Rearrangements of 3-aryl-substituted cyclopropenyl anions and the gas-phase acidity of 3-(4-Methylphenyl)cyclopropene

Article abstract of DOI:10.1021/jo001102a

3-(4-Methylphenyl)-3-trimethylsilylcyclopropene and 3-(4-trifluoromethylphenyl)-3-trimethylsilylcyclopropene react with fluoride ion in the gas phase to afford 6-substituted 3-indenyl anions via a spontaneous rearrangement of their corresponding cyclopropenyl anions. These isomerizations led us to reinvestigate the reported gas-phase generation of 1,2,3-triphenylcyclopropenyl anion, and contrary to the previous study, a similar rearrangement to 1,2-diphenyl-1-indenyl anion is observed. Despite the instability of 3-aryl-3-cyclopropenyl anions, we were able to measure the acidity of 3-(4-methylphenyl)cyclopropene at the allylic position (δH°_acid = 398.6 ± 1.4 kcal/mol) by the DePuy kinetic method. Ab initio calculations on the structures and energies of mono- and triaryl-substituted cyclopropenyl anions also are presented.

Full text of DOI:10.1021/jo001102a

J . Org. Chem. 2001, 66, 99-106
99
Rea r r a n gem en ts of 3-Ar yl-Su bstitu ted Cyclop r op en yl An ion s a n d
th e Ga s-P h a se Acid ity of 3-(4-Meth ylp h en yl)cyclop r op en e
Katherine M. Broadus, Sangdon Han,† and Steven R. Kass*
Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455
kass@chem.umn.edu
Received J uly 20, 2000
3-(4-Methylphenyl)-3-trimethylsilylcyclopropene and 3-(4-trifluoromethylphenyl)-3-trimethylsilyl-
cyclopropene react with fluoride ion in the gas phase to afford 6-substituted 3-indenyl anions via
a spontaneous rearrangement of their corresponding cyclopropenyl anions. These isomerizations
led us to reinvestigate the reported gas-phase generation of 1,2,3-triphenylcyclopropenyl anion,
and contrary to the previous study, a similar rearrangement to 1,2-diphenyl-1-indenyl anion is
observed. Despite the instability of 3-aryl-3-cyclopropenyl anions, we were able to measure the
acidity of 3-(4-methylphenyl)cyclopropene at the allylic position (∆H°acid ) 398.6 ( 1.4 kcal/mol)
by the DePuy kinetic method. Ab initio calculations on the structures and energies of mono- and
triaryl-substituted cyclopropenyl anions also are presented.
In tr od u ction
to the condensed phase and observed the first long-lived
cyclopropenyl anion, 3-methoxycarbonyl-1,2-diphenylcyclo-
propenyl anion, by UV spectroscopy.12
Antiaromatic molecules can be described as fully
conjugated 4n (n ) 1, 2, 3..) π electron monocyclic
compounds which are destabilized relative to their acyclic
counterparts. They also are characterized by paramag-
netic ring currents in their 1H NMR spectra and the
potential for unusually small singlet-triplet gaps.1,2
Cyclopropenyl anion is the smallest such system, and
consequently, numerous studies have been carried out
over the years on this ion and its derivatives. Breslow
and co-workers showed that the allylic position of cyclo-
propenes is very weakly acidic with some pKA’s in excess
of 50 while Borden and co-workers successfully trapped
1,2,3-triphenylcyclopropenyl anion as a transient inter-
mediate.3-9 In 1994, Sachs and Kass prepared the first
stable cyclopropenyl anion, 1, by reacting fluoride ion
with 3-methoxycarbonyl-3-trimethylsilylcyclopropene in
the gas phase (eq 1).10 The proton affinity of this anion
was compared to its saturated derivative (eq 2), and the
observed difference of 14 ( 6 kcal/mol is consistent with
the notion of antiaromatic destabilization. 3-Cyano-3-
cyclopropenyl anion also has been generated and shown
to be 15 ( 6 kcal/mol less stable than 1-cyanocyclopropyl
anion.11 More recently, we extended these experiments
In addition to generating highly stabilized cycloprope-
nyl anions in the gas phase, we sought to explore
derivatives which are more basic and would better reflect
the thermochemistry of the parent. High-level ab initio
calculations indicate that cyclopropenyl and 3-methyl-3-
cyclopropenyl anions are unstable with regard to electron
loss (i.e., the anions are unbound).13,14 Other simple alkyl
derivatives also are apt to be unstable, so we decided to
examine aryl-substituted cyclopropenes. Herein, we re-
port the results of this study along with computed
structures and energies of the anions. On the basis of
our findings, we also reinvestigated the reported gas-
phase generation of 1,2,3-triphenylcyclopropenyl anion.15
† Current address: Department of Chemistry, University of
CaliforniasBerkeley, Berkeley, CA 94720.
Exp er im en ta l Section
(1) Minkin, V. I.; Glukhovtsev, M. N.; Simkin, B. Y. In Aromaticity
and Antiaromaticity: Electronic and Structural Aspects; Wiley-Inter-
science: New York, 1994.
(2) Garratt, P. J . In Aromaticity; J ohn Wiley and Sons: New York,
1986.
(3) Breslow, R. Pure Appl. Chem. 1982, 54, 927-938.
(4) Wasielewski, M. R.; Breslow, R. J . Am. Chem. Soc. 1976, 98,
4222-4229.
(5) Breslow, R. Acc. Chem. Res. 1973, 6, 393-398.
(6) Breslow, R. Angew. Chem., Int. Ed. Engl. 1968, 7, 565-570.
(7) Breslow, R.; Brown, J .; Gajewski, J . J . J . Am. Chem. Soc. 1967,
89, 4383-4390.
(8) Breslow, R. Chem. Eng. News 1965, 43, 90-99.
(9) Koser, H. G.; Renzoni, G. E.; Borden, W. T. J . Am. Chem. Soc.
