Effect of copper ions on the formation of hydrogen peroxide from photocatalytic titanium dioxide particles

Article abstract of DOI:10.1016/S0021-9517(03)00197-0

Reactive radicals such as hydroxyl radical (OH^.), hydrogen peroxide (H₂O₂), and superoxide anion (O₂ ^-) are the main power driving titanium dioxide (TiO₂) photocatalytic reactions, for example, photokilling of biological cells. Here, the effect of copper ions on the formation of H₂O₂ over photocatalytic TiO₂ was investigated. Under an oxygen-purged solution, formation of H₂O₂ was increased dramatically up to 20 times by the addition of a small amount of copper ions. By using the Fenton reaction, the H₂O₂ formed can be converted into OH^., a highly reactive radical. In contrast, under nitrogen-purged solution, no H₂O₂ was formed even in the presence of an electron acceptor, silver ion (Ag⁺). These results clearly show that H₂O₂ was generated from the reduction site of the photoexcited TiO₂ and indicate an effective way to increase the photocatalytic efficiency.

Full text of DOI:10.1016/S0021-9517(03)00197-0

Journal of Catalysis 219 (2003) 214–218
www.elsevier.com/locate/jcat
Effect of copper ions on the formation of hydrogen peroxide
from photocatalytic titanium dioxide particles
Ruxiong Cai,a,∗ Yoshinobu Kubota,b and Akira Fujishima c
a
Singapore Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075
b
c
Department of Urology, Yokohama City University, School of Medicine, 3-9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan
Department of Applied Chemistry, School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
Received 30 December 2002; revised 1 April 2003; accepted 23 April 2003
Abstract
Reactive radicals such as hydroxyl radical (OH ), hydrogen peroxide (H O ), and superoxide anion (O ) are the main power driving
·
−
2
2
2
titanium dioxide (TiO ) photocatalytic reactions, for example, photokilling of biological cells. Here, the effect of copper ions on the formation
2
of H O over photocatalytic TiO was investigated. Under an oxygen-purged solution, formation of H O was increased dramatically up
2
2
2
2 2
·
to 20 times by the addition of a small amount of copper ions. By using the Fenton reaction, the H O formed can be converted into OH ,
2
2
a highly reactive radical. In contrast, under nitrogen-purged solution, no H O was formed even in the presence of an electron acceptor, silver
2
2
+
ion (Ag ). These results clearly show that H O was generated from the reduction site of the photoexcited TiO and indicate an effective
2
2
2
way to increase the photocatalytic efficiency.
2003 Elsevier Inc. All rights reserved.
Keywords: Photocatalytic reaction; Oxidation site; Reduction site; Reactive radical; Hydrogen peroxide; Cell killing; Fenton reaction; Photocatalytic
efficiency
1. Introduction
for pure anatase) excites an electron from the valance band
to the conduction band, leaving a positively charged hole in
the valence band. This positive hole will oxidize water −to
In recent years, photocatalytic reactions with TiO2 have
been investigated worldwide for various applications [1–5].
One idea is to use such photochemical reactions to kill bi-
ological cells [6–9]. Under light irradiation, we found that
biological cells can be killed effectively when TiO2 parti-
cles (300 Å) are incorporated into the cells, or when the
biological cells are cultured on the surface of a TiO2-coated
substrate. Using this effect, we suggested that ultra fine TiO2
particles can be used for the treatment of skin tumors or skin
disease, and that TiO2-coated substrates can be used as an-
tibacterial materials.
The mechanism of cell death by the TiO2 photochemical
reaction involves photogenerated holes and reactive oxygen
such as OH·, O2−, and H2O2 [6,7]. The formation of these
reactive species from the photoexcited TiO2 is generally
considered to occur as follows. Illumination of TiO2 with a
photon energy greater than the band gap (3.23 eV = 384 nm
OH·, and the negative electron will reduce oxygen to O2
.
Therefore, the valence band edge and conduction band edge
respectively represent the oxidation site and the reduction
site of a photocatalytic TiO2 particle. As the OH· formed is
very reactive, and the valence band edge of TiO2 was re-
ported to be 2.6 V (vs SCE) [10], the reaction occurring at
the oxidation site is thought to be very strong and important.
In relation, O2− generated from the reduction site is thought
to be less reactive, and not so important in the photocatalysis
process [11,12]. On the other hand, the formation of H2O2
from a photocatalytic TiO2 particle was proposed through
various processes [13,14]. One view suggests that it is pro-
duced at the oxidation site; another theory proposes that it is
generated at the reduction site [15,16].
