2770 AND S.

WINSTEIN
ARNOLDH. FAINBERG Vol. 78

Kinetic Measurements.-The usual sealed ampoule TABLE

I11
technique was employed. Development of p-toluene-
sulfonic acid in acetic acid was followed by titration with DOWNWARD
DRIFTSIN RATECONSTANTS
DURING
ACETOLY-
sodium acetate in acetic acid, as previously described .la7 SIS
Ethanolysis rates were followed acidimetrically as pre- --% Decrease in K t per 0.01 M reaction-
viously described.6 The new data reported in Table I were o-CHaOCsHnCHzCHz- p-CHaOC6H4CHzCHz-
based on an average of six points per run followed to 70 to (LiClO~)X 102, OTs on
'M ,500 750 750
90y0 completion. The observed kinetics for the runs in
-
BOO

ethanol were all first order within experimental error of ca. 0 4 9 3 i

~kl7~ .
However, in a number of the runs in acetic acid the
rate constants drifted down slightly in the course of the run.
In these cases, a plot of the integrated rate constant k t v s .
0.00100
,0100 10 11
2
0
-
9
I

per cent. reaction was treated as roughly linear. Extrapola- ,0400 1 6

tion to zero reaction gave the initial rate constants listed in .loo 4 6 0 5
Table I . Table I11 summarizes the magnitudes of these 1.00 0 2 0 0
downward drifts, expressed as the per cent. decrease from
kto per 0.01 M reaction. 3.00 0 2 0 0
6.00 0 0 0 0
(7) S Winstein, E. Grunwald and L I,. Ingraham, THISJOURNAL,
T O , 821 (1948). Los ANGELES24, CALIFORNIA

[ CONTRIBUTIOS FROM THE DEPARTMEXT

O F CHEMISTRY, UNIVERSITY OF CALIFORNIA AT L O S ASGELES]

Correlation of Solvolysis Rates. 1II.I &Butyl Chloride in a Wide Range of Solvent

Mixtures2
BY ARNOLD H. FAINBERG
.4ND s. WINSTEIN
RECEIVED
DECEMBER
1, 1955

This paper is one of a series concerned with a detailed examination of the basis, scope and limitations of the linear free
energy relationship for the correlation of solvolysis rates, log k = log ko +
m y . Solvolysis data for t-butyl chloride a t 25.0°,
and the Y values derived therefrom, are reported for a greatly extended range of solvent composition. The solvents include
mixtures of water with ethanol, methanol, formic acid, acetic acid, dioxane and acetone, and mixtures of formic acid with
acetic acid and dioxane. They cover the complete range of composition for each binary pair. For interpolation, the Y
values are expressed in analytical form as a function of the mole fraction of the more rapid component of each binary pair.
The variation of log k for t-butyl chloride with various functions of the dielectric constant of the solvent, including (D - 1)/
(20 ,+ l ) ,,l/D, and log D, is examined. The superiority of the mY correlation over correlations of log k with these functions
of dielectric constant is demonstrated for a number of compounds for which suitable solvolysis data were available in the
literature.

The development of accurate correlations of the vents was regarded as limiting (Lim.lb); in other
variation of rates of solvolysis of organic halides, words, the contribution of I1 to the hybrid transi-
benzenesulfonates, etc., with solvent change is im- tion state was regarded as negligible. This com-
portant for diagnosing and controlling reaction pound was made the standard substance for setting
mechanism in solvolytic substitution and rear- up a scale of Y values, these being set equal to log
rangement. In the first two papers] of this series, k l k , a t 25'.
the linear free energy relation 1 was explored. In
TABLE
I
this equation
log k = log ko + mY (1)
RATESO F SOLVOLYSIS O F &BUTYL CHLORIDEa AT 25 0" A Y D
Y VALUESFOR SOLVENTS
k and ko are the first order solvolysis rate constants Solvent, vol % b
10'k (sec -1):
M e t h o d C 25 00 =k 0 02 Y
for a certain substance in a solvent and in the stand-
ard solvent, SOY0 aqueous ethanol, respectively, and EtOH-H?O
nz is the compound parameter measuring sensitivity 100 x 0 00860e - 2 033
of solvolysis rate to Y, the measure of ionizing 98 s 0193" -1 681
power of the solvent. 95 x 0479" - 1 287
Our previous discussionlb of solvolysis was based 90 x 166' ' -0 747
on a continuous spectrum of transition states de- 80 x 926' 0 000
pending on structure and solvolytic conditions, 70 s 3 65h 0 595
structures I , I1 and I11 contributing to the hybrid 60 x 12 32% 1 124
transition state for thefirst step of a solvolysis. The 50 x 40 4
solvolysis of t-butyl chloride in the common sol- 50 H 41 9' 1 655
45 H 77 8 1 924
e a 3 a 40 X 141
SO: R--S SO-R :X SO: R :S 40 H 145 6 k 2 19h
H H H
37 5 H 202 2 338
I I1 I11
35 H 275 5 2 473
(1) I'reviuus papers in this series: (a) (I), E. Grunwald and S . 30 H 488 2 721
Winstein, THISJ O U R N A L , T O , 846 (1948); (b) (11), S.Winstein, E.
25 H 750 2 908
Grunwald and H . W. Jones, ibid., 7 3 , 2700 (1951).
(2) Research sponsored by the Office of Ordnance Research, U. S. 20 H 1043 3 051
Army 15 H 1431 3 189
June 20, 1956 SOLVOLYSIS CHLORIDE
OF &BUTYL 2771

