Applied Catalysis B: Environmental 60 (2005) 181–190

www.elsevier.com/locate/apcatb

Heterogeneous photo-Fenton degradation of phenolic aqueous

solutions over iron-containing SBA-15 catalyst
F. Martı́nez, G. Calleja, J.A. Melero *, R. Molina
Department of Chemical, Environmental and Materials Technology, ESCET, Rey Juan Carlos University,
‘C/Tulipán s/n, 28933 Móstoles, Madrid, Spain

Received 12 January 2005; received in revised form 10 March 2005; accepted 21 March 2005
Available online 16 April 2005

Abstract

A novel iron-containing mesostructured material has been successfully tested for the heterogeneous photo-Fenton degradation of phenolic
aqueous solutions using near UV–vis irradiation (higher than 313 nm) at room temperature and close to neutral pH. This catalyst is a
composite material that contains crystalline hematite particles embedded into the mesostructured SBA-15 matrix in a wide distribution of size
(30–300 nm) and well dispersed ionic iron species within the siliceous framework. The outstanding physico-chemical properties make this
material a promising photocatalyst leading to better activity than other unsupported iron oxides. An experimental design model has been
applied to assign the weight of catalyst and hydrogen peroxide concentrations in the photo-Fenton processes over this particular material. The
catalytic performance has been monitored in terms of aromatics and total organic carbon (TOC) conversions, whereas the catalyst stability
was evaluated according to the metal leached into the aqueous solution. Hydrogen peroxide concentration plays an important role in the
stability of the iron species, preventing their leaching out into the solution, in contrast to the effect shown in typical dark-Fenton reaction. The
homogeneous leached iron species result in very little contribution to the overall photocatalytic process. Catalyst loadings of 0.5 g/L and
concentration of hydrogen peroxide close to the stoichiometric amount have yielded a total abatement of phenol and a remarkable organic
mineralization.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Photo-Fenton; Heterogeneous catalysts; SBA-15; Experimental design and phenol

1. Introduction The oxidation system based on the Fenton’s reagent

(hydrogen peroxide in the presence of ferric ions) has been
During the last decade the abatement of aromatic used as a powerful source of oxidative radicals [3]:
compounds in wastewaters has been an important challenge Fe3þ þ H2 O2 $ FeðOOHÞ2þ þ Hþ (1)
due to their bio-toxicity and refractory behavior for the
microorganisms as well as their hazard carcinogenic effects FeðOOHÞ2þ ! Fe2þ þ HO2
(2)
for human being. The application of advanced oxidation Fe 2þ
þ H2 O2 ! Fe 3þ
þ HO 
þ HO

(3)
technologies (AOTs) like the one presented in this study
have emerged as an important alternative for the efficient Fe3þ þ HO2
! Fe2þ þ Hþ þ O2 (4)
removal of organic pollutants, being nowadays widely More recently, Fenton processes has been shown to be
described in the literature [1,2]. These novel destruction enhanced by light due to the decomposition of the photo-
techniques are based on the generation of hydroxyl radicals, active Fe(OH)2+ species, promoting an additional generation
which are able to mineralize refractory chemicals present in of OH
radicals in solution [4]:
effluent water.
FeðOHÞ2þ þ hn ! Fe2þ þ HO
(5)
* Corresponding author. Tel.: +34 91 4887087; fax: +34 91 4887068. However, it has been reported that Fenton homogeneously
E-mail address: juan.melero@urjc.es (J.A. Melero). catalysed reactions needs concentrations of ca. 50–80 ppm

0926-3373/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2005.03.004
182 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190

