Environ. Sci. Technol.
2007, 41, 1696-1703
Trichloroethene Degradation by
UV/H2O2 Advanced Oxidation
Process: Product Study and Kinetic
Modeling
K E L I , †,| M I H A E L A I . S T E F A N , * ,‡,⊥ A N D
JOHN C. CRITTENDEN§
Department of Civil and Environmental Engineering,
Michigan Technical University, Houghton, Michigan 49931,
Department of Chemistry, The University of Western Ontario,
London, Ontario, Canada N6A 5B7, and Department of Civil
and Environmental Engineering, Arizona State University,
P.O. Box 872803, Tempe, Arizona 85287-2803
The broadband UV irradiation of 1.1 mM trichloroethene
(TCE) aqueous solution in the presence of 10.4 mM H2O2
resulted in formic, oxalic, dichloroacetic (DCA), and
monochloroacetic (MCA) acids, as organic byproducts.
The organic chlorine was converted completely to chloride
ion as a final product. TCE and its degradation products
were completely mineralized in 30 min, under a volume-
averaged UV-C irradiant power of 35.7 W/L from a 1
kW medium-pressure mercury vapor arc lamp. TCE degraded
primarily through hydroxyl radical-induced reactions and
only to a low extent through direct UV photolysis and chlorine
atom-induced chain reactions. The experimental patterns
of TCE, H2O2, and detected reaction products combined with
the literature information on radical reactions in the
aqueous phase were used to postulate a degradation
mechanism and to develop a kinetic model to predict the
TCE decay, formation and degradation of byproducts,
and pH and oxygen profiles. The agreement between the
model calculations and the experimental data is satisfactory.
Introduction
Trichloroethene (TCE) is a volatile organic compound that
has been proven to cause liver damage and kidney failure in
humans and was assessed as carcinogenic to animals. TCE
is produced worldwide in very large quantities due to its
extensive use in the chemical industry. Leakage from storage
tanks and, sometimes, improper disposal practices led to
soil, groundwater, and surface water contamination with TCE.
Consequently, trichloroethene is listed as a priority pollutant
on the U.S. EPA Chemical Contaminant List and is strictly
regulated in drinking water to a maximum contaminant level
of 0.005 mg/L.
The UV/H2O2 process is currently successfully applied as
an effective treatment of trichloroethene in groundwater and
* Corresponding author phone: (519)457-3400; fax: (519)457-3030;
e-mail: mstefan@trojanuv.com.
† Michigan Technical University.
‡ The University of Western Ontario.
§ Arizona State University.
| Current affiliation: Department of Civil and Environmental
Engineering, Arizona State University, P.O. Box 872803, Tempe, AZ
85287-2803.
⊥ Current affiliation: Trojan Technologies, Inc., 3020 Gore Road,
London, ON, Canada N5V 4T7.
1696 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 5, 2007
reclaimed water. The treatment process relies primarily on
the •OH radical-induced oxidation of TCE. Trichloroethene
exhibits an absorption band in the wavelength range of 200-
250 nm [max(200 nm))7183 M-1 cm-1 (1)] and undergoes
direct photolysis with an overall quantum yield of φ ) 0.354
(1). Extensive studies have been conducted on TCE degrada-
tion by the UV/H2O2 process (2-8). Most of the published
literature focused on the kinetics of TCE decay in various
water qualities and on the associated treatment cost, and
only a few studies reported the intermediate products formed.
Hirvonen et al. (9) examined the formation of byproducts
during TCE degradation in the UV/H2O2 process applied to
groundwater remediation and detected di- and mono-
chloroacetic acids. Studies on γ radiation-induced degrada-
tion of TCE indicated also the formation of chlorinated acids
as reaction intermediates (10, 11). Haloacetic acids (HAAs)
are formed during the drinking water disinfection with
chlorine and are commonly known as disinfection byproducts
(DBPs), which could cause bladder, colon, and rectal cancer
(12). According to the U.S. EPA regulations for drinking water
the HAA total concentration must not exceed 60 µg/L.
In this work, we investigated the TCE degradation
mechanism in the presence of H2O2 under irradiation with
a 1 kW medium-pressure Hg lamp. Based on the proposed
reaction mechanism and the kinetic parameters available in
the literature, we have developed a kinetic model to predict
the dynamics of this system and found a satisfactory
agreement between the calculated and experimental data.
Experimental Section
Reagents. All chemical reagents were of analytical reagent
grade and were used as received. Catalase from bovine liver
(Sigma, 19 900 units/mg) was used to destroy any residual
hydrogen peroxide in the sample vials prior to IC analysis.
Apparatus. All experiments were carried out in a stainless
steel 1 kW bench-scale Rayox reactor described elsewhere
(13). The total reaction volume was 28 L, and the reactor was
airtight. In the wavelength range of 200∼300 nm, the volume-
averaged photon flow entering the reactor was 7.8 × 10-6
einstein s-1 L-1, as determined by potassium persulfate
actinometry (14). The physical path length of the reactor was
10.4 cm. Additional information on the experimental setup
is available elsewhere (1).
Analytical Methods. The analytical methods for TCE and
other organic acids were reported previously (1). The molar
absorption coefficients of TCE, H2O2, and organic acids are
provided in the Supporting Information. Hydrogen peroxide
was determined according to the triiodide method (15).
The oxygen level was monitored in the holding reservoir
with an ATI Orion Model 840 DO meter.
