摘要:

QUARTERLY REVIEWS ASYMMETRIC TRANSFORMATION AND ASYMMETRIC INDUCTION By E. E. TURNER M.A. D.Sc. F.R.I.C. F.R.S. and MARGARET M. URIS,* B.Sc. PH.D. (LECTURER IN CHEMISTRY BEDFORD COLLEGE UNIVERSITY OF LONDON) [PROFESSOR OF CHEMISTRY (BEDFORD COLLEGE) UNIVERSI!I!Y OF LONDON] Asymmetric Transformation Introductory ASYMMETRIC transformation involves no synthetic factor and is concerned with stereochemical changes only. The term was used by H. Leuchs and J. Wutke in 1913 to describe the observed fact that addition of brucine to dl-2-o-carboxybenzyl-a-hydrindone (VII) in acetone solution resulted in an “ asymmetric transformation ” of the initially formed an& dissolved base.dl-acid into solid base.d-acid in yield showing that practically all the base. I-acid in solution had been transformed into solid diastereoisomeric base.d-acid.Removal of the brucine gave a dextrorotatory acid which readily racemised. Other workers discovered analogous cases and occa- sionally the words “optical activation” were used thus the brucine was said to have activated the inactive Leuchs acid. Attention was con- centrated on what came out of solution rather than what was happening in solution. An important experimental observation of a different kind was made by J. Read and A. M. McMath,2 who found that the I-hydroxyhydrindamine salts (I) of I- and of dl-chlorobromomethanesulphonic acid exhibited in dry acetone solution a rotational change which could only be explained on the assumption of the existence of an equilibrium I-Base. 1-Acid + I-Base. &-Acid which waa greatly in favour of the Z.Z-salt.Again the idea of activation of a potentially active molecule (that of the acid) by a stably active molecule (that of brucine) was put forward although it waa found impossible to isolate an optically active specimen of the free acid. In 1928 W. H. Mills and K. A. C. Elliott 3 observed the partial “ activation ” of N-benzene- sulphonyl-8-nitro-1 -naphthylglycine (XII) by means of an approximate equivalent of brucine in chloroform solution; in the same research these authors obtained by processes depending on asymmetric transformation both the d- and the Z-acid which had low but appreciable optical stability. Ber. 1913 46,2420. J. 1926,127 1572. a J. 1928 1291. * (nCe Jamison.) 299 U 300 QUARTERLY REVIEWS The discovery that the salt of 4 4'-dinitrodiphenic acid (XXIII) with the laevorotatory base quinine was strongly dextrorotatory in solution and more- over was apparently a single individual but that all attempts to liberate an active acid failed led R.Kuhn4 in 1932 to introduce the expreasion " asymmetrische Umlagerung erster Art " to describe this and other cases which seemed to him similar (Read and McMath; P. Pfeif€er and his co- workers 5). This expression was translated " asymmetric transformation of the first order," but " sort " or " kind " or " type " would have been a happier rendering of " Art " since " order " raises thoughts of reaction kinetics. R. Kuhn proposed the term " asymmetric transformation of the second order " for cases such as that of Leuchs where one diastereoisomeride was obtained in preponderating quantity and removal of the activating base led to the isolation of an optically active acid.H. King,6 in 1933 made a useful survey of the subject up to that date. M. M. Jamison and E. E. Turner 7 re-defined first-order asymmetric transformation as a phenomenon relating only to the attainment of an equilibrium while second-order transformation denotes the appearance of a second phase. They showed that one essential condition for first-order transformation was the real existence of diastereoisomerides in solution the cause of any observed mutarotation being the approach to an equilibrium such as d-Base. d-Acid + d-Base . 1-Acid where the base is optically stable and the acid optically labile salts will only show the effect in solutions in which ionic dissociation is largely absent. There are however numerous examples of diastereoisomerism apart from those concerned with salts ; thus a- and ,8-sugars are in the present sense diastereoisomeric and their examination can be extended to aqueous solutions in which salt-diastereoisomerides would as *a rule be ionically dissociated.In fact the " mutarotation " of sugars which has been long studied must now be reconsidered in the light of the conception of first- order transformation as must also the interesting studies of '' asymmetric catalytic racemisation " made by A. McKenzie and his co-workers 8 which can be more clearly interpreted now that first-order transformation has been investigated using compounds deliberately synthesised for the purpose. Before discussing individual examples of asymmetric transformation it is pertinent to consider the kind of molecule which possesses unstable dis- symmetry.(Up to the present most experiments have been made in the normal range of temperatures but at temperatures considerably higher than the ordinary many optically stable molecules would become optically labile and the investigation of derivatives of such compounds as tartaric acid at temperatures a t which they racemise readily offers inkeresting problems.) A large proportiop of the experimental evidence in connection with asymmetric transformations has been obtained from a study of molecules Ber. 1932 66 49. Ann. Reports 1933 30 261. A. McKenzie and I. A. Smith J. 1924 125 1582 ; Ber. 1926 58 894. Ibid. 1931 64 2667 ; 1932 65 560 ; 1933 66 416. J . 1942 437. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC.301 owing their dissymmetry to restriction of rotation about a single bond. A second class of molecule which has contributed is that in which interconver- sion of antipodal forms is rendered possible by the operation of prototropic changes this class including some sugars. To these can be added a number of rather miscellaneous compounds among which are certain complex salts.; The total number of compounds the configuration of which can readily be inverted is relatively small ; what might be called the average asymmetric carbon atom offers considerable resistance to inversion. (We are not con- cerned in the present article with inversion during replacement reactions of the type associated with the name of Walden although cases can be fore- seen where the steric aspects of “ aliphatic substitution ” and Grst-order asymmetric transformation may well have common experimental material.) Compounds owing their optical activity to restriction of rotation about a single bond provide the most convenient material for a study of the two types of asymmetric transformations.They are suitable material for examination because their racemisation is of purely physical origin and therefore spontaneous that is not generally subject to acceleration by the action of catalysts. Some members of this class give active forms of very high optical stability e.g. 6 6’-dinitrodiphenic acid while the optically least stable compounds yet known to show measurable mutarotation are certain alkaloidal salts 9 of N-benzoyl-2 4-dichlorodiphenylamine-2’-carboxylic acid (B p. 325) and N-benzoyl-2 4-dimethyldiphenylamine-2’-carboxylic acid (C p.325). These acids belong to one of the most convenient and accessible series of compounds in the restricted rotation class. Substituted diphenyls offer a large field but particular individuals are difficult to prepare in any quantity. PeriDisubstituted naphthalenes of which the MUs-Elliott acid mentioned above is the best example provide a useful but limited field and are tedious to synthesise. Enantiomeric pairs which racemise by a tautomeric mechanism are inter- convertible in ways such as the following R H R R CO*R” \ \ / C 2 - C=C(OH)R” .+ C \ / / \ / / \ ( a ) H R’ CO-R” R’ R’ Inactive intermediate U B Skeletal f o m of a sugar R H R R S0,R” ( c ) C 3 - \ C=SO(OH)R” + ‘c’ \ / / \ / R‘/ ‘I3 R’ S0,R” R‘ OJ. 1938 1646.302 QUARTERtY REVIEWS These processes are subject to influence by catalysts and in some cases e.g. the (as it is now classified) first-order transformation H OH c1 the equilibrium process is too slow for measurement in absence of suitable catalysts.* The sugars provide examples of both types of asymmetric transformation. Among substances which do not fall under either of these two headings are the oxime (V) (p. 310) and benzoylphenylhydrazone of cyclohexanone-4- carboxylic acid ; 10 here the inversion mechanism is purely a configurational change depending on the stereochemical instability of the system R \ / \ C=N s C=N / R /- \ In complex salts particularly the chromioxalates (11) (p. 308),11 the mechan- ism of inversion is unknown although various obvious possibilities can be conjectured.It may be that optical instability arises from sheer chemical instability since asymmetric transformation was not observed by Werner in his exhaustive treatment of complex salts of the ammine type where the chemical stability is considerable. The " asymmetric tin " compounds of W. J. Pope and S. J. Peachey l2 must owe their optical instability largely to ease of inversion in the ion MeEtPrSn' or a solvated modification. Optical instability of a molecule owing its dissymmetry to " folding " was described by I. G. M. Campbell l3 in the case of 10-p-carboxyphenyl-2-methyl- phenoxstibine (VI) (p. 310). If a structure contains two " centres of asymmetry " X and Y each centre can have either a dextro- or a hvo-configuration so that under favourable conditions two pairs of diastereoisomerides are possible In the present discussion we are concerned either with ( A ) or with (B) what applies to the one equally applies to the other with all signs changed.Taking ( A ) therefore three classes are possible Class I. Both centres X and Y are configurationally stable under experimental conditions. This presents no problem since although d-X.d-Y and d-X.Z-Y must have different free energies the energy barrier which would have to be surmounted in order to bring the two compounds into mobile equilibrium is too high for attainment. ( A ) d-X.d-Y and d-X.Z-Y ( B ) Z-X.d-Y and Z-X.Z-Y. 10 W. H. Mills and A. M. Bain J. 1910 97 1866 ; 1914,105 64. l1 A. Werner B0r. 1912 45 3061. l e Proc. 1900 16 42 116. lS J. 1947 4. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC.303 Class 11. One centre say X has a high configurational stability while Y can undergo configurational inversion a t a measurable rate under the experimental conditions. An equilibrium can now be established d-X .d-Y + d - X . I-Y and its establishment may be capable of observatian as a mutarotation. Class 111. Both centres X and Y have low optical stability under experimental conditions. Starting with either d-X . d-Y or with d-X . I-Y the final result of equilibration will be a mixture of ( A ) and (B) the ratio of d - X . d-Y to d-X . I-Y (equal to that of I-X . I-Y to I - X . d - Y ) being determined by the relative free energies of the diastereoisomerides. Most of the known examples of asymmetric transformation relate to salts either the acidic or much more usually the basic part containing X the fixed asymmetric centre.In the sugar series Y represents the CH-OH group which can adopt either the a- or the p-configuration X representing the rest of the molecule and usually containing several asymmetric centres of high optical stability which can be thought of as acting together as one unit. The cases dealt with under " asymmetric catalytic racemisation " are esters. For second-order asymmetric transformation to occur with a Class I1 compound the two diastereoisomerides need have no real existence in solution what is necessary is that one salt should crystallise from solution. Thus if d-X is a stably active base and &I-Y is an optically labile acid even if in solution there are merely the ions corresponding to base and acid then provided e.g.d-X . d-Y begins to crystallise the acid ion I-Y can racemise and in this way provide continually more .of the d-Y ion and so more d - X . d - Y . For first-order transformation however the two optical centres concerned must be in combination and this means that the solvent must be one in which little ionisation occurs in particular a non-hydroxylic one. This condition fulfilled let us suppose the d-form of an optically stable base to be in solution. On adding an equivalent of the dl-form of an optically unstable acid there is formed at once in solution ( 1 ) &Base. d-Acid + &Base. I-Acid Owing to the different free energies of the two diastereoisomerides equilibra- tion (first-order asymmetric transformation) will occur until we reach the composition (2) d-Base.d-Acid + d-Base.1-Acid This can sometimes be followed polarimetrically considerable rotational changes being observed In other cases the equilibrium may be reached too quickly for observation ; or the difference between the rotation of the partial racemate and that of the equilibrium mixture may be very small.Since first-order transformation depends on a difference of free energy between diastereoisomerides in a particular sorvent and the difference d-X.Z-Y - d-X.d-Y may not be the same as d-X'.I-Y - d - X ' . d - Y using a second base X ' the stereochemist varies both activating agent and solvent in attempting to bring first-order asymmetric transformation We are here concerned only with Class 11. 50% 50 % x% (100 - x)% 304 QUARTERLY REVIEWS involving a specific labile group within the measurable range of magnitude and velocity.it was possible to prepare the pure I-base.I-acid and determine the rate constant for the change into the equilibrium mixture in addition to determining the rate constant for the “ activation ” process vix. the change of the 50 50 mixture of I-base.d-acid and I-base.I-acid into the same equilibrium mixture. The two rate constants were found to be equal k the measured value being the sum of two rate constants kd and k, for the partial inversion of base.d-acid and base.Z-acid. (By “ partial inversion ” we mean the change I-base . d-acid -+ I-base . I-acid or I-base . I-acid + I-base . d-acid.) These changes have been described as “ partial racemisation ” but this term is misleading since the partial racemate composed of equivalent weights of d-X .d-Y and d - X . I-Y is not the equilibration product. In partial racemisation the values of kd and k are equal it is their difference that accounts for first-order transformation. In the case raised above in which second-order asymmetric transforma- tions involve crystallisation of diastereoisomerides which become ionised in solution first-order transformation is excluded and partial rucemisalion accurately describes what happens in solution. In one instance To summarise the practical aspects of resolution second-order trans- formation and first-order transformation it is convenient to consider a hypothetical case in which a dl-acid (optically unstable) and a I-base (optically stable) are dissolved in a solvent in which the salts formed are not dissociated and to predict the results of applying various conditions on the solution and what crystallises or is precipitated from it 1-B 4- dl-A dissolved .1 kl I At moment I-B.I-A of solution { 50 % tiA} - By first-order + transformation I I-B. 1-A I-B d A { 50% } and { 50% } By resolution 4 kl I I-B.l-A= 1-B.d-A x% kd (100 - .)% In solution I Com’plete qulick precip litation .1 I-B.1-A f 1-B.d-A 2% (100 - .)yo Solid mixed Slow crylstallisation 3. loo yo I-B .d-A By second-order transfoimation This scheme which is based on practical experience,g~l4 shows that the appropriate treatment of an optically labile substance with one activating l4 M. M. Jamison,’Tram. Paraday SOC. 1945 41 696. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC. 305 agent and one solvent only can produce an interesting variety of results.There may be greater variety than is here indicated the difference in free energy between the two diastereoisomerides may be so slight that in second- order transformation one form or the other may crystallise without there being an apparent difference in procedure. Ordinary resolutions in which the question of optical stability does not normally arise are sometimes complicated 1 5 = ~ by the separation of the partial racemate in crystalline form and similarly partial racemates sometimes separat,e as alternatives to the normal products of second-order asymmetric transformation. It is understandable that decomposition of an optically pure salt obtained by second-order asymmetric transformation might give an optically inactive i.e. racemised acid. How in that case can the crystallisation be classed as second-order transformation ? Obtaining an active acid is the only entirely satisfactory proof of second-order transformation but it may be suspected when a solution made up to contain a g.of an optically stable base d-X is mixed with one containing the equivalent b g. of an acid &I-Y and crystallisation produces considerably more than (a + b ) / 2 g. of solvent-free apparently homogeneous salt with a molecular rotation different from that " calculated " for the partial racemate. The suspicion is heightened if crystallisation appears to be progressive rather than sudden and if it is accelerated by gentle heating. It becomes very nearly a certainty (1) if several crops are obtained each with the same rotation and which together weigh nearly (a + b) g.(2) if a solution of the salt in the same or a different solvent exhibits mutarotation or (3) if when the salt is dissolved in a different solvent a new and uniform salt crystallises which in turn gives rise to mutarotational changes when dissolved in the same or it different solvent. A striking difference even of sign between the rotation of &-X and that of the salt which crystallises is not enough to justify the assumption of asym- metric transformation. Second-order transformations have often been called resolutions this mistake has led to the judgment that diphenyl compounds are easy to resolve when in fact it has been found easy to obtain one antipodal form only by second-order transformation. A problem of great interest is this if d-X.d-Y and d - X .Z-Y (X stable Y unstable optically) are brought together in equivalent amounts in solution and first-order asymmetric transformation leads to the equilibrium d-Xed-1' + d-X.1-Y so that more d-X.Z-Y is finally present in solution than d-X.d-Y then if crystallisation begins which of the two diastereoisomerides will separate ? With a pair of solids such as are met with in a study of allotropy or poly- morphism we should usually have a stable form and an unstable form (as distinguished from a pair of diastereoisomerides in equilibrium). Generally lSu J. Meisenheimer and 0. Beisswenger Ber. 1932 65 32. J. Meisenheimer W. Theilacker and 0. Beisswenger Annakn 1932 495 249. 306 QUARTERLY REVIEWS speaking the stable form would be less soluble than the unstable form so that apart from chance inoculation the stable form would be the one to separate if time were given for stability to assert itself thermodynamically over instability.We should have to hesitate however before answering the above question by (apparent) analogy. The answer can be given " the stable form is the more soluble " in the case of some diastereoisomeric 319 32 some esters investigated by McKenzie,8 and the only example known 7 in which both first- and second-order asymmetric transformations have been observed with one pair of diastereoisomeric salts in one and the same solvent. Examination of Experimental Material Two cases of second-order asymmetric transformation appear in Pope and Peachey 's demonstration of optical activity in tin compounds.12 Methylethyl-n-propyltin d-camphorsulphonate crystallised from water in one form only [MID + 95" in water ; [MID + 45" is the calculated value for the basic radical from this the dextrorotation was retained when the camphorsulphonate was converted into the iodide.l2 Secondly the specific rotations of successive crops of methylethyl-n-propyltin d-a-bromocamphor- sulphonate from acetone solution were constant ([HID + 318" in water 12). Since the acid radical was known to have [.MID + 270" the authors attributed + 48" to the basic part and coniirmed it by conversion into d-methylethyl-n-propyltin iodide as before. The aqueous solution [MID + 318" was heated to 100" in a sealed tube for two hours by which time its rotation bad fallen to + 273" decomposition of this solution with potassium iodide gave the inactive iodide but evaporation to dryness gave the original d-a-bromocamphorsulphonate [MID + 315 ".They ascribed the fall in rotation on heating to partial racemisation so that the whole series of changes can be expressed Sealed tube loo" 1 yo solution KIY!3tO€€ '\ 2hours / K I ~ t O H Evaporate t o dryness d-B . &A dl-B . d-A [MID f 318" [MID -k 273" I I J. d-MeEtPrSnI J. dl-MeEtPrSnI Read and McMath were able to carry out a second-order asymmetric transformation using &I-chlorobromomethanesulphonic acid in either the d- or the I-direction by using the d- or the I-hydroxyhydrindamine. The I-hydroxyhydrindamine dl-chlorobro- mornethanesulphonate (I) M[aID- 72" in methyl alcohol crystallised from acetone containing a little methyl alcohol to give a salt which while it CH*OH Br had eventually the rotation - 72" had M[aID - 173 " when first observed.H Cl 0:"" CH*NH, \ / C / \ (1.1 TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC. 307 Proof of the activity of the acid part of the salt could only be obtained by the expedient of mixing equal quantities of the d-base.d-acid and Z-base . dZ-acid salts ; a residual [MI of + 49" was then observed attempts to replace the optically active base by benzidine or a-naphthylamine gave inactive salts. When the salt Z-base. Z-acid was dissolved in specially purified anhydrous acetone it had [MID - 256" three minutes after first wetting with solvent a value which changed to - 187" in less than an hour (Fig. 1). This change might have been considered as consequent on partial racemisation had it not been that the partial racemate itself Z-base .dZ-acid when dissolved in the same solvent had [MI - 71" initially changing to - 187" on standing. This latter observation has become the classical case of first-order asymmetric transformation. If the salts are not dissociated in solution and their rota- tions are constant over the concentration ranges employed the composition as a simple calculation shows a t equilibrium is B.1-A + B.d-A 81% 19% It was of course desirable to remove the Z-hydroxyhydrindamine from the equili- brated solution to prove that the muta- rotation was due to optical activation of the acid (particularly as Z-hydroxy- hydrindamine benzenesulphonate shows (unexplained) mutarotation in methyl alcohol [MI changing from - 100" to - 76" in 8 hours) but the authors were unable to accomplish this. The experi- ments described are not suitable for correlation of the directions of first- and second-order transformation since the first- order transformation was carried out in specially purified and dried acetone and the second-order from acetone-methyl 20 40 60 Time in MinuTes.FIQ. 1 alcohol also the alternative crystalline solid which can be obtained is not the diastereoisomeride but the partial racemate Z-B.dZ-A. An inter- esting recorded observation which would be worthy of further investigation is that " an acetone solution'' of the salt deposits crystals the acetone solution of which has [MI - 93" mutarotating to - 154" and yet depositing on evaporation the crystals with [&?ID - 93". A second series of experiments was made with the same base and chloro- bromoacetic acid.l6 A solution of equimoleculaz quantities of the Z-base and dl-acid was made in chloroform containing a little methyl alcohol; slow crystallisation gave Z-base.dZ-acid [a] - 50" in the same solvent while quick cooling of a hot solution to supersaturation gave Z-base . d-acid in 75% yield [&?ID changing from the first observed 0" to - 50" on standing 16 J . 1926 2183. See also H. J. Backer and H. W. Mook J. 1928 2125. 308 QUARTERLY REVIEWS I-base.dZ-acid was deposited from the mother liquor. Although the anf$; podal forms of the above pair of salts d-base.I-acid and d-base.&-acid wem prepared attempts by mixing to observe a rotation which was due to acid only failed-an inconclusive observation of - 0.1" wim made 1-6 minufe~ after wetting with solvent. The naming of the various types of crystal is therefore conjectural.The crystals deposited from a hot solution of potassium distrychnine chromioxalate in ethyl alcohol were shown by Werner l1 to contain agym- metrically activated chromioxalate ion. The rotation of the salt a tetra- hydrate was [alG + 430" in water the part due to the chromioxalate ion (+ 0.43" observed) mutarotating to zero in la hours ; the tripotassium salt obtained from a sample of it before mutarotation was dextrorotatory. A dilute solution of potassium distrychnine chromioxalate in water deposited crystals of tristrychnine I-chromioxalate (+ 4H20) all the crops measured being laevorotatory the specific rotation [aIG was - 300" in water and decomposition with potassium iodide gave I-potassium chromioxalate (11). Werner investigated the mother liquor from which crystallisation was taking place and found it "practically inactive "; this is what might be expected if asymmetric transformation of the second order were taking place whereas resolution would result in an increase of rotation in solution of the opposite sign from that of t h solid coming out.This work of Werner's enabled gPfeif€er and K. Quehl5 to put an inter- pretation on some results they obtained in crystallising zinc &camphor- sulphonate from water in presence of o-phenanthroline. A solution of zinc phar. (111.) @-camphorsulphonate itself had a D + 0.92" (1/1000-mol. in 25 C.C. of water) while the salt Zn(phen),0*S02*Cl,H,50,7H,0 (as 111) obtained in 80 yo yield by crystallising zinc p-camphorsulphonate from water containing 3-5 mols.of o-phenanthroline had aD 0.0". When 3 mols. of o-phenanthroline were added t o a solution of zinc /?-camphorsulphonate the rotation fell immediately from + 0-92" t o + 0.09" neither ammonia pyridine nor ethylenediamine produced this diminution. Replacement of the /I-camphorsulphonate ions by nitrate or TUaNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC. 309 bromide gave inactive products ; nevertheless Pfeiffer and Quehl accounted for their observations by assuming formation of Z-[Zn(phen),] ++ under the influence of the p-camphorsulphonate ions. The quinate and a-bromo-n- camphorsulphonate ions appeared to cause similar activation (first order) of zinc complexes large changes in rotation being observed on adding o-phenanthroline or 2 2'-dipyridyl to aqueous solutions of these zinc salts.One feels a little hesitant a t accepting the interpretation of these experimental observations in solution as first-order changes since the salts [Zn(phen),]X must be completely ionised that is the transforming agent i,s separated from the complex i t is activating. In comparison with examples of similar phenomena in other fields of molecular dissymmetry it is more surprising that a cation should be able to activate a cation as the authors suggest in the following cases a solution of cinchonine hydrochloride and zinc sulphate in water showed aD + 5-29' before the addition of 3 mols. of o-phenanthroline after which it changed immediately to - 1-89" and to - 2.46" on standing. Precipitation of the cinchonine by means of alkali left an optically inactive zinc salt.Results of the same type were obtained using strychnine sulphate instead of cinchonine hydrochloride and indeed there are many similar examples in the work of PfeSer and his collaborators. It is possible that these mutarotations are not first-order transformations a t all but result from the replacement of alkaloid by o-phenanthroline in a metal complex with consequent mutarotation. The activation of the ferrioxalate ion by means of d- or Z-a-phenylethyl- amine recorded by W. Thomas l7 rests on more slender evidence but would be worthy of repetition. J. J. Woldendorp 18 records a crystallisation of strychnine hydrogen chromimalonate (C,,H,,O,N,~H)H[Cr(malonate),] resulting in deposition of Z-base . Z-acid only. W. H. Mills and R. E. D. Clark's l9 work on the complex anion of mercury with 4-chlorobenzene-1 2-dithiol (IV) is particularly interesting on account of the arguments the authors use in favour of a tetrahedral disposition of the mercury valencies.The &quinine salt was formed and on crystallisation from chloroform {Quinine H }$ solution produced an a-form which on crystallisation from acetone deposited a p-form a process which could be repeated indefinitely each crystalline form has solvent of crystallisation but the two forms remain different after the solvent is removed. The fact that no two forms of me.tallic salts were discovered may be taken as exclud- ing the possibility that the a- and /?-forms are cis- and trans-forms of a L( flat mercury cation. In spite of the fact that the authors were unable Vers. K . Akad. Wetemch. Amsterdam 1919 27 1212.pgyJ&} (IV.) acetone solution p-fom 7 4 u-form chloroform solution 7 l7 J. 1921 119 1140. lo J . 1936 175 ; see on a practical point T. S. Patterson J . 1927 1717. 3 10 QUARTERLY REVIEWS to observe mutarotation of the substances in solution at temperatures as low as - 35" they present the evidence of solubility changes as indicating optical activation of a tetrahedral mercury cation by the quinine and assume that the mutarotations accompanying the transformations were too quick t o measure. Zinc and cadmium salts were prepared which showed similar behaviour. The oxime of cyclohexanone-4-carboxylic acid (V) was prepared by Mills and Bain lo in order to investigate the stereochemical configuration of the group \ /C=N-OH. When they attempted its resolution with quinine in 30 parts of water a salt quinine I-acid 24H,O crystallised in 80% @kid of the total quantity present ; the mother liquor instead of containing the diastereoisomeride as would have been H CHZ-CH OH expected in resolution was inactive and decomposition of the salt with sodium hydroxide yielded a sodium / \ / salt [MI - 91" (morphine effected a dextroasymmetric transformation similarly giving a salt from hot ethyl alcohol which could be decomposed by aqueous ammonia to give dextrorotatory ammonium salts).The authors ex- plained their results as being due to the configurational instability of the oximino-group which is involved in the dissymmetry of the whole molecule the aqueous solution contains the partial racemate and as the less soluble quinine d-acid salt is removed by crystallisation equilibrium is rapidly re-established by rncemisation of the quinine Z-acid salt remaining in solution.Similar behaviour was encountered in the crystallisation of the quinine and morphine salts of the N - benzoylphenylhydrazone of cycZohexanone-4- carboxylic acid ; lC with a mixture of methyl alcohol and water as a solvent the I-quinine d-acid salt crystallised first and was converted into the sodium potassium or ammonium d-acid salt. The authors stated that this crystal- lisation was an asymmetric transformation and not a resolution although the percentage of the salt crystallising was not recorded in support they did not obtain any of the I-base. Z-acid salt which makes resolution appear unlikely. I n the same way the semicarbazone of the parent acid was acti- vated by crystallising its morphine salt from aqueous methyl alcohol and converted into a dextrorotatory ammonium salt.\ / C=N \ / C HO2C CH2-CHz (V.) 0 :'; (> &ie A strychnine salt of lO-p-carboxyphenyl-2- rnethylphenoxstibine (VI) having [XI= - 18" in chloroform was converted almost completely by boiling with alcohol for 30 minutes into a salt with [.ID + 17". This was regarded by the author l3 as a case of second-order asymmetric transformation. 0 CO,H 2-o-Carboxybenzylindan-1-one (VII) and brucine w.1 TURNER AND HARRIS ASYMMETRIU TRANSFORMATION ETC. 311 crystallise from acetone t o give a 94% yield of a single salt the acid part of the salt in its ketonic form shows molecular dissymmetry and decom position of this salt with sulphuric acid gave a dextrorotatory acid [a]2,O0 + 64" which mutarotated in chloroform solution.Leucha and Wutke,' who carried out this work gave the explanation that the optically active forms of the ketonic acid were interconvertible through the inactive enolic form thus providing a mechanism for asymmetric transformation by the agency of the brucine Fraction. B . l-Acid + B . enolic Acid + B . d-Acid + Crystals Weight (g.). aD ( I = 0.5). CH I C0,H \ / C0,H co \ OH d- or Z-Form Inactive enolic form (VII.) By the use of another compound owing its optical instability to keto-enol tautomerism hydrocarbostyril-3-carboxylic acid (VIII) Leuchs 20 has provided a very clear example of a second-order asymmetric transformation. The accompanying table shows the weights of three crystalline fractions (y>& C@,j-COzH obtained from 2.4 g.of hydrocarbostyril-3- carboxylic acid and 4.07 g. of anhydrous quinidine in 40 C.C. of methyl alcohol. The dihydrated salt crystallised and was shown to be one form only by removing the quinidine in hydrochloric acid at - 10" and watching the mutarotation of the residual acid in glacial acetic acid solu- tion a t 18" ; the rotations of the separate preparations of acid are given (VIII.) 1 2 3 4 1-6 0-7 + 1.08" + 1.09 + 1-04 in the third column extrapolated to the time of wetting with solvent. The total weight of salt is seen to be 6.2 8. out of a CH,.CH possible 6.9 g. I W. C. Ashley and R. L. Shiner 21 found that Ph-SO2-C-CO2H a-phenylsulphonylbutyric acid (IX) underwent almost theoretical asymmetric tramformation of the second order under the influence of brucine in acetone solution to give the brucine Z-acid salt.1 H (1x4 10 Ber. 1921 54 830. 21 J . Amer. Chern. SOC. 1932 64 4410. 312 QUARTERLY REVIEWS Decomposition gave the I-acid a tautomeric mechanism being envisaged for the optical inversion. An 83% yield of optically pure brucine I-benzylmalonoanilic acid (X) separated from an alcoholic solution of brucine and the dl-acid a second-order asymmetric transformation which the discoverers E. M. Davidson and E. E. Turner,22 found to be accelerated by heating. (One of the most impressive things about a crystallisation which is taking place with asymmetric transformation is that the beaker of filtered solution can be left in a warm place rather than in a cold to accelerate deposition.) Cinchonidine in chloro- .,H form solution converted the dI-acid in 90% yield into the optically pure base.d-acid salt.\ First-order asymmetric transformations were R not observed in spite of very careful searching. The analogous benzylmalono-o-toluidic acid was activated similarly by cinchonidine from acetone solution to give an 83% yield of the base.d-acid salt. This was amply proved to be a second-order transformation by removal of the base to give the d-acid which had very considerable optical stability in formic acid solution and did not racemise a t a measurable rate in cold ethyl-alcoholic solution. OH I \ / / \ Ph*CH C=O C H CO-N (X. ) The papers of A. McKenzie and his school contain much interesting experimental work which is valuable in a study of fht-order asymmetric transformation that classified as " asymmetric catalytic racemisation " in particular.The I-menthyl esters of various acids &I-phenylbromoacetic dl-phenylchloroacetic and dl-mandelic when dissolved in ethyl alcohol and given a very small concentration of alcoholic potash undergo mutarotation.* Evidently the ethoxide ion acts as a catalyst for the inversion of the acid part of the ester and thereby provides a mechanism for the optical activation or first-order asymmetric transformation under the influence of the I-menthyl residue. In these cases the transforming agent and the labile group are in chemical combination. McKenzie and Smith for example studied in detail the changes in rotation undergone by I-menthyl phenylchloroacetate starting from the I-menthyl d-acid I-menthyl I-acid and I-menthyl dl-acid esters they 'came to the conclusion that " the velocity of the catalysis is greater with the I-menthyl d-phenylchloroacetate than with its diastereo- isomeride " by making a calculation based on the measurement of the per- centage of the original ester left after a certain time of mutarotation.This calculation assumes that the system is moving towards the partial racemate whereas in fact the reactions in which the speed of the two diastereoisomerides differ is partial inversion. The rate coastant for approach to equilibrium from either diastereoisorneride is the same being the sum of two rate con- stants of partial inversion which are different. McKenzie and Smith's figures for the mutarotations although they were measured without tempera- ture control show fairly good agreement when used to calculate the rate of sa J.1945 843. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ,ETC. 313 approach to equilibrium from either side. The equilibrium composition calculated on strict proportionality is 57 yo of I-menthyl Z-phenylchloro- acetate and 43 yo of I-menthyl d-phenylchloroacetate. 23 Ph*CHCI*CO,C,,H, + Ph*CHC1*CO,C1oHls i 50% 50 % mixture I + - Ph*CHCl~CO,C,,H, + Ph~CHCl~CO,Cl,Hl { 57% 43 yo I Ph *cHCl*CO 2CloH, f '{ Ph~HCl*CO,C,H, 100 yo 100 yo It was not possible to saponify the equilibrium mixture of esters in order to prove that first-order transformation was the cause of the mutarotations because of their sensitivity to alkali and also because they are saponified at unequal rates and on this account also would lose their equilibrium com- position during the reaction. As a result of their experience of many crystallisations carried out in the resolution of I-menthyl dl-phenylchloroacetate using rectified spirit as the solvent McKenzie and Smith found that the I-menthyl d-acid ester was the less soluble diastereoisomeride.This is interesting to note in con- nection with the direction of mutarotation towards an excess of the methyl Z-acid ester in the same solvent. I. A. Smith 24 has shown that similar mutarotation phenomena can be observed with amygdalin where the transforming agent is the gentiobiose residue. It is interesting to correlate the first- and second-order transformations with mutarotation and crystallisation behaviour in the sugar series. Most of the work to be discussed here is long established as describing the proper- ties of the various sugars but was not accepted by stereochemists as of general application probably because i t is not possible to remove the " activating agent " from a sugar molecule-it is the whole molecule apart from the labile group-and therefore to prove optical activation in any one case.Thus an investigation incorporating many of the features described under restricted rotation compounds was carried out by Hudson and his collaborators 25 26 for several sugars but owing to the strict formulation of the sugars and their mechanism of partial inversion being in doubt the results could not take a pioneer place in a study of optically labile compounds. d-Glucose has been the subject of the most extensive experiments pre- sumably because i t crystallises well and is not difficult to obtain. Its composition in aqueous solution at equilibrium has for a long time been 23 P.D. Ritchie " Asymmetric Synthesis and Asymmetric Induction " 1933 p. 83. 24 Ber. 1931 64 1115. 26 C. S. Hudson and L. K. Yanovsky J . Amer. Chem. SOC. 1917 39 1013. C. S. Hudson and J. K. Dale ibid. p. 320. 