Photocatalytic reforming of formic acid for hydrogen production in aqueous solutions containing cupric ions and TiO2 suspended nanoparticles under UV-simulated solar radiation

Article abstract of DOI:10.1016/j.apcata.2015.09.020

Hydrogen is the ideal candidate to fulfill the growing energy demand in a sustainable manner because of its high energy content and no emission of greenhouse gases from its combustion. Currently most of hydrogen generation techniques involve the employmen

Full text of DOI:10.1016/j.apcata.2015.09.020

Accepted Manuscript
Title: Photocatalytic reforming of formic acid for hydrogen
production in aqueous solutions containing cupric ions and
TiO2 suspended nanoparticles under UV–simulated solar
radiation
Author: Laura Clarizia Ilaria Di Somma Raffaele Marotta
Patrizia Minutolo Roberta Villamaina Roberto Andreozzi
PII:
S0926-860X(15)30153-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.09.020
APCATA 15552
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
18-7-2015
9-9-2015
16-9-2015
Please cite this article as: Laura Clarizia, Ilaria Di Somma, Raffaele Marotta, Patrizia
Minutolo, Roberta Villamaina, Roberto Andreozzi, Photocatalytic reforming of formic
acid for hydrogen production in aqueous solutions containing cupric ions and TiO2
suspended nanoparticles under UVndashsimulated solar radiation, Applied Catalysis
A, General http://dx.doi.org/10.1016/j.apcata.2015.09.020
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.

1
2
3
Photocatalytic reforming of formic acid for hydrogen production in aqueous
solutions containing cupric ions and TiO2 suspended nanoparticles under UV–
simulated solar radiation
4
5
Laura Clariziaa, Ilaria Di Sommab, Raffaele Marottaa, Patrizia Minutolob, Roberta Villamainaa and
Roberto Andreozzia
6
7
8
a Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Università di
Napoli Federico II, p.le V. Tecchio 80, 80125 Napoli, Italia.
b Istituto di Ricerche sulla Combustione, Centro Nazionale delle Ricerche IRC–CNR, p.le V. Tecchio
80, 80125 Napoli, Italia.
9
10
11
12
Graphical abstract
13
14
Highlights:
15
16
17
18
19

Assessment on the efficient photocatalytic system Cu2+/HCOOH/Cl-/nano-TiO2/UVSSR for H2
evolution


Evaluation of the system behavior at varying selected experimental variables
Setting of the best operating conditions to maximize the H2 generation rate
1

20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44

Depiction of a consistent reaction mechanism for the photocatalytic system proposed
Abstract
Hydrogen is the ideal candidate to fulfill the growing energy demand in a sustainable manner because
of its high energy content and no emission of greenhouse gases from its combustion. Currently most
of hydrogen generation techniques involve the employment of fossil fuels, with consequent production
of toxic greenhouse gases. The possibility to produce hydrogen by means of photocatalytic processes
using the solar radiation as energy source fits in perfectly with the switch to a more sustainable
energy production. The solar photocatalytic hydrogen generation can be achieved by reforming
organic substances contained in civil or industrial wastewaters. This could allow to combine water
decontamination with production of an energy carrier starting from a renewable source, the solar
radiation.
Within this perspective, a novel nano-TiO2 photocatalytic system based on the solar reforming of
formic acid in presence of cupric ions and chlorides has been investigated. The effect on hydrogen
generation rate of the initial concentrations of formic acid, chloride and cupric ion, and pH values
has been evaluated. For both formic acid and chloride ions, at least up to a starting concentration of
103 mM, the higher the initial concentration, the higher the rate of hydrogen generation. Hydrogen
production has turned out to be noticeably dependent on cupric ion concentration. An almost opposite
behavior has been found varying the starting cupric ion concentration in the range 2.5–20 mM, with
the highest value of hydrogen production rate recorded for Cu(II) initial concentration equal to 5.0
mM. The pH value has been identified to be a crucial parameter of the system. A decrease in hydrogen
production has been also observed rising pH of the solution from 1.0 to 4.0. A characterization of the
solid samples recovered at the end of the runs has been performed by X-ray Diffractometry. These
experimental outcomes have been rationalized within a consistent reaction mechanism able to predict
the system behavior under different operating conditions. This work opens the way to the development
2

