The benzylidenecarbene-phenylacetylene rearrangement: An experimental and computational study

Article abstract of DOI:10.1021/ja310440e

Benzylidenecarbene was generated from a new photochemical source, 1-benzylidene-1a,9b-dihydro-1H-cyclopropa[l]phenanthrene, in deuterated benzene at ambient temperature. The carbene undergoes a facile rearrangement to phenylacetylene and could not be trapped by olefins. Generation of the carbene bearing a ¹³C label at the β-carbon produced phenylacetylene in which the label was found exclusively at the carbon adjacent to the phenyl ring. This overwhelming preference for H shift is consistent with B3LYP and CCSD(T) calculations. The label distribution observed in this work, however, contrasts previously reported high-temperature flash vacuum pyrolysis results where the interconversion of carbene and alkyne leads to the scrambling of labels over both alkynyl (sp) carbons.

Full text of DOI:10.1021/ja310440e

Communication
pubs.acs.org/JACS
The Benzylidenecarbene−Phenylacetylene Rearrangement: An
Experimental and Computational Study
Kathryn A. Moore, Jesus S. Vidaurri-Martinez, and Dasan M. Thamattoor*
Department of Chemistry, Colby College, Waterville, Maine 04901, United States
S
* Supporting Information
aromatic compound synthesis.3 Herein we describe the solution
ABSTRACT: Benzylidenecarbene was generated from a
new photochemical source, 1-benzylidene-1a,9b-dihydro-
1H-cyclopropa[l]phenanthrene, in deuterated benzene at
ambient temperature. The carbene undergoes a facile
rearrangement to phenylacetylene and could not be
trapped by olefins. Generation of the carbene bearing a
13C label at the β-carbon produced phenylacetylene in
which the label was found exclusively at the carbon
adjacent to the phenyl ring. This overwhelming preference
for H shift is consistent with B3LYP and CCSD(T)
calculations. The label distribution observed in this work,
however, contrasts previously reported high-temperature
flash vacuum pyrolysis results where the interconversion of
carbene and alkyne leads to the scrambling of labels over
both alkynyl (sp) carbons.
chemistry of benzylidenecarbene, generated from a new
photochemical precursor, and show that under these
conditions, the phenylacetylene results exclusively from a 1,2-
hydrogen shift in the carbene. A detailed computational study
of the benzylidenecarbene−phenylacetylene system, which
agrees with experimental observations, is also presented.
Nearly 50 years ago, it was demonstrated that the
cyclopropanated phenanthrene derivative 5 (R1R2H)
could be photolyzed to produce methylene with the
concomitant loss of phenanthrene (eq 3).4 Remarkably, the
rown et al. have previously noted that flash vacuum
potential of this system for generating other carbenes remained
untapped for another 25 years before 5 (R1R2Cl) was used
in a landmark study to determine the absolute kinetics of
dichlorocarbene in solution.5 Recently, however, there has been
renewed interest in compounds based on 5 as legitimate
carbene sources, and several laboratories, including ours, have
used such systems to investigate the chemistry of a number of
carbenes.6
B
pyrolysis (FVP) of 13C-labeled benzylidene malonate 1, a
precursor to labeled benzylidenecarbene (2), at 560 °C and 0.1
mm Hg, produced isotopomeric phenylacetylenes 3 and 4 in a
3:1 ratio (eq 1).1 They were careful to point out, however, that
As there are currently no convenient, readily prepared
photochemical sources of alkylidenecarbenes with potential for
general applicability, we wondered if it might be feasible to
generate such unsaturated carbenes by modifying 5.7 To test
this idea, we carried out the synthesis of 1-benzylidene-1a,9b-
dihydro-1H-cyclopropa[l]phenanthrene 8, a potential precursor
to benzylidenecarbene, following a simple three-step procedure
described in Scheme 1. First, phenanthrene was treated with
bromoform and aqueous sodium hydroxide under phase
transfer catalysis conditions, following our improved proce-
dure,8 to produce the dibromocarbene adduct 6. Reaction of 6
with butyllithium at low temperature, followed by quenching of
the resulting lithio anion with benzyl bromide, gave the
benzylated derivative 7, likely via an SN2 process. The relative
stereochemical disposition of benzyl and bromo groups in 7
was determined by X-ray cystallography (see Supporting
Information, SI) and is consistent with our previous
observations of analogous reactions.6a Finally, dehydrobromi-
nation of 7 to give the desired precursor 8 was successfully
the observed ratio “cannot be taken as an accurate reflection of
the relative migratory aptitudes of hydrogen and phenyl as at
this temperature, the migration is reversible”.1a
This notion was supported by the fact that when 3 was
independently pyrolyzed at 550 °C and 0.05 mm Hg, the
pyrolysate contained 3 and 4 in a 9:1 ratio (eq 2).2 However,
when the pyrolysis was conducted at 700 °C and 0.02 mm Hg,
approximately equal amounts of 9 and 10 were produced.2
These experiments led the authors to conclude that under high-
temperature FVP conditions, benzylidenecarbene and phenyl-
acetylene were in equilibrium, which led to the observed
scrambling of labels. Indeed, this alkylidene−alkyne inter-
conversion, now known as the Brown rearrangement, has
evolved into an important process, especially for polycyclic
Received: October 22, 2012
Published: November 26, 2012
© 2012 American Chemical Society
20037
dx.doi.org/10.1021/ja310440e | J. Am. Chem. Soc. 2012, 134, 20037−20040