Results and Discussion
Product Study. Trichloroethene (1.1 mM) in distilled water
was irradiated in the presence of 10.4 mM H2O2. At these
concentrations, the fractions of UV light absorbed by TCE
and H2O2 within 200-300 nm range were 24.9% and 59.5%,
respectively. Under such conditions, TCE degraded through
both oxidative chain reactions initiated by the OH radicals
and chlorine atoms, and direct UV photolysis due to its strong
absorption band below 240 nm and large quantum yield
[φ)0.354 (1)]. TCE decayed by 2 orders of magnitude in less
than 10 min, and fewer byproducts with lower yields were
observed than during the direct UV photolysis alone (1).
Figure 1 shows the time profiles of TCE, H2O2, and degrada-
tion byproducts.
10.1021/es0607638 CCC: $37.00  2007 American Chemical Society
Published on Web 02/02/2007
 The major byproducts were formic and oxalic acids.
Mono- and dichloroacetic acids were detected at low levels.
No chlorinated ethynes and aldehydes were measured, likely
due to their low levels and high reactivity toward hydroxyl
radical. Chloride ion was released as a final product from
early stages of irradiation through the end of the degradation
of chlorinated byproducts, and the organic carbon total
mineralization was achieved within ∼30 min exposure time.
The chlorine mass balance was calculated and indicated that
about 6% chlorine was missing from the system, most likely
due to some TCE volatility losses in the headspace zone of
the reactor.
As the organic and hydrochloric acids were formed, the
solution pH dropped sharply from 5.9 (t)0 min) to 2.6 (t)5
min). Dissolved oxygen level decreased from ∼7.1 mg/L (t)0
min) to 1.7-2.0 mg/L at 4-5 min reaction time and is
associated with the mineralization of approximately 70% of
the organic carbon. After 4 min, the oxygen level increased
continuously up to ∼23 mg/L through the photodecompo-
sition of H2O2 and due to a lower chemical oxygen demand
than at short exposure times [under oxygen-saturated
conditions, maximum O2 level in aqueous solutions could
be as high as 40 mg/L [1.25 × 10-3 M (16)]. The pH and
oxygen patterns over the irradiation time are shown in Figure
2.
Reaction Mechanism for Photolysis of H2O2. The H2O2
photolysis and subsequent thermal reactions of the generated
radicals were extensively studied (17-21). The reactions
relevant to this system and considered in the kinetic model
are as follows:
H2O2 + hv f 2•OH φOH ) 1.0 (17) (1)
H2O2 + •OH f H2O + HO2• k ) 2.7 × 107 M-1 s-1 (22)
(2)
H2O2 + HO2•/O2•- f •OH + H2O/OH- + O2
k ) 3.6 M-1 s-1 (22) (3)
• • •- -
OH + HO2 /O2 f H2O/OH + O2
k ) 1 × 1010 M-1 s-1 (23) (4)
HO2• + HO2•/O2•- f H2O2/HO2- + O2
k ) 1 × 106 M-1 s-1 (20) (5)
At 254 nm, the primary quantum yield (φ) of hydrogen
peroxide photolysis in eq 1 is 0.5. The extended set of the
H2O2 photolysis reactions is incorporated in a complex kinetic
model, which was validated for UV/H2O2 applications in
completely mixed batch-reactors (24).
TCE Decay and Byproduct Formation. A simplified
representation of the proposed mechanism for TCE degra-
dation is given in Figure 3.
The three major routes that describe the TCE degradation
start with the following primary processes: OH radical
addition to the double bond, direct photolysis, and chlorine
atom addition to the double bond.
TCE Degradation through OH Radical-Induced Reactions.
The OH radical addition to the double bond of trichloroethene
yields a geminal chlorohydrin carbon-centered radical:
ClHCdCCl2 + •OH f ClCH(OH)-C•Cl2
k ) 2.4 × 109 M-1 s-1 (25) (6)
There are various rate constant values reported in the
literature for reaction 6: 2.4 × 109 M-1 s-1 (25); 2.9 × 109 M-1
s-1 (26); 3.3 × 109 M-1 s-1 (10); 4.0 × 109 M-1 s-1 and 4.3 ×
109 M-1 s-1 (27); and 2.6 × 109 M-1 s-1 (28). In our kinetic
model, a rate constant of 2.4 × 109 M-1 s-1 (25) was considered.
The geminal chlorohydrins eliminate HCl very rapidly
with first-order rate constants k > 7 × 105 s-1 (27, 29); a
similar reaction occurs from chlorohydrin radicals:
ClCH(OH)-C•Cl2 f OHC-C•Cl2 + H+ + Cl-
k ) 5.1 × 105 M-1 s-1 (30) (7)
The reactions between the carbon-centered radicals and
O2 are diffusion-controlled processes (k>109 M-1 s-1) forming
peroxyl radicals. The peroxyl radical generated in reaction
8 evolves primarily into oxyl radicals and molecular oxygen
(10, 11).
OHC-C•Cl2 + O2 f OHC-C(Cl2)OO•
k ) 3 × 109 M-1 s-1 (31) (8)
2OHC-C(Cl2)OO• f 2OHC-C(Cl2)O• + O2
k ) 4 × 108 M-1 s-1 (31) (9)
The oxyl radicals decay through fast unimolecular frag-
mentation either by a C-C bond scission or by elimination
of a chlorine atom with formation of molecular products,
which undergo rapid hydrolysis.
OHC-C(Cl2)O• f COCl2 + •CHO
k ) 1 × 106 M-1 s-1 (31) (10a)
f OHC-C(O)Cl + Cl• k ) 1 × 105 M-1 s-1 (fitted)
(10b)
Phosgene hydrolysis occurs with a pseudo-first-order rate
constant of 6 s-1 (32) - 9 s-1 (16), with formation of CO2 and
hydrochloric acid.
