Study of the photocatalytic activity of Pt-modified commercial for hydrogen production in the presence of common organic sacrificial agents

Article abstract of DOI:10.1016/j.apcata.2015.09.033

A detailed study is undertaken of the photocatalytic activity in hydrogen production of the commercialphotocatalyst Kronos vlp7000 (KR) surface-modified with photodeposited platinum particles (2.1 wt%Pt). The influence of different photocatalytic variable

Full text of DOI:10.1016/j.apcata.2015.09.033

Accepted Manuscript
Title: Study of the photocatalytic activity of Pt-modified
commercial TiO for hydrogen production in the presence of
2
common organic sacrificial agents
Author: E.Pulido Meli a´ n Cristina R. L o´ pez Dunia E. Santiago
Ra u´ l Quesada-Cabrera J.A.Ortega M e´ ndez J.M.Do n˜ a
Rodr ´ı guez O.Gonz a´ lez D ´ı az
PII:
S0926-860X(15)30166-6
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.09.033
APCATA 15565
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
28-6-2015
18-9-2015
22-9-2015
Please cite this article as: E.Pulido Meli a´ n, Cristina R.L o´ pez, Dunia E.Santiago, Ra u´ l
´
´
Quesada-Cabrera, J.A.Ortega M e´ ndez, J.M.Do n˜ a Rodriguez, O.Gonz a´ lez Diaz, Study
of the photocatalytic activity of Pt-modified commercial TiO2 for hydrogen production
in the presence of common organic sacrificial agents, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.09.033
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Study of the photocatalytic activity of Pt-modified commercial TiO
2
for hydrogen
production in the presence of common organic sacrificial agents
1
1
2
E. Pulido Melián *, Cristina R. López, Dunia E. Santiago , Raúl Quesada-Cabrera , J.A.
1
1
1*
Ortega Méndez , J.M. Doña Rodríguez , O. González Díaz
1Grupo de Fotocatálisis y Espectroscopia Aplicada al Medioambiente (FEAM, Unidad
Asociada al CSIC por el Instituto de Ciencias de Materiales de Sevilla y Universidad de
Sevilla), Centro Instrumental Químico-Físico para el Desarrollo de Investigación Aplicada
(CIDIA)-Dpto. de Química, Edificio Polivalente I del Parque Científico Tecnológico,
Universidad de Las Palmas de Gran Canaria, Campus Universitario de Tafira, 35017, Las
Palmas, Spain.
2
Department of Chemistry, Materials Chemistry Centre, University College London, 20
Gordon St. London WC1H 0AJ, United Kingdom.
Corresponding author: E. Pulido Melián (elisendapm80@hotmail.com); O. González Díaz
(ogonzalez@ulpgc.es)
Graphical abstract
Highlights
1

1
2
3
4
5
. Pt-Kronos vlp7000 is a good commercial photocatalyst for hydrogen production.
. Slightly acidic or neutral pH values are more suitable for alcohol photoreforming.
. Optimum catalyst loading is around 1.5 g∙L-1.
. Reuse capacity testing favors the Pt-KR photocatalyst over the Pt-P25 reference one.
. Preliminary hydrogen production results using solar radiation are very promising.
ABSTRACT
A detailed study is undertaken of the photocatalytic activity in hydrogen production of the
commercial photocatalyst Kronos vlp7000 (KR) surface-modified with photodeposited
platinum particles (2.1wt%Pt). The influence of different photocatalytic variables is studied:
photocatalyst loading, suspension pH value and the nature and concentration of the sacrificial
agent used. Optimum pH was found to be neutral or slightly acidic. The catalyst loading, which
-
1
ranged between 0.5 and 2 g∙L , had little effect on hydrogen production. In general, the results
for the tested sacrificial agents in terms of hydrogen production were as follows (in descending
order): methanol > ethanol > ethylene glycol > glycerol. In studies of reuse the commercial
photocatalyst Kronos shows better perform in its photoactivity than the commercial
photocatalyst Aeroxide TiO P25. Finally, studies were also undertaken of photoactivity under
2
-
1
solar radiation in which promising results were obtained (5.5-18 mL∙h ).
Keywords: Hydrogen, photoreforming, methanol, ethanol, ethylene glycol, glycerol.
1.Introduction
The use of hydrogen gas (H ) as an energy storage solution is an interesting and important
2
option since, its direct combustion does not generate pollutants or greenhouse gases [1]. So far,
the most common approach for hydrogen production is steam reforming using fossil fuels,
mainly methane. However, this process requires high temperatures and pressures. Alternative
energy-efficient technologies are required to ensure that the energy consumption demanded by
an industrialised country can be met without harmful effects to the environment. Much work
and effort is being invested in the development of technologies capable of producing hydrogen
from renewable sources (principally wind and solar) [2, 3]. Another option that is being
investigated is heterogeneous photocatalysis [4], one of the so-called Advanced Oxidation
Processes. While heterogeneous photocatalysis has mainly been studied and developed for use
2

