Mechanisms of the thermal cyclotrimerizations of fluoro- and chloroacetylenes: Density functional theory investigation and intermediate trapping experiments

Article abstract of DOI:10.1021/ja2021476

Theoretical studies of the mechanisms of the thermal cyclotrimerization of fluoro- and chloroacetylenes, which were reported by Viehe and Ballester, respectively, were conducted with the aid of density functional theory calculations of the (U)B3LYP functional, indicating that the thermal cyclotrimerizations of fluoro- and chloroacetylenes involve tandem processes of regioselectively stepwise [2+2] and stepwise [4+2] cycloadditions. These tandem processes generate 1,2,6-trihalo-Dewar benzenes and 1,2,4-trihalo-Dewar benzenes, which then isomerize to the corresponding benzenes when heated. The rate-determining step of the cyclotrimerizations of haloacetylenes is the dimerization step involving open-shell singlet diradical transition states and intermediates. The substituent effects in the thermal cyclotrimerization of haloacetylenes have been rationalized using frontier molecular orbital theory. The higher reactivity of fluoroacetylenes compared to that of chloroacetylenes is due to the fact that fluoroacetylenes have lower singlet-triplet gaps than chloroacetylenes and more easily undergo dimerization and cyclotrimerization. In this report, additional experiments were performed to verify the theoretical prediction about the cyclotrimerization of chloroacetylene and to trap the proposed 1,4-dichlorocyclobutadiene intermediate. Experiments revealed that the thermal reaction of phenylchloroacetylene at 110 °C gave 1,2,3-triphenyltrichlorobenzene and 1,2,4-triphenyltrichlorobenzene together with a tetramer, cis-1,2,5,6-tetrachloro-3,4,7,8-tetraphenyltricyclo[4.2.0. 0^2,5]octa-3,7-diene. The proposed 1,4-diphenyldichlorocyclobutadiene intermediate in the thermal cyclotrimerization of phenylchloroacetylene was successfully trapped using dienophiles of maleic anhydride and dimethyl acetylenedicarboxylate.

Full text of DOI:10.1021/ja2021476

ARTICLE
pubs.acs.org/JACS
Mechanisms of the Thermal Cyclotrimerizations of Fluoro- and
Chloroacetylenes: Density Functional Theory Investigation and
Intermediate Trapping Experiments
Zhong-Ke Yao and Zhi-Xiang Yu*
Beijing National Laboratory for Molecular Sciences, Key Laboratory of Bioorganic Chemistry and Molecular Engineering of the Ministry
of Education, College of Chemistry, Peking University, Beijing 100871, China
S Supporting Information
b
ABSTRACT: Theoretical studies of the mechanisms of the thermal
cyclotrimerization of fluoro- and chloroacetylenes, which were
reported by Viehe and Ballester, respectively, were conducted with
the aid of density functional theory calculations of the (U)B3LYP
functional, indicating that the thermal cyclotrimerizations of fluoro-
and chloroacetylenes involve tandem processes of regioselectively
stepwise [2þ2] and stepwise [4þ2] cycloadditions. These tandem
processes generate 1,2,6-trihalo-Dewar benzenes and 1,2,4-trihalo-
Dewar benzenes, which then isomerize to the corresponding
benzenes when heated. The rate-determining step of the cyclotri-
merizations of haloacetylenes is the dimerization step involving open-shell singlet diradical transition states and intermediates. The
substituent effects in the thermal cyclotrimerization of haloacetylenes have been rationalized using frontier molecular orbital theory.
The higher reactivity of fluoroacetylenes compared to that of chloroacetylenes is due to the fact that fluoroacetylenes have lower
singletꢀtriplet gaps than chloroacetylenes and more easily undergo dimerization and cyclotrimerization. In this report, additional
experiments were performed to verify the theoretical prediction about the cyclotrimerization of chloroacetylene and to trap the
proposed 1,4-dichlorocyclobutadiene intermediate. Experiments revealed that the thermal reaction of phenylchloroacetylene at
110 ꢀC gave 1,2,3-triphenyltrichlorobenzene and 1,2,4-triphenyltrichlorobenzene together with a tetramer, cis-1,2,5,6-tetrachloro-
3,4,7,8-tetraphenyltricyclo[4.2.0.02,5]octa-3,7-diene. The proposed 1,4-diphenyldichlorocyclobutadiene intermediate in the ther-
mal cyclotrimerization of phenylchloroacetylene was successfully trapped using dienophiles of maleic anhydride and dimethyl
acetylenedicarboxylate.
’ INTRODUCTION
reported by Viehe and co-workers.6a,b This reaction was found
to be spontaneous below 0 ꢀC, yielding Dewar benzene 2,
benzvalene 3, and an unidentified tetramer 4. The trimers 2
and 3 were roughly equal in yield, and their total yield accounted
for two-thirds of the entire reaction products. Tetramer 4 (its
structure was not determined) accounted for 1ꢀ3% of the entire
reaction products. When 2 and 3 were heated at a higher
temperature (100 ꢀC), they isomerized to 1,2,3-tri-tert-butyltri-
fluorobenzene 5 and 1,2,4-tri-tert-butyltrifluorobenzene 6, re-
spectively (Scheme 1).
Alkyne cyclotrimerization to form benzene derivatives is one
of the most intensely studied reactions because of the ubiquitous
presence of substituted benzenes in molecular sciences.1 Alkyne
cyclotrimerization reactions are typically performed in the pre-
sence of catalysts, which can be transition metals,2 Lewis acids,3
amines,4 or disilanes.5 Uncatalyzed thermal cyclotrimerization of
alkynes to give benzene derivatives was rarely reported,6 though
the first uncatalyzed thermal cyclotrimerization of acetylene7 was
discovered ∼80 years before the first report of its catalytic
version.2a The process of uncatalyzed thermal cyclotrimerization
of alkynes to benzene derivatives was usually conducted at a very
high temperature to overcome the high reaction barrier. In 1866,
Berthelot reported the first example of the thermal transforma-
tion of acetylene to benzene. The reaction was conducted at
400 ꢀC, giving a complex mixture of products.7 A concerted
pathway was proposed for the cyclotrimerization of acetylene,
and the computed activation energy of the reaction was higher
than 50 kcal/mol at different calculation levels.8 An interesting
exception to all uncatalyzed and catalyzed alkyne cyclotrimeriza-
tions is the cyclotrimerization of tert-butylfluoroacetylene 1,
The proposed mechanism9 of the thermal cyclotrimerization
described by Viehe is shown in Scheme 2. The cyclotrimerization
was believed to start with the dimerization of two fluoroalkynes,
giving rise to tetrahedrane 7a, which was proposed to be in
equilibrium with planar 1,4-difluorocyclobutadiene derivative
7b, and 1,3-difluorocyclobutadiene derivative 7c. Intermediate
7b/7c could also exist in a diradical form of 7b0/ 7c0. Then 7b and
7c reacted with the third fluoroalkyne to give trimers 2 and 3,
Received: March 9, 2011
Published: June 23, 2011
r
2011 American Chemical Society
10864
dx.doi.org/10.1021/ja2021476 J. Am. Chem. Soc. 2011, 133, 10864–10877
|

Journal of the American Chemical Society
ARTICLE
respectively. Finally, upon being heated, Dewar benzene deriva-
tive 2 and benzvalene derivative 3 isomerized to the aromatic
1,2,3-tri-tert-butyltrifluorobenzene 5 and 1,2,4-tri-tert-butyltri-
fluorobenzene 6, respectively.
In 1986, Ballester and co-workers6c found that perchlorophe-
nylacetylene 8 can also undergo cyclotrimerization upon being
heated in perchlorostyrene, giving perchloro-1,2,3-triphenylben-
zene 9, perchloro-1,2,4-triphenylbenzene 10, and an unidentified
isomeric C24Cl18 compound 11 (Scheme 3). Formation of 9 and
10 was also thought to proceed via a mechanism involving
diradical cyclobutadiene and terahedrane intermediates, similar
to the mechanism of fluoroacetylene cyclotrimerization pro-
posed by Viehe.
Scheme 1. Thermally Spontaneous Cyclotrimerization of
tert-Butylfluoroacetylene
In 1993, Hopf and Witulski reported that 1,2,4- and 1,2,3-
tricyanobenzenes were formed with other products when cyano-
acetylene was heated at 160 ꢀC in a sealed tube.6d Recently,
they gave the energies of the products and the proposed
intermediates, but no information about the transition states
was given.10
The interesting phenomena of these uncatalyzed thermal cyclo-
trimerizations of alkynes have attracted some attention.8,10ꢀ12
However, several questions remain unanswered because of the
scarce experimental and computational evidence. Why can these
alkyne cyclotrimerization reactions occur under thermal reaction
conditions? Why are fluoroacetylenes more reactive than chloro-
acetylenes? What factors control the regioselectivity in the
thermal cyclotrimerization that gives different benzene derivatives
(Schemes 1 and 3)? In particular, how does this process generate
Scheme 2. Mechanism Proposed by Viehe for the Cyclotrimerization of tert-Butylfluoroacetylene
Scheme 3. Thermal Cyclotrimerization of Perchlorophenylacetylene
10865
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Scheme 4. Mechanism of the Cyclotrimerization of Fluoroacetylene Supported by DFT Calculations
the uncommon 1,2,3-trisubstituted benzene, which is difficult to
produce by metal-catalyzed cyclotrimerizations of substituted
alkynes? Another interesting question remained unanswered. What
is the tetramer in Viehe’s reaction?
states and intermediates, the YJH spin-projection22 scheme has been
used to reduce spin contamination. Orbital energies were computed
using the HF/6-31G(d) method,23 and the orbital coefficients listed are
the 2p orbitals of carbons and fluorines and 3p orbitals of chlorines. ΔG,
ΔH, and ΔE0 are the calculated relative free energies, relative enthalpies,
and zero-point energy (ZPE) corrected relative electronic energies in the
gas phase, respectively. Unless otherwise specified, all discussed energies
in what follows refer to ΔG values.
Here we report the first computational study of the uncata-
lyzed cyclotrimerizations of haloacetylenes, aiming to answer the
questions listed above. We believe that answering these questions
is critical not only to our in-depth mechanistic understanding of
these reactions but also to guiding the future design of other
uncatalyzed cyclotrimerizations and studying the potential ap-
plication of these reactions in material sciences and other
disciplines. It is well-known that fluoroacetylene derivatives are
highly reactive (sometimes explosive) and difficult to handle.13
Therefore, experimental investigation of fluoroacetylenes’ reac-
tions and reaction mechanisms and the experimental trapping of
the proposed intermediates are very difficult. However, experi-
mental studies of cyclotrimerizations of chloroacetylenes, if they
can occur thermally, could be easily performed because chloro-
acetylenes are stable and can be easily prepared. Because
of these facts, we investigated experimentally the thermal cyclo-
trimerization of arylchloroacetylene and trapped one of the
proposed cyclobutadiene intermediates with the aim of further
confirming or complementing our computational study. Further-
more, we expected that these new experiments could also serve
to identify the possible tetramer in Viehe‘s reaction and
provide information about the unidentified product in Ballester’s
reaction.
’ RESULTS AND DISCUSSION
Density Functional Theory (DFT) Study of the Cyclotri-
merization of Fluoroacetylene. For the sake of calculation
efficiency, we chose the fluoroacetylene reaction as the model
reaction instead of the cyclotrimerization reaction of experimen-
tally used tert-butylfluoroacetylene in the first attempt to under-
stand the reaction mechanisms. Because the used fluoroacetylene
is unsymmetric, there are several possible pathways for the
cyclotrimerization of this fluoroacetylene, which had all been
investigated by our DFT calculations. To present the results of
our calculation clearly and concisely, we will first elucidate how
the cyclotrimerization occurs in its most favored reaction path-
way. Details of structures and energetics of the stationary points
involved in this favored pathway will also be given. Then, we will
present other competing but unfavored pathways, aiming to
explain the regioselectivities of the cyclotrimerizations of
haloacetylenes.
Favored Pathways for the Cyclotrimerization of Fluoroace-
tylene. The reaction pathways showing how the cylcotrimeriza-
tion gives two benzene derivatives, P-1 and P-2, are summarized
in Scheme 4. The computed energy surface for these processes is
given in Figure 1. Structures of reactant, intermediate, and
transition states are shown in Figure 2.
Our calculations suggest that cyclotrimerization starts with
dimerization of fluoroacetylene 1a to intermediate INT2a-1. The
dimerization step occurs stepwise and involves the formation of a
diradical intermediate INT1a-1 (Scheme 4). Then the dimeriza-
tion product INT2a-1 reacts with the third fluoroacetylene to
give a diradical intermediate INT3a-1, which undergoes ring
closure via either pathway A or pathway B to give Dewar benzene
INT4a-1 or INT4a-2. Finally, isomerizations transform these
’ COMPUTATIONAL METHODS
All calculations were performed with Gaussian 03.14 The hybrid
B3LYP functional15 in conjunction with the 6-31G(d) basis set16 was
applied for the optimization of all the stationary points in the gas phase
except for those singlet diradical transition states and intermediates,
which were located at the UB3LYP/6-31G(d) level. This computational
method has been successfully applied to study reactions involving singlet
diradical intermediates and transition states.17ꢀ20 Frequency calcula-
tions were performed to confirm that each stationary point is either a
minimum or a transition structure. Intrinsic reaction coordinate (IRC)21
calculations were used to confirm the connection among the reactants,
products, and their transition states. For all singlet diradical transition
10866
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Figure 1. Energy surfaces of the favored pathways in the cyclotrimerization of fluoroacetylene. The steps from INT1a-1 to INT2a-1 and from INT2a-1
to INT3a-1 are not simple and involve a series of processes, and these are given in Figure 4.
Dewar benzenes to the corresponding benezene derivatives P-1
and P-2 upon heating.
Z-isomer INT1a-1EZ through a hydrogen switching transition
state, TS1a-1EZ. Then another hydrogen switching through
transition state TS1a-1ZZ gives the Z,Z-isomer INT1a-1ZZ.
