摘要:

QUARTERLY REVIEWS FOREWORD TO QUARTERLY REVIEWS VOLUME X No. 3 AN exchange of review articles has been arranged between the U.S.S.R. Academy of Sciences and the Chemical Society the articles from the U.S.S.R. to be published in QuarterZy Reviews and the British articles published in Russian journals. The four art,icles in this issue are published under this arrangement. They are printed as received except for occasional alteration of nomenclature and symbolism the subjects and authors having been selected by the U.S.S.R. Academy of Sciences. R 259

ISSN:0009-2681    

DOI:10.1039/QR9561000259

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

OXIDATIVE-HYDROLYTIC SPWTTING OF CARBON-CARBON BONDS OF ORGANIC MOLECULES By M. M. SHEMYAKIN and L. A. SHCHUKINA (U.S.S.R. ACADEMY OF SCIENCES Moscow) OXIDATIVE processes accompanied by splitting of the bonds between carbon atoms are one of the commonest types of chemical and biochemical trans- formation of organic molecules. In considering such changes the final result and the intermediate stages of the reactions are usually treated as being due to the action of only one factor on the molecule 'uz'x. the oxidant. It is often neglected that in all cases where the reactions take place in aqueous solution especially in the presence of alkalis or acids another factor may exert a considerable influence namely the hydrolysing action of the medium. The effect of this factor should -be most pronounced a t intermediate stages of oxidative processes since the carbon-carbon bonds of partly oxidised molecules are especially subject to hydrolytic splitting.Hence if oxidation is carried out under conditions where such action is not excluded the occurrence of hydrolytic molecular changes should be taken into account along with changes due to oxidation. This important problem which was neglected for a long time attracted our attention more than 10 years ago. In 1944 together with a number of collaborators (Y. B. Shvetsov D. P. Vitkovsky A. S. Khokhlov and others) we began a systematic study of various reactions induced by the simultaneous or alternate action of oxidants and hydrolysing agents on organic molecules. A comparison of the results of these studies with data published in the literature led to the conclusion that when the oxidative cleavage of carbon- carbon bonds is brought about in the presence of hydrolysing agents this reaction should be regarded as an oxidative-hydrolytic and not a purely oxidative reaction.Indeed if the initial compounds are subjected to the action of oxidants under conditions excluding or unfavourable to hydrolysis of carbon-carbon bonds the result is most frequently merely the introduc- tion of various oxygen-containing substituents into the molecules without any splitting of carbon-carbon bonds or any changes in the carbon skeleton.* But the accumulation of such oxygen-containing substituents gradually gives rise to structural alterations determining the possibility of hydrolysis of carbon-carbon bonds a t certain levels of oxidation.Thus in many cases the oxidising agents are incapable of splitting carbon-carbon bonds by themselves but predetermine the subsequent cleavage by hydrolysing agents ; that is why changes of this type should be regarded as oxidative- hydrolytic. At present this type of reaction of organic compounds may be * Except)ions may be oxidative reactions taking place by a radical mechanism through stages of peroxide formation as well as reactions resulting in the formation of oxidation produots thermally unstable under the given temperature conditions. 261 262 QUARTERLY REVIEW8 considered as one of the important kinds of oxidative processes inasmuch as the latter are very often brought about under conditions which promote the hydrolytic splitting of mrbon-carbon bonds.An extensive study of oxidative-hydrolytic changes has made it possible gradually to elucidate not only the nature of this phenomenon but some of the laws it obeys as well. The solution of this problem was facilitated by the fact that one of the authors in collaboration with I. A. Red’Bin investigated as early as 1938-1941 the causes and mechanism of purely hydrolytic splitting of carbon-carbon bonds in organic compounds. On the basis of these studies we were able to determine the causes of the influ- ence exerted by preliminary oxidation of the molecules on their subsequent hydrolytic cleavage ; we also elucidated the relation between the degree of oxidation of the molecules and the ease of hydrolysis of their carbon- carbon bonds. This helped to explain the chief directions along which oxidative-hydrolyt,ic changes evolve most frequently as well as to under- stand the subsequent fate of the substances initially formed.All this permitted a new approach to an understanding of the nature and mechanism of many oxidation reactions both those described previously in the litera- ture and those newly studied. Since the oxidative-hydrolytic splitting of carbon-carbon bonds is a result of the simultaneous or alternate action of oxidants and hydrolysing agents on the molecules reactions of this kind can often be studied only by investigating the action of each of these two factors separately. The following procedures were found expedient for the study both of individual particular problems and of the phenomenon as a whole (1) st’epwise oxida- tion of the initial substances in the absence of hydrolysing agents ; (2) in- vestigation of the hydrolytic splitting of carbon-carbon bonds in the absence of oxidants (in particular of atmospheric oxygen) a t each level of oxidation of the original substances ; (3) oxidative-hydrolytic splitting of molecules under very mild conditions in order to induce as little change as possible in the primary fission products which are frequently very susceptible to further alteration.These methods of investigation have made it possible to elucidate many of the questions mentioned above and the following paragraphs of this article are devoted to a discussion of those problems. Hydrolytic Splitting of Carbon-Carbon Bonds The possibility of hydrolytic splitting of certain carbon-carbon bonds has repeatedly been nientioncd in the literature and is commonly known (for a survey of the most iniportaiit papers in this field see refs.1-4). However little was known about the causes and the mechanism of the reaction or of the scope of its occurrence. A special study of these 1 R. C. Fuson and B. A. Bull Chenz. Rev. 1934 15 376. 2 F . Krohnke and W. Heffe Ber. 1937 70 864. 3 M. M. Shemyakin and 1. A. Red’kin Zhur. obshchei Khim. 1941 11 1142. 4M. M. Shemyakin and L. A. Shchukina ibid. 1948 18 1925. SHEMYAKIN AND SHCHUICINA SPLITTING OF CARBON-CARBON BONDS 263 problems 3 5-9 showed that in general such a reaction is possible with all molecules in which an atomic group of the common type (I) is contained or may arise. Given certain structural prerequisites and t,he necessary external conditions such an a.tomic group can undergo splitting according to the following general scheme The range of compounds capable of such changes is very wide.It includes not only many types of substituted alcohols but also various kinds of snbstituted unsaturated compounds carbonyl compounds and other types since the grouping (I) is formed readily as a result of hydration of polarised double bonds ; it also arises by addition of water ammonia or alcohols to the carbonyl group in compounds of the type -COG- / and may be a constituent part of other groups such as HO,C-C- / ‘\ \ * * (2) t H 2 0 :c=c * >C(OH).CH<- >CO + H2C< -co.cc +H20b -C(OH),.C$ _+ -C02H + HCC * (3) H02C.Cf _* COz + HCf The nature of all these changes becmme much easier to understand when we applied E. H. Usherwood’s idea lo concerning the analogy between rever- sible splitting-addition reactions and the corresponding tautomeric changes.,4 comparison,3 based on this analogy of the processes of hydrolysis of carbon-carbon bonds with related tautomeric processes of the keto-enol type (scheme 5 ) gave grounds from which the first of them could be approached from the standpoint of the ionotropy theory developed by C. K. Ingold. This approach proved to be quite justified and very fruitful. Indeed nll the principal theses of the ionotropy theory l1 were found to hold good for the hydrolytic splitting of carbon-carbon bonds. This is borne out by c3xperimental evidence previously published in the literature (see re- views 1-4) by numerous data obtained in the course of our investigations I. A. Red’kin and M.M. Shemyakin Zh!ir. obshchei Khirn. 1941 11 1157. 6 M. M. Shemyakin and I. A. Red’kin zbid. p. 1163. 7 M. M. Shemyakin and N. I. Oranskii ibid. p. 1169. I. A. Red’kin and M. M. Shemyakin ibid. p. 1176. 9 M. M. Shemyakin and N. I. Oranskii ibid. 1943 13 175. loE. H. Usherwood Chem. and Irhd. 1923 1246 J. 1923 123 1717. l1 J . TV. Baker “ Tautomerism ” Routledge London 1934 ; C. K. Ingold “ Struc- tiire and Rlechanism in Organic Chemistry ” Bell London 1963. 264 QUARTERLY REVIEW8 of various hydrolytic 5-g and oxidatlive- hydrolytic (see below) reactions as well as in more recent studies by other authors.12-15 Hydrolytic splitting of carbon-carbon bonds in compounds of type (I) can actually be brought about in all cases where the hydrogen of the hydroxyl group exhibits marked tendency to dissociate owing to the influence of the substituents present in the molecule a s well as under the action of the medium catalysts and temperature and where owing to the same causes the carbon-carbon bond to be split simultaneously becomes sufficiently polarised in the direction shown in scheme 1 .At the present time we can predict to a considerable extent not only the general possibility of hydrolytic splitting of carbon-carbon bonds but also the necessary conditions for this process. This becomes possible because the influence of the nature number and positions of the substituents in the molecules on these processes as well as the effect of the external conditions can be foretold on the basis of a number of earlier derived relation^,^ although the latter are of a qualitative kind.Thus it was established that the presence of electron-donating sub- stituents (such as Me 0 - etc.) on the cc-carbon atom and of electron-acceptor substituents (such as CO,H CHO CCl, NO, R,N+ etc.) on the p-carbon atom of group (I) always promotes hydrolysis. The stronger the electron- donor or electron-acceptor properties of the substituents the more marked is this effect. If however the electronic character of the substituents on the cc- and the P-carbon atom in group (I) is similar ( L e . both substituents are electron-donors or both are electron-acceptors) their influence on the hydrolysis of the carbon-carbon bonds is neutralised and the more this is so the clossr is the magnitude of the polarising effect of the substituents. When the a- and the @-carbon atom carry identical substituents the influence of the latter is completely cancelled and cleavage of the carbon-carbon bond in group (I) becomes very difficult or completely impossible just as when there are no substituents a t all or when they are remote from the a- and the ,!I-carbon atom.With regard to the hydration preceding the hydrolysis of unsaturated or carbonyl compounds (schemes 2 and 3) it should be noted that this reaction usually proceeds much more readily than the splitting of the hydrated molecules. That is why the latter can in many cases be isolated or detected indire~tly.~? 4 9 8 15-18 It is also important that in the case of the addition of water to the double bonds )C=C( or >C-0 the lower t>he electron density a t the carbon atom to which the hydroxyl is to be attached the more readily will the reaction take place.This circumstance may sometimes determine the bond at which the molecule will split ; such 12 A. S. Sultanov Zhur. obshrhei Khim. 1946 16 1835. l3 I. J. Postovskii and A. M. Eidlin ibid. p. 2053. l4 V. V. Feofilaktov and N. K. Semenova ibid. 1963 23 460. It. G. Pearson anti l3. A. Mayerle J . Amer. Chem. SOC. 1951 73 926 ; R. G. 1 6 Yu. B. Shvetsov and M. M. Shemyakin Zhur. obshchei Khim. 1949 19 450. 1 7 Yu. R. Shvetsov I. A. Retl’kin and M. M. Shemyakin ibid. 1951 21 339. l a L. A. Shchiikina ibid. 1952 22 668. Pearsoii and A. C . Sandy ibid. p. 931. SHEMYAKIN AND SHCHUKINA SPLTTTING O F CARBON-CARBON BONDS 265 for instance is the case with some polycarbonyl compounds; see p. 271 and refs. 4 16 17 19-21. As regards the external conditions the greatest influence on the hydro- lytic splitting of carbon-carbon bonds is exerted by the pH of the solution and the temperature.The reaction is catalysed by both acids and bases but the latter have a much greater effect. The less the course of the hydrolytic process is facilitated by structural factors the higher must be the temperature and the acidity or basicity of the solution. In more favourable cases however simple heating with water is sufficient. Evi- dently the influence of the catalysts is of the same nature in this case as in prototropic changes where the reaction may proceed according to a uni- or a hi-molecular mechanism.I1 Thus the tendency of carbon-carbon bonds to he split hydrolyticslly is inherent in all compounds the molecules of which contain or can easily acquire the prototropic group (I) but such changes can actually take place only if the molecules possess the structural features discussed ahove and under a definite set of external conditions.* It will be appropriate to give several examples in illustration of this.It follows from the above considerations that in compounds (11) and (111) hydrolytic splitting according to scheme 2 is practically impossible because the carbon-carbon bonds to be split iiz these compounds and their hydration products are polarised very little or not a t all. On the other hand in monocarboxylic acids (IV) hydrolysis is possible although it pro- ceeds with great difficulty owing to the comparatively low electron-acceptor capacity of the carboxyl group (the reaction requires prolonged boiling of l9 Yu.B. Shvetsov L. A. Shchukina and M. M. Shemyakin Zhur. obshchei Kkim. 2o A. S. Khokhlov L. A. Shchukina and M. M. Shemyakin ibid. 1951 21 106. 2 1 M. M. Shemyakin L. A. Shchukina Y. B. Shvetsov D. P. Vitkovskii and A. S. Khokhlov ibid. p. 1667. c8 G . A. Holmberg Actn Acad. Aboensis Math. Phys. 1949 16 6 ; 1950 17 1 ; 1952 18 9. 2 3 A. N. Nesmeyanov E. G. Pcrevalova N. A. Vollrenau and I. F. Shalavina Izvest. Akad. Nauli S.S.S.R. Otdel khim. Nauk 1951 692 ; A. N. Nesmeyanov N. A Volkenau and E. G. Perevalove ibid. p. 699. * I t should be noted that Holmberg 22 recently described the hydrolytic splitting of carbon-carbon bonds in several triarylmuthylnielonates and 5-triarylmcthy1barbiti~i.i~ acids the molecules of which do not contain and cannot acqiiire the prototropic group 1949 19 498.(1) Ar3C.CH(CO-) t H30* -+ Ar3C.OH + CHz(CO-)z This observation however does not contradict the above considerations because Nesmeyaiiov and his co-workers z3 showed that in such compounds this carbon-carbon bond is specifically weak. I t is capable of being split and of combining wit,h reagents not only under the action of aqueous solutions of acids but also under the influence of x-arious nther substances. 266 QUARTERLY REVIEWS concentrated alkaline solutions). Hydrolysis occurs readily (upon heating with water) only in dicarboxylic acids (V) where the two carboxyl groups exert a sufficiently strong polarising influence on the bond t o be Similar results were obtained in other cases e.g. in the hydrolytic splitting of symmetrically and asymmetrically substituted nitro~tilbenes.~ The next example shows to what extent the electron-acceptor capacity of certain substituents affects the hydrolytic splitting of carbon-carbon bonds.Iln full conformance with the polarising strength of the substituents hydrolytic splitting according to scheme 2 occurs very readily in nitro- compounds (VI) (usually upon heating with water) less readily in ketones (VII) (sometimes upon heating with water more often with dilute acid or alkali) and with great difficulty in monocarboxylic acids (IV) (see A similar picture is observed in the hydrolytic splitting of analogously substituted carbonyl compounds of the type -CO*CX( (cf. scheme 3) and in the decarboxylation according to scheme 4 of the a-substituted acids HO,C*CX( (X = NO, CN CO CO,H R3N+).1-49 21 Another example illustrating the extent of influence of the nature and number of substituents is the pronounced difference between the conditions for hydrolytic cleavage of substituted alcohols (or unsaturated compounds) according to scheme 2 and the conditions for the hydrolysis of similarly substituted carbonyl compounds according to scheme 3.Compounds of the latter type (e.g. P-keto-acids) are always split incomparably more readily than the former (ap-unsaturated or P-hydroxy-acids) owing to the fact that the hydrated forms of carbonyl compounds contain two easily ionisable hydroxyl groups with high electron-donating capacity.3~ l5 Finally several cases should be considered which are directly connected with the problem of oxidative- hydrolytic changes in organic compounds.It is well known that the oxidation of the latter is very often accompanied by the formation of hydroxyl and carbonyl groups in the molecules result- ing in cc- and B-ketol as well as in a- and P-diketone structures. Hence it is important to know the tendency of the carbon-carbon bonds in such structures to hydrolyse in order to foresee the direction in which the molecules are apt to split if they contain both these groups. Although it has long been established that all the structures mentioned are capable of being split under the action of hydrolysing agents the comparative readiness of their splitting has not yet been investigated. It might be expected that p-ketols and p-diketones (or more exactly the hydration products of p-diketones) should undergo hydrolytic splitting much more readily than a-ketols and cc-diketones owing to the hyperconjugation of 0- and n-bonds in the former.This conjecture was confirmed by an ex- perimental investigation carried out with compounds containing jointly the types of structures to be compared. It was shown in a number of SHEMYAKIN AND SHCHUKINA SPLITTING OF CARBON-CARBON BONDS 267 examples 16-199 24-29 in triketones (VIII) containing both ana- and a 6-di- ketone group and in hydroxy-diketones (IX) containing an cc-ketol and a 8-di- ketone group that only the P-diketone group undergoes hydrolytic splitting. \ I \I -CO.CO.yCO- - -CO.C-C-$- -+ -CO.CO,H + HC.CO- p / A J (m) H - 0 OH 03 The above conception gives a uniform standpoint! for interpretation of the hydrolysis (as well as alcoholysis and ammonolysis) of carbon-carbon bonds in a wide range of organic compounds of various types and makes it possible to predict approximately the conditions and trends not only of hydrolytic but also of oxidative-hydrolytic transformations of molecules.The latter problem will be dealt with in the following section of this Review. Oxidative-hydrolytic Splitting of Carbon-Carbon Bonds As mentioned at the beginning of this Review the action of oxidants on organic compounds under conditions unfavourable for hydrolysis of carbon-carbon bonds is often limited to the introduction of various oxygen- containing groups into the molecule. Thus the oxidation of a number of olefins and several quinones under the specified conditions with chromic acid -per-acids hydrogen peroxide oxygen and other oxidants goes no further than the formation of the ~ x i d e s ~ O - ~ ~ while the oxidation of poly- hydroxy-quinones by lead tetra-acetate nitric acid etc.results only in the formation of cyclic polycarbonyl compounds. But if these oxidative reactions are performed in the presence of hydrolysing reagents or when the previously oxidised molecules are acted upon by hydrolysing reagents (as carbon-carbon bonds are easily split in a number of cases. 1948 18 2121. for instance in the case of cyclic polycarbonyl compounds 4 19-21 1 the 2 4 L. A. Shchukina A. P. Kondratieva and M. M. Shemyakin Zhur. obshchei Khim. 2 5 L. A. Shchukina and M. M. Shemyakin ibid. 1949 19 193. 26 D. P. Vitkovskii and M. M. Shemyakin ibid. 1951 21 540. 27 Idem ibid. p. 547. 28 L. A. Shchukina A. S. Khokhlov and M. M. Shemyakin ibid.p. 908. 29 M. M. Shemyakin and L. A. Shchukina Collection of papers “ Problems of Chemical Kinetics Catalysis and Reactivity ” U.S .S.R. Academy of Sciences MOSCOW 1955 p. 757. 30 A. Byers and W. J. Hickinbottom Nature 1947 160 402 ; J. 1948 285 1328 1331 1334 ; W. J. Hickinbottom and D. G. Wood Nature 1951 168 33 ; J. 1951 1600 ; 1953 1906 ; D. P. Archer and W. J. Hickinbottom J. 1954 4197 ; W. J. Hickinbottom D. R. Hogg D. Peters and D. G. Wood J. 1954 4400; W. J. Hickin- bottom D. Peters and D. G. Wood J. 1955 1360. s1 E. J. Gasson A. F. Millidge G. R. Primavesi W. Webster and D. P. Young J . 1954 2161 ; E. J. Gasson A. R. Graham -4. F. Millidge J. K. M. Robson W. Webster A. M. Wild and D. P. Young J. 1954 2170 ; A. R. Graham A. F. Millidge and D. P. Young J. 1954 2180 ; G.E. Hawkins J. 1955 3288. 32 L. A. Shchukina E. I. Vinogradova and M. M. Shemyakin Zhur. ob,?hchei Khim. 1951 21 1661. 268 QUARTERLY REVIEWS A special investigation of this question has been carried out mainly with carbocyclic compounds (see below). The results showed that the action of oxidants on organic molecules very often facilitates hydrolysis of the carbon-carbon bonds a t definite levels of oxidation of the molecules and that on the other hand partially oxidised molecules sometimes acquire an increased tendency to further oxidation after hydrolytic splitting. Thus in many cases the splitting of carlion-carbon bonds becomes possible only as a resnlt of the simultaneous or alternate action of oxidants and hydrolysing agents. For the understanding of the nature of such oxidative-hydrolytic changes in organic compounds it is important first to ascertain (1) why preliminary oxidation of the molecules may influence the subsequent hydrolysis of the carbon-carbon bonds and (2) what relation exists between the degree of oxidation of the molecules and the capacity of their carbon-carbon bonds to undergo hydrolytic splitting.The answer to both these questions follows logically from the conception expounded in the previous section of this Review. Indeed the oxygen- containing substituents 00 -OH )CO etc.) formed in the molecules when they are oxidised can not only take part in the formation of proto- tropic groups of type (I) prerequisite for the hydrolysis (see p. 263) but can also influence the polarisation of the carbon-carbon bonds in these groups and hence their tendency to hydrolytic splitting.Thus it be- comes clear why the action of oxidants on organic compounds influences the hydrolysis of their carbon-carbon bonds very greatly during definite stages of oxidation of these compounds. An explanation is also obtained for the relation between the level of oxidation of the molecules and the capacity of their carbon-carbon bonds for hydrolytic splitting. This relation may be formulated as follows the more strongly the oxygen-containing substituents formed in the molecules under the action of the oxidants polarise the carbon-carbon bonds subject to hydrolytic splitting the more readily will the latter process take place [provided of course that the carbon- carbon bonds undergo polarisation in the direction indicated in scheme 1 for compounds of type (I)].Indeed the results of investigations 4 3 1 6 - 2 1 1 2 4 - - 2 9 9 32-46 of the conditions 33 L. A. Shchukina A. P. Kondratieva and M. M. Shemyakin Zhur. obshchei Khisii. 3 4 Idem ibid. 1949 19 183. s5 Idem ibid. p. 468. aG Ye. I. Vinogradova Yu. R. Shvetsov and M. M. Shemyakin ibid. p. 507. 37 L. A. Shchukina Yu. B. Shvetsov and M. M. Shemyakin ibid. 1951 21 346. 38 L. A. Shchukina A. S. Khokhlov and M. M. Shemyakin ibid. p. 917. 39 D. P. Vitkovskii and M. M. Shemyakin ibid. p. 1033. 40 M. M. Shemyakin 1). A. Bochvar and L. A. Shchiikina ibid. 1952 22 439. 4 1 0. M. Shemyakina B. M. Bogoslovskii and M. M. Shemyakin ibid. p. 675. 4 3 D. P. Vitkovskii and M. M. Shemyakin ibid. p. G79. 4 3 L. A. Shchukina and E. P. Syomkin ibid. 1956 in the press. 4 4 L.A. Shchukina ibid. 4 5 L. A. Shchukina and M. M. Shemyakin ibid. 4 6 0. M. Shemyakina R. M. Bogoslovskii arid M. M. Shemyakin ibid. 1948 18 1945. SHEMYAKIN AND SBCHUKINA SPLITTING OF CARBON-CARBON BONDS 269 and nature of hydrolytic splitting of carloocyclic compounds a t various stages of oxidation show that this process is in direct relation with the polarisation of the carbon-carbon bonds subject to splitting. This polarisa- tion in its turn depends on the nature number and positions of the sub- stituents either contained in the original molecules or formed in them under the action of the oxidants. Thus oxidation of aromatic hydrocarbons and their derivatives to p-quinones fails to give rise to the peculiar structural features in the mole- cules required to promote to a sufficient degree the splitting of the ring systems in the presence of hydrolysing agents.The reason for this lies in the fact that the carbonyl groups of p-quinones (X) are unable to polarise the bonds between the carbon atoms of the ring. That is why e.g. 1 4- naphthaquinone cannot be split hydrolytically as such but only after preliminary oxidation. 33 2-Methyl- and 2-ethyl-1 4-naphthaquinone behave in a similar manner 33 owing to the fact that alkyl radicals are but weak polarisers.* Even further oxidation of these p-quinones to the oxides (XI) and hydration of the latter to glycols (XII) do not yet result in compounds with a tendency to hydrolysis in the presence of hydrolysing agents because the introduction of an oxide-oxygen atom or two hydroxyl groups at positions 2 and 3 of the quinone molecules cannot promote polarisation of the bond between these carbon atoms.This is why the oxides and the corresponding glycols just as the initial quinones do not undergo hydrolytic splitting unless their molecules are preliminarily altered.z5y 3 2 9 3 4 3 3 8 9 45 0 0 0 An entirely different result 2 4 7 28 4 3 9 4 7 48 is observed however in the case of hydroxy-quinones (XIII) which may be formed through dehydra- tion of glycols (XII). These hydroxy-quinones can easily be hydrated a t the double bond which is polarised under the influence of the hydroxyl group giving rise to the hydration products (XIV). In the latter the ordinary bond between C(2) and C(3) is also very highly polarised owing to the presence of two easily ionisable hydroxyl groups in position 3.Accordingly hydroxy- quinones (XIII) undergo hydrolytic splitting quite readily when boiled in aqueous solution a t pH 7 yielding compounds of type (XV Depending on the number and position of hydroxyl groups formed in XVI). 47 L. F. Fieser J . Amer. Chem. SOC. 1929 51 940 1896. 48 L. F. Fieser and A. Bander ibid. 1951 73 681. * Failure of alkyl and aralkyl radicals to influence perceptibly the hydrolytic splitting of carbon-carbon bonds and a weak influence in this respect of the my1 radical were observed in a number of cases for instance in 2-substituted 1 4-naphthaquinone 2 3-oxides in the glycols corresponding to these oxides and in 2-substituted 3-hydroxy- 1 4-naphthaquinones ; 2 4 2 6 28 34 3 8 9 4 3 cf. also the conditions of hydrolysis of the corresponding ap-unsaturated nitro-compounds and monocarboxylic acids.39 6 270 the molecule in the course of its oxidation the influences of the hydroxyl groups on the ease of fission of carhon-carbon bonds in the ring may be widely different. In this connection it will be of interest to conipare the properties of glycols of type ( X I ) with those of isomeric compounds of type (XIV) which differ in structure only as regards the position of the hydroxyl groups. It is owing to this difference that compounds of type (XII) as mentioned above are incapable of hydrolytic splitting under con- ditions in which compounds of type (XIV) are split quite readily because only in the latter do the hydroxyl groups polarise the bond between C(e) and C(3). Another instance is the difference in stability to hydrolysis of the ring systems in 2-hydroxy- (XVII ; X = H Y = OH) and 2 3-dihy- droxy-1 4-naphthaquinone (isonaphthazarin) (XVII ; X = Y = OH).The latter quinone is incomparably more resist,ant to hydrolytic splitting than the former since the 2- and 3-hydroxyl groups in the isonaphthazarin molecule are incapable of polarising the bond between C(*) and C(3).209 28 The bond between C(2) and C(3) in monohydroxy-quinones (XIII) can be depolarised not only by introducing a hydroxyl group in position 2 but also by int)roducing certain other substituents of a similar electronic nature. Indeed 2-amino-3-hydroxy-1 4-naphthaquinone (XVII ; X = NH, Y -= OH) is no less stable to hydrolysing agents than isonaphthazarin (XVIII ; X = Y = OH) ; both these quinones remain almost unchanged when boiled for many hours with 1% sodium hydroxide solution in the absence of atmospheric 39 Another compound which is fairly stable though t o a lesser degree is 1-[3(1)-hydroxy-l 4(3 4)-naphtha- quinon-2-yllpyridinium betaine (XVII ; X = C,H,N+ Y = 0-) ; in this compound the bond between and C(3) is also quite highly depolarised 2G (the pathways of the hydrolytic changes of this betaine are considered on p.276). 0 0 Relations similar to those just described are observed whenever an additional carbonyl group appears in the p-quinone ring owing to oxidation or any other cause. The appearance of such a group usually renders the molecules very unstable t,owards hydrolysing agents. Thus it was shown l69 199 2 7 3 5 5 3 7 9 44 (see also ref. 49) that a great variety of triketones 49 T. Zincke and E.Winzheimer Annulen 1896,290,321 ; T. Zincke and B. Francke ibid. 1896 293 120 ; Ber. 1896 29 965 ; T. Zinoke and C. Gerland Ber. 1887 20 3216; 1888 21 2379. SHEMYAKIN AND SHCHUKINA SPLITTING OF CARBON-CARBON BONDS 271 of the tetralin series of general type (XVJII) not only are readily hydrated a t the double bond of the 3-carbonyl group but also undergo subsequent hydrolytic splitting between positions 2 and 3 of the ring. It is interesting to note in this respect that it is always the 3-carbonyl group and not that in position 1 or 4 that is hydrated in triketones of the tetralin series (XVIII). This is important because in this way the place is determined where the subsequent hydrolytic splitting of the ring will occur. The cause of this regularity lies in the fact that the electron density of carbon atoms in carbonyl groups situated next to the aromatic nucleus is lowered less than in a 3-carbonyl group.It must also be stressed that although the hydrolytic splitting of these triketones occurs very readily the conditions under which this process can be effected also depend to a considerable extent on the nature of the substituents a t position 2. For example owing to the presence of both a pyridiniiim residue and a chlorine atom in the hydrated triketone chloride (XIX; X = C,H,N+ Y = Cl) the 2 3-bond is so highly polarised that it is split in aqueous solution even without heating.27 I n the hydrated triketone series (XIX; X = Me or Ph Y = OH OAc or C1) the same bond is less strongly polarised and hence this bond can undergo hydrolytic splitting only when aqueous solutions of these compounds are heated.169 179 199 259 45 The 2 3-bond should be still less strongly polarised in the hydrated triketones (XIX ; X = H Me CH,Ph Ph etc.Y = H) which are formed by hydration of the corres- ponding hydroxynaphthaquinones (XIII) . Accordingly the hydrated forms of the latter will readily undergo splitting only when boiled in aqueous solution at pH 7 (see p. 269 and refs. 18 24 28 43). Only the hydrated triketone (XIX ; X = Ph Y = H) can be hydrolysed by boiling water ; 43 this is due to the electron-acceptor properties of the phenyl residue. The occurrence of a still greater number of carbonyl groups in molecules of carbocyclic compounds should result in compounds with an especially high t,endency for hydrolytic cleavage. It might be expected for instance that the ring in tetralin tetraketone (XVIII ; XY = 0) will be split very readily because in this case hydration of the carbonyl group in position 2 or 3 causes strong polarisation of the 2 3-bond owing to the presence of two carbonyl groups in positions 3 4 or Z 2.Indeed,20 this tetraketone was found to be split very rapidly in hot aqueous solution ; in solution a t pH 7 this process is almost instantaneous even a t room temperature. Polycarbonyl compounds belonging to other types of csrbocyclic as well as of fatty-aromatic compounds vix. o-pyruvoylphenylglyoxylic acid 35 272 QUARTERLY REVIEWS triquinonyl,50 ninhydrin,20 and anthra-1 4-9 10-diquinone 51 behave in a similar manner. They are exceedingly unstable in the presence of hydrolys- ing agents although in +,he absence of the latter all these compounds are quite stable i From the above-mentiL-ed results the pathways of oxidative-hydrolytic changes in aromatic compounds leading to the splitting of carbon-carbon bonds in the ring become evident.Oxidation of aromatic compounds to p-quinones (X) oxides (XI) and glycols (XII) is not sufficient to induce splitting of the ring systems (see p. 269). Only upon dehydration of glycols (XII) to monohydroxy-quinones (XIII) can there arise the structural features in the molecules which favour hydrolytic splitting of the bonds according to the scheme (XIII)+ (XIV) -+ (XV + XVI). If however for some reason excessively drastic conditions are needed to bring about hydrolytic cleavage of the hydroxy-quinones (XIII) or if these hydroxy- quinones are not formed a t all it becomes necessary to subject the molecule to additional oxidation in order to convert it into more oxidised and more readily hydrolysable compounds to monohydroxy- quinone oxides (XXII) hydroxy-triketones or their hydrates (XVIII and XIX ; X = OH) tetra- ketones (XVIII ; XY = 0) etc.Thus 35 37 44 the oxidative-hydrolytic transformation of hydroxy-quinones (XIII ; X = Me Ph o-tolyl ,B-C,,H,) into the corresponding triketo-acids (XXIII) takes place under the action of cold alkaline potassium permanganate according to the scheme (XIII) -+ (XXII) + (XIXa) -+ (XXa) -+ (XXIII). The oxidative-hydro- lytic cleavage of hydroxy-quinones (XIII ; X = OH NH, C1 etc.) pro- ceeds in a similar way ; 39 4 2 in boiling alkaline aqueous solution in the presence of atmospheric oxygen these compounds are transformed into phthalide-3-carboxylic phthalonic and phthalic acid through the inter- mediate stages (XXII) + (XIXa) -+ (XVIIIa).oxidant ‘0 heating and to other modes of treatment. f- CO*CH(OH)- - ~ C 0 . C O 2 H It is important also that not only the above-mentioned hydroxy-quinones (XIII) but likewise their precursor glycols (XII) often have a very marked 50 R. Nietzki and T. Benckisser Rer. 1886 18 499 1833 ; 1886 19 293; T. Zincke Ber. 1887 20 1265 ; A. Hantzsch Ber. 1888 21 2421. 51 0. Dimroth and E. Schultze AnnuZen 1916 411 345. SHEMYAKIN AND SHCHUKINA SPLITTING OP CARBON-CARBON BONDS 273 tendency to further oxidat'ion owing to their special structura'l features and can readily be oxidised even with such weak oxidants as atmospheric However in certain cases transition to the next degree of oxidation may not result in the formation of a compound readily split under the action of hydrolysing agents.Indeed while the oxidation of %substituted glycols (XI1 ; X = Me or Ph) results in hydrated hydrosy-tri- ketones (XIXn) which are split readily upon mere heating with waCer into compounds of type (XXn + X X I U ) ~ ~ ~ 845 45 oxidation of the unsubstituted glycol (XI1 ; X = H) leads to the forrnat8ion of isonaphthazarin (XVII) which owing to its structural peculiarities shows but little tendency (see p. 270) for hydrolytic splitting.201 38 In tlhis compound hydrolysis can be brought about only by oxidising the molecule still further vix. by trans- forming the isonaphthazarin ( XVII) iiito tetrairetoile (XVlIIa) the ring system of whioh is very unstable in the presence of hydrolysing agents (see p.271 and ref. 20). 3 4 9 3 8 9 45 Q CO CHX.OH c /x HO' 'C02H --+ KC0.C02H * aC/'\OH ( m a ) C,H,(CO,H).C 0 C02H JFJ:; ( m a > (xglz) H \$OH OH 02 *..-(- 0 C*H,(CO,H) 0 (rn) CH.CO2H co 0 0 Crn) ( m a ) C,H/ )o Thus in many cases the greater the degree of oxidation of the molecules the more readily will thc hydrolytic splitting of the carbon-carbon bonds occur. But sometimes the introduced oxygen-conta'ining subst'ituents do not influence the hydrolysis of the bonds at all and in certain cases they even hinder this process ; in such cases further oxidation of the rnolecules usually results at certain stages in the formation of compounds in which carbon- carbon bonds will undergo hydrolytic splitting very readily.Secondary and Side Reactions The oxidative-hydrolytic transformation of organic compounds is often associated with secondary and side reactions since the substances formed during int'ermediate stages of these changes [e.g. types (XI XII XIII XV + XVI XVIII XX $ YXI XXII)] have in a number of cases a great tendency to undergo various changes. A considerable number of secondary and side reactions is usually con- nected with the further action of oxidants and hydrolysing agents upon the primary compounds produced. The pathways of these changes are trivial and need not be considered in detail. They include for in- stance,17* 2 0 ~ 2 4 9 35* 3 7 9 399 the common oxidation processes -C0CH2-+ -CO-CH(0H)- -3 -CO*CO- or -CO*CO,H -+ -CO,H 274 QUARTERLY REVIEWS as well as the hydrolytic cleavage of P-dicarbonyl P-hydroxycarbonyl and other compounds.In the course of oxidative-hydrolytic changes of alkenes cycloalkenes and some of their derivat,ives side reactions sometimes arise owing to the fact that the oxides primarily formed are capable of partial isoinerisation to carbonyl compounds under the influence of acids. I n turn the glycols formed from these oxides have a tendency for dehydration. These side reactions lead along with the formation of normal products of oxidative- hydrolytic cleavage of unsaturated compounds a t their double bonds to the appearance of so-called anomalous oxidation products-unsaturated alco- hols carboxylic acids and other substances retaining the initial number of carbon atoms.30 31 ;c=c< o_ >c-c; H2° >c-$< 0.‘0’ H b OH b30+ 1-H20 :$-CO2H >$-CO- >C (OH) - $= In the course of oxidative-hydrolytic changes of :s+ s 0 0 six-membered carbo- cyclic compounds the most characteristic secondary reactions are those which result in the formation of new five- and six-membered ring com- pounds from acyclic intermediates. The reason for this is that many of the intermediate subst’ances tend to tautomerism of the type (XV + XVI) and the tautomeric form (XVI) being readily subject to dehydration to decarboxylation and a number of other changes.* -CO / \ I HO C02H CO2H am (=I On the other hand if there are structures of type. (XXIV) in the inter- mediate substances six-membered cyclic compounds of t,ype (XXV) may be formed which also are readily subject to various changes (dehydration oxidative decarboxylation e t ~ .) . ~ 199 34-373 44 co1 0 6’’ fi \& HO COZH CO2H 0 cm> ( X X Y ) It is typical that in t’he course of oxidative-hydrolytic changes of organic npounds various secondary and side reactions often proceed simultane- * See ref. 4 ; cf. also more recent investigatioizs (refs. 16 20 24 27 28 36 38 39 42 43 48). YHEMYAKIN AND SHCHUKINA SPLITTING OF CARBON-CARBON BONDS 275 ously in different directions or alternate with one another. By way of illustration a number of reactions are considered below most of which were investigated by isolation or synthesis of the intermediates followed by a study of the conditions and directions of their subsequent changes. Thus it has been established 43 that boiling 2-hydroxy-3-phenyl-1 4- naphthaquinone in the presence of atmospheric oxygen in aqueous solution a t pH about 9 results in the annexed series of hydrolytic and oxidative reactions * II a c > H P h CO'CHZPh CO*COsH C 1-\ HO C02H $? b C02H Transformations of the tetraketone (XVIII) intermediate in the oxidative-hydrolytic cleavage which is formed as an of certain 2-substituted 3-hydroxynaphthaquinones of type (XIII) (see p.272) likewise proceed in several stages. In this case 20$ 39 the annexed sequence of reactions occurs under the influence of atmospheric oxygen and tt boiling solution of sodium hydroxide. 0 9 9 gp - a c > c = * C / \ / \ HO COZH HO H 0 An example of a secondary reaction's resulting in the formation of six- membered carbocyclic compounds is the cyclisation of o-lactylyhenyl- glyoxylic acid one of the hydroxydiketo-acids of type (XXa) which arise in the course of oxidative-hydrolytic changes of hydroxy-quinones or quinone * With 3-hydroxynaphthaquinones carrying H Me or CH2Ph in position 2 the hydro- lytic cleavage of the ring system when effected under similar conditions is not associated with oxidative changes.24~ 281 4 3 5 276 QUARTERLY REVIEWS oxides (MBC pp.277 280). When heated in aqueous alkali in the absence of atmospheric oxygen o-lactoylphenylglyoxylic acid is rapidly transformed into 3 4-dihydroxynaphthoic acid 3 4 Finally an instance of secoiidary reactions resulting in the formation of heterocyclic six-membered rings is the transformation of 1-[ 1 (3)-hydroxy- 3 4( 1 4)-naphthaquinon-%yl]pyridinium betaine into l-hydroxyi.soquino- line-4-carboxylic acid,26 brought a8bout by hot aqueous alkali in the absence of atmospheric oxygen.This change evidently proceeds through the annexed intermediate stages. 0 --+ CO;LH The secondary and side reactions considered above might seem quite different a t first sight. However in many cases their course is so uniform that on the basis of relations previously established the main directions of the processes under investigation and the structure of both the inter- mediates and the final compounds can often be predicted. Examples of Oxidative-hydrolytic Transformations of Organic Compounds Reside the examples given in the previous sections several further cases of oxidative-hydrolytic transforinations of organic compounds should be considered which characterise the complexity and peculiar features of reactions of this kind.Of particular interest from this standpoint is the mechanism of Hooker's reaction which has been specially studied many times. Hooker's re- action,52 first described in 1936 is the transformation of hydroxy-quinones (XIIIa) into hydroxy- quinones (XIIIb) under the action of an alkaline 0 0 5 2 S. Hooker J . Amer. Chem. Soc. 1936 58 1168 1174 1100 ; 8. Hooker and A. Steyermarck ibid. pp. 1179 1198. SHEMYAKIN AND SHCHUKINA SPLITTING 02' CAXEON-CARBON BUNDS 377 solution of potassiuiri permaiigaiiate. * It involves ~iot oiily the shortening of the side chain by one inethylene group but also a shift of the substituent,s in the quinone ring ; the latter was proved by Pieser aiicl his collaborators 5 3 as early as 1936. Both Hooker arid Fieser were right in considering that processes of this bind must be asfiociated with opening and subsequent recyclisation of the quinone ring but t'hey interpreted the mechanism of the intermediat,e stages of reaction diEerently ; 5 4 the hypothetical schemes they proposed in 1936 and 1944 were not verified experimentally.The results of our investigations 49 35 66 of the mechanisni of Hooker's reaction were published in 1946-1949. Our concept of oxidative- hydrolytic trans- formations of organic compouiids enabled us to explore the chief stages of t'his reaction experimentally and tjo elucidate the nature of the process as a whole. It has been established that Hoolcer's reaction proceeds according to the scheme given below. The reaction is a result of the siniultaneous action both of oxidants (potassium permeiiganate and atmospheric oxygen) and of the aqueous alkali on the original molecule and on intermediates ; the alkali brings about not only hydrolysis but also condensation.I n n re-investigation of the mechanism of Hooker's react ion Fieser arid Fiescr suggested in 1948 a scheme for this reaction 5 7 which in its first part is in essential agreement with our scheme but differs from it substantially in its (=I (mb) socond part beginning with the stage of decarboxylation of the hydroxy- acid (XXV) formed as an intermediate. However experimental data 63 L. P. Fieser J. Hartwell and A. Seligman J . Amer. Chem. Xoc. 1936 58 1223. 54 S . Hooker ibid. p. 1174. 65 L. F. Fieser and M. Fieser " Organic Chemistry " Heath Boston 1944 p. 755. 68 L. A. Shchukina Thesis Moscow 1946.57 L. F. Fieser and M. Fieser J . Amer. Chem. SOC. 1948 70 3215. * Beside permanganate the oxidant's may be hydrogen peroxide-.copper sulphate and in some cases apparently atmospheric oxygen.48~ b 2 278 QUARTERLY REVIEWS available to-day 4 35 3 7 44 459 527 537 55-5’ and illustrated in the formula? on page 277 are all in favour of our scheme. The validity of the first stages (XIIIa -+ XXII ++ XIX ++ XXa) is confirmed by the fact that the glycol (XIX ; n = 1 R = H) synthesised for this purpose in a different way was converted under the usual conditions of Hooker’s reaction first into the hydroxydiketo-acid (XXa) and then into the final hydroxy-quinone (XIIIb).lgY 35 Also in two cases it was possible to stop the reaction at the intermediate hydroxydiketo-acids ( X X a ; The next stage (XXa + XXIII) was verified by the same methods.It has been shown 3 5 7 45 that hydroxydiketo-acids (XXa ; n = 1 R = H ; n = 0 R = Ph) can be oxidised by potassium permanganate to the triketo-acids (XXIII) under the conditions of Hooker’s reaction. Further we succeeded in isolating under certain conditions a number of triketo-acids (XXIII ; n = 1 R = H ; n = 0 R = Ph o-tolyl and /l-C,,,H,) directly from the reaction solution.35~ 44 Finally with the individual triketo-acids (XXIII) on hand it was possible to study the last stages of Hooker’s reaction (XXIII+XXV++XIIIb). Here it was shown in the instance of the triketo-acid (XXIII ; n = 1 R = H) 3 4 7 35 that the final hydroxy-quinones (XIIIb) can be obtained only in the presence of oxidants (permanganate atmospheric oxygen) since the intcrmediate hydroxy-acids of type (XXV) can undergo only oxidative decarboxylation.This does not agree with the mechanism suggested by Fieser and Fieser 57 for the final stages of Hooker’s reaction which implies the possibility of decarboxylation of hydroxy-acids (XXV) in the absence of oxidants n = 1 R = CH:CMe and Ph) and to isolate them.44 57 (=)-+I G O H I - fiOb4 --c 6 O H - (m4 HO COzH HOH OH It will be useful to discuss briefly the oxidative-hydrolytic reactions in which the initial or intermediate compounds themselves act as oxidants. Transformations of this kind were noted when certain quinones,28 33 quinone oxides,25 32 34 389 40 nitro- compound^,^^ 4 6 ~ 58 etc. were subjected to the action of hydrolysing agents. All these substances were found capable of manifesting their oxidising properties in the course of t’he reactions.In all these cases the processes of oxidation-reduction and hydrolysis take an especially complicated course usually resulting in a considerable variety of final compounds. By way of example we may mention the behaviour of 2-chloro-3- hydroxy-1 4-na~ht~haquinone (XXVI) when boiled in aqueous alkali in the presence or absence of atmospheric 42 In both cases this quinone undergoes a long sequence of oxidative-hydrolytic changes resulting finally in the formation of phthalide-3-carboxylic7 phthalonic and phthdic 58 F. M. Rowe and S. Ueno J. SOC. Dyers and C01ourist.s~ 1931 47 35 ; F. M. Rowe and I?. H. Jowett ibid. p. 163; F. M. Rowe and C. H. Giles ibid. 1935 51 278. SHEMYAKIN AND SHCIIUKTNA SPLITTING OF CARBON-CARBON BONDS 279 acid (see p.272); in the absence of atmospheric oxygen these changes are possible because they are associated with a parallel series of reductive- hydrolytic reactions (see annexed scheme). A detailed study of these reactions showed that two compounds acted as oxidants in this system namely the initial quinone (XXVI) or rather the tautomeric triketone (XXVII) the chlorine atom of which possesses oxidising properties and intermediate oxoindenecarboxylic acid (XXVIII) which is capable of cffecting dehydrogenation. 0 =@-dLd HO CO2H CO2H CO2H ~xxom It is interesting that the oxidative-hydrolytic cleavage of carbon-carbon lbonds forms the basis for a number of methods of synthesising polyfunc- t ional compounds produced by splitting of the corresponding cyclic com- pounds.As an example we may cite the production of halogeno-acids of type (XXb + XXIb) from 2-substituted 3-hydroxynaphthaquinones (XIII). This reaction has been studied by Fieser and Pieser 57 and simultaneously by us.42 16 27 According to Fieser and Fieser it can be brought about by the action of sodium hypochlorite on the hydroxy-quinones (XIII) in the presence of sodium carbonate which results in the direct formation of halogeno-acids (XXb + X XIb). This procedure is interesting for pre- parative purposes but it throws no light on the mechanism of the reaction. We performed this reaction stepwise by first acting upon the hydroxy- quinones (XIII) with chlorine then exposing the intermediate triketones (XVJII) to the action of water and finally treating the hydrated triketones 280 QUARTERLY REVIEWS (XIX) with aqueous sodium hydroxide or carbonate.In this way we were able to ascertain that in this case just as in Hooker’s reaction the ring is split only after the necessary prototropic structure has been formed in the molecule and substituents have been introduced which are capable of’ polarising to a sufficient extent the bond to be split. It is worthy of note that the intermediate triketones (XVIII) are readily subject not only to hydrolysis but also to ammonolysis being converted in tlhe latter case int,o ainides of chlorodiketo-acids or the corresponding lactlains l 7 0 CO. CHRCl CO-CHR R .“.)-a“ NH2 - IC0.CQ.NH2 + x,,-zH 0 Among other synthetical methods based on reactions of oxidative- hydrolytic or purely hydrolytic cleavage of carbon-carbon bonds mention should be made of two methods for the synthesis of hydroxy-acids of type ( X X n + XXIn) which were published almost simultaneously.According to one of these methods 49 3 4 9 45 acids of this kind can be obtained by boiling nsphthaquinone oxides (XI) with water in the presence of atmospheric oxygen ; this reaction proceeds z5 through the intermediate stages (XI) (XII) (XIX) (XXa) and (XXI ; X = OH). The other method 489 57 consists in treating %substituted 3-hydroxynaphthaquinones (XIII) with hydrogen peroxide in the presence of sodium carbonate and is superior to the former as a preparative procedure.45 Its mechanism is identical 37 with the first three stages of Hooker’s reaction (see p. 277). Hooker’s reaction was also utjlised recently 3 5 9 ** for the preparation of trikoto-acids (XXIII) from 2-substi tuted 3-hydroxynaphthaquinones (XIII) according to the scheme (XIIIa) -+ (XXII) -+ (XIX) + (XXa) -+ (XXIII).This method is especially suitable when the substituents in position 2 of the hydroxy-yujnone do not contain a methylene group (e.g. XIIfa ; n = 0 R = Ar) because the reaction cannot then go beyond formation of t,he desired trjketo-acids. In other cases [e.g. with (XIIIa n -=- 2 2 etc.)l special precautions must be taken and even then it is some- times [e.g. with (XIITa ; n = 1 R = Ph)] impossible to stop the prooess at the desired intermediate stage. Mention should also be made of a convenient method for the synthesis of keto-acids of type (XV +- XVI) by hydrolytic cleavage of 2-alkyl(or aralkyl or aryl)-3-hydraxy-quinones (XTTI) of the benzene naphthalene and phenanthrene series,l*> 2 4 3 28 439 479 4a according to the scheme (XIII) -+ Various points of view concerning the conditions and mechanism of this reaction have been set forth.On the one hand we ~uggested,~ 24 as early as 1948 our usual scheme of hydrolytic splitting of ca,rbon-carbon bonds. According to this the hydroxy-quinones should undergo hydration more easily than fission ; in contrast t o the fission hydration does not require an alkaline medium. On the other hand Fieser and Bander in 1951 came to ( ~ 1 1 7 ) -+ (XV + XVI). SHEMYAKIN AND SHCHUKINA SPLITTING OF CARRON-CARBON BONDS 281 the conclusion 4* that only the hydroxy-quinone anion can be hydrated an alkaline medium being required for this purpose whereas the subsequent hydrolytic splitting (or rearrarigeinent of the benzilic acid type) can 1Je brought about without the influence of alkali.However a further investiga- tion of this reaction for the case of 2-hydroxy-5-methyl-p-benzoquinone l8 confirmed the validity of our viewpoint. With this quinone we were able to separate the process into two steps and to isolate the hydrated hydroxy- quinone (XIV) the formation of which is much easier than its cleavage hydration can be brought about in neutral solution while the hyclrolyfic splitting is possible only in alkali. In a number of cases it is advisable to carry out the hydrolytic splitting of the hydroxyquinones (XIII) by boiling them in aqueous buffer solutions a t pH values between 7.5 and 9.5. Under such conditions an equilibrium is usually set up between the resultant keto-acids (XV + XVI) and their dehydration products i.e.compounds of type (XXIX) ; the position of this equilibrium depends very markedly on the structure of the compounds produced.Z4 288 43 C ‘CO CO2H I $! CO. CHzR C Kc>CHR * X;CR / \ HO CO2H COpH (xn) It is noteworthy that the azo-compounds (whose tautomeric forms are hydroxynaphthaquinone arylhydrazones) are subject in some cases to hydrolytic cleavage under the same conditions as the hydrosynaphtha- quinones themselves. The azo-compounds are first split into arylhydra- zine and hydroxynaphthaquinone which then undergo tlhe subsequent changes.41 46 To conclude we shall mention a few additional problems in the solution of which the application of our conceptions concerning hydrolytic and oxidative-hydrolytic changes of organic compounds has played a part.These conceptions proved useful in developing the theory of amino-acid metabolism catalysed by pyridoxal enzymes published by Braunshtein and one of us 59 in 1952-1953 (similar ideas were developed in 1954 by Snell and his collaborators 60). The conceptions were utilised recently in elucidating pathways of biogenesis of branched-chain carbohydrates and the mechanisms of some of their transformations 61 and also in explaining the hydrolytic 59 A. Ye. Braunshtein and M. M. Shemyakin Dokllxdy A k a d . Nazik S.S.S.R. 1952 6o D. E. Metzler M. Ikawa and E. E. Snell J. Amw. Chent. SOC. 1954 76 648. 61 M. M. Shemyakin A. X. Khokhlov and M. N. Kolosov Dolcludy A k d Nauk 85 1115 ; Biokhirniya 1953 18 393. X.S.S.R. 1952 85 1301. 282 QTJARTERLY REVIEWS splitting of certain antibiotics (actidione citrinin 62) ; they were also applied to the study of hydrolytic transformations of compounds such as aurins,13 azo-substituted keto-acids l4 and others.12 It can be expected that our conception concerning the essential mechanism of oxidative-hydrolytic pro- cesses will prove helpful in clarifying the nature of certain other types of hydrolytic and oxidative transformations.This applies in particular to some reactions of splitting of different di- and tri-hydroxyanthraquinones 63 and of purpurogallin 64-reactions which have already been described but have not been studied sufficiently. G 2 &I. M. Shemyakin and A. S. Rhokhlov " The Chemistry of Antibiotic Substances " Moscow 1953 pp. 48 192. c3 E. Bamberger and A. Praetorius Monatsh. 1901 22 587 ; R. Scholl P. Dahll and E. Hansgirg Ber. 1923 56 2548 ; R. Scholl and A. Zinke Ber. 1918 51 1419; 1919 52 1142 ; 0. Dimroth and E. Schultze Annalen 1916 411 339 ; R . Scholl and P. Dahll Ber. 1924 5'7 80 ; K. Fries and E. Auffenberg Ber. 1920 53 23. 6 4 H. Wichelhaus Ber. 1872 5 846; R. D. Haworth B. P. Moore and P. L. Pauson J. 1948 1045 ; D. R. Haworth and J. D. Hobson J. 1951 561 ; W. D. Crow and R. D. Haworth J. 1951 1325; P. L. Pauson Chem. Rev. 1955 55 110.

