Gas-phase kinetics and activation parameters for thermal 1,5-hydrogen shifts interconverting monodeuterium-labeled 1,3-cyclohexadienes

Article abstract of DOI:10.1021/jo048441e

(Chemical Equation Presented). Thermal equilibrations among the three possible monodeuterium-labeled 1,3-cyclohexadienes have been followed in the gas phase at temperatures from 254 to 284°C. The temperature-dependent rate constants for the 1,5-shift of a

Full text of DOI:10.1021/jo048441e

SCHEME 1
Gas-Phase Kinetics and Activation
Parameters for Thermal 1,5-Hydrogen
Shifts Interconverting
Monodeuterium-Labeled
1,3-Cyclohexadienes
The 1,5-hydrogen shift reactions of monocyclic conju-
gated dienes such as 1,3-cyclohexadiene (1) give degener-
ate products. The reactions are detectable only with the
assistance of isotopically labeled variants. An experi-
mental activation enthalpy of ∆Hq ) 40 kcal/mol for the
1,5-hydrogen shift of 1,4-d2-1 was reported in 1975.8 In
1976 in a Ph.D. thesis providing experimental details
about this work, the activation enthalpy for the reaction
of 1,4-d2-1 to give 2,5-d2-1 was reported to be 39 kcal/
mol.9 No other kinetic work on 1,5-hydrogen shifts shown
by isotopically labeled 1,3-cyclohexadienes has appeared.
John E. Baldwin* and Bonnie R. Chapman
Department of Chemistry, Syracuse University,
Syracuse, New York 13244
jbaldwin@syr.edu
Received September 3, 2004
Thermal equilibrations among the three possible monodeu-
terium-labeled 1,3-cyclohexadienes have been followed in the
gas phase at temperatures from 254 to 284 °C. The temper-
ature-dependent rate constants for the 1,5-shift of a single
hydrogen lead to the activation parameters Ea ) (40.1 ( 0.8)
kcal/mol, log A ) (12.1 ( 0.3), and ∆Sq ) -(6.3 ( 1.3) e.u.
These activation parameters are reconciled with experimen-
tal values reported earlier for reactions starting with 1,4-
d2-cyclohexadiene.
Kinetic studies of 1,5-hydrogen shifts equilibrating
monodeuterium-labeled cis,cis-cyclooctadienes10,11 and
cis,cis-cycloheptadienes12,13 have provided activation pa-
rameters necessary for comparisons with theory-derived
values and for approaching better understandings of how
geometrical details in transition structures impact reac-
tion rates.14 The present work followed the equilibrations
among the three possible isotopomers of monodeuterium-
labeled 1,3-cyclohexadienes.
The three monodeuterium-labeled 1,3-cyclohexadienes
and the possible isomerization paths provided by 1,5-
hydrogen shifts are shown in Scheme 1. The rate constant
symbols in Scheme 1 reflect the path degeneracies
involved and neglect possible secondary deuterium ki-
netic isotope effects. Two hydrogens might shift to convert
2-d-1 into 1-d-1, and vice versa, or 1-d-1 into 5-d-1, so
on a per hydrogen shift basis the rate constants are
defined as 2kH. Isotopomer 5-d-1 can isomerize to 1-d-1
through the shift of but a single hydrogen, and hence the
appropriate rate constant is kH. Were a deuterium
migration to take place, 5-d-1 would give another version
of 5-d-1 and the reaction would go undetected. A kD shift
might well occur at a kinetically competitive rate, but it
would not complicate the kinetic situation at all.
Structural isomerizations through thermal 1,5-shifts
of hydrogens in conjugated dienes were encountered more
than 100 years ago, though it took some time for them
to be recognized individually and as representatives of a
specific type of reaction.1-4 Today they are well-known
as [1,5] sigmatropic rearrangements involving a suprafa-
cial transfer of a hydrogen from one end of a pentadienyl
π system to the other.5 The stereochemistry6 and very
substantial energy of concert characteristic of these
reactions7 mark them as concerted processes, transfor-
mations that take place without the intervention of any
short-lived reactive intermediate. Detailed studies of such
reactions thus offer an attractive opportunity for probing
into just how structure-reactivity relationships may
reflect the nature of transition structures of these
concerted reactions. Cyclic conjugated dienes seem well
suited for kinetic and computational studies, for the
structural variations they offer are related to geometrical
constraints rather than to major electronic effects intro-
duced by strongly electron-donating or -withdrawing
substituents.
