Photooxidation pathway of sulforhodamine-B. Dependence on the adsorption mode on TiO2 exposed to visible light radiation

Article abstract of DOI:10.1021/es001064c

The temporal course of the photooxidation of sulforhodamine-B (SRB) in aqueous media illuminated by visible wavelengths in the presence of TiO₂ has been examined to determine the nature of the intermediate species produced and to explore the operative reaction pathway(s). Two pathways are described to account for the differences in the final photooxidation products whose nature depends on the different modes of adsorption of the dye on the metal-oxide mediator. In the SRB/TiO₂ system, when SRB is adsorbed on the positively charged TiO₂ particle surface through a sulfonate group cleavage of the SRB chromophore structure predominates and N-de-ethylation occurs only to a slight extent with the major photooxidation products being diethylamine and carbon dioxide. In the presence of the anionic dodecylbenzenesulfonate surfactant DBS, when SRB is near the negatively charged DBS/TiO₂ interface through the positive diethylamine group N-de-ethylation occurs preferentially before destruction of the structure with the major products being acetaldehyde and carbon dioxide. The temporal course of the photooxidation of sulforhodamine-B (SRB) in aqueous media illuminated by visible wavelengths in the presence of TiO₂ has been examined to determine the nature of the intermediate species produced and to explore the operative reaction pathway(s). Two pathways are described to account for the differences in the final photooxidation products whose nature depends on the different modes of adsorption of the dye on the metaloxide mediator. In the SRB/TiO₂ system, when SRB is adsorbed on the positively charged TiO₂ particle surface through a sulfonate group cleavage of the SRB chromophore structure predominates and N-de-ethylation occurs only to a slight extent with the major photooxidation products being diethylamine and carbon dioxide. In the presence of the anionic dodecylbenzenesulfonate surfactant DBS, when SRB is near the negatively charged DBS/TiO₂ interface through the positive diethylamine group N-de-ethylation occurs preferentially before destruction of the structure with the major products being acetaldehyde and carbon dioxide.

Full text of DOI:10.1021/es001064c

Environ. Sci. Technol. 2000, 34, 3982-3990
(5, 6). However, the photocatalytic activity of TiO2 is limited
Photooxidation Pathway of
to irradiation wavelengths in the UV region, thereby limiting
effective utilization of solar energy to about 3-5% of the
total solar spectrum.
Sulforhodamine-B. Dependence on
the Adsorption Mode on TiO
Exposed to Visible Light Radiation
Textile and industrial dyes are becoming one of the largest
groups of chemicals manufactured and of consequence are
causing significant, albeit progressive pollution to ground-
waters in China and the rest of the world (7, 8). Effective
utilization of visible sunlight to degrade colored compounds
(e.g. dyestuffs) in the presence of TiO2 particulates promises
to be a novel and favorable approach with significant
implications to energy resources and environmental reme-
diation (9-13). The adsorbed dyes are excited by visible light
to produce excited states, 1,3dye*, which under suitable
circumstances are poised to inject electrons into the semi-
2
G U A N G M I N G L I U , X I A N G Z H O N G L I , A N D
J I N C A I Z H A O *
The Laboratory of Photochemistry, Center for Molecular
Science, Institute of Chemistry, The Chinese Academy of
Sciences, Beijing 100080, China
H I S A O H I D A K A
conductor particle to generate the dye radical cation, dye•+
.
Subsequent transformation of dye•+ results in the destruction
of the dye chromophore structure and in the formation of
small decomposed fragments after a series of complex
oxidative steps.
Frontier Research Center for the Earth Environment
Protection, Meisei University, 2-1-1 Hodokubo, Hino,
Tokyo 191, Japan
N I C K S E R P O N E
This study examines the photooxidative transformation
of the laser dye sulforhodamine-B (SRB) chosen as a prototype
molecule because of its high solubility in water and its strong
adsorption on the TiO2 particle surface owing to two possible
functions which interact with the positive particle surface
under the prevailing conditions used, namely the diethyl-
amine group (positively charged) and the sulfonate group
(negatively charged) located at different ends of the dye
molecule. Consequently, with the local molecular microen-
vironment of the SRB dye, adsorption on the TiO2 surface
can easily be changed by simple control of the experimental
conditions. The process of photooxidation of SRB in aqueous
TiO2 dispersions under visible illumination is rather complex
involving many aspects, such as (i) electron transfer occurring
between the excited dye and the conduction band of TiO2,
(ii) formation of radical cations after the electron transfer
event, and (iii) hydrolysis or deprotonation of the dye radical
cations, among others. These processes are usually affected
by the surrounding environment and are closely related to
the mode of adsorption. Consequently, the different adsorp-
tion modes of SRB are likely to manifest differences in the
photooxidative pathways. In this study, UV-visible, CODCr
Department of Chemistry and Biochemistry,
Concordia University, Montreal, Canada H3G 1M8
The temporal course of the photooxidation of sulforho-
damine-B (SRB) in aqueous media illuminated by visible
wavelengths in the presence of TiO has been examined to
determine the nature of the intermediate species produced
and to explore the operative reaction pathway(s). Two
pathways are described to account for the differences in
the final photooxidation products whose nature depends
on the different modes of adsorption of the dye on the metal-
2
oxide mediator. In the SRB/TiO system, when SRB is
adsorbed on the positively charged TiO particle surface
2
2
through a sulfonate group cleavage of the SRB chromophore
structure predominates and N-de-ethylation occurs only
to a slight extent with the major photooxidation products
being diethylamine and carbon dioxide. In the presence of
the anionic dodecylbenzenesulfonate surfactant DBS,
1
measurements, H NMR, infrared, and GC/ MS techniques
when SRB is near the negatively charged DBS/TiO interface
2
were brought to bear on defining the mechanism(s) of
photooxidation of SRB in aqueous TiO2 dispersions under
visible light illumination. We show that there are two different
photooxidation pathways, each of which leads to different
photooxidation products; these arise from the two different
adsorption modes of the substrate on the metal-oxide
mediator.
through the positive diethylamine group N-de-ethylation
occurs preferentially before destruction of the structure with
the major products being acetaldehyde and carbon
dioxide.
