Photocatalytic degradation of dye sulforhodamine B: A comparative study of photocatalysis with photosensitization

Article abstract of DOI:10.1039/b001573n

The direct photocatalytic degradation of dye pollutant sulforhodamine B (SRB) in aqueous TiO₂ dispersions has been examined and compared to the photosensitization process. The mineralization extent of SRB degradation, the formation of intermediates and final products were monitored to assess the degradation pathways caused by direct photocatalysis. In the initial stage of the direct photocatalysis, SRB is mainly oxidized by a positive hole upon band-gap excitation of TiO₂ by UV light (330 nm < λ < 380 nm) and subsequently undergoes the similar degradation pathways as occur in the photosensitization under visible irradiation (λ > 420 nm). Diethylamine, N,N-diethylacetamide, N-ethylformamide, N,N-diethylformamide, formic acid and acetic acid were identified as intermediate species; SO₄^2-, NH₄⁺, CO₂ and H₂O are final mineralized products produced in the direct photocatalytic process.

Full text of DOI:10.1039/b001573n

View Article Online / Journal Homepage / Table of Contents for this issue
Photocatalytic degradation of dye sulforhodamine B: a comparative
study of photocatalysis with photosensitization
Guangming Liu Jincai Zhao*¤
T he L aboratory of Photochemistry, Center for Molecular Science, Institute of Chemistry,
T he Chinese Academy of Sciences, Beijing 100080, China
Received (in Montpellier, France) 24th February 2000, Accepted 27th March 2000
Published on the Web 12th May 2000
The direct photocatalytic degradation of dye pollutant sulforhodamine B (SRB) in aqueous TiO dispersions has
2
been examined and compared to the photosensitization process. The mineralization extent of SRB degradation, the
formation of intermediates and Ðnal products were monitored to assess the degradation pathways caused by direct
photocatalysis. In the initial stage of the direct photocatalysis, SRB is mainly oxidized by a positive hole upon
band-gap excitation of TiO by UV light (330 nm \ j \ 380 nm) and subsequently undergoes the similar
2
degradation pathways as occur in the photosensitization under visible irradiation (j [ 420 nm). Diethylamine,
N,N-diethylacetamide, N-ethylformamide, N,N-diethylformamide, formic acid and acetic acid were identiÐed as
intermediate species; SO 2~, NH `, CO and H O are Ðnal mineralized products produced in the direct
4
4
2
2
photocatalytic process.
Elimination of dyes and other commercial colorants from con-
tinual wastewater effluents of textile and paper mills and other
colorant manufactures is now the subject of considerable
concern of environmental remediation. Because most of the
dyes are resistant to biodegradation, they are not easily
removed in biological wastewater treatment plants. In addi-
tion, many N-containing dyes undergo natural reductive
anaerobic degradation to potentially carcinogenic aromatic
amines.1,2 These chemicals, when discharged to environment,
are hazardous for aquatic species, have signiÐcant impact on
the aquatic environment and human health, and could result
in serious consequence to the surrounding ecosystem.3
Direct photocatalytic reaction by using semiconductor
powders has been shown to e†ectively degrade many kinds of
pollutants including dye pollutants and in many cases even
completely mineralize the compounds.4,5 But it requires ultra-
violet light to cause charge separation on the semiconductor
particles. The pollutants can be oxidized by the photogener-
ated holes either directly or indirectly via hydroxide radical
generation. Sensitized photocatalysis has been successfully
applied to degrade dye pollutants.6h10 With visible light exci-
tation of adsorbed dyes, electron transfer from the excited dye
to the conduction band of the semiconductor occurs and sub-
sequent transformation of dye cation radical is possible. These
photochemical transformations may enhance the biodegrad-
ability of dye pollutants by producing low molecular weight
oxygenated species, suggesting that sensitized photocatalysis
has a potential to be used as a pretreatment process in com-
bination with biological treatment.
catalysis with these of sensitized photocatalysis of SRB in
TiO aqueous dispersions. Major intermediates and Ðnal pro-
2
ducts were identiÐed and it appears to involve similar primary
oxidative steps in the initial photodegradation stage for both
types of photoreactions, however further decomposing the
degraded fragments to result in the complete mineralization of
the dye can only be obtained in the direct photocatalysis. A
possible mechanism of photodegradation is proposed on the
basis of the experimental results of IR, 1HNMR and GC-MS.
The primary experimental results suggest that the TiO /UV-
2
Vis photocatalytic process has promising applications in the
elimination of dye pollutants, especially under sunlight irra-
diation.
Experimental
Materials
P-25 TiO (ca. 80% anatase, 20% rutile; BET area ca. 50 m2
2
g~1) was kindly supplied by Degussa Co. Horseradish peroxi-
dase (POD) was purchased from Huamei Biologic Engineer-
ing Co. (China), whereas the N,N-dimethyl-p-phenyl-
enediamine (DPD) reagent was obtained from Merck (p.a.).
The dye Sulforhodamine-B (SRB) was of laser grade quality
(Across Co.); and all other chemicals were of analytical
reagent grade quality and were employed without further
puriÐcation. 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. 4.2). The structure of SRB molecule is illustrated below;
the small letters indicate the protons for NMR identiÐcation.
