Towards an improved process for hydrogen production: The chemical-loop reforming of ethanol

Article abstract of DOI:10.1039/c5gc02139a

M-modified ferrospinels with the formula M_0.6Fe_2.4O_y (M = Co, Mn or Co/Mn) were employed as ionic oxygen and electron carrier materials for an alternative sustainable route to produce hydrogen via chemical-loop reforming of ethanol. The new materials were tested in terms of both redox properties and catalytic activity to generate hydrogen by oxidation with steam, after a reductive step carried out with ethanol. In addition, the research includes in situ DRIFTS and in situ XPS studies that allowed the extraction of information at the molecular level and following surface changes within the reduction/re-oxidation processes during ethanol chemical-loop reforming. It was found that Co(ii)-incorporation in spinels effectively improves decomposition/oxidation of ethanol, however a greater amount of coke is accumulated. On the other hand, addition of Mn(ii) into the system helps to significantly reduce the amount of coke and hence to avoid fast deactivation of the material. Thus, the behavior of Co_0.3Mn_0.3Fe_2.4O_y was shown to be the most promising one, as this material forms less coke during the reduction step, and consequently less CO_x is generated during the re-oxidation step with water, nevertheless a high hydrogen yield is maintained.

Full text of DOI:10.1039/c5gc02139a

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: O. Vozniuk, S.
Agnoli, L. Artiglia, A. Vassoi, N. Tanchoux, F. Di Renzo, G. Granozzi and F. Cavani, Green Chem., 2015,
DOI: 10.1039/C5GC02139A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/greenchem

Page 1 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
Towards an improved process for hydrogen production: the chemical-loop reforming of
ethanol
a
b
b
a
c
Olena Vozniuk , Stefano Agnoli , Luca Artiglia , Andrea Vassoi , Nathalie Tanchoux , Francesco
c
b,
a, d,
Di Renzo , Gaetano Granozzi †, Fabrizio Cavani
†
a.
Dipartimento di Chimica Industriale “Toso Montanari”, Università di Bologna, Viale
Risorgimento 4, 40136 Bologna, Italy.
b.
c.
Department of Chemical Sciences, Via F. Marzolo 1, 35131 Padova, Italy
Institut Charles Gerhardt Montpellier, ENSCM, 8 Rue Ecole Normale, F-34296
Montpellier, France
d.
INSTM, Research Unit of Bologna, Via G. Giusti 9, Firenze, Italy
†
e-mail: fabrizio.cavani@unibo.it; gaetano.granozzi@unipd.it
x
Non-stoichiometric M-modified ferrospinels with formula M0.6Fe2.4O (M=Co, Mn or Co/Mn) were
employed as ionic oxygen and electron carrier materials for an alternative sustainable route to
produce hydrogen via chemical-loop reforming of ethanol. The new materials were tested in terms
of both redox properties and catalytic activity to generate hydrogen by oxidation with steam, after a
reductive step carried out with ethanol. In addition, the research includes in-situ DRIFTS and in-situ
XPS studies that allowed to extract information at molecular level and to follow surface changes
within the reduction/re-oxidation processes during ethanol chemical-loop reforming. It was found
that Co(II)-incorporation in spinels effectively improves decomposition/oxidation of ethanol,
however a greater amount of coke is accumulated. On the other hand, addition of Mn(II) into the
system helps to significantly reduce the amount of coke and hence to avoid a fast deactivation of the
y
material. Thus, the behavior of Co0.3Mn0.3Fe2.4O was shown to be the most promising one, as this
material forms less coke during the reduction step, and consequently less CO is generated during
x
the re-oxidation step with water, nevertheless a high hydrogen yield is maintained.
Introduction
Current processes for the industrial production of hydrogen are based on the reforming of natural
gas/naphtha or coil gasification. Such processes are highly energy demanding. Hence, less energy-
intensive and more sustainable technologies, exploiting renewable feedstock (such as biomass, bio-