Introduction
Experimental Section
The photooxidation of organic compounds on TiO2 particle
surfaces has been examined extensively and is being exploited
widely to mineralize a variety of environmentally harmful
organic compounds. Typically, the process is initiated by
band-to-band excitation of the TiO2 particles by ultraviolet
radiation (1-4) to generate •OH radicals derived from valence
band hole oxidation of terminal OH- groups and water of
hydration on the particle surface. These radicals have been
postulated as the dominant oxidizing agents which appear
to control the overall kinetics of the oxidative process
Materials. P-25 TiO2 (ca. 80% anatase, 20% rutile; BET area,
ca. 50 m2 g-1) was kindly supplied by Degussa Co. The spin-
trap reagent 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) was
procured from Sigma Chemical Co. The dye sulforho-
damine-B (SRB) was of laser grade quality (Across Co.). The
surfactant sodium dodecylbenzenesulfonate (DBS) and all
other chemicals were of analytical reagent grade quality and
were employed without further purification. Deionized and
doubly distilled water was used throughout. The pH of the
solution was adjusted with either dilute HCl or NaOH (the
original pH of the dye solution was ca. 4.0). The structure of
the SRB molecule is illustrated below; the small letters identify
the protons for NMR identification.
* Corresponding author fax: +86-10-6487-9375; e-mail:
jczhao@ipc.ac.cn. Present address: Institute of Photographic Chem-
istry, The Chinese Academy of Sciences, Beijing 100101, China.
9
3 9 8 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 18, 2000
10.1021/es001064c CCC: $19.00
2000 Am erican Chem ical Society
Published on Web 08/18/2000

0.1 M LiClO4 dried overnight in a vacuum desiccator. Cyclic
voltammograms were obtained on the stationary Pt electrode
using an EG&G potentiostat Model 273. The solution in the
cell containing 1.0 × 10-3 M SRB was flushed with high-
purity nitrogen; temperature was 301 K. The redox potential
of the SRB*/ SRB•+ couple was estimated from E(SRB*/ SRB•+
) E(SRB/ SRB•+) - E*.
The intermediate and final products generated during
the photooxidative process were analyzed by gas chroma-
tography/ mass spectrometric methods (GC-MS, Trio-2000,
equipped with a BPX 70 column, size 28 m × 0.25 mm).
Results and Discussion
1. UV-Visible Spectra. The prevailing pH of the solutions
can affect the mode and extent of adsorption of SRB on the
TiO2 surface and thus, indirectly, the rate of transformation
of the SRB dye; this arises because pH greatly influences the
surface charge properties of the metal-oxide particles. The
TiO2 surface is positively charged in acid media (pH below
ca. 6.8), whereas it is negatively charged under alkaline
conditions (pH greater than 6.8) as evidenced by ú-potential
measurements (pITiO2 ) 6.8) (15). Both the adsorption of SRB
and the initial rate of SRB conversion in the SRB/ TiO2 system
decreased with an increase in the pH of the dispersion in
alkaline media, whereas in acidic media they increased
simultaneously with decreasing pH. To better scrutinize the
adsorption modes and the photooxidation pathway(s), all
the experiments were carried out at pH 2.5 (unless noted
otherwise) at which the rate of conversion was greater than
at the higher pHs.
The UV-vis spectral changes taking place during the
photooxidation of SRB in the SRB/ TiO2 system are illustrated
in Figure 1a. The extent of adsorption of the dye (2 × 10-5
M) on TiO2 (100 mg) particles was ca. 25% (see spectrum 2).
Under visible light irradiation, the characteristic absorption
band of SRB at ca. 565 nm decreased rapidly, accompanied
by concomitant slight hypsochromic shifts indicating the
occurrence of some N-de-ethylation during the photooxi-
dation of SRB. These observations are in keeping with similar
earlier observations of an electron-transfer reaction in the
Rhodamine-B/ CdS system (16). The color of the dispersion
disappeared indicating that at least the chromophoric
structure of the dye was destroyed. Absorption spectra
revealed no new intermediates or products formed that would
have manifested new absorption features in the visible and
near-ultraviolet region. Control experiments established that
SRB did not degrade in TiO2 suspensions in the dark or when
illuminated with visible light in the absence of TiO2. Visible
light illumination of SRB in the presence of SiO2 or Al2O3
particles, substituted for TiO2 particulates, caused no dye
degradation. We infer from this that both visible light and
TiO2 semiconductor particles are indispensable to the
photooxidative conversion of SRB.
Photoreactor and Light Source. The visible light source
was a 500 W halogen lamp (Institute of Electric Light Source,
Beijing) positioned inside a cylindrical Pyrex vessel and
surrounded by a circulating water jacket (Pyrex) to cool the
lamp and maintain temperature ambient at 298 K and to
block unwanted IR radiation. An appropriate cutoff filter was
also placed outside the Pyrex jacket to ensure complete
removal of radiation below 420 nm and to ensure that
irradiation of the SRB/ TiO2 system occurred only by visible
light wavelengths. The photon flux of the 500 W halogen
lamp at 560 nm was 3.4 × 10-8 einstein s-1 as reported earlier
(13c).
Procedures and Analyses. Aqueous suspensions of SRB
(usually 50 mL, 2 × 10-5 M) and 100 mg of TiO2 particulates
were placed in a Pyrex vessel. Prior to irradiation, the
suspensions were magnetically stirred in the dark for ca. 30
min to establish an adsorption/ desorption equilibrium
between the dye and the TiO2 particle surface. At given
intervals of illumination, a sample of the TiO2 particulates
was collected, centrifuged, and then filtered through a
Millipore filter (pore size 0.22 µm). The filtrates were analyzed
by UV-vis spectroscopy using a Lambda Bio 20 spectro-
photometer. Assays of the chemical oxygen demand (CODCr
)
was carried out using the potassium dichromate titration
method (14). The CODCr of the suspensions (50 mL of 5 ×
10-5 M SRB, TiO2 loading100 mg) was measured directly
without removal of the TiO2 particles at various irradiation
time intervals. Ion chromatography (Shimadzu LC-10AS) was
used to determine the quantity of HCOO- ions formed during
the photooxidation. Electron paramagnetic resonance (EPR)
measurements were carried out at ambient temperature (298
K) using a Bruker ESP 300E spectrometer operating at a
microwave frequency of 9.82 GHz and a power level of 5.05
mW; the center field was 3486.7 G and the sweep width was
100.0 G. The modulation amplitude was 1.814 G. Irradiation
at λ ) 532 nm was secured from a Quanta-Ray Nd:YAG pulsed
(10 Hz) laser system.
The inset to Figure 1a shows the temporal concentration
changes of SRB illustrating a first-order exponential fit of the
decay with k ) 8.8 × 10-2 min-1. That rapid decomposition
occurred is consistent with the notion that relatively strong
adsorption enhances photooxidation on the TiO2 surface
(12b). The tendency of the SRB maximum absorption to blue
shift is also illustrated in the inset of Figure 1a. Two
competitive processes, namely N-de-ethylation and cleavage
of the SRB chromophore ring structure, occur concurrently
during photooxidation with the latter process predominating
(e.g. compare part a with part b of Figure 1).
When the anionic surfactant DBS is added to the SRB/
TiO2 system, de-ethylation predominates during the initial
stages (Figure 1b). Destruction of the chromophore ring
structure of the fully de-ethylated dye occurred only after ca.
60 min of visible irradiation. Moreover, only 19% of the fully
de-ethylated dye was converted after ca. 240 min of further
Proton NMR spectra were obtained with a Varian 300
nuclear magnetic resonance spectrometer. Samples were
prepared as follows: several dispersions containing 200 mL
of the SRB dye (1 × 10-4 M) and 400 mg of TiO2 were irradiated
at different time intervals, following which the TiO2 particles
were removed by centrifugation and filtration as described
above. Subsequently, the solvent of the filtrate was removed
under reduced pressure (below 323 K). The remaining residue
was dissolved in 0.5 mL of D2O. Samples for infrared (IR)
spectra (FTS-165 spectrophotometer) were also prepared by
a method similar to that for proton NMR, except that the
residue was used directly.
The electrochemical redox potential of the SRB/ SRB•+
couple was measured in water-free acetonitrile media
(distilled over P2O5) using a Pt working electrode and a Pt
counter-electrode; the reference electrode was Ag/ AgCl
(+0.22 V vs NHE), whereas the supporting electrolyte was
9
VOL. 34, NO. 18, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 8 3

FIGURE 1. (a) UV-visible spectral changes of SRB (2 × 10-5 M, 50 mL) in aqueous TiO dispersions as a function of irradiation time. Spectra
2
-1
from top to bottom correspond to cases where TiO was absent (0 time), after addition of TiO particles (2 g L ) and achievement of an
2
2
equilibrium concentration (0 time), followed by irradiation for 5, 10, 15, 20, 25, and 30 min, respectively. Inset: SRB concentration changes
and wavelength shifts of the maximal absorption as a function of with irradiation time. (b) UV-visible spectral changes of SRB (2 × 10-5
M, 50 mL) in aqueous TiO /DBS (cmc) dispersions with irradiation time. Spectra from right to left refer to those before addition of TiO
2
2
-1
particles; after addition of TiO particles (2 g L ) and DBS (1.2 mM) to reach an equilibrium concentration; then followed by irradiation
2
for 10, 20, 30, 40, 60, and 300 min, respectively. Inset: SRB concentration changes and wavelength shifts of the major absorption band
with irradiation time.
irradiation (see inset to Figure 1b). Under visible illumination,
the dye is de-ethylated in a stepwise manner with the color
of the dispersion changing from an initial red color to a light
green-yellow (i.e. ethyl groups are removed one by one as
confirmed by the gradual peak wavelength shifts toward the
blue region, see Figure 1b). The fully de-ethylated SRB
molecule has a major absorption band at λmax ) 506 nm, in
accord with standard spectra. As expected, the eliminated
ethyl groups were oxidized to acetaldehyde (see other results
below).