The mechanism of TiO photocatalyzed reactions under
2
UV irradiation has been extensively investigated, however, the
photocatalytic degradation mechanisms of dye pollutants with
larger and more complicated molecular structure, particularly,
the nature of the intermediates and the reaction pathways,
have not been clariÐed yet. The aim of our investigation is to
increase our understanding on the pathway of SRB degrada-
tion by comparing the experimental results of direct photo-
¤
Present address: Institute of Photographic Chemistry, The Chinese
Academy of Sciences, Beijing 100101, China. Fax: ]86-10-6487-9375;
E-mail: jczhao=ipc.ac.cn
DOI: 10.1039/b001573n
New J. Chem., 2000, 24, 411È417
411
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2000

View Article Online
Photoreactor and light source
The UV resource is a 100 W Hg lamp (j [ 330 nm, Toshiba
SHLS-1002 A). To prevent direct excitation of the dye SRB, a
cuto† Ðlter was introduced between the irradiation source and
the photoreaction vessel to eliminate radiation above 380 nm,
thus only 330 nm \ j \ 380 nm UV light from the Hg lamp
was available to cause band-gap excitation of TiO . The
2
visible light source was a 500-Watt 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. A cuto† Ðlter was also placed
outside the Pyrex jacket to ensure complete removal of radi-
ation below 420 nm and to ensure that irradiation of the
SRB/TiO system occurred only by visible light wavelengths.
2
Procedures and analyses
Fig. 1 UV-visible spectral changes of SRB (2 ] 10~5 M, 50 mL, pH
Aqueous suspensions of SRB (usually 50 mL, 2 ] 10~5 M)
4
.2) in TiO (100 mg) dispersions under UV irradiation (330
and 100 mg of TiO particulates were placed in a Pyrex vessel.
2
2
nm \ j \ 380 nm). Spectra 2È7 denote the irradiation times 0, 5, 10,
Prior to irradiation, the suspensions were magnetically stirred
1
5, 25 and 30 min, respectively. Spectrum 1 is the UV-visible spectrum
in the dark for ca. 30 min to establish an adsorption-
of SRB before addition of TiO to the solution. Inset: SRB concentra-
tion changes and wavenumber shifts of the absorption maximum as a
2
desorption equilibrium between the dye and the TiO surface.
2
At given intervals of illumination, a sample of the TiO dis-
2
function of irradiation time under the same conditions.
persion was collected, centrifuged and then Ðltered through a
Millipore Ðlter (pore size 0.22 lm). The Ðltrates were analyzed
by UV-Vis spectroscopy using a Lambda Bio 20 spectropho-
tometer (P.E. Co.). The chemical oxygen demand (COD) of
ating in the photodegradation of the dye, the wavelength of
UV irradiation is limited at 330 nm \ j \ 380 nm. Although
a very small absorption band of the dye appeared at 353 nm,
the excitation of the dye is negligible compared with the exci-
the suspensions (50 mL of 5 ] 10~5 M SRB, TiO loading
2
1
00 mg) was measured directly, without removal of the TiO
2
particles, after various irradiation time intervals using the pot-
tation of TiO under the present conditions. The extent of
2
assium dichromate titration method.11 NH ` ions were
adsorption of the dye (2 ] 10~5 M) on TiO (100 mg) par-
4
2
analyzed with an ion chromatograph (Shimadzu LC-10AS)
ticles was ca. 20% (see spectrum 2 in Fig. 1). The absorption
equipped with a CD-5 conductivity detector using a Y-521
peaks corresponding to the dye disappeared under UV light
irradiation, indicating degradation of the dye. More inter-
estingly, the characteristic absorption band of SRB at ca. 565
nm decreased rapidly, accompanied by a slight hypsochromic
shift, indicating the occurrence of some N-de-ethylation
during the photocatalytic oxidation of SRB, and similar
observations have already been conÐrmed in the Rhodamine-
B/CdS system.12
cationic column with a HNO solution (4 mM) as the eluent
3
and SO 2~ ions were monitored by ion chromatography with
4
an I-524 anionic column using an eluent of phthalic acid (2
mM). Total organic carbon (TOC) analyzer (Shimadzu TOC-
5000) was used to assay the changes in TOC during the course
of photodegradation of the dye.
Proton NMR spectra were obtained with a Varian 300
nuclear magnetic resonance spectrometer. Samples were pre-
pared as follows: several dispersions containing 200 mL of the
An aqueous dispersion (50 mL, 2 ] 10~5 M) of the dye con-
taining TiO (100 mg) was degraded (92.0%) within 30 min of
2
SRB dye (1 ] 10~4 M) and 400 mg of TiO were irradiated at
irradiation, giving a pseudo-Ðrst-order rate constant k of
2
di†erent time intervals, following which the TiO particles
2
7.43 ] 10~2 min~1 (inset of Fig. 1). The blue shift of the SRB
were removed by centrifugation and Ðltration as described
maximal absorption (expressed in wavenumber, i.e. reciprocal
of the wavelength of the maximal absorption peak in inset of
Fig. 1) indicates that the dye SRB is de-ethylated one by one.
Two competitive processes, namely N-de-ethylation and
cleavage of the SRB chromophore ring structure, occur con-
currently during photocatalytic oxidation with the latter
process predominating. Blank experiments indicated that the
disappearance of SRB was negligible without UV irradiation
above. Subsequently, the solvent of the Ðltrate was removed
under reduced pressure (below 323 K). The remaining residue
was dissolved in 0.5 mL D O. Samples for infrared (IR)
2
spectra (FTS-165 spectrophotometer) were prepared by a
method similar to that for proton NMR, except that the
residue was used directly. The intermediate and Ðnal products
generated during the photocatalytic degradation process were
analyzed by gas chromatography/mass spectrometric methods
or in the absence of TiO .
