Correlation of Solvolysis Rates. VI. t-Butyl and α-Phenylethyl Bromides1

tail in a preceding paperla; in general, the solvents employed Table I were based on an average of six points per run for the rate runs described in t...
0 downloads 0 Views 755KB Size
1.527618]. 'l'he average of 37 solvolysis infinity values indicated a purity of 98.7 0.3y0. Solvents.-Preparation of the solvents is described in detail in a preceding paperla; in general, the solvents employed for the rate runs described in this paper were from the same batches as those employed1a t o determine Y.

+

.-~

(18) (1040).

f . W. Bteitenbach and A. Maschin, 2. p h y s i k . chefs., ~ 1 8 7 181 ,

[C~NTRIRUTION FROX

THE

Kinetic Measurements and Experimental Results.-Tlie techniques employed for the kinetic runs and analyses have been described previously.1a The new d a t a reported iii Table I were based on an average of six points per run followed past 50-8O'jL reaction. The observed kinetics in all cases were first order within experimental error; the overall average deviation was &l.OYO of k . The estimated probable error, T , in log k is 0.005. Los ASGELES24, CALIFORNIA

DEPARTMENT OF CHEMISTRY OF

Correlation of Solvolysis Rates.

VI.

THE

UXIVERSITV OF CALIFORNIA A T Los ANCELES]

t-Butyl and a-Phenylethyl Bromides'

BY ARNOLDH. FAINBERG A I ~ DS. ~T-INSTEIN RECEIVED SEPTEMBER 21, 1056 As part of a detailed examination of the basis, scope and limitations of linear free energy relationships for the correlation of solvolysis rates, attention is now focused on effects arising from change of leaving group from chloride t o bromide. To this end, rates of solvolysis of t-butyl and a-phenylethyl bromides have been determined in an extensive series of solvent coml both series show a strikingly similar pattern; the points for the non-carboxylic positions. Plots of log kRBr us. log k ~ c for acid-containing solvent mixtures, e.g. mixtures of water with ethanol, methanol and dioxane, form a very good single straight line, while those for the carboxylic acid-containing solvents, e.g. acetic acid-water mixtures, form other lines for each binary pair, equally straight, but lying below the first. Possible reasons for the exceptionally low values for the bromide-chloride ratios of solvolysis rates in t h e carboxylic acid-containing solvents are discussed. The relative contributions of A( AH*) and A(AS*) t o increase in solvolysis rate arising from the change of chloride t o bromide are contrasted for the various solvent mixtures.

In the preceding paper,2 concerned with the correlation of solvolysis rates of a-phenylethyl chloride by the linear free energy relationship3 represented by equation 1 log k = log

KO

+ InY

(1)

a tendency toward dispersion of points into separate lines, one for each solvent pair, was noted. This illustrated one of the limitations3 of the nzY relation, arising froin one type of change of structure of the substrate. In the present work, we have focused our attention on effects arising from a change of leaving group from chloride to bromide. T o this end, we have measured the rates of solvolysis of t-butyl and a-phenylethyl bromides in a wide range of solvents. Results.-In Tables I and I1 are summarized all of the available first-order rate constants for the solvolysis of these compounds in mixtures of water with ethanol, methanol, acetic acid and dioxane. In addition, for t-butyl bromide rate constants for acetic-formic acid and acetone-water mixtures are listed. LIost of the data in Tables I and I1 are new; those which are cited from the literature are suitably footnoted, as are several entries which duplicate previously reported results. The Tables also list the values of the thermodynamic quantities of activation, AH* and A S * . mY Plots and Y Values.-When the data for tbutyl bromide a t %.o", in 28 solvent compositions, are fjtted to equation 1 hv the method of least scliiares, ~ a l u e sof wz = 0.934 arid log K O = -3.494 ( I ) Research sponsored by t h e Office of Ordnance Research, U. S. Army. (2) A . H. Fainberg and S. Winstein, THIS JOURNAL, 79, 1697 (1957). (3) (a) E. Crunwald and S. Winstein, ibid., 70, 846 (1948): (b) S WinPteiii, R. Grunwald and H. W. Jones, ibid., 73, 2700 (1951); ( c ) A . I f . I:ainiierg ani1 S \ V i n s t e i n . i b & f . , 78, 2770 (195G).

are obtained. However, the probable errorlo of the fit, 7 = 0.155, now obtained with this large variety of solvents is considerably larger than that previously listed3"vbfor data in five ethanol-water mixtures. Reference to the plot of log K for tbutyl bromide a t 25' as. Y, shown in Fig. 1, reveals that most of this error is contributed by the inclusion of the data for the solvent mixtures containing acids, acetic acid-water and acetic-formic acid mixtures. These two binary sets form excellent straight lines of their own, falling considerably below the line defined by the non-acid-containing solvent mixtures. The latter set, involving mixtures of water with ethanol, methanol, dioxane and ace tone, now gives a much improved fit, 7 b e k g 0.044 a t 25'. I t is nevertheless true, as previously observed* for a-phenylethyl chloride, that there is a still further improvement in the fit when each solvent pair is taken separately. Table 111lists the parameters m and log K O of equation 1, as well as the probable error of the fit, 7 , for each of these solvent groupings. The mean value of 7 for the five separate lines for t-butyl bromide corresponding to individual solvent pairs is 0.015 a t 0" and 0.010 a t 25". These compare favorably with the estimated probable error of 0.01 in log K for the experimental data for the bromide, and they demonstrate that the lines are quite linear. A further indication of the linearity of the separate lines for the data a t 25' is the fact that the lines all intersect each other a t very close to the same point and that this point corresponds to a value of Y near that of pure water.17 These equations thus furnish a ~iieansof estimating the solvolysis rate of t-butyl bromide in pure water a t 2c5.00. These extrapolations are summarized in ( 4 ) (a) M. L. Dhar, E. D. Hughes and C. K. Ingold, J. Chem. So?., 2065 (1948); (h) E. D. Hughes, C. K . Ingold, S. Masterman and IJ. J . hlcNulty, ibid., 899 ( I ! I ? O i

April 5 , 1957

1603

SOLVOLYSIS OF &BUTYL AND a-PRENYLETHYL BROMIDES TABLE I RATESO F

SOLVOLYSIS O F

&BUTYLBROMIDE^ AH*,

Solvent vol. % b

MethodC

0.00

X X X H

o.oo896 0.19ge 1.19h 4.22 13.1' 49.5 205 24OOp

105 k, sec. -Id 25.00

50.0'

kcal./mole (25')

AS*, e.u. (259

24.64/ 22.10 21.46' 21.67 21.17' 20.44 18.92 21.9

-0.4' -3.5 -2.3' +1.0 +1.4' +1.4 -1.4 14

EtOH-H20 100 90 80 70 60 50 40 Hz0 MeOH-H20 100 00 80 70 60 50 40

ACOH-HCOOH~ 100 50 HCOOHm ACOH-H~O~ 0 . 5 0 iM HnO 2.00 MH2O 4.00 M HzO 8.00 M H20 Dioxane-H20 90 80 70

io

60 50 40 30 Me2CO-H20" 83.0 mole % 78.2 mole % 68.5 mole % 5 8 . 9 mole yo 4 9 . 1 mole % 3 9 . 4 mole %

H H

x x 9 H H H H

x

0.440' 6.62* 35.8' 131.6 378' 1276 4170 770OOp

0 . 0845° 0.60e 11.2 37.9 111.2 282

3.44 19.5 87 314 940

0.0063'

0.302' 53.3 958

X

0.0524' .190" .99ge

1.98 6.14 27.6

X X X H H H H H

0.025e

0.63' 7.92 45.5 46.2 212 857 2807

X X

x

x x

1.82 8.84 37.1 128.6 358

0.012G" .043" .157 ,542 1.85

11.85/ 128.3

79 370

7.90'

12.8' 42.6 116 457

9.51

0.157 .304 1.06 3.55 12.5 42.4

+

23.40 21.9

-0.5 -2

20.97 20.2

+0.4 0

24.41

-1.9

22.91 21.89 20.90

-3.2 -4.4 -4.8

20.2

-15

20.36 19.99 19.73 19.36

-5.5 -3.7 -1.8 -0.7

20.0" 20. lo 19.67 19.67 19.63

- 17" - 14' -12.7 -10.2 -7.9

a Initial concentration 0.017-0.035 M where halide analysis was employed; 0.004-0.007 M where acidometric analysis was used. * x vol. Yo A - B means x volumes of A plus 100 - x volumes of B, each a t 25" before mixing. H = acidornetric The average deviation of all of the new rates herein reported that were constant was analysis; X = halide analysis. k l . 2 7 0 of k. Extrapolated from the data a t the other temperatures. f Calculation from data previously reported4 gives for 105ka t 25.0", 0.451; a t 50.0°, 11.2; AH* 23.94; AS* -2.7. 0 Previously reported6 7.17. Previously reported a t 0.1", 1.16,6 1 . 1 2 . 6 Calculation from data previously reported gives for lob k a t 25O, 37.8: 36.6: 35.8: 35.8,* 36.66; AH* 22.22; 21.7g5; AS* +0.4,8 - 1.2.6 f Calculated from data reported a t other temperatures by Bateman, ef uL6 IC Contained 0.065-0.068 M lithium salts (acetate and/or formate) plus 0.01 M AQO. These are initial rate constants; the rates drifted up in the course of these runs; see the Experimental section. Contained 0.068 M lithium formate. CalThese are rough estimates culated from data previously reported by Cropper, Spieth and OlsonQat other temperatures. based on data reported9 over a temperature interval of 5.2". p Extrapolated values; see text.

'

( 5 ) L. C. Bateman, K . A. Cooper, E. D. Hughes and C. K . Ingold, ibid., 925 (1940). ( 6 ) K. A. Cooper and E. D. Hughes, ibid., 1183 (1937). (7) 0 . T . Benfey, E. D. Hughes and C. K. Ingold, ibid., 2494 (1952). ( 8 ) H . C. Brown and A . Stern, T H I SJOURNAL, 72, 5068 (1950).

(9) W. H. Cropper, F . Spieth and A . R . Olson, i b i d . , 76, 6248 (1054). (10) See reference 3 c , footnote 13. (11) J. W. Baker and D. M. Easty. J . Chem. Soc., 1208 (1952). (12) G. R . Lucas and L. P. H a m m e t t , THIS J O U R N A L , 64, 1928 (1942).

1GOI

hIet11od.C

s s

0.0199'

S

.761/ 2.27 6.51 23.1 48.1 110 2600k

II H

Ii H If

11.62' 75.3 '789

(i2.0

175 539 974 1,547 56000k

*-I ,

1

!Hi

"2.8

'3 14

11.31 34. ij 98.6

2ii 1 760 lS9B 0.0i0'

s s s

0,0018' ,0101' ,0417'

21.71 20.30 2 0 , 23 "0.81 20.72 19.79 18.89

16,s'

I).92'

17. 0Q.'~

3314., hcal /mol< (25"j

19.2

I ) , 17'

s

s

0.628 4.92 19.1

.lo9

1 1 1 II I1 I1

S

20.00

0.00

2 I .-I 20.2 ,

81

.- 7.) -J ) :

0.0772 0,3ti4 1.X 7.5

I .OSh

1 . s92

-, i

-. 03

"3.8 125

I!) :!I IO. 11 20.18 24.3 23.90 22.63

21.06 ? ( I , 90

21). lii 0.0134 1.10i 0.00297/ 20.6 1.00 l.i.8 03Sf 20.16 6.92 ,282' H 103 1!1.07 1.44 34.6 I3 19.53 164 H 7.32 18.99 685 H 33.3 18.83 133.5 H 2680 Initial concentration 0.017-0.034 M where halide analysis was elnplnycd; 0.002-0.006 M where acidoinctric aiialysis v a s H = acidometric iised. b x vol. % -4 - B means x volumes of A plus 100 - 1c volumes of B, each at 25.0' before mixing. analysis; X = halide analysis. d The average deviation of all of the new rates reported herein that were constant was rt 1.0% of k . e Extrapolation to 54.9" gives 105 k = 19.5; previously reported4 a t this temperature, 19.2. 1 Extrapolated from the d a t a at the other temperatures. p At 74.82'. * These are initial rate constants; the rates drifted up during the Contained 0.068 MLiOAc f 0.01 A!fAcsO. I Contained 0.068 i~fLiOAc. course of these runs; see Experimental section, Extrapolated value; see text.

x

.

Table IV. The five aqueous solvent pairs yield a remarkably small range of values of log k , -0.05 to -0.17, averaging -0.1l5 0.026 ( k 0.77 =t 0.04 sec.-l). In order to make the corresponding extrapolatioiis a t O.O', it is necessary to calculate the least squares fits of log k for t-butyl bromide us. log k for t-butyl chloride a t this temperature. The parameters for these equations are listed in Table V. The values for log k in water for the bromide a t 0.0' extrapolated with these equations are listed in Table I V ; they average -1.62 0.04 ( k 0.024 f 0.002 sec.-l). Turning now to the data for a-phenylethyl bromide, a plot of log k for this compound against Y shows even greater dispersion of lines corresponding to separate solvent pairs than was previously observed? for the chloride. Again, treating the lines separately gives fairly good fits, as can be seen in Table 111. However, in exact analogy with the chloride, the lines for the ethanol-water and dioxane-water mixtures show noticeably more curva-

ture than those for methanol-water and acetic acid-water. Sufficient data are available for the solvolysis of t-butyl nitrate in ethanol-water mixtures" and in dioxane-water mixtures12 to make visible dispersion of lines arising from change of leaving group from chloride l o nitrate. The parameters for the equations for these lines are listed in Tablc I I I . These are also listed for t-amyl chlorideL5-" and t-amyl bromide4a*8.13a,*4 based on data from the literature. Among the data available from the literature on rates of solvolysis of the compounds treated in Table I11 are several rates in solvent compositions for which Y values based on solvolysis rates for t-butyl chloride are as yet undetermined. For these cases, the equations, the parameters for which are listed in Table 111, yield estimates of Y == -2.42 and -2.72 in 78.2 and 83.0 mole % Me%CO-H20, respectively (from solvolysis data for t-butyl bromideg); Y = -2.97 in 95 wt. yo = 79.5 mole Yo dioxane-water (from solvolysis data for

April 5 , 1957

1605

SOLVOLYSIS O F t-RUTYI, A N D a-PIIRNYI,ETIIYL nROMIDlTS

TABLE I11 CORRELATION OF SOLVOLYSIS RATESWITH Y NO.

Temp, OC.