Solvent, vol. % b
TABLE
6
I (Continued)
hlethodc
105k (sec.-l)do
25.00 =k 0.04 Y
50
50
40
H
X
H
23.9
22.4
88.7
i 1.398
1,981
10 H 1901 3.312 30 H 281 2.482
5 H 231OW 3.397 25 H 453 2.689
Hz0 H 2880" 3.493 20 H 758 2,913
MeOH-H20 10 H 1575 3.230
100 x 0,0753' -1.090 Dioxane-HCOOH"
90 X 0.464 -0.301 80 X 0.00469" -2.296
80 X 0.381 60 x ,195 -0.677
70 X 40 X 2.34 ,402
0,961
70 H 20 X 18.1 1.291
60 X Formamide-Hg0
1.492
GO H 100 X 3.72 0.604
50 X 86.6 80 X 22.4 1.383
1.972
50 H 87.2 97.5% AczO-AcOH H 0,00048" -3.29
40 H 228 2.391 x-C~F~COOH +1.7t
30 H 525 2,753 i-PrOH -2.73"
20 H 98 1 3,025 t-BuOH -3.26"
10 H 1760 3.279 a Initial concentration 0.02-0.04 M where halide analysis
A~OH-HCOOH" was employed; 0.001-0.005 M where acidimetric analysis was
100 X 0.0213 -1.639 used. X vol. yo A-B means x volumes of A plus 100 - x
100" ,0196" - 1,675" volumes of B, each a t 25.0" before mixing. c H = acidi-
X .log" -0.929 metric analysis; X = halide analysis. The average de-
90 viation of all of the new rates herein reported that are con-
75 x .62 - ,175 stant is & l . O ~ oof k; the probable error 7 in log k is esti-
50 x 5.3 .757 mated t o be 0.004. e Extrapolation to 25.0" of data re-
25 x 27.1 1.466 ported'" a t other temperatures in EtOH containing several
concentrations of KOAc, with simultaneous linear extrapo-
10 s 67.5 1.862 lation to zero salt concentration, gives 0.00916. f Pre-
HCOOH-Hz0" viously reported3 0.171. Previously reported 0.915,3
100 x 105° 2.054 0.854,4 0.917,6 0.906,6 0.892.1 Previously reported 3.56,6
95 x 135.0 2.163 4.03.3 Previously reported 12.6.3 i Previously reported
36.7.3 Previously reported 129.3 Previously reported
90 x 155 2,222
0.0811,8 0.0776.'0 Unless otherwise indicated, solutions
80 x 193 2.318 of AcOH-HCOOH, HCOOH-H20, dioxane-HCOOH, and
66'/3 x 265 2,456 AcOH-H20 contained 0.065-0.068 M lithium salts (acetate
50 X 408 2,644 and/or formate); see Experimental section for details.
X 828 2.951 '' Estimated value a t zero salt concentration. Previously
331/3 estimated l l O . l a p Initial rate constant; upward drift in
25 X 1170 3.100 the run; see Experimental section. * Estimatedg from data
14.29 x 1627 3,244 on t-butyl bromidelo in acetone containing 0.170 mole frac-
AcOH-HzOn tion HzO. Calculation from data reported3 a t other tem-
X 0.0369 -1.400 peratures gives 0.192. a Estimated by Grunwald and Win-
0.50 M HzO stein1%from data a t 70.0". Estimated from data on a-
2.00 M Hz0 X ,127 -0.863 phenylethyl chlorideg in this solvent. Estimated from
4.00 M HzO X ,366 - ,404 data on t-butyl bromideg in these solvents. Calculated
8.00 MHzO X 1.45 ,193 from data a t 50 and 75°.9 Previously reported" 2570 zt
X 8.93 ,984 300. Previously estimated'& 3300.
16.00 M HzO
60 X 30.6 1,519 Since the previous publications' on the correla-
50 X 80.4 1,938 tion of solvolysis rates, we have accumulated a sub-
40 X 190 2.312 stantial body of data probing the basis, scope and
25 X 046 2.843 limitations of linear free energy correlations. Be-
D ioxane-HzO cause of the important position occupied by t-butyl
90 X 0.00864' -2.030 chloride in some of the correlations, we have made
80 X . 136p -0.833 an especially thorough study of the solvolysis of
70 X ,955 ,013 this substance in a greatly extended range of sol-
00 X ,715 vents. The kinetic data a t 25' and the now exten-
50 x 1.361
sive set of Y values are presented and applied in the
50 H 21.8 present article. Also reported are the experimental
40 H 81.6 1.943 methods employed in the accumulation of the data
30 H 264 2.455 for this and a number of subsequent members of
20 H 699 2,877 this series of papers.
10 H 1529 3.217 (3) E. D. Hughes, J . Chem. Soc., 255 (1935).
MeZCO-HzO (4) K. A. Cooper, E. D. Hughes, C. K. Ingold, G. A. Maw and B. J .
McNulty, ibid., 2049 (1948).
95. 2' -2.76'
(5) W.C . Brown and R. S. Fletcher, THISJ O U R N A L , 71, 1845 (1949).
90 H (6) H . C. Brown and H. L. Berneis, ibid., 7 5 , 10 (1953).
-1.856
90 X ,0127" (7) V. J. Shiner, J r . , ibid., 7 6 , 1603 (19.54).
80 H .197' -0.673 ( 8 ) A. R . Olson and R . S . Halford, ibid., 59, 2644 (1937)
io H
, 130
(9) A . H. Fainberg, unpublished work.
(10) W. H . Cropper, F. Spieth and A. R . Olson, THISJ O U R N A L , 7 6 ,
70 X 1.25 6248 (1954).
60 H 5.80 796 (11) C . G . Swain and S. D . Ross, i b i d . , 68, 6;s (1!146).
27’72 ARNOLD AND S. WINSTEIN
H. FAINBERG Vol. 78