A nonequilibrium mixture of monodeuterium-labeled
cyclohexadienes was conveniently prepared through the
(8) de Dobbelaere, J. R.; van Zeeventer, E. L.; de Haan, J. W.; Buck,
H. M. Theor. Chim. Acta 1975, 38, 241-244.
(9) de Dobbelaere, J. R. Thermal Sigmatropic [1,j] Shifts in Cyclic
Systems: A Quantumchemical Study. Ph.D. Dissertation, Eindhoven
University of Technology, The Netherlands, 1976.
(10) Glass, D. S.; Boikess, R. S.; Winstein, S. Tetrahedron Lett. 1966,
999-1008.
(1) Roth, W. R. Chimia 1966, 20, 229-236.
(2) Spangler, C. W. Chem. Rev. 1976, 76, 187-217.
(3) Hasselmann, D. Stereoselective Synthesis; Thieme: Stuttgart,
1996; Vol. E21, pp 4421-4463.
(4) For a recent extensive gathering of references on 1,5-hydrogen
shifts, see: Alabugin, I. V.; Manoharan, M.; Breiner, B.; Lewis, F. D.
J. Am. Chem. Soc. 2003, 125, 9329-9342.
(11) Baldwin, J. E.; Lee, T. W.; Leber, P. A. J. Org. Chem. 2001, 66,
5269-5271.
(12) Mironov, V. A.; Chizhov, O. S.; Kimel’feld, Ya. M.; Akhrem, A.
A. Tetrahedron Lett. 1969, 499-500.
(5) Woodward, R. B.; Hoffmann, R. J. Am. Chem. Soc. 1965, 87,
2511-2513.
(13) Baldwin, J. E.; Raghavan, A. S. J. Org. Chem. 2004, 69, 8128-
8130.
(6) Roth, W. R.; Ko¨nig, J.; Stein, K.Chem. Ber. 1970, 103, 426-439.
(7) Compare: Doering, W. von E.; Roth, W. R.; Breuckmann, R.;
Figge, L.; Lennartz, H. W.; Fessner, W. D.; Prinzbach, H. Chem. Ber.
1988, 121, 1-9.
(14) (a) Hess, B. A., Jr.; Baldwin, J. E. J. Org. Chem. 2002, 67,
6025-6033. (b) Hess, B. A., Jr. Int. J. Quantum Chem. 2002, 90, 1064-
1070.
10.1021/jo048441e CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/01/2004
J. Org. Chem. 2005, 70, 377-380
377

TABLE 1. Mole % Relative Concentration Data for
Monodeuterium-Labeled 1,3-Cyclohexadienes as
Functions of Reaction Temperature and Time
T (°C)
time (s)
2-d-1(t) (%)
1-d-1(t) (%)
5-d-1(t) (%)
0
4800
51.8
45.0
43.7
39.4
31.7
29.0
26.7
47.0
43.4
41.1
34.0
27.9
47.6
43.8
38.8
35.9
31.5
27.2
47.1
42.0
40.4
34.7
32.8
27.6
42.6
37.0
30.5
28.5
26.9
25.2
25.8
34.9
30.2
29.9
27.7
25.1
34.4
30.6
29.7
26.7
25.6
25.1
34.6
32.2
28.4
28.5
26.3
25.8
5.6
18.0
25.8
32.1
41.3
45.8
47.5
18.2
26.4
29.0
38.3
47.0
18.0
25.6
31.5
37.4
42.8
47.7
18.4
25.8
31.2
36.8
40.9
46.6
254.2
9600
14 000
28 800
39 600
50 400
2400
4800
7200
264.2
274.5
10 800
25 200
1200
2400
3600
5400
7200
12 600
600
284.2
1200
1800
2700
FIGURE 1. Deuterium NMR spectra for the starting mixture
d-1(0) (51.8% 2-d-1, 42.7% 1-d-1, and 5.5% 5-d-1, below) and
for a representative product mixture derived from gas-phase
thermolysis of d-1(0) at 274.5 °C for 7200 s (above), isolated
without GC purification.
3600
6300
the top of Figure 1. The relative concentrations of iso-
topomers 2-d-1, 1-d-1, and 5-d-1 as functions of temper-
ature and reaction time are summarized in Table 1.
The set of linear first-order differential equations
appropriate to the kinetic situation summarized in
Scheme 1 may be solved exactly to provide functions of
2-d-1(t), 1-d-1(t), and 5-d-1(t), which depend on a single
rate constant (k ) kH) for the 1,5-shift of a single
hydrogen and the initial relative concentrations 2-d-1(0),
1-d-1(0), and 5-d-1(0).17 For the experimentally deter-
mined set of initial concentrations relevant here, the
three functions may be calculated readily with the aid
of a linear algebra package18 (eqs 1-3).
SCHEME 2
short synthetic sequence outlined in Scheme 2. Reduction
of cyclohex-2-en-1-one (2) with NaBD4 in the presence of
CeCl3 provided 1-d-cyclohex-2-en-1-ol (3),15 which, when
dehydrated with CeCl3/NaI in acetonitrile,16 afforded a
mixture of monodeuterium-labeled cyclohexadienes, d-1(0),
shown by careful integrations of 2H NMR spectra at 92.1
MHz to be 51.8% 2-d-1 (δ 5.95), 42.4% 1-d-1 (δ 5.85), and
5.7% 5-d-1 (δ 2.17) (Figure 1). The major dehydration
products were those expected from 1,2- and 1,4-elimina-
tions.
2-d-1(t) ) 25.0 + 31.74 exp(-1.438kt) -
4.94 exp(-5.562kt) (1)
1-d-1(t) ) 25.0 + 8.90 exp(-1.438kt) +
8.80 exp(-5.562kt) (2)
The 2H NMR spectra in Figure 1 demonstrate that the
thermal 1,5-hydrogen shifts take place without ap-
preciable complications from other reactions. The absorp-
tions other than those associated with isotopomers 2-d-
1, 1-d-1, and 5-d-1 present in d-1(0) and in product
mixtures are readily assigned. The small signals at δ 5.74
and 2.02 are due to 1-d-1,4-cyclohexadiene and 3-d-1,4-
cyclohexadiene, respectively, minor side-products from
the dehydration of 3. The signal at δ 1.46 derives from
naturally abundant deuterium in cyclohexane, the bath
gas. The peaks at δ 7.42 ppm and 7.26 stem from
benzene-d1, a synthetic side-product, and naturally abun-
dant CDCl3 in the CHCl3 employed as the NMR solvent.
Gas-phase thermal isomerizations were run using a
static reactor; product mixtures from each run were
analyzed by 2H NMR spectroscopy, as in the example at
5-d-1(t) ) 50.0-40.64 exp(-1.438kt) -
3.86 exp(-5.562kt) (3)
The data points gathered in Table 1 may be fit with
theory-based values calculated using a least-squares best-
fit program and a single-variable parameter (k ) kH ).
The temperature-dependent values of k found through
these calculations for 5-d-1(t), the concentration showing
the largest dynamic range, are summarized in Table 2.
A representative plot of observed time-dependent mole
percent relative concentrations, Ci(obs), and the calcu-
lated functions is provided in Figure 2 for isomerizations
at 284.2 °C. The average absolute value of Ci(obs) - Ci-
(cal) for the 69 kinetic points in Table 1 having t > 0
was 0.7%. The estimated uncertainties in the rate
(15) Davis, M. W.; Crabtree, R. H. J. Org. Chem. 1986, 51, 2655-
2661.
(16) Bartoli, G.; Bellucci M. C.; Petrini, M.; Marcantoni, E.; Sambri,
L.; Torregiani, E. Org. Lett. 2000, 2, 1791-1793.
(17) Baldwin, J. E.; Leber, P. A.; Lee, T. W. J. Chem. Ed. 2001, 78,
1394-1399.
(18) LinearAlgebra. In Maple 9.03; Waterloo Maple, Inc.: Waterloo,
Canada.
378 J. Org. Chem., Vol. 70, No. 1, 2005

TABLE 2. Rate Constants for the 1,5-Shift of a Single
Hydrogen in 1,3-Cyclohexadiene
T (°C)
rate constant k
254.2
264.2
274.5
284.2
(3.94 ( 0.12) × 10-5 s-1
(7.54 ( 0.23) × 10-5 s-1
(1.56 ( 0.05) × 10-4 s-1
(3.00 ( 0.09) × 10-4 s-1
FIGURE 3. Arrhenius plots for 1,5-hydrogen shifts starting
with 1,4-d2-1 (upper line)9 and with the d-1(0) mixture of
monodeuterium-labeled 1,3-cyclohexadienes (51.8% 2-d-1, 42.4%
1-d-1, 5.7% 5-d-1; lower line, rate constant k ) kH).
pages of text, an Arrhenius plot, and a table of the
20
parameters obtained from that plot.9
The required sample of 1,4-d2-1 was prepared following
a route reported by Franzus.21 1,4-Cyclohexanedione was
reduced with LiAlH4; the diols obtained were treated with
TsCl to provide 1,4-d2-1,4-di(tosyloxy)cyclohexanes, and
the ditosylates were heated with NaOH in n-propyl al-
cohol to give 1,4-d2-1.21 Preparative GC provided 1,4-d2-1
free of 1,4-d2-cyclohexa-1,4-diene, a synthetic byproduct.9
The 1,4-d2-1 was isomerized 303 to 330 °C using a 9.15
mL micro flow reactor 186.5 cm in length.22,23 Residence
times varied from 5 to 60 s. For each temperature and
residence time, 10 injections of 2.5 µL gave product
mixtures that were collected together at -60 °C, dis-
FIGURE 2. Mol % concentrations of 2-d-1(t), 1-d-1(t), and 5-d-
1(t) as functions of time at 284.2 °C. The functions are given
in eqs 1-3, and k ) (3.00 ( 0.09) ×10-4 s-1
.
constants (taken to be (2σk) ranged from 3.0 to 3.2%.
These temperatures and rate constants provide a good
linear Arrhenius correlation of log(k) ) log(k) versus 1/T
(K) having slope and intercept of -8.754 × 103 and
+12.12, respectively. The activation parameters are
Ea ) (40.1 ( 0.8) kcal/mol and log A ) (12.1 ( 0.3), with
estimated uncertainties in Ea and log A calculated
according to O’Neal and Benson.19 A nonlinear least-
squares exponential fit of the Arrhenius equation gave
Ea ) 39.7 kcal/mol and log A ) 12.0, values well within
the estimated uncertainties. Calculation of ∆Sq using
A ) eT(k/T) exp(∆Sq/R) and taking T to be 542 K gave
∆Sq ) -(6.3 ( 1.3) cal K-1 mol-1 (e.u.). An Erying plot of
log (k/T) versus 1/T led to ∆Sq ) -6.4 e.u.
This experimental value for the activation energy of a
1,5-hydrogen shift in 1,3-cyclohexadiene, 40.1 kcal/mol,
is slightly less than Ea ) 41.2 kcal/mol, a value one may
calculate from ∆Hq ) 40 kcal/mol at 315 °C reported in
1975,8 since Ea ) ∆Hq + RT. It is identical with the Ea
value recorded in de Dobbelaere’s dissertation, 40 ( 1
kcal/mol.9
The experimental details for the kinetic study of 1,5-
hydrogen shifts shown by 1,4-d2-1 have never appeared
in a journal article, but they can be found in the Ph.D.
dissertation of de Dobbelaere,9 a thesis concerned pri-
marily with applications of semiempirical quantum-
chemical methods to define transition structures and
activation parameters for 1,5-hydrogen shifts in cyclo-
pentadiene, 1,3-cyclohexadiene, and 1,3,5-cyclohep-
tatriene.8,20 The experimental work expended in synthe-
sizing 1,4-d2-1 and in following thermal isomerizations
as functions of temperature is described in less than two
1
solved in CDCl3, and analyzed by H NMR using a 60
MHz instrument. Five to eight spectra from each product
mixture were recorded, and the integrated intensities
were averaged. The NMR absorptions at δ 6.0 for olefinic
protons and at δ 2.2 for the allylic hydrogens were
digitized and integrated by computer methods.9
The kinetic data led to the parameters ∆Eq ) 40 ( 1
kcal/mol, ∆Hq 39 ( 1 kcal/mol, ∆Sq - 4 ( 2 e.u., ∆Gq 41.0
( 0.5 kcal/mol, and log A′ ) 12.7 ( 0.4.9 When the
present Arrhenius plot and de Dobbelaere’s are com-
pared, the identical slopes and very distinct intercepts
are vividly apparent (Figure 3). Yet the apparent conflict
between the two experimental studies could be mislead-
ing. While the experimental details provided in the
dissertation9 did not include specifications of measured
NMR integration intensity ratios for product mixtures
or residence times for kinetic runs or derived rate
constants at specific temperatures or a definition of the
rate constant used for the Arrhenius plot, one can
nevertheless plausibly infer the cause of the apparent
discrepancy and demonstrate a fine agreement between
the two kinetic investigations.
