Computation-guided development of Au-catalyzed cycloisomerizations proceeding via 1,2-Si or 1,2-H migrations: Regiodivergent synthesis of silylfurans

Article abstract of DOI:10.1021/ja910290c

A novel highly efficient regiodivergent Au-catalyzed cycloisomerization of allenyl and homopropargylic ketones into synthetically valuable 2- and 3-silylfurans has been designed with the aid of DFT calculations. This cascade transformation features 1,2-Si or 1,2-H migrations in a common Au-carbene intermediate. Both experimental and computational results clearly indicate that the 1,2-Si migration is kinetically favored over the 1,2-shifts of H, alkyl, and aryl groups in the β-Si-substituted Au-carbenes. In addition, experimental results on the Au(I)-catalyzed cycloisomerization of homopropargylic ketones demonstrated that counterion and solvent effects could reverse the above migratory preference. The DFT calculations provided a rationale for this 1,2-migration regiodivergency. Thus, in the case of Ph₃PAuSbF ₆, DFT-simulated reaction proceeds through the initial propargyl-allenyl isomerization followed by the cyclization into the Au-carbene intermediate with the exclusive formation of 1,2-Si migration products and solvent effects cannot affect this regioselectivity. However, in the case of a TfO^- counterion, reaction occurs via the initial 5-endo-dig cyclization to give a cyclic furyl-Au intermediate. In the case of nonpolar solvents, subsequent ipso-protiodeauration of the latter is kinetically more favorable than the generation of the common Au-carbene intermediate and leads to the formation of formal 1,2-H migration products. In contrast, when polar solvent is employed in this DFT-simulated reaction, β-to-Au protonation of the furyl-Au species to give a Au-carbene intermediate competes with the ipso-protiodeauration. Subsequent dissociation of the triflate ligand in this carbene in polar media due to efficient solvation of charged intermediates facilitates formation of the 1,2-Si shift products. The above results of the DFT calculations were validated by the experimental data. The present study demonstrates that DFT calculations could efficiently support experimental results, providing guidance for rational design of new catalytic transformations.

Full text of DOI:10.1021/ja910290c

Published on Web 05/17/2010
Computation-Guided Development of Au-Catalyzed
Cycloisomerizations Proceeding via 1,2-Si or 1,2-H
Migrations: Regiodivergent Synthesis of Silylfurans
Alexander S. Dudnik,† Yuanzhi Xia,‡,§ Yahong Li,*,‡,§ and Vladimir Gevorgyan*,†
Department of Chemistry, UniVersity of Illinois at Chicago, Chicago, Illinois 60607, Key
Laboratory of Organic Synthesis of Jiangsu ProVince, Department of Chemistry and Chemical
Engineering, Suzhou UniVersity, Suzhou 215006, China, and Key Laboratory of Salt Lake
Resources and Chemistry of Chinese Academy of Sciences, Qinghai Institute of Salt Lakes,
Xining 810008, China
Received December 5, 2009; E-mail: liyahong@suda.edu.cn; vlad@uic.edu
Abstract: A novel highly efficient regiodivergent Au-catalyzed cycloisomerization of allenyl and homopro-
pargylic ketones into synthetically valuable 2- and 3-silylfurans has been designed with the aid of DFT
calculations. This cascade transformation features 1,2-Si or 1,2-H migrations in a common Au-carbene
intermediate. Both experimental and computational results clearly indicate that the 1,2-Si migration is
kinetically favored over the 1,2-shifts of H, alkyl, and aryl groups in the ꢀ-Si-substituted Au-carbenes. In
addition, experimental results on the Au(I)-catalyzed cycloisomerization of homopropargylic ketones
demonstrated that counterion and solvent effects could reverse the above migratory preference. The DFT
calculations provided a rationale for this 1,2-migration regiodivergency. Thus, in the case of Ph3PAuSbF6,
DFT-simulated reaction proceeds through the initial propargyl-allenyl isomerization followed by the cyclization
into the Au-carbene intermediate with the exclusive formation of 1,2-Si migration products and solvent
effects cannot affect this regioselectivity. However, in the case of a TfO- counterion, reaction occurs via
the initial 5-endo-dig cyclization to give a cyclic furyl-Au intermediate. In the case of nonpolar solvents,
subsequent ipso-protiodeauration of the latter is kinetically more favorable than the generation of the common
Au-carbene intermediate and leads to the formation of formal 1,2-H migration products. In contrast, when
polar solvent is employed in this DFT-simulated reaction, ꢀ-to-Au protonation of the furyl-Au species to
give a Au-carbene intermediate competes with the ipso-protiodeauration. Subsequent dissociation of the
triflate ligand in this carbene in polar media due to efficient solvation of charged intermediates facilitates
formation of the 1,2-Si shift products. The above results of the DFT calculations were validated by the
experimental data. The present study demonstrates that DFT calculations could efficiently support
experimental results, providing guidance for rational design of new catalytic transformations.
Che3 reported Au-catalyzed4 cycloisomerizations of alkynyl and
allenyl ketones into furans.5 These cascades were proposed to
Introduction
Cycloisomerizations of alkynes and allenes arguably represent
the most versatile methodologies for the assembly of hetero-
cycles.1 Introduction of molecular rearrangements into these
processes often provides highly efficient routes toward densely
substituted heterocycles. Within this area, Hashmi2 and later
proceed via formation of the key putative Au-carbene 1 followed
by a formal 1,2-H shift (Scheme 1).6 Apparently, 1,2-migration
of groups other than H, proceeding via intermediate 1, would
allow for an expedient synthesis of diversely functionalized
furans with substitution patterns not easily accessible via the
† University of Illinois at Chicago.
‡ Suzhou University.
