The mechanic study of the Pd-catalyzed synthesis of diphenylcarbonate with heteropolyacid as a cocatalyst

Article abstract of DOI:10.1016/S0022-328X(03)00203-1

The reaction to synthesize diphenyl carbonate (DPC) by an oxidative carbonylation of phenol with CO and O₂ has been found to proceed through the second-order of phenol concentration. The activation energy E _a, Δ S and Δ H are 27.0 kcal mol^-1, -6.43 cal mol^-1 and 26.3 kcal mol^-1, respectively. The kinetic and additive data obtained agree with the proposed mechanism as follows: Pd(OAc)₂ reacts with an ammonium phenoxy salt to give AcO-Pd-OPh which then reacts with CO to form AcO-Pd-COOPh. This species leads to PhO-Pd-COOPh which undergoes reductive elimination to give DPC and Pd(0). This Pd(0) is reoxidized to Pd(II) by the help of a heteropolyacid very effectively.

Full text of DOI:10.1016/S0022-328X(03)00203-1

Journal of Organometallic Chemistry 674 (2003) 96Á100
/
www.elsevier.com/locate/jorganchem
The mechanic study of the Pd-catalyzed synthesis of
diphenylcarbonate with heteropolyacid as a cocatalyst
Itsuhiro Hatanaka a, Naho Mitsuyasu a, Guochuan Yin b, Yuzo Fujiwara a,
Tsugio Kitamura c, Katsumi Kusakabe a, Teizo Yamaji a,*
a Department of Chemistry and Biochemistry, Graduate School of Engineering, Kyushu University, 6-10-1, Hakozaki, Higasi ku, Fukuoka 812-8581,
Japan
b Department of Chemistry, University of Kansas, Lawrence, KS 66044, USA
c Faculty of Science and Engineering, Department of Chemistry and Applied Chemistry, Saga University, Honjo-machi, Saga 840-8502, Japan
Received 10 December 2002; received in revised form 14 March 2003; accepted 25 March 2003
Abstract
The reaction to synthesize diphenyl carbonate (DPC) by an oxidative carbonylation of phenol with CO and O2 has been found to
proceed through the second-order of phenol concentration. The activation energy Ea, DS and DH are 27.0 kcal molꢀ1, ꢀ
/
6.43 cal
molꢀ1 and 26.3 kcal molꢀ1, respectively. The kinetic and additive data obtained agree with the proposed mechanism as follows:
Pd(OAc)2 reacts with an ammonium phenoxy salt to give AcOÃPdÃOPh which then reacts with CO to form AcOÃPdÃCOOPh.
This species leads to PhOÃPdÃCOOPh which undergoes reductive elimination to give DPC and Pd(0). This Pd(0) is reoxidized to
/
/
/
/
/
/
Pd(II) by the help of a heteropolyacid very effectively.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Heteropolyacid; Phenol; Diphenyl carbonate; Oxidative carbonylation; Mechanism
1. Introduction
direct oxidation of Pd(0) to the Pd(II) species with
oxygen was confirmed to be very slow, which results in
an aggregation of Pd(0) to form palladium black that is
more difficult to reoxidize to Pd species from Pd(0).
Therefore, the efficient redox cocatalysts are essential to
achieve a rapid reoxidation of Pd(0). It was reported
that heteropolyacids (HPAs) could act as a redox
cocatalyst in some palladium-catalyzed reactions
[12,13]. We also reported the highly effective HPA
cocatalysts for the catalytic carbonylation of phenol
with CO and O2 [14,15].
Diphenyl carbonate (DPC), an important precursor
of polycarbonate, is industrially produced by the reac-
tion of phenol and phosgene with a formation of
chlorinated waste [1]. In recent years, there is an
increasing demand for a safer process of DPC manu-
facture. Oxidative carbonylation of phenol with CO and
O2 is one of the attractive processes in which the use of
toxic phosgene is avoided as in Eq. (1).
(1)
In this process, TON of Pd(OAc)2 increased diame-
trically [15]. The amount of Pd found as palladium black
decreased clearly in the high TON reactions compared
with low TON reactions. It was also known that Pd
catalyst system to use ammonium halides gave the best
efficiency [5]. However, it is not yet known why halides
gave the best efficiency. It is also known that an HPA
acts as a redox cocatalyst against Pd through Mn metals
[15]. To elucidate the optimum amount of Mn, ammo-
nium salts and HPAs are very important and requisite
for not only to improve the productivity of DPC and
In the previous literature, many metal compounds
were chosen as catalysts for the oxidative carbonylation
reaction [2Á11] and palladium was thought to be the
/
most active catalyst. In this reaction, regeneration of
Pd(II) from Pd(0) is the crucial step for the high yield of
DPC and high turnover number (TON) of Pd. However,
* Corresponding author. Tel./fax: ꢁ
/
81-926424374.
