Influence of pore and crystal size of crystalline titanosilicates on phenol hydroxylation in different solvents

Article abstract of DOI:10.1006/jcat.2001.3308

The hydroxylation of phenol into catechol and hydroquinone with 30% aqueous H₂O₂ and titanium-substituted molecular sieves (TS-1 and Al-free Ti-Beta) was studied to understand the role of the zeolite structure, crystal size, external

Full text of DOI:10.1006/jcat.2001.3308

Journal of Catalysis 203, 201–212 (2001)
doi:10.1006/jcat.2001.3308, available online at http://www.idealibrary.com on
Influence of Pore and Crystal Size of Crystalline Titanosilicates
on Phenol Hydroxylation in Different Solvents
�
Uwe Wilkenh o¨ ner, Gunther Langhendries,† Frederik van Laar,‡ Gino V. Baron,† David W. Gammon,§
�,1
Pierre A. Jacobs,‡ and Eric van Steen
�
†
Catalysis Research Unit, Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch 7701, South Africa;
Department of Chemical Engineering, Vrije Universiteit Brussel, Pleinlaan 2, B-1050 Brussel, Belgium; ‡Centre for Surface Chemistry
and Catalysis, Katholieke Universiteit Leuven, Kasteelpark Arenberg 23, B-3001 Heverlee, Belgium; and §Department of Chemistry,
University of Cape Town, Private Bag, Rondebosch 7701, South Africa
Received April 2, 2001; revised June 15, 2001; accepted June 19, 2001
shown to be an active catalyst in the epoxidation of propy-
The hydroxylation of phenol into catechol and hydroquinone lene and higher �-olefins, in the ammoximation of cyclo-
with aqueous H2O2 (30%) and titanium-substituted molecular hexanone, and in the hydroxylation of phenol. The hy-
sieves (TS-1 and Al-free Ti-Beta) was investigated to understand
the role of the zeolite structure, the crystal size, the external sur-
face of the zeolite, and the nature of the solvent on the product
selectivity. Comparing Al-free Ti-Beta and TS-1 samples with sim-
ilar pore lengths, the activity as well as the ratio of catechol to
and procedure for H2O2 addition (5) have been identified
hydroquinone was significantly higher for Al-free Ti-Beta, showing
diffusional constraints for the conversion of phenol and geomet-
droxylation of phenol yielding hydroquinone and catechol
using aqueous hydrogen peroxide has been commercial-
ized by Enichem (1). Several parameters such as calcina-
tion conditions (2), nature of solvent (3), temperature (4),
as key parameters. However, even today, various aspects of
the phenol hydroxylation, such as the reaction mechanism,
contribution of the external surface, and diffusion of the
ric constraints for the formation of catechol. The diffusional con-
straints in the conversion of phenol were confirmed by using small
crystallites of TS-1. The role of the external surface of TS-1 in the reactants and products, are not yet fully understood.
phenol hydroxylation was investigated by inertization of the ex-
According to van der Pol et al. (6) phenol hydroxylation
ternal surface, using cycles of low-temperature chemical vapor de- is governed by intracrystalline diffusion in large crystals of
position (CVD) of tetraethoxysilane followed by high-temperature
calcination. Consecutive CVD cycles led to a slight increase of the
selectivity toward hydroquinone for all tested solvents as well as
a reduction of the coke formation in methanol and water. The ra-
controlled by the external surface activity. Hydroquinone
tio of hydroquinone to catechol, however, did not change much,
and catechol are thought to be the preferred reaction prod-
indicating that catechol must also be formed inside the pore struc-
ture. A kinetic analysis of the reaction data with the parent and
surface-inertized TS-1 revealed that the role of the external surface
TS-1. These authors excluded a major influence of external
surface activity. On the other hand, Tuel et al. (3) concluded
that the product ratio of hydroquinone to catechol is mainly
ucts in the micropores of TS-1 and on the external surface
titanium sites, respectively.
The initial objective of this work was to gain further in-
in terms of both activity and selectivity is significant and dependent
on the solvent used. A reaction mechanism consistent with the ob- sight into the mechanism(s) controlling product selectiv-
served enhanced selectivity for hydroquinone in protic solvents is ity. To achieve this, the influence of the pore structure, the
proposed.
�c 2001 Academic Press
Key Words: phenol hydroxylation; solvent effect; Al-free Ti-Beta;
TS-1; CVD; silanization; kinetic analysis; sorption.
crystal size, and the contribution of the external surface of
titanium silicalite-1 to overall reaction parameters, such as
phenolconversion, hydroquinone selectivity, and tar forma-
tion, were investigated in different solvents such as water,
methanol, and acetone.
1
. INTRODUCTION
The contribution of the external surface can be investi-
gated by eliminating the active sites located on the external
surface of TS-1 by cyclic chemical vapor deposition (CVD)
with tetraethoxysilane (TEOS) (7). The chemical vapor de-
position of silanes on zeolitic aluminosilicates, especially
ZSM-5, is a well-established technique to inertize the acid
sites on the external surface (8). The inertization may lead
to improvement in selectivity toward the product with the
Over the past 20 years, numerous reports on the synthesis
and application of titanium-substituted silicates have been
published. In particular, titanium silicalite-1 (TS-1) was
1
To whom correspondence should be addressed. Fax: +27-21-689 7579.
E-mail: evs@chemeng.uct.ac.za.
201
0021-9517/01 $35.00
Copyright �c 2001 by Academic Press
All rights of reproduction in any form reserved.

¨
WILKENHONER ET AL.
2
02
smallest kinetic diameter, as was observed in the toluene typical synthesis, 40 g of tetraethylorthosilicate (TEOS,
disproportionation to benzene and xylenes over H-ZSM- Aldrich) is hydrolyzed under stirring at room temperature
5
. With increasing number of cycles the p-xylene content in a solution containing 45.37 g of tetraethylammonium hy-
in the fraction of xylenes increases up to 99% , which was droxide (TEAOH, 35 wt% , Alfa), 6.40 g of H2O2 (35 wt%
explained in terms of pore mouth narrowing (7). in water, Fluka), and 2.29 g of water. After addition of 1.75 g
TS-1 has a three-dimensional pore structure with pore of tetraethylorthotitanate (Ti(OEt)4, Alfa), the solution is
dimensions of 5.6 � 5.4 A˚ while the larger pore molecu- stirred at room temperature until the ethanol formed is
lar sieve Ti-beta has a two-dimensional pore structure with evaporated. Finally, 4.49 g of HF (48 wt% , Aldrich) is added
pore dimensions of 7.6� 6.4 A˚ . Thus, the two materials can and the yellow thick paste is loaded in Teflon-lined stain-
�
be used to investigate the influence of the pore geometry less steel autoclaves and heated at 140 C under rotation
on the product selectivity and to identify shape-selective (60 rpm) for 20 days. The product obtained after the crys-
effects induced by the zeolite channels.
tallization is filtered, washed with deionized water, dried,
�
and then calcined at 550 C in static air for 8 h.
+
+
The alkali (Na + K ) content in all templates was less
than 50 ppm as confirmed by AAS.
2
. MATERIALS AND METHODS
The externalsurface ofTS-1, synthesized accordingto the
method described by Tangaraj et al. (4), was silanized using
a low-temperature deposition of tetraethoxysilane (TEOS,
Aldrich) (7). Nitrogen carrier gas (100 ml(NTP)/min) was
2
.1. Synthesis and Modification of Titanium-Containing
Molecular Sieves
The titanium-substituted zeolites used in this work
�
led through a TEOS-containing saturator at 15 C and
were two TS-1 samples with different crystal sizes and
aluminium-free zeolite Ti-Beta. Small crystallites of TS-1
were synthesized (Si/Ti = 34) according to the method de-
scribed by Tangaraj et al. (4). In a typical synthesis, 24.4 g
of tetraethylorthosilicate (TEOS, Aldrich) was hydrolyzed
with 29.3 g of tetrapropylammonium hydroxide solution
�
passed for 1 h at 100 C in the up-flow mode over a fixed
bed of TS-1 (particle size 0.1 �m), the space velocity
being 2.43 mg/(g � min). Subsequently, TS-1 was flushed
with pure N2 (100 ml(NTP)/min) for 0.25 h, heated in air
�
�
(
100 ml(NTP)/min) at 5 C/min up to 500 C, and kept at
this temperature for 2 h. This procedure was repeated up to
(
TPAOH, 20% in H2O, Alfa). A solution of 1.23 g of
2
0 times.
tetrabutylorthotitanate (TBOT, Alfa) in 5.55 g of dry iso-
propanol (Merck) was then added dropwise. Subsequently,
9
.8 g of TPAOH and 22.76 g of deionized water were added.
2
.2. Characterization of TS-1 by Adsorption
Before autoclaving, the reaction mixture was heated to
8
�
0 C to remove the alcohol. The molar composition of
The adsorption of phenol, hydroquinone, and catechol in
the initial mixture was as follows: 0.03 TiO2 : 1 SiO2 :25 TS-1 was determined from water, methanol, and acetone as
H2O : 0.32 TPAOH : 0.77 i-PrOH. This gel was then heated solvent using a batch and a chromatographic technique.
statically in Teflon-lined stainless steel autoclaves at auto-
In the batch adsorption method, 0.1 g of zeolite catalyst
�
�
geneous pressure at 170 C for 48 h. The resulting solid was wassuspended at 25 C in 5ml of a 5 wt% phenol–solvent so-
recovered, washed with deionized water, dried, and cal- lution. After adsorption equilibrium was established (24 h),
�
cined in static air at 550 C for 8 h to remove the template. a liquid sample was analyzed by gas chromatography. This
Large crystals of TS-1 (Si/Ti = 34) were synthesized ac- allowed us to calculate the phenol partition coefficient, �,
cording to the procedure of Milestone and Sahasrabudhe as the ratio of adsorbed phenol in the molecular sieve and
(
(
9). A typical synthesis mixture consisted of 1 g of SiO2 solvent phase.
Aerosil 200, Degussa), 0.5 g of TPABr (Fluka), 1.6 ml of The experimental set-up and conditions for the liquid-
a 40% solution of TPAOH (Alfa), 0.6 g of NH4F (Merck), phase chromatographic method have been described else-
.05 g of (NH4)2TiF6 (Aldrich), and 20 ml of deionized wa- where (11).
ter. The salts and the TPAOH were dissolved in the water Pore mouth narrowing due to silanization was investi-
0
and Aerosil was added and mixed until a homogeneous gated by a gas-phase chromatographic technique using an
gel was obtained. The synthesis mixture was autoclaved in experimental set-up similar to that used by Denayer et al.
Teflon-lined stainless steel autoclaves and heated statically (12), taking cyclohexane, 3-methyl-pentane, and toluene as
�
at autogeneous pressure at 170 C for 48 h. After crystalliza- tracer components.
tion, the solid was recovered, washed with deionized water,
dried, and calcined in static air at 500 C.
A hydrocarbon pulse (typically 1–4 �mol of hydrocar-
bon vapor in N2) was injected in a carrier gas (argon,
�
Aluminum-free Ti-Beta (Si/Ti = 40) was synthesized ac- 30 ml(NTP)/min), which was led over a 1 cm column (4 mm
cording to a procedure adopted by Blasco et al. (10) using a internal diameter) containing 0.05 g of parent or inertized
starting gel of the following molar composition: 1 TiO2 : zeolite powder. The column was mounted in a GC equipped
2
5 SiO2 : 14 TEAOH : 8.6 H2O2 : 189 H2O :14 HF. In a with a FID detector. The response on the pulse was

PHENOL HYDROXYLATION IN DIFFERENT SOLVENTS
203
�
�
�
monitored at 175–275 C and the shifts in the first and sec- (60 ml(NTP)/min) at 5 C/min to 150 C to desorb wa-
ond moments of the peaks were taken as a measure for ter. Subsequently, the carrier gas was changed to air
�
changes in the sorption behavior of TS-1 on inertization.
(60 ml(NTP)/min) and the sample was heated at 5 C/min
to 550 C and kept at this temperature for 4 h.
�
2
.3. Physicochemical Characterization
2
.5. Kinetic Modeling of Phenol Hydroxylation
All samples were characterized for silicon–titanium con-
tent (AAS), crystallinity (XRD), crystal size and morphol-
The data obtained in the phenol hydroxylation experi-
ogy (SEM), extraframework titanium (UV–VIS), and sur- ments were fitted in concentration–time profiles with the
face area (BET).
following second-order kinetic expressions, first order in
Powder X-ray diffraction spectra were taken by means both phenol and hydrogen peroxide:
of a Philips X’PERT diffractometer generating CuK� ra-
1
dCp
diation (� = 0.1542 nm) operating at 40 kV and 25 mA.
=
=
� kp Cp CO
[1]
[2]
[3]
[4]
[5]
�
S dt
Scanning was done between 4 and 55 2� with a step size of
�
0
.02 2� and 2.0 s counting time.
1 dCO
� kp Cp CO � kd CO
N2 BET surface areas were determined on a Micromerit-
S
dt
ics ASAP 2000 apparatus. Prior to N2 adsorption, the sam-
1
dCc
�
ples were calcined at 500 C and degassed at 88 K.
= k C C
c
p
O
S dt
dCh
Electron micrographs were obtained using a Leica LEO
S440 scanning electron microscope (SEM) with the accel-
1
=
=
kh Cp CO
�
erating voltage set at 40 kV. The stage tilt was set at 0 ,
S dt
dCa
S dt
the working distance and aperture size being 7 mm and
0 �m, respectively. Samples were mounted on aluminum
stubs and coated with Au/Pd film.
1
(kp � kc � kh)Cp CO .
3
The following symbols are used: p = phenol, c = catechol,
h = hydroquinone, o = oxidant (H O ), a = tar (aselec-
The infrared spectra were run on a Nicolet 5ZDX FT-IR
�
1
spectrometer at a resolution of 4 cm using self-supporting
wafers of the samples.
2
2
tive reactions), d = nonselective decomposition of oxidant,
S = total surface area, and t = time.
Diffuse reflectance UV–VIS spectra were recorded from
00 to 200 nm on a UV–Vis–NIR Varian Cary 5 spectropho-
tometer with a certified standard of Labsphere (SRS-99-
8
3
. RESULTS
0
10) as reference.
Atomic adsorption spectroscopy was done on a Varian
3
.1. Characterization of the Titanium-Containing
Molecular Sieves
Spectra AA-100, using solutions of 0.2 g of zeolite dissolved
in HF–HCl mixtures.
Figures 1A–1C show the scanning electron micrographs
of the obtained crystallites. The physical characteristics of
the synthesized materials are listed in Table 1. All sam-
ples were highly crystalline and showed a single UV–VIS
absorption band at 215 nm pointing to the absence of ex-
traframework TiO2 in the samples. The small TS-1 crys-
tallites show a cubic morphology with an average diam-
eter of about 0.1 �m. The large crystallites of TS-1 were
2
.4. Catalytic Hydroxylation of Phenol
Phenol hydroxylation was usually carried out at 60 C
�
with magnetic stirring (1000 rpm) in a 24-ml glass batch
reactor equipped with a Teflon sample port. After dissolu-
tion of 1.3 g of phenol (Aldrich) in 5 ml of the solvent, 0.13 g
of TS-1 was added. To start the reaction 4.6 mmol of H2O2,
i.e., 0.5 g of a 30 wt% aqueous solution (Merck), was added
in a single portion. Representative samples (0.2 ml) of the
vigorously stirred reaction mixture were withdrawn and an-
alyzed using HPLC. The HPLC was equipped with a C18
reverse-phase column (Luna) using 25 vol% of acetonitrile
in water as mobile phase and a Beckman 168 diode array
detector operating at 280 nm. The resulting H2O2 content
was monitored by means of a standard iodometric titration.
After reaction, the catalyst was recovered by filtration on
TABLE 1
Physical Characteristics of TS-1 and Al-Free Ti-Beta
TS-1 (small)
TS-1 (large)
Ti-Beta
a
Ti/Si, mol/mol
0.03
100
ca. 0.1
442
360
82
0.03
100
0.025
100
2–5
514
506
8
b
XRD crystallinity, %
dcrystal, �m
BET surface area, m /g
c
45 � 10 � 3
2
422
417
5
2
a 0.45-�m millipore filter, extensively washed with deion- Internal surface area, m /g
2
External surface area, m /g
ized water and dried at room temperature. The tar con-
tent of the zeolite was then analyzed using a Stanton Red-
croft STA 780 series thermal analyzer with a sample mass of
a
Determined by chemical analysis.
Compared to 100% crystalline standard.
Determined by SEM.
b
c
2
5 mg. For this analysis, the sample was heated in nitrogen

¨
WILKENHONER ET AL.
2
04
FIG. 1. Scanning electron micrographs of obtained titanium-containing molecular sieves. (A) Small crystals of TS-1; (B) large crystals of TS-1;
C) Al-free Ti-Beta.
(
coffin-shaped with dimensions of about 45 � 10 � 3 �m. path length. The results obtained by BET for all sam-
With the straight channels of the zeolite structure be- ples are in agreement with literature (10, 16). For the
ing in the direction of the small axis ([010] direction) small TS-1 crystals only ca. 82% of the total surface area
(
13), the shortest diffusional path length is of the or- is located in the pores of the crystal whereas the larger TS-1
der of about 1.5 �m. The crystallites of Al-free Ti-Beta and the Al-free Ti-Beta crystals have 98.5% of their Ti sites
were well-faceted truncated square bipyramids of 2–5 �m. inside the pores.
The diffusional path length in the large crystallites of
The cyclic CVD procedure of TEOS is used to inertize
TS-1 and Al-free Ti-Beta are thus similar, whereas the the external surface of TS-1. However, depending on the
small TS-1 crystallites have a much shorter diffusional silanization conditions, the application of this technique

PHENOL HYDROXYLATION IN DIFFERENT SOLVENTS
205
FIG. 1—Continued
can affect the zeolite pore mouth as well as result in pore diffusional path length. The comparison of the phenol hy-
mouth narrowing or blocking, as has been reported for droxylation on both catalysts therefore yields information
ZSM-5 (7). The absence of pore mouth narrowing in the on the effect of the pore structure on the phenol hydroxy-
silanized TS-1 samples used in this work was confirmed by lation (Table 2).
the chromatographic pulsed gas-phase sorption technique
The phenol conversion over large crystals of TS-1 is lower
(
12). No change in sorption properties (activation energy, than that over Al-free Ti-Beta. This can be attributed to the
first and second moment of pulse response) was observed smaller pores of TS-1, showing that mass transfer controls
for cyclohexane, 3-methyl-pentane and toluene. This con- the reaction rate. For the larger pores of Al-free Ti-Beta,
firms that the pore openings of the TS-1 samples were not intracrystalline diffusion is expected to be faster. With Al-
significantly affected by the silanization procedure.
free Ti-Beta catechol is the preferred product. This shows
SEM pictures of parent and silanized TS-1 samples also that the zeolite pore size controls the product formation.
showed no change, indicating that the agglomeration de- In the absence of a catalyst, only a low phenol conversion
gree of the small crystallites also remained unchanged.
No significant changes in IR and UV–VIS spectra were
observed. EDX of parent and silanized TS-1 showed an
increase of the average Si/Ti ratio from 33 to 60, confirming
the decrease of the Si/Ti ratio of the external rim of the
(0.85% ) to tars was observed.
TABLE 2
Influence of the Zeolite Crystal Structure on the Phenol
�
crystals and thus the deposition of silica on the external Hydroxylation in Water at 60 C and after 6 h under Standard
surface.
a
Conditions
H2O2
Xo, % efficiency, %
Hydroquinone/
catechol ratio
Tar,e
mg/gcat
Catalyst
Xp, %
3
.2. Phenol Hydroxylation
No catalyst
0.85
5.8
19.5
—
18.3
93.1
—
71.5
62.8
Only traces
1.3
—
n.d. f
118.0
The large difference in crystal size and thus in diffusional
b
TS-1
Ti-Beta
path length of the two TS-1 samples allows an investigation
of the influence of diffusional constraints on the reaction.
Furthermore, the small crystals have a much higher fraction
of active Ti sites located on the external surface, which can
yield information of the influence of the external surface
on the reaction selectivities obtained. The large TS-1 and
the Al-free Ti-Beta samples used in this work have a similar
c
d
0.5
a
See Materials and Methods.
TS-1 crystals of 3 � 10 � 45 �m.
Al-free Ti-Beta of ca. 2–5 �m.
Determined at Xp of 5.8% .
Amount of tar deposited after 24 h reaction.
Not determined (conversion not complete after 24 h).
b
c
d
e
f

¨
WILKENHONER ET AL.
2
06
TABLE 3
Influence of Crystal Size and External Surface Activity on the
�
Phenol Hydroxylation over TS-1 in Water at 60 C and after 6 h
under Standard Conditionsa
dcrystal,
H2O2
Xo, % efficiency, %
Hydroquinone/
catechol ratio
Tard
mg/gcat
�
m
Xp, %
3
0
0
0
� 10 � 45
.1
.1 (sil.)
5.8
28.9
27.2
23.8
18.3
100
100
100
71.5
86.6
83.1
71.5
1.30
n.d.e
84.2
53.0
91.5
1.31c
b
c
1.54
1.39c
.1 (coked)
a
See Materials and Methods.
b
c
d
e
Silanized small TS-1 crystals (20 CVD cycles).
Values at 5.8% phenol conversion.
Amount of tar deposited on TS-1 after 24 h.
Not determined.
FIG. 3. Concentration–time profile in the phenol hydroxylation over
TS-1 (dcrystal � 0.1 �m; mcatalyst = 0.13 g; mphenol = 1.3 g; 4.6 mmol of H2O2
(
30% in water, 0.5 g of solution); 5 ml of H2O as solvent).
The effect of the crystal size and the role of the external
surface for the phenol hydroxylation over TS-1 with water
as a solvent are shown in Table 3. Silanization of small TS-1 crystal size (Table 3). Silanization of the external surface re-
crystals leads to a decrease in the amount of coke formed, sults in a further increase of the hydroquinone-to-catechol
pointing to a decreased activity of the external surface. Sim- ratio. Catechol was, however, always formed in significant
ilar observations are noted for a coked TS-1 sample, which amounts. Even after a large number of silanization cycles
represents another way to achieve surface deactivation. A the ratio of hydroquinone to catechol over TS-1 with water
decrease in the crystal size of TS-1 leads to an enhancement as a solvent never exceeded 1.8. This indicates that cate-
in phenol and H2O2 conversion. This increase in the con- chol is not only formed on the sites located on the external
version cannot be attributed to the increase in the external surface, but also in the pores of TS-1.
surface, since under comparable conditions the conversion
In Fig. 3 the typical concentration–time profiles of reac-
with the silanized sample of TS-1 is still significantly larger tants and products, as well as the fit of the kinetic model
than that with the large crystals of TS-1. This suggests again applied (Eqs. [1]–[5]), are given. A good fit was obtained
that the phenol hydroxylation over large crystals is strongly for all experiments, showing that the overall second-order
diffusion limited. Van der Pol et al. (6) obtained a similar kinetic model describes the reaction well.
conclusion by comparing the activity of different crystal
sizes of TS-1 in the phenol hydroxylation.
The influence of the solvent on the phenol hydroxyla-
tion is summarized in Table 4. The highest catalytic activ-
With increasing phenol conversion over TS-1, the ity, expressed as k , is obtained with water as a solvent.
p
hydroquinone-to-catechol ratio increases (Figs. 2 and 3). This can be ascribed to the strong adsorption of phenol
This can be the result of coke formation and poisoning of in TS-1 in this solvent, which is driven by the nonideality
external surface sites (1). At similar phenol conversions of the water–phenol solution. At the same mole fraction,
the ratio of hydroquinone to catechol is independent of the the activity coefficient of phenol calculated using UNIFAC
software (23) is much higher in water than in the other sol-
vents (Table 5). Hence, high intraporous phenol and, as a
result of competitive adsorption effects, low hydrogen per-
oxide concentrations are expected with water as a solvent.
Surprisingly, no selectivity in the adsorption phenomena is
observed in methanol and acetone. With methanol or ace-
tone as a solvent the concentration of phenol in the pores
is about the same, while the affinity of the zeolite phase is
not much higher than that of the solution, as seen from the
Henry’s constants determined with the chromatographic
technique (Table 6).
On the other hand, the catalytic activity of TS-1 in
methanol is significantly higher than that obtained using
acetone as a solvent, whereas peroxide efficiency is better in
acetone as supported by the lower kd values (Table 4). This
FIG. 2. Ratio of the amount of hydroquinone formed relative to the
amount of catechol formed as a function of phenol conversion (phenol
conversion varied by different reaction times); catalyst TS-1 (dcrystal �
0
.1 �m; mcatalyst = 0.13 g; mphenol = 1.3 g; 4.6 mmol of H2O2 (30% in
water, 0.5 g of solution); 5 ml of H2O as solvent).
might be attributed to the side reaction between acetone

PHENOL HYDROXYLATION IN DIFFERENT SOLVENTS
207
TABLE 5
Phenol Partition (�) and Activity Coefficients (� )
in the Different Solvents
(TS-1)a,b
�(S-1)a,c
�
d
�
Water
Methanol
Acetone
33
� 1
� 1
19
—
—
9.0535
0.5943
0.3047
SCHEME 1. Formation of alkylidene peroxides from acetone and hy-
drogen peroxide (14).
a
See Materials and Methods: determined via both methods.
TS-1 of 0.1 �m.
b
c
d
Silicalite-1 of 0.2 �m.
Calculated with UNIFAC under reaction conditions (60 C) (23).
�
and H2O2 yielding hydroxy hydroperoxy propane and
derived products (14, 15) (Scheme 1), causing a lower con-
centration of free hydrogen peroxide at or near the active
sites in the pores and a gradual release of free H2O2.
The hydroquinone selectivity is highest in methanol [see
also (3)]. With acetone as a solvent the hydroquinone-to-
catechol ratio is lowest and similar for the original TS-1
and Ti-Beta samples. This indicates that the selectivity of
the reaction is governed much less by geometric constraints
or otherwise stated that the reaction in acetone takes place
to a larger extent on the external surface.
lated to side reactions between hydrogen peroxide and tar
components.
The cyclic deposition of TEOS allows us to follow the
progress of the inertization of the external surface. After
a number of cycles, the catalyst was removed from the
silanization reactor and used for a phenol hydroxylation
experiment in water. Figure 4 shows the second-order rate
constants for the phenol hydroxylation in water versus the
number of silanization cycles. Due to the silanization the
rate constant for the consumption of phenol is reduced by
Tar formation is following a similar trend as phenol con-
version (water > methanol > acetone). The tar deposition
on the catalyst surface decreases when the tar components
or precursors are more soluble in the solvent. On the other
hand, the decrease in tar deposition upon silanization isonly
significant in water and negligible in acetone. The residual
tar formation and hence deposition on silanized TS-1 cata-
lysts is somewhat unexpected, but can be attributed to tar
formation on titanium sites in pore mouths or even inside
the micropores. The tar poisoning of active sites can be
used as an alternative external surface inertization. Thus,
a TS-1 sample was used for reaction in water (where the
tar formation is most pronounced) and washed extensively
with deionized water. When this coked sample was used for
phenol hydroxylation in water (Table 3), no significant addi-
tional amount of tar is deposited (7.3 mg of tar/g of catalyst).
4
2% . This is mainly attributed to the reduction in the for-
mation of tars. The rate constant for the tar formation is
reduced by 91% . The rate constant for hydroquinone for-
mation decreases less than the rate constant of catechol
formation (25 vs 35% reduction). This indicates that at the
external surface the formation of catechol is favored over
that of hydroquinone, confirming literature reports (3, 5).
These results further point to the formation of at least some
catechol inside the pores of TS-1, as its rate constant is not
reduced to zero upon silanization.
4
. DISCUSSION
4
.1. Role of the External Surface—Kinetic Modeling
The present kinetic data enable us to determine the
However, a drastic decrease in phenol conversion (23.8% ) contribution of the external surface to the overall catalytic
was observed due to active site poisoning. The lower hydro- activity. The contribution of the external surface activity
gen peroxide efficiency (71.5% ) observed is presumably re- and the contribution of the sites in the pore system of the
TABLE 4
Influence of Solvent on the Observed Rate Constants for Phenol and Hydrogen Peroxide Consumption, Catechol, Hydroquinone, and
Tar Formation in the Phenol Hydroxylation, as Well as the Tar Deposition of a Parent and Silanized TS-1 Sample
2
2
2
2
a
a
2
b
c
Solvent
kp [l/(mol � m � s)]
kd [l/(m � s)]
kc [l/(mol � m � s)]
kh [l/(mol � m � s)]
k
[l/(mol � m � s)]
Tar [mg/g]
Tar [mg/g]
6
� 3
� 3
� 4
� 6
� 7
� 7
� 6
� 7
� 7
� 7
� 7
� 7
Water
Methanol
Acetone
5.0 � 10�
2.0 � 10
1.6 � 10
2.5 � 10
8.0 � 10
84.2
57.0
47.0
53.9
46.1
46.7
�
6
7
1.3 � 10
1.1 � 10
3.2 � 10
6.9 � 10
3.2 � 10
8.0 � 10�
5.5 � 10
3.2 � 10
2.6 � 10
2.2 � 10
�
Reaction at 60 C over TS-1 under standard conditions (see Materials and Methods).
a
ka = kp � kc � kh.
b,c
Tar deposition after 24 h reaction on parent and silanized TS-1 catalyst, respectively.

¨
WILKENHONER ET AL.
2
08
TABLE 6
Henry’s Constants (Ki) for the Adsorption of Phenol (p), Hydro-
quinone (h), and Catechol (c) from Different Solvents in Small TS-1
Crystals As Determined by the Liquid-Phase Chromatographic
Technique (11)
Water
Methanol
Acetone
Kp
Kh
Kc
84.4
5.7
9.2
0.7
0.6
0.7
0.6
0.6
0.7
titanium containing molecular sieves can be evaluated as
follows:
FIG 4. Overall rate constants for the consumption of phenol and the
formation of catechol, hydroquinone, and tar in the phenol hydroxylation
over TS-1 as a function of the number of cycles of deposition of TEOS
kn S = kn,iSi + kn,e Se
S = Si + Se,
[6]
(dcrystal � 0.1 �m; mcatalyst = 0.13 g; mphenol = 1.3 g; 4.6 mmol of H2O2
[7] (30% in water, 0.5 g of solution); 30 ml of H O as solvent).
2
in which n stands for p, c, h, o, a; i for internal, e for
external.
The observed rate constants for sites at the external sur-
For the small TS-1 crystals, the total surface area (S) face are all larger than those at the internal surface. Prin-
and the external surface area (Se) were determined to be cipally, diffusional constraints could also account for the
2
2
4
2
42 m /g and 82 m /g, respectively. Assuming that after difference between the observed rate constants for sites at
0 cycles the role of the external surface in the phenol hy- the external surface and those in the intracrystalline voids.
droxylation can be neglected; i.e., the observed activity is The crystals of TS-1 used in these experiments were ca.
�
18
2
solely due to the activity of TS-1 in the pores, the contribu- 0.1 �m. Based on the diffusion coefficient of 2 � 10
m /s
tion of the internal and external surfaces can be evaluated. for phenol in TS-1 reported by van der Pol et al. (6), the
The assumption of full inertization of the external surface effectiveness factor should be close to 1. Other diffusion
of TS-1 after 20 cycles seems justified due to the relatively measurements of phenol in various solvents with large TS-1
small changes in the rate constants (within the error of the crystals, prior to exposure to hydrogen peroxide, yielded
method) between 15 and 20 cycles of TEOS deposition.
diffusion coefficients that are 2 orders of magnitude larger
Table 4 shows the influence of the solvent on the overall (17). Thus, the assumption of an effectiveness factor close
rate constants. All observed rate constants reported are a to 1 for the small crystals seems to be justified.
product of the intrinsic rate constants and the sorption con-
On the basis of the rate constants determined for parent
stants. The rate constants for the activity of the sites inside and inertized TS-1, the contributions of internal and exter-
the pores of TS-1, obtained with a sample that has been nal surface in different solvents can be evaluated (Table 7).
exposed to 20 cycles of TEOS deposition, and the rate con-
stants for the activity on the external surface obtained by on the external surface. The contribution of internal sur-
applying Eq. [6] are given in Table 7.
face to the overall rate of phenol consumption amounts to
In water, the consumption of phenol mainly takes place
TABLE 7
Influence of Solvent on the Observed Rate Constants for Phenol and Hydrogen Peroxide Consumption and the Observed Rate Constants
for the Formation of Catechol, Hydroquinone, and Tars in the Phenol Hydroxylation (Both Based on the Concentration in the Liquid) at
�
a
6
0 C over TS-1 (d
= 0.1 �m; mcatalyst = 0.13 g; mphenol = 1.3 g; 4.6 mmol of H O (30% in Water, 0.5 g of Solution); 5 ml of Solvent)
crystal
2 2
2
2
2
2
kab [(l/(mol � m � s)]
2
Solvent
Water
kp [l/(mol � m � s)]
kd [l/(m � s]
kc [l/(mol � m � s)]
kh [l/(mol � m � s)]
6
� 3
� 7
� 6
� 7
� 7
� 7
� 6
� 6
� 6
� 7
� 6
� 7
� 7
� 8
� 6
� 8
� 6
� 7
� 7
kni
1.9 � 10� (32)
1.1 � 10 (44)
7.2 � 10 (35)
1.2 � 10 (39)
1.1 � 10 (1)
�
5
� 3
kne
kni
kne
kni
kne
1.8 � 10 (68)
6.0 � 10 (56)
5.7 � 10 (65)
8.2 � 10 (61)
4.3 � 10 (99)
�
7
� 4
Methanol
Acetone
8.8 � 10 (54)
6.3 � 10 (44)
2.1 � 10 (54)
6.1 � 10 (73)
5.6 � 10 (13)
�
6
� 3
3.3 � 10 (46)
3.4 � 10 (56)
7.7 � 10 (46)
1.0 � 10 (27)
1.5 � 10 (87)
�
7
� 4
4.1 � 10 (42)
2.5 � 10� (58)
1.8 � 10 (26)
1.3 � 10 (31)
1.4 � 10 (44)
1.5 � 10 (54)
6
� 3
2.2 � 10 (74)
1.2 � 10 (69)
7.9 � 10 (56)
5.5 � 10 (46)
a
Values in parentheses represent percent contribution of total formation/consumption (S�k).
ka = kp � kc � kh.
b

PHENOL HYDROXYLATION IN DIFFERENT SOLVENTS
209
TABLE 8
catechol formation varies between 46 and 69% . For the for-
mation of hydroquinone, the contribution of the external
surface sites to the total rate of formation of hydroquinone
is dependent on the solvent nature. With water as a sol-
vent, the contribution of the external surface sites to the
Rate Constant for the Formation of Hydroquinone (k ) Rela-
tive to the Rate Constant for the Formation of Catechol (kc) on
the Internal Surface Sites and the External Surface Sites in Dif-
h
ferent Solvents (Catalyst, TS-1 (dcrystal = 0.1 �m); m
= 0.13 g;
crystal
mphenol = 1.3 g; 4.6 mmol of H O (30% in Water, 0.5 g of Solution); total rate of formation of hydroquinone amounts to 60%
2
2
�
5
ml of Solvent; T
= 60 C)
(for the formation of catechol this is 66% ). This clearly in-
dicates that with water as solvent the reaction mainly takes
place on the external surface sites, which can be explained
by the strong adsorption of phenol and the consequently
low hydrogen peroxide concentration in the zeolite pores.
With methanol or acetone as a solvent, the contribution of
the external surface sites to the total rate of formation of
hydroquinone amounts to 27 or 56% , respectively. The cat-
echol selectivity is 45 and 68% in methanol and acetone,
respectively.
The kinetic data also contain information on the ratio
of the activity of the internal and external surface sites for
the formation of catechol and hydroquinone (Table 8). In
all solvents, on the internal surface titanium sites hydro-
quinone is preferentially formed. This can be attributed to
a geometric constriction for the reaction taking place in
the pore system of a titanium silicalite-1 catalyst. This is in
line with the observation that at 5.8% phenol conversion
over large crystals of TS-1 and Al-free Ti-Beta in water,
a hydroquinone-to-catechol ratio of 1.3 and 0.5, respec-
tively, was observed. On the external surface sites, a par-
ticular solvent effect is observed. In protic solvents (water
and methanol) hydroquinone formation is preferred over
catechol formation. In acetone, as an aprotic solvent, the
observed distribution is statistical.
reaction
Water
Methanol
Acetone
khi/kci
khe/kce
1.68
1.43
2.90
1.29
1.07
0.66
only 32% . The small contribution of the internal surface
to the rate of consumption of phenol can be explained by
the strong adsorption of phenol in TS-1 using water as a
solvent. Due to competitive adsorption, the hydrogen per-
oxide concentration near the active sites inside the pore
system of TS-1 (internal surface) will be very low. This will
result in a low observed rate for the consumption of phenol
in the pores.
With both methanol and acetone as a solvent the inter-
nal surface contributes to the overall rate of consumption
of phenol for 54 and 42% , respectively. The difference
between methanol and acetone can be explained by the for-
mation of acetone peroxides and hydroperoxides (14, 15)
(
Scheme 1), which are too bulky to react with titanium
sites in the pores. This is supported by the fact that TS-1
is very poorly active for oxidation reactions when tert-
butylhydroperoxide is used as an oxidant. The formation
of these compounds thus lowers the concentration of free
hydrogen peroxide in the liquid phase and consequently
in the pores of TS-1. Thus, in acetone the internal surface
contributes less to the overall reaction compared to the re-
action done in methanol.
On the basis of the rate constants obtained for hydro-
gen peroxide decomposition on silanized and nonsilanized
TS-1 (Table 7), the contribution of the external surface to
the rate of the nonselective hydrogen peroxide decomposi-
Decoupling internal and external surface activity clearly
demonstrates that the majority of the tar formation is taking
place at the external surface of the zeolite. Especially with
water as a solvent the external surface contributes 99% to
the total rate of formation of tars.
4
.2. Mechanistic implications
TS-1- and Al-free Ti-Beta-catalyzed phenol hydroxy-
tion, kd, for both methanol and water as a solvent amounts lation are possible only with hydrogen peroxide as the
to 56% . With acetone as a solvent this amounts to 74% , oxidant. Even though the organic peroxides such as tert-
as a consequence of the formation of the acetone perox- butylhydroperoxide may not penetrate the pores of TS-1
ides resulting in a lower concentration of hydrogen perox- there is ample space in the pores of Ti-Beta to activate
ide in the pores of TS-1. The rate constant for the non- the Ti sites, as is well known for epoxidation reactions
selective decomposition of hydrogen peroxide with water (10). However, no phenol hydroxylation occurs when tert-
and methanol as a solvent is 5.5 times larger on the exter- butylhydroperoxide is used as the oxidant. This strongly
nal surface. With acetone as a solvent this rate constant is indicates that the aromatic hydroxylation over titanium-
1
2 times larger than that for the internal surface. In all sol- substituted molecular sieves follows a different reaction
vents, the homogeneous decomposition of hydrogen per- mechanism than epoxidation reactions. Hence, it is pro-
oxide was found to be negligible for the duration of the posed that the terminal OH of the titanium hydroperoxo
experiment.
group is the electrophile that attacks the aromatic ring,
The contribution of the external surface to the rate of for- similar to the mechanism for amine hydroxylation, as sug-
mation of catechol is significant. Depending on the solvent, gested by Reddy and Jacobs (18). However, the existence
the contribution of the external surface to the total rate of of other pathways, e.g., through the formation of radicals,

¨
WILKENHONER ET AL.
2
10
tween methanol and the peroxo group at the titanium site
has often been proposed as the active intermediate com-
plex for TS-1 catalyzed reactions (2, 21). Scheme 2 shows
possible configurations of the titanium active site without
(
species c) and with coordination of 1 or 2 protic solvent
molecules (species a and b).
The coordination of protic solvent molecules results
in an increase of the size of the active titanium site.
SCHEME 2. Possible configurations of the hydroperoxo-titanium ac- This narrows the TS-1 channels, leading to a geometric
tive site of TS-1: (a) hexacoordinate octahedral, (b) pentacoordinate trig- constraint for an approaching phenol molecule. Hydro-
onal bipyramidal, and (c) tetracoordinate tetrahedral.
gen bonds of the phenolic OH with solvent OH groups
will make the phenol molecule more bulkier. Phenol,
cannot be excluded. The differences in pore size of the two hydrogen-bonded to the solvent OH groups, will ap-
tested molecular sieves reflect the difference in the product proach the bulky titanium site with the OH group pointing
selectivity of phenol hydroxylation with H2O2. As reported away from the titanium site (Scheme 3) yielding hydro-
earlier, in the narrow pores of TS-1 hydroquinone forma- quinone. Additionally, the presence of coordinated protic
tion is favored over that of catechol, whereas for Al-free Ti- molecules close to the peroxo group could lead to hydro-
Beta the relative amount ofcatecholisclose to the statistical gen bond formation with the active site (species a and
distribution.
b in Scheme 2), destabilizing H-bonding with the phe-
When the external surface of TS-1 was inertized with nol molecule assisting in o-hydroxylation (as shown in
SiO2, a relatively high amount of o-hydroxylation was Schemes 4A and 4B).
detected. The ratio of hydroquinone to catechol is al-
On the other hand, in nonprotic solvents, phenol can take
ways lower at the external surface than at the internal over the role of the protic solvent molecule. Although water
surface. This can be rationalized by the fact that at the is always present, its concentration in the hydrophobic TS-1
external surface the geometric constrictions are low and pores is expected to be low, as indicated by the extremely
o-hydroxylation becomes more likely. This does not mean high K value of phenol in TS-1 with water as a solvent. Thus,
that in the narrow TS-1 pores catechol formation is ex- the existence of titanium sites without protic molecules co-
cluded.
ordinated is proposed (Scheme 2, species c). In that case
To understand para or ortho hydroxylation, attention another reaction pathway is opened, i.e., the conversion via
must be focused on the titanium sites. The interaction of pentacoordinated (trigonal bipyramidal) Ti site, involving
protic solvents, such as alcohols and water, with Ti sites coordination of phenol to Ti peroxo species yielding cate-
is well known and confirmed by spectroscopic studies. In chol (Scheme 4A). This will be the case in aprotic solvents,
literature, it is generally agreed that protic molecules coor- such as acetone. Protic solvents will compete with phenol,
dinate to titanium, expanding its coordination sphere to 5 so that this pathway is not dominant in these solvents.
or 6 (19, 20). An increase in the average Ti–O bond length
Formation of catechol without coordination of phenol
upon adsorption of methanol and water has been observed to the active titanium site takes place via a six-membered
as well (20). A five-membered ring with hydrogen bonds be- transition state involving phenol (see Scheme 4B).
SCHEME 3. Proposed reaction mechanism for the formation of hydroquinone in phenol hydroxylation.

PHENOL HYDROXYLATION IN DIFFERENT SOLVENTS
211
SCHEME 4. Proposed reaction mechanism for the formation of catechol in phenol hydroxylation.
SCHEME 5. Proposed titanium sites on the external surface of TS-1.

¨
WILKENHONER ET AL.
2
12
Summarizing, hydroquinone is formed in a more se-
ACKNOWLEDGMENTS
lective way under limiting geometric conditions, as in
the pores of TS-1, and strongly influenced by H-bonding
effects of polar solvents. On the other hand in apolar
solvents or for less restricted geometric Ti sites, as in the
We thank K. Vanlaere for some of the experiments on the phenol
hydroxylation. The cooperation between the Katholieke Universiteit
Leuven, Vrije Universiteit Brussel, and University of Cape Town was
sponsored by the Flemish and South African governments in the frame
outer surfaces or in wider-pore molecular sieves, catechol of an international scientific and technological cooperation. The Flemish
partners acknowledge IUAP-PAI for sponsoring as well. Financial assis-
tance from NRF and THRIP (South Africa) is greatly acknowledged. We
thank I.W.T (F.v.L.) and IUAP-PAI (G.L.) for fellowships.
will be the preferred product.
The formation of tars takes place mainly outside the
TS-1 channels, since only a limited decrease in catalytic
activity is observed in a regeneration experiment. Due
to the hydrophobic nature of the TS-1 pores, a relatively
REFERENCES
low level of H2O2 is expected inside the pores, especially
in water. At the external surface, titanium is exposed to
much higher peroxide concentrations and the formation of
a second OOH group at the same titanium atom may take
place; the existence of titanium di(hydro)peroxo species
has been postulated (22). Alternatively, H2O2 molecules
could compete for coordination at the titanium peroxo site
1. Romano, U., Esposito, A., Maspero, F., Neri, C., and Clerici, M., Stud.
Surf. Sci. Catal. 55, 33 (1990).
2
. Martens, J. A., Buskens, Ph., Jacobs, P. A., van der Pol, A., van Hooff,
J. H. C., Ferrini, C., Kouwenhoven, H. W., Kooyman, P. J., and van
Bekkum, H., Appl. Catal. A 99, 71 (1993).
3. Tuel, A., Moussa-Khouzami, S., Ben Taarit, Y., and Naccache, C.,
J. Mol. Catal. 68, 45 (1991).
4. Tangaraj, A., Kumar, R., Mirajkar, S. P., and Ratnasamy, P., J. Catal.
30, 1 (1990).
1
(
Scheme 5). These rather unstable complexes could liberate
singlet molecular oxygen that is quenched to its triplet state
22) and hence decrease the oxidant selectivity and lead to
5
6
. Thangaraj, A., Kumar, R., and Ratnasamy, P., J. Catal. 131, 294 (1991).
. van der Pol, A. J. H. P., Verduyn, A. J., and van Hooff, J. H. C., Appl.
Catal. A 92, 113 (1992).
(
overoxidation and thus coke formation. These particular
external sites are poisoned with the tars as the reaction
progresses, explaining the shift toward the production of
more hydroquinone relative to catechol with time in water
and methanol. Since acetone is a good solvent for tars, the
poisoning of these sites is prevented and no shift toward
more hydroquinone is observed.
7. R o¨ ger, H. P., Kr a¨ mer, M., M o¨ ller, K. P., and O’Connor, C. T., Micro-
porous Mesoporous Mater. 21, 607 (1998).
8
. Wang, I., Ay, C.-L., Lee, B. L., and Chen, M.-H., “Proceedings, 9th
International Congress on Catalysis, Calgary, 1988” (M. J. Philips and
M. Ternan, Eds.), p. 324. Chem. Institute of Canada, Ottawa, 1988.
. Milestone, N. B., and Sahasrabudhe, N. S., “Proceedings of the 12th
International Zeolite Conference” (M. M. J. Treacy, B. K. Marcus,
M. E. Bisher, and J. B. Higgins, Eds.), p. 1901. MRS, Warrendale, 1999.
9
1
0. Blasco, T., Camblor, A., Corma, A., Esteve, P., Guil, J. M., Martinez,
A., Perdigon-Melon, J. A., and Valencia, S., J. Phys. Chem. B 102, 75
(1998).
CONCLUSION
1
1
1. Langhendries, G., De Vos, D. E., Baron, G. V., and Jacobs, P. A.,
J. Catal. 187, 453 (1999).
The conversion of phenol in the presence of hydrogen
peroxide is determined by the pore geometry, the exter-
nal surface titanium sites, the crystal size of the titanium-
containing molecular sieve, and the nature of the solvent.
Smaller pore dimensions (TS-1 vs Al-free Ti-beta) lead to
a decreased conversion and an enhanced selectivity for hy-
droquinone. The phenol hydroxylation is strongly diffusion
limited and small crystals should be applied. With small
crystals the external surface sites contribute significantly to
the total rate of reaction and its selectivity.
2. Denayer, J. F., Souverijns, W., Jacobs, P. A., Martens, J. A., and
Baron, G. V., J. Phys. Chem. B 102(23), 4588 (1998).
3. Ruthven, D. M., Eic, M., and Richard, E., Zeolites 11, 647 (1991).
4. Davies, A. G., “Organic Peroxides.” Butterworth, Stoneham, MA,
1961.
1
1
1
1
5. Sauer, M. C. V., and Edwards, J. O., J. Phys. Chem. 75(19), 3004 (1971).
6. Mirajkar, S. P., Thangaraj, A., and Shiralkar, V. P., J. Phys. Chem. 96,
3073 (1992).
1
1
7. Wilkenh o¨ ner, U., Duncan, W., and van Steen, E., in preparation.
8. Reddy, J. S., and Jacobs, P. A., Catal. Lett. 37, 213 (1996).
19. Bonneviot, L., Trong On, D., and Lopez, A., J. Chem. Soc., Chem.
Commun. 685 (1993).
2
Especially in water as a solvent, a significant part of phe-
nol conversion is due to external surface activity. While in
the pores of TS-1, hydroquinone is clearly the preferred
product, the major product formed on the external surface
0. Davis, R. J., Liu, Z., Tabora, J. E., and Wieland, W. S., Catal. Lett. 34,
01 (1995).
1
21. Bellussi, G., Carati, A., Clerici, M. G., Maddinelli, G., and Millini, R.,
J. Catal. 133, 220 (1992).
is solvent dependent: catechol is preferred in acetone, hy- 22. van Laar, F. M. P. R., De Vos, D., Vanoppen, D. L., Pierard, F.,
Brodkorb, A., Kirsch-De Mesmaeker, A., and Jacobs, P. A., “Proceed-
ings of the 12th International Zeolite Conference” (M. M. J. Treacy,
B. K. Marcus, M. E. Bisher, and J. B. Higgins, Eds.), p. 1213. MRS,
Warrendale, 1999.
droquinone in protic solvents. In protic solvents, the coor-
dination of solvent molecules to the titanium peroxo site
leads to a geometric constraint in the TS-1 pores, inducing
a “transition-state shape selectivity” with the para isomer
preferentially formed.
2
3. Sandler, S. I., “Chemical and Engineering Thermodynamics,” 2nd Ed.
Elsevier, Amsterdam, 1992.