(2) Hashmi, A. S. K.; Schwarz, L.; Choi, J.-H.; Frost, T. M. Angew. Chem.,
Int. Ed. 2000, 39, 2285.
§ Qinghai Institute of Salt Lakes, Chinese Academy of Sciences.
(1) For selected reviews on transition-metal-catalyzed transformations of
alkynes and allenes, see: (a) Alonso, F.; Beletskaya, I. P.; Yus, M.
Chem. ReV. 2004, 104, 3079. (b) Kirsch, S. F. Synthesis 2008, 3183.
(c) Nakamura, I.; Yamamoto, Y. Chem. ReV. 2004, 104, 2127. (d)
Patil, N. T.; Yamamoto, Y. Chem. ReV. 2008, 108, 3395. (e) Jime´nez-
Nu´n˜ez, E.; Echavarren, A. M. Chem. Commun. 2007, 333. (f) Crone,
B.; Kirsch, S. F. Chem.sEur. J. 2008, 14, 3514. (g) Rubin, M.;
Sromek, A. W.; Gevorgyan, V. Synlett 2003, 2265. (h) Ma, S. Chem.
ReV. 2005, 105, 2829. (i) Bates, R. W.; Satcharoen, V. Chem. Soc.
ReV. 2002, 31, 12. (j) Zimmer, R.; Dinesh, C. U.; Nandanan, E.; Khan,
F. A. Chem. ReV. 2000, 100, 3067. (k) Abu Sohel, S. M.; Liu, R.-S.
Chem. Soc. ReV. 2009, 38, 2269. (l) Majumdar, K. C.; Debnath, P.;
Roy, B. Heterocycles 2009, 78, 2661.
(3) Zhou, C. Y.; Chan, P. W. H.; Che, C. M. Org. Lett. 2006, 8, 325.
(4) For selected recent reviews on Au-chemistry, see: (a) Arcadi, A. Chem.
ReV. 2008, 108, 3266. (b) Hashmi, A. S. K. Chem. ReV. 2007, 107,
3180. (c) Gorin, D. J.; Toste, F. D. Nature 2007, 446, 395. (d) Li, Z.;
Brouwer, C.; He, C. Chem. ReV. 2008, 108, 3239. (e) Fu¨rstner, A.;
Davies, P. W. Angew. Chem., Int. Ed. 2007, 46, 3410. (f) Jime´nez-
Nu´n˜ez, E.; Echavarren, A. M. Chem. ReV. 2008, 109, 3326. (g)
Widenhoefer, R. A.; Han, X. Eur. J. Org. Chem. 2006, 4555. (h) Shen,
H. C. Tetrahedron 2008, 64, 3885. (i) Shen, H. C. Tetrahedron 2008,
64, 7847. (j) Hashmi, A. S. K.; Hutchings, G. J. Angew. Chem., Int.
Ed. 2006, 45, 7896. (k) Belmont, P.; Parker, E. Eur. J. Org. Chem.
2009, 6075. (l) Hoffmann-Ro¨der, A.; Krause, N. Org. Biomol. Chem.
2005, 3, 387.
9
10.1021/ja910290c  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 7645–7655 7645

A R T I C L E S
Dudnik et al.
Scheme 1. 1,2-Group Migrations in the Synthesis of Furans via
Au-Carbene Intermediate 1
Scheme 3. Predicted Activation Energies for the 1,2-Si- versus
1,2-R Migration in the Au-Carbene Intermediate 1′
Results and Discussion
Recent DFT study revealed that the Au-catalyzed cycloi-
somerization of bromoallenyl ketones7j proceeds through the
Au-carbene 1, wherein the 1,2-Br migration is kinetically
favored over a facile 1,2-H shift.8 Naturally, we were interested
in identifying other migrating groups capable of undergoing a
selective 1,2-migration over the 1,2-H shift. Thus, we turned
our attention to a silyl group, which is known to undergo a
facile 1,2-migration to electron-deficient centers. Despite the
well-known 1,2-Si migration to cationic centers,9,10 little atten-
tion was devoted to 1,2-Si shifts to free11 or metal-stabilized
carbenes,12 especially with a focus on synthetic significance.
Also, few studies on migratory aptitudes in these systems have
been reported.11,12
We envisioned that DFT calculations could shed light on the
relative rates of 1,2-migrations of Si, H, and alkyl groups in
the Au-carbene 1′ (Scheme 3).13 Thus, the results of DFT
computations14,15 indicated that the 1,2-Si migration is strongly
favored over the 1,2-shifts of methyl, phenyl, and even H in 1′
for both AuCl3 and H3PAu+ (Scheme 3).14
Scheme 2. Regiodivergent 1,2-Si Migrations in the Synthesis of
Furans via Au-Carbene Intermediate
existing methodologies.7 Along this line, we recently reported
the Au-catalyzed regiodivergent synthesis of halofurans involv-
ing 1,2-Hal or 1,2-H migration in 1.7j Herein, we wish to report
a computation-designed regiodivergent Au-catalyzed cycloi-
somerization of homopropargyl and allenyl ketones into silyl-
furans, featuring 1,2-Si or 1,2-H migration in the Au-carbene
intermediate 1 (Scheme 2).
(5) For recent reviews on furan synthesis, see: (a) Brown, R. C. D. Angew.
Chem., Int. Ed. 2005, 44, 850. (b) Kirsch, S. F. Org. Biomol. Chem.
2006, 4, 2076. (c) Patil, N. T.; Yamamoto, Y. ARKIVOC 2007, 121.
(d) Hou, X. L.; Cheung, H. Y.; Hon, T. Y.; Kwan, P. L.; Lo, T. H.;
Tong, S. Y.; Wong, H. N. C. Tetrahedron 1998, 54, 1955. (e) Hou,
X.-L.; Yang, Z.; Yeung, K.-S.; Wong, H. N. C. Prog. Heterocycl.
