Kinetic of adsorption and of photocatalytic degradation of phenylalanine effect of pH and light intensity

Article abstract of DOI:10.1016/J.apcata.2010.03.054

Phenylalanine (Phe) was chosen to study the TiO₂ photocatalytic degradation of amino acids, which are at the origin of the formation of odorous compounds after chlorination. The photocatalytic degradation has been investigated in aqueous solutions containing TiO₂ suspensions as photocatalyst, in order to assess the influence of various parameters, such as adsorption, initial concentration, pH and radiant flux on the photocatalytic process. Results showed no correlation between dark adsorption and photocatalytic degradation. A multilayer kinetic was observed in the dark with a monolayer corresponding to less that 1% of OH covered, whereas Langmuir-Hinshelwood model seems to modelize the photocatalytic disappearance of Phe. However, even if the form of the curve is similar to L-H model, the degradation of phenylalanine is not a kinetic of L-H as we could plan it by considering the adsorption of the phenylalanine in the dark. The study of the mineralization of carbon and nitrogen showed that nitrogen atoms were predominantly photoconverted into NH₄⁺ and a total mineralization of nitrogen and carbon seems occur. The identification of the by-products by LC-MS reveal mono- and di-hydroxylation and nitrogen-carbon (N-C) cleavage. The effect of pH showed an increase of adsorption under acid pH but a decrease of disappearance rate. The more efficient degradation was found at basic pH. The evolution of hydroxylated compounds of phenylalanine as a function of conversion revealed the presence of more hydroxylated compounds at natural pH and at basic pH compared to acid pH suggesting a modification of mechanism with solution pH. The effect of the radiant flux evaluated under different initial concentration of phenylalanine allowed us to determine that Κ increases by increasing the radiant flux, whereas Κ decreases or remains constant from about a value of 3.5 mW/cm². The disappearance rate as a function of radiant flux has been showed to reach a maximal value corresponding to a maximal quantum yield of 1.6%.

Full text of DOI:10.1016/J.apcata.2010.03.054

Applied Catalysis A: General 380 (2010) 142–148
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Kinetic of adsorption and of photocatalytic degradation of phenylalanine
effect of pH and light intensity
L. Elsellamia,b, F. Vocansonc,d, F. Dappozzea, E. Puzenata, O. Païssee, A. Houasb, C. Guillarda,∗
a IRCELYON, CNRS UMR 5256/Université Lyon1, 2 av. Albert Einstein, 69626 Villeurbanne Cedex, France
b Laboratoire de Catalyse et Environnement, Faculté des Sciences de Gabès, Tunisia
c Université de Saint-Etienne, F-42000 Saint-Etienne, France
d Laboratoire Hubert Curien, UMR 5516, F-42000 Saint-Etienne, France
e service Central d’analyse du CNRS, Echangeur de Solaize B.P. 22, 69390 Vernaison, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Phenylalanine (Phe) was chosen to study the TiO2 photocatalytic degradation of amino acids, which are
at the origin of the formation of odorous compounds after chlorination. The photocatalytic degradation
has been investigated in aqueous solutions containing TiO2 suspensions as photocatalyst, in order to
assess the influence of various parameters, such as adsorption, initial concentration, pH and radiant flux
on the photocatalytic process. Results showed no correlation between dark adsorption and photocat-
alytic degradation. A multilayer kinetic was observed in the dark with a monolayer corresponding to less
that 1% of OH covered, whereas Langmuir–Hinshelwood model seems to modelize the photocatalytic
disappearance of Phe. However, even if the form of the curve is similar to L–H model, the degradation of
phenylalanine is not a kinetic of L–H as we could plan it by considering the adsorption of the phenylala-
nine in the dark. The study of the mineralization of carbon and nitrogen showed that nitrogen atoms were
predominantly photoconverted into NH4+ and a total mineralization of nitrogen and carbon seems occur.
