Effect of the leaving group solvation on solvolytic behavior of benzhydryl derivatives

Article abstract of DOI:10.1002/poc.1508

An effect of the leaving group (LG) solvation on reactivity of benzhydryl derivatives in S_N1 reactions has been investigated by using X, Y-substituted benzhydryl phenyl carbonates, methyl carbonates, 3, 5-dinitrobenzoates (DNB), and the corresponding benzhydryl chlorides as reference compounds. Reaction constants (s_f) derived from LFER equation log k (25°C)=sf (N_f+E_f) indicate that s _f parameters of carbonates and DNBs decrease as the fraction of the water in a given solvent/water mixture increases, while those of chlorides remain unchanged. This phenomenon is due to less important solvation and less charge separation in the TS. Effects of the solvents on the reaction rates were analyzed by Grunwald-Winstein correlations using various solvent-ionizing power scales. The m values obtained for carbonates and DNBs are considerably smaller than the m values for chlorides. Also, the solvolysis rate constants of substrates that have stronger electrofuges are less influenced by solvent (lower m) than those with weaker electrofuges. Values of m parameters obtained for a given substrate in a given binary solvent system correlate well with the electrofugality of the generated benzhydrylium ion. Abscissa at which m=0 represents the extrapolated critical electrofugality Ecrit f of the substrates whose solvolysis rates should not depend on the water fraction in the aqueous/organic solvent mixtures. Similar values for the critical electrofugality have also been obtained from extrapolated logk versus Ef plots. Copyright

Full text of DOI:10.1002/poc.1508

Research Article
Received: 13 October 2008,
Revised: 17 November 2008,
Accepted: 20 November 2008,
Published online in Wiley InterScience: 14 January 2009
(www.interscience.wiley.com) DOI 10.1002/poc.1508
Effect of the leaving group solvation on
solvolytic behavior of benzhydryl derivatives
Bernard Denegria and Olga Kronjaa
*
An effect of the leaving group (LG) solvation on reactivity of benzhydryl derivatives in SN1 reactions has been
investigated by using X,Y-substituted benzhydryl phenyl carbonates, methyl carbonates, 3,5-dinitrobenzoates (DNB),
and the corresponding benzhydryl chlorides as reference compounds. Reaction constants (sf) derived from LFER
equation log k (25 -C) ¼ sf(Nf R Ef) indicate that sf parameters of carbonates and DNBs decrease as the fraction of the
water in a given solvent/water mixture increases, while those of chlorides remain unchanged. This phenomenon is due
to less important solvation and less charge separation in the TS. Effects of the solvents on the reaction rates were
analyzed by Grunwald–Winstein correlations using various solvent-ionizing power scales. The m values obtained for
carbonates and DNBs are considerably smaller than the m values for chlorides. Also, the solvolysis rate constants of
substrates that have stronger electrofuges are less influenced by solvent (lower m) than those with weaker
electrofuges. Values of m parameters obtained for a given substrate in a given binary solvent system correlate well
with the electrofugality of the generated benzhydrylium ion. Abscissa at which m ¼ 0 represents the extrapolated
critical electrofugality Efcrit of the substrates whose solvolysis rates should not depend on the water fraction in the
aqueous/organic solvent mixtures. Similar values for the critical electrofugality have also been obtained from
extrapolated logk versus Ef plots. Copyright ß 2009 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this article.
Keywords: LFER reaction constant; electrofugality; nucleofugality; solvent effect; leaving group; benzhydryl derivatives;
critical electrofugality; kinetics
fundamentals for sþ values, so it is not surprising that Ef
INTRODUCTION
correlates with sþ (Ef ¼ 4.39 Ssþ ꢀ 6.14, r2 ¼ 0.992).[2,3] Con-
Nucleofugality of phenyl carbonate (C6H5OCOꢀ2 ; PhCarb) and
methyl carbonate (CH3OCOꢀ2 ; MeCarb) in SN1 reactions have
recently been examined thoroughly by studying the solvolysis
rates of the substituted benzhydryl carbonates in numerous
solvents and solvent mixtures.[1] It has been demonstrated that
excellent correlations of log k versus Ef were obtained
(0.996 ꢁ r ꢁ 1.000) if the special case of LFER was applied, which
has been developed on the benzhydrylium system.[2–5] In this
approach, solvolysis rate constants of benzhydrylium tosylates,
bromides, chlorides, trifluoroacetates, 3,5-dinitrobenzoates, and
4-nitrobenzoates were subjected to the least-squares fit accord-
ing to the correlation eqn (1)
sequently, slope parameters (rþ and sf) measure the same
phenomenon.[6] Also, the origin of the substituent effects of Ef is
the same as that which accounts for the magnitude of sþ.
However, the Hammett–Brown correlation is poor if it is applied
to the unsymmetrically substituted benzhydryl and other
a-R-diarylmethyl systems, due to nonadditivity and nonlinearity.
Therefore, it is justified to use the sf values instead of the rþ for
evaluation of the results obtained with benzhydryl substrates,
particularly since excellent correlation have been obtained for all
logk versus Ef plots studied.[1,2–5]
The outstanding feature of solvolysis of phenyl and methyl
carbonates is that the slopes of the log k versus Ef plots (sf) are
generally lower than those of the corresponding chlorides, and
they decrease as the fraction of the water in a given solvent/water
mixture increases. Also, sf parameters for the series of phenyl
carbonates are lower than for the series of methyl carbonates
(Table 1). On the other hand, correlations of numerous kinetic
data for solvolysis of benzhydryl chlorides in different solvents
give parallel linear plots with slopes of unity (Table 1) i.e., sf
log kð25 ꢂCÞ ¼ sf ðEf þ Nf Þ
(1)
where k is the first order rate constant of the SN1 reaction, sf
(slope of the correlation line) and Nf (negative intercept on the
abscissa) are the nucleofuge specific parameters, and Ef is the
electrofugality parameter. Ef is an independent parameter
determined with the substituents on the benzhydryl system,
while Nf and sf are defined for a combination of the leaving group
(LG) and the solvent. Predefined parameters are sf ¼ 1.00 for
chloride nucleofuge in pure ethanol and Ef ¼ 0.00 for dianisyl-
carbenium ion (1). The electrofugality parameters (Ef) of
*
Correspondence to: O. Kronja, Faculty of Pharmacy and Biochemistry,
ˇ ´
University of Zagreb, Ante Kovacica 1, 10000 Zagreb, Croatia.
