1508 LAPWORTH AND HA”: CXLIX.- The Mutcrrotation of Camphorquinonehydr- axone and Mechanism of Simple Desmotropic Change. By ARTHUI~ LAPWORTH and A. C. OSBORN HANN. OF the various types of isomeric change, that which involves a change of position of one hydrogen atom only, as in a simple desmo- tropic change, would, for various reasons, appear to be the most simple, and probably the most easy to investigate. We do not intend to review the many important contributions already made to this subject, in particular since such a course would involve, for the most part, a repetition of much with which Lowry has somemhat recently dealt in his suggestive paper on mutarotation andTHE MUTAROTATION OF CAMPHORQUINONEHYDRAZONE. 1509 reversible isodynamic change (Trans., 1899, 75, z l l ) , but a brief reference to some views which have already been advanced appears desirable.Perhaps the theory most commonly used with regard to the mechan- ism of reversible desmotropic transformations is that which assumes an alternate addition and elimination of the elements of water or of some other hydrogen compound in the following way : H,O+:C=&OH f-, : c H - ~ < ~ ~ OH +-+ :CH-*=O+H,O or Lowry (Trans., 1899, 75, 221) has assumed that in the acceleration which is almost invariably effected by bases the first change con- sists in the formation, in the case of a nitro-compound, of a salt of the iso.form, in this way : NaOEt + :CH*NO, -+ :C:NO*ONa + HOEt, and that the interaction of this compound with water or alcohol leads, on the one hand, to the free iso-form, and on the other, to the free normal form of the nitro-compound. This view possesses distinct ad- vantages over the foregoing one, but appears to disregard certain very important points.Laar’s theory (Bey., 1885, 18, 648), which involves the conception of a free hydrogen atom vibrating continuously between the two different positions of attachment, was not put forward to explain the occurrence of desmotropic change, and, in fact, the reasons for its advancement can no longer be said to exist. It need not, therefore, be taken into consideration a t present, although it bears a superficial resemblance to certain views which appear later in this paper. An altogether different view of the change of enols into ketones has been advanced by Bruhl (Sei-., 1899, 32, 2329), who has found that the velocity of the conversion of ethyl mesityloxideoxalate from the enolic into the ketonic form is closely related to the dielectric constant of the medium, and in such a manner as to support his prediction that the influence of the medium would be found to be a function of its ionising power.Briihl’s view of the change may be expressed briefly as follows. The enols are hydroxylic substances, weak acids, and electrolytes, and dissociate into hydrogen ions and a residue :C:d-O-; these residues have a tendency to reunite with hydrogen ions t o form a non-ionised compound of a ketonic type, :CH*k = 0. This view, if correct, contains, we venture to beIieve, the elements of a real explanation, lending itself t o experiment and development in1510 LAPWORTH AND HANN : a way which the others can scarcely be said to do, and provides, more- over, a means of ascertaining what is the probable part played by traces of catalytic agents in accelerating the velocity of change. Although it may always remain a matter of opinion whether the presence of catalytic agents is absolutely essential in effecting changes which, in their absence, do not progress with measurable velocities, it is generally recognised that it is difficult to overestimate the im- portance which m%y be attached to their influence, as Armstrong in particular has repeatedly insisted.It appears, therefore, not unlikely that a study of the induence of traces of different agents in accelerat- ing or retarding a change may be found to serve as one of the most satisfactory bases on which to form useful views of the mechanism of the change.Bruhl’s view, as above stated, is, however, by no means complete, as in many cases the enolic forms do not become completely converted into the ketonic forms, although the latter have no measurable conductivity, but in solution a state of equilibrium between measurable amounts of the two forms is finally reached. From the standpoint of a theory of isomeric change which involves the assumption of ionisation, there can scarcely be chosen a more suitable case of change for investigation than that in which the migrating part is a hydrogen atom, since there can be little or no doubt with regard to the charge of the migratory ion, whereas in the majority of other cases this difficulty is a very real one.Assuming, then, the broad principles of Briihl’s suggestion as altogether reasonable, it is a t least worth while endeavouring to discover to what they may lead, With respect to the transformation of the ketone to the enol, the condition of equilibrium between isodynamic forms is not disturbed by the presence of traces of impurities, a statement which is supported both by thermodynamic principles and by direct experiment, even though the impurities may be present in amount sufficient to acceler- ate very greatly the velocity with which the position of equilibrium is attained (compare Lowry, Zoc. cit., 222, 243). Hence any agent which affects the velocity of either change must affect the reverse one in the same manner, and in the case under consideration it seems difhult to avoid the conclusion that in the change of ketone to enol the same type of process must go on under any set of conditions as in the reverse one, that is to say that these compounds are also ionised to a minute, although quite immeasurable, extent, affording hydrogen ions.We thus arrive a t a view similar in a sense to those of Thiele (AnnaZen, 1899, 306, Il4), Henrich (Bey., 1898, 31, 2103), and others who regard the hydrogen of the group X = Y - ZH as ‘‘ directly replaceable,” and Vorliinder has pointedTHE MUTAROTATION OF CAMPHORQUINONEHYDRAZONE. 15 11 out that the grouping mentioned is characteristic of acids (Bey., 1901, 34, 1633). It must therefore be inferred from Briihl’s view that the mutual and spontaneous interconversion is to be represented as the result of the processes -t- X=Y-ZH ++ X=Y-Z+H ++ XH-Y=Z; but it would appear from this that the velocity of change should be simply dependent on the number of times per second that the hydrogen ions come in contact with the negative ions or “residues.” This would be proportional both to the concentration of the hydrogen ions and to that of the negative ions, and therefore to the product of these concentrations.This product, however, in any solvent, bears a constant ratio to the concentration of the undissociated part of the substance, hence the addition of a trace of an agent which only altered this ratio, as, for example, an acid or a base, should produce no marked alteration of the velocity. Or, representing the concentration of the individuals in the above scheme in order as c,, c2, cQ, c4, and the velocities as k,, Z1, kr2, k2, we have dc, = c2c3k, - c&’, and ’% = ~ 2 ~ 3 k 9 , - c4kt2 dt d t hence the rate of change from, say, undissociated enol to ketone is that is to say, a velocity which is of the same order as that of the ionisation, and therefore for an enol of measurable conductivity (where at least we are probably justified in assuming a practically instantaneous ionisation) very great.Again, addition of a small quantity of an acid or base does not alter ~ 2 ~ 3 , and produces no appreciable effect on c1 or c4, the concentration of the undissociated parts of the enol and ketone. Further, in equilibrium we should have dc 3 = 0 = -4 or c2c3k1 = clZl and cacsk‘2 = c4k2 ; dt dt hence 5 = ($)@) = $ or the ratio of the concentration of 04 the t,wo forms would be equal to the inverse ratio of their dissociation constants.Thus, when one form is a conductor and the other is not, the former would not be preseut in appreciable amount, a result which is altogether at variance with what is known of the facts. Some other time factor must therefore be involved, probably small in com-I512 LAPWORTH AND HA": parison with those above considered, and it seems obvious that this must be connected with the internal change of structure which the molecules undergo during mutual interconversion. I n our opinion, this velocity is best introduced in accordance with the suggestion already made by one of us (Trans., 1901, 79, 1266) that the internal change takes place only in the ion, and simply resuIts in a move- ment of the position of '' free affinity." Introducing this, we have the following scheme to represent the changes occurring : X:Y*ZH t-, X:Y*Z- +H x:y*z- .f-, -x*y:z -X*Y:Z+H t-, XH*Y:Z Here the velocity of change will be proportional to the absolute Concentration of the organic ion and will of course be increased as the concentration of the hydrogen ions is diminished and vice uersb.It is not difficult to show that it will be inversely proportional to the concentration of the hydrogen ions, and this is consistent with the properties of many organic compounds. Indicating the concentration of the second organic ion by c5 and the two new velocities introduced as k, and k', respectively, we should have in equilibrium Hence or, the ratio of the concentrations of the two forms is independent of the concentration of the hydrogen ions and therefore unaffected by catalytic agents, but is inversely proportional to the respective veloci- ties of internal rearrangement and also to the dissociation constants.It follows from the latter that where a measurable amount of conduct- ing enol is in equilibrium with an apparently non-conducting ketone, the ion of the latter is transformed into that of the former with a velocity which is very great in comparison with the reverse change. The conclusions to which the foregoing view leads are in accordance with the behaviour of certain isodynamic pairs of this kind. Thus, Lowry has found that the conversion of nitrocamphor into the &o- nitro-form is accelerated by bases but not by acids (Zoc.cit., p. 221) ; he has not stated however, whether any retardation was observed, whereas we find that in the case of camphorquinonehydrazone it is apparently possible to arrest the isomeric change a t any stage by the introduction of a trace of an acid. Moreover, it appears to be a very general rule that bases of all types accelerate the change, so that inTHE MUTAROTATION OF CAMPHORQUINONEHYDRAZONE. 