Solvent range

of points

m

log koa

r

27 28

1.001 0.924

-4.918 -3.494

0.166 ,155

23

,953

-4.821

,090

22 7 7 6 5 3 4

,917 1.017 0.941

,044 ,028 ,012 ,018 ,007 ,014 .OOj ,012 ,009

5

0.947 ,946 1.200 1.067 0.963 ,931 ,921 ,930 ,936

-3.383 -4.967 -3.455 -4.942 - 3.426 -3.972 -5.2313 -3.780 -2.537 -4.718 -3.336 -4.661 -4.359

4

,913

-3.308

,006

l-Butyl bromide All solvents All solvents All solvents except those containing acids All solvents except those 25.0 containing acids 40-100 vol. % EtOH-Hz0 0.0 40-100 VOI. % EtOH-HzO 25.0 0.0 40-100 vol. yo MeOH-HrO 6G.100 vol. % MeOH-HIO 25.0 25.0 0-100% AcOH-HCOOH 0.0 0-8 hl Hz0 in AcOH 25.0 0-8 .AI HrO in AcOH 5 0 , o 0-8 hl HzO in AcOH 3cI-90 vol. % dioxane-Hz0 0.0 40-90 vol. yo dioxane-Hz0 2.i.U 39-69 mole % MezCO-HzOb 0.0 5 . 1 2 20-60 mole % MezCO-HzOb 39-69 mole % MezCO25.0 HzO'sg

0.0 26.0 0.0

4

5 6 G 4

1.010

.Ollj

,008 ,006 ,013

a-Phenylethyl bromide 0.0 28.0 0.0 25.0 0.0 25.0 0.0 25.0 50.0 0.0 25.0

All solvents All solvents 40-100 vol. 5 EtOH-HTO 40-100 vol 7 0 EtOH-Hz0 50-100 vol. % MeOH-HzO 60-100 vol. % MeOH-HZ0 0-8 I\. HzO in AcOH 0-8 M 1 ~ 2 in 0 AcOH 0-8 M H r 0 in AcOH 30-90 vol. % dioxane-HrO 30-90 "01. % dioxane-HzO

25 26

8 8 5 13

1.123 1.031 0.872 ,817 ,873 ,843 1.375 1 245 1.135 1.038 1.014

0.321 -5.415 ,294 -3.955 ,078 - 5 048 ,047 -3.642 ,025 -4.797 ,011 -3.388 ,007 -5.825 ,003 -4.364 ,007 -3.129 - Fi , 5 13 .050 -4.1:10 ,036

"

-2

+2

0

+4

Y. Fig. 1.-Plot of log k. os. Y for solvolysis of t-butyl bromide a t 25.0' in: EtOH-H20, 0; MeOH-H?O, 0 ; dioxaneHzO, 8 ; Me2CO-H20, 8 ; AcOH-HCOOH, (3; AcOH-H~O, 0.

t-butyl nitrate12); and Y = -2.73 and -3.26 for i-PrOH and t-BuOH, respectively (from solvolysis data of t-amyl bromidel4). Effect of Leaving Group.-The data for the tbutyl system show that, even with the structure of R in RX held constant, change in X gives disper&Butyl nitrate sion of lines in the log k vs. Y plots. For the CY2 0.938 4 . 9 5 7 GO-90 vol. % EtOH-H,OC 0.0 phenylethyl system, the effects due onlv to change 2 ,892 - 3 . 5 0 3 GO-90 vol. % EtOH-HzOC 25.0 in leaving group can be isolated by plotting log k 3 ,820 - 3 . 4 1 2 0 . 0 0 1 60-85 wt. % dioxane-HzOd 25.0 5 ,839 - 3 . 4 4 2 ,040 All solvents 25.0 for the bromide os. log k for the chloride, as shown t-Aknylchloride in Fig. 2. Here i t is seen that the points corre6(1-80 vol. % EtOH-HrO' 2 0.871 - 4 . 8 1 8 25.0 sponding to mixtures of water with ethanol, methanol and dioxane form an excellent single t-Amyl bromide straight line, while those for the acetic acid-water MeOH, 80-100 vol. % 25.0 0.026 mixtures form another line, equally straight, lying Et OH-HnO' 4 0.862 -3.233 Units of k , set.-'. Based on the data of Cropper, below the first. The latter line has the higher slope Spieth and Olson ,g These equations yield Y values of -2.42 and is headed for an intersection with the first line and -2.42 for 78.2 and 83 0 mole 7 0 Me*CO-H*O, respectively. Based on data of Baker and Easty." Based on a t the point for pure water. The parameters for data of Lucas and Hammett.l2 This equation yields a Y the equations for these lines a t several temperavalue of -2.97 for 95 wt. yo dioxane-HZ0 (NH~o 0.2047). tures are listed in Table V ; the small values of r eBased on the average of data reported by several au- demonstrate the excellent linearity. As before, t h o r ~ . ' ~ -'Based ~~ on the average of data reported by several Assuming negligible dispersion, these equations permit the estimation of the sol5 5 5 7 7

