Gold-catalyzed cyclization of diynes: Controlling the mode of 5-endo versus 6-endo cyclization - An experimental and theoretical study by utilizing diethynylthiophenes

Article abstract of DOI:10.1002/chem.201302967

Herein, a dual-gold catalyzed cyclization of 3,4-diethynylthiophenes generating pentaleno[c]thiophenes through gold-vinylidenes and C-H bond activation is disclosed. Various new heteroaromatic compounds - substrate classes unexplored to date - exhibiting three five-membered annulated ring systems could be synthesized in moderate to high yields. By comparison of the solid-state structures of the corresponding gold-acetylides, it could be demonstrated that the cyclization mode (5-endo versus 6-endo) is controlled by the electronic and not steric nature of the diyne backbone. Depending on different backbones, we calculated thermodynamic stabilities and full potential-energy surfaces giving insight into the crucial dual-activation cyclization step. In the case of the 3,4-thiophene backbone, in which the initial cyclization is rate and selectivity determining, two energetically distinct transition states could be localized explaining the observed 5-endo cyclization mode by classical transition-state theory. In the case of vinyl and 2,3-thiophene backbones, the theoretical analysis of the cyclization mode in the bifurcated cyclization area demonstrated that classical transition-state theory is no longer valid to explain the high experimentally observed selectivity. Herein, for the first time, the influence of the backbone and the aromatic stabilization effect of the 6-endo product in the crucial cyclization step could be visualized and quantified by calculating and comparing the full potential-energy surfaces. Copyright

Full text of DOI:10.1002/chem.201302967

DOI: 10.1002/chem.201302967
Full Paper
&
Homogenous Catalysis
Gold-Catalyzed Cyclization of Diynes: Controlling the Mode of 5-
endo versus 6-endo Cyclization—An Experimental and Theoretical
Study by Utilizing Diethynylthiophenes
[a]
[a]
[a]
[a]
ˇ
Max M. Hansmann,* Svetlana Tsupova, Matthias Rudolph, Frank Rominger, and
A. Stephen K. Hashmi*[a, b]
Abstract: Herein, a dual-gold catalyzed cyclization of 3,4-di-
ethynylthiophenes generating pentaleno[c]thiophenes
step. In the case of the 3,4-thiophene backbone, in which
the initial cyclization is rate and selectivity determining, two
energetically distinct transition states could be localized ex-
plaining the observed 5-endo cyclization mode by classical
transition-state theory. In the case of vinyl and 2,3-thiophene
backbones, the theoretical analysis of the cyclization mode
in the bifurcated cyclization area demonstrated that classical
transition-state theory is no longer valid to explain the high
experimentally observed selectivity. Herein, for the first time,
the influence of the backbone and the aromatic stabilization
effect of the 6-endo product in the crucial cyclization step
could be visualized and quantified by calculating and com-
paring the full potential-energy surfaces.
through gold–vinylidenes and CꢀH bond activation is dis-
closed. Various new heteroaromatic compounds—substrate
classes unexplored to date—exhibiting three five-membered
annulated ring systems could be synthesized in moderate to
high yields. By comparison of the solid-state structures of
the corresponding gold–acetylides, it could be demonstrat-
ed that the cyclization mode (5-endo versus 6-endo) is con-
trolled by the electronic and not steric nature of the diyne
backbone. Depending on different backbones, we calculated
thermodynamic stabilities and full potential-energy surfaces
giving insight into the crucial dual-activation cyclization
Introduction
In the last decade, homogenous gold catalysis underwent
a dramatic development often entitled as the “XXI century
gold rush” offering various and highly inventive tools for or-
ganic synthesis.[7] In 2012, Zhang’s group[8] and our group[9] in-
dependently described a new reactivity pattern of gold cata-
lysts, in which two gold centres synchronously activate a diyne
system (dual activation).[10] Herein, one gold center acts as p
Lewis acidic compound activating one of the alkynes, whereas
the other gold center forms a gold–acetylide, which then
reacts as a nucleophile through the Ca or Cb atom (Scheme 2).
Initially, the 5-endo cyclization pathway via the Cb atom of the
gold–acetylide was experimentally observed generating
a gold–vinylidene species, which undergoes CꢀH activation of
nonactivated aliphatic[8,9b] and aromatic[9a,c] CꢀH bonds, as well
as intermolecular CꢀH bond activation processes (Sche-
me 2a).[9d] Following a theoretical prediction, we were able to
observe a new cyclization mode—the cyclization through the
Ca atom—and could show that the backbone of the diyne
The cyclization of diynes and yne–allenes plays a fundamental
role in organic, medicinal, and computational chemistry.[1] For
thermal and radical processes, the main CꢀC bond-forming
pathways containing a diyne system can be divided into the
1,6-cyclization (Bergman cyclization)[2] or 1,5-cyclization
(Schreiner and Pascal cyclization);[3] in the case of yne–allenes
into the 2,7-cyclization (Myers–Saito cyclization)[4] and 2,6-cycli-
zation (Schmittel cyclization; Scheme 1).[5] Beside these main
radical transformations, there is an increasing interest in the
development of transition-metal-catalyzed cyclizations of these
systems that can afford elsewise inaccessible new products
under mild conditions through nonradical mechanisms.[6]
+
ˇ
[a] M. M. Hansmann, S. Tsupova, Dr. M. Rudolph, Dr. F. Rominger,
Prof. Dr. A. S. K. Hashmi
Organisch-Chemisches Institut, Ruprecht-Karls-Universitꢀt Heidelberg
Im Neuenheimer Feld 270, 69120 Heidelberg (Germany)
Fax: (+49)6221-54-4205
system acts as
a switch in controlling the cyclization
mode.[11,12] With 2,3-substituted thiophene derived diyne sys-
tems, 6-endo cyclization initially leads to a 1,3-diaurated spe-
cies, which undergoes a gold shift by an aryne–gold transition
state to form a 1,4-diaurated species bearing a carbene moiety,
followed by CꢀH insertion (Scheme 2b).
