Solvolysis of benzylic chlorides with extended charge delocalization. α-tert-Butyl(2-naphthyl)methyl, 9-fluorenyl and monosubstituted benzhydryl chlorides

Article abstract of DOI:10.1002/(SICI)1099-1395(199803)11:3<223::AID-POC986>3.0.CO;2-6

Rate constants of solvolysis of α-tert-butyl(2-naphthyl)methyl chloride (1), 9-fluorenyl chloride (2) and a series of monosubstituted benzhydryl chlorides (3) in a wide range of solvents were measured. Grunwald-Winstein-type correlation analysis of log k for 2 and 3 against Y_BnCl, with or without nucleophilicity N, yielded less satisfactory linear correlations than that against log k(l). A new scale of solvent ionizing power, Y_xBnCl for the correlation of solvolytic reactivities of benzylic chlorides with extended charge delocalization based on log k(1) was developed. Application to the mechanistic study suggested the solvolysis of 2 and 4-nitrobenzhydryl chloride were non-limiting. Hammett plots against α⁺ constants exhibited more negative ρ values in less nucleophilic solvents. In a benzhydryl chloride containing a strong deactivating substituent, such as 4-nitro, the positive charge delocalizes mainly over the unsubstituted ring in the cationic transition state. The uneven charge distribution was also confirmed by Mulliken population analysis at the level of the RHF/6-31G*//RHF/3-21G(*) basis set for cations. Comparison of the results of correlation analysis using the equation log(k/k₀) = mY vs the equation log(k/k0) = mY + hI, and using the equation log(k/k₀) = mY +IN vs the equation log(k/ko) = mY + IN + hI indicated the use of Y_BnCl or YxBncl could give a better understanding of solvolytic mechanisms than the combinatorial use of Y_Cl and I.

Full text of DOI:10.1002/(SICI)1099-1395(199803)11:3<223::AID-POC986>3.0.CO;2-6

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)
Solvolysis of benzylic chlorides with extended charge
delocalization. a-tert-Butyl(2-naphthyl)methyl, 9-¯uorenyl
and monosubstituted benzhydryl chlorides
Kwang-Ting Liu,* Yen-Shyi Lin and Meng-Lin Tsao
1Department of Chemistry, National Taiwan University, Taipei 106, Taiwan, Republic of China
Received 24 June 1997; revised 31 July 1997; accepted 16 August 1997
ABSTRACT: Rate constants of solvolysis of a-tert-butyl(2-naphthyl)methyl chloride (1), 9-fluorenyl chloride (2) and
a series of monosubstituted benzhydryl chlorides (3) in a wide range of solvents were measured. Grunwald–Winstein-
type correlation analysis of log k for 2 and 3 against YBnCl, with or without nucleophilicity N, yielded less satisfactory
linear correlations than that against log k(1). A new scale of solvent ionizing power, YxBnCl, for the correlation of
solvolytic reactivities of benzylic chlorides with extended charge delocalization based on log k(1) was developed.
Application to the mechanistic study suggested the solvolysis of 2 and 4-nitrobenzhydryl chloride were non-limiting.
‡
Hammett plots against ꢀ constants exhibited more negative r values in less nucleophilic solvents. In a benzhydryl
chloride containing a strong deactivating substituent, such as 4-nitro, the positive charge delocalizes mainly over the
unsubstituted ring in the cationic transition state. The uneven charge distribution was also confirmed by Mulliken
population analysis at the level of the RHF/6–31G*//RHF/3–21G(*) basis set for cations. Comparison of the results of
correlation analysis using the equation log(k/k0) = mY vs the equation log(k/k0) = mY ‡ hI, and using the equation
log(k/k0) = mY ‡ lN vs the equation log(k/k0) = mY ‡ lN ‡ hI indicated the use of YBnCl or YxBnCl could give a better
understanding of solvolytic mechanisms than the combinatorial use of YCl and I. 1998 John Wiley & Sons, Ltd.
KEYWORDS: Solvolysis; Grunwald–Winstein-type analysis; Hammett-type analysis; YxBnCl scale; 1-aryl-1-
phenylmethyl cations; Mulliken population analysis
INTRODUCTION
ethanol systems by using equation(1),8,10 or from the
improvement of linear correlations by using equation (2)
with an appropriate N scale.11
Later work on the solvolysis of substituted naphthyl-
methyl tosylates indicated the dependence of solvation on
the extent of charge delocalization in the cationic
transition state.12 Another approach by Kevill et al.13
was to add a new term, the aromatic ring parameter I,
based on the solvolysis of (4-methoxybenzyl)dimethyl-
sulfonium ion, to the original Grunwald–Winstein
equations for correcting the observed dispersion:
Correlation analyses of solvent effects on solvolytic
reactivities with single- or dual-parameter Grunwald–
Winstein equation1,2
logꢀk=k0† ˆ mY
logꢀk=k0† ˆ mY ‡ lN
ꢀ1†
ꢀ2†
have been widely employed in mechanistic studies. The
empirical parameter Y, the solvent ionizing power has
been found to be leaving group dependent, and a large
number of YX scales have been determined.3 Observa-
tions of dispersion in log k–YX plots have been
interpreted by a possible variation of solvation effects,4,5
or by mechanistic changes, such as the presence of
general base catalysis6 or ion pair return.7 Recent studies
on the solvolysis of benzylic substrates led to the
necessity of establishing new YBnX scales for benzylic
bromides,8 chlorides,9 p-nitrobenzoates10 and tosylates.11
Nucleophilic solvent intervention could be concluded
from the observation of depression of the points
corresponding to log k measured in ethanol–trifluoro-
logꢀk=k0† ˆ mY ‡ hI
logꢀk=k0† ˆ mY ‡ lN ‡ hI
ꢀ3†
ꢀ4†
More recently, we developed a YxBnBr scale for the
correlation of solvolytic reactivities of bromides, such as
benzhydryl bromides, having extended charge delocali-
zation in the transition state.14 The advantage of using
YBnBr or YxBnBr [equation (1) and (2)] over YBr and I
[equation (3) and (4)] in correlation analysis has also
been demonstrated.14
In an extension of our studies on the solvolysis of other
benzylic substrates with extended delocalization in the
cationic transition state, we found it necessary to
establish a similar YxBnCl scale based on log k values
*Correspondence to: K.-T. Liu, Department of Chemistry, National
Taiwan University, Taipei 106, Taiwan.
Contract/grant sponsor: National Science Council, Republic of China.
1998 John Wiley & Sons, Ltd.
CCC 0894–3230/98/030223–07 $17.50

224
K.-T. LIU, Y.-S. LIN AND M.-L. TSAO
for a-tert-butyl(2-naphthyl)methyl chloride (1). Applica-
tion to the mechanism of solvolysis of 9-fluorenyl
chloride (2) and benzhydryl chlorides (3) is reported in
this paper.
Figure 1. Correlations of log k for (*) 1 and (D) 3d vs YBnCl
RESULTS
and spectral data for all substrates were found to be
compatible with the assigned structures. Conductimetric
or, in some cases, titrimetric first-order rate constants
Chlorides 1–3 were prepared from the corresponding
alcohols by treatment with thionyl chloride in carbon
tetrachloride. Elemental analyses for new compounds
Table 1. Pertinent rate constants for solvolyses of chlorides 1±3
k (25°C) (sꢁ1
)
Solventa
1
2
3ab
3b
3c
3d
3e
100E
90E
80E
70E
60E
90A
80A
70A
60A
7.10 Â 10ꢁ9c
4.84 Â 10ꢁ8c
1.99 Â 10ꢁ7c
