Correlation of the rates of solvolysis of n-octyl fluoroformate and a comparison with n-octyl chloroformate solvolysis

Article abstract of DOI:10.1039/b109169g

The specific rates of solvolysis of n-octyl fluoroformate have been measured at 24.2°C in 28 pure and binary solvents. For the 23 solvents for which both N_T(solvent nucleophilicity) and Y_Cl (solvent ionizing power) values are known, a correlation using the two-term Grunwald-Winstein equation leads to sensitivities towards changes in the two scales of 1.80 ± 0.13 (l value) and 0.79 ± 0.06 (m value), respectively. For seven solvents, a parallel study was made of n-octyl chloroformate solvolysis and F:Cl specific rate ratios were, in most instances, above unity, consistent with the association step of an association-dissociation (addition-elimination) pathway being rate-determining.

Full text of DOI:10.1039/b109169g

View Article Online / Journal Homepage / Table of Contents for this issue
Correlation of the rates of solvolysis of n-octyl fluoroformate and
a comparison with n-octyl chloroformate solvolysis
Dennis N. Kevill* and Malcolm J. D’Souza†
2
Department of Chemistry and Biochemistry, Northern Illinois University, DeKalb,
Illinois 60115-2862, USA
Received (in Cambridge, UK) 9th October 2001, Accepted 27th November 2001
First published as an Advance Article on the web 17th January 2002
The specific rates of solvolysis of n-octyl fluoroformate have been measured at 24.2 ЊC in 28 pure and binary solvents.
For the 23 solvents for which both NT (solvent nucleophilicity) and YCl (solvent ionizing power) values are known, a
correlation using the two-term Grunwald–Winstein equation leads to sensitivities towards changes in the two scales
of 1.80 0.13 (l value) and 0.79 0.06 (m value), respectively. For seven solvents, a parallel study was made of
n-octyl chloroformate solvolysis and F : Cl specific rate ratios were, in most instances, above unity, consistent with
the association step of an association–dissociation (addition–elimination) pathway being rate-determining.
Recently, we have reported on the application of the extended
(two-term) Grunwald–Winstein equation1,2 [eqn. (1)] to the
log (k/k0) = lNT ϩ mYCl ϩ C
(1)
solvolyses of several chloroformate esters [eqn. (2)], including
(2)
(3)
phenyl,3 methyl,4 ethyl,5 isopropyl,6 benzyl,7 and 4-nitrobenzyl.7
In eqn. (1), k and ko are the specific rates of solvolysis in a given
solvent and in the standard solvent (80% ethanol), respectively;
l is the sensitivity towards changes in solvent nucleophilicity
(NT); m is the sensitivity towards changes in solvent ionizing
power (Yx, for a leaving group X); c is a constant (residual)
term. The eqn. (2) represents the most general form of the
solvolysis reaction. The dialkyl or alkyl aryl carbonate is
formed by nucleophilic attack of alcohol (SOH) at the acyl
carbon. The alcohol or phenol product is formed either by a
parallel attack by water to give the hydrogen carbonate ester,
followed by loss of CO2,8 or by attack of water on Rϩ, formed
by loss of CO2 from an intermediate carboxylium ion
(ROCO)ϩ. Capture of the Rϩ by SOH leads to ester (ROS) and
capture by the chloride ion formed in the ionization will lead to
a decomposition pathway (for the chloroformate ester reactant)
to give RCl. When feasible, an alternative decomposition
pathway, involving extraction of a β-proton from the R group
and alkene formation, can also occur.
Two types of mechanism have been found to operate, postu-
lated as an addition–elimination mechanism [eqn. (3)], with
addition being rate determining, and an ionization mechanism
[eqn. (4)]. Replacement of either or both oxygens by sulfur
has been shown to lead to an increased tendency towards the
ionization pathway.5,9,10
The solvolyses of 1-adamantyl chloroformate were found11,12
to proceed almost entirely via the 1-adamantyl cation. Only in
100% ethanol was a trace of the mixed carbonate observed.
However, replacement of chlorine by fluorine led to the reac-
tion following the addition–elimination pathway in all but the
(4)
most ionizing and weakly nucleophilic solvents.13 Since the
chloroformate and fluoroformate reacted in almost all solvents
by different pathways, a consideration of the overall F : Cl
specific rate ratios was not a useful mechanistic indicator. For
other haloformate esters, a consideration of the F : Cl rate ratio
has provided some of the best evidence in favor of the addition
step within an association–dissociation (addition–elimination)
mechanism being rate determining.