in liquid and gaseous effluent purification [5, 6], its practical application in hydrogen
production is a challenge that remains in the research and development stage [7]. The main
obstacle lies in the low efficiency of the photosplitting of pure water into H
2
and O gases
2
components. One way of increasing efficiency is through the addition of sacrificial agents that
are organic compounds that are easily oxidized photocatalytically and so they act as hole (h+)
scavengers in the photocatalysis process. These sacrificial agents favour the separation of
photogenerated electron-hole pairs, increasing charge lifetimes and enabling the migration of
electrons to the photocatalyst surface for the reduction of water to hydrogen gas. Many of
these sacrificial agents constitute a significant proportion of the chemical composition of
industrial waste, or are by-products of the agrochemical and agro-food industries.
Consequently, this technology could be used for the controlled transformation of otherwise
potential environmentally harmful waste into an energy carrier. An extensive number of
organic compounds have been tested, with the best results apparently being obtained with C1-
C3 alcohols in water-alcohol mixtures [8-15]. Among these, methanol has provided the highest
rates of H
2
gas and it is considered a model molecule in hydrogen production [16-21].
In addition to its chemical stability, low cost and low toxicity, the high photoreactivity of TiO
2
makes it an ideal photocatalyst for this purpose [22]. The deposition of metal particles,
particularly platinum (Pt), considerably enhances its photocatalytic activity in hydrogen
production [17, 23]. In this case, separation of the photogenerated electron-hole pairs is
favoured as the metal particles deposited on the photocatalyst surface act as efficient traps for
the photogenerated electrons.
A detailed study is undertaken in the present work of the photocatalytic activity in hydrogen
production of Kronos vlp7000 (henceforth, KR) photocatalyst with a photodeposited surface
Pt-loading of 2.1wt% (henceforth, Pt–KR). This photocatalyst was chosen in view of results
obtained in previous studies [24-25] in which it was compared with other commercial and non-
commercial photocatalysts and showed one of the highest activities both in its bare and Pt-
modified state. The aim of the present study is to consider its photocatalytic activity in
hydrogen production in greater detail through an analysis and optimisation of different
variables including photocatalyst loading, initial suspension pH value and the nature and
concentration of the sacrificial agent used. Not only methanol is used, but also other sacrificial
agents such as ethanol that can be produced from biomass [26], and ethylene glycol and
glycerol that are present in difficult-to-treat waste materials including vehicle coolants and
biodiesel production by-products [27]. The reuse capacity of the photocatalyst is also tested as
well as its activity under solar radiation.
2. Experimental
3


 Reagents and materials
The commercial photocatalysts used in this study were Kronos vlp7000 (Patent
US2005/0226761A1) supplied by Kronos Worldwide, Inc. and TiO Aeroxide P25 (47.2 ± 0.1
2
2
-1
m g , 79%/21% anatase/rutile) supplied by Evonik. The reagents methanol, ethanol and
ethylene glycol with 99, 99.5 and 99% purity, respectively, were supplied by Panreac and the
glycerol reagent by Sigma-Aldrich (99.5% purity). The pH was adjusted using NaOH (≥ 97%,
Sigma Aldrich) and H
2
SO (96%, Panreac). All the solutions were made using ultrapure Milli-
4
Q water (Millipore Corporation).
2
.2. Photodeposition of Pt metal particles
The Kronos vlp7000 photocatalyst was subjected to a photodeposition process of Pt metal
particles from hexachloroplatinic acid (IV) (H PtCl 6H O, Merck 40% Pt), following a
modified method from the literature [28]. 2 g of catalyst were introduced in 400 mL of a 0.3 M
solution of isopropanol, together with H PtCl 6H O, in order to obtain a Pt loading of 2.1 wt%
2
6
2
2
6
2
on the catalyst surface. This metal loading was established as the appropriate amount for the
optimum performance of the Kronos sample [25]. The photodeposition was carried out in a
commercial photoreactor (Photochemical Reactors Ltd.), equipped with a medium-pressure
mercury lamp (400W). The lamp was inserted into a quartz immersion well. The system was
kept at constant temperature (25 ºC) in a thermostatic bath. The illumination was maintained
for 6 h under continuous N
2
gas flow and strong stirring conditions. The recovery of the sample
was performed by filtration and it was washed and dried overnight in an oven at 110 ºC.
2.3. Photocatalytic test
The method employed for H
2
gas generation comprised the introduction into a borosilicate
reactor of 75 mL photocatalyst suspension with a predetermined concentration of a sacrificial
agent, pH adjustment and continuous stirring at 500 rpm. Three Philips HB175 lamps were
-
2
used as illumination source, each with four 15 W CLEO fluorescent tubes (9 mW∙cm ; = 300–
00 nm, max= 365 nm), placed 15 cm from the reactor. Spectra and irradiance measurements
4
were recorded using an Ocean Optics HR2000+ spectrometer. The system was operated at
ambient temperature and pressure (approx. 25 ºC and 1 atm). A reactor of larger capacity with
-
1
1
80 mL of the catalyt suspension (1 g∙L ) was used for the solar radiation tests and was placed
on the roof of the university building where the research group of the present study is based
28° 04’ 49.13’’ N; 15° 27’ 9.04’’ W). The suspension was continuously recirculated at 1200
(
-
1
mL∙min using a Selecta Percom N-M peristaltic pump. The photoreactor was inclined at a tilt
angle equivalent to the latitude angle of the location (28°) so that the direction of radiation
incidence would be as perpendicular as possible and was oriented south. A dihedral mirror
4