ISSN:0009-2681    

DOI:10.1039/QR9561000261

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

LATTICE ENERGY OF IONIC CRYSTALS By A. F. KAPUSTINSKII (MEMBER-CORRESPONDENT OF THE U.S.S.R. ACADEMY OF SCIENCES Moscow) THE lattice energies of ionic crystals play a considerable role in diverse chemical processes and an extensive survey could be made of the achieve- ments in various directions by many researchers in this field. The present Review however pursues a much narrower aim to summarise chiefly some investigations carried out by the Reviewer and his colleagues in devel- oping various methods of evaluating lattice energies as well as analysing and generalising the results obtained by these methods. Crystal lattice energy and its calculation by means of ionic charges and radii The concept of crystal lattice energy 77 which as is known equals the heat of dissociation of one mole of solid into its structural components became widespread after Max Born first proposed the following equat,ion for the energy of the lattice of ionic crystals where ql q2 and e are the ionic and electronic charges N is Avogadro's constant r is the inter-ionic distance m is the electronic shells repulsion exponent and A is the structural coefficient (Madelung constant).Con- sidering the quantum-mechanical character of the repulsion forces acting between the electronic shells of ions Born and Joseph Mayer later im- proved eqmtion (1 as follows introducing in it the constant p which characterises these forces and which €or most crystals may be regarded a,s pra,ctically constant and having the value 0.345. For all the significance of this achievement which is an advance in the theory of solids the sphere of the calculations is somewhat limited by inadequate knowledge of the structure and parameters of crystals per- mitting the application of equations (1) and (2) only to solids previously investigated by X-rays for the establishment of r and A .In general chemistry and also in geochemistry however it was desirable to extend t>he sphere of the calculations of U by some simpler method than the use of equations (1) and (2) but without any essential loss of accuracy. M. Born Verhandl. deut. phys. Ges. 1918 20 202 ; 1919 21 13. M. Born and J. Mayer 2. P h y ~ i k 1932 75 1. 283 284 QUARTERLY REVIEWS For the solution of this probleni the Reviewer used his empirically found parallelism connecting the differences between A and the Madelung con- stant for crystals of the rock-salt type and the differences between r and the sun1 of the ionic radii of the given crystal when referred to the co- ordination number 6.This makes it possible to reconstruct imaginatively so to speak an " iso-energetic " rebinding of a crystal of any given type into a rock-salt lattice structure which can be easily computed since thanks to V. Goldschmidt the ionic radii (zc and zA) are well known for most elements. The aforesaid permits t'he following transforination of equation (1 ) The number of ions per mole of a crystal equals the product of Avogadro's constant and the number of ions in the molecule NZn. Let us rewrite eqn. (1) so as to separate in the right-hand part of the equation the work of removing one ion (cation or anion) from the lattioe and then multiply and divide this part by gNXn.We denohe the structural coefficient referred to one ion cc = A/(Cn/2) or T i = =n -.Cc- w 1 h 2 ( l - ;) 2 r Since the changes of ct are equal to the changes in the interionic distances it is obvioixfi that we obtain the same value for U by calculating it with the aid of r derived by X-ray measurements and the structural factor corres- ponding to the given lattioe and by taking the Goldschmidt ionic radii referred to t,he co-ordination number 6 (instead of r = zc + zA) together with the structural coefficient of rock-salt type lattices cc = 1-745. In the above equation m varies little around the mean value of 9 and the adoption of this average value little affects the accuracy of the U calculations. Since Ne2 = 329.7 kcal. per A we as a result obtain equation (3) in which the numerical factors are grouped in one coefficient of proportionality This permits us to calculate the energy of any type of lattice with the aid of charges and radii i .e . with the aid of well-known magnitudes. A defect of this formula is that it is associated with antiquated concepts of the character of repulsion forces. We therefore later similarly trans- formed the quantum-meohanical equation (2) for lattice energy. This equation which has been described for example by Partington 4 yields U r= 25G-1~n~,~2/(zc + zA) . * ( 3 ) where 287.2 is the numerical factor summing all constants. Such is the general expression. It is in accord with modern concepts of the nature of the solid state and a t the same time permits evaluation in advance of the lattice energies of avy ionic crystals not yet investigated by X-ray measure- ments i .e . with r and the structural factors A still unknown experimentally. A. Kapustinskii 2. phys. Chem. 1933 $38 B 267. J. R. Partington " An Advanced Treatise on Physical Chemistry " Vol. 111 Longmans London 1952 p. 383 ; A. Kapustinskii Acta Physicochim. 1943 18 No. 5 370. RAPUSTINSKTI LATTICE ENERGY OF IONIC! CR,YSTALS 285 Yet it is approximately as accurate as Born and Mayer’s equation and there- fore opens up some wider prospect’s for extending the evaluation of lattice energy to vwious scientific investigations. ‘‘ Thermal ’’ (“ thermochemical ”) ionic radii As we shall show in greater detail later to verify the accuracy of equations (3) and (4) we have compared the results of calculations with experimental results (with the so-called “ experimental ” evaluations of lattice energy) obtained hy summing the thermochernical values already known from the Rorn-Haher cycle.I n the Goldschmidt system when such data were available and the radius of one of the ions was lacking it was possible t o “ reverse ” the calculation and thus by using eqn. (3) or (4) obtain from the thermal data the radius of a simple QT complex i0n.5 At the Reviewer’s suggestion such Values have been named “ thermal ” or “ thermochemical ” ionic radii. As an example we may cite the calculation for a tin ion Sn2f lacking from the Goldschmidt table. Computed in three separate ways namely from “ experimental ” energies of stannous sulphide oxide and chloride lattices this calculation produces identical results fluctuating within the error range generally accepted in evaluating radii namely ZSn8+ = 1.04 & 0.02 A Our ahove-mentioned investigation which was noted by Goldschmidt,6 has permitted the number of known geometric characteristics of ions in the crystallochemistry of inorganic compounds to be increased by means of the Rorn-Haber cycle from purely energetic magnitudes.The entropy of aqueous ions was later added to them.’ This method is helpful to a certain extent also in finding the dimensions of more complex ions or radicals which may be assumed to have spherical symmetry. Thus for example the Reviewer and K. B. Yatsimirsky,8 proceeding from equation (4) have calculated the thermochemical radii of a large number of tetrahedral ions. We have found that they all obey the additivity principle and only for a few combinations of relatively large anions with small cations are there small deviations ; these are explained by the peculiar ‘‘ intrusion effect ” due to the fact that in this kind of structure the small outer cation is capable of penetrating into the “ cavities ” between the ions comprising the large anion owing to which the distances between the particles diminish.* Table 1 gives the dimensions of some ions.A. Kapustinskii 2. Krist. 1933 86 A 359. 6 V. M. Goldschmidt “ Kristallchemie ” Handworterbuch der Naturwissenschaften 2nd edn. Vol. 5 Jena 1934 (Russian translation Leningrad 1937 p. 25). 7 A. Kapustinskii Doklady Akad. Nauk X.S.S.R. 1941 30 625. * A. Kapustinskii and K. Yatsimirskii Zhz~r. obshchei Rhint. 1949 19 2191. * The “ intrusion effect ” can be observed not only in complex ions but also in simple ones which reduces the formally binary system to a monolithic spherical form.Such for example is the nearly spherical hydrogen sulphide ion HS- formed by placing a negligibly small proton in the electronic environment of a large anion S2- [:S:]+ + [HI+ -+ [:<&:I- .. .. which agrees well with the insignificant polarity (0.65 X A. Kapustinskii I. Makolkin and L. Krishtalik Zhur. $2. Khim. 1947 21 126). of such a system (cf. 286 QUARTERLY REVIEWS TABLE 1. ‘‘ Therrnochemical” radii of tetrahedral ions in Strictly speaking the later examples belong to complex compounds. Typical examples of complex ions will be treated in the following section. The energetics of complex compounds While some complex radicals especially those of octahedral configura- tion may be considered as spheres whose dimensions are directly and fully determined by their radii there are a good many ions whose structure is far from spherical.In all cases however one may speak of effective ionic radii as spherical particles which can equivalently i.e. without changing the energy of the lattice as a whole replace in the solid an ion of any con- figuration. lo Such effective thermochemical radii can evidently be found in the same way i e . from the Born-Haber cycle in combination with the crystal-lattice energy equation (4). They are constant and additive mag- nitudes to such an extent that it is clearly manifest even for such ions which in fact are very far from being spherical. A good example of this is the complex compound formed by two par- ticles of approximately equal size such as an ion of barium and a molecule of water which may be illustrated by crystalline monohydrates of various barium salts.Eley and Evans’s investigation permits us to calculate the thermal effect of the process Ba2+(gas) + H,O(gns) = [Ba(H20)12+(gns) ; Qo = 32 kcal./iuole From the cycle I [Ba(H20)]2+(gas) + BCI-(gas) = [Ba(H,O)]Cl,(cryst.) + U L / BaZ+(gas) + H,O(gas) + 2Cl-(gas) it follows that the sum of this thermal effect (Q0) and the already known thermal effects of the separate stages (Q1 Qz Q S anddH,, of part I1 of the D. Eley and M. Evans Trans. Faraday SOC. 1938 34 1093. KAPUSTINSKII LATTICE ENERGY OF lONIC CRYSTALS 287 cycle equa,ls the thermal effect of part I i.e. the energy of the lattice U of barium chloride monohydrate taken here as a concrete example.We can therefore by deriving the latter magnitude from the cycle and using equation (4) proceed to evaluate the effective ion radius z[B~(H,?)~P+ from v,arious monohydrates cont,aining anions of already known dimensions (see Table 2). TABLE 2. ‘‘ Thermochemical” radius in A Radius . . . 1 1.60 1 1-60 1 1.61 1 The average magnitude of T[Ba(H20)]2+ = 1-6 Here additivity is manifestly not less a t any rate than in the deter- mination of the dimensions of simple ions with the aid of lattice parameters determined by X-ray measurements. In the last three compounds in Table 2 the anions are also complex radicals. Yatsimirskii,’o in his book (which is the first monograph on the thermochemistry of complex compounds) has not only computed the thermo- chemical radii for 36 complex anions and 67 complex cations and shown their constancy and additivity but has based the theoretical part of his monograph mainly on various applications of eqn.(4). This has enabled him to extend the energy characteristics of ions quite widely in the chem- istry of complex compounds establish many hitherto unknown binding energies of gaseous addenda to the complexing agent (heat of hydration entropies of complex ions in the solution and alkyl affinities) calculate the heat of solution and solubility of salts and the thermal stability of complex compounds and lastly outline some theoretical principles of the use of complex compounds in chemical analysis. Lattice energy and chemical analysis It is fascinating for a chemist to see the realisation even if only in part of chemical syntheses predicted on the basis of crystallochernistry con- firmed by the production of new theoretically suggested compounds.Since for condensed states the entropy factor is negligible and the free- energy changes which determine the trend of the chemical reaction closely approach thermal effects the apparent problem is to find the free-energy changes if they are known by means of “ thermochemical ” radii and eqn. (4). The negativity of the thermal effect (& = - O F ) thus found indicates that the synthesis cannot occur while a positive effect shows that 10 K. B. Yatsimirskii “ Thermochemistry of CompIex Compounds ” ed. A. F. 0.01 A. Kapustinskii U.S.S.R. Academy of Sciences Moscow 1951. 288 QUARTERLY REVIEWS the planned reaction for obtaining the substance under consideration can be accomplished and t'he more positive the effect the more probable is its curresponding reaction.Pollowing this preliminary theoretical predic- tion a number of hitherto unknown compounds have been synthesised l1 Such as for example [Mg ( C0,N2H4)6][SnC16] [&( C0,N2H4),][ SnC1,] and [Co(C,H,N),CJ,][YbCI,]. Crystallochemical electronegativity of simple and complex ions The fairly long-known concept of electronegativity as the tendency of atvnis in the molecule t o attract electrons first received a strictly quantita- tive expression in the covalent-bond theory which Science owes to Pauling.12 He also used it to calculate heats of formation. Unfortunately these calculations are rather rough and for all compounds not containing nitrogen and oxygen even contradict the facts of the existence of endothermic compounds.Theoretically it would be preferable to take as a basis not the common standard of thermochemistry (the physical state of reagents at 25"c and 1 atm.) but the state of gaseous ions (reagents) and crystals (products) under equal thermodynamic conditions. The exothermic heat of reaction under such conditions is the energy of the crystal lattice M+(gas) + A-(gas) = MA(cryst.) + U If the charge of the ion is 7 then its potential V = ~ j / z and ion energy For the initial state of' the system (reagents) the energy equals the w = rp/22. total energy of the ions HI = Wcntion + Wanion = v2/2zc + ' Y ~ ~ / Z ~ A Such additivity does not take place in crystal lattices for which in accordance with Mulliken's l3 quantum-inechanical investigation i t inay be assumed that the energy of the ions in the lattice is proportional to the geometrical mean of the ionic potentials being the proportionality factor The quantity showing the potential and characterising the energy of the electron attraction to the ion being described by R (this energy increases with the increase of the ion charge and decreases with the increase of the ion radius) Then KC2 = q2/2zC a d R A 2 = ?12/2rA U = HI - H l1 K.Yatsimirskii Izvest. Akad. Nauk S.X.S.R. Otdel. Khim. Nauk 1948 263 ; A. Kapustinskii and V. Vaver ibid. 1951 631. l2 L. Pauling " The Nature of the Chemical Bond " Cornell Univ. Press Ithaca 1940. l3 R. Mulliken J . Chem. Phys. 1934 11 782. KAPUSTIN SKI1 LATTICE ENERGY OF IONIC CRYSTALS 289 or finally passing from one bond for which the conclusion has been drawn to the summation X of all the bonds U = X(Rc - IiA)' .- ( 5 ) Here the magnitudes Iz characterise the energy of the electron attmction to the ion in the crystal lattice and therefore in contrast to Pauling's electronegativities may be called " crystallochemical electronegativities ". It is most convenient to compute their values directly from experimental data i.e. from the lattice " experimental " energies the fluoride ion being assumed to have R = 1. Then with equation (5) it is not difficult t o obtain the corresponding data for the other ions (thus for chlorine R = 2.1 ; sulphur 0.4 ; oxygen 1.8 etc.). To what extent these values are really additive may be judged from the example of zinc corripounds (Table 3).TABLE 3. Crystallochemical electronegativity of zinc ion (kcak ./mole at 25 "c) Compound . . . . . . 1 ZnO 1 ZnS 1 ZnF ZnC1 Lattice " experimental " energy U . . . . ~ 970 ~ 862 ~ 672 ~ 630 R for Zn2+ . . . . . . 20.2 20.2 19.3 19.9 Despite the wide range of the lattice-energy changes R for Zn2+ remains The R thus found for various constant within 3% (mean 20.0 & 0.2). other ions are given in Table 4. TABLE 4. Ion crystallochemical electronegativity (kcal./mole at 2 5 " ~ ) (For the fluoride ion R is assumed to be unity) ~ €I - 1.8 Li + %e2+ 16.5 21.2 Na + Mg2+ 16.6 19.7 K+ Ca2+ 14.8 18.6 cu+ 17.5 cu2+ zn 2 + 19.6 20.0 Rb -t- Sr2+ 14.5 18.1 - Cd2+ 19.0 Ag + 16-8 Cs + Baa + 14.1 17.6 B3+ c4 + 24.2 30.2 23-0 26.2 Sc3+ Ti4 + 21.3 3 5 4 ~ 1 3 + Si4 t Ga3+ 22.8 1n3+ 22.3 La3 + 20.1 02- F - - 1.8 1.0 82- c1- - 0.4 2.1 s e 2 - Br- - 0.1 2.3 Fez+ 19.9 co2+ 20.0 20.1 Ni2 + Cr3+ I- 22.5 2.7 Mn4 + 26.1 Sn4+ Mn2+ 24.2 19.8 Pb4-k Tl3+ Pb2+ 24.8 15.2 18.0 290 QUARTERLY REVIEWS These new physicochemical constants l4 which permit an approximate evaluation of crystal lattice energy like the other constants appear to obey the Mendeleyev Periodic Law (Fig.1). fl- t.? I4 16 18 20 22 24 26 28 3yI -2-I 0 I 2 3 R FIG. 1 The Meibdeleyev Periodic System and i o n crystallochemical electronegativity. Lilies connect elements of like groups (and subgroups). As the Reviewer and Yakushevskii l5 have shown R has similarly been calculated for 79 complex ions and not only has the possibility of con- sidering the complex ion in the lattice as a sphere been confirmed but it has also been demonstrated that the influences of various neutral addenda such as the molecules of water and ammonia on the crystallochemical electronegativity of a central ion in a complex compound are practically equal.Lattice energy and the thermochemical logarithmic rule In Fig. 1 it has already been shown that crystal-lattice energy manifests periodicity in accordance with the Mendeleyev system. Let us examine this matter in somewhat greater detail. The approximate dependence established earlier l6 between the heat of formation of chemical compounds (calculated for the equivalent weight) and the place of elements in the Mendeleyev system determined by the atomic number the so-called therinochemical logarithmic rule in the mathematical expression indicates that there is a linear dependence between the heats of formation and the logarithmic atomic number in the series of similar compounds (halides oxides etc.) in both the groups and periods of the Mendeleyev system.This approximate concordance is observed for ions and atoms having analogous electronic shells. The above-mentioned rule which basically shows the connection between thermochemistry and l4 A. Kapustinskii Doktady Akad. 2L’cl.uk S.h”.S.R. 1949 3 467; 1949 4 663. l5 A. Kapustinskii and B. Yakushevskii Ixvest. Sekt. Platiny drug. blagorod. lMetall. l6 A. Kapustinskii Izvest. Akad. Nauk S.X.X.R. Otdel. Khim. Nauk 1948 668 581. 1952 152. KAPUSTINSKII LATTICE ENERGY OF IONIC CRYSTALS 291 the Mendeleyev periodic system and was recently discussed in Long's 1' excellent review extends also to lattice crystal energy.This is only natural since the latter is a most simple kind of thermochemical constant referred to the physical standard state ionic gas-crystal uniform for all ionic crystals. A detailed exposition of this question is given by the Reviewer in an article soon to appear." Isotopy and crystallochemistry We therefore would like to recall here our work jointly with Shamovskii and Bayushkina,l* dealing with this subject. I n it the elementary photo- chemical process in the ionic crystal is treated as corresponding to the direct " leap '' of the electron in the lattice from anion to cation with the formation of neutral atoms. I n this case from the cycle which may be called ' * photochemical " it is easy to obtain an expression for the maximum absorption band hv from which The thermochemistry of isotopes still receives insufficient attention.h v = E - I + U . - (6) Light absorption and electron leap h v [ Li+H-)( cryst. ) -+ Li(gas) + H(gas) / - I / + E I Li+(gas) + H-(gas) where E is electron affinity I is ionisation energy and U is lattice energy. In accordance with the example chosen to illustrate the cycle lithium hydride was used in the investigation. Subliming this substance in a high vacuum on a quartz plate and determining spectrophotometrically the maxi- mum absorption of the crystals in the ultraviolet part of the spectrum and then performing the same experiments with lithium deuteride we obtained results permitting us with the aid of equation (6) to calculate U for lithium compounds with light and heavy hydrogen isotopes ULiH = 219.2 kcal./mole and ULi = 220.8 lscnl./niole .(7) These values appeared to be in agreement with theoretical lattice-energy calculations carried out by Zintl and Hardner l9 with the aid of X-ray measurements. Substitution in the lithium hydride crystal of the light hydrogen isotope protium by the heavy isotope deuterium increases the crysta,l lattice energy by 0.5%. 17 L. H. Long Quart. Rev. 1953 7 134. 18 A. Kapustinskii L. Shamovskii and K. Bayushkina Acta Physicochim. 1937 6 19 E. Zintl and A. Hardner 2. phys. Chern. 1935 28 B 478 493. * A. F. Kapustinskii " Ionic Crystals Lattice Energy and the Thermochemical Logarithmic Rule " Transactions of Mendeleyev Institute of Chemical Technology Moscow. 799. T 292 QUARTERLY REVIEWS Errors in lattice-energy calculations and possibility of their greater accuracy It is known that the more complete expression for lattice energy contains addit’ional terms namely the zero-point energy e0 and the van der Waals forces c / r 6 with the constant c computed by the London method (a good exposition of this question is given by Glasstone 20).Comparison however with experimental data obtained by the Born-Haber cycle indi- cates that for purely ionic crystals the second and the third term have little significance and therefore attention should be directed chiefly to tlhe possibility of increasing the accuracy of the first term as playing the main role in the general balance of crystal energy. Since the crystal lattice energy equations considered in this Review [which are reduced to taking account precisely of the first term of eqn.(S)] have found certain currency in investigations of general chemistry and geochemistry it is important to have a clear idea regarding the accuracy of these expressions. Though the last question seems a simple elementary one it is unfor- tunately not always possible to get a satisfactory answer to it as direct experimental det’erminations subject to comparison with theoretically com- puted energies have been carried out only very rarely and the accuracy of the values in the Born-Haber cycle has usually been unknown. Neverthe- less the Reviewer and Weselowsky 21 had in 1933 verified equation (3) on then accessible thermochemical material and showed that the accuraey of equation (3) is of approximately the same order as that of eqn. (1). Later the Reviewer 5 had similarly verified the accuracy of the improved equation (4) which was already based on the quantum-mechanical concepts of the nature of repulsion forces.In Fig. 2 the points on the co-ordinates U/I=nq,q2 and l/(rc + rA) represent experimental data for metal halides oxides and chalcogenides ; the curve corresponds to equation (4) the accuracy of which is of the order of 2-3%. It will be noted that in most cases the ionic radii themselves are known with the same accuracy. Gradually with the progress of experimental thermochemistry the accuracy of various magnitudes in the Born-Haber cycle has increased and so has the number of direct experimental determinations of U . Rossini and Wagman together with Evans Levine and Jaffe of the National Bureau of Standards U.S.A. have carried out a critical selection 22 of the more reliable thermochemical constants though still not supplied as a rule S.Glasstone “ Tesilmol.; of Physical Chemistry ” Maemillan New York 1946. 3 1 A . Knpustinskii and B. Vcselovskii Z. phys. Chem. 1933 22 B 261. 32 It’. Itossiiii 13. Wagman IT-. Ekans S. Leviiit and I. Jaffe “ Selected Values of Chemical Thermodynamic Properties ” Circular of the National Bureau of Sbanciards No. 500 Washington D.C. 1952. KAPUSTINSKII LATTICE ENERGY OF IONIC CRYSTALS 293 with an accuracy limit. This work partially includes besides results of previous works by Mayer. and Helmholtz,23 more recent results of direct determinations of lattice energies by Saha Sriva,stava and Tandon. 24 As for the ionic radii values in equations (3) and (4) no essential changes and improvements have been made in them in the past period.720 700 90 80 0.3 Q.4 1/(% + ?4) FIG. 2 Compurisou of L' czperiruerztal " lattice energies (points) wit18 those (curue) calculated by means of equation 4. The Reviewer and Yatsimirskii have this year used new thermocheniical data for a reverification of equa,tioii (4). The work has been published in the Zhurnal obshchei Khimii. As an object for investigation we took halides of Groups I and I1 of the Mendeleyev system. All of them though they crystallise with different structures are typical ionic compounds and there are for all of them thermochemical data giving a clear idea of the accuracy of the measurements. At the same time since the ions under consideration belong to the noble-gas type the Goldschmidt ion radii used were not subject to doubt.If in equation (4) the accuracy of the value of the experi- mentally obtained coefficient for p = 0.30 is increased one may reach the 23 J. Mnyer Z . Physik 1930 61 798 ; 1,. Helmholz and J . Mnyer J . Chem. Phys. 2 4 M. Snlin mid A. Tandon Proc. Nut. Inst. Xci. Inditi 1937 3 288 ; B. Srivastava ibid. 1938 4 365 ; A ' h d o n I n d i m J . Phys. 1937 11 99 ; idem Proc. Nut. Inst. Sci. India 1937 7 102. 1934 2 246. 294 QUARTERLY REVIEWS conclusion that this formula continues true with the same accuracy for the experimental thermochemical material. When necessary this accuracy niay even be increased by the introduction of the empirically found dependence of p on the sum of ionic radii When the crystal lattice energy (4) is substituted into our equation we obtain the following particular form of this expression == 0.345 - 0.00435(~ + z A ) ~ .' (9) I n this case a comparison of calculated results with experimental data shows that the errors are distributed statistically and that the quadratic mean deviation for salts of the MX type is 0.8 lical. (i.e. about 0.5%) and for salts of the MX type 5.2 kcal. (i.e. about 1%). This more accurate expression is unfortunately more cumbersome. There is hardly any sense in any further improvement of the expression for lattice energy by means of ionic radii until the values of the ionic radii themselves are known with great,er accuracy. Whether this will be possible is hard to say. Thorough investigation of ionic radii from all aspects is needed. An account of such a theoretical investigation by the Reviewer will be found in the first issue of the new Zhurnal neorganicheskoii Khimii which began to appear in the U.S.X.R.in 1956. Electron affinity of oxygen sulphur and selenium The more accurate thermochemical data already mentioned above have in their turn enabled us to obtain greater accuracy for some of the values which play an essential part in general chemistry. Thus in combination with Pritchard's 25 critically selected data on the heats of formation of gaseous atoms of oxygen sulphur and selenium one can find the values of electron affinity with precise limits of possible errors (Table 5 ) . TABLE 5 . Electron afinity (kcal. at 25") Ogas + 2e = Oz-gas Segas + 2e = Se2-gas s,s + 2e = s2-g, AH,, = 172 & 5 AH,, = 100 & 2 AH,, = 117 2 2 5 H. Pritchard Chem. Rev. 1933 52 529.

