A proof of the direct hole transfer in photocatalysis: The case of melamine

Article abstract of DOI:10.1016/j.apcata.2015.11.012

The photoinduced transformation of 2,4,6-triamino-1,3,5-triazine (melamine) was studied by using different advanced oxidation technologies under a variety of experimental conditions. The systems involving homogeneous hydroxyl radicals, as generated by H₂O₂/hν, Fenton reagent, and sonocatalysis are ineffective. However, melamine is degraded under photocatalytic conditions or by SO₄^- (S₂O₈^2-/hν). The time evolution of long-living intermediates, such as 2,4-diamino-6-hydroxy-1,3,5-triazine (ammeline) and 2-amino-4,6-dihydroxy-1,3,5-triazine (ammelide), has been followed, being 2,4,6-trihydroxy-1,3,5-triazine (cyanuric acid) the final stable product. During both photocatalytic and S₂O₈^2-/hν experiments, in the early steps, a fairly stable intermediate evolving to ammelide is observed in a large extent. This intermediate was identified as 2,4-diamino-6-nitro-1,3,5-triazine. This indicates that the primary photocatalytic event is the oxidation of the amino-group to nitro-group through several consecutive fast oxidation steps, and that a hydrolytic step leads to the release of nitrite in solution. To elucidate the nature of the oxidant species hole scavengers such as methanol and bromide ions were added to the irradiated TiO₂. They completely stop the degradation, whereas chloride and fluoride ions decrease the degradation rate. The study of the photocatalytic degradation rate of melamine at increasing concentrations using two different commercial titanium dioxides, such as P25 and Merck TiO₂, showed an intriguing behavior. A drastic abatement of the melamine transformation rate was observed when coagulation of the P25 slurry occurs due both to the pH change and melamine concentration effect that increase melamine adsorption. In the presence of TiO₂ (Merck) the melamine initial degradation rates are significantly lower than those observed in the presence of P25 but are not depressed at larger concentrations. The experimental evidences (e.g.; absence of melamine adsorption onto TiO₂ surface at low concentrations or at acidic pH or due to the catalyst surface texture, and the lack of reactivity toward OH free and bound) suggest that the effective photocatalytic mechanism is based on an outer sphere direct hole transfer to the melamine. Its formal potential lies in the range 1.9-2.3 V vs NHE. Then, the photodegradation of melamine is an efficient tool to evaluate the direct hole transfer ability of a photocatalyst.

Full text of DOI:10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
Applied Catalysis A: General xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
A proof of the direct hole transfer in photocatalysis: The case of
melamine
Valter Maurino, Marco Minella, Fabrizio Sordello, Claudio Minero ,1
∗
Chemistry Department and NIS Centre of Excellence, University of Torino, Via P. Giuria 5, Torino 10125, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The photoinduced transformation of 2,4,6-triamino-1,3,5-triazine (melamine) was studied by using dif-
ferent advanced oxidation technologies under a variety of experimental conditions. The systems involving
homogeneous hydroxyl radicals, as generated by H2O2/h␯, Fenton reagent, and sonocatalysis are inef-
Received 28 August 2015
Received in revised form 3 November 2015
Accepted 6 November 2015
Available online xxx
•−
2−
fective. However, melamine is degraded under photocatalytic conditions or by SO4 (S2O8 /h␯). The
time evolution of long-living intermediates, such as 2,4-diamino-6-hydroxy-1,3,5-triazine (ammeline)
and 2-amino-4,6-dihydroxy-1,3,5-triazine (ammelide), has been followed, being 2,4,6-trihydroxy-1,3,5-
Keywords:
Photocatalysis
Melamine
Hydroxyl radicals
TiO2
Advanced Oxidation Technologies
Fluorinated titania
2−
triazine (cyanuric acid) the final stable product. During both photocatalytic and S2O8 /h␯ experiments,
in the early steps, a fairly stable intermediate evolving to ammelide is observed in a large extent. This
intermediate was identified as 2,4-diamino-6-nitro-1,3,5-triazine. This indicates that the primary pho-
tocatalytic event is the oxidation of the amino-group to nitro-group through several consecutive fast
oxidation steps, and that a hydrolytic step leads to the release of nitrite in solution. To elucidate the
nature of the oxidant species hole scavengers such as methanol and bromide ions were added to the
irradiated TiO2. They completely stop the degradation, whereas chloride and fluoride ions decrease the
degradation rate.
The study of the photocatalytic degradation rate of melamine at increasing concentrations using two
different commercial titanium dioxides, such as P25 and Merck TiO2, showed an intriguing behavior. A
drastic abatement of the melamine transformation rate was observed when coagulation of the P25 slurry
occurs due both to the pH change and melamine concentration effect that increase melamine adsorp-
tion. In the presence of TiO2 (Merck) the melamine initial degradation rates are significantly lower than
those observed in the presence of P25 but are not depressed at larger concentrations. The experimental
evidences (e.g., absence of melamine adsorption onto TiO2 surface at low concentrations or at acidic pH
•
or due to the catalyst surface texture, and the lack of reactivity toward OH free and bound) suggest that
the effective photocatalytic mechanism is based on an outer sphere direct hole transfer to the melamine.
Its formal potential lies in the range 1.9–2.3 V vs NHE. Then, the photodegradation of melamine is an
efficient tool to evaluate the direct hole transfer ability of a photocatalyst.
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
that retain the amino group (e.g., from atrazine ammeline is
formed) and no cyanuric acid is observed [4,5]. These evidences
suggest to study the behavior of melamine (2,4,6-triamino-1,3,5-
triazine) under photocatalytic conditions and in OH generating
systems.
The photocatalytic transformation of several s-triazine deriva-
•
tives has been extensively investigated [1–3]. s-Triazine derivatives
have the unique property to lose easily the substituents, but
to retain the s-triazine ring till to the formation of 2,4,6-
trihydroxy-1,3,5-triazine (cyanuric acid), which is stable to further
photocatalytic oxidation [1]. Interestingly, other oxidation pro-
cesses like UV/H O and Fenton reaction lead to stable products
A preliminary investigation about the UV/TiO /H O melamine
2
2
2
degradation was proposed [6] where a not complete mineraliza-
tion was reported because of the formation of cyanuric acid as
last product. An increase of the toxicity during melamine degra-
dation was observed suggesting an undesirable increase of toxicity
also during melamine degradation in living tissues. Nothing was
reported about the nature of intermediates during the first steps of
the photocatalytic process.
2
2
∗ Corresponding author. Fax: +39 011 6707615.
E-mail address: claudio.minero@unito.it (C. Minero).
Web: http://www.environmentalchemistry.unito.it.
1
http://dx.doi.org/10.1016/j.apcata.2015.11.012
0
926-860X/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
2
In the present paper the photocatalytic transformation of
melamine by using different oxidation processes under a variety
of conditions is reported. Attention is devoted to the mechanism
of oxidation, and in particular to the role of hydroxyl radical vs
order to avoid interference from organic impurities and inorganic
ions adsorbed on the photocatalyst. The physical and photochem-
ical properties of the used photocatalyst are reported elsewhere
[18].
direct hole oxidation mechanism. In TiO photocatalysis the role of
Aqueous stock solutions of Melamine (1000 mg dm−3), AN
2
•
−3
−3
−3
bound or free OH mediated oxidation vs direct hole transfer oxi-
(20 mg dm ), AD (20 mg dm ) and CYA (1000 mg dm ) are fairly
stable and last for weeks.
dation has been extensively debated [7] and the elucidation of the
significance of these pathways has a fundamental importance in the
understanding and control of photocatalytic processes. Evidences
in support of both mechanisms have been obtained [8,9]. As far as
mechanism is concerned, hydroxyl radical addition or direct elec-
tron abstraction can be undistinguishable based on the detected
intermediates for most of the substrates.