Chem. 2008, 19, 176. (f) Balme, G.; Bouyssi, D.; Monteiro, N.
Heterocycles 2007, 73, 87. (g) D’Souza, D. M.; Muller, T. J. J. Chem.
Soc. ReV. 2007, 36, 1095. For selected recent examples on transition-
metal-catalyzed synthesis of furans, see: (h) Zhang, M.; Jiang, H.-F.;
Neumann, H.; Beller, M.; Dixneuf, P. H. Angew. Chem., Int. Ed. 2009,
48, 1681. (i) Xiao, Y.; Zhang, J. AdV. Synth. Catal. 2009, 351, 617.
(j) Pan, Y.-M.; Zhao, S.-Y.; Ji, W.-H.; Zhan, Z.-P. J. Comb. Chem.
2009, 11, 103. (k) Zhang, G.; Huang, X.; Li, G.; Zhang, L. J. Am.
Chem. Soc. 2008, 130, 1814. (l) Xiao, Y.; Zhang, J. Angew. Chem.,
Int. Ed. 2008, 47, 1903. (m) Shibata, Y.; Noguchi, K.; Hirano, M.;
Tanaka, K. Org. Lett. 2008, 10, 2825. (n) Ji, K.-G.; Shen, Y.-W.; Shu,
X.-Z.; Xiao, H.-Q.; Bian, Y.-J.; Liang, Y.-M. AdV. Synth. Catal. 2008,
350, 1275. (o) Barluenga, J.; Riesgo, L.; Vicente, R.; Lo´pez, L. A.;
Toma´s, M. J. Am. Chem. Soc. 2008, 130, 13528. (p) Aurrecoechea,
J. M.; Durana, A.; Pe´rez, E. J. Org. Chem. 2008, 73, 3650. (q) Zhao,
L.-B.; Guan, Z.-H.; Han, Y.; Xie, Y.-X.; He, S.; Liang, Y.-M. J. Org.
Chem. 2007, 72, 10276. (r) Shu, X.-Z.; Liu, X.-Y.; Xiao, H.-Q.; Ji,
K.-G.; Guo, L.-N.; Qi, C.-Z.; Liang, Y.-M. AdV. Synth. Catal. 2007,
349, 2493. (s) Peng, L.; Zhang, X.; Ma, M.; Wang, J. Angew. Chem.,
Int. Ed. 2007, 46, 1905. (t) Arimitsu, S.; Hammond, G. B. J. Org.
Chem. 2007, 72, 8559. (u) Zhang, J.; Schmalz, H.-G. Angew. Chem.,
Int. Ed. 2006, 45, 6704. (v) Aponick, A.; Li, C.-Y.; Malinge, J.;
Marques, E. F. Org. Lett. 2009, 11, 4624. (w) Blanc, A.; Tenbrink,
K.; Weibel, J.-M.; Pale, P. J. Org. Chem. 2009, 74, 5342. (x) Blanc,
A.; Tenbrink, K.; Weibel, J.-M.; Pale, P. J. Org. Chem. 2009, 74,
4360. (y) Yoshida, M.; Al-Amin, M.; Shishido, K. Synthesis 2009,
2454. (z) Egi, M.; Azechi, K.; Akai, S. Org 2009, 11, 5002. (aa) Liu,
F.; Yu, Y.; Zhang, J. Angew Chem., Int. Ed. 2009, 48, 5505. (ab)
Suhre, M. H.; Reif, M.; Kirsch, S. F. Org. Lett. 2005, 7, 3925. (ac)
Patil, N. T.; Wu, H.; Yamamoto, Y. J. Org. Chem. 2005, 70, 4531.
(ad) Liu, Y.; Song, F.; Song, Z.; Liu, M.; Yan, B. Org. Lett. 2005, 7,
5409. (ae) Yao, T.; Zhang, X.; Larock, R. C. J. Am. Chem. Soc. 2004,
126, 11164. (af) Hashmi, A. S. K.; Sinha, P. AdV Synth. Catal. 2004,
346, 432. (ag) Kel’in, A. V.; Gevorgyan, V. J. Org. Chem. 2002, 67,
95.
Synthesis of Furans via Au(III)-Catalyzed 1,2-Si
Migration in Allenes. It occurred to us that the Au-catalyzed
cyclization of silyl allenes 4 could provide a straightforward
access to the Au-carbene 1′ via a well-established route (eq 1).2,6
(6) For Ag-catalyzed synthesis of furans via a formal 1,2-H shift in
allenones, see:(a) Marshall, J. A.; Robinson, E. D. J. Org. Chem. 1990,
55, 3450. (b) Hashmi, A. S. K. Angew. Chem., Int. Ed. Engl. 1995,
34, 1581.
(7) For selected examples on 1,2-shifts in the synthesis of aromatic
heterocycles, see: (a) Kim, J. T.; Kel’in, A. V.; Gevorgyan, V. Angew.
Chem., Int. Ed. 2003, 42, 98. (b) Sromek, A. W.; Kel’in, A. V.;
Gevorgyan, V. Angew. Chem., Int. Ed. 2004, 43, 2280. (c) Schwier,
T.; Sromek, A. W.; Yap, D. M. L.; Chernyak, D.; Gevorgyan, V. J. Am.
Chem. Soc. 2007, 129, 9868. (d) Dudnik, A. S.; Gevorgyan, V. Angew.
Chem., Int. Ed. 2007, 46, 5195. (e) Dudnik, A. S.; Sromek, A. W.;
Rubina, M.; Kim, J. T.; Kel’in, A. V.; Gevorgyan, V. J. Am. Chem.
