Mechanistic Insights into the Post-Cyclization Isomerization in Gold-Catalyzed 7-exo-dig-Hydroarylations

Article abstract of DOI:10.1002/chem.201501075

The subsequent double-bond isomerization in the synthesis of dibenzocycloheptenes and their heteroaromatic analogues was investigated. In the case of biphenyls, a basic additive completely prevented an isomerization to the thermodynamic product. With elec

Full text of DOI:10.1002/chem.201501075

DOI: 10.1002/chem.201501075
Full Paper
&
Synthetic Methods
Mechanistic Insights into the Post-Cyclization Isomerization in
Gold-Catalyzed 7-exo-dig-Hydroarylations
Daniel Pflästerer,[a] Sçren Schumacher,[a] Matthias Rudolph,[a] and A. Stephen K. Hashmi*[a, b]
Abstract: The subsequent double-bond isomerization in the
synthesis of dibenzocycloheptenes and their heteroaromatic
analogues was investigated. In the case of biphenyls, a basic
additive completely prevented an isomerization to the ther-
modynamic product. With electron-rich intramolecular heter-
oaromatic nucleophiles, the isomerization was still observed,
but the kinetic product can be obtained by careful control
of the reaction times in most cases. Mechanistic studies
demonstrated that a slow isomerization is also possible with
the gold catalyst at elevated temperatures, but much faster
isomerization rates were observed with acidic additives. An
observed initiation period for the gold-catalyzed isomeriza-
tion indicates that not the homogenous catalyst, but a de-
composition product of it may be the catalytically active
species.
vinyl–gold intermediates were favored.[5] d-Alkynylfuranes are
standard substrates for the gold-catalyzed phenol synthesis re-
ported by our group,[6] but a 7-exo-dig hydroarylation as side
reaction can also be observed under some conditions. Interest-
ingly, the primarily formed exo-methylene product isomerized
in CDCl3 without addition of any catalyst.[6a] Fitting into the
rather intransparent picture, a gold-catalyzed 6-exo-dig hydro-
arylation of N-aminophenyl propargyl malonates forming dihy-
droquinolines delivered, for some substrates, a mixture of the
initial and the isomerized product, whereas the addition of
5 mol% of p-TsOH led to complete isomerization of the pri-
mary product (Scheme 1).[7]
Introduction
In gold-catalyzed exo-dig cyclizations, the primary product
often undergoes a subsequent isomerization to the product
containing a thermodynamically more stable internal alkene.
Although during the construction of methylenecycloheptanes
by a gold-catalyzed 7-exo-dig cyclization no double-bond iso-
merization could be observed,[1] in the case of indoles as nucle-
ophiles in gold-catalyzed intramolecular hydroarylations,
a double-bond isomerization was observed in some cases.[2]
The AuCl3-catalyzed conversion of N-propargylindole-2-carbox-
amides by 6-exo-dig hydroarylation selectively led to the iso-
merized product.[3] Othman and co-workers obtained a mixture
of both primary and isomerized product by a 6-exo-dig hydro-
arylation. They could convert the mixture into the internal
alkene simply by heating in dichloroethane.[4] An intramolecu-
lar 7-exo-dig hydroamination reaction reported by Sawamura
and co-workers directly delivered the isomerized alkene for
most substrates.[5] In this case, the reaction of the isolated exo-
methylene compound with either a AuI catalyst or the catalyst
in combination with a mild proton source showed mainly de-
composition. These results led to the conclusion that the for-
mation of the exo-methylene compound followed by a subse-
quent isomerization cannot play a major role in the isomeriza-
tion pathway. Instead, pathways starting directly from the
[a] M. Sc. D. Pflästerer, S. Schumacher, Dr. M. Rudolph, Prof. Dr. A. S. K. Hashmi
Organisch-Chemisches Institut
Ruprecht-Karls-Universität Heidelberg
Im Neuenheimer Feld 270, 69120 Heidelberg (Germany)
E-mail: hashmi@hashmi.de
Scheme 1. Accessing two different double-bond isomers from the same sub-
strate.