Results.-In Table I are summarized the first- of the more rapid component, water in the case of
order rate constants for solvolysis of t-butyl chlo- the aqueous solvent mixtures and formic acid for
ride in mixtures of water with ethanol, methanol, the acetic-formic acid and dioxane-formic acid
formic acid, acetic acid, dioxane and acetone, and in mixtures. As mentioned previously,la such plots
mixtures of formic acid with acetic acid and dioxane. are convenient for interpolation to obtain Y values
Most of the rate constants in Table I are new, a few for solvents of intermediate composition. For this
representing duplication as indicated in the foot- purpose, the curves in Fig. 1 have been fitted by
notes. For the first time, these data cover the the method of least squares to the power series
complete composition range for each of the binary shown in equation 2 , where N is the mole fraction
solvent pairs. The very high solvolysis rates in of the more rapid solvent component. For mixtures
pure water and in the very highly aqueous mixtures Y = a + bN + cN2 + d N 8 + eN4 (2)
were measured by a new technique described in the
Experimental section. For the whole body of data of ethanol, dioxane or acetone with water, the
in Table I, the probable error yl* in the rate con- three solvent pairs for which the plots of Y vs. N
stants is estimated to be 0.004 in log k or 1% in k . exhibit the most complex behavior, some simplifica-
Table I also lists the set of Y values for the vari- tion of the analytical form, as well as an improve-
ous solvents based on the new data, as well as a ment in the fits, were obtained by expressing the
few Y values, indirectly derived, for solvents in curves in two parts, one for the low, and the other
which the solvolysis rate of t-butyl chloride has not for the high water range. The 40 parameters for
been measured. Where the present Y values du- the resulting eleven equations, summarized in Table
plicate the previous’ ones, the revisions are minor, 11, reproduce the 76 experimental Y values with a
except in the case of the aqueous acetones. For the pooled12bprobable errorI3 of 0.011 in log k or 2.5%
latter solvents, Y values were previously based on in k . Interpolated values of Y of sufficient accu-
t-butyl bromide, but Table I lists the set based on racy for correlation purposes can thus be calculated
t-butyl chloride, the present values being different with these equations. However, it should be
from the previous ones by as much as 0.3. noted that the residual errors in these fits are mostly
Y vs. Mole Fraction-The variation of Y with systematic, and that somewhat greater accuracy of
composition of the binary solvent mixtures is illus- interpolation can be obtained (estimated probable
trated in Fig. 1 with plots of Y vs. mole fraction error of 0.004 in log k ) from a large scale plot of Y os.
N.
The solvent mixtures range from very poorly
to highly nucleophilic media of widely varying ion-
izing power. For example, the series of acetic-
formic acid mixtures is one in which ionizing power
is varied over a wide range while nucleophilic char-
acter is held relatively constant a t a level far below
that of the aqueous alcohols.1b The series of
aqueous formic acid mixtures, on the other hand, is
one where nucleophilic character varies more than
ionizing power. Nevertheless, the variation of Y
with mole fraction of the more rapid component is
remarkably similar for all of the solvent mixtures
represented in Fig. 1. Table I1 lists, for each sol-
Y
vent pair, the composition of solvent a t which the
curves inflect. All of the curves have a t least one
inflection, and two of them, for ethanol-water and
acetone-water mixtures, show a definite second in-
flection in the high water region. A second inflec-
tion for the dioxane-water curve is not as obvious;
however, a t the very least, the curve becomes linear
in the high water region. Of the six aqueous sol-
vent pairs, the methanol-water set is the only one
for which the curve does not show an inflection in
the low water region.
The portions of the plots of Y vs. ilr in the neigh-
borhood of the points of inflection can be treated
3 ~

02
~

04 06 58
I
IO
as linear. Such treatment has actually been ap-
Mole Fraction plied by Cropper, Spieth and Olson10 to solvolysis
of Y us. mole fraction of fast component
Fig. 1 -Plot data for t-butyl chloride and bromide in a limited
(HzO or HCOOH) for HCOOH-H20 (A): MeOH-H20 range of aqueous acetone mixtures. These authors
( B j ; ACOH-HZO (C); AcOH-HCOOH ( D ) ; EtOH-H20

JG?!!!
(13) The probable error of the fit,” 7, is calculated with the equation
____ -_
(E); hIezCO-HzO (Fj ; dioxane-HzO (G j ; and dioxane-
I1COOH (H).
= 0.6745
(12) (a) A. H . Margenau and G . bl. Murphy, “Mathematics of
Physics and Chemistry,” D . Van Nostrand Co., Inc., New York, N. Y., where d is the difference between theexperimentaland calculatedvalues,
1943, p. 502; (b) W. J. Youden, “Statistical Methods for Chemists,” n is the numberof piecesof data,andfisthenumberof disposableparam-
John Wiley and Sons, Inc., New York, N. Y., 1951, pp. 12, 15. eters employed in the equation.
June 20, 1956 SOLVOLYSIS CHLORIDE
OF &BUTYL 2773

TABLE
I1
PARAMETERS SQUARES FITSOF Y TO THE POWER
FOR THE LEAST SERIES,
Y =a + bN + cN2 + dN3 + eN4
Range N a t in- (n ;
Solvent pair VOl. % No a b C d flectionb f) rl3

EtOH-Hz0 100-40 0-0.83 -2.042 6.086 - 6.209 6.024 { 0.34

.85a
6 0.013

EtOH-Hz0 40-0 0.83-1.00 -17.772 37.766 -16.509 7 ,006

MeOH-HzOd 100-0 0-1.00 -1.079 3.362 +2.173 -0,944 .77 7 ,016
HCOOH-Hz0 100-0 0-1.00 +2.065 0.986 -1.342 1.801 .25 6 ,015
AcOH-Hz0 100-0 0-1.00 -1.632 8.448 -13.714 15.205’ .40 6 .009
Dioxane-H20d 90-30 0.34-0.92 -5.641 15.837 -20.374 13.885
e::( .0°5
3 ,005
Dioxane-HZ0 40-0 0.88-1.00 -9.100 12.601
Me2CO-HzOd 90-30 0.31-0.91 -5.101 14.955 -18,542 12.459
e:::( .0°7
MezCO-H20d 50-0 0.80-1.00 -7.277 10.784 5 ,011
AcOH-HCOOH 100-0 0-1.00 -1.634 4.544 -1,896 1.056 .60 3 .(I12
Dioxane-HCOOH 80-0 0.36-1.00 -5.777 12.477 - 9.666 5.022 .64 1 .005
N is mole fraction of the fast component of each binary pair, H20 for the aqueous solutions and HCOOH for the AcOH-
HCOOH and dioxane-HCOOH mixtures. *
Unless otherwise noted, calculated from the equations. ( n - f) is the num-
ber of independent checks of each equation, equal to the number of pieces of data fitted ( n )minus the number of parameters
(f) employed in the equation. These equations reproduce the data of Cropper, Spieth and Olson,lo adjusted t o 25.0”.
for the solvolysis of t-butyl chloride in MeOH-HZO, N = 0-0.60, in dioxane-H*O, N = 0.70-0.80, and in Me2CO-H20,
N = 0.51-0.80, with a probable error of 0.018 in log k . e Estimated from the plot of Y os. N . e = -4.810.
also cited several other examples of such linear de- and substituted benzyI23 p-toluenesulfo-
pendence of log k on N . They did, however, recog- nates in aqueous acetone. Also, such plots have
nize that these linear dependencies are the result of been made for ethylene bromo- and iodohydrins2*
inflections in the plots. in aqueous ethanol, for p-nitrobenzyl bromidez6in
Rates VS. Dielectric Constant.-The availability aqueous dioxane, and, recently, for methyl chloro-
of a large body of rate data on t-butyl chloride methyl ether26in mixtures of ethanol with dioxane,
solvolysis provides an opportunity to scrutinize benzene and carbon tetrachloride, and of mbutanol
possible correlations based on dielectric constant of with benzene. I n most of the cases, the range of
the solvent. If we rewrite equation 1 in the form dielectric constant examined was not large, so that
of equation‘ 3, the term mY is replaced by linearity of the plot was tested over only limited
log k = log ko +
log C f ~ x / f + ) (3)
portions of the possible ranges. In spite of this,
definite departures from linearity are observable in
(fRX/f+), the ratio of activity coefficients of the a number of the plots, and in some, the curvature is
solvolyzing molecule and the transition state, re- considerable.
spectively. Treating the activity coefficient of a I t becomes of interest to plot log k VS. ( D - 1)/
+
dipole in terms of electrostatic forces, Kirkwood14 ( 2 0 1) for t-butyl chloride for all of the data in the
has derived a linear relation between the log of ac- solvent mixtures for which dielectric constants are
tivity coefficient and ( D - 1 ) / ( 2 D +
l), where D available. These solvent pairs include O-90yo
is dielectric constant of the solvent. One can ex- aqueous dioxane ( D = 78-6),270-95% aqueous ace-
press both fRXand f+ in this form. Assuming di- tone ( D = 78-22),28 O - l O O ~ o aqueous ethanol
pole moments and dimensions of the RX and transi- ( D = 78-24)z9and O-lOO% aqueous methanol ( D =
tion state dipoles to be independent of solvent, and 78-33).a0 Such plots are shown in Fig. 2, which
the influence of non-electrostatic terms to be negli- covers the range of dielectric constant from 78 to 22.
gible, a linear relation is expected16*16between log This range includes all of the above-mentioned
k and ( D - 1 ) / ( 2 D 1).+
+ solvent compositions except 65-90% aqueous diox-
A plot of log k VS. ( D - 1 ) / ( 2 D 1 ) was made ane ( D = 22-6). The portion of this curve which is
by Laidler and EyringlSafor t-butyl chloride in 40- not shown is very strongly curved.
90% aqueous ethanol ( D = 58 to 29). Similar The breakdown of the simple linear relation be-
plots have been made by others, for example, for
these plots show considerable curvature, which is not substantially
acetic anhydride, l7-I9 benzoyl chloride, 18,zo and changed when log k itself is used.
(14) J. G. Kirkwood, J. Chcm. P h y s . , 2 , 351 (1934). (22) J. K. Kochi and G. S. Hammond, THISJO U R NA L , 76, 3445
(15) (a) K. J. Laidler and H. Eyring, A n n . N. Y . Acad. Sci., 39, (1953).
303 (1940); (b) S. Glasstone, K. J. Laidler and H. Eyring, “The Theory (23) J. K . Kochi and G. S . Hammond, ibid., 76, 3452 (1953).
of Rate Processes,” McGraw-Hill Book Co., New York, N. Y.,1941, (24) H. D.Cowan, C. L. McCabe and J. C. Warner, ibid., 72, 1194
p. 419. (1950).
(16) A linear relation is similarly expected15 for a reaction between (25) J. W.Hackett and H. C. Thomas,” ibid., 72,4962 (1950).
two dipoles through a polar transition state. (26) P. Salomaa, A n n . U n i v . Turkucnsis, A14 (1953).
(17) V. Gold, Trans. F a r a d a y Soc., 44,506 (1948). (27) F. E.Critchfield. J. A. Gibson, J r . , and J. L. Hall, THIS JOUR-
(18) V. Gold, J. Hilton and E. G. JetTerson,*1 J. Chem. SOC.,2756 NAL, 76, 1991 (1953).
(1954). (28) H. S. Harned and B. B. Owen, “Physical Chemistry of Elec-
(19) Gold’s first plot,” covering the 0-50% acetone range, appears trolytic Solutions,” A. C. S. Monograph No. 95, 2nd E d . , Reinhold
t o be quite linear; in point of fact, however, considerable curvature Publ. Corp., New York, N. Y., 1950.
was found when an error in the drawing was corrected,la and t h e range (29) J. L. Hall, private communication t o Professor E. Grunwald
was extended t o include t h e 80% acetone point.” of Florida State University.
(20) B. L. Archer and R. F. Hudson,” J . Chcm. Soc., 3259 (1950). (30) P. S. Albright and L. J. Gosting, THIS JOWRNAI, 68, IO61
(21) These authors plotted log [k/(Ha0)1 ws. ( D - 1)/(2D + 1); (1946).
2774 ARNOLD13. FAINBERG
AND S. WIXSTEIN Vol. 7s
of the plots for thc aqueous ethanol and aqueous
acetone mixtures, and t h a t the plot for aqueous di-
oxane is curved all the way t o pure water. A sec-
ond set of inflections in the low water regions of
aqueous ethanol (70-90%) and aqueous acetone
/ (‘70-9570) is also responsible for .the apparent lin-
earity of the curves in these regions.
Using a double spherical ion model for the acti-
vated complex, Scatchard3*plotted log k for solvoly-
sis of t-butyl chloride in aqueous methanol and eth-
anol against another function of dielectric constant,
1i D . Electrostatic considerations s~iggest’~ that
this is the function of D against which log k should
be plotted for reactions between ions, or between
-6 // inns and neutral molecules. However, it needs to be
emphasized that it is impossible t o decide betwcen
the functions l / D and ( D - 1)/(2D 1) on the +
basis of linearity of log k plots for dielectric con-
stants greater than ca. s. This can readily be seen
by carrying out explicitly the division of ( D - 1) by
1 . (20 + l), as shown in equation 4. For D = S, the
0.47 0.48 0.49 error involved in neglecting terms in D 2 and higher
( D - 1)/(2D + 1).
Fig 2.-Plot of log k GS. ( D - 1)/(2D +
1) for solvolysis
of t-butyl chloride at 25.0’ in dioxane-H20 (A); EtOH- powers is less than one per cent. Therefore, ( D -
H20 ( B ) ; Me2CO-H20 ( C ) ; &4eOH-H20 ( D ) . 1):’(2D +
1) is linear in ( I I D ) to an excellent ap-
proximation. This is true in all solvent mixtures of
+
tween log k and ( D - l),’(L?D l), due to the fail- water with acetone, ethanol and methanol. This
ure of the approximations involved and also solvent approximation would become inexact only for such
‘ ‘ ~ o r t i n g , ’ is
’ ~emphasized
~,~~ by two features of the solvent mixtures as aqueous dioxane containing
plots in Fig. 2 . These features are large dispersion less than lOy0water. It is obvious, therefore, that
of the plot into separate lines for each solvent pair the plots of log k us. 1/D for t-butyl chloride show
and curvature of the separate lines. Such disper- curvature and dispersion just as do the plots e’s. ( D
sion has been noticed, for example, by Salomaa2‘j - 1),/(2D 1) in Fig. 2. +
for ethanolysis of methyl chloromethyl ether in Bohine, et aZ.,39340 plotted log k against log D for
ethanol-dioxane and ethanol-benzene mixtures. a-phenylethyl chloride, benzhydryl chloride, meth-
Further, there are many indications t h a t dielectric anesulfonyl chloride and benzoyl chloride in mix-
constant alone is a poor guide to solvolysis rate31of tures of water with dioxane, tetrahydrofuran, eth-
halides such as benzhydryl and benzoyl ylene glycol dimethyl ether and p,,B’-dimethoxydi-
chloride,36and to rates of the Menschutkin reac- ethyl ether. In all cases except benzoyl chloride,
t i ~ n or l ~the ~ reaction of dilbutylamine with diiso- the plots were quite linear.
propyl phosphorochloridate. 37 The plots of log k us. log D for t-butyl chloride in
The departures from linearity in most of the re- aqueous dioxane, acetone, ethanol and methanol
gions of the curves is considerable. The linear por- are complex curves, as is clear from Fig. 3. How-
tions of the plots extend only over the ranges of ever, as a result of the numerous inflections, major
solvent composition of dielectric constant greater portions of the curves can be approximated roughly
than ca. 33. These include O-30yo aqueous dioxane by straight lines. Thus, as far as linearity of plots is
and 0-5076aqueous acetone, ethanol and methanol. concerned, log D is the best function of D so far
However, close examination of a large scale plot of considered. Howcl-er, the magnitude of the dis-
the high water region shows that inflections in the persion of the separate lines remains as large as for
curves are responsible for the apparent linearity the other functions of D .
(31) I n this connection, the d a t a in heptnfluorobutyric acid a n d in mY Correlations.-The usefulness of the mY
formamide are particularly striking. The dielectric constant for correlation of solvolysis rates becomes apparent
heptafluorobutyric acid must be close t o t h a t of acetic acid (Dann- when the same data which led t o marked curvature
hauser and Cole32 reported D = 6.2 for acetic acid, and D = 8.3 for
trifluoroacetic acid; the value for t h e latter is probably somewhat of the plots of log k e’s. (D - 1 ) / ( 2 D +
1) are
hizher than t h a t for heptafluorobutyric acid). EIorvever, the solvolysis replotted against Y. Table I11 lists the parameters
rate of a-phenylethyl chloride in heptafluorobutyric acid exceeds t h a t m and log ko of equation 1, as well as the probable
in acetic acid by four powers of ten.O T h e solvolysis rate of t-butyl error of the f i t , 1 2 s 1 3 Y, for all but one41of these sets of
chloride in formamidc ( D = 103.5)33 is ca. one-thirtieth t h a t i n formic
acid (D = J6.1),34 and is somewhat less than that in 607, aqueous di- ( 3 8 ) Q. Scntchard, J . C h e n . I ‘ h y ~ . ,7 , ti37 ( l W 9 ) .
oxane ( D = 26.6.j.2: (39, H. Bhhme and H . ,I. Ifenning, Z X a 1 u 1
(32) W. Dannhauser and R. 13. Cole, l t i r s J O U R S A L , 7 4, 6103 (-10) H. Bohme and W.Schiirhoff, C h e w . Iici.., 84, 28 (I!l5lJ
(1952). (41) Benzoyl chloride hydrolysis has hren recently dissecte<i l ~ y
(33) G. K I.eader, i b i d . , 73, 836 (19.51). Gold, e t o l . , l $ into two distinct solvolytic processes, which he rejiards
(34) J . F. Johnson and R . H. Cole, ibid., 78, 4336 (1951). a s separate S y l and Sv2 substitutions, Our analysis of the d a t a ,
(35) N. T. Farinacci and L. P. Hammett, ibid., 6 9 , 2512 (1937). which supports the existence of two solvolytic processes, is included
(36) D . A . Brown and R . F Hudson. J . Chem. SOL in a later paper. Our interpretation, however, is that one of the proc-
esses is s!ihstit{!timi :inil thc o t h e r i i arl<lition
June 20, 1956 SOLVOLYSIS CHLORIDE
OF &BUTYL 2775

