Multiphoton infrared initiated thermal reactions of esters: Pseudopericyclic eight-centered cis-elimination

Article abstract of DOI:10.1021/ja804812c

Multiphoton infrared absorption from a focused, pulsed CO₂ laser was used to initiate gas-phase thermal reactions of cis- and trans-3-penten-2-yl acetate. By varying the helium buffer gas pressure, it was possible to deduce the product distribution from the initial unimolecular reactions, separate from secondary reactions in a thermal cascade. Thus, trans-3-penten-2-yl acetate gives 54 ± 5% of β-elimination to give trans-1,3-pentadiene, 40 ± 3% of [3,3]-sigmatropic rearrangement to give cis-3-penten-2-yl acetate and 6 ± 4% of cis-1,3-pentadiene. Similar irradiation of cis-3-penten-2-yl acetate gives 45 ± 1% of β-elimination to give cis-1,3-pentadiene, 32 ± 2% of [3,3]-sigmatropic rearrangement to give trans-3-penten-2-yl acetate and 23 ± 2% of trans-1,3-pentadiene. The latter process is an eight-centered δ-elimination, which is argued to be a pseudopericyclic reaction. Although β-eliminations have been suggested to be pericyclic, B3LYP/ 6-31G(d,p), MP2 and MP4 calculations suggest that both β- and δ-eliminations, as well as [3,3]-sigmatropic rearrangements of esters are primarily pseudopericyclic in character, as judged by both geometrical, energetic and transition state aromaticity (NICS) criteria. Small distortions from the ideal pseudopericyclic geometries are argued to reflect small pericyclic contributions. It is further argued that when both pericyclic and pseudopericyclic orbital topologies are allowed and geometrically feasible, the calculated transition state may be the result of proportional mixing of the two states; this offers an explanation of the range of pseudopericyclic and pericyclic characters found in related reactions.

Full text of DOI:10.1021/ja804812c

Published on Web 12/19/2008
Multiphoton Infrared Initiated Thermal Reactions of Esters:
Pseudopericyclic Eight-Centered cis-Elimination
Hua Ji, Li Li,† Xiaolian Xu,‡ Sihyun Ham,§ Loubna A. Hammad,¶ and
David M. Birney*
Department of Chemistry and Biochemistry, Texas Tech UniVersity, Lubbock, Texas 79409-1061
Received June 24, 2008; E-mail: david.birney@ttu.edu
Abstract: Multiphoton infrared absorption from a focused, pulsed CO2 laser was used to initiate gas-phase
thermal reactions of cis- and trans-3-penten-2-yl acetate. By varying the helium buffer gas pressure, it was
possible to deduce the product distribution from the initial unimolecular reactions, separate from secondary
reactions in a thermal cascade. Thus, trans-3-penten-2-yl acetate gives 54 ( 5% of ꢀ-elimination to give
trans-1,3-pentadiene, 40 ( 3% of [3,3]-sigmatropic rearrangement to give cis-3-penten-2-yl acetate and 6
( 4% of cis-1,3-pentadiene. Similar irradiation of cis-3-penten-2-yl acetate gives 45 ( 1% of ꢀ-elimination
to give cis-1,3-pentadiene, 32 ( 2% of [3,3]-sigmatropic rearrangement to give trans-3-penten-2-yl acetate
and 23 ( 2% of trans-1,3-pentadiene. The latter process is an eight-centered δ-elimination, which is argued
to be a pseudopericyclic reaction. Although ꢀ-eliminations have been suggested to be pericyclic, B3LYP/
6-31G(d,p), MP2 and MP4 calculations suggest that both ꢀ- and δ-eliminations, as well as [3,3]-sigmatropic
rearrangements of esters are primarily pseudopericyclic in character, as judged by both geometrical,
energetic and transition state aromaticity (NICS) criteria. Small distortions from the ideal pseudopericyclic
geometries are argued to reflect small pericyclic contributions. It is further argued that when both pericyclic
and pseudopericyclic orbital topologies are allowed and geometrically feasible, the calculated transition
state may be the result of proportional mixing of the two states; this offers an explanation of the range of
pseudopericyclic and pericyclic characters found in related reactions.
of 1 that favors the [3,5] product 2 (eq 1).1e We now report
Introduction
that the same is true of ester eliminations, namely that eight-
A pseudopericyclic reaction is a concerted reaction in which
bond changes occur around a ring, but which lacks the cyclic
orbital overlap that is characteristic of a pericyclic reaction.1,2
“A pseudopericyclic reaction may be orbital symmetry allowed...
regardless of the number of electrons involved.”1a In support
of this bold statement, we have reported ab initio1d and density
functional theory (DFT) calculations1e which indicate that [1,3],
[3,3], and [3,5] sigmatropic rearrangements of esters are all
pseudopericyclic and allowed. These are consistent with the
observed competition between [3,3] and [3,5] rearrangements
(2) (a) Huisgen, R. Angew. Chem., Int. Ed. Engl. 1980, 19, 947–973. (b)
Fabian, W. M. F.; Kappe, C. O.; Bakulev, V. A. J. Org. Chem. 2000,
65, 47–53. (c) de Lera, A. R.; Alvarez, R.; Lecea, B.; Torrado, A.;
Coss´ıo, F. P. Angew. Chem., Int. Ed. Engl. 2001, 40, 557–561. (d)
Rodr´ıguez-Otero, J.; Cabaleiro-Lago, E. M. Angew. Chem., Int. Ed.
2002, 41, 1147–1150. (e) de Lera, A. R.; Coss´ıo, F. P. Angew. Chem.,
Int. Ed. 2002, 41, 1150–1152. (f) Rodr´ıguez-Otero, J.; Cabaleiro-Lago,
E. M.; Hermida-Ramo´n, J. M.; Pen˜a-Gallego, A. J. Org. Chem. 2003,
68, 8823–8830. (g) Cabaleiro-Lago, E. M.; Rodr´ıguez-Otero, J.;
Gonza´lez-Lo´pez, I.; Pen˜a-Gallego, A.; Hermida-Ramo´n, J. M. J. Phys.
Chem. A 2005, 109, 5636–5644. (h) Zora, M. J. Org. Chem. 2004,
69, 1940–1947. (i) Cabaleiro-Lago, E. M.; Rodr´ıguez-Otero, J.; Varela-
Valera, S. M.; Pen˜a-Gallego, A.; Hermida-Ramo´n, J. M. J. Org. Chem.
2005, 70, 3921–3928. (j) Cabaleiro-Lago, E. M.; Rodr´ıguez-Otero,
J.; Garcia-Lo´pez, R. M.; Pen˜a-Gallego, A.; Hermida-Ramo´n, J. M.
Chem.-Eur. J. 2005, 11 (20), 5966–5974. (k) Fukushima, K.;
Iwahashi, H. Heterocycles 2005, 65, 2605–2618. (l) Lo´pez, C. S.; Faza,
O. N.; Coss´ıo, F. P.; York, D. M.; de Lera, A. R. Chem.-Eur. J.
2005, 11, 1734–1738. (m) Alajarin, M.; Ortin, M.-M.; Sanchez-
Andrada, P.; Vidal, A. J. Org. Chem. 2006, 71, 8126–8139. (n)
Sheikhshoaie, I.; Belaj, F.; Fabian, W. M. F. J. Mol. Struct. 2006,
794, 244–250. (o) Fukushima, K.; Iwahashi, H. Chem. Lett. 2006, 35,
1242–1243. (ii) Kruger, H. G.; Mdluli, P.; Power, T. D.; Raasch, T.;
Singh, A. J. Mol. Struct. THEOCHEM 2006, 771, 165–170. (p) Rode,
J. E.; Dobrowoski, J. C. J. Phys. Chem. A 2006, 110, 3723–3737. (q)
George, L.; Netsch, K. P.; Penn, G.; Kollenz, G.; Wentrup, C. Org.
Biomol. Chem. 2006, 4, 558–564. (r) Pen˜a-Gallego, A.; Rodr´ıguez-
Otero, J.; Cabaleiro-Lago, E. M. Tetrahedron 2007, 63, 4937–4943.
(s) Chamorro, E.; Notario, R.; Santos, J. C.; Perez, P. Chem. Phys.
Lett. 2007, 443, 136–140. (t) Pen˜a-Gallego, A.; Rodr´ıguez-Otero, J.;
Cabaleiro-Lago, E. M. J. Phys. Chem. A 2007, 111, 2935–2940. (u)
Rode, J. E.; Dobrowolski, J. C. Chem. Phys. Lett. 2007, 449, 240–
245. (v) Calvo-Losada, S.; Sanchez, J. J. Q. J. Phys. Chem. A 2008,
112, 8164–8178.
† Present address: Department of Chemistry, Purdue University, 560 Oval
Drive, West Lafayette, IN 47907-2084.
‡ Present address: Lexicon Pharmaceuticals, 350 Carter Road, Princeton,
New Jersey 08540.
§ Present address: Department of Chemistry, Sookmyung Women’s
University, Seoul 140-742, Republic of Korea.
¶ Present address: Department of Chemistry, Indiana University, 800 E.
Kirkwood Ave., Bloomington, IN 47405-7102.
(1) (a) Ross, J. A.; Seiders, R. P.; Lemal, D. M. J. Am. Chem. Soc. 1976,
98, 4325–4327. (b) Birney, D. M.; Wagenseller, P. E. J. Am. Chem.
Soc. 1994, 116, 6262–6270. (c) Birney, D. M.; Ham, S.; Unruh, G. R.
J. Am. Chem. Soc. 1997, 119, 4509–4517. (d) Birney, D. M.; Xu, X.;
Ham, S. Angew. Chem., Int. Ed. 1999, 38, 189–193. (e) Quideau, S.;
Looney, M. A.; Pouyse´gu, L.; Ham, S.; Birney, D. Tetrahedron Lett.
1999, 40, 615–618. (f) Shumway, W. W.; Dalley, N. K.; Birney, D. M.
J. Org. Chem. 2001, 66, 5832–5839. (g) Wei, H.-X.; Zhou, C.; Ham,
S.; White, J. M.; Birney, D. M. Org. Lett. 2004, 6, 4289–4292. (h)
Gudipati, I. R.; Sadasivam, D. V.; Birney, D. M. Green Chem. 2008,
10, 283–285. (i) Sadasivam, D. V.; Birney, D. M. Org. Lett. 2008,
10, 245–248.
9
528 J. AM. CHEM. SOC. 2009, 131, 528–537
10.1021/ja804812c CCC: $40.75
2009 American Chemical Society

Pseudopericyclic Eight-Centered cis-Elimination
A R T I C L E S
the Woodward-Hoffmann rules it was explicitly described as
an allowed σ2CH + π2CO + σ2CO reaction.5a In light of the
synthetic and theoretical importance of this reaction, it is not
surprising there have been both semiempirical6 and quantum
chemical7 calculations of ꢀ-elimination from ethyl formate (3)
and of more complex esters. The calculated transition state has
been described as “‘aromatic’ in that it involves six electrons
in a six-membered cyclic structure.”6a Although this mechanistic
description is familiar and seems reasonable, it is not consistent
with the fact that at all levels of theory, the transition state is
planar, or close to planar.6,7 This geometry is in qualitative
contrast to the calculated transition state for the hydrocarbon
retro-ene reaction of 1-pentene (5, eq 3), which has a nonplanar
transition state to allow for π-orbital overlap in a pericyclic
transition state, as illustrated in Figure 1A.8
Figure 1. (A) Qualitative transition state orbital overlap in the pericyclic
hydrocarbon retro-ene reaction of 1-pentene (5) and (B) in the planar,
pseudopericyclic ꢀ-elimination of ethyl formate (3). (C, D) HOMO (in-
plane) and HOMO-1 (out-of-plane) of the B3LYP/6-31G(d,p) transition
state for ꢀ-elimination of 3, (TS-3, after reference.7b (E) Ester resonance.9
Note that the ester ꢀ-elimination could alternatively involve orbital overlap
with the ester π-system as in Figure 1A; this would disrupt the ester
π-resonance that is preserved in the pseudopericyclic orbital topology in
1B.
The details of possible transition states will be discussed in
more detail below in the context of new computational results.
Briefly, a planar transition state geometry for an ester ꢀ-elimina-
tion is possible because the ester π-system is disconnected from
the in-plane σ-system as shown in Figure 1B. This planar
transition state transfers a proton to an ester carbonyl lone pair
and the breaking bond becomes a lone pair on the acid. The
HOMO and HOMO-1 for the planar transition state for
ꢀ-elimination from 3 (TS-3, B3LYP/6-31G(d,p)7b) are shown
in Figure 1C and D. The in-plane and out-of plane systems
remain orthogonal and therefore cannot have cyclic orbital
overlap involving the π-system;10 thus, the transition state can
neither be aromatic nor antiaromatic.11 Nor can it be discussed
in terms of the Woodward-Hoffmann rules12 because it has
neither suprafacial nor antarafacial orbital overlap. Rather, this
orbital topology corresponds to a pseudopericyclic transition
state, as first described by Lemal1a and as subsequently discussed
by us1b-i and others.2 The lack of cyclic orbital overlap in
pseudopericyclic reactions may have three experimental con-
sequences: (1) such reactions are allowed regardless of the
centered δ-eliminations (e.g., of 4, eq 2b) are competitive with
the more familiar ꢀ-eliminations (e.g., of 3, eq 2a).3
It is well-known that pyrolysis of esters with a ꢀ-hydrogen
(e.g., ethyl formate, 3) leads to stereospecific cis-elimination
via a six-centered transition state (ꢀ-elimination, eq 2a).4 This
synthetically useful, textbook reaction has often been described
as a six-electron concerted process that is fundamentally the
same as a retro-ene reaction.5 Shortly after the publication of
(6) (a) Lee, I.; Cha, O. J.; Lee, B.-S. J. Phys. Chem. 1990, 94, 3926–
3930. (b) Hamon, L.; Levisalles, J.; Pascal, Y.-L. Tetrahedron 1994,
50, 161–170.
(7) (a) Bohm, S.; Skancke, P. N. Int. J. Quantum Chem. 1991, 40, 491–
500. (b) Erickson, J. A.; Kahn, S. D. J. Am. Chem. Soc. 1994, 116,
6271–6276. (c) Speybroeck, V. V.; Martele, Y.; Schacht, E.; Waro-
quier, M. J. Phys. Chem. A 2002, 106, 12370–12375. (d) Notario, R.;
Quijano, J.; Quijano, J. C.; Gutie´rrez, L. P.; Sua´rez, W. A.; Sa´nchez,
C.; Leo´n, L. A.; Chamorro, E. J. Phys. Chem. A 2002, 106, 4377–
4383. (e) Hermida-Ramo´n, J. M.; Rodr´ıguez-Otero, J.; Cabaleiro-Lago,
E. M. J. Phys. Chem. A 2003, 107, 1651–1654. (f) Silva, A. M. Chem.
Phys. Lett. 2007, 439, 8–13.
(3) A preliminary communication of computational results on model
systems has been presented: Birney, D. M.; Xu, X.; Ham, S.
Pseudopericyclic Reactions of Esters: Eight-Centered Cis Elimination
is Allowed; Abstracts of National Meeting of The American Chemical
Society, Boston, MA, 1998.
(4) (a) DePuy, C. H.; King, R. W. Chem. ReV. 1960, 60, 431–457. (b)
Benson, R. E.; McKusick, B. C.; Grummitt, O.; Budewitz, E. P.;
Chudd, C. C. Org. Synth. Coll. 1963, 4, 746. (c) Smith, G. G.; Kelly,
F. W. Prog. Phys. Org. Chem. 1971, 8, 75–234.
(8) (a) Loncharich, R. J.; Houk, K. N. J. Am. Chem. Soc. 1987, 109, 6947–
6952. (b) Deng, Q.; Thomas, B. E. IV.; Houk, K. N.; Dowd, P. J. Am.
Chem. Soc. 1997, 119, 6902–6908. (c) Sakai, S. J. Phys. Chem. A
2006, 110, 12891–12899.
(9) Wiberg has advanced an alternative explanation for the planarity of
esters; see: Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1987,
109, 5935–5943, and references therein.
(5) (a) Caserio, M. C. J. Chem. Educ. 1971, 48, 782–790. (b) Alder, R. W.;
Baker, R.; Brown, J. M. Stabilized cyclically conjugated transition
states. Mechanism in Organic Chemistry, 1st ed.; John Wiley & Sons:
London, 1971; p 239, “stabilized cyclically conjugated transition-
states”. (c) Pearson, R. G., A six-electron TS. Symmetry Rules for
Chemical Reactions, 1st ed.; John Wiley and Sons: New York, 1976;
p 368, “a six-electron TS”. (d) Gilchrist, T. L.; Storr, R. C. Organic
reactions and orbital symmetry, 1st ed.; Cambridge University Press:
Cambridge, 1979; Section 7.8, p 292, “other six-electron pericyclic
processes”. (e) Fleming, I. It must be all-suprafacial and it is therefore
fundamentally a retro-ene reaction. Pericyclic Reactions, 1st ed.;
Oxford University Press: Oxford, 1999; p 87, “it must be all-suprafacial
and it is therefore fundamentally a retro-ene reaction”.
(10) Note that there is some cyclic overlap between non-adjacent atoms in
the in-plane system (Figure 1C).
(11) (a) Dewar, M. J. S. Angew. Chem. Int., Ed. Engl. 1971, 10, 761. (b)
Zimmerman, H. E. Acc. Chem. Res. 1971, 4, 272. (c) Schleyer, P. v.
R.; Maerker, C.; Dransfeld, A.; Jiao, H.; Hommes, N. J. R. v. E. J. Am.
Chem. Soc. 1995, 118, 6317–6318. (d) Morao, I.; Lecea, B.; Cossio,
F. P. J. Org. Chem. 1997, 62, 7033–7036. (e) Herges, R.; Geuenich,
D. J. Chem. Phys. A 2001, 105, 3214–3220.
(12) Woodward, R. B.; Hoffmann, R., The ConserVation of Orbital
Symmetry. ed.; Verlag Chemie, GmbH: Weinheim, 1970.
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 529

A R T I C L E S
Ji et al.
Scheme 1
Scheme 2
number of atoms involved,1e,f,i (2) the transition states can be
planar, in contrast to the nonplanar ones for pericyclic reactions,1f,g
(3) the barrier heights can be lower than a competing pericyclic
alternative.1b,c Pseudopericyclic reactions are expected not to
show the calculated transition state aromaticity associated with
pericyclic reactions.2d,m,11c-e
The pyrolyses of several allylic acetates have been studied
by Lewis et al. and the kinetics for racemization have been
measured.16 Of particular relevance to this work is the rear-
rangement of optically active trans-16 shown in equation 5.
Rearrangement leads to racemization with an Ea ) 38.8 ( 0.5
kcal/mol. This is faster than, but competitive with elimination
to form trans-17. Because of the double bond geometry in trans-
16, an eight-centered elimination is geometrically disfavored
and the formation of trans-17 can be assumed to involve only
ꢀ-elimination. Isomerization to cis-16 is also observed, but is a
minor pathway.
If the 6-centered transition state for ꢀ-elimination is pseudo-
pericyclic, then our qualitative prediction is that the vinylogous
δ-elimination is also allowed via an 8-centered transition state.
Prior to the Woodward/Hoffmann rules, there were a number
of experimental studies that gave serious consideration to the
possibility of δ-eliminations of esters, including some of the
earliest experimental studies of ester pyrolyses. In 1956, Bailey
et al. studied the products obtained by flash vacuum pyrolysis
of cis-2-buten-1,4-diol acetate (6, eq 4).13 They obtained a 26%
yield of the formal δ-elimination product (7) and a 20% yield
of rearranged diacetate (8), along with 24% of recovered 6. They
commented that 7 could be formed by rearrangement to 8,
followed by 6-centered elimination but that their results could
not rule out a small amount of a direct eight-centered elimina-
tion. Interestingly, there was none of the isomeric diene 9 that
could, in principle also be formed by elimination from 8.
Results and Discussion
The pyrolysis of cis-3-penten-2-yl acetate (cis-16) was studied
in this work because it provides a stringent test of the qualitative
prediction that δ-elimination via an eight-centered transition state
is allowed. This substrate (cis-16) has the potential to undergo
three reactions, both ꢀ- and δ-eliminations as well as a [3,3]-
sigmatropic rearrangement, with little or no steric or energetic
bias toward any one pathway (Scheme 2).17 A [3,3]-sigmatropic
rearrangement would be expected to give the more stable trans-
3-penten-2-yl acetate (trans-16) in preference to the degenerate
rearrangement to give cis-16; this expectation is confirmed by
the computational studies described below. ꢀ-Elimination of
acetate from cis-16 could only give cis-1,3-pentadiene (cis-17).
If the δ-elimination is allowed and competitive, the geometry
of cis-16 would permit it; thermodynamic control would also
be expected to give the more stable trans-1,3-pentadiene (trans-
17) in preference to cis-17; this is similarly in accord with
computations, below. However, observation of trans-17 from
a conventional pyrolysis does not require the concerted δ-e-
limination; trans-17 can be formed via sequential rearrangement
to trans-16 followed by ꢀ-elimination. Nevertheless, we antici-
pated that we could distinguish these two pathways for formation
A number of ester eliminations that lead to ortho-xylylene
and related species have also been studied, as shown in Scheme
1. These include eliminations to form ortho-xylylene (11),14 as
well as heteroaromatic analogs15 including the furyl derivative
13.14 The intermediates were not isolated, but were trapped with
a variety of dienophiles. The pyrrole derivative 14 also gave
trapping products that would be consistent with the radiallene
derivative 15 as an intermediate, although stepwise elimination,
trapping, elimination and trapping is also possible.15c Tranha-
novsky et al.14 recognized that there were two possible mech-
anisms for such a reaction, (1) direct δ-elimination via an eight-
centered transition state, or (2) a two-step process, consisting
of a [3,3] rearrangement followed by a more conventional
ꢀ-elimination. The rearrangement in the two-step process
disrupts the aromatic or heteroaromatic ring, favoring, but not
requiring the direct δ-elimination.
(13) Bailey, W. J.; Barclay, R., Jr. J. Org. Chem. 1956, 21, 328–331.
(14) Trahanovsky, W. S.; Cassady, T. J.; Woods, T. L. J. Am. Chem. Soc.
1981, 103, 6691–6695.
(16) Lewis, E. S.; Hill, J. T.; Newman, E. R. J. Am. Chem. Soc. 1968, 90,
662–668.
(15) (a) Potter, A. J.; Storr, R. C. Tetrahedron Lett. 1994, 35, 5293–5296.
(b) Chauhan, P. M. S.; Crew, A. P. A.; Jenkins, G.; Storr, R. C.;
Walker, S. M.; Yelland, Y. Tetrahedron Lett. 1990, 31, 1487–1490.
(c) Vessels, J. T.; Janicki, S. Z.; Petillo, P. A. Org. Lett. 2000, 2,
73–76.
(17) Additional transition states were considered in the computational
studies described below and in the Supporting Information. Their
energies are sufficiently high that these processes do not need to be
considered in this context.
9
530 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Pseudopericyclic Eight-Centered cis-Elimination
A R T I C L E S
Scheme 3. MP-IR Initiated Reaction of cis-16a
a Vibrationally excited molecules are shown in red and designated by
an asterisk.
of trans-17 by multiphoton infrared (MP-IR) photolysis/ther-
molysis, as described below.
Multiphoton Infrared Photolysis/Thermolysis (MP-IR). We18
and others19 have previously used multiphoton infrared pho-
tolysis/thermolysis to identify first-formed intermediates in other
thermal reaction cascades and it appeared to be a promising
technique to answer this question as well.18b This technique has
been described in detail elsewhere;20 briefly, a pulsed infrared
laser is tuned to a frequency where a compound of interest has
an IR absorption. In the gas phase, absorption of one IR photon
is followed by intramolecular vibrational relaxation (IVR); this
process is generally enhanced by molecular collisions. This
allows for the absorption of another photon; within a typical
laser pulse of approximately 100 ns, a molecule can thus absorb
multiple photons and be heated sufficiently to undergo relatively
high barrier reactions. This heating is generally assumed to give
a molecule with a random (Boltzmann) energy distribution.
However, any molecule that does not absorb the IR photons
cannot be heated. Thus a nonabsorbing buffer gas can rapidly
cool initial products in competition with a secondary thermal
reaction. MP-IR reactions can be carried out at pressures up to
approximately 100 torr, which balances the increased IR
absorbance due to collision-induced IVR with the increased
cooling at higher pressures.
Figure 2. Gas phase (3 torr) IR spectra of cis-16 and trans-16 in the region
of 1100 to 900 cm-1 (2 cm-1 resolution) and the laser line at 1057.3 cm-1
used for the MP-IR irradiation. The baselines are offset to show the spectra
more clearly.
process and the ꢀ-elimination to form trans-17 is unimolecular,
these two processes can be distinguished. In the simplest
analysis, if trans-16 does not absorb at the wavelength of the
IR photons, then it remains cool and unreactive. As will be
shown below, trans-16 does absorb at the IR wavelength and
does react, but extrapolation to the initial reaction of cis-16
removes this complication.
