The Grunwald-Winstein relationship in the solvolysis of crowded tertiary alkyl chlorides. Hindered hydration and hydrophobic effect

Article abstract of DOI:10.1002/poc.358

Various highly crowded tertiary alkyl chlorides having a neopentyl or a (1-adamantyl)methyl substituent on the reaction center were subjected to solvolysis rate studies, and the Grunwald-Winstein (GW) type relationship with respect to the Y_Cl scale was examined. Analyses of the plots showed that these bulky substituents efficiently preclude the nucleophilic solvent participation from the rear side and that the data points for non-aqueous protic solvents give linear GW type plots. On the other hand, increased crowding causes considerable downward dispersions of the data points in aqueous mixtures of ethanol, acetone and 1-propanol. The magnitude of the downward dispersion increases in this order, giving a curvature with a downward bulge in the GW type relationship. Aqueous mixtures of the smallest alcohol, methanol, on the other hand, give only slight downward dispersions of the data points, which constitute a linear GW type plot. These results can be explained in terms of two causes. First, structural crowding makes the transition state of ionization less susceptible to the Bronsted base-type hydration to the β-hydrogens than 1-chloroadamantane as the standard of the Y_Cl scale. Second, with highly hydrophobic substrates the first solvation shell in aqueous ethanol is expected to become more ethanol rich than the bulk phase, causing less easy ionization of the substrate. The rate data can be semiquantitatively analyzed by using Hansch's hydrophobicity parameters. The present anomalies found in solvolysis reactions are regarded as a kinetic version of Wepster's observations of the solvent effects on the magnitude of Hammett σ constants of bulky alkyl groups. Copyright

Full text of DOI:10.1002/poc.358

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2001; 14: 229–238
The Grunwald±Winstein relationship in the solvolysis of
crowded tertiary alkyl chlorides. Hindered hydration and
hydrophobic effect
Ken'ichi Takeuchi,* Masaaki Takasuka, Eiji Shiba, Hironobu Tokunaga, Tadasuke Endo, Takuhiro
†
Ushino, Kazuhiko Tokunaga, Takao Okazaki, Tomomi Kinoshita, Yasushi Ohga
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan
Received 23 October 2000; revised 11 December 2000; accepted 13 December 2000
ABSTRACT: Various highly crowded tertiary alkyl chlorides having a neopentyl or a (1-adamantyl)methyl
epoc
substituent on the reaction center were subjected to solvolysis rate studies, and the Grunwald–Winstein (GW) type
relationship with respect to the YCl scale was examined. Analyses of the plots showed that these bulky substituents
efficiently preclude the nucleophilic solvent participation from the rear side and that the data points for non-aqueous
protic solvents give linear GW type plots. On the other hand, increased crowding causes considerable downward
dispersions of the data points in aqueous mixtures of ethanol, acetone and 1-propanol. The magnitude of the
downward dispersion increases in this order, giving a curvature with a downward bulge in the GW type relationship.
Aqueous mixtures of the smallest alcohol, methanol, on the other hand, give only slight downward dispersions of the
data points, which constitute a linear GW type plot. These results can be explained in terms of two causes. First,
structural crowding makes the transition state of ionization less susceptible to the Brønsted base-type hydration to the
b-hydrogens than 1-chloroadamantane as the standard of the YCl scale. Second, with highly hydrophobic substrates
the first solvation shell in aqueous ethanol is expected to become more ethanol rich than the bulk phase, causing less
easy ionization of the substrate. The rate data can be semiquantitatively analyzed by using Hansch’s hydrophobicity
parameters. The present anomalies found in solvolysis reactions are regarded as a kinetic version of Wepster’s
observations of the solvent effects on the magnitude of Hammett s constants of bulky alkyl groups. Copyright  2001
John Wiley & Sons, Ltd.
Additional material for this paper is available from the epoc website at http://www.wiley.com/epoc
KEYWORDS: solvolysis rate; tertiary alkyl chloride; steric congestion; solvent effect; Grunwald–Winstein
relationship; solvation; hydrophobic effect
INTRODUCTION
represented by the equation
Tertiary chloroalkanes constitute a family of popular
substrates in solvolysis studies. However, systematic
work related to the solvent effects on their solvolytic
reactivities, in particular on the Grunwald–Winstein
logk=ko ˆ mY
where k and k are the first-order rate constants of
o
1
1
solvolysis of a given compound in a given solvent and in
(
GW) relationship, has been started only recently by
2
3
the Liu’s and our groups. This paper presents a full
account of the solvent effects in the solvolysis of various
crowded tertiary chloroalkanes 3–9 in comparison with
2-chloro-2-methylpropane (2).
The GW relationship was originally proposed as a
linear free energy relationship in solvolytic reactions as
*
Correspondence to: K. Takeuchi, Department of Energy and
Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto
University, Sakyo-ku, Kyoto 606-8501, Japan.
E-mail: ktake@scl.kyoto-u.ac.jp
†
Present address: Department of Chemistry, Faculty of Engineering,
Oita University, Oita 870-1192, Japan.
Contract/grant sponsor: Japanese Ministry of Education, Science,
Sports and Culture; Contract/grant number: 10440188.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

2
30
K. TAKEUCHI ET AL.
4
8
0% EtOH–20% H O (v/v) at 25.0°C, respectively. As
Table 1. Rate constants of solvolysis for 1-chloroadaman-
2
tane (1) in TFE±MeOH, TFE±DMSO, and 1-propanol±H O at
the reference compound 2 was selected, and the ionizing
power of a solvent, Y, was defined by placing m = 1.00 for
2
25.0°C, and the corresponding YCl values
4
2
in eqn (1).
a
5
� 1
b
Solvent
k  10 (s
)
YCl
In 1970, Schleyer and co-workers noted that even 2
would be subject to nucleophilic solvent participation
NSP), and in 1982, Bentley and co-workers redefined
as the standard
chloride whose rear side is geometrically precluded from
NSP [Eqn. (2)]:
c,d
e,f
8
6
9
0T20M
0T40M
0T10D
80T20D
70P30W
0P40W
0.0357
0.0041
0.0413
0.00385
0.0026
0.0073
1.65
0.71
1.71
0.68
0.51
0.96
5
a
(
c,g
5
b,c
c,h
YCl by using 1-chloroadamantane (1)
e,i
e,j
6
a
M, T, D, P and W denote methanol, 2,2,2-trifluoroethanol, dimethyl
logk=k ˆ mY
2
o
Cl
sulfoxide, 1-propanol, and water, respectively. The numbers mean
volume� % of each component at 25.0°C. The concentration of
�
3
substrate was 0.020 mol dm
.
