Correlation of the rates of solvolysis of the benzhydryldimethylsulfonium ion. Application of the aromatic ring parameter

Article abstract of DOI:10.1021/jo960909t

Values for the specific rates of solvolysis of the benzhydryldimethylsulfonium ion in 34 solvents have been analyzed using various forms of the extended Grunwald - Winstein equation. The specific rates are insensitive toward changes in solvent nucleophilicity (N_T) values, and they correlate best against a combination of Y⁺ values (based on the solvolyses of the 1-adamantyldimethylsulfonium ion) and aromatic ring parameter (I) values. Common-molecule return is observed, being especially powerful in solvents rich in fluoro alcohol; the logarithm of the associated mass law constant correlates inversely with the solvent NT values. The product selectivities in ethanol - water mixtures are also consistent with an S_N1 mechanism for the solvolyses.

Full text of DOI:10.1021/jo960909t

7256
J . Org. Chem. 1996, 61, 7256-7262
Cor r ela tion of th e Ra tes of Solvolysis of th e
Ben zh yd r yld im eth ylsu lfon iu m Ion . Ap p lica tion of th e Ar om a tic
Rin g P a r a m eter 1
Dennis N. Kevill,* Steven W. Anderson,† and Norsaadah HJ Ismail‡
Department of Chemistry, Northern Illinois University, DeKalb, Illinois 60115-2862
Received May 17, 1996X
Values for the specific rates of solvolysis of the benzhydryldimethylsulfonium ion in 34 solvents
have been analyzed using various forms of the extended Grunwald-Winstein equation. The specific
rates are insensitive toward changes in solvent nucleophilicity (NT) values, and they correlate best
against a combination of Y+ values (based on the solvolyses of the 1-adamantyldimethylsulfonium
ion) and aromatic ring parameter (I) values. Common-molecule return is observed, being especially
powerful in solvents rich in fluoro alcohol; the logarithm of the associated mass law constant
correlates inversely with the solvent NT values. The product selectivities in ethanol-water mixtures
are also consistent with an SN1 mechanism for the solvolyses.
Recently we developed,2 from a consideration of the
values;7 h is the sensitivity to changes in aromatic ring
parameter (I) values;2 and c is a constant (residual)
quantity.
differences in response to solvent variation of the specific
rates of solvolysis of the (p-methoxybenzyldimethyl)-
sulfonium ion and the 1-adamantyldimethylsulfonium
ion (Y+ values3 ), a scale of aromatic ring parameter (I)
values for the treatment of the dispersion involved in
Grunwald-Winstein plots of the specific rates of solvoly-
sis of benzylic derivatives proceeding with, at the transi-
tion state, appreciable development of positive charge.
The one-term Grunwald-Winstein equation (1)4 is thus
expanded to eq 2. When solvent nucleophilicity changes
Specific rates of solvolysis of benzylic,8,9 naphthyl-
methyl,10,11 anthrylmethyl,12 and benzhydryl10,13 deriva-
tives with anionic leaving groups (RX substrates) have
been successfully correlated using eq 2 or, occasionally,
eq 3.2,14-16 Even allylic and propargylic substrates9 show
improved correlations of their specific rates of solvolysis
when the hI term is incorporated, but with reduced h
values relative to benzylic substrates.2 Recently, support
for the incorporation of the hI term, based on molecular
orbital studies using the AM1 method, has been put
forward.17a However, there have been very few studies
of the effect of solvent variation on the solvolyses of
R-aryl-substituted derivatives with neutral molecule
leaving groups (RX+ substrates). Accordingly, we have
carried out a study of the solvolysis of the benzhy-
dryldimethylsulfonium ion (1), as the trifluoromethane-
sulfonate salt, in a wide variety of solvents.
log (k/ko)RX ) mYX + c
(1)
(2)
log (k/ko)RX ) mYX + hI + c
also partially control the variations of specific rates with
solvent, an lNT term can be added5,6 to give eq 3. In eqs
log (k/ko)RX ) lNT + mYX + hI + c
(3)
Me
1-3, k and ko are the specific rates of solvolysis of RX in
the solvent under consideration and in the standard
solvent, 80% ethanol; l is the sensitivity to changes in
solvent nucleophilicity, best expressed as NT values;6 m
is the sensitivity to changes in solvent ionizing power (YX)
–
C
H
S
SO3CF3
Me
1
From a similarity model consideration,17b the unimo-
lecular solvolyses of RSMe2+-type substrates should give
good correlations when the scales applied are the I scale,2
† Present address: Department of Chemistry, University of Wis-
consinsWhitewater, Whitewater, WI 53190-1790.
‡ Present address: Research and Consultancy Center, Institut
Teknologi MARA, Shah Alam, Selangor, Malaysia.
X Abstract published in Advance ACS Abstracts, October 1, 1996.
(1) Brief summaries of portions of this work have appeared previ-
ously: (a) Kevill, D. N.; Anderson, S. W.; Fujimoto, E. K. In Nucleo-
philicity; Harris, J . M.; McManus, S. P., Eds.; Advances in Chemistry
Series; American Chemical Society: Washington, DC, 1987; p 278. (b)
Kevill, D. N.; Anderson, S. W.; Ismail, N. HJ ; Fujimoto, E. K. Studies
Org. Chem. 1987, 31, 311.
(7) For a review with tables of values and leading references, see:
Bentley, T. W.; Llewellyn, G. Prog. Phys. Org. Chem. 1990, 17, 121.
(8) Liu, K.-T.; Sheu, H.-C. J . Org. Chem. 1991, 56, 3021.
(9) Bentley, T. W.; Dau-Schmidt, J . P.; Llewellyn, G.; Mayr, H. J .
Org. Chem. 1992, 57, 2387.
(10) Liu, K.-T.; Chin, C.-P.; Lin, Y.-S.; Tsao, M.-L. Tetrahedron Lett.
1995, 36, 6919.
(2) Kevill, D. N.; Ismail, N. HJ ; D’Souza, M. J . J . Org. Chem. 1994,
59, 6303.
