The carbene fragmentation - Ring expansion route to bridgehead carbocations

Article abstract of DOI:10.1021/ol020083j

(Matrix Presented) The products, kinetics, and activation parameters were determined for the fragmentation-ring expansions of 1-norbornylmethyloxychlorocarbene (11) and 3-noradamantylmethyloxychlorocarbene (17). Products from 11 (in dichloroethane) included 1-chlorobicyclo[2.2.2]octane (9, 56%) and 1-chlorobicyclo[3.2.1]octane (10, 37%); from 17, we obtained 1-chloroadamantane (15, 68%) and protoadamantyl chloride (16, 28%).

Full text of DOI:10.1021/ol020083j

ORGANIC
LETTERS
2
002
Vol. 4, No. 14
341-2344
The Carbene Fragmentation−Ring
Expansion Route to Bridgehead
Carbocations
2
Robert A. Moss,* Fengmei Zheng, Jean-Marie Fed e´ , and Ronald R. Sauers*
Department of Chemistry and Chemical Biology, Rutgers, The State UniVersity of New
Jersey, New Brunswick, New Jersey 08903
moss@rutchem.rutgers.edu
Received April 18, 2002
ABSTRACT
The products, kinetics, and activation parameters were determined for the fragmentation−ring expansions of 1-norbornylmethyloxychlorocarbene
11) and 3-noradamantylmethyloxychlorocarbene (17). Products from 11 (in dichloroethane) included 1-chlorobicyclo[2.2.2]octane (9, 56%) and
-chlorobicyclo[3.2.1]octane (10, 37%); from 17, we obtained 1-chloroadamantane (15, 68%) and protoadamantyl chloride (16, 28%).
(
1
6
The energy and stability of bridgehead carbocations are
directly related to the strain imposed upon them by enforced
nonplanarity of the cationic carbon. Accordingly, certain
(DCE) at 25 °C, representing a rate enhancement of 5 ×
1
1
2
10 over the triflate acetolysis.
1
bridgehead carbocations are highly strained, very unstable,
1
and difficult to form by solvolytic processes. For example,
acetolysis of 1-norbornyl triflate, via the 1-norbornyl cation
-
8
-1
q
(
2
1), proceeds at 70 °C with k ) 6.5 × 10
s
and ∆H )
2
8.2 kcal/mol. On the other hand, we find that the
Bridgehead carbocations can also arise by ring expansions.
For example, 1-adamantylmethyl tosylate undergoes con-
certed ring expansion-acetolysis to 1-homoadamantyl ac-
7
etate via the 1-homoadamantyl cation, 2. Analogously,
fragmentation of 1-adamantylmethoxychlorocarbene in MeCN
or DCE affords products largely derived from ring-expanded
5
,8
cation 2
3
fragmentation of alkoxychlorocarbenes in polar solvents
provides alkyl cations (as components of ion pairs) with low
activation energies (<10 kcal/mol);4 cf. eq 1. Thus, cation
Here, we examine the 1-bicyclo[2.2.2]octyl (3) and 1-ada-
mantyl (4) cations as generated by carbene fragmentation-
,5
6
ring expansion reactions or by direct fragmentations. We
1
could be generated from 1-norbornyloxychlorocarbene with
focus on two unusual features of the ring expansion reac-
tions: (1) due to the low activation energies associated with
fragmentation-ring expansions, competitiVe ring expansions
occur; (2) the ion pairs formed via ring expansion or direct
4
-1
k
frag ) 3.3 × 10 s (E
a
) 9.0 kcal/mol) in dichloroethane
(
1) (a) Abboud, J.-L. M.; Herreros, M.; Notario, R.; Lomas, J. S.; Mareda,
J.; M u¨ ller, P.; Rossier, J.-C. J. Org. Chem. 1999, 64, 6401. (b) Abboud,
J.-L. M.; Castano, O.; Della, E. W.; Herreros, M.; M u¨ ller, P.; Notario, R.;
Rossier, J.-C. J. Am. Chem. Soc. 1997, 119, 2262.
(6) Moss, R. A.; Zheng, F.; Fed e´ , J.-M.; Ma, Y.; Sauers, R. R.; Toscano,
J. P.; Showalter, B. M. J. Am. Chem. Soc. 2002, 124, 5258.
(7) Liggero, S. H.; Sustmann, R.; Schleyer, P. v. R. J. Am. Chem. Soc.
1969, 91, 4571.
(8) Moss, R. A.; Ge, C.-S.; Makismovic, L. J. Am. Chem. Soc. 1996,
118, 9792.
(
(
(
(
2) Bingham, R. C.; Schleyer, P.v. R. J. Am. Chem. Soc. 1971, 93, 3189.
3) Moss, R. A. Acc. Chem. Res. 1999 32, 969.
4) Yan, S.; Sauers, R. R.; Moss, R. A. Org. Lett. 1999, 1, 1603.
5) Moss, R. A.; Johnson, L. A.; Yan, S.; Toscano, J. P.; Showalter, B.
M. J. Am. Chem. Soc. 2000, 122, 11256.
1
0.1021/ol020083j CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/17/2002

fragmentation reactions in methanol are not identical.
Presumably they do not live long enough to equilibrate within
9 or 10, respectively. In competition with ring expansion,
chloride attack at the original methylene group provides 7%
and with their solvent cages.
of chloride 8 (which might also arise by S
chloride directly on carbene 11; see below).
N
2 attack of
9
17
1-Norbornylcarboxylic acid and 3-noradamantylcarboxy-
lic acid (Aldrich) were reduced with LiAlH
methanol and 3-noradamantylmethanol, respectively. These
alcohols were then converted to the isouronium salts 5 and
4
to 1-norbornyl-
Of special note is the competition between alternative ring
expansions a and b in Scheme 1. Expansion a (of the “short”
1-carbon bridge) exceeds expansion b (of the “long” 2-carbon
bridge) by a statistically corrected factor of 3.0 in the carbene
fragmentation reactions. In contrast, acetolysis of 1-nor-
bornylmethyl tosylate is reported to give only 1-carbon bridge
9
10
6
by reactions with 1 equiv each of cyanamide and
11
trifluoromethanesulfonic acid. Oxidations of 5 and 6 with
12
12% aqueous NaOCl afforded diazirines 7a and 7b, which
1
8
were the desired carbene precursors. The diazirines displayed
expansion to 1-bicyclo[2.2.2]octyl acetate.
UV maxima in DCE at 351 and 364 nm (7a), and 353 and
Similarly, photolysis of diazirine 7b in DCE afforded 4%
of 3-noradamantyl chloride (14), 68% of 1-adamantyl
chloride (15), and 28% of protoadamantyl chloride (16);
Scheme 2. Product 14 was identified by GC-MS and MS
363 nm (7b).
Scheme 2
Photolysis of diazirine 7a at 25 °C in DCE (λ > 320 nm,
356 ) 0.5) gave 7% of 1-norbornylmethyl chloride (8), 56%
A
of 1-chlorobicyclo[2.2.2]octane (9), and 37% of 1-chlorobi-
cyclo[3.2.1]octane (10). Product identities were confirmed
by GC-MS, NMR, or GC comparisons with authentic
materials.1
3-16
We formulate these reactions as shown in Scheme 1, where
photolysis of diazirine 7a generates 1-norbornylmethyloxy-
chlorocarbene (11).
comparisons to literature data,19 whereas 15 was identical
to an authentic sample (Aldrich) in GC-MS and GC spiking
experiments. The identity of protoadamantyl chloride, 16 (M
+
Scheme 1
at m/e 170, 172), was assigned in analogy to the rearrange-
ments of carbene 11 and by exclusion, because isomers 14
2
0
and 15 are already assigned. Additionally, treatment of
3
2
-noradamantylmethanol with SOCl and pyridine (120 °C,
1
5 min) gave 14, 15, and 16 in an 11:85:4 distribution.
These reactions can be formulated via 3-noradamantyl-
methoxychlorocarbene, 17 (Scheme 2), in parallel to the
reactions of 1-norbornylmethoxychlorocarbene, 11 (Scheme
1
). Ion pairs 18 and 19 are drawn in analogy, and the
statistically corrected“short-bridge/long-bridge” (a/b) ring
expansion ratio is ∼4.8 for the fragmentation of 17 (com-
pared to 3.0 in the fragmentation of 11). As in the case of
1
-norbornylmethyl tosylate, acetolysis of 3-noradamantyl-
methyl tosylate is reported to give only short-bridge ring
2
1
expansion to 1-adamantyl acetate.