Experimental Methods
cis-3-Penten-2-yl acetate (cis-16)16 was synthesized as outlined
in equation 6;22 full experimental details and spectroscopic
characterization are provided in the Supporting Information. Fol-
lowing Henry,22a propynylmagnesium bromide was added to
acetaldehyde to give the propargyl alcohol 1822a and subsequent
reduction with Lindlar’s catalyst gave cis-3-penten-2-ol (19).22a
Esterification22b with acetyl chloride gave cis-16.22c trans-3-Penten-
2-yl acetate22b,c (trans-16) was synthesized by acetylation of
commercially available trans-3-penten-2-ol.
Yuan T. Lee has used MP-IR to initiate the decomposition
of ethyl acetate in a molecular beam/TOF mass spectrometer
experiment.19c The major reaction channel is the ꢀ-elimination
to give ethene and acetic acid. Under the low-pressure conditions
of this experiment, the acetic acid undergoes some secondary
thermolysis to give ketene and water. A minor pathway is
homolytic fragmentation to give ethyl radical, methyl radical,
and CO2.
Applying the MP-IR technique to the thermal reactions of
cis-16 suggests a reaction scheme as shown in Scheme 3. MP-
IR heating of cis-16 leads to a molecule with sufficient internal
energy (designated by *, cis-16*) to react via the transition
states,21 3,3-c-t-TS, ꢀ-c-c-TS or possibly δ-c-t-TS, in competi-
tion with collisional cooling from the helium buffer gas.
Conservation of energy dictates that rearrangement of cis-16*
produces trans-16*, also with sufficient internal energy to likely
undergo ꢀ-elimination via ꢀ-t-t-TS. However, collisional cooling
of trans-16* will occur in competition with the elimination.
Since the collisional cooling of trans-16* is a bimolecular
Our apparatus for MP-IR experiments has been previously
described.18 Briefly, a pulsed TEA CO2 laser is focused by a ZnSe
lens into a 10 cm gas cell fitted with NaCl windows and filled with
0.7 torr of cis-16 and a known pressure of He buffer gas (0 to 20
torr). Comparison of the IR spectra of cis-16 and trans-16 with the
emission lines available from the TEA CO2 laser suggested that
1057.3 cm-1 would give stronger absorbance by cis-16 relative to
trans-16 (Figure 2), so this wavelength was selected for this study.
Based on B3LYP/6-31G(d,p) frequency calculations (vide infra),
this corresponds to a C-O single bond stretch. This frequency is
on the side of an absorption in trans-16 that also is calculated to
correspond to a C-O stretch. The laser frequency was measured
using a CO2 laser spectrum analyzer. The progress of the reaction
was followed by IR spectroscopy. After 100 laser pulses, nitrogen
was added to bring the cell to atmospheric pressure. The products
(18) (a) Unruh, G. R.; Birney, D. M. J. Am. Chem. Soc. 2003, 125, 8529–
8533. (b) Li, L. Elimination of cis-3-Penten-2-yl Acetate. M.S. Thesis,
Texas Tech University, Lubbock, TX, 2003.
(21) Transition states related to 16 are designated as follows: first, the type
of reaction is indicated, [3,3]-sigmatropic rearrangement, b-elimination
or d-elimination, then cis or trans as related to the double bond
geometries in reactants and products. For other transition states, the
compound number of the starting material is also included.
(22) (a) Hamed, O.; Henry, P. M. Organometallics 1997, 16, 4903. (b)
Bateson, J. H.; Quinn, A. M.; Smale, T. C.; Southgate, R. J. Chem.
Soc. Perkin Trans 1. 1985, 2219. (c) McKew, J. C.; Kurth, M. J. J.
Org. Chem. 1993, 58, 4589–4595.
(19) (a) Nguyen, H. H.; Danen, W. C. J. Am. Chem. Soc. 1981, 103, 6253–
6255. (b) Danen, W. C.; Rio, V. C.; Setser, D. W. J. Am. Chem. Soc.
1982, 104, 5431–5440. (c) Hintsa, E. J.; Wodtke, A. M.; Lee, Y. T.
J. Phys. Chem. 1988, 92, 5379–5387.
(20) (a) Bagratashvilli, V. N.; Letokhov, V. S.; Makarov, A. A.; Ryabov,
E. A. Laser Chem. 1984, 5, 53–105. (b) Lupo, D. W.; Quack, M.
Chem. ReV. 1987, 87, 181–216.
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 531

A R T I C L E S
Ji et al.
Scheme 4. MP-IR Initiated Reaction of trans-16
in the gas phase were directly injected into the GC and each
experiment was analyzed in triplicate. The product ratios were
corrected for FID response factors and are reported in the
Supporting Information.
Figure 3. Product composition versus percent reaction of trans-16 under
MP-IR irradiation at 1057.3 cm-1. Red squares are trans-17, blue diamonds
are cis-17 and purple circles are cis-16. The lines are a linear least-squares
fit to the data; trans-17, y ) (0.13 ( 0.09)x + (53.7 ( 4.6); cis-17, y )
(0.37 ( 0.07)x + (6.2 ( 3.6); cis-16, y ) (-0.51 ( 0.05)x + (40.1 ( 2.8).
See Scheme 4.
Experimental Results
The first experiment performed was a conventional flash
vacuum pyrolysis of cis-16 (containing 6% trans-16) at 400
°C. As expected, this gave a mixture of trans-16 (19%) cis-17
(13%) and trans-17 (11%) along with unreacted cis-16 (57%)
This is consistent with Lewis’ results,16 in that all three products
are formed. It does not address the question as to whether
δ-elimination occurs via an eight-centered transition state;
therefore the MP-IR experiments were undertaken. Significantly,
cis-3-penten-2-ol (19) was recovered unchanged when it was
submitted to the pyrolysis conditions, indicating that a 6-centered
δ-elimination of water does not occur under these conditions.
The MP-IR induced reactions of both cis-16 and trans-16
were then studied. The results for trans-16 are more straight-
forward and will be discussed first. The anticipated reactions
are outlined in Scheme 4. The only mechanism for the formation
of cis-17 is the sequential process, rearrangement of trans-16*
to cis-16* followed by ꢀ-elimination to give cis-17. In principle,
it should be possible to completely quench the formation of
cis-17 at sufficiently high He pressures. In practice, increasing
the He pressure also cools trans-16* and thus quenches all the
reactions. Nevertheless, it is possible to extrapolate to infinite
He pressure.
Figure 4. Product composition versus percent reaction of cis-16 under MP-
IR irradiation at 1057.3 cm-1. Red squares are trans-17, blue diamonds are
cis-17 and green triangles are trans-16. The lines are a linear least-squares
fit to the data; are trans-17, y ) (0.31 ( 0.03)x + (23.4 ( 1.5); cis-17, y
) (0.05 ( 0.02)x + (45.0 ( 1.0); trans-16, y ) -0.37 ( 0.03)x + (31.6
( 1.5). See Scheme 3.
When trans-16 (1.7 torr) was subjected to MP-IR irradiation
at 1057.3 cm-1 at varying He pressures (0 to 20 torr) the
anticipated products trans-17, cis-17, and cis-16 were formed.
In accord with Lee’s MP-IR results on ethyl formate,19c ketene
was also observed by in situ IR and by GC, however, because
this was only a small quantity of a secondary product, it is not
considered in the following discussion. The product distributions
of trans-17, cis-17, and cis-16 were analyzed as follows. The
percent reaction was taken as the sum of the three products,
divided by the sum of these products and unreacted trans-16
(eq 7) Higher pressures of He as buffer gas led to lower overall
percent reaction. The percent of each product was taken as the
proportion of that product (trans-17, cis-17, or cis-16) divided
by the sum of these three products (eq 8). The percent of each
product as a function of percent reaction is shown in Figure 3.
