Structural and solvent effects on rates of solvolysis of secondary alkyl substrates - An update

Article abstract of DOI:10.1002/poc.862

Rate constants for solvolyses of secondary alkyl tosylates in fluorinated media [including hexafluoropropan-2-ol (HFIP), hexafluoroacetone sesquihydrate, and trichloromethylbis(trifluoromethyl)carbinol] are reported. Rates of solvolysis of 2-adamantyl p-toluenesulfonate in 97% (w/w) HFIP-water at 25°C are neither retarded by the addition of NaOTs nor accelerated greatly by NaClO₄). The α-deuterium kinetic isotope effect for solvolyses of 1-(1-adamantyl)ethyl methanesulfonate in 20% acetone-water at 25°C is 1.14. Mechanistic implications of these results are discussed. Copyright

Full text of DOI:10.1002/poc.862

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2005; 18: 96–100
Published online 21 September 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.862
Structural and solvent effects on rates of solvolysis
of secondary alkyl substrates---an updatey
T. William Bentley* and Ian Roberts
Department of Chemistry, University of Wales, Swansea, Singleton Park, Swansea SA2 8PP, UK
Received 26 September 2003; revised 4 February 2004; accepted 31 May 2004
ABSTRACT: Rate constants for solvolyses of secondary alkyl tosylates in fluorinated media [including hexafluor-
opropan-2-ol (HFIP), hexafluoroacetone sesquihydrate, and trichloromethylbis(trifluoromethyl)carbinol] are re-
ported. Rates of solvolysis of 2-adamantyl p-toluenesulfonate in 97% (w/w) HFIP–water at 25 ꢀC are neither
retarded by the addition of NaOTs nor accelerated greatly by NaClO4. The ꢀ-deuterium kinetic isotope effect for
solvolyses of 1-(1-adamantyl)ethyl methanesulfonate in 20% acetone–water at 25 ꢀC is 1.14. Mechanistic implica-
tions of these results are discussed. Copyright # 2004 John Wiley & Sons, Ltd.
KEYWORDS: solvolysis; kinetics; fluorinated solvents; ꢀ-deuterium kinetic isotope effects
INTRODUCTION
have also been examined.8 Solvent effects on solvolytic
reactivity can be correlated with solvent ionizing power
(Y or YOTs),5 solvent nucleophilicity (N),7,9 and a para-
meter (I) which represents solvation effects adjacent to
the reaction site (often aromatic rings).10 As multipara-
meter correlations require data for a diverse range of
solvents, we now report new data for HFIP and explora-
tory kinetic studies of several other fluorinated alcohol
solvents.
Direct measurements of very rapid rates of reactions of
some aryl-stabilized secondary carbocations generated
by flash photolysis (e.g. photoheterolysis of halides) have
been made,11 and solvolysis of 1 in TFE recently pro-
vided the first example of an SN1 reaction in which both
the formation and decomposition of a cationic intermedi-
ate could be monitored.1 Also, lifetimes of less stable
secondary cationic intermediates, such as substituted 1-
phenylethyl cations, have been estimated using a trapping
procedure (azide clock method).12 However, for many
solvolyses, there is only indirect evidence for the forma-
tion of a cationic intermediate. Typical, long-established,
investigations of solvolytic reactions include data on
product composition and stereochemistry and structural
and solvent effects on rates, possibly with additional
kinetic data on isotope effects, salt effects, rates of
substrate racemization or 18O scrambling in carboxylate
or sulfonate esters.7 Some of these studies have led to
complex and controversial mechanistic interpretations, to
which we now add additional data and comments.