45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
of new competitive processes able to use waste organic streams for hydrogen generation through
photacatalytic system based on solar energy.
Keywords: nano-sized TiO2 photocatalyst; hydrogen production; photocatalytic reforming; solar
photocatalytic processes; copper “in-situ” photoreduction.
Corresponding author: laura.clarizia2@unina.it (L. Clarizia) Tel.:+390817682253.
Introduction
The possibility to produce hydrogen through photocatalytic processes from aqueous solutions
containing organics (photoreforming) or from water (photosplitting) has emerged in recent years as
a new field of investigation and has renewed the interests of researchers in photocatalysis [1–3].
However, the simple idea that it is sufficient to remove gaseous oxygen from the photocatalytic
systems [4,5] to favor protons reduction by photogenerated electrons – after radiation absorption by
TiO2 particles – appears only partially true. In fact, most of bare metal oxides used in photocatalysis
(TiO2, ZnO and WO3) exhibit low efficiencies in hydrogen generation mainly due to the relevant
occurrence of recombination reaction between photogenerated holes and electrons [6]. Two different
approaches are generally adopted to reduce the incidence of this recombination reaction:
1)
The use of metals, such as Pt, Pd, Au and Cu, as co–catalysts: once incorporated on the photo–
semiconductor surface, they help promote electrons transfer to protons, which are reduced to
hydrogen [7–9].
2)
The addition to the reacting solution of a sacrificial organic agent which irreversibly reacts
(oxidizes) with positive holes [10,11]. It is particularly interesting to remember that the hole
scavenger may be also represented by an organic aqueous pollutant: in this case, hydrogen generation
may combine with water decontamination [12,13, 34].
3

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
Literature shows that many organic species, such as alcohols [14, 35-37] and sugars [15], have been
so far proposed as hole scavengers.
Moreover, a great enhancement of hydrogen generation has been observed with copper–doped TiO2
catalysts both under UV and near–UV range of radiation [16,17], or under irradiation with both visible
and solar light [18,19]. The results of these investigations have indicated that copper–incorporated
TiO2 photocatalysts show a hydrogen production significantly higher than that obtained with bare
TiO2 [20,21].
Moreover,
copper-doped
TiO2
photocatalysts
have
been
used
for
other
photocatalytic applications too. For example, according to literature indications [33] these materials
are capable of enhancing the sunlight-mediated degradation of organic compounds and indicator
microorganisms in water.
Some of the Authors recently reviewed the intense research activity recorded during last years in the
field of copper–incorporated TiO2 photocatalysts for hydrogen production [22].
In a previous paper [11], a new photocatalytic system for hydrogen generation was proposed by some
of the Authors. The system is based on the photoreforming of an organic species (formic acid) in an
aqueous TiO2 suspension containing copper (in different oxidation states) and in the presence of
chloride ions, under artificial UV–Vis radiation. The adoption of such a system was stimulated by
literature indications [23] on the tendency of zero-valent copper to undergo a photocorrosive process,
at very low pH (0.0–2.0), in presence of chloride ions and under UV–solar simulated radiation. As
reported by Eisel et al. [23], the photo-oxidation of zero-valent copper in presence of chloride ions
(ꢀꢁ–ꢀꢂ) results into hydrogen generation through proton reduction (ꢀꢃ):
ꢎ
ꢊꢋ,ꢌꢍ
ꢀ
ꢁ
Clꢄꢅꢆꢇ + Hꢈꢅꢉ ꢏ⎯⎯⎯ꢐ Cl•ꢅꢆꢇ + H•
ꢅꢉ,ꢅꢆꢇ
Cl•ꢅꢆꢇ + Cuꢑ → CuClꢅꢉ,ꢅꢆꢇ
H•ꢅꢉ,ꢅꢆꢇ + H•
ꢀ
ꢂ
→ Hꢂ ꢒ
ꢀ
ꢃ
( )
ꢅꢉ,ꢅꢆꢇ
91
92
In addition, in this system hydrogen can be produced also by photolysis of cuprous complexes formed
in the solution (ꢀ –ꢀꢔ) [24]:
ꢓ
4

ꢄ
Cuꢁꢈ + 2Clꢄ CuClꢂ
ꢀ
ꢓ
ꢂꢄ
Cuꢁꢈ + 3Clꢄ ⇄ CuClꢃ
ꢀꢕ
ꢊν
ꢄ
ꢀ
CuClꢂ + Hꢈ ꢏ⎯⎯ꢐ Cuꢂꢈ + 2Clꢄ + eꢄ
ꢖ
ꢇ
ꢊν
ꢂꢄ
ꢀ
CuClꢃ + Hꢈ ꢏ⎯⎯ꢐ Cuꢂꢈ + 3Clꢄ + eꢄ
ꢗ
ꢇ
ꢘꢈ + eꢄ → ꢁꢂ Hꢂ(ꢒ)
ꢔ
ꢀ
ꢇ
93
94
A continuous production of hydrogen could be thus imagined coupling the last system (Figure 1,
Subsystem 2) with another one (Figure 1, Subsystem 1). The resulting complete system allows the
photocatalytic reduction of cupric ions to zero-valent copper in presence of solid TiO2 – and the
oxidation of an organic species, such as formic acid, used as sacrificial agent – followed by a copper
reoxidation with a simultaneous hydrogen generation.
95
96
97
98
The complete system makes possible, at least in principle, the generation of hydrogen through the
photoreforming of an organic species which could also be contained in wastewaters.
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
Although in the previous work [11] the possibility to generate hydrogen through the complete system
was presented as proof of concept, the data collected so far indicate that hydrogen production cannot
only be attributed to a simple combination of Subsystems 1 and 2. In fact, hydrogen generation was
observed for prolonged reaction times only for particular initial values of some experimental
variables, such as the following: formic acid concentration, chloride ion concentration, total dissolved
copper concentration, starting oxidation state of copper (zero-valent or cupric ions) and TiO2 load.
For all the experimental conditions tested in which initially zero-valent copper was used starting from
the complete system, no hydrogen production was recorded during the run, although its formation
only due to Subsystem 2 was surely observed (for this uncertainty a dotted line is adopted to indicate
this pathway in Figure 1). On the other hand, the formation of hydrogen was observed in some runs
starting from copper initially present as cupric ions and TiO2 load not higher than 100 mg/L. However,
in the previous work [11] no detailed investigations were carried out to clearly identify for these
variables the concentration range which can guarantee the best performances of the complete system.
In particular, no indications were reported on the system reactivity over a wide concentration range
5