314 QUARTERLY REVIEWS calculated from the rotations of the 'go* # 'q0g a- and ,&forms (XI) and of the equili- brated mixture a calculation which w m n w seems to be justified as any other a-d-Glucose @-d-G1ucose form must be present in negligible quantity z 7 9 25 (so long as the sugars are in the ring form the diastereo- isomerides are " real " the carbon atom marked is the unstable centre of asymmetry and the configuration of all the other carbon atoms is fixed). on o n o n (XI.) Some of the latest published figures 279 for the rotations are a-d-Glucose.Equilibrium. 8-d-Glucose. [a]z233 + 110.0" + 52.56" + 19.7" at 20" the equilibrium composition of 64% /?- and 36% a-form being unaltered between 0 O and 40 O. 28 Crystallisation from cold water invariably produces the a-form as the mon~hydrate,~~ but if the operation is carried out between 35" and 40" the anhydrous a-d-glucose crystallises. The a-form therefore crystallises out from a solution containing excess of the @-form. [In order to obtain the /?-form C. Tanret left the a-form for some hours a t 105" R. L. Whistler and B. F. Buchanan 3O obtained it by evaporating an 85% solution containing 50 g. of glucose in a vacuum a t 100" to a solid mass of crystals. " Second-order transformations " to a- or to ,!?-&glucose can be obtained in various other ways ; Hudson and Dale,269 31 for example found that if an aqueous acetic acid solution was allowed to crystalhe slowly in the cold 75-80y0 pure anhydrous a-glucose was produced while a hot quick crystallisation resulted in 93% of p-glucose.] Hudson and Dale 2s 31 recognised that the measured velocity constant k for approach to equilibrium from a- or from @-glucose was the same and that it was the sum of two constants k and kp.Certainly it would seem that a-glucose is the less soluble of the two forms although they are both too soluble in water for accurate measurement to be made. The solubilities could however be measured in 80% ethyl alcohol,25 100 C.C. of which dissolved 4.9 g. of the p-form and 2.0 g. of the a-form the a-form is the one crystallising as the hydrate while the rotation values show that the /?-form is in excess at equilibrium a-glucose + 115.5" [or]ioo B-glucose + 20.30} equilibrium + 59.3' In absolute methyl alcohol Andrews and Worley27 find that there is excess of the /?-compound at equilibrium [.I:" or -glucose /?-glucose 'ii640} equilibrium + 76.8" 27 J.C. hdrews and F. P. Worley J. Phyeical Chem. 1927 1880 ; J. C. Kendrew 28 E. A. Moelwyn Hughes " Kinetics of Reactions in Solution " 1933 edn. p. 46. 30 J . Bwl. Chern. 1938,125 667. and E. A. Moelwyn Hughes Proc. Roy. SOC. 1940 A 176 353. C. Tanret C m p t . rend. 1895 120 1061. s* J. 1904,85 1661. TURNER AND HARRIS ASYlKMETBIU TRANSFORMATION ETC. 316 T. M. Lowry 31 found the a-form crystallising from methyl alcohol and showed that the solubilities in this solvent were small enough not to interfere with the relationship Sm - fla K (the equilibrium constant) = 5 = - kP s a which was also propounded by Hudson and Dale where Sa is the initial solubility of the a-form and This relation- ship was used to calculate the rotations of missing b-compounds.d-Mannose was sufficiently insoluble in water for Hudson and Yanovsky Z s to use the solubility-rotation relationship to calculate the rotation of the then unknown a-mannose ; they obtained the value + 30" knowing that for /?-d-mannose to be - 17" and that for the equilibrated solution to be + 14.6". P. A. Levene six years late1-,~2 confirmed their prediction on isolating a-mannose he records that mannose crystallises in the a-form under conditions in which glucose appears in the /?-form and vice versa a fact which we should link with the equilibrated d-mannose solution contain- ing excess of the a-form while glucose contains excess of the #?-form.Hudson and Yanovsky also predicted the rotation of a-mannose in 80% ethyl alcohol to be + 35" which Levene confirmed later. Taken together with a value [ C ~ ] ~ ~ O " - 14.9" for the p-form and + 25.7" for the equilibrium this means an excess of a-form a t equilibrium in a solution under which the #?-form is stable. Similar relationships hold for the a- and /?-forms of lactose and galactose the hydrated a-forms crystallising from aqueous solutions in which the /?-forms are present in excess. 339 25 the solubility at equilibrium. PhSO CH,CO,H \ / t6 N-Benzenesulphonyl-8-nitro- 1 -naphthylglycine (XII) which owes its optical activity to restric- tion of rotation of the substituted amino-group by the nitro-group was shown by Mills and Elliott to undergo second-order asymmetric transformation with brucine in either direction according to the solvent used.It was the first acid found to show the two transformations in this way decomposition of each of the diastereoisomeric salts gave an active acid (XII.) acetone I MeOH B . d-A,SH,O .1 I I acetone B. l-A,H,O {ti;?:} - 98% yield 1% solution * 75% yield + Z-acid .(I d-acid The effect although not the process was as if a resolution had been performed. 8a J . Biol. Chem. 1932 329. ss C. Tanret Bull. SOC. chirn. 1871 16 195. 316 QlJABTEB&Y BEVlEWS From the point of view of the subject of this article the most important work which these authors carried out on this acid wm to prove their inter- pretation of the mutarotation of the brucine dZ-acid salt - an optical activa- tion (or first-order asymmetric transformation).This they did as follows 0.183 g. of the d-acid was dissolved in 25 C.C. of chloroform and 0.221 g. (1.18 mols.) of brucine in a further 25 C.C. of chloroform and the two solutions were mixed. The initial aSlal observed immediately changed from - 0.78" to - 0.22" (I = 4 ; T = 0.7-1*5") as the Z-base.d-acid + Z-base.Z-acid equilibrium established itself with the former in excess. A similar solution but containing 0.211 g. of brucine only after being left for 3 hours was extracted with ice-cold dilute sulphuric acid the brucine being thus removed. The remaining solution (to which a little acetone had to be added to keep the acid in solution) had an unmistakable dextrorotation which mutarotated almost to zero at 1.2".Thia dextrorotation could only be due to the acid which had been activated in solution by the brucine. At a later date other workers 34 were attracted to this acid and prepared the cinchonjdine Z-salt which mutarotated in chloroform from - 255.5" to - 87-3" and the cinchonidine dl-salt which mutarotated from - 35.5" to - 87.3" ; this means a composition at equilibrium of 62% Z-base. Z-acid and 38% Z-base. d-acid neglecting dissociation or the possi- bility of a change of specific rotation over the concentrations involved in the calculation. The 8-benzenesulphonethylamido-1 -ethylquinolinium ion (XIII) is structurally very like the substituted glycine which has just been con- sidered and its range of optical stability is such that it can be made to perform the crystallisations PhSo2 Et associated with second-order asymmetric tr4nsforma- N Et tions.W. H. Mills and J. G. Breckenridge35 found that 8- benzenesulphonethylamido- 1 -ethylquinolinium d-or-bromocamphor-n-sulphonate crystallised as the d-base.d-acid,2H20 salt from a mixture of ethyl acetate and acetone. This salt could be converted into the d-quinolinium iodide by shaking the chloroform solution with aqueous potassium iodide and it muta- rotated in the laevo-direction in water chloroform and ethyl alcohol. No firsborder asymmetric transformation was detectable with the bromo- camphorsulphonate and from the rotations 0% the two diastereoisomerides and that of their equilibrated solution it would appear that the equilibrium solution has the composition of the partial racemate.We should now explain this difference in behaviour from the alkaloidal salts of N-benzene- sulphonyl-8-nitro-l-naphthylglycine as being due to the fact that in the latter case the diastereoisomerides are " real " in non-dissociating solvents while the quinolinium salts must be dissociated into ions even in chloroform solution. Such diastereoisomerides are "real " in this sense only on cry stallisation. \ / (XIII.) (5 34 M. M. Jamison and E. E. Turner J . 1940 264. 36 J . 1932 2209. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC. 317 Some members of the dinaphthyl series bear a certain skeletal resem- blance to these compounds. Meisenheimer and Beisswenger 150 found that when ethyl hydrogen 1 1'-dinaphthyl-8 S'-dicarboxylate (XIV) waa crystalked with brucine from ethyl acetate containing a little methyl alcohol the brucine Z-acid,3H20 salt appeared in almost 100% yield the Z-acid could be obtained by decomposing the salt with dilute mineral acid.A similar acid (XV) lacking only the carbethoxyl group formed a mono- hydrate with brucine crystallising from ethyl acetate solution as base. d-acid or base. Z-acid on inoculation with the appropriate crystal. Meisenheimer Theilacker and Beisswenger l5* again describe activation by alkaloids of the ,8-oxime of 2-hydroxy-3-carboxy-l-naphthyl methyl ketone (XVI). The diphenyl nucleus has formed an obvious framework for investigating the effective sizes of groups by stability of potentially active &OH CO,H (XVI.) observing their influences on the optical structures.There are therefore aeveral NO Me0 / \ '\ D-0 \ OMe CO,H (XVII.) examples of what must be asymmetric transformations to be found in reading accounts of such work. Brucine dZ-2-nitro-2' 5'-dimethoxy- diphenyl-6-carboxylate (XVII) dissolved in water crystalliaed in three fractions 90% of the total weight all with the same specific rotation ; on decomposition with ice-cold hydrochloric acid they all yielded the t - a ~ i d . ~ ~ The base. Z-acid salt prepared by this second-order transformation mutaro- tated in chloroform from [aID - 167" to + 3-2" in 100 minutes extra- polation of the recorded readings to zero time (time of wetting salt with chloroform) gave -180" as the proper initial value of [aID. A solution of brucine and the dZ-acid in chloroform had an initial [.ID - 8.6" changing to + 3.3" in 80 minutes.This latter mutarotation has all the appearance a6 H. C. Yuan and R. Adame J . Arne?-. Chem. Soc. 1932 64 2966. 318 QUARTERLY RBVIEWS of a first-order transformation but Yuan and Adam after performing a precipitation experiment on the equilibrated solution conclude that it is not. Examination of their figures l4 shows that if the mutarotation is due to first-order transformation the equilibrium composition by the simple calculation is 63.5% I-base . d-acid and 4605% I-bwe. Z-acid ;. precipitation of such a solution in chloroform with light petroleum which would not be quantitative might well give a product which was indistinguishable from the partially racemic mixture. The same acid underwent second-order transformation with cinchonidine also.The whole series (XVIII) of 2-nitro-2'-methox~diphenyl-6-carboxylic acids with methyl chlorine bromine and nitro-groups in the 6 position have been shown by the same ar %ora3' to undergo what are clearly second-order NO OMe / \ \ \ 0 - C 3 CH3 CO,H c1 Br NO (XVIII.) asymmetric transformation with brucine from alcohol containing varying quantities of water. The authors who were interested in obtaining speci- mens of optically active acids for another purpose describe these crystalli- sations as resolutions it seems a pity to use this term which is best reserved for the separation of a racemic mixture into its stereoisomeric forms to imply conversion of it all into one of them. Brucine or quinine brings 2'-fluoro-2-nitro-5'-methyldiphenyl-6-carboxylic acid (XIX) out of ethyl- alcoholic solution as the base.d-acid salt.38 If the fluorine atom is replaced by chlorine or bromine the optical stability is so raised that the crystallisct- tion process with brucine from the same solvent is resolution the rotations of the crops increasing from negative to positive in the order in which they are deposited. The following evidence may be interpreted as showing that a first-order transformation takes place with the fluoro-acid in chloroform by the agency of brucine. The first crop in the cryatallisation of 2.75 g. of the acid and 3.94 g. of brucine weighed 5.1 g. and was identified as the salt I-B . d-A,@,O. The rotation in chloroform ( [a]go) was - 32" but if the solution was made up a t 0" [a]r was + 13" when first observed and mutarotated to - 3*4O.This may of course be due to a large temperature coefficient of rotation but if it is not then it would seem that + 13" is nearer to the rotation of the base. d-acid salt while - 3-4" represents an equilibrium compoeition which is unlikely to be that of the racemic mixture. s7 H. C. Yuan and R. Adams J . Amer. Chem. SOC. 1932 64 4434. a8 R. W. Stoughton and R. Adam a i d . p. 4426. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC. 319 The 2-acid could be obtained 39 from each of three fractions crystallised from 95% ethyl alcohol of dibrucine 2 2’-di-iododiphenyl-6 6’-dicarboxylic acid (XX) which weighed together 83% of the possible total of salt. When a similar crystallisation was carried out using methyl alcohol as solvent both diastereoisorneric I CO,H forms crystallised out but not as an intimate mixture.They formed discrete crystals which could be separated by hand-picking. R. Adams and N. Kornblum 40 found two cases of what now appears to be second-order asymmetric having the 5 5’-positions joined by ether link- ages to a hydrocarbon chain. When n is 10 brucine in methyl alcohol gives a 77% yield of a dibrucine salt in one fraction and a further quantity from the d-‘b \ / (XX.1 C0,H I transformati’on in diphenyl compounds (XXI) mother liquor all of which and removal of the brucine. c1 CH3 I c1 CH*CO,H OCH /V\ CH (XXII.) (XXI.) yielded dextrorotatory acid on decomposition When n is 8 cinchonine in ethyl alcohol pro- duces the Z-acid salt in 3 fractions totalling 91 % of the theoretical quantity as proved by preparation of the Z-acid from it.The sub- stituted benzene derivative of R. Adams and J. Gross 41 deposited a quinine salt from ethyl acetate in a series of fractions all of which had the same specific rotation and were decomposed to yield d-p-chloro-p- (5 - chloro- 2-methoxy-4 6-dimethylphenyl)acrylic acid R. Kuhn and 0. Albrecht 42 claimed an (XXII) . asymmetric transformation of 4 4’-dinitrodiphenic acid (XXIII) by quinine on crystallisation from 96% ethyl alcohol although removal of the base from the deposited salt gave an acid in which they were unable to detect optical to support their conclusion was as follows. The deposited crystals all the same substance represent 80% of the theo- retical yield-1st crop m.p. 207-208” [alga + 108.4” in chloroform ; 2nd crop C0,H activity.The evidence which they used / m<3)--<->NO / (XXIII.) CO,H ae N. E. Searle and R. Adrtms &bid. 1933 55 1649. 4 1 Ibid. 1942 64 1786. 40 Ibid. 1941 63 188. 48 Annalen 1927,486,272. 320 QUARTIDLY REVIEWS m.p. 207-208' [a]:" + 110.3' in chloroform. Secondly the acid is one of a series with two one and no nitroxyl in the 6 6'-positione in the diphenic acid 6 6'-dinitrodiphenic acid is resolvable with bruche and is optically stable ; 4 6-dinitrodiphenic acid is resolvable with quinine and shows racemisation a t an observable rate ; 4 4'-dinitrodiphenic acid shows (sup- posed) asymmetric transformation with quinine and the acid is too unstable to be active. Thirdly quinine m-nitrobenzoate and quinine phthalate have the specific rotations - 163.5' and - 168.2" in the same circumstances rotations very different from the salt under investigation and of the opposite sign.This last piece of evidence largely owing to the work of M. S. Kharasch J. K. Senior D. W. Stanger and 5. A. Chenicek 43 on the anomalous rotation of quinine salts has been shown not to afford the support it appeared to a t first. A. Corbellini and A. Angeletti reported in 1932 44 that 2'-(a-hydroxyiso- propyl)diphenyl-2-carboxylic acid (XXIV) crystallised as the brucine salt from ethyl alcohol in the Zawo-form in 83% yield. Jamison and Turner,' who were looking for a representative optically unstable diphenyl com- pound which cotdd be prepared relatively easily raised this figure to 97.6% and found also that a second-order transformation could be effected by evaporating a chloroform solution of the brucine dl-acid salt to dryness with stirring on a boiling water-bath.This salt yielded the laxorotatory acid on removal of the brucine by means of dilute acid. Thus second-order transformation takes place in the lmo-direction from chloroform at the boiling point and first-order transformation in the same solvent a t 25-15' takes place in the opposite direction. The brucine salt of the dZ-acid (i.e. a mixture of brucine and the dl-acid in equimolecular proportions) in chloroform mutarotates from [a]:::/' - 5-08" to + 1.90'. The brucine Z-acid salt (obtained by second-order asymmetric transformation from ethyl alcohol solution) mutarotates from - 47.04" to + 1-46". The velocity constants for these mutarotations k ~ ~ ; ~ ~ h o m s - 1 were respectively 0.0280 and 0.0277 the agreement being taken to show that the same process is being observed in each case.Assuming no dissociation the equilibrium composition calculated from these figures is 68% of the &-acid salt and 42% of the E-acid salt. Unless the equilibrium composition vanes sensibly between room temperature and the boiling point of chloroform it appears that the base.d-acid is more stable in solution while the base.Z-acid has the greater tendency to come out of solution. CMe,-OH -- \ \ C0,H (XXIV. ) N-Benzoyl-4 6 4'-tribromodiphenylamine-2-carboxylic acid (XXV) a member of a useful series showing optical activity due to restricted rota- tion was the first of its kind to be submitted to a thorough stereochemical See also M. S. Lesslie and E. E. Turner 43 J. Amer. Chem. Soc.1934 56 1646. 44 Atti R. A d . Lincei 1932,16 968. J . 1934 347 ; M. S. Lesslie E. E. Turner and E. R. Winton J. 1941 257. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION HTO. 321 investigation. With cinchonidine (1 mol.) in acetone solution it can be made to show first- and second-order transformation and resolution by appropriate choice- of conditions. In the second- order transformation which can be accelerated by heating theoretical the quantity crystals and deposited consist are of the 94% optically of the &(yQBr pure cinchonidine d-salt. The &-acid can be obtained from this by treatment with pyridine followed by mutarotation of this salt was measured a t several and E calculated which showed that k (k = BvE/RT) might be small enough for resolution to be possible at - 15".This was put to the test by dissolving dI-acid and cinchonidine in warm acetone and chilling to - 15" as soon as crystallisation began. The deposition of crystals instead of continuing until all was out of solution as in the second-order transformation stopped when almost exactly 50% of the total weight had come down. The salt deposited was I-base.d-acid while the mother liquor on cold evaporation under reduced pressure showed a rotation indicating that it contained two thirds of the I-base I-acid salt. &I - N - Benzoyl - 2'-chloro-2-methyldiphenylamine-6- carboxylic acid (XXVI) is converted by crystallisation as the brucine salt from a mixture of ethyl alcohol and ether into the I-form. The I-acid was obtained free from brucine on decomposition of the salt by dissolving it in formic acid and stirring with ice- cold dilute hydrochloric acid.With varying subs ti tuent s the N - benzo yldi phen ylamine- 6 - car box y li c acids provided material for many more first-order asymmetric transforma- tions. N - Benzo yl- 2 -met hyldip henylamine - 6 -car box y lic acid ( XXVII ) in chloroform containing 2.5% of ethyl alcohol by volume underwent muta- rotation when 1 mol. of d-nor-#-ephedrine was present. The related acid dilute hydrochloric acid.g The velocity constant of I temperatures the value of the Arrhenius constants B COPh (XXV.) "' COaH c1 (XXVI.) I COPh 0 cHao COPh COPh COPh (XXrn.) (XXVIII.) (DX.) substituted in the 2 2'-positions by methyl groups (XXVIII) mutarotated with cinchonidine in the same solvent and provided proof that the mutaro- tationa were not due to slowness of salt formation in this way-the originally laevorofatory cinchonidine solution became immediately more laevorotatory 322 QUARTERLY REVIEWS on addition of the acid and then mutarotated in the dextro-direction.N-Benzoyl-2 4-dichlorodiphenylamine-6-carboxylic acid (XXIX) showed mutarotation with d-nor-+-ephedrine in chloroform and with cinchonidine in chloroform containing 2.5% of ethyl alcohol by volume. The assumed optical activation was proved by extracting the equilibrated solution with mineral acid leaving a dextrorotatory acid in the chloroform solution. N - Benzoyl - 2' - chloro- 2 - methyldiphenylamine - 6 - carboxyljc acid (XXVI) showed mutarotation with quinidine and brucine in the same chloroform- alcohol solvent. Base.Z-acid . . .Ba8e.d-acid . . . Base.dZ-acid. . . First-order transformation was observed with cinchonidine in chloro- form solution in the lzevo-direction. It was possible to observe the approach to equilibrium from all three starting points base. d-acid base. Z-acid (the optically impure mixture from the mother liquor in the - 15" experiment) and base. dZ-acid - 105" + 194 (extrap.) - 40.4 Starting material. 18.0' Ia15161 (initial)' 18.0" [a15461 (anal). - 44.5" - 44.5 - 44.5 17.7" %og, min.-l. 0.0200 0.0206 (range too small for measurement ) The measured velocity constant k is the sum of the two velocities of inversion kd and 4 of the diastereoisomerides neglecting the possibility of dis- sociation in solution kd Jcl concentration I-B . I-A at equilibrium concentration I-B. d-A at equilibrium _ - - whence kd = 0.0105 and kl = 0.0101 ; the difference is very small but there is no doubt that it is real.First-order asymmetric transformation of this acid was also observed in the dextro-direction with d-nor-+-ephedrine in chloroform. Investigation of the First-order Transformation Equilibria When it was first observed that the rotation of an equilibrated solution containing equivalent quantities of acid and base the acid being optically unstable and the base optically stable was changed by adding an excess of the dZ-acid the authors' immediate thought was to attribute the effect to suppression of dissociation of the salt. But this explanation was quickly dispr~ved,~? 34 and so far no satisfactory one has been put in its place. The added acid may enhance or diminish the existing rotation in different cases and the effect has been used to explore realms of optical instability which were hitherto unattainable.N-Benzoyl-2'- chloro-2-methyldiphenylamine-6-carboxylic acid (XXX) and quinidine in chloroform containing 2.5% of ethyl alcohol by volume behave as follows : TURNER AND HARRIS ASYMMETRIC TRANSFORMATIOX ETC. 323 of + 4-8" I = 2. On the addition of 1 equivalent of the dl-acid this rotation changed immediately to + 4.35' and then mutarotated to + 2.99" ; 2 equivalents of 0.1620g. of quinidine in 20 C.C. of solvent showed a rotation COPh and quinidine in CHCl,-E t,OH (XXX.) acid caused an immediate change to + 4-30' mutarotating to + 3-32' ; with 3 equivalents + 4.32" changed to + 3.67'. All these solutions on decomposition with mineral acid afforded the lzvorotatory acid equilibrium was attained more quickly the greater the excess of acid.Another example selected from many showing this type of behaviour is N-benzoyl-2 4-dichlorodiphenylamine-6-carboxylic acid (XXXI) which with d-nor- +-ephedrine in chloroform mutarotates towards the value of the rotation of the base instead of away from it. In a case such as this the I I CO2H COPh and d-nor-y-ephedrine in CHCl (XXXI.) 10 difleerence between the curves of initial and final readings is considered to be due to optical activation the departurg of the initial curve from the vertical shows that there are other reasons (such as increased concentration) for a static change in rotation present also. (The initial curves described a m obtained by extrclpolating the observed mutarotations back to zero time.) 324 QUARTERLY REVIEWS The whole subject would have been much less intriguing had not some of the initial and final curves crossed over that is to say a t certain acid base ratios first-order transformation was in the Zamo-direction and at other ratios in the hxtro-direction.This was a new phenomenon and as it is well substantiated a lthough not explained worthy of further quotation. N-Benzoyl-2 4-dichlorodiphenylamine-6 -carboxylic acid (XXXII) with and cinchonidine in chloroform-eth yl alcohol (XXXII.) cinchonidine in chloroform-ethyl alcohol mutarotated in the dextro-direction at the 1 1 ratio and in the hvo-direction at the 3 1 ratio ; the equilibrated solutions on decomposition yielded d- and I-acid (not optically pure) respec- tively.The experiments do not of course indicate whether the activated acid is free or combined as salt. Very similar results were obtained with (XXXIII) and (XXXIV). and d-nor-y-ephedrine in chloroform (XXXIII.) c A further point of interest is added in another case in which decomposition of the 1 1 and 2 1 solutions gave laevorotatory acid the 3 1 solution inactive acid and the 4 1 dextrorotatory acid-the curves for N-benzene- sulphonyl-8-nitro-1-mphthylglycine (XZurV) and cinchonidine in chloro- TURNER AND ~ m s ASM~METRIC TRANSFORMATION ETC. 325 and cinchonidine in chloroform-ethyl alcohol (XXXIV.) form-ethyl alcohol show how this comes about. This means then that without the separation of a salt both &- and Z-acid could be obtained-not optically pure but distinctly active through optical activation in the same solvent and by the same alkaloid.PhSOz CHa COsH N NOa I I \ / and cinchonidine in chloroform-e t hyl alcohol (XXXV.) 6' The plotting of a series of these " addition curves " served to demon- strate the potential optical activity of a series of acids which were too un- stable for the observation of it at ordinary temperatures. The blocking which causes dissymmetry in these acids is very slight so that they will tolerate neither resolution nor observable first-order transformation under normal conditions. CO,H CO,H CO,H CO,H O,:c> I ciO$O c1 I cH30,)0 CH COPh I 0,B' COPh I COPh COPh A B C D 326 QUARTERLY REVIEWS The curves obtained by addition to d-nor-#-ephedrine in chloroform- they are " final " curves the " initial " ones presumably being too ephemeral for observation-are shown in the diagram.The salts of acids A B and C have equilibrium *rotations which are highly sensitive to excess of the acid (0.1051 g . d-nor-$-ephedrine in 14.5 C . C . CHCI,; 1 = 2.) the curve for acid D which has an effectively symmetrical molecule shows it to be in a different class. Experiments a t - 31' justified this distinction 4 equivalents of acid C and one of d-nor-#-ephedrine in chloroform solution showed a muta- rotation a5461 changing from - 4-03" to + 2-15" half-life period 2.4 minutes. Acid B mutarotated more quickly in the same circumstances acid A showed no such muta- rotation and its claim to optical activity until a lower-temperature technique is developed rests on the curve in the above diagram in con- junction with those of the other acids in the series.In a review of these excess acid phenomena W. H. Mills 4 5 said that they might " affect the diagnostic value of the activation process ". But as has already been pointed out l4 there is no indication that proper use of the method would lead to fortuitous results while it has materially extended the field of investigation of labile optically active compounds. Asymmetric Induction The term " asymmetric induction " was introduced by E. Erlenmeyer Jun. in 1912 in explanation of his alleged silccesses in " inducing " optical activity in various unsaturated compounds the molecules of which were not dissymmetric on classical theory. He claimed to have induced optical activity in such substances as benzaldehyde and cinnamic acid by heating them with tartaric acid either in presence or in absence of a solvent.E. Wedekind,46 and L. Ebert and G. Kortum 47 were unable to confirm Erlenmeyer's results the references to which are given in the Obituary Notice 48 to Erlenmeyer. Although Erlenmeyer had attempted the induction of activity in mole- cules which to us clearly could not be dissymmetric other workers were concerned with a different matter the more legitimate inquiry as to the possibility of inducing optical resolution of racemates not by the standard methods but by differential solvent action say of a dextrorotatory solvent on the two enantiomeric forms of a second substance. This was examined 45 Presidential Address J. 1943 194. 47 Ibid. 1931 84 342. (* Ber. 1914 47 3172. 48 Ibid. 1921 54 107. TURNER AND HARRIS ASYMMETRIC TRANSFORMATION ETC.327 in two ways (1) by determining the solubility of the two enantiomeric forms separately in an active solvent and (2) by crystallising or extracting the racemate by means of an active solvent. In spite of much careful work no differentiation of the kind sought was f0und.~3 d9 An interesting set of results was obtained by A. McKenzie and his co-workers,60 although owing to the rather complex mixtures used their significance cannot yet be properly assessed. Z-Malic acid (1 mol.) was added to an aqueous solution of potassium racemate (1 mol.) and a crop of crystals obtained consisting of potassium hydrogen racemate and potassium hydrogen d- tartrate similar results being obtained with sodium rubidium and czsium salts. No other active acid than malic produced the same kind of result and no acid that was examined other than racemic acid could be “activated ”.The reality of these observations cannot be doubted and a thorough phase-rule study of one of the systems would no doubt repay the effort. Generally speaking racemates cannot be even partially resolved by crystallisation from an optically active solvent. This is what might be expected unless one antipode crystallised with solvent of crystallisation. An example remains to be discovered in which association with an optically active solvent by hydrogen bonding for example is responsible for solu- bility differences in a pair of optical isomerides although it may well be that McKenzie’s case can be interpreted in this way. Some such loose association with preference for one isomeride must be responsible for cases of partial resolution by adsorption on optically active adsorbents.Another type of experiment to which the name asymmetric induction was attached was the attempted conversion of a symmetrical into a dis- symmetrical molecule in solution in an optically active solvent. As long ago as 1896 D. R. Boyd 51 reduced benzoylformic acid in an aqueous solution of tartaric acid and four years later F. S. Kipping 52 performed the benzoin synthesis in presence of d-camphor. In these and many subsequent investi- gations no activity was induced by the non-reacting asymmetric material which had been added. In 1932 G. Kortiim gave his interpretation of the meaning of the term “ asymmetric induction ” as follows the action of a force exerted by asymmetric molecules on molecules capable of changing from a symmetricrtl into an asymmetrical configuration.He further noted the division of the effect into inter- and intra-molecular types. The examples we have just dealt with are intermolecular and we now turn to the intramolecular one^. In 1936 A. McKenzie,53 commenting on Walden’s dismissal of the Erlenmeyer conception of asymmetric induction said “ Nevertheless whether the idea of asymmetric induction is right or wrong it has since proved itself of service in the study of asymmetric synthesis and to-day it ought not to be a t once dismissed as both useless and superfluous.” The 49 Kortum “ Samml. chem. und them.-tech. Vortrage ” Stuttgart 1932. so A. McKenzie J. 1915 10’4 440; A. McKenzie and N. Walker J. 1922 U1 340 ; A. McKenzie H. J.Plenderleith and N. Walker J. 1923 123 2876. 61 Inaug. Dissert. Heidelberg. 6 2 Proc. 1900 16 226. 63 Ergebn. Enzymforsch. 1936 5 49. 328 QUARTERLY REVIEWS view “ that in optically active esters of a-ketonic acids the carbonyl group in the a-position might assume a dissymetrical configuration under the influence of an optically active radical” enabled him to correlate the steric course of a long series of reactions between Z-menthyl benzoylformate or Z-menthyl pyruvate on the one hand and Grignard reagents on the other with the direction of mutarotation observed when the two esters mentioned were dissolved in ethyl alcohol. McKenzie suggested that these mutarofa- tions might be due to the establishment of an equilibrium of the type ( A ) (B) whilst in ethereal solution (in which the Grignard additions were carried out) equilibrium was established too quickly for observation but that neverthe- less the two above forms were present in unequal amounts this accounting for the success of the asymmetric synthesis and also for the sign of rotation of the resulting ct- hydroxy-acids all the benzoylformate reactions giving hvorotatory acids and all the pyruvate reactions dextrorotatory ones.It seems clear that it would also have been necessary to assume that the rate of addition of the Grignard reagent to the carbonyl group was greater than the rate of equilibration. Nevertheless the detailed experimental evidence deserves close scrutiny. M. M. Jamison and E. E. Turner,54 although their evidence did not justify a precise interpretation preferred to regard the mutarotations in the alcoholic solutions of the esters as due to first-order transformation between the diastereoisomeric hemiacetab formed by the very probable reversible combination of the esters with the solvent OH OEt EtOH I I I R-CO-CO *OC,oH, & R - C - C O *oC,,-,H1 + R-C-CO*OC&,~ OH I OEt (4 (4 (I) ( I ) ( I ) At the same time the absence of mutarotation in ether was ascribed to the lack of any real distinction between ( A ) and (B) the partial stereo-specificity of the many Grignard reagent syntheses then being attributed to first-order asymmetric transformation of optically unstable intermediates.The idea of asymmetric induction in the sense of a double bond made dissymmetric previous to approach of the reagent as the cause of an “asymmetric reaction ” was not accepted.Without apparently realising the mass of experimental material which McKenzie and his co-workers as well as others had accumulated in their studies of asymmetric synthesis and related matters T. M. Lowry and E. E. Walker 55 suggested “ that an unsaturated group in an asymmetric molecule e.g. the carbonyl group in camphor may acquire an induced asymmetry and thus itself become optically active ”. This conclusion which was re- considered by T. M. Lowry and J. 0. Cutter,56 was based on “ the fact that 64 J . 1941 538. titi Nature 1924 113 565. ti6 J . 1925 127 604. -NEB AND HAILBIS ASYMMETBXI ‘I?RANBFORHATION ETO. the &pemion-eqwLtione for txmphor and its denvativea are haunted by a low-frequency term the p e d of which is definitely cha;rcccterhtio of the ketonic group ”.Lowry and Cutter further said “ We therefore asaign this partial rotation to the ketonic group which is proved to be aspmetric by the unequal yields of two stereoisomeric (diastereoisomeric 2) products which am obtained when the double ia converted into two single bonds. This absence of symmetry in a double bond has already been proved in the camphor series by the unaymmetrical reduction of camphor to borneol and koborneol and of its oxime to bornylamine and neobornylamine. . . . Since the two links of a double bond in an asymmetric compound are clearly unequal fiom the chemical point of view it would be absurd to pretend that they must be equal from the physical point of view and no dditional justification need therefore be given for using this conception. in order to explain the optical properties of camphor or of .. certain other unsatur- ated compounds.” H. Phillips 57 saw in the then freshly discovered (but since abandoned) semi-polar sulphoxide bond a means of giving the optical activity of a carbonyl group a physical meaning; he pictured Z-b-octyl acetate as the equilibrium CtlH13 CH3 C6H13 CH3 C6H13 CH3 \ / \ / CH CH \ / CH I 0 I 0 2 t I 0 I @ G - O @ I CH I I CH3 o=c I @ O - C @ I CH3 T. M. Lowry and G. Owen,‘je followjng S. Sugden J. B. Reed and H. Wilkin~,~Q pointed out that a semi-polar bond with carbonyl would represent the activation limit of a polarisation and not the normal state of the group. They saw in such an activation the origin of the ultra-violet ketonic band shown by camphor. C. E. Wood and S. D. Nicholas,6o in a study of anomalous rotatory dispersion concluded that the carbonyl group “ need not be regarded as an asymmetric centre but rather as causing a deflecting and disturbing action on the electronic system round the asymmetric centre ”.Lowry’a view that the two bonds of a double bond in an asym- metric compound are unequal from the chemical point of view is untenable because it over-simplifies the picture of addition reactions. Part of the present problem is discussed by M. P. Balfe and J. Kenyon.61 The use of the term “ induced anisotropy ” instead of “ induced asymmetry ” is an advantage since it avoids the implication that the optical and the alleged chemical mechanisms are intimately related. As W. C. Price 62 has pointed out the n-molecular wave functions ara responsible for the production of the optical anisotropy ; they are also concerned with chemical (’ J .1925 127 2552. Go J . 1928 1671. 68 J. 1926 606. s9 J. 1926 127 1525. Ann. Reports 1942 39 125. 6 2 Ibid. 1939 36 52. 330 QUARTERLY REVIEWS addition reactions. In the optical sphere they function as part af the permanent state of the molecule ; in the chemical sphere for all we know to the contrary they play their normal part in permitting electronic activation of the double bond prior to its two-stage saturation. It seem8 probable that a t any rate a t the moment only confusion will result from correhting the chemical reactivity (“ asymmetric induction ”) of a carbonyl group with the rotatory dispersion effects (“ induced asymmetry ”1 associated with it. Until the two effects have been more closely investigated no useful conclusions can be drawn.In order that a fixed centre of asymmetry shall influence the steric course of an addition reaction at an unsaturated centre in the same molecule in an asymmetric synthesis there must be some stage a t which either stereo- selective addition occurs as an irreversible process or first-order asymmetric transformation takes place. There are a t present insufficient experimental data upon which t o base an analysis of even the simplest “ asymmetric reaction ” but some general lines of argument can be foreseen. Thus in the addition of XY to a carbonyl group of a molecule already containing a fixed centre of asymmetry (in group R) the first stage may be regarded as the approach of X - towards the positive end of the polarised carbonyl group R R la+ d- I - c......”....” X- c-0 ____3 x- ..............-0 I R’ I R’ The two tetrahedral arrangements represented by the plane diagrams R R c ____.- R’ and R’---- c are nossible before the addition of Yf. If the energy changes concerned in t i e formation of these two structures are equal there is no immediate asymmetric addition. If they are unequal (i.e. influenced by existing asymmetry) then we have asymmetric addition which appears to take place even in non-reversible asymmetric reactions of this type (e.g. Grignard reactions). On the other hand addition which is known to be chemically reversible (e.g. when X- is CN-) could be accompanied by &&-order asymmetric transformation of the newly forming molecule at this stage and it would be rash to say without further experiment whether the new asymmetry is introduced during or after the first addition or a t both stages. The illustrations in this article are reproduced by permission from Transactions of the Faraday Society 1945 41 pp. 696-717.

