur with unsaturated ~ o m p o u n d s . ~ ~ ~ ~ ~ ~ Kinetic studies have been made.lg6-lg9 In certain cases reaction is of the second order overaIl;lg6 first order in ketone first order in oxidant the addition to the carbonyl group is slow and rate-determining. Steric hindrance to addition is shown in alicyclic ketones.4Substituted cyclohexanones show acceleration in the order H < Me < t-butyl the authors suggesting a non-classical carbonium ion giving 1 ,4-stabilisationlg6 (LXXI) since analogous 1 +effects are known.lg7 However all the latter examples involve 4halogeno- or 4-oxy-containing groups. a-Halogeno-ketones also show reduced reaction rates.lgs IQ1 M. F. Murray B. A. Johnson R. L. Pederson and A. C. Off J. Amer. Chem. SOC. 1956 78,981 ; cf. D. Y . Curtin Rec. Chem. Progr. 1954,15 1 11 and also ref. 189. lv2 R. W. White and W. D. Emmons Tetrahedron 1962 17 31. lg3 A. Rassat and G. Ourisson Bull. SOC. chim. France 1959 1133. 19* R. R. Sauers and J. A. Beisler J. Org. Chem. 1964 29 210. lQ5 J. Meinwald M. C. Seidel and B. C. Cadoff J. Amer. Chem. SOC. 1958 80 6303; G.Buchi and I. M. Goldman ibid. 1957 79 4741. lg6 J. L. Mateos and H. Mechaca J. Org. Chem. 1964,29,2026. lg7 E.g. L. Owen J. Chem. SOC. 1949 320; E. L. Bennett J. Amer. Chem. SOC. 1952 74 5076; H. L. Goering ibid. 1957 79 6270; D. S. Noyce ibid. 1957 79 755; 1960 82 885 1246. lQ8 S. L. Friess and P. E. Frankenburg J. Amer. Chem. SOC. 1952,74,2679; S . L. Friess and R. Pinson ibid. 1952 74 1302; S. L. Friess and A. H. Soloway ibid. 1951 73 3968; 1949 71,2571; 1950,72 5518. lQO M. F. Hawthorne and W. D. Emmons J. Amer. Chern. Soc. 1958 80 6398. 451 Quarterly Reviews CR3 0- 0,COR (LXX I) Rates for different ring sizes in alicyclic and cycloalkyl methyl ketones were compared.lg8 A rough correlation with Hammett's equation obtains for sub- stituted acetophenones. Since the rateslgs were of the first order with electron- attracting substituents and intermediate for acetophenone a change in rate- controlling step is obviously occurring.In contrast other workerslg9 conclude that the acid-catalysed decomposition of the peracid-ketone complex is rate-controlling. Distinct acid catalysis suggestive of general acid catalysis is observed even with weak peracids and catalysis by strong acids is marked.186b Increased solvent polarity accelerates the reaction. It seems that the most reproducible reagent is trifluoroperoxyacetic acid199-201 and with this third-order (sometimes pseudo-first or -second order) kinetics are observed. This implies that the intermediate species is derived from all three reagents (that is ketone peracid and acid). R? I d r n . m R? OH R*' R"' (LXXII) (LXXII) (Lxx I I I) (LXXIV) (LXXV) Hawthorne and Emmonslg9 confirm this adduct formation.The protonated intermediate (LXXII) will be in rapid equilibrium with three further forms (LXXIII LXXIV LXXV) and of these two (LXXIV LXXV) appear most favourable to rearrangement since both would give rise directly to neutral acid and protonated ester.lg9 However in view of the strength of trifluoroacetic acid,202 protonation to give (LxXrV) or (LXXV) seem rather unfavourable and a cyclic intermediate would seem to us more likely namely (LXXVI). + $\ + HO,. H\ R'\ @o-'.H'm"o'?'c-~~ ,c=m + CRl ?+ Rz 0-H O R3/ 'o-o/ R3<?>0.. .__._ 0.;;' RO 04 'C/ 'C-R' [ ] U X V O 2oo W. D. Emmons and G. B. Lucas J. Amer. Chem. SOC. 1955,77,2287. 201 W. F. Sagar and A. Duckworth J .Amer. Chem. SOC. 1955 77 188. 202 Cf. M. Ussanovitch and V. Tartakovskaya Zhur. obshchei Khim. 1946 16 1987; T. Sumarokova and Z . Grishken ibid. p. 1991; L. P. Hammett and A. J. Deyrup J. Amer. Chem. SOC. 1933 55 1900. 452 Lee and Ufl Although the measured Hammett reaction constants gave negative values these are the result of composite effects and cannot be held to confirm con- clusively a concerted process.203 Catalysis by trifluoroacetic acid of the cyclohexanone reaction with peracetic and trifluoroperacetic acids respectively proceeded much less rapidly in the former case in keeping with a rate-determining rearrangement. In a slightly different context Winstein204 examined analogues of the postu- lated Baeyer-Villiger intermediate (LXXII) namely (LXXVLI). /OMe 'o* ?Me k-7-0-014 F - [..-C,g--OA] - Me<+ Me (LXXVI I) OMe MeC02Me Mo-C<-OMe OMe Alcoholysis of (LXXVII) unexpectedly gave acetate esters.204 While the Baeyer-Villiger reaction is essentially acid-catalysed some study Enolisable P-diketones do not undergo a normal Baeyer-Villiger oxidation,206 has been made205 of the use of alkaline hydrogen peroxide.and the peroxide attack on benzils proceeds also by a different route.207 D. The Elbs PersuIphate Oxidation.208-Treatment of a phenol in alkaline solution with persulphate followed by acidification gives mainly para hydro- xylation although ortho products (catechols) are formed as secondary products. The reaction has been used fairly extensively by several groups of ~ o r k e r s . ~ ~ ~ - ~ Application to anilines gives predominantly ortho substitution,214 although 2-aminobenzaldehyde gives the Dakin reaction product viz.2-arninophen01.~~~ It may be noted that benzoyl peroxide converts N-alkylanilines into 2-benz- amido-phenols.216 Baker and his co-workers210 suggest a radical mechanism 203 R. Stuart and K. Yates J. Amer. Chem. SOC. 1958 80 6355. 204 S. Winstein Tetrahedron Letters 1962 No. 13 567. 205 H. 0. House and R. L. Wasson J. Org. Chem. 1957,22 1157. 206 H. 0. House and W. F. Gannon J. Org. Chem. 1958,23 879. 207 H. Kwart and N. J. Wegemer J . Amer. Chem. SOC. 1961 83 2746. 208 K. Elbs J. prakt. Chem. 1893 48 179. 209 T. R. Seshadri Experientia Suppl. 2 14th International Congress of Chemistry Zurich 1955 p. 258; T. R. Seshadri and P. S . Rao Proc. Indian Acad. Sci. 1943 A 18,222. 210 W. Baker and N. C. Brown J.Chem. SOC. 1948 2303; W. Baker and R. I. Savage ibid. 1938 1602; 1941 662. 211 J. Forrest and V. Petrow J. Chem. SOC. 1950 2340. 212 F. E. King T. J. King and L. C. Manning J. Chem. SOC. 1957 563; E. J. Behrman and B. M. Pitt J. Amer. Chem. SOC. 1958 80 3717. 213 K. C. Roberts J. Chem. SOC. 1960 785. 214 E. Boyland D. Manson and P. Sims J. Chem. SOC. 1953,3613; E. Boyland and P. Sims ibid. 1954 980. 215 E. Bamberger Ber. 1903 36 2042. 216 J. T. Edward J. Chem. Soc. 1954 1464. 453 Quarterly Reviews largely on the basis of free-radical formation in persulphate decomp~sition.~~~ Support for this is the finding of some coupling products.218 Behrman and Walker219 investigated the reaction kinetically. They confirmed that phenoxide ion is involved which is further supported by the variation of this effect with phenol s t r u c f ~ r e .~ ~ ~ ~ ~ ~ ~ Addition of cupric or ferrous ions equally has no effect since electron-releasing substituents in the phenol accelerate the reaction,219 a nucleophilic attack on persulphate seems likely as first step. E. The Dakin Reaction.-Alkaline hydrogen peroxide oxidises 0- and p-hydroxy- benzaldehydes221 and -acetophenones222 to catechols and quinols. In mechanism it is closely parallel to the Baeyer-Villiger reaction. It is not surprising therefore to find that electron-releasing groups facilitate reaction.223 For benzaldehyde itself hydride migration is the main reaction only traces of phenol being Salicylaldehyde gives catechol almost q~antitatively,2~~ presumably via a cyclic intermediate (LXXIX) . ph PhCOOH cu.99% PhCHO H-?-OW O- 'N Pho-fH c-. 0.7% (UXVI I I) 217 For a review see S . M. Sethna Chem. Rev. 1951 49 91 ; see also K. B. Wiberg J. Amer. Chem. SOC. 1959,81,252; 1953,75 1439; E. Ben Zvai and T. L. Allen ibid. 1961,83,4352; D. L. Ball M. M. Crutchfield and J. 0. Edwards J. Org. Chem. 1960,25 1599. 218 I. M. Kolthoff and I. K. Miller J. Amer. Chem. SOC. 1951 73 3055; R. G. R. Bacon R. Grime and D. J. Munro J. Chem. SOC. 1954 2275. 219 E. J. Behrman and P. R. Walker J. Amer. Chem. SOC. 1962 84 3454. cf. also S. M. Sethna ref. 217. 220 I. M. Kolthoff Rec. Trav. chim. 1923 42 969. 221H. D. Dakin Amer. Chem. J. 1909 42 477. 222 W. Baker E. H. T. Jukes and C. A. Subrahmanyam J. Chem. SOC. 1934,1681 ; W. Brown H. F. Brady J. Gumb and D. Miles ibid. 1953 1615. 223 E. Spath M. Pailer and G.Gergeley Ber. 1940 73 935. 224 J. D'Ans and A. Kneip Ber. 1915 48 1143; J. Hawkins J. Chem. SOC. 1950 2169. 225 A. Wacek and H. 0. Eppinger Ber. 1940,73,644; A. Wacek and A. V. Bezard ibid. 1941 74 845. 454 Lee and Ufl Labelling of the carbonyl group in p-methoxyacetophenone followed by oxidation led to products in which all label was retained in the acetate carb0ny1.l~~ 4 Other Reactions in which Oxygen behaves ElectrophilicaIIy A. 0zonolysis.-This is a topic large enough to warrant separate review. There is a considerable volume of literature availablezz6 including some reviewszz7 and it is not proposed to discuss this further. B. Peracid Rearrangements under the Influence of Lewis Acids.-The rearrange- ment of peracids under the influence of Lewis acids has been discussed and probably involves2z8 a concerted carbon-to-oxygen migration with simultaneous attack on carbonyl carbon by the second peroxy-oxygen (LXXX).- C02 + PhCO*OPh I 9 __t [ fSbCL3B BSbC6 Ph-C. ,C\ /Ph 0 0 Ph-C fCC-Ph 0-0’2 (LXXX) Using labelled oxygen compounds Denney et showed direct attack by carbanion on peroxides in their reactions and reviewed work already done on Grignard reagents and a number of active-methylene compounds. The Friedel-Crafts hydroxylation of aromatic compounds by peracids was examined inter aka by Waring and Hart,230 who originally postulated attack by HO+ but an alternative of nucleophilic attack by the aromatic compound upon a peracid-Lewis acid complex is also possible (LXXXI). (LXXX I) In agreement electron-releasing substituents facilitate reaction.z30 Aromatic oxygenation can be achieved with aryl peroxides and with peroxydicarbonates a26 E.g.D. G. M. Diaper and D. L. Mitchell Cunad. J. Chem. 1965 43,319; P. S. Bailey J. Amer. Chem. SOC. 1956 78 3811; J. Org. Chem. 1958 23 1089; 1962 27 1192 1198; R. H. Callaghan et al. ibid. 1961 26 1379; J. Pasero et ul. Bull. SOC. chim. France 1960 1717; D. G. M. Diaper and D. L. Mitchell Cunad. J . Chem. 1962,40 1189; K. A. Pollart and R. E. Miller J. Org. Chem. 1962,27,2392; D. G. M. Diaper and D. L. Mitchell Canad. J. Chem. 1960 38 1976; P. S. Bailey and S. S. Bath J. Amer. Chem. SOC. 1951 73 3120; 1960 83 6136; G. Riezebos J. G. Grimelikhysen and D. A. van Dorp Rec. Trav. chim. 1963 82 1234; P. R. Story R. W. Murray and R. D. Youssefyeh J. Amer. Chem. SOC. 1966 88 3143 3144; F. L. Greenwood ibid.1966,88 3146. 227 R. Criegee Rec. Chem. Progr. Kresger-Hooker Sci. Lib. 1957 18 111; Annalen 1948 560 127; P. S. Bailey Chem. Rev. 1958 58 925. 228 D. Z. Denney T. M. Valega and D. B. Denney J. Amer. Chem. SOC. 1964 86,46. 228 D. Z. Denney S. 0. Lawesson C. Frisell and D. B. Denney Tetrahedron 1963,19 1299. 230 (a) A. J. Waring and H. Hart J. Amer. Chem. SOC. 1963 85 2177; 1964 86 1454. (b) P. Kovacic ibid. 1965 87 1566 4811. J. Org. Chem. 1966 31 2011. (c) P. Kovacic and M. E. Kurz J. Amer. Chem. SOC. 1966 88 2068; Chem. Comm. 1966 321; Tetrahedron Letters 1966 2689. 455 Quarterly Reviews using cupric or ferric chloride catalysts. Although a purely radical mechanism can be invoked Kovacic and K ~ r z ~ ~ ~ ~ have pointed out that some evidence notably isomer distribution supports possible involvement of an oxonium ion resulting from oxidation of an oxyradical as shown.B (RCO,) 4- CU' - RC-0' + Cun -I- RCO,' 9 F) RC-0' 4- CU' + RC-0' i- CU' R = aryl or O-alkyl C. Electrophilic Behaviour involving HypohaIites.-Formation of acetyl hypoiodite from iodine and peracid is postulated231 in the iodination of aryl compounds with these reagents but the mechanism is not clear. In an anomalous Hunsdieker reaction observed by Pandit and Dirk232 a possible mechanism involves also nucleophilic attack by aromatic electrons upon oxygen (in this case an acyl hypobromite). D. Rearrangement of Diepoxides.-Rearrangement of 1 ,ddiepoxide rings to epoxy-ketones or diepoxide occurs together with formation of allylic ketol (LXXMI) - (LXXXTV) in many and some other234 systems.Although usually stated to be base-catalysedB5 rearrangements these could possibly involve electrophilic behaviour of oxygen (LXXXV). OH 231 Y . Ogata and K. Nakajima Tetrahedron 1964 20 43 2751; Y. Ogata and K. Aoki J. Org. Chem. 1966 31 1625. 232 U. K. Pandit and I. P. Dirk Tetrahedron Letters 1963 891 ; cf. J. W. Wilt and D. D. Oathaudt J. Org. Chem. 1958,23,218; J. W. Wilt and J. L. Finnerty ibid. 1961,26,2173. 233 R. J. Conca and W. Bergman J. Org. Chem. 1953,18,1104; E. L. Skau and W. Bergman ibid. 1939 3 166; W. Bergman F. Hirschman and E. L. Skau ibid. 1940 4 29; M. Matic and D. A. Sutton J. Chem. Soc. 1953 349; R. B. Woodward and R. H. Eastman J. Amer. Chem. Soc. 1950 72 399; T. G. Halsall W. J. Rodewald and D. Willis Proc. Chem. SOC. 1958,231; A. Windaus and G.Brunkin Annafen 1928,460,225. a34 E.g. D. C. De Luca Diss. Abs. 1965 25 2225. 235 E. J. Agnello R. Pinson and G. D. Laubach J. Amer. Chem. SOC. 1956 78 4756. 456 Lee and Uf (LXXXVI) W The alternative mechanism which postulates initial proton removal (LXXXVI) is definitely not possible in certain cases (e.g. ergosteryl diep~xides)~~~ where the oxygen atoms are attached to tertiary carbons and it is difficult to explain diepoxide formation by this mechanism.237 B l a d ~ n ~ ~ ~ reports a similar rearrange- ment of a diepoxide (LXXXVII) which almost certainly involves electrophilic oxygen (LXXXVIII). (U<XXVl I) (UtXXVl I I) / 0 4P 0 The rearrangement of steroid hydroperoxides may go by a similar pathway.239 We are grateful to Professor Wesley Cocker of Dublin for his assistance.236 E.g. A. Windaus and J. Brunkin Annalen 1928 460 225. 237 Cf. also W. Bergman and M. B. Myers Annalen 1959 620,46; J. Org. Chem. 1960,25 1451; D. J. Giancola Dissertation Yale Univ. 1956. 23s G. 0. Schenk 0. A. Neumuller and W. Eisfeld Angew. Chem. 1960,70 595. P. Bladon and T. Sleigh Proc. Chem. SOC. 1962 183. 457