All the aqueous solutions were prepared employing ultrapure
water obtained with a MilliQ plus apparatus (TOC = 2 ppb, conduc-
tivity 18.2 M ꢀ cm).
2.2. Degradation experiments
Some experimental evidences reported on phenol, formate,
hydrogen peroxide and glycerol photocatalytic transformations
shed some light on these possible pathways [10,11]. Moreover,
they gave insights into the energetics of the surface traps for holes,
including also the states associated to adsorbed substrates, and
the possibility of water oxidation. It was outlined how the intrin-
sic and extrinsic surface properties can affect the selectivity of the
photocatalytic degradation toward different substrates [12]. More-
over, changes in the extrinsic surface properties, and in particular
the adsorption of redox stable ions, influence the relative role of
the different paths during the photocatalytic process. Particular
attention was devoted to the specific adsorption of fluoride ions
The slurries for photocatalytic experiments were prepared by
suspending with sonication the required amount of photocatalyst
powder and addition of the needed amount of the aqueous stock
solution of substrate. Being 5.5 the natural pH of a TiO suspen-
2
−
3
sion (500 mg dm ) the pH of the slurries before irradiation was
adjusted by adding drops of 1 M solutions of NaOH or HClO4 as
required.
The irradiation experiments were carried out in Pyrex glass
cut-off at 295 nm) or quartz cylindrical cells (4.0 cm diameter,
.3 cm height) containing 5 mL of the aqueous suspension of the
(
2
photocatalyst powder and substrate, using a Philips TLK 40 W/05
fluorescent lamp (Phillips, Eindhoven, Nederland) in standard con-
(
ligand exchange reaction between the surface hydroxy groups
−
2
ditions (33 W m ). This lamp emits a band 60 nm wide, centered
at 360 nm (the complete spectrum is reported in Ref. [19]). Homo-
and the fluoride ions [11]), which promotes the phototransfor-
•
mation of substrates that react predominantly via OH mediated
2−
geneous degradation with the H O2 or S O8 /UV systems was
2
2
oxidation (e.g., phenol) [7,13], while decreases the degradation rate
of substrates (e.g., hydrogen peroxide, catechol) that react pre-
dominantly by direct hole transfer mechanism because hinders
their specific adsorption [11,14–16]. Depending on the substrates,
hydroxyl radicals, direct hole transfer (inner or outer sphere) or
reductive pathways may operate as initial step and be active all
together during the overall degradation process.
carried out in quartz cells, using a Philips 20 W low pressure mer-
cury lamp, emitting at 254 nm. The cell apparatus was described
−
6
elsewhere [1]. The total photon fluxes in the cells were 7.1 × 10
−
8
−1
and 7.9 × 10 Einstein min in the 200–420 nm range with the
fluorescent and the mercury lamp, respectively (ferrioxalate acti-
◦
nometry). The cell temperature during irradiation was 30 ± 3 C.
The experiments in the absence of air were prepared purging with
He the closed irradiation cells containing either the TiO2 suspen-
sion or water. Then, the required volume of substrate stock solution
and, when relevant, the hydrogen peroxide solution were injected
in the cell with a microsyringe. During irradiation the slurries were
magnetically stirred.
A substrate with a high monoelectronic oxidation potential, not
•
reactive toward OH radicals and poorly adsorbed onto TiO2 can
help to probe a mechanism that involves a direct hole transfer
(
likely promoted by shallow surface traps) in photocatalysis. To get
information on the hydroxyl radical mechanism, experiments with
•
homogeneous OH generating systems were performed. More-
Sonication experiments were conducted with a Branson Sonifier
B-15 equipped with standard horn and tip, and a stainless steel
sealed 50 mL chamber with cooling jacket. The output setting was
adjusted to obtain a 65 W output at 20 kHz. The temperature was
over, attention was devoted to the study of the differences in
the photocatalytic activity of TiO2 P25 and TiO2 Merck, two
different commercial powders with marked differences in their
surface features, to highlight the role of the surface properties
on the photocatalytic process mechanism. Considering that cya-
nuric acid is the final product of melamine transformation, this
◦
maintained at 25 C.
Degradation runs with the Fenton reagent were carried out in
the presence of H O2 50 mM and FeSO4 1 mM at pH 2 (H SO ). The
2
2
4
allowed investigating on the mechanism of −NH substitution.
2
reaction was stopped by adding methanol.
Previous studies showed that under photocatalytic conditions
the substituents containing nitrogen are redox interconverted
(
−NH → −NHOH → −NO → −NO ) [17].
2.3. Analytical determinations
2
2
The HPLC determinations were carried out with a Hitachi Elite
Lachrom L2200, equipped with a Diode Array Detector (Hitachi
L-2455), on the filtered (Millex HV 0.45 ␮m, Millipore) irradiated
aqueous samples.
2
. Experimental
2.1. Materials and reagents
Melamine, AN, AD and CYA were quantified with ion pair
chromatography with a bonded phase octadecylsilica column
(LiChrospher R100-CH 18/2 by Merck, 250 mm length, 10 mm
i.d., 5 ␮m packing); the mobile phase was 0.01 M sodium hexane
sulfonate (Aldrich ion-pair reagent 99+%) and 0.014 M H PO dis-
Melamine (99+%, Aldrich), Cyanuric acid (CYA) (98%, Aldrich)
were used without further purification. Ammeline (AN) and
Ammelide (AD) were synthesized with the procedure described
elsewhere [1]. All other chemicals were commercially available,
with at least analytical purity, and used without further purifi-
3
4
−
1
solved in water/CH CN 95/5 at 1 mL min . Retention times were:
3
cation. TiO P25 by Evonik (formerly Degussa, SSABET area ca
melamine 7.8 min, AN 6.8 min, AD 3.1 min and CYA 2.3 min. The
detection was carried out at 200 nm for melamine and CYA, at
221 nm for AD and at 229 nm for AN to optimize the analytical
sensitivity for each compounds.
2
2
−l
5
0 m g , mixture rutile:anatase 20:80) and TiO by Merck (SSABET
2
2
−l
area 10 m g , anatase 100%) were irradiated in aerated aqueous
suspension for at least 12 h and washed with ultrapure water in
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
3
Fig. 1. Photocatalytic transformation of melamine: TiO2 500 mg dm−3; initial substrate concentration 1.0 × 10 M; irradiation at 360 nm. A: pH 2.0 (HClO4); B: pH 5.5 (nitrate
−4
concentration refers to the right axis); C: pH 12 (NaOH). The reported mass balance is the sum of concentration of all detected species.
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
4
The formation of nitrate and nitrite ions was monitored by sup-
pressed ion chromatography, utilizing a Dionex DX500 apparatus
equipped with a Dionex AS4-SC separation column and an alkaline
buffer eluent containing Na CO (1.8 mM) and NaHCO3 (1.7 mM)
2
3
1
−
at the flow rate of 1.5 mL min . Ammonium ion quantifications
were performed using the Berthellot reagent (Merck) following the
procedures suggested by the vendor.
An aqueous solution of 2,4-diamino-6-nitro-1,3,5-triazine
(
4
Compound I) was synthetized for the first time irradiating for
−
1
0 min 1000 mL of a suspension containing melamine (10 mg L
)
and the photocatalyst in the same conditions employed for the
melamine photodegradation. After irradiation the suspension was
filtered (HV 0.45 ␮m, Millipore) and freeze-dried. The solid was
dissolved in water (40 mL), the residual melamine was precipi-
tated with the stoichiometric amount of cyanuric acid. The solution
of I, impure for AN and AD, was purified by chromatography
employing a column (500 mm length, 10 mm i.d., Superformance,
Merck) packed with octadecylsilica (40–63 ␮m diameter, irregu-
larly shaped, Merck). The purification runs were carried out on 2 mL
aliquots of the solution of I, with ultrapure water as eluent (flow
−
1
rate 4 mL min ). The column eluate was monitored with an UV
detector (ꢁ = 220 nm, Hitachi model L4200). The fractions contain-
ing I (confirmation was obtained by analytical HPLC) were pooled
and freeze dried.
Mass spectra were recorded with a Finnigan-MAT 95Q (Bremen,
Germany) hybrid instrument, with BEQ geometry. The ions were
collected at the first dynode-electron multiplier detector, located
after the magnetic and electrostatic analyzers. The sample was
introduced with a direct insertion probe, heated until a constant
sample vapor pressure was produced in the ion source. Electron
impact (EI) mass spectra were obtained at 70 eV electron energy
and 1.0 mA electron current. The mass analyzer was scanned from
m/z 29 to m/z 300. Positive chemical ionization (CI) conditions were
established by admitting isobutane in the ion source, as the reagent
gas, at a pressure of 50 Pa (0.5 m bar). With this pressure, the ratio
between m/z 57 and m/z 43 peaks was 10:1. The electron energy
was set to 200 eV and the electron current to 0.2 mA. The resolution
was set to m/ꢂm 1000 (5% valley) and the ion source temperature
◦
to 220 C. Electron capture chemical ionization (EC-CI) experiments
were carried out in the same conditions except for the polarity of
the ion-source and detector voltages, which were reversed to detect
negative ions.
3
. Results and discussion
3.1. Identification of intermediates and final products of
melamine photocatalytic degradation
Melamine disappearance followed pseudo-first order kinetics
over at least two half-lives in all conditions studied. When persul-
fate acted either as electron scavenger or photoactivable oxidant,
or other reactive chemicals were added to the system their concen-
tration did not vary significantly during the time interval in which
melamine disappeared.
Fig. 2. Mass spectra of Compound I. (A) Electron impact ionization. (B) Positive ion
chemical ionization. (C) Negative ion chemical ionization. The most characteristic
feature of these spectra is the neutral loss of 30 amu fragment from the molecular
ion, indicative of the presence of a nitro substituent. The molecular ions suggest
a molecular weight of 156 g mol , compatible with an s-triazine ring bearing two
amino and one nitro group as substituents.
Fig. 1 shows the time profiles of melamine disappearance and
product formation under the conditions presently investigated
(
the photocatalytic experiments were carried out with P25 TiO2
−
1
if not differently specified). As already observed in the s-triazine
pesticides degradation through Fenton reagent [4], O [5], and het-
3
erogeneous photocatalysis [1], many transformation products are
formed, among which ammeline, ammelide and cyanuric acid are
the major intermediates and final product, respectively. Cyanuric
acid turned out to be stable under the present systems (e.g., no vari-
ation is observed after 10 h of photocatalytic conditions, as well as
2−
with Fenton reagent, S O
and sonocatalysis).
2
8
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
5
In the melamine transformation under photocatalytic con-
ditions an additional long-living intermediate (Compound I) is
formed in the early stages of degradation at acidic pH, and partic-
ularly at pH 5.5 (Fig. 1B). To investigate its nature, the compound
was synthesized by using TiO photocatalysis and purified. Its basic
2
hydrolysis (NaOH 1 M) gives ammeline as sole detectable product
−
and NO2 as inorganic nitrogen species. This evidence, together
with the mass spectra reported in Fig. 2 (neutral loss of 30 from
the molecular ion, presence of a peak at m/z 30 characteristic of
nitro compounds, a molecular weight of 156) lead to assign the
structure of 2,4-diamino-6-nitro-1,3,5-triazine to the Compound
I. Moreover, the formation of a nitro group is confirmed by the
appearance of an absorption band with the maximum located at
2
95 nm in the UV spectrum. The synthesis of Compound I allowed
its quantification. Compound I is present at significant concentra-
2−
tion also in S O
/h␯ system at acid and neutral pH (see later). In
2
8
basic media Compound I was not observed due to its rapid hydrol-
ysis (half-life of the order of few seconds in NaOH 1 M). At acid pH
the detected concentration of Compound I is lower than at quasi-
neutral pH due to acid hydrolysis.
The quantification of all long-living intermediates accounted
for almost 100% of the mass balance. Besides the time evolu-
tion of detected intermediates, Fig. 1 reports also the sum of all
by-products concentrations. As the mass balance calculated on
detected intermediates is almost stoichiometric, possible other
products are short-living and rapidly converted. A mass balance
defect is present only at pH 5.5 (Fig. 1B) at times before Compound
I appearance, suggesting that some short-living intermediates pro-
ducing Compound I are formed and not detected.
As inorganic nitrogen-containing species are concerned, the
−
evolution of NO3 was followed, which is by far the most impor-
tant species accounting for 90% of the total released nitrogen
after 400 min in the melamine degradation with TiO /h␯ (see
2
Fig. 1B). Even aniline in oxygenated solution and in the same con-
−
ditions gave ca 60% of ammonium ions, and NO3 is essentially
formed through NH3 photocatalytic conversion [17]. In the case of
melamine (and also AD and AN) carbon atoms are in their high-
est oxidation number (+4). In comparable cases (e.g., CCl4 [20],
C(NO ) [21]) reductive pathways initiate the photocatalytic degra-
2
4
dation, but this is not occurring for melamine as in the absence
of an electron scavenger no reaction occurs (see Table 1, entry 4)
and because from mass balance there is no evidence of melamine-
2−
derived reduced species. Conversely, the presence of S O8 greatly
2
enhances the degradation rate (entries 11 and 12 of Table 1). In this
2
−
case please note that S O8 is ca 200 times more concentrated than
2
•
−
•−
O2 and generates the very active SO
produced by reduction of S O8 is an oxidant species that could
species. The SO4 radical
4
2−
2
increase the reaction rate, and this agrees with the entries 3 and
2−
1
1 of Table 1. In the presence of S O8
no significant difference
2
is observed if O2 is present or not. It turns out that electrons in
the conduction band, which could back react with holes and with
−
NO2 or other oxidized transient species (e.g., the primary prod-
2−
uct of the electron transfer), are scavenged by O and/or S O8
.