2. Chem ical Oxygen Dem and. Changes in the chemical
oxygen demand (CODCr) of the dispersions reflect the degree
to which the dye (or any organic substance) is degraded or
mineralized during irradiation. Table 1 summarizes the
results of CODCr measurements. CODCr values for the
irradiated SRB/ TiO2 suspensions decreased with increasing
illumination time from an initial value of 42.0 mg O2 L-1
(before illumination) to 11.6 mg O2 L-1 after 12 h of irradiation.
Such significant CODCr changes (72.4%) indicate unambigu-
ously that photooxidation of SRB in the presence of TiO2
under visible light irradiation is not a simple discoloration
or a simple destruction of the chromophoric structure of the
dye, because a large number of carbon atoms in the dye
We focus first on the photooxidation process of SRB in
the SRB/ TiO2 system to elucidate the mechanistic details of
the process under visible light excitation.
9
3 9 8 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 18, 2000

TABLE 1. COD Changes of SRB (50 mL, 5 × 10-5 M) Dispersions (pH ) 4.0; TiO Loading, 100 mg) with Irradiation Time
Cr
2
irradiation time (h)
0
1
3
5
7
11
12
14
CODCr
[COD0-CODt]/COD0, %
42.0
0.0
34.8
17.1
31.2
25.7
21.6
48.6
17.2
59.0
13.2
68.6
11.6
72.4
11.6
72.4
FIGURE 3. Temporal variations in the IR spectra in the photodeg-
radation of SRB under visible light irradiation.
of the aromatic hydrogens Hc, Hd, He, Hf, Hg, and Hh are
located at δ 6.66, 6.83-6.86, 6.96-6.99, 7.62-7.65, 8.25-
8.28, and 8.52 ppm, respectively; those of Ha and Hb of the
N-diethyl group appear at δ 1.18-1.23 and 3.49-3.60 ppm,
respectively. During the course of the photooxidation (spectra
a-d in Figure 2a), a series of new signals appeared at δ 1.30-
1.50 ppm (hydrogens from the CH3 and analogous groups),
2.04 ppm, 3.10-3.30 ppm, and 8.46 ppm, whereas the
characteristic signals of SRB at δ 1.18-1.23 ppm (CH3), δ
3.49-3.60 ppm (CH2), and δ 6.60-8.60 ppm (aromatic
protons) disappeared. The signals at δ 8.46 ppm and δ 2.04
ppm increased with increasing illumination time. We at-
tribute the signal at δ 8.46 ppm to the CH proton of formic
acid (confirmed by ion chromatography) and the signal at
δ 2.04 ppm to the methyl protons of CH3CHO (also see the
IR and GC-MS results). Of particular relevance are the proton
signals at δ 1.30-1.50 ppm (new) and δ 3.10-3.30 ppm (new),
together with the methyl (CH3) protons at δ 1.18-1.23 ppm
and the methylene (CH2) protons at δ 3.49-3.60 ppm of SRB,
all of which exhibit peculiar features. The intensities of the
two new signals increased with irradiation time, whereas the
latter two signals decreased in intensity with time. More
interesting is the ratio of the integrated area of the first two
peaks which remained constant at 3:2 during the photo-
oxidation, the same as the ratio of the latter two signals. This
observation demonstrates that the ethyl group (CH2CH3) of
SRB is not destroyed in the photooxidation process but rather
exists in the final photooxidation products in the form of
diethylamine (actually, as diethylamine hydrochloride due
to a large quantity of chloride present from adjusting the pH
of the original solutions with HCl; also see results in Figure
3 and in Figure 4a below).
On the basis of the signal intensities from proton NMR
spectra (spectrum d of Figure 2a), we establish that diethyl-
amine is a major photooxidation product in the SRB/ TiO2
system (also see the GC-MS results, Figure 4a). However, the
NMR signals of the NH or NH2 functions of diethylamine
were not observed in this system owing to proton exchange
between the active proton and the solvent used, D2O. To
further confirm the nature of the principal product, we used
DMSO as the solvent in lieu of D2O (Figure 2b, spectrum a;
same sample as used to record spectrum b of Figure 2a,
except for the solvent). The broad peak appearing at δ 9.10-
9.20 ppm arises from the nitrogen proton of diethylamine
hydrochloride. Because of the nature of DMSO, the signals
1
FIGURE 2. (a) Temporal profile of the H NMR spectra at various
irradiation intervals during the photooxidation of sulforhodamine-
B. (The strong signal at δ 4.85 ppm is that of an impurity, probably
water, in the D O solvent). (b) (a) 1H NMR spectrum obtained after
4 h of irradiation of the SRB/TiO dispersion (solvent is DMSO instead
of D O); (b) spectrum obtained by reduced-pressure distillation of
the sample used to record spectrum a under alkaline conditions;
(c) spectrum obtained by subtracting spectrum b from spectrum a;
(d) spectrum of an authentic diethylamine hydrochloride sample.
2
2
2
molecule are converted to CO2. Details of the process were
further elucidated by H NMR, IR, and GC/ MS methods.