TABLEI11
CORRELATION RATESWITH Y
OF SOLVOLYSIS
Temp., No. of
Compound OC. Solvent range points m lop koa rlJ

Acetic 5.0 0-50 wt. %

' Me2CO-H20 5 0.587 -5.226: 0.012
25.0 0-80 wt. % Me2CO-Hz0 7 ,579 -4.636 ,015
Ethylene bromohydrinZ4 70.0 0-80 wt. %' EtOH-Hz0 4 ,226 -6.245 ,033
Ethylene iodohydrinZ4 70.0 0-60 wt. %' EtOH-Hz0 3 ,166 -6.457 .013
l-Chloro-2-methylpropan-2-0143 85.1 0-80 V O ~ .%
' EtOH-HnO 5 ,281 -6.138 ,016
97.1 0-80 V O ~ .% EtOH-H20 5 ,292 -5.646 ,016
p-Iiitrobenzyl bromidez5 5.0 50-90 wt. %' Dioxane-HZO 3 ,340 -7.554b ,024
60.0 50-90 wt. % Dioxane-HzO 3 ,472 - 5.256b ,011
Benzyl p-toluenesulfonatezZ 25.3 23-71 mole % Me2CO-H20C 8 ,650 -4.711 ,019
p-Methylbenzyl p-toluenesulfonatez3 25.3 23-71mole % Me2CO-H20C 8 ,724 -3.275 ,014
p-P-Styrylbenzyl p-toluenesulfonatez3 25.3d 38-55 mole % MeZCO-HzO' 3 ,669 -2.455 ,005
p-Phenylethynylbenzyl p-toluenesul-
fonate23 25.3d 23-38 mole % MeZCO-Hz0" 4 ,741 -4.638 ,007
Methanesulfonyl chlorided0 25.0 50-90 wt. % Dioxane-HzO 9 ,465 -4.568 ,028
35.0 50-90 wt. % Dioxane-HzO 9 ,445 -4.141 .026
Ethyl c h l o r ~ f o r m a t e ~ ~ 25.0 50-90 wt. yo Dioxane-H20e 9 ,405 -4.595 .Oll
cis-5-Methyl-2-cyclohexenyl acid 99.85 50-90 vol. % MezCO-HzO 4 ,429 -4.039' .009
phthalate44 99.85 50-90 vol. % MezCO-HzO 4 ,472 - 4,193' ,001
0-90 vol. % MezCO-HzO \
2.3-Di~hlorodioxane~~ 15.0 14 ,879 -5,069 ,036
0-75 vol. % Dioxane-HZO
0-90 vol. % Me2CO-Hz0
35.0 14 ,845 -3.991 ,030
0-75yol. % Dioxane-HzO
15.0 0-90 vol. % MezCO-H20 10 ,873 -5.037 ,016
35.0 0-90 vol. % Me2CO-H20 10 ,839 -3.964 ,017
15.0 0-75 vol. % Dioxane-Hz0 5 ,919 -5.187 ,020
35.0 0-75 vol. % Dioxane-HzO 5 ,877 -4.084 ,016
Trimethylsulfonium hydroxide'6 100 0 0-100% EtOH-H20 - 4 ,778 - 1.731h ,017
Triethylsulfonium hydroxide46 100.0 0-80 V O ~'%EtOH-HzO
. - 3 ,835 - 1 .i O g h ,026
a Unit of ko, sec.-l, unless otherwise specified. * Hackett and Thomas25 reported their results as k h = k/(HzO); these
were converted to first-order solvolysis rate constants by multiplying by the following calculated water concentrations :
27.8,17.3and 5.73moles/liter for 50,70 and 90 wt. % dioxane-Hz0, respectively. Kochi and Hammond22rz3 state that
the solvent mixtures contained 1.84M LiC104; from their description of the preparation of the solutions, the actual con-
centrations employed were ca. 0.18 M . Temperature not givenz3; assumed to be 25.3'. e Values in the range 45 to 0
wt. % dioxane-HzO fall increasingly below this line. f k,; polarimetric rate constant. kt; titrimetric rate constant
Bimolecular rate constant, in liters/mole-second.