1513 most cases the foregoing process may play an important part. Never- theless, there are other cases where, on the contrary, acids may accelerate the change. Thus, in converting dibenzoylmethane into the corresponding enolic form acids are used (Annalen, 1899, 308, 219j, and Forster found that formic acid was the most effective solvent in transforming enolic benzoylcamphor into the ketonic form (Trans 1901, 70, 997).These facts might possibly be explained rather as the result of the retarding influence of the acids on the mixture in a state of equilibrium preventing the reversion of the usually more unstable form as crystallisation occurred. No such ambiguity, however, exists in the case of menthyl acetoacetate which the authors have recently had under examination. Here a trace of an acid accelerates the change very greatly, hence it appears that in some cases a t least a new factor must be taken into account. One of the authors has already suggested (Proc., 1901, 17, 95) that hydrogen ions may themselves causg the change by forming complex ions with the compounds : in other words, that the latter may act as feeble bases.I n this case, as in the previous one, the process involves a structural change, and this change, as before, may be conventiooally introduced by making the assumption that the organic ions in the two cases are not identical but are mutually interconvertible by a simple internal rearrangement. The imaginary process is easy to understand by reference to a model, but it is not so easy to represent as that in the previous case. However, it may be expressed as XH-Y=Z+H t+ XH-YTZH XH-YTZK ++ XH;-Y=ZH, XHYY-ZH ++ XY-ZH+H where the dot indicates the direstion in which the ‘ I free affinity ” of Y is temporarily disposed. Here, as before, the velocity of change will be proportional to the concentration of the organic ion if the ionisation velocities are very rapid in comparison with the speed of structural change, aria there- fore proportional to the concentration of the hydrogen ions, It is thus possible that a desmotropic or tautomeric change may be the result of one or both of two superposed reactions, one due to ion- isation in the compound itself, accelerated by bases and retarded by acids and the other due to an additive phenomenon, accelerated by acids and retarded by bases.Both reactions in compounds which are not associated will be unimoleeular in type where the concentration of the hydrogen ions does not vary, as will also be the velocity obtained by their superposition. That the two actions are different in kind appears to be shown by the fact that compounds which are not1514 LAPWORTH AND HA": very different in their susceptibility to bases may be altogether different i n their behaviour towards acids.How far such views as these may bear the test of further investi- gation remains to be seen, but i t will certainly be interesting toascer- fain whether some isomeric changes of other types may not be explain- able by the aid of similar views. Clearly two types of catalytic agents at least may be sought for, one type acting by diminishing the concentration of ions representing the labile group without altogether removing it, the second increasing the concentration of these ions and effecting the change by a process which may be regarded as one of addition. EXPERIMENTAL.Desrnoh-opic Forms of C a m p ~ o r ~ u i n c n e h d ~ ~ z ~ e . Camphorquinonehydrazone may be prepared by two methods, namely, by the action of phenylhydrazine on camphorquinone or by the action of diazobenzene on camphorcarboxylic acid. Betti has found (Ber., 1899, 32, 1995) that the partially purified material ob- tained in the latter process melts indefinitely at temperatures varying between 154' and 1 6 5 O , but when purified by crystallisation from hot alcohol or from benzene and petroleum at 178-180'. We have obtained exactly similar results when preparing the substance from camphorquinone and phenylhydrazine. The melting point of the most highly purified material obtained was about 180-181', a melting point practically identical with that given by Betti.On analysis : 0.2824 gave 0*7790 CO, and 0,2099 H,O. C = 75.2 ; H = 8.2. C,6H2,0N, requires C = 75.0 ; H = 7.8 per cent. The substance obtained from alcohol and melting at 180-18lo presents all the appearance of a uniform substance to the naked eye, and under the microscope is seen to consist entirely of small, well-defined plates which are apparently hemihedral in character, being usually pointed' at one end and truncated a t right angles to their length at the other; from some other solvents, as, for example, benzene, it frequently separates in well-formed, flattened prisms. There is, in fact, every reason for regarding the material as a simple substance, free from any appreciable quantity of a second compound. Under certain conditions specified by Betti (Zoc.cit., p. 1998), a mixture of two compounds is obtained, notably when the material is precipitated from alcohol by means of alkalis, or when the substance of high melting point is kept for a short time in a fused state. The product then usually melts at about 1 6 5 O , and no longer presents the appearance of a pure compound, and the melting point is somewhatTHE MUTAROTATION OF CAMPHORQUINONEHYDRAZONE. 1515 indefinite. Similar mixtures are frequently deposited from hot benzene. Betti describes a third substance which separated when piperidine was added to the solution of the compound in benzene; he states that it separates a t once in small crystals and melts at 1 5 5 O , and regards it as a uniform substance as it differs from either of the fore- going materials in affording no appreciable coloration with ferric chloride in benzene solution.It is in this particular alone that we have not been able to confirm Betti’s results, as we have not succeeded in isolating a uniform substance other thnn that melting at 180°, although we have obtained specimens of material which melted at temperatures not far removed from 1 5 5 O . For various reasons, which will be indicated later, we believe that the process used by Betti and based on the mistaken view of Schiff (Re?.., 1898, 31, 601), that piperidine favours the production of the ketonic form, would actually tend to be disastrous so far as the isolation of the usually more elusive member of such a desmotropic pair is concerned. As the substance meeting at 180°, when dissolved in dry benzene, gives a reddish coloration on addition of an ethereal solution of ferric chloride which is not noticed when the material of lowest melting point is employed, it would appear that there is no fault to be found with Betti’s conclusion that it probably represents the pure enolic form C*H,,<8.C*N:NPh. *H This coloration, however, is never intense when the pure compound is used, although a small quantity of some un- known impurity occasionally renders it s o ; it is, in fact, easily masked by the yellow colour of the solution OF the pure enol, or by that of the ferric chloride when this substance is used in too large amount. For these reasons we think that it is possible to attach undue importance to the apparent non-production of the coloration with the material of low melting point.Mzcturotation of Camp~orpwinonehydru~one, AS was to be expected, the hydrazone in certain solvents exhibits the phenomenon of mutarotation (Lowry, Trans., 1899, 75, 211), due, no doubt, to isodynamic change as in the other cases. It is unfor- tunate that many of our experiments lose some of their value as it was not possible to extend them to both of the pure desmotropic forms, but certain interesting conclusions may be safely drawn from them, and they have at least helped to render it certain that piperidine assists only in establishing equilibrium between the two forms, and does not favour the existence of either form, as Betti’s conclusion would appear to assume. Observations of the rotatory power of the compound in alcohol1516 LAPWORTH AND HA": showed that mutarotation could not be detected in this solvent, either when the pure enolic form or mixtures of low melting point were used; in both cases the rotation of 1 per cent.solutions immediately after making up was [ a ] , +436* to +440°, hence i t would seem that equilibrium is here attained very rapidly. I n ethyl acetate, a t 1 per cent. concentration, the initial rotation was [.ID +405O, and fell rapidly to about + 380". I n benzene, which me have found t o be the solvent in which the mutarotation is most easily followed, the initial rotation of the pure enolic form was about [a], + 320' t o 325O, and fell slowlytovalueswhich varied widely with the concentration. When the original rotation in a 2 dcm. tube was 2*81', it fell in less than two days to a constant value, 2*62', whilst when the rotation was initially 0*98O it fell t o 0.53' in about the same time ; in an extreme case a solution having originally a rotation of 0.38' gave finally the number 0.13'.Thus the ratio of the initial and final rotation varies from about 1.1 in 0.5 per cent. solution to nearly 3 in 0.1 per cent. solution, We have noticed that, when the change is going on, the solution usually becomes darker in colour, but this is not invariably the case. Lowry has observed a somewhat similar effect with nitrocamphor (Trans , 1899, 75, 2). It is unfortunate that it cannot be determined whether the curves represent uni- or bi-molecular changes, as it is impossible to calculate even approximately the actual proportion of ketone and enol present.I n making up solutions in benzene, we met with considerable difficulty in determining the exact initial specific rotations, for owing to the fact that the material dissolved very slowly in cold liquids, attempts to bring weighed quantities rapidly and completely into solution were unsuccessful, so that by the time we were able to observe the rotation this had already fallen appreciably. We were therefore compelled to resort t o the method of shaking the substance with the solvent for a short time, filtering from undissolved matter, and reading a t once, afterwards estimating the amount dissolved by evaporaking a known volume t o dryness. Such a process, how- ever, was not very accurate, and also involved waste of carefully purified material. The somewhat remarkable variations of the final specific rotatory power of the solution with dilution would be explained if tha enolic form were largely polymolecular in benzene solution, or if the specific rotatory power of the ketonic form varied greatly with the dilution.As a determination of the moIecular weight of the enolic form in benzene by the cryoscopic method made as soon as possible after solution indicated that the solution had a normal molecular weight, the second view appeared to be the correct one, and received confirmation from the following facts,THE MUTAROTATION OF CAMPHORQUINONERYDRAZONE. 151 7 A strong solution of the enolic form in benzene was left until it showed a constant rotation, which required about 40 hours.A measured pdrtion of the solution was then diluted to a known bulk with benzene and the rotatory power of the resulting solution at once measured ; the number obtained was much lower than that calculated by multiplying the number for the original solution by the ratioof the concentrations in the two cases, and agreed fairly well with the number which had been obtained as the final rotatory power in a solution of similar concentration ; moreover, no mutarotation was noticed. Several experiments of this kind were made, and in every case similar results were obtained. It would therefore appear that the ratio of concentra- tion of the two forms when in equilibrium in benzene solution is nearly independent of the concentration. Having prepared various specimens of material of low melting point by various methods, inclucting that recommended by Betti as leading to the production of the pure ketonic form, it was thought that an examination of their behaviour in benzene solution would finally decide the question whether they were really composed largely of the enolic form, for had this been the case there was every reason to believe that their rotation in benzene would have been lower than that of material which contained both forms in equilibrium, and that a rise with time would be observable.As a matter of fact, it was found in every case that in dilute solution a very considerable fall instead of a rise occurred, and we conclude that our materiais, a t least, have always contained a considerable proportion of the enolic form.Efect of Catalytic Agents on the Mutarotation,. I n determining how traces of catalytic agents affect the speed of change, no attempts were made to obtain solutions of known concentra- tion; usually, about a gram of the pure hydrazone melting a t 180° was shaken for a short time with 100 C.C. of benzene, and the solution filtered and distributed between several tubes. The initial velocities of mutarotation in the various tubes were determined by a series of observations with each. Traces of different materials were then introduced into some of the tubes and the subsequent rate of change determined as before. By plotting the results graphically, any sudden alterations could easily be detected, although this was usually un- necessary, the effect of the agents where they were efficient being at once obvious. (u) Efeot of Buses.-It is well known that the hydroxides and ethoxides of sodium and potassium, and primary and secondary organic bases are usually effective in promoting the attainment of equilibrium between tautomeric forms, andt in this case also traces of sodium1518 LAPWORTE AND HA"; ethoxide, ammonia, aniline, and piperidine accelerate the change of rotation in benzene solution in a very marked mahner; the nitro- anilines, in very small quantities, produced very little effect.When solutions in alcohol, which exhibited no mutarotation, were used, the agents did not appreciably alter the rotation, 'so that it seems fairly clear that equilibrium is reached almost immediately in this medium. It is frequently assumed that these bases act by a process of addition in the manner indicated in the introductory part of this paper ; experiments with a number of tertiary bases where addition in that sense is impossible were therefore undertaken, and it was found that these acted precisely as did the others, traces of trimethylamine, pyridine, strychnine, and brucine bringing about a great acceleration, although the amounts of the two alkaloids used were too small t o be detected by the polarimeter.As was only to be expected, the final rotations of the solutions were always independent of the speed with which the change occurred and of the catalytic agent, the differences between the rotations of the same solution in two or more tubes being of an order which was not greater than that probably due to experimental error.(6) E'eeqt of Acids.-It does not appear that any careful experiments on the influence of acids on the speed of tautomeric change have been made. As this point appears to us of the utmost importance in connection with the question of the mechanism of the change, we have devoted a considerable amount of attention to it. I n all instances we have found that the mutarotation i's retarded by acids, whether these are present in mere traces or fairly large amounts; moreover, there is a certain amount of evidence that the retardation is dependent on the strength of the acid : thus a set of tubes were filled with the same solution, and the curves of mutarotation traced by a number of readings and found to be very similar ; approximately equivalent quantities of acetic acid and of trichloroscetic acid were introduced into two of these by adding one drop of highly dilute equivalent solutions of the two acids in benzene, mixing, and taking a new set of readings.A very distinct retardation was noticed with the former, but in the case of the trichloroacetic acid the subsequent mutarotation was too slow to be detected during the course of ten days; the solution containing acetic acid had by that time fallen to a constant value, whilst the solution which did not contain any impurity had shown a constant rotation after thirty hours. The apparently complete arrest wbicb could be effected by a trace of a strong acid can most easily be shown by dipping a platinum wire into trichloroacetic acid, and then into a solution of the hydrazone contained in the cibservation tube.Small quantities of benzoic, salicylic, and several other acids produced Pimilar effects,THE MUTAROTATION OF CAMPHORQUXNONEHYDRAZONE. 1519 It may be noted that by adding a trace of a base to the solution, the mutarotation of which has been arrested, it is possible, with care, to cause it to proceed slowly once more and to arrest it a second time. The final rotation of the solution is of course independent of this treatment. Conclusion. Betti found that the material melting a t 155O, and which is supposed to be the pure ketonic form, could be reerystallised from benzene without change, providing that a trace of piperidine were present, Now it is cIear,from the above observations that bases bring about the equilibrium between the two forms, so that, even if Betti’s supposition were oorrect, it would seem that the presence of piperidine in the solution would not tend to preserve the ketonic form, but the reverse, and the presence of an acid rather than a base would probably be desirable. I n a benzene solution of the two forms in equilibrium, the solution might be nearly concentrated for both, and on cooling a srnaIl quantity of either form alone might separate, when the equilibrium would be momentarily upset. In presence of a sufficient quantity of a base, the result would be the immediate formation and separation of more of this form, and so on, whilst in the absence of an accelerator, a mixture of both would naturally be obtained. The isolation of the pure ketonic form might thus depend on the initial accidental presence of a minute quantity of that form in the solid state, and it is possible that we have been unfortunate in this respect. Nevertheless, on carefully reviewing the whole of the evidence, we cannot but regard Betti’s conclusion as to the nature of the compound melting a t 155O with much suspicion. However, the following possibility still presents itself. We have to express our cordial thanks to Signor Betti for most cour- teously acceding to our request that we might undertake this in- vestigation, as well as for giving us all the assistance in his power. Our thanks are also due to the Research Fund Committiee of the Chemical Society for a grant which helped t o defray the cost of the materials required in the investigation. CHEMICAL DEPARTMENT, GOLDSMITHS’ INSTITUTE, NEW CROSS, 8.E.