A survey of sulfide ligands for allylic C-H oxidations of terminal olefins

Article abstract of DOI:10.1002/chem.201301787

Perfect to a THT! Screening a diverse library of thioether ligands led to the discovery of tetrahydrothiophene (THT) as a highly reactive and selective ligand for Pd-catalyzed allylic C-H oxidation reactions (see scheme). This novel ligand system provides some of the highest reported yields for the formation of (E)-linear allylic acetates through allylic C-H activation chemistry (BQ = 1,4-benzoquinone). Copyright

Full text of DOI:10.1002/chem.201301787

COMMUNICATION
DOI: 10.1002/chem.201301787
À
A Survey of Sulfide Ligands for Allylic C H Oxidations of Terminal Olefins
Chi “Chip” Le, Kamala Kunchithapatham, William H. Henderson,
Christopher T. Check, and James P. Stambuli*[a]
À
The allylic C H oxidation of terminal olefins with palladi-
um catalysts is a useful method to functionalize small mole-
cules.[1] The recent use of catalytic palladium is an advance-
ment over earlier allylic oxidations that employed selenium
or mercury.[2] Although great strides have been made to
reduce the quantity of metal used in the initial catalytic re-
actions, there are still many unmet challenges that lessen the
Scheme 2. Ligands previously employed in the allylic oxidation of termi-
utility of this chemistry: high catalyst loading, regio-, diaster-
eo- and enantioselectivity, low reaction yields, and function-
al group tolerance.
nal olefins.
Rappoport and co-workers originally reported the use of
stoichiometric PdACHTUNGTRENNUNG(OAc)2 with terminal alkenes in acetic acid
reactions, an ideal catalyst for this transformation is still un-
discovered.
to yield the vinylic acetate as the major product, or the allyl-
ic acetate as the major product when DMSO was employed
as a co-solvent.[3] The use of DMSO demonstrated the abili-
ty of weakly coordinating ancillary ligands to control the
product formation and regioselectivity of the allylic acetate
(Scheme 1). Following Rappoportꢀs work, several groups
Although sulfoxides are well known ligands for transiton
metals,[8] sulfides are less prevalent.[9] Our group reported
the first examples of allylic oxidation reactions of terminal
olefins using only 5 mol% of a catalyst consisting of Pd-
AHCTUNGTRENNUNG
(OAc)2 and ligand 4.[4] In that same paper, we reported the
first palladium-catalyzed allylic oxidation reactions of termi-
nal alkenes with oxygen as the super stoichiometric oxidant.
The success of our unusual palladium catalyst prompted us
to survey a library of sulfide ligands in order to improve re-
action yields and selectivities, and better understand what
properties of the sulfide ligand allow an active and selective
catalyst. This information could then be utilized to rationally
design the next generation palladium–sulfide catalysts for al-
lylic oxidation reactions.
The sulfide ligands screens were conducted using 1-dode-
cene (7) as the test substrate. Our novel mixed sulfide ether
ligand, 4, yielded 52% of the linear allylic acetate under our
previously optimized reaction conditions: 5 mol% Pd-
Scheme 1. The effect of DMSO on allylic oxidations of terminal olefins
reported by Rappoport.
AHCTUNGRTEG(NNUN OAc)2, 5 mol% ligand, and 2 equivalents of BQ (benzoqui-
have combined various ligands with palladium salts to gen-
erate catalysts that promote allylic oxidation reactions.
Many different classes of ligands have been employed in-
cluding: sulfides,[4] sulfoxides,[5] pincer ligands,[6] and aromat-
ic amines[7] (Scheme 2). Despite the great advances made
since the first report of palladium-promoted allylic oxidation
none) in AcOH with dodecene (Table 1, entry 1). To probe
the importance of the oxygen atom in 4, the ether was re-
placed with a methylene linker. (Table 1, entries 2–4) show
that the oxygen atom in ligand 4 is not required for activity.
