Photocatalytic Oxidation of Ethanol: Isotopic Labeling and Transient Reaction

Article abstract of DOI:10.1021/jp960587x

Transient reaction techniques were combined with isotope labeling to study the reaction steps for the room-temperature, photocatalytic oxidation (PCO) of ethanol on TiO2.Carbon-13 labeled ethanol (CH3(13)CH2OH) was adsorbed on the catalyst and photocatalytically oxidized in the absence of gas-phase ethanol.The amounts of species remaining on the surface after PCO were determined by temperature-programmed oxidation.During PCO, only CO2 and H2O formed for low coverages of ethanol, whereas acetaldehyde also desorbed for saturation coverage.Acetaldehyde forms rapidly from ethanol oxidation during PCO.At both low and high ethanol coverages, the α-carbon is preferentially oxidized and thus (13)CO2 forms faster than (12)CO2 at short illumination times.At longer times, the rates of (13)CO2 and (12)CO2 formation are nearly identical.The difference in behavior between (13)CO2 and (12)CO2 formation suggests two parallel reactions of ethanol, which may be due to two adsorption sites on TiO2.

Full text of DOI:10.1021/jp960587x

15886
J. Phys. Chem. 1996, 100, 15886-15889
Photocatalytic Oxidation of Ethanol: Isotopic Labeling and Transient Reaction
Darrin S. Muggli, Sheldon A. Larson,† and John L. Falconer*
Department of Chemical Engineering, UniVersity of Colorado, Boulder, Colorado 80309-0424
ReceiVed: February 27, 1996; In Final Form: June 10, 1996X
Transient reaction techniques were combined with isotope labeling to study the reaction steps for the room-
temperature, photocatalytic oxidation (PCO) of ethanol on TiO2. Carbon-13 labeled ethanol (CH313CH2OH)
was adsorbed on the catalyst and photocatalytically oxidized in the absence of gas-phase ethanol. The amounts
of species remaining on the surface after PCO were determined by temperature-programmed oxidation. During
PCO, only CO2 and H2O formed for low coverages of ethanol, whereas acetaldehyde also desorbed for saturation
coverage. Acetaldehyde forms rapidly from ethanol oxidation during PCO. At both low and high ethanol
coverages, the R-carbon is preferentially oxidized and thus 13CO2 forms faster than 12CO2 at short illumination
times. At longer times, the rates of 13CO2 and 12CO2 formation are nearly identical. The difference in behavior
between 13CO2 and 12CO2 formation suggests two parallel reactions of ethanol, which may be due to two
adsorption sites on TiO2.
Introduction
rapidly than partial oxidation when the O2 concentration
decreased during 2-propanol PCO.16 Temperature-programmed
oxidation (TPO) was used to monitor adsorbed compounds
before and after PCO. Isotopically labeled ethanol (CH313CH2-
OH) was used with the R-carbon labeled with 13C so that we
could follow the reactivities of each carbon, and a significant
difference in reactivity was observed. Less-than-saturation
coverage was used in some experiments so that intermediates
stayed on the catalyst surface and their cracking fractions did
not interfere with detection of 12CO2 and 13CO2.
Heterogeneous photocatalytic oxidation (PCO) has great
potential for completely oxidizing organic pollutants in dilute
waste streams. A wide range of organics can be destroyed
photocatalytically1-11 at room temperature on TiO2 catalysts in
the presence of UV or near-UV illumination, but the reaction
mechanisms are poorly understood. The catalyst absorbs UV
light, which excites electrons from the valence to the conduction
band and creates a hole in the valence band of the semiconductor
catalyst. The resulting electron/hole pairs can migrate to the
catalyst surface and initiate redox reactions with organic
pollutants. Ethanol was used in this study because it is a
common solvent in spraying and coating applications and a
major component of bakery and brewery emissions. Industrial
gas emissions often contain contaminants in low concentrations
(<100 ppm) and flow at high rates, and ethanol can be oxidized
by PCO under these conditions.12
Previous studies of ethanol PCO in the gas-phase identified
several reaction intermediates. Cunningham et al.13 illuminated
TiO2 with a pulse of UV light as ethanol flowed over the catalyst
and detected acetaldehyde, ethylene, and ethane as products.
Nimlos and co-workers14 recently studied ethanol PCO in a
recirculating batch reactor and identified acetaldehyde, acetic
acid, formaldehyde, and formic acid as important intermediates.