COCl2 + H2O f CO2 + 2H+ + 2Cl- (11)
The unstable chloroglyoxal OHC-C(O)Cl hydrolyzes to
glyoxylic acid (reaction 12), whereas the formyl radical •CHO,
after hydration and reaction with O2, generates formic acid
through a fast HO2• elimination reaction [k>1 × 106 s-1 (33)].
The formyl radical can also be oxidized to CO and HO2• (10)
[k)4 × 105 M-1 s-1 (fitted)].
OHC-C(O)Cl + H2O f OHCCOOH + H+ + Cl-
(12)
O2
•
CHO + H2O 98 HCOOH + O2•- + H+
k ) 1 × 106 M-1 s-1 (33) (13)
Therefore, the •OH-induced degradation of trichloro-
ethene results in formic and glyoxylic acids as intermediates
and chloride ion and CO2 as final products.
TCE Direct Photolysis. The fraction of incident light (200-
300 nm) absorbed by TCE at t ) 0 min is 24.9%, as compared
to 42.6% for the same initial concentration and light source,
in the absence of hydrogen peroxide (1); on the other hand,
H2O2 absorbs a significantly larger fraction of light than TCE
under the conditions used in this work (Fabs(H2O2))59.5%, t)0
min). Therefore, the contribution of the direct photolysis to
the overall product yield of TCE degradation in the presence
of H2O2 is expected to be significantly smaller than that of
the OH radical pathway.
The direct photolysis of TCE occurs with an overall
quantum yield of φ ) 0.354 (1) and is represented by the
VOL. 41, NO. 5, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1697
FIGURE 1. Time profiles for the decay of TCE, H2O2, and intermediates in UV/H2O2 treatment process (Solid lines represent the computer
modeling fits.).

following primary processes: The mesomeric form of oxyl radical formed in reactions
19 is a carbon-centered radical, which generates another
ClHCdCCl2 + hν f ClHCdC•Cl + Cl• φ ) 0.13 (1) oxyl radical via reaction with oxygen (reaction sequence 20),
(14) which decays through (a) chlorine elimination followed by
hydrolysis to glyoxylic acid (reaction sequence 20) and (b)
H2O, φ ) 0.1 (1) rapid intramolecular 1,2-H shift to a C-centered radical, which
ClHCdCCl2 + hν 98 leads to oxalic acid (reaction 22) through hydrolysis of
k ) 3.2 × 105 s-1 (fitted)
ClHC(OH)-CHCl2 98 dichlorogyoxal formed in reaction sequence 21. The rate
constant fitted for this reaction sequence was 1 × 106 s-1.
Cl2HC-CHO + H+ + Cl- (15)
O2
ClHC•-C(O)Cl 98
ClHCdCCl2 + hν f HCtCCl + Cl2 φ ) 0.032 (1) +H2O,-Cl•
•
(16) O(Cl)HC-C(O)Cl 98 OHCCOO- + Cl- + 2H+
(20)
ClHCdCCl2 + hν f ClCtCCl + HCl φ ) 0.092 (1)
(17) O2, -HO2•
•
O(Cl)HC-C(O)Cl f HO(Cl)C•-C(O)Cl 98
The geminal chlorohydrin formed in reaction 15 releases Cl(O)C-C(O)Cl (21)
HCl to form dichloroacetaldehyde, a precursor of dichloro-
acetic acid (reaction 18). Cl(O)C-C(O)Cl + H2O f HOOCCOOH + 2H+ + 2Cl-
(22)
O2
Cl2HC-CH(OH)2 + •OH 9 -1 -1
8 The quantum yield of dichloroethyne formation (reaction
9 k ) 2 × 10 M s (fitted)
17) was estimated as φ ) 0.092 (1). Hydrolysis and oxidation
k′ ) 1 × 107 M-1 s-1 (fitted) induced by the hydroxyl radical are the expected degradation
Cl2HC-C(OH)2OO• 98
pathways:
Cl2HC-COOH + HO2• (18)
H3O+ H2O
ClCtCCl 98 ClCH2C(O)Cl 98 ClCH2COOH (23)
Monochloroacetic acid was detected at levels as low as
10-6 M. The 1,2-dichlorovinyl radical generated in reaction •OH

14 is the major route to MCA formation, through the reaction ClCtCCl 9 -1 -1

8
9 k ) 3 × 10 M s (fitted)
with O2 with a rate constant k ∼ 4.3 × 109 M-1 s-1 (16), followed -ΗCl H2O
by chlorine atom elimination and rapid hydrolysis of Cl(OH)CdC Cl 98 OdCdC•Cl 98 HOOC-HC•Cl
•

chloroketene intermediate (reaction sequence 19). This (24)

pathway is supported by the experimental observation that O2
MCA was not detected after 14 min of reaction when the HOOC-HC•Cl 98 HOOC-HC(Cl)O• f
fraction of absorbed UV light by TCE was negligible (less CO2•- + H+ + HC(O)Cl (25)
than 3%). Also, MCA was not reported in the •OH-induced
degradation of TCE without UV irradiation (11).
Formyl chloride HC(O)Cl decomposes into HCl and CO
O2
with a rate constant as high as 104 s-1 (31).
ClHCdC•Cl 9 -1 -1
8 The photolytic path to monochloroethyne (reaction 16)
9 k ) 4.3 × 10 M s (16)
occurs with a quantum yield of φ ) 0.032 (1), and its
•
ClHCdC(Cl)OO 9 -1 -1
8 contribution to the TCE decay must be negligible.
9 k ) 2 × 10 M s (fitted)
Therefore, the major degradation intermediates expected
H2O to originate from TCE photolysis are mono- and dichloro-
ClHCdC(Cl)O• 9 8 ClH2CCOOH + Cl•
6 -1 k ) 1.1 × 10 s (fitted) acetic acids; oxalic and glyoxylic acids are predicted mecha-

1698 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 5, 2007

FIGURE 2. pH and O2 patterns (Solid line represents the computer modeling fits.).
FIGURE 3. Schematic representation of TCE degradation mechanism [/photolysis products (1)].
nistically, but their yields must be very low. Chloride ion is
formed as a final product.
TCE Degradation through Chlorine Atom-Induced Reac-
tions. The chlorine atoms are formed through the C-Cl
photolytic bond cleavage in TCE molecule (reaction 14) and
C-Cl bond scission in chlorinated organic radicals.
To our knowledge, the rate constant for the addition of
chlorine atom to the TCE double bond (reaction 26) is
unknown. The chlorine reactions with substituted olefins
occur with rate constants lower by approximately one order
of magnitude than those for the OH radical reactions:
kCl,Cl2C)CCl2 ) 2.8 × 108 M-1 s-1 (30); kCl,H2C)CH2 ) 2.0 × 108 M-1
s-1; kCl,H2C)CHCH2Cl ) 3.0 × 108 M-1 s-1 (34). Therefore, we
believe that a rate constant of k ) 1.9 × 108 M-1 s-1 for reaction
26 would be appropriate and used it in our kinetic modeling.
ClHCdCCl2 + Cl• f Cl2HC-C•Cl2
k ) 1.9 × 108 M-1 s-1 (fitted) (26)
The carbon-centered radical leads to dichloroacetic acid
through the reaction sequence 27:
O2
Cl2HC-C•Cl2 98
H2O [k ) 1 × 105 s-1 (fitted)]
Cl2HC-CCl2O• 98
Cl2HCCOOH + H+ + Cl-+ Cl• (27)
H2O, O2
Cl2HC-CCl2O• 98 CO + CO2 + H+ + Cl-
k ) 9 × 105 s-1 (fitted) (28)
The last step in reaction sequence 27 was also suggested
as a pathway to DCA in pulse radiolysis studies on TCE (11).
The oxyl radical can also undergo C-C bond cleavage
followed by hydrolysis of phosgene and oxidation of Cl2HC•
radical to inorganic products (reaction 28). Therefore, the
intermediate product resulted from the TCE degradation
through chlorine atoms is dichloroacetic acid.
The chlorine atoms are effectively scavenged by both H2O2
(reaction 29) and chloride ions (reaction 30) due to their
large concentrations relative to that of TCE as well as to the
large reaction rate constants. Bicarbonate ions resulted
through mineralization processes react with the chlorine
atoms, generating carbonate radical ions. Such competitive
VOL. 41, NO. 5, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1699
reactions minimize the role of chlorine atoms on the overall
TCE degradation rate. Dichlorine radical ion Cl2•- (reaction
30) reacts with both H2O2 (reaction 31) and TCE with rate
constants much smaller than those of chlorine atoms.
Therefore, slower degradation rates are expected for TCE
once the chloride ion builds up in the solution.
• + - •
Cl + H2O2 f H + Cl + HO2
k ) 3 × 108 M-1 s-1 (assumed) (29)
Cl• + Cl- f Cl2•- k ) 8.6 × 109 M-1 s-1 (35) (30)
Cl2•- + H2O2 f H+ + 2Cl- + HO2•
k ) 1.4 × 105 M-1 s-1 (36) (31)
The rate constant for reaction 29 is not known, but it is
expected to be similar to that of bicarbonate-chlorine atom
reaction [k)2.2 × 108 M-1 s-1 (30)]. Also, the rate constant
for the TCE reaction is unknown, but based on the literature
data for unsubstituted olefins k ) ∼5 × 106-5 × 107 M-1 s-1
(34), it is conceivable to assume it as ∼1 × 107 M-1 s-1. The
chlorine atom and chlorine radical ion reactions discussed
here were considered in the kinetic model.
Although no attempts were made to distinguish experi-
mentally among various pathways of TCE decay, based on
the intermediate product yields and patterns and the
competition kinetics for the free radicals generated during
the course of irradiation, the kinetic model predicts the
following order for the relative contributions of these
pathways to the degradation rate of TCE, integrated over the
first 10 min reaction time: •OH . UV photolysis > Cl2•- .
Cl• (Kinetic Modeling section).
Degradation of Organic Acids. The direct photolysis of
detected organic acids is negligible, their degradation oc-
curring through OH radical-induced reactions.
Glyoxylic acid reacts with both H2O2 and the OH radical.
Its degradation products are formic and oxalic acids, which
are oxidized further to CO2, H2O, and formoyl radical.
OHCCOOH + H2O2 f HCOOH + CO2 + H2O
k ) 0.3 M-1 s-1 (37) (32)
H2O, O2
OHCCOOH (OHCCOO-) + •OH 98
HOOCCOOH (HOOCCOO-) + HO2• (33)
The rate constants for the OH radical reactions of
protonated and unprotonated forms of glyoxylic acid were
determined recently as k ) 5.9 × 108 M-1 s-1 and k ) 1.3 ×
109 M-1 s-1, respectively (38).
HCOOH (HCOO-) + •OH f CO2•- + H+ + H2O (34)
HOOCCOOH (HOOCCOO-, -OOCCOO-) +
•
OH f CO2 + CO2•- + H+ + H2O (35)
The rate constants for the •OH reactions of protonated/
unprotonated forms of formic and oxalic acids are known
(22) and used in the kinetic model.
Both MCA and DCA degraded through H-abstraction
reaction by •OH, followed by the peroxyl radical chemistry
as shown below.
ClCH2COOH + •OH f ClC•HCOOH + H2O
k ) 4.3 × 107 (4 × 108) M-1 s-1 (25, 39) (36)
ClC•HCOOH + O2 f •Ο2(Cl)HCCOOH (37)
1700 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 5, 2007
2•Ο2(Cl)HC(Cl)COOH f 2•Ο(Cl)CHCOOH + O2 (38)
Cl2CHCOOH + •OH f Cl2C•COOH + H2O
k ) 2.75 × 107 (9.2 × 107) M-1 s-1 (39) (39)
Cl2C•COOH + O2 f •O2(Cl2)CCOOH (40)
2•O2(Cl2)CCOOH f 2•Ο(Cl2)CCOOH + O2 (41)
The oxyl radicals formed in eqs 38 and 41 can decay
through both C-C scission and elimination of a chlorine
atom:
•
Ο(Cl)CHCOOH f CO2•- + H+ + HC(O)Cl (42a)
f OHCCOOH + Cl• (42b)
•
Ο(Cl2)CCOOH f CO2•- + H+ + Cl2CdO (43a)
f Cl(Ο)COOH + Cl• f HOOC-COOH (43b)
From the photocatalytic degradation of TCE, Mao et al.
(40) concluded that the Cl-elimination was not a significant
route because the yields of glyoxylic and oxalic acids resulted
from the degradation of MCA and DCA represented less than
6% and 3% of the total carbon degraded, respectively.
Formoyl radical (CO2•-) is produced from various radical
or molecular intermediates and decays through the reactions
with O2 and H2O2:
CO2•- + O2 + H+ f CO2 + HO2•/O2•-
k ) 2 × 109 M-1 s-1 (34) (44)
CO2•- + H2O2 f CO2 + OH- + •OH
k ) 6.3 × 105 M-1 s-1 (41) (45)
Bimolecular decay of CO2•- does not occur given the strong
competition from the reactions 44 and 45.
Peroxyl radical chemistry is more complex than as
depicted above. Cross-termination reactions among them-
selves or involving HO2• which reaches concentrations as
high as 1 × 10-6 M are possible as well as other routes of
decay of the tetroxide intermediates. The mechanism sug-
gested above comprises the most relevant reactions in this
system and are characteristic to the halogenated peroxyl
radicals. Other reactions involving bicarbonate ion and the
resulting carbonate radical ion were not described in the
mechanism but were considered in the kinetic model.
Kinetic Modeling
The model predicts the destruction of TCE and the formation
and fate of MCA, DCA, and oxalic and formic acids, along
with the pH and oxygen time profiles (Figures 1 and 2). The
reactions and kinetic parameters used in the model are
provided throughout the paper and summarized in Table S1
in the Supporting Information.
Selection of the Species and Reactions Used in the Model.
A critical screening of the reactions occurring in this system
was performed based on the competition kinetics at given
reaction times. Some basic considerations are given below.
The ionized and nonionized forms of organic acids have
different reactivity toward the OH radical, and the kinetic
model accounts for that. The concentrations of the pro-
tonated/unprotonated forms were calculated according to
the appropriate acid/base equilibrium constants. The H2O2/
HO2- and HCO3-/CO32-equilibria were ignored given the pH
range observed during the irradiation and the corresponding
pKa values [11.75 (42) and 10.75 (43), respectively].
 The superoxide radical anion O2•- reactions were con-
sidered because the concentration of this radical is about 10
times larger than that of its conjugated acid HO2• at the
beginning of irradiation and cannot be ignored within the
first minute reaction time. The self-termination reaction of
HO2• with regeneration of H2O2 and its reaction with •OH
were included because the concentration of HO2• reaches
levels as high as 1 × 10-6 M. Such reactions are also among
the few ones that contribute to the O2 formation.
The carbonate species not only affect the pH but are also
scavengers of radical species, such as •OH and Cl•. The
electron-transfer reactions with these radicals lead to car-
bonate radical ion CO3•-, which disappears primarily through
the reaction with H2O2 [4.3 × 105 M-1 s-1 (34)].
The chlorine atom reactions were captured in the model.
The steady-state concentration of Cl• was calculated as 10-13-
10-14 M over the first 5 min exposure time and decreased to
10-18 M by 30 min. The major scavengers of Cl• are H2O2
(reaction 29), Cl- (reaction 30), and HCO3- (reaction 46).
HCO3- + Cl• f CO3•- + H+ + Cl-
k ) 2.2 × 108 M-1 s-1 (30) (46)
The dichlorine radical ion Cl2•- is an electrophile, which
reacts with organic and inorganic species through H-atom
abstraction and addition to the double bond, and electron
transfer, respectively (34, 36). H2O2, TCE, and HO2• were
considered the major sinks for Cl2•-, based on their con-
centrations at given reaction times and rate constants: 1.4
× 105 M-1 s-1 (34), 1 × 107 M-1 s-1 (assumed), and 4.3 × 109
M-1 s-1 (34), respectively. Formic acid accumulates to
approximately 4 × 10-4 M, but the rate constant for the Cl2•-
reaction is small [3 × 105 M-1 s-1 (36)], and thus, this reaction
was disregarded in the model.
The •OH concentration was estimated by the model as ∼4
× 10-12 M at ∼3.5 min reaction time and increased by 1 order
of magnitude by the end of the experiment. The rate constant
for the reaction between •OH and Cl- is pH-dependent, and
in acid solution it cannot be ignored. It leads to rapid
formation of Cl2•- through a complex mechanism, which
can be simplified as follows:
•
OH + Cl- f HOCl•-
k ) (4.3 ( 0.4) × 109 M-1 s-1 (44) (47)
HOCl•- f •OH + Cl- k ) (6.1 ( 0.8) × 109 s-1 (44)
(48)
HOCl•- + H+ f Cl• + H2O
k ) (2.1 ( 0.7) × 1010 M-1 s-1 (44) (49)
Cl• + Cl- f Cl2•- k ) 8.6 × 109 M-1 s-1 (35) (30)
Cl2•- f Cl• + Cl- k ) 5.3 × 104 s-1 (35) (50)
In the model, a set of consecutive steps were simulated
as a single reaction only when the rate-limiting step was
clearly identified and the corresponding rate constant was
known or could have been reasonably approximated; oth-
erwise, the model used the individual reactions steps, with
the rate constants taken from the literature (see the Sup-
porting Information).
Modeling Approach. The kinetic equations used in the
model were derived using the governing equation in a
completely mixed batch reactor (CMBR)
dCA
) rA, CA|t)0 ) CA0 (51)
dt
where CAo and CA are the concentrations of species A at time
0 and t, respectively, and rA is the overall kinetic rate
expression for species A in the reaction system. The full set
of kinetic equations was obtained by substituting the rate
expressions for all main species in the TCE-UV/H2O2 system
and is provided in the Supporting Information.
Twelve kinetic parameters were evaluated by fitting the
experimental data obtained for TCE, H2O2, MCA, DCA, formic
and oxalic acids, and pH using a genetic algorithm developed
by Carroll (45). A range of the possible minima and maxima
for each rate constant was provided to the genetic algorithm.
This range was determined based on the literature reported
values for each type of reaction, such that the resulting 12
fitting parameters exhibit reasonable values. For example,
the rate constant for O2 addition to the carbon centered
radicals was set within the range of 5 × 108-1 × 1010 M-1 s-1
(31) and that for β-scission of alkoxyl radicals from 1 × 104
to 1 × 107 s-1 (31). The detailed analysis of the algorithm is
available in the literature (46) and is briefly covered in the
Supporting Information.
TCE Decay Modeling. Figure 1 shows the comparison
between the predicted and experimental patterns of TCE
and its byproducts. The model accounts for the four
photolysis pathways and the reactions with •OH, Cl•, and
Cl2•- radicals. TCE decay exhibits two kinetic regimes. Over
the first 2.5 min it can be described by both zero- and first-
order kinetics, due to the contribution of the direct photolysis
to the decay rate. After the fraction of light absorbed by TCE
decreased by ∼35%, a first-order kinetics is obeyed over ∼1.6
orders of magnitude decay of TCE concentration, with a rate
constant of (9.4 ( 0.6) × 10-3 s-1.
The contribution of each major pathway to TCE decay
was calculated using the concentration profiles of •OH, Cl•,
and Cl2•- and the time step TCE photolysis rate (Supporting
Information).
The calculation shows that over the 12 min exposure time
where TCE decays by ∼2.5 orders of magnitude, the integrated
relative contribution of each pathway to the overall TCE
degradation is 87.4%, 8.5%, 3.7%, and 0.4%, for •OH, UV
photolysis, Cl2•-, and Cl•, respectively. The actual time profiles
of such contributions are given in Figure 4.
Given the complexity of radical reactions in this system
and the lack of kinetic data for TCE reactions with chlorine
and dichlorine radicals, we believe that the model predicts
relatively well the experimental data of TCE decay. The bigger
discrepancies lie in the lower oxygen level zone (4-5 min),
where the mechanism could be more complex than described
above and, thus, not accounted for by the model.
H2O2 Decay Modeling. The degradation of H2O2 follows
zero-order kinetics, with an experimental rate constant of
5.47 × 10-6 M s-1, in good agreement with the model predicted
data (5.73 × 10-6 M s-1). The model predicts well the H2O2
decay.
Intermediates Modeling. The discrepancies between the
model-calculated patterns of DCA, MCA, and formic acid
and the experimental data are less than 10%, 2%, and 7%,
respectively, and could be attributed to both experimental
analytical errors and model imperfections.
pH Modeling. Simulation of pH pattern was based on the
charge balance as described elsewhere (1). In the pH
calculations the model calculates the inorganic carbon
balance at each reaction time assuming that all inorganic
carbon exists either as H2CO3* or HCO3-. The concentration
of bicarbonate is then calculated according to the pKa of
H2CO3*. Carbonate ion was ignored due to the low experi-
mental pH data in the system. Since almost all organic and
inorganic acids were considered in the model, the predicted
pH pattern agrees well with the experimental data. Negatively
charged radicals, i.e., Cl2•-, CO2•-, and CO3•-, were also
included in the charge balance.
VOL. 41, NO. 5, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1701
FIGURE 4. Model predicted relative contributions of the various pathways to the overall decay of TCE (12 min integrated time).
O2 Modeling. The O2 pattern was simulated by calculating
the O2-consuming and generating rate for all oxygen-
involving reactions included in the model (Table S1 in the
Supporting Information). For example, the overall decay of
formic acid, DCA, MCA, and glyoxylic acid, is associated in
the model with the O2 loss rate of 0.5 mol per each mol of
compound, based on the stoichiometric O2 required for
peroxyl radical formation (1 mol) (e.g., reaction 37) and the
amount of O2 released through bimolecular termination of
the corresponding peroxyl radical (0.5 mol) (e.g., reaction
38). However, due to the complexity of the peroxyl radical
reactions, the actual O2 recovery from peroxyl radical
termination could be less than 0.5 mol per mol peroxyl radical
formed.
The predicted data follows the experimental data for the
first 10 min reaction time; after 10 min, the model does not
predict the measured O2 data probably due to the following
reasons or the their combination: not all peroxyl radical
reactions were included in the model, yet they should be
dominant primarily within the 10 min reaction time; the
mass transfer between the atmosphere (reactor headspace)
and the solution was not considered. To our knowledge, no
other similar study on UV/H2O2 applications attempted to
model the experimental O2 pattern, and we believe that the
experimental and predicted O2 concentration time profiles
in this work agree reasonably well.
Acknowledgments
The authors are very grateful to the reviewers of this
manuscript for their thoughtful comments. The modeling
work was supported partially by the Center for Clean
Industrial and Treatment Technologies (CenCITT) sponsored
by the U.S. EPA at Michigan Technological University. Partial
support was provided by the Arizona State University
Foundation and the Richard Snell Presidential Chair Funds.
Supporting Information Available
Calculation of UV photolysis rate, brief description of the
genetic algorithm, full set of the reactions and kinetic
equations used in the model, and the calculation of the
contribution factors of different TCE decay pathways as well
as the radical concentration time-profiles (0-30 min). This
material is available free of charge via the Internet at http://
pubs.acs.org.
1702 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 5, 2007
Literature Cited
(1) Li, K.; Stefan, M. I.; Crittenden, J. C. UV photolysis of trichloro-
ethylene (TCE): product study and kinetic modeling. Environ.
Sci. Technol. 2004, 38, 6685-6693.
(2) Hirvonen, A.; Tuhkanen, T.; Kalliokoski, P. Treatment of TCE-
and PCE-contaminated groundwater using UV/H2O2 andO3/
H2O2 oxidation processes. Water Sci. Technol. 1996, 33, 67.
(3) Miller, J. M.; Giggy, C. L.; Thornton L. M.; Prellberg, J. W.
Advanced oxidation case studies involving third generation UV/
hydrogen peroxide systems. A Symposium on Advanced Oxida-
tion Processes for the Treatment of Contaminated Water and
Air, Proceedings, Toronto, 1990.
(4) Sundstrom, D. W.; Klei, H. E.; Nalette, T. A.; Reidy, D. J.; Weir,
B. A. Destruction of halogenated aliphatics by ultraviolet
catalyzed oxidation with hydrogen peroxide. Hazard. Waste
Hazard. Mater. 1986, 3, 101.
(5) Symons, J. M.; Prengle, H. W.; Belhateche, D. Use of ultraviolet
irradiation and hydrogen peroxide for the control of solvent
contamination in small water utilities. Annual Conference and
Exhibition of AWWA: Los Angeles, CA, 1989.
(6) Zeff, J. D.; Barich, J. T. UV/oxidation of organic contaminants
in ground, waste, and leachate waters. A Symposium on
Advanced Oxidation Process for the Treatment of Contaminated
Water & Air: Toronto, Canada, 1990.
(7) Weir, B. A.; McLane, C. R.; Leger, R. J. Design of a UV oxidation
system for treatment of TCE-contaminated groundwater. En-
viron. Prog. 1996, 15, 179-186.
(8) Weir, B. A.; Sundstrom, D. W. Destruction of trichloroethylene
by UV light-catalyzed oxidation with hydrogen peroxide.
Chemosphere 1993, 27, 1279-1291.
(9) Hirvonen, A.; Tuhkanen, T.; Kalliokoski, P. Formation of
chlorinated acetic acids during UV/H2O2-oxidation of ground
water contaminated with chlorinated ethylenes. Chemosphere
1996, 32, 1091-1102.
(10) Getoff, N. Advancement of radiation induced degradation of
pollutants in drinking and waste water. Appl. Radiat. Isot. 1989,
40, 585-594.
(11) Gehringer, P.; Proksch, E.; Szinovatz, W.; Eschweiler, H. Radia-
tion-induced decomposition of aqueous trichloroethylene
solutions. Appl. Radiat. Isot. 1988, 39, 1227-1231.
(12) Richardson, S. D. Water analysis: emerging contaminants and
current issues. Anal. Chem. 2003, 75, 2831-2857.
(13) Stefan, M. I.; Hoy, A. R.; Bolton, J. R. Kinetics and mechanism
of the degradation and mineralization of acetone in dilute
aqueous solution sensitized by the UV photolysis of hydrogen
peroxide. Environ. Sci. Technol. 1996, 30, 2382-2390.
(14) Mark, G.; Schuchmann, M. N.; Schuchmann, H. P.; Von Sonntag,
C. A chemical actinometer for use in connection with UV
treatment in drinking-water processing. Aqua (Oxford) 1990,
39, 309-313.
(15) Klassen, N. V.; Marchington, D.; McGowan, H. C. E. H2O2
determination by I3--method and KMnO4 titration. Anal. Chem.
1994, 66, 2921-2925.
(16) Mertens, R. and von Sonntag, C. Determination of the kinetics
of vinyl radical reactions by the characteristic visible absorption
spectra of vinylperoxyl radicals. Angew. Chem., Int. Ed. Engl.
1994, 33 (12), 1262-1264.
(17) Hunt, J. P.; Taube, H. The photochemical decomposition of
hydrogen peroxide: quantum yields, tracer and fractionation
effects. J. Am. Chem. Soc. 1952, 74, 5999-6002.
(18) Baxendale, J. H.; Wilson, J. A. The photolysis of hydrogen per-
oxide at high light intensities. Trans. Faraday Soc. 1957, 53,
344-356.
(19) Bielski, B. H. J.; Allen, A. Q. Mechanism of the disproportion-
ation of superoxide radicals. J. Phys. Chem. 1977, 81, 1048-
1050.
(20) Bielski, H. J.; Benon, H. J.; Cabelli, D. E.; Ravindra, L. A.; Alberta,
A. B. Reactivity of perhydroxyl/superoxide radicals in aqueous
solution. J. Phys. Chem. Ref. Data 1985, 14, 1041-1100.
(21) Volman, D. H.; Chen, J. C. The photochemical decomposition
of hydrogen peroxide in aqueous solutions of allyl alcohol at
253.7 Å. J. Am. Chem. Soc. 1959, 81, 4141-4144.