They examined the GW type relationship of 2 against
YCl and found that the data points for water and aqueous
mixtures of organic solvents fall upward with respect to
those for fluorinated solvents, such as 2,2,2-trifluoroetha-
nol (TFE), 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) and
b
= 8 Â 10� 9
� 1
The YCl values were calculated by using k
0E20W and Eqn. (2) with m = 1.
Determined titrimetrically in a single run in the presence of
o
s
for
5b
8
c
�
3
0
.025 mol dm 2,6-lutidine within an experimental error Æ2%.
d
� 6 � 1 � 5 � 1
k = 7.54 Â 10
s
(50.0°C), 9.15 Â 10
s
(75.0°C); DH‡ =
� 1
� 1 � 1
2
2.3 kcal mol ; DS‡ = � 13.2 cal mol
K .
5
b
e
trifluoroacetic acid (TFA). Recently, we reported that
all the data points in many solvents examined for 2-
chloro-2,4,4-trimethylpentane (3) are well accommo-
dated by a single mYCl plot because the rear side of 3 is
Extrapolated from data determined titrimetrically in the presence of
�
3
0
.025 mol dm 2,6-lutidine at other temperatures.
f
� 6 � 1
� 5 � 1
k = 1.05 Â 10
s
(50.0°C), 1.69 Â 10
s
(75.0°C); DH‡ =
(75.0°C); DH‡ =
(75.0°C); DH‡ =
(75.0°C), 1.50 Â
� 1
� 1 � 1
2
4.3 kcal mol ; DS‡ = � 11.0 cal mol
K .
g
� 5 � 1
1
� 4 � 1
k = 1.05 Â 10
s
(50.0°C), 1.34 Â 10
s
�
� 1 � 1
3
a,c,d
2
3.3 kcal mol ; DS‡ = � 9.6 cal mol
K
.
effectively shielded by the tert-butyl group.
extended the study to the highly crowded 4-chloro-
,2,4,6,6-pentamethylheptane (8) and found that the data
points of aqueous organic solvents come below those of
We
h
� 7 � 1
� 5 � 1
k = 9.63 Â 10
s
(50.0°C), 1.61 Â 10
s
� 1
� 1 � 1
2
4.3 kcal mol ; DS‡ = � 11.0 cal mol
K .
i
� 7 � 1
� 5 � 1
2
k = 7.19 Â 10
s
(50.0°C), 1.33 Â 10
s
�
4
� 1
� 1
1
0
s
(100.0°C);
DH‡ = 25.0 kcal mol
;
DS‡ = � 9.4
�
1
� 1
3
b
cal mol
K
.
non-aqueous solvents. The result was interpreted as
suggesting that the bulky tert-butyl groups inhibit not
only the NSP toward the cationic center, but also the
Brønsted base-type hydration toward the b-hydrogens in
j
� 6 � 1
� 5 � 1
k = 2.08 Â 10
s
(50.0°C), 3.71 Â 10
s
(75.0°C), 4.38 Â
� 1
�
4
� 1
10
cal mol
s
(100.0°C);
.
DH‡ = 25.0 kcal mol
;
DS‡ = � 7.2
�
1
� 1
K
3
b
the transition state of ionization. Further studies on the
solvolysis of 8 revealed that a downward bulge was
binary solvent systems containing TFE and MeOH, TFE
and DMSO, and 1-propanol and water were obtained by
determining the solvolysis rates of 1 in these solvents.
The results are summarized in Table 1.
3
d
evident for the aqueous ethanol data points. dThis
3
phenomenon was ascribed to hydrophobic effects.
This paper summarizes the accumulated rate data for
the solvolysis of 3–9 and some additional rate data for 1
and 2, and again emphasizes the important factors of the
Brønsted base-type hydration and hydrophobic effects as
the cause of dispersions of data points in the GW type
2-Chloro-2-methylpropane (2) and 2-chloro-2,4,4-
trimethylpentane (3). New rate constants of solvolysis
of 2 were obtained in some binary solvent systems
containing TFE and MeOH, TFE and DMSO, and 60%
TFE–40% EtOH (60T40E), and are summarized in Table
relationship with respect to YCl.
2. Table 2 also gives some newly determined rate
RESULTS
constants for 3 in 50% MeOH–50% water (50M50W)
and three TFE–EtOH mixtures.
Substrates
Other congested tertiaryl chloroalkanes (4±9). The
rates of solvolysis for 4–9 were determined in various
solvent systems and are summarized in Table 3. The rate7
of 8 in 80E20W has also been reported in the literature.
Eleven out of 25 rate constants for the solvolysis of 8 in
The chlorides except for 5, 7 and 9 were reported
recently. The chlorides 5, 7 and 9 were prepared from
the corresponding known alcohols by hydrochlorination
with HCl gas. The new precursor alcohol corresponding
to 9 was obtained by treating 1-(1-adamantyl)-4,4-
dimethyl-2-pentanone with methyllithium.
6
3b
Table 3 were reported previously.