The identification of the by-products by LC–MS reveal mono- and di-hydroxylation and nitrogen–carbon
(N–C) cleavage.
Received 4 January 2010
Received in revised form 12 March 2010
Accepted 24 March 2010
Available online 1 April 2010
Keywords:
Phenylalanine
Adsorption
Photocatalysis
Kinetic
By-products
pH
Radiant flux
The effect of pH showed an increase of adsorption under acid pH but a decrease of disappearance
rate. The more efficient degradation was found at basic pH. The evolution of hydroxylated compounds
of phenylalanine as a function of conversion revealed the presence of more hydroxylated compounds at
natural pH and at basic pH compared to acid pH suggesting a modification of mechanism with solution
pH. The effect of the radiant flux evaluated under different initial concentration of phenylalanine allowed
us to determine that k increases by increasing the radiant flux, whereas K decreases or remains constant
from about a value of 3.5 mW/cm2. The disappearance rate as a function of radiant flux has been showed
to reach a maximal value corresponding to a maximal quantum yield of 1.6%.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
formation of N-chloroaldimines [6] which are very stable com-
pounds. They are stable enough to be found in tap water (their life
Amino acids concentration in drinking water range from 0.33 to
1 ␮g/L [1] and can vary between 0 and 30 ␮g/L in several drinking
water plants of western France [2]. They are found behind filtra-
tion process [3] and come from the natural organic matter or from
microorganisms.
Amino acids are non-toxic compounds. However, it is known
that amino acids [4] can be precursors of odorous by-products.
Amino acids lead to the production, under chlorination, of nitriles,
chloramines and aldehydes [5], known as odorous but also to the
time is around 50 h at 20 ◦C but hundreds of hours at 15 ◦C, while
monochloramines’ half-life time is around 1 h). Their threshold of
odor is lower than 1 ␮g/L.
Amino acids are also interesting compounds, because they
are the building blocks of proteins and can help us to under-
stand better the mechanism of microorganisms’ photocatalytic
degradation. Actually, biochemical applications of TiO2−mediated
photoxidations, under solar light, are promising for the elimina-
tion of microorganism, e.g. bacteria, fungi, mold, and virus [7–10]
since the reaction occurs at ambient temperature, with addition
of non-toxic semiconductor, TiO2 which can be activated by solar
light.
∗
Corresponding author. Tel.: +33 44 72 53 16; fax: +33 4 72 44 53 99.
E-mail addresses: chantal.guillard@ircelyon.univ-lyon1.fr,
In the last decade, TiO2 heterogeneous photocatalysis has
proved to be one of the most powerful techniques for remediating
chantal.guillard@univ-lyon1.fr (C. Guillard).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.03.054

L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
143
environmental pollution in both liquid and gas phase reactions [11].
However, several points remain understood such as the absence of
correlation between the adsorption in the dark and the adsorption
determined considering Langmuir–Hinshelwood model.
Our objectives are first to contribute to a better understand-
ing of photocatalytic treatment for removing amino acids which
are the simpler molecules constituting microorganisms (DNA, RNA,
proteins, etc.) by studying kinetics of adsorption in the dark under
different pH, the kinetics of photocatalytic degradation at different
concentrations under different pH and light intensities.
2. Experimental
2.1. Materials
Fig. 1. Amount of phenylalanine adsorbed per gram of catalyst as a function of Phe
concentration remaining in solution at the equilibrium.
Pure (99% purity) phenylalanine was provided by Janssen Chim-
ica. The other chemicals used, HNO3 and NaOH were purchased
from Aldrich and used as received. Degussa P-25 titanium dioxide
(mainly anatase, 50 m2 g−1, nonporous) was used as the photocat-
alyst. Water used for preparation of samples was ultra-pure water,
filtered through a milli-Q PLUS 185 water system.
capillary potential = 4000 V, auxiliary gas = N2, flow rate = 12 L/min,
pressure = 55 psig.
