Gas phase transesterification of dimethylcarbonate and phenol over supported titanium dioxide

Article abstract of DOI:10.1006/jcat.1999.2495

The transesterification of dimethylcarbonate and phenol has been studied in a continuous gas flow reactor at high temperatures which were found to be favorable thermodynamically for high yields of methylphenylcarbonate (MPC). Among various solid catalysts, TiO₂/SiO₂ showed the highest activity and selectivity for MPC. The structure and the chemical state of titanium species in TiO₂/SiO₂ have been investigated by means of X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and X-ray absorption near edge structure (XANES) of Ti K-edge. It was observed that the titanium species was highly dispersed on silica. Below 10 wt%Ti loading, the titanium phase was not observed by XRD, yet weak XRD peaks of anatase were detected at higher loadings. The Ti K-edge XANES spectra and XPS analyses indicated that Ti(IV) species in the form of a monolayer was dominant below 5 wt% Ti loadings and TiO₂ of the anatase structure appeared at higher loadings. The amount of the surface Ti(IV) species measured by XPS increased with Ti loadings and was saturated above 10 wt% in the same manner as the selectivity to MPC changed with Ti loadings. This suggested that surface Ti(IV) species was directly responsible for the selective synthesis of MPC. The crystalline anatase TiO₂ was also an active and selective catalyst for the transesterification, yet it contributed to decrease in activity by coking.

Full text of DOI:10.1006/jcat.1999.2495

Journal of Catalysis 185, 307–313 (1999)
Article ID jcat.1999.2495, available online at http://www.idealibrary.com on
Gas Phase Transesterification of Dimethylcarbonate and Phenol
over Supported Titanium Dioxide
Won Bae Kim and Jae Sung Lee1
Department of Chemical Engineering/School of Environmental Engineering, Pohang University of Science
and Technology (POSTECH), San 31 Hyoja-dong, Pohang 790-784, Korea
Received September 28, 1998; revised March 8, 1999; accepted March 19, 1999
The fact that environmental costs are an integral part of
process economics is having a profound influence on tech-
nical developments in the chemical, petroleum refining, and
power generation industries. Particularly important is the
application of catalytic technologies that result in processes
that use fewer toxic raw materials, eliminate by-products
that are hazardous and difficult to dispose of, and lower
emissions (7).
Diphenylcarbonate, a convenient intermediate for the
synthesis of polycarbonate without using phosgene, has of-
ten been synthesized via a two-step reaction from dimethyl-
carbonate (DMC) and phenol (8–11) since the direct syn-
thesis of DPC is limited due to low equilibrium constants
for the forward reaction (9). The first step, transesterifica-
tion of DMC and phenol proceeds as shown in reaction [1]
and formed methylphenylcarbonate (MPC) is dispropor-
tionated into DPC and DMC in reaction [2]:
The transesterification of dimethylcarbonate and phenol has
been studied in a continuous gas flow reactor at high temperatures
which were found to be favorable thermodynamically for high yields
of methylphenylcarbonate (MPC). Among various solid catalysts,
TiO2/SiO2 showed the highest activity and selectivity for MPC. The
structure and the chemical state of titanium species in TiO2/SiO2
have been investigated by means of X-ray diffraction (XRD), X-ray
photoelectron spectroscopy (XPS), and X-ray absorption near edge
structure (XANES) of Ti K-edge. It was observed that the titanium
species was highly dispersed on silica. Below 10 wt% Ti loading, the
titanium phase was not observed by XRD, yet weak XRD peaks of
anatase were detected at higher loadings. The Ti K-edge XANES
spectra and XPS analyses indicated that Ti(IV) species in the form
of a monolayer was dominant below 5 wt% Ti loadings and TiO2
of the anatase structure appeared at higher loadings. The amount
of the surface Ti(IV) species measured by XPS increased with Ti
loadings and was saturated above 10 wt% in the same manner as
the selectivity to MPC changed with Ti loadings. This suggested
that surface Ti(IV) species was directly responsible for the selective
synthesis of MPC. The crystalline anatase TiO2 was also an active
and selective catalyst for the transesterification, yet it contributed
c
==
CH3O(C O)OCH3 + C6H5OH
*
)
==
C H O(C O)OCH3 + CH3OH
[1]
6
5
to decrease in activity by coking.