ISSN:0009-2681    

DOI:10.1039/QR9672100429

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Chemical Applications of Oxygen-17 Nuclear and Electron Spin Resonance By Brian L. Silver and Zeev Luz ISOTOPE DEPARTMENT THE WEIZMANN INSTITUTE OF SCIENCE REHOVOTH ISRAEL 1 Introduction Of the three stable isotopes of oxygen l60 170 and l80 only 170 has a magnetic nucleus. It has a spin quantum number of and a gyromagnetic ratio of 577 c./sec. per gauss. Unfortunately the natural abundance of 1 7 0 is only 0.037% and for this reason despite the chemical importance of oxygen relatively little work has been reported on the effect of 170 on nuclear and electron spin reson- ance (n.m.r. and e.s.r.). Recent progress in the large-scale separation of oxygen isotopes has resulted in an increased availability of materials enriched in 170. During the last six years 1 7 0 n.m.r. and to a less extent e.s.r.have been used to study a number of fundamental problems in chemistry and this is a Review of the achievements of these studies. The coverage of the literature is essentially complete up to the beginning of 1966 and we have discussed only a part of the articles published during 1966. 170 n.m.r. signals were first observed by Alder and Yul in water and a number of organic liquids both in isotopically normal and slightly enriched material. The list of compounds was extended by Weaver et aL2 and Dharmatti et aZ.,3 both groups using isotopically normal compounds. Christ* and subsequently Christ Diehl Schneider and Dahn5 examined over a 100 compounds most of them organic. Shifts for inorganic compounds have been published by Figgis et aL6 and Bramley et aL7 and more recently by Jackson and Taube.* Very few cases have been reported of the effect of 1 7 0 on e.s.r.spectra. Bairdg in 1961 observed 170 hyperfine splitting in di-s-butylnitric oxide and subse- quently such splitting has been reported for several other free radicals.1°-15 F. Alder and F. C. Yu Phys. Rev. 1951,81 1967. H. E. Weaver B. M. Tolbert and R. C. LaForce J. Chem. Phys. 1955,23 1956. S . S. Dharmatti J. K. Sundara Rao and R. Vijayoraghan Nuovo cim. 1959,11 656. H. A. Christ Helv. Phys. Acta 1960 33 572. H. A. Christ P. Diehl H. R. Schneider and H. Dahn Helv. Chim. Acta 1961 44 865. B. N. Figgis R. G. Kidd and R. S. Nyholm Proc. Roy. Soc. 1962 A 269,469. R. Bramley B. N. Figgis and R. S. Nyholm Trans. Faruday SOC. 1962 58 1893. J. A. Jackson and H. Taube J. Phys. Chem. 1965 69 1844. J. C.Baird J. Chem. Phys. 1962 37 1879. lo A. Reiker and K. Schemer Tetrahedron Letters 1965 19 1337. l1 K. Dimroth F. Bar and A. Berndt Angew. Chem. 1965,77,217. l2 W. M. Gulick and D. H. Geske J. Amer. Chem. SOC. 1965 87,4049. l3 (u) B. L. Silver Z. Luz and C. Eden J. Chem. Phys. 1966 44,4258; (b) W. M. Gulick jun. and D. H. Geske J. Amer. Chem. SOC. 1966,88,4119. l4 Z . Luz B. L. Silver and C. Eden J. Chem. Phys. 1966,44,4421. l5 R. W. Fessenden and R. H. Schuler J. Chem. Phys. 1966,44,434. 458 Silver atui Luz Most of this Review will be concerned with 170 n.m.r. in solutions. The effect of 170 on e.s.r. spectra will be dealt with in Section 3. 170 n.m.r. measurements in dilute solutions accurate line-width studies and the detection of 170 hyperfine splitting in e.s.r. require compounds isotopically enriched in 170.Water containing up to about 10 atoms % of 170 has been commercially available for several years and recently 1 7 0 isotope has been enriched up to 65 atoms % by thermal diffusion of oxygen gas.16 Since 1 7 0 is normally supplied in the form of water the most convenient method for syn- thesising 170-labelled compounds is often an isotopic exchange reaction between the normal compound and solvent H?’O or a mixed solvent containing Hk70. The exchange of isotopic oxygen with organic compounds and methods of isotopic synthesis have been reviewed.17 2 Nuclear Magnetic Resonance A. Line-widths.-Studies of line-widths in n.m.r. have provided much insight into details of nuclear relaxation processes. In subsequent sections we will dis- cuss the effect on n.m.r.line-widths of chemical exchange and interactions with paramagnetic ions. In the absence of these factors the dominant effect on the 170 line-width in solution is quadrupole relaxation which depends on the electric field gradient at the 170 nucleus and the tumbling time of the molecule.l* The field gradient necessarily vanishes if the symmetry around the oxygen atom is cubic or higher or it could vanish owing to a suitable electron distribution in the orbitals closely associated with the oxygen atom. A case in which oxygen is at a site of cubic symmetry is not expected in molecules and is not known though it often occurs in crystalline lattices. In fact an n.m.r. signal has been detected in solid Mg170 in which the oxygen is located at a site with cubic symmetry.19 In molecules the field gradient at a nucleus is determined largely by the population of thep orbitals.20 A zero contribution to the field gradient at the 170 nucleus occurs when the total population of thep orbitals in the bonds surrounding the oxygen nucleus is six (02-) zero (06+) or (in the case of sp3 hybridisation) four (02+).Large formal positive charges on oxygen are very unlikely but an oxygen bonded to a highly electropositive atom may carry close to two electronic charges and have a very small field gradient. Christ and Diehl18 measured the 170 line-widths and calculated the quadrupole interaction constants for a number of diamagnetic organic compounds. The linewidths ranged from 100 to lo00 c./sec. and the quadrupole coupling constant varied from 3 to 10 Mc./sec. Measurements on a series of alcohols showed that the 170 line-widths varied linearly with the calculated tumbling time of the molecule which indicates that the field gradient in the hydroxyl group of different l6 I.Dostrovsky and F. S. Klein personal communication. 17(u) D. Samuel and B. L. Silver in ‘Advances in Physical Organic Chemistry’ vol.3 Academic Press New York 1965 p. 123; (b) I. Dostrovsky and D. Samuel in ‘Inorganic Isotopic Syntheses’ W. A. Benjamin New York 1962 ch. 5. H. A. Christ and P. Diehl 1 lth Colloque Ampere Eindhoven 1962 p. 224. la J. A. Jackson J . Phys. Chem. Solids 1963 24 591. W. J. Orville-Thomas Quart. Rev. 1957 11 162. 459 2 Quarterly Reviews alcohols is very nearly constant. It is generally observed that 1 7 0 line-widths decrease with increasing temperature21 as would be expected from the corre- sponding decrease in tumbling times.B. Chemical Shifts in Diamagnetic Molecules.-An extensive survey of the 1 7 0 chemical shifts of organic compounds was carried out by Christ Diehl Schneider and Daha5 Most of the measurements were made on isotopically unenriched materials. This work has provided the basic data for the systemisation of 170 chemical shifts. The shifts covered a range of up to lo00 p.p.m. and almost all compounds exhibited a low-field shift compared with H270. The most striking feature of the results is the marked division between the shifts of doubly bonded oxygen which usually fall below -250 p.p.m. relative to H&70 and oxygen involved in single bonds (singly bonded oxygen) which usually fall above -250 p.p.m. A h e r classification of the shifts is provided by the fact that they are grouped into small ranges characteristic of each functional group.Chemical shifts are conveniently considered as being the sum of two contribu- tions 8 = 8~ + 8 ~ where 8~ is the diamagnetic and 8~ the paramagnetic contribution. Order-of-magnitude considerations and substituent effects indicate that the dominant contribution to the 170 chemical shift is due to paramagnetic shielding. That this is so at least for doubly bonded oxygen was convincingly demonstrated by Figgis Kidd and Nyho1m.G The paramagnetic shift arises from the creation of electronic angular momentum about the nucleus via the mixing into the ground state of excited states having angular momentum. Mixing results from the perturbation due to the external magnetic field and is inversely proportional to the energy difference AE between the ground and excited states.In the case of doubly bonded oxygen e.g. the carbonyl group the dominant contribution most probably arises from mixing into the ground state of a state resulting from the excitation of a single electron from the non- bonding orbital n localised on the oxygen atom to the.rr* antibonding orbital of the carbonyl bond. This results in the creation of electronic angular momentum about the oxygen nucleus. The paramagnetic shift is then propor- tional to:s < IO(L21T*> AEO,, where < 0 I L2 I .rr* > is the matrix element of the square of the angular momen- tum operator between the ground state and the antibonding n* state and AE,,, is the n +T* transition energy.Figgis et aL6 showed that there is a linear correlation between the chemical shift of doubly bonded oxygen and the reciprocal of AE,,, (Figure 1). Thus the diamagnetic shift of 170 8 ~ appears to vary little between the different compounds. Figgis et aL6 measured the 1 7 0 shifts in a number of diamagnetic transition- metal oxyanions of general formula M04n-. The shifts again give a very good linear correlation with the inverse of the energy of the lowest observed electronic 21 S. W. Rabideau and J. A. Jackson J. Chem. Phys. 1965,41 3405. 460 Silver and Luz Fig. 1 Correlation between the chemical shift of doubly bonded oxygen and n+w* transition energy [From B. N. Figgis R. G. Kidd and R. S. Nyholm Proc. Roy. SOC. 1962 A 269,4691 1. (NH2)2C0 8. CH,CO.CI 15. Cyclopentanone 2.CH,CO.OH 9. Furfural 16. CHsNOa 3. HCO.OH 10. Cl,C CH.0 17. CSH,NO 4. CH,CO.NH 11. (CH&CO 18. (CBHJ2N.NO 5. HC0.NH2 12. CH,CH2CH.0 19. CSH,ONO 6. (CHsCO)2O 13. CH,CH.O 20. C4H,0N0 7. CHsCO.OCzH 14. Cyclohexanone transition. However a calculation on permanganate ion showed that the main contribution to the paramagnetic shift does not come from the lowest excited state and the correlation must therefore as the authors pointed out be regarded as somewhat fortuitous. Nevertheless the dominant contribution of the para- magnetic shift is clearly demonstrated. The fact that different functional groups are each characterised by a com- paratively narrow range of chemical shifts can sometimes be used in determining the chemical nature of the group in which the oxygen atom finds itself.For example acetone in water gives a single 1 7 0 resonance at -520 p.p.m. which 461 Quarterly Reviews falls in the range characteristic for carbonyl oxygen. On the other hand form- aldehyde in aqueous solution has a single resonance at - 50 p.p.m. characteristic of the hydroxyl group. This indicates that formaldehyde is completely hydrated as has also been shown by other methods. Consistently acetaldehyde which is known to be about 50% hydrated in aqueous solution at room temperature gives two signals one at -560 p.p.m. due to the unhydrated acetaldehyde and another at -70 p.p.m. from the hydrated (Figure 2). "0 ~ I00 p.p.m. - CH~CH(OH)~ t I !O F I Fig. 2 The 170 nuclear magnetic resonance spectrum of a concentrated solution of acetaldehyde in water enriched in 1 7 0 . The smaller height of the line due to the hydrate is due to its much greater width [From P.Greenzaid Z . Luz and D. Samuel J . Amer. Chem. Soc. 1967 89 7491 An interesting application of 170 n.m.r. shifts is in the study of the equilibria between the two enol forms in asymmetric P-diketone~.~~ /%Diketones are known to exist as mixtures of tautomeric forms. In asymmetric #3-diketones RC( :O)CHRIC( :O)R2 there are one keto and two rapidly interconverting en01 forms (I and 11) possible OH 0 0 OH I I1 II I R-C = CR1-C-R2 + R-C-CRl= C-R2 (1) (11) A given oxygen atom in the en01 form can exist in two different environments; either as a carbonyl oxygen with a chemical shift SC0 or as a hydroxyl oxygen with shift 8COI-r. Owing to fast intramolecular proton transfer the actual spectrum of the enol forms will consist of two lines due to the two nonequivalent oxygens the observed shift of a given oxygen being the weighted average of its shifts in the two forms.Thus the observed chemical shift of the oxygen adjacent to R is given by 8R=8COH x fraction of (I) + 8CO x fraction of (11). A similar equation holds for the oxygen adjacent to R2. By using the tabulated5 chemical 22(a) P. Greenzaid Z . Luz and D. Samuel J. Amer. Chem. SOC. 1967 89 749; (b) P. Greenzaid unpublished work. 23 M. Gorodetzky and Z. Luz and Y . Mazur J. Amer. Chem. Soc. 1967 89 1183. 462 Silver and Luz shifts for the keto and hydroxyl oxygen it is possible from the observed chemical shifts to determine the position of equilibrium between forms (I) and (11). Specific isotopic labelling of one of the two oxygens in the asymmetric &dike- tone enables the 1 7 0 peaks to be assigned.The equilibrium constants for a number of asymmetric p-diketones have been determined in this way.23 The shift of the H270 line was measured24 in solutions containing a large range of 1 1 diamagnetic electrolytes. Both positive and negative shifts occur and it was possible to show that the observed shifts can be broken down into ionic contributions due separately to the cations and anions. The results were interpreted in terms of a direct interaction between anions and water molecules. C. Contact Shifts-A special kind of chemical shift is that due to the scalar interaction between the nucleus and the magnetic moment of unpaired electrons in paramagnetic molecules. Often these shifts fall far outside the range of shifts covered by diamagnetic compounds.The contribution of this so-called contact shift to the total shift is where A is the isotropic hyperfine interaction constant which is a measure of the energy of the interaction between a nucleus and the unpaired spin in the s shells surrounding the nucleus. The hyperfine interaction also contributes to the n.m.r. line-width and often this broadening is so extreme as to prevent observa- tion of an n.m.r. resonance. Nevertheless if the electron spin relaxation time or some other characteristic exchange time is short compared with A-l it is sometimes possible to observe n.m.r. signals in paramagnetic molecules and such signals have been directly observed from the solvation shell of Ni2+ and Co2+ in aqueous solutions and Co2+ in methan01.~~~~~ Consistently for these ions the electron spin relaxation time is very short and at low temperatures the exchange of solvent molecules in and out of the solvation shell is sufficiently slow to allow separate resonances to be observed for the solvation shell and bulk molecules.A 1 7 0 resonance was also found for manganese(n1)-trisacetyl- acetone in benzene s0lution,2~ for which the electron spin relaxation time is known to be very short. In the solid state a 1 7 0 n.m.r. signal has been reported for MnO and COO powder.28 In all these cases a shift of the order of 1 % to low field was observed indicating a positive spin polarisation at the oxygen nucleus. The hyperfine shift of the hydration-shell molecules of paramagnetic ions can be deduced from the observed 1 7 0 shift of solvent water provided the chemical exchange of the hydration-shell molecules is fast compared with the frequency shift (see Section 2F).In this case only a single 1 7 0 signal is observed the shift of which is the weighted average shift of the bulk and solvation-shell molecules. 24 Z. Luz and G. Yagil J. Phys. Chem. 1966,70,554. 25 R. E. Connick and D. Fiat J. Chem. Phys. 1966,44,4103. 26 D. Fiat B. L. Silver Z. Luz unpublished work. 27 Z. Luz B. L. Silver and D. Fiat J. Chem. Phys. 1967 46,469. ** D. E. O’Reilly and T. Tsang J . Chem. Phys. 1964,40 734. 463 Quarterly Reviews Knowing the concentration of the paramagnetic ions and their solvation number one can deduce the contact shift of the 1 7 0 in the solvation shell. The hyperfine interaction constants between co-ordinated water and bivalent cations of the first transition series and the tervalent rare-earth series have been obtained in this way.29-32 In all these hydrates the source of the unpaired electron is a para- magnetic metal ion.That contact shifts of 1 7 0 occur at all is because unpaired spin is delocalised from the metal on to the ligand. More specifically it is necessary that unpaired spin resides in an s orbital of the 170 atom. For the first series of transition- metal ions all 1 7 0 contact shifts reported have been paramagneticF9 i.e. the external field at the 1 7 0 nucleus is augmented by the field due to the magnetic moment of the unpaired electrons. The mechanism by which spin is delocalised in these complexes is via the bonds formed by the overlap of the cation 3d orbitals with the oxygen orbitals resulting in the transfer of spin from the metal to the 170 2s orbitals.An estimate of the covalency of the metal-oxygen bond can be obtained from the magnitude of the contact shift.27 In the com- plexes of the transition-metal ions of the first series the orbital angular momentum is largely quenched so that the electron spin on the metal is aligned essentially parallel with the external magnetic field and therefore the 1 7 0 contact shift is paramagnetic. Lewis Jackson Lemons and Taubeso have reported the 1 7 0 shift in aqueous solution of all the rare-earth ions except promethium. The elements from Ce to Sm gave paramagnetic i.e. low-field shifts and the rest diamagnetic i.e. high- field shifts. These results are the opposite of those expected if spin delocalisation occurs via overlap of metal 4 f and oxygen sp hybrid orbitals.The observed shifts are only consistent with a spin polarisation on the ligand opposed to that on the cation. The following model has been suggested30 to account for the experi- mental findings. There is a slight overlap of the sp hybrid orbitals of the oxygen and the 6s orbital of the rare-earth ion giving bonding and antibonding orbitals. Configuration interaction promotes a bonding electron with spin parallel with the spin on the cation into the empty antibonding orbital thus leaving a spin of opposite polarisation on the oxygen atom. For the rare earths the crystal field of the ligands is too small to interfere significantly with the orbital angular momenta of the electrons and strong spin-orbit coupling tends to align the unpaired spins antiparallel to the external magnetic field for the first half of the series and parallel with the magnetic field in the second half of the series.It follows that for the first half of the rare-earth series the spin left on the oxygen will be parallel to the external field and paramagnetic shifts are expected. The opposite situation applies to the second half of the series and will result in dia- magnetic shifts as observed experimentally. In dealing with Eu3+ second-order effects become important and produce a small diamagnetic shift. 29 T. J. Swift and R. E. Connick J. Chem. Phys. 1962,37 307; 1964,41,2553. 30 W. B. Lewis J. A. Jackson J. F. Lemons and H. Taube J. Chem. Phys. 1962,36,695. 31 R. G. Shulman and B. J. Wyluda J. Chem. Phys.1959,30,335. 33 L. E. Orgel J. Chem. Phys. 1959 30 1617. 464 Silver and Luz In addition to the magnitude of contact shifts interesting information can also be obtained from the line-widths of the n.m.r. signals of paramagnetic molecules. Thus measurements of the line-width can be used to give values for electron spin relaxation times in particular for metal ions for which an e.s.r. signal is not observable. From the temperature-dependence of the 1 7 0 linewidths in aqueous solutions of the first series of transition-metal ions Swift and C o n n i ~ k ~ ~ were able to derive values or upper limits for the electron spin relaxation times of these ions. An interesting observation has been made by Jackson Lemons and Taubess on the shift of 1 7 0 in aqueous solutions of Cr2+ containing various anions.They found that the shift of the H,170 resonance becomes less paramagnetic upon addition of certain anions in particular C104-. They interpreted this phenomenon in terms of a complexing mechanism in which the axial water mole- cules of the Cr2+ hydrate are substituted by the anions. In agreement with this model was the observation that the resonance of CP704- is strongly dependent on the concentration of Cr2+ in the solution which indicates direct bonding between the anion and the metal. A similar but stronger effect of Co2+ on the 1 7 0 resonance of Mn04- ion has also been observed and explained in terms of direct bonding between the two species.8 D. Spinspin Couplings.-The comparatively large linewidths of 170 n.m.r. resonances often preclude the observation of splitting due to Ppin-spin coupling with other magnetic nuclei.Even when coupling might be expected to be strong the splitting can be averaged out by chemical exchange or quadrupole relaxa- tion of the neighbouring nucleus. A few spin-spin couplings have however been reported. Several organophosphorus and inorganic phosphorus compounds have been studied by Christ and Diehls and oxygen-phosphorus coupling constants were found to range from 100 to 200 c./sec. The 1 7 0 resonance of water diluted in acetone is a 1 2 1 triplet as would be expected from spin-spin coupling with 2 equivalent protons under conditions of slow proton exchange% (Figure 3). Fig. 3 The 170 spectrum of H p O dissolved in acetone. The structure is due to spin-spin coupling with the two equivalent protons of the water molecule [From J.Reuben A. Tzalmona and D. Samuel Proc. Chem. SOC. 1962 3531 33 J. A. Jackson J. F. Lemons and H. Taube J. Chem. Phys. 1963,38 836. 34 H. A. Christ and P. Diehl Helv. Phys. Acta 1963 63 170. 35 J. Reuben A. Tzalmona and D. Samuel Proc. Chem. Soc. 1962 353. 465 Quarterly Reviews A nice example of spin-spin coupling between oxygen and a nucleus having a spin higher than 4 is provided by the spectrum of l'o-enriched perchlorate ion.36 The spectrum of this ion has four lines of equal intensity with a splitting of 86 c./sec. due to spin-spin interaction with 35Cl. In the ClO ion the chlorine nucleus is at a centre of cubic symmetry and consequently has a long spin- lattice relaxation time. In contrast the 170-labelled chlorate ion CI03- shows only one resonance line because the expected splitting is cancelled by quadrupole relaxation of the chlorine.The 170 spectrum of Xe170F4 shows a triplet with a strong central peak and two symmetrically disposed satellite^.^^ The two satellites are due to spin-spin interaction with 129Xe which has a spin of 8 and a natural abundance of 26.2%. The central peak is from those molecules containing other xenon isotopes. The splitting due to spin-spin interaction with 131Xe (spin E; natural abundance 21.2%) is washed out because of fast quadrupole relaxation of 131Xe. No spin- spin coupling between Xe and 170 was seen in xenic E. Kinetic Measurements.-The n.m.r. method can be employed in two ways in the study of kinetics. (i) For reactions having half-lives of longer than about one minute it is possible to follow changes in the intensity of different resonances as the reaction proceeds; (ii) n.m.r.can also be used to study fast reactions at equilibrium. This method is based upon the fact that in general reactions pro- ceeding at a rate comparable with the chemical shift or line-width of the chemical species involved will affect the shape of the resonance lines. The method has been used extensively to study proton exchange reactions but is of more limited use in the case of 170 since processes involving the transfer of oxygen are not usually fast enough to fall in the range of rates which can affect the n.m.r. line-shapes. Several types of reaction can however be studied by the line-broadening method (ii) some of which will be discussed in this and the subsequent two sections.Considerations of price and availability favour la0 as the isotope of choice in studies of oxygen exchange. In general the use of l 8 0 necessitates the separa- tion and purification of the reacting species before their conversion into a gas suitable for mass-spectroscopic analysis. Often the time taken to separate the reaction components would be of the same order as the exchange half-life. In such cases 170 n.m.r. provides an advantage over conventional tracer methods. An example is the exchange of oxygen between acetone and water. Upon dis- solving acetone in H2170 oxygen exchange occurs and the intensity of the 170 resonance due to the carbonyl group increases with time. The reaction was found to be both general acid and general base catalysed and catalysis constants for several catalysts were obtained.22b In a similar way it has been possible to study oxygen exchange between telluric acid and water.39 When a molecule has two functional groups containing oxygen the study of 36 M.Alei jun. J. Chem. Phys. 1965 43 2904. 37 J. Shamir H. Selig. D. Samuel and J. Reuben J. Amer. Chem. SOC. 1965 87,2359. 38 J. Reuben D. Samuel H. Selig and J. Shamir Proc. Chem. SOC. 1963 270. 39 Z. Luz and I. Pecht J. Amer. Chem. SOC. 1966,88 1152. 466 Silver and Luz oxygen exchange would normally require tedious degradation procedures. 1 7 0 n.m.r. provides a straightforward technique to deal with such cases. A simple example is levulinic acid which has both a carbonyl and a carboxyl group the latter having been found to exchange much more slowly than the former.40 Another example is acetylacetone in water where there are two tautomeric forms in equilibrium 0 0 OH 0 II II I II CHg-C-CH&-CH + CH&=CH-C-CH Each tautomer gives one 1 7 0 resonance the keto-form at ca.-560 p.p.m. and the enol form at ca. -270 p.p.m. Using 1 7 0 labelled water one can follow independently the exchange of the keto and enol forms.*l The use of 170 line-broadening to study fast organic reactions has hardly begun. An interesting example of a fast intramolecular reaction is the rearrange- ment of benzofurazan oxide :42 0 At room temperature this molecule shows two 1 7 0 peaks from the two nonequi- valent oxygen atoms. As the temperature is raised the two peaks broaden and eventually collapse into a single peak. From the temperature at which the peaks coalesce and by comparison with proton magnetic resonance results the rate and activation parameters of the rearrangement reaction could be determined.42 Recently the line-broadening technique has been applied to the reactions of oxyanions.In a solution of periodic acid in water no signal due to periodate ion can be detected. However the water resonance is appreciably broader than that in pure water. This was explained by the existence of a fast oxygen exchange between periodate ion and solvent water. From the temperature-dependence of the line-width rate constants and activation parameters for this reaction could be determined.43 In an aqueous solution of dichromate ion separate resonances are observed for the bridging and terminal oxygen atoms. Upon the addition of strong acid or base the resonances broaden owing to breaking and re-forming of the chromium-oxygen bond.The reaction was studied in acidic solutions by Jackson and Taube,8 and interpreted in terms of the acid-catalysed hydrolysis of the dichromate ion. Figgis Kidd and Nyh01rn~~ examined aqueous solutions of a H. Dahn Proc. Conf. Prep. and Storing Marked Molecules Brussels 1963 Euratom 1964 p. 1303. 41 Z . Luz and B. L. Silver J . Phys. Chem. 1966,70 1328. 42 P. Diehl H. A. Christ and F. B. Mallory Helv. Chim. Acta 1962 45 504. 43 I. Pecht and Z. Luz J. Amer. Chem. SOC. 1965 87,4068. 44 B. N. Figgis R. G. Kidd and R. S. Nyholm Canud. J. Chem. 1965,43 145. 467 Quarterly Re views sodium dichromate to which varying amounts of NaOH were added. In these solutions both chromate and dichromate ions exist and the results were inter- preted in terms of the nucleophilic attack of Cr0,- on Cr,0,2-.F. Proton Transfer in Water.-The process of proton transfer in aqueous solu- tions is of fundamental importance in the chemistry of electrolyte solutions. M e i b ~ o m ~ ~ has shown that the kinetics of these reactions can be studied from the effect of 170 on the proton magnetic resonance of water. In normal water only one proton line is observed and since there is only one distinct magnetic environ- ment no broadening of this line is expected due to proton exchange between different water molecules. If the water contains a certain amount of 170 atoms they will provide additional magnetic environments through spin-spin inter- action between protons and 170 nuclei. In practice even in highly enriched water the proton magnetic resonance spectrum does not display the expected six satellites because the splitting is partially averaged out by the oxygen quadrupole relaxation and by proton exchange.45 However in pure water and neutral aqueous solutions exchange is still not fast enough completely to average out these lines and the excess of width of the proton line over the corresponding width in water unenriched in 1 7 0 provides a measure of the rate of proton exchange between water molecules.Addition of acid or base increases the rate of proton exchange via the reactions kl H,O + H30++ H30+ + H20 k2 H2O + OH-+ OH- 3- H2O which results in a narrowing of the proton line. From the pH-dependence of the proton line-width M e i b ~ o m ~ ~ derived values for k and k of 10.6 x lo9 l.mole-lsec.-' and 3.8 x lo9 l.mole-lsec.-l respectively.The activation para- meters for kl and k2 have also been determined by the same rneth0d.4~9~~ The technique of using 170-enriched water has been applied to studies of proton transfers between solvent water and solute molecules (trimethylam- monium It is particularly useful in cases where proton exchange is too fast to give separate proton signals for water and solute and has been used to study proton exchange in aqueous solutions of acetate phosphate and phenolate b~ffers.4~ Proton exchange is also expected to affect the 170 n.m.r. line-~hape.4~ At low rates of proton exchange the 1 7 0 resonance of water should be a triplet owing to spin-spin interaction with the two protons. However proton exchange partially averages this splitting and in pure water only one broad 170 line is observed 45 S.Meiboom J. Chem. Phys. 1961 34 375. 46 A. Loewenstein and A. Szoke J. Amer. Chem. Soc. 1962 84 1151. 47 Z. Luz and S. Meiboom J . Amer. Chem. SOC. 1964 86,4768. 48 Z. Luz and S. Meiboom J . Chem. Phys. 1963,39 366. 49 Z. Luz and S. Meiboom J. Amer. Chem. Soc. 1963,85,3923; 1964,86,4764; 1964,86,4766. 468 Silver and Luz although at low temperature the line-shape indicates the presence of some structure.50 This line narrows upon addition of acids or bases owing to the proton exchange reactions discussed above. In principle the changes in the 1 7 0 line-width could be used to derive values for proton transfer reactions. Direct measurements on the 170 resonance do not however have an advantage over proton resonance studies since the range of broadening of the oxygen resonance is limited.On the other hand the broadening of the proton resonance can within limits be fixed by the 170 concentration used and accurate line-width measurements can be done especially by the spin echo technique. G. Hydration of Metal Ions.-The 1 7 0 nucleus provides a unique chemical probe for studying the hydration of ions in aqueous solution in particular solvation numbers and the lifetime of water molecules in the solvation shell. In some cases the solvation number of the ion can be determined directly. The interpretation of the nuclear magnetic resonance results differ for paramagnetic and diamagnetic ions. In paramagnetic ions the interaction with the electron spin often results in large shifts and line-broadening.In contrast with the situation for paramagnetic ions the 1 7 0 resonance due to the solvation shell of diamagnetic ions is not appreciably broadened and is almost unshifted from the position of the resonance of the bulk solvent. Paramagnetic ions will be con- sidered first. The 1 7 0 resonance of water in dilute solutions of the bivalent ions of the first transition-metal series is broadened and shifted relative to the resonance of pure water.51 This effect can be u n d e r s t ~ o d ~ ~ ~ ~ ~ by considering a model in which there are two magnetic environments possible for the solvent molecules (i) free solvent and (ii) the co-ordination shell of the paramagnetic ion. In the latter environment the scalar and dipolar interactions between the 170 nucleus and the unpaired spin of the ion cause rapid relaxation and large contact shifts of the 1 7 0 resonance.In principle it would be expected that two separate 1 7 0 peaks would be observed but chemical exchange of solvent between the two environ- ments broadens the peaks. Swift and C ~ n n i c k ~ ~ have treated the case in which the number of nuclei in the bulk far exceeds the number in the co-ordination shell (dilute solutions) and derived expressions relating the line-width and shifts of the bulk resonance to the exchange rate. From the temperature-depend- ence of the 170 resonance in aqueous solutions of the ions Mn2+ Fe2+ Co2+ Ni2+ and Cu2+ they were able to derive values for the lifetime of water molecules in the first co-ordination sphere of the cations and the activation energy for the hydration-dehydration r e a c t i ~ n .~ ~ ~ From the rate constants found29 for the hydration of Ni2+ and Co2+ it could be predicted that separate peaks for the bulk and solvation shell molecules would be observable below room tempera- 50 J. A. Glasel Proc. Nut. Acad. Sci. 1966 55 479. 51 (a) R. E. Connick and R. E. Poulson J . Chem. Phys. 1959,30,759; (6) R. E. Connick and E. D. Stover J . Phys. Chem. 1961,65,2075. 5e Z. Luz and S. Meiboom J . Chem. Phys. 1964,40,2686. 53 See also W. B. Lewis M. Alie and L. 0. Morgan J. Chem. Phys. 1966,44,2409. 469 Quarterly Reviews ture and it was indeed possible to observe two separate 170 peaks in both case^.^^,^^ In aqueous solutions of diamagnetic salts one would expect to observe separ- ate 1 7 0 resonances due to the bulk solvent and the hydration shell of the cations provided the exchange rate between the two sites is small compared with the chemical shift between the two resonances and of course provided the two lines do not overlap.In fact in the spectrum of an aqueous solution of the ion [(NH3),CoOH2I3+ a separate I7O signal appears owing to water bonded in the complex.54 A number of other salts fail to show a separate signal from the hydration shell in aqueous solution although it is known that for some of them the hydration shell does not exchange rapidly with the bulk solvent. Jackson Lemons and T a ~ b e ~ ~ speculated that the signal of the cation hydration shell falls below that of the bulk solvent and had the ingenious idea of shifting the latter signal by the addition of a small amount of Co2+ ion. This ion as explained above will shift the resonance line of free bulk water towards low field but will not affect the position of the hydration shell signal provided this shell does not exchange water molecules rapidly with the bulk.By use of this technique it was possible to observe separate resonances at room temperature for the solva- tion shells of APf Be2+ and Ga3+. No separate signal was found for H+ Li+ Mg2+ Ba2+ Sn2+ Hg2+ or Bi3+ indicating that the solvation shells of these ions exchange quickly with the solvent. Where a comparison can be made these findings are consistent with other evidence on the exchange rates of the solvation shell. If highly enriched water is used the study of the solvation shell can be put on a more quantitative basis i.e. hydration number and the rate of exchange of the hydration shell can be determined One method is to determine the relative amounts of bound and free water from the integrated intensities of the n.m.r.signals corresponding to the two types of water. From the known concentration of electrolyte the hydration number of the cation can be calculated. This method has been applied to AP+ and Be2+ by Connick and Fiat,55 using water enriched to 11.48% in H2’O. Average values were found of 5.9 and 4.2 respectively for the hydration number of these ions. Recently these authors have also determined the hydration number of Gas which was found to be The shift caused by a paramagnetic ion having a labile solvation shell is inversely proportional to the amount of free water in the solution. Thus if the solution also contains a cation that ties up water in its solvation shell the amount of free water decreases and its shift for a given amount of paramagnetic ion will accordingly increase.54 This fact has been exploited by Alei and Jacksonss to study the hydration number of A13f Be2+ and C?+.These authors used Dy”+ to shift the water line and the values obtained for the hydration numbers were 54 J. A. Jackson J. F. Lemons and H. Taube J. Chem. Phys. 1960,32 553. 55 (a) R. E. Connick and D. Fiat J. Chem. Phys. 1963,39,1349; (6) D. Fiat and R. E. Connick J. Amer. Chem. SOC. 1966,88,4754; T. J. Swift 0. G. Fritz and T. A. Stephenson J. Chem. Phys. 1967 46,406. 56 M. Alei jun. and J. A. Jackson J. Chem. Phys. 1964,41 3402. 470 Silver and Luz 5.9 3.8 and 6.8 respectively. Considering the experimental error in the measure- ments the last figure does not agree with the value of 6.0 found by the isotopic dilution method.57 Alei and Jackson suggested that the discrepancy can be explained by the existence of a second solvation shell which does not exchange very quickly with the bulk.The technique of shifting the water line is general1y.applicable to the case of a signal suspected of lying beneath the solvent water line. For example in aqueous solution chloral is completely hydrated but fails to give a 1 7 0 signal. Addition of Dy3+ to the solution shifts the water line and reveals the resonance due to the gem-diol oxygen atoms of chloral hydrate.22Q 3 Electron Spin Resonance A. Hyperfine Splitting.-Electron spin resonance spectra of free radicals exhibit hyperfine structure caused by the Fermi contact interaction.The 1 7 0 nucleus has a spin quantum number of and will therefore split each e.s.r. resonance line into six components. A very simple example shown in Figure 4 is the spectrum Fig. 4 The e.s.r. spectrum of chloranil anion enriched in I7O. The strong central line is due to the unlabelled compound and the six satellites to the chloranil anion containing one I7O atom [From M. Broze Z. Lu and B. L. Silver J. Chem. Phys. 1967,46,4891] 57 J. P. Hunt and H. Taube J. Chem. Phys. 1951,19,603. 471 Quarterly Reviews of 170-labelled chloranil anion. A collection of splitting constants due to 170 in some free radicals is given in the Table. I7O Hyperfine coupling constants ao in free radicals Radical Solvent * 1 ,4-Benzosemiquinonea DMF 1,4-Naphthasemiquinone" DMF 9,1 0-Anthrasemiquinone" DMF 2,5-Dichloro-1 ,4-benzosemiquinoneo DMF Chloranil anion" DMF 2,5-Diphenyl-l ,4-benzosemiquinonea DMF 2,5-Dioxo-l,4-benzo~emiquinone~ H,O 3,6-DichIoro-2,5-dioxo- 2,4,6-Triphenylphenoxyl N.S.2,4,6-Tri-t-b~tylphenoxyl~ N.S. Nitrobenzene aniond CH,CN Di-s- bu t ylni t ric oxide C6H6 Fremy's salt H2O FOOf CF,(liq) 1,4-benzo~emiquinone~ H2O I a0 I (GI 9.53 8.58 7.54 9.28 8.98 9.07 4-57 4.79 9.7 10.23 8.66 19.71 20.6 22.17 14.50 Spin density on oxygen 0.24778 0.2224g 0.1964Q 0.2356g 0.22440 0.151 50 - 0.1 51 68 0*174c 0-195h 0~651~ 0.65 1 ' - *DMF = Dimethylformamide; N.S. = Not specified. a Ref. 58; Ref. 1 1 ; CRef. 10; d Ref. 12; e Ref. 9; f Ref. 15; Calculated50 by Huckel mole- cular orbital method using a0 = OL + 0.408 /3c0 = 1.388; hRef. 59b; *G. Bertier H.Lemaire A. Rassat and A. Veillard Theoret. Chim. Acta 1965 3 17. Hyperfine structure has proved to be a rich source of information on the electronic structure of free radicals. The basis for this information is the well- established correlation that has been shown to exist59 between spin densities and hyperhe interaction constants of lH 14N and 13C. A similar correlation between spin-density distribution and a0 has been shown to hold for a0 in p-semi- quinones :58 where PO and pc are thepn spin densities on the oxygen and neighbouring carbon atom respectively (as calculated by the Huckel molecular orbital method with the parameters given in footnote g of the Table). Although the value of the coefficient of pc is not very certain there can be no doubt that in this case the major contribution to a0 arises from the spin density on the oxygen atom itself.The signs of hyperfine coupling constants are of great theoretical importance. From the analysis of the line-widths of the e.s.r. spectra it was possible to show a0 = -40.5~0 + 6 . 4 ~ ~ 58 M. Broza Z. Luz and B. L. Silver J. Chem. Phys. 1967,46,4891. 59 (a) H. M. McConnell J . Chem. Phys. 1956,24 764; (b) P. H. Reiger and G. K. Fraenkel J. Chem. Phys. 1963,39 609; (c) M. Karplus and G. K. Fraenke1,J. Chem. Phys. 1961,35 1312. 472 Silver and Luz that the sign of (10 is the same as the (negative) sign of 70 the gyromagnetic ratio of oxygen in both benzosemiquinonel& and Fremy’s salt.14 In most other cases there are line-width variations which indicate a similar sign for a 3 0 . ~ ~ s ~ ~ s The physical significance of this is that the net spin polarisation at the 1 7 0 nucleus is parallel with the total spin of the radical.B. Electron Spin Resonance Line-widths.-E.s.r. spectra often show variation in line-width between different hyperfine components as seen for example for chloranil anion in Figure 4. The general theory for e.s.r. line-widths60 predicts marked effects on the line-widths of hyperfine components due to nuclei having a high spin density and high magnetic moment. Since 170 has a relatively high spin quantum number and often carries high spin densities it furnishes an excellent probe with which to test the theory. Good agreement with the theory of line-widths was in fact found for the 170 satellites of p-benzosemiq~inone~~~ and Fremy’s salt,14 in the spectra of which large line-width variations are seen. 6o J. H. Freed and G. K. Fraenkel J. Chem. Phys. 1963,39 326. 473

