Photocatalytic degradation of dyes in water: Case study of indigo and of indigo carmine

Article abstract of DOI:10.1006/jcat.2001.3232

The TiO₂/UV photocatalytic degradations of indigo and of indigo carmine have been investigated both in aqueous heterogeneous suspensions and in the solid state. In addition to prompt removal of the color, TiO₂/UV-based photocatalysis

Full text of DOI:10.1006/jcat.2001.3232

Journal of Catalysis 201, 46–59 (2001)
doi:10.1006/jcat.2001.3232, available online at http://www.idealibrary.com on
Photocatalytic Degradation of Dyes in Water: Case Study
of Indigo and of Indigo Carmine
Manon Vautier, Chantal Guillard, and Jean-Marie Herrmann1
Laboratoire de Photocatalyse, Catalyse et Environnement LPCE (IFoS UMR CNRS n 5621), Ecole Centrale de Lyon,
B.P. 163, 69131 Ecully cedex, France
Received October 24, 2000; revised March 13, 2001; accepted March 23, 2001; published online May 31, 2001
eutrophication, and perturbations in aquatic life. As in-
ternational environmental standards are becoming more
stringent (ISO 14001, October 1996), technological sys-
tems for the removal of organic pollutants, such as dyes,
have been recently developed. Physical methods, (adsorp-
tion (2)), biological methods (biodegradation (3, 4)), and
chemical methods (chlorination and ozonation (5)) are the
most frequently used.
Among the new oxidation methods or “advanced oxida-
tion processes” (AOP), heterogeneous photocatalysis ap-
pears as an emerging destructive technology leading to the
total mineralization of many organic pollutants (6–12), fol-
lowing the proposed mechanism:
The TiO2/UV photocatalytic degradations of indigo and of in-
digo carmine have been investigated both in aqueous heterogeneous
suspensions and in the solid state. In addition to prompt removal of
the color, TiO2/UV-based photocatalysis was simultaneously able
to oxidize the dye, with almost complete mineralization of car-
bon and of nitrogen and sulfur heteroatoms into CO2, NH+4 , NO3 ,
and SO24 , respectively. A detailed degradation pathway has been
determined by careful identification of intermediate products, in
particular, carboxylic acids, whose decarboxylation by photo-Kolbe
reactions constitutes the main source of CO2 evolution. The only
persistent organic compound was acetic acid, whose degradation
required a longer period of time. These results suggest that TiO2/UV
photocatalysis may be envisaged as a method for treatment of di-
luted wastewaters in textile industries. The irradiation of titania
with visible light did produce a photoinduced decolorization of the
dye, probably induced by the breaking of the double-bond conjuga-
tion system of the chromophoric group. However, this decoloriza-
tion was not accompanied by any degradation of the molecule since
no loss of total organic carbon (TOC) nor release of inorganic ions
were observed. This corresponded to a stoichiometric reaction of an
electron transfer from the dye molecule excited in visible irradia-
tion to titania. Because indigo is very poorly soluble ( 2 ppm), it
was tentatively degraded in its solid state, mixed with titania in a
photocatalytic solid-solid-type reaction. Observation of the decol-
orization and of the degradation of solid indigo constitutes a sur-
prising and encouraging result for the development of self-cleaning
titania-coated objects (glasses, steel, aluminium, metals, walls, etc.)
c
—absorption of efficient photons (h
EG = 3.2 eV) by
titania
(TiO2) + h → eCB + hV+B
;
[1]
—oxygen ionosorption
(O2)ads + eCB → O2
;
[2]
—neutralization of OH groups into OH by photoholes
(H2O ⇔ H+ + OH )ads + h+VB → H+ + OH ; [3]
—oxidation of the organic reactant via successive attacks
by OH radicals
fouled by solid dirt particles.
2001 Academic Press
R + OH → R0 + H2O;
—or by direct reaction with holes
[4]
Key Words: titania; photocatalysis; photocatalytic degradation;
dyes; indigo; indigo carmine; water decolorization; water purifica-
tion; solid indigo.
R + h+ → R+ → degradation products.
[5]
INTRODUCTION
As an example of the last process, holes can react directly
with carboxylic acids, generating CO2 according to the so-
called photo-Kolbe reaction:
Fifteen percent of the total world production of dyes is
lost during the dyeing process and is released in the textile
effluents (1). The release of those colored wastewaters in
the ecosystem is a dramatic source of esthetic pollution,
RCOO + h+ → R + CO2.
[6]
1 To whom correspondence should be addressed. Fax: (33) 4 78 33 03 37.
E-mail: jean-marie.herrmann@ec-lyon.fr.
A general list of various families of organic pollutants
that can be treated by photocatalysis has been given (13).
46
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
47
In most cases, the degradation is conducted for dissolved (20). Up to now, as far as we know, no studies have been
compounds in water with UV-illuminated titania. Possible devoted to the photocatalytic degradation of indigo (13)
limits of the technique concern the irradiation source and probably because of its too low solubility (21). In the
the physical state of the pollutant. Recently, some authors present work, we studied the photocatalytic decoloriza-
have reported the degradation of organic pollutants by vis- tion and degradation of indigo and of one indigoid dye,
ible light using photosensitization of the catalyst (14–17). indigo carmine (indigo disulphonic-5,50 acid (Scheme 1)),
The interest is to use solar light, which is free and inex- since its solubility in water enables one to perform a bet-
haustible. With respect to the physical state of the pollu- ter kinetic study of degradation with easier identification of
tant, very few studies have been devoted to the photo- degradation intermediates. We studied successively (i) the
catalytic degradation of a solid. The only (undesirable) photocatalytic degradation of indigo and of indigo carmine
example concerns the chalking of pigmented coatings and in water, (ii) the photoinduced decolorization of indigo
plastics (18).
carmine in water by irradiation of TiO2 in visible irra-
Indigo or 2,20-bis-indole is not only one of the oldest dyes diation, and (iii) the photocatalytic degradation of solid
known but still is one of the most importantly used: It repre- indigo.
sents 3% of the total production of dyes. Its major industrial
EXPERIMENTAL
application is the dyeing of clothes (blue jeans) and other
blue denim products (19). Its unusually high melting point
(390–392 C) and its very poor solubility can be explained by
the existence of strong intermolecular hydrogenbonds. In
the solid state, indigo forms a polymer in which each indigo
molecule is linked to four surrounding molecules. In non-
polar solvents, the indigo is present mainly as a monomer,
whereasin polar solvents, intermolecular association occurs
and the solutions are blue (1). The basic color-producing
structure is a cross-conjugated system or H-chromophore,
Materials
Indigo and indigo carmine were supplied by a textile
firm and used as received. The photocatalyst was titania
Degussa P-25 (anatase/rutile = 3.1; surface area = 50 m2g
,
1
nonporous particles). The slurry, which is not colloidal, con-
sists of large fractal aggregates of individual particles, each
one having a mean diameter, d, of ca. 30 nm, in agreement
(i) with the value calculated from the formula (given as a
function of the volumic mass and the specific surface area
S for homodispersed nonporous particles),
==
consisting of a single C C double bond substituted by two
NH donor groups and two CO acceptor groups (Scheme 1)
d = 6/ S,
and (ii) with transmission electron microscopy performed
on sonicated samples.
Photoreactor and Light Source
Photocatalytic tests were performed in a batch micropho-
toreactor of 100 mL made of Pyrex (transmittance:
>
290 nm), which is described in Fig. 1. It was equipped with a
high-pressure mercury lamp (Philips HPK 125 W). For visi-
ble irradiation, the same lamp was used but the light beams
were filtered by a Corning 3-71 optical filter ( > 440 nm)
incorporated in the water cell used to remove IR beams
from the lamp and to avoid heating. The UV irradiation en-
ters the reactor through the bottom optical window made of
Pyrex, which plays the role of a UV filter with transmittance
for
290 nm. The slurry is agitated with a magnetic stirrer
whose magnet spins in a perpendicular plane. The tempera-
ture is maintained at 295 K using thermostated water in the
jacket. The headspace of the photoreactor has three necks:
one for liquid or slurry sampling through a septum; a cen-
tral one with a vacuum-tight valve for gas sampling; and
another one for possible introduction of products. In gen-
eral, degradations were performed with the oxygen from
the air since the kinetics, followed by analyses of the liquid
phase, have indicated that atmospheric air in the headspace
SCHEME 1. Developed formulae of indigo and of indigo carmine.

