Photochemical generation and matrix-isolation detection of dimethylvinylidene

Article abstract of DOI:10.1021/jo001464i

We report the spectroscopic characterization of dimethylvinylidene, (CH₃)₂C=C:, generated within an argon matrix at 14 K from a bisperoxyester precursor. The carbene was identified by comparison of the experimental IR spectrum with vibrational frequencies computed at the B3LYP/6-31G(d) level. Chemical trapping of the carbene within a 9% CO/Ar matrix to form dimethylpropadienone supports this analysis. Additional products produced during photolysis were identified by comparison to the appropriate computed vibrational frequencies. The potential energy surface of dimethylvinylidene and its intramolecular rearrangement products, 2-butyne and methylcyclopropene, were also investigated computationally at the B3LYP/6-31G(d) level. A spin-state analysis of this carbene using a variety of computational methods (CCSD(T), B3LYP, MP2) indicates the singlet state is more stable than the triplet by ～45 kcal mol^-1. We anticipate the bisperoxyester precursor used here will be a convenient and general way for initiating future studies of alkylvinylidenes under matrix-isolation conditions.

Full text of DOI:10.1021/jo001464i

J . Org. Chem. 2001, 66, 287-299
287
P h otoch em ica l Gen er a tion a n d Ma tr ix-Isola tion Detection of
Dim eth ylvin ylid en e
,†
Sasha C. Reed,‡ Gregory J . Capitosti,§ Zhendong Zhu, and David A. Modarelli*,§
Departments of Chemistry, Knight Chemical Laboratory, The University of Akron,
Akron, Ohio 44325-3601, Colgate University, Hamilton, New York 13346, and The Ohio State University,
Columbus, Ohio 43210
dam@chemistry.uakron.edu.
Received October 10, 2000
We report the spectroscopic characterization of dimethylvinylidene, (CH3)2CdC:, generated within
an argon matrix at 14 K from a bisperoxyester precursor. The carbene was identified by comparison
of the experimental IR spectrum with vibrational frequencies computed at the B3LYP/6-31G(d)
level. Chemical trapping of the carbene within a 9% CO/Ar matrix to form dimethylpropadienone
supports this analysis. Additional products produced during photolysis were identified by comparison
to the appropriate computed vibrational frequencies. The potential energy surface of dimethylvi-
nylidene and its intramolecular rearrangement products, 2-butyne and methylcyclopropene, were
also investigated computationally at the B3LYP/6-31G(d) level. A spin-state analysis of this carbene
using a variety of computational methods (CCSD(T), B3LYP, MP2) indicates the singlet state is
more stable than the triplet by ∼45 kcal mol-1. We anticipate the bisperoxyester precursor used
here will be a convenient and general way for initiating future studies of alkylvinylidenes under
matrix-isolation conditions.
In tr od u ction
mol-1. Thus, vinylidene 1 is predicted to exist as a
reactive intermediate. Schaefer4,6 and others7,8 have also
computationally explored the energies and structures of
several substituted vinylidenes. Spin-state analyses in-
dicate the singlet states in all cases are more stable than
the corresponding triplet states by a substantial amount
(∼35-50 kcal mol-1).4 The large difference in spin-state
energies is not surprising considering the doubly occupied
orbital in the 1A1 singlet is sp-hybridized in origin.
Promotion of an electron to the nonbonding and higher
energy carbene p-orbital is unfavorable.9
Aryl, alkyl, and halocarbenes are by now well charac-
terized by theory and spectroscopic (UV-vis, IR, EPR,
time-resolved) techniques.1 The spectroscopy of alkylvi-
nylidenes (R2CdC:, i.e., 1) is not as well understood,
although the solution-phase chemistry of alkylvinylidenes
is relatively well established and is typical of carbene
chemistry.2 Considering the important role these reactive
species may play in the interstellar medium,3 the lack of
spectroscopic evidence for alkylvinylidenes is of current
interest in physical organic chemistry.
Vinylidene and its deuterated and fluorinated ana-
logues have been indirectly detected by gas-phase pho-
toelectron-detachment9 and dispersed-fluorescence spec-
troscopies.10 From Lineberger’s work9 the vibrational
spectrum of 1 was determined and the barrier to rear-
rangement to acetylene estimated from the 2r0 CH2-
rock mode to be ∼1.3 kcal mol-1. The lifetime of (RdH)
1 was also estimated to be ∼0.04-0.2 ps on the basis of
line-broadening simulations. The enthalpy of this reac-
tion has been deduced9,10 to be exothermic by ∼41-45
The theoretical debate of whether 1 (R ) H) actually
exists as a bound state is substantial.4 The highest level
calculations (CCSD(T)/TZ+2P, EOM-CCSD/TZ2P) to date
from Schaefer’s and Stanton’s groups4,5 indicate the
barrier height for isomerization to acetylene to be ∼3 kcal
kcal mol-1
.
Recently, matrix isolation has been used to isolate and
study two vinylidenes. Maier et al.11 irradiated cyclopro-
penylidene in an argon matrix to isolate a mixture of
§ The University of Akron.
‡ Colgate University.
The Ohio State University.
† Current Address: Institute of Cognitive and Computational Sci-
ence, Georgetown University Medical Center, Washington, DC 20007.
(1) Platz, M. S. Kinetics and Spectroscopy of Carbenes and Biradi-
cals; Plenum Press: New York, 1990.
(6) (a) Hu, C.-H.; Schaefer, H. F., III. J . Phys. Chem. 1993, 97, 10681.
(b) Collins, C. L.; Meredith, C.; Yamaguchi, Y.; Schaefer III, H. F. J .
Am. Chem. Soc. 1992, 114, 8694. (c) DeLeeuw, B. J .; Fermann, J . T.;
Xie, Y.; Schaefer, H. F. J . Am. Chem. Soc. 1993, 115, 1039.
(7) Apeloig, Y.; Schrieber, R.; Stang, P. J . Tetrahedron Lett. 1980,
21, 411.
(8) Modarelli, D. A.; Ichimura, A. S. Unpublished data.
(9) Ervin, K. M.; Ho, J .; Lineberger, W. C. J . Chem. Phys. 1989, 91,
5974. Also see the following for information on the corresponding work
with difluorovinylidene: Gilles, M. K.; Lineberger, W. C.; Ervin, K.
M. J . Am. Chem. Soc. 1993, 115, 1031.
(2) J ones, M., J r.; Moss, R. A. Carbenes; Krieger Publishing Co.:
Malabar, FL, 1983.; Kirmse, W. Carbene Chemistry; Academic Press:
New York, 1971.
(3) Vrtilek, J . M.; Gottlieb, C. A.; Gottlieb, E. W.; Killian, T. C.;
Thaddeus, P. Astrophys. J . 1990, 364, L53. Cernicharo, J .; Gottlieb,
C. A.; Gue´lin, M.; Killian, T. C.; Paubert, G.; Thaddeus, P.; Vrtilek, J .
M. Astrophys. J . 1991, 368, L39.
(4) Gallo, M. M.; Hamilton, T. P.; Schaefer, H. F. J . Am. Chem. Soc.
1990, 112, 7714-8719.
(5) Stanton, J . F.; Gauss, J . J . Chem. Phys. 1994, 101, 3001-3005.
(10) J acobson, M. P.; Silbey, R. J .; Field, R. W. J . Chem. Phys. 1999,
110, 845. Chen, Y.; J onas, D. M.; Kinsey, J . L.; Field, R. W. J . Chem.
Phys. 1989, 91, 3976.
10.1021/jo001464i CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/15/2000