Journal of the American Chemical Society
Communication
Scheme 1. Synthesis of 8
been a subject of intense computational scrutiny over the
years.11 Recent calculations, at the CBS/CCSD(T) level with
core correlation effects included, suggest a barrier of 2.08 kcal/
mol for conversion of vinylidene into acetylene and a ΔErxn of
−46.54 kcal/mol.11a
While the barrier for 1,2-hydrogen shift in 12 is small, it has
been suggested that it could be even smaller in 9, perhaps due
to the assistance provided by the phenyl ring.7a Our
computational investigation into the structure of 9 and its
rearrangement to 10 appears to support such a notion.12
B3LYP13 calculations (Table 1), including zero-point energy
Table 1. Computed Relative Energies in kcal/mol, with
Zero-Point Energy Corrections, of 9s and 9t
Benzylidenecarbene, Phenylacetylene 10, and Transition
States for the Conversion of 9s to 10 via H and Ph Shifts
a
B3LYP/6-
B3LYP/6-
B3LYP/6-
B3LYP/aug-cc-
pVDZ
entity
bc
31G*
31+G*
31+G**
,
9s
0
0
0
0
Figure 1. X-ray structure of 8.
cde
, ,
9t
+38.43
−46.15
−0.82
+8.79
+37.70
−45.01
−0.45
+8.92
+38.37
−44.64
−1.35
+9.30
+38.90
−43.71
−1.78
bc
,
10
accomplished at rt by treatment with potassium tert-butoxide in
DMSO. The single crystal X-ray structure of 8 is shown in
Figure 1.
Precursor 8 undergoes photolysis cleanly (∼315−400 nm) in
benzene-d6 at ambient temperature to afford the putative
benzylidenecarbene 9, as evident from the quantitative
formation of phenylacetylene 10 within 5 h (eq 4).
bf
,
TS1,2‑H
TS1,2‑Ph
bf
,
+10.24
a
b
c
Computational details are provided in the SI. RB3LYP. No
d
e
imaginary frequencies. UB3LYP. Spin expectation values <S2> =
2.000. One imaginary frequency.
f
corrections, indicate that the singlet carbene, 9s, is much more
stable than its triplet counterpart (9t), with ΔEST values
ranging from 37.70 [(U)B3LYP/6-31+G*] to 38.90 kcal/mol
[(U)B3LYP/aug-cc-pVDZ]. The ΔErxn for conversion of 9s to
10 varied from −43.71 (B3LYP/aug-cc-pVDZ) to −46.15 kcal/
mol (B3LYP/6-31G*). Furthermore, while the 1,2-phenyl shift
has to overcome a modest barrier, ranging from 8.79 (B3LYP/
6-31G*) to 10.24 kcal/mol (B3LYP/aug-cc-pVDZ), transition
state energies for the 1,2-hydrogen shift pathway are all just
below 9s, varying from −0.45 (B3LYP/6-31G*) to −1.78 kcal/
mol (B3LYP/aug-cc-pVDZ). The preference for 1,2-hydrogen
shift in 9s to give 10 is overwhelming, and other processes, such
as 1,2-phenyl shift, much less intermolecular reactions, are not
competitive under these conditions.
A similar situation is also encountered when energies are
computed at the CCSD(T)/cc-pVTZ//B3LYP/6-31+G**
level of theory (Figure 2). Stationary points on the potential
energy surface are all well behaved with little, if any,
multireference character (all T1 diagnostic values <0.02). The
transition state for the 1,2-hydrogen shift remains slightly below
that of the starting material by −1.43 kcal/mol, whereas the
1,2-phenyl shift needs to overcome a barrier of 8.45 kcal/mol.
The ΔErxn of −48.24 is comparable to the DFT calculations.
The observed trend may be rationalized by examining the
structure of 9s which already appears “predisposed” for the 1,2-
Furthermore, no trapping products were obtained when the
photolysis of 8 was carried out in cyclohexene, cis-3-hexene, or
2-methyl-2-butene, presumably because rearrangement of 9 to
10 occurs much more rapidly than intermolecular processes.
We believed that our precursor 8 could be useful in
accurately determining the H vs Ph migratory aptitudes in
benzylidenecarbene, given that the phenylacetylene derived
from 9 is formed under conditions where it cannot revert to the
carbene. Accordingly, we prepared 11, the labeled version of 8,
by using Ph13CH2Br in step 2 of Scheme 1. Photolysis of 11 in
benzene-d6 at rt afforded 3 but no 4, demonstrating the
exclusive preference for the migration of hydrogen to phenyl in
carbene 2 (eq 5).9
Experiment and theory both indicate that the simplest
alkylidene, vinylidene (12), is a ground-state singlet with a
ΔEST of ∼48 kcal/mol.10 Indeed, the potential energy surface
for the singlet vinylidene−acetylene rearrangement (eq 6) has
20038
dx.doi.org/10.1021/ja310440e | J. Am. Chem. Soc. 2012, 134, 20037−20040

