ORGANIC CHEMISTRY.I. INTRODUCTION.A DISCUSSION on the “ Mechanisms and Chemical Kinetics of Organic.Reactions in Liquid Systems” was held by the Faraday Society in Sep-tember, 1941, and the complete text of the various contributions is nowavailable.1 A wide field was covered, including aliphatic substitution,elimination reactions, esterifkation and hydrolysis, additions to olefiniccompounds, nuclear and side-chain substitution in aromatic compounds,condensations of carbonyl compounds, prototropic and anionotropic changes,the Cannizzaro and the Friedel-Crafts reaction, ring closure, and reactionsinvolving radicals. Many of these have been dealt with in the AnnualReports of 1 9 3 8 4 0 , and in Part 2 of the present section some recentinvestigations of condensations, alkylation reactions, rearrangements, andthe influences of groups in ortho-positions are summarised.The values ofdissociation constants in a single solvent and a t an arbitrarily fixed tem-perature have long been accepted as providing a correct series of relativestrengths of acids and bases, and hence of the polar effects of substituents;the justification for this conventional view is examined in this Report.Systematic organometallic chemistry has been extended by the dis-covery, in triethylscandium and triethylyttrium, of the first purely organicderivatives of the transitional elements ( L e . , those “ framed ” in Bohr’sPeriodic Table) showing the group valency. Gallium and indium com-pounds have been studied in more detail, the simple organic types R,Tland Me,Pt have been prepared, and organic derivatives of titanium, vanad-ium, tantalum, molybdenum, tungsten, manganese and rhenium are reported.The formulation of the curious phenylchromium compounds has been recon-sidered, and that of the dimeric trimethylaluminium presents a problem invalency theory.Numerous compounds of the heavy metals have beenprepared from R*B( OH),, R*SO,H, R*HgCI or especially R-N,CI, by treat-ment with the metallic halide; in this way it is possible to prepare organo-metallic compounds containing hydroxyl or other reactive substituents.Alkyl, and sometimes aryl, residues attached to mercury or elements of the tingroup undergo remarkably facile disproportionation in presence of catalysts ;e.g., Me,Hg + Et2HgZ2MeEtHg.All this work is dealt with in Part 3of this section.It is now possible to present a general picture of the chief structuralpatterns upon which the natural polysaccharides are built, and to pick outcertain definite types (see Part 4). The polysaccharides may be describedas “ simple ” or “ complex ” according to whether they are composed oft3he same or varied monosaccharide building units; thus, starch conld be1 Trans. Faraday Soc., 1941, 37, 601112 ORGANIU CHEMISTRY.described as “simple” in that it is composed solely of glucose, whereasarabic acid is “ complex,” being constituted of galactose, glucuronic acid,arabinose, and rhamnose. A second method of classification depends uponthe observation that in many well-known polysaccharides the glycosidiclinkages between monosaccharide units are mainly or entirely of one type.For example, only p-1 : 4-glycosidic links occur in cellulose, a-1 : 4-links inpectic acid, and 8-1 : 4-linkages in alginic acid.On the other hand, thearabans which are found in constant association with pectins are composedof arabofuranose residues combined by 1 : 5- as well as by 1 : 3-glycosidiclinks; yeast mannan is constituted solely of mannopyranose units, butthree types of linkage, 1 : 2, 1 : 3, and 1 : 6, are present; in srabic acid,1 : 3-, 1 : 6-, and 1 : 4-links are found. Damson gum, which is probablybuilt on the game structural pattern as arabic acid, is known to contain1 : 2-glycosidic linkages; this type of union is commonly associated withmannose residues, just as 1 : 3-links are often found associated with galactose(e.g., in agar) and 1 : 4-links with glucose.A third basis for the classi-fication of polysaccharides is illustrated as follows. Cellulose, starch, pecticacid, alginio acid, laminarin, and dextrans have unbranched chains ofmonosaccharide residues, but in many other polysaccharides the repeatingunits are more or Iess highly ramified structures, a feature which is shownparticularly by the plant gums (e.g., gum arabic, damson or cherry gum),and found also in the simple arabans, galactans, yeast mannan, in the poly-saccharide associated with /%amylase, and in the mucilages. Snail galactogen(containing d- and I-galactose) differs from agar in possessing a branched-chain repeating unit.The plant gums and the mucilages are thereforecomplex ” in two senses; they contain more than one type of mono-saccharide unit and these are united by more than one kind of glycosidiclinkage. In addition, these natural products are acid polysaccharides, theacid character being contributed by the uronic acid residues, glucuronicacid in the former and galacturonic acid in the latter. I n all the substancesmentioned above, the repeating units contain one residue with a free reduc-ing group and it is by virtue of this that aggregation takes place. A carbo-hydrate isolated from egg-albumin is of great interest in that, although itis constituted on a similar plan to the others, it appears not to be a truepolysaccharide; the smallest unit in its structure is composed of elevenmonosaccharide residues and does not display a free reducing group.Thiscompound is perhaps more correctly described as a non-reducing hendeca-saccharide; about 60% of it is composed of N-acetylglucosamine, theremaining residues being of mannose and galactose.The most recent development in the chemistry of starch has been therecognition of the probability that it contains a t least two componentswhich are structurally different. One of these, amylopectin (or erythro-amylose), forms 80-95~0 of the whole starch, and is composed of aggregationsof repeating units which are themselves constituted of chains of 24-30glucose residues linked by a-1 : 4-glycosidic bonds. It has now been demon-strated that the polymeric links between the repeating units are a-1 : 6-(INTBODUCTION.113glycosidic. The second component of starch (amylose or amyloamylose)appears to be composed of much longer chains (100-300 glucose members),which may not form aggregates; it is perhaps identical with Hanes’ssynthetic starch, the chain-length of which has been shown to be a t least80 to 90 glucose residues. Amylose and synthetic starch are each degradedcompletely to maltose by p-amylase, whereas with amylopectin thedegradation ceases a t 60% conversion.Part 5 of this section reviews recent studies of the synthesis of aliphaticand aromatic compounds containing a multiplicity of conjugated ethyleniclinkages (polyenes), which have led to the discovery of new syntheticroutes and also to the improvement and extension of the classical syntheticmethods.Investigations have also been directed towards the synthesisof the more complicated naturally occurring polyenes such as the caro-tenoids and vitamin A, and although no outstanding successes have as yetresulted, the preparation of material possessing some vitamin A activityhas been reported and a number of promising intermediates have beenobtained. Aldehyde condensations of fundamental importance in thesynthetic polyene field, particularly those of citral with aldehydes, havebeen investigated in some detail from the preparative aspect. Higherpolyene carboxylic acids have been made available, and from them directsyntheses of palmitic and stearic acids have been achieved.Two newroutes to the polyene dicarboxylic acids have also been developed, a biologicalpreparation presenting novel features.A considerable number of papers on polyterpenes have appeared duringthe period under review ; these are distributed fairly evenly between sesqui-terpenes and allied substances, diterpenes and triterpenes. In view oflimitations of space, Part 6 of this section deals only with abietic acid andthe p-amyrin and lupeol groups of triterpenes. The vexed question of thelocation of the ethylenic linkages in abietic acid has been finally settledby the elegant degradation experiments of Ruzicka and co-workers ; theacid has the structure (I). Considerable progress has been made in thep-amyrin group of triterpenes, although it cannot be claimed that the basicstructure of the group has been elucidated.It now comprises at least tenmembers, each of which has been transformed into p-amyrin or an estab-lished derivative. Betulin has been converted into lupeol, and the longsuspected difference between these triterpenes and the members of thep-amyrin group has been established by the proof that the former containan isopropenyl side chain. The presence of this unsaturated centremeans, according to Ruzicka and Rosenkranz, that the lupeol groupcannot be hydropicene derivatives, a conclusion which invokes the isoprenerule.Following the discussion of natural naphthaquinone pigments in theAnnual Reports for 1939, a review is now given (Part 7) of the presentstate of knowledge of the large group of natural colouring matters whichare derivatives of benzoquinone, naphthaquinone, phenanthraquinone, andanthraquinone.In the benzoquinone group, particular interest attache114 OR0 ANTC CHEMISTRY.to the simple quinones produced by certain fungi and to perezone, whichis closely related to the sesquiterpenes. Similarly, tanshinone I is note-worthy among the derivatives of phenanthrene, since the available evidencesuggests a relationship with the diterpenoids. Degradative evidence hasbeen adduced in favour of the view that the antihsmorrhagic vitamin K,is 3-difarnesyl-2-methyl- 1 : 4-naphthaquinone, and work on the naphtha-quinone pigments of sea urchin eggs has recently been extended. Echino-chrome A, the sols pigment isolated from fully mature ovaries of Arbacia,i q accompanied in ovaries collected a t different seasons by echinochromesB and C; echinochrome A appears to exist in the eggs as a complex ofhigh molecular weight, which is more potent than the free pigment inconferring motility on spermatozoa.Considerable attention has beendevoted in recent years to the production of anthraquinone derivatives bymoulds, and an account is given of pigments of this type isolated fromHelminthosporium, Penicillium, and Aspergillus species. Investigationson mould pigments from Penicilliopsis and on hypericin, the photodynamicpigment of St. John’s wort, have revealed striking resemblances betweenthem, and the interesting suggestion has been made that hypericin andoxypenicilliopsin may be helianthrone derivatives.The reactivity of heterocyclic nuclei towards cationoid and anionoidreagents, and of methyl side-chains towards aldehydes, has been dis-cussed in general terms, and much experimental work reported.Numerousnuclear transformations of furans, pyrroles and the azoles are recorded,most of which depend upon ring fission and re-synthesis to another hetero-cyclic type. An interesting new case of stereoisomerism depends upon thehindered rotation of the substituted benzene nucleus of a quinol poly-methylene ether (11) relatively to the (moderately) large ring. The simpleoxygen-ring compound dioxadiene has been prepared. Interest in corn -pounds related to vitamin E has led to a thorough study of methods forthe synthesis of chromans and coumarans.Some twenty quinoline homo-logues have been isolated from straight-run gasoline ; their structures havebeen established and show some curious regularities. Evidence is sub-mitted that a nucleus C,N,, consi&ing of three fused cyanuric rings, ispresent in melem, cyameluric acid and other long-known products of thepyrolysis of thiocyanates. The pterins, pigments occurring in the wingsof butterflies and in other insects, are now found to be relatively simplecompounds closely related to the purines ; leuco- and xantho-pterins havebeen synthesised from 2 : 4 : 5-triamino-6-hydroxypyrimidine with oxalicand dichloroacetic acids respectively. These advances in the chemistrWATSON : PHYSICO-ORGANIC TOPICS. 