Soc. 2008, 130, 1440. (f) Gorin, D. J.; Davis, N. R.; Toste, F. D. J. Am.
Chem. Soc. 2005, 127, 11260. (g) Li, G.; Huang, X.; Zhang, L. Angew.
Chem., Int. Ed. 2008, 47, 346. (h) Takaya, J.; Udagawa, S.; Kusama,
H.; Iwasawa, N. Angew. Chem., Int. Ed. 2008, 47, 4906. (i) Davies,
P. W.; Martin, N. Org. Lett. 2009, 11, 2293. (j) Sromek, A. W.; Rubina,
M.; Gevorgyan, V. J. Am. Chem. Soc. 2005, 127, 10500.
(8) Xia, Y.; Dudnik, A. S.; Gevorgyan, V.; Li, Y. J. Am. Chem. Soc.
2008, 130, 6940.
(9) For selected examples, see: (a) Danheiser, R. L.; Carini, D. J.; Basak,
A. J. Am. Chem. Soc. 1981, 103, 1604. (b) Danheiser, R. L.;
Kwasigroch, C. A.; Tsai, Y.-M. J. Am. Chem. Soc. 1985, 107, 7233.
(c) Becker, D. A.; Danheiser, R. L. J. Am. Chem. Soc. 1989, 111,
389. (d) Panek, J. S.; Yang, M. J. Am. Chem. Soc. 1991, 113, 9868.
(10) For selected examples of Si-shifts in the synthesis of aromatic
heterocycles, see: (a) Nakamura, I.; Sato, T.; Terada, M.; Yamamoto,
Y. Org. Lett. 2007, 9, 4081. (b) Seregin, I. V.; Gevorgyan, V. J. Am.
Chem. Soc. 2006, 128, 12050. (c) Danheiser, R. L.; Stoner, E. J.;
Koyama, H.; Yamashita, D. S.; Klade, C. A. J. Am. Chem. Soc. 1989,
111, 4407.
(11) (a) Creary, X.; Butchko, M. A. J. Org. Chem. 2001, 66, 1115. (b)
Creary, X.; Wang, Y.-X. Tetrahedron Lett. 1989, 30, 2493.
9
7646 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

DFT-Guided Regiodivergent Synthesis of Silylfurans
A R T I C L E S
Table 1. Au-Catalyzed Cycloisomerization of Allenyl Ketones 4
a Isolated yield of product for reactions performed on 0.5 mmol scale. b Isolated yield of 2e for reaction performed in MeNO2. c Isolated yield of 2e
for reaction performed with 1 mol % of catalyst.
The latter intermediate, upon the 1,2-Si shift, should furnish
the 3-silylfuran 2.
a highly selective 1,2-Si migration, affording 3-silylfurans 2a
and 2b in high yields (Table 1, entries 1-2). Likewise,
cycloisomerization of phenyl-substituted allenes 4c,d proceeded
very efficiently to give 3-silylfurans 2c,d as sole isomers (entries
3-4). To our delight, R-to-Si-unsubstituted allene 4e cyclized
Scheme 4. Potential Energy Surfaces for the AuCl3-Catalyzed
Cycloisomerization of Allenes 4 (R ) H)
To this end, we have tested the possible cycloisomerization
of allenes 4 into 3-silylfurans 2 in the presence of various Au
catalysts. Gratifyingly, it was found that R-to-Si alkyl-substituted
allenes 4a and 4b in the presence of AuCl3 catalyst underwent
(12) Creary, X.; Butchko, M. A. J. Org. Chem. 2002, 67, 112.
(13) For selected examples of DFT studies in Au-chemistry, see: (a)
Nevado, C.; Echavarren, A. M. Chem.sEur. J. 2005, 11, 3155. (b)
Comas-Vives, A.; Gonza´lez-Arellano, C.; Corma, A.; Iglisias, M.;
Sa´nchez, F.; Ujaque, G. J. Am. Chem. Soc. 2006, 128, 4756. (c) Nieto-
Oberhuber, C.; Lo´pez, S.; Mun˜oz, M. P.; Ca´rdenas, D. J.; Bun˜uel, E.;
Nevado, C.; Echavarren, A. M. Angew. Chem., Int. Ed. 2005, 44, 6146.
´
(d) Faza, O. N.; Lo´pez, C. S.; Alvarez, R.; de Lera, A. R. J. Am.
Chem. Soc. 2006, 128, 2434. (e) Straub, B. F. Chem. Commun. 2004,
1726. (f) Correa, A.; Marion, N.; Fensterbank, L.; Malacria, M.; Nolan,
S. P.; Cavallo, L. Angew. Chem., Int. Ed. 2008, 47, 718. (g) Nieto-
Oberhuber, C.; Mun˜oz, M. P.; Bun˜uel, E.; Nevado, C.; Ca´rdenas, D. J.;
Echavarren, A. M. Angew. Chem., Int. Ed. 2004, 43, 2402. (h) Shi,
F.-Q.; Li, X.; Xia, Y.; Zhang, L.; Yu, Z.-X. J. Am. Chem. Soc. 2007,
129, 15503. (i) Lemie`re, G.; Gandon, V.; Cariou, K.; Hours, A.;
Fukuyama, T.; Dhimane, A.-L.; Fensterbank, L.; Malacria, M. J. Am.
Chem. Soc. 2009, 131, 2993. (j) Nieto-Oberhuber, C.; Pe´rez-Gala´n,
P.; Herrero-Go´mez, E.; Lauterbach, T.; Rodr´ıguez, C.; Lo´pez, S.; Bour,
C.; Rosello´n, A.; Ca´rdenas, D. J.; Echavarren, A. M. J. Am. Chem.
Soc. 2008, 130, 269.
Scheme 5. Generation of Products 2 and 3 from Complexes
2-AuCl3 and 3-AuCl3
(14) B3LYP/6-31G*(LANL2DZ for Au) method was used for all the
calculations, and solvation effect was calculated by CPCM model.
(15) See Supporting Information for details.
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7647

A R T I C L E S
Dudnik et al.
Figure 1. Geometries for the intermediates and transition states given in Scheme 4; selected distances are in angstroms.
in a highly selective manner into a product of the exclusive
1,2-Si over 1,2-H shift 2e in both toluene or more polar
nitromethane solvent (entry 5). Thus, the experimentally
observed exclusive Si-migration is in excellent agreement with
the computation-predicted higher 1,2-migratory aptitude of Si
over H, Alk, and Ar groups. It is worth mentioning that allene
4e was not stable in the presence of Ph3PAuOTf complex when
performing the reaction in toluene or 1,2-dichloroethane (DCE)
solvent. In addition, similarly to AuCl3, employment of
Ph3PAuSbF6 catalyst provided exclusive formation of 1,2-Si
migration product 2e in DCE solvent, albeit notable amounts
of the desilylated 3-methyl-2-phenylfuran were observed.
stationary points along the reaction pathway for the cycloi-
somerization of allenes 4 are presented in Figure 1.
Counterion Effects on the Au(I)-Catalyzed 1,2-Si
Migration in Alkynes. Next, we envisioned that easily accessible
homopropargylic ketones 6 could also serve as precursors7e,j
for the allene 5 toward the common Au-carbene intermediate 1
(eq 2). Alternatively, cyclization of 6 into furyl-Au species 7,2,16
followed by its ꢀ-protonation at the vinyl-gold moiety,17 would
generate the Au-carbene 1. Based on the computation results,
it was predicted that a subsequent facile 1,2-Si shift in 1 would
furnish 3-silylfuran 2. On the other hand, a counterion-assisted8,18
1,2-H shift in 1 or direct R-protonation of the vinyl-Au moiety
The computed energy surface for the Au-catalyzed cycloi-
somerization of allene 4 (AuCl3) is provided in Scheme 4.14
Hence, a highly exergonic (24.7 kcal/mol) cyclization of
complex 4-AuCl3 into the carbene 1-AuCl3 required an activa-
tion energy of 2.6 kcal/mol only. A subsequent kinetically more
favorable 1,2-Si migration (TS1_2-AuCl3, activation energy of
11.1 kcal/mol) over the 1,2-H shift (TS1_3-AuCl3, activation
energy of 23.6 kcal/mol) led to the product complex 2-AuCl3.
A requisite ligand exchange reaction proceeded as illustrated
in Scheme 5 to complete the catalytic cycle, thus furnishing
the final products and regenerating the active allene complex
4-AuCl3. The free energy values indicated that the formation
of 2 and 3 via the reactions of allene 4 with 2-AuCl3 and
3-AuCl3 were both endergonic by 17.8 and 12.2 kcal/mol,
respectively (Scheme 5). The optimized structures of key
(16) (a) Sheng, H.; Lin, S.; Huang, Y. Synthesis 1987, 1022. (b) Fukuda,
Y.; Shiragami, H.; Utimoto, K.; Nozaki, H. J. Org. Chem. 1991, 56,
5816. (c) Shapiro, N. D.; Toste, F. D. J. Am. Chem. Soc. 2007, 129,
4160. (d) Wang, W.; Xu, B.; Hammond, G. B. J. Org. Chem. 2009,
74, 1640.
(17) For examples of the proposed ꢀ-electrophilic attack on vinyl-gold,
see: (a) Huang, X.; Zhang, L. J. Am. Chem. Soc. 2007, 129, 6398. (b)
Luzung, M. R.; Mauleo´n, P.; Toste, F. D. J. Am. Chem. Soc. 2007,
129, 12402.
(18) For review on counterion effects in Au-catalysis, see: (a) Gorin, D. J.;
Sherry, B. D.; Toste, F. D. Chem. ReV. 2008, 108, 3351. For selected
examples, see: (b) Lemie`re, G.; Gandon, V.; Agenet, N.; Goddard,
J.-P.; de Kozak, A.; Aubert, C.; Fensterbank, L.; Malacria, M. Angew.
Chem., Int. Ed. 2006, 45, 7596. (c) Bhunia, S.; Liu, R.-S. J. Am. Chem.
Soc. 2008, 130, 16488. (d) Gorin, D. G.; Watson, I. D. G.; Toste,
F. D. J. Am. Chem. Soc. 2008, 130, 3736. (e) Kova´cs, G.; Ujaque, G.;
Lledo´s, A. J. Am. Chem. Soc. 2008, 130, 853.
9
7648 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

DFT-Guided Regiodivergent Synthesis of Silylfurans
A R T I C L E S
in 7 might compete with the 1,2-Si migration, affording the
2-silylfurans 3 (eq 2).
proceeded with the simultaneous abstraction of a propargylic
H-atom by the TfO- counterion to give the furyl-Au T-IN1
and HOTf, requiring 10.9 kcal/mol of activation energy.
Subsequent endergonic association of HOTf (15.1 kcal/mol)
with the R-C-atom of the vinyl-Au moiety in T-IN1 followed
by ipso-protiodeauration of the T-IN2 intermediate (T-TS4,
activation energy of 4.9 kcal/mol) provided the product of
the formal 1,2-H migration, T-3. On the other hand, the
kinetically more favorable ꢀ-protonation at the vinyl-gold
moiety in T-IN2 (T-TS2, activation energy of 4.7 kcal/mol)
over the ipso-protiodeauration (T-TS4, activation energy of
4.9 kcal/mol) should lead to a competing generation of the
Au carbene T-1. Subsequent 1,2-Si migration from T-1
required 10.7 kcal/mol of free energy (T-TS3) to give 1,2-
Si migration product T-2. Alternatively, it was found that a
TfO- ligand dissociation reaction could occur efficiently from
carbene T-1 (6.0 kcal/mol) to give the more electrophilic
carbene C-1, wherein a more kinetically facile 1,2-Si shift
(C-TS1, activation energy of 10.1 kcal/mol) afforded 3-si-
lylfuran C-2. In both carbene intermediates T-1 and C-1,
direct 1,2-H migrations were predicted to have higher
activation energy barriers than those for direct 1,2-Si shifts
Thus, DFT calculations were performed to evaluate pos-
sible counterion effects8,18 and establish overall reaction
pathways for cycloisomerization of homopropargylic ke-
tones 6 in the presence of electrophilic Ph3PAuOTf and
Ph3PAuSbF6 catalysts in DCE or toluene solvents, respec-
tively. The potential energy surfaces for the predictions of
-
the effects of TfO- and SbF6 counterions are depicted in
Schemes 6 and 7, respectively. In the case of Ph3PAuOTf
catalyst in DCE solution (Scheme 6), a highly exergonic (21.3
kcal/mol) direct 5-endo-dig cyclization of complex T-C
Scheme 6. DFT Predictions for the Ph3PAuOTf-Catalyzed Reaction ([Au] ) AuPH3, L ) OTf, R ) H) in 1,2-Dichloroethane Solutiona
a Relative energies (∆GDCE, ∆G298, ∆H298, ∆E0) are in kcal/mol.
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7649

A R T I C L E S
Dudnik et al.
Scheme 7. DFT Predictions for the Ph3PAuSbF6-Catalyzed Reaction ([Au] ) AuPH3, L ) SbF6, R ) H) in 1,2-Dichloroethane and Toluene
Solutionsa
a Relative energies (∆GDCE, ∆GTol) are in kcal/mol.
(C-TS2 versus C-TS1, activation free energy of 22.3 versus
10.1 kcal/mol, respectively). In addition, an alternative
reaction pathway, involving generation of reactive allene-
Au complex T-5 via Ph3PAuOTf-assisted propargyl-allenyl
isomerization of T-C was found to be higher in energy (T-
TS6, 16.4 kcal/mol) than the direct cyclization route in DCE.
Thus, based on the predicted 0.2 kcal/mol difference between
R- (T-TS4) and ꢀ- (T-TS2) protonation paths, competing
formation of products 2 and 3 can be expected in the
Ph3PAuOTf-catalyzed cycloisomerization of homopropargylic
ketone 6. It should be mentioned that analogous high-energy
profiles were later found for the propargyl-allenyl isomer-
ization of 6 into 5 and for the direct 1,2-H shift in Au-carbene
intermediates 1 in toluene solvent (Vide infra). The optimized
structures of key stationary points along the reaction pathway
for the Ph3PAuOTf-catalyzed cycloisomerization of alkynes
6 are presented in Figure 2.
In the case of Ph3PAuSbF6 catalyst (Scheme 7) in DCE
solvent, direct 5-endo-dig cyclization of complex H-C
afforded cyclic zwitter-ionic intermediate H-IN1 via the
H-TS1 transition state, requiring an activation free energy
of 10.8 kcal/mol. However, we were not able to locate the
next transition state for the highly endergonic (23.6 kcal/
mol) formation of the furyl-Au intermediate T-IN1 and
HSbF6, probably due to the extreme acidity of HSbF6. An
alternative pathway, involving a [1,5]-hydride shift via
H-TS2, was also found to be high in energy (activation free
energy of 22.2 kcal/mol). We believe that the extremely high
acidity of HSbF6 and very low nucleophilicity of its conjugate
base SbF6 complicate DFT simulation of the elementary
-
reaction processes involving a counterion-assisted proton
transfer step. This reasoning prompted us to perform DFT
calculations of the Ph3PAuSbF6-catalyzed reaction in the
presence of a three water molecule cluster (H2O)3, which
could model reactivity of a counterion, which is more
nucleophilic than the SbF6- counterion and much milder than
HSbF6 acid in its monoprotonated form. Accordingly, initial
coordination of the water cluster to homopropargylic ketone
H-C provided HW-C1 endergonically (free energy of 10.4
kcal/mol in DCE).8,13h Similarly to the TfO- counterion, the
water cluster could also abstract the propargylic H-atom
during a subsequent 5-endo-dig cyclization of HW-C1 (HW-
TS1, activation free energy of 9.1 kcal/mol) to give the furyl-
Au HW-IN1 and protonated water cluster.19 However, a
lower energy reaction path was found for the complex HW-
C1. Thus, formation of the activated allene H-5 from the
latter occurred as a stepwise H-abstraction/donation process
via transition states HW-TS4 and HW-TS5, respectively. The
first step of this process required an activation free energy
of only 3.9 kcal/mol in DCE and 2.9 kcal/mol in toluene,
respectively, leading to the intermediate HW-IN2 exergoni-
cally. The second quite facile H-donation step to give HW-5
occurred with an activation energy of 7.3 kcal/mol (DCE)
or 8.4 kcal/mol (toluene) and was followed by exergonic
decoordination of the water cluster to produce allene-Au
complex H-5. A facile cyclization of the latter via H-TS3
9
7650 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

DFT-Guided Regiodivergent Synthesis of Silylfurans
A R T I C L E S
Figure 2. Geometries of the intermediates and transition states for the Ph3PAuOTf-catalyzed reaction; selected distances are in angstroms.