These two hydrogen switching processes are easy, with activation
energies of only a few kilocalories per mole. Similar hydrogen
switching is also involved in the course of formation of INT3a-1
from INT3a-1E. Intermediate INT1a-1ZZ is expected to give a
cis diradical, which then undergoes ring closure to give INT2a-1,
but the expected cis diradical is not a stationary point in the
energy surface. Calculations showed that this hypothesized
diradical directly undergoes ring closure to form the cyclobuta-
diene INT2a-1 under geometry optimization. This is similar to
the process of dimerization of nitrile oxide studied by Houk and
co-workers.17g In TS2a-1, the forming CꢀC bond is 3.41 Å long,
and it has a nearly perpendicular configuration with the
C2ꢀC3ꢀC4ꢀC1 dihedral angle being 101.3ꢀ. TS2a-1 has a
diradical character with a computed ÆS2æ of 1.01, indicating that
TS2a-1 is a typical singlet diradical transition state. The ring
closure process is easy with an activation energy of 14.6 kcal/mol
with respect to INT1a-1 and is exothermic by 33.0 kcal/mol.
Once the 1,4-difluorocyclobutadiene is formed through the
dimerization process, the next step in the cyclotrimerization is
the [4þ2] reaction between 1,4-difluorocyclobutadiene and the
third fluoroacetylene. We found that the [4þ2] reaction can
The first step of the cyclotrimerization corresponds to the
formation of the head-to-head diradical INT1a-1 (Figure 1).
This step is easy with an activation energy of 16.1 kcal/mol and
an activation free energy of 24.7 kcal/mol. In the CꢀC bond
forming transition structure TS1a-1, the forming CꢀC bond is
1.89 Å long and has a trans conformation, as evidenced by the
C3ꢀC1ꢀC2ꢀC4 dihedral angle being 180ꢀ (Figure 2). The
computed ÆS2æ of TS1a-1 is 0.02, suggesting that the diradical
character in the transition state has not been fully developed. In
contrast, the formed intermediate INT1a-1 is a typical singlet
diradical with a computed ÆS2æ of 0.93. This is a σ-type diradical,
as judged by its SOMO-1 orbitals. (Its SOMO orbitals are
π-orbitals. The R- and β-radical spin orbitals of INT1a-1 and
INT3a-1 are depicted in Figure 3.) INT1a-1 still adopts a trans
conformation as TS1a-1 does, with the formed CꢀC bond being
1.48 Å long. Formation of this diradical intermediate INT1a-1 is
exothermic by 6.9 kcal/mol, and this process is only slightly
endogonic by 3.6 kcal/mol when considering the entropy penalty
of this bimolecular process.
Calculations indicated that the ring closure from INT1a-1 to
INT2a-1 is not simple and involves a series of processes
(Figure 4). First, the E,E-isomer INT1a-1 isomerizes to the E,
10867
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Figure 2. DFT-computed structures of the reactants, intermediates, transition states, and products.
Figure 3. Spin densities and orbitals of INT1a-1 (SOMO-1) and INT3a-1 (SOMO).
10868
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Figure 4. Hydrogen switching processes in the cyclotrimerization of fluoroacetylene.
occur with both concerted and diradical mechanisms. However,
the concerted processes (see Table S5 of the Supporting
Information) are not favored compared to the stepwise path-
ways. The stepwise [4þ2] pathway starts from CꢀC bond
formation via transition structure TS3a-1, generating a diradical
intermediate, INT3a-1. (Hydrogen switching from another
intermediate IN3a-1E to INT3a-1 is also involved, and this step
is easier. This is shown in Figure 4.) This step requires an
activation energy of only 3.2 kcal/mol and an activation free
energy of 13.0 kcal/mol. The forming C1ꢀC2 bond in the
transition state TS3a-1 is 2.26 Å long. The ÆS2æ in TS3a-1 is 0.76,
indicating that the diradical character in this transition state has
already been formed. The formed C1ꢀC2 bond in INT3a-1 is
1.51 Å long. The ÆS2æ in INT3a-1 is 1.04, indicating that INT3a-1
is a typical singlet diradical intermediate. The SOMO orbitals of
INT3a-1 show that the free radical on the four-membered ring is
a delocalized allylic radical, which is characterized as a π-type
radical. In contrast, the vinyl radical in INT3a-1 is a σ-type radical
composed of an sp2 orbital, and this SOMO orbital is perpendi-
cular to the plane of the double bond (see these orbitals in
Figure 3). Calculations showed that formation of this diradical
intermediate INT3a-1 is exothermic by 25.7 kcal/mol and is
exogonic by 14.4 kcal/mol. Then INT3a-1 undergoes a ring
closure reaction via two competing pathways, pathway A and
pathway B, giving Dewar benzene derivatives INT4a-1 and
INT4a-2, respectively, with barriers of ∼2.5 kcal/mol. This step
is highly exothermic, which suggests that the formation of Dewar
benzenes is irreversible and the regioselectivity of benzene
derivative products is kinetically controlled. The two competing
ring closures described above have similar activation energies,
suggesting the ratio of INT4a-1 to INT4a-2 is 1:1. This agrees
with the experimental results of Viehe, who showed the final
benzene derivatives 5 and 6 were obtained in equal amounts
(Scheme 1).
Finally, Dewar benzene INT4a-1 isomerizes into the aromatic
1,2,3-trifluorobenzene P-1 with an activation energy of 21.5 kcal/mol
and an activation free energy of 21.7 kcal/mol.24 Similarly, Dewar
benzene derivative INT4a-2 isomerizes into aromatic 1,2,4-tri-
fluorobenzene P-2 with an activation energy of 23.6 kcal/mol
and an activation free energy of 23.6 kcal/mol. We can see
that the activation energies of the isomerizations of Dewar
benzenes to aromatic benzenes are higher than those of the
cyclotrimerizations, suggesting that the thermal reaction of
fluoroacetylene gives Dewar benzenes at lower temperatures,
but it can produce benzenes if a higher temperature is used. This
agrees with experimental findings, where Dewar benzenes were
observed experimentally when cooler reaction conditions were
used.6b
From the computational results described above, we can see
that the dimerization of two fluoroacetylenes to form cyclobu-
tadiene is the rate-determining step and this step is irreversible.