115of heterocyclic compounds are summarised in the final part (8) of thissection of the report.J.F. J. DIPPY.E. R. H. JONES.8. PEAT.F. S. SPRING.T. S. STEVENS.,4. R. TODD.H. B. WATSON.2. PHYSICO-ORGANIC TOPICS.(a) Mechanisms of Condensations and Alkylation Reactions.The mechanisms of condensations of the aldol, Knoevenagel, Perkin andClaisen types were discussed in the AnnuaZ Reports for 1939,l where referencewas made to the recent work of C. R. Hauser and his collaborators.2 W. G .Brown and K. Eberly have now investigated the base-catalysed exchangeof hydrogen for deuterium between a number of esters and deuteroalcohol,EtOD, and they find that the facility of exchange may be correlated with thereactivity of the ester in the Claisen condensation.For a series of estersR*CH,*CO,Et where R is varied, the order is Ph > H > Me > Et > Pra> C,,H,, > C,&, > cyclo-C6Hll > Prs; a second alkyl group as inCHMe,*CO,Et leads to a further decrease in reactivity. The correlationis not a quantitative one, however, and the authors suggest that the influenceof structure upon reactivity in the Claisen condensation is threefold, beingmanifested in the extent of the initial anion formation? the rate at whichthe anion reacts with the second ester molecule? and the extent to whichthe equilibrium is shifted by salt formation on the part of the p-keto-ester.A similar correlation between deuterium exchange and alkylation in malonicesters R*CH(CO,Et), is suggested.The investigations of C.R.'Hauser and co-workers have been extendedto include a study of reactions of the Michael type, i.e., the addition of acompound having incipiently-ionised hydrogen at an olefinic linkage whichis rendered susceptible to the attack of a nucleophilic reagent by an adjacentgroup :>c--$-x + >CH-X~ --+ >Y-hH-X(Component A) (Component B) - >c-X'In the most familiar examples, where X and X' are carbonyl or ester group-ings, this becomes :C. R. Hauser and B. Abraniovitch 4 represent this condensation (the revers-Soe also J . Amer. Chem. SOC., 1940, 82, 62, 593,Ibid., p. 1763.A ? m Reports, 1939, 36, 210.Ibid., p. 113116 ORGANIC CHEMISTRY.ibility of which was demonstrated twenty years ago 5, by a scheme whichis completely analogous to that usually accepted for the aldol, Knoevenageland Perkin reactions (B- = OEt’, CPh,’, or other basic catalyst) :>CH-V=O + B- >C-Y=O + HBThe factors which influence the Michael condensation, and the effectsof structure upon reactivity, have been summarised by R.Connor andW. R. McClellan.6 The groups which may function as X or X’ are CO,R,COR, CHO, CN, CO-NH,, NO,, S0,R (references to typical examples arequoted), and component A may be acetylenic rather than olefinic. Theoccurrence of other reactions (see below) is avoided by using as catalyst asecondary amine such as piperidine, but the change is slow, and long reflux-ing is necessary even in favourable cases; one-third t o one-sixth of anequivalent of sodium ethoxide may bring about condensation where aminesare ineffective (Hauser and Abramovitch used the still more powerful agentsodium triphenylmethyl), and long standing at room temperature now yieldsthe best results.As solvent, methyl and ethyl alcohols, benzene, ether anddioxan have been used satisfactorily. Substituents in either componentusually decrease rea~tivity,~ but there are exceptions and, as suggestedoriginally by C. K. Ingold, E. A. Perren, and J. F. Thorpe,8 both spatialand polar influences may be involved.Actually the products obtained under the conditions which lead to theMichael condensation are of three types : ( a ) in presence of a small quantityof sodium ethoxide or piperidine the product is normal; ( b ) one equivalentof ethoxide leads, in some instances, to a “ rearrangement product,” e.g.,CH,CH:CH*CO,Et CH,CH *CH (CH,) *CO,Et + + I* CH,*CH( CO,Et), CH( CO,Et),(c) there are sometimes “ rearrangement-retrogression products,” whichcould arise from the cleavage of a rearrangement product, e.g.,Ph*CH:CH*COPh + CH,*CH( CO,Et),PhCH CH(C0,Et)COPh+ I CH,*CH*CO,Et J.Ph*CH:C( CH,)*CO,Et + Ph*CO*CH,*CO,EtThe formation of “rearrangement products,’’ according to the views ofJ. F. Thorpe 10 and of A. Michael and J. Ross,11 depends upon the cleavageti C. K. Ingold and W. J. Powell, J., 1921,119, 1976.9 See R. Connor and D. B. Andrews, J . Amer. Chem. SOC., 1934, 56, 2713.lo J., 1900, 77, 923.J . Org. Chem., 1939, 3, 570. 7 See also ref.9. J., 1922, 121, 1771.11 J . Amer. Chem. SOC., 1930, 52, 4608WATSON : PHYSICO-ORGANIC TOPICS. 117of the substituted malonic ester into portions such as CH3 and CH(CO,Et),,but N. E. Holden and A. Lapworth la postulate a normal addition, followedby migration of carbethoxyl (or similar group), a view which has beensupported more recently by J. A. Gardner and H. N. Ryd0n.1~Although the most familiar condensations of the aldol, Knoevenagel,Perkin, Claisen and Michael types occur under the catalytic influence ofbases (electron-donators) , examples of catalysis by acids (electron-acceptors)are known. D. S. Breslow and C. R. Hauser l4 have now carried out anumber of condensations in presence of the electron-accepting agents boronfluoride and aluminium chloride.These include the condensation of aceticanhydride with acetophenone (Claisen type) and the condensation of benz-aldehyde with malonic ester and with acetic anhydride (only a small yieldof cinnamic acid was isolated, however). The reaction of benzaldehyde withethyl malonate was followed by a Michael addition, giving ethyl benzylidene-dimalonate. I n presence of boron fluoride, ethyl acetoacetate is alkylatedboth by benzyl chloride l4 and by isopropyl acetate l5 [givingand CH,*CO-CH( CHMe,)*CO,Et respectively].I n a general discussion of reactions of the types referred to above, C. R.Hauser and D. S. Breslow l6 point out that component B is always a com-pound having incipiently-ionised hydrogen, and component A may be analkyl halide (alkylation of ethyl acetoacetate, etc.), an aldehyde or ketone(aldol, Knoevenagel, Perkin), an ester, anhydride or acid chloride (Claisen),or a suitably activated olefinic compound (Michael).In accordance withthe usually accepted view,l they suppose that bases activate component Bby converting it into anion, whereas acidic (electron-accepting) catalystsactivate component A by forming a co-ordination complex. For a givencomponent A, the eage of condensation should follow the activity of thelabile hydrogen of component B, and they quote examples to 8how thatsuch is the case. Further, compounds which are not sufficiently reactiveto function as component A in presence of a base might do so under theinfluence of an acid catalyst, and it is found that both diisopropyl etherand isopropyl alcohol condense with acetoacetic ester in presence of boronfluoride.Hauser and Breslow envisage the possibility of an attack of thecatalyst upon both reacting substances ; for example, an acidic catalystmight, in addition to activating component A, also co-ordinate with carbonyloxygen in component By and a basic catalyst might form an addition com-plex with component A. The last possibility was considered in the AnnualReporb for 1939.l I f the base co-ordinated with component A to give acomplex possessing sufficient energy to react with component B, the energyof activation would be needed for the formation of this complex, and wouldbe influenced by constitutional changes in A, but almost indifferent tochanges in B.This is actually the case in the reaction of benzaldehydeCompare EL Meerwein, Ber., 1933, 66, 411.CH,*CO*CH( CH,Ph)CO,Etl2 J., 1931, 2368. 13 J., 1938, 48.14 J . Amer. Chem. Soc., 1940, 62, 2385.16 J . Amer. Chern. Soc., 1940, 62, 2611. l6 Ibid., p. 2389I18 ORGANTC CHEMISTRY.with acetophenone in 90% alcohol with sodium ethoxide as catalyst.17There are, however, objections to this mechanism of basic catalysis,lB andfurther experimental studies are necessary.The rate of the reaction of benzaldehyde with acetophenone is propor-tional to the concentrations of both; l7 on the other hand, the aldol con-densation of acetaldehyde is of the first order with respect to the aldehyde,and the corresponding reactions of acetone and isobutyraldehyde are of thesecond and an intermediate order respectively .l9The influence of aluminium chloride and boron fluoride upon thesecondensations, referred to on p. 117, finds an analogy in the catalytic effectof boron fluoride in the esterification of acetic, propionic and a number ofaromatic acids by various alcohols, observed by J. A. Nieuwland and co-workers,20 who suggest that the mechanism is doubtless similar to that ofthe hydrion-catalysed reaction ; addition compounds of boron fluoride withacids, alcohols, and esters are known to exist.21 Esters are also formedwhen alcohols react with amides in presence of boron fluoride,22 and by thereaction of acids with olefins under the influence of the same catalyst ; 23boron fluoride also catalyses the acidolysis of esters.=A consideration of some aspects of the Friedel-Crafts reaction forms anatural sequel to the above, although boron fluoride is apparently noteffective in nuclear alkylation by alkyl halides,25 nor has it been employedin acylations by the Friedel-Crafts method.In the AnnuaE Reports for1937,26 reference was made to the use of substances other than alkyl halides(alcohols, ethers, esters, olefins) for alkylation in presence of aluminiumchloride, and similar processes have been oarried out with boron fluorideas the catalyst. anisole 29 and all threehydroxybenzoic acids have been converted into nuclear isopropyl deriv-atives by propylene in presence of boron fluoride (in $he case of phenol andthe hydroxybenzoic acids, etherification or esterification may be followed1 7 (Miss) E.Coombs and D. P. Evans, J., 1940, 1295.18 See discussion on the Mechanism and Chemical Kinetics of Organic Reactions19 R. P. Bell, ibid., p. 716.20 H. D. Hinton and J. A. Nieuwland, J . Arner. Chem. SOC., 1932, 54, 2017; F. J.21 H. Bowlus and J. A. Nieuwland, ibid., 1931, 63, 3835.21 F. J. Sowa and J. A. Nieuwland, ibid., 1933, 55, 5052.23 T. B. Dorris, F. J. Sowa, and J. A. Nieuwland, ibid., 1934, 56, 2689; T. B. Dorris24 F. J. Sowa, ibid., p. 654.25 A. Wohl and E. Wertyporoch, Ber., 1931, 64, 1357.*6 P. 260. Compare N. 0. Calloway, Chena. Reviews, 1935, 17, 327; (Miss) D. V.27 S. J. Slanina, F. J. Sowa, and J. A . Nieuwland, J . Amer. Chern. SOC., 1935, 57,ee F.J. Sowa, H. D. Hinton, and J. A. Wieuwland, ibid., 1932, 54, 3694.38 W. J. Croxall, F. J . Sowa, and J. A. Nieuwland, ibid., 1934, 56, 2054; 1935,For instance, benzene,27in Liquid Systems, Trans. Faraday SOC., 1941, 37, 718.Sowa and J. A. Nieuwland, ibid., 1936, 58, 271.and F. J. Sowa, ibid., 1938, 60, 358.Nightingale, ibid., 1939, 25, 329.1547,Eidem, ibid., 1933,56, 3402.67, 1549WATSON : PHYSICO-ORGANIC TOPICS. 119by rearrangement), and under the influence of the same catalyst, benzene 31and naphthalene 32 have been alkylated by alcohols ; esters 33 and etherscan also alkylate. The Claisen rearrrangement of phenolic ethers, whichis catalysed by boron fluoride, aluminium chloride and other agents, isprobably an intermolecular process involving nuclear alkylati0n,~5 and theFries rearrangement similarly involves acylation.Other reactions whichare catalysed by boron fluoride include the addition of alcohols to acetyleneto give aceta1s,21 and the preparation of phenolphthalein and flu~rescein.~~The view that alcohols, esters and ethers are first converted into olefin,which is the active agent in alkylati~ns,~l* 33* has been shown to be un-tenable 32 on the following grounds : ( a ) although cyclohexanol alkylatesnaphthalene in presence of boron fluoride, the alcohol can be recovered un-changed after treatment with the catalyst under conditions more drasticthan those required for alkylation, and ( b ) olefin formation is not possiblefrom benzyl alcohol, which nevertheless alkylates.,’ Price and Ciskowskitherefore suggest that a carbonium ion is the active intermediate, e.g.,RT + R’O-BF,/\Y R’+ + RO-BF,g > O + BF, 7 $>O-+BF,where R = alkyl and R’ = alkyl or acyl.A similar mechanism is suggestedfor alkylation by olefiiis and for the catalysed polymerisation of olefins, thefirst step being>C=C< + BF, ;+ >C-O-BF,C. C. Price and M. Meister 38 consider that the catalysis of geometrical invcr-sion depends upon the co-ordination of the catalyst (e.g., BF,) a t one of theunsaturated carbon atoms, the freedom of rotation consequent upon thetransformation of the olefinic linkage to a single bond making the inversionpossible, e.g.,4$HPh H--M--PhPh-C-H + BF3 CHPh BFa -I- Ph-9-HPh-C-H-BF,It may be noted, however, that the occurrence of the hydrion-catalysedchange of maleic into fumaric acid in this way is unlikely, since the presenceof deuterium chloride does not lead to a product containing deuteriumY3931 J.F. McKenna and F. J. Sowa, J . Amer. Chem Soc., 1937, 59,470; N. F. Toussaintand G. F. Hennion, ibid., 1940, 62, 1145.a* C. C. Price and J. M. Ciskowski, ibid., 1938, 60, 2499.33 J. F. McKenna and F. J. Sowa, &id., 1937, 59, 1204.34 M. J. O’Connor and F. J. Sowa, ibid., 1938, 60, 123; A. J. Iiolka and H. H . Vogt,s5 See Ann. Reports, 1939, 36, 208.36 J. F. McKenna and F. J. Sowa, e J . Amer. Cheti?. Soc., 1038, 60, 124.3 7 Compare E. Bowden, ibid., p. 645.3g C. Horrex, Tram. Faraday SOC., 1937, 33, 570.ibid., 1939, 61, 1463.38 Ibid., 1939, 61, 1595120 ORGANIC CHEMISTRY.and K.Nozaki and R. Ogg 40 have put forward a mechanism in which theacid catalyst here co-ordinates at oxygen of carboxyl, and anions, whichalso appear to play a part, may become attached a t one of the olefiniccarbon atoms; inversion under the influence of amines is representedsimilarly .41The catalytic effects of boron fluoride and aluminium chloride must,of course, be dependent upon the ability of these molecules to accept anelectron pair, and the resulting production of alkyl ions is in harmony withthe strongly acidic qualities of BF3-alcohol complexes.21 The existence ofan ionised complex at an intermediate stage in the familiar Friedel-Craftsreactions involving akyl halides has frequently been postulated, and isrendered more probable by the work of E.Wertyporoch and of F. Fair-brother.42 Fairbrother has shown recently that, for a number of pairs ofinorganic and organic bromides, the ease of exchange of radioactive bromineis closely parallel to the reactivity in the Friedel-Crafts synthesis, and theformation of highly polar complexes is indicated by dielectric-constantmeasurement^.^^ The " activation " of the alkyl halide, alcohol, ether,ester or olefin by the catalyst must consist fundamentally in its conversioninto a complex which is a strongly electrophilic reagent, for ( a ) the orientationof nuclear alkylation is, with certain exceptions, the same as in nuclearhalogenation, nitration and sulphonation, and ( b ) alkylation of a nucleusalready containing a powerfully activating group such as hydroxyl mayoccur in presence of a less powerful catalyst, and a reagent such as benzylchloride which is already strongly electrophilic owing to the facility ofionisation, Ph*CH,-CI, can alkylate in absence of a catalyst. The co-ordination compounds of boron fluoride with alcohols, ethers and esters maybe represented as follows :0*>O-BF3 Rwhich means probably that ionised structures in which R+ is dissociatedfrom the remainder of the molecule participate in the mesomeric state,This accounts for the observation that a normal alkyl group of more thantwo carbon atoms usually isomerises (e.g., CH3*CH,GH, will exist as++CH,-CH*CH,).I n the case of an olefin, the alkylating agent is no doubt'.,TJ,,'I.I.that formulated by Price and Ciskowski, and it is easy to understand why,for example, propylene introduces an isopropyl group via the complexCH36H-CH2-BF,. The co-ordination complex of aluminium chloride40 J . Amer. Chern. SOC., 1941, 83, 2583.See Ann. Reports, 1937, 34, 252. V. N. Ipatieff, H. Pines, and L. Schmerling( J . Org. Chem., 1940, 5, 253) include similar complexes in their alkylation mechanisms;they also discuss the isomerisation of groups.4f Ibid., p. 2681.43 J . , 1941, 293; Trans. Faraday Soc., 1941, 37, 763WATSON : PHYSICO-OWANIC TOPICS. 121with an alkyl halide will be R-X+AlCl,, which will give rise to ionisedstructures such as [k-X-MC1, C1-1, [k xAiC13], [A XAlC1, C1-1, and+ +[R AICI, X-] ; the participation of [R AlC1, X-] accounts for the equivalenceof the four halogen atoms observed by Fairbrother.42 Since boron fluorideapparently does not catalyse alkylation by alkyl halides, the structures inwhich the alkyl group is ionised may not be of much importance in the+n complex R-X+BF3, i.e., the process R-X-GF, does not occur sufficientlyto render the complex strongly electrophilic.The entry of a second alkylgroup into the m-position has been ascribed to the reversibility of theFriedel-Crafts reaction, the 1 : 2 : 4-trialkylated compound being formedand the l-alkyl group then expelled.The view that the alkylating agent is a carbonium ion (or a mesomericform in which ionised structures participate to an important extent) appearsa t first sight to be a t variance with Hickinbottom's view that, in the Claisenrearrangement, the group migrates as a neutral radical.44 This view is based,however, upon observations of the rearrangement brought about by heatalone, and the mechanism of the catalysed change is not necessarily the same.(b) Rearrangements.(Continued from Ann.Reports, 1939, 36, 191.)Continuing previous work in which optically active hydratropamide,CHPhMe*CO*NH,, was shown to be converted into a-phenylethylaminewith a 95.8% retention of optical activity,45 J. Kenyon and D. P. Young 46find that the Curtius degradation of hydratropic azide gives an &mine of99.3% optics1 purity. The intramolecular character of the Curtius changeis thus confirmed.The authors suggest that the loss of optical activityobserved in the Hofmann rearrangement of the amide, which, thoughsmall, is reaI, may be attributed to some racemisation of the intermediateisocyanate, a view which is based upon an earlier observation by E. S.Wallis and R. D. Dripps,*' who found that an optically active isocyanateis racemised on alkaline hydrolysis (as in the Hofmann change) but not onacid hydrolysis (as in the Curtius reaction). The intramolecular nature ofthe Beckmann transformation is demonstrated by the retention of opticalactivity in the conversion of methyl y-heptyl ketoxime into aceto-y-heptyl-amide. Kenyon and Young point out that an optically active radicalwhich is transferred in a Hofmann, Curtius or Beckmann change retainsnot only its asymmetry but also its configuration, Le., no Walden inversionoccurs.This was formerly assumed,** and P.D. Bartlett and L. H. Knox 49showed that the Hofmann rearrangement could occur in a case whereinversion was not possible. Direct evidence of the absence of optical44 See Ann. Reporta, 1939, 36, 209.46 J., 1941, 263.Ber., 1933, 66, 684.4 5 Ibid., p. 193.d 7 J . Amer. Chem. Xoc., 1933, 55, 1701.E. S. Wallis and S. C. Nagel, ibid., 1931, 53, 2787 ; J. von Braun and E. Friehmelt,4s J . Amsr. Chem. SOC., 1939, 61, 3184I22 ORGANIC! CHEMISTRY.inversion in the Hofmann rearrangement is provided by some earlier resultsof W. A. Noyes and co-workers,a who converted the half' smide of camphoricacid into the corresponding amino-acid without inversion, and the proofof the identity in configuration between benzylmethylacetic acid anda- benzylethylamine hydrochloride of the same sign of rotation togetherwith the conversion of the former into the latter via the amide or the azide *52gives another demonstration.The stereochemical method has also been used in J.I?. Lane and E. S.Wallis's study of the Wolff rearrangement of diazo-ketones, which occurswhen these compounds are treated with ammoniacal silver nitrate and certainother reagents. The mechanismis as follows :R*CO*CHN,R CHsuggested some time ago by B. Eistert 54--+ R*CO*CH + N,R---CHR*CH:CO + AH --+ RCH,*CO*A(where AH = NH,, H,O, ReNH, or R-OH)A case of this rearrangement without accompanying racemisation had beenreported at an earlier date,55 and in order to obtain further evidence of itsintramolecular character Lane and Wallis rearranged the diazo-ketoneCMeBuPh*CO*CHN2 in boiling aniline and also in aqueous dioxan containingsilver oxide and sodium thiosulphate; in neither case did the opticallyactive diazo-ketone give a racemic product.The compound (I) was sub-mitted to the same treatment, this substance being chosen because F. Bellhad previously shown that when optically active specimens of the acid(11) are degraded by either the Hofmann or the Curtius reaction the resultingamine is active ; 56 again no racemisation was observed. It is concluded thathere, as in the Hofmann and the Curtius change, the group is never released,and it would appear that the Wolff rearrangement is quite similar to thesedegradations, being represented most probably as in (III)?'(>,,.,,NO, c /ca,C0 II(TII.)50 See S.Archer, J . Amer. Chem. SOC., 1940, 82, 1872.51 J. Kenyon, H. Phillips, and (Miss) V. P. Pittman, J., 1935, 1072.58 E. S. Wallis et at. See Ann. Reports, 1939, 36, 193.63 J . Org. Chem., 1941, 8, 443. 64 Ber., 1935, 68, 208.5 5 N. A. Preobrashenski, A. M. Poljakova, and V. A. Preobrashenski, ibid., p. 860.5 6 J., 1934, 836. 67 Compare Ann. Reports, 1939, 36, 194WATSON : PHYSICO-ORGANIC TOPICS. 123Optically active benzylmethyldiazoacetone, CH,Ph*CHMe*CO*CHN,, how-ever, rearranges to give a partially or completely mcemised product 58 andit is suggested that this may be due to the presence of an enolisable hydrogen.The catalytic effect of phenols on the Wagner-Meerwein rearrangementhas been discussed by P.D. Bartlett and J. D. Gill.59 They find that theefficiencies of four phenols as catalysts of the change of camphene hydro-chloride into isobornyl chloride stand in the order of their hydrogen-bondingpowers, and conclude that the phenol solvates the chloride ion ; the variationof the reaction rate with the concentration of the phenol indicates that theattack takes place in two ways, involving one and two molecules of the phenol.It is interesting that F. C. Whitmore, A. H. Popkin, H. I. Bernstein,and J. P. Wilkins 6O were unable to isolate any tert.-amyl derivatives fromthe products of the action of metallic sodium upon neopentyl chloride;the Wurtz reaction proceeds via free radica1sy61 and it appears that underconditions giving radicals the neopentyl group does not undergo the changeto tert.