I~ACNMIZATION OF 2,2’-DIMETHOXY-G,6’-DICARBOXYDIPHENYL
Sept., 1959
1461
SOLVENT EFFECTS IN THE RACEMIZATION OF 2$’-DIMETHOXY-6,6’DICARBOXYDIPHENYL AND ITS DERIVATIVES1” BY B. M. GRAY BILL'^ AND J. E. L E F F L E R ~ ~ Contribution j r o m the Department of Chemistry, Florida State Univemity, Tallahassee, Florida Received February 81, 1969
The racemization of 2,2‘-dimethox diphenyl with carboxyl, ester or amide groups in the 6,6’-positions is subject to medium effects both on the reaction rates a d o n the activation parameters. The effect OF a change in solvent on the rate constants can be as much as a factor of five, the effect on the enthalpy of activation as much as 9 kcal./mole, and the effect on the entropy of activation as much as 21 cal./mole-degree.
The racemization of the sodium salt of 2,2’dimethoxy-6,6‘-dicarboxydiphenylhas been found to be decelerated by salts or organic solvent components added to the aqueous solution.2 The present paper is an extension of that work to derivatives of the biphenyl in which the barrier groups and the solvent are non-ionic but possess varying degrees of ability as donors or acceptors of hydrogen bonds. The acid, the dimethyl ester and the diamide (I111) have been racemized in an extensive series of solvents a t two temperatures. 0
decreasing rates is methanol>ethanol>trifluoroethanol > ethylene glycol > benzyl alcohol > isopropyl alcohol>t-butyl alcohol>>N,N-dimethylethanolamine. The latter solvent is unique in having a negative sign for the rotation of the dacid. The rate in that solvent is somewhat slower even than that of the mono-sodium salt in water?; the activation energy is a little lower, but the activation entropy is very much more negative. The activation parameters for the racemization of the acid in N,N-dimethylethanolamine also deviate widely from the roughly linear relationship between the activation parameters in the other solvents. The activation entropy for the racemization of the disodium salt in pure methanol is also very large and negative. Racemization of the Acid in an Asymmetric Environment.-The dl-acid does not show any spontaneous resolution on heating in ~-2-methyl-lbutanol; hence the rates of racemization of the
Solvent Effects in the Racemization of the Acid. TABLEI -The rates and activation parameters for the racemization of o,o’-dimethoxydiphenic acid (I) RACEAIIZATIONO F (L,‘2’-DIMETHOXY-6,6’-DICARBOXYDIPHENYL^ I N VARIOUS SOLVENTS have been determined in a series of 24 solvents k x 109, (Table I). The solvent effects are quite compliaec. h AH AS *,tal./ cated. The enthalpy and entropy are roughly Solvent 100’ 79.4O kcal./molec inole-deg.c 16.55 2 . 5 2 23.24 f 0.00 -18.58 3 ~ 0 . 1 7 correlated with each other, to the extent that the Methanol 16.0 2.21 24.41 f . I 2 - 1 5 . 5 1 1 .34 spread in rates is much less than it would otherwise Acetic acid 15.4 2.03 26.07 i .13 - 1 3 . 8 1 f .3G be. However, the sequence of decreasing rates Acetone Ethanol 15.0 2 . 1 8 23.88 i . l l -17.00 i . 3 0 does not correspond to the sequence of decreasing Triflooroet hanol 14.3 2.16 23.39 f . 1 0 - 1 8 . 4 0 f . 2 8 14.1 1.95 or increasing activation parameters. Excluding Acetonitrile 24.45 f . l l - 1 5 . 6 3 f .31 the solvent N,N-dimethyle thanolamine, the range Dirnetliylformaniide 1 3 . 8 1 . 8 1 2 5 . 1 2 f . l l - 1 3 . 9 1 f . 3 1 glycol 13.0 1 . 7 8 25.14 f .12 -13.87 f . 3 4 in rate constants is about a factor of four, while Ethylene Benzyl alcohol 12.2 1.76 23.85 f .10 - 1 7 . 5 5 f .2C, AH+ and AS+ vary by 3.3 kcal. and 8 cal./mole-deg., Isopropyl alcohol 10.5 1.57 23.50 f . l l -18.77 f .30 respectively. There is some tendency for the re- Dimethylsulfoxide 1 0 . 4 1 . 2 5 2 0 . 1 7 f . 1 0 - 1 1 . 6 5 f . 2 7 8 . 9 9 1 . 3 2 2 3 . 7 2 f . 1 4 -18.51 * . 3 8 action to be faster in the more polar solvents, al- Tetrahydrofuran 8 . 4 3 1 . 1 5 2 1 . 6 6 f . l l - 1 0 . 1 1 3t . 3 0 though there is no correlation with macroscopic Pyridine 1,2-Dirnethoxyparameters such as the dielectric constant. Benethane 7.60 L O 4 2 4 . 0 5 f .07 - 1 0 . 3 4 f .20 zene added to methanol decelerates the reaction, &Butyl alcohol 5 . 9 3 0.709 28.31 f .11 -12.37 f .31 Dioxane 5 . 0 0 0.653 25.15 f .07 -15.85 f .20 while water accelerates it. Solubility limited the N-dirnethylamount of water that could be added, but the ac- N ,ethanolamine 0.01 1 . 5 2 22.14 f .27 - 2 7 . 2 9 i .73 celeration is less than that produced by water in hIethanol-benzene, the case of the disodium salt2; in both cases the wt. % benzene 27.0 14.3 2.21 2 2 . 9 8 f .11 - 1 0 . 5 7 f . 2 9 more aqueous solutions have the higher activation 12.7 1 . 8 9 2 3 . 5 2 f . l l -18.36 f . 3 0 52.7 enthalpies and the less negative activation en65.0 11.4 1.71 23.40 f . I 1 -18.9Of .32 tropies. The relationship between the activation 76.0 9 . 6 7 1 . 4 0 2 3 . 8 7 f . I 3 - 1 7 , 9 5 5 .36 enthalpy and entropy for the methanol-benzene 88.6 5.50 1.20 24.23 f .I1 -17.2,5f .30 MethanolTwater, mixtures is N-shaped. wt. % methanol Within the series of pure alcohols, the order of 70.3 1 8 . 5 2 . 6 1 24.14 f .09 -15.93 i .25 -1
(1) (a) Presented in part a t the Symposiuni “Solvent Effects and Reaction Mechanisms,” Queen Mary College, London, July 1957; (b) Based on the doctoral dissertation of B. $4. G.: (c) To whom requests for reprints should be sent. (2) J. E. Leffler and B. hf. Graybill, THISJ O U R K A L83,1457 , (1969).
44.1
19.5
2.68
*,
2 4 . 5 5 f .OS
-14.74f
.21
Diphenic acid concentration 0.013 M . b Mean rate constants. For the significance of the probable error figures and the method of calculation, see J . E. Leffler and W. H . Graham, THIS JOURNAL, 63, 687 (1959). a
1462
B. M. GRAYBILL AND J. E. LEFFLER
Vol. 63
d- and Lisomers in this solvent must be the same. The interaction between the acid and alcohol is therefore not very stereospecific. Solvent Effects in the Racemization of the Didimethyl ester I1 has been methyl Ester.-The racemized in 23 solvents and in the presence of HCl, sodium tetraphenylborate and trinitrobenzene (Table 11). The rates vary by about a factor of two, and the activation enthalpy and entropy have ranges of 2.1 kcal./mole and 4.4 cal./moledegree, respectively. The higher enthalpies of activation tend to be associated with the less negative entropies of activation. The more polar solvents, in which the rates a t 100" range from 12.9 X to 15.6 X 10-6sec.-1, give activation parameters falling on a single isokinetic line or linear enthalpyentropy relationship.a However, the solvents in to 8.28 which the rates are in the range 12.9 X X loF6sec.-l deviate from the relationship by having higher enthalpies of activation. The least polar of these solvents, for example, benzene, dioxane and carbon tetrachloride, in which the reaction is slowest, tend to give points about 0.4 kcal. above the isokinetic line. The racemization of the acid (Table I) also gives some points falling on the same isokinetic line, and the points representing activation parameters for the solvents in which
the racemization of the acid is slow again fall above the line. It is interesting to note that five solvents (methanol, acetic acid, acetone, ethanol and dimethylformamide) give points on the isokinetic line for both the ester and the acid. Of these five solvents, dimethylformamide gives the slowest reaction for the acid and the fastest for the ester, while methanol is fastest for the acid and slowest for the ester.6 That is to say, in spite of a general tendency for both the ester and the acid to racemize more rapidly in polar solvents, within a restricted group of solvents the racemization of the acid is slowest in solvents which are hydrogen-bond acceptors and the racemization of the ester is slowest in solvents which are hydrogen-bond donors. This suggests that hydrogen bonding between the ground state of the biphenyl and the solvent may be an important factor influencing the racemization rate. The ester racemizes much more slowly in trifluoroethanol than in ethanol, whereas the acid racemizes at about the same rate in both solvents. Symmetrical trinitrobenzene added to a chloroform solution of the ester accelerates the racemization. The acceleration by 0.23 M T N B amounts to about 13 to 16%, and is accompanied by a lowering of the entropy of activation. Although the effect is small it is in the expected direction. TABLE I1 A hindered biphenyl in its ground state will not RACEMIZATION OF THE DIMETHYL ESTEROF 2,2'-DIMETHform pi-complexes4 but a planar transition state OXY-6,6'-DICARBOXYDIPHENYLa IN VARIOUS SOLVENTS might well do so, especially if it does not possess k X 106, very much kinetic energy. Such a relatively static sec. -1b AH AS tal./ transition state is reasonable if the delocalization Solvent 100' 79.4' kcal./mdleC mole-deg.0 energy due to conjugation between the rings Dimethylformamide 1 5 . 6 2 . 1 9 24.25 f 0 . 0 8 - 1 6 . 0 0 f 0 . 2 2 14.9 2 . 0 3 24.65 f .10 -14.99 rt . 2 6 Nitrobenzene actually produces a metastable intermediate; 14.5 2 . 0 0 24.53 f . ' 1 - 1 5 , 3 7 f .31 Beneonitrile in that case the traiisition state can be regarded 1 4 . 4 2 . 0 2 24.30 2C .13 - 1 6 . 0 2 f .35 Isopropyl alcohol as consisting of moderately excited vibrational .18 Dimethyl sulfoxide 1 4 . 3 1 . 9 2 24.86 i: . 0 6 - 1 4 . 5 3 f levels of the intermediate rather than highly 1 4 . 2 2 . 0 0 24.21 * . l l - 1 6 . 2 7 f .31 Acetone 1 4 . 0 1 . 9 5 2 4 . 4 1 f .08 - 1 5 . 7 6 2 ~ . 2 2 Ethanol excited vibrational levels of the ground state. .31 &Butyl alcohol 13.7 2 . 0 7 23.32 f .11 - 1 8 . 7 3 * Solvent Effects in the Racemization of the DiHa030wt'% } 1 3 . 8 1 . 9 7 24.11 5 .12 - 1 6 . 5 9 f .32 amide 111.-The racemization of the diamide I11 is Methanol 70 wt. % Methanol 13.7 1 . 9 4 2 4 . 1 2 2C .10 - 1 6 . 5 8 f .27 considerably slower than that of the other derivaAcetic acid 13.6 1 . 9 1 24.26 f .09 -1G.243Z . 2 6 tives and the rates have been measured a t higher .37 Propionitrile 12.9 1 . 8 6 23.98 f .14 - 1 7 . 0 7 f temperatures (Table 111). The number of solEthylene glycol 12.9 1 . 8 1 24.30 i: .07 -16.234Z .18 vents studied was limited by the solubility of the Pyridine 11.5 1 . 6 0 2 4 . 3 5 f .15 - 1 6 . 3 4 f .41 1,2-Dimethoxyethane 1 1 . 3 1 . 6 4 23.86 f . l 2 - 1 7 . 6 8 f . 3 4 diamide. It is almost insoluble in benzene, carbon Trifluoroethanol 10.8 1 . 5 0 24.41 f . l l - 1 6 . 2 8 2 ~ . 2 9 tetrachloride and dioxane, and only slightly soluble Acetonitrile 1 0 . 4 1 . 4 2 2 4 . 6 3 2C . I 1 -15.803Z .31 in isopropyl alcohol, t-butyl alcohol, ethylene glycol Anisole 10.2 1.41 2 4 . 4 5 =t . 1 0 - 1 6 . 2 8 3 ~ .2G and chloroform. The range in rate constants a t Chlorobenzene 9 . 8 7 1 . 3 2 24.81 f .13 - 1 5 . 4 0 4 ~ . 3 6 .30 Benzene 9 . 4 5 1.31 2 4 . 4 0 2C .11 - l G , 5 7 f the temperatures used is about a factor of five, but Dioxane 9 . 1 4 1 . 1 9 25.25 * . l 2 -14.383Z . 3 3 the variation in the activation parameters is reChloroform 8 . 6 3 1 . 1 6 2 4 . 7 7 f . 0 8 - 1 5 . 7 6 A .21 markable: about 9 kcal./mole and 21 cal./moleCarbon tetrachloride 8 . 2 8 1 . 1 2 2 4 . 6 2 f . l 2 - 1 6 . 2 5 3Z .32 degree. The variation in rate constant as the Methanol, satd. with HCl 13.01 1.86 24.08 f .11 - 1 6 . 7 9 3Z . 3 0 solvent is 'changed would be extremely large if it plus 0.11 A4 sodium were not for the high degree of correlation betweeii tetraphenylborate 13.94 1 . 9 9 24.07 f .16 - 1G. 68 f . 4 5 the enthalpies and entropies of activation (Fig. 1). Chloroform plus As in the case of the acid and the ester, the five 0.012MTNBd 8 . 6 0 1 . 1 8 24.562C . l l - 1 6 . 3 3 3 ~ .31 8 . 9 0 1 . 2 2 24.60 f . 0 6 - 1 6 . 1 7 * .16 0.14 M TNB solvents acetic acid, methanol, ethanol, acetone 0.23 M TNB 9 . 7 2 1 . 3 5 24.40 * . 1 3 - 1 6 . 5 2 A . 3 5 and dimethylformamide give activation parameters Acetonitrile plus falling a t least roughly on an isokinetic line. The 0.07MTNF' 1 0 . 6 1.46 2 4 . 6 0 rt . I 7 - 1 5 . 8 0 * .46 Concentration of the ester 0.013 M . Mean rate con- isokinetic temperature, or slope of the AH*, AS* stants. e For the significance of the probable error figures relationship is about 120' or 20" higher than that and the method of calculation, see J. E. Leffler and W. H. for the ester and acid. The order of increasing Graham, THISJOURNAL, 63, 687 (1959). 1,3,5-Trinitro- rate for the racemization of the diamide in this henaene. * Tetranitrofluorenone.
*
~
(3) J. E. Leffler, J . Oro. Cham., 20, 1202 (1955).
*#
(4) C. E. Castro, L. J. Andrews and R. M. Keefer, J. A m . Chem. Soc., 80, 2322 (1058).
,
Sept., 1950
RACEhlIZATION O F
14G3
2,2'-DIMETHOXY-G,6'-DICARBOXYDIPHENYL
TABLE I11 RACEMIZATION O F THE DIAMIDE OF 2,2'-DIMETHOXY-6,6'DICARBOXYDIPHENYL~ IN VARIOUS SOLVENTS k X 107,
sec.-lb Solvent 108.9' 90.5' Acetic acid 20.3 4.60 16.8 2.73 Methanol 15.6 2.44 Ethanol Acetone 7.73 0.988 Dimethylformamide 7.68 1.49 Acetonitrile 7.11 1.27 Dimethylsulfoxide 5.41 0.950 52.6% benzene } d 14.1 2.25 47.4% methanol 69% benzene 13.5 2.15 31% methanol } d 55.9% water 27.9 4.24 44.1% methanol
AH
*
kcal./moiec 21.49 f 0.21 26.51 f .13 27.12 & .33 30.59 f .14 23.90 f .21 25.07 & .26 25.33 f .39
30
AS*, ca1.L
mole-deg. -28.81 f0.57 -16.04f .36 -14.61 f .88 - 6.91 =t .37 -24.45 & .55 -21.53k .69 -21.40 & 1.0
26.79 f .08 -15.68f
.31
20.89 f .12 -15.49 f .31 27.53 f .24 -12.37f
.63
4
2 25
a Concentration of the diamine 0.02 M . b Mean rate constants. 0 For the method of calculation and the significance of probable error figures, see J. E. Leffler and W. H. Graham, THISJOURNAL, 63,687 (1959). Weight per cent.