Methyl formate, ethyl formate, and methyl acetate were identi-
fied as minor products. Sauer and Ollis15 also studied ethanol
PCO in a recirculating stream of humidified air. They proposed
that ethanol oxidizes to acetaldehyde, which then oxidizes in
three parallel reactions to form CO2. The intermediates were
acetic acid, formaldehyde, and formic acid.
Experimental Methods
The apparatus used for PCO and TPO was described
previously.16 Approximately 30 mg of Degussa P-25 TiO2
catalyst was coated on the inside of an annular Pyrex reactor.
This procedure formed a thin layer with an average thickness
of less than 0.5 µm. Although the film thickness was not
uniform, all of the TiO2 was illuminated during PCO, as verified
by the eventual PCO of all adsorbed ethanol. The reactor was
constructed with a 1-mm annular spacing so that high gas flow
rates could be maintained across the catalyst to minimize mass
transfer effects and rapidly flush gas-phase products from the
reactor. A cylindrical quartz furnace wrapped with Kanthal wire
surrounded the reactor. The heating-wire coils were spaced at
0.5-1 cm intervals so that they did not block a significant
amount of the light from UV lamps that surrounded the furnace.
The six 4-W UV lamps (GE F4T5-BLB) were 2.5 cm from the
catalyst film. These tubes generate light with a maximum
intensity near 390 nm.17 A 0.5-mm OD chromel-alumel
thermocouple was inserted into the annulus so that the tip was
in contact with the catalyst film. The thermocouple provided
feedback to a temperature controller, which produced a constant
heating rate during TPO. Desorbed products were analyzed by
a quadrupole mass spectrometer (Balzers, QMA 125) housed
in an ultrahigh-vacuum chamber located immediately down-
stream of the reactor. A heated stainless steel capillary tube
diverted some of the effluent into the vacuum chamber.
Computer-controlled data acquisition simultaneously recorded
multiple mass signals, temperature, and elapsed time.
In this study, transient reaction techniques were used to gain
insight into the surface reaction steps during PCO since these
transient techniques can separate and identify adsorption, surface
reaction, and desorption steps. Adsorption was separated from
the other steps by adsorbing ethanol at room temperature and
then reacting it in the absence of gas-phase ethanol. A low O2
concentration (0.2% O2 in He) was used during PCO to slow
the reaction rate so that intermediate products were more readily
detected. The rate of complete oxidation decreased much more
Before each experiment, the reactor was heated to 723 K in
flowing 0.2% O2/He (100 standard cm3/min) to generate a clean
oxide surface. Labeled ethanol was injected upstream of the
† Present address: National Renewable Energy Laboratory, 1617 Cole
Boulevard, Golden, Colorado, 80401-3393.
X Abstract published in AdVance ACS Abstracts, September 1, 1996.
S0022-3654(96)00587-4 CCC: $12.00 © 1996 American Chemical Society

Photocatalytic Oxidation of Ethanol
J. Phys. Chem., Vol. 100, No. 39, 1996 15887
Figure 2. Temperature-programmed oxidation spectra of TiO2 after
PCO of adsorbed CH313CH2OH (50% of saturation coverage).
TABLE 1: Desorption Amounts Measured during PCO and
TPO of Ethanol
Figure 1. Product desorption rates measured during PCO in 0.2% O2
of adsorbed CH313CH2OH (50% of saturation coverage) on TiO2.
desorption amount (µmol/g of TiO2)
ethanol acetaldehyde ethylene 13CO2 CO2 13CO CO
reactor into the flowing gas and allowed to evaporate and adsorb
onto the room-temperature catalyst. To obtain saturation
coverage, ethanol was injected until no additional uptake was
seen. The ethyl-1-13C alcohol (CH313CH2OH) was obtained
from Isotec with a minimum 99% 13C enrichment purity.
Excess ethanol was flushed from the reactor by the flowing gas.
Before initiating PCO, a metal shroud was placed around the
reactor and the UV lights were switched on and allowed to reach
a constant output. Reaction was initiated by removing the
shields. After a specified reaction time, the lights were switched
off and TPO was carried out by heating the catalyst to 723 K
in 0.2% O2 at a constant rate of 1 K/s . The mass spectrometer
signals were calibrated by injecting known volumes of gases
or liquids into the flowing gas downstream of the reactor. Mass
signals unique to desorbed products were measured whenever
possible. Otherwise, signals were corrected for contributions
from ions of identical mass from other species because of
cracking in the mass spectrometer. The 13C-labeled species were
calibrated for by using the corresponding 12C species. Although
water is a reaction product, its signal is not plotted in the figures
because its calibration is less accurate and because water
appearance in the gas phase is limited by desorption from the
TiO2 surface so that the water signal does not provide kinetic
information.
Less-than-Saturation Coverage
PCO
TPO
103
32
74
38
<1
<1
25
45
45
62
Saturation Coverage
PCO
TPO
82
3.8
4.5
104
68
71
71
6.4
8.5
the initial ethanol coverage. At higher O2 concentrations, the
TOF is higher but the initial 13CO2 rate is still higher than the
12CO2 rate.
Many mass peaks were checked to be sure that the difference
between the 44 and 45 mass signals was due to a difference in
the rates of formation of 12CO2 and 13CO2 and not to a cracking
fraction at mass 45 from an undetected reaction product. In
addition, the ratios of the 12/44 and 13/45 mass signals were
constant. This indicates that the mass 12 and 13 signals resulted
only from cracking of 12CO2 and 13CO2, respectively, and no
other carbon-containing compounds were present in the gas
phase. If some â-carbon in the ethanol was labeled with 13C
in addition to the R-carbon, more 13CO2 than 12CO2 might be
expected to form. The shapes of the 13CO2 and 12CO2 signals
during PCO would be expected to be the same, however, if the
only difference was that more 13C than 12C was present in the
adsorbed ethanol. Previous experiments determined that identi-
cal amounts of 13C and 12C were present in the ethanol sample.18
Temperature-programmed oxidation was used to determine
the isotopic composition of the species remaining on the surface,
and the spectra in Figure 2 showed that more 12C-containing
species than 13C-containing species formed during TPO. This
is expected since more 13CO2 formed than 12CO2 during PCO,
and the ratio of 13C to 12C in the ethanol was 1. Table 1 lists
the amounts of the products detected during PCO and TPO.
During TPO, small amounts of labeled ethanol and acetaldehyde
desorbed, but the major products were carbon oxides and water.
Results and Discussion
Less-than-Saturation Coverage. As soon as the catalyst
was illuminated, adsorbed ethanol reacted at room temperature
in 0.2% O2 to form CO2. As shown in Figure 1 for half of the
saturation coverage of CH313CH2OH, both 13CO2 and 12CO2
formed. Water was also observed but is not included in the
figure; no other products were detected. The O2 signal also
immediately decreased when the lights were turned on. What
is most interesting about this transient reaction experiment is
that 13CO2 formed at a significantly faster rate than 12CO2
initially, and only at long times were their rates comparable.
That is, the R-carbon in ethanol (the one bonded to the OH
group) preferentially oxidized to CO2. Since CO2 is weakly
adsorbed on TiO2 at room temperature,16 its appearance is
reaction limited, not desorption limited.
Both carbon dioxide products went through maxima at
different illumination times (90 s for 13CO2 and 500 s for
12CO2), and PCO was still taking place after 3700 s. By this
reaction time, significantly more 13CO2 than 12CO2 had formed,
and thus the species remaining on the surface were enriched in
12C. When the lights were turned off, both CO2 signals rapidly
dropped to zero, as expected for a reaction-limited process. For
the low O2 concentration used (0.2%), the initial turnover
frequency (TOF) was estimated to be 6.0 × 10-4 s-1, based on
The 12CO signal was significantly larger than the 13CO signal
over the entire temperature range, whereas the 13CO2 and
-
12
CO2 signals were similar in amplitude. The total amount of
13C oxides was essentially the same as the total amount of 12C
oxides formed during PCO plus TPO (160 µmol of 13C/g of
catalyst and 157 µmol of 12C/g of catalyst). This mass balance
further shows that the differences in the 13CO2 and 12CO2 signals
during PCO were not due to an experimental artifact. As a
further check, TPO was carried out for adsorbed 13C ethanol
without PCO. In contrast to TPO after PCO, the 12CO and 13
-
CO signals were identical, and the 12CO2 and 13CO2 signals
were almost identical.
As seen in Figure 2, both 12CO and 13CO formed in two peaks
at 550 and 640 K during TPO, but the relative amplitudes were

15888 J. Phys. Chem., Vol. 100, No. 39, 1996
Muggli et al.
total amount of 12C leaving the surface during PCO plus TPO
was equal to the total amount of 13C. That is, a mass balance
was obtained.
