A solvolysis model for 2-chloro-2-methyladamantane based on the linear solvation energy approach

Article abstract of DOI:10.1021/jo049798l

Solvolysis/dehydrohalogenation rates of 2-chloro-2-methyladamantane (CMA) in 15 hydrogen-bond acidic and/or basic solvents are studied. The rates of reaction in these solvents have been correlated with the solvation equation developed by Kamlet, Abraham, and Taft. The linear solvation energy relationship (LSER) derived from this study is given by the following equation: log k = -5.409 + 2.219π*₁ + 2.505α₁ - 1.823β₁ where π*₁, α₁, and β₁ are the solvation parameters that measure the solvent dipolarity/polarizability, hydrogen-bond acidity (electrophilicity), and hydrogen-bond basicity (nucleophilicity). A high correlation coefficient (r = 0.996, SD = 0.191) was achieved. The cavity term, which includes the Hildebrand parameter for solvent cohesive energy density, δ_H, was not found to be statistically significant for this reaction substrate. The resulting equation allows calculated rates of reaction in other solvents and provides insight into the reaction pathway. In a previously reported correlation for another tertiary chloride, tert-butyl chloride (TBC), the coefficients for α₁ and π*₁ are significantly larger and the coefficient for δ_H² is statistically significant. In addition, the coefficient for β₁ in the TBC correlation is positive, rather than negative, indicating that the transition states for TBC and CMA are significantly different. These results demonstrate why the uses of simple solvolytic correlation methods may be invalid even for comparisons of similar type substrates, e.g., tertiary chlorides. Also, these results provide confidence in the use of multiple linear regression analysis for predicting solvolytic rates in additional solvents.

Full text of DOI:10.1021/jo049798l

A Solvolysis Model for 2-Chloro-2-methyladamantane Based on the
Linear Solvation Energy Approach
,†
‡
§
Samuel P. McManus,* Sunil Somani, J. Milton Harris, and R. Andrew McGill
Department of Chemistry, The University of Alabama in Huntsville, Huntsville, Alabama 35899
smcmanus@al.nektar.com
Received February 4, 2004
Solvolysis/dehydrohalogenation rates of 2-chloro-2-methyladamantane (CMA) in 15 hydrogen-bond
acidic and/or basic solvents are studied. The rates of reaction in these solvents have been correlated
with the solvation equation developed by Kamlet, Abraham, and Taft. The linear solvation energy
relationship (LSER) derived from this study is given by the following equation: log k ) -5.409 +
/
/
2
.219π + 2.505R
1
- 1.823â
1
where π , R
1
, and â
1
are the solvation parameters that measure the
1
1
solvent dipolarity/polarizability, hydrogen-bond acidity (electrophilicity), and hydrogen-bond basicity
(
nucleophilicity). A high correlation coefficient (r ) 0.996, SD ) 0.191) was achieved. The cavity
term, which includes the Hildebrand parameter for solvent cohesive energy density, δ , was not
H
found to be statistically significant for this reaction substrate. The resulting equation allows
calculated rates of reaction in other solvents and provides insight into the reaction pathway. In a
previously reported correlation for another tertiary chloride, tert-butyl chloride (TBC), the
/
2
coefficients for R
1
and π1 are significantly larger and the coefficient for δH is statistically
significant. In addition, the coefficient for â
1
in the TBC correlation is positive, rather than negative,
indicating that the transition states for TBC and CMA are significantly different. These results
demonstrate why the uses of simple solvolytic correlation methods may be invalid even for
comparisons of similar type substrates, e.g., tertiary chlorides. Also, these results provide confidence
in the use of multiple linear regression analysis for predicting solvolytic rates in additional solvents.
Medium effects have an extraordinary influence on
many chemical reactions. After early efforts of Berthelot
and Saint-Gilles (1862) and Menschutkin2 (1890) to
demonstrate that solvent media affect chemical reaction
rates, many important contributions have appeared
attempts to correlate solvolysis rates along lines proposed
by Grunwald and Winstein (GW). An important con-
clusion to be drawn from the many studies on correla-
tion of reaction rates in pure or mixed solvents is
that no single solvent parameter will satisfactorily cor-
relate reaction rate data over a wide range of solvent
types.
3
1
dealing with reactions induced by the solvent, i.e.,
solvolysis reactions.3
-11
Several studies have dealt with
The inability to generalize the GW approach has
attracted interest to correlations involving multiple linear
†
Current Address: Nektar Therapeutics, 1112 Church St., Hunts-
1
2
ville, AL 35801.
regression analysis (MLRA). Koppel and Palm devised
the first general MLRA specifically designed for the
correlation of the rate constants. In an attempt to
overcome deficiencies noted earlier, Kamlet, Abraham,
and Taft (KAT) and co-workers took the lead in devising
a more general solvation equation which includes pa-
rameters that describe the solvent dipolarity/polarizabil-
ity, hydrogen-bond acidity and basicity, and formation
of a solute cavity in a solvent.13 This treatment correlates
solvent effects with some solvent-dependent property.
‡
Current Address: Nektar Therapeutics, 409 Discovery Dr., Hunts-
ville, AL 35806.
§
Current address: Naval Research Laboratory, Chemistry Division,
Code 6170, Washington DC 20375
1) (a) Berthelot, M.; Pean De Saint-Gilles, L. Ann. Chim. 1862, 65,
. (b) Berthelot, M.; Pean De Saint-Gilles, L. Ann. Chim. 1862, 66, 5.
(
5
(
c) Berthelot, M.; Pean De Saint-Gilles, L. Ann. Chim. 1863, 68, 225.
(
2) Menschutkin, N. A., Z. Phys. Chem. 1890, 5, 589.
(
3) (a) Grunwald, E.; Winstein, S. J. Am. Chem. Soc. 1948, 70, 846.
(
7
b) Grunwald, E.; Winstein S.; Jones, H. W. J. Am. Chem. Soc. 1951,
3, 2700.
4) Kosower, E. M. An Introduction to Physical Organic Chemistry;
John Wiley and Sons: New York, 1968; pp 259-384.
(
1
4
The method continues to have wide application. Of
interest here is the use of this method for correlating
reaction rates with solvent parameters. As applied to
(
5) (a) Bentley, T. W.; Schleyer, P. v. R. J. Am. Chem. Soc. 1982,
1
1
04, 5741. (b) Bentley, T. W.; Schleyer, P. v. R. J. Am. Chem. Soc.
976, 98, 7658.