2
(
GC/MS, Trio-2000, equipped with a BPX 70 column, size
Comparing with our previous study of the sensitized photo-
2
8 m ] 0.25 mm). The preparation of samples was similar as
catalysis,13 2 ] 10~5 M SRB solution (50 mL, pH 4.2) can
above but the residue was dissolved in methanol to be applied
to the GC-MS analyses.
also be degraded with Ðrst-order kinetics (k \ 8.8 ] 10~2
min~1) in the presence of TiO (100 mg) under the visible light
2
(
j [ 420 nm) irradiation. Under otherwise identical condi-
tions, we obtained that the photoefficiency (m \ 0.012) for the
degradation induced by UV irradiation was ca. 5 times greater
than that (m \ 0.002) by visible irradiation as we noted below.
The absorption band of the dye disappeared irreversibly, indi-
cating that at least the chromophoric structure of the dye was
destroyed under visible irradiation. Control experiments
Results and discussion
UV-Vis spectroscopy
The absorption spectra of an aqueous solution of SRB record-
ed following the UV irradiation (330 nm \ j \ 380 nm) are
shown in Fig. 1. In order to prevent the excitation of the dye
established that SRB did not degrade in TiO suspensions in
2
the dark or when illuminated with visible light in the absence
by UV light, i.e. a charge transfer from excited SRB to TiO ,
a well-known sensitization phenomenon in semiconductor
of TiO . Under visible light irradiation, it is intriguing that a
slight hypsochromic shift of the maximal absorption band was
2
2
photochemistry that functions as another primary step oper-
also observed in the degradation process, indicating the
412
New J. Chem., 2000, 24, 411È417

View Article Online
occurrence of N-de-ethylation during the photosensitized oxi-
dation of SRB as occurred under UV irradiation.
reduction was slower than that of the degradation of the dye,
43.1% of TOC still remained in the dispersion at 10 h of irra-
diation after the dye solution was totally discolored. After 16
h of irradiation, the mineralization yield reached a value of
Photoefficiencies
9
7.5%. NH ` ions began to form after a 6 h induction period
In this work, photoefficiencies m of the photodegradation of
the dye SRB under both visible and UV irradiation were
determined at 560 nm (using a Japan Optical Co. Interference
Ðlter to cuto† wavelengths below 550 nm and above 570 nm)
and at 365 nm (using a similar Ðlter to cuto† wavelengths
below 355 nm and above 375 nm), respectively. The photon
Ñux of the 500 W halogen lamp at 560 nm was 3.4 ] 10~8
einstein s~1 determined by Reineckate actinometry
4
at a rate slower than that of formation of SO 2~. This time-
lag may be attributed to the repulsive interactions between the
positively charged nitrogen moiety in the SRB structure and
the positive surface of the illuminated TiO particulates under
the conditions used (pH \ 4.2; pI
another more important reason that formation of NH ` was
not observed before 6 h of irradiation is most probably
because the cleavage of the conjugated chromophore ring
structure is predominant in the initial stage as we noted
above. Subsequently, ammonium ions arose from further oxi-
dation of the nitrogen-containing fragments and increased
with irradiation time to reach 72.0% (2.88 ] 10~4 M) of the
theoretical quantity (4.00 ] 10~4 M) after 16 h of irradiation,
while nitrate ion formation from the photooxidative step was
negligible. It can be explained by the chemical structure of the
substrate, which inÑuences the production and relative quan-
tities of NH ` and/or NO ~ ions during the photocatalyzed
oxidation of nitrogen-containing substances at the TiO ÈH O
interface.17h19 In the formation of sulfate ions, electrostatic
adsorption of the substrate via its most electronegative atom
to the TiO surface is one of the important factors that inÑu-
ence the photooxidation pathway. AM1 (Austin model 1)
molecular orbital calculations carried out semi-empirically
using the MOPAC system20 indicated that the O atoms of the
sulfonate groups in the dye molecule bear the largest negative
charge and therefore interact with the positively charged TiO2
surface. Thus, the sulfur atom or the a-carbon adjacent to the
sulfur atom are readily attacked by the electrophilic ~OH rad-
icals formed on the TiO surface to produce SO 2~ concomi-
tant with the destruction of the chromophore ring structure.
As depicted in Fig. 2, the amount of sulfate ions increased
with irradiation time during the photocatalytic degradation
process, and reached 72.6% (2.90 ] 10~4 M) of the theoretical
value.
4
2
\ 6.8).16 However,
TiO2
4
(m \ 0.2714) and the photon Ñux of the 100 W Hg lamp at 365
nm was 5.4 ] 10~9 einstein s~1 also assessed by Reineckate
actinometry (m \ 0.2914). For the degradation of SRB under
visible or UV irradiation, photoefficiencies were measured by
observation of the amount of SRB degraded per photon
(
within 20 min of irradiation) using a 50 mL dispersion con-
taining 2 ] 10~5 M SRB and 100 mg of TiO under the same
2
conditions as those employed for the photon Ñow measure-
ments. The photoefficiency for SRB degradation via direct
photocatalysis was 0.012, which falls within the reported
quantum yields for the direct photocatalytic degradation of
organic compounds ranging from 0.005 to 0.06.15 The photo-
efficiency for the sensitized photocatalysis of SRB was 0.002,
signiÐcantly lower than that for direct photocatalysis.
Although the photoefficiencies for the SRB sensitized photo-
catalytic degradation are smaller than those for the direct
photocatalytic degradation, the sensitized photocatalytic deg-
radation of dye pollutant SRB may play the main role in the
overall degradation by sunlight because the quantity of
photons available for sensitized photocatalysis with sunlight is
much higher than that for direct photocatalysis (\5% of total
photons).
4
3
2
2
2
2
4
Production of inorganic nitrogen, sulfur species
(
NH ‘, SO 2—) and mineralization (CO )
4 4 2
The destruction of SRB conjugated chromophore ring struc-
ture, decreasing of total organic carbon (TOC), and formation
of ammonium and sulfate ions in the direct photocatalytic
degradation of a 50 mL of SRB solution (2 ] 10~4 M, pH 4.2)
According to the previous results of photosensitization,13
the degradation of the dye (2 ] 10~4 M, 50 mL) suspensions
containing TiO (100 mg) under visible irradiation follows
Ðrst-order kinetics with rate constant k \ 1.73 ] 10~3 min~1.
The TOC of the solution changed from the initial value of
in the presence of TiO (100 mg) are illustrated in Fig. 2. The
2
2
complete degradation of the dye required 10 h of irradiation
with pseudo-zero-order rate constant k \ 2.22 ] 10~7 M
min~1. The decreasing rate of TOC exhibited two di†erent
behaviors before and after 7 h of irradiation, respectively, indi-
cating that the mineralization of the dye goes through two
di†erent stages: the ring cleavage in the initial photocatalytic
degradation stage and subsequent oxidation of the fragments
in the latter stage (see discussion below). The rate of TOC
52.4 ppm before irradiation to 24.0 ppm after 38 h of irradia-
tion, indicating that only partial mineralization (54.2%) of the
SRB occurred in aqueous TiO dispersion under visible irra-
diation, no more mineralization was observed with further
irradiation after the dispersion was discolored completely.
Meanwhile, the concentration of SO 2~ ions formed increased
with the irradiation time, reaching a constant value of 5.2 ppm
2
4
after 38 h of irradiation, only 13.5% of the theoretically
expected value (38.4 ppm). Formation of ammonium and
nitrate ions could not be observed in the system, indicating no
further mineralization of the nitrogen-containing fragments.
All the above data indicate that the degraded fragments, espe-
cially the nitrogen-containing fragments, did not undergo
further mineralization after the chromophore ring structure of
the dye molecule was destroyed, which is di†erent from that of
direct photocatalysis.
COD and H O measurements
2
2
Fig. 3 shows the temporal changes of chemical oxygen
demand (COD) in the direct photocatalytic degradation of the
dye SRB, which reÑect the mineralization extent of the organ-
ics in the total dispersions (both in the bulk and on the TiO2
Fig. 2 The changes in SRB concentration and TOC, and the forma-
tion of SO 2~ and NH ` during the course of direct photocatalytic
surface). Under UV irradiation, the decrease in COD values of
4
4
degradation of SRB (2 ] 10~4 M, 50 mL) in the presence of TiO (100
5 ] 10~5 M of SRB dispersions (50 mL, pH 4.2) containing
2
mg) at pH 4.2.
TiO (100 mg) also shows two di†erent stages similar to the
2
New J. Chem., 2000, 24, 411È417
413

View Article Online
initial oxidization of the dye, resulting in the destruction of the
dye chromophore structure, as will be further demonstrated in
the following.
IR and 1HNMR measurements
In order to probe the nature of surface photochemical events
of the dye SRB adsorbed on the TiO particles and to identify
2
the intermediates and Ðnal products arising from direct
photocatalytic degradation of the dye, a series of IR and
1
HNMR spectra were carried out at various irradiation times.
The temporal variation of the IR spectra during the course of
the photocatalytic degradation of SRB by UV irradiation is
illustrated in Fig. 4.
The assignments for the principal bands in the IR spectra of
SRB (before irradiation):28 1590, 1558, 1530, 1510, 1490 and
1470 cm~1 correspond to aromatic ring vibrations; 1344
cm~1 is due to CÈaryl bond vibrations; 1120È1145 cm~1 and
Fig. 3 Variation in COD and formation of H O during the course
2
2
of direct photocatalytic degradation of SRB (5 ] 10~5 M, 50 mL,
TiO : 100 mg) at pH 4.2.