E-mail: hashmi@hashmi.de
hansmann@oci.uni-heidelberg.de
[b] Prof. Dr. A. S. K. Hashmi
Chemistry Department, Faculty of Science
King Abdulaziz University, Jeddah 21589 (Saudi Arabia)
[+] Crystallographic investigation.
Once established, the general question arises why and how
the backbone of the diyne system can control the extraordina-
ry high selectivity towards 5-endo or 6-endo cyclization. To
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201302967.
Chem. Eur. J. 2014, 20, 2215 – 2223
2215
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
Experimental part
We started our experimental in-
vestigation by employing an aro-
matic five-membered thiophene
core as a backbone for the diyne
system. In our previous work
2,3-disubstituted
thiophenes
were studied;[11] however, in this
report, the diyne fragment was
installed in the 3,4-position of
the thiophene ring. We studied
these systems to analyze the in-
fluence of the alkyne–alkyne
angle. From a first point of view,
the angle should behave similar
within the five-membered-ring
backbones compared with the
Scheme 1. Main cyclization modes: comparison between radical-cyclization modes by utilizing diynes (top), yne–
six-membered-ring
backbones
allenes (middle), and nonradical dual-gold-activated cyclizations of diyne systems (bottom).
(for a detailed structure discus-
sion, see below). We were curi-
date, only three different backbones (phenyl,[8,9] thiophene,[11]
and vinyl[12]) were employed in the gold-catalyzed diyne cycli-
zation, and the reason for the different cyclization modes re-
mains unclear. Previous explanations are qualitatively based on
aromaticity in the products[9,11,12] or diyne angles of the starting
materials.[11] Herein, we present a detailed experimental and
theoretical study on the influence of the diyne backbone for
the cyclization mode taking place. This is crucial information
for further exploration of gold-catalyzed dual-activation
chemistry and serves as a guideline for the rational prediction
of new transformations. As outlined in the following, these sys-
tems are highly interesting from a theoretical point of view,
because they exhibit complicated and flat potential-energy
surfaces, which meet the limits of classical transition-state
theory and cross the border between classical and nonstatisti-
cal kinetics.
ous if these systems would undergo a 5-endo cyclization to
generate pentaleno[c]thiophenes (2) or a 6-endo cyclization,
which would lead to indeno[c]thiophenes (3) or an unselective
reaction generating a mixture of both. Interestingly, these sys-
tems cyclize very selectively to give the pentaleno[c]thiophene
system (2) through 5-endo cyclization generating a core of
three five-membered annulated rings and we could not ob-
serve the corresponding indenothiophenes, which were exclu-
sively obtained in the case of a 2,3-thiophene substitution
(Scheme 3).
After a solvent and catalyst screening (see the Supporting
Information), we found optimized reaction conditions employ-
ing 2.5 mol% traceless dual-activation catalysts (TDAC) with
PF6 as counteranion in dichloroethane at 658C.[13] Under these
ꢀ
conditions, pentaleno[c]thiophenes were obtained in moderate
to high yields (Table 1). It is noteworthy to mention that all
Scheme 2. Dual-gold-catalyzed cyclizations of diyne systems through intramolecular CꢀH bond activation leading to fulvene or indane substructures.
Chem. Eur. J. 2014, 20, 2215 – 2223 www.chemeurj.org ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2216

Full Paper
access to the C(sp3)ꢀH methylene insertion, generating the
linear triquinane core (Table 1, entry 1,2). A tert-butyl group
leads in moderate yield to the C(sp3)ꢀH methyl activation
(Table 1, entry 3). Even the activation of a methine C(sp3)ꢀH
bond, such as in a cyclohexane ring (1d), leads to the spirocy-
clic compound 2d in high yield (entry 4), or in the case of 1e,
to the desired CꢀH activation product (entry 5). Employing
a gem-dimethyl group vicinal attached to a nitrogen (Table 1,
entry 6) or oxygen (entry 7) nucleophile did not lead to the de-
sired reaction outcome, instead only elimination to afford the
enyne 2g could be observed. Furthermore, CꢀH insertion into
a CH2Cl fragment could not be detected (Table 1, entry 8). The
reaction of diyne 1i led in good yield to the CꢀH insertion into
the cyclopentane fragment generating an annulated array of
Scheme 3. Gold-catalyzed cyclization of 3,4-diethynylthiophene systems.
products decompose in solution explaining some of the re-
duced yields. Various types of C(sp3)ꢀH activation processes
were feasible, offering a long alkyl chain selectively gave
Table 1. Scope of the gold-catalyzed cyclization of 3,4-thiophene substituted diyne-systems. Typical conditions: diyne (0.25 mmol), dichloroethane (DCE,
0.21m).
Entry
Substrate
Product
t
[h]
Yield[a] Entry
[%]
Substrate
Product
t
Yield[a]
[h] [%]
1
1a
1b
2a
2b
2.5 51[b]
7
8
1g
1h
2g 18 40[d]
[e]
2
3
3
82
60
–
18
3
–
1c
2c 18
9
1i
1j
2i
72
[e]
4
1d
2d
2e
1.5 92
10
–
18
–
[e]
5
6
1e
1 f
2.3 85
11
12
1k
1l
–
18
–
[c]
–
18
–
2l
4
62
[a] Isolated yield. [b] Mixture of CꢀH activation isomers (5.8:1). [c] No reaction. [d] Pentylisonitrile gold chloride (5 mol%) and AgNTf2 (5 mol%) were used
as catalyst. [e] Decomposition of starting diyne.
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2217
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
four five-membered rings (entry 9). The strained com-
pounds such as 1j and 1k decomposed under the
reaction conditions (Table 1, entries 10 and 11),
whereas compound 1l underwent CꢀH insertion
generating the desired compound 2l even in the
presence of a benzylic CꢀH bond (entry 12). The un-
symmetrical substituted 5,5,5-thiopheno[c]pentalene
ring is a very interesting motif, and to date no com-
pounds derived from this motif are known.[14] Espe-
cially in the context of material science, there
has been a broad interest in
Figure 1. ORTEP representation of the solid-state molecular structures of 2m (left) and
2n (right; shown at 50% probability level).
pentalene-motif-containing com-
pounds, such as dibenzopenta-
lenes,[15] whereas the heteroaro-
matic analogues have been
much less explored presumably
caused by their difficult synthetic
access.
Initial attempts to cyclize the
diyne systems followed by a Nꢀ
H or OꢀH insertion employing
compounds, such as 1 f or 1g
failed (Table 1, entry 6,7). There-
fore, we tested the nonsubstitut-
ed alcohols and sulfonamides
and analyzed if aromatization of
Figure 2. Left: comparison of angles and distances in the diyne system depending on the backbone. Right: over-
the products could generate lay of the solid-state structures of the gold–acetylides of 2,3-ethynylthiophene (4) and 3,4-ethynylthiophene (6) fo-
cusing on the diyne fragment (other atoms were omitted for clarity).
a driving force and/or increase
their stability. Interestingly, these
systems undergo cyclization and
aromatization generating the unusual and unknown thio-
phene/furane (2m) or thiophene/pyrrole (2n) systems contain-
ing three annulated five-membered rings (Table 2). We were
able to obtain single crystals suitable for X-ray diffraction anal-
ysis of these compounds and could clearly prove the connec-
tivity and molecular structure of these interesting compounds
(Figure 1).[16]
We were curious if the angle or distances in the starting ma-
terials already reflect a general trend explaining the observed
cyclization modes. Therefore, we synthesized the correspond-
ing gold–acetylides and analyzed the metric parameters based
on the obtained X-ray solid-state molecular structures
(Figure 2; see also the Supporting Information).[16]
All gold–acetylides exhibit a nearly straight CꢀCꢁC-Au-C–
(NHC) fragment. Although the CꢀC bond lengths in the back-
bone follow the general trend 2,3-thiophene (1.383 ꢁ)<phenyl
(1.402 ꢁ)<3,4-thiophene (1.424 ꢁ), consistent with the degree
of p-electron delocalization, the angle (a) embedded between
the two alkyne–alkyne axes are very similar (69.4 vs. 70.28)
within the thiophene systems independent from the substitu-
tion employed (see overlay in Figure 2). Because the two re-
gioisomeric thiophene diyne systems show opposed cycliza-
tion reactivity, it is clear that the diyne angle is not a suitable
criterion to predict the cyclization’s outcome. Equally, the two
CꢀC distances d1 and d2 do not allow predicting a clear trend
explaining the different cyclization modes. In conclusion, an ar-
gumentation based on sterics and angles is not suitable to ex-
plain the gold-catalyzed cyclization of diynes, and the electron-
ics of the diyne backbone must control the cyclization mode.
Table 2. Gold-catalyzed cyclization leading to new polycyclic heteroaro-
matic compounds. Typical conditions: 0.2 mmol of diyne, DCE (0.2m).
Entry
1
Substrate
Product
t
[h]
Yield[a]
[%]
1m
1n
2m
1.5 33
2
2n 38
54[b]
Theoretical part
After analysis of the structural parameters of the starting mate-
rials, we decided to perform a detailed theoretical analysis of
[a] Isolated yields. [b] In this case 5 mol% of catalyst were used.