(6) (a) Peterson, P. E.; Kelley, R. E., Jr.; Belloli, R.; Sipp, K. A. J.
Am. Chem. Soc. 1965, 87, 5169. (b) Peterson, P. E.; Vidrine, D. W.;
Waller, F. J.; Henrichs, P. M.; Magaha, S.; Stevens, B. J. Am. Chem.
Soc. 1977, 99, 7968.
(10) (a) Reichardt, C. Solvents and Solvent Effects in Organic
Chemistry, 2nd ed.; Verlag Chemie: Weinheim, 1988. (b) Politzer, P.;
Murray, J. S. Quantitative Treatments of Solute/Solvent Inter-
actions; Elsevier: Amsterdam, 1994.
(
7) Streitweiser, A., Jr. Solvolytic Displacement Reactions; McGraw-
Hill Publishing Co.: New York, 1962.
(
(
8) Abraham, M. H. Pure Appl. Chem. 1985, 57, 1155.
(11) (a) Parker, A. J. Quart. Rev. (London) 1962, 16, 163. (b) Parker,
A. J. J. Chem. Soc. 1961, 1328. (c) Parker, A. J. Adv. Phys. Org. Chem.
1967, 5, 173. (d) Parker, A. J. Chem. Rev. 1969, 69, 1.
9) Kamlet, M. J.; Abboud, J. L. M.; Taft, R. W. In Progress in
Physical Organic Chemistry; Taft, R. W., Ed.; John Wiley and Sons:
New York, 1981; Vol. 13, pp 485-630.
(12) Koppel, I. A.; Palm, V. A. Org React. 1971, 8, 291.
1
0.1021/jo049798l CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/13/2004
J. Org. Chem. 2004, 69, 8865-8873
8865

McManus et al.
substrate solvolysis, the linear solvation energy relation-
ship (LSER) is shown in eq 1.
SCHEME 1. Solvolysis Reaction Pathways for
2-Chloro-2-methyl-adamantane (CMA)
/
1
2
log k ) log k + sπ + dδ + aR + bâ + hδ
(1)
o
1
1
1
H
In solvation eq 1, k is the observed solvolytic rate
constant. The explanatory variables are the solvent
/
solubility parameters: π is the solvent dipolarity and
1
polarizability; δ
1
is a polarizability correction factor which
is only significant for halogenated and/or aryl solvents;
is the solvent hydrogen-bond acidity; â is the solvent
hydrogen-bond basicity; and δ is the Hilderbrand solu-
bility parameter. The term including δ models the
formation of the solute cavity in a solvent. Log k is the
R
1
1
H
H
o
constant resulting from the MLRA and represents the
log k value for a solvent with all solubility properties
equal to zero. Our interest in predicting reactivity in a
wide variety of solvent types of complex substrates such
as 2,2′-bischloroethyl sulfide, “mustard gas”, led us to
investigate MLRAs.15
the previous studies concluded that more reaction sys-
tems of varying types needed to be studied in order to
provide a better basis to evaluate the widespread ap-
plicability of the KAT solvation equation. Because they
have been so widely used in medium effects studies, the
application of the KAT solvation equation to the solvoly-
sis of 1-adamantyl chloride and bromide would be inter-
esting. Unfortunately, 1-adamantyl halides only react to
a slight extent, sometimes an immeasurably small extent,
in many of the solvents required to apply solvation eq 1
and maintain statistically viable results. A similar
reactant substrate that does react in a broad range of
solvents is 2-chloro-2-methyladamantane (CMA). More-
over, since both CMA and tert-butyl chloride (TBC) are
tertiary aliphatic chlorides, mechanistic similarities would
be expected. This article details our application of sol-
vation eq 1 to the solvolysis of CMA in a wide variety of
solvents. The results are discussed in light of other
available data.
The KAT method has been shown to provide mecha-
nistic insight in a variety of reaction types. For example,
its application to the study of the decarboxylation of
benzisoxazole-3-carboxylate revealed insight regarding
16
the relevance of catalysis in an antibody binding site.
These linear solvation energy relationships have also
been used to study the isomerization of cis- to trans-
stilbene which occurs via a photoinduced excited state.16
The KAT solvation equation has been applied to only
a few solvolytic reactions, including the solvolysis of tert-
1
8
butyl halides and a mustard simulant, 2-phenylthio-
19
ethyl naphthalenesulfonate. Each study has provided
indications that the method has promise for correlating
reaction rates and modeling reaction pathways. However,
(
13) (a) Taft, R. W.; Abraham, M. H.; Doherty, R. M.; Kamlet, M. J.
J. Am. Chem. Soc. 1985, 107, 3105. (b) Abraham, M. H.; Grellier, P.
L.; Abboud, J.-L. M.; Doherty, R. M.; Taft, R. W. Can. J. Chem. 1988,
6
6, 2673. (c) Kamlet, M. J. et al., J. Org. Chem. 1983, 48, 2877. (d)
Kamlet, M. J.; Doherty, R. M.; Abboud, J.-L. M.; Abraham, M. H.; Taft,
R. W. Chem Tech 1986, 9, 566. (e) Abraham, M. H.; Taft, R. W.; Kamlet,
M. J. J. Org. Chem. 1981, 46, 3053. (f) Abraham, M. H.; Grellier, P.
L.; Nasehzadeh, A.; Walker, R. A. G. J. Chem. Soc., Perkin Trans. 2
Results and Discussion
Previous studies by Shiner20 and Bentley21 and their
co-workers suggested that CMA undergoes solvolysis in
a wide variety of solvent types by either solvolytic
substitution or by elimination, Scheme 1. To evaluate the
KAT LSER methodology and best characterize the sol-
volysis of CMA, several solvent classes were chosen in
order to provide a broad range of π , R
values. In Table 1 the solvents selected are listed along
with their solvation parameters and log k values from
rate measurements at 60 °C. The π value of the solvent
1
988, 1717.
14) (a) Lu, J.; Brown, J. S.; Liotta, C. L.; Eckert, C. A. Chem.
(
Commun. 2001, 665-666. (b) Lagalante, A. F.; Bruno, T. J. J. Phys.
Chem. B 1999, 103, 7319. (c) Lagalante, A. F.; Hall, R. L.; Bruno, T. J.