Acetaldehyde Adsorption and Oxidation. During PCO of
2-propanol, the appearance of the partial oxidation product
acetone in the gas phase was limited by desorption and not by
reaction.16 Acetone desorbed as the surface became saturated,
apparently because it was displaced by water that formed during
PCO. Since acetaldehyde was observed during PCO for high
but not low ethanol coverages, the same process may take place.
Indeed, when 1.0 µL of water was injected into the carrier gas
following adsorption of saturation coverage of acetaldehyde,
acetaldehyde desorbed. With additional H2O injections, more
acetaldehyde desorbed but in decreasing amounts. Water
displaced approximately 30% of the adsorbed acetaldehyde,
suggesting that acetaldehyde may be adsorbed on two types of
sites, only one of which is sufficiently weakly bound that it
can be displaced by water. Approximately 26% of the adsorbed
ethanol desorbed as acetaldehyde during PCO.
Figure 3. Product desorption rates measured during PCO in 0.2% O2
of adsorbed CH313CH2OH (saturation coverage) on TiO2.
Photocatalytic oxidation of adsorbed acetaldehyde at low
coverage formed CO2 in a similar manner to PCO of ethanol.
Since adsorbed ethanol quickly reacts to form acetaldehyde
(Figure 3) and acetaldehyde is not displaced by water at low
coverages (Figure 1), PCO’s of ethanol and acetaldehyde are
expected to be similar. At saturation coverage of acetaldehyde,
not much acetaldehyde desorbed during PCO, whereas a large
amount of acetaldehyde desorbed during PCO of ethanol at
saturation coverage. Thus, during PCO at saturation coverages,
the initial rate of CO2 formation was greater for acetaldehyde
than that for ethanol, since more acetaldehyde remained on the
surface.
Figure 4. Temperature-programmed oxidation spectra of TiO2 after
PCO of adsorbed CH313CH2OH (saturation coverage).
different. The peak at 550 K was larger than the peak at 640
K for 12CO, but the opposite was true for the 13CO. The similar
peak temperatures of the 13CO and 12CO signals indicate that
the precursors that formed the carbon monoxide during TPO
contained both 13C and 12C or that a single carbon precursor
was present in both 12C and 13C forms. The difference in the
relative magnitudes of the two CO peaks for the two isotopes
suggests that the two peaks were formed from different surface
species. If a single species was responsible for both peaks, one
would expect that the ratio of the relative amplitudes of the
two peaks would be the same for 12CO and 13CO. The 12CO2
formed in one broad peak at 675 K, but 13CO2 appeared to have
a low-temperature shoulder. The carbon oxides signals were
not zero when the heating was stopped, but the catalyst was
held at 723 K until the rates dropped to the baseline.
Saturation Coverage. At saturation coverage, PCO of
ethanol formed 13CO2 faster than 12CO2 initially, as observed
at lower coverage, but additional products were seen. Acetal-
dehyde and ethylene appeared in the gas phase almost im-
mediately after UV illumination, whereas the maximum rates
of 13CO2 and 12CO2 were delayed in time. The large amount
of acetaldehyde in the gas phase (Figure 3) indicates that in
0.2% O2 the oxidation of ethanol to form acetaldehyde is much
faster than the subsequent oxidation of acetaldehyde to CO2
and H2O. The rates of CO2 formation were not greater than
those obtained at low coverage, at least partly because a lot of
acetaldehyde intermediate desorbed before it was oxidized
further. The acetaldehyde signal reached a maximum after 80
s of illumination and then decreased exponentially. A small
ethylene peak was also observed with a maximum rate after 80
s. In contrast, the 13CO2 and 12CO2 rates reached broad maxima
after 475 and 975 s of illumination, respectively, and 12CO2
and 13CO2 were still forming at significant rates when the lights
were turned off. The carbon oxide peaks observed during TPO
following PCO (Figure 4) were similar to those seen for low
coverage of ethanol, but the rates were greater. Small ethanol,
acetaldehyde, and ethylene peaks also were observed during
TPO. More 12CO formed than 13CO during TPO so that the
Parallel Reactions. Isotopic labeling showed that the
R-carbon in ethanol was more rapidly oxidized to CO2 than the
â-carbon during PCO. Since the maximum 12CO2 rate occurred
approximately 400 s after the maximum 13CO2 rate and thus
the 12CO2 rate was increasing when the 13CO2 rate was
decreasing, the two types of carbon do not appear to oxidize in
the same reaction step initially. After extended PCO times,
however, the 13CO2 and 12CO2 rates were almost identical. That
is, ethanol appears to oxidize by two parallel reaction pathways
that have different rates. Sauer and Ollis15 concluded from
kinetic measurements that ethanol first oxidizes to acetaldehyde,
which then oxidizes by three pathways. One pathway is direct
oxidation of acetaldehyde to CO2, the second pathway goes
through an acetic acid intermediate and then forms CO2, and
the third goes through formaldehyde to formic acid and then to
CO2. Our displacement experiment shows that acetaldehyde
is adsorbed in more than one form or on more than one type of
site since only 30% of a saturation layer could be displaced by
water. Acetaldehyde adsorbed on different sites might be
expected to follow different pathways.