Rates of solvolyses of neutral secondary substrates
(R1R2CHX) span a huge range ( >1020), e.g. from solvo-
lysis of bis(p-methoxyphenyl)chloromethane (1)1 to 7-
norbornyl derivatives;2 in addition, there is a >1016-fold
dependence on the leaving group (X).3 Rates depend
primarily on the stabilization of positive charge by the
R1R2CH unit, now supported by the link between intrin-
sic (gas-phase) thermodynamic stabilities of cations and
solvolytic reactivity, which has recently been extended to
secondary substrates.4 Solvent effects on rates of SN1
reactions of adamantyl tosylates (AdOTs or p-toluene-
sulfonate) in protic solvents, varying over 109-fold from
tert-butanol to aqueous sulfuric acid, depend primarily on
the solvent polarity or YOTs scale of solvent ionizing
power.5
Relative rates of solvolyses of 2-propyl and 2-AdOTs
(2) vary >105-fold, from 134 in 1,1,1,3,3,3-hexafluoro-
propan-2-ol (HFIP) to 1.1 ꢁ 10ꢂ3 in ethanol,6 and are so
strongly influenced by solvent nucleophilicity that un-
hindered secondary substrates react with inversion of
stereochemistry by an SN2 mechanism in more nucleo-
philic solvents.6,7 In addition to typical solvolyses in
ethanol, ethanol–water mixtures and acetic acid,7 solvo-
lysis research now includes studies of weakly nucleophi-
lic, fluorinated alcohols such as 2,2,2-trifluoroethanol
(TFE)1 and HFIP. Several other fluorinated alcohols
*Correspondence to: T. W. Bentley, Department of Chemistry, Uni-
versity of Wales, Swansea, Singleton Park, Swansea SA2 8PP, UK.
E-mail: cmsolvol@swan.ac.uk
ySelected article presented at the Seventh Latin American Conference
on Physical Organic Chemistry (CLAFQO-7), 21–26 September 2003,
RESULTS
´
Florianopolis, Brazil.
Salt effects on the solvolysis of 2-AdOTs in aqueous
HFIP are given in Table 1. As neither 0.03 M LiClO4 nor
Contract/grant sponsor: Science and Engineering Research Council;
Contract/grant number: GR/A/95536.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 96–100

RATES OF SOLVOLYSIS OF SECONDARY ALKYL SUBSTRATES
97
Table 1. Salt effects on rate constants for solvolysis of 2-
adamantyl tosylate in HFIP–water at 25 ꢀC
Table 3. Rate constants for solvolysis in hexafluoroacetone
sesquihydrate
Solvent (%, w/w)
Salt
[Salt] (M)
k (10ꢂ5 sꢂ1
)
Substrate
Monitoring
T ( ꢀC)
k (sꢂ1
)
97.0a
97.0a
97.0c
97.0a
89.3e
89.3e
None
7.52 ꢃ 0.03b
7.44 ꢃ 0.05
7.7 ꢃ 0.1
2-AdOTs (2) Spectrophometrica 25.6 (9.8 ꢃ 0.1) ꢁ 10ꢂ4
NaOTs
LiClO4
NaClO4
None
0.008
<0.03d
<0.03d
2-AdOTs (2) Conductimetricb
25.0 (8.93 ꢃ 0.03) ꢁ 1ꢂ0ꢂ6 4
PriOTs
Spectrophometricb 25.1 (7.3 ꢃ 0.3) ꢁ 10ꢂ2
7.8 ꢃ 0.1
ButCl
Conductimetrica
25.0 (1.9 ꢃ 0.1) ꢁ 10
1.88 ꢃ 0.05
2.04 ꢃ 0.05
a Duplicate measurements; errors shown are average deviations.
b Single measurements of rate constant.
NaClO4
0.03
a Monitored spectrophotometrically at least in duplicate at 24.85 ꢃ 0.1 ꢀC.
b
Single conductimetric result, 105k ¼ 8.69 ꢃ 0.02; lit. 8.7 (Ref. 6), 7.17
(Ref. 8); spectrophotometric result, 7.97 (Ref. 13).
c Single measurement, monitored spectrophotometrically.
d Incompletely soluble.
footnote d); rates increase on addition of water to HFIP
for MeOTs and EtOTs (Table 2), but decrease for 2-
AdOTs (Table 1).
e Monitored spectrophotometrically at 25.0 ꢃ 0.05 ꢀC.
Hexafluoroacetone hydrate, a 1,1-diol (i.e. a tertiary
alcohol), is a low-melting solid, but solvolysis rates in
hexafluoroacetone sesquihydrate [ca 95% (w/w) 1,1-
diol–water)15 at 25 ꢀC are given in Table 3. Preliminary
studies of other fluorinated tertiary alcohols were made.