J. Phys. Chem. B 1998, 102, 660. (d) Lagalante, A. F.; Wood, C.; Clarke,
A. M.; Bruno, T. J. J. Solution Chem. 1998, 27, 887. Abraham, M. H.;
Poole, C. F.; Poole, S. K. J. Chromatogr. A 1999, 842, 79-114. (e)
Abraham, M. H.; Le, J. J. Pharm. Sci. 1999, 88, 868-880. (f) Abraham,
M. H.; Chadha, H. S.; Martins, F.; Mitchell, R. C.; Bradbury, M. W.;
Gratton, J. A. Pest. Sci. 1999, 55, 78-88. (g) Platts, J. A.; Abraham,
M. H.; Hersey, A.; Butina, D. J. Chem. Inf. Comput. Sci. 2000, 40,
/
2
, â
1
, and δH
1
1
/
7
1-80.
15) (a) McManus, S. P.; Neamati-Mazraeh, N.; Hovanes, B. A.;
1
(
greatly affects the rate for any reaction involving the
formation of a dipolar transition state. Based on earlier
experiences which indicate CMA fits this reaction cat-
Paley, S. M.; Harris, J. M. J. Am. Chem. Soc. 1985, 107, 3393. (b)
Harris, J. M.; Sedaghat-Herati, M.; Neamati-Mazraeh, N.; Kamlet, M.
J.; Taft, R. W.; Doherty, R. M.; Abraham, M. H. In Nucleophilicity;
Harris, J. M., McManus, S. P., Eds.; Advances in Chemistry Series
No. 215; American Chemical Society: Washington, DC, 1987; pp 247-
egory and knowing the rates of reaction of CMA in some
highly polar solvents,2
0,21
we anticipated difficulty mea-
2
54.
/
(
16) Grate, J. W.; McGill, R. A.; Hilvert, D. J. Am. Chem. Soc. 1993,
suring the reactivity of CMA in solvents with low π1
values.
Reactions of CMA in most of the pure solvents were
carried out in sealed conductivity cells with platinum
1
15, 8577-8584.
(17) McGill, R. A.; Rice, J. K.; Baronavski, A. P.; Owrutsky, J. C.;
Lowrey, A. H.; Stavrev, K. K.; Tamm, T.; Zerner, M. C. Int. J. Quantum
Chem. Quantum Chem. Symp. 1996, 30, 1595-1606.
(18) (a) Abraham, M. H.; Doherty, R. M.; Kamlet, M. J.; Harris, J.
M.; Taft, R. W. J. Chem. Soc., Perkin Trans. 2 1987, 913. (b) Abraham,
M. H.; Doherty, R. M.; Kamlet, M. J.; Harris, J. M.; Taft, R. W. J.
Chem. Soc., Perkin Trans. 2 1987, 1097.
(20) Shiner, V. J., Jr.; Fisher, R. D.; Seib, R. C.; Szele, I.; Tomic M.;
Sunko, D. E. J. Chem. Soc. 1975, 97, 2408.
(21) Bentley, T. W.; Bowen, C. T.; Parker, W.; Watt, C. I. F. J. Chem.
Soc., Perkin Trans. 2 1980, 1244.
(
19) Harris, J. M.; Sedaghat-Herati, M. R.; McManus, S. P.; Abra-
ham, M. H. J. Phys. Org. Chem. 1988, 1, 359.
8866 J. Org. Chem., Vol. 69, No. 25, 2004

Solvolysis Model for 2-Chloro-2-methyladamantane
TABLE 1. Solvation Parameters and CMA Observed and Calculated Rate Constants at 60 °C
/
2
H
no.
solvents
δ1
π
R1
â1
δ
log kobs
log kcalc
1
1
2
3
4
5
6
7
8
9
watera
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.50
0.50
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.09
0.60
0.54
0.52
0.48
0.47
0.92
0.73
0.65
0.85
0.88
0.75
0.71
1.00
0.64
0.00
1.17
0.93
0.83
0.78
0.76
0.79
0.90
1.51
1.96
0.22
0.00
0.19
0.08
0.00
1.12
0.00
0.18
0.62
0.77
0.83
0.95
0.88
0.52
0.00
0.00
0.25
0.69
0.37
0.48
0.76
0.45
0.00
5.490
2.052
1.621
1.432
1.331
1.295
2.740
1.371
0.893
1.585
1.389
1.378
0.906
1.688
2.035
0.980
-0.247
-3.01
-3.81
-3.85
-3.98
-3.77
-2.31
-0.195
1.18
-0.388
-2.88
-3.54
-3.81
-4.17
-3.99
-2.06
-0.007
0.943
-3.43
-4.71
-3.94
-4.51
-4.57
-2.00
-5.41
methanol
ethanol
1-propanol
2-propanol
1-butanol
ethylene glycol
b
TFE
c,d
HFIP
1
1
1
1
1
1
1
0
1
2
3
4
5
6
nitromethane
dimethylformamide
acetonitrile
-3.30
-4.75
-4.04
-4.46
-4.46
-2.07
acetone
dimethyl sulfoxide
acetic acid
e
cyclohexane
ND
a
b
Extrapolated from rates at other temperatures using rates extrapolated to pure water rates; this work. 2,2,2-Trifluoroethanol.
c
d
Hexafluoro-2-propanol. Estimated by extrapolation of rates of reaction of CMA in 97% (w/w) aqueous HFIP measured at other
temperatures; ref 21. ND ) not determined.
e
electrodes to determine the pseudo-first-order reaction
reaction was found to be extremely slow even at temper-
atures up to 100 °C. After failing to obtain a satisfactory
pseudo-first-order rate constant, the solvent was dropped
from the experimental solvent set. Also, since acetolysis
kinetics of CMA could not be satisfactorily followed
conductometrically, the GC method was utilized to
provide rates in the solvent acetic acid. Sodium acetate
was added to buffer the acetic acid solutions although
other mechanistic roles for acetate cannot be ruled out.
2
2
rate constants. Reaction temperatures were held con-
stant with a thermostated water bath. To provide readily
measurable rates, most of the solvolytic experiments were
conducted at a temperature in the range of 50-100 °C.
For each solvent at each temperature, two or more rate
determinations were made. If the difference between the
two experimentally measured reaction rates was greater
than 10% of the lower value, additional solvolytic experi-
ments were made to improve the statistical confidence
in the measured values. Correlation coefficients for all
individual rate determinations were 0.995 or higher.
Activation parameters were computed after obtaining
average reaction rates at two or more temperatures.
Measured rate constants and activation parameters are
presented in Table 2.