288 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
propadienylidene (2) and propynylidene (3). McMahon et
al.12 cleanly generated 2 by direct photolysis of propy-
nylidene (generated from diazopropyne) in an argon
matrix. Identification of 2 was determined by its CdC
stretch at 1952 cm-1. Recently, Stanton, McMahon, and
co-workers13 were able to deconvolute the transitions in
the complex absorption spectrum of 2 using the equation-
of-motion coupled-cluster method (EOMEE-CCSD/TZ2P).
Difluorovinylidene (4) was recently observed by Sander
et al.,14 generated by irradiation of 1,2-difluoroacetylene
in an argon matrix. The stability of 4 relative to 1 (R )
D) was attributed to a high barrier to 1,2-F migration
(i.e., 35-40 kcal mol-1 vs ∼3 kcal mol-1).6c,14 The IR
spectrum recorded for 3 revealed an intense absorption
at 1267 cm-1 that was assigned to the asymmetric CF2
stretch (ν4, b1) vibrational state coupled by Fermi reso-
nance to the 1245 cm-1 absorption (ν2+ν5, b1). Significant
to the work presented here, Sander and co-workers
reported a second intense vibration at 1672 cm-1 that
could be assigned to the ν1 CdC stretch. The carbene was
trapped with CO and N2, yielding stable products that
decomposed upon further irradiation.14
Condensed phase (solution, solid state) spectroscopy
has not been performed at all on simple alkylvinylidenes
such as 5. The lack of spectroscopic evidence is a direct
result of the lack of a suitable precursor for photochemi-
cal generation. Typical carbene precursors such as diaz-
irines and diazo compounds either are not stable or do
not form, although attempts have been made to synthe-
size these precursors.15,16 Peroxyesters represent poten-
tially useful precursors for the generation of alkylvinyli-
denes, however. Such precursors have previously been
used as both thermal11 and photochemical17,18 sources of
carbenes. Chapman and Adam18 irradiated the cyclic
peroxyester 7 at 77 K in a glassy matrix and observed
production of R-lactone 8. Upon further irradiation, these
lactones were found to extrude CO, resulting in ketone
formation and in one case (8b) loss of CO2 to yield allene,
In this work we report the use of a peroxyester
precursor to generate and observe the matrix-isolation
IR spectrum of 5. In addition we have performed a
detailed theoretical interpretation of these results and
of the potential energy surface (PES) of the possible
rearrangement products of 5. The information presented
here represents, to the best of our knowledge, the first
spectroscopic study of a dialkylvinylidene.
Exp er im en ta l Section
Com p u ta tion a l Meth od ology. Geometry optimizations
were performed at the Hartree-Fock level with the 6-31G(d)
basis set, as well as by using Becke’s gradient corrected
(original20 and three-parameter21) exchange functionals with
the correlation functional of Lee, Yang, and Parr22 (BLYP and
B3LYP, respectively23) with both the 6-31G(d) and 6-311G-
(2d,p) basis sets. Single point energies were computed at
the B3LYP/cc-pVTZ//B3LYP/6-31G(d), BLYP/cc-pVTZ//BLYP/
6-31G(d), and CCSD(T)/cc-pVTZ//B3LYP/6-31G(d) levels. When
used with the 6-31G(d) basis set, the DFT functionals have
been shown to give excellent agreement with both experiment24
and higher levels of theory.24,25 DFT methods are also known
to recover some electron correlation. Transition state struc-
tures were located using the synchronous transit-guided quasi-
Newton (STQN) method developed by Peng and Schlegel26 and
were performed at the B3LYP/6-31G(d) level. IRC calculations
were performed at the B3LYP/6-31G(d) level. All calculations
were performed with the Gaussian9427 and Gaussian9828 sets
of programs. Molecular orbitals were visualized using Mac-
MolPlt29 and calculated with MacGAMESS.30 Imaginary vibra-
tions were observed using an animation program.
Ma tr ix-Isola tion Exp er im en ts. The APD matrix isolation
apparatus has been described earlier.31 In general, the liquid
(19) Thelen, M.-A.; Felder, P.; Frey, J . G.; Huber, J . R. J . Phys.
Chem. 1993, 97, 6220.
(20) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
(21) Becke, A. D. J . Chem. Phys. 1993, 98, 5648.
(22) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(23) A description of the Gaussian implementation of density
functionals can be found in the following: J ohnson, B. G.; Gill, P. M.
W. L.; Pople, J . A. J . Chem. Phys. 1993, 98, 5612.
(24) Matzinger, S.; Bally, T.; Patterson, E. V.; McMahon, R. J . J .
Am. Chem. Soc. 1996, 118, 1535-1542.
(25) Schreiner, P. R.; Karney, W. L.; von R. Schleyer, P.; Borden,
W. T.; Hamilton, T. P.; Schaefer, H. F., III. J . Org. Chem. 1996, 61,
7030-7039.
(26) Peng, C.; Schlegel, H. B. Isr. J . Chem. 1993, 33, 449.
(27) 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.; Petersson,
G. A.; Montgomery, J . A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.; Cioslowski, J .;
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. 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, Rev. E.2; Gaussian, Inc.: Pittsburgh, PA, 1995.
(28) 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.; Chen, W.;
Wong, M. W.; Andres, J . L.; Gonzalez, C.; Head-Gordon, M.; Replogle,
E. S.; Pople, J . A. Gaussian 98, Rev. A.6; Gaussian, Inc.: Pittsburgh,
PA, 1998.
presumably through a carbene intermediate. Similarly,
acyclic peroxide precursors have also been used as
carbene precursors.11 Thus, irradiation of tert-butylper-
oxyesters produces CO2 and tert-butoxy radicals, the
latter of which can undergo R-cleavage to produce methyl
radicals and acetone.19
(11) Maier, G.; Reisenauer, H. P.; Schwab, W.; Ca´rsky, P.; Hess, B.
A., J r.; Schaad, L. J . J . Am. Chem. Soc. 1987, 109, 5183. Maier, G.;
Preiss, T.; Reisenauer, H. P.; Hess, B. A., J r.; Schaad, L. J . J . Am.
Chem. Soc. 1994, 116, 2014.
(12) Seburg, R. A.; Patterson, E. V.; Stanton, J . F.; McMahon, R. J .
J . Am. Chem. Soc. 1997, 119, 5847-5856.
(13) Stanton, J . F.; DePinto, J . T.; Seburg, R. A.; Hodges, J . A.;
McMahon, R. J . J . Am. Chem. Soc. 1997, 119, 429.
(14) Breidung, J .; Bu¨rger, H.; Ko¨tting, C.; Kopitzky, R.; Sander, W.;
Senzlober, M.; Thiel, W.; Willner, H. Angew. Chem., Int. Ed. Engl.
1997, 36, 1983.
(15) Newman, M. S.; Okorodudu, A. O. M. J . Org. Chem. 1969, 34,
1220. Stang, P. J . Chem. Rev. 1978, 78, 383.
(29) Bode, B. M.; Gordon, M. S. J . Mol. Graphics Modell. 1998, 16,
133.
(30) 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.
(31) See, for example, the following: Bally, T. In Radical Ionic
Systems (Topics in Molecular Organization and Engineering); Lund,
A., Shiotani, M., Eds.; Kluwer: Dordrecht, 1991; Vol. 6, pp 3-54.
(16) Other photochemical precursors can be found in the following:
Gilbert, J . C.; Weerasooriya, U.; Giamalva, D. Tetrahedron Lett. 1979,
48, 4619. Gilbert, J . C.; Luo, T. J . Org. Chem. 1981, 46, 5237. Gilbert,
J . C.; Butler, J . R. J . Am. Chem. Soc. 1970, 92, 7493.
(17) Adam, W.; Ruckta¨schel, R. J . Chem. Phys. 1971, 93, 557.
(18) Chapman, O. L.; Wojtkowski, P. W.; Adam, W.; Rodriguez, O.;
Ruckta¨schel, R. J . Am. Chem. Soc. 1972, 94, 1365-1367.