Further, the photocatalytic efficiency is a key issue to
overcome in consideration of various applications because
TiO2 is excited by UV light only. In our previous work, we
successfully monitored the electron transfer from photoex-
cited TiO2 particle to dissolved oxygen (O2), and then to an
*
Corresponding author.
E-mail address: chriscai@simtech.a-star.edu.sg (R. Cai).
0021-9517/$ – see front matter
2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00197-0

R. Cai et al. / Journal of Catalysis 219 (2003) 214–218
215
active center of superoxide dismutase (SOD) [16]. We made
positive use of O2−, the primary product from the reduction
site, by converting it into H2O2 and OH· in the presence of
SOD and the Fenton reagent [17]. As a result, the photo-
catalytic cell-killing effect was increased [18]. However, the
weak point in this case is that the photoexcited TiO2 will ox-
idize the bioenzyme, resulting in it being unable to sustain
such conversions.
of 4.7 µM H2O2 reduced Iarb to almost its half value, and
10 µM H2O2 resulted in Iarb being decreased to zero. Us-
ing a standard curve, an unknown H2O2 concentration can
be determined by measurement of the fluorescence peak in-
tensity of scopolectin.
3. Results and discussion
In this paper, instead of a biological enzyme, copper ions
were used to investigate the formation of H2O2. The amount
of H2O2 produced in the TiO2 suspension was measured
under oxygen-purged and nitrogen-purged conditions both
with and without electron acceptors. This was carried out
to clarify if H2O2 was generated from the reduction site or
the oxidation site of TiO2 particles. We also converted H2O2
into the most reactive OH· in the presence of ethylenedi-
aminetetraacetic acid disodium salt (EDTA)–Fe. As copper
ion in this study is more stable compared with the previously
used SOD enzyme, we believe that this is an effective way
to increase the photocatalytic efficiency.
We investigated a series of control experiments. They
are: (1) UV irradiation of solution without the addition of
TiO2; and (2) addition of copper ions to the solution with-
out UV irradiation. All these experiments showed no effect
on scopolectin fluorescence measurement. In the presence of
TiO2 particles, the formation of H2O2 was studied with and
without UV irradiation. With TiO2 particles in the dark, Iarb
was 100 as seen in the inset of Fig. 1 (curve 1a), where var-
ious fluorescence spectra of scopolectin solutions are shown
under different conditions. With TiO2 particles plus 5 min
UV irradiation, Iarb was decreased to 92 as indicated in
curve 1b. Further addition of 40 µM copper (II) into the
above solution (with exposure to UV for 5 min) reduced
Iarb to 10 (curve 1c). Using the relationship curve shown
in Fig. 1, we found that the amounts of H2O2 formed in ex-
periments 1b and 1c were 0.4 and 8.0 µM, respectively. As
no H2O2 was detected in the dark, we can conclude that the
2. Experimental
Formation of H2O2 was carried out as in the previous
study [17]. At first, 300 mg of TiO2 particles (p-25; Nip-
pon Aerosil Co., Tokyo, Japan) was dispersed in 3 ml of
distilled water. Copper chloride (CuCl2; copper (II)) solu-
tion was added into it, making final concentrations of 10,
20, 30, and 40 µM copper (II) solutions, respectively. Fenton
reagents were prepared by mixing ferrous sulfate (FeSO4;
Tokyo Kasei Co.) and EDTA (Tokyo Kasei Co.) and then
were adjusted to the targeted concentration. All above solu-
tions were freshly prepared throughout the experiment. The
pH for the final solution was adjusted to 4.5. Pure O2 (or N2)
gas was bubbled through the TiO2 suspension for 30 min
and sealed by O2 (or N2) overflow. It was then irradiated
for 5 min by UV light from a 500-W high-pressure Hg lamp
(Ushio Co., Tokyo, Japan). During irradiation, a UV pass fil-
ter (UVD2, Toshiba Co.) was used to transmit a wavelength
of 300–400 nm. After irradiation, the TiO2 particles were
removed by filtration through a 0.22-mm membrane filter,
and the H2O2 content of the filtered solution was measured.
The concentration of H2O2 was measured by a fluorescent
assay reported by Perschke and Broda [19]. In this method,
scopolectin dye was oxidized by H2O2 (H2O2 is a quencher
of scopolectin). The intensity of the fluorescence was de-
creased in the presence of H2O2. When 0.1 ml of scopolectin
solution (5 × 10−5 M) was added to various standard con-
centrations of H2O2 (without TiO2 particles), the intensity
of fluorescence at 460 nm from each sample was measured.