(11) Liu, K.-T.; Hsu, H.-Y.; Yang, J .-S. Tetrahedron Lett. 1992, 33,
3327.
(3) Kevill, D. N.; Anderson, S. W. J . Am. Chem. Soc. 1986, 108, 1579.
(4) Grunwald, E.; Winstein, S. J . Am. Chem. Soc. 1948, 70, 846.
(5) (a) Winstein, S.; Grunwald, E.; J ones, H. W. J . Am. Chem. Soc.
1951, 73, 2700. (b) Schadt, F. L.; Bentley, T. W.; Schleyer, P. v. R. J .
Am. Chem. Soc. 1976, 98, 7667. (c) Kevill, D. N.; Lin, G. M. L. J . Am.
Chem. Soc. 1979, 101, 3916.
(6) (a) Kevill, D. N.; Anderson, S. W. J . Org. Chem. 1991, 56, 1845.
(b) Kevill, D. N. In Advances in Quantitative Structure-Property
Relationships; Charton, M., Ed.; J AI Press: Greenwich, CT, 1996; Vol.
1, pp 81-115.
(12) Allen, A. D.; Girdhar, R.; J ansen, M. P.; Mayo, J . D.; Tidwell,
T. T. J . Org. Chem. 1986, 51, 1324.
(13) (a) Winstein, S.; Fainberg, A. H.; Grunwald, E. J . Am. Chem.
Soc. 1957, 79, 4146. (b) Bunton, C. A.; Mhala, M. M.; Moffatt, J . R. J .
Org. Chem. 1984, 49, 3639.
(14) Kevill, D. N.; D’Souza, M. J . J . Chem. Soc., Perkin Trans. 2
1995, 973.
(15) Kevill, D. N.; D’Souza, M. J . Org. React. (Tartu) 1995, 29, 55.
(16) Kevill, D. N.; D’Souza, M. J . J . Chem. Res. Synop. 1996 286; J .
Chem. Res., Miniprint 1649.
S0022-3263(96)00909-7 CCC: $12.00 © 1996 American Chemical Society

Solvolysis of the Benzhydryldimethylsulfonium Ion
J . Org. Chem., Vol. 61, No. 21, 1996 7257
+
Ta ble 1. Sp ecific Ra tes of Solvolysis of th e
developed using the solvolyses of p-MeOC6H4CH2SMe2
Ben zh yd r yld im eth ylsu lfon iu m Ion a a t 25.0 °C
(2), in conjunction with the Y+ scale3 (from a study of
the solvolyses of the 1-adamantyldimethylsulfonium ion)
and, when eq 3 is employed, the NT scale developed from
a consideration of the specific rates of solvolysis of the
S-methyldibenzothiophenium ion.6 Although, when eq
3 was applied, solvolyses of benzhydryl bromides10 with
moderately electron-withdrawing substituents showed l
values of very close to zero and large probabilities that
the lNT term was statistically insignificant, the presence
of the more powerfully electron-withdrawing p-NO2 sub-
stituent led to an appreciable l value of 0.31 ( 0.09 with
only a very low (0.01) probability that the lNT term was
statistically insignificant.16 Accordingly, it is important
in the study of the solvolyses of 1 to see whether here
also an lNT term needs to be incorporated into the
extended Grunwald-Winstein treatment.
As an alternative to the use of the hI term in conjunc-
tion with the mYX term, one can develop a series of
similarity models for use as the Y scale when the
solvolyses under study involve substrates with R-aryl
substituents. Several scales of this type (YBnX) have been
developed by Liu and co-workers,8,18-23 and Bentley has
suggested the use of p-methoxybenzyl chloride9,24 for
correlations of benzylic chlorides. It has been recog-
nized10 that additional similarity model scales are needed
when two aromatic rings are in conjugation with the
charge developing at the R-carbon.
solventb
104k, s-1 c
solventb
104k, s-1 c
100% EtOH
80% EtOH
60% EtOH
40% EtOH
20% EtOH
100% H2O
100% MeOH
80% MeOH
60% MeOH
40% MeOH
20% MeOH
95% acetone
90% acetone
80% acetone
70% acetone
60% acetone
40% acetone
20% acetone
4.60 ( 0.06d
2.30 ( 0.04
1.77 ( 0.04
1.66 ( 0.12
2.03 ( 0.15
2.19 ( 0.08
7.88 ( 0.13
4.53 ( 0.09
3.16 ( 0.17
2.61 ( 0.09
2.38 ( 0.15
3.16 ( 0.15
2.81 ( 0.06
2.29 ( 0.06
1.70 ( 0.08
1.62 ( 0.03
1.40 ( 0.03
1.58 ( 0.07
95% dioxane
80% dioxane
60% dioxane
40% dioxane
20% dioxane
100% TFE
97% TFE
90% TFE
70% TFE
50% TFE
97% HFIP
70% HFIP
50% HFIP
80T-20E
60T-40E
40T-60E
0.75 ( 0.01e
1.02 ( 0.08
1.08 ( 0.04
1.02 ( 0.03
1.41 ( 0.08
27.3 ( 1.7f
20.8 ( 1.4g
15.6 ( 1.1h
9.25 ( 0.19
6.39 ( 0.26
57 ( 6i
40.3 ( 2.9j
24.1 ( 1.6j
20.1 ( 0.9
13.5 ( 1.1
7.62 ( 0.20
6.37 ( 0.21
20T-80E
a
As the trifluoromethanesulfonate salt and at a concentration
of 0.003-0.005 M, except the concentration was ca. 0.01 M in
b
HFIP-containing solvents. Prepared on a volume-volume basis
at 25.0 °C, except for TFE-H2O and HFIP-H2O mixtures, which
were prepared on a weight-weight basis. c All runs performed in
duplicate, and all of the integrated values are used to obtain an
d
average value and the associated standard deviation. Also values
for 104 k (s-1) of 21.1 ( 1.2 at 35.0 °C, 0.91 ( 0.04 at 15.0 °C, and
0.067 ( 0.003 at 0.0 °C; from the four determinations at different
temperatures, ∆Hq ) 26.6 ( 0.1 kcal mol-1 and ∆Sq ) 15.3
298
298
( 0.2 eu. e Not used in the correlations (no NT value available).
f Substrate concentration of 0.0050 M in the presence of 0.0045 M
pyridine and using eq 5 with an R value of 600; in the absence of
pyridine an equilibrium was established at 20.0 ( 0.8% reaction.