48
VAUTIER, GUILLARD, AND HERRMANN
Irradiation of the Solid Deposits
Two milliters of an aqueous suspension containing 0.79 g
1
1
L
(3 mmol L 1) indigo and 5.09 mg L 1 (64 mmol L
)
TiO2 were deposited on an optical disk made of Pyrex or on
the inner bottom window of the photoreactor, which was
closed for the gas phase analyses. Water was evaporated
before irradiation.
Analyses
The decrease in the concentration of the dyes was ob-
served from the characteristic absorption at = 610 nm for
indigo carmine in water and at = 710 nm for indigo, us-
ing a UV–visible spectrophotometer (2000 Safas model).
The adsorption of the dyes on the catalyst was character-
ized by a FT-IR spectrometer in attenuated total reflection
(Perkin–Elmer).
The identification of reaction intermediates by HPLC
and/or GC/MS required preconcentration because of the
high dilution. For HPLC/UV detection of carboxylic acids,
preconcentration was performed with an ion-exchange
cartridge (SAX from Varian). It was activated with
water buffered at pH 7. Some carboxylic acids were con-
FIG. 1. Scheme of the slurry photoreactor used for photocatalytic dye
degradation.
of the photoreactor provided enough oxygen for the centrated using the solid phase microextraction (SPME)
oxidative degradation of pollutants. In these conditions, the technique with a 85- m-thick polyacrylate film, commer-
necks were open to air.
cialized by Supelco. For some acids, the fiber of the sy-
ringe was derivated with a solution of pyrenyldiazomethane
(PDAM). The detection of aromatic compounds by GC/MS
required preconcentration by liquid–liquid extraction: after
Determination of the Efficient Photon Flux
The efficient photon flux was calculated from the radi- acidification at pH 1 with H2SO4 and saturation by Na2SO4,
ant flux (in mW/cm2) measured with a radiometer (United extraction was performed three times with 20 mL of solu-
Detector Technology; model 21 A), calibrated against a tion using 10 mL of diethyl ether, and the solution dried
microcalorimeter and by taking into account the lamp under vacuum and the resulting solid dissolved in 100
L
spectral distribution, the optical filter transmission, and the of N,O-bis(trimethylsilyl)trifluoroacetamid (BTFA) before
absorbance spectrum of titania, as commonly done in our injection in the GC/MS. The analyses of intermediate prod-
previous studies.
ucts were performed with a HPLC Sarasep column (length,
30 cm; inner diameter, 7.8 mm) equipped with a Waters
486 UV detector at 210 nm. The mobile phase was com-
posed of H2SO4, 5 10 3 M and the flow rate was 0.7 mL
Irradiation of the Slurry
The reactor was filled with 20 mL of an aqueous suspen- min 1. Some of the intermediate products were also identi-
sion ofTiO2 (50mg, optimum concentration ofcatalyst, asin fied byGC-MS, usinga HP 5890seriesII gaschromatograph
Ref. (22)) containing either indigo (total content: 200 ppm (Chrompack CP-SIL 5-CB column; length, 25 m; inner
(i.e., 763 mol L 1), both dissolved (2 ppm) and mainly as diameter, 0.25 mm; film thickness, 1.2 m), equipped with
a quasi-colloidal suspension) or indigo carmine (29 ppm or a HP 5971 mass selective detector operating in the elec-
64 mol L 1, totally dissolved). The initial corresponding tron impact mode. The injections were made in the splitless
pH’s were equal to 4.6 and 4.8, respectively. Prior to illumi- mode. The column program was 393–573 K (heating rate,
nation, the suspension was magnetically stirred in the dark 3 K min 1; hold time, 5 min).
for 30 min, corresponding to the time needed to establish
For the kinetics of gas products evolution, i.e., CO2, the
the adsorption/desorption equilibrium at room tempera- headspace was preliminarily filled with pure oxygen with
ture. During irradiation, stirring was maintained to keep all the necks closed. For gas sampling, the loop of the GC
the mixture in suspension. At regular intervals, samples of conductivity detector and the gas tubing connecting it to the
0.5 mL were withdrawn for analysis. Before analyses they valve of the photoreactor were evacuated and then main-
were centrifuged or filtered through 0.45- m Millipore fil- tained under static vacuum (P 10 2 Torr). The headspace
ter disks to separate TiO2 particles.
gas volume was then expanded for a few seconds in the

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
49
evacuated loop of the CD chromatograph. The valve was a HPLC Waters IC-Pak Anion column (length, 50 mm;
closed and the gas contained in the loop was injected into inner diameter, 4.6 mm) for the anions (mobile phase,
the chromatograph by flushing the vacuum-tight six-way borate gluconate; flow rate, 0.9 mL min 1) and with a HPLC
valve with the carrier gas. After injection, the loop and the Vydac column (length, 300 mm; inner diameter, 7.8 mm) for
tubing were evacuated until the next injection. The stainless the cations (mobile phase, HNO3 , 2.5 10 3 M; flow rate,
steel tubing connecting the photoreactor to the loop had 1.5 mL min 1).
a small diameter of 1/16 in. to provide a total volume of
(loop + tubing) limited to 2 mL. Since the headspace has a
volume of 80 mL, this means that every sampling caused an
expansion of 2.5% , i.e., a loss of 2.5% of gaseous products,
which was taken into account for corrections. This proce-
dure caused limited removal of gas for sampling without
disturbance of kinetics determination. In addition, there
was no contamination of the gas phase by the carrier gas.
RESULTS AND DISCUSSION
1. Photocatalytic Degradation of Indigo and of Indigo
Carmine in Water by TiO2/UV
1.1. Real Photocatalytic Nature of the Reaction
To obtain relevant information about the photocatalytic
CO2 was analyzed with an Intersmat IGC 120-MB gas chro- degradation, it was necessary to carry out experiments from
matograph equipped with a Porapack Q column (length, which any possible direct photolysis was excluded. Exper-
3 m; inner diameter, 1/4 in.), connected to a catharometer iments were made (i) in the absence of TiO2 (neat photo-
detector.
chemical regime) and (ii) in the presence of UV-illuminated
Mineralization into CO2 was also determined in parallel SiO2 to detect any possible contact mass effect. In both
by measurements of the total organic carbon (TOC) using cases, no disappearance of indigo carmine was observed.
a Shimadzu 5050-A TOC analyzer equipped with a SFM A complete disappearance of indigo carmine could only be
5000 A solid sample module.
observed with the simultaneous presence of titania and of
The formation of inorganic ions was detected with a UV light, as shown in Fig. 2. This indicates that the system
Waters 431 electrical conductivity detector, equipped with is working in a pure photocatalytic regime.
FIG. 2. Disappearance of indigo carmine (C0 = 61 mol L 1). Curve A: UV irradiation (
290 nm) alone. Curve B: UV irradiation with inert
SiO2. Curve C: UV irradiation with TiO2.

50
VAUTIER, GUILLARD, AND HERRMANN
FIG. 3. Kinetics of adsorption of indigo carmine in the dark.
1.2. Adsorption of Indigo and of Indigo Carmine on TiO2
partially on titania, and absolutely not on SiO2. This re-
veals that the adsorption requires electrostatic interactions
between the dye molecule and the hydroxyl groups of the
photocatalyst.
Therefore, the difference in the adsorption mode be-
tween indigo and indigo carmine may produce some
changes in the kinetics and in the mechanism of degradation
(see next paragraphs).
The kinetics of adsorption in the dark is presented
in Fig. 3. From the decrease in concentration of indigo
carmine, the amount adsorbed was determined and cor-
2
responded to a surface coverage of 0.13 molecules nm
.
In comparison with the maximum surface density of
titania’s OH groups (5 molecules nm 2 (23)), a relative sur-
face coverage of ca. 3% can be estimated.
1.3. Photocatalytic Degradation of Indigo Carmine
and of Indigo
FT-IR investigations have been performed on each dye
in the presence and in the absence of the photocatalyst. The
resultsare the following:the modification ofthe asymmetric
stretching vibrations of benzenesulfonic acids at 1081 and
1103 cm 1 caused by the absence or the presence of TiO2 in
indigo carmine indicates that the adsorption involved one
of the sulfonate groups of the dye (Fig. 4, top). Adsorp-
tion through sulfonate groups has already been observed
in the case of dye alizarin red (17). Because of the extreme
positions of the sulfonate groups, the flat indigo carmine
molecule may be considered as adsorbed perpendicularly
to the surface of titania or at least obliquely, depending
on its environment in the adsorbed layer. By contrast,
for indigo, the shifts of the aromatic stretching vibration
1.3.1. Kinetics of disappearance. The kinetics of disap-
pearance of indigo carmine is represented in Fig. 2, with
a previous adsorption period in the dark of 30 min. After
a short (2 min) induction period, the disappearance was
achieved within less than 8 min.
The initial catalytic quantum yield is defined as the ratio
of the number of molecules transformed per second to the
number of incident efficient photons (i.e., absorbable by
titania) per second.
= r0/ϕ.
[7]
0
The initial rate of disappearance of indigo carmine being
(1609 cm 1) and of the intense skeletal bands at 1172, 1125, equal to 0.90 mol L 1 min corresponds to 1.81 1014
1
1
and 1063 cm (Fig. 4, bottom) shows that indigo adsorp- molecules converted per second. The efficient photonic
tion on TiO2 mainly involves interactions with the aromatic flux ϕ, calculated as indicated above, was equal to 5.20
1
rings. It would imply that indigo molecules adsorb parallel 1016 photons s and the resulting initial quantum yield
to the surface of TiO2.
was found equal to
= r0/ϕ = 0.35% , in good agree-
0
The amounts of in indigo carmine adsorbed on various ment with previous results on other pollutants (12). It has
solids has been correlated to their point of zero charge to be mentioned that the kinetics r = f(C) generally fol-
(PZC): PZC = 1.8, 9.1, and 6.7 for SiO2, Al2O3, and TiO2, lows a Langmuir–Hinshelwood mechanism (12, 22), with a
respectively. The dye was completely adsorbed on alumina, plateau at C > 5 10 3 mol L 1 and a linear increase at low

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
51
FIG. 4. (Upper figure): Changes of infrared absorption bands of indigo carmine. Curve A: without TiO2. Curve B: in the presence of TiO2. (Lower
figure): Changes of infrared absorption bands of indigo. Curve A: without TiO2. curve B: in the presence of TiO2.
concentrations. This catalytic quantum yield could be con- reported in Fig. 5 for indigo carmine and for indigo. The
sequently higher for higher concentrations, but presently it difference between the observed and the expected quan-
is high enough to ensure a reasonable degradation of di- tities of CO2 is probably due to the persistence of some
luted indigo carmine solutions. This confirms the sensitivity intermediate products and to the loss of volatile ones. The
of this dye to oxidizing agents (24).
identification of the main organic intermediate products
will be presented further. The quantity of sulfate ions is
lower than that expected from stoichiometry, which can be
explained by the strong adsorption of SO24 at the surface
of titania, as previously indicated (25). Figure 5 shows a
difference in the initial formation of nitrogen-containing
ions between indigo and indigo carmine. This point will be
1.3.2. Mineralization. The total degradation leads to the
conversion of organic carbon into gaseous CO2, whereas ni-
trogen and sulfur heteroatoms are converted into inorganic
ions, such as nitrate and ammonium, and sulfate ions, re-
spectively. The formation of CO2 and of inorganic ions is

52
VAUTIER, GUILLARD, AND HERRMANN
FIG. 5. Evolution of sulfate, nitrate, and ammonium ions in the solution and of gaseous CO2 during photocatalytic degradation of indigo carmine
(A) and of indigo (B).
discussed further. For both dyes, ammonium ions formed allowing nitrogen atoms to conserve their initial ( 3)
are slowly oxidized into nitrate ions, corresponding to the oxidation state.
stable maximum oxidation state of nitrogen (+5).
1.3.4. Identification of intermediate products. The inter-
mediates generated during the degradation process were
1.3.3. Influence of sulfonate groups. Sulfonate groups in
analyzed both by HPLC and by GC-MS and identified by
indigo carmine are responsible for the differences between
comparison with commercial standards and by interpreta-
indigo and indigo carmine behaviors. As seen in Section 1.2,
tion of their fragment ions in the mass spectra. It has been
the adsorption mode is different for both dyes. From
attempted to identify (i) the main aromatic metabolites re-
Fig. 5A, SO24 ions appear as a primary product since its ini-
sulting from both indigo and indigo carmine and (ii) espe-
tial rate (d[SO24 ]/dt0 = 0.6 mol L 1 min 1) is close to the
cially the carboxylic acid intermediates since their decar-
overall initial rate r0 (r0 = d[IC]/dt = 0.9 mol L 1 min 1).
boxylation according to the “photo-Kolbe” reaction
The difference between both initial rates can be ascribed to
a partially irreversible adsorption of SO24 ions, already ob-
R COO + h+ → R + CO2
[6]
–
served for sulfur-containing pesticides but without being
detrimental for photocatalysis (25, 26). Therefore, the ini- is the main source of CO2 evolution all along the various
tial step of IC degradation can be ascribed to the cleavage of steps of the degradation process. These intermediates are
–
the C S bond, leading to monosulfonated indigo and sulfate listed in Table 1. The main aromatic metabolites for in-
ions. The adsorption mode ofindigo carmine through itssul- digo were 2-nitrobenzaldehyde and anthranilic (2-amino-
fonate groups would favor the attack by photo-generated benzoic) acid. This clearly indicates that indigo initially re-
oxidizing agents, such as h+ and/or OH .
acts at the level of the central double bond, in agreement
At the beginning of degradation, nitrate and ammo- with an adsorption parallel to the surface as indicated in
nium ions are simultaneously formed f+or indigo carmine Section 1.2.
1
1
(r0(NO3 ) = 0.2 mol L min 1, r0(NH4 ) = 1.9 mol L
Nitrobenzene could only be detected in IC degradation.
min 1), whereas indigo only produces ammonium ions This could be related to the initial degradation step of IC
(r0(NH+4 ) = 2.8 mol L 1 min 1). These results are related involving the attack of the aromatic ring by abstraction of
to the influence ofthe sulfonate groups. The electrophilicat- the sulfonic substituent linked to the surface of titania. The
tack of nitrogen atoms in the dye molecule may be favored subsequent photocatalytic degradation of these aromatic
by the presence of the sulfonate electro-attractor groups, intermediates has been well known since our previous stud-
which weaken the electronic density of the aromatic rings. ies (for instance, Ref. (27) for nitrobenzene and Ref. (28)
By contrast, the initial steps of the degradation of the flat for substituted benzoic acids).
molecule of indigo, supposed to adsorb parallel to the sur-
The identification of aliphatic fragments, mainly carboxy-
face as mentioned above, do not concern the amino groups, lic acids (Table 1), is particularly instructive. Considering