E-mail address: yamaji@cstf.kyushu-u.ac.jp (T. Yamaji).
0022-328X/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00203-1

I. Hatanaka et al. / Journal of Organometallic Chemistry 674 (2003) 96Á
/
100
97
Fig. 1. The reaction order (second-order). Conditions: phenol, 0.5
mol; Pd(OAc)2, 0.16 mmol; (TBA)4SiW6Mo6O40, 0.047 mmol; TBAB,
3.0 mmol; M.S. 3A, 7.0 g; total pressure, 8 atm; CO, 500 ml minꢀ1; O2,
35 ml minꢀ1; 80 8C, 3 h.
Fig. 3. Arrhenius plot of DPC. Conditions: phenol, 0.5 mol;
Pd(OAc)2, 0.16 mmol; TBAB, 3.0 mmol; (TBA)4SiW6Mo6O40, 0.047
mmol; total pressure, 8 atm; CO, 500 ml minꢀ1; O2, 35 ml minꢀ1
;
M.S., 7 g; 0.5 h. k is the production rate of DPC (mol sꢀ1).
TON of Pd, but also to find out a plausible active
catalyst and reaction mechanism for this reaction. In
addition to the kinetic data for this reaction, these
experiments were performed. By these experiments, we
could understand this reaction well. Here we report the
results and propose the reaction mechanism.
2.2. Arrhenius plot of DPC production
Arrhenius plots were obtained in the temperature
range 327Á338 K. These data were shown in Fig. 3 to
give Eaꢂ
27.0 kcal molꢀ1, DSꢂ 6.43 cal molꢀ1 and
DHꢂ
26.3 kcal molꢀ1, respectively. The negative en-
/
/
/
ꢀ
/
/
tropy obtained means that some type of complex is
formed at the rate-controlling step [16]. The positive
enthalpy means that this reaction can be effectively
performed even under 373 K in the industrial scale.
2. Results and discussion
2.1. Kinetic analysis of DPC production in the PdÁ
/
HPA
catalyst system
2.3. Effect of TBA salts
Kinetic study was done using the flow system under
the conditions of 80 8C and total pressures of 8 atm by
adding Pd(OAc)2, (NBu4)4SiW6Mo6O40 ((TBA)4SiW6-
Mo6O40), tetrabutylammonium bromide (TBAB) and
molecular sieves (M.S.) to the system. The DPC yield
(X) was obtained against time. These data were
analyzed to give the best-fit for the second-order
The effect of change of anion of TBA salts is also an
interesting aspect. Table 1 summarizes these results. It
can be concluded that only bromide causes this reaction
effectively. Other anions could not cause this reaction at
all. In order to understand this phenomenon clearly, we
plotted DPC yield against pKa of conjugated acids. The
result is shown in Fig. 4. This figure clearly shows that
pKa of a conjugated acid is important in this reaction
reaction against phenol with R2ꢂ
0.9628 as shown in
/
Fig. 1. This result means that the reaction proceeds in
the second-order of phenol; phenol reacting with
another phenol containing active species at the rate-
determining step. The calculated data fitted with the
experimental ones as shown in Fig. 2.
and should be lower than ꢀ
/
8. If the pKa of the
conjugated acid is lower than ꢀ
/
8, we would be able to
get a good DPC yield. HI has a lower pKa than ꢀ9.
/
However Iꢀ is easily oxidized by oxygen to I2 [2].
Therefore ammonium iodide is not appropriate. This
Table 1
Effect of TBA salt (anion part) a
Entry
TBA salt
Yield (%) b
TON c
1
2
3
4
5
TBA-Br
TBA-Cl
20.8
2.0
325
33
1
TBA-F×
/
3H2O
0.1
Trace
Trace
TBA-CN
TBA-OAc
Á/
Á/
a
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; (TBA)4SiW-
Mo11O40, 0.047 mmol; Mn(OAc)2, 0.32 mmol; TBA salts, 3.0 mmol;
Fig. 2. Comparison of experimental data with calculated data.
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; Mn(OAc)2, 0.32
mmol; (TBA)4SiW6Mo6O40, 0.047 mmol; TBAB, 3.0 mmol; M.S. 3A,
M.S., 7.0 g; total pressure, 8 atm, CO, 500 ml minꢀ1; O2, 35 ml minꢀ1
;
7.0 g; total pressure, 8 atm; CO, 500 ml minꢀ1; O2, 35 ml minꢀ1
80 8C, 3 h. Determined by GC. DPC yield (%)ꢂ2ꢃ(DPC produced,
mol)/(reacted phenol, mol)ꢃ100.
;
80 8C, 3 h.
b
GC yield based on charged phenol.
TON based on Pd.
/
/
c
/

98
I. Hatanaka et al. / Journal of Organometallic Chemistry 674 (2003) 96Á100
/
stand the reaction mechanism. Fig. 5 gives these results.
From this figure, one can see that Bu3N gives a very
slow and low values of TON. On the contrary, TBAB
gives a rather higher TON. This means that an easy
coordination of a salt to form a phenoxy salt helps to
attain a higher TON. It may be necessary first to form a
phenoxy salt to proceed the reaction. In this sense, Bu3N
seems difficult to form a phenoxy salt.
2.6. Influence of the amount of H3PW6Mo6O40,
Mn(OAc)2 and TBA salt
Fig. 4. DPC yield and pKa of conjugated acid of TBA salts.
Then we investigate the effect of the amount of
H3W6Mo6O40 on the DPC product. The results are
shown in Fig. 6. DPC production increased until the
amount was 0.047 mmol and stayed constant until 0.06
mmol, then decreased rather sharply with the increase of
the amount of HPA. The rather sharp decrease may be
caused due to the difficult solubility of HPA to the
system. The optimum amount of Mo to Pd is around
result suggests that the use of the ammonium salt of an
HPA with the conjugated acid’ pKa lower than ꢀ9 gives
/
a good result. When negative entropy and the addition
of a large amount of ammonium salt are considered in
this reaction, a plausible active complex may be formed
with Pd, Mn, HPA, tetraalkylammonium cation and
PhOꢀ. These results support that the best coordination
to form phenoxy salt for the production of DPC is only
achieved by bromide anion. Other simple anions are
inferior to bromide anion in this sense.
2.2Á2.8, a little bit smaller than 3. This number shows
/
that about three Pd may be combined with 1 mol of
HPA to form an optimum active catalyst. The optimum
amount of Mn to Pd is 2 as shown in Fig. 7. The
optimized amount of the ammonium salt to HPA is
around 100 as shown in Fig. 8. In combination with
these results, we may be able to induce a general idea of
the active complex. It is shown in Fig. 9. In this figure,
Pd shows Pd metal or a variety of Pd(II) complexes to
avoid complexity. Mn also shows Mn(II) or Mn(III)
complexes. More oxonium ions are thought to be bound
to Mo (or W) of HPA or Mn ions actually. In order to
get a high turnover of Pd catalyst and a high yield of
DPC, It is requisite that the catalyst system is dissolved
in the reaction system. This was attained by an HPA and
a tetraalkylammonium salt by the formation of an active
catalyst complex. An HPA and a tetraalkylammonium
salt are important. Pd metal in the complex does not
form Pd metal cluster which is not easily reoxidized to
Pd(II). Pd metal in the complex must be coordinated by
HPA. In this sense Pd(0) may be bound to O atoms in
HPA. Mn ion is also combined with HPA in the same
way as in Pd. So the electron pumping from Pd(0) to
2.4. Effect of NaBr
The addition of NaBr to the reaction system was
investigated as shown in Table 2. The addition of NaBr
instead of tetraalkylammonium salt decreased the pro-
duction of DPC dramatically. It almost prevents the
reaction. NaBr might decrease the formation of a
phenoxy ammonium salt which would be a very effective
intermediate. Therefore, the formation of phenoxy salt
is requisite for this reaction. Addition of a large amount
of NaBr caused a difficulty of agitation to give an
inferior result.
2.5. Effect of the amount of an amine and ammonium
salts on Pd TON
It is also important to investigate the effect of the
amount of an amine and an ammonium salt to under-
Table 2
Effect
of
amount
of
NaBr a
Entry
NaBr (mmol)
Yield (%) b
TON c
1
2
3
3
30
0.83
0.23
12.8
3.8
Á/
22.7
354
a
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; Mn(OAc)2,
0.32 mmol; (TBA)4SiW6Mo6O40, 0.05 mmol; TBAB, 3 mmol; M.S., 7
g; THF, 15 ml minꢀ1; 8 atm; CO, 670 ml minꢀ1; O2, 47 ml minꢀ1
;
Fig. 5. Relations of the amount of amines vs TON. Conditions:
phenol, 50 mmol; Pd(OAc)2, 0.016 mmol; Mn(OAc)2, 0.032 mmol;
80 8C, 9 h.
b
GC yield based on charge phenol.
TON based on Pd.
H3PWMo11O40, 0.0047 mmol; M.S., 0.7 g; mixture gas (CO:O2ꢂ
/
c
11:1), 40 ml minꢀ1; 80 8C, 3 h. GC yield based on Pd(OAc)2.

I. Hatanaka et al. / Journal of Organometallic Chemistry 674 (2003) 96Á
/
100
99
Fig. 8. The influence of the amount of TBAB on the reaction.
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; H3PW6Mo6O40
0.05 mmol; Mn(OAc)2, 0.32 mmol; M.S. 3A, 7 g; total pressure, 8 atm;
Fig. 6. The influence of the amount of H3PW6Mo6O40 on the reaction.
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; Mn(OAc)2, 0.32
mmol; TBAB, 3.0 mmol; M.S. 3A, 7 g; total pressure, 8 atm; CO, 500
ml minꢀ1; O2, 35 ml minꢀ1; 80 8C, 3 h. Determined by GC.
,
CO, 500 ml minꢀ1; O2, 35 ml minꢀ1; 80 8C, 3 h. Determined by GC.
Mn3ꢁ may be done through O atoms. The solubility of
the active complex may be controlled by the number of
the combined ammonium cations on the active complex.
The more the ammonium cations attach to the complex,
the more it becomes soluble in the system. The more the
number of cations around the complex, the more it may
attract the phenoxy ion around Pd and induces high
yield of DPC and high TON of Pd. The production of
very small amount of CO2 and high selectivity of DPC
in this system also support the coordination of Mn to
HPA anion. The oxidation of HPA may be occurring
through the combination of Mo (or W) and O2 in the
ordinary manner [15].
give PhOCOÃ
/
PdÃ
/
OPh. This complex further experi-
ences reductive elimination to give Pd(0) and DPC. The
Pd(0) would be reoxidized to Pd(II) by the process
shown in the literature [15]. The experiments to prove
the formation of phenoxy salt from phenol and ammo-
nium salt, and the proof of CO complex formation are
now under investigation.
3. Conclusion
The reaction to synthesize DPC by an oxidative
carbonylation of phenol with CO and O2 proceeds
through the second-order of PhOH concentration. The
Based on the results obtained so far, we propose the
following mechanism (Fig. 10). In this figure, only the
real catalytic part of Pd is shown. To understand the
catalytic situations, it is better to replace each Pd in Fig.
activation energy Ea, DS and DH are 27.0 kcal molꢀ1
,
ꢀ
6.43 cal molꢀ1 and 26.3 kcal molꢀ1, respectively. The
values obtained in experiments agree with the proposed
mechanism as follows: Pd(OAc)2 reacts with an ammo-
/
9 with each of the Pd complexes (A)Á(E) in Fig. 10
/
depending on the complex mentioned. The acetate
ligand anion in Pd(OAc)2 is replaced by a phenoxy
anion of a tetraalkylammonium phenoxide to give a
nium phenoxy salt to give AcOÃ
reacts with CO to form AcOÃPdÃ
leads to PhOCOÃPdÃOPh that undergoes reductive
elimination to give DPC and Pd(0). Pd(0) is reoxidized
to Pd(II) by the help of an HPA very effectively.
/
PdÃ
/
OPh which then
/
/COOPh. This species
PhOÃ
/
Pd salt [2]. It allows to induce carbonylation by
/
/
CO to give a PhOCOÃ
/
PdÃ/OAc complex. This species
undergoes a further replacement of an acetate anion to
Fig. 7. The influence of the amount of Mn(OAc)2 on the reaction.
Conditions: phenol, 0.5 mol; Pd(OAc)2, 0.16 mmol; H3PW6Mo6O40
0.05 mmol; TBAB, 3.0 mmol; M.S. 3A, 7 g; total pressure, 8 atm; CO,
,
500 ml minꢀ1; O2, 35 ml minꢀ1; 80 8C, 3 h. Determined by GC.
Fig. 9. Image diagram of the active catalyst.

100
I. Hatanaka et al. / Journal of Organometallic Chemistry 674 (2003) 96Á100
/
Fig. 10. Possible mechanism.