Solvolysis products
Solvolysis rates
Previously, we reported that the methanolysis of 3 at
50°C gave the corresponding methyl ether and alkenes in
1
-Chloroadamantane (1). New YCl values for some
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

SOLVOLYSIS OF CROWDED TERTIARY ALKYL CHLORIDES
231
Table 2. Newly determined rate constants of solvolysis for 2-
chloro-2-methylpropane (2) and 2-chloro-2,4,4-trimethyl-
pentane (3) in TFE±MeOH, TFE±DMSO, 50% MeOH or
TFE±EtOH at 25.0°C
out that all the data points for non-aqueous solvents with
the exception of formic acid are well accommodated by a
single straight line with m = 0.58 (r = 0.9967) (Fig. 1).
3d
The magnitude of this value is considerably smaller than
the m values (0.73–0.77) for the other crowded
compounds 3–9 employed in this work. This can be
ascribed to the greater NSP of ethanol and methanol than
the fluorinated solvents. It should also be noted that the
methanol data point deviates slightly upwards (Fig. 1).
The non-aqueous solvent systems, which we concluded
could be accommodated by the single linear relation,
include 97HFIP, TFA, TFE, TFE–EtOH, TFE–MeOH,
TFE–DMSO, AcOH, MeOH and EtOH. In this respect,
the marked upward deviation of the formic acid data
point suggests a significant contribution to cation
stabilization, which deserves careful examination
5
� 1
k  10 (s
)
a
Solvent
2
3
b,c
8
6
9
8
0T20M
0T40M
0T10D
0T20D
2.11
—
—
—
—
3520
52.1
19.9
b,d
0.595
2.12
b,e
b,f
0.39g1
0.40g8
—
h
50M50W
b
b
b
6
5
4
0T40E
0T50E
0T60E
b
7.45
a
E, M, T, D and W denote ethanol, methanol, 2,2,2-trifluoroethanol,
dimethyl sulfoxide and water, respectively. The numbers mean volume
�
% of each component at 25.0°C. The concentration of substrate was
�
3
� 4
� 3
The upward deviations of the data points for aqueous
0
.020 mol dm
in titrimetric runs or (2–3) Â 10 mol dm
in
solvent mixtures are obviously caused by the very fast
conductimetric runs.
b
5b,c
Determined titrimetrically in a single run in the presence of
rate of 2 in water (Y = 4.57 ). Clearly, the problem is
Cl
�
3
0
.025 mol dm 2,6-lutidine within an experimental error Æ2%.
c
� 4 � 1
� 1
focused on the question of why water accelerates
dramatically the solvolysis of 2 as compared with 1
(see below). It should also be pointed out that the aqueous
methanol data points are linear whereas the aqueous
ethanol and aqueous acetone lines show a slight upward
bulge.
k = 3.92 Â 10
s
(50.0°C); DH‡ = 21.8 kcal mol ; DS‡ = � 6.9
�
1
� 1
cal mol
K .
d
� 4 � 1
(50.0°C); DH‡ = 22.3 kcal mol� 1; DS‡ = � 7.7
(50.0°C); DH‡ = 22.2 kcal mol� 1; DS‡ = � 5.5
(50.0°C); DH‡ = 23.3 kcal mol� 1; DS‡ = � 5.0
k = 1.18 Â 10
s
�
1
� 1
cal mol
K .
e
� 4 � 1
k = 4.16 Â 10
s
�
1
� 1
cal mol
K .
f
� 5 � 1
k = 8.90 Â 10
s
�
1
� 1
cal mol
K .
g
In contrast, all the data points for the solvolysis of 3 in
both aqueous and non-aqueous solvents are well
accommodated by a single straight line (Fig. 1). Liu
and co-workers reported that 3-chloro-3-isopropylpen-
tane and 3-tert-butyl-3-chloropentane behave similarly to
See Table S1 in Supplementary Material.
Determined conductimetrically in the absence of a buffer within an
h
experimental error of Æ1%.
3
c
2
2
4
7
6 and 74% yields, respectively. In the methanolysis of
3 in the GW relationship. Phenomenologically, this
the yields of the methyl ether and alkenes were 30 and
0%, respectively, at 25°C (Scheme 1). With increase in
would have been caused by the close approach of the
water data point to the non-aqueous line. Most of the
solvolysis rates used for the plots in Fig. 1 are reported
ones; all the data are summarized in Table S1 in the
Supplementary Material.
The linear behavior of the GW relationship for 3
suggested that compounds 4–9 having greater congestion
would also behave similarly. Figures 2 and 3 show the
GW plots for 4–6 and 7–9, respectively. There are a few
new phenomena that have not been noted before in this
field, as follows.
crowding, the yield of methyl ether decreased; 6 (at
0°C) and 8 (at 25°C) gave the corresponding methyl
5
ether in 12 and 6% yields, respectively. In the less
nucleophilic solvent TFE the substitution product was
formed in 1% or less yield from 6 and 8.
DISCUSSION
The Grunwald±Winstein (GW) type relationship
(1) In the GW plots of 4, 5, 6 (Fig. 2) and 7 (Fig. 3), the
aqueous methanol data points are nearly accommodated
by the linear correlation for non-aqueous solvents. Since
the aqueous methanol lines are essentially linear, the
water data point would fall on the non-aqueous line. On
the other hand, the aqueous ethanol and aqueous acetone
data points fall below the non-aqueous line. Obviously,
these aqueous solvent mixtures show a curvature with a
downward bulge.
As mentioned above, Bentley and Carter found that the
GW type plot for 2-chloro-2-methylpropane (2) against
YCl showed dispersions of the data points for fluorinated
solvents below the aqueous ethanol, aqueous methanol
5
b
and aqueous acetone data points. The faster rates of
solvolysis of 2 in aqueous organic solvents such as
EtOH–H O, MeOH–H O, TFE–H O and acetone–H O
2
2
2
2
than expected from fluorinated solvent data points were
ascribed to nucleophilic assistance to ionization in
aqueous solvent mixtures.