1983, 105, 6359-6360.
(10) Sachs, R. K.; Kass, S. R. J . Am. Chem. Soc. 1994, 116, 783-
784.
3-Aryl-3-trimethylsilylcyclopropenes and their correspond-
ing indene isomers were synthesized according to previously
reported procedures.16 These compounds were purified by
preparative gas chromatography on a 12 ft × 0.25 in. 10% SE
30 on Chrom W column at 150 °C except for 3-(4-methylphen-
(12) Arrowood, T. L.; Kass, S. R. J . Am. Chem. Soc. 1999, 121, 7272-
7273.
(13) Merrill, G. N.; Kass, S. R. J . Am. Chem. Soc. 1997, 119, 12322-
12337.
(14) Han, S.; Hare, M. C.; Kass, S. R. Int. J . Mass Spectrom. 2000,
201, 101-108.
(15) Bartmess, J . E.; Kester, J .; Borden, W. T.; Koser, H. G.
Tetrahedron Lett. 1986, 27, 5931-5934.
(16) Han, S.; Kass, S. R. J . Chem. Soc., Perkin Trans. 1 1999, 1553-
1558.
(11) Arrowood, T. L. Ph.D. Thesis, University of Minnesota, 1999.
10.1021/jo001102a CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/07/2000

100 J . Org. Chem., Vol. 66, No. 1, 2001
Broadus et al.
Ta ble 1. Su m m a r y of th e Rea ctivity of th e [2-TMS]-,
[2-TMS]--d 1 Ion s a n d In d ep en d en tly P r ep a r ed 4 w ith
Refer en ce Acid s
yl)-3-trimethylsilylcyclopropene. This derivative was initially
purified by column chromatography; however, it was necessary
to further separate it from traces of coeluting trimethylsilyl
impurities which would have affected the ratio in the kinetic
acidity measurement. To this end, 3-(4-methylphenyl)-3-tri-
methylsilylcyclopropene was treated with 1 equiv of tetra-n-
butylammonium fluoride in THF at room temperature for 1 h
and then subsequently purified by medium-pressure liquid
chromatography using hexanes as the eluting solvent. 3-Tri-
methylsilyl-1,2,3-triphenylcyclopropene9 and 2,3-diphenylin-
dene17 were prepared as in the literature. Reference silanes
for the kinetic acidity measurement were prepared by standard
methods and purified by preparative gas chromatography
using a 10% SE 30 on Chrom W column. In the case of
trimethylsilylcyclopropane, a XF-1150 column was employed.
The gas-phase experiments were carried out in a dual cell
model 2001 Finnigan Fourier transform mass spectrometer
(FTMS) equipped with a 3.0 T superconducting magnet and
controlled by a Sun workstation running the Odyssey version
4.2 software package. Ions were generated by fluoride-induced
desilylation of the requisite neutral. Fluoride ion was prepared
upon electron ionization (6 eV) of carbon tetrafluoride. The ions
of interest were then transferred to the second cell and isolated
by applying a SWIFT waveform or chirp broad band excita-
tion.18,19 Argon was pulsed into the cell at pressures of 10-5
Torr in an attempt to thermalize the ions. All neutral reagents
were introduced via slow leak valves, and the subsequent
reactions were monitored over time. For the kinetic acidity
measurements, hydroxide was prepared in the FTMS upon
electron ionization of H2O or H218O at 6 eV and transferred to
the other cell where the requisite silane was present at a
constant pressure. After a cooling pulse of argon and an initial
reaction period (∼500 ms), all product ions were ejected in
order to reduce the effects of nonthermalized hydroxide. The
ratio of product ions (Me3SiO-/R(Me)2SiO-) was recorded as a
function of time (1 to 5 s) and found to be constant. Ion
intensities were isotopically corrected and an error analysis
which accounted for the uncertainties in the calibration data
was carried out.20,21
acid
∆H°acid (kcal/mol)a [2-TMS]- [2-TMS]--d1
4
H/D Exchangeb
D2O
CH3OD
(CH3)3COD
CF3CH2OD
CD3CO2D
(CH3)3CO2D
392.9 ( 0.1
383.5 ( 0.7
374.6 ( 2.1
361.8 ( 2.5
348.6 ( 2.9
344.7 ( 2.1
2c
1
-
-
-
0
2c
1
-
1
2
2
1
1
2
2
-1d
-1d
Proton Transfer
C2H5OCH2CO2H
HCl
342.0 ( 2.0
333.4 ( 0.1
no
yes
no
yes
no
yes
a
b
Values taken from ref 33. Deuteron transfer is not observed
for any of the deuterated reagents listed. c The two sites are
d
nonequivalent, d1:d2 ) 4:1. A value of -1 indicates the initial
deuterium can be exchanged for a hydrogen if protio reagent is
used.
species. Zero-point energies were scaled by 0.9135 (HF) and
0.9646 (MP2), and all energies were adjusted from 0 to 298 K
by scaling the harmonic frequencies by 0.8929 (HF) and 0.9427
(MP2).24 All acidities are reported at 298 K.
Resu lts a n d Discu ssion
We were interested in generating a minimally stabi-
lized cyclopropenyl anion in order to explore its thermo-
chemistry and reactivity. An aryl group seemed to be a
reasonable substituent since ab initio calculations predict
that methyl (G2+) and vinyl (MP2/6-31+G(d) via isogyric
reactions with cyclopropyl and allyl radicals) derivatives
are unstable with regard to electron loss, and efforts to
generate 3-methyl-3-cyclopropenyl anion were unsuc-
cessful.14,25 The same approach that was used to generate
3-methoxycarbonyl-3-cyclopropenyl anion (1) was at-
tempted, and 3-(4-trifluoromethylphenyl)-3-trimethylsi-
lylcyclopropene (2) was reacted with fluoride ion in a
Fourier transform mass spectrometer (eq 3). A few initial
experiments, however, revealed that the ion which cor-
responds to loss of the trimethylsilyl group, [2-TMS]-, is
not cyclopropenyl anion 3 as its proton affinity is too low.