2
2
The reaction in the presence of He (Table 1 entry 12) would not be
2−
Fig. 3. Proposed pathways for the oxidative transformation of melamine
possible if S O
did not act as an electron scavenger (compare
2
8
2
−
promoted by TiO2/UV and S2O8 /h␯. [I] = 2,4-diamino-6-nitro-1,3,5-triazine,
AN] = ammeline, [AD] = ammelide, [CYA] = cyanuric acid. The reported structural
with Table 1 entry 4). In addition the results indicate that oxygen
is not crucial as found in many other photocatalytic degradations
[
formula for AN, AD, and CYA do not correspond to the most stable tautomer (see
Ref. [35]).
[
22]. This does not exclude that when oxygen is present, it can react
with intermediate radical species.
Therefore melamine degradation merely proceeds through an
oxidation pathway. The process can only initiate through oxidation
of nitrogen (here in −3 oxidation number) and further oxidation
should operate for the formation of traces of NH + (always less than
4
2% of the stoichiometric amount) from AN and AD. Among other
nitrogen-containing species, hydroxylamine was not detected even
in trace amount. On the basis of the above considerations the
scheme of reaction reported in Fig. 3 is proposed for the degradation
of melamine in the presence of irradiated TiO2.
of the amino group to −NO with subsequent hydrolytic release
2
−
−
of NO that is rapidly oxidized to NO3 . Ammonium ions, which
2
are not originated photocatalytically from nitrate, are formed in
very low amount through nitrite reduction [17]. Similar mechanism
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
6
Table 1
−
4
−3
Initial rates of photocatalytic degradation of melamine under different experimental conditions. Melamine initial concentration 1 × 10 M, TiO2 0.5 g dm , CH3OH 0.1 M,
−
−
−2
2−
−2
M
Cl 0.1 mM, Br 0.1 M, total fluoride 1 × 10 M, S2O8 5 × 10
Entry #
Redox system
Main active species
Initial rate (M s−1)
•
•
•
•
•
•
•
+
−
−
−
−
−
−
−
−
−8 (a)
1
2
3
4
5
6
7
8
9
pH 2.0, air
pH 3.0, air
pH 5.5, air
pH 5.5, He
pH 11.0, air
pH 12, air
OH, h , e , O2
8.0 × 10
+
−8
OH, h , e , O2
4.5 × 10
+
−8 (a)
OH, h , e , O2
9.7 × 10
+
OH, h , e
no reaction
+
−
−
−
−9
OH, h , e , O2
4.5 × 10
+
−
−
−9 (a)
OH, h , e , O2
7.2 × 10
pH 5.5, CH3OH, air
CH2OH, e O2
no reaction
−
•
•−
−
−
−
−9
pH 2.0, Cl , air
Cl , Cl2 , e O2
2.0 × 10
−
•
•−
−
pH 2.0, Br , air
Br , Br2 , e O2
no reaction
−
•
•
•
+
−
−
−8 (b)
10
11
12
pH 3.0, F , air
OH, h , e , O2
1.1 × 10
2
−
−
+
−
−6
pH 5.5, S2O8 , air
OH, h , SO4, O2
1.2 × 10
pH 5.5, S2O82 , He
OH, h , SO4
+
1.3 × 10
−6
(
a)
Data from Fig. 1.
Data from Fig. 5.
(
b)
•
The photocatalytic experiment alone is silent on the mechanism
Experiments in which OH radical is formed (see entries 4–6 in
•
of the primary oxidation step, whichever it is due to an electron
abstraction by the valence band hole (shallow trapped) or to a sur-
face deep trapped hole forming first a radical cation, or to a hydroxyl
radical addition or hydroxyl radical promoted H-abstraction. The
detected first long living intermediate I can be formed by all of
these processes.
Table 2) demonstrate that oxidation of melamine by OH radical
attack (both addition and H-abstraction) is not possible. Electro-
chemical treatments with high oxygen overvoltage anodes (cell
voltage of 4.0 V) proved to be ineffective for melamine oxidation
[24], although these conditions produce relatively high concentra-
tion of hydroxyl radicals in the proximity of the electrode. Free
hydroxyl radical does not react through electron transfer, due to
the high reorganization energy associated with transfer of charge
3
.2. Evidences for direct hole oxidation
[
25].
When generated by H O photolysis, the hydroxyl radi-
2
2
The melamine initial degradation rates in the presence of other
cal concentration is predominantly controlled by the reaction
with H2O2. Using the value of the kinetic constant reported
different redox systems are summarized in Table 2.
2−
Experiments carried out using S O8 /UV at different pH are
2
7
−1 −1
s ) [25], the quantum yield of 0.5 and the pho-
(
k = 2.7 × 10 M
reported in Fig. 4 and the melamine disappearance rate at entries
–3 of Table 2. The melamine degradation is faster than with irradi-
−
6
−1
tonic flux in the cell (7.1 × 10 Einstein min ), concentrations of
1
•
−11
−11
OH ranging from 2 × 10
M to 5 × 10
M are calculated when
ated TiO2 (Fig. 1) due to the different concentration of the primary
oxidant species, but the type and time evolution of intermediates
are equivalent, also considering the effect of pH. Also in this case the
sum of the concentrations of all detected intermediates accounts
for almost 100% of the initial melamine. As for photocatalytic trans-
formation, a mass balance defect is present at neutral pH (Fig. 4B),
even more pronounced than before. This can be attributed to the
faster reaction rate. A careful comparison of data of cyanuric acid
evolution of Figs. 1 and 4 shows that at acid and basic pH (although
not exactly the same) the formation of cyanuric acid is very simi-
lar. A difference is present only at neutral pH (pH 5.5 for TiO2 and
H O initial concentration ranges from 0.01 to 0.1 M. Thus, under
these conditions if the second order rate constant of OH radical
2
2
•
4
with melamine would be at least of the order of 10 , the rate would
−
10
−1
M s ; if compared with values of Table 1, melamine
be 2–5 × 10
would slightly disappear in the time windows of the experiment.
Conversely no reaction was observed in the presence of H O /h␯,
2
2
•
indicating an insignificant reactivity of OH radical with melamine
k < 10 ). The Fenton reagent is usually quite effective under the
adopted condition to oxidize organic substrates. The Fenton reac-
tion yields both OH (about 60% at pH 2) and ferryl (about 40%)
radicals, where OH alone is reactive, as recently assessed by a
kinetic study [26]. Thus in conditions similar to those of H O /h␯,
the previous considerations also apply.
Application of ultrasounds to aqueous solutions induces acous-
tic cavitation (bubble nucleation, growth, and implosive collapse).
Inside the collapsing microbubbles, temperature and pressure
increase up to very high values, inducing the pyrolysis of water
vapor and of the volatile organic compounds that can be present
in the gas phase. Pyrolysis of water vapor yields hydroxyl radicals
and hydrogen atoms. Reactions can take place in the gas phase, at
the gas–liquid interface, and in the solution bulk after transfer of
gaseous radicals into the liquid phase. Although quantification of
OH radical is a difficult task, under similar conditions quite refrac-
tory molecules are degraded [27]. With melamine as substrate no
reaction is observed.
From the above experiments (entries 4–6 of Table 2) it clearly
emerges that only heterogeneous photocatalysis on TiO and sys-
4
(
•
•
7
for persulfate), for which the photocatalytic rate is almost twice
than that in homogeneous solution. Peroxydisulfate in hot water
2
2
•
−
or under UV dissociates into SO
radical, for which the stan-
4
2−
•
−
◦
dard redox potential SO
/SO4 is quite large (E = 2.4 V vs NHE).
4
•
−
Because SO4 radical is quite efficient in promoting the melamine
degradation, this implies that the monoelectronic formal potential
of melamine oxidation is <2.4 V vs NHE. It is worth pointing out that
2
−
, even in absence of light or metal ions which promote the
S O
2
8
•
−
production of SO4 , degrades melamine (half-life of 30 days). From
•
−
a mechanistic point of view, SO4 radical acts mainly as an electron
abstractor [23]. This radical in unable to oxidize water to hydroxyl
radical at acidic/neutral pH as it is evident from the standard redox
potentials reported in Table 3.
•
On the basis of the above considerations the scheme of reaction
reported in Fig. 3 is confirmed for the degradation of melamine with
2
−
S O
/UV, in which the crucial step is melamine oxidation to 2,4-
2
8
2
diamino-6-nitro-1,3,5-triazine through a mechanism of electron
•
−
2−
tems generating SO4
radical can initiate the degradation of
abstraction for S O
. Possibly the electron abstraction mecha-
2
8
melamine, and that in the presently investigated time-domain
and for the photocatalytic experiments, hydroxyl radicals, both
free in solution, if present, and adsorbed, are not able to react
with melamine or at least to generate a transient intermediate
nism is valid also for TiO2 photocatalysis. The first oxidation step
is possibly the formation of a radical cation, which via some inter-
mediates apparent from the mass balance defect, rapidly evolves
to Compound I by further oxidation.
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
7
Table 2
Initial melamine degradation rates with different oxidation processes. Melamine initial concentration 1 × 10 M; Fenton reagent: H2O2 5 × 10−2 M, FeSO4 1 × 10−3 M at pH
−4
−2
2−
−2
2
(H2SO4); H2O2 5 × 10 M and S2O8 1 × 10 M; all the experiments were carried out under air.
Initial rate (M s 1
−
)
Entry #
Redox system
Main active species
S2O82 /h␯ (ꢁ = 254 nm) pH 1.0
−
−
−
SO4
SO4
•−
•−
•−
3.5 × 10
−6 (a)
−5 (a)
1
2
3
4
5
6
2
S2O8 /h␯ (ꢁ = 254 nm) pH 7.0
S2O82 /h␯ (ꢁ = 254 nm) pH 13.0
Fenton reagent
1.0 × 10
−6 (a)
SO4
1.1 × 10
•
•
•
OH
OH
OH, H
no reaction
no reaction
no reaction
H2O2/h␯ (ꢁ = 254 nm), pH 2–6
Sonocatalysis (ꢁ = 20 kHz, 65 W)
•
(
a)
Data from Fig. 4.
Table 3
Standard reduction potential for selected hole scavengers [25,45].
investigate the effect of pH, melamine concentration, surface mod-
ifiers and two different TiO2 specimens to figure out the surface
species responsible for the electron transfer.
Redox couple
Standard reduction potential vs NHE
•
•
−
There are many reports dealing with the thermodynamic of the
adsorption on TiO2 surface. Chemical adsorption (surface complex
formation) on TiO2 surface occurs when the substrate has oxy-
gen containing groups like carboxylate or alcoholic functions. The
review by Pang et. al [29] reports many of these cases, and in partic-
ular the cases of HCOOH and C2–C8 alcohols. As this is not the case
F/F
OH, H /H2O
3.6 V
2.7 V
2.6 V
2.4 V
2.3 V
2.0 V
1.9 V
1.7 V
1.6 V
−0.9 V
+
•
−
Cl /Cl
SO4 /SO4
Cl2 /2Cl
•
−
2−
•
−
−
•
−
Br /Br
•
−
−
−
OH/OH (free)
Br2/2Br
for melamine for which only amino-groups are available, melamine
•
−1
OH/OH (bound)
is not chemisorbed onto TiO . Experiments carried out with 2 g L
2
+
•
CH2O,H / CH2OH
−4
TiO P25 and 1 × 10 M let in the dark for 12 h showed no apprecia-
2
ble decrement of the solution concentration. Then an outer sphere
electron transfer can be suggested as primary mechanism for the
first step of melamine photocatalytic transformation.
evolving toward other more stable species. This confirms that the
only possible first step of melamine oxidation is the formation of
the radical cation by direct electron abstraction.
Consequently, the mechanism for the primary photocatalytic
degradation step is a direct electron abstraction by TiO2 photo-
generated holes. This possibility is supported by Nosaka et al. when
they suggest that spin-trapping agents react with surface-trapped
holes [9] and not with supposed free OH radical. Further substan-
tiation of this statement comes out from the results obtained in
the photocatalytic experiments in the presence of hole scavengers
entries 7–9 in Table 1). Hole scavengers, added at concentration
00–1000 times higher than melamine, transform the photogen-
erated oxidant species in other radical species unable to set off the
melamine degradation. The presence of CH OH or bromide com-
pletely suppresses the process.
Fig. 5 shows the initial degradation rate of melamine as
a function of the initial substrate concentration under differ-
ent photocatalytic conditions in the presence of TiO2 P25. The
photocatalytic melamine transformation was studied in a wide
concentration range that covers two orders of magnitude (from
−
5
−3
•
1 × 10 to 1 × 10 M) at different pHs (3, 5.5 and 11), in the pres-
−
3
ence or in the absence of fluoride ions and NaClO 10 M.
4
At pH 3, 11 and at pH 3 in the presence of fluorides (total fluo-
ride concentration 10 mM) the common saturation-type kinetics
was observed. The increasing rate vs substrate concentration is
also noticed at low concentration at pH 5.5. The sudden decrease
at this pH will be discussed later. The modelling of all the
processes occurring at the semiconductor–electrolyte interface
is elsewhere reported [30,31] and critically reviewed [32]. The
observed Langmuir–Hinshelwood-like dependence of the rate is
not due to the substrate adsorption, but derives from the recombi-
nation of charge carriers. From the quoted kinetic models the rate
should follow a square root dependence on the substrate concen-
tration. In the inset of Fig. 5 a log/log plot is presented. It is apparent
that the found slope for data at pH 3 and 5.5 has the predicted
dependence for substrates that are not chemisorbed. In the case
of chemisorption the rate dependence on concentration is much
more complex. An enlightening example is glycerol [10], for which
after an initial increase of the rate at low concentration, this decays
and later increases again. Then the observed trend of the rate is
compatible with a substrate that does not chemically adsorb.
Coming back to the possible involved surface species, depending
on the TiO2 specimen, different types of terminal and bridging OH
groups could be present on TiO2 [11] that can trap photogenerated
holes.
(
5
3
Considering the big difference between the substrate and the
hole scavenger concentration and that the reaction with melamine
would be a minority path, the photo-formed holes are quickly scav-
enged by methanol or bromide. The formed radicals are unable
to react with melamine. The standard reduction potentials for the
◦
•
◦
•
couples E Br
= 2.0 V vs NHE and E
= 1.7 vs NHE are
/
Br−
Br2 −/2Br−
apparently too low for oxidizing the melamine (see Table 3). In
the presence of chlorides, for which the standard reduction poten-
tials are larger (E
melamine oxidation is still possible, although the degradation rate
is significantly reduced. This observation implies that the standard
reduction potential E MEL +
couple (1.7 V) and lower than that of Cl2 /2Cl couple (2.3 V).