We carefully examined the data points for fluorinated
solvents and further added some data points for the
solvents as shown in Table 2. In consequence, it turned
(2) In the highly crowded compounds 8 and 9, the
deviation of the aqueous ethanol, aqueous acetone and
two aqueous 1-propanol data points becomes enormous.
In these compounds, the hypothetical water data points
are supposed to fall below the non-aqueous line, since the
aqueous methanol lines are almost linear. This is also
5
b
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

2
32
K. TAKEUCHI ET AL.
Table 3. Rate constants of solvolysis of congested teritiary choloroalkanes
5
� 1
1
0 k (s ) at 25.0°C
a
b
Solvent
00E
YCl
4
5
6
7
8
9
c,d
c
c
c
c
c
1
� 2.5
� 0.9
0.800
7.78
32.0
1.17
1.26
3.92
21.7
156
527
37.7
233
708
1760
4190
c
c
c
c
e
e
9
8
7
6
5
4
1
9
8
7
6
7
5
4
0E10W
0E20W
0E30W
0E40W
0E50W
0E60W
00M
0M10W
0M20W
0M30W
0M40W
0A30W
0A50W
0A60W
10.1c
11.4c
31.8
c
e
e,f
e
e
e
e
0.00
40.0
46.5
113e
e
e
e
0.8
1.38
2.02
2.75
� 1.2
� 0.2
0.67
1.46
2.07
0.17
1.73
2.46
� 1.6
(2.83)
2.83
2.96
3.16
1.89
0.63
0.16
� 0.48
� 1.42
1.65
0.71t
1.71t
0.68
93.7
124e
144e
318e
1070
1500e
e
245
349
481e
4280
1600
9170
g
c
c
c
c
e
e
e
e
e
7.64
11.9e
13.9e
39.3
202e
370
1560
5280
e
e
39.6
61.3
70.7
174e
955e
e
e
e
e
164
568
207e
319e
703
3910
e
e,h
879
2180
1130
2830e
13900
21200
e
10500
e
c
e
e
e
e
e
34.7
51.0
55.4
135e
596e
6140
684
8280
e
e
e
632
1040
1210
5130
1910
g
i
AcOH
134
283000
j
k
k
k
k,l
k,m,n
1
9
7
5
8
6
5
4
2
8
6
9
8
7
6
00T
7390k
18800
13300
15100
7T3W
0T30W
0T50W
0T20E
0T40E
0T50E
0T60E
0T80E
0T20M
0T40M
0T10D
0T20D
0P30W
0P40W
8010
k
k,o
k,m,p
9800k
18700
19600
2220
285
227000
k,m,q
11100
r
k
k
k
k
k
k
k,m,s
k
922
1630
209
6750
43000
5100
1900
691
r
k
k
k
k
k
130
779
10100
3540
1410
r
k
k
k
k
46.9
22.0
81.9
299
r
r
t
t
k
k
k
112
k,m,u
k,v
k,w
k,x
g
36600
5920
42800
5100
431
t
0.51t
g
0.96
855
a
E, M, A, T, D, P, and W denote ethanol, methanol, acetone, 2,2,2-trifluoroethanol, dimethyl sulfoxide, 1-propanol and water, respectively. The
numbers mean volume � % of each component at 25.0°C, except for the T–W system, which is based on weight � %. The concentration of substrate
�
3
� 4
� 3
was 0.020 mol dm in titrimetric runs or (2–3) Â 10 mol dm in conductimetric runs.
b
Quoted from Ref. 5c unless noted otherwise.
c
d
e
f
� 3
Determined titrimetrically in a single run in the presence of 0.025 mol dm 2,6-lutidine within an experimental error Æ2% unless noted otherwise .
� 4 � 1
� 1
� 1 � 1
k = 2.05 Â 10
s
(50.0°C); DH‡ = 24.2 kcal mol ; DS‡ = � 0.5 cal mol
K
.
� 3
Determined conductimetrically in duplicate in the presence of 0.025 mol dm 2,6-lutidine within an experimental error of Æ1%.
� 3 � 1 7
A reported value is 5.25 Â 10
s .
g
h
i
Determined conductimetrically in duplicate in the absence of 2,6-lutidine within an experimental error of Æ1%.
� 1
� 1
� 1
� 1
� 1 � 1
k = 0.0734 s (18.38°C), 0.109 s (22.5°C), 0.136 s (24.7°C); DH‡ = 16.2 kcal mol ; DS‡ = � 8.1 cal mol
K .
� 3
Determined in the presence of 0.025 mol dm NaOAc.
j
The value of 97T3W was assumed.
k
l
� 3
Determined conductimetrically in duplicate in the presence of 0.00125 mol dm 2,6-lutidine within an experimental error of Æ1%.
� 1
� 1
� 1
� 1 � 1
k = 0.0481 s (9.7°C), 0.111 s (20.5°C); DH‡ = 12.0 kcal mol ; DS‡ = � 22.1 cal mol
K
.
m
n
o
p
q
r
Extrapolated from data at other temperatures.
� 1
� 1
� 1
� 1
� 1 � 1
k = 0.0547 s (� 20.0°C), 0.124 s (� 12.0°C), 0.351 s (� 0.6°C); DH‡ = 12.5 kcal mol ; DS‡ = � 14.4 cal mol
K .
� 1
� 1
� 1
� 1 � 1
� 1
k = 0.0459 s (12.1°C), 0.111 s (20.0°C); DH‡ = 17.8 kcal mol ; DS‡ = � 2.3 cal mol
K
.
� 1
� 1
� 1
� 1
� 1
� 1 � 1
k = 0.0575 s (� 10.5°C), 0.108 s (� 5.0°C), 0.180 s (� 0.6°C); DH‡ = 15.5 kcal mol ; DS‡ = � 4.8 cal mol
K .
.