Specifically, the ion does not react to give the conjugate
base of a reference acid until it is reacted with hydro-
chloric acid. This result along with those given in Table
1 lead to a bracketed proton affinity of 338 ( 5 kcal/mol.
This can be contrasted with 1 which is sufficiently basic
to deprotonate water (∆H°acid (1H) ) 391 ( 4 kcal/mol).
In addition, calculations predict the proton affinity of 3
and its vinyl anion isomer to be 386.8 and 359.8 kcal/
mol (MP2/6-31+G(d)//HF/6-31+G(d)), respectively, or
382.1 and 361.3 kcal/mol (B3LYP/6-31+G(d)//HF/6-31+G-
(d)), respectively (Table 3); these levels of theory will be
denoted as MP2//HF and B3LYP//HF, respectively.
All computations were carried out using Gaussian 9422 or
GAMESS23 on UNIX-based workstations or Cray supercom-
puters. Structures were optimized at the Hartree-Fock level
of theory using the 6-31G(d) or 6-31+G(d) basis set. In some
cases, geometries also were explored with Møller-Plesset
second-order perturbation (MP2) theory. Vibrational analyses
were carried out to ensure structures correspond to true energy
minima. Single point energy determinations were made at the
MP2 level of theory, which accounts for electron correlation,
and also were investigated with density functional theory
using the B3LYP hybrid functional. These two approaches
render acidities which agree within 1-2 kcal/mol of each other
for most of the compounds in this work. Larger differences
(3-5 kcal/mol) are observed for the trifluoromethyl substituted
(17) Bergmann, E. D.; Berthier, G.; Ginsburg, D.; Hirshberg, Y.;
Lavie, D.; Pinchas, S.; Pullman, B.; A., P. Bull. Soc. Chim. Fr. 1951,
661-669.
(18) Wang, T. C. L.; Ricca, T. L.; Marshall, A. G. Anal. Chem. 1986,
58, 2935-2938.
(19) Marshall, A. G.; Roe, D. C. J . Chem. Phys. 1980, 73, 1581-
1590.
(20) Wenthold, P. G.; Squires, R. R. J . Am. Chem. Soc. 1994, 116,
11890-11897.
(21) Bevington, P. R. In Data Reduction and Error Analysis for the
Physical Science; McGraw-Hill: New York, 1969; pp 92-118.
(22) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T.; Peterson, G.
A.; Montgomery, J . A.; Raghavachari, K.; Al- Laham, M. A.; Zakrewski,
V. G.; Ortiz, J . V.; Foresman, J . B.; Cioslowski, J .; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, R. Y.; Chen,
W.; Wong, M. W.; Andres, J . L.; Replogle, E. S.; Gomperts, R.; Martin,
R. L.; Fox, D. J .; Binkley, J . S.; Defrees, D. J .; Baker, J .; Stewart, J .
P.; Head-Gordon, M.; Gonzalez, C.; Pople, J . A. Gaussian 94 Revisions
A, B, C; Gaussian, Inc.: Pittsburgh, PA, 1995.
The observed reactivity of the [2-TMS]- ion is consis-
tent with that expected for vinyl anion 4. In particular,
reaction with deuterated acids ranging from methanol-
(23) Schmidt, M. W.; Baldridge, K. K.; Boatz, J . A.; Elbert, S. T.;
Gordon, M. S.; J ensen, J . H.; Koseki, S.; Matsunaga, N.; Nguyen, K.
A.; Su, S. J .; Windus, T. L.; Dupuis, M.; Montgomery, J . A. J . Comput.
Chem. 1993, 14, 1347-1363.
(24) Pople, J . A.; Scott, A. P.; Wong, M. W.; Radom, L. Isr. J . Chem.
1993, 33, 345-350.
(25) Kass, S. R., unpublished results.

Rearrangements of Cyclopropenyl Anions
J . Org. Chem., Vol. 66, No. 1, 2001 101
Ta ble 2. Da ta for th e Kin etic Acid ity Mea su r em en t
Usin g 3-(4-Meth ylp h en yl)-3-tr im eth ylsilylcyclop r op en e
(12) a n d 6-Meth yl-3-tr im eth ylsilylin d en e (13)a
Sch em e 1. P ossible Rea r r a gem en ts of
3-(4-Tr iflu or om eth ylp h en yl)cyclop r op en yl An ion
(3)
∆H°acid (RH)
b
R-TMS
ethyl
methyl
cyclopropyl
vinyl
phenyl
12
13
Me3SiO-/R(Me)2SiO-
ln(3*ratio)
kcal/mol
0.122 ( 0.01
1.00
-1.01 ( 0.13 420.1 ( 2.0
0.000 416.7 ( 0.7
0.966 ( 0.05 411.5 ( 2.0
1.94 ( 0.01 409.4 ( 0.6
3.28 ( 0.03 401.7 ( 0.5
4.12 ( 0.08 398.6 ( 1.4
5.62 ( 0.23 392.2 ( 2.9
0.876 ( 0.034
2.32 ( 0.02
8.84 ( 0.27
20.6 ( 1.4
91.8 ( 22.1
a
Acidity values for calibrant species are taken from ref 33.