The majority of kinetic experiments were carried out
at temperatures in excess of 50 °C. However, for highly
polar and hydrogen-bond acidic solvents such as ethylene
glycol and trifluoroethanol (TFE), reactions were con-
ducted at lower temperatures due to the high reactivity
of CMA in these solvents. To improve the validity of rate
comparisons, rate constants were determined at two or
more temperatures near 60 °C and rate constants for
solvolysis at 60 °C were calculated using the Arrhenius
equation. In all solvolytic runs, 2,6-lutidine, as a molar
equivalent to CMA, was added to scavenge the acid
produced in the reaction. Without lutidine we observed
considerably more scatter of the data points, which could
be related to an acid-catalyzed side reaction.
The rates reported for some solvents were determined
from extrapolated binary solution data. The solvolysis
rate in pure hexafluoro-2-propanol (HFIP) at 60 °C was
estimated by extrapolation using rates measured for
21
solvolysis of CMA in 97% (w/w) aqueous HFIP. The rate
of solvolysis of CMA in pure water was obtained by
extrapolation of the rates obtained for solvolysis of CMA
in a range of binary aqueous solvents. Using the Grun-
wald-Winstein method, estimated conductometric rates
at 25 °C and 40 °C in water were obtained by extrapolat-
ing to pure water a plot of log k values versus solvent Y
values3 constructed from conductometric rate constants
for solvolysis of CMA in binary mixtures of 40%, 50%,
60%, 70%, and 80% aqueous ethanol (v/v), Table 3. To
improve confidence levels for the extrapolated value for
water, a separate series of kinetic experiments in 50%,
60%, 70%, and 80% aqueous ethanol were carried out at
,4
2
5 °C by monitoring the spectrophotometric absorption
of methyl orange at 525 nm. The solutions were sampled
at timed intervals and the data stored automatically by
computer. The plot of log k versus Y values provided an
extrapolated value for the solvolysis rate of CMA in
water within 2% of the value obtained conductometri-
cally. Similarly, the rates of solvolysis for CMA in 40%,
Solvolysis of CMA in dimethylformamide (DMF) was
also carried out utilizing gas chromatography (GC) to
monitor the progress of the reaction. The reaction mix-
ture, contained in a sealed vial, was sampled via a serum
cap at intervals and analyzed by GC. The reaction rate
was calculated by monitoring the consumption of CMA
with respect to an internal standard, diphenyl ether. The
reaction rates obtained with the GC method were in good
agreement with the reaction rates obtained using the
conductivity method. The GC method was also used for
the solvent cyclohexane. Unfortunately, the cyclohexane
6
0%, and 80% aqueous acetonitrile at 25 °C, Table 3,
were determined by the spectrophotometric method and
plotted to determine an extrapolated rate for CMA in
pure water at 25 °C. The extrapolated rates of solvolysis
of CMA in pure water at 25 °C determined separately
from aqueous ethanol and aqueous acetonitrile data
showed good agreement. Using the extrapolated rates in
pure water from the two temperatures, the rate for
solvolysis of CMA in pure water at 60 °C, which was
(
22) Hovanes, B. A.; Harris, J. M.; McManus, S. P. Am. Lab. 1984,
1
6, No. 6, 22.
J. Org. Chem, Vol. 69, No. 25, 2004 8867

McManus et al.
TABLE 2. Conductometric Rate Data for Solvolysis of 2-Chloro-2-methyladamantane in Pure Solvents
q
q
solvent
methanol
T (°C)
k (s-1)
∆H (kcal/mol)
∆S (eu)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
4
4
3
5
5
5
5
4
5
4
4
5
5
4
4
4
4
3
3
5
4
4
50.0
(3.85 ( 0.04) × 10
(9.70 ( 0.17) × 10
(1.47 ( 0.04) × 10
(5.33 ( 0.54) × 10
(9.55 ( 0.04) × 10
(2.86 ( 0.06) × 10
(3.91 ( 0.07) × 10
(8.25 ( 0.06) × 10
(6.19 ( 0.59) × 10
(2.09 ( 0.12) × 10
(4.47 ( 0.23) × 10
(3.95 ( 0.10) × 10
(8.90 ( 0.01) × 10
(4.58 ( 0.09) × 10
(2.52 ( 0.03) × 10
(5.57 ( 0.17) × 10
(4.93 ( 0.02) × 10
(1.04 ( 0.01) × 10
(1.32 ( 0.32) × 10
(4.10 ( 0.06) × 10
(2.46 ( 0.01) × 10
(4.36 ( 0.08) × 10
(1.68 ( 0.01) × 10
(3.66 ( 0.13) × 10
(9.73 ( 0.05) × 10
(1.38 ( 0.05) × 10
(2.26 ( 0.06) × 10
(4.13 ( 0.05) × 10
(1.00 ( 0.01) × 10
(1.48 ( 0.01) × 10
(7.57 ( 0.24) × 10
(5.65 ( 0.01) × 10
(1.97 ( 0.01) × 10
(4.79 ( 0.09) × 10
(1.43 ( 0.09) × 10
(6.07 ( 0.05) × 10
6
0.0
5.0
18.8
-16.2
-7.33
6
ethanol
50.0
5
6
7
7
5.1
5.0
0.0
5.0
22.9
1-propanol
2-propanol
1-butanol
50.0
6
5.0
5.0
17.0
22.0
-25.3
-11.0
7
49.9
59.8
4.3
7
65.1
7
7
8
9
5.0
5.1
5.0
0.0
15.8
18.9
17.0
-28.6
-17.1
-29.8
nitromethane
34.8
50.1
0.0
6
dimethylformamidea
59.9
b
b
-5
-5
6
7
8
9
9.8
9.4
9.3
5.3
-
-
-
-
-
-
-
-
-
-
-
-
3
5
4
4
4
4
4
4
5
4
4
4
4
1
04.4
56.6
9.9
9.9
acetonitrile
DMSO
6
8
15.8
20.2
18.6
14.9
16.1
-29.8
-18.7
-13.6
-14.8
-30.7
65.0
7
8
9.9
9.9
ethylene glycol
25.1
3
4
4.9
5.1
-
(45.1 ( 0.20) × 10
-
-
-
-
-
-
-
3
3
3
4
4
4
4
2,2,2-trifluoroethanol
25.1
(3.94 ( 0.47) × 10
(6.07 ( 0.28) × 10
(1.09 ( 0.20) × 10
(1.23 ( 0.24) × 10
(1.89 ( 0.12) × 10
(4.38 ( 0.27) × 10
(6.24 ( 0.07) × 10
(4.73 ( 0.09) × 10
3
4
4.9
5.1
acetone
80.1
8
9
5.1
0.0
05.3
1
waterc
25.0
-2
-
1
4
0.0
1.12 × 10
16.2
5.44
-11.3
-51.9
acetic acidd
40.0
0.0
(4.73 ( 0.09) × 10
-3
-
3
6
(8.51 ( 0.47) × 10
a
Combination of rates determined by conductivity and by gas chromatography. b Rates determined using conductivity measurements.