1
3. Proton NMR Spectra. To identify the species produced
in the photooxidation of SRB, the temporal proton NMR
profiles of the dye, intermediates, and final products were
monitored by proton nuclear magnetic resonance spectros-
copy. Spectrum a in Figure 2a shows the typical proton NMR
signals of pure SRB and their assignments to the various
protons in the structure of SRB (see above). The NMR signals
9
VOL. 34, NO. 18, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 8 5

FIGURE 4. (a) Mass spectral data of the major GC component A. Inset: Gas chromatogram of the final photooxidation products of SRB
for the adsorption mode involving a sulfonate group. (b) Mass spectral data of the major GC component A′. Inset: Gas chromatogram of
the final photooxidation products of SRB for the adsorption mode whereby a diethylamine group is involved in adsorption of SRB to the
TiO surface.
2
of the methyl and methylene protons of SRB were observed
at δ 1.10-1.30 ppm and δ 2.70-2.90 ppm, respectively,
whereas the signals of the corresponding methyl protons of
diethylamine hydrochloride appeared at δ 1.18 ppm, over-
lapping with those of SRB; the signals of the methylene group
protons of diethylamine hydrochloride appeared at δ 3.62
ppm, a location different from that of the corresponding
group in the diethylamine/ D2O system.
Distillation of an irradiated aqueous SRB/ TiO2 sample
under alkaline conditions (pH ) 10.5) led to removal of
diethylamine as evidenced by the loss of the corresponding
proton NMR signals at δ 9.10-9.20 ppm, δ 3.62 ppm, and δ
1.18 ppm (Figure 2b, spectrum b). The proton NMR spectrum
of diethylamine hydrochloride (spectrum c) can be obtained
by subtracting spectrum b from spectrum a to give a spectrum
identical with that of an authentic sample of diethylamine
hydrochloride (spectrum d).
The above NMR findings support the suggestion made
previously that the breakup of the chromophoric structure
of SRB dominates during the initial irradiation stage (forma-
tion of a large quantity of diethylamine) with only a small
extent of de-ethylation taking place as attested by formation
of a small quantity of CH3CHO. This is consistent with the
results of Figure 1a and with the adsorption mode (Scheme
1a) whereby the SRB dye is adsorbed on the TiO2 surface
through a sulfonate group.
4. Infrared Spectra. The IR spectra illustrated in Figure
3 were also used to monitor the temporal course of the
photoinduced conversion of SRB illuminated by visible
irradiation and to provide additional evidence for the
operational pathway(s). Results available in the literature
(17) suggest the following assignments for the principal bands
in the IR spectra of SRB (before irradiation): the bands at
1590, 1558, 1530, 1510, 1490, and 1470 cm-1 correspond to
9
3 9 8 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 18, 2000

SCHEME 1. Proposed Photooxidation Pathways of SRB in Aqueous TiO Dispersions under Visible Light Irradiation: (a)
2
Adsorption Mode with a Sulfonate Group and (b) Adsorption Mode with a Diethylamine Group
a
b
aromatic ring vibrations, whereas the 1344 cm-1 band is due
to C-aryl bond vibrations. The bands at 1120-1145 cm-1 and
668-625 cm-1 are caused by vibrations in the -SO3- group,
whereas the band at 1649 cm-1 is attributed to vibrations of
the carbon-nitrogen bond and the heterocycle vibrations
are the origins of the band at 1530-1558 cm-1. During
photooxidation, bands of vibrations characteristic of the
carbon-nitrogen bond (1649 cm-1), the C-aryl bond (1344
cm-1), the aromatic ring, and heterocycle vibrations (1470-
1590 cm-1) decreased with irradiation time and disappeared
after about 17 h of irradiation. At the same time, a strong
new IR band appeared at 1720 cm-1 attributable to >CdO
groups. The IR results indicate convincingly that the large
conjugated chromophore structure of SRB is destroyed under
visible light irradiation in aqueous media through the
mediation of TiO2 particulates and is further decomposed to
9
VOL. 34, NO. 18, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 8 7

a
TABLE 2. Positive Ion Mass Spectra of the Final Photooxidation Products of SRB
A
B
C
D
E
F(C′)
G(A′)
B′
D′
[M]+
73(20%)
87(28%)
73(38%)
115(24%)
101(35%)
46(60%)
45(50%)
44(100%) 45(100%) 59(100%)
[M - H]+
43(55%)
29(85%)
44(25%)
29(30%)
[M - CH3]+
[M - NH2]+
[M - OH]+
[M - C2H5]+
[M - C2H5N]+
[M - C3H7]+
[M - C3H8]+
[M - C3H7N]+
[M - C4H9]+
[M - C4H9N]+
[M - C4H10N]+
58(95%)
72(8%)
58(15%)
100(5%)
86(15%)
44(75%)
43(65%)
29(100%)
44(30%)
30(100%)
44(22%)
30(100%)
86(4%)
72(8%)
72(17%)
58(63%)
43(70%)
30(100%)
58(100%)
43(66%)
44(42%)
30(100%)
a
The data in brackets represent relative intensities.
smaller organic species containing >CdO groups: e.g.
N-ethylacetam ide, N-ethylform am ide, N,N-diethylacet-
amide, N,N-diethylformamide, HCOOH, and CH3CHO, as
confirmed by the proton NMR evidence above and by the
GC-MS results below. In addition, a series of new IR
absorption bands belonging to diethylamine hydrochloride
appeared at 3000, 2500, 1600, 1450, 1400 cm-1, in accord
with standard IR spectra of an authentic sample.