For solvolyses of ethyl haloformates, F : Cl specific rate ratios
of 28.6 for 85% acetone at 0.0 ЊC14 and of 1.1 for ethanol
at 25.1 ЊC15 have been reported. Since the C–F bond is
much stronger than the C–Cl bond, ratios of well below unity
would have been anticipated if the carbon–halogen bond
was appreciably broken at the transition state of the rate-
determining step. Since an extensive tabulation of the specific
rates of solvolysis of ethyl chloroformate in a wide variety of
solvents exists,5 it was decided that a parallel study would be
made of ethyl fluoroformate solvolyses. One could then see
† Current address: Department of Chemistry, Wesley College,
120 N. State Street, Dover, Delaware 19901-3875, USA.
240
J. Chem. Soc., Perkin Trans. 2, 2002, 240–243
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b109169g

View Article Online
Table 1 Specific rates of solvolysis of n-octyl fluoroformatea in a
whether there was any appreciable variation in l and m values
[eqn. (1)] from those obtained for the analysis of the chloro-
formate solvolyses. In addition, F : Cl specific rate ratios would
be available for a wide range of solvent composition, such that
it would be possible to see whether the previously observed14,15
26-fold variation in going from 85% acetone to ethanol can be
extended when additional solvents are considered.
Initially, we attempted to follow the kinetics of the solvolyses
of ethyl fluoroformate by the procedure developed by Queen
and Nour16 for solvolyses of aryl fluoroformates. This involved
quenching in pentane and extracting the developed HF into
water prior to titration. It was found that the ethyl ester, being
less hydrophobic, was also to some degree extracted into
the water, followed by relatively rapid hydrolysis and erratic
and inaccurate titration values. To overcome this problem,
we increased the hydrophobic character of the substrate by
replacing the ethyl group by the n-octyl group [CH3(CH2)7-
OCOF]. Also, to allow an unambiguous comparison, we deter-
mined for seven of the solvents the specific rate of solvolysis of
the n-octyl chloroformate [CH3(CH2)7OCOCl].
variety of pure and mixed solvents at 24.2 ЊC and the NT, YCl, and YOTS
values for the solvents
Solventb
105k/sϪ1 c
NT
YCl
YOTS
d
e
f
100% EtOH
90% EtOH
80% EtOH
70% EtOH
60% EtOHh
50% EtOHh
40% EtOHi
100% MeOH
90% MeOH
80% MeOHh
70% MeOHh
95% Acetone
90% Acetone
80% Acetone
60% Acetone
95% Dioxane
90% Dioxane
80% Dioxane
60% Dioxane
40% Dioxanei
90% TFEk
1.48 0.03g
14.3 0.5
59.6 0.9
0.37
0.16
0.00
Ϫ2.52
Ϫ0.94
0.00
0.78
1.38
2.02
2.75
Ϫ1.17
Ϫ0.18
0.67
Ϫ1.75
Ϫ0.58
0.00
0.47
0.92
1.29
2.08
Ϫ0.92
Ϫ0.17
0.47
93.9 1.4
Ϫ0.20
Ϫ0.38
Ϫ0.58
Ϫ0.74
0.17
Ϫ0.01
Ϫ0.06
Ϫ0.40
Ϫ0.49
Ϫ0.35
Ϫ0.37
Ϫ0.52
141.8 7.6
197.7 10.8
256.1 9.7
8.12 0.08j
78.7 0.8
160.8 4.1
271.9 15.2
0.0409 0.0013
0.322 0.005
2.22 0.04
16.9 0.3
1.46
1.02
Ϫ3.19
Ϫ2.22
Ϫ0.83
0.95
Ϫ2.95
Ϫ1.99
Ϫ0.94
0.66
0.192 0.011
0.999 0.009
2.97 0.08
23.9 0.7
Ϫ2.06
Ϫ1.10
ϩ0.30
Ϫ0.46
Ϫ0.54
Ϫ0.84
Ϫ2.55
Ϫ2.19
Ϫ1.98
Ϫ1.73
Ϫ1.76
Ϫ0.94
Ϫ0.34
0.08
91.4 1.2
Results
0.332 0.011
2.21 0.03
4.30 0.06
11.5 0.8
0.117 0.005
0.879 0.009
1.65 0.05
2.30 0.08
2.85
2.90l
2.96
3.16
1.89
0.63
Ϫ0.48
Ϫ1.42
1.90
1.94
2.00
2.14
0.98
0.21
Ϫ0.44
Ϫ1.18
80% TFEk
The specific rates of solvolysis of n-octyl fluoroformate were
determined in 28 solvents at 24.2 ЊC. The solvents consisted of
ethanol and methanol, binary mixtures of water with ethanol,
methanol, 2,2,2-trifluoroethanol (TFE), acetone, and 1,4-
dioxane, and four binary mixtures of TFE and ethanol. The
70% TFEk
50% TFEk
80T–20Em
60T–40Em
40T–60Em
specific rates of solvolysis are reported in Table 1, together with
20T–80Em
the available NT,2,17 YCl,18,19 and YOTS
values. Specific rates
18,20
a Substrate concentration of 0.0100 mol dmϪ3, unless otherwise indi-
cated. b Unless otherwise indicated, binary solvents are on a volume–
volume basis at 25.0 ЊC. c With associated standard deviation; average
of all integrated first-order rate coefficients from duplicate runs, except
for 90% TFE, where the average of initial values obtained by extrapola-
tion of plots versus extent of reaction is reported. d From refs. 2 and 17.
e From refs. 18 and 19. f Values from refs. 18 and 20a, except those
values based on solvolyses of 1-adamantyl toluene-p-sulfonate are
adjusted as described in ref. 20b. g Value of 1.53 ( 0.03) × 10Ϫ5 sϪ1 at
25.0 ЊC. h Substrate concentration of 0.0050 mol dmϪ3. i Substrate con-
centration of 0.0033 mol dmϪ3. j Value of 8.53 ( 0.05) × 10Ϫ5 sϪ1 at 25.0
ЊC. k On a weight–weight basis. l From ref. 7. m T–E are TFE–EtOH
mixtures.
of solvolysis of n-octyl chloroformate were determined in seven
solvents at the same temperature. These values are reported in
Table 2, together with the F : Cl specific rate ratios (determined
in conjunction with values from Table 1).
Discussion
The comparisons of the specific rates of solvolysis of
n-octyl fluoroformate and n-octyl chloroformate at 24.2 ЊC
(Table 2) show only modest F/Cl leaving-group effects, with
values ranging from 0.62 in 100% ethanol to 15.1 in 60%
ethanol. In ethanol and methanol, the chloroformate, reacts
marginally faster but, for all five of the comparisons in
aqueous–organic solvents, the fluoroformate is somewhat
faster, despite the stronger carbon–fluorine bond. Similar ratios
have been observed previously for the solvolyses of other
haloformate esters.14–16 The small leaving-group effects have
been ascribed14–16,21 to the operation of an addition–
elimination pathway in which the addition step is rate-
determining [eqn. (3)]. The rate of attack by the solvent at the
carbonyl carbon, to give a tetrahedral intermediate without
the rupture of the carbon–halogen bond, would be expected to
be only slightly dependent on the nature of the halogen. Due to
the higher electronegativity of fluorine leading to an increase in
the electron deficiency at the carbonyl carbon, it is reasonable
to predict that the attack will be somewhat easier for the fluoro-
formate substrate. Parallel arguments have been applied to
situations where small F : Cl leaving-group ratios have been
found in nucleophilic aromatic substitutions.22 After sub-
sequent, or concurrent, deprotonation by a second solvent
molecule (acting as a general base), regeneration of the
carbonyl group, in a fast step, will now favor loss of halide ion.
In contrast, nucleophilic substitution reactions in which the
carbon–halogen bond is believed to be broken in the rate-
determining step have considerably lower values for the F/Cl
leaving-group effect. For SN1 reactions, values of about 10Ϫ5
for tert-butyl halide solvolyses,23 10Ϫ6 for triphenylmethyl
halide solvolyses,23 and 10Ϫ7 for 4-(N,N-dimethylamino)benzoyl
halide solvolyses21 have been reported. The values are higher,
Table 2 Specific rates of solvolysis of n-octyl chloroformatea in pure
and mixed solvents at 24.2 ЊC and the F : Cl specific rate ratios
solventb
105 k/sϪ1 c
F : Cld
100% EtOH
80% EtOH
60% EtOH
100% MeOH
80% Acetone
80% Dioxane
80% TFEf
2.39 0.03e
7.37 0.09
9.42 0.19
8.51 0.12
0.775 0.012
1.56 0.06
0.217 0.008
0.62 0.02
8.09 0.12
15.1 0.9
0.95 0.02
2.86 0.07
1.90 0.09
10.2 0.4
a Substrate concentration of 0.0100 mol dmϪ3, unless otherwise indi-
cated. b See footnotes in Table 1. c See footnotes in Table 1. d The specific
rates of solvolysis of n-octyl fluoroformate are from Table 1. e Value
of 2.70 ( 0.04) × 10Ϫ5 sϪ1 at 25.0 ЊC. f Substrate concentration of
0.0200 mol dmϪ3
.
but still well below unity, for SN2 reactions, with values for
reactions of methyl halides of 3 × 10Ϫ2 for hydrolysis,24 9 × 10Ϫ2
for reaction with hydroxide ion,25 and 1.2 × 10Ϫ3 for reaction
with iodide ion.26 It is of interest that acetyl halides undergoing
solvolysis in 75% acetone23 show a very low F/Cl leaving-group
effect of 1.3 × 10Ϫ4, supporting the concept27 of an appreciable
difference in mechanism for these solvolyses, relative to that for
haloformate solvolyses in similar solvents; for example, in the
very similar solvent 80% acetone, the n-octyl haloformates
J. Chem. Soc., Perkin Trans. 2, 2002, 240–243
241

View Article Online
Table 3 Coefficients from extended Grunwald–Winstein correlations of the solvolyses of n-octyl fluoroformate and a comparison with corre-
sponding values for the solvolyses of three chloroformate esters
Substrate
Scales
na
lb
mb
cc
l/m
Rd
Fe
OctOCOF
OctOCOF
NT, YCl
NT, YCl
NT, YCl
NT, YCl
NT, YCl
NT, YCl
NT, YOTS
NT, Y
23
19f
21
19
28
19
25
26
1.80 0.13
1.67 0.07
1.68 0.10
1.59 0.09
1.56 0.09
1.61 0.09
1.56 0.13
1.31 0.10
0.79 0.06
0.76 0.03
0.57 0.06
0.58 0.05
0.55 0.03
0.46 0.04
0.96 0.08
0.88 0.06
0.13 0.34
2.28
2.20
2.95
2.74
2.84
3.50
1.63
1.49
0.959
0.988
0.973
0.977
0.967
0.975
0.944
0.962
114
325
159
Ϫ0.08 0.18
0.12 0.41
0.16 0.07
0.19 0.24
0.04 0.22
Ϫ0.08 0.38
Ϫ0.18 0.31
PhOCOClg
MeOCOClh
EtOCOCli
4-NO2BzOCOClj
OctOCOF
179
157
91
OctOCOF
143
a Number of solvents included in the correlation. b From eqn. (1), with associated standard error. c From eqn. (1), accompanied by the standard error
of the estimate. d Correlation coefficient. e The F-test value. f Omitting the four TFE–ethanol solvents (see text). g Values from ref. 3. h Values from
ref. 4. i Values from ref. 5. j For 4-nitrobenzyl chloroformate; values from ref. 7.
show an F : Cl ratio of 10.2 (Table 2), higher by almost five
orders of magnitude.
these data points are omitted from the correlation (Table 3), the
l and m values are only very slightly reduced but considerably
improved values for the correlation coefficient (0.988 relative
to 0.959) and for the F-test value (325 relative to 114) are
observed.
Due to both a low reactivity (consistent with the very low F :
Cl ratios for SN1 reactions) and a strong autocatalysis by the
acid developed,13,28 it has not been possible to use 1-adamantyl
fluoride to develop a YF scale. However, if the currently con-
sidered mechanism is as in eqn. (3), with the association step
rate determining, this is not necessarily a handicap to the
present investigation. The ideal YX scale would be based on a
solvolysis which mimics the movement of the π electrons of the
carbonyl group onto the oxygen. In several studies of chloro-
formate esters reacting by this mechanism, the YCl scale
was found to be appropriate for use within the extended
Grumwald–Winstein equation (Table 3) and it is reasonable to
predict that this scale will also be appropriate for use within
analyses, when eqn. (3) applies, of the specific rates of solvolysis
of other haloformate esters.
We have also carried out the analyses (Table 3) with the YCl
values replaced by YOTS values. In the establishment of the YOTS
scale negative charge is developing on oxygen, but it is dispersed
over the three oxygens of the developing toluene-4-sulfonate
anion (4-CH3C6H4SO3Ϫ). The goodness of fit is slightly inferior,
indicating that the development on one atom (the incipient
chloride ion) is preferable, as a similarity model, to the
development on oxygen—but spread over three atoms. Sur-
prisingly, since it contains a nucleophilic component,2,19a,b the
original Y scale, based on tert-butyl chloride solvolysis,29 in
conjunction with the NT scale gives reasonable correlations
(Table 3). However, due to the nucleophilic contribution to the
solvent ionizing power scale, a low l value is associated with this
analysis.
It is found (Table 3 and Fig. 1) that the use of NT values in
Conclusions
A mechanism for the solvolyses of n-octyl fluoroformate
involving an association–dissociation (addition–elimination)
pathway, with the association step being rate determining
[eqn. (3)], is supported by two types of evidence obtained in
this study. Firstly, the F : Cl specific rate ratios obtained in a
comparison with the corresponding solvolysis of n-octyl
chloroformate are close to unity with, in most cases, the fluoro-
formate reacting slightly faster and, secondly, the l and m values
obtained from an extended Grunwald–Winstein equation
[eqn. (1)] treatment are very similar to those previously
observed for the solvolyses of several chloroformate esters
which have been shown to solvolyze with the association step
of an association–dissociation pathway being rate determining.
The correlations are best carried out using the NT scale in
conjunction with the YCl scale. Use of the YOTS scale gives
slightly inferior correlations. Use of the original Y scale gives
surprisingly good correlations but the l values are low. These
low l values result from a component governed by solvent
nucleophilicity being incorporated into the tert-butyl chloride
based Y scale.
Fig. 1 Plot of log (k/ko) for solvolyses of n-octyl fluoroformate at 24.2
ЊC against (1.80NT ϩ 0.79YCl); ᭹ EtOH (aq), ᭺ MeOH (aq), ᮀ
Acetone (aq), ꢀ TFE (w/w; aq), ᭝ TFE–EtOH.
conjunction with YCl values leads to acceptable correlations,
with l and m values similar to those obtained in analyses of
the specific rates of solvolysis of phenyl,3 methyl,4 and ethyl5
chloroformates. The major difference is a somewhat larger
m value for the n-octyl fluoroformate solvolyses, possibly
reflecting the need for increased solvation of the developing
negative charge on the carbonyl oxygen in the presence of the
more electronegative fluorine attached at the carbonyl carbon.
Visual inspection of Fig. 1 indicates that the measures of good-
ness of fit would probably be considerably improved if the
four solvolyses in TFE–ethanol mixtures were omitted from
the correlation. Also in earlier correlations of other halo-
formate esters, it was found that the data points for these
solvent systems usually lay below the correlation line.3–5,7 When
Experimental
The n-octyl chloroformate (Aldrich) was purified by fractional
distillation at reduced pressure. The n-octyl fluoroformate
was prepared by the stirred reaction for four days, under nitro-
gen and at room temperature, of n-octyl chloroformate (5.0 g,
0.026 mol), potassium fluoride (2.8 g, 0.048 mol), and 18-
crown-6 (0.34 g, 0.0013 mol).30 Fractional distillation gave
1.20 g (26%) n-octyl fluoroformate, bp 51–53 ЊC/2.5 mmHg.
Anal. Calcd. for C9H17O2F: C, 61.39; H, 9.65; F, 10.79. Found:
C, 61.42; H, 9.93; F, 10.21.
242
J. Chem. Soc., Perkin Trans. 2, 2002, 240–243

View Article Online
The solvents were purified as previously described.17 The
kinetic runs for the chloroformate17 and fluoroformate13,16
esters were also carried out as previously described.
The multiple regression analyses were carried out using
the ABSTAT statistical package (Anderson-Bell, Arvada,
Colorado, USA).
11 D. N. Kevill, J. B. Kyong and F. L. Weitl, J. Org. Chem., 1990, 55,
4304.
12 D. N. Kevill and F. L. Weitl, Tetrahedron Lett., 1971, 707.
13 D. N. Kevill and J. B. Kyong, J. Org. Chem., 1992, 57, 258.
14 M. Green and R. F. Hudson, J. Chem. Soc., 1962, 1055.
15 S. I. Orlov, A. L. Chimishkyan and M. S. Grabarnik, J. Org. Chem.
USSR (Engl. Transl.), 1983, 19, 1981.
16 A. Queen and T. A. Nour, J. Chem. Soc., Perkin Trans. 2, 1975,
935.
17 D. N. Kevill and S. W. Anderson, J. Org. Chem., 1991, 56, 1845.
18 T. W. Bentley and G. Llewellyn, Prog. Phys. Org. Chem., 1990, 17,
121.
Acknowledgements
D. N. K. thanks Professor H. Mayr (Universität München) for
19 (a) T. W. Bentley and G. E. Carter, J. Am. Chem. Soc., 1982, 104,
5741; (b) D. N. Kevill and M. J. D’Souza, J. Chem. Res. (S), 1993,
174; (c) J. S. Lomas, M. J. D’Souza and D. N. Kevill, J. Am. Chem.
Soc., 1995, 117, 5891.
20 (a) F. L. Schadt, T. W. Bentley and P. v. R. Schleyer, J. Am. Chem.
Soc., 1976, 98, 7667; (b) B. Allard and E. Casadevall, Nouv. J. Chim.,
1985, 9, 565.
his hospitality whilst this manuscript was being prepared.
References
1 S. Winstein, E. Grunwald and H. W. Jones, J. Am. Chem. Soc., 1951,
73, 2700.
2 D. N. Kevill, in Advances in Quantitative Structure-Property
Relationships, ed. M. Charton, JAI Press, Greenwich, CN, 1996,
vol. 1, pp. 81–115.
3 D. N. Kevill and M. J. D’Souza, J. Chem. Soc., Perkin Trans. 2, 1997,
1721.
4 D. N. Kevill, J. C. Kim and J. B. Kyong, J. Chem. Res. (S), 1999,
150.
5 D. N. Kevill and M. J. D’Souza, J. Org. Chem., 1998, 63, 2120.
6 J. B. Kyong, Y.-G. Kim, D. K. Kim and D. N. Kevill, Bull. Korean
Chem. Soc., 2000, 21, 662.
21 B. D. Song and W. P. Jencks, J. Am. Chem. Soc., 1989, 111, 8470.
22 (a) J. Miller, Nucleophilic Aromatic Substitution, Elsevier,
Amsterdam, 1968; (b) J. F. Bunnett, E. W. Garbish, Jr. and
K. M. Pruitt, J. Am. Chem. Soc., 1957, 79, 385.
23 C. G. Swain and C. B. Scott, J. Am. Chem. Soc., 1953, 75, 246.
24 I. Fells and E. A. Moelwyn Hughes, J. Chem. Soc., 1959, 398.
25 D. N. Glew and E. A. Moelwyn Hughes, Proc. R. Soc. London, 1952,
A211, 254.
26 R. H. Bathgate and E. A. Moelwyn Hughes, J. Chem. Soc., 1959,
2642.
7 J. B. Kyong, B.-C. Park, C.-B. Kim and D. N. Kevill, J. Org. Chem.,
2000, 65, 8051.
27 T. W. Bentley, G. Llewellyn and J. A. McAlister, J. Org. Chem., 1996,
61, 7927.
28 (a) C. G. Swain and R. E. T. Spalding, J. Am. Chem. Soc., 1960, 82,
6104; (b) C. A. Bunton and J. H. Fendler, J. Org. Chem., 1966, 31,
2307.
29 E. Grunwald and S. Winstein, J. Am. Chem. Soc., 1948, 70, 846.
30 J. Cuomo and R. A. Olofson, J. Org. Chem., 1979, 44, 1016.
8 C. Faurholt and J. C. Gjaldbaek, Dansk. Tids. Farm., 1945, 19, 255;
C. Faurholt and J. C. Gjaldbaek, Chem Abstr., 1946, 40, 513.
9 D. N. Kevill, M. W. Bond and M. J. D’Souza, J. Org. Chem., 1997,
62, 7869.
10 D. N. Kevill and M. J. D’Souza, Can. J. Chem., 1999, 77, 1118.
J. Chem. Soc., Perkin Trans. 2, 2002, 240–243
243