2630 WOOD AND COMLEY THE ROTATORY DISPERSION CCCLXI. -The Rotatory Dispersion of certain Normd A1 kyl Hexahydromandelates. By CHARLES EDMUND WOOD and MERVYN ARTHUR COMLEY. IN relation to the investigation of the rotatory dispersion of the n-alkyl lactates (J. 1923 123 600) the effect on rotation of sub-stitution of the cyclohexyl group for the methyl group of lactic acid was of interest for no alkyl ester series derived from an optically active hydroaromatic acid had so far been investigated. For this purpose hexahydromandelic acid was synthesised and resolved into its optical enantioinorphs by means of morphine and quinine. The optical rotation of the nine esters (methyl to n-nonyl) was taken over as wide a temperature range as possible in certain cases ZOO" this range being limited by vaporisation or racem-isation of the particular ester.Rotatory Dispersion. The dispersion curves do not show either maxima or minima or an apparent approach to maxima or minima. No intersections of the curves for a particular ester occur; further a continuous and distinct spreading of the curves takes place on passing from the red to the violet end of the spectrum that is to say the larger the rotation the steeper the tangential angle made with the axis of zero rotation a t any particular wave-length. The curves for the octyl Izevo-ester are shown in Fig. 1 an interest-ing feature being t'hat the curves for the highest temperature recorded cross the zero axis towards the red end of the spectrum. On comparing our results with the analysis of thc Drude equation by Hunter (this vol.p. 1198) cases III and IV which involve two terms of oppositc sign are applicable. At low temperatures Ic,> k when A >A ; the curves then lie completely in the rpgion of negativc rotation. With increasing tcmperature k, and k , vary k becoming equal to and finally less than lz,; the curve OF CERTAIN NORMAL ALKPL H~EXAHYDROiLJANn~LATES. 2631 then cross the axis into the region of positive rotation. The same conditions with reversal of sign hold for tlhe nonyl dextro-ester. In these esters the crossing of the axis has been realised in the visible region but we could not detect a maximum such as is to be expected in the red end of the spectrum owing probably to the limits of visual observation. The rotatory dispersion exhibited by all the esters is of a similar type normal aiid complex.The nearest approach to simplicity, if simplicity is realisa.ble (Hunter Zoc. c i t . ) is shown by the methyl FIG. 1. -1 -3 n 2 -5 - 7 - 9 4200 5000 6000 A in Lu. ester at - 12". As the series is ascended in general the numerical value of the rotation diminishes and t,he dispersion approaches the condition for anomaly. The dispersion of the Z-hexahydromandelates is closely allied to that of the lactates but a point of difference is that the curves for the former are convex on the negative side and for the latter convex on the positive side to the axis of zero rotation. In the first column of Table I the ratio on ascending the series tends sensibly to a constant value but this is not maintained in the other ratios given.Even at 30" where Dhe rotation values are more remote from the axis the ratio a6,08/x5461 is very variable 2632 WOOD AND COMLEY THE ROTATORY DISPERSION ,-Ester. A 4359. Ethyl ......... 2-00 Propyl ...... 2.04 Amy1 ......... 2.31 Methyl ...... 1-72 Butyl ......... 2-03 Hexyl ......... 2.17 Heptyl ...... 2.34 Octyl ......... 2.47 Nonyl ......... 2.30 TABLE I. Dispersion Ratios. At 30'. 5461. 1 1 1 1 1 1 1 1 1 - 6708. 0.500 0.576 0.518 0.523 0.465 0.392 0.501 0.372 0.442 At 140'. 7 4359. 5461. 1.80 1 1.98 1 2.44 1 2.24 1 2.58 1 3.09 1 2.69 1 3.46 1 2.50 L v 6708. 0.601 0.555 0.418 0.422 0-416 0-192 0.232 0.080 0.265 and much more is this the case a t the higher temperature.Less variable though larger in magnitude are the ratios a435g/aj4G1 at 140". The irregularities in the ratios are to be expected since the numerical values of the rotation are small for the curves approach and even cross the axis in the case of the octyl and nonyl esters. The ratios at these points pass through infinity to negative values. Thus the ratios a4359/a5461 for the octyl ester a t 190" and 200" are 7.42 and 19-20 respectively and for the nonyl ester 4.03 and 5.74; the ratios ~ ( ~ ~ ~ ~ / a ~ ~ ~ ~ for the octyl ester at 190" and 200" are - 10.5 and - 4.22 and for the nonyl ester 16.80 and 117.0. A further reason for irregularity is possibly an approach in form of the curves to the anomalous type.With such ratios the dispersion might be simple complex or anomalous. The esters which have the ratio cz4359/a5461 considerably higher than the minimum for simplicity (1.57) derived from the single-term Drude equation are not simple for the usual graphical test indicates complexity. This is confirmed in addition by the sensitiveness to temperature and the inconstant character of the dispersion ratios of the series. I n cases of obvious anomaly the dispersion ratios vary