5.90 Â 10ꢁ7c
1.71 Â 10ꢁ6c
1.27 Â 10ꢁ3
1.06 Â 10ꢁ2
4.20 Â 10ꢁ2
5.42 Â 10ꢁ5
4.79 Â 10ꢁ4
1.95 Â 10ꢁ3
6.03 Â 10ꢁ3
1.59 Â 10ꢁ2
4.40 Â 10ꢁ6c
7.17 Â 10ꢁ5
5.10 Â 10ꢁ4
2.76 Â 10ꢁ3
1.38 Â 10ꢁ2
8.10 Â 10ꢁ4
3.65 Â 10ꢁ3
1.25 Â 10ꢁ6d
3.56 Â 10ꢁ5
9.02 Â 10ꢁ5
3.02 Â 10ꢁ4
1.20 Â 10ꢁ6c
8.94 Â 10ꢁ6
4.80 Â 10ꢁ5
2.83 Â 10ꢁ4
1.93 Â 10ꢁ5
9.16 Â 10ꢁ5
3.14 Â 10ꢁ4
1.47 Â 10ꢁ2
1.77 Â 10ꢁ3
2.47 Â 10ꢁ4
4.28 Â 10ꢁ8d
3.43 Â 10ꢁ7c
1.30 Â 10ꢁ6c
4.59 Â 10ꢁ6c
2.49 Â 10ꢁ6c
1.06 Â 10ꢁ5c
2.96 Â 10ꢁ5
7.22 Â 10ꢁ5
3.01 Â 10ꢁ7c
9.23 Â 10ꢁ7c
2.93 Â 10ꢁ6c
5.86 Â 10ꢁ9c
5.09 Â 10ꢁ8c
3.19 Â 10ꢁ7c
2.31 Â 10ꢁ6c
4.23 Â 10ꢁ7c
1.48 Â 10ꢁ6c
2.13 Â 10ꢁ5
5.92 Â 10ꢁ5
9.54 Â 10ꢁ6c
1.38 Â 10ꢁ6c
2.08 Â 10ꢁ7c
,
4.84 Â 10ꢁ10c
6.80 Â 10ꢁ9c
4.94 Â 10ꢁ8c
2.72 Â 10ꢁ7c
1.93 Â 10ꢁ6c
1.16 Â 10ꢁ7c
6.34 Â 10ꢁ7c
2.26 Â 10ꢁ6c
2.09 Â 10ꢁ5
2.82 Â 10ꢁ4
2.53 Â 10ꢁ5
1.97 Â 10ꢁ6c
2.74 Â 10ꢁ7c
9.47 Â 10ꢁ5
1.64 Â 10ꢁ3
1.21 Â 10ꢁ2
6.16 Â 10ꢁ2
6.65 Â 10ꢁ1c
1.98 Â 10ꢁ2
9.16 Â 10ꢁ2
3.26 Â 10ꢁ1c
3.43 Â 10ꢁ6c
1.68 Â 10ꢁ5
8.07 Â 10ꢁ5
5.21 Â 10ꢁ6c
2.43 Â 10ꢁ5
8.56 Â 10ꢁ5
6.56 Â 10ꢁ4
2.99 Â 10ꢁ3
3.84 Â 10ꢁ4
6.19 Â 10ꢁ5
3.39 Â 10ꢁ7c
1.95 Â 10ꢁ6c
9.71 Â 10ꢁ6
5.40 Â 10ꢁ7d
2.51 Â 10ꢁ6c
7.91 Â 10ꢁ6c
7.60 Â 10ꢁ5
5.44 Â 10ꢁ5
1.17 Â 10ꢁ5
2.52 Â 10ꢁ6c
50A
100M
90M
80M
60M
100T
80T20E
60T40E
40T60E
1.19 Â 10ꢁ1
1.47 Â 10ꢁ2
2.10 Â 10ꢁ3
7.57 Â 10ꢁ1c
7.03 Â 10ꢁ2
a
E = ethanol; A = acetone; M = methanol; T = trifluoroethanol. Numbers represent percentage composition.
b Ref. 41.
c
From Arrhenius plots of data at other temperatures.
d Ref. 15.
Table 2. Correlation analyses of log k vs YCl and YBnCl
Parameter
1
2
3a
3b
3c
3d
3e
YCl
n
m
ꢀa
r
18
14
13
15
14
14
14
0.876
0.100
0.909
18
0.893
0.121
0.906
14
0.794
0.151
0.847
13
0.763
0.110
0.888
15
0.724
0.108
0.889
14
0.694
0.096
0.902
14
0.578
0.067
0.928
14
YBnCl
n
m
ꢀa
r
0.957
0.028
0.993
0.838
0.064
0.967
1.03
0.048
0.988
0.932
0.04
0.988
0.833
0.039
0.987
0.745
0.028
0.992
0.618
0.061
0.946
a
Standard deviation.
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)

SOLVOLYSIS OF BENZYLIC CHLORIDES
225
Table 3. Statistical analyses of the linearity of log k vs YBnCl for 1 and 3d
Substrate
Solvent
n
m
ꢀa
r
Degrees of freedom
4
jtj value
11.5
CL (%)
1
Aqueous acetone
All others
Aqueous acetone
All others
5
13
3
1.02
0.026
0.007
0.020
0.027
0.999
1.00
1.00
>99.9
>99.9
}
}
0.885
0.921
0.724
3d
6
13.9
11
0.994
a
Standard deviation.
b Confidence limit.
Table 4. YxBnCl values
that there is a separation of the line defined by aqueous
acetone and a line defined by other solvents for 1 and 3d.
This could be confirmed by statistical analysis,17 shown
in Table 3.
Solvent
YxBnCl
Solvent
YxBnCl
100E
90E
80E
70E
60E
90A
80A
70A
60A
ꢁ1.45
ꢁ0.614
0.000
50A
100M
90M
80M
60M
0.987
ꢁ0.234
0.503
1.06
2.02
3.15
2.10
0.996
0.139
Similarly to benzhydryl bromides,14,18 excellent linear
relationships were obtained from log k–log k(1) plots for
3a–d. Therefore, a new YxBnCl scale of solvent ionizing
powers based on log k values for a-tert-butyl(2-
naphthyl)methyl chloride (1) can be developed. The
values are given in Table 4 and the results of correlation
analysis for 2 and 3 against YxBnCl using equation (1) are
summarized in Table 5.
0.472
0.934
ꢁ2.61
ꢁ1.47
ꢁ0.605
0.136
100T
80T20E
60T40E
40T60E
The observed linear relationships for 3a–d against
YxBnCl were confirmed by statistical analysis,17 and
representative examples are given in Table 6 and Fig.
2. For both 2 and 3e, the log k–mYxBnCl plots deviated
from linearity and correlation analyses using the dual-
Table 5. Correlation analyses of log k vs YxBnCl
2
3a
3b
3c
3d
3e
n
14
13
15
14
14
14
m
ꢀa
r
0.875
0.050
0.981
1.02
0.954
0.021
0.997
0.889
0.020
0.997
0.812
0.025
0.994
0.682
0.068
0.945
0.034
0.994
a
Standard deviation.
were measured at least in duplicate. In general, experi-
mental errors within Æ2% were observed. Ethanolysis
and methanolysis rate constants for 3a–c are in good
agreement (within 5%) with literature data.15 A larger
deviation, however, was found for 3b in certain aqueous
acetone and aqueous ethanol solutions.5 Pertinent data
are given in Table 1.
3
Regression analyses of log k values against YCl
showed a poor correlation (r < 0.93) in every case.
Although 1 and 3d gave excellent linear correlations
(r > 0.990) with YBnCl,9b,16 the results for others were less
satisfactory (Table 2). Further, it is obvious from Fig. 1
Figure 2. Correlations of log k for (*) 3c vs YxBnCl
Table 6. Statistical analyses of the linearity of log k vs YxBnCl for 3a and 3b
Substrate
Solvent
n
m
ꢀa
r
Degrees of freedom
10
jt jvalue
0.770
CL (%)b
3a
Aqueous acetone
All others
Aqueous acetone
All others
5
8
3
1.05
1.03
0.860
0.820
0.022
0.068
0.037
0.029
1.00
<80
}
}
0.987
0.999
0.994
3d
3
1.73
<85
11
a
Standard deviation.
b Confidence limit.
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)

226
K.-T. LIU, Y.-S. LIN AND M.-L. TSAO
Table 7. Correlation analyses of log k vs different Y and N
Substrate
Y, N
n
m
ꢀa
l
ꢀa
r
2
Y
BnCl, NT
14
14
13
13
14
14
13
13
1.03
0.991
1.00
0.990
0.869
0.914
0.843
0.912
0.068
0.055
0.056
0.036
0.066
0.085
0.037
0.052
0.295
0.186
0.351
0.261
0.375
0.320
0.473
0.444
0.079
0.063
0.081
0.051
0.082
0.095
0.059
0.074
0.986
0.990
0.988
0.995
0.982
0.973
0.993
0.985
Y
xBnCl, NT
Y
BnCl, NOTs
xBnCl, NOTs
BnCl, NT
xBnCl, NT
Y
3e
Y
Y
Y
BnCl, NOTs
Y
xBnCl, NOTs
a
Standard deviation.
‡
Table 8. Hammett-type analyses against ꢀ constants
Table 10. Calculated atomic charges for cations derived
from 3a±e.
Solvent
n
ꢀa
r
100E
80E
4
5
4
5
5
5
5
5
4
4
5
ꢁ4.11
ꢁ4.20
ꢁ4.06
ꢁ4.21
ꢁ4.18
ꢁ4.40
ꢁ4.20
ꢁ4.19
ꢁ4.26
ꢁ5.12
ꢁ4.93
0.077
0.091
0.279
0.088
0.099
0.186
0.060
0.064
0.040
0.266
0.117
1.00
0.999
0.995
0.999
0.999
0.997
1.00
1.00
1.00
0.997
0.999
70E
70A
60A
50A
100M
90M
80M
80T20E
60T40E
X
Carbon
4-Me
0.221
H
3-Cl
3-CF3
0.243
4-NO2
0.250
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
Pha
0.232
0.241
ꢁ0.055 ꢁ0.048
ꢁ0.042 ꢁ0.054 ꢁ0.034
0.120
0.020
0.223
0.007
0.121
0.104
0.038
0.139
0.041
0.110
0.138
0.146
0.078
0.140
a
Standard deviation.
ꢁ0.079 ꢁ0.098
0.171
0.052
0.101
0.180 ꢁ0.157
0.044
0.114
0.146
0.089
Table 9. Comparison of k values in solvents with similar Y
but different N values
ꢁ0.045 ꢁ0.048
ꢁ0.050 ꢁ0.049 ꢁ0.051
0.088
0.036
0.126
0.040
0.097
0.342
0.436
0.111
0.041
0.140
0.037
0.105
0.385
0.383
0.114
0.039
0.151
0.043
0.121
0.418
0.341
0.114
0.040
0.153
0.044
0.124
0.425
0.332
0.142
0.045
0.173
0.043
0.136
0.489
0.261
k (25°C) (sꢁ1 a
)
Substrate
50Ab
60T40Ec
Rate ratio
3a
3b
3c
3d
3e
6.65 Â 10ꢁ1
1.38 Â 10ꢁ2
2.83 Â 10ꢁ4
8.07 Â 10ꢁ5
9.71 Â 10ꢁ6
7.57 Â 10ꢁ1
1.47 Â 10ꢁ2
2.47 Â 10ꢁ4
6.19 Â 10ꢁ5
2.52 Â 10ꢁ6
0.878
0.938
1.15
Arb
a
1.30
Total charge on the phenyl ring.
Total charge on the aryl ring.
b
3.61
a
From Table 1.
b YxBnCl = 0.987; YBnCl = 1.23.
c
YxBnCl = 0.996; YBnCl = 1.18.
results of Mulliken population analyses22 at the level of
the RHF/6–31G*//RHF/3–21G(*) basis set are presented
in Table 10. Similar results were obtained from molecular
electrostatic potential23 and natural population analy-
sis.24
Treatment of rate data for 1–3 with the multi-
parameter equations (3) and (4) yielded the results given
in Table 11.
19
parameter equation (2) against YxBnCl or YBnCl, and NT
or NOTs20 [equation (2)] were carried out. The results are
given in Table 7.
‡
Hammett plots of log k values for 3a–e against ꢀ
constants;21
‡
logꢀk=k0† ˆ ꢀ ꢁ
ꢀ5†
gave excellent linear correlations in a variety of solvent
systems (Table 8). Comparisons of solvolytic reactivities
in 60% trifluoroethanol–40% ethanol and in 50% acetone
are illustrated in Table 9. To estimate the positive charge
distribution in carbocation, a number of ab initio calcu-
lations for cations derived from 3 were carried out. The
DISCUSSION
Solvolysis of secondary a-tert-butylbenzyl substrates was
in general considered to proceed via a limiting SN1
mechanism,11,12,14,18,25–27 with some exceptions such as
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)

SOLVOLYSIS OF BENZYLIC CHLORIDES
Table 11. Correlation analysis against YCl, N and I
227
Substrate
Parameter
n
m
ꢀa
hb
ꢀa
lc
ꢀa
r
1
YCl, I
YCl, I, NT
YCl, I, NOTs
YCl, I
YCl, I, NT
YCl, I, NOTs
YCl, I
17
17
15
14
14
13
12
14
14
13
13
13
12
0.874
0.926
0.878
0.863
1.07
0.052
0.070
0.074
0.085
0.093
0.069
0.065
0.057
0.057
0.065
0.071
0.103
0.082
1.72
1.83
1.75
1.03
1.37
1.50
1.85
1.45
1.39
1.03
0.405
0.777
0.949
0.247
0.263
0.311
0.283
0.242
0.231
0.253
0.236
0.245
0.224
0.328
0.340
0.343
0.980
0.982
0.979
0.958
0.979
0.986
0.979
0.976
0.973
0.964
0.933
0.954
0.968
0.122
0.113
0.109
0.142
2
0.314
0.400
0.103
0.103
1.04
3a
3b
3c
3d
3e
0.898
0.801
0.726
0.705
0.581
0.749
0.749
YCl, I
YCl, I
YCl, I
YCl, I
YCl, I, NT
YCl, I, NOTs
0.277
0.393
0.136
0.150
a
Standard deviation.
phenolysis of a-tert-butyl(4-methoxy)benzyl p-nitro-
benzoate28 and acetolysis of the corresponding chlor-
ide.29 Similarly to the corresponding bromide14 and
tosylate,12 the solvolysis of a-tert-butyl(2-naphthyl)-
methyl chloride (1) is also likely to involve a limiting
mechanism with positive charge delocalization over both
rings in the transition state. Although the regression
analysis in Table 2 indicated a poor linear correlation
with YCl (r = 0.909) but an excellent linear correlation
with YBnCl (r = 0.993) for 1, an obvious splitting of lines,
one for aqueous acetone and one for all others, could be
found (Fig. 1). The separation of lines can be confirmed
by statistical analysis17 to have a confidence level greater
than 99.99% (Table 3). Similar analysis was also found to
be applicable for 3d. A difference in the solvation of the
cationic transition states with different extents of charge
delocalization, such as (2-naphthyl)methyl cation (4) and
benzyl cation (5), is the likely explanation as has been
suggested in our previous work.12,14
also the solvolysis of 9-fluorenyl bromide34 and tosy-
late.35 Our recent studies, however, indicated significant
nucleophilic solvent intervention in the solvolysis of
monosubstituted benzhydryl bromides containing a
strong electron-withdrawing group,14,18 and in the case
of 9-fluorenyl bromide.30 Moreover, a new YxBnBr scale
was found to be necessary to correlate the solvolytic
reactivities of 2 and 3.14 It seems to be justified to solve
the present problem by a similar approach. Excellent
linear relationships were obtained from log k vs log k(1)
plots for 3a–d, and a new YxBnCl scale of solvent ionizing
powers based on log k values for 1 could be developed
(Table 4).
The results of regression analyses of log k values for 2
and 3a–e against YxBnCl using equation (1) are given in
Table 5. Excellent linear correlations (r = 0.994–0.997)
were found for 3a–d. The results of statistical analysis17
in Table 6 confirm the linearity of the log k vs YxBnCl plot
for 3a and d. The limiting SN1 mechanism in the
solvolysis of 3a–d is confirmed by the observed m values
of 0.8–1. An example of a Grunwald–Winstein plot
[equation (1)] is illustrated in Fig. 2. The application of
equation (2) was then examined for 2 and 3e. It is
interesting that despite the claim of a general preference
for using NT in dual-parameter correlations,19 the data in
Table 7 indicate that the use of NOTs20 would give a better
linear correlation and offer a more plausible explanation.
For instance, the combination of NOTs and the newly
developed YxBnCl values yielded an excellent linear
relationship (r = 0.995) with a reasonable l value
(0.261) for 2. Extensive charge delocalization through
the whole ring system and the presence of significant
nucleophilic solvent intervention at the reaction center in
the transition state are likely, as have been found in the
solvolysis of 9-fluorenyl bromide.30
Interesting results were also observed from the
solvolysis of 9-fluorenyl chloride (2) and monosubsti-
tuted benzhydryl chlorides (3). Like the corresponding
bromides,14,18,30 good linear relationships could not be
found for log k vs YCl [equation (1)], as was shown in
Table 2. Deviations from linear correlations with YBnCl
were observed for 2,3a–c and 3e. A unimolecular
mechanism has generally been accepted for interpreting
the solvolysis of benzhydryl halides,15 p-nitrobenzo-
ates,31 mesylates32 and dimethylsulfonium ion,33 and
On the other hand, the use of either NT or NOTs with
YxBnCl in equation (2) gave less satisfactory linear
correlations (r = 0.97–0.98 in Table 7) for 3e. Regression
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)

228
K.-T. LIU, Y.-S. LIN AND M.-L. TSAO
tions derived from 3a–e are summarized in Table 10. The
positive charge on C-1 is nearly the same for different
cations. However, the charge distribution in the mono-
substituted aryl ring decreases from 0.436 for 4-methyl to
0.383 for 4-hydrogen, 0.341 for 3-chloro, 0.332 for 3-
trifluoromethyl and 0.261 for 4-nitro. In contrast, an
increasing order of change in charge distribution is
realized in the unsubstituted phenyl ring. More important,
more positive charge is found in the aryl ring (0.436) than
in the phenyl ring (0.342) for the cation containing an
electron-donating 4-methyl group, whereas more positive
charge is found in the phenyl ring (0.489) than in the aryl
ring (0.261) for the cation containing an electron-
withdrawing 4-nitro group. These results of calculations
are in harmony with the conclusion drawn from the
correlation analyses of solvolytic rate data (see above). In
addition, molecular electrostatic potential23 and natural
population analysis24 gave similar results.37
The dual-parameter equation (3) and the multi-
parameter equation (4) were also applied to examine
the solvolytic behavior of 1–3. Both NT and NOTs scales
were employed. Since in Table 11 only moderately good
linear correlations (r = 0.95–0.98) with YCl and I, or with
YCl, I and NT, were obtained, no close relationship
between the outcome of these regression analyses and
solvolytic reactivities could be found. Therefore, in
addition to the deficiencies already discussed,38 the
present results suggest another disadvantage for using
the aromatic ring parameter I in the correlation analysis
of reactivities.
‡
Figure 3. Correlations of log k for 3a±e vs ꢀ constants
analysis using NOTs and YBnCl resulted in a linear
relationship with r = 0.993. Again, similarly to the
corresponding bromide,14 the present finding of a better
correlation with YBnCl than with YxBnCl may be attributed
to the smaller extent of positive charge distribution. It is
likely that in a benzhydryl chloride containing strong
electron-withdrawing group, such as 4-nitro, the charge
delocalizes mainly over the unsubstituted phenyl ring in
the cationic transition state. Consequently, YBnCl, instead
of YxBnCl, is the choice for the correlations.
Previous work on the solvolysis of monosubstituted
benzhydryl chlorides showed an excellent linear Ham-
‡
mett-type plot against ꢀ constants, with r = ꢁ4.13
(propan-2-ol), ꢁ4.15 (ethanol) and ꢁ4.22 (methanol).15
In the present study, Table 8 also indicates linear log k vs
‡
ꢀ
plots [equation (5)] with similar r values in
EXPERIMENTAL
nucleophilic solvents (ꢁ4.10 to ꢁ4.26). The r value
became more negative (ꢁ4.93 to ꢁ5.12) in the less
nucleophilic ethanol–trifluoroethanol solvent systems.
Representative examples in Fig. 3 demonstrated increas-
ing reactivities for substrates with a more electron-
donating group. Moreover, comparison of the rate
constants k in solvents of similar solvent ionizing powers
but different solvent nucleophilicities (Table 9) also
indicates a greater rate enhancement in the more
nucleophilic solvent, 50A, for less reactive substrates.
The large influence of solvent nucleophilicity in the case
of 3e might account for the small m value for 3e in Table
5. Further, the present results give an additional example
Capillary melting points are uncorrected. Proton and
carbon-13 nuclear magnetic resonance spectra were
measured on a Bruker Model AM-300 or AC-200
instrument using tetramethylsilane as internal standard.
Infrared spectra were measured using a Perkin-Elmer
Model 983G spectrometer. Elemental analyses were
performed by the Microanalytical Laboratory at National
Taiwan University.
Materials. Solvents for kinetic studies were of spectral or
reagent grade and were purified according to standard
methods.39 Reagent-grade solvents and chemicals were
used for general purposes. Doubly deionized water was
used to prepare aqueous solvent systems for solvolysis.
a-tert-Butyl(2-naphthyl)methyl chloride (1) was pre-
pared from the reaction of thionyl chloride with the
corresponding alcohol, which had been obtained from the
Grignard reaction of 2-naphthylmagnesium bromide with
2,2-dimethylpropanoyl chloride in tetrahydrofuran. A
white solid of m.p. 74–75°C and correct elemental
analysis (found, C 77.12, H 7.22%; C15H17Cl requires C
77.41, H 7.36%) was obtained.
‡
that the linear relationships against ꢀ constants
[equation (5)] are not sensitive to nucleophilic solvent
intervention, as has already been noted in tert-cumyl
systems.36
To confirm the above-mentioned difference in positive
charge distribution for carbocations generated from
different substrates, a number of ab initio calculations
were carried out. Charges on hydrogens were summed
into the attached carbons. The results of Mulliken
population analyses22 at the level of the RHF/6–31G*//
RHF/3–21G(*) basis set for 1-aryl-1-phenylmethyl ca-
9-Fluorenyl chloride (2), m.p. 149–150°C (lit.40 147–
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)

SOLVOLYSIS OF BENZYLIC CHLORIDES
229
11. K.-T. Liu, J.-S. Yang, S.-M. Chang, Y.-S. Lin, H.-C. Sheu and M.-
L. Tsao. J. Org. Chem. 57, 3041 (1992).
12. K.-T. Liu, H.-Y. Hsu and J.-S. Yang. Tetrahedron Lett. 33, 3327
(1992).
148°C) was prepared by the treatment of commercially
available 9-fluorenol with thionyl chloride in carbon
tetrachloride.
Monosubstituted benzhydryl chlorides 3a–e were
obtained according to the following procedures. Com-
mercially available benzophenones were reduced by
sodium borohydride to the alcohols, which were con-
verted into the corresponding chlorides by treatment with
thionyl chloride followed by purification by column
chromatography or vacuum distillation. 4-Methylbenz-
hydryl chloride (3a), benzhydryl chloride (3b), 3-chloro-
benzhydryl chloride (3c) and 4-nitrobenzhydryl chloride
(3e) are known compounds.15 An acceptable elemental
analysis (found, C 61.91, H 3.67%; C14H10F3Cl requires
C 62.12, H 3.70%) was obtained for the new compound
3-trifluoromethylbenzhydryl chloride (3d).
13. D. N. Kevill, N. HJ Ismail and M. D’Souza J. J. Org. Chem. 59,
6303 (1994).
14. K.-T. Liu, C.-P. Chin, Y.-S. Lin and M.-L. Tsao. J. Chem. Res. (S)
18 (1997).
15. S. Nishida. J. Org. Chem. 32, 2692 (1967).
16. K.-T. Liu, L.-W. Chang, D.-G. Yu, P.-S. Chen and J.-T. Fan. J.
Phys. Org. Chem. 10, 879 (1997).
17. J. C. Miller and J. N. Miller. Statistics for Analytical Chemistry,
2nd ed., Chapt. 3. Wiley, New York (1988).
18. K.-T. Liu, C.-P. Chin, Y.-S. Lin and M.-L. Tsao. Tetrahedron Lett.
36, 6919 (1995).
19. (a) D. N. Kevill and S. W. Anderson. J. Chem. Res. (S) 356 (1991);
(b) D. N. Kevill and M. J. D’Souza. J. Chem. Res. (S) 174 (1993).
20. M. Fujio, T. Susuki, M. Goto, Y. Tsuji, K. Yatsugi, Y. Saeki, S. H.
Kim and Y. Tsuno. Bull. Chem. Soc. Jpn. 67, 2233 (1994).
21. H. C. Brown and Y. Okamoto. J. Am. Chem. Soc. 80, 1913 (1958).
22. R. S. Mulliken. J. Chem. Phys. 78, 1833 (1955).
23. (a) L. E. Chirlian and M. M. Francl. J. Comput. Chem. 8, 894
(1987); (b) C. M. Breneman and K. B. Wiberg. J. Comput. Chem.
11, 361 (1990).
Infrared, proton and carbon-13 NMR data for 1–3 were
found to be consistent with the assigned structures.
24. (a) A. E. Reed and F. Weinhold. J. Chem. Phys. 78, 4066 (1983);
(b) J. E. Carpenter and F. Weinhold. J. Mol. Struct. (Theochem)
169, 41 (1988).
Kinetic Measurements. Conductimetric and titrimetric
rate constants were measured as described.9b
25. S. Winstein and B. K. Morse. J. Am. Chem. Soc., 74, 1133 (1952).
26. (a) Y. Tsuji, M. Fujio and Y. Tsuno. Bull, Chem. Soc. Jpn., 63, 856
(1990); (b) Y. Tsuji, M. Fujio and Y. Tsuno. Tetrhedron Lett., 33,
349 (1992).
Acknowledgment
27. K.-T. Liu and C.-S. Tang. J. Org. Chem., 61, 1523 (1996).
28. T. Kinoshita, K. Komatsu, K. Ikai, T. Kashimura, S. Tanikawa, A.
Hatanaka and K. Okamoto. J. Chem. Soc., Perkin Trans. 2 1875
(1988).
Financial support from the National Science Council,
Republic of China, is gratefully acknowledged.
29. T. Kinoshita, H. Ueda and K. Takeuchi. J. Chem. Soc., Perkin
Trans. 2, 603 (1993).
30. K.-T. Liu and Y. S. Lin. Tetrahedron Lett. 38, 1419 (1997).
31. D. J. McLennan and P. L. J. Martin. J. Chem. Soc., Perkin Trans. 2,
1091 (1982).
32. T. W. Bentley, M. Christl and S. Norman. J. Org. Chem., 56, 6238
(1991).
REFERENCES
1. E. Grunwald and S. Winstein. J. Am. Chem. Soc. 70, 846 (1948).
2. S. Winstein, E. Grunwald and H. W. Jones. J. Am Chem. Soc. 73,
2700 (1951).
3. T. W. Bentley and G. Llewellyn. G. Prog. Phys. Org. Chem. 17,
121 (1990).
4. A. H. Fainberg and S. Winstein S. J. Am. Chem. Soc. 79, 1602
(1957).
5. C. A. Bunton, M. M. Mhala and J. R. Moffatt. J. Org. Chem. 49,
3639 (1984).
6. T. W. Bentley and H. C. Harris. J. Org. Chem. 53, 724. (1988).
7. S. Winstein, A. H. Fainberg and E. Grunwald. J. Am. Chem. Soc.
79, 4146 (1957).
8. K.-T. Liu and H.-C. Sheu. J. Chin. Chem. Soc. (Taipei) 38, 29
(1991).
9. (a) K.-T. Liu, H.-C. Sheu, H.-I Chen, P-F. Chiu and C.-R. Hu.
Tetrahedron Lett. 31, 3611 (1990); (b) K.-T. Liu and H.-C. Sheu. J.
Org. Chem. 56, 3021 (1991).
10. K.-T. Liu, H.-I Chen and C.-P. Chin. J. Phys. Org. Chem. 4, 463
(1991).
33. D. N. Kevill, S. W. Anderson and N. H. J. Ismail. J. Org. Chem.,
61, 7256 (1996).
34. G. W. Cowell and A. Ledwith. A.J. Chem. Soc. B 695 (1967).
35. G. W. Cowell, T. D. George, A. Ledwith and D. G. Morris. J.
Chem. Soc. B 1169 (1966).
36. (a) K.-T. Liu, L.-W. Chang and P.-S. Chen. J. Org. Chem 57, 4791
(1992). (b) K.-T. Liu, P.-S. Chen, C.-R. Hu and H.-C. Sheu. J.
Phys. Org. Chem. 6, 122, 433 (1993).
37. M.-L. Tsao and K.-T. Liu. unpublished results.
38. K.-T. Liu. J. Chem. Soc., Perkin Trans. 2, 327 (1996).
39. D. D. Perrin and W. L. F. Armarego. Purification of Laboratory
Chemicals, 3rd ed. Pergamon Press, New York, (1988).
40. R. Bolton, N. B. Chapman and J. Shoter. J. Chem. Soc. 1895
(1964).
41. J.-T. Fan. MS Thesis, Department of Chemistry, National Taiwan
University (1996).
1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 223–229 (1998)