reflector was used to maximize solar radiation exploitation. Instantaneous and accumulated
solar energy was measured with an Acadus 85-PLS (Ecosystem S.A.) radiometer with a
response range between 290 and 370 nm.
In both, outdoor and indoor experiments, the monitoring of H gas and gaseous intermediates
2
was performed using a 490 Agilent MicroGC gas chromatograph (GC) fitted with injection
valve, a Molsieve 5A column and TCD detector. The Molsieve 5A column also allowed
analysis of carbon monoxide. The device also includes a PoraPlot Q column enabling
-
1
monitoring of carbon dioxide, methane and ethane. A constant flow of He (10 mL∙min ) was
maintained using a mass flow controller (EL-FLOW Bronkhorst HI-TECH) through the reactor
headspace, which directed the generated H gas into the chromatograph. The chromatogram
2
areas were converted to hydrogen flow using a calibration performed beforehand under the
same test conditions.
Formaldehyde concentration in liquid phase was quantified by a spectrophotometric method
[
29] using a Heλios spectrophotometer (Thermo Electron Corporation). Formic acid and
acetic acid were quantified by high performance liquid chromatography (HPLC) using a Varian
Prostar system with a diode array UV-Vis detector (Supelcogel column C-610H, H PO
3
4
1
-
1
vol.% of phosphoric acid in water, 1 mLmin , 25ºC). Acetaldehyde was analyzed using a
Shimadzu GC2010 Plus chromatograph fitted with a CP-WAX 52 CB column (0.25 mm x 30
m) and an FID detector. The pH was measured and adjusted with a Crison Basic20 pH-meter.
2.4 Characterization techniques
XRD patterns were recorded on a Siemens D-501 diffractometer equipped with a Ni filter and
graphite monochromator using Cu Kα radiation (λ = 1.5418 Å) to determine phase composition
of the samples. Crystal sizes were estimated from line broadening of the corresponding X-ray
diffraction peaks by using the Scherrer equation. Diffuse reflectance spectra were recorded for
all samples on a Varian Cary 5 and the Kubelka–Munk function, F(R∞), was applied to obtain
band-gap energies following the method proposed by Tandon and Gupta [30]. BET surface
area, pore volume and pore size measurements were carried out by N adsorption at 77 K using
2
a Micromeritics 2010 instrument. Field emission scanning electron microscopy (FESEM)
images were obtained using a Hitachi S-4800 microscope. The morphology of the samples was
studied by high resolution transmission electron microscopy (HR-TEM) using a JEOL JEM
2100 microscope combined with energy-dispersive X-ray (EDX).
 Results and discussion
3.1. Characterization of the Kronos TiO
2
-based materials
5

The Kronos vlp7000 (KR) is a pure anatase TiO
2
catalyst with a large surface area (258.60 m2
g ), as inferred from the small particle size observed from microscopy analysis (Figure 1). As
-
1
∙
discussed in previous studies [25], the photodeposition of Pt metal particles (Pt–KR) had an
impact on the catalyst surface, resulting in a slight increase in particle size and pore volume
and even, in large-size aggregates in aqueous suspension which do not appear in pure TiO
2
photocatalyst. The size of these platinum particles has been estimated from HR-TEM images
obtaining values around 3.5-4 nm, although Pt aggregates of larger size were clearly observed.
The relevant characteristics of both KR and PtKR (2.1 wt% Pt) samples are compared in Table
1
[24, 25]. According to the diffuse reflectance spectrum (see Supplementary Information,
Figure S1), the presence of Pt particles also resulted in increased catalyst radiation absorption
in the visible range. Nonetheless, no modification can be observed of the crystalline structure
(see Supplementary Information, Figure S2) or of the energy threshold for photoactivity.
3
.2. Hydrogen Production
The photogeneration of H
efficiency of the systems is discussed in the following subsections.
+
2
gas was carried out using a range of hole (h ) scavengers and the
3.2.1 Methanol as sacrificial agent
Hydrogen production in this study is considered for a wide range of methanol concentrations,
-
1
including at 100% v/v (24.47 M). Figure 2A represents the mean values of H flow (µmol∙h )
2
after a steady production rate has been obtained. As reported by other authors, hydrogen flow
rises to a maximum rate at which it remains stable until beginning to fall upon depletion of the
sacrificial agent or deactivation of the catalyst [31] (Figure 2B). Maximum production was
attained after approximately 60 min of the 3.5 h duration of the test and remained more or less
stable for the remaining time, even for the 100% v/v methanol concentration. It can be seen that
hydrogen production increased with methanol concentration up to 7.34 M where the rate was
-
1
1
665 µmol∙h . With a 100% v/v concentration of sacrificial agent it can be seen that hydrogen
flow is even lower than for the lowest tested initial methanol concentration of 1.22 M.
The main by-products of methanol photoreforming are formaldehyde, formic acid and CO
2
.
Other intermediates, such as methane, methyl formate, acetaldehyde and dimethyl ether as well
as traces of ethane, have been found in methanol photoreforming reactions, [17] however these
were found insignificant in the overall process. The amounts of both formaldehyde and formic
acid increased with methanol concentration up to 6.12 and 9.79 M, respectively (Figure 2A).
Beyond these concentrations, the formation of formaldehyde remained constant whilst the
concentration of formic acid suddenly decreased. However, this trend was not maintained for
100% v/v methanol concentration (24.47 M) and the concentration of both intermediates
6