This is to be expected, as the only reaction that forms trans-17
is one of the two reactions of trans-16* that lead directly to
products, in a ratio determined by their respective unimolecular
rate constants, and would be equally quenched by collisional
cooling of trans-16* by He. In contrast, the percentage of the
rearrangement product cis-16 increases significantly; it extrapo-
lates to 40 ( 3% with decreasing percent reaction (increasing
He pressure). Concomitantly, the proportion of cis-17 decreases;
it extrapolates to close to zero (6 ( 4%) as the He pressure is
increased. This is as expected if all of the cis-17 is formed
sequentially by rearrangement of trans-16* to cis-16* followed
by ꢀ-elimination. This is significant because it demonstrates that
the MP-IR technique, along with collisional cooling of hot
intermediate species, is able to identify the contributions to the
product ratio from the initial, unimolecular reactions, separate
from a second, but likewise unimolecular step.
In light of this, a similar series of MP-IR experiments were
performed on cis-16. The results were analyzed analogously to
eq 7 and 8 for trans-16 and are shown in Figure 4. The
anticipated reaction mechanism is shown in Scheme 3. Briefly,
we anticipate that some trans-17 is formed in a stepwise process
via trans-16*; collisional cooling by added He buffer gas will
increase the proportion of trans-16 at the expense of trans-17.
A Stern-Volmer plot of the cis-16 reaction versus He pressure
was linear, indicating that collisional cooling by He is occurring
sum GC yields cis-16,cis-17,trans-17
% reaction )
sum GC yields trans-16,cis-16,cis-17,trans-17
(7)
GC yield product
sum GC yields cis-16,cis-17,trans-17
%product )
(8)
The results in Figure 3 are consistent with the qualitative
reaction mechanism in Scheme 4. The proportion of trans-17
as a percentage of products is relatively constant (extrapolating
to 54 ( 5% of the initial reaction) as the He pressure is changed.
9
532 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Pseudopericyclic Eight-Centered cis-Elimination
A R T I C L E S
Scheme 5. Transition States Calculated for Reactions of cis-16
and trans-1621a
Table 1. Relative Energies (kcal/mol) and Low or Imaginary
Frequencies (cm-1) of Calculated Minima and Transition States in
the Reactions of 16 at the B3LYP/6-31G(d,p) Geometry with
MP4(SDTQ)/cc-pVDZ Single Point Energies (See Scheme 5)
c
structure
B3LYPa
(kcal/mol)
frequencyb
(cm-1
∆G
MP4d
(kcal/mol) (kcal/mol) (kcal/mol)
∆G MP4e
)
cis-16
trans-16
1.2
0.0
45.1
44.9
1.7
0.0
1.0
0.0
1.5
0.0
δ-c-t-TS
41.2
47.3
43.3
41.8
39.1
44.9
36.2
15.3
13.7
930.2i
1295.8i
940.7i
935.5i
414.7i
432.2i
391.5i
37.5
44.0
38.3
36.9
38.6
44.5
35.4
1.4
51.6
55.9
52.3
51.3
43.5
48.8
41.5
18.5
17.1
47.9
52.6
47.3
46.6
43.0
48.4
40.7
4.6
δ-c-c-TS
ꢀ-c-c-TS
ꢀ-t-t-TS
3,3-c-t-TS
3,3-c-c-TS
3,3-t-t-TS
HOAc+ cis-17
HOAc + trans-17
a Relative free energies (B3LYP/6-31G(d,p), 298.2 K) are in kcal/mol.
(Figure S8). However, any proportion of trans-17 that may be
formed directly from cis-16* via a δ-elimination cannot be
affected by added He buffer gas. Thus, the extrapolation to
infinite He pressure and complete collisional cooling should
reveal the percentage, if any, of trans-17 formed directly from
cis-16*. Indeed, as Figure 4 shows, the percentage of trans-17
thus formed directly from cis-16* is 23 ( 2%. It is also
significant that the proportion of cis-17 as a percentage of
products is relatively constant across the reaction conditions.
This is again consistent with a reaction scheme in which the
first reactions of cis-16* to give products are solely unimolecular
and cis-17 is only formed from cis-16*. In summary, the data
in Figure 4 show that cis-16* can undergo collisional cooling
back to cis-16 in competition with three unimolecular reactions,
forming trans-16, cis-17 and trans-17. Extrapolation to 0.0%
reaction (full collisional cooling) indicates that the initial
unimolecular reactions of cis-16* give trans-16, cis-17 and
trans-17 in 32 ( 2%, 45 ( 1% and 23 ( 2% yield respectively.
Since cis-3-penten-2-ol (19) does not undergo elimination
under FVP conditions, a six-centered δ-elimination can be ruled
out for the direct formation of trans-17 from cis-16*. This leaves
the eight-centered δ-elimination as the only mechanistic pos-
sibility. This confirms the qualitative prediction that both the
six-centered ꢀ-elimination and the eight-centered δ-elimination
are allowed because they are pseudopericyclic. Indeed, the
results in Figure 4 show that these two reactions are competitive,
with ꢀ-elimination slightly favored over δ-elimination (32 (
2% vs 23 ( 2%).
-0.1
3.3
a B3LYP/6-31G(d,p) energy relative to trans-16. b Lowest or imaginary
frequency from B3LYP/6-31G(d,p) vibrational calculation. c Relative free
energy at 298.2 K, based on B3LYP/6-31G(d,p) frequencies.
d MP4(SDTQ)/cc-pVDZ/B3LYP/6-31G(d,p)
relative
energies.
e MP4(SDTQ)/cc-pVDZ/B3LYP/6-31G(d,p) relative energies with B3LYP/
6-31G(d,p) free energy corrections.
relevant transition states are presented in Figure 5. The other
structures are shown in the Supporting Information.
It is worthwhile to first compare these results with the known
experimental and computational results. The racemization of trans-
16 is faster than the competitive ꢀ-elimination to form trans-17
(eq 5) and rearrangement to cis-16 is even less favored.16 And
indeed, at the B3LYP/6-31G(d,p) level, ∆Gq (at 298.2 K) is
calculated to be 35.4 kcal/mol for the racemization of trans-16 by
rearrangement to its enantiomer via 3,3-t-t-TS (experimental Ea )
38.8 kcal/mol16) This compares with 36.9 kcal/mol for the 1,5
elimination via ꢀ-t-t-TS and 38.6 kcal/mol for the rearrangement
to cis-16 via 3,3-c-t-TS. This is reasonable agreement, considering
that one reaction involves hydrogen migration from C to O and
the other two reactions simply exchange one C-O bond for another.
Interestingly, the MP4(SDTQ)/cc-pVDZ calculations (with B3LYP/
6-31G(d,p) free energy corrections) give higher barriers for the
rearrangements (40.7 and 43.0 kcal/mol) and a much higher barrier
for the elimination (46.6 kcal/mol); these calculations are in poorer
accord with the experimental results in that the rearrangements and
elimination from trans-16 are competitive. The activation energy
for elimination of ethyl formate calculated at the B3LYP/
6-31G(d,p) level (Ea ) 47.4 kcal/mol) is in closer agreement with
the experimental barrier (48.3 kcal/mol) than MP2 calculations with
various basis sets as reported by Rodr´ıguez-Otero et al.7e
Computational Results
We had previously reported B3LYP/6-31G(d,p) calculations on
model systems, including 3 and 4 (eq 2a and b) that suggested the
eight-centered δ-elimination was allowed and might be competi-
tive.3 These gave us some confidence to pursue the MP-IR
experiments described above. We now report a computational study
of the potential energy surface for the reactions of cis-16. Density
functional theory (B3LYP/6-31G(d,p) geometry optimization) and
ab initio calculations (MP4(SDTQ)/cc-pVDZ single point energies)
were used to investigate the mechanistic possibilities outlined in
Scheme 5 and to explore the pericyclic or pseudopericyclic nature
of the reactions.23,24 The relative energetics are reported in Table
1; in the following discussion, the B3LYP/6-31G(d,p) free energies
will be used, unless otherwise indicated. The geometries of the most
Although there are many examples where the B3LYP/6-31G(d,p)
level fails to reproduce experimental trends,25 there are also
numerous examples where it gives better agreement with experiment
than other levels of theory, particularly for orbital symmetry allowed
(24) Frisch, M. J.; et al. Gaussian 03, ReVision C.02, Gaussian, Inc.:
Wallingford CT, 2004. For the full reference, see the Supporting
Information.
(25) For a few examples, see: (a) Petersson, G. A.; Malick, D. K.; Wilson,
W. G.; Ochterski, J. W.; Montgomery, J. A.; Frisch, M. J. J. Chem.
Phys. 1998, 109 (24), 10570–10579. (b) Lynch, B. J.; Truhlar, D. G.