3
Fragmentation of the carbene, mainly with ring expan-
sion,5 gives ion pairs 12 or 13, which collapse to products
,8
(
14) Denney, D. B.; Garth, B. H.; Hanifin, J. W., Jr.; Relles, H. M.
Phosphorus Sulfur 1980, 8, 275.
(
(
(
9) Bixler, R. L.; Niemann, C. J. Org. Chem. 1958, 23, 742.
10) Tae, E. L.; Zhu, Z.; Platz, M. S. J. Phys. Chem. A 2001, 105, 3803.
11) Moss, R. A.; Kaczmarczyk, G. M.; Johnson, L. A. Synth. Commun.
(15) Kostova, K.; Dimitrov, V. Synth. Commun. 1995, 25, 1575.
(16) Della, E. W.; Tsanaktsidis, J. Aust. J. Chem. 1989, 42, 61.
(17) Moss, R. A.; Johnson, L. A.; Merrer, D. C.; Lee, G. E., Jr. J. Am.
Chem. Soc. 1999, 121, 5940.
2
000, 30, 3233. The isouronium salts were characterized by NMR and
elemental analysis.
(18) (a) Chenier, P. J.; Kiland, P. J.; Schmitt, G. D.; VanderWegen, P.
G. J. Org. Chem. 1980, 45, 5413. (b) See also Wilt, J. W.; Dabek, H. F.,
Jr.; Berliner, J. P.; Schneider, C. A. J. Org. Chem. 1970, 35, 2402.
(19) Mueller, A. M.; Chen, P. J. Org. Chem. 1998, 63, 4581.
(20) Chloride 16 has been prepared by a lengthy sequence, but no
spectroscopic data were offered: Vogt, R. B. Tetrahedron Lett. 1968, 9,
1575.
(
12) Graham, W. H. J. Am. Chem. Soc. 1965, 87, 4396.
9
(13) Chloride 8 was prepared from 1-norbornylmethanol, SOCl2, and
1
pyridine (reflux, 30 min) and characterized by H NMR comparison to the
literature and GC-MS. Chlorides 9 and 10 were prepared from 1-nor-
bornylmethanol, concentrated HCl, and ZnCl2, which afforded an 85:15
mixture of 9 and 10. These were characterized by GC, GC-MS, and
NMR comparisons to the reported spectra.
1
4
1
5
13
C
1
5,16
(21) Schleyer, P. v. R.; Wiskott, E. Tetrahedron Lett. 1967, 8, 2845.
2342
Org. Lett., Vol. 4, No. 14, 2002

The kinetics of the fragmentations of carbenes 11 and 17
cally lower activation energies associated with carbene
fragmentation reactions, which empower them to explore
pathways not available in tosylate solvolysis. Related effects
are observed in the fragmentations of exo- and endo-2-
were determined by laser flash photolysis (LFP).3
,5,6,22
LFP
at 351 nm and 25 °C of diazirine 7a in DCE (A356 ) 0.5) in
the presence of pyridine23 produced an ylide absorption at
2
2
4
16 nm due to pyridine trapping of carbene 11. A correlation
norbornyloxychlorocarbenes.
of the apparent rate constants for ylide formation (kobs
)
Above, we suggested that the “unrearranged” carbene
fragmentation products 8 (Scheme 1) and 14 (Scheme 2)
6
-1
(
2.0-5.5) × 10 s ) vs [pyridine] (1.65-7.42 M) was linear
5
-1
(
seven points, r ) 0.998) with a slope of 5.43 × 10 M
could arise (at least in part) by S 2 reactions of chloride
N
-1
s , equivalent to the rate constant for ylide formation (k
y
),
ions with carbenes 11 or 17, respectively. Such reactions
5
-1
and a Y-intercept of 7.72 × 10 s , equivalent to the oVerall
rate constant for fragmentation (kfrag) of carbene 11. See
Figure 1 in the Supporting Information. Using the product
distribution shown in Scheme 1, we can partition the
are known for, e.g., n-butoxychlorocarbene, and avoid the
1
7
formation of primary carbocations. Computations (see
below) indicate that the unassisted fragmentation of carbene
11 to primary chloride 8 requires an activation energy of
∼15.3 kcal/mol, which would not be competitive with the
5
-1
aggregate kfrag into kfrag(a) ∼ 4.3 × 10 s for 11 f 9 and
5
-1
k
frag(b) ∼ 2.9 × 10 s for 11 f 10. Note that fragmentation
fragmentations of 11 to tertiary chlorides 9 or 10 (E ≈ 5
a
b is twice as likely as fragmentation a.
