Electroreduction of triphenylphosphine oxide to triphenylphosphine in the presence of chlorotrimethylsilane

Article abstract of DOI:10.1055/s-0031-1289612

Electroreduction of triphenylphosphine oxide to triphenylphosphine in an acetonitrile solution of tetrabutylammonium bromide in the presence of chlorotrimethylsilane was performed successfully in an undivided cell fitted with a zinc anode and a platinum cathode under constant current. A plausible mechanism involving, (1) one-electron reduction of triphenylphosphine oxide generating the corresponding anion radical [Ph₃P-O^-], (2) subsequent reaction with chlorotrimethylsilane affording the (trimethylsiloxy)triphenylphosphorus radical [Ph₃P-OSiMe ₃], and (3) further one-electron reduction followed by P-O bond fission leading to triphenylphosphine is proposed. In a similar manner, electroreduction of some triarylphosphine oxides and alkyldiarylphosphine oxides was executed to give the corresponding phosphine derivatives in good to moderate yields. Georg Thieme Verlag Stuttgart · New York.

Full text of DOI:10.1055/s-0031-1289612

PAPER
4091
Electroreduction of Triphenylphosphine Oxide to Triphenylphosphine in the
Presence of Chlorotrimethylsilane
a
b
b
b
b
E
H
lectroreduction
o
i
f
Triph
r
e
nylphos
o
phine
O
xide
m
to Triphenylphosph
u
ine Kawakubo, Manabu Kuroboshi,* Tomotake Yano, Kazuma Kobayashi, Syogo Kamenoue,
b
b
Tomomi Akagi, Hideo Tanaka*
a
API Department, Asahi Kasei Chemicals Corporation, Kanda Jinbocho 1-105, Chiyoda-ku, Tokyo 101-8101, Japan
The Graduate School of Natural Science and Technology, Okayama University, Tsushima-Naka 3-1-1, Okayama 700-8530, Japan
Fax +81(86)2518079; E-mail: tanaka95@cc.okayama-u.ac.jp
b
Received 11 August 2011; revised 21 September 2011
Reduction of triphenylphosphine oxide (2a) to triphe-
nylphosphine (1a) (Scheme 1, path A) has been intensive-
Abstract: Electroreduction of triphenylphosphine oxide to triphe-
nylphosphine in an acetonitrile solution of tetrabutylammonium
bromide in the presence of chlorotrimethylsilane was performed
ly investigated, and performed successfully with various
8
successfully in an undivided cell fitted with a zinc anode and a plat- reductants, such as metal hydrides (hydrosilanes, alumi-
9
10
inum cathode under constant current. A plausible mechanism in- num hydrides ), low-valent metals (SmI /HMPA,
2
1
1
volving, (1) one-electron reduction of triphenylphosphine oxide
TiCp Cl /Mg ), and organic reductants [hydrocarbon/ac-
2
2
·
–
generating the corresponding anion radical [Ph P –O ], (2) subse-
12
13
3
tivated carbon, (Et N) P/POCl ]. The P=O bond of 2a
2 3 3
quent reaction with chlorotrimethylsilane affording the (trimethyl-
is sufficiently strong (ca. 500 kJ/mol) that highly reactive
reductants are required to cleave it efficiently. The report-
ed procedures to date are not practical, since they always
·
siloxy)triphenylphosphorus radical [Ph P –OSiMe ], and (3) further
3
3
one-electron reduction followed by P–O bond fission leading to
triphenylphosphine is proposed. In a similar manner, electroreduc-
tion of some triarylphosphine oxides and alkyldiarylphosphine ox- required more than a stoichiometric amount of the reduc-
ides was executed to give the corresponding phosphine derivatives tant, and this generally expensive, explosive, and/or not
in good to moderate yields.
easy to handle.
Key words: reduction, electron transfer, phosphorus, silicon
As an alternative access to triphenylphosphine (1a), chlo-
rination of triphenylphosphine oxide (2a) leading to tri-
phenylphosphorus dichloride (3a) and subsequent reduction
In modern organic synthesis, triphenylphosphine (1a) is of 3a to 1a has been investigated thus far (Scheme 1, path
an important reagent for various organic reactions, such as B). Phosphorus dichloride 3a is easily prepared by treat-
1
2
the Wittig reaction, Mitsunobu reaction, Mukaiyama– ment of 2a with chlorinating reagents such as phosgene
3
4
14,15
Corey lactonization, Appel reaction, Staudinger reac- and oxalyl chloride.
It is reasonable to assume that re-
5
tion, and so on. In these reactions, however, triphe- duction of the P–Cl bond of 3a proceeds more smoothly
nylphosphine (1a) is turned into triphenylphosphine oxide than that of the P=O bond of 2a, since the P–Cl bond (ca.
16
(
2a), which is a stable and flame-resistant chemical; as a 310 kJ/mol) is far weaker than the P=O bond. Indeed, re-
result, a significant amount of 2a has been stored as a trou- duction of 3a to 1a was performed by hydrogenation with
blesome waste. Moreover, phosphorus is an exhaustible transition-metal catalysts (Pt, Rh, and Pd),1 and reduc-
7,18
6
19
20
resource. Therefore, facile methods for the reduction of tion with several metals such as sodium, aluminum,
2
1
22
2
a to 1a have been in great demand from the view points silicon, and iron. The indirect procedure is, however,
of waste treatment and the recycling of phosphorus re- not necessarily satisfactory for practical use since the re-
7
sources (Scheme 1).
quired chlorinating reagents are highly toxic and have to
be handled with special care.
Wittig reaction, Mitsunobu reaction, etc.
Electroreduction is performed by electron-uptake on a
cathode and subsequent chemical reaction without any re-
ductants, thereby offering an inexpensive and environ-
mentally benign procedure for modern organic
reactions.23 The electroreduction of triphenylphosphine
Al-H,
Si-H, etc.