ISSN:0009-2681    

DOI:10.1039/QR9470100299

出版商:RSC    

年代:1947

数据来源: RSC

摘要:

CHEMISTRY OF THE METAL CARBONYLS By J. S. ANDERSON PH.D. F.A.C.I. (SENIOR LECTURER IN INORGANIC CHEMISTRY UNIVERSITY OF MELBOURNE) I. General.-The special interest attaching to the chemistry of the metal carbonyls arises from several causes. Whilst quite distinct from the organometallic compounds they differ in physical properties (e.g. their volatility) from all other compounds of the transition metals. Chemically they constitute a group of compounds in which the formal valency of the metal atoms is zero and in this respect (apart perhaps from the ammoniates of the alkali metals) they are comparable only with the recently discovered compounds K,[Ni(CN),] and K,[Pd(CN),] .2 Their constitution and that of their derivatives accordingly raises a number of problems for interpre- tation by valence theory.As a class the carbonyls are reactive compounds and a number of new types of inorganic compound has been discovered largely through the work of W. Hieber and his school. The various classes of derivative obtain- able from iron pentacarbonyl-which with cobalt and nickel carbonyls has hitherto been most closely studied-are indicated in Table I. TABLE I 5CO + Fe ( " Carbonyl iron ") Since the discovery of nickel carbonyl by Mond and Langer in 1888 the carbonyls of the iron group and of tungsten and molybdenum have found important technical applications-e.g. in the Mond nickel process and for the preparation of the metals in a state of subdivision and of purity suitable for powder metallurgy for catalysts etc. The reaction mechanism of the processes developed for producing the carbonyls technically has only recently received its interpretation.Within the space of a review it is necessary to limit discussion to a few topics. Particular stress has accordingly been laid upon (a) the chemical J. W. Eastes and W. M. Burgess J . Amer. Chem. SOC. 1942 64 1187. J. T. Burbage and W. C. Fernelius ibid. 1943 65 1484. J . 1890 57 749. 331 332 QUARTERLY REVIEWS mechanism of the reactions undergone by the carbonyls and eapecially of synthetic reactions ; (b) the newly discovered carbonyls of the platinum metah; and (c) the present statmi of the structural problem. The known carbonyls are listed in Table 11. 11. The FomnatiOn and Properties of the. MetuZ Cwbonyyls.-(l) Direct synthesis of carbonyla. The original discovery of the carbonyle aa also the large-scale preparation of the compounds hhga on the direct reaction between carbon monoxide and the free reactive metrtls.Two adsorption processes are thereby involved activated adsorption of carbon monoxide promoted by preparing the metal in a highly active state by reduction at low temperatures ; 4 6 and dorption of the carbonyl which bloch accma of carbon monoxide to the metal surface. One effect of operating a t high temperaturea and high pressures is doubtless to change the relative import- ance of these adsorption equilibria aa well as to shiff the chemical 8qud.i- brium between metal carbon monoxide and metal carbonyl. Only nickel carbonyl can be readily obtained a t atmospheric pressure (cf. the original preparation of iron pentacarbonyl 6 ) and in the 1.-G. high-pressure proceee (200" 50-200 atm.CO) the gas is continuously circulated to keep the iron mbonyl concentration below 2% so avoiding exceeeive adi3orption. Although the direct synthesis reaction is inhibited by oxygen yet it is promoted even at ordinary pressures by traces of sulphur. At high pressures material consisting largely of iron or nickel sulphides may be converted directly and nearly quantitatively into the volatile carbonyls.' In the presence of sulphides and of iron or copper to act aa ultimate acceptor of the sulphur molybdenum or tungsten can be converted directly into the hexacarbonyls on the technical scale.* In contrast with this iridium osmium and rhenium metals do not react with carbon monoxide under any conditions of temperature or pressure though their carbonyls have been prepared by the action of carbon monoxide on their compounds (11 4).( 2 ) Preparation of carbon@ by the Grignard reuctim. A. Job and his co-workera 9 observed that the slow reaction between carbon monoxide and the Grignard reagent was accelerated by salts of the transition metah and particularly by anhydrous chromic chloride. The organic products of this reaction are complex but an ether-soluble compound of chromium could be isolated in small yield ; this proved to be chromium hexacarbonyl Cr(CO),. A. Job and J. Ro~villois~a subsequently prepad tungsten hexacarbonyl W( CO), in the same way from tungsten hexachloride. The mechanism of this process has not yet been worked out. W. H i e k and E. Romberg 10 showed that no chromium carbonyl is formed before 4 Cf. F. W. Laird Ra. Trav. chim.1927 46 177; C. F. van Duin ibid. p. 381 5 A. Mittaach 2. angew. C h . 1928 41 827 ; Z. phy&al. Chem. 1902,40,1. * M. Berthelot Cmpt. r e d . 1891 ll2 1343; L. Mond and F. Quincke J. 1891 'I. G. Patents D.R.-P. 636,437 (Chcm. Zentr. 1930 11 601); 618,108 (ibid. * E.P. 367,481 ; F.PP. 708,260 708,379 (ibid. 1931 II 2041). * Cmpt. rend. 1926,188 392 ; BuU. SW. Ah. 1 9 2 7 a 1041. 00 Cmpt. rend. 1928,187 664. 59,604. 1933 II 3606) ; E.PP. 394,906 438,893 (ibid. 1936 I 2168). lo 2. anorg. Chem. 1936 jBl 321. Ni (28). I I Ni( CO) m.p. - 2 5 O b.p. + 43" colourleae w (74). W(CO), eublimee rhombic colourlesa Cr (24). ~ I I . p aublunes rhombic colourleea - TABLE XI C a r h y l s and Carbony1 Hydridea Mo (42). - WCO), sublimes rhombic colourless Mn (25). - - (43). - Re (75).Fe (26). I ~ * ( C O ) I O ~ colourless m.p. 177" sublimes monoclinic ' ?e( CO)& m.p. - 200 ' b.p. + 103" yellow ; decomp. loo" yellow triclinic ; F"eS(CO)l* decomp. 140" green monoclinic Fe,(CO), Fe(CO)& m.p. - 70" colourless Ru (44). JWCO), m.p. - 22" colourless ; Ru*(CO),. orange monoclinic ; RU,(CO)l* green needles 0 s (76). OS(CO), colourless m.p. - 15" ; bright yellow sublimes m.p. 224" os*(p), 0s(C0)4H2 Co (27). WCO), m.p. 51" orange-red ; decomp. 60° black crystalline c04(c0)11 Rh (45). Ir (77). IrdCO)* yellow-green sublimes ; crystalline WCO)SL canary-yellow decomp. 210" rhombohedra1 Pd (46). Pt (78). s34 W-Y BsJARws the stage of hydrolysis of the Gri@ reagent so the first product must be Borne organo-chromium carbonyl which is decomposed by acid forming Cr8+ ion and Cr(CO)6 amongst other products-e.g.(2 1) a ( c o ) & + 6H+ --+ Cr(CO) + 2Cr8+ + 12R + 6H + organic products + H This type of &mutation is commonly observed in the acid decomposition of substituted carbonyls. F. Hein’s work l1 has shown that in such circum- stances argano-chromium compounds are indeed formed and do undergo the required type of valency disproportionation. Only the non-polar halidea (W, MoCl, W G ) enter into the Grignad reaction ; stable complex salts do not react even though non-electrolytes the metal atom being presumably blocked from some initia,l step of co-ordinative union with carbon monoxide. The valency disproportionation involved in the Grignard synthesis is analogous to a num- ber of reactions in aqueous solutions whereby metal carbonyla are formed.For instance the complex cyanide of univalent nickel K,[Ni(CN),] ,I% absorbs carbon monoxide presumably forming K,[Ni( CN),( CO)] .Is When this ia acidified or when NiCN itself is treated with carbon monoxide nickel carbonyl is formed. Nickel(I1) cyanide Ni(CN), suspended in sodium hydroxide solution also absorbs carbon monoxide,l4 as also do nickel sulphide and the mercaptide.ls Manchot and Gall l3 suggested that the reaction proceeded in two steps formation of a carbonyl complex of univalent nickel (2 2) followed by disproportionation (2 3) (2 2) (2 3) ( 3 ) Indirect formation of uzrbonyls by reactions in solution. 2NiX + 2nCO --+ 2Ni(CO),& + X 2Ni(I) -+ Ni(I1) + Ni(0) ZNi(CO)& + (4 - 2n)CO --+ Ni(C0). + Nix (X absorbed by alkali; X = -OH -SH -SR) Later work by W.Hieber A. Schlecht and H. Behrens l6 has shown the peculiar aptitude of organic and inorganic thio-salts for this reaction e.g. nickel thiosalicylate xanthate etc. or hexamminonickel thio-salts such as [Ni(NH,),]MoS,. Whereas however the maximum absorption amounts to 4 molecules of CO per atom of Ni they found only half the nickel appearing as nickel carbonyl. The reaction thus presents some features still needing explanation. No analogous reactions are available for the formation of iron penta- carbonyl but cobalt compounds react similarly forming not the carbonyl but cobalt carbonyl hydride Co(CO),H (cf. 111 4). In the reactions undergone by substituted carbonyls and polynuclear carbonyla a second type of disproportionation is typical-redistribution of CO groups to give a free carbonyl [e.g.Fe(CO),] and products poorer in CO (e.g. Fez+). Thus decomposition of the substituted iron carbonyls l1 J . pr. Chem. 1931,152. 69. la J. Belucci 2. anorg. Chem. 1914 88 88. la W. Manchot and H. Gall Ber. 1926 69 1060. 1 4 A. A. Blanchard J. R. Fhfter and W. B. Adams J . Amer. Chem. Soc. 1934 l6 W. Manchot and H. Gall Ber. 1929 82 678. 56 16. Not yet published ; cf. W. Hieber Die Chemie 1942 66.7. ANDERSON CHEMISTRY OF TRE METAL CARBONYLS 335 Fe(CO),,A (A = MeOH etc.) furnishes Fe(CO) and Fe*+ as if thereaction were that of a Fe(CO) radical :17 (2 4) 3Fe(CO) - Fe(CO) + 2Fe(CO) (2 5 ) Fe(CO) + 2H+ -+ Fez+ + H + 2CO The hexacarbonyls of Cr and Mo are formed similarly in the decomposition of their derivatives M(CO),py (py = pyribe),l* and Ni(CO) from Ni( CO),phenan (phenan = o-phenanthroline).l9 Similarly a considerable part of the carbon monoxide displaced from iron tetracarbonyl by reaction with pyridine methyl alcohol nitric oxide etc. appears as iron penta- carbonyl (2 6) 2[Fe(CO),I3 + 3py -+ 3Fe(CO),,py + 3Fe(CO) Reactions of this kind originate in the ready reversibility of the mtul uzrbonyl + substituted carbonyl + CO system. With some carbonyls [e.g. Ni(CO), Fe(NO),(CO),] substitution of CO by pyridine can be effected only if the carbon monoxide is removed continuously ; l9 at high temperatures and CO pressures Cr(CO) forms Cr(CO),,py, but if the back reaction of (2 7) is obviated Cr(CO),,py is formed (2 7 ) Cr(CO),,py + py + Cr(CO),,py + co (4) New carbonyls of the platinum metals the high-pressure synthesis. Ruthenium carbonyl.That compounds of the metals could be converted directly into carbonyls by carbon monoxide a t high pressures was shown by W. Manchot and W. J. Manchot 20 in their preparation of ruthenium pentacarbonyl Ru(CO) . Ruthenium iodide RuI, reacts with carbon monoxide even a t the ordinary pressure the carbonyl iodide Ru(CO),I, being formed by partial displacement of iodine. If mixed with an acceptor for iodine such as finely divided copper or silver the carbonyl iodjde under- goes further ready reaction with carbon monoxide-especially at high pressures-forming Ru(CO) . Although metallic ruthenium does react with carbon monoxide,21 it is unlikely that RuI or Ru(CO),I is reduced to free metal as the initial step for carbonyls of iridium and osmium-metals which are quite inert towards carbon monoxide-have been prepared similarly.The reaction is better regarded as involving a series of displacement equilibria whereby the formal valency of the metal is lowered at each step (2 8) ~ R u I + 4CO + 2Ru(CO),12 + I (2 9) RuI + Ru(CO),12 + Ru(CO)J + Ru(CO) was described earlier by W. Manchot and E. Enk.22 A carbonyl bromide Ru( CO)Br containing formally univalent ruthenium l7 W. Hieber and E. Becker Ber. 1930 63 1405. W. Hieber and F. Muhlbauer 2. anorg. Chem. 1935 221 337. W. Hieber F. Miihlbauer and E. A. Ehmann Ber. 1932 65 1090. 2o 2. anorg. Chem. 1936 226 385. 21 Contrast R. L. Mond and A. E. Wallis J . 1922 121 29. Ber. 1930 63 1635. Iridium wbonyls Ir8(CO)8 F.(CO)&. W. Hieber and IE. w d y m and W. Hieber H. Lagally and A. Mayr 24 confirm the mechanism dequ* tion (2 9) by their work on iridium carbonyls.The iridium halides react with carbon monoxide (110" ordinary pressure) the first stage being described by (2 10) or (2 11) (2 10) (2 11) 21r13 + 4CO + 2Ir(CO),I + I This reaction wits observed previously by W. Mitnchot and H. Gall 25 ; it may be followed successively by reactions (2 12) and (2 13) (2 12) 2Ir(CO),X + 3CO -+ 2Ir(CO),X + COX (2 13) 2Ir(CO),X + CO -+ [Ir(CO),]z + COX In these changes the exothermic formation of carbonyl chloride or bromide at each stage plays an essential part in determining the oourse of reaction under ordinary pressure of carbon monoxide. As COI is non-existent the reaction with IrI stops a t (2 11) ; only with IrC1 does the third stage (2 13) occur though the main product is Ir(CO),Cl. In the presence of silver or copper to combine with halogen and at high pressures of carbon monoxide ultimate formation of carbonyl is favoured and it may logically be inferred that the tricarbonyl formed in reaction (2 13) is the precursor of [Ir( C0)J 2 the final product.Under these conditions the over-all process is as equation (2 14) and the ease with which halogen is split out of the ZIrX + 5CO -+ 2Ir(CO),X + COX (X = C1 or Br) + co (2 14) IrX + 3Cu + 3CO -+ 3CuX + [Ir(CO),Iz ___+ [Ir(CO),] iridium halides is more important than the heat of formation of COX,; conversion into carbonyl then proceeds with increasing ease in the sequence chloride + bromide -+ iodide. Little or no formation of carbonyl takes place if reduction to metal occurs-as in mixtures of iridium halides with copper or silver powder.These have been prepared like the carbonyls of iridium by high-pressure synthesis from the osmium halides,28 by reactions hinging on the intermediate formation of the car- bony1 halides Os(CO),X,. These CO-richest compounds (see Table IV) are formed from the halides (OsCl, Os,Br9 osmium " oxyiodide ") in carbon mon- oxide (120° 200 atm. pre~sure).~' In the presence of silver or copper-from the lining of the autoclave-partial conversion into the pure carbonyl occurs increasing in extent in the order Os(CO),Cl < Os(CO),Br < Os(CO)J,. Osmium carbonyl is most readily obtained however by an exceptional reaction vix. the direct action of carbon monoxide on the covalent highest oxide Osmium carbonyls Os(CO), Os,(CO),. (2 15) OsO + 9CO + OS(CO)~ + 4C0 23 2. anorg. Chem.1940 245 321. 2c Ibid. 1941 246 138. a 6 W. Hieber and H. Stallrnann 2. Elektrochern. 1943 49 288. 26 Ber. 1925 58 232. Idem Ber. 1942 75,f11472 ; cf. W. Mctnchot and J. KGnig ibid. 1926 68 229. ANDERSON CHEMISTRY OF THE METAL CARBONYLS 337 This reaction occurs fairly readily (loo" 50 atm.) presumably by way of an intermediate oxy-carbonyl since metallic osmium formed by reduction in a side reaction is without action upon carbon monoxide. Rhodium carbonyls [Rh( CO),] , [Rh(CO),], [Rh,( CO),,],. The existence of rhodium carbonyls is to be expected since cobalt and iridium form compounds of similar nature. W. Hieber and H. Lagally28 have shown that rhodium does indeed stand between cobalt and iridium in its reaction with carbon monoxide. As with cobalt [Rh(CO)JP is most readily formed by direct union of carbon monoxide with metallic rhodium at high pressures.The halides RhX react with carbon monoxide like those of iridium forming the rhodium carbonyl halides Rh( CO),X and the carbonyls [Rh(CO),], Rh,(CO),. The pure carbonyls especially the vary stable lower carbonyl [Rh( CO),], are formed even at low temperatures in the presence of a halogen- acceptor (Ag or Cu) ; the use of zinc or cadmium as halogen-acceptors does not however lead to the formation of mixed carbonyls as it does with cobalt carbonyl (see 111 6). The superior stability of the tricarbonyl is striking and at higher temperatures (> 100") a yet more highly condensed carbonyl is formed in the high-pressure synthesis &x. Rh,(CO), or a polymer thereof. The carbonyl derivatives .of rhenium are of special systematic importance as bridging the gap between the complete series of hexacarbonyls M(CO) of Group VI (M = Cr Mo W) and the complete series of pentacarbonyls M'(CO) and carbonyl halides M'(CO),X of the iron group (M' = Fe Ru 0 s ) .The compounds are of the types expected 29 carbonyl halides Re(CO),X and dimeric carbonyl [Re(CO),],. Rhenium carbonyl halides are so stable and so readily formed that they are the sole and invariable products of high-pressure syntheses from the halides or complex halides of rhenium. Thus from ReCl, K,ReC1 or K,ReBr in the presence of copper Re(CO),Cl and Re(CO),Br respectively are formed (230" 50 atm. of CO) ; carbon monoxide forms Re(CO),I from K,ReI even at the ordinary pressure. This great tendency to form the carbonyl halides is demonstrated by their formation from metallic rhenium or its oxygen compounds in the presence of the reducible halides of transition metals ; e.g.(2 16) Re + CuX + 6CO + Re(CO),X + Cu(C0)X (2 17) 2Re + Nix + 14CO + 2Re(CO),X + Ni(CO) or by chlorination with e.g. carbon tetrachloride (2 18) KReO + CCl + 8CO + KCl + Re(CO),Cl + COCl + 3C0 Their stability is such that the cowersion of these carbonyl halides into rhenium carbonyl has not been observed under any conditions investigated. [Re(CO),] is formed however,. directly from the sulphide Re,S and the oxygen compounds Re,O, KReO (250" 200 atm. of CO). The dimeric carbonyl ( CO),ReCORe( CO), resembles the co-ordina- 2. anorg. Chem. 1943 $361 96. W. Hieber and H. Schulten &id. 1939 243 164; W. Hieber R. Schuh and Rhenium carbonyl [Re( CO) ,I2.co H. Fuche ibid. 1941 248,243 ; W. Hieber and H. Fuchs &d. p. 256. 338 QUARTERLY REVIEWS tively saturated hexacarbonyls [Cr( CO), etc.] in its stability and chemical inertness. It is not decomposed by alkalis or concentrated mineral acids and although as a polynuclear compound it is not very volatile i t can be sublimed in carbon monoxide at 200". It reacts with gaseous halogens forming the stable carbonyl halides. ( 5 ) High-pressure synthesk of the iron-group carbonyls. Numerous patent specifications 30 record the conditions found empirically to promote the formation of metal carbonyls under technical conditions. These can now be correlated with systematic studies by Hieber and his co-workers on the mechanism of the high-pressure synthesis. Carbonyls can be formed from the compounbs of iron cobalt and nickel with highly polarisable non-metals ; i.e.from solids in which the lattice forces are not of purely ionic type. Thus cobalt sulphide (NiAs structure) is quantitatively converted into Co,(CO) (200° 200 atm.) but cobalt oxide does not react 31 (2 19) 2CoS + 8CO + 4Cu (autoclave lining) -+ Co,(CO) + 2Cu,S Of the halides CoF (rutile structure purely ionic) is without reaction whereas the other halides (layer lattice structures) increase in reactivity in the order CoC1 < CoBr < CoI,. Thus in carbon monoxide a t 250° 200 atm. and without direct contact with a free metal (so that halogen must be bound by reaction of some volatile compound with the silver or copper autoclave lining) relative conversion of halides into carbonyl accord- ing to (2 20) is x = c1.Br. I. yo Co,(CO) formed = 3.5 9 100 (2 20) The volatile compounds involved might be (a) COCl, COBr, or (from CoI,) free I ; or ( b ) a volatile cobalt carbonyl halide. No formation of COC1 or COBr has been detected and where the reaction mechanism definitely involves these compounds the sequence of reactivity appears to be iodide < bromide < chloride (cf. iridium carbonyl). H. Schulten 32 has however found that a t the ordinary temperature under 100 atm. of carbon monoxide cobalt iodide reacts to form a cobalt carbonyl iodide Co(CO)I, a dark brown crystalline solid with a high dissociation pressure of carbon monoxide which is perceptibly volatile even a t room temperature. The relative efficiency of the cobalt halides as reactants in the carbonyl synthesis would be consistent with an increase in stability of cobalt carbonyl halides in the sequence chloride < bromide < iodide which holds for the iron carbonyl halides.A second factor involved in the reaction is the r61e of the metal added as halogen acceptor. In intimate mixtures of cobalt bromide with finely divided metals the efficiency of various metals (Au Ag Cu etc.) in removing halogen and forming cobalt carbonyl increases in the order of the heat of Cf. refs. (5) and (7) and the compilation of patent refs. by R. L. Mond ( J . Soc. Chern. Ind. 1930 49 2 8 8 ~ ) . 2CoX + 4Cu + 8CO + Co,(CO) + 4CuX 91 W. Hieber H. Schulten and R. Marin 2. anorg. Chsm. 1939 240 261. sa Ibid. 1939 243 145. ANDERSON CHEMISTRY OF THE METAL CARBONYLS 339 1 3 1 4 1 1.5 1 2 formation of their bromides.(2 21) CoBr + 2M -+ 2MBr + Co + Q cnl. the metals cited be endothermic. Endothermic processes involving only solid reactants and resultants can take place only if the formation of solid solutions provides an increase in entropy. There is in fact limited mis- cibility between AgBr or CuBr and CoBr, so that in mixtures of CoBr + Ag (or Cu) heated in argon under conditions comparable with those of the carbonyl synthesis some metallic cobalt is set free. However as is shown by the data of Table 111 the yield of reaction (2 21) in argon compared with that of the (over-all exothermic) reaction (2 22) in carbon monoxide suggests that something more than the reaction of carbon monoxide with freshly liberated metallic cobalt is involved in carbonyl formation.(2 22) The displacement reaction (2 21) would for CoBr + 4Ag + 8CO -+ Co,(CO) + 4AgBr + ( Q f - 15.2) kg.-cal. (Qf = unknown heat of formation of Co,(CO), in kg.-cal.] 250" 250 180 180 TABLE I11 "/b lteaction at 200 atm. 31. Ag cu Cd Zn Jn argon. 6.2 12.5 20 23 I n CO. 29 76 100 100 This conclusion is borne out by experiments with zinc and cadmium as halogen acceptors. With these reaction (2 21) is incomplete under the experimental conditions although exothermic. In carbon monoxide however dehalogenation is quantitative and cobalt is converted completely not into [Co(CO),], but into the mixed carbonyl compounds [Co( CO),],Zn Similar results were found by W. Hieber H. Behrens and U. Teller 33 in comparable experiments with iron and nickel halides. The chlorides and bromides FeX, Nix undergo but little reaction when heated in carbon monoxide without direct contact with a halogen acceptor.Quantitative conversion of NiI into Ni(CO), and 50-75% conversion of FeI into Fe(CO) occurred ; in the latter case the iron is largely left as Fe(CO),I which because of its stability slows down the last stage of carbonyl forma- tion. Although no nickel carbonyl iodide was isolated the results suggest the formation of such a compound stable only under high pressures of carbon monoxide. The catalytic r81e of iodine in carbonyl formation is now apparent since it can enter into a cyclic set of reactions whereby the metallic iodide is formed and converted a t each stage into the carbonyl iodide and thence [Co(CO),I,Cd* 3 3 Ibid. 1912 251 26. a40 QUARTERLY REVIEWS into carbonyl and metallic iodide.The function of sulphur m the technid carbonyl processes must be andogous. No metal carbonyl dphides are known but they presumably exist as intermediates and the h- lation of an iron carbonyl selenide Fe,Se,(CO)s,~ formulated by Hieber as (I) may be taken aa evidence that such is the cam. 111. Metal Carbonyl Hydrides and their Derivatives.-( 1) The hydrolysis of iron pentwrbonyl. It was early observed 36 that iron pentacarbonyl w a ~ soluble in alcoholic potassium hydroxide giving a solution which turned red in air. Such solutions have powerful reducing propertiea,86 and absorb oxygen from the air. W. Hieber and F. Leutert 37 found that in complete absence of oxygen a volatile very unstable iron carbonyl hydride Fe(CO),H, is liberated when the solutions are acidified.This is formed by the hydro- lysis reaction 38 (3 1) Fe(CO) + 20H- -.+ Fe(CO),H + C0,B- With caustic alkalis formation of carbonate approaches one molecule per molecule of iron pentacarbonyl when a suitable excess of base is used ; with ammonia ethylenediamine etc. carbamic acids NHR*CO,H are formed as primary products. The carbonyl hydride formed in the process can be oxidised and titrated with methylene-blue. Mild oxidising agents (e.g. manganese dioxide) oxidise iron carbonyl hydride in concentrated solutions to iron tetracarbonyl in almost quantitative yield 39 (3 2) Drastic oxidants (e.g. hydrogen peroxide) bring about complete oxidation to ferric hydroxide carbon monoxide and carbon dioxide. Iron carbonyl hydride is a pale yellow liquid m.p. - 70° with a nauseat- ing smell.Above - 10" it decomposes rapidly giving free hydrogen and by disproportionation of the resulting Fe( CO) radical iron pentacarbonyl and ill-defined products with a lower CO Pe ratio (3 3) Fe(CO),H -+ H + Fe(CO) (3 4) 2Fe(CO) -+ Fe(CO) + Fe(CO), etc. followed by reactions (2 4) (2 5 ) . The most important reactions of iron carbonyl hydride are those giving rise to metal derivatives. These together with derivatives of cobalt car- bony1 hydride are considered in 111 3. ( 2 ) Hydrolysis of cobalt carbonyl. The reaction of cobalt tetracarbonyl with bases although formally more complex than that of iron pentacarbonyl co (CO),Fe*Se-Fe*Se*Fe(CO) CO (1.) 3Fe(CO),H + 3M1-102 + 3H2S0 + 3Mnso4 + 3H20 + [Fe(CO),] 3 4 W. Hieber and 0. Geisenheimer unpubl. ; cf. ref. (16). 36 J.Dewar and H. 0. Jones Proc. Roy. Soc. 1905 A 76 558 ; 1906 A 79 66. 38 €3. Freundlich and W. Malchow Ber. 1923 56 2264 ; 2. anorg. Chem. 1924 37 Naturwiss. 1931 19 360. 39 W. Hieber ibid. p. 165. 141 317. W. Hieber and F. Leiitert 2. anorg. Chem. 1932 204 145. MDEBSON cBmrsTBu or JEEE YBLTAL CARBONYLS 341 confom to trhe m e patfenn.a With strong barns [&(OH), KUH] the reaction is (8 6) 3[Co(CO),] + 40H- -+ 4Co(CO),H + 2C0,'- + 2[Co(CO),] p o l p e r At lower hydroxyl-ion concentrations (ammonia) an alternative hydro- lysis reaction is favoured (3 6) 3[Co(CO),] + 4H,O -+ 4Co(CO),H + 2Co(OH) + 8CO The dilute solutions obtained by hydrolysis have properties similar to those of iron carbonyl hydride they reduce methylene-blue and are very easily oxidised by air or mild oxidants forming cobalt tetracarbonyl.From such a solution (baryta hydrolysis) W. Hieber and H. Schulten 4 1 isolated the free hydride by the action of phosphoric acid. Co(CO),H is extremely unstable and decomposes into cobalt tetracarbonyl and hydrogen at temperatures above its melting point (- 26.2"). ( 3 ) Derivatives of metal carbonyl hydrides. The carbonyl hydrides behave as very weak monobasic acids forming true salts only with the alkali metals and with bulky ammine cations; compounds formed with the heavier metals do not have the properties of typical salts. F. Feigl and P. Krumholz 42 isolated a salt Fe(CO),HNa,MeOH as the immediate product of hydrolysing iron pentacarbonyl with sodium methoxide ; calcium or magnesium hydroxides similarly form [Fe(CO),H] ,Ca and [Fe(C0),H],Mg.d3 Bmines e.g.pyridine and o-phenanthroline also react kith iron carbonyl hydride solutions giving stable addition compounds which can be formulated as salts,** e.g. [Fe(CO),H,] ,2C5H5N or [Fe(CO),][C,H,NH], [Fe(CO),] [phenan,H,] . The electrically conducting solution formed by iron carbonyl hydride in pyridine accords with this view. This behaviour is in contrast to the action of pyridine on the pure carbonyls and all other classes of derivative ; in general reaction leads to displacement of carbon monoxide and substitution of pyridine. From solutions of iron carbonyl hydride sparingly soluble derivatives may be precipitated by reaction with solutions of the ammines and salts of the heavy rnetal~.4~ 46 With solutions containing the hexammine cations [M(NH3),J2+ (M = MnII FeII CoII NP) the crystalline compounds [Fe(CO),H],[M(NH,),] are obtained; no analogous salt is formed by the hexamminocobaltic ion [Co(NH3),13+ as this acts aa an oxidant forming [Fe(CO),],.In the hydrolysis of iron pentacarbonyl with ammonia a little of the salt [Fe( CO),H],[Fe(NH,),] is invariably formed through reactions analogous t o (3 3) (3 4) (2 5 ) followed by (3 7 ) Fee+ + 6NH3 -+ [Fe(NH3)s]a+ Very sparingly soluble salts [Fe(CO),H],[M phenan,] (M = FeII CoII 40 Idem 2. Elektrochem. 1934 40 168. 41 2. anorg. Chem. 1937 232 29. 43 F. Rein and H. Pobloth 2. anorg. Chern. 1941 248 84. 44 W. Hieber and H. Vetter ibid. 1933 212 145. 46 F. Feigl and P. Kx-umholz ibid. 1933 215 242. 46 W. Hieber and E. Fack aid. 1938 288 83. r e Monatsh. 1932 59 314. 342 QUARTERLY REVIEWS Ni") are also formed by the trisphenanthroline cations [M phenan,]z+ ; these are precipitated almost quantitatively.Solutions of these carbonyl hydride salts of complex cations in acetone or methyl alcohol have conductivities typical of solutions of strong electro- lytes,ds and it may be inferred that the compounds are true salts. The reactions of cobalt carbonyl hydride solutions are analogous ; 47 with ammoniacal nickel or cobalt solutions the salts [Co(CO),],[Ni(NH,),J [CO(CO),],[CO(NH~>~] are obtained the latter being formed (by reaction 3 6) in small amounts in the hydrolysis of cobalt carbonyl with ammonia. Gaseous ammonia also reacts directly with cobalt carbonyl Derivatives of markedly different properties are formed by the metals of the zinc and copper groups. Although formally related to the carbonyl hydrides their properties are not those of typical salts and they can be regarded rather as polynuclear complex compounds.That iron pentacarbonyl reacts with mercuric salts was observed 48 before iron carbonyl hydride was discovered (3 9 ) Fe(CO) + HgSO + H,O -+ H,SO + C02 + Fe(CO),Hg With a further molecular proportion of mercuric salt compounds of the type Fe(CO),Hg,X are obtained. Fe(CO),Hg is *a very stable .yellow substance insoluble in all solvents. It is quite unchanged in air and does not react with boiling pyridine but reacts with iodine forming iron carbonyl iodide (3 10) Fe(CO),Hg + 21 + HgI + Fe(CO),I It is decomposed a t 150" into mercury iron and carbon monoxide. Its formal relation to Fe( CO),H is shown by its precipitation from iron carbonyl hydride aolutions (3 11) Fe(CO),H + HgCl -3 2HC1 + Fe(CO),Hg Similar reactions with ammoniacal zinc or cadmium solutions form the insoluble compounds [Fe(CO),][Zn(NH,),] [Fe(CO)J[Cd(NH,)J [Fe(CO)J[Cd py,].I n these derivatives of the Group IIB metals there is a clear gradation of properties HOAc HOAc HCl F W O ),Zn(NH,) + Fe(CO),H Fe(CO),Cd(NH,) ___+ Fe(CO),Cd + Fe(CO),H Similar in type are the copper and silver compounds Fe(CO),Cu,(NH,), Fe( CO),Ag,phenan. These compounds are not derived from stable ammine cations and in Fe(CO),Hg . . . unaffected by acids 47 W. Hieber and H. Schulten 2. anorg. C'hem. 1937 232 17. H. Hock and H. Stllhlmann BBT. 1928 61 2097 ; 1929,82,2690. m E B S O N CEBMIS'pBY OR RSBTfi CYABBONYLS 34s them-in contra& to its behaviour in truly ionic compounds-E'e(CO),H functions aa formally dibasic Some rather unstable organometa,l.lic derivatives have been deecribed by F.Hein and E. Heuser.'@ Methylmercuric hydroxide CH,*HgOH mete with iron cctrbonyl hydride forming Fe(CO)4(Hg*C13[,),. This subetance is soluble in organic solvents [contrast Fe(CO),Hg] and presumably has a normal molecular weight. It slowly disproportionatea into Fe( CO),Hg and Hg( CH,) ; the correaponding ethylmercury and phenylmercury compounds are too unstable to be isolated and decompose immediately in a similar manner. A diethyl-lead derivative Fe(CO),Pb(C,H,), has also been described ; 451 it is formed by the action of (C,H,),PbOH on iron carbonyl hydride and is soluble in organic solvents. Cobalt carbonyl hydride solutions yield heavy-metal compounds of similar character e.g.