2
668È625 cm~1 are caused by vibrations in the ÈSO ~ group;
3
the band at 1649 cm~1 is attributed to vibrations of the
carbonÈnitrogen bond; and the heterocycle vibrations are the
origins of the absorption band at 1530È1558 cm~1. During
the direct photocatalytic process, bands of vibrations charac-
teristic of the carbonÈnitrogen bond, the CÈaryl bond, and
nearly all of the aromatic skeletal and heterocycle vibrations
decreased with irradiation time and disappeared after about
variation tendency of TOC, before and after 135 min of irra-
diation. 91.2% of total COD (45.2 mg O L~1) was reduced
2
after 420 min of UV irradiation; while under visible irradia-
tion,13 only 74.3% of the total COD was decreased after 780
min of irradiation and no change was observed with further
irradiation after the dispersion was totally discolored. Of more
importance is that the COD value (12 mg O L~1) of the
5
2 h of irradiation. Meanwhile, some new IR absorption
bands appeared at 1720 cm~1 (attributable to [C2O groups),
400 cm~1 (attribute to vibrations of ÈCH or ÈCH È of
2
dispersion at 135 min of UV irradiation is almost the same as
that (11.6 mg O2 L~1) of the discolored dispersion under
visible irradiation, indicating that the dye SRB mainly under-
went the destruction of the chromophore structure before 135
min of UV irradiation (similar as occurred in the photo-
sensitized degradation process) and subsequent mineralization
of the fragments occurred after 135 min of UV irradiation as
noted above.
1
3
2
diethylamine), and 857 cm~1. The IR spectra of SRB in the
photosensitization13 also show the similar variation tendency
except that the Ðnal products are more complicated than
those in the direct photocatalysis; almost all the new absorp-
tion bands that appeared in the direct photocatalysis were
found in the photosensitized process. The IR results indicate
that the large conjugated chromophore structure of SRB is
destroyed under both UV and visible light irradiation in
Formation of H O as one of main intermediates in the
2
2
direct photocatalysis was detected by the spectrophotometric
DPD and catalase enzyme methods.21h24 As shown in Fig. 3,
the quantities of H O formation (a maximum value of
aqueous TiO dispersions, and the dye molecule undergoes
2
irreversible chemical changes and degrades to smaller organic
2
2
2
.8 ] 10~5 M) at pH 4.2 increased with the irradiation time
species containing [C2O groups. In the photosensitization,
the degraded products remained unchanged with further
irradiation after the dispersion was totally discolored. In
the direct photocatalysis, the degraded small species
with second-order rate constant k \ 1.76 ] 103 M~1 min~1
in the initial stage (before 40 min of irradiation), indicating
p
that the formation of H O is a radical-radical reaction under
2
2
the present conditions. H O can be formed as following reac-
(diethylamine, N-ethylformamide, N,N-diethylacetamide,
2
2
tions (1)È(6) in the direct photocatalysis:
N,N-diethylformamide, CH COOH and HCOOH, as con-
3
Ðrmed below by the 1HNMR and GC-MS) can be further
h` ] OH~(or H O) ] ~OH
(1)
(2)
(3)
2
decomposed to inorganic species (CO , NH `, H O, etc.).
2
4
2
~
OH ] ~OH ] H O k \ 5.2 ] 109 M~1 s~1 25
To further identify the species produced in the photooxida-
tion of SRB and compare the photodegradation pathways
under the two di†erent photocatalytic mechanisms, the tem-
poral proton NMR proÐles of the dye, intermediates and Ðnal
products were monitored. Spectrum a in Fig. 5 shows the
2
2
e~(cb) ] O ] O ~~
2
2
O ~~ ] H` ] HO ~ pK \ 4.69,
2
2
a
k \ 1 ] 1010 M~1 s~1 26,27
(4)
HO ~]HO ~]H O ]O k\8.3]105 M~1 s~1 26,27 (5)
2
2
2 2
2
O ~~ ] e~ ] 2H` ] H O
(6)
2
2 2
In the photosensitization H O can only be formed via reac-
tions (3)È(6), where the conduction band electrons come from
2
2
the excited dye rather than band-gap excitation of TiO .
2
Because little amounts of H O can be observed at pH 4.2 in
2
2
the photosensitized degradation of the dye,13 the amount of
H O formed in the direct photocatalysis mainly comes from
reactions (1)È(2) under the similar conditions, which is consis-
2
2
tent with the notion that formation of H O is a radicalÈ
2
2
radical reaction under the present conditions. And moreover,
H O was mainly produced before 135 min of UV irradiation
2
2
during the process of destruction of the dye chromophore
structure as depicted in Fig. 3, indicating that the hydroxyl
radicals (~OH) are used mainly to produce H O . It is mainly
Fig. 4 Temporal changes of the IR spectra in the direct photo-
catalytic degradation of SRB.
2
2
the positive holes (h`), but not ~OH that participate in the
414
New J. Chem., 2000, 24, 411È417

View Article Online
demonstrates that the ethyl group (CH CH ) of SRB is not
2
3
destroyed in the initial direct photocatalytic oxidation
process, but remains in the diethylamine photooxidation pro-
ducts (also see the results of GC-MS below).
On the basis of the signal intensities from proton NMR
spectra (spectrum d of Fig. 5), we establish that diethylamine
is a major photooxidation product in the direct photocatalytic
degradation of the dye. Additional evidence of the formation
of diethylamine, HCOOH and CH COOH can be obtained in
3
the following experiments. It is expected that distillation of an
irradiated aqueous SRB/TiO sample under alkaline condi-
2
tions (pH 10.5) would lead to removal of diethylamine while
distillation under acidic conditions (pH 2.5) would lead to
removal of HCOOH and CH COOH. It was indeed the case
3
evidenced by loss of the corresponding proton NMR signals
at d 1.23È1.40 and 3.00È3.20 [HN(CH CH ) ] or d 1.98
2
3 2
(
CH COOH) and d 8.46 (HCOOH) in the 1HNMR spectra
3
under alkaline or acidic conditions, respectively.
