Thermal and metal-catalyzed cyclization of 1-substituted 3,5-dien-1-ynes via a [1,7]-hydrogen shift: Development of a tandem aldol condensation- dehydration and aromatization catalysis between 3-en-1-yn-5-al units and cyclic ketones

Article abstract of DOI:10.1021/ja061203b

This work investigates the feasibility of thermal and catalytic cyclization of 6,6-disubstituted 3,5-dien-1-ynes via a 1,7-hydrogen shift. Our strategy began with an understanding of a structural correlation of 3,5-dien-1-ynes with their thermal cyclizati

Full text of DOI:10.1021/ja061203b

Published on Web 07/11/2006
Thermal and Metal-Catalyzed Cyclization of 1-Substituted
3,5-Dien-1-ynes via a [1,7]-Hydrogen Shift: Development of a
Tandem Aldol Condensation-Dehydration and Aromatization
Catalysis between 3-En-1-yn-5-al Units and Cyclic Ketones
Jian-Jou Lian, Chung-Chang Lin, Hsu-Kai Chang, Po-Chiang Chen, and
Rai-Shung Liu*
Contribution from the Department of Chemistry, National Tsing-Hua UniVersity, Hsinchu 30043,
Taiwan, Republic of China
Received February 20, 2006; E-mail: rsliu@mx.nthu.edu.tw
Abstract: This work investigates the feasibility of thermal and catalytic cyclization of 6,6-disubstituted 3,5-
dien-1-ynes via a 1,7-hydrogen shift. Our strategy began with an understanding of a structural correlation
of 3,5-dien-1-ynes with their thermal cyclization efficiency. Thermal cyclization proceeded only with 3,5-
dien-1-ynes bearing an electron-withdrawing C(1)-phenyl or C(6)-carbonyl substituent, but the efficiencies
were generally low (20-40% yields). On the basis of this structure-activity relationship, we conclude that
such a [1,7]-hydrogen shift is characterized by a “protonic” hydrogen shift, which should be catalyzed by
π-alkyne activators. We prepared various 6,6-disubstituted 3,5-dien-1-ynes bearing either a phenyl or a
carbonyl group, and we found their thermal cyclizations to be greatly enhanced by RuCl3, PtCl2, and
TpRuPPh3(CH3CN)2PF6 catalysts to confirm our hypothesis: the C(7)-H acidity of 3,5-dien-1-ynes is crucial
for thermal cyclization. To achieve the atom economy, we have developed a tandem aldol condensation-
dehydration and aromatization catalysis between cycloalkanones and special 3-en-1-yn-5-als using the
weakly acidic catalyst CpRu(PPh3)2Cl, which provided complex 1-indanones and R-tetralones with yields
exceeding 65% in most cases. The deuterium-labeling experiments reveal two operable pathways for the
metal-catalyzed [1,7]-hydrogen shift of 3,5-dien-1-ynes. Formation of R-tetralones d4-56 arises from a
concerted [1,7]-hydrogen shift, whereas benzene derivative d4-9 proceeds through a proton dissociation
and reprotonation process.
Scheme 1
Introduction
Thermal cyclization of 3,5-dien-1-ynes to benzene derivatives
normally proceeds above 200 °C, and this reaction has attracted
considerable theoretical and experimental interest.1 In the
presence of suitable metal catalysts, cyclization of 3,5-dien-1-
ynes occurs under mild conditions via exclusive formation of
either a metal-vinylidene (A)2,3 or a π-alkyne intermediate (B)4
as depicted in Scheme 1. There is no report of metal-catalyzed
aromatization of 3,5-dien-1-ynes via other mechanisms in the
(1) Thermal cyclization and theoretical calculations of 3,5-dien-1-ynes, see:
(a) Zimmermann, G. Eur. J. Org. Chem. 2004, 457 (a review). (b) Litovitz,
A. E.; Carpenter, B. K.; Hopf, H. Org. Lett. 2005, 7, 507. (c) Christl, M.;
Braun, H.; Muller, G. Angew. Chem., Int. Ed. Engl. 1992, 31, 473. (d)
Hopf, H.; Berger, H.; Zimmerman, G.; Nuchter, U.; Jones, P. G.; Dix, I.
Angew. Chem., Int. Ed. Engl. 1997, 36, 1187. (e) Bettinger, H. F.; Schreiner,
P. R.; Schaefer, H. F.; Schleyer, P. v. R. J. Am. Chem. Soc. 1998, 120,
5741.
(2) Review for metal-vinyldene catalysis, see: (a) Trost, B. M. Acc. Chem.
Res. 2002, 35, 695. (b) Bruneau, C.; Dixneuf, P. Acc. Chem. Res. 1999,
32, 311. (c) Puerta, M. C.; Valerga, P. Coord. Chem. ReV. 1999, 193-
195, 977. (d) Bruneau, C. Top. Organomet. Chem. 2004, 11, 125.
(3) For catalytic aromatization of 3,5-dien-1-ynes via metal-vinylidene inter-
mediates, see selected examples: (a) Merlic, C. A.; Pauly, M. E. J. Am.
Chem. Soc. 1996, 118, 11319. (b) Maeyama, K.; Iwasawa, N. J. Org. Chem.
1999, 64, 1344. (c) Iwasawa, N.; Shido, M.; Kusawa, H. J. Am. Chem.
Soc. 2001, 123, 5815. (d) O’Connor, J. M.; Friese, S. J.; Tichenor, M. J.