The overall activation energy is only 16.1 kcal/mol, and the
activation free energy is 24.7 kcal/mol. The low activation energy
of the rate-determining step explains why the cyclotrimerization
of tert-butylfluoroacetylene can happen spontaneously below
0 ꢀC. The overall reactions to form Dewar benzenes INT4a-1
and INT4a-2 are very exogonic by 97.3 and 94.3 kcal/mol,
respectively. The low activation energy of 16.1 kcal/mol and the
very high exothermicity of the cyclotrimerization of fluoroacety-
lenes are consistent with the fact that these reactions are prone to
explosion under uncontrolled experimental conditions.6a,b,13
Regioselectivity and Substituent Effects of the Dimerization
of Fluoroacetylene. In the favored pathway described above, we
have discussed only the formation of head-to-head INT1a-1 via
transition state TS1a-1. In theory, there are other possible
competing pathways to give other cyclotrimerization products
that were not observed experimentally. Therefore, we computed
these possible pathways and analyzed why these pathways are not
10869
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
The regioselectivity of the dimerization of fluoroacetylenes
can be explained well with frontier molecular orbital (FMO)
theory.25 According to the FMO theory, the smaller the
LUMOꢀHOMO energy gap, the easier the dimerization will
be. In addition, FMO theory indicates that the larger molecular
orbital coefficients in both LUMO and HOMO facilitate the
overlap of molecular orbitals involved in the bond formation
transition state. The calculated 2p molecular orbital coefficients
of the acetylenic carbon attached to fluorine (we label this carbon
as CR) are larger than those of the other acetylenic carbon (we
label this carbon as Cβ) in both LUMO and HOMO (Table 2).26
Therefore, the first forming CꢀC bond between two molecular
fluoroacetylenes prefers to occur on the CR atoms. This is why
the head-to-head pathway is the favored pathway in the dimer-
ization of fluoroacetlyenes. It is interesting to see that substitut-
ing the hydrogen atom with an alkyl or aryl group in the
fluoroacetylene increases the energies of the disfavored transition
structures, making the head-to-head dimerization much more
favored. On the other hand, alkyl and aryl substituents are more
efficient than hydrogen in favoring the head-to-head dimeriza-
tion of fluoroacetylenes (Table 1, entries 2ꢀ4), suggesting that
the forming vinyl radical centers could be stabilized by the alkyl
and aryl groups. We found that the tert-butyl group is more
efficient than the methyl group in lowering the barrier of the
head-to-head dimerization of alkylfluoroacetylenes. Interest-
ingly, the phenyl group is the most efficient group in lowering
the barrier of the head-to-head dimerization of substituted
fluoroacetylenes (Table 1, entry 4). The activation energy of
TS1c-1 is 8.4 kcal/mol, lower than those of TS1b-1 and TS1-1
by 5.9 and 4.5 kcal/mol, respectively. This is consistent with the
FMO theory in that the order of the LUMOꢀHOMO energy
gaps is as follows: fluoroacetylene (17.6 eV) > methylfluoroace-
tylene (16.9 eV) > tert-butylfluoroacetylene (16.8 eV) > phenyl-
fluoroacetylene (11.6 eV).
Figure 5. Energies of the competing transition states and diradical dimers.
Table 1. Activation Energiesa of the Competing Transition
States
Possibility of the Existence of Tetrahedrane 7a in the
Cyclotrimerization. Viehe and co-workers proposed that tetra-
hedrane 7a (Scheme 2) could be formed in the reaction system.
We found that formation of 7a from tert-butylfluoroacetylene 1 is
endothermic by 16.3 kcal/mol, much higher than the activation
energy of dimerization to form 7b (see Figure 6 and entry 3 of
Table 1). This suggests direct formation of 7a from 1 can be ruled
out. Furthermore, this analysis indicates that head-to-head
dimerization of 1 [through TS1-1 (Table 1)] is favored versus
head-to-tail dimerization [through TS1-2 (Table 1)], meaning
that the formation of 7c from 1 through a stepwise [2þ2]
cycloaddition is impossible. Therefore, 7a and 7c cannot be
involved in the cyclotrimeization of tert-butylfluoroacetylene.
Actually, this can be further supported by experiments of Maier
and co-workers,27 who showed that formation of tetrahedrane
can be achieved only by UV irradiation. A similar process of
formation of tetrakis(trimethylsilyl)tetrahedrane could happen
upon irradiation of tetrakis(trimethylsilyl)cyclobutadiene.28 The
four longer CꢀSi bonds and the four σ-donating silyl substitu-
ents were thought to be responsible for the high stability of
tetrakis(trimethylsilyl)tetrahedrane. In the present system, such
effects are absent, and therefore, tetrahedrane formation is not
favored under thermal reaction conditions.
a The energies of the transition states are relative to the energies of the
corresponding separated fluoroacetylenes.
favored energetically. DFT calculation showed that the head-to-
tail transition state TS1a-2 and the tail-to-tail transition state
TS1a-3 are the other two competing transition states in the
dimerization of fluoroacetylenes (Figure 5).
The relative activation energies of TS1a-1, TS1a-2, and TS1a-
3 are 16.1, 18.4, and 19.6 kcal/mol, respectively, indicating that
the pathway for generating a CꢀC bond with fluorine substitu-
tion (head-to-head) is favored. Because both hydrogen and
fluorine are not sterically demanding substituents, we can state
that the significant preference of TS1a-1 over TS1a-2 or TS1a-3
is not due to steric effects, but due to the stereoelectronic effects
of the fluorine substituent.
Experimentally, tert-butylfluoroacetylene was used in the
thermal cyclotrimerization reaction;6a,b it is necessary to inves-
tigate the substituent effects of substituted fluoroacetylenes in
this rate-determining step. To study both the stereoelectronic
and steric effects, we chose methyl, tert-butyl, and phenyl as the
substituents of fluoroacetylenes in the DFT calculations. The
activation energies of the competing transition states in the
dimerization of substituted fluoroacetylenes are listed in Table 1.
Calculations also strongly suggest that head-to-head diradical
dimers are formed exclusively, because the other competing
pathways have activation energies higher than those in the
head-to-head pathways by more than 6 kcal/mol.
Regioselectivity of the Reaction of 1,4-Difluorocyclobuta-
diene with Fluoroacetylene. For the [4þ2] reaction of 1,4-
difluorocyclobutadiene with fluoroacetylene, there are also sev-
eral possible pathways to give various [4þ2] cycloadducts.
Table 3 lists all four possible diradical pathways. Again, formation
10870
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Table 2. Molecular Orbital Coefficients and LUMOꢀHOMO Gaps of Haloacetylenes
a The energy gaps were calculated with optimized geometries in both triplet and singlet states.
molecular orbitals of the CR atoms on the four-membered ring of
1,4-disubstituted difluoroacetylenes are larger than those of the Cβ
atoms in their LUMOs, whereas the molecular orbital coefficients
of these atoms in their HOMOs are very close (Table 4).
Energetically, the energy gap between the HOMO of the
hydro/methylfluoroacetylene and the LUMO of the 1,4-difluor-
obutadiene/1,4-dimethyldifluorobutadiene is smaller than the
energy gap between the LUMO of the hydro/methylfluoroace-
tylene and the HOMO of the 1,4-difluorobutadiene/1,4-di-
methyldifluorobutadiene. Therefore, the diradical reaction of
the hydro/methylfluoroacetylene with the 1,4-difluorobuta-
diene/1,4-dimethyldifluorobutadiene will be mainly controlled
by the orbital interaction between the HOMO of the hydro/
methylfluoroacetylene and the LUMO of the 1,4-difluorobuta-
diene/1,4-dimethyldifluorobutadiene (Figure 7a,b). Therefore,
the larger molecular orbital coefficients of the CR atoms in
HOMO of the hydro/methylfluoroacetylene (Table 2) and
in the LUMO of the 1,4-difluorobutadiene/1,4-dimethyldifluor-
obutadiene (Table 4) facilitate the overlap of molecular orbitals
involved in the bond formation process, leading to the first CꢀC
bond in the reaction of hydro/methylfluoroacetylene and 1,4-
difluorobutadiene/1,4-dimethyldifluorobutadiene forming be-
tween the CR atom of the hydro/methylfluoroacetylene and the
CR atom of 1,4-difluorobutadiene/1,4-dimethyldifluorobutadiene.