The relationship between the fluorescence intensity of the
scopolectin and the amount of H2O2 is shown in Fig. 1. One
can see that upon addition of H2O2 to the scopolectin solu-
tion, Iarb was decreased dramatically. This decrease shows
a linear drop from 1 to 10 µM H2O2. For instance, addition
Fig. 1. Fluorescence intensity of scopolectin at 460 nm as a function of
H O concentration. Curves 1a, 1b, and 1c in the inset represent the flu-
2
2
orescence spectra of scopolectin under different conditions. In this case,
three types of solutions were prepared. The first sample (curve 1a) was with
TiO in the dark. The second sample (curve 1b) was with TiO after 5 min
2
2
irradiation. The third sample (curve 1c) was with TiO and 40 µM cop-
2
per (II) ion added followed by further irradiation for 5 min. All these three
samples were used for scopolectin fluorescence measurement as described
under Experimental.

216
R. Cai et al. / Journal of Catalysis 219 (2003) 214–218
Fig. 2. The effect of copper (II) ion concentration on formation of H O
2
2
produced from TiO particles under 5 min UV irradiation.
2
H2O2 is generated from photocatalytic TiO2, and its forma-
tion was accelerated by the addition of a copper (II) solution.
Here, almost 20 times as much H2O2 was produced in the
presence of a 40 µM copper (II) solution.
Fig. 3. Schematic illustration of the photocatalytic formation of H O and
OH in the presence of copper (II) ion and Fenton reagent.
2
2
·
The effect of copper (II) concentration on formation of
H2O2 was investigated. For this purpose, various concentra-
tions of copper (II) solution were added into a TiO2 sus-
pension, and the previous experiment was repeated. When
copper (II) solutions were added at 10, 20, 30, and 40 µM,
the concentrations of H2O2 formed were 4.8, 5.8, 7.8, and
8.0 µM, respectively. The relationship between copper (II)
solution concentration and the amount of H2O2 is shown
in Fig. 2. The amount of H2O2 leveled off when the con-
centration of copper (II) solution reached the range of 30–
40 µM. Considering the amount of H2O2 generated at the
highest level, it is worth noting that the amount of H2O2 in-
duced by copper (II) solution in the present study is twice
as high as that achieved by using the bioenzyme catalyst,
SOD [17]. The efficiency of H2O2 production by SOD was
found to be sensitive to the environment. It was deactivated
by the photoexcited TiO2, resulting in a loss of enzyme func-
tion. Solution age and the used surrounding temperature also
largely affect SOD activity. In contrast, the copper (II) so-
lutions in this case are very stable. The reason for stability
is probably due to reduction of oxygen prevailing over the
reduction of copper (II) ion under the oxygen atmosphere.
Foster et al. described that on illumination of an oxygen-
purged solution, Cu(II) ions concentrations were increased
by release from Cu–TiO2 particles. On the other hand, upon
purging with nitrogen, Cu(II) ions in the supernatant were
deposited onto the surface of TiO2 [20]. Similarly, we found
that copper (II) was only deposited on TiO2 particles in the
presence of radical scavengers under nitrogen-purged solu-
tion, but not deposited under oxygen-purged solution (un-
published data). It is reasonable to assume that Cu(II) ions
in the present study are not reduced. Instead, the dissolved
oxygen is t−hought to react with the photogenerated electron
to form O2 (see Fig. 3).
Formation of H2O2 from TiO2 particles has been pro-
posed by many researchers [11–16]. Two alternative mecha-
nisms are proposed. One process involves oxidation of water
by photogenerated holes via reactions (1) and (2). In the
other process, dissolved oxygen is redu+ced by photogen-
erated electron via reaction (3). Here, hvb and ec−b are the
photogenerated hole in the
TiO2 → h+vb + ec−b
(1)
(2)
(3)
2H2O + 2h+vb → H2O2 + 2H+aq
O2 + 2e− + 2Ha+q → H2O2
cb
valence band and the electron in the conduction band, re-
spectively. These reactions are expected to occur on the TiO2
surface. To clarify which process contributes to the H2O2
formation in this study, we deaerated TiO2 suspension by us-
ing nitrogen gas and then measured the H2O2 amount in the
presence of copper (II) solution (silver ions from AgCl were
even used as electron acceptor). It is particularly surprising
that no H2O2 was detected in the deaerated TiO2 suspen-
sion. Also, when the TiO2 suspension was kept without UV
irradiation, no H2O2 was detected. These facts indicate that
oxygen and UV irradiation are both necessary for the forma-
tion of H2O2. Therefore, we can say that H2O2 was formed
from the reduction of dissolved oxygen via reaction (3) and
was not produced at oxidation sites through reactions (1)
and (2). This conclusion agrees well with our previous re-
sult, in which SOD was employed [17].