Ph
Ph P:
Ph
Ph
Ph
P
O
[
path A]
Ph
1
2
4
oxide (2a) has been previously reported, but P–C bond
cleavage occurred predominantly to give diphenylphos-
phine oxide (4a), diphenylphosphine (5a), phenylphos-
phine oxide, benzene, etc., with no detectable amount of
a
2a
Ph
2 e–
Cl
Cl
(COCl)2
+
Ph
Ph
P
3
a
[
path B]
25
the desired product 1a (Scheme 2). It is likely that the
Scheme 1 Recycling of triphenylphosphine (1a)
P=O bond is sufficiently strong that P–C bond fission
mainly occurs to give 4a, benzene, etc.
Recently, we have reported that the electroreduction of
triphenylphosphorus dichloride (3a) to 1a proceeds
smoothly in the presence of aluminum trichloride wherein
SYNTHESIS 2011, No. 24, pp 4091–4098
x
x.
x
x
.
2
0
1
1
Advanced online publication: 18.11.2011
DOI: 10.1055/s-0031-1289612; Art ID: F78211SS
Georg Thieme Verlag Stuttgart · New York
©

4
092
H. Kawakubo et al.
PAPER
+
–
t-BuOH, [Bu3NMe] OMs
ethylenediamine
phines 1. A plausible reaction mechanism involving one-
electron reduction of 2a followed by reaction with chlo-
rotrimethylsilane leading to tetracoordinate phosphorus
radical intermediate 8 and subsequent further one-
electron reduction resulting in the P–O bond fission is dis-
cussed.
Ph
Ph
P
Ph
P
O
Ph
(
GC)-(GC), undivided cell
CPE (50 V), 4 F/mol
Ph
Ph
2a
1
a (not detected)
Ph
P
Ph
P
Ph
P
3
1
A typical procedure is as follows. An acetonitrile (5 mL)
solution of triphenylphosphine oxide (2a, 0.5 mmol),
chlorotrimethylsilane (1.5 mmol), and tetrabutylammoni-
H
O
+
H
+
H
O
+ benzene (25%)
Ph
Ph
H
cyclohexadiene (35%)
cyclohexene (1%)
+
+
32
4a (2%)
5a (4%)
(8%)
um bromide (1 mmol) was placed in the cathodic cham-
ber of a divided cell (Figure 1), and an acetonitrile (5 mL)
solution of tetrabutylammonium bromide (1 mmol) was
Scheme 2 Electroreduction of triphenylphosphine oxide (2a)
2
placed in the anodic chamber; a platinum (1.0 × 1.5 cm )
tetracoordinate phosphonium species 6a is generated in
2
cathode and a zinc (1.0 × 2.0 cm ) anode, respectively,
situ (Scheme 3, route A). It is reasonable to assume that
were immersed in these solutions. A constant current (50
+
tetracoordinate phosphonium [Ph P–Cl] is positively
3
mA) was supplied at room temperature and 3 F/mol-2a of
charged (more electron deficient than 3a) and efficiently
undergoes two-electron uptake at the cathode to give 1a.
One-pot transformation of 2a to 1a through reaction of 2a
31
electricity was passed. P NMR of the catholyte showed
that triphenylphosphine (1a) was obtained in 80% yield
(
72% isolated yield: silica gel column chromatography)
2
6
with oxalyl chloride leading to 3a and subsequent
31
together with a small amount of byproducts 4a ( P NMR
ratio: 1%) and 5a (9%) and recovered 2a (10%) (Table 1,
entry 1). When a platinum anode was used, the electrore-
duction of 2a similarly proceeded to give 1a in 60% yield
2
7
28
chemical and electrochemical reduction has been suc-
cessfully achieved. We have also reported that the similar
one-pot transformation of 2a to 1a is performed by elec-
troreduction of a mixture of 2a and trimethylsilyl triflate
via tetracoordinate phosphonium intermediates 7a
(
entry 2).
2
9
(
Scheme 3, route B).
a. divided cell
b. undivided cell
(
COCl)2
AlCl3
+ 2 e–
+
–
Ph3PCl2
[Ph3PCl] [AlCl4]
route A
3a
6a
Ph3P=O
Ph3P
2a
1a
Me3Si-OTf
+
2 e–
+
–
[
Ph3P-OSiMe3] [OTf]
route B
Ph P=O (2a)
3
7a
Bu4NBr
Me3SiCl
MeCN
Scheme 3 Electroreduction of triphenylphosphine oxide (2a) to tri-
phenylphosphine (1a) via 6a and/or 7a
In our continuing study on the reduction of triphenylphos-
phine oxide (2a) to triphenylphosphine (1a), we found
that electroreduction of 2a proceeded smoothly in the
presence of chlorotrimethylsilane to give 1a in good yield
Bu NBr
4
glass filter
MeCN
3
0
(
Scheme 4). Some details of this reaction was investi- Figure 1 Electrolysis cell
3
1
gated by means of P NMR and cyclic voltammetry
CV).
(
The electroreduction of triphenylphosphine oxide (2a) in
Herein is described the first practically significant proce- an undivided cell (Figure 1) was also performed efficient-
dure for direct electroreductive transformation of 2a to 1a ly by correct choice of the anode materials; thus, with
and application to the reduction of some triaryl- and zinc, magnesium, aluminum, and iron anodes, triphe-
alkyldiarylphosphine oxides 2 to the corresponding phos- nylphosphine (1a) was obtained in 69, 57, 53, and 38%
yields, respectively (Table 1, entries 3–6). These anodes
R1
R1
seem to work as a sacrificial anode, thus avoiding unde-
sired anodic reactions. On the other hand, with platinum,
nickel, tin, and stainless steel (SUS304) anodes, no appre-
ciable reduction of 2a was observed and the starting ma-
terial was recovered unchanged (entries 7–10).
+
2 e–
R2
P
R3
O
R2
P
R3
Me3SiCl, Bu4NBr, MeCN
2
(2 mmol)
R1
1
R2
P
O
SiMe3
R3
No significant effect of the cathode materials was ob-
served (Table 2). Thus the cathode materials examined,
e.g., platinum, SUS304, glassy carbon, copper, zinc, and
lead, gave 1a in 58–72% yields.
