Effect of water on formic acid photocatalytic decomposition on TiO 2 and Pt/TiO2

Article abstract of DOI:10.1016/j.jcat.2010.08.011

The effect of water on the adsorption and photocatalytic decomposition (PCD) of formic acid on TiO₂ and Pt/TiO₂ was investigated using transient reaction studies, temperature-programmed desorption (TPD), and Fourier transform infrare

Full text of DOI:10.1016/j.jcat.2010.08.011

Journal of Catalysis 275 (2010) 294–299
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Effect of water on formic acid photocatalytic decomposition on TiO and Pt/TiO
2
2
a
b
a
a,
⇑
Kristi L. Miller , Chul Woo Lee , John L. Falconer , J. Will Medlin
a
Department of Chemical and Biological Engineering University of Colorado, Boulder, CO 80309-0424, United States
Department of Chemical Engineering, Hanbat National University, Daejeon 305-719, South Korea
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of water on the adsorption and photocatalytic decomposition (PCD) of formic acid on TiO2 and
Pt/TiO was investigated using transient reaction studies, temperature-programmed desorption (TPD),
and Fourier transform infrared (FTIR) spectroscopy. Reaction studies indicated that physisorbed water
increased the PCD rate of formic acid to a small extent on TiO and to a major extent on Pt/TiO , but
the presence of only chemisorbed water did not. FTIR spectroscopy and TPD studies indicate that the
main effect of the addition of water to TiO that had adsorbed formic acid was the displacement of formic
Received 19 May 2010
Revised 11 August 2010
Accepted 20 August 2010
Available online 28 September 2010
2
2
2
2
Keywords:
Photocatalysis
Formic acid
acid. However, FTIR spectroscopy indicated that the addition of water caused a change in the adsorbed
structure of formate that may be associated with the higher reactivity. These transformations can have
an important influence on elementary steps in the photocatalytic decomposition of formic acid on TiO
and Pt/TiO
2
TiO
2
Water
2
.
Fourier transform infrared spectroscopy
Ó 2010 Elsevier Inc. All rights reserved.
(
FTIR)
1
. Introduction
too high, competitive adsorption decreases the photocatalytic rate
[
6,8,13].
In the current study, the accelerating effect of water was inves-
tigated in transient photocatalytic decomposition experiments
with formic acid. Both TiO and Pt/TiO catalysts were employed
Titania photocatalytically oxidizes low concentrations of organ-
ics at room temperature, and the anatase form of TiO has a higher
2
activity than the rutile form [1]. Photocatalytic reactions clean
waste air streams and indoor air, and are used for self-cleaning
windows. Ultraviolet light initiates reaction by exciting electrons
from the valence to the conduction band of a semiconductor to cre-
2
2
as catalysts. Weakly adsorbed water was found to have a signifi-
cant effect on PCD rate, but strongly chemisorbed water did not.
One possible mechanism for this rate acceleration is that co-
adsorbed water changes the configuration of formic acid adsorbed
ate an electron–hole pair. On TiO
sometimes completely to CO and H
photocatalytic oxidation, PCO), or it decomposes formic acid to
CO and H O by extracting a lattice oxygen in the absence of oxy-
gen (photocatalytic decomposition, PCD). On Pt/TiO , formic acid
decomposes to CO and H during PCD.
2
, the hole oxidizes formic acid,
2
2
O in the presence of oxygen
2
on TiO . For example, Vittadini et al. used DFT calculations to show
(
that water induces the dissociation of formic acid to formate on
anatase (1 0 1) [14]. To probe this possibility, we used Fourier
transform infrared spectroscopy (FTIR) and temperature-pro-
grammed desorption (TPD). The FTIR and TPD measurements were
2
2
2
2
2
Water, which is usually present during photocatalytic reactions,
promotes or inhibits reaction rates, depending on the reactant and
water concentrations, and thus the influence of water on species
applied to high surface area P25 TiO
. Experimental methods
.1. Reaction studies
2
(80% anatase and 20% rutile).
2
2
adsorbed on TiO is of interest [2–8]. Multiple studies show that
water increases the PCO and PCD rates of formic acid, but the rea-
sons for the rate increase are unclear [9–12]. Liao et al. [9] ob-
served from IR spectroscopy that the rate of formic acid and
formate PCO increased by a factor of two when water was above
2
For room-temperature PCD experiments, a thin layer of 45 mg
of P25 TiO (80% anatase and 20% rutile) or 30 mg of Pt/TiO cata-
2
2
6
2
0% saturation coverage, indicating the H O(a)/HCOO(a) ratio must
lyst was coated on the inner side of an annular Pyrex reactor to en-
sure that UV lights irradiated the entire catalyst surface. The
average film thickness was calculated to be 0.4 lm. To provide
adequate catalyst surface area with these thin films, the outer
diameter of the reactor was 2 cm and the reactor was 13 cm long.
The reactor consisted of two concentric cylinders that formed an
annular region with a 1-mm gap; the inner cylinder was evacuated
reach a certain value for water to substantially increase the photo-
oxidation rate. The optimum water concentration is vital because
previous work has shown that when the water concentration is
⇑
Corresponding author. Fax: +1 303 492 4341.
E-mail address: will.medlin@colorado.edu (J.W. Medlin).
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.08.011

K.L. Miller et al. / Journal of Catalysis 275 (2010) 294–299
295
and sealed. A diagram of the apparatus has been previously pre-
sented [15]. The annular reactor ensured a high gas flow rate
through the reactor, but allowed for a larger diameter reactor. High
flow rates guaranteed quick detection of products desorbing from
the surface. A thermocouple located on the inside of the reactor re-
corded the catalyst temperature.
reactor. A chromel–alumel thermocouple was placed in the center
of the catalyst bed, and a temperature controller was used to con-
trol a quartz tube furnace that ramped the catalyst temperature
from room temperature to 723 K at 1 K s . The temperature was
not increased above 723 K to avoid a phase change to rutile. A
Balzers quadrupole mass spectrometer sampled the effluent from
the reactor through a silica capillary. Before each experiment, the
ꢀ1
A 0.2 wt.% Pt/TiO
2
was prepared using 67 mg of H
in 100 ml HCl. First, Na CO
2
6
PtCl mixed
with 10 g Degussa P25 TiO
2
2
3
was added
2 2
TiO sample was oxidized for 30 min at 723 K in a 20% O /80%
until a pH of 7–8, and then 3 mL of acetic acid was added until the
solution reached a pH of 4. Next, 300 mL of water was added, and
He stream, and then cooled to room temperature. During TPD, he-
3
ꢀ1
lium flowed through the reactor at a rate of 100 cm (STP) min
.
N
2
was bubbled through the solution under an UV lamp for 6 h. The
Liquid formic acid and water were injected upstream of the
room-temperature catalyst with a syringe, and the liquids evapo-
rated in the flowing stream so that only vapor contacted the cata-
lyst. A TPD to 723 K was performed 1 h after the last injection. For
calibration, known amounts of formic acid, water, and carbon
monoxide were injected into the gas stream below the reactor.
solution was filtered and washed with H O. The catalyst was then
2
dried at 373 K for 24 h and coated on the inside of the reactor by
evaporative deposition of a sonicated slurry of catalyst and water.
Twelve 8-W UV lamps (BLB Korea, type F8T5BLB), placed in a circle
6
cm from the reactor, radiated the reactor. Radiometer measure-
ꢀ2
ments showed a UV intensity of 2.5 mW cm and a maximum
light intensity at 360 nm. Turning the lights on and off started
and stopped the PCD reaction.
3
. Results
A quartz tube furnace held the catalyst temperature at 723 K for
0 min in a 20% O /80% He stream to remove hydrocarbons and
2
3
.1. Photocatalytic decomposition of formic acid on TiO and Pt/TiO
2 2
3
deposits from previous reactions. The transparent furnace allowed
for the transmission of UV light during heating and cooling of the
During transient PCD at room temperature, physisorbed water
increased the rate of formic acid decomposition on both TiO and
Pt/TiO . On TiO , the CO formation rate only increased by 25%
Fig. 1), whereas on Pt/TiO , the CO formation rate increased a factor
2
reactor. The reactor was cooled to room temperature, and 2
liquid formic acid (Aldrich, 99%) was injected into the reactor.
For experiments with co-adsorbed water, 2 L of a mixture of for-
lL of
2
2
2
(
2
2
l
of 8, as shown in Fig. 2. During the initial period of the reaction in
this transient experiment, the reaction rate decays exponentially
as the coverage of formic acid on the catalyst decreases due to
reaction. Water was injected upstream of the reactor after 310 s
of PCD; at that time, the rate was 15% of its initial value. Water in-
creased the PCD rate a factor of eight, and it increased the rate
above the rate for the initially saturated surface, even though the
formic acid coverage was significantly lower after 310 s of reaction.