ISSN:0009-2681    

DOI:10.1039/QR9561000283

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

DIRECT MEASUREMENT OF MOLECULAR ATTRACTION BETWEEN SOLIDS SEPARATED BY A NARROW GAP By B. V. DERJAGUIN (CORRESPONDING MEMBER OF THE U.S.S.R. ACADEMY OF SCIENCES) I. I. ABRIKOSOVA and E. M. LIBSHITZ U.S.S.R. ACADEMY OF SCIENCES Moscow) (LABORATORY FOR SURFACE PHENOMENA INSTITUTE OF PHYSICAL CHEMISTRY Introduction BESIDES the usual valency forces with a comparatively short radius of action which practically disappear a t interatomic distances of several Angstrom units there are forces of attraction acting between any two atoms or molecules which decrease much more slowly with distance. These molecular forces form the basis of a wide field of fundamental problems in physical chemistry and molecular physics. Surface tension capillarity capillary condensation physical adsorption and most other surface phenomena are mainly explained by molecular forces ; they also determine most of the properties of liquids-their viscosity heat of evapora- tion and solubility in other liquids.The attraction between atoms and molecules naturally gives rise to analogous '' molecular attraction " between two macroscopic bodies whose surfaces have been brought within a short dist'ance of one another. As an aample we may cite coagulation processes in colloidal and aerosol systems caused by the molecular interaction between colloidal particles when they approach each other. The idea of the role of molecular long-distance forces was first put forward in a short article by Kallmann and Willstatter. Similar forces also form the basis of the quantitative theory of stability and coagulation of colloids,l.along with the repulsion forces of the double diffuse ionic layers of approaching particles. Some authors e . g . Langmuir,3 consider however that the theory of colloidal systems can be formulated without assuming the existence of long-distance attraction caused by the van der Waals forces between the molecules of the neighbouring colloidal particles. Thus the problem of calculating the force or energy of molecular jnter- action of particles large in relation to molecular dimension as the function of the distance between them is one of the basic problems in the theory of B. Derjaguin Bull. Acad. Sci. U.R.S.S. Classe Math. Nnt. SPr. Chim. 1937 1153 ; Acta Physicochim. 1939 10 333; Trans. Faraday Soc. 1940 36 203 730. 23. Derjaguin and L. Landau Acta Physicochim. 1941 14 633 ; J .Exp. Theor. Yhys. (Russia) 1941 11,802 reprinted 1945 15 662 ; Verwey and Overbeeck " Theory of' the Stability of Lyophobic Colloids " Elsevier Amsterdam 1948. I. Langmuir J . Chem. Phys. 1938 6 873. 295 296 QUARTERLY REVIEWS stability and coagulation of colloids. The same forces must play an im- portant part in thixotropic phenomena and the formation of tactoids and coacervates as well as in swelling phenomena syneresis and other colloidal processes. I n spite of the considerable importance both theoretical and practical of molecular forces the study of their nature and theoretical foundation is of recent origin. A correct idea of the nature of molecular forces was first put forward by the physicist P. N. Lebedeff. In 1894 Lebedeff,4 in discussing the mechanical action of waves on resonators wrote " In Hertz's researches in his interpretation of light oscillations as electromagnetic processes there lies another problem which has hitherto not been considered the problem of the sources of radiation of the processes which take place in a molecular vibrator when it radiates light energy into space." This problem takes us on the one hand int,o the field of spectral analysis and on the other quite unexpectedly into the theory of molecular forces one of the most complicated problems of modern physics. This follows from the following considerations. Prom the standpoint of the electro- magnetic theory of light i t must be admitted that between two light- emitting molecules as between two vibrators in which electromagnetic oscillations arise there exist mechanical forces caused by the electrodynamic interaction of the alternating electric currents in the molecules (according to AmpBre's law) or of the alternating charges in them (according to Coulomb's law).We must therefore admit that in this case there exist intermolecular forces whose origin is closely connected with radiation processes. . . . " . . Of special interest and difficulty is the process which takes place in a physical body when many molecules interact simultaneously the oscillations of the latter being interdependent owing to their proximity. If the solution of this problem ever becomes possible we shall be able from the results of spectral analysis to calculate in advance the values of the intermolecular forces due to molecular inter-radiation deduce the laws of their temperature dependence and by comparing the values obtained with experimental results solve the fundamental problem of molecular physics whether all the so-called ' molecular forces ' are confined to the already known mechanical action of light radiation mentioned above to electro- magnetic forces or whether some forces of hitherto unknown origin are involved.. . ." The first quantitative theories of molecular forces could be built only after the structure of atoms and molecules had been established. One of them is the general quantum-mechanical theory of molecular forces between isolated pairs of atoms or molecules formulated by L ~ n d o n . ~ When the distance between the molecules is great compared with their diameters (mainly jn the case of gases) theory leads to forces decreasing in inverse ratio to the seventh power of the intermolecular distance.However in P. N. Lebedeff Works Moscow 1913 pp. 5G-57; Wied. Ann. 1894 52 621. F. London %. Physik 1930 63 245 ; 2. phys. Chem. 1931 11 221. DERJAGUIN ABRTKOSOVA AND LIFSHITZ MOLECULAR ATTR,ACTION 297 solid bodies where the molecules do not rotate freely there may exist forces decreasing much more slowly with the distance. Simultaneously at the short intermolecular distances peculiar to condensed bodies forces which decrease more rapidly with distance as for instance those connected with quadrupole moments may have a strong influence. This being taken into account the existing attempts to verify the theory of molecular forces quantitatively must be considered to be and to have been in principle incapable of yielding accurate and conclusive results.Indeed all these attempts have been based on a comparison with the theory of the effects of integral nature in which those ternis prevail which depend on close intermolecular action i.e. on the interaction of molecules a t distances of the same order as their radii vix. the value of the constant a in the van der Waals equation the heats of sublimation and evaporation and the energies of adsorption and wetting. Accurate comparison with theory is rendered difficult in this case by the fact that not a single theory of molecular forces is strictly speaking applicable at such short distances. Moreover the result involves superposition of forces of different nature (for instance the quadrupole forces) which depend not only on the orientation of the mole- cules (which is frequently unknown) but also on the asymmetry of the molecular force fields.The current theories of molecular forces could be checked far more strictly if their action were measured a t distances much greater than molecular diameters. Of special interest from this point of view are measurements of the resultant molecular attraction of solid bodies separated by a gap many molecular diameters wide i.e. measurements similar to those in Cavendish’s experiments with gravitation forces and Coulomb’s experimerits with the forces acting between electric charges. Such experiments in contrast to those measuring adhesion forces upon contact,6 would if supplemented by a method of summing the forces between pairs of molecules of the solid concerned enable us to check the existing theories of intermolecular action at distances which involve only one order of forces and at which the existing theoretical limitations are swept away.First of all they would allow us to check the quantum-mechanical theory of dispersion forces and thus give ground for the elucidation of their true nature. At the same time i t is evident that such measurements are of the utmost interest especially if we take into consideration the wide application in the past of molecular long-distance forces to the basic problems of colloidal chemistry and surface phenomena on the one hand and the doubts as to their existence expressed for instance by Langmuir on the other. Finally the measurement of molecular attraction in bodies ‘of finite size is important for a verification of the methods of summation of molecular interactions.As far as we know until quite recently (1951) no experiments of this kind even of a qualitative nature were published. The reason doubt- less lies in the obvious experimental difficulties which we shall discuss later. R. Bradley Phil. Mug. 1932 13 853. 298 QUARTERLY REVIEWS The aim of the present Review is to discuss a method of direct rneasure- ment of tihe molecular attraction of two solid bodies as a function of the gap separating them and to apply the result to check the theories in question as well as to solve some problems of colloid chemistry and surface phen- omena. Simultaneously the current theories concerning the existence and magnitudes of molecular forces in objects of finite size are discussed. Critical review of the current theories of molecular forces Objects of Molecular Dimensions.-According to F.London's calcula- tions the energy of interaction between separate atoms or molecules a t distances much greater than their diameter is in inverse ratio to the sixth power of the distance between them. The London law can be formulated as follows where U is the energy of molecular interaction between two particles separ- ated by the distance r and C is a positive coefficient constant for each type of atom ; it is calculated by means of the matrix elements of the electric moments of both atoms. points out that for many simple mole- cules one may use the approximate formula London where hv is the characteristic energy term which can be deduced from an experimental formula for the optic dispersion of gases and a is the polaris- ability of the molecule.As the force is expressed by F = - dU/dr the van der Waals forces of attraction change in inverse ratio to the seventh power of the distance between the molecules. The London theory has its limitations calculations based on it becoming incorrect not only for very short interatomic distances when the wave functions of the atoms overlap but also for great distances when it is necessary to take into account electromagnetic retardation. Electromagnetic retardation was taken into account by Casimir and Polder 7 who made use of quantum electrodynamics for this purpose. They applied the same perturbation method as London but besides electrostatic attraction their perturbation operator contained the interaction of the radiation field of one atom with the othqr and vice versa.According t o these authors if r > 2 where 3 is the chief absorption wavelength of the atom in question the energy of interaction between two atoms with static polarisability cc is . - ( 3 ) 23 hca2 u=--.- 4n 2nr7 or U = C'/r2 where C' = 251e2cc2 ' ( a ) H. R. C. Casimir and D. Polder Phys. Rev. 1948 73,360 ; ( b ) H. B. C. Casimir Proc. k . ned. Akad. Wetenschap. 1948 60 793. DERJAGUIN ABRIKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 299 The attraction force between two atoms (- dU/dr) in this extreme case thus varies inversely with the eighth power of the distance. Thus a t present we have theories which explain the origin of mole- cular attraction and enable us to calculate the interaction of free atoms and molecules.The most detailed description of this problem is given in the reviews by London8 and Margena~.~ Objects Large compared with Molecular Dimensions.-Besides attraction between individual free atoms or molecules of special interest for colloid chemistry and for the theory of disperse bodies in general is the problem of interaction between macroscopic objects. By this we mean the interaction between bodies due to van der Wads attraction between their constituent molecules i .e. the resultant molecular attraction of two bodies. The London forces being assumed to be additive the attraction between bodies consisting of large numbers of molecules can be calculated by sum- ming the attraction energies of each pair of molecules constituting the body. Thus de Boer lo and Hamaker I1 find the interaction of two bodies con- sisting of q molecules per unit volume by integrating the elementary inter- actions which obey the London law.Hamaker deduces formulae for the energy and force of attraction between two equal spheres between a sphere and an infinite flat wall and between two flat parallel walls. If the minimum distance between the surfaces is much smaller than their radius of curvature p the interaction energy is in the first case expressed by U = - A p / l 2 h . * (4) and the force by F = Ap/12h2 . * (4’) u=-4 P/6h * * ( 5 ) F = Ap/6h2 . * ( 5 ’ ) u = - A/12nh2 . * (6) f = - A/6,h3. (6’) In the second case the energy and the force are respectively For two infinite plates the energy and the force f per unit area of surface are given by where h is the minimum distance between the bodies and A is the constant introduced by Hamaker which depends on the character of the bodies and equals n2q2C’.These formula? are usually used in calculating the interaction between colloidal particles and other macroscopic objects (the Hainaker-London interaction). If these calculations are carried out with allowance for the electro- magnetic retardation effect the following equation results in the limiting 8 F. London Trans. Faraday SOC. 1937 33 8. lo J. H. de Boer Trans. Faraday SOC. 1936 32 10. l 1 H. C. Hamaker Physica 1937 4 1058. H. Margenau Rev. Mod. Phys. 1939 11 1. 300 QUARTERLY REVIEWS case of sufficiently great distances. parallel plates i.e. when h > Am For the energy per unit surface of zc = - A'/30nh3 . (7) and the force per unit area of surface f = A'/lOnh,~ .- (7') where A' = n2qC' The applica,bility to the case of condensed bodies of methods of calcula- tion for which London forces are assumed to be additive has never been proved either theoretically or experimentally from the standpoint of quantum mechanics. This procedure would be justified only in the absence of any strong intermolecular interaction for example in the impossible case of two gases separated by a gap. Moreover in condensed systems the atomic and molecular character- istics particularly a and I' differ from the properties of isolated atoms and molecules owing to the influence of neighbouring particles on each other. The share of each separate molecule in the total molecular interaction depends on the co-ordination and concentration of the molecules and for surface molecules on the number of their neighbours.Thus strict addi- tivity being admitted we must in order to be consistent take the values a and F for isolated molecules which lead to a deliberate error. On the other hand if we do not the " true " values of cc and r are dificult to obtain as they are hard to determine for condensed systems and are often entirely unknown. Besides the lack of physical logic in the above method it should be pointed out that it is always difficult in practice to calculate the constants A and A' even for isolated atoms and molecules. I n most cases calculation does not lead to quantitative results and it is not clear hgw this could be achieved as for many atoms the values of a and r have not been deter- mined. I n such cases it remains only to find the polarisability from the refraction data of solid bodies and thus to use characteristics of condensed media in the London formulae which are true only for the interaction of separate atoms.Besides the difficulties connected with finding the parameters determin- ing the values of A we should point out the doubtful applicability in general cases of an approximate equation (2) which has been proved true by London only for a number of simple molecules. The theory of molecular attraction between bodies large compared with molecular dimensions It is possible to approach this problem from a quite different-a purely " macroscopical "-point of view in which the interacting bodies are con- sidered as continuous media. This approach is justified inasmuch as the distance between the surfaces of the bodies although small in other respects is assumed to be large compared with the interatomic distances in the solids.The basic idea of the proposed theory is to consider the interaction DERJAGUIN ABRIXOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 30 1 of the bodies as realised by means of the alternating electromagnetic field. Owing to thermodynamical fluctuations such a field is always present inside any material body and penetrates also into the surrounding space outside it. h well-known manifestation of this field is the thermal radiation of a body but it is to be emphasised that tlhe field of this radiation does not exhaust the whole alternating electromagnetic field outside the body. This is most clearly seen from the fact that the electromagnetic fluctuations exist also at the absolute zero of temperature when there is no thermal radiation at all ; at this temperature the fluctuations are of a purely quantum-mechanical nature and are connected with the so-called zero-point oscillations of the electromagnetic field.We imagine both bodies as two semi-infinite regions separated by a plane-parallel slit of the given width h . The procedure of the theory is to calculate the electromagnetic alternations in such a system and in particular to determine the electromagnetic field inside the slit. This being done we can determine the force f acting on each of the two surfaces (per of its area) by calculating the corresponding components of the Maxwellian stress tensor. It is to be stressed that the outlined method of approach to the problem is quite general and is applicable a t any temperature t o any two bodies independently of their molecular nature (ionic or homopolar crystals amorphous bodies metals dielectrics etc.).An important feature is also the fact that this method automatically takes into account the so-called retardation effects due to the finite propagation velocity of electromagnetic interactions; these effects become predominant when the distance h is large enough in which case h > Ao 1 being the characteristic wavelength of the absorption spectrum of the bodies in question. We shall discuss here briefly the final results of the theory without dwelling upon the details of the rather complicated calculations which can be found in the original papers.12 I n the following formulze enters F(CU) the dielectric permeability of the body as a function of the circular frequency of the field.* It is to be recalled that e(w) is in general a complex quantity [ E = ~’(co)) + ie”(ci~)] and its imaginary part ( E ” ) is always positive and determines the energy dissipation of a wave propagating in the body.F is connected with the refractive index n and the extinction coefficient K of the medium by means of the familiar relation %/le = n + iK. As it is known by considering ~ ( c r ) ) form- ally as a function of a complex variable cr) it is possible to obtain certain integral relations between ~’(cr)) and ~ ” ( c r ) ) (the so called Kramers’s relations). A particular consequence of these relations is the formula which determines the values of the function E for purely imaginary l2 E. M. Lifshitz Doklady AIsad.NauE S.S.S.R. 1954 97 643 ; 1955 100 879 ; Zhur. exp. teoret. Fiz. 1955 29 94. * We assume that the magnetic permeability p of the bodies can be put equal to unity as is usually the case. 302 QUARTERLY REVIEWS arguments by means of the function &”(a) for real values of w. &(it) is a real function and decreases monotonically from the value c, (the electrostatic dielectric permeability > 1) when = 0 to 1 when t = GO. Now the general expression for the attractive force f is * where E replaces &(i[) R = ( F - 1 + pz)z and p and are the two real integration variables. It is seen that the attraction force can be in prin- ciple calculated for any dista,nce h provided the function &(it) is known. But according to equation (8) &(it) is known if we have the function ~ “ ( w ) (in a sufficiently wide spectral region of the frequencies).Thus the imagin- ary part &”(a) of the dielectric polarisation of the bodies is the only one of their macroscopic properties which is needed for the calculation of the molecular attraction force. The complicated formula (9) is considerably simplified in certain limiting cases. First let h < 1,. In this case the general expression (9) can be reduced to the following - (10) x2dxdl f=- 32:3h301:{m [ ( F + l ) / ( e - 1)I2ez - 1 * or with sufficient accuracy in practice Thus we arrive in this case a t an inverse cube law for the distance-dependence of the force with a coefficient which can be calculated provided &(it) is known. It is of interest to perform in equation (11) a formal transition to the case when both bodies are “ gases ”.The macroscopic characteristic of a rarefied medium is that its dielectric polarisation is near unity. By making this assumption and using equation (8) it is possible to reduce (11) to the following form If we consider this force as a result oi interaction of individual pairs of molecules it corresponds to the interaction law with the potential energy 3h * In order to simplify formulae i t is assumed here and in the following that both The formuh can be written also for the case of bodies are of the same substance. different substances. DERJAGUIN ABRJXOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 303 Here r is the distance between two molecules q the number of molecules per ~ m . ~ and a known formula is used 2n 2e 2q$ (CU) cu.&"(CU) = - -___ m This relates ~ " ( c r ) ) (of a gas) to the " spectral density " of the oscillator strength $(cu). Now consider as a simple example two hydrogen atoms and write a summation over the discrete energy levels En of an atom instead of the integration in (13). By using the well-known expression for the oscillator-strength of the transition from the state En to E (Xon is the corresponding matrix element of the co-ordinate of the electron in the atom) we obtain which is exactly London's quantum-mechanical formula for the van der Waals forces (without the retardation effects). Thus we see how this formula is reproduced here from a purely " macroscopic " theory. We proceed to the opposite limiting case of " large " distances where h > Ao. In this case equation (9) can be reduced t o the following expression where so = ( E ~ - 1 + p2)$-.Here E~ is the electrostatic value of the dielectric polarisation of the substance. Thus in this case we have a reciprocal seventh-power law for the force with a coefficient which is determined solely by the electrostatic dielectric polarisation. The integral ( 14) can be computed nuinerically for any value of E ~ . We can write it in the form Some numerical values of + ( E ~ ) are as follows L I/% ' . 0 0.025 0.1 0-25 0.50 1.0 $44 . 1 0.53 0.41 0.37 0.35 0.35 .For the attraction between two metal surfaces we put 8 = infinity and from equation (15) hc n2 2nh4 240 f=-.- . This coincides with the result derived by H. B. C. Casimir by a different However it is to be stressed that it holds only for the case of h >A,.304 QUARTERLY REVIEWS For the case of 1 (a " rarefied'' medium) This force corresponds to interaction of individual molecules wit'h the potential energy where a = (Q - 1)/4nq is the static polarisability of the molecule. This is exactly Casimir and Polder's result derived by the quantum-electro- dynamical method taking into account the retardation effects. Thus this result too is derivable from the macroscopical theory. A question naturally arises as to what actua,lly is the magnitude of A with which h is compared (the characteristic wavelength of the absorption spectrum of the body). The answer cannot be given in a general way and depends on the actual form of the spectral distribution of the absorption in the body concerned [i.e. on the actual properties of its function ~ " ( c u ) ] .For instance for metals one can arrive a t a reasonable estimate of the region in which equation (16) is valid by taking e ( u ) in the forin &(a,) = - 43te2,/1/mcu2 so that e(iE) = 4ne2q/rnt2 where ; this formula is known to be satisfactory in the infrared region of the spectrum. By start- ing again from the general equation (9) and developing it in powers of h-l the following expression is obtained is the number of the conduction electrons per f = 480h4 *[l - ,.2(;)?] . Taking for example rI = 5-9 x ~111.1~ (for silver) we find that the second term in the bracket is small and thus equation 16 holds if h > 5500 8. The case of quartz with which the experiments of B. Derjaguin and I. Abrikosova were performed presents some special features owing to the peculiarities of its absorption spectrum.Quartz is known to exhibit strong absorption in the ultraviolet and the infrared region and between these regions is a broad gap in which quartz is transparent ; the distances h used in experiments fall roughly in the latter region of 3 values. An analysis based on the general equation ('3) shows that one can obtain a fair estimate of the attraction force for these distances by using equation (15) but taking for c0 not the electrostatic dielectric polarisation but the value of the dielectric polarisation in the region of optical transparency. The value off thus obtained is somewhat underestimated for larger and overestimated for smaller values of h. In all the formulae written above the influence of the finite temperature To K of the bodies on the attraction force is neglected i.e.the formulae DERJAGUIN ABRIKOSOVA AND LIFYHITZ RTOLECULAR ATTRACTION 305 are written for T = 0. justified. In practice this neglect is usually completely A necessary condition is that kT < hcu0,/2n where cr)o = 2nc/& This is certainly €ulfilled a t ordinary temperatures. However this condition is not sufficient and the analysis shows also that the inequality kT < hc/2nh must hold in order to justify the approximation T = 0. The latter condition is certainly violated for sufficiently large distances h. At such distances the temperature effect becomes significant ; this point is in principle important although in practice the force f for these values of h is already very small. We shall not reproduce the general formula for the attraction force for arbitrary values of h and T which is a generalisation of tlhe formula (9).We shall only mention in order to illustrate the temperature effect that in the opposite limiting case when h > hc/2nkT the attraction force turns out to be Thus for sufficiently large distances we have again an inverse third-power law but with a coefficient which depends on the temperature (and on the electrostatic dielectric polarisation). Method of measurement Prom what has been said above it is easy to understand the importance of working out a method for direct measurement of the attraction forces between objects large compared with molecular dimensions. In any method of direct measurement of molecular attraction between l)odies the experiment always reduces to the measurement of two quantities namely the force of interaction between the two bodies and the width of the gap between them.This task presents extreme experimental difficulties which can be over- come successfully only if the objects selected for the investigation are of suitable shape and substance. Objects of Measurement.-For a number of reasons it is advisable to have one of the objects of spherical shape and the other flat. Therefore we measured the force of attraction between a plate 4 mm. x 7 mm. and spherical lenses with curvature radii p = 10 cm. and p = 25 em. This facilitates adjustment of the surfaces which is more complicated in the case of two plates; besides the least distance between the bodies can be determined accurately enough by measuring the diameter of Newton’s rings.Moreover with such objects the dependence of the forces on the radius of curvature of the spherical surface can be studied in order to distinguish the molecular forces which are proportional to p from various camouflaging factors such as those connected with surface charges. The proportionality between the molecular attraction and the radius of the spherical surface is evident from the equation (13) from which can be derived F(h) = 2npu(h) . - (20) 306 QUARTERLY REVIEWS where P(h) is the force of attraction between the spherical surface and the plate and u(h) the energy of interaction between two infinite plates of the same nature in the same medium per cme2 and h the least gap between the surfaces. It follows from the above equation that measurements of the force of attraction between the spherical surface and the plate give the direct value of the energy of interaction between two infinite plates i.e.a result independent of the radius of curvature. Measurements were made both under atmospheric pressure and in vaczco. The interaction of two bodies should not depend on what fills the gap between them be it' air or vacuum. Nevertheless each of the cases has its own advantages and disadvantages from an experimental point of view and a comparison of the results obtained in each case is an important method of control of the accuracy of the measurements. The vacuum method proved to be more convenient and precise owing to the fact that the effect of the viscosity of the air in the gap on changes of its width even though they are brought about slowly may become comparable with the effect of the molecular forces in question making it necessary to wait until each reading became constant.This greatly prolonged the measuring pro- cedure and even then it was often impossible to " catch " a moment during which the reading was free from fluctuations. Moreover it was impossible completely to avoid trembling of the beam owing to air convection currents. Our experiments in vacuo were carried under residual air pressures between 0.1 and several mm. of mercury. Method of Measuring Short-range Interaction between Solid Bodies. The Feed-back Balance.-The main difficulty in measuring molecular attraction between bodies is due to the fact that the forces F becoming perceptible only with very small gap widths increase very rapidly as the gap grows smaller so that dF/dh is high and negative.Therefore if we'bring the surfaces close enough together they will adhere. Obviously this problem requires the use of a balance of considerable restoring moment on the one hand and of high sensitivity on the other these requirements being incom- patible in the case of an ordinary balance. This difficulty was overcome by using the negative feed-back method,14 the idea of which is as follows. The deflection y of the beam from its equilibrium position or its angle of shift 8 causes an electric current i in some way or other which gives rise to an electromagnetic moment M affecting the beam and tending to restore it to its equilibrium position. In the ordinary sense the balance becomes less sensitive the same overload producing smaller deflections of the beam but as feed-back enables estimation of the overload not by the deflection of the beam but by the current value measured with a micro- ammeter or galvanometer the sensitivity can practically be greatly raised.At the same time feed-back reduces the period of oscillations of the balance so that equilibrium in our case was established very rapidly. These perturbing effects were much smaller in vacuo. l3 B. Derjaguin KoZZoid Z. 1934 69 155. l4 Idem Doklady Akad. Nauk S.S.S.R. 1948 84 274. DERJAGUIN A BRIKOSOVA AND LIFSHITZ MOLECULAR ATTEACTION 307 It is easy to make the moment N determined by feed-back many times greater than the parallel moment caused by tlhe shifting of the centre of gravity of the beam from its lowest point of equilibrium. The latter moment usually determines the sensitivity period and otlicr met,rologicai peculiarities of an analytical balance.Thus the device which we shall call simply feed-back radically changes all the features of the balance. In view of the above considerations instead of considering t,he moment due to the force of gravity M’ = d,8 acting on the beam we need consider only the moment M of the electromagnetic forces acting on the beam and depending on the angle of deflection. This moment of electromagnetic forces is M = do the constant d (3 do) depending on the current yield of the beam tracking device arid on the coefficient of proportionality between the current i and the nioinent M . With t,he same overload in one pan the angle of deflection of the balance will be - d/do times smaller than for the same balance without feed-back ajnd the period of the balance will be (d/d,)& times shorter.The parameters of the feed-back and therefore the features of the balance are easily readjusted. This makes the method highly adaptable which is extremely important for the solution of the problem facing us because among other reasons the interaction of bodies with very small gaps between them varies widely with the gap width in magnitude and gradient. Design and Principle of Operatiofl of the Apparatus.15 Is-BaZance. The force of interaction between the flat surface of the plate P and the convex sur- face of the lens L was measured by means of a special beam balance (Fig. 1). The length of the beam K was 35 mm. and its weight 0.1 g. The plate P was placed on the end of the beam and the lens L on a balre not connected with the beam care being taken to leave a sufficiently small gap between the lower convex surfaoe of the lens and the upper surface of the plate.A mirror S was cemented to the other end of the beam arm. The beam had an agate prism A connected to it and resting on an agate fulcrum F. Rough balancing was done with a rider in the form of a. glass fibre C weighing l 5 B. Derjaguin and T. Abrikosova Zhur. exp. tcoret. .Fix. 1951 21 495. lG I. Abrikosova and B. Derjaguin Doklady Akad. Nauk S.S.S.R. 1953 90 1055 Diacusa. Faraday Soc. 1954 18 24. U 308 QUARTERLY REVIEWS 10-50 mg. riding along the beam. The beam was fixed rigidly to a frame 2 having 15-20 turns of wire in a constant field due to the magnet M (Fig. 2) (magnetic field induction 13 = 850 gauss).FIG. 2 Photo-electric beam-tracking device. Feed-back was accomplished by send- ing a current through the coil of the frame 2 from a highly sensitive photo- electric beam-tracking device which kept track of the deflection of the beam. The current was carried to the frame by Wollaston wires 6-10 ,u in diameter and about 30 mm. long. The tracking device consisted of a raster-type photo-relay and one- valve amplifier. The photo-relay was placed 35 mm. above the beam with its optical axis 00 (Fig. 3) parallel to the axis of rotation of the beam. 0 -Po FIa. 3 The light source L was a 50-w incandescent lamp with a small hair-type filament. Passing through a condenser K the light rays illuminated a linear raster P (a glass plate with alternating transparent and opaque bands of equal width) after which they passed through an objective 0 and were focused on a mirror S fixed to the beam (the paths of the light rays are shown by continuous lines).By means of the objective 0, the mirror X and a second objective 0 with tlhe same focal length (7.5 em.) the real image of the raster P was projected on to the plane of a second raster P having the same line spacing (60 per em.) (the paths of these rays are shown by broken lines). The size of the image of the raster P coincided DERJACUIN ABRIKOSOVA AND LIPSHITZ MOLECULAR ATTRACTION 309 with that of the raster P, as they were situated in the focal planes of two similar objectives. The insignificant difference between the focal lengths of the objectives 0 and 0 was easily compensated by shifting the rasters slightly out of the focal planes of their objectives.The planes of the rasters were normal to the plane of the beam and its axis of rotation. The lines of the rasters were perpendicular to the axis of rotation and to the beam itself. The slightest turn of the mirror would change the position of the image of the first raster with respect to the other widening or narrow- ing the gaps allowing the light through to the photocell. Fig. 4 represents FIG. 4 diagrammatically the changes in the light- transmitting surface area depend- ing on the respective positions of the image of the raster PI (hatched bands ) and of the raster P (black bands). After passing through the second raster the light fell on an antimony-casium vacuum photocell which con- trolled the grid of an amplifier valve. The current from the photocell was amplified by means of the simple circuit shown in Fig.2. A battery was included in the grid circuit of the pentode (GAC7) valve through a potentiometer to produce a negative potential on the grid. The resistance of the load R was 11 megohms. The negative feed-back in the amplifier (resistance in series with the cathode RK) guaranteed high stability. The anode current iA partly compensated by the current iB (from a 1.5-v dry battery) was sent to the coil of the balance. By regulating the current iB by means of a resistance box K the current in the coil could be controlled without leaving the steepest part of the amplifier valve characteristic. Principle of operation of the apparatus. At a certain zero position of tbhe beam the current in the frame coil is zero (i = 0).A slight deflection of the beam through the angle 8 changes the amount of light transmitted through the second raster and hence the illumination of the photocell pro- ducing a current i = EB where the current yield k is a constant of the device. Feed-back was accomplished by sending the current i through the frame coil on the balance in the required direction. This coil being situated in a magnetic field the balance beam was thus subjected to the action of the torque ill = ni = nk6 = d8 where n and d are constants n depending only on the nuinber and shape of the turns and on the intensity of the magnetic field. 310 Q,UARTERLY REVIEWS By bringing the spherical surface of the lens which was set on a platform with a fine lever-mechanism close to the surface of the plate the gap h could be reduced until there appeared an attraction force P.The latter deflected the beam to an angle (with the high current yield of the tracking device i.e. with “ rigid” feed-back this angle was very small) where the moment of this force PR (R being the “ arm” of the interaction force) was balanced by the feed-back moment M . The force was calculated by the formula Ir’ = w ~ / R . - (21) where i is the current intensity determined hy a microammeter included in the anode circuit (PA in Pig. 2) and R is the distance from the knife edge L4 to the point on the surface of tlhe plate P corresponding to the shortest distance between the surfaces under investigation (Pig. 1). The niethod of determining the constant n is described below. Thus thanks to the feed-back arrangement the molecular attraction was automatically balanced by a moment acting on the frame in the mag- netic field proportional to thc current intensity and could be determined by measuring the latter.As the glass rider C (Fig. 1) permitted only a rough lialancing of the beam the current) i corresponding to zero force was not actually equal to zero but corresponded to a certain zero monient M, which kept the beam in equilibrium with large gaps before the appearance of interaction between the surfaces. When the interaction appeared the equilibrium current i increased by Ai and the force was calculated according to the formula P = TLA~,/R . - (22) The appearance of repulsive forces as a result of hindrances of one kind or another would cause a decrease in the current i (di < 0).One of the most difficult tasks in our investigation was to obtain a gap of stable width between the surfaces of the order of magnitude of a fraction of a micron as well as fine and smooth adjustment of the gap width. Of all the methods of regulating the gap tried by us the best proved to be that making use of the same feed-back arrangement. A sufficiently fine adjustment of the gap (down to 0.01,~) was achieved by shifting the raster PI (Fig. 3) perpendicular to its lines with a micrometer. This gave rise to a moment of electromagnetic force (not compensated by gravity) which deflected the beam through a certain angle into a new position of equilibrium. The beam could be deflected to widen or narrow the gap between the sur- faces by moving the raster P in the necessary direction the equilibrium current i in the frame remaining constant until the gap became narrow enough for perceptible molecular attraction to appear.This was the sole method we used to regulate the gap width in the later models of the appar- atus which were therefore not provided with mechanisms for shifting the lens. The zero current i remains constant and equations (21) and (22) applic- Gap-width adjustment and regulation. DERJAGUIN ABRIKOSOV-4 AND LIFSHITZ MOLECULAR ATTRACTION 31 1 able only if the restoring moment of the beam itself in the absence of feed- back is negligible in relation to the moment H. It is clear that there will be no moment of the force of gravity at all when the distance between the centre of gravity and the knife edge 8 = 0. By making use of the feed- back arrangement which shortens the period of oscillation and can make it sufficiently small it is possible (contrary to an ordinary balance) to reduce s to zero.In our balance the centre of gravity wa8 situated practi- cally on the knife edge. The criterion of sufficient snidliiess of the distance s was confitancy of the current i in the frame coil for various positions of the beam over a wide range the gap width h the latter being of course wide enough to eliminate attraction between the plate and the lens. In our measurements the current i remained constant with h < 0.025 rnm. with an accuracy of 0.1-@05 p 4 which was quite sufficient for otir purposes. The distance s is connected with the period of oscillations to (in the absence of feed-back). ?Ve calculated so by experimentally determining to which proved to be approximately 6-8 sec.Values of t and s are very sensitive to the slightest change in the balance and therefore we had to control and reproduce the necessary values t = to (or s =I so) before each experiment. For this purpose we used a new glass rider C (Fig. 1) of suitable weight. Coincidence of the centre of gravity with the knife edge considerably decreases the sensitivity of the apparatus to vibrations of the support as the latter are transmitted mainly through the fulcrum. But inclinations of the support can be transmitted to the beam through a viscous eushion between the lens and the plate. Adaptability of the method. In view of the fact that the current intensity and the electromagnetic moment of the interaction between the frame coil and the magnet are in direct proportion (characterised by the coefficient n ) the sensitivity OC a balance with a feed-back arrangement does not depend on the current yield and other characteristics of the beam-tracking device.The balance secures a linear relation between the forces of interaction between the bodies (the " load ") and the current in the frame coil inde- pendently of whether the characteristic of the amplifier is linear or not. The sensitivity of such a balance can be regulated by changing the number of turns in the frame coil and the intensity of the magnetic field or by simply shunting the frame This possibility of regulating the sensitivity of the balance by a simple electric method is especially valuable. With constant sensitivity the value of the coeffioient of feed-back " rigidity " d depends on the current yield E which we were able to vary within a wide range through various parameters of the amplifier circuit by changing the cathode resistance R (Pig.2) by employing different parts of the amplifier anode Characteristic or finally by using the amplifier with a valve of a smaller slope and Iesg noise. Evaluation of the current yield in ampBres per radian was carried out as follows. The beam was held by a special arrest 5 (Fig. 5) leaving only a small gap between the plate and the lens. The change in current yield in amperes per radian corresponding to a small inclination of the beam by means of the arrest was measured The frame 2 was short-circuited. 