Phys. Chem. A 2001, 105, 2936–2941. (c) Schreiner, P. R.; Fokin,
A. A.; Robert, A.; Pascal, J.; Meijere, A. d. Org. Lett. 2006, 8, 3635–
3638. (d) Wodrich, M. D.; Corminboeuf, C.; Schleyer, P. v. R. Org.
Lett. 2006, 8, 3631–3634.
(26) For a few examples, see: (a) Houk, K. N.; Gonzalez, J.; Li, Y. Acc.
Chem. Res. 1995, 28, 81. (b) Wiest, O.; Montiel, D. C.; Houk, K. N.
J. Phys. Chem. A 1997, 101, 8378–8388. (c) Birney, D. M. J. Am.
Chem. Soc. 2000, 122, 10917–10925. (d) Guner, V.; Khuong, K. S.;
Leach, A. G.; Lee, P. S.; Bartberger, M. D.; Houk, K. N. J. Phys.
Chem. A 2003, 107, 11445–11459. It is also worth noting that a larger
basis set does not necessarily improve the accuracy of B3LYP
calculations. (e) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Pople,
J. A. J. Chem. Phys. 1997, 106, 1063–1079.
(23) Calculations were performed using Gaussian 03.24 All stationary points
were characterized by vibrational frequency analysis and all transition
states had a single imaginary frequency. ꢀ-SOSP-3 had two imaginary
frequencies, as required for a second-order saddle-point. Other
constrained geometries were not stationary points. Optimized geom-
etries, absolute energies and further details of the calculations are
available in the Supporting Information. Three conformations each
of cis-16 and trans-16 were calculated and are reported in the
Supporting Information; only the lowest energy ones are discussed in
the article.
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 533

A R T I C L E S
Ji et al.
Figure 5. Top, side, and end views of B3LYP/6-31G(d,p) optimized geometries of transition states (A) ꢀ-c-c-TS, (B) 3,3-c-t-TS, and (C) δ-c-t-TS for the
reactions of cis-16 and trans-16. Carbons are gray, hydrogens are white and oxygens are red. Distances for breaking and forming bonds are in Ångstroms.
The top view is looking above the approximately planar cyclic transition state, the side view is looking at the breaking and forming bonds, and the end view
is looking down the acetate CH3-C bond.
reactions.26 In this system, the calculated differences between the
rearrangements and eliminations at the MP2/6-31G(d,p) level are
significant and do not follow the experimental trends; the agreement
of the B3LYP/6-31G(d,p) free energies with the experimental
trends gives us some degree of confidence in the results.
pseudopericyclic based on transition state aromaticity considerations
(NICS11c,d or ACID11e analysis), although some have also discussed
“borderline pseudopericyclic reactions”.2e,j,l More recently, Chamor-
ro has proposed a scale of pericyclic/pseudopericyclic character
for transition states, based on his ELF analysis.2s The Atoms in
Molecules analysis of Bader27 has also been explored as a tool for
distinguishing the range of these reactions. The ellipticity of the
electron density at the bond critical point of the forming bond is
more exaggerated in some pericyclic transition states and less so
in related pseudopericyclic ones.2l,u The Laplacian of the electron
density (but not the ellipticity) at the bond critical points of the
forming bonds has been shown to be different in pericyclic and
pseudopericyclic transition states of [4 + 2] cycloadditions.2v
To address the pericyclic/pseudopericyclic question for ester
eliminations, we first compare the retro-ene reaction of 1-pentene
(5), a classic six-electron, suprafacial pericyclic reaction8,11,12 with
the ꢀ-elimination of formic acid from ethyl formate (3)4 which is
allowed either via suprafacial pericyclic (as described in the
literature5-7) or planar pseudopericyclic transition states.1-3 The
transition state geometries of these (ꢀ-TS-5 and ꢀ-TS-3) are
significantly different, as shown in Table 2 and Figure 6.
The B3LYP/6-31G(d,p) free energy calculations also agree with
the results of our MP-IR studies on the reaction of cis-16. Of the
five transition states from cis-16 (δ-c-t-TS, δ-c-c-TS, ꢀ-c-c-TS,
3,3-c-t-TS and 3,3-c-c-TS, Scheme 5) the eight-centered elimination
via δ-c-t-TS is calculated to have the lowest activation energy (∆Gq
) 37.5 kcal/mol). However, the free energy barriers for three
reactions that have been shown in these experiments to be
competitive (δ-c-t-TS, ꢀ-c-c-TS and 3,3-c-t-TS) are within about
1 kcal/mol at this level of theory. The other two transition states
from cis-16 (δ-c-c-TS and 3,3-c-c-TS) have higher barriers. The
eight-centered elimination to form cis-1,3-pentadiene via δ-c-c-
TS is calculated to have a barrier of 44.0 kcal/mol; the degenerate
3,3-rearrangement to form the enantiomer of cis-16 (3,3-c-c-TS)
has a barrier of 44.5 kcal/mol. These higher calculated barriers
justify ignoring these pathways in the discussion of the MP-IR
results above.
Experimentally, δ-elimination of acetic acid from cis-16 occurs,
as evidenced by our MP-IR studies; B3LYP/6-31G(d,p) calcula-
tions suggest this occurs via an eight-centered transition state (δ-
c-t-TS). this can be understood as a pseudopericyclic reaction, in
which there are one or more orbital disconnections in the cyclic
transition state. How well do the calculated transition states agree
with this qualitative picture? In the discussion below, we argue
that they agree quite well. Moreover, this series of transition states
illustrate a more general concept. For reactions where both
pseudopericyclic and allowed pericyclic transition state orbital
interactions are possible, these can be considered as two electronic
states that are of like symmetry. As is the case for other examples
of electronic states of like symmetry, they can mix and produce a
state of lower energy; the closer in energy the two states are, the
more they mix. Thus, if the transition state shows a deviation from
the ideal geometry for either a pericyclic or pseudopericyclic
reaction, it cannot be purely one or the other, but requires a
proportional contribution from each, as we have previously
suggested.1c There have been significant arguments in the
literature2e-o,p-u trying to classify reactions as either pericyclic or
The transition state (ꢀ-TS-5) for the retro-ene reaction of 5 has
been calculated at a number of theoretical levels8 including B3LYP/
6-31G(d)8b and CASPT2.8c The geometries are similar at all levels.
The B3LYP/6-31G(d)8b transition state is reproduced in Figure
6E; the barrier for the retro-ene is 58.1 kcal/mol at this level.8b
Selected occupied molecular orbitals (MOs) and natural bond
orbitals28 (NBOs) are shown in Figure S3 (Supporting Information)
and clearly reflect the qualitative orbital picture in Figure 1A. The
breaking σ-bonds are nearly coplanar, as required for the formation
of the new C4-C5 π-bond; the C3-C4-C5-H6 dihedral angle
is 11.7°. As also required for suprafacial orbital overlap in the allyl
π-system, the allyl fragment is tipped far out of the plane of the
breaking σ-bonds; the dihedral angles C1-C2-C3-C4 and
(27) Bader, R. F. Atoms in Molecules - A Quantum Theory; number 22 in
The International Series of Monographs on Chemistry; Oxford
University Press: Oxford, 1990.
(28) (a) Weinhold, F.; Carpenter, J. E. The Structure of Small Molecules
and Ions; Plenum: New York, 1988. (b) Reed, A. E.; Curtiss, L. A.;
Weinhold, F. Chem. ReV. 1988, 88, 899–926.
9
534 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Pseudopericyclic Eight-Centered cis-Elimination
A R T I C L E S
Table 2. Selected Dihedral Angles of Transition States Calculated
state ꢀ-TS-3 is primarily stabilized by the pseudopericyclic interac-
tions, even though the pericyclic interactions are slightly manifest
in the transition state geometry.
at the B3LYP/6-31G(d,p) Level, unless Otherwise Indicateda
dihedral
Returning to the transition states for the eight-centered δ-elim-
inations of acetic acid from cis-16 (δ-c-t-TS, Figure 5C and δ-c-
c-TS, Supporting Information), these would be forbidden if they
had the same orbital topology as the retro-ene of 1-pentene (ꢀ-TS-
5). They would be allowed if they were pericyclic and antarafacial
on one component or alternatively if they were pseudopericyclic.