In contrast, the reaction between phenylfluoroacetylene and 1,4-
diphenyldifluorobutadiene is dominated by the orbital interaction
between the LUMO of the phenylfluoroacetylene and the HOMO
of the 1,4-diphenyldifluorobutadiene (Figure 7c). Consequently,
the first CꢀC bond in this reaction will also form between the CR
atom of the phenylfluooacetylene and the CR atom of 1,4-
diphenyldifluorobutadiene, because larger molecular orbital coef-
ficients in LUMO and HOMO facilitate the overlap of molecular
orbitals involved in the bond formation transition state.
Figure 6. Energies of dimers of tert-butylfluoroacetylene 1.
of the first CꢀC bond in the reaction of 1,4-difluorocyclobuta-
diene and fluoroacetylene takes place on the CR centers of the
two molecules. Therefore, the head-to-head pathway via TS3a-1
is favored. This trend is observed for various substituted haloa-
cetylenes (Table 3, entries 2ꢀ4).
The fact that the formation of a new CꢀC bond between the
CR centers is lower in activation energy than the competing
pathways can also be rationalized by FMO theory (see the atom
labeling in Table 4).
Calculations show that regardless of whether the substituent is a
hydro, methyl, or phenyl group, the 2p coefficients of the frontier
10871
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
The [4þ2] reaction could also occur via a concerted pathway,
as demonstrated by Houk and Snapper.29 However, this system is
different from Houk and Snapper’s system due to the presence of
fluorine atoms. DFT calculations found that in our case, the
concerted pathways are higher in energy than the diradical
pathways shown in Table 3 (details can be seen in Table S5 of
the Supporting Information). This can also be understood by
invoking the mechanism of the reaction of tetrafluoroethylene
and 1,3-butadiene studied by Borden and co-workers.30 They
found that tetrafluoroethylene and 1,3-butadiene prefer to form a
1,4-diradical intermediate instead of undergoing a concerted
DielsꢀAlder reaction, because fluorine substituents have a
profound stabilizing effect on diradical formation. Formation of
the 1,4-diradical intermediate is also included in the stepwise
[4þ2] reaction of 1,4-difluorocyclobutadiene with fluoroacety-
lene. Therefore, it is not surprising that the concerted pathways
are higher in energy than the diradical pathways for the [4þ2]
reactions shown in Table 3.
noticed that tert-butylchloroacetylene is thermally stable, sug-
gesting that chloroacetylene is not reactive compared with its
fluoro counterpart. Later, Ballester and co-workers reported the
thermal cyclotrimerization of perchlorophenylacetylene at 110ꢀ
120 ꢀC, indicating that arylchloroacetylene is more reactive than
alkylchloroacetylene. Therefore, we computed the dimerization
processes of several chloroacetylenes to investigate their reactiv-
ities and regioselectivities (Table 5). Calculations show that
arylchloroacetylenes (Table 5, entries 1 and 2) strongly favor
the head-to-head dimerizations, whereas alkylchloroacetylenes
(Table 5, entries 3ꢀ5) have a very weak or no preference for
head-to-head dimerization. Again, this regioselectivity can be
explained by FMO theory. For arylchloroacetylenes, the calcu-
lated molecular orbital coefficients of the CR atoms are larger
than those of the Cβ atoms in both their LUMOs and HOMOs,
while this difference is lost on methylchloroacetylenes (Table 2).
That is why the formation of the head-to-head carbonꢀcarbon
bond is the favored pathway in the dimerization of arylchloro-
acetlyenes, while alkylchloroacetylenes have a very weak or no
preference for head-to-head dimerization.
Another conclusion drawn from the comparison of Tables 1
and 5 is that alkylchloroacetylenes are much more sluggish than
their fluoro counterparts in undergoing dimerization by more
than 12.8 kcal/mol in terms of the activation energy. Similar to
the fluoroacetylenes, chloroacetylenes with aryl substituents are
more reactive than alkyl chloroacetylenes during dimerization.
For example, the activation energy for the dimerization of
methylchloroacetylene is 27.1 kcal/mol, while those of phenyl-
chloroacetylene and perchlorophenylacetylene are 21.8 and
22.1 kcal/mol, respectively. It is interesting to note that phenyl-
chloroacetylene and perchlorophenylacetylene have the same
dimerization reactivity. Experimentally, perchlorophenylacety-
lene gave cyclotrimerization products at high temperatures.6c
This implies that phenylchloroacetylene could also undergo
cyclotrimerization under the same or similar reaction conditions.
Our experimental proof of this will be given below.
Cyclotrimerization of Chloroacetylenes. Substituent Effects
and Regioselectivity of the Dimerization of Chloroacetylenes in the
Rate-Determining Step. Experimentally, Viehe and co-workers
Table 3. Activation Energiesa of the Reactions of 1,4-Di-
fluorocyclobutadiene and Fluoroacetylene
To unveil the different reactivities of fluoroacetylene and
chloroacetylene, we initially attempted to explain this difference
with FMO theory. According to the FMO theory, the larger the
a The energies of the transition states are relative to the energies of the
corresponding difluorocyclobutadienes and fluoroacetylenes.
Table 4. Molecular Orbital Coefficients and LUMOꢀHOMO Gaps of 1,4-Difluorocyclobutadienes
10872
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Figure 7. Orbital energies of fluoroacetylenes and 1,4-difluorobutadienes.
Table 5. Activation Energiesa of the Transition States in the
Dimerization of Chloroacetylenes
and the preference of fluorine-substituted carbon centers for
pyramidal geometries.11d,32 This is analogy to the explanation of
the reactivity of tetrafluoroethylene by Borden and co-workers,
showing that tetrafluoroethylene has a weak π-bond and prefers
to generate fluorine-substituted radical centers with pyramidal
geometries.32
DFT Investigation of Cyclotrimerization of Phenylchloroace-
tylene. Figure 8 gives the DFT-computed energy surfaces of the
cyclotrimerization of phenylchloroacetylene 1d to yield 1,2,3-
triphenyltrichlorobenzene P-3 and 1,2,4-triphenyltrichloroben-
zene P-4. The other competitive but disfavored pathways are
given in the Supporting Information (Tables S6 and S8). DFT
calculations show that the cyclotrimerization energy surfaces of
phenylchloroacetylene are very similar to those of fluoroacety-
lene (Figure 1). The formation of INT2d-1 is stepwise and
involves diradical intermediate INT1d-1 with an activation
barrier of 21.8 kcal/mol, which is 13.4 kcal/mol higher than
that for the dimerization of its fluoro counterpart (Table 1).
This suggests that a higher temperature is required to achieve
the cyclotrimerization of phenylchloroacetylene. This ex-
plains why Ballester had to conduct the reaction of perchlor-
ophenylacetylene at 110ꢀ120 ꢀC. The subsequent reaction
between INT2d-1 and 1d to form INT3d-1 with an activation
barrier of 7.4 kcal/mol is also a regioselective process (Table S8
of the Supporting Information), followed by the formation of
Dewar benzenes INT4d-1 and INT4d-2 via two competing path-
ways with very low activation barriers of 0.3 and 0.5 kcal/mol, res-
pectively. Finally, the isomerization of INT4d-1 and INT4d-2 to
benzenes P-3 and P-4 requires activation barriers of 17.4 and 28.7
kcal/mol, respectively. These isomerization barriers are close to the
activation barrier of the dimerization process, suggesting that the
cyclotrimerization of phenylchloroacetylene cannot stop at the
Dewar benzenes but directly gives aromatic benzene derivatives
1,2,3-triphenyltrichlorobenzene P-3 and 1,2,4-triphenyltrichloro-
benzene P-4. This agrees with the results of Ballester’s experi-
ments.6c
a The energies of the transition states are relative to the energies of the
corresponding chloroacetylenes.
LUMOꢀHOMO energy gap, the higher the reaction barrier will
be. As we have presented above, FMO theory can explain the
substituent effects in both fluoro- and chloroacetylene systems
but fails to explain why fluoroacetylene is more reactive than
chloroacetylene. For example, the higher reaction barriers of
chloroacetylenes have smaller LUMOꢀHOMO gaps, while the
lower reaction barriers of fluoroacetylenes have larger LUMOꢀ
HOMO gaps (Table 2). This is in contradiction with FMO
theory. We attributed the relative reactivity in dimerization of
haloacetylenes to the energies required to convert haloacetylenes
to their triplet diradical structures. In the dimerization transition
states, the haloacetylenes adopt trans conformations, very similar
to the interaction between two triplet diradicals. Therefore, the
easier it is for a haloacetylene to form its triplet diradical, the
more reactive this haloacetylene will be in dimerization. Our
calculations found that the tripletꢀsinglet gap of methylfluor-
oacetylene (2.9 eV) and phenylfluoroacetylene (2.4 eV) is
smaller than that of methylchloroacetylene (3.2 eV) and phenyl-
chloroacetylene (2.5 eV) (see Table 2). Therefore, fluoro-
acetylene more easily reaches the diradical transition state than
chloroacetylene, making the former more reactive than the latter.
Such an explanation is very similar to the valence bond theory for
explaining reactivity.31
Experimental Test of the DFT Prediction and Trapping
Experiments. To verify our prediction that unsubstituted aryl-
chloroacetylene can undergo cyclotrimerization and to gain more
information about this process, we embarked on the experi-
mental studies of this reaction. In addition, we managed to trap
the 1,4-disubstituted cyclobutadiene intermediate by two die-
nophiles. When phenylchloroacetylene 1d was subjected to
the thermal reaction conditions at 110 ꢀC without using any
solvent, three products of phenylchloroacetylene were isolated
In addition, we can also attribute the high reactivity of fluoroa-
cetylenes to the weakening of the π-bond of the fluoroacetylenes
10873
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Figure 8. Calculated minimum energy pathway for the cyclotrimerization of phenylchloroacetylene.
(Scheme 5). Two of them, 1,2,3-triphenyltrichlorobenzene P-3
and 1,2,4-triphenyltrichlorobenzene P-4 are trimers with yields
of 7.3 and 14.7%, respectively. The third product is a ladderane,33
cis-1,2,5,6-tetrachloro-3,4,7,8-tetraphenyltricyclo[4.2.0.02,5]octa-
3,7-diene 12, with a reaction yield of 6.6%. The structures of P-3,
P-4, and 12 were confirmed by X-ray crystallographic analyses
(Scheme 5). Trimers P-3 and P-4 are the expected products from
the calculation study described above. Ladderane 12 is a dimer of
1,4-diphenyldichlorocyclobutadiene INT2d-1. From these ex-
perimental results, we propose that the unidentified tetramer in
Viehe’s experiment (Scheme 1) could be an analogue of ladder-
ane 12. The cis conformation of ladderane 12 is the result of the
syn dimerization of INT2d-1. This process is similar to the syn
dimerization of cyclobutadiene, which is thought to be a highly
favored process.34 In contrast, cyclohexylchloroacetylene 1h is
stable at 110 ꢀC for 24 h and can polymerize slowly into
unidentified dark polymers at 125 ꢀC. This suggests that
cyclohexylchloroacetylene is more sluggish than phenylchloro-
acetylene in cyclotrimerization, in good agreement with the
results of our calculation (Table 5).
To further reveal the intermediary existence of antiaro-
matic INT2d-1, we conducted trapping experiments. Heating
phenylchloroacetylene in the presence of maleic anhydride
without solvent gives cycloadduct 13 in 20% yield (Scheme 6).
In the presence of dimethyl acetylenedicarboxylate, cycload-
duct 14 is obtained in 13% yield. The structure of 14 was
confirmed by X-ray crystallographic analysis (Scheme 6).
This experimental evidence clearly confirms the fact that
INT2d-1 is formed as an intermediate in the cyclotrimerization
of phenylchloroacetylene.
’ CONCLUSIONS
We theoretically scrutinized the thermal cyclotrimerizations of
fluoro- and chloroacetylenes with computational and experimental
studies. On the basis of these calculations, 1,4-dihalocyclobutadiene
is regioselectively formed in a stepwise [2þ2] cycloaddition of two
haloacetylenes, which can be trapped regioselectively with another
haloacetylene to generate a 1,4-diradical trimer intermediate. Finally,
1,2,6-trisubstituted Dewar benzene and 1,2,4-trisubstituted Dewar
benzene are equally formed from the ring closure of the 1,4-diradical
trimer intermediate in two competing pathways. The calculations
suggest that the dimerization of the haloacetylene is the rate-
determiningstep withanoverallactivation energyof16.1kcal/mol
for fluoroacetylene and 21.8 kcal/mol for phenylchloroacetylene.
That is why the fluoroacetylene cyclotrimerization can occur at low
temperatures and the arylchloroacetylene cyclotrimerization can
occur at 110 ꢀC. The FMO theory was applied to explain the
regioselective dimerization of haloacetylenes and the regioselective
reaction of 1,4-dihalocyclobutadiene and haloacetylene. It predicted
that formation of a bond on the halogenated carbon centers was
always favored. It was found that the higher reactivity of fluoroace-
tylenes compared to that of chloroacetylenes is due to the fact that
fluoroacetylenes have smaller singletꢀtriplet gaps than chloroace-
tylenes and more easily undergo dimerization and cyclotrimeriza-
tion. In addition to the computational study, experiments were also
conducted to identify the products of the cyclotrimerization
and trap the proposed cyclobutadiene intermediate. The
expected 1,2,3-trisubstituted and 1,2,4-trisubstituted benzenes
were isolated from the thermal reaction of phenylchloroacety-
lene. In addition to these two expected trimers, a tetramer with a
10874
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Scheme 5. Thermal Cyclotrimerization of Phenylchloroacetylene and Cyclohexylchloroacetylene
Scheme 6. Trap INT2d-1 with Dienophiles
tricyclic skeleton was also isolated from the same reaction. In a
trapping experiment, the proposed 1,4-disubstituted cyclobutadiene
was successfully trapped with dienophiles of maleic anhydride and
dimethyl acetylenedicaroxylate. These experimental findings gave
concrete evidence to support the computational predictions.
’ ACKNOWLEDGMENT
We are indebted to the generous financial support from the
Natural Science Foundation of China (20825205, National
Science Fund for Distinguished Young Scholars) and the Na-
tional Basic Research Program of China-973 Programs
(2011CB808601 and 2010CB833203).
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures, spec-
b
’ REFERENCES
tral data of all new compounds, and crystal structure data of P-3,
P-4, 12, and 14 in CIF format. This material is available free of
charge via the Internet at http://pubs.acs.org.
(1) (a) McAlister, D. R.; Bercaw, J. E.; Bergman, R. G. J. Am. Chem.
Soc. 1977, 99, 1666. (b) Vollhardt, K. P. C. Acc. Chem. Res. 1977, 10, 1.
(c) Schore, N. E. Chem. Rev. 1988, 88, 1081. (d) Saito, S.; Yamamoto, Y.
Chem. Rev. 2000, 100, 2901.
(2) (a) Reppe, W.; Schlichting, O.; Klager, K.; Toepel, T. Justus
Liebigs Ann. Chem. 1948, 560, 1. (b) Mao, F.; Schut, D. M.; Tyler, D. R.
Organometallics 1996, 15, 4770. (c) Jhingan, A. K.; Maier, W. F. J. Org.
’ AUTHOR INFORMATION
Corresponding Author
yuzx@pku.edu.cn
10875
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
Chem. 1987, 52, 1161. (d) Gevorgyan, V.; Radhakrishnan, U.; Takeda,
A.; Rubina, M.; Rubin, M.; Yamamoto, Y. J. Org. Chem. 2001, 66, 2835.
(e) Kakeya, M.; Fujihara, T.; Kasaya, T.; Nagasawa, A. Organometallics
2006, 25, 4131. (f) Orian, L.; van Stralen, J. N. P.; Bickelhaupt, F. M.
Organometallics 2007, 26, 3816.
(18) It was demonstrated that the computational results from
(U)B3LYP and CISD, CASSCF, and full CI differ by less than 5 kcal/
mol.17 However, the results from (U)B3LYP calculations are very
reliable for understanding reaction mechanisms and reproducing the
experimentally observed regio- and stereoselectivities.
(3) (a) Sch€aer, W. Angew. Chem., Int. Ed. 1966, 5, 669. (b) Maitlis,
P. M. Acc. Chem. Res. 1976, 9, 93. (c) Jankovꢀa, S.; Draꢁcínskꢀy, M.;
Císaꢁrovꢀa, I.; Kotora, M. Eur. J. Org. Chem. 2008, 47.
(4) (a) Balasubramanian, K. K.; Selvaraj, S.; Venkataramani, P. S.
Synthesis 1980, 29. (b) Matsuda, K.; Nakamura, N.; Iwamura, H. Chem.
Lett. 1994, 1765.
(5) (a) Yang, J.; Verkade, J. G. J. Am. Chem. Soc. 1998, 120, 6834.
(b) Yang, J.; Verkade, J. G. Organometallics 2000, 19, 893.
(19) Calculations using larger basis sets show that the activation
energies of the dimerizations are higher by 4 kcal/mol than those using
the 6-31G(d) basis set, but the conclusions using the 6-31G(d) basis set
are the same as those obtained by using the 6-311þþG(d,p) basis set.
Therefore, we still prefer to use the results from the 6-31G(d) basis set.
The computed energy surfaces for the cyclotrimerizations using the
6-311þþG(d,p) basis set are given in the Supporting Information
(Figures S1 and S2).
ꢁ
(6) (a) Viehe, H. G.; Merꢀenyi, D.-I. R.; Oth, J. F. M.; Valange, P.
Angew. Chem., Int. Ed. 1964, 3, 746. (b) Viehe, H. G.; Merꢀenyi, D.-I. R.;
Oth, J. F. M.; Senders, J. R.; Valange, P. Angew. Chem., Int. Ed. 1964,
3, 755. (c) Ballester, M.; Castaner, J.; Riera, J.; Tabernero, I. J. Org. Chem.
1986, 51, 1413. (d) Hopf, H.; Witulski, B. Pure Appl. Chem. 1993, 65, 47.
(e) Hanamoto, T.; Koga, Y.; Kawanami, T.; Furuno, H.; Inanaga, J.
Angew. Chem., Int. Ed. 2004, 43, 3582. (f) Taniguchi, S.; Yokoi, T.;
Izuoka, A.;Matsushita, M. M.; Sugawara, T.Tetrahedron Lett. 2004, 45, 2671.
(7) Bertholet, M. C. R. Seances Acad. Sci. 1866, 905.
(20) The basis set superposition errors (BSSE) are small using the
6-31G(d) basis set and can be completely neglected using the
6-311þþG(d,p) basis set. See these results in the Supporting Informa-
tion (Table S3).
(21) (a) Fukui, K. J. Phys. Chem. 1970, 74, 4161. (b) Gonzalez, C.;
Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154. (c) Gonzalez, C.; Schlegel,
H. B. J. Phys. Chem. 1990, 94, 5523.
(22) Yamaguchi, K.; Jensen, F.; Dorigo, A.; Houk, K. N. Chem. Phys.
Lett. 1988, 149, 537.
(8) (a) Wagenseller, P. E.; Birney, D. M.; Roy, D. J. Org. Chem. 1995,
60, 2853. (b) Jiao, H.; Schleyer, P. v. R. J. Phys. Org. Chem. 1998, 11, 655.
(c) Morao, I.; Cossío, F. P. J. Org. Chem. 1999, 64, 1868.
(23) French, A. D.; Kelterer, A.-M.; Johnson, G. P.; Dowd, M. K.;
Cramer, C. J. J. Comput. Chem. 2001, 22, 65.
(24) For theoretical studies of similar isomerizations, see:
(a) Johnson, R. P.; Daoust, K. J. J. Am. Chem. Soc. 1996, 118, 7381.
(b) Norton, J. E.; Olson, L. P.; Houk, K. N. J. Am. Chem. Soc. 2006,
128, 7835. (c) Draꢁcínskꢀy, M.; Casta~no, O.; Kotora, M.; Bouꢁr, P. J. Org.
Chem. 2010, 75, 576.
(25) (a) Fukui, K.; Yonezawa, T.; Shingu, H. J. Chem. Phys. 1952,
20, 722.(b) Fleming, I. Molecular Orbitals and Organic Chemical Reac-
tions; John Wiley & Sons, Ltd.: Chichester, U.K., 2010.
(9) Viehe, H. G. Angew. Chem., Int. Ed. 1965, 4, 746.
(10) Hopf, H.; Mlynek, C.; McMahon, R. J.; Menke, J. L.; Lesarri, A.;
Rosemeyer, M.; Grabow, J.-U. Chem.—Eur. J. 2010, 16, 14115.
(11) (a) Helgeson, R. C.; Cram, D. J. J. Am. Chem. Soc. 1966, 88, 509.
(b) Arnett, E. M.; Sanda, J. C.; Bollinger, M. M.; Barber, M. J. Am. Chem.
Soc. 1967, 89, 5389. (c) Doyle, M. P.; West, C. T. J. Am. Chem. Soc. 1975,
97, 3777. (d) Lemal, D. M. J. Org. Chem. 2004, 69, 1.