3 12 QUARTERLY REVIEWS by the microammeter. The angle of deflection was considered equal to the ratio of the change in the gap width to the arm of the beam (the method of measuring the gap width is described below).I n investigating interaction forces which do not require a very high value of the coefficient d the current yield of the tracking device was reduced as this minimised the effect of vibrations on the beam and through it on the readings of the microammeter. When using the negative feed-back method with an ordinary analytical balance a slow creep of the current yield unlike short-period fluctuations is not injurious equilibrium being preserved by gradual deflection of the beam without disturbing the constant value of the anode current by which the load is estimated. For our apparatus both slow and rapid changes in the tracking device are dangerous as the beam has to remain in its equilibrium position long enough to enable measurement of both the gap between the surfaces and the corresponding current intensity.This is especially essential for measure- ments under atmospheric pressure as the motion of the beam causes resist- ance forces depending on the viscosity of the air and distorts the forces measured. To secure stable conditions throughout the experiment we used only storage batteries and dry cells as sources of current. The wires of the amplifier were screened and the platform on which the balance was placed was earthed. The above method can be used to measure interaction forces between solid bodies from 1-2 x up to ca. 20 dynes with a comparatively rapid decrease of the force with the gap width. Thus a foroe with a gradient of lo6 dynes per cm.can be measured with an accuracy of up to 0-02 dyne. With such a rigid feed-back (a current yield of about 500 A per radian) the period of the beam was about 5 x sec. The advantages of this balance made it possible to overcome the considerable and specific difficulties of our problem. It should be noted that self-oscillations of considerable amplitude may arise in a circuit with a feed-back arrangement. This is due to the inertia of the tracking device i.e. to the fact that the phase of the current yielded by the tracking device lags behind the deviation of the beam. This phenomenon can be avoided by merely including phase-shifting units in the amplifier circuit. When measuring molecular forces oscillations were generated only with wide gaps between the surfaces when the damping effect of the thin layer of air ceased to exist.Even when measuring in vucuo mm. Hg) this damping for narrow gaps was sufficient and we did not have to make use of the phase-shifting units. The design of the apparatus is shown in Fig. 5. A massive brass plate 1 bearing all the parts of the apparatus rests on three supports high enough to ensure free access to the adjusting screws under the platform. Directly on the platform are a plate 1' supporting the beam 2 a support Self-oscillations. Design of the apparatus. DERJAGUTN ABRTKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 31 3 for the lens 3 the mechanism regulating the niotion of the glass fibre 4 the arrest 5 a magnet 6 with a core 7 and supports for the raster relay 8. The plate 1’ rests on three fulcra two of which are adjustable by means of differential micrometric screws under the platform 1.The purpose of such an arrangement will be explained below. An agate rest ( A ) glued to the plate serves as the bearing for the agate knife edge (B) on the beam (Fig. 1). FIU. 5 The beam is 1-1 shaped in cross-section and is made of 0.16-mm. aluminium. The central notch holds the agate knife edge which is cemented in with shellac; the feed-back frame Z (Fig. 2) is cemented to the other two notches. The rectangular base of the frame is made of O.16-mm. aluminium and has ribs for rigidity. The ends of an enamelled copper wire 50 ,LA in diameter are wound on the frame (5-20 turns) and soldered to thin Wollaston wires connected to terminal posts (13) located under the mechanism for moving the glass rider.The frame is 20 mm. x 40 mm. x 2 nim. A mirror S and a quartz or glass plate P (Fig. 1) are It has three notches. 314 QUARTERLY REVIEWS fixed to the ends of the beam by means of aluminium couplingg. Fig. 6 shows the shapes of the plate and the lens. The plate 1' (Fig. 5 ) has an aperture through which the support with the lens can pass freely. The screws mentioned above serve to tilt the beam with the plate a t various angles to the lens which may be necessary to shift the contact point between the surfaces under investigation. The glass rider 9 (Fig. 5 ) used for rough balancing and for calibration lies in the groove of the beam and can be moved along it by the carrier 10. The carrier is brought into motion by turning a screw under the platform which conveys the motion to a horizontal slider 4.In order to prevent the lens from touching the plate and to separate them in case of deliberate or chance contact a special arrest 5 has been provided. Contact between the arrest and the beam is through the crossed edges of two corundum crystals in order to decrease the force of adhesion. One crystal is cemented to the left-hand coupling and the other to the plate of the arrest. The arrest is moved up and down by mepns of a differential screw under the platform of the apparatus. POhh ed PIG. 6 The magnet 6 and its core 7 rest freely on the platform of the apparatus. All the parts of the apparatus between the supports of the photorelay are covered by a low brass case. Small glazed windows have been cut opposite the lens and the mirror on the beam.In the apparatus for vacuum work an air-proof rubber gasket is inserted between the case and the platform and all the micrometer screws regulate the apparatus through grommets soldered to the underside of the platform. The support 8 holds the raster relay (see Fig. 5 ) which is assembled inside a brass tube 40 nim. in diameter and 300 mm. long. A picture of the apparatus (with the cover removed) is given in the Plate. The force of interaction between the lens and the plate is calculated from the equation BaZanee cazibration. F = ni/R = y i . (21) The methods of meaauring the current i and R were described above; it remains now to explain how the coefficient n was determined. Calibration of the balance (or the determination of n) was carried out by using the glass rider (C Pig. l) which could be moved along the beam.314 QUARTERLY REVIEWS fixed to the ends of the beam by means of aluminium couplingg. Fig. 6 shows the shapes of the plate and the lens. The plate 1' (Fig. 5 ) has an aperture through which the support with the lens can pass freely. The screws mentioned above serve to tilt the beam with the plate a t various angles to the lens which may be necessary to shift the contact point between the surfaces under investigation. The glass rider 9 (Fig. 5 ) used for rough balancing and for calibration lies in the groove of the beam and can be moved along it by the carrier 10. The carrier is brought into motion by turning a screw under the platform which conveys the motion to a horizontal slider 4. In order to prevent the lens from touching the plate and to separate them in case of deliberate or chance contact a special arrest 5 has been provided.Contact between the arrest and the beam is through the crossed edges of two corundum crystals in order to decrease the force of adhesion. One crystal is cemented to the left-hand coupling and the other to the plate of the arrest. The arrest is moved up and down by mepns of a differential screw under the platform of the apparatus. POhh ed PIG. 6 The magnet 6 and its core 7 rest freely on the platform of the apparatus. All the parts of the apparatus between the supports of the photorelay are covered by a low brass case. Small glazed windows have been cut opposite the lens and the mirror on the beam. In the apparatus for vacuum work an air-proof rubber gasket is inserted between the case and the platform and all the micrometer screws regulate the apparatus through grommets soldered to the underside of the platform.The support 8 holds the raster relay (see Fig. 5 ) which is assembled inside a brass tube 40 nim. in diameter and 300 mm. long. A picture of the apparatus (with the cover removed) is given in the Plate. The force of interaction between the lens and the plate is calculated from the equation BaZanee cazibration. F = ni/R = y i . (21) The methods of meaauring the current i and R were described above; it remains now to explain how the coefficient n was determined. Calibration of the balance (or the determination of n) was carried out by using the glass rider (C Pig. l) which could be moved along the beam. DERJAGTJIN ABRIKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 31 5 With a wide gap between the surfaces (in the absence of moleciilar attrac- tion) the current intensity i in the frame was measured by an ammeter lor various positions of the rider registered by means of a microscope with im ocular scale.The condition of equilibrium can be expressed thus FVAW = nAi0 where W is the weight of the rider Aur its travel and A; the corresponding current intensity variation. The coefficient in question equals For the determination of n the co-ordinates of the end of the glass rider Fig. 7 shows one of ths The weight of the rider W = 46.60 mg I Number of turns in the frame coil N = 15 n = WAw/Ai were plotted against the current i on a graph. I 1 0 0.5 7 - 0 i (m4 F I G . 7 calibration graphs in which n = 2.51 mg. cm./mA. arm R averaged 1.9 em.with a corresponding balance sensitivity of y == n/R = 1-32 mg./mA In our experiments the The coefficient n measured in this way agreed well with that deduced according to AmpBre’s law. M = BSN sin (2. K)i where B is the induction of the magnetic field N the number of turns in the frame coil i, the current S the frame area in the magnetic field K the unit vector of the normal t o the plane of the frame. The constant to be determined n is the proportionality coefficient between M and i ; thus The torque of the frame M, equals n = BNS sin (B.K). In our experiments B e 850 gauss ij’ = 2 sin (B.K) 1. With h’ = 15 turns we get n = 850 x 2 x 15 = 25,500 dyne cm./weber n = 2.55 dyne cm./mA or which agrees well with the results of direct calibration. 316 QUARTERLY REVIEWS The accuracy of the above calculation is unquestionably lower than that of determining n from it calibration graph *and it can be used only to check for gross errors in calibration.The balance was calibrated with the circuit prepared for experiment before and after measurements the current being measured always with the same instmiment. Thus all the factors which could affect the division value and sensitivity of the balance during measurements were in action and autonzatically taken into consideration during calibration. The insignificant scattering of the points on the calibration graph (Fig. 7) ensured determination of n with an accuracy of & 0.4%. The value of I2 was measured by rule and equalled 19 & 0.5 mrn. As a result of this the error in determining the sensitivity was & 3% (y = 1-32 & 0.04).The accuracy of determination of P depended almost entirely on the errors in current measurements as the vibrations of the stand made the value of the latter fluctuate in tlze most favourable cases within & 0.1 ,MA while the absolute value of the current intensity was between 0.2 and 2 PA. The lower limit of measurable forces was determined by these external vibrations which caused an error of & 0 . 1 ~ = & 0.13 pg. m 0-13 x dyne. method of Measuring the Gap Width.-The least width h of the gap between the lens L and the plate P was calculated from the diameters of the Newton rings measured through a microscope with an optical scale. The system was illuminated by a cinema lamp (300 w) through a mono- chromator of constant deviation and the vertical illuniinator of the micro- scope which provided normal incidence of tlze light on the surface o€ the FIG.8 plate P (Fig. 8). There was no difficulty in establishing the relation between the gap widfih h the diameter of the mth dark ring d, the wavelength of the monochromatic light A and the radius of the spherical surface p. The condition for the formation of the mth dark ring is 28 $- 2h + A/2 = (2m + 1)A/2 . . (23) Out of geometrical considerations the part of the thickness of the air DERJAGUTN ABRIKOSOVA AND LIFBHTTZ ; MOLECULAR ATTRACTION 31 7 gap designated 6, (Fig. 9) is related to the ring diameter am by the equation A = dm2/0p. FIG. 9 Substituting d for 6 in equation (23) we i ntersurface distance h from which it follows that m d, I and p must obtain for the minimum - (24) be known to determine h.To find the number of the ring m the surfaces may be brought into contact while watching the ring after which its ordinal number can be counted. For measuring the molecular attraction between bodies this method is inconvenient because of the possibility of contact electrification. Therefore we employed a different method based on the changes in diameter of the interference ring with variations of m and of the wavelength of mono- chromatic light 3 with a constant gap width h. Let us introduce the values A(dm2) and A(dA2) determined by the fo!Iow- ing equations and A(dm2) = dl+dil,m2 - d ~ ~ m ~ - - (25) A(dn2) = dl,m+dm2 - dl,rn2 * . (25’) Taking into consideration equation (24) we have A(dm2)/AA = 4 p m . * (26) and A ( d A 2 ) / A m = 4pI .. (26’) Dividing equation (26) by equation (26‘) we obtain Before commencing our main experiments i .e. before measuring the attraction force F and the corresponding gap width h we would determine the number of one of the rings m by use of equation (27) and from it the numbers of the rest of the rings. 318 QUARTERLY REVIEWS The radius of the spherical surface p was determined with the same optical set-up. Equation (24) shows a linear relation between dm2 and m with constant 3L and h. If we plot dm2 against m the tlangent of the angle between the corresponding straight line and the m axis divided by 4A gives the radius p. As A m and p were always determined before the main measurements the latter reduced to measurements of the current i and the diameter of one (occasionally 2 or 3) interference rings.This procedure contributed to the success of the experiment as it allowed us t o concentrate our attentlion only on two measurements. The accuracy of measurement of the gap h was almost completely dependent on the error in measuring the diameter of the wcth (usually the second) ring dm this error (in the case of d,) being & 1% which gave an accuracy of & 0.01 ,u in determining the gap width. Cleaning the Surfaces.-Of great importance is the method of cleaning the glass and quartz surfaces used in the experiment. For a successful experiment they must be absolutely free from films or dust particles of any kind. The usual methods of chemical cleaning as for instance rinsing with chromic acid mixture were not used as they sometimes damage the polished surface of the glass.The plate and lens were swabbed with cotton wool degreased in a Soxhlet appar- atus and soaked in distilled alcohol and ether and were then treated with a glow discharge i n vacuo. The cleanliness of the surfaces could be checked by the fact that after treatment the glass surfaces were totally wettable by water. Of all the devices tried by us such as dusting with a degreased soft brush cotton wool cloth chamois etc. the best was wiping the surfaces (after treatment with glow discharge) with degreased cotton wool slightly moistened in pure ether (not to make the surfaces cleaner of course). After such treatment the surfaces were clean and were totally wettable with water just as immedi- ately after treatment in the glow discharge. Before being placed on the balance the surfaces were examined under a binocular microscope with a magnification of 100.Principal DifIiculties of the Experiments.-The greatest difficulties were connected with (i) the sensitivity of the apparatus to vibrations of the stand (ii) contamination of the surface under investigation with dust and (iii) electrification of the surfaces during cleaning. (i) The practical sensitivity of the apparatus for the determination of interaction forces depends greatly on the vibrations of the stand. A special study of the vibrations and of the effect of various shock-absorbing devices was therefore undertaken. Vibrations of the stand give rise to current fluctuations in the feed-back circuit By consecutively eliminating each of the sources of vibrations we established that they were mainly of industrial and traffic origin trans- mitted through the ground.I n order to estimate the shock-absorbing We used the following method to get thoroughly clean surfaces. Elimination of dust particles proved to be much more difficult. DERJAGUIN ABRIKOSOVA AND LIE'SICITZ MOLECULAR ATTRACTION 319 effect of vasious types of stand t'he balance was excluded and the beam replaced on the platform by a large mirror (Fig. 5). This made a sensitive photoelectric traiismitter of t,he vibrations of the stand. As both while preparing and carrying out the experiments we could not help touching the apparatus shock-absorbing devices similar to those used for sensitive galvanonieters were not suitable for our purposes. The best results were obtained when the apparatus was placed on a special table supplied with shock absorbers (1 1 Pig.5 ) placed on a cement pillar sunk into the ground and isolated from the house foundation (12 Fig. 5). in order further to eliminate the influence of vibrations the optica.1 arrangement of the beam-tracking device was improved. The light reflected from the mirror 8 (Fig. 10) on the beam was again reflected from a mirror i "f I FIG. 10 8 fastened to the platform 1 (Fig. 5 ) which made the photocurrent inde- pendent of the oscillations of the beam together with the platform so that it depended only on the gap width h. The angle between the mirrors XI and 8 was approximately 90". could be regulated when adjusting the nppa.ratus by meaiw of a screw located uncler- neath the platform. (ii) The treatment of t,he surfaces before the measurements is very important as during elimination of the dust particles the surfaces become electrically charged and hence interact much more strongly than by mole- cular attraction thus camouflaging the latter.To remove the charges from the surfaces of the plate and lens it was necessary to enlarge the gap and to ionise the air surrounding the apparatus by some method. This often led to dust particles from the air settling on the ,surfaces which made it necessary to clean and discharge them again and again until both the dust particles and the electrostatic interaction were completely eliminated. Experience has shown that when the gap is very small dust particles do not get into it so that it is necessary to obtain a simultaneous absence of dust and charges on the surfaces only once.After that the surfaces should not be separated by more than 5-20p. (iii) Electric charges arise on clean dry surfaces very easily even by slight contact with a clean brush or a rubber glove. Each time before we could measure the molecular attraction it was necessary to eliminate the The position of the niirror 320 QUARTERLY REVIEWS surface charges by ionising the atmosphere. To discharge quartz surfaces we employed a radioactive sulphur isotope (35S). It was found impossible to eliminate the charges when the gap was very narrow. Only with very wide gaps (1 mm. or more) between the surfaces was the action of the ioniser effective ; this proved the electrostatic origin of the attraction forces hit herto observed. FIG. 11 The slow removal of the charges in case of a narrow gap is evidently due to the fact that too few ions penetrated into it.The difference between curves I and I1 in Fig. 11 is a result of high ionisation of the air in the gap. As we have already noted the sensitivity of the apparatus for measuring attraction forces is determined almost entirely by the current fluctuation in the feed-back circuit caused by the vibrations of the stand. Therefore a t first before the measures against vibrations described above had been taken the sensitivity was insufficient for the detection and measurement of the forces of molecular attraction. After elimination of the charges no interaction between the surfaces could be detected down to gap widths of It = 0.05 ,u the accuracy of the measurements being & lob2 dyne. Results The main result of the investigation was the detection and quantitative measurement of molecular attraction between surfaces of fused quartz with a gap width of 10-5-10-4 em.Fig. 12 and Table 1 demonstrate the dependence of the attraction force F on distance 11 measured under atmospheric pressure. The radius p of the spherical surface was 10 em. This dependence of F on I/ corresponds to the minimum attraction effect of all those observed by us for quartz glass. This fact together with the sufficiently good reproducibility of the effect in various experiments gave grounds to assume it to be of a molecular character. But in order to be DERJAQUIN ABRIICOSOVA AND LIPSHITZ MOLECULAR ATTRACTION 32 1 TABLE 1. Measurem>ents under atmospheric pressure h(P) tr 103F (dynes) h(P) 103F (dynes) h(V) 103F (dynes) 0.09 0.13 0.13 0-13 1.5 0.14 0.7 0.19 0.9 0.16 0.8 0.17 0.7 0.7 0.5 0-3 cluite sure of our conclusion we had to prove the absence of outside effects on our experiments.The most important hindrance in measuring mole- cular attraction was electrostatic interaction of the measured objects. It might therefore be assumed that the dependence represented in Fig. 12 i d u d e s an electrostatic component and that the molecular interaction sought is only a part of that measured in the experiment. 0.32 0.1 0.59 0-0 0.64 0.0 0-82 0.0 2.0 c FIG. 12 But if the electrostatic charges have been eliminated completely and the attraction observed is truly a molecular attraction then it must be ( i ) not affected by repeated ionisation of the surrounding air ; (ii) propor- tional to the radius of the spherical surface [see equation (9)] ; (iii) repro- ducible both in value and in the law by which it decreases as the gap width is increased from experiment to experiment ; (iv) reproducible in experi- ments where different parts of the surfaces are brought together; and (v) unaffected by the removal of the air from the gap.Our further experiments were aimed a t verification of the fulfilment of the above conditions in the case of the attraction which we had assumed to be molecular. The most accurate and reproducible proved to be the experiments in z'acuo. The results of these experiments are given in Fig. 13 and Table 2. (hrve I corresponds to a radius of p = 10 em. and curve I1 to p = 26 cni. Within the range of the error of measurement these results quite satisfy t lie conditions listed above.Fig. 14 gives a bilogarithmic representation of a large number of experi- ments carried out a t considerable time intervals with various quartz surfaces. 322 QUARTERLY REVIEWS TABLE 2. Measurements in vacuo I p = 10 em. I p = 26 cm. h (cc) 0.08 0.10 0.110 0.13 0.15 0.16 0.17 0.18 0.20 0.42 0.64 0.96 103F (dynes) 1.9 2.0 1.3 0.9 0.5 0.72 0.46 0.5 0 . 2 0 0 0 h ( P ) 0.13 0-14 0.17 0.18 0.20 0.22 0.25 0.28 0.31 0.42 0.62 0.71 0*9G 1OsF (dynes) 3.1 2.4 1.5 1.5 1.3 0.6 0.4 0 0 0 0 1.0 0.2 The lens radius was 11.1 cm. The open circles represent the valuek obtained a t atmospheric pressure. Before almost all measurements the surrounding air mas sub jectecl t o the repeated action of powerful ionisers 0 FIG. 13 FIG. 14 before being pumped out (if required).wag calculated according to Lifshitz's theory ; see below.) attraction energy of two infinite plates per cm.2 zc(h). full circles corresponding to the experiment with the lens radius p = 11.1 cm. (The straight line on the figure According to equation (20) the quotlient F(h)/Bnp is equivalent to the Fig. 15 shows the dependence of the energy u on the gap width h the DERJAGUIN RERTKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 323 the triangles to that with p = 10 cm. and the open circles to that with 3 = 25.4 em. This graph illustrates the linear relation between the attraction force and the sphere radius and thus shows that the experimentally obtained Oo L 0 4Lr-co-o 0.8 7.0 attraction energy for parallel plates u(h) does not depend on the nature of the lcns used in the experiment.* Thus all the conditions listed above are satisfied.Discussion Considering that the investigation was devoted to the detection and measurement of an effect the existence of which had not been proved previously by direct experiment we find it necessary to give a detailed analysis of the results of our measurements and their interpretation. Two questions should he disciisscd (i) it must be proved that the effect obtained was not due to any defects of method and (ii) it must be proved that the attraction between bodies measured in our experiments is independent of any non-moleculnr interaction. (i) The coincidence of the results obhined in vacuo and in the presence J. T. G. Overbeek and M. J. Sparnaay J . Colloid Sci. 1952 7 343. l8 N. Fuchs 2. Physik, 1934 89 736. l9 J. B.Petrjanov Acta Physiochim. 1942 IS 185. * We have not been able as yet to increase the range of different radii used as with p < 10 cni. the attraction forces are weak and with p> 26 cm. it is very difficult to avoid dust particles and tt higher vacuum is required owing to the high damping action of the air in the gap. x 324 QUARTERLY REVIEWS of air proves their independence of convection currents radiometric effect the presence of a viscous air layer between the bodies and of the water vapour present in tJhe air. The necessary precautions were undertaken to avoid errors of measure- ment due to any unforeseen mechanical influence of tJhe various parts of the apparatus such as the elastic effect of the wires supplying the current to thc frame coil friction between the knife edge of the beam and its bearing dust particles remaining on the surfaces under investigation etc.The elastic action of the wires was reduced to a minimum by the use of very thiri Wollaston wires and by annealing them in a Bunsen flame. The agate prism and bearing conformed with the requirements for the best grades of microniialytical balances. In a good aiialyticrtl balance with the beam and pans weighing several score grams friction is no impediment to attaining ail accuracy of to g. As we know friction is approximately pro- portional to the load. It can easily be understood therefore that in our balance the beam of which weighed 0.1 g. the friction between the knife edge and the bearing could be completely neglected in weighing with an accuracy of 1 or 2 x The presence on the surfaces of particles capable of affecting our measure- ments was always revealed by the appearance of repulsion forces upon narrowing the gap.These forces could be registered with our apparatus with the same sensitivity as attraction forces and they never changed smoothly with the gap width. Measurements were performed only when no forces except those of attraction were observed with gap widths down to We also believe that the attraction observed cannot be due to any film left on the surfaces after cleaning. The presence of adsorbed films of water unavoidable under any conditions whether in a low vacuum or a t atmo- spheric pressure did not affect the results of the measurements as first the distance between the surfaces was much larger than these films and secondly the dielectric constant closely connected with the value of the molecular attraction is approximately the same for adsorbed films as for quartz.2o If the dielectric constant of the film is close to that of the quartz the presence of an adsorbed film of approximately 10 A is equivalent t o a change in the gap width between the quartz surfaces of the same order i.e. 10 A which when measuring h with an accuracy of ca. 100 A cannot affect the results in any way. The graph in Fig. 12 for instance represents experiments in which the dust particles were removed by means of a degreased brush and those in Pigs. 13 and 14 with degreased cotton wool moistened with ether. If we were to admit that the attraction is due to films then the coincidence of the results of many experiments would lead us to assume the presence of identical filnis in all the experiments which is very improbable.(ii) The absence of any possible influence on the results of the experi- mcnts of contact charges on the quartz surfaces was verified experimentally. 2o F. Keurbatow Zhur. fiz. Khim. 1954 28 287. g. 0.05-0. 1p. Various methods of cleaning were used in different experiments. DERJAGUIN ABRIKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 325 It can be seen immediately without computation that the attraction forces represented for instance in Fig. 12 are not gravity forces as the latter cannot change so abruptly with such minute changes in the gap width in relation to the distance between the centres of gravity of the bodies. Magnetic forces are out of the question in the case of quartz specimens. As to forces of interaction connected with the electrical charges of the bodies they have been dealt with in detail above.Comparison with theory Comparison with calculations obtained by summhag interactions between eac7i pair of molecules. If we follow the current method of summing the interactions between each pair of molecules then for a sphere and a plane surface the following formula should be applied F = A p / 6 7 ~ ~ Substituting the results of our experiments in this formula we find the value of the constant A to be about 5 x lod1* erg. This constant for quartz however is approximately 10-12 erg or 20 times greater than the value obtained by experiment. This shows that the methods l1 of calculation hitherto current are quite unacceptable at least for distances of the order of em.It may therefore be concluded that the results of our experiments are of universal significance in spite of the fact that only one kind of material was used. If we apply the same method of summing with allowance for Casimir and Polder’s corrections the energy u( h ) should be calculated according to the formula u = - A’/30nh3 Substituting the results of our experiments in this formula we obtain for the constant A’ a value of approximately 3 x 10-18 erg cm. Calcula- tion according to thc formula ,4‘ = n2q2C’ gives A’ = 1 x 1O-l8 erg cm. (The polarisability 61 was taken from Margenau’s paper.9) Here again experiment does not agree with theory but the discrepancy is much smaller than those obtained by summation of London interactions. Comparison with the Lifshdtz theory. A precise comparison with the Lifshitz theory requires a sufficient knowledge of the optical characteristics of the substance in its absorption zones without which the function p(it) cannot be deduced.Nevertheless the nature of absorption in quartz per- mits an approximate theoretical estimation. Quartz is highly absorbent in the ultraviolet (beginning a t about 0.15 p) and in the infrared (beginning i L t several microns) region between which it is transparent. The results obtained in the experiment lie in the transparent zone and for the sake of estimation I& may be considered small in comparison with the value of A/2n of the low-frequency and great in comparison with the value of ;1/2n for the high-frequency boundary of absorption. The share of the ultra- violet absorption zone in the force f can be estimated by formula (15) I, being assumed equal to the coefficient of the square of the refractivity in the zone of optical- transparency.The share of the infrared zone is given 326 QUARTERLY REVIEWS by equation (11) ; in the order of its value it is hcu,/c times smaller (where coo are the infrared absorption frequencies) so that in a rough estimate off this value can be neglected. In order to compare the theoretical data with the values directly measured by experiment let us perform the following modifications. Inte- grating expression (15) for f ( h ) we get the following equation for the energy of attraction between two plates per em2 area In Fig. 15 the broken line represents the relation u(h) calculated by the above equation E~ being assumed equal to the square of the coefficient of refraction in the optical region.In order to pass from the energy u(h) to the force of interaction F(h) between a sphere of radius p and a plane surface we shall use the formula Fig. 14 gives the relation F(h) calculated in this way for p = 11.1 em. in a bilogarithmic scale. Taking into consideration thc roughness of such an estimate and the errors of measurement the agreement may be considered satisfactory. The coincidence of experiment with theory revealed in the graphs (Pigs. 14 and 15) should be considered as a confirmation of the Lifshitz theory on the one hand and as substantial evidence of the molecular nature of the measured experimentally attraction effect. The coincidence of experimental results with the theory which explains molecular interaction as interaction between electromagnetic fields existing in any absorbing medium and extending beyond its boundaries provides an answer to the question put by Y.N. Lebedeff in 1894 as to whether molecular attraction really reduces to electromagnetic forces “ and does not include other forces of hitherto unknown origin ”. It should be pointed out that the experimental results obtained by Overbeek and Sparnaay 17 21 differ from ours. They measured the attrac- tion forces between two plates of fuzed quartz by means of a special dynamometer in which the spring movements were measured by the capaci- tance method. The gap width between the plates was established by the interference colours in a narrow slit. At a discussion held in 1954 a t the Faraday Society they presented their paper 21 at the same time as we did.The results obtained by the Dutch investigators were presented in the form of a graph showing the dependence of the force on the gap width in a logarithmic scale and a calculation of the experimental value of the constant A which they found to equal 3.8 x 10-11 erg. If calculat’ions are carried out according to the Lifshitz theory for a gap width of 1200 A between two quartz plates the attractive force will be found to equal approximately 2 x dynes/cm.2 while in Overbeek and Sparnaay’s experiments the force for this gap width was 1 dyne. Thus their experimental data exceeded theory by the order of 10,000 times. F(h) = 2npu(h) 21 J. T. G. Oswrbeek and &I. J. Spamaay Discuss. Paraday SOC. 1954 18 12. DERJAGUIN ABRIKOSOVA ANP LIFSHITZ MOLECULAR ATTRACTION 327 The low reproducibility in Overbeek and Sparnaay 's experiments and the excessively high value of the attraction eff'ect observed is probably due t.0 electrical surface charges.Application to the theory of coagulation According to the theory of N. Fuchs l8 the rat'e of coagulation of a disperse system the particles of which of radius a are attracted with an energy U ( r ) depending on the distlance r between their centres increases in comparison with the case U(r) = 0 considered by Smoluchowski by a number of times given b y * exp[- U(r)/kT]dr = 1 * exp[ U(z)/kT]dz 1 2aS 2a r2 where z = ( r - Za)/2a. If a is sufficiently small (compared with the wavelengths of the chief bands in the absorption spectrum of the particles) then for values satis- fying the condition r - 2a < Za U may be expressed by the formula 1 A 24(r - Za) 24' z - __ - - A.2a U = Considering that A/24 and kT are usually values of the same order it is obvious that the coefficient of acceleration of coagulation is perceptibly greater than unity,* as U / k T will have a tangible value for a sufficiently large range of values of a beginning with zero.At the same time the coefficient of acceleration will not depend on a. But if a becomes sufficiently large the result will be different. With sufficiently large values of a the main part of the integral in equation (27) corresponds to values of Y which require correction for electromagnetic retardation which will lead to a decrease in U and therefore in the accelera- tion of coagulation. That is why for instance in aerosols with particles for which a > @3p the acceleration of coagulation under the influence of molecular forces will be very small.It is different with the coagulation of lyophobic sols in which repulsion energy plays a part alongside that of attraction. The repulsion energy is due to the overlapping of the ionic atmospheres of two particles. In this case for sufficiently large particles both components of the interaction energy ( a t the gap widths where they become perceptible) are proportional to the radius. Therefore the absence or presence of an energy barrier to t'he resulting interaction which determines the stability of the system in prac- tice will not depend on the radius of the particles but on the laws of decrease of both components of interaction energy with the gap width.Obviously with thin ionic atmospheres (i.e. with medium and high concentrations of electrolyte) less than 10-6 cm. thick only the behaviour of molecular forces at gap widths not requiring correction for electromagnetic retardation is significant. * This conclusion has been formulat'ed earlier in eqiiation ( I 9). 328 QUARTERLY REVIEWS Thus the previously developed theory of sol stability 1’ remains valid in particula,r the Hardy-Schulze sixth-power law of the coagulating action of the charges on counter-ions. The applicability of the law is therefore not limited by the radius of the particles,* but by their concentration. With very low concentrations considering the more rapid decrease (per power of the gap width) of the attraction forces a t long distances it is easy to demonstrate that the sixth-power rule must be substituted by an eighth- power rule.As very low coagulating concentrations can be observed only for heavily charged ions (ter- and quadri-valent) the corresponding effect should be expected only in these cases. Conclusions 1. A method has been developed which permits measurement of the inter- action forces between very smooth transparent solid substances depending on the width of the gap between them. The force of interaction is measured by means of a special beam-type microbalance with photoelectromagnetic negative feedback. The gap width between the bodies is calculated from the diameter of Newton’s rings. The measurable force range is froin 1-2 x to 20 dynes and the gap widths from to em. 2. The molecular attraction between two specimens of quartz glass has been detected and measured.The attractmion energy per between two plates u(h) changes with gap width h according to an inverse cube law and equals about 1 x erg when h = 1-5 x 3. It has been shown experimentally that the force of attraction between a spherical and a plane surface is proportional to the radius of the sphere ; this agrees with the conception of the molecular nature of these forces. 4. The investigation described i s the first direct experimental verifica- tion of the theories of molecular attraction between two condensed bodies. 5 . An analysis of the modern theories of molecular attraction has been given. It has been pointed out that the usually accepted assumption of the additivity of molecular attraction forces in condensed media has neither a theoretied nor an experimental foundation.6. It has been proved by experiment that molecular attraction between bodies large compared with molecular dimensions cannot be calculated by summing the interaction of each pair of molecules found by the London formula if the distance between the surfaces is em. or more. The best results are obtained when allowance is made for electromagnetic retarda- tion upon the propagation of the respective forces. This proves P. N. Lebedeff ’s hypothesis concerning the electromagnetic nature of molecular forces. 8. As the London-Hamaker inverse-square law for the energy is a par- ticular limiting case of Lifshitz’s theory for short gap widths a t which electromagnetic retardation plays no part the confirmation of this theory shows that this law is applicable to short distances.* For this reason we cannot agree with Overbeek’s assumption that for large particles em. 7. Our experimental results agree with E. M. Lifshitz’s theory. this law requires correction for electromagnetic retardation. DERJAGUIN ABRIKOSOVA AND LIFSHITZ MOLECULAR ATTRACTION 329 9. It can easily be shown that the deviatioii from this law towards a decreasc observed for greater distances ( em.) points to the insignifimnt influence of molecular forces on the rate of coagulation of aerosols with particles larger than 3 x 10. The results obtained confirm the existence of long-dist,ance mole- cular surface forces which is one of the €oundation stones of the current theory of stability and coagulation of colloids. 11. The results obtained show that deviations froin the sixth power-of- the-charge law in the Hardy-Schultze rule should be expected only with very low concentrations of electrolytes with highly charged counter-ions. 12. It has been pointed out that the values of intermolecular attraction hetween bodies obtained by Overbeek and Sparnaay exceed both the theoretical values and those obtained by us by 3 or 4 orders of magnitude which is apparently due to the fact that their measurements were influenced by effects other than molecular forces. em.

ISSN:0009-2681    

DOI:10.1039/QR9561000295

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

TETRA- AND TRI-CHLOROALKANES AND RELATED COMPOUNDS By A. N. NESMEYANOV 1%. KH. FREIDLINA and L. 1. ZAKIIARKIN (U.S.S.R. ACADEMY OF SCIENCES Rloscow) THE present Review is a brief account of investigations carried out by the authors during recent years in collaboration with Ye. J. Vasil’eva R. G. Petrova V. N. Kost Sh. A. Karspetyan N. A. Senienov A. B. Belyavsky and T. A. Kost on reactions of polychlorohydrocarboiis. In these investigations we were chiefly concerned with the changes in 27 A. N. Nesmeyanov R. Kh. Freidlina and V. I. Firstov Dolclady d k a d . Nauk 2 A. N. Nesmeyanov R. Kh. Freidljna and L. I. Zakharkin &id. 1951 81 199. 3 4 . N. Nesmeyanov R. Kh. Freidlina and V. I. Firstov Izvest. Aknd. Nuuk 4 A. N. Nesmcyanov and L. I. Zakharkin ibid. 1953 988. S.S.S.R. 1951 78 717.S.S.S.R. Otdel. khim. Nauk 1951 505. A. N. Nesmeyanov L. I. Zakharkin and R. Kh. Freidlina ibid. 1954 34. A. N. Nesmeyanov L. I. Zakharkin and It. G. Petrova ibid. p. 253. A. N. Nesmeyanov L. I. Zakharkin V. N. Kost aiicl R. Kh. Freidlina ibid. Idem ibid. p. 604. A. N. Nesrneyanov R. Kh. Freicllina and L. I. Zalrherkin DolcZudy Akacl. Nuuk p. 258. S.S.S.R. 1954 96 87. 10 Idem ibid. 1954 97 91. 11 Idem ibid. 1954 99 781. l2 A. N. Nesrneyariov L. I. Znkharkin and R I-Zh. Freidlina Izvest. AEad. Nnzck 13 A. N. Nesrneyanov arid 1,. I. Zakharkin ibid. p. 224. 14 R. IZh. Freidlina V. N. I h s t arid A. N. Nesmeyanov ibid. p. 233. l5 R. ICh. Froidlina and Ye. I. Vasil’eva Doklccdy Akad. Nuuk S.S.S.R. 1955 l6 A. N. Nesmeyanov L. I. Zakharkin aiid T. A. Kost I u e s t . Akad. Nut& S.S.S.R.l7 A. N. Nesmeyanov V. N. ICost and R. Kh. Freidlina Doldudy Akud. Naulc la A. N. Nesmeyanov R. Kh. Freidlina and V. N. ICost, Izvest. d k a d . NuuX S.S.S.R. l9 L. I. Zakharkin ibid. 1955 1009. 2o (a) A. N. Nesmeyanov R. Kh. Freidlina and N. A. Semenov ibid. p. 003; 21 L. I. Zakharkin Dolclady Alcad. Nauk S.S.S.R. 1955 105 985. 2 2 A. N. Nesmeyanov R. Kh. Freicllina L. I. Zakharkin and A. B. Belyavsky 23 L. I. Zakharkin Izvest. Akad. Naulc S.S.S.R. Otdel khim. NU& 1956 314. 2 4 A. N. Nesrneyanov Sh. A. Karapetyan and R. Kh. Freidlina Dolclady Akad. Nauk. S.S.S.R. 1956 in the press. ( a ) A. N. Nesmeyanov R. Kh. Freidlina and L. I. Zakharkin U.S.S.R. Pet. 98449/1054 ; ( b ) R. Kh. Freidlina and L. I. Zakharkin U.S.S.R. Pat. 99484/1954 ; S.S.S.R. Otdcl. khirn. Nauk 1955 40.100 85. Otdel. khim. Nauk 1955 657. S.S.S.R. 1955 103 1029. Otdel. khim. Nuuk in the press. ( b ) R. Kh. Freidlina and N. A. Semenov ibid. 1956 in the press. Zhur. obshchei Khim. 1956 26 130. 330 N E8P;IEYANOtr et U l . TETRA- AND TRI-CIILOROALKANES 33 1 ~cc~~-tetl.achloroalkanes and orcccc-trichloroalkanes which became readily available by the telomerisatioii of ethylene and carbon tetrachloride or ethylene and chloroform a reaction due to Joyce Hanford and Harmon. 28-30 In a number of cases we have investigated polyhalogeno-derivatives obta,ined by adding carbon tetrachloride or halogeno-derivatives t,o olefins and t o vinyl ethers as well as by condensing halogeno-derivatives with halogeno-olefins in the presence of aluminium chloride. Our aim has been to work out general methods of synthesis of various organic compounds starting with those involving for example the following radicals The investigation also involved the examination of some rearrangements in the series of unsatnruted polychlorohydrocarbons.In the course of our investigation we have synthesised a great number of substances sorne tlata being listed in Tables 3-9. Reactions of the Trichloromethyl Group in Saturated Compounds The Determination of the Character of the Trichloromethyl Group as an Orientant.-The inveatigation of the orienting action of the trichloromethyl group on the electrophilic substitution in the aromatic nucleus has led to ambiguous results. Thus tlhe trichloromethyl group in beiizotrichloride orients to the meta-position in nitration but to the Pam-position in chlor- i n a t i ~ n .~ ~ Kharaach and his co-workers 32 failed to determine the orienting influ- cnce of the trichloromethyl group on the electrophilic addition of hydrogen bromide to 3 3 3-trichloropropene7 these authors having dealt in their investigation with 1 1 2-trichloroprop- 1 -em mistakenly thought by them to be 3 3 3-trichloroprop-l-ene (see p. 339). Study of the reaction of the true 3 3 3-trichloropropene with hydrogen bromide showed that the reaction does riot take place in the absence of catalysta and i)hat when aluminium trichloride is present 3 3 3-trichloro- prop-l-ene is isornerised to 1 1 3-trichl0roprop-l-ene.~-~ Two of 11s and V. N. Kost have studied the conjugated addition of chlorine to 3 3 3-tlrichloropropene in gla,cial acetic acid or concentrated sulphuric acid 2 3 3 3-tetrachloropropyl acetate having been obtained ( c ) A.N. Ncsmeyanov R. Kh. Freidlina L. I. Zakhrkin et al. Trudy Vscsoyuz Soveshch. Kornpleksnoi Percrab. Naft. Gazov ; ( d ) It. Kh. Freidlina arid Ye. I. Vasil’eva. [J.S.S.R. Pat. Appl. ; (e) R. Kh. Preidlina and L. I. Zakharkin U.S.S.R. Pat. 100341/ 1955. 26 A. N. Nesmeyanov R. Iih. Freidlina and R. G. Petrova Izvest. Alcad. Nauk S.S.S.R. Otdel. khirn. Nauk 1956 in the press. 27Ye. I. Vasil’eva and R. Kh. Preidlina ibid. p. 177. 2 8 R. M. Joyce MT. F. Hanford and J. Harmon J. Arner. Chern. Xoc. 1948,70,2429. 2v R. M. Joyce and W. F. Hanford ibid. 1950 72 2213. 30 W. F. Hanford and R. M. Joyce U.S.P. 2,440,800/1948. 31 W. M. Latimer and C. W. Porber J . Amer. Chem. SOC. 1930 52 206. a2 M. 8. Kharasch E.Rossin and E. K. Fields ibid. 1941 63 2558. 332 QUARTERLY REVIEWS in the former and a corresponding sulphate in the latter. The structure of the acetate and sulphate were proved by hydrolysis to 2 3 3 3-tetra- chloropropaiiol identical with the alcohol synthesised according to the following scheme k8 AcOIf CH,.OH CCl,:CHCH,Cl + CH,CO,K > CCl,:CH-C€-I,.OAc --+ C1 CCI,:CHCH,.OH --+ CCl,*CHCl*CH,.OH HCI Carrying out the reactions of conjugated addition of chlorine to propene in glacial acetic acid or concentrated sulphuric acid gives the corresponding esters of 1-chloropropan-2-01.~~ Comparing the reactions one finds the orienting action of the trichloro- methyl group to be opposite to that of the methyl group the electron- attracting character of the trichloromethyl group being thereby proved CH,*CH:CH + C1 --+ CH,.CH(OAc).CH,CI -+ CH,*CH(OH).CH,Cl CC1,CH:CHz + C1 - * CCI,.CHCl*CH,*OAc + CCl,.CHClCH,-OH -4cOH BcOH 3 3 3-Trichloropropene also adds hypobroinous acid in the reverse order to propene CH,:CH-CCl + HOBr + HOCH,*CHBr*CCl CH,:CH*CH + HOBr -+ BrCH,*CH(OH)*CH The structure of 2-bromo-3 3 3-trichloropropanol was proved by dechlorination with alcoholic alkali to 2-bromo- 1 1 -dichloro-3-hydroxy- prop- 1 -ene.The electron-accepting inductive effect of the trichloromethyl TABLE 1. Dissociation constants in water. 1 Acid CCl,*[CH,],*CO,H . . . . . . CH,*[CH,],.CO,H . . . . . . CH,CI-[CH,],*CO,H . . . . . CF,*[CH,],CO,H . . . . . . CCl,.[CW,],CO,H . . . . . . CH,*[CH,],*CO,H . . . . . . CH2C1*[CH,],*C0,H . . . . . CF,fCH,],-CO,H . . . . . .Dissociation constants 6.2 x 1.53 x lo-' 3 x 10-6 6.98 x low6 3.0 x 10-6 2-04 x 3.2 x 1.51 x 10-5 Temp. (" C) 20 18 25 25 20 18 25 25 Ref. 12 34 35 36 12 34 35 36 group is shown again in the increased strength of the carboxylic acids con- taining a trichloromethyl group. The dissociation constants of the acids CCl,*[CH,];CO,H are greater than those of the unchlorinated carboxylic acids containing the same number of carbon atoms and are greater than those 33 A. I. Titov and F. 0. Maklyayev Zhzw. obshchei Khim. 1954 24 1860. 3 4 E. Larson and B. Adell 2. phys. Chem. 1931 158 352. 35 D. M. Lichty Annalen 1'301 219 36% 36 A. L. IIenne and Ch. J. Fox J . Anzer. Chem. Xoc. 1953 7'5 2323 5750. NESMEYANOV et al. TETR~A- AND TRI-CHLOROALIIANES 333 of corresponding 0)-chloro-carboxylic acids.12 The difference in the influ- ence of the trichloromethyl and trifluoromethyl groups decreases with the increase in the number of methylene groups and is already negligible with trihalogenovaleric acids (see Table 1).The Action of Electrophilic Reagents on Saturated Compounds containing a Trichloromethyl Group.-We have studied the action of sulphuric and nitric acid as electrophilic reagents causing hydrolysis of the trichloro- methyl to the carboxyl group as well as the action of aluminium and ferric chlorides leading to the splitting off of hydrogen chloride at the expense of the chlorine of the trichloromethyl group. Hydrolysis to the carboxyl group is the principal reaction of the trichloromethyl group as it permits passage from chlorine derivatives involving this grouping to the corresponding carboxylic acids.Previously the only way of effecting hydrolysis of the trichloromethyl group in saturated compounds was by heating with concentrated (92-95%) sulphuric acid.13 28 379 35 By this procedure tetrachloroalkaiies CH,Cl*[CH,];CCl (where n = 4 6 8) were converted into 5-chloropentanoic 7-chloroheptanoic7 and 9-chlorononanoic acid 25 37 and the corresponding higher aaaw-tetrachloroalkanes yielded 11 -chloroundecanoic 1 S-chlorotri- decanoic and 15-chloropentadecanoic acids.13 One must however note that whilst hydrolysis of lower tetrachloro- alkanes can be effected with almost quantitative yield of the corresponding acids hydrolysis of higher tetrachloroalkanes with sulphuric acid proceeds with marked " slurrying " or tar formation and the yields are greatly reduced.Thus the yield of 13-chlorotridecanoic acid amounted to 42% and that of 15-chloropentadecanoic acid to 24Y0.l3 Compounds containing a chlorine atom in the cc-position to the chloro- methyl group are only slowly attacked by concentratted sulphuric acid the reaction starting only a t 160-170" and being accompanied by con- sidcrable slurrying. One of us and Ye. J. Vasil'eva have now shown that nitric acid (s.g. 1.51-1.52) reacts with saturated polychloroalkanes containing a trichloro- methyl group even a t room temperature to give the corresponding carboxylic acids ; l5 to complete the reaction the mixture is heated at 60-90" for 1-3 hours. This procedure was employed l5 to prepare in high yields acids from tetrachloroalkanes containing 5 7 9 and 11 carbon atoms as well ag for the hydrolysis of 1 1 l-trichlorotridecane 1 1 1- trichloropentadecane and 1 1 1 -trichloroheptadecane.This method is particularly useful for obtaining the higher carboxylic acids as the reaction proceeds readily without slurrying. When trichloroalknnes are hydrolysed with nitric acid the yields of acids containing 13 15 and 17 carbon atoms amount to 61 66 and 40% of theory respectively. Those containing chlorine in the or-position to the tri- chloromethyl group also undergo hydrolysis rather readily when heated with fuming nitric acid. In this case the reaction is carried out a t 120-130". Hydrolysis. 37 R. 31. Joyce U.S.P. 2,398,430. ** H. J. Prins J . prakt. Chem. 1914 89 414. 334 QUARTERLY REVIEWS Thus froni 1 1 1 2 5-pentachloropentane was obtained 2 5-dichloro- pentanoic acid.With 60-G0yo nitric acid there is virtually no reaction ; 90yo nitric acid reacts with amw-tetrachloroalkanes but the yields of acids containing the sanie nuniber of carbon atoms are in this case lower. Unlike hydrolysis by sulphuric or nitric acid which takes place only with concen- trated acids percliloric acid as dilute as 7074 hydrolyses aaccw-tetrachloro- alkanes the yields being however substantially lower than with the procedures mentioned above. 25d Phosphoric acid does not hydrolyse the trichloromethyl group. Dehydrochlorination of aaacr,-tetrachloroalkanes and ccau-trichloroalkanes. Among chemical reactions of polychloro-derivatives an important place is to be allotted to dehydrochlorination as constituting a route to unsaturated polychl oro-derivatives.Dehydrochlorination of higher tetrachloroalkanes has been described in patent l i t ( e r a t ~ r e ~ ~ 40 where it is suggested that catalytic renioval of hydrogen chloride and the removal by alkali take place a t the expense of the chlorine in the trichloromethyl group and result in trichloroalkenes CCl,:CH* [CH,];Cl. Actually it has been found that dehydrochlorination with alcoholic alkali yields a mixture of products which is difficult to separate. One must also note that the constants for trichloroalkenes described in the patent literature e.g. those of 1 1 5-trichloropent-l-ene proved to be inaccurate as shown by a divergence between the values found and the calculated molecular refraction (MR) . The literature reports dehydrochlorination of polychloro-derivatives under the action of aluminium chloride to give e.g.hexachloropropene from lieptachloropropane,41 tetrachloroethylene from penta~hloroethane~4~ e t ~ . ~ ~ There are examples of dehydrochlorination by heating with an- hydrous ferric chloride,44 e.g. DDT. These reactions are however known to have been carried out at a comparatively high temperature the scope of the procedure being thereby limited. aaacu-Tetrachloroalkanes and ma-trichloroalkanes have now been found to split off hydrogen chloride under the action of a small quantity of aluminium chloride and particularly of anhydrous ferric chloride men a t room temperature ; 4 the reaction is brought to completlion by a short period of heating a t 40-60 O yielding dichlorovinyl derivatives Cl=[CH,];CH:CCI and CH,-[CH,];CH:CCl,.Under these conditions by- products are not formed nor does isomerisation of the paraffin chain take place. The method was used to give 1 1 5-trichloropent-l-ene 1 1 7 - triclilorohept-l-ene 1 1 9-trichloronon-l-ene7 1 l-dichloropent-l-ene and 1 l-dichlorohept-l-ene. The Action of Nucleophilic Reagents on Saturated Compounds containing a Trichloromethyl Group.-The trichloromethyl group proved inert towards 39 B.P. 581,899 ; Chem. A h . 1947 41 3477. 4 0 R. M. Joyce U.S.P. 2,410,541. 41 J. Boescken J. van du Scheer and J. G. Voogt Rec. Traw. chim. 1915 34 78. 4 2 H. J. Prins ibid. 1935 54 249. 4 3 Idenz ibid. 1946 65 455. 4 4 E. E. Fleck and H. L. Haller J. Amer. Chem. SOC. 1944 66 2095. NESMEYAKOV et d. TETRA- AND TRI-CIILORO AL~TANES 335 the action of nucleophilic reagents.Thus 1 1 1-trichloropentane does not exchange with ammonia (heated with alcoholic ammonia a t 140" for 10 hours or liquid ammonia at 140" for 5 hours) or with sodium iodide (refluxed in acetone for 18 hours) or with diethyl sodiomalonate. Other aaa-trichloroalkanes behave ~imilarly.~ Unlike these compounds benzo- trichloride and chloroform react with nucleophilic reagents e . g . when treated with ammonia they form benzonitrile 45 and hydrogen cyanide,46 respectively. In the action of nucleophilic reagents on aaam-tetrachloro- alkanes the trichloromethyl group also remains intact only the chloro- methyl group entering the rea~tion.~ Thus in the action of sodium iodide on 1 1 1 5-tetrachloropentane in acetone during 8 hours' heating 1 1 1-trichloro-5-iodopentane is formed in 90% yield.The structure of 1 1 1-trichloro-5-iodopentane was ascertained by converting it by means of sodium cyanide into the known 1 1 1-trichloro-5-cyanopentane.5 1 1 1 5-Tetrachloropentane when heated with potassium acetate in glacial acetic acid for 18 hours (preferably in the presence of a small amount of potassium iodide) forms 5-acetoxy-1 1 1-trichloropentane in 86% yield. The structure of 5-acetoxy-1 1 1-trichloropentane was proved by converting it into 5 5 5-trichloropentan-1-01 in quantitative yield. Ammonia diethyl s~diomalonate,~ potassium ~yanide,~ 47 and other nucleophilic reagents react with aaorm-tetrachloroalkanes similarly. It is to be noted that depending on the basicity of the nucleophilic reagent and the reaction conditions there takes place a varying extent of dehydro- chlorination a t the expense of the trichloromethyl group.9 Quite different is the behaviour in a number of reactions of 1 1 1 3- tetrachloropropane.Thus in reaction with sodium cyanide sodium stil- phide or other nucleophilic reagents it is not possible to bring about the exchange of chlorine in the chloromethyl group ; instead the dehydrochlorina- tioii reaction usually takes place with formation of a mixture of isomeric trichloropropenes and the products of their subsequent reaction. Oiily when rigid conditions of refluxing with an excess of aniline were eniployed could one obtain 1 1 1-trichloro-3-aniliimpropane in a low yield.'@ The trichloromethyl group being inert to nucleophilic reagents it is impossible to hydrolyse it in weakly acidic neutral or basic media.This is not the case in a strongly acid medium where the electrophilic qualities of the reagent come into play. Similarly the trichloromethyl group does not undergo exchange with bromine anion under the action of hydrogen bromide but such an exchange does take place under the concixrrent attack of electrophilic aluminium chloride which can be represented as c1 c1 Br- - - - C - - - -Cl.AlCl + Br-C + AlClh // c1 /,c1 I I I R I R 45 N. Limpricht Aunalen 1865 135 82. 46 A. Hofinann ibid. 1867 144 116. 47 R. Joyce U.S.P. 2,425,426. 336 QUARTERLY REVIEWS Thus by introduciirg hydrogen bromide into 1 1 1-trichloropentane in the presence of a small amount of aluminium chloride a t 4-5" 1 1 1- tribromopentane is formed in high yield.g Similarly in 1 1 1 &tetra- chloropentane the halogen exchange takes place initially in the trichloro- methyl group 1 1 l-tribromo-5chloropentane being f~rrned.~ It seems that the action of nucleophilic reagents on compounds con- taining the trichloromethyl group in concentrated acid or in the presence of an aprotic acid (AlCl etc.) can find a wider application.and of the cant- pounds CCl,*CHCl*CR,X l4 with akoholic alkali. The dehydrochlorination of 1 1 1 3-tetrachloropropane is of special interest as it led to the formerly unknown 3 3 3-trichloroprop- 1 -ene. The trichloropropene b.p. 115" described in the literature 3 2 9 48-50 as having the structure CCl,*CH:CH, has been shown by Kirrmsnn and Ostermann 51 to possess the structure CCl,:CCl*CH,. Reaction between alcoholic alkali and 1 1 1 3- tetrachloropropane in the cold leads to 3 3 3-trichloroprop-l-ene 1 1 3- trichloroprop-l-ene and 1 l-dichloro-3-ethoxyprop-l-ene ; 3 3 3-tri- chloropropene is readily isolated from the mixture by fractionation.The last two products are separated with difficulty and therefore it is better in some cases to carry out the reaction in ethyl cellosolve. Data concerning 3 3 3-trichloropropene are given below (p. 339). Polychloro-derivatives of the structure CCl,*CHCl*CH,X (where X = Ph OMe NEt, CN or CO,H) were dehydrochlorinated by alcoholic alkali to ascertain the influence of the type of substituent adjacent to the methylene group on the order of scission of hydrogen chloride from the particular molecules. l4 I n all the examples mentioned scission occurred i r i accordance with Saitzeff's rule l4 Dehydrochlorinution of 1 1 1 3-tetrachloropropane l ICOH CCl,*CHCl.CH,X -+ CCl,:CCl.CH,X (where X = Ph OMe NEt,) When the compounds CCl,*CHCl*CH,*CN and CCl3*CHC1*CH,*CO,H were dehydrochlorinated by alcoholic alkali the reaction ran contrary to that rule 7 KOH CCl,*CHCl*CH,X -+ CC1,CH:CHX These observations show that the dehydrochlorination of substances CCl,*CHCl*CH,X proceeds according to Saitzeff's rule when X behaves as an electron-releasing substituent whilst when this substituent is a pro- nounced electron-attracting one Sajtzeff's rule is not obeyed.l4 The starting materials with X = Ph CN or CO,H were obtained by chlorinating the corresponding compounds CCI ,:CH*CH,X in carbon tetrachloride at 0-5 O. l4 The compounds CCl,*CHCl*CH,-OMe and CCl,*CHCl*CH,*NEt were produced by chlorinating CCl,:CH*CH,*OMe and CC12:CH-CH,*NEt in ether and concentrated hydrochloric acid siinultane- 4 8 E.Vitoria Rec. Trav. chim. 1905 24 265. 4u J-. Henry ibid. p. 342. 5 0 A. L. Heniie and A. M. Whaley J. Arner. Clzem. SOC. 1942 64 1157. 51 A. Kirrmanii and J. Ostermann Bull. SOC. chim. Frunce 1945 15 168. NESMEYANOV et Crl. TETRA- AND TRI-CHLORO.IIL1iANES 337 ously saturating the mixture with chlorine and hydrogen chloride 14 (see page 348). The structures of the dehydrochlorination products were ascertained as follows. The compound CCl,:CCl*CH,Ph was identified as the product of the reaction AICl CCl,:CCl*CH,Cl + C,H __+ CCl,:CClCH,*C,H 'To the samples of pheiiyltrichloropropene obtained by following the two routes we added chlorine and determined the m.p.of the mixed sample of phenylpentachlovopropane l4 (CCl,*CCl,*C H,Ph). The conipound (:Cl,:CCl*CH,*OMe proved by its constants to be identical with that obtained in the reaction l4 CCl,:CCl.CI€,Cl + Me-ONa -+ CCl,:C'Cl.CH,.OMe The hydrochloride of the compound CCl,:CCl*CH,*NEt was identified by mixed 1n.p. determination 14 as the substance obtained by the reaction CCl,:CCl.CH,Cl + NHEt + CC12:CC1CE3,.NEt,,HC1 The acid from the nitrile obtained when CCl,*CHCl*CH,*CN was dehydro- chlorinated showed constants identical with those of the known 52 yyy- trichlorocrotonic acid obtained when pyyy-tetrachlorobutyric acid was tlehydrochlorinated and exhibited no depression of the melting point when it was mixed with authentic yyy-trichlorocrotonic acid. Attack by Radicals on the Trichloromethyl Group in Saturated Polychloro- hydrocarbons.-We have investigated hoinolytic reactions involving a trichloromethyl group with phenylmagnesium bromide in the presence of cobaltous chloride under the action of Raney nickel and finely ground copper.I n all cases reaction took place at the expense of the trichloro- methyl group the monochloromethyl group remaining ~nchanged.~ In the absence of cobaltous chloride 1 1 1 5-t~etrachloropentane does not react with phenylmagnesium bromide. In the presence of cobaltous chloride which is known 53 to direct the reaction of organomagnesium compounds with halogen derivatives along the radical mechanism 1 1 1 5- tetrachloropentane and phenylmagnesium bromide formed a mixture from which were isolated two main products dipheiiyl and 1 5 5 6 6 10- hexachlorodecane no products arising from reaction of the chloromethyl group having been found.According to Kharasch 53 the reaction runs as follows PhMgBr + CoC1 -+ Ph- + CoCl + MgClBr 2Ph- + Ph-Ph Cl*[CH,],.CCl -+ CoCl -+ Cl.[CH,],*CCl,* + COCI 2C1.[CH,],*CC12* + C1fCH,],~CCl,CCl~*[CH~]4*Cl Refluxing of 1 1 1 5-tetrachloropentane with Raney nickel in ethyl :~lcohol for 2 hours gives 1 5 5 6 6 10-hexachlorodecane along with some starting material. Finely ground copper when heated has the same 5 2 I<. Auwers and H. Wissebach Ber. 1923 56 731. 53 M. S. Kherasch and E. K. Fields J . Arner. Chein. SOC. 1941 63 2316. 333 QUARTERLY REVIEWS effect on 1 1 1 5-tetra~hloropentane.~ In the presence of platinum palladium or Raney nickel catalyst and bases hydrogen acts selectively on the trichloromethyl and does not affect the monochloroinethyl group resulting in hydrodiinerisation a t the expense of the former group,l6 549 55 with the formation of the compounds (11).ClfCH2lnCC13 + (Cl*[CH,],,CCl2*)2 -+ (Cl*[CH,],.CCl:) -+ (1) (11) (111) (Cl*[CH2]nCH2') 2 (IV) The next step of the hydrogenation has been shown 16 to be dechlorina- tion to form a compound involving a symmetrical dichlorovinyl group subsequently reduced to the disubstituted alkane. The reduction to the end product (IV) is of necessity carried out through the isolation of an intermediate compound of the type (11) as being carried out continuously in one step the process runs very slowly and results in a poor yield of end product. The higher tetrachloroalkanes containing 7 9 and 11 carbon atoms behave towards nucleophilic and radical reagents just as does 1 1 1 5-tetrachloropentane.Conclusions.-From the above account one can make conclusions about the chemical reactions of the trichloromethyl group in saturated polychloro- hydrocarbons. The trichloromethyl group is inert to nucleophilic reagents ; this seems to be due to the screening of the central carbon atom from nucleophilic attack by the three chlorine atoms. Electrophilic reagents behave in reactions with aaa-trichloroalkanes and ccaccu-tetrachloroalkanes oppositely to nucleophilic reagents in that they attack in the first place the trichloro- methyl and leave unaffected the monochloromethyl group. Radicad reagents also selectively attack the trichlorornethyl group.It is interesting to note that the heterolytic reactions of nucleophilic substitution of chlorine in the moiiochloroniethyl group and of attack by electrophilic reagents on t>he trichloroinethyl group in tlhe polychloro- hydrocarbons under study result in a high yield of product. Homolytic changes of the trichloroinethyl group are much inore complex a number of products being formed. The introduction of a chlorine atom into the a-position to a trichloro- methyl group considerably retards the attack by electrophilic reagents. Reactions of the Trichloromethyl Group in Compounds containing the Grouping cC1,*d-=c< Synthesis and Properties of 3 3 3-Tri~hloropropene.~9 3-Chemical changes which have been studied most thoroughly were those of the simplest compound of this class namely 3 3 3-trichloropropene.5 4 B.P. 652,768; Chenz. Abs. 1952 46 1577. 6 5 E. C. Ladd and H. Sargent U.8.P. 2,651,664. NESMEYANOV et d. TETRB- AND TRI-CHLOROALKANES 339 For a long time the trichloropropene b.p. 114-115" ng = 1.4827 d:O = 1.369 first obtained by dehydrating 3 3 3-trichloropropanol was mistakenly postulated to have the structure 3 3 3-trichloropropene. Actually it is the 1 1 2-tri~hloroprop-l-ene.~~ The mistaken assumption has led to a number of wrong suggestions as to the properties and chemical behaviour of 3 3 3-trichloropropene as well as to khe structures of many compounds related to 1 1 2-trichloroprop-l-ene. 3 3 3-Trichloroprop- ene was obtained by the action of potassium hydroxide on 1 1 1 3- tetrachloropropane a t 0-5" ; the reaction also yields 1 1 3-trichloro- prop-l-ene.3 3 3-Trichloropropene is a liquid b.p. 101-102" n; = 1.4680 d2p0 == 1.3292 (Found MR 30.37 ; calc. MR 30.20). The structure of this trichloropropene was proved by its yielding chloral when ozonised. Contrary to the prevailing literature reports that 3 3 3-trichloro- propene ~s,~O supposedly inert it proved to be a rather reactive substance. In particular it readily undergoes allylic rearrangement adds chlorine and bromine and also in the presence of benzoyl peroxide adds hydrogen Iiromide. It can be dimerised and polymerised by peroxides and condensed with benzene in tthe presence of aluminium chloride. Allylic rearrangement of 3 3 3-trichloroprop-l-ene into 1 1 3-trichloroprop-1-ene results when t(he former is heated in a steel tube up to 150" or when a small amount of aluminium chloride is added to it at 0".3 3 3-Trichloropropene readily adds chlorine when its solution in carbon tetrachloride is saturated with gaseous chlorine a t room temperature to give a liquid pentachloropropane b.p. 64-65"/8 mm. n g = 1.5105 (1:O = 1-6117 (Found MR 40.16; calc. MR 40.39). This pentachloro- propane must be 1 1 1 2 3-pentnchloropropane because of the route by which it was obtained. Its properties differ markedly from those of t,he crystalline pentachloropropane I1.p. 170-180" described in the litera- lure as having this structure. The latter compound obtained 487 49 by adding chlorine to 1 1 2- trichloropropene b.p. 115" is probably Z 1 1 2 2-pentachloropropane. Addition of bromine to 3' 3 3-trichloropropene gives a liquid dibromo- tz-ichloropropane b.p.76-70"/3 mm. nko = 1.5640 di0 = 2.1712 (Found MR 45-75 ; calc. MR 46-18) apparently 2 3-dibromo-1 1 l-trichloro- propane.19 3 The crystalline dibromotrichloropropane described in the literature as melting a t 210" and supposed to be the 2 3-dibromo-1 1 1- trichloropropane owing to its being obtained 48p 49 by adding bromine to the trichloropropene b.p. 115" is actually 1 2-dibromo-1 1 2-trichloro- propane. For the synthesis from 3 3 3-trichloropropene of a number of halo- geno-propanes and -propenes containing fluorine chlorine and bromine see references 1-3 56. The Action of Nucleophilic Reagents.