Photocatalytic degradation of chlorinated phenoxyacetic acids by a new buoyant titania-exfoliated graphite composite photocatalyst

Article abstract of DOI:10.1021/jp970132n

A new class of floating composite photocatalysts for water purification is introduced. The composite material is comprised of exfoliated graphite support, titania photocatalyst, and sol-gel-derived methyl silicate binder. The catalyst composition was optimized for the degradation of rhodamine B dye. The photodegradation performance of the optimized catalyst was compared with supported titania film and P-25 titania suspension for the degradation of chlorophenoxyacetic acids. The photodegradation of the coated catalyst was between 50percent and 95percent of the activity of supported photocatalyst film.

Full text of DOI:10.1021/jp970132n

J. Phys. Chem. B 1997, 101, 4623-4629
4623
Photocatalytic Degradation of Chlorinated Phenoxyacetic Acids by a New Buoyant
Titania-Exfoliated Graphite Composite Photocatalyst
Alexander Modestov, Victor Glezer, Iliya Marjasin, and Ovadia Lev*
DiVision of EnVironmental Sciences, Fredy and Nadine Herrmann School of Applied Science, The Hebrew
UniVersity of Jerusalem, 91904, Israel
ReceiVed: January 8, 1997; In Final Form: March 21, 1997X
A new class of floating composite photocatalysts for water purification is introduced. The composite material
is comprised of exfoliated graphite support, titania photocatalyst, and sol-gel-derived methyl silicate binder.
The catalyst composition was optimized for the degradation of rhodamine B dye. The photodegradation
performance of the optimized catalyst was compared with supported titania film and P-25 titania suspension
for the degradation of chlorophenoxyacetic acids. The photodegradation of the coated catalyst was between
50% and 95% of the activity of supported photocatalyst film.
Introduction
C is the concentration of the target compound in solution. Since
the bandgap of titania is about 3 eV, only radiation with
The photocatalytic oxidation of organic species in water by
illuminated titania involves two simultaneous processes: oxida-
tion of the target compound and reduction of dissolved
oxygen.1-3 To promote photocatalytic oxidation of organic
impurities in water, four componentsslight, oxygen, the target
compound, and the photocatalystsshould meet at one place.
Light is supplied either by intense artificial light sources or by
solar irradiation. Solar photochemical reactors usually contain
shallow reaction basins or flowing thin films of the contaminated
solution. The titania photocatalyst is present either in suspension
or in the form of coated films on the bottom of the reactor.4,5
However, for purification of water in deep water or waste-water
reservoirs and for purification of oil slicks and floating pollut-
ants, titania particles are not applicable because titania will sink
and disappear at the bottom of the aquatic system. Thus, to
prevent loss of the titania and to concentrate the titania where
it is most likely to be neededsat the liquid-air interfacesbuoyant
catalysts are needed. Serpone and co-workers6 and Heller and
co-workers7,8 introduced the coated hollow glass beads, Dagan
and Tomkiewicz proposed floating titania aerogels,9 and our
group proposed methyl silicate floating composites.10 An
apparent drawback of titania-coated beads is the gradual
detachment of titania particles from the beads by collisions and
hydrodynamic turbulence. Titania aerogels are not hydrophobic,
and thus once wetted, they will sink in water. In the present
article we introduce a novel buoyant photocatalyst and show
that the fact that the buoyant support is not transparent had
surprisingly little effect on the activity of the composite
photocatalyst.
wavelength shorter than 400 nm can be utilized in the process.
The effective reaction rate constant K reaches saturation at high
concentration of suspended titania, when the titania concentra-
tion is sufficient to capture all useful light. Further increase of
titania concentration results in absorption of light by the upper
layers of the suspension, leaving the deeper sections of the
reaction vial in darkness. Increasing the concentration of the
titania suspension therefore does not increase further the
degradation rate. The above approach enables one to compare
the degradation rates observed with titania suspensions and
supported titania on the basis of a given illuminated cross section
reactor area and maximal utilization of light. Since the light
absorption coefficient of titania is rather high (λ ) 350 nm, R
) 1.4 × 105 cm-1 12), only a small fraction of the titania in
both concentrated titania suspensions and supported titania films
is active, and therefore, an opaque support can be used instead
of the glass beads or methyl silicate supports for manufacturing
floating catalysts.
The goal of the present work was to present a new type of
buoyant titania-coated photocatalysts based on a coating of
porous, floatable matrix containing both closed and open pores.
The open pores provide surface roughness and penetration of
titania into the porous matrix and thus provide higher stability
against erosion. The closed pores are expected to provide
floatation. In this case, collisions of particles will not affect
the catalytic properties because erosion of the particle surface
will reveal new photocatalytically active TiO2 sites.