Furthermore, the identity of the alkyl group did not drasti-
cally change the outcome of the reaction until the trifluoro-
methylethyl group (14) was added to the ligand scaffold
(7% yield, entry 5). The decreased conversion when com-
[a] C. Le, K. Kunchithapatham, W. H. Henderson, C. T. Check,
Prof. J. P. Stambuli
bining ligand 14 with PdACTHNUTRGNE(NUG OAc)2 is likely caused by the de-
Department of Chemistry and Biochemistry
The Ohio State University
Evans Laboratory, 100 W. 18th Avenue, Columbus, OH 43210 (USA)
Fax : (+1)614-292-1685
creased coordination of the less basic ligand to the palladi-
um center. The decreased coordination of the external sul-
fide ligand would cause the palladium acetate and other
generated palladium salts within the reaction, to precipitate
from solution at a greater rate than in the presence of a
better coordinating ligand. Methylphenyl sulfide provided
E-mail: stambuli@chemistry.ohio-state.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301787.
Chem. Eur. J. 2013, 19, 11153 – 11157
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11153

Table 1. Alkyl–aryl sulfides as ligands in allylic oxidations of terminal olefins.[a,b]
the vinyl acetate product as well. Modifiying both
aromatic rings on the ligand (entries 11 and 12) did
not provide increased yields of allylic acetate. Re-
placing one of the phenyl groups with an indolyl
group (entry 13) also had no positive effect on the
reaction. Overall, the diaryl sulfides examined in
Table 2 were found to be poor ligands for the regio-
selective acetoxylation of terminal olefins, since
most all yields of linear (E)-allylic acetate were less
than 40% of the desired product. The reason for
the low yields was typically not the conversion,
rather the increased selectivity for the Wacker
product (vinyl acetate, 10) rather than the allylic
acetate products. This finding is in contrast with the
alkylaryl sulfide ligands from Table 1, which always
produced the (E)-linear allylic acetate as the major
product. This drastic change in selectivity is likely
caused by diminished coordination of the ligand to
the metal center, which allows for the vinylic ace-
tate to be formed through a competing process, sim-
ilar to the reaction reported by Rappoport.[3]
The relationship between the percent yield of
linear allylic aceate product and Hammett s values
was plotted (Figure 1). The implied trend in this
plot shows that electronics on the aromatic ring of
the ligand appear to correlate with the yield and se-
lectivity of the reaction. In general, electron-rich ar-
omatics on sulfur provided higher yields of linear
allylic acetate than electron-withdrawing aromatic
groups on sulfur. One major outlier, the electron-
withdrawing cyano group, produced greater yields
of the linear allylic acetate than expected likely be-
cause of coordination of the nitrile to palladium.
Ligand
Conv. [%]
85
8/9/10
E/Z
8a [%]
1
2
3
4
5
6
7
4
11
12
13
14
15
16
22:4:1
6:1
52
85
80
80
39
84
93
22:1:4
16:1:3
24:1:3
2:1:7
8:1
8:1
59
53
61
7
10:1
1:1
16:1:2
18:1:2
9:1
59
72
8:1
8
9
17
18
19
20
21
87
76
89
18
57
24:1:3
25:1:5
21:1:4
–
9:1
8:1
9:1
–
71
59
76
–
10
11
12
7:0:1
9:1
52
[a] Conditions: alkene
7 (1.0 equiv), PdACHTUNTGRENUGN(OAc)2 (5 mol%), ligand (5 mol%), BQ
(2 equiv), in AcOH (1.3m) at 408C. [b] Yields and selectivities determined by gas chro-
matography.
an active catalyst to promote the reaction similar to previ-
ous non-fluorinated alkyls (entry 6). Modification of the
electronics on the aromatic ring of the methylphenyl sulfide
(15) ligand showed that the more electron-rich aromatic
groups increase the reaction yield approximately 10% with
respect to an unsubstituted or electron-withdrawing aromat-
ic group on the ligand (entries 7–9). Replacing the methyl
group (entry 7) with benzyl had little effect on the reaction
outcome (entry 10). Finally, lower conversions were ob-
served when ortho-acetyl thioanisole and ethylnaphthyl sul-
fide were used as ligands (entries 11 and 12).
The next class of sulfide ligands under investigation were
the diaryl sulfides. Diphenyl sulfide (22) provided good con-
version, but also a large amount of the vinyl acetate com-
pound (10; Table 2, entry 1). Modification of the aromatic
ring did not decrease the yield of the vinyl acetate product.
For example, para- or meta-methoxy substituted aromatic
rings (entries 2 and 3) did not greatly affect the reaction out-
come. A bulkier aromatic group containing 2,6-dimethyl
substitution on one of the rings decreased the yield of the
desired linear allylic acetate even further (8; entry 4). Other
electronically modified ligands (entries 5–10) mostly favored
Figure 1. Plot of percent yield of linear allylic acetate versus Hammett s
values using ligands 22–24, 26–30 and yields from Table 2.
11154
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 11153 – 11157

À
Allylic C H Oxidations of Terminal Olefins
COMMUNICATION
Table 2. Diaryl sulfides as ligands in allylic oxidations of terminal olefins.[a,b]
and the corresponding ligands were tested. An elec-
tron-donating aromatic, such as anisole, lowered
the conversion and overall reaction yield (entry 2)
with respect to the unsubstituted bis-sulfide ligand
(entry 1). However, the use of p-trifluoromethylat-
ed aromatic (entry 3) generated a catalyst that pro-
duced a similar yield of (E)-allylic acetate 8a to
bis-sulfide 36. Naphthyl groups were installed in
place of the phenyl groups on sulfur (38), which
lowered the conversion and yield of allylic acetate
(entry 4). The effect of the methylene backbone be-
tween the two sulfur atoms was also investigated.
The one-carbon linker ligand 39 showed good selec-
tivity for the linear allylic acetate (19:1:3 of 8/9/10);
however, the diastereoselectivity was poor (6:1;
entry 5). In comparison, the four-carbon linker (40)
showed lower selectivity for the linear allylic ace-
tate (8), but higher diastereoselectivity, and an over-
all high yield for the desired linear allylic acetate 8
in 85% yield. Replacing the aromatic groups at
sulfur to benzyl groups proved costly (41), as the re-
action failed to give an appreciable conversion
(entry 7). As a general trend, the diminished con-
versions likely emanate from the overly strong co-
ordination of the bidentate sulfide to the metal
center, inhibiting the reaction from occurring.
While the longer alkyl-chain ligand was the best
bis-sulfide ligand for the formation of linear allylic
acetates, it is still inferior to the monodentate alkyl
aryl sulfides previously screened because of the re-
quirement for double the catalyst loading. The ne-
cessity for this higher loading is likely due to the
longer reaction times required for bidentate sul-
fides. The longer the reaction times, the more likely
the palladium salts will be deactivated by precipita-
tion from the reaction. Hence, additional catalyst is
required to compensate for the loss in palladium.
Ligand
Conv. [%]
86
8/9/10
E/Z
8a [%]
1
2
3
22
23
24
6:1:5
8:1
28
95
89
9:1:4
6:1:7
4:1
5:1
40
25
4
25
86
2:1:6
6:1
14
5
6
26
27
89
73
4:1:5
2:1:7
3:1
22
6
10:1
7
8
9
28
29
30
91
85
83
5:1:5
3:1:7
2:1:8
5:1
4:1
6:1
19
11
11
10
11
12
31
32
33
54
83
95
5:1:9
2:1:6
7:1:3
2:1
4:1
6:1
9
12
58
13
34
88
2:1:3
6:1
10
[a] Conditions: alkene
7 (1.0 equiv), PdACHTUNTGRENUGN(OAc)2 (5 mol%), ligand (5 mol%), BQ
(2 equiv), in AcOH (1.3m) at 408C. [b] Yields and selectivities determined by gas chro-
matography.
Although palladium-catalyzed allylic oxidation re-
actions of terminal olefins have not worked particu-
larly well with more basic ligands, such as phos-
phines, it was clear that the diaryl sulfides were not
basic enough under these reaction conditions to
promote a selective and high-yielding amount of
linear allylic acetate product.
The increased selectivity and yield obtained from
electron-rich sulfides was believed to come from
the increased coordination to palladium. Therefore,
we investigated the use of bidentate sulfides that
would increase the binding to the palladium center,
which may increase the selectivity of the reaction.
The use of 1,2-bis(phenylthio)ethane as a ligand
greatly decreased the amount of vinyl acetate prod-
uct (10), but lower conversions to the desired prod-
uct were obtained than with ligand 4 (Table 3,
entry 1). Modifications of the electronics on the ar-
Table 3. Bidentate sulfides as ligands in allylic oxidations of terminal olefins.[a,b]
Ligand
Conv. [%]
78
8/9/10
E/Z
8a [%]
1
2
35
36
24:2:1
9:1
72
62
88
52
11:1:1
12:1:2
14:1:2
17:1
12:1
11:1
49
72
47
3
4
37
38
5
6
39
40
90
95
19:1:3
11:1:1
6:1
65
85
10:1
7
41
14
–
–
–
[a] Conditions: alkene 7 (1.0 equiv), Pd
ACHTUNGRTEN(NUNG OAc)2 (10 mol%), ligand (10 mol%), BQ
(2 equiv), in AcOH (1.3m) at 408C. [b] Yields and selectivities determined by gas chro-
omatic groups on the bis-sulfide ligand were made matography.
Chem. Eur. J. 2013, 19, 11153 – 11157
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11155

J. P. Stambuli et al.
Table 5. Substrate scope of allylic oxidation reactions with tetrahydro-
thiophene as the ligand.
The majority of the aryl–alkyl sulfide ligands provided
good conversions and selectivity for the linear allylic acetate
product. Alternatively, the diaryl sulfide ligands usually pro-
vided the vinyl acetate as the major product. Finally, the bi-
dentate sulfide ligands typically provided the linear allylic
acetate as the major reaction product, but the conversions
were lower and the reaction required longer times and
higher catalyst loadings. The order of binding to the metal
center of these classes of sulfide ligands should go from bi-
dentate sulfides> aryl–alkyl sulfides> diaryl sulfides. There-
fore, it was hypothesized that the dialkyl sulfides, which
should bind more strongly to the palladium center than
aryl–alkyl sulfides but more weakly than the bidentate sul-
fide ligands, may provide a selective and active palladium
catalyst for these allylic oxidation reactions.
Substrate
Product
t [h]
Yield [%]
87
1
2
7
8a
48
47
57
24
80
3
4
48
49
58
59
20
12
77
79
5
6
50
51
60
61
18
16
61
48
7
8
52
53
62
63
16
16
85
82
Several dialkyl sulfides were screened for activity as li-
gands in allylic oxidation reactions (Table 4). Dimethyl sul-
fide was active in this chemistry when combined with palla-
dium acetate, and provided the allylic acetate product 8a in
62% yield (entry 1). Increasing the size of the alkyl groups
to diethylsulfide (entry 2) slightly increased the yield and
conversion compared to dimethylsuflide. However, increas-
ing the steric of the alkyl group to the di-isobutyl sulfide
drastically lowered the selectivity of the reaction by produc-
ing the linear allylic acetate and vinyl acetate in a 1.2:1
9
54
55
64
65
12
24
71
76
10
use of commercially available tetrahydrothiophene
as ligand and palladium acetate at 5 mol% loading
provided some of the highest yields reported to
date for the highly selective formation of the (E)-
linear allylic acetate compounds. The isolated yields
in the table are for the (E)-linear allylic acetate
Table 4. Dialkyl sulfides as ligands in allylic oxidations of terminal olefins.[a,b]
Ligand
Conv. [%]
8/9/10
E/Z
8a [%]
1
2
42
43
88
98
10:1:3
11:1:2
8:1
10:1
62
70
À
product only (8a). Dodecene was C H activated
and acetoxylated in 87% isolated yield over 48 h
(entry 1). Allylbenzene and 1-allyl-4-fluorobenzene
were excellent substrates as the corresponding ace-
toxylated products in 80 and 77% yield, respective-
ly (entries 2 and 3). Olefins appended with a silyl
ether or a benzyl ether reacted to provide 61% and
82% of the corresponding products (entries 5 and
6). Esters were also tolerated and gave high yields
as the para-methoxybenzyl ester and the benzoate
appended olefins gave 48 and 85% of the acetoxy-
3
4
44
45
85
20
6:1:5
–
11:1
–
28
–
5
46
95
20:1:1
10:1
88
[a] Conditions: alkene
(2 equiv), in AcOH (1.3m) at 408C. [b] Yields and selectivities determined by gas chro-
matography.
7 (1.0 equiv), PdACHTUNTGRENUGN(OAc)2 (5 mol%), ligand (5 mol%), BQ
ratio. Increasing the ligand sterics even further to the di-tert-
buylsulfide generate a poorly active catalyst, which provided
20% conversion of a mixture of products. From these re-
sults, it appeared that steric hindrance had a pronounced
affect on product formation with dialkyl sulfide ligands.
Therefore, the cyclic dialkylsulfide, tetrahydrothiophene
(THT, 46) was tested. Using tetrahydrothiophene with Pd-
lated products in high yields (entries 7 and 8). Finally, the
pyrrolidine dione compounds were reacted and provided 71
and 76% of the corresponding acetoxylated compounds
under these reaction conditions.
In summary, a library of electronically and sterically di-
verse mono- and bidentate sulfide ligands were combined
with palladium acetate and examined in the allylic acetoxy-
lation of terminal olefins. In general, aryl–alkyl sulfides pro-
vided good yields and selectivities for the linear allylic ace-
tate products, while diaryl sulfides typically promoted the
near equivalent formation of vinyl acetate and linear allylic
acetate under the reaction conditions. Bidentate sulfide li-
gands favored formation of the linear allylic acetate, but re-
quired higher catalyst loadings and longer reactions times
ACHTUNGTRENNUNG(OAc)2 at only 5 mol% catalyst loading gave the highest
yield of linear allylic acetate product (8a) at 88% (entry 5).
Overall, the dialkyl sulfides led to greater product formation
than aryl–alkyl sulfides and diaryl sulfides, which is pre-
sumed to emanate from increased basicity on sulfur.
With this new reactive and selective catalyst in hand, the
terminal olefin substrate scope was examined (Table 5). The
11156
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 11153 – 11157

À
Allylic C H Oxidations of Terminal Olefins
COMMUNICATION
[2] a) M. A. Umbreit, K. B. Sharpless, J. Am. Chem. Soc. 1977, 99, 5526;
b) Z. Rappoport, P. D. Sleezer, S. Winstein, W. G. Young, Tetrahedron
Lett. 1965, 6, 3719.
[3] W. Kitching, Z. Rappoport, S. Winstein, W. G. Young, J. Am. Chem.
Soc. 1966, 88, 2054.
[4] W. H. Henderson, C. T. Check, N. Proust, J. P. Stambuli, Org. Lett.
2010, 12, 824–827.
[5] a) M. S. Chen, M. C. White, J. Am. Chem. Soc. 2004, 126, 1346;
b) M. S. Chen, N. Prabagaran, N. A. Labenz, M. C. White, J. Am.
Chem. Soc. 2005, 127, 6970.
than the monodentate sulfides. Dialkyl sulfides generated
the most active and selective catalysts under these reaction
conditions. From this group of dialkyl sulfides, we discov-
ered that simple and inexpensive tetrahydrothiophene pro-
vided one of the most active and selective catalysts for allyl-
ic oxidation reactions of terminal olefins.
Acknowledgements
[6] a) J. Aydin, K. J. Szabo, Org. Lett. 2008, 10, 2881–2884; b) L. T. Pilar-
ski, N. Selander, D. Bose, K. J. Szabo, Org. Lett. 2009, 11, 5518–5521.
[7] a) A. N. Campbell, P. B. White, I. A. Guzei, S. S. Stahl, J. Am. Chem.
Soc. 2010, 132, 15116–15119; b) B. L. Lin, J. A. Labinger, J. E.
Bercaw, Can. J. Chem. 2009, 87, 264.
The authors thank The Ohio State University for funding and the Ohio
BioProduct Innovation Consortium for mass spectrometry support.
[8] M. Calligaris, Coord. Chem. Rev. 2004, 248, 351–375.
[9] M. R. Dubois, Chem. Rev. 1989, 89, 1–9.
Keywords: allylic compounds · ligand effects · oxidation ·
Received: May 9, 2013
Published online: July 9, 2013
sulfur heterocycles
[1] a) E. M. Stang, M. C. White, Nat. Chem. 2009, 1, 547; b) E. M. Stang,
M. C. White, Angew. Chem. 2011, 123, 2142; Angew. Chem. Int. Ed.
2011, 50, 2094.
Chem. Eur. J. 2013, 19, 11153 – 11157
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11157