E-mail: kronja@pharma.hr
X,Y-substituted benzhydrylium ions are taken from Reference [2,3]
.
If one compares the Hammett–Brown correlation
(log k ¼ rþsþ þ log k0) with eqn (1), it is evident that the
fundamentals of Ef parameters are essentially the same as the
a
B. Denegri, O. Kronja
Faculty of Pharmacy and Biochemistry, University of Zagreb, Ante Kovacica 1,
10000 Zagreb, Croatia
ˇ ´
J. Phys. Org. Chem. 2009, 22 495–503
Copyright ß 2009 John Wiley & Sons, Ltd.

B. DENEGRI AND O. KRONJA
Table 1. Solvolysis rate constants and activation parameters of some X,Y-substituted benzhydryl phenyl carbonates, methyl
carbonates, 3,5-dinitrobenzoates, and chlorides in various solvents
Compounda
Solventb
90E10W
t (8C)
k (sꢀ1 c
)
DHz (kJ molꢀ1 d
)
DSz (J Kꢀ1 molꢀ1 d
)
3-PhCarb
15
25
35
20
25
35
40
25
35
40
45
15
25
35
15
25
35
10
15
25
10
15
20
25
25
35
45
25
25
25
35
45
25
40
50
25
35
45
25
45
55
50
60
70
15
25
35
30
35
42
50
35
45
(2.33 ꢃ 0.01) ꢄ 10ꢀ3,e
(7.17 ꢃ 0.13) ꢄ 10ꢀ3,e,g
(2.08 ꢃ 0.01) ꢄ 10ꢀ2,e
(7.08 ꢃ 0.08) ꢄ 10ꢀ4,e
(1.27 ꢃ 0.03) ꢄ 10ꢀ3,e,g
(4.06 ꢃ 0.07) ꢄ 10ꢀ3,e
(7.10 ꢃ 0.00) ꢄ 10ꢀ3,e
(7.22 ꢃ 0.17) ꢄ 10ꢀ5e,g
(2.58 ꢃ 0.01) ꢄ 10ꢀ4,e
(4.83 ꢃ 0.11) ꢄ 10ꢀ4,e
(8.59 ꢃ 0.09) ꢄ 10ꢀ4,e
(1.51 ꢃ 0.03) ꢄ 10ꢀ3,e
(5.00 ꢃ 0.15) ꢄ 10ꢀ3,e,g
(1.51 ꢃ 0.03) ꢄ 10ꢀ2,e
(1.12 ꢃ 0.01) ꢄ 10ꢀ4,e
(4.17 ꢃ 0.10) ꢄ 10ꢀ4,e,g
(1.48 ꢃ 0.01) ꢄ 10ꢀ3,e
(1.68 ꢃ 0.01) ꢄ 10ꢀ3,e
(3.00 ꢃ 0.07) ꢄ 10ꢀ3,e
(8.46 ꢃ 0.06) ꢄ 10ꢀ3,e,f
(5.27 ꢃ 0.00) ꢄ 10ꢀ3,e
(9.30 ꢃ 0.00) ꢄ 10ꢀ3,e
(1.56 ꢃ 0.01) ꢄ 10ꢀ2,e
(2.55 ꢃ 0.05) ꢄ 10ꢀ2,e,f
(1.53 ꢃ 0.05) ꢄ 10ꢀ3,e,f
(4.84 ꢃ 0.10) ꢄ 10ꢀ3,e
(1.41 ꢃ 0.03) ꢄ 10ꢀ2,e,f
(2.39 ꢃ 0.06) ꢄ 10ꢀ2,e
(3.11 ꢃ 0.06) ꢄ 10ꢀ3,e
(8.83 ꢃ 0.07) ꢄ 10ꢀ4,e
(2.86 ꢃ 0.08) ꢄ 10ꢀ3,e
(8.48 ꢃ 0.17) ꢄ 10ꢀ3,e
(1.54 ꢃ 0.02) ꢄ 10ꢀ4,e
(8.82 ꢃ 0.10) ꢄ 10ꢀ4,e
(2.57 ꢃ 0.09) ꢄ 10ꢀ3,e
(1.98 ꢃ 0.05) ꢄ 10ꢀ3,f,h
(6.16 ꢃ 0.00) ꢄ 10ꢀ3,f
(1.76 ꢃ 0.04) ꢄ 10ꢀ2,f
(3.50 ꢃ 0.10) ꢄ 10ꢀ4,f,h
(3.51 ꢃ 0.06) ꢄ 10ꢀ3,f
(9.59 ꢃ 0.17) ꢄ 10ꢀ3,f
(4.87 ꢃ 0.10) ꢄ 10ꢀ4,f
(1.49 ꢃ 0.01) ꢄ 10ꢀ3,f
(4.13 ꢃ 0.06) ꢄ 10ꢀ3,f
(8.87 ꢃ 0.01) ꢄ 10ꢀ4,e
(2.99 ꢃ 0.03) ꢄ 10ꢀ3,h
(8.91 ꢃ 0.01) ꢄ 10ꢀ3,e
(2.14 ꢃ 0.01) ꢄ 10ꢀ4,e
(3.95 ꢃ 0.03) ꢄ 10ꢀ4,e
(9.00 ꢃ 0.14) ꢄ 10ꢀ4,e
(2.24 ꢃ 0.02) ꢄ 10ꢀ3,e
(6.62 ꢃ 0.16) ꢄ 10ꢀ5,e,h
(2.19 ꢃ 0.03) ꢄ 10ꢀ4,e,h
78.3 ꢃ 0.3
85.6 ꢃ 0.6
ꢀ23.2 ꢃ 1.0
ꢀ13.1 ꢃ 1.9
4-PhCarb
5-PhCarb
90E10W
90E10W
95.3 ꢃ 0.5
ꢀ4.6 ꢃ 1.6
4-PhCarb
5-PhCarb
1-MeCarb
1-MeCarb
70E30W
70E30W
90E10W
70E30W
82.5 ꢃ 0.3
92.8 ꢃ 0.8
73.0 ꢃ 1.2
71.3 ꢃ 0.9
ꢀ12.2 ꢃ 1.1
þ 1.7 ꢃ 2.7
ꢀ39.7 ꢃ 4.1
ꢀ36.3 ꢃ 3.2
3-MeCarb
70E30W
85.0 ꢃ 0.3
ꢀ13.6 ꢃ 0.8
1-DNB
2-DNB
3-DNB
90E10W
90E10W
86.7 ꢃ 0.4
87.6 ꢃ 0.1
83.6 ꢃ 0.4
87.4 ꢃ 0.9
95.8 ꢃ 0.9
82.7 ꢃ 1.0
93.0 ꢃ 0.2
ꢀ12.7 ꢃ 1.2
ꢀ24.0 ꢃ 0.4
ꢀ16.3 ꢃ 1.1
ꢀ18.0 ꢃ 2.9
ꢀ12.5 ꢃ 2.8
ꢀ16.0 ꢃ 3.4
ꢀ8.6 ꢃ 0.8
4-DNB
3-DNB
4-DNB
5-DNB
1-DNB
3-DNB
90E10W
80E20W
80E20W
80E20W
80A20W
80A20W
4-DNB
94.7 ꢃ 0.2
ꢀ17.8 ꢃ 0.5
(Continues)
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 495–503

EFFECT OF THE LEAVING GROUP SOLVATION
Table 1. (Continued)
Compounda