-amyl which occurs in processes where it appears as a positive ion.A few studies of certain migrations from the side chain to the nucleusof aromatic compounds have been recorded since the Annual Report for1939.P. J. Dmmm, W. F. O'Connor, and J. Reilly e2 have examined theproducts of the Hofmann-Martius rearrangement of dibenzylaniline hydro-chloride, and fmd p-aminodiphenylmethane, 1 -amino-2 : 4-dibenzylbenzeneand also an aminotribenzylbenzene, probably the 2 : 4 : 6-compound; thisappears to be the first example of the introduction of more than two groupsby this reaction.The electronic mechanism of the benzidine transformationhas been discussed by Sir R. Robinson in his Presidential Address to theChemical Society,63 and E. D. Hughes and C. K. Ingold 6* have publishedsonie comments on this reaction. A summary of our knowledge of the Friesreaction, with a discussion of proposed mechanisms, has appeared.65 A. W.Ralston, M. R. McCorkle, and S. T. Bauer 66 find that variations in thequantity of aluminium chloride and in the solvent influence the o/p ratioin the Fries rearrangement and in the Friedel-Crafts reaction similarly,indicating an analogy between the two processes.D. S. Tarbell and J. F. Kincaid 67 have shown that the Claisen rearrange-ment of 2 : 6-dimethylphenyl allyl ether to 2 : 6-dimethyl-4-allylphenol iskinetically unimolecular.There is further evidence that in the p-rearrange-nient the a-carbon, and not the y-carbon as in the o-rearrangement of arylallyl ethers, becomes linked to the nucleus,6* and stereochemical consider-ations indicate that the former change must be intermolecular. Neverthe-5 8 J. P. Lane, J. Willenz, A. Weissberger, and E. S . Wallis, J . Org. Chenb., 1940,5, 276.5g ? J . drner. Chem. SOC., 1941, 63, 1273.G 1 See Ann. Reports, 1940, 37, 286,F,s J . , 1941, 220.6 5 A. H. Blatt, Chem. Heviewu, 1940, 27, 413.6 8 . I . Org. Chem., 1940, 5, 645.6 B 0. Mumm and J. Diederichsen, Ber., 1939, 73, 1523; E. Spath and I?. Kuffrier,6O Ibid., p. 124.62 J . Amr. Chent. SOC., 1940, 62, 1241.lbid., p.608.6 7 J . Amw. C'hem. SOC., 1940, 62, 7'38.ibid., p. 1580124 ORUANIC OHEMISTRY.less, W. I. Gilbert and E. S. Wallis 69 detected no migration of a group toa, foreign nucleus when phenyl isopropyl ether and p-tolyl sec.-butyl etherwere rearranged in the same solution in presence of sulphuric acid, and theyconclude that the mobile group is at all times within the sphere of influenceof the molecule. There are instances, however, where this group has beenfound linked to another molecule.70 A comprehensive account of the factsrelating to the Claisen rearrangement has been published by D. S. Tarbell.71Compounds such as vinyl allyl ether, which contain the essential partof the aryl allyl ether skeleton, undergo a change exactly similar to theClaisen rearrangement, and an analogous change in certain three-carbonsystems has now been discovered by A.C. Cope and c o - ~ o r k e r s . ~ ~ Onheating to temperatures between 135" and 200" the following occurs (X andY = CN or C0,Et) : >c=b*; _3 >C-b=qy Xb 5 C3HSIThe rearrangement becomes less easy in the order malonitriles > cyano-acetic esters > malonic esters, and is kinetically unimolecular. In the re-arrangement of the related crotyl compounds, it has been shown that themigrating group becomes linked through the y-carbon atom; e.g.,CHR:CMe><; 9 HR-CMe , y XCHMe:CH*CH, CHMe-CH:CH, Yand if two esters containing severally allyl and crotyl groups are rearrangedtogether, no interchange of migrating groups occurs.There can thereforebe little doubt that the change is intramolecular.(c) The ortho-Effect ; Xteric Inhibition of Mesomerisrn.(Continued from Annual Reporb, 1939, 36, 215.)A reduction of the mesomeric effects of the nitro- and dimethylamino-groups by two methyl groups standing in o-positions with respect to them,as in dimethylmesidine, nitrodurene and nitrodimethylaminoduene, waspostulated by Hampson and co-workers 73 in their discussion of the dipolemoments of these compounds, which are considerably lower than those ofthe corresponding compounds in which the methyl groups are absent; itwas considered that the o-methyl groups would make it dif5cult for the NO,or NMe, to come into the plane of the nucleus, thus producing conditionsunfavourable to mesomerism.The study of cases of this kind has now beenextended, particularly by workers at the University of Chicago. R. G.Kadesch and S. W. Weller 74 find that the dipole moments of acetylmesitylene( 2 * 7 1 ~ . ) and acetyldurene (2-68) are almost identical with those of aliphatic' 0 J . Org. Chem., 1940, 5, 184.71 Chew&. Revkw8, 1940, 27, 496.74 J . Aw. Chm. SOC., 1940, 62, 441; 1941, 63, 1843, 1862. '' J., 1937, 10; 1939, 981.7O See Ann. Reports, 1939, 30, 207.14 J . Amer. Chem. SOC., 1041, 88, 1310WATSON : PHYSIOO-ORQBNI13 TOPIOS. 126ketones, and suppose that the o-methyl groups inhibit the meaomeric effectof the carbonyl group which operates in acetophenone (p = 2.88); ob-servations of the absorption spectra of acetylmesitylene and 2 : 4 : 6-tri-isopropylacetophenone point to the same concl~sion.~~ On the other hand,o-methyl groups produce no diminution of the moment of benzaldehyde(2.92; p for mesitylaldehyde = 2-96>, and models show that, whereas thesegroups interfere very considerably with methyl of the CO-CH, in acetophen-one, any such interference with the hydrogen of the aldehyde is almostnegligible.The moment of 2 : 4 : 6-trimethylbenzoyl chloride (2-95), again,is appreciably less than that of the unsubstituted acid chloride (3-32).Hampson's conception of a steric inhibition of mesomerism has stimulatedinterest in the chemical aspects of the " ortho-effect," and was adopted byG. Baddeley 76 as the basis of the interpretation of a number of the peculi-arities associated with compounds in which two groups stand in o-positionswith respect to each other.An inhibition of mesomerism is also postulatedby G. W. Wheland and A. A. Danish 77 in order to explain the reduction ofthe acidic character of 4 : 4' ; 4"-trinitrotriphenylmethane by methyl groupsin the six positions ortho to the nitro-groups. R. T. Arnold, G. Peirce, andR. A. Barnes 7t3 find, too, that 4-nitrodimethyl-a-naphthylamine is a muchstronger base than 4-nitro-a-naphthylamine, and they suppose that theinterference of the second ring with the bulky dimethylamino-group hererenders it difiicult or impossible for this group to become coplanar with thenucleus, thus reducing the mesomerism with a consequent increase in basicstrength ; confirmation is found in the lower melting points of the N-dialkylcompounds as compared with the primary amine (R-NH,, 191". R-NMe,, 65".R*NEt,, liquid), which indicate a less polar character.On the other hand,R. D. Kleene, F. H. Westheimer, and G. W. Wheland 79 have observed thatthe relative strengths of substituted cis- and trans-cinnamic acids cannotbe accounted for on the basis of the steric inhibition of mesomerism; forinstance, the dissociation constant (in 40% acetone) of trans-3 : 4 : 6-tri-methylcinnamic acid is greater by a factor of 2.5 than that of the cis-isomeride,although the former is probably planar whereas the latter is certainly not.From measurements of the reactions of piperidine with some nitro- andcyano-bromobenzenes, W.C . Spitzer and G. W. Wheland 80 conclude thatmesomerism can be inhibited in suitably substituted aromatic nitro-com-pounds but not in the corresponding cyano-derivatives, although in theformer case the effect is smaller than would be expected from the dipole-moment measurements of R. H. Birtles and G. C. Hamps0n.7~ F. H.Westheimer and R. P. Metcalf's study of the alkaline hydrolysis of a numberof substituted ethyl benzoates having NO,, NH, or NMe, in the p-position 81has shown that the effects of these groups (of which NO, accelerates and theothers retard the reaction) are rendered smaller by methyl groups at the75 M. T. O'Shaughnessy and W. H. Rodebush, J . Amer. Chm. SOC., 1940, 62, 2906.7 6 Nature, 1939, 144, 444.7 8 Ibid., p. 1627.80 Ibid., 1940, 62, 2996.77 J.Amer. Chern. SOC., 1940, 62, 1125.79 Ibid., 1941, 63, 791.Ibid., 1941, 63, 1339126 ORCI ANICl CHEMTSTRY.3 : 5-positions; this is particularly striking in the case of ethyl 4-dimethyl-a mino-3 : 5-dimethylbenzoate, as illnstrated by the velocity coefficientsbelow :ethyl 4-dimethylarninobenzoate .................. 0.00152 1 = Ratio 30 4ethyl benzoate 0.052ethyl 3 : 5-dimethylbenzoate ..................... 0.373 Ratio""{ethyl 4-dimethylamino-3 : 5-dimethylbenzoate 0406 1 = 1-8A reduction in the reactivity of the p-position owing to the presence ofa group placed ortho to NMe, was observed by von Braun in 1916,82 and inan investigation of the deuteration of derivatives of dimethylaniline, W. G.Brown, A. H.Widiger, and N. J. Letang 83 find a similar effect in o-bromo-and o-chloro-dimethylaniline, but very little in the o-fluoro-compound.When the nitrogen is linked to the ortho-carbon atom to form a five-, six-or seven-membered ring system, as in N-methylindoline (I), N-methyl-tetrahydroquinoline (11), and IY-methylhomotetrahydroquinoline (111),however, the reactivity is dependent upon the size of the second ring; in(I) and (11), where it is coplanar with the benzene ring (or almost so), thereactivity is high, exceeding that in dimethylaniline itself, but the puckeredseven-membered ring in (111) leads to a relatively slow reaction........................................ a t 30°{(\AVH3 i'lF2-g\/\ PH, N I \,A w P H 2 KCH, CH,(111.)CH3(1.) (11.)The authors consider that " these results provide rather convincing evidencethat the key to the situation really lies in the ability of the dialkylamino-group to come into the plane of the benzene ring " ; ie., the mesomeric (and,in a reaction, the electromeric) effect of the group is inhibited when it isnot able to do so.I n a, recent paper,s* Brown and Letang have demonstrateda reduction of reactivity (in the deuteration reaction) in dimethyl-a-naphthyl-amine by chlorine or a nitro-group in the 8-position, and also a mutualhindrance of dimethylamino-groups in the peri-positions. Extension ofthis work to carbazole derivatives (V), however, has given results which areless easy to interpret, for these compounds are less reactive than the cor-responding diphenylamine derivatives (IV) in spite of their favourable planarstructure, Dihydroacridine derivatives (VI) also show a depression of82 See Ann, Reports, 1939, 36, 218.84 Ibid., 1941, 63, 358.83 J .Artier. Chan. L!OC., 1839, 61, 2597DIPPY : PHYSICO-ORGANIC TOPICS. 127reactivity as coinpared with (IV), but this may be due to a folded structuresimilar to that of dihydroanthracene and NN'-dimethyldihydrophenazine. 85These new results appear to confirm the suggestion made in the AnnualReports for 1939 that the various manifestations of the " ortho-effect ''cannot be interpreted on the basis of a single conception; steric inhibitionof mesomerism, interaction between groups in o-positions, and perhaps alsogeometrical steric hindrance as envisaged by Victor Meyer wiIl probablyall find their place in a final and comprehensive interpretation of the observedphenoniena when such an interpretation is achieved.H.B. W.