1999 Academic Press
==
2 C6H5O(C O)OCH3
Key Words: transesterification; gas phase; dimethylcarbonate;
phenol; TiO2/SiO2; XANES; active phase.
*
)
==
==
C H O(C O)OC6H5 + CH3O(C O)OCH3. [2]
6
5
Usually, the transesterification of DMC and phenol is
carried out in the liquid phase using homogeneous cata-
lysts such as organic Pb, Sn, or Ti compounds. Recently,
Fu et al. (11) reported that molybdenum oxide supported
on silica showed the highest activity for both the transes-
terification and the disproportionation in the liquid phase
reaction. This is one of a few reports on the development
of active solid catalysts for the reaction. However, there ex-
ists a critical thermodynamic limitation for the synthesis of
DPC from DMC and phenol, especially in the reaction 1 (9).
Tundo et al. (12) reported the reaction of DMC with phe-
nol under continuous flow of gaseous reactants over a solid
bed supporting a liquid phase-transfer catalyst. A spherical
macroporous -alumina (3 mm in diameter, 0.03 m2 g 1 sur-
face area) was coated with 5 wt% of potassium carbonate
and 5 wt% of PEG 6000 and used as support for the base
INTRODUCTION
Polycarbonate has been conventionally produced by the
interfacial polycondensation of bisphenol-A and phosgene.
Drawbacks of the conventional phosgene process include
environmental and safety problems involved in the use
of a copious amount of methylenechloride as the solvent,
which is 10 times the weight of the products, and a highly
toxic phosgene as the reagent (1). Hence, phosgene-free
processes for polycarbonate have been proposed that em-
ploy melt transesterification (2–5) or solid state polymeriza-
tion (1, 6) using bisphenol-A and diphenylcarbonate (DPC)
with the latter synthesized in a phosgene-free process.
1 To whom correspondence should be addressed. Fax: 82-562-279-5799.
E-mail: jlee@postech.ac.kr.
307
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

308
KIM AND LEE
and the catalyst. However, anisole was the sole product. Characterization of the TiO2/SiO2 Catalyst
They suggested that the transesterification between DMC
The specific surface areas of the catalysts were deter-
and phenol was not favored thermodynamically with Keq of
mined on a constant volume adsorption apparatus (Mi-
crometrics ASAP 2021C surface area analyzer) by the N2
BET method at the liquid nitrogen temperature. Powder
X-ray diffraction (XRD) measurements were conducted on
a MAC Science Co. Model M18XHF X-ray diffractome-
ca. 3 10 4 at 453 K (12).
In this work, the transesterification of DMC and phe-
nol (reaction [1]) was performed in a continuous gas flow
fixed bed reactor over solid catalysts at high temperatures
which were found to be more favorable from the thermody-
namic calculation of the reaction [1] (13). Among various
catalysts tested, titanium oxide supported on SiO2 was the
most active catalyst for the transesterification of DMC and
phenol in the gas phase. The same catalyst was also found
to be highly efficient in the disproportionation of MPC in
the liquid phase reaction (13, 14). We employed several
spectroscopic techniques such as X-ray diffraction (XRD),
X-ray photoelectron spectroscopy (XPS), X-ray absorption
near edge structure (XANES), and N2 adsorption to in-
vestigate the structure of the active components and the
correlation between the structure and the activity of the
TiO2/SiO2 catalyst in the gas phase transesterification of
dimethylcarbonate and phenol.
˚
ter using a radiation source of Cu K ( = 1.5405 A) at
40 kV and 200 mA with a scanning rate of 4 min 1. X-
ray photoelectron spectroscopy (XPS) measurements were
carried out usinga Perkin-Elmer PHI 5400ESCA spectrom-
eter with monochromatic Mg K radiation (1253.6 eV) at
15 kV and 20 mA. The vacuum in the measurement cham-
ber, during the collection of spectra, was maintained below
9
5
10 Torr. The pass energy of the measurement was
89.45 eV. Charging effect was circumvented by referenc-
ing binding energies to that of C1S peak at 284.6 eV. X-
ray absorption near edge structure (XANES) spectra were
taken of the K-edge of Ti in a transmission mode at beam-
line 3C1 of the Pohang Accelerator Laboratory (PAL) in
Pohang, Korea, operating at 2.0 GeV with ca. 100–150 mA
of the stored current. Double-crystal Si(111) monochro-
mators were used for the measurement. The detector gases
were N2 for both the incident and the transmitted beams.
The obtained data were analyzed using the UWXAFS 3.0
package licensed from University of Washington. The stan-
dard procedure is described in detail elsewhere (15).
EXPERIMENTAL
Catalysts Preparation
To prepare TiO2/SiO2 catalysts, SiO2 (Aldrich 23683-7)
was impregnated with a solution of tetrabutoxytitanium
(Aldrich 24411-2) dissolved in toluene. Samples of different
metal loadings given in weight percent by metal were dried
in an oven at 383 K for 12 h to remove the organic solvent
RESULTS
and calcined in a quartz reactor at 773 K for 4 h with an air
Gas Phase Transesterification of DMC and Phenol
over TiO2/SiO2 Catalysts
1
stream of 89 mol s
.
In catalyst screening experiments in a fixed bed, continu-
ous gas flow reactor, it was found that the most efficient
catalyst was TiO2 supported on activated carbon or sil-
ica (14). The catalyst performed significantly better than
MoO3/SiO2 which was the catalyst recommended by Fu
et al. (11) for the liquid phase transesterification of DMC
and phenol. Hence, TiO2/SiO2 catalyst was chosen for fur-
ther study and its properties were investigated in more de-
tail in the following. Phenol conversion and selectivity to
methylphenylcarbonate (MPC) was determined at a time
on stream of ca. 8.5 h when a steady state reaction was es-
tablished. The MPC selectivity (% ) is defined as the moles
of MPC produced per 100 mole of consumed phenol and
MPC yield (% ) is obtained from multiplication of phenol
conversion by MPC selectivity.
Gas Phase Transesterification of DMC and Phenol
The catalysts were pretreated under a flow of nitrogen
1
gas at a rate of 15 mol s at 773 K for 1 h just prior to
the delivery of the mixture of DMC and phenol. The re-
actant mixture was a solution of DMC (Aldrich D15292-7)
and phenol (Aldrich 18545-0) with a ratio of 5 mol DMC to
1 mol phenol. The reactants were introduced into the flow
1
of nitrogen gas of 15 mol s via a syringe pump (Sage
Instruments Model 361) at a pumping rate of 1.0 cm3 h 1 at
atmospheric pressure. A vaporization chamber filled with
glass beads was positioned upstream of a Pyrex U type reac-
tor to ensure a complete mixingofgaseousreactants. Identi-
fication and quantitative analysis of reaction products were
carried out by an on-line HP 5890II gas chromatograph
(GC) equipped with a flame ionization detector (FID) and
by a GC-Mass spectrometer of HP 5890II-HP5972MSD.
An HP PONA capillary column (50 m 0.2 mm 0.5 m)
was used to separate products for GC analysis. A sample
was injected into GC every 20 min by a six-port sampling
valve attached to the effluent stream line.
It was observed that the major by-products were anisole
and cresol, which could be produced via methylation ofphe-
nol with DMC. Generally, more anisole was formed than
cresol by a few times with all catalysts through a nucle-
ophilic attack of methyl group in DMC by the phenoxy
compound. However, little diphenylcarbonate (DPC) was

TRANSESTERIFICATION OF DIMETHYLCARBONATE AND PHENOL
309
as our standard reaction temperature. The phenol conver-
sion increased continuously as the concentration of DMC
in the feed was increased at a fixed phenol feed. However,
MPC selectivity showed a maximum at DMC/phenol ratio
of 5 and this ratio became a standard reaction condition.
Figure 1 shows the steady state activities of the transester-
ification reaction at 703K asa function ofweight percentage
of titanium in TiO2/SiO2 catalysts. The conversion of phe-
nol and the yield of MPC both showed maxima at 10 wt%
Ti loading. The selectivity toward MPC increased with Ti
loading up to 10 wt% and remained constant at about 87%
with higher Ti loadings. It should be noted that silica it-
self (points at 0% Ti loading) had appreciable activity, yet
selectivity was low at about 52% .
X-ray Diffraction
The structure and the chemical state of titanium ox-
ide species on silica have been investigated by means of
several spectroscopic methods such as X-ray diffraction,
X-ray photoelectron spectroscopy, and X-ray absorption
fine structure. Figure 2A shows the results of X-ray diffrac-
tion (XRD) analyses for the fresh TiO2/SiO2 catalysts with
different Ti loadings. At low Ti loadings, the titanium phase
was not observed. Weak, but characteristic peaks of anatase
type TiO2 were detected for TiO2/SiO2 catalysts with Ti
loadings greater than 10 wt% . The used catalysts showed
FIG. 1. The activities at steady state for the transesterification of
DMC and phenol as a function of weight percentage of titanium loaded.
Reactant mixture feeding rate, 1.0 cm3 h 1; mole ratio of DMC to phenol, 5;
N2 flow rate, 15 mol s 1; TiO2/SiO2 catalyst loading, 0.48 g; reaction tem-
perature, 703 K.
produced, indicating that there was an insufficient contact
time for the consecutive transesterification of MPC with
phenol to occur. It was found that the production of DPC
became appreciable as the space velocity of reactants de-
creased further.
Gas phase reactivity of DMC over TiO2/SiO2 catalysts
was more complicated than that of liquid phase as discussed
elsewhere (16). When gas phase DMC itself is supplied,
without phenol over TiO2/SiO2, dimethylether, methanol,
CO2, and carbon deposit were produced. At the initial
stage of the transesterification of DMC and phenol, DMC
was rapidly and completely converted to dimethylether,
methanol, CO2, and carbon deposit as in the case without
phenol. At the steady state of the reaction, however, most
of converted DMC reacted with phenol to produce MPC
and by-products such as anisole and cresol while very little
DMC converted to dimethylether. Overall, the MPC selec-
tivity based on DMC was more than 90% at the steady state
in all cases.
We carried out a preliminary study on the effects of reac-
tion variables on the performance of the catalysts and a part
of the results including temperature effects were reported
elsewhere (14). Briefly, the conversion of phenol increased
with increasing temperature below 703 K, but it showed a
much slower increase above 703 K. The selectivity for MPC
did not show much variation with reaction temperatures of
FIG. 2. X-ray diffraction patterns of TiO2/SiO2 catalysts with the dif-
640–750 K and remained at 86 5% . Hence, we chose 703 K ferent Ti loadings before (A) and after the reaction (B) at 703 K for 8.5 h.

310
KIM AND LEE
X-ray Absorption Near Edge Structure
X-ray absorption near edge structure (XANES) is a pow-
erful tool used to investigate the local structure around an
absorbing atom. For this reason the coordination state of Ti
has often been probed by XANES at the Ti K-edge. Ti(IV)
is in a d0 configuration corresponding to A1 or A1g states de-
pending on tetrahedral or octahedral symmetry (17). The
first feature in XANES of Ti K-edge is the pre-edge corre-
sponding to the excitation of 1s electrons into empty bound
states derived from the d and p states of Ti and O atoms.
In tetrahedral symmetry, since the final states are T2 and
E, the transitions of A1 into T2 and E occur. As consid-
erable mixing of 3d and 4p orbitals occurs, the A1-into-T2
transition is allowed and the corresponding pre-edge peak
became particularly strong. This transition overwhelms the
weaker A1-into-E transition and therefore tetrahedral sym-
metry of Ti(IV) shows a strong single peak in the pre-edge
region. In the case of octahedral symmetry, the transition
of 1s electrons occurs into empty d orbitals and the fi-
nal states are T2g and Eg. Hence, expected transitions are
A1g-into-T2g and A1g-into-Eg, which are both Laporte for-
bidden. Consequently, the octahedral symmetry of Ti(IV)
shows weaker pre-egde absorption peaks than that for the
FIG. 3. The surface concentration of Ti as a function of Ti loading in
TiO2/SiO2 before reaction determined from XPS quantitative analyses.
exactly the same patterns of XRD as shown in Fig. 2B,
implying that the sintering of catalyst during the reaction
had not occurred.
X-ray Photoelectron Spectroscopy
The investigation of binding energy and intensity of some tetrahedral symmetry. Usually, three weak peaks are ob-
surface elements by X-ray photoelectron spectroscopy served. Two of them are associated with A1g-into-T2g and
(XPS) gives the information of the chemical states and A1g-into-Eg transitions as described above. The third peak,
relative quantities of the outermost surface compounds. which is at the lowest energy, is of uncertain origin. Anal-
The Ti 2p3/2 binding energies of the TiO2/SiO2 catalysts yses of XANES around the Ti K-edge are presented here,
were invariant with Ti loading in the range of 458.3 0.2, comparing the spectra of the catalysts with those of refer-
which corresponds to the binding energy of titanium diox- ence compounds.
ide. Binding energies of Ti 2p3/2 electrons are 454.0 0.2,
Figure 4 shows the XANES spectra of reference Ti com-
455.0 0.2, and 458.8 0.3 eV for titanium metal, titanium pounds, namely titanium metal, anatase type and rutile type
monoxide, and titanium dioxide, respectively.