greatly when the dispersion curve crosses the axis of zero rotation which: up to the present for compounds in the homogeneous condition, has been observed in the violet region. With t,he octyl and nonyl members the dispersion curves exhibit a gradual change in shape on approaching and ultimately crossing the axis in the red end of the spectrum and the condition corresponds closely to certain curves given by Lowry (J.1915 107 1200). His diagram shows the gradual change of the " complex but normal " to the '' complex and anomalous '' type on crossing the zero axis from the positive to the negative region of rotation. I n the case of the two esters mentioned the curves are complex but do not show actual anomaly OF CERTAIN KORMAL ALKYL IIEXAHYDROMANDELATES. 28.33 Actual anomaly is to be expected but cannot probably be attained in this series owing to experimental difficulties. After reaching a maximum in the red end of the spectrum the curves would have ultimately the zero axis as asymptote; on increasing the tem-perature the axis would be crossed further to the left (octyl ester), and the maximum would move in the usual manner towards the violet.The high ratios in the violet and the low ratios in the red for the hexyl and octyl esters are brought out particularly with increase of temperature and show the combined effects of anomaly and low rotation value. In the Patterson modification of the Armstrong diagram for the series of hexahydrornandelic esters a t various temperatures the values obtained for particular wave-lengths (taking rotations for along the horizontal reference line) roughly lie on straight lines. These lines produced to the right intersect the reference line about the region + 7.5" (compare Patterson J. 1916 109, 1181) for tJhe lzvo-esters. The diagram shows clearly t'hat thc " rational dispcrsioii ratio " for any two wave-lengths is approxim-ately constant for the homologous series undcr the temperature conditions described that is change of constitution (by increasing the length of the carbon chain) and increase of temperature are similar in effect.The values of the molecular rotations for thc 2-lactates in a similar diagram lie practically on straight lines for particular wuve-lengths. The lines converge in this case towards the left intersecting the axis in the rieighbourhood of + 2.2". The angle made between the reference line and the line for any wave-length is slightly greater (especially for short wave-lengths) in the case of the hexahydro-mandelates than the corresponding angle in the case of the lactates. The tangents of the angles between the lines X 5461 and h 6708, 6152 4861 for the former are respectively 0-343 0-234 0.335 and for the latter are 0.349 0.223 and 0.300.I n both cases increasing temperature or lengthening of the carbon chain produces a movement along the lines towards the right except in certain lactates where maxima occur in the temperature-rotation curves. The diagram correlates phenomena in both ester series and shows that whilst the rotation is largely affected molecular dispersion is little changed by substitution of the methyl for the cyclohexyl group. Whilst only an approximation for the esters the diagram applies best to molecular and not to specific rotation. The molecular dispersion is shown thereby to be chiefly dependent on the groups immediately surrounding the asymmetric centlre temperature and lengthening of the chain producing only secondary effects 2634 WOOD AND COMLEY THE ROTATORY DISPERSSIC)~~ Eflect of Temperature.It is notable that in the case of the lactates increase in temperature causes an increase in rotation. I n both cases the temperature coefficient of change of rotation is greater the smaller the wavc-length of light used and it is decidedly greater for the hexahydro-mandelates than for the lactates. The teniperuture-rotation curves for light of different wave-lengths for the former series are lines of small curvature convergent witlEi increase of temperature (the lactates show a divergence) and inclined a t an angle to the axis of zero rotation. The general direction tends to become more parallel to this axis for the higher estms the curves of two of these, the octyl and nonyl cutting the axis.No maxima or minims occur in the ternperaturc range considcrcil. Iiicrease of temperature causes a decrease in rotation. TABLE 11. Molecular Rotatory Power." [ M ] r . Temp. 30". Alkyl Ester. A -6708. 6563. Ethyl ......... 17.4 18.2 Butyl ......... 10.2 10.8 Methyl ...... 25-7" 28.2' Propyl ...... 10.6 11.3 Amyl ......... 6.2 6-6 Hexyl ......... 5.7 6.2 Heptyl ...... 6-2 6.3 Octyl ......... 4.3 4.8 Nonyl ......... 5.0 6.5 5893. 5086. 4455. 4359. 36-8" 54.4" 73.7'' 77.6" 24-4 37.7 57.3 60.6 15.5 25.7 39.1 41.4 15.4 24.5 37.3 39-7 9.7 17-5 28-9 30.7 9.3 17.5 27.2 28.7 9.1 16.2 26-9 28.9 8.2 15.0 25.4 27.1 8.3 15.9 24.9 264 [M]',40".Temp. 140". Methyl ...... 19-6 20.3 26.3 40.7 56.1 58.9 Ethyl ......... 10-0 10.4 14.5 22-6 33.2 35.6 Propyl ...... 