Ratios are isotopically corrected.
b
Ta ble 3. Su m m a r y of Exp er im en ta l a n d Com p u ted
Acid itiesa
matic species ring opens to give vinylcarbene intermedi-
ate 7 which then readily reacts by C-H bond insertion
to afford vinyl ion 4 (pathway a′). Upon reaction with an
acid, this species isomerizes to indenyl anion 5. The
transformation of neutral aryl cyclopropenes to indene
derivatives is a known thermal and photochemical pro-
cess in solution.16,26 Alternatively, one could envision 7
generating aryl-substituted deprotonated allenes or
alkynes (pathway a′′). Initial pseudorotation of 3 prior
to ring opening also could give rise to a similar product
via pathway b. Given our results, these latter two
reaction pathways (a′′ and b) do not appear to be
occurring to a significant extent, if at all, as these product
ions are expected to be more basic and/or incorporate two
deuteriums when only one is observed with the [2-TMS]-
ion.27 Moreover, in the reaction with carbon disulfide, the
known reaction pathways for allenyl and propargylic
derivatives are not detected.28-30
cmpd
expt
MP2//HFb
B3LYP//HFc
3H (C3)
(C1)
4H (C3)
(C1)
15H (C3)
(C1)
16H (C1)
indene (C3)
8H (C3)
386.8
359.8
382.1
361.3
386.1
339.9
399.0
370.5
386.3
-
338 ( 5
383.3
335.2
398.6 ( 1.4d
398.0 (398.1)e
368.6
390 ( 3f
386.9 (387.0)e
392.0
392.2 ( 2.9g
-
389.5h
a
b
All values in kcal/mol. MP2/6-31+G(d)//HF/6-31+G(d), 298
d
K. c B3LYP/6-31+G(d)//HF/6-31+G(d), 298 K. Experimental value
for 3-(4-methylphenyl)cyclopropene (14H). e MP2/6-31+G(d)//MP2/
6-31+G(d) values given in parentheses. f Reference 38. g Value for
h
the 3 position of 6-methylindene. Calculated by eq 10, see text.
d1 to 2,2,2-trifluoroethanol-d1 induces a single hydrogen-
deuterium exchange and an isomerization to the more
stable indenyl ion 5 (eq 3, Table 1). The latter species is
capable of undergoing a second H/D exchange with
stronger acids such as acetic acid-d4 and trimethylacetic
acid-d1. The significantly faster rate for the first exchange
with these acids supports the idea of two nonequivalent
sites. Evidence for a rearranged ion also was obtained
by observing that the deuterium atoms incorporated in
the [2-TMS]- ion upon reaction with D2O could not be
washed out with H2O. To verify our proposed structure
for the [2-TMS]- ion, 4 was independently prepared by
fluoride-induced desilylation of 3-trimethylsilyl-6-trifluo-
romethylindene (6, eq 4a). The reactivity of this ion was
found to be identical to that observed for the [2-TMS]-
ion (Table 1). The bracketed proton affinity of 5 (338 ( 5
kcal/mol) also is in agreement with theory, 335.2 kcal/
mol (MP2//HF) and 339.9 kcal/mol (B3LYP//HF). Carbon
disulfide and carbonyl sulfide also are effective probe
reagents for distinguishing 4 from 5. In the former case,
the ion abstracts a sulfur atom from both reagents while
no products are detected in the latter case (eq 4). The
ion generated from 2 and its d1 species display similar
reactivity differences with these reagents.
Given the propensity for rearrangement of 3-aryl-
substituted cyclopropenyl anions, we decided to reinves-
tigate the chemistry of the triphenyl derivative 8. Bar-
tmess and co-workers reported generation of this ion
upon thermal electron impact of 9 in an ICR; no reaction
was observed with water, and proton transfer was noted
with methanol and stronger acids which suggested a
proton affinity for 8 of 385 ( 5 kcal/mol.15 In our study,
9 was introduced into the FTMS via a heated (45 °C) solid
probe inlet and reacted with fluoride ion.31 The chemistry
of the resulting thermalized desilylated ion was explored,
and in contrast to the previously reported work, proton
transfer was not observed upon reaction with methanol-
d1, 2,2,2-trifluoroethanol-d1, or acetic acid-d4, but was
with hydrochloric acid. Acetic acid-d4 induces a single
hydrogen-deuterium exchange. Identical behavior is
observed if the [9-TMS]- ion is generated by electron
(26) Klett, M. W.; J ohnson, R. P. J . Am. Chem. Soc. 1985, 107, 3693-
3971.
(27) The proton affinities of the alkynyl and allenyl derivatives are
not known, but they are estimated to be larger (more basic) than that
for 5 based on the following acidities: ∆H°acid ) 380 (allene), 371
(phenylacetylene), and 368 kcal/mol (3-phenylpropene).
(28) DePuy, C. H.; Bierbaum, V. M.; Robinson, M. S.; Davico, G. E.;
Gareyev, R. Tetrahedron 1997, 53, 9847-9856.
(29) DePuy, C. H. Org. Mass Spectrom. 1985, 20, 556-559.
(30) An additional ion that was considered, but which can be ruled
out, is the vinyl anion derived from 3. Acid-induced isomerization of
cyclopropenyl anions is known, but this species is inconsistent with
the observed H/D exchange behavior (one should not be able to wash
out the label); it is predicted to be much more basic (359.8 kcal/mol at
the MP2//HF level), and the reactivity of this type of ion (see ref 40) is
different from the species we observed.
(31) In the experiments with 9 an M-1 ion at m/z 339 also is
generated. This ion arises from electron ionization not from reaction
with F-. The 9-TMS ion displays the same reactivity when it is not
isolated and the entire experiment is done in 1 cell.
The following mechanism is proposed for the formation
of 4 from 3 (Scheme 1). The initially generated antiaro-

102 J . Org. Chem., Vol. 66, No. 1, 2001
Broadus et al.
ionization and stray electrons are kept out of the cell.32
Our results lead to a bracketed proton affinity of 341 (
8 kcal/mol for the [9-TMS]- ion. This value is too low for
a cyclopropenyl anion and strongly suggests the initially
formed anion rearranges to 1,2-diphenyl-1-indenyl ion
(10) (eq 5). To support this hypothesis, 10 was indepen-
dently prepared by deprotonation of its conjugate acid,
and its reactivity with the reference acids mentioned
above was found to be identical to that of the [9-TMS]-
ion. In addition, the thermochemistry is in accord with
the reported gas-phase acidities of 1,3-diphenylindene
(335.2 ( 2.1 kcal/mol) and 1,2,3-triphenylindene (335.5
( 2.1 kcal/mol).33
but can represent transition state structures. In fact, ab
initio calculations indicate that explusion of an alkyl
anion is facilitated by the simultaneous deprotonation
of the alcohol.36 In any case, the natural logarithm of the
statistically corrected ratio of these products correlates
linearly with gas-phase acidities. In this work, various
substituted trimethylsilanes were used to construct a
calibration curve which yielded the following equation:
∆H°acid(RH) ) -4.29 × ln(Me3SiO-/R(Me)2SiO-) + 416.3,
r2 ) 0.984. A complete listing of the data is given in Table
2.
The calculated gas-phase acidity of 3-(4-trifluorometh-
ylphenyl)cyclopropene suggests this compound would not
be suitable for the described kinetic measurement as the
∆∆H°acid relative to methane is larger than that of the
calibrant species and is too big to give an accurate Me3-
SiO-/R(Me)2SiO- ratio. Therefore, 3-(4-methylphenyl)-3-
trimethylsilylcyclopropene (12) was used. It should be
mentioned that the [M - TMS]- ion generated upon
fluorodesilylation of 12 was briefly explored and found
to rearrange, presumably, to the indenyl derivative; the
[M - TMS]- ion undergoes 2 H/D exchanges with D2O,
1 with MeOD, and 2 with CF3CH2OD, and proton
transfer is observed with CD3CO2D, which is consistent
with the acidity of indene (∆H°acid ) 352.0 ( 2.1 kcal/
mol). The reaction of hydroxide with 12 affords two
product ions at m/z 201 and m/z 129 in addition to the
desired R(Me)2SiO- and Me3SiO- ions (eq 7). The ion at
m/z 201 corresponds to deprotonation of 12 at the vinyl
position (∆H°acid ) 369 kcal/mol (MP2//HF) for the
equivalent site in 3-phenylcyclopropene). The presence
of this ion complicates measurement of the R(Me)2SiO-
intensity due to its significant 13C, 29Si, and 30Si isotope
peaks at the same nominal mass. The use of oxygen-18
labeled hydroxide eliminates this interference by shifting
the R(Me)2SiO- signal from m/z 203 to m/z 205. Overall,
a ratio of 46:1 was observed for the m/z 201 and m/z 205
ions, and thus it would have been difficult to reliably
measure the desired ratio without removing the isotopic
overlap (Figure 1).
Results for the [9-TMS]- ion indicate that the initially
observed species is not vinyl ion 11. Specifically, the
indicative behavior of one hydrogen-deuterium exchange
with weaker acids and sulfur-atom abstraction with
carbon disulfide is not observed. In the absence of an acid-
induced isomerization, as we have proposed for 3-aryl-
3-cyclopropenyl anions, two sequential phenyl migrations
are postulated to account for the formation of 10 from
11. Such rearrangements are precedented under pho-
tolytic conditions for 1,2,3-triphenylcyclopropene26 and for
2-phenylethyl anion in the gas phase.34 Isomerization of
9 while it was heated (45 °C) during the course of the
experiments did not occur since 9 is stable in refluxing
THF (65 °C) and higher temperatures were required to
induce its isomerization (i.e., 9 rearranges at 190-200
°C to an undetermined isomer with a half-life of ∼90 h.)
Although 3-aryl-3-cyclopropenyl anions appear to be
unstable with respect to rearrangement in the gas phase,
their experimental proton affinities (or, equivalently,
∆H°acid of their conjugate acids) might be obtained using
kinetic methods. DePuy and co-workers have measured
such thermochemical data for various alkanes whose
conjugate bases are unbound.35 This approach does not
require the anion to be generated as a free species, but
rather determines an unknown’s acidity relative to that
of methane based on the fragmentation of a pentacoor-
dinate siliconate anion (eq 6). Specifically, the intermedi-
ate formed upon reaction of a substituted trialkylsilane
The other product, m/z 129, corresponds to loss of a
trimethysilyl group from 12, and its origin must be
accounted for. The structure of this ion was determined
to be the conjugate base of 6-methylindene based on
observed proton transfer with acetic acid but not weaker
acids and the absence of hydrogen-deuterium exchange
(33) All acidities unless otherwise noted come from the following
reference: Bartmess, J . E. NIST Chemistry WebBook, NIST Standard
Reference Database Number 69. In Secondary NIST Chemistry Web-
Book, NIST Standard Reference Database Number 69; Mallard, W. G.,
Linstrom, P. J ., Eds.; National Institute of Standards and Technol-
ogy: Gaithersburg, MD 20899 (http://webbook.nist.gov).
(34) Maas, W. P. M.; van Veelen, P. A.; Nibbering, N. M. M. Org.
Mass. Spectrom. 1989, 24, 546-558.
(35) DePuy, C. H.; Gronert, S.; Barlow, S. E.; Bierbaum, V. M.;
Damrauer, R. J . Am. Chem. Soc. 1989, 111, 1968-1973.
(36) Davis, L. P.; Burggraf, L. W.; Gordon, M. S. J . Am. Chem. Soc.
1988, 110, 3056-3062.
such as RSiMe3 with hydroxide can dissociate by expul-
sion of either CH3 or R-. The incipient ion then depro-
-
tonates the hydroxysilane to generate R(Me)2SiO- or
Me3SiO-, respectively. While this is the commonly given
explanation for this reaction, it is important to recognize
that the species in brackets need not be intermediates,
(32) The yield of the [9-TMS]- ion by electron ionization (1.3 eV, 25
s) is low.

Rearrangements of Cyclopropenyl Anions
J . Org. Chem., Vol. 66, No. 1, 2001 103
Sch em e 2. P ossible Rea ction P a th w a ys in th e Kin etic Acid ity Mea su r em en t w ith 12
kcal/mol. The p-methyl group in 14H should have a
negligible impact on the thermochemistry; thus, 14H and
15H will be considered equivalent for further discussion.