Green Chemistry
Page 2 of 33
View Article Online
DOI: 10.1039/C5GC02139A
alcohols and water)and renewable primary energy sources (e.g. sunlight, wind, wave or hydro-
power), appear to be very attractive for both industry and consumer applications. A variety of new
technologies offering a non-fossil based route for hydrogen production are in a different stage of
1
–10
development, and each offers unique opportunities, benefits and challenges.
Chemical-loop steam reforming is one of the possible ways to produce H
conventional or renewable sources. Differently from classical reforming, separation costs can be
avoided by splitting the process into two alternated steps in order to detach H and CO streams.
One of the advantages of this technology relates to its feedstock flexibility. The chemical-loop
2
starting from either
2
x
11
process can be performed using a number of different reductants, such as gas resulting from coal
12–14
15
16–18
or biomass
gasification, light hydrocarbons reforming, methane,
CH and CH
4
/CO
2
4
/H
2
18,19
20,21
22
23,24
mixture,
pyrolysis oil,
methanol and pure H
2
(as a method for H storage).
2
Furthermore, the nature of the oxide used as the ionic oxygen and electron carrier and the reaction
conditions are important parameters since they determine the potential for low costs and high
efficiency of this process in order to have a commercial impact. Among the various proposed metal
oxides for the chemical-loop approach, the mostly studied ones are pure and mixed iron oxides.
Several research groups have been exploring modifications of simple iron oxide (Fe
3
O
4
2 3
and Fe O )
25
26
in order to prevent deactivation, to lower the operating temperature and to increase the structural
2
7,28
29,30
stability and reducibility,
the reaction rate for oxidation and total efficiency of the process.
3
1,32
Several studies were dedicated to different metal additives to iron oxide.
In addition, ternary
33,34
metal systems have also been considered in the search for a better synergetic effect.
This paper deals with an investigation of spinel type mixed oxides (modified ferrospinels), as the
carriers for the production of clean hydrogen in the so-called chemical-loop reforming of ethanol
(Scheme 1). The choice of ethanol as reducing agent has several advantages: its renewable origin
together with the possibility to decompose at a relatively lower temperature with the formation of
hydrogen-rich mixture.
35-39
In our previous studies,
we were dealing with stoichiometric binary ferrospinels MFe O (M =
2 4
Fe, Co, Ni or Cu), as carrier materials, which have shown promising features for hydrogen
production. The initial experiments were carried out with methanol as reducing agent. According to
the obtained results we concluded that the morphology of the CoFe O spinel (calcined at 450 and
2
4
o
7
50 C) led to differences in methanol anaerobic decomposition and oxidation; however, after a first
35
complete cycle, the sintering effect resulted in a strong modification of these properties. In
addition, a complete re-oxidation of CoFe O to its original oxidation state was not possible using
2
4
only water as the oxidant; thus a third step, with the partial re-oxidation of the material with air was
36
37
needed for a completion of the cycle. The sintering phenomena of magnetite and reaction