The exfoliated graphite13,14 material of very high porosity and
superb floatability was chosen as a model of porous carrier for
titania. Exfoliated graphite (EG) is produced by rapid heating
of an intercalated graphite compound to 700-1200° C. Inter-
calation is usually performed by treatment of natural graphite
in a mixture of concentrated sulfuric and nitric acids. Heating
of intercalated graphite results in evaporation of the intercalate
and consequent expansion of graphite crystallites along their
c-axes by up to several hundred times. Exfoliated graphites
are widely used as thermal insulators and grease composition
components, and they can also be pressed to form foils, gaskets,
and seals. So despite the fact that light reflection by graphite
is relatively poor (approximately 30% at 350 nm wavelength15),
it was chosen as a model floating support. Titania (P-25 of
Degussa) was chosen as a model semiconducting material
The rate of the overall photocatalytic process is determined
by light intensity distribution in the reactor. It was shown in
ref 11 that photodegradation process can be described as a
pseudoheterogeneous reaction between the light and the species
in solution with an effective reaction rate constant K:
dC/dt ) -K(S/V)C, where K ) kIn
(1)
The effective reaction surface in this case is the illuminated
cross section of the solution (S). V stands for the volume of
solution, I is the radiation intensity (<400 nm), n is an effective
reaction order with respect to light intensity (1 > n > 0.5), and
X Abstract published in AdVance ACS Abstracts, April 15, 1997.
S1089-5647(97)00132-6 CCC: $14.00 © 1997 American Chemical Society

4624 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Modestov et al.
SCHEME 1
top. Results are reported per illuminated cross-sectional area
of vial or per unit of apparent surface area of titania film. The
test solutions were buffered by 0.01 N Borax (pH 9.2). The
Xe lamp was equipped with dichroic mirror and a glass filter
to produce a 300-400 nm band-pass filter. Light intensity
hitting the target cell was adjusted by use of an Ophir thermopile
light intensity monitor, Nova 03A-P-CAL (Jerusalem, Israel).
Unless otherwise stated, the illuminated vial was kept at a
constant temperature of 25° C regulated by a water jacket and
a thermostat.
In all cases, to ensure lack of direct UV photolysis, blank
experiments were carried out by irradiation of the test solution.
The concentrations of the chlorinated phenoxyacetic acids and
rhodamin B were monitored by periodic sampling of 2 mL
samples during the course of the experiment and determination
of their concentrations by a Varian, Cary 1E UV-vis spectro-
photometer equipped with a diffuse reflectance accessory.
Concentrations of the target compounds were determined by
monitoring the light absorption peaks at 269, 279, 283, and 288
nm for PhAc, Cl-PhAc, 2,4-D, and 2,4,5-T, respectively. The
concentration of Rh-B was determined by monitoring the
absorbance at 543 nm. Total organic carbon (TOC) measure-
ments were performed using a Dohrmann DC-80 automated
laboratory TOC analyzer. The initial solution was used as a
standard. Concentration of Cl- was determined by titration of
the acidified solution with 0.01 N AgNO3, using a Metrohm
673 titroprocessor with combined silver electrode Model
6.0404.100. The quantitation of TOC and Cl- was done by
destructive tests and required large volumes. Therefore, a sepa-
rate test was conducted for each data point and the whole test
solution was taken for analysis at the end of every experiment.
because of its high photocatalytic reactivity. Methyl silicate
(Ormosil) was used as a binder of the titania to the EG support.
The activity of buoyant catalyst was measured for the
photocatalytic oxidation of rhodamine B (Rh-B) and chlorinated
phenoxyacetic acids, namely, 2,4,5-trichlorophenoxyacetic acid
(2,4,5-T), 2,4-dichlorophenoxyacetic acid (2,4-D), and 4-chlo-
rophenoxyacetic acid (Cl-PhAc) (see Scheme 1). The degrada-
tion mechanisms of 2,4-D on floating photocatalysts was
compared to P-25 Degussa catalysts and previous studies.
2,4,5-T and 2,4-D are widely used broad-leaf herbicides. The
degradation of halogenated phenoxyacetic acids and their
reaction intermediates by irradiated P-25 titania,16-18 by Fenton
reagent,19,20 and by UV irradiation16,21 was recently studied.
Rh-B is often used as a model compound in studies of
photocatalytic activity.22 The selection of these compounds as
target compounds was at least partly motivated by their stability
under near-UV radiation and their easy quantitation by spec-
trometric methods.
Determination of the organic intermediates of the photooxi-
dation of 2,4-D was conducted using a straightforward solvent
extraction procedure and by diazomethane esterification fol-
lowed by GC/MSD analysis. Briefly, 4 g NaCl (Frutarom,
Israel) was added to 20 mL sample. The sample was then
acidified with nitric acid (Palacid Ltd., Israel) to pH 2 and mixed
with 2 mL of methyl tert-butyl ether (MTBE, Sigma), and 69
ppb of 1,2,3-trichloropropane was added as an internal standard.
The organic phase was collected in a 3 mL vial, dried over
sodium sulfate (Frutarom, Israel), and concentrated under helium
flow to 0.1 mL. A quantity of 5 µL of the concentrated solution
was injected for GC analysis. Diazomethane esterification was
conducted using 50 µL of the concentrated solution and 50 µL
of diazomethane-saturated MTBE solution, which was prepared
from N-methyl-N-nitroso-p-toluenesulfonamide according to
standard methods.23 An RTX-1 column, 30 m × 0.32 mm, 1
µm film thickness, was used. The oven temperature program
was as follows: initial temperature ) 40 °C, 1 min; first ramp,
5 °C/min to 180 °C; second ramp, 10 °C/ min to 210 °C; third
ramp, 30 °C/min to 240 °C; final hold time ) 4 min. The on-
column injector was kept 3 °C higher than the oven temperature
at all times.
Experimental Section
Reagents. Rhodamin B and methyltrimethoxysilane were
purchased from Aldrich. 2,4,5-Trichlorophenoxyacetic acid and
2,4-dichlorophenoxyacetic acid were supplied by Applied Sci-
ence Lab. Corp., and 4-chlorophenoxyacetic acid (Cl-PhAc) and
phenoxyacetic acid (PhAc) were supplied by Aldrich. Titania
photocatalysts, P-25 Degussa, were donated by Degussa Corp.
The powder was suspended in a test solution and sonicated for
5 min before use. Triply distilled water was used throughout.
Apparatus. Scanning electron microscope (SEM) studies
and electron probe analysis were conducted using a JEOL 840
microscope. Energy dispersive spectrum was taken on EDS
LINK 10000 calibrated on Co and operated on 20 kV. BET
surface area measurements were carried out using a Micromer-
itics Gemini II 2370 surface area analyzer. Measurements were
carried out using nitrogen as the adsorbent. Desorption
isotherms were used for quantitation. Instron 4502 was used
for strain-stress studies of powders. A HP5890 gas chromato-
graph equipped with an electron impact (EI) mass selective
detector HP5971 and a cool on-column injector was used for
determination of reaction intermediates.
Preparation of Supported Titania Films. TiO2 films were
prepared by repeated dip coating of glass slides with a titania
P-25 suspension mixed with sol-gel precursors of methyl
silicate binder. The coated films were dried at 80 °C for 3 min
after each dipping step. The following protocol was used. Glass
slides were degreased in methanol and chromic acid and dried
before the coating step. The coating sol contained 20 mL
methanol, 0.2 g of TiO2 P-25, an appropriate amount of
methyltrimethoxysilane (MTMOS), and 1 N HCl solution. The
amount of acid was set to provide a 4.5:1 water-to-Si ratio. The
sol was allowed to age for 3 h in a closed vessel before coating
was performed.
Photooxidation Experiments. The photooxidation experi-
ments were carried out by exposing 100 mL of aerated, stirred,
buffered aqueous solution of the target substances and the
photocatalyst to the irradiation of a 300 W Oriel ozone-free Xe
lamp. The catalysts were introduced either as sonicated P-25
titania sol, as a coated glass slide, or as a layer of floating
photocatalyst. The test vial was uniformly illuminated from

Phenoxyacetic Acids
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4625
Preparation of Exfoliated Graphite-Titania Composite
Photocatalysts. Exfoliated graphite was used as a porous
support for the photocatalyst. It was prepared by thermal
exfoliation of commercial intercalated graphite compound
(bisulfate graphite). Exfoliation of bisulfate graphite (Asbury
Graphite Co., U.S.) was carried out by heating it to 1000 °C
for about 20 s. Material produced this way had approximately
20 g/L bulk density, 20 m2/g BET surface area, and 10-15%
volume of closed pores, as it was estimated in ref 24. The
volume of closed pores was calculated by subtracting the volume
of open pores from the total volume of powdered particles.
The buoyant photocatalysts were prepared by coating the
exfoliated graphite particles with TiO2 P-25 and methyl silicate
binder. The coating protocol was as follows. A quantity of
25 mL of methanol, an appropriate amount of TiO2 P-25, and
an appropriate amount of methyltrimethoxysilane were mixed
and sonicated for 5 min to ensure dispersion of titania. Then
0.5 g EG was added and the mixture was stirred. A quantity
of 1 M HCl was added to hydrolyze the methyltrimethoxysilane
at a 4.5:1 water-to-silicon ratio. The mixture was allowed to
hydrolyze for 3 h in a closed vessel and then was cast in a Petri
dish. The color of the particles was changed from black to gray
by the coating step. After drying at room temperature for 20
h, the photocatalysts were heat treated for 30 min at 300° C to
remove residual methanol. When a high concentration of titania
(more than 33%) was used, particles formed agglomerates that
could be separated only by mechanical chopping. It takes 1 g
of 1:1:1 Ormosil:TiO2:EG catalyst to cover 60 cm2 surface area
of a solution with one layer of particles. The observed
degradation rates are related per unit surface area of solution
covered by particles of photocatalyst.