(d) The Strengths of Organic Acids and Bases.The classical dissociation constants of Wilhelm Ostwald provided theearliest numerical data which reflected the changes in reactivity attendantupon systematic variations in chemical structure, and the rise of the theoryof interionic attraction has led more recently to the computation of thermo-dynamic dissociation constants ( K ) which furnish a better measure of thestrengths of organic acids than do the older values. The later data aresuperior not only because the corrections made necessary by modern theoryare taken into account in their derivation, but also by reason of improve-ments in experimental technique. Wide use has been made of the dis-sociation constants for aqueous solution in discussion of the polar influencesof substituent groups, and several investigators have been able to relatethem quantitatively with reactivities and other characteristics of organicmolecules.1, 2 A collection of the more reliable dissociation constants ofmonobasic organic acids and the strengths of organic bases (for one tem-perature) has been published re~ently,~ and to this may now be addedfurther values for formic, n-butyric and cyanoacetic acids.4Most of the available data relate to aqueous solution a t 25", and quiterecently doubts have been expressed as to the correctness of basing com-parisons of acid strengths upon the values of dissociation constants for anarbitrarily fixed temperature and a single selected solvent.The factors governing acid dissociation will be affected by elevation oftemperature, both the solvent and the electrolyte being directly concerned.The dielectric constant of a liquid diminishes with increasing temperature,and so with the decreasing electric field there will be a smaller tendencyfor ions to separate. Again, there is less solvation and also loss of com-plexity in both solute and solvent.There exist comparatively few data relating to the variation of dis-sociation constants of organic acids with temperature, and those of a reliable85 (Miss) I.G. M. Campbell, (Mrs.) C.G . Le Fevre, R. J. W. Le FBvre, and l4. E.See Ann. Reports, 1937, 34, 52; 1938, 35, 239; 1939, 38, 216.H. 0. Jenkins, J., 1940, 1447.J. F. J. Dippy, Chem. Reviews, 1939, 25, 151.B. Sexton and L. S. Darken, J . Amer. Chem. SOC., 1940, 82, 846.Turner, J., 1938, 404128 ORGANIC OEEMISTRY.character are still further limited.* The last-mentioned refer in the mainto some common monocarboxylic acids (including ampholytes), mostlyin aqueous solution over a temperature range of usually 0-60" (determinedlargely by the e.m.f. method perfected by Harned), although some measure-ments on partially aqueous solutions have been performed.It is plain from the available values that the dissociation constants ofuncharged acids pass through maxima with rise of temperature, and variousattempts to relate K with temperature have been made with acids of thischarge type.fitted their earlierresults to the four-constant equation,In the first place Harned and co-workerslogK=-a/T+blogT+cT+d . . . - (1)where 27 is the temperature in degrees absolute. This they replaced Iaterby a further empirical equation: for temperatures in the vicinity of themaximum, which has the general form,log K -logK, = - p ( t - 0)2 . . . . (2)where K , is the maximum dissociation constant, 8 the corresponding tem-perature ('a), K the dissociation constant a t some other temperature t("c), and p a constant.1° An excellent account of the applicability of thisequation to the experimental data has been published by H.S. Harned andB. B. Owen,5 and a qualitative interpretation provided by R. W. Gurney.llSince then a relationship arising from theoretical treatment, and applicable,like equation (2), in the neighbourhood of K,,,, has been put forward byJ. L. Magee, T. Ri, and H. Eyring,12 vix.,In K - In K, = p(t - O)a + q(t - e)3 + . . . (3)in which the constants p and q are dependent largely on the propertiesof the solvent, in the case of water. This is similar to equation (2) apartfrom the cubic and higher terms; the inclusion of the cubic term leads to aslightly different curve in the plot of log K against t, which the authorsconsider to fit the experimental points better.* These have been listed by :H. S. Harned and B. B. Owen, Chent. Reviews, 1939, 25, 31.D.H. Everett and W. F . K. Wynne-Jones, Trams. Faraday SOC., 1939, 35, 1380.7 J. F. 5. Dippy and H. 0. Jenkins, ibid., 1941, 37, 366.Additional measurements have lately been provided by : W. F. K. Wynne- Jones andG. Salomon, ibid., 1938, 34, 1321; H. S. Harned, J . Physical Chem., 1939, 43, 275;H. Suter and K. Lutz, Helv. Chirn. A&, 1940,23, 1191 ; J. E. Ablard, D. S. McKinney,and J. C. Warner, J . Amer. Chern. SOC., 1940, 62, 2181; F. C. Hickey, ibid., p. 2916;D. H. Everett and W. F. K. Wynne-Jones, Proc. Roy. SOC., 1941, A, 177, 499; J. H.Elliott and M. Kilpatrick, J. Physical Chem., 1941, 45, 466; H. S . Herned and R. S .Done, J . Amer. Chem. SOC., 1941, 63, 2579.H. S. Harned and R. W. Ehlers, J . Amer. Chem. SOC., 1933, 55, 2379.9 H.S. Harned and N. D. Embree, a i d . , 1934, 56, 1060.lo See Ann. Reports, 1937, 34, 101-105.11 J . Chem. Pity&, 1938, 6, 499.la Ibid., 1941, 9, 419DIPPY : PHYSICO-ORGANIC TOPIUS. 129K. S. Pitzer l3 has claimed superiority for the equation. . . . (4)A AC, In T I n K = g j + B +(where A and B are constants) in that it has wider applicability, althoughit is agreed that, within limits, this is essentially the same as the Harnedand Embree relationship. In this equation Pitzer assumes for the fattyacids a value of - 40 cals./degree for AC,, the heat capacity change, invariantwith respect to temperature within the experimental range.D. H. Everett and W. I?. K. Wynne-Jones 6 propose the equationwhich they believe to be an improvement on others; this is similar toPitzer's equation except that significance is given to constants A and B.The quantities AHo, A.Xoo and AC, (the heat and entropy of ionisation a tabsolute zero and the change of heat capacity at constant pressure, re-spectively) are considered to be unaffected by temperature. It has beenpointed out,' however, that the temperature invariance of ACp in chemicalreactions has certainly not been proved, and such an assumption is, at best,an approximation applying over a very limited temperature range.In a review of the position, H.S. Harned and R. A. Robinson 1* havecompared equation (2) with three other equations, all considered likely toaccount for change of K with temperature [one being essentially that pro-posed by Everett and Wynne-Jones, i.e., equation ( 5 ) ] .They concludethat, whereas the Harned-Embree expression is only a first approximation,the other three equations are all capable of representing the data withinthe limits of experimental accuracy, and state that on the present evidencethey are of equal merit. It is emphasised by Harned and Robinson thatexperiments over the present limited temperature ranges fail to decidewhether or not AC, is independent of temperature, because, although thethree equations mentioned seem to be of equal applicability, two of thempredict a heat capacity term proportional to the tFmperature, whilst accord-ing to the other equation (that of the form proposed by Everett and Wynne-Jones) it is independent of temperature.For practical purposes Harnedand Robinson decide in favour of the equation which lends itself best tocalculation, and thus the dissociation constant is expressed aswhere A , B and C are empirical constants derived from the experimentaldata.15 H. S. Harned and R. S. Donels have lately shown that thisequation represents well their observed values of K for formic acid in fourwater-dioxan mixtures at temperatures ranging from 0" to 50".*logK=--A/T+B-CT . . . . . (6)lS J . Amw. Chem. SOC., 1937, 59, 2365. l4 Tram. Faraday SOL, 1940,36, 078.l5 Compare with the equation of E. C. Baughan (J. Chem. Physics, 1930, 7, 951).l6 J . Amer. Chern. SOC., 1041, 03, 2579.* H. S. Harned and T. R. Dedell (J. Arner. Chem. SOC., 1941, 83, 3308) heve sinoeshown that the ionisation constants of aoetic and propionic acids in dioxan-watmmixtures can also be expressed by the Harned-Robinson equation.REP.--vOL.XXXVIII. 130 ORGANIC CHEMISTRY.At the present time, therefore, there exist a number of different equations,embodying by no means identical premises, but all capable of accountingreasonably well for the experimental facts ; their reliability for extrapolationpurposes is uncertain, however. This forms the basis of a criticism 7 of Everettand Wynne-Jones's use of the values of AHo as representing the relative" intrinsic strengths " of series of acids. Thus, it is contended that theorder of strengths n-butyric > propionic > acetic > formic, arrived a t bythese authors, is opposed t o the abundant evidence from other fields ofinquiry [which shows that alkyl attached to carbon is electron-repulsive(+ I , + M ) ] , and that it is probably the outcome of unjustifiable extra-polation.It is evident, nevertheless, that in the case of n-butyric acid thecurve obtained by plotting log K against t is displaced with respect to thecurves for the other aliphatic acids so far studied,* and that, aa a consequencethe strengths of the simpler fatty acids might not present the same sequenceat all temperatures.? This anomaly is identified with a suggested restrictingpotential arising from an attraction between the C-CH, and C=O dipoles(hydrogen- bonding), particularly in the ani0n.l' Independent evidencefavouring this suggestion has been supplied by J.P. McReynolds and J. R.Witmeyer 18 in a study of the stabilities to racemisation of certain saltscontaining aliphatic acid radicals. Also, Magee, Ri, and Eyring,l2 dis-cussing the enhanced K for n-butyric acid a t 25", indicate that for a regulargradation of dissociation constants in the fatty series the heat of ionisationof n-butyric acid would be about - 250 cals., and not - 691 cals., andthey give the postulated hydrogen-bonding as the probable reason for thislarge value. It appears probable, therefore, that the dissociation constantsof the simple fatty acids at any fixed temperature will present an orderwhich is consistent with the known influences of alkyl groups, provided thatallowance is made for an additional spatial interaction in n-butyric acidand higher acids. The use of AHo as a measure of true acid strength seemsto apply no more successfully in the benzoic acid series, where it leads toa number of conclusions which conflict with the well-defined polar effectsof substituents such as iodine and nitroxyl.Another basis of comparing acid strengths, regarded as less arbitrarythan the use of dissociation constants a t some fixed temperature, has beenproposed by Harned and Embree.g They recommend that the values of1 7 J.F. J. Dippy, J., 1938, 1222; H. 0. Jenkins and J. F. J. Dippy, J. Amer. ChenX.18 J . Amer. Chern. SOC., 1940, 62, 3148.* B. W. Gurney (ref. 11) has already indicated that there is no correlation betweenthe degree of dissociation of an acid and the value of 8.He regards the value of thelatter as dependent on the relative magnitudes of the electrostatic and non-electro-static pads of the dissociation energy. t Only in the cases of ampholytes have different sequences of acid strength beenobtained in practice by varying the temperature (compare J. F. J. Dippy and H. 0.Jenkins, ref. 7, and D. H. Everett and W. F. K. Wynne-Jones, Trans. Faruduy Soc.,1941, 38, 374), although it is not certain to what extent this is accounted for byconstitutional changes.SOC., 1940, 62, 483; see Ann. Reports, 1938, 35, 2bl.See also Everett and Wynne-Jones (ref. 6)DIPPY : PHYSTCO-ORGANIC TOPICS. 131dissociation constants at their maxima ( L e . , where AH is zero) should beselected. It seems, however, on examination of the available data (ex-cluding ampholytes), that no different sequence of strengths is exhibitedwhere this method of comparison is ad0pted.1~ Everett and Wynne-Jonesconsider that this method is inadequate.L.P. Hammett 20 has called attention to the fact that different conclu-sions regarding the influences of substituent groups would be derived fromconsideration of heats of ionisation (AH = RT2d In K/dT) on the one hand.and free energies of ionisation (AF = -RT In K ) on the other, since theeffect of a substituent on the two quantities is far from comparable. It isclear, however, that values for the free energy change can be ascertainedwith much the greater certainty, and actually AF does represent the maximumwork which the system is capable of performing.In this connexion it isnoteworthy that Harned and Done,lG in an estimate of the accuracy of thedeterminations of the thermodynamic functions evaluated from dissociationconstant-temperature data, emphasise that the value of the heat of ionisationis subject to large errors because of the difficultyof determining a quantityby differentiation.It is significant, however, that J. G. Kirkwood, F. H. Westheimer, andcollaborators 21 have successfully elaborated N. Bjerrum’s original pro-position g2 that the effect on K due to the introduction of a polar substituentinto an organic acid is primarily electrostatic in origin. On this basis, theratio of the strengths of two acids has been calculated from the electrostaticwork done in transferring a proton from one acid to the anion of the other ;the formulation takes into account the position of the substituent and theshape of the molecule.Monobasic and dibasic aliphatic and monobasicaromatic acids and certain phenols have been examined, and, on the whole,a close correspondence between predicted and observed strengths has beennoted. A similar fundamental assumption is implicit in H. 0. Jenkins’scorrelations.In short, the foregoing criticisms have to contend with the fact thatdissociation constants for a fixed temperature display pronounced regularitieswhich are in harmony with a mass of observations proceeding from otherinvestigations, and that such agreement would scarcely have existed unlessthe method of comparing acid and base strengths was, in the main, a goodapproximation to the truth.There remains the question as to whether identical conclusions would bereached regarding the relative polar effects of substituent groups if data fororganic acids in solvents other than water were taken into account, It wasindicated by W.F. K. Wynne-Jones that examination of the data for acollection of acids in alcohols and water showed differing orders of strengths,1s See J. F. J. Dippy, J., 1938, 1222. 2o J . Chem. Physics, 1936, 4, 613.21 Ibid,, 1938, 6, 50’7, 513; J . A m r . Chena. SOC., 1939, 61, 555, 1977; cornparsA. Eucken, Angew. Chem., 1932, 45, 203; G . Schwarzenbach and H. Egli, Helv. Chini.Acta, 1934, 7, 1183.a2 Z . physikal. Chena., 1923, 108, 219. 29 Chem, and I n d ., 1933, 52, 273132 ORaAKIU UHEMISTRY.but G. N. Burkhardt pointed out that the irregularities were due to theinclusion of certain acids already recognised as exhibiting abnormally highatrengths. Nevertheless, by means of these data Wynne-Jones 25 tested satis-factorily his linear relationship, log K, oc 1 /D (where K, is the ratio of the dis-sociation constant of a given acid to the dissociation constant of a referenceacid, often the parent acid, and D is the dielectric constant of the solvent),deduoed on the basis of electrostatic theory and involving simplifyingassumptions such &B neglect of non-electrostatic factors.26 He suggested,for purposes of comparison, the use of '' intrinsic strengths " derived byextrapolation to i&nite dielectric constant ; it is noteworthy, however,that these intrinsic strengths show the same sequence as the dissociationconstants of the acids in water.3A bibliography of measurements on acid strengths in non-aqueoussolvents up to the year 1931 was provided by N.F. Hall; 27 a list of thereferences to later data for acids and bases in non-aqueous and partiallyaqueous solutions is included in the footnotes below.2848 Almost always thetemperature of experiment was 25".I n a number of these investigations the Wynne- Jones relationship hasbeen tested and found to be applicable 379 39,42*45~46, 473 49 (in the case ofpartially aqueous solutions the dielectric constant of the mixed solvent isvaried by adjusting the proportions of the components). M.Kilpatrick24 Chem. and Ind., 1933,62, 330.26 Proc. Roy. SOC., 1933, A , 140, 440; see also L. J. Minnick and M. Kilpatrick,20 Compare Ann. Reports, 1934, 31, 7 8 .28 J. 0. Halford, J . Amer. Chem. SOC., 1931, 53, 2944.80 V. K. LaMer and H. S. Domes, J . Amer. Chem. SOC., 1933, 55, 1840; Clieijz.31 G. E. K. Branch, D. L. Yabroff, and collaborators, J . Amer. Chem. Xoc., 1933,32 J. W. Murray and N. E. Gordon, ibid., 1935, 57, 110.s3 L. A. Wooten and L. P. Hammett, ibid., p. 2289.3J, G. M. Bennett, G. L. Brooks, and S. Glasstone, J., 1935, 1821.36 S. KiIpi and H. Warsila, 2. physikal. Chem., 1936, A , 177, 427.30 F. H. Verhoek, J . Amer. Chem. SOC., 1936, 58, 2577.37 R. B. Mason and M. Kilpatrick, ibid., 1937, 59, 572.3 8 D.C. Griffiths, J., 1938, 818.39 C. C. Lynch and V. K. LaMer, J . Amer. Chem. SOC., 1938, 60, 1252.40 W. C. Davies, J., 1938, 1866.41 H. H. Hodgson and R. Smith, J., 1939, 263.42 L. J. Minnick and M. Kilpatrick, J . Physical Chem., 1939, 43, 259.43 N. A. Izmailov, M. B. Schustova, and N. Vorodez, J . Qen. Chent. Russia, 1939,44 J. N. Beliaev, Kolloid Schurn., 1940, 6, 531.4 5 13. Adell, 2. physikal. Chem., 1940, 186, 27.4 6 M. Kilpatrick and W. H. Mears, J . Amer. Chem. SOC., 1940, 62, 3047, 3051.4 1 J. H. Elliott and M. Kilpatrick, J . Physical Chem., 1941, 45, 454, 466, 472, 485.48 R. D. Kleene, F. H. Westheher, and G. W. Wheland, J . Amer. Chem. Suc.,4' H. S . Harned, J . Physical Chem., 1939, 43, 275.J . Physical Chem., 1939, 43, 259.21 Chem.Reviews, 1931, 8, 191.M. Kilpatrick and M. L. Kilpatrick, Chem. Reviews, 1933, 13, 131.Reviews, 1933, 13, 47.55, 2935; 1934, 56, 937, 1850, 1865.9, 698.1941,63, 791DIPPY PHYSICO-ORQBNIC TOPICS. 133and co-workers 42p47 and C. C. Lynch and V. K. LaMer 39 have demonstrated,nevertheless, that that the relationship does not hold when the mediumpossesses a dielectric constant of less than 20-25. J. H. Elliott and M.Kilpafrick 47 believe that this may be due to the larger part played by dipoleinteractions between acid and solvent, and E. C. Baughan 50 considers thatsuch a breakdown might be expected. The first-mentioned authors alsoencountered lack of linearity with substituted benzoic acids in dioxan-water mixtures of D varying from 55 to 15, and they suggest that thispossibly arises from a preferential orientation of dioxan molecules aroundthe solute, which would cause a lower dielectric constant in the immediatevicinity of the acid molecules; L.A. Woofen and L. P. Hammett 33 alsoobtained little better than qualitative agreement. It is noteworthy thatorharily the slopes of the straight lines obtained in the plot of log K ,against 1/D have a positive slope, e.g., almost all m- and p-substitutedbenzoic acids, and are often roughly parallel; with o-substituted benzoicacids the lines are inclined in the opposite direction, although salicylic acidproves exceptional in this respect. Elliott and Kilpatrick attribute thisfeature to chelation between the substituent and the hydrogen of carboxylin the case o-chloro- and o-nitro-benzoic acids, a view which conflicts withH.0. Jenkins’s suggestion 51 that the strengths of these acids (in water)can be adequately accounted for simply by ascribing to the groups theirordinary polar characteristics.There can be no doubt that, despite the success which the Wynne-Jonesrelation has achieved, the chemical r61e of the solvent is by no means aninsignificant factor governing the extent of acid dissociation. The infer-vention of this factor has been mentioned in earlier Reports and elsewhere.52It is interesting to note that, although Minnick and Kilpatrick have statedthat the relative strengths of carboxylic acids are the same in two solventsof identical dielectric constant, vix., methyl and ethyl alcohols, on the onehand, and dioxan-water mixtures on the other, Elliott and Kilpatrick47now contradict this claim after further measurements with similar solventsbut using a different method of procedure.Actually, it has been stressedmore than once 37, 53 that in order to test the Wynne-Jones expressionsatisfactorily, data referring to solvents of similar chemical type should beselected, and, moreover, the data should all be derived from a consistentexperimental method.42* 47 It is possible that in these circumstances thechemical factor introduced by the solvent will be cancelled out. In supportof this stipulation a few instances may be quoted in which a change in thechemical character of the solvent brings about a distinct increase in thestrength of the dissolved acid, which cannot be attributed to the alterationGo J .Chem. Physics, 1939, 7 , 951.62 L. P. Hammett, J . Amer. Chem. SOC., 1928, 50, 2666; “ Physical OrganicChemistry,” New York, 1940, p. 256; Ann. Reports, 1930,27,326-356; J. 0. Half’ord,J . Amer. Chem. SOC., 1931, 53, 2939; C. A. Kraus, J . Franklin Inst., 1938, 225,702-7 0 7 ; J. F. J. Dippy, Chem. Reviews, 1939, 25, 166; J., 1941, 650; W. F. Luder, Chem.Reviews, 1940, 27, 555-568.ti1 J., 1939, 640.6s J. F. J. Dippy, J., 1941, 660134 ORGANIC CHEMISTRY.in the dielectric constant. Thus the acidity of hydrochloric acid in dioxanis increased by small additions of certain phenols and alcohols; * this isthought to be due to hydrogen-bond formation between hydroxyl and chlorine.Again, the strengths of seven common monocarboxylic acids have beenfound to be slightly greater in 20% aqueous sucrose than in water, althought,he dielectric constant of the former solvent is appreciably smaller thanthat of the latter.53 It is also noteworthy that acids of the ammonium-iontype (BH+), the dissociation of which should be scarcely affected by a changeof dielectric constant, show an increased acidity in ethyl alcohol as comparedwith water.55 *The various solvents employed so far include methyl, ethyl, and n-butyl alcohols, glycol, benzene, chlorobenzene, formamide and acetonitrile,and also aqueous methyl and ethyl alcohols, glycerol, dioxan, acetone andsucrose (the solvent in certain cases also contained some inert electrolyte,cq., lithium chloride).Among the acids and bases examined have beenaliphatic acids, substituted benzoic, cinnainic and phenylboric acids, halo-geno-phenols and -anilines, and substituted dimethylanilines. For the mostpart, potentiometric and colorimetric methods of measurement have beenused.In discussing the relative strengths of acids in a pair of solvents, J. 0.Halford 56, 28 has pointed out that, although variations in relative strengthwith change of solvent are only minor among acids of one charge type, itmight be unwise to draw conclusions, especially of a quantitative nature,where differences in absolute strength are less than 1pK unit, i.e., a, factorof 10 in K.57 Nevertheless, examination of the present evidence showsthat the order of strengths is quite well preserved from solvent to solventamong uncharged acids differing in K by much less than this factor.Anumber of investigators have actually stated that their measurements onseries of acids (and bases) in a given solvent reveal a close correspondencewith the order of strengths in water. In some cases quantitative agree-ment has been noted ; e.g., N. F. Hall 58 has shown that the relative strengthsof a large number of organic bases in acetic acid and in water are nearlyproportional. P. H. Verhoek 36 obtained straight lines on plotting p~ forsolutions of a variety of carboxylic acids and phenols in formamide againstvalues for aqueous solutions, and Wooten and Hammett 33 arrived a t asimilar result with m- and p-substituted benzoic acids in m-butyl alcohol andwater. Again, V.K. LaMer and H. C. Downes30 record that unchargedorganic acids in benzene retain the same differences in strength among them-selves that exist in water (salicylic acid is exceptional), thus going further54 P. D. Bartlett and H. J. Dauben, J. Amer. Chem. SOC., 1940, 62, 1339.6 6 See L. P. Hammett, " Physical Organic Chemistry," New York, 1940, p. 260.6 6 J. Amer. Chem. SOC., 1931, 53, 2939. 67 Compare Ann. Reports, 1934, 31, 78.s0 J . Amer. Chem. SOC., 1930, 52, 5115.* F. J. Moore and S. B. Johns ( J . ArneT. Chem. SOC., 1941, 83, 3336) have recentlyrecorded that the ionisation constants of picric acid in acetone, methyl ethyl ketone,acetophenone, propionitrile, and benzonitrile depend less on the dielectric constant ofthe medium than on the electron-sharing ability of the radicals in the solvent moleculeDIPPY : PHYSICO-ORGAMC! TOPTCS.135than J. N. Bronsted,59 who showed good qualitative correspondence forbenzene solutions; a similar conclusion warns reached by I). G. G r i f f i t h ~ , ~ ~who used chlorobenzene as solvent. It is interesting that agreement isfound with the measurements in aprotic solvents, i.e., inert diluents havingneither proton-accepting nor proton-donating character, in which it is tobe expected that acids will exhibit their true relative strengths. This givesstrong support for the practice of employing dissociation constants foraqueous solutions of acids in discussions concerning the polar influences ofsubstituents.Indeed, the conclusion has been drawn 3, 53 that existingdata show that, in general, organic acids (uncharged) exhibit the samerelative strengths in proceeding from solvent to solvent, provided thatcertain well-defined acids are excluded, particularly those in which thereexists some specific interaction of groups. The qualification might be addedthat it is safer to restrict correlations to acids of similar c1assJ6O i.e., where thereacting groups are alike. For instance, i t is better to consider mono-carboxylic acids as apart from phenols ; thus Verhoek 36 found that thedata for these two classes of acid gave separate straight lines in the plotof for formamide and aqueous solutions. Notable anomalies are salicylicand o-toluic acids, where the hydroxyl and methyl snbstituents are believedto form a hydrogen-bond with the carboxyl group (on this basis n-butyricand higher aliphatic acids should also be anomalous).Wooten and Ham-mett 33 have made the important observation that, in transferring fromwater to n-butyl alcohol, any substituent whidh causes a rise in absolutestrength causes an increase in relative strength (hence the positive slope inthe plot of log K, against l/D), the converse being true for a substituentwhich depresses absolute strength. They exclude from this generalisationnot only o-substituted benzoic acids but a-substituted aliphatic acids as well(proximity effect). o-Substituted benzoic acids were among those acidswhich Burkhardt described as anomalous, and furthermore, they provideexamples of negative slopes in the log Kr-l/D plot.In such cases theabnormality caused by the operation of the additional factor might notbe reproduced systematically in all sovents,3 because, as Hammett 61indicates, the chemical nature of the solvent can influence the extent ofthe abnormality. It is noteworthy, therefore, that whereas Elliott andKilpatrick 47 have asserted that the relative strengths of o-, m- and p -nitro-, -halogeno-, -methyl-, -hydroxy- and -methoxy-benzoic acids inmethyl, ethyl, and n-butyl alcohols and ethylene glycol do not present thesame sequence as in water, a detailed examination of their results 63 revealsthat, when the above anomalous acids are excluded, a very close correspond-ence with the data for aqueous solutions exists.I n conclusion, it may be said that, until an ideal method of comparing thestrengths of weak acids is forthcoming, i t would appear that the selection59 Ber., 1928, 61, 2049.*O See G. M.Bennett, G. L. Brooks, and S. Glasstone (ref. 34), and J. F. J. Dippy81 L. P. Hammett, “ Physical Organic Chemistry,” New York, 1940, p. 207.(ref. 3)136 0RUA.NIC OHEMISTRY.of thermodynamic dissociation constants referring to the solvent waterand a temperature of 25” is reasonable. The pronounced regularities dis-closed by inspection of the existing data certainly seem to give it justifimtion,and changes in the sequences of acid strengths occasioned by variations oftemperature and solvent appear likely to be only of a minor character andto admit of a simple explanation consistent with the present views regardingthe polar effects of substituents.J. F. J. D.3. ORGANOMETALLIC COMPOUNDS.Since the appearance of the last reports1 on this subject, continuedattention has been given to compounds of metals from all parts of theperiodic table, largely in order to systematise on a broader basis their existence,composition, and reactions, and partly in view of their importance in medicineand the arts. Compounds of the less typically metallic elements have beenexcluded from the following account except for purposes of comparison.General.2* 3The preparation of organic compounds of further elements or of newtypes is significant in view of attempts to define the kinds of derivativespossible to elements in the various regions of the periodic table.Triethyl-scandium and triethylyttrium * appear to be the first exceptions to A. vonGrosse’s generalisation that the transition elements do not give compoundsof the type R,M, where n is the group valency of M towards hydrogen.Among newer types may also be cited the reactive tri-alkyl or -arylderivatives of gallium,6* 7. indium,*# and thalliurn,lO~ l1 together with the“ mixed ’’ gallium compounds Me,GaCl and MeGaC1,; compounds ofunivalent thallium may also exist.ll V. M. Pletz reports highly unstablebutyltitanium ethoxides,12 and there are indications of the formation ofAnn. Reporta, 1928, 25, 92; 1932, 29, 96, 98; see also 1937, 34, 243.2 E.Krause and A. von Grosse, “ Chemie der metall-organischen Verbindungen ”(1937).K. Gilman, “ Organic Chemistry,” Chap. 4 (1938).2. anorg. Chm., 1926, 152, 145.4 V. M. Pletz, Compt. rend. Acad. Sci. U.R.S.S., 1938, 20, 27.6 G. Renwanz, Ber., 1932, 65, 1308; C. A. Kraus and F. E. Toonder, Proc. Nut.Acad. Sci., 1933,19, 292, 298; J . Amer. Chem. Soc., 1933, 65, 3547; L. M. Dennis andW. Patnode, %id., 1932, 54, 182.7 H. Gilman and R. G. Jones, ibid., 1940, 62, 980. * A. W. Laubengayer and W. F. Gilliam, ibid., 1941, 63, 477.s L. M. Dennis, R. W. Work, E. G. Rochow, and E. M. Chamot, ibid., 1934, 56,1047; W. C. Sohumb and H. I. Crane, ibid., 1938,60, 306; H. Gilman and R. G. Jones,ibid., 1940, 82, 2353.10 S. F. Birch, J., 1934, 1132; E.G. Rochow and L. M. Dennis, J . Amer. Chem.~ o c . , 1935, 57, 486.11 H. Gilman and R. G. Jones, ibid., 1939, 61, 1613; 1940, 62, 2357.1’ J . Ben. Chem. Russia, 1938, 8, 1298; cf. L. G. Makarova and A. N. Nesmejanov,aid., 1939, 9, 771STEVENS : ORGANOMETALLIC COMPOUNDS. 137organo-vanadium l3 and -tantalum c0mpounds.1~ Phenyl derivatives ofmolybdenum,15 tungsten,ls and manganese 1' have been recorded, but fulldescription is still lacking ; tri( ?)methylrhenium is described in a brief note.18H. Gilman and M. Lichtenwalter l9 have prepared the first " simple "organoplatinum compounds, Me,Pt and Me6Pt,.Et,Sc,Et20 ......Et,Y ,Et20 .........Me,Ga ............Et,T1 ...............PhMnI ............Me,Re ............Me,Pt ...............Me,h ...............M.p.-- 19"89- 63solidcryst.-TABLE I.B. p. Action of: air. water. cold HCI.170-172" ++ +222-225 + +55.7 ++ ICH, ZCH,50/24 mm. + 2CH,54.8/1.5 mm. - 1C2H, 1C2H,ca. 60 -- ICH,- + - ++ I+ + spontaneously inflammable.Physiwcherniuzl Properties.-The vapour density of trimethylaluminiumat 70" corresponds to the formula Al,Me,, and the values at higher temper-atures indicate a heat of dissociation of some 20 cals., comparable with thatof the dimeric aluminium halides. The usual formulation (I) for the latteris not applicable to the alkyls, and the possibility of AI-A1 bonding is con-sidered20 to be supported by determinations of the dipole moments oftrimethylaluminium and the methylaluminium halides.x\Al,x, Triethylaluminium also is associated in the vapourAn elaborate study 21 has been made of the molecular Xvolumes, heats of combustion, and refractivities ofnumerous alkyl derivatives of mercury, tin, lead, and the group V metals.A.von Grosse 2* discusses the physical constants of organic compoundsof the elements in relation to the periodic table.Electrolysis of a series of alkylmagnesium halides in ethyl or butylether 22 gave one equivalent of magnesium per Faraday at the cathode,and one molecule of magnesium halide at the anode. With methylmagnesiumhalides ethane is the predominant gaseous product at high current densities ;l3 C. C. Vernon, J . Amer. Chem. SOC., 1931, 53, 3831; A. V. Kirsanov and T. V.l4 B. N. Afanasyev, Chem. and Jnd., 1940, 59, 631.l6 F.Hein, Angew. Chem., 1938, 51, 503.l6 F. Rein and E. Nebe, Naturwiss., 1940, 28, 93.l7 H. Gilman and J. C. Bailie, J . Org. Chem., 1937, 2, 84.la J. G. F. Druce, J., 1934, 1129; cf. H. Gilmrtn, R. G. Jones, F. W. Moore, and/ kx/' \x state, but triethylgallium and trimethylindium are(1.1Sazonova, J . Gen. Chem. Russia, 1935, 5, 956.M. J. Kolbezen, J . Amer. Chem. SOC., 1941, 63, 2525.Ibid., 1938, 60, 3085.2o R. H. Wiswall and C. P. Smyth, J . Chem. Physics, 1941, 9, 352; of. L. 0.Brockwey and N. R. Davidson, J . Amer. Chem. Soc., 1941, 63, 3287.21 W. J. Jones et al., J . , 1931, 2109; 1932, 2284; 1935, 39; Bull. SOC. chim., 1931,49, 187 ; J . P h y k l Chem., 1933, 37, 583.22 W. V. Evans et al., J .Amer. C k . Soc., 1934, 56, 654; 1936, S7, 489; 1936, 58,720, 2284; 1939,61,898; 1940,62, 534; 1941,63, 2514138 ORGANIC CHEMISTRY.its nearly complete supersession by methane (derived from reaction with thesolvent) a t low c. d.’s suggests that the methyls are not liberated in pairs.The higher alkylmagnesium compounds give the corresponding alkane andalkylene with quantities of dialkyl increasing from traces with the ethyland isopropyl through tert.-butyl and n-propyl to nearly exclusive productionof octanes from the other butyl derivatives and of dodecane from the n-hexyl.The main reactions are formulated :Arylmagnesium halides give much polyaryl, some diaryl, and much styrenefrom interaction with the solvent.Conductivity measurements and general considerations led K.A.Jensen 23 to attribute covalent structures to mixed organometallic com-pounds, their conductivity in water and in many cases their solubilitydepending on the formation of aquo-complexes-R,SnX + H,O +[R,Sn*OH,]X. C. P. S m ~ t h , ~ , from studies of the dipole moments, con-cludes that many metal-halogen bonds are largely ionic, although the carbon-metal linkages are covalent.The compounds formulated as Ph,CrX, Ph,CrX, and Ph,CrX havemagnetic moments approximating to 1.73 Bohr magnetons and are believedto contain quinquevalent chromium. It is suggested 25 that these substances,some of which yield much diphenyl on decomposition, may really containdiphenylyl groups and be [C,H,Ph*CrPh,]X, [C,H,Ph*CrPh,H]X, and[C,H,Ph*CrPhH,] X.Mercury-oZeJin Complexes.