Dimethylvinyldiene
J . Org. Chem., Vol. 66, No. 1, 2001 289
Ta ble 1. Rela tive En er gies (in k ca l/m ol) of Dim eth ylvin ylid en e (5) Sin glet a n d Tr ip let Sta tes by Differ en t Qu a n tu m
Ch em ica l Meth od sa
method
5S
5T
∆EST
HF/6-31G(d)
MP2/6-31G(d)
MP2/6-311+G(2d,p)
BLYP/6-31G(d)
BLYP/6-311+G(2d,p)
BLYP/cc-pVTZ//BLYP/6-31G(d)
B3LYP/6-31G(d)
B3LYP/6-311+G(2d,p)
B3LYP/cc-pVTZ//B3LYP/6-31G(d)
CCSD(T)/cc-pVTZ//B3LYP/6-31G(d)
-155.897782535
-155.325245973
-155.459868284
-155.808033637
-155.864446076
-155.873120493
-155.897782536
-155.950018034
-155.959075875
-155.602948790
-155.826004495b
-155.251702809c
-155.381212492c
-155.735447255d
-155.788282013
-155.792895414
-155.826420731
-155.874954653
-155.884580361
-155.530680150
45.0
46.1
49.4
45.5
47.8
50.3
44.8
47.1
46.7
45.3
a
b
All structures corresponded to minima. Spin contamination for this spin state was 2.397. c Spin contamination for this spin state
d
was ∼2.33-2.34. Values of ∼2.03-2.05 were obtained at all DFT levels.
precursor was stored in a special U-type glass tube at room
temperature and connected to an argon line and a vacuum
chamber. To deposit the sample on the IR salt plate, ultrahigh
purity argon was passed through the U-tube, and the gaseous
mixture of the precursor and argon was deposited at 14 K on
the surface of a CsI window within the closed-cycle cryogenic
system. The argon matrix was kept at 14 K during photolysis.
UV-vis spectra were measured using a Lambda 6 UV-vis
spectrophotometer, and IR spectra were recorded on a Perkin-
Elmer 2000 FT-IR spectrometer with 2 cm-1 resolution.
Precursor irradiation was performed using a Rayonet photo-
reactor with 254 nm bulbs.
Syn th esis. 1. Gen er a l. NMR spectra were recorded on a
Varian Gemini-300BB NMR spectrometer in CDCl3, unless
reported otherwise, and are reported in ppm (δ) downfield from
TMS. Infrared spectra were measured on a Perkin-Elmer 1605
FT-IR spectrometer. Melting points were obtained on either a
Mel-Temp or Thomas-Hoover melting point apparatus in open
capillary tubes and are uncorrected. Solvents (Aldrich or
Fisher, reagent grade) were used as received unless otherwise
specified.
2. Isop r op ylid en e Ma lon ic Acid . A two-necked 100 mL
round-bottom flask was charged with KOH (2.3414 g, 0.042
mol) in a solution of 95% ethanol (5.0 mL) and water (1.5 mL).
To this stirred solution, diethyl isopropylidene malonate (2.00
g, 0.010 mol) in 10 mL ethanol was added dropwise under
nitrogen, and the resulting solution was refluxed overnight.
Upon cooling to 0 °C, concentrated HCl was added, making
sure the temperature in the reaction flask never went above
0 °C, until a pH of 4 was reached. This solution was extracted
with ether, and the ether layers were dried over MgSO4 and
evaporated under reduced pressure to yield a yellow solid.
Recrystallization of the yellow solid from chloroform afforded
a white solid (0.605 g, 42%; mp 168 °C, lit. 169 °C):32 1H NMR
(acetone-d6, δ) 1.99 (s); 13C NMR (acetone-d6, δ) 22.47, 125.81,
153.43, 166.57, 206.30; IR (1715 cm-1).
3. Isop r op ylid en e Ma lon yl Ch lor id e. To a 50 mL round-
bottom flask containing 0.2917 g of isopropylidene malonic acid
was added ∼20 mL of SOCl2, and the reaction was allowed to
stir overnight under N2 at room temperature. The excess SOCl2
was removed under vacuum, leaving behind a yellow oil
(0.2919 g, 79.7%). The oil was dissolved into pentane in a
preweighed vial, transferred to the reaction flask, and used
without purification: 1H NMR (CDCl3, δ) 2.16 (s); IR (1772
cm-1).
dried over Na2SO4, filtered, and evaporated under reduced
pressure to yield a colorless oil (0.1526 g, 32.8%). Purification
was accomplished by column chromatography (30% ethyl
acetate/hexane): 1H NMR (CDCl3, δ) 2.14 (s), 1.34 (s); 13C NMR
(acetone-d6, δ) 23.68, 26.12, 38.68, 68.11, 84.13, 117.55, 161.40;
IR (1776 cm-1).
Resu lts a n d Discu ssion
Spin-state analyses of simple vinylidenes (i.e., 1, R )
H) have indicated the singlet state is substantially more
stable than the triplet.6 Previous computational studies
regarding vinylidene spin states did not include alkyl
substitution (i.e., 5), so we decided to investigate the
effects of alkyl substitution. Three possible conformations
of 5 were thus considered: (1) an “out-out” C2v geometry
where the in-plane hydrogens face away from one other
and toward the carbene carbon (exo to the CH3-C-CH3
bond angle); (2) an “in-out” Cs geometry where one
methyl group was rotated 180° so that one C-H bond
points inward and one outward; and (3) an “in-in” C2v
geometry where both in-plane C-H bonds point inward
toward one another.
The Cs “in-out” geometry was found at all theoretical
levels to have one imaginary vibration corresponding to
a rotation of one methyl group, and the C2v “in-in”
geometry was found to have two imaginary vibrations.
One vibration corresponds to a “conrotatory” double
rotation and the second corresponds to a “disrotatory”
double rotation, both vibrations leading to the C2v “out-
out” structure. All calculations were therefore performed
with the C2v “out-out” geometry, which was found to
correspond to a minimum (i.e., no imaginary frequencies)
on the potential energy surface at all levels of theory.
Relative energies of the optimized out-out singlet and
triplet carbenes 5S and 5T were computed at the Har-
tree-Fock, Møller-Plesset, density functional, and
coupled-cluster (CCSD(T)) levels and are tabulated in
Table 1.
4. ter t-Bu tyl Isop r op ylid en e P er ester . Isopropylidene
malonyl chloride (0.2919 g, 0.00161 mol) was dissolved in
pentane and cooled under N2 to 0 °C with the use of a cryo-
cooler. To this solution was added dropwise over a 2 h period
a pentane solution of 90% tert-butyl hydroperoxide (0.4103 g,
0.00455 mol) and pyridine (0.3602 g, 0.00455 mol). The
addition caused the coloration of the clear solution to become
orange, and the temperature was cooled to -30 °C and stirred
for 3 days. The reaction mixture was then filtered to remove
pyridinium hydrochloride and washed with 10% HCl, satu-
rated NaHCO3, and distilled water. The organic layer was
Consistent with previous calculations,6 we find a strong
preference (∼45-50 kcal mol-1) for the singlet state at
all levels of theory used in the spin-state calculations.
The basis for this preference is the energetic difference
between the occupied orbitals in the two states. The
HOMO in the singlet state is localized on the carbene
(32) Meyenberg, J . Chem. Ber. 1895, 28, 786.

290 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
repulsion between the methyl groups and the coplanar,
singly occupied p-orbital in the triplet state. The net
result in either case is a shorter CdC bond in the singlet.
The progression in the C4-C1-C3 bond angles (CH3-C-
CH3: 120.1° in the singlet and 116.6° in the triplet) is
consistent with repulsion of the C-CH3 bonds with the
in-plane, singly occupied p-orbital in the triplet state. The
change in the C-CH3 bond distances is not so noticeable
and is 1.517 Å for the singlet and 1.520 Å for the triplet.
The larger basis set calculations at B3LYP/6-311G(2d,p)
show a slightly more pronounced change, so the CdC
distances are 1.297 and 1.320 Å for the singlet and triplet
states, respectively. The C4-C1-C3 bond angles are
nearly equivalent to the B3LYP/6-31G(d) results and are
120.4° for the singlet and 116.4° for the triplet state.
Ma tr ix Isola tion . Diperoxyesters have been reported
as thermal11,17 and photochemical18 precursors for car-
benes, indicating these molecules might also prove to be
suitable precursors for dialkylvinylidenes. Thus, we
prepared diperoxyester 6 from isopropylidene diethyl-
malonate by the route shown in Scheme 1. Matrix-
isolation experiments were performed in both argon and
Nujol matrixes; similar results were obtained with both
matrixes, and the results in the Ar matrixes will be
discussed here. A variety of irradiation wavelengths were
initially attempted using filters and a broad-band Xe arc
lamp; the results reported here were acquired by lowering
the displex head, after cooling to 14 K, into a Rayonet
reactor containing sixteen 254 nm bulbs.
Irradiation of 6 at 14 K for 1 min in an Ar matrix
results in major changes in the IR spectrum (Figure 3).
Precursor depletion is predominantly noted at 1776,
1179, 1078, and 1003 cm-1, corresponding to the precur-
sor CdO, C-O, CdC, and CdC stretches, respectively.
New absorptions in the spectrum are observed at 2340,
2139, 2124, 2102, 1924, 1721, 1695, 1473, 1457, 1363,
1148, 1121, 1050, 663, and 651 cm-1. Increasing the
irradiation time to 5 min resulted in notable intensity
increases in the absorptions at 2340, 2139, 2124, 2102,
1721, 1473, 1457, 1363, 1121, 663, and 651 cm-1, while
the absorptions at 1924, 1695, 1148, and 1050 cm-1 did
not change. Similar changes were observed upon further
irradiation (20 min) at 254 nm. No other absorptions
were observed in either set of spectra.
When 6 was irradiated for 5 min with 254 nm light in
an argon matrix doped with 9% CO (2139 cm-1), a similar
decrease in precursor intensities was observed (Figure
4). Interestingly, the number of absorption bands changed
so that only absorptions at 2340, 2139, 2124 (shoulder),
2102, 1924, 1706, 1695, 1148, 1050, 665, and 651 cm-1
were observed. Increasing the irradiation time to 20 min
resulted in an increase in the intensity of the absorptions
at 2340, 2124 (shoulder), 2102, 1706, 665, and 651 cm-1
(Figure 4) and a decrease in the intensity of the band at
2139 cm-1 but did not affect the other absorptions (1924,
1695, 1148, and 1050 cm-1).
In ter p r eta tion of IR Resu lts. The IR spectra shown
in Figure 3 are somewhat complicated due to the pres-
ence of the multiple photoproducts that are produced
upon irradiation of 6. From the spectra presented in
Figures 3 and 4, it is clear that increasing irradiation
times results in the preferential growth of certain ab-
sorptions. In particular, we found extended irradiation
in an Ar matrix produced increasingly larger absorptions
at 2340, 2139, 2124, 2102, 1721, 1363, 1121, 663, and
651 cm-1, while increasing irradiation times in the Ar/
F igu r e 1. An orbital diagram showing the in-plane unhy-
bridized p-orbital and the sp-hybridized orbital, at the carbene
carbon atom, in 5S and 5T. Hyperconjugation is found to be
more important to the singlet state than the triplet.
F igu r e 2. Bond lengths and C2-C1-C3/4 bond angle for
carbenes 5S and 5T. The values given are for geometries
calculated at the B3LYP/6-31G(d) and BLYP/6-31G(d) (in
parentheses) levels of theory. Bond lengths are given in
angstroms and are indicated in the figure by full arrows; bond
angles are indicated using half-arrows.
carbon and is sp-hybridized. The LUMO is also localized
on the carbene carbon and is an unhybridized p-orbital,
oriented in the molecular plane (Figure 1). This electronic
arrangement is different from “typical” carbenes due to
the change in the nature of the occupied nonbonding
orbital from sp2-hybridization to sp-hybridization. This
change results in a lower energy HOMO for vinylidenes
relative to normal carbenes, and a subsequent change
in the relative energies of the HOMO and LUMO. Thus,
the doubly occupied HOMO in the singlet is substantially
lower in energy than the unoccupied p-orbital. In the
triplet state both of these orbitals are occupied, and
placement of one of the electrons into the p-orbital
substantially increases the energy of this state relative
to the singlet.
Increasing basis set size results in small increases in
∆EST (Table 1). This basis set effect probably results from
the interaction in the singlet state of the unhybridized
p-orbital on the carbene carbon with the H3C-C σ-bonds
by hyperconjugation. No similar interaction is observed
in the triplet state. Increasing the amount of electron
correlation, on the other hand, has the opposite effect and
lowers the triplet state relative to the singlet. The effect
is only slight, however, and probably is a result of the
fact that electron-correlation effects are most important
when the electrons are delocalized over several atoms in
the molecule. Because the only pi-bond in the molecule
(the CdC bond) is orthogonal to the unhybridized p-
orbital, delocalization is not possible and large contribu-
tions to the energies of the singlet and triplet states are
not expected with increasing correlation. Thus, only a
small change in the relative energies is observed going
from HF/6-31G(d) to CCSD(T)/cc-TZP//B3LYP/6-31G(d).
The singlet and triplet carbenes differ primarily in the
CdC bond length due to the different occupancies of the
unhybridized p-orbital (Figure 2). Singlet 5S shows a
slightly shorter CdC bond distance of 1.309 Å compared
to 1.328 Å in the triplet, consistent with either a
hyperconjugative stabilization between the C-CH3 σ-or-
bitals and the empty p-orbital in the singlet state or