of iron ions in solution, which is quite above the 2 ppm by some authors [12–15] has been substituted by UV near
permitted by the European Community Directives [5]. visible irradiation (>313 nm) in this contribution, which
Therefore, in order to remove the iron ions from solution, represents a significant advantage for future application in
precipitation and re-dissolution techniques are necessary solar photodegradation processes. Additionally, hydrogen
with the additional operational costs. This reason has peroxide to iron ratio has been proven to play an important
promoted the development of dark- and photo-Fenton role in homogeneous photo-Fenton processes [16] but few
processes based on heterogeneous catalytic systems, details are reported for heterogeneous photo-Fenton
which provide an easy separation and recovery of the systems. In the present work, multivariate analysis has
catalyst from the treated wastewater. Unlike homogeneous been used to assess the conditions (oxidant and catalyst
Fenton reaction, where a strict control of pH around 2–3 is concentrations) that yield the best result in terms of organic
required for a good catalytic performance, immobilized degradation and catalyst stability.
Fenton catalytic systems provide the possibility of work-
ing in a wider pH range. The crucial point of heteroge-
neous catalytic systems is linked to the resistance of 2. Experimental
immobilized iron species to be leached out into the solu-
tion under the typically acidic and strong oxidizing con- 2.1. Preparation of catalyst
ditions in which Fenton reactions occur. In this way,
research efforts are currently addressed to design new Iron-containing SBA-15 mesostructured material has
photocatalysts with high stability and allowing an efficient been prepared by co-condensation of iron (FeCl36H2O;
use of the hydrogen peroxide due to its relative high cost. Aldrich) and silica (tetraethoxysilicate, TEOS; Aldrich)
Thus, one of the aims in this research field should be to sources under acidic conditions and templated with Pluronic
prepare stable materials that require the lowest oxidant to 123 as described elsewhere [17]. The precipitation of
catalyst ratio for obtaining good degradation rates. crystalline iron oxide particles was promoted by ageing of
In the literature, different organic and inorganic materials the resultant solution at 110 8C for 24 h under static
have been already reported as supports for the immobiliza- conditions and pH 3.5. After ageing step, the solid product
tion of active iron species in heterogeneous photo-Fenton was recovered by filtration and air dried at room temperature
processes. Nafion, as films or pellets, has been employed as overnight. The template was removed by calcination in air at
organic support due to the presence of sulphonic groups 550 8C for 5 h.
allowing the effective anchoring of Fe ions [6–8]. However, Commercial hematite with a BET surface area of 60 m2/g
Nafion materials show a low photocatalytic activity and are was purchased from Aldrich with the purpose of compar-
not very attractive due to their relative high cost. More ison.
recently, a novel Fe/C structured fabric has been success-
fully used for the complete discoloration of Orange II, but a 2.2. Characterization
high hydrogen peroxide to substrate ratio was needed [9].
Among inorganic supports, some type of zeolites and X-ray powder diffraction data were acquired on a
pillared clays have been also used. Iron(III) doped-Y zeolitic PHILIPS X-PERT diffractometer using Cu Ka radiation.
material, prepared by ion exchange techniques, was used for The data were collected from 2u ranging from 0.58 to 908
degradation of poly(vinyl alcohol) [10]. However, a low with a resolution of 0.028. Nitrogen adsorption and
catalytic activity was achieved in terms of the remaining desorption isotherms at 77 K were measured using a
levels of dissolved organic carbon after 2 h of treatment. Micromeritics Tristar 3000 equipment. The data were
These Fe exchanged zeolites have been also tested in the analysed using the BJH model and the pore volume (Vp) was
photo-assisted degradation of phenol with a low initial taken at P/P0 = 0.975 single point. Transmission electron
phenol concentration of 100 ppm and a stoichiometric microscopy (TEM) microphotographs and energy dispersive
amount of hydrogen peroxide [11]. Nanosized Fe2O3 X-ray (EDX) microanalysis were carried out on a PHILIPS
particle oxides intercalated as pillars between layered clays TECNAI-20 electron microscope operating at 200 kV. Bulk
such as laponite and bentonite have revealed a high iron content of the prepared sample was obtained by atomic
photocatalytic activity for discoloration and mineralization emission spectroscopy with induced coupled plasma (ICP–
of different dyes, but using high energetic 254 nm UV AES) analysis collected in a Varian Vista AX system.
irradiation [12–14].
The purpose of this study is the assessment of a novel 2.3. Photo-Fenton reactions
composite Fe-containing SBA-15 mesostructured material
for the heterogeneous photo-Fenton degradation of phenolic Photo-Fenton catalytic runs were carried out in a Pyrex
aqueous solutions. Power irradiation light and hydrogen batch cylindrical reactor containing 1 L of the aqueous
peroxide consumption are the main parameters to be dispersion of the catalyst and the pollutant. The reactor was
considered in photo-Fenton processes due to their econom- provided on the top with three ports for inflow and outflow of
ical importance. High energetic 254 nm UV radiation used gases and withdrawal of aliquots for their analysis. The
 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190 183

irradiation was performed with a 150 W medium pressure

mercury lamp (Heraeus TQ-150) inside a quartz jacket,
immersed in the mixture and coaxial with the photoreactor.
The lamp was surrounded by a cooling tube in which a
copper sulphate aqueous solution was circulating to prevent
overheating of the suspension and cutting off radiation with
a wavelength shorter than 313 nm.
Prior to photo-Fenton experiments, the corresponding
amount of catalyst was dispersed in 980 mL of deionised
water and introduced in the reactor. Simultaneously, the
lamp was immersed in the reactor and switched on during
15 min before adding the phenol and hydrogen peroxide
reagents, in order to avoid the inherent induction time of the
irradiation source. Thereafter, a concentrated phenol
solution with the required amount of hydrogen peroxide
was added to the reactor up to obtain 1 L of aqueous
solution. The temperature was maintained at 25 8C during
all the reaction time. The initial concentrations were
500 ppm for phenol and 4100–800 ppm for hydrogen
peroxide. The stoichiometric oxidant concentration required
for the total oxidation of phenol is 2550 ppm, according to
the following reaction:
C6 H6 O þ 14H2 O2 ! 6CO2 þ 17H2 O (6)