Figure 3 represents the Ti atomic concentration obtained
from quantitative XPS analyses of the outermost surface of
TiO2/SiO2 catalysts as the weight percentage of Ti is in-
creased. The surface concentration of Ti increased rapidly
up to 5 wt% Ti. Between 5 and 10 wt% Ti, the surface
concentration increased at a slower rate. Above 10 wt%
the surface concentration of Ti reached a constant value of
4.2% . This result indicates the mode of TiO2 dispersion on
the support as Ti loading is raised. Below 5 wt% , TiO2 is
highly dispersed probably as monolayer, but in higher load-
ings it begins to be crystallized into bulk titania. Especially
above 10 wt% Ti, further added titania contributes to the
growth of crystal phase, but not to the dispersion of tita-
nia on the silica surface. The surface compositions (% ) of
other components measured by XPS with different Ti load-
ings are as follows in the order of O, Si, and C: 70.4, 25.6,
and 3.2 for 1 wt% catalyst, 67.4, 23.2, and 6.3 for 5 wt% cata-
lyst, 68.1, 24.2, and 3.6 for 10 wt% catalyst, 67.7, 21.1, and
6.9 for 13 wt% catalyst, and 66.5, 18.6, and 10.7 for 20 wt%
TiO2/SiO2 catalyst. As Ti loading increased, surface con-
centration of Si decreased while that of carbon increased.
FIG. 4. X-ray absorption near edge structures (XANES) of Ti K-edge
for reference Ti compounds.

TRANSESTERIFICATION OF DIMETHYLCARBONATE AND PHENOL
311
titania. The pre-edge peak of Ti metal was very sharp and
high. The high and single peak in the pre-edge region indi-
cates that the Ti in the metal has a tetrahedral symmetry.
However, this spectrum of Ti(0) species could be differen-
tiated from that of tetrahedral Ti(IV) species which shows
similar pre-edge peak shape but its absorption edge is ob-
served at a higher energy. Anatase and rutile showed three
weak pre-edge peaks, indicating an octahedral symmetry. In
the post-edge region, the resonance patterns of the absorp-
tion coefficient for anatase and rutile titania are somewhat
different. The resonances present in the rutile phase de-
rived from constructive interference at ca. 4992 eV and ca.
5020 eV were not present in anatase.
Figure 5 represents Ti K-edge XANES spectra of TiO2/
SiO2 catalysts with different Ti loadings. In pre-edge re-
gion, only 5 wt% catalyst showed a single peak although it
is not as sharp as typical tetrahedral Ti(IV) species (17, 18).
Thus, 5 wt% TiO2/SiO2 catalyst appears to contain both
tetrahedral Ti(IV) species and octahedral Ti(IV) species.
Catalysts with higher Ti loadings ( 10 wt% ) showed the
pre-edge feature of typical octahedral symmetry. In post-
edge region of spectra, the characteristic peaks of rutile
type titania were not observed, such as resonance peaks
at 4992 and 5020 eV, indicating titanium oxide species on
silica were Ti(IV) dioxide of anatase form. With increas-
ing Ti loadings, the peak intensity increased, especially that
at ca. 5004 eV. This feature of XANES spectra indicates
that the crystallinity of titania is augmented as Ti loading is
increased above 10 wt% of Ti.
FIG. 6. BET surface area of the TiO2/SiO2 catalysts as a function of
Ti loading before and after reaction at 703 K for 8.5 h.
an almost linear decrease of BET areas as Ti content in-
creased up to 20 wt% . The slope was very small and change
was within 10% for whole Ti loadings. However, the used
catalysts showed a more rapid decline of specific surface
area as Ti loading was increased. Thus, the surface area dif-
ference of the fresh and the used catalysts was greater for
higher Ti content.
DISCUSSION
In this work, a new method for the synthesis of diphenyl-
carbonate (DPC) from dimethylcarbonate (DMC) and
phenol was proposed. Methylphenylcarbonate (MPC) is
the inevitable intermediate in the synthesis of DPC from
DMC and phenol, but is difficult to be produced in a high
yield through the conventional liquid phase reaction due to
the low thermodynamic equilibrium constants. By using the
functional group contribution method (19), the standard
heat of reaction [1] in the gas phase was estimated to be
23.5 kJ mol 1 (13). Equilibrium yields of this endothermic
reaction should be more favorable as reaction temperature
increases. In this new scheme, TiO2/SiO2 was found to be
an efficient catalyst, giving a higher yield of MPC in a fixed
bed gas flow reactor at a relatively high temperature. DPC
could be obtained effectively through the liquid phase dis-
proportionation of MPC over the same catalyst (13). In the
present work, we investigated the structure and chemical
state of the active components of the TiO2/SiO2 catalyst sys-
tem and their relationship with catalytic activity in the gas
phase transesterification of dimethylcarbonate and phenol.
There are a large number of studies in the literature on
the structure of supported TiO2. Stranick et al. (20) em-
ployed XPS, XRD, and laser Raman spectroscopy (LRS) to
characterize the dispersion and chemical state oftitanium in
a series of TiO2/Al2O3 catalysts. Titania was well dispersed
on alumina for Ti loadings less than 14 wt% . The formation
Specific Surface Area
Figure 6 illustrates change in specific surface area for
TiO2/SiO2 catalysts at increasing Ti loadings before and
after the reaction. The fresh TiO2/SiO2 catalysts showed
FIG. 5. X-ray absorption near edge structures (XANES) of Ti K-edge
for the TiO2/SiO2 catalysts as a function of Ti loading.

312
KIM AND LEE
of a TiO2 :Al2O3 surface phase occurred in 1 to 6 wt% Ti
Quantitative analysis of surface titanium measured by
range. Formation of anatase titania occurred, in addition to XPS is presented in Fig. 3. Rapid growth of surface Ti
the surface phase, at loadings in excess of 6 wt% Ti, with concentration below 5 wt% Ti loading suggests that most
the amount of anatase increasing with Ti loadings. XRD added titanium contributes to the monolayer formation.
analysis of TiO2/Al2O3 revealed that segregation of crys- The slope of the surface Ti concentration vs Ti loading
talline anatase particles at a loading of 14 wt% Ti and the curve decreases above 5 wt% Ti loading and becomes flat
amount of crystalline anatase present was approximately from a Ti loading of 10–13 wt% . This shape of the curve
1.5 to 2 wt% , as estimated by XRD and LRS. Using the could be interpreted as Ti added above 5 wt% contributing
samples prepared from the grafting of silica with tetraiso- to the growth of both monolayer and bulk TiO2 and then
propoxytitanium in isopropanol, Castillo etal. (21) reported only to the three-dimensional growth above 10–13 wt% Ti
that a very good dispersion of TiO2 over silica (BET sur- loading.
face area of 320 m2 g 1) approaching monolayer dispersion
Three characterization techniques of XRD, XANES, and
was observed for TiO2 content lower than about 15 wt% . XPS provide a consistent picture on the evolution of tita-
At higher TiO2 content, TiO2 crystallites corresponding to nium structure in TiO2/SiO2 as the amount of titanium in-
octahedral coordination were formed. Lassaletta et al. (22) creased; a monolayer containing tetrahedral Ti(IV) at low
studied TiO2 thin film prepared by evaporation of Ti in the Ti loadings and bulk TiO2 crystals with anatase structure
presence of oxygen onto silica substrates. The supported at high loadings. A simple calculation shows that a Ti load-
titania phase showed a significant blue shift in its band gap ing of 14.5 wt% is needed to physically cover SiO2 with a
observed by UV–VIS absorption spectra which was higher surface area of 300 m2 g 1 with a monolayer of tetrahedral
than the value expected based on the thickness of the TiO2 TiO2 species. Figure 3 suggests that monolayer coverage is
layer. They ascribed such shifts to the fact that the coverage reached around 10 wt% Ti loading and the amount of Ti
layer was constituted from TiO2 crystallites smaller than the actually used for monolayer should be less than 10 wt%
thickness of the deposited layer. These previous works are because a part of it has formed three dimensional TiO2 par-
consistent with each other in TiO2 forming a well-dispersed ticles above 5 wt% Ti loading. Since the estimate assumes
phase (surface or monolayer phase) at low Ti loadings and a complete close-packed coverage of whole SiO2 surface
a crystalline TiO2 at high loadings.
with TiO2, it would overestimate the amount of Ti required
From the XRD results of Fig. 2, it was found that the for a monolayer coverage. Hence, it could be concluded
titanium species in our TiO2/SiO2 catalysts was highly dis- that agreement is reasonably good between estimation and
persed on silica support in agreement with previous reports observation of the saturation surface coverage.
discussed above. Below 5 wt% Ti loading, the titanium
It is interesting to see how this structural information
phase was not observed by XRD. The species in the cata- could be related to the reactivity of TiO2/SiO2 in the gas
lysts seemed to be a highly dispersed Ti phase due to the phase transesterification of phenol and dimethylcarbonate.
strong interaction between silica and titanium(IV) dioxide. Figure 1 shows that phenol conversion shows a maximum
However, only the catalyst of 5 wt% Ti loading showed a at 10 wt% Ti loading. The selectivity to MPC increases first
single pre-edge peak of Ti K-edge XANES, the signature and becomes flat at 10 wt% Ti loading as the loading in-
of a tetrahedral symmetry of Ti coordination although its creased. The behavior of selectivity at increasing Ti load-
spectrum represented a combined pre-edge peak of some ing is similar to that of surface Ti concentration shown in
fraction of octahedral Ti(IV) symmetry. The fine structure Fig. 3. The behavior of phenol conversion is also similar
of postedge spectrum of 5 wt% sample is also distinguished below 10 wt% loading. The results suggest that the sur-
from those of other spectra. Hence, this phase could be face titanium oxide is the active catalytic sites for the gas
assigned as a surface phase of TiO2 : SiO2 forming a mono- phase transesterification of phenol and dimethylcarbonate.
layer of TiO2 on SiO2 as in the case of TiO2/Al2O3 (20).
Both monolayer and bulk TiO2 appear to serve as active
At titanium loadings of 10 wt% and above, TiO2 par- sites because the good MPC selectivity is maintained above
ticles with the anatase structure is formed on SiO2. This 10 wt% . As mentioned, silica itself has a modest catalytic
three-dimensional growth of titania is unequivocally man- activity for the reaction showing phenol conversion of 14%
ifested in XRD (Fig. 2) and XANES (Fig. 5) results. As and MPC selectivity of 52% . Thus the activity and selectiv-
more titanium is added above 10 wt% , the additional ity of TiO2/SiO2 below 10 wt% loading should be combined
titanium contributes only to the growth of anatase parti- contributions of surface TiO2 and less active and less selec-
cles and sharper XRD peaks are observed. XANES fea- tive SiO2. This accounts for the continued increase of MPC
tures, on the other hand, do not change as Ti loading in- selectivity at low Ti loadings because added TiO2 would
creased above 10 wt% because the technique represents cover the less selective SiO2 surface.
the local structure surrounding the absorbing atom (Ti) and
is not sensitive to three-dimensional growth of the TiO2 creases as Ti loading goes above 10 wt% . This appears to be
crystals. due to the decrease in surface area at the high loadings as
Unlike selectivity to MPC, the conversion of phenol de-

TRANSESTERIFICATION OF DIMETHYLCARBONATE AND PHENOL
CONCLUSION
313
Silica-supported titania was the active catalyst for the
synthesis of methylphenylcarbonate (MPC) from dimethyl-
carbonate (DMC) and phenol in the gas phase. The amount
of surface Ti(IV) species highly dispersed due to a strong
interaction of titania and silica increased with Ti loadings
and got saturated above 10 wt% . This surface Ti(IV) species
was directly responsible for the selective synthesis of MPC
by the gas phase transesterification of DMC and phenol.
The crystalline TiO2 anatase phase was also an active and
selective catalyst for gas phase transesterification, yet it also
decreased the catalyst activity through the loss of surface
area caused by coking.
FIG. 7. The surface area of Ti obtained by the multiplication of the
steady state BET surface area and the surface concentration of Ti by quan-
titative XPS analysis.
ACKNOWLEDGMENT
This work has been supported by LG Chemicals Ltd. and Korean Sci-
ence and Engineering Foundation through the Research Center for Cata-
lytic Technology of Pohang University of Science and Technology.
