1290
F. Yang et al. / Tetrahedron Letters 43 (2002) 1289–1293
1
After full characterization by IR, MS, H, 13C NMR
and EA and X-ray crystallographic analysis (Fig. 1),6
this proved to be the homocoupling product 3a. That
means benzaldehyde does not take part in the reaction.
A series of propargyl esters as substrates have been
tested for this new intermolecular homocoupling reac-
tion. The results are shown in Table 1.
taining no substituent on the terminal triple bond gave
moderate yields (entries 12–14) with concomitant for-
mation of 1-allenyl-5-alkyne, but alkynes containing a
Me3Si group on the terminal triple bond carbon gave
much better results (entry 7). For most 1-monosubsti-
tuted propargyl carbonates a mixture of meso and dl
isomers in various ratios were obtained. The ratio can
be roughly estimated based on their 1H NMR (for
compounds 3g, 3h, 3j, 3l, 3m, 3n, 3o) or 13C NMR (for
compound 3f), but we have not determined which is the
meso or dl isomer except for 3j;7 (c) ethyl 3-ethoxycar-
bonyl-1,1-dimethylpropargyl carbonate 1p was trans-
formed to the b-elimination product without any
coupling product being isolated (entry 16); (d) the
coupling reaction of the 3-phenylpropargyl carbonate,
in which there is no 1-substituent, did not occur at all
(entry 17). In that case most starting material was
recovered even though the reaction period was
extended to 24 h. When a more active propargyl mesy-
late was used, only propargyl chloride was isolated.
The results in Table 1 using a variety of propargyl
esters as substrates show that: (a) 1,1-disubstituted
propargyl esters produce the homocoupled compounds
smoothly with high yields (entries 1–5); (b) the reac-
tions with 1-mono-aryl propargyl carbonates also result
in homocoupled products in good yields (entries 6, 7,
11) but the coupling reactions of acetates (entry 10) or
1-monoalkyl propargyl carbonates (entries 8 and 9)
usually proceed slowly with moderate yields and a
considerable amount of recovered starting materials.
There was always a certain amount of 1-allenyl-5-
alkyne in the product of the coupling reactions. The
coupling reaction of 1-aryl propargyl carbonates con-
In order to explore the mechanism we carried out the
following experiments. (1) The coupling reactions of
compounds 1a and 1k proceeded smoothly with com-
parable yield using 0.5:1 (mole ratio) of Ti(OiPr)2Cl2:
substrate 1a (or 1k) instead of a 1:1 ratio. That means
only one mole of Ti(OiPr)2Cl2 is needed for the cou-
pling of two moles of propargyl carbonate. (2) A cross
coupling reaction of 1,1-dimethyl-3-phenylpropargyl
carbonate 1d: 3-phenylpropargyl carbonate 1q:
Ti(OiPr)2Cl2 (1:1:1 molar ratio) was carried out but no
cross coupling product was detected except the homo-
coupling product of 3d. This excludes a mechanism
whereby a monoallenyltitanium intermediate, which
could be produced in our system,8 attacks the elec-
trophilic C1 of a second 1d molecule to form the
homocoupling products since the primary 3-phenyl-
propargyl carbonate 1q should be more active to SN2
attack than the tertiary 1,1-dimethyl-3-phenylpropargyl
carbonate 1d. (3) We have carried out an experiment to
verify the role of Mg as follows: Ti(OiPr)2Cl2 was
treated first with Mg powder in dry ether under N2 at
0°C, and then the supernatant solution was transferred
under N2 to a propargyl carbonate solution 1d in ether
without Mg present at 0°C, but only a small amount of
coupling product was isolated. When the supernatant
solution was transferred to a propargyl carbonate 1d
solution in ether with Mg present in the same manner,
the homocoupling product 3d was formed in normal
yield. This means that the monoallenyltitanium inter-
mediates 2 do not couple each other, but they couple
with each other in the presence of Mg.
Table 1. Homocoupling of propargyl esters mediated by
Ti(OiPr)2Cl2/Mga
Entry
Propargyl
ester
Products
Yield (%)b
Ratioc (3/4)
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
1q
3a
3b
3c
3d
89d,l
94l
76
81
70
3e, 4e
3f, 4f
3g, 4g
3h
3i, 4i
3j, 4j
3k, 4k
3l, 4l
3m, 4m
3n, 4n
3o
52:48
54:46
87:1
82
89d
45e
42f
53g
83
9
70:30
71:29k
81:19k
50:50
52:48
55:45
10
11
12
13
14
15
16
17
48
46.6
49
55h
43i
Littlej
a Reaction conditions: propargyl ester:Ti(OiPr)2Cl2:Mg=1:1.2:2
(mmol); 0°C for 1 h; then room temperature for 3–4 h.
b Isolated yield after purification on silica gel column chromatogra-
phy (petroleum ether).
c The ratio was determined by the isolated yields of 3 and 4 after
purification on column chromatography.
d Homocoupling product crystal was obtained by recrystalization in
CH3CN and used for X-ray structure analysis.
e 30% starting material was recovered.
f 28% starting material was recovered.
A radical coupling reaction mechanism is tentatively
suggested as shown in Scheme 3, based on the fact that
the more substituents at C1 the propargyl carbonates
possess, the higher the yield of the coupling product.
The oxidative insertion of low-valent Ti(OiPr)2, formed
from Ti(OiPr)2Cl2 and Mg in ether at 0°C, to the triple
bond of the propargyl carbonate produces an allenylti-
tanium intermediate. A propargyl radical could then be
formed on the Mg surface through a SET process and
two propargyl radicals couple with each other to form
the homocoupling product. In a radical detection test
g 29% starting material was recovered.
h Reaction for 24 h at room temperature and 38% starting material
was recovered.
i Only 1,1-dimethyl-3-ethoxycarbonylallene was isolated and 45%
starting material was recovered.
j Reaction for 24 h at room temperature and most starting material
did not change.
k The ratio was estimated by the 1H NMR of the product mixture,
since the relevant 1-allenyl-5-alkyne was unstable and difficult to be
separated in pure form.
l It contains one sixth 1-(1%-cyclopentenyl)-2-phenylethyne.