(I

this equation yields Y values of -2.73 and -3.26 for i-PrOH and t-BuOH? respectively, based on data of Brown and Moritani." 0 This equation reproduces the data of Tommila, et aZ.,l6 over the range of 10 to 38 mole % Me2CO-H20 with an average deviation of 0.033 in log k . However, in the range of 50 to 85 mole yoMe*CO-HgO, the data of Tommila, et aZ.,I6 are in very marked disagreement with those calculated from the d a t a of Olson, et aZ.,O listed in Table I. The latter are supported by data reported by Hughes, et al.,13b,o in 90 and 95 vol. yoMe2CO-H20. (13) E. D. Hughes and B. J. McNulty, J . Chem. SOC.,1283 (1937); (b) K . A. Cooper a n d E. D. Hughes, ibid., 1183 (1937); (c) L. C . Bateman, E. D. Hughes and C. K. Ingold. ibid., 960 (1940). (14) H. C. Brown and I. Moritani, TIUSJOURNAL, 7 6 , 455 (1954). (15) (a) V. J. Shiner, Jr., ibid., 7 5 , 2925 (1953); 7 6 , 1603 (1954); (b) H . C. Brown and H . L. Berneis, i b i d . , 7 5 , 10 (1953); (c) H . C . Brown and K. S. Fletcher, ibid., 71, 1845 (1949): ibid., 7 3 , 1317 (1951). (16) E. Tommila. M. Tiilikainen and A. Voipio, A n n . Acad. Sci. Fennicae, 66, 3 (1955). (17) This is i n contrast with a n earlier conclusion of Cropper, Spieth and

TABLE IV ESTIMATIONO F RATE O F SOLVOLYSIS O F I-BUTYL PHENYLETHYL BROMIDES IN WATER

AND

a-

log k (estimatedI4 a t 0.0

Solvent pair

a t 25.0'

log (kza.o/ka.o)

.!-Butyl bromide EtOH-HsO MeOK-HzO Dioxane-HI0 AcOH-Hz0 MeKO-HrO Average

-1 . 6 9

-1.63 - 1.61 - 1.56

-1.62

zk 0 . 0 4

-0.17 - .12 - ,12 - .05 - .I2 -0,115

1.52 1.51 1.49 1 50

+ 0.025

1 . 5 0 5 + O 01

a-Phenylethyl bromide Aq. E t O H , MeOH, -1.62 dioxane 1.53 AcOH-HzO -1.58 i 0.04 Average

-

a

log

-0,265 ,230 -0.25 i 0.02

-

-1.36 -1.30 - 1 . 3 3 =t0.03

Estimated by extrapolation, employing the equations k R B r = a log k R C l b , and the log k ~ values. ~ l

+

1GOB

ARNOLDH. FAINBERG AND

s. WINSTEIN

TABLE V CORRELATION OF SOLVOLYSIS RATESOF ALKYLBROMIDESWITII CHLORIDES, via Temp.,'"

R

t-Butylb

OC.

0.0

.o .0 .0 .0

EQvArIow log

TIE

s o .of ]mints

Solvent range

b

n

19

Aq. EtOH, MeOH, dioxane 40-100yo EtOH-H20 40-100yo MeOH-H20 30-90y0 dioxane-HzO 0-8 M H?O in AcOH

0.94" ,944 945 ,922 1.007

-,

0 G 4

k1113,

1.333 1.270 1.338 1 ,286 1.791

Aq. EtOH, AleOH, dioxane 17 0,910 0.81," 25.0 Aq. EtOH, MeOH, dioxane' 18 930 ,957 50.0 Aq. EtOH. hleOH, dioxane' 16 ,943 1. 0 5 i 0.0 0-S Ill H?O in AcOH 5 1.074 1,340 25.0 0-8 M H20 in -4cOH 5 1.077 1,192 50.0 0-8 Ad H?O in AcOH 5 1,081 1.007 a For the solvolysis rates of both bromides and chlorides. The corresponding parameters a t 25.0" are readily calculated from the proper entries in Table 111, noting that n = nz and h = 5,033 7% log ko c Excludes 10 vol. c; EtOIT-TI?O, which appears to be far out of line, a-Phen yleth yl

0.0

+

volysis rates of a-phenylethyl bromide in water a t these temperatures; these extrapolations are summarized in Table IV. Figure 2, in which a-phenylethyl bromide is compared with the chloride, shows a marked resem0-

ionizing power; for the a-phenylethyl system, it is less than one-sixth. COMPARISOX O F VENTS AT 25.0'. AND

TABLE VI BROMIDEn ' I I H CHLORIDE

IS P U R E SOLCOSTRIBUTIOVS O F CHANGE I N ENTROPY

ENTHALPY OF .ICTIVATIOS TO CHASGE r w FREEE W R G Y -T

l a P E

vent

0 1 m -2-

-%

1

$

1

f

%6 b

-4

I

c

c ~

z -

3

1

1

H?O MeOH EtOH iZcOH

kRcl

27 46 51 14

1-1~0 12 hZeOH EtOH .4cOH

23 29 4.6

e.11.