The formation of carboxylic acids has been analyzed by LC
using a Waters 600 pump, a Waters 486 UV detector (detection
at 210 nm), and a Sarasep CAR-H (300 mm × 4.6 mm) column. The
flow rate was 0.7 mL/min. The injection volume was 100 ␮L and the
mobile phase was H2SO4 (5 × 10−3 mol/L).
2.2. Reactor and light source
The formation of nitrate ions has been followed using ionic
chromatography with a Dionex DX-120 pump and conductivity
detector, and an IonPac AS14A (250 mm × 4 mm) column. The flow
rate was 1 mL/min and the mobile phase was an alkaline buffer
(NaHCO3 (1.0 mmol/L) + Na2CO3 (8.0 mmol/L)).
The aqueous suspensions were irradiated in 100 mL open cylin-
drical reactor whose base contained an optical window with a
surface area of about 19.6 cm2. The output of a Philips HPK 125 W
high mercury lamp was filtered through a circulating water cell
(thickness = 2.2 cm), which avoided the solution warming by IR, and
a filter Corning 0.52 mW/cm2 allowing to cut off the wavelengths
below 340 nm. The radiant flux was measured using a VLX-3W
radiometer with a detector CX-365 (355–375 nm). The use of a grid
permits to decrease the intensity of the light in order to obtain the
intensity selected.
The formation of ammonium ions has also been followed using
ionic chromatography with a Dionex DX-120 pump and conduc-
tivity detector. In this case a CS 12A (250 mm × 4 mm) column. The
flow rate was 1 mL/min and the mobile phase was H2SO4 solutions
containing 610 ␮L of pure sulfuric acid).
3. Results and discussion
2.3. Photocatalytic experiments
3.1. Adsorption
A volume of 20 mL of phenylalanine solution with different
concentrations was used. 1.25 g/L of TiO2 is sufficient to absorb
all photons entering the photoreactor [12]. The degradation was
carried out at room temperature and at neutral pH (pH = 6.5)
unless otherwise stated, in particular for the study of pH influence.
Then, the suspension was first stirred in the dark until equilibrium
adsorption. Samples taken at different times of irradiation were
filtered through 0.45 ␮m Waters filters to remove TiO2 particles
before analyses. During the irradiation, aliquots of 200 ␮L of the
phenylalanine solutions were collected at regular times.
In order to ensure that the adsorption process had reached
equilibrium, the mixture (TiO2/water/phenylalanine) ([phenylala-
nine = 121 ␮mol/L]) was stirred in the dark and analysed as a
function of time. The phenylalanine adsorption reached equilib-
rium after about 30 min. The amount of Phe adsorbed at equilibrium
corresponds to less than 3% of the initial phenylalanine con-
centration which was measured from the remaining solution.
This amount corresponds to an area density of adsorbed Phe of
0.031 molecules/nm2. By considering that adsorption of phenylala-
nine occurs on OH group present on the TiO2 solid surface, which
represent about 5 OH/nm2 [13], less than 1% of OH covered.
This low coverage of OH could partially be explained firstly (i)
by taking into account the work of Tran et al. [14] which sug-
gest that effective adsorption for phenylalanine should be only on
basic terminal OH on the solid surface and secondly (ii) by the
small concentration of Phe used in our experimental observation
(120 ␮mol L−1). It is why we determined the effect of the initial
concentration of phenylalanine on the adsorption. In all cases, the
initial reaction mixture was maintained in the dark for 30 min
under stirring to reach equilibrium adsorption as reported in the
first part. The amount of Phe adsorbed per gram of TiO2 as a func-
tion of the Phe concentration presenting in solution (Ceq) indicates
that Langmuir model is not applicable for this type of molecule
(Fig. 1).