data. I n all cases, there is considerable improve-

ment in the linearity of the fits of log k vs. Y over r
that obtained for log k vs. (D - 1)/2 D
Also listed in Table I11 are a number of additional
1). + -2 t

c
cases from the literature which correlate well with
Y, and, therefore, very poorly with (D - 1)/(2D + -3
1). In particular, i t should be noted that, in the
case of 2,3-dichlorodioxane, where data are avail-
able in both aqueous acetones and aqueous diox-
anes, the enormous dispersion that exists when the -4 !
data are plotted against (D - 1)/(2D +
1) is re-
duced very considerably when plotted against Y.
I t is nevertheless true, however, that the fits are
3
still better when seDarate lines for each of the sol-
vent pairs are used. -Such dispersion of lines is treated
in subsequent papers of this series. The pooledIzb
probable error of fit for all of the data in Table 111,
employing separate lines for different solvent pairs,
is 0.018 in log k , which corresponds to less than 5y0
in k . This is typical of the fits which have been ob-
tained with the mY relation.
I I 1
(42) A. R . Emery and V. Gold, J . Chem. SOL., 1455 (1950).
(43) G. J. Harvey, N. V. Riggs and V. R . Stimson, ibid., 3267 1.5 1.7 1.9
(1955).
(44) H . L. Goering and E. F. Silversmith, THIS JOURNAL, 71, 1129 Log D.
(1955).
(45) P. Salomaa, Acta Chem. Scond., 8 , 744 (1954).
Fig. 3.-Plot of log k 8s. log D for solvolysis of t-butyl
(46) J. L. Gleave, E. D. Hughes and C. K. Ingold, J . Chem. SOC., chloride a t 25.0"in dioxane-H20 ( A ) ; EtOH-HzO ( B ) ;
236'(1935). Me2CO-H20 ( C ) ; and MeOH-H20 (D).
 2776 .IRNOLD AND S.WINSTEIN
H. FXINBERG Vol. 78