increased markedly. The build-up of formaldehyde at 100% v/v is due to the fact that its
subsequent oxidation into formic acid is strongly inhibited in the absence of water [32]. Figure
3
A shows how the CO
2
flow rates increased over time for all methanol concentrations except
for the 100% v/v, where it reached a steady rate of production. It can also be seen how CO
2
flow rates fall as initial methanol concentration rises. The mineralisation rate for 100% v/v
methanol is very low, as inferred from the accumulation of formaldehyde observed in Figure
2A. Furthermore, traces of methane were also detected (Figure 3B), the formation of which is
competitive and detrimental to hydrogen production as all the hydrogen atoms available in the
initial methanol will not result in the formation of hydrogen.
The influence of pH and catalyst loading was studied while maintaining a constant methanol
concentration of 25% (v/v). As it can be seen in Figure 4A, an optimum production of H
2
gas
was found under slightly acidic and neutral pH conditions, whilst the H rates decreased
2
significantly at strongly acidic or alkaline pH values. This is in agreement with previous
literature reports and it is a common observation for a range of alcohols, including ethanol and
ethylene glycol [33-35]. This behaviour is expected for all the alcohols considered in this work
since their acidity constants are very similar (pKa methanol= 15.5, pKaethanol= 15.9, pKaethylene
glycol= 14.22 and pKaglycerol= 14.2).
The concentration of formaldehyde ranged between 65 and 80 mM, following the same trend as
the H rates. However, the concentrations of formic acid were much lower (< 1.50 mM) and
2
the trend increased with increasing pH, as expected, since its formation is inhibited in acid
conditions [35].
The CO
2
flow rates (results not shown) for 3.5 h had the following order for the different pH
-
1
-1
-1
values: pH 5 (17.35 µmol∙h ) > pH 7 (6.37 µmol∙h ) > pH 3 (5.00 µmol∙h ) > pH 9 (2.44
-
1
µmol∙h ) with higher initial induction periods for pH 9 and pH 3 of 135 min. This is in
accordance with hydrogen flow rates and reported formaldehyde and formic acid
concentrations. Specifically at pH 9 it was expected that the formic acid accumulation (in
formate form) resulted in low mineralization and at pH 3 the low CO flow is well related to
2
the lower hydrogen production rate and to the inhibition of the formation of formic acid in
acidic conditions.
Little variation in hydrogen production was observed for different catalyst loadings (0.52.0 g∙L-
1)
(Figure 4B), ranging between 1396 and 1558 µmol∙h . A slight increase in hydrogen
-1
-
1
production can be observed up to a catalyst loading of 1.5 g∙L . At higher loading, the rate
decreased likely due to catalyst particle aggregation, with reduction of active surface area, and
to radiation scattering, which impedes transmission of the radiation throughout the
photoreactor and therefore impedes the radiation from reaching the active centres of the
catalyst particles [36-38]. Formic acid and formaldehyde concentrations also decreased with an
7

-
1
increase in catalyst loading above 1 g∙L . The effect of catalyst loading on the activity is in
agreement with that reported in the literature and the optimum loading is found within the range
considered in the present study [39, 40].
3.2.2. Ethanol as sacrificial agent
Figure 5A represents hydrogen flow as a function of initial ethanol concentration and the
concentrations of the main intermediates of ethanol photoreforming, acetaldehyde and acetic
acid. It can be seen how the hydrogen production increased to an initial ethanol concentration
-
1
of 4.89 M where it reached a flow rate of 1378 µmol∙h . This flow rate is comparable to that
obtained when using similar concentrations of methanol (< 4.89 M) (Figure 2A). The H rates
2
from ethanol at low concentrations can be relatively high compared to those from methanol
solutions [18, 32]. An indirect oxidation mechanism via hydroxyl radicals prevails at low
concentrations of alcohol [41], reducing the difference in H
ethanol solutions. However, high alcohol concentrations involve direct oxidation mechanisms
and methanol is a better hole scavenger [42], which explains why the H rates increased upon
increasing concentrations of methanol whilst it remained constant for ethanol concentrations
between 4.89 and 7.34 M (Figure 5A). At higher concentrations (>7.34 M), the H rate
dropped drastically for methanol and a surprisingly high rate was found in saturated ethanol
2
production between methanol and
2
2
-
1
-1
solution (1252 µmol∙h ) compared to the equivalent methanol system (822 µmol∙h ). The high
efficiency of Pt-KR with saturated ethanol solutions will need further investigation, however,
so far, preliminary tests up to 24 h (not shown) have confirmed these result whilst other
catalysts undergo important deactivation within 3.5 h in this study [32].
The concentration of acetaldehyde in the suspension was between 24 M and 61 mM, varying
with the same tendency as hydrogen flow, while the concentration of acetic acid generally
decreased (from 3 mM to 0.5 mM) as the initial ethanol concentration increased. At the same
time, as it happened in the case of methanol, in general the CO levels drop upon increasing
2
concentration of alcohol (Figure 5B). However, in this case, there is significant mineralization
with 100% v/v ethanol, which is in accordance with the high rate of hydrogen production
observed for this concentration and the low concentrations of intermediates when compared
with the initial ethanol concentration of 1.22 M.
It is important to note that, whereas neither methane nor ethane were found for the
corresponding pure alcohol solutions (100% v/v), they were detected in the diluted systems
(Figure 6). The formation of these two sub-products would appear to be related to the presence
of water. In the case of the ethanol-containing solutions, the production of methane was
favoured at low initial ethanol concentrations whereas the ethane levels were high at high
ethanol concentrations. The flows of these by-products are rather significant when using
8