4.9 5.4 8-3 15.7 27.1 29.0 Butyl ......... 4.8 5.3 7.8 14.2 22-1 23.3 Amyl ......... 3.1 3-3 4.9 10.2 17.3 18-9 Hexyl ......... 0.9 1.3 2.9 7.7 14.2 15.3 Heptyl ...... 1.3 1.6 3.6 8.1 13.9 14-9 Octyl ......... 0.4 0.6 2.4 7.6 14.7 16.2 Nonyl ......... 1-4 1.6 3.2 7.6 12.2 13.1 * Numerical values only. The curves (Fig. 2) are obtained by interpolation of the molecular rotation-temperature curves for the wave-lengths in the above table. In general the rotation decreases on ascending the homo-logous series ; the curves gradually converge approach and tend to become parallel to the axis of zero rotation. This means that the molecular rotation approaches a constant value for those esters investigated above the arnyl ester.It is to be noticed that ther OF CERTAIN NORMAL ALKY-T~ TTRXAHY~ROMAN~ELAT~S. 2635 is a marked irregularity a t the butyl ester and a slight deviation in the form of the curve especially for light of short wave-length, a t the octyl ester which may be due to the ester approaching the anomalous condition. Comparison of our results with those for the esters derived from I-isopulcgol (Pickard Hunter Lewcock and Pennington, J. 1920 117 1248) shows that the elevation of molecular rotation occurs in the latter case a t the rL-valerate of isopulegol which corremonds in the chain to the n-butyl ester of hexahydromsndelic I acid. soo 70 GO t 50 '2 40 3 30 20 10 0 FIG.2, On ascending the n-alkyl series the negative rotation shown by the E-hexahydromandelic esters is reduced the positive rotation shown by the I-lactates is increased; the movement in rotation due to the lengthening alkyl chain is in the same direction in each case. Again in the higher members of these series as in many others investigated (Pickard and Kenyon J. 1914 105 830; Phillips J. 1923 123 23 etc.) advance from one member to the next produces little change in molecular rotation. This is to be expected if Rule's suggestion (this vol. p. 1121) that '' the influence of a substituent on the optical rotation of a parent compound is dependent in sign and magnitude on the change in electrostati 2636 WOOD AND COMLEY THE ROTATORY DISPERSION moment " is correct.The difference in electrostatic moment a t the asymmetric centre would become progressively smaller from one member to the next on lengthening the chain but the electro-static moment pari passu would increase (compare Thomson, Phil. Mag. 1923 [vi] 46,497) except when the helical chain returns upon itself; here a depression in molecular rotation (when the rotation increases up the series) or an elevation (when the rotation decreases up the series) is to be expected and not vice versa. I n accordance with this an elevation is shown by the l-hexahydro-mandelic series and depressions by the Z-lactat.es; in the former case the elevation occurs at the butyl ester in the latter depres-sions occur a t the amyl and octyl members.Instances of molecular rotation-molecular weight curves which exhibit increase of rotation on ascending the series and show depressions a t certain carbon atoms are numerous. The following may be quoted The esters of I-isopulegol (Zoc. cit.) the westers of d-benzylmethylcarbinol, d-p-octanol and d-y-nonanol (Pickard Ken yon and Hunter J. , 1923 123 l) esters of the type CB,*CH(OH)*R (Pickard and Kenyon J. 1914 105 S30) ethers of benzylmethylcarbinol (Phillips Zoc. cit.) the n-aliphatic ethers of d-p-octanol (Kenyon and McNicol J. 1023 123 14) and d-7-nonanol (Kenyon and Barnes this vol. p. 1395). An exception is shown by the n-secondary alcohols in the homo-geneous condition (Pickard and Kenyon J. 1913 103 1923). These show elevations of molecular rotatory power a t certain carbon atoms in a series in which the rotation increases on lengthening the chain.The effect of substitution of methyl for cyclohexyl in the hexa-hydromandelic series is to move the rotation-dispersion curves as a whole from the negative to the positive side of the axis of zero rotation. E x P E R I M E N T A L . Synthesis and Resolution of Hexahydromandelic Acid. Zelinsky (Ber. 1908 41 2677) described as the acid a compound which appears to have been the amide. The yield obtained by his method however was too poor for our requirements hence cyclohexyl bromide was changed directly to hexahydrobenzaldehyde, which gave through the cyanohydrin a satisfactory yield of the acid. Ethyl orthoformate was prepared by the authors' method ( J .Xoc; Chem. Ind. 1923 42 4 2 9 ~ ) . Esterification of cyclohexanol by the use of 48% hydrobromic acid (Scott J. 1900 77 648) having proved unsatisfactory (the crude products showed unsaturation and the best yield obtained was 570/,) cyclohexyl bromide was prepared by Kohler an OF CERTAIN NORMAL ALKYL HEXAHYDROMANDELATES. 