(20) (a) de Dobbelaere, J. R.; de Haan, J. W.; Buck, H. M.; Visser,
G. J. Theor. Chim. Acta 1973, 31, 95-99. (b) de Dobbelaere, J. R.; Van
Dijk, J. M. F.; de Haan, J. W.; Buck, H. M. J. Am. Chem. Soc. 1977,
99, 392-397.
(21) Franzus, B. J. Org. Chem. 1963, 28, 2954-2960.
(22) Cramers, C. A. M. G.; Keulemans, A. I. M. J. Gas Chromatog.
1967, 5, 58-64.
(23) Cramers, C. A. Some Problems Encountered in High-Resolution
Gas Chromatography. Ph.D. Thesis, Eindhoven University of Technol-
ogy, The Netherlands, 1967.
(19) Benson, S. W.; O’Neal, H. E. Kinetic Data on Gas-Phase
Unimolecular Reactions; National Bureau of Standards: Washington,
DC, 1970; p 9.
J. Org. Chem, Vol. 70, No. 1, 2005 379

TABLE 3. Experimental and Calculated Ea Values (in
SCHEME 3
SCHEME 4
kcal/mol) for 1,5-Hydrogen Shifts in Cyclic 1,3-Dienes
diene
Ea (exptl) Ea(calcd)a ∆Ea(calcd - exptl)
1,3-cyclohexadiene
1,3-cycloheptadiene
cis-1,3-cyclooctadiene
40.1
41.9
33.7
32.2
1.8
6.2
3.2
27.5b
29.0c
a Ref 14. b Ref 13. c Ref 11.
show reasonable agreements (Table 3). The agreement
for cyclohexadiene, the main concern of the present work,
is the best of this limited set. Further experimental and
theoretical work on degenerate thermal 1,5-hydrogen
shifts in these and related cyclic dienes should serve to
uncover or rule out possible variations in contributions
from tunneling as functions of geometries of transition
structures.
Experimental Section
The kinetic scheme for interconversion of 1,4-d2-1 and
2,5-d2-1 given in Scheme 3 will serve well as long as there
is no significant participation of deuterium shifts. Were
they to intrude, the kinetic situation outlined in Scheme
4 would be relevant, and the challenges posed by data
reduction requirements would be substantially more
difficult.
The ratio of absorption intensities for allylic to vinyl
hydrogens at 1,4-d2-1 is 2:1, and for 2,5-d2-1 it is 1:1. The
reversible interconversion of 1,4-d2-1 and 2,5-d2-1 start-
ing from 1,4-d2-1 at 100 mol % would be governed by a
single-exponential function when expressed as in eq 4,
if kH . kD so that kD may be neglected.
1-Deuterio-2-cyclohexenol (3).15 To a solution of cyclohex-
enone (3.84 g, 0.04 mol) and CeCl3‚7H2O (14.92 g, 0.04 mol) in
20 mL of MeOH was added NaBD4 (1.68 g, 0.04 mol) over a 10
min period. The reaction mixture was stirred for 3 h at room
temperature and then quenched with H2O (20 mL) and extracted
with pentane (5 × 15 mL). The combined organic extracts were
washed with H2O (20 mL) and dried over Na2SO4. Filtration and
concentration by patient fractional distillation at atmospheric
pressure gave 3.84 g (97% yield) of 3 as a slightly yellow oil. 1H
NMR δ 1.6-2.1 (m, 6H), 5.7 (d, 1H), 5.8 (m, 1H).