+14 -0.5 - .4 -1.9

e

11.

niole

t-Butyl -2.2 0.7 .,Y -2.6 -2.8 .8 .2 -0.6

~-Phenylethyl +5 -4.0 1.2 -2.8 0.8 -6 -1.6 .5 -9.4 ,I5 -6 -0.5

mole

1.3 I..?

t.5 1.4 0.4

1.0 1.*5 fl.75

inole

2.0 2.3 3.3 1 fi

1.6 1.8 2.0 0.9

Several possible reasons can be visualized for the low bromide-chloride ratios in acetic acid solvent: IL_-L--i--L (i) The blend of specific and general solvent in-8 -6 -4 -2 fluences which make up ionizing power3 for correlaLog k for a-Phenyleihyl Chloride tion of rates of solvolysis of chlorides is not quite Fig. 2.-Plot of log k for solvolysis of a-phenylethyl bro- suitable for correlating solvolysis rates of bromide at 25.0' v s . log k for solvolysis of a-phenylethyl chlo- mides; ;.e., there is some specificity associated with leaving group. ride a t 25.0" in: EtOH-H?O, 0; MeOH-H20, 0 ; dioxane(ii) The titrimetric rate constants, k t , are not HzO, 8 ; AcOH-H~O, 0 . equal t o the rate constants for ionization to the blance to Fig. 1 in which t-butyl bromide is com- first intermediate, kl,values of which are more apt pared with t-butyl chloride. In both cases, i t is the to follow a simple linear free energy relationship acid-containing solvents which are chiefly respon- than composite rate constants such as kt. To the sible for the dispersion of lines arising from change extent that ion pair return18 diminishes the titrimetin leaving group. This resemblance is not acciden- ric rate from the rate of ionization to some lower tal: similar examples will be presented in subse- value, we can expect i t to be more serious in acetic quent papers in this series. This result can be acid than in the other solvents and also more seristated in another way: the ratios of solvolysis rates ous for bromides than for chlorides. This comfor bromides to chlorides are much lower in the acid- bination would thus contribute to the dispersion containing solvents than in the nucleophilic sol- observed in the plots of log k R B r vs. log k ~ ~ i . vents. These ratios are listed for the pure solvents (iii) Nucleophilic character of the solvent is imfor the t-butyl and a-phenylethyl systems in Table portant t o solvolysis rate of the compounds in VI. The bromide-chloride ratio for the t-butyl question because their solvolysis is not limiting. sb system in acetic acid is hardly more than one(18) S. Winstein, E. Clippinger, A . H. Fainberg, R . Heck a n d G . C. quarter that in ethanol, a solvent of comparable Robinson. Tats JODRNAL, 78, 328 (1956).

April 5, 1957

SOLVOLYSIS OF L-BUTYL

This would contribute to the dispersion observed in log k R B r vs. k R C l plots if nucleophilic character of solvent were more important for bromide than chloride. I n the next paper in this series, a bromidechloride comparison is made for a system where bromide-chloride dispersion can be due only to (i) above. AH* and As*.-In a previous paper in this series,Ig an ABC classification was set up to describe the relative contributions of changes in AH* and A S * to change in rate arising from variation in solvent composition. I n Table VI1 a r t listed these classifications for t-butyl and a-phenylethyl bromides. Comparison with those previously noted for t-butyllg and a-phenylethy12 chlorides reveals a marked similarity in the pattern of the variation of the thermodynamic quantities of activation with solvent composition. TABLE VI1 CLASSIFICATION on THERMODYNAMIC BEHAVIOR Compound

Solvent range

AND

a-PHENYLETHYL BROMIDES

ever, the essential difference between acetic acid and the other solvents, the reason for the low kt-BuBr/kt-BuCl ratio and thus for the dispersion of ~ k&BuCl ~ ~ plot, is clearly lines in the log k t - vs.~ log seen to be the greatly decreased contribution of the A(AS*) term in acetic acid (0.6 e.u.) as compared with the value of 2.5 i= 0.3 e.u. for the other three solvents. One consequence of this difference is that A(AH*) is ca. 65% of A(AF*) for water, methanol and ethanol, while it is ca. 85% for acetic acid. The situation for the a-phenylethyl system is somewhat more complex, in that there is considerable variation in both the magnitude and relative contributions of the changes in AH* in the four pure solvents to change in AF*. Compensating variations in A(AS*) for water, methanol and ethanol tend to bring these solvents into line with each other. However, A(AS*) for acetic acid is much too small to compensate. Again, i t is this factor which is principally responsible for the low k B r / k c 1 ratio.