6. Initial Photooxidation Pathway. After photoexcitation
of the SRB/ TiO2 system at λ > 420 nm, injection of an electron
from the excited dye (1,3dye*) to TiO2 takes place since the
thermodynamics are appropriate {E(SRB*/ SRB•+) ) -1.30 V
vs NHE and ECB (TiO2) ) -0.5 V vs NHE (6)}; the TiO2(e-) is
then scavenged by preadsorbed molecular oxygen to yield
active oxygen radicals and the cation radicals SRB•+. The
role of active oxygen radicals produced in the photooxidation
of SRB was also examined.
5. GC-MS Measurem ents. Further convincing evidence
for the operational pathway(s) was obtained by GC-MS
spectroscopic methods for the two different systems (see
Scheme 1 below), which would likely lead to two different
photooxidation pathways as inferred from the results dis-
played in Figure 1a,b. The GC-MS results of the final
photooxidation products for the SRB/ TiO2 system are
presented in Figure 4a. The predominant peak A in the GC
chromatogram (inset to Figure 4a) was analyzed in detail by
mass spectroscopy. The major component A corresponds to
diethylamine (i.e. diethylamine hydrochloride) in accord with
results from proton NMR and IR spectroscopy. Mass spectral
analyses confirmed that the other (minor) final products were
N-ethylacetamide (B), N-ethylformamide (C), N,N-diethyl-
acetamide (D), N,N-diethylformamide (E), formic acid (F),
and acetaldehyde (G). The major features of the positive ion
mass spectra are summarized in Table 2. Because the
amounts of components B-E are relatively small compared
to Aand because the positions of their NMR signals are similar
to those of A, their identification was precluded by the 1H
NMR spectral method (Figure 2).
The presence of B and C indicates that de-ethylation is
incomplete, whereas formation of A (main product), D, and
E demonstrates that destruction of the chromophore struc-
ture is the predominant process in the photooxidation of
SRB, in accord with the adsorption mode depicted in Scheme
1a (see below). More important, according to the relative
intensity of A-G in the gas chromatograms it is possible to
predict the approximate theoretical changes in the COD
values of SRB during its photooxidation (i.e. from the SRB
molecule to the final photooxidation products). We obtain
ca. 65.4%, in relatively good agreement with the experimental
results from COD measurements (72.4%, see Table 1).
Figure 4b illustrates the GC-MS results of the final products
of SRB photooxidation in the SRB/ DBS/ TiO2 system (Scheme
1b). From the results of mass spectral analysis, we confirmed
that the major component A′ in the gas chromatograms is
acetaldehyde (the de-ethylated product, Figure 4b); the other
(minor) components are B′, formamide; C′, formic acid; and
D′, acetamide. The major features of the positive ion mass
spectra are summarized in Table 2. The presence of the
products A′, B′, and D′ and the absence of A-E indicate that
the N-de-ethylation process occurs preferentially before
destruction of the structure at this adsorption mode.
Although the hydroxyl radical was detected as the DMPO-
•OH adduct in the EPR measurements, it is not the principal
and sole species that leads to photooxidation of SRB in the
presence of TiO2 under visible irradiation. This was confirmed
by control experiments. Photooxidation of the dye was not
inhibited under conditions otherwise identical to those above
except for the presence of 2-propanol (1%), known to be a
•
good OH radical quencher (18). Moreover, when added to
the dispersion (SRB: 2 × 10-5 M; TiO2 loading: 2 g L-1
)
superoxide dismutase (SOD: 2.5 mg mL-1, activity: 10,000
U mg-1), which catalyzes the disproportionation of O2 to
•-
H2O2 and O2, had no effect on the initial degradation rate.
This precludes any effect of the active superoxide radical
species on the photooxidation process.
Surface-adsorbed oxygen plays the important role of
scavenging the TiO2(e-) electron thereby preventing its
recombination with Dye•+ (19). However, if the exclusive role
of O2 was merely to scavenge the TiO2(e-), then the presence
of an added oxidizing agent in an O2-free SRB/ TiO2 dispersion
should lead to the same N-de-ethylation and photooxidation
of SRB on excitation of the surface-adsorbed dye. To test this
hypothesis, CrO42- was selected as the oxidizing agent (20).
A 50 mL aqueous dispersion containing SRB (2 × 10-5 M, pH
) 4.0), CrO4 (1.6 × 10-4 M), and TiO2 (100 mg) was
2-
deoxygenated for 5 h using high purity (99.999%) N2 gas. The
2-
pH of the dispersion was kept at 4.0 to permit CrO4 to
access the metal oxide surface and scavenge TiO2 (e-).
Illuminating the dispersion for 5 h with visible light led to
negligible degradation in the first 30 min arising from some
residual chemisorbed oxygen. After oxygen was totally
consumed, no further photooxidation or de-ethylation took
place, clearly suggesting an overall inefficiency of CrO42- to
promote N-de-ethylation or photooxidation of the dye.
Therefore, the role of molecular oxygen is significant and
cannot be substituted in the N-de-ethylation and photo-
oxidation of SRB. Consequently, chemisorbed O2 on the
metal-oxide mediator acts not only as an electron trap to
suppress SRB•+/ TiO2 (e-) recombination but also as an
essential oxidizing agent which participates in the oxidative
process leading to de-ethylation and to destruction of the
chromophore ring structure, and ultimately to formation of
smaller intermediates and CO2.
9
3 9 8 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 18, 2000

In the photooxidation of SRB, different final photooxi-
dation products (as verified experimentally by UV-vis, H
structures are preceded by generation of carbon-centered
radical (12, 13). Consistent with these, degradation of SRB
must occur via two different photooxidation pathways
(destruction of the chromophore structure and N-de-
ethylation) due to formation of different radicals (either
carbon-centered radicals or nitrogen-centered radicals).
There is no doubt that electron injection from the dye to the
conduction band of TiO2 yields dye cation radicals, a process
which is determined by the nature of the HOMO orbitals of
the excited dye, 1,3dye*. After this step, the cation radical,
dye•+, can undergo hydrolysis and/ or deprotonation. Because
these two processes are sensitive to the molecular surround-
ings, the question as to which radical is formed (carbon-
centered radicals or nitrogen-centered radicals) is dictated
by the nature of the different hydrolysis or deprotonation
pathways of the dye cation radicals, which in turn are
determined by the different adsorption modes of SRB on the
TiO2 particle surface.
1
NMR, IR, and GC-MS) arise, no doubt, from different
photooxidation pathways, which are seeming defined by two
entirely different adsorption modes. Under our experimental
conditions (pH 2.5), the positively charged surface of the
TiO2 particles in the SRB/ TiO2 system permits SRB to be
chemisorbed strongly on TiO2 through the sulfonate groups
by electrostatic forces because the presence of two sulfonic
acid groups in the dye molecule ensures its existence in the
form of a singly charged anion in aqueous solutions within
the pH range 2-12 (17). When the SRB dye adsorbs on the
TiO2 surface through sulfonate groups then (Scheme 1a),
cleavage of the SRB chromophore structure seems to
predominate. By contrast, N-de-ethylation occurs only to a
slight extent. The predominant final products were diethyl-
amine and carbon dioxide. Other (minor) products detected
were N-ethylacetamide, N-ethylformamide, N,N-diethylacet-
amide, N,N-diethylformamide, formic acid, and acetaldehyde
(from incomplete de-ethylation).
On the basis of all the above experimental results (different
photooxidation products depend on the different adsorption
modes), we tentatively propose the two photooxidation
pathways depicted in Scheme 1a,b distinguished by the
different adsorption modes. In the first (Scheme 1a), the
excited dye in the SRB/ TiO2 system is adsorbed to the particle
surface through the sulfonate group. Electron injection to
the conduction band of TiO2 produces the dye radical cations,
after which either nucleophilic interactions between SRB•+
and H2O or deprotonation occur to produce neutral carbon-
centered radicals. The location of the radical in the otherwise
complex dye structure then becomes the point of attack by
adsorbed molecular oxygen (see Scheme 1a) leading to a
sequence of complex reactions: rapid decomposition via
cleavage of the ring rupture and further oxidative reactions
to yield a multitude of products. Consequently, cleavage of
the SRB chromophore structure predominates with the main
products being diethylamine and carbon dioxide; de-ethyl-
ation is a smaller process which yields the minor components
N-ethylacetamide, N-ethylformamide, and acetaldehyde.
In a previous study (15), we established an adsorption
model for the anionic surfactant DBS on the surface of TiO2
particles at different pHs by measurement of the ú-potential
of TiO2 particles. Anionic DBS molecules are easily adsorbed
on the positively charged TiO2 surface under acidic conditions
(pH 2.5) through the negative sulfonate groups owing to
electrostatic effects. This results in a drop of the ú-potential
of TiO2 particles from 55 mV to -20 mV at the DBS critical
micellar concentration of 1.2 mM (cmc). The physical location
of the SRB in SRB/ DBS/ TiO2 system was evidenced by the
fluorescence spectral changes of SRB in different media.
It is well-known that the fluorescence maxima of organic
compounds are sensitive to the polarity of the solvent (21);
that is, the more polar the solvent is the more the fluorescence
maximum of the organic compound shifts to the red. When
DBS (1.2 mM) was added to an SRB (5 × 10-7 M) solution
(pH 2.5), the maximum absorbance of SRB increased
suggesting that the presence of DBS helps to separate the
aggregated SRB molecules, whereas the red shift of the
fluorescence maximum of SRB shows that SRB interacts
strongly with DBS; note that the environment formed by
DBS is surely more polar than neat water. On the other hand,
the maximum absorbance of SRB also shifted red concomi-
tant with SRB fluorescence quenching after addition of
colloidal TiO2 (5 × 10-3 M) to an SRB solution. However,
when TiO2 was added to the SRB/ DBS system, except for
some fluorescence quenching, there was no shift of the
fluorescence maximum of SRB compared with that in the
TiO2-free system. This infers that the SRB molecules are
located, to some extent, far from the TiO2 surface in the SRB/
DBS/ TiO2 system (see Scheme 1b), so that the TiO2 surface
has little influence on the shift of the fluorescence maximum
of SRB. On addition of the anionic DBS surfactant, it results
that the SRB dye can adsorb to the negative TiO2/ DBS
interface only through the positively charged diethylamine
groups owing to favorable electrostatic interactions.
