Carbocation-forming reactions in dimethyl sulfoxide

Article abstract of DOI:10.1021/jo026468x

Mesylate derivatives of 3-aryl-3-hydroxy-β-lactams and thiolactams react in DMSO-d₆ by first-order processes to give alcohol products. Substituent effect studies implicate carbocation intermediates (ion-pairs) that are captured by DMSO-d₆ to give transient oxosulfonium ions. Rapid reaction of the oxosulfonium ions with trace amounts of water leads to the alcohol product and regenerates DMSO-d₆. H₂¹⁷O labeling studies show that ¹⁷O is incorporated into the DMSO. The mesylate derivatives of endo- and exo-2-hydroxy-2-phenylbicyclo[2.2.1]heptan-3-one also react in DMSO-d₆ to give the alcohol products. Ion-pair intermediates that capture DMSO giving unstable oxosulfonium ions are again proposed. Exo-2-phenyl-endo-bicyclo[2.2.1]heptyl trifluoroacetate readily eliminates trifluoroacetic acid in DMSO-d₆ via a cationic mechanism involving loss of the endotrifluoroacetate leaving group as well as an exo-hydrogen. The O-methyl oxime derivative of α-chloroα,α-diphenylacetophenone reacts in DMSO-d₆ to give 1-methoxy-2,3-diphenylindole, a product derived from cyclization of a cationic intermediate. A common ion rate suppression provides further evidence for a cationic mechanism. The triflate derivative of pivaloin reacts by a cationic mechanism in DMSO-d₆ to give rearranged products. The rate is even faster than in highly ionizing solvents such as trifluoroethanol or trifluoroacetic acid. 1-Adamantyl mesylate reacts in DMSO-d₆ by a first-order process (Y_OMs = -4.00) to give a long-lived oxosulfonium ion, 1-Ad-OS(CD₃)₂⁺, which can be characterized spectroscopically. This oxosulfonium ion reacts only slowly with water at elevated temperatures to give 1-adamantanol. DMSO is therefore a viable solvent for k_s, k_C, and k_Δ cationic processes.

Full text of DOI:10.1021/jo026468x

Ca r boca tion -F or m in g Rea ction s in Dim eth yl Su lfoxid e
Xavier Creary,* Elizabeth A. Burtch, and Ziqi J iang
Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, Indiana 46556
creary.1@nd.edu
Received September 23, 2002
Mesylate derivatives of 3-aryl-3-hydroxy-â-lactams and thiolactams react in DMSO-d6 by first-
order processes to give alcohol products. Substituent effect studies implicate carbocation intermedi-
ates (ion-pairs) that are captured by DMSO-d6 to give transient oxosulfonium ions. Rapid reaction
of the oxosulfonium ions with trace amounts of water leads to the alcohol product and regenerates
DMSO-d6. H217O labeling studies show that 17O is incorporated into the DMSO. The mesylate
derivatives of endo- and exo-2-hydroxy-2-phenylbicyclo[2.2.1]heptan-3-one also react in DMSO-d6
to give the alcohol products. Ion-pair intermediates that capture DMSO giving unstable oxosulfo-
nium ions are again proposed. Exo-2-phenyl-endo-bicyclo[2.2.1]heptyl trifluoroacetate readily
eliminates trifluoroacetic acid in DMSO-d6 via a cationic mechanism involving loss of the endo-
trifluoroacetate leaving group as well as an exo-hydrogen. The O-methyl oxime derivative of R-chloro-
R,R-diphenylacetophenone reacts in DMSO-d6 to give 1-methoxy-2,3-diphenylindole, a product
derived from cyclization of a cationic intermediate. A common ion rate suppression provides further
evidence for a cationic mechanism. The triflate derivative of pivaloin reacts by a cationic mechanism
in DMSO-d6 to give rearranged products. The rate is even faster than in highly ionizing solvents
such as trifluoroethanol or trifluoroacetic acid. 1-Adamantyl mesylate reacts in DMSO-d6 by a first-
order process (YOMs ) -4.00) to give a long-lived oxosulfonium ion, 1-Ad-OS(CD3)2+, which can be
characterized spectroscopically. This oxosulfonium ion reacts only slowly with water at elevated
temperatures to give 1-adamantanol. DMSO is therefore a viable solvent for ks, kC, and k∆ cationic
processes.
In tr od u ction
from activation of DMSO with reagents such as trifluo-
roacetic anhydride, 3, or oxalyl chloride can also lead to
oxidation processes. Such reagents have proven to be the
method of choice for conversion of certain alcohols to
aldehydes and ketones.4
Dimethyl sulfoxide (DMSO), 1, is a common solvent
and reagent in organic chemistry.1 Organic compounds
tend to be quite soluble in this solvent, and the deuter-
ated analogue CD3SOCD3 (DMSO-d6) is a commonly used
solvent for recording NMR spectra. DMSO is also char-
acterized by its high dielectric constant (E ) 48.9), and
many ionic substances have relatively high solubility in
this solvent. Due to the inability to hydrogen bond, many
anions have extraordinary nucleophilic reactivity as well
as basicity in DMSO. SN2 and E2 reactions tend to occur
much more rapidly in this solvent. DMSO is quite
hygroscopic, and it is therefore difficult to exclude all
traces of water. In fact, DMSO at high temperature is
an effective dehydrating agent for certain alcohols.2
DMSO can also act as an oxidizing agent, and the
Kornblum oxidation, which proceeds via the alkylated
DMSO 2, is a facile method for conversion of alkyl halides
and tosylates to carbonyl compounds.3 Reagents derived
Over the years, we have carried out numerous sol-
volytic studies in protic solvents on substrates that react
readily at room temperature by carbocationic mecha-
nisms. Recording NMR spectra of these substrates in the
aprotic solvent DMSO-d6 was often problematic since
these substrates also react in this solvent by a mecha-
nism that we now conclude is carbocationic in nature.
Indeed, reactions that involve capture of tropylium
cation5 and tert-butyl cation6 by DMSO have been pro-
posed in the past. However, there has been no systematic
study of reactions that proceed via carbocationic inter-
mediates in DMSO. The compounds selected for study
(1) For reviews and leading references, see: (a) Buncel, E.; Wilson,
H. Adv. Phys. Org. Chem. 1977, 14, 133. (b) Martin, D.; Weise, A.;
Niclas, H.-J . Angew Chem., Int. Ed. Engl. 1967, 6, 318.
(2) (a) Traynelis, V. J .; Hergenrother, W. L.; Livingston, J . R.;
Valicenti, J . A. J . Org. Chem. 1962, 27, 2377. (b) Traynelis, V. J .;
Hergenrother, W. L.; Hanson, H. T.; Valicenti, J . A. J . Org. Chem. 1964,
29, 123. (c) Traynelis, V. J .; Hergenrother J . Org. Chem. 1964, 29, 221.
(3) (a) Kornblum, N.; Powers, J . W.; Anderson, G. J .; J ones, W. J .;
Larson, H. O.; Weaver, W. M. J . Am. Chem. Soc. 1957, 79, 6562. (b)
Kornblum, N.; J ones, W. J .; Anderson, G. J . J . Am. Chem. Soc. 1959,
81, 4113. (c) J ohnson, A. P.; Pelter, A. J . Chem. Soc. 1964, 520.
(4) Mancuso, A. J .; Swern, D. Synthesis 1981, 165.
(5) Garfunkel, E.; Reingold, I. D. J . Org. Chem. 1979, 44, 3725.
(6) Dossena, A.; Marchelli, R.; Casnati, G. J . Chem. Soc., Chem.
Commun. 1979, 370.
10.1021/jo026468x CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/15/2003
J . Org. Chem. 2003, 68, 1117-1127
1117

Creary et al.
1
F IGURE 1. Evolving H NMR spectra during reaction of 4 with 0.10 M NaN3 in DMSO-d6 at 50 °C.
in this paper were readily available in our laboratory,
having been examined by us in the past in protic solvents
where they have relatively high reactivity. Reported here
are some reactions of mesylates, trifluoroacetates, and
chlorides in DMSO-d6, many of which proceed via car-
bocation intermediates.
Resu lts a n d Discu ssion
â-Th iola cta m a n d â-La cta m System s. We recently
reported on the reaction of mesylate 4 with azide ion in
dimethyl formamide, which proceeds to give azide sub-
stitution product 5 with clean inversion.7 We have now
employed DMSO as a solvent and found that the reaction
of 4 with excess sodium azide gives varying amounts of
azide product 5 along with alcohol 6. Figure 1 shows a
typical reaction using 0.1 M NaN3, where comparable
amounts of 5 and 6 are formed.
F IGURE 2. Plot of the observed pseudo first-order rate
constant for reaction of 4 with NaN3 in DMSO-d6 at 50.0 °C
versus [NaN3].
As the NaN3 concentration is increased, the pseudo first-
order rate constant for the disappearance of 4 increased
(Figure 2) and less of alcohol 6 was formed. The limiting
slope of the plot shown in Figure 2 is 3.06 × 10-4 M-1
s-1 and corresponds to the second order rate constant for
(7) Creary, X.; Zhu, C.; J iang, Z. J . Am. Chem. Soc. 1996, 118, 12331.
1118 J . Org. Chem., Vol. 68, No. 3, 2003

Carbocation-Forming Reactions in Dimethyl Sulfoxide
TABLE 1. Ra tes of Rea ction of Su bstr a tes in DMSO-d 6
F IGURE 3. First-order kinetic plot for reaction of mesylate
4 in DMSO-d6 at 60.0 °C.
the azide reaction. The small deviations from linearity
at higher azide concentrations are probably due to
aggregation phenomena in the more concentrated solu-
tions.
When mesylate 4 was reacted in DMSO-d6 without
added NaN3, 4 disappeared in a clean pseudo first-order
process8 (Figure 3), and alcohol 3 was the only product
formed. A study on the optically active mesylate 4 showed
that the alcohol product was 94% inverted (6% race-
mized). To shed further light on the origin of this alcohol
product, a limited substituent effect study on the rate
was carried out. p-CH3 analogue 7 reacted 43 times faster
than 4, and p-CF3 analogue 8 reacted 333 times more
slowly than 4 (Table 1). The p-OCH3 substituted mesylate
analogue could not be prepared due to its high reactivity,
but the corresponding chloride 9 can be prepared and
reacted in DMSO-d6. All of these substrates react in
DMSO-d6 to give the corresponding alcohol products. If
one assumes a mesylate/chloride rate ratio9 of 3 × 104,
then the Hammett F+ value based on 4 and 7-9 is -4.9
(correlation coefficient ) 0.998). If data for 9 are omitted,
then the F+ value is -4.6.
It is proposed that these data are consistent with a
carbocation mechanism, where the solvent DMSO-d6 has
an “ionizing power” that is not insignificant. This belief
(8) Reaction is easily monitored by integration of the signals in
Figure 1 or by adding Et3N and monitoring the changing shift of the
methyl group of Et3N. See Experimental Section for details.
(9) Noyce, D. S.; Virgilio, J . A. J . Org. Chem. 1972, 37, 2643.
J . Org. Chem, Vol. 68, No. 3, 2003 1119

Creary et al.
F IGURE 4. Evolving 17O NMR spectra during reaction of 4 in DMSO-d6 containing 0.5% H217O at 60 °C. The peak at δ 11.5 at
time ) 0 is due to naturally occurring 17O in the DMSO-d6 solvent.
is based partially on the F+ value of -4.9. This value is
quite large and consistent with a transition state with
significant positive charge. The stereochemical course of
the reaction of optically active 4 suggests that a minimum
of 6% of this reaction proceeds via a carbocation mech-
anism (racemization). In fact, we suggest that the entire
reaction proceeds via a tight ion-pair mechanism where
nucleophile capture occurs preferentially from the op-
posite side of the departing mesylate leaving group. This
ion-pair is short-lived and strongly solvated by the
DMSO-d6 solvent from the rear of the leaving group.
Solvolysis reactions proceeding through such ion-pair
mechanisms, which give large amounts of net inversion,
have been observed in the past.10 Hence, the formation
of the largely inverted product 6 is not inconsistent with
a carbocation mechanism.
The rates of these substrates (Table 1) are all some-
what less than the previously reported rates6 in acetic
acid, a very common solvent in solvolysis studies. Car-
bonyl analogue 10 was also examined and reacts in a
similar fashion in DMSO-d6 to give the alcohol product
at a rate that is slightly slower than that of 4. p-CH3
analogue 11 reacts 37 times faster than 10, and this
substituent effect is also consistent with an analogous
ion-pair mechanism.
incorporated into the DMSO. Control experiments show
that (in the absence of mesylate 4) the oxygen atom of
DMSO-d6 does not undergo exchange with H217O even
when heated under the reaction conditions and even
when a trace of CH3SO3H is added. Examination of
alcohol product 6 reveals no 17O incorporation, while an
authentic sample of the 17O-labeled 6 showed an 17O
signal at δ 39.1. The source of the oxygen atom in alcohol
product 6 is therefore not residual water in the DMSO-
d6 but the DMSO-d6 itself.
A reasonable mechanism accounting for the results of
the labeling study involves capture of cationic intermedi-
ate 12 by DMSO-d6 to give oxosulfonium intermediate
13.11 Intermediate 13 could then react with water to give
14, which then breaks down to give 17O-labeled DMSO-
d6 and unlabeled alcohol product 6. Alternatively, 13
could react with water by nucleophilic attack at sulfur
and directly displacing the unlabeled oxygen, thereby
avoiding hypervalent intermediate 14.