It is well known that reactive radicals such as O2−, H2O2,
and OH· are the main contributors to TiO2 photocatalytic
reactions. Among them, OH· is the most reactive. On the
other hand, EDTA–Fe is a well-known Fenton reagent for

R. Cai et al. / Journal of Catalysis 219 (2003) 214–218
217
16 times in the presence of copper (II) ion [21]. However,
if H2O2 alone were added, it was chemisorbed onto TiO2 at
low pH, to give a surface peroxo complex [27]. This surface-
adsorbed peroxo complex significantly inhibited photooxi-
dation of methyl orange [28]. Thus, one new finding from
the present study is to combine the H2O2-enhanced effect on
copper (II) ion with the Fenton reaction to achieve a higher
photocatalytic efficiency.
In the past, rare metals (platinum, silver, palladium) and
copper have generally been deposited onto the surface of
TiO2 particle in order to increase the photocatalytic effi-
ciency [29,30]. In this case, the deposited metal acts as an
electron acceptor, resulting in efficient charge separation, to
intensify the hole and OH· effect. In the present study, as
copper ions will not be reduced under oxygen-purged condi-
tions, we can say that the role of the copper ions is different
from that of deposited copper.
At this stage, it is clear that H2O2 production is acceler-
ated by copper (II) ions, but the mechanism by which the
copper (II) ion catalyzes H2O2 production is still unclear.
We assume that it may proceed as with SOD through the
following reactions:
O2− + Cu(II) → O2 + Cu(I)
(5)
(6)
Fig. 4. Fluorescence spectra of scopolectin under different conditions:
O2− + 2H+ + Cu(I) → H2O2 + Cu(II)
curve 4a, TiO particle alone; curve 4b, TiO + 40 µM copper (II) ion +
2
2
20 µM EDTA–Fe + 5 min irradiation; curve 4c, TiO + 40 µM copper (II)
2
ion + 5 min irradiation.
Indeed, we found that pH changes during reaction and con-
firmed that reactions (5) and (6) are possible as the pathway.
In another experiment, when oxygen was totally removed,
we found that the TiO2 photocatalytic oxidation of H2S gas
was almost stopped (unpublished data). This fact indicates
that oxygen is a key component in the photochemical reac-
tion. In the past, Heller and co-workers studied the role of
oxygen in the photooxidation and reported that the rate of
photooxidation of organic compounds on the surface of a
catalyst is limited by the rate of electron transfer to the oxy-
gen [2]. Here, we show that various types of reactive oxygen
are produced from photocatalytic TiO2 particles under an
oxygen atmosphere, especially in the presence of copper (II)
ion and EDTA–Fe together. Less reactive species can con-
vert into more reactive ones, which in turn oxidize organic
substances. Like this, controlling the conversion of radicals
generated from the reduction site can alter the rate of the
photocatalytic reaction. Our experimental evidence suggests
that chemical reactions occurring at the reduction site of
TiO2 particles are equally important as those occurring at
the oxidation site.
converting H2O2 into OH· via reaction (4) [21], and its effect
in the
H2O2 + Fe2+ → OH· + OH− + Fe3+
(4)
TiO2/copper (II) ion system was further studied. In Fig. 4,
with TiO2 particles in the dark, no H2O2 was formed and
Iarb was 100. When 40 µM copper (II) solution was added
to TiO2 suspensions under UV irradiation, Iarb was de-
creased to 10 (curve 4c in Fig. 4). However, the intensity
of Iarb recovered up to 58 upon addition of 20 µM EDTA–Fe
(curve 4b in Fig. 4). This suggests that H2O2 was used as a
source to generate OH· via reaction (4). If more OH· is pro-
duced, higher photocatalytic efficiency is expected. In fact,
it was found that the rate of cell killing was increased−3-fold
by using SOD and the Fenton reagent to convert O2 into
H2O2, and H2O2 into OH·, respectively [18]. To increase the
TiO2 photocatalytic efficiency, some researchers have added
H2O2 into TiO2 suspension [22–24]. For example, photoca-
talytic oxidation of phenol was reported to be significantly
accelerated by addition of H2O2 together with transition-
metal ions such as Fe3+ and Cu2+ [22,23]. They reported
that Cu2+ did not cause the intensive short-circuiting prob-
lem which usually happens in the photocatalytic oxidation
process when adding transition ions alone [24]. Walling et al.
described that H2O2 was decomposed into hydroxyl radicals
by UV irradiation [25]. This conversion was also acceler-
ated in the presence of ferric ions [26]. Similarly, Fujihira
et al. reported that the quantum yield of photocatalytic oxi-
dation of aromatic compounds by TiO2 was increased about
4. Conclusions
The results in this study are s−ummarized in Fig. 3. In the
presence of copper (II) ions, O2 generated from reduction
sites on TiO2 particles was effectively converted into H2O2.
The H2O2 formed was also converted into a highly reactive
OH· radical in the presence of Fenton reagent. Our exper-
iment also suggests the possibility that Cu(I) can also be

218
R. Cai et al. / Journal of Catalysis 219 (2003) 214–218
formed by reduction by the conduction band electrons of
irradiated TiO2 in competition with reduction of molecular
oxygen to give the superoxide radical anion. Compared with
biological enzyme, the copper (II) ion used in this study is
more stable. We believe that such a system can be potentially
applied to various applications where TiO2 photocatalysts
are currently being employed for destruction of organic pol-
lutants in water.
[10] H. Gerischer, in: B.O. Seraphin (Ed.), Solar Energy Conversion, in:
Topics in Applied Physics, Vol. 31, Springer, Berlin, 1979.
[11] J.R. Harbour, J. Tromp, M.L. Hair, Can. J. Chem. 63 (1985) 204.
[12] A.P. Hong, D.W. Bahnemann, M.R. Hoffmann, J. Phys. Chem. 91
(1987) 2109.
[13] See for example, M.V. Rao, K. Rajeshwar, V.R.P. Verneker, J. Dubow,
J. Phys. Chem. 84 (1980) 1987.
[14] H. Muraki, T. Saji, M. Fujihira, S. Aoyagui, J. Electroanal. Chem. 169
(1984) 323.
[15] M.L. Hair, J.R. Habour, Adv. Chem. Ser. 184 (1980) 173.
[16] R. Cai, R. Baba, K. Hashimoto, Y. Kubota, A. Fujishima, J. Electro-
anal. Chem. 360 (1993) 237.
References
[17] R. Cai, K. Hashimoto, Y. Kubota, A. Fujishima, J. Electroanal. Chem.
326 (1992) 345.
[18] R. Cai, K. Hashimoto, Y. Kubota, A. Fujishima, Chem. Lett. (1992)
427.
[19] H. Perschke, E. Broda, Nature 190 (1961) 257.
[20] N.S. Foster, D.N. Richard, A.K. Carl, Environ. Sci. Technol. 27 (1993)
350.
[21] M. Fujihira, Y. Satoh, T. Osa, Nature 293 (1981) 206.
[22] T. Wei, C. Wan, J. Photochem. Photobiol. A 69 (1992) 241.
[23] K. Tanaka, T. Hisanaga, K. Harada, J. Photochem. Photobiol. A 48
(1989) 155.
[1] R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Ki-
tamura, M. Shimohigoshi, T. Watanabe, Nature 388 (1997) 431.
[2] C.M. Wang, A. Heller, H. Gerisher, J. Am. Chem. Soc. 114 (1992)
5230.
[3] A. Fujushima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C 1
(2000) 1.
[4] R.Q. Long, R.T. Yang, J. Catal. 207 (2002) 158.
[5] D.F. Ollis, H. Al-Ekabi, Photocatalytic Purification and Treatment of
Water and Air, Elsevier, Amsterdam, 1993.
[24] T. Wei, Y. Wang, C. Wan, J. Photochem. Photobiol. A 55 (1990) 115.
[25] C. Walling, Acc. Chem. Res. 8 (1975) 125.
[26] F. Haber, J. Weiss, J. Proc. R. Soc. A 147 (1934) 332.
[27] A.H. Boonstra, C.A.H.A. Mutsaers, J. Phys. Chem. 79 (1975) 1940.
[28] T.B. Graham, R.D. James, J. Phys. Chem. 88 (1984) 4955.
[29] W.W. Dunn, Y. Aikawa, A.J. Bard, J. Am. Chem. Soc. 103 (1981)
6893.
[6] R. Cai, Y. Kubota, T. Shuin, K. Hashimoto, A. Fujishima, Cancer
Res. 52 (1992) 2346.
[7] R. Cai, K. Hashimoto, Y. Kubota, A. Fujishima, Bull. Chem. Soc. Jpn.
64 (1991) 1268.
[8] A. Fujishima, R. Cai, K. Hashimoto, Y. Kubota, Electrochim. Acta 38
(1993) 153.
[9] H. Sakai, E. Ito, R. Cai, T. Yoshioka, Y. Kubota, K. Hashimoto, A. Fu-
jishima, Biochem. Biophys. Acta 1201 (1994) 259.
[30] J.C.S. Wu, C.Y. Yeh, J. Mater. Res. 615 (2001) 16.