Aliquots of 8 mL were taken at selected reaction times

and filtered by means of 0.22 mm nylon membranes before
being analysed. The phenol conversion was monitored with
a HPLC chromatograph Varian Prostar apparatus equipped
with a Waters Spherisorb column (250 mm  4.6 mm) and
an UV detector at 215 nm. The analysis of other by-products
coming from the incomplete mineralization of phenol was
also performed by this technique. Total organic carbon
(TOC) content of the solutions before and after reaction was
measured with a combustion/non dispersive infrared gas
analyser model TOC-V Shimadzu. Aromatic removal (XA)
in percentage of carbon was calculated as follows:
initial TOCðppm of CÞ  ppm of
C from aromatics measured by HPLC
XA ð%Þ ¼ 100
initial TOCðppm of CÞ

Hydrogen peroxide concentration after reaction was eval- Fig. 1. Characterization results of iron-containing SBA-15 material: (a)
XRD pattern at low angle range; (b) XRD pattern at high angle range; (c)
uated by iodiometric titration. Iron content in the filtered
nitrogen adsorption/desorption isotherm (pore size distribution in top left);
solution after reaction was measured by ICP–AES analysis (d) TEM microphotograph and iron content (wt.%) measured by EDX
collected in a Varian VISTA AX system. microanalysis.

These crystalline entities show the typical pattern of

3. Results and discussion crystalline hematite (see Fig. 1b, top right). Fig. 1c illustrates
the nitrogen adsorption/desorption isotherm of the calcined
3.1. Catalyst properties material as well as the pore size distribution in the mesopore
range. The adsorption results indicate a type IV isotherm
Iron-containing SBA-15 material was first characterized with a H1 hysteresis loop typical of mesoporous materials
by XRD. Low angle XRD spectrum depicted in Fig. 1 [19]. The calculated BET area for the calcined material is
clearly shows (1 0 0), (1 1 0) and (2 0 0) reflections typical around 468 m2/g, which is slightly lower than that usually
of hexagonally mesostructured SBA-15 materials [18]. found for pure silica SBA-15 materials [18]; this might be
Additionally, the presence of crystalline iron oxides particles attributed to the presence of supported hematite entities.
has been detected by high angle XRD analysis (Fig. 1b). Finally, a narrow pore size distribution (see Fig. 1c, top left)
184 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190

confirms the presence of uniform pores of the same size and conditions (curve 2) leads to just 5% of TOC degradation,
the high mesoscopic ordering previously inferred from low which clearly evidences the important role of the hydroxyl
angle XRD spectrum. radicals coming from photocatalytic decomposition of
TEM images also confirm the presence of both hematite hydrogen peroxide. Similar TOC reductions are observed
entities (a broad distribution particle size among 30–300 nm in the experiment carried out in absence of UV light and
has been established in a previous study [20]) and well- presence of catalyst and hydrogen peroxide (2450 ppm;
ordered channels with 2D hexagonal symmetry typical of curve 3). This fact demonstrates the restricted production of
SBA-15 materials. Microanalysis measurements show zones hydroxyl radicals by catalytic dark-Fenton reactions at these
with a high concentration of Fe, which are related with the particular conditions. Note that UV light in absence of
micro-aggregates of hematite and low contents for the catalyst and presence of hydrogen peroxide (2450 ppm;
mesostructured zone (Fig. 1d). The bulk iron content of the curve 4) yielded a relevant enhancement of TOC conversion,
sample measured by ICP–AES was 16 wt.%. Moreover, which reveals that hydroxyl radical production by photolytic
analysis of local iron environment in this material by means decomposition of hydrogen peroxide under UV–vis irradia-
of Mössbauer experiments has also confirmed the presence tion higher than 313 nm cannot be ruled out (reaction (7)):
of hematite particles embedded into the mesostructured H2 O2 þ hn ! 2HO
(7)
matrix and iron present in ionic dispersion in the
mesostructured siliceous framework [20]. Finally, the photocatalytic activity of Fe–SBA-15 material
has been compared to that obtained with a commercial
3.2. Preliminary experiments hematite iron oxide under analogous reaction conditions
(curve 5). A low TOC conversion is shown by the commer-
The efficiency of iron-containing SBA-15 catalyst in the cial iron oxide, which indicates beyond doubt the advantages
photo-assisted degradation of phenolic aqueous solutions of supporting small iron oxides particles over the surface of
was preliminarily evaluated under different reaction condi- hexagonal SBA-15 mesostructured materials. However, the
tions combining the presence of hydrogen peroxide and UV role of ionic Fe species in Fe–SBA-15 catalyst could also
light. Fig. 2 shows the evolution of TOC conversion within account for the striking enhancement of TOC removal over
the reaction time for these experiments. The photocatalytic this material.
activity of Fe–SBA-15 material in presence of a close
stoichiometric hydrogen peroxide concentration (2450 ppm; 3.3. Factorial design of experiments for photo-Fenton
curve 1) shows a significant TOC reduction achieving a heterogeneous system
conversion value of ca. 50% after 4 h of reaction. The
absence of hydrogen peroxide in analogous reaction Preliminary experiments have clearly evidenced that
immobilized iron species over mesostructured SBA-15
materials present an acceptable catalytic performance for the
photo-assisted degradation of phenolic aqueous solutions
through Fenton processes. In this second part of the
manuscript the influence of significant variables on the
reaction has been studied by means of an experimental
design methodology. The study was mainly focused on the
influence of catalyst and hydrogen peroxide concentrations
on the heterogeneous photo-Fenton degradation of phenol
over this novel iron-containing material.
A factorial experimental design based on 32 [21] was
carried out considering low and high levels for Fe–SBA-15
catalyst loadings (0.5–1.5 g/L) and hydrogen peroxide
concentration (800–4100 ppm). The choice of this particular
range of catalyst concentration was based on the literature
data, where similar ranges are generally considered
[12,13,22]. Optimal concentration of 1 g/L has been in
particular reported for photo-Fenton processes catalysed by
iron-containing clays [13]. However, other authors like
Fig. 2. Evolution of TOC conversion along the time under different reaction Cuzzola et al. [23] have also tested higher catalyst loadings,
conditions. [TOC]o = 380 ppm; [catalyst] = 0.5 g/L; initial pH = 5.5; up to 30 g/L, for the treatment of linear alkylbenzene
T = 25 8C. Photocatalytic activity of Fe–SBA-15 catalyst in presence (1) sulphonic acids with Fe(III) supported over g-Al2O3 or SiO2
and absence (2) of hydrogen peroxide; dark-Fenton reaction over Fe–SBA-
15 catalyst (3); photolysis in presence of hydrogen peroxide (4); photo- under solar irradiation. With respect to the hydrogen
catalytic activity of commercial crystalline hematite in presence of hydro- peroxide concentration, the range was selected according
gen peroxide (5). to other studies described in literature, taking into account
 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190 185