In a separate experiment, water and formic acid were co-adsorbed
and the reactor flushed before PCD. The PCD rate was essentially
the same as for formic acid alone. Water that remained adsorbed
in the absence of gas-phase water did not accelerate the PCD rate
mic acid and distilled water was injected. The reactor was flushed
for 2 h, and then turning on the UV lights initiated isothermal reac-
tion at room temperature. For experiments with physisorbed
water, 1 lL of water was injected during PCD. During transient
3
ꢀ1
PCD, 100 cm min (STP) of He or Ar flowed through the reactor,
and a Balzers quadrupole mass spectrometer (QMA 125) detected
products as a function of time. A computer recorded the ampli-
tudes of multiple mass peaks and the catalyst temperature simul-
taneously. Known volumes of gases injected downstream from the
2
reactor were used to calibrate the mass spectrometer. The O con-
centration in the feed stream, estimated be 0.3 ppm, remained be-
low the detection limit of the mass spectrometer [16]. After PCD,
2
on Pt/TiO . Physisorbed water apparently enhanced the formic acid
2
TPD (He or Ar flow) and TPO (20% O flow) were performed by
PCD rate, but strongly adsorbed water did not. Liao et al. [9] did not
distinguish between physisorbed and chemisorbed water, but
determined that 60% saturation coverage of water was required
ꢀ1
heating the catalyst to 723 K at a rate of 1 K s while monitoring
the mass signals with the mass spectrometer.
2
to increase the formic acid PCO rate on TiO .
2
.2. Infrared spectroscopy
The mechanism for the acceleration of the PCD rate by water is
not immediately clear. There are many complex aspects of surface
chemistry; for example, the P25 catalyst consists of a mixture of
A Nicolet 6700 FT-IR Spectrometer recorded infrared spectra at
ꢀ1
4
cm
resolution. Titania samples were prepared by pressing
powder into a tungsten grid with a hydraulic
2
TiO crystal structures (80% anatase and 20% rutile), and even
Degussa P25 TiO
2
press. A chromel–alumel thermocouple spot-welded to the tung-
sten grid measured the temperature, and the grid was attached
to a copper holder and placed in a small stainless steel vacuum
0
0
.16
.12
2 2
chamber with CaF windows for the IR beam. The TiO catalyst
was cleaned by heating in vacuum for 4 h at 673 K and then in oxy-
gen at 473 K for 2 h. Formic acid was adsorbed by exposing the
sample to saturated vapor at room temperature. The cell was then
evacuated for 20 min to remove weakly adsorbed formic acid.
Saturated water was added to the cell at room temperature, and
IR spectra were recorded at various water pressures as water was
slowly removed from the system. The IR spectra for gas-phase
water was subtracted from the IR spectra recorded to correct for
water in the gas phase.
0
0
.08
.04
0
Water
injection
0
400
800
1200
2.3. Temperature-programmed desorption
Time, s
A packed-bed tubular reactor (7-mm i.d.) was used for TPD and
Fig. 1. The CO
2
formation rate during formic acid PCD on TiO
2
. The UV lights were
TPO measurements. The catalyst (106 mg of Degussa P25 TiO
2
) was
turned on (open triangles) and off (solid triangles) as indicated. Water was injected
after 650 s.
placed on quartz wool, which rested on a quartz frit in a quartz

2
96
K.L. Miller et al. / Journal of Catalysis 275 (2010) 294–299
Table 1
Vibration frequencies from FTIR of formic acid molecularly and dissociatively
adsorbed on TiO
0
0
0
.3
.2
.1
2
.
Assignments
HCOOH
2
HCOOH and H O
COO and CH Formate
CH Formic Acid
CH Formate
C=O p-Formic Acid
c-Formic
COO- asym, Formate
COO- sym, Formate
COO- sym, Formate
CH Formate
2945
2921
2867
1787
1675
1550
1378
1360
1323
1263
1105
2945
2921
2861
1565
1360
1360
1323
Water
injection
CO or CH c-Formic Acid
CO or CH p-Formic Acid
0
0
200
400
600
Time, s
c denotes chemisorbed and p denotes physisorbed.