� 1
� 1
� 1 � 1
k = 0.0454 s (9.0°C), 0.0753 s (14.0°C), 0.120 s (19.5°C); DH‡ = 14.6 kcal mol ; DS‡ = � 12.9 cal mol
K
Ref.13a.
s
k = 0.0537 s (5.0°C), 0.0919 s (10.0°C), 0.204 s (17.5°C); DH‡ = 16.6 kcal mol� 1; DS‡ = � 4.6 cal mol� 1
� 1
� 1
� 1
� 1
K
.
t
See Table 1.
u
v
w
x
� 1
� 1
� 1
� 1
� 1 � 1
k = 0.0790 s (9.5°C), 0.130 s (14.3°C), 0.216 s (19.5°C); DH‡ = 15.9 kcal mol ; DS‡ = � 7.1 cal mol
K .
� 1
� 1
� 1
� 1
� 1 � 1
� 1 � 1
k = 0.0288 s (18.3°C), 0.0334 s (19.7°C), 0.0423 s (22.0°C); DH‡ = 17.9 kcal mol ; DS‡ = � 4.0 cal mol
K .
� 1
� 1
� 1
k = 0.0973 s (10.5°C), 0.215 s (18.1°C); DH‡ = 16.6 kcal mol ; DS‡ = � 4.6 cal mol
K .
� 1
� 1
� 1
� 1
� 1
� 1
k = 0.0272 s (19.32°C), 0.0328 s (20.95°C), 0.0406 s (22.91°C), 0.0594 s (26.40°C), 0.255 s (40.00°C); DH‡ = 19.1 kcal mol
;
�
1
� 1
DS‡ = � 0.3 cal mol
K .
supported by the slower rate of 8 in 70T than in 100T
highly alkyl-substituted substrates. It is suggested that 8
and 9 may be less nucleophilically assisted than the
standard 1-chloroadamantane (1). We have found that 1-
bromo-3,5,7-triisopropyladamantane (10) also shows
5
b,c
despite the greater Y value of the former (2.96)
the latter (2.83).
than
Cl
5
b,c
The above two features appear to be characteristic of
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

SOLVOLYSIS OF CROWDED TERTIARY ALKYL CHLORIDES
233
Scheme 1
3d
very similar behavior in the GW plot. We ascribe the
downward positioning of the water data point below the
non-aqueous solvent line and the bulge found for aqueous
ethanol and aqueous acetone data points to different
causes as discussed below.
Steric hindrance to Brùnsted-base type hydration
1-Chloroadamantane (1) cannot be attacked by a
nucleophile from the rear side. Therefore, the result that
1
9
is more susceptible to solvent nucleophilicity than 8 or
may be interpreted as a sign that 1 is more strongly
solvated than 8 or 9 on the cation side in the transition
state of ionization. In the early work by Bentley, and co-
workers, they postulated that the NSP as indicated by
using Eqn. (2) is considered to involve the interactions
with both the reacting carbon atom and the hydrogen
8
atoms, in particular the b-hydrogens. Recently, Richard
et al. pointed out that Brønsted base-type solvation to b-
Figure 1. Grunwald±Winstein plots with respect to YCl for
the solvolysis of 2 and 3 at 25°C. The points for 3 are shifted
upward by 2 units for clarity. The m value for 2 in non-
aqueous solvents is 0.58 (r = 0.9967) and that for 3 in all
solvents is 0.75 (r = 0.9977)
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

2
34
K. TAKEUCHI ET AL.
Figure 3. Grunwald±Winstein plots with respect to YCl for
the solvolysis of 7, 8 and 9 at 25°C. The points for 8 and 9
are shifted upward by 2 and 4 units, respectively, for clarity.
The m values for 7, 8 and 9 in non-aqueous solvents are 0.73
(r = 0.9994), 0.77 (r = 0.9987) and 0.76 (r = 0.9992), respec-
tively
Figure 2. Grunwald±Winstein plots with respect to YCl for
the solvolysis of 4, 5 and 6 at 25°C. The points for 5 and 6
are shifted upward by 2 and 4 units, respectively, for clarity.
The m values for 4, 5 and 6 in non-aqueous solvents are 0.74
(r = 0.9976), 0.76 (r = 0.9959) and 0.75 (r = 0.9995), respec-
tively
hydrogens would be more important than nucleophilic
assistance toward carbenium carbon in the ionization step
use. Several factors could be behind the behavior of the
solvents to give curvatures, including hydrophobic
effects, changing cluster structure, possibility of solvent
sorting or formation of water pools, difference in ground-
state solvation between substrates, and even mechanistic
differences between the crowded alkyl systems and
standard 1-chloroadamantane (1). Among these possibi-
lities, we prefer the hydrophobic effect (see below).
The marked curvatures of the plots of aqueous ethanol
data (Figs. 2 and 3) stem from significant deviations from
9
of cumyl substrates. According to Monte Carlo and
‡
�
RISM calculations on the hydration of a t-Bu Cl ion
pair, the average primary hydration number is about 1 for
carbenium carbon, but it amounts to 3 for each methyl
1
0
group. The latter type of solvation to b-hydrogens in the
transition state of ionization of 8 or 9 should be severely
blocked owing to the presence of bulkyl tert-butyl or
adamantyl groups. The downward positioning of the
water points below the non-aqueous line may be ascribed
to greater steric hindrance of hydration toward b-
hydrogens in 8 or 9 than in 1.
linearity in the log k vs water mole fraction (f ) plots
H O
2
[Fig. 4(a)]. More comprehensible plots are given in Fig.
4(b), where the deviations (Dlog k) in Fig. 4(a) are plotted
against f . The deviation increases in the order
H O
2
2
< 1 ꢀ 3 < 6 < 8 < 9.