Am. Chem. Soc. 2002, 124, 3506. (e) Martin-Matute, B.; Nevado, C.;
Cardenas, D. J.; Echavarren, A. M. J. Am. Chem. Soc. 2003, 125, 5757. (f)
Shen, H.-C.; Pal, S.; Lian, J.-J.; Liu, R.-S. J. Am. Chem. Soc. 2003, 125,
15762.
literature.5 Here, we report a new metal-catalyzed cyclization
of 3,5-dien-1-ynes via a [1,7]-hydrogen shift of π-alkyne
intermediate C (path a, eq 2).
(4) For catalytic aromatization of 3,5-dien-1-ynes via metal-π-alkyne intermedi-
ates, see: (a) Mamane, V.; Hannen, P.; Fu¨rstner, A. Chem.-Eur. J. 2004,
10, 4556 and references therein. (b) Dankwardt, J. W. Tetrahedron Lett.
2001, 42, 5809.
9
10.1021/ja061203b CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 9661-9667
9661

A R T I C L E S
Scheme 2
Lian et al.
Table 1. Cyclization of 1-Substituted 3,5-Dien-1-ynes by a
Thermal and Ruthenium-Catalyzed Process ([TpRu] )
TpRuPPh3(CH3CN)2PF6)
thermala
time (yields)c
catalystb
time (yields)c
dienynes
products
X ) CH2
RuCl3
[TpRu]
(1) R ) TMS (1)
(2) R ) nPr (2)
(3) R ) Ph (3)
X ) (CH2)0
72 h (N.R.d) dec.e
dec.e
dec.e
72 h (N.R.d) dec.e
8
9
48 h (26%)
36 h (31%)
24 h (71%) 24 h (75%)
24 h (78%) 24 h (83%)
12 h (92%) 12 h (86%)
(4) R ) Ph (4)
X ) CO
(5) R ) TMS (5)
(6) R ) nPr (6)
(7) R ) Ph (7)
10
11
12
36 h (58%)
72 h (12%)
72 h (33%)
dec.e
30 h (74%)
18 h (81%) 18 h (61%)
a Xylene, 120 °C, [substrate] ) 0.5 M. b 5 mol % catalyst, [substrate] )
0.5 M, xylene, 120 °C. c Yields are given after separation from a silica
column. d Starting 1 and 2 were recovered in 58% and 62% yields,
respectively. e No starting substrates were recovered with the same reaction
period as the thermal process.
the best of our knowledge, such a process was proposed only
as a possible route for flash vacuum pyrolysis (500 °C) of
2-methylstyrylalkynes to 2-alkenylnaphthalenes,10 and acid-
catalyzed thermal cyclization of acetylenic enones to benzene
products (200-250 °C).11 Both reactions proceeded sluggishly
at elevated temperatures.10,11 We first investigated the thermal
feasibility of this process, with an emphasis on the structural
correlation of substrates with their thermal activities. Table 1
summarizes the results for thermal cyclization (xylene, 120 °C,
36-72 h) of model molecules 1-7 via alternation of their C(1)-
alkynyl and C(6)-alkenyl substituents; species 5-7 have E-
configuration. Entries 1-3 show the C(1)-phenyl effect of 3,5-
dien-1-yne 3, which was the only active species in thermal
cyclization. The C(6)-carbonyl groups of dienynes 5-7 also
show a significant improvement in both thermal and catalytic
cyclizations as compared to their cycloalkyl analogues 1-4.
These results indicate the importance of the C(7)-H acidity of
3,5-dien-1-ynes, which is crucial for both thermal and catalytic
cyclization. Accordingly, we selected electrophilic π-alkyne
Recently, we reported6 a catalytic cyclization of cis-3-en-1-
ynes using TpRuPPh3(CH3CN)2PF6 catalyst, and this process
actually mimics the thermal [1,5]-hydrogen shift of cis-1-allen-
4-enes because the terminal alkyne of substrate was transformed
into ruthenium-vinylidene species E (eq 1, Scheme 2). We
thought that this catalytic process was extensible to the
aromatization of 6,6-disubstituted 3,5-dien-1-yne through an
analogous [1,7]-hydrogen shift as represented by intermediate
D (path b, eq 2, Scheme 1); instead, we observed a competitive
6-π-electrocyclization of intermediate F to give highly structur-
ally reorganized products with two selected instances shown in
Scheme 2 (eq 2). The mechanism was supported by deuterium-
labeling experiments.7 We applied this catalytic cyclization to
2′,2′-disubstituted o-(ethynyl)styrenes, but the ruthenium-vi-
nylidene intermediate G underwent 5-endo-dig cyclization to
give indene products with a considerable skeletal rearrangement
(eq 3).8 After unsuccessful efforts with metal-vinylidene ap-
proaches, we seek an alternative solution using π-alkyne
activators, which ultimately lead to aromatization of 3,5-dien-
1-ynes via a [1,7]-hydrogen shift.
(10) Aitken, R. A.; Boeters, C.; Morrison, J. J. J. Chem. Soc., Perkin Trans. 1
1997, 2625.
(11) Collindine p-toluenesulfonate (CPTS) was attempted to catalyze the
cyclization of acetylenic enones to benzene products, but the efficiencies
were very low (4-19%) under drastic temperatures of 200-250 °C. This
cyclization is thought to proceed via initial formation of 3,5-dien-1-ynes
(a) by tautomerization, followed by a second acid-catalyzed rearrangement
to 3,5-diene-1-allene. The low yields of products were attributed to the
kinetically instability of enol intermediates a,b caused by enol-ketone
tautomerization. See: Jacobi, P. A.; Kravitz, J. I. Tetrahedron Lett. 1988,
29, 6873.