On one type of site, 13CO2 forms quickly from oxidation of
the R-carbon, and a relatively stable intermediate containing
â-carbon remains on the surface. On another type of site,
oxidation is slower and 13CO2 and 12CO2 form at the same rate.
When UV lights are turned on, PCO takes place on both types
of sites, but the total rate of 13CO2 is greater because reaction
is faster on the type of site that preferentially oxidizes the
R-carbon. After the acetaldehyde on these sites reacts to
completion, reaction on the less-active sites continues to form
13CO2 and 12CO2 at equal rates. The differences between the
amounts of 13CO2 and 12CO2 formed are 26 and 29 µmol/g of
catalyst for the two coverages used, indicating that the more
active sites are almost completely occupied during both the low-
and high-coverage experiments. The more active sites comprise

Photocatalytic Oxidation of Ethanol
J. Phys. Chem., Vol. 100, No. 39, 1996 15889
(8) Wada, K.; Yoshida, K.; Takatani, T.; Watanabe, Y. App. Catal. A
approximately 9% of the total sites available for ethanol
adsorption. Our isotope experiments can distinguish two
reaction pathways, but three pathways are possible with either
two of the pathways having the R-carbons react faster or two
pathways having the two carbons react at the same rate.
1993, 99, 21.
(9) Ait-Ichou, I.; Formenti, M.; Pommier, B.; Teichner, S. J. J. Catal.
1985, 91, 293.
(10) Aupo, M.; Aikawa, N.; Kubokawa, Y. J. Phys. Chem. 1984, 88,
3998.
(11) Jacoby, W.; Blake, D.; Noble, R. D.; Koval, C. A. J. Catal. 1995,
157, 87.
Acknowledgment. Acknowledgment is made to the donors
of the Petroleum Research Fund, administered by the American
Chemical Society, for the support of this work.
(12) Miller, A.; Fox, R. In The Proceedings of the 1st International
Conference on TiO2 Photocatalytic Purification and Treatment of Water
and Air; Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier: Amsterdam, 1993; p
573.
References and Notes
(13) Cunningham, J.; Finn, E.; Samman, N. Faraday Disuss., Chem.
Soc. 1974, 58, 160.
(1) Sampath, S.; Uchida, H.; Yoneyama, H. J. Catal. 1994, 149, 189.
(2) Jacoby, W.; Nimlos, M.; Blake, D. EnV. Sci. Technol. 1995, 29,
1223.
(3) Pichat, P.; Herrman, J. M.; Disdier, J.; Mozzanega, M. J. Phys.
Chem. 1979, 83, 3122.
(4) Raupp, G. B.; Junio, C. T. App. Surf. Sci. 1993, 72, 321.
(5) Sauer, M. L.; Ollis, D. F. J. Catal. 1994, 149, 81.
(6) Sopyan, I.; Marasawa, S.; Hashimoto, K.; Fujishima, A. Chem. Lett.
1994, 4, 723.
(7) Stafford, U.; Gray, K. A.; Kamat, P. V.; Varma, A. Chem. Phys.
Lett. 1993, 205, 55.
(14) Nimlos, M. R.; Wolfrum, E. J.; Brewer, M. L.; Fennell, J. A.;
Bintner, G. To be published.
(15) Sauer, M. L.; Ollis, D. F. J. Catal. 1996, 158, 570.
(16) Larson, S. A.; Widegren, J. A.; Falconer, J. L. J. Catal. 1995, 157,
611.
(17) Dibble, L. A.; Raupp, G. B. EnViron. Sci. Technol. 1992, 26, 492.
(18) Cordi, E.; Falconer, J. L. Catal. Lett. 1996, 38, 45.
JP960587X