Page 3 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
38
39
mechanism together with deactivation phenomena of methanol and ethanol anaerobic
decomposition/oxidation over mixed ferrospinels were also investigated.
The present work is focused on new non-stoichiometric binary and ternary M-modified ferrospinels
(i.e. mixed solid solutions) for the chemical-loop reforming of bio-ethanol that can operate in mild
temperature range for both reduction and re-oxidation steps. Substitution of Fe centers with other
transition metals leads to the formation of an inverse, normal, or mixed spinels with different degree
of inversion. Besides this, such an introduction may strongly modify the redox properties of the
40
resultant ferrospinels.
Several research groups have investigated the effect of various M-additives on the stability and
redox behavior of iron oxide for chemical hydrogen storage using, amongst others, Pd, Pt, Rh, Ru,
4
1
42
Al, Ce, Ti, Zr and Al, Cr, Zr, Ga, V , Mo . For example it was found out that Pd, Pt, Rh and Ru
additives have an effect of promoting the reduction and lowering the re-oxidation temperature of
iron oxide. At the same time, Al, Ce, Ti, Zr, Cr, Ga and V additives prevent deactivation and
sintering of iron oxide during repeated redox cycles. However, according to our knowledge, a
systematic study on the role of dopants and their effect on the chemical-physical properties of the
resulting mixed M-modified ferrites as future perspective materials for chemical-loop reforming of
ethanol has not been done until now. Thus, the aim of this work is to investigate the behavior of
these new materials in the two-step chemical-loop reforming of bio-ethanol. Particularly the focus is
on the reactivity behavior of ternary materials based on Co-Mn ferrospinels with different Co:Mn
atomic ratios. The special interest in the Co and/or Mn doping is explained by their ability to form
thermodynamically stable spinel oxides which allowed us to both re-obtain the initial spinel phase
upon cycling and increase the stability of the looping material.
Experimental
Materials preparation
M-modified non-stoichiometric ferrospinels with different Co:Mn atomic ratio (where Co:Mn =
1
0
2
1
:0; 0.9:0.1; 0.5:0.5 and 0:1) and nominal composition M
M Fe O were prepared by the co-
.6-x x 2.4 y
precipitation method. Nitrate precursors were used for all of the synthesized materials. The mixed
.
solutions of metal precursors containing 100mL of 1M Fe(NO
3
)
3
9H
2
O, 50mL of 0.5M
.
1
.
2
.
M(NO ) xH O (or 50mL of 0.25M M (NO ) xH O and M (NO ) xH O) were put into a separation
3
2
2
3 2
2
3 2
2
funnel and then drop-by-drop added into the reaction vessel containing 0.5L of 2M NaOH aqueous
solution at the temperature of 50°C under vigorous stirring with a constant speed. The suspensions
were stirred for 2h at the constant temperature of 50°C adding NaOH 3M in order to maintain the
pH above 13. The precipitates were separated by vacuum filtration by means of a Büchner funnel

Green Chemistry
Page 4 of 33
View Article Online
DOI: 10.1039/C5GC02139A
and then washed with about 1.5L of distilled water at room temperature (RT), to remove sodium
and nitrates ions. The washed samples were dried at 120°C in air for 12 h, milled in agate mortar
o
and then calcined in static air at 450°C for 8h with the heating rate of 10 C/min.
Materials characterization
The specific surface area was measured, applying the single-point BET method. The instrument
used was a Carlo Erba Sorpty 1700. For the analysis, around an amount of 0.5 g of the sample was
placed inside the sample holder and then heated at 150°C under vacuum (2 MPa) to release water,
air, CO or other molecules adsorbed. Subsequently the sample was placed in liquid nitrogen, and
2
2
the adsorption of the gaseous N was carried out.
For TEM/EDX characterisation, samples were suspended in ethanol and homogenized using
ultrasonication (for 5 min). The suspension was deposited on a holey-carbon film Cu-grid for TEM
analysis, then dried at 100°C. The analyses were carried out by a microscope TEM/ STEM FEI
TECNAI F20 at 200 keV.
The instrument used for Thermal-Programmed-Reduction (TPR) experiments was the TPR 1100
o
analyzer from Thermo Quest. The samples were first pre-treated at 450 C for 45 min in He flow (30
2
mL/min). Reduction (TPR) was carried out with 5 % H in the Ar flow (30 mL/min).Temperature
o
program started at 50 C and then the temperature was increased up to 800°C (with ramp 10°C/min)
and maintained for 40 min.
Raman spectra were recorded using a Bruker spectrometer with a 633 nm He/Ne laser source (filter:
Plaser ÷10; slit 100 ꢀm) equipped with a microscope (50X lens). Each spectrum was taken with 300
sec exposure time and 1 accumulation.
The powder X-ray diffraction (XRD) patterns were measured on Bruker D8 Advance diffractometer
(
using Cu K
α
radiation at ʎ = 1.5418 Å) with a Bragg-Brentano geometry and equipped with a
o
Bruker Lynx Eye detector. The data were recorded from 5 to 80 2Ɵ diffraction angle with an
acquisition time of 1 s per each step.
Diffuse-Reflectance Infrared FT spectroscopy, coupled with a quadrupole mass spectrometer
(DRIFTS-MS) was used to characterise the surface features of materials. During a standard
procedure, the sample was pre-treated at 450°C in a He flow (5mL/min) for 1 h in order to remove
molecules possibly adsorbed on the material, mainly carbon dioxide and water. Then the sample
o
was cooled down to 30 C, and the IR background was collected. Afterward ethanol was fed at 0.6
mL/min until saturation was reached (as seen by IR and MS; around 20 min) and whereupon the He
flow was maintained until the weakly adsorbed ethanol was evacuated from the cell with loaded
catalyst. During evacuation, the DRIFT and mass spectra were taken to follow stabilization of the