114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
neither of formic acid, chloride or total cupric ions (separately or simultaneously added to the reacting
system), nor of pH of the solution.
The present work aims at fulfilling this lack of information in order to provide the best conditions
which maximize the hydrogen generation rate, which is proportional to the maximum efficiency of
hydrogen production.
Experimental
Procedures and analytical methods
Photocatalytic runs have been carried out in an annular glass batch reactor (V=300 mL) covered with
a layer of aluminum foil. On the top of the reactor, an inlet has allowed to feed reactants and nitrogen
gas, and an outlet has been used to collect liquid and gaseous samples at different reaction times.
The reactor has been endowed with a high–pressure mercury vapor lamp by Helios Italquartz (power
input: 125 W), principally emitting at 305, 313 and 366 nm (manufacturer’s data), located inside a
glass cooling jacket in the center of the reactor and surrounded by the reacting solution. The effective
radiative powers of the lamp (Pλ) at 305, 313 and 366 nm are 4.69•10-5, 8.11•10-5 and 2.59•10-4 E/min,
respectively. The reactor has been cooled at 25 °C during each run by means of a thermostatic bath
(Falc GTR 90).
For each photocatalytic run, a proper amount of commercial nano-TiO2 anatase has been added
directly to an unbuffered aqueous solution (V=300 mL), magnetically stirred, containing formic acid
as sacrificial agent.
The pH of the solution has been successively regulated by means of perchloric acid and sodium
hydroxide additions and monitored by means of an Orion 420Aþ pH–meter (Thermo). The resulting
suspension has been fed into the batch reactor under magnetic stirring.
In order to avoid the undesired reaction of dissolved oxygen with photogenerated electrons, before
starting the photocatalytic runs, a nitrogen stream has been bubbled into the solution for 30 minutes
for removing atmospheric oxygen.
6

140
141
After this period, cupric sulfate pentahydrate and sodium chloride have been quickly added to the
mixture. Throughout the experiments, nitrogen has been continuously fed at a flow rate (ꢙꢚ ) of 0.3
ꢛ
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
L/min to prevent any entrance of air into the reactor. Indeed, as previously reported [37], some Cu-
TiO2-based materials suffer from a dramatic copper leaching under oxidizing and acidic conditions.
The liquid samples, collected at different reaction times, have been quickly filtered on regenerated
cellulose filters (pore diameter 0.20 μm, Scharlau) and the filtrate has been used to measure dissolved
copper and formic acid concentrations. The gaseous samples have been recovered from the reactor
outlet in Tedlar gas sampling bags and then used to evaluate hydrogen concentration.
Dissolved copper concentration has been measured by means of a colorimetric method using an
analytical kit (Macherey–Nagel) based on oxalic acid bis–cyclohexylidene hydrazide (cuprizone) [4].
A UV/Vis spectrometer (Cary 100 UV–Vis, Agilent) has been employed for the measurements at a
wavelength of 585 nm. The concentration of formic acid has been measured by means of HPLC
analysis. For this purpose, the HPLC apparatus (Agilent 1100) has been equipped with a UV–Vis
detector (λ=210 nm) and a Alltech OA–1000 column, using a mobile phase of sulphuric acid 2.5•10-
3 mol/L, flowing at 0.8 mL/min.
Hydrogen concentration has been measured by a gas-chromatograph (Agilent 7820A) equipped with
a HP–PLOT Molesieve 5A column (Agilent) and a TCD detector using argon as carrier gas. Each
photocatalytic run has been carried out in duplicate.
Materials
Cupric and chloride ions have been introduced in the system as cupric sulphate pentahydrate,
(CuSO4•5H2O) and sodium chloride, respectively. Formic acid, copper sulphate, perchloric acid,
sodium perchlorate and potassium sulfate have been purchased from Sigma Aldrich and used as
received.
TiO2 catalyst has been purchased from Sigma Aldrich (pure anatase phase, average size 25-70 nm).
7

166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
Catalyst recovery, cleaning and characterization
At the end of the run, the suspension has been allowed to settle overnight under inert atmosphere
producing two distinct layers: aqueous solution on the top and solid particles deposited on the bottom.
After removal of most part of the aqueous solution by decanting it, the remaining part has firstly been
washed with degassed water and then evaporated by means of a gentle inert flow.
The X-ray diffraction (XRD) patterns of the solid samples have been recorded by a Bruker D2
Phaser Diffractometer employing a CuKα source (40 kV, 30 mA) and operating at diffraction angles
ranging between 20 and 80° 2 with a scan rate of 0.02° 2 s−1.
Results and discussion
In Figure 2 is shown an example of the results collected during a typical photocatalytic run starting
from a suspension containing cupric ions (5.0 mM), chloride ions (103 mM), formic acid (103 mM)
and TiO2 (102 mg/L). The hydrogen production rate increases until a value of about 8 μmol/min is
reached and slightly reduces after the first hour of reaction time, whereas the total dissolved copper
concentration keeps approximately constant throughout the experiment. The idea that the trend
observed is due to a decrease of formic acid concentration cannot be demonstrated, since no
appreciable decreasing of formic acid concentration has been detected by HPLC analysis (data not
shown).
Effect of formic acid concentration
As reported in Figures 3a–b, the effect of changing the initial formic acid concentration towards the
hydrogen production rate and Cu(II) consumption has been investigated between 10 mM and 5.0103
mM. It is evident that, keeping constant initial cupric ion and chloride concentrations, TiO2 load and
pH, the higher the formic acid concentration, the higher the maximum hydrogen production rate
observed. The values of hydrogen production rate vary from 0.7 to 8.3 μmol/min, at least until formic
acid concentration is equal to 103 mM. Since hydrogen production proceeds through an initial
8

192
193
194
reduction of cupric ions (ꢀꢜ), made possible by a contemporary oxidation of formic acid (ꢀꢁꢑ), a
monotonic increase of hydrogen generation rate is clearly expected for higher concentrations of
formic acid according to the results previously reported [11]:
ꢞ
ꢞ
ꢝ
ꢝ
ꢟꢠ
ꢟꢠ
ꢀ
ꢜ
( )
( )
Cu II ꢏ⎯⎯⎯⎯ꢐ Cu I ꢏ⎯⎯⎯⎯ꢐ Cu(0)
ꢢ
ꢊ
ꢀ
ꢡꢠ
•ꢄ
ꢁꢑ
HCOOH ꢏ⎯⎯⎯⎯⎯ꢐ COꢂ + 2Hꢈ
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
As previously reported, the concentration of the organic species influences the rate of cupric ion
reduction during the sacrificial photocatalytic process [25]. It can be thus supposed that reduced
copper species, i.e. cuprous ones formed through reaction ꢀꢜ, in presence of chlorides and under UV
irradiation immediately reoxidize (ꢀ − ꢀ ), therefore ejecting electrons (eꢄ ) in the solution and
ꢇ
ꢖ
ꢗ
allowing a prompt reduction of protons (ꢀꢔ).
Only a moderate reduction, if any, of total dissolved copper and low hydrogen generation rates have
been recorded during the runs with the lowest initial formic acid concentration (Figure 3, full
diamonds and triangles). For a starting formic acid concentration of 103 mM, a reduction in dissolved
copper of about 20% has been recorded within a reaction time of 300 minutes (Figure 3, full circles).
It is worth considering that the formation of hydrogen proceeds through two consecutive steps: a first
one (ꢀꢜ) in which a preventive reduction of cupric ions to cuprous ones occurs with a contemporary
formic acid oxidation (ꢀꢁꢑ), and a second one consisting in the photolysis of some chloride complexes
of cuprous ions (ꢀꢖ − ꢀꢗ). It can be thus hypothesized that any increase in initial concentration of
formic acid may result into a faster reduction of cupric ions. The results collected indicate that, for
formic acid concentrations up to 103 mM, the rate determining step is represented by cupric ion
reduction. It is therefore enough to accelerate the determining step to have a faster hydrogen
generation for formic acid concentrations between 10 mM and 103 mM.
At the highest formic acid concentration tested, equal to 5.0103mM (Figure 3, cross symbols),
dissolved copper has been completely reduced after 180 minutes of reaction, casting doubt on the
catalytic role of copper. In other words, it seems that the complete system (Figure 1) is characterized
by a certain tendency of zero-valent copper to precipitate and be not capable of undergoing further
9