Class

t-Butyl bromide

40-100% EtOH-H20 60-100% MeOH-HZO 0-8 M H20 in AcOH 40-90% diox.-H20 39-83 mole yo Me2CO-H20

0 25 70 85 0 20 60 0 35 35 A 65 90

a-Phenylethyl bromide

40-10001, EtOH-Hz0 50-100% MeOH-H2O 0-8 M Ha0 in AcOH 30-90% diox.-H20

0 0 0 35

20 A 30

25 A 20 10 A 55

60 60 80 60 A 75 35 90

For both the t-butyl and a-phenylethyl systems, in all of the solvent compositions examined, the increase in solvolysis rate arising from change of leaving group from chloride to bromide is made up of a decrease in AH* plus an increasez0in AS*, the AH* contribution being a little larger than that of AS*. Thus, for the t-butyl system, on the average, A H * for the bromide is lower by 1.2 f 0.3 kcal./mole and A S * is higher by 2.9 f 1.2 e.u. than for the chloride. For the a-phenylethyl system, the corresponding comparison shows that AH* is lower by 0.8 f 0.4 kcal./mole and A S * is higher by 2.5 f 1.4 e.u. A more detailed examination of these relative contributions reveals significant differences between those found for the carboxylic acid-containing solvents as opposed to the others. This comparison is made for the pure solvents in Table VI. These differences provide a thermodynamic basis for the dispersion of lines in the log k R B r vs. log KRCI plots. Thus, for the t-butyl system, the change in AH* is quite comparable in magnitude for all four pure solvents, water, methanol, ethanol and acetic acid-ca. 1.4 f 0.1 kcal./mole. How(19) A. H. Fainberg and S. Winstein, THISJOURNAL, 79, in press. (20) Evans and Hamann*' previously noted that the values for A S for several bromides were larger than those for the corresponding

*

chlorides. They also called attention t o a similar difference between the entropies of solution of the gaseous bromide and chloride ions in aqueous methanol.'z (21) A. G. Evans and S. D. Hamann, Trans. Faraday Soc.. 47, 40

(1951). (22) W.L. Latimer and C. M. Slansky, THISJOURNAL, 62, 2019 (1940).

1607

Experimental Part &Butyl Bromide.-Anhydrous hydrogen bromide was passed into t-butyl alcohol t o yield t-butyl bromide, b.p. 72.6-72.7' (752 mm.), %*OD 1.4299, #D 1.4265; reportedZ3 b.p. 73.3' (760mm.),TPD 1.4306. a-Phenylethyl Bromide.-Anhydrous hydrogen bromide was passed into a-phenylethyl alcohol, b.p. 102.5-103.0' (20 mm.), to yield a-phenylethyl bromide, b.p. 97" (20 mm.), TPD1.5636, @D 1.5610, reported b.p. 98-99' (20 mm.),24 ~ P 1.5612.26 D This material was 94.3 f 0.7% pure, based on solvolysis infinities. Solvents.-Preparation of the solvents was described in detail in a previous ~ a p e r . ~ oIn most cases, the solvents employed for the rate runs reported in this paper were from the same batches as those employed30to determine Y. Kinetic Measurements and Experimental R e s u l k The techniques employed for the kinetic runs and analyses have been described previously.ao The new data reported in Tables I and I1 were based on an average of six points over a reaction range past 50 t o 80% reaction. The average deviation for all of the rate constants which were constant was 1 1 . 2 % of k for t-butyl bromide and 1 1 . 0 % of k for aphenylethyl bromide. The observed kinetics were first order within experimental error for all of the solvent compositions employed except those specifically footnoted in Tables I and 11. For those that drifted in the course of the runs, initial rate constants were estimated by linear extrapolation to zero reaction of plots of integrated rate constant us. percentage reaction; these are the values listed in Tables

TABLE VI11 UPWARDDRIFTSIN RATE CONSTAXTS DURING SOLVOLYSIS

Compound

&Butyl bromide

a-Phenylethyl bromide

Solvent

Temp., OC.

90% diox.-H20 80% diox.-H20 AcOH AcOH AcOH + O . 5 M H z O

25.0 50.0 25.0 25.0 50.0 50.0

-4cOH AcOH

50.0 75.0

90% diox.-H20

Increase in k per 0.01.M reaction,

7%

6 4 3

4.3 4.4 4

2 3

(23) J. Timmermans and Y. Delcourt, J. ckim. phys., 31, 100

(1934).

(24) C. L. Arcus, A. Campbell and J. Kenyon, J. Chem. Sac., 1510 (1949). (25) W.Reppe, 0.Schlichting, K . Klager and 7'. Toepel, Anrz., 660 1 (1948).