Solventb
t (8C)
k (sꢀ1 c
)
DHz (kJ molꢀ1 d
)
DSz (J Kꢀ1 molꢀ1 d
)
50
25
25
25
25
15
25
30
25
25
35
40
45
50
25
35
45
ꢀ10
ꢀ7
5
(3.86 ꢃ 0.03) ꢄ 10ꢀ4,e,h
(6.35 ꢃ 0.08) ꢄ 10ꢀ3,e
(7.25 ꢃ 0.10) ꢄ 10ꢀ4,e
(2.82 ꢃ 0.03) ꢄ 10ꢀ4,e
(5.60 ꢃ 0.02) ꢄ 10ꢀ5,e
(3.65 ꢃ 0.01) ꢄ 10ꢀ3,e
(1.20 ꢃ 0.02) ꢄ 10ꢀ2,e
(2.11 ꢃ 0.02) ꢄ 10ꢀ3,e
(1.52 ꢃ 0.03) ꢄ 10ꢀ3,e
(6.74 ꢃ 0.11) ꢄ 10ꢀ4,e
(2.14 ꢃ 0.01) ꢄ 10ꢀ3,e
(3.81 ꢃ 0.00) ꢄ 10ꢀ3,e
(6.66 ꢃ 0.08) ꢄ 10ꢀ3,e
(1.09 ꢃ 0.01) ꢄ 10ꢀ2,e
(1.38 ꢃ 0.02) ꢄ 10ꢀ4,e
(4.61 ꢃ 0.12) ꢄ 10ꢀ4,e
(1.43 ꢃ 0.00) ꢄ 10ꢀ3,e
(3.65 ꢃ 0.01) ꢄ 10ꢀ3
(5.23 ꢃ 0.00) ꢄ 10ꢀ3
(2.28 ꢃ 0.00) ꢄ 10ꢀ2
(3.27 ꢃ 0.09) ꢄ 10ꢀ2
(7.17 ꢃ 0.01) ꢄ 10ꢀ3
(2.64 ꢃ 0.04) ꢄ 10ꢀ2
(4.98 ꢃ 0.01) ꢄ 10ꢀ2
1-DNB
2-DNB
3-DNB
4-DNB
1-DNB
70A30W
60A40W
82.5 ꢃ 0.1
87.1 ꢃ 0.7
ꢀ5.0 ꢃ 0.2
2-DNB
3-DNB
60A40W
60A40W
ꢀ13.5 ꢃ 2.2
4-DNB
5-Cl
60A40W
90E10W
89.6 ꢃ 0.0
72.8 ꢃ 0.6
ꢀ18.2 ꢃ 0.1
ꢀ13.8 ꢃ 2.2
8
5-Cl
80E20W
80A20W
ꢀ15
ꢀ5
0
73.4 ꢃ 0.5
ꢀ0.3 ꢃ 2.0
6-Cl
5-Cl
83.2 ꢃ 0.5
72.1 ꢃ 0.9
ꢀ17.6 ꢃ 1.8
ꢀ32.9 ꢃ 3.0
0
10
25
(1.79 ꢃ 0.02) ꢄ 10ꢀ3
(5.53 ꢃ 0.15) ꢄ 10ꢀ3
(2.79 ꢃ 0.04) ꢄ 10ꢀ2
a Abbreviations defined in Scheme 1.
b Binary solvents are on a volume–volume basis at 25 8C. A ¼ acetone, E ¼ ethanol, W ¼ water.
c Average rate constants from at least three runs at 25 8C and two runs at other temperatures. Errors shown are standard deviations.
d Errors shown are standard errors.
e Buffered with Proton sponge base [1,8-bis(dimethylamino)naphthalene].
f Buffered with 2,6-lutidine.
g From Reference [1].
h From Reference [2,3].
parameters seem independent of the solvent. Similarly, bromides
also produce parallel logk versus Ef plots, but the slopes are
slightly lower (sf ¼ 0.95 ꢃ 0.03; calculated from 12 entries).[4]
Generally a lower slope of Hammett plot in comparison to some
reference SN1 reaction has been rationalized in different ways: as an
increase in the SN2 character, as an earlier transition state due to
relief of steric strain of tertiary derivatives, or as a neighboring
group participation in the rate determining step.[7–13] Essentially, in
all cases effects come from less positive charge on the reaction
center in the transition state. According to above consideration, the
same phenomena can account for lower sf parameters.
The negative charge generated in heterolysis of phenyl and
methyl carbonates is distributed almost equally over all three
oxygen atoms because of the resonance and the negative
(inverse) hyperconjugation effects.[1] Thus, for phenyl carbonate
ion, the NBO charges at B3LYP/6-311þþG(d,p) level are: ꢀ0.70,
ꢀ0.73 for oxygen atoms involved in resonance, and ꢀ0.66 for that
involved in inverse hyperconjugation (ꢀ0.80, ꢀ0.77, and ꢀ0.66
for methyl carbonate ion). Our first assumption was that lower
slopes obtained in solvolysis of substituted benzhydryl phenyl
and methyl carbonates than those of benzhydryl chlorides come
from different solvation, due to dispersed negative charge
generated in the transition state.
Finally, it should be mentioned that McLennan and Martin
investigated the solvolytic behavior of various substituted
benzhydryl p-nitrobenzoates in 50–80% aq. ethanol at
100 8C.[7] The corresponding rþ values showed that the reaction
constants were essentially invariant to the solvent change, i.e.,
almost parallel log k versus sþ plots were obtained (rþ in the
range between ꢀ3.18 and ꢀ3.11 in 80, 70, 60, and 50% aq.
ethanol). Slight trends in decrease of rþ values as the fraction of
the water increases are in the limits of experimental error, so they
cannot be taken as a valid prove of lack of the solvent influence.
In order to test systematically how the solvation of the LG
influences the solvolytic behavior of substrates that produce
carbocations which are highly stabilized by resonance, we carried
out kinetic measurements with substituted benzhydryl phenyl
carbonates, methyl carbonates, 3,5-dinitrobenzoates (DNB), and
chlorides in various solvents. Also, our kinetic data obtained
previously for substituted benzhydryl carbonates in numerous
J. Phys. Org. Chem. 2009, 22 495–503
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

B. DENEGRI AND O. KRONJA
Scheme 1.
solvents[1] were used to elucidate the influence of the LG on the
reactivity of benzhydryl derivatives.
In order to calculate slope parameters (sf) and the correspond-
ing nucleofugality parameters (Nf) for DNB, the logarithms of the
first-order rate constants (at 258C) measured here for 1–5-DNB in a
given solvent were plotted against Ef. Excellent linear correlations
obtained in aqueous acetone and ethanol for DNB are presented in
Fig. 1 (r > 0.997). In order to avoid overlaps, the plots measured in
the least polar solvents are not presented in Fig. 1 (Supporting
Information). The slopes of the logk versus Ef plots represent the sf
parameters of DNB in the given aqueous binary mixtures, while the
negative intercepts on the abscissa correspond to its nucleofug-
ality parameters, as defined in eqn (1). The nucleofuge specific
parameters for DNB are presented in Table 1 along with some
referent values obtained earlier with PhCarb, MeCarb, and
chlorides.[1] The results show that the slope parameters sf for
DNB are decreased as the fraction of the water in the binary
solvent increases in a similar manner as those for carbonates.
RESULTS
Electrofuges used here (1–6, Scheme 1) differ in reactivities up to
six orders of magnitude (electrofugalities are as follow:
Ef(1) ¼ 0.00,
Ef(2) ¼ ꢀ0.81,
Ef(3) ¼ ꢀ1.29,
Ef(4) ¼ ꢀ2.06,
Ef(5) ¼ ꢀ3.47, Ef(6) ¼ ꢀ6.05.[13] Benzhydryl DNBs were designed
by selecting an appropriate benzhydryl electrofuge to enable
measurements in the given solvents at 25 8C. Solvolysis rate
constants of 1–5-DNB were measured conductometrically at
25 8C in aqueous acetone and aqueous ethanol. Activation
parameters of some substrates 1–5 with carbonates, DNB, and
chloride as LGs in some solvents were determined from data
collected at at least three different temperatures. Because of the
problems with solubility of the substrates at lowered tempera-
tures and with high reactivities at elevated temperatures, and
also because of low solubility of the substrates in solvents with
higher content of water, the activation parameters have not been
determined for all possible combinations of the substrates and
the solvents. The first order rate constants for solvolysis of
substrates 1–5 and the corresponding activation parameters
along with some reference data are presented in Table 2.
DISCUSSION
It has been demonstrated repeatedly that the different extent of
solvation of the transition state, due to dispersal of a positive
charge in the electrofuge[14–16] or a negative charge in the LG,
influences the solvolytic behavior.[17–23] In order to study the
Table 2. Nucleofuge specific parameters (Nf and sf) for 3,5-dinitrobenzoate (DNB), and sf parameters for phenyl carbonate (PhCarb)
and methyl carbonate (MeCarb) in various solvents at 25 8C
DNB
PhCarb
MeCarb
Cl
Solventa
sfb
Nfb
sfc
sfc
sfc
100E
1.10 ꢃ 0.03c
1.07 ꢃ 0.02
0.99 ꢃ 0.02c
ꢀ2.04 ꢃ 0.09c
ꢀ1.53 ꢃ 0.06
ꢀ1.43 ꢃ 0.04c
0.98 ꢃ 0.02
0.93 ꢃ 0.02
0.87 ꢃ 0.02
0.83 ꢃ 0.02
0.79 ꢃ 0.02
0.76 ꢃ 0.06
1.15 ꢃ 0.04
1.04 ꢃ 0.02
0.99 ꢃ 0.02
0.94 ꢃ 0.03
0.90 ꢃ 0.04
0.87 ꢃ 0.06
1.00
90E10W
80E20W
70E30W
60E40W
50E50W
90A10W
80A20W
70A30W
60A40W
0.98 ꢃ 0.01
0.98 ꢃ 0.01
0.96 ꢃ 0.04
0.97 ꢃ 0.06
1.18 ꢃ 0.02c
1.07 ꢃ 0.03c
0.99 ꢃ 0.05
0.93 ꢃ 0.05
ꢀ2.57 ꢃ 0.04c
ꢀ2.39 ꢃ 0.11c
ꢀ2.27 ꢃ 0.18
ꢀ2.12 ꢃ 0.18
0.99 ꢃ 0.01
1.02 ꢃ 0.01
0.99 ꢃ 0.01
0.98 ꢃ 0.02
0.96 ꢃ 0.00
0.85 ꢃ 0.02
0.80 ꢃ 0.02
0.97 ꢃ 0.09
0.92 ꢃ 0.01
aBinary solvents are v/v; A ¼ acetone, E ¼ ethanol, W ¼ water.
bData from Reference [1].
cData from Reference [2,3].
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 495–503

EFFECT OF THE LEAVING GROUP SOLVATION
generating positive charge is delocalized, and the negative is
concentrated on relatively small chloride ion. Because of
less pronounced solvation, the solvolysis rate constants of
carbonates and DNBs increase less with increasing electrofugality
than those of chlorides. The net effect is lower sf in all solvents
(the exceptions are the sf values for methyl carbonate and DNB in
the least polar solvents).
Why would the log k versus Ef slopes (sf) decrease when the
fraction of the water in the binary solvent increases? While for the
series of benzhydryl chlorides small differences in solvation of
the electrofuge are negligible due to strong solvation of the
chloride ion, for the series of substrates in which delocalization
occurs also in the nucleofuge, different solvation of the
electrofuge is important. Increasing electrofugality in the series
of given benzhydryl derivatives (carbonates or DNBs) comes from
more extensive positive charge delocalization and thus less
stabilization of the activated complex by more polar solvent. On
the other hand, in the case of benzhydryl substrates with weaker
electrofuges, the solvation has more influence on the reactivity
since the positive charge is less dispersed. Thus, in the solvent
with a higher fraction of water (more polar solvent) rate
enhancement becomes less strong if the substrate have a
stronger electrofuge than that with a weaker electrofuge, and the
result is a lower log k versus Ef slope (sf) in a more polar solvent.
This consideration is in accord with experimental findings. For
example, the ratio between the solvolysis rates constants of
benzhydryl carbonate 5-PhCarb (Ef ¼ ꢀ3.46) in 60% aq. ethanol
and 90% aq. ethanol is 11.8, while that of 3-PhCarb (Ef ¼ ꢀ1.29) in
the same solvents is only 5.7.[1]
Figure 1. Plots of logk (25 8C, from Table 2) versus Ef for the solvolysis
reactions of X,Y-substituted benzhydryl 3,5-dinirobenzoates in aqueous
acetone and aqueous ethanol
effects of simultaneous solvation of the LG and the benzhydryl
electrofuge, solvolytic behavior of benzhydryl chlorides have
been compared with behavior of other benzhydryl derivatives. In
chlorides the developed negative charge in the TS is concen-
trated only on chloride atom. In benzhydryl 3,5-dinitrobenzoates
(1–5-DNB) the negative charge is delocalized to two carboxylate
oxygen atoms by resonance, and to two nitro groups on the
phenyl ring by negative inductive effect. In benzhydryl phenyl
carbonate (1–5-PhCarb) and benzhydryl methyl carbonates
(1–5-MeCarb), as mentioned above, intense charge delocaliza-
tion occurs to three oxygen atoms.
Because of similar structures of phenyl and methyl carbonates,
the influence of the charge delocalization and of charge
separation on sf values may be analyzed separately. The reaction
constants shown in Table 1 indicate that the sf values for the
series of phenyl and methyl carbonates decrease for almost the
same amount if the water content increases. For example, if
Analysis of the log k versus Ef plots
The major factor controlling variation of rate with solvent
composition of the reference chlorides is the solvation of the
developing chloride ion, regardless of the barrier of the
heterolysis reaction, i.e., shifts of transition states toward
reactants in faster reactions do not diminish the importance of
the solvation. Since the chloride solvation is much more
important than the solvation of the substituted benzhydrylium
ion because of the charge delocalization in the latter, the changes
of the water content have a similar effect on the reactivity for all
substrates, so parallel log k versus Ef plots are obtained, i.e., same sf
parameters in all solvents used.[4]
Even though consistent explanation of dependence between
the solvent polarity and sf is lacking, two major factors can be
considered to influence the reaction constant (sf): the degree of
the charge separation (earlier or later transition state) and the
degree of the charge delocalization which mainly determines the
importance of solvation. Clear differentiation between those two
influences is difficult, since the net effect is combination of both.
As mentioned above, the negative charge in carbonate anions
and in DNB anion is delocalized, so the charge delocalization
occurs simultaneously in both parts of the activated complex, an
electrofuge and a nucleofuge, while in chlorides only the
Figure 2. Correlation plots of log k versus YOTs for solvolysis of
(a) benzhydryl phenyl carbonates 1-Phcarb, 4-PhCarb, and 6-PhCarb
and in binary aqueous solvents (v/v); solvents: W ¼ water, E ¼ ethanol,
M ¼ methanol, A ¼ acetone; m values are given in parentheses (abbrevi-
ations defined in Scheme 1)
J. Phys. Org. Chem. 2009, 22 495–503
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

B. DENEGRI AND O. KRONJA
Table 3. Values of m parameters from Grunwald–Winstein correlations for solvolysis of X,Y-substituted benzhydryl phenyl
carbonates (PhCarb), methyl carbonates (MeCarb), 3,5-dintirobenzoates (DNB), and chlorides, and the corresponding critical
electrofugalities Efcrit
m
b
Leaving group
PhCarb
Solventa
E-W
Scale
1
2
3
4
5
6
Efcrit
YCl
YOTs
0.24d
0.33d
0.33
0.45
0.36
0.50
0.46
0.64
0.60d
0.83d
4.1
4.0
4.0 W 0.0c (ꢅ4.0)
A-W
E-W
A-W
YCl
YOTs
0.45d
0.59d
0.53
0.70
0.57
0.74
0.70
0.92
0.87d
1.14d
6.2
6.2
6.2 W 0.0c (ꢅ6.0)
MeCarb
YCl
YOTs
0.29
0.40
0.30
0.42
0.39
0.53
0.41
0.55
0.65d
0.89d
4.7
4.7
4.7 W 0.0c (ꢅ4.7)
7.6
YCl
YOTs
0.52
0.69
0.59
0.78
0.93d
1.22d
7.6
7.5 W 0.1c (ꢅ7.5)
DNB
Cl
E-W
A-W
YCl
YOTs
YCl
0.34
0.43
0.33
0.37
0.35
0.45
0.36
0.40
0.41
0.53
0.44
0.49
0.42
0.53
0.48
0.54
0.48
0.62
0.59f
0.66f
0.58
0.75
0.79f
0.88f
(ꢅ2.9)
4.2
4.2
YOTs
4.2 W 0.0c (ꢅ4.2)
E-W
A-W
YCl
YOTs
YCl
0.53e
0.72e
0.91e
0.59e
0.79e
0.92e
aBinary solvents are v/v at 25 -C; A ¼ acetone, E ¼ ethanol, W ¼ water.
bAbscissa of the m versus Ef plot where m ¼ 0.
cValue of the average Efcrit W SD obtained from seven different scales (Supporting Information); the approximative values Efcrit given in
parentheses are derived from intersections of log k versus Ef plots in given binary mixture.
dRate constants employed in Grunwald–Winstein correlations were calculated from eqn (1) by using Nfs and sf for carbonates given in
Reference [1] and Efs given in References [2,3].
eRate constants employed in Grunwald–Winstein correlations were calculated by using Nfs and sf for chlorides given in Reference [4].
fRate constants employed in Grunwald–Winstein correlations were calculated from eqn (1) by using Nfs and sf for
3,5-dinitrobenzoates (Table 1).
switching from 90% aq. ethanol to 60% aq. ethanol, sf is
decreased for the same value for both series of carbonates
(Dsf ¼ 0.14; Dsf ¼ 0.10 and 0.11, respectively when switching from
80% to 60% aq. acetone). This phenomenon could be the
consequence of negative charge delocalization, since in both LGs
the charge is mainly distributed to three oxygen atoms.[1] On the
other hand, consistently smaller sf values for phenyl carbonates
than for methyl carbonates for each solvent is likely to come from
less charge separation in the TS, which is in accord with greater
reactivity of phenyl carbonates.
Even though the influence of the negative charge delocaliza-
tion and the charge separation in TS on the importance of
solvation cannot be differentiated for the series of benzhydryl
DNBs, the trends in sf values in the series of a given organic
solvent with various water contents, indicate that diminished
solvation occurs in the TS in a similar manner as for benzhydryl
methyl carbonates.
approach and analyzed the solvolytic behavior of a single substrate
separately in the series of organic solvents/water mixtures.
Analysis of the logk versus Y plots
Using various solvent ionizing power scales we plotted log k of
1–6-LG versus Y values from various ionizing power scales (YCl, YBr,
YOTs, YBnCl, YxBnCl, YxBnBr, Y)[24–28] (Supporting Information). It
should be pointed out that all kinetic data are obtained in the
organic-rich solvents in which linear relationship between log k
and Y exists.[29]
It has been reported repeatedly that kinetic data obtained in
different types of solvents cannot be treated with a single line
expressed by the Grunwald–Winstein equation because of
dispersion of the data, but the rates in each binary aqueous
solvent pair should be treated separately.[17–19,25,30–32] This
phenomenon was proposed to be mainly a consequence of
differential solvation, electrophilic solvent assistance, and also of
different extent of ion-pair return in different binary sys-
tems.[18,30–33] Figure 2 shows the lines obtained by plotting
the logarithms of solvolysis rates constants (experimental or
The above conclusions have been drawn by analyzing the
kinetic data for the series of benzhydryl derivatives in a given
solvent. In order to provide further support about the importance
of solvation in the LG for benzhydryl substrates, we used a different
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 495–503

EFFECT OF THE LEAVING GROUP SOLVATION
calculated by eqn 1) of 1-Phcarb, 4-PhCarb, and 6-PhCarb against
YOTs (for correlation parameters for carbonates and DNBs refer to
Supporting Information). Basically equally good correlations for
separated lines were obtained regardless which solvent ionizing
power scale was used. Table 3 presents the m values for methyl
and phenyl carbonates, DNBs, and chlorides derived log k versus
YCl and log k versus YOTs plots (for complete correlation analyses
and other details refer to Supporting Information).
As expected, all m values for benzhydryl carbonates and DNBs
with better electrofuges are lower than those for chlorides
(Table 3), indicating lower sensitivity of these substrates for the
solvent change.
Low m values have usually been coupled with relatively
positive DSz, as was demonstrated for 2-adamantyl perchlorate
and 1-adamantyl picrate.[20–22] However, DSz values obtained
with benzhydryl carbonates and DNBs are not considerable more
positive than those for chlorides. This might be due to more
restricted internal motion in the transition states than in
the ground state of the benzhydryl system, which is not the
case with, for example, the rigid adamantyl system. The reaction
rates of carbonates and DNBs are mostly determined with the
magnitude of DHz, while no obvious correlations between the
reaction rates and DSz could be found. The same was observed
earlier for solvolysis of substituted benzhydryl chlorides.[34–37]
Table 3 shows that the m values for carbonates and DNBs
decrease as the electron donating ability of the substituted
benzhydryl ring increases, while those of chlorides are essentially
the same. The lower sensitivity of the substrates with better
electrofuges toward solvent polarity is consistent with the above
conclusions based on the log k versus Ef plots.
Figure 4. Extrapolated log k (25 8C) versus Ef correlation plots for the
solvolysis of phenyl carbonates in aqueous ethanol (v/v; W ¼ water,
E ¼ ethanol)
carbonates and DNBs in aqueous ethanols and/or aqueous
acetones (for other correlations refer to Supporting Information).
At one point the extrapolated m versus Ef plots reach m ¼ 0.
Abscissa at that point indicates the critical electrofugality value
Efcrit at which the solvolysis rate of a given benzhydryl derivative in
a given series binary solvent mixture should not depend on the
water content. Such derivatives are those with highly stabilized
electrofuges.
In order to provide independent support that the substrates
are less sensitive to change of the solvent polarity if the charge is
more delocalized, we plotted the m parameters obtained for the
series of phenyl carbonates (1–6-PhCarb), methyl carbonates
(1–6-MeCarb), and 3,5-dinitrobenzaotes (1–5-DNB) in a given
aqueous binary solvent series, against the corresponding
electrofugalities (Ef), since for benzhydryl system the charge
delocalization is closely related to electrofugality. Correlations of
m versus Ef have been performed with m parameters derived from
seven different solvent-ionizing power scales.
Extrapolated critical electrofugality values from two m versus Ef
correlations (m derived from log k versus YCl and log k versus YOTS
plots) are summarized in Table 3 (for others refer to Supporting
Information). Regardless which ionizing power scale has been
used for derivation of the m parameter, essentially the same
results for the extrapolates critical electrofugality are obtained.
To verify the existence of the extrapolates critical electro-
fugality established by analyzing m versus Ef plots, we have
returned to analyze log k versus Ef plots based on eqn (1). The
extrapolated log k versus Ef plots constructed for a given LG
(PhCarb, MeCarb, or DNB) in a given binary solvent system
intersect in a certain interval of Ef. Assuming that log k versus Ef
plots are still linear in the extrapolated region, abscissa above this
interval indicates that the solvolysis of a benzhydryl substrate
with a given electrofugality in a given binary solvent mixture
series system should not depend on the water content, which
concept corresponds to above-established critical electrofugality
Efcrit based on the intersection of the extrapolated m versus Ef
lines. Figure 4 shows the extrapolated plots that correspond to
phenyl carbonates in aqueous ethanol. Extrapolates Efcrit derived
from extrapolated k versus Ef plots for phenyl and methyl
carbonates and DNBs in all series of solvents measured are
presented in Table 3 (for the extrapolated plots refer to
Supporting Information). It is rather outstanding that the
extrapolated Efcrit derived from m verus Ef plots and that derived
from log k versus Ef plots are practically the same (Table 3 and the
Supporting Information). It should be mentioned that if instead of
assuming the model with linear behavior of log k versus Ef plots in
the region of high electrofugality, the curved plots that approach
an asymptote are considered (also for m vs. Ef plots), the concept
of existance of critical electrofugality is unchanged, only the
values of Efcrit may be slightly shifted toward higher values.
Lines (m vs. Ef) show good correlations regardless to the
ionizing power scale used for calculating the m parameter.
Figure 3 presents the mOTs versus Ef correlation lines for phenyl
Figure 3. Correlation plots of m (obtained in log k vs. YOTs correlation)
versus Ef in binary aqueous solvents (solvents: W ¼ water, E ¼ ethanol,
M ¼ methanol, A ¼ acetone)
J. Phys. Org. Chem. 2009, 22 495–503
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

B. DENEGRI AND O. KRONJA
The values of the critical electrofugalities for phenyl carbonate
are consistently lower than for methyl carbonate in all aqueous
binary mixture systems. The lowest Efcrit is obtained for DNB in
aqueous acetone.
NMR and 13C NMR spectra are presented in the Supporting
Information.
4-Methoxy-40-methylbenzhydryl phenyl carbonate (3-PhCarb),
4-methoxybenzhydryl phenyl carbonate (4-PhCarb), and
4,40-dimethylbenzhydryl phenyl carbonate (5-PhCarb) were
prepared from the parent benzhydrol and phenyl chloroformate
in the presence of pyridine in dry benzene. 4,40-Dimethoxy-
-Dimethoxybenzhydryl methyl carbonate (1-MeCarb) and
methoxy-40-methylbenzhydryl methyl carbonate (3-MeCarb)
CONCLUSION
Benzhydryl derivatives in which substantial simultaneous
delocalization of the positive charge in the benzhydryl moiety
and the negative charge in the LG occur in the TS, generally have
lower reaction constants (sf) than the compounds that lack the LG
which can delocalize the evolving negative charge. For such
substrates the reaction rate is enhanced less strong if the polarity
of the solvent of the given binary mixture is increased (higher
fraction of water), causing decrease of sf values. Compounds that
show the above phenomena are X,Y-substituted benzhydryl
phenyl and methyl carbonates and X,Y-substituted benzhydryl
3,5-dinitrobenzoate.
were prepared from the parent benzhydrol and methyl
[1]
chloroformate. Detailed procedures are given in Reference
.
Kinetic methods
Solvents were purified and dried according to the standard
procedures. Solvolysis rate constants at various temperatures
(ꢃ0.05 8C) were measured conductometrically. Typically,
10–40 mg of substrate was dissolved in 0.10 ml of dichloro-
methane and injected into the fresh mixed aqueous solvent.
Increase of conductivity during solvolysis was monitored
automatically by means of WTW LF 530 conductometer using
Pt electrode LTA 1/NS. Individual rate constant were obtained by
the least squares fitting of the conductivity to the first order
kinetic equation for 3–4 half-lives. Rate constants at 25 8C were
averaged from at least three measurements, while the rate
constants at higher and lower temperatures are obtained from at
least two measurements.
Liberated weak acids (3,5-dinitrobenzoic, phenyl carbonic, and
methyl carbonic) were ionized by addition of the proton sponge
base [1,8-bis-(dimethylamino)naphthalene]. Calibration showed
a linear response of the conductivity toward concentrations of
liberated 3,5-dinitrobenzoic acid in the presence of a base. The
ranges of concentrations for each given aqueous binary mixture
for DNBs are presented in Table 4.
Because of the decreasing slopes of log k versus Ef plots with
increasing water content in the binary organic/aqueous solvent,
the extrapolated plots intersect in the narrow range and the
abscissa of that extrapolated intersection indicates the critical
electrofugality Efcrit. It seems that if the electrofugality of
benzhydryl derivative is above the critical electrofugality Efcrit
,
its solvolysis rate may be independent on the water content, at
least in solvents rich in organic component. To design such
substrates, the extended electrofugality scale is needed and also
the critical Efcrit for a given LG. Nevertheless, it might be predicted
that the ethanolysis rate of benzhydryl phenyl carbonate with
Ef ꢅ 3 depends only marginally on the fraction of water, while that
of the corresponding benzhydryl DNB still reacts with unchanged
rate in all ethanol/water mixtures.
Some rate constants for solvolysis DNBs in 80% ethanol were
taken from the Reference [2,3]. In order to keep experimental
conditions unchanged, kinetic measurements with DNBs in 80%
ethanol in this work were also performed in the presence of
2,6-lutidine. Test measurement with the proton sponge base
conformed that in 80% ethanol the rate constants obtained in the
presence of 2,6-lutidine and those obtained in the presence of
the proton sponge base were all within the range of the
experimental error. However, it turned out that in less polar
solvents, 2,6-lutidine did not dissociate the liberated
3,5-dinitobenzoic acid completely. Therefore, solvolysis rates of
DNBs in other solvents were measured in the presence of the
proton sponge base. Test measurements were carried out
EXPERIMENTAL SECTION
Substrate preparation
4,40-Dimethoxybenzhydryl
3,5-dinitrobenzoate
(1-DNB),
4-methoxy-40-phenoxybenzhydryl 3,5-dinitrobenzoate (2-DNB),
4-methoxy-40-methylbenzhydryl 3,5-dinitrobenzoate (3-DNB),
4-methoxybenzhydryl 3,5-dinitrobenzoate
(4-DNB),
and
4,40-dimethylbenzhydryl 3,5-dinitrobenzoate (2-DNB) were pre-
pared from the parent benzhydrols and 3,5-dinitrobenzoyl
chloride in the presence of pyridine in dry benzene solution
according to procedure presented in details in Reference [2,3]. 1H
Table 4. Experimental initial concentrations of the substrates and the bases for solvolysis of the substituted 3,5-dinitrobenzoates
Solventa
Baseb
c0(DNB), (mM)
c0(Base), (mM)
c0(Base)/c0(DNB)
90E10W
80E20W
80A10W
70A30W
60A40W
PSB
2.3–3.8
1.6–3.5
1.6–2.7
1.9–2.5
1.5–2.5
15.6–31.2
8.0–12.0
10.9–15.6
12.5–18.7
3.1–6.2
5.0–10.0
3.0–7.0
4.5–6.8
5.0–8.0
2.0–4.0
Lutidine
PSB
PSB
PSB
a Binary solvents are on a volume–volume basis at 25 8C. A ¼ acetone, E ¼ ethanol, W ¼ water.
b PSB ¼ proton sponge base [1,8-bis(dimethylamino)naphthalene].
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 495–503

EFFECT OF THE LEAVING GROUP SOLVATION
titrimetrically. Details and ranges of base concentration for
phenyl and methyl carbonates are given in Reference [1], Table 6.
[10] D. J. McLennan, P. L. Martin, J. Chem. Soc. Perkin Trans. 1982, 2,
1999–2105.
[11] I. Malnar, K. Humski, O. Kronja, J. Org. Chem. 1998, 63, 3041–3044.
´
´
[12] S. Juric, A. Filipovic, O. Kronja, J. Phys. Org. Chem. 2003, 16, 900–904.
[13] S. Juric, O. Kronja, J. Phys. Org. Chem. 2005, 18, 368–372.
[14] K.-T. Liu, H.-C. Sheu, J. Org. Chem. 1991, 56, 3021–3025.
[15] K.-T. Liu, Y.-S. Lin, M.-L. Tsao, J. Phys. Org. Chem. 1998, 11, 223–229.
[16] K.-T. Liu, C.-P. Chin, Y.-S. Lin, M.-L. Tsao, Tetrahedron Lett. 1995, 36,
6919–6922.
´
SUPPORTING INFORMATION
Correlation analyses of solvolyses of substituted benzhydryl
carbonates and chlorides, correlations of
m values with
[17] T. W. Bentley, G. E. Carter, K. Roberts, J. Org. Chem. 1984, 49,
5183–5189.
[18] T. W. Bentley, K. Roberts, J. Org. Chem. 1985, 50, 4821–4828.
[19] T. W. Bentley, G. Llewellyn, Z. H. Ryu, J. Org. Chem. 1998, 63,
4654–4659.
[20] D. N. Kevill, M. S. Bahari, S. W. Anderson, J. Am. Chem. Soc. 1984, 106,
2895–2901.
[21] D. N. Kevill, S. W. Anderson, J. Org. Chem. 1985, 50, 3330–3333.
[22] D. N. Kevill, D. C. Hawkinson, J. Org. Chem. 1990, 55, 5394–5399 (and
references cited therein).
electrofugality parameters Ef, corresponding critical electrofug-
ality values.
Acknowledgements
We gratefully acknowledge the financial support of this research
by the Ministry of Science, Education and Sport of the Republic of
Croatia (Grant No. 006-0982933-2963).
[23] P. R. Luton, M. C. Whiting, J. Chem. Soc., Perkin Trans. 1979, 2,
1507–1511.
[24] A. H. Fainberg, S. Winstein, J. Am. Chem. Soc. 1956, 78, 2770–2777.
[25] T. W. Bentley, G. Llewellyn, Prog. Phys. Org. Chem. 1990, 17, 121–159.
[26] K.-T. Liu, H.-C. Sheu, H.-I. Chen, P.-F. Chiu, C.-R. Hu, Tetrahedron Lett.
1990, 31, 3611–3614.
[27] K.-T. Liu, Y.-S. Lin, M.-L. Tsao, J. Phys. Org. Chem. 1998, 11, 223–229.
[28] K.-T. Liu, C.-P. Chin, Y.-S. Lin, M.-L. Tsao, J. Chem. Res (S) 1997, 18–19.
[29] T. W. Bentley, C. T. Bowen, H. C. Brown, F. J. Chloupek, J. Org. Chem.
1981, 46, 38–42.
[30] A. H. Fainberg, S. Winstein, J. Am. Chem. Soc. 1957, 79, 1602–1608.
[31] A. H. Fainberg, S. Winstein, J. Am. Chem. Soc. 1957, 79, 1608–1612.
[32] S. Winstein, A. H. Fainberg, E. Grunwald, J. Am. Chem. Soc. 1957, 79,
4146–4155.
REFERENCES
[1] B. Denegri, O. Kronja, J. Org. Chem. 2007, 72, 8427–8433.
´
[2] B. Denegri, A. Streiter, S. Juric, A. R. Ofial, O. Kronja, H. Mayr, Chem. Eur.
J. 2006, 12, 1648–1656.
[3] B. Denegri, A. Streiter, S. Juric, A. R. Ofial, O. Kronja, H. Mayr, Chem. Eur.
´
J. 2006, 12, 5415–5415.
[4] B. Denegri, A. R. Ofial, S. Juric, A. Streiter, O. Kronja, H. Mayr, Chem. Eur.
´
J. 2006, 12, 1657–1666.
[5] B. Denegri, S. Minegishi, O. Kronja, H. Mayr, Angew. Chem Int. Ed.
2004, 43, 2302–2305.
[6] T. W. Bentley, Chem. Eur. J. 2006, 12, 6514–6520.
[7] D. J. McLennan, P. L. Martin, Aust. J. Chem. 1979, 32, 2361–2370.
[8] D. J. McLennan, Tetrahedron 1978, 34, 2321–2341.
[9] D. J. McLennan, P. L. Martin, J. Chem. Soc. Perkin Trans. 1982, 2,
1991–1998.
[33] D. N. Kevill, M. J. D’Souza, J. Phys. Org. Chem. 1992, 5, 287–294.
[34] C. Schade, H. Mayr, Tetrahedron. 1988, 5761–5770.
[35] S. Nishida, J. Org. Chem. 1967, 32, 2692–2695.
[36] S. Nishida, J. Org. Chem. 1967, 32, 2695–2697.
[37] S. Nishida, J. Org. Chem. 1967, 32, 2697–2701.
J. Phys. Org. Chem. 2009, 22 495–503
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc