Selective oxidation of phenol and benzoic acid in water via home-prepared TiO2 photocatalysts: Distribution of hydroxylation products

Article abstract of DOI:10.1016/j.apcata.2012.07.019

The hydroxylation of phenol (a substrate containing an electron donor group) and of benzoic acid (a substrate containing an electron withdrawing group) has been carried out by the photocatalytic method in aqueous suspensions containing commercial or home prepared TiO₂ samples. The aim of the work was to study the distribution of hydroxylation products when different photocatalysts were used and to correlate the selectivity to some physico-chemical features of the powders. The samples were characterized by X-ray diffraction, thermogravimetry, determination of crystalline phase percentage, specific surface area and zero charge point. The photoreactivity results indicate that the products of the primary oxidation of phenol are the ortho- and para-mono-hydroxy derivatives while those of benzoic acid are all the mono-hydroxy derivatives independently of the catalyst. The selectivity toward mono-hydroxy derivatives shows a strong dependence on catalyst hydroxylation and crystallinity degrees: the highest selectivity values were obtained by using the commercial samples that resulted the least hydroxylated and the most crystalline ones. A kinetic model, taking into account the mineralization and the partial oxidation reaction routes, is proposed by using the Langmuir-Hinshelwood model.

Full text of DOI:10.1016/j.apcata.2012.07.019

Applied Catalysis A: General 441–442 (2012) 79–89
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Selective oxidation of phenol and benzoic acid in water via home-prepared TiO2
photocatalysts: Distribution of hydroxylation products
Marianna Bellarditaa,∗, Vincenzo Augugliaroa, Vittorio Loddoa, Bartolomeo Megnab,
Giovanni Palmisanoa, Leonardo Palmisanoa,∗∗, Maria Angela Pumaa
a “Schiavello-Grillone” Photocatalysis Group, Dipartimento di Ingegneria Elettrica, Elettronica e delle Telecomunicazioni, di tecnologie Chimiche, Automatica e modelli Matematici
(DIEETCAM), University of Palermo, Viale delle Scienze, 90128 Palermo, Italy
b Dipartimento di Ingegneria Civile, Ambientale, Aerospaziale, dei Materiali, University of Palermo, Viale delle Scienze, 90128 Palermo, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 June 2012
Received in revised form 9 July 2012
Accepted 11 July 2012
Available online 23 July 2012
The hydroxylation of phenol (a substrate containing an electron donor group) and of benzoic acid (a
substrate containing an electron withdrawing group) has been carried out by the photocatalytic method in
aqueous suspensions containing commercial or home prepared TiO2 samples. The aim of the work was to
study the distribution of hydroxylation products when different photocatalysts were used and to correlate
the selectivity to some physico-chemical features of the powders. The samples were characterized by X-
ray diffraction, thermogravimetry, determination of crystalline phase percentage, specific surface area
and zero charge point. The photoreactivity results indicate that the products of the primary oxidation of
phenol are the ortho- and para-mono-hydroxy derivatives while those of benzoic acid are all the mono-
hydroxy derivatives independently of the catalyst. The selectivity toward mono-hydroxy derivatives
shows a strong dependence on catalyst hydroxylation and crystallinity degrees: the highest selectivity
values were obtained by using the commercial samples that resulted the least hydroxylated and the
most crystalline ones. A kinetic model, taking into account the mineralization and the partial oxidation
reaction routes, is proposed by using the Langmuir–Hinshelwood model.
Keywords:
Selective hydroxylation
Product distribution
Benzoic acid
Phenol
Heterogeneous photocatalysis
Reaction pathways
Kinetics
TiO2
© 2012 Published by Elsevier B.V.
1. Introduction
Several papers have been published dealing with the pho-
tocatalytic degradation of benzene derivatives such as phenol
Photocatalysis by TiO2 has been mainly applied to degrade
organic and inorganic pollutants both in vapor and in liquid phases
[1–3]. Its main advantage consists not only in the mild experimental
conditions under which the process is carried out, but also in the
possibility to abate refractory, very toxic and non biodegradable
molecules. TiO2 has been elected as the most reliable photocat-
alyst due to its low cost, high activity and (photo)stability under
irradiation [2].
Investigation on photocatalytic oxidation of aromatic com-
pounds has been focused on the interactions of these species with
the catalyst surface, the mechanism of photoreaction and the reac-
tion kinetics [4–19]. The affinity of the aromatic substrates to the
TiO2 surface determines the occurrence of adsorption and photoad-
sorption in a more or lesser extent.
[14,20–25] and benzoic acid [14,26–28,13] and the reaction mech-
anism in the presence of TiO2 dispersion is also well known
[4,7,20,27–32]. Only few papers deal with selectivity toward the
hydroxylated species [33,34].
In a study of the photocatalytic degradation of phenol and ben-
zoic acid, Vione et al. [14] report indirect evidence of considerably
different interactions of these two species with the TiO2 sur-
face. The principal oxidation route of phenol degradation involve
hydroxyl radicals adsorbed on TiO2 surface, whereas for benzoic
acid the main process is the adsorption of substrate on the sur-
face of the catalyst and the degradation of the substrate takes place
through electron transfer process involving surface species. This
finding suggests that the interaction of benzoic acid with the TiO2
surface is stronger than that of phenol as the former behaves as
a bidentate species. When the surface of the TiO2 was modified
by replacing the hydroxyl groups with fluoride ions, the adsorp-
tion phenomena resulted different and the degradation process
occurred in the solution bulk.
∗
Corresponding author. Tel.: +39 091 23863722; fax: +39 091 23863722.
Corresponding author. Tel.: +39 091 23863746; fax: +39 091 23863722.
E-mail addresses: marianna.bellardita@unipa.it (M. Bellardita),
By considering the intermediates produced during the first
oxidation steps, mechanistic studies suggest that hydroxyl rad-
icals are the species involved in oxidation processes carried
∗∗
leonardo.palmisano@unipa.it (L. Palmisano).
0926-860X/$ – see front matter © 2012 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.apcata.2012.07.019

80
M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
out in the presence of O2 [4–8]. In particular, phenol oxidation
was studied in the presence of anatase TiO2 and hydroquinone,
catechol and 1,2,4-trihydroxybenzene were found as the main
intermediates [4,5,7]. In order to explain the formation of mono-
hydroxy derivatives substituted only in ortho (catechol) and para
(hydroquinone) positions, Peiró et al. [7] proposed that phenol
molecule reacts with a hydroxyl radical, giving rise to an adduct
that transforms into a phenoxy radical possessing resonant struc-
tures only when the attack occurs in ortho- and para-positions.
These resonance structures account for the achievement of the
two different mono-hydroxy derivatives and other more complex
intermediates deriving from reaction between two phenoxy rad-
icals. Hydroquinone, formed by either phenol or 4-chlorophenol
hydroxylation, participates to a keto-enol equilibrium by forming
p-benzoquinone, as previously described [8].
Generally many mechanistic studies have been performed by
investigating the degradation of only one aromatic compound
while only a few investigations have been aimed to determine
whether the nature of the substituent group in the aromatic ring
influences the production of the hydroxylated intermediates. A
study on selectivity of photocatalytic hydroxylation of aromatic
species indicates that the electronic nature of the substituent
groups is linked to the formation of one or more monohydroxylated
species [35]. The study was carried out by using only commercial
anatase TiO2 sample and the general rule that the meta-substituted
hydroxy-species is never formed in the presence of an electron
donor group (EDG) was found. This feature has been also confirmed
by in a lot of research papers where many types of TiO2 photo-
catalysts were used while no papers have been published where
meta-substituted hydroxy-species were reported to be formed in
the presence of an EDG.
It is indeed well known that the oxidation of aromatic species
occurs through two main pathways, i.e. direct mineralization
(through a series of adsorbed intermediates which do not des-
orb into the bulk of solution) and partial oxidation giving rise
to different aromatic, aliphatic and ionic (both organic and inor-
ganic) species [36,37]. In this paper TiO2 photocatalysts with
different bulk and superficial features have been used for per-
forming the oxidation of two aromatic compounds, i.e. phenol
(PH) bearing an EDG and benzoic acid (BA) bearing an electron
withdrawing group (EWG). The investigation has been aimed to
find a correlation between the photocatalytic selectivity toward
the hydroxylated compounds and some physico-chemical prop-
erty of the different powders. Different kinds of home-prepared
anatase, rutile and brookite TiO2 samples have been used along
with Merck and Degussa P25 commercial samples and a SiO2 sup-
ported sample for the sake of comparison. A kinetic study by using
the Langmuir–Hinshelwood model has been also performed and
the model parameters have been determined.
resulting solid was filtered and washed until the washing water had
a neutral pH. Then it was calcined at 673 K for 3 h and the sample
is named HP298-673.
1 5BS373 [40]: 10 mL of TiCl4 were slowly added to 50 mL of
distilled water at room temperature. After ca. 15 min of continuous
stirring, a clear transparent solution was obtained. This was heated
in a closed bottle and aged at 373 K in an oven for 48 h. The bottle
was allowed to cool to room temperature, and the resultant solid
was recovered using a rotary evaporator at 323 K. The sample is
identified with the notation 1 5BS373.
Brookite [41]: 15 mL of TiCl4 were added dropwise to a solution
containing 630 mL of demineralized water and 240 mL of concen-
trated hydrochloric acid. The solution obtained after continuous
stirring was heated in a closed bottle and aged at 373 K in an oven
for 48 h. The resultant precipitate contained a mixture of brookite
and rutile. Pure brookite nanoparticles were separated by peptiza-
tion by removing many times the supernatant and adding water to
restore the initial solution volume. After few washes, a dispersion
of brookite particles was obtained whilst the rutile phase remained
as a precipitate. The dispersions containing the brookite particles
were collected and dried in a rotary evaporator at 673 K.
S5 [42]: A commercial SiO2 (Cabot) was used as support. 10 g of
silica were added to 110 mL of the TiCl4 solution and the obtained
suspension was boiled for 2 h. The final suspension was dried in a
rotary evaporator. The code used for this sample is S5.
2.2. Preparation of the home prepared samples starting from
TiOSO4
Titanium (IV) oxysulfate (Sigma–Aldrich) was used as the pre-
cursor. 20 g of TiOSO4 were added to 90 mL of distilled water. After
ca. 2 h of continuous stirring at room temperature, a clear solution
was obtained. This solution was heated in a closed bottle and aged
at 373 K in an oven for 48 h. The resultant precipitate was washed
by withdrawing many times the supernatant liquid and by adding
pure water to restore the initial solution volume and to eliminate
most of the sulfate ions. The resultant solid was recovered using a
rotary evaporator at 323 K and calcined at 873 K for 10 h. The code
used for this sample is HP873S.
2.3. Catalysts characterization and reactivity set up
X-ray diffractometry (XRD) patterns of the powders were
recorded by a Philips diffractometer (operating at a voltage of 40 kV
and a current of 30 mA) using the Cu K␣ radiation and a 2ꢀ scan rate
of 1.28◦ min−1. The crystalline sizes of the samples were deter-
mined by using the Scherrer equation [43]. The specific surface
areas (SSA) of the powders were determined in a FlowSorb 2300
apparatus (Micromeritics) by using the single-point BET method.
The samples were degassed for 0.5 h at 523 K prior to the measure-
ment by using a N2/He mixture 30/70 (v/v).
2. Experimental
The crystallinity of all the samples was evaluated following the
procedure reported by Jensen et al. [44]. XRD diffractograms were
recorded for mixtures of TiO2 and CaF2 (50%, w/w) and the areas of
the 100% peaks of anatase (1 0 1), rutile (1 1 0) and CaF2 (2 2 0) were
determined. By comparing the ratio between the areas of (1 0 1) and
(2 2 0) peaks or of (1 1 0) and (2 2 0) peaks to the ratios obtained by
using the pure phases (1.25 for anatase and 0.90 for rutile), the
amount of crystalline and amorphous phases present in the sam-
ples was determined. No reference values are reported for brookite
TiO2 crystalline phase.
The zero charge point (ZCP) of the various catalyst samples was
calculated by using the method of mass titration [45–47] which
involves finding the limiting pH value of an oxide/water slurry as
the oxide mass content is increased. Varying amounts of powders
were added to water and the resulting pH values were measured
2.1. Samples preparation starting from TiCl4
Titanium tetrachloride (Fluka 98%) was used as the precursor
without any further purification. It was slowly added to distilled
water (molar ratio Ti/H2O 1:60; volume ratio 1:10) at room tem-
perature. After ca. 12 h of continuous stirring at room temperature,
a clear solution was obtained.
HP0.5 [38]: The clear solution was boiled for 0.5 h under agita-
tion. This treatment produced a milky white TiO2 dispersion; the
solid was dried in a rotary evaporator at 323 K. The code used for
this sample is HP0.5.
HP298-673 [39]: 360 mL of 1 M NaOH aqueous solution were
added to clear TiCl4 solution to induce the precipitation of TiO2. The

M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
81
after 24 h of equilibration. For Degussa P25 sample the determi-
(h)
(g)
nation of the ZCP value was repeated three times by preventively
washing the sample under UV-irradiation in order to clean the sur-
face from all impurities.
Simultaneous thermal gravimetric (TG/DTG) and differential
thermal analysis (DTA) measurements were performed on cata-
lysts by using a PerkinElmer STA 6000 system, in the 30–750 ◦C
range in a nitrogen flux of ca. 20 mL min−1. The temperature pro-
gram consists of three steps: temperature scan from 30 to 120 ◦C at
10 ◦C min−1, 15 min in isothermal condition at 120 ◦C, temperature
B
B
(f)
(e)
scan from 120 to 750 at 10 ◦C min−1
.
A Pyrex batch photoreactor of cylindrical shape, containing 0.5 L
of aqueous suspension, was used for performing the reactivity
experiments. The photoreactor was provided with ports in its upper
section for the inlet and outlet of gases and for sampling. A magnetic
stirrer guaranteed a satisfactory suspension of the photocatalyst
and the uniformity of the reacting mixture. A 125 W medium pres-
sure Hg lamp (Helios Italquartz) was axially immersed within the
photoreactor and it was cooled by water circulating through a Pyrex
thimble; the temperature of the suspension was about 300 K. The
radiation energy impinging on the suspension had an average value
of 10 mW cm−2; it was measured by using a radiometer UVX Digital,
at ꢁ = 360 nm.
(d)
(c)
R
(b)
(a)
A
A
A
A
Different amounts of the catalysts (ranging from 0.2 to 0.8 g L−1
)
were used for the reactivity runs; they were determined in order
to get the same photon flux absorbed by the suspension for each
catalyst. That amount was determined by slowly adding the powder
to the suspension and by measuring the photon flux transmitted by
the suspension; when the transmitted radiation was less than 10%
with respect to the incident one, catalyst addition was stopped.
The starting concentrations of PH and BA ranged from 0.1 to 5 mM.
The initial pH of the suspensions was ca. 6 for PH and ca. 3 for BA.
When necessary, pH was adjusted by adding some drops of NaOH
1 M solution. The suspensions were saturated by bubbling O2 at
atmospheric pressure for 0.5 h in the dark and throughout all the
runs.
Samples for analyses were withdrawn at fixed intervals of time;
the catalyst was immediately separated from the aqueous solution
by filtering through 0.2 ␮m Millex Millipore filters. For each sam-
ple a particular care was devoted to determine the concentrations
of hydroxylation products (ortho, meta and para). The substrates
and the intermediates produced during the reactions were ana-
lyzed with a HPLC Beckman Coulter (System Gold 126 Solvent
Module and 168 Diode Array Detector), equipped with a Phe-
nomenex Synergi 4 ␮ Hydro-RP column (250 mm long × 4.60 mm
i.d.). The HPLC eluent consisted of 45% (v/v) 1 mM aqueous solution
of trifluoroacetic acid and 55% (v/v) methanol. The flow rate was
0.4 mL min−1 and the identification was carried out by comparison
with authentical standard samples.
20
30
40
50
60
2θ
Fig. 1. X-Ray diffractograms of the used samples: (a) Merck; (b) P25; (c) HP0.5; (d)
HP298-673; (e) 1 5BS373; (f) Brookite; (g) HP873S; (h) S5.
with the references where the detailed preparation methods can be
found. The various peaks corresponding to the three main phases
of TiO2 have been identified. HP0.5 and S5 are the least crystallized
samples with the smallest crystallite size. Both catalysts show the
presence of anatase as the main phase; they were not subjected to
a strong thermal treatment, but were only warmed at 100 ◦C for
0.5 or 2 h. Experimental evidence of the virtual absence of anatase
and rutile in brookite sample has been confirmed by Raman spec-
troscopy [41]. It can be noticed that the specific surface areas of the
home prepared samples are higher than those of the commercial
samples, whereas the crystallite sizes are smaller. The SSA of Merck
is the lowest among all the samples, indicating a strong particle
agglomeration.
The commercial samples showed an amorphous phase percent-
age of 10% (P25) and 26% (Merck). Among the home prepared
samples, the highest crystallinity values were found for calcined
samples HP298-673 and HP873S, indicating that the crystallinity
increases by increasing the temperature of the thermal treatment.
All the other samples revealed a low percentage of crystalline
phase. The percentage of crystalline phase in the brookite sample
was not determined because no reference values are reported in lit-
erature and it is not a trivial task to prepare a pure brookite sample
with a high degree of crystallinity (hopefully near to 100%) since
calcination treatments easily transform brookite into rutile.
Zero charge points of catalysts ranged from 0 to 4.5, the highest
figures being those of commercial samples; it must be reported
that the ZPC value of the used P25 sample resulted lower than
that reported in the literature [47]. The difference found can be
probably ascribed to differences between P25 batches that are also
characterized by different ratios between anatase and rutile.
According to literature [48] the weight losses of hydroxylated
TiO2 samples in the 30–120 ◦C range are due to the physically
adsorbed water and they mainly depend on residual humidity after
the preparation of the powders. The losses in the 120–300 ◦C and
Total organic carbon (TOC) analyses were carried out by using
a 5000A Shimadzu TOC analyzer. These analyses allowed to deter-
mine the amount of organic carbon mineralized to CO2 in the course
of the reaction.
The quantitative determination of oxalate, acetate and formate
ions was performed with a Dionex DX 120 ionic chromatograph,
using a Dionex Ion Pac AS14 column (250 mm long × 4 mm i.d.).
The eluent was an aqueous solution of NaHCO3 (1 mM) and Na2CO3
(8 mM) at a flow rate of 1 mL min−1
.
3. Results and discussion
3.1. Catalysts characterization
Fig. 1 reports the X-ray diffractograms of all the samples used
in this work; Table 1 summarizes the main features of the catalysts

82
M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
Table 1
Some morphological, structural and surface features of the used catalysts.
c
Photocatalyst
Phases
SSAa (m2 g−1
)
Crystallinityb (%)
Crystallite size (nm)
Tmax,p (K)
ZCPd
Referencee
Merck
P25
HP0.5
HP298-673
1 5BS373
Brookite
S5
A
10.0
50.0
235
35.0
87.0
82.0
177
74
60.0 (A)
–
–
4.5
4.0
0.30
3.7
0.45
0.70
0
–
–
A, R
A, R
A
R
B
72 (A) 18 (R)
6.8 (A)–(R)
26
12
25.0 (A) 33.0 (R)
4.60 (A) 2.10 (R)
8.80
3.50
6.60
373
673
373
373
373
873
[38]
[39]
[40]
[41]
[42]
–
–
A, R, S
A
6 (A)–(R)
40
12.7 (A) 2.80 (R)
23.6
HP873S
44.0
1.5
A: anatase; R: rutile; B: brookite; S: silica.
a
BET specific surface area.
Percentage of crystallinity.
Highest temperature to which the photocatalysts were heated during their preparation.
Zero charge point.
References reporting details of the preparation method.
b
c
d
e
300–600 ◦C ranges can be related to weakly bonded OH groups and
strongly bonded OH groups, respectively. In this work the ther-
mal analysis of P25 sample has been taken as reference for the
other catalysts. The P25 graph, reported in Fig. 2(a), shows that
two clear steps in weight loss are present in relation to the differ-
ent OH groups: the first one at ca. 300 ◦C and the second one at ca.
600 ◦C. Table 2 reports the OH groups weight losses together with
the humidity loss for all the catalysts; the values of global losses of
OH groups are also reported. It can be seen that the two commer-
cial samples are the least hydroxylated ones with values ranging
between 0.1 and 0.78% (w/w), whereas the home-prepared ones are
the most hydroxylated ones with the highest value equal to 10.76%
(w/w) for HP0.5. It can be noticed that S5 sample is characterized
by a high amount of humidity, probably due to the high porosity of
silica used as the support. Home-prepared samples calcined at high
temperature have a low OH content, which decreases by increas-
ing the calcination temperature and HP873S shows values close to
those found for P25.
3.2. Photoreactivity
Before performing the photocatalytic runs, the extent of adsorp-
tion of the substrate was measured after 0.5 h mixing, keeping
the reactor under dark conditions. The differences between the
initial concentration value (0.5 mM for both substrates) and that
measured after adsorption were negligible in the presence of
commercial TiO2 samples, whereas in the presence of HP873S
and HP298-673 samples, adsorption was negligible only for PH.
Conversely, adsorption of both PH and BA in the presence of home-
prepared samples was significant, ranging from 2% to 21%. The
highest values were recorded with BA in the presence of HP0.5
(12%), 1 5BS373 (19%) and HP873S (21%). These data indicate that
the adsorption is neither straightforwardly related to the SSA and
ZCP values nor to the hydroxyl groups content and that PH has a
week affinity with the catalysts surface, although all of the above
parameters can in principle influence the adsorption phenomena.
It is known [35] that the additional electron density, induced in
the aromatic ring by the presence of OH substituent, is not ben-
eficial for the adsorption of the molecule, while the presence of
COOH, by reducing the electron density of the aromatic ring and
by imparting the property of a bidentate species to BA, favors the
adsorption of this molecule onto the catalyst surface. In our opinion
the above effects should be taken into account rather than the slight
100.2
(a)
100.0
99.8
Y = 0.7531 %
99.6
99.4
99.2
Y = 0.4170 %
99.0
98.8
Y = 0.3643 %
98.6
98.4
98.2
98.0
30
100
200
300
400
500
600
700 750
Temperature (°C)
difference between the 1-octanol-water partition coefficients (Kow
)
100
98
96
94
92
90
88
86
84
82
80
reported by Gumy et al. [49] for PH (1.46) and BA (1.50). Notably a
higher Kow value should suggest a stronger affinity of the molecule
to the TiO2 particles.
The photoreactivity runs indicated that PH and BA were
oxidized in all cases and the concentrations of their oxidation
products increased with irradiation time. Two parallel routes
(b)
Y = 7.9878 %
Y = 6.6565 %
Table 2
OH groups weight percentage (%, w/w) determined by TGA.
Sample
Humidity
OHweak
OHstrong
OHtotal
Merck
P25
HP873S
HP298-673
S5
Brookite
1 5BS373
HP0.5
0.3000
0.7500
1.300
3.440
50.45
4.720
5.430
7.990
0.060
0.420
0.370
0.890
2.90
2.01
2.40
6.66
0.0400
0.360
0.350
0.920
0.940
2.29
0.1000
0.7800
0.7200
1.810
3.840
4.300
Y = 4.0972 %
25.72 100
200
300
400
500
600
700 747
Temperature (°C)
2.22
4.10
4.620
10.76
Fig. 2. Weight percentage as a function of temperature for samples (a) P25 and (b)
HP05.

M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
83
Table 3
Photoreactivity in the presence of PH and BA.
Photocatalyst
Merck 0.8 g L−1
cPH (mM)a
tX=15% (min)b
MinX=15% (%)c
Mint=90 (%)d
cBA (mM)a
tX=15% (min)b
MinX=15% (%)c
Mint=90 (%)d
0.10
0.25
0.50
1.0
3.0
5.0
7.40
13.0
22.0
38.2
81.5
120
1.80
1.60
Negligible
Negligible
1.00
25.0
6.60
Negligible
1.00
0.500
0.500
0.10
0.25
0.50
1.0
9.20
17.1
25.3
64.6
2.00
1.20
3.20
1.00
23.5
13.5
9.00
1.00
0.500
5.0
144
2.50
0.200
P25 0.2 g L−1
0.10
0.25
0.50
1.0
5.00
7.30
11.0
17.6
133
6.70
6.20
6.00
2.10
3.80
92.0
82.0
47.6
9.60
2.60
0.10
0.25
0.50
1.0
4.50
5.20
10.5
15.0
108
8.30
4.80
3.70
3.30
7.60
100
83.7
50.0
20.6
5.0
5.0
6.50
HP0.5 0.4 g L−1
0.10
0.25
0.50
1.0
30.0
57.5
95.0
145
7.00
5.00
12.0
10.0
12.5
14.0
13.0
11.0
7.00
2.00
0.10
0.25
0.50
1.0
10.0
12.7
15.0
70.0
300
8.60
6.50
4.50
9.80
8.60
77.4
44.5
23.0
14.0
5.0
780
5.0
3.10
HP298-673
0.6 g L−1
0.50
1.0
60.0
78.3
7.60
4.70
31.0
24.0
0.50
16.4
4.90
27.0
1 5BS373 0.6 g L−1
Brookite 0.6 g L−1
S5 0.8 g L−1
0.50
0.50
0.50
0.50
65.0
41.7
60.0
15.0
8.50
5.00
7.80
4.50
19.0
13.0
8.30
26.0
0.50
0.50
0.50
0.50
9.50
12.8
38.0
10.5
5.90
4.60
12.2
4.20
55.0
26.0
29.0
37.6
HP873S 0.6 g L−1
a
Substrate initial concentration.
Time needed to reach 15% conversion.
Mineralization at 15% conversion.
Mineralization at t = 90 min.
b
c
d
[35] were operative from the starting of irradiation: the first one
giving rise to the direct mineralization of the substrate through
adsorbed intermediates that do not desorb into the bulk of solution
and the second one leading to partial oxidation to aromatic and
aliphatic molecular and ionic species. Times required to reach
15% conversion (tX=15%) along with mineralization percentages for
the same conversion (MinX=15%) and for reaction time of 90 min
(Mint=90) are reported in Table 3.
The tX=15% parameter refers to the disappearance rate of the
substrate, regardless of what the produced species are; conversely
the MinX=15% and Mint=90 ones have been calculated by consider-
ing the TOC decrease in the reacting solutions. Provided that no
organic volatile compounds were formed during the reaction, TOC
measurement amounts to the formed CO2. This assumption has
been supported by gas-chromatographic analyses carried out with
a flame ionization (FID) detector, that showed the absence of any
peaks.
The PH photodegradation rate was higher than that of BA in the
presence of Merck, whereas it was lower in the presence of all the
other catalysts (see the longer and the shorter times, respectively,
in the tX=15% column). This insight can be correlated to the poor
surface hydroxylation [50] and to the low SSA of Merck.
Merck, Degussa P25 and HP873S are the most oxidizing catalysts
with the shortest times needed to reach 15% conversion. Accord-
ingly all these catalysts are well crystallized as shown by Fig. 1 and
contain anatase as the main phase.
The highest degree of mineralization is observed in the presence
of P25 for PH and of 1 5BS373 for BA (by taking the value of 0.5 mM
for the substrates starting concentration and 90 min as reaction
time). The other home prepared samples showed a mineralization
power between P25 and Merck. TiO2 Merck indeed gave rise to a
low extent of mineralization both in the presence of PH and BA.
Apart from CO2, other species found during PH oxidation were
the monohydroxylated derivatives, formed by insertion of an
OH group into the aromatic ring, the p-benzoquinone, the dihy-
droxylated species 1,2,4-trihydroxybenzene and traces of some
polyaromatic species (2,2ꢀ-biphenol, 4,4ꢀ-dihydroxybiphenol) in
accord to previous studies [7]. Aliphatic compounds deriving from
breakage of aromatic ring (mucic, malonic and oxalic acid) were
detected during the oxidation of both PH and BA in the present
study.
3.3. Selectivity
Selectivity determination was aimed to analyze the regios-
electivity corresponding to the formation of ortho-, meta- and
para-monohydroxy derivatives upon oxidation. During phenol oxi-
dation, hydroquinone, that is the para-monohydroxy derivative,
gives rise to p-benzoquinone since it participates to a keto-enol
equilibrium as previously reported [8]. Some photocatalytic runs
(not reported for the sake of brevity) were carried out by using
hydroquinone or p-benzoquinone as the starting substrate: in both
cases the two species were found in the irradiated solutions. In
the present investigation we consider the para-monohydroxy phe-
nol derivative as the sum of hydroquinone and p-benzoquinone.
Notably the latter compound was present in relevant amount in
the runs carried out by using home prepared catalysts whereas
only traces were detected in the presence of Merck and P25.
The low quantity of p-benzoquinone detected in the presence of
commercial catalysts could be explained by the high degradation
rate shown by these catalysts. This hypothesis was corroborated
by photocatalytic runs carried out starting from p-benzoquinone
in the presence of P25 and HP0.5 samples: from the beginning
of irradiation an aliquot of the substrate was degraded while
the remaining one was transformed to hydroquinone and 2-
hydroxy-benzoquinone. The degradation rate of p-benzoquinone
and hydroquinone by P25 was two times higher than that by HP0.5.
Moreover other oxidation derivatives came from PH, as previously
highlighted.
Table 4 shows the selectivity values at 15% conversion and the
o–m–p distribution of the monohydroxylated compounds. In par-
ticular, selectivity ranged from 1.9% to 74%. Phenol oxidation did
not give rise to attack of hydroxyl radical in meta-position. This
behavior, independent of the used catalyst, has been reported for all

84
M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
Table 4
Selectivity to mono-hydroxylated derivatives in the presence of PH and BA.
Photocatalyst
Merck 0.8 g L−1
cPH (mM)a
SX=15% (%)b
o–m–pc
cBA (mM)a
SX=15% (%)b
o–m–pc
0.1
0.25
0.5
1
3
5
65
64
74
59.8
72
55–0–45
54–0–46
56–0–44
55–0–45
56–0–44
50–0–50
0.1
0.25
0.5
1
43
46
39
57
41–31–28
30–36–34
37–35–28
40–34–26
74
5
34
35–36–28
P25 0.2 g L−1
0.1
0.25
0.5
1
50
48
50
39.7
65
22–0–78
27–0–73
43–0–57
42–0–58
42–0–58
0.1
0.25
0.5
1
21
24
27.4
27.5
33.4
21–43–36
14–47–39
22–43–35
23–45–32
23–51–26
5
5
HP0.5 0.4 g L−1
0.1
0.25
0.5
1
30
25
26.8
14.5
12.7
0–0–100
0–0–100
0–0–100
6–0–94
0.1
0.25
0.5
1
3.2
1.9
2.32
4.5
8.9
0–42–58
0–39–61
0–41–59
0–40–60
14–33–53
5
11–0–89
5
HP298–673
0.6 g L−1
0.5
1
27
19.4
0–0–100
5–0–95
0.5
22
17–43–40
1 5BS373 0.6 g L−1
Brookite 0.6 g L−1
S5 0.8 g L−1
0.5
0.5
0.5
0.5
23.2
25.4
35.6
54
8–0–92
6–0–94
2–0–98
3–0–97
0.5
0.5
0.5
0.5
3.2
9.1
9.8
12–60–38
15–40–45
12–42–46
19–44–37
HP873S 0.6 g L−1
33.3
a
Substrate initial concentration.
Selectivity at 15% conversion.
o–m–p distribution of monohydroxylated oxidation compounds at 15% conversion.
b
c
substituted benzenes bearing an electron donor group (EDG), such
as PH [15,35]. In the case of electron withdrawing groups (EWGs),
instead, all the three regio-isomers are obtained, as confirmed dur-
ing the oxidation of BA regardless of the catalysts used.
On this ground TiO2 Merck, that is the least hydroxylated sam-
ple, and HP0.5, that is the most hydroxylated one, are the most and
the least selective catalysts, respectively. This finding is corrobo-
rated by a previous study [22] showing that a higher hydroxylation
degree of the catalyst leads to a lesser oxidation of PH, because of
limited surface coverage and hole transfer to the organic molecule.
Moreover PH oxidation produced higher amounts of ortho-isomer
with respect to the para-one when Merck was used. Conversely
P25 gave rise to an opposite trend. By increasing the initial PH con-
centration, however, the selectivity difference between ortho- and
para-monohydroxy derivatives, especially in the presence of P25,
decreased to a few percentage units. This behavior can be ascribed
to the high coverage of catalytic sites by PH molecules, causing a
lower selectivity of the hydroxyl radical attack.
The ortho-isomer was absent or its concentration was much
lower than that of the para-one in the presence of home prepared
catalysts especially when the PH initial concentration was lower
than 1 mM. The maximum regio-selectivity recorded was 11% in
the presence of HP0.5 and an initial concentration of 5 mM. Such
values can be explained by considering the strong interaction of
ortho-monohydroxy PH derivative, i.e. catechol, with catalyst sur-
face, determining a fast degradation before desorption. An increase
of PH initial concentration favors desorption because of the greater
competition on the catalytic sites. Hence, selection of appropriate
catalyst and initial concentration can avoid the release of ortho-
monohydroxy derivative into the solution bulk.
BA oxidation carried out by commercial TiO2 samples gave rise
to a quite uniform distribution in ortho-, meta- and para-isomers.
A slight predominance of ortho-derivative was observed in the
presence of Merck, whereas P25 produced prevalently the meta-
derivative. Home prepared HP0.5 sample gave rise to the only meta-
and para-isomers with total absence of the ortho-isomer up to an
initial substrate concentration of 1 mM, whereas 14% ortho-isomer
appeared when 5 mM initial concentration was used. An analogous
explanation to that given for ortho-monohydroxy PH derivative can
be provided. Using all the other home prepared catalysts yields
BA ortho-monohydroxy derivative in minor amounts, ranging from
The highest selectivity values were obtained with commercial
well crystallized samples for both substrates, accordingly to the
results of partial oxidation of glycerol [51]. Conversely, in the partial
oxidation of aromatic para-substituted alcohols to the correspond-
ing aldehydes [38,39] an opposite behavior was observed, with the
highest selectivities obtained in the presence of home-prepared
catalysts. In the last cases the highly crystalline powders degrade
both the substrate and its intermediates owing to their elevated
oxidant power. The substrates nature and their interaction with
catalyst surface determine different adsorption degree and reaction
routes and it is not possible to generalize the catalyst behavior with
respect to selectivity without considering the type of photoreaction
and the experimental conditions used. In particular Merck, which
is less oxidant than P25 (as showed in Table 3, columns tX=15%), was
more selective. HP873S was the most selective catalyst among the
home prepared ones for both the substrates (Table 4), and it was
also more selective than P25. The selectivity was quite low in the
presence of the other poorly crystalline home-prepared powders.
Fig. 3 reports the selectivity data as a function of the crystallinity
degree: it may be noted that by increasing the crystallinity, the
selectivity increases. The higher is the percentage of amorphous
phase, the greater is the amount of defects and therefore the prob-
ability of entrapment of the hole–electron pairs, thus reducing the
photocatalytic activity.
TGA analysis allowed to correlate the amount of total
hydroxylation of catalysts (Table 2) with selectivity toward mono-
hydroxylated derivatives. A sharp decrease in selectivity was found
by increasing the total hydroxylation degree, as shown in Fig. 4.
A similar correlation was found independently of the two differ-
ent OH groups (weekly and strongly adsorbed), indicating that the
kind of interaction between OH and TiO2 is not essential in deter-
mining reaction selectivity, whereas the total amount is a crucial
parameter.

M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
85
80
70
60
50
40
30
20
10
0
100
a)
b)
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
100
0
20
40
60
80
100
Crystallinity (%)
Crystallinity (%)
Fig. 3. Selectivity to mono-hydroxylated species vs. crystallinity of the used photocatalysts. (a) PH 0.5 mM; (b) BA 0.5 mM.
12% to 19%. A possible reason which may account for difference in
selectivity to ortho-isomer is that molecules can adsorb on differ-
ent TiO2 surface centers and the used catalysts contain a diverse
distribution of these various sites.
The results indicate that the production of monohydroxy deriva-
tives coming from PH usually is higher than that from BA,
suggesting that the presence of either EDG or EWG influences
not only the regioselectivity of hydroxyl radical, but also the rel-
ative amount of the monohydroxy derivatives. There was a lack
in the carbon mass balance when the concentration of the species
reported in the figures was considered and compared with the TOC
values, experimentally determined. This is due to the presence of
unknown compounds deriving from the aromatic ring opening.
Interestingly, dark adsorption of catechol (2OH-PH) and 2-
hydroxy-benzoic acid (2OH-BA) showed to be significant for both
compounds only on HP0.5: an experimental run carried out in this
study showed that starting with a concentration of 0.5 mM and cat-
alyst amount of 0.4 g L−1, 20% of the initial 2OH-PH was adsorbed on
TiO2 surface, whereas starting with 0.5 mM 2OH-BA and 0.4 g L−1
catalyst, 33.3% of substrate was adsorbed. This strong interaction
can be explained by considering the affinity of OH and COOH
groups with the hydroxylated high surface area of HP0.5. The spa-
tial proximity of two of these groups on the aromatic ring can be
probably responsible for the measured strong adsorption on the
catalyst. In fact, the meta- and para-hydroxy derivatives did not
show analogous behavior. Aran˜a et al. [52,53] studied the photo-
catalytic degradation of phenol and phenolic compounds by using
P25 as the catalyst. They observed a different way of adsorption
of the various molecules onto TiO2 surface. Catechol adsorption
was much higher with respect to other compounds and its inter-
action occurred mainly through the formation of a catecholate
monodentate. Starting from a 0.5 mM concentration, the percent-
age of adsorbed catechol on P25 and Merck surface was only 3.8%.
Figs. 5 and 6 show the evolution of TOC, concentration of PH,
BA and their monohydroxy-derivatives in the course of photo-
catalytic runs carried out by using P25 and HP873S. Moreover
p-benzoquinone concentration was reported when significant. CO2
concentration has been divided by the carbon atoms present in PH
(6) and BA (7) in order to be easily comparable with the disappeared
substrate expressed in molar concentration.
3.4. Kinetics
The data obtained in this work indicate that all the three mono-
hydroxy derivatives were obtained starting from BA, whereas the
main monohydroxylated compounds from PH were ortho- and
para-isomers being the meta-isomer totally absent.
The kinetics were studied considering the two parallel path-
ways previously discussed, as already highlighted in literature for
other substrates [54]. Photoreactivity results modeling was based
on the assumption that all the elementary reactions of BA and PH
oxidation, both partial and total, occur on the catalyst surface and
involve adsorbed species, which can interact with hydroxyl groups.
The rate-determining step of the photooxidation process on the
catalyst surface is hypothesized to be the second order reaction
between the hydroxyl radical and the aromatic molecule adsorbed
onto the catalyst surface. This model is commonly used to analyze
the kinetics of photocatalytic reactions and it generally provides a
satisfactory prediction of the progress of the species concentration
in liquid [55] or gas [56] phase.
The different kinds of sites hypothesized to exist onto the cat-
alyst surface are the following: the first one adsorbing O2, the
80
50
a)
b)
70
40
30
20
10
0
60
50
40
30
20
10
0
0
2
4
6
8
10
12
0
2
4
6
8
10
12
OHtotal (%)
OHtotal (%)
Fig. 4. Selectivity to mono-hydroxylated species vs. samples global hydroxylation determined by TGA for the used photocatalysts. (a) PH 0.5 mM; (b) BA 0.5 mM.

86
M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
Fig. 5. PH (a) and BA (b) photocatalytic oxidation runs by using Degussa P25. (a)
♦ TOC, ꢀ phenol, hydroquinone, and catechol; (b) ♦ TOC, ꢀ benzoic acid,
2-hydroxy benzoic acid, 3-hydroxy benzoic acid, and 4-hydroxy benzoic acid.
Fig. 6. PH (a) and BA (b) photocatalytic oxidation runs by using HP873S. (a) ♦ TOC,
ꢀ phenol,
hydroquinone,
catechol, and
p-benzoquinone; (b) ♦ TOC, ꢀ ben-
4-hydroxy
zoic acid,
2-hydroxy benzoic acid,
3-hydroxy benzoic acid, and
benzoic acid.
second one adsorbing substrates by producing mono-hydroxylated
species and the third one adsorbing the substrates and their inter-
mediate products by eventually producing their mineralization. As
the adsorbed oxygen acts as an electron trap thus hindering the
electron-hole recombination, the hydroxyl radical concentration
depends on the fractional sites coverage by O2 [54].
Under these assumptions the disappearance rate of substrate for
second order surface reactions in parallel (monohydroxylation and
mineralization) may be written in terms of Langmuir–Hinshelwood
kinetics as [54]:
in which COx, CS and CI are the oxygen, the substrates and the mono-
hydroxy derivatives concentrations in the aqueous phase, whereas
KOx, KMH and KMin are the equilibrium adsorption constants of oxy-
gen and of substrate on monohydroxylating and mineralizing sites,
respectively. Eqs. (3) and (4) are written under the hypothesis that
the values of equilibrium adsorption constants of the substrates and
monohydroxy derivatives are similar, both for the hydroxylation
site and for the mineralization one.
Owing to the fact that all the experiments were carried out in
a batch reactor by continuously bubbling oxygen into the liquid
phase, the fractional site coverage of oxygen is constant during the
occurrence of the run. By substituting Eqs. (3) and (4) in Eq. (1), one
obtains:
1 dNS
(−rS) ≡ −
= k1ꢀOxꢀ1 + k2ꢀOxꢀ2
(1)
S
dt
where S is the surface area of the catalyst, NS the substrate moles, t
the irradiation time and ꢀOx the fractional site coverage of oxygen.
k1 and k2 are the second order rate constants and ꢀ1 and ꢀ2 the
fractional site coverages of the substrate evolving to monohydroxy
derivatives and of the substrate that is directly oxidized to CO2,
respectively.
It is reasonable to hypothesise that the monohydroxy interme-
diates, once formed, compete with the starting substrates for the
adsorption on the mineralizing sites, thus the fractional site cover-
ages, given by the Langmuir relationship, are [54]:
V dCS
KMHCS
KMinCS
II
ꢀ
ꢂ
ꢀ
ꢂ
−
= kI
+ k
(5)
ꢁ
ꢁ
S
dt
1 + KMH CS +
ICI
1 + KMin CS +
ICI
where V is the reaction volume, kI = k1ꢀOx and kII = k2ꢀOx
.
The production rate of intermediates can be expressed as fol-
lows:
ꢁ
ꢁ
KMin ICI
ꢀ
d
ICI
V
S
KMHCS
II
= kI
− k
(6)
ꢀ
ꢂ
ꢂ
ꢁ
ꢁ
dt
1 + KMH CS +
ICI
1 + KMin CS +
ICI
K
OxCOx
The production rate of CO2, normalizing its concentration to the
substrates one, can be expressed as:
ꢀOx
=
(2)
(3)
1 + KOxCOx
ꢀ
KMin CS +
ꢀ
ꢂ
ꢁ
KMHCS
V d(Cꢀ)
ICI
ꢀ
ꢂ
ꢂ
ꢀ1
ꢀ2
=
ꢁ
= kII
(7)
ꢂ
ꢁ
1 + KMH CS +
ICI
S
dt
1 + KMin CS +
ICI
KMinCS
where Cꢀ is the normalized carbon dioxide, i.e. it is equal to CCD/6
ꢀ
=
(4)
ꢁ
1 + KMin CS +
ICI
in the case of PH and to CCD/7 in the case of BA, CCD being the

M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
87
Table 5
35
PH 0.5mM
BA 5mM
(a)
Equilibrium adsorption constants and kinetic constants for phenol and benzoic acid
photocatalytic partial oxidation (KMH, kI) and mineralization (KMin, kII).
30
25
20
15
10
5
Phenol
Benzoic acid
P25
KMH (mM−1
)
)
2.0
25
5.1
1.8
30
0.11
4.3
KMin (mM−1
kI ×106 (mmol m−2 s−1
)
)
kII ×106 (mmol m−2 s−1
2.83
Merck
KMH (mM−1
)
)
1.1
15
5.0
0.23
2.5
15
KMin (mM−1
kI ×106 (mmol m−2 s−1
kII ×106 (mmol m−2 s−1
HP0.5
)
)
0
0.83
1.4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
C' (mM)
KMH (mM−1
KMin (mM−1
)
)
0.38
2.8
0.052
0.15
9.1
8.2
0.033
0.28
8
7
6
5
4
3
2
1
0
kI ×106 (mmol m−2 s−1
)
)
PH 0.25mM
BA 0.1mM
(b)
kII ×106 (mmol m−2 s−1
measured carbon dioxide concentration. The global molar balance
can be described by the following relationship:
ꢃ
CS,0 = CS +
CI + Cꢀ
(8)
(9)
I
ꢃ
CS +
CI = CS,0 − Cꢀ ≡ y
I
in which CS,0 is the initial substrate concentration. By substituting
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Eq. (9) in Eq. (7), the following differential equation is obtained:
C' (mM)
V dy
S dt
K
Miny
−
= kII
(10)
45
40
35
30
25
20
15
10
5
1 + KMin
y
PH 1mM
BA 5mM
(c)
Integration of Eq. (10) with the limiting conditions that y = CS,0
at t = 0 and y = CS,0 − Cꢀ at t = t gives:
ꢄ
ꢅ
V
S
1
CS,0
CS,0 − Cꢀ
t =
ln
+ KMinCꢀ
(11)
kIIKMin
A least square best fitting procedure of Eq. (11) to all the experi-
mental data of Cꢀ versus t yields the values of kII and KMin (R2 > 0.95)
for BA and PH. This non linear fit was carried out by using the curve-
fitting function available in TK Solver 5.0. These values are reported
in Table 5 together with the 90% confidence intervals. In order to
compare the experimental data with the kinetic model (Eq. (11))
which gives t as a variable dependent on Cꢀ, for some representa-
tive runs, Fig. 7 shows the experimental values of t versus Cꢀ in the
case of P25, Merck and HP0.5 samples. The lines drawn through the
experimental points represent Eq. (11) in which the fitted parame-
ters have been substituted; a satisfactory fitting of the model to the
experimental data may be noted. Moreover, the obtained results
suggest that the mineralization pathway is similar for both sub-
strates. In fact the lines indicating the kinetic model are very close
each other, thus suggesting that the nature of substituent group
does not influence the mineralization reaction.
0
0
0.2
0.4
0.6
0.8
1
1.2
C' (mM)
Fig. 7. Experimental values of irradiation times versus Cꢀ: ꢁ, BA; ꢂ, PH. The con-
tinuous and dotted lines represent the kinetic model (Eq. (11)) for BA and PH runs,
respectively. (a) P25; (b) Merck; (c) HP0.5.
By substituting the values of kII and KMin in Eq. (13), the least
square best fitting procedure for all the runs, at different initial con-
centrations, yielded the values of kI and KMH for BA and PH. They are
reported in Table 5 for P25, Merck and HP0.5 catalysts together with
the 90% confidence intervals. Partial oxidation kinetic constants (kI)
are higher by using Merck sample for both PH and BA, conversely
In order to determine kI and KMH, Eq. (10) has been divided by
Eq. (5) giving rise to the following relation:
dy
dCS
Ay(1 + KMHy)
k
II figures are higher when using HP0.5. P25 shows different kinetic
=
(12)
CS[(1 + KMiny) + A(1 + KMHy)]
behavior for PH and BA oxidation: in the former case, kI is higher
than kII, contrarily to values obtained with BA. The last insight is
in agreement with the stronger mineralization generally obtained
from BA oxidation with respect to PH oxidation in the presence of
P25 (see Table 3). By using the same catalyst sample, moreover,
selectivity to mono-hydroxy derivatives was higher starting from
PH, in accord with the kinetic constants figures.
where
KMinkII
A =
K
MHkI
By integrating Eq. (12) with the limit conditions that y = y0 for
CS = CS,0 and y = y for CS = CS, one obtains the following equation:
ꢄ
ꢅ
As far as the equilibrium adsorption constants are concerned,
the highest ones are those referring to PH and BA mineralization
carried out by P25 sample. The lowest figures are instead the ones
ꢆ
CS,0
CS
y
y0
KMin
1 + KMHy0
1
A
y0 1 + KMH
y
ln
=
ln
+
ln
(13)
1 + KMHy0
AKMH
1 + KMH
y
y

88
M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
The photocatalytic results indicate that the products of the pri-
PH 0.5mM
BA 1mM
(a)
mary oxidation of phenol were the ortho and para-mono-hydroxy
derivatives while those of benzoic acid were all the mono-hydroxy
derivatives. For both the substrates the ortho-isomer was absent
or its concentration was much lower than that of the para-one
in the presence of home prepared catalysts because of the strong
interaction of ortho-monohydroxy derivative with catalyst surface,
especially when the substrate initial concentration was lower than
1 mM. Hence the selection of the appropriate catalyst and of the
substrate concentration avoid the release of ortho isomer in the
solution facilitating the separation operations. The highest selec-
tivity values were obtained with the commercial sample Merck for
both substrates. As expected, the selectivity toward mono-hydroxy
derivatives and the distribution of the regio-isomers are strongly
influenced by the photocatalyst properties. In particular, a depen-
dence of the selectivity on the percentage of catalyst crystallinity
and on the total amount of OH groups present on the catalyst sur-
face has been observed. Crystalline and poorly hydroxylated TiO2
Merck sample, and HP0.5, the most hydroxylated one, are the most
and the least selective catalysts, respectively.
0
0.2
0.4
0.6
0.8
1
y (mM)
5
4.5
4
PH 5mM
BA 5mM
(b)
3.5
3
A kinetic Langmuir–Hinshelwood model, taking into account
the two parallel pathways of mineralization and of partial oxida-
tion, was developed for P25, Merck and HP0.5 samples. Equilibrium
adsorption constants and kinetic constants found for the mineral-
ization route were similar for both substrates, thus indicating that
the nature of the substituent group does not influence the min-
eralization reaction. Opposite trends were observed for the partial
oxidation constants. Finally it can be stated that the fitting of exper-
imental data to the kinetic model is generally satisfactory.
2.5
2
1.5
1
0.5
0
4
4.1
4.2
4.3
y (mM)
4.4
4.5
4.6
1.1
PH 1mM
BA 0.5mM
(c)
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
References
[1] M. Schiavello (Ed.), Photocatalysis and Environment. Trends and Applications,
Kluwer, Dordrecht, 1988.
[2] A. Fujishima, K. Hashimoto, T. Watanabe, TiO2 Photocatalysis. Fundamentals
and Applications, BKC Inc., Tokyo, 1999.
[3] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C: Photochem. Rev.
1 (2000) 1–21.
[4] K. Okamoto, Y. Yamamoto, H. Tanaka, M. Tanaka, Bull. Chem. Soc. Jpn. 58 (1985)
2015–2022.
[5] V. Augugliaro, L. Palmisano, A. Sclafani, C. Minero, E. Pelizzetti, Toxicol. Environ.
Chem. 16 (1988) 89–109.
[6] I. Ilisz, A. Dombi, Appl. Catal. A: Gen. 180 (1999) 35–45.
[7] A.M. Peiró, J.A. Ayllón, J. Peral, X. Doménech, Appl. Catal. B: Environ. 30 (2001)
359–373.
[8] (a) S. Parra, J. Olivero, L. Pacheco, C. Pulgarin, Appl. Catal. B: Environ. 43 (2003)
293–301;
(b) X. Li, J.W. Cubbage, T.A. Tetzlaff, W.S. Jenks, J. Org. Chem. 64 (1999)
8509–8524.
[9] N.N. Rao, A.K. Dubey, S. Mohanty, P. Khare, R. Jain, S.N. Kaul, J. Hazard. Mater.
B101 (2003) 301–314.
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
y (mM)
Fig. 8. Experimental values of CS versus y (defined in Eq. (9)): ꢁ, BA; ꢂ, PH. The
continuous and dotted lines represent the kinetic model (Eq. (13)) for BA and PH
runs, respectively. (a) P25; (b) Merck; (c) HP0.5.
[10] C. Guillard, J. Disdier, J.-M. Herrmann, C. Lehaut, T. Chopin, S. Malato, J. Blanco,
Catal. Today 54 (1999) 217–228.
[11] D.S. Bhatkhande, V.G. Pangarkar, A.A.C.M. Beenackers, Water Res. 37 (2003)
1223–1230.
[12] G. Marcì, M. Addamo, V. Augugliaro, S. Coluccia, E. García-López, V. Loddo,
G. Martra, L. Palmisano, M. Schiavello, J. Photochem. Photobiol. A: Chem. 160
(2003) 105–114.
[13] A.H.C. Chan, C.K. Chan, J.P. Barford, J.F. Porter, Water Res. 37 (2003) 1125–1135.
[14] D. Vione, C. Minero, V. Maurino, M.E. Carlotti, T. Picatonotto, E. Pelizzetti, Appl.
Catal. B: Environ. 58 (2005) 79–88.
corresponding to mono-hydroxylation of BA and PH by using P25
and HP0.5, respectively.
For some representative runs, Fig. 8 reports the values of CS as
a function of y; the lines drawn through the experimental points
represent the kinetic model (Eq. (13)) in which the fitted param-
eters have been substituted; also in this case a satisfactory fitting
of the model to the experimental data may be noted. The results
indicate that the production of monohydroxy derivatives is higher
for PH than for BA suggesting that the presence of either EDG or
EWG influences not only the regio-selectivity of hydroxyl radical,
but also the relative amount of the monohydroxy derivatives.
[15] G. Palmisano, M. Addamo, V. Augugliaro, T. Caronna, E. García-López, V. Loddo,
L. Palmisano, Chem. Commun. 9 (2006) 1012–1014.
[16] P. Piccinini, C. Minero, M. Vincenti, E. Pelizzetti, Catal. Today 39 (1997) 187–195.
[17] V. Augugliaro, A. Bianco Prevot, V. Loddo, G. Marcì, L. Palmisano, E. Pramauro,
M. Schiavello, Res. Chem. Intermed. 26 (2000) 413–426.
[18] G. Mailhot, L. Hykrdová, J. Jirkovsky´, K. Lemr, G. Grabner, M. Bolte, Appl. Catal.
B: Environ. 50 (2004) 25–35.
4. Conclusions
[19] A. Di Paola, V. Augugliaro, L. Palmisano, G. Pantaleo, E. Savinov, J. Photochem.
Photobiol. A: Chem. 155 (2003) 207–214.
[20] S. Ahmed, M.G. Rasul, W.N. Martens, R. Brown, M.A. Hashib, Desalination 261
(2010) 3–18.
[21] K. Chhor, J.F. Bocquet, C. Colbeau-Justin, Mater. Chem. Phys. 86 (2004) 123–131.
[22] T.R.N. Kutty, S. Ahuja, Mater. Res. Bull. 30 (1995) 233–241.
[23] C. Gomes Silva, J.L. Faria, J. Mol. Catal. A: Chem. 305 (2009) 147–154.
The photocatalytic hydroxylation of a substrate containing an
electron donor group (phenol) and one containing an electron
withdrawing group (benzoic acid) in the presence of commer-
cial and home prepared TiO2 photocatalysts has been reported.

M. Bellardita et al. / Applied Catalysis A: General 441–442 (2012) 79–89
89
[24] C. Adán, J. Carbajo, A. Bahamonde, A. Martínez-Arias, Catal. Today 143 (2009)
247–252.
[40] A. Di Paola, M. Bellardita, R. Ceccato, L. Palmisano, F. Parrino, J. Phys. Chem. C
113 (2009) 15166–15174.
[25] M. Mungmart, U. Kijsirichareonchai, N. Tonanon, S. Prechanont, J. Panpranot,
T. Yamamoto, A. Eiadua, N. Sano, W. Tanthapanichakoon, T. Charinpanitkul, J.
Hazard. Mater. 185 (2011) 606–612.
[26] T. Velegraki, D. Mantzavinos, Chem. Eng. J. 140 (2008) 15–21.
[27] Y. Deng, K. Zhang, H. Chen, T. Wu, M. Krzyaniak, A. Wellons, D. Bolla, K. Douglas,
Yuegang Zuo, Atmos. Environ. 40 (2006) 3665–3676.
[28] V.G. Gandhi, M.K. Mishra, M.S. Rao, A. Kumar, P.A. Joshi, D.O. Shah, J. Ind. Eng.
Chem. 17 (2011) 331–339.
[41] (a) A. Di Paola, G. Cufalo, M. Addamo, M. Bellardita, R. Campostrini, M. Ischia,
R. Ceccato, L. Palmisano, Colloids Surf. A 317 (2008) 366–376;
(b) A. Di Paola, M. Addamo, M. Bellardita, E. Cazzanelli, L. Palmisano, Thin Solid
Films 515 (2007) 3527–3529.
[42] M. Bellardita, M. Addamo, A. Di Paola, G. Marcì, L. Palmisano, L. Cassar, M. Borsa,
J. Hazard. Mater. 174 (2010) 707–713.
[43] P. Scherrer, Nachr. Ges. Wiss. Göttingen Math. Phys. Kl. 2 (1918) 96–100.
[44] H. Jensen, K.D. Joensen, J.E. Jørgensen, J.S. Pedersen, E.G. Søgaard, J. Nanopart.
Res. 6 (2004) 519–526.
[29] A. Sobczyn´ ski, Ł. Duczmal, W. Zmudzin´ ski, J. Mol. Catal. A: Chem. 213 (2004)
225–230.
[45] J.S. Noh, J.A. Schwarz, J. Colloid Interface Sci. 130 (1989) 157–164.
[46] T. Mahmood, M. Tahir Saddique, A. Naeem, P. Westerhoff, S. Mustafa, A. Alum,
Ind. Eng. Chem. Res. 50 (2011) 10017–10023.
[47] T. Preocˇanin, N. Kallay, Croat. Chem. Acta 79 (2006) 95–106.
[48] R. Mueller, H.K. Kammler, K. Wegner, S.E. Pratsinis, Langmuir 19 (2003)
160–165.
[30] M. Salaices, B. Serrano, H.I. de Lasa, Chem. Eng. Sci. 59 (2004) 3–15.
[31] A. Ortiz-Gomez, B. Serrano-Rosales, H. de Lasa, Chem. Eng. Sci. 63 (2008)
520–557.
[32] A. Quintanilla, J.A. Casas, A.F. Mohedano, J.J. Rodríguez, Appl. Catal. B: Environ.
67 (2006) 206–216.
[33] S. Huixian, Z. Tianyong, L. Bin, W. Xiao, H. Meng, Q. Mingyan, Catal. Commun.
12 (2011) 1022–1026.
[49] D. Gumy, S.A. Giraldo, J. Rengifo, C. Pulgarin, Appl. Catal. B: Environ. 78 (2008)
19–29.
[34] J. Gao, M. Liu, X. Wang, X. Guo, Ind. Eng. Chem. Res. 49 (2010)
2194–2199.
[50] V. Augugliaro, S. Coluccia, V. Loddo, L. Marchese, G. Martra, L. Palmisano, M.
Schiavello, Appl. Catal. B: Environ. 20 (1999) 15–27.
[35] G. Palmisano, M. Addamo, V. Augugliaro, T. Caronna, A. Di Paola, E. García-
López, V. Loddo, G. Marcì, L. Palmisano, M. Schiavello, Catal. Today 122 (2007)
118–127.
[36] G. Palmisano, V. Augugliaro, M. Pagliaro, L. Palmisano, Chem. Commun. (2007)
3425–3437.
[51] V. Augugliaro, H.A. Hamed El Nazer, V. Loddo, A. Mele, G. Palmisano, L.
Palmisano, S. Yurdakal, Catal. Today 151 (2010) 21–28.
[52] J. Aran˜a, E. Pulido Melián, V.M. Rodríguez López, A. Pen˜a Alonso, J.M. Don˜a
Rodríguez, O. González Díaz, J. Pérez Pen˜a, J. Hazard. Mater. 146 (2007)
520–528.
[37] G. Palmisano, E. García-López, G. Marcì, V. Loddo, S. Yurdakal, V. Augugliaro, L.
Palmisano, Chem. Commun. 46 (2010) 7074–7089.
[38] G. Palmisano, S. Yurdakal, V. Augugliaro, V. Loddo, L. Palmisano, Adv. Synth.
Catal. 349 (2007) 964–970.
[53] J. Aran˜a, J.M. Don˜a Rodríguez, O. González Díaz, J.A. Herrera Melián, C. Fernán-
dez Rodríguez, J. Pérez Pen˜a, Appl. Catal. A: Gen. 299 (2006) 274–284.
[54] G. Palmisano, V. Loddo, V. Augugliaro, L. Palmisano, S. Yurdakal, AIChE J. 53
(2007) 961–968.
[39] V. Augugliaro, T. Caronna, V. Loddo, G. Marcì, G. Palmisano, L. Palmisano, S.
Yurdakal, Chem. Eur. J. 14 (2008) 4640–4646.
[55] A. Gora, B. Toepfer, V. Puddu, G. Li Puma, Appl. Catal. B: Environ. 65 (2006) 1–10.
[56] H. Ibrahim, H. de Lasa, AIChE J. 50 (2004) 1017–1027.