Phosphine-catalyzed stereoselective dimerizations of ketenes

Article abstract of DOI:10.1016/j.tet.2020.131838

Full details of optimisation studies of the phosphine-catalyzed ketene homodimerization reaction and the detailed development of an asymmetric variant are discussed. Studies towards the development of a phosphine-catalyzed ketene heterodimerization reaction are revealed. A discussion of possible reaction mechanisms for the dimerization reactions, supported by spectroscopic analysis of intermediates and trapping experiments, is also presented.

Full text of DOI:10.1016/j.tet.2020.131838

Tetrahedron 78 (2021) 131838
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Phosphine-catalyzed stereoselective dimerizations of ketenes
Ahmad A. Ibrahim a, Pei-Hsun Wei a, Gero D. Harzmann a, Divya Nalla a,
,
*
Mukulesh Mondal a, Kraig A. Wheeler b, Nessan J. Kerrigan c
a Department of Chemistry, Oakland University, 2200 N. Squirrel Road, Rochester, MI 48309-4477, USA
b Department of Chemistry, Whitworth University, 300 W. Hawthorne Road, Spokane, WA 99251, USA
c School of Chemical Sciences, Dublin City University, Glasnevin, Dublin 9, Ireland
a r t i c l e i n f o
a b s t r a c t
Article history:
Full details of optimisation studies of the phosphine-catalyzed ketene homodimerization reaction and
the detailed development of an asymmetric variant are discussed. Studies towards the development of a
phosphine-catalyzed ketene heterodimerization reaction are revealed. A discussion of possible reaction
mechanisms for the dimerization reactions, supported by spectroscopic analysis of intermediates and
trapping experiments, is also presented.
Received 30 September 2020
Received in revised form
23 November 2020
Accepted 28 November 2020
Available online 3 December 2020
© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
Keywords:
Phosphine
Catalysis
Lactone
Asymmetric synthesis
Enantioselective
1. Introduction
introduced by our group [3,9]. Recently, we determined that
phosphines (trialkyl- and dialkyaryl-) were excellent catalysts for
b
-Lactones have historically been important molecules which
promoting a range of formal [2 þ 2] cycloadditions of ketenes,
including with aldehydes to give highly substituted trans- -lac-
tones, with N-tosylimines to give trans- -lactams, and catalytic
homodimerization of disubstituted ketenes to provide access to
ketene homodimer -lactones [3,10,11]. We have noted that phos-
phines often provide superior substrate scope compared to N-
heterocyclic carbenes and azaferrocene catalysts, especially for
difficult substrate classes, such as chiral aldehydes [12]. Further-
more, they often provide complementary access to a given isomer
have seen extensive use as intermediates in the synthesis of com-
plex molecules, and moreover function as integral structural fea-
tures of many biologically active molecules [1,2]. Important routes
b
b
to
b
-lactones include nucleophile catalyzed-[2 þ 2] cycloaddition of
b
ketenes with aldehydes, homodimerization of ketenes, and heter-
odimerization of ketenes [3e6]. The alkaloid-catalyzed enantiose-
lective reaction of methanol with disubstituted ketenes, developed
by Pracejus in the 1960s was the first truly high level organo-
catalytic process [7]. However, in the following years few alkaloid-
catalyzed reactions of ketenes were developed, with Wynberg’s
of product, e.g. the trans-isomer of b-lactam from N-tosylimines,
relative to the cis-isomer provided by azaferrocene catalysis [10].
In the 1960s, Elam and, shortly after, Bentrude showed that the
homodimerization of dimethylketene could be promoted by tri-
alkylphosphite nucleophilic catalysts [4b-4d]. In 1996, and more
comprehensively in 2003, Calter showed that a nucleophilic cata-
lyst (TMS-quinine or TMS-quinidine) could catalyze the homo-
dimerization of a number of in situ-generated alkyl substituted
asymmetric synthesis of
b-lactones from ketene and chloro-
substituted aldehydes and ketones (e.g. chloral) being the most
notable achievement [8]. Over the last two decades, the develop-
ment of catalytic ketene reactions has exploded, and a number of
new classes of catalyst have emerged [9]. The most successful of
these new catalyst classes for disubstituted ketenes have been
azaferrocenes introduced by Fu’s group, N-heterocyclic carbenes
reported by the groups of Ye and Smith, and phosphine catalysts
aldoketenes with high enantioselectivity [5]. Ketene dimer b-lac-
tones have been used extensively in the synthesis of poly-
propionate natural products by Calter and co-workers, and by
Romo’s group in the synthesis of fatty acid synthase inhibitors and
tetronic acids [5,13]. In 2009 our group reported that a tri-
alkylphosphine catalyst could provide a versatile system for
* Corresponding author.
E-mail address: nessan.kerrigan@dcu.ie (N.J. Kerrigan).
https://doi.org/10.1016/j.tet.2020.131838
0040-4020/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
promoting the homodimerization of disubstituted ketenes [11]. We
also demonstrated that Binaphane and, later, Josiphos were excel-
lent chiral catalysts for promoting the enantioselective homo-
dimerization of disubstituted ketenes [11]. Around the same time
Ye’s group published their work on the N-heterocyclic carbene-
catalyzed asymmetric ketene homodimerization reaction [14].
However the latter method was found to be unsuitable for the
homodimerization of ortho-substituted alkylarylketenes and dia-
lkylketenes, whereas the Josiphos system worked well for those
substrates, in addition to the more commonly explored alkylar-
ylketenes (methylphenylketene, ethylphenylketene, etc.) [11c]. In
this paper, we discuss the extensive optimisation involved in the
development of the phosphine-catalyzed enantioselective homo-
dimerization process, the scope and limitations of the phosphine-
catalyzed homodimerizations (both non-enantioselective and
enantioselective variants), phosphine-catalyzed ketene hetero-
dimerization, in situ ketene generation studies, and mechanistic
studies of the phosphine-catalyzed dimerization reactions.
promoting the desired homodimerization reaction (entries 3e5).
Methylphenylketene homodimer 2a was formed, along with trimer
3a, with complete conversion of ketene to product when 20 mol% of
the tri-n-butylphosphine catalyst was employed. Careful optimi-
sation revealed that LiI as an additive effectively suppressed trimer
formation (entry 6). We propose that the main effect of the LiI
additive is to stabilize enolate intermediate 2aa favouring cyclisa-
tion to dimer rather than further reaction with ketene to give
trimer (Scheme 1).
Other Lewis acids were evaluated, e.g. LiClO4, LiOTf, TiCl4, AlCl3,
and B(c-Hex)2Cl, but all afforded lower yield of homodimer 2a than
when LiI was employed. The amount of lithium iodide used was
reduced to 0.3 equiv without any significant increase in trimer
formation and importantly with an increase in the yield of homo-
dimer being observed (entry 12). At the same time the tri-n-
butylphosphine catalyst loading was dropped to 10 mol% (0.1
equiv) without loss of reaction efficiency. Ultimately the best iso-
lated yield for 2a of 82% was obtained after purification through a
plug column of neutral silica, rather than through a standard sized
column (entry 14). By way of comparison, triphenylphosphine was
found to perform moderately well (47%) under identical reaction
conditions (entry 13), but clearly the superior nucleophilicity of the
trialkylphosphine catalyst is needed for optimal ketene
homodimerization.
2. Results and discussion
2.1. Optimisation of the phosphine-catalyzed homodimerization
reaction (non-enantioselective variant)
Optimisation of the catalytic homodimerization of disubstituted
ketenes initially revolved around methylphenylketene as the test
substrate (Table 1). A number of nucleophilic catalysts were
investigated for their ability to promote the desired homodimeri-
2.2. Substrate scope
We then proceeded to evaluate the scope of the tri-n-butyl-
phosphine-catalyzed ketene homodimerization (Table 2). In many
cases (for R2 s Me) it was not necessary to use lithium iodide as an
additive in order to minimise trimer formation. This was due to the
lower reactivity of enolate intermediates 2aa in some cases
(because of resonance stabilisation, e.g. R1 ¼ R2 ¼ Ph for entries 6
and 7, or steric bulk of R2 alkyl substituent e.g. for R2 ¼ Et, n-Bu, i-
Bu), compared to those derived from methylphenylketene
(R2 ¼ Me).
zation in forming ketene homodimer b-lactone 2a. Interestingly
alkaloid catalysts were found to be ineffective with only ca. 5e10%
conversion to homodimer being observed (e.g. Table 1, entry 1).
Phosphites, such as triethylphosphite, were also found to perform
poorly (e.g. entry 2). On the other hand, the trialkylphosphine, tri-
n-butylphosphine, was found to be an excellent catalyst for
Table 1
Effect of reaction conditionsa on yield and selectivity of methylphenylketene
homodimerization.