Perfluoro-2-methyl-2-pentanol (3)16 reacted very slowly
with 2-AdOTs (t >5000 min), but acetate buffers did not
appear to be soluble; adequately buffered solutions con-
taining 1–2% water gave t >7000 min. Supporting data
from single spectrophotometric measurements for mix-
tures of 3 and HFIP gave the following results for 105k
(sꢂ1) at 25 ꢀC: HFIP, 13.2 ꢃ 0.5; 51.5% (w/w) 3–HFIP,
2.8 ꢃ 0.2; 74.5% (w/w) 3–HFIP, 0.70 ꢃ 0.02; hence k ꢄ
10ꢂ6 sꢂ1 in 100% 3. Conductimetric studies of 97%
(w/w) 3 for tert-butyl bromide were inconclusive, but
1-adamantyl bromide gave k ꢄ 3 ꢁ 10ꢂ4 sꢂ1 at 25 ꢀC
(but not reproducible). Other disadvantages are that 3 is
toxic16 and difficult to recycle by distillation (even from
concentrated H2SO4) after kinetic runs.
Table 2. Rate constants for solvolysis of alkyl tosylates in
HFIP and HFIP–water
Rate constant (10ꢂ6 sꢂ1
EtOTsa PriOTsb
)
Solvent
MeOTsa
97% HFIPc
HFIPf
6.2 ꢃ 0.5d
2.93 ꢃ 0.1
0.98 ꢃ 0.1e
1.69 ꢃ 0.23
1.87 ꢃ 0.07
0.99 ꢃ 0.07g
a Determined at 100.05 ꢃ 0.05 ꢀC by spectrophotometric analysis of
quenched aliquots.
b Determined at 25.01 ꢃ 0.02 ꢀC by spectrophotometric analysis of
quenched aliquots.
c Duplicate measurements for MeOTs and EtOTs.
d Literature values determined conductimetrically: 5.5 ꢃ 0.3 (Ref. 14); 1.02
and 1.69 (5 ꢁ 10ꢂ3 M and 10–1 M unbuffered substrate, Ref. 8).
e Literature value determined conductimetrically: 1.55 (Ref. 14, extrapo-
lated).
f Duplicate measurements for EtOTs, single for others; 10ꢂ2 M NaOAc
added as buffer.
g In agreement with the value of 1.1 obtained by continuous spectro-
photometric monitoring (Ref. 6).
Perfluoro-tert-butanol is available commercially, but is
expensive, has a small liquid range (b.p. 45 ꢀC)16 and is
fairly toxic, whereas trichlorobis(trifluoromethyl)carbi-
nol (4) is less volatile and has a higher liquid range (m.p.
ꢂ2 ꢀC and b.p. 136–138 ꢀC).17 Various preparations of 4
were attempted (see Refs 17 and 18 and Experimental),
but a very pure sample was not obtained—the product
absorbed with A > 2 up to 325 nm (hence precluding
spectrophotometric monitoring at 273 nm), and conducti-
metric monitoring of reactions at 25 ꢀC buffered with
acetamide gave k ꢄ 3 ꢁ 10ꢂ2 sꢂ1 for ButCl and k ꢄ
6 ꢁ 10ꢂ3 sꢂ1 for 2-AdOTs. A solution of PriBr (0.05 M)
at 100 ꢀC, buffered with 2,6-lutidine (0.1 M), monitoring
NaClO4 was completely soluble in 97% HFIP–water,
90% HFIP–water was also investigated. Various batches
of HFIP were used to study solvolysis rates; a typical
value in the presence of NaOAc buffer was
k ¼ 1.4 ꢁ 10ꢂ4 sꢂ1, in satisfactory agreement with our
earlier data6 (k ¼ 1.47 ꢁ 10ꢂ4 sꢂ1). Inclusion of NaOAc
buffer led to reliable data for solvolyses of alkyl tosylates
in HFIP (Table 2), but two other buffers (2,6-lutidine and
proton sponge) gave unreliable results. Results for solvo-
lysis of MeOTs in 97% HFIP agree with our previous
work,14 and contrast with lower published values (Table 2,
Table 4. ꢀ-Deuterium kinetic isotope effects for solvolysis of 1-(1-adamantyl)ethyl sulfonates (5) at 25.0 ꢀC
Solvent
X
kH (sꢂ1
)
kD (sꢂ1
)
kH/kD
97% HFIPa
20% Ac
OTs
(5.2 ꢃ 0.2) ꢁ 10ꢂ3
(4.8 ꢃ 0.3) ꢁ 10ꢂ3
1.1 ꢃ 0.1b
OMsd
(7.70 ꢃ 0.05) ꢁ 10ꢂ4
(6.74 ꢃ 0.05) ꢁ 10ꢂ4
1.14 ꢃ 0.02e
a Determined spectrophotometrically from five measurements for kH and one for kD.
b Value in 98% HFIP is 1.116, determined conductimetrically for the pentamethylbenzenesulfonate (Ref. 20).
c Determined conductimetrically in 20% (v/v) acetone–water; data quoted are for the three most reliable measurements for kH and two for kD.
d Ms ¼ methanesulfonate.
e If all measurements are used (six for kH and five for kD), the value is 1.14 ꢃ 0.03.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 96–100

98
T. W. BENTLEY AND I. ROBERTS
Table 5. Rate ratios for 1-(1-adamantyl)ethyl (5)–pinacolyl
(6) either directly observed for solvolysis (kt) or calculated for
ionization (k1)a of sulfonates (X) at 25 ꢀC
the change in the 6H doublet (isoprene is the product) by
NMR, showed k ꢄ 6 ꢁ 10ꢂ6 sꢂ1 [slower than in 97%
HFIP (4.5 ꢁ 10ꢂ5)];19 as reaction of ButBr in 4 is likely
to be faster than for 97% HFIP (2.2 ꢁ 10ꢂ2),19 the rate
ratio ButBr/PriBr (ꢀ-Me/H ratio) is likely to exceed the
high value of 106.2 for 97% HFIP.19
Rate constants and ꢀ-deuterium kinetic isotope effects
(kH/kD) for solvolysis of 1-(1-adamantyl)ethyl sulfonates
(5) in relatively polar solvents (Table 4) supplement other
published studies.20
a
b
Solvent
X
(k5/k6)t (k5/k6)1
kH/kD
80% (v/v) EtOHc
97% (w/w) TFEc
97% (w/w) TFEe
20% (v/v) Af
OBs
OBs
OTs
OMs
OTs
0.81
3.8
3.7
2.7
8.1
2.3d
7.7d
1.147
1.111
1.107
1.14
(2.7)g
97% (w/w) HFIPh
(1.116)i
a Ionization rates are calculated for formation of a contact ion pair.
b
For solvolyses of 1-(1-adamantyl)ethyl sulfonates (5).
c Data from Ref. 20.
d
From a steady-state treatment of secondary kinetic isotope effects; see
Ref. 20.
e Data from Ref. 25; kH/kD refers to brosylates.
f Data for 20% acetone–water from Table 4 and Ref. 26.
g
Ratio of observed rate constants, assuming negligible ion pair return for
solvolyses in a highly aqueous medium (see also Refs 27 and 28).
h
Data from Table 4 and Ref. 6.
For the pentamethylbenzenesulfonate in 98% HFIP (Ref. 20).
i
pinacolyl (6); the 10-fold variation in relative rates (Table 5)
can be explained simply by the greater sensitivity of 5 to Y,
and possibly also to a lower sensitivity to N (to explain
the new results for 20% acetone).25 An increased sensi-
tivity to Y is often associated with a decreased sensitivity
to N, within the SN2–SN1 spectrum.6 Alternatively, results
can be explained by the general model of Gregoriou and
Varveri,29 which includes the formation of an intermedi-
ate ‘with extensive covalencies with the leaving group
and with the incoming solvent’.
DISCUSSION
Because rapid rearrangement prevents ion pair return
for solvolysis of 6, observed rates (kt) are equal to
ionization rates (k1); a comprehensive and complex
analysis of ꢀ- and ꢁ-deuterium kinetic isotope effects,
and products for solvolysis of 5 in various solvents, with a
simplex optimization of 46 parameters from 86 experi-
ments20 gave ionization rates (k1) up to three-fold greater
than observed rates (kt), with more ion pair return in 80%
ethanol than in 97% TFE. The lower kH/kD for 5 (com-
pared with 1.15 for 6 and 1.22 for 2) was explained by a
proportion of the reaction proceeding through a transition
state having a rearranged structure.20
Both the simple and complex interpretations (see the
previous two paragraphs) are in agreement that 5 can
ionize up to eight times faster than 6, but the complex
interpretation does not provide insights as to why. The
simple explanation6,25,27,29 is that there are enhanced
electronic effects because (i) additional carbon atoms in
5 donate electrons and (ii) contributions from nucleophi-
lic solvation30,31 are less in weakly nucleophilic media.
Presumably owing to changes in acidity which pro-
vide electrophilic assistance to ionization, the tertiary
alcohol 4 (pKa ¼ 5.1)17 is an even more highly ionizing
solvent (YOTs ꢄ 5.4), than the secondary alcohol HFIP
(pKa ¼ 9.3;17 YOTs ¼ 3.82)5, continuing a trend from the
primary alcohol TFE, (pKa ¼ 12.4;17 YOTs ¼ 1.77).5 Not
surprisingly, the 1,1-diol HFA hydrate (pKa ¼ 6.6)32 is a
Salt effects (Table 1) show that rates for 2 are unaffected
( ꢃ 2%) by the addition of 0.008 M tosylate anion, so the
common ion effect is small or absent and the small rate
acceleration in the presence of 0.03 M perchlorate appears
to be a normal salt effect. The results are consistent with
the accepted interpretation that products are formed by
nucleophilic attack before formation of a free cation.6
Furthermore, extrapolations of measured cation life-
times11 indicate that even the tert-butyl cation may
have a lifetime shorter than the time for one bond
vibration.21 Secondary alkyl cations lacking stablizing
groups such as 2-adamantyl will have shorter lifetimes;
even the resonance-stabilized secondary 2-deoxyglucosyl
oxacarbenium ion may not be solvent equilibrated in
typical alcoholic solvolysis media.22
Short or negligible lifetimes of cations formed during
the solvolysis of unactivated secondary substrates are
probably the main factor underlying a long-running
controversy:6,20 e.g. does 18O scrambling of oxygen in
‘unreacted’sulfonates indicate ion pair return on the main
reaction pathway20 or a side reaction via a concerted
polar mechanism?23,24 Indirect mechanistic probes lead
to different interpretations or emphases, illustrated by the
results in Table 5.
Observed rates (kt) for solvolysis of 1-(1-adamanty-
l)ethyl sulfonates (5) are similar to or faster than those of
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 96–100

RATES OF SOLVOLYSIS OF SECONDARY ALKYL SUBSTRATES
99
more ionizing solvent (YOTs ¼ 4.6) than HFIP. However,
rates of solvolysis of 2 in a series of primary alcohol
solvents (CnF2n þ 1CH2OH, with n ¼ 1, 2, 3 and 7) are
within a factor of 3,8 in contrast to our observation that 3
(YOTs ꢄ 2) is less highly ionizing than 4 (YOTs ꢄ 5.4).
fraction at 56–57 ꢀC. This fraction was then redistilled
under N2 through a triple-pass Widmer column either
from 3A molecular sieves (activated at 300 ꢀC/1 mmHg
for 2 days) or from P2O5 (3 g lꢂ1), collecting the middle
80% fraction (b.p. 58 ꢀC; lit.19 58–59 ꢀC). The pure
solvent was stored under N2 in the collection vessel
(fitted with a tap adaptor and septum cap) and transfers
were made by syringe. Binary solvents mixtures were
stored in a glass-stoppered flask, sealed with Parafilm at
5 ꢀC for up to 6 weeks; under these conditions (and in
contrast to storage at room temperature), the solvent
composition did not change sufficiently to affect the
solvolysis rate constant for 2.
Hexafluoroacetone sesquihydrate (HFAꢅ1.5H2O), a
commercial sample (PCR), and acetone (AR grade)
were used without further purification. Water was freshly
distilled in an all-glass apparatus and was passed through
an ion-exchange column (Amberlite MB-1) immediately
prior to use.
Perfluoro-2-methylpentan-2-ol (3) was supplied by
K. V. Scherer (b.p. 93 ꢀC, 728 mmHg).16 Trichloromethy-
lbis(trifluoromethyl)carbinol (4) was prepared from
CCl4, BuLi and (CF3)2CO (from HFA hydrate–H2SO4)
in THF at ꢂ110 ꢀC: b.p. 136–138 ꢀC, lit.17 136–138 ꢀC;
yield, ca 30% maximum three (with frequent failures),
lit.17 50%, using (CF3)2CO or ‘low yield’;18 even after
three distillations the sample contained 1H NMR signals
at ꢂ 1–3 (various signals), 6 (s) and 9 (s), possibly because
of plasticizers from the PVC tubing.
CONCLUSIONS
Structural effects3,4 and solvent effects (including both
electrophilic and nucleophilic solvation effects)5 make
dominant contributions to the observed rates of typical
secondary solvolyses, and the nature of the initial hetero-
lysis step may change because of nucleophilic solvent
assistance or nucleophilic solvation (appropriate termi-
nology is still under discussion30,31). Other effects on
reactivity such as a change in rate-determining step20 or
solvation effects adjacent to the reaction site10 are rela-
tively small. The evidence for a mechanistic change to
rate-determining separation of contact ion pairs (i.e.
return from contact ion pairs)20 relies on ambiguous
18O scrambling data23,24 and on complex interpretations
of secondary deuterium kinetic isotope effects,20 in cases
where an ion pair may not survive long enough to
undergo one vibration.21
EXPERIMENTAL
Materials. Methyl and ethyl tosylates were commercial
samples (BDH) and 2-propyl and 2-adamantyl tosylates
(2) were prepared by standard methods;27 all were
recrystallized and dried before use. 1-(1-Adamantyl)etha-
nol derivatives were prepared from the 1-adamantyl-
methyl ketone (Aldrich) using lithium aluminium
hydride or deuteride, were converted to sulfonates by
standard methods27 and were recystallized from light
petroleum: 1-(1-adamantyl)ethanol (m.p. 79–80 ꢀC;
lit.33 79.8–80.0 ꢀC); tosylate (m.p. 122.5 ꢀC; lit.33
123.0–124.2 ꢀC); mesylate (m.p. 54–55 ꢀC); ꢀ-D-alcohol
(m.p. 77–78 ꢀC; lit.33 79.5–80.3 ꢀC), with >99.8% iso-
topic purity shown by mass spetrometry; ꢀ-D-tosylate
(m.p. 123–124 ꢀC; lit.33 123.8–124.6 ꢀC); ꢀ-D-mesylate
(m.p. 53.5–54.5 ꢀC). The purity and structure of all
samples were confirmed by 1H NMR spectroscopy.
Sodium p-toluenesulfonate was prepared by neutraliz-
ing the acid with an excess of NaHCO3, followed by
salting out the product with NaCl and recystallizing from
ethanol. Methanesulfonyl chloride was fractionally dis-
tilled from P2O5 (b.p. 58 ꢀC/13 mmHg; lit.34 55 ꢀC/
11 mmHg). Triethylamine was heated under reflux
with phthalic anhydride (1 g/30 ml) and then fractionally
distilled.
Kinetic methods. Spectrophotometric measurements
were made at 273 nm, using a Phillips SP1800 instru-
ment. Kinetic runs at 25 ꢀC were carried out directly in
1.1 ml cuvettes in a water-cooled cell block; cuvette
temperatures were monitored before and after kinetic
runs using a calibrated bare thermistor in a water-filled
cuvette, and varied slightly with ambient temperature.
Substrate concentrations were (1–3) ꢁ 10ꢂ3
M
and
[NaOAc] was ca. 10ꢂ2 M. The cuvette was sealed with
an all-glass stopper, incorporating a sample boat contain-
ing a pre-weighed amount of substrate; the thermally
equilibrated cuvette was shaken vigorously twice to
dissolve the sample. Kinetic data for simple alkyl tosy-
lates (Table 2) were obtained after reactions in thermo-
stated baths, by analysis of quenched, sealed, 5 ml
ampoules (7 ꢁ 1.5 ml, with two ampoules reserved for
checking the infinity value). Rate constants were com-
puted using LSKIN.35
Acknowledgements
This research was supported by the SERC (UK) with
the award of a postdoctoral fellowhip. We are very
grateful to K. V. Scherer for supplying samples of the
fluorinated alcohol (3) and to P. v. R. Schleyer for helpful
discussions.
HFIP, containing 0–3% water from previous studies,
was recycled as follows. Addition of K2CO3 until alka-
line (and then 20 g lꢂ1 more) was followed by distillation
through a Vigreux column, collecting the middle 80%
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 96–100

100
T. W. BENTLEY AND I. ROBERTS
17. Filler R, Schure RM. J. Org. Chem. 1967; 32: 1217–1219.
REFERENCES
18. Kotun SP, Anderson JDO, DesMarteau DD. J. Org. Chem. 1992;
57: 1124–1131.
19. Bentley TW, Bowen CT, Parker W, Watt CIF. J. Chem. Soc.,
Perkin Trans. 2 1980; 1244–1252.
20. Wilgis FP, Neumann TE, Shiner VJ. J. Am. Chem. Soc. 1990; 112:
4435–4446.
1. Mayr H, Minegishi S. Angew. Chem. Int. Ed. 2002; 41: 4493–
4495.
2. Winstein S, Shatavsky M, Norton C, Woodward RB. J. Am. Chem.
Soc. 1955; 77: 4183–4184.
3. Bentley TW, Christl MC, Kemmer R, Llewellyn G, Oakley JE.
J. Chem. Soc., Perkin Trans. 2 1994; 2531–2538.
4. Abboud JLM, Alkorta I, Davalos JZ, Mu¨ller P, Quintanilla E,
Rossier JC. J. Org. Chem. 2003; 68: 3786–3796.
5. Bentley TW, Llewellyn G. Prog. Phys. Org. Chem. 1990; 17: 121–
158.
6. Bentley TW, Bowen CT, Morten DH, Schleyer PvR. J. Am. Chem.
Soc. 1981; 103: 5466–5475.
7. Bentley TW, Schleyer PvR. Adv. Phys. Org. Chem. 1977; 14: 1–67.
8. Allard B, Casadevall A, Casdevall E, Largeau C. Nouv. J. Chim.
1979; 3: 335–342.
9. Kevill DN. In Advances in Quantitative Structure—Property
Relationships, vol. 1, Charton M (ed). Jai Press: Greenwich,
CT, 1996; 81–115.
10. Kevill DN, Ismail NHJ, D’Souza M. J. Org. Chem. 1994; 59:
6303–6309.
11. McClelland RA. Tetrahedron 1996; 52: 6823–6858.
12. Richard JP, Rothenberg ME, Jencks WP. J. Am. Chem. Soc. 1984;
106: 1361–1372.
13. Maskill H, Thompson JT, Wilson AA. J. Chem. Soc., Perkin
Trans. 2 1984; 1693–1703.
21. Bentley TW, Llewellyn G, Ryu ZH. J. Org. Chem. 1998; 63:
4654–4659.
22. Zhu J, Bennet AJ. J. Org. Chem. 2000; 65: 4423–4430.
23. Dietze PE, Wojciechowski M. J. Am. Chem. Soc. 1990; 112:
5240–5244.
24. Hammett LP. Physical Organic Chemistry (2nd edn). McGraw
Hill: New York, 1970; 165.
25. Bentley TW, Liggero SH, Imhoff MA, Schleyer PvR. J. Am.
Chem. Soc. 1974; 96: 1970–1973.
26. Bentley TW, Bowen CT, Brown HC, Chloupek FJ. J. Org. Chem.
1981; 46: 38–42.
27. Bentley TW, Bowen CT. J. Chem. Soc., Perkin Trans. 2 1978;
557–562.
28. Richard JP, Jencks WP. J. Am. Chem. Soc. 1984; 106: 1373–
1383.
29. Gregoriou GA, Varveri FS. J. Chem. Soc., Perkin Trans. 2 1981;
985–990.
30. Richard JP, Toteva MM, Amyes TL. Org. Lett. 2001; 3: 2225–
2229.
31. Kevill DN, Miller B. J. Org. Chem. 2002; 67: 7399–7406.
32. Middleton WJ, Lindsey RV Jr. J. Am. Chem. Soc. 1964; 86:
4948–4952.
33. Seib RC, PhD Thesis, Indiana University, 1978.
34. Weast RC (ed.). Handbook of Chemistry and Physics. Chemical
Rubber Company: Cleveland, Ohio, 1971: C370.
35. DeTar DF. In Computer Programs for Chemistry, vol. 1, DeTar DF
(ed). Benjamin: New York, 1968; 126–173.
14. Schadt FL, Bentley TW, Schleyer PvR. J. Am. Chem. Soc. 1976;
98: 7667–7674.
15. Gambaryan NP, Rokhlin EM, Zeifman YU, Ching-Yun C,
Knunyants IL. Angew. Chem. Int. Ed. Engl. 1966; 5: 947–956.
16. Scherer KV Jr, Terranova TF, Lawson DD. J. Org. Chem. 1981;
46: 2379–2381.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2005; 18: 96–100