Dimethylvinyldiene
J . Org. Chem., Vol. 66, No. 1, 2001 291
Sch em e 1
CO matrix results only in stronger signals at 2340, 2124,
2102, 1706, 663, and 651 cm-1. The absorptions at 1924,
1695, 1148, and 1050 cm-1 did not appear to change with
increasing irradiation times in either matrix. The ab-
sorption at 2340 cm-1 results from CO2 in the matrix.
Wentrup and Lorencak33 have shown residual CO2 in the
displex head generally deposits on the colder parts of the
displex coldfinger upon cooling to ∼10 K and not on the
IR windows; the presence of CO2 on the IR windows is
generally observed only during warm-up cycles. The
growth of the absorption at 2340 cm-1 in these experi-
ments occurs with increasing irradiation times, and its
presence can therefore be attributed to the production
of CO2 during irradiation, consistent with precursor
depletion. That is, irradiation of 6 results in cleavage of
a peroxide bond followed by decarboxylation. The inten-
sity of this band was shown to increase with the con-
comitant growth of the negative absorptions associated
with precursor depletion. The absorption observed at
2139 cm-1 in both matrixes is due to CO. The observation
that several of the absorptions increase with increasing
irradiation times, while others do not, is indicative that
more than one photoproduct is produced and that one or
more of these photoproducts in turn undergo further
photoreaction. The introductory discussion of peroxyester
photochemistry and vinylidene chemistry indicates that
irradiation of 6 potentially produces products 9-13
(Scheme 2). Due to the extremely labile nature of the
precursor, and the obvious dependence of product forma-
tion on irradiation time, we performed frequency calcula-
tions at the B3LYP/6-31G(d) level to aid in deconvolution
of the admittedly complicated IR spectra. These calcula-
tions, together with our rationalization of the IR spectra
absorptions and a proposed mechanism for formation of
the products responsible for the absorptions, are de-
scribed below.
The computed vibrational frequencies and intensities
of 5S and potential photoproducts 9-13 are shown in
Tables 2 and 3. The experimental spectrum obtained
upon 254 nm photolysis of 6 in an Ar matrix (20 min) is
shown together with the calculated IR spectrum of 5S
in Figure 5 (expanded to maximize the relevant spectral
features). The strongest computed frequency for 5 occurs
at 1701 cm-1, a value that corresponds reasonably well
with the observed absorption at 1721 cm-1 in the matrix
spectrum. This absorption is consistent with the “carbo-
nyl-like” structure of the carbene and relatively close in
energy to the CdC: stretch observed for F2CdC: by
Sander et al. (1672 cm-1).14 Four of the remaining five
computed frequencies have analogous absorptions in the
experimental spectrum (1473, 1457, 1363, and 1121 cm-1
observed vs 1459, 1449, 1381, and 1055 cm-1 computed),
although the correlation between the computed and
experimental values is not as good in some cases. The
absorption at 1363 cm-1 is quite close to the calculated
absorption at 1381 cm-1 (a methyl group umbrella
pucker), although the intensity (relative to the 1721 cm-1
absorption) is akin to the predicted frequencies of 1459
or 1449 cm-1, while the absorption at 1120 cm-1 differs
from the predicted IR band at 1055 cm-1 (an in-plane
C-CH3 symmetric rock) by 65 cm-1. Nonetheless, as we
will show later, these absorptions can also be attributed
to 5.
Of the remaining absorptions found in Figure 3, the
absorption at 2340 cm-1 is readily attributed to CO2 and
that at 2139 cm-1 is attributed to CO. However, the
absorptions observed at 2124, 2102, 1924, 1695, 1148,
1050, 663, and 651 cm-1 must result from either a
product(s) produced from an intermediate generated prior
to carbene formation or from reaction of the carbene
through an intramolecular rearrangement or an inter-
molecular pathway. The possible rearrangement products
methylcyclopropene (12) and 2-butyne (13) are ruled out
both by frequency calculations (see Table 3) and by
comparison to the published absorptions.35,36 We have
therefore identified 9-11 as possible sources of these
absorptions. Figure 6 shows a comparison between the
computed spectra for 9-11 with the experimentally
obtained spectrum.
Chapman and Adam18 reported formation of R-lactone
8 upon irradiation of the cyclic peroxyester 7 at 77 K.
These lactones were found to have CdO absorptions in
the 1895-1935 cm-1 region, the exact position of which
depended upon the substitution at the R-carbon. Our
calculations on 8a and 8b at the B3LYP/6-31G(d) level
(Figure 7) indicate the CdO stretches occur at 1932 and
1966 cm-1, respectively.
These values compare favorably with Chapman’s ex-
perimental data for 8a (1900 cm-1) and 8b (1935 cm-1),18
providing a benchmark for comparison to 9. The IR
spectra generated upon photolysis of 6 in both Ar and
Ar/CO matrixes (Figures 3 and 4) show an absorption at
1924 cm-1 that is consistent with Chapman’s observed
results for the coupled CdC-CdO stretch in R-lactones.
The calculated value for 9 assigned to this vibration is
somewhat high at 2013 cm-1 but is nonetheless consis-
tent with the trend exhibited by 8a -b, for which the
(33) Wentrup, C.; Lorencak, P. J . Am. Chem. Soc. 1988, 110, 1880-
1883.
(34) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391-399.
(35) Butler and Newbury (Butler, I. S.; Newbury, M. L. Spectrochim.
Acta 1980, 36A, 458.) have reported IR absorptions (with the corre-
sponding relative intensities in parentheses) at 2963 (vs, br), 2925 (vs),
2862 (s), 2743 (m), 2534 (w), 2468 (w, br), 2414 (w), 2059 (m), 1840
(m), 1450 (vs, br), 1372 (m), 1251 (m, br), 1142 (s), 1037 (s, br), 803
(w), and 607 (w) cm-1 in the solid state (77 K).
(36) Mitchell and Merritt (Mitchell, R. W.; Merritt, J . A. Spectro-
chim. Acta 1969, 25A, 1881.) have reported IR absorptions (with the
corresponding relative intensities in parentheses) at 3130 (w), 2960
(s), 2916 (w), 2877 (s), 1780 (m), 1478 (w), 1437 (m), 1400 (vw), 2370
(w), 1320 (vw), 1151 (m), 1093 (vw), 1056 (m), 1026 (vs), 959 (m), 921
(m), 698 (s), and 667 (w) cm-1 in the solid state (77 K), where w )
weak, m ) medium, s ) strong, vs ) very strong, and br ) broad. See
also the following: Wiberg, K. B.; Nist, B. J . J . Am. Chem. Soc. 1961,
83, 1226.

292 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
F igu r e 3. Experimental difference IR spectra obtained upon irradiation of 6 with 254 nm light in an Ar matrix at 14 K. Irradiation
times were 1 min (a), 5 min (b) and 15 min (c); positive absorptions (up) represent growth of new absorptions, negative absorptions
(down) represent precursor depletion. (a) IR spectrum from 2500 to 500 cm-1. (b) IR spectrum expanded to 2200-1000 cm-1
.
calculated values are also high by ∼30 cm-1. Three other
absorptions that can be attributed to 9 are observed
under both matrix conditions, at 1695, 1148, and 1050
cm-1 (predicted: 1710, 1156, and 1051 cm-1, correspond-
ing to the CdC, asymmetric bending CH3 mode, and an
in-plane deformation of the lactone ring, respectively).
The other predicted absorptions for 9 are much weaker
in intensity (see Table 3) and are therefore not expected
to be observed. The exception is the predicted absorption
at 707 cm-1, which should be comparable in intensity to
the weak absorption at 1148 cm-1. We do not at present
have an explanation for the absence of this absorption.
Two absorptions are visible at 663 and 651 cm-1, but the
intensity of these bands increases with increasing ir-
radiation times. Because the intensities of the absorp-
tions already attributed to 9 do not increase with
increasing irradiation, we are not able to assign these
absorptions to 9.
Chapman et al. also demonstrated R-lactones such as
8 could undergo photodecarbonylation to yield ketones.

Dimethylvinyldiene
J . Org. Chem., Vol. 66, No. 1, 2001 293
F igu r e 4. Experimental difference IR spectra obtained upon irradiation of 6 with 254 nm light in an Ar/CO (9% CO) matrix at
14 K. Irradiation times were 5 min (a) and 20 min (b); positive absorptions (up) represent growth of new absorptions, negative
absorptions (down) represent precursor depletion. (a) IR spectrum from 2500 to 500 cm-1. (b) IR spectrum expanded to 2200-
1000 cm-1. The absorption at 2124 cm-1 is observed as a shoulder growing into the 20 min spectrum. The growth of the band at
2102 cm-1 is minor in this particular spectrum.
The analogous reaction in R-lactone 9 is the formation
of dimethyl ketene, 10. Dimethyl ketene has an intense
absorption at 2130 cm-1 but has only minor absorp-
tions elsewhere.37 Our calculated values for 10 (Table
3) are in good agreement with the experimental val-
ues. On the basis of the available experimental and
computational data, we have assigned the absorption at
2102 cm-1 to dimethyl ketene 10. Consistent with this
observation, we have observed a growth in the signal
at ∼2139 cm-1 in an Ar matrix with increasing irradia-
tion times, corresponding to the loss of CO by 9 to form
10.
(37) Parker, J .K.; Davis, S. R. J . Phys. Chem. A 1999, 103, 7280-
7286. These authors have reported absorptions (with the corresponding
relative intensities in parentheses) at 2902 (1.3), 2130 (100), 2065 (1.5),
1455 (0.5), 1337 (0.6), and 729 cm-1 (0.5).

294 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
Sch em e 2
Ta ble 2. Exp er im en ta l a n d Ca lcu la ted (B3LYP /6-31G(d ))
In fr a r ed Absor p tion s for Dim eth ylvin ylid en e 5
addition to producing 11, in the disappearance of the
absorptions attributed to 5S. As shown in Figure 4, the
absorptions we have attributed to 5S (1721, 1473, 1457,
1363, 1120 cm-1) have completely disappeared under
these conditions, while the absorption at 2124 cm-1 has
increased. In addition, we observe a depletion of the CO
peak at 2139 cm-1 in the IR difference spectra. These
results are interpreted as evidence for the chemical
trapping of 5S and add further support for the presence
of 5S in the matrix.
exptla
calcdb
3013 (12)
2982 (37)
2924 (26)
1701 (36)
1459 (19)
1449 (13)
1381 (2)
1055 (10)
762 (3)
2973 (s)
1721 (s)
1473 (w)
1457 (w)
1363 (m)
1121 (m)
One of the inevitable byproducts of tert-butylperoxide
homolysis is the formation of the tert-butoxy radical,
which can undergo R-cleavage and lose a methyl group
to yield acetone (absorption ∼1710 cm-1).33 Because the
position of our assigned CdC: stretch for 5 is at 1721
cm-1, we sought to confirm that the appearance of this
peak was not simply due to the presence of acetone.
Literature reports41 on the gas-phase decomposition of
the tert-butoxy radical indicate Arrhenius parameters of
logA ≈ 14.6-14.9 and Ea ≈ 15-16 kcal mol-1 for the C-C
bond cleavage reaction. More recent results using the 248
nm output of a KrF excimer laser to generate the radical
from tert-butyl hypochlorite in a molecular beam42 showed
Ea ≈ 20 kcal mol-1, indicating a slight temperature
dependence to fragmentation. Extrapolation of these
results to 14 K indicates kcleavage should be quite small.
However, the radicals generated in the molecular beam
experiments were translationally hot, with enough in-
ternal energy to undergo bond cleavage. To determine
whether bond cleavage to form acetone is likely under
the experimental conditions used in this work, we irradi-
ated tert-butylperoxide under conditions analogous to
those used for irradiation of 6 (i.e., 254 nm light, 14 K,
1-30 min irradiation time). tert-Butylperoxide was cho-
a
Frequencies in cm-1, relative intensities in parentheses; argon
matrix, 14 K. Frequencies in cm-1 (corrected, 0.9613 scaling
b
factor),34 absolute intensities in parentheses.
Wentrup33 and Dailey38 generated dimethylpropadi-
enone 11 under matrix-isolation conditions by flash
vacuum pyrolysis. The CdCdCdO band in these experi-
ments was reported at 2088 and 2115 cm-1, respectively,
and corresponds to a molecule having a bent geometry
(Cs symmetry). Calculations on 11 at the B3LYP/6-31G-
(d) level in C2v symmetry reveal a CdCdCdO absorption
at 2255 cm-1 (scaled, one imaginary frequency at 63i
cm-1), while calculations performed within Cs symmetry
(corresponding to a bent geometry) result in an absorp-
tion at 2154 cm-1 that corresponds reasonably well to the
experimentally determined absorptions. Thus, we assign
the absorption observed at 2124 cm-1 to 11. The matrix
data obtained previously33,38 differs from our own by 36
and 9 cm-1. The discrepancies among the three sets of
results can be attributed to matrix effects. Interestingly,
this absorption was found to be present in both the Ar
and Ar/CO matrixes, increased in intensity in both cases
with longer photolysis times, and was shown to be more
intense in the Ar/CO matrix (Figure 4). Brahms and
Dailey38 reported additional weaker absorptions in their
sen for this control experiment because its bond dissocia-
43
tion energy (36-39 kcal mol-1
)
is close to the bond
IR spectrum, at frequencies of 1777, 1765, and 1362 cm-1
.
dissociation energy of tert-butyl benzoylperoxide (34.1
kcal mol-1),44 which is a reasonable model for 6. The
acetone carbonyl stretch generated under these condi-
tions is found at 1715 cm-1 and is quite weak (Supporting
Information). While this absorption is relatively close to
that assigned to 5S (1721 cm-1) and was initially wor-
risome, we would like to point out several facts that lead
us to discount formation of this molecule as being the
The frequencies calculated here for 11 are shown in Table
3 and indicate that these absorptions are expected to be
less intense than the 2124 cm-1 absorption by at least a
factor of 20. Given the intensity of signal observed by us
at 2124 cm-1, we are not surprised that these additional
absorptions were not detected under the present condi-
tions.
The reaction of carbenes with CO under matrix-
isolation conditions has been used previously to provide
evidence for carbene formation.39,40 We reasoned that
irradiation of 6 in the presence of CO should result, in
(41) Nakamura, T.; Busfield, W. K.; J enkins, I. D.; Rizzardo, E.;
Thang, S. H.; Suyama, S. J . Org. Chem. 2000, 65, 16-23. Batt, L.;
Hisham, M. W. H.; Mackay, M. Int. J . Chem. Kinet. 1989, 21, 535.
Heicklen, J . Adv. Photochem. 1988, 14, 177.; Batt, L.; Robinson, G. N.
Int. J . Chem. Kinet. 1982, 14, 1053. Birss, F. W.; Danby, C. J .;
Hinshelwood, C. Proc. R. Soc. London, Ser. A 1957, 239, 154.
(42) Thelen, M.-A.; Felder, P.; Frey, J . G.; Huber, J . R. J . Phys.
Chem. 1993, 97, 6220-6225.
(38) Brahms, J . C.; Dailey, W. P. Tetrahedron Lett. 1990, 31, 1381-
1384.
(39) Seburg, R. A.; McMahon, R. J . J . Am. Chem. Soc. 1992, 114,
7183.
(43) Cubbon, R. C. P. Prog. React. Kinet. 1970, 5, 29.
(44) Tuleen, D. L.; Bentrude, W. G.; Martin, J . C. J . Am. Chem. Soc.
1963, 85, 1938.
(40) Zhu, Z.; Bally, T.; Stracener, L. L.; McMahon, R. J . J . Am. Chem.
Soc. 1999, 121, 2863.

Dimethylvinyldiene
J . Org. Chem., Vol. 66, No. 1, 2001 295
Ta ble 3. Exp er im en ta la a n d Ca lcu la ted b (B3LYP /6-31G(d )) In fr a r ed Absor p tion s
9
10
11
12c
13d
exptl
calcd
exptl
calcd
exptl
calcd
calcd
calcd
3019 (13)
2970 (32)
2924 (35)
2013 (521)
1710 (102)
1459 (18)
1452 (6)
3006 (20)
2952 (62)
2910 (54)
2907 (31)
2131 (635)
1464 (12)
1455 (14)
1396 (15)
1366 (11)
1174 (3)
3042 (9)
2975 (33)
2918 (72)
1452 (11)
1387 (10)
1042 (3)
2999 (20)
2969 (84)
2920 (20)
2917 (90)
1805 (11)
1496 (10)
1453 (8)
1449 (7)
1059 (11)
1039 (31)
916 (13)
704 (26)
3038 (12)
2964 (27)
2919 (27)
2154 (1252)
1714 (19)
1458 (17)
1452 (9)
1363 (18)
1185 (62)
686 (8)
1924 (vs)
1695 (m)
2102 (s)
2124 (s)
1379 (3)
1148 (w)
1050 (m)
1156 (60)
1051 (174)
934 (7)
1005 (8)
710 (4)
707 (53)
614 (20)
635 (22)
628 (15)
a
b
Frequencies in cm-1, relative intensities in parentheses; argon matrix, 14 K. Frequencies in cm-1 (corrected, 0.9613 scaling factor),34
absolute intensities in parentheses. c See ref 35 for experimental values. See ref 36 for experimental values.
d
F igu r e 5. Comparison between the computed B3LYP/6-31G(d) IR spectrum of 5 and the experimental difference IR spectra
obtained upon irradiation of 6 with 254 nm light (20 min) in an Ar matrix at 14 K.
carrier of the 1721, 1473, 1457, 1363, and 1120 cm-1
absorptions. First, under the irradiation conditions used
in the control experiment, we observed the production
of only minor amounts of acetone, despite substantial
depletion of tert-butylperoxide (see Supporting Informa-
tion). While this result at first appears surprising com-
pared to the molecular beam experiments, R-cleavage
may not be efficient under the conditions used here due
to differences in the recoil energies available to each
radical upon O-O vs O-Cl bond scission. Second, O-O
bond scission in 6 proceeds after an initial redistribution
of energy from the nfπ* excited state, and so the amount
of remaining energy, that is, the internal energy available
for accessing the transition state for cleavage to acetone,
will be low. This reasoning necessarily implies the
amount of acetone produced upon irradiation of 6 should
be less than that observed in the control experiment.
Third, the position of the carbonyl absorption is ∼5 cm-1
lower in energy than the band assigned to 5, within the
2 cm-1 resolution of the spectrometer used in this work.
Fourth, matrix EPR experiments45 in the 12-77 K
temperature range on a series of tert-butyl phenyl per-
oxyoxlates, a molecule that also yields the tert-butoxy
radical in a mechanism similar to 6, yielded no signals
that could be identified with the characteristic signature
of the methyl radical. Fifth, and most important, the
spectra obtained in the Ar/CO matrix are completely
devoid of absorptions in the 1710-1730 cm-1 region.
Generation of acetone is not precluded by inclusion of CO
in the matrix, and so if acetone is the carrier of these
absorptions, these bands should still be present in Figure
4. The fact that we do not observe absorptions at any of
these frequencies is a strong indicator that acetone is not
(45) Modarelli, D. A. Ph.D. Dissertation, The University of Mas-
sachusetts, Amherst, MA, 1991.