Journal of the American Chemical Society
Communication
8 releases carbene 9, and the complete absence of precursor
rearrangements, is particularly noteworthy. This may be
attributed to the fact that extrusion of the carbene from 8 is
accompanied by aromatization of the central ring in
phenanthrene, which provides an important driving force for
the reaction. Such a driving force is unavailable to 16 and 17.
We are currently investigating the scope and limitations of
analogs of 8 to determine whether such precursors could serve
as a general source of alkylidenecarbenes.
In conclusion, benzylidenecarbene, generated from a new
photochemical precursor, undergoes a facile rearrangement to
phenylacetylene. The carbene could not be trapped by
cyclohexene, cis-3-hexene, or 2-methyl-2-butene. Using a 13C-
labeled precursor, it was demonstrated that phenylacetylene
results exclusively from a 1,2-hydrogen shift in the carbene, and
the 1,2-phenyl shift is not competitive. Calculations show that
the carbene is a ground-state singlet that is more stable than the
triplet by ∼38 kcal/mol. Furthermore, the barrier for hydrogen
migration in the singlet carbene to form the alkyne is essentially
nonexistent. The activation energy for phenyl shift, however, is
much larger by comparison (∼ 8−10 kcal/mol).
Figure 2. PES for the rearrangement of singlet benzylidenecarbene
(9s) to phenylacetylene (10) at CCSD(T)/cc-pVTZ//B3LYP/6-
31+G** including energies, T1 diagnostic values (CCSD), and
imaginary frequencies (for transition states).
hydrogen shift (Figure 2). The angles around C2 are
substantially distorted from what might be expected for a
formally sp2 carbon. In particular, the H1−C2−C1 angle is
severely compressed to 94°, whereas the C1−C2−C3 angle is
quite large at 146°. In TS1,2‑H, the C1−C2−C3 has essentially
become linear at 173°, and the H1−C2−C1 angle has
decreased further to 62° as the migrating hydrogen leaves C2
and approaches the C1 terminus. Thus, very little motion
seems required to tip the hydrogen over from C2 to C1 as 9s
proceeds toward 10. On the other hand, shifting the phenyl is
considerably more difficult. In 9s, the phenyl ring is coplanar
with the C1−C2 internuclear axis. In TS1,2‑Ph, however, the
plane of the phenyl ring is essentially orthogonal to the C1−C2
bond (Figure 2). To achieve such a geometry, the phenyl ring
will need to twist out of plane in 9s, presumably disrupting
stabilizing conjugation with the C1−C2 π system.
These observations provide an interesting contrast to an
earlier report by Stang et al. who showed that 14C-labeled (E)-
and (Z)-2-phenylpropenyltriflates (13), when individually
treated with potassium tert-butoxide in pentane/glyme for 24
h at −20 °C, produced a common carbene intermediate 14 and
resulted in exclusive phenyl migration to afford [2-14C]-1-
phenylpropyne (15) (Scheme 2).14 Thus, our present work,
together with Stang’s observations, provides confirmatory
evidence for the H≫Ph≫CH3 trend in migratory aptitudes
in the rearrangement of vinylidenes.
ASSOCIATED CONTENT
* Supporting Information
Procedures, characterization data, computational details, and
full ref 12. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
dmthamat@colby.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge support by the National Science
Foundation (CHE-1012914), and a summer research award
from the Clare Booth Luce Foundation to K.A.M.
REFERENCES
■
(1) (a) Brown, R. F. C.; Eastwood, F. W.; Harrington, K. J.;
McMullen, G. L. Aust. J. Chem. 1974, 27, 2393. (b) Brown, R. F. C.;
Harrington, K. J. J. Chem. Soc., Chem. Commun. 1972, 1175.
(2) Brown, R. F. C.; Harrington, K. J.; McMullen, G. L. J. Chem. Soc.,
Chem. Commun. 1974, 123.
Scheme 2. Stang’s Demonstration of the Exclusive Migration
of Phenyl, Relative to Methyl, in Carbene 14
(3) (a) Tsefrikas, V. M.; Scott, L. T. Chem. Rev. (Washington, DC,
U.S.) 2006, 106, 4868. (b) Brown, R. F. C. Eur. J. Org. Chem. 1999,
3211.
(4) Richardson, D. B.; Durrett, L. R.; Martin, J. M., Jr.; Putnam, W.
E.; Slaymaker, S. C.; Dvoretzky, I. J. Am. Chem. Soc. 1965, 87, 2763.
(5) Chateauneuf, J. E.; Johnson, R. P.; Kirchhoff, M. M. J. Am. Chem.
Soc. 1990, 112, 3217.
(6) (a) Graves, K. S.; Thamattoor, D. M.; Rablen, P. R. J. Org. Chem.
2011, 76, 1584. (b) Farlow, R. A.; Thamattoor, D. M.; Sunoj, R. B.;
Hadad, C. M. J. Org. Chem. 2002, 67, 3257. (c) Thamattoor, D. M.;
Snoonian, J. R.; Sulzbach, H. M.; Hadad, C. M. J. Org. Chem. 1999, 64,
5886. (d) Robert, M.; Likhotvorik, I.; Platz, M. S.; Abbot, S. C.;
Kirchoff, M. M.; Johnson, R. J. Phys. Chem. A 1998, 102, 1507.
(e) Nigam, M.; Platz, M. S.; Showalter, B. M.; Toscano, J. P.; Johnson,
It has been reported before that photofragmentation of
methylenecyclopropanes (16, R = H, D, CH3) affords the
corresponding vinylidenes but in yields <5%.15 Furthermore,
photolysis of 16 is complicated by the methylenecyclopropane
rearrangement that converts it into 17, a source of acetylene via
vinylidene (12). In this context, the ease with which precursor
20039
dx.doi.org/10.1021/ja310440e | J. Am. Chem. Soc. 2012, 134, 20037−20040

Journal of the American Chemical Society
Communication
R.; Abbot, S. C.; Kirchhoff, M. M. J. Am. Chem. Soc. 1998, 120, 8055.
(f) Ruck, R. T.; Jones, M., Jr. Tetrahedron Lett. 1998, 39, 2277.
(g) Ruck, R. T.; Jones, M., Jr. Tetrahedron Lett. 1998, 39, 4433.
(h) Glick, H. C.; Likhovorik, I. R.; Jones, M., Jr. Tetrahedron Lett.
1995, 36, 5715.
(7) For reviews, see: (a) Knorr, R. Chem. Rev. (Washington, DC, U.S.)
2004, 104, 3795. (b) Stang, P. J. Pure Appl. Chem. 1983, 55, 369.
(c) Stang, P. J. Chem. Rev. 1978, 78, 383.
(8) Nguyen, J. M.; Thamattoor, D. M. Synthesis 2007, 2093.
(9) It has been reported that treatment of Ph13CHCBr2 with
butyllithium produces phenylacetylene with the label adjacent to the
phenyl ring, but it is likely the process involves a lithium carbenoid.
See: Xiang, J. S.; Fuchs, P. L. Tetrahedron Lett. 1996, 37, 5269.
(10) Sander, W.; Kotting, C. Chem.Eur. J. 1999, 5, 24.
(11) Recent articles with references to prior work, see: (a) Joseph, S.;
Varandas, A. J. C. J. Phys. Chem. A 2010, 114, 13277. (b) Greenman,
L.; Mazziotti, D. A. J. Phys. Chem. A 2010, 114, 583.
(12) Frisch, M. J. et al. Gaussian 09, rev. A.02; Gaussian, Inc.:
Wallingford, CT, 2009.
(13) (a) Becke, A. D. Phys. Rev. A Gen. Phys. 1988, 38, 3098. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (c) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648.
(14) Stang, P. J.; Fox, D. P.; Collins, C. J.; Watson, C. R. J. Org. Chem.
1978, 43, 364.
(15) Gilbert, J. C.; Butler, J. R. J. Am. Chem. Soc. 1970, 92, 7493.
20040
dx.doi.org/10.1021/ja310440e | J. Am. Chem. Soc. 2012, 134, 20037−20040