Table 1
Numerical data and results of the factorial design of phenol degradation in photo-Fenton reactions
Exp. Codified values Y = TOC conversion (%)
X1 [catalyst] (g/L) X2 [H2O2]o (ppm) Y1 (30 min) Y2 (75 min) Y3 (150 min) Y4 (240 min)
Exp. Calc. Exp. Calc. Exp. Calc. Exp. Calc.
1 0.5 (1) 800 (1) 23.3 25.7 28.7 31.4 27.4 28.5 29.7 30.0
2 0.5 (1) 4100 (+1) 34.4 36.8 42.2 44.9 49.7 50.8 59.1 59.4
3 1.5 (+1) 800 (1) 8.4 10.8 12.1 14.8 23.1 20.1 29.0 23.3
4 1.5 (+1) 4100 (+1) 46.7 49.1 48.9 51.6 55.2 56.3 82.4 82.8
5 1 (0) 2450 (0) 13.1 15.2 16.7 18.4 43.3 42.6 56.1 56.5
6 1 (0) 2450 (0) 12.4 15.2 12.6 18.4 40.3 42.6 55.3 56.5
7 1 (0) 2450 (0) 10.5 15.2 15.3 18.4 39.7 42.6 56.9 56.5
8 0.5 (1) 2450 (0) 35.9 31.1 39.7 34.3 44.5 42.1 48.1 47.5
9 1.5 (+1) 2450 (0) 40.0 35.2 45.1 39.7 52.6 50.4 61.5 60.9
10 1 (0) 800 (1) 11.0 6.2 15.7 10.3 35.6 33.4 50.0 49.4
11 1 (0) 4100 (+1) 23.9 19.1 29.3 24.0 39.3 37.1 53.6 53.0