It has been demonstrated2,14-16 in terms of eq 2 (or eq
3) that the important consideration is not similarity in
structure but similarity in the h/m ratio for the solvolyses
of the substrate under investigation and the solvolyses
of the substrate chosen as the similarity model. It has
been shown that the h value of the ratio increases in
value not only on introduction of a second aromatic ring
at the R-carbon but also on introduction at this carbon
of an electron-withdrawing group or on introduction
within the aromatic ring of an electron-supplying sub-
stituent.
g
Substrate concentration of 0.0050 M and using eq 5 with an R
value of 130; equilibrium was established at 92 ( 1% reaction.
h
No evidence for fall-off of experimental integrated first-order rate
coefficients, and equilibrium was established at 97 ( 1% reaction.
Average value of determinations with 0.0100 M substrate and
0.0076, 0.0151, and 0.0227 M pyridine present (six runs) and using
eq 5 with an R value of 1500; in the absence of pyridine an
equilibrium was established at 11.9 ( 1.1% reaction. In the
absence of pyridine and using eq 5 with an R value of 200 for 70%
HFIP and an R value of 150 for 50% HFIP.
i
j
Although one would expect a similarity model that is
closely related to the substrate undergoing solvolysis, in
regard to both h/m ratio and actual structure, to give
excellent correlations, it has been shown2,14-16 that in
almost all instances the improvements relative to use of
the appropriate YX scale7 in conjunction with the I scale2
are marginal. It follows that the considerable extra effort
involved in first establishing what would be a good
similarity model (no nucleophilic component and similar
h/m ratio) and then, if necessary, experimentally deter-
mining the YBnX scale would only rarely25 be justifiable.
In the present paper, we report on the solvolyses of 1
with respect to specific rates of solvolysis in a wide range
of solvents, temperature effects on the specific rates of
ethanolysis, product partitioning in aqueous ethanol
mixtures, and common-molecule retardation (including
studies with initially added dimethyl sulfide). The
specific rates of solvolysis are found to be very well
correlated using I and Y+ values within eq 2.
Resu lts
Kin etic Stu d ies. All kinetic studies were carried out
with 0.003-0.006 M concentration of 1, except for a ca.
0.01 M concentration being used in 1,1,1,3,3,3-hexafluoro-
2-propanol (HFIP)-containing solvents. The solvents
were prepared, at 25.0 °C, on a volume-volume basis,
except for 2,2,2-trifluoroethanol (TFE)-H2O and HFIP-
H2O mixtures, which were prepared on a weight-weight
basis.
Specific rates of solvolysis, determined at 25.0 °C, are
reported in Table 1 for 35 solvents. For determinations
in water, ethanol, and methanol and in aqueous mixtures
with ethanol, methanol, acetone, and dioxane, as well as
in TFE-ethanol mixtures, the values reported are aver-
age values of integrated first-order rate coefficients from
duplicate runs. Similar analyses were possible for sol-
volyses in 70% and 50% TFE.
(17) (a) Lee, I.; Koh, H. J .; Chang, B. D. Bull. Korean Chem. Soc.
1995, 16, 1104. (b) Zalewski, R. I.; Krygowski, T. M.; Shorter, J .
Similarity Models in Organic Chemistry, Biochemistry and Related
Fields; Elsevier: New York, 1991.
(18) Liu, K.-T.; Sheu, H.-C.; Chen, H.-I.; Chiu, P.-F.; Hu, C.-R.
Tetrahedron Lett. 1990, 31, 3611.
(19) Liu, K.-T.; Sheu, H.-C. J . Chin. Chem. Soc. 1991, 38, 29.
(20) Liu, K.-T.; Chen, H.-I.; Chin, C.-P. J . Phys. Org. Chem. 1991,
4, 463.
(21) Liu, K.-T.; Yang, J .-S.; Chang, S.-M.; Lin, Y.-S.; Sheu, H.-C.;
Tsao, M.-L. J . Org. Chem. 1992, 57, 3041.
(22) Liu, K.-T. J . Chin. Chem. Soc. 1995, 42, 607.
(23) For a review commentary, see: Kevill, D. N.; D’Souza, M. J . J .
Phys. Org. Chem. 1992, 5, 287.
(24) Bentley, T. W.; Koo, I. S.; Norman, S. J . J . Org. Chem. 1991,
56, 1604.
Runs in other fluoro alcohol-containing solvents needed
analyses taking into account an appreciable common-
molecule rate depression. Also, the runs in 100% TFE
(25) Kevill, D. N.; D’Souza, M. J . J . Chem. Res., Synop. 1993, 332

7258 J . Org. Chem., Vol. 61, No. 21, 1996
Kevill et al.
Ta ble 2. Effect of Ad d ed P yr id in e a n d Dim eth yl Su lfid e
u p on th e Sp ecific Ra te of Solvolysis of th e Ben zh yd r yl-
d im eth ylsu lfon iu m Ion a in 100% TF E a t 25.0 °C
Ta ble 4. Effect of Ad d ed Dim eth yl Su lfid e u p on th e
Exp er im en ta l Sp ecific Ra te of Solvolysis of th e
Ben zh yd r yld im eth ylsu lfon iu m Ion a a t 25.0 °C
[pyridine], M
[Me2S], M
104kexpt, s-1 b
104k, s-1 c
solvent
[Me2S], M
104kexpt, s-1
Rb (M-1
)
0.00453d
0.00650
0.00905d
0.01300
0.01359d
0.00453
0.00453
0.00453
0.00453
0.00453
27.3 ( 1.7
25.0 ( 1.5
29.4 ( 2.2
30.0 ( 2.2
31.0 ( 1.7
24.9 ( 1.1
28.0 ( 1.4
31.6 ( 1.0
23.7 ( 1.2
25.0 ( 1.6
95% acetone
0.0000
0.0500
0.1000
0.0000
0.0500
0.1000
3.00 ( 0.04c
2.67 ( 0.05
2.47
2.50
2.40 ( 0.05
95% dioxane
0.750 ( 0.012
0.701 ( 0.007
0.666 ( 0.012
1.40
1.26
0.0025
0.0050
0.0100
0.0200
0.0400
a
As the trifluoromethanesulfonate salt and concentration of ca.