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
53
TABLE 1
associated with that of C3-carboxylic (acrylic, tartronic, and
pyruvic) acids.
Main Intermediate Products Detected during Photocatalytic
Degradation of Indigo Carmine and of Indigo
==
The central C C double bond has been found very reac-
tive since the previous study of gas phase photo-oxidation
of propene producing an epoxide as a primary product (29).
Analytical
techniques
Retention
time (min)
Molecular
weight (g/mol)
Compounds
==
Presently, the initial step of the reaction could be the C
C
double-bond cleavage, thus explaining the initial evolution
of CO2 (d[CO2]/dt > 0 at tUV = 0) resulting from anthranilic
acid formation followed by its decarboxylation.
2-Nitro-benzaldehyde
and/or 2,3-dihydroxy-
indoline
GC-MS
151
Anthranilic acid
HPLC-UV
GC-MS
HPLC-UV
GC-MS
HPLC-UV
GC-MS
GC-MS
HPLC-UV
GC-MS
HPLC-UV
HPLC-UV
GC-MS
GC-MS
HPLC-UV
GC-MS
21.8
7.9
137
150
134
The part of intermediates originating from the nondisso-
==
ciation of the central C C double bond concerns fumaric
Tartaric acid
Malic acid
and 2,3-dihydroxyfumaric acids as well as two amino prod-
ucts: 3-amino-glyceric acid and 1,2-diaminoglycol. The pho-
tocatalytic degradation of malic acid and of its metabolites
has previously been accurately studied as a polycarboxylic
compound found in the biomass and in degradation prod-
ucts of complex molecules (30). Its present identification in
indigo degradation products enables us to account for the
C2–C4 carboxylic metabolites observed.
The kinetics of formation of intermediate acids during
the degradation of indigo carmine and of indigo are rep-
resented in Fig. 6. The first acids formed after 20 min of
UV irradiation are oxalic, anthranilic, and malic acids, in
line with the complete bleaching of the dye in 8 min. The
shorter carboxylic acids appear subsequently, with a maxi-
mum amount obtained after ca. 30 min of UV irradiation.
All the intermediate products, except acetic acid, are de-
graded within 1 h of UV irradiation, in agreement with the
time of formation of CO2 (1 h). Acetic acid requires a longer
time for mineralization, as generally observed. A tentative
mechanism of degradation inferred from these data is re-
ported in Scheme 2.
8.9
Amino-fumaric acid
Pyruvic acid
131
88
8.4
Malonaldehydic acid
Malonic acid
9.3
9.3
104
3-Amino propenoic acid
Glycolic acid
87
76
11.3
6.4
Oxalic acid
HPLC-UV
GC-MS
HPLC-UV
GC-MS
90
Acrylic acid
17.9
59
121
3-Amino; 2,3-dihydroxy-
propanoic acid
Acetic acid
HPLC-UV
13.8
60
the developed formulae, it can be conceived that the for-
mation of nitrobenzene produces unsaturated trans-C4 fu-
maric acid, which by oxidation, generates malic and tartaric
acids. In parallel, the formation of anthranilic acid can be
FIG. 6. Evolution of the main intermediate acids detected during the photocatalytic degradation of indigo carmine and of indigo. (A) Cn acids
(n 4); (B) C2 acids.

54
VAUTIER, GUILLARD, AND HERRMANN
SCHEME 2. Mechanism of the photocatalytic degradation of indigo carmine and of indigo.