[b] Prof. Dr. A. S. K. Hashmi
Chemistry Department, Faculty of Science
King Abdulaziz University, Jeddah 21589 (Saudi Arabia)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501075.
Chem. Eur. J. 2015, 21, 11585 – 11589
11585
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
In our recent report on the synthesis of dibenzocyclohepta-
trienes,[8] we found that for many catalyst systems and espe-
cially low-catalyst loadings, a mixture of the primary 3a and
the rearranged product 2a was obtained. Interestingly, 3a was
stable at ambient conditions and the obtained ratio 2a/3a did
not change by dissolution in CDCl3 or during performance of
column chromatography with silica gel. On the other hand, ad-
dition of 5 mol% of HNTf2 as additive successfully gave 2a in
perfect selectivity. This led to the hypothesis that in this case
after the initial gold-catalyzed hydroarylation, a subsequent
acid-catalyzed isomerization should be the major pathway.
Small amounts of strong acid, which are present during the re-
action may be, in fact, responsible for the isomerization in the
first place. To further investigate this hypothesis and, if possi-
ble, to utilize it for a selective synthesis of 3a, we performed
the reaction in the presence of a suitable base to trap traces of
acid.
Figure 1. AuI catalysts used in the optimization.
16 h (Table 1, entry 5). Switching from N-heterocyclic carbene
(NHC) to phosphine ligands (PPh3 and SPhos) gave higher
yields again after 16 h of reaction time. Compound PPh3AuNTf2
produced 3a in 94% yield (Table 1, entry 6), SPhosAuNTf2 gave
an almost identical yield (95%; Table 1, entry 7).
For the optimization of the reaction conditions initially the
original conditions for the cyclization were tested in the pres-
ence of 5 mol% 2,6-di-tert-butylpyridine in dichloromethane
(Table 1, entry 1). Unfortunately, no conversion was observed,
which can be explained by the coordination of the base to the
active site of the catalyst. To circumvent this problem, higher
temperatures were applied in benzene as solvent. To our de-
light, the use of the complexes B and C (Figure 1) under the
elevated temperatures gave a clean conversion to the non-iso-
merized product in reasonable reaction time and high yield
(99% yield after 3 h; Table 1, entries 2 and 3). When IPrAuNTf2
(5 mol%) was used in combination with 5 mol% of base,
a slightly lower yield of 88% was obtained after 17 h (Table 1,
entry 4). With the related Tedii-AuINTf2 (Tedii=(1-cyclopenta-
decyl-3-(2,6-diisopropyl-phenyl)-1,3-dihydro-2H-imidazol-2-yli-
dene)); A) 3a was obtained in slightly lower yield of 84% after
It is noteworthy that after five days of continued stirring at
808C with catalyst B, traces of 2a were observed, whereas in
the case of C, no isomerization at all was detected. Decreasing
the reaction temperature to 408C in C6D6 was not possible,
and only incomplete conversion and 23% yield after 48 hours
were obtained (Table 1, entry 8). Finally, we conducted one re-
action in the absence of base in benzene as solvent. To our
surprise, even in the absence of a base, a high yield of 95% of
the desired product could be obtained. This indicates that the
polarity of the solvent plays an important role for the post cyc-
lization processes in gold catalysis (entry 9). Attempts to per-
form the reaction at room temperature without a base failed.
The scope of the transformation was then investigated
under the optimized conditions. First, compound 1a was con-
verted under preparative conditions, which furnished 3a in an
excellent yield of 95% after 3 hours reaction time (Table 2,
entry 1). N-Heterocycles as nucleophiles were used next. Com-
pound 1b containing a tert-butyloxycarbonyl (Boc)-protected
pyrrole showed a complete conversion after only ten minutes,
providing the desired product 3b in 71% yield (Table 2,
entry 2). For this type of substrate, a significant drop in yield
and selectivity was observed in the absence of base (32% vs
71%), which indicates that for more electron-rich starting ma-
terials the use of the base is crucial. Nevertheless, after six
hours reaction time, the internal alkene 2b can be obtained in
65% yield (Table 2, entry 3) even in the presence of the base.
