摘要:

QUARTERLY REVIEWS TBE CYANINE DYES By F. M. HAMER Sc.D. D.Sc. F.R.I.C. (IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY LONDON S.W.7) Early Work.-In 1856 C. G. Williams obtained the blue dye cyanine by the action of caustic alkali on a quaternary quinolinium sa1t.l Years later it was recognised that this quinoline contained its 4-methyl homologue lepidine,2 and shown 3 4 that it was the mixture of ethiodides that gave rise to cyanine according to the equation A mixture of the ethiodides of quinoline and its 2-methyl hornologue,- quinaldine gave the redder iso~yanine.~~~ The structures of the molecules were considered 5-7 and it was proved that in isocyanine formation the %methyl group of quinaldine and the 4-position of quinoline 9 are involved. The present-day structures (I11 and 11) for cyanine and isocyanine were suggested as a possibility in 1906,lO though open-chain formulse were regarded as more probable.lf Practical Applications.-Cyanine has no value for dyeing fabrics because of its instability to light but in 1873 H.W. Vogel discovered the phenomenon of photographic sensitisation by dyes,l2 amongst them cyanine,13 and all the modern achievements of photography are its outcome. It is remarkable that except for erythrosin sensitisers of value are drawn only from the cyanine and related groups out of the myriad known dyes. A gelatino silver chloride photographic emulsion such as is used for papers is sensitive according to its wedge spectrogram only from 3500 to 4500 A. (Plate 1A) ; a bromide emulsion such as is used for films and plates is sensitive only from 3500 to 5300 A.(Plate lB) but bathing it in a solution of cyanine conferred sensitivity to green light (Plate 1C). As cyanine also fogged emulsions it appeared of solely academic interest until 1903 when Trans. Rog. SOC. Edinburgh 1856,21,377 ; Chem. News 1859,1 15 ; 1860,2,219. S. Hoogewerff and W. A. van Dorp Rec. Trau. chim. 1883 2 28 41. Idem ibid. p. 317 and 1884 3 337. W. Spalteholz Ber. 1883 18 1847. E. Vongerichten and C. Hofchen ibid. 1908 41 3054. A. Kaufmann and E. Vonderwahl ibid. 1912 45 1404. CgH,N,C,HJ + C,,HgNYC2HJ = C23H23N21 + H2 + HI a A. UT. Hofmann Proc. Roy. SOC. 1863 12 410. * H. Decker ibid. 1891 24 690 ' A. Miethe and G. Book ibid. 1904 37 2008. lo W. Konig J . p r . Chem. 1906 [ii] 73 100. 11 Idem ibid. 1912 [ii] 86 166. l2 Ber. 1873 8 1302. '3 Ibid. 1875 8 1635.327 328 QUARTERLY REVIEWS isocyanines (Plate 1D) were observed to sensitise without fogging.14 On this followed German patents largely the work of E. Konig ; they covered isocyanines into which substituents had been introduced in order to improve sensitising or in which the acid radical had been altered so as to increase the solubility. l5 By carrying out an isocyanine condensation in the presence of formaldehyde there resulted pinacyanol l6 the first sensitiser to red light (Plate 1E). As observed by H. W. Vogel,17 the sensitising maximum of a dye approxi- mates to the absorption maximum of its solution but lies at a somewhat longer wave-length ; it has recently been explained that the sensitising maximum is the same as the absorption maximum of dyed silver bromide,l8 and close correspondence between the sensitising curve and this absorption curve has been established.l 9 By means of sensitising dyes in conjunction with light filters plates can now be made having the same sensitivity as the eye can be made equally sensitive to the whole visible spectrum can be sensitised to any desired part of it which fact is utilised in colour photography or may even be sensitised to the near infra-red. In distance photography as in taking photographs from aeroplanes sensitisation to red or infra-red light is essen- tial because the atmosphere scatters badly light of shorter wave-length. At the outbreak of the first World War however all sensitisers were made in Germany; through submission of the problem to Sir William Pope those required by the Allies were prepared in the University of Cambridge and the chemistry of the cyanines was placed on a scientific basis by the work of W.H. Mills and his pupils. Scientific Foundation.-On the scientific foundation laid in the decade beginning with 1920 a worthy superstructure was built in the following twenty years. By oxidative breakdown the constitution of isocyanine had been established as (II),20 and received confirmation,21 22 the open-chain formula being disproved ; 23 it followed that the closely related cyanine must be represented by (III).20 Unlike the 2 4’- and 4 4’-cyanines (I1 and 111) which had so long been known the first 2 2’-cyanine (I ; R = Me) was only prepared in 1920 by condensing quinaldine methiodide with 2-iodo- quinoline methiodide ; 24 the reaction was subsequently extended this type also being found to comprise photographic sensitisers.22 The con- stitution (IV) suggested for pinacyanol 25 was established by oxidative 14A. Miethe and A. Traube G.P. 142,926/1902. 16 Farbenfabr. vorm. F. Bayer and Go. G.P. 158,078 ; 170,048-9/1903 ; Farbw. 16 Farbw. vorm. Meister Lucius and Briining G.P. 172,118/1905. l7 Ber. 1874 7 976. 18 S. V. Natanson Nature 1937 140 197 ; J. Phys. Chem. Russia 1938 11 157. l9 J. A. Leermakers J. Chem. Physics 1937 5 889. 2o W. H. Nills and R. S. Wishart J. 1920 117 579. 21 E. Q. Adams and H. L. Haller J . AWW. Chem. SOC. 1920 42 2389 *2 F. M. Hamer J. 1928 206. 23 W. H. Mills and P. E. Evans J . 1920 117 1035. 2 4 0. Fischer and G. Scheibe J . pr. Chem. 1920 [ii] 100 86. 26 W. H. Mills and VV. J. Pope Phot. J. 1920 60 253.vorm. Meister Lucius and Briining G.P. 167,159 ; 167,770/1903. IIAMER THE CYANINE DYES 329 breakdown ; 26 it was confirmed by the relationship of pinacyanol to rnethylenediquinaldine (1 3-di-2'-quinolylpropane) diethiodide 27 and by its synt,hesis from quinaldine ethiodide and ethyl orthoformate. 28 The unknown 2 4'- and 4 4'-isomerides (V) and (VI) were predicted.26 Later (VI) was synthesised by the action on lepidine ethonitrate in pyridine of the disulphide (VII) one function of which is to provide the central methin group (*CH:) of the trimethin chain; 29 and (V) was similarly synt'hesised from the ethonitrates of lepidine and quinaldine.30 (VI) was identified with " kryptocyanine ",29 which had been made by the action of alkali with formaldehjfde or chloroform on a lepidinium quaternary CHO kHO salt 31 and had proved a valuable sensitiser to infra-red light.Of ( ( di- cyanines " similarly made from 2 4-dimethylquinolinium salts and described as infra-red sensitisers a commercial sample was shown to be spectro- scopically similar to (V). 30 A purple dye first prepared in 1887 by the action of ammonia on the alkiodides of benzthiazole and 2-methylben~thiazole,~2 was obtained by W. H. Mills in better yield by use of pyridine and was accompanied by a yellow product both dyes possessing sensitising properties. Mills established 26 W. H. Mills and F. M. Ha.mer J. 1920 117 1550. 27 F. M. Hamer J. 1923 123 246 ; 1925 127 211. 28 W. Konig Ber. 1922 55 3293. 29 W. H. Mills and W. T. K. Braunholtz J. 1923 123 2804. 30'W. H. Mills and R. C. Odams J. 1924 125 1913.31 E. Q. Adams and H. L. Haller J . Amer. Ckm. Soc. 1920 42 266'1 32 A. W. Hofmann Ber. 1887 20 2251. 330 QUARTERLY REVIEWS fhat the yellow substance was the methincyanine (VIII) and the purple dye the trimethincyanine (IX ; R' = H).33 The latter was also prepared from 2-methylbenzthiazole ethiodide and (VII) in pyridine ; besides provid- ing a methin group (VII) had a second function in providing a benzthiazole t xs (xr) nucleus which united with the quaternary salt present so that (IX ; R' = H) was accompanied by (VIII).z9 Similarly by the method involving (VII) (IV) was accompanied by (X) whilst (VI) was accompanied by (XI) ; 29 representatives of type (XI) were also prepared by the action of alkali on a mixture of quaternary salts of 2-methylbenzthiazole and of quinoline and were found to be sen~itisers.~~ The ethyl orthoformate method 2s was applied to the synthesis of (IX ; R' = H),35 (XI1 ; Y = 0),35 and (XI1 ; Y = CMe,).36 By introduction of pyridine as condensing agent this became the best practical route to trimethin~yanines,~~ and was applied to the preparation of (XI1 ; Y 2 Se) 3s and by use of ethyl orthoacetate to that of (IX; /yD R' = Me).39 All these proved to be sensi- a 1 ) c H .w cH-c prised both symmetrical members such as (I) (111) and (VIII) in each of which there are two similar nuclei similarly linked and unsymmetrical members such as (11) (X) and (XI) but the tri- methincyanines were all of the symmetrical type with the exception of (V) where two similar nuclei are differently linked. By 1929 intermediate com- pounds had been prepared leading to unsymmetrical trimethincyanines which comprised photographic sen~itisers.~O A wide field now lay open and one problem in a review such as this is to separate the grain from the chaff.Definition and Nomenclature.- %4 R t i s e r ~ . ~ ' - ~ ~ So far the methincyanines com- (xu.) 1 - - - - - - - - - - - I I - -- -- - -1 33 J. 1922 121 455. 36 W. Konig and W. Meier J. pr. Chem. 1925 [ii] 109 324. a* W. Konig Ber. 1924 57 685. 3* L. M. Clark J. 1928 2313. 34 W. T. K. Braunholtz and W. H. Mills ibid. p. 2004. 37 F. M. Hamer J. 1927 2796. 39 F. M. Hamer ibid. p. 3160. H A. Piggott E. H. Rodd and I.C.I. Ltd. B.P. 344,409/1929 ; 354,898/1930. HAMER THE CYANINE DYES 331 A cyanine is a monoacid salt in which two heterocyclic nitrogen-contain- ing nuclei are linked by an odd-numbered methin chain as above (where rn = 0 1 2 or 3 and n and n’ = 0 or 1).Although one nitrogen atom is tertiary and the other quaternary it was realised years ago that the acid radical cannot be regarded as attached to one nitrogen atom rather than t o the other.41 Each cyanine is regarded as a resonance hybrid of two canonical structures; although no one single formula is a complete repre- sentation in the following pages for brevity only one is given. The nomenclature in the early days was comparatively simple and trivial names such as thia-2’-cyanine for (X) and thia-.l’-cyanine for (XI) still retain a certain usefulness. Cyanines in which the nuclei were linked by a tri- methin chain were conveniently named “ carbocyaniiies ’7,25 whence followed the term “ dicarbocyanines ” for those with a pentamethin chain,82 and “ tricarbocyanines ” for those with a heptamethin ~hain.~O As an example (IX ; R = R‘ = Et ; X = I) was described as 3 3‘ 9-triethylthiacarbo- cyanine iodide.As the subject developed however this type of nomen- clature came to involve the interpretation of a whole host of trivial names besides the difficulties of naming simpler dyes such as (XV) very complex ones such as (LXXVI) and cyanines with anomalous chain-lengths such as (XXXIX). There was thus built up a systematic nomenclature in which only a knowledge of the fundamental structure of a cyanine being pre- supposed each name specified the nature of the nuclei their positions of linking and the length of the chain joining them. According to this (IX ; R = R’ = Et ; X = I) is described as bis-2-(3-ethylbenzthiazole)-~-ethyltri- methincyanine iodide.Other examples of both types of nomenclature as applied to fairly simple types of cyanine are to be found elsewhere.234 Tri- nuclear cyanines have been named as substituted dinuclear cyanines.2051 206 Methincyanines.-According to the historic method a quaternary salt of lepidine or quinaldine was condensed with one of quinoline in alcoholic solution to give (111) 3 respectively according to the equation on p. 327 the liberated acid being taken up by the caustic alkali and the two hydrogen atoms by the customary excess 9 42-44 of quinolinium salt. An 82% yield of 2 4’-cyanine 45 and a 63% of thia-4I-cyanine (XI) 46 were attained. 4’-Cyanines containing a simple thiazole 479 48 or a A2-thia- zoline 49 nucleus were also made and in general any heterocyclic quaternary ammonium salt having a 2- or a 4-methyl group can condense through it with a quinolinium salt at the 4-position.If the 4-position is blocked condensation may occur at the 2-position as in the preparation of 2 4’- cyanines by the action of alkali on lepidine alkiodides.21 or ( 1 1 ) 7 4 9 4i W. H. Mills and W. T. K. Braunholtz J. 1922 121 1189. 42 W. H. Mills and W. J. Pope Phot. J . 1920 60 183. 43 H. Barbier Bull. Soc. chim. 1920 27 427. 44 F. M. Hamer J. 1921 119 1432 ; Phot. J 1922 62 8. 45 Idem J. 1939 1008. 46 G. H. Keyes and L. G. S. Brooker J . Amer. Chem. SOC. 1037 59 74. 47 W. H. Mills and J. L. B. Smith J. 1922 121 2004. 48 I.G. Farbenind. A.-G. B.P. 386,903/1931. d9 L. G. S. Brooker J . Amer.Chem. SOC. 1936 58 662. 332 QUARTERLY REVIEWS A second method consists in condensing a heterocyclic quaternary ammonium salt having a 2- or 4-methyl group with a similar reactive iodo-salt in the presence of absolute alcohol two molecules of acid being removed by the added alkali. Thus from 2-iodoquinoline alkiodide there was first prepared type (I),24 and subsequently (11) (X) (XIII ; Y = CMe,),22 and (XIV).49 Up to a point the other reactant also could be varied alkiodides used being those of 2-iodopyridine and 2-iodo-~-naphthoquinoline (2-iodo-5 fS-benzq~inoline).~~ Both non-ionic iodine atoms of 2 4-di- iodoquinoline ethiodide can take part in this type of reaction.51 The observation that increased yields are obtainable by using a strong organic base such as triethylamine as condensing agent made practicable the preparation of (XIII; Y = 0) (XV),52 and (XVI; Y = 0 S or Se).53 A limitation of the general method is that it is inoperative with 2-iodo-salts which have more than one hetero-atom in the ring carrying the iodine atom e.g.with 2-iodobenzthiazole alkiodides. R R X' Xnr.1 (=IS A third method was on the one hand described as consisting in reaction of a heterocyclic quaternary ammonium salt having a 2- or a 4-methyl group with a heterocyclic base having in the 2- or the 4-position a group such as :S /a11ky!54 On the other hand the second reactant was regarded as a quaternary heterocyclic ammonium salt having a 2- or a 4-alkylthio- group.55 A recommended procedure consisted in heating together a base having a 2- or a 4-alkylthio-group7 one having a 2- or a 4-methyl group and an alkyl tol~ene-p-sulphonate.~~~ 56 One molecule of acid and one of alkanethiol were eliminated.By means of alkylthio-compounds five- and six-membered diazole nuclei were introduced into methincyanine molecules. 57 In such syntheses the alkyl of the alkylthio-group should be identical with that of the ester used for salt-forrnati~n,~~ since a rearrangement of alkyl \X 50 F. M. Hamer and M. I. Kelly J. 1931 777. 51 L. G. S . Brooker and L. A. Smith J. Amer. Chem. Xoc. 1937 59 67. 52 L. G. S. Brooker and G. H. Keyes ibid. 1935 57 2488. 53 L. G. S. Brooker G. H. Keyes and F. L. White ibid. p. 2492. 54 I.G. Farbenind. A.-G. B.P. 423,792/1932. 5 5 J. D. Kendall B.P. 424,559/1933 ; 438,420/1934. 5 6 J. D. Kendall and H. G. Suggate J. 1949 1503. 57 J. D.Kendall B.P. 425,609/1933. 58F. M. Hamer J. 1940 799. HAMER THE CYANINE DYES 333 groups occurred when ethyl iodide acted on 2-methylthioquinoline 59 or on 2 -met hy It hio benzt hiazole . 6O R'eaction of a 4- aryl t hiop yridinium salt a 4-methylpyridinium salt and a strong base yielded the 4 4'-isomer 6 1 A fourth method leading solely to symmetrical methincyanines con- sisted in heating the alkotoluene-p-sulphonate of a heterocyclic base having a 2-alkylthio-group7 with malonic acid and pyridine. 62 A rather curious fifth method of preparing methincyanines consisted in heating a heterocyclic quaternary ammonium salt having a reactive methyl group with amyl nitrite and acetic anhydride ; the original patent 6 3 covered 2 3 3-trimethylindoleninium salts but the products were formu- lated as B-azatrimethincyanines ! In applying the method to 2-methyl- benzthiazolium salts the products were established as (VIII) ; 64 in its day this was the best method for their preparation but it has been superseded by the alkylthio-method.It was applied to the preparation of (XVII; Y = 0) which were interesting as being the first colourless cyanines,65 and of (XV). Me clod- ( =.> of (XVII ; Y = Se).66 The dyes from indoleninium salts were also estab- lished as (XVII ; Y = CMe,) and in this series there was isolated an inter- mediate compound (XVIII) and hence t'he unsymmetrical (XIX). 67 Sixthly the synthesis of (VIII ; R = Et) from ethyl malonate and o-aminothiophenol through di-2-benzthiazolylmethane its ethiodide and the base formed by elimination of acid,33 though not a recommended preparative method is a classic without which any account of methincyanines would be incomplete.I n this connection an attempted synthesis through methanes substituted by two heterocyclic nuclei of which at least one was a substituted quinoxaline e.g. (XX) is interesting as is also its failure because of the resistance of such methanes to quaternary salt formation. 59 B. Beilenson and F. M. Hamer J. 1939 143. 6o W. A. Sexton ibid. p. 470. 61 L. G. S. Brooker and Eastman Kodak Co. U.S.P. 2,202,827/1940; 62 J. D. Kendall B.P. 431,141/1933. 6 3 I.G. Farbenind. A.-G. B.P. 291,888/1927. 13* N. I. Fisher and F. M. Hamer J. 1930 2502. 66 I.G. Farbenind. A.-G. B.P. 380,702/1931. 67 R. Kuhn A. Winterstein and G . Balser Ber. 1930 63 3176. 2,231,657/1941. e6 Idem J.1934 962. 334 QUARTERLY REVIEWS However two methincyanines e.g. (XXI) each containing one quinoxaline nucleus were obtained by the alkylthio-method.68 (=I S m e t r i c d Trimethincy&es.-The method of preparing symmetrical trimethincyanines by the action of alkali on an alcoholic solution of quin- aldine alkiodide and quinoline alkiodide in t he presence of formaldehyde 16 was the standard procedure of thirty years ago 25 69-71 and is now obsolete. The quinolinium salt did not enter into the dye molecule 69 and its function may have been to take up two hydrogen atoms which in addition to water and acid were eliminated ; its usefulness has however been denied.72 A second early method was based on the use of a trihalogenomethane 7 3 t o provide the central unit of the chain.It is recorded that treatment of an alcoholic solution of the appropriate picolinium iodide with potassium hydroxide and chloroform did lead to (XXII) and its 4 4'-ana10gue774 whereas (XXII) could not be obtained by the ethyl orthoformate E t E t method.37 74 Until recently use of a trihalogenomethane in the presence of alkali alk~xide,'~ was the only known way of preparing trimethincyanines of the benziminazole series such as (XXIII). That method in which the central carbon atom of the chain was supplied by a disulphide (VII) was invaluable in establishing the constitution of (V) (VI) and (IX ; R' = H) 299 30 but is not a convenient preparative one. The standard method of preparing symmetrical trimethincyanines is by zxse of ethyl orthoformate and pyridine.28t 35-38 A fifth way differs from the others in that the whole trimethin chain is furnished by glutaconic acid or by crotonic anhydride which is condensed in the presence of pyridine with two moles of the alkotoluene-p-sulphonate of a heterocyclic base having a reactive alkylthio-group.62976 A .H. Cook and R. F. Naylor J. 1943 397. 6 9 0. Fischer J . pr. Chem. 1918 [ii] 98 204. 70 L. E. Wise E. Q. Adams J. K. Stewart and C. H. Lund I?d. Eng. Chem. 1919 7 1 W. T. K. Braunholtz J. 1922 121 169. 7 2 K. Lauer and M. Horio J . pr. Chem. 1935 [ii] 143 305. 7 3 Farbw. vorm. Meister Lucius and Briining G.P. 200,207/1907. 7 4 E. Rosenhauer and F. Barlet Ber. 1929 62 2724. 75 I.G. Farbenind. A.-G. B.P. 521,165/1937 ; cf. A. van Dormael with J. Libeer 76 J. D. Kendall B.P. 431,186/1933. 11 460. Sci. et I d . Phot.1949 [2] 20 451. HAMER THE CYANINE DYES 335 Unsymmetrical Trimethincyanines.-Historically it is interesting that (VII) was used in synthesising the unsymmetrical (V) from a mixture of quaternary salts,30 and that similar dyes were made commercially from 2 4-dimethylquinolinium salts where both methyl groups are reactive. In general however the use of a mixture of salts is undesirable because the unsymmetrical dye is liable to be contaminated by two symmetrical ones. The first practical preparation of trimethincyanines having two different nuclei is an instance of valuable work which has been published only in the form of patents.40 Diarylformamidines ArNH*CH:NAr were used either to give symmetrical trimethincyanines by condensation with two moles of het erocyclic quaternary ammonium salt having a reactive methyl group or to give intermediate compounds by condensation of equimolecular proportions.If obtained by simple fusion the intermediate was of the anilinovinyl type e.g. (XXIV ; R = H) but if prepared in the presence of acetic anhydride it was an acetanilidovinyl compound e.g. (XXIV ; R = Ac). By its condensation with a molecule of a different salt having a reactive methyl group unsymmetrical trimethincyanines were produced. It was recognised 7' that better yields result from the acetanilido- than from the anilino-type of intermediate when the medium is other than acetic ( xxnr.) ( m.) (V anhydride. The use of ethyl N-phenylformimidate (ethylisoformanilide) PhN:CH*OEt as a substitute for diphenylformamidine was an improvement in giving higher yields of the #?-anilinovinyl intermediate compounds.78 Other intermediate compounds claimed were the semicarbazone etc.of aldehydes e.g. (XXV) which was obtained by hydrolysis of (XXIV; R = Me) ; '9 also (XXVI) prepared either by use of ethyl trithio-ortho- formate or from an aldehyde and phosphorus pentasulphide followed by an ester.81 Symmetrical and Unsymmetrical Pentamethincyanines.-The first penta- methincyanines were meso-substituted e.g. (XXVII ; Y = C1 Br or NO,). They were prepared 82 by condensing one mole of a-chloro-#?-anilinoacralde- hyde anil NHPh*CH:CCl*CH:NPh or the corresponding a-bromo- or a-nitro-compound with two moles of a quaternary salt having a reactive methyl group. Sometimes the acid salts of these anils were used some 77 T. Ogata Bull. Inst. Phys. Chem.Res. Japan 1934 13 549 ; Proc. Imp. Acad. Japan 1937 13 325. 78 E. B. Knott J. 1946 120. 79 I.G. Farbenind. A.-G. B.P. 486,780/1936 ; 510,242/1938. 81 K. Kumetat and 0. Riester vested in Alien Property Custodian U.S.P. 82 S. Beattie I. M. Heilbron and F. Irving J. 1932 260. J. D. Kendall and J. R. Majer J. 1948 687. 2,349,179/1944. 336 QUARTERLY REVIEWS reactions being effected by piperidine in pyridine and others by potassium acetate in acetic anhydride. The unsubstituted p-anilinoacraldehyde anil hydrochloride gave dyes such as (XXVII ; Y = H) but also by reaction with the usual 2- or 4-methyl salts in the presence of acetic anhydride intermediate compounds such as (XXVIII ; R = Ac) whence unsym- metrical dyes e.g. (XXIX) could be obtained.83 Non-acetylated inter- mediate compounds e.g.(XXVIII ; R = H) were best prepared by using a strong organic base as condensing agent with alcohol as medium ; unlike their lower vinylene homologues where the acetylated were more useful than the non-acetylated compounds in this series the non-acetylated salts were applied to the preparation of unsymmetrical pentamethincyanines for use as photographic sensitisers for the deep red and i n f r a - ~ d . ~ ~ (-3 (=.) By condensing quaternary salts having a reactive methyl group with ,&ethoxyacraldehyde diethyl acetal in the presence of acetic acid it was possible to make ethoxy-compounds such as (XXX) which could be used in preparing unsymmetrical pentamethinc~anines.~~ The ethoxy-compounds were found useful in the crude state as a source of the pure anilinobuta- dienyl type of intermediate.86 It is recorded that p-alkylthioacraldehyde acetal or dithioacetal may also be ~ s e d .8 ~ A method proposed for symmetrical pentamethincyanines and consisting in the condensation of a quaternary salt having a Z-/?-acetanilidovinyl group with nialonic acid 88 proved useful specifically for those having two benzoxazole nuclei. 89 For symmetrical pentamethincyanines a suggested method resembles the fifth for symmetrical trimethincyanines ( q . ~ . ) in that the whole chain is provided as a unit in this instance by sorbic anhydride.76 Symmetrical and Unsymmetrical Heptamethincyanines.-Jmt as tri- and 83H. A. Piggott E. H. Rodd and I.C.I. Ltd. B.P. 355,393/1930. 84 I.G. Farbenind. A.-Q. B.P. 434,234-5/1933. 85R. H. Sprague and Eastman Kodak Co. U.S.P. 2,269,234/1912.86 F. M. Hamer J. 1949 32. 87 J. D. Kendall and H. D. Edwards B.P. 562,565-8/1933. 88 T. Ogata Bull. Inst. Phys. Ghem. Res. Japan 1937 16 631. 89 F. M. Hamer and R. J. Rathbone J. 1945 595. HAMER THE CYANINE DYES 337 penta-methincyanines were prepared by use of diphenylformaniidine and p-anilinoacraldehyde anil hydrochloride respectively’ so symmetrical hepta- methincyanines were prepared by condensing “ glutaeonaldehyde dianilide ” (PhNH*CH:CH.CH:CH*CH:NPh) hydrochloride with various quaternary salts having 5t reactive methyl group ; those made had two %linked quino- line benzthiazole naphtho( 1’ 2’-4 5)thiazole benzselenazole A2-thiazo- line or 3 3-dimethylindolenine nuclei ; the condensing agent was usually caustic alkali in alcohol but for (XXXI) was sodium acetate in acetic anhydride.g0 Dyes of this last type were also made by two other sets of w~rkers,~lt 92 all three recording that they sensitise to infra-red light.Spec- trograms showing the sensitisation were published 939 94 and it was pointed out 94 that penta- and hepta-methincyanines had been independently dis- covered in Germany chiefly by W. Konig and put into commercial use. Use of triethyladne as condensing agent at a low temperature made it possible to prepare the symmetrical heptamethincyanines containing RN CH.CH:CH.CH:CH*CH.CH simple thiazole and 4-linked quinoline nuclei.g5 The latter (XXXII) has been used for infra-red photography e.g. of arc spectra in the region A scientifically interesting method of preparing heptamethincyanines consisted in treating the quaternary salt having a reactive methyl group with alkali and 2 4-dinitrophenylpyridinium chloride in the presence of alcoh01,~O~ 91 whereupon the pyridinium ring opened to give glutacon- aldehyde,97 which condensed with the salt.The first intermediate compound of this series (XXXIII ; R’ = Me ; U = CMe,) was prepared in acetic anh~dride.~l Later as with the next lower vinylene honiologues an alkaline medium was recommended for making (XXXIII ; R‘ = H ; Y = *CH:CH* S or Se).9* Use of a strong organic base as condensing agent made available intermediate compounds of even 9000-1 0,000 A. 96 N. I. Fisher and F. M. Hamer J . 1933 189. 91 0. Wahl and I.G. Farbenind. A.-G. G.P. 499,967/1928. 92 H. A. Piggott E. H. Rodd and I.C.I. Ltd. B.P. 355,693/1930. 9a L. G. S. Brooker F. M. Hamer and C.E. K. Mees Phot. J. 1933 73 258. * p W. Dieterle H Diirr and W. Zeh 2. wiss. Phot. 1933 32 145. 95 L. G. S. Brooker and Eastman Kodak Co. U.S.P. 2,095,856/1937; 96 W. F. Meggers and C. C. Kiess Bur. Stand. J . Res. 1932 9 309. 97 T. Zincke Annalen 1904 330 361. 98 I.G. Farbenind. A,-G. B.P. 438,449-50 ; 438,484/1933. 2,165,337,4939. 338 QUARTERLY REVIEWS the benzoxazole series and hence heptamethincyanines with benz- or naphth- oxazole n~clei.9~ Another method of preparing unsymmetrical heptamethincyanines con- sisted in condensing the usual quaternary salts having a reactive methyl group with heterocyclic pentamethin aldehydes (XXXIV ; Y = *CH:CH* S or Se) obtained by the action of alkali on the corresponding N-substituted hexamethin salts (XXXIII). loo (==&I ( aKxm.) Polymethincyanines.-Much as the pyridinium nucbus had been rup- tured giving the hydrochloride of glutaconaldehyde dianilide so rupture of the furfuraldehyde nucleus gave e.g.(XXXV ; n = 0) and of 2-/?-furyl- acraldehyde gave e.g. (XXXV ; n = 1) and (XXXVI ; 12 = 1) * ; lol subse- quent,ly dyes such as (XXXVI ; n = 2 and 3) were prepared by rupture of (m.) Ph*NH.CH:CH-CH:C.[CH:CH-l CH:NPh,HBr I n OH “2/C-C,H* H*/C-C\H2 N .CH:CH-CH:C.[CH:CH;J CH=N I OH I-s the nuclei of 5-furylpenta-2 4-dien-l-a1 and 7-furylhepta-2 4 6-trien-l-al respectively but the instability increased as the chain was lengthened. By acyl chlorides or anhydrides in the presence of pyridine the hydroxyl group could be acylated giving more stable compounda.102 These could be con- densed with various quaternary salts having a reactive methyl group to ( Zrmxmr.1 give cyanines having a chain of at least seven carbon atoms with acyloxy as substituent.Examples in a patent 103 include hepta ennea- and hen- deca-methincyanines e.g. (XXXVII ; n = 0 1 or 2) but it would seem 99 G. H. Keyes and Eastman Kodak Co. U.S.P. 2,251,286/1941. loo I.G. Farbenind. A.-G. B.P. 499,318 ; 501,005/1936. 101 W. Konig J. pr. Chem. 1905 [ii] 72 555; loa W. Konig with K. Hey F. Schdze F. Silberkweit and K. Trautmann Ber. 1°3 I.G. Farbenind. A.-G. B.P. 441,624/1933. * In formula (XXXVI) a positive charge has inadvertently been omitted from the 1913 [ii] 88 193. 1934 67 1274. right-hand nitrogen atom. B ayea with isocyaniue D Wave-lengths (cm.-E) PLATE 1 Wedge spectrograms. Immide tniulsion dyed with triniettiin- cyariiiie 13 1)rorriide elllulsiol1 dyed with heptainet Iiin- cyaiiiue D Wave-lengths (em.-*) PLATE 2 Wedge spectrograms of semitising by cyanines having two benzthiazole nuclei.HAMER THE CYANINE DYES 339 that this eleven-membered chain is the longest reached no cyanine prepared from (XXXVI ; n = 3) having been recorded. A limiting factor to the infra-red sensitising properties of the polymethin- cyanines 1049 105 is the instability of the dyes. Thus the keeping properties of infra-red sensitised plates are poor and the sensitivity decreases with increasing wave-length,l06 so that for plates sensitised to 8500 9500 and 10,500 A. respectively the attainable sensitivities are as 100 10 l.107 It has been stated that both symmetrical and unsymmetrical ennea- methincyanines having an unsubstituted chain are obtainable by con- densing (XXXVIII ; n = 3) with quaternary salts having a reactive methyl group ; the preparation and infra-red sensitising properties of four sym- - 4- (XXXVIII.) RR’N*[CH:CH*]nCH :NRR’ X metrical dyes of the benzthiazole or benzselenazole series have been described ; lo8 a symmetrical hendecamethincyanine of the benzthiazole series similarly obtainable from (XXXVIII ; n = 4) is recorded as sensitis- ing three times as strongly as its chain-substituted acetoxy-derivative.1O9 Effects of Lengthening the Methin Chain.-In the cyanines a chain of conjugated double bonds connects the two nitrogen atoms of which only one is quaternary this implying an odd-numbered chain of niethin groups.For given nuclei the length of this chain affects the position of the absorption and therefore of the sensitising maxima which shift about 1000 A.into the region of longer wave-length for each vinylene (8CH:CI.I.) increment. Thus for example a methincyanine having two benzthiazole nuclei gives an alcoholic solution with its absorption maximum at 4220 A. and may be used for sensitising gelatino chloride photographic plates (Plate 2A). Passing the trimethincyanine with its sensitising maximum in the green (Plate 2B) and the pentamethincyanine sensitising in the deep red and near infra-red (Plate 2C) the heptamethincyanine gives a solution with its absorption maximum at 7620 A. and sensitises to infra-red light (maximum at 7950 A.) (Plate 20) the ennea- and hendeca-methincyanines sensitising still further into the infra-red.The effects of chain length on absorption and sensitising have been studied by chemists throughout the world,ll0-114 and soiiie interesting data recorded. The valuable work of L. G. S. Brooker and his co-workers 113 has already been discussed in Quarterly Reviews.115 loo L. G. S. Brooker and G. H. Keyes Phot. J. 1935 75 191. lo5 W. Dieterle and W. Zeh 2. wiss Phot. 1935 34 245. lo* J. Eggert Chern. Z. 1934 58 397. 107 A. Seyewetz La Nature 1935 No. 2962 300. lo8 I.G. Farbenind. A.-G. B.P. 485,623-4 ; 503,337/1936 ; 512,470/1937. lo9 W. Dieterle and 0. Riester 2. wiss Phot. 1937 36 141. l 1 0 N. I. Fisher and F. 31. Hamer Proc. Roy. Xoc. 1936 A 154 703. ll1 B. Beilenson N. I. Fisher and F. M. Hamer {bid. 1937 A 163 138. 112 T. Ogata Proc. Imp. Acad. Japan 1932 8 421. 113 L. G. S. Brooker F.L. White G. H. Keyes C . P. Smyth and P. F. Oesper J . Arner. Chern. SOC. 1941 63 3192. 114 S. M. Soloviev J . Phys. Chem. Russia 1945 19 451. 116 A. Maccoll Quart. Reviews 1947 1 16. 340 QUARTERLY REVIEWS Zinc dust in py-ridine in the presence of a trace of acetic acid reduced the mono- tri- penta- and hepta-methincyanines of the indolenine series to leuco-bases ; the tri- but not the hepta-methincyanine could be regener- ated by oxidation.l16 With these same dyes chromatographic adsorption increased step by step on ascending the series from met hin- to hepta-niethin-cyariine . I 1 7 Certain cyanines with @-linked pyrrole nuclei e.g. (XXXIX ; n = 1 or Z ) are ex- we[J-j--rcH:cH3 n c ' O D \,A I- Et Et ceptional in having the nuclei connected by dimethin and tetramethin chains though the - chains joining the nitrogen atoms are odd-numbered.l18 Variations connected with the Nuclei.-The nature of the nuclei has a profound influence on the position of the absorption maximum of a cyanine of given type. Thus the methincyanine with two benzoxazole nuclei absorbs in the ultra-violet that with two benzthiazole nuclei absorbs in the blue and cyanine itself with two 4-linked quinoline nuclei in the yellow. The 4-linked quinoline nucleus is the most strongly bathochromic. As discussed in an earlier review,l15 L. G. S. Brooker and his collaborators studied the absorption of unsymmetrical dyes and using resonance as a basis of classification arranged the various nuclei in order of basicity.119 120 In the series having only one hetero-atom in the nucleus the use of 1 -iodoisoquinoline alkiodide gave methincyanines ; l2 subsequentsly not only 1 - but also 3-methylisoquinoliniuni salts were used as starting points.122 Four dyes prepared from 9-methylphenanthridine which is a derivative of both quinoline and isoquinoline gave no sensitisation.123 The various polycyclic nuclei may be regarded as substituted simple nuclei.After cyanines with heterocyclic nuclei other than quinoline had been made patenting activity followed. This literature contains for instance statements that improved sensitising is produced in specified types by the introduction of groups such as amino alkylthio and alkylseleno into the benzthiazole and of 5-phenyl into the benzoxazole niicleus.124 It is ah present not possible to assign a generally beneficial influence to particular substituents.On the other hand the generalisation that nitro-groups depress sensitisation seems acceptable an example being the conversion of the sensitiser (IX ; R' = H) into a desensitiser by introduction of two 6 - n i t r o - g r ~ u p s . ~ ~ ~ It has interested scientists t o introduce a given sub- 11* R. Kuhn and A. Winterstein Ber. 1932 65 1737. 117 P. Ruggli and P. Jensen HeEv. Chim. Acta 1935 18 624. 11* L. G. S. Brooker and R. H. Sprague J . Amer. Chenz. XOG. 1945 67 1869. 119 L. G. S. Brooker G. H. Keyes and W. W. Williams ibid. 1942 64 199. 120 L. G. S. Brooker A. L. Sklar H. W. J. Cressman G. H. Keyes L. A. Smith R. H. Sprague E. van Lare G. van Zandt F. L. White and W. W. Williams ibid. 1945 67 1875. lz1 N. I. Fisher and F. M. Hamer J. 1934 1905. lZ2 F.L. White L. G. S. Brooker and Eastman Kodak Co. B.P. 633,873-4/1946 ; lZ3 L. G. S. Brooker and G. H. Keyes J . Amer. Chent. Soc. 1936 58 659. lZ4 I.G. Farbenind. A.-G. B:P. 400,951/1931 ; 420,971 ; 421,015/1932 ; U.S.P. 2,466,523/1949. 