The photocatalysts were used in several repeated degradation
tests with no loss of activity during consecutive series of tests.
The photocatalyst retained full floatability even after 3 weeks
of solar illumination. Freshly prepared particles were hydro-
phobic, and visual inspection could not reveal wetting by the
solution. After 2-4 h of irradiation partial wetting of the
particles occurred. This was also accompanied by increased
photoactivity. This change of activity is ascribed to an increase
of the active surface by wetting. The illumination oxidized some
of the methylsiloxane groups that are close to the titania active
sites and thus reduced the hydrophobicity. Similar increase of
the water contact angle and wettability was indeed observed
also for titania-Ormosil films that were deposited on glass
slides.
Figure 1. SEM micrograph of a floating catalyst particle taken in back-
scattering mode: (A) general view; (B) cross section of an exfoliated
graphite-titania composite particle. Composition is TiO2:Ormosil:EG
) 1:1:1.
is much larger than the variability in the distribution of the
binder. Microanalysis of a bright site (TiO2-rich site, Table 1)
indicates enrichment of TiO2 over the Ormosil binder, while
analysis of a typical dark site at the cross section (TiO2-poor,
Table 1) shows enrichment of Ormosil. Exfoliated graphite is
highly hydrophobic, but it is readily wetted by methanol.
During the coating step open pores of the exfoliated graphite
are filled with the TiO2-Ormosil-methanol suspension. Hy-
drolysis and condensation of the methyltrimethoxysilane leaves
a network of Ormosil + TiO2 coating on the surface of the
macropores as well as on the outer surface of the graphite
particles.
The Ormosil plays several roles in this catalyst. First, it
serves as a binder for TiO2, both inside and outside the graphite
particle. Second, the Ormosil interpenetrates the catalyst and
increases its rigidity. Thus, on one hand it increases the erosion
resistance and on the other hand, once the catalyst is eroded or
broken, a fresh TiO2-Ormosil layer is exposed, all that without
losing the closed pores and without affecting the buoyancy of
the composite material. Figure 2 depicts results of Instron
compression tests of unmodified EG, Ormosil-modified EG (1:
1.5), and floating catalyst (TiO2:Ormosil:EG ) 1:1:1), which
demonstrate the increased toughness by the coating procedure.
Table 2 presents the BET surface areas of the P-25 titania, EG,
and the coated EG. The table suggests that most of the surface
area of the composite catalyst is contributed by the methyl
silicate binder and that the titania contributed less than 30% of
the surface area of the composite photocatalyst. Stability of
the methyl-Ormosil binder under UV radiation was investigated
by sunlight exposure (noon time UV radiation ) 3.7 mW/cm2
for <400 nm wavelength) of a powdered composite catalyst
(titania:Ormosil ) 1:4) for 7 days. The FTIR peak at 1276
cm-1, characteristic of the methyl-silicon bond remained
Results and Discussion
Physical Characterization. SEM micrographs of exfoliated
graphite-titania photocatalyst particles TiO2:Ormosil:EG )
1:1:1 are presented in Figure 1. Figure 1A presents typical
micrographs of composite photocatalysts taken in a back-
scattered mode. Figure 1B presents a micrograph of the cross
section of a typical particle. The brightness of the picture is
roughly proportional to concentration of TiO2 (as verified by
X-ray microanalysis). The very bright sites are composed of
droplets of TiO2-Ormosil composite, and the darker sites
correspond to bare graphite surfaces or deep cracks. Examina-
tion of the cross section of the particle, Figure 1B, shows that
TiO2 coating is not confined to the outer surface of the exfoliated
graphite and that the coating is also taking place on the inner
surface of the macropores. Results of X-ray microanalysis are
presented in Table 1.
Table 1 indicates that the average concentrations of TiO2 and
Ormosil at the outer surface and inside the particle are roughly
the same, though the variability in concentration of the titania

4626 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Modestov et al.
TABLE 1: X-ray Microanalysisa
site of analysis
outer surface
average
outer surface
TiO2-rich spot
cross-section
average
inner section
TiO2-poor spot
compound
Ormosil, % weight
TiO2, % weight
20.6
15.8
28.7
45.2
22.1
16.6
23.9
9.4
a Sum of TiO2 + Ormosil content is less than 100%; graphite and interference of surface topology make up the rest.
TABLE 2: Specific Surface Area of Catalysts and Composite Materials
floating catalyst
TiO2:Ormosil:EG
1:1:1
EG/Ormosil
composite
BET surface area (m2/g)
TiO2 P-25
55
EG
23
1:1.5
106
63
Figure 3. Dependence of the Rh-B photodegradation rate constant
(per unit of illuminated surface area) on composition of floating catalyst.
Composition is Ormosil:EG ) 1 with variable TiO2 content. Irradiation
intensity is 60 mW/cm2.
Figure 2. Curves from Instron compression tests of powders: (1)
exfoliated graphite; (2) Ormosil-coated exfoliated graphite (EG:Ormosil
) 1:1); (3) floating catalyst (Ormosil:EG:TiO2 ) 1:1:1). The cylinder
cross section is 2 cm2, the full piston path length is 50 mm, and the
velocity is 10 mm/min.
unaffected. This indicates that degradation of the Si-C bond
is (at most) confined to the vicinity of the titania particles and
the bulk methyl-Ormosil remained intact after prolonged
exposure.
Photocatalytic Activity. Optimization of a three-component
photocatalyst was based on the photocatalyzed oxidation rate
of Rh-B. Two sets of compositions were prepared. The first
contained a constant ratio of EG:Ormosil ) 1 and variable TiO2
content, and in the second, the TiO2:EG ) 1 was kept constant
and the Ormosil content was changed. Photodegradation rates
for Rh-B are presented in Figures 3 and 4. In all cases the
photodegradation of Rh-B followed first-order rate laws. Typi-
cal degradation curves are shown in the insert of Figure 4. Figure
3 shows that increasing the TiO2 content in the photocatalyst
improves the photocatalytic activity. The relative increase is
sharp at low concentrations of TiO2 but levels off at higher
titania content. The activity saturation indicates full coverage
of the particle surface by titania. Similar saturation of the
photocatalytic activity at high titania coverage was observed
for titania-coated glass beads.8 The data scattering for the
titania-rich catalysts is ascribed to agglomeration of catalyst
particles during the preparation step, which causes poor
reproducibility. The dependence of the photocatalytic activity
on the Ormosil concentration is shown in Figure 4. The increase
of activity of the floating catalyst when the Ormosil content in
the photocatalyst was increased is ascribed to better coating of
the surfaces of particles by TiO2 + Ormosil and thus smaller
losses of light by absorption by the uncoated graphite. Based
Figure 4. Dependence of the Rh-B photodegradation rate constant
(per unit of illuminated surface area) on composition of floating catalyst.
Catalyst composition is EG:TiO2 ) 1 with variable Ormosil content.
Irradiation intensity ) 60 mW/cm2. Insert shows concentration-time
curves. Volume is 100 mL; initial concentration is 0.02 mM; illuminated
surface area ) 18 cm2.
on the above measurements, the optimal composition of the
photocatalyst was taken as TiO2:Ormosil:EG ) 1:1:1.
Demonstration of the capability of the buoyant photocatalyst
to perform useful photooxidation and comparative evaluation
of the floating photocatalyst relative to titania films and titania
suspensions were performed taking the photodegradation of the
chlorophenoxyacetic acid family as a test case. To compare
the performance of the suspended catalysts and the two forms
of supported photocatalysts (in buoyant and film forms) based
on a common basis, it was important to ascertain first that the
degradation of the phenoxyacetic acids on these catalysts follows
the same reaction rate laws (eq 1) and has a similar reaction

Phenoxyacetic Acids
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4627
Figure 7. Concentration changes during irradiation of 100 mL of 50
ppm 2,4,5-T and of 2,4-D or Cl-PhAc acids in presence 0.3 g floating
catalyst. Photocatalyst composition is TiO2:Ormosil:EG ) 1:1:1.
Irradiation intensity ) 60 mW/cm2. Slopes of the curves are 1/min:
[2,4-D] - 0.0027, R2 ) 0.994; [2,4,5-T] - 0.002, R2 ) 0.999; [Cl-
PhAc] - 0.0014, R2 ) 0.999.
Figure 5. LN-LN plot of the concentration derivative with time as a
function of light intensity for photodegradation of 60 mL of 50 ppm
2,4-D in presence of 0.3 g floating catalyst with TiO2:Ormosil:EG )
1:1:1.
Figure 6. Arrhenius plot of the degradation rate of 2,4-D by titania-
coated glass slide. Composition is TiO2:Ormosil ) 6; S ) 9.2 cm2;
volume is 60 mL; light intensity ) 60 mW/cm2; slope of the curve )
3350 K; correlation coefficient R2 ) 0.92.
Figure 8. Concentration changes during irradiation of 100 mL of 50
ppm 2,4,5-T and of 2,4-D, Cl-PhAc, and Ph-Ac acids in the presence
of 4 mg TiO2 P-25. Illuminated surface area ) 18 cm2. Irradiation
intensity ) 60 mW/cm2. Slopes of the lines are (1/min): [2,4-D] -
0.0011, R2 ) 0.997; [2,4,5-T] - 0.004, R2 ) 0.95; [Cl-PhAc] - 0.0066,
R2 )0.998; [Ph-Ac] - 0.0022, R2 ) 0.98.
mechanism (the latter will be addressed in the next section).
Indeed, for all three forms of catalysts, the degradation rate
obeyed a linear dependence on the concentrations of the
chlorophenoxyacetic acids and evaluation of the degradation
dependence on light intensity gave for all three types of catalysts
approximately 1/2-order dependence. In all cases the LN-
(degradation rate) vs LN(illumination intensity) yielded straight
lines with correlation coefficients R > 0.94 (Figure 5). Half-
order dependence on the irradiation intensity is often encoun-
tered in low quantum yield photodegradation processes.2,3,25,26
The energy of activation was determined for the degradation
of 2,4-D on thin titania film (TiO2:Ormosil ) 6). Arrehnius
type dependence with a relatively high activation energy (E )
27.8 kJ/mol) was observed (Figure 6). This value is com-
mensurate with reported values for the activation energy of the
degradation of 4-chlorophenol (20.6 kJ/mol27). Agitation of the
solution did not improve the kinetics, which confirmed that the
photodegradation is not dependent on outer diffusion control.
Figures 7-9 depict the photodegradation of the chlorophe-
noxyacetic acids on the three types of photocatalysts. Photo-
degradation curves of 2,4,5-T, 2,4-D, and Cl-PhAc by illumi-
nated floating catalyst are shown in Figure 7. The composition
of the catalyst was TiO2:EG:Ormosil ) 1:1:1 by weight. The
weight of floating catalyst (0.3 g) was adjusted to cover entirely
the water-air interface of the reaction vial (18 cm2) with
approximately one layer of particles. It can be seen that the
degradation rates follow a first-order rate law for all compounds
(linear regression coefficients greater than 0.99 was observed
in all cases). The rate of photodegradation of the chlorinated
phenoxyacetic acids obeyed the following order: 2,4-D >
2,4,5-T > Cl-PhAc. The kinetics of photodegradation of
chlorinated phenoxyacetic acids on P-25 suspension is shown
in Figure 8. It can be seen that the first-order reaction rate law
is still valid, but the order of activity of the phenoxyacetic acids
is 2,4-D > Cl-PhAc > 2,4,5-T. Figure 9 demonstrates that the
photodegradation of the three model compounds on titania film
proceeded at approximately the same rate (K ) 0.016 cm/min).
All three catalysts exhibit a deviation from the Hammett law.
The Hammett equation28 predicts that the reactivity of the
compounds increases with the number of chlorine substituents
on the aromatic ring. However, similar deviations from
Hammett’s prediction were reported for the degradation of
halophenols. Pichat and co-workers27 reported the following
order of activity for the degradation of chloro-substituted phenols
by illuminated P-25 titania: 2,4-dichlorophenol > 2,4,5-
trichlorophenol > 4-chlorophenol. El-Ekabi and co-workers29
reported that an equimolar unbuffered mixture of 2,4,5-T, 2,4-
dichlorophenol, and 4-chlorophenol exhibited approximately the
same degradation rate by illuminated film of titania P-25.
The first-order rate law for degradation of chlorinated
phenoxyacetic acids, which was observed in our work for both
suspended and supported TiO2, indicates that the surface
concentration in all cases was proportional to volume concentra-

4628 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Modestov et al.
Figure 9. Concentration changes during irradiation of 60 mL of 50
ppm solution of 2,4,5-T and of 2,4-D or Cl-PhAc with titania-coated
film. Composition is TiO2:Ormosil ) 3. S ) 8.7 cm2. Irradiation
intensity ) 60 mW/cm2. Slopes of curves are (1/min): [2,4,5-T] -
0.0023, R2 ) 0.998; [2,4-D] - 0.0023, R2 ) 0.999; [Cl-PhAc] -
0.0026, R2 ) 0.999.
Figure 10. Relative concentration changes of 2,4-D, Cl- and TOC
during irradiation of 100 mL of sample with S ) 18 cm2. Initial
conditions are 50 ppm 2,4-D acid, 4 mg TiO2, irradiation intensity of
60 mW/cm2.
tion. However, that does not mean that the proportionality
constant is the same for all acids. So deviation from the
Hammett law may be ascribed to preferential adsorption of the
2,4-D on titania. However, the adsorption of all three com-
pounds on titania at pH 9 is very low. Our attempts to evaluate
the adsorption isotherms of chlorinated phenoxyacetic acids on
TiO2 at concentrations that are relevant to this report were
unsuccessful because of their very low adsorption (<1 µg/g P-25
titania). This is understandable, since the point of zero charge
(PZC) of titania is pH 6.3, and thus, both the target compounds
and the titania surface are negatively charged at basic pH.
The changes of the relative degradation rates due to changes
in the form of the photocatalysts are probably caused by changes
in adsorption of the compounds or their reaction intermediates
on the hydrophobic catalysts. At this time we do not know
whether the changes in the relative lability are caused by the
graphite support or by the Ormosil, but it is clear that the support
is affecting to some extent the mechanism of the reaction.