2637 Burnley's modification ( J . Amer. Chem. Xoc. 1910 43 412) of Freundler and Damond's method (Compt. rend. 1905 141 593). With mechanical stirring 148 g. of cydohexanol (b. p. 75"/20 mm.) were allowed to fall (1 drop in 2 secs.) into 152 g. (15% excess) of phosphorus tribromide below 0". The product was left in the freezing mixture over-night kept for 2 days at the ordinary temperature and then poured on ice. After purification by Kohler and Burnley's method (loc.cit.) the bromide had b. p. 61-62"/20 mm. (yield 207 g.; 86%). It is important that no excess of cyclo-hexanol be used otherwise unsaturated products result. Efficient stirring during addition of the alcohol to the bromide raised the yield from 69 t o 86%. The yields of hexahydrobenzaldehyde obtained by the methods of Sabatier and Mailhe (Compt. r e d . 1904 139 343) Bouveault (Bull. S'oc. chirn. 1903 [iii] 29 1049) Zelinsky and Gutt (Ber., 1907 40 3051) arid Wallach ( A m a l e . ~ ~ 1906 347 316) are poor. Details of the method adopted by us are given ( J . Xoc. Chem. Ind., 1923 42 429T). The aldehyde is a colouriess oil b. p. 159.3". Bouveault (Zoc. cit.) gives b. p. 159". Preparation of I~exahydromandelonitrile and its Hydrolysis.-Hesahydrobenzaldehyde bisulphite compound (100 g .) was well shaken with a saturated solution of 42 g.of potassium cyanide, and the nitrile was extracted with small portions of ether the product being at the same time gradually diluted with water. The nitrile was heated under reflux for 1 hour with concentrated hydro-chloric acid (d 1.16 ; 4 vols.). The nitrile remaining unchanged was separated while hot and again hydrolysed for 1 hour with fresh acid. This process was repeated but complete hydrolysis was never attained owing to the formation of isonitrile. The nitrile should be hydrolysed immediately after its formation. Yield of acid 34-5 g. (47%); m. p. 134.7" (corr.). The acid thus obtained was very soluble in hot water or ether, did not crystallise well from acetone was moderately soluble in alcohol or benzene and contained no water of crystallisation.The acid obtained by Zelinsky had m. p. about 16G" was readily soluble in hot water from which it was obtained in silvery scales difficultly soluble in ether and crystallised well from acetone [Found by authors C = 60.85 ; H = 8-53. Zelinsky found C = 60.64 ; H = 9-10. C,H,,O requires C = 60.72 ; H = 8.91%. Equiv. wt. = 157-6 (by silver salt) 157.8 (by titration). HexahyclromandeEamide was isolated in quantity after a certain hydrolysis of hexahydrobenzaldehydecyanohydrin with concen-trated hydrochloric acid. The product after treatment with dilutc caustic potash in the cold gave m. p. 164" crystallised from benzene Calc. 158*1] 26338 WOOD AND COMLRY THE ROTATORY DISPERSION in a different manner from the acid and when recrystallised from water had m.p. 165.2" (con.). It was readily soluble in hot water or alcohol (from the former it was obtained in scales on cooling), difficultly soluble in ether or ligroin and might be recrystallised from acetone (Found N = 9.11. C,H1,O,N requires N = 8.92%). The above properties correspond to those given by Zelinsky for the acid. If the ethereal solution of the cyanohydrin be hydrolysed the hydrolysis to the acid is not complete and a quantity of hexahydro-rnandelamide results. Zelinsky did not state that the nitrile was isolated from ethereal solution before hydrolysis. The best con-ditions for hydrolysis to the acid are isolation of the cyanohydrin as rapidly as possible and hydrolysis for the time stated otherwise a mixed product results.Resolution of Hexahydromandelic Acid.-Comple te resolution of the acid into the optically active enantiomorphs was effected by the use of quinine and morphine. A good criterion of optical purity was given by the fairly high value of the rotation of the methyl ester derived from both forms. Purther evidence of completeness of resolution was obtained from the rotation of esters other than the methyl of both forms of the acid and of the dextro-and lzevo-ammonium salts. The outline of the methods of resolution adopted is similar to that for the resolution of lactic acid by morphine (Wood Such and Scarf Zoc. cit.). Hexahydromandelic acid (20 g.) exactly neutralised with N-potassium hydroxide was added to 20 g.of the acid in 3000 C.C. of hot water containing 41 g. of quinine. The quantity of quinine added was sufficient to give the basic salt; the normal salt crystallises with difficulty. The whole was brought gradually to the boiling point; a quantity of the dark-coloured syrupy quinine salt remained undissolved which became solid on cooling. The solution was left over-night in ice, the mass of long colourless needles separated the liquid again brought to the boiling point in contact with the quinine salt and a further crop of crystals obtained. This process was repeated eight times. Crystallisation took place below 45" and was assisted by inoculation. The quinine salt was recrystallised twice from 4000 C.C. and 3000 c.c. respectively of water.To the hot solution was added excess of dilute ammonia the precipitated quinine separated and the solution reduced to 250 C.C. and rendered just alkaline with ammonia. Traces of quinine were separated and the acid was obtained from the ammonium d-hexahydromandelate by acidification with dilute hydrochloric acid filtration and ether-extraction. The acid was purified by redissolving in warm water, Isolation of optically pure dextro acid OF CERTAIN NORMAL ALKYL HEXAHYDROMANDELATES. 2639 leaving until cold extracting the supersaturated solut'ion with ether and crystallising from ether The resolution gave 58% of the theoretical quantity of d-acid. Recovery of the partly resolved hexahydromandelic acid from the mother-liquors was effected by addition of excess of dilute ammonia filtration reduction in volume to 300 c.c.further addition of ammonia until just alkaline and further procedure as detailed above. Resolution by means of quinine proved more difficult than by the use of morphine. d-Hexahydromandelic acid crystallises from ether in prisms which appear to be monoclinic m. p. 129-7" (corr.); [a];:" = + 13.51" (K = + 1.68' for 0.8570 g. of acid in 6-9 C.C. of solution in absolute alcohol). Quinine d-hexahydromandelate crystallises from water in long needles m. p. 45"; [ax = - 123.24" (a = - 5.21" for 0.4366 g. of salt in 10.3 C.C. of solution in absolute alcohol). Ammonium d-hexahydromandelate is of little value as a criterion of the optical purity of the acid owing to its small rotatory power : [a];? = - 7-78" (a = - 0.56" for 0-5 g.of acid and 2.0 C.C. of ammonium hydroxide d 0-88 made up to 7.7 C.C. of solution with distilled water). The rotations of the sodium and potassium salts also are of opposite sign t'o t,hat of the acid. Isolation of optically pure lcrvo-acid by means of morphine was carried out as follows Half bhc available acid (27 g.) was neutral-ised by ,%'-potassium hydroxide solution and the other half by tlhe equivalent quantity (51.5 g.) of morphine in 900 C.C. of boiling water. The solutions were mixed cooled and crystals of morphine 1-hexahydromandelate were precipitated. On evaporation to half volume an additional crop was obtained. The morphine salt was recrystallised twice from water the rotation then being constant. To the boiling solution (900 c.c.) of t'he salt ammonia was added, and the morphine filtered off.The solution kept just alkaline with ammonia was reduced in volume to 250 C.C. and left over-night. The small quantity of precipitated morphine was separated, and the I-hexahydromandelic acid isolated by acidification of the solution of its ammonium salt in a similar manner to that for the dextro-acid. The resolution gave 85% of the theoretical quantity of Z-acid. The mother-liquors were worked up to obtain partly resolved d-acid for rcsolution with quinine. Z-Hexahydromandelic acid crystallises from ether in hexagonal clusters m. p. 129.7" (identical with that of the dextro-enantio-morph) ; [a]F = - 13-62' ( m = - 1-04' for 0.5047 g. of the acid in 6.6 C.C. of solution ill absolute alcohol) 2640 WOOD AND COMLEY THE ROTATORY DISPERSION Morphine 1-hexahydromandelate crystallises from water in short needles m.p. 127-6" (decomp.); [a]"; = - 63-73" (a = - 3-51' for 0.6392 g. in 11.6 G.C. of solution in absolute alcohol). The rotation of the salt was constant after the second recrystallisation. Ammonium I-hexahydromandelate gave [u]";" = + 7-51' (a = + 0.54"; 0.50 g. of acid 2 C.C. ammonium hydroxide d 0.88 made up to 7.7 C.C. of solution wit'h water). Both alkaloids gave as the less soluble salts in water and acetone those derived from the dextro-acid. The partly resolved strychnine salt is precipitated from water as a syrup. The brucine salt is similar but it may with difficulty be crystallised from acetone as extremely deliquescent feathery needles.After three recrystallisations from acetone the acid derived from this salt was not completely resolved, Methyl 1-Hexahydromande1ate.-In general esterification without racemisatlion can be effected by both the sulphuric and hydrochloric acid methods but the higher-boiling fraction of esters produced by the latter method frequently showed an elevation of rotation which probably indicated lactide formation. Kahlbaum's methyl alcohol (4 mols. ; 6.2 g. dried with anhydrous sodium sulphate) was mixed with 1.5 g. of concentrated sulphuric acid cooled and 7.5 g. of I-hesahydromandelic acid were added; the mixture was heated on the water-bath at 65-70' for 29 hours, and allowed to cool. The excess of sulphuric acid was removed by dilution with dry ether (4 vols.) addition of potassium carbonate, filtration from the salts and removal of the ether.After four distillations in a vacuum the ester had b. p. 108.7'/ 5 mm. (oil-bath temperature 143") and m. p. + 4.7". Yield 6-5 g (80%). A fifth distillation caused no change in rotation. Since the specific rota-tion of this ester was much higher than that of the m-propyl ester, two further specimens of the acid dextro- and levo-rotatory respectively were esterified by the sulphuric and hydrochloric acid methods. (1) Methyl d-hexalhydroinandelate was prepared by adding 5.0 g. of the d-acid to a mixture of 4-1 g. of methyl alcohol and 1.5 g . of concentrated sulphuric acid. The mixture was heated at 65-70' for 3 hours. The ester isolated as in the above prepar-ation had b.p. 110~4-110~9"/7 mm. (oil-bath temperature 150") and ay + 23.11'. (2) Methyl I-hexahydromandelate was prepared by passing 0.258 g. (6%) of dry gaseous hydrogen chloride into a mixture of 5 g. of I-hexahydromandelic acid and 4.1 g. of methyl alcohol, and heating a t 65' for Z h hours. The ester treated as in the above Resolution by strychnine and brucine. The ester gave rotation a$' = - 23.00" OF CERTAIN NORMAL ALKYL KEXAHYDROMANDELATES. 