Results and Discussion
Structure-Activity Correlation in the [1,7]-Hydrogen
Shift. Although a [1,7]-hydrogen shift is well documented for
1-substituted 1,3,5-trienes,9 little is known about the thermal
cyclization of 3,5-dien-1-ynes via a 1,7-hydrogen shift.10 To
(5) Very recently, O’Connor and co-workers reported an unusual aromatization
of 3,5-dien-1-ynes via a η6-coordinated mode in a stoichiometric reaction.
See: O’Connor, J. M.; Friese, S. J.; Rodgers, B. L.; Rheingold, A. L.;
Zakharov, L. J. Am. Chem. Soc. 2005, 127, 9346.
(6) Datta, S.; Odedra, A.; Liu, R.-S. J. Am. Chem. Soc. 2005, 127, 11606.
(7) Lian, J.-J.; Odedra, A.; Wu, C.-J.; Liu, R.-S. J. Am. Chem. Soc. 2005, 127,
4186.
(8) Madhushaw, R. J.; Lo, C.-Y.; Su, M.-D.; Shen, H.-C.; Pal, S.; Shaikh, I.
R.; Liu, R.-S. J. Am. Chem. Soc. 2004, 126, 15560.
(9) Selected examples for a thermal 1,7-hydrogen shift of 1,3,5-trienes: (a)
Hess, B. A., Jr. J. Org. Chem. 2001, 66, 5897. (b) Baldwin, J. E.; Reddy,
V. P. J. Am. Chem. Soc. 1987, 109, 8051. (c) Okamura, W. H.; Hoeger, K.
J.; Miller, K. J.; Reischl, W. J. Am. Chem. Soc. 1988, 110, 973. (d) Steuhl,
H.-M.; Bornemann, C.; Klessinger, M. Chem.-Eur. J. 1999, 5, 2404.
9
9662 J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006

Cyclization of 1-Substituted 3,5-Dien-1-ynes
A R T I C L E S
activators12-14 to enhance their C(7)-H acidity and aromati-
zation efficiencies. Among substrates 1-7, thermal cyclization
performs most efficiently with compound 5 bearing a trimeth-
ylsilyl (TMS) and a carbonyl group at its C(1)- and C(6)-
carbons, respectively. The yield of benzene derivative 10 was
Table 2. Cyclization of Various 3,5-Dien-1-ynes by a Thermal and
Ruthenium-Catalyzed Process ([TpRu] ) TpRuPPh3(CH3CN)2PF6)
15
58% (entry 5), but was increased significantly using RuCl3
(92%) and TpRuPPh3(CH3CN)2PF616 (89%) at 5 mol % catalyst
loading. The retention of the trimethysilyl functionality of
benzene derivative 10 precludes the participation of proton in
12
the catalytic transformation. In entry 5, PtCl2 gave benzene
10 (xylene/DMF ) 2/1, 120 °C, 16 h) equally efficiently with
86% yield but at a 10% loading; AuCl3 (5 mol %)13 produced
the desilylated derivative of benzene 10 in 78% yield (DMF,
120 °C, 18 h). Although thermal cyclization of internal alkynes
3, 4, 6, and 7 proceeded unsatisfactorily (12-33%), their
15
efficiencies can be significantly improved with RuCl3 and
TpRuPPh3(CH3CN)2PF6,16 which gave desired benzenes 8, 9,
11, and 12 with yields up to 75-92%. The molecular structure
1
of benzene 10 was confirmed by H NOE spectra.
The Scope of a Catalytic [1,7]-Hydrogen Shift. To test our
hypothesis that the cyclization efficiency is determined by the
C(7)-H acidity of substrates, we prepared additional 3,5-dien-
1-ynes 13-22 in Table 2, which contained exclusively either a
C(1)-phenyl or C(6)-carbonyl substituent. Similar to the preced-
ing examples, 3,5-dien-1-ynes 16 and 22 are more active than
other substrates in thermal cyclization because of their activating
C(1)-TMS and C(6)-carbonyl groups. Again, we observed an
obvious enhancement of cyclization using RuCl3 and TpRuPPh3-
(CH3CN)2PF6 catalysts (5 mol %) when substrates 13, 14, and
18 performed inefficiently in the thermal process. The resulting
products 23, 24, and 28 were obtained in 26-42% yields in
hot xylene, but were greatly increased to 81-91% using either
RuCl3 or TpRuPPh3(CH3CN)2PF6 (entries 1, 2, and 6). For the
remaining substrates 15-17 and 19-22 bearing a cycloketone
group, the two ruthenium species showed a similar catalytic
improvement, and gave the observed benzenes 25-27 and 29-
32 with yields up to 77-93%, much higher than their thermal
yields (entries 3-5, 7-10) in most instances. Although the
structure-activity relationship follows the same trend as in Table
a Xylene, 120 °C, [substrate] ) 0.5 M. b 5 mol % catalyst, [substrate] )
0.5 M, xylene, 120 °C. c Reaction time for RuCl3 (5%) and TpRuPPh3(CH3-
CN)2PF6 (5%). d Yields were given after separation from a silica column.
Table 3. Thermal and Catalyzed Cyclization of 3,5-Dien-1-ynes
Bearing Acyclic Substituents
b
c
dienynes
products
thermala
AuCl3
PtCl2
(1) R1 ) nBu, R2 ) CO2Et (33)
(2) R1 ) nBu, R2 ) COPh (34)
(3) R1 ) Ph, R2 ) CO2Et (35)
(4) R1 ) Ph, R2 ) COPh (36)
(5) R1 ) H, R2 ) CO2Et (37)
(6) R1 ) H, R2 ) COPh (38)
(7) R1 ) TMS, R2 ) CO2Et (39)
(8) R1 ) TMS, R2 ) COPh (40)
N.R.
26%
30%
37%
41%
45%
27%
25%
N.R.
68%
74%
76%
88%
89%
d.s.e
d.s.
N.R.
66%d
61%
80%
78%
82%
76%
71%
41
42
43
44
45
46
47
(12) Selected reviews for π-alkyne activators including various platinum and
gold catalysts: (a) Hashimi, A. S. K. Angew. Chem., Int. Ed. 2005, 44,
6990. (b) Bruneau, C. Angew. Chem., Int. Ed. 2005, 44, 2328. (c) Ma, S.;
Yu, S.; Gu, Z. Angew. Chem., Int. Ed. 2006, 45, 200.
a [substrates] ) 0.5 M, DMF, 140 °C, 72 h. b 5 mol % AuCl3, [substrate]
) 0.5 M, DMF, 24-32 h, 140 °C. c 10 mol % PtCl2, [substrate] ) 0.5 M,
140 °C, 30-36 h. d Product yields are reported after separation from a silica
column. e Desilylation occurred with formation of products 44 and 45 in
80% and 81% yields, respectively.
(13) For recent catalytic reactions, using PtCl2 salts as π-alkyne activators, see:
(a) Fu¨rstner, A.; Davies, P. W.; Gress, T. J. Am. Chem. Soc. 2005, 127,
8244. (b) Qian, H.; Han, X.; Widenhoefer, R. A. J. Am. Chem. Soc. 2004,
126, 9536. (c) Nakamura, I.; Mizushima, Y.; Yamamoto, Y. J. Am. Chem.
Soc. 2005, 127, 15022. (d) Chatani, N.; Furukawa, N.; Sakurai, H.; Murai,
S. Organometallics 1996, 15, 901. (e) Mendez, M.; Munoz, M. P.; Nevado,
C.; Cardenas, D. J.; Echavarren, A. M. J. Am. Chem. Soc. 2001, 123, 10511.
(f) Nevado, C.; Cardenas, D. J.; Echavarren, A. M. Chem.-Eur. J. 2003, 9,
2627. (g) Fu¨rstner, A.; Szillat, H.; Stelzer, F. J. Am. Chem. Soc. 2000,
122, 6785. (h) Fu¨rstner, A.; Stelzer, F.; Szillat, H. J. Am. Chem. Soc. 2001,
123, 11863. (i) Marion, F.; Coulomb, J.; Courillon, C.; Fensterbank, L.;
Malacria, M. Org. Lett. 2004, 6, 1509.
(14) For recent catalytic reactions using Au(I) and Au(III) as π-alkyne activators,
see selected examples: (a) Kennedy-Smith, J. J.; Staben, S.; Toste, F. D.
J. Am. Chem. Soc. 2004, 126, 4526. (b) Gorin, D. J.; Davies, N. R.; Toste,
F. D. J. Am. Chem. Soc. 2005, 127, 11260. (c) Nieto-Oberhuber, C.; Munoz,
M. P.; Bunuel, E.; Nevedo, C.; Cardenas, D. J.; Echavarren, A. M. Angew.
Chem., Int. Ed. 2004, 43, 2402. (d) Nieto-Oberhuber, C.; Lopez, S.;
Echavarren, A. M. J. Am. Chem. Soc. 2005, 127, 6178. (e) Mamane, V.;
Gress, T.; Krause, H.; Fu¨rstner, A. J. Am. Chem. Soc. 2004, 126, 8654. (f)
Zhang, L.; Kozmin, S. A. J. Am. Chem. Soc. 2005, 127, 6962. (g) Georgy,
M.; Boucard, V.; Campagne, J.-M. J. Am. Chem. Soc. 2005, 127, 14180.
(h) Shi, Z.; He, C. J. Am. Chem. Soc. 2004, 126, 13596.
1, we notice that dienynes 4, 13, and 19 bearing a five-
membered carbocyclic ring performed better in catalytic aro-
matization than did their six-membered analogues 3, 14, and
20.
We further extended this catalytic cyclization to 3,5-dien-1-
ynes 33-40 bearing two acyclic substituents (Table 3); their
cyclizations require higher temperature (140 °C) probably due
16
to the conformational flexibility of acyclic substituents. AuCl3
(5 mol %, DMF, 18-30 h) and PtCl2 (10 mol %, DMF/xylene
) 1, 24-46 h) are far superior to RuCl3 and TpRuPPh3(CH3-
CN)2PF, which gave approximately the same yields as those in
thermal activation. DMF solvent serves to prevent AuCl3 from
a rapid formation of a gold mirror as observed in hot 1,2-
(15) We selected RuCl3 as an activator because it was reported to be the best
catalyst in the literature for aldol condensation and dehydration between
aldehydes and cycloalkanones, see: Iranpoor, N.; Kazemi, F. Tetrahedron
1998, 54, 9475.
9
J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006 9663

A R T I C L E S
Lian et al.