216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
reoxidations; therefore, zero-valent copper can be deemed completely unreactive for increasing
reaction times. It could be reasonably considered that, under the adopted conditions (formic acid
concentration equal to 5.0103mM), the photocatalytic reduction rate of cuprous ions is by far higher
(
)
than the reoxidation rate of zero-valent copper ꢀꢁ − ꢀꢂ , provided that this reoxidation occurs.
Alternatively, it should be taken into account the possibility that the complete system (Figure 1)
inhibits zero-valent copper oxidation together with hydrogen formation. An attempt to throw light on
this specific point has been made by investigating hydrogen generation in the system zero-valent
copper/formic acid/chlorides/nano-TiO2/UV irradiation at pH=1.0. The data indicate values of
hydrogen generation as lower as those recorded in the system formic acid/chlorides/nano-TiO2/solar
UV radiation at pH=1.0, without any significant accumulation of dissolved copper as Cu(I) or Cu(II)
ions. This result clearly indicates that a photoxidation of zero-valent copper in presence of chloride
ions, formic acid and nano-TiO2 does not occur at all.
Effect of chloride ion concentration
The effect of changing the initial chloride ion concentration has been investigated in the range
102−2.0103 mM (Figures 4a–b).
For fixed formic acid concentration, any increase in addition of chloride ions to the reacting solution
results into a higher hydrogen generation rate (the maximum value recorded is 8.3 μmol/min for a
chloride ion concentration of 103 mM). However, it is interesting to stress that, contrary to what
observed in the runs with variable formic acid concentration, for the lowest value of chloride ion
tested (102 mM) a significant reduction of total dissolved copper has been recorded, along with the
precipitation of a reddish–brown solid on TiO2.
These results may be explained by supposing that, if chloride ion concentration is reduced, Cu(I)
reoxidation through chloride complexes photolysis (ꢀꢖ − ꢀꢗ) is slowed down and becomes the rate
determining step. As a consequence of this behavior, the rate of hydrogen generation decreases too.
10

241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
On the other hand, a reduction in hydrogen production rate has been observed for further increases in
chloride ion concentration from 103 mM up to 2.0103 mM (Figure 4, full triangles). It is known that
at rising chloride ion content both Cu(I) and Cu(II) may form different complexed species, such as
-
2-
-
CuCl2 and CuCl3 (for cuprous ions) [26] or CuCl+, CuCl2 and CuCl3 (for cupric ions) [27]. The
formation of Cu(II)–chloride complexes results into the occurrence of a shielding effect which
reduces the radiation absorption by Cu(I)–chloride complexes. Indeed, both cuprous and cupric
complexes are capable of absorbing in the same radiation range of the lamp used (305–366 nm) with
comparable extinction coefficients, although those belonging to Cu(I) complexes type are slightly
higher (Figure 5).
For higher chloride ion concentrations, an increase in concentration of all chloride complexes occurs,
-
causing two opposite effects: (i) a rise in concentration of cuprous-chloride complexes (CuCl2 and
2-
CuCl3 ), which absorb and photolyze at wavelengths higher than 290 nm [26]; (ii) an increase in
-
concentration of cupric–chloride complexes (CuCl+, CuCl2 and CuCl3 ) and, consequently, of their
shielding effect. It can be considered that the shielding effect prevails for chloride concentrations
higher than 103 mM, with a consequent reduction in hydrogen generation rate. However, only the
availability of a suitable kinetic model could help to completely throw light on this issue.
So as to determine the nature of the precipitate, a sample recovered at the end of a run at pH=1 washed
accordingly with the procedure reported in the experimental has been submitted to a XRD
measurement (Fig 6). Two peaks at 2 equal to 43°3’ and 50°5’ which can be ascribed to zero-valent
copper have been observed. Therefore, it can be concluded that the solid recovered is represented by
zero-valent copper.
Effect of cupric ion concentration
The influence of initial cupric ion concentration on hydrogen production rate and dissolved copper
concentration is reported in Figure 7. Distinct behaviors have been recorded when initial cupric ion
concentration changes from 2.5 mM to 5.0, 10.0 and 20.0 mM (Figures 7a–b).
11

267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
A delayed hydrogen production and a lower decrease in dissolved copper have been observed at rising
initial cupric ion concentration from 5.0 to 20 mM. A complete reduction of cupric ions and a reddish
solid precipitation within 300 minutes of reaction have been recorded only for the run starting from
a cupric ion concentration of 2.5 mM. XRD measurements have demonstrated that the solid is
represented by zero-valent copper (data not shown).
To explain these observations it has to be considered that at increasing cupric ion concentration
between 5.0 and 20 mM, the shielding effect due to the formation of Cu(II)–chloride complexes
becomes more significant and lowers the rate of hydrogen production. When cupric ion concentration
is hugely reduced through the reaction between cupric ions and photoelectrons (ꢀꢜ ) (with a significant
increase in concentration of Cu(I) in the form of Cu(I)–chloride complexes), the rate of hydrogen
production reaches a maximum value (Figure 7).
For initial cupric ion concentration equal to 2.5 mM, the less important shielding effect promotes a
faster hydrogen production, but also a rapid photocatalytic reduction of Cu(I) to zero-valent copper
(ꢀꢜ ) which does not reoxidize. Indeed, in this case the maximum value of hydrogen production rate
is lower with respect to that recorded starting from a cupric ion concentration of 5.0 mM.
282
283
A more careful examination of these data indicates that different quantities of hydrogen are produced
per moles of precipitated copper (ꢣꢤ / ꢣꢥꢦ(ꢑ), Table 1) for reaction times of 300 and 600 minutes
ꢛ
284
285
respectively, depending on the initial concentration of cupric ion adopted. The best results, consisting
in highest values of ꢣꢤ / ꢣꢥꢦ(ꢑ), have been recorded in the run with a starting cupric ion concentration
ꢛ
286
287
288
289
290
of 5.0 mM for both reaction times. The results collected indicate that a complete precipitation of
dissolved copper happens at about 600 minutes of reaction. This demonstrates a certain tendency of
the catalytic system to deactivate, as already observed in the run starting from a cupric ion
concentration of 2.5 mM. In other words, it is clear that, even for different reaction times depending
on the initial cupric ion concentration and still in presence of significant quantities of formic acid, the
12

