Communication
of metal complexes. Then, extrusion of another molecule of ni-
trogen from IV, V, or VI would generate olefin VII. Secondly,
the Bamford-Stevens process would also effect selective deto-
sylation to afford intermediate II (path b), in which one diazo is
formed and the other tosylhydrazone remains. Direct thermoly-
sis of II would then generate carbene intermediate VIII
(Scheme 2, b1) and subsequent intramolecular NÀH insertion
would yield heterocycle X. The presence of a metal catalyst
would afford metal carbene IX, which could also form X by NÀ
H insertion (Scheme 1, b2).
Table 1. Optimizations of the reaction conditions.[a]
Entry
Catalyst
Base
2a [%][b] 3a [%][b]
1
2
3
4
5
6
–
–
–
–
–
–
LiOtBu 30
K2CO3
Li2CO3
Cs2CO3 75
KOtBu 78
15
10
–
56
–
5
0
Moreover, there is another competitive process, namely the
classical base-induced Shapiro[12] or Bamford–Stevens reaction,
which would yield alkene III as dominant product of the elimi-
nation of the hydrogen atom from the hydrazone and the
acidic proton a to the hydrazone carbon (Scheme 2, path c).
Based on the above analysis, we rationalized that the intra-
molecular coupling of bis-N-tosylhydrazones would furnish
two major types of products, namely the olefin VII and hetero-
cycle X, either in the absence of or in the presence of a metal
catalyst (Scheme 2). Thus, the focus of this investigation was to
establish an effective approach that would allow selective and
controllable decomposition of N-tosylhydrazones. Inspired by
our previous studies,[13] especially those on gold-catalyzed in-
termolecular diazo cross-coupling,[13c,d] herein we describe effi-
cient and practical procedures to deliver cyclic olefins and het-
erocycles with excellent selectivity, with or without the pres-
ence of a gold complex (Scheme 2B).
NaH
58
0
7
8
9
[IPrAuNTf2]/NaBArF
[Ph3PAuCl]/NaBArF
[Et3PAuCl]/NaBArF
LiOtBu 38
LiOtBu 38
LiOtBu 45
LiOtBu 44
LiOtBu 40
LiOtBu 45
LiOtBu 30
LiOtBu 48
LiOtBu 35
LiOtBu
LiOtBu
KOtBu
LiOtBu
5
35
7
3
7
6
31
9
10
11
12
13
14
15
16
17
18
19
[Cy3PAuCl]/NaBArF
[(PhO)3PAuCl]/NaBArF
[(2,4-tBu2C6H3O)3PAuCl]/NaBArF
[(C6F5)3PAuCl]/NaBArF
[XantPhos(AuNTf2)2]/NaBArF
[XPhosAuCl]/NaBArF
[tBuXPhosAuCl]/NaBArF
[tBuXPhosAuCl]
[tBuXPhosAuCl]/NaBArF
[tBuXPhosAuCl]/AgSbF6
10
85
23
45
55
3
0
3
9
[a] Reaction conditions: To a solution of 1a (0.2 mmol) in DCE (5 mL) was
added base (3 equiv) in the absence/presence of gold complex
(2.5 mol%) and NaBArF (2.5 mol%) at RT. Then the mixture was heated to
reflux and stirred for 12 h; [b] yields of isolated product. DCE=1,2-di-
chloroethane.
At the outset, bis-N-tosylhydrazones 1a was used as the
model substrate to establish the optimal reaction conditions
(Table 1). First, treatment of 1a with LiOtBu in 1,2-dichloro-
ethane at reflux for 12 h afforded diazepine 2a in 30% yield,
alongside a 15% yield of cyclopentene 3a (Table 1, entry 1).
The use of K2CO3 gave 2a and of 3a in 56% and 10% yield, re-
spectively (Table 1, entry 2). When Cs2CO3 was employed, 75%
yield of 2a and 5% yield of 3a were isolated (Table 1, entry 4).
Further using KOtBu produced 2a as single product in 78%
yield (Table 1, entry 5), whereas NaH furnished 2a in moderate
yield (entry 6). Next, various gold catalysts were evaluated
combined with NaBArF (BArF =[B{3,5-(CF3)2C6H3}4]À) and LiOtBu.
A gold–N-heterocyclic carbene catalyst afforded 2a as the
major isomer, albeit in lower yield (Table 1, entry 7) and phos-
phine or phosphite ligand-derived gold complexes gave poor
reactivities and selectivities (Table 1, entries 9 to 15). Gratifying-
ly, the combination of tBuXphosAuCl/NaBArF and LiOtBu deliv-
ered 3a as the major isomer in 85% yield (Table 1, entry 16).
Consistent with our former observations on gold-catalyzed
diazo coupling,[13] the reaction was sluggish in the absence of
NaBArF (Table 1, entry 17). Moreover, using KOtBu as base and
replacing NaBArF with AgSbF6 resulted in low conversion
(Table 1, entries 18 and 19). The different outcomes with or
without the gold catalyst [tBuXPhosAuCl] (tBuXPhos=2-di-tert-
butylphosphino-2’,4’,6’-triisopropylbiphenyl) were probably
a resultant of the reactions following different reaction path-
ways; path a of Scheme 2 is preferred in the presence of [tBuX-
phosAuCl], whereas path b is dominant without the gold cata-
lyst. Moreover, Wang and co-workers reported a seminal rhodi-
um-catalyzed approach towards the olefin formation.[11]
Based on the above optimization, we next investigated the
substrate scope under transition-metal-free conditions (Table 1,
entry 5; condition A) and with gold catalyst(Table 1, entry 16;
condition B). For most cases, the desired heterocycles and ole-
fins were obtained in moderate to high yields (Table 2). For the
unsymmetric diaryl bis-tosylhydrazone 1e, heterocycle 2e was
obtained as the major isomer under metal-free condition (75%
yield; Table 2, entry 5).[14] For bis-tosylhydrazone 1 f, removal of
the aldehyde-derived tosylhydrazone has been prioritized, and
the corresponding product was obtained in 75% yield (Table 2,
entry 6). However, for the aryl alkyl bis-tosylhydrazone 1g, the
tosylhydrazone adjacent to the inner aryl group was selectively
removed and produced 2g as single isomer (68% yield;
Table 2, entry 7). For the above-mentioned unsymmetric bis-N-
tosylhydrazones, only one regioisomer was isolated, which can
probably be attributed to the faster decomposition of one to-
sylhydrazone in relation to the other moiety under the reaction
conditions. For 1,5-bis-tosylhydrazones, the corresponding het-
erocycles (2h–j) were isolated in high yields (Table 2, entries 8
to 10; trans/cis=2:1 or 1:1). The 1,6-bis-tosylhydrazone was
also applicable and the corresponding products 2m and 3m
were isolated in high yields. Notably, the 12-membered hetero-
cycle 2n and olefin 3n were obtained in 76% and 48% yields,
respectively (Table 2, entry 14). The similar phenomenon was
observed for bis-tosylhydrazone 1o. The 14-membered hetero-
cycle 2o and 12-membered olefin 3o were obtained in moder-
ate yields (Table 2, entry 15). These results indicate that this is
a practical and efficient approach to preparing large-ring het-
Chem. Eur. J. 2015, 21, 12871 – 12875
12872
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim