Photocatalytic oxidation and decomposition of acetic acid on titanium silicalite

Article abstract of DOI:10.1021/es001400p

Transient reaction of adsorbed monolayers of acetic acid was used to characterize the photocatalytic properties of titanium silicalite zeolites (TS-1). The TS-1 zeolites having Si/Ti ratios of 5, 12.5, and 50 are effective catalysts at room temperature fo

Full text of DOI:10.1021/es001400p

Environ. Sci. Technol. 2001, 35, 1252-1258
structure, which is the same structure that ZSM-5 and silicalite
zeolites have.
Yamashita et al. (2) found that titanium oxides anchored
in the Y-zeolite cavity had high activity for photocatalytic
Photocatalytic Oxidation and
Decomposition of Acetic Acid on
Titanium Silicalite
decomposition (PCD) of NO into N
2 2 2
, O , and N O at 275 K
and high selectivity for N formation. The best catalysts were
2
prepared by ion exchange, and they attributed the high
activity to highly dispersed, tetrahedral titanium oxide
species. The specific photocatalytic activities, normalized
G U N D A E L E E , † V U A . T U A N , A N D
J O H N L . F A L C O N E R *
Department of Chemical Engineering, University of Colorado,
Boulder, Colorado 80309-0424
by the amount of TiO
for bulk TiO
Anpo et al. (3) used titanium oxides anchored within
Y-zeolite for photocatalytic reduction of CO with H O. The
2
in the catalyst, were much higher than
2
.
2
2
titanium oxide species that were highly dispersed within the
zeolite cavities existed in a tetrahedral coordination, and
Transient reaction of adsorbed monolayers of acetic acid
was used to characterize the photocatalytic properties
of titanium silicalite zeolites (TS-1). The TS-1 zeolites having
Si/Ti ratios of 5, 12.5, and 50 are effective catalysts at
room temperature for both photocatalytic oxidation (PCO)
and decomposition (PCD) of acetic acid. The rates of
PCO are higher than the rates of PCD for each catalyst.
Acetic acid oxidized photocatalytically in 0.2% O2 to form gas-
phase CO2 and CH4 and adsorbed H2O on the TS-1
catalysts, whereas no CH4 formed on Degussa P25 TiO2.
Isotope labeling showed that, on both TiO2 and TS-1 catalysts,
the R-carbon formed CO2 whereas the â-carbon formed
CH4 and CO2. The rates of oxidation of the two carbons have
different dependencies on UV intensity. The catalysts
with higher Si/Ti ratios adsorbed significantly more acetic
acid, and the PCO rates per gram of titanium are highest
on the TS-1 catalyst with the lowest Ti content, apparently
because a larger fraction of the Ti atoms are surface
atoms on this catalyst. During PCD in an inert atmosphere,
CO2, CH4, and C2H6 formed on TiO2 and on the catalyst
with a Si/Ti ratio of 5, but C2H6 was not detected on the
other catalysts. The CO2/CH4 selectivity during PCD increased
with increasing Si/Ti ratio. The first step in PCO and
PCD on TS-1 catalysts appears to be similar and involves
formation of a CH3 radical.
they had high selectivity for CH
3
OH formation. In contrast,
catalysts with aggregated, octahedrally coordinated titanium
oxide and powdered TiO catalysts had high selectivities for
2
CH formation.
4
Yamagata et al. (4) used TS-1 with a Si/ Ti ratio of 33.2 as
a photocatalyst to form CH by H reduction of CO . They
4
2
2
concluded that all Ti atoms in the TS-1 framework acted as
photocatalytic sites. From IR measurements, they determined
that -CH and -CH were left on the catalyst after reaction.
2 3
Yamashita et al. (5) used TS-1 (Si/ Ti )100) with some titanium
oxide dispersed within zeolite cavities as photocatalysts for
CO
observed that Ti-mesoporous molecular sieves exhibited
much higher photocatalytic activity for CH OH formation
2 2 4 3
reduction by H O to produce CH and CH OH. They
3
than powered TiO . The high activity was attributed to both
2
high dispersion of titanium oxide species and the large pore
size.
In the current study, PCO and PCD of gas-phase acetic
acid were compared on three TS-1 catalysts with different
Si/ Ti ratios. These catalysts are of interest because of their
activity in selective oxidation of organic compounds with
hydrogen peroxide under mild conditions (6) and because
they have been used for other photocatalytic reactions as
described above. The titanium in TS-1 is in the framework
or present as extra framework oxide, and TS-1 zeolites have
high surface areas for adsorption. Thus, these catalysts were
studied to determine their effectiveness for complete oxida-
tion of organic pollutants in air and to determine if they have
2
any advantages over TiO photocatalysts. Photocatalytic
Introduction
reaction of acetic acid is of interest because it is a common
product of biological digestion (7). Also, acetic acid is a
reaction intermediate during oxidation of ethanol and
acetaldehyde, and it exhibits similar rates to these molecules
Heterogeneous photocatalytic oxidation (PCO) of organic
pollutants is a promising process for air decontamination.
A wide range of organics in dilute systems can be oxidized
2 2
to CO and H O at ambient temperature and pressure on
2
during PCO on TiO . Photocatalytic oxidation rates were
semiconductor photocatalysts under UV illumination. The
rates of PCO are high for many organics so that PCO has the
potential to be an efficient method for removing low
concentrations of organics. Band gap excitation of semi-
conductors generates electron-hole pairs that can initiate
redox processes. An efficient semiconductor photocatalyst
must have suitable band-gap energy and stability over long
irradiation times, and TiO is the most active catalyst for
2
organic oxidation (1-3). Putting titanium within the zeolite
cavity or the zeolite framework, as in titanium silicalite (TS-
measured, and subsequent temperature-programmed oxi-
dation (TPO) helped determine the saturation coverages.
Transient photocatalytic reactions were carried out by
adsorbing a monolayer of acetic acid on the catalyst, flushing
excess organic from the gas phase, and then exposing the
catalyst to UV illumination in both the presence (PCO) and
absence (PCD) of gas-phase O
detected by a mass spectrometer. For comparison, PCO and
PCD experiments were carried out on powdered TiO
2
. The reaction products were
2
13
(
(
Degussa P25). In some experiments, C-labeled acetic acid
1
), may have advantages however because zeolites have
13
CH
3
COOH) was used to monitor the reactivity of the R-
nanoscale pores, high adsorption capacities, and ion-
exchange capacities (2, 3). Titanium silicalite has the MFI
and â-carbons separately.
Experimental Methods
*
Corresponding author e-mail: john.falconer@colorado.edu; fax:
Catalyst Preparation and Characterization. Three titanium
silicalite (TS-1) catalysts with Si/ Ti atomic ratios of 5, 12.5,
and 50 were prepared by in-situ crystallization. To obtain
(
303)492-4341.
†
Permanent address: Division of Chemical Engineering, Pukyong
National University, Pusan 608-739, Korea.
1
2 5 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
10.1021/es001400p CCC: $20.00

2001 Am erican Chem ical Society
Published on Web 02/10/2001

multiple mass peaks simultaneously and the output from a
thermocouple in contact with the catalyst. The mass spec-
trometer was calibrated by injecting known volumes of gases
into the flow downstream of the reactor, and signals were
corrected for cracking in the mass spectrometer. After PCO
or PCD, TPO was performed by heating the catalyst at 1 K/ s
TABLE 1. Molar Compositions of Gel for TS-1 Synthesis
Si/Ti
Ti
sample ratio wt % TPAOH 2-propanol TEOT TEOS H2O
TS-1-A
TS-1-B 12.5
TS-1-C 50
5
21
9.6
2.6
0.5
0.5
0.5
8.0
5.0
3.0
0.20
0.08
0.02
1.0
1.0
1.0
36.0
36.0
36.0
2
to 723 K in 20% O and holding it at 723 K until no desorption
products were detected. The TPO was used to determined
how much acetic acid, reaction products, and intermediates
remained on the surface after PCD or PCO.
high dispersion of Ti within the MFI framework, clear
synthesis solutions were used. The Si source was tetraethyl
orthosilicate (TEOS), and the Ti source was tetraethyl
orthotitanate (TEOT). Tetrapropylammonium hydroxide
Results and Discussion
Characterization of TS-1 Catalysts. Figure 1 shows the XRD
patterns of the three TS-1 catalysts. The XRD pattern for the
Si/ Ti ) 50 catalyst matches that reported for the pure MFI
structure with good crystallinity. The upper limit for inclusion
of titanium in the TS-1 framework is around 2.5 mol % (6).
Above that value, the excess is extra framework titanium,
which is usually present as anatase particles. Since catalysts
TS-1-A and TS-1-B have loading higher than 2% (Table 1),
they should have extra framework titanium, and thus the
XRD pattern for the TS-1A (Si/ Ti )5) catalyst is a combination
of TS-1 and anatase patterns. A number of studies reported
that, at Si/ Ti ratios of 5-12, a mixture of TS-1 and anatase
was obtained (17-20).
As shown in Figure 2, SEM analysis indicates that the size
and shape of crystallites changed with the Si/ Ti ratio. The
crystal size increased with increasing Ti content, probably
due to gel dilution. For the catalyst with Si/ Ti ) 50, most
crystals were approximately 0.7 µm in diameter, whereas for
the catalyst with Si/ Ti ) 5, the crystals were rectangular with
an approximate length of 5 µm and an aspect ratio of about
2.5. More 2-propanol was used in the synthesis as the Si/ Ti
ratio decreased (Table 1), and this lowered the gel viscosity,
which seemed to favor the development of zeolite crystals
along the a and c axes. Similarly, the crystal size of Ge-ZSM-5
zeolite has been reported to increase with decreasing Si/ Ge
ratio (14).
(
TPAOH) was used as the template. The required amount of
TEOT was diluted with 2-propanol while stirring, TEOS was
then added, and the solution was stirred for 30 min at room
temperature. A solution of water and TPAOH was then added
to the mixture, which was stirred for an additional 2 h at
room temperature to obtain a clear, homogeneous gel. The
gel was crystallized in a Teflon-lined autoclave at 448 K for
5
days. The zeolite was separated from the liquid phase by
centrifugation, dried at 373 K, and then calcined at 773 K for
h. Gel compositions are presented in Table 1. Under these
8
synthesis conditions, the Si/ Ti ratio in the TS-1 crystals is
expected to be the same as in the synthesis mixture. Degussa
P25 TiO
2
was used as a reference TiO
2
catalyst.
X-ray diffraction (XRD) analyses of the TS-1 samples were
preformed on a Siemens diffractometer using Cu KR radia-
tion. Scanning electron micrographs (SEM) were obtained
on a JEOL 8600 microscope.
Transient Reaction Measurem ents. The apparatus used
for PCD, PCO, and TPO was described previously (8).
Approximately 30-50 mg of catalyst was coated as a thin
layer on the inside surfaces of an annular Pyrex reactor so
that all the catalyst was exposed to UV light. The annular
reactor had a 1-mm annular spacing to minimize mass
transfer effects and rapidly flush gas-phase products from
the reactor. The outside diameter of the reactor was 2 cm,
and the reactor was 13 cm high so that sufficient catalyst
mass was present to allow detection of reaction products by
a mass spectrometer. Twelve 8-W UV lamps (Johnlite,
F8T5BlB) surrounded the reactor.
Photocatalytic Oxidation. In previous transient PCO
studies of a monolayer of acetic acid on P25 TiO
CO and formaldehyde formed immediately upon illumina-
tion of the catalyst, and the water product remained on the
TiO surface (9). The same behavior was seen in the current
2
, gas-phase
2
Transient reaction of acetic acid was carried out at room
temperature with mass spectrometric detection. Before each
experiment, the catalyst was held at 723 K for 30 min in 20%
2
experiments, but a higher light intensity was used so that
rates are higher and the details are different. The higher
intensity creates electron-hole pairs at a higher rate, and
different steps in the reaction can be affected differently by
O
2
in He and then cooled to room temperature to create a
reproducible surface. Depending on the catalyst, two to four
1
3
1
-µL pulses of acetic acid (Sigma, 99%) or C-labeled acetic
the higher concentrations of electrons and holes. As shown
1
3
13
acid (CH
3
COOH, Isotec, 99+% atom enrichment) were used
in Figure 3, PCO (0.2% O
2
) of a monolayer of CH
, and CH
not shown because water adsorbs strongly on TiO
3
COOH on
12
to saturate the catalyst in the dark at 300 K prior to PCD or
PCO. All experiments started with the surface saturated. After
acetic acid was injected, the reactor was flushed for 2 h to
remove it from the gas phase so that only reaction of the
adsorbed monolayer was studied. Photocatalytic decompo-
P25 formed 13CO
2
,
CO
2
2
O. The water product is
, whereas
2
CO
2
is weakly adsorbed.
The rate of 13CO
formation in Figure 3 rapidly reached
2
a maximum after the P25 catalyst was illuminated, but the
3
12
sition was carried out at room temperature in 100 cm / min
2
rate of CO formation increased slowly to a maximum after
(
STP) of He flow by turning the UV lights on and observing
460 s of illumination. As previously reported (9), this indicates
that the R-carbon (labeled with 13C) in acetic acid prefer-
the products that formed with a Balzers QMG 421C quad-
rupole mass spectrometer. Reaction was stopped and started
again by turning the lights off and on. A small m/ e ) 32
signal was detected by the mass spectrometer, but that signal
did not change when the lights were turned on with acetic
acid adsorbed on the surface. Thus, this signal was due to
background gas in the vacuum chamber, and O
stream was below the detection limit, which we estimate to
be 0.3 ppm (9). Photocatalytic oxidation was carried out in
0
13
entially oxidizes to CO
2
. The rates of 12CO
2
and CO
2
formation were only comparable at long times. The time
delay before 12CO
formation reached a maximum indicates
that the â-carbon oxidized through an intermediate species,
2
1
2
such as formate or formaldehyde. Indeed, CH
2
O but no
COOH on P25 TiO . The
mass spectrometer was not calibrated for CH O. The amount
O was estimated from a carbon mass balance between
1
3
13
2
in the feed
CH
2
O formed during PCO of CH
3
2
2
of CH
2
1
2
13
.2% O
2
flow at room temperature for most experiments,
C and C. During PCO for 3600 s, the rate of total CO
2
1
2
13
although a few used 20% O
2
.
( CO
2
+
CO
2
) formation decreased to 0.045 µmol (g of
s . This final rate was 7% of the initial rate, and
-
1
-1
The mass spectrometer monitored the reactor effluent
immediately downstream of the reactor. A computer inter-
faced to the mass spectrometer recorded the amplitudes of
catalyst)
the acetic acid coverage was approximately 8% of a mono-
layer. The amount of acetic acid adsorbed on P25 was 417
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1 2 5 3

FIGURE 1. X-ray diffraction patterns of TS-1 powders with different Si/Ti ratios.
µmol/ g of TiO . This was estimated from the amount of formation indicate a competition between hydro-
in 12CO
2
2
carbon dioxide formed during PCO and during a subsequent
genation and oxidation of adsorbed CH
3 2
, with CO forming
TPO.
through an intermediate.
A comparison of the PCO results in Figure 3, which was
The 12C/ 13C ratio for the gas-phase products in Figure 5
is less than 1, indicating that some C-enriched products or
2
for a UV intensity of 2.5 mW/ cm , to transient PCO on P25
12
2
at a lower UV intensity (0.3 mW/ cm ) but also in 0.2% O
2
(9)
2
intermediates were left on TS-1-B catalyst after PCO. This
shows several differences. The maximum rate of 13CO
12
13
was confirmed by a subsequent TPO; the C/ C ratio during
formation was a factor of 2.5 higher for the higher light
12
13
TPO was greater than 1, and thus the ratio of C/ C in the
total products during PCO and TPO was close to 1.
1
2
2
intensity, whereas the maximum rate of CO formation was
1
2
13
4
2 2
.6 times higher. Thus, CO formed faster than CO formed
The rate dependence on time during transient PCO of
acetic acid was different for the TS-1-C catalyst. As shown
in Figure 6, the rates were much smaller than for the other
catalysts. Note, however, that these rates are per gram of
catalyst, and TS-1-C contains much less titanium than the
for a short time period in Figure 3, but this was not observed
at the lower light intensity. The oxidation rates of the two
types of carbons have different dependencies on light
intensity, and this may be because the oxidation rates of
intermediates that form from the â-carbon have different
1
2
other catalysts. The CO
increased upon UV illumination, but the rates and the CO
CH selectivity were almost constant during 3600 s of PCO.
2 4
and CH formation rates rapidly
2
dependencies. Moreover, the CO rate reached a maximum
2
/
after 460 s in Figure 3, but at the lower light intensity, the
maximum was not reached until 1050 s. This also indicates
that both acetic acid and its intermediates oxidized faster at
the higher UV intensity.
4
The constant rate may arise because most of the acetic acid
was not adsorbed on active sites (i.e., those that contain
titanium). As acetic acid on the titanium sites reacted, more
diffused from unreactive sites to keep the active sites saturated
and thus the rate approximately constant.
In contrast to its behavior on P25 TiO
2
, acetic acid oxidized
photocatalytically on TS-1 catalysts to form gas-phase CO
2
and CH
4
, and CO
2
formed significantly faster than CH
4
. As
shown in Figure 4, on TS-1-A the rates of CO
2
and CH
4
The P25 had a higher initial rate per gram of catalyst than
the TS-1 catalysts, whose rates decreased with increasing
Si/ Ti ratio. However, the initial rates per gram of TiO (Table
2
2) for the TS-1 catalysts were 2-3 times higher than that for
P25. This difference may be due mostly to the higher
dispersion of titanium in the TS-1 catalysts; i.e., more of the
titanium is on the surface of the TS-1 catalysts than for P25.
Since some titanium is not in the framework for the TS-1
formation reached maxima immediately after UV illumina-
tion, decreased relatively rapidly, and then decreased more
slowly. The same behavior was observed for TS-1-B, but the
1
3
rates were lower. For PCO of CH
3
COOH on TS-1-B (Figure
1
3
5
), the R-carbon oxidized exclusively to CO , whereas the
2
1
2
2 4 2 4
and CH and CH
. The 13CO
12
12
â-carbon formed both CO
formation rates reached maxima almost as soon as the
1
2
catalyst was exposed to UV. In contrast, the CO
2
formation
catalysts with lower Si/ Ti ratios, the TiO
be expected to increase as the Si/ Ti ratio increased. The initial
rate, per gram of TiO , increased with Si/ Ti ratio, as shown
in Table 2. The acetic acid adsorbs on both titanium sites
2
dispersion might
rate did not reach a maximum until approximately 2000 s,
and the maximum was broad. The immediate maximum in
2
1
2
4
the rate of CH formation and the delay before the maximum
1
2 5 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

FIGURE 4. Product formation rates during transient photocatalytic
oxidation of a monolayer of CH COOH on TS-1-A (Si/Ti ) 5) in 0.2%
in He flow.
3
O
2
FIGURE 2. Scanning electron micrographs of TS-1 powders with
different Si/Ti ratios.
FIGURE 5. Product formation rates during transient photocatalytic
13
oxidation of a monolayer of CH
.2% O in He flow.
3
COOH on TS-1-B (Si/Ti ) 12.5) in
0
2
FIGURE 3. Product formation rates during transient photocatalytic
13
3 2 2
oxidation of a monolayer of CH COOH on TiO in 0.2% O in He
flow. The UV lights were turned on (open triangles) and off (solid
triangles) as indicated.
and on other sites in the zeolite since it adsorbs on silicalite
(
15).
As shown in Figure 7, when 20% O
FIGURE 6. Product formation rates during transient photocatalytic
oxidation of a monolayer of CH COOH on TS-1-C (Si/Ti ) 50) in 0.2%
in He flow.
2
was used instead of
2
rates were higher on TS-1-B
1
3
12
3
0
.2% O
2
, the CO
2
and CO
1
2
O
2
and no CH
concentration increased the CO
4
was detected. A 100-fold increase in the O
2
1
3
2
rate by only 25-50% and
1
2
rate by 100%. The 12CO
groups form on the surface and then react by parallel
pathways to CH and CO . Methane is not being formed and
then oxidized at the higher O concentration; it has a low
reactivity during PCO (16). Higher O concentrations shifts
the selectivity toward CO formation.
the CO
because no CH
2
2
rate increased more partly
1
2
4
formed at the higher O
2
concentration.
rates were still not equal; the
R-carbon oxidizes faster than the â-carbon. The absence of
CH formation indicates that during PCO on TS-1, methyl
4
2
1
2
13
Note that the CO
2
and CO
2
2
2
4
2
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1 2 5 5

TABLE 2. Initial Rates of CO
Acetic Acid on TiO and TS-1
2
and CH
4
Formation during PCO of
2
initial rates
µmol (g of TiO2) s-1
)
-
1
(
catalyst
CO2
CH4
total
TiO2
TS-1-A (Si/Ti ) 5)
TS-1-B (Si/Ti ) 12.5)
TS-1-C (Si/Ti ) 50)
0.64
1.05
1.35
1.54
0
0.64
1.24
1.56
1.92
0.19
0.21
0.38
FIGURE 8. Product formation rates during transient photocatalytic
decomposition of a monolayer of CH COOH on TiO in He flow.
3
2
then 10 times as much CO
2
formed on TS-1-C as on P25 in
3
600 s. The rate per gram of titanium is so much higher on
TS-1-C because more of its titanium is on the surface than
for P25.
Photocatalytic Decom position. Since the TS-1 catalysts
are less active per gram of catalyst, they may not have any
2
advantages over P25 TiO in applications to PCO of organic
contaminants. However, the TS-1 catalysts have much higher
adsorption capacities for organics, and under conditions
where the organic concentrations vary widely, the TS-1
catalysts might have some advantages.
The rates of product formation during PCD (no O in the
2
gas phase) were lower than during PCO, and additional
products were seen. As shown in Figure 8, acetic acid
FIGURE 7. Product formation rates during transient photocatalytic
13
oxidation of a monolayer of CH
0% O in He flow.
3
COOH on TS-1-B (Si/Ti ) 12.5) in
2
2
decomposed photocatalytically on P25 to form CO
and C . Upon UV illumination, the rates of CO and C
formation reached maxima and decreased quickly. The C
formation rate decreased to almost 0 after 250 s of PCD (the
baseline apparently drifted slightly upward in Figure 8), but
2 4
, CH ,
TABLE 3. Amounts Formed during 3600-s PCO and Subsequent
TPO of Acetic Acid on TiO and TS-1
2
H
6
2
2
H
H
6
2
2
6
amt formed
during PCO
CO
2
amt
total acetic
(
µmol/g of
during TPO acid adsorbed
catalyst)
2 4
the CO and CH formation rates were similar after 250 s. All
(
µmol/g of
(µmol/g of
catalyst)
rates dropped to 0 when the UV lights were turned off. When
they were turned back on after 1920 s, the rates of formation
of products were higher than before the lights were turned
catalyst
CO
2
4
CH HCHO catalyst)
TiO
TS-1-A (Si/Ti ) 5)
2
723
414 56
42a
69
450
1177
1376
417
460
797
775
off. The CO and CH
2
4
rates increased by factors of 1.7-1.8,
TS-1-B (Si/Ti )12.5) 369 48
but the C H rate increased a factor of 10.
2
6
TS-1-C (Si/Ti ) 50)
144 29
2
This behavior, for a UV intensity of 2.5 mW/ cm , is similar
to that reported on P25 for a UV intensity of 0.3 mW/ cm (9).
a
Estim ated from carbon m ass balance between 12C and 13C.
2
2 6 4
However, the initial ratio of C H to CH rates was higher at
Table 3 shows the amounts of gas-phase products detected
during 3600 s of PCO and during a subsequent TPO for the
four catalysts. Because of calibration difficulties, the amount
the higher UV intensity. Moreover, the CH rate was higher
4
after a dark time than before the lights were turned off, as
shown in Figure 8. No increase in CH rate was seen during
4
1
2
13
of CH
2
O formed was estimated so that the C/ C ratio was
a dark time following PCD at the lower UV intensity. As
1
in the products of PCO plus TPO. The total acetic acid
reported previously (9), acetic acid decomposed on TiO by
2
adsorbed (417 µmol/ g of catalyst for P25) agrees well with
the value of 415 µmol/ g of catalyst measured previously using
TPD and subsequent TPO (17).
parallel pathways to form CO , CH , C H , and H O during
PCD:
2
4
2
6
2
As shown in Table 3, the TS-1 catalysts have higher
adsorption capacities than P25. This might be expected since
CH3COOH f CO2 + CH4
2CH COOH + O(l) f C H + 2CO + H O
(1)
(2)
2
the TS-1 zeolites have higher surface areas than P25 (50 m /
3
2
6
2
2
g). The two TS-1 catalysts with higher Si/ Ti ratios have higher
adsorption capacities than TS-1-A. The presence of extra
framework titania in TS-1-A may block access to adsorption
Reaction 1 does not require oxygen, whereas lattice oxygen
is extracted from the TiO surface in reaction 2. As the TiO
surface is reduced, the rate of C formation rapidly
decreases because the remaining lattice oxygen is not as easy
to remove. Methane and CO still form on the reduced surface,
2
2
sites (18). The amounts of CO
2
that formed during 3600 s of
2 6
H
PCO decreased as the Ti content decreased, as shown in
Table 3. The fraction of the adsorbed acetic acid that reacted
2
during PCO decreased dramatically. For P25 TiO
2
, 92% of the
although their rates appear to be affected by the reduced
surface since they increase after a dark time. During the dark
acetic acid oxidized in 3600 s, whereas on TS-1-C only 9%
oxidized. Much of the acetic acid was not adsorbed on titania
sites on TS-1-C, and thus the fraction that reacted was lower.
When normalized by the amount of titanium in the catalyst,
time, lattice oxygen diffuses from the TiO
the surface oxygen vacancies (16), and thus the C
formation rate increased dramatically after a dark time.
2
bulk to replenish
2
H
6
1
2 5 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

FIGURE 9. Product formation rates during transient photocatalytic
decomposition of a monolayer of CH COOH on TS-1-A (Si/Ti ) 5)
in He flow.
FIGURE 11. Product formation rates during transient photocatalytic
decomposition of a monolayer of CH COOH on TS-1-C (Si/Ti ) 50)
in He flow.
3
3
than just before the lights were turned off, although the
increase in the CH rate was slight.
The PCD behavior on TS-1-B was similar to that on TS-
4
1
3
1
-A. On TS-1-B, CH
3
COOH was used so the pathways of the
two carbons could be followed. Acetic acid decomposed
photocatalytically to 13CO
This is similar to the behavior during PCO; R-carbons were
exclusively oxidized to CO , and â-carbons only produced
CH . The rates of product formation on TS-1-B were less
than half those on TS-1-A, and neither CO
12
2
4
and CH , as shown in Figure 10.
2
4
1
2
13
2
4
nor CH were
detected. A lower rate is expected since TS-1-B has only half
as much titanium as TS-1-A. Ethane, if it formed, had too
low a rate to be detected. The rapid increase and decrease
in the 13CO
much higher than the CH
on TS-1-A) indicate that C
rate and the fact that the CO
rate (the same behavior observed
probably formed on TS-1-B.
ratio was less than 1.
2
2
rate was initially
4
2
H
6
Note also that the 12CH
/ CO
4 2
13
FIGURE 10. Product formation rates during transient photocatalytic
As on TS-1-A, the rates decreased quickly on TS-1-B when
the lights were turned off, but they did not drop to 0. The
slow decrease in rates over 1000 s is unlikely to be due to
catalytic reaction at room temperature in the absence of UV
decomposition of a monolayer of CH
3
COOH on TS-1-B (Si/Ti ) 12.5)
in He flow.
Indeed, when O
was much higher when the lights were turned back on (19).
2
was injected in the dark, the rate of PCD
illumination. Instead, the detection of CO
dark is probably because these species adsorb more strongly
on TS-1 than on TiO and also because they are limited by
diffusion out of the small TS-1 pores.
On TS-1-C, the rates of CO and CH
than on TS-1-B since TS-1-C contains less titanium. No C
was detected, although as on catalyst B, C may have
formed in low concentrations and not been detected. More
CO than CH formed, and as shown in Figure 11, the CO
2 4
and CH in the
The TS-1-A catalyst exhibited similar PCD behavior to
2
P25. The main gas-phase products were CO
as shown in Figure 9. The rates of formation of all products
rapidly reached maxima upon UV illumination. The CO rate
decreased rapidly initially, but then its rate was almost
constant. In addition, when the UV lights were turned off,
2 4 2 6
, CH , and C H ,
2
4
formation were lower
2
2 6
H
2 6
H
the CO
2
and CH
4
rates rapidly decreased at first, but then
2
4
2
slowly dropped to 0. When the UV lights were turned on
after 1920 s, the formation rates of all products were higher
rate had two distinct maxima with time. When the lights
were turned off, the rates dropped more slowly than for the
2 2
TABLE 4. Products Distribution of PCO in 0.2% O and PCD of Acetic Acid on TiO and TS-1
product distribution (%)
CH4
catalyst
CO2
C2H6
HCHO
TiO2
PCD
PCO
PCD
PCO
PCD
PCO
PCO
PCD
PCO
52 (0.20)a
95 (0.64)
57 (0.09)
88 (0.22)
60 (0.03)
88 (0.13)
100 (0.20)
69 (0.01)
83 (0.04)
40 (0.11)
8 (0.08)
5
TS-1-A (Si/Ti ) 5)
29 (0.05)
12 (0.04)
40 (0.02)
12 (0.02)
14 (0.02)
TS-1-B (Si/Ti ) 12.5)
b
TS-1-C (Si/Ti)50)
31 (<0.005)
17 (<0.01)
a
Num ber in parentheses is initial form ation rate in µm ol (g of catalyst)-1 s-1. b PCO in 20% O
.
2
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1 2 5 7

other catalysts, but the rates were also lower, and the slow
drop probably has the same origin as for TS-1-B.
partial support by donors of the Petroleum Research Fund,
administered by the American Chemical Society. G.D.L. also
thanks Pukyong National University for support.
2
During PCD on TiO and TS-1-A, when UV illumination
resumed after the dark period, the product formation rates
were higher than the rates measured just before the lights
were turned off. In contrast, no increase in product formation
after the dark time was observed on TS-1-B and TS-1-C, and
Literature Cited
(1) Anpo, M.; Chiba, K. J. Mol. Catal. 1992, 74, 207.
no C
-A, which has the highest titanium content of the TS-1
catalysts used, has more extra framework TiO and that
removing surface lattice oxygen (reaction 2) from TiO should
be easier than removing oxygen from the TS-1 framework.
The oxygen vacancies created in the TiO surface are
replenished by oxygen diffusion from the bulk. The reaction
rate thus increases after a dark time because the surface
oxygen concentration is higher (9).
2
H
6
was detected. These results may indicate that TS-
(2) Yamashita, H.; Ichihashi, Y.; Anpo, M.; Hashimoto, M.; Louis,
C.; Che, M. J. Phys. Chem. 1996, 100, 16042.
1
(3) Anpo, M.; Yamashita, H.; Ichihashi, Y.; Fuji, Y.; Honda, M. J.
Phys. Chem. B 1997, 101, 2632.
2
2
(
4) Yamagata, S.; Nishijo, M.; Murao, N.; Obta, S.; Mizogushi, I.
Zeolite 1995, 15, 490.
2
(
5) Yamashita, H.; Fuji, Y.; Ichihashi, Y.; Zhang, S. G.; Ikeue, K.;
Park, D. R.; Koyano, K.; Tatsumi,T.; Anpo, M. Catal. Today 1998,
4
5, 221.
6) Millini, R.; Previde Massara, E.; Perego, G.; Bellussi, G. J. Catal.
992, 137, 497.
(
Com parison of PCO and PCD on TS-1 and TiO
lists the product distributions and their initial rates of
formation during PCO and PCD. The first step in PCO and
PCD may be similar on TS-1 and involve formation of CO
and CH radicals since CO and CH form in both reactions
and the selectivity for CO increased with O in the gas phase.
Aqueous acetic acid was reported to form CH in both PCO
20) and PCD (21). In the absence of gas-phase O , CH
radicals mainly combine with H to form CH , but in the
presence of O , methyl radicals also oxidize to CO . Isotope
labeling showed that R-carbon does not require O to form
CO whereas â-carbon does. Gas-phase O suppresses CH
and C
oxidized (20).
Adding O
2
. Table
1
4
(
(
7) Fox, M. A.; Dulay, M. T. Chem. Rev. 1993, 93, 341.
8) Larson, S. A.; Widegren, J. A.; Falconer, J. L. J. Catal. 1995, 157,
2
6
11.
3
2
4
(
9) Muggli, D. S.; Keyser, S. A.; Falconer, J. L. Catal. Lett. 1998, 55,
2
2
129.
3
(10) Millini, R.; Perego, G.; Bellussi, G. J. Catal. 1990, 137, 503.
(
2
3
(
11) Fricke, R.; Kosslick, H.; Tuan, V. A.; Pilz, W. Stud. Surf. Sci. Catal.
1994, 83, 57.
4
2
2
(12) Pilz, W.; Penker, C.; Tuan, V. A.; Kosslick, H. Ber. Bunsen Phys.
Chem. 1993, 97, 1037.
2
2
2
4
2
(13) Grohmann, C.; Pilz, W.; Kosslick, H.; Tuan, V. A. Surf. Interface
Anal. 1994, 22, 403.
2
H
6
formation because the radical intermediate is
(14) Kosslick, H.; Tuan, V. A.; Fricke, R.; Peuker, C. H.; Pilz, W.; Strorek,
W. J. Phys. Chem. 1993, 97, 5678.
2
to the gas phase increased the rate of CO
(
15) Sano, T.; Shigeyuki, E.; Yamada, K.; Kawakami, Y.; Yanagishita,
formation from both â- and R-carbons. During PCD of
1
3
13
H. J. Membr. Sci. 1997, 123, 225.
CH
3
COOH on TS-1-B (Figure 10), the initial rate of CO
2
-
1
-1
12
(16) Muggli, D. S.; Falconer, J. L. J. Catal. 1999, 187, 230.
formation was 0.028 µmol (g of catalyst)
formed. In contrast, both CO
PCO (Figure 5), and the initial rate of CO
.12 µmol (g of catalyst)
s
, and no CO
2
1
3
12
(17) Muggli, D. S.; McCue, J. T.; Falconer, J. L. J. Catal. 1998, 173,
2
and CO
2
formed during
2
formation was
4
70.
1
3
-
1
-1, whereas the initial CH
(18) Vayssilov, G. N. Catal. Rev. Sci. Eng. 1997, 39, 209.
0
s
4
(
(
(
19) Muggli, D. S.; Falconer, J. L. J. Catal. 2000, 191, 318.
20) Kraeutler, B.; Bard, A. J. J. Am. Chem. Soc. 1978, 100, 5985.
21) Nosaka, Y.; Koenuma, K.; Ushida, K.; Kira, A. Langmuir 1996,
formation rate was essentially the same. Table 4 also shows
that the rate and selectivity for CO formation, especially
2
during PCD, depend on the catalyst, although the percentages
were similar.
1
2, 736.
Received for review June 20, 2000. Revised manuscript re-
ceived November 20, 2000. Accepted December 8, 2000.
Acknowledgments
We gratefully acknowledge support by the National Science
Foundation (Grant CTS 9714403). We also acknowledge
ES001400P
1
2 5 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001