series of five solvents is almost the same as that for the acid and the reverse of that for the ester, suggesting that the ability of the amide hydrogen atoms to form hydrogen bonds is an important factor.6 There also seems to be some correlation between the rate and the polarity of the medium, since, as is also the case for the acid and the ester, dilution of a methanolic solution with benzene decelerates the reaction while dilution with water accelerates it. The most drastic medium effect on the racemization of the diamide is the effect of its own crystal lattice. Thus, the I-diamide has a characteristic melting point, 227-228", but quickly resolidifies and remelts at 270-271°, the melting point of the dl-diamide.6 I n any fluid medium above 200" the diamide would racemize instantaneously. Experimental The d- and I-2,2'-dimethoxy-6,6'-dicarboxydiphenylswere made as described in the previous paper of this series.3 The purification of the solvents and the details of the kinetic measurements have also been reported previously .7 Dimethyl Ester of 2,2 '-Dimethoxyd ,6 '-dicarboxy diphen yl . -To 25 cc. of 50% aqueous KOH and 80 cc. of ether cooled to 5" is added 7 g. (0.068 mole) of N-nitrosomethylurea with shaking. The diazomethane and about two-thirds of the ether are distilled over into a receiver containing 4.0 g. of the optically active acid and 40 cc. of ether. The acid (5) These rates do not fall in the same order as the enthalpies of activation because the working temperature is so close to the approximate isokinetic temperature that the differences in rate are largely the result of deviations from a perfect isokinetic relationship. (6) W. Stanley, E. McMahan and R. Adams, J . Am. Chem. Soe., 55, 706 (1933). (7) J . E. Leffler and W. H. Graham, T H r s JOURNAL, 68,687 (1959): W. H.Graham and J. E. Leffler, ibid., 63, 1274 11959).
3 -30
I
I
-
- 20
10 0 ASS. Fig. 1.-The correlation between the activation parameters for the racemization of the diamide in various solvents. and ethereal diazomethane are allowed to stand until all of the acid has dissolved and the yellow color is gone. Filtration and removal of the ether gives the dimethyl ester which is recrystallized from methanol-water to 3.9 g. of fine white D (acetone), +134O needles, m.p. 100-101', [ o ~ ] ~ , ~+152O (methanol). The racemic dimethyl ester melts at 134136 Diamide of 2,2 '-Dimethoxy-6,6'-dicarboxydiphenyl.-The I-diamide6 after recrystallization from ethyl acetate mel? at 227-228' but quickly resolidifies and remelts at 270-271 . Anal. Found: C, 63.71; H, 5.46. Calcd.: C, 63.99; H, 5.37. [ o 1 I z 4 ~-79" (methanol), -69' (acetic acid). Asymmetric Environment Experiments:-The dl- acid, 0.15 g., is diluted to 20 cc. with ~-2-methyl-l-butanol,filtered, the rotation measured and the solution heated for 70 hours a t 100". Initial polarimeter oreading 167.0992' at 24.2'; finalreading 167.0950' at 24.3 Error Analysis.-For a discussion of the expected errors and the tabulated measures of precision see references 2 and 7. The precision of the data for the acid and ester is comparable to that of the data for the salts and agrees with the expected error. The data for the diamide are somewhat less precise than expected on the basis of the known sources of error; however, the times of heating were longer and the temperatures higher than in the case of the acid and the ester.
.
.
Acknowledgment.-This investigation was supported in part by the Office of Ordnance Research, U. S. Army and by a Tennessee Eastman Company fellowship held by B. M. G.