In the second pathway (Scheme 1b), the dye molecule in
the SRB/ DBS/ TiO2 system is adsorbed through the positively
charged diethylamine function. Following electron injection
from the excited dye to the TiO2 particle surface and
subsequent hydrolysis (or deprotonation) yields a nitrogen-
centered radical, which is subsequently attacked by molecular
oxygen to lead ultimately to de-ethylation. The mono-de-
ethylated dye, SRB1, can also be excited by visible light and
be implicated in other similar events (electron injection,
hydrolysis or deprotonation, and oxygen attack) to yield a
bi-de-ethylated dye derivative, SRB2. The de-ethylation
process as described above continues until formation of the
completely de-ethylated dye, SRB4 (see UV-visible spectral
results in Figure 1b). When SRB4 is excited by visible light,
the subsequent transformation of the carbon-centered cation
radicals generates formamide, acetamide, and formic acid
products as confirmed by the GC-MS results of Figure 4b.
When the SRB dye molecules are located near the TiO2
surface through the diethylamine group and assisted by the
negative ends of the DBS molecules (cmc: 1.2 mM) (Scheme
1b), which de facto neutralize somewhat the surface, the
N-de-ethylation process predominates during the initial
stages. Destruction of the chromophore ring structure occurs
mostly only after full de-ethylation of the dye. The final
products were acetaldehyde and carbon dioxide (major
products) along with formamide, formic acid, and acetamide
as the minor products.
Acknowledgments
The generous financial support of this work from the National
Natural Science Foundation of China (No. 29677019, No.
29725715, and No. 29637010), the Foundation of the Chinese
Academy of Sciences, and the China National Committee
for Science and Technology is gratefully acknowledged. The
work in Tokyo is sponsored by a Grant-in-Aid for Scientific
Research from the Japanese Ministry of Education (No.
10640569 to H.H.) and in Montreal by a grant from the Natural
Sciences and Engineering Research Council of Canada (No.
A5443 to N.S.). Finally, the authors thank the reviewers for
valuable and useful suggestions on the proposed mechanism.
According to earlier reports (22-24), most oxidative N-de-
alkylation processes are preceded by formation of nitrogen-
centered radical, whereas destruction of dye chromophore
9
VOL. 34, NO. 18, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 9 8 9

(13) (a) Wu, T.; Lin, T.; Zhao, J.; Hidaka, H.; Serpone, N. Environ. Sci.
Technol. 1999, 33, 1379. (b) Liu, G.; Wu, T.; Zhao, J.; Hidaka, H.;
Serpone, N. Environ. Sci. Technol. 1999, 33, 2081. (c) Liu, G.;
Zhao, J. New J. Chem. 2000, 24, 411.
(14) Chinese National Standard: GB 11914-89, 1989.
(15) Zhao, J.; Hidaka, H.; Takamura, A.; Pelizzetti, E.; Serpone, N.
Langmuir 1993, 9, 1646.
Literature Cited
(1) Fox, M. A.; Dulay, M. T. Chem. Rev. 1993, 93, 341.
(2) Photocatalysis-Fundamentals and Applications; Serpone, N.;
Pelizzetti, E., Eds.;Wiley-Interscience: New York, 1989.
(3) Ollis, D. F.; Pelizzetti, E.; Serpone, N. Environ. Sci. Technol. 1991,
25, 1522.
(4) Legrini, O.; Oliveros, E.; Braun, A. M. Chem. Rev. 1993, 93, 671.
(5) Turchi, C. S.; Ollis, D. F. J. Catal. 1990, 122, 178.
(6) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W.
Chem. Rev. 1995, 95, 69.
(7) Tincher, W. C. Text. Chem. Color. 1989, 21, 33.
(8) Kulkarni, S. V.; Blackwell, C. D.; Blackard, A. L.; Stackhouse, C.
W.; Alexander, M. W. U.S. Environ. Prot. Agency, Res. Dev. [Rep.]
1985, EPA-600/ 2-85/ 010.
(9) (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.
(16) Watanabe, T.; Takizawa, T.; Honda, K. J. Phys. Chem. 1977, 81,
1845.
(17) Mchedlov, P. N. O.; Shapovalov, S. A.; Egorova, S. I.; Klesh-
chevnikova, V. N.; Cordova, E.A. Dyes Pigm. 1995, 28, 7.
(18) Richard, C. J. Photochem. Photobiol. A: Chem. 1993, 72, 179.
(19) Vinodgopal, K.; Wynkoop, D. E.; Kamat, P. V. Environ. Sci.
Technol. 1996, 30, 1660.
(20) Fu, H.; Lu, G.; Li, Sh. J. Photochem. Photobiol. A: Chem. 1998,
114, 81.
(21) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum
Press: New York, 1983; Chapter 7, and references therein.
(22) Shaefer, F. C.; Zimmermann, W. D. J. Org. Chem. 1970, 35, 2165.
(23) Laube, B. L.; Asirvatham, M. R.; Mann, C. K. J. Org. Chem. 1977,
42, 670.
(10) Ross, H.; Bendig, J.; Hecht, S. Solar Energy Mater. Solar Cells
1994, 33, 475.
(11) (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. Applied
Catal. B: Environ. 1998, 15, 147. (c) Zhao, J.; Wu, K.; Wu, T.;
Hidaka, H.; Serpone, N. J. Chem. Soc., Faraday Trans. 1998, 94,
673.
(12) (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.
(24) Galliani, G.; Rindone, B.; Scolastico, C. Tetrahedron Lett. 1975,
1285.
Received for review March 2, 2000. Revised manuscript re-
ceived June 12, 2000. Accepted June 21, 2000.
ES001064C
9
3 9 9 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 18, 2000