-(a) On 3 3 3-trichloropropene. In all cases studied the action of nucleophilic compounds on 3 3 3- trichloropropene takes place with allylic rearrangement giving products je R.N. Haszeldine J. 1953 3371. Y 340 QUARTERLY REVIEWS identical with those obtained by reaction of the same reagents with 1 1 3-trichloropsop- 1 -ene. As nucleophilic reagents diethylamine diethyl sodiomalonate sodium sulphide and sodium methoxide were used. I n reactions with 1 1 3- trichloroprop- 1 -ene the ally1 chlorine was substituted. These reactions can be illustrated + Et2NH + CCI,:CH*CH,.NEt CCl,CH:CH + NaCH (CO ,E t ) -+ CCl,:CH-CH,*CH (CO pE t ) CCl,:CH*CH,Cl + Na,S -+ (CCI,:CHCH,),S i + MeONa + CCl,:CH*CH,-OMe The identity of the dichlorodiethylaniinopropenes obtained from the two trichloropropenes by reaction with diethylsniine is proved by mixed ni .p. determination of the hydrochlorides. The structure and ident,ity of the products of the reaction of diethyl sodiomalonate and the trichloropropenes were proved by their conversion into glutaric acid by hydrolysis and decarboxylation.The identity of the bisdichloropropenyl sulphide derived from the two trichloropropenes was indicated by the boiling point of a mixture of the sulphones obtained from the two sulphides. The action of sodium methoxide on either 3 3 3- trichloropropene or 1 1 3-trichloroprop-l-ene gave the same compound apparently 1 1 -dichloro-3-methoxyprop- 1 -ene. It is to be stressed that the reactions of nucleophilic reagents with 3 3 3-trichloropropene give good yields under conditions which exclude its preliminary isomerisation into 1 1 3-trichloroprop-l-ene. One can suppose that the centre of the nucleophilic attack in 3 3 3-trichloro- propene is the methylene group the carbon atom of the trichloromethyl group being strongly screened by chlorine atoms and consequently these reactions of 3 3 3-trichloropropene belong to the type taking place with " transfer of reaction centre ".57 The reaction of 3 3 3-trichloropropene with say diethylamine may be shown to take place as follows Similar results were obtained in the reaction of nucleophilic reagents with 3 3 3-trichloro-2-methylprop-l-ene.58 de la Mare and Vernon 58 found that when 3 3 3-trichloro-2-methylprop-l-ene reacts with sodium thiophenoxide there takes place a second-order reaction only one com- pound with the structure CCl,:CMe*CH,*SPh being formed.This led the authors to conclude that the reaction was exclusively of X,2' type.The 5 7 ( a ) A. N. Nesmeyanov Uch. Zap. Mosk. Uwiw. 1950 No. 132 5 ; ( b ) A. N. Nes- meyanov and M. I. Kabachiiik Zlaur. obshchei Khim. 1955 25 41 ; ( c ) A. N. Nesmoyanov R. Kh. Freicllina and A. Ye. Borisov YuBi1einy.i Sborizilc Akad. Nauk S.S.S.R. 1947 p. 658. t8 P. B. 1). cle la Mare and C. A. Vernon J. 1952 3628. NESMEYANOV t?t d. TETRA- AND TRI-CHLOROALKANES 341 same authors 59 have also found that the reaction of 3 3-dichloroprop-l-ene with nucleophilic reagents follows two paths-with and without isomerisa- tion. For the overall reaction a second order having been found the authors believe the reaction to follow XN2' and &"2 mechanisms. Judging from the reported behaviour of 3 3 3-trichloropropene toward nucleophilic reagents one would expect the reactivity of the compounds CC1,GH:CRR' to be greatly influenced by the character of substituents R and R' directly bound to the centre of nucleophilic attack.The influence of these substituents has been studied by us in collabora- tion with A. B. Belyavsky by using CCl,*CH:CHMe CCl,*CH:CHPh CCl,*CH:CMe, and CCl,*CH:CHCMe,. The synthesis and proof of struc- ture of these compounds and their allylic isomers CCl,:CH*CHCl*Me and CCl,:CH*CClMe has been given. 22 As nucleophilic reagents ammonia and amines alcohols in the presence of alkali sodium alkoxides and potassium acetate were used among others. In all cases studied the reactions pro- ceeded with transfer of the reaction centre according to the scheme On compounds CC1,aCH:CRR'. 22 0-b $21 XY + RR'C'-CH--C-+Cl -+ Xc'l 4 YRR'C'CH CCI b l the following peculiarities being noted 1 1 l-trichloro-4 4-dimethyl- pent-2-ene does not react with diethylamine and only extremely slowly with sodium methoxide ; this seems to result from steric hindrance due to the tert.-butyl group directly adjacent to the centre of nucleophilic attack.The reactions of Me*CH:CH*CCl and Ph*CH:CH-CCl with diethylamine in alcoholic media result in mixtures of the corresponding alkoxy- and diethylamino-derivatives ; in the case of Me,C:CH*CCl only alkoxy- derivatives are formed whereas under the same conditions CH,:CH*CCl forms only diethylamino-derivatives. When the reaction with diethylamine is carried out in the absence of alcohol Me*CH:CH*CCl and PhCI-I:CH*CCl react in the usual way giving diethylamino-derivatives whereas Me,C:CH*CCI does not react even at 100-llO" slurrying taking place at ,b higher teniperature.The reaction of Me,C:CH*CCl with ammonia and piperidine in alcohol leads to negligible quantities of amino-derivatives. The same is true of the reaction with sodium sulphide in alcoholic solution CICl,*CH:CH yielding only the sulphide (CCl,:CH*CH,),S and CCl,*CH:CMe yielding only an alkoxy-derivative. The results are listed in Table 2 (overleaf). As is seen from the Table the substances investigated can be arranged in the series CCl,*CH:CH, CCl,*CH:CHMe CCl,*CH:CHPh CCl,=CH:CH*CMe, CCl,*CH:CMe, in which the ability of the compound to be alkylated on the nitrogen atom is decreasing and to be alkylated on the oxygen atom increasing. 5 9 P. B. D. de la Mare and C. A. Vernon J. 1952 3325. 342 QUARTERLY REVIEWS TABLE 2.The action of nucleophilic reagents on CCl,*CH:CRR’. Compound CCl,*CH:CH . . . . . CCI,*CH:CHMe . . . . . CC1,CH:CHPh . . . . CC1,-CH:CMe . . . . . CCI,.CH:C€T-CMe . . . N N N does n o t substitute - 0 denotes formation of CCl,:CH.CRR’.OR” ; N formation S formation of (CCl,:CH*CRR’),S. Et,NH in alc. N 0 and N 0 and N 0 Na,S in ale. s 0 and S 0 - O I - of CCI,:CHCRR’:NR” ; and If one considers the electrophilicity of tjhe y-carbon atom in the above series of trichloromethyl derivatives to decrease being the least with the tl.ichlorodimethylpropene (see X) then the observed relation can be explained as follows as the electrophilicity of the y-carbon atom decreases the rate of reaction of alkylstlion of the oxygen atom increases and that of the nitrogen atom decreases.It must however be borne in mind that cy n*7+H3 X (JtC‘-X>H :- C c1 J \c. H the initial trichloromethyl compounds being compared differ not only in their electrophilicity but also in steric hindrance a t the carbon atom the influence of each of these factors taken separately being unknown. We now consider some other reactions of these substances.22 The reaction of CCl,*CH:CMe with alcoholic potassium hydroxide diethylamine or piperidine gives in addition to the main product CCl,:CH*CMe,X a small quantity of 1 1-dichloro-3-methylbuta-1 3-diene possibly as a re- sult of the isomerisation of the trichloromethyl compound to 1 1 3-tri- chloro-3-methylbut- 1-ene followed by dehydrochlorination or as a result of a direct attack of the nucleophilic reagent on the om^ conjugated system according to the scheme p1 X y -t- ~ ~ - C ‘ M d ~ - C ‘ - + C I + YH + CH,=CNleCH=CCI + XCI H’ LCl The latter suggestion is preferred for whilst CCl,*CH:CMe under the action of alcoholic alkali gives mainly 3-alkoxy- 1 1 -dichloro-3-methyl- but-1-ene and only a little 1 1-dichloro-3-methylbuta-1 3-diene under the same conditions CCl,:CH*CMe is wholly converted into 1 l-dichloro- 3-methylbuta- 1 3-diene which suggests that preliminary isomerisation does not take place.The reactions of the trichloromethyl group in Ph*CH:CH*CCl are in some respects similar to those of the same group in benzotrichloride. Thus, NESMEYANOV d Ul. TETRA- AND TRI-CHLOROALKANES 343 the compound was hydrolysed readily when heated for 30 minutes with 90% acetic acid to give cinnamic acid in 95% yield.It is easily disproportionated when heated with chloroacetic acid yielding chloroacetyl chloride. The Attack of Radicals on 3 3 3-Trich1oropropene.-I7he rearrangement of the radical CCl,*CH*CH,X (X = Br or CCl,) in When studying the addition of bromotrichloromethane and hydrogen bromide to 3 3 3- trichloropropene in the presence of benzoyl peroxide we observed a re- arrangement which we interpret to be the rearrangement of the free radical. The addition of bromotrichloromethane to 3 3 3-trichloroprop- ene in the presence of benzoyl peroxide should lead to the compound CCl,*CHRr-CH,*CCl,. Actually the reaction was more complex and yielded a number of products some of which we isolated and identified as the com- pounds CCl,*CH,*CH:CCl, CCl,*CH,*CHClCCl,Br and ClCH2*CHCI*CCl2Br.The formation of 1-bromo-1 1 2 4 4 4-hexachlorobutane can be understood when allowance is made for rearrangement in the intermediate radical CCl3*6H*CH,*CC1,. The reaction can be represented (PhCO,) -+ Ph* + CO + Ph*CO*O* * ( 1 ) Ph* + BrCC1 + PhBr + CCI,. . - ( 3 ) CCl,* + CH,:CH.CCl -+ CCl,.CH2.i:H.CC1 . - (3) C(.'l,*CH2*6H.CC1 -+ CCl,*CH,CHClCCl,* . * (4) CCI,CH2-CHC1.CC1,* + RrCCI + CCl,CH,*CHCl.CCl,Br -f- CCl (5a) CCl,*CH,*CH:CCI -I- CICH,.61-I.CC13 (5b) CCl,~CH2CHC1~CC12* + CH,:CH.CCl -+ CICH~.~HCCI -+ CICH,-CHC~~I'C~ . * (6) ClCH,CHC1*&21 + BrCC1 + ClCH,*CHdl~CCl,Br 4- CC1,- (7) 2CCl,* -+ C,Cl * (8) It will be seen that step (4) suggests isomerisation of the radical CCl,*CH2*6H*CCI to CCl,*CH,*CHCl*~Cl,. An alternative apparent mech- anism would have consisted in the phenyl radical formed during the de- composition of the peroxide reacting not with bromotrichloromethane but with trichloropropene and abstracting the labile chlorine from the trichloro- methyl group according to the scheme CH,:CH*CCl + Ph* + CH,:CH&l + PhCI giving rise to the radical CH,:CH*CCl,* which could have been the starting point for the same products as the first mechanism.I n a special study of the decomposition of a 50 g. sample of benzoyl peroxide in a mixture of bromotrichloromethane and trichloropropene bromobenzene was the only halogenobenzene formed ; that is the phenyl radical takes up bromine from bromotrichloromethane which result defin- itely indicates against the second reaction mechanism. The structure of the pentachlorobutane has been confirmed by its hydrolysis to succinic acid.The structure of the bromohexachlorobutane as 344 QUARTERLY REVIEWS I-bromo-1 1 2 4 4 4-hexachlorobutane has been indicaked by the reactions H804 --+ HO,C*CH:CH*CO,H C!Cl,CH,CHCl*CCI,Br- CCl,:CH*CCl:CCl 1 =: - CCl,*CH,*CH:CCl -+ HO,C.[CH,],CO,H The structure of 1-bromo-1 1 2 3-tetrachloropropane was proved by its conversion by the action of alcoholic alkali into ClCH,*CCl:CCl, which was identified as the known 1 1 2-trichloro-3-diethylaminoprop-1-ene hydrochloride. When hydrogen bromide is added to 3 3 3-trichloropropene in the presence of benzoyl peroxide there is a ready formation in good yield of a product C,H,Cl,Br which proved to be 3-bromo-1 1 2-trichloropropane. This could also have been formed by isomerisation of the intermediate radical Br + CH,:CHCCl + BrCH,*6H.CCl3 -j.BrCH,.CHClCCl,* BrCH,*CHCl*CCl,* + HBr + BrCH,*CHCl.CHCl + Br Dechlorination of the compound with alcoholic alkali a t 0" is accom- panied by removal of hydrogen bromide to form a compound b.p. 126-127" ng = 1.4840 di0 = 1.3843 (Found MR 30.07; calc. for C3H,Cl, MR 30.18). The resulting trichloropropene differs in physical constants from the four known of the six possible trichloropropenes. The two unknown tri- chloropropenes have the structure CH,:CCl*CHCl and CHCl:CH*CHCI,. The trichloropropene obtained was ozonised and the ozonide decom- posed by water without further oxidation to yield an acid (b.p. 92-94"/ 13 mm. ; dichloroacetic acid has b.p. 91-92"/12 mm.) which was converted through the chloride into the anilide m.p.116-117". The mixture with m-dichloroacetanilide melted a t 1166-1 17 ". Thus the trichloropropene had the structure CH,:CClCHCI, and the parent bromotrichloropropane would seem to be CH,Br*CHCl*CHCl,. Such behaviour of the trichloromethyl group adjacent to the carbon atom carrying a free valency is formally similar to its behaviour in the dehydration of 1 1 l-trichloropropan-2-ol which proceeds with rearrange- ment and formation of 1 1 2-trichloroprop-1-ene. The reaction can be represented by a cationic rearrangement CCl,.CH(OH)*CH -+ CCl,*CH.CH + CC1,*CHC1*CH3 + CC1,:CC1*CH3 Other reactions of 3 3 3-trichloropropene of homolytic type.7 When butylmagnesium bromide reacts either with 3 3 3-trichloropropene or with 1 1 3-trichloroprop-1-ene the main product is 1 l-dichlorohept- 1-ene the structure of which has been proved by hydrolysis to heptanoic acid.When the reaction was carried out with phenylmagnesium bromide in addition to 1 1-dichloro-3-phenylprop-1-ene some diphenyl was pro- duced. It is to be noted that these reactions as is usually the case with + + NESMEYANOV et al. TETRA- AND TRI-CHLOROALKANES 34.5 hornolytic reactions proceed with formation of a number of other products not investigated in detail. These reactions can be represented by the following scheme 2RMgX + R*R +2MgX . * (1) CH,:CH-CCl + MgX -+ CH,:CH*eCl + MgXCl . (2) CH,:CH6C12 + RMgX + R*CH,CH:CCI + MgX . (3) Under the action of Raney nickel in ethanol 3 3 3-trichloropropene yielded a tetrachlorohexadiene which apparently has the structure [CC1,:CH*CH2I2 since when hydrolysed in the presence of concentrated sulphuric acid it gave adipic acid though in a low yield.Thus reactions which are likely to be homolytic also proceed with 1 1 3-trichloroprop-1-ene a t the expense of allylic chlorine and in the case of 3 3 3-trichloropropene with allylic rearrangement.' Consequently in homolytic reactions 3 3 3-trichloropropene appears to undergo two types of rearrangement ( a ) allylic and ( b ) with shift of a chlorine atom from the trichloromethyl group to the adjacent carbon atom. One may suppose the type of rearrangement to depend on the mechanism of the reaction. Reaction with allylic rearrangement takes place when the radical attacks the trichloromethyl group and that with chlorine shift when the radical attacks the methylene group.The Electrophilic Reagent Attack on 3 3 3-Trichloropropene.- Friedel-Crafts cataZysts.lP Under ordinary conditions 3 3 3-trichloro- propene is not in tautomeric equilibrium with its allylic isomer 1 1 3- trichloroprop- 1-ene. Small quantities of such electrophilic reagents as aluminium chloride ferric chloride and antimony pentachloride produce isomerisation of 3 3 3-trichloropropene to 1 1 3-trichloroprop-1-ene. This isomerisation induced by the action of e.g. aluminium chloride can be represented CH,:CH.CCl + AlCI + (CH,:CHCCl,)+AlCl,- + Cl*CH,.CH:CCl + AlC1 The reverse isomerisation of 1 1 3-trichloroprop-1-ene to 3 3 3- trichloropropene is unknown. The reaction of compomds CCl,*CH:CRR' with aromatic compounds in the presence of aluminium chloride.3 3 3-Trichloropropene condenses extremely readily with benzene in the presence of small quantities of a'luminium chloride a t 0-5" giving in good yield a product of the com- position C6H,*C,H,C12 b.p. 93-94"/6 mm. n? = 1.5490 di0 = 1.2032 (Found MR 49.45 ; calc. MR 49.43). In the presence of aluminium chloride the reaction has thus proceeded with allylic rearrangement and the product ha's the structure PhCH,*CH:CCl,. Indeed the reaction of benzene with 1 1 3-trichloroprop-1-ene under the same conditions yielded the same product.1 Hydrolysis of this substance with 70% perchloric acid gave P-phenylpropionic acid.25d This reaction may be of int'erest as a synthetical method permitting the introduction into the aromatic molecule of the reactive grouping C€I,*CH:CKl, which. in particular is readily 346 QUARTERLY REVIEWS converted into the propionic acid residue.Two of us in collaboration with N. A. Semenov have investigated the reaction of 3 3 3-trichloropropene with chlorobenzene bromobenzene anisole phenol aniline and methyl- and dimethyl-aniline. 20a Bromobenzene and chlorobenzene react violently with 3 3 3-trichloro- propene in the presence of aluminium chloride with evolution of heat the main products being 3-p-bromophenyl- and 3-p-chlorophenyl-1 1 -dichloro- prop-1-ene respectively. The structure of these products has been proved by hydrolysing them with concentrated sulphuric acid giving in good yield /?-p-bromophenyl- and P-p-chlorophenyl-propionic acid. 1 1 -Dichloro-3- p-chlorophenylprop-1 -ene and 3-p-bromophenyl- 1 1 -dichloroprop-1 -ene add chlorine yielding 1 1 1 2-tetrachloro-3-p-chlorophenylpropane and 3-p- bromophenyl-1 1 1 2-tetrachloropropane.Anisole and phenol react less readily with 3 3 3-trichloropropene in the presence of aluminiun chloride than do chloro- and bromo-benzene requiring several hours a t 80-90" to bring the reaction to completion. Anisole gave 1 l-dichloro-3-p- methoxyphenylprop- 1 -ene the structure being ascertained by oxidation with 5% potassium permanganate solution t o anisic acid. 3 3 3-Tri- chloropropene condenses with phenols when heated even in tjhe absence of aluminium trichloride; it is better however to carry out the reaction in the presence of aluminium chloride obtaining a mixture of 1 l-dichloro- 3-0- and 1 1-dichloro-3-p-hydroxyphenylprop-1 -ene. The structure of these compounds has been proved by alkylating them with dimethyl sulphate to the corresponding methoxyphenyl compounds which were oxidised to o-methoxybenzoic acid and anisic acid respectively.Hydro- lysis of 1 1 -dichloro-3-p-methoxyphenylprop- 1 -ene with concentrated sul- phuric acid yielded sulphonated P-p-methoxyphenylpropionic acid obtained as barium salt. The reaction of 3 3 3-trichloropropene with aqueous sodium phenoxide gave products of both C- and 0-alkylation resulting in a mixture of 1 1 -dichloro-3-o- 1 l-dichloro-3~p-hydroxyphenylprop-1-eney and 1 1 -dichloro-3-phenoxyprop-1-ene. The same products were obtained when the above mentioned aromatic compounds reacted with 1 1 3- trichloroprop-1-ene but the reaction then proceeded less readily and the yields were lower than with the reaction with 3 3 3-trichloropropene.Such trichloromethyl derivatives as CCl,.CH:CHMe CCl,*CH:CHPh and CCI,*CH:CMe also condense with benzene in the presence of aluminium chloride yielding the compounds Ph*CHMeCH:CCl, Ph,CH*CH:CCI, and Ph*CMe,*CH:CC1,.22 The structure of the last wars proved by oxidising it with potassium permanganate to cccc-dimethylphenylacetic acid. As distinct from the reactions of 3 3 3-trichloropropene with electrophilic reagents we have just discussed the conjugated addition of chlorine to 3 3 3-trichloro- propene in glacial acetic and concentrated sulphuric acid proceeds as has been already shown (p. 331) without isomerisation yielding in addition to 1 1 1 2 3-pentachloropropane 2 3 3 3-tetrachloropropyl sulphate and acetate.Hydrolysis of these esters gave 2 3 3 3-tetrachloropropanol. In this case absence of isomerisation may be due to the fact that it was Chlorination of 3 3 3-trichloropropene in acids.l8 NEESMEYANOV et UI?. TETRS- AND TRI-CHLOROALKANES 347 not the trichloromethyl group that has been subjected to electrophilic attack but the carbon atom situat,ed in the centre of the chain. Conclusions.-The investigation of the reactivity of compounds involving the system of linkages shown in ( A ) toward nucleophilic electrophilic and radical reagents has thus shown these reactions to proceed with rearrange- ments in all cases when one can assume that the first or the fourth atom of the chnin is are attacked strikingly the subjected to the attack. there is no rearrangement. If the central atoms of this system These relations demonstrate occurrence of GZ conjugation in the system which can be represented as in (B).Because of the screening of the carbon atom of the trichloromethyl group the only centre for nucleophilic attack in this par- ticular system is a t the first atom which is supplied along the chain of on conjugation with electrophilicity from the carbon atom of the trichloro- methyl group ; reaction of nucleophilic reagents with compounds of this type therefore always proceeds with allylic rearrangement. Depending on the character of the reagent electrophilic attack can take place either at the second or the fourth atom of the system. The second atom being attacked reactions proceed without isomerisation ; when the fourth atom is attacked there is allylic rearrangement.Radical reagents when attacking the first atom of the system cause rearrangement with shift of chlorine from the trichloromethyl group to the adjacent carbon atom. Attack on the fourth atom leads to allylic resrrange- ment. Reactions of the 2 2-Dichlorovinyl Group Compounds containing the 2 2-dichlorovinyl group (CCl,:CH*) are now readily available. They may be produced by dehydrochloririation of aaa-trichloroalkanes and aaccco-tetrachloroalkanes (see p. 334) as well as by the action of Grignard reagent on 3 3 3- or 1 3 3-trichloropropene. Coxizpounds of the type CCl,:CH*CHCl*OR are readily prepared by dehydro- chlorination of the products of the addition of carbon tetrachloride to alkyl vinyl ethers,lO 60 61 and compounds of the type CC1,:CHCRR’:X by both nucleophilic and some other reagent attack on the compounds CC1,CH:CRR’ (see p.341) as well as in other ways. Hydrolysis-Hydrolysis of compounds containing the 2 2-dichlorovinyl group with concentrated sulphuric acid yields carboxylic acids. This reaction is carried out under similar conditions to the hydrolysis of the trichloromethyl group. In contrast with the action of fuming nitric acid Go S. A. Glickman U.S.P. 2,560,219. 61 E. Lewas and E. Lewas Conzpt. rend. 1950 230 1670. 348 QUARTERLY REVIEWS qn the trichloromethyl group l5 this acid acts on 2 2-dichlorovinyl com- pounds to yield neutral nitr- enous compounds. Hot 70% perchloric acid hydrolyses the dichlovovinyl group as has been shown by two of us in col- laboratioii with Ye. J. Vasil'eva by obtaining co-chlorovaleric and 7-chloro- heptanoic acid from 1 1 5-trichloropent-l-ene and 1 1 7-trichloro- hept-l-ene.The reaction is however accompanied by slurrying and gives moderate yields. 1 1 -Dichloro-3-phenylprop- I -ene and 70% perchloric acid yield ,!I-phenylpropionic acid. Since hydrolysis of aromatic compounds containing the trichloromethyl or dichlorovinyl group by sulphuric acid is often accompanied by nuclear sulphonation hydrolysis of such compounds with perchloric acid is of some preparatory value.25d Oxidation. 25u-Compouiids containing the 2 2-dichlorovinyl group evolve hydrogen chloride on storage and acquire a strong pungent odour but if a little quinol is added the compounds do not undergo decomposition. It is thus evident that dichloroalkenes gradually oxidise when stored.A number of dichloroalkenes having been saturated with oxygen a t 100-1 10" furnished acids in 40-50% yield a-chlorovaleric acid from 1 l-dichloro- pent-l-eiie a&dichlorovaleric acid from 1 1 5-trichloropent-l-ene and 2 7-dichloroheptanoic acid froin 1 1 7-triclilorohept-l-ene. The oxidation by oxygen of compounds containing the dichlorovinyl group is already known as illust,rated by several simple instances to give dichloro-carboxylic acids.62 Chlorination.14 18-In collaboration with V. N. Kost we have shown that addition of chlorine to the double bond of the compounds CCl,:CK*[CH,];Cl and CCI,:CH*[CH,];CH in a neutral medium is usually accompanied by a varying amount of chlorination of the saturated part of the molecule. We have studied the reaction between chlorine and compounds of the structure CCl,:CX*CH,Y where X = H or Cl and Y = Et,14 OMe,14 NEt,,14 N02,14 CN,7 or CO,H,' in mild conditions a t 0-5" in carbon tetra- chloride.Compounds having Y = Ph NO, CN or C0,H have been found to add chlorine smoothly according to the scheme CCl,:CX.CH,Y + C1 --+ CCl,.CXCl*CH,Y But with compounds in which Y = OMe or NEt the reaction is accom- panied by an energetic evolution of hydrogen chloride indicating that the saturated part of the molecule is undergoing chlorination.14 Thus chlorina- tion in the dark of 1 l-dichloro-3-methoxyprop-l-ene under the mildest conditions produces a pentachloro-derivative b. p. 81-82 O / 1 - 5 mm. ng = 1-5070 dzo = 1.5713 (Found MR 46-66 ; calc. for C4H,0C1, MR 46.65). This compound when heated in methyl alcohol in the presence of hydro- chloric acid gave 2 3 3 3-tetrachloropropan-1-01 in 92% yield identified by the melting point of its mixture with an authentic sample.Consequently the structure of the pentachloride obtained can be represented as l8 CC~Z,*CHCl*CH,*O*CH,Cl. The previously suggested l4 substitution of chlorine for allylic hydrogen does not take place in this case under the conditions studied. 6 2 F. W. Kirkbrick U.S.P. 2,292,129 ; G.P. 391,674. NESMEYANOV et Ui. TETRA- AND TRI-CHLOROALKANES 349 The same 1 1 1 2-tetrachloro-3-cliloromethoxypropane has also been obtained according to the scheme 18 CH,O HCl CC1 *CHCICH,*OH + CC1,CHCI CH -0 *CH,Cl and then had b.p. 81-82"/1-5 mm. ng = 1.5065 di0 = 1.5711. The reaction between chlorine and 1 1 5-tricliloropent-l-ene 1 1- dichloro-5-cyanopent-l-ene and 1 1-dichloro-7-cyanohept-1-ene a t 0-15 O in chloroform also proceeds with strong evolution of hydrogen chloride and formation of a mixture of products.l* In all cases it has proved possible to suppress chlorination of the saturated part of the molecule by carrying out the chlorination in a mixture of ether and concentrated hydrochloric acid simultaneously saturated with chlorine and hydrogen chloride.14 18 It will be recalled that diallyl ether allyl alcohol allyl acetatle and allyl benzoate are already known 63 to add chlorine smoothly in hydrochloric acid.Chlorination of Compounds containing the 2 2-Dichlorovinyl Group in Sulphuric Acid. A New Synthesis of a-Chloro-carboxylic Acids.17 Is- Direct ch1orinat)ion of carboxylic and m-chloro-carboxylic acids in the presence of catalysts yields a number of corresponding a-chloro- and am- dichloro-carboxylic acids.The method for obtaining ad-dichlorovaleric acid by chlorination of 6-chlorovaleric acid has in particular been des- cribed. 