Scheme 3
dichloroethane, toluene, and xylene, which gave products in a
messy mixture.17,18 AuCl3 and PtCl2 efficiently catalyzed
cyclization of dienynes 34-40 bearing a C(6)-carbonyl sub-
stituent and gave good yields (68-89%) of benzene products
41-47 even though their thermal yields were generally low
(25-45%). Notably, species 33 bearing a C(6)-ester substituent
is totally inactive; this observation emphasizes the importance
of a functional group (R1 ) TMS, H, and Ph) at the C(1)-carbon
as in species 35, 37, and 39. A drawback in the AuCl3-catalyzed
reaction is the occurrence of desilylation as shown in entries 7
and 8; this reaction is caused probably by partial hydrolysis of
AuCl3 with residual water to generate a small amount of HCl,
which catalyzed the cleavage of the reactive C-Si bond of
benzenes 46 and 47. To understand the role of HCl, in separate
experiments, we heated 3,5-dien-1-yne 36 and 39 with HCl (10
mol %) that led to a serious polymerization.19
furyl derivatives. To accomplish this process, we focused on
3-en-1-yn-5-als bearing a C(1)-TMS group because their result-
ing 3,5-dien-1-yn-7-ones gave thermal cyclization in moderate
yields (ca. 50-60%), which could be equally improved by weak
π-alkyne activators. Furthermore, weakly acidic catalysts are
expected to be less active in the self-cyclization of 3-en-1-yn-
5-al, but more active in aldol condensation.22-25
Development of Tandem Catalytic Aldol Condensation-
Dehydration and Aromatization Reactions. In this study, the
synthesis of final 1-indanones and R-tetralones requires four-
step operations beginning with 3-en-1-yn-5-als and cycloal-
kanones; three-step operations served to prepare 3,5-dien-1-yn-
7-ones through enol silane formation, Mukaiyama aldol
condensation, and a dehydration protocol20 (Scheme 3). To
achieve an atom economy, development of a tandem aldol
condensation-dehydration and aromatization catalysis between
3,5-dien-1-als and certain ketones is highly desired, but this
approach seems formidable because of an eminent problem:
electrophilic π-alkyne activators such as AuCl3,18b,c PtCl2,21 and
TpRuPPh3(CH3CN)2PF616c were reported be very active in self-
cyclization of 3-en-1-yn-5-als to form undesired benzo[c]-
pyrylium or 2-(furyl)carbenoid species (Scheme 3), which were
catalytic intermediates for syntheses of special aromatic and
Although catalytic aldol reaction and its subsequent dehydra-
tion have attracted considerable attention,22,23 many studies
focused on Mukaiyama aldol-dehydration reactions using enol
ethers.21 RuCl3 was reported to be the most active catalyst using
starting cycloketones,15 but it only gave moderate yields of
benzene products in our targeted tandem reactions because of
its strong acidity to induce side reactions as described in Scheme
4 (eq 2). Similarly, other strong π-alkyne activators including
PtCl2, AuCl, AuCl3, and TpRuPPh3(CH3CN)2PF6 gave a
complicated mixture of products. On screening weakly acidic
ruthenium catalysts24 (5 mol %) including RuCl2(PPh3)3, H2-
Ru(CO)(PPh3)3, [RuCl2(p-cymene)]2, CpRu(PPh3)2Cl, and TpRu-
(PPh3)2Cl, we found that the only active CpRu(PPh3)2Cl shows
greater efficiency than RuCl3 in the aldol condensation-
dehydration as depicted in a test in Scheme 4 (eq 1). In the
cyclization of 3-en-1-yn-5-al 49 with cyclohexanone (5 equiv,
eq 2), we found that the two ruthenium species not only
catalyzed the aldol reaction-dehydration sequence, but also the
subsequent aromatization with CpRu(PPh3)2Cl (78%) being
more efficient than RuCl3 (48%) in production of R-tetralone
10. The thermal cyclization of unsaturated ketone 5 into
R-tetralone 10 proceeds with 58% yield (Table 1, entry 5), less
than the tandem reaction yield (78%) using the CpRu(PPh3)2Cl
catalyst; this result is indicative of its catalytic activity in the
final aromatization reaction. Actually, treatment of ketone 5 with
CpRu(PPh3)2Cl (5 mol %) in hot xylene (120 °C, 12 h) gave
R-tetralone 10 in 89% yield. Equation 3 shows a one-step CpRu-
(16) TpRuPPh3(CH3CN)2PF6 is active in catalytic reactions via generation of
not only ruthenium-π-alkyne intermediates15a-c but also ruthenium-vi-
nylidene species;6-8 see examples for the former case: (a) Odedra, A.;
Wu, C.-J.; Pratap, T. B.; Huang, C.-W.; Ran, Y.-F.; Liu, R.-S. J. Am. Chem.
Soc. 2005, 127, 3406. (b) Lin, M.-Y.; Maddirala, S. J.; Liu, R.-S. Org.
Lett. 2005, 7, 1745. (c) Shen, H.-C.; Liu, R.-S. Tetrahedron Lett. 2004,
45, 9245. (d) Yeh, K.-L.; Liu, B.; Lo, C.-Y.; Liu, R.-S. J. Am. Chem. Soc.
2002, 124, 6510. (e) Datta, S.; Chang, C.-L.; Yeh, K.-L.; Liu, R.-S. J. Am.
Chem. Soc. 2003, 125, 9294. (f) Madhushaw, R. J.; Lin, M.-Y.; Abu Sohel,
S. M.; Liu, R.-S. J. Am. Chem. Soc. 2004, 126, 6895.
(17) In Table 3, when AuCl3 catalyst was heated with 3,5-dien-1-ynes in dry
toluene, benzene, and 1,2-dichloroethane at 100 °C, a gold mirror was
gradually deposited from solution after several hours, and no catalytic
activity was observed in these solvents. The resulting solution was black
because polymerization of substrates occurred, presumably caused by HCl
during formation of gold mirror. We found that DMF seemed to stabilize
gold complexes, so that the metallic gold mirror formed more slowly even
at 140 °C; this dark brown DMF solution showed a cyclization activity.
Laguna17a reported that Au(III) complexes are stabilized by acid to avoid
decomposition to Au(0) in catalytic hydration of terminal alkyne with water
or alcohol. We believed that residual water in DMF reacted with AuCl3 to
give HCl in small amount, which might stabilize Au(III) species in the
course of catalytic reactions.
(18) We noticed that several catalytic reactions were performed at high
temperatures4b,13h,17 using AuX3 (X ) Cl, Br) and other Au(I) complexes;
see selected examples: (a) Casado, R.; Contel, M.; Laguna, M.; Romero,
P.; Sanz, S. J. Am. Chem. Soc. 2003, 125, 11925. (b) Asao, N.; Takahashi,
K.; Lee, S.; Kasahara, T.; Yamamoto, Y. J. Am. Chem. Soc. 2002, 124,
12650. (c) Asao, N.; Aikawa, H.; Yamamoto, Y. J. Am. Chem. Soc. 2004,
126, 7458. (d) Mizushima, E.; Sato, K.; Hayashi, T.; Tanaka, M. Angew.
Chem., Int. Ed. 2002, 41, 4563. (e) Fukuda, Y.; Utimoto, K. J. Org. Chem.
1991, 56, 3729.
(19) In these experiments, 3,5-dien-1-yne 36 and 39 (0.42 M) were heated with
concentrated HCl (10 mol %) in hot xylene (140 °C, 24 h) to give a black
solution of a complicated mixture of products in addition to polymer
formation, from which the desired benzene derivative 43 and desilylated
benzene product 44 were isolated in only 15% and 7% yields, respectively.
These data, however, cannot exclude the possibility of using a proton source
to catalyze this [1,7]-hydrogen shift.
(22) For the catalytic Mukaiyama aldol-dehydration sequence, see selected
examples: (a) Yanagisawa, A.; Coudu, R.; Arai, T. Org. Lett. 2004, 6,
4281. (b) Ishihara, K.; Kurihara, H.; Yamamoto, H. S. Synlett 1997, 597.
(23) (a) Wang, W.; Mei, Y.; Hao, L.; Wang, J. Org. Lett. 2005, 7, 601. (b)
Kreher, U. P.; Rosamilia, A. E.; Raston, C. L.; Scott, J. L.; Strauss, C. R.
Org. Lett. 2003, 5, 3107. (c) Abello´, S.; Medina, F.; Rodr´ıguez, X.; Cesteros,
Y.; Salagre, P.; Sueiras, J. E.; Tichit, D.; Coq, B. Chem. Commun. 2004,
1096.
(24) Formation of a metal enolate requires a base, whereas an acid favors
dehydration of the aldol product. We envisage that CpRu(PPh3)2Cl
undergoes dissociation in solution to form CpRu(PPh3)Cl and free PPh3,
which are actually Lewis acid and base, respectively. We use this guideline
to screen weakly acidic ruthenium catalysts containing a phosphine ligand.
A catalyst with an acid-base pair is reported to be very active in the aldol
reaction-dehydration sequence; see the examples in refs 26a-c.
(25) The X-ray crystallographic data of anthrone 50 are provided in the
Supporting Information.
(20) The procedure for the synthesis of 3,5-dien-1-yn-7-one 5 using 3-en-1-yn-
5-al and cyclohexanone is provided in the Supporting Information.
(21) Miki, K.; Nishino, F.; Ohe, K.; Uemura. J. Am. Chem. Soc. 2002, 124,
5260.
9
9664 J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006

Cyclization of 1-Substituted 3,5-Dien-1-ynes
A R T I C L E S
Scheme 4 a
Scheme 5
a Conditions: (a) cyclohexanone (5.0 equiv), aldehyde (1.0 equiv), neat,
120 °C, 16 h; (b) tetralone (2.0 equiv), aldehyde (1.0 equiv, 0.5 M), xylene,
120 °C, 96 h.
Scheme 6
Table 4. Ruthenium-Catalyzed Cyclization of 3-En-1-yn-5-als with
Cyclic Ketones
products
(yields)c
aldehyde
ketonea,b
(1) R1 ) R2 )
-(CH2)3- (49)
(2) 49
X ) (CH2)0
53 (72%)
X ) CHMe
X ) CH(tBu)
X ) CO
X ) (CH2)0
X ) CH2
54 (68%)
55 (65%)
56 (38%)
57 (78%)
16 (72%)
58 (68%)
59 (75%)
31 (82%)
60 (72%)
(3) 49
(4) 49
to elucidate the mechanism of the 1,7-hydrogen shift catalyzed
by TpRuPPh3(CH3CN)2PF6. We prepared deuterated d4-4
containing 94% deuterium contents at the two allylic methylene
protons; this species was catalytically transformed into benzene
d4-9 with deuterium contents located at the two methylene
carbons (X ) 0.94 D; Y ) 0.73 D) adjacent to the C(3)- and
C(1)-carbons, respectively. The loss of 20% deuterium at the
CY2 group of species d4-9 is attributed to the proton exchange
with residual water during the cyclization process. Because of
the isotope effect, a longer period (xylene, 120 °C, 48 h) is
required to complete this catalytic reaction relative to its
undeuterated d0-4 (24 h). For species d4-9, its benzyl CY2
deuterium content was increased to 87% (Y ) 0.87 D) in the
presence of D2O (1.0 equiv), but was decreased to 60% in the
case of H2O (1.0 equiv). The CY2 deuterium value remained at
50% for a large excess of H2O (5.0 equiv). In the absence of
catalyst, we did not observe a loss of deuterium content at the
CY2 of species d4-9 in thermal activation, even though water
was present. In the ruthenium-catalyzed transformation depicted
in eq 3, the methylene CX2 proton of d6-56 was unaffected by
the presence of H2O or D2O in various proportions, indicative
of a direct [1,7]-hydrogen shift. We observed negligible loss of
the benzyl deuterium contents of d6-56 upon heating the 3-en-
1-yn-5-al 49 with d8-1,4-cyclohexanedione in the presence of
CpRu(PPh3)2Cl. This benzyl deuterium content remained the
same even though excess water (5 equiv) was present.