3-(4-Methylphenyl)cyclopropene represents the simplest
cyclopropene whose acidity has been measured experi-
mentally in the gas phase. Cyclopropene, itself, is com-
puted to be 20 kcal/mol less acidic, 418.9 kcal/mol at the
CCS D(T)/6-311 +G(2d f,2p d )//MCS CF (10(9),9)/6-
31+G(d,p) level.13 Comparison of the proton affinity of
15 with that of its saturated analogue 16 (390 ( 3 kcal/
mol)38 indicates that the cyclopropenyl system is desta-
bilized by 9 ( 3 kcal/mol (11 kcal/mol MP2//MP2) (eq 8).
This quantity is slightly larger than an equivalent
isodesmic reaction for the parent system (i.e., 3-cyclo-
propenyl anion is destabilized relative to cyclopropyl
anion by 5.8 kcal/mol at the CCSD(T)/6-311+G(2df,2pd)//
MCSCF(10(9),9)/6-31+G(d,p)) level.13 The difference (3
kcal/mol) can be attributed, at least partially, to the
preferential stabilization (∼2 kcal/mol) of the cyclopropyl
group by the phenyl substituent as a result of improved
resonance delocalization.39 The energetics of eq 8 are
predominantly due to the antiaromatic interaction in 15;
however, subtle differences in strain energy, polarizabil-
ity, and inductive effects must be considered as well.
3-Phenyl-3-cyclopropenyl anion also can be compared
with a linear analogue, 2-phenylallyl anion (eq 9). This
model indicates comparable destabilization within ex-
perimental error (13 ( 5 kcal/mol) for 15 upon cyclic
conjugation. One can compare 15 to 1-phenylallyl anion
(∆H°acid(3-phenylpropene) ) 369 ( 4 kcal/mol)38 too, in
which case the difference (30 ( 4 kcal/mol) is much larger
F igu r e 1. Abundance of the m/z 201 and m/z 205 ions
generated in the reaction of 12 with oxygen-18 labeled
hydroxide.
F igu r e 2. Experimental gas-phase acidites obtained by the
Depuy kinetic method.
with D2O; deuterium incorporation is expected if the
-
C10H9 ion corresponds to a cyclopropenyl anion or a
3-indenyl anion derivative. Two possible pathways (2 and
3) for the formation of the m/z 129 ion are shown in
Scheme 2. The distinct ratios for the loss of CH4 and
C10H10 observed for the siliconates of 12 and 6-methyl-
3-trimethylsilylindene (13) (Table 2, Figure 2), the good
agreement with computed acidities (Table 3), and the fact
that pathway 3 does not take place with other 3-trim-
ethylsilyl-substituted cyclopropenes giving rise to stabi-
lized cyclopropenyl anions10,11 argues against this reac-
tion channel. Therefore, we attribute the formation of the
but no different than for the parent system (i.e., ∆∆H°acid
-
(cyclopropene - propene) ) 418.9-389.1 ) 29.8 kcal/
mol).
-
C10H9 ion (m/z 129) to pathway 2b, and consequently
its intensity was added to that of trimethylsiloxide for
the acidity measurement. Overall, the correction is minor
and results in a smaller (more acidic) ∆H°acid for 3-(4-
methylphenyl)cyclopropene (14H) by 2 kcal/mol.37
Based on our measurements with 12, a value of 398.6
( 1.4 kcal/mol can be assigned for the gas-phase acidity
of 14H. This is in favorable agreement with ab initio and
density functional calculations for 3-phenylcyclopropene
(15H) which predict a deprotonation energy of 398.0
(MP2//HF), 398.1 (MP2//MP2), and 399.0 (B3LYP//HF)
We have used theory to gain insights into the struc-
tures of 3-aryl-3-cyclopropenyl anions. The geometries
discussed here were optimized at the HF/6-31+G(d) level
(37) The error limits in our measured acidity come from an error
analysis which includes the uncertainty in the measured ion ratios
and the reference acids. As in any kinetic determination, unaccounted-
for systematic errors can lead to larger uncertainties in the final result.
There is no apparent reason there should be a problem in this case
and, indeed, experiment and theory are in excellent agreement.
(38) Chou, P. K.; Dahlke, G. D.; Kass, S. R. J . Am. Chem. Soc. 1993,
115, 315-324.

104 J . Org. Chem., Vol. 66, No. 1, 2001
Broadus et al.
Ta ble 4. HF /6-31+G(d ) Geom etr ies for P h en yl-
cyclop r op en e (15H), 3-(4-Tr iflu or om eth ylp h en yl)cyclo-
p r op en e (3H) a n d Th eir C3 Con ju ga te Ba sesa
R ) H
15H (Cs)
R ) CF3
3H (C1) 3 (Cs)b
parameter
C1-C2
15 (Cs)
1.277 (1.306) 1.291(1.322) 1.276
1.497 (1.512) 1.511 (1.510) 1.497
1.497 (1.512) 1.511 (1.510) 1.498
1.505 (1.494) 1.449 (1.426) 1.503
1.392 (1.404) 1.416 (1.428) 1.389
1.388 (1.398) 1.382 (1.394) 1.388
1.386 (1.400) 1.394 (1.408) 1.383
1.388 (1.400) 1.394 (1.408) 1.391
1.386 (1.398) 1.382 (1.394) 1.381
1.394 (1.403) 1.416 (1.428) 1.396
1.076 (1.088) 1.077 (1.090) 1.502
1.069 (1.081) 1.073 (1.085) 1.069
1.069 (1.081) 1.073 (1.085) 1.069
1.303
1.452
1.452
1.331
1.473
1.355
1.424
1.424
1.355
1.473
1.456
1.070
1.070
53.3
C1-C3
C2-C3
C3-C4
C4-C5
C5-C6
C6-C7
C7-C8
C8-C9
C4-C9
C7-R10
C1-H11
C2-H12
C1-C3-C2
C2-C1-C3
C3-C1-C2
Rc
F igu r e 3. HF/6-31+G(d) structures of phenylcyclopropene,
3-(4-trifluoromethylphenyl)cyclopropene, and their conjugate
bases.
bond length relative to 3H (1.331 vs 1.503 Å). This
difference is more than three times that computed for
15 and 15H. The olefinic bond in 3 lengthens a modest
amount (1.276 to 1.303 Å) as is typical for cyclopropenyl
anions, and the paraffinic bonds decrease (1.497(8) to
1.452 Å) as also noted for 3-formyl- and 3-methoxycar-
bonyl-3-cyclopropenyl anions.11,25,40 This last structural
change presumably arises because of the hybridization
change at C3.