-Mercuric salts in water or alcohol combinewith ethylene, yielding products which are usually formulated asRO*CH2*CH,*HgX (Ia) and O[CH2*CH2*HgX], ; owing to the ready regener-ation of ethylene they have also been regarded as Werner complexes:26[RO*Hg*C,H,]X (Ib) and [C2H4*Hg*O*Hg*C2H4]X2. After it had been dis-covered 27 that diaryltin dichlorides react with inorganic mercury corn-pounds to give diarylmercury but with alkylmercury halides furnish mixedalkylarylmercury, the reaction was applied 26 to the compound (I ; R = H,X = Br).This gave only hydroxyethyl-p-tolylmercury (which affordedethylene quantitatively with hydrochloric acid !). On the other hand,28 2. anorg. Chem., 1937, 230, 277.24 J .Org. Chern., 1941, 6, 421; G. L. Lewis, P. F. Oesper, and C. P, Smyth,J . Arner. Chem. SOC., 1940, 62, 3243.26 W. Klemm and (Frl.) A. Neuber, Z. anorg. Chem., 1936, 227, 261; F. Hein,ibid., p. 272.26 Summary: A. N. Nesmejanov and R. C . Freidlina, Compt. rend. Acad. Sci.U.R.S.S., 1940, 26, 60; Ber., 1936, 69, 1631; R. N. Keller, Chem. Reviews, 1941,28, 229.27 A. N. Nesmejanov and K. -4. Kotscheschkov, Ber., 1934, 67, 317; idem andR. C. Freidlina, Ber., 1936, 68, 665STEVENS : ORQANOblETALLIC COMPOUNDS. 139the substances HgCl,,C2H2 and HgCl2,2C2H, yielded diarylmercury, andtriphenylphosphine displaced acetylene from them, giving HgCl,,PPh3.26I n an attempted rational synthesis of (Ia, R = Et),28 the saItEtOCH,*CH2*S0,*HgCl was boiled with water, but ethylene was producedquantitatively. Since ethylene is formed from EtOCH,*CH,Br and Mg,29and from CPh,*CH,*CH,I and Na,m and the volatile and no doubt correctlyformulated '' Lewisite, ' ' CHCI:CH*AsCl,, affords acetylene on treatmentwith alkali, it seems dangerous to base co-ordination formulae on the mereready regeneration of unsaturated hydrocarbon.R. C. Freidlina and A. N.Nesmejanov 26 emphasise the continuity in behaviour from admittedprincipal valency compounds through substances of controversial structureto recognised Werner complexes, and offer spectroscopic evidence thatresonance between the two types occurs in mercuric chloride-acetylenecompounds.General Chemical Behaviour.-With compounds containing activehydrogen, organometallic substances undergo fission, RM + HX --+ RH +MX, which proceeds with very different degrees of facility, the more easily,broadly speaking, the baser the metal (compare Tables I and 11).Freehydrogen at room temperature decomposes reactive organometallic com-pounds analogou~ly,~~ PhNa + H, + PhH + NaH, the speed increasingin the order PhCaI, PhLi, PhNa, PhK, PhRb, PhCs. The reaction may beinfluenced by other, more specific, properties of the reagent as well as by itsacidity. Thus tetraethyl-lead and triethyl- bismuth are unaffected by organicacids under mild conditions, but yield ethane quantitatively with thiols,and have been recommended for the " Zerevitinov " determination of SHin presence of OH or NH.33Different organic radicals, moreover, are not detached equally readily.The partial scission of unsymmetrical mercury compounds enables radicalsto be arranged in a series l* 34 of decreasing lability which is regarded as aseries of diminishing electronegativity : a-thienyl > o- and p-anisyl >a-naphthyl > o-, m-, andp-tolyl > phenyl and halogenophenyl > n- andmany other alkyls > benzyl > tert.-butyl and neopentyl.Compounds ofother metals furnish series very similar but less elaborately investigated-germanium : 35 p-tolyl > m-tolyl > phenyl > benzyl; tin : 36 a-thienyl >H. J. Lucas and S. Winstein have expressed similar views.312 8 J. D. Loudon and N. Shulman, J., 1939, 1066.*@ R. C. Tallman, J . Amer. Chem SOC., 1934, 58, 126.3O C.B. Wooster and R. A. Morse, ibid., p. 1735.31 Ibid., 1938,80, 836; 1939, 81, 3102 (with F. R. Hepner).32 H. Gilman, A. L. Jacoby, and (Miss) H. Ludeman, J . Amer. Chent. SOC., 1938,33 H. Gilman and J. F. Nelson, ibid., 1937,59, 935.34 M. S. Kharasch, H. Pines, and (Miss) J. H. Levine, J . Org. Chem., 1938, 3, 347;M. S. Kharasch and S. Swartz, ibid., p. 405; M. S. Kharasch, R. R. Legault, andW. R. Sprouls, ibid., p. 409; F. C. Whitmore and H. Bernstein, J . Amer. Chem. SOC.,1935, 80, 2626.60, 2336 ; see also W. H. Zartrnan and H. Adkins, ibid., 1932,54, 3398.36 J. K. Simons, ibid., 1935, 57, 1299.a * T. S. Bobaschinskaja and K. A. Kotscheschkov, J . Qen. C'henb. Russia, 1938,8, 1860140 ORGANIC CHEMISTRY.p-anisyl > a-naphthyl > phenyl > cyclohexyl ; Z e d : 37 a-fury1 > a-thienyl,p-ltnisyl > phenyl > ethyl, benzyl; or-naphthyl > phenyl.A. N. Nesme-janov and K. A. Kotscheschkov3s describe a transformation in whichelimination of a radical as hydrocarbon is in competition with an alternativemode of reaction :(A) Hg + R2SnC12+-R2Hg + SnC1, (+EtOH)+Hg +2RH + (EtO),SnCI, (B)Dominance of the mode (A) is associated with " electronegativity " of R,and decreases in the order : p-C6H,*NH2, pC,H,*OH, o-anisyl, p-naphthyl,o- and p-tolyl, p-halogenophenyl, phenyl, p-C6H4*C02Et, benzyl, ethyl.Much elaborate work has been done on the relation between the reactivityof an organometallic compound and the nature of the metal, especially itsposition in the periodic table. Mutually compatible results are obtainedonly with important restrictions as to the reactions employed in assessingreactivities.Thus ease of thermal decomposition and of atmosphericoxidation show little connection with one another or with sensitiveness towater and acids, but the last property runs roughly parallel with thesynthetically important power of addition to multiple linkages. Table I1shows the behaviour of derivatives of a number of metals towards differentreagents ; more precise comparisons, often with benzonitrile as substrate,have led H. Gilman to the following inequalities in " reactivity " : RCs >RRb > RK > RNa > RLi > RMgBr (R = CPhiC.) ; 39 RLi > RMgXor R a g ; 4O PhLi > PhCaI > PhMgI ; 41 RMgX > R3AI > R,B > R&n ; 42R2Zn > R2Cd > R2Hg ; 43 R3A1 >> R3Tl ; Ph,In > Ph3Ga 9 Ph3T1.7*TABLE 11.2 + G + a, 0 0ON ; B o o 6 g z v 0, G s d i w r n ON 9++ ++ RNa ++ ++ ++ ++ ++ ++ RCu46 ++ ++RBEX ++ ++ ++ R,Zn + + + + + + + + +RCaX44 ? ++ - ? ++ ++ RzCd43047 + + ++ ++ - + - ++- ++ RMgX -46 ++ ++ ++ ++ ++ RZHg ++ ++ R4Pb- ++ + R4SnRa-R3Ga 6* 7 + + - ++++ RSbR,In 0 + + - ++++R3T1 11 + ++ ++ ++m d m xRLi ++ ++ ++ ++ ++ ++ RAg4;* + +++ - + + +--- --- - - ++4a++ +- ?$?149 + + + + + + + + + R3BiRapid action at or near room temp.is indicated by ++, no action under ordinary8 7 H. Gilman and E. B. Tome, Rec. Trav. chim., 1932, 61, 1054; J . Amer. Chem.38 Ber., 1930, 63, 2496; Sci. Rep. M08eW State Univ., 1934, No.3, 283.3s H. Gilman and R. V. Young, J . Org. Chern., 1936, 1, 315.40 H. Gilman and R. H. Kirby, J . Amer. Chem. Soc., 1933, 55, 1265; H. Gilman4 1 H. Gilman, R. H. Kirby, M. Lichtenwalter, and R. V. Young, ibid., p. 79.48 H. Gilman and K. E. Marple, ibid., p. 133.43 H. Gilrnan and J. F. Nelson, ibid., p. 518.conditions by - ; mere addition of NH, is disregarded.Soc., 1933, 55, 4689 (with H. L. Jones) ; 1939, 61, 739.See also ref. 3.and M. Lichtenwalter Rec. Trav. chim., 1936, 55, 561STEVENS : ORCIANOMETALLIC COMPOUNDS. 141H. ailman associates the alternative courses of some reactions withthe reactivity of the organometdlic participant. Derivativea of the alkalimetals and calcium undergo predominantly normal addition t o the carbonylgroup of benzylideneacetone and the like, whereas many others affordfinally the saturated ketone CHPhR*CH,*COMe-" I : 4-addition ".Qp 49The former reagents similarly convert benzophenoneanil normally intotriphenylmethylaniline,(9 and phenylmagnesium bromide attacks thenucleus, giving o-phenylbenzhydrylaniline, C,H,Ph*CHPh*NHPh. Withazobenzene,17 aryl compounds of potassium and calcium afford triphenyl-hydrazine, whereas those of other metals, including sodium and lithium,yield hydrazobenzene and/or aniline by reduction.H.Gilman and J. F. Nelson3*43 relate the reactivity of organometalliccompounds to the position of the metal in the periodic table by a series ofrules which somewhat outrun the experimental data. Broadly speaking,in each group, Li-Cs, Be-Ca, B-Al, G-Pb, and N-Bi, the reactivityof the organic derivatives rises with the atomic weight of the metal, and fallswith increasing group number from I to IV. In each B group, Cu-Ag,Zn-Hg, and Ga-T1, the reactivity is less than in the corresponding Agroup and appears to fall with increasing atomic weight of the metal. Inmany cases the mixed organic compounds of a metal are a little less reactivethan the simple ones, but substances of the types RBX, and R,TlX areextraordinarily inert in comparison with R3B or R,TI.Attempts to correlate by various methods the reactivity of Grignardreagents with the nature of the hydrocarbon radical have given contra-dictory results. 60A redistribution of organic radicals between metallic atoms, as in thecase Me,Hg + Et,Hg 2MeEtHg, has been occasionally encountered,and now forms the subject of an elaborate study by G. Calingaert and co-w0rkers.~1 In the example cited, the equilibrium mixture attained fromeither direction contains 50 molecules yo of the mixed dialkylmercury and25% of each of the simple ones, corresponding to a random distribution ofradicals between the metallic atoms. If excess of one dialkylmercury istaken, the new equilibrium corresponds to a new random distribution,and the same holds for the more complicated case of a pair of alkyl-leadswhich gives a mixture of R,Pb, R,R'Pb, R,R',Pb, RR',Pb, and R',Pb.STEVENS : HETEROCYCLIC COMPOUNDS. 227L. Pauling and J. H. Sturdivant63 suggested that cyameluric andhydromelonic a,cids, melam, melem, and melon, compounds derived directlyor indirectly from pyrolysis of thiocyanogen derivatives,were related to a nucleus C6N7 in the same way as the/ \ cyanuric compounds to the C,N, ring. For this nucleusthey proposed the structure (XI) with the large calculated ' resonance energy of 150 Caltls. C. E. Redemann andH. J. Lucasa show that hydromelonic acid, regarded asC6N7(NH*CN),, like tricyanomelamine, C,N,( NH-CN),,is too strong an acid for its dissociation constant to bemeasured, and cyameluric acid, C6N,( OH),, like cyanuric,is much weaker. Hydromelonic acid gives on hydrolysiscyameluric acid which has been converted into thechloride C6N,C1, and into (mainly N - ) alkyl derivatives and which yieldsone molecule only of cyanuric acid on further hydrolysis. Melam is nowbelieved to be an imide (NH2),C3N3*NH*C,N3(NH,),, and melem to becyameluramide. Melon is probably a mixture of intermediate stages of" deammonation " between melem and a graphite-like (C3N&.Pterins.-These pigments of butterflies, wasps, and other insects, arenot easily purified and characterised and are also difficult to combust. Acareful revision of the analyses removes the necessity of assigning to themformulae containing nineteen carbon atoms with three pyrimidine or purinenuclei.65 Leucopterin is now formulated C6H503N5 (the trimethyl derivativegives the expected molecular weight in phenol)70 and has been synthesisedby fusing 2 : 4 : 5-triamino-6-hydroxyyrimidine with oxalic acid.66 Theformulse (XII) and (XIII) are suggested for the s~bstance.~5~6*~ 69 Con-densation under milder conditions with dichloroacetic in place of oxalic acidgives xanthopterin C6H302N5,6g which yields on oxidation leucopterin (not,as previously reported, an iminoleucopterin) ,65 " Anhydroleucopterin '' isshown to be a deoxyleucopterin, distinct from 6-deoxyleucopterin preparedfrom the 6-chloro-compound.67 Guanopterin is identical with isoguanine.66(XII.) HN=V fi-NH>c.Co2H HN=V G-NH-70 (XTlJ.)