Figure 11. Dynamic evolution of the 2,4-D photodegradation inter-
mediates. Numbers corresponds to compounds in Scheme 2 and are
the following: 1 ) 2,4-D; 2 ) 2,4-dichloro-6-hydroxyphenoxyacetic
acid, 3 ) 5,7-dichloro-1,4-benzodioxan-2-one, 5 ) 2,4-dichlorophenol,
6 ) 3,5-dichloro-1,2-hydroquinone
SCHEME 2
Comparison of the photocatalytic activity of the coated plate
with the activity of the floating catalyst shows that activities of
particles of composition TiO2:Ormosil:EG ) 1:1:1 range
between 48% and 94% of the activity of the film (for 2,4-D
and Cl-PhAc, respectively). The relatively small difference in
activity shows that most of the outer surface of the particles is
covered by TiO2 and that only a small fraction of the light is
lost in gaps between particles of floating catalyst and by
absorption by bare graphite. The activity of P-25 suspension
was only 2-5 times larger than that of the supported catalysts.
Mechanism and Reaction Intermediates. Figure 10 pre-
sents concentration changes during photocatalytic oxidation of
2,4-D by a P-25 suspension. The photocatalytic oxidation is a
complicated multistage process. The decay of the UV absorp-
tion peak indicates only the disappearance of the initial
compound. This process under conditions of Figure 10 is
characterized by a half-life of approximately 20 min. The half-
life of Cl- release is approximately 110 min, and the half-life
of TOC removal is approximately 13 h, that is, approximately
40 times longer than the half-life of the initial 2,4-D acid.
Scheme 2 demonstrates the degradation pathways of 2,4-D
on floating catalyst based on aromatic intermediates identified
by GC/MSD. Figure 11 depicts the dynamics of formation of
the most abundant aromatic reaction intermediates. Since in
most cases standards were unavailable, quantitation was based
on the abundance of the products relative to the abundance of
2,4-D. Our studies were conducted at pH 9 and compared to
results at pH 4.2. The same intermediates were identified for
the floating and suspended catalysts.
Pathway 1 f 2 f 3 represents the oxidative hydroxylation
of 2,4-D, which is expected at the R position, which allows
further lactone formation. Pichat and co-workers16 also found
compound 3 (5,7-dichloro-1,4-benzodioxan-2-one) but attributed

Phenoxyacetic Acids
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4629
References and Notes
it to an artifact of the high-temperature gas chromatography.
Pignatello20 also identified compound 3, but both researchers
did not identify intermediate 2. Our findings clearly indicate
that compound 3 is not an artifact, since the methylation reaction
employed in our studies impedes the cycle-closing reaction.
Additionally, the use of cool on-column injection precludes high-
temperature reactions in the injector. Theoretically, such
reactions could occur also in the capillary column itself, but
that would have led to the formation of a very broad peak, which
was not observed in our studies. In fact, the lactone peak had
a regular shape. Pichat,16 Pignatello,20 and also Barbeni et al.18
(who studied the degradation of 2,4,5-T) correctly inferred the
formation of the hydroxylated intermediate, compound 2, though
this was identified only in the current studies probably because
of the higher sensitivity gained by the cool on-column injector.
A second oxidation pathway involves the oxidative decarboxyl-
ation of the aliphatic chain either directly to dichlorophenol
(compound 5) or most probably to a 2,4-dichlorophenol formate
intermediate (compound 4). The mechanism of photooxidative
decarboxylation was originally suggested for aliphatic acids by
Yoneyama, Bard, and co-workers30 (see eq 2, R ) X1X2X3Ph,
where X1, X2, and X3 are H or Cl).
(1) Gerischer, H.; Heller, A. J. Phys. Chem. 1991, 95, 5261.
(2) Gerischer, H. In Photocatalytic Purification of Water and Air; Ollis,
D. F., Al-Ekabi, H., Eds.; Elsevier: Amsterdam, 1993; pp 1-18.
(3) Gerischer, H. Electrochim. Acta 1995, 40, 1277.
(4) Matthews R. W. In Photocatalytic Purification of Water and Air;
Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier: Amsterdam, 1993; pp 121-138.
(5) Turchi, C.; Mehos, M.; Pacheco, J. In Photocatalytic Purification
of Water and Air; Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier: Amsterdam,
1993; pp 789-794.
(6) Serpone, N.; Bogarello, A.; Harris, R.; Cahill, P.; Bogarello, M.
Sol. Energy Mater. 1986, 14, 121.
(7) Rosenberg, I.; Brock , J. R.; Heller, A. J. Phys. Chem. 1992, 96,
3423.
(8) Heller, A.; Nair, M.; Davidson, L.; Luo, Z.; Schwitzgebel, J.; Norrel,
J.; Brock, J. R.; Lindquist, S.-E.; Ekerdt J. G. In Photocatalytic Purification
of Water and Air; Ollis, D. F., Al-Ekabi, H., Eds.; Elsevier: Amsterdam,
1993; pp 139-153.
(9) Dagan, G.; Tomkiewicz, M. J. Phys. Chem. 1993, 97, 12651.
(10) Dagan, G.; Sampath, S.; Lev, O. Chem. Mater. 1995, 7, 446.
(11) Modestov, A.; Lev, O. In preparation.
(12) Eagles, D. M. J. Phys. Chem. Solids 1964, 25, 1243-1251.
(13) Chung, D. D. L. J. Mater. Sci. 1987, 22, 4190.
(14) Dowell, M. B.; Howard, R. A. Carbon 1986, 24, 311.
(15) Gmelins Handbuch der anorganischen Chemie, 8. Auflage, Kohlen-
stoff, Teil B-Lieferung 2, System-Nummer 14; Verlag Chemie: Weinheim,
1968; p S697.
(16) Pichat, P.; D’Oliveira, J.-C.; Maffre, J.-F.; Mas, D. In Photocatalytic
Purification and Treatment of Water and Air; Ollis, D. F., Al-Ekabi, H.,
Eds.; Elsevier: Amsterdam, 1993; p 683-688.
•
R-OCH2COO- + p+ f R-OCH2 + CO2
•
R-OCH2 + O2 f R-OCH2OO• f R-OC(O)H +OH•
(17) Trillas, M.; Peral, J.; Domenech, X. Appl. Catal. B 1995, 5, 377.
(18) Barbeni, M.; Morello, M.; Pramauro, E.; Pelizzetti, E.; Vincenti,
M.; Borgarello, E.; Serpone, N. Chemosphere 1987, 16, 1165.
(19) Sanchez, L.; Peral, J.; Domenech, X. Electrochim. Acta 1996, 41,
1981.
R-OC(O)H + H2O f R-OH + HCOO- + H+ (2)
(20) Sun, Y.; Pignatello, J. J. J. Agric. Food Chem. 1993, 41, 1139.
Pignatello, J. J. EnViron. Sci. Technol. 1992, 26, 944.
(21) Chamarro, E.; Esplugas, S. J. Chem. Technol. Biotechnol. 1993,
57, 273.
(22) Anderson, C.; Bard, A. J. Phys. Chem. 1995, 99, 9882.
(23) Eaton, A. D., Clesceri, L. S., Greenberg, A. E., Eds. Standard
methods for the examination of water and wastewater, 19 ed.; APHA:
Washington, DC, 1995.
(24) Marjasin, I.; Sandbank, E.; Shelef, G. U.S. Patent N5282975, 1.02,
1994.
(25) Serpone, N.; Pelizzetti, E.; Minero, C; Fox, M. A.; Draper, B.
Langmuir 1989, 5, 250.
(26) Hofstadler, K.; Ruppert, G.; Bauer, R.; Heisler, G; Novalic, S. In
Photocatalytic Purification of Water and Air; Ollis, D. F., Al-Ekabi, H.,
Eds.; Elsevier: Amsterdam, 1993; pp 777-782.
(27) Pichat, P.; Guillard, C.; Maillard, C.; Amalric, L; D’Oliviera, J-C.
In Photocatalytic Purification of Water and Air; Ollis, D. F., Al-Ekabi, H.,
Eds.; Elsevier: Amsterdam, 1993; pp 207-223.
(28) March, J. AdVanced organic chemistry, 4th ed.; John Wiley: New
York 1992; pp 278-286.
(29) Al-Ekabi, H.; Serpone, N. In Photocatalysis. Fundumentals and
applications; Serpone, N., Pelizzetti, E., Eds.; Wiley: New York, 1989;
pp 457-488.
Compound 4 detected by Pichat,16 Pignatello,20 and Barbeni18
in acid media was not found under our experimental conditions
(pH 9), but we identified it in pH 4 studies. This suggests that
the hydrolysis of the phenolic ester, 4 to the corresponding
phenol, 5 is much faster under basic condition than in acid
media. Pichat and co-workers16 identified also 2,4-dichlo-
romethoxybenzene as a reaction byproduct, but in our studies
it was found only after methylation of the reaction mixture as
a methylated derivative of 5, while unmethylated samples did
not show even traces of the 2,4-dichloromethoxybenzene. Com-
pound 5 can be further oxidized to form 3,5-dichloro1,2-
hydroquinone (6) as anticipated by the previous studies.16
Note that the initial degradation pathway from 1 to 3 does
not lead to TOC reduction, which explains the lag time in the
reduction of TOC (Figure 10). In contradistinction chlorides
start to accumulate from time zero because of the reductive as
well as the oxidative pathways.
Acknowledgment. We thank the KFK, Germany and the
Ministry of Science, Israel for supporting this research.
(30) Yoneyama, H.; Takao, Y; Tamura, H.; Bard, A. J. J. Phys. Chem.
1983, 87, 141.