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2218
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Calculated energy differences of 5-endo and 6-endo cyclization products employing different backbones and different ligand systems (PH3, NHC) at
the B3LYP/cc-pVDZ level of theory.
the dual-gold-catalyzed cyclization event. The first idea to find
a general trend analyzing the natural charges and orbital coef-
ficients of Ca and Cb in the gold–acetylides as a function of the
diyne backbone did not lead to any pronounced differences.
Therefore, we switched to the two-gold-atom-containing inter-
mediates. Depending on the backbone of the diyne system,
a thermodynamical analysis of the stability of the two products
generated in the 6-endo and 5-endo cyclization was performed
(Figure 3). We chose five different backbones (vinyl, 2,3-thio-
phene, phenyl, 3,4-thiophene, and ethyl) and calculated the
energy differences between the cyclization products at the
B3LYP/cc-pVDZ level with 1) PH3 as a ligand on both gold
atoms and 2) an NHC ligand (with N-methyl substitution to
reduce computational cost). Interestingly, independent of the
ligand system, the general thermodynamic stability trend re-
flects the experimentally observed reactivity; for example, in
the case of the 2,3-diethynylthiophene system, the cyclization
towards 6-endo (IV) is favored by DG= +11.6 kcalmolꢀ1 (PH3)
over the 5-endo (III), whereas this trend inverts with the 3,4-di-
ethynylthiophene system (DG=ꢀ0.4 kcalmolꢀ1; VII vs. VIII),
which is consistent with the observed selectivity. When em-
ploying the NHC ligand on both gold centers, which is a more
realistic experimental set-up, all energy differences shift by ap-
proximately 4 kcalmolꢀ1 towards the stabilization of the gold–
vinylidene species (5-endo pathway), presumably caused by
the better donor ability to stabilize the Ca-electron deficiency
of the gold–vinylidene. With this ligand correction, the trend is
even more pronounced and also the phenyl system, which ex-
perimentally undergoes the 5-endo cyclization, fits excellently
into this trend (DG=ꢀ3.8 kcalmolꢀ1). The reason for this pro-
nounced shift in energy differences can be nicely explained by
aromatic stabilization, which is only gained in the 6-endo cycli-
zation mode. Although the generation of the vinylidene spe-
cies is—independent of the backbone—an energetically nonfa-
vored process, the extent of this aromatic stabilization, howev-
er, is strongly dependent on the backbone. Although in the
case of the vinyl fragment the full aromatic stabilization is
gained (II), the 2,3-thiophene and phenyl systems show a re-
duced aromatic stabilization effect, because the formed benzo-
thiophene (IV) and naphthyl-systems (VI) already contain an
annulated aromatic ring. In the case of the 3,4-thiophene back-
bone, there is only very little aromatic stabilization (it is not
possible to delocalize the p electrons of VIII without generat-
ing non favored zwitterionic species), whereas in the ethyl
case (X), no aromatic stabilization is given.
Although this thermodynamical trend is in good agreement
with the observed reactivity, the question of the origin of se-
lectivity remains unexplained. In the specific case of the experi-
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2219
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 4. Comparison of the two cyclization pathways with L=PH3; upper path: cyclization by gold–vinylidene (XIII) leading to indeno[c]fulvenes (XVI);
lower path: 6-endo cyclization leading to indeno[c]thiophenes (XXII); energy values referenced in respect to the starting dual-gold compound XI.
mentally studied 3,4-thiophene backbone, we calculated the
two possible reaction pathways after 5-endo and 6-endo cycli-
zation (Scheme 4). This information is crucial to see if the selec-
tivity is gained in the initial cyclization event, or if a Curtin–
Hammett type pre-equilibration is possible, and the selectivity
is explained through the lowest CꢀH activation pathway.
deno[c]thiophenes (DG=22.7 kcalmolꢀ1 XVI versus XXII). Be-
cause we cannot experimentally detect these cyclization com-
pounds (XXII), it must be concluded that the selectivity and
rate-determining step is clearly given in the initial cyclization
event, and that there is no equilibration of 5-endo (XIII) and 6-
endo (XVII) products.
To understand the selectivity of the initial cyclization event
and to gain a full picture of the backbone influence, we picked
the extreme cases (vinyl, 2,3-thiophene and 3,4-thiophene) and
calculated the potential-energy surfaces as a function of the
CꢀC distances d1 and d2 (Figure 4).
In the case of the 5-endo cyclization, gold–vinylidene XIII un-
dergoes CꢀH insertion in a strong exergonic reaction though
TS-IV (DG= +1.5 kcalmolꢀ1) to generate the five-membered
annulated ring system XIV. Gold migration generates gem-di-
aurated compound XV and a catalyst-transfer step regenerates
the dual-activated substrate XI and closes the catalytic cycle.
Importantly, the slightly less stable 6-endo cyclization product
XVII can undergo a fast 1,2-Au shift via gold–aryne transition
state TS-V, exhibiting a very low energy barrier of DG=
+1.6 kcalmolꢀ1 consistent with earlier reports on this
shift.[11,12] The generated 1,4-diaurated benzene compound
XVIII, which exhibits a strong distortion towards a bent allene,
undergoes CꢀH bond activation via TS-VI. This process re-
quires a small free-activation enthalpy of DG= +5.8 kcalmolꢀ1
and is strongly exergonic. The generated gold–carbene XIX un-
dergoes a 1,2-hydride shift (XX) followed by gold migration to
generate the gem-diaurated species XXI. Importantly, by calcu-
lating this pathway, it could be demonstrated that after 6-endo
cyclization there is no significant activation barrier (all com-
pounds are energetically lower than the 6-endo transition-state
energy) to generate the energetically even more stable in-
Figure 4. Parameterization of the potential-energy surface E(d1,d2).