ethanol as sacrificial agent, and in the case of methane they are even higher than those of CO
2
gas. These two gaseous intermediates, methane and ethane, are not commonly investigated in
the literature, with the exception of Bamwenda et al. [33]. These authors detected methane,
although not quantified, and explained its formation based on the decarboxylation of acetic acid
or hydrogenation of CO . Both methane and ethane were detected and quantified in a previous
2
publication in our group [32] and they have been confirmed here. Ethane could be formed from
the combination of two methyl radicals generated as a result of acetic acid decarboxylation
[43].
As a result of the presence of these by-products the mineralization rates observed for ethanol
are much lower than for methanol. Ethanol appears to lose efficiency in hydrogen production
due to the formation of these intermediates. Higher hydrogen production rates have accordingly
always been reported in the literature for methanol [18, 19], though only justified as a
consequence of the size of the alcohol chain and so to steric problems hindering approach to the
catalyst surface and not to the generation of photoreduction by-products as it is evidenced here.
Furthermore, these alkanes seem to form significantly from those alcohols where not all carbon
atoms are linked to hydroxyl groups [44].
 Ethylene glycol as sacrificial agent
Contrary to the observations described for methanol and ethanol solutions (vide supra), no
steady rate of hydrogen flow was reached when using ethylene glycol as sacrificial agent, with
the flow rate beginning to slowly descend after reaching a maximum value (Figure 2B).
Figure 7A represents maximum flow rates as a function of initial concentration of ethylene
glycol and absolute hydrogen production and formaldehyde and formic acid concentrations
after 3.5 h. It can be seen how hydrogen production falls as ethylene glycol concentration rises,
with the same trend observed for the intermediates. Unfortunately, photoreforming studies
using ethylene glycol are lacking in the literature. Bamwenda et al. [34] quantified CO and
2
detected traces of methane from ethylene glycol solutions. In the current work, the presence of
both formaldehyde and formic acid could be confirmed and quantified, in agreement with our
previous publication [32]. It was interesting to observe that the maximum flow rate (1347
-
1
µmol∙h ) was higher than the flow rates recorded from methanol and ethanol for the same
concentration. When comparing ethylene glycol with ethanol, both of which are two-carbon
alcohols, it appears that the effect descripted by Fu et al. [44] about the enhancement of the
efficiency of the process due to the presence of more OH groups is only decisive for low
concentrations.
Of the gaseous intermediates, CO flow rates dropped drastically upon increasing alcohol
2
concentration (Figure 7B) and traces of methane were only found for the lowest concentrations
9

of ethylene glycol. This is again evidence of the fundamental role of water in the formation of
alkanes (vide supra).
3
.2.4. Glycerol as sacrificial agent
The maximum H flow rate in glycerol-containing solutions (Figure 8A) was almost as high
1300 µmol∙h ) as that recorded from ethylene glycol systems. However, contrary to the latter,
2
-
1
(
a steady rate of hydrogen flow was reached when using glycerol as sacrificial agent. Hydrogen
flow rate remained stable in a wide range of glycerol concentrations (up to 8.53 M), after
which it decreases. The 100% v/v concentration was not tested in the case of glycerol given the
extremely high viscosity of this substance.
As glycerol is a more complex alcohol, a large number of intermediates can be formed, among
these acetic acid, methanol, carbon dioxide, carbon monoxide, acetol, acetaldehyde, ethanol,
glyceraldehyde, glycolaldehyde, acetone and acrolein have been reported in the literature [45].
Using the methodologies employed in the present study, formaldehyde and formic acid were
detected in the liquid phase. These intermediates decreased upon increasing concentrations of
glycerol (Figure 8A) and they would originate from methanol, following the mechanism
reported by Panagiotopoulou et al. [45]. As for the gaseous intermediates, traces of methane
could be detected in all the solutions and the CO
2
flow rates were as high as those observed
from methanol and ethylene glycol systems (Figure 8B).
3.2.5. Catalyst reuse
The Pt-KR catalyst was subjected to various reuse cycles in order to evaluate its ability to
maintain its photocatalytic performance in continuous use (Figure 9). A low sacrificial agent
load (0.011 M), either methanol, ethylene glycol or glycerol, was used in the cycles so that the
2
H flow rate would reach a zero value within reasonable time periods. For such small
concentrations, the production patterns reached a maximum value and dropped immediately
without maintaining a steady flow rate.
A stable production (around 2200-2300 moles) from one cycle to the next can be seen in the
case of methanol, with a cycle time of 15 hours. These results were comparable to those
obtained for another reference commercial photocatalyst, Aeroxide TiO P25, modified with
2
photodeposited platinum particles (2.1 wt% Pt, Pt-P25) under the same conditions [32].
However, if we compare the Pt-KR results using ethylene glycol and glycerol with those for the
Pt-P25 catalyst (Figure 10), the Pt-KR clearly performs better.
Using ethylene glycol as sacrificial agent the cycles with the Pt-KR photocatalyst are 20 hours
long whereas with the Pt-P25 some 45 hours or more of illumination are required for the flow
-
1
rate to fall below 5 molh . Of the three sacrificial agents, glycerol required the most time for
1
0

its complete elimination. This concurs with the observations of Panagiotopoulou et al. [45]
who reported that more than 300 h were required to eliminate the glycerol contained in 60 mL
of a 20 mM glycerol solution. The activity of the Pt-P25 photocatalyst decreased from one
cycle to the next as it required longer time periods in each successive cycle to reach the same
absolute production (from 62 h in the first cycle to more than 140 h in the third). In contrast,
the cycles using the Pt-KR photocatalyst displayed more reproducibility in terms of activity
with cycle times of around 30 h. This has been explained in terms of specific surface area,
2
-1
2 -1
which is 5 times larger in Pt-KR (280 m g ) compared to Pt-P25 (49.2 m g ), and thus the
former would be less affected by active centre blockage from the adsorption of organic
intermediates during degradation mechanism of the more complex alcohols. Likewise, a
catalyst with small surface area studied in a previous publication [32], SG750-2.1 wt%Pt (18
2
-1
m g ), showed rather poor efficiency compared to Pt-P25 using these sacrificial agents.
The hydrogen production rates reported from the degradation of organic compounds in the
literature are typically explained based on the photoreforming of the sacrificial agent only [17,
46-48], which give us the following equations:
[eq 1]
[
eq 2]
eq 3]
[
The amounts of hydrogen and carbon dioxide obtained from methanol and ethylene glycol
solutions containing the Pt-KR photocatalyst, were those expected from a stoichiometric basis.
However, glycerol did not reach full mineralization, despite de long reaction times used, and the
2
H rates obtained represented the 85% of the total balance.
3.2.6. Solar Photoactivity
Preliminary tests were carried out to evaluate the performance of the Pt-KR catalyst for
hydrogen production under solar irradiation conditions. These tests were carried out using a
-
1
concentration of 1 g∙L of Pt-KR catalyst, pH natural conditions and a concentration of 7.34
M of the corresponding sacrificial agent. The H flow rates were represented as a function of
accumulated energy (Figure 11A) and as a function of time (Figure 11B) until a total
2
-
2
accumulated energy of 44.7 W∙h∙m was reached. In the solar tests, the highest absolute
production was determined for glycerol-containing dispersions (4183.87 µmol), closely
followed by those containing methanol (3428.7 µmol) and those with ethanol lagging some way
behind (1029.4 µmol). These absolute rates were estimated from the areas under the curves in
Figure 11B and did not follow the trends observed in the laboratory experiments using artificial
light (vide supra). This difference was attributed to the variable nature of solar irradiance. The
1
1

solar tests were carried out on different days over the same week with considerable variation in
instantaneous energy, which was clearly a critical factor in the hydrogen production rates
(Supplementary Information Figure S3).
During the experiments using methanol, the instantaneous energy values remained generally
-
2
high and only 4.5 h were required to reach the 44.7 W∙h∙m value. However, there was only
one period of intense solar radiation for the experiments with ethanol and 7.3 h were required
to reach the same accumulated energy.
While glycerol was not the best sacrificial agent according to the results of the laboratory tests,
the hydrogen production under solar radiation was almost the same as that obtained in the lab
tests though requiring a longer time period (6 h as opposed to 3.5 h).
The values were also low since only 3–5 % of the total solar emission is within the UV range.
In terms of irradiance, this means that even under optimum solar radiation conditions, the
outside irradiance levels were 6 times below those provided by the lamps in the laboratory.
Based on tests using solar simulators, Chowdhury et al. [36] reported that the hydrogen
-
2
production dependency with irradiance was linear in the range 30–100 mW∙cm . These
-
2
energies however would have to be obtained with concentrators since 100 mW∙cm is the
maximum irradiance of solar radiation that reaches the surface of the planet [49].
-
1
Studies by Bowker et al. [50] reported mean production rates of 0.3 mL∙h at latitude N 51º 35
over a 6 h period of a sunny day in June whereas the rates reported in the present study
´
-
1
ranged between 5.518 mL∙h , in Gran Canaria (N 28 4). The results obtained with solar
radiation are encouraging and optimization of the reaction variables (pH, photocatalyst
loading…), reactor design and radiation collection would further improve the results, especially
if efficient catalysts were available to capture a wider solar radiation spectrum.
4
. Conclusions
The H production was notably influenced by the pH conditions, with the best results obtained
for slightly acidic and neutral pH values. Optimum catalyst loading for the system under study
2
-
1
was determined as 1.5 g∙L , though its influence was less important than that of pH conditions.
From the maximum rates obtained, it could be concluded that methanol was the most
-
1
-1
convenient sacrificial agent (1665 µmol∙h at 7.34 M), followed by ethanol (1378 µmol∙h at
-
1
-1
4
.89 M), ethylene glycol (1347 µmol∙h at 1.22 M) and glycerol (1300 µmol∙h a 1.22 M-8.51
M) with similar H rates. The total H production obtained from photoreforming of methanol
2
2
and ethylene glycol at low concentration (0.011 M) are those expected from reaction
stoichoimetry, resulting in full mineralisation within 15 and 20 h, respectively for methanol and
ethylene glycol. In the case of glycerol, the H
2
rates were 85% that of the theoretical value
within 30 h.
1
2

The activity and reusability of the Pt-KR catalyst exceeded those of the platinised standard, Pt-
P25, particularly in the case of ethylene glycol and glycerol. This has been explained in terms
of the high surface area of the former.
Traces of methane were observed from methanol, ethylene glycol and glycerol whilst both
methane and ethane were detected in significant quantities in ethanol solutions. The presence of
these by-products in ethanol hindered the efficiency of the system. These intermediates seem to
form abundantly in alcohols containing a limited number of carbon-hydroxyl groups. In
addition, water seems to play a crucial role in the formation of these intermediates.
Encouraging results were obtained in preliminary tests carried out using solar radiation.
However, an increase in the amount of useful photoactive energy not only implies enhancement
of reactor design but also of the materials used as catalysts. Such an improved design could
enable the system to be used for household supply.
Acknowledgements
The authors thank the Spanish Ministry of Economy and Competitiveness for their support
through the project IPT-120000-2010-033 (GESHTOS) and the infrastructure project
UNLP10-3E-726, co-financed with ERDF funds. Cristina R. López would also like to thank
the Spanish Ministry of Science and Innovation for the pre-doctorate grant of the Research
Training Program (Spanish initials: FPI) and E. Pulido Melián would like to express her
gratitude to the University of Las Palmas de Gran Canaria for her post-doctorate contract.
References
[
[
[
[
[
(
1] J.D. Holladay, J. Hu, D.L. King, Y. Wang, Catal. Today 139 (2009) 244-260.
2] D. Honnery, P. Moriarty, Int. J. Hydr. Energy 34 (2009) 727-736.
3] T. Abbasi, S.A. Abbasi, Renew. Sustainable Energy Rev. 15 (2011) 3034-3040.
4] C.H. Liao, C.W. Huang, J.C.S. Wu, Catalysts 2 (2012) 490-516.
5] C.C. Kaan, A.A. Aziz, S. Ibrahim, M. Matheswaran, P. Saravanan, in: M. Kumarasamy
Ed.), Heterogeneous Photocatalytic Oxidation an Effective Tool for Wastewater Treatment –
A Review, InTech, 2012, pp. 219-236.
[
[
[
6] A.O. Ibhadon, P. Fitzpatrick, Catalysts 3 (2013) 189-218.
7] A.A. Ismail, D.W. Bahnemann, Sol. Energ. Mat. Sol. Cells 128 (2014) 85-101.
8] M.P. Languer, F.R. Scheffer, A.F. Feil, D.L. Baptista, P. Migowski, G.J. Machado, G.J.
Machado, D.P. de Moraes, J. Dupont, S.R. Teixeira, D.E. Weibel, Int. J. Hydr. Energy 38
1
3

(
[
[
[
[
2013) 14440-14450.
9] I. Rossetti, ISRN Chem. Engineer. 2012 (2012) Article ID 964936, 21 pages.
10] A. Galinska, J. Walendziewski, Energy Fuels 19 (2005) 1143-1147.
11] M.R. Stjohn, A.J. Furgala, A.F. Sammells, J. Phys. Chem. 87 (1983) 801-805.
12] X. Fu, J. Long, X. Wang, D.Y.C. Leung, Z. Ding, L. Wu, Z. Zhang, Z. Li, X. Fu, Int. J.
Hydr. Energy 33 (2008) 6484-6491.
[
[
[
13] T. Kawai, T. Sakata, Nature 286 (1980) 474-476.
14] Y.X. Li, Y.Z. Me, S.Q. Peng, G.X. Lu, S.B. Li, Chemosphere 63 (2006) 1312-1318.
15] A. Patsoura, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 64 (2006) 171-
1
79.
16] T.A. Kandiel, R. Dillert, L. Robben, D.W. Bahnemann, Catal. Today 161 (2011) 196-
01.
[
2
[17] G.L. Chiarello, M.H. Aguirre, E. Selli, J. Catal. 273 (2010) 182-190.
[18] N. Strataki, V. Bekiari, D. I. Kondarides, P. Lianos, Appl. Catal. B: Environ. 77 (2007)
1
84–189.
19] M. Ni, M.K.H. Leung, D.Y.C. Leung, K. Sumathy, Ren. Sust. En. Rev. 11 (2007) 401–
25.
20] J.G. Highfield, M.H. Chen, P.T. Nguyen, Z. Chen, Energy Environ. Sci. 2 (2009) 991-
002.
[
4
[
1
[
[
[
21] H.-J. Choi, M. Kang, Int. J. Hydr. Energ 32 (2007) 3841-3848.
22] O. Carp, C.L. Huisman, A. Reller, Prog. Solid State Chem. 32 (2004) 33-177.
23] A. Naldoni, M. D’Arienzob, M. Altomare, M. Marelli, R. Scotti, F. Morazzoni, E. Selli,
V. Dal Santo, Appl. Catal. B: Environ. 130–131 (2013) 239-248.
24] E. Pulido Melián, O. González Díaz, A. Ortega Méndez, C. Rodríguez López, M. Nereida
Suárez, J.M. Doña Rodríguez, Int. J. Hydr. Energy 38 (2013) 2144–2155.
25] E. Pulido Melián, A. Ortega Méndez, Cristina R. López, O. González Díaz, M. Nereida
Suárez, J.M. Doña Rodríguez, J.A. Navío, D. Fernández Hevia, Int. J. Hydr. Energ. 38
[
[
(
[
[
[
[
2013) 11737-11748.
26] E. Gnansounou, A. Dauriat, J. Sci, Ind. Res. 64 (2005) 809-821.
27] J.V. Gerpen, Fuel Proc. Technol. 86 (2005) 1097-1107.
28] B. Kraeutler, A.J. Bard, J. Am. Chem. Soc. 100 (1978) 2239-2240.
29] E.R. Kennedy, in: P.M. Eller, M.E. Cassinelli (Eds.), Formaldehyde: method 3500, in:
NIOSH manual of analytical method (NMAM), Atlanta, 1994, pp. 243-247.
[30] S.P. Tandon, J.P. Gupta, Physica Status Solidi, 38 (1970) 363–367.
[31] J. Puskelova, L. Baia, A. Vulpoi, M. Baia, M. Antoniadou, V. Dracopoulos, E. Stathatos,
K. Gabor, Z. Pap, V. Danciu, P. Lianos, Chem. Eng. J. 242 (2014) 96-101.
1
4

[
32] C.R. López, E. Pulido Melián, J.A. Ortega Méndez, Dunia E. Santiago, J.M. Doña
Rodríguez, O. González Díaz, J. Photochem. Photibiol. A: Chem. 312 (2015) 45-54.
33] G.R. Bamwenda, S. Tsubota, T. Nakamura, M. Haruta, J. Photochem. & Photobiol. A:
Chem. 89 (1995) 177-189.
34] G.R. Bamwenda, S. Tsubota, T. Kobayashi, M. Haruta, J. Photochem. & Photobiol. A:
Chem. 77 (1994) 59-67.
[
[
[
(
[
(
[
35] G. Wua, T. Chena, W. Sua, G. Zhoua, X. Zonga, Z. Leia, C. Lia, Int. J. Hydr. Energy 33
2008) 1243-1251.
36] P. Chowdhury, G. Malekshoar, M.B. Ray, J. Zhu, A.K. Ray, Ind. Eng. Chem. Res. 52
2013) 5023-5029.
37] E. Pulido Melián, O. González Díaz, J.M. Doña Rodríguez, J. Araña, J. Pérez Peña,
Applied Catalysis A: General 455 (2013) 227– 233.
[38] S. Kaur, V. Singh, J. Hazard. Mater. 141 (2007) 230-236.
[39] S. Oros-Ruiz, R. Zanella, R. López, A. Hernández-Gordillo, R. Gómez, J. Hazard. Mater.
263 (2013) 2-10.
[
[
[
40] H.M.G. Tambago, R.L. de Leon, Int. J. Chem. Eng. Appl. 6 (2015) 220-227.
41] G.L. Chiarello, D. Ferri, E. Selli J. Catal. 280 (2011) 168–177.
42] Y. Tamaki, A. Furube, M. Murai, K. Hara, R. Katoh, M. Tachiya, J. Am. Chem. Soc.
1
28 (2006) 416–417.
43] K. Rajalakshmi, V. Jeyalakshmi, K.R. Krishnamurthy, B. Viswanathan, Indian J. Chem.
1 (2012) 411-419.
44] X. Fu, X. Wang, D. Y. C. Leung, Q. Gu, S. Chen, H. Huang, Appl. Catal. B: Environ.
06 (2011) 681-688.
[
5
[
1
[
[
(
[
45] P. Panagiotopoulou, E.E. Karamerou, D.I. Kondarides, Catal. Today 209 (2013) 91-98.
46] O. Rosseler, M.V. Shankar, M.K.L. Du, L. Schmidlin, N. Keller, V. Keller, J. Catal. 269
2010) 179-190.
47] D.I. Kondarides, V.M. Daskalaki, A. Patsoura, X.E. Verykios, Catal. Lett. 122 (2008)
2
6-32.
48] V.M. Daskalaki, P. Panagiotopoulou, D.I. Kondarides, Chem. Eng. J. 170 (2011) 433-
39.
49] M. Romero, D. Martínez, E. Zarza, Terrestrial solar thermal power plants: on the verge of
commercialization, In: 4th International Conference on Solar Power from Space, 2004.
[
4
[
[
50] M. Bowker, H. Bahruji, J. Kennedy, W. Jones, G. Hartley, C. Morton, Catal. Lett. 145
2015) 214-219.
(
1
5

Table
Table 1. Properties of the KR and Pt–KR (2.1 wt% Pt) photocatalysts, including crystallite
size, band-gap energy (E ), surface area (SBET) and pore volume.
g
Photocatalysts Anatase
%)
Size
(nm)
7.0
E
g
S
BET
-1
Pore Volume
2
3
-1
(
(eV) (m ∙g )
3.24 258.60
3.20 280.65
(cm ∙g )
0.392
KR
100
100
Pt–KR
8.0
0.407
Figure Capitons
Figure 1. (A) HR-TEM image and (B) back-scattering SEM image of the Pt-modified Kronos
sample (PtKR).
Figure 2. (A) Hydrogen flow rates (bars) and formaldehyde (o) and formic acid (■)
concentrations as a function of methanol concentration. (B) Hydrogen patterns vs. time for
7.34 M concentrations of the tested sacrificial agents.
1
6

Figure 3. Evolution of (A) CO
2
and (B) CH in the photocatalytic runs with methanol used as
4
sacrificial agent.
1
7

Figure 4. Hydrogen production (bars) and concentrations of formaldehyde (●) and formic acid
■) by-products as a function of (A) pH and (B) catalyst loading.
(
1
8

Figure 5. (A) Hydrogen flow rates (bars) and acetaldehyde (o) and acetic acid (■)
concentrations as a function of ethanol concentration. (B) Evolution of CO
2
vs. time in the
photocatalytic runs with ethanol used as sacrificial agent.
1
9

Figure 6. Evolution of (A) methane (CH
4
) and (B) ethane (C
2
6
H ) in the photocatalytic
degradation with ethanol used as sacrificial agent.
Figure 7: (A) Hydrogen flow rates (bars) and formaldehyde (o) and formic acid (■)
concentrations as a function of ethylene glycol concentration (total micromoles of hydrogen
obtained in 3.5 h are also shown). (B) Evolution of CO
2
vs. time in the photocatalytic runs with
ethylene glycol as sacrificial agent.
2
0

Figure 8. (A) Hydrogen flow rates (bars) and formaldehyde (o) and formic acid (■)
concentrations as a function of glycerol concentration. (B) Evolution of CO
2
vs. time in the
photocatalytic runs with glycerol as sacrificial agent.
2
1

Figure 9. Reuse cycles with (A) methanol, (B) ethylene glycol and (C) glycerol for Pt-KR.
Total micromoles of hydrogen are also shown.
2
2

Figure 10. Reuse cycles with (A) ethylene glycol and (B) glycerol for TiO
2
Aeroxide P25
2.1wt%Pt. Total micromoles of hydrogen are also shown.
2
3

Figure 11. Hydrogen flow as a function of (A) accumulated energy and (B) time (1 g∙L-1
catalyst, pH natural and a concentration of 7.34 M sacrificial agent).
2
4

2
5