Surprisingly, the temporal proton NMR spectroscopy pro-
Ðles of the dye, intermediates and Ðnal products in the photo-
sensitized degradation13 were exactly similar to that observed
in the direct photocatalytic degradation, which indicates that
the dye SRB probably underwent the same photodegradation
pathway at least in the initial stages for the two di†erent deg-
radation mechanisms. The exclusive di†erence is that the
degraded species (such as diethylamine) can be further decom-
posed to inorganic species (CO , NH `, H O) (as in Fig. 5
2
4
2
spectra e) under UV irradiation while under visible irradia-
tion, no degradation can be observed after the dispersion was
totally discolored.
The above NMR Ðndings together with the results of IR
support the notion that the same degradation pathway occurs,
that is the breakup of the chromophoric structure of SRB and
dominates during the initial irradiation stage in both the
direct and sensitized degradation of the dye (formation of a
large quantity of diethylamine) with only a small extent of
de-ethylation taking place, as depicted in Fig. 1. The degraded
fragments can be further decomposed to mineralized products
in the direct photcatalysis while they remain unchanged in the
photosensitization.
IdentiÐcation of intermediates by GC-MS measurements
GC-MS was used to analyze the degraded intermediates and
Ðnal products formed in the direct photocatalytic degradation
of the dye SRB. Fig. 6 shows the gas chromatogram of the
degraded SRB dispersion obtained after 30 h of UV irradia-
tion. Seven peaks, marked with AÈG, appeared in the chro-
matogram. Mass spectra of main peaks (A, C, D, E, G) are
shown in Fig. 7. The compounds corresponding to the GC
spectral peaks were determined to be as follows: (A)
diethylamine (t \ 13.6 min), (C) N,N-diethylacetamide (t \
Fig. 5 Temporal variation of the 1HNMR spectra during the course
of direct photocatalytic degradation of SRB.
typical proton NMR signals of pure SRB and their assign-
ments to the various protons in the structure of SRB (see
above). The NMR signals of the aromatic hydrogens H , H ,
c
d
H , H , H and H were located at d 6.66, 6.83È6.86, 6.96È
6
e
f
g
h
.99, 7.62È7.65, 8.25È8.28 and 8.52 ppm, respectively; those of
r
r
H and H of the N-diethyl group appeared at d 1.18È1.23 and
11.6 min), (D) N-ethylformamide (t \ 8.7 min), (E) N,N-
a
b
r
diethylformamide (t \ 7.5 min). The mass spectra revealed
r
3
.40È3.60 ppm, respectively. During the course of the direct
photocatalytic oxidation of SRB (spectra a to e in Fig. 5), a
series of new signals appeared at d 1.23È1.40 ppm (hydrogens
from CH3 and analogous groups), 1.98, 3.00È3.20, and 8.46,
that the main peak (A) had a fragmentation pattern consistent
with diethylamine, gave m/z 73 [M]` (the molecular ion
peak),
58
[MÈCH ]`,
44
[MÈCH CH ]`,
30
3
2
3
whereas the characteristic signals of SRB at d 1.18È1.23 (CH ),
[MÈNCH CH ]` (the base peak). The chromatogram peaks
3
2
3
3
.40È3.60 (CH ), and 6.60È8.60 (aromatic protons) disap-
C and E contained an ion peak at m/z 72, which is assignable
to a fragment of dehydrogenated diethylamine ion. The peak
D (N-ethylformamide) has the same molecular mass and
similar fragmentation pattern as peak A (diethylamine), they
were separated from each other according to the di†erent
retention times in the column and were identiÐed by compari-
son with retention times of standards. A species having the
largest molecular mass of 179 of all the compounds observed
in the GC was detected as chromatogram peak G, precise
identiÐcation has not yet been done.
2
peared. The new signal at d 8.46 is attributed to the CH
proton of formic acid and the signal of d 1.98 to the methyl
protons of CH COOH by comparison with 1HNMR signals
3
of the standards, both of them increased with increasing illu-
mination time. The intensities of the two new signals at d
1.23È1.40 and 3.00È3.20 increased with the irradiation time
whereas the methyl (CH ) protons at d 1.18È1.23 ppm and the
3
methylene (CH ) protons at d 3.40È3.60 ppm of SRB
2
decreased in intensity with irradiation time. More interesting
is the ratio of the integrated area of the Ðrst two peaks which
remained constant at 3 : 2 during the photooxidation, the
same as the ratio of the latter two signals. This observation
In addition to the intermediates shown in Fig. 7, two lower
intensity intermediates having the molecular mass of 96 and
164 were also detected (peaks F and B appeared in Fig. 6).
New J. Chem., 2000, 24, 411È417
415

View Article Online
formamide, HCOOH and CH CHO were also conÐrmed
3
by the 1HNMR and GC/MS results as the end products.13
Photocatalytic degradation pathways
In the direct photocatalysis, both positive holes and hydroxyl
radicals have been proposed as the oxidizing species
responsible for initiating the degradation of organic sub-
strates. Some researchers proposed that when photocatalytic
oxidation is conducted in the presence of water, indirect hole
transfer, which involves water molecules and surfacial
hydroxyl radicals, is
a more efficient hole-scavenging
pathway.29,30 While others argued that substrates with the
proper oxidation potentials can undergo direct oxidation, gen-
erating radical species,31,32 in fact, a number of species, espe-
cially among organic compounds strongly adsorbed on the
Fig. 6 GC chromatogram obtained by GC-MS analysis from the
direct photocatalytic degradation of SRB after 30 h of irradiation.