kcal/mol; see above), and further suggests that the 11 f 8
Similarly, we measured the aggregate kfrag for carbene 17
conversion may be an S
expect the S 2 reaction to have E
to be competitive with fragmentation-rearrangement path-
ways. Note, however, that S 2 chloride attack on carbene
11 (a “neopentyl” system) will be more difficult than
comparable attack on n-BuOCCl, for which E
≈ 3 kcal/
N
2 reaction of 11 with chloride. We
in DCE: kfrag ) 2.0 × 10 s (7 points, r ) 0.981).24 In
6
-1
4,17
N
a
≈ 4-6 kcal/mol and
this case, the appropriate partition of kfrag (Scheme 2) leads
5
-1
to kfrag(a) ≈ 1.4 × 10 s for 17 f 15 and kfrag(b) ≈ 0.6 ×
N
5
-1
1
0 s for 17 f 16. Again, fragmentation b is statistically
favored by a factor of 2.
a
1
7
An Arrhenius study of the fragmentation of carbene 11
was carried out in DCE from -40 to 30 °C; the resulting
fair (r ) 0.959) correlation of ln(kfrag) vs 1/T is shown in
mol.
In support of these ideas, we find that decomposition of
+
-
11 in the presence of 1.2 M added n-Bu N Cl in DCE
4
Figure 2; see Supporting Information. The derived (ag-
affords an 8/9/10 product mixture that is enhanced in 8
q
gregate) values of E
a
(5.1 kcal/mol) and ∆S (-17 eu, ln A
(distribution 20:46:34), relative to the “normal” distribution
(7:56:37, see Scheme 1, above). Moreover, LFP of diazirine
7a in DCE with a constant concentration (5.77 M) of pyridine
-
1
)
21.8 s ) are quite comparable to the analogous data for
the fragmentation of 1-adamantylmethoxychlorocarbene, for
q
5
+
which E
We also determined E
eu for the fragmentation of carbene 17 in DCE (seven points,
40 to +30 °C, r ) 0.944). Although the data are no more
a
) 4.6 kcal/mol and ∆S ) -18 e.u.
and varying quantities (0.042-0.504 M) of added n-Bu N -
4
q
-
a
) 5.2 kcal/mol and ∆S ) -14
Cl gives a linear correlation (seven points, r ) 0.999) of
-
k
obs for ylide formation vs [Cl ]. The slope of this correlation,
6
-1 -1
-
8.2 × 10 M s , can be taken as the second-order rate
constant (k ) for quenching of carbene 11 by chloride; i.e.,
fragmentation of the carbene induced by chloride.
Parallel results are obtained for the fragmentation of
than adequate (significant linearity at the 99% confidence
level), it is clear that the aggregate activation parameters for
the fragmentations of 1-adamantylmethoxychlorocarbene,
2
17,23,25
+
-
carbene 11, and carbene 17 are similar and that E
mol in each case.
a
≈ 5 kcal/
carbene 17. In the presence of 1.26 M added n-Bu N Cl ,
4
the 14/15/16 distribution (Scheme 2) changes from 4:68:28
These low activation energies allow the fragmentations
of 11 and 17 to explore the competitive ring expansion
pathways depicted in Schemes 1 and 2. In contrast, the triflate
or tosylate analogues of these carbenes undergo acetolysis
with expansions only of their “short” bridges. The substantial
(no added chloride) to 8:61:27 (with a 4% of an unknown
7
-1 -1
product), while k
induced fragmentation.
2
) 2.5 × 10 M
s
for the chloride-
2
6
Photolysis of diazirine 7a in MeOH gave methyl ethers
20, 21, and 22, in addition to the analogous chlorides 8, 9,
and 10. A substantial quantity of 1-norbornylmethanol (23)
also formed.
The structures of 20-22 were confirmed by GC, NMR,
and MS comparisons with authentic samples made by the
activation enthalpies for these reactions (28.2 kcal/mol for
2
1
-norbornylmethyl triflate and 20.7 kcal/mol for 3-norada-
21
mantylmethyl tosylate ) permit sufficiently large differential
activation enthalpies between short- and long-bridge ring
expansions in favor of exclusive short-bridge rearrangements.