ISSN:0009-2681    

DOI:10.1039/QR9672100458

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

The Elementary Particles By B. H. Bransden THE UNIVERSITY DURHAM 1 Introduction Just twenty years ago in 1947 the list of sub-atomic or 'elementary' particles that were known or believed to exist was quite short (Table 1) and of deceptive simplicity. The two heavy particles the neutron and proton were known to be the constituents of atomic nuclei interacting together with a force that was very strong compared with the well-understood electromagnetic interactions that are responsible for non-nuclear atomic and molecular phenomena. The nuclear force also differs from the electromagnetic force in the range of interaction. The electromagnetic force is of long range but the neutrons and protons only interact at distances closer than 10-13cm.; that is the nuclear force is of short range. The only other strongly interacting particles called the n mesons were considered to be the quanta of the field producing the nuclear forces just as photons are the quanta of the electromagnetic field.It can be shown that the range of the interaction produced by the exchange of a particle of mass M is R = fi/Mc where is h/277 and h is Planck's constant ; c is the velocity of light. For then meson mass R = ca. lO-l3crn. as observed. The long-range character of the Coulomb force is consistent with this expression if it is remembered that the exchanged photon has zero mass. The strongly interacting particles which are collectively called the hadrons can be distinguished sharply from the lighter particles or leptons comprising the electron e- the positron e+ the pf and Table 1 The elementaryparticles known in 1947 Particle" Mass I b Yb JPc Weak- ( Mev/c2) decay Photon y Neutrino vd Electron e- p-Meson p- n-Meson no n-Meson n+ Proton p Neutron n 0 0 0.51 105.7 135 140 938.2 939.2 Stable Stable Stable eTv+tr Y+YC p++v Stable p+e-+G Mean Antiparticle" Iyetime (sec.) - Photon y - Antineutrino rd - Positron e+ p-Meson p+ 10-le w-Meson no +Meson n- - Antiproton d lo3 Antineutron Ed a The superscript f denotes that the particles possess charge f ; The isotopic spin I the third component 13 and the hypercharge Yare defined in the text.Y and I3 for an antiparticle are of equal magnitude to but of opposite sign from those quantities for the corresponding particle; C J is the internal angular momentum or spin measured in units of h. P = + for even and - for odd parity particles; These particles had not been discovered in 1947 but were believed to exist.They have been detected subsequently; e This decay is electromagnetic. 474 Bransden p- mesons and the neutrino tl because these do not share in the strong nuclear interaction. Of the particles shown in Table 1 only the photon the electron the proton and the neutrino (and the corresponding antiparticles) are completely stable. The remainder ultimately decay into the stable particles. For example a free neutron decays into a proton an electron and a neutrino with a mean life of 17 minutes while a p- meson decays into an electron a neutrino and an antineutrino with a mean life of lod6 sec. The spontaneous decay times of these particles are all longer than 10-la sec. which may be compared with the time required for a typical strong interaction which is of the order see.so that as far as the strong interactions alone are concerned all the particles of Table 1 can be considered to be stable. Some force must be responsible for these slow decays and this force must be extremely weak compared with the strong inter- actions in view of the long life-times of the particles on the nuclear time scale. We shall call this force the weak nuclear interaction. One further interaction is known and is very familiar to us that due to gravity but this is many times weaker than the weak nuclear interaction and plays no important role in ele- mentary-particle interactions (see Table 2). The importance of gravitational Table 2 The interactions Strengths Range Strong nuclear interaction 1-10 10-13cm.Weak nuclear interaction 10-14cm. Electromagnetic interaction e2/fic = 1/137 Long-range inverse-square law Gravitational interaction Long-range inverse-square law phenomena in the macroscopic world arises from its long-range inverse-square law in contrast to both the weak and strong interactions which vanish when the particles concerned are separated by distances much in excess of lO-l3cm. In 1927 P. A. M. Dirac was able to show by combining quantum mechanics with the theory of special relativity that for every particle another should exist of the same mass and angular momentum but of opposite charge known as its anti-particle. In particular he predicted that the electron should possess a positive counterpart (the positron) which was subsequently discovered by C. D. Anderson in 1933.On this basis it was confidently believed that the proton neutron and neutrino would possess corresponding antiparticles and indeed negatively charged antiprotons were observed for the first time in the summer of 1954. Table 1 illustrates further interesting features. It can be seen that the neutron and proton have nearly the same mass and also that the no meson has nearly the mass of the T+ and T- mesons. From the study of the atomic nuclei it had be- come apparent that the neutron and proton both took part in the nuclear inter- actions in the same way and this suggested that the neutron and proton were two charge states of the same particle called the nucleon. The small mass difference between the neutron and the proton was attributed to the different action of the electromagnetic field on the charged proton and on the neutral 475 Quarterly Reviews neutron.In the same way the threen mesons could be considered to be different states of one particle and again the difference in mass between then* mesons and the 7ro meson could be attributed to electromagnetic effects. These remarks will be elaborated below. 2 Discovery of the ‘Strange Particles’ The relatively simple picture presented by Table 1 was soon disturbed by the discovery of further particles known as ‘strange particles’ in cosmic radiation by C. Butler and G. D. Rochester. The characteristics of the new particles were finally determined a few years after the initial discovery when the large accelera- tors first at Brookhaven (1952) and later at Berkeley (1954) and at C.E.R.N. Geneva (1959) came into operation.These machines accelerate protons by electric and magnetic fields to energies of up 30 Gev (1 Gev = 1000 MeV; 1 Mev = 10gev). The high-energy proton beam can be directed at a target so that nuclear interactions take place between the incident protons and the nuclei of the target atoms. Just as an electron deflected by the Coulomb field of a charged particle will radiate light (photons) so during the scattering of protons by the strong interaction particles that take part in the strong interaction may be produced. For example the result of a collision between two protons might be a proton a neutron and an+ meson p + p-+ p + +n+ Another possible reaction if sufficient energy were available wouId be the creation of an antiproton 6) by P + P - + P + P + P + P (2) Apart from the primary proton beam by focusing of the products of these collisions secondary beams of various particles can be produced.These second- ary beams can also be made to interact with a target and it is by examining the products of such collisions that most of the new particles have been discovered. So far some thirty particles have been discovered that are stable under the strong interaction but which decay slowly under the weak interaction. The properties of these particles are summarised in Table 3. The new particles are also found to be grouped into multiplets each member of which possesses the same internal angular momentum or ‘spin’ and approximately the same mass but differ in charge. Those with a mass near 500 Mev/c2 and ofzero spin are called the K mesons K+ and KO and the corresponding antiparticles are the K and K O mesons while a further meson the 7 is found with a mass of 548 Mev/c2 also having zero spin.The particles in Table 3 with mass greater than the nucleon are known as hyperons and comprise the / 1 O at 11 15 Mev/c2 the Z-t 2- and CO at 1190 Mev/c2 the E- and So at 1320 Mev/c2 and the & at 1675 Mev/c2. Each of these particles possesses a corresponding antiparticle. The particles stable under the strong interactions can generally travel a sufficient distance to be detected directly in a photographic emulsion or in a 476 Bransden Table 3 The particles stable under the strong interactions discovered since 1947 ParticleQ Mass la Isb Yb Jpc PrincQle ( Mev/c2) weak-decay Mesons (B = 0) K+ KO 497.7 Q -9 1 0- r1O 548.7 0 0 0 0- y+ye Baryons (B = 1) A0 1115 0 0 0 &+ N+n C-t 1189 1 1 0 $+ N+n co 1192 1 O O Q+ Ao+ye 2- 1197 1 -1 0 &f N+T EQ 1314 8 8 -1 &+ AO+nO 1321 & -4 -1 &+ AO+n- a- n a- 1675 See footnotes to Table 1.Mean Anti- life-time particlea - 10-8 K- - lo-* KO ? TIo 10-10 A0 1 0-l0 2- 10-14 co 10-10 c+ - 10-10 a+ bubble chamber before decaying under the weak interaction but in the last few years it has been realised that other particles exist that are not stable under the strong interaction and which decay too rapidly to be observed directly as they possess mean life-times of the order sec. For instance it is observed that ifn mesons are scattered from protons then at certain well-defined energies a great increase in the flux of scattered particles takes place. This may be inter- preted by supposing that these ‘resonance’ energies are the energies at which then meson and proton combine to form an unstable compound particle called the N* that subsequently decays rapidly into a proton andn meson again.p + n + N * - t p + v (3) Resonance phenomena of this kind have been seen and studied in scattering processes between atoms or ions and between nuclei corresponding to the formation of excited molecular or nuclear states respectively. In experiments of a different type the resonant particle may be produced along with other particles and its presence can be detected by examining the correlations among the end products of the reaction. The p meson which decays into twon mesons can be studied in this way through the reaction + p + p + p - + n + n + p (4) In this general way a great number of particles unstable under the strong inter- actions have been discovered (some of these are shown in Table 4).It is now quite clear that the name ‘elementary’ is undeserved and in fact that we are 477 Quarterly Reviews Table 4 Some of the Ikhter particles unstable under the strong interactions Particle" P K* X0 f 0 + N*q(1236) Y * 1 385) N*%( 151 8) Y* o( 1 520) s*( 1530) N*,( 1688) Mass (Mev/c2) I b Yb JPC Mesons (B = 0) 769 1 0 1- 783 0 0 1- 1- 0- 959 0 0 1020 0 0 0- 1253 0 0 2+ 891 4 1 Baryons (B = 1) 3 9- 3+ 1236 2 1 1385 1 0 2 1518 4 1 3- 1520 0 0 ;- 1530 4 -1 3+ - 1 s_+ 1688 1 Principle strong decays a The (21 + 1) charge states of each multiplet have not been listed separately. A complete list of all particles known is given by A. H. Rosenfeld A.B. Gallieri W. H. Barkas P. L. Bastien J. Kirz and M. Ross Rev. Mod. Phys. 1965 37 633. b ~ c See footnotes (b) and (c) of Table 1 looking at a mass spectrum of particles comparable with the energy spectrum of atoms or molecules and it is quite possible that there is an infinite number of such excited states. As in the case of the atomic spectrum the full description of the system of particles requires two steps first the states must be classified by assigning appropriate quantum numbers and secondly the underlying dynamical explanation must be sought. Before considering the classification of the particles we note that with one exception the list of leptons in Table 1 has not increased. This is consistent with the idea that the mass spectrum is a consequence of the strong interactions in which the leptons play no part.The single exception is that it is not known from a study of neutrino interactions that them are two kinds of neutrino one asso- ciated with the electron and one with the p meson. It is still a complete puzzle why the p meson which has the same properties as the electron but is of grsater mass exists at all. 3 The Classification of Elementary Particles All reactions whether caused by the electromagnetic the weak or the strong interactions obey the fundamental laws of conservation of energy momentum angular momentum and of electrical charge. Both in classical and quantum mechanics these laws (with the exception of conservation of charge) arise from certain symmetries of space and time. In fact conservation of energy is a neces- sary consequence of the application of the invariance of laws of physics to time displacements conservation of momentum follows from invariance under spatial displacements and conservation of angular momentum corresponds to invari- 478 Bransden ance under rotations of the co-ordinate systems.It can be chtcked experimentally that reactions between elementary particles conserve angular momentum if each particle is assigned a definite intrinsic angular momentum or spin J. According to the general principles of quantum mechanics all angular momenta must be of the form nh where n is an integer or half-integer. In quantum theory the wave function of a particle must either be even or odd under reflection of the co- ordinate system in which the position vector of the particle r becomes -r.States for which the wave function is even are said to be of parity P = + 1 and those in which the wave function is odd of parity P = - 1. If a system is com- posed of several parts the total parity is found by multiplying together the parities of the separate parts. It can be shown that the total parity of a molecular system bound by coulomb forces is conserved. Reactions involving the strong interactions (but not the weak) are also found to conserve parity but only if it is assumed that each particle has a definite internal parity associated with it in addition to the ordinary parity of the wave function familiar in atomic and molecular physics. For example the nucleon is found to have panty P = + 1 while the anti-nucleon and the 7~ mesons have parity P = - 1. So far we have associated three numbers J P and Q (the charge) with each particle.Of these the conservation law of charge is particularly simple; each particle in a reaction has a definite charge whether positive or negative and in a reaction such as the algebraic sum of the charges on each side of the equation must be the same. Are there any further such ‘additive’ quantum numbers? A study of all known reactions whether strong or weak shows that if we define the baryon number B of each baryon (the collective name given to nucleons hyperons and other heavier particles possessing half-odd integer spin in units of h; J = # gh,. . .) to be +1 and that of each antibaryon to be - 1 and that of all other particles to be zero then the total baryon number is conservtd for example the initial and final systems in eqns.(l) (3) (4) and (5) have B = 1 while in eqn. (2) B = 2. A reaction such as where B = 1 on the left-hand side and B = 2 on the right-hand side is never observed. A less obvious quantity conserved in the strong interactions only is the hyperchange Y which is defined as twice the average charge (in units of e the charge on the electron) in any charge multiplet. The average charge of the neutron-proton multiplet is Be so the nucleon multiplet is said to have hyper- charge Y = 1. The K+ KO multiplet similarly has Y = 1 but the / 1 O and the C+,Z:-,co hyperons possess Y = 0 and the .r” Eo multiplet Y= - 1. Con- servation of Y was originally introduced to explain the puzzling fact that hyperons were always produced in association with K mesons in 7~ meson- 479 Quarterly Reviews proton reactions such as that of eqn.(5). As a consequence of the definition the hypercharge of an antiparticle is always equal in magnitude but opposite in sign to that of the corresponding particle. We have suggested that as far as the strong interactions are concerned all the particles in a charge multiplet (in which each particle possessing the same J P and Yare of approximately the same mass) are just different states of the same particle the different states being labelled by the charge Q. If such a multi- plet contains (21 + 1) particles then I is called the isotopic spin and is also a characterising quantum number (non-additive) associated with each member of the multiplet. To summarise each particle can be classified by specifying J P B Y Q and I and of these B Y and Q are additive quantum numbers while J adds like an ordinary angular momentum and P obeys a multiplicative rule.The significance and addition rule of the isotopic spin I will now be discussed. An alternative to Y often used is the ‘strangeness’ S defined by S = B - Y. 4 Internal Symmetry Schemes A. The Isotopic Spin.-Studies of nuclear forces in the early 1930’s suggested that if the Coulomb attraction between protons were ignored the proton-proton neutron-proton and neutron-neutron forces were identical in corresponding states. W. A Heisenberg showed that this fact could be expressed by a con- servation law or symmetry principle. It is supposed that there is an abstract three-dimensional space called isotopic space and that the nucleons are asso- ciated in this space with a vector quantity the isotopic spin vector with com- ponents 11 I, and I,.It is further postulated that 11 12 and I obey the same algebraic relations (the commutation relations) as do the components of the ordinary angular momentum vector in ordinary space. Physical processes are now assumed to be invariant under rotations in this abstract three-dimenrional space and this implies that systems of nucleons are described by wave-functions which correspond to definite values of the magnitude of the isotopic spin vector and of its third component I, just as invariance to rotation in ordinary space implies that systems must possess a definite value of the total angular momentum and a definite value of the component of angular momentum along a given direction. For a given value of total isotopic spin I [which must be a multiple of (*)I the multiplicity of different states with different 1 is (21 + l) so that as the nucleon exists in two states it can be assigned the value I = 4 with I3 = 8 for the proton and I = -8 for the neutron.For systems of nucleons the allowed states are found by adding the isotopic spins by use of the same rules by which angular momentum is added in quantum mechanics. The invariance condition then requires that the total isotopic spin (and the third component) found in this way is conserved in any nuclear reaction. It is now important to notice a fundamental difference between conservation of angular momentum and of isotopic spin. The former law is always exactly satisfied in all reactions whatever the mechanism as it arises from the structure of space itself.In the latter case the law is satisfied by the strong nuclear inter- 480 Bransden action only to the extent that the Coulomb forces between protons can be neglected. It is more accurate for systems with small numbers of protons and less accurate for systems with large numbers. Nevertheless because the Coulomb forces are so small compared with the nuclear forces the idealisation in which isotopic spin is exactly conserved in strong interactions (and in which the mass differences between particles within multiplets is ignored) is most useful. The extension of this concept to the other elementary particles has been extra- ordinarily successful. Then mesons form a triplet with I = 1 the K+ KO mesons a doublet with I = 4 and so on (see Tables). The value of I is not independent of the quantum numbers that we have previously introduced; in fact I = Q -QY.The mass differences within any multiplet are of a magnitude that is consistent with their being of electromagnetic origin breaking the complete isotopic spin symmetry slightly. With the assignments shown in the Tables the hypothesis that the strong interactions conserve isotopic spin may be tested by noting that all processes that violate this principle occur extremely slowly on the nuclear time scale and can be explained as the result of either the weak or the electromagnetic interactions. This hypothesis also predicts explicit relationships between reaction rates for different members of a multiplet which after correc- tion for Coulomb forces appear to be accurately satisfied. B. Higher Symmetry Schemes.-The successfu1 grouping into isotopic spin multiplets posed the question as to whether further groupings could be achieved.The algebra satisfied by the isotopic spin components (and the ordinary angular momentum components) allows one of these components I, to be an additive quantum number. There is a further additive quantum number the hypercharge Y involved in the strong interactions so a natural extension is to examine algebras for which both I3 and Y can be additive quantum numbers. The most successful algebra achieving this is known as SU, and contains in addition to 11 I, J3 five further independent quantities one of which can be identified with Y. We have seen that the isotopic spin algebra (known as SU,) contains multiplets containing (21 + 1) particles where I is a multiple of ($) but the corresponding SU multiplets have not quite such a simple structure.The lowest multiplets possible are those containing 1 3 6 8 10 15 and 27 different states. If these are to be associated with particles each particle in a multiplet must have the same values of J P and B. In addition the isotopic spin multiplets contained within each SU multiplet are not arbitrary but are determined. For example the eight-dimensional multiplet or octet contains an isotopic spin singlet with I = 0 Y = 0 a doublet with I = & Y = 0 a doublet with Y = 1 I = Q and a triplet with I = 1 Y = 0 The eight baryons f l o ; n p; B- 8 O ; and C+ C- co fall exactly into this pattern (and the corresponding antiparticles can also be fitted just as well in an octet pattern). These states are shown in Figure 1 plotted in a ‘weight’ diagram against the allowed values of I3 and Y.If the strong inter- actions were exactly invariant under transformations in the abstract SU, space then each particle in a multiplet would have exactly the same mass apart from the small splitting among the isotopic spin multiplets caused by the electro- 481 Quarterly Reviews Y +I 0 -I n D 11 ' I- 0 I I I I I - -I -t - 0 ++ 'I 1 1 Fig. 1 The baryon SUI octet. magnetic interaction and also many relations between the reaction rates for collisions between the various particles could be predicted. Examinations of the baryon masses show that they lie between 938 and 1132 Mev/c2 so that some interaction of medium strength must be present which produces a sizable departure from strict SU, invariance.By consideration of the simplest pos- sibilities for the SU properties of this medium-strength interaction a mass formula giving the masses in terms of I and Y has been developed. The most convincing evidence for this scheme came in 1962. At that time nine baryon resonances with J = and P = +1 were known consisting of an isotopic spin quartet I = with Y = 1 a triplet with I = 1 and Y = 0 and doublet with I = 9 Y = 1 (Figure 2). These would fit into a ten-dimensional multiplet if one further particle could be found with Y = -2 and I = 0 and on the basis of the mass formula this was predicted to have a mass of about 1680 Mev/c2. The subsequent discovery of this particle G- with a mass of 1676 Mev/c2 must be considered to be a triumph of the theory. The mesons also fit neatly into the SU multiplets.The nine mesons with J = 0 and P = - 1 divide into an SU octet (Figure 3) and an SU singlet (the single particle is Xo of Table 4) and so do the nine mesons with J = 1 and P = -1 (Figure 4); the physical w 4 particles appear to be mixtures of the 482 Bransden MC3.s 123b 1385 + I 0 -I o lb7b Fig. 2 The SU3 decouplet of baryon resonances. K O K+ I - t I I I 1 1 0 +f +I 5 - - I -1 Fig. 3 The SU octet of mesons stable under the strong interactions. 483 QuarterZy Reviews tI 0 -I I * I I = 0 L I= 2. I I I I I + 0 t+ +I 1 3 -L -I - 2 Fig. 4 The SU3 octet of mesons with J = 1. singlet state with Y = 0 and Z = 0 and the octet state with the same quantum numbers. Mathematically the wave functions for particles of any isotopic spin can be formed by combining the wave functions of a suitable number of doublets with isotopic spin I = Q and Z3 = *&.In the same way all SU wave functions can be built by combining triplets. The basic SU triplets are composed of an isotopic doublet together with an isotopic spin singlet. The rules for adding SU states show that an octet is to be built from three basic triplets and because B = 1 for the baryons and the baryon charges are integral in e this requires that the quantum numbers of the basic triplet states gi must be q1 with B = 4 Y = 4 Q = #e; q2 with B = 6 Y = 5 Q = -&e; and 4 with B = Q Y = -8 Q = - Qe. A daring speculation of M. Gell-Mann was that one way of account- ing for the observed symmetries would be to postulate the existence of three real particles called the quarks,* qi of spin J = Q and P = +1 with the basic SU triplet quantum numbers q.Antiquarks qi would also exist with J = 8 P = - 1 and with the opposite signs of B Y and Q. All elementary particles would then *‘Three quarks for Muster Mark! Sure he hasn’t got much of a bark’; James Joyce ‘Finnigans Wake’ Faber and Faber 1939 p. 383. 484 Bransden be composed of quarks and antiquarks bound together by some basic interaction. The baryon octet would exist of three bound quarks the meson octets would consist of a bound quark and an antiquark and so on. Further groupings of particles result from considering the spin and orbital angular momentum of parts of the quark wave functions. Similar results can be obtained by consider- ing higher algebras than SU, in particular by considering SU which combines ordinary spin with SU,.From three quarks with zero orbital angular momentum each of spin 4 states can be built with total spin either of J = or J = $. The spin J = Q states form an SU octet and the J = states an SU decouplet so that the existence of the baryon octet and the decouplet of spin J = resonances finds an explanation. The quark-antiquark states with zero orbital angular momentum give rise to states with J = 1 or J = 0; in each case the SU part of the wave function corresponds to an octet plus a singlet so the existence of nine mesons with J = 0 and nine with J = 1 is again explained. Many other particles must be connected with excited states of non-zero orbital angular momentum of the quark systems but this identification is not yet complete. The mass differences between the various isotopic spin multiplets within an SU multiplet can also be explained on this model.It is only necessary to suppose that the quark qs possesses a different mass from the quarks q1 and q2 to account for the mass splitting within octets. By considering basic interactions between quarks that depend on spin or on the SU operators the different masses of the difftrent octets and decouplets may perhaps also be explained. 5 The Search for Quarks If quarks have a real existence as the basic building bricks out of which the elementary particles are constructed then it should be possible to detect them because in contrast to all other particles they possess a charge which is a fraction of the charge on the electron. It seems certain that quarks are not among the particles produced by the giant accelerators and this implies that they must be extremely heavy with a mass greater than 4 Gev/c2.Searches for quarks in cosmic rays have proved to be negative so far and although these searches are continuing it is natural to consider other physicochemical methods for quark detection. Consider the fate of a quark produced by very high-energy cosmic rays in our atmosphere; cosmic ray events are observed in which the primary particle responsible possessed an energy of thousands of Gev so that quarks of great mass should be produced from time to time. The quarks of charge -Qe will ultimately be captured by atoms into a Bohr orbit. Because of the great mass of the quark this orbit will be inside the nucleus of the atom. The nucleus will then possess a fractional charge (in units of e) and will remain so for all time because any mechanism such as ,8 decay which changes the charge of a nucleus can only change it in units of e.The chemical properties of an atom with a frac- tionally charged nucleus are not yet known in detail; but they will not be very different from the normal atom because the number of electrons attached to the ‘quarked’ atom will be the same as for a normal atom. The positively charged quarks must eventually capture an electron and subse- 485 Quarterly Reviews quently the electron-quark combination will behave like a highly reactive electro- positive hydrogen atom. For a quark with charge ++e this hydrogen-like atom will possess an ionisation potential of 1.51 ev and it could exist in water as a hydrated ion and would be evaporated as a positive ion.For the quark with charge +ge the ionisation potential is 6.04 ev and this could be evaporated mainly in association with an electron or negative ion and would therefore be manifested as a negatively charged object. In the very interesting experimental work of Chupka et aL,l which we closely follow in this section large samples of various materials were passed in gaseous form through an electric field strong enough to detach fractionally charged particles and to concentrate then on a platinum filament. The concentrated specimen could either be examined with a mass spectrometer or alternatively connected to an apparatus that measured the concentration of negatively charged particles by accelerating them through a field of 15 kv on to an electron multiplier.In this way when the ions were evaporated by heating the filament the concentration could be measured as a function of time. This concentration decreases cxponentially the decay constant depending on the species. All known negative ions are evaporated in the form of neutral atoms so the accderating field of the detector has no effect on the rate of evaporation and if the field is reversed and then restored the concentration of negative ions continues to follow the same exponential curve (see Figure 5). Positive ions evaporate in the positively charged state so that evaporation ceases while the field is reversed. An example of this behaviour is also shown in Figure 5. As quarked atoms never can be neutralised they will behave like positive ions in this respect and should be easily detected.The production rate of quarks by cosmic rays is expected to be inversely proportional to their mass so a measurement of quark concentration also provides a measure of this mass. The first material in which quarked atoms might be found is the atmosphere and Chupka et al. sampled large quantities of air containing ca. los molecules. In addition in case the quarked atoms adhered to dust particles the material collected by the air filters of the laboratory was also examined the amount corresponding to about air molecules. Some anomalous behaviour was found but unfortunately this was not repeatable and it was concluded that if the quark mass was less than 10 Gev/c2 then the con- centration must be less than 1 quark in los nucleons. The experiments were repeated in sea water and on meteorite material but no definite evidence for the existence of quarks has been obtained.Another line of attack is to repeat Millikan’s famous oil-drop experiment. It is interesting to note that Millikan himself reported ‘one uncertain and unduplicated observation . . . giving a value of charge on the drop some 30% lower than the final value of e’. The quest for the quark has now spread outside the Earth to the solar atmo- spheie. It has been argued2 that -Qe quarks in the galaxy would be found pre- ferentially attached to carbon nitrogen and oxygen atoms in the stellar atmo- spheres. These quarked atoms should show distinctive and easily calculated 1 W. A. Chupka J. P. Schiffer and C. M. Stevas Phys. Rev. Letters 1966,17,60. 486 Y. B. Zel’dovich L. B. Okun’ and S.B. Pikel’ner Uspekhi Fiz. Nuuk 1965,87 113. 1 1 t '\ \ '\ I&. 0 20 40 60 TIME (sex) I0 Id z Cbl I I i HOLD P I I I 0 20 40 60 Fig. 5 (a) The evaporation rate of a sample containing negative ions. The rate is determined by the evaporation of neutralised particles and is unaflected by reversing during the 'hold' period. (b) A corresponding result for a positive-ion species. The evaporation is inhibited during the reversed field 'hold' period. Quarterly Reviews electronic spectra. A recent search in the ultraviolet region3 has proved negative but the limits on the quark concentration are not very stringent. 6 Bootstraps At the beginning of this Review the original idea of field theory was noted in which nuclear forces are considered to be due to the exchange of n mesons.This requires that a nucleon should be able to convert into a nucleon plus an meson and vice versa and similarly that a meson should be able to convert into a nucleon-antinucleon pair and vice versa N + N +=; = +i + N (7) These elementary processes are virtual because energy and momentum con- servation cannot be satisfied simultaneously but real scattering occurs through a sequence of elementary processes such as Another way of looking at the elementary reactions of eqn. (7) is to consider that the nucleon partly consists of a nucleon and a n meson (and partly of all other combinations of particles into which it can convert). Similarly every particle can be considered to be composed of combinations of other particles the most important contribution coming from the combination of lightest mass.For example the p meson would be predominantly composed of two n mesons the N*l resonance of a nucleon and a n meson and so on. Sets of mathematical equations based on rather general principles such as conservation of probability and relativistic invariance can be written embodying these ideas which express the relationships between the masses of particles and scattering amplitudes for processes like eqn. (8). It is the hope of the ‘bootstrap’ school of theorists that these equations have a unique solution determining the observed masses. As there are probably an infinite number of these equations relating all possible partides and scattering processes some rather drastic approximations must be made but such calculations as can be done are quite encouraging.For example the N* resonance can consistently be described as n meson-nucleon bound state. The most extensive calculations by G. F. Chew and his colleagues at Berkeley on the system of the twon mesons have occupied some six years and still have not produced conclusive results but some encouraging ideas have emerged. For example one prediction is that particles can be grouped into families so that for each family the masses are a smooth function of the spin of the particles. These mass-spin plots are known as Regge trajectories and it does seem that the known particles do lie on such curves. Further the existence of these trajectories has been shown to imply certain behaviours of scattering processes at very high energies which also appear to be satisfied. How can the bootstrap model be reconciled with the quark model? How far can the N* be described both as a meson-nucleon state and also as a bound 0.Singanoglu B. Skutruk and R. Tousey Phys. Rev. Letters 1966 17,785. 488 Bransden state of three quarks? These questions are far from being resoIved and it may well be that the two models are incompatible. This does not of course mean that the higher symmetries cannot be incorporated into the bootstrap theory but that these symmetries are not consequences of the existence of real quarks bound together in some potential. In this Review only the strong interactions have been discussed because these are responsible for the proliferation of elementary particles. The weak interaction and their connection with the strong interactions have been also extensively studied with some success and the interested reader might consult ref. 4 for further study. C. E. Swartz ‘The Fundamental Particles’ Addison-Wesley London 1965; R. D. Hill ‘Tracking Down Particles’ Benjamin New York 1963. 489

ISSN:0009-2681    

DOI:10.1039/QR9672100474

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

The Carbanion Mechanism of Olefin-forming Elimination By D. J. McLennan CHEMISTRY DEPARTMENT UNIVERSITY OF AUCKLAND NEW ZEALAND Introduction This Review is primarily a summary and evaluation of evidence pertaining to the carbanion or ElcB mechanism1 of olefin formation. The mechanism differs from the more usual bimolecular (E2) mode of elimination in that it is a step- wise process involving the intermediacy of a discrete carbanion whereas an E2 elimination is concerted and passes through a single transition We illustrate the case for a neutral basic species. Scheme 1 Carbanion mechanism (El cB) B H &+(I I 1 I 1 B + -F-F- - - ~ = y - - BH+ + Scheme 2 Bimolecular mechanism (E2) X x 6- By assuming a steady-state carbanion concentration in anism we obtain k,k,[substrate] p] Rate = k-1[BH+I + k >c=c( + x- the carbanion mech- (1) so that second-order kinetics rate = kobs [substrate] [B] as will be exhibited by all non-solvolytic E2 reactions will be observed under the following condi- tions (assuming that [substrate] and [B] are comparable) (i) The first step is rate-determining being essentially the bimolecular irre- versible formation of the carbanion and the second step is the relatively rapid ejection of the leaving group from the a-carbon atom (k,>>k-JBH+]).[According to current terminology concerned with elimination reactions the leaving group (halogen quaternary ammonium arenesulphonate etc.) departs from the a-carbon atom and the proton is removed from the /%carbon atom.] (ii) The first step is a rapidly-attained equilibrium and the second is the rate- C.K. Ingold ‘Structure and Mechanism in Organic Chemistry’ Cornell University Press Ithaca New York 1953 p. 423. * Ref. 1 p. 422. a J. F. Bunnett Angew. Chem. Internut. Edn. 1962 1 225. 490 McLennan limiting unimolecular decomposition of the carbanion (kkl[BH+]> > k2). If in this case the base B is not the lyate ion of the solvent its conjugate acid BH+ must be present in excess to allow the observation of second-order kinetics. Carbanionic eliminations thus fall into two distinct mechanistic categories one in which carbanion formation is irreversible and rate-determining which we shall call the ‘irreversible’ type of mechanism although we are considering the irreversibility of the first step only and another in which the carbanion and substrate are in equilibrium.We shall call the latter the ‘pre-equilibrium’ type of carbanion mechanism. A kinetic distinction between these categories will be discussed later. Under the most commonly used conditions of base and solvent (alkoxide ions in alcohol) there are obtained second-order kinetics for both classes of carbanion mechanism and for the bimolecular mechanism so that no kinetic criterion of mechanism is available. Recent developments of the theory of the bimolecular mechanism view this process as comprising a spectrum of sub-mechanisms which differ subtly in transition-state character. Although C-H and C-X bond-breaking processes are necessarily coupled in that electrons partially freed by the former facilitate the latter and impart double-bond character to the transition state the stretching of one bond may outpace that of the other in forming the transition state.3 Thus we can conceive three general types of €2 transition state which are illustrated in Scheme 3.The transition state may resemble a carbonium ion the olefinic product or a carbanion depending on the timing of the covalency changes El-like central El cB- I ! ke Scheme 3 at the 01- and /%carbon atoms. The carbonium ion and carbanion intermediates of the pure El and ElcB mechanisms are thus seen to be limiting structures in the spectrum of E2 transition states. Reference should be made to Bunnett’s papel.3 for the development of the theory and a discussion of its consequences and applications. In the absence of considerable stabilisation (which is the case in most of the reactions we discuss) the carbanion in an ElcB reaction will be formed in an endo-energetic step so that the transition state leading to it should have car- banionic characteristic^.^ Thus structural and environmental effects will be similar to those expected for E2 reactions utilising El cB-like transition states making experimental demonstration of the nature of the intervening species in an elimination reaction difficult.The mechanistic tests we describe in the follow- ing sections are therefore not completely unambiguous. G. S. Hammond J . Anier. Chem. SOC. 1955 77 334. 49 1 3 Quarterly Reviews The list of specific applications of mechanistic criteria will not be exhaustive as experimental results for many reactions that may be ElcB are too meagre to allow more than speculation. The reader is referred to Banthorpe5 for mention of several such cases.We confine our discussion to reactions mentioned in the title. Acetylene formation from vinylic halides and the like (which almost invariably involves a carbanionic intermediate)* has been adequately reviewed by Kobrich? and while the stepwise’ and concerted8 mechanisms of halogeno- form hydrolysis are analogous to the ElcB and E2 mechanisms respectively as has been demonstrated by some of the mechanistic tests described herein we will not discuss them in any detail. @Hydrogen Exchange accompanying Elimination If a pre-equilibrium carbanionic elimination is proceeding in a protic solvent P-hydrogen atoms of the substrate will be exchanging with solvent protons at a rate which is rapid compared with that of olefin formation. Such exchange can be demonstrated by using either the substrate or solvent suitably labelled with isotopic hydrogen allowing the reaction to proceed to partial completion and examining recovered substrate for changes in isotopic composition.If exchange is not detected an irreversible-type carbanion mechanism or an E2 mechanism is indicated. Hine Wiesboeck and Ramseys have shown that 1,1,l-trifluoro-2,2-dihalo- genoethanes are dehydrofluorinated much more slowly than they undergo hydrogen exchangelo in alkaline methanol. The activating effects that the a-fluorines and /%halogens have on the /%hydrogen together with the well- known reluctance of fluorine to depart as an anion from saturated carbon (and more so from a carbon bearing other halogensll) make the intermediacy of carbanions in these eliminations highly likely and in fact they provide us with the only examples of the ElcB mechanism being involved in reactions of satur- ated halogeno-alkanes.Breslow12 has raised the formally correct point that *In some cases it has been found that trans elimination from a vinylic halide proceeds by a concerted mechanism whereas the analogous cis elimination is carbanionic (and slower) (See S. J. Cristol and W. P. Norris J. Amer. Chem. Soc. 1954 76 3005; S. J. Cristol A. Begoon W. P. Norris and P. S. Ramey ibid. p. 4558; G. Marchese G. Modena and F. Naso Cliem. Comm. 1966,492). This fact does not vitiate the arguments we advance on the stereo- chemistry and its mechanistic consequences of olefin-forming elimination since the mech- anisms and stereochemistry of displacements at unsaturated centres have little in common with those of saturated displacement reactions [see D.E. Jones R. 0. Morris C. A. Vernon and R. F. M. White J. Chem. SOC. 1960,2349; J. F. Bunnett ‘Theoretical Organic Chemistry’ (Kekult Symposium) Butterworths London 1958 p. 1441. D. V. Banthorpe ‘Elimination Reactions’ Elsevier Amsterdam 1963 chap. 4. G. Kobrich Angew. Chem. Internat. Edn. 1965 4 49. J. Hine J. Amer. Chem. SOC. 1950,72,2438; J. Hineand A. M. Dowell ibid. 1954,76,2688. J. Hine and J. J. Porter J. Amer. Chem. SOC. 1957 79 5493; J. Hine and P. B. Langford J. Hine R. Wiesboeck and 0. B. Ramsey J. Amer. Chem. SOC. 1961,83 1222. ibid. 1957 79 5497. lo J. Hine R Wiesboeck and R. G. Ghirardelli J. Amer. Chem. Soc. 1961,83 1219. l1 J. Hine C. H. Thomas and S. J. Ehrensen J. Amer.Chem. Soc. 1955 77 3886; J. Hine S. J. Ehrensen and W. H. Brader ibid. 1956 78 2282. l2 R. Breslow Tetrahedron Letters 1964 NO. 8 399. 492 McLennan depending on the relative energetics rapid hydrogen exchange may be an irrelevant side-reaction to E2 olefin formation. Hine and his co-workersg anti- cipated this argument by pointing out that if carbanions are formed they are most likely to be intermediates in the elimination sinm transfer of their free electron pairs should be more effective in promoting olefin formation than should transfer of a partially free electron pair in an E2 transition state. Pertaining to this concept is the observation of Paquette and Wid3 that pyrolysis of the ammonium hydroxide (1) is accompanied by hydrogen exchange on the carbon a to the carbonyl group.This is not unexpected in view of the presence of the keto-function but the question remains as to whether the car- banion is actually involved in the elimination. In the following section we discuss a clear case of an a-keto-group’s assisting an ElcB reaction and by inference it appears that the reaction of (1) is also ElcB. The poorness of the leaving group3 would also make this pathway feasible. Carbanions are formed readily by removing a proton from the 9-position of fluorene and its derivatives as shown by exchange experiments. When a tri- fluoromethyl substituent is present at the 9-position7 the carbanion (generated in alkaline methanol) slowly ejects fluoride ion to form the substituted f~1vene.l~ The usual factors [carbanion stability (caused in this case by the aromaticity of the carbanion as well as the electron-withdrawing power of the trifluoromethyl group) and the tightness of binding of the leaving group] are responsible for the El CB mechanism’s being preferred.Iskander and Riad have found that the formation of p-nitrostyrene from 2-(p-nitrophenethyl)thioacetate ion is accompanied by hydrogen exchange.16 However the relative rates of the exchange and elimination reactions are not available and further studies would be desirable before a mechanism is con- fidently assigned to reactions of this unusual type. The case of p-benzene hexachloride (1,2,3,4,5,6-hexachlorocyclohexane with all chlorines equatorial) has aroused much interest. Cristol and Fix16 observed a very small amount of hydrogen exchange accompanying its dehydrochlorina- tion and contending that this was significant invoked the ElcB mechanism.Their reaction scheme involved the rate-determining formation of a carbanion which inverted its configuration to that of the &isomer carbanion more rapidly than it abstracted a proton from the solvent. The 8-carbanion after reprotona- l3 L. A. Paquette and L. D. Wise J. Org. Chem. 1965 30 228. l4 A. Streitwieser A. P. Marchand and A. H. Pudjaatmaka J . Amer. Chem. SOC. 1967 89 693. l5 Y . Iskander and Y. Riad J Chem. SOC. 1961,223. l6 S. J. Cristol and D. D. Fix J . Amer. Chem. SOC. 1953 75 2647. 493 3+ Quarterly Reviews tion then could react by E2 elimination to form the products (a mixture of trichlorobenzenes with the 1,2,4isomer predominating) more rapidly than the b-isomer for stereochemical or conformational r e a s ~ n s .~ ~ J ~ As pointed out by Cram,lD this mechanism requires that the products contain deuterium when the reaction is run in a deuterated medium but subsequent investigations have shown that such deuteration is insignificant.20 Thus the mechanism can be only E2 or irreversible ElcB. Isotopic demonstrations of reversible carbanion formation in the course of elimination are rare However several substrates with structural features con- sidered favorable for carbanionic elimination may have failed the hydrogen exchange test for the reason that under the conditions of alkalinity employed in the reaction the inequality kJBH+]> > k did not apply. Base Catalysis Although elimination reactions are not base-catalysed in the true sense of the words they adhere to rate laws consistent with recognised classes of base catalysis.The kinetics of base-catalysed reactions have been fully discussed by Be1121 but we shall consider them from the point of view of eqn. (1). Let us assume that a carbanionic elimination is proceeding under pseudo-first-order conditions (observed rate coefficient k+) in a B-BH+ buffer system. In the pre- equilibrium case k,[BH+]:.>k, so k* is given by eqn. (2) and changes of base concentration at constant buffer ratio should have no effect on k+. This is specific base catalysis and reversible protonation of the carbanion should be observable. In the other extreme case k-l[BHs]< <k and we obtain eqn. (3) so k, is linearly dependent on base concentration at constant buffer ratio and general base catalysis is observed.As buffer concentration is increased at constant ratio so that the kinetics change from those given by eqn. (3) to those of eqn. (2) the situation depicted in the Figure should result. On the other hand bimolecular reaction should be general base-catalysed throughout the range of buffer concentrations at a given pH so in theory at least we have a kinetic criterion of mechanism. There is the complication that a general base-catalysed reaction may exhibit kinetics qualitatively consistent with specific base catalysis if the catalytic efficiency of the lyate ion is much greater than that of any other base present as in the E2 reaction of DDT with benzenethiolate in methanol22 or if the base is strong enough to produce a k$b = kl[BI (3) l7 E. D. Hughes C. K. Ingold and R. A. Pasternak J.Chem. Soc. 1953 3832. l8 S. J. Cristol N. L. Hause and J. S. Meek J. Amer. Chem. Soc. 1951 73 614. lS D. J. Cram ‘Steric Effects in Organic Chemistry’ ed. M. S. Newman Wiley New York 1956 p. 321. 2o J. Hine R. D. Weimer P. B. Langford and 0. B. Ramsey J . Amer. Chem. Soc. 1963 85 3984; 1966,88 5522. a1 R. P. Bell ‘The Proton in Chemistry’ Cornell University Press Ithaca New York 1959 p. 134. 494 McLennan kinetically significant concentration of lyate ions as when phenoxides are used as bases in E2 reaction^.^^,^^ I Figure The dehydrochlorinations of the diastereoisomeric 4,4’-dichlorochalcone dichlorides (2 and 3 ; Ar = p-Cl-C,H,) are not stereo~pecific.~~ The threo dich- loride yields chiefly the trans olefin as expected for a ‘trans’ elimination through an anti-periplanar transition state (see section on stereochemistry) while the erythro substrate gives a 2 1 mixture of the trans- and cis-olefins.This in itself is not sufficient evidence for the formation of a carbanion with sufficient lifetime to permit partial rotation about the central carbon-carbon bond but it was also H H Arc Ar@ a Cl (2) (threo) (3) (wyihro) shown that the kinetics of the reactions in buffered ethanolic solutions were consistent with a refined form of eqn. (1). A clear demonstration of the carbanion mechanism was thus constituted. Other halogenochalcones eliminate in a non- stereospecific fashion25 although some reactions of this type reported by South- wick et u I . ~ ~ are not direct eliminations. The decomposition of the Michael adduct of 4-nitrochalcone and malono- nitrile in neutral and acidic methanol exhibits kinetics that are consistent with a carbanion me~hanism.~’ Similar results for the decomposition of 1,1,1,3- tetranitro-2-phenylpropane have been reported by Hine and Kaplan2* and again the ElcB mechanism was invoked.Related to mechanisms of base catalysis are the basicity functions (H-) of 22 B. D. England and D. J. McLennan J . Chem. SOC. (B) 1966,696. a3 R. F. Hudson and G. Klopman J . Chem. Soc. 1964 5 . 24 T. I. Crowell A. A. Wall R. T. Kemp and R. E. Lutz J. Amer. Chem. SOC. 1963,85,2521. 25 R. E. Lutz D. F. Hinckley and R. H. Jordan J. Amer. Chem. SOC. 1951,73,4647. 26 P. L. Southwick A. K. Colter R. J. Owellen and Y.-C. Lee J . Amer. Chem. SOC. 1962 84,4299. 27 S. Patai S. Weinstein and Z . Rappoport J . Chem. Soc. 1962 1741.28 J. Hine and L. A. Kaplan J. Amer. Chem. SOC. 1960 82 291 5. 495 Quarterly Reviews alkaline solutions which are analogous to the Hammett acidity functions (HJ. Ridd and More O'Ferra1129 have obtained H- values for concentrated methanolic solutions of sodium methoxide and have found that log k$ for the hydrolysis of chloroform is proportional to the H- value of the medium. This was expected since proton transfer in this carbanionic a-elimination is under thermodynamic control and rates of such specific base-catalysed reactions should correlate with On the other hand the rates of the E2 dehydrochlorination of 2- phenethyl chloride and 1 -chloro-3,3-dimethylbutane (general base-catalysed re- actions) were proportional to the base concentrations and not to the H- values of the alkaline solutions.This mechanistic criterion is not widely applicable at present as H- scales have been established for only a few solutions. In any case the same objections that have been raised against drawing mechanistic con- clusions from H, correlation^^^ may apply to the use of H-. The catalytic constants (k) of constitutionally related bases reacting by general base catalysis with a given substrate can be related to their base strengths (&) by the Bronsted equation (4) in which the p parameter which normally has values between zero and unity is usually taken as a measure of the degree of transfer of the proton from the substrate to the base in the transition state.31 If an irreversible type of carbanion mechanism (general base-catalysed) in which the transition state resembles the carbanion is operative /3 should be close to unity and should ideally be unity if a pre-equilibrium type of ElcB reaction occurs as rate constants of specific base-catalysed reactions are directly proportional to the basicity constants of the attacking bases.32 These predictions await experi- mental verification in so far as elimination reactions are concerned.Bronsted exponents have recently been used to diagnose the nature of E2 transition states. It is found that relatively high values of /3 (0-6-0-9) are asso- ciated with carbanion-like transition states while lower /3 parameters (0.24.4) result when an E2 transition state has partial carbonium-ion c h a r a ~ t e r . ~ ~ ~ ~ Isotope Eff&s A. Deuterium Isotope Effects.-The deuterium isotope effect (kH/kD) of a threecentre proton transfer should be the theoretical maximum (about seven at room temperature in the absence of tunnelling) when the donor and acceptor molecules exert equal control over the proton in the transition state.= Several E2 reactions which probably pass through central-type transition states exhibit this maximum isotope effect.35 If the character of the transition state is shifted 29 J.H. Ridd and R. More O'Ferrall J. Chem. SOC. 1960 5030 5035. 30 J. F. Bunnett J . Amer. Chem. SOC. 1961 83,4956 4968,4973 4978. 31 J. F. Bunnett Ann. Rev. Phys. Chem. 1963 14 271. 32 R. P. Bell ref. 21 ch. 10. 33 D. J. McLennan J . Chem. SOC. (B) 1966,705 709. 34 F. H. Westheimer Chem. Rev. 1961 61 265. 35 (a) V. J. Shiner J. Amer. Chem. SOC. 1952 74 5285; (b) V. J. Shiner and M. L.Smith J. Amer. Chem. SOC. 1961 83 593; (c) W. H. Saunders and D. H. Edison J . Amer. Chem. SOC. 1960 82 138 496 Me Lennan towards the carbanion or carbonium ion extremes the isotope effect should decrease and should be at or near unity if proton transfer is either practically complete or has hardly commenced when the transition state is attained as in the ElcB and El mechanisms respectively. A low isotope effect should also be observed when the pre-equilibrium carbanion mechanism is operative for a /3-deuterated substrate will exchange most of its deuterium for hydrogen from a protic solvent before olefin formation becomes appreciable. Measurement of the deuterium isotope effect has enabled the irreversible- type carbanion mechanism to be identified as the pathway of the dehydro- fluorination of 1 1,l -trifluoro-2-methyl-3-phenyl propane in alkaline alcoholic solution.38 The value of 1-2 is the lowest yet reported for an olefin-forming elimination.The pre-equilibrium mechanism is invalidated by the lack of /%hydrogen exchange and the fact that optically active substrate can be recovered unracemised after partial reaction. The isotope effects in the pre-equilibrium carbanionic dehydrofluorination of CF,-CHCl and CF,-CHBrCl are 1-26 and 1.41 respectively,1° and these values are consistent with the assigned mechanism. Eliminations involving expulsion of fluorine from a trifluoromethyl group thus appear to be particularly favourable for the observation of the carbanion mech- anism although the resulting olefins are prone to is~merise~~ or suffer addition of ~ o I v e n t .~ ~ J * ~ ~ ~ There remains much scope for the measurement of deuterium isotope effects in reactions of doubtful mechanism some of which we will mention later in the section on stereochemistry. B. Leaving-group Isotope Effects.-These have not been exploited as much as deuterium isotope effects as mechanistic criteria and there are only two examples of interest. The nitrogen isotope effect for elimination from the 2-phenethyltrimethyl- ammonium ion in aqueous alkaline solution is experimentally indistinguishable from yet when the medium is alkaline ethanol the isotope effect is about 30% of that expected for complete C-N rupture in the transition state.39 This could mean that the mechanism changes from E2 to ElcB as the base-solvent system is changed from OEt-EtOH to OH-H,O but arguing against this interpretation is the fact that the deuterium isotope effect (3.0) is the same in both alkaline ethanol and 50% aqueous ethanolzc (where the base is largely present as hydroxide40).A more likely explanation is that the difference in nitrogen isotope effect is associated with the greater solvation energy to be over- come when the C-N bond is stretched in water. The bimolecular transition state will shift further in the carbanion direction with increasing solvent with a consequent decrease in the degree of C-N rupture. 36 D. J. Cram and A. S. Wingrove J . Amer. Chem. SOC. 1964,86 5490. 37 W. T. Miller E. W. Fager and P. H. Griswold J. Amer. Chem. Soc. 1948,70 431. 38 S. AYperger et al. quoted by D. V. Banthorpe and J. H. Ridd Proc. Chem. SOC. 1963,225.39 G. Ayrey A. N. Bourns and V. A. Vyas Canad. J. Chem. 1963,41 1759. 40 R. G. Bums and B. D. England Tetrahedron Letters 1960 No. 24 1. 41 D. V. Banthorpc E. D. Hughes and C. K. Ingold J . Chem. SOC. 1960,4056. 497 Quarterly Reviews The second case involves elimination from the isomeric 2-phenylcyclohexyl- trimethylammonium ions to yield 1-phenylcy~lohexene.~~ anti-Elimination from the cis isomer(4) has a nitrogen isotope effect of 1.2% (theoretical maximum for NMel dH (5) H Ph complete C-N rupture in the transition state would be about 3%) which is indicative of an E2 mechanism. However the isotope effect for syn elimination from the trans isomer ( 5 ) is only 0-2% so it appears that the antielimination is E2 and the syn elimination is ElcB. Cristol and Stermitz had earlier assigned these mechanisms on stereochemical grounds.& Carbon-14 isotope effects can be measured for displacement at both the a- and the /%carbon atom but as yet only investigations on E2 reactions have been carried out .44 C.Solvent and Secondary Isotope Effects.-Deuteroxide ion in D,O is a stronger base than hydroxide in water by a factor of about 1.6 at 80" and Stcffa and T h ~ r n t o n ~ ~ have reported solvent isotope effects [k(D,O)/k(H,O)] of 1-57 and 1-79 for the respective E2 reactions of the 2-phenethyl-dimethylsulphonium and -trimethylammonium ions at this temperature. These effects demonstrate the change in E2 transition state character that has been so amply illustrated in the 2-phenethyl s e r i e ~ 3 ~ ~ ~ ~ and such a method should be capable of detecting a carbanion mechanism.The solvent isotope effect for the hydrolysis of chloro- form4' (an a-elimination proceeding via a reversibly-formed carbanion) is 1-74 and this figure could be taken as a guide for the ElcB reactions of neutral sub- strates. A secondary (a) deuterium isotope effect of unity is expected for a carbanion mechanism but since elimination from the 2-phenethyldimethylsulphonium ion a known E2 reaction has an a-isotope effect of 1.00,4* this particular criterion holds little promise. The Hammett Equation Rates of elimination of substrates bearing suitably substituted a- or p-bound yhenyl groups can be correlated by the Hammett equation. For /&bound aryl 42 G. Ayrey E. Buncel and A. N. Bourns Proc. Chem. SOC. 1961,458. 43 S. J. Cristol and F. R. Stermitz J. Amer.Chem. SOC. 1960 82,4962. 44 H. Simon and G. Mullhofer Pure Appl. Chem. 1964 8 385. 45 L. Steffa and E. R. Thornton J. Amer. Chem. Sac. 1963 85,2680. 46 W. H. Saunders and R. A. Williams J. Amer. Chem. SOC. 1957,79,3712 C. H. DePuy and D. H. Froemsdorf ibid. 1957,79,3710 C. H. DePuy and C. A. Bishop ibid. 1960,82,2532 2535. 47 J. R. Jones Trans. Faraday Soc. 1965,61 95. 48 S. ASperger N. Ilakovac and D. PavloviE J. Amer. Chem. Soc. 1961 83 5032. 498 McLennan groups the value of the reaction parameter p increases as structural and environ- mental conditions are changed so as to shift E2 transition states in the ElcB Thus a high positive value of p is expected to obtain for a carbanionic elimination since negative charge density on the p-carbon atom in the transition state will be at a maximum.The question of how high a p value to expect for a ElcB reaction is in doubt. DePuy et al.49 suggested that the value of 5.0 obtained from rates of carbanionic polymerisation be used as a guide and while this seems reasonable a model process more similar to elimination reactions would be desirable. Stereochemistry of Elimination Reactions If it is accepted that all E2 reactions follow an anti (‘trans’)* stereochemical course then any elimination proceeding with different stereochemistry (e.g. syn or ‘cis’ elimination) must necessarily have a different mechanism and the carbanion mechanism is most frequently cited in this regard. It is the purpose of this section to examine critically the empirical and theoretical basis of the above assumption. In doing so we will show that it is only partially correct by present- ing evidence in favour of the availability of a syn-stereochemical course to E2 reactions which thereby removes the necessity of suggesting another (e.g.carbanion) mechanism for reactions that are not anti-stereospecific. First based on experimental observation^,^^ the ‘E2 Rule’ has recently been rationalised on the basis of a conceptual mechanistic dissection51 and by mole- cular orbital descriptions and calculation^.^^ It is clear that in an E2 transition state the partial bonds joining the p-hydrogen and the leaving group to the central carbons should be as nearly coplanar as possible in order that p-orbital overlap to form the developing n-bond be maximised but this simple considera- tion leaves the question of anti against syn coplanarity open.In a recent treat- ment anti elimination is justified by a consideration of the magnitude of the resonance integral between the atoms or groups whose departure effects double- bond formation.= The experimental evidence for the ‘E2 Rule’ usually involves the fact that eliminations from a suitable diastereoisomeric pair are anti-stereospecific or that the formation of 1-R-cycloalkene by anti elimination of HX from cis- 1 -R-2-X-cycloalkane is faster than the corresponding syn elimination from the trans isomer. In the acyclic case anti elimination can proceed through as many staggered *We prefer to discuss the steric relationships of groups across a single bond in terms of the nomenclature of W. Klyne and V. Prelog Experientia 1960 16 521 and to reserve the cis- trans nomenclature for cases of geometric isomerism.49 C. H. DePuy G. F. Morris J. S. Smith and R. J. Smat J. Amer. Chem. SOC. 1965 87 242 1 . 50 C. K. Ingold ref. 1 p. 467; D. J. Cram ref. 19 chap. 6. 51 C. K. Ingold Proc. Chem. SOC. 1962 265. 52 G . H. Stewart and H. Eyring J. Chem. Educ. 1958 35 550; K. Fukui and H. Fujimoto Tetrahedron Letters 1965 No. 48 4303. 53 E. L. Eliel N. L. Allinger S. J. Angyal and G. A. Morrison ‘Conformational Analysis’ Interscience Publishers New York 1965 p. 482. 499 Quarter 1y Re views (low-energy) conformations of the initial and transition states as there are hydrogens on the /%carbon atom but for syn elimination an eclipsed (higher energy) conformation must lie on the reaction co-ordinate. If we take the 3 kcal./ mole potential barrier to rotation of ethane as the minimum difference in energy between the anti-periplanar and syn-periplanar transition states for elimination from a substrate where anti elimination will produce one olefin and syn elimina- tion another we find that the former reaction will be favoured by a rate factor of at least 160.In other words without having made any assumptions regarding stereochemical rate preferences in elimination we have calculated that con- formational factors will cause the product mixture to contain 99-4 % of the olefin arising from anti elimination. More sensitive techniques than those commonly used in the past would be necessary to detect the minor product. It will be noted that the foregoing only applies to fully concerted E2 reactions and that alternative activation processes and transition-state conformations may be available when either El-like or ElcB-like transition states are involved.51 We turn from product comparisons in the acyclic case to the rate determina- tions that have demonstrated the preference for anti elimination from alicyclic compounds.For cyclohexanes it is known that the leaving group must be axial for easy anti elimination to occur,54 and if by analogy we assume that the leaving group in a syn elimination must also be axial we find that anti elimina- tion is diaxial and syn elimination is axial-equatorial.* The two initial state conformations for reaction are thus (6) and (7) respectively in which R is a hydrogen-activating group (e.g. Ar ArSO,). It can be seen immediately that syn elimination from the trans isomer (7) will be comparatively unfavorable X X A (6) unii-elimination (7) syn-elimination (both leading to I - R-cyclohexene) because of the necessity of having two bulky groups in the axial positions in the reactive conformation which will thus be sparsely populated.Again a rough calculation is informative. If we assume that both the above conformations eliminate HX at the same rate that eliminations from the conformations with equatorial X can be neglected and that the difference in free energy between (6) and (7) is 3-5 kcal./mole (reasonable for large R and X) we find that the mea- sured rate of anti elimination from the cis isomer is about 370 times greater *Recent work in a conformationally frozen 2-p-tolyl sulphonylcyclohexyl toluene-p-sulphonate system has shown that the leaving group in a syn elimination indeed prefers the axial orienta- tion for departure (W.M. Jones T. G. Squires and M. Lynn J. Amer. Chem. SOC. 1967 89,318). No further evidence supporting the operation of the El cB mechanism can be adduced from these authors’ results (cf. refs. 56 57 59 and 60). 54 D. H. R. Barton J . Chem. SOC. 1953 1027. 500 McLennan than that of syn elimination from the trans isomer. This observed rate ratio is not a consequence of the inherent preference for anti elimination. It arises out of differences in the position of conformational equilibria. Banthorpe= has considered this matter from a slightly different viewpoint but the same con- clusion can be drawn from his illustrative data namely that the observed pre- ference for anti elimination may be wholly or partly an artefact arising out of initial-state conformational differences.syn-Periplanar transition states for a cyclohexyl syn elimination are available only when boat-form conformations are attainable and the even greater energy of these can further accentuate the apparent preference for anti elimination. (See N. 0. Brace J. Arner. Chem. SOC. 1964 86 2428.) Implicit in the above is the assumption that E2 syn elimination is possible and we now cite evidence to that effect. Deuterium isotope effect and Hammett studies by DePuy and his co-worker~*~ have shown that the formation of 1- phenylcyclopentene from cis- and trans-2-phenylcyclopentyl toluene-p-sul- phonates (by anti and syn eliminations respectively) follows the E2 mechanism. The important point was that kanti/ksyn had the low value of 14 whereas with the cyclohexyl analogues (6 and 7; R = Ph X = OTs) this kinetic preference for anti elimination was greater than 10,OOO.The difference which appears in other cyclopentyl-cyclohexyl reaction has been rationalised in terms of E2 elimination rates' being at a maximum when the dihedral angle between the H-C and C-X bonds is either 180 O (anti elimination) or 0" (syn-periplanar elimination) and decreasing to a minimum as the angle approaches 90". The dihedral angle in cyclopentyl syn eliminations is clearly less than in the corre- sponding cyclohexyl cases thus affording a greater chance for partial 7-r-bond formation to occur without undue torsional strain. When neutral bases are used there is also an electrostatic contribution to the preference for cyclopentyl syn elimination.56b It has already been mentioned that when the leaving group is NMe,+ (6 and 7; R = Ph X = NMe,+) the ElcB mechanism apparently prevails in the syn elimination while the anti elimination is E2.42,43 This is not necessarily a general result since the positively charged leaving group will be more able to stabilise carbanionic charge than will a neutral atom or group.When the arylsulphonyl group is used to activate the /%proton (6 and 7; R = ArSO, X = OTs) the situation is unchanged and the preference for anti elimination is again less strong with the corresponding cyclopentyl corn pound^.^^ The pre-equilibrium carbanion mechanism was ruled out by the finding of general base and rates of carbanion formation from the cyclopentyl compounds were shown to be too slow to account for the syn elimination,57 which must therefore be E2 When trimethylamine was used as base stereo- chemical preference was almost completely absent from the cyclopentyl series.66b 55 D.V. Banthorpe ref. 5 p. 90. 56 (a) J. Weinstock R. G. Pearson and F. G. Bordwell J. Amer. Chem. Soc. 1956 78 3468; (b) J. Amer. Chem. SOC. 1956 78 3412. 57 J. Weinstock J. L. Bernoulli and R. G. Pearson J. Amer. Chem. SOC. 1958 80 4961. 501 Quarterly Re views Hine and Ramsey using a Taft equation6* approach have shown indirectly that the rates of carbanion formation and syn elimination from the cyclohexyl trans isomer are about equalF9 and have thereby secured proof of the carbanion mechanism albeit from a three-point linear free-energy plot. The corresponding trans-chloride is about as reactive in syn elimination as the sulphonate,6O and while this is difficult to understand in terms of an E2 mechanism where second- ary sulphonates are invariably more reactive than halides the ElcB mechanism affords a ready explanation since the activation step involves proton ionisation and OTs and C1 have similar electron-withdrawing powers.69 However the relative rates of displacement of halide and sulphonate ions are variable and there are aspects of sulphonate eliminations that are not well understood.61 Turning now to the 2,3-dihalogenonorbornanes we find that syn elimination from the trans-dihalides (8; X = Hal) is faster than anti elimination from the endo-cis isomers (9; X = Hal).62963 Measurements of deuterium isotope effects show that the E2 mechanism in contradiction of an earlier suggestion of the carbanion mechanism.62 Removal of exo groups in the syn elimination is the preferred stereochemical c0urse,6~ and the tentative explanation is in terms of steric effects.DePuy’s hypothesis explains the unusual steric preference nicely as coplanarity can be achieved in the transition state for syn elimination but the geometric rigidity of the system prevents this in the anti dehydrohaloge- nation. The carbanion mechanism has also been invoked to explain the same reactivity order syn > anti in the dehydrochlorination of the 11,12-dichloro-l1,12- dihydr0-9,lO-ethanoanthracenes~~ (10 and 11) but the necessary tests to confirm this have not been performed. The effect of substituting an acid-strengthening p-tolylsulphonyl group for one of the chlorines is to increase the rates of de- hydrochlorination about a million-fold and to cause anti elimination to pre- dominate by a small rate factor.64 This is intelligible in terms of the carbanion 68 R.W. Taft ‘Steric Effects in Organic Chemistry’ ed. M. S. Newman Wiley New York 1956 ch. 13. 59 J. Hine and 0. B. Ramsey J. Amer. Chem. SOC. 1962 84 973. 6o H. L. Goering D. I. Relyea and K. L. Howe J. Amer. Chem. SOC. 1957 79 2502. 61 (a) H. M. R. Hoffmann J. Chem. SOC. 1965,6753 6762; (b) D. H. Froemsdorf W. Dowd and K. E. Leimer J. Amer. Chem. Soc. 1966 88 2345; J. F. Bunnett and R. A. Bartsch unpublished results. 62 S. J. Cristol and E. F. Hoegger J. Amer. Chem. SOC. 1957 79 3438. 63 N. A. LeBel P. D. Beirne E. R. Karger J. C. Powers and P. M. Subramanian J. Amer. Chem. SOC. 1963 85 3199; N.A. LeBel P. D. Beirne and P. M. Subramanian ibid. 1964 86 4144; H. Kwart T. Takeshita and J. L. Nyce ibid. p. 2606. S . J. Cristol and N. L. Hause J. Amer. Chem. SOC. 1952 74 2193; S. J. Cristol and R. P. Arganbright ibid, 1957 79 3441. 502 McLennan ~ A c ?Ac mechanism as discussed by Cristol and his co-workers but an E2 explanation is also available. The substituent could cause an E2 transition state to shift in the ElcB direction where the bond changes are relatively uncoupled and this may result in the stereochemical consequences of geometric rigidity being overcome. If the mechanism is in fact E2 DePuy's hypothesis receives further confirmation. Bimolecular syn elimination has been suggested in order to explain apparently anomalous cis trans olefin ratios in the reactions of secondary arenesul- phonates65 and 'onium salts.66 Other syn eliminations recently shown to be operative in the reactions of cycloalkyl 'onium saltss7 could from the pattern of results very well be caused by the incursion of an a'-P (ylide)6* mechani~m,~~ although no elimination reaction in alkaline solution has yet been shown to proceed by this mechanism.There is also one example of syn elimination being faster than anti elimina- tion in a conformationally mobile system. Bordwell Arnold and Biranowski'O have reported that removal of acetic acid from compound (12) by a syn-clinal elimination (60" dihedral angle) is 3.5 times faster than anti elimination (180" dihedral angle) from its isomer (13) when piperidine is used as base in an ethanol- chloroform solvent. This fact is inexplicable in terms of the E2 mechanism as neither conformational considerations nor DePuy's hypothesis can accom- modate it.Using kinetic parameters Hammett reaction constants and the effects of 4,4-dimethyl substituents as supplementary evidence Bordwell and his co-workers invoked the ElcB mechanism noting that the strongly electron- withdrawing nitro-group would facilitate this. At first sight the deuterium isotope effect for elimination from both compounds (4.9) seems to be too high 65 C. H. Snyder and A. R. Soto Tetrahedron Letters 1965 No. 37 3261. For dficulties in interpreting product results from arenesulphonate eliminations in dimethylsulphoxide see ref. 616. 66 J. Zavada and J. Sicher Proc. Chem. Soc. 1963 96; Coll. Czech. Chem. Comm. 1965 30 438. 67 J. Sicher J. Ztivada and J.KrupiEka Tetrahedron Letters 1966 No. 16 1619; J. Zgvada M. Svoboda and J. Sicher ibid. 1966 No. 16 1627. 69 J. F. Bunnett personal communication. 70 F. G. Bordwell R. L. Arnold and J. B. Biranowski J. Org. Chem. 1963 28 2496. D. V. Banthorpe ref. 5 p. 101. 503 Quarterly Reviews for carbanionic elimination but Lewis has found that such isotope effects in the formation of carbanions by pyridine-catalysed deprotonation of nitroalkanes can be as high as 25.'l This is attributed to a proton tunneling effect. We now consider so-called stereoconvergent26 eliminations in which a pair of diastereoisomers eliminate one necessarily by a syn elimination to give the same olefin as illustrated in Scheme 4. Cram Greene and DePuy have dis- covered such a reaction in the 1,2-diphenyl-l-X-propane system (R = Me; R' = Ph) where in the medium t-butoxide-t-butyl alcohol the trimethyl- ammonium salts (X = NMe,+) both yield the trans-olefin (their scheme is cis - olef in onti R;C,H I! (erytbro) / - R R/C'$ R trans- o I e f i n H*:.(threo) Scheme 4 opposite to that shown in our Scheme 4) and at virtually identical rates.72 As the starting materials are of approximately equal the transition states must be of comparable energy making the carbanion mechanism also favoured by structural and medium features a strong possibility. Whether the carbanion in the overall syn elimination is formed in a syn-stereospecific step or whether the proton anti to the leaving group is removed to be followed by rotation is not known. When ethoxide in ethanol is the base unti-stereospecificity is restored and the threo-ion becomes 57 times more reactive than the erythro as a result of the E2 transition state's shifting towards the central region where phenyl eclipsing effects are important.The reactions in t-butyl alcohol warrant further investigation and studies of isotope and phenyl-substituent effects could prove fruitful. Cristol and pa pya^'^ report that the threo- and erythro-chlorides (R = ArSO, R' = Ph X = C1 in Scheme 4) both yield the cis-olefin on dehydrochlorination with anti elimination from the erythro-chloride being much faster than syn elimination from the threo-substrate. This could be a case of E2 anti elimination 71 E. S. Lewis and J. D. Allen J . Amer. Chem. SOC. 1964,86,2022; L. Funderburk and E. S. Lewis ibid. 1964 86 2531. 72 D. J. Cram F.D. Greene and C. H. DePuy J. Amer. Chem. SOC. 1956,78,790. 79 F. A. Abd Elahafez and D. J. Cram J . Amer. Chem. SOC. 1953,75 339. 74 S. J. Cristol and P. Pappas J. Org. Chem. 1963 28 2066. 504 McLennan and carbanionic syn elimination but further data were not presented to justify the latter mechanism. In particular it would be interesting to know the relative stabilities of the two diastereoisomers so that conformational effects on the rate difference could be accounted for. The hydrogen and chlorine atoms in the threo-substrate will certainly be predominantly in syn-clinal positions relative to each other in the initial state and presumably syn-periplanar in the transition state were the mechanism E2 and the consequent eclipsing of phenyl groups could be the cause of the decrease in rate.However in the absence of more data we are hesitant to comment further on this reaction. Hughes and Maynard75 have studied the dehydrochlorination of the racemic (14) and meso (1 5 ) 1,Zdichlorosuccinic acids in aqueous alkaline solution and confirmed earlier reports of stereoconvergency. Both compounds yield chloro- fumarate ion although anti elimination from the meso substrate should give H -0,c cYx:- H / - 0,c LC H c1’ ‘COT II C (c h lor of u ma rate ) H p 2 - C C II Cl’ ‘COT (I 5) (chloroma leate) chloromaleate. Tests for hydrogen exchange were negative and the apparent anomaly was explained in terms of ElcB-like bimolecular transition states. It was suggested that the carboxyl groups acted as reservoirs for the electron pair partially released by the stretching of the C-H bond by virtue of the potential prototropy they imported into the system.This being the case electrostatic repulsion between them would cause transition states leading to chlorofumarate to be favoured. This factor should not arise when the reactions are conducted in neutral aqueous solution (where they were shown to be E2) and in fact the meso-dichloride yields comparable amounts of chlorofumarate and chloro- maleate. The anti-stereospecificity expected of an E2 reaction is illustrated by the formation of cis- and trans-olefins from the erythro- and zhreo-2-X-3-aryl- sulphonylbutanes re~pectively~~ (16 and 17; X = OBs or I). However the stereo- specificity in this particular system is not necessarily indicative of a concerted mechanism for carbanions generated from a-arylsulphonyl compounds are 75 E.D. Hughes and J. C. Maynard J. Chem. SOC. 1960 4087. 76 P. S. Skell and J. H. McNamara J . Amer. Chem. Soc. 1957 79 85; F. G. Bordwell and P. S. Landis ibid. p. 1593. 505 Quarterly Reviews known to retain their config~ration,~~ and if formed in these eliminations would be expected to decompose by an anti-stereospecific pathway. ArSO H*: ArSO Conclusion It can be seen that clear examples of the carbanion mechanism are rare. Of the criteria we have discussed stereochemistry of elimination is at present the least reliable since there is now evidence that E2 reactions are not necessarily anti stereospecific. This means that it no longer suffices to conclude from stereo- chemical findings alone that a mechanism other than E2 is operative when anti elimination does not prevail. Many of the reactions that proceed with unusual stereochemistry and to which the carbanion mechanism has conse- quently been assigned must therefore be re-examined by use of other available criteria before the mechanistic question can be considered solved. I thank Professors J. F. Bunnett and B. D. England and Drs. R. A. Bartsch and M. D. Can for interest and helpful criticism. I also thank the Victoria University of Wellington New Zealand and the University of California Santa Cruz U.S.A. for facilities. '' D. J. Cram D. A. Scott and W. D. Neilsen J. Amer. Chem. Soc. 1961 83 3696.

ISSN:0009-2681    

DOI:10.1039/QR9672100490

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Electronic Properties of Binary Compounds of the First-row Transition Metals By A. T. Howe* and P. J. Fenshamf CHEMISTRY DEPARTMENT UNIVERSITY OF MELBOURNE AUSTRALIA 1 Introduction Binary Compounds The first-row transition metals by combining with non-metal atoms (particularly those of Group VI) form solid binary compounds which now number more than one hundred and are listed in the Table. Despite the similarities of the composition and structure of these compounds a quite remarkable range of electronic properties is displayed. Their electronic conductivities even for the stoicheiometric compounds range from metallic through semiconducting to insulating ; their magnetic behaviour includes examples of Pauli paramagnetism and ferro- antiferro- and ferri-magnetism. Many of these compounds are the products of tarnishing of the parent metals and their conducting properties influence the progress of these reactions and have been basic to their theoretical exp1anation.l Likewise the compounds appear very commonly as heterogeneous catalysts and again their electronic properties are involved in the basic chemisorptive steps that underlie their catalytic activity.Indeed for a considerable period the conducting properties or often the models that were thought to apply to them determined the direction of much catalytic re~earch.~,~ In the last few years after some disillusionment with the results of this approach interest has turned to the relation between activity and other electronic properties such as spectra6e6 and magnetic behaviour.' Besides their involvement in these chemical processes the compounds (or modifications of them formed by doping non-stoicheiometry or combination as in the ferrites) are of increasing interest in the applied areas of solid-state physics and electronics.Notwithstanding the significance of these compounds inorganic chemistry has tended to neglect them in favour of more complex co-ordination compounds of these metals. Essentially this is because the d electron properties of the latter compounds are similar in solution and in the solid state whereas in the simple * Present address Chemistry Department University of Newcastle-upon-Tyne. Present address Faculty of Education Monash University Clayton Victoria Australia. K. Hauffe Progr. Metal Physics 1953 4 71. G. Parravano and M. Boudart Adv. Catalysis 1955 7 47. K. HauEe Adv.Catalysis 1955 7 213. (a) J. D. Cotton and P. J. Fensham Trans. Faraday SOC. 1963,59,1444; (b) H. B. Charman R. M. Dell and S . S . Teale Trans. Faraday SOC. 1963 59 453. D. A Dowden and D. Wells Actes du Deuxitme Congr. Int. de Catalyse Technip Paris 1961 p. 1499. 6 J. Haber and F. S . Stone Trans. Faraday SOC. 1963 59 192. 7 K. S. De M. J. Rossiter and F. S. Stone Proc. 3rd Internat. Congr. Catalysis North Holland Amsterdam 1965 p. 520. 507 Quurterly Reviews binary compounds co-operative electronic properties exist as a special feature of the solid compounds. Nevertheless an understanding of the features of these compounds will be increasingly important as interest turns to metal-cluster compounds* and such chain systems as copper bromide. As these compounds exhibit remarkable transitions in properties while retain- ing structural simplicity they appear to offer a unique theoretical opportunity for the elucidation of these various phenomena.Although the data and the theoretical models are still somewhat limited they do warrant overall prescnta- tion and assessment. 2 The 3d Electrons and Theoretical Models In general the electronic properties are a consequence of two facts the 3d shells of the transition-metal ions are incompletely filled and there is some measure of overlap of 3d orbitals with the surrounding anion and cation orbitals. A consideration of electronic levels indicates that upon compound formation the 4s electrons from the metal atoms will transfer to fill the p orbitals of the anionic atoms thus leaving the partially filled 3d orbitals outermost.The degeneracy of such 3d orbitals is lifted by interaction with the negatively charged anion ligands in the crystal.1° Figure 1 shows a cation octahedrally surrounded by anions as in the case of the NaCl structure exhibited by the oxides TiO-NiO where the d electrons have lower energy if they face between the anions (tZg orbitals) and higher energy if they face towards them (e orbitals). However unlike the co-ordination compounds of these transition metals the binary compounds have interionic distances which may permit extended overlap amongst these cation orbitals and also favour overlap between the orbitals of the anions. Accordingly two theoretical models have been used in discussing their electronic properties. When appreciable overlap occurs between orbitals band theory describes the delocalised electrons in terms of wave functions (Bloch functions) which extend throughout the solid and have equal amplitudes at equivalent lattice sites.On the other hand the Heitler-London approach de- scribes localised orbitals which only overlap to a very small extent in terms of a collection of wave functions each of which is centred on one particular ion in the 1attice.ll The example of Ti0 (d2 if the above electron transfer is assumed to be com- plete) best illustrates the effect of delocalisation on the electronic properties. The observed metallic conductivity and temperature-independent susceptibility (Pauli paramagnetism)12 are characteristic of extensively delocalised electrons which result from the overlap of the partially filled tzg orbitals on one cation * F.A. Cotton Quart. Rev. 1966 20 389. lo D. E. O'Reilly J. Chem. Educ. 1961 38 312. l1 J. B. Goodenough. 'Magnetism and the Chemical Bond' ed. F. A. Cotton Interscience C. G. Barraclough and C. F. Ng Trans. Furaduy SOC. 1964 60 836. New York 1963 ih.. 1. l2 (a) S. P. Denker J . Appl. Phys. 1966 37 142; (b) A. A. Samokhvalov A. G. Rustamov Soviet Phys. Solid State 1963 5 877; (c) F. J. Morin Phys. Rev. Letters 1959,3 34. - 508 Howe and Fensham Fig. 1 The NaCl lattice showing some of the orbitals present in the transition-metal oxides Ti0 to NiO. The tZ8 orbitals which lie between the anions are favourably oriented for cation- cation overlap such as is found in TiO. When the eg orbitals are occupied overlap with the p orbitals of oxygen causes an antiferromagnetic coupling of each cation spin dipole with those on rhe next-nearest-neighbour cations as found in NiO and indicated on the diagram with those on the twelve adjacent cations.The overlap is sufficient to enable delocalisation analogous to that found for the 3s electrons in metallic sodium. Band theory predicts that for such a collection of electrons moving in the periodic potential of the cation lattice the energy levels which are specified by a linear momentum quantum number k will be very closely spaced up to the number of levels necessary to accommodate two 3s electrons from each atom present. Beyond these levels which are referred to as the ground state or valence band a forbidden region exists until another series of levels termed the excited- state band begins.13 Figure 2 shows the energy of the electrons plotted against k (drawn as if k were continuous) together with a representation of the half-filled band ofsodium.Under the influence of an external electric field high electronic conduction is possible owing to the presence of empty levels in the band. In the alkaline-earth l3 C. Kittel ‘Introduction to Solid State Physics’ 3rd edn. Wiley New York 1966 ch. 9. 509 Quarterly Reviews metals this band is filled by two s electrons from each atom but conduction is still possible because other empty bands in fact overlap the filled band. When two types of atom constitute the lattice a more complex situation exists. For covalent non-transition-metal compounds such as InAs which are also described by the band model the valence electrons lose their s or p character and form one filled band delocalised over both the indium and arsenic cores as represented in Figure 2c.14 E x c i t e d - ~ ' o n ' y state band 1 1 Empty band +]Empty band (6) Na a metallic conductor (c) InAs,a semiconductor Fig.2 Plot of energy (E) against wave vector (k) and schematic energy levels in band-type compounds In the transition-metal compound considered here the valence electrons (s and p ) again form a band but in the oxides which are predominantly ionic the band has a high density of states around the anions which could be described as a purely anion band of overlapping filled 2p orbitals. The d electron levels lie between this and the excited-state band which now has a very high energy. The extent of overlap of the d levels control magnetic and conductivity charac- teristics as is evident from the example of TiO the electronic structure of which is indicated in Figure 3.@)Ti0 (d2) Cb,V,OJd2) (c) N i0 (0'') above transition temp. 2p(filled) Fig. 3 Schematic energy levels showing the narrow partly filled d bands present in Ti0 and V,O (above transition temperature) and the discrete d levels in NiO. The empty excited-state bands lie above the e levels l4 0. Madelung 'Physics of 111-V Compounds' Wiley New York 1964 ch. 2. 510 Ho we and Fenshanz As will be seen in Section 7 NiO shows markedly different conductivity characteristics from Ti0 owing to localisation of both the e and t2g electrons which are best described by the Heitler-London approach. Morin15 has attributed this reduction in overlap of the tzs orbitals to the contraction of the orbitals onto the cation sites as the nuclear charge increases across the series resulting in a narrowing of the tag band-width until discrete levels are present (Figure 3c).Conductivity will now occur by jumps of electrons from one discrete level to another unoccupied one. 3 Magnetic Properties The known magnetic properties of these compounds shown in the Table and derived primarily from Goodenough’s tabulations,16 fall into two main groups those without and with spontaneous magnetic moments. Except for VO which has Curie-Weiss paramagneti~rn,~’ the first group consists of those with Pauli paramagnetism such as TiO TiN Tic TiSe and TiTe. The absence of a magnetic moment except in an applied field is due to the absence of magnetic interactions between electrons in an extensively delocalised band.Many of this group are superconductors below l O 0 ~ l 8 and all are metallic conductors above it. The second group the remainder of the compounds exhibit their spontaneous magnetisation as antiferro- ferro- or ferri-magnetism owing to the strong magnetic interactions which can arise either between electrons which are not as delocalised as in the first group while still permitting metallic conduction (the chalcogenides) or between localised electrons (most of the oxides). Indeed for the whole of the second group it is not possible to deduce apriori the extent of orbital overlap from the occurrence of spontaneous magnetisation. In the oxide series MnO to NiO magnetic interactions occur first between the 3d electrons on each cation to produce a high spin configuration (Hund’s rule) and secondly between these localised dipoles to give a periodic antiparallel align- ment throughout the lattice.The origin of both these effects lies in the quantum mechanical exchange energy which is due to the change in the Coulomb energy resulting from different spatial distributions of charge for different spin states. For alignment of the cation dipoles the energy is given by -2JcSi.Sj where Si and Sj are the electron spins and J is the exchange integral which is negative for antiferromagnetism and positive for ferromagnetism (parallel coupling).lg Ferrimagnetism results from a non-equivalent antiparallel coupling. Neutron diffraction20 reveals that for these oxides the exchange interaction is not between t2s electrons on adjacent cations which are evidently too localised but between the e electrons on next-nearest neighbours via the intervening oxide ion.Each cation thus has six next-nearest-neighbour dipoles aligned antiparallel l5 F. J. Morin Bell System Tech. J. 1958 37 1047. l6 Ref. 11 p. 98. la B. T. Mathias Rev. Mod. Phys. 1963 35 1. l9 A. H. Monish ‘The Physical Principles of Magnetism’ Wiley New York 1965 p. 275. 2o C. G. Shull W. A. Strauser and E. 0. Wollan Phys. Rev. 1951 83 333. S. Kawano K. Kosuge and S. Kachi J. Phys. SOC. Japan 1966,21 2744. 51 1 Crystal structure und electronic properties of binary (and sunie ternary) transition-metal campounds --_- - --. - afCr,S6 Cr,Se6 ‘Cr,Te# Fe,Se PPTiSe,. --. - $3. afCr,Se ,Cr,Te4 Fe,S “Fe,Se CoSe,. _. _---. Cr,Ss Cr,Ses Yk2Te3 - I NiAs I NiSl.,p NiSe,. 4 Crystal structure Magnetic properties Conduction properties Where known these are indicated by underline - metallic conductors ; --..metallic to semiconducting transition observed; *A magnetic transition is exhibited (refer to Section 3 of text). +A. D. Wadslcy in “Non-Stoichiometric Compounds” ed. L. Mandelcorn Academic Press New York 1964 ch. 3. The formulae for the oxides and the chalcogenides with the NiAs structure represent distinct phases which have been identified by X-ray diffraction. and which onl! exist over a small composition range unless otherwise indicated? (by a line joiiing the phase limits). The nominal formulae appear for the remaining compounds of whicl only a selection are tabulated. The Magneli series TinOqn- and VnOZn- (n = 4 to 8) have been omitted. Where !mown these are indicated by d diamagnetic; p Curie-Weiss paramagnetic; ’pp Pauli paramagnetic; f ferromagnetic; af antiferromagnetic; fi ferrimagnetic.-_. semiconductors. Ti,O is metallic or intrinsic semiconducting with very low Eq. Howe and Fensham as shown in Figure 1. Anderson21 has called this mechanism superexchange and it arises from a small amount of overlap of the e orbitals and the neighbouring oxygenp orbitals such that the spins of the two eg orbitals are coupled owing to spin polarisation of the lobes of the oxygen orbital. Particularly when the M-O-M angle is 180" the exchange energy is consider- able as shown by the Nkl points (TN'K temperature above which thermal vibrations destroy the magnetisation) which range from 122"~ for MnO to 523"~ for NiO. Calculations based on superexchange can give TN within 10" for these oxides which supports the use that Goodenough22 has made of the mechanism to explain insulating oxides in general.For compounds whose structures are more complicated than NaCl he has suggested that the exchange will be made up of a combination of M-O-M superexchange (either (T type overlap of e,-p-e orbitals or n- type overlap of tzg-p-tzg orbitals) direct cation- cation exchange of t2,-t2 orbitals and indirect exchange through a mutually adjacent anion (M-O-M angle = 90"). Thus in the corundum series the anti- ferromagnetism of a-Cr203 is seen to arise from mainly cation-cation exchange linkages since there is no eg electron while that of oc-Fe,03 with half-filled e orbitals is due to M-O-M super-exchange at 135". Similarly the same con- siderations are used to explain the behaviour of oxides with spinel (e.g.Fe,04) and perovskite (e.g. NiO,Fe,O,) structures which for the examples quoted display ferromagnetism and ferrimagnetism respectively. However it is difficult to see what such exchanges mean in the complex NiAs structure which is assumed by most of the simple chalcogenides like MnTe (antiferromagnetic) Cr,Te8 (ferromagnetic) and Fe,S8 (ferrimagnetic). Good- enough has recognised that greater delocalisation of 3d electrons exists in some of these compounds than is allowed in his exchange mechanism between essen- tially localised electrons and has postulated that spontaneous magnetisation can also arise from exchange between electrons some of which are localised and some delocalised. This may be the beginnings of a theory which will explain the conducting and magnetic properties of such metallic antiferromagnets as Cr,.,,Te.23 Some successful calculations have been done by Vonsovsky and T ~ r o v ~ ~ for metallic iron with this model involving direct exchange of cation- cation linkages as well as delocalised electronic levels.However for metallic a purely band approach involving interacting spin bands has proved more successful and it must be concluded that the understanding of these magnetic interactions is still at a very elementary stage. 4 Conductivity General Features At present conductivity measurements provide the most direct information about the electronic structures of the compounds. The Table gives the type of conduc- a1 P. W. Anderson in 'Solid State Physics' vol.14 ed. F. Seitz and D. Turnbull Academic Press New York 1963. aa (a) Ref. 11 ch. 3; (b) J. B. Goodenough Phys. Rev. 1960 117 1442. a3 F. Gronvold and E. F. Westrum Z . anorg. Chem. 1964 328,272. 24 S. V. Vonsovsky and E. A. Turov Soviet Physics 1953,24,419. Ref. 19 p. 300. 513 Quarterly Reviews L_3 Nitrides and Carbides of TI and v series L=- __j L-. J Group Vand VI Compounds with NlAs or MnP Structures Group V and VI Compounds with Morcasite,Arsenopyrites OT Pyrites Structures T i 0 V O ; and V203 VO above transition temperature C-ZIZ3 i-__-__1 T1O,(dO) v20 (do) L - T - ~ 1 - 1 - ~ _ ~ 1 V,O,VO below transition temperature t-T- - __l C ~ O (duo). ZnO ( d " ) FeO ,COO N I 0 CFZ--ET X 7 - - - - - - _ ? ! $--I- &-I -+ - c 4- - - -L-&- A- 5 4 3 2 I 0 -I -2 -3 - 4 - 5 -6 -7 -8 -9 -10 ;I1 -12 - 1 -14 €3 Alkali metals log CM") L I Z 3 Transition metals --_ - __ -ZIIZIEl Si,Ge,lnAs Q u a r t z .porcelain LY METALLIC CONDUCTORS . SEM I CONDUCTORS INSULATORS Conduction vio bonds El117 Conduction viu localised levels Fig. 4 Room-temperature conductivity of pure substances and those for which the conductivity has been increased by the solid solution of impurity ions (up to 10% in some cases) or by the presence of non-stoicheiometric defects.a Transition-metal compounds appear above the scale aFor a treatment of this topic see F. A. Kroger 'The Chemistry of Imperfect Crystals' North Holland Amsterdam 1964 tivity exhibited and Figure 4 shows the observed conductivities at 2 9 8 " ~ ( C T ~ ~ ~ " K ) of the salts both when pure and when doped either by the solid solution of impurity ions (up to 10% in some systems) or by non-stoicheiometric defects.The temperature-dependence of these values shows certain marked charac- teristics the significance of which is discussed below. If the conductivities could be measured near O'K the compounds would fall into two groups. Those referred to as metallic conductors would have a conductivity greater than their room-temperature value while all other compounds would be insulators. As the temperature is raised the conductivity of the group labelled semi-conductors increases owing to thermal excitation of electrons from trapping centres or from the normal ground state into current-carrying levels. At high temperatures or with heavily doped samples the conductivity may approach that of the metals and it is this wide range of possible conductivities which characterises the semi- conductors from the metallic conductors and insulators as can be seen in Figure 5 for Ti0 and NiO.At any temperature o = n l e l p (1) where n is the number of charge carriers present e is the electronic charge and p the mobility is the rate at which they move under unit field gradient. n can be determined from measurements of the potential (VH) induced across a conductor carrying a current I by the application of a magnetic field H. This potential is known as the Hall voltage and its magnitude is given by where the Hall constant VH = &I X H (2) 514 Ho we and Fensham where c is the velocity of light and the + and - signs correspond to electrons or positive holes as the current carriers.26 The combination of a and VH also enables p ~ the Hall mobility to be calcu- lated and it is from n and p and their temperature-dependence that the electronic structure is inferred.Metallic conductors have values of n which are independent of temperature and equal to the number of electrons in the partially filled band while semiconductors have n values which increase exponentially with tempera- ture from zero at OOK as thermal excitation of the electrons produces current carriers. In discussing the significance of the mobilities two models of transport have to be considered movement through bands and jumps between discrete elec- tronic levels Transport in metallic conductors can only be explained by the band model but in semiconductors both can apply. In the following sections the conductivities and mobilities will be interpreted in terms of the electronic structure described by these models.5 Conductivity in Metallic Conductors Of the many metallic conductors in the Table the series TiO TiN and Tic has received most attenti~n.~' They are referred to as refractory hard metals because in addition to their conductivity and magnetism they have physical similarities to the parent metals,28 for example Tic and TiN like titanium melt at above 1 8 0 0 " ~ . However they are not to be regarded as metallic alloys because the bonding involves distinct ionic and covalent character. In the band model of conduction which applies to metallic conductors such as these the acceleration of the electrons through the energy levels of the crystal is described by the concept of the effective mass ratio rn*/rno which relates the acceleration possible in the periodic potential of the lattice to that of a free electron.A large effective mass ratio is associated with a small acceleration as found in narrow bands where the interaction of the electrons with the periodic potential is appre~iable.~~ The other concept used to characterise these band conductors is the relaxation time T which refers to the frequency of interruption of the electronic accelera- tion by scattering. Mobility relaxation time and effective mass are then related by er m c L = < (4) Finally the product of the velocity and T specifies the mean free path A of the current carriers. In general two main scattering processes occur in metallic conductors. At low temperatures (usually not greater than 2 0 " ~ for metals) p is proportional to TP owing to scattering by impurities in the lattice.At higher temperatures the 26 P. N. Shive 'Simiconductor Devices' Van Nostrand New York 1959 ch. 24. 27 (a) T. Tsuchida Y. Nakamura M. Mekata J. Sakurai and H. Takaki J. Phys. SOC. Japan 1961 16 2453; (6) W. S. Williams Phys. Rev. 1964 135 A 505 t* E. Dempsey Phil. Mag. 1963 8 285. 2D Ref. 26 ch. 15. 515 Quarterly Reviews I L 6 -A to- 1 I - 1 2 1000 I t - .- .. - 1 I 1 (single I crystal) j I Fig. 5 Electronic conductivity as a function of temperature for various oxides [From the data of S. P. Denker (TiO);lZa F. J. Morin (VO, V203);12c and S. Koide J. Phys. SOC. Japan 1965 20 123 (doped NiO)] When it is assumed that the relaxation time of the conduction electrons in Ti0 is the same as for the metals an effective mass ratio of 10 is deduced from the calculated Hall mobility of 1.0 cm.2 v-l s e .~ . - ~ ~ ~ which supports transport in 30 S. P. Denker J . Phys. and Chern. Solids 1964 25 1397. 31 Ref. 12b. 516 Howe and Fenshnrn a narrow band (cf. m*/m = 0-98 p = 21 cm,2 v-l sec.-l for free electrons in the wide band of sodium). 6 Transition from Delocalised to Localised Electrons As indicated in Section 3 the change from Pauli paramagnetism to spontaneously magnetised systems does not simply correspond to a transition from delocalised to localised electronic states. However in this section it will be seen that such a correspondence can be established for the change from metallic to non-metallic conduction. M ~ t t ~ ~ has predicted that as the nuclear separation of a delocalised system is increased a sharp transition to the localised state occurs with a con- sequent drop in conductivity.Two examples provide evidence for this relation- ship. The first has already been implied in Section 2 namely the transition from metallic conduction in Ti0 to semiconducting or insulating behaviour in NiO. Successive localisation occurs owing to the increase in interaction of the electrons with the nuclear charges across the series and M ~ r i n ~ ~ has shown that the metallic compounds Ti0 and the high-temperature forms of VO V,03 and VO, have more tzg-t2 and more n-type t2,-p overlap than the insulators MnO FeO and Cr203. Secondly such delocalisation in V20 and VO is abruptly destroyed at 160"~ and 3 4 0 " ~ respectively as shown by a transition to semiconducting behaviour and a change in the magnetic state.Figures 5 and 6 summarise the experimental evidence. Goodenougha attributes this change to a crystallo- graphic distortion involving the pairing of the cations so that the electrons become trapped in metal-metal pair bonds. Fig. Magtietic susceptibility as a function of temperature for various oxides [From t..e data of A. A. Samokhvalov and A. G. Rustamov (Ti0);126 K. Kosuge (VOa);36 P. H. Carr and S. Foner J . AppI. Phys. Sirppl. 1960 31 3443 (V,O,); and J. R. Singer Phys. Rev. 1956 104,929 (NiO)] 32 N. F. Mott Phil. Mag. 1961 6 287. 33 F. J. Morin J. Appl. Phys. Suppl. 1961 32 2195. 34 Ref. 1 1 p. 256. 517 Quarterly Reviews In the corundum structure assumed by Ti203 as well as V203 the trigonal ligand field splits the tzs orbitals into those which lie in the basal planes of the cation octahedra (r,,) and those which point above and below these planes in the direction of the c axis (r,,) as shown in Figure 3.Pairing can occur in both these directions. In the case of Ti203 (dl) the c-axis electrons are not de- localised even at high temperatures and pairing which occurs over a wide temperature range (500"-350"~) only alters the conductivity slightly. How- ever V,03 (#) in addition to having one electron per cation in the localised c-axis bonds has one electron per cation in the basal-plane orbitals. Hence c-axis pairing which occurs between 500°K and 400'~ does not alter the con- ductivity but the transition of electrons in the basal plane from delocalised to paired at 1 6 0 " ~ causes the conductivity to drop by a factor of lo6 within a degree and the magnetic properties to change from paramagnetic to antiferromagnetic.Before such detailed crystallographic data were available the transition in conductivity was interpreted by M ~ r i n ~ ~ as due to the onset of antiferrornag- netism which split the partially occupied d bands into non-conducting spin or magnetic sub-bands. However in the case of VO (dl) K ~ s u g e ~ ~ observed that the Mossbauer line of Fe doped into VO crystals remained unsplit below the transition temperature (340"~) revealing that the magnetic transition is from paramagnetic to diamagnetic rather than to antiferromagnetic. This supports the contention of Go~denough~~ and Adler and Feinleib3' that cation pairing causes the transition in both magnetic and conductivity behaviour.The characteristics of the semiconductivity observed below the transition temperature in these oxides provide a clue to the electronic structure of the nominally unoccupied higher-energy d levels. Owing to an excess of 1 part in lo5 of vanadium in VO, n-type semiconductivity occurs as determined from the negative sign of both the Hall3* and the Seebeck c~efficient.~~ It can be accounted for by the thermal promotion of electrons from their trapping centres into an unfilled band which Gooden~ugh~~ attributes to 7r type overlap between tzs-p-tzg orbitals in the plane perpendicular to the V-V pairs. The room-tem- perature mobility of 0.1 cm.2 v-l sec.-l indicates that this is a narrow band. For non-stoicheiometric semiconductors electrons are said to be trapped when they are localised in the positive potential of a vacant anion site or localised on an interstitial cation forming V3+ in this case.The number of untrapped electrons available for current carrying is given by where ED is the trapping energy given by the slope of the plot of log CJ against 1 IT the temperature-dependence of the mobility being assumed small compared 35 F. J. Morin Phys. Rev. Letters 1959 3 34. 36 K . Kosuge J . Phys. SOC. Japan 1967 22 551. 37 D. Adler and J. Feinleib Phys. Rev. Letters 1964 12 700. 38 I. Kitahiro T. Ohashi and A. Watanabe J. Phys. SOC. Japan 1966 21 2422. 39 I. Kitahiro and A. Watanabe J. Phys. SOC. Japan 1966 21 2423. * O J. B. Goodenough Bull. SOC. ckim. France 1965 1200. 518 Howe and Fensham with that of nmobile.41 Figure 5 shows this plot to have the expected negative gradient for the semi-conducting regions of the oxides considered.TiO (do) rutile does not show a crystallographic transition like that of VO,. When it is slightly oxygen-deficient the n-type conductivity which occurs can be interpreted in terms of electron promotion from the traps first to one then to a still higher unfilled band.42 This is evidence of an electronic structure similar to that of VO, where two such bands were indicated by transport in the metallic and semiconducting regions respectively. Detailed information is not available for the Magneli series V,O2,_ (n = 4 to 8) which exhibit semiconductivity without a transition to the metallic ~tate.4~ An abrupt change in the magnetic susceptibility occurs in some cases.44 Similar information regarding the electronic structure of already occupied d levels is obtained from p-type semiconduction which occurs in oxygen-excess structures where the valence of the cations can easily be increased.For example Cu20 @lo) as distinct from those p-type oxides with localised d electrons dis- cussed in the next section exhibits p-type conduction in a band as deduced from the positive sign of the Hall coefficient and the negative ternperature- dependence of the rn0bility.4~ The current is carried by the positive holes Cp) formed in the otherwise full d band by the absence of electrons which when trapped can be represented as Cu2+ ions adjacent to cation vacancies. 7 Compounds with Localised 3d Electrons Oxides having a 3d structure which cannot be described by the band model are best accounted for by the localised model.This group has been reviewed by Jonker and van Ho~ten.*~ The presence of localised electrons is deduced from the absence of metallic cond~ctivity.~~ A minimum conductivity of 10 0hm-l cm.-l would be expected if delocalised electrons were present. However the series MnO to NiO which have high-spin d-electron configurations (d5 to d8) shows a conductivity of the order of ohm-l cm.-l when pure and stoicheio- metric. In addition localised states have been inferred from the similarity of the visible and ultraviolet spectra of the oxides to those of the complexed species in solution?* When non-stoicheiometric the oxides MO1+ are p-type semiconductors and this is ascribed to an exchange of electrons between M2+ ions and the M“ ions present as chemical defects as in p-type Cu,O.The conductivity should indicate features of the electronic ground state if a suitable model for transport through localised energy levels were used. Two approaches have been adopted. 41 N. F. Mott and R. W. Gurney ‘Electronic Processes in Ionic Crystals’ Clarendon Press Oxford 2nd edn. 1948 ch. 5. 43 J. H. Becker and W. R. Hosler Phys. Rev. 1965 137 A 1872. 43 S. Kachi T. Takada and K. Kosuge J. Phys. SOC. Japan 1963 18 1839. 44 K. Kosuge T. Takada and S. Kachi J . Phys. Sac. Japan 1963 18,318. 45 M. O’Keefe Y. Ebisuzaki and W. J. Moore J . Phys. SOC. Japan Suppl. 2 1963,18 131. 46 G. H. Jonker and S. van Houten ‘Halbleiterprobleme VI’ ed. F. Sauter Friedr. Vieweg und Sohn Braunschweig 1961 p. 118. 47 N. F. Mott Proc. Phys. SOC. 1949 62 A 416.48 R. Newman and R. M. Chrenko Phys. Rev. 1959,114,1507. 519 4 Quarterly Reviews Heikes and Johnston49 recognised that when a Ni2+-NiS+ pair exchange an electron the distortion of the surrounding anion charges alters so as always to be centred about the Ni3+ ions and this contributes an activation energy to the electron jump. The mobility p is then given by where a is the closest cation-cation distance; vo is the jump frequency of the order 1 Ola sec.-l the frequency of lattice vibrations suggesting that jumps occur when vibrations cause the mutual approach of the cations. On the other hand Holsteinso has presented a formal treatment of a current carrier together with its accompanying lattice polarisation and for this pair the name polaron is used. At temperatures greater than half the Debye temperature of the lattice charge transport is predicted to be by activated jumps and mobility to increase with temperature (as for Heikes and Johnston's approach).Lattice vibrations are such that polarisation effects are restricted to a few lattice spacings around the Ni3+ and the polarons in such circumstances are described as small. However at lower temperatures the polarisation extends over a large distance and the polarons interact with each other to form what is called a band of polaron states so that charge transport now has a mobility which increases with decreasing temper at ure. When attempts were made to use experimental evidence to distinguish these two approaches,5l difficulties arose. First transport measurements on powdered oxides are difficult and not always reproducible particularly when the con- ductivity is low.To minimise these difficulties the conductivity in compounds such as NiO has been increased by high dope concentrations of Li20 added to form a solid solution of Li+ and the complementary Ni3+ ions in the NiO lattice.52 However the presence of Li+ raises other uncertainties. In addition the theore- tical significance of the Hall voltage and the Seebeck effect in these jump-type and antiferromagnetic compounds has only recently been investigated and is not as clearly understood as the band-type conductor^.^^ Improved experimental methods and the use of single crystals of NiO with only low Li20 concentrations have now led to results which can be reliably disc~ssed.6~ Seebeck measurements above 100"~ together with conductivity data have been interpreted by Heikes and Johnson to favour the simple jump process.After some approximations the Seebeck coefficient is found to vary inversely as the number of charge carriers and its constancy above 5 0 0 " ~ on this theory 49 R. R. Heikes and W. D. Johnston J. Chem. Phys. 1957,26,582. 50 T. Holstein Ann. Phys. New York 1959 8 325 343. 51 A. J. Bosman and C. Crevecoeur Phys. Rev. 1966,144,763. 52 E. J. W. Verwey 'Semiconducting Materials' Butterworths London 1951 p. 151. 53 (a) L. Friedman and T. Holstein Ann. Phys. 1963 21 494; (6) L. Friedman Phys. Rev. 1963,131 2445; (c) E. R. Heikes in 'Informal Proceedings of the Buhl International Confer- ence on Materials' Pittsburgh 1963 ed. E. R. Schatz Gordon and Breach New York 1966 p. 1. 54 S. P.Mitoff J. Chem. Phys. 1961 35 882. 520 Howe and Fensham means that any activation energy present in the conductivity (see Figure 5 ) must arise from the mobility term. Amobility of lo4 r1 sec.-l with an activation energy q ranging from 0.1 to 0.5 ev was deduced. The jump process also finds support from the presence of high-frequency relaxation losses in NiO samples over the range 4" to 300"Kt5 although these results could be due to jump con- duction involving impurity defects rather than the Ni2+-Ni% exchange. However there is some experimental evidence that favours Holstein's polaron view if the difficulty of interpreting Hall constants in such complex magnetic materials is set aside. For NiO the temperature coefficient of the Hall mobility is negative between 200" and 5 0 0 " ~ which accords with movement in a polaron band.56 There is a discontinuity in the Hall voltage through the Niel tempera- ture and above TN it takes on as yet unexplained negative values.The conduc- tivity on the other hand shows only a slight change in slope at TN and Heikes5' accounts for this in terms of the magnetic effects on the jump mobility. 8 Chdcogenides and Pnictides Of the chalcogenides and pnictides (Group V) of the transition metals those having the NiAs and the closely related MnP structure (Table) have been most inve~tigated.~~ All exhibit spontaneous magnetisation and their characteristically high room-temperature conductivity lies between loa and 105 0hm-l cm.-l (Figure 4). These values indicate conduction in bands since they can only be explained by assuming high mobilities (1 to 100 cm.2 v-l sec.?) such as are only possible in the band scheme.As these compounds all possess a large natural deviation from stoicheiometry (of the order of l%) the conductivity may be high even though it may not be of the metallic type so it is difficult to distinguish between metallic conduction and semiconduction in a band with a large number of current carriers. Hall measurements cannot as yet readily be interpreted to enable the number of charge carriers to be calculated as there is an additional component to the Hall voltage in such magnetic corn pound^.^^ It has been inferred however that of compounds investigated only Cr3Te4,6* MnTe,S1 FezTe3,62 and Fe,S23 are semiconductors while some sulphidess4 and selenide~~~ of chromium exhibit a semiconducting to metallic transition (Table).s6 (a) S. van Houten J. Phys. and Chem. Solids 1962 23 1045; (b) S. van Houten and A. J. Bosman Ref. 53c p. 123. s6 I. G. Austin A. J. Springthorpe and B. A. Smith Phys. Rev. Letters 1966 21,2O. 67 Ref. 53c. 68 (a) W. Albers and C. Haas in 'Proceedings of the Seventh International Conference on the Physics of Semiconductors' Paris 1964 Academic Press New York 1965 p. 1261 ; (6) S. Fujime M. Murakami and E. Hirahara J. Phys. SOC. Japan 1961 16 183; (c) K. van Con and J. P. Suchet Compt. rend. 1963 256 2823. Kg (a) R. Karplus and J. M. Luttinger Phys. Rev. 1954 95 1154; (6) M. Nogami Jap. J. Appl. Phys. 1966 5 134. 6o J. P. Suchet and P. Imbert Compt. rend. 1965 260 5239. 61 (a) J. D. Wasscher A. M. J. H. Seuter and C. Haas Ref. 58a p. 1269; (b) H. Yakada T.Harada and E. Hirahara J. Phys. SOC. Japan 1962,17,815. 6e F. Aramu and P. Manca Nuovo Cimento 1964,33 1025. A. Theodossiou Phys. Rev. 1965 137 A 1321. T. Kamigaichi K. Masumoto and T. Hihara J. Phys. SOC. Japan I960,15 1355. K. Masumoto T. Hihara and T. Kamigaichi J. Phys. SUC. Japan 1960,1!5,12#. 521 QuarterZy Reviews The origin of the high mobility and metallic conductivity in most of the com- pounds is not yet well understood. The possibility exists of metallic conduction through bands formed by the overlap of the partially filled tzs levels which are split into two sub-levels as in the corundum structure of V,O (Figure 3). Moreover Goodenoughs6 has shown that overlap is favoured when the cations are displaced slightly into the easily accessible tetrahedral sites which may be occupied by up to 10 % of the cations resulting in a high conductivity.However this explanation alone could not apply to the nickel salts which have filled tzg levels and Suchet6' and Albers and Haas68 recognising the marked covalence of these compounds postulate an overlapping of either the broad 4s cation or the broad p anion band with the narrow d bands. This leads to a redistribution of the electrons amongst the bands so that transport can occur with a greatly increased mobility. Hulliger and M o o ~ e r ~ ~ have studied the series formed from transition metals and anions such as S Z - PS3- and P2& a selection of which is shown in the Table. From consideration of crystal-field splittings they have been able to relate both the crystal structure and the conductivity to the number of d electrons present.These compounds like the previous group have very high conductivity ranging from 1 to lo4 ohm-l cm.-l at room temperature again indicative of movement in overlapping d-s or d-p bands. On the basis of the Seebeck co- efficient and the temperature-dependence of conductivity Hulliger and Mooser inferred that only the structures with the d7 configuration are metallic. Here there are six non-conducting tzg electrons above which lies an unfilled band containing the additional one electron per cation. Superconductivity is found amongst this group,18 which are expected to be Pauli-paramagnetic. Semiconductivity occurs in the remaining compounds where metallic conductivity is prevented by the accommodation of the d electrons in exactly half-filled high-spin bands as for MnS (d5 pyrites structure) or in completely filled d sub-bands (d4 d6 marcasite structures and d5 arsenopyrites structure).In the marcasite compounds a dia- magnetic state results from the large crystal-field splitting caused by the covalent anions. In the d5 arsenopyrites structures additional cation-cation pairing exists. Both these and the NiAs type compounds provide a fertile field for investiga- tions towards a more definite description of their unique conductivity patterns. 9 Halides Halides of the bi- and ter-valent transition metals have antiferromagnetic NCel temperatures below 100"~ and for some such as MnF, the magnetic structure has been extensively in~estigated.'~ However very little has been reported on their electronic conductivity so their electronic structure is unknown.66 Ref. 11 p. 276. 67J. P. Suchet Physica Status Solidi 1966 15 639. 68 W. Albers and C. Haas Phys. Rev. Letters 1964 8 300. 6* F. Hulliger and E. Mooser J . Phys. and Chem. Solids 1965 26 429. H. Bizette and B. Tsai Compt. rend. 1954 238 1575. 522 Howe and Fensham 10 Limitations of Current Theoretical Models Theories of electrical transport and magnetic exchange can qualitatively account for the general features of conductivity and magnetism of transition-metal com- pounds in terms of either the delocalised or localised model of the electronic structure. We must now see to what extent calculations of the energy levels can reproduce the electronic structure indicated by experiment. For electrons known to be localised the method of Heitler and London (Sec- tion 3) based on atom-centred functions should enable reasonably accurate calculations to be performed but they have only been successful in the case of the simple hydrogen molecule.When extended to a solid lattice of heavy atoms with localised electrons such as NiO the difficulty of including the large number of electron-electron and electron-nuclear interactions renders the calculations extremely diffi~ult,~~ and success is not yet in sight. These difficulties however are not inherent in the band approach where electrons are described as free and move in a periodic potential provided by all the ions in the lattice. Accurate calculations have been performed for the outer s electrons in sodium metal where use of the band approach is justified by the fact that the experimentally determined effective mass ratio is close to unity (m*/rno = 0.98 p = 21 cm.2 v-l sec.-l band-width ca.9 e ~ ) . ~ ~ For the chalcogenides and pnictides where the mobility is likewise high use of the band approach would also be valid although the complications caused by the presence of many overlapping bands and intricate crystal forms have pre- vented calculations from being performed. However for metallic oxides such as TiO low mobilities of 0.1-1 SO cm.2 v - ~ sec.-l and effective mass ratios of 10-100 indicate a significant deviation from the free-electron model as caused by an increased interaction of the electrons with the lattice potential. Although these situations can be described formally in terms of a narrow band (ca. 3 ev wide) the concept of a band breaks down when band theory is used to calculate the mean free path,73 which comes to be between 0-3 and 3-0 A.This is less than the lattice spacing in these compounds and cannot be accounted for on the basis of known scattering processes. Despite the apparent inapplicability of some band concepts to these com- pounds the qualitative success of the theory in describing the conduction and magnetic properties has been responsible for the increasing use of band calcula- tions especially since no other approach is available by which even approximate calculations can be made for solid lattices. replaced the Plane Wave Bloch functions used for sodium by a set of atomic functions modified by a periodic function (Tight Binding or Linear Combination of Atomic Orbitals appr~ach’~) and after approximating for many of the parameters obtained a crude represen- tation for TiO TiN and Tic.Ern and S~itendick~~ chose a function which 71 J. B. Goodenough Ref. 53c p. 65. 72 D. Pines Phys. Rev. 1954 95 1090. 73 A. F. Joffe ‘Physics of Semiconductors’ Infosearch London 1960 p. 424. 74 H. Bilz 2. Physik 1958 153 338. 75 J. Callaway ‘Energy Band Theory’ Academic Press New York 1964 p. 102. 76 V. Ern and A. C. Switendick Phys. Rev. 1965 137 A 1927. 523 Quarterly Reviews approximates to an atomic function within spheres around each ion but which represents a Plane Wave function between the spheres (Augmented Plane Wave APW method77). For TiO TiN and Tic these calculations indicated a 7-11 ev-wide tzs band which in the case of TiO overlaps with the 4s band and for TiN and Tic overlaps with the 2p anion band permitting a redistribution of electrons between the bands.Although these calculations do not indicate a narrow d band they explain the fractional number of conduction electrons per Ti atom found from the Hall coefficient and conductivity data.78 The band structure of Cu,O has also been estimated by this appr~ach.'~ It would not be expected that band calculations could provide a description of NiO and this has been confirmed by Yamashitaao who performed a Tight Binding calculation which predicted metallic conductivity at variance with the facts. In an attempt to include magnetic effects Switendicksl performed a more detailed APW calculation on NiO and incorporated antiferromagnetic inter- actions into the band structure using the concept of spin bands. Although this approach in which the d band is split into two spin bands containing electrons of opposite spin has had some usefulness for transition metals it does not yield real predictions of magnetic properties.However for NiO these calculations predicted the observed insulating properties. Five electrons occupy one spin band while the remaining three electrons of opposite spin occupy the other spin band which because it is split by the crystal field has full fag levels thus prevent- ing conductivity. This argument is nevertheless unsatisfactory since it cannot be applied to COO which has one less d electron but also shows insulating properties. It is clear that only very limited success has been achieved in predicting the observed electronic properties and a more accurate description of the electronic structure may need new concepts.First these must be more applicable to orbitals which do not overlap as much as in the metals and secondly these should enable magnetic interactions to be incorporated adequately. Goodenoughs2 has ex- plored the possibility of using an extended Molecular Orbital scheme with magnetic interactions between localised and delocalised electrons while Hubbardg3 has suggested that a unified many-body approach may be possible by use of Green's functions. Financial support of one of us (A.T.H.) from the University of Melbourne is gratefully acknowledged. 77 Ref. 75 p. 95. 7s J. P. Dahl and A. C. Switendick J. Phys. and Chem. Solids 1966 27 931. 8o J. Yamashita J . Phys. SOC. Japan 1963 18 1010. Group Massachusetts Inst. of Tech. 1963 p. 41. 8a Ref. 11 p. 158; Ref. 40. 83 J. Hubbard Ref. 53c p. 99. Ref. 276. A. C. Switendick Quarterly Progress Report No. 49 Solid-state and Molecular Theory 524

ISSN:0009-2681    

DOI:10.1039/QR9672100507

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

The Chemistry of the C,-Diterpene Alkaloids By S. W. Pelletier DEPARTMENT OF CHEMISTRY UNIVERSITY OF GEORGIA ATHENS GEORGIA U.S.A. 1 Introduction Diterpene alkaloids1 are widely distributed throughout the plant world and have long been of interest because of their pharmacological properties and complex structures. These alkaloids can be divided into two broad categories. The first group which is the subject of this Review includes a series of comparatively simple and relatively non-toxic amino-alcohols (alkamines) which are modelled on a C,,-skeleton. (This class is sometimes loosely referred to as the ‘atisines’. In this Review a more restricted classification will be used. The term ‘atisines’ will refer only to those compounds modelled on an atisine skeleton.) These compounds are not extensively oxygenated and contain at most one methoxy group.One of the distinguishing chemical features of this group is the formation of phenanthrenes when subjected to selenium or palladium dehydrogenation. A few of these compounds occur in the plant as monoesters of acetic or benzoic acid. The second group comprises the highly toxic ester bases (aconitines) which are heavily substituted by methoxy and hydroxy groups. Hydrolysis of the aconitines furnishes the relatively non-toxic aconines which are modelled on a hexacyclic C1,-s keleton. The Cz0-diterpene alkaloids are derived from tetracyclic diterpenes in which carbon atoms 19 and 20 are linked with the nitrogen of a molecule of /%amino- ethanol methylamine or ethylamine to form a heterocyclic ring. Thus far two different types of skeleton have been encountered the veatchine (1) and the atisine (3) types.The veatchinc skeleton which occurs in the Garrya alkaloids incorporates a phyllocladene skeleton (2) and obeys the isoprene rule. The atisine skeleton (3) differs from the veatchine type in that ring D is six- rather than five- membered; it does not obey the isoprene rule. All the Cz0-diterpene alkaloids yet encountered in Nature are constructed on these two skeletal types. In certain alkaloids such as songorine and kobusine one or more additional ring fusions are present. Detailed reviews on various aspects of diterpene alkaloid chemistry have been published by (a) S. W. Pelletier Experientia 1964 20 1; (b) S. W. Pelletier Tetrahedron 1961 14 76; (c) H. G. Boit ‘Ergebnisse der Alkaloid-Chemie bis 1960’ Academie-Verlag Berlin 1961 pp.851-905 1009-1001 ; (d) A. R. Pinder in ‘Chemistry of Carbon Compounds’ vol. IVc ed. E. R. Rodd Elsevier Amsterdam 1960 pp. 2019-2033; (e) E. S. Stem in ‘The Alkaloids’ vol. VII ed. R. H. F. Manske Academic Press New York 1960 pp. 473-503 ; (f) K. Wiesner and Z. Valenta ‘Progress in the Chemistry of Organic Natural Products’ Springer-Verlag Vienna 1958 pp. 26-89. 525 Quarterly Reviews Careful chemical investigation of the alkloids of species of Aconiturn and Delphinium was initiated by Walter A. Jacobs and his collaborators at the Rockefeller Institute in 1942. These workers carried out a number of meticulous degradations which laid the foundation for subsequent contributions in this field. In later work on the Garrya alkaloids K. Wiesner paralleled Jacobs's early studies on atisine and pointed out the striking similarity in the chemistry of the two groups of alkaloids.Advances in one field then greatly assisted pro- gress in the other. Since structures of the Garrya alkaloids were the first to be elucidated these compounds will be considered first. 2 The Garrya Alkaloids A. Vetachine and Garryine (Garrya veatchii Kellogg).2-Veatchine (4) C22H,,N02 is a strong tertiary base (pK,' 11.5) containing one acylatable hydroxyl a C-CH group and an exocyclic methylene group. On treatment with base veatchine is converted into the isomeric garryine (5) (PIC,' 8.7). Reduction of either alkaloid with LiAlH results in hydrogenolysis of the ether bridge to give the same dihydroveatchine (6) and catalytic hydrogenation of either alkaloid gives tetrahydroveatchine (7).Both dihydro- and terrahydro-veatchine show two active hydrogens and give 0,O'-diacetate derivatives.2 Selenium dehydrogenation of veatchine or garryine yields 7-ethyl-1-methyl- phenanthrene and 7-ethyl-1-methyl-3-azaphenanthrene (8).2 On the basis of these products and other degradations to be described structure (6) was suggested for dihydroveatchine. Veatchine and garryine were then formulated as the isomeric cyclic ethers (4) and (5) respectively. These structural assignments were confirmcd by pK studies and by the results of degradations which will be briefly summarised. Pyrolysis of veatchine or garryine with selenium at 290' gave the isomeric pyrolysis bases (9) and (10). Reduction of (10) with LiAlH gave the secondary base (11) which was alkylated with ethylene chlorohydrin to give dihydroveat- See ref.l(f) for a series of papers on the Garryu alkaloids. 526 Pelletier chine (6). Oxidation of the latter with one equivalent of osmium tetroxide in ether afforded garryine (9 cyclisation occurring at the less hindered 19-position. These reactions demonstrate the presence of an oxazolidine moiety in veatchine and garr~ine.~,* Oxidation of veatchine with potassium permanganate in acetone furnishes a y-lactam oxoveatchine-A (1 2) and an isomeric &lactam oxoveatchine-B (13). Under the same conditions garryine gives the &lactam oxogarryine (14). Reduction of oxogarryine with LiAlH furnishes dihydroveatchine (6).3,6 (4) - (12) R'=o R'=H (13) R'=Y,R*=o 0 Under more vigorous conditions of permanganate oxidation veatchine furnishes the lactam dicarboxylic acids (15) and (16) formed by cleavage of the five-membered ring bearing the -CHOH group.These acids are readily con- verted into the corresponding anhydrides (17) and (18) the infrared spectra of which are compatible with the presence of the y-lactam (1700 cm.-l) and 6- lactam (1640 cm.-l) respectively and of a glutaric anhydride grouping in each (1800 1762 and 1800 1775 cm.-l respectively). The dimethyl esters (no infrared hydroxyl absorption) of (15) and (16) on hydrolysis give monomethyl esters K. Wiesner W. I. Taylor S. F. Figdor M. F. BartIett J. R. Armstrong and J. A. Edwards K. Wiesner R. Armstrong M. F. Bartlett and J. A. Edwards J. Amer. Chem. SOC. 1954 K. Wiesner S. K. Figdor M. F. Bartlett and D. R. Henderson Canad. J. Chem. 1952,30 Chem.Ber. 1953,86,800. 76 6068. 608. 527 Quarterly Reviews which are resistant to further hydrolysis an indication of the hindered nature of one of the carboxyl groups in each c~mpound.~ Ac,O (4) - CO H c=o (17) R'=O R'=H (IS) ~'=y R,= o Dehydrogenation of acid (1 6) with selenium afforded l-methylphenanthrene- 7-carboxylic acid (19) and pimanthrene (20). Dehydrogenation of the trio1 (21) formed by lithium aluminium hydride reduction of the dimethyl ester of (16) also led to pimanthrene! B. Cuauchichicine and Ganyfoline.6-These alkaloids of Garrya ZaurifoZia Hartiv. are isomeric with veatchine and garryine and can be separated from each other by countercurrent distribution. The infrared spectrum of cuauchi- chicine shows the absence of -OH or -NH absorption but a strong carbonyl band at 1730 cm.? which can be attributed to a cyclopentanone ring.Kuhn- Roth determinations demonstrated the presence of two C-methyl groups and the lack of an exocyclic methylene group was confirmed by ozonisation. Selenium pyrolysis of cuauchichicine at 290" yielded a base (9) identical with that obtained earlier from veatchine by similar treatment. Moreover reduction of the alkaloid with LiAlH4 gives tetrahydroepiveatchine (23) a compound which has been synthesised from imine (9) by reduction to (24) followed by alkylation with ethylene chlorohydrin. Since the pK value of cuauchichicine (11.15) is com- parable with that of veatchine (11.5) rather than that of garryine (8.7) in cuau- chichicine the oxazolidine ring is fused at C(20) and the structure of the alkaloid may be formulated as (22).s dMe 19 C .Djerassi C . R. Smith A. E. Lippman S. K. Figdor and Y. Herran J. Amer. Chem. SOC. 1955 77,4801. 528 Pelletier Garryfoline is a tertiary base containing a C-methyl group a secondary hydroxyl and an exocyclic methylene group. The skeleton of garryfoline is the same as that of cuauchichicine (22) since it can easily be isomerised by acid to the latter compound. Garryfoline is reduced with LiAlH4 to dihydrogarryfoline (26) a transformation analogous to the conversion of veatchine (4) into dihydro- veatchine (6) and invoIving reductive cleavage of the oxazolidine ring. Like garryfoline the dihydro-derivative (26) is isomerised with acid to a ketone dihydrocuauchichicine (27) which is also obtained by catalytic reducion of cuauchichicine (22).On reduction with LiAIH4 dihydrocuauchichicine yields tetrahydroepiveatchine (28) also obtainable by catalytic hydrogenation of garryfoline itself. These transformations indicate that garryfoline has the structure of 15-cpiveatchine (25). The attachment of the oxazolidine ring at C(20) follows from a comparison of the pK,' of garryfoline (1 1.8) with veatchine (1 1 ~ 5 ) . ~ Base-catalysed isomerisation comparable with the conversion of veatchine into garryine also occurs with cuauchichicine and garryfoline. (26) d= CH d= f:, (27) R'=-Me,R2= 0 (28) R'=-Me,R*= CH The acid-catalysed isomerisation of garryfoline (25) to cuauchichicine (22) and of dihydrogarryfoline (26) to dihydrocuauchichicine (27) requires comment in view of the stability of veatchine to acidic conditions. Reduction of dihydro- cuauchichicine (27) with lithium and alcohol in liquid ammonia conditions known to lead to the more stable epimeric alcohol affords as the sole product tetrahydroepiveatchine (28) ; thus garryfoline is thermodynamically more stable than veatchine.Yet garryfoline is the epimer which undergoes rearrangement to cuauchichicine. The isomerisation can proceed only when the four centres involved are coplanar and the non-classical carbonium ion (30) from garryfoline (29) has been invoked as an intermediate in this isomerisation to the methyl ketone (31). In this cation the quasi-axial (with respect to the 5-membered D-ring) and C(15) hydrogen atom is coplanar with the bond joining C(12) and C(16).* This leads to the relative stereochemistry for garryfoline as (29); the hydroxyl groups at C(15) in garryfoline and veatchine are respectively quasi- equatorial and quasi-axial with reference to ring D? *The mechanism of this rearrangement has recently been investigated by M.F. Barnes and J. MacMillan [I. Chem. Soc. (C) 1967 3511 using the epimeric (-)-kaur-16-en-15-01~ as models. The 15/3-01 rearranges in mineral acid to 16R-( -)-kaur-15-one by a 15,16-hydride shift. The 16a-01 like veatchine is stable under these conditions. 529 Quarterly Reviews H+ __t C. Napellhe Songorine and Luciculine.-Napelline C,,H,,NO, and songorine C22H31N03 occur in Aconitum napellus L. and A. soongaricum. Songorine is a ketonic base which on reduction with LiAlH yields the corresponding secondary alcohol napelline. Each base contains an N-ethyl C-methyl and exocyclic methylene group.The carbonyl group of songorine is part of a six-membered ring (1 71 2 cm.-l). Napelline is readily isomerised to isonapelline which contains a cyclopentanone system (1740 cm.-l) and two C-methyl groups and in these respects resembles garryfoline in its tendency to undergo ketonisation in the presence of acid. The presence of a veatchine-type skeleton in these bases is indicated by formation of 7-ethyl-l-methyl-3-azaphenanthrene (8) on dehydro- genation of isonapelline. Selenium dehydrogenation of songorine gives 7-ethyl- 1,9-dimethylphenanthrene (35). The chemistry of napelline (32) songorine (33) and isonapelline (34) will be discussed in terms of the structures ultimately derived?',' Silver oxide converts dihydronapelline (36) into a weakly basic carbinolamine ether (37) pK,' 6-77 which forms a ternary iminium salt (38).Reduction of (37) with LiAlH regenerates dihydronapelline. Oxidation of (36) with Cr0,-pyridine gives a triketolactam (39).'f Isosongorine (40) available from the allylic rearrangement of songorine (33) gave on Huang-Minlon reduction the deoxy-derivative (41). Oxidation of the latter with Cr0,-pyridine gave a keto-a-lactam (42) [1707 1635 cm.-l]. The latter gave an amorphous monobenzylidine derivative and exchanged 1.82 atoms of deuterium when equilibrated in the presence of NaOD in MeOD and D,O. These results indicate the presence of one methylene group adjacent to the keto- function and therefore limit the site of the ring-A hydroxyl in napelline to either T. Sugasawa Chem. and Pharm. Bull. (Japan) 1961,9 889 897. 530 Pelletier Cr03- Pyr 1 C(l) or C(3).To distinguish between these possibilities the dioxolactam (43) was treated with isopentyl nitrite to give the isonitroso-derivative (44). Cleavage of the latter with benzenesulphonyl chloride and alkali gave a nitrile (45) which was stable at its melting point and did not decarboxylate in boiling ethanolic hydrogen chloride. These results permit assignment of the hydroxyl in ring A to C(1); napelline and songorine are therefore represented by structures (32) and (33) respectively.' Reduction of songorine with LiAlH gives luciculine (33a) the alkamine of lucidusculine (33b). The structure and absolute configuration of the latter has been determined by an X-ray analysis.a (42) R= H 0 (43) R - 0 D. The Structure of Atisine.-(i) General structural considerations.Atisine C22H,3N02 the predominate alkaloid of Aconitum heterophyllum is a tertiary base which contains an exocyclic methylene a secondary hydroxyl and a C-methyl group. N-Alkyl determinations indicate the presence of an N-ethyl 8 T. Okamoto M. Natsume Y. Iitaka A. Yoshino and T. Amija Chem. and Pharm. Bull. (Japan) 1965 13 1270. 53 1 Quarterly Reviews group or suitable precursor in the m~lecule.~ Atisine is a very strong base (pK,' 12.8) which can easily be isomerised to isoatisine (pK,' 10.3). Reduction with hydride gives a dihydro-derivative analogous to dihydroveatchine. Oxida- tion of isoatisine with potassium permanganate gives a lactam oxoisoatisine C2,H3,N03 analogous to oxogarryine (14). Atisine contains one active hydrogen and on hydrogenation gives a mixture of tetrahydro-derivatives.(ii) Skeleton of atisine. One important difference noted between atisine and the Garrya alkaloids is their behaviour on selenium dehydrogenation. While veat chine gives 7-et hyl- 1 -met h ylphenan t hrene and 7-et hyl- 1 -methyl- 3 -=a- phenanthrene atisine gives l-methylphenanthrene 6-ethyl-1 -methylphenan- threne and 6-ethyl-l-methyl-3-azaphenanthrene (49).loJ1 This behaviour prompted the suggestion that the difference between atisine and the Garrya alkaloids lies in the position of attachment of ring D. In particular structure (46) was suggested for dihydroatisine and structures (47) and (48) for atisine and isoatisine re~pective1y.l~~~~ The relative positions of the secondary alcohol and exocyclic methylene group are determined by two independent studies.Dvornik and Edwards14 converted the azomethine alcohol (50) into the bisnor-keto-acid (51). Dibromination of the latter followed by dehydrohalogenation gave a crystalline phenol (52) and a ketoy-lactone (53) [bridging of ether oxygen either at C(11) or C(13)l. Forma- tion of the latter involving displacement of bromine a to the ketone by carboxy- late anion demonstrates a 1,4-relationship of the ketone and carboxyl groups in (51). These results taken with dehydrogenation evidence for substitution at a C. F. Huebner and W. A. Jacobs J . Biol. Chem. 1948 174 1001. lo C. F. Huebner and W. A. Jacobs J. Biol. Chem. 1947 170 203. 11 D. M. Locke and S. W. Pelletier J . Amer. Chem. SOC. 1959 81 2246. 1s K. Wiesner R. Armstrong M. F. Bartlett and J. A. Edwards Chem. and I d .1954 132. 18 S. W. Pelletier and W. A. Jacobs J . Amer. Chem. SOC. 1954 76,4496. 14 D. Dvornik and 0. E. Edwards Canad. J . Chem. 1964,42,137. 532 Pelletier C(12) [C(6) on phenanthrene nucleus] show clearly that a bicyclo[2,2,2]octane system is present and that the exocyclic methylene is located at C(16) [or C(13)] and the secondary hydroxyl at C(15) [or C(14)] respectively. (Since the bicyclo- octane system is symmetrical the allylic alcohol system could be located on either the cis- or trans-branch. This point is considered later.) Additional evidence on this point is available from the following sequence of reactions. Isomerisation of isoatisine (48) with ethanolic hydrochloric acid gave a mixture of epimeric 16-methyl ketones (54). Wolff-Kishner reduction of (54) gave the deoxy-derivative (55) and dehydrogenation of the latter with selenium afforded 6-isopropyl-1 -methylphenanthrene (56).The formation of the isopropyl derivative constitutes additional evidence for assigning the methylene to C( 16) [or C( 1 31. f$goMe (54) 0 Me (56) (iii) Oxazolidine ring system. Evidence for the existence of an oxazolidine ring in atisine and isoatisine is derived from two experiments.13J5 Oxidation of atisine (47) with permanganate in acetone gives lactam dicarboxylic acids (57) and (58) whose methyl esters (59 and 60) show no infrared absorption for a hydroxyl group and give negative tests for active hydrogen. A similar oxidation of isoatisine furnishes oxoisoatisinedicarboxylic acid (61) [compare oxidation of garryine (5) to oxogarrine (14)] which has been related to (57) by catalytic reduc- tion of the latter.The second piece of evidence is the oxidation of both atisine S. W. Pelletier and P. C. Parthasarathy J . Arner. Chern. Suc. 1965 87 777. 533 Quarterly Reviews Atisine - (47) and isoatisine to conjugated enones [(62) (Am, 229 mp 9500) and (63) (Amax 227 mp 8070) respectively] which show no infrared hydroxy absorption and give negative tests for active hydrogen. Refluxing (62) in methanol gives a good yield of the iso-enone (63) a reaction which parallels the easy veatchine 4 garryine and atisine + isoatisine isomerisations. Final confirmation of the presence of an *To I soa tis i ne 1 oxazolidine moiety has been provided by reconstitution of atisine from the i mino-alco hol (64). l6 (iv) Stereochemistry of atisine. At the time when the selenium dehydrogenation results mentioned above were available it was assumed that atisine possessed the trans-anti skeleton which is common to most diterpenes.Conforrnational arguments1' relative to the structure of the related alkaloid ajaconine have shown that ,+trans stereochemistry is present in atisine. Further reduction of ketones (65) and (67) with sodium borohydride gives in each case a pair of epimeric alchols (66a,b and 68a,b) each of which readily forms an O-acetate. This rela- tively unhindered character of the hydroxyl groups supports location of the allylic alchol group on the trans bridge of the bicyclo-octane system for if the group were on the cis bridge only one epimer would be expected on reduction owing to the severe crowding at C(14).l5~l8 Moreover this view is confirmed by a study of the pK,'s of the epimeric N-ethyl compounds (69a,b) and their 1% S .W. Pelletier and W. A. Jacobs J . Amer. Chem. SOC. 1956 78 4144. 17 A. J. Solo and S. W. Pelletier Chem. and Znd. 1960 1108. 18 D. Dvornik and 0. E. Edwards Chem. and Ind. 1958 623. 534 Pelletier acetates. Detailed studies indicate that the complete relative stereochemistry of atisine is expressed by structure 70.1bJ9 The assignment of configuration of the secondary hydroxyl in atisine is made difficult by the high degree of symmetry of the bicyclo[2,2,2]octane system. A tentative assignment of the @configuration as in (70) has been made on the basis of the difference in the absorption of the epimeric alcohols (68a,b) on alumina.lb l5 The absolute configuration of atisine is considered under Section 4C.U E. The Structure of Atidhe.-Atidine C22H33N03 another constituent of Aconitum heterophyllum is a tertiary base possessing two hydroxy groups (diacetate) a ketone function (oxime) in a six-membered ring an exocyclic methylene a C-methyl and an N-CH,CH,OH group. Atidine is shown to be an oxodihydroatisine (71) by Huang-Minlon reduction to dihydroatisine (46). A correlation of atidine and ajaconine via the dihydro-derivative (see below) allows the keto-function to be assigned to C(7) in atidine.20 F. The Chemistry of Aja~onine.l~-~~-Ajaconine C,,H,,NO, an alkaloid of Delphinium ajacas L. has been shown to have the same carbocyclic skeleton as atisine by conversion into the oxygen-free azomethine base (72) obtained earlier from atisine.21 That the allylic alcohol system of ajaconine has the same position and stereochemistry as in atisine was shown by reduction of atidine (71) (pre- viously correlated with dihydroatisine) to a mixture of epimers one of which is dihydroajaconine (73).20 19 D.Dvornik and 0. E. Edwards Tetrahedron 1961 l.4 54. *O S . W. Pelletier J . Amer. Chem. SOC. 1965 87 799. 21 D. Dvornik and 0. E. Edwards Chem. and Ind. 1957,952. 535 QuarterIy Reviews The work of Dvornik and Edwards has shown that ajaconine (74) contains a carbinolamine ether involving an oxygen atom at C(7).19 The N-methylcarbino- lamine (75) derived from ajaconine was transformed by methanolic alkali into a mixture of the iso-compound (76) and a hydroxy-lactam (77). This unusual product results from an intramolecular Cannizzaro-type reaction involving a transannular hydride transfer from C(20) to C(7)./ (76) I (77) RP-W (78) R=--OH The absolute configufation indicated for ajaconine (74) and its derivatives follows from its correlation with atisine and atidine.19s20 4 Correlations and Absolute Stereochemistry of Atisine and Gurrya Alkaloids A. Correlation of Atisine and Gurrya Alkaloids.-The Atisine and Gurryu alkaloids have been interrelated by converting both atisine and veatchine by a parallel sequence of degradations into the same N-acetyl ester (84).22 The respective N-acetyl derivatives (79a) and (79b) derived from atisine [12-16 bond] and veatchine [13-16 bond] were converted by oxidation with permanganate- periodate to the respective carboxylic acids (80a,b). Hydrolysis of the dimethyl esters (81a,b) gave (82a,b) which were transformed into the corresponding monobromides (83a,b) by the Hunsdiecker method.Reductive debromination of (83a) and (83b) with zinc dust in acetic acid gave the same acetyl ester (84). This correlation demonstrates that the Garrya alkaloids have the same stereo- chemistry of ring fusions as atisine. B. Correlation of the Atisines with Diterpenes.-(i) RemovaZ of the nitrogen from diterpene alkaloids. Degradation of diterpene alkaloids to diterpenes of established configuration was delayed by the lack of a suitable method of remov- ing the nitrogen atom. Since both positions p to the nitrogen are quaternary Hofmann-type degradations are ineffective. Success in removing the nitrogen from (72) has been achieved through a mild reaction with aqueous nitrous ** S.W. Pelletier and D. M. Locke J. Amer. Chem. Soc. 1965 87 761. 23 J. W. Ap Simon and 0. E. Edwards Canad. J. Chem. 1962,40 896. 536 Pelletier (79b) 13- 16 bond 16 bond I I The major product is the hemiacetal (85). This hemiacetal has been converted successively into a primary alcohol (86)’ aldehyde (87) and carboxylic acid (88). The same procedure has been applied with success to the azomethines derived from the Garrya alkaloids (see below). (85) Hd (88) R=CO,H (ii) Correlation of garryfoline with kaurene and ~tevane-B.~* The azomethine (89) derived from garryfoline was converted into the hemiacetal (90) by treatment with nitrous acid. Wolff-Kisner reduction of (90) effected simultaneous reduc- tion of the 15-0x0- and masked 19-aldehydo-functions. Oxidation of the resulting primary alcohol (91) with Cr0,-pyridine gave aldehyde (92) which was transformed by vigorous Wolff-Kishner conditions into the hydrocarbon (93).This hydrocarbon was identical with (-)-‘/i?’-dihydrokaurene the minor hydrogenation product of (-)-kaurene (95) and with ‘stevane-B’ a degradation product of steviol (96). The evidence for structure (90) for the hemiacetal is based on the extremely hindered nature of the derived aldehyde (92) and acid (94) [pK*Mcs 9.491. (iii) Correlation of atisine and the resin acids. The aldehyde (87) derived from atisine has been converted by Wolff-Kishner reduction into the hydrocarbon (97) which is enantiomorphic with a hydrocarbon (98) prepared by a long degradative sequence from abieta-6’8-diene (99).25 This work represents the first correlation of the tetracarbocyclic ring system of atisine with the resin 24 H.Vorbrueggen and C. Djerassi J . Amer. Chem. SOC. 1962,84,2990. 25 W. A. Ayer C. E. McDonald and G. G. Iverach Tetrahedron Letters 1963 No. 17,1095. 537 Quarterly Reviews ___) (91) R = W20H (92) R = CHO (93) R = Me (94) R = C02H (95) R’-H,R‘=M~ (96) R’=OH R ’ = C ~ H acids and confirms the mirror-image relationship at C(5) C(9) and C(10). Hydrocarbon (99) has also been synthesised from abietic acid via maleopimaric acid.26 C. Absolute Configuration of A tisine and Gurryu Alkal~ids.~*-Vorbrueggen and D j e r a ~ s i ~ ~ have converted veatchine (100) and garryfoline (101) into the 17- nor-16-ketones (102) and (103). These compounds exhibit a positive Cotton effect of amplitude similar to that observed for the 17-nor-16-ketone (104) from phyllocladene (105).Since the absolute configuration of phyllocladene has been established and since the configuration at C(9) should not effect the sign of the Cotton effect the absolute configuration at C(8) and C(13) of the Garrya alkaloids [and hence of C(8) and C(12) of atisine] is established. The correlation of garry- foline with (-)-‘p’ dihydrokaurene (93) has been discussed previously. These results lead to the complete absolute configurational representations (100) and (101) for veatchine and garryfoline. In view of the correlation22 of atisine and veatchine this absolute configurational assignment also applies to atisine (70) and its relatives such as ajaconine (74) and atidine (71). Independent evidence for the absolute configuration of atisine is provided by a synthesis from podocarpic acid (106) of the antipode of the phenol (52) origin- ally obtained by degradation of atisine.The synthesis involved photolysis of the azide (107) derived from the methyl ether of podocarpic acid. One of the products 26 L. H. Zalkow and N. N. Girotra J . Org. Chem. 1964,29 1299. 538 Pellet ier (101) R=(& (103) R= Ac (105) R=CH of the photolysis reaction was the &lactam (108) which was converted by con- ventional procedures into the enantiomer of (52). N3F @,) 0 5 The Ternary Iminium Salts of the Atisine and Garrya Alkaloids Atisine veatchine garryfoline and cuauchichicine are isomerised by dilute base to the iso-type bases isoatisine garryine isogarryfoline and isocuauchi- chicine respe~tively,l~~~~~ The isomerisation (109) -f (11 1) proceeds even at room temperature in alcohol without external base.Members of these pairs of isomers manifest a remarkable difference in basic strengths. In 50% methanol atisine has a pK,' of 12.8 while isoatisine has a value of 10.35. Similar differences exist for the Garrya alkaloid pairs. The salts of these alkaloids exist in the ternary iminium form. In hydroxylic solvents the 'normal' bases (109) exist almost completely as the ternary iminium hydroxides (110; X = OH) whereas the 'iso'-bases exist mainly in the oxazoli- dine form (111). The reasons for the preponderance of the ternary iminiurn 539 Quarferly Reviews hydroxide in the ‘normal’ base equilibria have been reviewed.l“Jf~~~ In solution isomerisation of the ‘normal’ bases proceeds through the ternary iminium forms (110) and (112) by prototropy.Since steric factors are responsible for the ‘iso’- bases having a lower free energy than the ‘normal’ bases the equilibrium is shifted toward the sterically more favoured ‘iso-’ f o r m ~ . ~ ~ J f ~ ~ ~ ~ In the case of the salts of the ‘normal’ and ‘iso’-bases the reverse of the situation described for the bases prevails. The ‘normal’ salts are more stable than the iso-~alts.~~ Thus isoatisinium chloride (112; X = Cl 12-16 bond) can be converted into atisinium chloride (110; X = C1 12-16 bond) by refluxing in such solvents as dimethyl sulphoxide dimethylformamide diethylformamide or high-boiling alcohols. Garryinium chloride (112; X = C1 13-16 bond) in similar conditions is isomerised to veatchinium chloride (110; X = CI 13-16 bond).Since the normal salts can be readily converted without isomerisation into the corresponding bases by treatment with cold aqueous alkali this thermal isomerisation of the salts provides a convenient practical method of reversing the easy ‘normal’ base (109) -+ ‘iso’-base (1 11) isomerisation. Detailed rate studies on this reaction in several organic solvents have been carried 6 The Chemistry of Alkaloids with a Modiiied Alkaline Skeleton In recent years a wide range of Aconitum species native to Japan and India have been examined for alkaloids. Among those encountered are several which are modelled on an atisine skeleton but possess additional ring fusions. This section will survey the chemistry of these compounds A. The Chemistry of Hetisine.-Hetisine CzoHa7N0, a minor constituent of Aconitum heterophyllmz,3° represents an interesting variant of the atisine skeleton.The alkaloid has one hydrogenatable double bond three active hydrogen atoms an exocyclic methylene group one C-methyl group and a tertiary nitrogen atom. N-Alkyl and methoxyl determinations are negative. Since dihydrohetisine shows no adsorption in the near-ultraviolet region it is clear that hetisine must have a heptacyclic skeleton.3l The nature of the oxygen functions is indicated by formation of a crystalline diacetate and an amorphous triacetate both of which regenerate hetisine on hydrolysis. Further the alkaloid is inert to both periodate and lead tetra-acetate and does not form an acetonide. It therefore possesses three acylatable hydroxyls which are non-vicinal and are not in a 1,3-cis-diaxial relationship.Dehydrogenation of hetisine yields a complex mixture of hydrocarbons from which pimanthrene (20) has been isolated. The fact that hetisine lacks a free N-alkyl group and compares in basicity (pK,’ 9.85) with quinuclidine (10~3)~ 27 K. Wiesner and J. A. Edwards Experientia 1955 11 255. 2D S. W. Pelletier K. Kawazu and K. W. Gopinath J . Amer. Chem. SOC. 1965 87 5229. 30 W. A. Jacobs and L. C. Craig J . Biol. Chem. 1942 143 605. 31 A. J. Solo and S. W. Pelletier J. Amer. Chem. Soc. 1959 81 4439; J. Org. Chem. 1962 27,2702. S. W. Pelletier K. W. Gopinath and K. Kawazu Chem. and Ind. 1966 28. 540 Pelletier suggests a quinuclidine-type structure with bonding from the nitrogen to either C(1) C(2) C(3) C(6) or C(7). One additional ring and three hydroxyl groups are necessary to complete the structure.An X-ray diffraction has estab- lished the correct structure of hetisine as (113). Under mild conditions of quaternisation hetisine forms a methiodide (1 14) which undergoes Hofmann degradation to give demethylhetisine. The latter possesses an exocyclic methylene group and a new double bond both of which can be hydrogenated. Under more vigorous conditions the methiodide re- arranges with participation of the original exocyclic methylene group. This rearrangement has been interpreted as proceeding via the ketone (115) to the hemiketal structure (1 16).= B. Ignavine and Anhydroignavinol.a-This alkaloid occurs in the roots of Aconitum sanyoense Nakai A . tasiromontanum Nakai and A. japonicum. Ignavine C,,H,,NO, lacks methoxy methylenedioxy- or N-methyl groups and is unreactive toward the usual carbonyl reagents.Hydrolysis of ignavine gives one mol. of benzoic acid and anhydroignavinol C,,H,,NO,. Of the six oxygens in ignavine two occur in the benzoyloxy-group and four in hydroxyls. One of these hydroxy groups is adjacent to the benzoyloxy-group since the hydrolysis product is susceptible to periodate cleavage while ignavine is not. Diacyl derivatives are formed which contain a mol. of water less than calculated and are therefore anhydroignavinol derivatives. Hydrolysis of these derivatives does indeed afford anhydroignavinol. That the exocyclic methylene group in ignavine is involved in a secondary allylic alcohol system as in atisine was demonstrated by catalytic isomerisation to a methyl ketone (1692 cm.-l) and by oxidation to a conjugated enone (1615 1687 cm.-l).a 32 M.Przybylska Canad. J. Chem. 1962,40 566; Acta Cryst.. 1962. 16 871. a3 K. Wiesner Z. Valenta and L. G. Humber Tetrahedron Letters 1962 No. 14 621. 34 E. Ochiai and T. Okamoto Chem. and Pharm. Bull. (Japan) 1959,7 550; ibid. p. 556. 541 Quarterly Reviews Dehydrogenation of anhydroignavinol yields a complex mixture of hydro- carbons from which lY7-dimethyl-6-n-propylphenanthrene 3-ethyl-l,8-dimethyl- phenanthrene and 3-isopropyl-lY8-dimethylphenanthrene have been isolated. These products account for 19 of the 20 carbon atoms of anhydroignavinol. The assumption of a bicyclo[2,2,2]octane-allyl alcohol system in ignavine is given credence by oxidation of de-N-methyloxoanhydroignavinol to a related dicarboxylic acid in which one of the carboxyl groups is tertiary.Further the three phenanthrene dehydrogenation products suggest a bond between C(14) and C(20) and a methyl group at C(1). Numerous degradations have established the position of the nitrogen and shown the presence of hydroxyls at C(2) and C(3) in anhydroignavinol. Since ignavine has no N-alkyl group a third bond must extend from nitrogen to one of the rings. The data suggest C(6) as the most likely site. The published data clarify the nature of four of the six oxygens of ignavine and three of the four of anhydroignavinol. The fourth oxygen in anhydroig- navinol is probably an ether since no hydroxyl band is observable in the infrared spectra of tribenzoylanhydroignavinol and certain other anhydro-derivatives. It is therefore likely that loss of water accompanying many of the reactions of ignavine involves ether rather than double-bond formation.Possible positions for the hydroxyls involved in the elimination assuming a /?-glycol moiety are C(11-13) C(7-14) C(7-9) and C(5-9). Since ignavine has a normal pK,' value (7-7) for a tertiary amine it is unlikely that a hydroxyl is at C(19) or C(20). In view of the above anhydroignavinol and ignavine have been provisionally represented by partial structures (1 17) and (1 18) respectively. (The absolute configuration shown is based on analogy to that of the other diterpene alkaloids.) C. Hypognavine and Hypognavin~l.~~~-Certain varieties of A. sanyoense contain an ester alkaloid which has one less oxygen atom than ignavine and is in many respects similar to it. This alkaloid hypognavine C,,H,,NO, is a benzoyl ester has no methoxly or N-methyl group fails to react with the usual carbonyl- test reagents and contains one hydrogenatable double bond.Like ignavine it has two acylatable hydroxy groups and an exocyclic methylene group. In con- 35 E. Ochiai T. Okamoto s. Hara s. Sakai and M. Natsume Pharm. Bull. Japan 1958 6 327. 36 S. Sakai J . Pharm. SOC. Japan 1956,76 1054. 37 S. Sakai Chem. and Pharm. Bull. (Japan) 1958 6 448. 38 S. Sakai Chem. and Pharm. Bull. (Japan) 1959 7 50 55. 542 Pelletier trast to ignavine hypognavine can be hydrolysed to an alkamine (hypognavinol CzoH2,N0,) without loss of a mol. of water. Acylation reactions and methiodide formation are also straightforward in the case of hypognavinol. Reactions analogous to those previously described afford clear evidence for the existence of a secondary ally1 alcohol system such as occurs in atisine ignavine and songorine.Selenium dehydrogenation of hypognavinol furnishes 1,8-dirnethyl-phenan- threne 1 ,7-dimethy1-6-n-propylphenanthrene7 and 3-ethyl-l,8-dimethylphenan- threne the latter two being characteristic products of anhydroignavinol. This suggests that hypognavine has the same skeleton as derived for ignavine. Formula (1 19) will be used as a basis for interpreting the chemistry of hypognavinol. Hofmann degradation of hypognavinol methiodide gave the de-N-base (1 20) and the isomerisation product (1 21). The resistance of de-N-methylhypognavinol toward a second Hofmann degradation is well accommodated by (120) for the /3 positions to the nitrogen at C(4) and C(10) bear no hydrogen and formation of a double bond between C(14) and C(20) would violate Bredt's rule.The size of the heterocyclic ring is indicated by oxidation of hypognavinol derivatives to 8-lactams. The three remaining oxygen atoms of hypognavinol exist as hydroxyls of which two are acylatable. The non-acylatable hydroxyl is shown to be tertiary by oxidation experiments. Detailed transformations have shown that the a-glycol system in hypognavinol is masked by a benzoyloxy-group in hypog- navine. The location of the a-glycol system in ring-A of hypognavinol is fixed by oxidation of de-N-methylhypognavinol (1 22) with silver oxide or alkaline ferrocyanide to a carbinolamine ether (123) from which (122) can be regenerated by reduction with NaBH,. Salts (124) of the carbinol-amine ether show infrared absorption typical of the :C = N+ group (1686-1679 cm.-l) and regenerate the parent base on treatment with alkali.Models show that an axial hydroxyl at C(2) is most favourable for ether formation. 543 QuarterZy Reviews The assignment of the second hydroxyl of the glycol system at C(l) or C(3) is not yet settled. Periodate cleavage experiments indicate that a trans-glycol system is present. The benzoyloxy-group in hypognavine is accordingly assigned a 1p- or 3/%configuration. The site of the tertiary hydroxyl is unknown. Hypo- gnavinol is thus represented (125b) or (126b). by structure (125a) or (126a) and hypognavine by RO (126a) R-H (l26b) R- COPh (125b) R= COPh D. Kob~sine?~-~~--This alkaloid C,,H,,NO, has been isolated from A. Kamtschaticum (fischeri) A.sachalinense Fr. Schmidt A. Iucidusculum Nakai and A. yesoensis Nakai. It possesses two secondary hydroxy groups in six- membered rings one of which is involved in an allyl alcohol grouping. Selenium dehydrogenation furnished 1 ,7-dimethyl-6-n-propylphenanthrene a charac- teristic product obtained also from ignavine and hypognavine. These results suggest that kobusine has the same skeleton as ignavine and hypognavine. The chemistry of kobusine will be discussed in terms of the ultimately derived structure (127). When kobusine (127) is warmed with dilute HCl compounds (128)-(130) are formed. Compounds (129) and (130) are reducible to the same glycol (131). Oxidation of (131) with Cr0,-pyridine gave a y-lactone (132) showing the proximity of the two hydroxyls. Treatment of kobusine with sodium in propanol gave the dehydroxyl derivative (134) which was related to the methyl ketone (128) via (133).Oxidation of (134) with Os04-HI04 followed by Cr0,-pyridine gave a y-lactone (135) thus establishing the relationship between the two hydroxyl groups. E. Pseudokobusine.41~42-Isolated from A. yesoensis Nakai and A. Zucidusculum Nakai this alkaloid C,,H,,NO, is closely related to kobusine. It is a tertiary base containing three acylable hydroxyls (tribenzoate) an exocyclic methylene involved in an allyl alcohol group a C-methyl group and no methoxy. Oxida- tion experiments show that the allylic hydroxyl and one other are on six-mem- bered rings and the third hydroxyl is tertiary since it is inert to Kiliani's reagent. Selenium dehydrogenation of pseudokobusine gave lY7-dimet hyl-6-n-propyl- phenanthrene a result which suggests that pseudokobusine has the same 3D M.Natsume Chem. and Pharm. Bull. (Japan) 1959 7 539. 40 T. Okamoto Chem. and Pharm. Bull. (Japan) 1959 7,44. 41T. Okamoto M. Natsume H. Zenda S. Kamata and A. Yoshino Abstr. I.U.P.A.C. Symposium Kyoto 1964 115. 42 M. Natsume Chem. and Pharm. Bull. (Japan) 1962,10 879. 544 Pelle tie + + Ni ' 9 ° C SEt 1 (133) < (134) - (l)Os0,-HI04 (ii) Cr0,- pyr structure as kobusine with the addition of one tertiary hydroxy group. The chemistry of pseudokobusine will be discussed in terms of structure (136). Acetylation of pseudokobusine gave besides the normal 0-acetate an N- acetyl-seco-derivative (1 37). An analogous N-cyano-seco-derivative (1 38) was prepared with cyanogen bromide.Both seco-compounds regenerated kobusine when hydrolysed with 20 % potassium hydroxide. Pseudokobusine methiodide (139) gave with ammonium hydroxide the N-inethylketone (140) (1675 cm.-l) which regenerated (139) when treated with hydriodic acid. (i) HJPd -C G'iICrO -AcOH (137) R=Ac (138) R=CN (140) R - Me The location of the third hydroxyl group was determined by the following sequence. Oxidation of N-acetyl-seco-pseudokobusine (1 37) with Os0,-NaIO gave a hemiacetal monocarboxylic acid (142) (1720 1650 cm.-l) the methyl ester of which was oxidised with CrO,-pyridine to the y-lactone (143) (1778 545 Quarterly Reviews 1728 1699 1630 cm.?). The same y-lactone was obtained by oxidation of (144) with Os0,-NaIO, followed by esterification. These transformations limit the position of the third hydroxyl group to position C(11) or C(13).That C(11) is the correct locus for the hydroxyl was shown by correlation of pseudokobusine with kobusine. t F. Is~hypognavine.~~~-This alkaloid C,,H,,NO (145b) occurs in the roots of A. majimai Nakai and A. japonicum Thumb and is a benzoate of the alkamine isohypognavinol (145a). The latter has three acylatable hydroxyls one of which is involved in the typical ally1 alcohol system. Isohypognavine (145b) has been correlated with kobusine (127). Reduction of isohypognavinol (145a) with sodium in propanol gave a deoxy-derivative (146) the ethiodide of which afforded on Hofmann degradation a tertiary base (147). Nuclear magnetic resonance studies on (147) support the assignment of an a-hydroxyl group at C(2). RO-. - OH (145b) R=Bz 7 Synthesis of Diterpene Alkaloids There has been such a flurry of activity in this area in the past few years that only a few significant developments will be cited.The N-acetyl ester (148) a key intermediate in the correlation of the Atisine and Garrya alkaloids has been converted via (149) into atisine (70) by a twelve- step sequence.45 The first complete stereospecific synthesis of (&)-atkine was reported by Nagata et aZ.46 starting from ketone (150) which by a 23-step process 43 E. Ochiai T. Okamoto S. Sakai and S. Inoue J. Pharm. SOC. Japan 1955,75,638. 44E. Ochiai 1". Okamoto S. Sakai M. Kaneko K. Fujisawa U. Nagai and H. Tani J. Pharm. SOC. Japan 1956,76 550. 46 S. W. Pelletier and P. C. Parthasarathy Tetrahedron Letters 1963 No. 4 205. 46 W. Nagata T. Sugasawa M. Narisuda T.Wakabayashi and Y . Hayase J. Amer. Chem. SOC. 1963 85 2342. 546 Pellitier was also converted into ketone (149). A synthesis of (&)-garryine and (5)- veatchine involving intermediates from the atisine synthesis has also been reported by Nagata et ~ 1 . 4 ~ A totally different synthesis of the diterpene alkaloids has been reported by Masamune starting from (151).** He has also converted compound (152) obtained from veatchine azomethine acetate by a multistep procedure into the monoester carboxylic acid (148). Since (148) has already been converted into atisine,46 this work completes in a formal sense the synthesis of atisine also. Still a third synthesis of the Garrya alkaloids has been reported by Valenta Wiesner and Wong starting with 5-metho~y-2-tetraIone.~~ In recent papers the New Brunswick group50 has described two new synthetic sequences which can be used to elaborate both atisine- and Garrya-type structures.Several other interesting approaches to the synthesis of the diterpene alkaloids have been rep~rted.~l-'l The Reviewer's work was supported by grants from the National Institutes of Health U.S. Public Health Service. 47 W. Nagata M. Narisuda T. Wakabayashi and T. Sugasawa J. Amer. Chem. SOC. 1964 86 929. 40 S. Masamune J. Amer. Chem. Soc. 1964 86 288 290 291. 49 Z. Valenta K. Wiesner and C. M. Wong Tetrahedron Letters 1964 No. 36 2437. 50 R. W. Guthrie A. Philipp Z . Valenta and I<. Wiesner Tetrahedron Letters 1965 No. 34 2945. 51 I. Iwai A. Ogiso and B. Shimizu Chem. and Znd. 1962 1288. 52 I. Iwai and A. Ogiso Chem. and Znd.1963 1084. 63 B. Shimizu A. Ogiso and 1. Iwai Chem. and Pharm. Bull. (Japan) 1963,11 333 766. 54 A. Ogiso B. Shimizu and I. Iwai Chern. and Pharm. Bull. (Japan) 1963 11 770 774. 55 A. Ogiso and I. Iwai Chem. and Pharm. Bull. (Japan) 1964 12 820. 66 A. A. Othmann and N. A. J. Rogers Tetrahedron Letters. 1963 No. 20 1339. 57 W. L. Meyer and A. S. Levinson Proc. Chem. Soc. 1963,15; J. Org. Chem. 1963,28,2859. ssL. H. Zalkow and N. N. Girotra J . Org. Chem. 1963 28 2037; 1964 29 1299; Chem. and Ind. 1965 704. 547 Quarterly Reviews 59 N. N. Girotra and L. H. ZaIkow Tetrahedron 1965 21 101. 6o T. Matsumoto and A. Suzuki Bull. Chem. SOC. Japan 1961,34 274. R. A. Bell and R. E. Ireland Tetrahedron Letters 1963 No. 4 269. 62 R. A. Finnegan and P. L. Bachman J . Org. Chem. 1965,30,4145. 63 D. H. R. Barton and J. R. Hanson Chem. Comm. 1965 No. 7 117. 64 K. Wiesner K. K. Chan and C. Demerson Tetrahedron Letters 1965 No. 33 2893. 65 A. Tahara K. Hirao and Y . Hamazaki Tetrahedron 1965,21,2133; Chem. and Ind. 1965 850. 66 K. Wiesner and A. Philipp Tetrahedron Letters 1966 No. 14 1467. 67 A. Tahara and K. Hirao Tetrahedron Letters 1966 No. 14 1453. 68 R. W. Guthrie Z. Valenta and K. Wiesner Tetrahedron Letters 1966 No. 38 4645. ‘O A. A. Othmann M. A. Qasseem and N. A. J. Rogers Tetrahedron 1967,23 87. 71 R. W. Guthrie W. A. Henry H. Immer C. M. Wong 2. Valenta and K. W. Wiesner Coll. Czeck. Chem. Comm. 1966 31 602. K. Wiesner and J. Santroch Tetrahedron Letters 1966 No. 47 5939. 548

ISSN:0009-2681    

DOI:10.1039/QR9672100525

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Quarterly Reviews ERRATUM Page 300. Scheme 6 (Structure 72) should read as above and not as printed. 548

ISSN:0009-2681    

DOI:10.1039/QR9672100548

出版商:RSC    

年代:1967

数据来源: RSC