Since Y correlates log k for certain reactions for Molar concentrations were calculated at 25.0’.
which electrostatic considerations suggest the vari- Kinetic Measurements.-The usual sealed ampoule tech-
able ( D - 1 ) / ( 2 D +
l), i t may also be employed
to correlate certain reactions for which the sug-
nique, using 5-ml. aliquots, was employed for rate measure-
ments above 25”, and also for the slower rates a t 0 and 25’.
Special precautions were taken to minimize possible error
gested variable is 1 /D. This follows because (D - due t o the volatility of the t-butyl chloride. Thus, for the
+
1 ) / ( 2 D 1) is linear in 1/D. One such reaction is
the bimolecular decomposition of trialkylsulfonium
slower rates, in preparing the solvolysis solution, the t-butyl
chloride was weighed into a 100 ml. volumetric flask pre-
viously made up almost to the mark with solvent at room
hydroxides in aqueous ethanol mixtures.46 Here, temperature. After mixing, the flask was cooled to just
passage t o the transition state involves a decrease slightly above the dew point (generally ca. 10-15’) and the
in charge; hence, change to a more strongly ioniz- entire contents were poured into a 100-mI. all-glass hypo-
dermic syringe. The solution was dispensed into the am-
ing medium decreases rate of reaction. Laidler poules from the syringe. In this way, the difference in com-
and Eyring16aplotted the logarithms of the bimolec- position between the first and last ampoules was minimized.
ular rate constants for the decomposition of tri- The vapor space above the solutions in the ampoules was
methylsulfonium hydroxide against 1/D; the re- kept as small as possible. With these precautions, the ex-
perimental infinity titers obtained after ten half-lives of re-
sulting plot is markedIy curved. In contrast, an action, or, in the case of very slow rates, the analytically
excellent correlation with Y was obtained] as shown determined initial concentrations (obtained by pipetting an
in Table 111. aliquot into 25 ml. of 40% aqueous ethanol, allowing i t t o
solvolyze to completion, and then analyzing for chloride
Experimental Part ion), were in most cases found t o agree with the calculated
&Butyl Chloride.-Treatment of t-butyl chloride with value within 1 to 2%.
phosphorus pentoxide, followed by fractional distillation, For most of the fast rates, the solvolysis was carried out
gave a center cut, b.p. 50.8’ (760 mm.), naaD1.3855, @D directly in the hypodermic syringe immersed in the constant
1.3819, d2S4 0.8370 (reported b.p. 50.7” (760 rnn~.),~~ nZo~ temperature bath, aliquots being taken by forcing the solu-
1.3855,48 dZ540.83534’). tion up into the pipet. I n such cases the vapor space was
Solvents.-Methanol was dried by the method of Lund completely eliminated.
and Bjerrum4$;d2540.78651-0.78658 (reported60 d2540.7866). Rates at 0’ were carried out by immersing the hypodermic
Ethanol was dried b y the method of Lund and B j e r r ~ m ~ ~syringe ; containing the reaction mixture in a mixture of
d2Sa0.78505-0.78513 (rep0rted5~.d*~r 0.78507). crushed ice and distilled water in a Dewar flask, in a cold
Acetone was dried by refluxing over Drierite and frac- room held at 0 i 2’. In this way, i t was possible to em-
tionally distilling therefrom; dZ540.7844-0.7847 (reportedK0 ploy a pipet which was approximately a t the solvolysis tem-
dZ6a0.7850). perature.
Dioxane’was purified by the method described by Fieser5’; Aliquots were analyzed for acid by titration with meth-
d254 1.0261-1.0279 (reported50d25( 1.0280). anolic sodium hydroxide, using brom thymol blue indicator.
Formamide was fractionally distilled, b.p. 99-100” (9.5 For fast rates, the aliquots were effectively quenched by
mm.), ? z Z 5 ~1.4445 (reported6O n Z 51.4459).
~ pipetting into 25 ml. of acetone. I n this solvent, the indi-
Formic acid (B. and A., C.P. 98-100%) was dried by the cator employed, several drops of a 0.1% neutralized alco-
method described by Winstein and Marshall52; d25, 1.21381- holic solution of brom phenol blue, gave a remarkably sharp
1.21397 (reported50 d254 1.2140). Lithium formate in dry end-point at very low acid concentrations to M).
and aqueous formic acid was prepared by solution of the While this end-point did not correspond exactly to the
calculated amount of anhydrous lithium carbonate. A stoichiometric point (ca. 4% low), its position was found
solution 0.068 IM in lithium formate therefore contained to be substantially independent of acid concentration at
0.038 M added water. Such a water concentration is with- levels below 0.03 M ; this “error” therefore had no appreci-
out significant effect on solvolysis rate. Solutions of lithium able effect on the accuracy of the rate data. Using the ace-
formate in the dioxane-formic acid mixtures were prepared tone quench, satisfactory rates were obtained with concen-
by dissolving therein the calculated amount of anhydrous trations of t-butyl chloride of 0.001 M and lower.
lithium formate. The latter was prepared by reaction of Aliquots were analyzed for chloride ion by a modification
equimolar quantities of lithium carbonate and formic acid of the procedure described by Bateman and Hughes.55 An
in aqueous solution, followed by quantitative dehydration aliquot was pipetted into 25 ml. of pentane in a 50-1111.
of the salt under vacuum a t 110’. separatory funnel. This solvent is superior to the carbon
Anhydrous acetic acid was prepared by distillation from tetrachloride previously employedK6because i t forms the
triacetyl borate,’3 or by addition of the calculated amount upper phase, thus simplifying the extraction procedure, and
of pure acetic anhydride, refluving for 12 hours, and distilla- because i t separates more rapidly and cleanly from the
tion. The amount of acetic anhydride required was deter- aqueous phase. The mixture was extracted with two suc-
mined by Karl Fischer t i t r a t i ~ n . ~ Acetic
‘ acid solutions con- cessive 10-ml. portions of water. The procedure was
taining lithium acetate and water or excess acetic anhydride standardized a t 40 “shakes” for the first extraction and 30
were made up by weighing in the calculated amounts of for the second. The aqueous extracts were combined,
lithium carbonate and distilled water or pure acetic anhy- acidified with nitric acid, and titrated for chloride ion by the
dride. I n the case of the acetic anhydride, the resulting Volhard procedure, nitrobenzene being employed.6B The
solution was refluxed for 12 hours. extraction procedure was shown to detect at least 99.8% of
Other mixed solvents were prepared by mixing the re- the halide ion present in the solvolysis solution. Hydroly-
quired volumes of the components which had previously sis during the aqueous extraction was found to be negligible
been equilibrated a t 25.0’. Density checks on a number of for t-butyl chloride.
the aqueous methanol and ethanol solvents showed that the The following technique was developed for the determiria-
compositions agreed with the nominal values within 0.04 tion of very fast rates with a half-life between several min-
wt. yo. utes and CQ. 20 seconds. The material being solvolyzed
was roughly measured into the solvent which had previously
(47) J . Timmermans and Y . Delcourt, J . Chem. Phys., 81, 98 (1934). been equilibrated a t the solvolysis temperature. The solu-
(38) A. Audsley and F. R. Gosa, J . Chem. Soc., 873 (1941). tion was rapidly mixed, immersed in the thermostat, and
(49) H. Lund and J. Bjerrum, Ber., 64, 210 (1931). the pulling of points started at once. Each aliquot, taken
(50) J. Timmermans, “Physico-Chemical Constants of Pure Or- with a fast-draining automatic pipet, was discharged either
ganic Compounds,” Elsevier Pub. Co., Inc., New York, N. Y., 1950. into 25 ml. of acetone for acid determination, or into a sepa-
(51) L. F. Fieser, “Experiments in Organic Chemistry,” D. C. ratory funnel, containing 25 ml. of pentane and 10 ml. of
Heath and Co., New York, A*. Y., 1941, p. 369. water at the solvolysis temperature, for halide determina-
(52) S. Winstein and H. Marshall, THIS JOURNAL, 74, 1126 (1952). tion. I n the latter case, the separatory funnel was then
(53) W. C. Eichelberger and V. K. La Mer, i b i d . , 66, 3633 (1933). shaken, time being measured a t the first “shake.” Separa-
(54) I n t h e Karl Fischer titration of acetic acid, it was found neces- -
sary t o carry out t h e titration in the presence of a t least a fivefold ( 5 5 ) 1,. C. Bateman and E. D. Hughes, J. Chcm. SOC..1190 (1937).
excess of methanol; otherwise, the reaction with water was erratic, (56) J. R . Caldwell a n d H. V. Moyer, I n n . Eng. Chcm., Anal. E d . , 7 ,
s l o w and sometimes completely inhibited. 38 (193.5).
June 20, 1956 SALTEFFECTSIN ACETOLYSIS
SPECIAL 2777