From the measured kfrag values of carbenes 11 and 17, and
the reported rate constants for acetolyses of their tosylate
4
AgClO -mediated methanolysis of 1-norbornylmethyl io-
2
7
dide.
analogues,2 we estimate kfrag/kacet ≈ 1 × 10 in the
,21
16
11
1
-norbornylmethyl series and ∼2 × 10 in the 3-norada-
mantyl series. These huge enhancements reflect the dramati-
(
22) Moss, R. A., Zheng, F.; Sauers, R. R.; Toscano, J. P. J. Am. Chem.
Soc. 2001, 123, 8109.
23) Jackson, J. E.; Soundararajan, N.; Platz, M. S.; Liu, M. T. H. J.
Am. Chem. Soc. 1988, 110, 5595.
24) A353 ) 0.8 for diazirine 7b in DCE. The pyridinium ylide of carbene
(
Alcohol 23, which forms from carbene 11 in 25% yield
at even low concentrations (øMeOH ) 0.093) of methanol in
(
5
-1 -1
1
7 had λmax ) 320 nm and ky ) 1.1 × 10 M
s .
Org. Lett., Vol. 4, No. 14, 2002
2343

DCE, and in 30% yield in pure MeOH, is a carbene trapping
product: MeOH captures 11 in competition with fragmenta-
tion, producing ROCH(Cl)OMe, which yields orthoformate
the LYP correlation functional of Lee, Yang, and Parr.33
Computed gas-phase energies (unscaled) were corrected for
thermal effects at 298.15 K and for zero-point energy
differences. Normal coordinate analyses confirmed the nature
of the computed stationary points as either ground state (no
imaginary frequencies) or transition structures (one imaginary
frequency). Solvation effects in dichloroethane (ꢀ ) 10.36)
were simulated via single-point energy calculations and
6
,28
ROCH(OMe)
2
and HCl by acid-catalyzed hydrolysis.
Products RCl and ROMe, however, represent competition
+
+
between chloride return to R vs methanol capture of R in
ion pairs 12 and 13. These products (9, 10, 21, and 22)
represent ring expansion-fragmentation of 11, whereas ether
3
2
2
0 may stem from solvolytic fragmentation of the carbene.
In pure methanol, the product distribution from carbene
1 is chlorides 8 (3.0), 9, (26.0), and 10 (15.1); ethers 20
PCM methodology.
The fragmentation transition structures for the rearrange-
ments of carbene 11 to chlorides 9 and 10 are shown in
Figure 3; see Supporting Information. The computed activa-
tion energies at 298.15 K, converted from activation enthal-
pies, and corrected for zero-point energies, are 12.5 (9) and
13.8 (10) kcal/mol in vacuo and 4.84 (9) and 5.94 (10) kcal/
1
(1.3), 21 (10.5), and 22 (13.7); alcohol 23 (30.3). We note
first that, even in pure methanol, alkyl chloride formation
3
from carbene 11 persists; the total chloride/ether ratio is
44:26. Thus, the lifetimes of the putative ion pairs formed
upon carbene fragmentation (12 and 13 in Scheme 1) must
mol in simulated DCE solution. The computed E
rearrangements, 4.8-5.9 kcal/mol, are in excellent agreement
with the experimental (aggregate) E for the disappearance
of carbene 11 (5.1, kcal/mol, see above).
The computed E for the “unimolecular” fragmentation
of 11 to unrearranged 8 is 21.2 kcal/mol in vacuo and 15.3
kcal/mol in simulated DCE. The calculations therefore
suggest that the 11 f 8 pathway should not be competitive
with either the 11 f 9 or 11 f 10 rearrangements. The
finding of 7% of 8 from the fragmentation of 11 is most
a
’s for these
be comparable to the diffusive events that would remove
-
29,30
Cl and fully solvate the carbocations.
a
Furthermore, the bicyclo[2.2.2]octyl cation-chloride anion
pair (12) generated by carbene fragmentation-ring expansion
of carbene 11 in MeOH (Scheme 1) is not identical to the
a
“
1
same” ion pair generated by the direct fragmentation of
-bicyclo[2.2.2]octyloxychlorocarbene. Thus, the RCl/
6
ROMe product ratio (9/21) is 26:10.5 ≈ 2.5 from fragmenta-
tion-rearrangement of carbene 11, but only 19:48.6 ≈ 0.39
5
from fragmentation of 1-bicyclo[2.2.2]octyloxychlorocarbene.
likely the result of chloride ion-induced S
of the carbene, as suggested above.
N
2 fragmentation
Ion pair 12 “remembers” its carbene origin; its choice
between reaction with chloride or methanol depends on its
origin.
The migrating carbons of the transition structures for 9
and 10 lie midway along the reaction coordinates in contrast
to the “early” transition structures computed for the cyclo-
3
1
Similar results obtain for ion pair 18 in Scheme 2: the
RCl/ROMe ratio (15/1-Ad-OMe) is 37:15 ≈ 2.5 from the
fragmentation-ring expansion of carbene 17 but 54.2/37 ≈
3
4
propylmethyl analogue (E
1-adamantylmethyl system (E ) 5.6 kcal/mol). Whereas
a
) 3.0 kcal/mol) and the
a
5
,8
1
.5 from the direct fragmentation of 1-adamantyloxychloro-
strain relief provides a significant driving force for the
1-norbornylmethyl rearrangements, strain increases for the
1-adamantylmethyl rearrangement. On the other hand, incipi-
ent homoconjugation may provide significant stabilization
of the cycloproplymethyl system with relatively little geo-
metric change. What makes the present systems unique are
the locations of the chloride ion in the transition structures.
In most other computed transition structures for these
fragmentation reactions, the Cl- - -C terminus distance ranges
5
carbene. That the RCl/ROMe ratios are more nearly equal
when the selecting carbocation is more stable (i.e., 4 vs 3),
1
seems reasonable, but a more extensive study of these
phenomena is clearly needed.
In analogy with previous computational studies of alkoxy-
halocarbene fragmentation,4,6 we computed energies for cis-
carbene 11 and the three competing fragmentation transition
structures leading to 8, 9, and 10. All structures were fully
optimized by analytical gradient methods at the B3LYP/6-
4
,5,6,34
from 3.0 to 3.6 Å,
but in the present case it is more
32
tightly bound at distances of 2.6-2.8 Å. The contrast reflects
an unexpected variability in the geometry of “early” frag-
mentation transition structures.
3
1G(d) level using the Gaussian98 suite of programs. DFT
calculations used Becke’s three-parameter hybrid method and
(
25) The values of k2 for carbene 11 (in DCE) and n-butoxychlorocarbene
1
7
(
in MeCN) are identical.
(
26) This result was obtained as described for carbene 11. The correlation
Acknowledgment. We are grateful to the National
Science Foundation for financial support and to the Center
for Computational Neuroscience of Rutgers University
-
of kobs with [Cl ] had r ) 0.971 for 6 points.
27) Kropp, P. J.; Poindexter, G. S.; Pienta, N. J.; Hamilton, D. C. J.
Am. Chem. Soc. 1976, 98, 8135.
28) Smith, N. P.; Stevens, I. D. R. J. Chem. Soc., Perkin Trans. 2 1979,
298. The orthoformate cannot be isolated; see pp 1302-1303.
29) From carbocation trapping experiments with trimethoxybenzene,5
(
(
(Newark) for computational support.
1
,30
11
(
-
we could estimate the lifetimes of the cations of Scheme 1 as ∼7 × 10
Supporting Information Available: Figures 1-3. This
material is available via the Internet at http://pubs.acs.org.
s in DCE. Thus, in methanol, the cations of the ion pairs react either with
their chloride counterions or with solvent molecules in the first shell of the
solvent cage.
OL020083J
(
30) Pezacki, J. P.; Shukla, D.; Lusztyk, J.; Warkentin, J. J. Am. Chem.
Soc. 1999, 121, 6589.
31) For examples of “memory effects” in carbocation rearrangements,
see: Berson, J. A. Angew. Chem., Int. Ed. Engl. 1968, 7, 799.
32) Gaussian 98, revision A.9; Gaussian, Inc.: Pittsburgh, PA, 1998.
(
(33) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B. 1998, 37, 785.
(34) Moss, R. A.; Zheng, F.; Johnson, L. A.; Sauers, R. R. J. Phys. Org.
Chem. 2001, 14, 400.
(
2344
Org. Lett., Vol. 4, No. 14, 2002