1605

ARNOLDH. FAINRERG A N D S. WINSTHIN

I and 11. The magnitudes of the drifts expressed in percentage change in the integrated rate constant per 0.01 mole of reaction are listed in Table WIT. These drifts appear

[CONTRIBUTION FROM THE

Vol. 7!)

to be salt effects and will be treated further it] a subsequent paper. Los ANGELES24, CALIFORNIA

DEPARTMENT OF CHEMISTRY OF

THE UNIVERSITY OF CALIFORNIA AT LOS ANGELES]

Correlation of Solvolysis Rates. VII. Neophyl Chloride and Bromide' BY fiRNOLD

H. FAINBERG AND

s. LTTINSTEIN

RECEIVED SEPTEMBER 21, 1956 Rates of solvolysis of neophyl chloride and bromide are reported in a variety of solvents. For the neophyl system, in which solvolysis proceeds by way of an anchimerically assisted ionization, the observed first-order titrimetric solvolysis rate constant, k t , is equal to the ionization rate constant, kl. Therefore, any failure of the linear free energy relationship, b, t o correlate rates of solvolysis of neophyl bromide with those of neophyl chloride must be due t o log krtx = a log k R y genuine failure of the relationship t o account for rates of ionization t o the first intermediate, rather than t o the fact t h a t kt is a composite quantity, dependent on the extent of ion-pair return. The plot of log k for neophyl bromide vs. log k for the chloride does, in fact, show a dispersion of the data into a pattern of lines quite analogous to t h a t previously observed for the bromide-chloride comparison for the t-butyl and a-phenylethyl systems. This dispersion is ascribed t o a leaving group specificity t o which the principal contribution arises from specific hydrogen bonding electrophilic assistance t o ionization.

+

+

log k = log KO mY (1) In other work, i t has been shown that solvolysis of neophyl derivatives I proceeds very predomi- for all solvent mixtures. However, i t is also obnantly by way of an anchimerically assisted ioniza- vious that there is a strong tendency for the data tion involving aryl parti~ipation.2,~Ion pair re- for each solvent pair to form a separate line. This turn4 of cationic intermediates I1 to the covalent dispersion of the mY plot into separate lines for condition would give rise t o very reactive tertiary each binary solvent mixture is similar to that prederivatives 111 which will survive only briefly. viously observed for a-phenylethyl ~ h l o r i d e . ~ Therefore, the observed first-order titrimetric sol- Therefore, as before, the data for each solvent pair volysis rate constant, kt, measuress the ionization rate constant, k,. For this reason, as in the study of salt effects,e neophyl derivatives are rather unique control substances for testki -; \ e / ing the scope and limitations of linear free \ ( CHa)2C-CH~ energy relationships in the correlation of sol- (CH&C--CH2 -+ (cH~)&='cEI~ --+ volysis rates. Any failure of a linear free en\Y8 Ye

0

@ I

Y

0

.

IIa

1Ib

ergy relationship for the correlation of rates I of solvolysis of a neophyl derivative must be due to a genuine failure of the relationship J. to account for rates of ionization to the first intermediate rather than to the fact that Kt is very J. (CHs)zC-CHz Products a composite quantity. I rapid In the present paper we report and discuss Y I11 the results of a study of rates of solvolysis of neophyl chloride and bromide in a variety of sol- are treated separately. Table I11 lists the parameters m and log ko of equation 1, as well as the vent mixtures. Results.-In Tables I and 11 are summarized probable error of the fit, r , for each set of binary the first-order rate constants for the solvolysis of solvent mixtures for neophyl chloride. On the neophyl chloride and bromide in a large variety of whole, the fits obtained by this modification of the solvent mixtures a t several temperatures. All original8 mY relation are quite satisfactory for of the data in these tables are new; their over-all many purposes. probable error is estimated to be less than 0.01 in A similar dispersion of a plot of log k for neophyl log k or 2% in k . Also listed are the values of the bromide US. Y into lines corresponding t o separate thermodynamic quantities of activation, AH* and solvent pairs is obtained; these data are also AS*. treated analytically in Table 111. Correlation of So~vo~ysis Rates with Y.-h Bromide-Chloride C~mparison.-~k before,g efFig. 1 is plotted log k for the solvolysis of neophyl fects due only to change in leaving group can be chloride OS. Y. It is clear from this plot that the isolated by plotting log k for neophyl bromide vs. data cannot be treated satisfactorily with a single log k for neophyl chloride, as shown in Fig. 2 for line expressed by equation 1 the data a t 50.0'. I n this plot, the points corre(1) Research sponsored by the Office of Ordnance Research, U. S. sponding to mixtures of water with ethanol, _d

pJ -

Army. (2) S. Winstein, et al., THISJOURNAL, 74, 1113 (1952). (3) R . Heck, unpublished work. (4) S. Winstein, E. Clippinger, A. Fainberg, R . Heck and G. C. Rohinson, THIS JOURNAL. 78, 328 (1956). (5) S . Winstein a n d K . C . Schreiber, ;bid,, 74, 2171 (1952). (fj) A . 1 3 . Fainlierg and S. Winstrin, i b i d . , 78, 2763 (1956).

(7) A. H . Fainberg and S. Winstein (Paper V), ibdd., 79, 1597 (1957). (8) (a) E. Grunwald and S . Winstein, ibid., 70, 846 (1948); (b) S. Winstein, E. Grunwald and H. W.Jones, ibid., 73, 2700 (1051). (9) A . FI. f ~ i i i i i t x r g a n d S . Winstein (Paper V I ) . i b i d . . 79, I (XI2 ( 1 957).