TiO2/graphene-like photocatalysts for selective oxidation of 3-pyridine-methanol to vitamin B3 under UV/solar simulated radiation in aqueous solution at room conditions: The effect of morphology on catalyst performances

Article abstract of DOI:10.1016/j.apcata.2014.09.002

Graphene-like layers, synthesized through a two-step oxidation/reduction wet treatment of a high surface carbon black, have been used to prepare composites with TiO₂ nanoparticles by liquid phase deposition, followed by calcination at 200 °C. The photocatalytic activity of the TiO₂/graphene-like composites has been tested for the selective conversion of 3-pyridine methanol to 3-pyridine carboxyaldehyde and nicotinic acid (vitamin B3), under de-aerated and UV/solar simulated conditions, in the presence of cupricions. Two different composite morphologies have been explored and a dependence of the photocatalytic activity has been assessed. An enhanced photocatalytic activity, with respect to the neat TiO₂, has been observed and attributed to the broader variety of stable free-radical species generated, at a given photo-catalyst morphology, within the delocalized π-electron systems.

Full text of DOI:10.1016/j.apcata.2014.09.002

Applied Catalysis A: General 487 (2014) 91–99
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
TiO /graphene-like photocatalysts for selective oxidation of
2
3
-pyridine-methanol to vitamin B3 under UV/solar simulated
radiation in aqueous solution at room conditions: The effect of
morphology on catalyst performances
Michela Alfè , Danilo Spasiano , Valentina Gargiulo , Giuseppe Vitiellob,c,
a
b,∗
a
d
b
Roberto Di Capua , Raffaele Marotta
Istituto di Ricerche sulla Combustione, Consiglio Nazionale delle Ricerche (IRC)-CNR, p.le V. Tecchio, 80, 80125 Napoli, Italy
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Università di Napoli “Federico II”, p.le V. Tecchio, 80, 80125 Napoli, Italy
CSGI, Consorzio Interuniversitario per lo Sviluppo dei Sistemi a Grande Interfase, Italy
a
b
c
d
Dipartimento di Fisica, Università di Napoli “Federico II”, and SPIN-CNR UOS, via Cintia, 80126 Napoli, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 May 2014
Received in revised form 26 August 2014
Accepted 2 September 2014
Available online 16 September 2014
Graphene-like layers, synthesized through a two-step oxidation/reduction wet treatment of a high
surface carbon black, have been used to prepare composites with TiO2 nanoparticles by liquid phase depo-
sition, followed by calcination at 200 C. The photocatalytic activity of the TiO2/graphene-like composites
has been tested for the selective conversion of 3-pyridine methanol to 3-pyridine carboxyaldehyde and
nicotinic acid (vitamin B3), under de-aerated and UV/solar simulated conditions, in the presence of cupric
ions. Two different composite morphologies have been explored and a dependence of the photocat-
alytic activity has been assessed. An enhanced photocatalytic activity, with respect to the neat TiO2, has
been observed and attributed to the broader variety of stable free-radical species generated, at a given
photo-catalyst morphology, within the delocalized ␲-electron systems.
◦
Keywords:
Graphene-like
TiO2-composites
Selective photocatalytic oxidation
Vitamin B3
© 2014 Elsevier B.V. All rights reserved.
1
. Introduction
At the present time, there is a great interest to carry out the
oxygen, the HO radical production is strongly limited and the pro-
cess selectivity increases [8,9]. In particular, when Cu(II) ions accept
the photo-generated conduction band electrons, they are reduced
first to Cu(I) and then to Cu(0), which precipitate. Moreover, since
the zero-valent copper could be simply re-oxidized to Cu(II), at the
end of process, with an air flow blown into the solution in dark con-
ditions, it is possible to consider also Cu(II) ions as co-catalysts [10].
However, even if the presence of oxygen is avoided through inert
gas inlet during the photocatalytic process, the formation of HO
radicals is not totally quenched, just resulting from the reaction of
water molecules or surface adsorbed hydroxyl groups with positive
holes [11]:
production of fine chemicals, under very mild conditions, using
renewable energy sources, such as the solar radiation, and eco-
friendly solvents, such as water [1,2]. To this aim the use of
TiO -based photocatalysts appears to be advantageous as TiO2 is
easily available, stable to photocorrosion, cheap, and nontoxic sub-
stance [3–7].
2
When molecular oxygen accepts the TiO photo-generated elec-
2
trons in aqueous solution, it leads to the formation of HO radicals,
which play highly negative role in these processes since they un-
selectively oxidize organic substrates:
+
◦
+
TiO (h ) + H O
→ TiO + HO + H
2
2
(ads)
2
TiO (e ) + O2 → O•2−. . . → HO•
−
vb
2
cb
On the other hand, when metal ions, such as Cu(II) ions, are
used in de-aerated conditions as electron acceptor instead of
TiO2(h ) + H2O−
+
→ TiO2 + HO◦
vb
(ads)
The produced hydroxyl radicals, attack the organic substrates
and desired products, thus lowering the selectivity of the processes
[8,10]. Moreover, one of the main problems associated with the use
∗ Corresponding author. Tel.:+39 0817682253; fax: +39 0815936936.
E-mail address: danilo.spasiano@unina.it (D. Spasiano).
http://dx.doi.org/10.1016/j.apcata.2014.09.002
0
926-860X/© 2014 Elsevier B.V. All rights reserved.

9
2
M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
◦
of TiO2 as a photocatalyst is the fast recombination reaction of the
electron–hole pair:
distilled water, vacuum dried for 1 h and then calcined at 200 C for
1 h.