296 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
F igu r e 6. Comparison between the computed B3LYP/6-31G(d) IR spectra of 9-11 and the experimental difference IR spectra
obtained upon irradiation of 6 with 254 nm light (20 min) in an Ar matrix at 14 K.
the carrier of the 1721 cm-1 absorption. We have observed
the growth of a band at 1706 cm-1, which may be due to
acetone; however, if this absorption results from acetone,
it is clearly not the 1721 cm-1 absorption we attribute
to 5.
We were somewhat surprised not to observe 2-butyne
in the matrix experiments. Dailey et al.38 identified
2-butyne as a photoproduct from irradiation at 10 K of a
sample of dimethyl ketene (10), implying the interme-
diacy of 5S. Due to the small concentration of 10
generated under the conditions used in the matrix
experiments reported here, the concentration of 2-butyne
may simply be too small to detect by IR.
An alternative route of formation of 9 (Scheme 3, path
B) involves the loss of a second tert-butoxy fragment from
the acyloxyl monoradical prior to decarboxylation by
scission of the second O-O bond. The diacyloxyl diradical
generated in this process can then monodecarboxylate
to yield the acyloxyl-vinyl diradical 14 and, ultimately,
9. Regardless of the exact mechanism for formation,
Chapman’s work with R-lactones18 (i.e., 8) has indicated
a sensitivity to irradiation, with decarbonylation to yield
ketones a predominant path. Assuming analogous be-
havior to 8, photoinduced decarbonylation of 9 produces
dimethyl ketene (10). This result is consistent with the
observed results under extended irradiation times, where
we have observed little change in the absorptions at-
tributed to 9 (1924, 1695, 1148, and 1050 cm-1), while
the absorption attributed to 10 (2102 cm-1) and CO (2139
cm-1) increase with increasing irradiation times. These
results are internally consistent: irradiation of 6 pro-
duces a diradical intermediate that ring-closes to form
lactone 9, while irradiation of 9 produces 10. Due to its
probable instability under ambient conditions, we have
not attempted to isolate 9 by another route.
On the basis of the results described above, we propose
a mechanism describing the photolysis of 6 that accounts
for the presence of 5 and 9-11 (Scheme 4). Irradiation
at 254 nm produces carbene 5, probably by first forming
acyloxyl-vinyl diradical 14 upon the loss of two tert-
butoxy radicals and CO2. This point represents the
branching point in the mechanism. Loss of CO2 produces
vinylidene 5 (Scheme 4, path A), while cyclization of the
diradical produces R-lactone 9 (Scheme 4, path B).
Irradiation46 of 9 results in the loss of CO, producing
dimethylketene, 10. We suspect formation of dimethyl-
propadienone 11 occurs upon capture of CO from the
We have not been able to positively assign several
absorptions (1623, 660, and 651 cm-1), either by com-
parison with the current literature or by comparison to
computed structures.
We have thus far assigned structures 9-11 as the
photoproducts resulting from irradiation of 6 under
matrix-isolation conditions. Reaction of the carbene with
the argon matrix is clearly unlikely, and so these absorp-
tions must therefore arise from either (a) a product
produced through a noncarbenic path from an intermedi-
ate generated prior to carbene formation, or (b) reaction
of the carbene with CO2 or CO (produced during irradia-
tion of 6). We will see that both processes occur. Irradia-
tion (254 nm) of 6 results in preparation of an nfπ*
excited state that leads to cleavage of the peroxide O-O
bond (Scheme 3). Assuming that fragmentation of 6 is a
stepwise process, the resulting acyloxyl monoradical can
either (a) decarboxylate or (b) undergo a secondary O-O
bond cleavage. Decarboxylation of the acyloxyl mono-
radical produces a vinyl monoradical (Scheme 3, path A).
Scission of the second O-O peroxide bond produces an
acyloxyl-vinyl diradical (14) that can cyclize to produce
the R-lactone (9).
(46) Presumably the heat-induced loss of CO2 occurs at higher
temperatures, although we have not verified that prediction.

Dimethylvinyldiene
J . Org. Chem., Vol. 66, No. 1, 2001 297
Sch em e 3
in an argon matrix doped with 9% CO, the main absorp-
tion bands observed, aside from those attributed to CO2,
were those at 1924 and 2124 cm-1, consistent with
formation of 9 and 11, respectively (Table 3). Importantly,
no IR bands consistent with 5 were observed. Because
CO is known to trap carbenes under matrix-isolation
conditions,39,40 we surmise 5 is trapped to yield 11.
Formation of 9 occurs through a noncarbenic intermedi-
ate (Schemes 3 and 4), and so its formation is not
precluded by the presence of CO. Assignment of the other,
less intense absorptions for 9-11 is not possible at this
time. These absorptions are all calculated to be weaker
than the observed stretches by at least one order of
magnitude, and therefore, these absorptions are not
expected to be visible due to the low concentration of
these photoproducts.
F igu r e 7. Bond lengths and relevant bond angles for lactones
8a , 8a , and 9. The values given are for geometries calculated
at the B3LYP/6-31G(d) level of theory. Bond lengths are given
in angstroms and are indicated in the figure by full arrows;
bond angles are indicated using half-arrows.
Th eor etica l Con sid er a tion s of th e In tr a m olecu la r
Rea r r a n gem en ts of Dim eth ylvin ylid en e. The main
intramolecular rearrangement pathway of vinylidene is
the 1,2-H migration to form acetylene. The barrier to this
rearrangement has been estimated4,5,9 to be as high as
∼3 kcal mol-1 and should be quite rapid even at cryogenic
temperatures. The intramolecular rearrangement prod-
ucts of alkylvinylidenes are expected to be similar, and
alkyne formation is anticipated to be a major rearrange-
ment product. However, the competing insertion reaction
to form a cyclic alkene is also possible (Scheme 5). To
address the viability of these two rearrangement path-
ways for 5, we performed density functional calculations
at the B3LYP/6-31G(d) level for the rearrangement of 5
to the corresponding alkyl shift (12) and C-H insertion
(13) products shown in Scheme 5. The relevant structural
information and relative energies of these products are
shown in Figure 8 and Table 4. At the B3LYP/cc-pVTZ//
B3LYP/6-31G(d) level, 2-butyne (12) and 1-methyl-1-
cyclopropene (13) are more stable than 5 by ∼52 and ∼23
kcal mol-1, respectively. Alkyne formation is thus a more
energetically favorable intramolecular process than C-H
insertion, although this result may not be general due
to the strained nature of the cyclopropene ring, which is
matrix by carbene 5 (Scheme 4, path A). When 6 is
irradiated in a CO/Ar matrix, carbene 5 is trapped and
observation of its IR signals suppressed. Dimethylpro-
padienone 11 again is the product of this trapping
reaction; R-lactone 9 and dimethyl ketene 10 are pro-
duced under these conditions as well due to the different
route of their formation (path B).
Comparison with Wentrup33 and Dailey’s38 spectra
indicates the absorption at 2124 cm-1 is due to 11. We
have been unable to rationalize the formation of this
molecule from an intramolecular mechanism; however,
CO is produced under the experimental conditions by
photoinduced decarbonylation of 9 to form 10. We there-
fore surmise the CO produced under these conditions
reacts with carbene 5 to produce dimethylpropadienone,
11.39,40 Irradiating 6 at 14 K in an Ar matrix containing
9% CO confirms these conclusions, as the absorption
assigned to 11 increases in intensity and we do not
observe the absorption assigned to 5. The mechanism of
this addition is presumably similar to the addition of CO
to arylcarbenes under matrix-isolation conditions.40
Several additional observations support the assign-
ments of 9-11. When 6 was irradiated with 254 nm light