4. Experimental
Acknowledgements
(TBA)SiW6Mo6O40 and (TBA)SiMo11O40 were
synthesized according to the reported method [17]. All
other reagents were obtained from commercial sources
and used as received. Molecular sieve (M.S.) was
activated at 450 8C in air for 10 h before use. All the
reactions are run in an autoclave by flow system except
the one done for Pd TON using tributylamine and
TBAB under a normal pressure.
T.Y., K.K. and T.K. thank RITE (Research Institute
of Innovative Technology for the Earth) for the support
of this research. HPAs were generously gifted by
Nippon Inorganic Color and Chemical Co. Ltd.
References
A typical reaction procedure is as follows: Pd(OAc)2
4H2O (0.078 g, 0.32
(0.036 g, 0.16 mmol), Mn(OAc)2×
/
[1] D. Freitag, U. Grico, P.R. Muller, W. Nouvertne, in: H.F. Mark
(Ed.), Polycarbonate in Encyclopedia of Polymer Science and
Engineering, 11, Wiley, New York, 1987, p. 649.
[2] J.E. Hallgren, G.M. Lucas, J. Organomet. Chem. 212 (1981) 135.
[3] M. Takagi, H. Miyagi, T. Yoneyama, Y. Ohgoinori, J. Mol.
Catal. A 129 (1998) L1.
mmol), (TBA)SiW6Mo6O40 (0.047 mmol), TBAB (1 g, 3
mmol), 3A powder M.S. (7 g) and hot phenol (47 g, 0.5
mol) were charged in a 100-ml stainless steel autoclave
which was kept at 60 8C. In case of Arrhenius plotting
experiments, the reactor was first heated to and kept at
the reaction temperature. The autoclave was then sealed.
With mechanical stirring at 800 rpm, a gas mixture (6.5
mol% O2 and 93.5 mol% CO) was introduced into the
autoclave at a flow rate of 535 ml minꢀ1 and the
autoclave was heated to 80 8C. The pressure of the
autoclave was kept at 8 atm with continuous gas flow
for the desired time. After the reaction, the pressure was
reduced to an atmospheric pressure, and the gas mixture
was exchanged with nitrogen. Then, the hot reaction
solution was transformed into a receiver and 30 ml of
EtOAc was used to wash the autoclave. Finally, the
solutions were combined and analyzed by GC with a
[4] M. Goyal, R. Nagahata, J. Sugiyama, M. Asai, M. Ueda, K.
Takeuchi, Catal. Lett. 54 (1998) 29.
[5] H. Ishii, M. Ueda, K. Takeuchi, M. Asai, J. Mol. Catal. 144
(1999) 369.
[6] M. Goyal, R. Nagahata, J. Sugiyama, M. Asai, M. Ueda, K.
Takeuchi, J. Mol. Catal. A 137 (1999) 147.
[7] A. Vavasori, L. Toniolo, J. Mol. Catal. A 139 (1999) 109.
[8] E.J. Pressman, S.J. Shafer, US Patent 5,898,079 (1999).
[9] H. Kezuka, F. Okuda, JP open, 1995, p. 39247.
[10] H. Buysch, C. Hesse, J. Jentsch, J. Rechner, E. Zirngiebl, Eur.
Patent 0,736,324 A2 (1996).
[11] J. Tsuji, Palladium Reagents and Catalysts, Wiley, New York,
1995.
[12] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199.
[13] K. Bergstad, H. Grennberg, J. Backvall, Organometallics 17
(1998) 45.
capillary column (DB-1, 15 mꢃØ 0.53 mm, J & W
/
Scientific). The products were identified by NMR
spectra (JEOL, JNM-AL300). The major side products
were phenylsalicylate and 4-phenoxy phenol. The yield
of DPC is based on the charged phenol and determined
by GC using diphenyl as an internal standard and is
listed in the tables. The selectivity is based on the DPC
percentage in the products evaluated from the peak
areas on the GC chart.
[14] G.C. Yin, C.G. Jia, T. Kitamura, T. Yamaji, Y. Fujiwara, Catal.
Lett. 69 (2000) 89.
[15] G.C. Yin, C.G. Jia, T. Kitamura, T. Yamaji, Y. Fujiwara, J.
Organomet. Chem. 630 (2001) 11.
[16] Md. Asadullah, Ph.D. thesis, Department of Chemistry and
Biochemistry, Graduate School of Engineering, Kyushu Univer-
sity, 2000.
[17] E. Yoshisato, Private communication, Polymer Research Insti-
tute, Teijin Ltd.