+
−
TiO (h ) + TiO (e ) → heat
2.3. GL layers synthesis
2
2
vb
cb
For this reason, many researchers are pointing their atten-
GL layers in water suspension were obtained through a two
steps oxidation/reduction method [24] starting from a high surface
tion to the synthesis of different TiO -based catalysts containing
2
inorganic and organic compounds, such as metal oxides, noble
metals, graphene and copper sulphide [11–13], that decrease the
electron–hole recombination. For example, it has been proven
that the electron accepting and transport properties of graphene
provide a convenient way to direct the flow of photo-generated
charge carriers, which thus increases the lifetime of electron–hole
pairs generated by TiO2 upon light irradiation [14,15].
Direct TiO2 growth on graphite oxide (GO) sheets followed by
reduction of GO to graphene [16–20] has been proposed as possible
strategy to produce composites that combine the electron accept-
ing features of graphene and the TiO2 photocatalytic activity. The
use of the hydrophilic GO as intermediate step to obtain graphene
is a convenient approach adopted to overcome the graphene aggre-
gation driven by strong van der Waals forces when used in water
and in polar organic solvents [14,21].
carbon black (CB). Briefly, 500 mg of CB powder (15–20 nm primary
particles diameter, specific BET area 139 m2/g) was oxidized with
10 mL of nitric acid (67 wt.%) at 100 ◦C under stirring for 90 h. The
oxidized carbon nanoparticles were recovered by centrifugation
and washed until acid traces were successfully removed. Follow-
ing the oxidation step, the nanoparticles (20 mg) were dispersed in
20 mL distilled water and treated with 450 ␮L of hydrazine hydrate
(50%) at 100 ◦C under reflux for 24 h.
At the end of the reaction the excess of hydrazine was neutral-
ized with nitric acid (4 M) and the resulting black solid recovered
by centrifugation (3000 rpm, 30 min). The solid was washed with
distilled water, recovered by centrifugation three times in order
to remove traces of unreacted reagents and acid and named
GL. The pH of the GL water suspension was 3.70. The dried GL
resulted to be insoluble in water and in the most common organic
solvents, both polar and apolar (water, ethanol, N-methyl pirrolidi-
none, dichloromethane, heptane, dimethylformamide (DMF) [23]).
It was demonstrated by transmission electron microscopy (HRTEM)
and atomic force microscopy (AFM) that the water-insoluble GL
undergo to self-assembling in thin film on surfaces after drying
[24]. The self-assembled film is made by pliant sheets that easily
conform to any feature of that surface. The formation of flat film is
associated with the decrease in the polar functionalities in GL lay-
ers as a consequence of the reduction step and subsequent intimate
self-assembling interaction between the restored graphitic layers,
as previously observed for the reduced graphite oxide [22].
In the present investigation, direct TiO growth on water-stable
2
conductive graphene-like (GL) layers to produce TiO -based com-
2
posite photocatalytic materials has been proposed. The GL layers,
prepared in the present investigation, have a good stability in
aqueous solutions and, undergo to self-assembling in flat blocks
if dried on a flat surface thanks to the instauration of hydrophobic
interactions between the graphenic layers. This behavior, typically
observed in reduced graphite oxide [16], allowed to investigate dif-
ferent morphological arrangement of TiO /GL composites with the
2
aim of studying the relationship between morphology and pho-
tocatalytic activity for a cleaner chemical production of vitamin
B3.
Two strategies were adopted to produce the TiO -based com-
2
The photocatalytic activities of the composites were tested by
means of the selective oxidation of 3-pyridine methanol to 3-
pyridine carboxyaldehyde and nicotinic acid, under de-aerated and
UV/solar simulated conditions, in presence of cupric ions in aque-
ous solution at ambient temperature. Nicotinic acid (also known as
niacin, vitamin PP or B3) is an important vitamin in B group and
it is considered as one of the 80 essential human nutrients, largely
used for the prevention and treatment of pellagra disease [23].
posites, differing from the morphology of the GL layers: (1) freshly
prepared GL layers in water suspension (GLW) were readily used;
(2) GL layers (10 mg/mL) were allowed to dry at room temperature
on a flat surface (glass surface). The resulted self-assembled prod-
uct was scraped from the glass surface obtaining flat platelets of
assembled GL layers (GL platelets, GLP).
2.4. Synthesis of TiO /GLW and TiO /GLP composites
2
2
Two composites were prepared by directly growing TiO2 on
GLW and GLP. TiO /GLW and TiO /GLP composites were prepared
2
. Experimental
2
2
adapting the strategy reported by Jiang et al. for the prepara-
tion graphene oxide/TiO2 composites [16]: 2.5 g of (NH ) TiF and
2.1. Materials
4
2
6
2
.35 g of H BO3 were dissolved in 125 mL of the GL (or GLP) aque-
3
Cupric ions were introduced in the system as cupric sulphate
◦
ous suspension (75 mg/L) and stirred for 2 h at 60 C. After cooling
a grey solid was recovered by filtration (Anodisc, pore size 0.2 ␮m),
washed with distilled water, vacuum dried for 1 h and then calcined
pentahydrate, (CuSO ·5H O, ACS grade). 3-Pyridine methanol
4
2
(
(
3-PMA), 3-pyridine carboxyaldehyde (3-PCA), nicotinic acid
NA), copper sulphate, acetonitrile, ammonium acetate, hydrazine
◦
at 200 C for 1 h. The expected GL layers percentage is 5 wt.%
monohydrate (50 wt.%), thionin acetate (THA) salt (powder),
ammonium hexafluorotitanate ((NH ) TiF ), boric acid, nitric acid
This approach aims to develop two composites with different
4
2
6
morphology, as illustrated in Fig. 1. In the case of TiO /GLW, the
GL and the TiO2 precursors are well mixed in the reactive mixture
allowing a good dispersion of GL into TiO . In the case of TiO /GLP,
the GL layers are assembled in flat blocks offering a flat surface for
the growing of TiO2 crystal.