that the iron to oxidant ratio is an important variable giving Y2 ð%TOC removalÞ
to different mechanisms [24]. Moreover, it should be noted
¼ 18:41ð4:1Þ þ 2:72ð4:9ÞX1 þ 6:79ð4:3ÞX2
that the oxidant concentration is also strongly dependent on
the organic content to be degraded. In this sense, hydrogen þ5:81ð2:82ÞX1 X2 þ 18:57ð2:8ÞX12  1:31ð2:06ÞX22
peroxide dose has been set in accordance with the initial  5:2ð2:31ÞX1 X22 þ 5:79ð2:19ÞX2 X12 (9)
phenol concentration, being also within the typical iron to
oxidant ratio for homogeneous Fenton reactions (1 part of Y3 ð% TOC removalÞ
iron per 5–25 parts of hydrogen peroxide (w/w)) [16,25].
¼ 42:56ð1:8Þ þ 4:16ð2:18ÞX1 þ 1:84ð2:95ÞX2
Finally, the efficient use of hydrogen peroxide has also
attracted the attention of others in order to minimize the þ3:46ð1:26ÞX1 X2 þ 3:67ð1:27ÞX12  7:31ð0:89ÞX22
addition of the oxidant [26].  4:87ð1:02ÞX1 X22 þ 12:79ð0:97ÞX2 X12 (10)
Table 1 describes the factorial design of the experiments,
including the codified and real experimental values selected Y4 ð% TOC removalÞ
in this work. Low and high levels are denoted by 1, +1,
¼ 56:48ð0:57Þ þ 6:7ð0:68ÞX1 þ 1:8ð0:61ÞX2
respectively, whereas the central point as 0. Central point
experiments were repeated three times in order to check the þ 7:51ð0:39ÞX1 X2  2:3ð0:4ÞX12  5:3ð0:28ÞX22
reproducibility and obtain the standard deviation of the  2:54ð0:32ÞX1 X22 þ 20:40ð0:30ÞX2 X12 (11)
experimental response. The percentage of TOC removal at
different reaction times (minutes), 30 (Y1), 75 (Y2), 150 (Y3) Concerning the equation coefficients shown above, it is clear
and 240 (Y4), was chosen as the response factor. The that quadratic terms are dominant, resulting in a significant
experimental values of TOC conversion as well as the curvature of the response surface of TOC removal. The
calculated values obtained from the model are also depicted catalyst concentration ðX12 Þ is important for initial reaction
in Table 1. From the matrix generated by the experimental times, whereas the synergic effect of the catalyst concentra-
TOC data and assuming a second order polynomial model, tion ðX12 Þ and the hydrogen peroxide (X2) becomes dominant
Eqs. (8)–(11) were obtained for different reaction times for longer reaction times (150 and 240 min) in detrimental to
using a Levenberg–Marquad algorithm for nonlinear the effect of every individual variable.
regression, where X1 and X2 represent the catalyst and 3D-graphical representation of Eqs. (8)–(11) are shown
hydrogen peroxide concentrations, respectively. The mag- in Fig. 3. 3D-response surface at 30 min (Fig. 3a) clearly
nitude of the interaction between the variables is related with shows a decrease of TOC conversion with catalyst loading
the value of the coefficients of the polynomial expression when the initial hydrogen peroxide concentration is small.
used to fit the experimental data. Values in parenthesis This detrimental effect of the catalyst concentration can be
describe the relative error associated to each coefficient. attributed to the presence of a significant amount of solid
particles, which would reduce the transmission of near UV–
vis light and produce scattering of the incident light. A
Y1 ð%TOC removalÞ similar effect has been also reported in photocatalytic
¼ 15:22ð3:6Þ þ 2:04ð4:33ÞX1 þ 6:42ð3:37ÞX2 processes with titanium oxide [1]. However, a significant
þ 6:78ð2:5ÞX1 X2 þ 17:92ð2:5ÞX12  2:56ð1:76ÞX22 increase of TOC conversion is obtained using the highest
catalyst and oxidant concentrations. Both variables have led
 2:68ð2:04ÞX1 X22 þ 5:92ð1:93ÞX2 X12 (8) to an enhancement of the catalytic performance for
186 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190

Fig. 3. 3D response surfaces of TOC conversion for (a) 30, (b) 75, (c) 150 and (d) 240 min. [TOC]o = 380 ppm; initial pH = 5.5; T = 25 8C.

heterogeneous like-Fenton reactions in dark conditions [27]. 80% were attained with 1.5 g/L of catalyst and 4100 ppm
This fact may suggest the presence of a possible coupling of of H2O2.
dark-Fenton and photo-Fenton processes as responsible for A possible explanation of these results, as early
the overall mineralization of organic matter when high mentioned, might be the contribution of dark-Fenton
concentrations of oxidant and catalyst are present in the reactions under strong oxidant conditions. In order to
reaction medium. A similar behavior is observed after corroborate this hypothesis two additional catalytic experi-
75 min of reaction (Fig. 3b). For longer reaction times (150 ments were carried out with high and low concentration
and 240 min; Fig. 3c and d respectively), the increase of levels of both oxidant and catalyst in the absence of UV
catalyst loadings at low hydrogen peroxide concentration radiation. Thus, for oxidant and catalyst concentrations of
exhibits a convex-like 3D surface response curves of TOC 4100 ppm and 1.5 g/L, without UV radiation, 22% of TOC
conversion, in contrast to the concave shapes shown for removal was achieved, whereas hardly 3% of TOC removal
lower reaction times. was obtained for the lowest values of 800 ppm and 0.5 g/L.
The dramatic decrease of TOC conversion evidenced The decrease of TOC concentration by adsorption of the
after 240 min of reaction for low oxidant concentration organic compounds on the catalyst surface has been up to
and high catalyst concentration (Fig. 3d) is due to the 8% for catalyst loadings of 1.5 g/L. These results clearly
absence of hydrogen peroxide as promoter of hydroxyl reveal a significant contribution of dark-Fenton reactions to
radicals—almost complete conversion of hydrogen per- the overall catalytic degradation when both high catalyst
oxide was detected after 75 min of reaction. However, a loadings and strong oxidant conditions are present.
larger amount of hydrogen peroxide with high catalyst Conversely, for very low catalyst loadings (0.5 g/L), TOC
loadings (1.5 g/L) leads to an outstanding enhancement of removal is not significantly enhanced with hydrogen
TOC conversion, which is observed even at initial reaction peroxide concentration. This detrimental effect of high
times. Thus, after 240 min, TOC conversions higher than hydrogen peroxide could be explained in terms of reaction
 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190 187

Fig. 4. Aromatic removal and oxidant conversion for different initial hydrogen peroxide concentrations at two catalyst loadings. [TOC]o = 380 ppm; initial
pH = 5.5; T = 25 8C.

(12). An excess of oxidant can favour its reaction with material is attained (240 min; 800 ppm of H2O2; 1.5 g/L
hydroxyl radicals (responsible for the direct oxidation of the catalyst).
organic matter) precluding the extent of mineralization. This
negative effect has been also reported for iron-based 3.4. Induced-photo iron leaching
homogeneous Fenton and photo-Fenton systems [16]:
Leaching experiments studies of iron species from Fe–
H2 O2 þ OH
! HO2
þ H2 O (12) SBA-15 catalyst during photo-Fenton degradation of phenol
Fig. 4 shows the degradation of aromatic compounds, i.e. were performed in order to ascertain the strength of the iron
phenol, cathecol and hydroquinone (calculated as described species supported into SBA-15 mesoporous matrix and the
in the Section 2) and the hydrogen peroxide conversion for possible contribution of dissolved iron ions in the
the three levels of oxidant concentration at both minimum homogeneous catalytic reactions. Fig. 5a illustrates the iron
and maximum catalyst loadings (0.5 and 1.5 g/L, respec- concentration dissolved into the aqueous solution after
tively). Aromatic degradation values above 70% are 240 min of reaction for the set of experiments carried out in
achieved for both high and low catalyst concentration after our factorial design. Iron concentrations ranging from ca. 2.5
75 min of reaction regardless hydrogen peroxide concentra- to 11.5 ppm are detected into the aqueous solution
tion. On the other hand, just a slight increase in aromatic depending on the studied reaction conditions. The resistance
removal is observed as the catalyst concentration increases. of iron species to be leached out seems to be related with the
Values of aromatic removal close to 100% are attained after oxidant to catalyst ratio. In this sense, several hypothesis can
240 min of reaction independent of the oxidant and catalyst be suggested to figure out the relationship between the
concentration. experimental conditions and the dissolution of iron species
Hydrogen peroxide conversion increases with the from the catalyst.
reaction time as expected, being this effect more evident In the set of reactions carried out at different hydrogen
at higher catalyst concentrations. Plots in Fig. 4 also show a peroxide loadings for a given catalyst concentration, iron
proportional decrease of H2O2 conversion with the increas- leaching shows a general decreasing trend as oxidant
ing of initial concentration of oxidant for each catalyst concentration is increased, a trend that is more accentuated
concentration. It is also observed that for the lowest initial for higher catalyst concentrations. This behavior of lower
oxidant concentration and the highest catalyst loading, a metal leaching for stronger oxidant conditions appears to be
complete consumption of hydrogen peroxide is produced, different to that shown by heterogeneous catalytic Fenton
conditions at which complete removal of the aromatic like-reactions under dark conditions in which the more
188 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190

sites again. From this mechanism, the general decrease of

iron leaching for high oxidant concentrations could be
attributed to the shielding effect of surrounding metal–
H2O2 complex species which prevents photo-induced leach-
ing phenomenon by UV–vis irradiation.
Other researches have also proposed the reasonable
hypothesis of leaching of FeII from FeIII containing catalysts
as a result of photochemical reduction of FeIII species
[29,30] being accumulated in the aqueous phase unless it
becomes reoxidized by an excess of hydrogen peroxide. This
hypothesis is also in agreement with our results.
Looking at the stability of catalyst in terms of percentage
of iron dissolved with respect to the initial content present in
the fresh catalyst (Fig. 5b), a significant increase of iron
leaching degree is observed for catalyst loadings of 1 g/L in
comparison with the lower and higher contents for the three
levels of oxidant concentration. These results indicate a
complex behavior of the stability of catalyst in photo-Fenton
systems, being strongly dependent on the reaction condi-
tions. As it has been mentioned, UV irradiation and the role
of hydrogen peroxide in the shielding effect by iron–H2O2
complex or in the reoxidation of dissolved Fe(II) ions into
the solution can be considered as main arguments to justify
the stability of the catalyst.
Therefore, in the case of a catalyst loading of 0.5 g/L, in
which a low catalyst to oxidant ratio is present for the three
levels of oxidant concentration, the effect of hydrogen
peroxide excess seems to be the main reason to explain the
lower iron leaching degrees shown in comparison to set of
experiments carried out at 1 g/L. On the other hand, as
irradiation might be another important factor of photo-
induced leaching, the scattering UV–vis light phenomenon
mediated by greater catalyst concentrations may account for
the lower iron leaching degrees observed for 1.5 g/L catalyst
Fig. 5. Stability of Fe–SBA-15 catalyst in terms of (a) iron concentration
detected into the aqueous solution and (b) percentage of iron leached out loading as compared to that found for 1 g/L. It must be also
from the catalysts, after 240 min of reaction. [TOC]o = 380 ppm; initial noteworthy that the influence of pH in the final solution
pH = 5.5; T = 25 8C. could be another important point for the explanation of the
stability of catalyst which will be subject in future studies.
significant metal leaching phenomenon was observed for Finally, it must be also noted the progressive increase of
high hydrogen peroxide concentrations [27]. In this sense, it iron leaching with reaction time for all the tested reactions,
has been reported that catalytic decomposition of hydrogen although in different extension depending on the reaction
peroxide over iron oxides is based on the formation of conditions (data for particular reaction conditions are shown
peroxide complex species with Fe(III) active sites on the in Fig. 6). This behavior is very different to that found by Feng
catalyst surface following a series of reactions in which this et al. [13] with iron laponite RD clays as catalysts in which
metal surface complex undergoes different electronically maximum concentrations of iron ions were detected after
excited states until dissociation in peroxide radical [28]. The 45 min of reaction followed by a continuous decrease up to
schematic path of reactions would be the following: 90 min. Although a reasonable explanation has not been
BBFeIII OH þ H2 O2 , ðH2 O2 ÞS (13) reported for these results yet, it can be deduced that the reason
of these different behaviors in iron leaching for photo-Fenton
ðH2 O2 ÞS , ðBBFeII 
O2 HÞ þ H2 O (14) systems should be related to the specific nature of the catalyst,
probably involving different mechanisms.
ðBBFeII 
O2 HÞ , BBFeII þ HO2
(15)

BBFeII þ H2 O2 ! BBFeIII OH þ HO
(16) 3.5. Influence of homogeneous photo-Fenton reactions

According to this reaction scheme, reduced iron sites can In order to ascertain the influence of the iron species
react with hydrogen peroxide to regenerate oxidized iron leached out into the aqueous solution, additional experi-
 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190 189

iron species coming from the heterogeneous catalyst is much

lower for initial reaction times, where much higher TOC
removal differences are produced. These results evidence
that Fenton homogeneous contribution might occur in
parallel with immobilized Fenton process but the remark-
able contribution of iron-containing SBA-15 catalyst to the
overall photocatalytic process is also clearly demonstrated,
especially at initial reaction times where TOC degradation
rate is higher.
Finally, an experiment of phenol adsorption was carried
out using 1.5 g/L of catalyst concentration in absence of
oxidant with the purpose of checking the TOC reduction
attributed to the adsorption phenomena. The results shown
in Fig. 6 clearly indicate the negligible contribution of
organic adsorption in the pollutant degradation.

4. Conclusions

Iron-containing SBA-15 catalyst consisting of crystalline

hematite particle oxides supported onto a mesostructured
silica matrix has been shown as a promising catalyst for the
treatment of phenolic solutions through photo-Fenton
processes. The outstanding physico-chemical properties
make this material more attractive than unsupported
commercial hematite iron oxide catalyst, leading to a better
overall photocatalytic performance. The experimental
Fig. 6. Contribution of homogeneous iron species vs. heterogeneous cat- design model carried out for different levels of catalyst
alytic systems in photo-Fenton processes. [TOC]o = 380 ppm; and hydrogen peroxide concentrations has demonstrated that
[H2O2]o = 2450 ppm; initial pH = 5.5; T = 25 8C. the increase of catalyst loadings from 0.5 to 1.5 g/L does not
promote a significant enhancement of TOC removal except
for strong oxidant conditions. The stability of the catalyst
ments were carried out using iron(III) chloride as metal towards the leaching of iron species is strongly dependent on
source for homogeneous testing reactions. The iron the oxidant to catalyst ratio. For the highest catalyst and
concentration determined in aqueous solution after hydrogen peroxide concentrations (1.5 g/L and 4100 ppm,
240 min for the heterogeneous system was found to range respectively), the amount of iron species detected into the
from 2.5 to 11.5 ppm (data corresponding to Fig. 5a). In aqueous solution does not exceed 8 ppm. This low leaching
particular, heterogeneous experiments carried out with a is accompanied with a complete removal of aromatic
hydrogen peroxide concentration of 2450 ppm and catalyst compounds and a TOC conversion of ca. 80% after 240 min
loadings of 0.5 and 1.5 g/L have led to dissolution of iron of reaction, which is a remarkable result. Finally, the
species up to ca. 2.7 and 9.7 ppm, respectively, after 240 min contribution of homogeneous photo-Fenton reactions from
of reaction. With the purpose of comparison, iron the iron species leached out from the catalyst has been
concentrations of 3 and 11 ppm were chosen for the proven to be poorly relevant.
homogeneous reaction tests, keeping 2450 ppm as initial
hydrogen peroxide concentration. The evolution of TOC and
iron concentration in aqueous solution for the homogeneous Acknowledgements
and heterogeneous runs are shown in Fig. 6.
Increasing iron concentration in the homogeneous runs The authors thank ‘‘Spanish Science and Technology
reveals a slight enhancement of TOC removal up to reaction Ministry’’ for the financial support of this research through
times of 150 min. However, a remarkable increase of TOC the project PPQ-2003-03984.
conversion is observed for the homogeneous run with
11 ppm iron concentration at 240 min. Interestingly, the
evolution of TOC degradation for homogenous systems is References
significantly lower than that shown by the heterogeneous
ones, especially for the initial reaction times. Moreover, it [1] P.R. Gogate, A.B. Pandit, Adv. Environ. Res. 8 (2004) 501–552.
should be pointed out that the concentration of dissolved [2] P.R. Gogate, A.B. Pandit, Adv. Environ. Res. 8 (2004) 553–597.
190 F. Martı́nez et al. / Applied Catalysis B: Environmental 60 (2005) 181–190