(22) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B.
Critical review of rate constants of hydrated electrons, hydrogen
atoms, and hydroxyl radicals (•OH/•O-) in aqueous solution. J.
Phys. Chem. Ref. Data 1988, 17 (2), 513-886.
(23) Elliott, A. J.; Buxton, G. V. Temperature dependence of the
reactions hydroxyl-superoxide and hydroxyl-hydroperoxy in
water up to 200°C. J. Chem. Soc., Faraday Trans. 1992, 88, 2465-
2470.
(24) Crittenden, J. C.; Hu, S.; Hand, D. W.; Green, S. A. A Kinetic
model for H2O2/UV process in a completely mixed batch reactor.
Water Res. 1999, 33, 2315-2328.
(25) Farhartaziz; Ross, A. B. Selected specific rate constants of transient
from water in aqueous solution; U.S. Department of Com-
merce: Washington, DC, 1977.
(26) Getoff, N. Radiation- and photoinduced degradation of pol-
lutants in water. A comparative study. Radiat. Phys. Chem. 1991,
37, 673-680.
(27) Koester, R.; Asmus, K.-D. Die Reaktionen chlorierter Äthylene
mit hydratisirten Elektronen und OH-Radikalen in wäβriger
Lösung. Z. Naturforsch. 1971, 26b, 1108-1116.
(28) Koester, R.; Asmus, K.-D. Pulse radiolysis studies of halogenated
organic compounds in aqueous solutions. In Proceedings of the
3rd Tihany Symposium on Radiation Chemistry; Dobo, J., Hedvig,
P., Eds.; Akademiai Kiado: Budapest, Hungary, 1972.
(29) Mertens, R.; von Sonntag, C. Reaction of the OH radical with
tetrachloroethene and trichloroacetaldehyde(hydrate) in oxygen-
free aqueous solution. J. Chem. Soc., Perkin Trans. 2 1994, 10,
2181-2185.
(30) Mertens, R.; von Sonntag, C. Photolysis (λ)254nm) of tetra-
chloroethene in aqueous solutions. J. Photochem. Photobiol.,
1995, 85, 1-9.
(31) von Sonntag, C.; Schuchman, H.-P. Peroxyl Radicals in Aqueous
Solution. In Peroxyl Radicals; Alfassi, Z., Ed.; John Wiley & Sons:
New York, 1997; pp 172-234.
(32) Manogue, W. H.; Pigford, R. L. The kinetics of the absorption
of phosgene into water and aqueous solution. Am. Inst. Chem.
Eng. J. 1960, 6, 494-500.
(33) Neta, P.; Gordkowski, J.; Ross, A. B. Rate constants for reactions
of aliphatic carbon-centered radicals in aqueous solution. J.
Phys. Chem. Ref. Data 1996, 25, 709.
(34) Neta, P.; Huie, R. E.; Ross, A. B. Rate constants for reactions of
inorganic radicals in aqueous solution. J. Phys. Chem. Ref. Data
1988, 17, 1027-1284.
(35) Yu, X.-Y.; Bao, Z.-C.; Barker, J. R. Free radical reactions involving
Cl•, Cl2-•, and SO4-• in the 248 nm photolysis of aqueous
solutions containing S2O82- and Cl-. J. Phys. Chem. 2004, 108,
295-308.
(36) Hasegawa, K.; Neta, P. Rate constants and mechanism of reaction
of Cl2- radicals. J. Phys. Chem. 1978, 82, 854-857.
(37) Leitzke, A.; Reisz, E.; Flyunt, R.; von Sonntag, C. The reactions
of ozone with cinnamic acids: formation and decay of
2-hydroxyperoxy-2-hydroxyacetic acid. J. Chem. Soc., Perkin
Trans. 2, 2001, 793-797.
(38) Minakata, D.; Mezyk, S. P.; Cooper, W. J.; Crittenden, J. C. 2006,
unpublished data.
(39) Maruthamuthu, P.; Padmaja, S.; Huie, R. E. Rate constants for
some reactions of free radicals with haloacetates in aqueous
solution. Int. J. Chem. Kinet. 1995, 27, 605-612.
(40) Mao, Y.; Schoneich, C.; Asmus, K.-D. Identification of Organic
Acids and Other Intermediates in Oxidative Degradation of
Chlorinated Ethanes on TiO2 Surfaces en Route to Mineraliza-
tion, A Combined Photocatalytic and Radiation Chemical Study,
J. Phys. Chem. 1991, 95, 10080-10089.
(41) Schwarz, H. A. Reaction of the hydrated electron with water. J.
Phys. Chem. 1992, 96, 8937-8941.
(42) Encyclopedia of Chemical Technology, 4th ed.; Kirk-Othmer, Ed.;
John Wiley and Sons: New York, NY, 1991; Vol. 1.
(43) CRC Handbook of Chemistry and Physics, 86th ed.; Lide, D. R.,
Ed.; CRC Press, Inc.: Boca Raton, FL, 2005-2006.
(44) Jayson, G.; Parsons, B.; Swallow, A. J. Some simple, highly
reactive, inorganic chlorine derivatives in aqueous solution. J.
Chem. Soc., Faraday Trans. 1 1973, 69, 1597-1607.
(45) Carroll, D. L. http://cuaerospace.com/carroll/ga.htm (accessed
2006).
(46) Goldberg, D. E. Genetic Algorithms in Search, Optimization and
Machine Learning; Addison-Wesley: Reading, MA, 1989.
Received for review March 30, 2006. Revised manuscript
received October 16, 2006. Accepted November 2, 2006.
ES0607638
VOL. 41, NO. 5, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1703