8
Scheme
phosphine oxides 2
4
Chlorotrimethylsilane-promoted electroreduction of
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

PAPER
Electroreduction of Triphenylphosphine Oxide to Triphenylphosphine
4093
Table 1 Electroreduction of Triphenylphosphine Oxide (2a) to Triphenylphosphine (1a); Effect of the Anodea
b
c
Entry
Cell
Anode
Current
Ratio (%)
Yield (%) of 1a
(
mA)
1a
2a
10
4a
1
5a
9
6
6
0
0
0
0
0
0
0
1d
2d
3d
4e
5e
6e
7e
8e
9e
divided
Zn
Pt
50
80
74
74
59
55
42
0
72
60
69
57
53
38
–
divided
50
16
16
1
undivided
undivided
undivided
undivided
undivided
undivided
undivided
undivided
Zn
Mg
Al
100
25
25
25
25
25
25
25
4
29
12
11
1
34
Fe
57
Pt
100
100
100
100
0
Ni
0
0
–
Sn
SUSf
0
0
–
1
0e
0
0
–
a
Conditions: 2a (0.5 mmol), Me SiCl (1.5–6 mmol), (anode)–(Pt), Bu NBr (1 mmol), MeCN (5 mL), divided or undivided cell, 25–100 mA,
3
4
2
–3 F/mol, r.t.
b
Determined by 31P NMR.
Isolated yields.
c
d
e
f
Me SiCl (1.5 mmol), 3 F/mol-2a.
3
Me SiCl (6 mmol), 2 F/mol-2a.
3
SUS304.
Table 2 Electroreduction of Triphenylphosphine Oxide (2a) to
Triphenylphosphine (1a); Effect of the Cathode
Table 3 Electroreduction of Triphenylphosphine Oxide (2a) to
Triphenylphosphine (1a); Effect of the Amount of Chlorotrimethyl-
silanea
a
b
c
Entry Cathode Ratio (%)
Yield (%) of 1a
b
c
Entry Me SiCl
Ratio (%)
Yield (%) of 1a
3
1
a
2a
28
27
31
22
23
16
4a
3
5a
3
(
mmol)
1a
2a
4a
78
48
16
4
5a
0
1
2
3
4
5
Pt
66
67
64
75
71
78
67
68
58
72
62
71
1
2
3
4
5
0
0
6
–
SUSd
GCe
Cu
4
2
0.5
1.0
1.5
3.0
6.0
9
38
15
16
17
14
3
–
3
1
66
74
72
72
5
70
69
68
67
2
1
6
Zn
4
2
2
5
6
Pb
2
2
6
1
7
a
Conditions: 2a (0.5 mmol), Me SiCl (6 mmol), (Zn)–(cathode),
a
3
Conditions: 2a (0.5 mmol), Me SiCl (0–6 mmol), (Zn)–(Pt),
3
Bu NBr (1 mmol), MeCN (5 mL), undivided cell, 100 mA, 2 F/mol,
4
Bu NBr (1 mmol), MeCN (5 mL), undivided cell, 100 mA, 3 F/mol,
4
r.t.
r.t.
b
Determined by 31P NMR.
Isolated yields.
SUS304.
Glassy carbon.
b
Determined by 31P NMR.
Isolated yields.
c
c
d
e
The product selectivity also varied with the intensity of
the electric current (Table 4). Triphenylphosphine (1a)
was obtained in good yields when the electrolysis was car-
ried out at 25–100 mA (entries 2–5). When the current
was reduced to 10 mA, a considerable amount of diphe-
nylphosphine oxide (4a) (P–C bond cleavage) was formed
and the yield of 1a reduced to 41% (entry 1). Upon in-
crease of the current up to 150–200 mA, the yield of 1a
decreased to 48–55% and significant amounts (40–45%)
of 2a were recovered (entries 5 and 6).
It is interesting to note that product selectivity was highly
dependent on the amount of chlorotrimethylsilane
(
Table 3). The desired product 1a (P=O bond cleavage)
was obtained as the major product when more than 1
mmol (2 equiv) of chlorotrimethylsilane was used (entries
3
–6). In contrast, diphenylphosphine oxide (4a) (P–C
bond cleavage) was obtained as the sole product when the
electroreduction of 2a was carried out in the absence of
chlorotrimethylsilane (entry 1).
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

4
094
H. Kawakubo et al.
PAPER
Table 4 Electroreduction of Triphenylphosphine Oxide (2a) to
Triphenylphosphine (1a); Effect of the Current
Table 6 Electroreduction of Triphenylphosphine Oxide (2a) to
a
Triphenylphosphine (1a); Effect of Additivesa
b
c
b
c
Entry Current
Ratio (%)
Yield (%) of 1a
Entry Additive
Ratio (%)
Yield (%) of 1a
(
mA)
1a
56
74
78
66
55
47
2a
12
13
14
28
40
45
4a
32
13
6
5a
0
1
a
2a
28
4a
3
5a
3
1
2
3
4
5
10
41
67
73
67
55
48
1
2
3
4
5
6
7
Me SiCl
66
67
65
–
3
25
0
t-BuMe SiCl 68
26
35
1
5
2
50
2
i-Pr SiCl
62
24
0
1
2
3
100
150
200
3
3
Me SiCl
76
0.5
0
0
–
2
2
3
0.6
2
TsCl
100
100
100
100
0
–
6
2
MsCl
BnCl
0
0
0
–
a
Conditions: 2a (0.5 mmol), Me SiCl (6 mmol), (Zn)–(Pt), Bu NBr
1 mmol), MeCN (5 mL), undivided cell, 10–200 mA, 2 F/mol, r.t.
Determined by P NMR.
Isolated yields.
3
4
0
0
0
–
(
b
31
8
Ac O
0
0
0
–
2
c
a
Conditions: 2a (0.5 mmol), additive (6 mmol), (Zn)–(Pt), Bu NBr (1
4
mmol), MeCN (5 mL), undivided cell, 100 mA, 2 F/mol, r.t.
Determined by P NMR.
Isolated yields.
The reduction of 2a to 1a is a two-electron reduction pro-
cess and, theoretically, requires 2 F/mol-2a of electricity.
After passage of 2 F/mol-2a of electricity (25 mA), 1a
was obtained in 67% yields (current efficiency 67%)
b
c
31
sumably because these reagents would be reduced more
easily than phosphine oxide 2a.
(
Table 5, entry 3). The yield of 1a increased up to 89%
when 3 F/mol of electricity was passed (entry 4). Further
increase of electricity up to 4 F/mol-2a resulted in no
more increase of 1a, but contamination with a small
amount of 5a (12%) which seems to form from over-
reduction of 4a (entry 5).
31
In the P NMR of 2a [Figure 2, (a)], a singlet peak was
observed at d = 27.3 (downfield from H PO ). When one
3
4
equivalent of trimethylsilyl triflate was added to triphe-
nylphosphine oxide (2a), the singlet peak shifted to d =
5
2.1 [Figure 2, (c)], indicating that trimethylsilyl triflate
Table 5 Electroreduction of Triphenylphosphine Oxide (2a) to
Triphenylphosphine (1a); Effect of the Amount of Electricitya
immediately reacted with 2a to form the tetracoordinate
+
–
phosphonium
species
[Ph P–OSiMe ] OTf
7a
3
3
(
Scheme 5).
b
c
Entry
Q
(
Ratio (%)
Yield (%) of 1a
F/mol) 1a
2a
86
63
13
6
4a
2
5a
0
Ph
Ph
P
OSiMe3 OTf– (1)
1
2
3
4
0.5
1
12
31
74
87
85
8
34
67
89
–
Ph
P
O + Me3Si-OTf
Ph
Ph
Ph
7a: δ 52.1
6
0
2
a: δ 27.3
2
13
3
0
Scheme 5
3
4
In contrast, addition of chlorotrimethylsilane to triphe-
nylphosphine oxide (2a) resulted in no appreciable
change in the chemical shifts in the 31P NMR of 2a
[Figure 2, (b)], suggesting that tetracoordinate phosphoni-
5
4
2
1
12
a
Conditions: 2a (0.5 mmol), Me SiCl (6 mmol), (Zn)–(Pt), Bu NBr
1 mmol), MeCN (5 mL), undivided cell, 25 mA, 0.5–4 F/mol, r.t.
Determined by P NMR.
3
4
(
b
31
+
–
33
um [Ph P–O–SiMe ] Cl was not formed. Similar re-
3
3
c
Isolated yields.
sults were obtained with tert-butylchlorodimethylsilane
d = 27.3 ppm), chlorotriisopropylsilane (d = 28.9 ppm),
(
As mentioned above, the presence of chlorotrimethylsi- and dichlorodimethylsilane (d = 31.9 ppm), indicating
lane is essential for the direct electroreduction of triphe- that these silylating reagents do not give the correspond-
ing tetracoordinate phosphonium species. No appreciable
nylphosphine oxide (2a) to triphenylphosphine (1a). In
place of chlorotrimethylsilane, tert-butylchlorodimethyl-
silane, and chlorotriisopropylsilane were also used to pro-
mote efficiently the electroreduction of 2a to give 1a in
31
shift in the P NMR of 2a was observed by addition of p-
tosyl chloride (d = 27.5 ppm), mesyl chloride (d = 27.7
ppm), and benzyl chloride (d = 27.5 ppm).
6
0–70% yields (Table 6, entries 2 and 3). Dichlorodi-
Cyclic voltammetry of 2a in an acetonitrile solution of tet-
rabutylammonium bromide (0.10 M) in the presence and
absence of chlorotrimethylsilane was investigated. The
cyclic voltammogram of 2a (Figure 3, green line) shows a
methylsilane was less effective (entry 4). Tosyl chloride,
mesyl chloride, benzyl chloride, and acetic anhydride did
not promote the electroreduction at all (entries 5–8), pre-
+
single, irreversible wave at –3.0 V vs. Ag/Ag . The cyclic
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

PAPER
Electroreduction of Triphenylphosphine Oxide to Triphenylphosphine
4095
2
a
+
e–
Ph
Ph
C–P bond cleavage
–
•
O–
Ph
P
Ph
Ph
P
O
O
1
1a
Ph
Ph •
9
a
Me3Si Cl
Ph
P
Ph
•
Ph
Ph
1
P
O
SiMe3
H
Ph
8a
4
a + benzene
3
1
Figure 2
P NMR of (a) Ph P=O, (b) Ph P=O/Me SiCl (1:1), (c)
3
3
3
Ph P=O/Me SiOTf (1:1) in MeCN/C D .
3
3
6
6
e–
+
5
a
Ph
–
P
+
cyclohexadiene
cyclohexene
voltammogram of chlorotrimethylsilane alone shows no
reduction wave in the range of –2.0 to –3.5 V. Addition of
chlorotrimethylsilane to the solution of 2a (purple line)
brought about no appreciable change of the redox poten-
tial of 2a, but a remarkable increase in the reduction cur-
rent.
+
O
SiMe3
Ph
1
0a
Me3Si Cl
P–O bond cleavage
a + Me3Si
SiMe3
O
Scheme 6 A plausible reaction mechanism
The present chlorotrimethylsilane-promoted electrore-
duction was successfully applied to the reduction of some
triarylphosphine oxides and alkyldiarylphosphine oxides
2
to the corresponding phosphine derivatives 1 in good to
moderate yields (Table 7). The electroreduction was car-
ried out at 45 °C. The reduction of triarylphosphine oxides
2
c–e proceeded smoothly to give the corresponding tri-
arylphosphines 1c–e in 80–96% yields after 4 F/mol of
electricity was passed (entries 3–5), whereas tri-2-
tolylphosphine oxide (1b) was reduced less effectively to
give 2b in rather low yield, presumably due to steric hin-
drance of the ortho-substituent (entry 2). The chlorotri-
methylsilane-promoted electroreduction of alkyldiaryl-
phosphine oxides 2f and 2g was also performed efficiently
to give the corresponding phosphines 1f and 1g in 79 and
Figure 3 Cyclic voltammogram of triphenylphosphine oxide with
chlorotrimethylsilane. [Working electrode, Pt; counter electrode, Pt
+
wire; reference electrode, Ag/Ag ; scan rate 200 mV/s; Bu NBr (0.10
4
M); MeCN. BG = background (without Ph PO and Me SiCl);
3
3
Me SiCl = Me SiCl (0.01 M); Ph PO = Ph PO (0.01 M); Ph PO +
3
3
3
3
3
Me SiCl = Ph PO (0.01 M) + Me SiCl (0.01 M)].