Rates obtained by extrapolation to pure water using the Grunwald-Winstein method of rates of CMA in aqueous ethanol and aqueous
c
acetonitrile (25 °C only). d Rates determined using gas chromatography.
TABLE 3. Rates of Solvolysis of CMA in Binary
Aqueous Solvents
Products of Reaction of CMA in Pure Solvents.
20
Shiner et al. determined the products of solvolysis of
4
-1
solvent
T (°C)
10 k, s
CMA in 50%, 70%, and 80% aqueous ethanol finding 7%,
5%, and 33% elimination, respectively, to 2-methylene-
2
1
8
7
6
5
4
1
6
4
3
1
8
6
5
00% EtOH
25
25
25
25
25
25
25
25
25
25
40
40
40
40
0.0245 ( 0.002
1.45 ( 0.16
3.63 ( 0.05
12.0 ( 0.7
0% aq EtOH (v/v)
0% aq EtOH (v/v)
0% aq EtOH (v/v)
0% aq EtOH (v/v)
0% aq EtOH (v/v)
00% MeCN
adamantane (MA). 2-Hydroxy-2-methyladamantane
(HMA) and 2-ethoxy-2-methyladamantane (EMA) were
also products with their amounts dependent on the
solvent composition. As expected, the amount of HMA
increased with increasing water content of the solvent.
Since CMA is sterically incapable of undergoing backside
solvolytic displacement, the substitution product must
form by frontside displacement, probably involving solvent-
29.5 ( 2.8
51.0 ( 3.8
0.0309 ( 0.002
8.71 ( 0.21
32.4 ( 2.4
0% aq MeCN (v/v)
0% aq MeCN (v/v)
0% aq MeCN (v/v)
00% EtOH
0% aq EtOH (v/v)
0% aq EtOH (v/v)
0% aq EtOH (v/v)
47.9 ( 2.9
0.141 ( 0.012
2.45 ( 0.20
79.3 ( 1.6
23
separated ion pairs. We determined the products of
reaction of CMA in some pure solvents by GC-MS
analysis. As one might predict, in polar aprotic solvents,
the product of elimination (MA) is the sole product
113 ( 22
needed for the MLRA, was obtained by extrapolation.
These estimates are shown in Table 2.
(
23) For example, see Adcock, W.; Trout, N. A.; Vercoe, D.; Taylor,
D. K.; Shiner, V. J., Jr.; Sorensen, T. S. J. Org. Chem. 2003, 68, 5399.
8868 J. Org. Chem., Vol. 69, No. 25, 2004

Solvolysis Model for 2-Chloro-2-methyladamantane
TABLE 4. Regression Results for Reaction of CMA at 60 °Ca
/
2
step
1
regression information
log k0
π
R1
â1
δ1
δ
1
H
15, all parameters
T-test
-5.374
0.9999
2.084
0.9772
2.409
0.9999
0.996
0.209
2.487
1.0000
0.996
0.199
2.505
1.0000
0.996
0.191
-1.753
0.9995
(0.408)
0.2972
(0.039)
0.2057
overall corr coefficient
std deviation
2
2
3
15, exclude h (dH )1
-5.466
2.257
0.9998
-1.761
0.9998
(0.174)
0.2492
T-test
1.0000
overall corr coefficient
std deviation
2
H
15, exclude h, δ , dδ
1
-5.409
1.0000
2.219
0.9999
-1.823
0.9999
T-test
overall corr coefficient
std deviation
a
Values in parentheses are not statistically significant at 95% confidence levels.
TABLE 5. Regression Results Obtained by Sequentially and Randomly Deleting One Solvent Set at a Timea
/
step
3
solvent set size
15
log ko
π
R1
â1
overall corr coeff
0.996
SD in calcd log k values
1
-5.409
1.0000)
2.219
(1.0000)
2.265
(1.0000)
2.200
(0.9999)
2.098
(0.9996)
1.913
(0.9952)
1.921
(0.9891)
1.941
(0.9986)
2.505
(0.9999)
2.485
(0.9999)
2.495
(1.0000)
2.438
(1.0000)
2.393
(1.0000)
2.392
(0.9999)
2.441
(1.0000)
-1.823
(1.0000)
-1.913
(1.0000)
-2.043
(1.0000)
-2.157
(0.9999)
-2.181
(0.9998)
-2.187
(0.9994)
-1.884
(1.0000)
0.191
(
4
14
13
12
11
10
14
-5.388
0.9999)
-5.273
1.0000)
-5.083
1.0000)
-4.195
0.9999)
-4.919
0.9998)
-5.416
1.0000)
0.996
0.996
0.996
0.996
0.996
0.995
0.187
0.180
0.182
0.189
0.204
0.190
(
5
(
6
(
7
(
8
(
9b
(
a
Values in parentheses indicate the calculated probability of nonrandom occurrence (T-test). b Obtained by excluding water from the
regression analysis.
2
observed. Although we were unable to obtain a good rate
for the solvolysis of CMA in cyclohexane, we determined
the sole observable product of its reaction to be MA.
Reaction products were not determined for all of the
solvents studied because the available data provided good
information about the nature of the products and, to some
extent, how they may have formed. Furthermore, there
is no compelling evidence to indicate that the product-
forming step in the reaction sequence influences the
kinetics of the reaction and, hence, the application of
solvation eq 1.
and δ are not statistically significant so it is reason-
H
able to eliminate these variables from the solvation
equation and repeat the MLRA. However, we do so by
deleting the parameters sequentially and repeating the
regression analysis. After the elimination of the term for
2
H
δ
the rerun MLRA, shown in step 2, Table 4, is little
changed from the result of step 1 but the overall standard
deviation is slightly improved. Again the term for δ is
statistically insignificant. The MLRA is again repeated
after deleting δ, in step 3, Table 4. After this iteration
there is little change in the magnitude of the coefficients
but the regression quality was further improved. This
results in eq 2:
Application of Solvation eq 1 to CMA. The KAT
solvation equation was used to correlate the solvolysis
rates of CMA at 60 °C in pure solvents. MLRA was used
to analyze the log K values in terms of their dependence
on the solvent solubility properties. The resulting regres-
sion coefficients were considered statistically important
if application of the student’s T-test produced a value
greater than 0.99.1 In the present work, we found that
coefficients either give T-tests of better than 0.99 or less
than 0.70, indicating confidence levels of >99% and
/
1
log k ) -5.409 + 2.219π + 2.505R - 1.823â
1
1
(
2)
This result is the most suitable solvation equation for
modeling the solvolysis rates of CMS in all the solvents
studied. The regression correlation coefficient is 0.996
and the standard deviation for calculated log k values is
0.19. Considering the magnitude of log k values, which
include rate constants over a range of about 6 orders of
magnitude, and the inherent errors associated with the
experimental measurements, and the extrapolated pro-
cessing of some data, the statistical quality of solvation
eq 2 is excellent. The solvation rate eq 2 indicates that
dipolar and hydrogen-bond acidic solvents significantly
accelerate the solvolysis reaction of CMA whereas hy-
drogen-bond basic solvents retard the reaction. No sig-
nificant dependence on a cavity effect is observed.
3,18
<
70%, respectively. Therefore, decisions regarding sta-
tistical significance are not controversial. We find that
only three solvent parameters are statistically significant
at the 99% confidence level for the CMA solvolyses.
Results of the multiparametric analysis are summarized
in Tables 4 and 5.
We completed our analysis of the CMA data as follows.
The first attempted MLRA, step 1, Table 4, was obtained
using all 15 solvents and by including all the solvent
parameters from eq 1. Notice that the coefficients for δ
1
J. Org. Chem, Vol. 69, No. 25, 2004 8869

McManus et al.
To evaluate the size and quality of the solvent set, the
MLRA for solvation eq 1 was repetitively carried out by
randomly deleting one solvent at a time. This process was
repeated until the solvent set was reduced to 10 solvents.
The resulting regression equations (steps 4-8) can be
written from the data in Table 5 and indicate that there
is very little change in the regression coefficients or the
quality of the regression as indicated by the student
T-test results, the overall correlation coefficient, and
overall standard deviation. As a final test of the MLRA
the data for water was excluded. Water has the largest
/
2
H
π value and by far the largest δ value and as a result
1
could dominate a regression analysis. The resulting
solvation equation (from data in Step 9) is again little
changed from eq 2, which utilizes the full set of solvents
available.
Another important factor is the choice of solvent set
in the experimental design. We sought to use a wide
range of solvents and solvent types to provide as wide a
range in solvent-solute interactions as possible. Indeed,
for statistically meaningful results, the cross correlation
between the different variables, i.e., the solvent solubility
parameters, must not be significant. For the full solvent
set, none of the parameters are correlated with coef-
ficients higher than 0.48. In addition, we sought to
include solvents that would provide experimentally mea-
sured variables, k values, which maximized the range
used in the MLRA. By the careful adoption of some rates
extrapolated from those measured at other temperatures,
experimental rate constants covering approximately 6
orders of magnitude were included. Ultimately, it appears
that the experimental design, including the choice of
solvent set, is sound and the statistical quality of the
regression is excellent. Thus, a plot of experimentally
determined log k and calculated log k values, using eq 2,
Figure 1, illustrates the quality of the regression. As a
result of the extended data set, the resulting solvation
equation has provided chemically reasonable conclusions.
The resulting solvation eq 2 can now be used to
calculate reaction rates for solvents that were not used
and those for which it was difficult to get experimental
data. The calculated rate for the solvolysis of CMA in
cyclohexane at 60 °C gives a log k value of -5.409. This
indicates that the reaction rate in this solvent is about
an order of magnitude slower than the slowest reaction
rate for which useful experimental data was obtained.
It is perhaps, in hindsight, not surprising that during
our experimental efforts to determine rate data in
cyclohexane, difficulties were found in collecting repro-
ducible or reliable data.
FIGURE 1. Plot of log k experimental rate constants versus
log k calculated rate constants for the reaction of 2-chloro-2-
methyladamantane (CMA) in aprotic and protic solvents at
60 °C; r ) 0.996.
using only three parameters. The cavity term is not
statistically significant with CMA while its significance
2
is statistically justified for TBC. The term with δH
models any change attributed to the solvent cavity during
the rate determining step of the reaction. In the solvolytic
reactivity of aliphatic chlorides, the rate determining step
is very likely to correspond to ionization to form the
carbocation/chloride ion pair. The implication for a lack
of dependence in the cavity term for the solvolysis of
CMA is that there is no significant change in spatial size
of the molecule during the rate-determining step of the
reaction. This is in direct contrast to results obtained for
TBC. Of course, upon ionization, TBC can lead to a
flattened (planar) carbocation, something that is impos-
sible with the rigid adamantane structure. Abraham et
24
al. have indicated that for TBC the minor accelerating
effects of solvent cohesive energy density can be explained
by the electrostriction of surrounding solvent molecules
by the dipolar transition state. For CMA we could infer
that the transition state is less dipolar reducing such
effects to an inconsequential amount. Such might be the
case if the CMA solvolysis transition state involves
formation of a tight or intimate ion pair of the carbocation
and chloride ion this could account for the cavity terms
results. However, we will consider the possibility that the
medium effects in these reactions may be temperature-
dependent making a comparison of two substrates stud-
ied at different temperatures risky.
Mechanistic Significance of the Correlation. To
place in perspective the mechanistic significance of our
multiparameter correlation of CMA, it is useful to
consider some other examples using solvation eq 1 for
solvolysis studies. In fact, few examples exist, but fortu-
nately one of them is a comprehensive solvolytic study
of TBC at 25 °C in 21 pure solvents.1 The correlation
for TBC is
8b
In general, when considering that the reaction rates
for CMA were determined at 60 °C and those for TBC
were determined at 25 °C it is not surprising that the
rates are generally higher for CMA as indicated by the
/
1
2
H
log k ) -14.6 + 5.10π + 4.17R + 0.73â + 0.48δ
1
1
(3)
o
higher log k value. It is also not surprising that the
Significantly, the solvolysis of TBC requires four
parameters, whereas the solvolysis of CMA is correlated
(24) Harris, J. M. Prog. Phys. Org Chem. Vol. 1974, 11, 89-173.