Dependence of curvature on substrate and sol-
vent
We examined similar plots for aqueous methanol and
TFE–EtOH systems, but the former showed only slight
deviations of <0.2 log k unit, and the latter gave an
essentially linear plot of log k vs. TFE mole fraction as
shown in Fig. 5 for 1, 2, 3 and 8.
Marked deviations with a curved shape in the GW type
plot, which are observed for the solvolysis of crowded
substrates in aqueous ethanol, aqueous acetone and
aqueous 1-propanol (Figs. 2 and 3), call for caution and
place a limitation on the use of these solvent systems.
Aqueous methanol is a good solvent system in this
respect, but poor solubilities of substrates restrict its wide
Hydrophobic effects
The deviation in the linear free energy relationship
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

SOLVOLYSIS OF CROWDED TERTIARY ALKYL CHLORIDES
235
Figure 4. (a) Plots of log k against mole fraction of water in the solvolysis of 1, 2, 3, 6, 8 and 9 in EtOH±H O at 25°C. The rate
2
constants in water for 1 and 2 were taken from Refs 5b and 4, respectively, and those for 3, 6, 8 and 9 were estimated by
extrapolation of the MeOH±H O data points to Y of water (4.57) in Figs 1±3. (b) Plots of deviations of log k from a straight line
2
in Fig. 4(a) against mole fraction of water
Cl
P
caused by solvent effects has been reported in the
Hammett rule. Hoefnagel and Wepster found that the
Hammett s constants of bulky alkyl groups, such as tert-
p. For example,
considering that the three hydrogens of 2 were replaced
by a tert-butyl (p = 1.98) and two methyl (p = 0.56)
p of 6 was evaluated as 3.10 by
12
12
P
butyl and 1,1-diethylpropyl (Et C), become significantly
groups. Table 4 summarizes the p, log(k70E/k60T40E),
and log(k80M/k60T40E) values, and Fig. 6 shows the
relationships between the two logarithmic rate ratios and
3
more negative (less ionizing) in aqueous ethanol or
aqueous tert-butyl alcohol than in water, and the results
P
1
1
were ascribed to the hydrophobic effect. They applied
p.
1
2
Hansch’s hydrophobic constant (p) for substituents to
A comparison of the slopes, � 0.25 for 70E and � 0.09
for 80M, indicates that the former solvent system is about
2–3 times more sensitive to the hydrophobicity of the
substrate. The upward deviations of the values for 2 from
the correlation lines in Fig. 6 would be ascribed to NSP
improve the Hammett relationship in aqueous organic
1
1
solvents. The present results in solvolysis reactions
constitute a kinetic version of Hoefnagel and Wepster’s
observations. We assume that the more alkylated the
substrate, the more water molecules are expelled from the
solvation shell. Presumably, the first solvation shell
becomes more ethanol-rich than the bulk phase. This
would lead to less easy ionization of the substrate.
We wished to examine semiquantitatively the sensi-
tivity of hydrophobic effect in aqueous methanol and
aqueous ethanol. Since the extent of downward deviation
from the non-aqueous line indicates the magnitude of the
hydrophobic effect, we examined the rates in 70E30W
and 80M20W relative to that in 60T40E as a standard for
each substrate. These solvent systems have similar
5
b,c
5b,c
13a
respective YCl values, 0.8,
0.67
and 0.63 (Table
3). Therefore, the quantities defined by log(k70E/k60T40E
)
and log(k80M/k60T40E) may be regarded as a parameter of
the hydrophobic effect exerted by 70E and 80M at a YCl
value of approximately 0.7. The magnitude of the
hydrophobic effect with respect to the substrate structure
may be evaluated by making a plot of log(k70E/k60T40E
)
and log(k8 /k60T40E) vs relative Hansch’s hydrophobic
0M
P
constants ( p). We selected 2-chloro-2-methylpropane
P
P
(2) as a standard ( p = 0), and the p values of other
Figure 5. Plots of log k against mole fraction of TFE in the
solvolysis of 1, 2, 3 and 8 in TFE±EtOH at 25°C
compounds were calculated by assuming the additivity of
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

2
36
K. TAKEUCHI ET AL.
P
Table 4. Total p ( p) relative to 2 and logarithmic rate ratios
between the rate in 70% EtOH or 80% MeOH and that in
6
0%TFE±40% EtOH
P
1.98
3.30
3.00
3.10
3.51
3.96
5.28
a
b
b
Compound
p
Log(k70E/k60T40E
)
Log(k80M/k60T40E)
2
3
4
5
6
7
8
9
0
0.94
0.10
0.74
0.18
0.10
0.00
0.05
� 0.05
� 0.12
� 0.28
� 0.14
� 0.23
� 0.30
� 0.39
� 0.54
� 0.76
a
Relative to 2. For p, see Ref. 12.
For rate data, see Tables 2, 3 and S1.
b
including the Brønsted base-type solvation. Recently,
Tada et al. reported the hydrophobic effects on the
Figure 6. Plots of log(k70E/k60T40E) (*) and log(k80M/k60T40E
)
P
(
*) values against p for the solvolysis of 2±9 at 25 °C
solvatochromism (E ) in various aqueous organic
T
1
4
solvents.
EXPERIMENTAL
1
13
Boiling and melting points are uncorrected. H and
C
NMR spectra were recorded in CDCl3 at 270 and
8 MHz, respectively. Medium-pressure liquid chroma-
tography (MPLC) was conducted on Merck silica gel 60
230–400 mesh). The known alcohols corresponding to 5
6
Possibilities of differential ion-pair return or
mechanistic change
(
and 7 were prepared by treating 4,4-dimethyl-2-penta-
none (Aldrich) with propylmagnesium bromide and
isobutyllithium, respectively. Most of the solvolysis
1
3b
Recently, Kevill and D’Souza
applied Eqn. (3) to a
limited number of rate data for 8 that we reported in
997:
3
b
18
1
solvents were purified as described previously. 1-
Propanol was distilled over CaH . DMSO was distilled
2
over CaH under reduced pressure. HFIP was dried over
2
logk=k ˆ mY ‡ hI
3
o
Cl
molecular sieves 3A and distilled. Solvent mixtures based
on volume-% were prepared by mixing weighed
19
Equation (3) was originally developed to correct the
charge delocalization to an aromatic ring, but it was
also shown to be usable for correction of ion-pair
return. Compounds 8 and 9 mainly give alkenes as
solvolysis product. This means that the leaving group,
once ionized, mainly attacks one of the neighboring
hydrogen atoms, and the fraction of ion-pair return would
be very small. On the basis of the fairly good fit of the
plot, they suggested a possibility of reduced ion-pair
components by using the following densities at 25°C;
water (0.99705), acetone (0.78440), ethanol (0.78493),
methanol (0.78637), TFE (1.3686, an extrapolated value
from 1.4106 and 1.3736 at 0 and 22°C, respectively) and
DMSO (1.09537).
1
5
1
6
1-(1-Adamantyl)-2,4,4-trimethyl-2-pentanol. 1-(1-Ada-
20
mantyl)-4,4-dimethyl-2-pentanone (1.59 g, 6.39 mmol)
was treated with methyllithium (9.54 mmol) in diethyl
ether. Usual work-up of the reaction mixture followed by
1
3b
return.
against Y
However, the behavior of 10 in the GW plot
parameters, which is very similar to that of
and 9, lessens the possibility of reduced ion-pair
5
b,c
MPLC [SiO , hexane–diethyl ether (97:3)] afforded the
Br
3
2
d
8
alcohol (1.63 g) in 96% yield. Colorless crystals, m.p.
return as a cause of deviations in the GW relationship in
Figs 2 and 3.
55.5–56.5°C. Found: C, 81.63; H, 12.41. C H O
18 32
requires C, 81.75; H, 12.20%. ꢀH, 1.03 (9H, s), 1.21
(1H, s), 1.36 (3H, s), 1.34 (1H, d, J 14.8 Hz), 1.39 (1H, d,
J 14.8 Hz), 1.49 (1H, d, J 14.8 Hz), 1.53 (1H, d, J
14.8 Hz), 1.67–1.70 (12H, m), 1.89–1.98 (3H, m); ꢀC,
28.9 (CH), 29.7 (CH ), 31.6 (C), 31.8 (CH ), 34.0 (C),
It was also suggested by Bentley et al. that the
unexpectedly slow rates of solvolysis of congested alkyl
compounds in aqueous ethanol might be caused by
1
7
mechanistic changes to concerted elimination. How-
ever, this possibility would be low since 10, which cannot
undergo the elimination reaction, also shows marked
deviations in the GW relationship similarly to the
3
3
37.1 (CH ), 44.2 (CH ), 57.1 (CH ), 58.7 (CH ), 76.4 (C).
2
2
2
2
General procedure for the preparation of 5 and 7. The
corresponding known alcohols (ca 1.5 g) were treated
3
d
congested substrates used in the present study.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

SOLVOLYSIS OF CROWDED TERTIARY ALKYL CHLORIDES
237
with dry HCl gas in pentane at � 40°C for 5–10 min. The
tween alkenes and substitution product was determined
by GLC. The methoxyl H signal appeared at ꢀH 3.12.
1
solution was dried with a small amount of CaCl , flushed
2
with N and filtered. Pentane was evaporated and the
Detailed analysis of alkenes was not carried out.
2
chloride was distilled under reduced pressure. The
chlorides were used for rate studies without being
subjected to microanalysis because of their instability.
Methanolysis and tri¯uoroethanolysis of 4-chloro-
2,2,4,6,6-pentamethylheptane (8). The product ratios
were determined by H NMR. The methyl ether showed
1
4
-Chloro-2,2,4-trimethylheptane (5). B.p. 42.0–42.5°C
at 6 mmHg. ꢀH, 0.93 (3H, t, J 7.3 Hz), 1.05 (9H, s), 1.45–
.60 (2H, m), 1.63 (3H, s), 1.74–1.82 (2H, m), 1.85 (1H,
d, J 15.2 Hz), 1.89 (1H, d, J 15.2 Hz); ꢀC, 14.2 (CH3),
the methoxyl signal at ꢀH 3.12. No indication for the
formation of the trifluoroethyl ether was found. The ratio
between the endo- (presumably a mixture of Z- and E-
forms) and exo-alkenes was determined based on the
respective signals at ꢀH 5.11 (br s) and 4.79 (s).
1
1
8.2 (CH ), 31.5 (4 Â CH ), 32.3 (C), 48.8 (CH ), 55.7
2
3
2
(CH ), 75.8 (C).
2
Kinetic Methods. The titrimetric and conductimetric rate
18
4-Chloro-2,2,4,6-tetramethylheptane (7). B.p. 43.0°C
constants were determined as described previously.
at 4 mmHg. ꢀH, 1.00 (6H, d, J 6.6 Hz), 1.06 (9H, s), 1.65
3H, s), 1.68 (1H, dd, J 5.6, 14.8 Hz), 1.78 (1H, d, J
(
1
1
3
7
5.2 Hz), 1.79 (1H, dd, J 5.6, 14.8 Hz), 1.91 (1H, d, J
5.2 Hz), 1.89–2.00 (1H, m); ꢀC, 25.0 (CH), 25.1 (CH3),
1.4 (CH ), 31.7 (CH ), 32.3 (C), 55.4 (CH ), 56.7 (CH ),
Supplementary Material
A table of the rate constants for 2 and 3 used for the plots
3
3
2
2
1
5.8 (C).
in Fig. 1 and H NMR spectra for 9 are available at the
epoc website at http://www.wiley.com/epoc.
1-(2-Chloro-2,4,4-trimethylpentyl)adamantane
(9).