494,715/1937. J. D. Kendall B.P. 543,993/1940. ECAMER THE CYANINE DYES 341 stituent into different positions in the dye molecule and observe the effect on absorption or sensitising maxima. Thus at first this was done with isocyanines *% 44 and 2 2'-trimethincyanines ; 259 71 more recently sym- metrical trimethincyanines having two 4- or 5-methylbenzthiazole 126 or two 4- 5- 6- or 7-methoxybenzthiazole nuclei,127 and tri- and penta- methincyanines having two 5- or 7-phenylbenzoxazole nuclei 128 have been examined. Dyes having 6-chloro- and 6-bromo-benzthiazole nuclei were compared,l29 and into each of the 6-positions in (IX ; R' = H) a series of electro-positive substituents 130 and one of electro-negative substituents l3I were introduced.With the cyanines an effective way of introducing heavy substituents and thus according to Nietzki's rule shifting the absorption (and sensitising) maximum towards the region of longer wave-length is to increase the complexity of the nuclei ; e.g. trimethincyanines containing two naphtho- (1' 2'- or 2' l'-4 5)thiazole nuclei absorb at a longer wave-length by 350 A. than does that containing two benzthiazole n~clei.1~2 Absorption in a desired region does not by any means guarantee a corresponding sensitising action but cyanines containing these naphthothiazole nuclei did comprise valuable sensitisers ,132 133 Although patents cover the prepara- tion of cyanines from thiazoles of still greater complexity,l34 135 it has been denied that trimethincyanines containing two anthraceno( 1' 2'- or 2' 1'-4 Eifthiazole nuclei sensitise more strongly than does their homologue having two benzthiazole nuclei; 136 nor do cyanines with two anthra- ceno(9' 10'-4 5)thiazole nuclei sensitise as strongly as those with two of the naphthothiazole nuclei mentioned above.46 Turning to simpler nuclei we find that quaternary salts of 2-methylthiazole were used both early 29 47 and in later work 48p 53 64t 121 whilst 2-methyloxazole and 2-methylselenazole salts were also used.53 The absorption maxima of dyes from quaternary salts of 2-methyl-A2-thiazoline lay at a shorter wave-length than those of dyes from salts of 2-methylthiazole.49 Similarly the absorption maxima of cyanines having tetrahydronaphthothiazole nuclei were hypsochromic as compared with those having naphthothiazole nuclei.l3' Replacement of the A2-thiazoline by the dihydro-1 3-thiazine nucleus produced a batho- chromic shift of ab~orption.1~8 The fact that cyanines with sulphur- 126 A. I. Kprianov and E. D. Sitsch Trav. Inst. Chim. Charkov 1936 2 25. l a 7 I. I. Levkoev N. N. Sveshnikov and S. A. Kheifets J . Gen. Chem. Russia lzsA. T. Troschtschenko J . Gen. Chem. Russia 1939 9 1661. lZ9 B. BeiIenson and F. M. Hamer J. 1936 1225. 130 A. I. Kiprianov I. K. Uschenko and E. D. Sitsch J. Qen. Chem. Russia 1945 132 F. M Hamer J. 1929 2598. lS3 L. G. S. Brooker and Eastman Kodak Co. U.S.P. 1,846,300-1/1932; 134 I.G.Farbenind. A.-G. B.P. 396,217 ; 400,401/1931. 136Generd Aniline and Film Gorp. B.P. 697,612/1944. 138 I. I. Levkoev and V. V. Durmaschkina J. Ben. Chem. Russia 1945 15 215. 13' I.G. Farbenind. A.-G. B.P. 427,887/1932. 136F. M. Hmer and R. J. Rathbone J. 1943 243. 1946 16 1489. 15 200. 131 A. I. Kiprianov and I. K. Uschenko ibid. p. 207. 1,969,444/1934. A A 342 QUARTERLY REVIEWS containing nuclei were sometimes vastly superior to those having quinoline nuclei in that they might combine powerful sensitising with anti-fogging a~tion~139-1~1 was doubtless a factor contributing to their thorough investiga- tion. An attempt to obtain improved sensitisers by the introduction of certain sulphur-containing substituents was however unsuccessful.141 The original cyanines carried an alkyl group on each of the two cyclic nitrogen atoms but members with hydroxyalkyl 142 or aralkyl 142 143 groups have been recommended whilst N-aryl-2-methyl-A2-thiazolinium salts have been synthesised making available N-aryl-substituted cyanines of that series.144 Instead of a 3-alkyl group certain cyanines have been made with a 3 4-trimethylene group this depends on the synthesis of a salt (XL; Y = 0 S or Se).145-147 (XL.) m-3.) In place of NN’-dialkyl compounds in the methincyanine series the corresponding NN’-alkylene derivatives e.g.(XLI) have been synthesised. 14* The possibility of cyanhes having heterocyclic imino- instead of EtO C Ph Ph CO,Et (Ems iJlJi- c H Me H H Br- N-alkyl groups was early realised in the pyrrole and indole series and dyes such as (XLII ; R == 0 or 1) * were described ; 149 the sulphate corresponding 139 E.Fuchs Chem. Z. 1933 57 853. 140 M. M. Sobolev M. V. Bondareva and M. F. Evteeva J . Appl. Chem. Russia 1 4 1 A. I. Kiprianov 2. P. Sitnikov and E. D. Sitsch J . Cien. Chem. Russia 1936 1 P 2 L. G. S. Brooker L. A. Smith and Eastman Kodak Co. W.S.P. 2,213,238/1940 ; 143 W. Mees M. Schouwenaars and G. Schwarz Ilijds. Vlaams. Ins. Vereen 1937 144 L. G. S. Brooker and Eastman Kodak Go. W.S.P. 2,441,558/1948. 145 W. Konig W. Kleist and J. Gotze Ber. 1931 64 1664. 146 G. Schwarz M. Schouwenaars and Gevaert Photo-Producten N.V. B.P. 147 L. G. S. Brooker and H. W. J. Cressman J . Amer. Chem. Soc. 1945 67 2046. 148 L. G S. Brooker R. H. Sprague and Eastman Kodak Co. U.S.P. 2,478,367/1949. 149 W. Konig J. pr. Chem.1911 [ii] 84 194; 2. angew. Chem. 1925 38 743. * In formula (XLII) a positive charge has inadvertently been omitted from the 1936 9 335. 6 576. 2,241,237/1941. 6 89. 625,245/1940. right-hand nitrogen atom. HAMER THE CYANINE DYES 343 to (XLII ; n = 1) has recently been made by another method. Methin- cyanines such as (XLIII) mere prepared by condensing a pyrrole having a free a-position with ethyl orthoformate in the presence of acid.150 Cyanines with Substituents on the Chain.-In the methincyanine series 2-iodoquinoline alkiodide may be condensed with cyclic ammonium quaternary salts having a 2-ethyl or 2-benzyl instead of a 2-methyl group giving dyes with absorption at a longer wave-length.l5l* 152 In the trimethincyanine series the preparation through ethyl ortho- acetate of symmetrical meso-methyl dyes having two benzthiazole nuclei 39 was followed by that of others with various higher meso-alkyl as well as -benzyl and -phenyl groups and of some analogues with two benzoxazole or two benzselenazole nuclei.153 Other groups such as meso-2-thienyl 154 and -alkylthiomethyl,l41 were also introduced.The ortho-ester method is applicable to dyes with A2-thiazoline,49 A2-selenazoline,l55 and thiazole 156 nuclei but notable exceptions whether by this or by any other method are those containing quinoline nuclei. ii x - (=.I (=) (=3 A very intriguing reaction is that of a quaternary salt (XLIV ; Y = S or Se) with a benzoxazolium salt (XLV) and sodium dissolved in alcohol whereby the 2-R' group of (XLV) becomes the meso-R' group of the resultant symmetrical trimethincyanine 15' (IX or its selenium analogue).Symmetrical trimethincyanines of three series having in the meso- position an o-carboxyphenyl or a 2-carboxyethyl group were prepared by condensing phthalic anhydride or succinic anhydride respectively with a quaternary salt of type (XLIV; Y = 0 S or Se) in pyridine.158 Into the meso-position of symmetrical trimethincy anines of two series the alkyl or aryl group R' could be introduced by condensing a quaternary 150 A. H. Cook and J. R. Majer J. 1944 482 486. 151 J. Gotze 2. angew. Chem. 1936 49 563. 152 J. Gotze and W. Schulte Chem. Z . 1938 62 458. 153 L. G. S. Brooker and F. L. White J . Amer. Chem. Soc. 1935 57 2480. 154 I.G. Farbenind. A,-G. B.P. 403,845/1932. 155 F. L. White and Eastman Kodak Co. U.S.P. 1,957,869/1934 ; f,990,682/1935.156 L. G. S. Brooker and Eastman Kodak Co. U.S.P. 1,973,462/1934; 157 I.G. Farbenind. A.-G. B.P. 439,359 ; 439,807/1933. 158 L. G. S. Brooker R. H. Sprague and Eastman Kodak Co. U.S.P. 2,226,166/1940. 1,994,563/1935. 344 QUARTERLY REVIEWS salt (XLIV ; Y = S or Se) with an alkylthio-imide such as PhN:CR'*SMe ; intermediate compounds (XLVI ; Y = S or Se) could also be prepared.159 By condensing quaternary salts (XLIV ; Y = S or Se) with an acyl chloride in pyridine intermediate compounds (XLVII) were prepared. These were convertible into halogenovinyl compounds such as (XLVIII)," whence (XLVI ; Y = S or Se) could be obtained ; thioketones (XLIX ; Y = S or Se) could be prepared from (XLVII) or (XLTIIII).160 By use of appro- priate thioketones meso-cyclbalkyl groups were introduced.161 Trimethincyanines of the pyrrole series such as (L) with an a-R sub- stifuent were prepared by condensing two moles of a pyrrole carrying an imino-group and having the 2- or the 2- and the 5-position free with one each of ethyl orthoformate and a ketone ROCOMe in the presence of acid.Trimethincyanines of t.he indole series e.g. (LI) with a-R and #l-carbethoxy- substituents were prepared by condensing an indole carrying an imino-group and having its 3-position free with a #l-keto-ester followed by reaction $O,Et C-CMe k - CH-C a N > M e MeC7 \$ H H Br- H Br' H fL.1 (ns of the product with a 2-methylindole-3-aldehyde (similarly with a pyrrole-2- aldehyde). 50 In the pentamethincyanine series derivatives with various electro- negative meso-substituents were the earliest known and have been discussed.82 Recently a meso-cyano-group has been introduced by condensing (XXVI) with cyanoacetic acid.162 A methyl group has been introduced into every possible position in the chain 8 4 p 899 163 164 and the effect on absorption and sensitisation studied.Hepta- and poly-methincyanines with acyloxy-substituents have been c0nsidered.1~~ With the heptamethincyanines the effect of introducing a methyl group into various positions in the chain has been studied,98~ 165 and heptamethincy anines having y -halogeno- 166 8-alkoxy- ,167 and 6- cyano- substituents 162 have been made. Certain symmetrical trimethincyanines e.g. (LII) and (LIII) in which 159 I.G. Farbenind. A.-G. B.P. 412,309/1932. 160 L. G. S. Brooker F. L. White and Eastman Kodak Co. U.S.P. 2,112,139-40/1938 ; 2,231,659 ; 2,233,509-10 ; 2,243,081/1941 ; 2,315,498/1943 ; 2,369,646-7/1945.L. G. S. Brooker G. H. Keyes and Eastman Kodak Co. U.S.P. 2,441,529/1948. l*SF. P. Doyle J. D. Kendall and Ilford Ltd. B.P. 620,801/1947. 163 Z. P. Sitnik and B. S. Steingardt J . Appl. Chem. Russia 1936 9 1842. 164 J. D. Kendall H. W. Wood and J. R. Majer B.P. 553,143-4/1941. 165 F. M. Hamer and R. J. Rathbone J. 1947 960. 1~ A. Corbellini and R. FUSCO Red. Ilst. Lomb. Sci. Lett. 1935 68 961. 167 J. D. Kendall B.P. 526,684/1939. * In formula (XLVIII) the double bonds of the benzene ring have inadvertently been omitted. XAMER !FHE CPANINE DYES 345 the chain forms part of a homocyclic nucleus were prepared by condensing quaternary salts having a reactive alkylthio-group with cyclopentadiene or indene ; subsequently intermediate compounds and hence unsymmetrical trimethincyanines with the cyclopentadiene nucleus were c1aimed.lss The cyclohexene ring was inserted as part of the pentamethin chain giving e.g.(LIV) by condensing a quaternary salt having a reactive methyl Et Br' a z ! a Et M e I' Me I- fir I' E t group with a cyclohexane-1 %&one or with a 3-alkylthiocyclohex-2- enone.ls9 Pentamethincyanines in which the chain formed part of a benzene nucleus e.g. (LV) were prepared by eliminating acid from a diquaternary salt in which two benzthiazole nuclei were linked by *(p-)CH,*C,H,*.170~ 171 Cyclisation of the pentamethin chain had a hypso- chromic effect upon absorption. 1- Special types of chain-substituted methin- and trimethin-cyanines e.g. (LVI) and (LVII) were prepared through cyclic ammonium quaternary salts in which a 2 3-di- 2 3-tri- or 2 3-tetra-methylene group played Is* Idem B.P.431,142 ; 431,187/1933 ; 526,892-3/1939. 16* J. D. Kendall F. P. Doyle and Ilford Ltd. B.P. 595,783-5; 604,217/1945. 170 L. G. S. Brooker R. H. Sprague and Eastman Kodak Go. U.S.P. 2,356,445/1944. A. I. Kiprianov I. K. Uschenko and A. L. Gerschun J. #en. Chem. Russia 1944 14 865. 346 QUARTERLY REVIEWS the part of the 2-methyl and N-alkyl 173 Other special types of chain-substituted trimethincyanines such as (LVIII) and (LIX) were prepared by making use of the reactive homocyclic methylene group of quaternary salts of dihydro-p-quinindene. 17* my"* "*YyJ-J C - CH:CH - C N C-CH='C \; Mt? I- Me I- Me (ax.) MycD fJ-JCY% Me (ma Azacyanines.-The possibility of preparing methincyanine analogues with the nuclei linked by nitrogen was realised when the first azamethin- cyanine (LX) was synthesised through a quaternary salt of N-acetyldi-2- quinolylamine.As compared with the methincyanine its absorption was hyp~ochromic.1~~ A general method for preparing azamethincyanines consisted in condensing a quaternary salt having an alkylthio-group either with one having an amino-group in the reactive position or with ammonia ; a base having a reactive amino-group could be used instead of a salt giving a base of which the azacyanine was the quaternary ~a1t.l'~ The first a/?-diazatrimethincyanines e.g. (LXI) synthesised by condens- ing a nascent 2-formyl quaternary heterocyclic ammonium salt with the hydrazone of a heterocyclic ketone (such as 3-ethyl-2-benzthiazolone hydrazone) were found to have desensitising pr~perties,l~~ and the method was extended to others of this type.178 aa'-Diazatrimethincyanines e.g.(LXII) were synthesised from a heterocyclic base having a 2-amino-group f m.1 cm.) and ethyl orthoformate via a di-heterocyclic-substituted formamidine or else from a quaternary salt having n reactive alkylthio-group and an amidine ; 176 others were prepared by the action of ethyl orthoformate and 172 G. Schwarz Natuurwetensch. Tijds. 1937 19 243. 173 P. de Smet and G. Schwarz ibid. 1939 21 271. 17p A. B. La1 and V. Petrow J. 1948 1895. 176 F. M. Hamer J. 1924 125 1348. 176 J. D. Kendall B.P. 447,038 ; 447,109/1934 ; 461,668/1936. 177 K. Fuchs and E. Grauaug Ber. 1928 61 57. 178 N. I. Fisher and F. M. Hamer J.1937 907. HAMER TBlE CYANINE DYES 347 pyridine on 2-aminoquinoline alkiodide.178 As compared with the trimethin chain a@-diaza-groups produced a hypsochromic shift of about 800 A. but an cca'-diaza-structure caused one more than thrice as great .178 Symmetrical and unsymmetrical meso-alkyl-a-azatrimethincyanines e.g. (LXIII) were made by condensing a heterocyclic base having a reactive amino-group with an ortho-ester and treating the quaternary salt of the product with a salt having a reactive methyl group and with p ~ r i d i n e . l ~ ~ Some aza- cyanines possessed photographic sensitising properties.176 The generalisa- tion was made that desensitisation occurs when the nitrogen atom of the chain is connected to one or more of the heterocyclic nitrogen atoms by an even-numbered carbon chain.180 Several azadimethincyanines of a type subsequently prepared e.g.(LXIV) having two such even-numbered carbon Me I- a N > P h C-N m a Me I- f=Jyp H p Et EL (Lxp.) chains did in fact desensitise. They were synthesised by condensing a 1 2-disubstituted 3-nitroso-indole with certain quaternary salts having a reactive methyl group,l81 or alternatively a 1 2-disubstitu ted %amino- indole with e.g. quinoline-2-aldehyde followed by conversion of the base into a quaternary salt .I82 Cyclisation of 3-acylamido-l-alkyl-(or aryl-)2-phenyl- indoles gave indoloisoquinolines ; their methiodides had a reactive methyl group whence azadimethincyanines such as (LXV) were obtainable.183 Bases of which Cyanines are the Quaternary Salts.-In the synthesis of (VIII) the penultimate step was the base of which this cyanine is the (Em3 (=.I quaternary ~ a l t .~ 3 Some years later the base (LXVI ; Y = S ; R = Me ; n = 0) of which (X) is the quaternary salt was also made.184 A general method of preparing bases such as (LXVI) consisted in condensing a 2- or 4-methylquinoline with a heterocyclic quaternary salt having a reactive alkylthio- or ,&acetanilidovinyl group ; some of them were ~ensitisers.18~ The scope of the method was extended in that bases e.g. (LXVII) were synthesised having the alkyldihydro-structure in the quinoline nucleus. 179 J. D. Kendall and D. J. Fry B.P. 544,646/1940. 180 J. D. Kendall Proc. 9th Internat. Congr. Phot. Paris 1935 227. 181 F. G. Mann and R. C. Haworth J. 1944 670. Huang-Hsinmin and F. G. M m J. 1949 2903. Idem ibid.p. 2911. 18* L. M. Clark J. 1936 607. lS6 J. D. Kendall B.P. 456,362/1935; M. Brtrent and J. D. Kendall B.P. 477,983,4936. 348 QUARTERLY REVIEWS The absorption maxima of such bases compared with those of the related cyanines seemed erratic but were usually hypsochromic.186 The study of this subject in connection with the theory of colour has already been re~iewed.11~ It is interesting that 2-methylperinaphtho-1 3-thiazine resisted quaternary salt formation but by its reaction with quaternary salts having a 2-/?-acetanilidovinyl group gave bases of trimethincyanines. l87 Similarly Z-alkylthio- and 2-methyl-4 5-benz-1 3-thiazine resisted quaternary salt formation and were used for synthesising bases of the methin- and trimethin-cyanine series. ls8~ lS9 Certain symmetrical cyanines e.g.(IV) (VIII) or (IX; R' =H) when heated with a high-boiling base such as diethylaniline lost RX giving the dye bases.lgO By condensing a heterocyclic nitrogenous base having a 2-cyanomethyl group with a heterocyclic quaternary ammonium salt having a suitable reactive group a-cyano-substituted bases e.g. (LXVIII ; n = 0 1 2 or 3) were obtainable; on acid hydrolysis the cyano-group was replaced by hydrogen. 191 Similarly methin and trimethin bases having an a-COR group were synthesised and hydr0lysed.1~2 Some Salts related to Cganines.-Quaternary heterocyclic ammonium salts having a reactive methyl group are capable of condensation with aldehydes and the patent literature dealing with cyanines from these various salts contains numerous references to styryl compounds some of which are sensitisers.When an aldehyde such as p-dimethylaminobenz- aldehyde is used the products e.g. (LXIX) are indeed closely related to the cyanines but because only one nitrogen atom is cyclic are not regarded as true cyanines. Again condensation of a quaternary heterocyclic ammonium salt having a 2-/?-acetanilidovinyl group or of either of the two higher vinylene homologues with a primary or secondary non-aromatic l*@ L. G. S. Brooker R. H. Sprague C. P. Smyth and G. L. Lewis J. Amer. Crhem. 187 F. M. Hamar and R. J. Rathbone J. 1943 487. la* B. Beilenson and F. M. Hamer J. 1942 98. lag B. Beilenson F. M. Hamer and R. J. Rathbone J. 1945 222. Gevaert Photo-Producten N.V. B.P. 477,990/1935. lD1 L. G. S. Brooker R. H. Sprague and Ewtman Kodak Co. U.S.P. 2,345,094/1944; le2 A.van Dormael and J. Nys Bull. SOC. chim. Belg. 1948 57 547. Soc. 1940 62 1116. 2,393,743/1946. HXMER THE CYANINE DYES 349 amine gave polymethin compounds e.g. (LXX ; 9% = 1 2 or 3) which comprise sensitisers and are closely related to the cyanines.lg3 meroCyanines.- 1-----t I - - - - 1 &R[*CH:CH],*b[ :CH*CH],:b *CO*kR' merocyanines of the above constitution (m = 0 1,2 or 3 ; rt = 0 or 1) l94 must be considered before trinuclear cyanines can be dealt with. They are non-ionic dyes prepared by condensing intermediate compounds which would give cyanines with substances having a reactive cyclic methylene group ; they comprise valuable photographic sensitisers and were discovered in two countries ; lg5 lg6 they are intermediate in structure between the cyanines and the oxonols formulated below (m = 0 1 2 or 3) and have \ / \ / I II * I/ I C*CH[ :CH*CH] :C H O G C*OH oc C:CH[*CH:CH],,*C 7" \ / \ been named ~ystematically.~~~ As with cyanines so also with mero- cyanines various unusual intermediate compounds have been used and chain-substituted dyes made.Polpudear merocyanines have been built up. !hinuclear Cyanines.-Trinuclear cyanines prepared by taking advantage of both reactive groups in a 2 4-disubstituted quinoline alkiodide have already been mentioned. 51 A very interesting observation was that of the reactivity of the meso- methyl group in (IX ; R' = Me) e.g. with p-dimethylaminobenzaldehyde it gave the chain-substituted trimethincyanine (LXXI) .I97 Subsequently (IX ; R' = Me) with heterocyclic quaternary ammonium salts having a reactive alkylthio-group etc.gave trinuclear cyanines such as (LXXII) .l9* X' 6 i N Me 1- SEt (LXXL.) ( LXXII.) Neocyanine was announced in 1926 as sensitising more powerfully beyond 8000 A. than any dye then known ; lg9 it had been made by treating 193 F. L. White G. H. Keyes and Eastman Kodak Co. U.S.P. 2,166,736/1939; 2,263,749/ 1941. 184 F. M. Hamer and B. S. Winton J. 1949 1126. 196 J. D. Kendall B.P. 426,718/1933. lS6 L. Cr. S. Brooker and Eastman Kodak Co. U.S.P. 2,078,233/1937. 198 L. G. S. Brooker F. L. White and Eastman Kodak Co. U.S.P. lS9 M. L. Dundon A. L. Schoen and R. M. Briggs J. Opt. SOC. Amer. 1926,12 397. T. Ogata Bull. Chem. Soc. Japan 1936 11 262. 2,282,115/1942. 350 QUARTERLY REVIEWS a lepidinium salt with iodoform and alkali.200 It also arose as by-product in making a 4 4'-trimethincyanine by the ethyl orthoformate and pyridine method,37 which on modification gave greatly improved yields of neo- cyanines from lepidinium salts.Analyses showed that the neocyanine molecule had been formed from three molecules of lepidine alkyl halide with elimination of hydrogen halide and that either one or two additional carbon atoms had entered. Based on the former supposition the first neocyanine formula showing it as a @-substituted 4 4'-trimethincyanine was proposed.201 This formulation was applied to other dyes made from alkiodides of various heterocyclic bases with orthoformic ester or diphenyl- formamidine in the presence of different condensing agents,202 but when neocyanines of the thiazole series were described an alternative was suggested 2*3 which had the advantage of explaining the stability of the neocyanines to alkali.In 1935 neocyanines were formulated as having an unbranched penta- methin chain with three similar nuclei symmetrically attached so that I' I' (=I two trimethin chains were also present.204 According to the modern equivalent of this formula the two units of positive chargo of the cation are distributed over the three basic groups the molecule being regarded as a resonance hybrid of three canonical structures of which (LXXIII) is one. The synthesis of such trinuclear cyanines was effected starting from various heterocyclic methylene bases e.g. (LXXIV) or their anilo- methyl derivatives which on heating with ethyl N-phenylformimidate and zinc chloride gave intermediate dianilo-compounds e.g (LXXV) which were condensed with the usual salts having a reactive methyl group.The synthetic dyes having three similar heterocyclic nuclei were in three instances identified with neocyanines which had arisen empirically. By condensing a dianil with a salt having a reactive methyl group but a different nucleus there resulted trinuclear cyanines with two similar nuclei and one other e.g. (LXXVI) the two similar ones being at the ends of the penta- 2oo H. T. Clarke and Eastman Kodak Co. U.S.P. 1,804,674/1931. 201 F. M. Hamer J. 1928 1472. 202 T. Ogata Proc. Imp. Acad. Tokyo 1932 8 503 ; 1933 9 602 ; Bull. Inst. Phys. 203 L. G. S. Brooker and Eastman Kodak Co, U.S.P. 1,969,445/1934; 204 W. Konig Z. wiss. Phot. 1935 34 15. Chem. Res. Jupan 1934 13 497. 1,994,562/1935. HAMER THE CYAKINE DYES 351 methin chain.Under milder conditions the dianils reacted with quaternary salts to give anilomethyl intermediate compounds such as (LXXVII)," having two nuclei; from these were prepared first neocyanines with two similar nuclei and one other the two similar nuclei lying at the ends of a trimethin chain and secondly neocyanines with three dissimilar nuclei. I' E t ( Lxxzn.) (=) By condensing one mole of a dianil with two of a substance having a reactive cyclic methylene group there resulted dyes e.g. (LXXVIII) having the trimethin chain of an oxonol and two of the chains characteristic of dimethinmerocyanines. They also gave intermediate compounds e.g. (LXXIX) which are anilomethyl-dimethinmerocyanines. These could be used to give a dye of the same type as (LXXVIII) but with the three nuclei all different.On the other hand condensation with a quaternary salt having a reactive methyl group gave dyes e.g. (LXXX) characterised by This kind of dye which is simultaneously a tri- the chain :C methincyanine a dimethinmerocyanine and a tetramethinmerocyanine could also be spthesised from the other type of intermediate compound . /CH \CH:CH. .- I Et CH=C--S I # I 1 oc ,cs (-3 E t (=s I' Et (LXXVII). the neocyanine having the chain :C In each of thirty-seven instances the absorpt,m max++num of lay at a shorter wave-length /CH:CH* \CH:CH~ * In formula (LXXVII) a positive charge has inadvertently been omitted from the right-hand nitrogen atom. 352 QUARTERLY REVIEWS than that of the corresponding pentamethincyanine ; in forty-five of forty- seven instances that of the neocyanine lay at a longer wave-length than that of the corresponding trimethincyanine-in the two exceptions the A2-thiazoline nucleus was concerned.As sensitisers the neocyanines were disappointing which is partly explained by the fact that since the announce- ment of the first the standard of infra-red sensitisers has soared through the discovery of the penta- hepta- and poly-methincyanines.205 Another preparation of neocyanines consisted in heating a trimethin- cyanine with a quaternary salt having a reactive methyl group and ethyl trithio-orthoformate in acetic anhydride ; thus (XI1 ; Y = 0) with 2-methylbenzoxazole ethiodide gave the benzoxazole analogue of (LXXIII). From unsymmetrical trimethincyanines neocyanines with two or three different nuclei could be prepared. The predominant formation of only one neocyanine from an unsymmetrical trimethincyanine was remarkable entailing the working out of the constitution of each product by an alterna- tive synthesis.206 However the fact of there being only one product makes this method a convenient preparative one.Inclusion of acid amongst the reactants led to the production of intermediate compounds such as (LXXXI ; n = 0) which led to neocyanines.207 SEt Intermediate compounds e.g. (LXXXI ; n = l) similarly prepared from pentamethincyanines yielded with quaternary salts having a reactive methyl group higher vinylene homologues of neocyanines.207 Photographic Sensitisation.-The optimum quantity of sensitiser varies from 5 to 100 parts per million of photographic emulsion; above the optimum desensifisation and fogging may become serious.The longer the wave-length of sensitisation the less the amount of sensitiser required ; this has been attributed 19 to the efficiency of energy-transfer from dye to silver halide at longer wave-lengths. Contrary to the theory that the sensitising process is a true photochemical reaction of a dye unstable to light it seemed probable that the sensitiser facilitated transfer of an electron from Br- to Ag+ without itself undergoing decomposition.20* A determina- tion of the number of silver atoms formed during exposure of infra-red sensitised emulsions led to the conclusion that at wave-lengths of 7000 8500 9500 and 10,500 A. each sensitiser molecule acted at least 5,90,160 and 8 times respectively in causing decomposition of one silver bromide molecule.209 206 F. M. Hamer R. J. Rathbone and B. S. Winton J. 1947 954 1434 ; J. 1948 1872; J. 1949 1113. 206 J. D. Kendall and J. R. Majer J. 1948 690. 207 J. D. Kendall F. P. Doyle and Ilford Ltd. B.P. 638,023-4/1947. 208 S. E. Sheppard R. H. Lamberti and R. D. Walker Nature 1937 140 1096; 1938 142 478; J. Chem. Phy&s 1939 7 426. J. Eggsrt W. Meidinger and H. hens Helv. Chim. Acta 1948 31 1163. HARIER "HE CYANINE DYES 353 Since the amounts of dye producing optimum sensitisation were found to be directly proportional to the silver halide surface it was concluded that the dye molecules were adsorbed in a unimolecular layer."O According to another view the first layer of dye was adsorbed through the polar nitrogen atoms the molecules projecting upwards from the surface whilst a second layer might be oppositely orientated.211 Resonance within conjugated systems results in a tendency towards planarity of the molecule.For optical sensitising it appeared that a planar configuration of the dye molecule was necessary.212 With one symmetrical cyanine it was found that the replacement of two imino- by two methylimino- groups necessitated a departure from planarity ; this required energy whence the observed shift of the absorption to longer wave-lengths was e~plained.21~ Subsequently other examples were added of such a shift of absorption caused either by crowding of the nuclei out of a plane or by increased distortion in a molecule already non-planar. Unsymmetrical dyes behaved differently but in both groups non-planar dyes absorbed less intensely than the corresponding planar ones.21* In certain chain-substi- tuted trimethincyanines the planar principal resonance structure was determined by comparison of the absorption spectra with those of simpler related dyes which were planar.215 Supersemitisation.-Photographic emulsion makers know only too well the difficulties of using mixtures of sensitising dyes and with a pair of sensitisers to achieve sensitisation approximating to the sum of two separate sensitisations seemed until fairly recently as much as could be aimed at.However it now appears that in some instances mixing gives a result actually superior to the sum of two separate factors which effect is known as supersensitisation. The first recorded observation of it was in 1918,216 in the improvement of the sensitisation of pinacyanol by addition of aura- mine but this remained an isolated instance.In the earliest of the com- paratively recent patents pinacyanol or a 2 2'-cyanine 217 was named as sensitiser the action of which was improved by adding as supersensitiser a dye having a dialkylaminostyryl group. There followed many instances in which cyanines of one specific type were used in conjunction with cyanines of another or with dyes of a different class; later the distinction as to which was sensitiser and which supersensitiser was not made. Moreover certain sensitising dyes could be improved by addition of supersensitisers which themselves possessed no sensitising power. Most of the dyes which can undergo supersensitisation are those capable of producing two types J. A. Leemakers B. H. Carroll and C. J.Staud J. Chem. Physics 1937 5 893. S. E. Sheppard R. H. Lambert and R. D. Walker ibid. 1939 7 265. 213 Idem Nature 1940 145 386. 213 K. J. Brunings and A. H. Corwin J . Amer. Chem. Soc. 1942 64 593. L. G. S. Brooker F. L. White R. H. Sprague S. G. Dent and G. van Zandt,. Chem. Reviews 1947 41 325. 34 703. 215 P. Bruylants A. van Domael and J. Nys Bull. Acad. Belg. Cl. Sci. 1948 0. F. Bfoch and F. F. Renwick Phot. J. 1920 60 145. 217 C. E. K. Mees and Eastman Kodak Co. U.S.P. 2,075,046-7/1937. 354 QUARTERLY REVIEWS of sensitising according to their existence in the molecular or in the aggre- gat;ed condition. Aggregation of Cyanine Dyes.-The first indication of aggregation in the cyanme group was the observation that Beer’s law did not hold for certain members in aqueous solution.218 Thereafter followed observations that such solutions of 2 2’-cyanine chloride might exhibit an intense narrow absorption band accompanied by resonance fluorescence and methods for inducing this condition were 220 On the one hand it was regarded as a molecular phase intermediate between an earlier one of true solution and a later formation of nematic aggregates of dye molecules that phase in turn being succeeded by crystallisation.221 As the fluorescence and narrow absorption were accompanied by abnormally high viscosity the phenomenon was alternatively explained as due to polymerisation following ionic dissociation.220 From the absorption curves of pinacyanol chloride at different concentrations the presence of mono- di- and poly-molecular forms was inferred ; absence of high viscosity in this instance was explained by postulating spherical particles.222 With the 2 2’-methincyanine adsorption experiments on mica led to the conclusion that the individual cations in the polymer were held parallel with staggering of alternate units by the inductive forces of the p-electrons of the nuclei and of the chain.223 Alternatively direct parallel apposition of the units accounted for the absence of aggregation in certain unsymmetrical cyanines the forces postulated being intermolecular resonance linkages through hydrate water molecules; a new type of nematic phase consisting of plurimolecular filaments was proposed.22* Measurements of the decay of fluorescence were used to determine the degree of polymerisation of methincyanines.225 The absorption curves of a 2 2’-cyanine and of isoQuinoline Red supported conclusions as to their degree of aggregation as determined by osmotic pressure and conductivity measurements.226 Under suitable conditions a cyanine may confer sensitivity in different spectral regions corresponding with the molecular and aggregated states.227 Thus whereas all sensitising cyanines show that type of sensitisation in which the maximum lies 250-500 A. beyond the absorption maximum of an alcoholic solution meso-alkyltrimethincyanines under conditions favour- ing adsorption may show sensitisation at a longer wave-length 228 the capacity of numerous dyes to behave thus has been e~amined.~~g Absorp- tion spectra of trimethincyanines indicated the existence of molecular 218 G. Scheibe with E. Rager Angew. Chem. 1936 49 563.219 E. E. Jelley Nature 1936 138 1009. 220 G. Scheibe L. Kandler and H. Ecker Naturwiss. 1937 25 75. 221 E. E. Jelley Nature 1937 139 631. 222 G. Scheibe A. Mareis and H. Ecker Naturwiss. 1937 25 474. 223 G. Scheibe Angew Chem. 1939 52 631. 224 S. E. Sheppard Science 1941 93 42. 226 F. Katheder Kolloid Z. 1940 92 299 ; 93 28. 226 H. 0. Dickinson Trans. Paraday Soc. 1947 43 486. 227 J. A. Leermakers B. H. Carroll and C. J. Staud J . Chem. Physics 1937 5 878. 228 G. Schwarz Sci. et I n d . Phot. 1939 (2) 10 233. 229 I. I. Levkoev and S. Natanson Acta Physiochem. RUSS. 1946 21 437. HAMER THE CYANINE DYES 355 polymeric and aggregated states.230 Certain cyanines were adsorbed to silver halide in two forms and sensitising data also led to the conclusion that both single and aggregated dye molecules e x i ~ t .~ ~ l In leaving the subject of aggregation we may note claims concerning the preparation of certain polymeric cyanines as sensitisers ; 2329 2a3 starting points were the quaternary salts of bis-heterocyclic bases in which the two cyclic nitrogen atoms were linked by a hydrocarbon. radical and there was a reactive group in the cc-position to each. In this review an attempt has been made to cover primarily the different types of cyanine dyes and the methods used in their synthesis with special reference to their use in photography. The wide range of cyanines of various kinds offers an incomparable opportunity for developing and testing theories about colour. A probable future development is the synthesis of complex cyanines with new types of branched polymethin chains.On the applied side progress may depend not so much on the discovery of new complex dyes as in finding how to make the best of comparatively simple ones which end must surely be reached by fundamental work on sensitisation. 230 S. M. Soloviev J . Phys. Chem. Russia 1945 19 459. 2s1 E. P. Davey Trans. Faraduy SOC. 1940 36 323. 232 C. D. Wilson and E. I. du Pont de Nemours and Co. U.S.P. 2,393,351/1946 ; 233 L. G. S. Brooker and Eastman Kodak Co. U.S.P. 2,461,137/1949. 234 Thorpe’s Dictionary of Applied Chemistry 4th edition “ Cyanine dyes ”. 2,425,772-4/1947 ; 2,465,774/1949.