2.4. Analyses
The degradation of phenylalanine and the formation of hydrox-
ylated compounds (o, m, p hydroxyl-phenylalanine) were followed
by HPLC instrument with a Waters System which was comprised a
Waters 600 isocratic pump and a Waters 486 UV detector adjusted
at 212 nm. A Hypersil BDS C18 reverse phase column (125 mm long,
4 mm diameter), was used. The mobile phase was consisted of 3.8 g
ammonium acetate acidified with acetic acid at pH = 5. The flow
rate was 0.3 mL/min
The HPLC–MS identification of photoproducts was performed
with electrospray ionization (ESI) in positive and negative mode
using a Hewlett-Packard HP 1100 series LC–MSD using a col-
umn Hypersyl BDS C18 (125 mm × 4 mm, particle size 5 ␮m).
The mobile phase composition was H2O buffered at pH = 5 with
CH3–COO–NH4/acetonitrile at a ratio 5/95. The flow rate was
0.3 mL/min. The parameters of nebulization were the following:
The curve suggests that adsorption of Phe was a multi-
layer model. A first plateau corresponding to the monolayer was
observed at a value of about 0.43 mg of phenylalanine per gram

144
L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
Fig. 4. Evolution of coverage rate as a function of Phe equilibrium concentration.
Fig. 2. Kinetics of phenylalanine disappearance for different initial concentrations.
from a Langmuir–Hinshelwood type kinetic, r = kKC/1 + KC, since
the product of the adsorption constant, K, by the concentration,
cannot be neglected with respect to one in the denominator of r at
higher concentrations.
From this curve we can determine the rate constant k and
the adsorption constant K and determine the coverage rate
(ꢁ = KC/1 + KC) under UV.
The values of k and K are respectively 0.38 ␮mol/min and
0.26 L/␮mol. It is important to note that k is very important com-
pared to the value determined at a concentration of 240 ␮mol/L
whereas the disappearance rate is already nearly independent of
the concentration and that coverage rate is near 100% (Fig. 4). In
these conditions k ≈ r0 (k ≈ 0.2 ␮mol/min).
This result shows that even if the form of the curve is similar
to L–H model, the degradation of phenylalanine is not a kinetics of
L–H as we could plan it by considering the adsorption of the pheny-
lalanine in the dark. This result shows the importance to study the
adsorption of pollutant in the dark to check that the adsorption
obey to Langmuir model as recommended by Ohtani et al. [15].
However, we do not have to forget that under irradiation the surface
of solids can be modified leading to the modifications of adsorption.
An interesting value in photocatalysis is the quantum yield (ꢂ). It
is defined as the ratio of the initial rate (number of molecules trans-
formed per second (r)) divided by the number of incident efficient
photons absorbable by titania (ϕn) per second.
Fig. 3. Evolution of initial rate of disappearance of phenylalanine as a function of
the concentration of phenylalanine remaining in solution (Ceq).
of TiO2, that is to say 0.031 molecules/nm2. As the Phe adsorp-
tion is not modelized by Langmuir equation, we can assume that a
molecule of Phe is not adsorbed on one OH site but on several OH
groups or that OH groups are not the adsorption site.
3.2. Phenylalanine disappearance
The complete disappearances of Phe in aqueous solutions at
initial concentrations of 60, 90, 120, 181 and 240 ␮mol/L were
obtained upon illumination at ꢀ > 340 nm and a radiant flux ˚
equals to 3.5 mW/cm2 at 25, 35, 40, 50 and 55 min respectively
(Fig. 2). The direct photolysis at ꢀ > 340 nm was negligible indicat-
ing that disappearance observed in the presence of TiO2 and UV at
this wavelength is only due to photocatalytic process.
The initial rate of disappearance of phenylalanine is reported
as a function of the concentration of phenylalanine remain-
ing in solution (Ceq) (Fig. 3). Until a concentration of about
120 ␮mol/L, the initial rate is proportionnal to the Phe con-
centration in solution (r = kKCeq). Then, the initial rate became
independent of Phe concentration. This phenomenon is expected
r
(˚/E)S
NA
ꢂ =
,
ϕn
=
ϕn
with the radiant flux ˚ in W cm−2 (J s−1 cm−2), the radiant flux
energy E in J photons−1, NA the Avogadro’s number (NA) and the
surface of optical window S (19.6 cm2).