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
55
The different metabolites duly identified are logically re- hydroquinone known as the last aromatic molecule found
ported accordingto their decreasingmolecular weight. Pho- before the aromatic ring opening. Presently, no C5 to C10
tocatalytic oxidation in water is not selective by contrast aliphatic acids were detected, corresponding to fragments
with selective mild oxidation in a pure gaseous or liquid consecutive with the ring opening. Either they were in
phase of aliphatic or substituted aromatic hydrocarbons as minute undetectable amounts or they remained mainly at
discussed in Ref. (12). Two oxidative agents can be con- the surface or in the double layer without transient desorp-
sidered: the photo-produced holes h+ and/or OH radicals, tion in water.
which are known as strongly active and degrading but non-
2. Decolorization of Indigo Carmine in an Aqueous
Suspension of TiO2 Irradiated by Visible Light
selective agents. Holes are mainly involved in the decar-
boxylation reaction (“photo-Kolbe”) (Eq. [6]). OH radi-
cals can be generated by various reactions:
Irradiation by visible light induces changes in the decol-
orization mechanism. Although TiO2 cannot be directly
activated by visible irradiation with an energy <3.1 eV
(band gap energy of anatase), visible light can induce the
excitation of the dye molecule. Previous works have re-
ported the electron injection from excited dye molecules
into polycrystalline TiO2 electrodes (30). Electron transfer
from adsorbed dye molecules to surface-acceptor sites was
proved to be effective for molecules in the excited state
since (i) the energy level of the donor orbital of the excited
dye molecule is sufficiently high above the conduction band
edge of the oxide and (ii) the density of acceptor sites often
increases toward the edge of the conduction band (31, 32).
Such a transfer has already been observed for cresyl violet
(32).
—(i) oxidation of water by holes
(H2O ⇔ H+ + OH ) + h+ → H+ + OH ;
[8]
—(ii) transient formation of hydroperoxide radicals
O2(g) + e → O2
[9]
[10]
[11]
[12]
O2 + H+ → HO2
2HO2 → H2O2 + O2
H2O2 + e → OH + OH .
In the present case, the OH radicals can break the var-
ious C–N and C–C bonds of the chromophore group as
described in Scheme 2, thus accounting for the various
metabolites identified.
The aromatic intermediates found (anthranilic acid,
nitrobenzaldehyde, nitrobenzoic acid, and nitrobenzene)
undergo successive attacks by OH radicals, giving hydrox-
ylations and/or substitutions generally leading to hydroxy-
2.1. Kinetics of Decolorization
The kinetics of bleaching is represented in Fig. 7. The
total decolorization of indigo carmine by visible-irradiated
FIG. 7. Kinetics of decolorization (bleaching) of indigo carmine using visible light. Curve A: without any solid. Curve B: with photo-inactive SiO2.
Curve C: with TiO2 photocatalyst.

56
VAUTIER, GUILLARD, AND HERRMANN
titania was reached within 16 h, following a first-order ki- species at the origin is the photo-produced hole h+. The
netics (k = 8 10 3 min 1). A solution without titania and subsequent oxidizing and degradation chemical agent is the
a blank suspension of photo-inert SiO2 were irradiated for OH radical mostly formed by oxidation of water (Eq. [8]).
16 h in the same conditions (see Fig. 7). The decrease in Presently, the absence of holes (see Scheme 3B) implies the
concentration was equal to 7 and 17% , respectively. From absence of OH radicals and subsequently of any degrada-
the two blank tests, it was inferred that visible-light-induced tion, as observed for quinones (34). Since no subsequent
decolorization necessarily involves titania. However, tita- degradation occurs, this means that this electron transfer
nia being quite photo-inactive at > 440 nm, the process of should be stoichiometric. This has been checked by using
the photo-bleaching has to involve the dye as a sensitizer.
an initial number of indigo carmine molecules 2.5 times
higher than the maximum number of O2 sites deduced
from a saturation coverage of 5 1018/m2 (23): the discol-
orization did not occur even after 36 h of continuous visible
illumination.
Eventually, it has to be mentioned that our results are
in contrast to the degradation obtained in the visible with
alizarin red (17) and with acid orange 7 (33).
2.2. Nature of the Reaction
No intermediates and no final mineralization products
were found in the solution, indicating that the dye is only
decolorized but not degraded. This was confirmed by TOC
analyses, which revealed not only that no degradation oc-
curred but also that there remained the same quantity of
IC adsorbed after visible irradiation (0.15 molecule/nm2
as before irradiation at the end of the adsorption period
in the dark (0.13 molecule/nm2). Scheme 3 represents an
energy diagram under UV and visible irradiation. Under
UV, titania gets excited and indigo carmine can cede an
electron to its valence band either directly or indirectly
via an attack by OH radicals. Under visible light, only
indigo carmine becomes excited at > 440 nm with the
promotion of an electron in an upper orbital, whose rel-
ative energy level enables the transfer of one electron into
the conduction band of the catalyst. This transfer could be
favored by a driving force consisting of a subsequent elec-
tron transfer to an adsorbed oxygen molecule. The present
photo-bleaching of indigo carmine only corresponds to de-
colorization by loss of conjugation of the double bonds in
the whole molecule induced by the electron transfer to ti-
tania. In titania-based photocatalysis, the actual oxidizing
3. Photocatalytic Degradation of Solid Indigo
The poor solubility of indigo prompted us to assess a
possible photocatalytic reaction in a biphasic solid–solid
system. The possible lack of water in the photocatalytic
degradation of solid indigo preventing the formation of
OH radicals (Eqs. [8]–[12]) has been taken into consid-
eration. Actually, the preparation of the layer constituted
by a mixture of titania and of solid indigo consisted merely
of evaporation at room temperature of a slurry deposited
on the optical disk support. This enabled full coverage of
titania in water at tUV = 0. In addition, photodegradation
produces water, which maintains a sufficient hygroscopic
level at the surface and (ii) ambient air moisture always
keeps titania’s surface (which is highly hydrophilic) in an
hydrated state.
The results obtained on optical disks (Fig. 8) clearly indi-
cated a photobleaching of solid indigo mechanically mixed
with powder titania.
3.1. Kinetics of Disappearance
The disappearance of solid indigo is reported in Fig.
9. It follows an apparent first-order kinetics (k = 9
10 3 min 1). The final level is reached within 9 h. This much
longer period of time compared with that of the disappear-
ance of indigo carmine in water has to be attributed to the
slow diffusion of the oxidizing agents (h+, OH ) on the
surface of the photocatalyst, diffusion becoming the rate-
limiting step of the whole process.
3.2. Mineralization
The formation of CO2 was chosen as experimental proof
for degradation. The maximum amount of CO2 formed
was achieved in 20 h and corresponded to a conversion of
33% , as shown in Fig. 9. Three reasons may explain this
result: (i) the loss of volatile species formed during the
SCHEME 3. Mechanism of TiO2-assisted photodecolorization (A)
under UV irradiation and (B) under visible light.