The acetate fragment is nearly coplanar with the breaking C-O
bond and the forming O-H bond; the O1-C2-O3-C4 dihedrals
are 7.9 and -2.8° and the H8-O1-C2-O3 dihedrals are 4.1 and
-4.4° respectively. Interestingly, the acetate fragments are very
slightly twisted toward the geometry that would be required for
antarafacial overlap12 (Mobiu¨s aromaticity)11 of the acetate fragment
with the breaking and forming bonds, as seen in the end view of
δ-c-t-TS in Figure 5C. Constraining these dihedral angles to 0.0°
in δ-c-t-TS raises the energy by an inconsequential amount, 0.1
kcal/mol (Table S4, Supporting Information); the calculated free
energy of δ-c-t-TS is lower by 0.1 kcal/mol at this nonstationary
point. Based on these energetics, this is essentially a purely
pseudopericyclic reaction. The near planarity of the transition states
is consistent with the qualitative geometry expected for a pseudo-
pericyclic transition state in which the breaking and forming bonds
are in the plane of the acetate fragment. The molecular orbitals are
likewise consistent with a pseudopericyclic transition state, showing
a clear separation between the acetate π- and σ-systems (Figure
S4, Supporting Information). The breaking C-O and C-H σ-bonds
are close to coplanar; the O3-C4-C7-H8 dihedral angles are 10.9
and -1.6° respectively, as expected for the formation of the new
diene π-bond system.
The geometries of the transition states for the ꢀ-eliminations (ꢀ-
c-c-TS and ꢀ-t-t-TS) are not quite as planar as the pseudopericyclic
δ-eliminations (δ-c-c-TS, and δ-c-t-TS). The O1-C2-O3-C4
dihedrals are -5.5 and -6.3° and the H6-O1-C2-O3 dihedrals
are 7.1 and 8.1° in ꢀ-c-c-TS and ꢀ-t-t-TS respectively, indicating
a distortion toward suprafacial orbital overlap on the acetate
fragment. This can be seen in the end view of ꢀ-c-c-TS in Figure
5A; note that both of the breaking and forming bonds are to the
same side of the acetate fragment. This must mean that the transition
states are not purely pseudopericyclic. However, these geometries
are far from being as significantly out-of-plane as the retro-ene
reaction transition state (ꢀ-TS-5), as discussed above, and so ꢀ-c-
c-TS and ꢀ-t-t-TS are not purely pericyclic either.
The geometries of the [3,3]-sigmatropic rearrangements are more
distorted toward a pericyclic transition state than are the ꢀ- or
δ-eliminations. For 3,3-c-t-TS, 3,3-c-c-TS, and 3,3-t-t-TS, the
O1-C2-O3-C4 dihedrals are 37.3, 43.6, and 33.3° and the
C6-O1-C2-O3 dihedrals are -34.8, -35.1, and -33.3°, respec-
tively. These compare well with the angles calculated at the MP2/
6-31G(d) level for allyl formate [3,3]-rearrangement boat transition
state.32 However, these angles are still far from those in a
hydrocarbon Cope reaction (65.1° at the B3LYP/6-31G(d) level).29
As might be expected, constraining 3,3-c-t-TS to a planar geometry
raises the energy a bit more, this nonstationary point is 1.6 kcal/
mol higher in energy (Table S4, Supporting Information).
In the discussion above, geometry and energy have been used
to compare pericyclic and pseudopericyclic contributions to the
transition states. Are the calculated transition state aromaticities
consistent as well? The nucleus independent chemical shift (NICS)
has been used to provide a quantitative measure of aromaticity in
both ground and transition states.11c,d In the simplest systems, the
Structure
1234
2347
3478
8123
δ-c-t-TS
δ-c-c-TS
7.9 -13.8
-2.8
10.9
4.1
-4.4
4.8 -1.6
1234 2345 3456 4561 5612 6123 3461
g
g
g
g
ꢀ-c-c-TS
-5.5 -3.1 -6.3
-6.3 -3.1 6.7
37.3
43.6 -0.6 -59.2 66.4 -16.4 -35.1 6.6
33.3
-63.5
0.0
7.1
8.1
ꢀ-t-t-TS
3,3-c-t-TS
9.4 -64.7 66.4 -14.0 -34.8 1.7
3,3-c-c-TS
3,3-t-t-TS
13.9 -65.9 65.9 -13.9 -33.3 0.0
37.6
0.0
8.1
0.0
ꢀ-TS-5b
11.7
0.0
8.9
0.0
1.0
57.3
0.0
30.2
0.0
g
g
g
g
g
g
g
g
g
g
ꢀ-TS-3c B3LYP
ꢀ-TS-3d MP2
ꢀ-SOSP-3e MP2
ꢀ-ene-3f MP2
-32.7
0.0
-63.5
27.3
57.3
a Atom numbering is as in Figures 5 and 6. b B3LYP/6-31G(d)
optimization, from ref 8b. c B3LYP/6-31G(d,p) tight optimization. d MP2/
6-31G(d,p) geometry from ref 7d. e This planar structure is a second-order
saddle point at the MP2/6-31G(d,p) level. f MP2/6-31G(d,p) optimized
with the 1234 and 6123 dihedral angles constrained to -63.5° and 57.3°
respectively (from ꢀ-TS-5). g Because the partial bonds to H6 are nearly
linear, the 4561 and 5612 dihedral angles are not particularly meaningful.
C3-C2-C1-H6, are -63.5 and 57.3°, respectively. These are
similar to the corresponding dihedral angles in other typical
hydrocarbon transition states. For example, in the Cope reaction
of 1,5-hexadiene this angle is 65.1° at the B3LYP/6-31G(d) level29
and in the transition state for Diels-Alder reaction between 1,3-
cyclohexadiene and quinone, this dihedral angle is 64.6° at the
B3LYP/6-31G(d,p) level.30
The transition state for ꢀ-elimination of ethyl formate at the
B3LYP/6-31G(d,p) level has been variously reported to have a
planar7b or a slightly nonplanar geometry.7e We started from a
nonplanar geometry and performed an optimization at the B3LYP/
6-31G(d,p) level with tight optimization cutoffs; this reproduces
theplanarstructure7b(ꢀ-TS-3,O1-C2-O3-C4andH6-O1-C2-O3
dihedrals of 0.0°) shown in Figure 6A. The HOMO and HOMO-1
are shown in Figure 1C,D; other selected occupied MOs and
NBOs28 are shown in the Supporting Information and clearly reflect
the qualitative orbital picture in Figure 1B. The calculated barrier
is 50.4 kcal/mol; this is lower than that calculated for 5, as expected
when comparing pseudopericyclic and pericyclic reactions. The
MP2 method with a variety of basis sets predicts slightly nonplanar
transition states3,7e,31 that are close in energy to the planar second-
order saddle point (ꢀ-SOSP-3), only 0.2 kcal/mol more stable at
the MP2/6-31G* and G2(MP2) levels.3 Figure 6 shows the MP2/
6-31G(d,p) nonplanar structure7e (ꢀ-TS-3, Figure 6C) and the
planar structure (ꢀ-SOSP-3, Figure 6B); the latter is still only 0.2
kcal/mol higher in energy.
We also performed a constrained transition state optimization
(ꢀ-ene-3, Figure 6D) at the MP2/6-31G(d,p) level that is analogous
to the retro-ene reaction of 1-pentene, fixing the O1-C2-O3-C4
and H6-O1-C2-O3 dihedral angles to -63.5° and 57.3°, which
are those calculated for ꢀ-TS-5 at the B3LYP/6-31G(d,p) level.
This nonstationary point (ꢀ-ene-3) is calculated to be 4.2 kcal/mol
higher in energy (without free energy corrections) than ꢀ-TS-3.
Distorting toward the nonplanar Woodward-Hoffmann suprafacial
orbital topology in ꢀ-ene-3 requires substantially more energy than
does the distortion to the planar, pseudopericyclic orbital topology
in ꢀ-SOSP-3. This suggests that the MP2/6-31G(d,p) transition
(32) Pascal, Y. L.; Levisalles, J.; Normant, J.-M. Tetrahedron 1993, 49,
7679–7690. A chair transition state is also found for the rearrangement
of allyl formate; this is 1.9 kcal/mol higher in energy than the boat
transition state at the MP2/6-31G* + ZPE level. Attempts to locate
a chair transition state corresponding to 3,3-c-t-TS were unsuccessful
at the B3LYP/6-31G(d,p) level; the optimizations proceeded smoothly
to give 3,3-c-t-TS.
(29) Wiest, O.; Black, K. A.; Houk, K. N. J. Am. Chem. Soc. 1994, 116,
10336–10337.
(30) Birney, D.; Lim, T. K.; Koh, J. H. P.; Pool, B. R.; White, J. M. J. Am.
Chem. Soc. 2002, 124, 5091–5099.
(31) MP2 calculations on pericyclic reactions often give activation energies
that are lower than experiment.26
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 535

A R T I C L E S
Ji et al.
Figure 6. Side and top views of calculated transition states. (A) ꢀ-TS-3 (B3LYP/6-31G(d,p), tight optimization, compare with refs 7b [planar] and 7e
[nonplanar]), (B) MP2/6-31G(d,p) second order saddle point (ꢀ-SOSP-3, this work), (C) MP2/6-31G(d,p) transition state (ꢀ-TS-3, reproduced from ref
7e), (D) MP2/6-31G(d,p) constrained to the 1234 and 3216 dihedral angles of -63.5° and 57.3° respectively, corresponding to those of ꢀ-TS-5. This is a
transition state within the constraints, but is not a stationary point. (E) ꢀ-TS-5 (B3LYP/6-31G(d), reproduced from ref 8b). See Figure 5 for key. Barrier
heights are at the corresponding level of theory without any vibrational corrections.
at this geometry as well. The ꢀ-and δ-eliminations of the esters
have low transition state aromaticitity (minimum NICS value e-3
and e-5, respectively). The δ-eliminations have more negative
NICS values on the side near the alkene carbons (C6 and C6); the
NICS value at the center of the ring is e -3), comparable to the
ꢀ-eliminations. The 3,3-rearrangements (3,3-c-c-TS, 3,3-c-t-TS, and
3,3-t-t-TS) show slightly more transition state aromaticity and
are more distorted out of plane than the eliminations. Even here,
the most negative points are near the central allyl carbon (C5, Figure
5); the NICS values at the center of the ring are smaller, while
constraining the reaction to be planar on the acetate (3,3-c-t-TS-
planar) makes the reaction substantially less aromatic (minimum
NICS value -1, Figure S7, Supporting Information). When the
slightly nonplanar elimination transition states (ꢀ-SOSP-3 and δ-c,t-
TS-planar) are constrained to be planar on the acetate, the NICS
curves are quite close (Figure S7, Supporting Information). These
NICS results are consistent with the argument that the pseudoperi-
cyclic reactions lack the transition state aromaticity of pericyclic
reactions. They also tend to support the suggestion that pericyclic
and pseudopericyclic orbital topologies can mix, giving rise to
transition states that are likewise mixed.
Conclusions
Figure 7. NICS values calculated for the indicated transition states. See
(1) MP-IR of trans-16 at 1057.3 cm-1 gives rearrangement
to cis-16 and elimination to a mixture cis- and trans-17.
text for a description of the calculations. Key ꢀ-TS-5 black (, ꢀ-TS-3 blue
(, ꢀ-ene-3 blue ×, δ-c-t-TS red b, δ-c-c-TS red 9, 3,3-c-t-TS burgandy
2, 3,3-c-c-TS burgandy (, 3,3-t-t-TS burgundy 9, ꢀ-c-c-TS green 2, ꢀ-t-
t-TS green b.
However, adding He buffer gas increases the yield of cis-16,
and concomitantly reduces the formation of cis-17. Extrapolation
to complete cooling indicates that essentially all of the formation
of cis-17 is stepwise, by rearrangement from trans-16* to cis-
16* and then elimination to cis-17, while all of the trans-17 is
formed directly from trans-17*.
(2) Similarly, MP-IR of cis-16 at 1057.3 cm-1 gives rear-
rangement to trans-16 and elimination to a mixture cis- and
trans-17. Adding He buffer gas increases the yield of trans-16,
and concomitantly reduces the formation of trans-17. However,
extrapolation to complete cooling indicates that approximately
23% of the observed trans-17 is formed directly from cis-16,
while 45% of the reaction directly forms cis-17 and 32% forms
trans-16. The remainder of the observed trans-17 is formed by
the stepwise process, rearrangement from cis-16* to trans-16*,
chemical shift is evaluated at the center of the aromatic ring and at
points above and below the plane of that ring. This is complicated
by the lack of symmetry and nonplanar geometry of many transition
states. In this work, the center was taken as the geometric center
of the four heavy atoms (C, O) with partial σ-bonds. The chemical
shift was then evaluated at five points 0.2 Å apart above and below
the approximate plane defined by these four atoms. The results
(RHF/6-31+G(d) GIAO) are summarized in Figure 7. The retro-
ene reaction of 1-pentene (ꢀ-TS-5) shows a typical pattern for a
pericyclic transition state,11c,d with significant aromaticity (a
minimum NICS value of -28). Constraining the ethyl formate
elimination to a similar geometry (ꢀ-ene-3) shows less, but still
significant transition state aromaticity (a minimum NICS value of
-12); this is consistent with some pseudopericyclic orbital overlap
9
536 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Pseudopericyclic Eight-Centered cis-Elimination
A R T I C L E S
followed by elimination to trans-17. Significantly, this work
demonstrates that cis-16* undergoes rearrangement in competi-
tion with both ꢀ- and δ-eliminations. Because alcohol 19 does
not eliminate under the flash vacuum conditions that decompose
cis-16, the δ-eliminations arguably occurs via an eight-centered
transition state. Qualitative considerations suggest that both the
ꢀ- and δ-eliminations are pseudopericyclic.
the furthest from a planar, pseudopericyclic geometry. However,
these are far from the nonplanar transition state geometries
calculated for the hydrocarbon Cope rearrangement. Constrain-
ing 3,3-c-t-TS to planarity raises the energy by a modest 1.6
kcal/mol. The NICS values are still relatively small. These
calculations are also consistent with primarily pseudopericyclic
transition states.
(3) The general trends in the barrier heights for the competing
reactions of cis-16 and trans-16 are better reproduced by
transition state calculations at the B3LYP/6-31G(d,p) level than
at the MP4(SDTQ)/cc-pVDZ//MP2/6-31G(d,p) level. The
transition state geometries for the δ-eliminations, δ-c-c-TS, and
δ-c-t-TS, are essentially planar on the acetate fragment. The
molecular orbitals and the Natural Bond Orbitals show a clear
σ, π-separation. Constraining δ-c-t-TS to a planar geometry
raises the energy by an inconsequential 0.1 kcal/mol. The NICS
values are small, as compared to more pericyclic ꢀ-eliminations
(ꢀ-TS-5 and ꢀ-ene-3). These are all consistent with δ-elimina-
tions being primarily pseudopericyclic.
(4) The calculated transition state geometries for ꢀ-elimina-
tions of esters are either planar, as for ethyl formate (ꢀ-TS-3)
at the B3LYP/6-31G(d,p) level, or slightly nonplanar, as for
ꢀ-TS-3 at the MP2/6-31G(d,p) level and ꢀ-c-c-TS, and ꢀ-t-
t-TS at the B3LYP/6-31G(d,p) level. The NICS values are very
small. The planar ethyl formate second order saddle point (ꢀ-
SOP-3) is only slightly higher in energy (0.2 kcal/mol) than
the nonplanar transition state, but constraining the dihedral
angles to those of the hydrocarbon ene reaction raises the energy
by 4.0 kcal/mol. The geometric, energetic and aromaticity
criteria are consistent with these as pseudopericyclic reactions.
(5) The calculated boat transition state geometries for the
[3,3]-rearrangements (3,3-c-c-TS, 3,3-c-t-TS and 3,3-t-t-TS) are
(6) We suggest that in general, when both a pseudopericyclic
and a pericyclic orbital topology are possible for a reaction,
the transition state can involve mixing of these two electronic
states. The degree of mixing can be manifest in the energy of
the transition state, as compared to that of either of the ideal
transition states. The geometry can likewise provide an indica-
tion of the degree of mixing of prototypical pseudopericyclic
(planar) and pericyclic (nonplanar) allowed transition states.
Indeed, we suggest that any transition state property should
reflect this mixing.
Acknowledgment. This work was supported by grants from
the Robert A. Welch Foundation and the National Science Founda-
tion (0415622). We thank Ms. Emily Jenkins for assistance with
preliminary calculations. D.M.B. thanks Professor Peter Chen and
his group members for hospitality during a development leave and
Dr. Erik Pieter Adriaan Couzijn for helpful discussions.
Supporting Information Available: Experimental details for
synthesis and MP-IR of cis-16 and trans-16. Computational
details including absolute energies and Cartesian coordinates
of all optimized structures. Complete ref 24. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA804812C
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 537