50.5 (51.2)
64.7 (64.4)
64.7 (64.4)
53.5 (54.8)
1.6 (2.3)
50.6 (52.0)
64.7 (64.0)
64.7 (64.0)
64.2 (49.2)
9.8 (12.4)
50.5
64.8
64.8
53.7
63.4
63.4
0.5
âd
1.0(0.9) 0.0(0.0)
149.6 1.0
C2-C3-C4-C5 149.6 (149.6) 62.3 (58.8)
a
Bond lengths are in angstroms, and angles are in degrees.
Parenthentical values correspond to MP2/6-31+G(d) geometries.
b
3 was optimized using no symmetry constraints but has Cs
symmetry. c R is the displacement of the aryl group from the plane
d
of the cyclopropene ring. See Figure 3. â is the deviation of H11
To predict the gas-phase acidity of triphenylcyclopro-
pene (8H) we turned to theory. A reliable computed
number requires an optimized structure for the conjugate
base which includes diffuse functions since they are
needed to properly describe the electron distribution of
anions. We were unable to obtain the energy (or a
geometry) of 8H with the 6-31+G(d) basis set at the HF,
B3LYP, and BVWN5 levels of theory, however, which
precluded a direct determination of the acidity with this
basis set. We, therefore, computed the acidity of 8H
relative to 3-phenylcyclopropene as shown in eq 10 and
adjusted this difference using our experimental number
for 3-(4-methylphenyl)cyclopropene (see above). This
approach yields a value of 389.5 kcal/mol for the acidity
of 1,2,3-triphenylcyclopropene and indicates that the
additional phenyl groups at the vinyl positions stabilize
the anion by 9 kcal/mol or slightly less than half of the
effect of the first phenyl group. In contrast, the acidity
of 8H in solution is only 11 pKA units (or 15 kcal/mol)
more acidic than cyclopropene (50 vs 61),41,42 and al-
though the pKA of 3-phenylcyclopropene has not been
measured, it is clear that the remote phenyl groups have
less of an impact in solution (at least in an absolute
sense). Nevertheless, based on a variety of trapping
experiments, it is apparent that remote phenyl groups
facilitate the generation of the corresponding cyclopro-
penyl anions and that incorporation of p-trifluoromethyl
and H12 from the plane of the cyclopropene ring. In 3H and 15H,
H11 and H12 are on the same side of the three-membered ring as
the aryl group whereas the opposite is true for 15.
of theory unless otherwise noted. Deprotonation of phen-
ylcyclopropene (15H) at the 3 position (i.e., formation of
15) induces several structural changes (Table 4). The
phenyl group which is aligned with the Walsh orbitals
of 15H rotates 90° to be conjugated with the newly
formed charge. The C3-C4 bond length decreases by
0.056 Å (0.068 Å MP2), indicating delocalization of the
charge into the phenyl group. This interaction is slightly
counterbalanced by movement of the phenyl group out
of the plane of the cyclopropene ring (R increases from
53.5° to 64.2°, Figure 3), but this parameter is sensitive
to the computational method (i.e., R decreases by 5.6° at
the MP2 level). The vinyl hydrogens in 15 move 11.4° to
the opposite side of the cyclopropene plane, and the
paraffinic bonds (C1-C3 and C2-C3) lengthen by 0.014
Å. Both distortions serve to disrupt overlap of the C1-
C2 π bond with the electron lone pair. Elongation of the
C4-C5 and C4-C9 bonds in the phenyl ring by 0.024
and 0.022 Å, respectively, also is noteworthy. Introduc-
tion of a p-trifluoromethyl substituent on the phenyl ring
provokes a markedly different anion structure; 3 is nearly
planar, R ) 0.5° (Figure 3). The increased delocalization
in 3 is also reflected in the considerably shorter C3-C4
(39) The value of 2 kcal/mol comes from acidity differences: ∆∆H°acid
(cyclopropane - phenylcyclopropane) ) 22 ( 4 kcal/mol (expt, see refs
33 and 38) and ∆∆H°acid (cyclopropene - phenylcyclopropene) ) 20.3
kcal/mol (calc, see refs 13 and this work). Similar preferential stabi-
lizations for the methyl ester (7 kcal/mol, ref 40) and cyano (7 kcal/
mol, ref 11) derivatives also have been found.
(40) Kroeker-Sachs, R. Ph.D. Thesis, University of Minnesota, 1993.
(41) Breslow, R.; Balasubramanian, K. J . Am. Chem. Soc. 1969, 91,
5182-5183.
(42) Breslow, R.; Chu, W. J . Am. Chem. Soc. 1973, 95, 411-418.

Rearrangements of Cyclopropenyl Anions
J . Org. Chem., Vol. 66, No. 1, 2001 105
Ta ble 5. Ha r tr ee-F ock Geom etr ies for
Tr ip h en ylcyclop r op en e (8H, 6-31G(d )) a n d Its Con ju ga te
Ba se (8, 6-31+G(d ))
parameter
8H (Cs)a
8 (Cs)
1.302
C1-C2
1.288 (1.293)
1.495 (1.513)
1.495 (1.518)
1.506 (1.470)
1.390
1.386
1.384
1.387
1.384
F igu r e 4. Computed structures for 8H (HF/6-31G(d)) and 8
(HF/6-31+G(d)).
C1-C3
1.505
1.505
1.463
1.411
1.384
1.392
1.392
1.384
1.411
1.451
1.451
51.3
64.3
64.3
56.0
14.5
C2-C3
C3-C4
pene plane while the phenyl rings at C1 and C2 adopt a
nearly planar configuration with the cyclopropene ring
in order to be optimally conjugated with its double bond.
Upon anion formation, the 3-phenyl group moves into
conjugation with the electron lone pair and slightly bends
away from the plane of the cyclopropene ring (R ) 53.2°
vs 56.0°). The C3-C4 bond length exhibits the largest
alteration, decreasing by 0.043 Å. The vinyl phenyl
substituents move out of the plane of the cyclopropene
ring by 14.5° as well as undergoing a dramatic rotation
about the C1-C10 and C2-C11 bonds (Figure 4). The
magnitude of the latter distortion is reflected in the
change in the C11-C10-C1-C2 (176.3° vs. -141.6°) and
C12-C10-C1-C2 (-3.9° vs 38.0°) dihedral angles be-
tween 8H and 8. The out-of-plane distortion (the increase
in â) is consistent with ab initio calculations for related
systems and results so as to minimize the unfavorable
cyclic 4π electron interaction.13 The rotation of the C1
and C2 phenyl groups results in diminished conjugation
with the cyclopropene double bond, apparently, because
of steric congestion between the o-aryl hydrogens in all
three phenyl rings. As it is, the hydrogen-hydrogen
distances between the C3 and C1 (or C2) and the C1 and
C2 rings are 2.72 and 2.36 Å, respectively. If the phenyl
groups did not rotate then these two distances would be
2.04 and 1.85 Å, respectively (Figure 4). In 3-methoxy-
carbonyl-1,2-diphenylcyclopropenyl anion, where this
steric interaction is not present, the phenyl rings do not
rotate.25 Finally, it is worth noting that a MNDO semiem-
pirical geometry was previously reported for 8; the
structure is nearly C2v with the C3 phenyl group orthogo-
nal to the three-membered ring and the other two
phenyls nearly in the ring plane.45 This geometry is
entirely different from the ab initio one and most likely
is the result of MNDO’s now known problems in repro-
ducing carbanion geometries.
C4-C5
C5-C6
C6-C7
C7-C8
C8-C9
C3-C9
1.393
C1-C10
C2-C13
C1-C3-C2
C1-C2-C3
C2-C1-C3
Rb
1.454 (1.422)
1.454 (1.443)
51.0 (50.5)
64.5 (65.0)
64.5 (64.6)
53.2
0.6 (5.4)
149.2
176.3
âc
C2-C3-C4-C5
C11-C10-C1-C2
C12-C10-C1-C2
62.2
-141.6
38.0
-3.9
a
b
X-ray data from ref 43 are given in parentheses. R is the
displacement of the 3-phenyl group from the plane of the cyclo-
propene ring. c â is the deviation of the C1 and C2 phenyl groups
from the plane of the cyclopropene ring. In 8H, the C1 and C2
phenyl rings are on the same side of the three-membered ring as
the C3 phenyl group whereas the opposite is true for 8.
substituents increases the reaction rate even further.9,43
Con clu sion
The structural changes upon deprotonation of 1,2,3-
triphenylcyclopropene proved to be most intriguing (Table
5, Figure 4). The optimized geometry of 8H at the HF/
6-31G(d) level of theory was found to be in reasonable
accord with experimental X-ray crystallography data44
(Table 5) with minor deviations observed in the C1-C10
and C3-C4 bonds lengths; theory predicts the bonds to
be longer than experiment by 0.032 and 0.036 Å, respec-
tively. The 3-phenyl group is orthogonal to the cyclopro-
3-Aryl-substituted cyclopropenyl anions were found to
be unstable in the gas phase as they spontaneously
rearrange to 6-substituted 3-indenyl anions. 1,2,3-Tri-
phenylcyclopropenyl anion undergoes a similar gas-phase
isomerization, contrary to a previous report describing
its preparation. Despite the instability of its conjugate
base, the acidity of 3-(4-methylphenyl)cyclopropene was
(43) Han, S. Ph.D. Thesis, Unversity of Minnesota, 1997.
(44) Domnin, I. N.; Kopf, J .; Keyaniyan, S.; DeMeijere, A. Tetrahe-
dron 1985, 41, 5377-5382.
(45) Winkelhofer, G.; J anoschek, R.; Fratev, F.; Spitznagel, G. W.;
Chandrasekhar, J .; Schleyer, P. v. R. J . Am. Chem. Soc. 1985, 107,
332-337.

106 J . Org. Chem., Vol. 66, No. 1, 2001
Broadus et al.
determined by the DePuy kinetic method (∆H°acid ) 398.6
( 1.4 kcal/mol), and the value is in good agreement with
ab initio calculations. This result represents the simplest
cyclopropene whose acidity has been measured in the gas
phase. The structures and energies of 3-(4-methylphenyl)-
cyclopropene, 3-(4-trifluoromethylphenyl)cyclopropene,
1,2,3-triphenylcyclopropene, and their conjugate bases
were explored with theory. Formation of 1,2,3-triphenyl-
cyclopropenyl anion is predicted to be accompanied by
dramatic changes in the orientation of all three phenyl
substituents at the HF/6-31+G(d) level.
Ack n ow led gm en t. Support from the National Sci-
ence Foundation, the donors of the Petroleum Research
Foundation, as administered by the American Chemical
Society, and the Minnesota Supercomputer Institute are
gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Computed energies
and structures (x,y,z coordinates). This material is available
free of charge via the Internet at http://pubs.acs.org.
J O001102A