ISSN:0009-2681    

DOI:10.1039/QR9500400327

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

MELTING AND CRYSTAL STRUCTURE By PROFESSOR A. R. UBBELOHDE MA. D.SC.(OXON.) F.R.I.C. (QUEEN’S UNIVERSITY BELFAST) SIR HUMPHRY DAVY’S original suggestion that melting involves some kind of increase in the heat motion of the molecules has been greatly clarified by the discovery that unlike liquids solids normally consist of long regular arrays of molecules and by the development of thermodynamical and other methods of studying the nature of “increases in heat motion”. 1. Thermodynamical Studies on Melting 1.1 Entropy Changes on Fusion.-For different solids the phenomenon of melting can occur at very different temperatures and can involve very different heats of fusion. For example solid hydrogen melts at 14.0’~. with a heat of fusion of 28.2 kcals./g.-mol. whereas solid platinum melts at 2028”~.with a heat of fusion of 5310 kcals./g.-atom. However the respective entropies of fusion are closely similar (2.01 and 2.62) and indicate that the melting process is very similar for the two solids. The Clausius- Clapeyron equation for phase changes dPf/dTf = ASf/AV’ . . (1.1) relates the temperature T and the pressure PI at which the phase change occurs with and AVf the entropy and volume changes respectively per unit mass (conveniently per g.-mol. of substance). The equation indi- cates that these two thermodpamic parameters might give significant information about the nature of the phase change. This is supported by Boltzmann’s statistical relationship between (a) the entropy of fusion and ( b ) ?Ts and ITL the number of ways of realising the solid and the liquid state respectively though no corresponding generalisation has been proposed with regard to the volume change on melting.The statistical interpretation of melting is discussed in greater detail in Section 4. Empirical correlation of entropies of melting with crystal structure suggests that the entropy of fusion is approximately constant and ranges between 1.5 and 3 entropy units for the following types of compound crystals composed of simple units e.g. single atoms or ionic crystals such as NaCl (Table I) metals (Table II) and crystals of certain polyatomic molecules (Table 111). Even quite large molecules can have comparatively low entropies of fusion when the molecular shape is compact so that intermolecular attrac- tions depend essentially on van der Waals forces acting at the molecular surface (cf.the data in Table IV from ref. 2a). However other polyatomic molecules even when fairly similar in character have very much larger 356 ASf = R In FL/WX per mole . . (1.2) UBBELOHDE MELTING AND CRYSTAL STRUCTURE TABLE I Entropies of fusion of crystals composed of simple (spherically symmetrical) units (V = molar volume of crystal just below the m.p.) 357 Substance. Ne . . . A r . . . Kr . . . X e . . . NaCl. . . KCI . . . Ref. 1 la 1 la 1 1 1 1 3-26 3.35 3.36 3.4 3.35 3.08 24.6 83.6 11 5.95 161.3 1073 1042 0.151 0.144 0.151 0.151 0.30 0.23 TABLE I1 Melting parameters of metals (cf. ref. l b ) C = specific heat at constant pressure. V = specific volume of solid. a = coefficient of expansion. Metal. Li . Na * K . Rb . cs . cu . Ag * Au . Mg * Hg * Zn . Cd . A l .Ga . In . TI . Sn . Pb . Sb . Bi . Fe . Ni . Co-ordina- tion no. in solid. 8 8 8 8 8 12 12 12 12 6 + 6 6 + 6 6 + 6 12 1 4 6 4 + 8 12 4 + 2 12 3 3 - 3 3 + 3 8 12 ASP 1.53 1-70 1.70 1.68 1-65 2.29 2.22 2.29 2.25 2.48 2-57 2.57 2.70 4.42 1-82 1.79 3.35 1.98 5.25 4.78 2.01 2-45 0.0165 0.025 0-0255 0-025 0.026 0.0415 0.038 0.051 0.041 0.042 0.047 0.037 0.060 0.027 0.032 0.028 0.035 -0*0095* - 0*0335* -0*032* 0*007/0*044 - 'p aolidlTf cals. deg.-* x 103. 15.3 19.0 20.8 23.0 24.4 5.5 6.05 5.5 8.2 10-2 11.6 29.0 8.4 21.1 15.4 12.4 14.2 11.6 (7.1) (12.6) 5.8 (4.6) 'p liquid/Tf cals. deg.-l x 103. 15.4 20.2 22-54 25.2 26.4 5.7 6.7 5.2 8.8 11.1 12.0 28.3 8.7 22.5 16.4 11.7 13.1 12.5 7.9 14-0 5.2 _. %lid* deg.-l x 10%. 0.1 8 0.22 0.25 0.27 0.29 0.070 0.081 0.058 0.110 0.113 0.126 0-171 0.099 0.054 0.125 0.126 0,095 0.12 (0.033) (0,040) - 0-057 aliquid* deg.-' x 10%.- 0.275 0.29 0.34 0.37 0.095 0.105 0.069 0-125 0.154 0.164 0.182 0.122 0.126 0.150 0,115 0.13 0.10 0.12 _. - - It may be noted that as the co-ordination number increases or the crystal structure * Cf. Section 2.5 below. becomes more complex the entropy of fusion also increases. Data quoted by A. Eucken 2. angew. Chem. 1942 55 163. J. E. Lennard-Jones and A. F. Devonshire Proc. Roy. Xoc. 1939 A 169 317 ; 170 464. I* 0. Kubackewski Trans. Puruduy SOC. 1949 45 931. B B 358 QUARTERLY REVIEWS TABLE I11 Examples of polyatomic molecules with small entropies of fusion ASr. 3.3 2-2 1.3 1.0 3.2 Substance. 1 Ref. Tf("K.). 437 279 298 241 230 - Diatomic Substance. Camphttne . . . isoCamphane . . . Camphene . . .Pinene hydrochloride Camphor . . . . Camphorquinone . 0 . . N . * co * . HC1 . . HBr . . HI . . H . . ASf. 3.9 2.0 2.8 3.0 3.6 3.4 I PoZvatomic H,S . . D,S . . H,Se . . D,Se . . PH . . CH . . CH,D . CD . . SiH . . CF . . cc1 . . C(CH,) * AS,. 2-0 2-7 2-9 3.0 3.1 3.1 2.06 3.0 3.0 2.9 2.9 1.97 2.5 2-4 2.4 1.8 2.0 2.3 3.03 Tf(" K.). 54.3 63.1 68.1 158.9 186-2 222.3 13.95 187.6 187.1 207.4 206.2 139.4 90-6 90.6 89.2 88.5 84.5 250.3 256.5 Transitions in the solid between m.p. and the lowest temperature of measurement. No of transitions. two one one one three two - two two two two two one one two one one one one Total AStram. TABLE IV Entropies of fusion of large compact organic molecules Tf(" x.). 427 338 324 397 45 1 472 Substance. Bornyhmine . . cycloHexane . . . cycZoHexano1 . . cycloHexanone .. cycZoHsxy1 chloride 4.9 1.5 2.5 2.9 2-1 1.8 - 4.4 4.6 6.12 7.05 5.09 0.8 2.6 3.1 2-3 4.6 4.8 4.4 entropies of fusion. Table V gives values for some of the simpler poly- atomic molecules which show this contrast with the values of A&' in the previous Tables. A clue to this contrast in behaviour in the melting of crystals of the molecules recorded in Tables 111 and IV and of the molecules recorded in Table V is found to lie in the differences in mode of heat intake of the solids below the melting point. Solids recorded in Table V generally show '' normal " vibrational specific heat-temperature curves with unbroken trend in contrast with solids recorded in Tables I11 and IV which generally A. R. Ubbelohde refs. detailed in Ann. Reports 1939 36 159. UBBELOHDE MELTING AND CRYSTAL STRUCTURE 359 TABLE V Examples of simple polyatomic molecules with entropies of fusion greater than 4 e.u.per g.-mol. in order of increasing number of atoms (cf. refs. 1 2 2a) Substance. c02 . cs * Br . Cl . (CN) * HCN . C2H4 f so . % C2H6 * CHCl . SiC1 . ASf. 9.25 8-58 8.41 6.51 9.66 8.89 7.90 7.73 7.70 7-60 8-95 9-06 10.8 216-5 182.3 134.3 161.1 267 172.1 245.3 259.8 103.97 89.9 197.6 210 203.4 0.285 0.116 0.067 - - - - - 0.1 16 0.12 - - 0.14 Substance. CBH . . . . . C,H,Br . . . . . S F . . . . . . (-)-Fenchone . . . a-Aminocamphoric acid Associated in the liquid 1 H,O. . . . . . CH,*OH . . . . tert.-Butyl alcohol . N,H . . . . . NK,. . . . . . N O . . . . . . C,H,.NH2 . . . . C,H,'CH . . . . ASf. 8.44 8.25 9.44 8-67 5-32 12 11.2 Lase 5.25 4.38 5.5 -3.73 6-9 1 5.03 Tf(" K.).278.5 2424 266.9 177.9 218 279 171.2 273.15 172.2 295 274.5 195.4 109.4 show one or more specific-heat maxima below the melting point. These maxima are associated with transitions in the solid which are processes whereby the molecules in the solid increase their average symmetry of orientation. For example oscillation of the molecule between two orienta- tions in the crystal separated by a potential barrier would increase the average symmetry of orientation. In certain cases the molecules probably rotate at temperatures higher than the transition point. Various types of transition have been discussed.3 Solids containing polyatomic molecules which fail to acquire sufficient freedom in the crystal to attain such increased average symmetry do so on melting. On this basis the sum of the entropies of transition and fusion for molecules which have specific-heat maxima in the solid should be of the same order as the entropy of fusion when the whole process occurs at the melting point.That this is so may be seen by summing columns 3 and 7 of Table I11 and comparing the sum with AS in Table V. Intermediate energy intakes below the freezing point may lead to rotation only about certain axes in the crystal. In such cases complete spherical symmetry is not attained by the molecules in the solid and the entropy of fusion has an intermediate value.4 5 A detailed discussion of the cryoscopic behaviour of carbon tetrachloride has been given on this basis.6 K. SchBfer 2. angew. Chem. 1943 56 99; L. A. K. Staveley Quart. Reviews For cyclohexane see 0. Hrtssel L.0. Fischer Bull. SOC. chim. Belg. 1940 49 129. 1949 3 65 ; J. Jaffray Ann. Physique 1948 3 5. and A. M. Sommerfeldt 2. physikal. Chem. 1938 40 B 391 and ref. 4. F. W. Thompson and A. R. Ubbelohde Trans. Faruday Soc. 1950,46 349. For partial rotation in benzene crystals see A. Rowset Compt. rend. 1944 219 6A. W. Davidson W. J. Argersinger and C. I. Michaelis J . Phys. Coll. Chem. 485 546 and A. Kastler and A. Rousset Physical Rev. 1944 71 455. 1948 52 332. 360 QUARTERLY REVIEWS In view of the theoretical considerations below it is interesting to deter- mine whether the entropy of fusion per g.-mol. shows any continuing trend with increasing molecular weight. This is most conveniently studied by examining homologous series in cases where the crystal structure shows the same general features as the molecular weight is increased.One such series is presented by the polymethylene compounds CH,*[CH,],*X. The crystal structure consists essentially of an alignment of molecules with the long chains parallel to one another and normal to or making a small angle with the crystal planes passing through the ends of the chains.' For this series with X = H I etc. there is a trend towards higher entropies as n increases. Entropies and heats of fusion increase uniformly according to the empirical equations ASf =So +nS . . (1.3) A H f = H o + n H . . (1.4) This observation implies that the freezing point Tf = AHf/ASj . . (1.5) must converge to a limiting value or " convergence temperature " Tlb. = Hl/,!j'l above which the substance is liquid however great the molecular weight.For normal paraffins empirical equations proposed . are 8 or Calculated convergence temperatures are illustrated in Table VI. Tg = (0.6085n - 1*75)/(0*001491n + 0.00404) l/Tj = 0.002395 + O.O171/n X. H (paraffis) . . CO,H (neven). . CO,H (nodd) . . TfC K.). 408-417.5 (according to formula) 389 385 When X is CO,H,l* acids with odd and even members for fall on separate curves owing to a transition below the freezing point in the odd- membered series. Similar empirical equations have not yet been obtained for other homologous series which are less easily obtained synthetically and for which the crystal structure depends in a more complex way on n. For polyenes 11 it is interesting to note lla that whereas AHf probably increases ?A. Miiller PEC. Roy. Soc. 1928 A 120 437; 1930 A 127 417.@ K. H. Meyer and A. A. van der Wyk Helv. Chim. Acta 1937 20 1313. lo MT. El. Garner and F. C. Randall J. 1924 125 881. llR. Kuhn and C. Grundmann Bw. 1936 69 224. A. R. Ubbelohde Ann. Reports 1940 37 172. W. E. Garner K. Van Bibbsr and A. M. King J. 1931 1533. UBBELOHDE MELTENG AND CZtYSTAL STRUCTURE 36 1 uniformly with increasing n Ah) remains small so that much higher melting points can be attained. On a semi-empirical basis the origin of this difference has been attributed to the increased freedom of the flexible chains of paraffin derivatives on melting. Conjugation stiffens chains in the polyenes and lessens the gain of freedom on melting. A limiting case for the melting of homologous series is presented by certain polymers. On the basis of X-ray data 12 and from determinations of densities l3 and heat capacities l* as a function of temperature it seems likely that when the molecules of a polymer are flexible and cross linking can be neglected freezing can be described in terms of the mutual alignment of portions of adjacent molecules.Melting takes place over quite a wide range of temperatures and within this range any particular chain molecule may traverse several successive “ crystalline ” and “ amorphous ” or liquid regions. The notion that such molecules melt in more or less independent segments has been put forward to explain the entropies and heats of fusion of homologues (for refs. see ref. lla). Theories of the range of melting have been proposed by various authors.2 15 l6 l7 Attempts have been made to discuss the general relationship between heat of fusion and melting point for different molecular structures but definite conclusions appear to be hampered by values of doubtful accuracy.18 184 1.2 Volume Changes on Fusion,-As has been stated above the magni- tude of the volume change on fusion is less easy to systematise than the entropy change.Volume changes are however significant in the molecular theory of melting which deals with the entropy changes resulting from lattice flaws (section 6). Values of AV’/Vo are illustrated in some of the Tables above. 2. Thermodynamic Properties of the Solid and Liquid Phases in the Immediate Neighbourhood of the Melting Point A study of such properties as the specific heat and thermal expansion of the two phases in the immediate neighbourhood of the transition point can be informative in two different ways.In so far as these propertiesare much the same for the liquid a few degrees above the melting point as for the solid a few degrees below the melting point they lend support to the molecular picture (see below) according to which a liquid near its freezing point is to be regarded as quasi-crystalline; but very near to the actual melting point the values for the solid of and c p = (aH/aT) . . (2.1) a p = ( l / l q ( w / a T ) p . . (2.2) lac. W. Bunn and T. C. Akock Trans. Paraday SOC. 1945 41 317. l3 E. Hunter and W. G. Oaks ibid. p. 49. lP €€. C. Raine R. B. Richards and H. Ryder ibid. p. 56. 1sE. M. Frith and R. F. Tuckett ibid. 1944 40 251. 10 P. Flory J . Chem. Physics 1949 17 223 (where earlier references are given). 17 R. B. Richards Trans.Puraday Soc. 1945 41 127. l 8 J. Pirsch Ber. 1937 70 12. laaIdem ibid. 1936 69 1328. 362 QUARTERLY REVIEWS rise steeply for certain solids as if part of the volume change and heat change on melting took place in a “premelting ” region. 2.1 Interrelation of’the Solid and Liquid Phases on Melthg.-According above the freezing point. The experimental GTT curve stops in this region (cf. Fig. 2). ( b ) Monophase premelting (see below) implies that the GrT curve begins to bend round as it approaches the melting point. As long as a distinction can be made be- tween solid and liquid at the ~ 7 Fra. 1 Liquid to the classical theoryof phase t r a n s i t i o n s equilibrium between two phases is reached at temperatures and pressures where the phases S and L attain the same free energy per unit mass Gs = GA say.Diagrammatically a plot of the free energy G against tem- perature at constant pressure should give quite distinct curves for the two phases with intersection at a sharp angle at the transition point. Classical theory does not ex- UBBELOHDE MELTING AND CRYSTAL STRUCTURE 363 The difference between the tangents at the point of intersection gives However if there is monophase premelting in the solid this implies an abnormally high specific heat C immediately below the freezing point. Since Cp = - T PCLs/aT2 abnormally high values of Cp must mean a curling round of the c;ls curve as the solid phase approaches the melting point 8s illustrated in Fig. 2. This behaviour is very similar to the Cp-T curves in other thermo- dynamic transitions taking place wholly in the solid and emphasises the similarity between the solid and the liquid phase.2.2 Thermodynamic Evidence €or the Quasi-crystalbe Structure of Liquids within a Few Degrees of the Xelting Point.-As an empirical fact it is interesting to note that the values of C a few degrees below the melting point for the solid agree in many cases approximately with values a few degrees above the melting point for the liquid (ref. 11 and cf. Table I1 for metals). This indicates that the thermal molecular processes responsible for heat intake are much the same for the two phases a fact which has to be explained in any molecular theory of melting. Unfortunately data are limited owing to the comparatively small number of calorimetric and dilato- metric observations in the immediate neighbourhood of the melting point.The properties of supercooled liquids are of considerable interest in this connection. If the solid and liquid phases have entirely independent p V T curves the supercooled liquids should not show any break in properties at the freezing point. In a number of cases measurements of density in the supercooled Ziquid show no major break at the freezing point when the liquid is associated-for instance liquid resorcin0l.1~ For liquids which solidify to glasses,20 a rapid change in the properties of the supercooled liquid occurs in the temperature region where the viscosity q reaches about 1013 poises. Such a break is generally far removed from the melting point. However in the case of supercooled liquid diphenyl ether it has recently been claimed 21 that a plot of log Similar effects are claimed for m-chloronitrobenzene and for sodium and potassium.21a If the break in the q-T curve of the liquid can be correlated with any marked change in the slope of the p V T curves this would provide further evidence for the interrelation of the solid and the liquid phase near the melting point. However the interpretation of the observations is not yet clear. When a liquid is capable of being supercooled below the freezing point this must imply a considerable molecular rearrangement on formation of the crystals. The break in the q-T curve can hardly be attributed to this kind of rearrangement in the liquid since if it occurs spontaneously it is difficult to see how supercooling can persist. An empirical correlation has been proposed between the molar entropy of fusion and the molar AHJT.against 1/T shows a break at the melting point. Is J. M. Robertson and A. R. Ubbelohde Proc. Roy. SOC. 1938 A 187 138. 2o G. 0. Jones Reports Progr. Physics 1949 12 133. 21 C. Dodd and Hu Pak Mi Proc. Physical SOC. 1949 62 454. 21a Y. S. Chiong Proc. Roy. SOC. 1936 A 157 264. 364 QUARTERLY REVIEWS difference AC in specific heats between solids and liquids 22 for simple solids such as A H, N, O, Cl, Cd and Sn namely For more complicated solids a second parameter must be introduced in the formula for At$. 2.3 Thermodynamic Evidence €or PremeIting.-FVhen a calorimetrically determined specific-heat curve is carried through the freezing point so as to include the properties of the liquid a sharp increase in heat content is generally observed just below the freezing point.This is conveniently termed '' premelting ',. A corresponding volume anomaly may be observed by dilatometry. One rather trivial cause for the steep rise appears to have been first discussed by H. C. Dickinson and N. S. OsborneYz3 who pointed out that if the liquid contains a small amount of an impurity not soluble in the solid phase then on freezing the concentration of the impurity in the liquid becomes progressively enhanced as more solid separates. As a result the temperature of equilibrium (Tf - AT) between solid and liquid becomes progressively lower as more solid freezes. On the assumption that normal freezing-point depression laws hold the solid freezes over a range of tempera- tures such that the fraction x melted at (Tf - AT) is given by the equation AT/A~'E~ = l/x .. (2.3) This means that part of the latent heat of fusion is taken up below the limiting freezing point (Tf - ATlim.) when the whole mass is just liquid. The additional specific heat AC due to this cause at the temperature (Ti- AT) is AC = AHf(dx/dT) i.e. AQ/AHf = ATIim./AhT' . . (2.4) There -can be no anomaly in specific heat above Tf due to this effect. Simi- larly the anomalous coefficient of expansion due to this effect would be ASf = AC or dV/dT = AVf(dx/dT) (dV/dT)/AVj = ATIi,./A!P . A recent application of the assumption that premelting phenomena are due to impurities soluble in the liquid but not in the solid phase has been worked out for the hydrocarbons cyclopentane methylcycZopentane and methylcyclohexane. On the assumption that the observed increase in specific heat just below the melting point is caused solely by the formation of solution premelting curves have been calculated 24 (cf.Table VII). Actually the limiting freezing point observed does not agree with the value extrapolated on the above hypothesis indicating that factors other than heterophase impurities contribute to the high specific heat below the freezing point or that Raoult's law does not apply rigorously. As equations (2.4) and (2.5) illustrate such heterophase segregation of 23 S. Procopiu Compt. rend. 1948 226 1001. 23 Bull. U.S. Bur. Stand. 1915 12 69. g4 D. R. Douslin and H. M. Huffman J . Amer. Chem. Soc. 1946 68 173. UBBELOHDE MELTING AND CRYSTAL STRUCTURE TABLE VII Computed premelting curues 9.5 28.0 75.7 94.2 365 146.552’ 146.565 146.570 146.573 eyeloPentane.% melted. 37.2 59.5 90-6 100 (Tf - AT) a K. 179.666” 179.678 179.690 179.713 (extrap.) Methylcyclopentane. % melted. 21-5 51.1 70.5 89.9 100 (Tf - AT) K. 130.660O 130,689 130.699 130.705 130.726 (ex trap. ) Methyleyclohexane. % melted. 1 (Tf - AT) a K. impurities in the first part of the solid to melt or the last part of a liquid to freeze can make a large numerical contribution to the specific heat and coefficient of expansion immediately below the limiting freezing point and can also affect other properties of the solid (see section 3). However in certain systems premelting phenomena are observed both in the solid below the freezing point and in the liquid immediately above the freezing point which cannot be attributed to the segregation of impurities in the liquid phase.One example of “monophase” premelting is presented by the paraffins C€€,*[CH,];CH of medium and high chain length where any likely impurities would remain in solid solution and for which deliberate addition of likely impurities does not substantially affect the specific-heat curves. Fig. 3 gives the type of curve obtained.25 Calorimetric evidence for pre- melting with carefully purified materials also includes experiments on trichtoroethane 26 and dimethyla~et~ylene.~~ Similar premelting anomalies have been observed for zinc and cadmium by dilatometry.2s The most likely interpretation is that for certain types of crystal lattice flaws which ultimately lead to greater stability’ of the liquid phase (cf. section S) make an important contribution to the heat content and volume of the solid below the melting point.Residual order in the liquid in the immediate neighbourhood of the melting point may likewise make a contribution to the heat content and volume of the liquid. For such solids freezing may be compared with the aggregation of micelles into a gel with the boundaries between the micelles gradually “ squeezing out ” lattice defects as the temperature falls (cf. also J. W. Oldham and A. R. Ubbelohde 29). Measurements of the thermal conductivity of various organic materials likewise indicate anomalous behaviour near the melting point which is most %li A. R. Ubbelohde Trans. Paraday SOC. 1938 54 289. Dilatometric observations z 6 T. R. Rubin B. H. Levedahl and D. M. Yost J. Amer. Chem. Soc. 1944 66 27 D. M. Yost D. W. Osborne and C. S. Garner ibid.1941 63 3492. 28 W. F. Hachkovsky and P. G. Strelkov Nature 1937 139 715. 2* J. W. Oldham md A. R. Ubbelohde P ~ o c . 303. SOC. 1940 A 176 50. have been made by A. Van Hook and L. Silver J . Chem. Physics 1942 10 686. 279. 366 7500 1400 1300 7200 1'100 4 $ 7000- -? $ 900- 2 QUARTERLY BEVIEWS - - - - - 400 500[ A $ % 20 '40 '-40 ' 4 0 Temperature ' \ r 40 60 80 FIG. 3 Molar specijic heat of octadecane at various temperatures. The first series of points refers to standard octrcdecane the second series to standard octadecane purijied by molecular distillation and also to a sample prepared by a modified @@nard reaction. The setting point is marked with the awow. [Reproduced by permission from Tram. Farachy ~ Q O . 1938 34 292.1 UBBELOHDE MELTING AND CRYSTAL STRUCTURE 367 readily accounted for on the hypothesis of a partial breakdown of the lattice structure by premelting just below the freezing point and by residual struc- ture in the liquid just above the freezing point.30 Elasticity moduli of single crystals e.g.of sodium chloride show a steep decrease as the freezing point is approached.31 Indirect evidence of monophase premelting includes obser- vations such as the absence of a break in a solubility curve on melting of the solid.32 Monophase premelting effects are probably of widespread occurrence and are particularly evident in crystals where the molar entropy of fusion is large. As might be expected on account of the very high molecular weights mono- phase premelting is of dominant importance for polymers such as " Poly- 2.4 Cryoscopic Evaluation of Entropies of Fusion.-It is important to recognise that the cryoscopic evaluation of entropies of fusion from measured freezing-point depressions gives a series of values of ASf which need not correspond exactly with calorimetric entropies of fusion evaluated for the pure substances.Two major factors which may lead to a significant difference in certain cases are (i) If the specific heats of the solid and the liquid are abnormally high around the freezing point of the pure solid (T,),p=o the difference in the heat content AH 7 HIi+ - HsoEd may change to an appreciable extent as the freezing point is lowered by the addition of a second component at a concentration n2 (cf. ref. 29). (ii) If the second component enters into solid solution this may alter AH, particularly if the orientational entropy is affected (cf.ref. 4 which includes references to methods of cryoscopic evaluation of Empirical observations on the change in freezing-point depression with concentration of solute have been discussed for solids such as camphor and benzene.34 When the molecular configuration of the solute molecule lies sufficiently close to that of the solvent formation of solid solutions can be so extensive that no depression of freezing point of the lower-melting by the higher-melting substance is observed. The solidus curve can approximate fairly closely to a straight line joining the freezing points of the two pure substances. Examples include n-hexadecane and n-hexadec-l-ene (cf. refs. in ref. 29) and a wide range of pairs of organic molecules with high symmetry.35 18a Other molecules have been described which form extensive solid solutions.36 Where ordered solid solutions are formed the phase-rule diagram indicates the formation of a crystal compound as for bromocamphor (cf.refs. in ref. 36). 2.5 Effect of Pressure on the Freezing Point.-Within recent years a 30 J. H. Read and D. M. G. Lloyd Tram. Paraday Soc. 1948 44 721. 31 L1 Hunter and S. Siegel Physical Rev. 1942 61 84. 32 Solubility curve of octanoic acid in water F. H. Constable and S. Tegul Nature 33 P. J. Flory J . Chem. Physics 1947 15 6 8 4 ; 1949 17 223. 34 A. V. Brancker S. J. Leach and V. A. Daniels Nature 1944 153 407. 35 L. 0. Fischer Bull. Soc. chim. Belg. 1940 49 129. 36 A. Kofler 2. Elektrochem. 1944 50 104. thene "-15 17 33 1946 15'9 735. 368 QUARTERLY REVIEWS considerable number of freezing parameters have been determined 37 up to 50,000 kg./cm.2 The results are of interest in connection with a number of fundamental problems of melting and crystal structure.(i) The values of some typical freezing parameters are listed in Table VIII. P kg./cm.'. 15,000 . . . 30,000 . . . 15,000 . . . 30,000 . . . TABLE VIII aTfiaP. 0.0064 0.0055 0.0068 0.0047 15,000 . . . 30,000 . . 0.0062 0.0039 I I 15,000 . . 30,000 . . . I I 0.0088 0.0078 AHp kg.cm./g. 1850 2700 2050 3040 845 996 890 1040 900 1100 Asp kg.cm./g./deg. 6.93 7-60) E} Substance. Ethyl alcohol - n-Butyl aIcohol Ethyl bromide n-Propyl bromide Carbon disulphide Whereas AH increases with increasing compression in all the examples given A&'$ decreases for non-hgdrogen-bonded molecules but increases for the alcohols.This suggests that the energy required to produce flaws in the solid always increases with increasing compression (cf. section 6). But apparently compression decreases the dZference in order between solid and liquid for (' normal " molecules and increases it when there are directed bonds such as hydrogen bonds in the solid and the liquid. I n view of the importance of these parameters for correlation of the configurational entropy' of melting with molecular structure (section 6) extension of these observations should eventually give valuable theoretical information on melting. (ii) As the compression is increased the temperature below which a substance is solid normally rises. By analogy with the critical boiling point for the transition liquid -+ gas it might be expected that the transition solid -+ liquid would become continuous above a certain temperature i.e.that no solid could exist above a " critical melting point ". But no definitive evidence has so far been obtained for such a limitation to the phase region for the solid state which would have important geophysical impli~ations.~8* 399 48bt 49b Observations on the melting of helium may 37 P. W. Bridgman Amer. Acad. Sci. 1942 74 399 ; Amer. Scientist 1943 31 23 ; 38 Cf. P. W. Bridgman Amer. Scientist 1943 31 1. s9 J. E. Lennard-Jonw and A. F. Devonshire Proc. Roy. Soc. (a) 1939 A 169 Rev. Mod. Physica 1946 18 27. 317; (b) 170 464, UBBELOHDE MELTMQ AND CRYSTAL STRUCTURE 369 permit investigation of the possibility of a critical melting point over a wider effective temperature range.40 Only four common substances have been found which contract on melting wix.bismuth antimony gallium and water 41 (cf. Table 11). It seems likely that this kind of change is associated with directed bonds in the crystal which form a framework that undergoes partial collapse on melting (cf. ref. 19). 2.6 Kinetics of Melting.-As has been indicated in section 2.1 the forward and the reverse reaction in the rate processes Crystal 7- Melt are only balanced when both phases are present. Under these conditions the change from melt to crystal takes place at the crystal face by a “two- dimensional ” process.42$ 4 3 5 44 The fact that a crystal cannot be readily superheated implies that this rate process proceeds rapidly and without difficulty at the surface of a solid.But in the absence of a crystal nucleus the reverse process of spontaneous crystallisation may be delayed for many days. This has been attributed 43 to the requirement for a ‘‘ co-operative fluctuation ” in the melt for production of a crystal nucleus. Such co-opera- tive fluctuations have a very low probability of occurrence because of the decrease of entropy involved. A further consequence is that when crystal- lisation occurs from the melt spontaneously in the absence of nuclei the least stable of a number of possible polymorphs tends to be produced most readily. This ‘‘ law of successive states ” is explained 43 on the basis that formation of a less stable solid involves smaller entropy fluctuations relative to the melt. Various discussions of the kinetics of melting and crystallisation indicate that growth of a crystal by deposition of layers of atoms or molecules on preformed crystals is quite distinct from the production of fresh crystallisation nuclei.Although it has not proved possible to superheat a crystal appreciably without melting occurring claims that a small but definite concentration of ‘‘ crystallisation nuclei ” persist in the melt are based on a large if some- what vague body of evidence. Such crystallisation nuclei dissolved in the melt cannot apparently pass continuously into the solid but must be separated from it by a region of metastable states. Though such nuclei are important for theories of crystal growth the experimental evidence is as yet too indeterminate to provide much information about 2.7 Melting in Adsorbed Layers.-2.7.1 Melting of Liquids Adsorbed in CapiZZaries.Various experiments indicate that the absorption of a liquid in a sorbent with capillaries leads to a lowering of the melting point which 2.6.1 Persistence of Crystal 2VucEei in the Melt. melting.42 45 46 ; also refs. 14 ref. lla O0 F. A. Holland J. A. W. Huggill G. 0. Jones and F. E. Simon Nature 1950 *l P. W. Bridgman Amer. J . Sci. 1945 243 A 90. O2 Faraday SOC. Discussion “ Crystal Growth ” 1949 5; 4a A. R. Ubbelohde Trans. Faraday Soc. 1937 33 1198. 44 S. S. Penner J . Phys. Coll. Chem. 1948 52 949. 46 “ Supersaturation limits of solutions ’ 7 R. Gopal J . Indian Chem. Soc. 1948 46 L. G. Carpenter L. J. C. Connell and A. B. Osborn Nature 1949 188 23. 165 147. 25 87 443. 370 QUBRTERLY BEVIEWS may extend over several degrees.*‘* 48 The effect might in principle be correlated with the change of internal pressure in a capillary but in addition adsorption may stabilise the liquid relative to the solid phase by imposing a preferred orientation on the molecules in the liquid phase.For example studies of the dielectric constant of ethyl chloride adsorbed on silica gel show that the molecules have their rotation hindered in the adsorbed layer. 49 2.7.2 Melting of Xurface Films on Liquids. Application of the Lang- muir-Harkins-Adam film-pressure technique has revealed that in a limited number of cases films which are “rigid” below a certain temperature become ‘‘ liquid ” above a “ melting point ”. Such “ melting points ” are not far from but have not been quantitatively related with the melting points of the bulk phases.Since molecules in the films are less constrained than in a three-dimen- sional crystal melting might be expected at a lower temperature (but see ref. 50). Actually the lessened constraint may decrease both A€€$ and Ah) and no unique difference of sign between Ti for the crystal and for the film can be predicted. Structural observations (cf. ref. lla) suggest that in the crystals of long-chain compounds ‘‘ lateral ” melting with a large increase of lateral freedom appears to precede “ long-spacing ” melting. Lateral packing in unimolecular films of long-chain compounds is closely similar to the packing in bulk crystals so that much the same melting phenomena might be expected. 2.7.3 Grain Boundary Nelting. It is not yet clear whether the earlier melting at the boundaries of crystal grains e.g.in very pure a1uminium,61 is caused by segregation of impurities at these boundaries or whether the more disordered solid in this region passes into a liquid at a lower temperature than the more highly ordered atoms at the centres of the grains. 3. Other Properties of the Solid and Liquid near the Melting Point in Relation to Molecular Structure As with other transitions in the solid state the origin of the energy intake on melting has been investigated in special cases by making use of a whole range of molecular properties in addition to the thermodynamic parameters described in sections 1 and 2. 3.1 &Ray Investigations on the Structure of Solid and Liquid near the Melting Point.--3.1.1 For crystals composed of arrays of single atoms X-ray studies show that in the liquid the molecules have approximately the same close packing as in the solid (short-range order) but that there is 47 H.Reiss and I. B. Wilson J . Colloid Sci. 1948 3 551. 48 ( a ) W. T. Richards J . Amer. Chem. SOC. 1932 54 479; (b) P. W. Bridgman Physical Rev. 1914 3 126 153; 1915 6 1 94; 1934 46 930. *@ (a) R. McIntosh H. S. Johnson N. Hollies and L. McLeod Canadian J . Em. 1947 25 B 566; (b) F. Simon M. Ruhemann and W. A. Edwards 2. physikul. Chem. 1929 2 B 340 ; 1930 6 B 62,331 ; Simon Trans. Paraday SOC. 1937 33 65. 6o N. K. Adam “ Physics and Chemistry of Surfaces ” Oxford 1941 p. 55. s1 G. Chaudron P. Lacombe and N. Yannaquis Compt. rend. 1948 226 1372. UBkELOHDE MELTING AND CRYSTAL STRUCTURE 371 no evidence of long-range order in a liquid (for recent studies of liquid *metals see refs.52 and 53). Other evidence on premelting suggests that if X-ray measurements on atomic liquids were made sufficiently near to the freezing point rudimentary order of longer range might be observed even in the liquid. Careful tempera- ture control and monochromatic radiation would be required and such experiments do not appear to have been reported. 3.1.2 For crystals made up of more complex molecules X-ray evidence frequently indicates an approximation of the structure of the solid to that of the liquid as the melting point is approached (cf. ref. lla and refs. for cyclohexane in ref. 4). In the case of sodium palmitate the solid “melts” so far as lateral packing is concerned whilst it preserves the long spacing to higher temperatures.64 Studies of the melt really near the freezing point would also be of interest but do not appear to have been carried out.3.2 Mechanical Properties of the Solid and the Liquid near the Melting Point.-A description of the transition from the solid to the liquid state in terms of the change in mechanical properties is complementary to the description in terms of the change in thermodynamic properties. In this connection the most significant mechanical property is the rigidity. Whereas solids can assume finite shear strains in liquids the shear stresses F are relaxed at a rate determined by their viscosity according to the Maxwell relationship where E is the modulus of elasticity and X is the strain (cf. ref. 20) z is the relaxation time dP ds P dt dt z ___ Empirical formula have been proposed for the temperature coefficient of shear strain of solids according to which the rigidity modulus ,u vanishes a t the melting point for example P = Pol3 - (~/53)21 (refs.in ref. lla p. 174). Studies on the viscosity of melts which suggest some survival of crystal- line structure very near the melting point have been referred to in section 2.2. The temperature coeficient of viscosity of mesomorphic liquids is also of interest in giving information about the structure near the melting p0int.54~ No discontinuity is apparent in the compressibility of a substance on melting.37 3.2.1 Acoustic Vibrations in Xolid and Liquid Phases. Closely related to the postulated change in mechanical rigidity there should be a difference in the type of acoustic vibrations which can be present in the liquid com- 63 H.Hendus Naturforsck. 1947 2a 605. 63 0. Kubackewski Tram. Faraday SOC. 1949 45 931. 64 H. Nordsieck F. B. Rosevear and R. H. Ferguson J . Chem. Physics 1948 18 175. 64a R. Schenck 2. physikal. Chem. 1898,27,167 ; cf. the second report on Viscosity and Plasticity prepared by the Committee for the Study of Viscosity of the Academy of Sciences at Amsterdam. 372 QUARTERLY REVIEWS pared with those which are present in the solid. However the fact that C for the liquid a few degrees above the melting point lies close to the value of C for the solid a few degrees below the melting point (see above) suggests that no very clear theoretical distinction can be made between solid and liquid on the basis of different acoustic vibrations (cf. refs. 1 and lla). 3.3 Scattering of Light by Solid and Liquid Phases.-For an increasing number of crystals studies on infra-red absorption or Raman scattering give evidence about the structural changes which occur on melting.These are of two kinds. (a) The orientations assumed by individual molecules show increased freedom on melting and frequently in the premelting region. An illustration of this change is presented by observations on single crystals of benzene in which a limited number of Raman lines of long wave-length (35-100 cm.-l) attributed to a vibration of the molecules in the crystal disappear on melting.56 (b) For molecules capable of changing their shape by intramolecular rotation configurational isomers may be formed in the liquid on melting whereas the crystal structure may incorporate only one isomer.Evidence in favour of such a process has been obtained by comparing the Raman spectra of solid and liquid polymethylene hydrocarbons (n-paraffins). When the solid melts additional Raman lines appear in the liquid phase. These have been attributed to " rotational " isomerism. In the solid only the stretched zigzag form of the molecule is present but on melting a fraction of the molecules isomerise by rotation about C-C bonds to give more-or-less coiled molecule^.^* An analogous possibility has been tentatively suggested for cycibhexane. 57 (c) The Brillouin scattering of monochromatic light by stationary waves in the medium gives a broadened central line for gases a doublet for solids and both a doublet and a central line for liq~ids.~8 Light scattering attributed to rotation of molecules in crystals has been discussed,59 but the correlation with melting is not yet very complete.3.4 Dielectric Constant and Dielectric Losses in the Losses in the Solid and Liquid.-When the molecules in a crystal are carriers of permanent dipoles the onset of rotation or of increased freedom of orientation in the solid is normally accompanied by a large increase in the dielectric constant. For dipole carriers where there is no rotational anomaly in the solid the increase in dielectric constant would be expected at the melting point. Comparatively few investigations have been reported of measure- ments taken through the melting point. However studies on the dielectric constant and dielectric relaxation of long-chain molecules in crystals either 66 A. Friihling J . Physique et Radium 1948 9 88.66 J. G. Aston D. H. Rank N. Sheppard and G. J. Szasz J . Amer. Chem. SOC. 1948 70 3525; D. H. Rank N. Sheppard and G. J. Szasz J . Chem. Physics 1949 17 83 ; N. Sheppard and G. J. Szasz ibid. pp. 86 93 ; San-iehiro Mizushima and Hiroatsu Okazaki J . Amer. Chem. SOC. 1949 71 3411. 67 F. W. Thompson and A. R. Ubbelohde see ref. 4. 68 P. Debye 2. Elektrochem. 1939 45 174. 59 E. Gross and A. Raskin Acta Physicochim. U.R.S.S. 1942 17 127. UBBELOHDE MELTMGF AND CRYSTAL STRUCTURE 373 pure60 or in solid solution,61 give evidence for the increased torsional or rotational mobility' about the long axis of the chain which accompanies other premelting phenomena. 3.5 Diamagnetic Susceptibility.-The diamagnetic susceptibility of many organic substances increases by 5% or more on melting usually with evidence of premelting phenomena G 2 (cf.Fig. 4). No definitive explanation of this effect has been given in terms of the change of structure though it appears to indicate some mutual interference of the bound electron orbits of the molecules in the crystal. This interference is lessened in the liquid. In metals the conduction electrons make an additional contribution to the atomic magnetism which may be positive or negative. A decrease in diamagnetism is observed on melting of metals such as zinc tin (grey) lead gallium or bismuth and an increase for gold and gerrnani~m.6~ This must be correlated with changes of co-ordination number of the metal atoms on passing from solid to liquid,52 but no detailed explanation has been proposed. When the molecules pass through a " liquid-crystal " phase on melting inter- mediate between the solid and the isotropic liquid applied magnetic fields can orient the anisotropic mole- cular clusters in the nem- atic phase.For example in p-azoxyanisole the dia- magnetism falls through a I I t Tem,mrutwe FIU. 4 (Xeneral retationship between susceptibility and temp- erature for a diamagnetic substance in the neigh- bourhood of the melting point. The change in sus- ceptibility may amount to 5% or more. minimum and then rises again to the temp- erature where the liquid becomes isotropic. This is because in the liquid the long dimension of the molecule which is least diamagnetic orients along the lines of force.64 3.6 Ionic Conductivity.-For ionic crystals the big jump in ionic con- ductivity which is generally observed on melting is illustrated in Table IX.cr is the conductivity of the liquid and a that of the solid. This jump in conductivity is directly associated with the large increase in the number of lattice flaws on melting and with the fact that the expansion 6o A. Muller Proc. Roy. Soc. 1937 A 158 403 ; 1938 A 166 316; 1940 A 61 V. Daniel Nature 1949 163 725. 6 2 A. E. OxIey Phil. Trans. 1914,214 109 ; 1915 215 79 ; 1920 220 247 ; Science Progr. 1920 14 588; B. Cabrera and H. Fahlenbrach 2. Physilc 1933 85 568; 1934 89 682; B. Cabrera J . Chim. physique 1940 37 86. K. Honda Ann. Physique 1910 32 1027 ; M. Owen ibid. 1912 37 657 ; K. Honda and T. Ishiwaxa Sci. Rep. Tohoku Imp. Univ. 1915 4 215. 64 C. Mauguin Compt. rend. 1911 152 1680 ; G. Foex J . Physique et Radium 1929 10 421; Trans.Paraday Soc. 1933 29 958. 174 137. cu 374 QUARTERLY REVIEWS TABLE I X Conductivity c h a q e on melting of ionic crystals [a = A exp. (- E/kT)] Substance. KC1. . . NaCl . . K B r . . . AgCl . . AgBr . . PbC1 . . 10,000 3000 7000 30 20 900 Asoiid' 2 x 107 9.5 x 104 2.8 x 103 8.5 x 103 1 x 106 1-4 36.9 27.5 31-3 11.2 10.6 7.4 I I 6.5 7.4 6.3 6.8 4.9 28.6 Emeft x ergs/g.-mol. 1.6 1.1 1.9 0.5 0-4 3.3 ftererences are given in rer. 00 ana ma. in volume on fusion reduces the activation energy for migration of the ions (cf. columns 3 and 6). 3.7 Electrical Conductivity.-In contrast with ionic conductors for TABLE X Effect of melting on the speci$c conductivity of metals Section of ref. 66. a a a b C C C d e fd $7 d d g d Metal. Li Na K cs cu Ag Au Zn Cd Hg A1 Gat Sn Pb Sbt Bit Type.* b.c.c.b.c.c. b.c.c. b.c.c. f.c.c. f.c.c. f.c.c. h.c.p. h.c.p. f.c*c. - - - f.c.c. - - Crystal structure. a A. 3.46 4.24 5-25 6.05 3.609 4.078 4.070 2.65 2.97 4.04 - - - 4.93 - - 1 aab0vem.p. i u below m.p.' _- ______ d A. _~ 3.00 3.67 4-54 5.24 2.55 2.88 2.87 2.65 2.97 2.86 - __. - 3-48 - _. 0-51 0-75 0-72 0-61 0.48 0.51 0.44 0.48 0.51 0.24 0.61 1.72 0-495 0-49 1.43 2-33 * b.c.c. = body-centred cubic ; f.c.c. = face-centred cubic ; h.c.p. = hexagonal t Cf. Table 11. a = edge length of fundamental cube in cubic lattices ; d = nearest interatomic close-packed. distance ; c = other parameter in hexagonal close-packed lattices. - 6* S . E. Rogers and A. R. Ubbelohde Trans. Faraday SO~. 1950 46 1051. 65a AgCl + AgBr W. Lehfeldt 2. Physik 1933 85 717. 66 (a) A. Bernini Cim.1903 [v] 6 21 289 ; Physikal. Z. 1904 5 241 406 ; Nuouo Cim. 1904 [5] 8 262 ; Physikal. Z. 1905 6 74 ; (b) L. Hackspill Compt. rend. 1910 151 305; (c) E. F. Northrup J. Franklin Inst. 1914 177 1 287; 178 86; (d) H. Tsutsumi Sci. Rep. Tohoku Univ. 1918 7 93 ; ( e ) G. Vassura Cim. 1892 [3] 31 25 ; (f) C. L. Weber Wied. Ann. 1885 25 245 ; (9) P. W. Bridgman Proc. Amer Acad. 1922 57 41. UBBELOHDE MELTING AND CRYSTAL STRUCTURE 375 metals there is normally a decrease in electrical conductivity on melting. This is generally attributed to the increased scattering of electrons in the molten metal owing to the greater disorder of the atoms and is thus analogous to the effect of cold working or straining of the metal. The order of magnitude of the change is illustrated in Table 2% and should be compared with the temperature coefficient of electrical conductivity of the solid.For the few abnormal metals (bismuth antimony gallium) the electrical conductivity increaso,s on melting. This is related to the presence of a proportion of unusual “ non-metallic ” bonds in the crystal which are broken on melting and may be correlated with the anomalous sign of AV,lV (referred to in Table 11) for these crystals. 4. Phenomenological Description of Melting From the volume increase which is observed in all simple cases of melting it seems plausible that the packing in the liquid must involve lattice defects or holes. The various lines of evidence indicate that the proportion of defects increases to such an extent on passing from solid to liquid that long- range order breaks down.Simple types of lattice defects may be accom- panied by co-operative defects in some ~rystals,~g Both types of change may be described as a decrease in positional order. When the molecules are not spherically symmetrical orientational order can also decrease on melting. Finally in crystals where the linkages between the atoms are strong as for example in quartz melting may leave strings and networks of molecules linked together in the liquid In such cases the viscosity of the liquid may show an anomalous increase on cooling and in the absence of crystal nuclei the liquid may “ polymerise ” to a glass as an alternative to freezing to a crystal.20 5. Special Types of Transition from Crystal to Disordered State 5.1 Mesomorphic States intervening between Three-dimensional Crystals and Noxi-crystalline Liquids.-Various crystals “ melt ” to “ liquids ” in which optical and thermal evidence indicates residual crystallinity which may be attributed to bundles or clusters of molecules in the liquid with an arrangement within the cluster more orderly than that which prevails in the usual liquids.In such cases a second temperature or ‘cclearing point ” can usually be observed above the normal melting point at which the liquid becomes optically isotropic.67 Both the melting point and the 8’ (a) D. Vorliinder et ul. Ber. 1938 71 B 501 ; ( 6 ) 2. Krist. 1937 97 485 ; (c) P. Chatelain Cornpt. rend. 1936 203 1169; (d) B. Jones J. 1935 1874; ( e ) C. Weygand and R. Gabler Ber. 1938,71 B 2399 ; (f) 0. A. Knight and B. D. Shaw J. 1938 682 ; (9) H. Stolzenberg and M. E. Huth 2.physikul. Chem. 1910 71 641. For other examples cf. D. Vorliinder Trans. Furaduy SOC. 1933 29 907 ; C. Weygand 2. physiku2. Chern. 1943 53 B 75 ; C . Weygand R. Gabler and J. HofCinann ibid. 1941,50 B 124 ; C. Weygand R. Gabler and N. Bircan J . pr. Chem. 1941,158 266 ; C. Weygand and R. Gabler 2. physikal. Chern. 1940 46 B 270 ; C. Weygand and W. Lanzendorf J . pr. Chem. 1938,155,221 ; D. Vorlander et al. Ber. 1938,71 B 501 ; 376 QUARTERLY REVIEWS clearing point are associated with increases in heat,68 entropy and volume. Typical values are illustrated in Table XI. Though the change from crystal to isotropic liquid by stages is unusual it does not seem to present major difliculties of interpretation in terms of structural changes on the basis of the previous sections.TABLE XI Melting and clearing temperatures of Some liquid crystals Section of ref. 67. a b c e d 9 f Compound. 4-(p-Benzylideneamino)diphenyl E t 4- (p-methoxybenzylideneamino) , (ii) p - Azoxyanisole p-Amyloxybenzoic acid 1 -Tetradecylpyridinium chloride Silver bromide cinnamate form (i) M.P. 243-245' 83-85 106 115-116 122 77 259 Clearing point. 254" 139 139 134 148 205 398 5.2 Crystallisation on Stretching or Freezing of Rubber-like Solids.-In certain polymers with rubber-like properties the molecules are normally present in coiled form and give X-ray reflection patterns similar to those obtained from gases or liquids. When they are strongly cooled a crystal structure develops. This change from disorder to order on cooling is analogous to those described in the previous sections except that it occurs over a rather wider range of temperatures.Rubber-like solids present a new phenomenon however in that they also crystallise when stretched a t constant temperature. As with cooling crystallisation by stretching involves an increase in order. The statistical interpretation of this phenomenon has been discussed in numerous papers in view of its technological importance.69 6. Theories of MeIting Any theory of melting must in principle account quantitatively for the coexistence of two condensed phases at the same temperature and pressure. A major aim is to relate the temperature of melting the volume change on melting and the latent heat directly to the intermolecular forces. Calcu- lations on melting in terms of a critical temperature of vibrational instability 193?,70 B 2096.For reviews see Ann. Reports 1931,28 280 ; " General Discussion on Liquid Crystals " Trans. Paraday Soc, 1933 29 881 ; L. S. Orstein Nederl. Tijds. Natuurk. 1940 7 373 ; Proc. Acad. Sci. Amsterdam 1938 41 1046 ; A. S. Lawrence J . Ray. Microscop. SOC. 1938,58 30 ; W. Kast Physikal. Z. 1937 38 627; W Bragg Nature 1934 133 445 ; Proc. Roy. I n s t . 1934 28 57. 68 C . Kreutzer Ann. Physik 1938 33 192 ; K. Kreutzer and. W. Kast Naturwiss 1937 25 233. 6g Cf. the review by K. H. Meyer Chem Reviews 1939 25 137. UBBELOHDE MELTING AND CRYSTAL STRUCTURE 377 of the crystal lattice,70 or a critical temperature of mechanical instability of the crystal under vanishingly small external shear,71 cannot account directly for the thermodynamic parameters which govern the coexistence of two phases.Even when such non-thermodynamic theories contain interesting and valid concepts about the effect of temperature on the properties of the crystalline state the account they give of melting must be basically incomplete. Examination of the experimental evidence (sections 1-5) indicates that many factors may contribute to a thermodynamic parameter such as the entropy of fusion. Typical examples are presented by molecules of complex shape. Calculations on melting also become more complicated when several kinds of interatomic forces affect the stability of condensed phases as in the solids containing hydrogen bonds or in metals. No general thermo- dynamic theory of melting has yet been proposed which covers all types of entropy increase in the Boltzmann expression (equation 1.2).An interesting and suggestive model has been proposed 39a for the limited problem of the melting of a crystalline face-centred cubic solid com- posed of particles with spherically symmetrical force fields. Such a model solid can account for the quantitative aspects of melting of simple crystals like solid argon. It also gives semi-quantitative information about changes of positional entropy on melting of more complex crystals. All the evidence suggests that in a simple crystalline solid even near its melting point the number of lattice flaws is only a small fraction of the total number of atoms and is insuffcient to affect the long-range order appreciably. In the liquid the number of defects is comparable with the number of atoms so that no long-range order can be recognised.In their theory of melting Lennard- Jones and Devonshire 39a3 b simplify the formal treatment by considering only the " interstitial " type of lattice defect. For convenience interstitial " defect " positions are considered to lie on a face-centred cubic lattice which interpenetrates the " normal " lattice like the Na+ and Cl- lattices in sodium chloride. In this model each " normal " or a-site is surrounded by 2 defect or @-sites and vice versa A fundamental parameter for melting is the interaction energy W of each pair of atoms occupying adjacent a- and p-sites ; W is assumed to be a function only of the specific volume. The degree of order at equilibrium is calculated by processes similar to those for order-disorder in alloys. The degree of order Q being defined as Q = NJN where Nu is the number of atoms on a-sites and N the total number of atoms the partition function FQ of the disordered assembly is shown to be FQ = fN x Y(&) exp[- ZNW&(f - &)/kT] .. (6.1) where fN refers to the partition function of atoms in a state of perfect order and the second factor takes account of the disorder. Y(Q) is the number of ways of distributing Nu= NQ atoms on a sites and NB = N( 1 - Q ) 'OE.q. F. A. Lindemam Physikal. Z. 1910 11 609; W. Braunbek 2. Physik 1926 38 549. 7 1 R. Lucas Compt. rend. 1938,207 1408. For other aspects of mechanical theories of melting cf. M. Born Nature 1940 145 741 and ref. l l a p. 174. 378 QUARTERLY REVIEWS atoms on @ sites. Y(Q) = (N!/[NQ!(l - @!]I2. E’or a given volume temperature and energy of interaction PQ has a maxmium given by (2& - 1 ) = tanhZW(2Q - l)/4kT .. (6.2) This equation is always satisfied by Q = 4. When 25W/4kT > 1 there is another root greater than 8. If it exists it corresponds to the maximum of the partition function but when it does not exist the maximum is given by Q = 9. Large values of W/kT give Q w 1 i.e. order nearly perfect and values W/kT < 4/25 give Q = & Le. complete disorder. This property of equation (6.2) gives the essential possibility of equilibrium between a highly ordered and a disordered condensed phase. In view of the factorising of the partition function in equation (6.1) the free energy internal energy and entropy’ can be separated into the contribution from the order (single prime) and from the disorder (double prime) A = A’ + A” U = U’ + U” etc.Since - A/NkT = (1/N) log (FQ max.) - A’/NkT = log f - A”/NkT = - [ZW&(1 - Q)/kT] - 2(1 - &) log ( 1 - Q) - 2Q log Q (6.3) and similarly 77’‘ = ZNW&(l - &) . . (6.4) S“/Nk = - 2(1 - &) log (1 - &) - 2& log Q . (6.5) The pressure p = - (dA/dV) can similarly be expressed as the sum of the pressure p’ of an ordered assembly plus the additional pressure p” due to a state of disorder. . The important result follows that p” = - ZN(dW/dV)&(l - &) . . (6.6) or since n’‘ the number of pairs of atoms in adjacent a- and /?-sites is n“ = ZNQ(1 - Q ) . . (6.7) and U” = n”W (from 6.4) p’’ = - %”(dW/d‘CT) . . (6.8) If W depends on volume according to an inverse-power law so that it can be written in the form w = W&/V). the simple formula for p” is obtained $3“ = +- x U N p (x = 4 corresponds with an inverse twelfth power of the repulsive potential; this gives reasonable values for argon).At small volumes 77’’ vanishes owing to the state of high order. At large volumes U” tends to QNZW but W diminishes at large volumes so that 77’’ vanishes at large volumes also. Thus p” passes through a maximum (cf. Fig. 5 ) . This new term p” in the equation of state of the disordered condensed assembly has a profound effect on the shape of the isotherm. UBBELOHDE MELTING AND CRYSTAL STRUCTURE 379 The corresponding free energy as a function of volume is shown in Curve I1 shows the actual isotherm of the solid (argon) at the temperature at which it melts at zero pressure. The comparatively ordered phase corresponding with A coexists with the expanded disordered phase C.The original papers must be con- sulted for further details and refine- Fig. 6. ~ l \ _ . I \ FIG. 5 FIG. 6 The pressure as a function of volume for The a given temperature ; the lower curve curves I I I and III correspond to A'/NkT gives p' the pressure for a state of A/NkT and A"/NkT .respectively. The order the upper one p the sum of p' points A B C correspond to the same and p". points as in Fig. 2. ments of the theory but the following numerical results may be noted. The melting temperature at zero pressure is taken as an empirical para- meter to determine the value of Wo in the above equations. Results which are directly calculated from this value of Wo are The free energy as a function of volume. Argon A V,/ V on melting at zero pressure .. ASf at zero pressure (83.8' K.) . . . Pressure of melting at 90.3" K. (mega- dynes/cm.a) . . . . . . . . Coefficient of expansion of liquid in immediate neighbourhood of the m.p. Calc. (two methods). 13.5 and 12.Syo 1.70k and 1.74k 286 and 294 0.0040 and 0.0049 0 bserved . 12% 1.66k 291 0.0045 Other semi-empirical relationships follow from a correlation between Wo and #o where #o is the potential energy of a pair of particles at the 380 QUARTERLY REVIEWS equilibrium distance ro which can be calculated from the known force fields. The temperature of melting at zero pressure is given by Tf = ,8#o/k and from argon the empirical parameter Other simple crystals may be compared on this basis (see table). is calculated to be 0.7. Neon . . . N . . co . . . CH . . . H . . .0,. . . . I 35.3 96 96.4 142-4 30.7 122.5 ! Tf,calc. O K. I TI obs. I<. I- . _ ~ _ _ 24.7 67.2 67.5 99.5 21.5 85.8 24-4 63-2 73 89 14 46 I Except for hydrogen where quantum effects begin to be important and for oxygen where the magnetism probably affects the force fields the correlation between Tj and #o is very promising. The theory predicts that the coefficient of expansion of the liquid just above the melting point should be a = 0*48k/$, which appears to hold fairly well for argon. The Lindemann correlation between the character- istic temperature $ for lattice vibrations and the melting temperature Ts is shown to follow from the fact that the same force field controls the atomic vibrations and the energy required to form lattice defects. An attempt has been made to extend the theory to include the melting of crystals of molecules with non-spherically symmetrical force fields,l by introducing entropy terms corresponding with hindered rotators in the crystal.This leads to a semi-quantitative treatment of orientation entropy in the solid and of the changes in orientation entropy on melting. S-W Experimental methods of investigating melting include (1) a study of entropy and volume changes a t the melting point ; (2) studies of the thermodynamic properties of solid and liquid phases in the immediate neighbourhood of the melting point such as the specific heat and the coefficient of expansion; and (3) investigation of other properties related to structure including the modulus of elasticity and the viscosity the scattering of light by the solid and liquid phases the electrical resistance (for ionic solids and for metals) the dielectric constant and dielectric losses and the diamagnetic suscepti- bility; and X-ray studies on solid and liquid.The results lead to a convenient experimental classification of factors determining the transition from solid to liquid including (i) the decrease in positional order which is the dominant effect for the simplest crystals made up from inert-gas atoms or spherically symmetrical molecules ; (ii) the decrease in orientational order which accompanies the melting of crystals of non-spherically symmetrical molecules ; (iii) for crystals in which there are strong linkages between molecules, UBBELOHDE MELTMG AND CRYSTAL STRUCTURE 381 as in quartz various glasses or graphite the breaking of these linkages which is of major significance for liquefaction ; and (iv) for molecules of very abnormal shape special melting phenomena.In polymers such as " Polythene " there is very marked premelting. In rubber-like substances a transition from the amorphous to the crystalline state can be induced by cooling or by stretching. For crystals containing rod-like molecules e.g. with several aromatic nuclei mesomorphic states intervene between the crystal and the fully amorphous liquid. Recent molecular theories of melting are amenable to detailed quanti- tative treatment only in the case of the simplest and most general change of positional order.

ISSN:0009-2681    

DOI:10.1039/QR9500400356

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

THE "RA!PION OF HETEROCYCLIC NITROGEN COMPOUNDS By K. SCHOFIELD B.Sc. PH.D. F.R.I.C. (LECTURER IN CHEMISTRY UNIVERSITY COLLEGE OF THE SOUTH-WEST EXETER) I. Introduction OUR knowledge of aromatic nitration is not entirely unsatisfactory as shown by the recent comprehensive survey by Gillespie and Millen.1 Much quantitative work has been done not only in determining the precise pro- portions of isomers formed when a given compound is nitrated but also in establishing the intimate mechanism of substitution. Unfortunately the same cannot be said of the problem of substitution particularly nitration in heterocyclic compounds. Not only are quantitative studies of orientation lacking but the effect of changes in reaction medium of great importance with compounds which are almost always basic in nature has rarely been evaluated.Many compounds are to some extent protonised by sulphuric acid,l but this becomes a major consideration with basic substances. Fortunately much of our knowledge of the mechanism of nitration of aromatic compounds can be carried over unchanged to heterocyclic sub- stances but the gaps indicated above place the problem of substitution in heterocyclic compounds in much the same position as was that in the aromatic series before the work of Holleman. It is the purpose of the present Review to collate the known facts largely qualitative about the nitration of heterocyclic nitrogen compounds providing a sort of extended footnote on this topic t o Gillespie and Millen's article. II. Five-membered Monocyclic Compounds The monocyclic nitrogen compounds present interesting contrasts be- tween pyrrole on the one hand and pyridine on the other with pyrazole and glyoxaline occupying intermediate positions.As is well known pyrrole (I) is very feebly basic (Table I) this being attributed to the need for the nitrogen atom to contribute electrons to the aromatic structure thereby becoming incapable of salt formation. As a result strong acids by destroying their aromatic character cause pyrrole derivatives to polymerise a fact which renders satisfactory nitration difficult. However the peculiar struc- ture of pyrrole besides making it sensitive to acids renders it very susceptible to electrophilic substitution. Dewar points out that the transition complex (11) involved in such substitution contains the same number of double bonds as pyrrole itself' so the activation energy is small.Further discussion leads to the conclusion that substitution will occur preferentially at the a-carbon atom. 1 Quart. Reviews 1948 2 277. 2 " The Electronic Theory of Organic Chemistry " Oxford Univ. Press 1949. This monograph will frequently be referred to because it is the only one on the subject which contains a reasonably extensive discussion of heterocyclic compounds. Some of Dewar's views are also to be found in Research 1950 3 154. 382 SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 383 The results of nitration experiments in the pyrrole series fit into this description satisfactorily so far as they go. Early workers had little success with pyrrole itself for the reasons given above but Rinkes,3 by treating pyrrole in acetic anhydride at - 10" with a small excess of nitric acid isolated 21 % of 2-nitropyrrole.Similarly he converted 2-acetylpyrrole into a mixture of its 4- and 5-nitro-derivatives and the same isomers were obtained from methyl pyrrole-2-carboxylate and pyrrole-2-carboxylic acid. Rinkes pointed out that pyrrole derivatives containing a powerful meta- directing group (NO,) behave like similar thiophen compounds in giving the 4-isomer as the principal nitration product whilst with a less powerful group (C0,Me) the 5-isomer predominates. Clearly when one of the TABLE I Values of s m e heterocyclic nitrogen compounds Compound. Pyrrole . . Pyrazole . . Glyoxaline . Indazole . . Benziminazole Pyridine . . Aniline . . pK,. 1; Compound. 0.4 Cinnolins. . 2.53 I' Quinazoline . 7.03 1' Quinoline.. 1.3 11 isoQuinoline . 5.23 // 4.5746 ' ~ ' I 5.53 / / PXa. /I Compound. 4-Hydroxyquinoline . 4-Hydroxyquinazoline 4-Hydroxycinnoline . 11 2.41 2.07 1.77 a-positions is already blocked there is little to choose in ease of substitution between the a'- and p'-positions. This is not surprising even in pyrrole itself a consideration of the two transition states (11) and (111) suggests that the energy difference between them is small. The ease with which the nitro-group replaces other groups already present in pyrrole derivatives is striking. Such replacements were observed by early worker^,^ who used more severe conditions than those described Rec. Trav. chim. 1934 53 1167 ; 1941 60 650 ; s0s also Anderlini Ber. 1889 22 2503. (a) Albert Goldacre and Phillips J . 1948 2240 ; ( 6 ) Keneford Morley Simpson and Wright J.1949 1356 ; (c) Hall and Sprinkle J . Amer. Chem. SOC. 1932 54 3469. Ciamician and Silber Ber. 1885 18 1456 ; 1886 19 1079. 384 QUARTERLY REVIEWS by Rinkes and this author has also noticed example^.^ For instance some 2-nitropyrrole was formed during the nitration of pyrrole-2-carboxylic acid and hot nitric acid converted 4-nitropyrrole-2-carboxylic acid into 2 4- dinitropyrrole. The acetyl group in 2-acetylpyrrole suffers similar displace- ment. In this respect pyrrole derivatives resemble benzene compounds activated towards electrophilic reagents by the presence of several alkyl or alkoxy-groups. Pyrrole shows acidic properties the anion being stabilised by resonance with respect to pyrrole itself. This anion should be even more susceptible to electrophilic attack than is the parent compound and it is of interest that in ethereal solution in the presence of sodium pyrrole reacts with ethyl nitrate to give a small amount of 2-nitr0pyrrole.~ Although the mechanism of such a nitration is not certain it seems likely that the reaction is an example of electrophilic substitution into the pyrrole anion.Glyoxaline (IV) is especially interesting. Probably because of the symmetry of the cation (V) it is a considerably stronger base than pyridine (Table I). By comparison with pyrrole glyoxaline is itself more sym- metrical and whilst still activated to some extent because of the NH group is less susceptible than the former to electrophilic attack. Consideration of the transition states (VI) and (VII) led Dewar to conclude that in glyoxaline electrophilic substitution in neutral and alkaline media should proceed at C(2, in support of which he quoted the instances of bromination and of coupling with diazonium salts.Further since in the transition states the salts of glyoxaline would acquire positive charges on both nitrogen atoms electrophilic substitution in these salts should be more difficult than in glyoxaline itself. The facts available on the bromination of glyoxaline do not warrant the inclusion of this reaction as an example of electrophilic substitution at C(2,. The evidence relating to electrophilic substitution in glyoxaline is of great interest in connection with our present discussion of nitration and will be briefly outlined. On bromination glyoxaline and its derivatives readily give di- or tri- bromo-compounds but in chloroform solution a monobromo-derivative can be produced from 4-methylglyoxaline and it has been proved to be 5-bromo-4-methylglyoxaline.8 1 4- and 1 5-Dimethylglyoxaline likewise give 5-bromo-1 4- and 4-bromo-1 5- dimethylglyoxaline but Langenbeck @ obtained 2-bromo-1 4-dimethyl- glyoxaline by treating 1 4-dimethylglyoxaline with cyanogen bromide in ether.Pauly and Arauner 10 isolated small proportions of 2-iodo-derivatives from the iodination of glyoxaline and 4-methylglyoxaline and similarly glyoxaline couples with diazonium salts in alkaline media at C(21:11 In contrast to this somewhat confused position in halogenation experiments the nitration of glyoxaline and its derivatives gives consistently the 4- or * (Miss) Nightingale Chem. Reviews 1947 40 117.Angeli and Alessandri Atti R. AcccLd. Lincei 1911 20 I 311 ; Hale and Hoyt J . Amer. Chem. SOC. 1915 37 2538. 8Pyman J. 1910 97 1814; Pyman and Timmis J. 1923 123 494. J. pr. Chem. 1928 119 77. l1 Pyman and Timmie J. SOC. Dyers Col. 1922 38 269. lo Ibid. 1928 118 33. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 385 5-nitro-derivatives. Glyoxaline and 4-methylglyoxaline l Z a give the &nitro- compounds. Purthermore nitration of glyoxalines is only possible if either the 4- or the 5-position is free Fargher and Pyman 12a having found that they could not nitrate 4 5-dimethylglyoxaline. The situation is thus seen to be complicated and it has occasionally been suggested that the substitution at C(2) which occurs during iodination may proceed by other than a normal electrophilic reaction.13 Whether this is so or not it seems probable that with a strong base such as glyoxaline the experimental conditions will affect the orientation of substitution.In alkaline or neutral conditions glyoxaline itself would be the entity under- going substitution but in nitration (and sulphonation 12*) the directive effect observed may be that of the glyoxalinium cation. The case is quali- tatively similar to the nitration of aniline,l* but quantitatively distinguished since aniline a weaker base (Table I) provides some p-nitroaniline along with the m-compound even when nitration is conducted in a very large excess of sulphuric acid. The actual position of nitration in the glyoxalinium cation is not easy to explain in qualitative terms. Enumeration of the various forms which might be expected to contribute to the transition state (Inn.) (=.I (X3 leaves little to choose between the 2- and 4(5)-positions.Perhaps the nitronium ion approaches the glyoxalinium cation at the point most remote from the positively charged heterocyclic nitrogen atoms. The nitration of glyoxaline 12@ was effected by hot mixed acids acting for several hours and although electrophilic substitution in these circumstances is clearly more difficult than in the case of pyrrole yet it is much less difficult than with pyridine (see below) as indicated by the yield of 4-nitroglyoxaline obtained (63%). Pyrazole (VIII) is a much weaker base than is glyoxaline (Table I) and even if it were not so it is doubtful whether electrophilic substitution would proceed differently in the strongly acid conditions used in nitration as compared with the circumstances in say chlorination for the expected point of attack is in a peculiarly symmetrical situation with respect to the two ring nitrogen atoms which distinguishes this case from that of glyoxaline.Thus consideration of the transition states (IX) and (X) leads us to expect electrophilic substitution at Ct4). This is the case not only in nitration l5 but also in chlorination l6 and bromination.15~ 17 The conditions used in l2 (a) Rung and Behrend Annalen 1892 271 28; Behrend and Schmitz ibid. 1893 277 338 ; Fargher and Pyman J. 1919 115 217 ; Fargher J. 1920 117 668 ; (b) Barnes and Pyman J. 1927 2711. l3 Brunings J . Amer. Chem. SOC. 1947 69 205. l4 Holleman “ Die direkte Einfuhrung von Substituenten in den Benzolkern ” Leipzig 1910.Buchner and Fritsch Annalen 1893 273 262. l6 Knorr Ber. 1895 28 715. l7 Buchner ibid. 1889 22 2166. 386 QUARTERLY REVLEWS this nitration do not permit us to judge of the relative ease of nitration of pyrazole and glyoxaline but it seems clear that the reaction proceeds more readily than with pyridine. III. Pyidine and its Derivatives The nitrogen atom in pyridine differs from the NH group of the com- pounds discussed so far in that the latter is an activating factor whilst the former lowers the availability of electrons at all points in the ring,l* rendering pyridine less susceptible to electrophilic attack than is benzene. The situation is therefore similar to that in nitrobenzene. Consideration of the three transition states (XI) (XII) and (XIII) leads to the con- clusion 2 that (XII) would be the most stable since in (XI) and (XIII) " the nitrogen atom occupies a more highly (positively) charged position in the mesomeric cation ".Thus we should expect electrophilic substitution in pyridine to occur at C(3, with more difficulty than in benzene and in the pyridinium cation with even more difficulty but at the same position. In the last respect pyridine resembles pyrazole ; Le. in both cases protonisation is not expected to change the site of nitration. . ,' - + (XIS (xn.) ( m.1 Pyridine has proved extremely diacult to nitrate. Friedl,19 by distilling pyridine in a current of air with fuming sulphuric acid and potassium nitrate obtained 15% of 3-nitropyridine and later prepared the same derivative by treating pyridine with sulphuric acid and potassium nitrate at 330".20 Kirpal and Reiter,21 in repeating this method found traces of iron to be essential to success and in the light of this finding raised the yield to 22%.However den Eertog and Overhoff,22 in the most careful examination so far described were unable to realise the yield claimed by Kxpal and Reiter but noticed an important new point. They nitrated pyridine in lOOyo sulphuric acid with a mixture of potassium and sodium nitrates at tem- peratures varying between 300" and 450". Fifty % of the pyridine was recovered and 6% of nitropyridines isolated. This product proved to be a mixture of 2- and 3-nitropyridine the proportions of which varied with the temperature being at 300" 0.5% and 4.5% ; at 3?0° 2% and 4% ; at 450° 2.5% and O% respectively.Similar variations though with higher yields are observed in the halogenation of pyridine at different temperature^,^^ and in both the nitration and the halogenation experiments it is likely that the mechanism of substitution changes from the electrophilic 1* Longuet-Higgins and Coulson Trans. Paraday Soc. 1947 43 87. l9 Ber. 1912 45 428. 2o Chern.-Ztg. 1913 36 589 ; Monatsh. 1913 34 760. 21 Ber. 1925 58 699. aa Rec. Trav. chim. 1930 49 552. 23 The results of several investigations are summarised by Wibaut Experientia 1949 5 337. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 387 to the free-radical type as the temperature rises. This being so it is not possible to attach such significance to the formation of 2-nitropyridine as is attributable to the production of a proportion of o-substitution in benzene derivatives containing meta-directing groups.Dinitrogen tetroxide nitrates pyridine in the vapour phase to give low yields of 3-nitropyridine but the mechanism is uncertain.24 The nitration of alkylpyridines described by Plazek 25 shows no new features but as would be expected the introduction into pyridine of a hydroxyl group greatly facilitates nitration the hydroxyl group controlling the orientation. For instance mixed acids convert 2-hydroxypyridine into the 3-nitro- 3 5-dinitro- and 5-nitro-derivativesY in that order of pre- dominance,26 whilst 3-hydroxypyridine provides 3-hydro~y-Z-nitropyridine~~~ and 4-hydroxypyridine gives 4-hydroxy-3-nitropyridine.28 The first two cases are worthy of comment. The predominant formation of Z-hydroxy- 3-nitropyridine from 2- hydroxypyridine may be an illustration of chelation in the transition complex between the entering nitro-group and the hydroxyl group affecting the o :p-ratio as suggested by Waters,Z9 whilst the case of 3-hydroxypyridine might suggest that the ring nitrogen atom (perhaps protonised) can also contribute to such an effect.It is interesting that 3-ethoxypyridine is converted by mixed acids to 3-ethoxy-2-nitropyridine in 75-8070 ~ield.~O These examples in the pyridine series make interesting comparisons with similar benzene compounds. For example m-nitrophenol on nitration gives mainly 3 4-dinitrophenol together with smaller amounts of 2 3- and 2 5-dinitrophenol,l4 whilst m-nitroanisole gives 2 3- (51.2y0) 2 5- (40.6%) and 3 4-dinitroanisole (8.270).14 The effect o f the ring nitrogen atom in pyridine clearly differs considerably in these cases from the effect of the nitro-group with which it is so frequently compared.Protonisation and some form of chelation may be responsible for this and the steric factors must be very different in the two cases. As is well known the properties of arnino-groups in heterocyclic com- pounds vary considerably with their positions in the molecule. 2- and 4-Aminopyridine are considerably removed from the category of primary aromatic amines into which 3-aminopyridine falls. It is therefore not surprising that in their behaviour towards nitric acid these two amines represent the extreme examples of the change in character noticeable in moving from anilines containing electron-releasing groups to those contain- ing say several nitro-groups.2-Aminopyridine is readily converted into 2-pyridylnitramineY31 which rearranges in the presence of sulphurie acid 24 Shorigin and Topchiev Ber. 1936 69 1874. 26 Binz and Maier-Bode Angew. Chem. 1936 49 486. Plazek and Rodewald Rocz. Chem. 1936 16 502. 28 Bremer Annalen 1937 529 290. 30 den Hertog Jouwersma van der Waal and Willebrands-Schogt Rec. Trav. chim. 1949 68 275. 31 (a) Tschitschibabin and Rasorenow J. Russ. Phys. Chem. SOC. 1915 47 1286 ; (b) Tschitsehibabin and Builinkin ibid. 1920 50 47 ; Phillips J. 1941 9 ; Caldwell and Kornfeld J . Amer. Chem. Soc. 1942 64 1696 ; (c) Plazek and Sueharda Ber. 1928 61 1813. 25 Ibid. 1939 72 577. 29 J. 1948 727. 388 QUARTERLY REVIEWS producing 2-amino-3- and -li-nitropyridine the latter being the major product.The mechanism of the rearrangement has not been examined and the evidence in the case of analogous phenylnitramines is not clear cut.32 One difference to be noted is that although a surprisingly large amount of o-nitroaniline is formed in the benzene series the 5-nitro-com- pound is the main product from 2-pyridylnitramineY although higher tem- peratures in the rearrangement are said to favour the formation of Z-amino- 3-nitropyridine A halogen or nitro-group at C(5) does not prevent the formation and rearrangement of nitramines from substituted Z-amino- p y r i d i n e ~ ~ ~ ~ ? 33 and 4-aminopyridine behaves similarly.34 It is a curious fact that 2-acetamidopyridine resists nitration.31c 3-Aminopyridine also forms the related nitramine when it is treated with nitric acid in sulphuric acid solution but unlike the isomeric com- pounds already mentioned 3-nitraminopyridine cannot be isomerised to a 3-amino-nitropyridine.35a This stability has not been explained. In con- trast the nitramine from 3-methylaminopyridine rearranges to 3-methyl- amino-2-nitropyridine. 35* In 2-dimethylaminopyridine where nitramine formation is impossible nitration occurs mainly para (90%) to the dimethylamino-group together with some ortho (10%) substitution ; 36i.e. the dimethylamino-group behaves as an op-directing group rendering nitration easier by comparison with pyridine itself. This is not surprising when it is remembered that proto- nisation probably occurs on the nuclear nitrogen atom rather than on the NMe group,4a and the result is in contrast to that found with dimethylaniline which when nitrated in concentrated sulphuric acid gives mainly m-nitro- dimethylaniline,14 The results with 2-dimethylaminopyridine suggest that direct nitration of 2-aminopyridine in the work already discussed is at least a possibility.IV. Bicyclic Compounds (a) The Work of Fries.-A review of the known facts of the nitration of heterocyclic compounds would not be complete without reference to the long sequence of publications by Fries and his co-~orkers.~~ He set out to study the properties of numerous heterocyclic systems of which we shall mention only those containing nitrogen in an attempt to decide whether they should be regarded as resembling benzene or naphthalene the more closely. The tests chosen by Fries to distinguish benzenoid from naphth- alenoid compounds include some typical electrophilic reactions such as nitration and the work has revealed an interesting variety of behaviour 82 Bradfield and Orton J.1929 915. 33 Tschitschibabin and Tjashelowa J. Rzcss. Phys. Chem. SOC. 1920 50 483; 3* Koenigs Kinne and Wsiss ibid. 1924 57 1172 ; Koenigs Mields and Gurlt 35 (a) Tschitschibabin and Kirssanow Ber. 1927 60 2433 ; (b) Plazek Marcini- 86 Tschitschibabin and Knunyantz Ber. 1929 62 3053. 37 (a) AnnaEen 1912 389 305 ; ( b ) $bid. 1914 404 50 ; (c) ibid. 1927 454 121 ; Magidson and Menchikov Ber. 1925 58 113. ibid. p. 1179; Rath and Prange Annalen 1928 467 1. kow and Stammer Rocx. Chem. 1935 15 365. (d) ibid. 1934 511 213 ; (e) ibid. 1937 527 38 ; (f) ibid. 1941 550 31. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 389 Compound.Quinoline . . . . . . . Azimidobenzene Me #-Azimidobenzene . . . . Benziminazole . . . . . 2 3-Dimethylindole . . . Indazole . . . . . . . . . . . . . -~ . TABLE I1 Position of nitration. Conditions. Mixed acids ? 9 ,7 9 1 Fuming HNO * N. = Naphthalenoid ; B. = benzenoid. TABLE I11 Compound. 5-Nitroquinoline . . . . . . . 6-Nitroquinoline . . . . . . . 7 . Ni troquinoline . . . . . . . 8-Nitroquinoline . . . . . . . H ______- Position of nitration. 7 5 and 8 5 and 8 6 6 5 Type." Ref. 38 3 7d 37d 3 7d 40 41 42 39 Ref. 43 44 44 45 44 45 45 46 3 7c 37f 38 J . Amer. Chem. SOC. 1940 62 1640 ; J. 1947 1613. sB von Auwers and Kleiner J. pr. Chem. 1928 118 67. 40 Bamberger and BerlB Annalen 1893 273 340 ; Fischer and Hess Ber. 1903 4 2 Plant and (Miss) Tomlinson J.1933 955. 43 Claus and Hartmann J. pr. Chem. 1896 53 199. 4 4 Kaufmann and Hussy Ber. 1908 41 1735. 4 6 Kaufmann and Decker ibid. 1906 39 3648. 46 Kym and Ratner ibid. 1912 45 3248. 36 3967. 41 Bauer and Strauss ibid. 1932 65 308. DD 390 QUARTERLY REVIEWS in heterocyclic compounds. The facts discovered or quoted concerning nitrations fall into three groups the position of mononitration of bicyclic compounds the orientation of the second nitro-group in the dinitration of similar systems and the behaviour of chloro- or bromo-hydroxy-derivatives of these heterocyclic systems on nitration. TABLE IV Compound. Product. H O W / N HO Br H Conditions. HNOa-CHCl 9 ) HNO HNOa-AcOH 97 Type. N. 9 ) 79 B. Intermediate B. Ref. 37a 7 ) t 9 37c 37b 37b The positions of mononitration of some of the nitrogen compounds examined by Fries are indicated in Table 11.A distinction is made between those which react at ‘‘ a ” positions and those initially nitrated at “/I ” carbon atoms. The former are regarded as showing naphthalenoid character the latter as being benzenoid. Table I11 summarises the facts relating to the dinitration of some nitrogen compounds. It is striking that in the quinoline series the second nitro-group enters mainly meta to the first whilst in several of the bicyclic compounds with a five-membered ring fused to a benzene ring the second nitro-group is substituted adjacently to the first. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 391 The third type of nitration reaction examined by Fries is illustrated in Table IV.Some heterocyclic compounds containing adjacent hydroxyl and halogen groups behave like similar halogeno-naphthols in forming with nitric acid chloronitro-ketones whilst others react like ordinary members of the benzene series and give normal nitro-compounds. Once again on the basis of this reaction systems are classified as naphthalenoid or benzenoid. Clearly this work has produced a mass of interesting facts which future theory must accommodate but for several reasons its present value as a contribution to comparative heterocyclic chemistry is limited. The classi- fication into naphthalenoid or benzenoid types is vague and often it is not clear whether the nature of the parent heterocyclic compound or one of its derivatives is indicated. All of the evidence is qualitative and the choice of criteria for classifying compounds arbitrary.The latter point is clearly seen in the different behaviour of say 7-chloro- and 7-bromo-6-hydroxy- indazole on nitration ; according to the particular case chosen the nature of the indazole system would require different description. The fact that diEerent conditions in nitration experiments may lead to diEerent results is not taken into account and the mechanisms of the various reactions are uncertain. ( b ) IndoEe and its Derivatives.-The indole series demands further mention because of the variety of its reactions with nitric acid under different con- ditions. Indole itself like pyrrole is very susceptible to electrophilic attack but consideration of the transition states leads us to expect substitution in this case to proceed initially a t C,, Ct2 coming next in reactivity.Indole itself and its derivatives unsubstituted at C!z and C,, are so reactive as to render difficult the observation of monolutration. Although 3-nitroindole cannot be obtained by the action of nitric acid upon indole it is formed when the latter is treated with ethyl nitrate and sodium ethoxide in ether.47 2-Methylindole with the same reagents provides 2-methyl-3- nitr~indole,~~ although an apparently Werent mononitroindole is obtained with mixed acids.48 Warm nitric acid converts 2-methylindole into a dinitro-derivative in which one o f the nitro-groups is probably at C(3).48 49 The work of Perkin and of Plant has shown that with indoles of the type (XIV) nitric acid reacts differently according to the nature of the substituents.Nitration with mixed acids of tetrahydrocarbazole (XIV ; R = H R’R”- - <[CH,]4),50a 1 2 3 4-tetrahydro-3-methylcarbazole 47 Angelic0 and Tielardi Atti R. Accad. Lime; 1904 13 I 242 ; Gazzetta 1904 48 von Walther and Clemen J. p. Chem. 1900 61 268. 4@ Mathur and Robinson J. 1934 1415 ; see also Zatti Gazzetta 1899 19 260. so (a) Perkin and Plant J. 1921,119 1825 ; (b) idem J. 1923 123 676 ; (c) idem ibid. p. 3242; (d) Manjunath and Plant J. 1926 2260; (e) Plant and Rosser J. 1928 2454; (f) Plant and Rutherford J. 1929 1970; (9) Plant ibid. p. 2493; (h) Fsnnsll and Plant J. 1932,2872 ; (i) Massey and Plant J. 1931,1990 ; ( j ) Plant J. 1936,899 ; (k) Moggridge and Plant J. 1937,1125 ; ( I ) Plant and (Miss) Wilson J. 1939 237 ; (m) Plant and Whitaker J. 1940 283 ; (7i) Gaudion Hook and Plant J.1947 1631 ; ( 0 ) Bannister and Plmt J. 1948 1247. 84 60. 392 QUARTERLY REVIEWS (XIV ; R = H R’R” = -[CH2]2*CHMe*CH2-),5* and of 2 3-dimethylindole (XIV ; R = H R’ = R” = Me) 41 42 gives the corresponding derivatives of type (XV). Only when the indicated position is already occupied does nitration proceed elsewhere. 50e~k 1 2 3 4-Tetrahydro-N-methylcarbazole likewise undergoes nitration at C(o).50a m; N*2fJ--Jg f+JJ; 02 R’ R n R R ( ¶.) ( XY.) (xnr.) ( ¶.) (xsm.f (=.) ( xx.) ( =.) In contrast compounds in which the indole nitrogen atom is acylated show a surprising degree of variety. Almost all the nitrations of these acyl derivatives have been affected in acetic acid-nitric acid and the results are outlined in Table V. TABLE V Products formed from nitric acid and substituted indole derivatives i 1 2 3 4-Tetra- 1 hydrocarbazole.* N-Acyl derivative.Acetyl . . Cinnamoyl . . Carbethoxy . . Benzyl . . . Benzovl . . . p-Toluoyl p-Chlor obenzo y I} o- and m-Toluovll benzo yl Heterocyclic series. 2 3-Dimethyl- indole. XVI + XVII (42 50m) - - - XIX (42) - 2 3-Diphenylindole. XVI (5072) XVI + complex products (50h) XVI (50h) XVI (5072) Dihydro- pentindole. XVI + XIX (50c,g,i,m) XVI + XIX XVI + XIX (50g,i) XVI + XIX (509) (50;) - - Broadly speaking N-acetyl N-cinnamoyl and N-carbethoxy-derivatives of all the series examined except the dihydropentindole one give derivatives of the type (XVI)* and (XVII) whilst the corresponding dihydropentindoles give products of the types (XVI)* and (XIX). The proportions of the two products formed in any case may vary with the relative amounts of acetic and nitric acids and with the temperature.Minor aberrations from the above general scheme have been observed ; e.g. 3-methyl- and N-carbethoxy- * A 6-nitro-group should have been shown in (XVI). SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 393 6-chloro-1 2 3 4-tetrahydrocarbazole give unexpectedly compounds of the types (XVI)* and (XIX).5&9k Further 5-chloro-8-cinnamoyldihydro- pentindole provides as well as the expected (XVI)* and (XIX) a small proportion of (XX).50{ Also noteworthy is the fact that N-acetyl derivatives of 2 3-dimethylindoles containing electronegative substituents give products of type (XIX).5m9n Little discussion or explanation of the large number of observations of this kind has yet been attempted.The fact that the unacylated indoles are nitrated to give compounds o f the type (XV) whilst acylated systems give rise to compounds (XVI),* was a t first considered to be anomalous but it was suggested that the nitrogen atom exerts its influence through the very reactive indole double bond in the sense of the expression (XXI),psl 50@ from which point of view the NH group becomes meta- and the N-acyl groups para-directing as was expected. Such an explanation whilst it stresses the interesting character of the indole double bond is not entirely adequate for present-day theories which consider the molecule as a whole rather than isolating particular elements in it. Whatever the true signific- ance of these results certain points may be stressed. First in no case has a reaction of the present type been studied quantitatively.Such a study would present great difliculties and it is perhaps significant that in many of the examples mentioned the yields obtained are not quoted. Secondly little is known of the mechanism whereby solutions of nitric acid in acetic acid are able to effect additions to double bonds and whether the variety of addition products noted arises from different modes of direct addition of the medium or from subsequent changes in one primary addition product is not clear. The indole double bond is of course not unique in its ability to add on the elements of nitric acid. Several examples are known of such additions to ethylenic linkages and were responsible for the mistaken notion that aromatic nitration proceeds by addition of nitric acid followed by the elimination of water.51 ( c ) Quinoline and Related Compounds.-The series-quinoline isoquino- line quinazoline and cinnoline (no data are available for quinoxaline and phtha1azine)-presents a more homogeneous group for discussion than the bicyclic compounds already mentioned.A very large number of quinoline derivatives has been nitrated from time to time and we shall not attempt to deal with these exhaustively since many are of no special significance. The nitration of quinoline in sulphuric acid has been described by several workers most recently by Fieser and Hershberg and by Curd Graham (Miss) Richardson and Rose.38 Although the conditions have varied from case to case it has always been found that 5- and 8-nitroquinoline are formed in roughly equal proportions.Dufton 52 obtained a by-product which was later shown to be 5-hydroxy-6 8-dinitroquinoline. 53 Bacharach Haut and Car0line,~4 by nitrating quinoline with lithium nitrate and a 51 Gilman " Organic Chemistry An Advanced Treatise " 2nd sdn. Vol. 11 p. 175 52 J. 1892 61 782. 63 Bennett and Grove J. 1945 378. 64 Rec. Trav. chim. 1933 52 413. * A 6-nitro-group should have been shown in (XVI). New Pork 1943. 394 QUARTERLY REVIEWS TABLE VI Nitration of quinoline derivatives and related compounds a s1 Compound. Quinoline . . . . . . isoQuinoline . . . . Quinazoline . . . . Quinaldine . . . . . Lepidine . . . . . 4-Methylcinnoline . . . 2-Chloroquinoline . . . 2-Bromoquinoline . . . 3-Chloroquinoline . . . 3-Bromoquinoline . . . 4-Chloroquinoline . . . 4-Chlorocinnoline .. . 2 4-Dirnethylquinoline . 2-Chlorolepidine . . . 4-Chloroquinaldine . . 4-Chloro-3-methylquinoline 2 4-Dimethylquinazoline . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . Carbostyril . . . . . . . 2-Alkoxyquinolines . . . . N-Alkylcarbostyrils . . . . 4-Hydroxyquinoline . . . . 4-Hydroxyquinazoline . . . 4-Hydroxycinnoline . . . . 1-Methyl-4-cinnolone . . . . 4-Hydroxyquinaldine. . . . 2-Hydroxylepidine . . . . 4-Hydroxy-3-methylquinoline . 2 4-Dihydroxyquinoline . . 7-Chloro-4-hydroxyquinoline . 4-Hydroxyquinaldine N-oxide . 4-Hydroxy -2-methylquinazoline 2-Aminoquinoline . . . . . 4-Aminoquinoline . . . . . 4 6-Diaminoquinoline . . . Product. b 5 8 7 5 8 6 5 8 8 5 (?) 8 8 5 5 8 5 8 (?) 5 8 8 5 8 8 6 5 (?) 8 6 5 (?) 8 5 6 5 8 4-Hydroxy-%methyl- 6-nitroquinazoline 6 6 6 6 8 3 6 6 8 6 3 (?) 8 6 3 6 6 8 3 3 6 66 68 3 3 (?I Conditions.@ M L a o 3 / Cu(NO3)2/ Ac,O M M M M M M M M M M M M M M M M M M N M N M M N M M N M M N N/AcOH M M M (&) _______ Ref.38 54 24 55 56 57 58 56b 59 60 61a,b 62a,b 63 62b 64 65 66 67 68 69 70 71 72 73 60 74 61b 563,75 56b 76 77 77 56b 56b 78 69 70 79 80 71 81 82 83 76 84 84 85 86a a Only the initial mononitration product is indicated. c M = " mixed acids " ; N = nitric acid. dThe N-oxide when treated in hot acetic acid solution with nitric acid gave 4-hydroxy-3-nitroquinaldine. The N-oxide of the latter resulted when nitric acid was added t o a dilute sulphuric acid solution of the parent compound. e Through the nitramine. f Cinnoline has given two nitro-compounds differing from 6-nitrocinnoline.is identical with 8-nitrocinnoline (Morley and Alford and Schofield unpublished). As far as possible the major product is named fist. One 55 Claus and Hoffmann J. p. Chem. 1893 47 253 ; Fortner Monatsh. 1893 14 146 ; Le FBvre and Le FBvre J. 1935 1470 ; Andersag Chem. Zentr. 1934 I 3595. 56 (a) Elderfield Williamson Gensler and Kremer J. Org. Chem. 1947 12 405 ; ( 6 ) Schofield and Swain J. 1949 1367. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 395 trace of cupric nitrate in acetic anhydride obtained 7-nitroquinoline and the same product together with some 3 7-dinitroquinoline resulted from the reaction between liquid quinoline and dinitrogen tetroxide.2P These results and those for some quinoline derivatives and for the other heterocyclic compounds mentioned above are summarised in Table VI.It is clear that in the quinoline series (and probably in the isoquinoline group also though the evidence here is not extensive) except for the cases of hydroxy- and amino-derivatives there is a powerful tendency for nitration to occur at C(5) and at C(*) where that is possible. This is true of Bz- substituted quinolines as well as we have already seen in the case of the further nitration of 6- and 7-nitroquinoline (Table 111) some substitution occurring ortho to the nitro-group in each case. 6- and 7-Halogenoquinolines 67 Doebner and von Miller Ber. 1884 17 1698 ; Gerdeissen {bid. 1889 22 245 ; Decker and Remfrey ibid. 1905 38 2773; Harmnick J. 1926 1302. 68 Busch and Koenigs Ber. 1890,23,2687 ; Johnson and Hamilton J . Amer. Chem. Soc.1941 03 2864; Buchman et d. ibid. 1947 09 380. 6s Fischer and Guthmann J . pr. Chem. 1916 93 378 ; Deinet and Lutz J . Amer. Chem. SOC. 1946 08 1325. Bo Bennett Crofts and Hey J. 1949 277. (a) Claus and Pollitz J . pr. Chem. 1890 41 41 ; (b) Decker and Pollita ibicE. 62 (a) Edinger and Lubberger ibid. 1896 54 340 ; (b) Baker et al. J . Amer. Chem. e3 Claus et al. J . pr. Chem. 1889 39 301 ; 1893 48 157 ; 1896 53 413 ; Decker 64 Gowley et al. J . Amer. Chem. SOC. 1947 69 303 ; Simpson and Wright J. 1948 86 Keneford Morley and Simpson ibid. p. 1702. 1901 04 85. Soc. 1946 08 1532. Ber. 1905 38 1274. 1707. Price Velzen and Guthrie J . Org. Chem. 1947 12 203 ; Vaughan J . Amer. Ochiai and Shimizu J . PJtarm. SOC. Japan 1943 63 398. Chem. SOC. 1948 70 2294. 68 Johnson and Hamilton J .Amer. Chem. Soc. 1941 83 2867 ; Krahler and Burger 6sAdams and Hey J. 1949 3185. 70 Halcrow and Kermack J. 1945 415. ?IAdams and Hey J. 1950 2092. 79 Friedlander and Lazarus Annalen 1885 229 233 ; Decker and Kasatkin J . pr. 74Kaufmann and de Petherd Ber. 1917 50 336. 75Morley and Simpson J. 1948 2024; Adam and Hey J. 1949 255. 76 Bogert and Geiger J . Amer. Chem. Soo. 1912 34 524 ; Morley and Simpson 77 Schofield and Simpson J. 1945 512; Simpson J. 1947 237. ?* Kermack and (Miss) Weatherhead J. 1939 563. 79 Conrad and Limpach Ber. 1887 20 948. 8o Balaban J. 1930 2346. 81 Gabriel Ber. 1918 51 1500 ; cf. Ashley Perkin and Robinson J. 1930 382. 82 Bmslow et al. J . Amer. Chem. SOC. 1946 08 1232. 83 Gabriel and Gerhard Ber. 1921 54 1067 1613. 84 Tschitschibabin Witkowski and Lapschin ibid.1925 58 803. 85 Simpson and Wright J. 1948 2023 ; Jensch Annalem 1950 568 73. 88 (a) C.P. 613,065 (1935) ; F.P. 779,092 (1935) ; (b) Kaufmann and Zeller Ber. ibid. 1942 64 2417. Tomisek and Christensen J . Amer. Chem. SOC. 1948 70 2423. Chem. 1901 64 85. J. 1948 360. 1917 50 1630. 396 QUARTERLY REVIEWS probably behave similarly.87 The tendency to nitrate at Ct5) and C(*) clearly persists in quinoline substituted by one methyl or halogen group in the heterocyclic ring (Table VI) and it is usually supposed that the predominance of the 8-nitro-compound in the nitration product of lepidine is due to a steric effect (although the methyl group in 1-methyl-4-cinnolone appears to exert no hindrance). 4-Methylcinnoline seems to be directly comparable. In view of all these results the case of quinazoline is of out- standing interest being the only one so far encountered in this series where the parent heterocyclic compound is nitrated at The production of 4-hydroxy-2-methyl-6-nitroquinazoline from the action of mixed acids upon 2 4-dimethylquinazoline stresses the peculiarity of the quinazoline nucleus.The orientation of substituents entering heterocyclic molecules has rarely been discussed. Roberts and Turner 88 considered the problem in terms of Thiele’s theory and related the behaviour of quinoline and isoquinoline to that of naphthalene. has compared these com- pounds with 1- and 2-nitronaphthalene respectively. It is interesting to More recently Dewar ( y j Y . . . ......... bj+ I . .I.. .............. ; ...... i ........ r’ ................. ......v + examine Dewar’s views in some detail since they illustrate a difficulty always encountered by qualitative discussion of this problem. In examining further substitution into a monosubstituted naphthalene Dewar enumerates the transition states (XXII -+ XXV) and (XXVI -+ XXVIII) the sub- stituent Y already present being electron releasing. These transition states are supposed to contain elements o f structure resembling benzyl and phenyl- ally1 cations and Dewar concludes that an electron-releasing 1 -substituent should direct further electrophilic substitution mainly to and C(*) and less powerfully to C(51 and C(,). [Dewar’s analogy to the benzyl cation entirely neglects the portion of the system represented as a circled double- bond in (XXIV) and (XXVII).] A similar 2-substituent likewise activates the adjacent l-position strongly and the 6- and 8-positions weakly.With an electron-attracting first substituent the position is reversed and Dewar is led to state that in l-nitronaphthalene for example the 5- and 7-positions 87 Claus and Schedler J . pr. Chem. 1894 49 359 ; Gilman et al. J . Amer. Chem. SOC. 1946 88 1577 ; Claus and Junghanns J . pr. Chem. 1893 48 253 ; La Coste Ber. 1885 18 2940 ; Andersag Chem. Ber. 1948 81 499. 88 J. 1927 1832, SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 397 will be deactivated more strongly than the 6- and 8-positions whilst in 2-nitronaphthalene the 6- and 8-positions will be more strongly deactivated than Thus nitration of 1-nitronaphthalene should give 1 8- and 1 6-dinitronaphthalene whilst 2-nitronaphthalene should provide 1 6- (2 5)- and 2 7-dinitronaphthalene.This expectation is only partly in accordance with experiment. In fact 1-nitronaphthalene when nitrated in sulphuric acid yields 1 8- (64%) and 1 5-dinitronaphthalene (36%),s9 whilst 2-nitronaphthalene under the same conditions gives 1 3 8-trinitro- naphthalene but in glacial acetic acid it provides 1 6- and 1 7-dinitro- n~phthalene.~O Proceeding then to discuss quinoline and isoquinoline Dewar states that “both undergo substitution in the benzene ring ex- clusively and in the 5- and $-positions for the same reasons that mono- nitronaphthalenes give 5- and 8-substitution products ”. The previous argument implies rather that quinoline should give 6- and 8-nitroquinoline and that isoquinoline should be converted into 5- and 7-nitroisoquinoline.The same difficulty arises if instead of considering transition states one simply writes down the various resonance forms of the types (XXIX) and (XXX).gl Waters 29 has referred to the transition-state theory of aromatic and C(7). 0.958 0.772 0.996 0.938 0.978 0.940 0-942 02 N 1.594 0 * 984 a+ .- a- a 0 * 7 8 9 1,003 1.633 0.948 0.767 IXXTXS (Xxx.) (-1 f xxxn.) substitution in this connection. Discussing naphthalene he points out that since the transition state for 1-substitution is related to that for 2-substitution as are 1 4- and 1 2-naphthaquinone the former should be the more stable and 1-substitution should be favoured. Similar arguments apply to quin- oline and isoquinoline although the limitations of this viewpoint are severe since like those already discussed it tells us nothing of the relative ease of substitution of C(5) compared with C(s).Measurements of oxidation- reduction potentials of heterocyclic quinones related to this question are satisfactory as far as they g 0 . ~ 2 Quinazoline is interesting in the light of this suggestion and it may be that in this c a ~ e ~ ~ b as well as with 2 4- dimethylquinoline 2-chlorolepidine and 4-chloroquinaldine (see Table VI) the amphi-quinonoid form of the activated complex is of importance. This account of attempts to explain the orientation of substituents enter- ing quinoline and related molecules would be incomplete without reference to recent quantum-theory calculations which promise eventually to provide a quantitative estimate of the reactivities of the several positions in hetero- cyclic nuclei.Earlier workers concentrated on evaluating the net charges (n-electron distributions) on the nuclear carbon atoms. Depending on whether molecular-orbital or valence-bond theory is used and also on the 89Hodgson and Whitehurst J. 1945 202. 91 Schofield and Swain Nature 1948 161 690. 9* Fieeer and Martin J Amer. Chem. SOC. 1935 57 1840 Vesely and Jake; BuEE. SOC. chim. 1923 33 952. 398 QUARTERLY REVIEWS somewhat arbitrary choice of parameters slightly different results are obtained but on the whole the theories agree satisfactorily about the relative values of the charges at the nuclear positions. The earlier calculations of Longuet-Higgins and Coulson 93a from molecular-orbital theory might be used as an illustration. They give for quinoline and isoquinoline the n-electron densities shown in (XSXI) and (XXXII).Taken alone these results indicate that in quinoline and isoquinoline susceptibility to electro- philic attack would decrease in the order C(* > C(6) > C(3) > C(s, etc. and C(51 > C(,) > C(s) > C(3! etc. respectively (availability of electrons being assumed to be the controlhng factor g3b) These sequences do not wholly fit the facts of nitration (or sulphonation) and it soon became clear that since differences between charges at digerent positions are small other factors must be taken into account. Other quantities now used are the bond order the free-valency index the polarisability and the potential 0.980 0,873 0.444 0.433 2-40 2-55 0.401 0.400 0.397 2-61 2.33 1.01 3 0- 970 1-024 barrier.93b*c It is beyond the scope of this article and the competence of the Reviewer to discuss these quantities closely but it can be said that the last three are of significance in Substitution reactions.The potential barrier g3b is a quantitative means of accommodating the relative stabilities of transition states associated with substitution at various points and the polarisability measures the ease with which the electron density at a par- ticular point may be augmented during a substitution reaction. As an example we might consider a recent paper by Sandorfy and Y ~ a n . ~ % These authors give figures for the charge (XXXIII) the polarisability (XXXIV) and the potential barrier (XXXV) for quinoline in electrophilic substitutions as shown. Charges being considered the decreasing order of reactivity to Qa (a) Trans.Faraday SOC. 1947 43 87 ; (b) Wheland J. Amer. Chem. SOC. 1942 64 900 ; (c) Pading Brockway and Beach ibid. 1935 57 2705 ; Penney Proc. Roy. SOC. 1937 A 158 306 ; Daudel and Pullman J. Physique 1946 7,59 74 105 ; Coul- son and Longuet-Higgins R o c . Roy. SOC. 1947 A 191 39 ; ( d ) Compt. rend. 1949 229 715 ; Bull. SOC. chim. 1950 131 ; ( e ) Pullman Rev. Sci. 1948 86 219 ; Daudel Compt. rend. 1948 227 1241 ; Coulson Daudel and Daudel Bull. SOC. chim. 1948 1181 ; Daudel Buu-HOT and Martin ibid. p. 1202; Sandorfy {bid. 1949 615; Longuet-Higgins and Coulson J. 1949 971 ; Yvan Compt. rend. 1949 229 622 ; Buu-HOT and Daudel Bull. SOC. chim. 1949 801 ; Sandorfy VrQejq& Yvm Chalvet and Daudel ibid. 1950 304. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 399 electrophilic reagents would be C(s) > Ct3) > C(6) > C(5, etc.but the differences are small and the potential-barrier diagram gives the sequence > C(!) > Ct6) > C(3, etc. Only by considering the lower potential barrier (or the hzgher polarisability) of C(5) is that position given its correct place. In the manner indicated a fairly complete picture of the reactivities of heterocyclic nuclei is being built up.93e None of the theoretical work so far published has allowed for protonisation of these heterocyclic molecules so important in nitration and sulphonation. In view of the success of the theory and of the behaviour of the nitronaphthalenes it seems likely that protonisation does not change the site of substitution in these cases but merely increases the degree of 5-substitution in quinoline and of 8-substitu- tion in isoquinoline.Furthermore free quinoline would nitrate more quickly than the quinolinium ion and the two would be in equilibrium in solution. It is therefore interesting that nitration of the methylquinolinium ion which cannot dissociate gives the 5- and the 8-nitro-compound the former appar- ently predominating. Nitration of some substituted methylquinolinium ions also appears to proceed mainly at C(5).94 It should be mentioned however that in many halogenations some of which may be electrophilic substitutions but in which protonisation would not be important initial substitution occurs at C(3).95 There can be no doubt that the nitrations discussed above proceed by electrophilic substitution involving in most cases the nitronium ion.It therefore seems likely that in the reactions which convert quinoline into 7-nitroquinoline (Table VI) some other mechanism is involved. There is no evidence on this point. The hydroxy-heterocyclic compounds (Table VI) present an interesting problem. Various facts discussed in this Review make it doubtful whether the nitration of these weakly basic compounds (Table I) in sulphuric acid can be referred solely to protonised forms [e.g. (XXXVI)] in which sub- stitution would be expected to occur a t C(6) or C(s) as has been previously 70 and the =me facts make unlikely the suggestion 56* that 4- hydroxyquinoline derivatives for example are extensively protonised in nitric acid alone. However if some of the nitration of these compounds in sulphuric acid is referred to unprotonised forms it is difficult to explain the absence in the products of appreciable quantities of 3-nitro-compounds the major products when nitric acid alone is used.Halcrow and Kermack 70 suggested that the 3-nitro-compounds arose as the result of the directing properties of the hydroxyl group in the unprotonised heterocyclic com- pounds but the above discussion makes this uncertain and the effect of nitrous acid may be important.56b It is noteworthy that other electrophilic substitutions namely halogenations in which protonisation is unlikely proceed readily at C(3) in 4-hydroxyquinoline derivatives and somewhat less readily at the same position in 4-hydro~ycinnolines.~~ g4 Decker Ber. 1905 38 1274. 96 Manske Chem. Reviews 1942 30 137 ; Beilstein's '' Hmdbuch der Organischen Chemie " 4th edtn.Vol. 20 p. 342. B6 Schofield and Swain J. 1950 384. 400 QUARTERLY REVIEWS The entries in Table VI concerning 2- and 4-arninoquinoline recall the behaviour of the aminopyridines. It isinteresting that 4 6-diaminoquinoline and its 6-acetyl derivative should give 3-nitro-compounds s6@ (it is not clear whether this proceeds through nitramine formation or not) whilst 6-toluene- p-sulphonamidoquinoline when nitrated with nitric acid alone gives the 5-nitro-deri~ative.*~b V. Phenyl-substi~ted Heterocyclic Compounds Several workers have described the nitration o f phenyl-substituted heterocyclic compounds. This makes possible an interesting comparison of different heterocyclic groups with phenyl itself as substituents in the benzene ring throwing light on the problem of conjugation between linked aromatic systems and on the question of the effect of protonisation already raised in this Review.The cases so far reported are summarised in Table VII. Unfortunately not all of the nitrations were effected under the same conditions. - It is profitable to consider first the related nitrodiphenyls and these compounds are therefore included in Table VII. They are a striking example of the '' constancy of type of substitution ",97 o- m- and p-nitrophenyl showing the same op-directing properties as phenyl itself. Dewar 98 has included phenyl among the + E substituents 99 and has deduced that for this category the o p-ratio should fall as the electron affinity of the group rises. 2 4-Dinitrophenyl appears to be misplaced in the series.Dewar also suggests that steric influences tend to decrease the ratio forcing us to conclude that in the present case such influences are of minor importance. This is strange since if we assume the substituents to be of the + E type we imply that planar terms of the type (XXXVII) contribute to the struc- tures of the diphenyl derivatives which would lead us to expect a degree of steric interference in such cases as that of 2-dinitrodiphenyl. Early workers loo concluded from the absence of meta-substitution in this series that the nuclei behaved independently i.e. that conjugation between them was unimportant and further since the o :p-ratio for 2- and 4-dinitro- diphenyl was very similar that steric factors were not important here. The problem is complicated and the inadequacy of naive resonance pictures [e.g.(XXXVIII)] which would lead us to expect meta-substitution is stressed. Gull and Turner 100 suggested that in the case of diphenyl derivatives if a sufficiently strongly polarising group could be introduced into one of 97 Waters Ohem. Reviews 1930 7 407. 99 Ingold's nomenclature is used in this Review. loo Gull and Turner J 1927 491. ** J. 1949 463. SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 401 TABLE TI1 Nitration of phen~l-su~st~tuted heterocyclic compounds R in R*CBH,. 2-Pyridyl . . 3-Pyridyl . . . 4-Pyridyl . . . 2-Quinolyl . . 4-Quinolyl . . f 2-CSHQNMe . . Z-C,H,N.CH . . 3-C,H,N*CHz . . 4*C,HpN*CH . . 2-C,H4N*[CH,] . 2-Benzopyrylium . 4-Glyoxalinyl . . 2-Glyoxalinyl . . 2-Glyoxaline-4 5- dicarboxylic acid 2-Glyoxaline-4- carboxylic acid .2-14 5-Dihydro- gl yoxalinyl) 2-Nitrophenyl . 4-Nitrophenyl . 2 4-Dinitrophenyl 2-C,H,N*CHCH . . * . . . . . . . . . . ’ . . . * . . . . . * . . . . . . I . . . . . . Isomers % 0-. 5 ? 12-7 small 5 - - - - 16-0 40.1 25 1-5 - - - - 39 37 45 m-* _____- 34.9 ? 28.5 30 97 40 10.4 4.8 3.4 nil 86 0.2 52 52 80 - - - - - P-. 42.3 64.3 38.0 60 - - 50 66.7 63.1 70.2 64.5 49.0 69 50 19 19 - - 61 63 55 Conditions. Nitrate added to H,SO Ba:e addedi to HNO ? 9 ? ? 7 9 1 1 7 7 9 HNO alone or with H,SO Ratios changed but little Variety of conditions Perchlorate added to HNO Nitrate added to H,S04 Mixed acids 99 9 9 ? f f Nitrate added to H,SO HNO 9 Y Ref. ~- 101 101 101 102 102 103 104 104 104 105 105 106 107 108 108 108 109 100 100 100 the rings some meta-substitution might occur in the other.In passing from the diphenyl to the phenylpyridine series this state of affairs appears to have been reached with respect to nitration for whilst in this reaction substitution st ill occurs predominantly ortho-para considerable quantities of m-nitro-compounds now appear. Phenylpyridines and phenylquinolines appear therefore to fall into the category of anomalous compounds which undergo meta-para rather than ortho-para substitution but the explanation is presumably that we are observing the nitration of both protonised and unprotonised forms of the compounds the former being responsible for the appearance of m-nitro-compounds. As Flurscheim indicated 101 it is difficult to predict the orientation accurately in such cases because nothing is known of the relative rates of substitution of protonised and unprotonised forms (except that the former will be the greater).Thus whilst predictions of This paper contains interesting notes by Fliirscheim Ingold and Robinson. Iol Forsyth and Pyrnan J. 1926 2912. Le Fevre and Mathur J. 1930 2236. Koenigs and Nef Ber. 1887 20 622 ; Koenigs ibid. 1893 26 713 ; Besthorn and Jaegle ibid. 1894 27 907 ; Besthorn Banzhof and Jaegl6 ibid. p. 3035. lo4 Bryans and Pyman J. 1929 549. Io6 Le Fhvre J. 1929 2771. Io8 Pyman and Stanley J. 1924 2484. IoB Forsyth Nimka and Pyman J. 1926 800. lo6 Shaw and Wagstaff J. 1933 79. Grant and Pyman J. 1921 1893. 402 QUARTERLY REVIEWS the direction of nitration of phenylpyridines in a general sense were possible,lOl the significance of the observed o :p-ratios is a t present obscure.Similar remarks apply to the phenylquinolines and it is interesting to see that when the dissociation [(NR3H)+ + NR + H+] is not possible as is the case with methyl-2-phenylquinolinium meta-substitution occurs exclusively (see 2-phenylbenzopyrylium also) As would be expected with the benzylpyridines the amount of meta- substitution is less than with the phenylpyridines. The meta-substitution must be attributed to protonisation but comparison with the case of Ph*CH,*NMe (88% meta) shows that this cannot be regarded as efifec- tively complete. Again as would be expected the amount of meta-nitration varies in the order 2 > 4 > 3. Z-Z’-Phenylethylpyridine is nitrated to an even smaller extent in the meta-position than are the benzylpyridines because of the further removal from the phenyl group of the meta-directing positive charge in the protonised form and it is noteworthy that the variation in o :p-ratio with change in nitrating medium was found to be very small.lo5 This is not surprising since the inductive eEect will fall oE rapidly along the ethyl chain and the phenyl group will consequently be insensitive to small changes in the degree of protonisation of the pyridyl residue.[Even with Ph*CH,*CH,*NMe, (0 + p)-compounds account for 81% of the p r o d ~ c t . ~ l ~ ] 2-Stilbazole (trans-?) is the only compound of the type now under dis- cussion for which thorough studies of the variation in o p-ratio with change in nitrating medium have been made.lo5 The ratio varies considerably (0-37-0.56 for i o p ) . In the absence of sulphuric acid increase in con- centration combined with decrease in quantity of nitric acid decreases the ratio.Addition of soluble nitrates or acetic acid to the nitric acid increases the ratio but to a smaller extent than an equal weight of water. Nitration in sulphuric acid gives ratios (0.42-0.46) equal to those obtained with 87-92% nitric acid. These variations were attributed to changes in the base-salt-ion equilibria but a satisfactory theory of the o :p-ratio was not available a t the time. In terms of Dewar’s calculations if we attribute to the pyridylethylene group a + E effect 2-stilbazole is seen to bear a striking resemblance to a~etanilide.~~ As the concentration of nitric acid increases or when sulphuric acid is used the degree of protonisation of the pyridyl group will be increased; i.e.the electron-affinity of the sub- stituent will be increased and the o :p-ratio would be expected to fall as is actually the case. The observed ratios (which must be regarded as the over-all figures for nitration of both protonised and unprotonised forms) are in the same range as those for other + E substituents. These results provide a clear demonstration of the intervention of free base-protonised base equilibria in determining the proportions of isomers formed from com- pounds of this type a factor which has likewise been shown t o be of importance in oxygen compounds capable of protonisation.111 110 Ingold Rec. Trav. chirn. 1929 48 805 ; Ann. Reports 1926 23 131. 111 Baker J. 1931 307; Baker and Hey J. 1932 1236 2917. -f- + SCHOFIELD NITRATION OF HETEROCYCLIC NITROGEN COMPOUNDS 403 The phenylglyoxalines (Table VII) seem to represent border-line cases between the nitrodiphenyls and the phenylpyridines.Although in the diphenyl series a nitro-group is not powerful enough to change the directive character of the phenyl group it appears that in the glyoxaline series carboxyl groups are able to some extent to convert the heterocyclic nucleus containing them into a rneta-directing entity. VI. Conclusion The field reviewed is clearly in need of quantitative investigation not only kinetically but also with regard to the proportions of isomers formed in the nitration of a given compound with due consideration of the effect on these proportions of variations in reaction media. The evident importance of polarisability factors in some of the nitrations discussed is an interesting vindication of the general ideas of the English school of theoreticians.l12 Accurate experimental testing of quantum- mechanical calculations regarding heterocyclic compounds is a field as yet untouched.112 Remick " Electronic Interpretations of Organic Chemistry " 2nd edtn. New York 1949 Chapter V.

ISSN:0009-2681    

DOI:10.1039/QR9500400382

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

ANIONOTROPY By E. A. BRAUDE PH.D. A.R.C.S. D.I.C. (LECTURER IN ORGANIC CHEMISTRY IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY LONDON S.W.7) THE term anionotropy conventionally refers to molecular rearrangements involving the simultaneous transposition of a multiple bond and of groups or atoms such as OH OR and C1 which can exist as stable anions. This nomenclature dates from a time when such reactions were generally thought to involve the separation of the migrating group as a free anion in contrast to cationobropy involving the migration of groups or atoms such as H which can exist as stable cations. It has since become recognised however that molecular rearrangements like other organic reactions can proceed by homolytic as well as heterolytic bond fission and that heterolytic rearrange- ment does not necessarily involve the separation of free ions.Moreover there is no difference in principle between rearrangements involving the simultaneous transposition of two groups and systems involving the simul- taneous migration of one group and of an electron pair belonging to a multiple bond. Also the description in classical terms of rearrangements resulting in the migration of groups such as alkyl which do not form stable ions is ambiguous. For these reasons it will be convenient to re-define anionotropy to include all rearrangements in which the migrating or departing group or groups retain the electron pairs by which they were originally linked to the rest of the molecule. An account of some of the early work in this field was given by J. W. Baker 1 in 1934. At that time the subject was still one of mainly academic interest but during the last ten years many new anionotropic reactions have been discovered and have found numerous synthetic applications.Much progress has also been made in the elucidation of the mechanism of anionotropy and anionotropic systems have been shown to provide a useful basis for the study of other problems in theoretical organic chemistry such as the influence of substituents on reactivity. Diad Anionotropy According to the generally accepted ideas of Lowry,2 Whitm~re,~ and others diad rearrangements such as the pinacol-pinacone and oxime-amide (Beckmann) rearrangements involve a retention of complete electron octets by the migrating hydroxyl and alkyl or aryl groups and therefore represent anionotropic changes within the wider meaning of the term.It is not intended here to discuss in detail the numerous and well-known reactions 1 " Teutomerism " London 1934. 2 Institut de Chimie Solvay IIme Conseil de Chimie Paris 1926 p. 150. 8 J. Amer. Chem. Soc. 1932 54 3274. 404 BRAUDE ANIONOTROPY 405 of this type,** 5 but rather to draw attention to certain features common to diad and higher anionotropic systems. The most fully investigated example is the Beckmann rearrangement .* This has been studied under widely varied conditions generally giving rise to complicated catalytic influences but in aqueous solution the rearrange- ment of an oxime is a simple acid-catalysed first-order process and the rate constants are dependent on the acidity function of the medium.‘ There is no evidence for the intermediate formation of oxime esters under these conditions.8 The other main facts concerning the reaction are (i) it is facilitated by electron-donating substituents in the aryl or alkyl groups (ii) the migrating groups origina,lly occupy trans-positions about the C:N link and (iii) a migrating alkyl group retains its optical configurati~n.~ These facts are in accord with the mechanism (1) R R f H+ \ C=N _1\ C=N \ OH,+ OH \ - / \ R’ OH / R’ R’ R \ / -+ C-=N __+ R’-C-NHR .(1) OH ,+ / ti The first step is the reversible addition of a proton to the :N*OH group. The protons will be mostly attached to the more basic nitrogen atom to give the imoniurn ion :+NH*OH but this will be in equilibrium with the oxonium ion :N*OH,+ formation of which will weaken the N-0 link and facilitate the separation of the hydroxyl group as a neutral water molecixle.The second and rate-determining step is a double nucleophilic substitution in which the R and OH groups (the latter in the form of OH,) exchange positions each carrying their own electron pairs. The migration of It must be intramolecular ; the migration of OH on the other hand may be intra- molecular as shown or it may be intermolecular and take place by an attack of a solvent water molecule at the carbon atom. The third step is the fast reversible loss of a proton followed by enol-keto prototropy to give the amide or anilide. The characteristic features of the pinacol-pinacone rearrangement are very similar. The reaction is facilitated by electron-donating substituents and the rearrangement of benzpinacol in aqueous media is a first-order reaction and the rate-constants are dependent on the acidity function.10 Mechanism (2) which is analogous to (l) is in accord with all the facts a t 4 Watson Ann.Reports 1939 36 191 ; 1941 38 121. 6 Wallis and Gilman “ Organic Chemistry ” New York 1943 p. 965. Blatt Chem. Reviews 1933 12 215 ; Chapman J. 1935 1223 ; B. Jones Chem. Hammett Chem. Reviews 1935 16 67. Pearson and Ball J . Org. Chem. 1949 14 118. Reviews 1944 35 335; Nature 1946 157 519. 9 Campbell and Kenyon J. 1946 25. lo Braude unpublished. E E 406 QUARTERLY REVIEWS present known but as before the rate-determining migration may be inter- molecular with regard to the OH group instead of intramolecular as shown The rearrangement again takes place in the oxonium ion formed by addition of a proton to one of the hydroxyl groups; the C(,)-O and C(,,-R links undergo simultaneous fission and during this process the positive charge travels continuously from 0 to C(,) from C0) to C(z) and from C(z) back to 0.The centre of location of the positive charge in the transition state will be approximately midway between carbon atoms C(,) and C(z) but the above method of representation is adopted in order to indicate the process of migration and to distinguish mechanism (2) from those * postulating free mesomeric carbonium ions as intermediates. In an asymmetrical glycol the direction of the rearrangement will be determined both by the relative “ basicities ” of the two hydroxyl groups and by the ease of separa- tion of the migrating group R and it is therefore not surprising that no simple generalisations can be made concerning the “ relative migratory aptitudes ” of different groups.lI There is a close resemblance between these diad systems and triad anionotropy in which a hydroxyl or similar group exchanges positions with an electron R*CH*CH=CHR‘ 1 OH pair H+ 3 7 r CH l+ -H+ R.CH=CH*CHR‘ R*CH=CH*CHR’ .(3) I I OH ,+ OH Such reactions are discussed in detail in the following section. If R’ is an ethoxyl group rearrangement is followed by loss of ethanol to give 1 2 the aldehyde R*CH:CH*CHO and the analogy to changes of the pinacol- pinacone type is even more complete. l1 Bennett and Chapman Awn. Repor& 1930 27 114 ; cf Hatt Pilgrim and l2 Arms and van Dorp Nature 1947 160 189 ; Rec. Trav. chim. 1948 67 973 ; Stephenson J. 1941 478. Heilbron E. R. H. Jones Julia and Weedon J.1949 1823. BRAUDE ANIONOTROPY 407 Triad Anionotropy Rearrangements of Ally1 Alcohols (Three-carbon Oxotropy). -The only type of triad anionotropic system at present known is the so-called three- carbon system and the majority of examples are rearrangements of un- symmetrically substituted allyl alcohols. The special term oxotropy l3 is used for such anionotropic rearrangements involving only the migration of a hydroxyl group corresponding to the term prototropy for cationotropic rearrangements in which the migrating group is a hydrogen atom. Whereas triad prototropy is one of the commonest and longest-known types of molecular rearrangement triad oxotropy has been studied in detail only during the last two decades. Oxotropic changes are invariably brought about by acidic reagents (proton-donors) in contrast to prototropic changes which are catalysed by either acids or bases or by both.Initially neutral reagents such as acetic anhydride or acid chlorides have occasionally been employed but under these conditions partial or complete acylation takes place and rearrange- ment occurs in virtue of the free acid produced; if the reaction mixture is kept neutral e.g. by working in pyridine solution at low temperatures only the derivative of the original alcohol is obtained. With simple allyl alcohols containing only alkyl substituents oxotropic mobility is relatively low and isomerisation to the equilibrium mixture requires acid treatment at elevated temperatures or for prolonged periods. Thus 1 -methylally1 alcohol (I) and 3-methylallyl alcohol (crotyl alcohol) (11) are interconverted by 1% sulphuric acid at 95" (5 hours) to a mixture containing about 30% of the primary alcohol.l* The next higher homologues 1 l-dimethylallyl =,SO (I.) OH*CHMe*CH:CH CHMe:CH*CH,*OH (11.) alcohol and 3 3-dimethylallyl alcohol are similarly interconverted by treatment with 1% sulphuric acid at room temperature for 60 hours,15 l6 the equilibrium mixture containing nearly equal proportions of the two isomers.Although caution must be exercised in interpreting data obtained under non-homogeneous conditions in view of the large effect of solubility differences these two examples clearly show that alkyl substituents increase anionotropic mobility but have little influence on the relative stability of the two isomers. This conclusion is amply confirmed by kinetic studies in more complex systems (cf.below). In the presence of a chloro- or ethoxyl group in position 3 however rearrangement is followed by elimination of hydrogen chloride or ethanol and the equilibrium is disturbed.12 1' This provides a useful method for the synthesis of unsaturated aldehydes; thus 3-chloro-1 l-dimethylallyl alcohol (111) obtained by condensing Braude and E. R. H. Jones J. 1944 436 ; Braude ibid. p. 443. l4 Hearne and La France U.S.P. 2,373,956 ; Chem. Abstr. 1945 39 4081. I6 Locquin and Sung Wouseng Compt. r e d . 1922 174 1711 ; 175 100; Sung l* Nazsrov Azsrbaev and Rakchseva Bull. A~ad. SG~. U.S.S.R. 1946 419. I7 E. R. H. Jones and Weedon J. 1946 937 ; Toogood md Weedon J. 1949 Wouseng Ann. Chim. 1924 1 386. 3123; Bruun Heilbron Weedon and Woods J, 1950 633.408 QUARTERLY REVIEWS 2-chlorovinyl methyl ketone with methylmagnesium bromide is converted into ,&methylcrotonaldehyde (IV) when shaken with 10% sulphuric acid in ether.17 (111.) OH*CMe,*CH:CHCI CMe,:CHCHCl*OH --+ HaSO - HCI CMe,:CH.CHO (IV.) I n aryl-substituted allyl alcohols oxotropic mobility is considerably enhanced and the equilibrium is displaced in favour of the more highly conjugated arylvinyl isomer. The first example of this type was described by Valeur and Luce l8 who showed that l-phenylallyl alcohol (V) is con- verted into cinnamyl alcohol (VI) on treatment with dilute sulphuric acid whereas 1-cyclohexylallyl alcohol remains unchanged under these conditions. The rearrangement of 1 -phenylallyl alcohol into cinnamyl alcohol has acid (V.) OH*CHPh*CH:CH + CHPh:CH*CH,*OH (VI.) recently been studied in more detail by Braude Jones and Stern,ls who found that the isomerisation goes practically to completion in homogeneous aqueous-dioxan solution in the presence of dilute mineral acid.If aqueous ethanol is employed as a solvent a mixture of cinnamyl alcohol and cinnamyl ethyl ether is formed although cinnamyl alcohol is not etherified under these conditions. Detailed analysis shows that ether formation occurs mainly during and simultaneously with rearrangement but that 1 -phenyl- allyl alcohol also undergoes some etherification before rearrangement and that ethyl l-phenylallyl ether rearranges less readily than the alcohol. Numerous other examples of this type have been d e ~ c r i b e d . l ~ - ~ ~ The 1-arylallyl alcohols can be obtained by the condensation of an arylmagnesium halide with an ccg-ethylenic aldehyde or ketone by the condensation of an aryl aldehyde or ketone with an alkenylmagnesium halide or alkenyl-lithium derivative or by the selective reduction or an alkenyl aryl ketone as illustrated below for CMe :CHCOMe 7 ““..8 L Ph.COMlle CMe :CHLi A Ph*CMe*CH :CMe Grignard condensation of 1 3 3-trimethyl-l-phenyl- acids + Ph*CMe:CH*CMe I OH CMe,:CH*COPh /’ allyl alcohol.21 22 The 3-arylallyl alcohols obtained on rearrangement be independently synthesised by similar methods. 18 Bull. Xoc. chim. 1920 27 611. l9 Braude E. R. HI. Jones and Stern J. 1946 396 ; 1947 1087 ; Braude Stern J. 1947 1096 ; Braude Faweett and Newman J. 1950 793 ; Braude Fawcett J. 1950 800 and unpublished work.can and and 2* Burton and Ingold J. 1928 904 ; Burton J. 1928 1650 ; 1929 455 ; 1930 21 Braude and Timmons J. 1950 2000 2007 and unpublished work. 22 Braude and Coles J. 1950 2012 2014 and unpublished work. 248; 1931 759. BRAUDE ANIONOTROPY 409 The rearrangement of 1 -ary;lallyl to 3-arylallyl alcohols and similar reactions can be followed conveniently and quantitatively by taking advan- tage of the accompanying change in ultra-violet light absorption ; 1 3 9 l9 thus cinnamyl derivatives exhibit an intense absorption band near 2510 A. associated with the conjugated (styryl) system whereas l-phenylallyl deriva- tives show only weak absorption in this region. By; measuring the change in intensity of the 2510-~. band the rate of isomerisation can readily be determined. Hardly any kinetic data concerning three-carbon anionotropy were available before the introduction of the spectrometric technique but during the last few years detailed investigations have been carried out by this means.Oxotropic rearrangements in solution are invariably found to be simple first-order reactions and the rate constants ( E ) are dependent on Hammett’s acidity function I€ which is a measure of the tendency of the medium to convert a neutral base (B) into its conjugate acid @HI).’$ 23 At low acid concentrations and for a given solvent H is proportional to the hydrogen-ion concentration but at high acid concentration H increases more rapidly and both H and li also show a characteristic non-linear dependence on solvent composition in organic solvent-water mixtures which can be explained in terms of the changes in solvent “ structure ” and the mode of solvation of the hydrogen ions.13 23 The proportionality between log E and H means that the rate of rearrangement depends on the fraction of the allyl alcohol (ROH) which is converted into its conjugate acid and that the actual rearrangement is preceded by a fast reversible forniation of the oxonium ion (ROH,+).The subsequent migration may be intra- molecular as represented in (3) (p. 406) or it may take place intermolecularly through an attack by a water or similar molecule at as shown in (4). c-c=c c-c=c ___j. C=C=C I OH H+ 4- - H 2 0 fast 1 RoH slow [ ] -xz OH,+ OH OHR - H+ c=c-c ,- c=c-c . (4) OHR OR 1 I + fast Competition between different solvent molecules thus leads to partial etherification in alcoholic solvents and under such conditions the inter- and intra-molecular reactions probably occur side by side.In other sol- vents e.g. anhydrous dioxan on the other hand the migration must be largely intramolecular. The function of the hydrogen-ion catalyst is to weaken the existing carbon-oxygen bond ; 13 instead of the separation of a charged hydroxyl anion the migration involves the energetically much more favourable separation of a neutral water molecule. In the rate- determining step the positive charge undergoes a continuous transfer from the departing oxygen atom via the allyl group to the entering oxygen atom while a continuous electron displacement ta.kes place simultaneously in the 23Braude J. 1948 1971; Braude and Stern J. 1948 1976 1982; Nature 1948 161 169. 410 QUARTERLY REMEWS opposite direction.Both the formation of the oxonium ion and the fission of the carbon-oxygen bond are facilitated by electron accession at the reaction centre and oxotropic mobility is therefore increased by electron- donating substituents and decreased by electron-attracting substituents in the allyl and phenyl groups (Tables I and 11). The effect of substituents is exerted mainly through changes in the energy of activation. For nuclear substituents the first-order rate constants (k) at one temperature vary in the sequence p-Br < p-C1< p-F < H < m-Me < o-Me < pMe < p-Me0 and show the usual linear relation to the dissociation constants of the cor- responding substituted benzoic acids. l9 The result for the halogen deriva- TABLE I Rate constants and energies of activation in three-carbon oxotropic sgstems HO*CHX*CH:CHMe + CHX:CH*CHMe*OH 139 199 21* 30 X.Phenyl . . . O-Tolyl . . . m-Tolyl . . p-Tolyl . . . p-Fluorophenyl . p-Chlorophenyl . p-Bromophenyl . p-Methoxyphenyl 1-Naphthyl . . 2-Naphthyl . . 9-Anthryl . . 9-Phenanthryl . 2-Thienyl . . Vinyl. . . . Ethynyl . . . 2-Furyl . . . Acetyl . . . ~ ~~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . kao a (nun.-*). 1.84 3.59 2.17 1-50 0.43 0.34 3.45 3-20 8.05 1.46 16.1 I89 166 63.6 1.75 0.00036 0.15 x 10-6 EAlT.6 (kcals./g.-mol.). 19.5 19.5 20.4 19.6 19.4 20.3 20.4 17.9 19.9 20.2 19.5 20.3 16.8 17.4 19.9 26.9 - a First-order rate constants divided by acid concn.for 60% aqueous-dioxan hydro- * Arrhenius energies of activation. Extrapolated from measurements on 1-acetyl-1 3 3-trimethylallyl alcohol (XXIX p. 416). tives also confirm the '' reversed " order of electromeric effects I? > C1 > Br frequently encountered in other reactions. The ready accessibility of aryl- allyl alcohols the mild conditions under which isomerisation takes place the absence of side-reactions and the comparatively high sensitivity of the reaction towards structural changes combine to render oxotropic rearrange- ment of this type an excellent basis for investigating the electronic properties of substituents which have been little studied in this respect because of a lack of suitable reactions.In polycyclic arylallyl alcohols k- increases in the order 9-phenanthryl< phenyl < 2-naphthyl N l-naphthyl < 9-anthryl and except for the phenanthryl derivative which is anomalous a linear relation holds between E and the expected number of polar resonance forms con- tributing to the electronic effect of the polynuclear substituent .I9 The chloric acid solutions at 30". BRAUDE ANIONOTROPY 41 1 large increase in k observed with l-fury1 and l-thienyl substituents can be discussed in similar terms.? The greater stability of the 3-arylallyl as compared with the l-arylallyl alcohols indicated by the almost complete displacement of the oxotropic equilibrium must clearly be ascribed to the additional resonance energy associated with the conjugated styryl or arylvinyl system.If aryl groups are present in both the 1- and the 3-position the difference in thermo- dynamic stability between the two isomers is much reduced and the equilibrium is again restored. Reversible systems of this type were first investigated by Burton and Ingold,2* who found that the equilibrium mix- ture obtained by heating the phenyl-p-tolylallyl alcohols (VII VIII ; X = p-tolyl) with acetic anhydride contained about equal proportions of the two acetates while with the phenyl-p-chlorophenylallyl alcohols the p-chlorostyryl isomer (VIII ; X = p-C,H,Cl) predominated. The phenyl- p-nitrophenylallyl alcohols (VII VIII ; X = p-N0,*C,H4) and phenyl- naphthylallyl alcohols (VII VIII ; X = naphthyl) are similarly inter- (VII.) HO*CHX*CH :CHPh + CHX:CH*CHPh*OH (VIII.) converted in the presence of dilute mineral acids.13* 24 By means of the change of equilibrium with temperature the heat of rearrangement which should be approximately equal to the difference in resonance energy of the styryl system in (VII) and the arylallyl system in (VIII) can be deter- mined and since the resonance energy of the styryl system is known from thermal data this provides a promising method for estimating reson- ance energies of other conjugated systems.A case of oxotropy in a 1 3-diarylallyl system in which one of the isomers undergoes further reaction occurs in the formation of 2 3-dimethyl-l-phenylindene from a-benz ylidenepropiophenone Ph*CO*CMe:CHPh ___j) Ph*CMe*CMe:CHPh --+ MsMgBr RCl I OH In the 1 1 3-triphenylallyl alcohol system the equilibrium again lies entirely on one side and only the diphenylvinyl derivatives are isolated under acidic conditions 26 Ph.CHCH*MgBr ___j.OH-CPh,*CH.CHPh ____I_ CPh,:CR.CHPh*OMe A displacement of oxotropic equilibrium similar to that shown by aryl- allyl alcohols might be expected in allyl alcohols containing other types of unsaturated substituents. Hardly any such systems were known however PhSCO HCI-MeOH ** Braude and Waight unpublished. 2b Smith and Hanson J. Amer. Chern. Soc. 1935 57 1326. as K. H. Meyer and Schuster Ber. 1922 65 815 ; Straw and Ehrenstein AnnaZen 1925 442 93; K. Ziegler Ber. 1925 58 369. 412 QUARTERLY REVIEWS until the discovery by E. R. H. Jones and McCombie 27 of the acid-catalysed rearrangement of 1 -ethynylallyl alcohols (e.g, IX) to 3-ethynylallyl alcohols (e.g. X) HCi CNa acids CHMe:CH*CHO - CHiCGH(0H)CH:CHMe + CH ;CGH :CH*CHMe*OH This reaction proved to be quite general and has found many important applications particularly in the synthesis of vitamin A and of vitamin A analogues from acetylenic precursors.28p 29 The 1 -alkynylallyl alcohols are readily accessible by the condensation of acetylenic Grignard or alkali- metal derivatives with ethylenic aldehydes or ketones. Isomerisation to the 3-alkynylallyl alcohols which are characterised by high-intensity light absorption near 2200 A. associated with the conjugated vinylacetylene (1x4 (X.1 The efSect of systems X. - R1. - H Me H H H Me H Me H Me Me ___ TABLE I1 methyl substituents on the *mobility of three-carbon oxotropic .~ RS. __ H H Me H H H H H H H H - 0*00000015 0~00009 0~0000011 0.00058 0,133 4.5 0.000054 0.0164 0.81 5-60 354 x.- H H H Me H __ H H H __ R 2. __ H Me H H H .~ H H H -_ RS. ~ H H Me Me Me __ H Me Me R4. H H H H Me H H Me 0.0018 0.0033 1.73 22 150 0.0043 2.21 79 0 First-order rate constants divided by acid concentration for 60% aqueous-ethanolic hydrochloric acid solutions at 30". chromophore can usually be effected in high yield by treatment with dilute mineral acids. Although the oxotropic equilibria again lie far on the side of the conjugated isomers oxotropic mobility in the alkynylallyl system is much lower 30 than in the corresponding arylallyl system (Table I) as might be expected in view of the strong electron-attracting properties of acetylenic groups. As before rearrangement is greatly facilitated by alkyl J. 1943 261. 28 Heilbron E. R. H.Jones and Raphael J. 1943 264 ; Heilbron Johnson E. R. H. Jones and Raphael ibid. p. 265 ; Heilbron E. R. H. Jones and Weedon J . 1944 140; Cymarman Heilbron and E. R. H. Jones J. 1944 144. For summaries see Sir Ian Heilbron J. 1948 386 ; Johnson Ann. Reports 1949 46 168 ; E. R. H. Jones J. 1950 754. 30 Braude and E. R. H. Jones J. 1946 122 128. BRAUDE ANIONOTROPY 413 substituents and the relative rates of isomerisation of alkynylallyl alcohols containing up to four such substituents cove? a range of over lo9 (see Table 11). The differences in reactivity can be analysed into rate factors which are of the order of 102 10 and lo3 for methyl substituents attached to the 1- and 2- and 3-ally1 positions respe~tively.~O The results clearly demonstrate the operation of two different substituent effects the induc- tive effect which rapidly decreases with increasing distance between the substituent and the reacting centre and the much larger tautomeric effect which comes into play only on the interposition of the highly polarisable ethylenic group.By partial catalytic hydrogenation alkynylallyl alcohols can be con- verted into the corresponding alkenylallyl alcohols (dialkenylcarbinols) and it was shown by Heilbron E. R. H. Jones McCombie and Weedon 31 that the latter undergo oxotropic rearrangement with much greater ease than their acetylenic analogues in accordance with the weaker electron-attracting properties of ethylenic as compared with acetylenic groups. Thus vinyl- ethynylcarbinol obtained by condensing acraldehyde with sodium acetylide is selectively hydrogenated to divinylcarbinol which rearranges to buta- dienylcarbinol (XI).CH; C*CH( OH) *CH :CH ___j. CH :CH*CH( OH)*CH :CH __j H 2-Pd acids CH :CH*CH:CH*CH,*OH (XI.) Dialkenylcarbinols can also be prepared from dialkenyl ketones but this method is rather limited in scope as the required ketones themselves are not usually accessible. However diisobutenylmethylcarbinol (XII) can be obtained from the readily available diisobutenyl ketone (phorone) and methylmagnesium bromide. The rearrangement of highly substituted dialkenylcarbinols such as (XII) occurs so readily that unless careful preclautions are taken to exclude all traces of acids during their prepara- tion and isolation only the isomeric butadienylcarbinols are obtained. The latter easily undergo dehydration to the corresponding substituted hexa-1 3 5-trienes (e.g.XIV).21* 22 (CMe,:CH),CO -+ (CMe,:CH),*CMe*OH -+ CMe,:CH*CMe:CH*CMe,*OH -+ CMe :CH*CMe :CH*CMe :CH (XII.) (XIII.) (xrv.) A more direct and in some respects more general route to dialkenyl- carbinols as well as to other anionotropic systems which has recently been developed 2~ 22 consists of the condensation of alkenyl organometallic derivatives with carbonyl compounds. With the exception of 2-arylated vinyl bromides such as p-bromostyrene vinyl halides do not usually form Grignard derivatives but the corresponding lithium alkenyls can be obtained in good yields under appropriate conditions. Thus isobutenyl bromide with lithium in ether gives isobutenyl-lithium which with acraldehyde 3L J. 1945 84 88. 414 QUARTERLY REVIEWS yields isobutenylvinylcarbinol (XV).Rearrangement of such unsym- metrical dialkenylcarbinols' could theoretically proceed in two directions but it is found 21s 31 that the hydroxyl group always migrates initially to CH :CH*CHO CMe :CHLi ! CH :CH*CH( OH) *CH CMe \ \ \ 'u (XV4 CH CHCH :CH-CMe 2*OH I (XVI. ) I H O *CH,*CH :CH*CH :CMe (XVII.) the more alkyl-substituted C(al and that the resulting butadienylcarbinols (e.g. XVI) then undergo a much slower reversible further rearrangement to the other conjugated isomer (e.g. XVII). The final equilibrium mixture contains approximately equal proportions of the tertiary and the primary carbinol and it may therefore at first sight appear surprising that the initial rearrangement should proceed exclusively in one direction. How- ever the influence of alkyl substituents on the relative rates of two possible rearrangements of the same carbinol will be similar to their effect on the relative rates of rearrangement in two different carbinols.Since a 3-methyl substituent increases the rate of rearrangement by a factor of about 103 (XVI) will be formed from (XV) at least 1000 times faster than (XVII) and in conditions leading to the rearrangement into (XVI) at a measurable rate the amount of (XVII) formed will be indetectably small. The equili- brium position in the subsequent five-carbon oxotropy on the other hand is governed by the almost equal stability of the two conjugated isomers (cf. below). The carbinols (XVI) and (XVII) can be distinguished by their high-intensity absorption maxima at 2235 and 2360 A. respectively charac- teristic of a mono- and a tri-alkylated butadiene chromophore and their structures are rigidly proved by complete catalytic hydrogenation to the saturated alcohols.The effect of alkyl substituents in determining the direction of three-carbon oxotropy is also clearly shown22 in the cyclo- hexenylcarbinols rearrangement of which gives exclusively the allylidene- cyelohexanol (XVIII) when R = H but the cyclohexenylbutenol (XIX) when R = Me. CHR:CH*CHO -+ 7 / CH(OH)*CH :CHR @ o ~ ~ ~ (XVIII. ) :CH*CHR*OH (XIX.) In substituted butadienylvinylcarbinols such as (XX) which represent potential three- as well as five-carbon oxotropic systems triad rearrange- ment takes place preferentially. Reactions of this type form a key step in one of the methods successfully employed for the synthesis of vitamin A BRAUDE ANIONOTROPY 415 (XXI) and vitamin A analogues.32 Rearrangement and simultaneous dehydration of the glycol (XX) (in the form of its primary acetate) can be effected in high yield by treatment with a trace of iodine in boiling xylene a method which probably depends on the formation of a small amount of hydrogen iodide which acts as a catalyst. CH,R-CH :CMe*CHO + CH i C-CMe :CH*CH,*OH CH,R*CH :CMe*CH(OH)*CiC*CMe :CH*CH,*OH CH ,R *CH :CMe *CH( OH) CH :CHCMe CH-CH ,*OH (XX. 1 [ CH,R* CH (OH) *CMe :CH* CH CH*CMe :CH*CH ,*OH] J - H*O CHR:CH*CMe :CH*CH:CH*CMe :CH*CH,*OH (XXI.) (R = 2 6 6-trimethylcyclohex-l-enyl.) The systems so far discussed demonstrate that unsaturated substituents such as aryl alkynyl and alkenyl groups invariably displace the oxotropic equilibrium towards one side but increase or decrease mobility depending on their electron-donating or electron-attracting properties.Low oxotropic mobility would therefore be expected for allylic systems terminated by carbonyl carboxyl or nitrile groups. Probably the first example of three- carbon anionotropy to be found in the literature occurs in the conversion of crotonaldehyde cyanohydrin into Iaxwlic acid 33 CHMe :CH*CH( OH) *CN ka [CH ,Me*CH :C( OH)*CN] CH,Me*CH,*CO*CN HY (XXII.) CHMe :CH*CH(OH)*CO ,H (XXIII.) HO*CHMe*CH :CH*CO ,H [HO-CMe :CH*CH,*CO ,HI 4 J. 4 .1 .1 CH,Me*CH,*CO *CO ,H (XXV. ) (XXIV. ) Me-CO *CH,*CH,*CO ,H Hydrolysis of the cyanohydrin with concentrated hydrochloric acid yields 1 -hydroxybut-2-ene- 1 -carboxylic acid (XXII) which undergoes oxotropic 32 Isler Huber Ronco and Kofler Helv.Chim. Acta 1947 30 1911 ; Cheeseman Sir Ian HeiIbron E. R H. Jones Sondheimer and Weedon J. 1949 1516. 33 Fittig Annalen 1898 299 37 ; Bougault Compt. rend. 1913 156 1469 ; 157 377 403; J. Pharm. Chirn. 1913 8 393 406. 416 QUARTERLY REVIEWS rearrangement to 3-hydroxybut-1-ene-1-carboxylic acid (XXIII). The latter can be isolated in the form of its lactone but easily undergoes further double prototropic rearrangement to lavulic acid (XXIV) via the unstable enolic form. Similar but reversible oxotropy takes place in l-hydroxy- 3-phenylprop-2-ene- I-carboxylic acid obtained by hydrolysis of cinnam- aldehyde cyanohydrin. 33 In the crotonaldehyde cyanohydrin itself migra- tion of a hydrogen atom occurs in preference to that of the hydroxyl group.Thus treatment with phosphorus tribromide followed by water yields not the conjugated nitrile corresponding to the acid (XXIII) but butyroyl cyanide and a-ketovaleric acid (XXV).34 The fact that oxotropy is retarded while prototropy is facilitated by electron-attracting groups is illustrated even more clearly by the corre- sponding hydroxy-ketones. 4-Hydroxyhept-5-en-3-one (XXVI ; R = Et) obtained by the interaction of crotonaldehyde cyanohydrin with ethyl- magnesium bromide undergoes prototropy to heptane-3 4-dione (XXVII ; R = Et),21 and hydration of l-ethynyl-3-methylallyl alcohol (IX) leads to hexane-2 3-dione (XXVII ; R = Me) the intermediate hydroxy-ketone (XXVI ; R = Me) undergoing immediate rearrangement under the acidic conditions employed 35 acid CHMe:CH*CH(OH)*COR + (XXVI.) [ CH,Me*CH :C( OH) *COR] -+ CH,Me*CH ,420 *COR (XXVII.) Oxotropic rearrangement in systems of this type can however be realised if prototropy is precluded by replacing the hydrogen atoms concerned by alkyl groups. 3-Hydroxy-3 5-dimethylhex-4-en-2-one (XXIX) obtained by the condensation of isobutenyl-lithium with diacetyl dimethyl monoketal (3 3-dimethoxybutan-2-one) followed by careful hydrolysis of the ketal (XXVIII) is converted by 1% sulphuric acid at 60" into the isomeric 5-hydroxy-3 5-dimethylhex-3-en-%one (XXX) the constitution of which is proved by oxidative degradation to diacetyl and a-hydroxyisobutyr- aldehyde. 21 (MeO) ,CMe*CO*Me -+ (MeO) ,CMe*CMe( OH)*CH $Me + (XXVIII.) Me*CO*CMe( OH) *CH :CMe -+ Me*CO*CMe :CH*CMe ,*OH Rearraagements of Propargyl Alcohols.-A type of three-car bon oxotropy less common than that so far discussed occurs in substituted propargyl (a-acetylenic) alcohols.If an ally1 group is also present as in l-ethynyl- 3-methylallyl alcohol (IX) ordinary allylic rearrangement takes place but in other cases the hydroxyl group can exchange positions with one of the electron pairs of the triple bond resulting in the formation of an allene (XXIX.) (XXX.) 34 Rambaud Compt. rend. 1933 197 689. Heilbron E. R. H. Jones Smith and Weedon J. 1946 54. BRAUDE ANIONOTROPY 41 7 alcohol which immediately undergoes further prototropic rearrangement to an unsaturated aldehyde or ket0ne.3~~ 37p 38 As might be expected in view of the lack of electron accession at the hydroxyl group and in view of the steric strain associated with the formation of the allene system such rearrangements take place much less readily and much less smoothly than allylic oxotropy.Thus whereas 1 -phenylallyl alcohol is almost quantita- tively converted into cinnamyl alcohol by dilute mineral acid the isomerisa- tion of 1 -phenylpropargyl alcohol (XXXI) into cinnamaldehyde (XXXII) requires prolonged treatment with moderately concentrated sulphuric acid and is accompanied by side-reactions 37 HO*CHPh*CiCH -+ [CHPh:C:CH*OH] + CHPh:CH*CHO (XXXI.) (XXXII.) Rearrangements of Ally1 Esters.-Esterification of allylic alcohols usually yield3 a mixture of the isomeric allyl esters unless special precautions are taken to neutralise the acid produced in the replacement reaction (cf. p. 407). Numerous examples of this kind are to be found in the earlier literature.In a well-known group of papers,20 Burton and Ingold reported some quali- tative experiments on the degree of rearrangement accompanying replace- ment-acetylation 9-nitrobenzoylation and -bromination of arylallyl alco- hols and concluded that the ease of rearrangement increases in the order alcohol < acetate < p-nitrobenzoate < bromide Le. in the order of the stability of the anions OH- OAc- NO,*C,H,*CO,- and Br-. These authors also examined the isomerisation of 1 -phenylallyl p-nitrobenzoate in different solvents and observed that the rate of rearrangement increases in the sequence p-xylene < chlorobenzene < acetic anhydride < benzonitrile i.e. in the order of increasing dielectric constants and ionising properties of the media. Mainly on the basis of this evidence Burton and Ingold suggested that anionotropic rearrangement proceeds by way of a rate-determining ionisation into a carbonium ion and a fully dissociated anion and that the mesomeric carbonium ion then rapidly undergoes further reaction ( 5 ) 4- - A‘- A*C-C=C + A - + C x C z C ___ A-+C=C-CA’ .( 5 ) Further support for this view has been adduced in later papers by Hughes Ingold and their collaborators.3Q More recent investigations however have shown that Burton and Ingold’s observations receive a more complex explanation than could perhaps at first have been surmised and make it unlikely that mechanism (5) plays any significant part in the aniono- tropic rearrangement of allyl esters. (For it more detailed discussion see reference 40.) 36 K. H. Meyer and Schuster Ber.1922 55 819 ; Rio Cornpt. rend. 1949,228 690. 37MacGregor J . Amer. Chem. SOC. 1948 70 3953; 1950 72 2501. 38 Chanley ibid. p. 244 ; Hennion et al. ibid. 1949 71 2813 ; 1950 72 3542. 39 Catchpole and Hughes J. 1948 1 4 ; Catchpole Hughes and Ingold J. 1948 8 ; cf. de la Mare England Fowden Hughes and Ingold J . Chim. physique 1948 45 236. 4o Braude Ann Reports 1949 40 125. 418 QUBRTERJLY REVIEWS A careful study of the rearrangement of allyl esters in various solvents was carried out by Meisenheimer and his co-~orkers,~l who made the important observation that the rearrangement of 1 -phenylallyl p-nitroben- zoate is " autocatalytic " and is accompanied by the formation of p-nitro- benzoic acid. Later Kenyon Partridge and Phillips42 found that the rearrangement of allyl esters is catalysed by dilute acids and that the hydrogen phthalates of phenylallyl alcohols undergo self-rearrangement whereas the corresponding sodium salts or neutral esters are quite stable.These facts clearly suggest that the rearrangements of allyl esters like those of allyl alcohols are hydrogen-ion catalysed and take place by mechanisms analogous to (3) or (4). This has been confirmed by recent investigations. The rates and energies of activation of the rearrangement of 1-ethynylallyl acetate in aqueous acid media are very similar to those of the alcohol; 43 rearrangement of the acetate is accompanied by hydrolysis and a detailed ' fcso 1 (mn. l). 1 fcso 1 EBn. ' (mm. l). (kcals./g.-mol.). I I ___ I_ --___I TABLE I11 Rmrrangements of 1 - ~ h e n y ~ ~ l y l derivatives by 0.OlM-hydrochbric acid in 60% a>queous diozan at 9O0.Z4 E A ~ 1 (kcals./g.-mol.).~ Alcohol . . Acetate . . 1.39 25.5 Benzoate . . 0.27 1 26.5 2.97 1 24.5 f 0.5 l a First-order rate constants divided by acid concentration. p-Nitrobenzoate I 0.79 1 26.5 Methyl ether 1 1.28 ~ 27.0 i analysis indicates that the following reactions occur side by side (i) intra- molecular rearrangement followed by hydrolysis (ii) simultaneous inter- molecular rearrangement and hydrolysis through attack by solvent water molecules at Ct3) and (iii) hydrolysis of the acetate without rearrangement followed by rearrangement of the alcohol produced. Similar results have been obtained with 1-phenylallyl esters and ethers 24 (see Table 111). Contrary to the requirements of mechanism ( 5 ) the relative mobilities clearly bear no relation to the anionic stabilities of the migrating groups but they are fully in accord with mechanism (6) in which relative mobility will be a complex function of the combined effects of the ester group on the proton-addition equilibrium and on the ease of fission of the carbon- oxygen bond in the oxonium ion.In the intramolecular rearrangement of the esters the transition state may involve a six-membered ring with the acyl-oxygen atom becoming attached to C, ( 6 ~ ) . ~ 2 The slow rearrangement of allyl esters in neutral solvents is due to the formation of carboxylic acid which then catalyses the rearrangement. In aqueous or alcoholic solvents the carboxylic acid is produced by hydrolysis or alcoholysis ; thus 3-methyl-1 -phenylallyl p-nitrobenzoate is converted 4 1 Meisenheimer Schmidt and Schafer Annalen 1933 501 131.J. 1937 207. 4s Braude J. 1948 794. BRAUDE AMONOTROPY 419 in methanol solution into methyl 1-methyl-3-phenylallyl ether and p-nitro- benzoic acid.42 In non-hydroxylic solvents such as chlorobenzene the carboxylic acid is produced 24 by a side-reaction the nature of which has not .C. / \. c c HO 0 'CRY. c-c-c I O*CO*R 11H' &C=C ..... fl.....c ~-R-CO,H c==c-c I OH,+ I HO+*CO*R 11 - H+ 11 -+ c=G-c C==C-C AH I O*CO*R yet been established with certainty ; in the case of 1-phenylallyl esters it may be a direct 1 2-elimination to give the unstable phenylallene (e.g. XXXTII) R*CO*O*CHPh*CH:CH -+ CHPh:C:CH + R*CO,H (XXXIII. ) Definite proof that the rearrangement under these conditions is acid- catalysed and not spontaneous is provided by the fact that the values of the rate constants in the absence of added acid fall on the straight lines obtained by plotting the rate constants in the presence of added acid against the total acid concentration.24 Comparisons of anionotropic mobility in neutral solvents thus have little significance and some of the earlier observa- tions of Burton and Ingold (see above) arise from the fact that the amount of carboxylic acid eliminated varies with the ester and with the solvent and that p-nitrobenzoic acid is a better catalyst for rearrangement than is the weaker acetic acid.Rearrangements of Ally1 Halides.-The preparation of allyl halides by replacement reactions of allyl alcohols with hydrogen halides phosphorus 420 QUARTERLY REVIEWS halides or similar reagent^,^*-^' by addition reactions of conjugated dienes,M or by substitution reactions of olefins with N-halogen compounds 49 generally gives rise to a mixture of the two isomers.Recent work has shown that contrary to earlier suggestions,20 the formation of mixtures in the case of allyl chlorides is not caused by the lability of the halogen derivatives but rather by rearrangement of the allyl alcohol under the influence of the acid reagent or by simultaneous occwrence of 1 2- and 1 4-addition. The chlorides can be separated by fractionation and undergo rearrangement in solution rather less readily than do the corresponding alcohols or esters. The isomeric rearrangement of allyl chlorides has been little studied but methylallyl and ethylallyl chlorides have been found to isomerise slowly in boiling acetic acid 44 or under the influence of hydrochloric acid.45 It thus appears that anionotropy in allyl halides is subject to hydrogen-ion catalysis and may possibly involve the intermediate formation of halonium ions RHalH+ analogous to oxonium ions ROH,+.The possible occurrence of another mode of heterolytic catalysis is indicated by the interesting observation 4 5 ~ 50 that the isomerisation of ethylallyl chlorides in the pres- ence of hydrochloric acid is accelerated by ferric chloride zinc chloride and other multivalent metal chlorides. The effect of the metal chloride may well be due to co-ordination with the allylic chloride atom and to an ionisation process RC1 + MC1 fi R+ + MC1,- analogous to that believed to be operative in Friedel-Crafts reactions.In contrast to allyl chlorides isomeric allyl bromides are difficult to separate and readily undergo interconversion. 4 5 p 47 Here rearrangement is accelerated by peroxides and similar agents 45 and is probably a t least in part a free radical process involving homolytic fission of the C-Br bond and migration of bromine atoms. Non-isomeric rearrangement in allyl halides has been more fully investi- gated Whereas replacement reactions of allyl esters under basic conditions invariably yield only the unrearranged derivatives,40 alkaline hydrolysis alcoholysis and similar reactions of allyl halides are often accompanied by partial or complete rearrangement. The first observation in this field was made by Charon,51 who found that the treatment of einnamyl chloride 44 Meisenheimer and Link Annalen 1930 479 211 ; Meisenheimer and Beutter 46 Kharash Margolis and Mayo J.Org. Chem. 1936,1,393 ; Kharash Kritchevsky 48 Martin and Trinh Compt. rend. 1949 228 688. 47 Winstein and Young J. Amer. Chem. SOC. 1936 58 104; Young and Lane ;bid. 1937 59 2051 ; 1938 60 847 ; Young and Nozaki ibid. 1940 62 311. 48 Petrov J . Qen. Chem. Russia 1943 13 741 ; UltBe J. 1948 530 ; Rec. Trav. chim. 1949 68 125 352 483. Karrer and Ringli Heh. Chim. Acta 1947 30 863 1776 ; Braude and Waight Nature 1949 164 241 ; Bateman Cuneen and Koch ibid. p. 242. Pudovik and Arbuzov Bull. SOC. chim. U.R.S.S. Cl. Sci. Chirn. 1946 427 ; Pudovik Nikitina and Aigistova J. Gen. Chem. Russia 1949 19 67 279 ; Pudovik ibid. p. 1179. 61 Charon Bull. SOC. chim. 1910 7 86 ; Dupont and Labaume Chew. Zentr. 1910 ibid.1934 508 58. and Mayo ibid. 1937 2 489. 11 734. BRAUDE ANIONOTROPY 421 with methanol in the presence of sodium carbonate yielded mainly the methyl ether of 1-phenylallyl alcohol (V). Some of the more important reactions which have been examined in this respect are summarised in Table IV. TABLE IV Replacement reactions of allyl chlorides Y*CHR-CH.CHR’ (a-substitution) CHX:CH*CHR’Y (y-substitution) I’ Cl*CHR*CH:CHR’ Reaction. Hydrolysis (H20-EtOH-Na,C03) . Ethanolysis (Et0I-I-NaOEt) . . Acetolysis (AcOH-KOAc) . . . Amination (NHE$) . . . . . R. Me H H Me H H c1 Me H H Et H H Me R’ H Me Ph H Me Ph Ph H Me Ph H Et CjCH CiCH Product. U + Y y+ Y Malnly y a*+ Y Mamly a Mainly a Mainly y % + Y U + Y U + Y Y U OL Y Ref. 44 52 44 52 44 39 62 39 52 44 63 44 52 44 52 44 44 44 54 54 The extent of rearrangement accompanying substitution is intimately dependent both on the structure of the allyl halide and on the reagent.Few valid generalisations can be made but “ abnormal ” or y-substitution is usually favoured by neutral rather than strongly alkaline conditions 399 * ‘ 9 55 and kinetic investigations 39 61 on the hydrolysis alcoholysis and acetolysis of the methyl- and ethyl-ally1 chlorides indicate that second-order reaction with charged anions proceeds in the normal manner and that y-substitution is associated with first-order reaction with the solvent. Hughes Ingold and their collaborators 39 have concluded that y-substitution takes place exclusively by mechanism ( 5 ) and involves a unimolecular ionisation of the halide to give the mesomeric carbonium ion which then undergoes reaction at the a- and the y-position at comparable speeds.This view does not readily account for the fact however that the thermodynamically less stable deconjugated products are obtained from cinnamyl and other con- jugated halides ; if the mesomeric carbonium ion were formed the positive charge would be expected to remain concentrated at C,,,. For these and other reasons it appears much more likely 4% 52 that y-substitution in allyl b2 Roberts Young and Winstein J. Amer. Chem. SOC. 1942 64 2157 ; Young and 63 Andrews ibid. 1946 68 2684 ; 1947 69 3062 ; Andrews and Linden ibid. p. :tE. R. H. Jones Lacey and Smith J. 1946 940. o w w u -d Sultanbawa Nature 1949 163 997; J. 1949 3089. Andrews ibid. 1944 66 421 ; Kepner Winstein and Young ibid. 1949 71 115.2091. FF 422 QUARTERLY REVIEWS halides is analogous to intermolecular isomeric rearrangement in ally1 alcohols and esters (reactions 4 and 6b) and usually involves attack by a neutral reagent at C(,,). In alcohols and esters this type of substitution is successful under acidic conditions only when the carbon-oxygen link is weakened by the formation of the oxonium ion whereas fission of the carbon-halogen bond occurs sufficiently readily in neutral solution. Conclusive evidence for this mode of reaction has been obtained by Kepner Winstein and Young,52 who found that the reaction of 1 -ethylally1 chloride with diethyl sodiomalonate in ethanol is strictly of the second order i.e. bimolecular and yet gives 23% of the 3-ethylallyl derivative. As would be expected the effectiveness of neutral reagents for y-attack increases with their basicity in the order EtOH < H20 and NH < NH,Et < NHEt2.54 Pentad Anionotropy Fiveucarbon Anionotropy.-It has already been mentioned (p.415) that asymmetrical butadienylcarbinols formed by three-carbon oxotropy of alkenylallyl alcohols can subsequently undergo further five-carbon oxo- tropy.21 56 If saturated (e.g. methyl) substituents only are present the stability of the two conjugated isomers differs only by the relatively small hyperconjugation energy of the groups directly attached to the butadiene system and such rearrangements like the corresponding three-carbon rearrangements are therefore reversible. If on the other hand an un- saturated substituent is present in one of the terminal positions the equilibrium is displaced entirely in favour of the triply conjugated deriva- tive.The first example of this type was described by Heilbron Jones and McCombie,57 who showed that 1-ethynylpentadienyl carbinol (XXXIII ; X = CHiC R = Me) obtained by condensing sorbaldehyde with sodium acetylide is readily converted into octa-3 5-dien-7-yn-2-01 (XXXIV ; X = CHiC R = Me) on treatment with dilute mineral acids. Analogous rearrangements have been carried out with several other substituted acetyl- acids HO*CHX-CH :CHCH :CHR + CHX :CH*CH :CH*CHR*OH (XXXIII. ) (XXXIV.) enic pentadienylcarbinols 58 as well as with butadienylphenylcarbinol (XXXIII ; X = Ph R = H) 59 and pentadienylphenylcarbinol (XXXIII ; X == Ph R = Me).6* Such poly-carbon rearrangements might conceivably take place by two consecutive allylic rearrangements and it has been claimed 56 that the partly rearranged styrylvinylcarbinol CHPh:CH*CH( OH)*CH:CH is an isolatable intermediate in the isomerisation of butadienylphenylcarbinol.Nazarov and Fischer Bull. Acad. Sci. U.R.X.S. CI. Sci. Chim. 1945 631 ; 1948 311 427; 1949 112. 57 J. 1944 134. 68 Sir Ian Heilbron E. R. H. Jones and Richardson J . 1949 287 ; idem Lewis and Weedon J. 1949 742 2023. 6s Salkind and Kulikov J . @en. Chem. Russia 1945 15 643 ; Woods and Sanders J . Amer. Chem. Soc. 1947 69 2926 ; Nazarov and Fischer Bull. Acad. Sci. U.R.S.S. CE. Sci. Chim. 1948 436. 6o Barany Braude and Coles unpublished. BRAUDE ANIONOTROPY 423 Other investigations 22 60 have failed however to reveal any measurable concentration of partly rearranged intermediates and it is very probable that the double bond migration takes place in one step by an intermolecular mechanism analogous to (4) (p.409). A few isolated examples of five-carbon anionotropy in cyclic systems are also known. 1 Thus 10- benz ylidene-9 LO-dih ydro-g - phenylant hran- 9- ol (XXXV) obtained by condensing benzylideneanthrone with phenylmag- nesium bromide undergoes rearrangement to the isomer (XXXVI). A 0 Ph OH Ph (yJJ MgPhBr* & H*SOlt (* II 1 CHPh*OH II CHPh CHPh (XXXV.) (XXXVI.) somewhat similar transformation occurs with tertiary fury1 carbinols. 62 On treatment with hydrogen chloride in ethanol furyldiphenylcarbinol (XXXVII) is converted into 2-diphenylmethylene-5-ethoxy-2 5-dihydro- furan which undergoes ring opening to give ethyl 99-diphenyl-lsevulate (XXXVIII) ___ 10 (,,!bHPhz EtO-(0)=CPha + CHPh2*CO*CH2*CH2*C0 2Et OH (XXXVII.) (XxxTrIII.) These rearrangements are of interest in that they involve the participation of double bonds belonging to an aromatic or a heterocyclic ring. Replacement reactions of butadienyl halides can give rise to non-isomeric five- carbon anionotropy analogous to the three- carbon rearrangements accompanying replacement reactions in allylic halides. Thus the action of aqueous potassium cyanide on sorbyl chloride gives l-methylpenta-2 4- dienyl cyanide 63 KCN H 8 0 CHMe :CH*CH :CH-CH2C1 CN-CHMeCH :CH*CH:CH + HO,C*CHMe*CH :CH*CH :CH2 Other Anionotropic Pentad Systems.-Whereas triad and higher-mem- bered prototropic systems involving nitrogen atoms are well known none of the corresponding open-chain oxotropic systems has yet been described.It is very probable however that the acid-catalysed rearrangement of phenylhydroxylamine and allied reactions proceed by way of successive 61 J. W. Cook J. 1928 2798 ; Julian Cole Diemer and Schafer J. Amer. Chem. SOC. 1949 71 2058; Badger and Peace J. 1950 2311 2314. 6* Ushskov and Kutscherov J. Gem. Ckm. Russia 1944 14 1073 1080 1087. 8s Reichstein and Trivelli Helv. Chim. Acta 1932 15 254. F F* 424 QUARTERLY REVIEWS pentad anionotropic and prototropic changes in the C=C-C==C-N skele- ton as originally suggested by Bamberger.64 The rearrangement may occur in two steps with the migrating group first becoming attached to the ortho- carbon atom ; in the presence of certain p-substituents the rearrangement in fact stops at this point. On the other hand in the case of p-tolyl- hydroxylamine the p-methyl substituent does not prevent para-migration which is followed by migration of the methyl group to the meta-position or by loss of water to give the p-methyleneimine.H H H H OH,+ OH I I 0 H+ d$+()-&() I I1 I NH*OH NH,+*OH NH*OH,+ NH NHZ A different mechanism has been put forward by D e ~ a r ~ ~ who has postulated that the phenylhydroxylamine and similar rearrangements are entirely cationotropic and involve migration of .the hydroxyl or similar group as a positive ion which remains attached to the rest of the molecule in the form of a 3t-complex. The main argument in favour of this theory is the unproven assumption that such reactions are mainly intramolecular. Even if this assumption should prove to be correct it would not necessarily conflict with the anionotropic mechanism and any final decision will have to await kinetic investigations of these reactions.Heptad and Higher Anionotropic Systems By condensing sodium acetylide with octatrienal Heilbron Jones and McCombie 57 obtained the acetylenic carbinol (XXXIX) which readily undergoes heptad oxotropy to give the fully conjugated trienyne (XL). This type of reaction has recently been employed in the synthesis of vitamin A homologues from 1 -ethynylcycZohexenes 66 acids CH i CGH( OH)*[CH :CHI ,*Me ___+ CH CfCH :CHI ,*CHMe*OH (XXXIX.) (XL. 1 The corresponding glycol (XLI ; n = 3) obtained from acetylenebismag- nesium bromide similarly undergoes the interesting double heptad rearrange- ment to the hexenyne glycol (XLII).67 This elegant method of polyene synthesis has also been applied to the two lower vinylogues (XLI ; n = 1 and 2) and to /?-ionone analogues,6s with a view to the total synthesis of carotenoids (which has since been achieved) 68 (XLI.) acids + Me fCH :CHIn*CH(OH) *C i C*CH( OH) *[CH :CH],*Me HO*CHMe*[CH:CH ],*CiC*[CH :CH],*CHMe*OH (XLII.) 64 Ber.1900 33 3600; 1901 34 61. 67 Heilbron E. R. H. Jones and Raphael J. 1943 268; 1944 136. 881nhoffsn Pommer and Bohlmann Annulen 1948 561 26; 1950 569 237. 65 J. 1946 406. Cheeseman Sir Ian Heilbron E. R. H. Jones and Weedon J. 1949 3120. BRAUDE ANIONOTROPY 425 An example of undecad anionotropy involving simultaneous migration of five double bonds probably occurs in the conversion of vitamin A (XLIII) into anhydrovitamin A (XLV) by hydrochloric acid.gs The structure of anhydrovitamin A has not however been rigidly proved and the inter- mediate (XLIV) has not been isolated. Me Me H :CH*CMe :CH*CH :CH*CMe :CH*CH,*OH (XLIII.) CH*CH :CMe*CH :CH*CH :CMe*CH :CH I - - H a 0 (XLIV.) * Me Me flH*CH :CMe*CH :CH*CH :CMe*CH :CH a (XLV.) Shantz Embree and Cawley J. Amer. Chem SOC. 1943 65 901 ; Meunier Compt. rend. 1943 216 907; 1948 226 128.

ISSN:0009-2681    

DOI:10.1039/QR9500400404

出版商:RSC    

年代:1950

数据来源: RSC

摘要:

ERRATA Page 282 366 368 369 376 1948 Vol. 11 No. 4 Line 24 23 23 294 Eqn. and 295 Eqn. for " 4.4 " read " 4.0 '7. In the top central formula for " CH " read " CN 7'* for (' 3-phenyl-1 2 3 4-tetrahydroiso- quinolines and 4-phenyl-1 2 3 4- tetrahydroisoquinolines " read " 4-phenyldecahydroquinolines and a 4-phenyldecahydroisoquinoline ". for " 2-dimethylamin0propy1~' read '' 2-dimethylaminoethyl ". In formula (XLVIII) the long horizontal bond should join the two long vertical bonds as in formula (XLVII). 1949 Vol. 111 No. 4 426

ISSN:0009-2681    

DOI:10.1039/QR9500400426

出版商:RSC    

年代:1950

数据来源: RSC