3.68 ( 0.14
1.65 ( 0.07
0.91 ( 0.04
0.005 M. Calculated from eq 4. c An additional determination,
b
independent of the one reported in Table 1.
a
Ta ble 5. Illu str a tive Ru n for Solvolysis of 0.00621 M
Ben zh yd r yld im eth ylsu lfon iu m Ion a in 100% TF E, a t 25.0
°C in th e P r esen ce of 0.00650 M P yr id in e
As the trifluoromethanesulfonate salt and concentration of ca.
b
0.005 M. Integrated first-order rate coefficients fall off ap-
preciably with increasing extent of reaction, except at the higher
[Me2S] concentrations. c Calculated on a point-by-point basis from
the experimental integrated first-order rate coefficients according
t (min)
0
3.79 7.08 9.67 12.55 15.38 18.47 1620
1.05 2.26 3.00 3.25 3.57 3.98 4.20 6.01
12.3 11.8 10.2 9.4 9.7 9.1
25.0 27.2 24.6 23.8 26.1 25.4
12.2 11.2 10.2 9.6 9.4 8.9
103x, M
104kexpt
104kb
d
to eq 5, with an R value of 600. Infinity acid titer of 95 ( 3%, 91
( 3%, and 85 ( 3% of maximum possible with [pyridine] of
0.004 53, 0.009 05, and 0.013 59 M, respectively.
104kexpt
(calcd)
Ta ble 3. Effect of Ad d ed P yr id in e a n d Dim eth yl Su lfid e
u p on th e Sp ecific Ra te of Solvolysis of th e Ben zh yd r yl-
d im eth ylsu lfon iu m Ion a in 97% HF IP a t 25.0 °C
a
Footnotes as in Table 3, plus x represents extent of formation
b
of acid. Using R ) 600 in eq 5 (with c ) 0).
[pyridine], M
[Me2S], M
104k, s-1 b
Rx
t
) kexpt[1 + R(a + c)]
(5)
0.0076
0.0151
0.0227
0.0076
0.0076
56 ( 6c
51 ( 4
k +
63 ( 6
0.0200
0.0400
55 ( 6c
69 ( 6c,d
initial concentration of substrate and x represents the
extent of solvolysis at time t. The k values, calculated
according to eq 5, are given in Tables 2 and 3, which also
show the essentially negligible effect of variation of the
pyridine concentration. The k values are obtained from
the kexpt values by determining the best fit value for the
mass law constant (R), and this R value is then used in
the estimation (eq 5 with c ) 0) of the k values reported
in Table 1. For runs in 97% TFE and 70% and 50%
HFIP, the R value is smaller and the variation in kexpt
values during the solvolyses in the absence of added Me2S
can be analyzed directly to get the best fit value for R
and the appropriate k value. An illustrative run and the
accompanying analysis for the solvolysis of 1 in 100%
TFE is presented in Table 5.
In the consideration of the specific rates of ethanolysis
of 1, values were also determined at 35, 15, and 0 °C and
these values, together with the enthalpy and entropy of
activation values obtained from an Arrhenius equation
treatment of the four measurements, are presented in
footnote d of Table 1.
Deter m in a tion of Selectivities in Aqu eou s Eth a -
n ol. From a gas chromatographic determination of the
relative amounts of benzhydrol and benzhydryl ethyl
ether formed, one can determine selectivity values (S),
defined in eq 6, where the ratio of the concentrations of
a
As the trifluoromethanesulfonate salt and concentration of ca.
b
0.010 M. Calculated on a point-by-point basis from the experi-
mental integrated first-order rate coefficients according to eq 5,
with an R value of 1500. c Followed to 40-50% of possible acid
d
formation. Experimental first-order rate coefficient value of 1.07
((0.10) × 10-4 s-1, showing slight fall-off with increasing extent
of reaction.
and 97% HFIP reached early equilibria, which could be
avoided by addition of at least 1 equiv of pyridine (Tables
2 and 3). The infinity acid titers then approached the
theoretical values but remained somewhat less in 100%
TFE, presumably because of some capture by pyridine
to give the benzhydrylpyridinium ion (without acid
development); the percentages of the theoretical infinity
titer in the presence of up to 0.0136 M pyridine are
reported as a footnote to Table 2.
The common-molecule rate depression is very small in
95% acetone or 95% dioxane, and it can be detected as a
modest retardation only in the presence of a large excess
(0.05-0.10 M) of added dimethyl sulfide (Me2S). The
specific rates and the mass law constants (R), calculated
according to eq 4, are presented in Table 4. In eq 4, kexpt
k ) kexpt(1 + Rc)
(4)
is the specific rate determined in the presence of a
concentration c of Me2S and k is the experimental value
in the absence of added Me2S.
kE [(C6H5)2CHOEt] [H2O]
S )
)
‚
(6)
kW
[EtOH]
[(C6H5)2CHOH]
In contrast, extremely large common-molecule retarda-
tions are observed for solvolyses, even in the absence of
initially added Me2S, in 100% TFE and 97% HFIP. In
these solvents, a very pronounced initial fall-off in
integrated first-order rate coefficient values is followed
by a series of fairly constant values, which could, errone-
ously, be averaged to give an apparent (but much too low)
k value. When concentrations of added Me2S are not in
large excess, the kexpt values fall with extent of reaction
and the true k values are given by eq 5, where a is the
benzhydryl derivatives is determined after, at least, 10
half-lives of reaction, and the concentrations of water and
ethanol are those present in the solvent mixture. De-
terminations were carried out with from 96 to 20%
(volume-volume) of ethanol in the binary solvent. The
selectivities are reported in Table 6, together with (for
comparison) previously determined values for benzhydryl
bromide,26 chloride,26-29 2,4-dinitrobenzoate,28 and p-
nitrobenzoate.28

Solvolysis of the Benzhydryldimethylsulfonium Ion
J . Org. Chem., Vol. 61, No. 21, 1996 7259
Discu ssion
ing secondary alkyl groups that are much more suscep-
tible to nucleophilic attack than the benzhydryl group.
While the N-benzhydrylpyridinium ion will not be
formed via a concerted bimolecular route, it can be
formed by capture of the intermediate carbocation, in
competition with solvolysis. A competition of this type
was found to be present when pyridine was added to
solvolyses of the (p-methoxybenzyl)dimethylsulfonium
ion.2 For solvolyses of 1 in 100% TFE, the infinity titers
were 95, 91, and 85% of the theoretical value with
additions of pyridine of 0.0045, 0.0090, and 0.01359 M,
respectively (Table 2, footnote d). These observations
correspond to kN/kS ratios of 12, 11, and 13 M-1, where
kN is the second-order rate coefficient for capture of the
carbonium ion by pyridine and kS is the first-order rate
coefficient for capture by solvent. The constancy of the
value obtained is consistent with the proposed mecha-
nism, and the average value corresponds to a selectivity
value (S) for capture by a pyridine molecule relative to a
TFE molecule of 160.
Com m on Molecu le Retu r n a n d th e Effect of
Ad d ed Dim eth yl Su lfid e. In 95% acetone and 95%
dioxane, although no fall-off in values could be detected
during kinetic runs in the absence of additives, the
specific rates could be depressed to a lower, essentially
constant, value by the addition of a fairly large excess of
dimethyl sulfide (Table 4). Defining the mass law
constant (R) in the usual way, as the k-1/k1 ratio within
eq 7, one arrives at a rate equation at any instant as
Sp ecific Ra tes for Solvolysis of 1. From the Ar-
rhenius plot of the specific rates of ethanolysis at 0-35
°C, one can calculate a value of 1.6 × 10-2 s-1 at 50 °C,
about 200 times greater than the value for the (p-
methoxybenzyl)dimethylsulfonium ion (2) at that tem-
perature.2 A linear free energy relationship (LFER) plot
of the specific rates of solvolysis of 1 at 25.0 °C against
the corresponding specific rates of solvolysis of 2 at 50.0
°C gives, for 34 solvents, a linear plot with a slope of 1.01
( 0.02, intercept of 0.01 ( 0.07, correlation coefficient of
0.991, and F-test value of 1843.
As in two previous studies of solvent effects upon the
specific rates of solvolysis of substrates believed to
solvolyze by the SN1 mechanism and involving dimethyl
sulfide as the leaving group, the highest specific rate for
the series of solvents studied was in 97% HFIP and the
lowest was in 95% dioxane. The ratio of the specific rates
in these solvents was 76 (at 25.0 °C), compared to 50 for
solvolyses of 2 (at 50.0 °C)2 and 5 for solvolyses of the
1-adamantyldimethylsulfonium ion (at 70.4 °C).3 A very
small variation of the specific rate of solvolysis with
solvent variation has also been found30,31 for the solvoly-
ses of the 1-adamantylpyridinium ion. This once again
illustrates the larger variation observed, due to differ-
ences in solvation changes in going from the ground state
to the transition state, upon the introduction of R-aryl
substituents.2,9,22,32,33 Variations with solvent of the
extent of ion-pair (in this case, ion-molecule) return have
also been considered as a possible source of disper-
sion.13,23,34,35
In almost all instances the reaction proceeded es-
sentially to completion, but as indicated in the footnotes
to Table 1, an equilibrium was established in solvents
rich in fluoro alcohol. With a substrate concentration of
0.005 M, the solvolysis proceeded to 20% in 100% TFE,
92% in 97% TFE, and 97% in 90% TFE. With a substrate
concentration of 0.0100 M, the solvolysis proceeded to
12% in 97% HFIP.
Effect of Ad d ed P yr id in e. In order to follow the
kinetics of solvolysis with reasonable precision, uncom-
plicated by a movement toward equilibrium, pyridine was
added to the solvolyses in 100% TFE and 97% HFIP. In
both of these solvents essentially unchanged specific rates
were obtained after addition of up to 2.5 equiv of pyridine
(Tables 2 and 3). A second-order component involving
attack by pyridine is not to be expected because it has
been shown36 that additions of a moderate concentration
of pyridine do not significantly increase the rates of
solvolysis of N-isopropyl-, N-2-pentyl-, or N-3-pentylquin-
olium salts in 100% TFE or 100% HFIP, cations contain-
Ph2CHSMe2+ yk1z Ph2CH+
+
k-1
SMe2 SOH8 Ph2CHOS + SOH2+ + X- (7)
ks
expressed in eq 8, where a is the initial concentration of
d[Ph2CHOS]
) kinst(a - x)(1 + Rx)-1
(8)
dt
substrate, x is the amount of solvolysis product that has
been formed at time t, and kinst is the instantaneous rate
coefficient for development of solvolysis product. With
a large excess of added dimethyl sulfide, integration leads
to eq 4 with, using the data of Table 4, R values of 2.5 (
0.3 M-1 in 95% acetone and 1.3 ( 0.2 M-1 in 95% dioxane.
The value in 95% acetone corresponds to slightly less
external return than for the corresponding solvolysis of
the (p-methoxybenzyl)dimethylsulfonium ion (R value of
4.8 ( 0.6, at 50.0 °C).
In solvents rich in fluoro alcohol, the nucleophilicity
of the solvent is considerably reduced relative to 95%
acetone or dioxane6 and, accordingly, in the competition
for the benzhydryl carbocation between solvolysis and
return of dimethyl sulfide the extents of return are
considerably magnified.
Under these conditions, the integrated first-order rate
coefficients fall off with extent of reaction, as dimethyl
sulfide accumulates. An illustrative run with 0.0062 M
substrate in 100% TFE is presented in Table 5. The fall-
off in values for kexpt can be reproduced using an initial
value (zero [Me2S]) of 25.4 ( 1.2 s-1 for kexpt and an R
value of 600 M-1, with the analysis in terms of eq 5 and
c (the added [Me2S] concentration) being zero. The
addition of equimolar pyridine is to prevent an early
(26) Karton, Y.; Pross, A. J . Chem. Soc., Perkin Trans. 2 1977, 1860.
(27) (a) Harris, J . M.; Becker, A.; Clark, D. C.; Fagan, J . F.; Keenan,
S. L. Tetrahedron Lett. 1973, 3813. (b) Harris, J . M.; Clark, D. C.;
Becker, A.; Fagan, J . F. J . Am. Chem. Soc. 1974, 96, 4478.
(28) McClennan, D. J .; Martin, P. L. J . Chem. Soc., Perkin Trans. 2
1982, 1099.
(29) Bentley, T. W.; Ryu, Z. H. J . Chem. Soc., Perkin Trans. 2 1994,
761.
(30) Katritzky, A. R.; Brycki, B. J . Am. Chem. Soc. 1986, 108, 7295.
(31) Katritzky, A. R.; Brycki, B. Chem. Soc. Rev. 1990, 19, 803.
(32) Fainberg, A. H.; Winstein, S. J . Am. Chem. Soc. 1957, 79, 1597.
(33) Fujio, M.; Susuki, T.; Yatsugi, K.; Saeki, Y.; Goto, M.; Kim, S.
H.; Tsuji, Y.; Rappoport, Z.; Tsuno, Y. Bull. Chem. Soc. J pn. 1995, 68,
2619 and references therein.
(34) Fainberg, A. H.; Winstein, S. J . Am. Chem. Soc. 1957, 79, 1602.
(35) McManus, S. P.; Sedaghat-Herati, M. R.; Karaman, R. M.;
Neamati-Mazraeh, N.; Cowell, S. M.; Harris, J . M. J . Org. Chem. 1989,
54, 1911.
(36) Katritzky, A. R.; Lopez-Rodriguez, M. L.; Marquet, J . J . Chem.
Soc., Perkin Trans. 2 1984, 349.

7260 J . Org. Chem., Vol. 61, No. 21, 1996
Kevill et al.
Ta ble 6. Selectivity Va lu es (kE/kW) for th e Solvolysis of
th e Ben zh yd r yld im eth ylsu lfon iu m Ion in Aqu eou s
Eth a n ol a t 25.0 °C a n d Com p a r ison w ith Deter m in a tion s
for th e Cor r esp on d in g Solvolyses of Oth er Ben zh yd r yl
Der iva tives (P h 2CHX)
Ta ble 7. Ma ss-La w Con sta n t (r) for Solvolyses of 1 in
Solven ts of Va r yin g Solven t Nu cleop h ilicity (NT) Va lu es
a t 25.0 °C
solventa
R, M-1
log R
NT
b
97% HFIP
70% HFIP
50% HFIP
100% TFE
97% TFE
1500c
200
150
600
130
2.5
3.18
2.30
2.18
2.78
2.11
0.40
-5.26
-2.94
-2.49
-3.93
-3.30
-0.49
kE/kW values
% EtOH (v/v) in aqueous EtOH
X
96 95 90 86 80 70 60 50 40 30 20
2.6 2.3 2.7 2.7 2.5 4.3 4.0
+ a
SMe2
Brb
95% acetone
3.1 3.3
2.3 2.5
2.2
3.7 4.1
a
On a weight-weight basis, except for 95% acetone, which is
Clb
2.8 3.3c
b
on
a volume-volume basis at 25.0 °C. Values from ref 6.
Cld,e
Cle,f
2.6 3.0 3.2 3.8
c Calculated, using eq 5, from runs in presence of an excess of
dimethyl sulfide.
2.5
2.8 3.3 3.7 4.0 4.4 4.6
2.3 2.8 3.2 3.8
DNBd,g
PNBd,h
2.2 2.6 3.3 3.8
because the k-1/k ratio, which R represents, is dependent
not only on solvent nucleophilicity but also upon the
nucleophilicity of dimethyl sulfide, which one would
expect to be solvent dependent. However, the nucleo-
philicity of the dimethyl sulfide will be to a large degree
influenced by the ability of the solvent to interact
electrophilically with the sulfur atom, and these reduc-
tions in sulfide nucleophilicity will be approximately
related to the reductions in solvent nucleophilicity.
Exten d ed Gr u n w a ld -Win stein Equ a tion Tr ea t-
m en t w ith Ap p lica tion of th e Ar om a tic Rin g P a -
r a m eter (I) Sca le. For 34 of the 35 solvents listed in
Table 1 (NT value not available for 95% dioxane), the
variation in specific rate with solvent composition has
been analyzed in terms of eqs 1, 2, and 3, and, also, in
terms of linear regression against I values and regression
against a combination of Y+ and NT scales. The results
of these five analyses are presented in Table 8.
a
b
This work. At 25.0 °C, from ref 26. c A value of 5.0 has also
been reported (ref 27). From ref 28. e At 25.0 °C. f From ref 29.
The 2,4-dinitrobenzoate ester at 100 °C. The p-nitrobenzoate
ester at 100 °C.
d
g
h
(20%) equilibrium. This pyridine concentration would
not be expected to seriously perturb the kinetics since
interpolation within data presented as footnote d in Table
2 leads to an infinity titer of 93% of theoretical. This
indicates that pyridine at this concentration competes
only weakly with the solvent, in contrast to dimethyl
sulfide, which competes very effectively. The effective-
ness of the dimethyl sulfide intervention is indicated by
the observation (Table 5) that, when the initial titer is
at 17% solvolysis and a subsequent titer at 38% solvoly-
sis, the calculated kexpt value is only 50% of the true initial
value.
When dimethyl sulfide is added initially to solvolyses
in 100% TFE at 25.0 °C, the fall-off in specific rates is
dramatic. With 0.01 M additive, the specific rate is about
12% of the value in the absence of dimethyl sulfide, and
with 0.04 M additive the corresponding ratio is only about
4%. Using the previously estimated R value of 600 M-1
one can convert, using eq 5, the observed kexpt values to
the unperturbed specific rates of solvolysis (Table 2).
In 97% HFIP, the common-molecule rate depression
is even more pronounced, and it was not possible to
obtain even an approximately accurate value for R from
runs in the absence of added dimethyl sulfide. An
enormous fall-off in rate with development of the first
small amounts of dimethyl sulfide was observed. The R
value could, however, be accurately estimated (at a value
of 1500 M-1) from the very slow runs in the presence of
initially added dimethyl sulfide. To illustrate this re-
tardation, in the presence of 0.040 M dimethyl sulfide,
the true solvolysis rate (in the absence of perturbation
by dimethyl sulfide) was reduced by 98.5%. The R value
obtained from these experiments was then applied, using
eq 5, to solvolyses in the absence of initially added
dimethyl sulfide. A value for the specific rate of solvolysis
in the absence of any perturbation induced by dimethyl
sulfide of about 60 × 10-4 s-1 (at 25.0 °C) was obtained
both for runs in the absence of initially added dimethyl
sulfide and for runs when up to 0.04 M dimethyl sulfide
was added (Table 3).
It can be seen, from analyses either in terms of Y+ and
NT scales (l ) -0.01 ( 0.10) or Y+, I, and NT scales (l )
0.03 ( 0.02), that the specific rates of solvolysis are
insensitive to changes in solvent nucleophilicity. For
linear regression analyses, it is found that correlation
against Y+ values is unsatisfactory (r ) 0.795; F-test
value of 55) and correlation against I values is consider-
ably superior (r ) 0.935; F-test value of 224).
The best results (see Figure 1) are obtained by multiple
regression analysis against a combination of Y+ and I
values (R ) 0.991; F-test value of 897). In this correla-
tion, the m value of 1.35 ( 0.10 and h value of 0.99 (
0.04 are essentially identical to those for the solvolysis
of the (p-methoxybenzyl)dimethylsulfonium ion at 50.0
°C [h value of 1.00 (standard system) and m value of 1.3].2
This correspondence in h and m values is consistent with
the previously discussed good LFER plot with unit slope
for the specific rates of the two solvolyses. It appears
+
that, for solvolyses of the RSMe2 derivative, the two
phenyl substituents within the benzhydryl(diphenyl-
methyl) group have the same influence upon the changes
in the specific rates of solvolysis, as they result from
variation of both Y+ and I, as the p-methoxyphenyl
substituent within the p-methoxybenzyl (p-methoxyphe-
nyl)methyl group.
P r od u ct Selectivity in Aqu eou s Eth a n ol. For
solvolyses of 1 at 25.0 °C in aqueous ethanol mixtures,
the product partitioning between benzhydrol and ben-
zhydryl ethyl ether has been determined and converted
into selectivity values (S), defined as kE/kW, where kE and
kW are the second-order rate coefficients for capture by
ethanol and water, respectively. The previously noted
absence of any sensitivity of the specific rates of solvolysis
toward changes in solvent nucleophilicity suggests that
the capture is of a carbocation previously formed in the
The R values obtained for solvolyses in six solvents,
varying in solvent nucleophilicity (NT) values from -0.49
(95% acetone) to -5.26 (97% HFIP),6 are reported in
Table 7. It is found that the log R values are quite well
correlated with solvent nucleophilicity values, with a
slope of -0.58 ( 0.08, intercept of 0.39 ( 0.30, and
correlation coefficient of -0.960. It could be considered
as rather surprising that the correlation is this precise,

Solvolysis of the Benzhydryldimethylsulfonium Ion
J . Org. Chem., Vol. 61, No. 21, 1996 7261
Ta ble 8. Cor r ela tion of th e Sp ecific Ra tes of Solvolysis, in 34 Solven ts a t 25.0 °C,a of th e Ben zh yd r yld im eth ylsu lfon iu m
Ion , Usin g th e Gr u n w a ld -Win stein Ap p r oa ch a n d Va r iou s Com bin a tion s of Y+, NT, a n d I P a r a m eter s
scale(s)b
lc
mc
hc
cd
Re
Ff
Y+
2.71 ( 0.37
-0.17 ( 0.31
0.21 ( 0.18
0.795
0.935
0.795
55
224
27
I
1.30 ( 0.09
Y+, NT
-0.01 ( 0.10
2.60 ( 0.85
-0.17 ( 0.31
(0.89)
Y+, I
1.35 ( 0.10
1.56 ( 0.19
0.99 ( 0.04
1.00 ( 0.04
0.00 ( 0.07
0.00 ( 0.07
0.991
0.992
897
612
Y+, I, NT
0.03 ( 0.02
(0.20)
a
b
Data from Table 1. Values from refs 3, 6, and 7. c With associated standard error; values in parentheses are the probabilities that
d
the term is not statistically significant (recorded if greater than 0.005). Residual (constant) term, accompanied by the standard error of
the estimate. e Correlation coefficient. f The F-test value.
tivities toward changes in both Y+ and I are virtually
identical, and it appears that, within these correlations,
the influence of two phenyl groups mimics that of a
p-methoxyphenyl group. The utility of the aromatic ring
parameter (I) in these correlations gives yet further
evidence for its generality in correcting for dispersion in
Grunwald-Winstein plots for the solvolyses of benzylic
and related derivatives.
Common-molecule return is found. While it constitutes
a weak effect in 95% acetone (detectable only on addition
of appreciable concentrations of dimethyl sulfide), the
effect is considerably magnified in solvents rich in fluoro
alcohol. In 97% HFIP, the rate depression is so pro-
nounced that an accurate value for the mass law constant
(R) can be obtained only from runs carried out in the
presence of a fairly large initial concentration of dimethyl
sulfide. The log R values, available for six solvents,
correlate quite well with NT values (r ) -0.960).
F igu r e 1. Plot of log (k/ko) for 1 against (1.35Y+ + 0.99I).
rate-determining step. The modest preference for reac-
tion with the more nucleophilic ethanol component (Table
6) suggests capture of a free carbocation, and the
somewhat larger S value in the more aqueous solvents
is consistent with previously observed29,37 trends for such
a capture. In contrast, capture at a solvent-separated
ion pair is usually accompanied by a modest preference
for reaction with the water and an S value of below unity.
Similar behavior was observed for the product partition-
ing both from the 1-adamantyldimethylsulfonium ion and
from 1-adamantyl derivatives with anionic leaving groups,
suggesting that capture at a solvent-separated ion-
molecule stage has very similar characteristics to capture
at a solvent-separated ion pair stage.3
If the capture is at the free carbocation stage then the
competition must be independent of the source of the
carbocation, and not surprisingly, the selectivities ob-
served for 1 parallel closely those previously observed,
and listed in Table 6, for the corresponding bromides,26
chlorides,26,28,29 2,4-dinitrobenzoates,28 and p-nitroben-
zoates.28
Selectivity values (kE/kW) in aqueous ethanol show a
small preference for reaction with ethanol, and they rise
slightly in value in the water-rich solvents. This pattern
of behavior is usually taken to indicate competition for a
free carbocation, and it contrasts with that for capture
at the solvent-separated ion-pair (or ion-molecule) stage,
which is usually accompanied by a slight preference for
reaction with the less nucleophilic water.
Exp er im en ta l Section
Ma ter ia ls. The purification of acetone, dioxane, ethanol,
and methanol were as described previously.38 The purifica-
tions of 1,1,1,3,3,3-hexafluoro-2-propanol39 and 2,2,2-trifluo-
roethanol40 also used previously reported procedures. Ben-
zhydryl chloride, dimethyl sulfide, and silver trifluorometh-
anesulfonate were used without further purification. Pyridine
was refluxed over barium monoxide and then fractionally
distilled.41 Benzhydrol (Aldrich) was recrystallized from li-
groin (bp 60-80 °C), with decolorization using Norit (activated
carbon), to give white needles. Benzhydryl ethyl ether was
prepared using the method previously reported for the prepa-
ration of 1-adamantyl ethyl ether38 but with benzhydryl
chloride replacing 1-adamantyl iodide: bp 122-122.5 °C (1.4
mmHg) [lit.42 bp 128-132 °C (1.3 mmHg)].
Con clu sion s
The logarithms of the specific rates of solvolysis of 1
in 34 hydroxylic solvents of widely varying character can
be very well correlated against a combination of Y+ and
I parameters (R ) 0.991). It is found that the sensitivity
of the specific rates of solvolysis toward changes in
solvent nucleophilicity (NT) values is essentially zero,
consistent with the considerable evidence for SN1 sol-
volyses of benzhydryl derivatives. The ion 1 undergoes
ethanolysis at 50.0 °C about 200 times faster than the
(p-methoxybenzyl)dimethylsulfonium ion, but the sensi-
Ben zh yd r yld im eth ylsu lfon iu m Tr iflu or om eth a n esu l-
fon a te. The procedure followed was identical to that previ-
ously used2 to prepare the (p-methoxybenzyl)dimethylsulfo-
nium salt, but with benzhydryl chloride substituted for
(38) Kevill, D. N.; Kolwyck, K. C.; Weitl, F. L. J . Am. Chem. Soc.
1970, 92, 7300.
(39) Bentley, T. W.; Bowen, C. T.; Parker, W.; Watt, C. I. F. J . Chem.
Soc., Perkin Trans. 2 1980, 1244.
(40) Rappoport, Z.; Kaspi, J . J . Am. Chem. Soc. 1974, 96, 4518.
(41) Fieser, L. F. Experiments in Organic Chemistry, 3rd rev. ed.:
D. C. Heath: Boston, 1957; p 291.
(37) Ta-Shma, R.; Rappoport, Z. Adv. Phys. Org. Chem. 1992, 27,
239 and references therein.
(42) Cheeseman, G. W. H. J . Chem. Soc. 1957, 115.

7262 J . Org. Chem., Vol. 61, No. 21, 1996
Kevill et al.
p-methoxybenzyl chloride. A 73% yield of white crystals was
obtained: mp 117.5-121.5 °C;43 IR (KBr disk) 3010, 2930,
1490, 1420, 1270, 1160, 1030, 640 cm-1; PMR (CD3CN) δ 2.71
Chromosorb W (non-acid-washed) at a programmed temper-
ature span of 100-200 °C, 5° per min, temperature held
initially at 100 °C for 10 min. A Packard instrument (Model
GC-430), with flame ionization detector and microprocessor
integrator, was used. All solutions were allowed to react for
a little over 10 half-lives, and determinations were at least in
duplicate, with an error of ( 5% estimated for the product
ratio.
(s, 6H), 5.76 (s, 1H), 7.2-7.7 (m, 10H). Anal. Calcd for C16
-
H17F3O3S2: C, 50.78; H, 4.53; S, 16.94. Found: C, 50.64; H,
4.57; S, 16.88.
Kin etic P r oced u r es. The kinetic runs and subsequent
multiple regression analyses were performed as previously
described.2
Deter m in a tion of P r od u ct Com p osition . The product
division between benzhydrol and benzhydryl ethyl ether was
determined for solvolysis at 25.0 °C in seven aqueous ethanol
mixtures by use of a response-calibrated GLPC analysis with
nitrogen as the carrier gas and a 3.1 m × 6 mm glass column
packed with OV-17 (3%) on 80-100 mesh DMCS-treated
Ack n ow led gm en t. N. HJ Ismail thanks the Institut
Teknologi MARA (Selangor, Malaysia) for financial
support. The research was supported, in part, by the
award of a Doctoral Research Fellowship to S.W.A. by
the NIU Graduate School. D.N.K. thanks Professor R.
Gompper, Institut fu¨r Organische Chemie, Universita¨t
Mu¨nchen, for hospitality during the time that this paper
was being prepared.
(43) It has been pointed out previously that for sulfonium salts,
melting points are
a poor test of purity. The usually observed
decomposition points depend very much upon the rate of heating:
Saunders, W. H.; Asperger, S. J . Am. Chem. Soc. 1957, 79, 1612.
J O960909T