291
292
system becomes no more capable of producing hydrogen. This deactivation of the system can be
considered as a consequence of the complete precipitation of zero-valent copper.
ꢣꢤ / ꢣꢥꢦ(ꢑ)
ꢣꢤ / ꢣꢥꢦ(ꢑ)
[Cu(II)]0
mM
2.5
ꢛ
ꢛ
300 min
1.8
600 min
1.8
5.0
3.8
3.5
10
2.7
1.8
-
20
1.7
293
294
Table 1 – Effect of cupric ion concentration: correlation between micromoles of H2 generated (ꢣꢤ )
ꢛ
295
296
and micromoles of Cu(0) precipitated (ꢣꢥꢦ(ꢑ)) after 300 and 600 minutes of reaction at pH=1.0, T=25
°C, [HCOOH]0=103 mM, [NaCl]0=103 mM, TiO2 load=102 mg/L.
297
298
299
300
301
302
303
304
305
306
307
308
309
Effect of pH of the solution
Since a dependence of hydrogen generation rate on protons concentration (and pH) may be assumed,
the best results could be expected under low pH condition. Indeed, proton reduction is a fundamental
reaction through which the process develops. However, in most of the researches appeared in
literature on copper–based TiO2 photocatalysts for hydrogen generation, the best performances are
reported for basic or strong basic condition [28–30]. The prevailing explanation seems to rely on the
detrimental effect that acidic conditions exert on the catalyst stability, with a continuous copper
leaching into the solution [31,32].
To assess the effect on the system reactivity of pH in the solution, two additional photocatalytic runs
have been performed at pH 2.5 and 4.0 respectively, for both chloride ion and formic acid initial
concentrations equal to 103 mM, cupric ion concentration at 5.0 mM and TiO2 load equal to 102 mg/L.
The results collected during these runs are shown in Figure 8a-b.
13

310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
It is interesting to observe that both at pH 2.5 and 4.0 hydrogen production appears to be depressed
and reaches zero after about 120 minutes of reaction. For the same reaction time total dissolved copper
reduces to zero and a black solid is formed on TiO2 nano-particles.
Also in this case, an attempt to characterize the nature of copper deposited on TiO2 nanoparticles has
been made. The solid sample has been firstly washed as reported in the experimental and then
submitted to X-ray diffraction analysis. XRD spectrum (Figure 9) shows two peaks at 2 equal to
43°3’ and 50°5’, indicating the presence of zero-valent copper in agreement with what reported by
Li et al. [38].
As discussed above, in the system studied hydrogen production proceeds through a complex pathway
made up of several reaction steps, the last and faster being proton reduction by solvated electrons.
When pH is increased in the range 1.0-4.0, the resulting reduction in proton concentration slows down
the rate of solvated electrons capture by protons. Therefore, in such cases solvated electron capture
by protons becomes the rate-controlling step, thus lowering the hydrogen generation rate.
In these conditions, a higher availability of solvated electrons may be expected in solution with a
consequent enhancement in Cu(I) and Cu(II) reduction. In other words, copper is continuously
separated from the solution through a reduction of dissolved copper species until the system becomes
completely unreactive.
The behavior of the system investigated may be depicted by means of the scheme illustrated in Figure
10.
Hydrogen formation in a system containing copper in the initial form of cupric ions is strictly
connected to the reduction of cupric ions to cuprous ones (Step A, Figure 10) and the photolysis of
Cu(I)–chloride complexes (Step B, Figure 10). In fact, during the photolytic process electrons (ꢧꢄ )
ꢨ
are ejected into the solution and protons promptly react with them but not with the photogenerated
ones (ꢧꢄꢥꢩ). Chloride complexes provide an important reservoir of Cu(I) in solution; therefore, in
the presence of low Cl– concentrations, Cu(I) more easily reduces to Cu(0) which precipitates (Step
C, Figure 10). When pH is increased, the low proton concentration makes available a higher
14

336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
concentration of solvated electrons which reduce both Cu(II) and Cu(I) and lead to the deposition of
a black solid, formed by zero-valent copper (Step C, Figure 10).
Conclusions
The present investigation proposes a new competitive process for hydrogen generation through the
system cupric ions/formic acid/chloride ions/nano-TiO2/UV-solar radiation. The efficiency of the
process in terms of hydrogen production has been studied at varying the initial concentrations of
formic acid, chloride and cupric ion and pH.
Hydrogen generation has revealed similar dependences on formic acid and chloride ion
concentrations. The rate of hydrogen production increases up to a formic acid concentration of 103
mM. On the other hand, a huge decrease in the maximum value of hydrogen production rate has been
observed for a formic acid concentration equal to 5.0103 mM, with a contemporary precipitation of
zero-valent copper. With regard to chloride ions in solution, an increase in hydrogen generation rate
has been observed up to a chloride concentration of 103 mM, with a clear decrease for higher
concentrations.
Hydrogen production is also strongly dependent on cupric ion concentration. Initial increases in
cupric ion concentration have resulted in a reduction of hydrogen generation rate. An early loss of
reactivity of the system has been recorded during the run at the lowest value of cupric ion tested (2.5
mM), with a complete precipitation of cupric ions as zero-valent copper within a reaction time of 300
minutes. A similar behavior has been observed for higher starting concentration of cupric ion, but at
longer reaction times. No re-oxidation of zero-valent copper has been observed in the complete
system during the present investigation.
The increase in pH of the solution from 1.0 to 2.5 to 4.0 has shown a decrease in hydrogen generation
rate and the precipitation of zero-valent copper.
15

360
361
362
363
A consistent reaction mechanism able to predict the system behavior under different operating
conditions has been proposed. This study lays the groundwork for the future development of a suitable
kinetic model that will completely throw light on this issue.
364
Acknowledgments
365
366
The authors wish to thank Dr. Luciano Cortese from IRC-CNR, Naples, for the help in performing
XRD analysis.
16

367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
References
[1] Y. Ma, X. Wang, Y. Jia, X. Chen, H. Han, C. Li, Chem. Rev. 114 (2014) 9987–10043.
[2] M. Ni, M.K.H. Leung, D.Y.C. Leung, Y.C. Sumathy, Renew. Sust. Energ. Rev. 11 (2007) 401–
25.
[3] R. Abe, J. Photochem. Photobiol. C. 11 (2010) 179–209.
[4] M. Canterino, I. Di Somma, R. Marotta, R. Andreozzi, Water. Res. 42 (2008) 4498–506.
[5] N.S. Foster, R.D. Noble, C.A. Koval, Environ. Sci. Technol. 27 (1993) 350–356.
[6] X. Chen, S. Shen, L. Guo, S.S. Mao, Chem. Rev. 110 (2010) 6503–6570.
[7] G. Chiarello, E. Selli, L. Forni, Appl. Catal. B 84 (2008) 332.
[8] Y. Li, G. Lu, S. Li, Appl. Catal. A 214 (2001) 179–185.
[9] T. Sreethawong, S. Yoshikawa, Catal. Commun. 6 (2005) 661–668.
[10] D.I. Kondarides, V.M. Daskalaki, A. Patsoura, X.E. Verykios, Catal. Lett. 122 (2008) 26–32.
[11] V. Lanese, D. Spasiano, R. Marotta, I. Di Somma, L. Lisi, S. Cimino, R. Andreozzi, Int. J.
Hydrogen Energ. 38 (2013) 9644–9654.
[12] R. Priya, S. Kanmani, Environ. Technol. 34 (2013) 2817–2823.
[13] W. Zhang, Y. Li, C. Wang, P. Wang, Q. Wang, Water. Res. 47 (2013) 1480–1490.
[14] V.M. Daskalaki, P. Panagiotopoulou, D.I. Kondarides, Chem. Eng. J. 170 (2011) 433–439.
[15] M.R. St. John MR, A.J. Furgala, A.F. Sammells, J. Phys. Chem. 87 (1983) 801–805.
[16] H.J. Choi, M. Kang, Int. J. Hydrogen Energ. 32 (2007) 3841–3848.
[17] J.P Yasomanee, J. Bandara, Sol. Energy Mater. Sol. Cells 92 (2008) 348–352.
[18] L.S. Yoong, F.K. Chong, B.K. Dutta, Energy 34 (2009)1652–1661.
[19] K. Lalitha, G. Sadanandam, V. Durga Kumari, M. Subrahmanyam, B. Sreedhar, N.Y. Hebalkar,
J. Phys. Chem. C 114 (2010) 22181–22189.
[20] J. Bandara, C.P.K. Udawatta, C.S.K. Rajapakse, Photochem. Photobiol. Sci. 4 (2005) 857–861.
[21] M.K. Jeon, J.W. Park, M. Kang, J. Ind. Eng. Chem. 13 (2007) 84–91.
[22] L. Clarizia, D. Spasiano, R. Marotta, I. Di Somma, R. Andreozzi, D.D. Dionysiou, Int. J.
Hydrogen Energ. 30 (2014) 16812–16831.
[23] C. Eisel, M.W. Tausch, J. Photoch. Photobio. A 128 (1999) 151–154.
[24] O. Horváth, Coord. Chem. Rev. 135/136 (1994) 303–324.
[25] M.I. Litter, Appl. Catal. B 23 (1999) 89–114.
[26] D.D. Davis, K.L. Stevenson, C.R. Davis, J. Am. Chem. Soc. 100 (1978) 5344–5349.
[27] N. Zhang, Q. Zhou, X. Yin, D. Zeng, J. Solution Chem. 43 (2014) 326–339.
17