A question arises as to the origin of the alcohol
products in DMSO-d6 as a solvent. Substitution product
6 has incorporated an OH group (and not OD) as can be
1
clearly observed in the H NMR spectrum in DMSO-d6
before workup. The logical source of 6 is the trace of water
(0.04%) present in the commercial “dry” DMSO-d6. To
probe this question in more detail, the reaction of 4 was
carried out in DMSO-d6 containing an added trace of
H217O (0.6%) and some buffering Et3N. The reaction was
then monitored by 17O NMR. Figure 4 shows that 17O is
Oth er Ca r boca tion -F or m in g Su bstr a tes. To fur-
ther probe DMSO as a solvent for carbocation-forming
(11) Oxosulfonium ions analogous to 9 have been proposed as
intermediates in reaction of R2SO and triflic anhydride with carbohy-
drates. These intermediates react with nucleopliles at carbon to
displace R2SO. See: Nguyen, H. M.; Chen, Y.; Duron, S. G.; Gin, D. Y.
J . Am. Chem. Soc. 2001, 123, 8766.
(10) (a) Weiner, H.; Sneen, R. A. J . Am. Chem. Soc. 1965, 87, 287.
(b) Streitwieser, A.,J r.; Walsh, T. D.; Wolfe, J . R., J r. J . Am. Chem.
Soc. 1965, 87, 3682. (c) Hughes, E. D.; Ingold, C. K.; Martin, R. J . L.;
Meigh, D. F. Nature 1950, 166, 679.
1120 J . Org. Chem., Vol. 68, No. 3, 2003

Carbocation-Forming Reactions in Dimethyl Sulfoxide
reactions, mesylate 15 was examined. Solvolysis reactions
of 15 in protic solvents proceed via R-keto carbocationic
intermediates that lead to exo-substitution products.12
Mesylate 15 disappears in a clean pseudo first-order
process in DMSO-d6, and rate data are given in Table 1.
Mesylate 15 in DMSO-d6 is 8 times less reactive than in
ethanol and 40 times less reactive than in acetic acid.
The product formed in DMSO-d6 (containing a slight
excess of triethylamine) is exclusively exo-alcohol 16. No
buildup of intermediates could be detected spectroscopi-
cally over the course of the reaction. An analogous
labeling experiment using 0.5% H217O in DMSO-d6
showed that 17O is again incorporated into the DMSO-
d6 and not into the alcohol product. The suggested
mechanism to account for the 17O labeling result again
involves formation of an oxosulfonium ion intermediate
derived from capture of DMSO-d6 from the exo-face. The
resultant oxosulfonium ion then reacts with the trace of
water present in the DMSO-d6, which results in the
observed alcohol product and label incorporation into the
DMSO-d6. The stereochemistry of 16 is consistent with
either an SN2 process utilizing DMSO-d6 as the nucleo-
phile or an ion-pair process where the cation preferen-
tially captures DMSO-d6 from the exo-face.
ratio of 14.4, although on the small side,15 is completely
consistent with ion-pair mechanisms for reactions of 15
and 17, where the ion-pairs are quite short-lived. The
complete inversion in the solvolysis of 15 suggests, as
expected, complete capture of ion-pair 19 from the exo-
face. On the other hand, solvation of ion-pair 20 from the
endo-side is not as effective and some solvent finds its
way to the exo-side. These stereochemical studies suggest
that ion-pairs 19 and 20 are quite short-lived and readily
collapse to covalent oxosulfonium ions 21.
Reaction of chloride 22 provides further evidence for
carbocation formation in DMSO-d6. At room temperature,
22 is cleanly converted to a mixture of indole 23 (83%)
and alcohol 24 (17%). As before, the minor alcohol product
is suggested to be derived from rapid reaction of the trace
amount of water with intermediate 27. Indole 23 is a
product of cationic cyclization and indicative of the
intermediacy of the resonance-stabilized carbocation 25.
This type of resonance stabilization of R-oximino cations
has been previously proposed,16 and the observation of
23 supports this notion. Cyclization of 25 also provides
an interesting route leading to the indole nucleus. Thus,
treatment of 24 with catalytic CF3CO2H in chloroform
also leads to a high yield of indole 23.
When the reaction of 22 in DMSO-d6 is monitored by
1H NMR spectroscopy (0.01 M in 22), the rate is subject
to a common ion rate suppression,17 i.e., the apparent
first-order rate constant for disappearance of 22 de-
creases as the reaction proceeds. Also, the addition of
tetramethylammonium chloride (0.1 M) at the beginning
of the reaction greatly suppresses the rate. This provides
additional evidence for the involvement of a cationic
intermediate. Intermediate 25 must live long enough to
arrive at the dissociated ion-pair stage, where externally
The reaction of exo-mesylate 17 in DMSO-d6 sheds
further light on the mechanism. Mesylate 17 reacts 14.4
times faster than endo-analogue 15. However, the inver-
sion of stereochemistry in this reaction is not clean and
an 84:16 ratio of alcohols 18 and 16 is produced. The exo/
endo rate ratio of 14.4 is inconsistent with SN2 processes.
In fact, exo-2-norbornyl brosylate is slightly less reactive
than endo-2-norbornyl brosylate in SN2 reactions.13 The
significant amount of exo-alcohol 16 derived from 17 is
also inconsistent with an SN2 mechanism. On the other
hand, an ion-pair mechanism analogous to that proposed
for 4 nicely accounts for the rate and stereochemical
results. It is well established that exo-2-norbornyl deriva-
tives solvolyze via cationic intermediates more rapidly
than endo-2-norbornyl derivatives.14 Our exo/endo rate
(14) For leading references, see: Brown, H. C. The Nonclassical Ion
Problem; Plenum Press: New York, 1977.
(15) Although the exo/endo ratio of 20.x for 17 and 15 might be
considered small, comparable values are seen in other 2-norbornyl
systems solvolyzing via cationic mechanisms. See: Wilcox, C. F.;
J esaitis, R. G. Tetrahedron Lett. 1967, 2567. (b) Wilcox, C. F.; J esaitis,
R. G. Chem. Commun. 1967, 1056. (c) Traylor, T. G. Perrin, C. L. J .
Am. Chem. Soc. 1966, 88, 4934.
(16) (a) Creary, X.; Wang, Y.-X.; J iang, Z. J . Am. Chem. Soc. 1995,
117, 3044. (b) Creary, X.; J iang, Z. J . Org. Chem. 1996, 61, 3482.
(17) Smith, M.; March, J . March’s Advanced Organic Chemistry;
Wiley-Interscience: New York, 2001; p 395.
(12) Creary, X. J . Org. Chem. 1979, 44, 3938.
(13) Banert, K.; Kirmse, W. J . Am. Chem. Soc. 1982, 104, 3766.
J . Org. Chem, Vol. 68, No. 3, 2003 1121

Creary et al.
added chloride can trap 25 and thereby return to starting
chloride 22.
Fluorenyl system 28 is another example of a substrate
that reacts in DMSO-d6 to give the corresponding alcohol
as the exclusive product. The rate of reaction of chloride
28 is also subject to a common ion rate suppression, and
addition of tetramethylammonium chloride at the begin-
ning of the reaction greatly slows the rate. Presumably,
ring strain precludes indole formation from the cationic
intermediate involved in this reaction. Although strict
first-order rate behavior is not observed due to the
common ion effect, the rate of reaction of 28 is qualita-
tively much slower than that of analogue 22. The time
required for 50% of 22 (0.01 M) to react at 50 °C is 32
min,18 while a temperature of 112 °C is required to
achieve comparable reactivity in 29. The features leading
to decreased solvolytic reactivity in fluorenyl systems
analogous to 28 have been discussed.19
using CD3 analogue 35-d3. Replacement of the methyl
hydrogens by deuterium slows the rate by a factor of 2.11
( 0.06. Previously determined R-CH3/CD3 isotope effects
for 2-chloro-2-methyladamantane in protic solvents ranged
from 1.48 to 1.68.22 Our value of 2.11 in DMSO-d6 is
considered to be too large for simple rate-determining
cation formation, where the expected value is 1.46.22 It
is therefore suggested that the reaction of 35 in DMSO-
d6 proceeds via the E2C+ mechanism,23 where the rate
limiting step is proton loss from the reversibly formed
2-methyladamantyl cation.
Trifluoroacetate 37 and chloride 39 also react readily
at room temperature in DMSO-d6 to give elimination
products 38 and 40, respectively, with no alcohol products
being formed. It therefore appears that, when feasible,
â-proton loss from cationic intermediates can be more
facile than capture of the carbocation by DMSO.
Solvolytic Elim in a tion s. Trifluoroacetates 30 and 31
undergo facile solvolytic eliminations in DMSO-d6 at
room temperature to give exclusively alkenes 32 and 33.
To probe these elimination processes in more detail, the
endo-deuterated substrates were also prepared and al-
lowed to react in DMSO-d6. The exclusive products were
the deuterated alkenes, where the exo-hydrogens are lost
in the elimination processes. These products are consis-
tent with stepwise mechanisms proceeding via carbocat-
ions 34, which preferentially lose the exo-hydrogen.20
Products 32 and 33 are inconsistent with concerted
eliminations of trifluoroacetic acid from 30 and 31.
Concerted eliminations such as E2 or ester pyrolysis-type
mechanisms should be cis elimination processes.21 Rate
data (Table 1) are also consistent with the intermediacy
of cation 34 since p-CF3 analogue 31 reacts 146 times
more slowly than unsubstituted analogue 24.
Trifluoroacetate 35 solvolyzes readily in DMSO-d6 at
slightly elevated temperatures to give elimination prod-
uct 36 exclusively. No alcohol product or oxosulfonium
ion intermediate could be detected. By way of contrast,
solvolyses of 2-chloro-2-methyladamantane in protic sol-
vents give mainly solvent capture products and only
smaller amounts of the elimination product.22 To shed
further light on this elimination reaction of 35, we have
measured the â-deuterium isotope effect in DMSO-d6
Oth er Ben zylic Su bstr a tes. ks Su bstr a tes. Atten-
tion was next turned to the reaction of mesylate 41 in
DMSO-d6. Our previous studies have shown that solvoly-
ses of this substrate in polar alcohol solvents proceed via
carbocationic intermediates.24 We wanted to evaluate the
potential for carbocation formation in DMSO-d6. Also of
interest is the relationship between our observed car-
bocation-forming reactions in DMSO-d6 and the Korn-
blum oxidation. This classic oxidation involves reaction
on benzylic chlorides with DMSO and presumably in-
volves nucleophilic displacement of halide with DMSO,
(18) Rates of reaction of 16 and 22 are extremely concentration
dependent due to the common ion effect. Chloride 16 (0.0015 M) is
50% reacted in 70 min at 25 °C.
(19) Creary, X.; Wolf, A. J . Phys Org. Chem. 2000, 13, 337.
(20) Protonation of norbornenes from the exo-face is well established.
Microscopic reversibility requires that exo-deprotonation of norbornyl
cations be much faster than endo-deprotonation. See: (a) Kropp, P. J .
J . Am. Chem. Soc. 1973, 95, 4611. (b) Stille, J . K.; Hughes, R. D. J .
Org. Chem. 1971, 36, 340. (c) Brown, H. C.; Liu, K.-T. J . Am. Chem.
Soc. 1975, 97, 2469. (d) Brown, H. C.; Kawakami, J . H. J . Am. Chem.
Soc. 1975, 97, 5521.
(23) (a) Ingold, C. K. In Structure and Mechanism in Organic
Chemistry, 2nd ed.; Cornell University Press: Ithica, New York, 1969;
p 955. (b) J ansen, M. P.; Koshy, K. M.; Mangru, N. N.; Tidwell, T. T.
J . Am. Chem. Soc. 1981, 103, 3863. (c) Creary, X.; Geiger, C. C.; Hilton,
K. J . Am. Chem. Soc. 1983, 105, 2851.
(21) Coke, J . L.; Cooke, M. P., J r. J . Am. Chem. Soc. 1967, 89, 6701.
(22) Fisher, R. D.; Seib, R. C.; Shiner, V. J ., J r.; Szele, I.; Tomic,
M.; Sunko, D. E. J . Am. Chem. Soc. 1975, 97, 2408.
(24) Creary, X.; Geiger, C. C. J . Am. Chem. Soc. 1982, 104, 4151.
1122 J . Org. Chem., Vol. 68, No. 3, 2003

Carbocation-Forming Reactions in Dimethyl Sulfoxide
followed by â-elimination of dimethyl sulfide from the
oxosulfonium intermediate. Mesylate 41 reacts readily
in DMSO-d6 at 70 °C to give oxidation product 43 (83%)
as well as the substitution product, methyl mandelate,
44 (17%). The starting mesylate disappears in a clean
pseudo first-order process and, contrary to the previously
discussed substrates, oxosulfonium intermediate 42,
which builds up to a maximum concentration of 17%, can
be observed by 1H NMR.25 Unlike the reactions of
substrates 4, 7-11, 15, and 17, intermediate 42 is
probably formed not by a carbocationic process but by
an SN2 process. This intermediate partitions readily
between elimination of dimethyl sulfide and alcohol
formation. The pseudo first-order rate constants shown
can be extracted by fitting the NMR data to a kinetic
scheme involving consecutive pseudo first-order pro-
cesses. Of interest is the reaction of 42 with the water in
the DMSO-d6 solvent, which leads to 44. Adding 0.5%
water to the DMSO-d6 increased the amount of 44 formed
to 38%. The rate of reaction of 42 with water is many
orders of magnitude slower than the reaction of inter-
mediate 13 with water. While intermediate 42 lives long
enough to allow spectroscopic identification, no trace of
analogous intermediate 13 (or intermediates derived from
7-11, 15, 17, or 28) can be detected. This is presumably
due to the high rate of reaction of 13 with water. The
origin of this large rate difference for reactions of 42 and
13 with water is unclear.
F IGURE 5. Conversion of PhCH2OMs 45 to 46 and 47 in
DMSO-d6 at 60 °C.
mesylate with DMSO-d6. In other words, DMSO-d6 can
be displaced from 46 by solvent DMSO-d6, where the
sulfur is now the nucleophilic end of the solvent. The
formations of both 46 and 47 are probably SN2 processes.
When Et3N is added to the DMSO-d6 at the beginning
of the reaction, no buildup of 46 is observed and the
observed product of the reaction is then benzaldehyde,
48. If there is no Et3N present and larger amounts of
water (1-2%) are added, then 46 can be intercepted and
converted to benzyl alcohol, 49, in competition with
conversion to 47. Thus, in DMSO-d6 containing 2% H2O,
46 partitions to give comparable amounts of 47 and 49.
However, the reactivity of 46 with water is many orders
of magnitude lower than that of 13 with water. While
46 can be easily observed in the NMR even with 2% water
present, oxosulfonium intermediate 13 cannot be detected
even when water is present to the extent of only 0.04%.
We have now prepared benzyl mesylate, 45, and found
that the products formed when this substrate reacts with
DMSO-d6 are highly dependent on the reaction condi-
tions. When benzyl mesylate is dissolved in DMSO-d6
containing no added base, oxosulfonium intermediate 46
is initially formed (k ) 6.2 × 10-5 s-1). This intermediate
does not react readily with trace amounts (0.04%) of
water present, and this allows 46 to be characterized in
situ by 1H and 13C NMR spectroscopy. Under the reaction
conditions, this intermediate converts slowly to sulfonium
salt 47 (k ) 6.7 × 10-6 s-1). Figure 5 shows the buildup
of intermediate 46 as a function of time as well as the
slower conversion of 46 to the final product 47. This
figure shows that the initial rate of formation of 47 is
quite slow, but the rate of formation of 47 increases as
intermediate 46 builds up. This kinetic behavior is
consistent with a mechanism where 47 is formed via
intermediate 46 and not by direct reaction of benzyl
1-Ad a m a n tyl Mesyla te: A kC Su bstr a te. 1-Ada-
mantyl and 2-adamantyl systems have become standards
for evaluation of reactions proceeding via kC (carbocat-
(25) An oxosulfonium intermediate [Ph2S-O-sugar]+ has been
spectroscopically characterized at low temperatures. See: Garcia, B.
A.; Gin, D. Y. J . Am. Chem. Soc. 2000, 122, 4269.
J . Org. Chem, Vol. 68, No. 3, 2003 1123

Creary et al.
ionic) mechanisms.26 These systems react in solvolysis
reactions without backside assistance due to nucleophilic
solvent properties. A series of solvent ionizing power
values (Y values) have been developed on the basis of
adamantyl systems. It was therefore of interest to deter-
mine the reactivity of 1-adamantyl mesylate in DMSO-d6.
DMSO-d6 is of little value in predicting the reactivity of
triflate 54 in DMSO-d6. The products of reaction of 54 in
DMSO-d6 are also of interest. The major product (63%)
is the rearranged elimination product 55, and this attests
to the cationic nature (k∆) of the reaction in DMSO-d6.
Also formed is a smaller amount of oxosulfonium salt 56,
which is derived from solvent capture of the rearranged
cation 57. Of note is the fact that 56 can be easily
detected under the reaction conditions and none of the
corresponding rearranged alcohol product is observed.
This represents a rare example where DMSO capture
successfully competes with â-proton loss from the cationic
intermediate.
1-Adamantyl mesylate, 50, reacts in DMSO-d6 at 50
°C to give oxosulfonium salt 51, which is relatively
unreactive under the reaction conditions. The activation
parameters calculated from the data in Table 1 (∆Hq )
24.4 kcal/mol; ∆Sq ) -5.8 eu) are unexceptional. The
extrapolated rate constant for disappearance of 50 (4.21
× 10-7 s-1 at 25 °C DMSO-d6) is 104 times slower than
reaction of 50 in 80% aqueous ethanol,27 and this corre-
sponds to a YOMs value of -4.00 for DMSO-d6. The rate
of reaction of mesylates 4, 7, 10, and 11 in DMSO-d6 are
all somewhat faster than predicted on the basis of this
YOMs value for DMSO-d6 derived from 50. It is suggested
that 50, which is quite reactive in protic solvents, is
relatively unreactive in DMSO-d6 due to the inability of
DMSO-d6 to solvate the developing mesylate anion as
effectively as hydrogen-bonding solvents.
The fact that oxosulfonium salt 51 does not react
rapidly with water in DMSO-d6 at 50 °C allows it to be
characterized in situ by 1H as well as 13C NMR. However,
46 is slowly converted to 1-adamantyl alcohol, 52, at 100
°C, presumably by reaction with the trace of water
present. Thus, adding 1% water to the DMSO-d6 and
heating at 100 °C for 4 h gives complete conversion of
51 to 52.
Tr ifla te Der iva tive of P iva loin : A k∆ Su bstr a te.
Reaction of the triflate derived from pivaloin, 54, is of
special interest due to the high reactivity of this substrate
in DMSO-d6. Triflate 54 disappears in a clean first-order
process with a half-life of only 15 min at 25 °C in this
solvent. Previous solvolysis studies on 54 have been
carried out in solvents ranging from the relatively “poorly
ionizing” solvents ethanol and acetic acid (half-life of 54
) 46 h in HOAc) to the “highly ionizing” solvents formic
acid, hexafluoroisopropyl alcohol, and trifluoroacetic acid
(half-life of 54 ) 58 min in CF3CO2H).28 Triflate 54
solvolyzes in DMSO-d6 at a rate that exceeds rates in all
of these solvents. Hence the YOMs value of -4.00 for
Com p u ta tion a l Stu d ies. To shed further light on the
reaction of substrates with DMSO and water, computa-
tional studies were carried out at the B3LYP/6-31G*
level.29 Shown is the geometry-optimized structure of the
adduct [Ph-CH2-O-S(CH3)2]+, 46-H6, derived from O-
benzylation of DMSO. Covalent product 58 potentially
formed by reaction of water with 46-H6 could not be
located as an energy minimum at any computational level
attempted. Starting with a geometry where water is
within covalent bonding distance of sulfur, attempted
geometry optimization led to the energy minimum struc-
ture 59 shown. The H2O-sulfur distance is quite long
(3.250 Å) and indicative of an ion/dipole attractive inter-
action. The covalent form represented by 58 contains a
hypervalent sulfur, and this computational study casts
doubt on the involvement of such species in reactions of
13 and 27 (and other oxosulfonium intermediates) with
water.
To probe for subtle structural differences that might
lead to facile reaction with water, computational studies
were also carried out at the B3LYP/6-31G* level on
unlabeled oxosulfonium ion 13-H6, as well as on oxosul-
(29) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.7; Gaussian, Inc.: Pittsburgh, PA,
1998.
(26) For reviews, see: (a) Bentley, T. W.; Schleyer, P. v. R. Adv.
Phys. Org. Chem. 1977, 14, 1. (b) Kamlet, M. J .; Abboud, J . L. M.;
Taft, R. Prog. Phys. Org. Chem. 1981, 13, 485.
(27) Bentley, T. W.; Carter, G. E. J . Org. Chem. 1983, 48, 579.
(28) Creary, X.; McDonald, S. R. J . Org. Chem. 1985, 50, 474.
1124 J . Org. Chem., Vol. 68, No. 3, 2003

Carbocation-Forming Reactions in Dimethyl Sulfoxide
fonium ion intermediate 62 derived from mesylate 15.
Optimized structures are shown. As before, the compu-
tational studies indicate that covalent water adducts 60
and 63 are not energy minima because water is not
within covalent bonding distance of sulfur in minimized
forms 61 and 64. How then do oxosulfonium ions react
with water? The obvious alternative is a concerted
pathway where nucleophilic attack of water on sulfur
occurs with simultaneous sulfur-oxygen bond breaking.
An additional water molecule may well be hydrogen-
bonded to oxygen in the departing alcohol to avoid
formation of anionic RO-. However, these calculations
do not offer an obvious reason as to why certain oxosul-
fonium intermediates such as 13 and 27 react so much
more readily with water than oxosulfonium ions 42, 46,
51, and 56.
Con clu sion
Certain mesylates, triflates, trifluoroacetates, and
chlorides react at reasonable rates in DMSO-d6 to give
J . Org. Chem, Vol. 68, No. 3, 2003 1125

Creary et al.
1 in 2.31 g of DMSO-d6 containing 11.2 mg of H217O (20.1 atom
%
products derived from carbocationic intermediates. These
carbocations can suffer proton loss leading to elimination
products or they can be captured by DMSO-d6 to give
oxosulfonium intermediates. These oxosulfonium inter-
mediates then undergo rapid reaction with trace amounts
of water present in the DMSO-d6 to give alcohol products.
Reaction in the presence of H217O leads to 17O incorpora-
tion into the DMSO-d6 and no 17O in the alcohol product.
Reaction of the O-methyl oxime derivative of R-chloro-
R,R-diphenylacetophenone in DMSO-d6 gives an indole
as the major product, and this finding supports the
involvement of a delocalized R-oximino carbocation. Sol-
volysis of 1-adamantyl mesylate in DMSO-d6 allows
determination of a YOMs value (-4.00) for DMSO-d6.
Capture of the 1-adamantyl cation by DMSO-d6 gives a
long-lived oxosulfonium intermediate which can be char-
17O) and 28 mg of Et3N. This solution was sealed in two
NMR tubes, and the tubes were immersed in a water bath at
60.0 °C. One of the tubes was periodically withdrawn from the
water bath and analyzed by 17O NMR spectroscopy. The
evolving NMR spectra are shown in Figure 4. The signal at δ
0.0 is due to the H217O, while the signal at δ 11.5, which
increases with time, is due to (CD3)2S17O. After 22 h at 60 °C,
the contents of the NMR tubes were added to water and
extracted into ether. The ether extract was washed with water
and saturated NaCl solution and dried over MgSO4. After
solvent removal using a rotary evaporator, the residue was
analyzed by 1H NMR, which showed 6 as the sole product. 17
NMR spectroscopy of 6 showed no signals.
O
P r ep a r a tion of a n Au th en tic Sa m p le of 6-17O. The
preparation of 6-17O was analogous to the preparation of
unlabeled 6 from acetophenone as shown below. Thionyl
chloride (2.20 g) was added to 410 mg of acetophenone-17O (20
at. % 17O)30 containing 5.6 mg of pyridine.
1
acterized by H and 13C NMR. This intermediate reacts
slowly with residual water in the solvent at 100 °C.
Hence, oxosulfonium intermediates possess remarkably
differing reactivities toward water. In some cases, they
cannot be detected at room temperature, while in other
cases they survive for significant periods of time at 100
°C.
Exp er im en ta l Section
The mixture was stirred for 2.25 h at room temperature, and
the excess SOCl2 was removed using a rotary evaporator. The
crude residue was dissolved in 5 mL of ether, and the ether
solution was added dropwise to a stirred solution of 2.0 g HN-
(CH3)2 in 10 mL of ether at -78 °C. The mixture was slowly
warmed to room temperature with stirring and then taken up
into ether and water. The ether extract was washed with water
and saturated NaCl solution and dried over MgSO4. The ether
solvent was removed using a rotary evaporator, and 537 mg
(81%) of thioamide i was obtained as a yellow solid. The 17O
NMR spectrum of i (CDCl3) showed a signal at δ 497.4.
A solution of 489 mg of thioamide i in 3 mL of THF was
added dropwise to a solution of lithium diisopropyl amide
(LDA) prepared from 329 mg of diisopropylamine and 1.90 mL
of 1.6 M n-BuLi in 3 mL of THF at -78 °C. The mixture was
slowly warmed to room temperature, and an aqueous workup
followed with ether extraction. The ether extract was dried
over MgSO4 and filtered, and the solvent was removed using
a rotary evaporator. The crude product was chromatographed
on 5 g of silica gel, and 236 mg (48%) of 6-17O eluted with 30%
ether in hexanes. The 17O NMR spectrum of 6-17O (DMSO-d6)
showed a signal at δ 39.1.
Rea ction of (S)-(+)-Mesyla te 4 in DMSO-d 6. (S)-(+)-
Mesylate 4 (15 mg) (prepared from optically pure (S)-(+)-6)7
was dissolved in 1.9 mL of DMSO-d6 containing 10.8 mg of
Et3N. The solution was sealed in a tube and heated at 60.0 °C
for 17 h and then at 64 °C for 5.5 h. After a standard aqueous
workup, the crude alcohol product was passed through a small
amount of silica gel in a pipet and then analyzed by 600 MHz
1H NMR. The ratio of the enantiomers present was determined
by addition of the chiral shift reagent Eu(hfc)3. The ratio of
(R)- to (S)- enantiomers was 97.1:2.9 as determined by
integration of the o-hydrogens of the phenyl group, which
appeared at δ 7.86 for (R)-6 and δ 7.99 for (S)-6. This
corresponds to 94.2% inversion and 5.8% racemization.
P r ep a r a tion of Ch lor id e 22. A mixture of 1.78 g of
R-hydroxy-R,R-diphenylacetophenone31 and 0.670 g of meth-
oxylamine hydrochloride in 5 mL of pyridine was sealed in a
tube and heated at 120 °C for 48 h. The contents of the tube
were then added to 30 mL of water and extracted into ether.
The ether extract was washed with dilute HCl solution,
saturated NaCl solution, and dried over MgSO4. After solvent
Rea ction of Mesyla te 4 w ith Na N3 in DMSO-d 6. Kin etic
Stu d ies. Mesylate 47 (2.4 mg) was dissolved in 2.0 mL of 0.1
M NaN3 in DMSO-d6, and a portion of this solution was sealed
in an NMR tube. The tube was immersed in a water bath at
50.0 °C for a given amount of time. The tube was then cooled
to room temperature and analyzed by 600 MHz 1H NMR. The
amounts of unreacted 4 and products 5 and 6 were determined
by integration of signals at δ 4.63, 4.27, and 4.11. Figure 1
shows some representative spectra as a function of time at
50.0 °C. Pseudo first-order rate constants for disappearance
of 4 were calculated by standard least squares procedures.
Correlation coefficients were all greater than 0.9999. Kinetic
studies in 0.20 and 0.30 M NaN3 were carried out in a
completely analogous fashion.
R ea ct ion of Mesyla t es 4, 7, 8, 10, 11, 15, a n d 17 in
DMSO-d 6. Kin etic Stu d ies. Rates of reaction of these me-
sylates7,12 were determined by 1H NMR using two methods.
Method 1: approximately 5 mg of the appropriate mesylate
was dissolved in 1 mL of 0.06 M Et3N in DMSO-d6, and the
sample was sealed in an NMR tube. The tube was placed in a
constant-temperature bath at the appropriate temperature or
in the probe of the NMR at 25.0 °C. At appropriate time
1
intervals, the tube was analyzed by H NMR, and the shift of
the upfield triplet due to the Et3N was determined. The shift
of the Et3N, which moves downfield as a function of time, was
monitored. After 10 half-lives, a final reading was taken. First-
order rate constants for disappearance of mesylates were
calculated by standard least squares procedures. Correlation
coefficients were all greater than 0.9999. Rates of 4, 7, 9, 15,
and 17 were measured by this method.
Method 2: approximately 5 mg of the appropriate mesylate
was dissolved in 1 mL of DMSO-d6, and a portion of the
solution was sealed in an NMR tube. The tube was placed in
a constant-temperature bath at the appropriate temperature.
At appropriate time intervals, the tube was analyzed by 600
1
MHz H NMR and areas due to starting mesylate and alcohol
product were measured. First-order rate constants for disap-
pearance of mesylates were calculated by standard least
squares procedures. Correlation coefficients were all greater
than 0.9999. Rates of 8, 10, and 11 were measured by this
method.
Rea ction of Mesyla te 4 in DMSO-d 6 w ith 0.5% H217O.
A solution was prepared by dissolving of 40.3 mg of mesylate
(30) Creary, X.; Inocencio, P. A. J . Am. Chem. Soc. 1986, 108, 5979.
(31) Greene, J . L.; Zook, H. D. J . Am. Chem. Soc. 1958, 80, 3629.
1126 J . Org. Chem., Vol. 68, No. 3, 2003

Carbocation-Forming Reactions in Dimethyl Sulfoxide
removal using a rotary evaporator, the residue was chromato-
graphed on silica gel and the column was eluted using 10%
ether in hexanes. The (Z)-isomer of the O-methyloxime eluted
first (250 mg) followed by 1.612 mg of the pure (E)-O-
methyloxime 24. 1H NMR of 24 (CDCl3): δ 7.40 (m, 4 H), 7.29
(m, 6 H), 7.24 (t, J ) 7.4 Hz, 1 H), 7.19 (t, J ) 7.4 Hz, 2 H),
6.82 (m, 2 H), 4.23 (s, 1 H), 3.86 (s, 3 H). 13C NMR (CDCl3): δ
160.2, 143.0, 132.1, 128.63, 128.60, 128.3, 127.85, 127.81,
127.7, 81.5, 62.6.
mmol) in 4 mL of pyridine was stirred at 60 °C for 10 h. The
mixture was taken up into ether and extracted with water,
2% hydrochloric acid, and saturated NaCl solution. The ether
extract was dried over MgSO4 and filtered, and the solvent
was removed using a rotary evaporator to give 160 mg (71%
yield) of O-methyloxime derivative 29. 1H NMR of 29
(CDCl3): δ 7.59-7.53 (m, 2 H), 7.46-7.39 (m, 2 H), 7.33-7.26
(m, 4 H), 7.03 (t of t, J ) 7.5, 1.2 Hz, 1 H), 6.93 (t of m, J )
7.5, 2 H), 6.44 (d of m, J ) 8.2 Hz, 2 H), 5.18 (br, 1 H), 3.99 (s,
3 H). 13C NMR of 29 (CDCl3): δ 157.7, 145.7, 140.8, 130.6,
129.4, 128.2, 127.9, 127.6, 127.2, 124.5, 120.0, 83.8, 62.8. Anal.
Calcd for C21H17NO2: C, 79.98; H, 5.43. Found: C, 79.77; H,
5.47.
A mixture of 280 mg of alcohol 24, 157 mg of SOCl2, and
468 mg of Na2CO3 in 3 mL of ether was stirred at room
temperature for 3 days. The mixture was filtered, and the
solvent was removed using a rotary evaporator. The NMR of
the crude solid showed 6% of the (Z)-chloride along with 94%
of 22. The solid was slurried with ether, and the ether was
decanted. The solid was collected using a Bu¨chner funnel,
washed with a small amount of ether, and dried under
A mixture of 108 mg of O-methyloxime derivative 29
prepared above and 100 mg of Na2CO3 in 3 mL of ether was
stirred as 133 mg of thionyl chloride in 0.5 mL of ether was
added. After the mixture was stirred for 10 h at room
temperature, the Na2CO3 was removed by filtration and the
solvent was removed using a rotary evaporator to give 93 mg
1
vacuum. H NMR of 22 (CDCl3): δ 7.50 (m, 4 H), 7.39-7.32
(m, 6 H), 7.30-7.22 (m, 3 H), 6.95 (m, 2 H), 3.79 (s, 3 H). 13C
NMR of 22 (CDCl3): δ 159.5, 141.7, 133.1, 129.2, 128.6, 128.4,
1
(81% yield) of chloride 28, mp 81-82 °C. H NMR (CDCl3): δ
128.1, 127.7, 127.5, 79.7, 62.7. Exact mass (EI) calcd for C21H18
NOCl 335.1077, found 335.1059.