[Co(CO)J,Cd [Co(CO)J,Hg. These (compare the iron compounds) are monomeric and soluble in organic solvents but insoluble in water. The tendency to form cobalt carbonyl hydride is so great that it may be formed from carbon monoxide and cobalt(I1) compounds by reactions parallel to those described in 11 3 for the preparation of nickel carbonyl. Cysteine SH*CH,CH(NH,)*CO,H (= H,SR below) forms with bivalent iron and cobalt complex salts of the type (I). Alkaline solutions (4) Fomnation of mrbonyl hydrides by reactions in solution. K,[M<&YHs H,*CH*CO2- ) l o or 2H,0 K2[(C0)*Fe<YS H,*CH*CO2- ),I (1.) (11.) of these inner complex salts are sensitive to oxygen and also absorb carbon monoxide.60 From the ferrous dicysteine complex a dicarbonyl derivative (11) is formed.The cobdt(II)-cysteine complex absorbs one molecule of CO per atom of cobalt but no addition compound can be isolated ; the complex undergoes disproportionation the cysteine being recovered in the form of a cobalt(II1) complex K,[Co(SR),] ,3H,O and the complementmy product of reaction being cobalt mbonyl hydride (3,12) 9[Co(SR),la- + 8CO + 2H,O -+ S[CO(SR),]~- + Co(OH) + 2Co(CO),H Prolonged action of carbon monoxide forms the carbonyl hydride in increased amount by reaction with the cobaltitriscysteinate formed in reaction (3 12) (3,13) [Co(SR),I3- + 6CO + 70H- + 2C0,4- + 3SRS- + 3H,O + Co(CO),H The net effect of (3 12) and (3 13) is to regenerate the cysteine so that as was found by G. W. Coleman and A. A. Blan~hard,~~ a small amount of cysteine suf%ces to bring about absorption of carbon monoxide correspond- ing to nearly complete conversion of the cobalt salt.2. anorg. Chrn. 194.2 240,293. 6o M. P. Schubert J . Amer. Chern. Soc. 1933 66,4663. u1 Ibid. 1936 58 2160. 344 QUARTERLY REVIEWS The cobalt carbonyl hydride formed in these processes was chtlrac- terised by precipitating the silver and mercury derivatives. Formation of cobalt carbonyl by decomposition when the solution is acidified provides a convenient means of preparing that substance. In alkaline solutions cobalt carbonyl itself reacts with free cysteine forming the carbonyl hydride (3,14) 2[C0(C0)4] + 3HzSR + 30H-+ 3Co(CO),H +[CO(SR),]~- +4CO + 3H,O Analogous reactions take place with cobalt(I1) salts and other thio- compounds that form inner complex ~alts.5~ Absorption of carbon mon- oxide first forms a substituted cobalt carbonyl derivative 53 (3 15) which furnishes cobalt carbonyl hydride when decomposed with acid (3 16).Thus with potassium xanthate KXa (Xa = C,H,*O-CS*S-) (3 15) (3 16) 6CoCl + 12KXa + 5CO + EtOH- Co,(CO),,EtOH + 2H++EtOH + Coz+ + Co(CO),H + CO + iH2 12KC1 + 4CoXa + Co,(CO),,EtOH The foregoing reactions are applicable only to the preparation of cobalt carbonyl hydride but if the sequence of properties in the pairs Ru-Rh Os-Ir recapitulates that of the pair Fe-Co similar methods may be available for the preparation of Rh(CO),H and Ir(CO),H. Reaction (3 16) however is typical of the acid decomposition of substituted carbonyls involving a disproportionation of valencies and redistribution of carbonyl groups. Formation of ferrous ion and iron carbonyl hydride occurs to some extent in the acid decomposition of amine- or alcohol-substituted iron carbonyls especially those with the ratio CO Fe = 3 1 or 2 1.Thus the ethylene- diamine compound Fe,(CO),en reacts quantitatively according to (3 17) 64 (3 17) Fe,(CO),en + 2H+ -+ Fez+ + Fe(CO),H + 3 en With tricarbonyl derivatives the corresponding reaction occurs to a smaller extent possible products 55p 56 being (3 18) (3 19) 3Fe(CO),,X + 2H+ + Fez+ + Fe(CO),H + Fe(CO) + 3X 2Fe(CO),,X + 2H+ + Fez+ + Fe(CO),H + 2CO + 2 X It may be noted that although the quantitative formation of iron carbonyl hydride in reaction (3 17) has led to the suggestion45 that the amine compounds could be formulated as carbonyl hydride salts yet no salts with similar complex cations @I en3]2+ could be prepared syn- theti~ally.4~ The nickel and cobalt hexammine salts of iron carbonyl hydride react with carbon mono~ide,~2 the final products being iron carbonyl hydride M W.Hieber 2. Elektrochem. 1937 43 290; Angew. Chrn. 1936 49 463; DiS os Cf. ref. (19). 6 p W. Hieber and F. Leutert Ber. 1931 04 2832. ss Cf. ref. (17). 6e W. Hieber and H. Vetter Ber. 1931 64 2340. Chernb 1942 55 7. ANDERSON CHElldZgTaY OF !l'HE METAL UARBONYLS 345 and either nickel carbonyl or cobalt carbonyl hydride. IIieber coneidem that mixed carbonyls not yet isolated are formed as intermediates [Fe(CO),HI,[Co(NH,),I -+ [Fe(CO),Hla[CdCO)41+ Fe(CO),H + [Fe(CO)& + co(Co),H * [Co(co)aI* + Hii ( 5 ) Formation of metal mrbonyl hydrides in the high-pressure syntkis. Hieber's recent work on the high-pressure synthesis has disclosed a number of new reactions by which the carbonyl hydrides may be formed.These are of interest not only in themselves but also in that they emphasise the peculiar ease of formation of cobalt carbonyl hydride and suggest similar properties for the compounds of the related elements rhodium and iridium. In their study of the synthesis of cobalt carbonyl W. Hieber El. Schulten and R. Marin 31 found that the gases blown off from the autoclave frequently contained a volatile carbonyl compound although reagents that could lead to contamination with nickel carbonyl or iron penhcaxbonyl were caremy avoided. By precipitation of Co(CO),HgCl,QH,O from mercuric chloride solution the volatile compound w&s shown to be Co(CO),H which formed when the reactants contained traces of moisture.When cobalt aulphide or iodide was deliberately moistened formation of the carbonyl hydride took place very readily by some process such as (3 20) 2C0S + Ha0 + 9CO + 4Cu -+ 2Co(CO),H + CO + 2-S The formation of Rh(CO),H Ir(CO),H and probably Os(CO),H culd Re(CO),H occurs under similar conditions. Total synthesis of cobalt carbonyl hydrick. Hieber Schulten and Bkrh found further that cobalt carbonyl hydride is formed in circumshnoes which amount to a total synthesk from ita constituents. Thus cobalt carbonyl is partly converted into the carbonyl hydride when it is heated in hydrogen (120 atm.) and carbon monoxide (160 atm. to prevent decom- position) at 165' (3 21) [CO(CO)J2 -j 2CO(CO),H Reaction (3 21) is thus reversible the synthesia being the converae of the spontaneous decomposition of the hydride at ordinary pressures.Partial converaion into hydride also OCCWB when metallic cobalt or cobalt aulphide is heated in hydrogen (50 atm.) and carbon monoxide (3 22) 2Co + 8CO + H,+ ~CO(CO)~H (3 23) 2CoS + 8CO + H + 4Cu+ ~CO(CO)~H + 2Cu,S Rather better conversion into the hydride waa found in the action of carbon monoxide on T. Weichselfelder's cobalt hydride COH~.~' Hieber attzibub the ready formation of cobalt carbonyl hydride in this w e fo the pre- existence of the Co-H bond but this view is not compatible with current models for the constitution of the carbonyl hydrides. 67 Annalen 1926 447 64. 346 QUARTERLY REVIEWS Rhodium carbonyl hydride has also been obtained by reaction (3 22). By contrast iron carbonyl hydride is formed only in solution by the reactions cited in I11 (1) and I11 (4).Careful investigation 5 8 has shown that none of the methods cited in this section yields any iron carbonyl hydride neither is any formed when iron tetracarbonyl or iron carbonyl iodide is heated in hydrogen and carbon monoxide. In every case the product of reaction is iron pentacarbonyl. The ease of formation of cobalt carbonyl hydride is reflected by the direct formation of its heavy- metal derivatives a t high pressures. As was discussed in I1 ( 5 ) the r61e of silver or copper in the synthesis turns upon the coupling of the exothermic formation of carbonyl with the endothermic reduction of cobalt sulphide or halide. Using metals baser than copper so that the reduction also is exothermic reaction with carbon monoxide becomes quantitative ; the product is not [Co( CO)4], however but the corresponding heavy-metal derivative e.g.[Co(CO),],Zn. The compounds Co(CO),Tl (yellow) [Co(CO),],MII [Co(CO),l3MI1I [MI11 = Ga (?) In (red) T1 (violet)] (6) High-pressure synthesis of mixed carbonyls. [MI1 = Zn Cd Hg Sn Pb (?)I yellow or orange have been obtained by three types of reaction 59 (a) The reaction just discussed typified by (3 24) ZCoBr + 3Zn + 8CO + BZnBr + [Co(CO),],Zn The thermochemistry of the process makes it possible to form the mercury compound by a reversal of the r6les of halogen acceptor and free metal (3 25) HgX + 3c0 + 8CO + COX + [Co(CO),],Hg (X = C1 Br I) (b) Total synthesh ( c ) From pre-formed cobalt carbonyl and zinc cadmium or mercury (3 26) in carbon monoxide at hgh pressures (3 27) [Co(CO)& + Cd -+ ICo(CO),I,(=d The ability to form heavy-metal derivatives in this way is restricted to the metals in the portion of the Periodic Table shown inset the tendency to do so being greatest for mercury.The heavy-metal compounds ob- tained in this way are identical with some prepared from solutions of cobalt At3 Sb carbonyl hydride (cf. 111 3). All are soluble in organic solvents with normal Au (Pb) Bi molecular weights. They are distinctly more stable then cobalt carbonyl itself may be sublimed and melt with some decomposition above 70". They are decomposed by an excess of halogen but with one molecular proportion cobalt carbonyl and 2Co + Zn + 8CO + [Co(CO),]Zn M g cu pr 68 W. Hieber and U. Teller 2. anorg. Chem. 1942 249 58. Idsrn ibid.p. 43. ANDERSON CHEMISTRY OF THE METAL CARBONYLS 347 the halide of the metal are formed. Nitric oxide reacts to form cobalt nitrosocarbonyl ; with [Co(CO)p12Cd just half the cobalt is converted into Co(CO),NO. Pyridine does not displace carbon monoxide but forms addition compounds in which it is presumably linked to the second metal. In view of their properties these compounds are formulated hy Hieber co co co and Teller as polynuclear complexes assigned to the amminated derivatives of iron and cobalt carbonyl hydrides tetracarbonyl Fe(C0)4Hg could obvi- ously acquire a structure essentially similar to (I) by a process of infinite polymerisation and it is significant that it is in fact an insoluble completely non-volatile compound in contrast to Metal Curbonyl Halides.-( 1) The halogen compounds of the metals with co-ordinated carbon monoxide are of interest in that they link the chemistry of the carbonyls with the more familiar chemistry of co-ordination compounds especially of 4 / 4 r C / \ .) ( co e.g. (I). Similar constitutions can be co Hi3 c o obtained fiom solution. Mercury iron (1.) 7cof b o f \CO CO [co(co)4l,Hg- IV. the platinum metals. compounds does not coin- cide exactly with the for- T& Metals enclosed - form carbonyla. mation of pure carbonyls as is indicated by the inset. An extension of the range of conditions experimentally accessible would doubtless increase the number of known compounds as the recent isolation of Co(CO)I shows. The experimental evidence points to the existence of an analogous nickel carbonyl iodide at high pressures ; at the ordinary pressure the existence of K,[Ni(CN),(CO)] has been established.13 The sulphate Ag,SO,,CO is the only known carbonyl derivative of silver.s0 For the metals Pd Pt Cu Au which form no pure carbonyls the sequence of stability of the carbonyl halides appears to be iodide < bromide < chloride.61 The stability ease of formation and volatility of the compounds of the carbonyl-forming metals however all trend in the opposite sequence running parallel with the polarisability of the halogen and the increasingly covalent character of the M-X valency forces.The formation of the platinum-metal carbonyl halides by direct com- bination of carbon monoxide with the metallic halides was observed before systematic inGestigation coupled their reactions with those of the carbonyls proper (e.g.the platinum carbonyl halides,s2 ruthenium carbonyl Metals enclosed - - - - form carboayl halides. 6o W. Manchot and J. Konig Bw. 1927 60 2183. 61 Cf. 0. H. Wagner 2. anorg. Chsm. 1931 196 364. dl P. Schutzenberger Ann. Chim. 1868 15 100; 1870 21 360; F. Mylius and F. Foerster Ber. 1891 24 2424 3751. Z 348 QUABTERLY B3cvIIEw8 bromide and the carbony1 chlorides of p&lladi~m,~~ gold,66 iridium,@6 and osmium67). The general chemistry of compounds of this type has been summarised and those aspects which are relevant to the chemistry of the carbonyls have received notice in diacusaing the synthesis of the platinum-metal carbonyls (11 4). Table IV collects together the known types of carbonyl halides and makes apparent the tendency to form compounds in which the stable co- ordination numbers 6 or 4 (probably square planar configuration) are attained either in monomeric compounds-e.g.Re(CO),X Os(CO),X, Ir(CO),X Pt(CO)&,-or by formation of a binuclear complex aa in [oS(co)4x] 2 [Rh(co)2xI 2- TABLE IV Metallic Carbony1 Halides and Reluted Ccnnpounds K,[Co(CN),CO] [Rh(CO),XIa M n - K2[Ni(CN),CO] [Pd(C0)Cl,lYI H[Pd( CO)Cl,] ~ Re(CO),X (2) Iron carbony1 halides. Co(C0)I2 1Ni - I c=u(CO)X Au( C0)Cl An adequate discussion of the chemistry of the carbonyl halides is possible only for the iron compounds. The range of known compounds including the substituted carbonyl halides is sum- marised in Table V ; the ratio CO Fe varies in these from 5 1 to 1 1 the TABLE V I I I W. Manchot and E. Enk ibid. 1930 68 1635. 84 W. Manchot and J.Konig ibid. 1926 59 883. 6s W. Manchot and H. Gall ibid. 1925 58 2176 ; M. S. Kharagch and H. S. Isbell J . A ~ T . Chm. SOC. 1930 52 2918. W. Manchot and H. Gall BBT. 1926 58 232. H. J. Emel6us and J. S. Anderson " Modern Aspects of Inorganic Chemistry " 6' W. Manchot and J. Konig ibid. p. 229. Routledge London 1938 Chap. 12. ANDERSON CHEMISTBY OF TEE lKE!l'AL CARBONYLS w substances poorer in carbon monoxide being formed d k d y d e r mom drastic conditions or formed and stabilised through the replacement of CO groups by amine molecules. Whereas nickel and cobalt carbonyls are completely decomposed by free halogens iron pentacarbonyl forms fist the addition o0mpound.s Fe(CO),X, isolable only at low temperatures and then by e4pontttneonls loss of carbon monoxide the relatively stable tetracarbony1 halides." The reaction of iron tetracarbonyl with bromine is more complex,'@ and i t has been inferred that a mixture of Fe(CO)*Br with [Fe(CO),Br,J is formed although the second compound was not isolated.The iron carbonyl halides are non-electrolytes soluble in inert organic solvents. Their stability rises progressively in the order chloride < bromide < iodide as is evident fkom the comparison of decomposition temperatures photochemical sensitivity and thermochemical data 7l shown in Table VI where Q and &' are given in kg.-cal. TABLE vr X = Decomposition temp. of Decomposition temp. of FeX + 4CO -+ Fe(CO),X,. . . . . . Fe(CO),X,. . . . . . Fe(CO)*X + Q Fe(CO)& + CO + Q' Fe(CO) + X,(g) -+ Stability of Fe(C0),py2X . . Stability of Fe(CO),phenanX c1.- 35" + 10" Q = 17.9 Q' = 45.8 Non-existent Decomp. - 10" Br. - 10" + 55" 28.3 43.4 Decomp. 0" photosensitive Stable I. 0" + 76" 38-9 23.1 Stable Stable not photosensitive photosensitive Iron tetracarbonyl iodide is stable enough to sublime unchanged in a vacuum. If it is heated in hydrogen the lower carbonyl iodide Fe(CO),I is obtained as a bright red sublimate although complete decomposition (to Fe + 4CO) occurs extensively al~0.72 The same lower iodide is formed directly by the action of iron pentacarbonyl on iodine dissolved in boiling benzene. Another mode of partial decomposition hm been observed on heating Fe(CO),I in carbon dioxide. In addition to iron and carbon monoxide FeI and the unstable Fe(CO),I [compare Fe(NO),I '9 are formed in small amount. Decomposition of the latter compound furnishes the hitherto unknown bright red unstable FeI.The decomposition reaction (4 1) proceeds to completion at room temperature and is b'rought about by water or by pyridine in excess; it 69 W. Hieber and G. Bader Ber. 1928 61 1717. 7Q Idem Z. anorg. Chsm. 1931,201 329. 71 W. Hieber and A. Woerner 2. Elkktrochern. 1934 40 287. 7 2 W. Hieber and H. Lagally 2. awq. Chrn. 1940 $245 296. 73 J. S. Anderson and W. Hieber ibirt. 1933 all 132. 350 QUARTERLY REVIEWS may be compared with (4,2) which is typical of the aqueous decomposition (49 1) Fe(CO),I -+ FeI + 4CO (49 2) Pt(CO)Cl + H,O -+ 2HC1 + CO + Pt of the platinum-metal carbonyl halides. At high pressures ( 4 l ) is revers- ible ; a t room temperature the equilibrium pressure is probably about 6 stm. and ethereal solutions of ferrous iodide are slowly and completely converted into Fe(CO),I by carbon monoxide at high pressures.74 Copper and other metals react with iron carbonyl iodide forming cuprous iodide and the products of disproportionation of the Fe(CO) radical.75 Every step involved in the mechanism of the high-pressure synthesis (11 4 ; 11 5 ) can thus be realised for the iron compounds.Pyridine and other amines can effect partial replacement of carbon monoxide the substituted compounds 75 being relatively stable (Table VI). These reactions with others mentioned previously 76 are summwised in Table VII. TABLE VII #e(CO)py,I - FeI Mixed carbonyl halides Fe(CO),XY (X = I Y = C1 or Br) have been obtained by W. Hieber and A. Wirsching 77 by the action of IC1 or IBr on iron pentacarbonyl.These are intermediate in stability and properties between the corresponding simple carbonyl halides ; their chemical individu- ' Sn ' ality is perhaps not certain. Compounds such as SbCl or SnCl, which can act as halogenating agents react with iron pentacarbonyl forming (1.1 Fe( CO),SbCl and Pe( CO),SnCl respectively. The latter with a normal molecular weight in benzene may be represented as (I). The former has in benzene only half the expected molecular weight and as the electrical conductivity of its nitrobenzene solutions is negligibly small it probably dissociates into Fe(CO),Cl + SbCl,. c1 c1 (CO )4Fe C1 \ P \ c1 74 W. Hieber and H. Lagally 2. anorg. Chern. 1940 245 305. 7 s W. Hieber 2. Ekktrochern. 1937 43 390. ?ti W. 'ELieber and G. Bader 2. anorg. Chern. 1930,190,193.77 Ibid. 1940,24S 38. ANDERSON CHEMISTEtY OF !FHE METAL CARBONYLS 361. Sulphuryl chloride reacts additively with iron pentacarbonyl at low temperatures but effects complete decomposition at 0". It also mots with Fe(C0)JZ displacing both iodine and carbon monoxide. and forming the previously unknown compound Fe(CO),Cl,. The Constitution of the Metal Carbonyb.-(l) The constitution of the metal carbonyls has since the discovery of nickel carbonyl presented a problem in every formulation of a theory of valency. Their non-polar molecular constitution (which renders them volatile) and the diamagnetism of the simple carbonyls and all their substitution products stand in contra& with the properties of other classes of compound formed by the transition elements. Although the formal valency of the metals is not immediately apparent some simple systematic relation clearly exists between the atomic numbers of the metals and the composition of their simplest carbonyls.It is in the first place clear that the CO groups exist as such within the molecules and that these groups retain on the whole the bond character of the free carbon monoxide molecule. The former point is assured bythe ease with which carbon monoxide is liberated as such and by the partial replacement of CO by neutrd molecules-every stage of progressive replacement being capable of realisation with certain classes of compounds (cf. Table V). The second contention is supported by several lines of evidence. Nickel carbonyl has a zero dipole m~ment,~s which implies collinearity of the Ni-G-0 bonds.Unless the Ni-C and G O linkages are both double bonda (Ni=C=O) which would involve very high covalenoiiee in Fe(CO) or Cr(CO), the carbon and oxygen must be effectively triply linked (Ni-CkO) aa in carbon monoxide itself. That thb is so is con- firmed not only by the bond length but also by the strongest Raman frequency of nickel carbonyl 7Q9 (2039 cm.-l; cf. 2155 cm.-l in the Raman spectrum of carbon monoxide and the usual magnitude of double- and triple- bond frequencies). The CO group in the carbonyls can accordingly be regarded w a little- modified carbon monoxide molecule co-ordinated to a central metal atom as other neutral molecules or ions are linked in most types of co-ordination compound to central cations. The essential equivalence of CO and of ammino-groups is demonstrated in the chemistry of the carbonyls and also by the simplest carbonyl compound borine carbonyl BH,,CO which is precisely analogous to BH,,NMe or to BF,,NH:,.On this hypothesis the composition of the monometallic carbonyls is determined by the requirement that the effective atomic number (E.A.N.) of the central atom shall be made up to that of the next inert gas Fe(CO) Fe = 26 E.A.N. = 26 + 6 x 2 = 36 V. Ni(CO) Ni = 28 2 8 + 4 ~ 2 = 3 6 Mo(CO) M o = 42 42 + 6 x 2 = b4 Ru(CO) RU = 44 44 + 6 x 2 = 64 Os(CO) 0 s = 76 76 + 5 x 2 = 86 '* L. E. Sutton and J. B. Bentley J. 1933 652. '@ J. S. Anderson Nalure 1932 180 1002. J . Chern. Pibyah 1934 $3 636. 352 QUARTERLY REVIEW3 It is significant that the elements of odd atomic number-& Rh Re Ir- form no monometallic carbonyla.The requisite condition of molecular stability is reached by some further formation of co-ordinate links within bi-nuclear molecules. (2) Stereochmisty and b d character in the carbonyls. Pauling’s hypothesis that 4 or 6 equivalent a-type bonds are formed in co-ordination compounds by suitable hybridisation of d- s- and p-type orbitals makes the spatial arrangement of co-ordinated groups an index of the bond type. Particular interest therefore attaches to the detailed interpretation of the structure of the carbonyls. of the solid hexacarbonyls of chromium molybdenum and tungsten and electron-diffraction investigatiom of their vapours 82 have established the octahedral configuration of CO groups around the central atom. Few complexes of co-ordination number 5 are known and the configuration of iron pentacarbonyl was for some time in doubt.A trigonal bipyramidal structure should have zero dipole moment whereas that of iron pentacarbonyl is finite though small,83 suggesting either a tetragonal pyramidal molecule or non-equivalence of the FeCO bonds. The regular trigonal bipyramid as in PF, appears to be the only structure in accord with the electron-diffraction data. 84 For 4-co-ordinate compounds based on the Ni2+ ion a square planar configuration has been rigorously established. A. B. F. Duncan and J. W. Murray80 wrongly inferred from the Pvaman spectrum that the same steric arrangement obtained in nickel carbonyl but electron-diffraction measurements 85 clearly indicate that the molecule is tetrahedral as also are the nitrosocarbonyls and carbonyl hydrides isoelectronic with it (see below).Bond lengths deduced from the electron-diffraction data are summarised in Table VIII. The G O distances are those expected for :C:::O (cf. 1.13 A. in carbon monoxide itself). The M-C distances are however shorter in every case than the sum of the covalent radii of carbon and the central atom. Even if the covalent radius of triply-bonded carbon is uncertain by as much as 0.08 A. (cf. ref. 82) the residual discrepancies are siMcant. Brockway and his co-workers have accordingly regarded this as indicative of a real bond shortening due to resonance between the bond arrangements (I) and (11). Structure (11) has the merit of avoiding an accumulated nega- Crystal-structure studies M t - S O M=C=O tive charge on the metal atoms which have low electronegativities.Only in nickel carbonyl could all the CO groups be so linked to the metal; the W. Rudorff and U. Hofmann 2. physikal. Chem. 1935 B 28 351. 8 2 L. 0. Brockway R. V. G. Ewens and M. W. Lister Trans. Faruday SOC. 1938 8 3 W. Graffunder and E. Heymann 2. physikal. Chem. 1932 B 15 377 ; E. 84 R. V. G. Ewem and M. W. Lister Trans. Faraduy SOC. 1939 35 681. 86 L. 0. Brockway and P. C. Cross J. Chem. Physics 1935 3 828. 34 1350. Bergmann and L. Engel ibid. 1931 B 13 236. ANDERSON (JHEMISTRY OF THlc METAL CARBONYLS 353 TABLE VIII Inter-atomic Distances in -Metal Carbonyls from Electron-diflraction 1.15 1.15 1-16 1.16 1.13 1.14 1-16 1.16 1.15 Data (in 1.82 1.84 1.92 2.08 2.06 1-83 1 4 4 1-83* 1-82* A. units) Ni(CO) . . . . Fe(CO) . . . . Cr(CO) . . . . Mo(CO) . . . . W(CO) .. . . Co(CO),NO . . . Fe(CO),(NO) . . Co(CO),H. . . . Fe(CO),H . . . C-0. 1 M-C. I- I N-0. M-N. M-CY calc. 1.10 1.12 1.76 1.77 1 1-98 2.00 ' 2.02 ? ? 1.99 2.00 1-99 I 2.00 Bond shortening. - 0.16 - 0.16 - 0.10 ? ? - 0.16 - 0.16 - 0.16 - 0.18 * Average bond lengths taken t o be weighted means of M-GO and M-C-OH groupings. type (I) bonding-which determines the stereochemical propertiea in any case -must be increasingly important in Fe(CO) and the hexacarbonyls as the last column of Table VIII shows. Hieber 86 has suggested that the evidence for effective octavalency of nickel is not cogent and that the bond shortening might arise from a secondary interaction between the n-electrons of the C k O bond and the 3d electrons of the nickel atom. The accurate structure determination of [Fe(C~~-CH,),]C1,,3H~o,87 in which (see below) the C_N-CH group plays a similar r61e to that of C_O in the carbonyls does however substantiate the idea of resonance between (I) and (11).In deducing the bond type in the carbonyls and their derivatives isosteric relations with other classes of co-ordination compound may be noted. The :C:::O group the CN- ion :C:::N and the NO+ group :N:::O are isosteric and isoelectronjc. The Cr atom is (by the displacement rule) similarly isoelectronic with the Fez+ and Co3+ ions. Hence the electro- statically neutral Cr(CO) has the same electronic configuration and steric arrangement as EL number of diamagnetic complex cyano-a~ons (Table IX) all based on the formation of a-type (d%p3) covalencies about a IS core. I 06 W. Hieber Die Chemis 1942 M 25. H.M. Powell and G. W . R. Bartindale J . 1945 799. 354 QUARTERLY REVIEWS In the case of Ni(CO), the central atom is Nio 1S(KL39231,63d10) isoelectronic with Cu+ and Zn++. No 3d orbits are available for hybridisa- tion [to give four co-planar &pa bonds] so that in Ni(CO), &s in the complex ions [Cul(CN) J 3- and [Znn(CN)J a- a tetrahedral configuration is imposed by the formation of four (434ps) bonds. If the central metal atom cannot otherwise attain a closed electronic configuration (e.g. with metals of odd atomic number or with lower CO metal ratios than the maximum) polynuclear carbonyls are formed. The constitution of these has probably not yet been finally settled. N. V. Sidgwick and R. W. Bailey 88 suggested a general principle for formulating polynuclear carbonyls and nitrosyls based on the hypotheses (a) that all the metal atoms should acquire the effective atomic number of an inert gas and ( b ) that the CO group can form two collinear co-ordinate links.&'e,(CO) then receives the constitution (I). The isoelectronic CN- group can undoubtedly form two such collinear links-e.g. in crystalline AgCN,89 and in [(C,H,),AuCNI QO-but participation of CO in similar struc- tures would involve residual positive charges on the adjacent carbon and oxygen atoms and so is less likely. Brill's inference tn that the molecule of Fe,(CO) has trigonal symmetry (3) Polynucleur curbonyls. (CO),Fe f- CEO + Fe(CO) 88 Proc. Roy. SOC. 1934 A 144 521. go R. F. Phillips and H. M. Powell Proc. Roy. SOC. 1939 A 173,147. R. D. West 2. Krist. 1934 88 173 ; 1935 90 555.R. Brill 8. Krist. 1927 65 85. ANDERSON CHEMISTRY OF THE METAL CARBONYLS 355 haa been confirmed by a detailed X-ray crystallographic study,ee which ha,s shown conclusively that the structure is that shown in (11) ; the Sidgwick- Bailey rules are not apparently valid. The terminal CO groups are collinear with the co-ordinate links but the bridge CO groups are linked by the carbon atom to both iron atoms. Their G O distance (1-3 A.) is longer than that in the terminal CO groups and approximates to that of the C=O group. Whilst the geometry of the molecule is thereby established inferences as to the bond character must be critically examined for their wider implications. The obvious conclusion that each iron atom forms three ordinary covalencies [in place of the effective zero-valency of iron in Fe(CO),] gives to it the effective atomic number 35 and.leaves it with one unpaired electronic spin. The diamagnetism of Fe,(CO),93 can then be explained only by the supplementary hypothesis of spin coupling between the unpaired electrons of the two relatively closely spaced iron atoms. An analogous formulation of Co,(CO) as (111) would present the same difficulty. 0 II 0 (111.) 0 0 0 II I1 C C C / \ 7 1 Ill I M M M M M M If however the bond pattern advanced for Fe,(CO) is extended to Fe,(CO),, Co,(CO),, etc. the problem is more acute and supplementary hypotheses are less plausible. Detailed structure analyses for these com- pounds are lacking but R. Brill considers that the space-group data for Fe,(CO), indi- cate the two-fold symmetry of structure (VII).(CO),Fe Fe Fe(CO) This structure on the Powell and Ewens model would confer a lS configuration on the outer iron atoms but the central iron atom would be quadrivalent with E.A.N. = 30; the molecule would certainly be paramag- netic. Similar difficulties arise for the “ mixed carbonyl ” derivatives of iron and cobalt carbonyl hydrides (cf. 111 6). Any distinction between two substantially different h d s of metal-CO binding within the polynuclear carbonyls seems out of harmony with the general chemistry of these compounds as compared with tho simple carbonyls. In this connection the considerations of K. A. Jensen 95 on the constitution of polynuclear complexes are of interest. He suggests for the linking of the co co . . . . . . . . co co (VII.) O 2 H. M. Powell and R. V. G. Ewem J.1939 286. 9s W. Klemm H. Jacobi and J. Tilk 2. anorg. Chem. 1931 201 1. m4 2. KriBt. 1931 77 36. 96 2. a w q . Chem. 1944 252 234. 356 QUARTERLY REVIBWS bridge groups a resonance between the forms (IV) (V) and (VI) ; Fe,(CO) would then be regarded as a resonance hybrid of (VIII) and (IX) which would fit the observed bond lengths and would be compatible without supplementary huypotheses with the observed diamagnetism. 0 0 O r C / \A/ C 2 CrO II 0 (VIII.) I/ 0 (IX.) (4) Nitrosocarbonyls and carbonyl hydrides. The structural principle underlying the existence and composition of the two series Fe(CO),(SO) Co( CO),SO Si(CO) Fe( CO),H Co( CO),H Si(CO) is clearly the attainment of the same closed electronic configuration through- out-that of the nickel carbonyl molecule. In the nitrosocarbonyls (:X:::O:) + groups replace (:C:::O:) groups progressively as the diamagnetic susceptibilities show.The KO+ groups arise from the transfer of the odd electron of the nitric oside molecule to the iron or cobalt atoms which thereby become “ pseudo-nickel ” atoms (OEC),-Si (OrC),-CO-X~O ( OEX),-F+(S_O) 8 + 28 = 36 6 f 27 + 1 + 2 = 36 4 + 2 6 + 2 + P = 3 6 Electron-diffraction data 96 (Table VIII) show that the nitrosocarbonyls are closely isosteric with nickel carbonyl. The somewhat enhanced polarity of the nitrosocarbonyl structure is reflected in their lovcer volatility 97 and other physical properties. The carbonyl hydrides have also been proved to have the tetrahedral Ni(CO) structure but the mode of linkage of the hydrogen 84 is perhaps not certainly established by the electron-diff raction data.Hieber’s view 38 86 that the hydrogen atoms are incorporated in some way as protons within the core of the cobalt and iron atoms (which become “ pseudo-nickel ” atoms) is difficult to accept as a physical model though it is not very Werent in principle from K. S. Pitzer’s recently advocated borane structure.98 Ewens and Lister 84 suggested that an electron is transferred to the metal atom and that a proton is linked to the lone pair of electrons of the CO oxygen atom. The resulting group (:C:::O:H)+ would be isoelectronic with and not too different sterically from the *O L. 0. Brockwsy and J. S. Anderson Trans. Famday Soc. 1937 33 1233. O7 W. Hieber and J. S. Anderson 2. anmg. Ckm. 1932 208 232. J . Amer. Chm. Soc. 1946 67 1126. ANDERSON CHEMISTRY OF THE METAL CARBONYLS 357 (:N:::O:)+ group 80 that the carbonyl hydrides are the analogues of the nitrosocarbonyls.However Hieber has pointed out that the formulation involves in effect a quaternary oxonium group which would be unique ; this should be stabilised by alkylation whereas the carbonyl hydrides form no Bkyl derivatives under any conditions yet found. Moreover no cogent evidence hrts been advanced that there are two types of M-C and C-0 bond dimension within the molecules. This question should therefore be regarded &g open. Critical consideration of the structural problems presented by this whole group of compounds serves to lay salutary emphasis on the caution that is needed in drawing conclusions from purely metrical data as to the intimate constitution of molecules of unusual type.