(5) R1 ) R2 ) -(CH2)4- (51)
(6) 51
(7) 51
X ) CHMe
X ) (CH2)0
X ) CH2
(8) R1 ) Me, R2 ) Ph (52)
(9) 52
(10) 52
X ) CHMe
a [Ru] ) CpRu(PPh3)2Cl, conditions for entries 1, 2, and 5-10: aldehyde
(1 equiv), ketone (5 equiv), neat, 120 °C, 16 h for entries 1 and 2, 24 h for
entries 5-10. b For entries 3 and 4, aldehyde (1 equiv), ketone (2 equiv,
0.60 M), xylene, 140 °C, 36 h for entry 3 and 96 h for entry 4. c Yields are
given after separation from a silica column.
(PPh3)2Cl-catalyzed synthesis of complex anthrone 50 (76%
yield) according to our tandem catalytic protocol; its molecular
structure has been characterized by X-ray diffraction study.25
Table 4 shows the generalization of this tandem process
through annulation of 3-en-1-yn-5-als 49, 51, and 52 with
various cycloketones using CpRu(PPh3)2Cl (5 mol %) catalyst.
Entries 1-3 provide additional examples for aldehyde 49, which
reacted smoothly with cyclopentanone, 4-methylcyclohexanone,
and 4-(tert-butyl)cyclohexanone to give cyclized ketone products
53-55 in 65-72% yields. Its reaction with 1,4-cyclohexanedi-
one was less efficient, giving the corresponding benzene 56 in
38% yield (entry 4). The same cyclizations also worked
satisfactorily for 3-en-1-yn-5-al 51 bearing a cyclohexyl ring,
which provided cyclized products 57, 16, and 58 in 68-78%
yields following the same protocol (entries 5-7). Acyclic 3-en-
1-yn-5-als 52 maintained the same catalytic efficiency and
produced cyclized benzenes 59, 31, and 60 with yields exceeding
72% (entries 8-10).
The labeling experiment shown in Scheme 5 has deuterium
distributions that are different from those observed in ruthenium-
catalyzed cyclization of 3,5-dien-1-yne bearing a terminal alkyne
as depicted in Scheme 6. In the latter case, the two benzylic
protons of the product arise primarily from the remote allylic
Mechanism of the Thermal and Catalytic [1,7]-Hydrogen
Shift. Scheme 5 (eq 1) depicts our deuterium-labeling results
9
J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006 9665

A R T I C L E S
Scheme 7
Lian et al.
However, this process is not supported by deuterium-labeling
experiments for formation of d4-9 (Scheme 5, eq 1), which
confirms that external water is also the source for one of its
benzyl CY2 protons. Accordingly, we propose that the major
role of species I is to enhance the dissociation of the C(7)-H
proton (path d) to form metal-allenyl intermediate J, and to
give 4,6-dien-1-allene species K through subsequent reproto-
nation and reductive elimination. Species K ultimately gives
the observed benzene product via 6-π-electrocyclization30 and
a [1,3]-hydrogen shift.
The proposed mechanism in Scheme 7 emphasizes the
importance of the methylene C(7)-H acidity of 3,5-dien-1-ynes,
and this mechanistic hypothesis is verified by the rate-accelera-
tion effects of added 2,6-lutidine in both thermal and ruthenium-
catalyzed cyclization of 3,5-dien-1-yne 4 according to the results
in a separate experiment.30 This phenomenon suggests that
dissociation of a proton of intermediate H (or I) is the rate-
determining step in the metal-catalyzed cyclization. It also
suggests that base-catalyzed cyclization of 3,5-dien-1-ynes is
likely to occur.31
protons of starting 3,5-dien-1-ynes. These deuterium-label-
ing results reveal a methylene transfer process, consistent with
the example shown in Scheme 2 (eq 2). Such a cyclization is
thought to proceed through a 6-π-electrocyclization of initial
metal-vinylidene intermediate with key steps shown in Scheme
6.6
Scheme 7 rationalizes the mechanistic nature of the thermal
aromatization of 3,5-dien-1-ynes bearing internal alkynes. As
species 5, 16, and 22 are more intrinsically active than other
analogues in thermal cyclization, we proposed that their alkynyl
C(1)-TMS carbon is reactive toward electrophiles26 and that their
C(6)-ketone group increases the adjacent C-H acidity to
facilitate a “protonic” hydrogen shift as represented by species
G. Such a “protonic” hydrogen shift is also favored by the
presence of an electron-withdrawing C(1)-phenyl and C(6)-
carbonyl group as reflected by the substituent-activity trends
in Tables 1-3. This model also rationalizes the superior
cyclization efficiencies of dienynes 4, 13, and 19 bearing a five-
membered carbocyclic ring relative to their six-membered ring
analogues 3, 14, and 20 as the C-H acidity is higher for a
smaller ring because of a greater s-character.27 The [1,5]-
hydrogen shift of 3-en-1-ynes represented by intermediate E
(Scheme 2, eq 1) has been characterized as a “hydridic”
hydrogen shift according to our recent findings.28
