Evidence for H2O2 generation during the TiO2-assisted photodegradation of dyes in aqueous dispersions under visible light illumination

Article abstract of DOI:10.1021/jp9846678

Photodegradation of a series of dyes (Rhodamine B, Orange II, Sulfo-rhodamine B, Fluorescein, Alizarin red, Squarylium cyanine, and Eosin) in the presence of TiO₂ particles under air-equilibrated controlled conditions and visible light illumination led to the formation of hydrogen peroxide. Combined with chemical oxygen demand (COD) measurements and in situ adsorption of added H₂O₂, the results reveal that H₂O₂ is a perfidious intermediate species in the process of dye photodegradation. The formation rate of H₂O₂ depends on the rate of dye degradation. It can also be decomposed by TiO₂ particles under visible light irradiation because of a TiO₂/H₂O₂ surface complex. H₂O₂ was detected because its formation rate was greater than its decomposition rate; failing this, H₂O₂ would not be observed. Its decomposition rate depends on the amount of substrates or intermediates formed during dye degradation and adsorbed on the surface of the TiO₂ particles. The greater the degradation rate of the dye or the greater the quantity of substrates or intermediates adsorbed on the TiO₂ surface, the less is the depletion of H₂O₂ with the consequence that a large quantity of H₂O₂ accumulates in the solution bulk permitting its facile detection. Also, pH appears to play a beneficial role in the observation of H₂O₂ formed.

Full text of DOI:10.1021/jp9846678

4862
J. Phys. Chem. B 1999, 103, 4862-4867
Evidence for H2O2 Generation during the TiO2-Assisted Photodegradation of Dyes in
Aqueous Dispersions under Visible Light Illumination
Taixing Wu, Guangming Liu, and Jincai Zhao*
Institute of Photographic Chemistry, Chinese Academy of Sciences, Beijing 100101, China
Hisao Hidaka
Frontier Research Center for the Earth EnVironment Protection, Meisei UniVersity, 2-1-1 Hodokubo,
Hino, Tokyo 191, Japan
Nick Serpone
Department of Chemistry & Biochemistry, Concordia UniVersity, Montreal, Canada H3G 1M8
ReceiVed: December 8, 1998; In Final Form: March 30, 1999
Photodegradation of a series of dyes (Rhodamine B, Orange II, Sulfo-rhodamine B, Fluorescein, Alizarin
red, Squarylium cyanine, and Eosin) in the presence of TiO2 particles under air-equilibrated controlled conditions
and visible light illumination led to the formation of hydrogen peroxide. Combined with chemical oxygen
demand (COD) measurements and in situ adsorption of added H2O2, the results reveal that H2O2 is a perfidious
intermediate species in the process of dye photodegradation. The formation rate of H2O2 depends on the rate
of dye degradation. It can also be decomposed by TiO2 particles under visible light irradiation because of a
TiO2/H2O2 surface complex. H2O2 was detected because its formation rate was greater than its decomposition
rate; failing this, H2O2 would not be observed. Its decomposition rate depends on the amount of substrates or
intermediates formed during dye degradation and adsorbed on the surface of the TiO2 particles. The greater
the degradation rate of the dye or the greater the quantity of substrates or intermediates adsorbed on the TiO2
surface, the less is the depletion of H2O2 with the consequence that a large quantity of H2O2 accumulates in
the solution bulk permitting its facile detection. Also, pH appears to play a beneficial role in the observation
of H2O2 formed.
Introduction
whereas Ross et al.11 have examined the degradation of
terbutylazine under visible irradiation on TiO2 particles sensi-
tized by Rose Bengal.
Recently we reported the photodegradation of several dyes
under exposure to visible light in aqueous TiO2 dispersions.12-14
The visible irradiation mechanism described by eqs 1-6 has
been shown to be different from the UV irradiation pathway as
portrayed previously;14a the dye and not the semiconductor TiO2
is excited by Visible light:
Much attention has been focused in the past two decades on
the photocatalytic degradation of organic pollutants mediated
by TiO2 particles in aqueous dispersions irradiated by UV light
irradiation. The large body of evidence that has been collected
suggests that this method is undoubtedly a potential and effective
approach toward the degradation or mineralization of a wide
variety of harmful/toxic organic pollutants in wastewaters and
toward the purification of drinking water.1-6 Mechanistically,
it is now commonly accepted that the photocatalyst TiO2 is first
excited by UV light and subsequently initiates the photodeg-
radation processes. However, artificial UV light sources tend
to be somewhat expensive, and the UV light component in
sunlight reaching the surface of the earth and available to excite
TiO2 is a relatively small component (ca. 3-5%) of the AM1
solar spectrum. Therefore, recent efforts have been focused on
exploring means to utilize the inexpensive visible light sources
more effectively or to use the inexhaustible sunlight for treating
polluted waters.