Fig. 2. The CO
2
formation rates during formic acid PCD on Pt/TiO
2
. The UV lights
were turned on (open triangles) and off (solid triangles) as indicated. Water was
injected after 310 s.
ing of chemisorbed formic acid [9]. It is noted that OH stretching
frequencies for adsorbed hydroxyls have also previously been re-
ꢀ1
ported close to 1263 cm [25,26]. The peaks at 2945, 2921, and
ꢀ
1
within a single type of crystal, multiple facets will be exposed at
the surface. Furthermore, the accelerating effect of water is much
greater on Pt/TiO , suggesting that chemistry at Pt–TiO interfaces
2 2
may be important to the mechanism. To make progress in investi-
gating how water may play a role in this and other related reac-
tions, the investigations below focus on the effect of water on
the adsorbed states of formic acid and a key initial decomposition
2
867 cm correspond to formic acid and formate CH stretching.
ꢀ1
The large peaks at 1550 and 1378 cm correspond to formate
COO– asymmetric and symmetric stretching. Previous studies indi-
cated that formic acid adsorbs molecularly on anatase (1 0 1) and
dissociatively (to produce formate) on rutile and anatase (0 0 1)
[13,27–31]. The most abundant surface of P25 TiO2 is anatase
(1 0 1) [32], which is the most stable anatase surface [14,33–35],
but other anatase or rutile surface planes or defects could explain
the presence of formate. Gong et al. used DFT to determine that for-
mic acid adsorbs as bidentate formate on anatase (0 0 1) [34].
Adsorption geometries of formate are identified by examining
the frequency difference between COO– asymmetric and symmet-
2 2
intermediate, formate, on TiO . In simulations, TiO is in most cases
modeled as an anatase (1 0 1) single crystal, the most abundant
crystal face, and the role of Pt is not directly considered. In Sec-
tion 4, we address the possible limitations of the model assump-
tions made in this investigation and the implications of this
study for the promoting effect of water in the catalysis.
ric stretches, denoted as
Dvas-s, relative to the aqueous ionic
ꢀ1
formate (
Dvionic = 201 cm ) [18,36–40]. Aqueous ionic formate
3
3
.2. Formic acid adsorption
peaks for COO– asymmetric and symmetric stretching are at
ꢀ
1
1
567 and 1366 cm [18]. The correlations between
Dvas-s and
.2.1. Infrared Spectroscopy
adsorption geometry have been described as follows [39,40].
Spectra for both molecular formic acid and formate were ob-
served when formic acid adsorbed on P25 TiO
2
at room tempera-
D
D
D
v
v
v
as-s
as-s
as-s
>
<
D
D
v
v
ionic: monodentate coordination.
ionic: chelating or bidentate bridging.
ture, as shown in Fig. 3. The corresponding assignments (Table 1)
were based on a combination of DFT results and previous assign-
ments. The spectrum in Fig. 3 is similar to that reported previously
ꢁ Dvionic: bidentate chelating.
ꢀ1
[
1
1
6,9,12,17–23]. Some studies reported an additional peak at
The Dvas-s was 172 cm in Fig. 3, indicating that monodentate
ꢀ1
413 cm , which was assigned to formate [17,19,24]. Peaks at
formate was not on the surface. The bridging and chelating modes
cannot be distinguished from IR spectra. Previous computational
studies determined that the chelating formate structure is less sta-
ble then the bidentate bridging structure [14,27]. Thus, apparently
formate adsorbs in the bidentate bridging configuration on P25
ꢀ1
787 and 1675 cm have been assigned to the C@O stretch of
gas phase or physisorbed and chemisorbed formic acid, respec-
ꢀ1
tively [9]. The peak at 1105 cm has been assigned to CO stretch-
ing or CH bending for gas phase or physisorbed formic acid, and the
ꢀ1
peak at 1263 cm has been assigned to CO stretching or CH bend-
2
TiO at room temperature.
(
a)
1565
1550
(b)
0
.3
0.15
2
921
867
2861
1
360
323
2
945
1
675
2
1
378
0
0
.2
0
.10
.05
0
1
1
787
1105
1263
.1
0
0
1
800
1600
1400
1200
-1
1000
3100
3000
2900
2800
2700
Wavenumber (cm )
Wavenumber (cm-1)
at room temperature: (a) from 1000 to 1800 cm 1, (b) from 2700 to 3100 cm . The solid line shows a spectrum
ꢀ
ꢀ1
Fig. 3. Infrared spectra of formic acid and water on P25 TiO
2
collected after a saturating exposure of formic acid, and the dotted line shows a spectrum collected after water addition to the formic acid saturated surface.

K.L. Miller et al. / Journal of Catalysis 275 (2010) 294–299
297
3
.2.2. Temperature-programmed desorption
1550
Formic acid was adsorbed to saturation coverage on TiO
2
, and
lmol g of formic acid desorbed in a broad
1565
1360
ꢀ1
TPD results showed 52
1378
peak at 428 K and a smaller peak at 525 K (Fig. 4). Most of the for-
ꢀ1
mic acid decomposed to form 306
l
mol g
2
H O, which desorbed
ꢀ1
in peaks at 468 and 525 K, and 450
l
mol g CO, which desorbed
in overlapping peaks at 488, 525, and 570 K (Fig. 5). Kim et al.
and Muggli et al. identified the same desorption products for for-
mic acid TPD, but with a different number of desorption peaks
and slightly different temperatures. On P25 TiO
served formic acid and water desorption peaks at 400 K and attrib-
uted CO and H O desorption peaks at 525 and 600 K to formate
decomposition. From another study on the anatase (0 0 1), formic
acid and water desorbs at 390 K, and the remaining formate
2
, Kim et al. [41] ob-
2
1800
1700
1600
1500
1400
1300
-1
Wavenumber(cm )
decomposes to CO and H
et al. observed that formic acid desorbs and dehydrates to CO
and H O in a single peak at 525 K on P25 TiO [43].
2
O in a single peak at 570 K [42]. Muggli
2
Fig. 6. Infrared spectra of formic acid and water on P25 TiO . The dashed line shows
a spectrum collected after water addition to the formic acid saturated surface, and
the solid line shows a spectrum collected after heating that surface to 383 K.
2
2
3
3
.3. Effect of water on formic acid adsorption
ꢀ1
and 1105 cm decreased. Due to the large intensity of the OH
stretching peak for water, changes in the high-frequency region
are difficult to discern. The COO– asymmetric peak shifted from
550 to 1565 cm , and the COO– symmetric peak shifted from
378 to 1360 cm ; these changes increased the
.3.1. Infrared spectroscopy
Infrared studies show that water displaced formic acid and
changed the formate adsorption geometry on P25 TiO . When
ꢀ1
ꢀ1
1
1
2
D
v
as-s from 170
also shifted
water was added at its room-temperature saturation pressure to
the IR cell, the formic acid peaks decreased and the formate
COO– symmetric and asymmetric peaks shifted and increased in
amplitude (Fig. 3, dashed line). Chemisorbed formic acid peaks at
ꢀ1
ꢀ1
to 210 cm . The CH formate peak at 2867 cm
ꢀ1
slightly to 2861 cm . Vibrational calculations from DFT, described
in a separate contribution [44], and previous literature aided in
interpreting the peak shifts. Heating the sample to 383 K removed
all the adsorbed water, and largely reversed the COO– asymmetric
and symmetric peak shifts. The formate peak at 1565 cm shifted
back to 1550 cm , the frequency observed before the water addi-
ꢀ1
1
675 and 1263 cm and physisorbed formic acid peaks at 1787
ꢀ1
0
0
.4
.3
ꢀ1
ꢀ1
tion, and the peak at 1360 cm broadened to include a smaller
peak at 1378 cm , as shown in Fig. 6.
ꢀ
1
The changes in the vibrational spectrum associated with for-
mate after dosing of water suggest a change in the structure of
the coadsorbate. One possible transition consistent with the higher
reactivity of the water-dosed surface is the conversion of bidentate
formate to a solvated form of monodentate formate, the latter of
which has been found to be more reactive toward decomposition
in prior investigations on model surfaces. This possibility is ex-
plored in more detail in a separate contribution [44].
0
0
.2
.1
0
300
400
500
600
700
Temperature (K)
3
.3.2. Temperature-programmed desorption
Fig. 4. Formic acid TPD recorded after formic acid and water adsorption on P25
TiO . The solid line shows a spectrum collected after a saturating exposure of formic
acid, and the dashed line shows a spectrum collected after water addition to the
formic acid saturated surface.
The TiO
posed to 5
2
surface was saturated with formic acid and then ex-
L of water. The water displaced 70 ± 7% of the formic
2
l
acid that otherwise would have desorbed intact during TPD
(Fig. 4). Water displaced the weakly adsorbed formic acid, so that
the rest that desorbed intact had a peak temperature of 525 K.
Thirty percent less CO desorbed after water adsorption, and the
CO peak at 488 K was absent, whereas the other two CO peaks
were unchanged (Fig. 5). The CO peak temperatures did not shift
because monodentate formate reverted to bidentate formate above
4
3
3
83 K, as seen from IR results (Fig. 6). The water peak temperature
did not change, but the amount desorbed increased to
2
1
0
ꢀ1
ꢀ1
3
26
l
mol g . The water increase, 20
l
mol g , is comparable to
ꢀ
1
the formic acid decrease, 36
l
mol g . Even though formic acid
competes for sites favorably with water, it can be displaced if for-
mic acid is not in the vapor phase. Henderson et al. [2] reported
that water removes acetone from rutile TiO
adsorbs more strongly than water.
2
, even though acetone
300
400
500
600
700
Temperature (K)
4
. Discussion
Fig. 5. Carbon monoxide TPD recorded after formic acid and water adsorption on
P25 TiO . The solid line shows a spectrum collected after a saturating exposure of
2
This discussion will relate the results presented above to
observations of the promotional effects of water on photocatalysis.
formic acid, and the dashed line shows a spectrum collected after water addition to
the formic acid saturated surface.

2
98
K.L. Miller et al. / Journal of Catalysis 275 (2010) 294–299
Photocatalytic decomposition of formic acid forms CO
TiO , but it forms CO and H on Pt/TiO . The reaction steps for formic
acid PCD on TiO and Pt/TiO are listed below in Eq. (4.1)–(4.7).
2
2
and H O on
limiting. Finally, it is important to note that rate acceleration could
be caused by the participation of species not directly invoked in
the reaction scheme shown above. For example, previous studies
have shown that UV exposure causes water to form hydroxyl
radicals that may accelerate the reaction [48,49]. The possible
importance of hydroxyls in this surface chemistry is further under-
scored by computational studies reported elsewhere [44].
From characterization experiments and DFT calculations re-
ported previously by Vittadini et al., water appears to influence
both formic acid dissociation and formate decomposition on
2
2
2
2
2
2
HCOOH $ HCOOHðsÞ
HCOOHðsÞ $ HðsÞ þ HCOOðsÞ
HCOOðsÞ $ CO þ HðsÞ
HðsÞ þ OðlatticeÞ ! H
OðsÞ $ H
HðsÞ $ HðPtÞ
HðPtÞ $ H
In the above equations, (s) denotes a species adsorbed on TiO
ð4:1Þ
ð4:2Þ
ð4:3Þ
ð4:4Þ
ð4:5Þ
ð4:6Þ
ð4:7Þ
2
2
2
OðsÞ
TiO
2
. There may be a larger effect on Pt/TiO
2
because the rate-lim-
it is
iting step is formic acid or formate decomposition, and on TiO
2
H
2
2
O
the extraction of lattice oxygen. Liao et al. [9] reported that formic
acid PCO was 53 times faster than formate PCO. This indicates the
effect of water on formate decomposition may have a greater effect
on the overall rate of PCD.
The possible higher reactivity of monodentate carboxylates is
supported by several previous studies. Henderson et al. [50] deter-
mined that monodentate carboxylates are more reactive then
2
2
2
and
(
Pt) denotes a species adsorbed on Pt. Thus, during formic acid PCD
on TiO , formic acid extracts a lattice oxygen from TiO to produce
CO and H O at room temperature. This creates an oxygen vacancy
that decreases the PCD rate [45,46]. Based on findings by Muggli
et al., [45], the rate-limiting step for formic acid PCD on TiO is
2
bidentate carboxylates on TiO (1 1 0). Using FTIR, Wu et al. [51]
2
2
showed that the monodentate PCO rate is 1.5 times higher than
the bidentate formate PCO rate for methoxy and ethoxy groups.
These results are consistent with an accelerated reaction rate
resulting from the conversion of bidentate formate to monodentate
formate. Also, Miura et al. [52] determined that formate decom-
poses on NiO (1 1 1) by first changing from bidentate to monoden-
tate, then formate rotating around the CO axis, and finally
hydrogen atom transferring from the formate to the surface atom.
2
2
2
likely the extraction of lattice oxygen. They saw a large rate de-
crease during PCD, but after a period of dark time that replenished
the oxygen lattice, the rate increased. Muggli et al. concluded that
lattice oxygen extraction causes the slow deactivation of TiO
Water did not re-oxidize TiO in their experiments, even during
UV illumination. One possible reason for the higher PCD rate on
Pt/TiO is that formic acid decomposition does not require lattice
oxygen removal since H forms through Eq. (4.7) preferentially over
O formation by Eq. (4.5). The Pt acts as a site for hydrogen atoms
to recombine to H , which then desorbs. The rate-limiting step for
formic acid PCD on Pt/TiO would then likely be either formic acid
or formate decomposition because formic acid adsorption (Eq. (4.1))
and H formation on Pt (Eq. (4.7)) are rapid [47].
Water has varying effects on steps 4.1–4.7. The reaction mech-
anism on TiO and Pt/TiO are the same through Eq. (4.3), and thus
water would have the same effect on steps 4.1–4.3 on Pt/TiO
From IR and TPD, water displaces formic acid on TiO , making
2
.
2
2
If formate decomposes by the same mechanism on TiO , the in-
creased conversion by water from bidentate to monodentate for-
mate would increase the rate of formate decomposition.
2
2
H
2
5
. Conclusions
2
2
Transient reaction studies showed that physisorbed water dra-
matically increased the rate of photocatalytic decomposition (PCD)
on Pt/TiO and mildly increased the PCD rate on TiO , whereas
chemisorbed water did not. Infrared spectroscopy showed both
molecular formic acid and bidentate formate on P25 TiO at room
2
2
2
2
2
.
2
2
temperature. Water displaced formic acid and shifted formate
COO– asymmetric and symmetric peaks, indicating a transition
from bidentate formate to monodentate formate. This transition
may relate to the increase in PCD rate.
2
the adsorption of formic acid in step 4.1 more difficult. Water
may increase the decomposition rate of formic acid to formate
(
[
Eq. (4.2)), as suggested by the DFT calculations of Vittadini et al.
14]. Water may also speed up formate decomposition (Eq. (4.3)).
The IR studies showed that water appeared to perturb the structure
of adsorbed formate, perhaps to a more reactive form, e.g., through
the conversion of bidentate formate to monodentate formate as
shown in step 4.8, with monodentate formate being more reactive,
as discussed below:
Acknowledgment
We gratefully acknowledge support by the National Science
Foundation, Grant CBET0730047.
References
H
H
[1] A.L. Linsebigler, G.Q. Lu, J.T. Yates, Chem. Rev. 95 (1995) 735.
[2] M.A. Henderson, J. Catal. 256 (2008) 287.
O
H
[
3] L.F. Liao, C.F. Lien, D.L. Shieh, M.T. Chen, J.L. Lin, J. Phys. Chem. B 106 (2002)
11240.
H
O
+
H O
O
CO +H O
O
O
2
2
2
[
[
4] S. Sato, JCS-Chem. Comm. (1982) 26.
5] S. Sato, J. Phys. Chem. 87 (1983) 3531.
Ti
Ti
Ti
Ti
[6] M. El-Maazawi, A.N. Finken, A.B. Nair, V.H. Grassian, J. Catal. 191 (2000) 138.
7] T.N. Obee, S.O. Hay, Environ. Sci. Technol. 31 (1997) 2034.
[8] C. Hagglund, B. Kasemo, L. Osterlund, J. Phys. Chem. B 109 (2005) 10886.
9] L.F. Liao, W.C. Wu, C.Y. Chen, J.L. Lin, J. Phys. Chem. B 105 (2001) 7678.
[10] S. Sato, K. Ueda, Y. Kawasaki, R. Nakamura, J. Phys. Chem. B 106 (2002) 9054.
[
ð4:8Þ
[
2
On TiO , Muggli and Falconer determined that water does not affect
the rate of lattice oxygen extraction (Eq. (4.4)), [45] and the desorp-
tion of water (Eq. (4.5)) is rapid. On Pt/TiO , water may increase the
[
[
11] D.S. Muggli, J.L. Falconer, J. Catal. 187 (1999) 230.
12] T. Chen, G.P. Wu, Z.C. Feng, G.S. Hu, W.G. Su, P.L. Ying, C. Li, Chin. J. Catal. 29
2
(
2008) 105.
hydrogen transport rate to Pt (Eq. (4.6)) if the hydrogen recombina-
tion rate is diffusion limited. Water’s effect on the hydrogen
recombination rate and desorption from Pt (Eq. (4.7)) was not inves-
tigated, but this step is sufficiently fast that it should not be rate
[13] H. Idriss, A. Miller, E.G. Seebauer, Catal. Today 33 (1997) 215.
[
[
14] A. Vittadini, A. Selloni, F.P. Rotzinger, M. Gratzel, J. Phys. Chem. B 104 (2000)
300.
15] A. Larson, J.A. Widegren, J.L. Falconer, J. Catal. 157 (1995) 611.
1
[16] D.S. Muggli, S.A. Keyser, J.L. Falconer, Catal. Lett. 55 (1998) 129.

K.L. Miller et al. / Journal of Catalysis 275 (2010) 294–299
299
[
[
17] T. Kecskes, J. Rasko, J. Kiss, Appl. Catal. A 268 (2004) 9.
18] J.R.S. Brownson, M.I. Tejedor-Tejedor, M.A. Anderson, J. Phys. Chem. B 110
2006) 12494.
19] Z.Q. Yu, S.S.C. Chuang, J. Catal. 246 (2007) 118.
20] R.C. Millikan, K.S. Pitzer, JCS 80 (1958) 3515.
21] T. van der Meulen, A. Mattson, L. Osterlund, J. Catal. 251 (2007) 131.
22] C.C. Chuang, W.C. Wu, M.C. Huang, I.C. Huang, J.L. Lin, J. Catal.185 185 (1999) 423.
23] G.Y. Popova, T.V. Andrushkevich, Y.A. Chesalov, E.S. Stoyanov, Kinet. Catal. 41
[34] X.Q. Gong, A. Selloni, A. Vittadini, J. Phys. Chem. B 110 (2006) 2804.
[35] A. Tilocca, A. Selloni, J. Phys. Chem. B 108 (2004) 4743.
[36] J.R.S. Brownson, M.I. Tejedor-Tejedor, M.A. Anderson, Chem. Mater. 17 (2005)
6304.
[37] F.P. Rotzinger, J.M. Kesselman-Truttmann, S.J. Hug, V. Shklover, M. Gratzel, J.
Phys. Chem. B 108 (2004) 5004.
[38] I. Dolamic, T. Burgi, J. Phys. Chem. B 110 (2006) 14898.
[39] G.B. Deacon, F. Huber, R.J. Phillips, Inorg. Chim. Acta 104 (1985) 41.
[40] G.B. Deacon, R.J. Phillips, Coordin. Chem. Rev. 33 (1980) 227.
[41] K.S. Kim, M.A. Barteau, Langmuir 4 (1988) 945.
(
[
[
[
[
[
(
2000) 805.
24] C.-C. Chuang, W.-C. Wu, M.-C. Huang, I.C. Huang, J.-L. Lin, J. Catal. 185 (1999)
23.
[
4
[42] K.S. Kim, M.A. Barteau, Langmuir 6 (1990) 1485.
[
[
[
[
25] A. Yee, S.J. Morrison, H. Idriss, J. Catal. 186 (1999) 279.
26] J.L. Davis, M.A. Barteau, Surf. Sci. (1990) 235.
27] P.R. McGill, H. Idriss, Surf. Sci. 602 (2008) 3688.
[43] D.S. Muggli, J.T. McCue, J.L. Falconer, J. Catal. 173 (1998) 470.
[44] K.L. Miller, J.W. Medlin, J.L. Falconer, J. Catal. (2010), submitted for publication.
[45] D.S. Muggli, J.L. Falconer, J. Catal. 191 (2000) 318.
28] H. Uetsuka, M.A. Henderson, A. Sasahara, H. Onishi, J. Phys. Chem. B 108 (2004)
[46] G.D. Lee, J.L. Falconer, Catal. Lett. 70 (2000) 145.
1
3706.
[47] K. Mogyorosi, A. Kmetyko, N. Czirbus, G. Vereb, P. Sipos, A. Dombi, React. Kinet.
Catal. Lett. 98 (2009) 215.
[
[
[
29] M.A. Henderson, J. Phys. Chem. B 101 (1997) 221.
30] B.E. Hayden, A. King, M.A. Newton, J. Phys. Chem. B 103 (1998) 203.
31] Y. Uemura, T. Taniike, M. Tada, Y. Morikawa, Y. Iwasawa, J. Phys. Chem. C 111
[48] K. Takahashi, H. Yui, J. Phys. Chem. C 113 (2009) 20322.
[49] J. Peral, X. Domènech, D.F.J. Ollis, Chem. Technol. Biotechnol. 70 (1997) 117.
[50] M.A. Henderson, J.M. White, H. Uetsuka, H. Onishi, J. Catal. 238 (2006) 153.
[51] W.C. Wu, C.C. Chuang, J.L. Lin, J. Phys. Chem. B 104 (2000) 8719.
[52] T. Miura, H. Kobayashi, K. Domen, J. Phys. Chem. B 105 (2001) 10001.
(
2007) 16379.
[
[
32] J.F. Porter, Y.-G. Li, C.K. Chan, J. Mater. Sci. 34 (1999) 1523.
33] A. Vittadini, A. Selloni, F.P. Rotzinger, M. Gratzel, Phys. Rev. Lett. 81 (1998)
2
954.