Having this standard redox potential melamine could be thermo-
dynamically oxidized (see Table 3) by persulfate (and this happens)
and by OH radicals (and this does not happen for kinetic restric-
tions as above discussed).
◦
◦
•
−
•
= 2.6 V vs NHE, E
= 2.3 vs NHE),
Cl− /Cl−
Cl2 /2Cl−
◦
•
•−
is larger than that of Br
2
−
/2Br
/
MEL
•
−
−
•
The schematic Ti(OH)(OH) for the (TiO ) unit at the sur-
2
face resumes many of the possible surface forms that can trap
the hole [28]. The unit surface Ti(OH)(OH) can trap holes as
3
.3. The hole transfer at the TiO surface
2
•
•
(
Ti(O )(OH)) ꢀ (Ti(OH)(O )), where the first (OH) refers to linear
hydroxyl groups and the right (OH) to bridged hydroxyl groups
28]. In the framework of water oxidation, several other forms have
been proposed, like [Ti = O Ti]s and [Ti Ti]s that refer to a triply
The electron transfer across the interface is at the heart of redox
photocatalysis as recently reviewed [28]. Electron transfer is an
interfacial phenomenon between a surface trapped charge carrier
and a chemisorbed or physisorbed species. The complex inter-
play of the surface texture and the system pH, which affects the
main surface species and the melamine charge, prompted us to
[
O
2
coordinated O atom and a bridging O atom at the surface, respec-
•
tively [33]. For the trapped hole the form [Ti O HO Ti]s was also
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
8
proposed suggesting the possibility that surface oxygen is included
in oxidation products. The schematic picture of hole trapping as
•
•
(
Ti(O )(OH)) ꢀ (Ti(OH)(O )) is useful only for a general discussion
of the electron transfer process and loses many details as above.
The hole trapping as Ti(O )(OH) is assumed important only at pH
3 and the most probable hole localization is Ti(OH)(O ), i.e., on
•
•
1
bridged oxygen.
The Ti(OH)(OH) can be protonated/deprotonated at various
extent depending on solution pH. Detailed studies on surface oxy-
gen species in a pH range 2.3–11.7 by internal reflection FTIR
spectroscopy reported that terminal hydroxyl Ti OH is present in
+
a pH range from 4.3 to 10.7 (maximum = 8), that Ti OH2 exists
+
in a pH range below 5, and Ti OH Ti exists in a pH range below
4
.3 [34]. At low pH the surface will be protonated with a positive
charge repelling positively charge protonated melamine. The pKa of
melamine is 5.0 [6,35] indicating that at pH below 4 the prevalent
form is the protonated one.
The above reported hole trapping scheme outlines the role of
solution pH on the electron transfer rate.
The maximum rate observed at pH 5.5 can be explained as the
result of the interplay among the valence band edge potential shift,
the substrate acid–base properties and the amphoteric properties
of the TiO2 surface (for P25 TiO2 pzc ≈ 6.3) [36]. The TiO2 CB edge
exhibits Nernstian dependence with pH, shifting about −60 mV per
pH unit in a wide range of pH [37] starting from E (pH = 0) = 3.0 V
vb
vs NHE for anatase [38]. As the surface is positively charged, the
+
electron abstraction from Ti(OH2 )(OH) is more difficult, imply-
ing that the hole trap Ti(OH2+)(O ) is shallower than Ti(OH)(O )
•
•
[
28]. From this it follows a greater potential difference with the
redox potential of the substrate to be oxidized and a greater reac-
tion rate, if additional effects due to the electrostatic interaction
with the surface are not present. Because the electron abstraction is
quite disfavored from the protonated melamine, the redox reaction
must involve neutral melamine. Considering the pKa of melamine
[
6,35], at pH below 4 the prevalent form (more than 99%) is the pro-
tonated one. The decrease in the degradation rate at pH 2.0 with
respect to pH 5.5 would follow the melamine speciation and the
rate is expected to be dramatically depressed. As this is not the
case, it must be hypothesized that the melamine near the surface
is in the not protonated form. This is possible only if melamine is
not chemisorbed, but resides in the double layer faced to the pos-
itive surface where the actual concentration of protons is lower
than in solution. In addition, the protonated surface is richer in
hydrogen atoms, which increase the possible hydrogen bonding
with melamine. These effects partially counterbalance the electro-
static repulsion between the catalyst surface and the protonated
melamine, and add to the increase in the driving force discussed
above (shallower trap) and guessed by the Marcus–Hush theory
[
8].
For basic pH the rate is depressed with respect to acid pH.
−
•
At basic pH the Ti(O )(O ) is stabilized, as an electron is easily
abstracted from a negative surface. This implies a lower oxidiz-
ing potential and reduced trapped hole reactivity. No data on the
second dissociation of melamine are available [35], and a further
dissociation is hardly envisaged leading to a negative charged imide
form. The possible additional adverse effects due to the electro-
static repulsion with the negative surface can be present implying
a lower rate. For the extreme investigated pH Evb= 2.29 V at pH 12
and Evb = 2.88 V at pH 2.0. Since at pH 12 melamine transforma-
tion was still observed (entries 1,3–6 in Table 1), even if with a rate
decrease of an order of magnitude with respect to pH 5.5, it follows
that the redox potential for electron abstraction from melamine
is ≤2.3 V, as already suggested above from the effect of chloride
addition.
2
−
and irradiation at
Fig. 4. Melamine transformation in the presence of S2O8
2
1
54 nm. Initial S2O82 concentration: 10 mM; initial substrate concentration:
−
−
4
.0 × 10 M. A: pH 1.0 (H2SO4 0.05 M). B: pH 7.0 (phosphate buffer 0.10 M); C:
pH 13 (NaOH 0.10 M). The reported mass balance is the sum of concentration of all
detected species.
The rate dependence on melamine concentration at pH 5.5
−
3
shows a sudden decrease. In the presence of NaClO4 10 M at pH
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
9
Fig. 5. Initial photocatalytic transformation rate of melamine as a function of the substrate concentration at pH 5.5, pH 3, pH 11, pH 3 in the presence of 10 mM of total
−3
−3
fluoride and pH 5.5 in the presence of NaClO4 10 M (air, TiO2 P25 0.5 g dm ). Inset: logarithmic plot of rates at pH 3 and 5.5 (substrate concentration <0.2 mM).
5
.5 the same is observed also at very low melamine concentra-
general fluorination increases the rate for substrates that react with
•
tion. The effect is not observed at lower pH. Thus the TiO2 surface
should have low adsorption ability toward the perchlorate anion,
and the role of NaClO4 would be limited to form a more compact
diffusion layer (Guy–Chapman layer) with the decrease of the col-
loid stability, similarly to other anions. The counterion presence
in the double layer strongly changes the water structure at tita-
nium dioxide interface [39,40]. In Fig. 5 the initial rate evaluation
at pH 5.5 was carried out in the absence of a pH buffer so that the
pH slightly changes with the increase of the melamine concentra-
OH radicals and depresses the rate for substrates that react by
•
direct electron transfer. As melamine does not react with free OH,
this experiment would support the original [7] and still debated
•
hypothesis, that surface fluorination produces free OH in solution
•
and impedes hole trapping as Ti(O )(OH). However an alternative
explanation more consistent with all the above discussion is possi-
ble.
For substrates that chemisorb, fluoride has a beneficial effect as
the back reaction of the adsorbed oxidized substrate with photo-
generated electrons is impeded. By means of photopotential decay
measurements it was demonstrated that TiO2 surface fluorina-
tion retards the reactivity of photogenerated electrons, both for
recombination with surface trapped holes and for transfer to oxy-
gen, upward shifts the electronic levels in the potential energy
scale, changes the mechanism from direct to indirect for strong
adsorbing species, and inhibits the adsorption of intermediates
that could serve as recombination centers [42,43]. Chemisorption
depresses the photocatalytic rate as demonstrated for glycerol for
which different degradation products were observed at large sub-
strate concentration [10]. In the absence of chemical adsorption of
melamine at acidic pH (adsorption by hydrogen bonding was sup-
posed above), no displacement of the substrate from the surface
can be invoked by fluoride ions. In addition, for the purpose of the
−
5
tion, ranging from 5.5 (melamine initial concentration 1 × 10 M)
−
3
to 6.4 (melamine initial concentration 1 × 10 M). The apparent
trivial change in the pH values around the TiO pzc and the concur-
2
rent adsorption of neutral melamine [6] to the surface by hydrogen
bond between lateral amino groups and terminal surface hydrox-
yls cause a shift from the colloid stability regime to a condition
where the primary particles coagulate. The possibility of a strong
hydrogen bonding with surface hydroxyls is supported by the for-
mation of an insoluble complex with cyanuric acid [41] used for
the first time in this work for purification of Compound I, and also
suggested above in the discussion of pH effect on the rate. A rele-
vant consequence of the colloid instability is the dramatic decrease
of the initial degradation rate of almost an order of magnitude, due
mainly to a reduced light harvesting ability of the slurry and pos-
sibly to the increased recombination at the particle surface. The
effect of colloid size and its light absorption was already reported
•
present discussion it must be noted that the surface trap Ti(F)(O )
•
is more shallow than Ti(OH)(O ), and its energy level is more res-
[
31]. In addition, the coagulation of the primary particles generates
onant with free holes in the valence band. Then at least none or
a moderate positive increase of the reaction rate was expected as
for phenol and benzoic acid [44]. The rate decrease must then be
ascribed to the reduced surface accessibility induced by fluoride
adsorption which impedes the adsorption by hydrogen bonding
and a close contact with surficial bridge oxygen on which holes are
shallowly trapped. Then this experiment together with the effect
of pH supports the mechanism of direct hole transfer.
Finally, the melamine degradation rate in the presence of P25
and Merck TiO2 is here compared with the aim to ascertain that
what before discussed is not specific of P25. By a comparative FT-
IR analysis under various conditions on P25 there are a variety of
surface OH, that is at least 3 types of linear hydroxyl groups (in
surface contact points that are active as recombination centers as
demonstrated by Gomez et al. [42]
The addition of fluoride ion to the TiO2 slurry at pH 3 (com-
pare the plots in Fig. 5, and entries 2 and 10 in Table 1) significantly
reduces the rate for the whole range of melamine concentrations. At
−4
C = 1 × 10 M and pH 3, rpristine/rfluorinated = 4. Under the adopted
0
conditions the saturation of the fluoride adsorption was presum-
•
−
ably achieved [7]. The estimated redox potential of the couple F/F
is 3.6 V vs NHE and makes fluoride stable against oxidation by TiO2
valence band holes, so that its role is a merely surface modifica-
tion. After the first paper on surface fluorination [7], many paper
appeared that have been summarized in a recent review [28]. In
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
1
0
sible to suggest that the primary degradation step of melamine is an
outer sphere direct electron transfer. This feature makes melamine
degradation a useful tool to evaluate the direct hole transfer abil-
ity of a photocatalyst. The estimated formal reduction potential of
melamine is quite large. Then it follows that on TiO surface there is
2
the presence of shallow hole traps with redox potential larger than
2
.3 V vs NHE. The surface fluorination decreases the physisorption
of melamine on the surface and then the reaction rate.
An incomplete mineralization was always observed because
the last product of the melamine transformation is the stable and
non-toxic cyanuric acid. The formation of 2,4-diamino-6-nitro-
1
,3,5-triazine (Compound I) as first long living intermediate was
observed, indicating that oxidation proceeds on the amino-group
till to nitro-group through several consecutive fast oxidation steps.
The hydrolysis of Compound I results in the release of nitrite in
solution and initiates processes that gives ammeline, ammelide and
cyanuric acid.
In conclusion, the melamine degradation is a useful tool to eval-
uate the direct hole transfer ability of a photocatalyst. The use of
this peculiar substrate can benefit more fundamental studies on the
photocatalytic mechanism and also various applications, compris-
ing tests to evaluate the activity of titania photocatalysts.
Fig. 6. Initial photocatalytic transformation rate of melamine as a function of the
substrate concentration at pH 5.5 and pH 3 (air, TiO2 Merck 0.5 g dm ). Inset:
logarithmic plot of rates at pH 3 and 5.5.
−
3
which OH is bound to a surface Ti, let say Ti OH), and 3 types of
bridged hydroxyl groups (Ti OH Ti) [11,18]. Fig. 6 shows the ini-
tial transformation rate at pH 5.5 and 3 as a function of the initial
substrate concentration in the presence of TiO2 Merck. The pre-
viously described saturative profile was observed at both pHs for
the whole concentration range presently investigated. In the inset
of Fig. 6 a log/log plot is presented. It is apparent that the found
slope for data at pH 3 and 5.5 has a square root dependence on the
melamine concentration as already observed for P25. Comparing
Acknowledgments
The financial support from project Ricerca Locale—Torino Uni-
versity, the project PHOTORECARB—Progetti di Ateneo/CSP 2012
–
Call 03 – Università di Torino & Compagnia di S. Paolo, and
EU Commission project SETNanoMetro #604577 are gratefully
acknowledged.
the rate of both specimens (Figs. 5 and 6), it is evident that on TiO
References
2
−
5
Merck the rate is markedly lower by a factor 52 (C = 1×10 M) to
0
−
3
[1] E. Pelizzetti, V. Maurino, C. Minero, V. Carlin, E. Pramauro, O. Zerbinati, M.L.
Tosato, Environ. Sci. Technol. 24 (1990) 1559–1565.
2
3 (C = 1 × 10 M).
0
Ratios similar to 1 were observed previously with substrates
[
2] C. Minero, V. Maurino, E. Pelizzetti, Res. Chem. Intermed. 23 (1997) 291–310.
(
(
e.g., phenol) that react predominantly with surface trapped hole
also referred as hydroxyl radical mediated oxidation), while ratios
[3] G. Goutailler, J.-C. Valette, C. Guillard, O. Païssé, R. Faure, J. Photochem.
Photobiol. A: Chem. 141 (2001) 79–84.
[
4] S.M. Arnold, W.J. Hickey, R.F. Harris, Environ. Sci. Technol. 29 (1995)
083–2089.
significantly higher than 1 were observed during the photocat-
alytic transformation of compounds that react predominantly by
direct hole oxidation (hydrogen peroxide, formic acid/formate)
2
[5] G. Zong, W.R. Lusby, M.T. Muldoon, R. Waters, C.J. Hapeman-Somich, J. Agric.
Food Chem. 40 (1992) 2294–2298.
[
6] A. Bozzi, M. Dhananjeyan, I. Guasaquillo, S. Parra, C. Pulgarin, C. Weins, J. Kiwi,
J. Photochem. Photobiol. A: Chem. 162 (2004) 179–185.
[
11]. Considering the different specific surface area between the
two investigated catalysts (SSAP25/SSAMerck = 5) it emerges a con-
siderable intrinsic difference in the ability of the two catalysts to
promote the degradation, because of the different density of sur-
face traps and surface hydroxyls available for hydrogen bonding.
These sites in low coordination conditions are more present on the
P25 surface that on Merck TiO2 [18]. The holes injected at surface
defects (borders, corners, steps. . .) and on (0 0 1) plane possibly
have both higher redox potential and/or faster transfer kinetics.
Then the observed rates agree once again with the direct hole trans-
fer mechanism, as the degradation rate of substrates that react
predominantly by direct hole injection are larger on specimens
with high density of surface defects.
[7] C. Minero, G. Mariella, V. Maurino, E. Pelizzetti, Langmuir 16 (2000)
632–2641.
2
[
[
8] M.A. Henderson, Surf. Sci. Rep. 66 (2011) 185–297.
9] Y. Nosaka, S. Komori, K. Yawata, T. Hirakawa, Y.A. Nosaka, Phys. Chem. Chem.
Phys. 5 (2003) 4731–4739.
[
[
10] C. Minero, A. Bedini, V. Maurino, Appl. Catal. B: Environ. 128 (2012) 135–143.
11] M. Minella, M.G. Faga, V. Maurino, C. Minero, E. Pelizzetti, S. Coluccia, G.
Martra, Langmuir 26 (2010) 2521–2527.
[12] J. Ryu, W. Choi, Environ. Sci. Technol. 42 (2008) 294–300.
[13] C. Minero, G. Mariella, V. Maurino, D. Vione, E. Pelizzetti, Langmuir 16 (2000)
8
964–8972.
14] V. Maurino, C. Minero, G. Mariella, E. Pelizzetti, Chem. Commun. (2005)
627–2629.
[15] H. Park, W. Choi, J. Phys. Chem. B 108 (2004) 4086–4093.
[
2
[
[
16] M. Mrowetz, E. Selli, New J. Chem. 30 (2006) 108–114.
17] P. Piccinini, C. Minero, E. Pelizzetti, M. Vincenti, J. Chem. Soc. Faraday Trans.
93 (1997) 1993–2000.
4
. Conclusions
[18] G. Martra, Appl. Catal. A: Gen. 200 (2000) 275–285.
[
19] S. Chiron, S. Barbati, S. Khanra, B.K. Dutta, M. Minella, C. Minero, V. Maurino, E.
Pelizzetti, D. Vione, Photochem. Photobiol. Sci. 8 (2009) 91–100.
The comparison of different oxidation processes (photocatal-
[
20] P. Calza, C. Minero, E. Pelizzetti, Environ. Sci. Technol. 31 (1997) 2198–2203.
2
−
ysis, Fenton process, sonocatalysis, H O /h␯, S O
that only photocatalysis and the generation of SO4
/h␯) showed
[21] C. Minero, P. Piccinini, P. Calza, E. Pelizzetti, New J. Chem. 20 (1996)
1159–1164.
2
2
2
8
•
−
are able to
[
[
22] E. Pelizzetti, C. Minero, Colloids Surf. A 151 (1999) 329–338.
23] P. Neta, R.E. Hule, A.B. Ross, J. Phys. Chem. Ref. Data 17 (1988) 1027–1284.
transform the melamine. Noticeably, the homogeneous hydroxyl
radicals are unable to turn on the degradation process. This makes
melamine a peculiar substrate, quite different from the previously
tested molecules which react both with free and surface bound
[24] S. Stucki, R. Kötz, B. Carcer, W. Suter, J. Appl. Electrochem. 21 (1991) 99–104.
[
25] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, J. Phys. Chem. Ref. Data
7 (1988) 513–886.
26] C. Minero, M. Lucchiari, V. Maurino, D. Vione, RSC Adv. 3 (2013) 26443–26450.
1
[
•
OH, and with valence band holes. From the detailed analysis of the
[27] N. Serpone, R. Terzian, P. Colarusso, C. Minero, E. Pelizzetti, H. Hidaka, Res.
Chem. Intermed. 18 (1992) 183–202.
melamine degradation rates under different conditions, it was pos-
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
11
[
28] C. Minero, in: D.W. Bahnemann, P.K.J. Robertson (Eds.), Environmental
Photochemistry Part III, The Handbook of Environmental Chemistry, 35,
Springer-Verlag, Berlin Heidelberg, 2013, pp. 23–44, 2015.
[35] Y.H. Jang, S. Hwang, S.B. Chang, J. Ku, D.S. Chung, J. Phys. Chem. A 113 (2009)
13036–13040.
[36] J. Augustynski, Struct. Bond. 69 (1988) 1–61.
[37] A. Fujishima, X. Zhang, D.A. Tryk, Surf. Sci. Rep. 63 (2008) 515–582.
[38] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995)
69–96.
[39] S. Kataoka, M.C. Gurau, F. Albertorio, M.A. Holden, S.-Mi. Lim, R.D. Yang, P.S.
Cremer, Langmuir 20 (2004) 1662–1666.
[
[
[
[
29] C.L. Pang, R. Lindsay, G. Thornton, Chem. Soc. Rev. 37 (2008) 2328–2353.
30] C. Minero, Catal. Today 54 (1999) 205–216.
31] C. Minero, D. Vione, Appl. Catal. B: Environ. 67 (2006) 257–269.
32] C. Minero, V. Maurino, D. Vione, Photocatalysis and Water Purification. From
Fundamentals to Recent Applications, in: P. Pichat (Ed.), Wiley VCH Verlag
GmbH, Weinheim, 2013, pp. 53–72.
[40] B.P. Nelson, R. Candal, R.M. Corn, M.A. Anderson, Langmuir 16 (2000)
6094–6101.
[
[
33] A. Imanishi, T. Okamura, N. Ohashi, R. Nakamura, Y. Nakato, J. Am. Chem. Soc.
1
29 (2007) 11569–11578.
34] P.A. Connor, K.D. Dobson, A.J. McQuillan, Langmuir 15 (1999) 2402–2408.
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012

G Model
APCATA-15639; No. of Pages12
ARTICLE IN PRESS
V. Maurino et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
1
2
[41] A. Ranganathan, V.R. Pedireddi, C.N.R. Rao, J. Am. Chem. Soc. 121 (1999)
1
752–1753.
[
[
42] D. Monllor-Satoca, R. Gómez, J. Phys. Chem. C 112 (2008) 139–147.
43] D. Monllor-Satoca, T. Lana-Villarreal, R. Gómez, Langmuir 27 (2011)
1
5312–15321.
[
[
44] D. Vione, C. Minero, V. Maurino, M.E. Carlotti, T. Picatonotto, E. Pelizzetti,
Appl. Catal. B: Environ. 58 (2005) 81–90.
45] P. Wardman, J. Phys. Chem. Ref. Data 18 (1989) 1637–1755.
Please cite this article in press as: V. Maurino, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.11.012