dye + hV f dye*
dye* + TiO2 f dye+• + TiO2 (e)
(1)
(2)
-•
TiO2 (e) + O2 f TiO2 + O2
(3)
(4)
(5)
O2-• + TiO2 (e) + 2H+ f H2O2
H2O2 + TiO2 (e) f •OH + OH-
dye+• + O2 (or O2-• or •OH) f
peroxylated or hydroxylated intermediates f f
Dyestuffs represent a class of organic pollutants that absorb
visible light. Electron transfer processes occurring between dyes
and semiconductors, especially TiO2, have been examined and
found to have practical potential.7-9 Kamat and co-workers10
reported the photodegradation of Acid Orange 7 and Naphthol
Blue Black dyes preadsorbed on the surface of TiO2 particles,
degraded or mineralized products (6)
Subsequently, TiO2 plays the pivotal role of an electron carrier
leading to separation of injected electrons and dye cation radi-
cals. The processes described by eqs 1-5, e.g., electron trans-
fer,7-9 generation of the superoxide radical anion,14c H2O2,14c
* Correspondence author fax, +86-10-6487-9375; e-mail, jczhao@ipc.ac.cn.
10.1021/jp9846678 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/22/1999

H2O2 Generation during Photodegradation of Dyes
J. Phys. Chem. B, Vol. 103, No. 23, 1999 4863
and •OH radical formation14b,c have all been demonstrated
earlier. Equation 6 is supported by recent evidence.15 Therefore,
TiO2-assisted photoprocesses provide an attractive route to treat
or pretreat dye pollutants using either artificial visible light or
sunlight. To the extent that studies involving visible light
irradiation to mediate the degradation of pollutants by this
advanced oxidation technology are less reported than others,
more extensive and detailed investigations seem appropriate and
necessary to gain a further understanding of these processes.
On the basis of the visible light irradiation mechanisms (eqs
1-6), H2O2 should be generated. In fact, when H2O2 formation
was probed, it was surprising that not in all cases was the
photogenerated H2O2 observed. To determine the causes of H2O2
generation, we systematically examined what decisiVe factors
might affect the formation and thus observation of H2O2 in TiO2
dispersions during the photodegradation of various dyes under
visible light irradiation. UV-visible spectra and chemical
oxygen demand (COD) of the degraded solution were also
examined to disclose some of the details of the H2O2 formation
process. The results provide a determining contribution to our
understanding of the photoassisted and TiO2-mediated pathway
for the degradation of dyes under visible irradiation.
Procedures and Analyses
Aqueous TiO2 suspensions were prepared by addition of a
given weight of TiO2 powder to a 50-mL aqueous dye solution.
Prior to irradiation, the suspensions were magnetically stirred
in the dark for ca. 30 min to ensure the establishment of an
adsorption/desorption equilibrium. The dispersions were kept
under constant air-equilibrated conditions before and during
irradiation. At given irradiation time intervals, 3-mL aliquots
were collected, centrifuged, and then filtered through a Millipore
filter (pore size 0.22 µm) to remove the TiO2 particulate. The
filtrates were analyzed by recording variations at the wavelength
of absorption band maximum in the UV-vis spectra of the dyes
using a Lambda Bio 20 spectrophotometer (Perkin-Elmer Co.).
COD was assayed using the potassium dichromate titration
method.17 COD values of both the dye suspensions (45 mL)
containing the TiO2 particulates and those of the bulk solution
after removal of TiO2 particles were determined. The concentra-
tion of total peroxides (including organoperoxides and H2O2)
formed during irradiation of a 50-mL solution of dye containing
TiO2 was determined immediately after irradiation at various
time intervals and removal of the TiO2 particles by centrifugation
and filtration. The spectrophotometric DPD method18 was
employed (λ ) 551 nm, ꢀ ) 21,000 M-1 cm-1) wherein the
DPD reagent is oxidized by either H2O2 and/or the organoper-
oxides on the basis of the peroxidase catalyzed reaction.18 It is
known that the catalase enzyme can eliminate H2O2 from the
mixture of H2O2 and organoperoxides.19 If catalase is added to
the degraded solution that contained H2O2 and organoperoxides
before the DPD method is used, then this DPD method assays
only the organoperoxides. Thus, we can discriminate between
H2O2 and organoperoxides and determine whether one or both
are formed during the temporal course of dye photodegradation.
Experimental Section
Materials. TiO2 nanoparticulates (P25, ca. 80% anatase, 20%
rutile; BET area, ca. 50 m2 g-1) were kindly supplied by
Degussa Co. Horseradish peroxidase (POD) was purchased from
Huamei Biologic Engineering Co. (China), and the N,N-
dimethyl-p-phenylenediamine (DPD) reagent from Merck (p.a.).
The dyes (see structures above) Orange II (Baker Co.),
Rhodamine B (Beijing Chemicals Co.), Alizarin red (Beijing
Chemicals Co.), Eosin (Beijing Chemicals Co.), and other
chemicals used in the experiments were all of analytical reagent
grade quality and used without further purification. The Sulfo-
rhodamine B dye was laser grade quality (Across Co.). The
Squarylium cyanine dye was synthesized according to a method
reported earlier.16 Deionized and doubly distilled water was used
throughout this study.
Photoreactor and Light Source. A 500-Watt halogen lamp
(Institute of Electric Light Source, Beijing) was positioned inside
a cylindrical Pyrex vessel surrounded by a circulating water
jacket (Pyrex) to cool the lamp. A cutoff filter was placed
outside the Pyrex jacket to remove radiation below 420 nm and
to completely ensure irradiation of the dispersion only by visible
light wavelengths.
Results and Discussion
The photodegradation of a series of dyes was carried out in
TiO2 dispersions under visible light irradiation during which
the generation of peroxides was examined. The DPD method
employed for peroxide measurements was used for the detection
of both H2O2 and any hydroperoxy organic intermediate that
formed during the photodegradation of the dyes. Through the
discrimination of total peroxides formed (see Experimental
Section), the only peroxide detected was H2O2; no organoper-
oxides formed. For example, in the photodegradation of
Squarylium cyanine the maximal concentration of peroxide
appeared after ca. 30 min of irradiation (curve a, Figure 1); after

4864 J. Phys. Chem. B, Vol. 103, No. 23, 1999
Wu et al.
adsorption ability; dye adsorption should increase with an
increase in the acidity of the dispersion) and to accelerate the
degradation rate. The results plotted in Figure 2 show that H2O2
was generated, as expected, owing to the strong adsorption (Γ,
17.0 × 10-9 mol mg-1 TiO2) and a fast degradation rate (tend
,
350 min for 8 × 10-5 M) after adjustment of pH. Similar results
were obtained under otherwise similar changes in pH (adjusted
from 4.2 to 2.5) for the TiO2 dispersions containing Sulfo-
rhodamine B (Figure 3). More interestingly, addition of the
anionic surfactant DBS (at its cmc of 1.2 mM) to the Rhodamine
B (1.0 × 10-4 M, pH 4.2) dispersions with TiO2 particles
(previous studies showed that adsorption and the rate of
degradation of Rhodamine B are not significantly enhanced by
adjustment of pH but are greatly improved in the presence of
an anionic surfactant such as DBS14b) led to a striking increase
in both the extent of adsorption (∼ 100%) and in the rate of
degradation of the dye (tend ) 80 min for complete discoloration
of 1 × 10-4 M of dye); concomitantly, a considerable quantity
of H2O2 formed with maximal [H2O2] ) 6.0 × 10-5 M after
ca. 80 min of irradiation. These observations demonstrate that
strong adsorption and fast degradation rates of the dyes are
prerequisite for generation and detection of H2O2.
Figure 1. Detection of H2O2 and organoperoxides in the photodeg-
radation of Squarylium cyanine (8.3 × 10-5 M)/TiO2 dispersions (TiO2
20 mg/50 mL). Curve a was obtained by addition of phosphate buffer
(pH ) 6.8), DPD, and POD to the degraded solution after ca. 30 min
of irradiation; curve b was obtained by the same procedure except that
catalase (0.1 mL, 0.5%; activity greater than 2.0 U mg-1) was added
prior to the other reagents (DPD, POD).
Failure to observe formation of H2O2 during the degradation
of the Sulfo-rhodamine B dye at the original pH (4.2) is
somewhat enigmatic since the quantity of dye adsorbed (Γ )
7.0 × 10-9 mol mg-1) compares with that of Fluorescein (10.0
× 10-9 mol mg-1) and the extent of adsorption is greater than
that of Fluorescein by a factor of 2. Compared to the degradation
of Fluorescein (maximal [H2O2 ] ) 3.0 × 10-5 M, ∆COD
changed from an initial 7.7 mg L-1 to 27.0 mg L-1), the
degradation of Sulfo-rhodamine B necessitated a longer irradia-
tion time for its complete discoloration at the original pH of
4.2. Furthermore, addition of H2O2 had no effect on the
degradation rates of all the dyes tested in the present experi-
ments, indicating that H2O2 did not interact with the various
substrates. The above two considerations suggest that H2O2
formed must have been photodecomposed by the TiO2 particu-
lates during the degradation process and that its rate of
decomposition was greater than its rate of formation. To confirm
these assertions, we examined the photodecomposition of
hydrogen peroxide in TiO2 dispersions irradiated with visible
light containing only H2O2. The results depicted in Figure 4
illustrate that H2O2 decomposes under these conditions (note
that no change in [H2O2] occurred either in the presence of TiO2
particles in the dark or in the absence of TiO2 under visible
light irradiation). As well, addition of the surfactant DBS to
the H2O2/TiO2 dispersions decreased the rate of decomposition
of H2O2, which demonstrates that DBS covered the surface of
TiO2 particles to some extent, thereby retarding the interaction
between H2O2 and the TiO2 particles.
addition of catalase to remove H2O2, no other peroxide was
detected (curve b, Figure 1). Measurements of peroxides in other
dye/TiO2 dispersions also showed that only H2O2 formed. When
necessary, control experiments were also carried out. In all cases
examined, except for Eosin, no dye photodegraded and no H2O2
formed in the absence of TiO2 particles, in the presence of
visible light illuminated Al2O3 or SiO2 particles, or in TiO2
dispersions in the dark. For Eosin, degradation (initial concen-
tration 2 × 10-4 M) and H2O2 formation were observed both
in visible light irradiated aqueous dye solutions and in visible
light irradiated dye/TiO2 dispersions. However, the degradation
of Eosin with TiO2 particles was faster (only ca. 100 min for
complete discoloration) than in a homogeneous media (ca. 250
min for complete discoloration). Evidently, the two relevant
degradation mechanisms must be different. The presence of TiO2
significantly accelerates the degradation of Eosin. In this case,
formation of H2O2 in the presence of TiO2 particles should be
ascribed to photoassisted processes.The experimental conditions
and the corresponding findings from H2O2 measurements are
summarized in Table 1. The results show that for the Rhodamine
B (pH ∼ 4.2) and Orange II dyes (pH ∼ 6.5), which adsorb
little on the surface of TiO2 particles, the complete discoloration
of the dyes necessitated longer irradiation times, and any H2O2
that may have formed was not observed. However, for the
Fluorescein, Alizarin red, and Squarylium cyanine dyes, which
strongly adsorb on TiO2, their complete discoloration required
shorter irradiation times and H2O2 was observed; maximal
values detected were 3.0 × 10-5 M, 5.3 × 10-5 M, and 6.6 ×
10-5 M, respectively. For Rhodamine B, Orange II, or Sulfo-
rhodamine B, regardless of the dye concentration and TiO2
particle loading used, there was no evidence that any H2O2
formed at the original pH of 4.2, 6.5, and 4.2, respectively.
The mechanism of dye degradation mediated by semiconduc-
tor TiO2 nanoparticles under visible light irradiation (eqs 1-5)
indicates that the photodegradation of the dyes should generate
the intermediate H2O2. That no hydrogen peroxide formed
during the degradation of Orange II (for Rhodamine B and
Sulfo-rhodamine B see below) is likely due to the little quantity
of dye adsorbed on the TiO2 surface. Consequently, the pH of
the Orange II/TiO2 dispersions was adjusted from 6.5 to 2.5 to
enhance adsorption of Orange II on the TiO2 surface (the
sulfonate moiety in this molecular structure has a strong
Consideration of the quantity of H2O2 that should form in
the Sulfo-rhodamine B/TiO2 suspensions at pH 2.5 suggests that
equivalent concentrations of Sulfo-rhodamine B in dispersions,
albeit containing different loadings of TiO2 particles, should
nonetheless produce equal concentrations of H2O2 after complete
discoloration of the dye if the H2O2 generated is not decomposed
by the TiO2 particles (however, see curves e and f in Figure 3).
Sulfo-rhodamine B dispersions containing three different load-
ings of TiO2 (100, 200, and 300 mg) produced maximal
concentrations of H2O2 of 7.6 × 10-5 M, 4.2 × 10-5 M, and
2.5 × 10-5 M, respectively. After confirming adsorption of H2O2
on TiO2 particles in different dispersions by an in situ assay
(20% for 100 mg, 28% for 200 mg, and 38% for 300 mg of
TiO2), the maximal concentrations of hydrogen peroxide formed

H2O2 Generation during Photodegradation of Dyes
J. Phys. Chem. B, Vol. 103, No. 23, 1999 4865
TABLE 1. Data on the Formation of H2O2 in the TiO2-Mediated Photodegradation of Dyes
c
max d
conc.
adsorp
(%)
TiO2
mg/50 mL
Γ (×10-9)b
tend
CH2O2
dye
(mol L-1
)
pHa
(mol mg-1 TiO2)
(min)
(mol L-1
0
)
Rhodamine B
Rhodamine Be
Orange II
1.0 × 10-5
1.0 × 10-4
2.0 × 10-5
8.0 × 10-5
5.0 × 10-5
5.0 × 10-5
8.0 × 10-5
2.0 × 10-4
8.3 × 10-5
2.0 × 10-4
4.2
4.2
6.5
2.5f
4.2
2.5g
6.1
3.8
4.2
4.8
8
∼100
10
100
100
25
25
83
83
50
50
20
83
0.4
50.0
4.0
66.0
7.0
11.0
10.0
76.0
80.0
13.0
480
80
6.0 × 10-5
0
600
350
600
100
160
120
60
Orange II
41.0
23
36
12.5
38
39
3.4 × 10-5
0
Sulfo-rhodamine B
Sulfo-rhodamine B
Fluorescein
Alizarin red
Squarylium cyanine
Eosin
7.8 × 10-5
3.0 × 10-5
5.3 × 10-5
6.6 × 10-5
11.0 × 10-5
11
100
a Original pH when the dye dissolved in doubly distilled water, adjusted by neither acid nor base. b Γ, Adsorption amount per milligram of TiO2.
c tend, Irradiation time for the complete decoloration of dyes. CH2O2max, maximal concentration of H2O2 formed. e With addition of surfactant DBS
d
at its critical micelle concentration. f ,gAfter adjusting from pH 6.5 and 4.2, respectively.
Figure 2. Concentration changes of Orange II (8 × 10-5 M; curve a),
Figure 4. H2O2 decomposition (initial concentration 6.0 × 10-5 M)
H2O2 formation (curve b) and COD changes (curve c, total COD of
dispersions containing TiO2; curve d, COD changes of bulk solution
after removal of TiO2 particles) as a function of irradiation time during
degradation in the presence of TiO2 particles (25 mg/50 mL) under
visible light irradiation after adjustment of the pH to 2.5.
with irradiation time in TiO2 dispersions (25 mg/50 mL) under visible
light: (a) pH ) 2.5; (b) pH ) 6.0 with added surfactant DBS (at cmc
of 1.2 mM); (c) pH ) 6.0; (d) pH ) 9.5.
COD Measurements. Changes in COD with illumination
times represent the degree of dye degradation or mineralization
from which we can gain some insights into the mechanism of
H2O2 generation during visible light irradiation. Differences in
COD (i.e., ∆COD) between the total COD of the dye/TiO2
dispersion and the COD of the bulk solution after removal of
TiO2 particles, at the same irradiation times reflect the total
organic carbon (substrates plus intermediates) adsorbed on the
TiO2 surface. The greater this ∆COD, the greater is the quantity
of organic carbon adsorbed on the TiO2 surface, which results
in less or no direct interaction between H2O2 formed and TiO2
particulates, thereby favoring accumulation of formed H2O2 in
the bulk solution and its subsequent detection. Indeed, when
H2O2 was easily observed in the cases of Sulfo-rhodamine B
(pH 2.5) and Fluorescein media, ∆COD changes were relatively
large or even increased during the degradation of the dyes
(except for Orange II and Alizarin red, see discussion below).
In particular, in the case of Eosin degradation (see Figure 5), a
greater final ∆COD (52 mg L-1) was obtained despite the small
initial ∆COD (6.0 mg L-1). This demonstrates that an increase
in ∆COD is an important contribution to the observation of H2O2
because the greater quantity of organic carbon (intermediates
or degraded end products) adsorbed on the TiO2 surface hinders
the reaction of H2O2 formed and TiO2 particles, even though
the extent of adsorption is initially small (ca. 10%). Thus, H2O2
accumulated in solution and decayed slowly (note the plateau
in the plot of H2O2 formation versus irradiation time). It is
noteworthy that coverage of the surface of TiO2 particulates by
a substrate and its intermediates varies in the different cases
Figure 3. Concentration changes of Sulfo-rhodamine B (5 × 10-5
M) (curve a), H2O2 formation (curve b), and COD changes (curve c,
total COD of dispersions containing TiO2; curve d, COD changes of
bulk solution after removal of TiO2 particles) as a function of irradiation
time during degradation in the presence of TiO2 particles (100 mg/60
mL) under visible light irradiation after adjustment of the pH to 2.5.
Curves e and f represent H2O2 concentrations formed in the degradation
under otherwise identical conditions except that the loading of TiO2
particles was 200 mg and 300 mg, respectively.
were reestimated, respectively, at 9.1 × 10-5 M, 5.4 × 10-5
M, and 3.5 × 10-5 M. This shows that the greater the loading
of TiO2 particles, the greater the quantity of H2O2 that is
decomposed, thereby emphasizing the photoassisted interaction
between H2O2 and TiO2 particles.

4866 J. Phys. Chem. B, Vol. 103, No. 23, 1999
Wu et al.
TABLE 2. In Situ Adsorption of H2O2 Added to the
Dye/TiO2 Dispersions before Irradiation and after Complete
Decoloration of the Dyes
TiO2
before
after complete
decoloration
dye
Orange II
Fluorescein
Sulfo-rhodamine B 2.5
Alizarin red
Eosin
pH (mg/50 mL) irradiation
2.5
6.1
25
50
83
50
83
5% (8.8)a
50% (7.7)
27% (10.0)
11% (20.0)
38% (6.0)
5% (9.0)
13% (27.0)
20% (31.2)
37% (3.6)
14% (52.0)
3.8
4.8
a Data in parenthese are the ∆COD values in the corresponding
stage.
time during the first 15 h.14a Complete discoloration of Sulfo-
rhodamine B and Rhodamine B needs too long an irradiation
time, with the consequence that any H2O2 formed is decomposed
during the photodegradation and is not detected. These above
results demonstrate that when ∆COD is small and/or the rate
of degradation of the dye is slow, the H2O2 is formed slowly
but is rapidly and extensively decomposed.
Figure 5. Concentration changes (curve a), H2O2 formation (curve
b), and COD changes (curve c, total COD of dispersions containing
TiO2; curve d, COD of bulk solution after removal of TiO2 particles)
as a function of irradiation time during the degradation of Eosin (2 ×
10-4 M, pH ) 4.8) in the presence of TiO2 particles (100 mg/60 mL)
under visible light irradiation.
Diagnosis for an In Situ Adsorption of H2O2. The above
results suggest that a large ∆COD should correspondingly lead
to less adsorption of H2O2 since the ∆COD reflects the coverage
of the TiO2 surface by substrates or intermediates. The in situ
adsorption of H2O2 on the TiO2 surface was verified by addition
of H2O2 to the dye/TiO2 dispersions before irradiation and after
complete discoloration. Note that the final in situ adsorption of
added H2O2 was measured after the TiO2 dispersions had been
completely discolored and after the photogenerated H2O2 had
been completely decomposed for prolonged irradiation times.
The results are collected in Table 2. It is evident that the in situ
adsorption of H2O2 depends on the ∆COD value in the
degradation of a dye. When the initial ∆COD values were
smaller than the final ∆COD, the in situ adsorption was greater
before irradiation than after complete discoloration of the dye;
conversely, the in situ adsorption was less before irradiation
than after complete discoloration of the dye. Adsorption of H2O2
on the TiO2 surface did not vary (always at 5%), whereas the
∆COD remained constant (10 mg L-1) before irradiation and
after complete discoloration of Orange II, which indicates that
adsorption of H2O2 is closely related to ∆COD or to adsorption
of substrates or intermediates on the surface of TiO2 particles.
That is, adsorption of substrates or intermediates on the surface
of TiO2 particles (reflected by ∆COD) directly affects the
interaction of H2O2 with the TiO2 particles and hence signifi-
cantly affects the decomposition rate of the H2O2 produced.
Figure 6. Concentration changes of Alizarin red (2 × 10-4 M) and
H2O2 formation as a function of irradiation time during degradation in
the presence of TiO2 particles (50 mg/50 mL) under visible light
irradiation. Inset: total COD changes containing TiO2 particles (curve
c) and the COD changes of the bulk solution (curve d) after removal
of TiO2 particles with illumination time under otherwise identical
conditions.
examined, with different degrees of interaction between H2O2
and TiO2. Strong interactions (i.e., small ∆COD) led to rapid
decomposition of the photogenerated H2O2. As a demonstration
of this notion, we note the case of Alizarin red. The initial
adsorption of substrate and ∆COD were remarkably large;
subsequently, ∆COD decreased gradually with increase in
irradiation time (see curves c and d in Figure 6). Note the sharp
peak in the plot of H2O2 concentration versus irradiation time
(curve b, Figure 6) because a large quantity of substrates was
adsorbed on the TiO2 surface to hinder H2O2 decomposition
during the initial stages; however, to the extent that the
intermediates adsorbed less in the later stages, the H2O2 present
in the solution bulk decomposed rapidly.