tion of the aqueous layer and re-extraction of the pentane formic acid mixtures, even in the presence of a twofold ex-
layer with the second 10-ml. portion of water was deferred cess of lithium acetate. As shown in Table IV, in the sol-
until after all of the points were pulled, i t having been dem- volysis of 0.034 M t-butyl chloride in acetic acid containing
onstrated that the reaction was effectively quenched by the 0.068 M lithium acetate, the experimental infinity, taken
first extraction. In this way, the first point could be pulled after ten half-lives and accurate to & l % , was found to be
within 20 to 30 seconds of mixing, and subsequent points 6.2% smaller than the analytically determined initial con-
could be pulled every 15 to 20 seconds. centration of t-butyl chloride. A much higher concentra-
A microphone feeding a tape recorder was employed for tion (1 M ) of lithium acetate was found to suppress the back
the time measurement in the fast rates, the quench time for reaction completely, as did the addition of 2 M water. The
each point being announced orally. The elapsed times be- addition of formic acid was only partly effective. Even in
tween points were later obtained from the tape on playback, pure formic acid containing 0.068 M lithium formate, a
using a stopwatch. Accuracy was ensured by reading a measurable equilibrium was observed.
time signal into the tape before and after each run, and
employing this interval as a calibration time check. In this TABLE IV
way, timing accuracy of 2t0.270 was readily attained. PER CENT.CHLORIDEIONAFTER 10 SOLVOLYSIS
HALF-LIVES
The new data reported in Table I were based on an aver-
age of six points per run followed past 50 to 90% reaction. USDER VARIOUS CONDITIONS
The average deviation for all of the rates which were con- (Lithium % c1-
Temp., (I-BuCI), salt), after 10
stant was =tl.O%; in only four cases did the average de- Solvent O C . M M half-lives
viation reach &2%: 50% HCOOH-H20 a t 25.0°, 16 M
AcOH f 0.01 M A c ~ O 50.0 0.034 0.068 93.8
Hz0 in AcOH a t 50.0°, and 5070 AcOH-HCOOH a t 25.0
and 50.0'. The observed kinetics were first order within AcOH +
0 . 0 1 M Ac2O 50.0 ,032 1.000 100
experimental error for all of the solvent compositions em-
ployed except 80 and 90 vol. % aqueous dioxane a t both
AcOH +
2 M H20 50.0 ,033 0.068 99
75% AcOH-HCOOH 50.0 ,038 .066 95.6
25.0 and 50.0". I n the former solvent, the integrated
first order rate constant calculated from the true zero time 50% AcOH-HCOOH 25.0 .035 .065 97.3
rose 2.6% for the reaction of the first 0.01 M t-butyl chloride 50% AcOH-HCOOH 50.0 .035 .065 97.1
a t both temperatures; in the latter solvent, the correspond- 25% AcOH-HCOOH 0.0 ,051 ,065 96
ing rise was 5.6%. 25% AcOH-HCOOH 25.0 ,051 ,065 97.6
Equilibrium in Acetolysis of t-Butyl Chloride.-The acetol-
ysis of t-butyl chloride comes to an equilibrium position far HCOOH 25.0 ,042 ,068 97.8
short of completion unless the acetolysis is carried out in the In computing the rate constants for these solvolyses, the
presence of excess base.",5'*5* The indications in the present initial analytically determined concentrations of t-butyl
work were that solvolysis of t-butyl chloride fell slightly chloride were employed in place of the experimentally de-
short of completion in acetic acid, formic acid and acetic-
termined infinities. In all cases, the drop in the integrated
(57) J. Stiegman and L. P. Hammett, THISJOURNAL, 59, 2540
rate constants as the equilibrium position was approached
(1 937).
was not perceptible until 60 to 70% of reaction; later points
were not averaged.
(58) W.G. Young, S . Winstein and H. L. Goering, ibid., 73, 1958
(1951). Los ANGELES24, CALIFORNIA

[CONTRIBUTION FROM THE DEPARTMENT

O F CHEMISTRY, UNIVERSITY OF CALIFORNIA AT LOS ANGELES]

Salt Effects and Ion Pairs in Solvolysis and Related Reactions. VI. Additional
Examples of Special Salt Effects in Acetolysis1
H. FAINBERG,
BY ARNOLD AND S. WINSTEIN
GENEC. ROBINSON
RECEIVED
JANUARY 30, 1956

trans-2-p-Anisylcyclopentyl p-bromobenzenesulfonate, trans-2-anisylcyclohexylp-bromobenzenesulfonate, l-p-anisyl-2-

propyl p-toluenesulfonate and erythro-3-p-anisyl-2-butyl p-bromobenzenesulfonate are shown to exhibit the same special
salt effect of lithium perchlorate in acetolysis as was previously demonstrated for 2-0- and 2-p-anisylethyl p-toluenesul-
fonates. The numerical quantities characterizing the magnitude of both the normal and special salt effects are tabulated
and compared.

I n the preceding paper2in this series, i t was shown present paper is reported a study of such salt ef-
that, in the acetolysis of 2-0- and p-anisylethyl p - fects in acetolysis of trans-2-p-anisylcyclopentyl p-
toluenesulfonates, lithium perchlorate exerts a bromobenzenesulfonate, trans-2-p-anisylcyclohexyl
special salt effect superimposed on the normal salt p-bromobenzenesulfonate, 1-p-anisyl-2-propyl p-
effect previously demonstrated3 for the neophyl toluenesulfonate, and erythro-3-p-anisyl-2-butylp -
esters. I n this special salt effect, a steep rise in the bromobenzenesulfonate. For all of these sub-
titrimetric rate constant, kt, from the base value, stances, acetolysis involves predominantly or ex-
k:, is produced by low concentrations of lithium clusively anchimerically assisted i o n i z a t i ~ n ~to~ , ~ . ~
perchlorate. This phenomenon, apparently con- a bridged carbonium ion, the anisyl group partici-
nected with the reduction of ion pair r e t ~ r n , ~pating,~ as in the 2-anisylethyl cases.8
has been noted in a number of other cases.6 I n the I n Table I are summarized the first-order rate
(1) Research sponsored by the Office of Ordnance Research, U. S. Fainberg and G. C. Robinson, Chemisfry a n d Industry. 664 (1954);
Army. (c) G.C. Robinson, E. Clippinger and S. Winstein, p. 21-N,Abstracts
(2) A. H. Fainberg and S. Winstein, THISJOURNAL, 78, 2767 of Papers, 125th meeting, A.C.S., Kansas City, Mo., March, 1954;
(1956). (d) G. C.Robinson, unpublished work.
(3) A. H. Fainberg and S. Winstein, <bid., 78, 2763 (1956). (6) S. Winstein, M. Brown, K. C. Schreiber and A. H. Schlesinger,
(4) S. Winstein, E. Clippinger, A. H. Fainberg, R. Heck and G. C. THIS JOURNAL, 74, 1140 (1952).
Robinson, ibid., 78, 328 (1956). (7) R. Heck, unpublished work.
(5) (a) S.Winstein, E. Clippinger, A. H. Fainberg and G. C. Robin- (8)S. Winstein, C. R. Lindegren, H. Marshall and L. L. Ingraham,
son, ibid,, 76, 2597 (1954); (b) s. Winstein, E. Clippinger, A. H. THISJOURNAL, 75. 147 (1953).