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2220
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ty at the thiophene ring but
show experimentally opposed
cyclization selectivities, they
seem to be the ideal system to
compare their potential energy
surfaces. All potential-energy
surfaces can be analyzed by uti-
lizing three important transition
states: 1) 5-endo cyclization; 2) 6-
endo cyclization; 3) a ridge tran-
sition state connecting the two
cyclization products. Interesting-
ly, in the case of the 3,4-thio-
phene, potential-energy surface
(Figure 5, top) the shape of the
surface clearly favors a 5-endo
cyclization mode in accordance
with the observed reactivity.
The free-enthalpy difference be-
tween the two 5-endo and
6-endo transition states is
DDG(TS)=1.9 kcalmolꢀ1 towards
the favor of the 5-endo pathway
(for a discussion of the absolute
value, see below). In the case of
the vinyl and 2,3-thiophene
backbone, the large aromatic
stabilization of the 6-endo prod-
uct decreases this barrier for the
6-endo pathway (although there
is no significant change of the 5-
endo potential energy surface),
which in total leads to a more
symmetrical and flat potential-
energy surface around the ridge
(Figure 5 middle and bottom). A
comparison of the position of
the 6-endo transition state (see
dotted line in Figure 5) shows
that d2 shifts to larger values
(more in the direction of the
starting material) with increasing
stability of the 6-endo product,
which is in accordance with an
Figure 5. 2D and 3D representation of the relaxed potential-energy surfaces E(d1,d2) calculated at the B3LYP/cc-
pVDZ level of theory; top: 3,4-thiopene backbone; middle: vinyl backbone; bottom: 2,3-thiophene backbone.
The quantitative potential-energy surfaces obtained at the
B3LYP/cc-pVDZ level of theory with L=PH3 are shown in
Figure 5 as 3D and 2D plots dependent on the diyne back-
bone. In all cases, each point (900 points on a 36ꢂ25 grid with
step sizes of 0.05 ꢁ in each direction) on the surface was calcu-
lated under fully relaxed conditions with the only constrain
being the two CꢀC distances d1 and d2. For all backbones, the
grid has the same dimensions, and the energy is referenced in
each case to the energy of the geometry being closest to the
starting material (d1=2.77 ꢁ; d2=3.13 ꢁ), allowing a direct
comparison of the three potential-energy surfaces.
earlier transition state in agreement with Hammond’s princi-
ple.[17]
In the case of the 2,3-thiophene backbone, the energy differ-
ences for the 5-endo and 6-endo transition states invert, which
is in good agreement with the observed selectivity and
strengthens the herein outlined mechanism. Furthermore, the
precise calculation of the potential-energy surfaces gives in-
sight into an interesting property of the potential-energy sur-
face known as bifurcation. The potential-energy surfaces of the
2,3-thiophene and vinyl backbone exhibit very similar charac-
teristics of a “two-step no intermediate mechanism”.[18,19] Al-
though in the ideal model picture there should be one transi-
tion state leading to the ridge transition state without any in-
Because the 2,3-thiophene and 3,4-thiophene systems are
isomeric structures with the only difference in their connectivi-
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2221
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tervening minimum, we observed a more classical scenario, in
which the first transition state is split into two transition states
(Figure 5 middle). It could be a question, for example, of basis-
set quality/functional choice, how strict these two transition
states can be identified or if the surfaces shift further together
to lead to a single transition state.[8] Because both 5-endo and
6-endo transition states are more or less equal in energy and
are embedded in a very flat potential-energy surface, it is clear
that normal transition-state theory fails to rationalize the high
selectivity for the product formation.[18] Because the gradient
on the ridge connecting the two cyclization products is steep-
er towards the 6-endo product than the 5-endo product, it
seems crucial to consider dynamic effects and to take the
shape of the potential-energy surface into account to explain
the observed selectivity for these systems. Ab initio molecular
dynamics with this fairly complex molecule (N>25 atoms) was
out of reach to quantify the product distribution, but is focus
of further studies.
tion-state energies. As was expected, the NHC ligand favors
the 5-endo cyclization mode and increases the energy differ-
ence at the same level of theory from DDG(TSjPH3)=ꢀ1.9 kcal
molꢀ1 to DDG(TSjNHC(Me))=ꢀ3.3 kcalmolꢀ1 and even up to
DDG(TSjNHC(Ph))=ꢀ4.3 kcalmolꢀ1. This energy difference is
pronounced enough to exclude dynamical effects in the case
of the 3,4-thiophene system and to explain the cyclization
preference based on classical transition-state theory.