400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
[28] R. Brahimi, Y. Bessekhouad, A. Bouguelia, M. Trari, J. Photoch. Photobio. A 186 (2007) 242–
247.
[29] Z. Jin, X. Zhang, Y. Li, S. Li, G. Lu, Catal. Commun. 8 (2007) 1267–1273.
[30] Y. Wu, G. Lu, S. Li, Catal. Lett. 133 (2009) 97–105.
[31] S. Xu, D.D. Sun, Int. J. Hydrogen Energ. 34 (2009) 6096–6104.
[32] S. Xu, J. Ng, X. Zhang, H. Bai, D.D. Sun, Int. J. Hydrogen Energ. 35 (2010) 5254–5261.
[33] M.B. Fisher, D.A. Keane, P. Fernández-Ibánez, J. Colreavy, S.J. Hinder, K.G. McGuigan, S.C.
Pillai, Appl. Catal B: Environ 130–131 (2013) 8-13.
[34] M. Cargnello, A. Gasparotto, V. Gombac, T. Montini, D. Barreca, P. Fornasiero, Eur. J. Inorg.
Chem. (2011) 4309–4323.
[35] T. Montini, V. Gombac, L. Sordelli, J.J. Delgado, X. Chen, G. Adami, P. Fornasiero,
ChemCatChem 3 (2011), 574–577.
[36] D. Barreca, P. Fornasiero, A. Gasparotto, V. Gombac, C. Maccato,T. Montini, E. Tondello,
ChemSusChem 2 (2009) 230–233.
[37] V. Gombac, L. Sordelli, T. Montini, J.J. Delgado, A. Adamski, G. Adami, M. Cargnello, S.
Bernal, P. Fornasiero, J. Phys. Chem. A 114 (2010) 3916–3925.
[38] Z. Li, J. Liu, D. Wang, Y. Gao, J. Shen, Int. J. Hydrogen Energ. 37 (2012) 6431–6437.
418
419
420
421
Captions
Figure 1 – Schematic illustration of the complete system for hydrogen generation through nano-TiO2
photocatalytic reforming of formic acid in presence of copper species and chloride ions.
422
423
424
425
426
427
428
429
430
Figure 2 – Hydrogen production rate () and dissolved copper normalized concentration () at pH=1.0
(with HClO4), [HCOOH]0=103 mM, [Cu(II)]0=5.0 mM, [NaCl]0=103 mM, TiO2 load=102 mg/L.
Figure 3 – Effect of formic acid concentration: hydrogen production rate (a) and dissolved copper
normalized concentration (b) at pH=1.0 (with HClO4), [Cu(II)]0=5.0 mM, [NaCl]0=103 mM, TiO2
load=102 mg/L, [HCOOH]0 (mM): (+) 5.0103, () 103, () 5.0102, () 102, (▲) 10.
18

431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
Figure 4 - Effect of chloride ion concentration. Hydrogen production rate (a) and dissolved copper
normalized concentration (b) at pH=1.0 (with HClO4), [Cu(II)]0=5.0 mM, [HCOOH]0=103 mM, TiO2
load=102 mg/L, [NaCl]0 (mM): (▲) 2.0103 mM, () 103 mM, () 5.0102 mM, () 102 mM.
Figure 5 – UV spectra of Cu(I) and Cu(II) chloride complexes (data source: Davis et al., 1978 [25]).
Figure 6 – XRD patterns registered for a solid sample recovered at the end of a photocatalytic run at
pH=1, bare TiO2 (anatase) and Cu(0) powders.
Figure 7 – Effect of cupric ion concentration. Hydrogen production rate (a) and dissolved copper
normalized concentration (b) at pH=1.0 (with HClO4), [HCOOH]0=103 mM, [NaCl]0=103 mM, TiO2
load=102 mg/L, [Cu(II)]0 (mM): (▲) 2.5, () 5.0, () 10, () 20.
Figure 8 – Effect of pH. Hydrogen production rate (a) and dissolved copper normalized concentration
(b), [Cu(II)]0=5.0 mM, [HCOOH]0=103 mM, [NaCl]0=103 mM, TiO2 load=102 mg/L, pH:  1.0, 
2.5,  4.0.
Figure 9 – XRD patterns registered for a solid sample recovered at the end of a photocatalytic run at
pH=4, bare TiO2 (anatase) and Cu(0) powders.
Figure 10 – Schematic illustration of the reaction mechanism for the investigated system.
455
456
457
19

458
459
460
Figure 1
461
462
Figure 2
463
464
20

465
466
467
Figure 4
468
21

469
470
Figure 5
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
22

486
487
488
489
Figure 6
Zero valent copper
TiO anatase
2
Catalyst (pH = 1)
20
30
40
50
60
70
80
2

490
491
23

492
Zero valent copper
TiO anatase
2
Catalyst (pH = 4)
20
30
40
50
60
70
80
2

493
494
Figure 9
495
24

496
497
25