In the experimental conditions the quantum yield increases
until about 1.6%. At low concentration of Phe adsorbed on TiO2,
active species formed on TiO2 surface cannot react and recombine
leading to a lower quantum yield.
Fig. 5. TOC disappeared (a) and evolution of nitrate and ammonium (b) as a function of irradiation time.

L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
145
Table 1
Structure of Phe by-products identified by LC–MS with electrospray (ES).
Compounds (m/z)
Fragments positive mode
198/159/152/119/104/97
Fragments
negative mode
tR (min)
Structure
A-197
(MS): 196
4.3
(MS/MS): 196/157
B-197
C-181
198/159/152/119/104/97
(MS): 196
(MS/MS): 196/157
6.2
6.8
(MS): 182/165/147/136/180
(MS/MS):
(MS): 180
(MS/MS):
182/165/147/136/123/119/95/ 91
182/163/119/93
D-181
(MS): 182/136 (MS/MS):
182/136/91
(MS): 180
(MS/MS):
8.9
182/163/119
E-181
F-150
(MS): 182/164/136 (MS/MS):
182/164/147/136/119/91
(MS): 180
(MS/MS):
180/186/119/107
11.9
16.9
151/150/119/111/97
3.3. Mineralisation and transformation products
teric structure) as mentioned below and suggested by Möning et
al. [19].
Total mineralization is the ultimate step of the photocatalytic
degradation of organic molecules. In order to quantify this pro-
cess, total organic carbon (TOC) and the evolution of nitrate and
ammonium ions are followed as a function of time.
The TOC disappearance (TOC measured) and the evolution of
nitrate and ammonium are represented in Fig. 5.
The difference between Phe disappearance and its mineral-
ization reflects the presence of several intermediate compounds
during its oxidation.
Main intermediate products detected were identified by LC–MS
and are listed in Table 1.
TOC disappeared from the beginning of irradiation indicating
initial decarboxylation reaction. After complete elimination of Phe
(40 min), around 65% of the value of TOC expected from total miner-
alization was obtained and 1 h is necessary to mineralize about 95%
of carbon. The presence of about 0.5 ppm of TOC at the end is not
significant due to the precision of TOC measure. We can consider a
total mineralization of carbon.
These different structures correspond to the following reac-
tions:
NH4+ ions appear from the beginning of the degradation and are
the major one. The formation of nitrates occurred only after total
disappearance of Phe indicating that nitrate ions are not primary
oxidation by-products and come from intermediate compounds
Fig. 5(b). After an irradiation time of 4 h, total mineralization of
nitrogen was observed.
(i) Hydroxylation on the aromatic ring (compounds C–E),
(ii) dihydroxylation (compounds A and B),
(iii) deamination of heterocycle (compound F).
No decarboxylation compound has been evidenced whereas the
decarboxylation seems to be important regarding the evolution of
TOC. The absence of decarboxylation products could be explained
considering that amine intermediates formed could have important
interaction with residual OH of the C18 column and not be eluted
in our experimental conditions.
The more important formation of NH4+ ions is in agreement with
the work of Hidaka et al. [17] on different amino acids and with our
previous work [16]. Actually, oxidation degree of amine group and
ammonium are both equal to +3. The formation of ammonium can
be explained considering the reaction of e− on protonated amine (in
our pH conditions, more than 99% of Phe is present under ampho-
Five linear organic acids have also been identified i.e., oxalic,
oxamic, lactic, acetic and propanoic acids. Lactic and oxamic appear

146
L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
Table 2
Molecular fraction of different forms of phenylalanine.
Phenylalanine
pH = 10
pH = 6.5
pH = 2
0%
0%
80%
14%
100%
20%
86%
0%
0%
from the beginning of degradation suggesting that they come from
direct break of aliphatic chain. The three other acids are not primary
intermediates.
3.4. Influence of pH
The pH influences both the surface state of titania and the ion-
ized state of Phe (Table 2).
For pHs higher than the PZC (point zero charge) of titania, which
is around 6.5 for TiO2 Degussa P-25, the surface becomes negative
charged. At pH < PZC the surface is positive charged according to
the following equilibrium:
+
pH < PZC : Ti − OH + H+ ⇔ TiOH2
,
pH > PZC : Ti − OH + OH− ⇔ TiO− + H2O,
Considering the pKa values of Phe (pKa1 = 2.6 and pKa2 = 9.2)
the percentage of molecular fraction of cationic, amphoteric and
anionic forms are reported in Table 1 with three different pH values
(pH = 2, 6.5 and 10).
It can be seen that the predominant forms are varied with three
pH values.
Fig. 7. Evolution of initial rate of Phe as a function of (a) Phe concentration and (b)
radiant flux.
3.4.1. Influence of pH on sorption isotherm
Firstly, we observed the evolution of Phe adsorbed on TiO2 in
the dark with three pH values as (Fig. 6). Even if the pH values are
different, the forms of these curves suggest a multilayer adsorption.
The more important adsorption is observed at acid pH. At basic
pH, the adsorption of phenylalanine is about 30% higher than this
one found at neutral pH.
This behaviour is not understandable considering the surface
state of titania and the ionization state of phenylalanine if we con-
sider that phenylalanine adsorbed in the same way at three pH.
This behaviour could suggest that considering the pH of the solu-
tion, the part of phenylalanine adsorbed on TiO2 could be different.
At acid pH, the adsorption will be made mainly by carboxylic groups
whereas at basic pH adsorption predominantly will occur by NH2
groups.
3.4.2. Influence of pH on the disappearance
Initial disappearance rates of Phe at acid, neutral and basic pH
were determined at different initial concentration of phenylala-
nine. The order of reactivity of phenylalanine under three pH values
is the following:
r0(pH = 10) > r0(pH = 2) > r0(pH = 6.5)
Considering Table 1, Phe seems to be more easily degraded under
anionic or cationic forms but not under amphoteric form.
It can be suggested that different types of mechanisms occur
depending the ionic form of Phe. At acid pH NH3+ are reduced by e−
as proposed by Horikoshi et al. [18] whereas decarboxylation will
be the main reaction at basic pH. Actually, decarboxylation reaction
(photo-kolbe reaction) is favoured in the presence of carboxylate
group which will be less important at acid pH. This hypothesis is in
agreement with the work of Möning et al. [19].
R–COO− + h+ → R◦ + CO2
Hydroxylation reactions should also be considered to explain the
higher efficiency at basic pH.
+
h
OH−−→ OH◦
At acid pH the number of OH◦ should be lower due to the presence
of hydronium ions favouring the presence of hole.
H2O + h+ ꢀ OH◦ + H+
Considering these two types of reaction, we can explain the
higher efficiency observed at basic pH. Actually, carboxylate
group are present and the number of OH◦ will be more
important.
Fig. 6. Amount of phenylalanine adsorbed per gram of catalyst as a function of Phe
concentration remaining in solution at the equilibrium.

L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
147
Fig. 8. Evolution of rate constant k and adsorption constant K as a function of radiant flux.
The hypothesis of a modification of the number of OH◦ radi-
cals depending of the pH is in agreement with the quantification
of hydroxylated compounds. Actually a more important concentra-
tion of hydroxylated compounds is formed at basic pH compared
to this one detected at acid pH. However, it is at neutral pH that the
more important amount of hydroxylated compounds is detected.
This result could be explained considering repulsion between OH−
and the surface of TiO2 which is negatively charged. However, car-
boxylate group should also repulse from TiO2 surface leading to less
decarboxylation reaction. More studies should be done to under-
stand better this behaviour.