Entry Catalyst (Equiv)
LiI Equiv Solvent
Yieldb (Conv) 2a: 3ac
1
TMS-quinine (0.1) 0.3 CH2Cl2/Et2O <10
e
2
P(OEt)3 (0.1)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.2)
PBu3 (0.1)
e
0.3
0
0
CH2Cl2/Et2O 10
e
3
PhCH3
PhCH3
CH2Cl2
Et2O
- (>99)
- (>99)
- (>99)
45
1.4:1
2.5:1
4:1
>16:1
>16:1
4:1
>16:1
10:1
e
4d
5
0
6
7
8
9
10
11
12
13
14e
1.0
1.0
1.0
1.0
0.2
0.3
0.3
0.3
0.3
PhCH3/Et2O 41
THF 50
CH2Cl2/Et2O 50
CH2Cl2/Et2O - (>99)
CH2Cl2/Et2O
0
PBu3 (0.1)
PPh3 (0.1)
PBu3 (0.1)
CH2Cl2/Et2O 60
CH2Cl2/Et2O 47
CH2Cl2/Et2O 82
16:1
16:1
16:1
a
Reactions carried out at [0.5 M] of ketoketene in solvent.
b
c
Isolated yield of Z-dimer after flash column chromatography on neutral silica.
Determined by 1H NMR analysis of Me signals in 1H NMR of crude product
mixture.
d
Reaction conducted at [0.1 M] of ketoketene in solvent.
Purified by passing through a plug column of neutral silica (iatrobeads).
e
Scheme 1. Role of lithium salt additive in ketene homodimerization and trimer
formation.
2

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
Table 2
Yields and diastereoselectivities for the formation of 2a-4o.
2
Entry
PR3
R1
R
Yielda
Z:Eb
Product
1
PBu3
PBu3
PBu3
PBu3
PBu3
PBu3
PMe3
PBu3
PBu3
PBu3
PBu3
PBu3
PBu3
PBu3
PMe3
PBu3
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Me
Et
n-Bu
i-Bu
Bn
82
>16:1
37:1
nd
28:1
nd
2a
2b
2c
2d
2e
2f
2
84/98c
99
3c
4c
84
5c
(64)d
67
6c
Ph
Ph
e
7c
61/78e
99
e
2f
Fig. 1. Catalytic homodimerization of isobutylphenylketene at different concentrations
of PBu3. Standard conditions refers to a starting concentration of 0.1 M for iso-
butylphenylketene in CH2Cl2 and 0.01 M (0.1 equiv.) for PBu3 in CH2Cl2.
8c
2-ClC6H4
2-tolyl
4-tolyl
4-tolyl
6-MeO-2-Naphthyl
3-Thienyl
Me
Me
Me
Me
Et
Me
Et
Me
Ph
Et
nd
nd
39:1
nd
>16:1
20:1
e
2g
2h
2i
2j
2k
2l
2m
2n
4o
9c
97
90
99
62
10
11c
12
13
14c,f
15c,e
16h
regioisomer, rather than the b-lactone regioisomer, as the major
78
product. We attribute that result to a change in the electronics
arising from dialkyl-substitution, causing the putative enolate in-
termediate to react through the C-atom rather than through the O-
atom (Scheme 2). When a phenyl substituent is present, electron
delocalization, by resonance, can help to stabilize the enolate in-
termediate, rendering the C end of the enolate less reactive, with
the enolate displaying a tendency to react through the O-atom
(which is also a less sterically hindered nucleophilic site). However,
73 (94)g
60
c-Pentyl
c-Hexyl
3:1
e
87
a
Isolated yield after passing through a plug column of neutral silica (iatrobeads).
Purity ꢀ95% in all cases with the exception of 2n (90% purity) as determined by
GCMS and HPLC analysis.
b
Determined by 1H NMR analysis of Me signals in 1H NMR of crude reaction
mixture or by HPLC analysis on an AD column.
c
No LiI was used.
Conversion as determined by GC-MS analysis.
when dialkyl-substitution is present at the a-position of the eno-
d
e
late, no such electron delocalization at the C-atom end of the
enolate is possible, and so the enolate displays a preference for
reacting through the C-atom of the enolate to give the 1,3-
cyclobutanedione regioisomer.
0.2 equivalents of PMe3 was used.
f
Reaction carried out in THF at ꢁ25 ꢂC.
g
Isolated yield for b-ketoalcohol derivative.
h
1,3-Cyclobutanedione regioisomer.
2.3. Kinetics studies
As a result, a preference for intramolecular O-enolate ring
closure to give dimer rather than trimer-forming reaction of 2aa
was observed (Scheme 1). Variation of the alkyl substituent was
tolerated with high yields obtained for R1 ¼ Me, Et, n-Bu and i-Bu
(entries 1e4). Although lower efficiency was noted for R2 ¼ Bn
(entry 5), the very stable ketene, diphenylketene, could be sub-
jected to homodimerization, albeit with longer reaction times
involved (up to 3 days for entry 7). It was notable that ortho-
substituted aromatics for R1 could be accommodated (entries 8 and
9) as well as heteroaromatic substituents (e.g. thiophene, entry 13),
without any loss of reaction efficiency. Dialkylketenes such as
dimethylketene and ethyl-c-hexylketene also performed well un-
der the optimised conditions. Interestingly, homodimerization of
ethyl-c-hexylketene led to the formation of a 1,3-cyclobutanedione
In order to measure the rate of product formed per unit time
with differing reaction conditions, kinetic experiments were car-
ried out by subjecting isobutylphenylketene to homodimerization
catalyzed by tri-n-butylphosphine. These reactions were carried
out by varying the concentration of the catalyst and concentration
of ketene in different experiments. The rate of 2d formation was
monitored by GC-MS every 30, 70,115 and 200 min. Fig.1 shows the
results for the formation of ketene homodimer 2d at various con-
centrations of tri-n-butylphosphine (0.01e0.2 equiv) with constant
Fig. 2. Catalytic homodimerization of isobutylphenylketene at different concentrations
Scheme 2. Regioselectivity in phosphine-catalyzed ketene homodimerization.
of ketene, with constant equivalents of PBu3 (0.1 equiv) in CH2Cl2.
3

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
Table 3
Catalyst
methylphenylketene.
initial concentration of isobutylphenylketene (0.1 M) in CH2Cl2.
Fig. 1 shows that the rate of 2d formation increases with the in-
crease in concentration of the catalyst (PBu3), with 0.2 equiv
(0.02 M) of PBu3 being optimal (initial concentration of PBu3/initial
concentration under standard conditions ¼ 2, Fig. 1).
screening
for
catalytic
asymmetric
homodimerization
of
Fig. 2 indicates the rate of product formed at different concen-
trations of ketene in CH2Cl2 while at constant equivalents of PBu3
(0.1 equiv). The optimal concentration of ketene for best reaction
rate was found to be 0.25 M (initial concentration of ketene/initial
concentration under standard conditions ¼ 2.5, Fig. 2).
The kinetic data showed that the rate of the homodimerization
reaction was dependent upon the catalyst concentration with 0.2
equiv (0.02 M) being optimal, although 0.1 equiv (0.01 M) and even
0.05 equiv (0.005 M) of PBu3 also worked effectively. The rate of the
reaction was also found to be dependent upon the concentration of
the ketene with 0.25 M proving optimal, although 0.15 M and 0.5 M
also gave good results in terms of rate.
2.4. Optimisation of the phosphine-catalyzed enantioselective
homodimerization reaction
Having identified tri-n-butylphosphine as the optimal nucleo-
philic catalyst, we then proceeded to investigate a range of chiral
phosphines in order to identify a catalyst for enantioselective
ketene homodimerization. We were initially attracted to phos-
phines exhibiting atropisomerism, such as phosphepines, MOP
Table 4
Scope of Josiphos-catalyzed asymmetric homodimerization of ketoketenes.
Entry
Phosphine
Conc (M)
2a:3aa
% Conv (% Yield)b,c
% eed
1
2
3
5a
5b
5c
5d
5d
6a
6b
7a
7a
7b
8a
8b
9
10
11
12a
12a
12a
12b
12c
12d
12e
12e
12e
12e
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.25
0.125
0.125
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
4:1
nd
(30)
(15)
(20)
(70)
(78)
30
(63)
(60)
nd
<5
20
3
nd
9
14
26
26
33
4
0
9
10
0
8
0
0
9
13
73
64
80
78
2
4
5e
6
Entry
R1
R2
Product
% Yielda
% eeb
Config
7
1c
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
2-ClC6H4
2-tolyl
4-ClC6H4
4-tolyl
3-thienyl
3-thienyl
c-hexyl
c-hexyl
Me
Me
Me
Et
Et
Et
Bn
n-Bu
i-Bu
Me
Me
Et
Et
Et
Et
Me
Et
(¡)-2a
(þ)-2a
(¡)-2b
(¡)-2b
(þ)-2b
-
65
74
93
65
86
(10)h
90
0
45
80
81
99
68
61
97i
77i
75
94
90
86
80
90
nd
89
e
S
R
S
S
R
-
8
9f
2d
3e
10
11
12
13
14
15
16g
17e
18h
19
20
21
22
23
24
25i
4f
5g
6e
7g
(þ)-2c
R
-
30
10
8
-
9c
(þ)-2g
(þ)-2h
(¡)-2p
(¡)-2j
(¡)-2l
(¡)-2l
4q
96
94
94
85
46
78
na
na
na
S
R
S
S
S
S
>99
12
>99
>99
>99
0
>99
>99
>99
>99 (74)
10d
11e
12e
13e
14f
15d
16e
17d
2.1:1
2.3:1
1.8:1
e
e
2.3:1
5:1
32:1
32:1
82
nd
80
90
4o
2m
Me
a
Isolated yield (%). Dimer:trimer ꢀ32:1 in all cases. Z:E > 32:1 as determined by
a
Dimer:trimer ratio determined by 1H NMR or GC-MS analysis of crude product.