2641 preparations had b. p. 116.5'/12 mm. (oil-bath temperature 138"). After four distillations it gave rotation a:' = - 23.01O. The ester exhibits complex rotatory dispersion. Densities d determined 1.0992 a t - 14.6"; 1.0698 a t + 17.6" 1.0313 a t 60.3"; 1.0081 at 85.9"; 0.9686 a t 128.7'. 4" 4" di!5c d:F' d:!8'50 di?' dl 9 '6' $"" 1.0970.1.0683. 1.0296. 0.9906. 0.9743. 0.9596. A. 6708 ...... 6563 ...... 6152 ...... 5893 ...... 5461 ...... 5324 ...... 5086 ...... 4861 ...... 4455 ...... 4359 ...... [ u ] y . - 17.10° 17.60 20.59 23.04 28.49 30.33 34.32 38-31 46-38 48-49 -15*80° -16-70 19-69 21.95 26-69 28.65 32.29 36.18 43.79 45.95 [ u]?. ' 14.72" 15.38 18.03 19-94 24.32 25-89 29.26 32.97 40.07 42.26 [cz]lhO"*. - 13.08" 13.49 15.88 17.62 21-49 23.05 26.29 29.59 36.20 38.06 [ u y . - 12-17' 12.71 14.80 16.45 20.17 21-79 24-88 28.15 34.39 36.1 1 10.93 12.67 14.07 17.74 19.27 22.24 25-20 30.56 32.1 8 &12" ,. - - X 1.0970; = X 1.0683; a? = [a]?' x 1.0296; 10' c czy = [.IF X 0-9906; = x 0.9743; =s x 0-960.[MI = [cz]; x 1.72. Ethyl 1-Hexahydromande1ute.-The ester was prepared by heating 7-0 g. of Z-hexahydromandelic acid with a mixture of ethyl alcohol (8.2 g . ) and concentrated sulphuric acid (2.17 g.) for 3 hours at 60-65'. The product was treated in the same manner as the methyl ester. After three distillations in a vacuum the ester formed a mass of needles m. p. 43-1' (corr.) b. p. 133"/20 min. Yield 6.4 g. (78%). The ester exhibits complex rotatory dispersion. 0.9677 a t 94.6"; 0.9293 a t 138.9"; 0.9048 a t 167.3". Densities d$ determined 1.0160 a t 39-3"; 0.9868 a t 73.1" d Y O d:?' d F U d26.5" &1'1" [US? 1.4268. 0-&52. 0.9724. 0.9426. 0.9176. [ u]lg '".[ u]:~,"". [ cz JF"". I 5.590 I 5.200 A. laly. 6708 . . . . . . . . . - 9.52" - 7.87" - 6-69' 6563 ......... 10.00 8.2 1 7.03 5-85 5.42 6152 ......... 5893 . . . . . . . . . 5461 .... .... . 5324 . . . . . . . . . 5086 . . . . . . . . . 4861 .... .. . .. 4455 4359 . . . . . .. . . . . . .. . . 11.71 13.34 16-40 17.97 20.64 23.90 31.32 33.31 0.77 11.08 13.79 15-10 17-42 19.93 25-98 27.97 8.54 9-84 12.20 13.24 15.45 17-51 22.73 24.51 7.35 8.43 10.43 11-32 13.19 14-93 19.24 20-71 6-55 7.40 9-14 9-95 11.53 12-98 16.80 18.01 n-Propyl Hexahydromande1ate.-Since the methyl ester showed unusual propertics in the lactate series (Wood Such and Scarf 2642 WOOD AND CONLEY THE ROTATORY DISPEESION Zoc.cit.) methods of esterification were investigated in the prepar-ation of the propyl ester. Three specimens of the ester prepared as below were examined, (1) The ester prepared from the E-acid by the hydrogen chloride method (described under the methyl ester) had rotation cxg = (2) A specimen of the d-ester prepared by the use of 12.9% hydrogen chloride under similar conditions to those of the last experiment had rotation ag" = + 13.06" after the first fractional di~tillat~ion ; this probably indicated lactide or anhydride formation. After five distillations the ester had b. p. 129.5-130-0"/5 mm. (oil-bath temperature 162") and gave constant rotation a:" = + 8.40". (3) To a cold mixture of 7.8 g. 01 propyl alcohol and 1-5 g. (sufficient to form dihydrated sulphuric acid with the water pro-duced) of concentrated sulphuric acid 5-07 g.of Z-hexahydro-mandelic acid were added. The solution was heated for 3 hours at 60" cooled and subsequently treated as for the methyl ester. Yield 4.8 g. (73%). After four vacuum distillations the ester had b. p. 130-0"/5 mm. (oil-bath temp. 155") rn. p. 4.6" and gave rotation = - 8.35". - 8.31". The ester exhibits complex rotatory dispersion. Densities d$ determined 1.0334 at 0.0"; 1.0038 at 34.9"; 0.9776 at 64.6"; 0-9433 at 103.7"; 0.9257 at 124.5"; 0.9015 at 152.2". d:;" &" &4'2 1 d;?' d:!G'4 c p s " 1.0312. 1*0170. 0-9918. 0.9739. 0-9500. 0.9065. A. 6708 ...... 6563 ...... 6152 ...... 5893 ...... 5461 ...... 6324 ...... 5086 ......4861 ...... 4455 ...... 4359 ...... [a]?. - 6.12' 6.53 7.77 8.87 11-69 12-53 14.55 16.87 21.77 23-29 [o.]ph9'". 5-91 5-09 7.19 6.18 8.22 7.09 10.63 9-37 11.57 10.16 13-24 11-90 15-67 13.75 20.42 18-22 21-60 19.19 -5.58' -4.79" [a]?'". -4.12' 4.36 5.43 6.27 8.3 9 9-13 10.70 12.61 16-91 17.89 [ a1y. -3.46' 3.67 4.60 5.29 7.19 7-80 9.25 11-05 15-29 16-28 [ a l y p . -2.36' 2.54 3.33 4.03 5.75 6-23 7.71 9.48 13.30 14-26 n-But$ l-Hexahydromande1ate.-The ester was prepared by the sulphuric acid method using 4 rnols. of n-butyl alcohol and sufficient sulphuric acid to form the dihydrated acid with the water formed. The purified ester was a colourless oil b. p. 147"/11 mm.(oil-bath tcmpuraturo 176") and m. p. - 0.7". The solution was warmed at 70-75" for 3$ hours OF CERTAIN NORMAL ALKYL HEXAHYDROMANDELATES. 2643 Densities d$ determined 1.0213 at - 8.0"; 1.0118 at + 4.3"; 0.9841 a t 38.3"; 0.9357 a t 97.4"; 0.8956 a t 147.0". d'529 & 5" &;' ' d?" &'$' 4" 1.0203. 1.0038. 0.9742. 0.9459. 0.8915. A. [a]?. [a~?'''. [ a ] ~ ' " . [ u]?. 6708 . . . . . . . . . . . . - 6.04' - 5.27' -4.06" - 3 ~ 2 4 ~ -2.10' 6563 .. . . . . . . .. . . 6152 ..... . . . . . . . 5893 .. .... .... . . 5461 ... . . . . .. . . 5324 . . . . . . . . . . . . 5086 . . . . . . . . . . . . 4861 ..... . . . . . . 4455 * ........... 4359 . . . . . . . . . . . . 6.42 7.90 8.69 11.04 11.94 13.94 15.93 20.53 21-83 6-66 6-82 7.94 9.88 10.76 12.47 14.51 18.74 19-91 4.3 7 5-38 6.31 7.99 8.74 10.62 12.09 15-93 16.93 3.49 4-34 5-08 6-62 7.27 8.48 10-04 13.28 14.3 1 2-25 2-82 3.36 4.72 5.31 6.27 7.4 8 9-78 10.36 a:'" = x 1.0203; a? = [air x 1.0038; a:7 = [a]?' X 0.9742; ah 83 :1" = [ a ~ i 5 ' x 0.9459; a:" = [ a ] y L x 0.8915.[MI = [u] X 2-14. n-Amyl d-hexahydromandelate was prepared under similar con-ditions to those employed for the butyl ester from Kahlbaum's n-amyl alcohol and had b. p. 164.5"/12 mm. (oil-bath temperature 204"); it crystallised on cooling in small clusters of prisms m. p. - 2.5". After a fourth distillation the rotation was unchanged. Complex rotatory dispersion is exhibited by this ester, Densities d$ determined 1.0135 at - 8.0"; 0.9926 at + 16.6"; 0.9640 at 51.3"; 0.9133 a t 114.6"; 0.8745 a t 159.1".&! 3' d:3' &"' dll d:26 5o 1.0111. 0.9873. 0.9508. 0.9088. 0.8767. A. Carn.* [ ayy. [a] lJ. G708 .......... ,. +3.14" +2*80° +2*24O +1-60° +1*15O 6563 ............ 3.38 2.99 2-37 1.66 1.21 6152 ...... . .. . . . 4.25 3-69 2.99 2.16 1.53 5893 ............ 5-01 4.43 3.51 2-54 1.89 5461 ..... ....... 6-73 6.01 4-86 3.65 2.90 5324 ............ 7.36 6-61 5-48 4.14 3-32 5086 ............ 8-83 7-94 6.51 5.02 4.07 4861 .... ........ 10-58 9-54 7-78 6-05 5.08 4456 ............ 14.85 13-06 10.67 8.43 7.00 4359 ............ 16.07 14.01 11.46 9.13 7.64 al'y" = ra]y x 0.9088; a;b*j = E a ] 7 5 X 0.8767. a:*$o = [a]A530 x 1.011; = x 0.9873; =-= x 0.951; [M] = [a] x 2-28.n-Hexyl 1-hexahydromandelate was prepared by the sulphuric acid method from n-hexyl alcohol b. p. 157-0"/760 mm. ; the latter was prepared from pure n-amyl alcohol by the cyanide synthesis (J. 1923 123 600). The ester boiled a t 1EO.5"/7 mm. (oil-bath temperature 190") and melted sharply at - 1.5". Complex rotatory dispersion is exhibited. Densities d$ determined 1.0091 a t - 12.0"; 0.9861 a t + 14.3": 0.51518 at 57.2" ; 0.0136 a t 101.3" ; 0.8893 at 133.0" ; 0.8704 at 154% 2644 WOOD AND COMLEY THE ROTATORY DISPERSION ds'50 d?''? d:!'5c d:!l'Go &7'20 1.0045. 0,9816. 0.9278. 0-8979. 0.8768. A. [u1i6'B0 [ u 3 ~ ' Y . [ u]y'"". [a1y'60. [ ,]:?"2'. 6708 ......... ... -3.05O -2.59' -1.29" -0*71° -0.30° 6563 ............ 3.26 2-79 1-47 0.85 0.46 6152 .. . . . ... . . . . 4.03 5893 . . . ....... .. 4.69 5461 . .. .... . . .. . '6.44 5324 ............ 7.08 5086 ....... ..... 8-32 4861 ............ 9-64 4455 ............ 12.67 4359 ............ 13.16 3-51 4.15 5.72 6.33 7-48 8.69 11.61 12-13 2-00 2.52 3.86 4.37 5.40 6.48 8.77 9-33 1-29 1-67 2-45 2.89 3.92 4.87 6-92 7.43 0.80 1.01 1.78 2.12 2.83 3.58 5.42 5.83 ah -6.5" = [ u ] ~ ' " X 1.0045; u:.'~ = [u]:" x 0.9816; u:'~' = [ u ] ~ " ' x 0.9278; a,\ 121'60 = x 0.8979; u y ' " = [u]:"~' x 0.8768. [MI = [a] x 2.42. n-Heptyl d-hexahydromandelute was prepared by the sulphuric acid method from n-heptyl alcohol b. p. 174.6-174.9"/745 mm. The conditions were as described under the butyl ester.The ester had b. p. 182.5"/12 mm. (oil-bath temperature 215") and m. p. - 7.1". Densities d: determined 0.9952 a t - 10.5" ; 0.9780 a t + 12.5" ; 0.9351 a t 66.8"; 0.5966 a t 115.2"; 04563 a t 165.5". It exhibited complex rotatory dispersion. &5* 42 lo 40 d:i0 3* d:22'5* $5'89 0.9867. 0.9697. 0.9360. 0.9004. 0.8586. A. [ u]""'. [ uJy'l* [ u]y 8.. [ ,]?'%*. [ a]y'5°a 6708 .... ........ +3-04' +2.58' fl.67O +O-9So +0-19O 6563 ............ 3-19 2-72 1.80 1.1 1 0.30 6152 ............ 3.74 3.20 2.29 1-55 0.79 6893 ............ 4-33 3.74 2-71 1.98 1.07 5461 ... .. ..... . . 5-69 5.02 3.78 2.82 1.75 5324 . ........... 6.26 5.54 4.22 3.22 2.00 6086 .. . .... .. . .. 7-34 6.61 5-16 3-95 2.61 4861 ............ 8-75 7-81. 6.23 4.79 3.19 4455 ............11-92 10.85 8.7 1 6.75 4.53 4359 ........... . 12.89 11.68 9.50 7.32 4.84 u1'5* = [u]pjO x 0.9867; = [ u ] ~ . ' O x 0.9697; = [uz'8* x 0.9360; ay0'30 = [,~,]f:""' x 0.9004; c ~ f ~ ~ ~ = x 0.8586. [MI = [a] x 2.56. The est'er was found to be partly racemised after heating a t 183" for 2 hours. n-Octyl 1-hexahydromandelate was prepared under similar con-ditions to those employed for the heptyl ester (except that the time of heating was 44 hours) from n-octyl alcohol b. p. 195*5-195.8"/ 758 mm. After three fractional distillations the ester had b. p. 184"/7 mm. (oil-bath temperature 220") and m. p. + 7.8". Com-plex rotatory dispersion is exhibited and the rotation-dispersion curves cross the axis of zero rotation at high temperatures and for light of long wave-length.Densities-d: determined 09361 at - 9.6"; 0-9738 at + 7.0" OF CERTAIN NORMAL ALKYL €IESA€rYDROl\.IA~DELATES. 2645 0.9421 a t 48.9"; 0.8933 a t 111.8"; 0.8684 a t 166.9"; 0.8283 at 197.2"; 0.8240 a t 203.1". d193'3. dt? d?"" d:f''5' 9 ~ 7 ' 6 . 4' 0.9845. 0.9612. 0.9130. 0-8668. 0.8261. A. 6708 . . . . . . .. . . . . 6563 .. . ....... .. 6152 . . .. . .. ... . . 5893 . ........... 5461 . . . . . . . . . . . . 5324 . . . . . . . . . . . . 5086 . . . . . . . . . . . . 4861 . . ... .. .. . . . 4455 4359 . . . .. . . . . . . . . . .. . . . . * . . [ u p . - 1.98' 2.25 3.01 3.63 4.80 5-29 6-37 7.47 10.34 11-05 [ U l y . - 1-62' 1.89 2.56 3-08 4.19 4.61 5.75 6.77 9-54 10.22 [ u 1 y .-0*83' 1-05 1.57 1.97 2.98 3.35 4.28 5.22 7.74 8.34 [ u ] y 5 " . - 0 . 0 2 O 0.11 0.51 0.7 1 1.57 1-92 2.61 3-42 5.09 5.62 0.89 0.73 0.53 -0.15 0.41 0.80 1.23 1.89 2-35 C . 0 ai'6 = [a]:'6' x 0.9845; a:' == [a]?' x 0.9612; a~"" = [u]:'" X 0.9130; 147'5' = ah [a]y5" x 0-8668; a ~ " ' = x 0.8261. [ill]; = [u] x 2.7. n- NonyE d-hexahydromandelate was prepared by the sulphuric acid method from n-nonyl alcohol b. p. 213.1-213.4"/757 mm., obtained by the cyanide synthesis from n-octyl alcohol (see under n-hexyl Z-hexahydromandelate). The time of heating at 65" was 5 hours. The ester had b. p. 193-3-194.5"/16 mm. (oil-bath temperature 234") and m. p. + 5.4"; it crystallised in short, colourless prisms.Complex rotatory dispersion is exhibited and, similar to those for the octyl ester the rotation-dispersion curves cross the axis of zero rotation a t high temperatures for light of long wave-length. Densities d$ determined 0.9757 at - 5.5"; 0.9600 a t + 14.7"; 0.9087 a t 85.0" ; 0.8791 a t 121.3" ; 0.8446 a t 167.5" ; 0,8192 at 201.5". A. 6708 ... . . . . . . . . . 6563 . . . . . . . . . . . . 6152 ...... . . . . . . 5893 .. .. . .. . . . .. 5461 . . .. . . . . . . . . 6324 . . . . . . . . . . . . 5086 ..... . .. . . . . 4861 .. .. . . . . . . . . 4455 4359 . * . . . . . . . . . . .. . . * . . . . . . dil.4' [ a 3 y . 0.9742. 3-2-03' 2.26 2.83 3.35 4.71 5.35 6-56 7-79 10.39 11-09 d'?.5* 0.9543. [ a]?'". + 1-84O 1.99 2.60 3-04 4.19 4-74 5.78 6.93 9-06 9.48 d$'Bo 0.9210.[ a]y'6°. + 1.43' 1-54 1-98 2.35 3-24 3.63 4-55 5.40 7.31 7.76 d Y 4 * [ a ] y 4 0 . 0.87 15. + 0.59' 0.69 0.96 1.26 2-08 2-30 2.87 3.52 4.61 4.95 d:c1'8' 0.8192. 0.37 -0.14 + 0.02 0-42 0.61 1.00 1-39 2.20 2.42 q 4 ' = [a]:'40 x 0.9742; &F'5* = x 0.9543; aG6'60 = [ ~ ] f ; ' ~ X 0.9210; a132'4° I - [u]?'~' x 0.8715; = x 0.8192. [MI = [a] x 2.84. The rotations of the esters did not change on keeping; from this i t was inferred that the tendency for lactide formation is small. Of the nine esters examined (methyl to n-nonyl) the five esters, methyl to n-amyl have an apple-like odour. They are liquids a 2646 CLASSTONE STUDIES OF ELECTEOLYTIC POLARISATION. the ordinary temperature except the ethyl ester (m. p. 43-1") and melt sharply within a few degrees of zero. While the esters have the same sign of rotation as the acid from which they are derived the sodium potassium and ammonium salts are of opposite sign. Summary. 1. Hexahydromandelic acid has been synthesised from cyclo-hexanol by means of the Grignard reagent and resolved by the use of morphine and quinine. 2. I n the homogeneous condition the n-alkyl hexahydromandel-ates (methyl to n-nonyl inclusive) exhibit complex rotatory dis-persion of a similar type. The rotation-clispersion curves for the octyl and nonyl members cross the axis of zero rotation inferring an approach to anomaly. 3. The effect of substitution of the cydohexyl group for the methyl group of lactic acid is that the esters generally retain the sign of rotation of the acid from which they are derived the molecular dispersion is slightly greater than for the lactates and the rotatory power decreases with rise of temperature. Temperature has a greater effect on the rotation of the hexahydromandelates than on that of the lactates. 4. The molecular rotation appears to decrease to an approxim-ately constant value on ascending the series. An elevation of rotation occurs a t the n-butyl and a deviation a t the n-octyl ester, The latter may be due to anomaly. 5. Dispersion is due to the groups immediately surrounding the asymmetric centre temperature and lengthening of the carbon chain producing secondary effects. The progress of this research has been facilitated by a grant to one of us (M.A.C.) from the Department of Scientific and Industrial Research to whom we desire to express our thanks. UNIVERSITY OF BIRMTNGHAN, EDGBASTON. [Received July 22nd 1924.