(12) Very recently, Danheiser and co-workers discovered that the
intramolecular cyclotrimerization of alkynes occurs via a propagylic ene
reaction/[4þ2] reaction mechanism. See: (a) Robinson, J. M.; Sakai, T.;
Okano, K.; Kitawaki, T.; Danheiser, R. L. J. Am. Chem. Soc. 2010,
132, 11039. For the other studies, see: (b) Saaby, S.; Baxendale, I. R.;
Ley, S. V. Org. Biomol. Chem. 2005, 3, 3365. (c) Kral, K.; Hapke, M.
Angew. Chem., Int. Ed. 2011, 50, 2434.
(13) Hopf, H.; Witulski, B. In Modern Acetylene Chemistry; Stang, P. J.,
Diederich, F., Eds.; Wiley-VCH Verlag GmbH: Weinheim, Germany, 1995.
(14) Frisch, M. J.; et al. Gaussian 03, revision B.02; Gaussian, Inc.:
Pittsburgh, PA, 2003. Detailed citation can be found in the Supporting
Information.
(26) This difference is slightly reversed in the HOMO of 1a, perhaps
because of the strong interaction between the π-orbitals of the triple
bond and lone pairs of the fluorine (Table 2).
(27) (a) Maier, G.; Pfriem, S.; Sch€afer, U.; Matusch, R. Angew. Chem.,
Int. Ed. 1978, 17, 520. (b) Maier, G. Angew. Chem., Int. Ed. 1988, 27,
309.
(28) Maier, G.; Neudert, J.; Wolf, O.; Pappusch, D.; Sekiguchi, A.;
Tanaka, M.; Matsuo, T. J. Am. Chem. Soc. 2002, 124, 13819.
(29) (a) Limanto, J.; Tallarico, J. A.; Porter, J. R.; Khuong, K. S.;
Houk, K. N.; Snapper, M. L. J. Am. Chem. Soc. 2002, 124, 14748.
(b) Limanto, J.; Khuong, K. S.; Houk, K. N.; Snapper, M. L. J. Am. Chem.
Soc. 2003, 125, 16310.
(15) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.;
Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(16) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio
Molecular Orbital Theory; Wiley: New York, 1986.
(30) Getty, S. J.; Borden, W. T. J. Am. Chem. Soc. 1991, 113, 4334.
(31) In the traditional VB theory, the triplet energies are computed
on the basis of the ground state closed-shell geometry. However, this
traditional VB model cannot explain the different reactivities of halo-
acetylenes because chloroacetylenes have smaller singletꢀtriplet gaps
that fluoroacetylenes (see Table S9 of the Supporting Information). We
think that the traditional VB theory does not consider the distortion
energy in the transition state. Therefore, we called our explanation a
modified version of the VB theory, where the triplet energies in this
theory are computed on the optimized structures so that distortion
energies in the transition states can be taken into consideration. For
discussion of traditional VB theory, see: (a) Shaik, S.; Schlegel, H. B.;
Wolfe, S. In Theoretical Aspects of Physical Organic Chemistry; John Wiley
& Sons Inc.: New York, 1992. (b) Pross, A. In Theoretical and Physical
Principles of Organic Reactivity; John Wiley & Sons Inc.: New York, 1995.
For discussions of distortion energy in the transition state, see:
(c) Kitaura, K.; Morokuma, K. Int. J. Quantum Chem. 1976, 10, 325.
(d) Nagase, S.; Morokuma, K. J. Am. Chem. Soc. 1978, 100, 1666. (e) Ess,
D. H.; Houk, K. N. J. Am. Chem. Soc. 2007, 129, 10646. (f) de Jong, G. T.;
Bickelhaupt, F. M. ChemPhysChem 2007, 8, 1170. (g) Ess, D. H.; Houk,
K. N. J. Am. Chem. Soc. 2008, 130, 10187. (h) Bento, A. P.; Bickelhaupt,
F. M. J. Org. Chem. 2008, 73, 7290. (i) Xu, L.; Doubleday, C. E.; Houk,
K. N. Angew. Chem., Int. Ed. 2009, 121, 2784. (j) Braida, B.; Walter, C.;
Engels, B.; Hiberty, P. C. J. Am. Chem. Soc. 2010, 132, 7631.
(17) (a) Goldstein, E.; Beno, B.; Houk, K. N. J. Am. Chem. Soc. 1996,
118, 6036. (b) Wittbrodt, J. M.; Schlegel, H. B. J. Chem. Phys. 1996,
105, 6574. (c) Nendel, M.; Sperling, D.; Wiest, O.; Houk, K. N. J. Org.
Chem. 2000, 65, 3259. (d) Cremer, D.; Filatov, M.; Polo, V.; Kraka, E.;
Shaik, S. Int. J. Mol. Sci. 2002, 3, 604. (e) Zhang, D. Y.; Hrovat, D. A.;
Abe, M.; Borden, W. T. J. Am. Chem. Soc. 2003, 125, 12823. (f) Yu, Z.-X.;
Houk, K. N. J. Am. Chem. Soc. 2003, 125, 13825. (g) Yu, Z.-X.;
Caramella, P.; Houk, K. N. J. Am. Chem. Soc. 2003, 125, 15420.
(h) Abe, M.; Adam, W.; Borden, W. T.; Hattori, M.; Hrovat, D. A.;
Nojima, M.; Nozaki, K.; Wirz, J. J. Am. Chem. Soc. 2004, 126, 574.
(i) Zhao, Y.-L.; Suhrada, C. P.; Jung, M. E.; Houk, K. N. J. Am. Chem. Soc.
2006, 128, 11106. (j) Winter, A. H.; Falvey, D. E.; Cramer, C. J.;
Gherman, B. F. J. Am. Chem. Soc. 2007, 129, 10113. (k) Zhou, H.; Wong,
N. B.; Lau, K. C.; Tian, A.; Li, W. K. J. Phys. Chem. A 2007, 111, 9838.
(l) Leach, A. G.; Houk, K. N.; Foote, C. S. J. Org. Chem. 2008, 73, 8511.
(m) Penoni, A.; Palmisano, G.; Zhao, Y.-L.; Houk, K. N.; Volkman, J.;
Nicholas, K. M. J. Am. Chem. Soc. 2009, 131, 653. (n) Siebert, M. R.;
Osbourn, J. M.; Brummond, K. M.; Tantillo, D. J. J. Am. Chem. Soc. 2010,
132, 11952. (o) Lovitt, C. F.; Dong, H.; Hrovat, D. A.; Gleiter, R.;
Borden, W. T. J. Am. Chem. Soc. 2010, 132, 14617.
10876
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877

Journal of the American Chemical Society
ARTICLE
(32) (a) Wang, S. Y.; Borden, W. T. J. Am. Chem. Soc. 1989,
111, 7282. (b) Nicolaides, A.; Borden, W. T. J. Am. Chem. Soc. 1991,
113, 6750.
(33) For syntheses and applications of ladderanes, see: Nouri, D. H.;
Tantillo, D. J. Curr. Org. Chem. 2006, 10, 2055.
(34) Li, Y.; Houk, K. N. J. Am. Chem. Soc. 1996, 118, 880.
10877
dx.doi.org/10.1021/ja2021476 |J. Am. Chem. Soc. 2011, 133, 10864–10877