For comparison, changes in COD with irradiation time were
also examined at the original pH of 4.2 for Sulfo-rhodamine B.
The COD changes in these dispersions display the same situation
(large initial ∆COD ∼ 20.0 and final ∆COD near 0 after 6 h
of irradiation) as that of Alizarin red. In the case of Rhodamine
B/TiO2 dispersions at the original pH 4.2, our previous study
had shown that ∆COD increased with increase in irradiation
Implication of pH Effects. Another factor that influences
the adsorption and decomposition of H2O2 is the prevailing pH
of the dispersions. As depicted in Figure 4, the adsorption and
decomposition rates of H2O2 are smaller at the lower pH of
2.5. By contrast, at the higher pH of 9.5, the opposite results
are obtained. These findings indicate that H2O2 displays a slower
decomposition rate in higher acidic media, even though it can
be adsorbed on the TiO2 surface. Accordingly, pH makes a
nonnegligible contribution to the formation and thus observation
of H2O2. A good example of this is the result from the Orange
II dye (at pH 2.5): a large quantity of H2O2 (near the theoretical
value) was detected after a relatively long irradiation time
(Figure 2), albeit the ∆COD values are not large. Thus, it is
possible that the slow decomposition rate of H2O2 at pH 2.5
played a positive role in the detection of H2O2 formed. The
large extent of adsorption and decomposition of H2O2 in the
TiO2 dispersion at the higher pH of 9.5 (Figure 4) implies that
there is an interaction between H2O2 and TiO2 particles, and

H2O2 Generation during Photodegradation of Dyes
J. Phys. Chem. B, Vol. 103, No. 23, 1999 4867
even a stronger interaction by the complexing of titanium ions
and H2O2 in the alkaline solution than seen in acidic media, in
keeping with the results reported by Kiwi and Graetzel.20 We
ascribe the decomposition of H2O2 formed to the photodecom-
position by visible light of the (yellow) titanium(IV) peroxy
surface complex, TiO2/H2O2, formed on the particles.21
Foundation of China (No. 29677019, No. 29725715); the
Foundation of the Chinese Academy of Sciences and the China
National Committee for Science and Technology; by a Grant-
in-Aid for Scientific Research from the Ministry of Education
(No. 066400757) of Japan; and by the Natural Sciences and
Engineering Research Council of Canada (No. A5443).
Earlier, Hoffmann et al. reported the formation of H2O2 in
semiconductor ZnO and TiO2 aqueous dispersions during the
photocatalytic degradation of acetate under UV light irradia-
tion.19b,22 It is relevant to note that the mechanism of H2O2
formation under UV irradiation is different from the present
one under visible irradiation. In the UV irradiation case, H2O2
may be formed through both H2O and OH- ions by surface
oxidation by the photogenerated holes and, in part, by the
disproportion of the superoxide radical anion, whereas decom-
position of H2O2 is caused by reduction by the conduction band
References and Notes
(1) Fox, M. A.; Dulay, M. T. Chem. ReV. 1993, 93, 341.
(2) Ollis, D. F., Al-Ekabi, H., Eds.; Photocatalytic Purification and
Treatment of Water and Air, Elsevier Science Publishers: Amsterdam, 1993.
(3) Ollis, D. F.; Pelizzetti, E.; Serpone, N. EnViron. Sci. Technol. 1991,
25, 1523.
(4) Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr. Chem. ReV. 1995, 95,
735.
(5) (a) Hidaka, H.; Zhao, J.; Pelizzetti, E.; Serpone, N. J. Phys. Chem.
1992, 96, 2226. (b) Zhao, J.; Hidaka, H.; Takamura, A.; Pelizzetti, E.;
Serpone, N. Langmuir 1993, 9, 1646.
(6) Hidaka, H.; Asai, Y.; Zhao, J.; Pelizzetti, E.; Serpone, N. J. Phys.
Chem. 1995, 99, 8244.
(7) Rossetti, R.; Brus, L. E. J. Am. Chem. Soc. 1984, 106, 4336.
(8) (a) Vinodgopal, K.; Kamat, P. V. J. Photochem. Photobiol. A:
Chem. 1994, 83, 141. (b) Kamat, P. V.; Das, S.; Thomas, K. G.; George,
M. V. Chem. Phys. Lett. 1991, 178, 75.
(9) (a) He, J.; Zhao, J.; Shen, T.; Hidaka, H.; Serpone, N. J. Phys.
Chem. 1997, 101, 9027. (b) Zang, L.; Qu, P.; Zhao, J.; Shen, T.; Hidaka,
H. Chem. Lett. 1997, 791. (c) Qu, P.; Zhao, J.; Shen, T.; Hidaka, H. Colloids
Surf. A 1998, 138, 39.
(10) (a) Vinodgopal, K.; Wynkoop, D.; Kamat, P. V. EnViron. Sci.
Technol. 1996, 30, 1660. (b) Nasr, C.; Vinodgopal, K.; Fisher, L.;
Hotchandani, S.; Chattopadhya, A. K.; Kamat, P. V. J. Phys. Chem. 1996,
100, 8436.
(11) Ross, H.; Bendig, J.; Hecht, S. Sol. Energy Mater. Sol. Cells 1994,
33, 475.
(12) (a) Zhang, F.; Zhao, J.; Zang, L.; Shen, T.; Hidaka, H.; Pelizzetti,
E.; Serpone, N. J. Mol. Catal. A, Chem. 1997, 120, 173. (b) Zhang, F.;
Zhao, J.; Shen, T.; Hidaka, H.; Pelizzetti, E.; Serpone, N. Appl. Catal., B:
EnViron. 1998, 15, 147.
(13) (a) Qu, P.; Zhao, J.; Shen, T.; Hidaka, H. J. Mol. Catal. A, Chem.
1998, 129, 257. (b) Zhao, J.; Wu, K.; Wu, T.; Hidaka, H.; Serpone, N. J.
Chem. Soc., Faraday Trans. 1998, 94, 673.
electron, oxidation by the valence band holes, or by reaction
22b
with TiIII.
However, in the visible light irradiation mecha-
nism, H2O2 is formed through continuous reduction of dioxygen
adsorbed on the surface of TiO2 particles;14c H2O2 decomposition
occurs mainly through the photodecomposition of the Ti(IV)/
H2O2 surface complex (see above). To our knowledge, formation
of H2O2 during the photodegradation of dyes in the presence
of TiO2 particles under visible light irradiation has not received
much attention.
Conclusions
Photogenerated H2O2 has been observed during the TiO2-
mediated degradation of dyes by direct detection or by changing
experimental conditions such as pH or by addition of a surfactant
to the dye/TiO2 dispersions. The concentration of H2O2 depends
on both the coverage of substrates and/or intermediates on the
TiO2 surface and the photodegradation rate of the dye. Mecha-
nistically, two moles of dyes produce one mole of H2O2 by
(14) (a) Wu, T.; Liu, G.; Zhao, J.; Hidaka, H.; Serpone, N. J. Phys.
Chem. B 1998, 102, 5845. (b) Zhao, J.; Wu, T.; Wu, K.; Oikawa, K.; Hidaka,
H.; Serpone, N. EnViron. Sci. Technol. 1998, 32, 2394. (c) Wu, T.; Liu,
G.; Zhao, J.; Hidaka, H.; Serpone, N. EnViron. Sci. Technol. 1999, 33, 1379.
(15) He, J.; Zhao, J.; Hidaka, H.; Serpone, N. J. Chem. Soc., Faraday
Trans. 1998, 94, 2375.
-•
reducing O2 via O2 or •OOH, and subsequently the •OH
radical is formed by reduction of H2O2 generated. On the basis
of the present findings that a large quantity of H2O2 did form
and that both discoloration and degradation of dyes occurred,
we can unambiguously infer that electrons injected onto the TiO2
particles from excited state(s) of the dyes are depleted principally
for the formation of H2O2. The dye cation radicals formed after
the electron injection stage are subject to oxidation by molecular
oxygen.
(16) Lin, T.; Peng, B. Dyes Pigm. 1997, 35, 331.
(17) Chinese National Standard: GB 11914-89 (1989).
(18) Bader, H.; Sturzenegger, V.; Hoigne, J. Water Res. 1988, 22, 1109.
(19) (a) Zuo, Y.; Hoigne, J. Science 1993, 260, 71. (b) Carraway, E.
R.; Hoffmann, A. J.; Hoffmann, R. EnViron. Sci. Technol. 1994, 28, 786.
(20) Kiwi, J.; Graetzel, M. J. Mol. Catal. 1987, 39, 63.
(21) Thompson, R. C. Inorg. Chem. 1984, 23, 1794.
(22) (a) Hoffmann, A. J.; Carraway, E. R.; Hoffmann, M. EnViron. Sci.
Technol. 1994, 28, 776. (b) Kormann, C.; Bahnemann, D.; Hoffmann, M.
EnViron. Sci. Technol. 1988, 22, 798.
Acknowledgment. The authors appreciate the generous
financial support of this work by the National Natural Science