Conclusion
We demonstrated that diynes based on a 3,4-substituted thio-
phene backbone cyclize under gold catalysis to generate via
CꢀH bond activation new pentaleno[c]thiophene systems con-
taining an interesting linear heterocyclic triquinane core. The
selectivity for this process is an electronic and not steric effect
given in the initial cyclization event. We analyzed the stability
of the 5-endo and 6-endo cyclization products for a variety of
backbones and calculated for three cases the crucial potential-
energy surfaces. In the case of the experimentally studied
system (3,4-thiophene backbone), the cyclization selectivity
can be explained based on traditional transition-state theory
through two separate transition states, which differ in energy
by up to approximately 4.3 kcalmolꢀ1 (NHC ligand). A compari-
son of the potential-energy surfaces clearly shows that with in-
creasing aromatic stabilization of the 6-endo product, the tran-
sition states for 5-endo and 6-endo cyclization shift closer to-
gether on the surface making an energetic distinction impossi-
ble and outcompeting classical transition-state theory. Even if
the larger gradient towards the 6-endo product would explain
the preferred cyclization mode, it is focus of further studies
how to quantify selectivity on these flat bifurcated surfaces.
It is interesting to highlight that the three different back-
bones represent the three converging scenarios from favor of
one cyclization transition state (up) over bifurcation (middle)
to slight favor of the other cyclization transition state (down).
Although the flat potential-energy surfaces in the case of the
vinyl and 2,3-thiophene backbone demand a further quantifi-
cation based on ab initio molecular dynamics, the general
trend of the backbone influence in this bifurcation area is for
the first time illustrated and fits to the experimentally observed
reactivity.
In the case of the herein experimentally studied system em-
ploying a 3,4-thiophene backbone, we were curious if the
product distribution can be exclusively explained by transition-
state theory and, therefore, calculated the energy differences
between the 6-endo and 5-endo transition states with various
ligands on both gold atoms (Figure 6). Increasing the ligand
size from PH3, PMe3 to PPh3 favors the 5-endo cyclization and
increases the free-energy difference between the two transi-
Acknowledgements
M.M.H. is grateful to the Fonds der Chemischen Industrie for
a Chemiefonds scholarship and to the Studienstiftung des
deutschen Volkes. S.T. is grateful for a Kristjan Jaak scholarship
from the Archimedes Foundation and the Ministry of Educa-
tion and Research of Estonia. The computational studies were
supported by bwGRiD, member of the German D-Grid initia-
tive, funded by the Ministry for Education and Research and
the Ministry for Science, Research and Arts Baden-Wꢃrttem-
berg. The authors would like to thank PD Dr. Markus Pern-
pointner for helpful discussions. The authors thank Umicore
AG & Co. KG for the generous donation of gold salts.
Keywords: alkynes · bifurcation · CꢀH activation · density
functional calculations · gold
[1] a) K. K. Wang, Chem. Rev. 1996, 96, 207–222; b) J. W. Grissom, G. U. Gu-
nawardena, D. Klingberg, D. Huang, Tetrahedron 1996, 52, 6453–6518;
c) A. Basak, S. Mandal, S. S. Bag, Chem. Rev. 2003, 103, 4077–4094;
d) D. S. Rawat, J. M. Zaleski, Synlett 2004, 3, 393–421; e) U. Galm, M. H.
Hager, S. G. V. Lanen, J. Ju, J. S. Thorson, B. Shen, Chem. Rev. 2005, 105,
739–759; f) P. R. Schreiner, A. Navarro-Vazquez, M. Prall, Acc. Chem. Res.
2005, 38, 29–37; g) M. Kar, A. Basak, Chem. Rev. 2007, 107, 2861–2890;
Figure 6. Comparison of transition-state energy differences and geometries
obtained with five different ligand systems at the B3LYP/cc-pVDZ level of
theory. Vibration modes visualized with arrows.
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2222
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
h) R. K. Mohamed, P. W. Peterson, I. V. Alabugin, Chem. Rev. 2013, 113,
10.1021/cr4000682.
[2] a) R. R. Jones, R. G. Bergman, J. Am. Chem. Soc. 1972, 94, 660–661;
b) R. G. Bergman, Acc. Chem. Res. 1973, 6, 25–31.
[3] a) M. Prall, A. Wittkopp, P. R. Schreiner, J. Phys. Chem. A 2001, 105,
9265–9274; b) C. Vavilala, N. Byrne, C. M. Kraml, D. M. Ho, R. A.
Pascal Jr., J. Am. Chem. Soc. 2008, 130, 13549–13551; c) P. R. Schreiner,
M. Prall, V. Lutz, Angew. Chem. 2003, 115, 5935–5938; Angew. Chem. Int.
Ed. 2003, 42, 5757–5760; d) B. Kçnig, W. Pitsch, M. Klein, R. Vasold, M.
Prall, P. R. Schreiner, J. Org. Chem. 2001, 66, 1742–1746.
Wieteck, M. Rudolph, F. Rominger, Angew. Chem. 2012, 124, 4532–
4536; Angew. Chem. Int. Ed. 2012, 51, 4456–4460; c) A. S. K. Hashmi, M.
Wieteck, I. Braun, P. Nçsel, L. Jongbloed, M. Rudolph, F. Rominger, Adv.
Synth. Catal. 2012, 354, 555–562; d) A. S. K. Hashmi, M. Wieteck, I.
Braun, M. Rudolph, F. Rominger, Angew. Chem. 2012, 124, 10785–
10789; Angew. Chem. Int. Ed. 2012, 51, 10633–10637.
[10] For further applications, see: a) P. Nçsel, T. Lauterbach, M. Rudolph, F.
Rominger, A. S. K. Hashmi, Chem. Eur. J. 2013, 19, 8634–8641; b) D. D.
Vachhani, M. Galli, J. Jacobs, L. Van Meervelt, E. V. Van der Eycken, Chem.
Commun. 2013, 49, 7171–7173.
[4] a) A. G. Myers, E. Y. Kuo, N. S. Finney, J. Am. Chem. Soc. 1989, 111, 8057–
8059; b) R. Nagata, H. Yamanaka, E. Okazaki, I. Saito, Tetrahedron Lett.
1989, 30, 4995–4998.
[11] M. M. Hansmann, M. Rudolph, F. Rominger, A. S. K. Hashmi, Angew.
Chem. 2013, 125, 2653–2659; Angew. Chem. Int. Ed. 2013, 52, 2593–
2598.
[5] a) M. Schmittel, M. Strittmatter, S. Kiau, Tetrahedron Lett. 1995, 36,
4975–4978; b) B. Engels, C. Lennartz, M. Hanrath, M. Schmittel, M.
Strittmatter, Angew. Chem. 1998, 110, 2067–2070; Angew. Chem. Int. Ed.
1998, 37, 1960–1963; c) M. Schmittel, C. Vavilala, M. E. Cinar, J. Phys.
Org. Chem. 2012, 25, 182–197.
[12] Very recently, Zhang and co-workers independently reported on a simi-
lar 6-endo cyclization utilizing vinyl backbones, see: Y. Wang, A. Ye-
premyan, S. Ghorai, R. Todd, D. H. Aue, L. Zhang, Angew. Chem. 2013,
125, 7949–7953; Angew. Chem. Int. Ed. 2013, 52, 7795–7799.
[13] A. S. K. Hashmi, T. Lauterbach, P. Nçsel, M. Højer Vilhelmsen, M. Rudolph,
F. Rominger, Chem. Eur. J. 2013, 19, 1058–1065.
[14] We could only find a single publication, in which two symmetrical pen-
taleno[c]-dithiophene products were synthesized in 18 and 22% yield
by using Ni0 catalyst, see: A. Konishi, T. Fujiwara, N. Ogawa, Y. Hirao, K.
Matsumoto, H. Kurata, T. Kubo, C. Kitamura, T. Kawase, Chem. Lett.
2010, 39, 300–301.
[15] a) T. Kawase, T. Fujiwara, C. Kitamura, A. Konishi, Y. Hirao, K. Matsumoto,
H. Kurata, T. Kubo, S. Shinamura, H. Mori, E. Miyazaki, K. Takimiya,
Angew. Chem. 2010, 122, 7894–7898; Angew. Chem. Int. Ed. 2010, 49,
7728–7732; b) M. Saito, Symmetry 2010, 2, 950–969; c) see referen-
ce [9c] and citations therein.
[16] CCDC-948117 (2m), CCDC-948118 (2n), CCDC-948119 (4), and CCDC-
948120 (6) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data request/cif.
Compound 5 was reported in reference [9a].
[6] a) Y. Wang, M. G. Finn, J. Am. Chem. Soc. 1995, 117, 8045–8046; b) K.
Ohe, M. Kojima, K. Yonehara, S. Uemura, Angew. Chem. 1996, 108,
1959–1962; Angew. Chem. Int. Ed. Engl. 1996, 35, 1823–1825; c) T.
Manabe, S. Yanagi, K. Ohe, S. Uemura, Organometallics 1998, 17, 2942–
2944; d) A. Odedra, C.-J. Wu, T. B. Pratap, C.-W. Huang, Y.-F. Ran, R.-S.
Liu, J. Am. Chem. Soc. 2005, 127, 3406–3412; e) B. P. Taduri, Y.-F. Ran, C.-
W. Huang, R.-S. Liu, Org. Lett. 2006, 8, 883–886; for gold-catalyzed cycli-
zation of diynes, which involve a nucleophilic attack and not a dual-ac-
tivation mode, see: f) K. Hirano, Y. Inaba, T. Watanabe, S. Oishi, N. Fujii,
H. Ohno, Adv. Synth. Catal. 2010, 352, 368–372; g) Y. Inaba, S. Oishi, Y.
Takemoto, N. Fujii, H. Ohno, J. Org. Chem. 2011, 76, 9068–9080; h) S.
Naoe, Y. Suzuki, K. Hirano, Y. Inaba, S. Oishi, N. Fujii, H. Ohno, J. Org.
Chem. 2012, 77, 4907–4916; i) G. Ferrara, T. Jin, K. Oniwa, J. Zhao, A. M.
Asiri, Y. Yamamoto, Tetrahedron Lett. 2012, 53, 914–918; j) W. Hou, Z.
Zhang, F. Kong, S. Wang, H. Wang, Z.-J. Yao, Chem. Commun. 2013, 49,
695–697.
[7] For representative reviews on gold catalysis, see: a) A. S. K. Hashmi,
Gold Bull. 2004, 37, 51–65; b) N. Krause, A. Hoffmann-Rçder, Org.
Biomol. Chem. 2005, 3, 387–391; c) A. S. K. Hashmi, Angew. Chem. 2005,
117, 7150–7154; Angew. Chem. Int. Ed. 2005, 44, 6990–6993; d) A. S. K.
Hashmi, G. Hutchings, Angew. Chem. 2006, 118, 8064–8105; Angew.
Chem. Int. Ed. 2006, 45, 7896–7936; e) D. J. Gorin, F. D. Toste, Nature
2007, 446, 395–403; f) A. Fꢃrstner, P. W. Davies, Angew. Chem. 2007,
119, 3478–3519; Angew. Chem. Int. Ed. 2007, 46, 3410–3449; g) E. Jimꢄ-
nez-NuÇez, A. M. Echavarren, Chem. Commun. 2007, 333–346; h) A. S. K.
Hashmi, Chem. Rev. 2007, 107, 3180–3211; i) Z. Li, C. Brouwer, C. He,
Chem. Rev. 2008, 108, 3239–3265; j) A. Arcadi, Chem. Rev. 2008, 108,
3266–3325; k) D. J. Gorin, B. D. Sherry, F. D. Toste, Chem. Rev. 2008, 108,
3351–3378; l) R. Skouta, C.-J. Li, Tetrahedron 2008, 64, 4917–4938;
m) H. C. Shen, Tetrahedron 2008, 64, 3885–3903; n) H. C. Shen, Tetrahe-
dron 2008, 64, 7847–7870; o) A. S. K. Hashmi, M. Rudolph, Chem. Soc.
Rev. 2008, 37, 1766–1775; p) A. S. K. Hashmi, Angew. Chem. 2010, 122,
5360–5369; Angew. Chem. Int. Ed. 2010, 49, 5232–5241; q) C. Nevado,
Chimia 2010, 64, 247–251; r) S. Sengupta, X. Shi, ChemCatChem 2010,
2, 609–619; s) A. Corma, A. Leyva-Pꢄrez, M. J. Sabater, Chem. Rev. 2011,
111, 1657–1712; t) J. Xiao, X. Li, Angew. Chem. 2011, 123, 7364–7375;
Angew. Chem. Int. Ed. 2011, 50, 7226–7236; u) M. Rudolph, A. S. K.
Hashmi, Chem. Soc. Rev. 2012, 41, 2448–2462.
[17] G. S. Hammond, J. Am. Chem. Soc. 1955, 77, 334–338.
[18] For selected examples, see: a) D. A. Singleton, C. Hang, M. J. Szymanski,
M. P. Meyer, A. G. Leach, K. T. Kuwata, J. S. Chen, A. Greer, C. S. Foote,
K. N. Houk, J. Am. Chem. Soc. 2003, 125, 1319–1328; b) S. C. Ammal, H.
Yamataka, M. Aida, M. Dupuis, Science 2003, 299, 1555–1557; c) T.
Bekele, C. F. Christian, M. A. Lipton, D. A. Singleton, J. Am. Chem. Soc.
2005, 127, 9216–9223; d) M. Schmittel, C. Vavilala, R. Jaquet, Angew.
Chem. 2007, 119, 7036–7039; Angew. Chem. Int. Ed. 2007, 46, 6911–
6914; for an overview on bifurcation reactions, see: e) B. K. Carpenter,
Acc. Chem. Res. 1992, 25, 520–528; f) B. K. Carpenter, Chem. Soc. Rev.
2006, 35, 736–747; g) D. H. Ess, S. E. Wheeler, R. G. Iafe, L. Xu, N. C¸elebi-
ꢆlÅꢃm, K. N. Houk, Angew. Chem. 2008, 120, 7704–7713; Angew. Chem.
Int. Ed. 2008, 47, 7592–7601; h) J. Rehbein, B. K. Carpenter, Phys. Chem.
Chem. Phys. 2011, 13, 20906–20922.
[19] Bifurcations in transition-metal-catalyzed reactions are rarely described
to date, but could be observed in gold-catalyzed transformations, see:
a) G. Garayalde, E. Gomez-Bengoa, X. Huang, A. Goeke, C. J. Nevado, J.
Am. Chem. Soc. 2010, 132, 4720–4730; b) Z. J. Wang, D. Benitez, E.
Tkatchouk, W. A. Goddard III, F. D. Toste, J. Am. Chem. Soc. 2010, 132,
13064–13071; c) E. L. Noey, X. Wang, K. N. Houk, J. Org. Chem. 2011, 76,
3477–3483.
[8] L. Ye, Y. Wang, D. H. Aue, L. Zhang, J. Am. Chem. Soc. 2012, 134, 31–34.
[9] a) A. S. K. Hashmi, I. Braun, M. Rudolph, F. Rominger, Organometallics
2012, 31, 644–661; b) A. S. K. Hashmi, I. Braun, P. Nçsel, J. Schꢅdlich, M.
Received: July 27, 2013
Published online on January 16, 2014
Chem. Eur. J. 2014, 20, 2215 – 2223
www.chemeurj.org
2223
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim