An efficient Pd(II)-based catalyst system for carboxylation of aromatic C-H bond by addition of a phosphenium salt

Article abstract of DOI:10.1016/j.tetlet.2004.12.027

Addition of a phosphenium dramatically improved the reaction yields in the carboxylation of arenes by formic acid catalyzed by Pd(II). Control experiments revealed that the majority of the phosphenium triflate was converted to a mixed anhydride of phosphonic acid and formic acid (7), which however did not substitute for the phosphenium to improve the reaction yield.

Full text of DOI:10.1016/j.tetlet.2004.12.027

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 959–962
An efficient Pd(II)-based catalyst system for carboxylation
of aromatic C–H bond by addition of a phosphenium salt
Ken Sakakibara, Makoto Yamashita and Kyoko Nozaki*
Department of Chemistry and Biochemistry, Graduate School of Engineering, The University of Tokyo,
7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 Japan
Received 20 November 2004; revised 7 December 2004; accepted 8 December 2004
This article is dedicated to Professor Iwao Ojima on the occasion of his 60th birthday
Abstract—Addition of a phosphenium dramatically improved the reaction yields in the carboxylation of arenes by formic acid cata-
lyzed by Pd(II). Control experiments revealed that the majority of the phosphenium triflate was converted to a mixed anhydride of
phosphonic acid and formic acid (7), which however did not substitute for the phosphenium to improve the reaction yield.
ꢀ 2004 Elsevier Ltd. All rights reserved.
cat. Pd(OCOCF3)2
K2S2O8
Catalytic activation of aromatic C–H bonds leading to a
new C–C bond formation is of considerable interest in
organic synthesis.1–5 Compared with Friedel–Crafts
reaction, it would provide simple, clean, and economic
methods for making many useful substituted aromatic
compounds directly from simple arenes. Fujiwara and
co-workers developed the electrophilic substitution of
an aromatic hydrogen atom by Pd(II) to produce an
arylpalladium species.6 The arylpalladium was further
allowed to react with CO,7–9 CO2,10 O2,11,12 alkenes,13,14
alkynes,15 and nitrile16 to form aromatic carboxylic
acids, phenols, alkylarenes, alkenylarenes, and ketones,
respectively. Recently, Grushin reported similar carbox-
ylation of arenes by Rh-based catalyst system.17 We
recently reported the Pd-catalyzed carboxylation of
aromatic C–H bonds by using formic acid as a carbonyl
source (Eq. 1).18 Given that CO and CO2 are gaseous,
formic acid is more favorable for handling to produce
aromatic carboxylic acids.19 Here we report an efficient
catalyst system for carboxylation of aromatic C–H bond
by using phosphenium salts as additives to achieve high
yields in the Fujiwara-type carboxylation with formic
acid as a carbonyl source.
COOH
21%
ð1Þ
+
CF3COOH
72 h, 80 °C
HCOOH
Originally, phosphenium salts have been studied in com-
parison with its isoelectronic species, such as carbene
and silylene since its discovery.20–25 Phosphenium may
behave as a r-donor (Lewis base) and also as a p-accep-
tor (Lewis acid) due to the sp2 lone pair electrons and a
vacant p-orbital, respectively. In spite of their interesting
properties, however, only few catalytic reactions by the
use of phosphenium salts have been reported so
far.26,27 In fact, Baker showed his perspective22 that
some reactions whose mechanisms include the cationic
transition-metal complexes would be accelerated by
using strong electron-withdrawing ligands, such as pho-
sphenium ions. BakerÕs insight prompted us to use pho-
sphenium salts for our Pd-catalyzed carboxylation (Eq.
1), whose key stepis the electrophilic substitution of
aromatic hydrogen atom by Pd(II).18 The catalytic activ-
ity was improved by the addition of phosphenium ion
but no evidence was obtained to support the formation
of palladium–phosphenium complex.
Crystallographic study of phosphenium triflate 1 was
carried out to confirm the geometry around the central
phosphorus atom of 1.28 Baker reported applications
of phosphenium triflates for isolation of phosphenium-
ligated rhodium22 and platinum25 complexes. In their
Keywords: Electrophilic substitution; Phosphenium salt; Palladium;
Carboxylation; Catalyst.
*
Corresponding author. Tel.: +81 3 5841 7261; e-mail: nozaki@
chembio.t.u-tokyo.ac.jp
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.12.027

960
K. Sakakibara et al. / Tetrahedron Letters 46 (2005) 959–962
Table 2. Optimization of conditions for carboxylation reactiona
tBu
tBu
Pd(OCOCF3)2
1
K2S2O8
CF3COOH/(CF3CO)2O
HCOOH
COOH
Yieldb (%)
30 °C
Run
1/Pd
Time (h)
1
2
3
4
5
6
7
8
—
48
24
24
24
24
48
48
48
18c
32c
69c
22c
24c
84d
22c,e
2c
0.55
1.1
2.2
3.3
1.1
1.1
1.1f
a 5 mol % Pd(OCOCF3)2. 1.5 equiv K2S2O8, CF3CO2H/(CF3CO)2O
solvent (10:1).
b Sum of the two isomeric carboxylic acid (meta/para = 1:3).
c Yields were determined by 1H NMR using C2H4Cl2 as an internal
standard because of the presence of unseparable byproducts.
d Isolated yield.
Figure 1. ORTEP drawing of phosphenium triflate 1 with the triflate
anion of the other molecule. Dashed line is indicating the interactions
between the central phosphorus atom and two oxygen atoms. (A minor
part of disordered three fluorine atoms in triflate anion and hydrogens
were omitted for clarity.)
e Pd(OAc)2 was used.
f 2 was used instead of 1.
AgOTf
OMe
MeO
N
N
1
P
Cl
2
reports, they suggested that the structure of phosphe-
nium triflate 1 may consist of the weak coordination
of the triflate to the phosphorus atom based on a previ-
ously reported structure of phosphenium triflate 10.29
Single crystals suitable for X-ray analysis of 1 were
grown from a toluene solution diffused with hexane at
ꢀ30 ꢁC under argon. As shown in Figure 1, solid state
structure of 1 indicates that the central phosphorus
atom has two kinds of weak interactions between the va-
cant p-orbital of cationic phosphorus and oxygen atoms
For the carboxylation of arenes, the reaction conditions
were optimized with/without phosphenium salt 1.30
According to our previous report,18 arenes and formic
acid were used as substrates in the presence of Pd-
(OCOCF3)2 as a catalyst, K2S2O8 as an oxidant, phos-
phenium salts as additive, and
a
mixture of
CF3COOH/(CF3CO)2O as a solvent. Results are sum-
marized in Table 2. In the absence of the phosphenium
additive, the carboxylated product was obtained in 18%
yield in 48 h (run 1). The 1/Pd(OCOCF3)2 ratio was
optimized in runs 2–5 and the highest yield of 69% in
24 h was achieved with 1.1 equiv of 1 to Pd (run 3).
Longer reaction time of 48 h improved the yield up to
84% (run 6).31 The use of Pd(OAc)2 in place of Pd-
(OCOCF3)2 caused the dropof the yield to 22% (run
7). Addition of chlorophosphine 2 instead of phosphe-
nium 1 resulted in almost no reaction. Thus, in the
following studies, the reaction condition of run 6 was
chosen as the standard condition.
˚
of two triflate anions (P1–O4 = 2.4210(18) A, P1–
˚
O3* = 2.8322(19) A where O3* is one of oxygen atoms
of triflate anion in the other molecule) (Table 1) to form
pseudo-trigonal bipyramidal structure with a remaining
lone pair as the 5th substituent in the equatorial posi-
tion. Thus, the weak coordination of the triflate to the
central phosphorus atom, that is similar to 10,29 has been
confirmed.
MeO
N
N
OMe
N
N
The scope and generality of this reaction have been ex-
plored by using various commercially available arenes
(Table 3). The reaction exhibited good yields and gave
no byproducts for benzene and alkyl-substituted benz-
enes although electron-poor arenes, such as benzonitrile,
nitrobenzene, and a,a,a-trifluorotoluene, did not react
at all. Benzene was converted to benzoic acid as a sole
product in 53% yield (run 1). In the cases of mono-
and di-substituted arenes, carboxylation at the sterically
less-hindered position took place predominantly (runs
2–5). Different steric factor of the alkyl groupin the
mono-substituted arenes did not affect the total yield
of the carboxylated arenes, but did change the isomeric
ratio of the products (runs 2 and 3). In the case of m-xyl-
ene, even sterically hindered C–H bond was activated
P
P
OTf
1
OTf
1'
˚
Table 1. Selected interatomic lengths (A) and angles (ꢁ) of 1
P(1)–N(1)
P(1)–O(4)
N(1)–C(1)
1.628(2)
2.4210(18)
1.479(3)
93.74(10)
114.32(16)
124.64(16)
94.40
P(1)–N(2)
P(1)–O(3*)
N(2)–C(4)
1.634(2)
2.8322(19)
1.485(3)
114.01(16)
122.47(16)
93.48
N(1)–P(1)–N(2)
P(1)–N(2)–C(2)
P(1)–N(2)–C(10)
N(2)–P(1)–O(4)
N(2)–P(1)–O(3*)
P(1)–N(1)–C(1)
P(1)–N(1)–C(3)
N(1)–P(1)–O(4)
N(1)–P(1)–O(3*)
90.89
101.24

K. Sakakibara et al. / Tetrahedron Letters 46 (2005) 959–962
961
Table 3. Reaction scope of the carboxylationa
tBu
Pd(tfa)2
31P NMR
δ 11
(1JPH = 735 Hz)
1
Run
Arene
Product
Yield
53b
ð2Þ
K2S2O8
CF3COOH/(CF3CO)2O
HCOOH
COOH
1
Table 4. The 31P NMR chemical shift of phosphenium 1 with additives
2
3
4
5
6
89c
86d
83e
71f
80b
in CF3COOD/(CF3CO)2Oa
COOH
tBu
additives
tBu
31P NMR
Ar
N
N
Ar
P
CF3CO2D
(CF3CO)2O
OTf
1
COOH
Run Additives
dP (ratio)
1JPH
Hb
Ha
(Hz)
COOH
COOH
1
2
3
4
5
None
Pd(tfa)2
132, 108, 4.2 (1:8:1)
132, 108, 3.4 (1:7:1)
—
—
b
Ha
Pd(tfa)2,b C6H6,c HCOOHd 11 (single peak)
739
739
739
C6H6,cHCOOHd
HCOOHd
11 (single peak)
11 (single peak)
Hc
Hb
a 1 55 lmol (24.8 mg), TFA-D 0.5 mL, TFAA 0.05 mL.
b 50 lmol.
COOH
COOH
c 1 mmol.
d 6 mmol (121 equiv).
7
93b
Ar
N
Ar
N
CF3CO2H
(CF3CO)2O
a11 mol % 1, 1/Pd = 1.1. 30 ꢁC, 48 h, 1.5 equiv K2S2O8, CF3CO2H/
(CF3CO)2O solvent (10:1).
b Single product was obtained.
c o/m/p = 18:31:51.
P
P OCOCF3
N
Ar
N
Ar
OTf
1
3
CF3SO3H
tfa
Ar = p-MeOC6H4
d o/m/p = 0:25:75.
e Ha/Hb = 83:17.
f Ha/Hb/Hc = 14:9:77.
O
P
tfa
O
N
F3C
Ar
O
F3C
(CF3CO)2O
N
Ar
N
N
P(tfa)3
Ar
4
CF3
+
Ar
6
5
under the reaction condition (run 5). The reaction of
mesitylene afforded 2,4,6-trimethylbenzoic acid as a sole
product in a high yield even though mesitylene has only
hindered C–H bonds. Hence, the present carboxylation
reaction widely covers a range of substituted aromatic
hydrocarbons with high yields.
H
HCOOH
P
(OCOH)2
7
5
O
Scheme 1. Sequential transformation of phosphenium in the reaction
media.
In order to elucidate the role of phosphenium 1 in the
catalytic reaction, an aliquot taken from the reaction
mixture was analyzed by 31P NMR spectroscopy. The
COCF3)2 and benzene (runs 3, 4, and 5). This peak was
assigned to be mixed anhydride 7 as follows. Dissolving
PCl3 in HCOOH afforded a single peak at d 11. The
peak at d 11 was also observed by an addition of
HCOOH to P(OCOCF3)3 (5). Considering the P–H cou-
pling constant and chemical shift of HP(@O)(OAc)2 (dP
3, 1JPH = 620 Hz),35 one can expect that the peak of d 11
is assignable to mixed anhydride 7, which consists of
phosphonic acid and formic acid although 7 was not
able to be isolated.36 Thus, phosphenium salt 1 seems
to have transformed into 5 through 3 and 4 in
CF3COOH/(CF3CO)2O solvent.
1
only visible peak was at d 11 with JPH = 735 Hz (Eq.
2). For the purpose of characterization of this peak, sev-
eral control experiments by 31P NMR spectroscopy were
carried out (Table 4). Although 1 exhibits a singlet peak
of d 185 in C6D6, a higher-field shift of the peak of 1 to d
108 (accompanied by two small peaks of d 132 and 4) in
31P NMR spectrum was observed when the phosphe-
nium was dissolved in the reaction media, that is,
CF3COOD/(CF3CO)2O (run 1). Considering that the
reaction of tris(dialkylamino)phosphines with acetic
anhydride formed P(OAc)3 via the P–N bond cleav-
age,32 the peak at d 108 could be assigned to
P(OCOCF3)3 (5) (Scheme 1).33,34 No change in chemical
shifts and intensities of the three peaks in the existence
of palladium, which implied that the formation of any
phosphorus-containing palladium complex is less prob-
able (run 2). An addition of formic acid made a signifi-
cant chemical shift change to d 11 with 1JPH of 739 Hz in
31P NMR spectrum regardless to the existence of Pd(O-
For the catalytic reaction, however, the use of mixed
anhydride 7 instead of phosphenium 1 did not repro-
duce the yield achieved with 1. The carboxylation
reaction by using phosphenium 1 took place in 69%
yield (Table 5, run 1, the same as Table 2, run 3). In
run 2, P(OCOCF3)3 (5) generated from PCl3 and AgO-
COCF3 didnÕt lead to high yield although mixed anhy-
dride 7 should be generated from 5 and formic acid.
To assimilate a reaction condition to the optimized

962
K. Sakakibara et al. / Tetrahedron Letters 46 (2005) 959–962
Table 5. Examination of the transformation of phosphenium 1
9. Lu, W. J.; Yamaoka, Y.; Taniguchi, Y.; Kitamura, T.;
Takaki, K.; Fujiwara, Y. J. Organomet. Chem. 1999, 580,
290–294.
10. Sugimoto, H.; Kawata, I.; Taniguchi, H.; Fujiwara, Y. J.
Organomet. Chem. 1984, 266, C44–C46.
11. Jintoku, T.; Taniguchi, H.; Fujiwara, Y. Chem. Lett. 1987,
1865–1868.
Additives
Pd(tfa)2
tBu
tBu
K2S2O8
CF3COOH/(CF3CO)2O
HCOOH
COOH
12. Jintoku, T.; Takaki, K.; Fujiwara, Y.; Fuchita, Y.; Hiraki,
K. Bull. Chem. Soc. Jpn. 1990, 63, 438–441.
13. Moritani, I.; Fujiwara, Y. Tetrahedron Lett. 1967, 1119–
1122.
14. Fujiwara, Y.; Moritani, I.; Danno, S.; Asano, R.; Tera-
nishi, S. J. Am. Chem. Soc. 1969, 91, 7166–7169.
15. Jia, C. G.; Piao, D. G.; Oyamada, J. Z.; Lu, W. J.;
Kitamura, T.; Fujiwara, Y. Science 2000, 287, 1992–
1995.
Run
Additives
Yields (%)
1
1
69
2a
3
5
Trace
Trace
Trace
6
5, 6
4a
t-Butylbenzene 4 mmol, Pd(tfa)2 0.2 mmol, TFA-H 2.0 mL, TFAA
0.2 mL, HCOOH 1.0 mL, K2S2O8 6.0 mmol,
0.22 mmol, 6 0.22 mmol, 24 h, 30 ꢁC.
1 0.22 mmol, 5
a 5 was generated from PCl3 + AgOCOCF3.
16. Zhou, C. X.; Larock, R. C. J. Am. Chem. Soc. 2004, 126,
2302–2303.
17. Grushin, V. V.; Marshall, W. J.; Thorn, D. L. Adv. Synth.
Catal. 2001, 343, 161–165.
18. Shibahara, F.; Kinoshita, S.; Nozaki, K. Org. Lett. 2004,
6, 2437–2439.
19. Morimoto, T.; Kakiuchi, K. Angew. Chem., Int. Ed. 2004,
43, 5580–5588.
one, independently synthesized diamide 6 was added to
the reaction system (runs 3 and 4), but only trace
amount of product was obtained. Thus, it ascertained
that adding phosphenium 1, not 7, to the reaction sys-
tem had some special effects.
20. Cowley, A. H.; Kemp, R. A. Chem. Rev. 1985, 85, 367–
382.
21. Gudat, D. Coord. Chem. Rev. 1997, 163, 71–106.
22. Abrams, M. B.; Scott, B. L.; Baker, R. T. Organometallics
2000, 19, 4944–4956.
23. Nakazawa, H.; Miyoshi, K. J. Synth. Org. Chem. Jpn.
2001, 59, 52–61.
24. Nakazawa, H. Adv. Organomet. Chem. 2004, 50, 108–
143.
In conclusion, we found a new phosphenium usage as an
additive for the carboxylation reaction. Various arenes
were converted to the corresponding arenecarboxylic
acids in high yields. The mixed anhydride 7 was revealed
to be the major species containing phosphorus atom un-
der the carboxylation reaction. However, this was not
an active species, and thus the origin of the phosphe-
nium effect towards the better yields is now under
investigation.
25. Hardman, N. J.; Abrams, M. B.; Pribisko, M. A.; Gilbert,
T. M.; Martin, R. L.; Kubas, G. J.; Baker, R. T. Angew.
Chem., Int. Ed. 2004, 43, 1955–1958.
26. Breit, B. Chem. Commun. 1996, 2071–2072.
27. Breit, B. J. Mol. Catl. A 1999, 143, 143–154.
28. Crystal data for 1 (CCDC No 255646) colorless prism,
0.50 · 0.50 · 0.20 mm3, C17H18F3N2O5PS, M = 450.36,
Acknowledgements
We thank Professor Takayuki Kawashima and Professor
Kei Goto for X-ray crystallography and glovebox use.
˚
T = 120(2) monoclinic, P21/c a = 13.531(3) A, b =
˚
˚
9.111(2) A, c = 15.898(4) A, b = 96.2910(9)ꢁ.
29. Jones, V. A.; Sriprang, S.; Thornton-Pett, M.; Kee, T. P.
J. Organomet. Chem. 1998, 567, 199–218.
Supplementary data
30. Addition of two other phosphenium salts, which have
2,4,6-trimethylphenyl or 2,6-diisopropylphenyl group on
each nitrogen atom instead of 4-methoxyphenyl group, to
the reaction system resulted in poorer yields.
31. The yield of t-butylbenzoic acid was increased by time to
time upto 48 h (see Supplementary data). During the
reaction, no change in isomeric ratio of the products was
observed.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.
2004.12.027.
References and notes
32. Hargis, J. H.; Mattson, G. A. J. Org. Chem. 1981, 46,
1597–1602.
1. Shilov, A. E.; ShulÕpin, G. B. Chem. Rev. 1997, 97, 2879–
2932.
2. Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507–514.
3. Kakiuchi, F.; Murai, S. Acc. Chem. Res. 2002, 35, 826–
834.
33. Garner, C. D.; Hughes, B. Inorg. Chem. 1975, 14, 1722–
1724.
34. Although Ref. 33 did not reported the 31P NMR chemical
shift of P(OCOCF3)3, we confirmed it made from PCl3 and
AgOTf based on the procedure described in Ref. 33 as d
108.
4. Ritleng, V.; Sirlin, C.; Pfeffer, M. Chem. Rev. 2002, 102,
1731–1769.
35. SalÕkeeva, L. K.; Nurmagambetova, M. T.; Kurmanaliev,
O. S.; Gazizov, T. K. Rus. J. Gen. Chem. 2003, 73, 183–
186.
36. The solution of 7 evolved small bubbles, which likely to be
CO gas due to the formyl group on the phosphorus. As the
evolution of the bubbles proceeded, the S/N ratio of 31P
NMR spectrum changed to lower, indicating continuous
decomposition of phosphorus-containing species in this
condition.
5. Kakiuchi, F.; Chatani, N. Adv. Synth. Catal. 2003, 345,
1077–1101.
6. Fuchita, Y.; Hiraki, K.; Kamogawa, Y.; Suenaga, M.;
Tohgoh, K.; Fujiwara, Y. Bull. Chem. Soc. Jpn. 1989, 62,
1081–1085.
7. Jintoku, T.; Fujiwara, Y.; Kawata, I.; Kawauchi, T.;
Taniguchi, H. J. Organomet. Chem. 1990, 385, 297–306.
8. Taniguchi, Y.; Yamaoka, Y.; Nakata, K.; Takaki, K.;
Fujiwara, Y. Chem. Lett. 1995, 345–346.