POSITION-ISOMERISM AND OPTICAL ACTIVITY. I583 CV.--Positioiz-isomerisrrL and Optical Activity ; the Comparative Rotatoy P o i v e u 9 f the Dibenxoyl and Ditoluyltnrtru t es. By PERCY FRANKLAND, Ph.D ., F.R.S., and FREDERICK MAr,coLw mTHARTON, A.1.c. WE hare recently shown (Trans., 1896, 1309) that the rotatory effect of the para-toluyl radicle is greater than that of the meta-, and this, again, greater than that of the ortho-toluyl grouping; and we have also pointed out that this relationship is in harmony with the relative position of the centre of gravity in these sereral groups ; for, assum- ing that the centre of gravity of the benzene ring is the geometrical centre of a regular hexagon, it is obvious that in the ortho-arrange- ment of the tolnyl group the centre of gravity is somewhat nearer,, in the meta-arrangement somewhat further, and, in the para-arrange- mcnt, still further than that geometrical centre from the carbon atom (the carbonyl carbon), by means of which the ring is attached to the asymmetric carbon atom of the tartaric acid.1584 FRANKLAND AND WHARTON : Thus, in Fig.1, with the CH3 group in the ortho-position relatively to the CO group, the centre of gravity of the toluene ring would be at g', with the CH3 group in the nieta-position i t would be at g", and, with the CH3 group i n the para-position it would be a$ gttr. Now, the moment ofthe mass of the toluene group around CH3(0) the carbonyl carbon atom will obviously be greatest when the centre of gravity i s at g"', and least when i t is at 9'. If, then, the op- tical activity is affected by the moment of this mass, the rotatory power of the para-toluyl compound should be the greatest, that of the ortho-toluyl compound the least, and that of the meta-toluyl compound intermediate between those of the other two.This rela- tionship was, indeed, shown to hold good in the case of the methyl aiid ethyl salts of ortho-, meta-, and para-ditolayltartaric acid. The question naturally arises as t o how the rotation of these tolujl compounds will be related to the correspondiiig benzoyl deri- vatives, I n the latter we have a smaller mass (by CH,) ; but this mass, which has its centre of gravity practically coincident with the geometrical centre of the hexagon, acts through a longer arm than in the case of the larger mass of the ortho-toluyl arrangement, and through a shorter arm than in either the meta- or para-toluyl arrangements.In order to institute this comparison, the rotatory power of methylic and ethylic dibenzoyl tartrates was required. Both of these com- pounds have already been prepared (Pictet, Guge aAnd Fayollat, Freundler ; Thesis, Paris, 1894), but in both cases their rotation had only been determined in solution, and the results were, therefore, of no value for the purpose we had i n view. We have, consequently, prepared both of these substances again, and have determined their rotation over a wide range of temperature, i n the liquid 01' superfused state, for comparison with the rotation of the methylic and ethylic ditoluyltartrates recently described by us.Preparation of Meth ylic Dibenzoy1tart~ate.-Fifteen grams of pure FIQ. 1. co I CH3(m)POSITION-ISOMERIShf AND OPTICAL ACTIVITY. 158 3 methylic tartrate (obtained by the hydrochloric acid method) were run from a dropping funnel into 38 grams of benzoic chloride heated to 1 4 5 O , the temperature being gradually raised to 180'. When the evolution of hydrogen chloride had ceased, the excess of benzoic chloride was distilled off under reduced pressure, the ethereal salt subsequently passing over at 275-282' (about 1'3 mm. pressure) ; 21 grams were obtainel which crystallised immediately on cooling. The substance was repeatedly recrystallised from methylated spirit until of constant, melting point; it is but very slightly soluble irt this solvent, and is obtained as long needles melting at 135.5'.On combustion, the following results were obtained. 