3
3
3
4
6% yields, respectively (entries 6 and 7). Further appli-
A plausible reaction mechanism is shown in Scheme 6. In
the initial stage of the electroreduction of 2a both in the
presence and absence of chlorotrimethylsilane, one-elec-
tron reduction would occur at the same potential (–3.0 V
cation to electroreduction of triphenyl phosphate (2h) was
not successful and no appreciable reduction product 2h
was obtained (entry 8).
+
·
–
Electroreduction of triphenylphosphine oxide (2a) to
triphenylphosphine (1a) was performed successfully by
passage of 2–3 F/mol of electricity through a solution of
vs. Ag/Ag ) to form an anion radical [Ph P –O ] 9a. In the
3
presence of chlorotrimethylsilane, 9a would be trapped to
·
give radical [Ph P –OSiMe ] 8a. Further one-electron re-
3
3
2
a and chlorotrimethylsilane (~3 equiv) in acetonitrile
duction of 8a leading to anion 10a, and subsequent P–O
–
containing tetrabutylammonium bromide. A plausible
ECEC mechanism involving siloxyphosphorus radical 8
is proposed: Application of the chlorotrimethylsilane-pro-
moted electroreduction of some triarylphosphine oxides
and alkyldiarylphosphine oxides 2 to the corresponding
phosphines 1 was successfully performed. Further appli-
cation to other phosphorus compounds is under investiga-
tion.
bond cleavage followed by silylation of Me SiO with
3
chlorotrimethylsilane would give 1a together with Me Si–
3
O–SiMe3.
In the absence of chlorotrimethylsilane, C–P bond fission
from 9a would occur predominantly to give an anion 11a
and the phenyl radical. Subsequent several reaction steps,
e.g., further electroreduction and protonation, would give
4
a, 5a, benzene, cyclohexadiene, and cyclohexane.
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

4
096
H. Kawakubo et al.
PAPER
Table 7 Chlorotrimethylsilane-Assisted Electroreduction of Phosphine Oxides
R1
P
R3
R1
P
R3
Me3SiCl (6 mmol)
R2
O
R2
(
Zn)–(Cu); Bu4NBr (2 mmol)
MeCN (5 mL), undivided cell
2
1
1
00 mA , 4 F/mol, 45 °C
(
2 mmol)
R1
R2
R3
Ratio (%)
b
Yield (%) of 1
a
Entry
Substrate
1
2
1
2
3
4
5
6
7
2a
2b
2c
2d
2e
2f
Ph
Ph
Ph
91
68
91
80
96
79
46
0
2
89
58
92
82
93
82
2-Tol
3-Tol
4-Tol
2-Tol
3-Tol
4-Tol
2-Tol
3-Tol
4-Tol
32
5
11
3
4-MeOC H
4-MeOC H
4-MeOC H
6
4
6
4
6
4
Ph
Ph
Bu
7
2g
2h
Ph
Ph
i-Pr
OPh
42
48
8
OPh
OPh
a
Isolated yields.
b
c
Determined by 31P NMR.
(
Pt)–(Pt), divided cell, 25 mA, 2 F/mol, r.t.
1
13
H and C NMR spectra were recorded on a Varian Mercury 300
ditions (100 mA, 129 min, 4 F/mol). The mixture was poured into
1
13
1
(
(
H: 300 MHz, C: 75 MHz) spectrometer in CDCl [ H: TMS
d = 0 ppm), C: CDCl (d = 77.0 ppm)]. P NMR spectra were re-
5% aq HCl and extracted with EtOAc (3 × 10 mL). The combined
3
13
31
organic layers were washed successively with sat. aq NaHCO and
3
3
corded on a Varian Mercury 300 (120 MHz) spectrometer in CDCl3
and/or MeCN/C D [85% H PO (d = 0 ppm) as an external stan-
dard]. IR spectra were recorded on a Jasco FT/IR-4100 spectropho-
tometer. Elemental analysis was executed on Perkin-Elmer PE 2400
Series II CHNS/O analyzer. Gas chromatography (GC) was per-
formed with Shimadzu GC-2014 instrument equipped with two
packed columns (Silicon SE-52 (5%), Chromosorb W 80-100 AW-
DMCS, 2.1 m × 3.2 mm i.d.) using N gas as an eluent. Analytic
TLC was performed on Merck, pre-coated plate silica gel 60 F254
(
brine, dried (Na
P NMR (120 MHz, CDCl
2
SO
4
), and concentrated under reduced pressure.
) analysis of the residue showed Ph P
3
3
1
6
6
3
4
3
1a (d = –5.3 ppm), 2a (d = 29.5 ppm), and Ph PH=O 4a (d = 21.7
2
ppm) in a ratio of 91:2:5. The residue was chromatographed (silica
gel, hexane–EtOAc, 5:1) to afford 1a (464.1 mg, 1.77 mmol, 89%).
Electroreduction of Triphenylphosphine Oxide (2a) Using
Me SiCl in a Divided Cell
2
3
Into a divided cell equipped with magnetic stirring bars and Pt
2
0.25 mm thickness). Column chromatography was performed on
anode and cathode (1.5 × 1.0 cm ) was placed 2a (139.2 mg, 0.500
Kanto Chemical silica gel 60N (40–50 mm).
mmol) in the cathodic chamber. The whole apparatus was purged
with argon. A soln of Bu NBr (647.9 mg, 2.01 mmol) in MeCN (5
4
Cyclic voltammetry (CV) was measured on the BAS 100B/W po-
tentiostat (BAS, Tokyo, Japan). A Pt (3 mm diameter), a Pt wire
electrode (1 mm diameter), and an Ag/Ag electrode (Ag wire in
mL) was added to both chambers. Me SiCl (0.19 mL, 1.5 mmol)
3
was added to the cathodic chamber at r.t. The resulting mixture was
electrolyzed under constant current conditions (50 mA, 48 min, 3 F/
mol). The catholyte was poured into 5% aq HCl and extracted with
EtOAc (3 × 10 mL). The combined organic layers were washed suc-
cessively with sat. aq NaHCO and brine, dried (Na SO ), and con-
+
0
.01 M MeCN soln of AgNO and Bu NClO ) were used as a work-
3 4 4
ing, a counter, and a reference electrode, respectively. The working
electrode was polished with 5 mm diamond slurry and with 0.5 mm
alumina slurry, successively, and washed with deionized H O and
3
2
4
2
31
centrated under reduced pressure. P NMR (120 MHz, CDCl3)
analysis of the residue showed 1a, 2a, 4a, and 5a in a ratio of
80:10:1:9. The residue was chromatographed (silica gel, hexane–
EtOAc, 5:1) to afford 1a (94.1 mg, 0.338 mmol, 72%).
acetone.
Unless otherwise noted, all materials were purchased from commer-
cial suppliers, and reagent grade materials were used without fur-
ther purification. All reactions were performed in anhyd solvents
under an argon atmosphere. MeCN and DMF were distilled from
CaH . THF and 1,4-dioxane were distilled from Na benzophenone
ketyl under argon atmosphere.
Triphenylphosphine (1a)
2
Colorless solid; R = 0.80 (hexane–EtOAc, 5:1).
f
–
1
IR (KBr): 3066, 1638, 1308, 1281, 1155 cm .
1H NMR (300 MHz, CDCl
): d = 7.28–7.35 (m, 15 H).
Electroreduction of Triphenylphosphine Oxide (2a) Using
3
Me SiCl in an Undivided Cell
13
3
C NMR (75 MHz, CDC1 ): d = 128.45 (d, J = 6.8 Hz), 128.66,
3
Into an undivided cell equipped with a magnetic stirring bar, a Zn
1
33.69 (d, J = 19.5 Hz), 137.10 (d, J = 9.8 Hz).
2
2
anode (1.5 × 1.0 cm ), and a Cu cathode (1.5 × 1.0 cm ) were placed
3
1
P NMR (120 MHz, CDCl ): d = –4.91.
Ph P=O (2a, 556.7 mg, 2 mmol) and Bu NBr (646.8 mg, 2 mmol);
3
3
4
the whole apparatus was purged with argon. To the mixture were
Tri-2-tolylphosphine (1b)
added MeCN (5 mL) and Me SiCl (0.76 mL, 6.0 mmol) at 45 °C.
3
Colorless solid; R = 0.63 (hexane–EtOAc, 5:1).
The resulting mixture was electrolyzed under constant current con-
f
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

PAPER
Electroreduction of Triphenylphosphine Oxide to Triphenylphosphine
4097
IR (KBr): 3056, 3002, 2969, 2942, 2913, 2362, 1955, 1919, 1843,
1382, 1364, 1327, 1305, 1267, 1240, 1181, 1156, 1095, 1070, 1046,
1025, 999, 968, 927, 911, 879, 848, 742, 697 cm .
–
1
1
1
810, 1695, 1624, 1587, 1565, 1467, 1451, 1377, 1282, 1271, 1201,
161, 1032, 985, 942, 869, 800, 751, 716, 680, 667, 555, 515, 458
1
H NMR (300 MHz, CDCl ): d = 7.57–7.47 (m, 4 H), 7.38–7.26 (m,
–
1
3
cm .
6
H), 2.45 (d sept, J = 6.9, 6.9 Hz, 1 H), 1.08 (dd, J = 15.6, 6.9 Hz,
1
H NMR (300 MHz, CDCl ): d = 7.30–7.21 (m, 6 H), 7.11–7.05 (m,
6 H).
3
3
H), 6.75–6.71 (m, 3 H), 2.40 (s, 9 H).
13
C NMR (75 MHz, CDC1 ): d = 137.31 (d, J = 24.8 Hz), 133.11 (d,
3
1
3
C NMR (75 MHz, CDC1 ): d = 142.77, 142.43, 134.42, 134.23,
J = 19.0 Hz), 128.20, 127.89 (d, J = 6.9 Hz), 24.65 (d, J = 8.9 Hz),
3
1
32.96, 130.04, 129.98, 128.63, 126.12, 21.30, 21.01.
19.39 (d, J = 18.1 Hz).
3
1
31
P NMR (120 MHz, CDCl ): d = –29.0.
P NMR (120 MHz, CDCl ): d = 1.70.
3
3
Tri-3-tolylphosphine (1c)
Cyclic Voltammetry of 2a (Figure 3)
Colorless solid; R = 0.63 (hexane–EtOAc, 5:1).
The working, counter, and reference electrodes were dipped into a
MeCN soln of Ph P=O 2a (10 mM) containing Bu NBr (0.1 M) as
f
3
4
IR (KBr): 3074, 3046, 3020, 2994, 2951, 2917, 2360, 1952, 1884,
a supporting electrolyte in a beaker-type electrochemical cell. The
soln was degassed with argon for 10 min before measurement. The
cyclic voltammogram was measured in a potential range from
1
9
1
693, 1589, 1474, 1449, 1400, 1305, 1269, 1216, 1169, 1103, 1038,
98, 884, 781, 696, 542, 519 cm .
–
1
H NMR (300 MHz, CDCl ): d = 7.36–7.17 (m, 12 H), 2.41 (s, 9 H).
+
3
–2.0 V to –3.5 V vs. Ag/Ag at a scan rate of 200 mV/s.
1
3
C NMR (75 MHz, CDC1 ): d = 137.74, 137.64, 137.16, 137.01,
3
1
34.44, 134.15, 130.69, 130.46, 129.31, 128.23, 128.15, 21.26.
References
3
1
P NMR (120 MHz, CDCl ): d = –7.38.
3
(
1) (a) Maercker, A. In Organic Reactions, Vol. 14; John Wiley
& Sons: New York, 1965, Chap. 3. (b) Boutagy, J.;
Thomas, R. Chem. Rev. 1974, 74, 87.
Tri-4-tolylphosphine (1d)
Colorless solid; R = 0.63 (hexane–EtOAc, 5:1).
f
(
2) (a) Hughes, D. L. In Organic Reactions, Vol. 42; John
Wiley & Sons: New York, 1992, Chap. 2. (b) Mitsunobu,
O.; Yamada, M. Bull. Chem. Soc. Jpn. 1967, 40, 2380.
IR (KBr): 3068, 3013, 2971, 2920, 2865, 2732, 2359, 1917, 1814,
1
1
5
1
658, 1595, 1496, 1450, 1394, 1352, 1307, 1272, 1211, 1186, 1118,
104, 1090, 1038, 1019, 971, 846, 811, 711, 640, 624, 605, 516,
05, 476 cm .
(
c) Mukaiyama, T.; Matsueda, R.; Maruyama, H. Bull.
–1
Chem. Soc. Jpn. 1970, 43, 1271.
H NMR (300 MHz, CDCl ): d = 7.55–7.13 (m, 12 H), 2.34 (s, 9 H).
(3) (a) Mukaiyama, T.; Araki, M.; Takei, H. J. Am. Chem. Soc.
1973, 95, 4763. (b) Corey, E. J.; Nicolaou, K. C. J. Am.
Chem. Soc. 1974, 96, 5614. (c) Corey, E. J.; Nicolaou, K.
C.; Melvin, L. S. J. Am. Chem. Soc. 1975, 97, 653.
(d) Modification: Gerlach, H.; Thalmann, A. Helv. Chim.
Acta 1974, 57, 2661.
3
1
3
C NMR (75 MHz, CDC1 ): d = 138.38, 134.63, 134.38, 134.14,
3
1
32.46, 130.18, 128.19, 20.82.
3
1
P NMR (120 MHz, CDCl ): d = –4.73.
3
Tris(4-methoxyphenyl)phosphine (1d)
(
4) (a) Appel, R. Angew. Chem., Int. Ed. Engl. 1975, 14, 801.
(b) Calzada, J. G.; Hooz, J. Org. Synth. 1974, 54, 63.
Colorless solid; R = 0.60 (toluene–CHCl , 4:1).
f
3
IR (KBr): 3067, 3009, 2956, 2929, 2896, 2834, 1643, 1581, 1565,
(5) (a) Staudinger, H.; Meyer, J. Helv. Chim. Acta 1919, 2, 635.
(b) Gololobov, Y. G.; Zhmurova, I. N.; Kasukhin, L. F.
Tetrahedron 1981, 37, 437. (c) Scriven, E. F. V.; Turnbull,
K. Chem. Rev. 1988, 88, 297. (d) Gololobov, Y. G.;
Kasukhin, L. F. Tetrahedron 1992, 48, 1353. (e) Shah, S.;
Protasiewicz, J. D. Coord. Chem. Rev. 2000, 210, 181.
(6) Phosphorus is obtained economically from limited regions
of the world: Kuroda, A.; Takiguchi, N.; Kato, J.; Ohtake, H.
J. Environ. Biotech. 2005, 4, 87.
1
1
1
487, 1474, 1462, 1439, 1401, 1362, 1307, 1284, 1185, 1178, 1099,
030, 832, 798 cm .
–1
H NMR (300 MHz, CDCl ): d = 7.34–7.29 (dd, J = 8.2, 6.3 Hz, 6
3
H), 7.03–6.93 (m, 6 H), 3.87 (s, 9 H).
1
3
C NMR (75 MHz, CDC1 ): d = 159.94, 134.95 (d, J = 20.6 Hz),
3
1
28.85 (d, J = 7.8 Hz), 114.05 (d, J = 7.6 Hz), 55.02.
3
1
P NMR (120 MHz, CDCl ): d = –9.73.
3
(
7) As an another approach, catalytic Wittig reaction using
3
-methyl-1-phenylphospholane 1-oxide was reported:
Butyldiphenylphosphine (1e)
O’Brien, C. J.; Tellez, J. L.; Nixon, Z. S.; Kang, L. J.; Carter,
A. L.; Kunkel, S. R.; Przeworski, K. C.; Chass, G. A. Angew.
Chem. Int. Ed. 2009, 48, 6836.
Colorless liquid; R = 0.63 (hexane–EtOAc, 5:1).
f
IR (neat): 3070, 3052, 3014, 3000, 2956, 2928, 2870, 2674, 1951,
1
1
9
1
884, 1810, 1739, 1655, 1585, 1570, 1542, 1480, 1463, 1432, 1416,
376, 1326, 1304, 1271, 1244, 1199, 1182, 1156, 1096, 1069, 1026,
99, 968, 893, 842, 781, 738, 696 cm .
(
8) (a) Fritzsche, H. Chem. Ber. 1965, 98, 171. (b) Coumbe, T.;
Lawrence, N. J.; Muhammad, F. Tetrahedron Lett. 1994, 35,
–
1
625. (c) Marsi, F. M. J. Org. Chem. 1974, 39, 265.
H NMR (300 MHz, CDCl ): d = 7.47–7.40 (m, 4 H), 7.38–7.30 (m,
(9) (a) Horner, L.; Hoffmann, H.; Beck, P. Chem. Ber. 1958, 91,
1583. (b) Imamoto, T.; Tanaka, T.; Kusumoto, T. Chem.
Lett. 1985, 14, 1491. (c) Griffin, S.; Heath, L.; Wyatt, P.
Tetrahedron Lett. 1998, 39, 4405. (d) Nelson, G. E. US
3
6
H), 2.09–2.03 (m, 2 H), 1.48–1.40 (m, 4 H), 0.94 (t, J = 7.2 Hz, 3
H).
1
3
C NMR (75 MHz, CDC1 ): d = 138.90 (d, J = 13.6 Hz), 132.42 (d,
3
4,507,502, 1985; Chem. Abstr. 1985, 103, 37617.
J = 18.4 Hz), 128.07 (d, J = 5.8 Hz), 27.91 (d, J = 15.9 Hz), 27.59
(
e) Busacca, C. A.; Raju, R.; Grinberg, N.; Haddad, N.;
(
d, J = 11.6 Hz), 24.07 (d, J = 13.1 Hz), 13.62.
James-Jones, P.; Lee, H.; Lorenz, J. C.; Saha, A.;
Senanayake, C. H. J. Org. Chem. 2008, 73, 1524.
3
1
P NMR (120 MHz, CDCl ): d = –15.51.
3
(
f) Malpass, D. B.; Yeargin, G. S. US 4,113,783, 1978;
Isopropyldiphenylphosphine (1f)
Chem. Abstr. 1979, 90, 23256.
Colorless liquid; R = 0.60 (hexane–EtOAc, 5:1).
f
(10) Handa, Y.; Inanaga, J.; Yamaguchi, M. J. Chem. Soc., Chem.
Commun. 1989, 298.
11) Mathey, F.; Maillet, R. Tetrahedron Lett. 1980, 21, 2525.
IR (neat): 3139, 3070, 3052, 3027, 3015, 3000, 2950, 2923, 2864,
(
1
952, 1883, 1812, 1738, 1661, 1585, 1570, 1543, 1480, 1461, 1432,
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York

4
098
H. Kawakubo et al.
PAPER
(
(
12) Dockner, T. Angew. Chem. 1988, 100, 699.
(24) (a) Santhanam, K. S. V.; Bard, A. J. J. Am. Chem. Soc. 1968,
90, 1118. (b) Raju, T.; Kulangiappar, K.; Kulandainathan,
M. A.; Muthukumaran, A.; Krishnan, V. IN 2002-DE793,
2007; Chem. Abstr. 2007, 147, 235305.
13) Timokhin, B. V.; Kazantseva, M. V.; Blazhev, D. G.;
Rokhin, A. V. Russ. J. Gen. Chem. 2000, 70, 1310; Chem.
Abstr. 2001, 134, 311265.
(
14) (a) Iorga, B.; Camichael, D.; Savignac, P. C. R. Acad. Sci.,
Ser. IIC 2000, 3, 821. (b) Hermeling, D.; Hugo, R.;
Lechtken, P.; Rotermund, G. W.; Siegel, H. DE 19,532,310,
(25) CRC Handbook of Chemistry & Physics, 74th ed.; Lide, D.
R., Ed.; CRC Press: London, 1993.
(26) Since 3a is highly moisture sensitive, 3a was used without
purification.
1997; Chem. Abstr. 1997, 126, 199670. (c) Masaki, M.;
Fukui, K. Chem. Lett. 1977, 6, 151.
(27) Yano, T.; Hoshino, M.; Kuroboshi, M.; Tanaka, H. Synlett
2010, 801.
(28) Yano, T.; Kuroboshi, M.; Tanaka, H. Tetrahedron Lett.
(
15) Triphenylphosphorus dichloride (3a) is also prepared by
treatment of 2 with chlorinating reagents such as
diphosgene, triphosgene, and phosphorus pentachloride.
2010, 51, 698.
(
a) Rao, V. J.; Reddy, A. M. IN 1996-DE1812, 2005; Chem.
(29) Kuroboshi, M.; Yano, T.; Kamenoue, S.; Kawakubo, H.;
Tanaka, H. Tetrahedron 2011, 67, 5825.
Abstr. 2007, 146, 402090. (b) Li, H.; Chen, Z.; Wu, L.;
Wang, C.; Hu, B. CN 1,660,862, 2005; Chem. Abstr. 2006,
(30) Tanaka, H.; Yano, T.; Kobayashi, K.; Kamenoue, S.;
Kuroboshi, M.; Kawakubo, H. Synlett 2011, 582.
(31) Among thus far examined solvents, MeCN was the only
solvent effective for the present purpose; thus, electro-
reduction of 2a to 1a hardly occurred in THF, DMSO, DMF,
and 1,4-dioxane. The electroreduction proceeded in a
mixture of THF and MeCN (4:1 to 1:4).
144, 488805. (c) Horner, L.; Hoffmann, H.; Beck, P. Chem.
Ber. 1958, 91, 1583.
(
16) Theoretical calculation: Mo, O.; Eckert-Maksic, Y. M.;
Maksic, Z. B.; Alkorta, I.; Elguero, J. J. Phys. Chem. A.
2005, 109, 4359.
(
(
(
(
(
(
(
17) Masaki, M.; Kaketani, N. JP 53-034,725, 1978; Chem.
Abstr. 1978, 89, 109953.
18) Fukui, K.; Kaketani, N.; Kita, J.; Fujimura, S. JP 55-149,293,
(32) As a supporting electrolyte, Bu NOTf, Bu NBF , and
4
4
4
Bu NClO could be used to give 1a in slightly lower yields
4 4
1
980; Chem. Abstr. 1981, 94, 175259.
19) Horner, L.; Beck, P.; Hoffmann, H. Chem. Ber. 1959, 92,
088.
20) Hermeling, D.; Siegel, H.; Hugo, R.; Rotermund, G. W. EP
25,073, 1996; Chem. Abstr. 1996, 125, 195993.
21) Wettling, T. EP 5,48,682, 1993; Chem. Abstr. 1993, 119,
39547.
(42, 38, and 52%, respectively). The electroreduction of 2a
proceeded without supporting electrolyte to give 1a in 41%
yield.
2
(33) One singlet signal appeared at d = 32.5 ppm in the 31P NMR
of a mixture of 2a, Me SiCl, and Bu NBr. On the other hand,
7
3
4
a mixture of 2a and Me SiBr showed a singlet peak at d =
3
1
48.7 ppm. These results suggest that Me SiCl would react
3
22) Young, D. A.; Brannock, K. C. US 3,780,111, 1973; Chem.
Abstr. 1974, 80, 60039.
23) (a) Organic Electrochemistry, 4th ed.; Lund, H.;
Hammerich, O., Eds.; Marcel Dekker: New York, 1991.
partially with Bu NBr to give Me SiBr, which would inter-
4
3
+
–
act with 1a to form a trace amount of [Ph P –O–SiMe ]Br .
3
3
However, since the electroreduction of Ph P=O (2a)
3
–
proceeded without Br source (ref. 32), the electroreduction
of 2a seems to proceed mainly through the ECEC
mechanism (Scheme 6).
(
b) Torii, S. In Electroorganic Reduction Synthesis;
Kodansha & Wiley-VCH: Tokyo/Weinheim, 2006. (c) New
Developments in Organic Electrosynthesis; Fuchigami, T.,
Ed.; CMC: Tokyo, 2004. (d) Electroorganic Chemistry,
Kagaku Zokan, Vol. 86; Osa, T.; Shono, T.; Honda, K., Eds.;
Kagaku Dojin: Kyoto, 1980.
Synthesis 2011, No. 24, 4091–4098 © Thieme Stuttgart · New York