[3] C. Walling, A. Goosen, J. Am. Chem. Soc. 95 (1973) 2987–2991. [18] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka,
[4] V. Nadtochenco, J. Kiwi, Inorg. Chem. 37 (1998) 5233–5238. G.D. Stucky, Science 279 (1998) 548–552.
[5] EEC List of Council Directives 76/4647, European Economic Com- [19] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscow, R.A. Pierotti, J.
munity, Brussels, Belgium, 1982. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
[6] J. Fernandez, J. Bandara, A. Lopez, P. Albers, J. Kiwi, Chem. [20] K. Lazar, G. Calleja, J.A. Melero, F. Martinez, R. Molina, in:
Commun. (1998) 1493–1494. Proceedings of the 14th International Zeolite Conference, Cape Town,
[7] S. Sabhi, J. Kiwi, Water Res. 35 (2001) 1994–2002. South Africa, 2004, pp. 805–812.
[8] J. Fernandez, J. Bandara, A. Lopez, P. Buffat, J. Kiwi, Langmuir 15 [21] G.E.P. Box, W.G. Hunter, J.S. Hunter, Statistics for Experiments, An
(1999) 185–192. Introduction to Design, Data Analysis and Model Building, Wiley,
[9] T. Yuranova, O. Enea, E. Mielczarski, J. Mielczarski, P. Albers, J. New York, 1978.
Kiwi, Appl. Catal. B: Environ. 49 (2004) 39–50. [22] Z.W. Zheng, L.C. Lei, S.J. Xu, P.L. Cen, J. Zhejiang, Univ. Sci. 5
[10] S.H. Bossmann, E. Oliveros, S. Göb, M. Kantor, A. Göppert, L. Lei, (2004) 206–211.
P.L. Yue, A.M. Braun, Water Sci. Technol. 44 (2001) 257–262. [23] A. Cuzzola, M. Bernini, P. Salvadori, Appl. Catal. B: Environ. 52
[11] H.E. Feng, S. Xue-you, L. Le-Cheng, J. Environ. Sci. 15 (2003) 351– (2004) 117–122.
355. [24] E. Neyens, J. Baeyens, J. Hazard. Mater. 98 (2003) 33–50.
[12] J. Feng, X. Hu, P.L. Yue, H.Y. Zhu, G.Q. Lu, Water Res. 37 (2003) [25] F. Torrades, M. Perez, H.D. Mansilla, J. Peral, Chemosphere 53 (2003)
3776–3784. 1211–1220.
[13] J. Feng, X. Hu, P.L. Yue, Environ. Sci. Technol. 38 (2004) 269– [26] F.J. Rivas, F.J. Beltrán, O. Gimeno, P. Alvarez, J. Hazard. Mater. 96
275. (2003) 277–290.
[14] P.L. Yue, J.Y. Feng, X. Hu, Water Sci. Technol. 49 (2004) 85–90. [27] J.L. Sotelo, G. Ovejero, F. Martinez, J.A. Melero, A. Milieni, Appl.
[15] A. Bozzi, T. Yuranova, E. Mielczarski, J. Mielczarski, P.A. Buffat, P. Catal. B: Environ. 47 (2004) 281–294.
Lais, J. Kiwi, Appl. Catal. B: Environ. 42 (2003) 289–303. [28] S.S. Lin, M.D. Gurol, Environ. Sci. Technol. 32 (1998) 1417–1423.
[16] M. Pera-Titus, V. Garcia-Molina, M.A. Banos, J. Gimenez, S. Esplu- [29] M. Rios-Enriquez, N. Shahin, C. Duran-de-Bazua, J. Lang, E. Oli-
gas, Appl. Catal. B: Environ. 123 (2003) 1236–1245. veros, S.H. Bossmann, A.M. Braun, Solar Energy 77 (2004) 491–501.
[17] F. Martı́nez, Y. Jhan, G. Stucky, J.L. Sotelo, G. Ovejero, J.A. Melero, [30] A. Bozzi, T. Yuranova, J. Mielczarski, J. Kiwi, New J. Chem. 28
Stud. Surf. Sci. Catal. 142 (2002) 1109–1116. (2004) 519–526.