64 But chlorination at the a-position to the carboxy-group necessi- tates relatively vigorous heating to 120-150" and often leads to isomeric chloro-acids as well as to formation of by-products. Direct chlorination of dicarboxylic acids usually gives a mixture of products difficult to separate. One of us and Ye. J. Vasil'eva,15 using 1 1 1 2 5-pentachloropentane have shown compounds contnining the CCl,CHCl* group and not hydrolysed by concentrated sulphuric acid to be smoothly hydrolysed by fuming nitric acid to the corresponding a-chloro-carboxylic acids.The synthesis of amp-tetrachloro-derivatives from trichloromethyl derivatives involves a two-step process the first step consisting of dehydrochlorination of the trichloromethyl derivative to 1 1 -dichloroalk- 1 -enes the second in chlorina- tion of the 1 1-dichloroalk-1-enes. Though the simple conditions for carrying out in a high yield both the dehydrochlorination and chlorina- tion l7 have been found we thought it advisable to look for a means of direct conversion of 1 1 -dichloroalk- 1 -enes into a-chloro-carboxylic acids. Two of us and V. N. Kost l7 have worked out a new synthesis of chloro- carboxylic acids by chlorinating 1 1 -dichloroalk- 1-enes in 92--93% sul- phuric acid a t 0-20". The reaction is accompanied by evolution of hydro- gen chloride. After the reaction mixture has been decomposed by water a-chloro-carboxylic acids are obtained in high yield.In this way we ob- tained ay-dichlorovaleric a-chlorovaleric 2 7-dichloroheptanoic and 2- chloroheptanoic acid starting from 1 1 5-trichloropent-l-ene 1 l-di- chloropent-1-ene 1 1 7-trichlorohept-l-ene and 1 1-dichlorohept-1-ene respectively.17 For the synthesis of these starting products see ref. 4. 63H. Ing J . 1948 1393. 6 4 R. Gaudry and L. Berlinquet Canad. J . Res. 1949 27 B 282. 350 QUARTERLY REVIEWS I n some cases we observed as side reaction the addition of chlorine to the double bond but the separation of or-chloro-carboxylic acids from these neutral products presented no difficulty. The reaction proved t o be applicable t o the synthesis of a-chloro-di- carboxylic a ~ i d s .1 ~ Thus from 6 6-dichlorohexen-5-oic 7 7-dichloro- hept-6-enoic and 8 8-dichloro-oct-7-enoic acid were obtained in a good yield K-chloroadipic or-chloropimelic and a-chlorosuberic acid respectively. For the syntheses of the initial acids see ref. 12. It is to be noted that urider the conditions of this reaction a t 0-20" there is no hydrolysis of the dichlorovinyl group to carboxylic acid. When the reaction is carried out a t a higher temperature or in more dilute sulphuric acid the yield of a-chloro-carboxylic acid drops. Thus a t 30-40 O hydrolysis of the dichlorovinyl group is rather pronounced whilst when chlorination takes place in 70% sulphuric acid only the product of addition of chlorine to the double bond is formed.18 The mild conditions of the reaction enable one to obtain a-chloro- carboxylic acids with a variety of substituents.In particular under the action of chlorine on 1 1 5 5 5-pentachloropent-l-ene in concentrated sulphuric acid a t 20--25" the trichloromethyl group is retained 2 5 5 5- tetracliloropentanoic acid being formed. l8 The reaction is presumed to pass through the intermediate formation of compounds of the type R*[CH2],*CHC1*CC1,*O*S03H. Reactions of this type are well 65 A. J. Titov and I?. L. Maklyayev 33 have recently shown that reaction of chlorine with olefins in concentrated mineral acids produces a-chloro- esters according to the scheme R*CH:CH + C1 + HA + ClCH,*CHRA + HC1 where A = S03H H,PO, etc. The reactions investigated allowed us to show that compounds contain- ing the 2 2-dichlorovinyl group are also liable to undergo conjugated addition.Action of Sodium on Compounds containing the 2 2-Dichlorovinyl Group. Synthesis of Monosubstituted Acetylenes.-Pinner 66 67 was the first to investigate the action of sodium on the unsymmetrical dichlorovinyl group as exemplified by 1 1 -dichloroprop-1 -ene. He demonstrated that decoinposition with water of the product of the reaction of 1 l-dichloro- prop-l-ene and sodium gave inethylacetylene 67 in low yield and under the action of carbon dioxide propiolic acid.68 Pinner suggested that the initial product was C3€I,C1,Na2 and that the reaction took the following course CW,*CH:CCI + 2Na + C,H,CI,Na + CH,*CZCH + 2NaC1 The reaction has not been further investigated. It appeared to us of interest to investigate the reaction as a preparative H20 6 5 C.K. Ingold " Structtxre and Mechanism in Organic Chom.istry " Cornell 1953. 6 6 A. Pinncr Annalen 1875 179 49. 67 Iclena Ber. 1875 8 1282 6 8 Idem ibid. 1881 14 1081. NESMEYANOV et d. TETRA- AND TRI-CHLOROSLKANES 361 route as well as to elucidate the mechanism of the conversion of the dichloro- vinyl into the ethynyl group. In accordance with existing data it has been found that when treated with carbon dioxide the product of the reaction of 1 l-dichloropent-l-ene with sodium results in pentynecarboxylic acid and the action of benzaldehyde on the product of the reaction of 1 1- dichloro-5-diethylaminopent-1 -ene and sodium leads to B-diethylamino-l- hydroxy- 1 -phenylhex-%yne. These findings clearly show that tlhe product of the reaction of a dichlorovinyl derivative with sodium contains RCrCNa and that the dichlorovinyl group gives up two chlorine atoms and one hydro- gen atom.It is quite obvious that chlorine is eliminated as sodium chloride. It was found that in the course of the reaction only a negligible amount of hydrogen is eliminated (2-5%) and that only when the product is decom- posed with water is the necessary amount of hydrogen liberated (0.5 mol.). Also one must use 4 g.-atoms of sodium per mole of dichlorovinyl derivative in contradistinction to the 2 g.-atoms suggested hy Pinner. These results can be explained on the assumption that the hydrogen from the dichloro- vinyl group is taken up as sodium hydride and the reaction of the dichloro- vinyl derivative with sodium can then be represented by the equ cz t' ion R*CH:CCI + 4Na -+ R-C-CNs + NaH + 2NaCI Confirmation of this suggestion is seen in the formation of some mono- substituted acetylenes and sodium and lithium hydrides in some reczcbions of sodium or lithium with monohalogeno-olefins R*CH:CHX.G 9 5 70 The substitution of sodium for chlorine in R*C%Xl takes place extreniely readily. 71 Investigation of the action of sodium on dichlorovinyl deriva- tives has shown the reaction to proceed smoothly and in most investigated cases to produce monosubstit'uted acetylenes in 40-80% yield. Data for monosubstituted acetylenes so produced are listed in Table 9. With /?p-dichlorostyrene we have not been able to obtain cz good yield of phenyl- acetylene because of failure to bring the reaction to completion. With dichlorostyrene a mixture of prodncts consisting mainly of phenylnllene Ph*CH:C:CH, has been obtained.1 1 3-Trichloro-5-diethylaminopent-l-ene being taken as example the trichlorovinyl group has been converted into the ethynyl group by the action of sodium Et,N.[CH,],*CCl:CCl + 4Na + Et,N*[CH,],*CzCH -t- 3NsCI The yield of 5-diethylaminopent-l-ene is good. 5-Ethoxypent-l-ene as the bromomagnesium derivative and benzaldehyde gave 1 -hydrosy- 1 -phenyl-6-ethoxyhex-Z-yne in a good yield. The Action of Alkali on pp-Dichloroacraldehyde Acetals. 21 Synthesis of Chloropropiolaldehyde Acetals.-S-Alkoxy- 1 1 1 3-tetrachloropropanes are easily obtained by the reaction of carbon tetrachloride and alkyl vinyl ethers in the presence of radical initiators. 3-Alkoxy-1 1 1 3-tetra- chloropropanes a t 130-150" readily lose hydrogen chloride and give in 6 9 A.Kirrmann Cornpt. rend. 1925 181 671. ' 0 E. A. Braude and J. A. Coles J. 195.0 179. 7 1 R. Truchet Ann. Chim. 1931 16 349. 352 QUARTERLY REVIEWS high yield 3-alkoxy-1 1 3-trichloroprop-l-ene. These with an equivalent amount of the appropriate sodium alkoxide in alcohol in the cold result in @@-dichloroacraldehyde acetals which are converted without isolation from the reaction mixture by hot potassium hydroxide into chloropropiol- aldehyde acetals as represented by the scheme CCl,~CH,-CHCl.OR -+ CC1,:CHCHCl.OR + CCI,:CH*CH(OR) -+ CCkC.CH(OR) The diethyl and dibutyl acetals of chloropropiolaldehyde are liquids stable when stored and distilling in mcuo without decomposition. Sodium is readily substituted for chlorine in these acetals by the action of sodium in ether to give sodiopropiolaldehyde acetals.CCkC*CH(OR) + 2Na -+ NaC_C*CH(OR) 4- NaCl Sodiopropiolaldehyde acetals enter into the usual reaction with carbonyl compounds e.g. with cydohexanone NaCrC*CH(OEt) + ()=o ~ (jj:c~c.c"(0Eq2 and can also be alkylated e.g. by dimethyl sulphate NaC-C-CH(OBu) + (MeO),SO -+ CMe.C=C.CH(OCBu) Organomagnesium compounds and diethyl and dibutyl acetals of chloro- propiolaldehyde do not react as one might have expected at the acetal group but mainly at the chlorine atom to form the acetals of substituted propiol aldehydes ClCrC*CH(OR) + R'MgX + R'CzC*CH(OR) This reaction takes place with both aliphatic and aromatic organo- magnesium compounds ; we have studied the action of organomagnesium compounds of the following bromine derivatives n-C,H 7Br i-C,H,Br n-C,H,Br i-C,H,Br n-C,H,,Br n-Cg€I19Br PhBr and cc-C,,H,Br.The reaction was carried out by adding the organometallic compound to an ethereal solution of chloropropiolaldehyde acetal the heat of reaction being sufficient to maintain boiling. Data concerning the products are summarised in Table 9. Proof of the structure of the products is exemplified by butylpropiolaldehyde diethyl acetal which was hydrogenated to heptanal diethyl acetal the latter being identified as heptanal 2 4- dinitrophenylhydrazone. The yields in the cases investigated were 50-70% of theory. AUylic Rearrangements in the Series of Polychlorobutyl Acids and some Mistakes made by Auwers and Wissebach 7 2 3 52 we have examined the relationship of the acids CHCI,*CH:CH*CO,H and CCl,:CH*CH,*CO,H and that of their derivatives ; i.e.the relationship of the prototropic allylic rearrangement. In collaboration with V. N. Kost K. Auwers and H. Wissebach Ber. 1923 56 23 715. NESMEYANOV et al. TETRA - AND TRI-CH1,OROALKANES 353 There are two yy-dichlorocrotonic acids described in the literature one m.p. 42-43' obtained by Auwers and Wissebach 7 2 v 6 2 by the reduction of yyy-trichlorocrotonic acid (see also ref. 73) a,nd the other m.p. 101-102" obtained from dichloroacetaldehyde and malonic acid. 74 It was decided to prepare the hitherto unlinown 4 4-dichlorobut-3-enoic acid by mildly hydrolysing its nitrile obtained by action of cuprous cyanide on 1 1 3-trichloroprop-1-ene CCl,:CH.CH,Cl + CuCN + CCl,:CH*CH,.CN + C~Cl,:CH*CH,*CO,H Both 3 3 3- and 1 1 3-trichloroprop-1-ene give the same 1 1-dichloro- 3-cyanoprop-1 -ene when treated with cuprous cycznide the yields being high.The identit,y of both products has been proved by their yielding the same crystalline tetrachloride (mixed melting point) on addition of chlorine. The unsaturated nitrile is proved to be 1 1-dichloro-3-cyano- prop- 1 -ene by hydrolysing it to succinic acid. The crystalline tetrachloride was proved to be pyyy-tetrachlorobutyronitrile by obtaining trichlorocrotono- nitrile when it was dehydrochlorinated with alcoholic alkali.' was hydrolysed to dichlorobut-3-enoic acid by heating it with a 2 1 1 mixture of acetic acid concentrated hydrochloric acid and water. The 4 4-dichlorobutenoic acid obtained melts at 42-43",' just as does the acid described by Auwers and Wissebach as yy-dichlorocrotonic.A mixture of our acid and that obtained by following Auwers and Wisse- bach's method had the same melting point. The products of chlorine and bromine addition to acids which had been synthesised following both methods also proved identical as shown by the absence of depression of the melting point of mixtures. As evidence that the acid that has been obtained by both routes has the 4 4-dichlorobut-3-enoic acid structure is the fact that according to Auwers and Wissebach its esters do not exhibit any exalta- tion of the molecular refraction and its hydrolysis in the presence of con- centrated sulphuric acid gives succinic acid. On the other hand the sug- gestion is opposed by the reduction of the acid with sodium amalgam to crotonic acid which was effected by Auwers and Wi~sebach.5~ The point a t issue is the more entangled by Auwers and Wissebach's having also obtained from their acid through the acid chloride an amide and a nitrile with constants differing from those of 4 4-dichlorobut-3-enonitrile and exhibiting an exaltation of the molecular refraction.The repetition of the syntheses of these derivatives according to the procedure of these authors starting from the acids obtained in both ways gave the same acid chloride amide and nitrile notwithstanding the origin of the parent acid ; but only the constants for the acid chloride agreed fully with those given by Auwers and Wissebach those of the amide and nitrile being different. As the structure of the starting materials for the syntheses of the unsaturated dichloro-acid by both methods has been proved (the structure of yyy-trichlorocrotonic acid by its acid hydrolysis to furmaric acid that The nitrile 7 3 G.Braun J . Amer. Chem. SOC. 1930 52 3172. 7 4 G. TV. Deodhar J. Indian Chem. SOC. 1934 11 83. 354 QUARTERLY REVIEWS of 4 4-dichlorobut-3-enonitrile by its acid hydrolysis to succinic acid) one has to assume that one and the same acid could have been obtained only by rearrangement or in the reduction of trichlorocrotonic acid or in the hydrolysis of the nitrile as is represented by the following scheme The answer has bcen found in the following correlation. On the one hand the addition of chlorine to 4 4-dichlorobut-3-enonitrile gave pyyy-tetra- chlorobutyronitrile and then in the usual way the corresponding acid acid chloride and amide.During the first step of these changes isomerisa- tion is hardly likely in the other ones it is impossible. The structure of t~etrachlorobutyric acid has been proved by dechlorination to yyy-trichloro- crotonic acid identical with that described by Auwers and Wissebach and of unambiguous structure this excluding the possibility of any isomerisa- tion taking place in the first step as well. On the other hand the unsaturated dichloro-acid invest)igated obtained by acid hydrolysis of the 4 4-dichlorobut-3-enonitrile was converted into acid chloride amide and nitrile the last having been found to be identical with the parent nitrile. The addition of chlorine to this acid its amide and nitrile resulted in compounds identical with /3yyy-tetrachloro- butyric acid and its derivatives described above.The results are correlated as follows CCl,:CH.CH,~CO,H 3. .1 CCl,:CH*CH,.COCl CCl,:CH*CH,*CN CCl CHClCH CN 7 * CCl,*CHCl.CH2*C0,H + CCl,*CH:CH*CO,H .1 J CCl,*CHCl*CH,*COCl CL 4 CCI,:CH.CH,CO*NH -+ CCl,*CHCl*CH,*CO*NH and show that the above-mentioned acid has the structure 4 4-dichloro- but-3-enoic acid no isomerisation taking place either when it is being pro- duced from nitrile or undergoing other changes indicated in the above scheme. Hence the reduction of yyy-trichlorocrotonic acid by zinc and glacial acetic acid in ethyl alcohol does not lead to yy-dichlorocrotonic acid a.s has been presumed by Auwers and Wissebach but results in 4 4-dichloro- but-3-enoic acid that is the reduction takes place with rea.rrangement Zn-HOAc CCl,*CH:CHCO,H - CCl,:CHCH,*CO,H NESMEYAXJOV ef al.TETRA- AND TRI-CHLOR0.ALKA4NES 355 Dichlorobuteiioic acid does not undergo further reduction under the same conditions. 72 It has also proved possible to elucidate the reason for this a t first sight obscure point in the reduction of the 4 4-dichlorobut-3-enoic acid to the crotonic acid with sodium amalgam. We have found that 4 4- dichlorobut-3-enoic acid its amide and nitrile are readily isomerised by a base (e.g. triethylamine) to yy-dichlorocrotonic acid and its derivatives. The resulting yy-dichlorocrotonic acid was identical with the acid synthesised from dichloroacetaldehyde and malonic acid 74 its structure being thereby ascertained. As the reduction of 4 4-dichlorobut-3-enoic acid had been effected by Auwers and Wissebach by the action of sodium amalgam it provided conditions for isomerisation to yy-dichlorocrotonic acid and its further reduction to crotonic acid.The conversion of yyy-trichlorocrotonic acid on one hand by hydrolysis to funiaric acid and on the other hand by reduction to crotonic acid is taken in the current reviews and textbooks as rigid proof of the trans- configuration of the crotonic acid and as an example of the determination of geometrical configuration by conversion into a derivative of known con- figuration by reactions without effect on the olefinic bond. Now we are however able to see that the two-step reduction effected '19 Auwers and Wissebach of the yyy-trichlorocrotonic acid t o crotonic acid proceeds with twofold isomerisation which they had failed to notice that it cannot therefore serve to determine the configuration of the crotonic xid and that allylic rearrangement substantially restricts the method of determining configuration by conversion into a derivative of known steric configuration.That we obtained acid chlorides amides and nitriles of both 4 4-dichlorobut-3-enoic acid and yy-dichlorocrotonic acid enabled us to determine that the acid chloride described by Auwers and Wissebach is the 4 4-dichlorobut-3-enoyl chloride their amide and nitrile being on the other hand yy-dichlorocrotonic acid derivatives. It is evident that the conversion of the 4 4-dichlorobut-3-enoyl chloride into the amide was itccompanied by isomerisation to give the amide of yy-dichlorocrotonic acid which had passed unnoticed by them. It is to be pointed out that both series of derivatives CCl,:CH*CH,X and CHCl,*CH:CHX fail to isomerise in acids.Thus hydrolysis of 4 4- tlichlorobut-3-enonitrile in hydrochloric and acetic acid gave 4 4-dichloro- but-3-enoic acid and that of yy-dichlorocrotononitrile under the same conditions gave yy-dichlorocrotonic acid. The same is true of yyy-trichloro- t-rotonic acid which in strong acid hydrolyses without isomerisation to yield fumaric acid. It follows then that the production by Auwers and iViseebach of 4 4-dichlorobut-3-enoic acid by reducing yyy-trichlorocrotonic :.cid in acid cannot be explained either by the preliminary isomerisation of the parent acid or by intermediate formation of yy-dichlorocrotonic acid followed by isomerisation. Rearrangement seems to be taking place in the very process of reduction. IYe believe this reaction also t o follow the '' transfer of reaction centre " mechanism.2 356 QUARTEBLY REVIEWS Allylie Rearrangements in the Substituted Polyhalogenoaml Alcohol Series Allylic anionotropic rearrangements of substituted allyl alcohols involving a halogen atom a t the double bond have been investigated in detail only in the case of monochloro-derivatives HO*CRR’*CH:CHCl producing in the course of rearrangement unsaturated aldehydes. 75-77 As far as dichloro-derivatives of allyl alcohols are concerned it has been noted that for example 1 l-dichloro-3-hydroxy-3-methylnon-l-ene-4-yne fails to rearrange to the corresponding unsaturated acid.76 1 1 3-Tri- chloroprop-l-ene 1 and 1 1 3-trichlor0-2-methylprop-l-ene~~~ involving the >CCl*b:CCl system also did not undergo allylic rearrangement.It seemed of interest to investigate in detail the possibility of allylic rearrange- ment of polyhalogenoallyl alcohols (or their ethers) according to the fol- lowing scheme (1) (11) (111) HO*CRR’*CH:CX -+ (CRR’:CH.CX,.OH) + CRR’:CHCO,H HO *CRR’*CX:CHX + (CRR’:CX*CHX*OH) -+ CRR’:CX*CHO HO*CRR’*CX:CX -+ ( CRR’:CX*CX2.0H) + CRR’:CX-C0,H Compounds (I) containing alkyl groups as substituents did not re- arrange even under vigorous conditions neither 1 1 -dichloro-3-hydroxy- hept- 1 -ene after prolonged heating with 10% sulphuric acid in aqueous alcohol or aqueous dioxan nor 1 1 -dichloro-3-ethoxyhept-l-ene when heated in acetic acid in the presence of sulphuric acid was changed. 1 1 -Dichloro-3-hydroxy-3-methylbut- 1 -ene in acid medium easily loses water giving 1 l-dichloro-3-methylbuta-1 3-diene and 1 l-dichloro-3- methoxy-3-methylbut-1 -ene when heated with 10% sulphuric acid in aqueous alcohol remains intact.Compounds (I) where R = aryl and R’ = hydro- gen readily rearrange in acid to arylacrylic acids 11 H+ H& ArCH(OH)*CH:CCI + ArCK:CH*CO,H The reaction is carried out by heating the aryl derivative in acetic acid in the presence of hydrochloric acid. As aryl substituents we used phenyl p-tolyl a-naphthyl and pchlorophenyl. 1 l-Dichloro-3-hydroxy- 3-phenylbut-1 -ene does not however rearrange under these conditions or with 10% sulphuric acid in aqueous alcohol only losing water to give 1 1 -dichloro-3-phenylbuta-l 3-diene H+ HO*CPhMe*CH:CCl + CH,:CPh*CH:CCI The latter when heated dimerises with loss of a hydrogen chloride 75 E.R. H. Jones and B. C. L. Weedon J. 1946 937. 76 I. M. Heilbron E. R. H. Jones and 31. Julia J. 1949 1430. 7 7 M. Julia Ann. Chiim. 1950 5 595. ‘*A. Kirrmann and R. Jacob Bull. SOC. chirn. France 1940 7 586. X-ESMETA4NOV d at!. TETRB- AND TRI-CHLOROALHAXES 357 niolecule to yield a product of the composition C,,H1,C13. 3-Uoxy-3-aryl- 'I l-dichloroprop-1-enes when hea,ted in acetic acid in the presence of hydrochloric acid are smoothly converted into arylacrylic acids possibly witah no niigration of the ethoxy-group taking place the reaction proceeding by the .' transfer of reaction centre " inecha,nism L(7 Et03 H f OH AreCH-CH=CCl -+ ArCH:C'H*CO,H t EtOH + HCl + 1 P -Dichloro-3-phenylprop-l -ene however reinains unaltered when heated with acetic acid in the presence of hydrochloric acid.The presence of an electron-accepting substitueiit in compounds (I) hinders anionotropic rearrangement even if owing to double- bond transfer there is the possibility of a conjugated system's being formed. Thus 1 1- dichloro-3-cyano-3-ethoxy(or hydroxy)prop-1-ene when heated in acid does not undergo aiiionotropic allylic rearrangement which would have pro- duced fumaric acid only hydrolysis of the nitrile group to carboxyl taking place. The prototropic allylic rearrangement of 1 1 -dichloro-3-cyano-3-ethoxy- prop-1-ene takes place readily however with triethylamine (to be compared witlh ref. 8). The product when heated with dilute hydrochloric acid undergoes an interesting conversion into chloroacrylic acid H+ H*O CCI ,:CH.CH( OEt ) C N -+ CHCl,*CH:C (OE t)*CN -F (CHCl,*CH,*CO*CN) -j.CHCl:CH*CO,H Synthesis of compounds (I) was carried out by the following routes 2- 3 1 1 -dichloro-3-hydroxy-3-methylbut-l-ene by the action of methylmag- nesium iodide on 4 4-dicl1lorobut-3-en-2-one ; and 1 l-dichloro-3- hydroxyhept-l-ene 1 l-dichloro-3-ethoxyhept-l-ene 3-aryl-1 l-dichloro- 3 - h ydr oxypr op- 1 -ene and 1 1 -dichloro- 3 - ethox y - 3 - pheny lpr op - 1 - ene by the action of the corresponding organomagnesium derivative on pj3-di- chloroacraldehyde or 1 1 3-trichloro-3-ethoxyprop-1-ene CCl,:CH*CHCl*OEt 3- RMgX + CCl,:CHCHR*OEt 1 l-Dichloro-3-p-chlorophenyl-3-hydroxyprop-l-ene was produced in the following way CBH5Cl + CCI,:CH*COCl -+ Cl*C,H,*CO*CH:CCl AICI Al( OYr') Cl*C,H,*CH( OH) *CH:CCl :,Ye prepared PP-dichloroacrylic acid needed for this reaction by oxidising /3/3-dichloroacraldehyde with chromic anhydride in acetone.1 1-Dichloro- 3-cyano-3-ethoxybut-1-ene was obtained by the action of cuprous cyanide on 1 1 3-trichloro-3-ethoxyprop-I-ene. As far as allylic rearrangements are concerned compounds (11) behave as do compounds (I). With alkyl substituents rearrangement does not occur even under vigorous conditions. 358 QUARTERLY REVIEWS With 10% sulphuric acid in aqueous alcohol 1 2-dibromo-3-hydroxyhex- 1-ene 1 2-dibromo(and 1 2-dichloro)-3-hydroxy-3-methyl-pent-l-ene and 1 - (1 2-dibromoviny1)- 1-hydroxycyclohexane are recovered unchanged a t room temperature and when heated lose water or form a slurry. 1 2-Dibromo-3-hydroxy-3-phenylprop-l-ene and 3-acetoxy-1 2-di- chloro-3-phenylprop- 1 -ene readily rearrange to the corresponding a-halo- genocinnanialdehyde when heated in acetic acid containing hydrochloric acid Ph*CH( 0H)eCBr:CHBr + Ph*CH:CBr.CHO 1 2-Dibromo-3-hydroxy-3-phenylbut-l-ene like 1 l-dichloro-3-hydroxy-3- phenylbut-1-ene does not rearrange in acid medium but loses water.It is noteworthy that 1 -chloro-3-hydroxy-3-phenylbut-l-ene rearranges to the aldehyde. The behaviour of 1 2-dibromo-3-hydroxy-3 3-diphenylprop- 1 -ene in acid medium presents some peculiarities. With hydrochloric or hydrobromic acid in acetic acid it yields first 1 2-dibromo-3-chloro- and 1 2 3-tribromo-3 3-diphenylprop-l-ene respectively which when heated in 90% acetic acid produce 1 2-dibromo-3-phenylindene rather than a-bromo-@-phenylcinnamaldehyde though the two trihalogeno-derivatives and 1 2-dibromo-3-hydroxy-3 3-diphenylprop- 1-ene with 2 4-dinitro- phenylhydrazine in alcohol-sulphuric acid yield the 2 4dinitrophenyl- hydrazone of a-bromo-/3-phenylcinnamaldehyde.1 2-Dibromo-3-phenyl- indene is produced immediately from 1 2-dibromo-3-hydroxy-3 3-di- phenylprop-1-ene when its acetic acid solution is treated with sulphuric acid or better perchloric acid. HC H2O Ph HRr HClO /\/\ Ph,CBr.CBr:CHBr +- Ph,C(OH)*CBr:CHBr .-> jl \i pr HOAc HOAc v - - B r Ph,C:CBr.CH:N.NH.C,H,(PU'Q,) The formation of 1 2-dibromo-3-phenylindene can be best represented as follows The structure of 1 2-dibromo-3-phenylindene follows from t'he o-carb- oxybenzophenone obtained by oxida'tion with potassium permanganate. Compounds (11) were produced by adding bromine to the corresponding ncetylenic alcohols.NiGSJIEYANOV et Cd. TETRA- AND TRI-CIILOR0~4LKASES 359 The possibility of rearrangement in the system HO*CRR'*CX:CX has investigated in the case of 1 1 3-trichloro-3-hydroxy-3-phenylprop- :-ene. When heated in acetic acid in the presence of hydrochloric acid ;S was converted into cx-chlorocinnamic acid 23 rearrangement taking place inuch less readily than with 1 l-dichloro-3-hydroxy-3-phenylprop-1 -ene ; IiO CHPhCC'I CC'I + PhCH CCI.CO,H To obtain a 750/ conversion of the trichloro-compound into a-chlorocinnamic x i d the former has to be heated for 25 hours whilst the 1 l-dichloro- ilerivative is converted into cinnaniic acid in 30-35 minutes. 1 1 2- 'i'richloro-3-hydroxy-3-phenylprop-l-ene was obtained by reduction of the 1 Ihenyl cx&3-trichlorovinyl ketone with aluniiniuni isopropoxide.€l+ H,Q AICI Al(OPr') ( '6H6 + c ~ l ~ c c 1 ' c o c ~ C,H,'CO*CC1:CCI C,H,-CH( OH).CCI:CCI I t will be noted that 3 3-clialkoxy-1 l-dichloroprop-l-enes are converted j rito @@-dialkoxypropionic esters by simply boiling them with alcohols. lo This change seems to be due to allylic isomerisation according to the + cheme 'CI,:CH*CH(OR) -+ (ROCH:CHCCl,*OR) + ROH RC)-CH:CH*COCl + RC1 + (RO),CH*CIS,*CO,R ('r through addition of alcohol to the dichlorovinyl group ['Cl,:CH*CH(OR) + ROH + [(R0)2CH*CH2*CCl,.0R] + (RO),CH.CH,*CO,R ii:~-nialkoxypropioiiic esters can also be produced by boiling 3-alkoxy- II 1 1 3-tetrachloropropane or 3-alkoxy-1 1 3-trichloroprop-l-ene in Elcohol.10 Supplement Synthesis of Higher aaao-Tetrachloroalkanes and aaa-Trichloroalkanw by the Telomerisation Reaction.