When indole 1c was used, the reaction needed 30 minutes to
completion. The tetracyclic compound 3c was formed in excel-
lent yield of 98% (Table 2, entry 4). In contrast to 1b, a com-
plete isomerization was not observed under these conditions
even after five days. Furan derivatives were investigated as
well. 3-Substituted furan derivative 1d furnished 3d after only
30 minutes in an excellent yield of 93% (Table 2, entry 5). Simi-
lar to the above-described examples, after a prolonged reac-
tion time, an isomerization to thermodynamic product 2d was
observed, although a base was used in the reaction. After
14 hours at 808C, 2d could be isolated in 82% yield (Table 2,
Table 1. Optimization of the reaction conditions.
Entry[a]
Solvent
Catalyst
T [8C]
t [h]
Yield [%][b]
1
2
3
4
5
6
7
8
9
CD2Cl2
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
B
B
C
RT
80
80
80
80
80
80
40
80
48
3
3
17
16
16
16
72
3
n.c.[c]
99
99
88
84
94
95
23
IPrAuNTf2
A
PPh3AuNTf2
SPhosAuNTf2
C
B[d]
95
[a] Reactions were carried out with 0.05 mmol of substrate. [b] Yields
were determined by 1H NMR spectroscopy with tri-tert-butylbenzene as
internal standard. [c] No conversion. [d] No base was used.
Chem. Eur. J. 2015, 21, 11585 – 11589
www.chemeurj.org
11586
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Scope of the reaction.
Table 2. (Continued)
Entry[a]
Substrate
t
Product
Yield [%]
95
Entry[a]
Substrate
t
Product
Yield [%]
–
1[b]
3 h
19
72 h
n. c.[d]
[a] Reaction were carried out under the optimized conditions at 0.2 mmol
scale in C6H6. [b] 0.5 mmol substrate were used. [c] Yield in brackets for the
reaction without base. [d] No conversion.
2
3
10 min
6 h
71 (32)[c]
65
entry 6). The corresponding 2-furan derivative 1e was also
completely converted after 30 minutes, but 3e was obtained
only in moderate yield of 53% (Table 2, entry 7). Interestingly,
no product of a gold-catalyzed phenol synthesis was ob-
served.[9] In the case of benzofuran 1 f, the tetracyclic com-
pound 3 f again was obtained in an excellent yield of 91%
after 30 minutes (Table 2, entry 8). After a very long reaction
time of 72 hours, again, the isomerized product 2 f was ob-
tained in excellent yield of 90% (Table 2, entry 9). Next, we
turned our attention to thiophene derivatives. Surprisingly, ini-
tial GC-MS analysis after two hours revealed the presence of
starting material 1g, as well as both possible products 2g and
3g. This made the selective synthesis of 3g impossible. After
six hours, 1g was completely converted but still as mixture of
2g and 3g. After 32 hours, 3g was completely transformed to
2g, which was then isolated in 84% yield (Table 2, entry 10).
For compound 1h, the observed results were similar as for 1h,
finally giving 2h after 32 hours in 76% yield (Table 2, entry 11).
Without the addition of base, decomposition was observed for
furane and thiophene derivatives.
4
1 h
98
5
6
7
30 min
14 h
93
82
53
30 min
8
9
30 min
72 h
91
90
10
11
32 h
32 h
84
76
As a next step, we tested differently substituted arene nucle-
ophiles. An electron-donating methoxy group at the non-react-
ing aromatic part of the biaryl substrate 1i produced 3i within
3 h, but only in a moderate yield of 66%. This can be ex-
plained by the formation of the 6-endo-dig hydroarylation
product 4 in 33% yield (Table 2, entry 12). Shifting the elec-
tron-donating substituent to the other part of the biaryl
moiety in 1j delivered 3j in 65% yield after 22 hours in a clean
reaction (Table 2, entry 13). The longer reaction time and lower
yield might be explained by the -I effect of the additional me-
thoxy group that has no positive +M effect onto the reaction
center. Both fluorine containing substrates 1k and l delivered
the desired products 3k (Table 2, entry 14) and 3l (Table 2,
entry 15) after three hours in 81% (3k) and 77% (3l), respec-
tively. The shift of one of the donating groups to a non-conju-
gated position in respect to the reacting carbon (1m) led to
an extensively longer reaction time of 48 h and 3m was ob-
tained in a moderate yield of 41% yield (Table 2, entry 16).
Only traces of 3n could be observed for acetal 1n even after
72 h reaction time (Table 2, entry 17). Substrates 1o and 1p,
containing naphthyl groups as possible nucleophiles, did not
show any conversion even after extended reaction times of
72 h (Table 2, entries 18 and 19).
12
3 h
66+33
13
14
15
22 h
3 h
65
81
77
3 h
16
48 h
41
17
18
72 h
72 h
trace
–
n. c.[c]
The fact that even in the presence of a base still slow iso-
merization takes place for some of the substrates led us to the
Chem. Eur. J. 2015, 21, 11585 – 11589
www.chemeurj.org
11587
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
purpose, trifluoroacetic acid (TFA)-protected aniline
was tested as possible weak base (but still strong
enough to capture the proton of the in situ formed
HNTf2), which does allow reactions at room tempera-
ture.[10] Catalyst
B (0.5 mol%) and the aniline
(5 mol%) were used in CD2Cl2 at room temperature.
After 24 hours, complete conversion was achieved
giving a mixture of 2a and 3a in 99% yield, deter-
mined by NMR spectroscopy, and a selectivity for 3a
of 8:1. In contrast to this result, the same reaction at
room temperature without the additive gave 2a as
major product (Scheme 3).[8]
Conclusion
Scheme 2. Test reactions over the influence of acid during the isomerization step.
The isomerization step in the gold-catalyzed synthe-
sis of dibenzocycloheptenes and their analogues with
one heteroaromatic ring was investigated. Former ex-
conclusion that besides the acid-catalyzed pathway for double-
bond isomerization, another pathway might exist. Thus, differ-
ent test reactions were conducted to shed some light to this
issue (Scheme 2). All reactions were carried out at 0.05 mmol
scale in deuterated solvents, and the yield was determined by
1H NMR spectroscopy by using tri-tert-butylbenzene (TTBB) as
internal standard. First, compound 3a was reacted with
5 mol% HNTf2 in CDCl3 at room temperature. A fast conversion
to 2a was observed after five minutes with 98% NMR yield.
The reaction of 3a with 5 mol% of C in C6D6 showed no con-
version at all even after three days at room temperature. The
same result was observed by heating 3a in C6D6 at 808C for
two days in the absence of any catalyst. But when 5 mol% of
C was used in C6D6 at 808C after 48 h, 100% conversion was
observed in GC-MS analysis. Interestingly, after six hours, no
conversion was detected, whereas already 90% of 3a was con-
verted into 2a after 22 hours. The NMR yield after 48 hours
was calculated to be 95%. This clearly shows that competing
gold-catalyzed pathways for the isomerization step cannot be
completely ruled out, although these only seem to operate at
elevated temperatures. In this context, it should be noted that
whenever isomerized products were obtained, clear signs of
catalyst decomposition were visible on the reaction vial. A pos-
sible reason for the observed induction period in the isomeri-
zation could be the isomerization not being catalyzed by the
homogenous catalyst itself but by a decomposition product of
the catalyst. Thus, the involvement of heterogeneously cata-
lyzed pathways is also possible.
periments in our group led to the conclusion that the isomeri-
zation step is not catalyzed by the homogenous gold catalyst
but by strong acid, which is formed during the course of the
reaction. As a consequence, basic additives could be used to
successfully prevent the isomerization of the primary hydroary-
lation product to the thermodynamic products in the case of
biphenyls. In contrast to acidic conditions, electron-rich hetero-
aromatic systems can also work as nucleophile under these
conditions. More electron-rich five-membered heterocycles as
nucleophiles led to much higher isomerization rates, and
therefore, for some cases even under basic conditions, only
the thermodynamic products could be obtained. Mechanistic
studies indicate that besides a fast proton-catalyzed pathway,
a slower gold-catalyzed isomerization is also conceivable. Most
probably, the active species for this isomerization is a degrada-
tion product of the catalyst as an induction period was ob-
served for the isomerization. The use of weak nucleophilic
bases made it possible to conduct the reaction at room tem-
perature with good selectivity for the kinetic product.
Acknowledgements
The authors thank Umicore AG & Co. KG for the generous don-
ation of gold salts.
Keywords: additives · cycloheptenes · gold · hydroarylation ·
isomerization
As a resume, working at lower temperatures seems to be
necessary to completely suppress the isomerization. For this
[1] H. Ito, H. Ohmiya, M. Sawamura, Org. Lett. 2010, 12, 4380–4383.
[2] a) C. Ferrer, A. M. Echavarren, Angew. Chem. Int. Ed. 2006, 45, 1105–
1109; Angew. Chem. 2006, 118, 1123–1127; b) C. Ferrer, C. H. M. Amijs,
A. M. Echavarren, Chem. Eur. J. 2007, 13, 1358–1373.
[3] D. B. England, A. Padwa, Org. Lett. 2008, 10, 3631–3634.
[4] L. Boiaryna, M. K. El Mkaddem, C. Taillier, V. Dalla, M. Othman, Chem. Eur.
J. 2012, 18, 14192–14200.
[5] H. Ito, T. Harada, H. Ohmiya, M. Sawamura, Beilstein J. Org. Chem. 2011,
7, 951–959.
´
[6] a) A. S. K. Hashmi, M. C. Blanco, E. Kurpejovic, W. Frey, J. W. Bats, Adv.
Synth. Catal. 2006, 348, 709–713; b) A. S. K. Hashmi, M. Wçlfle, F. Ata,
M. Hamzic, R. SalathØ, W. Frey, Adv. Synth. Catal. 2006, 348, 2501–2508;
Scheme 3. Selective reaction providing 3a at RT.
Chem. Eur. J. 2015, 21, 11585 – 11589
www.chemeurj.org
11588
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
c) A. S. K. Hashmi, F. Ata, E. Kurpejovic, J. Huck, M. Rudolph, Top. Catal.
2007, 44, 245–251; d) A. S. K. Hashmi, M. Rudolph, J. W. Bats, W. Frey, F.
Rominger, T. Oeser, Chem. Eur. J. 2008, 14, 6672–6678; e) A. S. K.
Hashmi, F. Ata, P. Haufe, F. Rominger, Tetrahedron 2009, 65, 1919–1927;
f) A. S. K. Hashmi, M. Hamzic, F. Rominger, J. W. Bats, Chem. Eur. J. 2009,
15, 13318–13322; see also g) A. S. K. Hashmi, G. J. Hutchings, Angew.
Chem. 2006, 118, 8064–8105; Angew. Chem. Int. Ed. 2006, 45, 7896–
7936.
[8] D. Pflästerer, E. Rettenmeier, S. Schneider, E. de Las Heras Ruiz, M. Ru-
dolph, A. S. K. Hashmi, Chem. Eur. J. 2014, 20, 6752–6755.
[9] a) A. S. K. Hashmi, T. M. Frost, J. W. Bats, J. Am. Chem. Soc. 2000, 122,
11553–11554; b) A. S. K. Hashmi, T. M. Frost, J. W. Bats, Org. Lett. 2001,
3, 3769–3771.
[10] D. Pflästerer, P. Dolbundalchok, S. Rafique, M. Rudolph, F. Rominger,
A. S. K. Hashmi, Adv. Synth. Catal. 2013, 355, 1383–1393.
[7] C. Gronnier, Y. Odabachian, F. Gagosz, Chem. Commun. 2011, 47, 218–
220.
Received: March 18, 2015
Published online on June 26, 2015
Chem. Eur. J. 2015, 21, 11585 – 11589
www.chemeurj.org
11589
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim