Chemical and physical characterization of a TiO2-coated fiber optic cable reactor

Article abstract of DOI:10.1021/es960047d

Practical application of metal oxide photocatalysts for the remediation of contaminated wastestreams often requires immobilization of the photocatalyst in a fixed-bed reactor configuration that allows the continuous use of the photocatalyst by eliminating the need for post-process filtration. A novel optical fiber cable reactor (OFR) is used to transmit UV light to solid-supported TiO₂ in order to investigate the photocatalytic degradation of pentachlorophenol (PCP), 4-chlorophenol (4-CP), dichloroacetate (DCA), and oxalate (OX). The distribution of light as a function of fiber diameter and the quantum efficiencies as a function of incident light intensity are investigated. Light propagation down individual fibers is found to increase with increasing fiber diameter. An increased linear transmission of light results in increased quantum efficiencies, while a 2-order of magnitude reduction in incident light intensity results in a 4-fold increase (φ = 0.010-0.042) in quantum efficiency for the degradation of 4-chlorophenol. The rates of degradation of dichloroacetate and oxalate have strong pH dependencies. Relatively high apparent quantum efficiencies of φ = 0.010, 0.015, 0.08, and 0.17 for PCP, 4-CP, DCA, and OX, respectively, and complete mineralization to CO₂, H₂O, and HCl are observed in the OFR system. Practical application of metal oxide photocatalysts for the remediation of contaminated wastestreams often requires immobilization of the photocatalyst in a fixed-bed reactor configuration that allows the continuous use of the photocatalyst by eliminating the need for post-process filtration. A novel optical fiber cable reactor (OFR) is used to transmit UV light to solid-supported TiO₂ in order to investigate the photocatalytic degradation of pentachlorophenol (PCP), 4-chlorophenol (4-CP), dichloroacetate (DCA), and oxalate (OX). The distribution of light as a function of fiber diameter and the quantum efficiencies as a function of incident light intensity are investigated. Light propagation down individual fibers is found to increase with increasing fiber diameter. An increased linear transmission of light results in increased quantum efficiencies, while a 2-order of magnitude reduction in incident light intensity results in a 4-fold increase (φ = 0.010-0.042) in quantum efficiency for the degradation of 4-chlorophenol. The rates of degradation of dichloroacetate and oxalate have strong pH dependencies. Relatively high apparent quantum efficiencies of φ = 0.010, 0.015, 0.08, and 0.17 for PCP, 4-CP, DCA, and OX, respectively, and complete mineralization to CO₂, H₂O, and HCl are observed in the OFR system.

Full text of DOI:10.1021/es960047d

Environ. Sci. Technol. 1996, 30, 2806-2812
2
(
3), glass plates (24), reactor walls (22, 25), fiberglass mesh
26), and porous films on glass substrates (27, 28).
In order to meet this reactor design challenge, we have
Chemical and Physical
Characterization of a TiO -Coated
2
developed a fixed-bed system that employs a bundled
optical fiber cable array as a means of light transmission
to solid-supported TiO2 (29). Ollis and Marinangeli (30-
Fiber Optic Cable Reactor
N I C O L A J . P E I L L A N D
3
2) and Gapen (33) conducted preliminary investigations
of this type of system. This reactor configuration enhances
the uniformity and distribution and, potentially, the il-
luminated surface area of the light-activated catalyst within
a given reaction volume relative to conventional fixed-bed
reactor designs. This should result in reduced mass
transport limitations to photochemical conversion and
allow for higher processing capacities. The system can be
operated in both gas- and aqueous-phase modes (29). In
addition, the optical fiber reactor system (OFR) allows for
the remote delivery of light over a distance of up to several
kilometers. An OFR system could be used for the in situ
treatment of contaminated sites in the subsurface environ-
ment.
M I C H A E L R . H O F F M A N N *
Environmental Engineering Science, W. M. Keck Laboratories,
California Institute of Technology, Pasadena, California 91125
Practical application of metal oxide photocatalysts for
the remediation of contaminated wastestreams often
requires immobilization of the photocatalyst in a
fixed-bed reactor configuration that allows the continu-
ous use of the photocatalyst by eliminating the need
for post-process filtration. A novel optical fiber
cable reactor (OFR) is used to transmit UV light to solid-
In our previous study (29), we found apparent quantum
efficiencies of the optical fiber reactor comparable to those
reported for slurry-phase reactors for the oxidation of
supported TiO
lytic degradation of pentachlorophenol (PCP),
-chlorophenol (4-CP), dichloroacetate (DCA), and
2
in order to investigate the photocata-
4
-chlorophenol. The OFR was optimized by using a coating
4
thickness of 7 µm that results in 100% absorption of radially-
refracted light out of the fibers and into the photocatalyst
layer and by maximizing the propagation of light down the
optical fibers. Optimization of the OFR resulted in an
increase in the activated photocatalytic surface area and in
a corresponding increase in the apparent quantum ef-
ficiency. Enhancement of linear light transmission was
achieved by maximizing the incident angle of light intro-
duction into the fibers, θi, and by minimizing interfacial
contact area of TiO2 coating particles on the quartz fiber.
In this paper, we explore the effects of fiber diameter on
linear light propagation and absorbed light intensity on
the photochemical quantum efficiency. In addition to a
physical characterization of the OFR, the reactor was used
to investigate the photodegradation of pentachlorophenol,
oxalate (OX). The distribution of light as a function of
fiber diameter and the quantum efficiencies as a
function of incident light intensity are investigated.
Light propagation down individual fibers is found to
increase withincreasingfiberdiameter. Anincreased
linear transmission of light results in increased quantum
efficiencies, while a 2-order of magnitude reduction
in incident light intensity results in a 4-fold increase
(φ ) 0.010-0.042) in quantum efficiency for the
degradationof4-chlorophenol. The rates ofdegradation
of dichloroacetate and oxalate have strong pH de-
pendencies. Relatively high apparent quantum efficien-
cies of φ ) 0.010, 0.015, 0.08, and 0.17 for PCP, 4-CP,
DCA, and OX, respectively, and complete
4
-chlorophenol, dichloroacetic acid, and oxalic acid as a
function of pH and ionic strength.
2 2
mineralization to CO , H O, and HCl are observed in
the OFR system.
Experimental Section
The optical fiber bundled array reactor employed for all
photooxidation experiments has been described in detail
(
29). The basic components are a xenon-arc UV-radiation
Introduction
source, a 310-375 nm band pass filter, a focusing lens, a
light transmitting TiO2-coated fiber optic bundle consisting
of 72 1 mm diameter, 1.2 m long fibers (3M Power-Core
FT-1.0-UMT), and a 227 mLcylindricalPyrexreaction vessel
with a bottom-fritted gas inlet. The TiO2-laden fibers were
prepared by stripping the terminal 20 cm of the fibers to
expose the quartz core, coating with a 13 wt % aqueous
slurry of TiO2 (Degussa P25), and firing at 300 °C maximum.
Input and absorbed light intensities for the single-fiber
light distribution and fiber bundle photooxidation experi-
ments were measured by chemical actinometry using
R-(2,5-dimethyl-3-furylethylidene) (isopropylidene) suc-
cinic anhydride (Aberchrome 540). The experimental
methods have been previously described (29).
Photocatalysis employing wide band-gap metal oxide
semiconductors has been shown to be an effective method
for the remediation of chemical waste streams (1-4). In
the laboratory, photocatalysis with TiO2/ UV has been
studied extensively (5-19). Legrini et al. (20) have argued
that the chemical viability of TiO2 photocatalysis has been
established and that the next stage of development of the
process will depend on new ideas for the fixation of TiO2
and on new designs for photochemical reactors. Fixation
of a photocatalyst in a reactor system is necessary for the
continuous use of the catalyst by eliminating the need for
filtration. A variety of heterogeneous supports have been
explored including glass beads (21, 22), lamp casings (9,
*
To whom correspondence should be addressed; e-mail address:
Light flux measurements were made for uncoated and
coated fibers. The difference between the two was taken
mrh@cco.caltech.edu; fax: 818-395-3170.
2
8 0 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 9, 1996
S0013-936X(96)00047-8 CCC: $12.00
 1996 Am erican Chem ical Society

TABLE 1
Reaction Conditions, Initial Rates, and Relative Quantum Efficiencies, O, for Photooxidations of 4-CP, PCP,
DCA, and OXa
substrate
µ (M)
pHo
[C]i (mM)
Iabs (µEinstein min-1)
initial rate (mM h-1
)
O
Ionic Strength
4
4
4
-CP
-CP
-CP
0.005
0.05
0.5
5.5
5.5
5.5
0.100
0.100
0.100
6.1
6.1
6.1
18.0 × 10-
17.8 × 10-
3
3
3
0.009
0.009
0.009
-
17.0 × 10
pH
DCA
DCA
DCA
DCA
OX
OX
OX
OX
0.001
0.001
0.001
0.003
0.0005
0.0005
0.0005
0.0005
3.0 pHstat
5.0 pHstat
8.0 pHstat
11.0 pHstat
4.25
5.25
6.6 pHstat
8 pHstat
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
5.7
5.8
6.0
5.8
9.3
9.3
9.3
9.3
0.11
0.08
0.0
0.061
0.044
0.0
0.0
0.0
0.20
0.16
0.04
0.01
0.034
0.028
0.007
0.001
Kinetics and Stoichiometry
4
-CP
5.5
0.100
0.062
1.0
5.3 (θi ) 84°)
17.5 × 10-
3
0.010
0.015
0.095
0.017
PCP
DCA
OX
5.5-6.0 pHstat
3.0
4.5-5.0 pHstat
3.9 (θi ) 84°)
4.2 (θi ) 84°)
6.6
18.5 × 10-3
0.001
0.0005
0.125
0.35
0.5
Incident Angle (θi)
DCA
DCA
0.001
0.001
3.0
3.0
1.0
1.0
7.0
0.15
0.125
0.070
0.095
4.2 (θi ) 84°)
a
The solution pH was unadjusted, and an incident angle of light into the fiber bundle of θ
i
) 76° was used unless otherwise noted.
to be the photon flux absorbed by the TiO2 layer, Iabs. The
input light intensities, Iinput, were found to be a function of
the focusing lens used and the corresponding coupling
efficiency of the light beam to the fiber/ fiber-bundle face.
Fiber Diam eter Optim ization. Light from the UVsource
was focused into a single, 1 mm diameter quartz optical
fiber (3M) at a minimum incident angle of θi ) 76° using
a plano-convexquartz lens (Newport-Klinger Rolyn Optics).
The fiber was stripped and coated at increasing lengths. At
each length, uncoated total, coated total, and coated tip
light flux measurements were made. The difference be-
tween uncoated total and coated tip flux measurements
was taken to be the portion of the input flux that was re-
fracted out of the fiber. Fiber diameters of 400, 600, and
for 30 min. Nylon (Nalgene 0.45 µm) or Teflon (Acrodisc
4CRPTFE 0.45 µm) syringe filters were used to filter all
samples.
Chloride production was followed by chemical poten-
-
tiometry using a Cl -specific electrode (Orion 9617BN, 1.0
M NaNO3 matrix) and by ion chromatography (Dionex
BIOLC 4500i series). The pH was determined with a
Radiometer PHM85 pH meter while constant pH was
maintained with a pH-stat (Radiometer PHM85 pH meter,
a Radiometer TTT80 titrator, and a Radiometer ABU80
autoburet). 4-CP, PCP, and associated intermediates were
analyzed spectrophotometrically and by HPLC (λ ) 224/
280 and 224/ 250 nm, respectively). TOC was determined
with a TOC analyzer (Shimadzu TOC-5000). Oxalate and
DCA were quantified by ion chromatography (Dionex
BIOLC 4500i series). Pure oxygen was used at saturation
as the primary electron acceptor and to mix the reactor via
continuous sparging, which was passed through an in-line
gas purifier (Alltech 8132).
The photooxidation rates for 0.5 mM solutions of DCA
and OX at pH 3, 5, 8, and 11 and at pH 4.25, 5.25, 6.6, and
8, respectively, were determined. pH adjustments were
made by the addition of HNO3 or KOH using the pH stat
system described above. In a separate set of experiments,
the oxidation of a 100 µM 4-CP solution was carried out at
ionic strengths of µ ) 0.005, 0.05, and 0.5 M (KNO3) at an
initial pH of 5.5. Reaction rates were determined over 2 h.
1
000 µm and stripped lengths of 5, 10, and 15 cm were
studied.
Light Intensity Dependence. Photooxidation of 4-chlo-
rophenol (4-CP) was carried out to test the relationship
between the incident light intensity in the OFR and the
resulting quantum efficiency. The input light intensity was
reduced to 37%, 10%, and 1% of the maximum using neutral
density filters (Oriel Corp.) with [4-CP]o ) 100 µM. In order
to factor out concentration effects, photooxidation was
allowed to take place until [4-CP]t ) 86 µM. This con-
centration level was obtained after 1 h of irradiation for the
highest-intensity case. The 4-CP concentration was fol-
lowed spectrophotometrically (Shimadzu UV-2101PC) and
by HPLC (HP SeriesII 1090) at λ ) 224 and 280 nm. No
changes in 4-CP concentration due to the volatilization of
Results
4
-CP or water evaporation were observed for a 100 µM
Fiber Diam eter Optim ization. The UV light distribution
down a single quartz fiber as a function of length for fiber
diameters of 400, 600, and 1000 µm was determined. The
percentage of light remaining in the fiber as a function of
the TiO2-coating lengths are shown in Figure 1. For a given
fiber diameter, the amount of light remaining in the fiber
decreases for each coating length. For a given coated fiber
length, less light is transmitted down the fiber as the fiber
diameter is decreased. For example, all of the light has
solution that was sparged with O2 for 30 h in the dark.
Ionic Strength, pH, and Stoichiom etry. Sample solu-
tions (190 mL) of 4-chlorophenol (Aldrich), pentachlo-
rophenol (PCP) (Aldrich), dichloroacetate (DCA) (Spec-
trum), and oxalate (Sigma) of varying initial concentrations,
pH, and ionic strength (Table 1) were prepared from stock
solutions. Before irradiation, the coated fiber bundle was
immersed in the reaction solution and allowed to equilibrate
VOL. 30, NO. 9, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 8 0 7

a
FIGURE 1. Percent of light flux leaving fiber tip as a function of
coated length for quartz optical fibers with diameters of 400, 600,
and 1000 µm. 13 wt % Degussa P25 coating, θ ) 76°.
i
b
FIGURE 3. (a) Oxidation rate of DCA as a function of pH. Reaction
conditions are given in the text and in Table 1. (b) Photooxidation
rate of oxalate as a function of pH. Error bars indicate range of pH
during experiment. Reaction conditions are given in the text and in
Table 1.