298 J . Org. Chem., Vol. 66, No. 1, 2001
Reed et al.
Sch em e 4
path was followed by the IRC method. The calculated
structures of TS1 and TS2 are shown in Figure 9. Single-
point energy calculations on these structures at the
B3LYP/TZP level show that TS1 and TS2 are 10.7 and
12.1 kcal mol-1 higher in energy than 5, respectively.
Examination of TS1 indicates an early transition state,
with the migrating methyl group not quite midway along
its migration axis when the transition state forms. The
C-C bond undergoing scission is stretched to 1.794 Å
(from 1.517 Å in the carbene), while the distance between
the two carbon atoms that form the new C-C bond is
1.821 Å; importantly, the migrating carbon atom must
still travel ∼2.3 Å until the equilibrium 2-butyne geom-
etry is attained. The methyl group that is not migrating
forms a 166.8° angle with the forming triple bond (Figure
9). A break in symmetry from Cs to C1 occurs along the
path, resulting from methyl group rotation prior to
migration. Although the migration along this path results
in formation of 2-butyne with C2h symmetry (the ground
state is found to have C2v symmetry), rotation around the
C-CH3 single bond has a nearly negligible energy
barrier.50 A second transition state ∼2 kcal mol-1 higher
in energy, and having two imaginary vibrations, was also
found along the reaction path.
F igu r e 8. Bond lengths and relevant bond angles for C4H6
rearrangement products 12 and 13. The values given are for
geometries calculated at the B3LYP/6-31G(d) level of theory.
Bond lengths are given in angstroms and are indicated in the
figure by full arrows; bond angles are indicated using half-
arrows.
Sch em e 5
The transition state corresponding to C-H insertion,
TS2, is also found rather early on the PES. TS2 is found
in Cs symmetry and exhibits a CdC bond distance of
1.322 Å, close to the 1.309 Å in 5S, and a distance of 1.842
Å between the carbon atoms that will form the new C-C
single bond. The energies of TS1, TS2, 12, and 13 at the
various computational levels used in this work are shown
in Table 4.
estimated using Benson equivalents47 to be ∼53.7 kcal
mol-1 less stable than an acyclic molecule.
Given that these reactions are both substantially
exothermic, it is therefore surprising that neither product
was observed in any quantity in the matrix isolation
experiments,48 yet intermolecular trapping products were
observed.49 To help explain the efficient intermolecular
trapping and lack of rearranged products, we examined
the transition states leading to the expected rearrange-
ment products. Transition state structures for the 5f12
(TS1) and the 5f13 (TS2) rearrangements were calcu-
lated using Schlegel’s26 STQN method, and the reaction
The computational results indicate the rate constants
for formation of 12 and 13 from 5 should be fairly slow
at 14 K. Although this assumption is consistent with our
experimental observations, it does not necessarily pre-
clude formation of either 12 or 13 from 5. The experi-
mental IR spectrum obtained upon irradiation of 6 is
quite complex, and the signals associated with 12 and
13 might simply be unobservable due to other absorp-
tions. However, 2-butyne35 has an absorption of medium
intensity at 1840 cm-1 while 1-methylcyclopropene36 has
an absorption at 1780 cm-1, also of medium intensity.
Both of these absorptions occur in a region of the IR that
is transparent to 5 and 9-11, and we are therefore
inclined to believe if either of these products is formed,
it is produced in such small amounts as to preclude
detection. We have argued that R-lactone 9 extrudes CO
(47) Benson, S. W.; Cruickshank, F. R.; Golden, D. M.; Haugen, G.
R.; O’Neal, H. E.; Rodgers, A. S.; Shaw, R.; Walsh, R. Chem. Rev. 1969,
69, 279.
(48) The lack of the presence of the IR bands associated with
methylcyclopropene and 2-butene is not necessarily proof regarding
their formation. The intensity of these absorptions in the IR is not
large, and so their detection in the observed spectrum is complicated
by the variety of products formed during photolysis.
(49) Raman experiments were not performed on the photolyzed
matrixes due to the lack of experimental facilities. Such an experiment
might have enabled detection of small amounts of 2-butyne.
(50) The imaginary frequency due to methyl group rotation in C2h
2-butyne is -16i cm-1
.

Dimethylvinyldiene
compound
J . Org. Chem., Vol. 66, No. 1, 2001 299
Ta ble 4. Ca lcu la ted En er gies of th e P r od u cts Resu ltin g fr om In tr a m olecu la r Rea r r a n gem en t of 5a
B3LYP/
6-31G(d)
B3LYP/cc-pVTZ//
B3LYP/6-31G(d)
B3LYP/cc-pVTZ//
B3LYP/6-31G(d)+ZPVE
HF/6-31G(d)
5S
-154.8411037
-155.8977825
-50.9
-155.9590759
-51.6
-155.8757479
-50.8
2-butyne (12)
TS1(6 f 12)
1-methylcyclopropene (13)
TS2 (6 f 13)
-42.8
22.5
-16.9
b
11.8
11.4
10.7
-29.4
14.3
-28.2
13.7
-23.5
12.1
a
b
Relative energies are shown (in parentheses) at the B3LYP/cc-pVTZ//B3LYP/6-31G(d) level of theory. Geometry did not converge.
Con clu sion s
Matrix-isolation irradiation of a diperoxyester precur-
sor (6) in an argon matrix was used to produce dimeth-
ylvinylidene 5, which was subsequently identified by IR
spectroscopy and DFT calculations. Spin-state analysis
of 5 at a variety of correlated levels indicate the singlet
state is more stable than the triplet by ∼45 kcal mol-1
.
Several additional photoproducts were identified in these
experiments by correlation with their computed IR
frequencies and consisted of an R-lactone (9), dimeth-
ylketene (10) and dimethylpropadienone (11). Neither of
the expected intramolecular rearrangement products,
2-butyne or 1-methyl-1-cyclopropene, were identified in
the matrix IR spectrum. DFT calculations were used to
identify the potential energy surface for intramolecular
rearrangement of the carbene and show a relatively high
barrier to intramolecular rearrangement.
F igu r e 9. Bond lengths and relevant bond angles for the
transition states connecting carbene 5 to rearranged products
12 and 13. The transition states are denoted as TS1 and TS2,
respectively. The values given are for geometries calculated
at the B3LYP/6-31G(d) level of theory. Bond lengths are given
in angstroms and are indicated in the figure by full arrows;
bond angles are indicated using half-arrows.
Ack n ow led gm en t. The authors would like to thank
The University of Akron for their support of this work
in the form of a Faculty Research Grant (FRG-1441).
We would like to thank Monica Cerro-Lopez for gener-
ating the tert-butylperoxy radical under matrix-isolation
conditions and Professor Matthew Platz (Ohio State
University) for providing a critical review of the manu-
script and for the use of his matrix isolation system.
We would also like to thank Dr. Allan East (The
University of Akron) for helpful comments on the
calculations for transition state TS1.
upon excitation, resulting in dimethylketene, which was
used by Brahms and Dailey to produce 12 and 13,
presumably through 5.38 However, direct irradiation of
dimethylketene should produce 5 in a vibrationally hot
state that is capable of further rearrangement. Irradia-
tion of 6, however, probably generates 5 via a series of
stepwise bond cleavages resulting in a vibrationally
relaxed carbene, which does not rearrange rapidly in the
matrix. We surmise dimethylketene is produced in such
small quantities that loss of CO produces only minor
amounts of the vibrationally hot carbene, and this
amount in turn produces 12 and 13 in such small
quantities so as to be undetectable.
Su p p or tin g In for m a tion Ava ila ble: Experimental IR
spectrum resulting from irradiation of tert-butylperoxide in an
Ar matrix at 14 K, the IR difference spectrum obtained upon
subtraction of the Ar and Ar/CO matrix spectra shown in
Figures 3 and 4, and energies and Cartesian coordinates for
5, 9-13, and TS1 and TS3. This material is available free of
charge via the Internet at http://pubs.acs.org.
J O001464I