The precursor alcohol (414 mg, 1.57 mmol) in CH Cl2
2
was treated with dry HCl gas at � 40°C for 2 min. The
reaction mixture was treated with a small amount of
CaCl , flushed with N , and filtered. Evaporation of the
Acknowledgements
2
2
Fruitful discussions with Professor Emeritus John Shorter
are gratefully acknowledged. Support for this work was
partially provided by a Grant-in-Aid for Scientific
Research (10440188) from the Japanese Ministry of
Education, Science, Sports and Culture.
1
solvent afforded a yellow liquid, which was found by H
1
3
and C NMR to contain three kinds of alkenes and 9
showing the chlorinated carbon at 76.0 ppm. The NMR
spectra showed that 9 was present in about 40%, and no
indication was found for the formation of other chloride
isomers. All of the four tert-butyl signals and all the
1
olefinic protons in H NMR spectra were unambiguously
assigned (see Supplementary Material). The mixture was
used for rate studies without further purification.
REFERENCES
1
2
. Streitwieser A Jr. Solvolytic Displacement Reactions. McGraw-
Hill: New York, 1962.
. (a) Liu KT, Sheu HC, Chen HI, Chiu PF, Hu CR. Tetrahedron Lett.
1990; 31: 3611; (b) Liu KT, Hou SJ, Tsao ML. J. Org. Chem.
General procedures for product studies. The solvolysis
�
3
was conducted for a solution of 0.04 mol dm substrate
�
3
1
9
998; 63: 1360; (c) Liu KT, Tsao ML. J. Org. Chem. 1998; 63:
in the presence of 0.05 mol dm 2,6-lutidine for 10 half-
lives at appropriate temperatures. The reaction solution
was mixed with saturated aqueous NaCl and successively
extracted with pentane. The combined pentane layer was
washed with saturated aqueous NaCl three times and
552.
3
. (a) Takeuchi K, Ohga Y, Ushino T, Takasuka M. J. Phys. Org.
Chem. 1996; 9: 777; (b) Takeuchi K, Ohga Y, Ushino T, Takasuka
M. J. Org. Chem. 1997; 62: 4904; (c) Takeuchi K, Ohga Y, Ushino
T, Takasuka M. J. Phys. Org. Chem. 1997; 10: 717; (d) Takeuchi
K. Pure Appl. Chem. 1998; 70: 2025.
dried (MgSO ). The product distributions were deter-
4. Grunwald E, Winstein S. J. Am. Chem. Soc. 1948; 70: 846.
4
5
. (a) Raber DJ, Bingham RC, Harris JM, Fry JL, Schleyer PvR. J.
Am. Chem. Soc. 1970; 92: 5977; (b) Bentley TW, Carter GE. J.
Am. Chem. Soc. 1982; 104: 5741; (c) Bentley TW, Llewellyn G.
Prog. Phys. Org. Chem. 1990; 17: 121.
mined by GLC (PEG 20M, 2m  3 mmf id) for the
1
extract, or by H NMR for the crude product after
distillation of pentane on a 30 cm Vigreux column.
6
. Takeuchi K, Takasuka M, Shiba E, Kinoshita T, Okazaki T,
Abboud JLM, Notario R, Casta n˜ o O. J. Am. Chem. Soc. 2000; 122:
Methanolysis of 1-(2-chloromethylpropyl)adamantane
4). GLC analyses showed the ratio between alkenes and
7
351.
(
7. Brown HC, Berneis HL. J. Am. Chem. Soc. 1953; 75: 10.
1
8
9
. (a) Bentley TW, Schleyer PvR. J. Am. Chem. Soc. 1976; 98: 7658;
b) Bentley TW, Bowen CT. J. Chem. Soc., Perkin Trans. 2 1978;
57; (c) Bentley TW, Bowen CT, Parker W, Watt CIF. J. Chem.
Soc., Perkin Trans. 2 1980; 1244.
. Richard JP, Jagannadham V, Amyes TL, Mishima M, Tsuno Y. J.
Am. Chem. Soc. 1994; 116: 6706.
the methyl ether to be 70:30. H NMR analysis showed
the formation of the endo-alkene [ꢀH 4.84 (br s)] and the
exo-alkene [ꢀH 4.50 (m), 4.76 (m)] in a ratio of 5:65.
(
5
Methanolysis and tri¯uoroethanolysis of 4-chloro-4-
ethyl-2,2-dimethylhexane (6). The product ratio be-
1
0. Jorgensen WL, Buckner JK, Huston SE, Rossky PJ. J. Am. Chem.
Soc. 1987; 109: 1891.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238

2
1
1
1
1
1
38
K. TAKEUCHI ET AL.
1. Hoefnagel AJ, Wepster BM. J. Chem. Soc., Perkin Trans. 2 1989;
77.
2. Hansch C, Leo A, Unger SH, Kim KH, Nikaitani D, Lien EJ. J.
Med. Chem. 1973; 16: 1207.
3. (a) Kevill DN, D’Souza MJ. J. Chem. Res. (S) 1993; 174; (b)
Kevill DN, D’Souza MJ. Tetrahedron Lett. 1998; 39: 3973.
4. Tada EB, Novaki LP, El-Seoud OA. J. Phys. Org. Chem. 2000; 13:
16. Kevill DN, Ismail NHJ, D’Souza MJ. J. Org. Chem. 1994; 59:
9
6303.
1
1
7. Bentley TW, Llewellyn G, Ryu ZH. J. Org. Chem. 1998; 63: 4654.
8. Takeuchi K, Ikai K, Shibata T, Tsugeno A. J. Org. Chem. 1988; 53:
2
852.
1
9. Riddick JA, Bunger WB, Sakano TK. Organic Solvents (4th edn).
Wiley: New York, 1986.
20. Reetz MT, Maier WF. Chem. Ber. 1980; 113: 3741.
6
79.
5. Kevill DN, Anderson SW. J. Org. Chem. 1991; 56: 1845.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 229–238