THE CHARAC‘PERISA~ION OF “ RACEMIC ” LIQUIDS. 1119 By FREDEBIC: s r i 7 m L E Y ~ ~ I P P I N G aid \ J 7 ~ ~ ~ ~ ~ ~ f JACKSON POPE. IN a paper published early this year (this vol., p. 36), we proved that Ladenburg’s general method for distinguishing between a racemic sub- stance and a mixture of enmtiomorphously related compounds (Ber., 1894, 27, 3065) is fallacious and cannot afford the criterion desired. Shortly afterwards, in a reply communicated to this Society (this vol., p. 466), Ladenburg, after admitting the justice of our criticisms, gave what he described as a ‘‘ different form ” to his previous statement of the method in question, but, as was pointed out by one of us at the time (Proc., 1899,15, 73), the alteration in form was so profound that the new statement had no principle in common with the original one, This new method, which was also described in a German version of the paper (Bey., 1899, 32, 864)-where it appeared as a spontaneous effort on the part of its author-consists in determining the solubility of the externally compensated substance with, and without, the addition of a small proportion of one of its optically active components; if the solubilities are different, then the substance is racemic, whilst, if they are the same, it is a mere mixture of the two enantiomorphs.This method, so far as it is applied to crystalline substances, is, except 4 1’ 21120 HIPPING AND POPE: in isolated cases, perfectly valid, and is merely the logical outcome of Kenrick’s work (Ber., 1897, 30,1749) and our own (Ioc.cit.). Recently, however, Ladenburg has attempted to apply it to externally compen- sated liquids with the object of ascertaining whether they are racemic or mere mixtures (Bey., 1899, 32, 1822); and in some of these experi- ments, instead of determining the solubility of the externally com- pensated substance alone and in admixture with a small proportion of one of its optically active components, he adopts the better and more convenient plan of examining polarimetrically the solutions obtained in the two cases. H e takes, for example, an optically inactive mixture of d- and I-limonene, shakes it with dilute alcohol insufficient to dis- solve the whole, then adds a little d-limonene, and shakes again ; on subsequently examining the alcoholic solution, he finds that it is optic- ally inactive, and so he concludes that d- and I-limonene do not form a racemic liquid at the ordinary temperature.Now the results obtained by such an application of the method afford no evidence whatever of the non-racemic nature or otherwise of the externally compensated liquid, and consequently the conclusions which Ladenburg draws from them have not the slightest value. The method is valid with crystalline substances, because in the case of a non-racemic mixture of optical antipodes two solid phases in contact with the solu- tion are being dealt with; whilst, in the case of a racemic substance, t o which a small proportion of one active isomeride bas been added, there is the question only of one solid phase (a racemic one), and this is in contact with a solution saturated with respect to it and partially saturated with respect to the active component.A liquid mixture of d- and I-isomerides, whether they form a racemic liquid or not, only constitutes one phase in the system, and never two, as in the case of a crystalline non-racemic mixture. The error into which Ladenburg has fallen is the more surprising, since Bakhuis Roozeboom has recently contributed a very clear discussion of the subject from the standpoint of the theory of equilibrium (Ber., 1899,32, 637 ; Zeit. physikcd, Cheni., 1899, 28, 494). We have shown, then, by the foregoing argument, that Ladenburg’s method of dealing with liquid externally compensated substances is in disagreement with the very principles from which its author professes to deduce it, namely, the principles of equilibrium.Since, however, the examination of optically active and externally compensated sub- stances in as many directions as possible is urgently desirable at the present time, in order to make certain that the laws of equilibrium have been properly applied, we have made a fresh investigation of some substances of this kind, with results which are quite in accordance with our theoretical argument.THE CHARACTERISATION OF " RACEMIC " LfQUIDS 1121 Pseudorcccemic and Dextro-Cu~nphorsu~3~~onic Chlorides. We have previously shown that the sulphonic chloride obtained from the product of the sulphonation of d-camphor with anhydrosulphuric acid, consists of a mixture of the d- and I-isomerides, the former being present in the larger proportion.For many reasons, discussed in earlier papers (Trans., 1893, 63, 548 ; 1895, 67, 354; 1807, 71, 989), we have also concluded that the externally compensated substance is not truly racemic, but that the antipodes crystallise together, forming a pseudoracemic substance. We defined a pseudoracemic substance as one in which the enantio- morphously related components are twinned together, and we may therefore take this opportunity of pointing out that the term seems to be understood in a different sense by Roozeboom (Zeit. phpsikal, Cliem., IS99, 28, 404), who attributes to pseudoracemic substances the proper- ties of isomorphous mixtures or solid solutions. As we alone, so far, have worked with pseudoracemic substances, and have observed no marked analogy between isomorphism and pseudoraccmism, we cannot share Roozeboom's views as to the nature of such substances ; possibly, however, the pinonic acids, recently examined by Fock (Zeit.Krpt. illin., 1899, 31, 479), afford a case of pseudoracemism in the sense in which Roozeboom understands it. However, according to our views of pseudoracemism, a substance such as crystalline externally compensated camphorsulphonic chloride should behave towards solvents in a manner geometrically similar to that of a non-racemic mixture of optical antipodes, and consequently, on extracting with a solvent a sample of this substance containing a small proportion of one or other isomeride, an optically inactive solution should be obtained; this we find to be the case.A considerable quantity of crude camphorsulphonic chloride was purified by recrystallisation from ethylic acetate, and a sample was ultimately obtained having the specific rotatory power [ a ] D + 13' in chloroform solution ; since cl-camphorsulphonic chloride has the specific rotatory power [ alu + 1 2 8 O , the sample contained about 45 per cent. of I - and about 55 per cent. of the cl-sulphonic chloride. Portions of 5 grams of this mixture, when agitated for 5-7 hours at constant temperature with 30 c,c. of various mixtures of light petroleum (b. p. 40-60') and chloroform, yielded solutions devoid of optical activity when examined in 200 mm. tubes in a polarimeter reading to 0*01'. Our previous conclusion is thus confirmed; crystalline externally compensated camphorsulphonic chloride is not a racemic substapce, but is pseudoracemic in the sense of our defi- ni tion.1122 KIPPING AND POPE: But, since externally compensated camphorsulphonic chloride is not racemic in the solid state, there is no reason for expecting that it would be racemic in the pure liquid state at the same temperature, and the probability that it mould exist as a racemic substance in dilute solution is even more remote, because, so far as they have been investigated, compounds proved to be racemic in a crystalline con- dition are known to be wholly resolved into their components in solution.In order, however, to study the behaviour of mixtures of d- and I-camphorsulphonic chloride in a dissolved condition, the following experiments were made : portions of 3 gramr of the same sample as before were completely dissolved in mixtures of chloroform (10 c.c.) and light petroleum (20 c.c.) ; a mixture of alcohol (30 c.c.) and water (7 c.c.) was then added to the solution, and the whole shaken during 3-4 hours at the ordinary temperature. The liquid, which at the end, as at the commencement of the experiment, mas free from crystals, separated when left at rest into two layers; the lower one (about 32 c.c.), we may call the alcoholic, the upper one, the petroleum, solution of camyhorsulphonic chloride, Both these solutions were found to be optically active when examined in a 200 mm.tnbe, the alcoholic solution showing a rotatory power of about uD+0*6O, the petroleum about ~ ~ + 0 * 9 ~ . A repetition of these experiments with 2 grams of the same sample of sulphonic'chloride afforded similar results.Now, if Ladenburg's application of the method referred to above to liquids were valid, we should have to conclude that dissolved, externally compensated camphorsul phonic chloride is a racemic sub- stance-a conclusion which, as indicated above, is scarcely within the bounds of possibility ; there are, moreover, other arguments which lead equally to the conclusion that the sulphonic chloride in the state of solution does not show the behaviour of a racemic compound, so far as the disputed method is concerned. I n the first pl;bce, the result of shaking together the two solutions of unequal quantities of the two optically active sulphonic chlorides is quite different from that ob- tained on shaking a solid mixture of n mcemic substance and one of its optically active components with a solvent ; in the former case, both solutions (the extract and the extracted) remain optically active, whereas in the latter the optical activity is wholly confined to the liquid extract ; even granting the existence of a racemic sulphonic chloride in solution, i t seems to us that, according to Ladenburg's views, one of the solutions should become optically inactive.In the second place, or rather, putting this same argument differently-if pseudo- racemic camphorsulphonic chloride become racemic when it is dis- solved in a mixture of chloroform and petroleum, it could not possibly yield anoptically inactive solution when a mixture of unequal quantitiesTHE CHARACTERISATION OF (( RACEMIC ” LIQUIDS.1123 of the two antipodes is shaken with such a solvent, whereas our experiments have led to this result. We have therefoTe no hesitation in concluding that the experi- mental evidence which we have brought forward suffices to bear out the truth of the arguments which we have advanced, and that Ladenburg’s application to liquids (or wholly dissolved solids) of the principle used in the case of solids is theoretically unsound and in- capable of yielding practical results of the slightest value. One statement by Ladenburg, and one only, Seems to be at variance with them conclusions, namely, that a mixture of unequal quantifies of d- and I-limonene yields an optically inactive extract when shaken with dilute alcohol insufficient to dissolve the whole.So far as we can judge from the very brief description of the experiment given in his paper (Zoc. cit.), it seems more probable that the alcoholicsolution contained so much water that it dissolved too small a quantity of the mixed limonenes to afford observable optical activity. Whatever the explanation of this particulnr experiment, it seemed probable on b p’iori grounds that the results obtained on extracting an externally compensated liquid, containing a small proportion of one of the antipodes, should be subject to the ordinary law of dis- tribution, just as in the case of optically inactive compounds. This view might perhaps have been put to tho test of experiment with the aid of the sample of cnmphorsulphonic chloride employed above, but this substance does not seem a very suitable one for such a purpose; partly because it gradually hydrolyses in dilute alcoholic solution (but, as we satisfied ourselves, far too slowly to appreciably affect the qualitative results already described), partly because af the necessity for employing four different volatile liquids.For these reasons, we used the ennntioniorphously related pinenes for a series of quctn titative experiments on the distribution of the components between two liquids. Dextro- and Lavo-pinene. The pinenes which we employed mere obtained from dextrorotatory American turpentine and lmvorotntory French turpentine respectively ; both were repeatedly distilled in a current of steam with addition of a little sodium carbonate and finally fractionated twice under atmo- spheric pressure, the fractions boiling at 161-162’ being taken as pinene.Repeated distillation with steam is necessary because, after mixing the two pinenes so as to obtain an optically inactive liquid and then extracting with methylic alcohol, the alcoholic extract becomes optically active unless the purification has been carefully1124 KIPPING AND POPE: carried o u t ; in such cases, the activity is doubtless due to the presence of some oxidation product, such as sobrerol. The rotatory powers of the d- and Z-pinene thus obtained having been determined, the hydrocarbons were mixed in such proportion as to give a liquid which appeared optically inactive when examined in a 200 mm. tube. Quantities of 25 C.C. of this inactive oil were then mixed with different small proportions of Z-pinene in stoppered bottles and abont 25 C.C.of 75 per cent. ethylic alcohol added to each ; the sample of Z-pinene used had the rotatory power uD-63.33' in a 200 mm. tube at 2 1 O . The bottles were then agitated during four hours at the constant temperature of 22', by which time it was judged that equilibrium would have been attained ; the liquids were then poured into separat- ing funnels, and as soon as the separation into two layers was com- plete, samples of each layer mere run into 200 mm. tubes and examined polarimetrically. The results are given in the following table : Pinene and 75 p e ~ cent. ethplic alcol~ol. - No. 1 2 3 4 5 6 7 a - a, of oil, 3;. 0 - 2-64' - 5.53 - 7'11 - 8'25 - 11.03 - 13.75 - 58 7 3 a, of solution, Y.0 - 0.19" - 0.34 - 0.46 - 0'49 - 0.75 - 0.08 - 2.66 - 13'9 16.3 1 5 5 16'8 1473 14'0 22 *1 It will be seen that in every case both the alcoholic solution and the mixture of pinenes in equilibrium with i t are optically active; further, the ratio of the optical activity of the two liquids is approxi- mately constant. Since the pinenes are miscible in all proportions with absolute ethylic alcohol and dilution with water is necessary in order to make the miscibility imperfect, i t seemed desirable to make another series of determinations with a, solvent which is not completely miscible with the pinenes. For this purpose me employed methylic alcohol, in which the oil is only moderately soluble, and carried out the experi- ments just as before, the methylic alcohol and piaene being shaken together during about 5 hours at 20'.The values of uD in 200 mm. tubes for the oils and the alcoholic solntions are given in the following table :No. 1 2 3 4 5 6 7 8 9 10 - THE CHARACTERISATION OF '' RACEMIC " LIQUIDS. 1126 Pinene and methylic nZcohoZ. aD of oil, 0 - 0'90" - 1.81 - 5.29 - 4-23 - 4.91 - 6.17 - 6'62 - 9'46 - 10'49 a,, of solution, Y. 0 - 0.16" - 0.33 - 0'63 - 0.78 - 0'82 - 1.07 - 1.1s - 1 *73 - 1.92 X - Y' - 5.62 5-48 5 *22 5.42 5 -99 5 -77 5-61 5 '47 5 *46 The same kind of result is obtained in this as in the previous series of experiments, and strong evidence is thus obtained that the ratio of the observed rotmatory powers of the oil and of the alcoholic solution in contact with it is constant.This result proves on analysis to be of considerable interest. Let us assume in the first place that externally compensated pinene is not racemic. Then in any two determinations with the same solvent, the oils contain, in unit volume, quantities which we may call xZ1 and xZ2 of Z-pinene and quantities xd, and xd, of d-pinene; similarly, unit volumes of the alcoholic solutions in equilibrium with them contain the quantities yl, and yZ2 of Z-pinene and pd, and yd, of the d-isomeride. Then, according to the distribution law, the equalities xz, X I 2 - xd, xd, & - yd, and - & - should hold, provided that the molecular weight of the pinene is the same in the two solvents. Further, since d- and Z-pinene are enantiomorphously related and the differences between the quantities of each present are small com- pared with the actual quantities, we should expect that xtz, xd, x7, xz, ?/d, ?/d, yl, yz2 d, - zd, E - -.- -- - whence xz, - xcz, $2 - 9 4 = const. - - But the differences between the quantities of d- and Z-pinene present in unit volume are directly proportional to the algebraic sum of the rotations due to the two active components, the length of the polarimeter tube remaining constant. Hence the values of x/p i s the1126 THE CHARACTERISATION OF " RACEMIC " LIQUID6. above t,able should be constant, as they are actually found to be within the limits of experimental error. It should be pointed ont that, since the concentrations xd and X Z in the oil are very high, the ratios xd/yld and xZ/yZ would not be expected t o maintain their constancy independent of the total concentration ; nevertheless, since the quantities of dextro- and Isvo-pinene used in any experiment differ by a t most 15 per cent.and the unknown causes affecting the ratios xdlp? and xZ/yZ are therefore operative t o about the same extent on each, the difference of the two ratios x Z ~ -xd, 12: -- - - - approximately preserves constancy. Yll - Yd, 9 Again, if we assume that externally compensated pinene is racemic in both of the liquid phases, the same result is arrived at. For, replace the quantities d throughout the mithemsticnl statement given above by the qmntities 9' representing racemic material; the coin- positions of the oils in any two experiments are then ml, xZ1, and ~ 7 ' ~ and xZp, whilst the compositions of the solutes in the corresponding alcoholic solutions are yl, 9Zl, and y~~ and yZ2, and from the dis- tribution law xz, xl, 99.1 v 2 94 Yl2 - xr1 I x 2 and - - - Since the components of the first equality are inactive, the latter does not affect the rotations, and the ratios x/p in the table are pro- portional to the xZ divided by the $, &c.A third case may of course be imagined, namely, that the externally compensated pinene is racemic in one solution and not in the other ; we should then have the eqnalities involving rooh of the component terms and the ratio x/y would not be independent of the total con- centration. This eventuality, however, is excluded by the results of the experiments. We have thus shown that the behaviour of mixtures of unequal quantities of the enantiomorphously related pinenes towards solvents is quite in accordance with the ordinary distribution law and indicates that an opticnlly active pinene exists in the same st.ate of molecular aggregation when dissolved in its optical isomeride as solvent as when dissolved with this isomeride in methylic alcohol; the results of these experiments as they stand are quite independent of any assumption as to the racemic or non-racemic nature of the externally compensated hydrocarbon. Obviously then, these results, in conjunction with those obtained with the camphorsulphonic chlorides, show the futility of attempting to decide between a racernic and a non-racemic liquid by We would point out in conclusion that our use of the term racemic the method employed for t.his purpose by Ladenburg.ASYMMETRIC OPTICALLY ACTIVE NITROGEK COMPOUNDS. 1127 as applied to liquids must not be construed as an admission of the existence of such compounds; the only evidence favouring the recog- nition of racemic liquids is Ladenburg’s statement that a change of temperature occurs on mixing dextro- and laevo-coniine. We tender our thanks to the Government Grant Committee of the Royal Society for funds defrqing the expenses incurred in this investigation. UNIVERSITY COLLEGE, KOTTINGH AM. GOLDSMITHS’ INSTITUTE, NEW Cnoss.