ISSN:0009-2681    

DOI:10.1039/QR9470100331

出版商:RSC    

年代:1947

数据来源: RSC

摘要:

THE ALIPHATIC NITRO-COMPOUNDS I Nitromethane . 101.2" Nitroethane . . ~ 114.0 l-Nitropropane . . 131-6 I By N. LEVY PH.D. and J. D. ROSE B.A. B.Sa. (IMPERIAL CHEMICAL INDUSTRIES LIMITED) Methyl nitrite Ethyl nitrite n-Propyl nitrite ALTHOUGIH dinitroethane was fist prepared in 1864 and a general method for nitro-paraffins devised in 1872 progress was rather restricted until recently by the difficulty and expense of the preparative methods none of which was suitable for large-scale manufacture. The development during the past decade of a process for the vapour-phase nitration of the lower paraflhs due primarily to the work of H. B. Hass and his collaborators in the United States has stimulated renewed interest in the chemistry and applications of nitro-paraffins. Comparison of the reviews of T.W. J. Taylor and W. Baker and H. B. Hass and E. Riley 2 shows the substantial advances made. Since the latter review publication has been sustained and in particular the hitherto obscure reaction between olefins and oxides of nitrogen has been elucidated by recent work in this country. New nitro- aliphatic compounds for example the lower dinitro-paraffi have thus become available and some novel reactions including those of addition to the nitro-olefins disclosed. A fresh outline embodying later work therefore seems desirable. Nitro-parafins Prepccration.-(a) Indirect metAocE8. Until the advent of the vapour- phase methods described below the preparation of the lower nitro-paraflfins largely depended on the methods of Victor Meyer 3 of which were publkhed in 1872. The Victor Meyer the reaction between alkyl iodides and silver nitrite and H.Kolbe,' both method makes use of R*NO + AgI Ir RI + AgNO' ROONO + AgI 'I and was h s t applied with amyl iodide. The method is a general one yielding a mixture of nitro-paraffin and alkyl nitrite in proportions which depend on the nature of both the halogen and the alkyl group. Separation of the isomerides is easy because of the appreciably higher boiling point of the nitro-paraffin as illustrated in Table I TABLE I Nitro-paramn. Alkyl nitrite. 1 R.p. I-.-.-.- I I Sidgwick'a " Organic Chemistry of Nitrogen " Oxford 1937 pp. 229247. C'hem. Revkw8 1943 32 373. * V. Meyer and 0. Stuber Ber. 1872 5 203. J . pr. Chem. 1872 6 427. 358 LEVY AXD ROSE THE AUPHATIC NITRO-COMPOUNDS 369 The method which has been extended by R.B. Reynolds and H. ad kin^,^ is applicable to the synthesis of nitro-paraffin derivatives such as nitro- ketones nitro-esters nitro-alcohols and nitro-olefins other than Aa-nitro- o l e h e.g. 1 -nitroprop-Z-ene from ally1 bromide. The reported preparation of 1 2-dinitroethane from ethylene iodides is now questionable.47 Although silver nitrite is commonly used the method has been applied to mercury and alkali nitrites. Apart from halides other than iodides mono- and di-alkyl sulphates and alkyl p-toluenesulphonates have been employed without cheapening the process sufficiently for commercial scale operation. Kolbe's method is based on the reaction between a-halogeno-acids and sodium nitritein aqueous solution and was fist applied to nitromethane H,O CH,Cl*CO,Na + NaNO + NaCl + CH,(NO,)*CO,Na --+ CH,*NO + NaHCO Nitromethane can be obtained thereby in 55-70% yield on chloro- acetate depending on the nitrite chloroacetate ratio.' The main sources of loss appear to be hydrolysis of the chloroacetate to glycollate and decom- position of the nitroacetate first formed.V. Auger has reported 50% yields of the homologous nitro-parah using a-bromopropionic a-bromobutyric and a-bromoheptoic acids. The reaction fails when applied to acids in which the bromine is attached to a tertiary carbon atom small yields of the pseudo- nitrole being obtained for example from a-bromoisobutyric acid (CH,),CBr*CO,H + (CH,),C(NO,)*NO Other early methods include the oxidation of primary amines (methyl- amine ethylamine and tert.-butylamine) CH,*NH,+ CH,*NH*OH-+ CH2=N*OH + CH,=N -+ CH,*NO C H and the reaction of halogeno-nitro-paraffin8 with zinc alkyls ; for example the formation of tert.-nitrobutane from chloropicrin and dimethylzinc.lo l1 The ease of preparing alkyl nitrites would make isomerisation methods attractive but the claims made 12 13 have not been substantiated by extensive repetition of the work both in the vapour phase and in solution.14 Means of preparing the higher nitro-paraffins from the lower homologues are described in a later section. Now that alternative routes to nitro- olefins are available it should be mentioned that they can be converted into lower nitro-paraffins by oxidative or hydrolytic fission \ \ / / C==CR*NO2 + H,O + CO + R*CH,*NO For example nitromethane is formed in almost quantitative yield from J . Amer. Chem. Soc.1929 51 279. V. Ipatov J . Russ. Phys. Chem. SOC. 1917 49 297. I.C.I. private communication. @ E. Bamberger and R. Seligman Ber. 1902 35 4299. lo I. Bevad ibid. 1893 26 129. l2 P. Neogi and T. Chowduri J . 1916,109 701. lS Ihm J. 1917 111 899. * Bull. SOC. chim. 1900 23 336. u Idem J . pr. Chem. 1893 48 346. 1' I.C.I. private communication. 360 QUARTERLY REVIEWS a-nitrotkobutene,m nitroethane from 2-nitrobut-2-ene and nitroneopentane from 3-nitro-2 4 4-trimethylpent-2-ene.60 The only remaining method of any importance for the introduction of the nitro-group into aliphatic compounds is nitration of active methylenic substances with ethyl nitrate and sodium methoxide. The synthesis of phenylnitromethane is the best example of this process 15 Et .O.NO * NaOH Ph*CH,.CN ___+ Ph*C(CN):NO*ONrt + NaOEt acid Ph*C(CO,Na):NO*ONa + Ph*CH,*NO + CO and the reaction has also been used for the synthesis of nitrosulphones.16 An interesting paper by W.Steinkopf and M. Kuhnel 17 describes the addition of nitryl chloride (N0,Cl) to unsaturated compounds. Although ethylene is merely chlorinated by this reagent vinyl bromide is converted into 1 -chloro-l-bromo-2-nitroethane with such vigour that unless the reaction is carried out at low temperature and in dilute solution violent explosion occurs. Similarly 1 2-dichloroethylene gives 1 1 2-trichloro- 2-nitroethane and tetrachloroethylene yields pentachloronitroethane. Phenylacetylene and nitryl chloride afford 1- chloro-2-nitro-2-phenylethylene. The direct nitration of liquid paraffins was first reported with petroleum fractions by F.Beilstein and A. Kurbatov in 1880 since when a great deal of work has been reported on paraffins containing five or more carbon atoms. Procedures have been extremely varied and some confusion has therefore arisen the main products and side-reactions .being affected by the temperature and nitrating agent employed. The reaction has been carried out with various strengths of nitric acid with sulphuric acid-nitric acid mixture and with oxides of nitrogen and the reagents have been refluxed together a t atmospheric pressure or heated in sealed tubes at higher temperatures. Certain regu- larities have nevertheless emerged. Tertiary carbon atoms are most readily attacked appreciable reaction occurring at little above room temperature l8 and in general ease of forma- tion follows the order tertiary >secondary > primary nitro-paraffin.Usually the reaction is slow due to the low mutual solubility of the acid and the paraffin. Reaction times vary from a few hours to several days depending on the technique (temperature and pressure) and the hydrocarbon. As would be expected the longer reaction times are associated with atmospheric pressure nitration of the lighter normal paraffins l9 and the shorter reaction times with the higher temperatures obtainable in the nitration of heavier paraffins or in nitration under pressure.20 2l The mononitro-paraffin first formed is more soluble particularly in concentrated acid and undergoes further reaction to form on the one hand polynitro-paraffins and on the other hand decomposition products of ( b ) Liquid-phase nitration of parapns.Org. Synth. 19 73. 17 Ber. 1942 75 1323. 19 R. A. Worstall J . Arner. Chem. SOC. 1898 20 202. 20 M. Konovalov J . Russ. Phys. Chem. SOC. 1894 25 472. 21 Idem J . 1894 65 265. 16 F. Arndt and J. D. Rose J. 1935 1. 18 W. Markovnikov ibid. 1899 32 1441. LEVY AND ROSE THE ALIPHATIO NITRO-COMPOUNDS 361 secondmy and primary nitro-parafhs. Some direct oxidation &o OCCWB. Hence the product consists of a mixture of mono- and poly-nitro-paraffins together with oxides of carbon fatty acid alcohols etc. arising from oxidation and hydrolysis reactions. Through these side reactions part of the nitric acid is lost a,s nitrogen and nitrous oxide whereas in vapour- phme nitration practically quantitative recovery as nitro-paraffin and nitric oxide is obtained.Dilute nitric acid seems preferable to concentrated since it facilitates the use of pressure and the attendant higher reaction tempera- tures faster nitration and better yields being obtained for example 60% yield of nitrohexanes from treatment of n-hesane a t 10'5 in a sealed tube with dilute nitric acid 20 21 compared with 6% by refluxing a t atmospheric pressure with concentrated acid.lg Mixed acid is less reactive than nitric acid 1% 22 and its sulphuric acid content presumably promotes more rapid hydrolysis of the mononitro-paraffins. Concentrated nitric acid favours oxidation and polynitro-compound formation as well as the formation of primary and secondary nitro-paraffins a t the expense of tertiary.23 24 This lends some support to the claims of R. Worstall 1 9 9 25 to have obtained exclusively I-nitro-paraffins from C,-C normal paraffins ; on the other hand the identification has been criticised,26 and the relatively low tempera- tures employed should favour secondary derivatives though these might not persist as mononitro-paraffins .An interesting method providing a link with fully vapour-phase nitra- tion has been employed by C. Grundmann 27 for ClO-Cl8 normal paraffins as well as higher petrol fractions. The paraffin is caused to react a t 140- 210" with aqueous nitric acid vapour or nitrogen dioxide preheated to the same temperature ; up to 60% of the paraffin is thus converted into nitro- compounds and only l-8% into fatty acid for hydrocarbon acid ratios of 2 1 to 1 2 and reaction times of 3-5 hours. The proportion of mono- nitro-paraffin in the product increases sharply with the hydrocarbon acid ratio.The original claim that the method gave largely 2-nitro-paraffins was later contested and retracted.28 29 It is interesting to speculate how much of the reaction in this case and in the earlier work under pressure actually occurred in the vapour-phase. The use of solvents is limited because of attack by the acid a t tempera- tures required for nitration. Nitrates increase the reaction velocity by raising the boiling point of the nitric and reducing agents have been claimed as catalysts for liquid-phase nitrati0n.3~ The introduction of negative groups into a hydrocarbon residue facilitates nitration. M. Konovalov 3 2 9 33 showed that chloro-derivatives of butane and pentane are 2 2 W. Markovnikov J . pr. Chem. 1899 59 556.a3 M. Konovalov J . Russ. Phys. Chem. SOC. 1899 31 57. 24 Idem ibid. 1906 38 1 109 124. 2s J . A m r . Chem. SOC. 1899 21 210 223 233 237 ; 1900 22 164. 26 L. Henry Rec. Trav. chim. 1905 24 352. ae F. Asinger ibid. p. 323. 30 C. Denison Thesis Purdue University 1940. 32 M. Konovalov J . Rws. Phys. Chem. SOC. 1904 36 220 537. 33 Idem ibid. 1906 38 607. 27 Die Chemie 1943 56 159. 29 C. Grundmann loc. cit. 3 t R. Senkus U.S.P. 2,332,491. 362 QUARTERLY REVIEWS more reactive than the parent hydroccwrbons. Acetoacetic ester is readily nitrated with fuming nitric acid in acetic anhydride to give ethyl nitro- a~etate,~4 whilst malonic ester with fuming nitric acid alone gives nitro- malonic ester in good yield. The greater reactivity of tertiary carbon atome and the effect of sub- stituents and of activating groups suggest that the nitration reaction is electrophilic in character.That fiee radicals are not involved is indicated by the formation of optically active 3-nitro-3-methyloctane by the liquid. phase nitration of Z-3-methyloctane.86 ( c ) Liquid-phase nitration ofolefins. Since the work of A. Semenov 8' over 80 years ago on the reaction between ethylene and apparently pure dinitrogen tetroxide persistent attempts have been made to achieve satis- factory preparation of nitro-aliphatic compounds from o l e b and oxides of nitrogen or nitric acid. H. Wieland 3 8 9 39 has suggested that nitric acid adds to the double bond as HODNO, giving rise to the nitro-alcohol which in turn forms the nitrate ester and the nitro-olefin HO.NO8 H NO/ (CH,),C=CH __+ (CH,),C(OH)*CH,*NO + -HNO (CH,),C( O=NO,)*CH,*NO __+ (CH,),C=CH*NOt Although this is supported by the formation of 2-nitroethanol from ethylene and fuming nitric acid,40 an alternative mechanism based on addition of nitrous gases formed in situ by the accompanying oxidation reactions has been proposed by A.Michael and G. H. Carls0n.4~ Satisfactory yields with nitric acid still remain to be achieved. The position arising from the use of oxides of nitrogen was until recently no less confused as indicated in previous reviews.1 2 4 Semenov undoubtedly prepared 1 S-dinitro- ethane although calling it " ethylene dinitrite." Dinitro-compounds have also been isolated from the product with fully-substituted ethylenes.43,*4 The formation of nitrosites nitrosates and nitro-olefins has been reported and that of dinitro-paraffins inferred,d6 d6 but in general the greater part of the product resisted identification and dinitro-parafis had not been previously isolated from the simple olefins.This situationhas been transformed by a recent series of papers 4794*,49950,51 s4 L. Bouveault and A. Wahl Bull. SOC. chim. 1904 31 847. sK P. G. Stevens and R. W. Schiessler J . A m r . Chem. SOC. 1940 62 2886. ae Jahresbericht 1864 480. 37 2. Chem. Phrm. 1864 7 129. 39 Ibid. 1921 54 1770. 40 3. McKie J . 1927 962. 4a J. L. Riebsomer Chm. Revkwe 1946 36 157. 48 H. Biltz Ber. 1902 35 1628. 44 A. Michael and G. H. Carlson J . Org. Chem. 1940 5 14. 46 Idem J . Amer. Chem. SOC. 1937 59 843. 47 N. Levy and C. W. Scaife J. 1946 1093. 48 N. Levy C. W. Scaife and A. E. W. Smith J. 1946 1096.40 N. Levy and C. W. Scaife J. 1946 1100. Ber. 1920 53 201. J . Amer. Chern. Soc. 1935 57 1268. I h J . Org. Chem. 1939 4 169. Idem in the press. 61 Idem in the press. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 363 dealing with the additiorl of dinitrogen tetroxide to ethylene propylene isobutene but-1-ene but-2-ene 2 4 4-trimethylpent-1- and -2-ene and cyclohexene in the liquid phase. The work has shown that under appro- priate conditions high overall yields of dinitro-compound nitro-aloohol and nitro-nitrate can be obtained and isolated in pure condition by safe methods. It has also been shown that much of the earlier work was vitiated by (i) the use of impure tetroxide or " nitrous fumes " which gave rise to nitroso- compounds and increased the instability of the mixed product (ii) failure to control oxidative side-reactions resulting in formation of dinitrogen trioxide and subsequent addition (iii) the variable effect of reaction medium in certain cases decisive and (iv) decomposition of products ofhn with violence during attempted separation of the products rendering recovery and identification unsatisfactory.The methods used to overcome these difficulties are described below. The reaction is best carried out in the liquid phase at - 10" to + 25". Gaseous olefins are absorbed in pure dry dinitrogen tetroxide as liquid or concentrated solution the reaction mixture being cooled and liquid olefins simply mixed by dropwise addition. Ethylene reacts fairly slowly 10 hours being required for the optimum absorption of 0.4 mole per mole of tetroxide.Higher olefms react more rapidly and molar equivalents can be made to react completely in 1-2 hours. Propylene isobutene and the normal butenes require an ether or ester type of solvent since otherwise oxidation supervenes and no dinitro-paraffin can be isolated. Oxidation reactions and interference by dinitrogen trioxide addition can always be avoided by use of excess tetroxide or where convenient the addition of gaseous oxygen. Under such conditions addition of tetroxide to the double bond takes only two forms (i) as two NO groups to give dinitro-compounds and (ii) as one NO group and one 0-NO group to give nitro-nitrites. Some oxida- tion of the latter to nitro-nitrate occurs in the presence of oxygen but the product always contains a substantial proportion of nitro-nitrite which being unstable tends to decompose sometimes explosively during removal of excess of tetroxide or solvent.It has been found however that the nitro- nitrites are readily converted into the quite stable nitro-alcohols by treat- ment in the cold with water or a lower aliphatic alcohol and this is the most essential feature in safe and efficient recovery of nitrated products. Prefer- ably solvent and excess of tetroxide are removed by continuous evaporation from a warm falling film the stripped material falling immediately into cold water or methanol I I NO NO I ' I O*NO' NO oxidn. C < + I I 0-NO NO >C I I OH NO (nitro-alcohol) (dinitro-paraffin) (nitro-nitrate) 364 QUABTEBLY REVIEWS Subsequent separation of the dinitro-paraffid nitro-alcohol and nitro- nitrate is not =cult and varies to some extent with the olefin ; for example dinitroethane and dinitroisobutane are crystallised by cooling the methyl alcohol solution of the product water extraction is used for recovery of all nitro-alcohols up to CI and vacuum fractional distillation covers all other separations and ultimate purification.Total yields of separated products are 65-85% the deficiency being mainly due to separation losses inherent in the small scale. Except where dinitrogen trioxide is deliberately used or allowed to accumulate nitroso-compounds or their derivatives are not formed. As a result of this work the following generalisations can be made (i) The two modes of initial addition occur to a similar extent and only minor changes are involved in passing from one olefin to another.(ii) In the formation of nitro-nitrites from unsymmetrical olefins (propylene kobutene but- 1-ene and the trimethylpentenes) the nitro- group invariably attaches to the carbon atom with the greater number of hydrogen atoms and the nitrite group to the other side of the double bond ; for example CH3*CH=CH2-CH *CH( 0 *NO )*CH,*NO ( CH,),C=CH2--+( CH3),C( 0-NO )*CH,*NO W " O .NO) It is therefore suggested that the reaction is electrophilic and concerns polar forms of dinitrogen tetroxide ; initial attack is by the positive nitro-group followed by release and attachment of the remaining group as -NO or -O=NO n + >C=CH + NO,-X-+ >C-CH*NOg---- -X -+ >CX-CH*NO I I 1 I (X=NOp 0-NO or NO.) (iii) Parallel experiments on the behaviour of dinitrogen trioxide in ether showed that i t behaves as (NO,)(NO) with the nitroso-group like the nitrite group attaching to the carbon atom with + - fewer hydrogen atoms.The nitrosonitro-paraffin is usually isolated as the solid dimer although in the Me-C-C*NO It II N N case of propylene it appears to be converted into \ / 4-nitro-3-methylfurazan oxide (I). At the same time the trioxide gives rise to tetroxide by dissociation and hence to dinitro-paraffin and nitro- nitrite. (iv) Th e solventeffect noted above was most completely studied with isobutene. Liquid dinitrogen tetroxide alone hydrocarbons and chlorinated solvents promote oxidation among the products of which the nitrate ester of a-hydroxyisobutyric acid [(CH,),C( 0-NO,) *C02H] is usually found and the &nitro-paraffin cannot be recognised. Symmetrical ethers dioxan benzyl methyl ether methylal acetal tetrahydropyran tetrahydrofuran (1.) LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 365 and acetate esters all permit normal addition.Formic esters act as inter- mediate types allowing some oxidation as well as normal addition. Substi- tuted ethers with reduced electron availability at the oxygen atom are ineffective for normal addition. This gradation of solvent effects and the isolation of a solid 1 1 complex of dioxan with dinitrogen tetroxide indicate that effective solvents operate through molecular association with the tetroxide and thereby moderate its oxidising tendency. (v) Nitro-nitrate is formed in relatively small amount unless oxygen js deliberately added to promote stability and obviate the formation of dini- trogen trioxide.Nitro-nitrate appears to be formed at the expense of dinitro-compound as the oxygen rate is increased. Apart from the importance of this work in clearing up a confused field in organic chemistry it has made available a series of compounds which were formerly unknown or extremely difiicult of access. Their value in further syntheses will be illustrated later in this review. (d) Vapour-phase nitration of parafins. Until 1934 when the fist vapour-phase nitration of a saturated paraffin was claimed by H. B. Hass and his ~ollaborators,5~ a process of this type seemed unlikely the only other positive result being a 2% yield of primary nitro-paraffins from reaction between normal C,-C paraffins and nitric oxide over an alumina-ceria- thoria catalyst.63 The Hass process operated since 1940 by Commercial Sol- vents Corporation on a moderate works scale originated with trial of liquid- phase nitration of isobutane under pressure with ultimately greater succesa at temperatures above the critical for isobutane.2 Succeedjng investigations by the Hass school are described in numerous papers since 1936 and a series of patents covering process detail ; the more important of these are referred to below. Developments in this country are indicated in patents assigned to Imperial Chemical Industries Ltd. Following the nitration of isobutane under pressure ethane propane and the butanes were nitrated in a flow apparatus at atmospheric pressure and 420° using 68% nitric acid a hydrocarbon acid ratio of 2 1 and a contact time of 2 seconds.64 Pass conversions of nitric acid into mixed mononitro-paraffins ranged from 9% for ethane to 28% for isobutane ; methane was apparently unattacked.It was later found by G. K. Landon 65 66 that pressures above atmospheric were required for the nitra- tion of methane. Reaction under pressure also gave improved results with ethane and pressures of 10 atmospheres or more are employed in the commercial process with propane. Nitration of the lower paraffins gives rise to all the mononitro-paraffins possible by G-H and C-C fission ; thus ethane gives nitromethane as well as nitroethane propane gives both these as well as 1- and 2-nitropropane n-butane gives nitromethane nitroethane 1 -nitropropane and 1 - and 69 H. B. Hms E. B. Hodge and B. M. Vanderbilt U.S.P. 1,967,667 (1934). 6 8 Platonov and Shaikind J . Gen. Chem.Russia 1934 4 434. 64 H. B. Ham E. B. Hodge and B. M. Vanderbilt Ind. Eng. Chem. 1936 28,339. 6 5 U.S.P. 2,161,475 (1939). 67 H. J. Hibshman E. H. Pierson and H. B. Hass Ind. Eng. Chem. 1940 31 427. 68 U.S.P. 2,164,774 (1939). 366 QUABTERLY REVIEWS 2-nitrobutane and kobutane give nitromethane 2-nitropropne and 1 - and 2-nitroisobutane. Similar results obtain with normal pentane ~i13 and iso- ~ e n t a n e . ~ ~ The proportions of the various nitro-paraffi in the product are affected by temperature the reaction being more selective at lower temperatures when ease of substitution follows the order tertiary > secondary > primary hydrogen atoms. As the temperature is raised more primary and lower nitro-paraffi are formed. This is illustrated in the nitration of propane with nitrogen tetroxide at temperatures of 250-800°.g0 Over this range the 2-nitropropane content diminishes from 72% to 23.5% while the other components increase as follows nitromethane from 14% to 236y0 nitroethane from 3% to 23+iY0 and l-nitropropane from 11% to 29.5%.In manufacturing practice this degree of flexibility is limited by considerations of yield the competing oxidation reaction becoming more marked with rising temperature and on this account it is necessary to control temperature precisely. Operation under pressure requires higher hydro- carbon ratios and shorter contact times for the same reason ; for example the nitration of ethane at 7 atmospheres and 450-470" was carried out with an ethane nitric acid ratio of 10 1 and a contact time of 0.23-0.33 second 67 compared with a ratio of 2 1 and a contact time of 2 seconds a t atmospheric pres~ure.~4 Dinitrogen tetroxide can be used under similar conditions but better control and conversions can be achieved at substantially lower temperatures and longer contact times together with higher partial pressures of nitrating agent.Sl The main advantage of these conditions is that the lower tempera- tures and moderate rate of heat evolution permit a more positive control of reaction temperature for example by heat transfer to a boiling liqud.The process is a cyclic one that is after passage of the mixed vapours through the reactor nitro-paraffins and water are condensed and the spent gas (containing excess of hydrocarbon and nitric oxide oxides of carbon and other minor constituents) is treated to remove oxides of carbon.Paraffin and nitric oxide after re-oxidation by suitable means are returned to the converter with fresh paraffin and nitrating agent. The crude nitro-paraffin mixture may be steam-distilled first and then fractionated in a series of stills operated under partial vacuum. When propane and nitrogen tetroxide are used at about 360" and 10 atmospheres pressure the product contains 20-25% of nitromethane 5-10yo of nitroethane 45-55% of 2-nitropro- pane and 20% of 1-nitropropane in total yield of 7 5 - 8 0 ~ 0 on the propane consumed in the converter and over 90% on the tetroxide.sl Acetone is also formed in small amount and up to 1% of the nitro-paraffi mixture is 2 2-dinitropropane. The main source of loss is oxidation of the hydro- carbon ultimately to carbon monoxide and dioxide though the reduction of acid or tetroxide stops short at nitric oxide.The oxidation reaction is a chain process with the reactor walls playing an important part and it 68 H. B. Hass and J. Patterson I d . Eng. Chem. 1938 30 67. ' 0 L. W. Siegle and H. B. Hass ibid. 1939 31 648. 60 H. B. Hass J. Dorsky and E. B. Hodge ibid. 1941 33 1138. 61 I.C.I. private communication. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 367 is possible that nitro-paraffins of lower carbon number than the original hydrocarbon are by-products of branching oxidation chains. The use of stainless steel reaction vessels with nitric acid although initially satisfactory leads to progressive increase in the oxidation reaction at the expense of nitration. In a number of patents assigned to Commercial Solvents Corporation intermittent washing or coating with a number of inorganic salt solutions 62 63 or the injection of small amounts of sodium or potassium nitrate with the acid 64 is claimed to control this effect while G.K. Landon 5 5 9 56 mentions precious metals and " Pyrex " for the reaction vessel. The precise r6le of catalysts in this reaction is not known but arsenic- or antimony-containing glasses,65 lead or lead-containing g1ass,e6 and borosilicate glass 68 have all been claimed to give improve- ments. Recent data 69 on German war-time research on the synthesis of nitro-paraffins show no new features. The formation of dinitro-paraffins in vapour-phase reactions may well be a function of the use of dinitrogen tetroxide and the associated longer contact times albeit a t lower temperature,70 '1 though it is unlikely that solely l-nitro- and 1 l-dinitro-paraffins are formed.s0 72 Reaction between isobutane and dinitrogen tetroxide a t 200-250" yields a product up to half of which is 1 2-dinitroisobutane the remainder being tertiary nitrobutane mixed with smaller amounts of lower mononitro-paraffins and 2 2-dinitro- propane.73 If the mixed product is steam-distilled much of the dinitro- butane is lost by conversion via a-nitroisobutene into acetone and nitr~methane.~~ More recently Hass and his school have extended their investigations to miscellaneous higher paraffins and cycloparaffins up to A free-radical mechanism has been proposed for the nitration reaction 76 which accounts for the specific mononitro-paraffins formed in all cases ; small amounts of the expected olefins are also found.On the other hand it seems unlikely to explain the proportions of the various products or the fact that the nitration of isobutyl chloride 77 leads to mononitro-derivatives of the parent substance (l-chloro-3-nitro- l-chloro-2-nitro- and l-chloro-l- nitro-2-methylpropane) l-chloro-2-nitropropane and nitromethane but no q . 7 4 75 6% U.S.P. 2,236,905 (1941). U.S.P. 2,260,258 (1941). 66 K. W. Gee and I.C.I. Ltd. B.P. 576,129. 67 M. P. Applebey and I.C.I. Ltd. B.P. 678,044. 68 N. Levy A. E. Rout K. W. Gee and I.C.I. Ltd. B.P. 586,203. 6s U.S.P. 2,236,906 (1941). 6 6 N. Levy and I.C.I. Ltd. B.P. 676,733. S. Masterman and D. B. Clapp C.I.O.S. Item 22 File No. XXXIII-41 'O T. Urbanski and M. Slon Rocznicki C h m . 1936 16 466 ; Compt.rend. 1936 71 Idem Rocznicki Chem. 1937 17 161. 7 2 M. H. Danzig and H. B. Hass J. A m . Chem. Soc. 1944,66,2017. 73 I.C.I. private communication. 74 A. P. Howe and H. B. Hass Id. E n g . Chem. 1946 38,251. 76 R. T. Bickerstaff and H. B. Ham J . Amer. Chem. SOC. 1946 68 1431. 76 R. F. McCleary and E. F. Degering I n d . E n g . Chem. 1938 30 64. 77 T. J. Oleszko and E. T. McBee Thesis Purdue University 1939. H.M.S.O. 1947. 203 620. 368 QUARTERLY REVIEWS chloronitromethane or 2-nitropropane. by Hass for example H -C +HzO -C- The alternative theory mentioned of complex formation between nitric acid and the hydrocarbon \ I I / \ I I/ -c + c- - C d \ I \ I I I 1 I I 1 ; + I NO and ; - I NO OH I NOz-OH NO,-OH while providing for the formation of alcohols small amounts of which survive also fails to explain the above case unless complex formation is itself electrophilic and affected by neighbouring substituents.It would be interesting to see this tried and the decisive test applied by Stevens and Schiessler 35 to liquid-phase nitration used with vapour-phase examples. N. L. Nitromethane CHI 'NO,. 1.139 - 29" 101.2" Physical Properties.-A summary of the physical properties of the commoner nitro-paraffins is given in Table II.78 Nitroethane CHj *CHa*NOa. 1-052 - 90" 114.0" TABLE II 27-8 1.3818 9.5 2.2 4;:. . . . . . 15.6 1.3916 4.5 0.9 M.p. . . . . . . 7-65 1.4015 1.4 0.5 B.p. . . . . . . 12.9 1.3941 1.7 0.6 Vap. press. (mm. at 20") . . . n2O0 . . . . . . Solubility at 20" C.C. in 100 C.C. of water C.C. of wafer in 100 C.C. 1 I 1.003 0.992 - 108" I - 93" I I I 131.6" I 120.3" I All the mononitro-paraffins are colourless when pure and have dipole moments of about 3-2.They are miscible with most organic solvents and possess considerable solvent power for many organic substances. This wiII be dealt with more fully in the section of this,review dealing with their u8e.s. Of the mononitro-paraffins only nitromethane can be detonated with a cap ; the dinitro- and polynitro-compounds being more in oxygen balance are more sensitive and explode more easily. In view of the known considerable toxicity of many aromatic nitro- compounds it is surprising that the nitro-paraffins are comparatively From " The NitroparafEns " Commercial Solvents Corporation New York. LEVY AND ROSE THE ALWHATIU NITBO-OOMPOONDS 369 innocuous ; they are said to be of the same order of toxicity aa the normal para& of the same boiling point.Structure of the Nitro-parafins and their Salti?.-Nitro-parafih are clasei- fied as primary secondary or tertiary accordmg fo the number of hydrogen atoms carried by the carbon atom adjacent to the nitro-group. Thus nitromethane with three and nitroethane and l-nitropropane with two a-hydrogen atoms are primary nitro-compounds 2-nitropropane with one a-hydrogen atom is secondary and tertiary nitrobutane haa no a-hydrogen atoms. Primary and secondary nitro-paraffins dissolve in aqueous or aqueous alcoholic solutions of strong alkalis but tertiary nitro-compounds do not. This difference in behaviour is due to the fact that primary and secondary nitro-paraffins can exist in an OH aci- or iso-form (known as nitronic acids) and many \ / of the reactions of the aliphatic nitro-compounds are C=N due to the reactivity of this aci-form.The generally / 4 (n.1 accepted structure of the aci-nitro-paraffi is (11) and the arguments in favour of this structure are excellently reviewed by Taylor and Baker. The structure of the salts has until recently been a matter of consider- able controversy. In 1927 it was reported by R. Kuhn and H. Albrecht 79 that optically active 2-nitrobutane gives an active sodium salt when heabd with sodium methoxide ; this sodium salt with hromine gave an optically active 2-bromo-2-nitrobutane. A similar study on active 2-nitro-octane w&s carried out by R. L. Shriner and J. H. Young 80 in 1930 ; they demon- strated that an optically active sodium salt was formed which gave an active 2-bromo-2-nitro-octane and in which the acti,vity was retained to the extent of 24% and 71% when the parent nitro-paraffin was regenerated at - 10" and - 70" respectively.The acceptance of these results rendered untenable the accepted structure (111) of the nitro-paraffin salts and many theories were advanced,l 789 8ot 819 82 none of which was completely satis- factory to account for these phenomena. As far as Shriner and Young's results are concerned the problem has been solved by a recent re-investigation of their It has been shown that their active 2-nitro-octane prepared from active 2-bromo-octane and silver nitrite contains some active 2-nitro- octyl nitrate which cannot be separated by +stillation. Naturally the activity due to this ester is retained throughout the cycle of sodium salt formation and regeneration and accounts for the 24% residual activity in the experiments at - 10".The 71% activity in the lower temperature work is demonstrated as due to (a) the octyl nitrate and ( b ) incomplete salt formation at the lower temperature. When pure optically active 2-nitro- 0 R' ' 9 Ber. 1927 60 1297. B1 " Organic Chemistry " ed. H. Gilman John Wiley & Sons New York 1943 ed. 8' H. B. Haw and E. Riley Chem. Revkw8 1043,88,381. Ea N. Kornblum N. N. Lichtin J. T. Patton and D. C. IfFland J . Amer. Ohm. Eo J . A w . Chem. Soc. 1930,62 3332. Vol. I p. 391. Soc. 1047 69 307. 3 70 QUARTERLY REVIEWS octane is converted completely into its salt and regenerated the resulting nitro-paraffin is inactive. The structure (111) can therefore be accepted with confidence and it is postulated by L.P. Hammet,*' that the contribu- tion of form (IIIB) is more important than that of (IIIa). a b (111.) Primary secondary and tertiary aliphatic nitro-compounds can be differentiated one from another by reaction with nitrous acid in alkaline solution the so-called " red white and blue " reaction. Primary compounds give a red colour due to the formation of a nitrolic acid salt (IV) secondary give a blue colour due to a peudonitrole (V) whilst tertiary do not react. NO2 / Action of Acids and Bases.-Cautious acidification of the aqueous solution of the sodium salt of a nitro-paraffin liberates the free uci-form which can often be isolated in state of purity. Thus phenylnitromethane an oil after dissolution in sodium hydroxide and acidification a t O" gives aci-phenyl- nitromethane a white crystalline solid (m.p.84") which can be recrystallised from light petroleum ; this reverts slowly to the normal liquid nitro-paraffin. Regeneration of nitro-compounds from alkaline solutions is best carried out either a t low temperature or with a weak acid (carbon dioxide or acetic acid). At higher temperatures alkaline solutions of primary nitro-paraffins may rearrange to hydroxamic acids s5 which by hydrolysis give carboxylic acids and hydroxylamine R-CH:NO*OH -+ R*C(OH):N*OH Excellent yields of hydroxylamine are obtained by heating primary nitro- paraffins with 85 yo sulphuric acid.86 Under these conditions secondary nitro-paraffins give only tars. An alternative mode of decomposition of the alkali aci-nitro-paraffins is to an aldehyde (from primary) or ketone (from secondary nitro-compounds) and nitrous oxide.87 This reaction predominates when the aqueous sodium salt is added to excess of dilute sulphuric acid and aldehydes and ketones ~34 " Physical Organic Chemistry " McGraw Hill Book Co.New York 1940 p. 67. s6 H. L. Yale Chem. Reviews 1943 33 226. 88 S . B. Lippincott and H. B. Ham I n d . Eng. Chem. 1939 31 118. J. U. Nef Annalen 1894 280 263. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 37 1 have been obtained by K. Johnson and E. F. Degering ** in 85% yield by this method ZCH,*CH.NO,Na + 2H,SO,-+ 2CH,.CHO i- N20 + NaHSO + H,O 2(CH,),C:N02Na + 2H,SO,+ B(CH,),CO + N20 + XaHSO + H20 The alkali-metal salts of the nitro-paraffins are unstable to heat or shock the potassium salts being more sensitive than the sodium salts.Heavy metals form insoluble precipitates with aqueous solutions of the alkali salts but nitromethane is exceptional in giving mercury fulminate by spontaneous dehydration of the intermediate mercury salt r With strong alkalis nitromethane gives salts of methazonic acid which by further hydrolysis gives salts of nitroacetic acid thus CH,:NO*ONa + CH,*NO -3 [NO,*CH,-CII,*N( OH)*ONa] _L - HZO N(OH):CH*CH:NO,Na -+ [CN*CH:SO,Ka] -+ CO,Na*CH:NO,Na Dehydration of the intermediate methazonic acid with thionyl chloride affords nitromethyl ~yanide.8~ The action of basic substances on primary nitro-paraffins higher than nitromethane was investigated by W. R. Dunstan and his co-workers who demonstrated the formation of trialkylisooxazoles. More recently S.B. Lippincott 91 has carried out this reaction under milder conditions and shown that ,&dioximes are the initial products formed by the action of organic bases and that these on hydrolysis with dilute acids give the trialkylisooxazoles. sodium methazonate The following reaction mechanism is postulated X-OH X*OH I I! - HKO 2R*CH:NO*OH + R*CH,*NO + CR-CR-CR - +H,O I so 0 / \ N-OH N.OH K*OH N CR I1 II II II II CR*CHR*CR -+ CR*CIIR.COR + CR-CR Reduction.-The reduction products of nitro-paraffins and their deriva- tives vary according to the conditions and reducing agent employed. A large number of methods of reduction have been described and an extensive review of these methods is given by H. B. Hass and E. Riley.2 K. Johnson J . Org. Chcm. 1943 8 10. W. Steinkopf and L. Bohrmann Ber.1908 41 1048. \Ir. R. Dunstan and T. S. Dymond. J. 1891 59 410; W. R. Dunstan and E. Goulding J. 1900 77 1262. Dl J . A m r . Chem. SOC. 1940 62 2604. 372 QUARTERLY REVIEWS and E. F. Degering 92 conclude that the beat yields of rtmines from nitro- parafis are obtained with iron and hydrochloric acid or by ca,ta,lytio reduction over Raney nickel under pressure. For the reduction of %nitro- alcohols Raney nickel and hydrogen ia the preferred method since the alcohola are sensitive to acids and alkalis ; a diluent is necessary to prevent the reduction from getting out of hand and it is usually preferable to work a t superatmospheric pressure and room temperature. Secondary aminea have been prepared by reduction of a nitro-compound with zinc dust and dilute acetic acid in presence of an aldehyde.Thus reduction of nitro- methane with beiizaldehyde gives ben~ylmethylamine.~3 Hydroxylamines are isolated as major products when nitro-paraffins are reduced with zinc dust and water zinc dust and ammonium chloride,9* zinc and dilute acetic acid,95 or sodium amalgam in neutral s o l u t j ~ n . ~ ~ Carbonyl compounds have been obtained by reduction of primary or secondary nitro- paraffins with acid reducing agents such as zinc dust and acetic acid ; 97 these are probably formed by hydrolysis of an intermediate oxime since K. Johnson 92 has shown that zinc dust and glacial acetic acid is a general method for the reduction of nitro-paraffins to oximes. Ketones are ah0 formed by catalytic reduction of a-chloronitro-paraffins with palladium on barium ~ulphate,~* and oximes are obtained in excellent yield by reducing the sodium salt of the nitro-paraffin with cold stannous chloride and hydro- chloric acid.99 Action of Diwo -cmpounds .-The condensation of aryldiazonium salts with nitro-paraffins was discovered by Victor Meyer,lW who obtained an orange solid from benzenediazonium sulphate and sodium nitroethane ; similar substances were also prepared from 2-nitropropane.The compounds derived from secondary nitro-compounds are regarded as true azo-compounds (VI) but those from primary nitro-compounds with one molecular proportion of diazo-salt correspond better with the hydrazone structure (VII) for the free nitro-compounds and the azo-structure (VIII) for the salts.lO1 102 103 RR'C*N:NPh R*C:N*NHPh R* CON NPh Me*C(N,Ph) I1 I NO*ONa NO (VIII.) (IX. 1 The salts (VIII) react with a further molecule of aryldiazonium chloride and the h a 1 product of the reaction between e.g. nitromethane and benzene- oaJ. Amer. Chem. SOC. 1939 61 3194. Oa S. Kanao J. Phurm. SOC. Japan 1929 49 42 ; Chem. Zontr. 1929 I 2974. O4 E. Beckmann Annalen 1909 365 206. W. R. Dunstan and E. Goulding J. 1900 77 1262. 96 W. Charlton and J. Kenner J. 1932 760. O7 T. Urbanski and M. Slon Compt. rend. 1937 204 870. E. Schmidt and A. Ascherl Ber. 1925 58 366. J. v. Braun and E. Danziger ibid. 1913 46 103. loo V. Meyer and G. Arnbuhl ibid. 1876 8 761 1073. 101 V. Meyer ibid. 1888 21 11. 108 E. Bamberger Ber. 1894 27 167. loa J. 1930 919. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 373 diazonium chloride is “ nitroformazyl,” NHPh*N:C(NO,) *N,Ph lo8 and of nitroethane the bisazo-compound (IX).E. C. S. Jones and J. Kenner lo2 have shown that interaction of %nitro- propyl alcohol and benzenediazonium chloride gives the phenylhydrazone (VII ; R = Me) formaldehyde being eliminated. Similarly the alcohol and glycol derived from nitromethane and formaldehyde give “ nitro- formazyl ” on treatment with benzenediazonium chloride. C. F. Feasley and E. F. Degering lo4 have recently described the preparation of a series of azo-compounds of type (VI) from 2-nitropropane and 2-nitrobutane with various aromatic amines. If the original amine contains an acidic auxo- chromic group the product will dye wool and silk directly and dyeings are also achieved by coupling of the components on the fibre. Such dyes however have no commercial v a l ~ e .~ ~ s Reactions with Aldehydes and Ketones.-Primary and secondary aliphatic nitro-compounds condense readily with aldehydes and ketones. The product from a primary nitro-compound is usually the corresponding alcohol (X) Bometimes particularly in condensations with aromatic aldehydes the nitro-olefin (XI) and occasionally the dinitro-compound (XII) formed by addition of a further mole of nitro-paraffin to the nitro-olefin (see later) R*CHO + CH,R’*NO + R*CH(OH)*CHR’-NO + (X4 R*CH:CR’*NO + R’*CH*NO1 I R=CH-CHR’*NO I (XI. 1 (XII.) The condensing agents which have been used to effect these reactions are normally basic ones such as alkali carbonates hydrogen carbonates hydroxides or alkoxides primary aliphatic amines and calcium hydroxide but occasionally with aromatic aldehydes acidic agents such as zinc chloride have been employed.In general formaldehyde in excess under mildly basic conditions can replace all the a-hydrogen atoms of a nitro-paraffin by hydroxymethgl groups. Thus nitromethane gives tris(hydroxymethyl)nitromethane nitro- ethane gives 2-nitro-2-methylpropane-1 3-dio1 and 2-nitropropane gives 2-nitro-2-methylpropanol. The nitromethane reaction cannot be restricted to the formation of the diol (2-nitropropane-1 3-diol) ; this is prepared by removal of a molecule of formaldehyde from the trihydroxy-compound by treatment with sodium alkoxide.1°6 With the higher primary nitro-paraffins alkaline condensation with formaldehyde gives a mixture of the alcohol and the glycol and yielde of either can be controlled within certain limits by correct choice of proportion of reactants and experimental conditions.It has recently been found in the laboratories of I.C.I. Ltd. Dyestuffs Division lo’ that the yield of alcohol lo4 J . Org. Chem. 1943 8 12. Io8 E. Schmidt and R. Wilkendorf Bsr. 1919 52 389. lo’ A. Lowe private communication. lo5 I.C.I. Ltd. private communication. 374 QUARTERLY REVIEWS can be considerably increased at the expense of the diol by carrying out the reaction in the presence of a full molecular proportion of caustic alkali. The condensation of aliphatic aldehydes higher than formaldehyde with nitro-paraffins becomes progressively more difficult with increasing mole- cular weight of the aldehyde. Acetaldehyde and nitromethane give 2-nitroisopropyl alcohol (XIII) in good yield the diol [3-nitropentane- 2 4-diol (XIV)] being only a minor by-product in the normal condensation of equimoles of the reactants.lO* Three molecules of acetaldehyde cannot be condensed with nitromethane but the remaining a-hydrogen atom in the diol (XIV) can be replaced by a hydroxymethyl group thus CH,*NO + CH,*CHO + CH,-CH(OH)*CH,*NO -+ (XIII.) CH,-OH H-CHO / [CH,-CH(OH)],CH*NO - [CH,*CH(OH)],C.(XIV.) Aromatic aldehydes and nitro-paraffins condensed together under alkaline conditions yield mcnohydric alcohols which often dehydrate spontaneouely to the nitro-olefin unless special precautions are taken. Thus benzalde- hyde and nitromethane yield o-nitrostyrene directly when the alkaline reaction product is acidified with a mineral acid ; if acetic acid is used the nitro-alcohol is precipitated log and can be isolated.Ph*CHO + CH,*NO + Ph*CH(OH)*CH,-NO + Ph.CH:CH*NO Nitromethane reacts with ketones such as acetone or methyl ethyl ketone under the influence of basic catalysts to give 1 3-dinitro-paraffins.ll0 H. B. Hass ll1 has shown that in the reaction product of acetone and nitro- methane l-nitro-2-methylprop-l-ene (XVI) is present and that this can react with nitromethane giving 1 3-dinitro-2 2-dimethylpropane (XVII). An important reaction product isolated by Hass and not reported by Fraser and Kon,llo is 5-nitro-4 4-dimethylpentan-2-one (XVIII) formed by addition of excess of acetone to the nitro-olefh. By the use of a large excess of acetone this can be made the main product. Hass poetulatea these reactions as following the course below but describes nitro-tert. -butanol (XV) as unknown.It has recently been shown '2 that when sodium methoxide or hydroxide a quaternary ammonium hydroxide or trimethylamine is used as catalyst nitro-tert.-butanol is formed in high yield. This nitro-alcohol kept at room temperature with secondary amines gives a mixture of the unchanged alcohol the nitro-olefin (XVI) and the dinitro-paraffin. This indicates that the reaction mechanism proposed by Hass (see above) is essentially correct but the formation of the nitro-alcohol from acetone and nitromethane lo* G. D. Buckley and (Mrs.) J. L. Charlish J. 1947 1472. loo K. W. Rosenmund Ber. 1913 46 1037. ll0 H. B. Fraser and G. A. R. Kon J . 1934 604. 1 n Ind. Eng. Chsrn. 1943,35,1160. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 376 is reversible.ll2 The rectctions of nitromethane with cyclohexttnone and methyl ethyl ketone follow a similar course.(CH,)&O + CH,*N08 + (CH,),C(OH)*CH,*NO -+ (XV.) acetone (CH,),C:CH*NO + CH,*CO*CH,*C( CH,),*CH,.NO (XVI.) (XVIII. ) CH,*NO I (CH3)2C(CH2*N02)2 (XVII.) Nitro-alcohols and glycols derived from nitro-paraffins and aldehydes or ketones are readily reduced to the corresponding amino-alcohols which have been extensively investigated as intermediates in the preparation of emulsify- ing agents detergents and dispersing agents fields in which the ethanol- amines (prepared from ethylene oxide and ammonia) have already found considerable application. The Mannich Reaction.-The Mannich reaction on nitro-paraffins was first investigated by L. Henry,l13 who showed that N-hydroxymethyl- piperidine condensed with nitromethane and nitroethane yielding respectively 2-nitro- 1 3-di-N-piperidinopropane (XIX) and 2-nitro- 1 3-di-N-piperidino- 2-methylpropane (XX).ZC,H,,N*CH2*OH + CH,*NO -+ N0,*CH(CH2*NC5H,,) (XIX.) SC,H,,N*CH,*OH + CH,*CH,*NO + N0,*C(CH,)(CH,~~C5Hl,) (XX. ) As a result of later work 1 1 4 9 115 it has been established that although under normal conditions nitromethane and nitroethane condense with two molecules of a secondary hydroxymethylamine higher nitro-paraffins (nitropropane, l-nitrobutane) condense with only one. The formation of the nitro- monoamines (XXI) from e.g. nitroethane and piperidine requires long reaction periods at low temperatures C,H,,N*CH,*OH + CH,-CH,*NO --+ C5H,,N-CH,*CH(CH,)*N02 (XXI.) and the products are unstable ; (XXI) for example on being kept precipi- tates the nitro-diamine (XX).llS The nitro-diamine (XX) can also be prepared from the nitro-glycol (XXII) by reaction with piperidine and NO,*C(CH,)(CH,*OH) + C5H,,N -+ NO,*C(CH,)(CH,*NC5Hl0) CH,*CH( OH)*CH,*NO + 2C,H,,N*CH,~OH+NO,*CH( CH,*NC,H,,) +CH,*CHO (XXII.) lla A.Lambert and A. Lowe J. 1947 1617. 113 Bull. Acad. roy. Belg. 1896 32 33 ; Ber. 1906 38 2027. 114 Cerf de Mauney Bull. SOC. chim. 1937 4 1461 1460. 115 M. Zief and J. P. Mason J. Org. Chem. 1943 8 1. llLI A. Lambert and J. D. Rose J . 1947 1511. 376 QWAR!l'ERLY BEVIEWS (XIX) is formed from hydroxymethylpiperidine and 2-nitrokopropyl aloohol with elimination of acetaldehyde.116 Addition to Activated Ethylenic Compnds.-Closely parallel to the reaction of nitro-paraffins with aldehydes is the addition of primary and secondary aliphatic nitro-compounds to substances containing an activated ethylenic linkage.The reaction is similar to the Michael condensation of the sodium derivative of malonic ester with activated unsaturated com- pounds. By this method E. P. Kohler and his co-workers 117 condensed nitromethane with benzylideneacetophenone using an alkaline catalyst and obt,ained 3-nitro-2-phenylbutyrophenone (XXIII) and the bis-adduct (XXrV) CH,*NO + Ph*CH:CH*COPh -+ Ph*CH*CH,*COPh -P Ph*CH*CH,*COPh I CH*NO I 1 CHI*NO Ph*CH*CH,*COPh (XXIII.) (=. 1 A. Sonn 11* obtained 3-nitrobutyrophenone (XXV) from phenyl vinyl ketone and nitromethane whilst C. F. Allen and A. B. Bell 11* obtained tria-(2~ benzoylethy1)nitromethane (XXVI) fiom the same reactants CH,*NO + CH,:CH*COP -+ NO,*C(CH,*CH,*COPh) (XXVI.) \ N0,*CH,*CHa*CK,6COPh (XXV.) Mannich bases derived fiom methyl ketona react with nitromethane under alkaline conditions and form similar products ; t h m 2-dimethylamino- propiophenone affords the nitro-ketone (XXV).140 H.Bruson has recently demonstrated the addition of nitro-para5~ to vinyl cyanide. Nitromethane in presence of a quaternary ammonium hydroxide catalyst gives tris-(2-~yanoethyl)nitrometha4e (XXVII) nitro- ethane gives 3-nitro-1 5-dicyano-3-methylpentax~e (XXVHI) and 2-nitm- propane adds only one molecule of cyanide giving 3-nitro-l-cyano-3- methylbutane (XXIX) CH,*NO + SCH,:CH.CN -+ NO,*C(CH,*CH,*CN) (XXVII.) CH,*CH,*NO + ZCH,:CH.CN -+ NO,*C(CH,)(CH,*CH,CN) It ie difEcult to restrict this reaction to the condeneation of number of vinyl cyamide molecules less than the number of active hydmgen atome in the nitro-paraffin but recently the bis-adduct of vinyl cyanide and nitro- methane has been prepared by carrying out the condensation in the presence of a full molecular proportion of crtustic dkdi.lrn (XXVIII.) (CH,),CH*NOa + CH,:CH.CN -+ NO,*C(CH,),*CHS*CH,*CN (XXIX.) 117 J .Amer. Ohm. Soc. 1916 88 889; 1919 U 1044; 1922 024 2144; 114 Ber. 1936 88 148. 11* Cunudiun J . Bss. 1934,11 -6. l * O B. Reichert and H. Posemann A d . P-. t957,816,67. lal J . A m . Ohen. Soe. 1943 06 23. 1'1 G. D. Buckley T. 3. Elliott F 42. Hunt and A. Lowe J. lM7,lM)Ei. 1924,16 609 1274 ; 1926 48 2426. An interesting phenomenon haa been observed in attempts to reduce the nitro-cyanide (XXIX) to the corresponding diamine. Reduction with iron and hydrochloric acid gave 5-jmino-2 2-dimethylpyrrolidine (XXXIII) and a white crystalline substance C,H,,ON, formulated &B 5-amino-2 2- dimethylpyrroline N-oxide (XXXII) or its tautomeride 5-imino- 1 -hydroxy- 2 2-dimethylpyrrolidine (XXXI).128 This substance is also obtained in good yield by reduction of the nitro-cyanide with hydrogen over Raneynickel or in excellent yield by reduction with zinc dust and ammonium chloride.CHa-CH I I \ / (CH,)gC*CH,*CH,*CN + (CH,)gC*CH,*CH,.CN -+ (CH3)aC N dH NO9 I II LOH 1 (=.) (x=4 (XXXI.) or i CH,*CH(NOa) *CHaCH,*CN 1 (xxxrv-) CHa-CH I I \ / J. CH,*CH C-NH N 0 CHa-CH CH,*CHs.CN \ / C / CH 0 (XXXVII.) This last method of reduction gives a clue to the mechanism of form ti n of the N-oxide ; zinc and ammo&um chloride are specific reagents in the nitro-para5 series for reduction of the nitro- to the hydroxyla;mino-group and it is probable that in this case the intermediate product is 3-hydroxyl- amino-3-methylbutyl cyanide (XXX).A similar case has been reported l** G. D. Buckley and T. J. Elliott J. 1947 1508. 378 QUABTERLY REVIEWS by K. H. Bcluer 12* who hydrogenated o-nitroatyryl cyanide over a palladium catalyst and obtained 2-aminoquinoline N-oxide in 60% yield. Similar behaviour on reduction was observed with analogues of (XXIX) ; 83 thus 3-nitrobutyl cyanide (XXXIV) gave 5-amino-2-methyl- pyrroline N-oxide (XXXV) and 5-amino-2 2-pentamethylenepyrrole N - oxide (XXXVII) resulted from the reduction of 1 -nitro-l-( 2-cyanoethyl) cycbhexane (XXXVI) . The behaviour of nitro-parafks with vinyl cyanide is closely paralleled by that with vinyl ~ulphones.l2~ In the presence of alkaline catalysts all the a-hydrogen atoms of the nitro-paraffin are replaced giving nitro- monosulphones from secondary and nitro-disulphones from primary nitro- paraffins.Nitromethane and methyl vinyl sulphone give tris-(e-rnethyl- sulphony1ethyl)nitromethane (XXXVIII) and nitroethane with butyl vinyl sulphone in the presence of a small amount of potassium hydroxide gives %nitro- 1 5-di-(n-butylsulphonyl)-3-methy.lpentane (XXXIX) ; if the quan- tity of potassium hydroxide is increased to a full molecular proportion the main product of this reaction is 3-nitrodibutyl sulphone (XL). 2-Nitro- propane and divinyl sulphone afford 3 3’-dinitro-3 3'-dimeth yldibutyl sulphone (XLI). 3CH3*S0,*CH:CH2 + CH,*NO -+ NO,*C( CH,*CH,*SO,*CH,) ZC,H,*SO,*CH:CH ’+ CH3CH2*N0 -+ NO,*C( CH,*CH2*S0,-C4K,) (XXXVIII .) I CH (XXXIX.) C,H,*SO,*CH:CH + CH,.CH,*NO + NO,*CH*CH,*CH,*SO,*C,H I CH (XL.) CH,:CH*SO,*CH:CH + 2(CH,),CH*N02 -+ N0,*C*CH,*OH,*S0,*CH,-CH2*C( CH,) Reduction of the nitro-sulphones to amino-sulphones proceeds readily and smoothly with Raney nickel and hydrogen at ordinary temperature and pressure. (see p. 370) of nitro-paraffins is unusually facile in the nitrosulphone series and the free nitro-compounds must be regenerated from aqueous solutions of their alkali metal salts with acetic or carbonic acid. Acidification of the potassium salt of (XL) with mineral acid for example yields 3- ketodibut yl sulphone .I25 Reactions with Organometallic Reagents.-The action of organometallic compounds on nitro-paraffins was f i s t examined by C.Moureu126 who observed that nitro-ethane and two molecules of ethylmagnesium iodide gave NN-diethylhydroxylamine. J. Bevad 12’ studied the reaction of primary and secondary nitro-paraffins with several alkylzinc and alkyl- lZ4 Ber. 1938 71 2226. 186 G. D. Buckley (Mrs.) J. L. Charlish and J. D. Rose J. 1947 1514. lZ6 Compt. rend. 1901 132 837. The Nef hydrolysis 12’ Ber. 1907 40 3066. LEVY AND ROSE THE ALIPHATIC NITRO-UOMPOUNDS 379 magnesium iodides and found that nitroethane reacted with ethylmagnesium iodide with evolution of ethane and a little ethylene to give a complex which on hydrolysis gave NN-diethylhydroxylamine together with minor quan- tities of diethylamine and ethyl-sec.-butylhydroxylamine. Ethylzinc iodide reacted similarly but gave ethyl-sec.-butylhydroxylamine as the main product.Similar results were obtained from 1 - and 2-nitropropane. Bevad postulated reaction of the Grignard reagent with the mi-nitro- paraffin to give the complex (XLII) followed by addition of one or two molecules of Grignard reagent to the ethylenic linkages OH OMgI (XLII.) / / L L + CH,*CH:N 0 0 T. Zerewitinoff l2* showed that primary and secondary nitro-parafks reacted with methylmagnesium iodide to give slightly less than one equiva- lent of methane and considered that this arose from reaction of the Grignard reagent with the mi-nitro-paraffin. A. B. Wang,129 from a detailed study of the reaction products of phenylmagnesiurn bromide and nitromethane put forward a reaction mechanism differing only in detail from that of Bevad and also involving a complex of type (XLII).These mechanisms are unlikely since they postulate the preliminary tautomerism of the nitro-paraffin to its mi-form a change normally requiring strongly ionising conditions. The reaction has been recently re-examined by G. D. Buckley 130 who considers that the first stage is addition of the Grignard reagent to the N=O bond of the nitro-group giving a complex (XLIII). Complexes of this type CH,*CH:N OMgBr / CH,*CH,*N=O + CH,*CH,*N-Et (XLIII.) J. 0 J. 0 have been isolated by interaction of equimoles of the reactants at low temperature. They are white hygroscopic powders which on reduction with zinc and hydrochloric acid give a secondary amine e . g . diethylamine from (XLIII). The complex may then react with a second molecule of ethyl- magnesium bromide with evolution of a molecule of ethane and formation of a second complex which yields NN-diethylhydroxylamine on hydrolysis.The mechanism of the second stage of this reaction is still obscure. ChZoro-nitro-~aru~ns.-The most important of the chlorinated nitro- paraffins is chloropicrin (trichloronitrornethane CC1,*N02). It is formed when almost any organic compound is treated destructively with " aqua regia " and was first prepared by R. Preibisch 131 by the action of bleaching powder on nitromethane. Preibisch thought his product was monochloro- nitromethane but it has since been shown that he obtained the azeotrope 128 Ibid. 1910 43 3593. lZ9 Trans. Sci. SOC. China 1932 7 253. J . pr. Chern. 1873 8 309-327 ; J . 1874 27 462. 130 J . 1047 1492. B B 380 QUARTERLY REVIEWS of nitromethane and chloropicrin.The literature on chloropicrin up to 1933 has been comprehensively reviewed by K. E. Jackson 132 and little needs to be added. A process for the quantitative conversion of nitro- methane into chloropicrin by chlorination in an aqueous suspension of calcium carbonate has been worked out by W. D. Ramage,133 and similar results are claimed using an alkali or alkaline-earth hypochlorite.134 Chloropicrin is an extremely active lung irritant fatal to man a t con- centrations of 1 part in 20,000 of air. Although used in 1916 in gas shell it has been rendered obsolete for military purposes by the efficiency of modern respirators ; it is however finding important applications as a soil-sterilising agent. Other halogenated nitro-paraffins are made very easily by addition of halogen to the aci-form or sodium salt of the nitro-compound and all the a-hydrogen atoms can be substituted in this way.The a-chloro- and a-bromo-nitro-paraffins are more stable than the a-iodo-analogues some of which decompose on attempted distillation. 135 Little is known of the chemical reactivity of the a-halogeno-nitro- compounds a.lthough the formation of 1 2-dinitro-paraffins by interaction of 2-chloro-2-nitropropane with the sodium salts of other nitro-paraffins has been reported. Chlorination of t,he nitro-paraffins under the influence of phosphorus pentoxide and intense illumination causes halogenation on carbon atoms other than the a- ; nitro-alcohols can be converted into chloronitro- cornpounds by treatment with phosphorus pentachloride.Polyn itro-parafins .-Mi phat ic dinitro-compounds in which the two nitro-groups are attached to the same carbon atom can be prepared by chromic acid oxidation of the psewdonitroles which in turn are formed by the action of nitrous acid on secondary nitro-paraffins 136 or by treatment of oxiniea with dinitrogea tetroside 137 HNOo NO2 /No H,CrO / R,CH*NO R2C __+ R2C NO2 \ Certain ketones notably ethyl ketones give dinitro-compounds on treatment with concentrated nitric acid. Thus diethyl ketone methyl ethyl ketone and ethyl propyl ketone all yield 1 l-dinitroethane.138 139 The V. Meyer method can also be applied to the synthesis of gem-dinitro-compounds and 1 l-dinitropropane has been produced by the interaction of l-bromo-l- nitropropane and sodium nitrite. l40 Nitroform (trinitromethane) is formed in small yield by passing acetylene 132 Chem.Rsviews. 1934 14 251. 134 B. &I. Vanderhilt U.S.P. 2,181,411. 1 3 j L. W. Seigle and H. B. Hass J. Org. Chena. 1940 5 100. V. MIcyer RmaaZm. 1875 175 120. 138 31. G. Chancel. Bull. SOC. chim. 1879 31 504. 31. Fileti and G. Ponzio J . pr. Chein. 1897 55 195. I * " E. ter >Jeer AnnuZen 1570 181 1. 133 U.S.P. 1,996,388. 137 G. Born. Ber.. 1896 29 90. LEVY AND ROSE lpHID -TI0 WfiBO-OOMPOENDS ss1 into a mixture of nitrio aoid and edphudc wid 141 or from e t k y h and fuming nitric scid.142 Mercuric nitrate is d 1u1 a cafalyst aud fnom ethylene mme dinitroethy1 alcohol is formed. Nitroform on being h d with fuming sulphuric acid givea tetranifrOrnetht1-ne,~4~ from which nitro- form am be regenerated by treatment with potassium hydroxide or ethoxide reaction which appears from the dewription of an explosion given by A.K. Macbeth to be extremely dangerous. Safer produma however have been worked out by F. D. Chattaway l44 and by A. K. Macbeth.1'6 Tetranitromethane can be prepared in 80% yield by interaction of nitrio acid and acetic anhydride for several days a t room temperat,~re.~4~ The dinitro-compounds with the exception of dinihomethane are relatively stable substances which can be distilled without decomposition. The primary ones (i.e. those containing an a-hydrogen atom) are strong acids which almost certainly exist in solutions in the aci-form. On reduc- tion one nitro-group is readily lost giving an oxime which on further reduc- tion-yields an amine l4' The products of heating dinitromethane with potassium hydroxide are potassium nitrate ammonia and a fatty acid.This easy loss of a nitrogen atom has been adduced by G. Ponzio 148 as evjdence that dinitromethane is in fact a nitro-nitrite ; this is unconvincing since a similar loss of a nitro- group occurs even more readily with tetranitromethane where the evidence for four nitro-groups is incontrovertible. Its aqueous solutions and salts are yellow suggesting that in ionising solvents it exists in the mi-form. A solid form m.p. 50" can be obtained by cautious acidification of the potassium salt with ice-cold concentrated sulphuric acid ; this is probably the solid mi-form (NO,),aC NO*OH the form m.p. 23" being the true trinitro-compound CH(N02),. The mercury and silver salts are readily soluble in organic solvents and it is probable that the silver salt exists in the form of the chelate ring compound (XLIV) R*CH(NOa) -.+ R*CH:N*OH + R*CH,*NH Nitroform is a colourless liquid which solidifies at 23".0 N-0 NO + / NO2.C \ I NO,*C*O*NO I \ I N=O NO J. 0 (XLN.) (XLV.) Tetranitromethane is a liquid boiling unchanged at 126" and melting at 13". It forms violently explosive mixtures with aromatic hydrocarbons and 141 K. J. P. Orton and P. V. McKie J. 1920,117 283. 14' P. V. McKie J. 1927 962. 144 F. D. Chattaway and J. M. Harrison J. 1916 109 171. 146 A. K. Macbeth and W. B. Om J. 1932 638. u6 F. D. Chattaway J. 1910 97 2099. l'' a. Ponzio J. p. Chem. 1902 G 197. 149 Ber. 1913 40 2537. Ibi&. 1903 07 137. B B* 382 QUARTERLY REVIEWS fatalities fiom euch mixturea h v e been reported.149 A marked yellow colour the origin of which is not known appears when tetranitromethane is mixed with an unsaturated or aromatic compound and this colour reaction is often used as a test for unsaturation.The dipole moment of tetra- nitromethane in carbon tetrachloride is zero ; 150 the calculated dipole of a molecule with the structure (XLV) deduced from the individual dipoles of nitromethane and amyl nitrite would be 0.8 D a value which is far beyond the limits of error of measurement. The symmetrical tetranitro-structure can therefore be accepted with confidence. The ease with which tetranitromethane loses a nitro-group makes possible its use as a nitrating agent. Dimethyl-p-toluidine and tetranitromethane in alcohol give m-nitrodimethyl-p-toluidine but non-basic substances e.g.phenol are nitrated only in the presence of a base such as pyridine. The formation of the 1 2-dinitro-paraffins has been referred to earlier in this review (p. 363). These substances show properties which are very different from those of the 1 3- and 1 4-analogues ; in fact the latter behave so " normally " that it is not proposed to consider them at this point but to deal with them under the reactions of nitro-olefins and nitroparaffins by which they are formed. The abnormal properties of the 1 2-dinitro-paraffins are largely due to the ease with which one nitro-group is lost. In unsymmetrical substances such as 1 2-dinitroisobutane the NO group attached to the carbon poorer in hydrogen is readily removed. Thus by refluxing 1 2-dinitroisobutane with alcohol 1 -nitro-2-methylprop- 1 -ene is formed ethyl nitrite being evolved 50 151 (CH3),C(N02)*CH2*N02 + C2H,*OH + (CH,),C:CH*NO + C2H,*O*N0 + H20 Alternatively the nitro-olefin may be formed by stirring the dinitroiso- butane with a weak alkali such as calcium hydroxide zinc oxide or sodium carbonate.152 This process is however limited to the formation of non- polymerisable olefins and treatment of e.g.1 2-dinitroethane with weak alkalis gives only a polymer of nitr0ethylene.4~9 153 1 2-Dinitroethane is a white crystalline solid m.p. 39-40' b.p. 8S0/ 1 mm. ; it slowly decomposes on storage giving gome nitroethylene and nitrous fumes. This decomposition is probably alkali-catalysed and can be completely inhibited by incorporation of 0.5% of an aromatic sulphonic acid such as naphthalene-1 5-disulphonic acid.48 154 The structure of 1 2-dinitroethane as a true dinitro-compound is rigidly established by reduction t o ethylenediamine and by mineral acid hydrolysis to oxalic acid and hydroxylamine 4 8 9 155 10H NH2*CH2*CH2*NH2 t NO2CH2*CH,*NO2 --+ (C02H) + 2NH2*OH A.Stettbacher J. Ind. Hyg. (Abstr.) 1943 25 49. lS0 A. Weissberger and R. Sangewald Ber. 1932 65 701. 161 A. E. W. Smith R. H. Stanley C. W. Scaife and I.C.I. Ltd. B.P. 583,468. lS2 A. E. W. Smith C. W. Scaife and I.C.I. Ltd. B.P. 580,256. ls3 A. E. W. Smith and I.C.I. Ltd. B.P. 572,891. 164 C. W. Scaife and I.C.I. Ltd. B.P. 678,169. 166 A. E. W. Smith and I.C.I. Ltd. B.P. 573,630. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 383 Other 1 2-dinitro-paraffi are similar to dinitroethane ; 1 2-dinitro- propane dB and 1 2-dinitrobutane 50 are liquids which on treatment with alkali yield respectively poly-l-nitroprop-l-ene and l-nitrobut-l-ene.Reduction gives the diamines and hydrolysis hydroxylamine. Nitro-OleJinS Preparation.-The formation of the non-polymerisable olefins and of polymers of nitroethylene and nitropropene by fission of HNO from the 1 2-dinitro-paraffins has already been mentioned. The 2-nitroalkyl nitrates behave in the same way on treatment with alkali and the nature of the product (monomer or polymer) is determined by the degree of sensitivity of the nitro-olefin to alkali-catalysed polymerisation. Thus 2-nitroethyl nitrate on treatment with weak alkalis gives polynitroethylene 156 and 2-nitroisopropyl nitrate yields poly- l-nitropropene.40 Nitro-&rt.-butyl nitrate however on treatment with sodium hydroxide gives an 83% yield of 1 -nitro-2-rnethylprop- l-ene,60 as this nitro-olefin is relatively insensitive to alkali.(CH,),C( O*NO,)*CH,*NO + (CH,),C:CH*NO The direct formation of nitro-olefins by interaction of an aldehyde or .ketone with a nitro-paraffin has been mentioned on p. 373. This is limited in its applicability and usefulness to nitro-olefins derived from ketones with nitromethane and those from aromatic aldehydes and primary nitro-paraffins. Examples of these two classes are l-nitro-2-methylprop-1-ene (nitroiso- butene) from acetone and nitromethane,112 and o-nitrostyrene (2-phenyl- nitroethylene) from benzaldehyde and nitromethane. The main standard method of preparation of nitro-olefins is by dehydra- tion either directly or indirectly of the 2-nitro-alcohols which are formed in good yield from aldehydes (and some ketones) and nitro-paraffins.The direct dehydration is not generally a good method; powerful dehydrating agents such as phosphoric oxide thionyl chloride potassium hydrogen sulphate or zinc chloride are necessary and often the nitro- olefin does not survive the conditions necessary for its formation. Yielda are poor and polymeric by-products are common. H. Wieland and E. Sakellarios 15' prepared nitroethylene by dehydration of 2-nitroethyl alcohol with potassium hydrogen sulphate but this method is suited only to small-scale preparations and fails in larger experiments.158 A better method of direct dehydration is by heating the nitro-alcohol with phthalican hydride or a substituted phthalic anhydride the nitro-olefin being removed by dis- tillation as it is formed.158 Possibly the best method for all-round applicability and general usefulness is based on that of E.Schmidt and G. Rutz who reported 169 that the acetates lS6 A. E. W. Smith R. H. Stanley C. W. Scaife and I.C.I. Ltd. B.P. 673,786. lS7 Ber. 1919 52 898. 16* GF. D. Buckley and C. W. Scaife J. 1947 1471. Ber. 1928 61 2142. 384 QUARTERLY REVIEWS of 2-nitro-alcohols on refluxing with potassium carbonate or hydrogen carbonate in ether solution lost acetic acid yielding nitro-olefins. Thia was modified by H. Schwarz and G. Nelles,lso who obtained excellent yielda of a-nitro-olefins by interaction of a 2-nitroalkyl acetate with sodium acetate or potassium carbonate and distillation of the product. R-CH(NO,)*CR'R".O*CO.CH + R*C(NO,):CR'R* + CH,*CO,H Nitro-dienes.This last method has recently been applied to the ayn- thesis of conjugated nitro-dienes ; 3-nitro-2 4-diacetoxypentane (XLVI) prepared by acetylation of the glycol (XLVII) from nitromethane and acetaldehyde on heating with sodium acetate gives a 64% yield of $nitro- penta-1 3-diene (XLVIII) CH,*NO + ZCH,*CHO + NO,*CH[CH(OH)*CH,J -+ (XLVII.) CH:CH / \ NO,*CH[CH(OAc)*CH,] -+ N0,mC (XLVI.) CH*CH (XLVIII.) 2-Nitro-3-methylbuta-1 3-diene (LII) was prepared from l-nitro-2-methyl- prop-2-ene (XLIX) by condensation with formaldehyde acetylation of the nitro-alcohol (L) and distillation of the acetate (LI) with sodium acetate NO,*CH,*C(CH,):CH + NO,*CH*C( CH,):CH -+ I (XLIX. ) CH,*OH (L.) NO,*CH*C( CH,):CH --+ NO,*C*C(CH,):CH II CH (LII.) I CH,*OAc (LI.) Starting with 1 -nitromethylcyclohex-l-ene (LIII) a similar series of reactions afforded l-a-nitrovinylcyclohexene (LIV) an unstable oil which on keeping gave a crystalline dimeride of unknown structure.CHa-CH CH,-CH CH \ II / C-C*NO \ / / \ C-CH2*N0 _+ CH / CH2 \ CHZ-CH CHZ-CH (LIII.) (J-JN.) Properties.-The lower or-nitro-olehs are when pure almost colourleae or pale yellow liquids the boiling points of which are of the same order aa those of the saturated compounds. Many of the lower members are strongly lachrymatory and polymerise readily. The polymerisation is catalysed by water and especially by alkalis and in some cases may be almost explosive 160 Assgrs. to General Aniline and Film Corporation U.S.P. 2,257,980. 1 6 1 G. D. Buckley and (Mrs.) J.L. Charlish J. 1947 1472. LEVY AND ROSE THE ALIPHATIC NITRO-COMPOUNDS 385 in its violence. Organic bases catalyse the polymerisation of some sub- stituted nitro-olefins of which nitrostyrene is a good example. Little is known of the structure of the polymers which are usually pale yellow or brown amorphous powders with low solubility in organic solvents. It has recently been reported lea that catalytic hydrogenation of the polymer of 2-nitropropene yields a water-soluble polymer containing primary amino- groups. Rdwtion.-There is no general method for the reduction of the nitro- o l e b to the corresponding nitro-paraffins although C. de Mauney 163 has reported the quantitative reduction of l-nitro-oct-l-ene to l-nitro-octane with hydrogen and platinum. Hydrogenation of nitrostyrene and piperonyl- idenenitromethane with platinum and hydrogen affords the " hydrodimers " 1 4-&nitro-2 3-diarylbutanes 164 (LV).R*CH:CH*NO + R*CH*CH,*NO R*CH*CH,*N02 I (LV4 This was codrmed by E. P. Kohler and N. L. Drake,ls5 who found however that if the reduction with hydrogen and platinum was conducted in the presence of dry hydrogen chloride the yield of the dinitrodiphenyl- butane (LV ; R = Ph) was greatly reduced and the main product was a mixture of the two isomerides of phenylacetaldoxime. Reduction of nitro- styrene with hydrogen and platinum in acetic acid is greatly influenced by the presence of sulphuric acid ; lee 2-phenylethylamine is obtained in 84% yield by using a mixture of glacial acetic and concentrated sulphuric acid but in only very low yield if the sulphuric acid is omitted.This pheno- menon is attributed by Kinder to the formation of a molecular complex pf nitrostyrene and sulphuric acid and the method has been successfully applied to the reduction of piperonylidenenitromethane to homopiperonyl- amine.ls7 . The reduction of a series of substituted nitrostyrenes to the oximes by hydrogenation in pyridine with a palladium or charcoal catalyst has been reported by B. Reichert and W. Koch.ls* The yields of oximes were high and substituted phenylethylamines were obtained from them by further reduction with platinum and hydrogen in alcohol containing oxalic acid. Good yields of substituted phenylethylamines have also been obtained from the corresponding nitrostyrenes by electrolytic reduction. lB9 Catalytic hydrogenation of 2-nitro-l-phenylprop- l-ene (LVI) to " benzedrine " (2-phenylisopropylamine) (LVII) is easier than reduction of 162 A.T. Blomquist W. J. Tapp and J. R. Johnson J . A m r . Chem. Xoc. 1946 67 1519. 163 Bull. SOC. chim. 1940 7 133. 164 A. Sonn and A. Schellenberg Ber. 1917 50 1513. 166 K. Kinder E. Brandt and E. Gehlhtaar Annalen 1934 511,209. 16' 0. Schalee Ber. 1936 08 1579. 168 Arch. Phrm. 1936 2'73 265. lo@ K. H. Sloth and G. Szyazka Ber. 1935 68 184. J . A m r . Chm. SOC. 1923 45 1281. 386 QUARTEBLY BEVIEWS nitrostyrene to phenylethylsmine and has been successfully carried out over Raney ni~ke1.l~~ Ph*CB:C*NO --+ Ph.CHa*CH*NH I I I CHS CH (LVI.) (LVII.) Zinc and acetic acid at 0" are reported 171 to reduce nitro-olefins to the oximes and iron and hydrochloric acid yield a mixture of ketones and ketoximes the latter being favoured if the amount of acid present is strictly limited.172 Since direct reduction of ketoximes or reduction of ketones in presence of ammonia yields the corresponding amines this represents a two-stage conversion of the nitro-olefin into the amine.Nitko-olefins in which the nitro-group is not attached to a doubly-bound carbon atom present no difEculties in reduction either to the saturated nitro-compound or to the saturated a-mine. Nitropentenyl cyanides (LVIII) can be reduced in high yield with hydrogen and palladium on calcium car- bonate to the saturated nitroamyl cyanides (LIX) with iron and acid to the aminopentenyl cyanides (LX) and with nickel and hydrogen to the amino- pent yl cyanides (LXI) . Fe NO2*CR%"*CH2*CH:CH*CHa*CN __+ NH,*CR'R"°CH,*CH:CH*CHg*CN (LVIII.) IPd,& \ (LX.1 NO,*CR%"*[CH,],*CN NH,*CR'R"*[CHa],*CN (LIX.) (LXI.) LIydrcbtion.-a-Nitro-olefins are hydrolysed by water dilute acids or alkalis to the original aldehyde or ketone and nitro-paraffin. L. Haitinger 174 showed that l-nitro-2-methylprop-l-ene (LXII) prepared by nitration of isobutene on hydrolysis with water gave nitromethane and acetone and B. Priebs l75 demonstrated a simjlar hydrolysis of co-nitrostyrene and of 2-nitro- 1 -phenylpropene to benzaldehyde and nitromethane and nitro- ethane respectively. (CH,),C:CH*NO -+ (CH,),CO + CH,*NO (LXII.) H. Wieland and E. Sakellarios 157 established that the first step of such hydrolysis is the addition of water to the ethylenic linkage and by treatment of nitroethylene with cold dilute sulphuric acid they were able to isolate 2-nitroethyl alcohol CH,:CH*NO + CH,(OH)*CH,*NO Addition of Halogens.-The simple a-nitro-olefins add chlorine and bromine readily in the cold forming stable 1 2-dihalogeno-nitro-paraffins.170 Unpublished work by J. B. Tindall quoted in ref. 2 p. 412. 171 L. Bouveault and A. Wahl Compt. rend. 1902,134 1145 ; 135,41. 17s (Mrs.) J. L. Charlish W. H. Dravies and J. D. Rose (in the pram). 17' Anwkn 1878 193 374. A. G. Susie Thesis Purdue University 1939 ; quoted in ref. 2 p. 412. lI6 Ibid. 1884 225 319. LEVY AND ROSE THE ALIPHATIU NITRO-OOMPOUNDB 387 L. Haitinger 17* prepared 1 2-dibromo-l-nitro-2-methylpropane from nitroisobutene and B. Priebs 175 prepared the &bromide and &chloride from nitrostyrene. Nitro-olefins carrying negative substituents react less readily than the simple ones ; thus attempts to add bromine to l-bromo-l- nitro-2-phenylethylene have failed completely.176 If the addition product of bromine and a nitro-oleh contains an a-hydrogen atom hydrogen bromide can be removed by mild treatment with alkalis ; thus 1 2-dibromo-l- nitro-2-phenylethane with sodium acetate affords 1 -bromo-l-nitro:2-phenyl- ethylene 1 7 7 Ph*CH:CH*NO + Ph*CHBr*CHBr-NO + Ph-CH:CBr*NO From this product a further molecule of hydrogen bromide can be abstracted with diethylamine giving the unstable phenylnitroacetylene.176 Ph*CH:CBr*NO -+ Ph*CiC*NO Addition of Hydrogen Chloride.-The early literature on the action of hydrochloric acid on the a-nitro-olefins is not extensive. L. Haitinger 178 reported that l-nitro-2-methylprop-l-ene (nitroisobutene) on treatwnt with hydrogen chloride at 20" or on being boiled with concentrated hydro- chloric acid gave hydroxylamine hydrochloride carbon dioxide ammonia and an unidentified hydroxy-acid m.p.65" which can now be recognised aa impure a-hydroxyisobutyric acid. Similar treatment of nitrostyrene enabled B. Priebs 175 to isolate phenylchloroacetic acid. The subject has recently been reinvestigated by R. L. Hmth and J. D. Rose.179 It is shown that if a nitro-oleh containing an a-hydrogen atom (R,C:CH*NO,) is treated with anhydrous ethereal hydrogen chloride the products are hydroxylamine hydrochloride and an a-chloro- or a-hydroxy- acid R,C(OH)*CO,H or R,CCl-CO,H. If a nitro-olefin with no a-hydrogen atom (R,C:CR*NO,) -is so treated the product is a dichloronitroso- compound RR'CCl-CRCl-NO.The following mechanism is suggested R,C:CH*N // 4 + [ R,C*CH:N ~1 "s k,ri3-NH<:H] 2 0 0 R,C-CH*NO + R,C-C:N*OH --+ R,C*COCl + NHaoOH,HC1 I c1 I I c1 c1 I I (LXIII.) (LXIV.) c1 c1 The initial step is postulated as 1 4-addition ; evidence for this is as follows. If the initial addition were 1 2- the product &om e.g. l-nitropropene lTa J. Loevenich and H. Gerber Ber. 1930 68 1707. 177 J. Thiele and S. Heeckel Annden 1902 326 7. 171 L. Haithger itfonatsh. 1881 2 287; Wkn. A M . Ber. 1878 7'7 420; A, 17* J. 1947 1486. 1879 700. would be 2-chloro-l-nitropropane CHs*CHCl.CH,*NOa. This produot e y n t h ~ d unambiguously from the corresponding nitro-alcohol by tire&- mmt with phosphaus pentachloride ia completely inert to hydrogen chloride under conditiona which convert 1-nitropropene ibelf into a-chlm- propionic acid and hydroxylamine hydrochloride.In this mechanism an important etep i~ the rearrangement of the dichloro-nitroso-compound (LXIII) to the dichloro-oxime (LXIV). If the remtion is carried out on a nitro-olefin R,C:CReNOa containing no a- hydrogen atom the dichloro-nitroso-compound R,CCl*CRCI*NO cannot rearrange and is therefore the end product. Addition of Alcohols and Thiols.-!l?here are many examples of the addition of alcohols to nitro-oleh ; thus w-njtrostyrene,lW w-bromo-o-nitro- styrene,177 1 -bromo- l-nitrobut-l-me and l-bromo- l-nitropent-l-ene 181 all add alcohols in the presence of basic catalysts to give 2-nitroalkyl ethers R,C:CR*NO ___+ R,C(OR’)*CHR*NO The formation of 2-nitroalkyl ethers from l-nitrobut-l-ene and some higher nitro-olefms has also been recently described by C.T. Bahner.182 A. Lam- bert et aZ.lsa have described the formation of 2-nitroalkyl ethers from nitro-olefins and alcohols usually in the presence of basic catalysts. In the case of nitroetbylene a base causes polymerisation but the addition of alcohol proceeds even in the presence of phosphoric acid present m a stabiliser for the nitroethylene. The same products are formed from the 1 2-dinitro-paraffins on heating them with alcohol one nitro-group being eliminated as akyl nitrite and from 2-nitroalkyl nitrates and sodium alkoxides. The nitro-ethers are smoothly reduced by hydrogen over Raney nickel to the corresponding alkyl 2-aminoalkyl ethers. A logical extension of the synthesis of 2-nitroalkyl ethers to the prepara- tion of alkyl 2-nitroalkyl sulphides by interaction of a thiol and a nitro- olefin is described by R.L. Heath and A. Lambert.184 The reaction is a general one and in many cases the 2-nitro-alcohols or their esters e.g. nitrates can be used as convenient laboratory substitutes for the nitro- olefins. R‘OH R,C:CR*NO + R’*SH + R,CCHR*NO I I SR‘ R,C*CHR*NO + R’oSNa + R,C*CHR*NO + NaNO SR’ I 0-NO The thiols add to nitro-olefins with greater ease than do alcohols and the reaction carried out in an alcoholic solvent yields almost exclusively the sulphide. lSo K. W. Rosenmund Ber. 1913 46 1034 ; J. Meisenheimer and F. Heim ibid. lS1 J. Loevenich J. Koch and U. Pucknat ibid. 1930 63 636. lS1 U.S.P. 2,391,815. lS3 A. Lambert C. W. Scaife and A. E. W. Smith J.1947 1474. 1s4J. 1947 1477. 1905 38 467. LEVY AND BOSE THE AlXPEA!IXfY NITRO-COMPOUNDS 389 Hydmgem dphide mcts with nitro-olefins in alcoholic solution giving a mnixture of 2-nitroaJkylthiol and di-(2-nitroalkyl) sulphide,l44 thus R,C:CR*NO + H,S + R,C(SH)*CHR*NO + R,C*CHR-NO I I S I R,C*CHR*NO The nitroalkyl sulphides can be oxidised with hydrogen peroxide to nitro- alkyl sulphones or reduced catalytically to aminoalkyl sulphides; the latter on oxidation give aminoalkyl sulphones also formed by catalytic reduction of the nitroakyl sulphones. H,O * R’*S=CR,*CHR*NO + R’*SO,*CR,*CHR*NO I I iHa H*O* iH1 R’*S*CR,*CHR*NH + R’*SO,*CR,*CHR*NH Addition of Sodium hTydrogen Su1phite.-Although the addition of sodium hydmgen sulphite to such ethylenic systems as acrylic acid styrene and ally1 alcohol has often been reported it is only recently that such addition to nitro-olefins hm been described.A simple synthesis of 2-nitroalkane- sulphonic acids and by reduction of 2-aminoalkanesulphonic acids ha,s been achieved by R. L. Heath and H. A. Piggott.la5 The reaction of nitro- olefins with sodium hydrogen sulphite is general and proceeds in aqueous or aqueous-alcoholic solution without catalyst to give good yields of sodium 2 -nitroalkanesulphonates. As an example 1 -nitro- 2 -met hylprop - 1 -ene (LXV) affords sodium l-nitro-2-methylpropane-2-sulphonate (LXVI) and the reaction is successful with nitroethylene the isomeric nitropropenes nitrostyrene and furylnitroethylene (CH,),C:CH*NO + (CH,),C( SO,Na)*CH,*NO Sodium sulphite reacts similarly giving the disodium salt (Le.the di-salt of the mi-form of the nitro-sulphonic acid) .whilst sulphurous acid affords the free nitro-sulphonic acid. (LXV.) (LXVI.) Na8Oa H*SO NO,*CH:CH*CH + NaO*NO:CH*CH(CH,)*SO,Na NO,-CH:CH*CH + N0,*CH2*CH(CH,)*S0,H Catalytic reduction of the nitrosulphonic acids in aqueous solution gives aminosulphonic acids. For example taurine (2-aminoethanesulphonic acid) is formed by reduction of the sodium 2-nitroethanesulphonate resulting from nitroethylene and sodium hydrogen sulphite. Addition of Amines and Ammonia.-The addition of amines to nitro- olefins has been described in the older literature. H. Wieland and E. Sakellarios l5’ added a d h e to nitroethylene and obtained 2-nitroethyl- aniline. D. E. Worrall186 attempted to add some forty bases including lS6 J. 1947 1481.186 J . Amer. Chem. Soc. 1927 49 1598. 390 QUARTERLY REVIEWS aliphatio and aromatic aminea and hydrazinee to o-nitrostyrene but found that only thirteen gave isolable adducts. Aniline and p-toluidine were the only common aromatic bases and piperidbe the only secondary amine which gave characterised adducta. A recent investigation 187 of the addition of ammonia and primary and secondary aliphatic and aromatic amines to nitro-olefins hw shown that in general the reaction p r o c d ectsily but the yields are variable owing to the inherent instability of the 2-nitrotlllrylamines. The products from aromatic amines are weaker bases than those from ammonia or aliphatic amines and are therefore more stable. Hydrogena- tion of the nitroamines over Raney nickel afforded derivatives of ethylene- diamine.Ammonia gave only mononitroalkylamines but di-2-nitroethyl- aniline was prepared in two stages from nitroethylene and aniline ; it is a very unstable weak base the hydrochloride of which on being heated with water loses a molecule of nitroethylene and regenerates 2-nitroethylaniline. CH,:CH*NO + Ph*NH --+ Ph.NH*CH,*CH,*NO CH,SH*NOI Ph*N( CH,*CH,*NO,) Addition of Nitro-parafiins.-Products derived from the addition of nitro-paraffins with nitro-ole& have occasionally been reported in the literature. D. E. Worrall lBS recorded the addition of phenylnitromethane to a-nitrostilbene giving 1 3-&nitro-1 2 3-triphenylpropane (LXVII). Ph CH .NO Ph*CH:C*Ph Ph*CH*CHPh*NO I I No2 Ph-CH*NO (LXVII.) H. B. Hass 189 has recently shown that one of the products from acetone and nitromethane is 1 3-dinitro-2 2-d.infethylpropane and demonstrated that its formation was due to the addition of the nitromethane to l-nitr0-2- methylprop-l-ene formed in an earlier stage of the reaction sequence out- lined on p.375. The results of a systematic investigation 190 have shown that the addition of nitro-paraffins to nitro-olefins is a general reaction in which yields are extremely variable and dependent on the nature of the nitro-olefin nitro-paraffin and conditions employed. Thus 2-nitrobut-2- ene (LXVIII) and 2-nitropropane give 47% of 2 4-dinitro-2 3-dimethyl- pentane (LXIX) in the presence of sodium ethoxide but yields are usually much lower. CH,*CH:C(CH,)*NO + (CH,),CH*NO --+ CH,*CH*CH(CH,)*NO I (LXVIII.) (CH,),C*NO (LXIX.) The properties of the 1 1- and 1 2-dinitro-paraffins have been shown earlier in this review to be " abnormal ".The 1 3-dinitro-paraffins how- ever have all the normal properties of nitro-paraffins. They dissolve in alkalis and give di-a- bromo-derivatives which are useful for characterisation la' R. L.- Heath and J. D. Rose J. 1947 1486. la8 J . Amer. Chem. Soc. 1935 57,2299. lag In&. Eng. Chem. 1943 86 1151. A. Lambed and H. A. Piggott J. 1947 1489. LEVY AND ROSE "HE ALIPHATIC NITRO-COMPOUNDS 392 purposes. Reduction preferably with hydrogen over Raney nickel affords 1 3 diaminee. Addition of at.ignurd Reage&.-The formation of nitro-pmai3i.n deriva- i h s in high yield by the action of Grignard reagents on di- and tri-phenyl- nitm%.lenm has been described by E. P. Kohler and J. F. Stone.1*1 Under skyar conditions the lower nitro-olefins give only poor yields of nitro-paraf€hqlB2 but by carrying out the reaction below 10" and avoiding an excess of oyganometallic halide secondary reactions are almost com- pletely suppressed and yields are much improved.lB3 If an excess of Crignard reagent is used nitro-paraffins oximes and reducing bases are formed the last being probably hydroxylamines.The initial steps consist of Addition chiefly 1 4- of the Grignard reagent to the conjugated system produce a complex (LXX) which on hydrolysis with water gives the nitro- EtwaiXn ; but since (LXX) contains the system C = N -+ 0 addition of a s%ond molecule of Grignard reagent occurs to give the complex (LXXI) wilich on hydrolysis yields the oxime (LXXII). The formation of reducing bamA in the reaction can be explained only by the assumption that part of the cTiginal nitro-olefin undergoes 1 2-addition of Grignard reagent to give a complex (LXXIII) which on further reaction with the Grjgnard reagent give 3 NN-dialkylhydroxylamines a reaction quite analogous to the reaction of Clrignard reagents on nitro-paraffins (see p.378). The scheme below outlines the suggested course of the reaction of ethylmagnesium bromide with 1 -nitro-2-methylpropene. 0 / MgEtBr + Me,C:CH*N 0 7 Me,C.CH:N Me,C:CH*N Me,C'*CH,*NO Me,C-CH*N( 0 *MgBr ) Dialkylh @ox ylamines I I (LXXI.) Et Et I E t Me,C--C=N*OH I I Et Et (LXXII.) By restricting the reaction outlined above to low temperatures and equimoles of the reactants a series of new nitro-parailim hw been J . Amer. Chern. Soc. 1930 53 761.10a G. D. Buckley J. 1947 1494. ma Idem J. 1947 1497. 392 QUARTERLY BEVIEWS synthesised. These on reduction with hydrogen over Raney nickel afford the corresponding amines in high yield and the chief value of this synthesis is in the preparation of a wide range of hitherto inaccessible primary aminw. Addition of Hdrogen Cyanide.-The addition of potassium cyanide to a nitro-olefin was described by M. H0llemann,~~4 who showed that o-nitro- styrene and potassium cyanide gave two stereoisomerides of I 4-dinitro- 2-cyano-2 3-diphenylbutane Ph*CH:CH*NO + HCN + [Ph*CH(CN).CH,*NO,J - Ph*CH:CH-NO (LXXIV.) Ph *C( CN).CHPh*CHa*NOs I CHa*NO (LXXV.) Although the intermediate nitro-cyanide (LXXIV) was not isolated by Hollemann it is clear that this must be formed as a first step in the reaction.A recent investigation has shown that such addition of hydrogen cyanide to nitro-olefhs is a general reaction,lg5 and is best carried out in the presence of a few moles per cent. of potassium cyanide as catalyst hydrogen cyanide alone will not react. By this method a series of 2-nitroalkyl cyanides has been synthesised from nitropropenes and nitrobutenes. A very interesting feature of these substances is their behaviour on reduction. Taking nitro- tert.-butyl cyanide (LXXVI) as an example catalytic reduction with nickeland hydrogen at ordinary temperature and pressure afforded a mixture of four products (a) amino-tert.-butyl cyanide (LXXVII) (1.5%) ( b ) 1 3-diamho- 2 2-dimethylpropane (LXXVIII) (1~5%)~ (c) p-amino-aa-dimethyl- propionamide (LXXIX) (!joy0) and (a) 5 5-dimethyl-2-( l-carbamyliso- propy1)hexahydropyrimidine (LXXX) (20%) ‘(CH,),C(CN)*CHa*NH (LXXVII.) I (CH3)aC(CH,*NHJz (LXXVIII.) 4 ( CH,),C=CH*NO €EN] H (CH,),C(CN)*CHg*NOa -+ (LXXVI.) (CH3 ,p<CH,NHa (LXXIX. ) CO*NH NH CO*NHa CH CH-C( CH,) I I I(cH31P,cH{ &I3 (L-4 The hexahydropyrimidine (LXXX) was recognised by the fact that with reagents for amines it gave the picrate hydrochloride and benzoyl deriva- tive of the dimethylpropylenediamine (LXXVIII) whilst with 2 4-dinitro- phenylhydrazine hydrochloride it gave the derivative of 8-carbamyliso- Rec. Trav. chim. 1904 aS 283. lo6 G. D. Buckley R. L. Heath and J. D. Rwe J. 1947 1600. LEVY AND ROSE THfc ALIPHATIU NITRO-COMPOUNDS 393 butaldtthyde. The mode of formation of this hexahydropyrimidine and of the arride (LXXIX) is not clear.The latter cannot be formed by a straightforward hydration of the cyano-group by the water formed from reduction of the nitro-group since an authentic specimen of the amino- cyanide (LXXVII) on treatment with Raney nickel catalyst and two molea of water in methyl alcohol (Le. the conditions of the hydrogenation without the hydrogen) was completely unchanged. It is probable that an 60- oxazole is formed as an intermediate product and is split by hydrogen to the amino-r mide. Applications of the Ajiphatic Nitro-compounds and their Derivatives The four commonest nitro-paraflins-nitromethane nitroethane l-nitro- propanc and 2-nitropropane-have been commercially available in large quantities for the last six years in America and have recently become available in pilot-plant quantities in this country.Six yam is a short time in chemical development and it is much too early to forecast the extent of their ultimate utilisation. Certain uses have however already been describl d. Dircct Uses of the Nitro-parafins.-The nitro-par&ns are good solvents for a wide range of organic substances. According to C. L. Gabriel lD6 of the Commercial Solvents Corporation nitro-paraffi have considerable solvent power for organic eaters of cellulose hcluding the difficultly soluble cellulose triacetate and cellulose acetobutyrate-the solvent power of 1- and 2-nitropropanes for the latter being about the same as that of butyl acetate for nitrclcellulose. The nitro-paraffi are among the most powerful solventa known for vinyl chloride-vinyl acetate co-polymers (“ vinylito ” resins) aad the soliitiona have a much lower viscosity than solutions in the more commoilly used higher ketones such as methyl isobutyl ketone.Solutions of nitrclcellulose in the nitro-parah have in the presence of alcohols a high tolerance for diluents. Othcr solvent uses are for oils fats and waxes and for synthetic rubbers such as “Buna N” “Chemigum” and “Hycar OR”. Advanfrtgea in their uee are low toxicity mild and non-persistent odour a medium rate of evaporakion (1-nitropropme has the mme evaporation rate as n-butyl acetate and comparatively low inflammability. Shell Development Co.lD7 have pz,tented their use &B selective solvents in the refining of petroleum. The use of nitro-parafbs in preparing accelerated rubber camants haa been described by A. W.Campbell lQ8 and patented by the Commercial Solvent 1 Corporation. lQQ The use of the nitrates of the nitro-alcohols as explosives will be referred to latei. Hercules Powder Co. have patented as explosives mixtures containing simple nitro-paraffins.200 Nitr+ahhoZs.-A great deal of work has been done on the use of the ”’ C16WL. In&wrtTkS 1939 45 664. lg7 U.S.P. 2,023,376 ; 2,019,772. 198 In& Eng. CicSm. 1941 88 809; 1942 84 1106. leg U.S.P. 2,261,220; 2,297,871. #O0 U.8.P. 2,326,064 ; 2,326,OM. 394 QUARTERLY REWEWS nitrates of the dtro-alcohols nitro-glycols and nitro-triols w explosives. Their close structural relationship to glyceryl trinitrate nee& no further emphasis. However the trinitrate of tris(hydroxymethy1)~tromethane is now considered of no value,201 in spite of the great attention it has received it does not meet the stability requirements of a modern explosive.Nitro- isobutyl nitrate has been described as a sensitiser for safety explosives containing ammonium nitrate,202 and the dinitrate of 2-nitro-2-methyl- propane-1 3-diol was patented as an explosive as long ago as 1928.203 The esters-particularly the acetates propionah butyrates and lactates-of the nitro-alcohols have been mentioned as plasticiser~,~4 and special mention has been made of methyl 2-nitroisobutylphthaltlttte rn a plasticiser for cellulose acetate.205 Other direct applications of the nitro-alcohols which have been described are as heat sensitisers for rubber latex198 and as finishing agents for textiles. 206 Amino-dcohols.-The amino-alcohols amino-glycols and ,amino-triols are made from the nitro-alcohols by simple catalytic reduction; of them 2 -aminobutanol 2- amino -2- methylpropanol 2 - amino-2-methylpropane- 1 3- di 01 2 - amino-2-ethylpropane- 1 3 - diol and trk( hydroxymethy1)methyl- amine are commercially available in America.The main applications of the amino-alcohols are as emulsifying agents and considerable use was made in America during the war of 2-amin0:2- methylpropanol in the preparation of emulsion camouflage paints. m6 Amino-alcohol soaps (laurates oleates stearates) are soluble in water and in a number of organic solvents such as alcohols benzene glycol and act &s stabilisers for either oil-in-water or water-in-oil emulsions,e01 and have been used in the compounding of floor polishes cosmetics automobile cleansers and leather dressings.Chloronitm-parafins.-The pesticidal activity of chloropicrin is well known and since the commercial production of nitro-para5 began chloropicrin in the United States has been made from nitromethane. Any aa-dichloronitro-paraffin has insecticidal activity,209 and an insecticide based on 1 l-dichloro-l-nitroethane is marketed under the trade name ‘‘ Ethide ”. It is non-injurious to man most fabrics and many foodstuffs but is as effec- tive as chloropicrin against most The chloronitro-paraffins are good solvents for a number of synthetic 201 H. B. Hm Ind. Eng. Chem. 1943,35 1146. m* Hercules Powder Co. U.S.P. 2,330,112. 204 Commercial Solvents Corporation U.S.P. 2,233,607 ; 2,177,767 ; B. M. Vender- ketones. Such soaps which are powerful emulsifying agent8,207 may 211 F.H. Bergeim U.S.P. 1,691,956. bilt a m . to Purdue Research Foundation U.S.P. 2,233,666. W. E. Scheer Chem. Industrim 1943 52,473. H. Robinette A m r . Dyestuff8 Reporter 1942 31 676. 207 B. M. Vanderbilt assr. to Commercial Solvents Corporation U.S.P. 2,281,177. 208 Idem assr. to Purdue Research Foundation U.S.P. 2,247,106. z10 W. C. O’Kane and H. W. Smith J . Econ. Entmnology 1941,84,439. H. B. Haos assr. to Commercial Solvents Corporation U.S.P. 2,281,239. C. E. Woodworth M. 1943 86 338. LEVY AND BOSE THE ALIPHATIC NITRO-COMPOUNDS 395 plastics. A list of such substances soluble in l-chloro-l-nitropropane is given by A. W. Campbell and J. W. B~rns,~la but there is no evidence of commercial application. A patent assigned to Universal Oil Products 21s descrih the u ~ e of certain chloronitro-parah as solvents for the alkylation of aromttic hydrocarbons by olefins catalysed by aluminium chloride.The chloronitro-paraffin dissolves both the aluminium chloride and the aromatic hydrocarbon. l-Chloro-l-nitropropane has also been described as a stabi- liser for rubber cements.lQQ Hydi*oxyZumine.-Most of the hydroxylamine sold in the United States is prepared by acid hydrolysis of primary nitro-paraffins chiefly 1 -nitropropane and hydroxylamine can logically be considered as a nitro-parafi derivative. Possible outlets for it are discussed by C. L. Gab~iel.2~4 The propionic acid simultaiieously produced in the hydrolysis of 1 -nitropropane has probably found a ready outlet as calcium propionate an anti-mould agent or been used in the production of solvents. J. D. R. sls Indiu Rubber World 1942 107 169. 418 U.S.P. 2,302,721. a14 Ind. Eng. Chrn. 1940 82 887.

ISSN:0009-2681    

DOI:10.1039/QR9470100358

出版商:RSC    

年代:1947

数据来源: RSC