REFERENCES
shown in Fig. 6. In order to demonstrate this point, the mul-
tiplication product of surface titanium concentration mea-
sured by XPS and BET surface area of used catalysts was
plotted against Ti loadings in Fig. 7. The product should be
proportional to the absolute amount of surface TiO2 taking
part in the catalytic reaction. The plot showed a remark-
able resemblance with the curve of phenol conversion vs
Ti loading. This supports our proposition that total surface
titanium sites are responsible for the catalytic transesterifi-
cation of phenol and dimethylcarbonate. The active Ti(IV)
species are present in the forms of a monolayer and bulk
crystals of the anatase structure or in tetrahedral and octa-
hedral symmetries.
In Fig. 7, it is evident that the decrease in BET sur-
face area is responsible for the reduced phenol conversions
when Ti loading is increased above 10 wt% . As shown in
Fig. 6, surface area of TiO2/SiO2 catalyst decreases as Ti
loading is increased. However, the decrease in the surface
area with Ti loading is not that great for fresh catalysts be-
fore the reaction. However, the decrease becomes much
greater after the transesterification at 703 K for 8.5 h and
thus the surface area difference between the fresh and used
catalysts has become greater for higher Ti loadings. In a
separate study to be published elsewhere, it has been found
that the cause of the surface area reduction for used cata-
lysts is coking (16). The presence of coke was confirmed by
surface analysis with XPS, temperature-programmed reac-
tion with H2, and the measurement of pore-size distribution
by N2 adsorption. Thus, Fig. 6 indicates that coking is more
favorable on catalysts with higher Ti loadings. Since these
catalysts contain more bulk TiO2 than catalysts with lower
Ti loadings, it could be concluded that bulk TiO2 promotes
coking more effectively than monolayer Ti phase.
1. Chem. Br. 30(12), 970 (1994).
2. Curtius, U., Bottoenbruch, L., and Schnell, H., U.S. patent 3 442 854
(1969).
3. Yamana, H., Kuni, T., Furusawa, T., Nakai, H., and Hiro, Y., U.S. patent
3 888 826 (1975).
4. Brunelle, D., U.S. patent 4 321 356 (1982).
5. Starr, J. B., and Ko, A., U.S. patent 4 383 092 (1983).
6. Komiya, K., Fukuoka, S., Aminaka, M., Hasegawa, K., Hachiya, H.,
Okamoto, H., Watanabe, T., Yoneda, H., Fukawa, I., and Dozono,
T., in “Green Chemistry: Designing Chemistry for the Environment”
(P. T. Anastas and T. C. Williamson, Eds.), p. 20. ACS, Washington,
1996.
7. Cusumano, J. A., Chemtech 22(8), 482 (1992).
8. Fukuoka, S., Tojo, M., and Kawamura, M., U.S. patent 5 210 268 (1993).
9. Harrison, G. E., Dennis, A. J., and Sharif, M., U.S. patent 5 426 207
(1995).
10. Murata, K., Kawahashi, K., and Watabiki, M., U.S. Patent 5 380 908
(1995).
11. Fu, Z.-h., and Ono, Y., J. Mol. Catal. A 118, 293 (1997).
12. Tundo, P., Trotta, F., Moraglio, G., and Ligorati, F., Ind. Eng. Chem.
Res. 27, 1565 (1988).
13. Kim, W. B., MS thesis, Pohang Univ. of Science and Technology, 1998.
14. Kim, W. B., and Lee, J. S., Catal. Lett., in press.
15. Teo, B. K., “EXAFS: Basic Principles and Data Analysis.” Springer,
Berlin, 1985.
16. Kim, W. B., Kim, Y. G., and Lee, J. S., Appl. Catal. A, submitted for
publication.
17. Bordiga, S., Coluccia, S., Lamberti, C., Marchese, L., and Zecchina,
A., J. Phys. Chem. 98, 4125 (1994).
18. Rhee, C. H., and Lee, J. S., Catal. Today 38, 213 (1997).
19. Liley, P. E., Reid, R. C., and Buck, E., in “Perry’s Chemical Engineers’
Handbook” (R. H. Perry and D. Green, Eds.), 6th ed., Section 3, p. 276.
McGraw-Hill, Singapore, 1984.
20. Stranick, M. A., Houalla, M., and Hercules, D. M., J. Catal. 106, 362
(1987).
21. Castillo, R., Koch, B., Ruiz, P., and Delmon, B., J. Catal. 161, 524 (1996).
22. Lassaletta, G., Fernandez, A., Espinos, J. P., and Gonzalez-Elipe, A. R.,
J. Phys. Chem. 99, 1484 (1995).