provided Au-carbene H-1 with an activation free energy of
5.3 kcal/mol in DCE and 4.9 kcal/mol in toluene. A
subsequent 1,2-Si migration in H-1 via H-TS4 was strongly
favored over the 1,2-H shift (H-TS5) in both DCE and
toluene solvents, leading to the 1,2-Si migration product
complex H-2. The optimized structures of key stationary
points along the reaction pathway for the Ph3PAuSbF6-
catalyzed cycloisomerization of alkynes 6 are presented in
Figure 3.
the DFT predictions cannot explain a selective formation of the
1,2-Si migration product 2l during the cycloisomerization of 6l
(entry 10). In addition, it was found that AuCl3 could also
catalyze cycloisomerization of alkynes 6; however, the regi-
oselectivity of this reaction toward the 1,2-Si migration product
2f was poor in a variety of solvents (entries 4-6). Unfortunately,
an analogous study on counterion effect in allenyl systems was
not possible since allenes 4 were not stable in the presence of
Ph3PAuOTf and even less electrophilic Ph3PAuOBz complexes.
Solvent Effects on the Au(I)-Catalyzed 1,2-Si Migration in
Alkynes. Remarkably, it was found that the solvent effect
could dominate over the counterion effect, as evidenced by
a reverse regioselectivity observed in the Ph3PAuOTf-
catalyzed cycloisomerization of 6 (Table 2, entries 2 vs 3
and 9 vs 10).20 Accordingly, when the reaction was conducted
To examine the above DFT-based prognosis, we investigated
cycloisomerization of the homopropargylic ketones 6 in the
presence of various Au catalysts (Table 2). It was found that
the experimental results overall are in good agreement with the
computational predictions.14 Thus, the Au-catalyzed cycloi-
somerization of alkynes 6 in the presence of the non-nucleophilic
-
SbF6 counterion in DCE gives exclusive formation of 1,2-Si
migration products 2f, l, and k (entries 1, 7, 11). As suggested
by the DFT calculations, switching to Ph3PAuOTf catalyst in
DCE with a more nucleophilic counterion leads to a competing
formation of 1,2-Si and 1,2-H migration products 2f and 3f,
respectively, for the cycloisomerization of 6f (entry 2). However,
(19) Simulation of a base in the Ph3AuOTf-catalyzed transformation using
water cluster demonstrated that the 5-endo-dig cyclization path has
the lowest energy profile, similarly to that for TfO--counterion base.
See Supporting Information for details.
(20) For a single example of a solvent effect in the Au-catalyzed synthesis
of pyrroles, see ref 7i.
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7651

A R T I C L E S
Dudnik et al.
Figure 3. Geometries of the intermediates and transition states for the Ph3PAuSbF6-catalyzed reaction; selected distances are in angstroms.
in DCE solution, the preferred formation of 1,2-Si migration
products was observed. However, when toluene was em-
ployed as a solvent, the 1,2-H migration products were the
major ones. To shed light on these quite uncommon results,20
DFT calculations for evaluation of the solvent effects on the
1,2-H and 1,2-Si migrations were performed. DFT predictions
for the Ph3PAuOTf-catalyzed reaction in DCE solvent were
provided in Scheme 6. The results for toluene solvent are
outlined in Scheme 8. According to the calculated potential
energy surfaces, employment of toluene solvent leads to the
intermediate T-IN2 via 5-endo-dig cyclization of T-C with
a reaction energy profile similar to that for DCE solvent.
However, in the case of toluene solvent, there is no difference
in activation energy barriers for R- (T-TS4) and ꢀ- (T-TS4)
protonation pathways of furyl-Au complex T-IN2. Most
importantly, while solvents do not affect the relative barriers
of the subsequent 1,2-H and 1,2-Si migrations from Au-
carbene intermediates T-1 or C-1 much, they alter the
stability of the dissociated noninteracting Au-carbene C-1
and TfO- anion relative to the associated Au-carbene
complex T-1. For instance, when the reaction is conducted
in toluene solution, which has a lower polarity than DCE,
the stabilization effects of the solvent on the charged species
C-1 and TfO- are limited (compare relative energies of C-1
intermediates in Schemes 6 and 8). On the other hand, an
alternative direct 1,2-Si migration from carbene T-1 via
T-TS3 has an activation free energy barrier of 10.7 kcal/
mol. Since complex T-1 does not undergo dissociation of
9
7652 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

DFT-Guided Regiodivergent Synthesis of Silylfurans
A R T I C L E S
Scheme 8. Solvation Effect of Toluene on the Ph3PAuOTf-Catalyzed Reaction (R ) H)a
a Relative free energies in toluene (∆GTol) are in kcal/mol.
the TfO- ligand in toluene solution easily, its formation
Table 2. Optimization of the Au-Catalyzed Cycloisomerization of
Homopropargyl Ketones 6
becomes reversible, forcing the reaction to proceed via the
R-protonation route T-IN2 - T-TS4 - T3, which is 3.9 kcal/
mol lower in activation free energy than the direct 1,2-Si
shift in T-1. The DFT calculations suggest that the Ph3PAuOTf-
catalyzed cycloisomerization of 6 in toluene solvent should
lead to the 1,2-H migration product 3, thus supporting the
experimentally observed reversed regioselectivity of the
reaction (Table 2, entries 2 vs 3 and 9 vs 10). In contrast to
the above case, no solvent effect was observed in the DFT-
simulated Ph3PAuSbF6-catalyzed reactions (Scheme 7). Thus,
formation of the allene-Au complex H-5 and its cyclization
into the Au-carbene H-1 were found to be more efficient than
other possible processes in both DCE and toluene solvents.
In addition, energies of the direct or base/SbF6--assisted
1,2-H migrations were much higher than that of the 1,2-Si
migration from H-1, regardless on the solvent (compare
entries 7 and 8 in Table 2). Thus, the experimentally observed
migratory preference (Table 2, entries 1, 7, 8, and 11) for
the Ph3PAuSbF6-catalyzed reactions was in a good agreement
with the performed DFT calculations. Finally, it is worth
mentioning here that, in contrast to the cycloisomerization
of alkynes 6, employment of solvents more polar than toluene
did not affect the AuCl3-catalyzed cycloisomerization of
allenyl ketone 4e (Table 1, entry 5).
Entry
R1
Catalyst
Solvent (C, M)
2:3 Ratioa
1
2
3
4
5
6
7
8
Ph (6f)
Ph3PAuSbF6
Ph3PAuOTf
Ph3PAuOTf
AuCl3
AuCl3
AuCl3
Ph3PAuSbF6
Ph3PAuSbF6
Ph3PAuOTf
Ph3PAuOTf
Ph3PAuSbF6
Ph3PAuOTf
DCE (0.05)
DCE (0.05)
PhMe (0.1)
PhMe (0.1)
DCE (0.1)
MeNO2 (0.05)
DCE (0.05)
PhMe (0.05)
PhMe (0.1)
DCE (0.05)
DCE (0.05)
PhMe (0.1)
100:0
66:34
7:93
60:40
72:28
84:16
100:0
100:0
12:88
98:2
n-C8H17 (6l)
t-Bu (6k)
9
10
11
12
100:0
17:83
a H1 NMR ratio of products for reactions performed on 0.5 mmol
scale.
PhMe2Si (entry 9) groups could be successfully employed
in this transformation. In the case of cycloisomerization of
substrates 6 bearing electron-deficient aryl substituents
(entries 3-5), the 1,2-H shift competed with the 1,2-Si
migration, which resulted in formation of isomeric silylfurans
2 and 3 with good to high degrees of regioselectivity favoring
3-silylfurans 2. It was also demonstrated that a variety of
functional groups such as methoxy (entry 2), bromo (entry
3), cyano (entry 4), and nitro (entry 5) could be tolerated
under these reaction conditions.
Synthesis of Furans via Au(I)-Catalyzed 1,2-Si Migration
in Alkynes. Next, we investigated the scope of the above 1,2-
Si migration cascade reaction of alkynes (Table 3). Thus,
cycloisomerization of a variety of 4-silyl homopropargylic
ketones 6, possessing aryl (entries 1, 2), primary (entries
7-9), and tertiary (entry 6) alkyl groups, proceeded smoothly
affording the C2-unsubstituted 3-silylfurans 2 as sole regioi-
somers in good to excellent yields (Table 3). It was found
that alkynes bearing TMS (entries 1-7), TES (entry 8), and
Conclusion
In summary, we developed a regiodivergent Au-catalyzed
cycloisomerization of allenyl and homopropargylic ketones into
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7653

A R T I C L E S
Dudnik et al.
Table 3. Au-Catalyzed Cycloisomerization of Homopropargyl Ketones 6
a Isolated yield of product for reactions performed on 0.5 mmol scale. b 2f and 2l contained 8 and 6% of the corresponding desilylated furan,
respectively. c 14:1 ratio of 2h:3h. d 10:1 ratio of 2i:3i. e 8:1 ratio of 2j:3j. f GC Yield, 2k is a volatile compound.
2- and 3-silylfurans featuring 1,2-Si or 1,2-H migrations in a
common Au-carbene intermediate. The present methodology
constitutes a general and efficient route for the synthesis of
3-silylfurans, important synthons,21 not easily available via the
existing methodologies. Both experimental and computational
results indicate that the 1,2-Si migration is kinetically more
favored over the 1,2-shifts of H, alkyl, and aryl groups in the
ꢀ-Si-substituted Au-carbenes. However, according to the ex-
perimental data obtained for the cycloisomerization of homopro-
pargylic ketones, counterion and solvent effects may reverse
this migratory preference. The DFT calculations provided a
rationale for this 1,2-migration regiodivergency. Accordingly,
in the case of Au(I) catalyst possessing non-nucleophilic SbF6
-
counterion, DFT-simulated reaction affords 1,2-Si migration
products exclusively via the initial propargyl-allenyl isomer-
ization followed by cyclization into the Au-carbene intermediate
regardless of the solvent employed in the reaction. However,
in the case of the TfO- counterion in nonpolar solvents, a
pronounced counterion effect is observed. Thus, reaction occurs
via the initial 5-endo-dig cyclization to give a cyclic furyl-Au
intermediate followed by a subsequent ipso-protiodeauration,
which is kinetically more favorable than the generation of the
common Au-carbene intermediate and leads to the formation
of formal 1,2-H migration products. In contrast, when polar
(21) For review on silylfurans, see: Keay, B. A. Chem. Soc. ReV. 1999,
28, 209.
9
7654 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

DFT-Guided Regiodivergent Synthesis of Silylfurans
A R T I C L E S
Acknowledgment. This work was supported by the Hundreds
of Talents Program of CAS, Natural Science Foundation of China
(20872105), and NIH of the US (GM-64444).
solvent is employed, ꢀ-to-Au protonation of the furyl-Au
species, leading to the Au-carbene intermediate, competes with
the ipso-protiodeauration. Subsequent dissociation of the triflate
ligand in the latter carbene in polar media due to efficient
solvation facilitates formation of the 1,2-Si shift products. The
above results of DFT calculations were validated by the
experimental data. Finally, the present study highlights that DFT
calculations could efficiently complement experimental results,
providing guidance for design of new transformations and
offering a better understanding of this chemistry.
Supporting Information Available: Computational and ex-
perimental details as well as full citation of computational
methods. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA910290C
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7655