8870 J. Org. Chem., Vol. 69, No. 25, 2004

Solvolysis Model for 2-Chloro-2-methyladamantane
SCHEME 2. Representation of Elimination from
the Tight Ion Pair
solvents and thus the transition state species is, by this
determination, highly dipolar. Assuming that the large
charge separation of the transition state in TBC is
correct, our data suggests significantly less charge de-
velopment in the transition state for the solvolysis of
CMA. In physical terms, one could suggest that, in the
transition state, CMA appears to produce tight or
intimate ion pairs while TBC gives either more highly
disassociated (loose) ion pairs or solvent-separated ion
pairs. As we discuss below, there is other data that agrees
with this interpretation.
Shiner and co-workers20 have studied CMA and its R
and â (CD ) deuterated derivatives in a few binary
3
solvent mixtures including 50%, 70%, and 80% aqueous
ethanol. They added the relevant data for CMA and its
relevant analogues into a previously linear correlation
found between the log k the rate effect of R-methyl
substitution for R-H in carbocationic reaction processes
and the logarithms of the rate effect of R-methyl deu-
FIGURE 2. Plot of log kCMA (60 °C) versus log kTBC (25 °C);17
r ) 0.924.
3
teration. The â-d isotope effect found in 50% aqueous
dependence on the dipolarity and hydrogen-bond acidity
is less because at elevated temperatures intermolecular
forces are considerably reduced. However, since both the
TBC and CMA correlations were highly correlated even
though many rates were extrapolated from rates deter-
mined at temperatures other than 25 and 60 °C, respec-
tively, there is no compelling evidence that a significant
mechanistic change occurred with either substrate over
this modest temperature range. A referee suggested
extrapolation of the CMA data to 25 °C to allow a better
comparison with TBC. While we appreciate the desire
to do this, we believe that this approach would risk
extrapolation errors that would potentially weaken the
correlations and thus hinder the comparison.
The most significant difference between the two sol-
vation equations, (2) and (3), is the magnitude and sign
of the coefficient for the solvent hydrogen-bond basicity.
This is strong evidence that the reaction pathway is
markedly different for CMA and TBC. That the two alkyl
chlorides indeed behave differently is simply demon-
strated by the correlation of the rates of one substrate
against those of the other. The least squares treatment
of the correlation of log kTBC versus log kCMA, shown
graphically in Figure 2, is relatively poor, indicative of
mechanistic differences.
ethanol is close to that predicted by the correlation line
while the values for 70% and 80% aqueous ethanol fall
off the line. Further, in 50% ethanol, CMA yielded only
7
% elimination products, while in 70% and 80% aqueous
ethanol, increasing fractions of elimination were found
to be associated with increasing â-d rate effects. Shiner
3
interpreted these data as indicating that at least some
fraction of the reactant substrate reacts with rate-
determining elimination from a tight ion pair with the
chloride ion acting as the base, as shown in Scheme 2.
In their aqueous ethanol solutions the fraction proceeding
by this pathway is suggested to be up to 33% of the
reaction. However, their data and conclusion explicitly
requires that most of the reaction is not proceeding by
rate-limiting elimination.20
Our results are consistent with Shiner’s conclusion that
a tight ion pair is involved. It is significant to recall that
our results are inconsistent with the involvement of
solvent in a rate-limiting elimination pathway. If rate-
limiting elimination were occurring, one would anticipate
that solvents with enhanced basic or nucleophilic char-
acter (i.e., those with higher â values) would speed the
reaction. In this scenario, a positive value for â would be
expected. However, in solvation eq 2 for CMA the large
The relatively small a and s coefficients for solvolysis
of CMA (a ) 2.51 and s ) 2.22), when compared with
the corresponding coefficients for the solvolysis of TBC
measured previously (a ) 4.17 and s ) 5.10), are probably
not fully explained by temperature differences for the
reaction data. It has been previously argued that the
magnitude of the a and s coefficients for TBC suggests
that there is significant charge separation in the transi-
1
negative coefficient for â (-1.823) indicates that the rate
of reaction of CMA decreases with increasing solvent
hydrogen-bond basicity. This indicates that the solvent
is not involved in a rate-limiting elimination pathway.
As pointed out above, however, Shiner’s conclusion was
that chloride ion, and not the solvent, was acting as the
base. Since the AKT correlation is focused on the various
roles of the solvent, it is probable that the correlation is
unable to directly elucidate the role of the chloride ion
in the mechanism of the reaction.
1
8b
tion state. The charge separation has been approxi-
mated at about 0.80-0.84 (1.0 ) total separation) in polar
J. Org. Chem, Vol. 69, No. 25, 2004 8871

McManus et al.
SCHEME 3. Mechanistic Scheme for the Reaction
of CMA
rate kinetic data obtained for the reaction of CMA in a
wide range of solvents. With a set of fifteen solvents
including protic and aprotic solvents, the overall correla-
tion coefficient was excellent. The solvation equation for
the solvolysis of CMA at 60 °C allows further values to
be calculated for a wide variety of solvents not studied,
but for which the relevant solvent parameters are avail-
able.
The KAT solvation equation provides chemical insight
into the behavior of the reactant substrate, although care
has to be taken in interpreting the individual regression
terms. There is an obvious advantage in comparing
similar substrates such as TBC and CMA, which show
significant differences in their behavior from similar
solvent sets. Adding additional reaction types, such as
2-adamantyl derivatives, which are thought to involve
Bentley and co-workers21 have studied solvolysis of
CMA in 97% aq hexafluoro-2-propanol (HFIP) at various
temperatures and reach conclusions totally in agreement
with those drawn from our correlations. They too suggest
that reaction in HFIP proceeds via the rate-determining
formation of tight (contact) ion pairs. If the mechanism
for solvolysis of CMA involves the formation of tight ion
pairs, one would expect that solvents with stronger
hydrogen-bond basicity would slow the reaction because
the solvent could compete with the chloride ion for
stabilization of the carbocation. Of course, one could
argue that CMA could minimize the need for stabilization
of the chloride ion by crossing over to a mechanism, i.e.,
rate-limiting elimination, where the chloride ion attacks
a â-H as ionization occurs. Although solvent hydrogen-
bond basicity is not indicated in the rate-determining
step, the more basic solvents, by their competition with
the chloride ion, may promote chloride ion-induced
elimination.
Based on our data and previous studies, the reaction
of CMA seems to follow the mechanistic scheme dia-
grammed in Scheme 3. The rate-determining step for the
reaction of CMA is the first or ionization step, which
leads to the formation of tight ion pairs. After the
transition state is reached, other intermediates, such as
solvent-separated ion pairs and free ions, may form if the
particular solvent stabilizes the separated ions.
2
3
rate-limiting formation of solvent-separated ion pairs,
would add to the confidence of using this method mecha-
nistically.
From our analysis of the solvolysis of CMA in fifteen
solvents having wide ranging properties, we conclude
that the solvolytic rate-determining step is the ionization
step. Since there appears to be a relatively moderate
amount of charge separation in the transition state, we
can conclude that the principal intermediate formed is a
tight or intimate ion pair. In the more polar solvents, and
especially in the alcoholic solvents which are known to
stabilize solvent-separated ion pairs,24 the insertion of
solvent molecules, after the transition state is reached,
is anticipated. These results demonstrate why the uses
of simple solvolytic correlation methods are invalid even
for comparisons of tertiary chlorides. Also, these results
suggest a greater reliability in the use of multiple linear
regression analysis for predicting solvolytic rates in
additional solvents.
Experimental Section
Substitution may only occur from the solvent-separated
ion pairs or from free ion pairs, if they form. If SOH in
Scheme 3 is a binary aqueous hydroxylic solvent, a
mixture of products may form, as shown. Elimination
may occur from either intermediate species with it likely
occurring from the tight ion pair when solvents of low
polarity are used. In nonpolar solvents like cyclohexane,
very little charge development is likely and the elimina-
tion reaction may be concerted or nearly so. This sug-
gestion is not inconsistent with our conclusion that the
rate-detemining step for the solvolytic reactions studied
is from the formation of a tight ion pair since the extent
of carbon-chlorine bond breaking is expected and shown
to be dependent on the solvent.
Since all of the solvent types chosen provide rate
constants that fit solvation eq 2, we can assume that the
transition-state formation is the same process being
modeled for all the solvents studied. However, note that
the hydroxylic solvents studied, but not the non-hydroxy-
lic solvents, give both substitution products and the
elimination product. Therefore, the products clearly occur
from different intermediates and by different mecha-
nisms. The substitution products must occur in a reaction
step after the transition step since solvent insertion
between the ions must occur.
Reagents. All reagents and reaction solvents were pur-
chased from standard sources. CMA was prepared from
20
2-methyl-2-adamantanol in 82% yield. Recrystallization from
2
0
hexane yielded pure CMA, mp 174-175 °C (lit. 176-176.5
C).
Solvolytic Reaction Solvents. The solvents, which were
°
unavailable in their highest purity form, were carefully
25,26
purified by distillation.
All solvents were desiccated under
dry nitrogen prior to use.
1
Instrumentation. H NMR spectra were obtained with an
IBM 200 MHz, high-resolution NMR spectrometer. All the
conductometric kinetic data were measured using an auto-
mated system consisting basically of the following compo-
nents: a data acquisition unit, a microcomputer, and a
conductivity bridge. Spectrophotometric data was collected
using a UV-vis spectrophotometer fitted with a programmable
cell changer and the associated software and a personal
computer system for system control and data storage. A gas
chromatograph fitted with a 25 m × 0.2 mm OV-101 column
and equipped with either a flame ionization detector or a mass
selective detector was used for all gas chromatographic and
GC-MS measurements.
Kinetic Measurements. The kinetic data were obtained
by the various methods described below. All of the kinetic runs
were followed through at least one half-life and in most cases
(25) Vogel, A. I. Practical Organic Chemistry; Wiley: New York,
1
956.
(
26) Riddick, J. A.; Toops, Jr. Organic Solvents; Physical Properties
Conclusions. The KAT solvation equation developed
is a very powerful technique for correlating the reaction
and Methods of Purification, 2nd ed.; Weissberger, A., Ed.;Techniques
of Organic Chemistry, Vol. 7; Interscience Publishers: New York, 1955.
8872 J. Org. Chem., Vol. 69, No. 25, 2004

Solvolysis Model for 2-Chloro-2-methyladamantane
through more than 1.5 half-lives. The typical concentrations
of reactant substrate and 2,6-lutidine were 10-3 M. The
required amounts were weighed on a 4-figure electronic
balance and the appropriate solvent was added to prepare the
solutions in either 25 or 50 mL volumetric flasks. A conductiv-
80 ( 0.05 °C and 100 ( 0.05 °C. The tubes were removed at
timed intervals and quenched in an ice-bath. The contents of
these ampules were then analyzed by GC-MS. After analysis
the ampules were returned to the thermostated oil bath.
Kinetics of CMA in aqueous acetonitrile and aqueous
2
2
ity method, previously described in detail, was employed
unless otherwise stated. The cells were filled until the elec-
trodes were covered and then were sealed either by rubber
caps or glass. The cells were placed in an oil bath and the leads
were connected to an electrometer and the data collection
system. The conductance was measured at specific times by
means of a computer-actuated conductivity bridge. The con-
ductivity data were analyzed by entering them into a computer
program, LSKIN, to allow computation of the rate constant
and to obtain calculated infinity values and correlation coef-
ficients. The temperature of the bath was recorded for each
measurement and the average temperature was computed for
the run. Three or more cells were normally used to determine
the rate constant at a temperature for each solvent. Multiple
temperatures were used to allow computation of activation
parameters and for extrapolation of rates to a single temper-
ature for rate comparisons.
ethanol solutions were determined at 25 °C using an auto-
mated spectrophotometric kinetics system previously described
in detail.27 The reactions were followed by observing the
change in the methyl orange absorbance at 525 nm.
Product Analysis. Using a GC-MS with an injector
temperature of 130 °C and a programmed column temperature
(
70-120 °C at 10°/min), the products of some of the reactions
were determined using solvolytic reaction mixtures held at an
appropriate temperature to provide at least five reaction half-
lives. For the solvents, acetone, DMF, and cyclohexane only
2
-methyleneadamantane was determined as the solvolytic
product.
Acknowledgment. We gratefully acknowledge the
financial support provided by the National Science
Foundation (CHE-8719885) and the U.S. Army Re-
search Office (DAAL03-92-G-0393).
Since the conductometric method was found to be unaccept-
able for acetic acid solutions, a GC method was employed to
follow the acetolysis of CMA. Aliquots of approximately 1 mL,
ca. 10 M in reactant substrate, were sealed into ampules and
placed in an oil bath. In the two acetolysis experiments, two
different temperatures were chosen to monitor the reactions,
JO049798L
(
27) Hovanes, B. A.; Sedaghat-Herati, M. R.; Harris, J. M.; Mc-
Manus, S. P. Am. Lab. 1986, 17 (3), 52.
J. Org. Chem, Vol. 69, No. 25, 2004 8873