2
(
(
67 wt.%), and perchloric acid, were purchased from Sigma–Aldrich
ACS grade) and used as received. Carbon black (CB) type N110, pro-
duced with the furnace process, was obtained by Sid Richardson
Carbon Co.
2
2
2.2. Preparation of neat TiO2
2.5. Characterization procedure
Neat TiO2 (in anatase form) was prepared according to the pro-
The thermal stability of the prepared samples was evaluated by
using thermogravimetric analysis (TGA) performed on a Perkin-
Elmer Pyris 1 Thermogravimetric Analyzer. The materials were
cedure reported by Jiang et al. [16]. 2.5 g of (NH ) TiF and 2.35 g
4
2
6
of H BO were dissolved in 125 mL of distilled water, sealed and
3
3
◦
−1 ◦
heated in an oxidative environment (air, 30 mL min ) from 50 C
stirred for 2 h at 60 C. After cooling, the white solid was recov-
◦
◦
−1
ered through filtration (Anodisc, pore size 0.2 ␮m), washed with
up to 750 C at a rate of 10 C min .

M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
93
(0.1 ␮g/mL) onto freshly cleaved mica substrates (grade V-1, Elec-
tron Microscopy Sciences), which were then allowed to dry in air.
2.6. Photocatalytic reactor equipment and analytical procedure
The photocatalytic runs were performed, in duplicate, in an
annular batch glass reactor (V 0.280 L, external diameter 6.5 cm,
height 40 cm and optical path 1.1 cm), equipped with a mercury
vapor high pressure lamp (nominal power 125 W, Helios Italquartz)
mainly emitting at 305, 313 and 366 nm (manufacturer data). The
photon flows of the lamp at 305, 313 and 366 nm, determined
Fig. 1. Representation of the hypothesized morphologies of TiO2/GLW and TiO2/GLP
composites.
−
5
−5
through actinometric analysis, were 4.85 × 10 , 8.34 × 10 and
The Brunauer–Emmet–Teller (BET) specific surface areas of
the samples were measured by N2 adsorption at 77 K using a
Quantachrome Autosorb 1-C. The samples were outgassed under
−
4
2
2
.71 × 10 E/min, respectively. The reactor was thermostated at
◦
5 C by using a thermostat bath (Julabo ED-13) and stirred mag-
netically. The apparatus device has been reported elsewhere [28].
Before turning on the lamp, the aqueous solution, con-
taining the organic substrate ([3-PMA]o = 1.5 mM), cupric
◦
vacuum at 110 C before the analysis. Adsorption/desorption data
were processed in accordance with the Barrett–Joyner–Halenda
(
BJH) model.
salts ([CuSO ]o = 1.5 mM) and
a
proper load of catalyst
4
X-ray diffraction (XRD) analysis was carried out using a Philips
(
TiO = 570 ␮g/mL, TiO2 composites = 600 ␮g/mL) was purged
2
PW1710 diffractometer operating between 5 and 80 2ꢀ with a Cu
K␣ radiation. A diffraction experiment was run on standard glass
slide for the background correction.
The determination of surface acidic functionalities was per-
formed according to the fluorimetric test reported by Visentin et al.
with a nitrogen stream for 30 min at dark (flow rate = 0.3 L/min).
The different catalyst load has been chosen to account the effective
TiO2 content in TiO2 composites. (30 ␮g/mL being the GL amount,
5
% of 600 ␮g/mL).
The photocatalytic runs were carried out under inert atmo-
[
25]. The carbonaceous material (0.01 mg) was treated with 1.5 mL
sphere to prevent the inlet of oxygen and the external walls of
the reactor were covered with an aluminum foil to prevent the
dispersion of the light emitted by the lamp.
The pH was adjusted at the desired value (pH 3.0) with perchlo-
ric acid (pH-meter Orion 420A+ by Thermo).
−
6
of a solution of THA in ethanol (4.3 × 10 M) for 30 min under con-
tinuous stirring at room temperature. The suspension was filtered
on a 20 nm filter unit (Anotop 25, Whatman) and the fluorescence
emission was measured with a Perkin-Elmer LS 50 B spectrofluo-
rimeter (scan speed 100 nm/min, emission recorded in the range
between 570 and 700 nm upon excitation at 594 nm).
The samples, taken from the reactor at different reaction
times, were filtered on regenerated cellulose filters (0.45 ␮m,
Teknokroma) and analyzed by high performance liquid chromatog-
raphy (HPLC) using an Agilent 1100 instrument equipped with a
diode array UV–vis lamp (265 nm) and a Synergy 4u MAX-RP 80A
Scanning Electron Microscopy (SEM) was performed on a FEI
TM
Inspect S50 Scanning Electron Microscope. Analysis was realized
on a powder sample previously dried and sputter coated with a thin
layer of gold to avoid charging.
◦
column (flow rate 1.0 mL/min, 30 C). The adopted gradient elu-
Electron Paramagnetic Resonance (EPR) Spectroscopy was
carried out by means of X-band (9 GHz) Bruker Elexys E-500
spectrometer (Bruker, Rheinstetten, Germany), equipped with a
super-high sensitivity probe head. Solid samples were transferred
to flame-sealed glass capillaries which, in turn, were coaxially
inserted in a standard 4 mm quartz sample tube. Measurements
were performed at room temperature. The instrumental settings
were as follows: sweep width, 100 G; resolution, 1024 points;
modulation frequency, 100 kHz; modulation amplitude, 1.0 G. The
amplitude of the field modulation was preventively checked to be
low enough to avoid detectable signal overmodulation. Prelimi-
narily, EPR spectra were measured with a microwave power of
tion was 95% of buffer solution (ammonium acetate 20 mM) and
5
% of acetonitrile for 5 min, rise to 75% of buffer solution and 25% of
acetonitrile in 10 min. The concentration of dissolved copper, was
measured by means of a colorimetric method with an analytical kit
(based on oxalic acid bis-cyclohexylidenehydrazide, cuprizone®)
purchased from Macherey-Nagel. A UV–vis spectrometer (Unicam)
was used to measure the dissolved copper at a wavelength of
5
85 nm. A detection limit of 0.1 mg/L was found during this inves-
tigation.
3. Results and discussion
∼
0.6 mW to avoid microwave saturation of resonance absorption
curve. Several scans, typically 16, were accumulated to improve
the signal-to-noise ratio. Successively, for power saturation experi-
ments, the microwave power was gradually incremented from 0.02
to 160 mW. The g-factor value was evaluated by means an internal
standard (TEMPOL ethanol solution) [26] which was inserted in the
quartz sample tube co-axially with the capillary containing the GL
in water suspension or composite samples. Free radical concen-
tration in the sample was estimated by using a TEMPOL ethanol
solution as reference, whose concentration was determined by UV
spectroscopy [27]. The area under the EPR absorption curves was
estimated by double integration of their first derivatives.
3.1. GLW and GLP characterization
GL layers in water suspension (GLW) consist of small flat
nanoparticles decorated at the edge with oxygen-containing
groups (mainly carboxylic/carbonylic and hydrazones) with
graphenic basal plane untouched [23]. The carboxylic functional
groups, residual from the reduction treatment of strongly oxi-
dized CB, have been quantified by a sensitive and fast fluorimetric
method [25]. The amount of acid groups was estimated to be
−
8
9.4 × 10 molCOOH/mg. The presence of residual carboxylic func-
tional groups allows a good stability and dispersion of GL sample
in aqueous solution.
Non-contact Atomic Force Microscopy (NC-AFM) was employed
to characterize the GL building-blocks. NC-AFM morphological
characterization was carried out by means of an XE100 Park instru-
ment operating in non-contact mode (amplitude modulation, sili-
con nitride cantilever from Nanosensor) at room temperature and
in ambient conditions. Samples for NC-AFM imaging were prepared
by drop-casting a very low-concentrated GL water-suspension
When GL layers in water suspension were allowed to dry on
a flat surface, they underwent to self-assembling forming a flat
film whose height depends upon the concentration of the starting
solution. NC-AFM morphological characterization of ultrathin
films (<20 nm) obtained from GL drying has been reported else-
where [24], providing a surface characterization and roughness

9
4
M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
Fig. 2. AFM characterization: (a) NC-AFM topographic image, 4 ␮m × 4 ␮m, showing the presence of GL nanoparticles on a mica substrate; (b) NC-AFM topographic image,
1
.5 ␮m × 1.5 ␮m, zoomed on some representative nanoparticles, reporting the analyzed profiles positions; (c) height profiles on GL nanoparticles, taken along the lines
highlighted in panel (b).
estimation. Here, the observation of the GL layers, which are at
the basis of the GLP formation, is reported. The freshly cleaved
mica used as substrate guarantees an atomically flat underlying
surface, which is needed for the correct observation and size
measurements of particles on a nanometer scale. In addition, a
very low-concentrated water-suspension (0.1 ␮g/mL) was used
for drop-casting onto mica substrate. Such a low value was chosen
in order to prevent the formation of aggregates on the substrate,
allowing instead the achievement of rather individual GL layers.
The concentration of water suspension, however, was high enough
to allow the presence of a certain number of nanoparticles per unit
area of the substrate, so that to be easily detectable in the scanning
area of the microscope.
interactions favor the nucleation of TiO2 nanoparticles on GLW
and GLP surface.
3.2. TiO2 and composites characterization
3.2.1. XRD analysis
The X-ray diffraction patterns of the prepared samples are
reported in the 20–80 2ꢀ range (Fig. 4). The diffraction pattern of
neat TiO2 confirms that the sample is present in the anatase form
(JCPDS 21-1272). The XRD path of both composites showed that the
production of TiO in the anatase form is not influenced by the addi-
2
tion of GL layers. Nevertheless, low intensity signals attributed to
rutile contamination (as 2ꢀ = 27.5, (1 1 0) diffraction peak) in TiO2-
GLW XRD pattern were also detected. There were no observable
peaks of the GL layers in both the composite samples, possibly due
both to the relatively low content of carbonaceous material in the
composites and to their low intensity, compared to the prevailing
TiO2 pattern. The peak at 2ꢀ = 25.2 (1 0 1 diffraction peak) for the
TiO /GLP sample is higher with respect to the neat TiO , whereas
Fig. 2 shows the presence of nanoparticles deposited on the mica
flat surface. A detailed characterization of such particles reveals a
size distribution, which can be ascribed to different aggregation
amounts of elementary constituents. Selected height profiles cross-
ing the particles demonstrate first of all, the flatness of the substrate
(
the measured roughness, of the order of 1 Angstrom, is undistin-
guishable from the noise level of the instrument, and anyway well
below the height of the particles under study), and its suitability for
a reliable characterization of the nanoparticles. From such profiles
we observed vertical sizes ranging from about 1 nm or less to few
nanometers; the larger lateral dimensions, a few tens of nanome-
ters (about 50 nm), confirms that at the chosen concentration of
water suspension some aggregations are still occurring.
2
2
the TiO /GLW exhibits a lower intensity peak. This result indicates
2
a higher crystallinity of TiO /GLP. It can be speculated that the
2
presence of well dispersed GL layers into the TiO /GLW composite
2
interferes with the TiO2 crystallization process whereas GLP offers
a flat surface advantageous for TiO2 growing and crystallization.
When a concentrated GL water suspension (10 mg/mL) was
allowed to dry on a flat surface, a carbonaceous film is formed. The
production of flat blocks about 1–2 ␮m thick (Fig. 3) was achieved
by scraping the carbonaceous film. These flat blocks, here named
GL platelets (GLP), offer a larger surface for TiO2 growing.
3.2.2. SEM analysis
The morphology of the TiO /GLW and TiO /GLP composites was
2
2
observed with SEM (Fig. 5). Neat TiO2 is also reported for compar-
ison. The dimensions of the composites range in the micrometer
scale: the granulometric analysis performed on the samples shows
that the particle size distribution is lower than 20 ␮m which allows
During the hydrolysis of (NH ) TiF , used as TiO precur-
4
2
6
2
sor, titanium hydroxide complex ions were gradually produced
and further form TiO2 nanoparticles [16]. The residual oxygen-
containing groups on the GLW and GLP surface, not present on
the raw CB surface and introduced after the oxidation-reduction
treatment, can interact with titanium hydroxide complex ions and
the growing TiO2 nanoparticles through hydrogen bonds. These
an easy separation of the composites by filtration. Neat TiO anatase
2
exhibits rather spherical primary particles which aggregate form-
ing larger particles. The morphology of the composites is similar
to that exhibited by neat TiO2 and no significant differences are
observable.

M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
95
Table 1
EPR spectral parameters for TiO2/GLW, TiO2/GLP and neat GLP powders.
GLP
TiO2/GLP
TiO2/GLW
g-factor
B (G)
Spin × g
2.0040 ± 0.0004
3.3 ± 0.2
2.0035 ± 0.0004
2.9 ± 0.2
2.0037 ± 0.0004
1.6 ± 0.2
ꢁ
−
1
14
13
14
1.4 × 10
8.5 × 10
4.9 × 10
3.2.3. BET analysis
N2 adsorption/desorption measurements with BET method
were carried out to determine the surface area of the samples.
2
TiO /GLP composite had a surface area of 106.9 m /g, higher than
2
the values measured for the neat TiO2 and GLP that are 85 and
2
7
6 m /g, respectively. The increase of the surface area is not
deducible by a weighted average of the two neat components (GLP
and TiO ) indicating the occurrence of non-additive effects due
2
to the incorporation of the GLP component. A higher surface area
2
was exhibited by TiO /GLW (138 m /g) and was interpreted as a
2
consequence of a more defective and irregular surface due to the
interference of GL in the TiO crystallization process (Figs. 2 and 3).
2
3.2.4. TGA analysis
Fig. 6 shows the TGA curves in oxidative environment (air)
for the TiO /GLW and TiO /GLP composites. Neat GLP was also
2
2
reported for comparison. The GLP sample exhibits two weight
◦
losses, the first one between 150 and 500 C (∼40% of total weight
loss) due to the progressive decomposition of residual oxygen-
containing groups (carbonylic/carboxylic), and the second one at
◦
5
50 C due to the bulk oxidation of GLP. Interestingly, the curve for
the TiO /GLW composite does not exhibit the clear steps of mass
2
◦
loss at about 550 C as TiO /GLP suggesting that the GL layers are
2
well-mixed within the composite and stabilized by the deposited
TiO2 through the interaction between oxygen-containing groups
on the GL and TiO2 particles. According to the TGA curves, in both
composites the weight content of GL layers was roughly evaluated
to be 3–4 wt.%, slightly lower than the expected percentage (5 wt.%).
3.2.5. EPR analysis
EPR measurements on TiO /GLW and TiO /GLP composites were
2
2
performed by following an experimental procedure recently used
in the characterization of melanins [29–31]. Analysis of this EPR
signal provides significant information on the nature of the para-
magnetic centers and on the supramolecular properties of solid
materials. In this study, we undertook an EPR characterization of
powdered samples synthesized as previously described. Neat TiO2
and GLP were used as references. All EPR spectra are reported in
Fig. 7, whereas the spectral parameters calculated from them are
summarized in Table 1.
Fig. 3. SEM images of GLP.
By observing Fig. 7, no signal was detected for neat TiO2 (spec-
trum A), indicating no oxygen vacancy or Ti3+ in this powder which
is consistent with a previous report that indicates no EPR signal for
pure TiO [32]. In Fig. 7, the spectra B, C and D represent the spectra
2
of neat GLP, TiO /GLP and TiO /GLW respectively. They present a
2
2
similar lineshape: a single, roughly symmetric signal at a g value
of ∼2.0035–2.0040, (Table 1), typical of carbon-centered radicals.
It is in agreement with the literature [33,34] and corresponds to
the samples with a high content of carbon which have more para-
magnetic centers. In particular, Nakamura et al. attributed the EPR
signal at this position to an oxygen vacancy [35], whereas no other
detectable signal, such as Ti3+ or O , appears in our tests. Since the
signals of the two composites (Fig. 7, spectra C and D) present the
same lineshape observed for the GLP (Fig. 7, spectrum B), the emer-
gence of the oxygen vacancy in the hybrid TiO2 samples is directly
dependent on the presence of GL layers in the composite synthesis,
in according to the literature findings [34].
2−
Fig. 4. XRD patterns of TiO2/GLW and TiO2/GLP composites and neat TiO2.

9
6
M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
Fig. 5. SEM images for TiO2/GLW and TiO2/GLP composites and neat TiO2.
However, a deeper analysis of the spectra of composites shows
that the TiO /GLP spectrum is evidently broader than the TiO /GLW
␲-electron systems of the TiO /GLP. The reduced “mobility” of free
radicals, generated under irradiation, is interpreted as an interrup-
2
2
2
one. The difference in lineshapes was corroborated by the quanti-
tative determination of the signal amplitude (ꢁB). The ꢁB value
tion of the electron “mobility” within the graphene-like, dispersed
over the TiO2 matrix but not grouped in platelet as in the case
determined from the spectrum of TiO /GLP is significantly lower
of TiO /GLP [34]. This property, added to the presence of oxy-
2
2
than those derived from the spectrum of TiO /GLP which is com-
gen vacancies, probably contributes to improve the photocatalytic
activity of the catalyst, as also suggested in literature [33]. Differ-
ent studies have proposed that the defects in the material structure
have a significant effect on the electronic states of the entire flake of
graphene. These defects can increase the electrons mobility which
is directly correlated to the material conductivity [36]. Finally, a
2
parable to GLP spectrum (Table 1). To ascertain physical reasons for
this experimental evidence, the EPR spectra of the same samples
were acquired setting a higher microwave power. An inspec-
tion of the normalized power saturation profiles, reported in
Fig. 8, shows that, as the microwave power is increased, the spin
density increases. Also, the power saturation curve of TiO /GLP
2
(
and similarly of pure GLP) does not go to a plateau, in contrast
to that observed in the case of TiO /GLW. The comparatively low
2
ꢁ
B value and the power saturation profile of TiO /GLW would
2
suggest a relatively homogeneous free-radical species distribution
that is spatially confined within restricted region of the catalyst
supramolecular structure, in contrast to the broader variety of
free-radical species that could be generated within the delocalized
Fig. 7. EPR spectra of neat TiO2 (A), neat GLP (B), TiO2/GLP (C) and TiO2/GLW (D)
Fig. 6. TGA curves of TiO2/GLW and TiO2/GLP composites and GLP.
composites.

M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
97
quantitative analysis of the spectra intensities shows that all sam-
ples present a similar high spin density (Table 1).
3.3. Composites photocatalytic activity
Composites were tested to check their photocatalytic activity for
the partial oxidation of 3-pyridine methanol (3-PMA) to 3-pyridine
carboxyaldehyde (3-PCA) and nicotinic acid (NA) under illuminated
and deaerated conditions in the presence of cupric ions, used as
photoelectron scavengers.
The normalized concentration profiles for 3-PMA and dissolved
copper, and the reaction yields for 3-PCA and NA, under differ-
ent experimental adopted conditions, are reported in Figs. 9a–d
respectively.
The homogenous system, in the only presence of CuSO4 (empty
triangles), shows a low reactivity probably ascribed to a photocat-
alytic activity of the Cu(II)/3-PMA complexes. The addition of the
sole GL or GLP to the aqueous solution, containing cupric ions and
3
-PMA, does not improve the system reactivity with respect the
previous one (empty squares and diamonds).
A marked increase of 3-PMA and Cu(II) consumption, and 3-PCA
and NA production rates was achieved in presence of cupric ions
with neat TiO2 catalyst at a load of 570 ␮g/mL (full squares).
Fig. 8. Influence of microwave power on amplitudes of free radicals of GLP (ꢀ),
TiO2/GLP (᭹) and TiO2/GLW (ꢁ) composites.
Fig. 9. Selective photo-oxidation of 3-PMA at different experimental conditions under deaerated conditions: (a) 3-PMA consumption; (b) Cu(II) photoreduction; (c) 3-PCA
◦
yield; (d) NA yield. [3-PMA]o = 1.5 mM, [CuSO4]o = 1.5 mM; pH 3.0; T = 25 C. In presence of TiO2/GLP, load 600 ppm (᭹); neat TiO2, load 570 ppm (ꢁ); TiO2/GLW, load 600 ppm
(ꢂ); GLP only, load 50 ppm (ꢀ); GLW only, load 50 ppm (♦); CuSO4 only (ꢀ).

9
8
M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
Table 2
Photocatalytic selective oxidation of 3-PMA in presence of CuSO4 and different solid samples (neat TiO2, neat GL, neat GLP and composite catalysts) for a reaction time of 5 h.
Conversion (X), yield (Y), selectivity (S).
Run
Catalyst
X(3-PMA) (%)
Y(3-PCA) (%)
S(3-PCA) (%)
Y(NA) (%)
S(NA) (%)
Carbon mass
balance (%)
1
2
3
4
5
6
Only CuSO4
17.3
12.4
8.44
50.2
35.8
63.3
13.4
9.14
4.08
30.7
24.1
39.5
77.4
73.7
48.3
61.1
67.3
62.4
2.23
1.63
1.21
14.3
7.28
21.9
12.8
13.1
14.3
28.5
20.3
34.6
98.0
98.2
96.3
94.6
94.4
97.9
CuSO4 and neat GL
CuSO4 and neat GLP
CuSO4 and neat TiO2
CuSO4 andTiO2/GLW
CuSO4 and TiO2/GLP
◦
[
3-PMA]o = 1.5 mM, [CuSO4]o = 1.5 mM; pH 3.0; T = 25 C.
The composites, present in the reacting system at a load of
00 ␮g/mL to account the effective TiO2 content (95 wt.%), show
very different catalytic activities depending on the morphology of
GL layers (GLW or GLP) added to the system.
It has been reported that the photocatalytic activity of TiO2 is
strongly affected by several parameters such as the phase structure,
crystal size, specific surface area and crystallinity [37,38]. In order
to achieve better photocatalytic performances, a good crystallinity
is often required to reduce the formation of electron traps, which
might affect the photocatalytic efficiency [39]. Moreover, a catalyst
with a large specific surface area may be also desirable since the
contact between the active sites and the reacting substances is
enhanced.
6
When TiO /GLW was used as catalyst (full diamonds), the reac-
2
tivity is even lower than that in which neat TiO2 is present. On the
other hand, the photocatalytic oxidation rates of 3-PMA, the cupric
ions reduction rates and the yields to 3-PCA and NA were rather
increased in the presence of TiO /GLP (full particles) at the same
2
adopted experimental conditions.
The lower photocatalytic activity of TiO /GLW sample was
2
The collected results indicated that, under the adopted exper-
imental conditions, the production of NA for photocatalytic
oxidation of 3-PMA, proceeds through the following steps:
driven by its lower crystallinity degree than TiO /GLP one, in
spite of its slightly larger specific surface area (138 m /g). In the
2
2
case of TiO /GLP, the improved photo-catalyst performances are
2
mainly ascribed both to the higher crystallinity degree and to the
presence the broader variety of “long-live” free-radical species
photo-generated within the delocalized GLP ␲-electron systems, as
indicated by EPR analyses. Moreover the lower free-radical density,
1
) photogeneration of holes and electrons
hv
−
+
TiO based catalyst −→ TiO based catalyst(e , h
)
2
2
cb
vb
evidenced in TiO /GLP, can account for the decrease of the occur-
2
rence of oxidation mechanisms of HO radical type that, producing
hydroxylated undesired by-products, inevitably would lower the
selectivity of the system and yields to desired products (3-PCA and
NA).
2
) formation of 3-PCA, as primary chemical intermediate, by reac-
tion of holes with 3-PMA
+
TiO based catalyst(2h ) + 3-PMA → TiO based catalyst
2
vb
2
+
3-PCA
4. Conclusion
3
) conversion of 3-PCA to NA
Water stable GL layers, produced through a two-step oxida-
tion/reduction wet treatment of carbon black, have been used in
two different forms (in water suspension and assembled in flat
blocks) to prepare composites with TiO2 nanoparticles by liq-
+
TiO based catalyst(2h ) + 3-PCA → TiO based catalyst + NA
2
vb
2
uid phase deposition (TiO /GLW and TiO /GLP, respectively). The
4
) photoreduction of cupric ions to copper zero-valent
2
2
adopted experimental approach allowed to explore the effect of
morphology on catalyst performances toward selective conversion
of 3-pyridine methanol to 3-pyridine carboxyaldehyde and nico-
tinic acid (vitamin B3), under de-aerated and UV/solar simulated
conditions, in presence of cupric ions. An enhanced photocat-
alytic activity of TiO /GLP, with respect to the neat TiO , has been
−
TiO based catalyst(2e ) + Cu(II) → TiO based catalyst + Cu(0)
2
2
cb
A summary of the conversion degrees for 3-PMA and selectivi-
2
2
ties and yields to 3-PCA and NA, along with the carbon mass balance,
after 5 h of reaction time, is reported in Table 2. First, the presence
of GL layers in water suspension or GLP only (runs 2 and 3) does
not provide any improvement compared to the homogeneous sys-
tem (run 1). On the contrary, the addition of neat TiO2 only (run
observed and attributed both to the broader variety of free-radical
species stabilized within the delocalized ␲-electron systems of the
TiO /GLP, as indicated by EPR analyses and to a reduction of the
2
hydroxyl radicals formation. In the case of TiO /GLW, the presence
2
of well dispersed GL layers during TiO2 nucleation seems to inter-
fere with the TiO2 crystallization process lowering the composite
photocatalytic activity, in spite of its higher surface area. Moreover,
beneficial effects driven by the presence of highly p-conjugated
systems are not observed due to the spatially confined free-radical
4
) involves an increase of the conversion degree up to 50% and
yields to 3-PCA and NA up to 31% and 14%. Whereas the overall
selectivity (87.3%) and the carbon mass balance (94.4%) are almost
the same, the use of TiO /GLW as photo-catalyst (run 5) involves a
reduction of the conversion degree (35.8%) and yield (24% and 7.3)
if compared to the use of neat TiO2.
2
species in TiO /GLW.
2
On the contrary, the use of TiO /GLP as photo-catalyst (run 6)
2
Acknowledgments
results in improved of the 3-PMA conversion degree (63.3%) and
yield (39.5% for 3-PCA and 22% for NA). Moreover, higher values
for the global selectivity (97%) and the mass carbon balance (98%)
Authors thank Mr. Luciano Cortese for SEM imaging and XRD
analysis and Dr. Paola Giudicianni for BET analyses. The work was
partially financed by Accordo CNR-MSE “Utilizzo pulito dei com-
bustibili fossili ai fini del risparmio energetico” 2011-2012.
are achieved. The TiO /GLP catalyst, compared to neat TiO2 and
2
TiO /GLW catalyst, promotes the selective oxidation of 3-PMA to
2
3
-PCA and of 3-PCA to NA.

M. Alfè et al. / Applied Catalysis A: General 487 (2014) 91–99
99
References
[21] S. Stankovich, R.D. Piner, X. Chen, N. Wu, S.T. Nguyen, R.S. Ruoff, J. Mater. Chem.
6 (2006) 155–158.
1
[
22] S. Stankovich, D.A. Dikin, R.D. Piner, K.A. Kohlhaas, A. Kleinhammens, Y. Jia, Y.
Wu, S.T. Nguyen, R.S. Ruoff, Carbon 45 (2007) 1558–1565.
23] R. Chuck, Appl. Catal. A: Gen. 280 (2005) 75–82.
24] M. Alfè, V. Gargiulo, R. Di Capua, F. Chiarella, J.N. Rouzaud, A. Vergara, A. Ciajolo,
ACS Appl. Mater. Interfaces 4 (2012) 4491–4498.
25] S. Visentin, N. Barbero, S. Musso, V. Mussi, C. Biale, R. Ploeger, G. Viscardi, Chem.
Commun. 46 (2010) 1443–1445.
26] M.F. Ottaviani, J. Phys. Chem. 91 (1987) 779–784.
[
[
[
1] C. Walling, R.W.R. Humphreys, J. Org. Chem. 46 (1981) 1260–1263.
2] C. Parmeggiani, F. Cardona, Green Chem. 14 (2012) 547–564.
3] M. Addamo, V. Augugliaro, M. Bellardita, A. Di Paola, V. Loddo, G. Palmisano, L.
Palmisano, S. Yurdakal, Catal. Lett. 126 (2008) 58–62.
[
[
[
[
[
[
[
[
4] V. Augugliaro, T. Caronna, V. Loddo, G. Marcì, G. Palmisano, L. Palmisano, S.
Yurdakal, Chem. Eur. J. 14 (2008) 4640–4646.
5] S. Yurdakal, G. Palmisano, V. Loddo, V. Augugliaro, L. Palmisano, J. Am. Chem.
Soc. 130 (2008) 1568–1569.
6] G. Palmisano, E. Garcıa-Lopez, G. Marcì, V. Loddo, S. Yurdakal, V. Augugliaro, L.
Palmisano, Chem. Commun. 46 (2010) 7074–7089.
7] S. Yurdakal, V. Loddo, G. Palmisano, V. Augugliaro, H. Berber, L. Palmisano, Ind.
Eng. Chem. Res. 49 (2010) 6699–6708.
[
[
[
27] N.D. Yordanov, K. Ranguelova, Spectrochim. Acta Part
73–378.
A 56 (2000)
3
[
[
28] R. Andreozzi, V. Caprio, A. Insola, R. Marotta, Water Res. 34 (2000) 463-472.
29] L. Panzella, G. Gentile, G. D’Errico, N.F. Della Vecchia, M.E. Errico, A. Napolitano,
C. Carfagna, M. d’Ischia, Angew. Chem. Int. Ed. 52 (2013) 12684–12687.
30] A. Pezzella, L. Capelli, A. Costantini, G. Luciani, F. Tescione, B. Silvestri, G. Vitiello,
F. Branda, Mater. Sci. Eng. C 33 (2013) 347–355.
8] R. Marotta, I. Di Somma, D. Spasiano, R. Andreozzi, V. Caprio, Chem. Eng. J. 172
[
[
[
[
[
[
[
[
[
[
(
2011) 243–249.
9] R. Marotta, I. Di Somma, D. Spasiano, R. Andreozzi, V. Caprio, J. Chem. Technol.
Biotechnol. 88 (2013) 864–872.
31] E. Cesareo, L. Korkina, G. D’Errico, G. Vitiello, M.S. Aguzzi, F. Passarelli, J.Z.
Pedersen, A. Facchiano, Plos One 7 (2012) e48849.
32] F. Zuo, L. Wang, T. Wu, Z. Zhang, D. Borchardt, P. Feng, J. Am. Chem. Soc. 132
[
[
10] D. Spasiano, L.P.P. Rodriguez, J.C. Olleros, S. Malato, R. Marotta, R. Andreozzi,
Appl. Catal. B: Environ. 136–137 (2013) 56–63.
11] W.T. Chen, V. Jovic, D. Sun-Waterhouse, H. Idriss, G.I.N. Waterhouse, Int. J.
Hydrogen Energy 38 (2013) 15036–15048.
12] K. Maeda, J. Photochem. Photobiol. C: Photochem. Rev. 12 (2011) 237–268.
13] Q. Wang, N. An, Y. Bai, H. Hang, J. Li, X. Lu, Y. Liu, F. Wang, Z. Li, Z. Lei, Int. J.
Hydrogen Energy 38 (2013) 10739–10745.
14] X. An, J.C. Yu, RSC Adv. 1 (2011) 1426–1434.
15] S. Bai, X. Shen, RSC Adv. 2 (2012) 64–98.
16] G. Jiang, Z. Lin, C. Chen, L. Zhu, Q. Chang, N. Wang, W. Wei, H. Tang, Carbon 49
(
2010) 11856–11867.
33] E.A. Konstantinova, A.I. Kokorin, S. Sakthivel, H. Kisch, K. Lips, Chim. Int. J. Chem.
1 (2007) 810–814.
34] M. Long, Y. Qin, C. Chen, X. Guo, B. Tan, J. Phys. Chem.
6734–16741.
6
[
[
C
117 (2013)
1
35] I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara, K. Takeuchi, J. Mol.
Catal. A: Chem. 161 (2000) 205–212.
36] L. C´ iri c´ , A. Sienkiewicz, R. Gaál, J. Ja c´ imovi c´ , C. Vâju, A. Magrez, L. Forrò, Phys.
Rev. B 86 (2012) 195139.
37] E.B. Li, X.Z. Li, M.F. Hou, K.W. Cheah, W.C.H. Choy, Appl. Catal. A 285 (2005)
[
[
[
(
2011) 2693–2701.
[
[
[
[
17] K. Zhou, Y. Zhu, X. Yang, X. Jiang, C. Li, New J. Chem. 35 (2011) 353–359.
18] D. Wang, D. Choi, J. Li, Z. Yang, Z. Nie, R. Kou, ACS Nano 3 (2009) 907–914.
19] Y. Liang, H. Wang, H.S. Casalongue, Z. Chen, H. Dai, Nano Res. 3 (2010) 701–705.
20] J. Liu, H. Bai, Y. Wang, Z. Liu, X. Zhang, D.D. Sun, Adv. Funct. Mater. 20 (2010)
1
81–189.
38] J. Yu, J.C. Yu, M.K.P. Leung, W. Ho, B. Cheng, X. Zhao, J. Zhao, J. Catal. 217 (2003)
9–78.
39] J. Ovenstone, J. Mater. Sci. 36 (2001) 1325–1329.
6
4
175-4181.