-
7.52 (d of m, J ) 7.6 Hz, 2 H), 7.46 (d of m, J ) 7.5 Hz, 2 H),
7.32 (t of d, J ) 7.8, 1.5 Hz, 2 H), 7.24 (t of d, J ) 7.2, 1.2 Hz,
2 H), 7.19-7.09 (m, 3 H), 6.80 (d of m, J ) 8.1, 2H), 3.89 (s, 3
H). 13C NMR (CDCl3): δ 155.6, 144.8, 139.4, 131.7, 129.5,
128.6, 128.4, 127.8, 127.5, 136.7, 120.0, 72.1, 62.6. Exact mass
(EI) calcd for C21H16NOCl 333.0920, found 333.0927.
P r ep a r a tion of In d ole 23. A solution of 237 mg of alcohol
24 in 5 mL of CDCl3 was stirred as 58 mg of CF3CO2H was
added. After 4 days at room temperature, the solvent was
removed using a rotary evaporator and the residue was passed
through a short silica gel column with ether elution. Solid
indole 23 (224 mg; 95% yield) was collected, mp 134-135 °C.
1H NMR of 23 (CDCl3): δ 7.72 (d of t, J ) 8.0, 0.9 Hz, 1 H),
7.52 (m, 3 H), 7.38-7.29 (m, 8 H), 7.24 (t, J ) 7.2 Hz, 1 H),
7.18 (t of d, J ) 8.0 Hz, 1.0 H), 3.694 (s, 3 H). 13C NMR of 23
(CDCl3): δ 134.5, 133.3, 132.8, 130.4, 130.1, 129.7, 128.4, 128.3,
128.0, 126.1, 123.8, 123.0, 121.0, 119.8, 112.5, 108.9, 64.2.
Exact mass (EI) calcd for C21H17NO 299.1310, found 299.1311.
Anal. Calcd for C21H17NO: C, 84.25; H, 5.72; N, 4.68. Found:
C, 84.69; H, 5.94; N, 4.48.
P r ep a r a tion of Ch lor id e 28. A solution containing 2.271
g of 9-cyano-9-trimethylsiloxyfluorene (8.13 mmol)32 in 5 mL
of ether was stirred while 9.8 mL of 1.0 M phenylmagnesium
bromide (9.8 mmol) in ether was added dropwise. After 15 min
at room temperature, the reaction was quenched with 10 mL
of saturated NH4Cl solution. The mixture was taken up into
ether, and the ether extract was washed with water and
saturated NaCl solution. The ether phase was dried over
MgSO4 and filtered, and the solvent was removed using a
rotary evaporator. The crude residue was dissolved in 16 mL
of tetrahydrofuran, and 4 mL of water was added followed by
0.716 g of H2SO4. The mixture was stirred at room temperature
for 7 h and then taken up into ether. The ether extract was
washed with saturated NaHCO3 solution and saturated NaCl
solution and then dried over MgSO4. Filtration and removal
of solvent by rotary evaporator gave 1.753 g of an oil, which
was purified by chromatography on 30 g of silica gel. The
column was eluted with increasing amounts of ether in
hexanes. 9-Benzoyl-9-hydroxyfluorene (1.00 g; 57% yield), mp
132-134 °C, eluted with 6-15% ether in hexanes. 1H NMR
(CDCl3): δ 7.74 (d of m, J ) 7.5 Hz, 2 H), 7.41 (t of d, J ) 7.2,
1.2 Hz, 2 H), 7.35 (d of m, J ) 8.1, 2 H), 7.31-7.20 (m, 5 H),
7.07 (t of m, J ) 7.5, 2 H), 5.67 (br, 1H). 13C NMR (CDCl3): δ
199.7, 146.0, 141.2, 133.3, 133.0, 129.9, 129.2, 128.7, 128.2,
124.5, 120.8, 86.52.
P r ep a r a tion of Tr iflu or oa ceta tes 35, 35-d 3, a n d 37.
Gen er a l P r oced u r e. A solution of the appropriate alcohol (1.0
equiv) and 2,6-lutidine (1.5 equiv) in ether was cooled to -10
°C, and 1.3 equiv of trifluoroacetic anhydride was added
dropwise to the stirred solution. The mixture was warmed to
0 °C, and after 5 min, water was added. The mixture was
rapidly transferred to a separatory funnel using ether, and
the ether extract was washed with dilute HCl solution,
NaHCO3 solution, and saturated NaCl solution and dried over
MgSO4. After filtration, the ether solvent was removed using
a rotary evaporator to give the crude trifluoroacetates that
were used without further purification. The trifluoroacetates
decompose upon prolonged standing at room temperature and
were therefore used immediately after solvent removal. The
following procedure is representative.
Reaction of 97 mg of 2-methyl-2-adamantanol22 and 121 mg
of 2,6-lutidine in 2 mL of ether with 161 mg of trifluoroacetic
anhydride gave of 2-methyl-2-adamantyl trifluoroacetate, 35.
1H NMR (CDCl3); δ 2.36 (m, 2 H), 1.99 (br d, 2 H), 1.93-1.78
(m, 6 H), 1.74 (m, 2 H), 1.707 (s, 3 H), 1.63 (br d, 2 H). 13C
NMR (CDCl3); δ 155.9 (q, J ) 41 Hz), 114.6 (q, J ) 287 Hz),
94.0, 37.9, 36.0, 34.6, 32.8, 27.1, 26.4, 22.0.
Com p u ta tion a l Stu d ies. Ab initio molecular orbital cal-
culations were performed using the Gaussian 98 series of
programs.29 Structures were characterized as minima via
frequency calculations that showed no negative frequencies.
Ack n ow led gm en t. Acknowledgment is made to the
National Science Foundation for partial support of this
research.
Su p p or tin g In for m a tion Ava ila ble: B3LYP calculated
structures, energies, and Cartesian coordinates of 46-H6, 59,
1
13-H6, 61, 62, and 64, H and 13C NMR spectra of compounds
22, 24, 28, and 35, as well as experimental procedures for
reactions of 22, 28, 30, 31, 35, 37, 39, 41, 45, 50, and 54. This
material is available free of charge via the Internet at
http://pubs.acs.org.
A solution of 205 mg of 9-benzoyl-9-hydroxyfluorene (0.715
mmol) and 71.8 mg of methoxylamine hydrochloride (0.857
(32) Gassman, P. G.; Talley, J . J . Organic Syntheses; Wiley: New
York, 1990; Collect. Vol. VII, p 20 .
J O026468X
J . Org. Chem, Vol. 68, No. 3, 2003 1127