An other important point to underline is the absence of correla-
tion between adsorption in the dark and disappearance. This result
is in agreement with a modification of adsorption under UV light.
Fig. 9. Quantum yield as a function of initial concentration of phenylalanine.
3.5. Influence of radiant flux
of phenylalanine initially degraded divided per the number of
photons absorbed during the same time. In our experimental con-
ditions, the concentration of TiO2 used allows to absorb all photons
emitted.
The effect of radiant flux has been studied by considering dif-
ferent concentrations of phenylalanine. The initial disappearance
rate of Phe as a function of remaining concentration in solution
(Fig. 7(a)) and a function of radiant flux (Fig. 7(b)).
r
S
=
× ꢂ
˚
NAE
Whatever the initial concentration of amino acid,
a
With the radiant flux ˚ in W cm−2 (J s−1 cm−2), the radiant flux
energy E in J photons−1, the Avogadro’s number (NA), the surface of
optical window (S) and the quantum yield (ꢂ).
Langmuir–Hinshelwood curve can modelize the kinetic observed.
At low Phe concentration the curve is linear and tends towards a
plateau. The proportionality occurs when the Phe concentration is
varied with the radiant flux.
This phenomenon can be explained by considering that the
number of active species (e−, h+) formed increases as a function of
radiant flux, but these species recombined themselves if no organic
molecule is present to react.
From Langmuir–Hinshelwood curves, the value of rate constant
k and adsorption constant K can be determined. These both values
are represented as a function of radiant flux in Fig. 8. The rate con-
stant proportionally increases with radiant flux while the opposite
effect is observed on adsorption constant.
The proportionality observed between k and ˚ means that the
number of active species linearly increases with radiant flux, in
other word that in our experimental conditions, photocatalytic
regime occurs. This result can also be seen in Fig. 8(b).
The decrease of adsorption constant K by increasing radiant flux
means that the rate of adsorption (Kads) decreases or that desorp-
tion rate increases (Kdes) under UV (K = Kads/Kdes).
The evolution of the quantum yield as a function of initial con-
centration at different radiant flux is represented on Fig. 9. The
maximal quantum yield obtained is 1.6%, it means that if the reac-
tion is performed under a higher initial concentration or radiant
flux, in 1 s, the maximum of molecules degraded by 100 photons
emitted will be 1.6. This result is very important to the industrial-
isation of the process because the percentage of degradation can
be estimated. From this figure it can be seen that the quantum
yield increases as a function of Phe concentration before reach-
ing a plateau indicating that recombination of (e−, h+) pairs is more
important.
4. Conclusions
In spite of
a multilayer adsorption of phenylalanine in
the dark, the effect of initial concentration of phenylalanine
on its photocatalytic degradation seems to be modelized by
Langmuir–Hinshelwood model. However, the rate constant found
by using this model is too important compared to the experimental
value. This behaviour shows that this mathematical model must be
used with caution and that kinetic of adsorption in the dark has
to be considered, although it can be suggested a modification of
adsorption site under UV.
K
ads
Phe + TiO2−→ Pheads
with rads = kads [Phe] [TiO2]
Kdes
Pheads−→ Phe + TiO2
with rdes = kdes [Pheads
This result is in agreement with a modification of adsorption
under UV underlined in Section 3.2.
From Fig. 8(b), it is interesting to notice that above 180 ␮mol/L,
the slope r/˚ remains constant. This slope is directly proportional to
quantum yield (ꢂ) which corresponds to the number of molecules
]
The study of mineralization of carbon and heteroatom show
that a total mineralization seems to be obtained after 4 h and that
ammonium ions are the major inorganic ions formed. Six aromatic
products were identified by LC–MS techniques corresponding to
mono-, di-hydroxylated compounds and to deamination. Five lin-
ear organic acids are also identified by LC–UV analysis.

148
L. Elsellami et al. / Applied Catalysis A: General 380 (2010) 142–148
Whatever the pH, a multilayer adsorption is observed. In the
[2] F. Dossier Berne, B. Panais, N. Merlet, B. Cauchi, B. Legube, Environ. Technol. 15
(1994) 901–916.
dark, adsorption is more important in acid solution, while basic
solution favour the degradation under UV. This result suggests that
the degradation mechanism could differ depending on pH. Actually,
first the presence of carboxylate group at basic pH should favour the
decarboxylation by photo-Kolbe reaction, and secondly the number
of OH◦ should be lower at acid pH due to the presence of hydrogen
ions favouring the presence of hole. This last hypothesis is in a good
agreement with the effect of pH on the formation of hydroxylated
compounds.
[3] K.M. Agbekodo, J.P. Croué, S. Dard, B. Legube, Rev. Sci. Eau 4 (1996) 535–
555.
[4] L. Hureiki, J.-P. Croué, B. Legube, Water Res. 28 (1994) 2521–2531.
[5] C. Le Cloirec, G. Martin, R.L. Jolley, W.A. Brungs, J.D. Johnson, R.B. Cumming
(Eds.), Ann Arbor Science Publishers Inc., MI, 5, 1984, p. 821.
[6] I. Freuze, S. Brosillon, A. Laplanche, D. Tozza, J. Cavard, Water Res. 39 (2008)
2636–2642.
[7] A.K. Benabbou, Z. Derriche, C. Felix, P. Lejeune, C. Guillard, Appl. Catal. B: Envi-
ron. 76 (2007) 257–263.
[8] P. Fernandez, J. Blanco, C. Sichel, S. Malato, Catal. Today 101 (2005) 345–
352.
[9] S. Parra, S. Malato, C. Pulgarin, Appl. Catal. B: Environ. 36 (2002) 131–
144.
[10] K.P. Khün, I.F. Chaberny, K. Massholder, M. Stickler, V.W. Benz, H.-G. Sonntag,
L. Erdinger, Chemosphere 53 (2003) 71–77.
[11] M. Kaneko, I. Okura (Eds.), Photocatalysis: Science and Technology, Springer,
2002.
[12] M. El Madani, C. Guillard, N. Perol, J.-M. Chovelon, M. El Azzouzi, A. Zrineh, J.-M.
Herrmann, Appl. Catal. B: Environ. 6 (2006) 70–76.
[13] H.P. Boehm, M. Hermann, Z. Anorg. Allg. Chem. 352 (1967) 156.
The study of the effect of radiant flux allowed us to determine
the maximal quantum yield which can be reached, 1.6% for this
amino acid. This value is very important for dimensioning the reac-
tor. Our results also underlined that at low coverage of TiO2 surface,
the quantum yield is lower, showing the important part of recom-
bination of active species.
[14] T.-H. Tran, A.Y. Nosaka, Y. Nasaka, J. Phys. Chem.
25531.
B 110 (2006) 25525–
Acknowledgement
[15] B. Ohtani, Chem. Lett. 37 (2008) 217–229.
This work was supported by the France–Tunisia cooperation
CMCU program.
[16] T.-H. Bui, M. Karkmaz, E. Puzenat, C. Guillard, J.-M. Herrmann, Res. Chem.
Intermed. 33 (2007) 421–431.
[17] H. Hidaka, S. Horikoshi, K. Ajisaka, J. Zhao, N. Serpone, J. Photochem. Photobiol.
A: Chem. 108 (1997) 197–205.
References
[18] S. Horikoshi, N. Serpone, J. Zhao, H. Hidaka, J. Photochem. Photobiol A: Chem.
118 (1998) 123–129.
[19] J. Möning, R. Chapman, K.D. Asmus, J. Phys. Chem. 89 (1985) 3139–3144.
[1] D. Kasiske, D. Klinkmuller K, M. Sonneborn, J. Chromatogr. A 149 (1978)
703–710.