Gas-phase photooxidation of trichloroethylene on TiO2 and ZnO: Influence of trichloroethylene pressure, oxygen pressure, and the photocatalyst surface on the product distribution

Article abstract of DOI:10.1021/jp972351e

Transmission Fourier transform infrared spectroscopy has been used to identify gas-phase and surface-bound products and intermediates formed during the gas-phase photooxidation of trichloroethylene (TCE) on TiO₂ and ZnO. Several factors are found to influence the gas-phase product distribution for this reaction. On clean TiO₂ and ZnO surfaces and at high TCE and O₂ pressures, gas-phase CO, CO₂, COCl₂, CCl₂HCOCl, CHCl₃, C₂HCl₅, and HCl are produced, whereas at low TCE and O₂ pressures, TCE is converted to gas-phase CO and CO₂ only. In addition to TCE and O₂ pressure, the product distribution of the photooxidation of TCE is strongly dependent upon the coverage of adsorbed species on the surface of the photocatalyst. It is shown here that the complete oxidation of adsorbed TCE can occur on clean photocatalytic surfaces whereas only partial oxidation of adsorbed TCE occurs on adsorbate-covered surfaces. The role of adsorbed surface products in TCE photooxidation is discussed.

Full text of DOI:10.1021/jp972351e

J. Phys. Chem. B 1998, 102, 549-556
549
Gas-Phase Photooxidation of Trichloroethylene on TiO2 and ZnO: Influence of
Trichloroethylene Pressure, Oxygen Pressure, and the Photocatalyst Surface on the Product
Distribution
M. D. Driessen,† A. L. Goodman, T. M. Miller, G. A. Zaharias, and V. H. Grassian*
Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
ReceiVed: July 18, 1997; In Final Form: NoVember 6, 1997X
Transmission Fourier transform infrared spectroscopy has been used to identify gas-phase and surface-bound
products and intermediates formed during the gas-phase photooxidation of trichloroethylene (TCE) on TiO2
and ZnO. Several factors are found to influence the gas-phase product distribution for this reaction. On
clean TiO2 and ZnO surfaces and at high TCE and O2 pressures, gas-phase CO, CO2, COCl2, CCl2HCOCl,
CHCl3, C2HCl5, and HCl are produced, whereas at low TCE and O2 pressures, TCE is converted to gas-phase
CO and CO2 only. In addition to TCE and O2 pressure, the product distribution of the photooxidation of
TCE is strongly dependent upon the coverage of adsorbed species on the surface of the photocatalyst. It is
shown here that the complete oxidation of adsorbed TCE can occur on clean photocatalytic surfaces whereas
only partial oxidation of adsorbed TCE occurs on adsorbate-covered surfaces. The role of adsorbed surface
products in TCE photooxidation is discussed.
Introduction
been suggested that differences in product distribution of the
photooxidation of TCE on TiO2 can, in some cases, be attributed
to the reaction and hydrolysis of these chlorinated partial
oxidation products on the surface of TiO2 particles that, because
of reactor design, are not exposed to light.7
Trichloroethylene (TCE), a human health hazard, is the most
common halogenated contaminant found in groundwater sup-
plies.1 Therefore, a great deal of effort has been put forth in
developing methods to degrade or transform this contaminant
into more environmentally benign compounds.2-13 Several
approaches have been taken; some of the more promising
approaches involve the aqueous and gas-phase photooxidation
of TCE using semiconductor photocatalysts such as TiO2.2-13
These catalysts are activated by absorption of light with
wavelengths contained in the solar spectrum.
In this study, it is shown that gas-phase products formed in
the photooxidation of TCE on both TiO2 and ZnO are dependent
upon the reaction conditions employed. The product distribution
is found to depend on TCE pressure, O2 pressure, and the nature
of the photocatalyst surface. On clean TiO2 and ZnO surfaces
and under conditions of low pressures of TCE (144 mTorr) and
molecular oxygen (721 mTorr), CO and CO2 are the only
carbon-containing, gas-phase products formed. After the reac-
tion is run at low pressure, the catalyst surface is covered with
several adsorbed species including water and dichloroacetate.
Subsequent photooxidation of TCE on the adsorbate-covered
photocatalyst results in the formation of gas-phase CO, CO2,
and additional carbon-containing, gas-phase products. These
products include COCl2, DCAC, C2HCl5, CHCl3, and HCl. At
high initial pressures of TCE (P > 144 mTorr), CO and CO2
as well as COCl2, DCAC, CHCl3, and C2HCl5 are produced in
the gas phase when either clean or adsorbate-covered ZnO and
TiO2 photocatalysts are used. As discussed here, the different
product distributions for this reaction are caused by changes in
the photocatalyst surface due to adsorbed photoproducts.
Reaction 1 shows the ideal photooxidation reaction where
TCE is completely mineralized into CO2 and HCl. Either
adsorbed water or surface-bound hydroxyl groups present on
the semiconductor surface may participate in the reaction.
TiO2 or ZnO
Cl2CdCHCl + 3/2O2 + H2O
8
hν (E > bandgap, 3.2 eV)
2CO2 + 3HCl (1)
However, the complete mineralization of TCE is often times
not realized as other products have been identified in the
reaction. Recently, there has been some debate concerning the
product distribution and the mechanism of the gas-phase
photooxidation of TCE on TiO2.3-13 There are a few studies
Experimental Section
that report the complete mineralization of TCE during the
photooxidation of TCE on TiO2.