-Joyce and Hanford 28-30 who discovered the telomerisation reaction of ethylene and carbon tetrachloride and of ethylene and chloroform have shown that a mixture of cxcxaw-tetrachloro- alkanes and cxxcx-trichloroalkanes are formed and have isolated the corre- snonding individual compounds containing 3-9 carbon atoms in the former a tid 3-11 carbon atoms in the latter.As it is most difficult t o synthesise organic molecules of an average niolecular weight and more than about 10 carbon atoms it seenis worth- 13 Me to determine the conditions for the telomerisation reaction which ~ " i ould provide for higher tetra- and tri-chloroalkanes and for their isolation s.-; pure substances. Two of us and Sh.A. Karapetyan 24 have slio1~7n that higher tetra- and tri-chloroalkanes can be obtained by the reaction of ethylene with carbon t4rachloride and chloroform under comparatively low pressures from 100 360 QUARTERLY REVIEWS to 150 atmospheres. The yield of higher polychloroalkanes is about equally dependent on the initial pressure at which the reaction is being run and on the initial molar ratio of ethylene to halogenomethane. The higher these two parameters are the higher is the yield of polychloralkanes. This relation is due to the fact found by G. D. Yefremova and G. G . Le~nt'eva,'~ that a t 100" and pressures above 105 atmospheres the ethylene- carbon tetrachloride system is homogeneous whatever its composition. Thus the reaction between ethylene and carbon tetrachloride taken at 20 1 mole ratio a t 150 atmospheres and go" in the presence of azo- bisisobutyronitrile produced a mixture of aaacco-tetrachloroalkanes consisting of tletrachloropentane (9%) tetrachloroheptane (12%) a fraction (2474) of tetrachloroalkanes with 9-15 carbon atoms a paraffin-like fraction (440 &) of tetrachloroalkanes soluble in acetone with an average molecular weight of 420 and a tetrachloroalkane fraction (11yL) insoluble in acetone with an average molecular weight 720.The molecular weight of the fractions was ascertained by chlorine determination. Increase of the molar ratio of ethylene to carbon tetra- chloride other things being equal is accompanied by the increase in yield of higher tetrachloroalkanes as well as by increase in the average molecular weight of the higher fractions.The same is observed in the reaction with chloroform this leading to the suggestion that in the ethylene-chloroform system critical phenomena also take place under the indicated conditions. From the mixtures obtained by telomerisation we have isolated higher tetrachloroalkanes with 13 and 15 carbon atoms l3 and trichloroalkanes with 13 15 and 17 carbon atoms,24 their constants being given in Table 3. Synthesis of Compounds containing Two and Wee Functional Groups starting with a a aw -Tetr achloroalkanes .--MMXW - Tet ra c hloroalk an es have a1 - ready been shown to undergo chemical changes by the action of nucleo- philic electrophilic and radical reagents reaction taking place selectively. either a t the chloromethyl or a t the trichloromethyl group. By combining successively reactions of the two types it is possible to effect the synthesis of various compounds involving two functional groups.Thus reaction of aacw-tetrachloroalkanes and the readily available ccaacco-trichloroalk- 1 -enes with ammonia 5 resulted in aminotrichloroallranes CCl,*[CH,];NH and aminodichloroalkenes CCl,:CK*[CH,];NH,. Hydro- genation of the polychloro-nitriles CCl,. [CH,],.CN and CCl,:CH*[CH,],.CN with hydrogen under pressure in the presence of Raney nickel produced the amines CHC1,*[CH,],*NR2 and CCl,:CH*[CH,],*NH,. Tetrachloroalkanes and trichloroalkenes reacting with sodium sulphide gave (CCl,*[CH,],),S and (CCl,:CH*[CH,],),S which on being hydrolysed gave cli- (co-carboxyalkpl sulphides (see Table 5). Another route for obtaining compounds with two functional grocps con- sis ts in the hydrolysis of a a ~ u - t e tra c hloroalkanes to CL - chloro- carbox ylic acids followed by chlorine substitution under the action of nucleophilic reagents.This procedure resulted in co-amino-carboxylic acids 2sc with 7 9 G. D. Yefremova and G. G. Leont'eva Trudy Q.I.A.P. Sh. 1954 5. XESMEYAYOV et al. TETRA- ASD TRI-C~LOROALKAXES 361 7 9 and 11 carbon atoms in a molecule di-(cu-carboxyalkyl) sulphides,6 cu-hydroxy-carbosylic acids a,nd many other compounds some of which are listed in Tables 5 7 and 8. The conjugated addition of chlorine to the dichlorovinyl group in con- centrated sulphuric acid leading to a-chloro-carboxylic acids,l7 reveals new possibilities for the synthesis of compounds involving three functional groups. Thus two of us and R.G. Petrova,26 starting with 1 1 5- trichloropent-1 -ene have obtained DL-proline and DL-ornithine according to the scheme R-[CHz],*CHC1*CO,H -+ ~U'H,*[CH,],.CH(NH,).CO,H (R = pht.halimidoj The conversion of 1 1 5-trichloropent-1-ene into 3 5-dichloropentanoic acid and that of 1 1-dichloro-5-phthaliinidopent-1-ene into 2-chloro-5- phthalimidopentanoic acid was achieved by chlorine addition in sulphuric acid a t 0-5" the reaction giving good yields (cf. ref. 17). When chlorine was added under the same conditions to 5-amino-1 1- ctichloropent-1-ene and the mixture neutralised with 5% ammonia the main product was DL-proline along with a small amount of 5-amino- 1 1 1 2-tetrachloropentane and 5-amino-2-chloropentanoic acid. The formation of proline seems to indicate that the intermediate 5-amino-2- chloropentanoic acid readily cyclises in weakly alkaline solutions.The literature describes the preparation of DL-prOline by ammonolysis of 2 5-dichloropentanoic acid with 25o/d aqueous ammonia for one hour a t 130" the yield of proline being 350/,.80 It has now been found 26 that 50-550/ yields are obtained when the ainmonolysis of 2 5-dichloropentanoic acid with 2.5:; ammonia solution is carried out a t room temperature for 14 hours. The isolation of DL-proline was as follows After removal of ammonia the aqueous solution was passed through S.D.V. cation-exchange resin to liberate chloride ion (cf. ref. 2 7 ) . Proline was eluted from the resin with S?/ aqueous ammonia. After removal of ammonia the residue was dissolved in alcohol and precipitated with dioxan.The resulting proline was purified by recrystallisation from alcohol. Taking into account the exceedingly ready availability of 1 1 1 5tetrachloropentane and the high yields obtained when synthesising the intermediate products-1 1 5-trichloro- pent-1-ene and 2 5-dichloropentanoic acid-this synthesis is to be con- sidered the easiest and the most conveaient one a t present available. 8o R. Gaudrp and L. Berlinquet Canad. J . Res. 1949 27 B 262. 362 QUARTERLY REVIEWS DL-Ornithine was obtained by the ammonolysis of 2-chloro-5-phthal- imidopentanoic acid with 2.57; aqueous ammonia in the presence of am- monium carbonate and was isolated as its monohydrochloride after refluxing the product with concentrated hydrochloric acid. Ornithine was identified as ornithuric acid or picrate.Along with ornitliine there was formed proline whose yield (30%) was determined by isolation of the product of condensation with isatin. 2G The instances shown undoubtedly do not cover all the possibilities of synthesis of natural and other cc-amino-acids from ccccrxm-tetrachloroalkanes and cccccc- trichloroalkanes. o-Chloro-carboxylic Acids and some of their Reactions. 13-The higher o-bromo-carboxylic acids Br*[CH,];CO,K have been investigated in detail ar,d are widely used in various syntheses. This is not true of the corre- sponding m-chloro-carboxylic acids which have received comparatively little attention presumably because the chlorine atom was thought to be fairly unreactive and because of the difficulty of production. We have now investigated some reactions of oJ-chloro-carboxylic acids and found that they can be successfully used in place of the corresponding bromo- acids.The corresponding alkoxy-derivatives are smoothly formed from the reaction of sodium allioxide and ethyl 6-chlorovalerate and ethyl 7-chloroheptanoate. The reaction of ethyl 6-chlorovalerate and ethyl 7-chloroheptanoate with diethyl sodiomalonate in the presence of sodium iodide gives 1 1 5-triethoxycarbonylpentane and 1 1 7-triethoxycar- bonylheptane in a good yield ; these are hydrolysed with dilute hydrochloric acid to pimelic and azelaic acid respectively. Salts of 7-chloroheptanoic and 9-chlorononanoic acid readily react in aqueous solutions with sodium phenoxide or sodium cyanide or are hydrolysed by alkali to co-hydroxy-carboxylic acids. Both these chloro- acids with sodium sulphide in aqueous solution readily yield the corre- sponding di-(m-carboxyalkyl) suIphide The yield of ornithine was 30:h.PhOfCH21n.C02H k PhOSa \ Na,S / ClfCH,],.CO,X'n RCY \a,Co3 L HO*[CH2]n*C02H J NC*[CH,],,*CO,H a'eid acid Oxidation of 7-hydroxyheptanoic acid with concentrated nitric res pimelic acid in good yield ; hydrogenation of 7-cyaaolieptanoic animonia solution with nickel catalyst leads to 8-amino-octanoic acid. The action of sodium phenoside or sodium cyanide in aqueous solut>ion on s&s of 6-chlorovaleric acid does not give the corresponding derivatives but' leads to 6-valerolactone (or 6- hydroxyvaleric acid) which under these conditions does not react further. On the ot'her hand sodium thiophenoside with 6-chlorovaleric acid gives 5-(pheny1thio)valeric acid in high yield.XES&IEYANOV et Ul. TETRA- SND TRI-CHLOROALIIANES 363 Phenosyvaleric acid may be obtained from 6-chlorovaleric acid oiily through 6-valerolactone which is heated with anhydrous sodium phenoxide a t a high temperature just as y-phenosybutyric acid is obtained via y-butyro- lactone.81 Heating 6-chlorovaleric acid with anhydrous ammonia a t 230-250 O or its ethyl ester with alcoholic ammonia a t 120-140" gave a-piperidone in good yield. 7-Chloroheptanoic acid was converted into hept-6-enoic acid through 7-trimefhylaminoheptanoie acid betaine the betaiiie being split by alkali C1.[CH2],C02H + Me,N*[CH,],.CO,L -+ CH2:CH.[CH2],.C0,H + Me3N \?7e have investigated the reaction of co-chloro-carboxylic acids with benzene in the presence of aluminium chloride; it has been previously shown that y-chlorobutyric acid with benzene gives y-phenylbutyric acid.82 Similarly 6-chlorovaleric acid gives 8-phenylvaleric acid in a high yield when aluminium chloride is used in the molar ratio 1 1.The amount of aluminium chloride being increased in addition to phenylvaleric acid one obtains a-ben~osuberone.1~ i 0 The yield of 6-phenylvaleric acid obtained from 6-chlorovaleric acid is higher than stated to be obtainable from reaction with 6-valerolactone.s3 Unlike the lower acids 7-chloroheptanoic 9-chlorononanoic and 11 -chloro- undecanoic acids react with benzene in the presence of aluminium chloride undergoing isomerisatioi t ; the phenylheptanoic phenylnonanoic and phenylundecanoic acids obtained have been proved to be different from the known 7-phenylheptlanoic,s4 9-phenylnonanoi~,~~ and 11 -phenylundeca- noic acids.S6 Oxidation of phenylnonanoic and phenylheptanoic acids with chromic anhydride in acetic acid led to the isolation of acetophenone," which suggests the presence in the acids of the C,H,*CHMe grouping.Condensation of 6-chlorovaleroyl and 7-chloroheptanoyl chlorides with benzene in the presence of aluniinium chloride proceeds as usual and gives the corresponding o-chloroalkyl phenyl ketones SlC1 C',H + Cl*CO*[CHJ,,.Cl + C,H,*~O'[CH2]n*CI where n = 4 or 6. 81 G.P. 711,687 (Zeitt. 1944 1 907). a 2 I. Eykinann C'hem. Weekblad 1907 4 727. E3 R. Christian J . -4wer. Chetn. SOC. 1952 74 1591. 84 J. von Braun Ber. 1911 44 2S78. 8 5 H. G. Raper and E. J. Wayne Biocizem. J. 1088 22 194. * Similar oxidation of S-phenylvaleric acid gives 7-benzoylbutyric acid.E. Fourneau aid P. Baranger Bull. Xoc. c h h . Frunce 1931 49 1161. 364 ~ io. - 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 16 .7 18 9 !O !1 ! 2 !3 ?4 !5 !6 37 38 39 30 31 32 13 34 35 16 17 38 39 $0 11 k2 k3 14 15 $6 17 18 19 50 51 52 53 34 55 T6 ~ Compound QUARTERLY REVIEWS TABLE 3. Polychloro-hydrocarbons. Me.CH:CCl . . . . CH,:CH*CC1 . . . . CCl,:CH.CH,Cl . . . CH,:CCl*CHCl . . . CH,:CCl*CCI . . . . CH,Cl*CCl:CCI . . . CHCl:CCl.CCl . . . . CH,ClCHCl*CCl . . . CH,Cl.CCl,.CCl . . . CH,:CBr*CCl . . . . CH,BrCHCl.CHCl . . CH,Br*CHBr.CCl . . . CC1,CH:CHMe . . . CC1,:CHCHMeCl . . . CCl,-CH,.CH:CCI . . . CCl,:CH.CCl:CCl . . . CBrCl,CHCl*CH,*CCl . CBrCl,*CHBr*CHMeCl . Me*CH,-CH,.CH:CCl . . CCl,:CH*CMe:CH . . . CH,Cl*CH,*CH,*CH:CCl . CCl,CH:CMe .. . . CHCl,.[CH,],CH,Cl . . CCl3*CHC1fCH,],*31e . . CCl,:CC1~[CH2],CH,Cl . CCl,:CHCH,*CH:CCl . . CCl,:CH*CH,CH,*CCl . CCl,:CHCH2CHC1*CHClz CCl,.CHClfCH,],.CH,Cl . CCl,:CH*CH,*CHCl-CCl . CBr,*[CH,],CH,Cl . . CBr,*[CH,],*Me . . . CC1,fCH,],*CH21 . . . Me.[CH,],.CH:CCl CCi,:CH*[CH,],*CH:CCi CH,Cl*[CH,],*CH:CCl . CHC1,.[CH2],*CH,C1 . . CCl,.CH:CMCMe . . . CCl,*CH,*CHBrCMe . . CCl,:CH-[CH ],.Me . . CH,C1fCH2]6*CH:CC12 . CH,C1*[CH,]8*CH,CI . . (CH,Cl*[CH,],.CCl:) . H.[CH21,,.CC1 . . . . Cl.[CH,],,.CCl . . . . CH,Cl*rCH2],,*CH2C1 . . (CH,CI.[CH,],CCl:) . . (CH,CI.[CH,],*CCl,-) . . H*[CH,J,,*CCI . . . . H*[CH,J,,*CCl . . . . Ph*CH,.CH:CCl . . . CC1,CH:CHPh. . . . Ph*CH,.CCI:CCI . . . CCl,:CHCMe,Cl . . c1* [CH,~,~LCCl,],*[CH,],~Ci CCl,~[CH,],,~Cl . . . M.p.0 84-% 38 -39 57-58 B. p.o,,’nm~. 77-78 101-102 131-132 1 26-1 2 7 54-55/30 6 8-69 / 30 59-60;s 64-65/8 101--102/15 54-55/10 76-7 7 /3 92-93/24 6 7-57.5/49 44-5-46 /3 54-55/3.5 93-94/3 S i / l 68/52 137-125 30-3 P ;S Cj8-69/7 45-46/8 68-58.5/ 15 84/8 72-7318 101-104/12 121-122/13 101-1o2/2 92-93,/8 57/1*5 106-107/7 119-120/lO 85-86 18 7 8-791 1 *5 68-691 14 98-100/’8 66-6i/1 74-751 1.5 64-65/ 10 85-SS/7 9 1-92/16 105-1 0611.5 152-154/3 103-108/0.3 141-142/5 143-144/1*5 90/5 152-153/1*5 178-180/2 123-125/0*3 1 3 3-1 43 /0*3 93-94/6 91-92/1 121-1 22/43 1.4450 1.4680 1493s 1,4840 1.5000 1.5 160 1.5282 1.5105 1.5382 1.632i 1.5290 1.5640 1-4810 1.4816 1.5172 1.5620 1-547s 1.5590 1.4548 1.5027 1.4892 1.4822 1.4847 1.4788 1.4825 1.5113 1.5197 1.5125 1.5285 1-5135 1-5291 1.5655 1.5390 1.5480 1.4589 1.5149 1.4850 1.4776 1.4725 1.5030 1.4697 1-4834 1.4620 1.1687 1.3292 1.3940 1.3843 1.5099 1.5409 1.6449 1.6117 1.7187 1.8442 1.8332 2.1712 1.2972 1.3026 1,6607 1.6142 1.8859 2.0466 1.0899 1.153i 1.2724 1.2497 1.2527 1.2438 1.3339 1.4121 1.4307 1.4707 1.5077 1.4807 1.5817 2.0902 1.9882 1.8086 1.0430 1.3828 1.1902 1.1744 1.1403 1.4793 1.0106 1.7352 0.9992 1.5055 1.2202 1.4649 1.0339 1.4842 I 1.1290 1.4522 1.1558 14980 1.1393 14658 1.4663 1.5490 1.5710 1.5630 1.0142 0.9992 1.2032 1.921; 1.3232 Ref.4 I- 3 3 2 1 3 1 3 I 3 1 3 1 3 1 3 1 2 1. 3 *> 9 i l i ) 9 A- 2 2 32 4 23.23 4. 16 23 22 16 9 9 18 7 7 9 17 9 9 0 4 7 4 16 22 19 4 16 16 1 6 24 1 3 13 16 16 16 24 24 1 3 1 3 14 9 1 n 0 3 Y d 99 - TABLE 3.-continued. 23-CI.C,H,.CH,.CH:CClz . CCl,.CHClCH,Ph . . . CCI,*CCl,-CH,Ph . . . CCl,*CHCI*CH,*C,H,Cl-~ .CCl,:CH-CPh:CH . . . CCI,:CHCHPh&k . . . CCI,.CHCi.CHPhMe . . CC1,:CH-CPh59e2 . . . CC!,:CH-CHPln . . . p-Br.C,N,.CH,.CH:CCI . ;U-Br.C,H4.CH,.CHC1.CCI,. CHBr:CBr.CPh,Cl . . . CHBr:CBr.CPh,Rr . . 1 2-Dibromo-3 -phenyl- indene 2,I.p.O $6-77 90 89 137-138 152-l53 82-83 E.p.o/'mm. 115-1 16/6 111-112/2 86-8$/ 1.5 S0-81/1 142-143/5 73-74/ 1.5 107-108/1-5 142-143/1 117~5-118/5 rig 1.5630 1.5535 1.5829 1.5423 1.5568 1-1540 1,5951 1.5830 1.3208 1.3867 1.2048 1.1702 1.3634 1.541 1 1.2 180 1.5532 20 14 14 20 23 22 22 22 22 20 20 23 23 23 TABLE 4. Chlo.ro-deri;z/atives contuining nitro- hydroxy- or 0x0-groups ; chloro-estm ~ x u 1 2 3 4 5 6 8 9 10 11 12 13 14 140 15 16 17 18 19 20 21 32 23 24 25 26 27 28 29 30 I- ___ C'onipound CCI,:CH*CH,.NO . . . CCI,*CHClCH,.NO . . CCl,:CH.CH,*OH .. . CCI,:CH.CX-I,.OMe . . . PhO*CH,*CH:CCl . . . CCI,:CH*CH,.OAc . . . CCl,:CCl*CH,*OMe . . . CCI,:CH-CHCl*OEt . . CCl,.CHCI*CH,.OR . . CCl,.CHClCH,.OMe . . CCI,CH,.CHCl*OAr . . CC1,~CHC1*CH2*OAc . . (CC1,*CHCI.CH,),S04 . . CCI,.CCI,~CH,.OBIe CC1,CHC1*CH,*OCH2C1 CCl,:CH*CHO . . . . CCI,:CH,.CH( OEt j3 . . CCI,:CH*CH( OBu) . . CCI,:CH*CHMe*OH . . CC1,:CHCHMe.OMe . . CC1,:CH-CHMeB-lc . . Bz.~CN,],*Cl . CCl,:CH*[CH,],*bH CCI,:CH*[CH,],.OE t GCI,:CH.[CH,] ,*OPh CCl,:CH*[CH,],-Ohe CCI,:CH.CNle,*OH . CCI~:CH*CMe~.OMe CCl,:CH.CMe,OEt . CCl,fCH,],*OH . CCI,.[CH,],.O,lc . . . . . . . . . . . . . . . . . . . . . M.p.0 39-40 46-47 49-50 B . p . * ,'nini. 47-48 / 2 73-74/ 15 56-57/4 P 31-1 32 103-104/7 53-54/10 71-i2/14 i6-7 i/ 14 5 8*5-59/4 i0-7 1 /2 8 1-82/2 189-190/2 63-64/2 sn-s2/1.5 1 24-1 2 5 83-84/15 124-125/9 i 2 / 1 0 64/57 83-84/26 56-60/2*5 112-113/2 58/1-5 GO-6 1/20 5 1-52! 10 i9-81 ill '72-i3/24 113-113/10 99-100/3.3 n20 1.4879 1.5005 1.4945 1-4558 1.9534 1.4650 1.4871 1.4797 1.4868 1.4721 1.4848 1.5085 1.5070 1.5067 1-44i8 1.4530 14792 1.1722 1.4590 1.4923 1.4642 1.5375 14690 147-80 1.4628 1.4616 1.4897 1.4700 D continued a"," 14494 1.6355 1.3763 1.3112 1.2718 1.2846 1.3782 1-2866 1.4568 14455 1.4905 1.5765 1.5713 1.1293 1.0495 1.2745 1.4580 1.2234 1.2452 1.1101 1-1914 1.2095 1.2220 1.1418 1.1101 1.3431 1.2859 Ref.14 14 18 7 20 18 14 LO 21 18 14 10 1s 18 7 14 18 11 LO 21 10 19 2," 22 22 13 12 19 19 12 23 22 22 13 12 ?n next page. 366 QUARTERLY REVIEWS TABLE 4.--co?ztinued. NO. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 - _- NO.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Compound CCl,.[CH,],-CHO . . . CHBr:CBr*CHPr*OH . . CHBr:CBr*CJfeE t*OH CCl,:CHCH( OH)fCH,],*l\ic CCl,:CH*[CH,],*CH,.OEt . 3C1,:CH*CH(OEt)fCH2],-Mt CCl :CH-CH( OMe).CMe3 . Cl-[CH,],.COPh . . . ClfCH,],*CPh:N*NH. CCI,:CH*[CH,],*OEt . . [CH,] > C(0HjCBr:CHBr PhCH:CCl-CHO . . . CHCl:CClCHPh*OH . . CC1,:CH.CHPh.OH . . CC1,:CH.CHPh.OEt . . CCl,:CHCH,*C,H,.OH-o . CCl,:CH*CH,.C,H,.OMe-p CC1,:CH~CH2~C,H,~OMe-o CC1,:CClCHPh.OR . p-Cl-C,H,*CH( OH)*CH:CCI p-Cl-C,H,*CO.CH:CCl . Ph*CBr:CH*CHO . . . CHBr:CBr.CHPh*OAc . CCl,:CH*CPhMe*OH . . CHBr:CBrCPhMe.OH . CCl,:CH.CH( OE t)C,H,Me-$ CPh,:CBr*CH:N*NH* CHBr:CBrCPh,*OH . . CCl,:CH*CH( OEt)C,,H -a a-Suberone . . . . . C6H3(N02)2 CCl,:CHCH,C,H,*OH-p . C6H3(N02)2 TABLE 5.C Coinpound (CCl,:CH.CH,),S . . . (CCl,:CH.CH,),SO . . (CCl,:CH.CHMe),S . . (CCl,:CH-[CH,],),S . . (CCl,.CH,.[CH,],),S . (GCl,-CH,*[CH,],),SO . . (CCl,.CH,*[CH,],),SO,. . (CC!,:CH-[CH,] j,S . . CCI,:CH*[CH,]6-CH,*SEt . CCl,:CH.CMe,.SPh . . (HO,C.[CH,],',,S . . . (EtO,C.[CH,],),S . . . (BuO,C*[CH,],),S . . . (HO,C-[CH,],),SO . . (HO,C*[CH,],),S . . . (EtO2C-[CH,1,),S . . . (BuO,C.[CH,j,),S . . . (HO2C~[CH,],),SO . . 3I.p." 34-35 1 10-1 1 1 73-74 158-159 5 7-5 8 40.5-41 51-52 70-71 245-216 112-1 13 1\I.p.O __. ~ ~ 114-115 35-36 67-68 114-1 15 96-97 156-157 95-99 37 22-23 148-149 B.p."/mm. 103-104/8 7s-79/1 98-99/7 103-104/8 114-116/15 84-S5/9 60-61 /9 147-1 48,' 1.5 88-S9/1 128-129/5 132-1 33 /9 105-106/2 90-91 /1 130-1 31 /3 1 16-1 17/3 1 15-1 19/5 112-1 13/1.5 142-1 43/3 168-1 69/8 139-1 40/ 1 '5 106-1 07/2 134-1 35/2 106-107/1*5 109/4*5 l59-160/4 124-125/7 contuining SB ~ _ _ _ _ B.p."/ mm.92-94/1 104/5 117/1 203-205/5 186-1 87/1.5 102-1 03/ 1.5 183-1 84/1.5 120-1 2 1 / 1 2 14-215/ 1.5 31 9-2!30/2 239-111/1.5 1.4890 1.5380 1.5380 1.4791 1-4622 1.4530 1.4620 1.4669 1.5733 1.5308 1-5732 1-5727 1.5486 1.5525 1.5820 1-5730 1.5752 1.5574 1-6061 1.5310 1.5987 1.5618 phur. rg _____ 1.5630 1.5345 1.5368 1.5214 1.4991 1.5705 1.4660 1-4641 1.4660 (1240 1.3662 1.7414 1.7417 1-1431 1-0603 1.0408 1.0755 1.1054 1.3238 1-1822 1.3057 1.3050 1.2307 1-2372 1.4225 1.3999 1.6705 1.2567 1-7217 1.1663 1.2299 1.0780 1.4481 1.3156 1,2818 1.1853 1-0672 1.1988 1.9951 0.9678 Rtlf. 12 23 23 23 19 19 22 13 13 19 23 23 23 11 11,22 20 20 20 20 23 23 23 23 23 23 23 11 23 23 11 13 Ref.6 7 6 7 12 6 6 6 6 6 19 22 6 6 6 6 6 6 6 6 - N E S ~ P A S O V et ul. TETRA- AXD TRI-CHLOROALKANES 367 No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 TABLE 6. Chloro-compounds containing the amino-group. Compound CCI,:CHCH,.NEt . . CCl,:CH*CH,*NEt,,HCL . CCl,:CCI.CH,-NEt . . CCl,:CCI-CH,.NEt,,HCl . CCl,*CHCI*CH,*NEt . . CCl,:CHCHMe.NEt . . CCl,:CH-CHMe.NE t 2 HC1 CCl,:CH*[CH,],*?SH . . CCl,:CH.[CH,j,*NHBz . CCl,:CH.[CH,],.NEt . . CCI,:(~H.CMe,.NH,,HCI . CCl,:CH.CMe,*N€I . . CCI 2:CH*CMe,-NC,H,o,HCl CCl,*[CH,],*NH . . . CC1,.[CH,],.NH2,HC1 . . CCl,*[CH,],*XHBz . . . CCIz:CC1~[CH,],*NEt . . CCI,:CH-[CH,],-NH . . CCI,:CH*[CH,],-KHBz . CHCI,*[CH,],.NH . . . CHCl,*[CH,],*NHBz . . CCI,*[CH,],.NH . . . CCl,*[CH,],.NHBz . . . CCl,:CH*[CH2],*NH .. CCl,:CH*[CK,],*NHBz . CCl,:CH*[CHJ,.NEt . CC1,fCH2],.NH . . . CCl,fCH,],-NHBz . . . CCl,:CH*[CH,],.NH . . CCI,:CH.[CH,],-NHBZ . CCl,:CH.CHPh.NEt . . M.p." 139-140 168-169 108.5- 109.5 167.5 55-56 180-151 218-249 229-230 95-96 36.5-37 56-57 TO*5-71.5 36.5-37.5 91-92 14-5-45.5 149-150 125-126 T3.p. O / m m . 65-66/7 8 1 -82/1 79*5-80/14 68-69/7 63-64/52 64-65/12 60-70/1.2 84-S5/2 83-81 93-94/7 75-79/1.2 102-103/7 S9-90/ 1.5 95-96/1.5 100-101/3 98-99/1 14708 1.4894 14690 1.4899 1.4719 1.4785 1.4862 1.4886 1.4865 1-4730 1.4543 1.4842 1.4730 1.4838 1-4822 1-5335 1.0693 1.1942 1.0470 1-1736 1.0349 1.1488 1.2619 1.1378 1-1331 1.1088 1.2192 1.1039 1-0097 1.1857 1.0599 1.1116 Ref. 7 7 14 14 14 22 22 5 5 19 22 22 22 5 5 5 19 5 5 5 5 5 5 5 5 19 5 5 5 5 22 22 19 QUiVITEREY REVIEWS TABLE '7..$fonocarbo,xy&c acids a?zd derivcrfiws. ~ 0. ~ 1 2 3 4 5 6 7 8 9 0 1 2 3 .4 5 .6 !7 IS 19 !O !1 !2 23 24 25 26 37 38 19 30 3 1 32 33 34 35 36 37 3s 39 40 41 42 43 44 45 46 47 4s 49 50 51 52 53 54 55 - Coinpound CHCl:CH-CO,H . . . 2Cl2:CH-CO,H . . . . (EtO),CH*CH,*CO,Et . . ( BuO ),CH*CH,*CO ,Bu . CCl,*CH,*CH,*CO*NHPh . CH,Cl*[CH,],-CO*NH2 . CH,Cl*[CH,],CO*NHPh . CHCl,*CH:CH*CO,H . . CHCl ,CH CH .C 0 *NH . CHCl,.CH:CH.CN . . . CHCl,.CH:C(OEt)*CN . . CHCI,CH:C(OBu).CN . CCl,:CH*CH,*CO,H . . CCl,:CH*CH,*COCl . . CCI,:CH*CH,*CO*NH . . CCI,:CH*CH,CO*NHPh . CCl,:CH.CH,CN CCl,:CHCH( OEt)*CO,H CCI,:CHCH(OEt)*CN . . CCl,:CH.CH( OBu)-CN . CCl,.CH:CH.CN . . . CCl,*CHCl*CH,*CO,H . . CCl,*CHCI-CH,COCl . . CCI,*CHC1.CH2.CO-NH . CCl,*CHCl.CH,*CN . . CCl,-CH( 0Me)-CH,*CN . CCl,Br*CHBr.CH,.CO,H .H0,C*[CH,],*NH2 . . . Me*[CH,],*CHCI-CO,H . Me.[CH,],CHCl.COCl . . MefCH,],.CHCl.CO*NHPh PhO-[CH,],-CO,H . . . PhS*[CH,],*CO,H . CH,Cl*[CH,],*CHCl.Cd,H ' CH2C1*[CH,],.CHClCOC1 . CH,Cl*[CH,],*CHCl. CO*NHPh CCl,:CH*[CH,],.CO,H . . CCl,:CH~[CH,],-CO~NHPh CCl,*[CH,],*CO,H . . . CCl,*[CH,],*CO*NHPh . CCl,.[CH,],-CO,* [CH,],.CCl CCl,:CH*[CH,],CO,H . . CCl,:CH.[CH,],*CO SNHPh CCl,:CH*[CH,],.CN . . CCl,*[CH,],.CO,H . . . CCl,*[CH,],*CO.NHPh . CC1,:CH.CHMe*CH2*CO,H CCl,*CHCl.[CH,],*CO,H . CH2:CH-[CH,l4.CO- NH-C,H,Me CH,:CH-[CH J,.CO,H . . CH,Cl*[CH,],*CO,H . . CH,CI.[CH,],*CO.PU'H . CH,Cl.[CH,],.CO.NHPI . EtO*[CH,],CO,Et . . . PhO*[CH,],*CO,H . . . 1I.p." S5-S6 76-77 159-160 7 8-7 9 10s-109 99-100 81-82 42-43 93-94 82-S3 108-109 1 3 8-1 38. 4 3-5-44 121-122 155-1 56 63-64 65-66 63-64 58-59 72-73 G5-66 117-1 1$ 63-64 50-51 109-11( 47-4s 60-61 82-83 85-86 56-57 B.p.";mm.95-96/12 112-1 14/ 1 lOl-l02/4 90-91/8 9 6-9 7 /4 67-68/13 86/10 Ill-ll2j7 77-78/11 114-115/2 84-8518 1 02- 103 15 6S-69/7 69-70/2.5 105-107/6.5 99-5 j6.5 93-94/5 61-62/28 129-131/5 s0/5 93-94/1 1 G4-165/1.5 1 3 9- 140/S 80-8113 114-1 15/1 102/1 118-130/14 122-123/11 77-78jP.5 1.4170 1*4280 1.4970 1.4797 1.4760 1.4831 1.4889 1 -4642 1-4635 1.5052 1-5130 14820 14442 1.4485 1.4835 1.4840 1.4898 1.5000 1.4895 1.4815 1.4800 1.44OC 1.4399 1.4295 pJ 0.9779 0.9239 1.3055 1.2074 1-1347 1.3122 1-3445 1.2160 1.1443 1.4237 1.6129 1.3879 1.1445 1.1765 1.3421 1.3513 1.3546 1-4060 1.2967 1.2018 1.2739 0.9500 1*0110 0*929@ ~~ Ref. 23 23 10 10 12 13 13 5 8 8 23 23 8 12 8 8 12 7 23 23 23 7 8 8 8 7 8 7 5 5 17 17 17 13 13 15 1' 17 17 12 13 1% 12 12 12 1 2 5 1% 12 22 18 13 13 13 13 13 13 13 TABLE 7.-continued.KO. ~ 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 8 10 11 12 13 14 15 16 17 18 19 Compound Me-[CH,],CHCl.CO,H . Me*[CH,],*CHCl.COCl . . CCl,:CH-[CH,],CO,H . CCI,:CH*[CH,],*CO-NHPh CH,Cl*[CH,],*CIICl-CO,H CH,Cl*[CH,],.CHCl*CO* NHPh CH2C1~[CH,],~CHC1*COC1 . CC1,.[CH21,*CO,H . . . CCl,*[CH,],*COCI . . . CCl,.[CH,],.CO*NH . . CCl,.[CH,],*CO*NHPh . CCl,:CH-[CH,],*CO,H . . CCl,:CH*[CH,],.CO.NHPh CCl,:CH-[CH,],.CN . . CCl,.[CH,],.CO,H . . . CCl,*[CH,],*CO*NHPh . CC1,-[CHJ,.CN . . . CH,CI*[CH,],CO,H . . CK,ClfCH,],*CO,Et . . CH,CI.[CH,],*COCI . . CH,C1*[CH,]7*C0.SH . CH,Cl*[CH,],*CO.NHPh . CC13*CHC1~[CH,],*COzH . NH,.[CH,],.CO,H .. . PhO*[CH,],*CO,H . . . CCI,:CH.[CH,],*CO,H . CCI,:CH*[CH,],*CO*NHPh Ph*CH:CCl.CO,H . . . p - Cl*C,K,*CH,*CH,-CO ,H p-Cl-C,H,*CH:CH*CO,H . p-Br*C,H,*CH,CH,.CO,H p-Me.C,H,*CH:CH.CO,H . Ph-[CH,],.CO,H . . a-CloH ,-CH:CH*CO,H . M.p.0 68-60 22-24 42-43 36-37 78-79 98-99 62-63 38-39 108-109 29-30 76-77 95-96 187-188 69-70 54-55 139-140 122*5-123 244-245 135 195-1 96 59-60 205-206 TABLE 8. Other carboxyl Compound H0,CfCH,],CHC1*CO2H . CCl,:CH.CH,.CH( C0,Et ) H 0 ,C*[ CH,] ,.CHCl.C 0 H . CCl,:CHCHMe-CH( C0,Et ) NC*CH,.[CH,],*CO,H . . HO2C~[CH,],CHC1~CO,H . CCl,:CH~[CH,]3*CH(C02Et) CCl,*[CH,],CH(CO,Etj . (CC12:CH-CH,),C(C0,Et) CCl,:CH-[CH,],*CH(C02Et)~ (M0,CfCH,],*CC1,)2 . . NC.[CH,],,.CN . . . . (HO,C*[CH,],.CCl:) . . (NC*[CH,],*CCl:)2 . . . NCfCH,],,.CN . . . . (H02CfCH,]6.CCl:) .. Et0,CfCH,],-CH(C02Et) (NCfCH,]6*CCI:) . . . EtO,C.[CH,I,.CM(CO,Et) M.p.0 104-105 88-89 98-99 39-40 223 108*5-. 109.5 48-49 05-96 B.p."jinm. 92 -93/1 ?6-77/13 120-181/1 128-130/1 1 o 4 p 120-121/0*5 91-92/1 12 8- 1291 1 99-1 00/ 1.5 139-140/ 1 123-1 25/2*5 136-137/8 100-1 0 1 /3 158-160/1 132-133/1 132-1 33/ 1.5 1.4485 1.4408 1-48i2 1-4804 1.4817 1.4859 1.4840 1.4787 1.4434 1.5018 1.4848 d20 1.0830 1.1006 1.9479 1.2441 1.2557 1.2120 1.1410 1.2097 0.9854 1-1806 acids and derivatives. B.p."/nim. 102-1 031 1.5 107/1 145-147/1.5 122-1 231 1.5 141-142.5/2*5 138-140/ 1.5 142-143/1 156-167/2 212-214/45 2 12-2 13/1 147-149/2 169-170/1*5 1.4633 1.4605 1.4650 1.4624 1.4663 1.4943 1.4895 1.4389 14419 1.2135 1.1829 1.1693 1.2130 1.1341 1.1295 1.0704 1.0568 1,0316 Ref. 17 17 12 12 17 17 17 5 12 5 5 12 12 12 12 12 12 12 13 13 13 13 13 18 13 13 12 12 23 20 23 20 11 13 11 Ref.17 17 22 13 17 12 5 7 12 16 16 16 16 16 16 16 13 13 n 370 QUARTERLY REVIEWS TABLE 9. Acetylenic compounds. ~ NO. ~ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Compound ClCiC.CH(OEt) . . . ClCjC*CH(OBu) . . . CHiC*CH(OBu) . . . CHiC.[CH,],.OEt . . . CH~C*[CH,],.OPh . . . CHiC-[CH,l,.NEt . . . CHiC*[CH,],-OEt . . . CHiC*[CH,],.NEt . . . CHiC*[CH,],*OEt . . . CHiC*[CH,],.SEt . . . CHjC*CHBun*OEt . . . [CH,] > C( OH)*CiCCHQ . [CH,] > C( OH)*CiCCH N-NH-C6H,( NO,) Ph*CH( OH)CiCfCH,],. CH,.OE t Ph*CH( OH)CiC.[CH,],. CH,.NEt Me-CiC*CH(OBu) . . . Prn.CiC.CH(OEt) . . . Prl*C{C.CH:N*NH. Bun*CiC.CH(OEt) . . Bun *CiC*CH:W-NH. Bui-CiC*CH(OEt) . . . Bui*CiC*CH:N*NH- n-C,H,,*CiC*CH( OEt) .n-C6H,,.C/CCH(OBu) . n-C,W,,CiC*CH(OEt) . Ph*CiC*CH(OEt) . . . Ph*C{C.CH(OBu) . . . c(-C,,H,.C~C.CH( OBu) . Pr'.CIC*CH(OEt) . . . . C6H3(x02)2 C,H,(~O,) C6H3(N02)2 M.p." 136-137 119-120 71-72 55-58-5 B.p."jmm. 55-56/10 1096-1 10*5/8 118-120/27 126-1 2 7 108-109/9 54/10 63-64/10 84-85/10 94-95/ 10 129-130/12 45-46/9 1 17.5- 118.5/1*5 185-1 S6/3 172-1 7 3 / 2 109.5-1 11/53 79-SO/S 73-75/8 99-100/10 91-92,'s 123-124/8 156-158/8 158-1 60/9 136-1 37/8 128-129/2 178-179/2 rl;o 1.4307 1-4450 1.4280 1.4204 1.5210 1.4412 1.4289 1.4460 14350 1.4771 1.4206 1.4741 1.5265 1.5327 1.4380 1.4338 1.4321 1.4328 1.4397 1.4395 1.4453 1.4419 1.5219 1.5058 1.5610 dfo 1.0300 0.9820 0.5752 0.8268 0.9845 0.8061 0.8318 0.8128 0.8336 0.8854 0.8229 1.0047 1.0284 1.006 0-8820 0.8898 0.8902 0.8771 0.8768 0.8760 0.8653 0.8669 0.9956 0-9540 1.0108 Ref.21 21 19 19 19 19 19 19 19 19 19 21 21 19 19 21 21 21 21 21 21 21 21 21 21 21 21 21 21

ISSN:0009-2681    

DOI:10.1039/QR9561000330

出版商:RSC    

年代:1956

数据来源: RSC