% Yield is isolated yield for 2a.
Z:E > 32:1 in most cases as determined by GC-MS analysis, and a comparison of
GC-MS analysis, and a comparison of spectroscopic data with that of ref 14.
b
c
b
% ee determined by chiral HPLC analysis.
(S,Rp)-12e was used as catalyst (Method C).
(R,Sp)-12e was used as catalyst (Method C).
(S,Rp)-12b was used as catalyst (Method D).
(R)-Binaphane was used as catalyst and reaction conducted at ꢁ78 ꢂC for 48 h
c
spectroscopic data with that of ref 14.
d
d
% ee of dimer determined by chiral HPLC analysis.
e
e
Reaction conducted at ꢁ78 ꢂC.
f
f
Conducted in the presence of 0.3 equiv LiI.
(Method E).
g
Reaction conducted at 0 ꢂC.
g
(R,Sp)-12b was used as catalyst (Method D).
Conversion as determined by GC-MS analysis.
Obtained as a mixture of cis- and trans-isomers.
h
Reactions in entries 18e25 conducted at ꢁ25 ꢂC.
h
i
i
Catalyst solution was also cooled to ꢁ25 ꢂC.
4

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
derivatives, and BINAP. Phosphepines, in particular, were consid-
ered to be promising catalysts due to their track record of success in
reactions involving allenoates, as demonstrated by Fu’s group [15].
Once again methylphenylketene was utilized as the standard test
substrate.
a significant deleterious effect on reaction efficiency was observed.
Those substrates possessing R2 ¼ CH3 were homodimerized with
optimal conversion and enantioselectivity (90e96% ee) using 12e
(Method C in Experimental). On the other hand, less reactive sub-
strates, that possess more sterically demanding alkyl substituents
(Et or n-Bu), were homodimerized more efficiently through the use
of the less sterically hindered 12b (Method D in Experimental). The
latter reactions were conducted at a slightly higher concentration
(0.25 M), without any increase in trimer formation (dimer:trimer
ꢀ32:1).
Monophosphepines such as 5a-5d were found to give modest
enantioselectivity (up to 33% ee, Table 3, entries 1e5). Interestingly
(R)-Binaphane 5d was found to be a good enantioselective catalyst
for the homodimerization of ethylphenylketene (80% ee), but it
proved to be less successful with methylphenylketene as a sub-
strate (entries 4 and 5) [11a,11c]. As many complex molecule targets
of interest (e.g. LY426965) possess a methyl group at a quaternary
stereocenter, we were motivated to develop a more successful
method for the enantioselective homodimerization of methyl-
phenylketene [16]. We next turned our attention to chiral phos-
phines possessing planar chirality [17]. Commercially available
Josiphos derivatives 12a-12e were examined for reactivity and
enantioselectivity in the homodimerization of methylphenylketene
(entries 16e25) [18].
Variation of the aromatic R1 substituent was also tolerated with
both ortho- and para-substituents (electron-donating or electron-
withdrawing groups) working equally well in terms of affording
good enantioselectivity. Significantly, the only other high-level
enantioselective method for disubstituted ketene homodimeriza-
tion, the N-heterocyclic carbene-catalyzed method of Ye’s group,
does not tolerate ortho-substituted substrates [14]. Heteroaromatic
substituents such as thiophene were also tolerated, albeit with
significantly lower enantioselectivity (46% ee, entry 13) presumably
due to smaller steric size of R1 substituent. In the case of the 3-
thienyl substituent it was necessary to employ Binaphane in or-
der to achieve useful levels of enantioselectivity (78% ee, entry 14).
Dialkylketenes were also found to undergo homodimerization
efficiently although the achiral 1,3-cyclobutanedione regioisomer
was found to be the major isomer in most cases (entries 15e17),
just as was observed with PBu3 catalysis. Homodimerization of
methyl isopropyl ketene also gave the 1,3-cyclobutanedione
Josiphos derivative 12e was determined to be the optimal
catalyst with very good enantioselectivity of 90% ee being obtained
(Table 1, entry 25). Excellent dimer:trimer selectivity was obtained
after the reaction was diluted to 0.125 M concentration (ketene in
CH2Cl2) (entries 24 and 25). The introduction of a more electron
donating and sterically bulky t-Bu group at R1 on Josiphos (12d)
inhibited the reaction completely (entry 21). Interestingly, intro-
duction of electron withdrawing and/or sterically bulky sub-
stituents at R2 on the cyclopentadienyl phosphino group led to a
small improvement (4% ee) in enantioselectivity (entry 22 vs entry
19). Cooling of the catalyst solution to ꢁ25 ꢂC prior to addition to
regioisomer, with only trace b-lactone being formed. The change
in regioselectivity (relative to alkylarylketene homodimerization)
likely arises from electron-donating dialkyl-substitution on the
enolate, causing the enolate intermediate to cyclise through the C-
atom rather than through the O-atom (Scheme 2).
Ready access to both enantiomers of a given disubstituted
ketene dimer through use of commercially available antipodes of
Josiphos is a clear advantage of the phosphine-catalyzed method-
ology (Table 4, entry 1 vs entry 2, entry 3 vs entry 4) [18]. The major
olefin isomer in each case (2a-2p) was determined to be the Z-
isomer on the basis of previously reported X-ray crystal structure
the cooled methylphenylketene solution led to
a significant
improvement in enantioselectivity (entry 25 vs entry 24). Lower
and higher temperatures (e.g. ꢁ78 ꢂC and 0 ꢂC) than ꢁ25 ꢂC had a
negative effect on enantioselectivity (entries 16 and 17). Disap-
pointingly, the employment of LiI as an additive was found to lead
to an unclean and sluggish reaction (46% conv. after 24 h), in
contrast to our earlier findings with the PBu3/LiI catalytic system
[11a,11c].
2.5. Scope of the phosphine-catalyzed enantioselective
homodimerization reaction
The substrate scope of the Josiphos-catalyzed methodology was
then evaluated (Table 4). Simple extension of the R2 alkyl substit-
uent (to Et, and n-Bu) had little effect on reaction enantioselectivity
or efficiency, provided the optimal Josiphos derivative was
employed (entry 2 vs entries 4 and 7). However, if the R2 substit-
uent incorporated branching (e.g. Bn and i-Bu, entries 6 and 8), then
Scheme
4. Initial
mechanism
proposed
for
Josiphos-catalyzed
ketene
Scheme 3. Regioselectivity in phosphine-catalyzed ketene heterodimerization.
homodimerization.
5

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
2.8. Mechanism of the Josiphos-catalyzed dimerization reaction
We had previously proposed that Josiphos catalyzes the
homodimerization reaction through a mechanism involving phos-
phonium enolate 15a (Scheme 4) [11c]. Phosphonium enolate 15a
would be formed through addition of the di-c-hexylphosphino
moiety of Josiphos 12b/12e to the less sterically hindered side of the
ketene 1b to form an enolate 15a in stereoselective fashion. With
PBu3 as catalyst, the stereoselectivity of enolate formation had been
verified previously by 31P NMR monitoring of the reaction with
ethylphenylketene 1b to give enolate 15b (Scheme 5). An E:Z ratio
of 7:1 for 15b was measured, while kinetic trapping (in a separate
experiment) with TMSCl provided a TMS enol ether derivative 18
with high stereoselectivity (dr 16:1, Scheme 5) [11d].
Phosphonium enolate 15a would then react, through the C-atom
or O-atom of the enolate, with another molecule of a ketene. In our
initially proposed mechanism (Scheme 4), 15a was anticipated to
react with another molecule of ketene through C- of the enolate to
give acylphosphonium 16. 4-Exo-trig cyclisation would then afford
ketene homodimer 2b and allow for regeneration of catalyst 12b
[11c].
Scheme 5. Mechanistic experiments.
However, in the PBu3-catalyzed homodimerization of diphe-
nylketene (0.1 equiv of PBu3 employed), a strong signal at 13.6 ppm
was observed. The signal was subsequently revealed to be due to
phosphonium enolate 15c, after the enolate was generated in
stoichiometric fashion and characterized (Scheme 5). Subsequently,
strong evidence was obtained for a reaction mechanism involving
15c as the true catalyst, where it reacts through the O-atom of the
analysis of racemic 2a (CCDC 794919), and for other compounds, by
a comparison of spectroscopic data with that reported by Ye and co-
workers [11d,14].
2.6. Development of the phosphine-catalyzed heterodimerization
reaction
We noted from our homodimerization studies that Josiphos 12b
and 12e did not affect homodimerization of isobutylphenylketene
or diphenylketene (e.g. Table 4 entry 8). Indeed, 31P NMR analysis
revealed that there was no interaction between the phosphine
catalyst 12b and diphenylketene. Therefore, a strategy for hetero-
dimerization involving slow addition (over 8 h) of a more reactive
ketene (e.g. methylphenylketene) to a solution containing the
catalyst and less reactive ketene (e.g. isobutylphenylketene) was
explored (Scheme 3) [6]. This approach was successful to an extent,
in that moderate regioselectivity (r.s. up to 5:1) in heterodimer
formation was observed when there was a significant difference in
reactivity between the proposed donor and acceptor ketenes, e.g.
methylphenylketene vs isobutylphenylketene (Scheme 3, upper
eq.). However, when the two ketenes possessed similar reactivity
(electrophilicity), little or no regioselectivity was observed (Scheme
3, lower eq.).