0.2004 gave 0.4561 COz and 0.0855 H20. The density taken at 150' and at 160°, was C = 62.07 ; H = 4.74. C20H1808 requires C = 62.18 ; H = 4-66 per cent. ai500/4~ = 1 . 1 ' ~ ~ . d160°/4' = 1.1191, from which, by extra- and in tra-polation, were calculated d 100°/4" = 1.1755. d 137'/4' = 1.1407. d 183"/4' = 1.0975. and, at these latter temperatures, the rotation was determined with the following results : Eotat ion of Methy lic Di benzoy ltartrnte. (Length of polarimeter tube in each case was 44 mm.). Preparation of Ethylic Dibenzoy1tartrate.-The method pursued was the same as that described above for the methyl compound; after distilling off the excess of benzoic chloride, the ethereal salt passed over at 270-280' (about 10 mm.pressure). The distillate did not solidify, nor could crystallisation be induced by employing the most varied solvents ; it was, therefore, redistilled, 2nd washed in ethereal solution with sodium carbonate, after which, on standing for about * [FJD is the rotation constant which Guye ha3 introduced under the name of " niolecular deviation " ; it is calculated from the formula a = observed rotation; E = length of tube; M = molecular weight; d = density. Repeated reference has been made to this constant by one of us in previous papers.1586 FRANKLAND AND WHARTON : t,hree weeks in the vacuum desiccator, crystah began to appear, and then, on stirring, the whole cryst,allised. The solid was subsequently purified by repeated recry stallisation from methylated spirit until of constant melting point, 62.5'.03618 gave 0.8437 CO, and 0.1772 H20. C = 63.60 ; H = 5.44. 0.1958 ,, 0.4571 ,, ,, 0.0957 ,, C = 63.67; H = 3.43. C2,H2,08 requires C = 63.77 ; H = 5.31 per cent. The density was determined at 70', 90°, loo', and 136*5O, d 70°,/40 = 1,1537, d 9Oo/4O = 1,1368, d 100'/4' = 1.1280, d 136*5'/4O = 1.0970, from which by extra- and intra-polation were calculated : d 1*3'/4O = 1.2122. d 18'/4' = 1.1979. d 38'14' = 1.1809. d 4!a0/4' = 1.1758. d 53.5°/40 = 1.1677. d 60'/4' = 1.1622. d 77*5"/4' = 1,1472. d 109.5'/4' = 1.1199. d 18S~5°/40 = 10571. and at these temperatures the rotation was determined, with the following results : Rotation of Ethylic Dibeitzoylta?-trate. (Length of polarimeter tube in each case was 44 mm.). -30.06° - A t 1.3' [a,]= = - - -56.36'; [S], = -477.5, 0.44 x 1.2121 -31.29' - - -59.36 - 61.70 - At 18.0 ,, - At 38.0 ,, - At 44.0 ,, - At 535 ,, - 0.44 x 1.1979 0.44 x 1.18- - U.44 x 1.1758 0.44 x l.lm7 - 0.4h x 1.1622 - 0.44 x 1.1472 - - -3206' - -32.10' - - -G2*@5 - - -32*000 - -62.28 At 60.0 ,, - -31*850 - -63.28 - 62.15 - -31.37' - At 77.5 ,, - At 100.0 ,, - -60.77 - -30.16' - 0.M x 1.1280 - - -29.51" - 0.44 x 1.1199 - -27.38' - 0.44 x 1.09iO - At 109.5 ,, - - -59.89 - 56.i2 At 136.5 ,, - A t 182.5 ,, - -24.03' - - -51.G6 - 0-44 x 1.0571 - ,, - -4499.0, - ,, - -514.0.- ,, - -515.0. - ,, - -514.5. - 5130. - Y 2 - -- ,, - -507.0. - ,, - -490.5. - ), - -481.0. - ,, - -4500. ,, = -3399.5.POSITION-ISOMERISM AND OPTICAL ACTIVITY.1587 This is a very remarkable series of rotations, exhibiting, as it does, a phenomenon which, as far as me are aware, has not hitherto been observed, namely, the passage through a maximum in the chaltge of rotation, brought about by change of temperatwe, or, in other words, the occurrence of a change in the s i g n of the seizsitivelzess of the rotation to the temperature, Thus, from the above figures it mill be seen that the greatest observed rot ation was attained between the temperatures of 38" and 53*5", whilst the maximum specific rotation was reached and remained practically constant between 53.5' and 77.5' ; again, the maximum molecular deviation is found, and remains practically constant between 38" and 60'. It is noticesble that this maximum rotation occurs in the vicinity of the melting point of the substance, namely, 62.5'.Indeed, from the figures and curves given in our recent communication on the inethylic and ethylic ditoluyltartrates, and which are reproduced below, it will be seen that in those case3 in which we examined the rotation of these compounds in the super- fused state, the sensitiveness of the rotation to temperature had a tendency to change io the vicinity of and below the temperature of fusion ; in fact, in the cage of the methylic orthoditoluyltartrate there was some evidence of a maximum rotation, although not nearly so pronounced as in the case of this ethylic dibenzoyltartrate. Thinking that possibly the above results might be due to the ethylic dibenzoyltnrtrate possessiiig an abnormal density in a state of superfusion, and as the densities employed in the above calcula- tions for specific rotsation at temperatures below the melting point had all been arrived at by extrapolation, we subsequently made the following direct determinations of the density in the superf used state : By experiment.By extrapolat,ion. d42*0'/4' = 1.1751 1.1 775 d 535'/4' = 1.1658 1.1677 d 59*0'/4' = 1.1602 1.1630 These differences, however, produce so little eifect on the specific rotation that the recalculation of the specific rotations on the basis of these direct density determinations is quite unnecessary, and the attainment of the maximum rotation referred to above is not influ- enced thereby, thus : Using extrapo!ated Using direct density density.determination. [a],, at 44.0' = -62.05' - 62.1 5'" ,, ,, 53.5 = -62.28 -62.38 ,, ,, 60.0 = -62.28 -62.44 In the following diagram we have, f o r comparison, graphically represented the sensiti-ceness to temperature of the specific rotation1558 FRANKLAND AND WHARTON : will be seen tbat the specific rotation of the methylic dibenzoyltar- trate is markedy greater than that of the methylio orthoditoluyl- tartrate for the same temperatures, and similarly the ethylicPOSITION-ISOMERISM AND OPTICAL ACTIVITY. 158 9 dibenzoyltarbrate has a greater specific rotation than the ethyltc orthoditoluyltai-trate, excepting at those lower temperatures at which t'he rotation of the ethylic dibenzoyltartrate becomes abnormal in the sense indicated above. In order to ascertain whether the same relationship between the several compounds i n question is maintained when the rotation pro- duced by an equal number OF moleciiles of each is submitted to com- parison, we have also calculated the " molecular deviation " (see note, p.1585) for the six ditolnyltartrates. Thus, using the formula MoleculaT Deriation of Methylic Ortho-ditolu~ltartrate. = -dm - -668.0. -42.30' 414 - At 12.0" At 19.0 ,, ____ 414 - At 33.5 ,, .- --G53.5. At 54.5 ,, = -625.0. At 100.0 ,, -- - -34.470dz = -557.0. 0.4 4 At 100.0 ,, - - - -0.44-dm 34.55' 414 = -558.0. At 136.0 ,, - - ___- --30*13*d= - - -492.0. - -4495.0. 0.44 424 - - At 137.0 ,, - At 1830 ,, - Td& -24.89' = -412.0. Molecular Deriation of Ethylic Ortho-diioluyltartrate. VOL. LXIX. 5 P1590 FRANKLAND AND WHARTON : - 30.30" drnSrn 442 = -500.At 48-49.5' [a], = - 0-44 At 70.0 ,7 At 100.0 ), -26.7 7" - -405. - 24.01' Molecular Deviation of Slethglic ~~eta-ditolzL?lltartrnte. -39~62~ 414 - - 34.44" At loo'oo [ 6 ] ~ = - - = -564.0. At 136.0 ,, At 183.0 ,) - rd m2 - -475.0. Molecular Deztiation of Ethylic nleta-ditolu yltartrate. - 583.5. -~ - -28.68' 414 - _- -35.43O 448 - At 20.5" [ = ~4 m7 - At 24.5 ,) - 379.0. - -34.75" 3/ 442 - -575.5, At 44-5 ,, - ~ - 34.54" At 50.0 ,, - - o.44.- 0.44 2. FlFO - dgi = -573.0. 442 c -522.0, -- At 100.0 ,, - =d& -30.76' = -516.0. At 136.0 ,, - =dm -27.5f" 442 = -467.0. Xolecula~ Der id ion of Met h y 1 ic Para- dito I 169 I f (1 r t ra t e . - 51.57" At 100.0" [a], = -~ - 44.68O 414 - At 135-5 ,, - .~ = - i 3 1 .0.44 2 1.1095 At 153.0 ,, - - -mdFoB7 -36.16" 4 14 = -599.POSITION-ISOMERISM AND OPTICAL ACTIVITY. 1591 A. Methylic para-ditoluyltartrate. ,, meta- ), ,, ortho-ditoluyltartrate. 7, d i benaoy 1 tartrate. Ilfolecular Deviation of Ethylic Pam-ditoluyltart,rate. B. E thylic para-ditoluyl t str trate. ), meta- 9 9 ,) dibenzoyltartrate. ,, ortho-ditoluyltartrate. Xethylic Dibenzoyltartrate. At 100 '0' [MID = - 280 '1" At 137.0 ) ) = -258 '0 At 183.0 ,) = -227.5 Eth.yZic DibenzoyItarlmte. At 1'3" [bl], = -233.3" At 18.0 )) = -245'7 At 38.0 ,) = - 2 j j . 4 At 4 . 0 ) ) =- --256 '9 At 53.5 ), = -257'8 At 60.0 ), = -257'8 At 77.5 ) ) = -257.3 At 100'0 ), = -251 *G At 109.5 ,, = -247.9 At 136.5 ,) = -234.8 At 182.5 )) = -213'9 2fethyZic 3-Ditoluyltart?-ate. At 12-Oo[M]D = -32'3'2O At 33.5 ,) 2 -318'8 At 54.5 ), = -30'7 '3 Ati 70'0 ), -- --298 -2 At 19.0 ,, = -324'7 At100.O ,, = -2281.6 A t 100.0 ) ? = -282.3 At 136.0 ,) = -253.7 At 137'0 ), = -255.6 At 183'0 ) ) = -218.4 Ethylic o - DitoLuyEtart,rate.At At 30.0 ), = -266'7 At 48-49 '5 ,) = -263 *1 At 70.0 .) = -256.2 At 100'0 ,, = -241.9 At 135'0 ,, = -222.6 11 *O" [MID = -266.8" Methylic p-Bitoluyltartrate. At 100'0" [MID = -425 *7" At 135'0 )) = -378.9 At 183.0 ) ) = -318.4 Ethylic p-Ditoluyltartyate. At 100'0" [nix]D = -39'7.7" At 137'0 ,) = -3360.0 At 183'5 ), = --307'2 5 P S15 92 POSITION-ISOMERISM AND OPTICAL ACTIVITP. Methylic m- Ditolzcyltavtrate. I Ethylic m- DitoZuyZtartrate. At 100'0" [MID = -327.1' At 136.0 ,, = -292'2 At 183.0 ,, = -252.4 At 20'5' [MID = -306'4° At 24-5 ,, - -304.4 At 44.5 ,, = -305.7 At 50'0 , l = -305.0 At 39.5 , l = -284.6 At 100'0 ,, = -281.7 At 136'0 ,, = -259.5 - These figures show that the relationship between the molecular rotations of the dibenzoyl- and o-ditoluyltartrates is somewhat more complex than in the case of the specific rotations and the moleculai- deviations.Thus, at high temperatures, the molecular rotation of the methylic and ethylic dibenzoyltartrates is markedly greater than that of the corresponding o-ditoluyltartrates, whilst a t low tempera- tures the latter compounds have the higher rotation. This relation- ship is best observed by means of the accompanying diagram, from which it will be seen that above 1 0 8 O , the methylic dibenzoyltartrate has a greater molecular rotation than the met,hylic orthoditoluylfar- trate, and that above 62' the ethylic dibenzoyltartrate has a greater molecular rotation than the ethylic orthoditoluyltartrate ; now the methylic dibenzoyltartrate melts a t 135*5', and the cthylic dibenzoyl- tartrate at 62*5", so that it is in both cases at temperatures below their melting points that they exhibit a molecular rotation inferior to that of the corresponding ortho-ditoluyl compounds.But it is just below the melting point that we should expect to find, and in some of the cases referred to we have actually found, abnormal rotations occurring. These results furnish, therefore, an answer to the question raised at the beginning of this paper, showing, as they do, that the smaller mass present in the benzoyl compound, but with the more distant centre of gravity, and acting, therefore, through a longer arm, has, in this particular case, a greater rotatory effect than the larger mass present in the ortho-toluyl compound with its centre of gravity nearer to the carbonyl carbon atom, by means of which, in each of these componnds, the groups in question are attached to the asym- metric atom of carbon. The result is, moreover, the same, whether we measure t.he " rotatory effect " by specific rotation, by molecular deviation, or by molecular rotation, provided that the comparisons are only made at temperatures above the melting points of the com- pounds in question. Mason College, B iwniningh am.PIG. 3.-Influence of Temperature on the Molecular Rotation of Methylic and Ethylic Diberlzoyltartrates and Ortho-, Meta-, and Para-ditoluyltartrates. Temperature. X.B.-The diagram also indimtee the melting points in t,he eltees of those compounds known in the solid &ate ; t,liose only known 38 liquids are marked “ liquid.”