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
57
FIG. 8. Decolorization of solid indigo.
FIG. 9. Kinetics of the photocatalytic degradation of solid indigo. Curve A: disappearance of indigo. Curve B: evolution of CO2 formed during
UV irradiation. Insert: semi-log linear transform of curve A.

58
VAUTIER, GUILLARD, AND HERRMANN
TABLE 2
CONCLUSIONS
Main Intermediate Products Detected during Photocatalytic
Degradation of Solid Indigo
The photocatalytic degradation of indigo and of in-
digo carmine has been successfully demonstrated when us-
ing UV-irradiated titania-based catalysts. In addition to a
prompt removal of the color, photocatalysis was simulta-
neously able to oxidize the dye, with an almost complete
mineralization of carbon and of nitrogen and sulfur het-
eroatoms into innocuous compounds. A detailed degrada-
tion pathway, based on careful identification of intermedi-
ate products, is proposed.
The irradiation of titania in the visible light produces a
photoinduced decolorization of the dye but without any
degradation, corresponding to a stoichiometric electron
transfer from the dye, excited in the visible irradiation, to
titania. The positive decolorization and degradation of solid
indigo, mechanically mixed, constitutes an encouraging re-
sult for self-cleaning titania-coated objects (glasses, steel,
walls, etc.) fouled by solid dirt particles.
Analytical
techniques
Retention
time (min)
Molecular weight
(g/mol)
Compounds
2-Nitrobenzoic acid
Nitrobenzene
Malic acid
GC-MS
GC-MS
HPLC-UV
GC-MS
167
123
134
8.8
Fumaric acid
HPLC-UV
GC-MS
GC-MS
15.8
116
148
Dihydroxyfumaric
acid
Glycolic acid
HPLC-UV
GC-MS
11.5
76
Oxalic acid
Acetic acid
HPLC-UV
GC-MS
HPLC-UV
6.4
13.8
90
60
The ensemble of these results clearly suggest that
TiO2/UV photocatalysis may be envisaged as a method for
treatment of diluted colored waste waters in textile indus-
tries, especially in sunny semi-arid countries where water
has to be recycled. Large solar pilot experiments are pro-
grammed for this purpose.
degradation, (ii) the persistence of non-photodegraded by-
products, and (iii) the heterogeneity of the deposit, which
prevents some solid indigo particles from being directly in
contact with titania crystallites.
ACKNOWLEDGMENTS
3.3. Identification of Intermediate Products
This work was supported by a CMCU Project (99F1201). The authors
are grateful to Michelle Besson and Marie-Laure Moulut for the TOC
analyses.
The same procedures and analytical techniques as those
in Section 1.3.4. were used to identify the intermediate
products. The main molecules detected are presented in
Table 2. They are similar to those obtained in water sus-
pensions so that the same degradation mechanism can be
proposed. Two additional nitro compounds (Table 2) were
identified by GC/MS, suggesting a transient accumulation
of these products in the absence of water, thus favoring their
detection. They have been included in the general degra-
dation pathway in Scheme 2.
The photocatalytic degradation of solid indigo, mechan-
ically mixed with powder titania and placed in a fixed cata-
lytic bed, constitutes one of the first examples (or maybe
the first one) of a photocatalytic reaction using a solid re-
actant. The only—and undesirable—known example is the
photo-(solar-)degradation of plastics or paints, pigmented
with deficientlypassivated titania. Naturalweatheringgrad-
ually produces a “chalking effect,” which is caused by the
progressive elimination ofthe organicpolymericbinder and
via a (slow) photocatalytic combustion of the organic mat-
ter in contact with an ill-protected titania-based opacifier
(18, 35).
REFERENCES
1. “Color Chemistry. Synthesis, Properties and Applications of Organic
Dyes and Pigments” (H. Zollinger, Ed.), 2nd revised ed. VCH, New
York, 1991.
2. Dejohn, P. B., and Hutchins, R. A., Tex. Chem. Color. 8, 69 (1976).
3. Patil, S. S., and Shinde, V. M., Environ. Sci. Technol. 22, 1160 (1988).
4. More, A. T., Vira, A., and Fogel, S., Environ. Sci. Technol. 23, 403
(1989).
5. Slokar, Y. M., and Le Marechal, A. M., Dyes Pigments 37, 335 (1998).
6. “Photocatalysis and Environment. Trends and Applications” (M.
Schiavello, Ed.), Kluwer, Dordrecht, 1988.
7. “Photocatalysis. Fundamentals and Applications” (N. Serpone, and
E. Pelizzetti, Eds.), Wiley Interscience, New York, 1989.
8. Guillard, C., Herrmann, J. M., and Pichat, P., Catal. Today 17, 7 (1993).
9. “Photocatalytic Purification and Treatment of Water and Air” (H. A.
Al-Ekabi, and D. Ollis, Eds.), Elsevier, Amsterdam, 1993.
10. Bahnemann, D. W., Cunningham, J., Fox, M. A., Pelizzetti, E., Pichat,
P., and Serpone, N., in “Aquatic Surface Photochemistry” (R. G. Zeep,
G. R. Helz, and D. G. Crosby, Eds.), p. 261. F.L. Lewis, Boca Raton,
FL, 1994.
11. Legrini, O., Oliveros, E., and Braun, A. M., Chem. Rev. 93, 671 (1993).
12. Herrmann, J. M., chapter 9 in “Environmental Catalysis” (F. Jansen
and R. A. van Santen, Eds.), Catalytic Science Series, pp. 171–194.
Imperial College Press, London, 1999.
The positive decolorization and degradation of solid in-
digo, illustrated by Fig. 8, constitutes an encouraging re-
sult for the use of outdoor self-cleaning titania-coated sur-
faces such as glasses (36–38), stainless steel (37), walls, etc.)
fouled by solid dirt particles.
13. Blake, D. M., “Bibliography of Work on the Photocatalytic Removal
of Hazardous Compounds from Water and air,” NREL/TP-430-22197.
National Renewable Energy Laboratory, Golden, CO, 1997 and 1999.

PHOTOCATALYTIC DEGRADATION OF DYES IN WATER
59
14. Pulgarin, C., Pajonk, G. M., Bandara, J., and Kiwi, J., “Meeting 27. Maillard-Dupuis, C., Guillard, C., and Pichat, P., New J. Chem. 18, 941
ACS Division of Environmental Chemistry, Anaheim, CA, 1995,”
Paper No. 232, p. 767. Am. Chem. Soc., Washington, DC, 1995.
15. Zhang, F., Zhao, J., Shen, T., Hidaka, H., Pelizzetti, E., and Serpone,
N., Appl. Catal. B 15, 147 (1998).
16. Wu, G., Wu, T., Zhao, J., Hidaka, H., and Serpone, N., Environ. Sci.
Technol. 33, 2081 (1999).
(1994).
28. Herrmann, J. M., Tahiri, H., Guillard, C., and Pichat, P., Catal. Today
54, 131 (1999).
29. Pichat, P., Herrmann, J. M., Disdier, J., and Mozzanega, M. N., J. Phys.
Chem. 83, 3122 (1979).
30. Vlachopoulos, N., Liska, P., Augustynski, J., and Gra¨tzel, M., J. Am.
Chem. Soc. 110, 1216 (1988).
31. Spitler, M., and Calvin, M., J. Chem. Phys. 66, 4294 (1976).
32. Kietzmann, R., Willig, F., Weller, H., Vogel, R., Nath, D. N.,
Eichberger, R., Liska, P., and Lehnert, J., Mol. Cryst. Liq. Cryst. 194,
169 (1991).
33. Vinogdopal, K., Wynkoop, D. E., and Kamat, P. V., Environ. Sci. Tech-
nol. 30, 1660 (1996).
34. Nasr, C., Vinogdopal, K., Ficher, L., Hotchandani, S., Cattopadhyay,
A. K., and Kamat, P. V., J. Phys. Chem. 100, 8436 (1996).
35. Kaempf, G., J. Coating Technol. 51, 51 (1979).
36. Fujishima, A., Look Jpn. 41, 47 (1995).
37. Fernandez, A., Lassaletta, G., Jimenez, V. M., Justo, A., Gonzalez-
Elipe, A. R., Herrmann J. M., Tahiri, H., and Ait-Ichou, Y., Appl.
Catal. B 7, 49 (1995).
17. Liu, G., Li, X., Zhao, J., Horikoshi, S., and Hidaka, H., J. Mol. Catal.
A 153, 221 (2000).
18. Ka¨mpf, G., and Vo¨lz, H. G., Farbe Lack 74, 37 (1968).
19. “The Sigma Aldrich Handbook of Stains, Dyes and Indicators” (F. J.
Green, Ed.), pp. 403–406. Aldrich Chemical, Milwaukee, WI, 1990.
20. Rademacher, P., and Kowski, K., Chem. Ber. 125, 1773 (1992).
21. Yang, Y., Wyatt, D. T., and Bahorsky, M., Textile Chem. Color. 30, 27
(1998).
22. Herrmann, J. M., Catal. Today 24, 157 (1995).
23. Boehm, H. P., Adv. Catal. 16, 179 (1966).
24. Roberts, E. L., Burguieres, S., and Warner, I. M., Appl. Spectrosc. 52,
1305 (1998).
25. Herrmann, J. M., Guillard, C., Arguello, M., Agu¨era, A., Tejedor,
A., Piedra, L., and Fernandez-Alba, A., Catal. Today 54, 353
(1999).
38. Heller, A., Acc. Chem. Res. 28, 503 (1995).
26. Kerzhentsev, M., Guillard, C., Herrmann, J.-M., and Pichat, P., Catal. 39. Romeas, V., Pichat, P., Guillard, C., Chopin, T., and Lehaut, C., New
Today 27, 215 (1996). J. Chem. 23, 365 (1999).