2.7. In situ-generated ketene studies
The availability of an in situ ketene generation variant would
allow chemists to use commercially available acyl chlorides to
assemble racemic and enantioenriched ketene homodimers, and
thus bypass the need to isolate, purify and handle moisture-
sensitive ketenes. To that end, the possibility of utilizing in situ-
generated methylphenylketene for phosphine-catalyzed ketene
homodimerization studies was investigated [6]. Methyl-
phenylketene was generated through the addition of excess
Me2NEt to 2-phenylpropionyl chloride in THF (or toluene) at 0 ꢂC.
PBu3 (0.2 equiv) was then added slowly to the reaction mixture. The
reaction was then allowed to warm to room temperature overnight.
Disappointingly, the isolated yield of desired ketene homodimer 2a
was modest, ca. 20e30%, and significantly lower compared to when
pre-generated methylphenylketene was used (Tables 1 and 2).
Scheme 6. Revised mechanism proposed for Josiphos-catalyzed ketene
homodimerization.
6

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
enolate [11d]. The in situ-formed catalyst 15c was proposed to
promote the homodimerization of diphenylketene by adding
though the O-atom of the enolate, rather than through the C-atom
of the enolate, to another molecule of ketene. Evidence for the
propensity of phosphonium enolate 15 (15b or 15c) to react
through the O-atom with carbonyl substrates was obtained by ex-
amination of the reaction of enolates 15b and 15c with 2-
phenylbutanoyl chloride. In both cases, the reaction proceeded
primarily through the O-atom, as was confirmed by characteriza-
tion of the ester products 19 and 21, especially with the observation
of an IR absorption band at 1761 and 1765 cmꢁ1 respectively [11d].
The presence of the ester group in both derivatives was further
confirmed by 13C NMR spectra with signals observed at 170.0 and
171.2 respectively. 15c was also independently generated and
demonstrated to act as a catalyst (at 20 mol% loading) for the
homodimerization of ethylphenylketene (95% yield).
Furthermore, TMSCl trapping of an intermediate formed in the
reaction of PBu3 with ethylphenylketene revealed the presence of a
new O-acylated (rather than C-acylated) derivative 20. This result
strongly points towards a mode of reactivity involving reaction of
15 with ketenes through the O-atom of the enolate rather than
through the C-atom.
In addition, acylphosphonium 16 (Scheme 4) would be expected
to appear between 28 and 32 ppm in 31P NMR spectra during re-
action monitoring. The absence of such a signal, in PBu3 and
Josiphos-catalyzed ethylphenylketene homodimerization re-
actions, strongly suggested that the reaction does not proceed
through an acylphosphonium intermediate [19].
Therefore, our revised proposal for the reaction mechanism is as
follows: reaction of phosphonium enolate 15a through the O atom
of the enolate (O-acylation) with the carbonyl of another molecule
of ketene leads to the formation of a second enolate intermediate
22 (Scheme 6). The latter ‘ester’ enolate intermediate adds to a third
molecule of ketene to generate a new enolate intermediate 23,
which undergoes 4-exo-trig cyclisation to afford the desired ketene
homodimer 2b and regenerate enolate 15a/15d. The assertion that
the Josiphos-catalyzed homodimerization would involve a similar
mechanism to that of the PBu3-catalyzed reaction is justified when
one considers the similar electronics of PBu3 and the tri-
alkylphosphine sidechain of Josiphos (Schemes 4e6). Moreover, the
observation that isobutylphenylketene homodimer was formed as
a side product (ca. 3e5%) during the heterodimerization reaction
(Scheme 3), whereas none (0%) was formed during an attempted
Josiphos-catalyzed homodimerization of isobutylphenylketene
(Table 4 entry 8) supports the idea of an alternative catalytic cycle
(i.e. not 15a to 16). A reasonable explanation would be that the
phosphonium enolate catalyst 15 is more readily generated
through the reaction of Josiphos with methylphenylketene than
with the more sterically hindered isobutylphenylketene, and that
the methylphenylketene-derived catalyst subsequently catalyzes
the homodimerization of isobutylphenylketene. We have previ-
ously demonstrated the use of a phosphonium enolate 15c, derived
from diphenylketene and tri-n-butylphosphine, in promoting the
homodimerization of ethylphenylketene, as well as the reaction of
methylphenylketene with 4-chlorobenzaldehyde [11d,20].
diphenylphosphino group at ꢁ28.1 ppm (d)). The signals between
40 and 41 ppm are consistent with tetravalent phosphonium spe-
cies [19,21,22]. We propose that the signals at 40.4 ppm and
40.1 ppm represent the two possible enolate isomers 15a and 15d
(Scheme 6). Both signals appeared approximately equal by inte-
gration (ratio ¼ ca. 1.3:1) throughout the time of reaction moni-
toring at ꢁ30 ꢂC, but the most intense signals were due to the free
Josiphos catalyst 12b (31P NMR
d
15.5 (d), and ꢁ25.5 (d) ppm), with
the ratio of 12b: (15a þ 15d) ¼ ca. 2:1. This shows that free Josiphos
12b is the resting state of the catalyst (at least for the formation of
ethylphenylketene dimer 2b), and that conversion of 12b to 15a
and 15d represents the slow step of the reaction.
The formation of 15 as a mixture of E/Z-isomers does not mean
that the ketene homodimer must be formed with low enantiose-
lectivity. This is because the key enantioselectivity-determining
step is that of ‘ester’ enolate 22 with ketene 1b to give 23. The
Josiphos catalyst structure in 22 blocks the same face of the ester
enolate, regardless of the olefin isomer geometry generated pre-
viously in 15. In other words, interchanging the Et and Ph sub-
stituents in the phosphonium enol moiety of 22 does not affect
sense of enantioselectivity significantly or change which enantio-
face becomes sterically hindered. Reaction of ester enolate 22 with
ketene 1b proceeds in a non-reversible fashion to afford interme-
diate 23, which undergoes cyclisation to provide ketene dimer 2b
with high enantioselectivity (86e90% ee) and diastereoselectivity
(Z:E > 32:1).
Finally, the absence of signals from the ꢁ50 to ꢁ120 ppm range
in all 31P NMR reaction monitoring experiments, conducted at a
variety of temperatures, strongly suggests that the phosphine-
catalyzed reactions avoid intermediacy of pentacovalent phos-
phorane species, such as 17a/17b (Schemes 4 and 6), or at least that
they are not formed in high enough concentrations to be detected
by NMR spectroscopy. The electron-donating alkyl substituents
(e.g. c-Hex) on the phosphonium centre of 22 would be expected to
stabilize the positively charged phosphorus atom, making it less
electrophilic, and hence disfavouring cyclisation to a pentacovalent
phosphorane intermediate (Scheme 6) [21]. This is in contrast to an
earlier mechanistic proposal by Bentrude’s group for the P(OMe)3-
promoted homodimerization of dimethylketene, which involved a
pentacovalent phosphorane intermediate of type 17b [4d,4e]. The
observation of such an intermediate in Bentrude’s studies may be
understood by taking into account the greater electrophilicity of a
phosphite-derived phosphonium centre, and hence the greater
susceptibility to nucleophilic attack by enolate oxygen.
3. Conclusion
Details of extensive optimisation studies revealed that tri-n-
butylphosphine is the optimal catalyst for promoting non-
enantioselective homodimerization of a range of alkylarylketenes
and dialkylketenes. On occasion lithium iodide was found to be a
useful additive in suppressing ketene trimerization. For the asym-
metric ketene homodimerization reaction, Josiphos derivatives
were found to give the best results, frequently providing the
31P NMR data obtained through reaction monitoring at ꢁ30 ꢂC
and at ꢁ78 ꢂC is also consistent with this new picture of the
Josiphos-catalyzed homodimerization reaction. The 31P NMR
spectrum of Josiphos 12b provided signals at 15.5 ppm (d, -P(c-
Hex)2) and at ꢁ25.5 ppm (d, -PPh2). 31P NMR monitoring at ꢁ78 ꢂC
of the 12b-catalyzed homodimerization of ethylphenylketene
revealed new signals at 40.5 ppm (d) and ꢁ28.4 ppm (d). On
warming to ꢁ30 ꢂC, 31P NMR monitoring revealed signals at
40.4 ppm (d) (along with free diphenylphosphino group
at ꢁ27.9 ppm (d)) and at 40.1 ppm (d) (along with free
desired
opment of
b
-lactone ketene dimer in ꢀ90% ee. Studies on the devel-
a
phosphine-catalyzed heterodimerization of two
disubstituted ketenes were less successful, with the desired ketene
heterodimer being formed in modest yield and with moderate
regioselectivity. Mechanistic studies involving intermediate trap-
ping experiments and extensive 31P NMR analysis suggested that an
in situ-generated phosphonium enolate was acting as the true
catalyst of the dimerization reactions but we cannot rule out the
possibility that an equilibrium amount of free phosphine is the true
active catalyst.
7

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
4. Experimental section
assigned on the basis of a comparison of specific rotation values
with literature values [11c]. the absolute configurations of 2c and
2g-2l were assigned by analogy. The dimer: trimer ratios
(Tables 1e4) were determined for the crude products by 1H NMR
analysis or by comparing the peak areas of GC-MS data. The major
diastereomer of 2a-2p was determined to be the Z-isomer through
comparison with literature spectroscopic data, and by X-ray crystal
structure analysis of racemic 2a (CCDC 794919) [11d]. The Z:E olefin
ratio for 2a-2p was determined to be > 32:1 by GC-MS analysis of
4.1. General information
All reactions were carried out in flame dried glassware under a
nitrogen atmosphere using standard inert atmosphere techniques
unless otherwise stated. Diethyl ether and THF were dried using
sodium benzophenone stills, CH2Cl2 and toluene were dried using
calcium hydride stills, methanol was dried using a magnesium
methoxide still, and N,N-dimethylethylamine was distilled from
potassium hydroxide under nitrogen [23]. (R)-1-[(Sp)-2-(Diphe-
nylphosphino)ferrocenyl]-ethyldicyclohexyl phosphine, (S)-1-
[(Rp)-2-(Diphenylphosphino)ferrocenyl]ethyl-dicyclohexylphosphi
ne, (S)-1-{(Rp)-2-[Bis[3,5-bis(trifluoromethyl)phenyl]phosphino]fe
rrocenyl}ethyldicyclohexylphosphine, (R)-1-{(Sp)-2-[Bis[3,5-bistrifl
uoromethyl)phenyl]phosphino]ferrocenyl}ethyldicyclohexylphosp
hine were purchased from Aldrich Chemical Co. and Strem Chem-
icals and used as received. (R)-Binaphane, tri-n-butyl phosphine,
lithium iodide, n-butyllithium (2.5 M in hexane), and LiAlH4 (1.0 M
in Et2O) were purchased from Aldrich Chemical Co. and used as
the crude b-lactones.
4.3. Synthetic procedures
4.3.1. Synthesis of racemic
ketoketenes
b
-lactones by homodimerization of
homodimerization of ketoketenes.
4.3.1.1. Method
A
for
Ketoketene (1.39 mmol, 1.0 equiv) was dissolved in CH2Cl2 (1.4 mL).
LiI (0.42 mmol, 0.3 equiv) was dissolved in Et2O (1.4 mL) and was
then transferred to the flask containing the ketoketene solution and
the resulting solution (0.5 M of ketoketene in solvent) was cooled to
0 ꢂC. Tri-n-butylphosphine (0.14 mmol, 0.1 equiv.) was added in one
portion, and stirred for the indicated time at the indicated tem-
perature. The reaction was then diluted with CH2Cl2 (5 mL) and
quenched by adding deionized water (10 mL). The layers were
separated, the aqueous layer was extracted with CH2Cl2 (2 ꢃ 5 mL),
and the combined organics were dried over anhydrous sodium
sulfate. Following filtration of the drying agent, aqueous H2O2 so-
lution (50%, 2 drops) was added to the CH2Cl2 solution and stirred
for 10 min at room temperature. The solvent was removed under
reduced pressure to provide the crude product for 1H NMR/GC-MS
analysis. 10% EtOAc/hexane (20 mL) and dichloromethane (5 mL)
were added to the crude residue, which was passed through a plug
column of neutral silica (iatrobeads, 2 ꢃ 2 cm, 10 g) and was eluted
with 10% EtOAc/hexane (100 mL). Finally solvent was removed
under reduced pressure to yield the desired ketoketene dimer in
high purity (ꢀ95%), in most cases, as determined by GC-MS and
HPLC analysis and confirmed by 1H and 13C NMR spectroscopy.
received. Iatrobeads (Bioscan, 6RS-8060, 60
TLC plates (Sorbent Technologies, UV254, 250
m
M particle size) and
mM) were used as
received. Methylphenylketene, methyl-4-tolylketene, methyl-2-
tolylketene, ethyl-4-tolylketene, methyl-2-chlorophenylketene,
ethylphenylketene, ethyl-4-chlorophenylketene, ethyl(3-thioph
enyl)ketene, n-butylphenylketene, isobutylphenylketene, benzyl-
phenylketene, diphenylketene, c-hexyl methyl ketene, ethyl c-
hexyl ketene and dimethylketene were prepared according to
literature procedures [24].
NMR spectra were recorded on a Bruker DPX Avance 200
spectrometer (200 MHz for 1H and 50 MHz for 13C) and on a Bruker
Biospin AG 400 spectrometer (400 MHz for 1H and 100 MHz for
13C). NMR chemical shifts were reported relative toTMS (0 ppm) for
1H and to CDCl3 (77.23 ppm) for 13C spectra. IR spectra were
recorded on a Bio Rad FTS-175C spectrometer. Optical rotations
were measured on
polarimeter.
a Rudolph DigiPol 781 TDV automatic
Low resolution mass spectra were recorded on a GC-MS Hewlett
Packard HP 6890 GC instrument with a 5973 mass selective de-
tector. High resolution mass spectra were obtained from the College
of Sciences Major Instrumentation Cluster at Old Dominion Uni-
versity. Analytical high performance liquid chromatography (HPLC)
4.3.1.2. Method B for homodimerization of ketoketenes. To a solution
of ketoketene (0.62 mmol, 1 equiv) in CH2Cl2 (1.2 mL, 0.5 M), tri-n-
butylphosphine (0.06 mmol, 0.1 equiv) was added at 0 ꢂC. The re-
action was stirred at the indicated temperature for the indicated
time. The reaction was then quenched by the addition of aqueous
H2O2 solution (50%, 2 drops) and stirred for 10 min at room tem-
perature. The solvent was removed under reduced pressure to
provide the crude product for 1H NMR/GC-MS analysis. 10% EtOAc/
hexane (20 mL) and dichloromethane (5 mL) were added to the
crude residue, which was passed through a plug column of neutral
silica (iatrobeads, 2 ꢃ 2 cm, 10 g) and was eluted with 10% EtOAc/
hexane (100 mL). Finally the solvent was removed under reduced
pressure to yield the desired ketoketene dimer in high purity
(ꢀ95%), in most cases, as determined by GC-MS and HPLC analysis
and confirmed by 1H and 13C NMR spectroscopy.
was performed using
a
a
Daicel Chiralpak AD column
Daicel Chiralpak AS-H column
(0.46 cm
(0.46 cm
ꢃ
25 cm),
ꢃ
25 cm) and a Daicel Chiralcel OD-H column
(0.46 cm ꢃ 25 cm) (Daicel Chemical Ind., Ltd.) on a PerkinElmer
235C instrument attached with diode array detector (deuterium
lamp, 190e600 nm) with HPLC-grade isopropanol and hexanes as
the eluting solvents.
Racemic ketene dimers 2a-2n and achiral cyclobutanediones 4q
and 4o were prepared and characterized as previously described
[11a,11d]. Enantioenriched 2a-2p were prepared and characterized
as previously described [11c]. Phosphonium enolate trapping
products 18e21 were prepared and characterized as previously
described.11d 31P NMR Spectra for 15c and 18e21 are found in the
Electronic Supplementary Information.
4.3.1.3. Method
C for asymmetric homodimerization of methyl
substituted ketoketenes. Ketoketene (0.34 mmol, 1.0 equiv)
was dissolved in CH2Cl2 (2.2 mL) and cooled to ꢁ25 ꢂC. (S)-1-{(Rp)-
2-[Bis[3,5-bis(trifluoromethyl)phenyl]phosphino]ferrocenyl}ethyl-
dicyclohexylphosphine or (R)-1-{(Sp)-2-[Bis[3,5-bis(trifluorometh
yl)phenyl]phosphino]ferrocenyl}ethyldicyclohexylphosphine (0.03
mmol, 0.1 equiv) was dissolved in CH2Cl2 (0.5 mL), and was then
transferred via syringe to the flask containing the ketoketene so-
lution. The resulting solution (0.125 M of ketoketene in solvent)
was stirred for 24 h at ꢁ25 ꢂC before being briefly warmed to room
temperature. The reaction was then quenched by the addition of
4.2. Determination of enantiomeric excesses, dimer: trimer ratios
(Tables 1e4) and Z: E olefin ratios
Enantiomeric excesses were determined by assaying the
b
-lac-
tones 2a-2p using chiral HPLC analysis (at
l
¼ 254 or 225 nm;
details given for each compound). Authentic racemic samples for
chiral HPLC analysis were generated through the PBu3-catalyzed
reaction [11a]. the absolute configurations of 2a, 2b, 2j, and 2p were
8

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
aqueous H2O2 solution (50%, 2 drops) at room temperature. After
stirring for 10 min at room temperature, the solvent was removed
under reduced pressure. The crude product was dissolved in 10%
EtOAc/hexane (5 mL) and dichloromethane (1 mL). The resulting
solution was passed through a plug column of neutral silica
(iatrobeads, 2 ꢃ 2 cm, 4 g) [50 x weight of reaction mixture]. The
plug column was eluted with 10% EtOAc/hexane (100 mL), and the
solvent was removed under vacuum to furnish the desired keto-
ketene dimer with ꢀ95% purity in most cases (as determined by
GC-MS and 1H NMR analysis). Further purification was carried out
in some cases as specified.