TiO surface, and appear able to compete efficiently for posi-
2
tive holes with OH groups adsorbed at the TiO surface.33 In
2
this system, we have conÐrmed that the degradation pathway
of the dye SRB in the initial stages of the direct photocatalysis
is exactly the same as that in the photosensitization on the
basis of UV-Vis spectra, IR, 1HNMR and GC-MS experimen-
tal results. It is well known that in photosensitization the dye
is excited by visible light and the excited dye injects an elec-
tron to the conduction band of TiO to form a cation radical,
2
subsequent transformation of the cation radical results in the
degradation of the dye.
In Scheme 1, we show a proposed mechanism with the reac-
tion paths for positive holes in the direct photocatalysis, the
TiO particles absorb UV photons to produce electron-hole
2
pairs, which may be competitively trapped by adsorbed H O,
2
OH~, O and dye molecules. The dye SRB adsorbed on TiO
is mainly oxidized by a photogenerated hole localized at the
2
2
surface of the irradiated semiconductor TiO , generating an
2
adsorbed cation radical, but not attacked by hydroxyl radicals
to produce hydroxylated intermediates. The dye cation
radicals formed combine with molecular oxygen adsorbed
and result in the destruction of the dye chromophore
structure to form small species (such as diethylamine,
N,N-diethylacetamide,
N-ethylformamide,
N,N-diethyl-
formamide, formic acid and acetic acid), these degraded
fragments can be further decomposed to give mineralized
products (SO 2~, NH `, CO , H O etc.) by the direct
4
4
2
2
Fig. 7 Mass spectra of the species observed in the GC chromato-
gram shown in Fig. 6.
The latter species gave a mass spectrum showing a base peak
at m/z 133, some ion peaks at m/z 39, 51, 77, indicating that
the intermediate contains a benzene ring, which was produced
in the initial stages as a result of destruction of the conjugated
chromophoric structure of the dye. Their precise assignments
were not made in this study. Some other kinds of interme-
diates having smaller molecular weights (such as formic acid
and acetic acid conÐrmed in the 1HNMR section) should have
been produced during the course of the mineralization of SRB
to CO , but they are difficult to detect by GC-MS if interme-
2
diates having a high boiling point such as carboxylate salts
were produced in the suspension, they were not easy to gasify
in the GC-MS measurement; or if intermediates with too low
boiling point were produced, they would be removed from the
sample before it was used for GC-MS analysis because rotary
evaporation had to be used to concentrate the samples.
In the photosensitized degradation of the dye SRB, similar
degraded products such as diethylamine, N-ethylacetamide,
Scheme 1 Proposed photodegradation pathways of the dye SRB in
the direct photocatalysis and in the photosensitization processes.
N-ethylformamide,
N,N-diethylacetamide,
N,N-diethyl-
416
New J. Chem., 2000, 24, 411È417

View Article Online
7
8
H. Ross, J. Bendig and S. Hecht, Sol. Energy Mater. Sol. Cells,
oxidation of positive holes or the attack of hydroxyl radicals.
On the contrary, in the photosensitization, the transformation
of dye cation radicals produces the Ðnal degraded products
1
994, 33, 475.
(a) F. Zhang, J. Zhao, L. Zang, T. Shen, H. Hidaka, E. Pelizzetti
and N. Serpone, J. Mol. Catal. A: Chem., 1997, 120, 173; (b) F.
Zhang, J. Zhao, T. Shen, H. Hidaka, E. Pelizzetti and N. Serpone,
Applied Catal. B: Environ., 1998, 15, 147; (c) J. Zhao, K. Wu, T.
Wu, H. Hidaka and N. Serpone, J. Chem. Soc. Faraday T rans.,
(e.g. diethylamine, N-ethylacetamide, N-ethylformamide, N,N-
diethylacetamide, N,N-diethylformamide, HCOOH and
CH CHO), which would not be further mineralized under
3
1
998, 94, 673.
visible light irradiation, the mechanism of their degradation is
9
0
(a) T. Wu, G. Liu, J. Zhao, H. Hidaka and N. Serpone, J. Phys.
Chem. B, 1998, 102, 5845; (b) J. Zhao, T. Wu, K. Wu, K. Oikawa,
H. Hidaka and N. Serpone, Environ. Sci. T echnol., 1998, 32, 2394.
(a) T. Wu, T. Lin, J. Zhao, H. Hidaka and N. Serpone, Environ.
Sci. T echnol., 1999, 33, 1379; (b) G. Liu, T. Wu, J. Zhao, H.
Hidaka and N. Serpone, Environ. Sci. T echnol., 1999, 33, 2081.
Chinese National Standard: GB 11914È89, 1989.
probably related to the active oxygen radicals produced in the
photosensitization process, while independent of the medi-
ation of the positive holes as occurs in the direct photo-
catalysis.
1
1
1
1
2
T. Watanabe, T. Takizawa and K. Honda, J. Phys. Chem., 1977,
Conclusions
81, 1845.
The dye SRB can be e†ectively degraded by both the direct
and sensitized photocatalysis. Similar intermediates (such as
diethylamine, N,N-diethylacetamide, N-ethylformamide, N,N-
13 G. Liu, X. Li, J. Zhao, H. Hidaka and N. Serpone, Environ. Sci.
T echnol., submitted. ““In this paper, the photooxidation of sul-
forhodamine B has been examined under visible light illumi-
nation, two di†erent oxidation pathways are described to
account for di†erences in the Ðnal photooxidation products
whose nature depends on the di†erent modes of adsorption of the
diethylformamide, formic acid) and Ðnal products (SO 2~,
4
CO ) were identiÐed by UV-Vis spectra, TOC, COD, IR,
1
2
HNMR and GC-MS techniques in the two di†erent degrada-
dye sulforhodamine B on the metal-oxide mediator (TiO ).ÏÏ
2
tion processes, but their distribution is dependent on the
photocatalytic mechanism. Direct and sensitized photo-
catalyses appear to involve similar photodegradation path-
ways in the initial stages, leading to a rapid destruction of
SRB chromophore structure to form smaller organic species
except that direct photocatalysis is capable of achieving com-
plete mineralization with further irradiation. Ammonium is
the exclusive primary inorganic nitrogen species produced in
the direct photocatalysis, which is not found in the sensitiza-
tion. The application of direct and sensitized photocatalysis
has the potential for the treatment of dye pollutants because it
can fully utilize solar radiation.
14 E. E. Wegner and A. W. Adamson, J. Am. Chem. Soc., 1966, 88,
3
94.
1
5
6
U. Sta†ord, K. A. Gray and P. V. Kamat, J. Phys. Chem., 1994,
9
8, 6343.
1
J. Zhao, H. Hidaka, A. Takamura, E. Pelizzetti and N. Serpone,
L angmuir, 1993, 9, 1646.
17 K. Nohara, H. Hidaka, E. Pelizzetti and N. Serpone, Catal. L ett.,
1996, 36, 115.
18
19
20
21
K. Nohara, H. Hidaka, E. Pelizzetti and N. Serpone, J. Photo-
chem. Photobiol. A, 1997, 102, 265.
H. Hidaka, K. Nohara, J. Zhao, K. Takashima, E. Pelizzetti and
N. Serpone, New J. Chem., 1994, 18, 541.
M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902.
H. Bader, V. Sturzenegger and J. Hoigne, W ater Res., 1988, 22,
1
109.
2
2
2
3
Y. Zuo and J. Hoigne, Science, 1993, 260, 71.
E. R. Carraway, A. J. Ho†mann and R. Ho†mann, Environ. Sci.
T echnol., 1994, 28, 786.
Acknowledgements
The generous Ðnancial support of this work from the National
Natural Science Foundation of China (No. 29725715, No.
2
4
5
T. Wu, G. Liu, J. Zhao, H. Hidaka and N. Serpone, J. Phys.
Chem. B, 1999, 103, 4862.
59772033 and No. 29637010), the Foundation of the Chinese
2
K. Sehested, O. L. Rasmussen and H. Fricke, J. Phys. Chem.,
Academy of Sciences and the China National Committee for
Science and Technology is gratefully acknowledged.
1
968, 72, 626.
26 B. H. J. Bielski, D. E. Cabelli, R. L. Arudi and A. B. Ross, J.
Phys. Chem. Ref. Data, 1985, 14, 1041.
2
7 D. T. Sawyer and J. S. Valentine, Acc. Chem. Res., 1981, 14, 393.
References
28 P. N. O. Mchedlov, S. A. Shapovalov, S. I. Egorova, V. N. Klesh-
chevnikova and E. A. Cordova, Dyes Pigm., 1995, 28, 7.
29 M. R. Ho†mann, S. T. Martin, W. Choi and D. W. Bahnemann,
Chem. Rev., 1995, 95, 69.
30 W. Choi and M. R. Ho†mann, J. Phys. Chem., 1996, 100, 2161.
31 (a) P. V. Kamat, Chem. Rev., 1993, 93, 267; (b) M. A. Fox and
M. T. Dulay, Chem. Rev., 1993, 93, 341.
32 M. A. Fox, R. B. Draper, K. E. OÏShea, in Photochemical Conver-
sion and Storage of Solar Energy, eds. E. Pelizzetti, and M. Schia-
vello, Kluwer, Dordrecht, 1991, p. 323.
33 A. Wahl, M. Ulmann, A. Carroy, B. Jermann, M. Dolata, P. Ked-
zierzawski, C. Chatelain, A. Monnier and J. Augustynski, J. Elec-
troanal. Chem., 1995, 396, 41.
1
2
D. Brown and P. Laboureur, Chemosphere, 1983, 12, 397.
J. McCann and B. N. Ames, Proc. Natl. Acad. Sci. USA, 1975, 73,
9
50.
3
4
W. C. Tincher, T ext. Chem. Color., 1989, 21, 33.
D. F. Ollis, E. Pelizzetti and N. Serpone, Environ. Sci. T echnol.,
1
991, 25, 1522.
5
6
R. W. Matthews, J. Catal., 1988, 111, 264.
(a) K. Vinodgopal, D. Wynkoop and P. V. Kamat, Environ. Sci.
T echnol., 1996, 30, 1660; (b) C. Nasr, K. Vinodgopal, L. Fisher, S.
Hotchandani, A. K. Chattopadhya and P. V. Kamat, J. Phys.
Chem., 1996, 100, 8436.
New J. Chem., 2000, 24, 411È417
417