FIGURE 2. Quantum efficiency for the oxidation of 4-chlorophenol
as a function of absorbed light intensity. For each oxidation, the
4
≈
-chlorophenol concentration was reduced from 100 to 86 µM. pH
i
Ionic Strength and pH Dependencies. The reaction
conditions and results for variations of ionic strength and
pH are summarized in Table 1. No reduction in the reaction
rate for the photooxidation of 4-CP was observed over the
ionic strength range of 0.005-0.5 M. The reaction rate was
found to be constant (7.6 ( 0.5 µM h ) over the wide range
of ionic strength.
The pH dependence of the photochemical oxidation of
DCA is shown in Figure 3a. The observed photooxidation
rate is the highest at pH 3 and decreases progressively to
zero at pH > 8.
5.5, 13 wt % coating.
been refracted out of the 400-µm fiber within the first 5 cm
of coated fiber. While 11% and 14% of the initial light flux
is transmitted further down the 600- and 1000-µm fibers,
respectively. A small fraction of light is transmitted through
the entire 15-cm coated sections for the 600 and 1000 µm
diameter fibers.
Light Intensity Variation. The apparent quantum
efficiency for the photooxidation of 4-CP was determined
as a function of absorbed light intensity. The quantum
efficiency, φ, is defined as the change in 4-CP concentration
divided by the time to achieve a predetermined level of
degradation, t, divided by the absorbed light flux, Iabs
-1
Astrong, discrete pH dependence for the photooxidation
of C2O42 is shown in Figure 3b. Over the pH range of
-
-
4
.0-4.5, photooxidation of C2O42 (0.5 mM) was complete
-
1
within 2.5 h with -d[OX]i/ dt ) 0.2 mM hr . However,
when the pH range was increased slightly to 5.0-5.5, the
d[4CP]
dt
d[hν]abs
initial rate of photodegradation of a 0.5 mM solution was
(
[4CP] - [4CP] )/ t
o t
-1
reduced to 0.16 mM h
.
φ ≡
=
(1)
Iabs
The pH and ionic strength conditions also had an affect
on the stability of the TiO2 coating. During irradiation of
reaction solutions of 0.5 M ionic strength, pH 3 and pH 11,
the reaction solution became slightly cloudy. This indicates
that a delamination of the coating had occurred. At pH <
3 and at pH > 11, the attachment of the TiO2 coating to the
R-SiO2 fibers was altered.
Kinetics and Stoichiom etry. In order to determine the
ability of a bundled array OFR system to degrade effectively
various hydrocarbon species of interest, the complete
degradation of 4-CP, PCP, DCA, and oxalate were inves-
tigated. The reaction conditions and results of these
experiments are listed in Table 1. The time-dependent
dt
Absorbed light intensities of 23.1 ( 4.8, 9.0 ( 2.1, 2.4 ( 0.5,
-1
and 0.2 ( 0.1 µEinstein (L min) were measured. Irradia-
tion times of 1, 2, 6, and 30 h were required to achieve 14%
degradation of 4-CP for the respective absorbed light
intensities. These experiments yield quantum efficiencies
of φ ) 0.010 ( 0.002, 0.013 ( 0.002, 0.019 ( 0.005, and 0.042
(
0.006 for the decreasing absorbed light fluxes. A 4-fold
increase in the quantum efficiency was achieved for a
-order of magnitude reduction in the absorbed light
intensity (Figure 2).
2
2
8 0 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 9, 1996

FIGURE4. Concentrationvs time plots for4-chlorophenoldegradation
-
1
at pH
The final pH was 4.4.
i
2
) 5.5, Iabs ) 5.3 µEinstein min at a 13 wt % TiO coating.
FIGURE 6. Concentration vs time plots for DCA degradation at pH
i
1
-
)
3.0 for Ilow (θ
at a 13 wt % TiO
respectively.
i
) 84°) ) 4.2 and Ihigh (θ
2
i
) 76°) ) 7.0 µEinstein min
coating. The final pH values were 2.7 and 2.4,
FIGURE 5. Concentration vs time plots for pentachlorophenol
5.5-6.0, Iabs ) 3.9 µEinstein min- at
1
degradation in pH range pH
a 13 wt % TiO coating.
i
2
chemical concentration profiles for the oxidation of 4-CP
and for the production of intermediates and products are
shown in Figure 4. Complete degradation of 4-CP was
achieved after 10 h of irradiation at Iads ) 5.3 µEinstein
min and pH 5.5 with a stoichiometric production of Cl .
No intermediates were detected, although both benzo-
quinone and hydroquinone were detected in other pho-
tooxidation experiments at concentrations less than 10 µM.
However, theywere subsequentlyphotooxidized. Complete
disappearance of total organic carbon (TOC) was observed
within 12 h with a net quantum efficiency of φ ) 0.010 for
FIGURE 7. Decrease in reflections per unit length with an increase
in fiber diameter: (a) base case 7 reflections; (b) increased fiber
diameter with 3 reflections.
-1
-
higher incident angle irradiation was higher (φ ) 0.095)
than the higher intensity irradiation (φ ) 0.07).
Complete mineralization of C2O42 (1.0 mM) over the
pH range of 4.5-5.0 was observed after 5 h of irradiation.
TOC reduction was only slightly out of phase compared to
-
4
-CP degradation.
2
-
2-
2-
-
d[C2O4 ]/ dt. In the case of C2O4 , φC2O4 ) 0.18.
PCP was totally degraded after 6 h with mineralization
achieved after 13 h of irradiation with a stoichiometric
production ofCl (Figure 5). Tetrachlorohydroquinone was
Discussion
-
Enhancing the activated photocatalytic surface area of the
coated fibers by maximizing light propagation down each
fiber appears to result in an increased quantum efficiency
for an OFR (29). Light propagation down a single fiber was
found to be extended by increasing the incident angle of
light introduced into the fiber (i.e., to 90° relative to the
normal to the fiber wall, Figure 7) and minimizing the
interfacial SiO2-TiO2 contact surface area.
detected as an intermediate at a maximum concentration
of 30 µM. A photochemical reaction half-life of 3.0 h and
a quantum efficiency of φ ) 0.015 were determined.
The photooxidation of DCA (Figure 6) was examined at
two different absorbed light intensities (i.e., 7.0 and 4.2
µEinstein min-1) that correspond to the minimum incident
angles of light introduction into the fibers of θi ) 76° and
8
4°, respectively. In general, larger incident angles were
shown to result in higher quantum efficiencies (29).
Complete degradation of DCA was achieved after 10 h of
irradiation at the higher intensity while 96% degradation
was achieved after 14 h at the lower intensity. Chloride
production was 100% and 88% of the stoichiometric values
for the higher and lower intensity irradiations, respectively.
In addition, 98% and 91% of the TOC levels were eliminated
after 13.5 and 14 h, respectively. However, the quantum
As the fiber thickness is increased, photons undergo
fewer reflections at the quartz-TiO2 interface for a fixed
incident angle and a given length. With a larger diameter
fiber, the probability that a photon will be refracted through
the quartz-TiO2 interface is reduced, and thus, the light
propagation down the fiber is extended. The results shown
in Figure 1 are consistent with this interpretation, which
is illustrated schematically in Figure 7. If findings from our
previous study (29) are applied to an optimal fiber diameter,
-
1
efficiency for the lower intensity (i.e., 4.2 µEinstein min ),
VOL. 30, NO. 9, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 8 0 9

the relative quantum efficiencies. Although these two
factors are inversely related, in general, reaction rates
increase with an increase in the total absorbed power. For
example, an order of magnitude increase in the light
intensity resulted in a 30-fold increase in the reaction rate
(
Figure 2). These results suggest that a large total light flux
into an OFR system, similar to those used in slurry reactor
studies, will result in rapid reaction rates since a high fiber
number density will minimize the light intensity per
individual fiber and thus allow the OFR to operate in the
high φ domain.
The pH and ionic strength ofcontaminated wastestreams
are highly variable. These solution characteristics can
influence the speciation of compounds, the nature and
magnitude ofelectrostatic interactions at the semiconductor
photocatalyst (metal oxide) surface, and the competition
for reactive surface sites. Variations in pH, µ, and T have
demonstrable effects on the oxidation rates measured in
FIGURE 8. Increase in fiber number density decreases the number
of photons to which a fiber’s photocatalytic coating is exposed: (a)
Ifiber ) Iinput/14; (b) Ifiber ) Iinput/140.
then the chemical efficiency will be enhanced in an OFR
employing larger diameter fibers in the fiber bundle.
Previous studies have shown an inverse relationship
between the absorbed light flux and the quantum efficiency
TiO / UV systems (6, 7, 39). In the OFR, the oxidation rates
2
of both DCA and oxalate were investigated and found to
(
6, 29, 34-38). In illuminated semiconductor photocata-
be highly pH dependent (Figure 3). Oxalate (pKa1 ) 1.2;
lysts, the charge-carrier recombination and interfacial
charge transfer are second- and first-order processes,
respectively. The charge-carrier density increases with
an increase in the absorbed light intensity. As a result, at
a high absorbed light intensity, the rate of charge-carrier
recombination is increased relative to interfacial charge
transfer, and this leads to a lower relative quantum
efficiency. However, a lower absorbed light flux yields
slower overall reaction rates as the rate of charge pair
generation becomes rate determining. A unique feature of
the OFR system is its ability to minimize the negative light
intensity effects on φ while maintaining a high input light
intensity necessary for rapid overall degradation. This
optimization can be achieved readily by increasing the fiber
number density in the fiber bundle as illustrated in Figure
pKa2 ) 4.0) and DCA (pK ) 1.5) have been found to form
a
inner-sphere surface complexes with TiO at low pH values
2
(40, 41). The reaction rates are highest in the pH region
where the adsorbate and TiO2 surface are oppositely
charged. As the pH approaches the pHzpc of TiO , 6.8 (6),
2
the repulsive forces between the negatively charged
-
TiO surface and adsorbate (e.g., Cl CHCO ) overcome
2
2
2
the tendency toward inner-sphere complexation with a
resulting loss of photoreactivity. While the reaction rates
for DCA appear to be highly correlated with the surface
charge density of the TiO surface as given in Table 2, the
2
oxalate oxidation rates are nonzero at and above the
pHzpc of TiO . This suggests that inner-sphere complexa-
2
tion of oxalate with the TiO2 surface is significant. In
the case of 4-CP, however, ionic strength did not influ-
ence the observed oxidation rate since 4-CP has weak
inner-sphere and electrostatic interactions at low pH
with the TiO2 surface (36). Therefore, the influence of
increasing ionic strength on the spatial extent of the
electrical double layer does not affect the oxidation rates
of 4-CP at low pH.
8
. In this case, the same input photon flux can be divided
among a larger number of fibers and distributed over a
greater photocatalytic surface area. For a given fiber bundle
cross-sectional area, smaller fibers support a higher pho-
tocatalytic surface area.
In order to test this hypothesis, in part, the input light
intensity was successively reduced to 37%, 10%, and 1% of
the maximum input value, and the corresponding quantum
efficiencies were determined (Figure 2). The measured
relative quantum efficiencies were compared at an equiva-
lent level of degradation (14%) rather than for a given
reaction time in order to factor out concentration effects.
A 4-fold increase in the quantum efficiency is achieved for
a 2-order of magnitude reduction in light intensity. The
apparent quantum efficiency of φ ) 0.042 (4.2%) for the
aqueous-phase, photocatalytic oxidation of4-chlorophenol
on TiO2 is high relative to previously reported values for
slurry suspension reactors. In our previous study (29), we
measured an average apparent quantum efficiency of φ )
The observed partial delamination of the coating at the
highest ionic strength, 0.5 M, and at pH values below pH
3 and above pH 11 suggests that repulsive electrostatic
forces cause destablization of the coating. At these pH
values and ionic strength, the net charge on the TiO2
particles is either quite positive or negative relative to
the mid-pH range as shown in Table 2 (42). At high ionic
strengths, the electrical double layer would be compressed
supporting a higher surface charge density at the same
pH allowing the coating to destabilize. These conditions
may produce a repulsive force sufficient to cleave TiO2
particles from each other and from the R-SiO2 surface.
Destablization of the interfacial layer may also be due to
+
-
0
.065 for slurry-phase photooxidation of 4-CP at low light
acid- (H ) and base- (OH ) catalyzed hydrolysis of the >Si-
O-Ti< bonds.
-
1
intensity (e.g., 7.6 µEinstein min ) and a TiO2 loading
Degussa P25) of 1 g L-1. Al-Sayyed et al. (36) reported
quantum efficiencies of φ ) 0.009 and 0.013 for light
(
4-CP, PCP, DCA, and OX undergo photocatalytic oxida-
tion via several mechanistic pathways including direct and
indirect hole transfer and inner- and outer-sphere electron
transfer (eqs 2-5) These compounds have moderate to high
water solubilities and low volatilities. 4-CP and PCP have
been reported to undergo hydroxyl radical insertion as an
initial step in the overall oxidation (35, 36, 39). DCA and
oxalate are believed to undergo hydrogen abstraction and
-
1
intensities of 1000-10 000 µEinstein min , respectively,
-
1
and a TiO2 loading (Degussa P25) of 2 g L . Thus, the
relative quantum efficiency may be significantly enhanced
in OFR systems without a net reduction in overall reac-
tion rates. Reaction rates in a given photocatalytic system
are a function of the total absorbed light intensity and
2
8 1 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 9, 1996

respectively, and for I ) 33 µEinstein min- and a 2 g L-1
TiO2 loading were determined. It should be noted that, in
general, relative quantum efficiencies determined for slurry-
phase systems represent lower limits due to scattering of
some fraction of the input light by the photocatalyst (46-
1
TABLE 2
TiO
2
(Degussa P25) Surface Charge Density as a
Function of pH and Ionic Strength
charge density (C/m2)
pH
µ (M)
+
-
4
8). Total mineralization (i.e., conversion to CO2, H , Cl )
-
-
-
-
-
-
-
-
-
0
3
3
3
3
3
3
3
2
1
4
.3
.5
.3
.9
.9
.7
.7
.5
.5
.5
10
10
10
10
10
10
10
10
10
10
+0.026
+0.009
-0.004
-0.005
-0.017
-0.028
-0.047
+0.015
+0.025
+0.030
was observed. The relatively long irradiation times, up to
14 h, were due to the low light intensities employed (e.g.,
5
6
6
7
8
9
5
5
5
-1
-1
6 compared to 500 µEinstein L min ). Higher light
intensities can be achieved by employing a larger fiber cable
collector with a greater number of fibers and with more
efficient light coupling interfaces.
We have demonstrated that an OFR is a promising
configuration for the practical application of TiO2 photo-
catalysis for the remediation ofcontaminated wastestreams.
The OFR system has the inherent advantages of a fixed-
bed reactor, coupled with the photochemical quantum
efficiencies of a slurry-phase reactor. The OFR is operable
over a broad range of pH and ionic strength, and it can
catalyze the degradation of a broad range of chemical
compounds. In conclusion, the OFR system has the
potential to substantially enhance photocatalytic efficien-
cies and to be applied in remote locations such as in
groundwater wells or underground storage tanks.
direct hole transfer, respectively, according to the following
generally accepted mechanisms (1, 43):
>
>
Ti
OH•+
ecb–
Cl
OH
Ti
OH
OH
–
δ
+
Ti OH
O
O
Ti
O
OH
Cl•
+
+ Cl– (2)
O
+
•
O–O
Acknowledgments
O
OH
HQ
We are grateful to ARPA and ONR (NAV 5 HFMN
N000149J1901) for financial support and to Tim Wu, Janet
Kesselman, Scot T. Martin, and Wonyong Choi for scientific
support. We would also like to thank 3M and Degussa for
their donations of the optical fiber samples and the P25
photocatalyst, respectively.
BQ
Cl
Cl
OH
–δ
H2O
>
Ti
OH•+
ecb–
Cl
Ti
Ti
OH
+
Cl
Cl
>
Ti OH
O
O
OH
Literature Cited
Cl
+
Cl
Cl•
(1) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W.
Chem. Rev. 1995, 95, 69-96.
OH + H+
(3)
(4)
OH
(2) Ollis, D. F.;Al-Ekabi, H. PhotocatalyticPurification and Treatment
of Water and Air; Elsevier: Amsterdam, 1993.
(3) NATO. In NATO Advanced Research Workshop on Homogeneous
and Heterogeneous Photocatalysis, Maratea, Italy, 1985; NATO:
Cl
Cl
O2•–
1
985 p 721.
>
>
Ti
Ti
OH•+
Ti OH2
H2O
(4) Fox, M. A.; Dulay, M. T. Chem. Rev. 1993, 93, 341.
(5) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. J. Phys. Chem.
1988, 92, 5196-5201.
H
Cl
•
–
+
2 2
Cl C CO
C
Ti
OH2
O
C
Cl
(
6) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environ. Sci.
Technol. 1991, 25, 494-500.
O
(
7) Abdullah, M.; Low, G. K. C.; Matthews, R. W. J. Phys. Chem.
1
990, 94, 6820-6825.
TiIV
O
e–C O 2H O
(8) Prairie, M. R.; Evans, L. R.; Stange, B. M.; Martinez, S. L. Environ.
Sci. Technol. 1993, 27, 1776-1782.
(9) Sabate, J.; Anderson, M. A.; Aguado, M. A.; Gimenez, J.;
Cerveramarch, S.; Hill, C. G. J. Mol. Catal. 1992, 71, 57-68.
10) Herrmann, J. M.; Disdier, J.; Pichat, P. J. Phys. Chem. 1986, 90,
2
h+
C
O
TiIV
O
Ti
Ti
OH
OH
(
CO2•– + O2•– + CO2 + 2H+
+
(5)
6
028-6034.
(11) Matthews, R. W. Water Res. 1986, 20, 569-578.
(
(
(
(
12) Pelizzetti, E.; Carlin, V.; Minero, C.; Gr a¨ tzel, M. New J. Chem.
991, 15, 351.
13) Hoffmann, A. J.; Yee, H.; Mills, G.; Hoffmann, M. R. J. Phys. Chem.
992, 96, 5540.
14) Choi, W.; Termin, A.; Hoffmann, M. R. J. Phys. Chem. 1994, 98,
13669-13679.
15) Gr a¨ tzel, M. Heterogeneous Photochemical Electron Transfer; CRC
Press: Boca Raton, FL, 1989.
16) Fox, M. A. Photochem. Photobiol. 1990, 52, 617.
17) Matthews, R. W. J. Catal. 1988, 111, 264.
1
Apparent quantum efficiencies are comparable to those
reported for slurry-phase oxidations. For example, the
apparent quantum efficiencies of 0.065 (29) and 0.013 (36)
for slurry-phase oxidation of 4-CP have been reported.
Hofstadler et al. (44) achieved apparent quantum efficien-
cies of only 0.0002 in a fixed-bed reactor employing TiO2-
coated quartz fiber rods. An apparent quantum efficiency
of φ ≈ 0.013 was reported by Mills and Hoffmann (35) for
1
(
(
(18) Bahnemann, D. W. Isr. J. Chem. 1993, 33, 115.
(19) Matthews, R. W. J. Catal. 1988, 111, 264-272.
(20) Legrini, E.; Oliveros, E.; Braun, A. M. Chem. Rev. 1993, 93, 671.
(
(
(
-
1
slurry-phase oxidation of PCP for I ) 9.1 µEinstein min
-
1
and a TiO2 loading of 0.2 g L at pH 5. Based on data
presented by Barbeni et al. (45) for the photocatalytic
degradation of PCP, apparent quantum efficiencies ranging
from φ ) 0.022 to 0.037 for pH values of 3.0 and 10.5,
21) Serpone, N. Sol. Energy 1986, 14, 121-126.
22) Al-Ekabi, H.; Serpone, N. J. Phys. Chem. 1988, 92, 5726-5731.
23) Sabate, J.; Anderson, M. A.; Kikkawa, H.; Edwards, M.; Hill, C.
G. J. Catal. 1991, 127, 167-177.
VOL. 30, NO. 9, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 8 1 1

(
(
(
(
(
(
24) Anderson, M. A.; Gieselmann, M. J.; Xu, Q. Y. J. Membr. Sci.
988, 39, 243-258.
25) Matthews, R. W.; Abdullah, M.; Low, G. K. C. Anal. Chim. Acta
990, 233, 171-179.
26) Al-Ekabi, H.; Safarzadey-Amiri, A.; Sifton, W.; Story, J. Intl. J.
Environ. Pollut. 1991, 1, 125-136.
27) Sabate, J.; Anderson, M. A.; Kikkawa, H.; Xu, Q.; Cerveramarch,
S.; Hill, C. G. J. Catal. 1992, 134, 36-46.
28) Aguado, M. A.; Anderson, M. A; Hill, C. G. J. Mol. Catal. 1994,
demic Publishers: Dordrecht, The Netherlands, 1991; pp 593-
622.
(39) Tseng, J. M.; Huang, C. P. Water Sci. Technol. 1991, 23, 377-387.
(40) CRCHandbook ofChemistryand Physics;CRC Press: Boca Raton,
1990-91.
(41) Hug, S. J.; Sulzberger, B. Langmuir 1994, 10, 3587-3597.
(42) Martin, S. T.; Kesselman, J. M.; Park, D. S.; Lewis, N. S.; Hoffmann,
M. R. Environ. Sci. Technol. 1996, 30, 2535-2542.
(43) Choi, W.; Hoffmann, M. R. Environ. Sci. Technol. 1995, 29, 1646-
1654.
(44) Hofstadler, K.; Bauer, R.; Novalic, S.; Heisler, G. Environ. Sci.
Technol. 1994, 28, 670-674.
1
1
8
9, 165-178.
29) Peill, N. J.; Hoffmann, M. R. Environ. Sci. Technol. 1995, 29,
974.
2
(
(
(
30) Ollis, D. F.; Marinangeli, R. E. AIChE J. 1977, 23, 415-425.
31) Ollis, D. F.; Marinangeli, R. E. AIChE J. 1980, 26, 1000-1007.
32) Marinangeli, R. E.; Ollis, D. F.; Alfano, O. M.; Negro, A. C. AIChE
J. 1982, 28, 945-955.
33) Gapen, D. C. Photocatalytic Degradation of Chlorinated Hy-
drocarbons. M.S. Thesis, University of Wisconsin, 1991.
34) D’Oliveira, J. C.; Al-Sayyed, G.; Pichat, P. Environ. Sci. Technol.
(45) Barbeni, M.; Pramauro, E.; Pelizzetti, E. Chemosphere 1985, 14,
195.
(
(
(
46) Cabrera, M. I.; Alfano, O. M.; Cassano, A. E. Ind. Eng. Chem. Res.
994, 33, 3031-3042.
47) Cabrera, M. I.; Alfano, O. M.; Cassano, A. E. Ind. Eng. Chem. Res.
995, 34, 500-509.
48) Cabrera, M. I.; Cassano, A. E. Ind. Eng. Chem. Res. 1995, 34,
88-499.
1
(
(
(
(
(
(
1
1
990, 24, 990.
35) Mills, G.; Hoffmann, M. R. Environ. Sci. Tech. 1993, 27, 1681-
689.
4
1
Received for review January 18, 1996. Revised manuscript
36) Al-Sayyed, G.; D’Oliveira, J. C.; Pichat, P. J. Photochem. Photobiol.
A: Chem. 1991, 58, 99-114.
37) Albery, W. J.; Brown, G. T.; Darwent, J. R.; Saievariranizad, E. J.
Chem. Soc. Farad. Trans. 1 1985, 81, 1999-2007.
38) Ollis, D. F. In Photochemical Conversion and Storage of
Solar Energy; Pellizzetti, E., Schiavello, M., Eds.; Kluwer Aca-
X
received April 22, 1996. Accepted May 7, 1996.
ES960047D
X
Abstract published in Advance ACS Abstracts, July 1, 1996.
2
8 1 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 9, 1996