Monodeuterio-1,3-cyclohexadiene (2-d-1, 1-d-1, 5-d-1).16
To a stirred solution of 3 (2.54 g, 0.026 mol) and CeCl3 (14.54 g,
0.039 mol) in acetonitrile (20 mL) was added NaI (1.95 g, 0.039
mol). The reaction mixture was heated to reflux for 5 h and then
cooled, diluted with pentane (10 mL), and quenched with 0.5 N
HCl (10 mL). The layers were separated, and the aqueous layer
was extracted with pentane (3 × 10 mL). The combined organic
material was washed with aq NaHCO3 (10 mL) and brine (10
mL) and dried over Na2SO4. Filtration, concentration, and micro-
spinning band column distillation at atmospheric pressure gave
a mixture of 2-d-1, 1-d-1, and 5-d-1 contaminated with several
minor impurities (see Figure 1). 1H NMR δ 1.7-2.0 (m, 6H), 5.73
1.25 × [1,4-d2-1(t)] - 25 ) [1,4-d2-1(0)] exp(-5kt)
(4)
The rate constant for the observable first-order decay
would be the sum of the forward and reverse rate
constants in Scheme 3. As the reaction progressed toward
equilibrium (20% 1,4-d2-1, 80% 2,5-d2-1) the allylic:vinyl
ratio of NMR intensities would progress from 2 toward
1.14. At any reaction time, the mole fraction of 1,4-d2-1
would be equal to 3 (r - 1)/(r + 1), where r is the
observable ratio of these NMR intensities.
2
(d, 1H), 5.81-5.87 (m, 1H). H NMR δ 7.41 (s), 5.93 (s) (51.8%
2-d-1), 5.84 (s) (42.4% 1-d-1), 5.72 (d), 2.16 (s) (5.7% 5-d-1).
Thermal Isomerizations of d-1(0).24 The mixture of mono-
deuterium-labeled 1,3-cyclohexadienes d-1(0) (51.8% 2-d-1, 42.7%
1-d-1, and 5.5% 5-d-1 ) (0.278 g, 3.5 mmol) was diluted 1:15 by
volume with cyclohexane. A 150 µL portion of this dilute solution
containing d-1(0) (0.112 g, 1.4 mmol) was injected into an
evacuated 300 mL kinetic bulb stabilized to ( 0.1 °C and allowed
to react for a defined time (Table 1). The product mixture and
cyclohexane bath gas were actively transferred to a U tube and
trapped in liquid nitrogen. The trapped hydrocarbons were then
warmed and passively transferred on a vacuum line into an
evacuated NMR tube containing degassed chloroform cooled in
For a 60 s kinetic run at 330 °C, the instance of max-
imum conversion of starting material, the rate constant
calculated from the Arrhenius parameters (1.6 × 10-2 s-1
)
would correspond to about 1.4 half-lives, or a reduction
of 1,4-d2-1 from 100 to 51% of the mixture or of r from 2
to 1.4. The kinetic runs thus were restricted to relatively
low conversions so as to avoid complications from kD
shifts which, at high conversion factors, might have been
problematic.
2
liquid nitrogen. The NMR tube was sealed off, and a H NMR
spectrum was recorded (see Figure 2).
Acknowledgment. We thank the National Science
Foundation for support of this work through REU
Grants CHE-9987838 and CHE-0211120; The Camille
and Henry Dreyfus Foundation for a Jean Dreyfus
Boissevain Undergraduate Scholarship for B.R.C. in
2002 and 2003; and Dr. H. M. Buck, Professor of
Organic and Theoretical Chemistry, Emeritus, and Ms.
Anneke Vriens, Eindhoven University of Technology, for
helpful correspondence.
The rate constants used in the Arrhenius plot (Figure
3) for reactions starting from 1,4-d2-1 may have been for
the exponential approach to the equilibrium proportions
of the two isomers and thus been equal to 5kH (Scheme
3; eq 4). Then, log A′ ) (12.7 ( 0.4)9 should be equal to
(log 5 + log A) ) 0.7 + (12.1 ( 0.3), which is indeed the
case. The apparent inconsistency of Figure 3 is most
probably just a consequence of different definitions of rate
constant. This reading of the kinetic analysis leads to an
agreement in Ea and log A values secured in Eindhoven
and at Syracuse that could hardly be better.
JO048441E
Comparisons between calculated14 and experimentally
determined Ea values for hydrogen shifts in cyclic dienes
(24) Compare: Baldwin, J. E.; Burrell, R. C. J. Org. Chem. 2002,
67, 3249-3256.
380 J. Org. Chem., Vol. 70, No. 1, 2005