3-5
However, a greater number
The IR cell used in these experiments has been described
previously.14,15 The cell consists of a 23/4 in. stainless steel cube
with two differentially pumped barium fluoride windows and a
sample holder through which thermocouple and power feed-
throughs are connected to a photoetched tungsten sample grid.
The sample holder design is such that the sample may be cooled
to near liquid nitrogen temperatures and heated resistively up
to 1300 K. The temperature is monitored using thermocouple
wires spot-welded to the top of the sample grid. The cell is
attached to an all stainless steel vacuum chamber through a 2
of studies that report the formation of chlorinated partial
oxidation products such as phosgene, dichloroacetyl chloride
(DCAC), monochloroacetyl chloride (MCAC), dichloroacetic
acid (DCAA), and monochloroacetic acid (MCAA);6-13 some
of these compounds are even more toxic than TCE.11 It has
† Present address: Department of Chemistry, Southwest Missouri State
University, Springfield, MO 65804.
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, December 15, 1997.
S1089-5647(97)02351-1 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/15/1998

550 J. Phys. Chem. B, Vol. 102, No. 3, 1998
Driessen et al.
ft bellows hose. The vacuum system is pumped using an 80
L/s ion pump after being rough pumped with a turbomolecular
pump.
Samples are made by spraying a slurry of the powdered
semiconductor (TiO2, Degussa P25; ZnO, Aldrich) and deion-
ized water onto a tungsten grid (Buckbee-Mears) which is held
at approximately 573 K. A template is used to mask half of
the grid so that one side can be coated with the semiconductor
powder, and the other side is left blank. Approximately 60 mg
of the semiconductor powder is evenly coated onto a 3 cm ×
1 cm area of the grid.
Once the sample is prepared it is mounted inside the IR cell,
after which the cell is evacuated. The sample is then resistively
heated to 673 K for 21/2 h under vacuum and then oxidized at
the same temperature by introducing 100 Torr of O2 into the
IR cell for 30 min. After evacuating the sample cell for 15
min, 100 Torr of oxygen is admitted into the cell, and the sample
is cooled to room temperature. Once the sample has reached
room temperature, the cell is evacuated for several hours or
overnight. This cleaning procedure removes adsorbed hydro-
carbon impurities, including carbonates, from the surface. The
sample contains isolated hydroxyl groups with infrared frequen-
cies of 3715 and 3672 cm-1, in agreement with literature
values.16
Figure 1. Infrared spectra of TCE adsorbed on TiO2 as a function of
TCE pressure. The inset displays a plot of the integrated area of the
930 cm-1 band of adsorbed TCE as a function of TCE pressure. The
data show that the TiO2 surface saturates above a TCE pressure of
approximately 1.4 Torr.
TABLE 1: Vibrational Assignment of TCE Adsorbed on
The IR cell is then placed on a linear translator inside the
sample compartment of a Mattson RS-1 FT-IR spectrometer
equipped with a narrowband MCT detector. The linear transla-
tor allows each half of the sample grid to be translated into the
infrared beam. This permits the investigation of gas-phase and
adsorbed species on the photocatalyst surface under identical
reaction conditions. Each spectrum was recorded by averaging
TiO2 (All Values in cm-1
)
mode
gas-phase
adsorbed
TCE on TiO2
adsorbed
descriptiona
TCEb
TCE on ZnOb
b
2ν(CdC)
ν(CH)
3168
3096
3158
3067
1602
1587
1255
930
3085
1602
1586
1254
935
845
782
2ν(C-Cl)
ν(CdC)
δ(CHCl)
ν(C-Cl)
ν(C-Cl)
ν(C-Cl)
1575
1254
940
848
783
1000 scans at an instrument resolution of 4 cm-1
. Each
absorbance spectrum shown represents a single beam scan
referenced to the appropriate single beam scan of the clean
photocatalyst or the blank grid prior to gas introduction, unless
otherwise noted.
n.o.c
n.o.c
a Mode description taken from ref 18. b This work. c n.o. ) not
observed, below absorptions of TiO2.
A 500 W mercury lamp (Oriel Corp.) with a water filter was
used as the light source in these experiments. The 300 nm long
pass filter (% T ) 0 at 300 nm) was placed in front of the lamp.
The broadband light was then reflected off of an aluminum-
coated mirror and then turned by a 1 in. quartz prism onto the
sample. The quartz prism is mounted inside of the FT-IR
sample compartment so that the dry air purge was not broken
during irradiation. The power at the sample was measured
before each experiment and was typically 190 mW/cm2. The
temperature of the sample did not exceed 325 K during these
experiments.
TCE (Aldrich, 99+%) and DCAC (Aldrich, 99%) were
transferred to a glass sample bulb and subjected to several
freeze-pump-thaw cycles before use. Hydrogen (Air Products,
Research Grade) and oxygen (Air Products, 99.6%) were used
as received. Gas pressures were initially measured in a volume
of 823 mL and then expanded into the infrared cell, which is
an additional 320 mL in volume. A valve between these two
portions of the vacuum chamber was then closed before
irradiation so that the total amount of reactants available in the
infrared cell are the pressures specified in the text contained in
the 320 mL volume of the infrared cell.
introduction of gas-phase TCE into the IR cell, bands due to
adsorbed TCE are observed near 3158, 3067, 1602, 1587, 1255,
and 930 cm-1. The intensities of these bands increase with
increasing TCE pressure until the TiO2 surface becomes
saturated at equilibrium pressures greater than 1.4 Torr (see inset
of Figure 1). The vibrational bands of adsorbed TCE can be
assigned by comparison with the vibrational spectrum of gas-
phase TCE (Table 1). A shift of 12 cm-1 to higher frequencies
is observed for the CdC stretching mode, along with shifts to
lower frequencies of the C-H and C-Cl stretching modes.
Hydrogen bonding is observed as a loss in intensity of the ν-
(OH) band near 3715 and 3672 cm-1 for free hydroxyl groups
on the TiO2 surface with a concomitant growth in the associated
hydroxyl stretching region near 3598 cm-1. There are probably
several sites for adsorption of TCE on the TiO2 surface. The
IR data show that there is some interaction between TCE and
the OH groups on the TiO2 surface.
Infrared data obtained for the adsorption of TCE on ZnO
show much less TCE adsorption on ZnO compared to TiO2.
This in part can be attributed to the substantial difference in
the surface areas of ZnO and TiO2 (the ZnO particles used in
these studies have an estimated surface area of approximately
5 m2/g whereas Degussa P25 TiO2 has a surface area near 50
m2/g) and in part to a weaker interaction of TCE with ZnO. At
high pressures (greater than 1.4 Torr) where TCE adsorption
saturates on the surface of TiO2, small infrared absorptions due
Results
TCE Adsorption on TiO2 and ZnO at Room Temperature.
The infrared spectrum of adsorbed TCE as a function of TCE
pressure on TiO2 is shown in Figure 1. Gas-phase contributions
to the spectra shown in Figure 1 have been subtracted out. After

TCE Photooxidation on TiO2 and ZnO
J. Phys. Chem. B, Vol. 102, No. 3, 1998 551
TABLE 2: Vibrational Assignment of Gas-Phase Products
from TCE Photooxidation on TiO2 and ZnO
a
obsd freq (cm-1
)
assigned species
mode description
e
3001
C2HCl5
ν(C-H)
ν(H-Cl)
ν(CdO)
ν(CO)
2885 (envelope center) HCl
2349
2144
1827
1794
1677
1220
1220
1076
1013
992
CO2
CO
COCl2
b
ν(CdO)
ν(CdO)
2ν(C-Cl2)
δ(CH)
Cl2HCCOCl (DCAC)c
b
COCl2
e
C2HCl5
CHCl3,dC2HCl5
δ(CH)
e
Cl2HCCOCl (DCAC)c
ν(C-C)
ν(C-C)
ν(C-C)
ν(C-Cl2)
νas(CCl3)
e
C2HCl5
Cl2HCCOCl (DCAC)c
b
853
776
COCl2
CHCl3,dC2HCl5
d
a This work. b Reference 19. c Reference 20. d Reference 9. e Refer-
ence 21.
Figure 2. Infrared spectra recorded of the gas phase during the
photooxidation of TCE. The spectra labeled a and b show gas-phase
absorptions following irradiation of 144 mTorr of TCE and 721 mTorr
of O2 on clean ZnO and TiO2, respectively. The spectra labeled c and
d show gas-phase absorptions following irradiation of 2.9 Torr of TCE
and 7.2 Torr of O2 on ZnO and TiO2, respectively. These spectra show
the product distribution of the photooxidation of TCE on both clean
ZnO and TiO2 is dependent upon the reaction conditions employed.
to adsorbed TCE on ZnO are observed. The frequencies of
these absorptions are given in Table 1.
Photooxidation of TCE on Clean TiO2 and ZnO Surfaces.
Figure 2 displays the infrared spectra of the gas-phase photo-
oxidation products following irradiation of two different reaction
mixtures of TCE and O2 on both ZnO and TiO2. The gas-phase
infrared spectra taken following photooxidation of TCE on ZnO
and TiO2, starting with 144 mTorr of TCE and 721 mTorr of
O2, are shown in parts a and b, respectively. Gas-phase infrared
absorption bands due to photooxidation products with frequen-
cies near 2147 and 2349 cm-1 are observed and assigned to
gas-phase CO and CO2, respectively. Under these conditions,
complete conversion is achieved in 2 min on TiO2 and 20.5 h
on ZnO.
The infrared spectra shown in parts c and d of Figure 2 were
recorded following the photooxidation of TCE on ZnO and TiO2,
respectively, starting with 2.9 Torr of TCE and 7.2 Torr of O2.
Several infrared absorption bands due to gas-phase species are
observed near 3001, 2349, 2144, 1827, 1794, 1677, 1413, 1220,
1076, 1013, 992, 853, and 776 cm-1. From IR standards and
literature data,9,19-21 these infrared absorption bands can be
assigned to a combination of the following: COCl2, DCAC,
CO, CO2, CHCl3, and C2HCl5 (see Table 2) with the exception
of the band at 1413 cm-1, which is left unassigned. CHCl3
and C2HCl5 have some spectral overlap, and therefore, some
of the same bands (1220 and 776 cm-1) are assigned to both of
these molecules. Additional absorptions are present in the
infrared spectra due to unreacted parent TCE (85% conversion
of TCE on TiO2 after 3 h of irradiation and 40% conversion on
ZnO after 21.5 h of irradiation). These data illustrate how the
reactant pressures can influence the product distribution for this
reaction. They also illustrate that similar processes occur on
ZnO and TiO2.
Figure 3. Plot of the normalized integrated area of the infrared band
for gas-phase CO2 following irradiation of TCE on TiO2 (left) and ZnO
(right) as a function of wavelength. Broadband filters were used in
these experiments, and the data points plotted are at the cutoff
wavelength for each filter used.
10 min. The next longest wavelength filter was then put in
place, and irradiation was continued for an additional 10 min.
This procedure was followed for all five filters used in this
experiment. The increase in integrated area of the CO2 infrared
band after irradiation with each filter was then corrected for
differences in light output between filters. The TiO2 wavelength
dependence was determined in a similar manner. As can be
seen, the onset wavelength of CO2 production coincides with
the bandgap of TiO2 and ZnO, and at shorter wavelengths, above
the bandgap, CO2 production increases even more. The
wavelength dependence indicates that the photooxidation of TCE
is initiated by the excitation of the photocatalyst and not by
direct absorption of TCE, which absorbs UV light with λ <
250 nm.17
Photooxidation of TCE on Adsorbate-Covered TiO2
Surfaces. The infrared spectra shown in Figure 2 provide clear
evidence that gas-phase COCl2, DCAC, and CHCl3 do not form
under conditions of low TCE and O2 pressure on clean ZnO
and TiO2 samples; however, chlorinated partial oxidation
products do form at low pressures of TCE and O2 on used TiO2
samples. Figure 4 displays the gas-phase infrared spectra
following the successive irradiation of 144 mTorr of TCE in
the presence of 721 mTorr of oxygen on TiO2 for 5 min. Each
successive run was performed without cleaning the TiO2 sample
in between; however, the IR cell was evacuated, and a fresh
reactant mixture (TCE and O2) was introduced before each run.
Initially, the only gas-phase product absorptions are found near
2147 and 2349 cm-1 (Figure 4a), corresponding to the formation
of CO and CO2, respectively. Other infrared bands that are
The wavelength dependence of the photooxidation of TCE
was investigated on TiO2 and ZnO. Figure 3 displays a plot of
integrated area of the infrared band for gas-phase CO2 as a
function of the wavelengths used during irradiation. Starting
with the longest wavelength first, the ZnO was irradiated for

552 J. Phys. Chem. B, Vol. 102, No. 3, 1998
Driessen et al.
Figure 6. Integrated absorbance of infrared bands for the specified
gas-phase products from the photooxidation of 2.9 Torr of TCE and
7.2 Torr of O2 on TiO2 as a function of irradiation time. The inset
displays an enlarged view of the first 20 min of irradiation.
Figure 4. Infrared spectra of the gas phase during the photooxidation
of 144 mTorr of TCE and 721 mTorr of O2. Four successive
experiments were performed on the same TiO2 sample. Initially, for
run 1, CO and CO2 are the only two gas-phase products. Other products
are formed in successive runs, runs 2-4. The infrared absorption bands
can be assigned to the following gas-phase products: CO2, CO, Cl2-
CO, DCAC, CHCl3, C2HCl5, and HCl (see Table 2). The inset shows
that the amount of phosgene produced after each run eventually levels
off.
(shoulder), 1419, 1375 (shoulder), and 1229 cm-1
. These
spectra show that the titania surface becomes nearly saturated
with adsorbed species by the third run as there is almost no
change in the infrared bands associated with the surface-bound
species between runs 3 and 4 (Figure 5, c and d). There is
clearly a correlation between the infrared data for gas-phase
and surface-bound products. The infrared data show that COCl2,
DCAC, C2HCl5, CHCl3, and HCl are produced in increasing
amounts as the TiO2 surface becomes more saturated with
adsorbed species. The production of phosgene is seen to reach
a plateau after three runs (see inset Figure 4) where the TiO2
surface has become saturated with adsorbed species.
A consequence of the accumulation of surface-bound products
on the TiO2 surface is that the amount of TCE that adsorbs on
the TiO2 surface decreases before each run. This is illustrated
in the left panel of Figure 5. The infrared spectra of adsorbed
TCE prior to runs 1-4 are shown. The spectra show that, after
the initial photooxidation of 144 mTorr of TCE; there is a
substantial decrease in the amount of adsorbed TCE; the
integrated area of the TCE absorption band in the C-H
stretching region shows that less than 10% of the initial amount
of adsorbed TCE can adsorb (Figure 5, left panel). Following
run 2 (Figure 5c), less than 1% of the initial amount of TCE
can adsorb due to the buildup of adsorbed photoproducts on
the catalyst surface. It is clear from the data presented in Figure
5 that adsorbed photoproducts block surface sites so that the
amount of adsorbed TCE decreases following each run.
The influence of adsorbed species on the TiO2 surface on
the gas-phase product distribution is also evident in the higher
pressure reaction. Figure 6 shows the integrated area of the
infrared bands due to several gas-phase products (CO2, CO,
COCl2, and CHCl3) formed during the photooxidation of 2.9
Torr of TCE with 7.2 Torr O2 on initially clean TiO2 plotted as
a function of irradiation time. The inset shows an expanded
view of the irradiation time between 0 and 20 min. The plot
shows that the product distribution changes over time. The inset
shows that only CO and CO2 are formed in the first 5 min of
reaction. At later times phosgene and chloroform are formed
as well. The induction period for chlorinated partial-oxidation
products can be correlated with the buildup of adsorbed surface
products from the photooxidation of TCE; i.e., only after some
critical surface coverage of adsorbed products is reached are
chlorinated products formed.
Figure 5. Right panel: infrared spectra of the adsorbed photoproducts
on TiO2 following each of the successive runs shown in Figure 4 (runs
1-4). Left panel: infrared spectra of adsorbed TCE prior to each of
the successive runs shown in Figure 4 (runs 1-4). The amount of TCE
that can adsorb prior to each run decreases as the coverage of adsorbed
products increases.
seen in the spectrum (e.g., at 942 cm-1) are due to unreacted
parent TCE. After 5 min of irradiation, approximately 85% of
the TCE has been consumed in each of the runs. Successive
runs (runs 2-4) show the appearance and growth of several
new infrared absorptions. A progression of bands near 2885
cm-1 becomes apparent after runs 3 and 4. These bands
correspond to the rovibrational spectrum of HCl. Other
absorption bands near 2349, 2144, 1827, 1794, 1677, 1413,
1220, 1076, 1013, 992, 853, and 776 cm-1 are also observed in
runs 2-4. As before, it is possible to identify these gas-phase
products from their infrared absorptions as CO2, CO, COCl2,
DCAC, CHCl3, C2HCl5, and HCl (Table 2).
Infrared spectra of the TiO2 surface were also recorded after
each of the four runs. These spectra are shown in the right
panel of Figure 5. After run 1, infrared absorptions due to
surface-bound species are observed near 1655, 1575, 1456
The spectrum recorded of the TiO2 surface after photooxi-

TCE Photooxidation on TiO2 and ZnO
J. Phys. Chem. B, Vol. 102, No. 3, 1998 553
Figure 8. Plot of the integrated area of the infrared absorptions for
gas-phase products formed from the photooxidation of TCE on TiO2
as a function of initial TCE pressure.
Figure 7. (a) Infrared spectrum of the TiO2 photocatalyst following
photooxidation of 2.9 Torr of TCE in the presence of 7.2 Torr of O2.
(b) Infrared spectrum of TiO2 following adsorption of DCAC. The
frequencies of the bands that are underlined are assigned to adsorbed
dichloroacetate. See text for further discussion.
adsorbed carbonyl on the TiO2 surface, respectively. The
adsorbed carbonyl is most likely adsorbed phosgene.9 The main
bands in the photoproduct spectrum have shoulders, this may
be due to the presence of other adsorbates of the form COx,
TABLE 3: Vibrational Assignment of Adsorbed
-
including CO2-, HCO2-, CO32-, and HCO3
.
Photoproduct Dichloroacetate (Cl2HCCO2) from TCE
Photooxidation on TiO2 (All Values in cm-1
)
TCE Photooxidation as a Function of O2 and TCE
Pressure. The data presented in Figure 1 show that the product
distribution of the photooxidation reaction of TCE changes as
a function of the partial pressures of the reaction mixture (TCE
and O2). To further investigate this pressure dependence,
experiments were performed as a function of TCE and O2
pressure. Figure 8 shows a plot of the integrated area of the
photooxidation products of TCE versus initial pressure of TCE,
in the presence of 7.2 Torr of molecular oxygen. In each case,
7.2 Torr of O2 is more than the stoichiometric amount needed
for complete oxidization of TCE, with the exception of the 2.9
Torr of TCE in which case it is exactly the required pressure to
fully oxidize the TCE present. The data shown were collected
from the same titania sample that was cleaned between
experiments according to the procedure outlined in the Experi-
mental Section. In each case more than 85% of the initial TCE
was converted to products. The distribution of products from
TCE photooxidation is strongly dependent upon the initial
amount of TCE present. It is only at the higher pressures of
TCE, at 721 mTorr and above, that detectable amounts of
phosgene, DCAC, HCl (not plotted), and CHCl3 are produced.
Figure 9 displays a plot of the ratio of the integrated areas of
the infrared absorptions for two gas-phase photooxidation
products of TCE photooxidation, CHCl3 and CO, plotted as a
function of oxygen pressure used. Each experiment was done
on an adsorbate saturated sample using 144 mTorr of TCE. This
plot illustrates that the product distribution of the photooxidation
of TCE is also highly dependent upon oxygen pressure once
the surface is saturated with adsorbed species.
a
dichloroacetate-TiO2
mode assignment
3011
1581
1418
1228
ν(C-H)
νas(COO)
νs(COO)
δ(CHCl2)
a From adsorption of dichloroacetyl chloride on TiO2. b Mode
assignments are based on comparison to literature spectra of carboxy-
lates (see ref 23).
dation of TCE at high pressures (2.9 Torr of TCE with 7.2 Torr
O2) is shown in Figure 7a. There are several bands apparent
in the spectrum due to adsorbed photoproducts. The assignment
of the surface-bound products on the TiO2 surface is more
complicated than the identification of gas-phase products. The
frequencies of the infrared bands observed between 1000 and
1800 cm-1 are in the region where carbonate, carboxylate, and
water absorptions occur.22-25 Each band in the spectrum is quite
broad and may be comprised of several absorption bands. It
has been recently determined from NMR studies that the
predominant carbon-containing, surface-bound product from
TCE photooxidation on TiO2 is dichloroacetate.26 It was
postulated that dichloroacetate formed from the reaction of
dichloroacetyl chloride and hydroxyl groups to yield HCl and
dichloroacetate. The IR spectrum of the TiO2 surface following
TCE photooxidation at high pressures is consistent with an
adsorbed carboxylate.23 To confirm that dichloroacetate is a
surface-bound photoproduct, the surface chemistry of DCAC
on TiO2 was investigated.
The infrared spectrum recorded following adsorption of
DCAC on a clean TiO2 surface is shown in Figure 7b. DCAC
was introduced into the IR cell at a pressure of 0.500 Torr and
then evacuated. The two spectra shown in Figure 7sthe
photoproduct spectrum and the DCAC spectrumsare very
similar and show that dichloroacetate is a surface-bound product
formed during TCE photooxidation. Absorptions near 3016,
1580, 1420, and 1229 cm-1 are associated with dichloroacetate.
Additional bands at 1650 and 1740 cm-1 present in the
photoproduct spectrum shown in Figure 7a are assigned to the
bending mode of adsorbed water and the CdO stretch of an
Discussion
From the data presented here, it has been shown that, in the
presence of molecular oxygen, TiO2 and ZnO are effective in
the photooxidation of TCE using light of wavelengths 390 nm
> λ > 300 nm. Differences in the rate of conversion of TCE
on TiO2 relative to ZnO are in part related to differences in the
surface area of the two samples used (50 m2/g vs 5 m2/g),
although, a full study of the differences in the kinetics of TCE
photooxidation on TiO2 and ZnO has not been done. The TiO2

554 J. Phys. Chem. B, Vol. 102, No. 3, 1998
Driessen et al.
SCHEME 1: Proposed TCE Photooxidation Mechanism
on an Adsorbated Covered Surface
Figure 9. Plot of the ratio of the integrated area of the infrared
absorption bands of gas-phase CHCl3 and CO formed during the
photooxidation of TCE on TiO2 as a function of oxygen pressure. The
amount of CO increases relative to CHCl3 as the oxygen pressure
increases.
not well understood, requires available surface sites for the
complete photooxidation of TCE.
and ZnO data show that the reaction conditions influence
product distribution in this reaction. Factors that influence TCE
photooxidation are discussed in detail below. Based on the data
presented here, a reaction mechanism for TCE photooxidation
is given.
hν
C2HCl3(a) + O2(a)
8 C2HCl3(a) + O*(a) surface8 reactions8
TiO2
ClH2CCO2(a) + COx(a) + H2O(a) + CO(g) + CO2(g) (2)
Proposed Reaction Mechanisms for TCE Photooxidation
on TiO2 and ZnO. As discussed in the Introduction section,
there have been several studies of the gas-phase photooxidation
of TCE on TiO2.3-13 A number of different mechanisms and
reactive species have been proposed for this reaction. Nimlos
et al. propose that because several of the photooxidation products
detected contained more Cl atoms than the parent, Cl atoms
must be responsible for the chain propagation of the photooxi-
dation of TCE.12 Nimlos et al.6 have raised the question as to
whether some of the reactions that occur during TCE photo-
oxidation over TiO2 occur in the gas phase or on the TiO2
surface itself. Fan and Yates suggest that an excited adsorbed
oxygen species forms and reacts directly with adsorbed TCE
to form DCAC. DCAC can then undergo continued photooxi-
dation to yield COCl2.9 Phillips and Raupp propose a mech-
anism that includes hydroxyl and hydroperoxide radicals as the
chain propagators.8 Hydroxyl radicals as the primary reactive
species have also been discussed by Anderson and co-workers.13
It is clear from the above discussion that there are several
different proposals for the identity of the reactive species which
initiates the photooxidation of TCE. Here we propose a
mechanism that encompasses all of these as contributing to the
photooxidation of TCE over TiO2. It is proposed that the TCE
photooxidation mechanism depends on the coverage of adsorbed
products and the availability of surface sites and hydroxyl
groups.
Since the initial conditions used in these experiments are
similar to those used by Fan and Yates, it is proposed that the
initial step in the photooxidation is the excitation of adsorbed
oxygen, as they do. The identity of the reactive oxygen species
is not known, but for purposes of this discussion, the reactive
oxygen species will be represented as O* without specifying
molecular or atomic oxygen. The next step involves reaction
O* with adsorbed TCE. On a clean surface, surface reactions
involving O* and adsorbed chlorinated partial oxidation products
continue until the surface is covered with adsorbed products.
CO2 and CO are the only two gas-phase, carbon-containing
products formed. This reaction sequence (reaction 2), although
As the surface becomes covered with adsorbed photoproducts,
the reaction probability of continued surface reactions decreases.
Although the TiO2 surface is covered with adsorbed photoprod-
ucts, the photooxidation of TCE still proceeds, but the products
are different; they are in fact the same products observed in the
homogeneous gas-phase photooxidation of TCE.27-29
On an adsorbate-covered TiO2 surface, it is proposed that
the initial photooxidation step occurs on the surface, but
subsequent chemistry differs. Initially, O* reacts with TCE to
form [C2HCl3O]*, an excited adduct (reaction 3).
hν
C2HCl3(a) + O2(a)
8 C2HCl3(a) + O*(a) f
TiO2, adsorbate covered
[C2HCl3O]*(a) f CO(g) + COCl2(g) + CO2(g) + CHCl3(g)
+
C2HCl5(g) + HCl(g) + CHCl2CClO(g) (3)
C2HCl3O* has been postulated to form in the homogeneous gas-
phase photooxidation of TCE with O(3P) and is most likely a
diradical with the oxygen bonded to the more positive end of
the TCE molecule, making the C-C bond predominantly σ in
nature.27 C2HCl3O* is unstable and breaks down to yield
reactive species similar to that which forms in the gas phase. A
mechanism that accounts for all of the observed species is shown
in Scheme 1.27-29 The reaction mechanism shown in Scheme
1 is essentially the same as the mechanism proposed for the
homogeneous gas-phase photooxidation of TCE.27-29 One
additional reaction that has been added to Scheme 1 involves
the photooxidation of CO on TiO2 to form gas-phase CO2. The
gas-phase photoproducts detected in this study are marked with
a box.