Conclusions
Before this study, little information was available on the [1,7]-
hydrogen shifts of 3,5-dien-1-ynes. In assessing the feasibility
in thermal and metal-catalyzed cyclization of 6,6-disubstituted
3,5-dien-1-ynes via a 1,7-hydrogen shift, we first studied the
thermal efficiency of model molecules via alternation of the
C(1) and C(9) substituents of substrates; we concluded that the
observed 1,7-hydrogen shift is nearly a “protonic” hydrogen
shift. This structure-activity relationship indicates that π-alkyne
activators can catalyze suitably functionalized 3,5-dien-1-ynes
efficiently although their thermal yields were generally low. We
prepared various 6,6-disubstituted 3,5-dien-1-ynes containing
either a C(1)-phenyl, a TMS, or a C(6)-carbonyl group, and we
found their thermal cyclizations to be greatly enhanced with
RuCl3, PtCl2, and TpRuPPh3(CH3CN)2PF6 catalysts to confirm
our hypothesis: the C(7)-H acidity of 3,5-dien-1-ynes deter-
mines the cyclization efficiency. To highlight the synthetic utility
of this cyclization, we developed a tandem aldol condensation-
dehydration and aromatization catalysis between cycloalkanones
and special 3-en-1-yn-5-als using weakly acidic CpRu(PPh3)2-
Cl catalyst; the yields generally exceeded 65% yield. Deuterium-
labeling experiments show two different behaviors for metal-
Scheme 7 also depicts a plausible mechanism for metal
catalysts to accelerate the 1,7-hydrogen shift. These metal
complexes may have two roles: (1) an alternation of the alkyne
sp-character toward sp2-character13c as in intermediate H, and
(2) the enhancement of the C(7)-H hydrogen acidity. The
π-alkyne bonding is also characterized by the resonance
structure I, which is thought to be responsible for several
+ 14a,b
reactions catalyzed by PtCl2,13 AuPPh3
,
AuCl3,13 and
AuCl.14 The polar nature (I) of this π-alkyne moiety is supported
by theoretic calculations.13c This π-alkyne bonding may facilitate
a 1,7-hydrogen shift of 3,5-dien-1-ynes because of a closer
C(1)-C(7) distance; this actually mimics the 1,7-hydrogen shift
of 1,3,5-trienes. Such a direct [1,7]-shift pathway (path c) is
compatible with formation of d6-56, of which the benzyl CY2
deuterium contents arise exclusively from d8-1,4-cyclohexadi-
enone (Scheme 5, eq 2) even though water was present.
(29) For 6-π-electrocyclization of 3,5-dien-1-allenes, see selected examples: (a)
Bross, H.; Schneider, R.; Hopf, H. Tetrahedron 1979, 2129. (b) Reischl,
W.; Okamura, W. H. J. Am. Chem. Soc. 1982, 104, 6115.
(30) The cyclization yields of 3,5-dien-1-yne 4 to benzene derivative 9 in hot
xylene (120 °C) were obtained in 31% (36 h) and 83% (24 h), respec-
tively, in thermal and TpRuPPh3(CH3CN)2PF6-catalyzed reactions (see
Table 1, entry 4). In the presence of 2,6-lutidine (5 mol %), the yields of
these two processes improved to 51% (12 h) and 93% (12 h) with shorter
reaction periods. We thank one reviewer for his very helpful suggestion.
(31) The observation that 2,6-lutidine accelerates the thermal cyclization of 3,5-
dien-1-yne 4 suggests the feasibility for a cyclization of 3,5-dien-1-ynes
using organic bases as depicted below. We will conduct related reactions
toward this direction.
(26) (a) Fleming, I. In ComprehensiVe Organic Synthesis: Addition to C-X
π-Bonds, Part 1; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
1991; Vol. 2, Chapter 2.2, p 575. (b) Panek, J. S. In ComprehensiVe Organic
Synthesis: Addition to C-X π-Bonds, Part 1; Trost, B. M., Fleming, I.,
Eds.; Pergamon Press: Oxford, 1991; Vol. 2, Chapter 2.2, p 580. (c)
Langkopf, E.; Schinzer, D. Chem. ReV. 1995, 95, 1375.
(27) (a) Bordwell, F. G.; Matthews, W. S. J. Am. Chem. Soc. 1974, 96, 1214.
(b) Streitwieser, A., Jr.; Reuben, D. M. E. J. Am. Chem. Soc. 1971, 93,
1974.
(28) Datta, S.; Odedra, A.; Liu, R.-S., unpublished results.
9
9666 J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006

Cyclization of 1-Substituted 3,5-Dien-1-ynes
A R T I C L E S
catalyzed aromatization of 3,5-dien-1-ynes. Formation of d4-
56 arises from a direct [1,7]-hydrogen shift mechanism, whereas
d4-9 proceeds through a proton dissociation and reprotonation
process.
Supporting Information Available: Experimental procedures
for the synthesis of 3,5-dien-1-yne substrates and catalytic
operations, NMR spectra, spectral data of compounds 1-60,
and X-ray structural data of anthrone 50. This material is
available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We wish to thank the National Science
Council, Taiwan, for supporting this work.
JA061203B
9
J. AM. CHEM. SOC. VOL. 128, NO. 30, 2006 9667