4.3.1.7. Z-3-Ethyl-3-phenyl-4-(1-phenyl-propylidene)-oxetan-2-one
(( )-2b). Method A was followed. Ethylphenylketene 1b (160 mg,
1.09 mmol), stirred for 4 h at 0 ꢂC, was isolated as a colorless oil
(135 mg, 84%); IR (thin film): 1857, 1699, 1140 cmꢁ1 1H NMR
;
(200 MHz, CDCl3, TMS):
d 7.42e7.20 (m, 10H), 2.38e2.02 (m, 4H),
1.15 (t, J ¼ 7.4 Hz, 3H), 0.83 (t, J ¼ 7.4 Hz, 3H); 13C NMR (50 MHz,
CDCl3):
d 170.9, 143.9, 135.4, 134.7, 129.3, 128.6, 128.6, 128.2, 127.6,
126.5, 116.5, 70.1, 26.3, 22.9, 12.8, 10.0; MS (EI 70 eV): m/z 292, 146,
117, 103, 91, 77; (M þ Na)þ HRMS m/z calcd for C20H20O2Na:
315.1355; found: 315.1351.
4.3.1.8. Z-3-Butyl-3-phenyl-4-(1-phenylpentylidene)-oxetan-2-one
(( )-2c). Method B was followed. n-Butylphenylketene 1c (108 mg,
0.62 mmol), stirred for 1.5 h at 0 ꢂC, was isolated as a colorless oil
4.3.1.4. Method D for asymmetric homodimerization of ketoketenes.
Ketoketene (0.31 mmol, 1.0 equiv) was dissolved in CH2Cl2 (1 mL)
and cooled to ꢁ25 ꢂC. (R)-1-[(Sp)-2-(Diphenylphosphino)ferro-
cenyl]ethyldicyclohexylphosphine or (S)-1-[(Rp)-2-(Diphenyl-
phosphino)ferrocenyl]ethyldicyclohexylphosphine (0.03 mmol, 0.1
equiv) was dissolved in CH2Cl2 (0.35 mL), cooled to ꢁ25 ꢂC in most
cases (exceptions: formation of 2l and 4o), and was then trans-
ferred via syringe to the flask containing the ketoketene solution.
The resulting solution (0.25 M of ketoketene in solvent) was stirred
for 24 h at ꢁ25 ꢂC before being briefly warmed to room tempera-
ture. The reaction was then quenched by the addition of aqueous
H2O2 solution (50%, 2 drops) at room temperature. After stirring for
10 min at room temperature, the solvent was removed under
reduced pressure. The crude product was dissolved in 10% EtOAc/
hexane (5 mL) and dichloromethane (1 mL). The resulting solution
was passed through a plug column of neutral silica (iatrobeads,
2 ꢃ 2 cm, 4 g) [50 x weight of reaction mixture]. The plug column
was eluted with 10% EtOAc/hexane solvent system (100 mL), and
the solvent was removed under vacuum to furnish the desired
ketoketene dimer with ꢀ95% purity in most cases (as determined
by GC-MS and 1H NMR analysis). Further purification was carried
out in some cases as specified.
(108 mg, >99%); IR
(thin film): 1869, 1703 cmꢁ1 1H NMR
;
n
max
(400 MHz, CDCl3, TMS):
d 7.38e7.19 (m, 10H), 2.30e2.19 (m, 2H),
2.12e2.04 (m, 2H), 1.61e1.36 (m, 2H), 1.28e1.02 (m, 2H), 0.90 (t,
J ¼ 7.2 Hz, 3H), 0.86e0.74 (m, 4H), 0.69 (t, J ¼ 7.1 Hz, 3H); 13C NMR
(100 MHz, CDCl3):
d 171.1, 144.8, 135.6, 135.1, 129.3, 128.6, 128.5,
128.2, 127.6, 126.5, 115.3, 69.3, 32.9, 30.4, 29.5, 27.6, 23.0, 22.8, 14.1,
14.0; MS (EI 70 eV): m/z 348, 264, 205, 174, 131, 117, 103; (M þ Na)þ
HRMS m/z calcd for C24H28O2Na: 371.1982; found: 371.1984.
4.3.1.9. Z-3-Benzyl-4-(1,2-diphenylethylidene)-3-phenyloxetan-2-
one (( )-2e). Method B was followed. Benzylphenylketene 1e
(111 mg, 0.53 mmol), stirred for 5 h at room temperature. Crude 2e
was obtained with 64% conversion as determined by GC-MS anal-
ysis; MS (EI 70 eV): m/z 416, 281, 208, 179.
4.3.1.10. Z-3-(2-Chlorophenyl)-4-(1-(2-chlorophenyl)ethylidene)-3-
methyloxetan-2-one (( )-2g). Method
chlorophenylketene 1g (60 mg, 0.36 mmol), stirred for 1.5 h at
ꢂC, was isolated as a colorless oil (60 mg, >99%); IR
(thin
B was used. Methyl-2-
0
n
max
film): 1891, 1843, 1725, 1647 cmꢁ1; 1H NMR (400 MHz, CDCl3, TMS):
7.48e7.12 (m, 8H), 2.06 (s, 3H), 1.68 (s, 3H); 13C NMR (100 MHz,
CDCl3): 169.8, 145.7, 136.0, 134.9, 133.4, 132.0, 131.5, 130.7, 130.3,
130.0, 129.3, 128.8, 127.5, 127.1, 108.5, 62.1, 20.0, 16.1; MS (EI 70 eV):
m/z 332, 166, 138, 103; (M
Na)þ HRMS m/z calcd for
18H14Cl2O2Na: 355.0263; found: 355.0266.
d
4.3.1.5. Method E for asymmetric homodimerization of ketoketenes.
Ketoketene (0.34 mmol, 1.0 equiv) was dissolved in CH2Cl2
(0.35 mL) and was cooled to ꢁ78 ꢂC. (R)-Binaphane (0.03 mmol, 0.1
equiv) was dissolved in CH2Cl2 (0.35 mL) and cooled to ꢁ78 ꢂC. The
phosphepine solution was then transferred via syringe to the flask
containing the ketoketene solution. The reaction was stirred
at ꢁ78 ꢂC for 48 h, after which the solvent was removed under
reduced pressure. The crude product was treated with dry iso-
propanol (3 mL) to dissolve the crude ketoketene dimer and pre-
cipitate the phosphepine. The mixture was filtered under nitrogen
using a Schlenk filter funnel. The precipitate was washed with dry
isopropanol (2 ꢃ 3 mL), the phosphine was recovered, and the
solvent was removed from the filtrate under reduced pressure. The
crude product was then dissolved in 5% EtOAc/hexane (6 mL) and
CH2Cl2 (1.5 mL), before being passed through a plug of neutral silica
(3.3 g). Elution with 5% EtOAc/hexane (60 mL), followed by solvent
removal under reduced pressure furnished the desired ketoketene
dimer with ꢀ95% purity in most cases (as determined by GC-MS
and 1H NMR analysis). Further purification was carried out in
some cases as specified.
d
þ
C
4.3.1.11. Z-3-Methyl-3-o-tolyl-4-(1-o-tolylethylidene)oxetan-2one
(( )-2h). Method B was used. Methyl-2-tolylketene 1h (50 mg,
0.34 mmol), stirred for 1 h at 0 ꢂC, was isolated as a colorless oil
(49 mg, 97%); IR n max (thin film): 1883, 1844, 1717, 1456, 1112 cmꢁ1
;
1H NMR (400 MHz, CDCl3, TMS):
2.15 (s, 3H), 2.10 (s, 3H), 1.79 (s, 3H); 13C NMR (100 MHz, CDCl3):
171.2, 145.3, 137.7, 136.8, 136.0, 133.1, 132.7, 130.6, 128.7, 128.5,
d 7.43e7.10 (m, 8H), 2.26 (s, 3H),
d
128.0, 127.1, 126.6, 126.1, 110.3, 63.3, 20.6, 20.5, 19.9, 17.5; MS (EI
70 eV): m/z 292, 264, 249, 146, 117, 91; (M þ Na)þ HRMS m/z calcd
for C20H20O2Na: 315.1356; found: 315.1359.
4.3.1.12. Z-3-Ethyl-3-thiophen-3-yl-4-(1-thiophen-3-yl-propyli-
dene)-oxetan-2-one (( )-2l). Method
thiophenylketene 1l (162 mg, 1.06 mmol), stirred for 7 h at 0 ꢂC,
isolated as a colorless oil (127 mg, 78%); IR (thin film): 1859,
A
was used. Ethyl-3-
n
max
1695, 1411 cmꢁ1 1H NMR (200 MHz, CDCl3, TMS):
; d 7.30e7.16 (m,
4.3.1.6. Z-3-Methyl-3-phenyl-4-(1-phenyl-ethylidene)-oxetan-2-one
(( )-2a). Method A was followed. Methylphenylketene 1a (270 mg,
2.05 mmol), stirred for 2 h at 0 ꢂC, and was isolated as a colorless oil
(221 mg, 82%); IR (thin film) 1881, 1844, 1699, 1140 cmꢁ1; 1H NMR
4H), 7.05e7.02 (m, 2H), 2.39e2.01 (m, 4H), 1.09 (t, J ¼ 7.3 Hz, 3H),
0.92 (m, J ¼ 7.4 Hz, 3H); 13C NMR (50 MHz, CDCl3):
d 170.1, 144.5,
136.1, 135.4, 127.4, 127.1, 126.1, 125.5, 122.6, 122.4, 111.3, 67.5, 26.1,
22.6,13.5, 9.7; MS (EI 70 eV): m/z 247,152,123, 97; (M þ Na)þ HRMS
m/z calcd for C16H16O2S2Naþ: 327.0484; found: 327.0485.
(200 MHz, CDCl3, TMS):
d
7.52e7.16 (m, 10H), 1.92 (s, 3H), 1.86 (s,
171.4, 146.9, 136.2, 135.2, 129.4,
3H); 13C NMR (50 MHz, CDCl3):
d
128.6, 128.6, 127.6, 127.4, 126.3, 108.6, 64.4, 19.6, 15.6; MS (EI 70 eV):
m/z 264, 132, 104, 78; (M þ Na)þ HRMS m/z calcd for C18H16O2Na:
287.1043; found: 287.1039.
4.3.1.13. 2,4-Dicyclohexyl-2,4-diethylcyclobutane-1,3-dione
(4o).
Method A was used. Ethyl c-hexyl ketene 1o (52 mg, 0.34 mmol),
stirred at 0 ꢂC to room temperature overnight. Crude 4o was
9

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
purified by passing through a plug of neutral silica (iatrobeads),
eluting with 3% EtOAc/hexane. The solvent was removed under
reduced pressure to yield 4o as a colorless oil (45 mg, 87%), and as
an inseparable mixture of diastereomers (cis:trans ¼ 1:1 as deter-
N,N-dimethylethylamine (0.66 mL, 6.0 mmol) was then added
slowly to the cooled solution over 10 min to give a yellow-coloured
mixture. Tri-n-butylphosphine (76 mL, 0.3 mmol) was added slowly
to the cooled reaction mixture over 1 min. The reaction was then
allowed to warm to room temperature overnight. The reaction
mixture was filtered through a plug of silica gel eluting with 30%
EtOAc/hexane to afford crude product solution. Further purification
of the crude product through flash column chromatography eluting
with a gradient system (hexane to 1%e2% EtOAc/hexane), followed
by solvent removal under reduced pressure furnished the desired
ketoketene dimer 2a (47 mg, 24%). A similar yield (42%, composed
of dimer:trimer ¼ ca. 2:1) was obtained when the reaction was
conducted in toluene.
mined by NMR analysis); IR
(thin film): 1754, 1714 cmꢁ1
;
1H
2.11e2.05
(m, 4H), 1.74e1.51 (m, 32H), 1.19e1.02 (m, 16H), 0.88 (t, J ¼ 7.4 Hz,
12H); 13C NMR (100 MHz, CDCl3) for both diastereomers:
171.5,
n
max
NMR (400 MHz, CDCl3, TMS) for both diastereomers:
d
d
171.5, 55.0, 55.0, 40.0, 40.0, 31.3, 31.3, 30.6, 30.6, 26.5, 26.5, 22.3,
22.3, 12.2, 12.2; MS (EI 70 eV): m/z 304, 153, 125, 83; HRMS analysis
obtained for 1,3-diol derivative (through LiAlH4 reduction of 4o)
(M þ Na)þ HRMS m/z calcd for C20H32O2Na: 331.2608; found:
331.2607.
4.3.1.14. Z-3-Ethyl-3-p-tolyl-4-(1-p-tolyl-ethylidene)-oxetan-2-one
(( )-2j). Method B was used. Ethyl-4-tolylketene 1j (77 mg,
0.48 mmol), stirred for 5 h at 0 ꢂC, was isolated as a colorless oil
(77 mg, >99%); IR n max (thin film): 1862, 1734, 1700 cmꢁ1; 1H NMR
4.3.1.18. Procedure for phosphine-catalyzed ketene heterodimeriza-
tion. Methylphenylketene 1a (51 mg, 0.38 mmol) in CH2Cl2
(0.8 mL) was added over 8 h to a solution of isobutylphenylketene
1d (55 mg, 0.38 mmol) and phosphine 12b (25 mg, 0.04 mmol) in
CH2Cl2 (3 mL) at ꢁ25 ꢂC. The reaction was then stirred at ꢁ25 ꢂC for
15 h. GC-MS analysis of the crude product showed 30% ketene
heterodimer (5:1 ratio of heterodimers 13b:13a), and 66% meth-
ylphenylketene homodimer 2a; 1H NMR (400 MHz, CDCl3, TMS) for
(400 MHz, CDCl3, TMS):
d 7.30e7.12 (m, 8H), 2.34e2.22 (m, 2H),
2.29 (s, 3H), 2.29 (s, 3H), 2.11e2.05 (m, 2H), 1.14 (t, J ¼ 7.4, 3H), 0.84
(t, J ¼ 7.4, 3H); 13C NMR (100 MHz, CDCl3):
d 171.1, 143.8, 138.3,
137.3, 132.5, 131.8, 129.9, 129.3, 128.1, 126.3, 116.1, 69.8, 26.2, 22.8,
21.4, 21.2, 12.8, 9.9; MS (EI 70 eV) m/z 320, 263, 160, 132, 117;
(M þ Na)þ HRMS m/z calcd for C22H24O2Na: 343.1669; found:
343.1667.
major regioisomer
d
7.55e6.91 (m, 10H), 2.27 (dd, J ¼ 14.3, 6.3 Hz,
1H), 2.05e1.99 (m, 1H), 2.01 (s, 3H), 1.64e1.51 (m, 1H), 0.77 (d,
J ¼ 6.6 Hz, 3H), 0.60 (d, J ¼ 6.7 Hz, 3H); 13C NMR (100 MHz, CDCl3)
for major regioisomer d 171.6, 147.4, 135.5, 135.0, 129.3, 128.6, 128.5,
4.3.1.15. (S,Z)-3-Methyl-3-phenyl-4-(1-phenyl-ethylidene)-oxetan-2-
one ((ꢁ)-2a). Method C was used. To a cooled solution of meth-
ylphenylketene (510 mg, 3.85 mmol) in CH2Cl2 (28 mL) was added
128.3, 127.6, 126.4, 113.9, 63.6, 38.4, 26.1, 22.4, 22.2, 19.6; MS (EI
70 eV): m/z 306 (13b).
a
solution of (S)-1-{(Rp)-2-[bis[3,5-bis(trifluoromethyl)phenyl]
Declaration of competing interest
phosphino]ferrocenyl}ethyldicyclohexylphosphine (333 mg, 0.39
mmol) in CH2Cl2 (3 mL). (ꢁ)-2a was isolated as a colorless oil
(331 mg, 65%); HPLC analysis: 94% ee [Daicel Chiralpak AD col-
umn; 1 mL/min; solvent system: 2% isopropanol in hexane;
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
retention times: 5.0 min (minor), 6.6 min (major)]; [
a
]
¼ ꢁ68.3
D
(c ¼ 0.41, CHCl3); IR (thin film) 1881, 1844, 1699, 1140 cm-1; 1H
Acknowledgments
NMR (400 MHz, CDCl3, TMS):
d 7.50e7.10 (m, 10H), 1.90 (s, 3H),
1.85 (s, 3H); 13C NMR (100 MHz, CDCl3):
d 171.3, 146.9, 136.2,
Support has been provided by the National Science Foundation:
Grant No. CHE-1463728 to N.J.K.
135.2, 129.3, 128.6, 128.6, 127.6, 127.4, 126.2, 108.6, 64.4, 19.6, 15.5;
MS (EI 70 eV): m/z 264, 132, 104, 78; (M þ Na)þ HRMS m/z calcd
for C18H16O2Na: 287.1043; found: 287.1039.
Appendix A. Supplementary data
4.3.1.16. Kinetic experiments on tri-n-butylphosphine-catalyzed
homodimerization of isobutylphenylketene. To a constant concen-
tration of isobutylphenylketene in CH2Cl2 (0.1 M) was added a
varied number of equivalents of PBu3 (0.01, 0.035, 0.05, 0.1, and 0.2
equiv.) at room temperature. The reaction was monitored every 30,
70,115, 200 min by GC-MS analysis to determine the rate of product
2d formation (Fig. 1).
Tri-n-butylphosphine (0.1 equiv.) was added to different con-
centrations of isobutylphenylketene (0.05, 0.1, 0.15, 0.25, 0.5 M) in
CH2Cl2 at room temperature. Each reaction was monitored indi-
vidually for the amount of product 2d formed every 30, 70, 115,
200 min by GC-MS analysis (Fig. 2).
From the GC-MS traces, the relative areas of the ketene 1d (a
known starting amount of ketene was used as the standard for
calculations) to that of the dimer product 2d were used to deter-
mine the number of mmol of ketene and ketene dimer in the
mixture at any given time. See the Supporting Information for
Tables showing details of kinetic experiments such as the average
rate of product formed (mmol/min) at the various time intervals.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.tet.2020.131838.
References
[1] (a) J. Taunton, J.L. Collins, S.L. Schreiber, J. Am. Chem. Soc. 118 (1996)
10412e10422;
(b) Z. Wan, S.G. Nelson, J. Am. Chem. Soc. 122 (2000) 10470e10471;
(c) S.G. Nelson, W.S. Cheung, A.J. Kassick, M.A. Hilfiker, J. Am. Chem. Soc. 124
(2002) 13654e13655;
(d) X. Shen, A.S. Wasmuth, J. Zhao, C. Zhu, S.G. Nelson, J. Am. Chem. Soc. 128
(2006) 7438e7439;
(e) Y. Wang, R. Tennyson, D. Romo, Heterocycles 64 (2004) 605e658.
[2] (a) A. Pommier, J.-M. Pons, Synthesis (1995) 729e744;
(b) H.W. Yang, D. Romo, J. Org. Chem. 62 (1997) 4e5;
(c) B.W. Dymock, P.J. Kocienski, J.-M. Pons, Synthesis (1998) 1655e1661;
(d) L.R. Reddy, P. Saravanan, E.J. Corey, J. Am. Chem. Soc. 126 (2004)
6230e6231.
[3] Leading catalytic asymmetric approaches to b-lactones from ketenes and al-
dehydes: (a) C. Zhu, X. Shen, S.G. Nelson, J. Am. Chem. Soc. 126 (2004)
5352e5353;
(b) J.E. Wilson, G.C. Fu, Angew. Chem. Int. Ed. 43 (2004) 6358e6360;
(c) M.A. Calter, O.A. Tretyak, C. Flaschenriem, Org. Lett. 7 (2005) 1809e1812;
(d) L. He, H. Lv, Y.-R. Zhang, S. Ye, J. Org. Chem. 73 (2008) 8101e8103;
(e) T. Kull, R. Peters, Angew. Chem. Int. Ed. 47 (2008) 5461e5464;
(f) M. Mondal, A.A. Ibrahim, K.A. Wheeler, N.J. Kerrigan, Org. Lett. 12 (2010)
1664e1667;
4.3.1.17. In situ generation of methylphenylketene for ketene homo-
dimer synthesis. 2-Phenylpropionyl chloride (253 mg, 1.5 mmol)
was dissolved in THF (3 mL), and the solution was cooled to 0 ꢂC.
(g) J. Douglas, J.E. Taylor, G. Churchill, A.M.Z. Slawin, A.D. Smith, J. Org. Chem.
10

A.A. Ibrahim, P.-H. Wei, G.D. Harzmann et al.
Tetrahedron 78 (2021) 131838
78 (2013) 3925e3938.
(c) R.J. Duffy, K.A. Morris, R. Vallakati, W. Zhang, D. Romo, J. Org. Chem. 74
(2009) 4772e4781.
[14] H. Lv, Y.-R. Zhang, X.-L. Huang, S. Ye, Adv. Synth. Catal. 350 (2008)
2715e2718.
[4] (a) J.C. Sauer, J. Am. Chem. Soc. 69 (1947) 2444e2448;
(b) R.H. Hasek, R.D. Clark, E.U. Elam, J.C. Martin, J. Org. Chem. 27 (1962)
60e64;
(c) E.U. Elam, J. Org. Chem. 32 (1967) 215e216;
[15] R.P. Wurz, G.C. Fu, J. Am. Chem. Soc. 127 (2005) 12234e12235.
[16] (a) K. Rasmussen, D.O. Calligaro, J.F. Czachura, L.J. Dreshfield-Ahmad,
D.C. Evans, S.K. Hemrick-Luecke, M.J. Kallman, W.T. Kendrick, J.D. Leander,
D.L. Nelson, C.D. Overshiner, D.B. Wainscott, M.C. Wolff, D.T. Wong,
T.A. Branchek, J.M. Zgombick, Y.-C. Xu, J. Pharmacol. Exp. Therapeut. 294
(2000) 688e700;
(d) W.G. Bentrude, W.D. Johnson, J. Am. Chem. Soc. 90 (1968) 5924e5926.
[5] (a) M.A. Calter, J. Org. Chem. 61 (1996) 8006e8007;
(b) M.A. Calter, R.K. Orr, Org. Lett. 5 (2003) 4745e4748;
(c) V.C. Purohit, R.D. Richardson, J.W. Smith, D. Romo, J. Org. Chem. 71 (2006)
4549e4558.
[6] (a) A.A. Ibrahim, D. Nalla, M. Van Raaphorst, N.J. Kerrigan, J. Am. Chem. Soc.
134 (2012) 2942e2945;
(b) A. Huang, J.J. Kodanko, L.E. Overman, J. Am. Chem. Soc. 126 (2004)
14043e14053.
(b) S. Chen, A.A. Ibrahim, N.J. Peraino, D. Nalla, M. Mondal, M. Van Raaphorst,
N.J. Kerrigan, J. Org. Chem. 81 (2016) 7824e7837;
(c) B. Bhaskararao, G. Jindal, R.B. Sunoj, J. Org. Chem. 82 (2017) 13449e13458.
[7] H. Pracejus, Liebigs Ann. Chem. 634 (1960) 9e22.
[8] H. Wynberg, E.G.J. Staring, J. Am. Chem. Soc. 104 (1982) 166e168.
[9] (a) R.K. Orr, M.A. Calter, Tetrahedron 59 (2003) 3545e3565;
(b) D.H. Paull, A. Weatherwax, T. Lectka, Tetrahedron 65 (2009) 6771e6803;
(c) A.D. Allen, T.T. Tidwell, Arkivoc (part(i)) (2016) 415e490;
(d) H. Guo, Y.C. Fan, Z. Sun, Y. Wu, O. Kwon, Chem. Rev. 118 (2018)
10049e10293.
[17] Azaferrocenes possessing planar chirality have been extensively used to
catalyze asymmetric reactions of disubstituted ketenes, see: (a) B.L. Hodous,
G.C. Fu, J. Am. Chem. Soc. 124 (2002) 1578e1579.(b) ref 3b.
(c) ref 10b.
[18] Josiphos Derivatives Are Available from Sigma Aldrich and from Strem.
[19] (a) E. Vedejs, S.T. Diver, J. Am. Chem. Soc. 115 (1993) 3358e3359;
(b) R.-H. Fan, X.-L. Hou, Tetrahedron Lett. 44 (2003) 4411e4413.
[20] S. Chen, M. Mondal, A.A. Ibrahim, K.A. Wheeler, N.J. Kerrigan, J. Org. Chem. 79
(2014) 4920e4929.
[21] X.-F. Zhu, C.E. Henry, O. Kwon, J. Am. Chem. Soc. 129 (2007) 6722e6723.
[22] (a) H.R. Hudson, K.B. Dillon, B.J. Walker, 31P NMR data of four coordinate
phosphonium salts and betaines, in: J.C. Tebby (Ed.), Handbook of
Phosphorus-31 Nuclear Magnetic Resonance Data, CRC Press, Boca Raton, FL,
1991, pp. 181e226;
[10] (a) S. Chen, E.C. Salo, K.A. Wheeler, N.J. Kerrigan, Org. Lett. 14 (2012)
1784e1787;
(b) E.C. Lee, B.L. Hodous, E. Bergin, C. Shih, G.C. Fu, J. Am. Chem. Soc. 127
(2005) 11586e11587.
[11] (a) A.A. Ibrahim, G.D. Harzmann, N.J. Kerrigan, J. Org. Chem. 74 (2009)
1777e1780;
(b) L.W. Dennis, V.J. Bartuska, G.E. Maciel, J. Am. Chem. Soc. 104 (1982)
230e235;
(b) A.A. Ibrahim, S.M. Smith, S. Henson, N.J. Kerrigan, Tetrahedron Lett. 50
(2009) 6919e6922;
(c) A.A. Ibrahim, P.-H. Wei, G.D. Harzmann, N.J. Kerrigan, J. Org. Chem. 75
(2010) 7901e7904;
(c) D.B. Chestnut, L.D. Quin, Tetrahedron 61 (2005) 12343e12349.
[23] W.L.F. Armarego, D.D. Perrin, Purification of Laboratory Chemicals, fourth ed.,
Butterworth Heinemann, 2002.
[24] For preparation of disubstituted ketenes (ketoketenes) see: (a) B.L. Hodous,
G.C. Fu, J. Am. Chem. Soc. 124 (2002) 10006e10007;
(d) P.-H. Wei, A.A. Ibrahim, M. Mondal, D. Nalla, G.D. Harzmann, F.A. Tedeschi,
K.A. Wheeler, N.J. Kerrigan, Tetrahedron Lett. 51 (2010) 6690e6694.
[12] M. Mondal, S. Chen, N. Othman, K.A. Wheeler, N.J. Kerrigan, J. Org. Chem. 80
(2015) 5789e5794.
(b) S.L. Wiskur, G.C. Fu, J. Am. Chem. Soc. 127 (2005) 6176e6177;
(c) A.D. Allen, L.M. Baigrie, L. Gong, T.T. Tidwell, Can. J. Chem. 69 (1991)
138e145;
[13] (a) M.A. Calter, W. Song, J. Zhou, J. Org. Chem. 69 (2004) 1270e1275;
(b) M.A. Calter, W. Liao, J. Am. Chem. Soc. 124 (2002) 13127e13129;
(d) J.E. Wilson, G.C. Fu, Angew. Chem. Int. Ed. 43 (2004) 6358e6360.
11