Chlorine atoms are produced in this reaction mechanism. Cl
atoms can undergo reaction with TCE as discussed previously
by Nimlos et al.9 Although Nimlos et al. have suggested that
Cl atoms are the active species in the photooxidation of TCE
on TiO2, it is not clear how they may be produced to initiate
the reaction. Here it is suggested that chlorine atoms are formed
through the decomposition of [C2HCl3O]* (Scheme 1) and then

TCE Photooxidation on TiO2 and ZnO
J. Phys. Chem. B, Vol. 102, No. 3, 1998 555
can initiate further photooxidation of TCE, as is observed in
gas-phase reactions. The products for the gas-phase, chlorine-
initiated oxidation of TCE are DCAC, COCl2, CHCl3, and CO,30
similar to the products observed in this study for the photooxi-
dation of TCE on TiO2 at high pressures. It is also clear from
Scheme 1 why quantum yields greater than one have been
measured for this reaction, as atomic chlorine is produced in
several steps and chlorine atom initiated oxidation is a radical
mechanism. Experiments done to trap chlorine atoms and other
radical species in the gas phase were unsuccessful, suggesting
that the radical mechanism occurs on the surface.31
Side reactions of DCAC and phosgene also occur in addition
to the reactions shown in Scheme 1. Fan and Yates have
demonstrated that DCAC is a reaction intermediate that can
undergo continued photooxidation on TiO2 to yield COCl2, CO,
and HCl (reaction 4).9
oxidation temperature-programmed desorption (TPD), they
identified changes in the surface chemistry between fresh TiO2
catalysts and used or completely deactivated TiO2 catalysts.
They found that a greater amount of CO and CO2 desorbed
from used catalysts during TPD than from fresh TiO2. They
postulated that this was due to the decomposition of larger
adsorbates formed from the photooxidation of TCE. It was also
proposed that DCAC was an intermediate in the photooxidation
of TCE, and therefore the effect of DCAC adsorption on TiO2
samples was also investigated. Larson and Falconer found that
the decomposition of DCAC on the TiO2 surface decreased the
number of adsorption sites and caused the loss in activity of
the TiO2 photocatalyst.32 The infrared data presented here show
that a stable product, dichloroacetate, forms on the surface from
reaction of DCAC with TiO2.
There is a product distribution dependence on the TCE
pressure. This dependence correlates with the buildup of
surface-bound products from the photooxidation of TCE. The
higher the initial pressure of TCE, the higher the ratio of gas-
phase chlorinated partial-oxidation products relative to CO2 (see
Figure 8). The data in Figure 9 show that the photooxidation
of TCE on TiO2 also has a strong dependence on O2 pressure.
At low pressures of oxygen, there is a larger probability that
the CCl2 radical and HCl will react to form CHCl3. At higher
oxygen pressures, CCl2 has a higher probability of reacting with
a molecule of oxygen to go on to form COCl2, CO, and HCl,
as shown in Scheme 1.
TiO2
Cl2HCCOCl + 1/2O2
8 COCl2 + HCl + CO (4)
hν
It has also been established that phosgene reacts with TiO2
surfaces in the dark to undergo hydrolysis to yield HCl and
CO2. The surface hydrolysis reaction of phosgene to form CO2
is therefore dependent on the amount of adsorbed water or
hydroxyl groups present on the TiO2 surface.
The role of hydroxyl groups on the TiO2 surface in the
hydrolysis of phosgene was investigated here. A TiO2 sample
that had been previously used as a photocatalyst was cleaned
of surface-bound products as described in the Experimental
Section. The infrared spectrum of the sample showed that
approximately 50% of the surface hydroxyl groups had reacted.
This dehydroxylated sample was then used to photooxidize 144
mTorr of TCE in the presence of O2. The infrared spectrum of
the TiO2 surface following the photooxidation reaction showed
a band due to adsorbed phosgene at 1740 cm-1. This same
photooxidation was performed on a TiO2 surface that contained
a significant portion of hydroxyl groups, and neither adsorbed
(1740 cm-1) nor gas-phase (1827 cm-1) phosgene was observed.
The formation of adsorbed phosgene before saturation of the
TiO2 surface with adsorbed products supports a surface-mediated
hydrolysis reaction with adsorbed OH groups on the surface of
clean TiO2.
The participation of hydroxyl groups does not imply that
hydroxyl radicals are the active species that initiate the photo-
oxidation of gas-phase TCE under the reaction conditions used
in this study as Phillips and Raupp8 and also Anderson and co-
workers13 have proposed. Fan and Yates have shown that, under
similar reaction conditions as the ones used here, adsorbed H2O
and hydroxyl groups are not the active species that initiate the
photooxidation of TCE but are involved in further reactions of
adsorbed species.9
Conclusions
The data presented here show that the photooxidation of TCE
on ZnO and TiO2 is a complicated process. The product
distribution can be influenced by several factors including the
initial pressures of TCE and molecular oxygen and the adsorbate
coverage on the photocatalyst surface. When certain reaction
conditions are employed, in particular, low pressures of TCE
and clean photocatalyst surfaces, CO and CO2 are produced as
the only two carbon-containing gas-phase products. Additional
gas-phase products form at higher initial pressures of TCE and
an adsorbate covered photocatalyst surfaces (dichloroacetate and
adsorbed water). Two photooxidation mechanisms are proposed
to explain these observations. One mechanism involves a
multistep surface reaction sequence that can occur on a clean
photocatalyst surface to produce predominantly two gas-phase
products CO and CO2. The second mechanism involves a
complex series of reactions that follows the mechanism for the
homogeneous gas-phase photooxidation of TCE. This second
mechanism results in the formation of gas-phase COCl2, DCAC,
CHCl3, C2HCl5, HCl, CO, and CO2 as well as chlorine atoms
that can further initiate photooxidation of TCE.
Influence of Adsorbed Products and Initial Partial Pres-
sures of TCE and O2. The data presented here clearly
demonstrate that the product distribution of TCE photooxidation
is dependent upon the surface coverage of adsorbed products.
The data herein support the postulated mechanisms showing
that the photooxidation occurs initially on the TiO2 surface to
produce only gas-phase CO and CO2. As the photooxidation
proceeds, active sites on the surface of the TiO2 become covered
with adsorbed photooxidation products. The dominant reaction
pathway changes from one that occurs on the TiO2 surface to
produce only gas-phase CO and CO2 as the primary carbon-
containing photooxidation products to one that produces several
chlorinated products in addition to gas-phase CO and CO2.
Changes in the TiO2 surface due to TCE photooxidation have
been observed by Larson and Falconer.32 Using post-photo-
Acknowledgment. The authors gratefully acknowledge the
National Science Foundation (Grant CHE-9614134) for support
of this work.
References and Notes
(1) Dyksen, J. E.; Hess, A. F., III J. Am. Water Works Assoc. 1982,
74, 394.
(2) Kenneke, J. F.; Ferry, J. L.; Glaze, W. H. Photocatalytic Purification
and Treatment of Water and Air; Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier
Science Publishers: Amsterdam, 1993; p 179.
(3) Yamazaki-Nishida, S.; Nagano, K. J.; Phillips, L. A.; Cervera-
March, S.; Anderson, M. A. J. Photochem. Photobiol. A: Chem. 1993, 70,
95.
(4) Dibble, L. A.; Raupp, G. B. Catal. Lett. 1990, 4, 345.
(5) Pruden, A. L.; Ollis, D. F. J. Catal. 1983, 82, 404.

556 J. Phys. Chem. B, Vol. 102, No. 3, 1998
Driessen et al.
(6) Nimlos, M. R.; Jacoby, W. A.; Blake, D. M.; Milne, T. A. EnViron.
Sci. Technol. 1993, 27, 732.
(7) Jacoby, W. A.; Nimlos, M. R.; Blake, D. M.; Noble, R. D.; Koval,
C. A. EnViron. Sci. Technol. 1994, 28, 1661.
(8) Phillips, L. A.; Raupp, G. B. J. Mol. Catal. 1992, 77, 297.
(9) Fan, J.; Yates, J. T., Jr. J. Am. Chem. Soc. 1996, 118, 4686.
(10) Lichtin, N. N.; Avudaithai, M. EnViron. Sci. Technol. 1996, 30,
2014.
(11) Jardim, W. F.; Alberici, R. M.; Takiyama, M. M. K.; Huang, C. P.
Hazard. Ind. Wastes 1994, 26, 230.
(12) Nimlos, M. R.; Jacoby, W. A.; Blake, D. M.; Milne, T. A.
Photocatalytic Purification and Treatment of Water and Air; Ollis, D. F.,
Al-Ekabi, H., Eds.; Elsevier Science Publishers: Amsterdam, 1993; p 387.
(13) Yamazaki-Nishida, S.; Cervera-March, S.; Nagano, K. J.; Anderson,
M.; Hori, K. J. Phys. Chem. 1995, 99, 15814.
(14) Basu, P.; Ballinger, T. H.; Yates, J. T., Jr. ReV. Sci. Instrum. 1988,
59, 1321.
(15) (a) Miller, T. M.; Grassian, V. H. J. Am. Chem. Soc. 1995, 117,
10969. (b) Driessen, M. D.; Grassian, V. H. J. Catal. 1996, 161, 810.
(16) Tanaka, K.; White, J. J. Phys. Chem. 1982, 86, 4708.
(17) Berry, M. J. J. Chem. Phys. 1974, 61, 3114.
(18) Bernstein, H. J. Can. J. Res. 1949, 28B, 132.
(19) Hannus, I.; Kiricsi, I.; Tasi, G.; Fejes, P. Appl. Catal. 1990, 66,
L7.
(20) Miyake, A.; Nakagawa, I.; Miyazawa, T.; Ichishima, I.; Shiman-
ouchi, T.; Misushima, S. Spectrochim. Acta 1958, 13, 161.
(21) NIST Standard Reference Database 69, Aug 1997.
(22) Little, L. H. Infrared Spectra of Adsorbed Species; Academic: New
York, 1966.
(23) Davydov, A. A. Infrared Spectroscopy of Adsorbed Species on the
Surface of Transition Metal Oxides; Wiley and Sons: Chichester, 1990.
(24) Saussy, J.; Lavalley, J.-C.; Bovet, C. J. Chem. Soc., Faraday Trans.
1 1982, 78, 1457.
(25) Bocuzzi, F.; Ghiotti, G.; Chiorino, A. Surf. Sci. 1985, 162, 361.
(26) Hwang, S.-J.; Petucci, C.; Raftery, C. J. Am. Chem. Soc. 1997,
119, 7877.
(27) Sanhueza, E.; Hisatsune, I. C.; Heicklen, J. Chem. ReV. 1976, 76,
801.
(28) Sanhueza, E.; Heicklen, J. Int. J. Chem. Kinet. 1974, 6, 553.
(29) Bertrand, L.; Franklin, J. A.; Goldfinger, P.; Huybrechts, G. J. Phys.
Chem. 1968, 72, 3926.
(30) Huybrechts, G.; Meyers, L. Trans. Faraday Soc. 1966, 62, 2191.
(31) Several experiments were done with C2H4 and C2H6 added to the
gas mixture in order to trap chlorine atoms in the gas phase.
(32) Larson, S. A.; Falconer, J. L. Photocatalytic Purification and
Treatment of Water and Air, Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier
Science Publishers: Amsterdam, 1993; p 473.