Page 5 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
system. When the mass and DRIFT spectra were stable, the temperature was increased up to 150°C
for 2 min and then the system was cooled down to 30°C to record the spectra. This last step was
o
38
repeated for the other temperatures (150, 200, 250, 300, 350 and 400 C).
X-ray photoemission spectroscopy (XPS) data were acquired using non-monochromatized Al K
(1486.6 eV) X-ray source, and an Omicron EA125 electron analyser. The measurements were
-10
carried out in a custom built ultra-high-vacuum system operating at the base pressure of 5x10
mbar equipped with facilities for sample heating and controlled gas dosing. The spectra were
acquired at RT on drop casted films of powder samples suspended in water. Different sample
o
o
treatments have been carried out: annealing in UHV at 450 C, reduction with ethanol at 450 C
-6
o
-4
(
5x10 mbar) and re-oxidation at 450 C in water ambient (2x10 mbar). For each of four M-
modified ferrospinels Fe 2p, O1s and Co 2p and/or Mn 2p photoemission spectra were recorded
sequentially after introduction in the UHV systems (“as received”), after annealing in UHV for 30
2
min at 450°C, after exposure to ethanol at 450°C and after H O exposure at 450°C.
Reactivity tests of ethanol chemical-loop reforming
Reactivity experiments were carried out by loading 400 mg of the pelletized sample (with particles
diameter ≈0.25 to 0.6 mm), in the fixed-bed quartz flow reactor with an internal diameter of 12mm
and a length of 30 cm; the catalytic bed height was less than 2 mm. The reduction reaction was
carried out by feeding continuously a stream of ethanol vapors (15.6 mol%) in N
also used as a standard; contact time was 0.25 s. For the re-oxidation step, water steam (29.3 mol%)
was fed in a N flow. Reactivity experiments were carried out at different cycling conditions with
2
; the latter was
2
chosen duration of the reduction/re-oxidation steps (20 min red·1cyc/20 min ox·1cyc). The reaction
o
temperature for both steps was 450 C. The products were monitored on-line by an Agilent 3000A
micro-GC with3 parallel columns: (a) a PlotQ column, with He carrier, for the separation of CH
4
,
CO , H O and ethanol; (b) a OV1 column, with He carrier, for the separation of CO , acetaldehyde,
2
2
2
H O, ethanol; (c) a Molecular Sieve 5A column, Ar carrier, for the separation of H , O , N , CH ,
2
2
2
2
4
2 2
and CO. A Plot U backflash column was installed in order to avoid CO and H O poisoning in the
third column. Yields of each product were calculated by the same method as described in detail in
3
9
Ref. .
Results and discussion
Bulk samples characterization
o
XRD patterns of the calcined materials at 450 C (for 8 h) are shown in Figure 1. All M-modified
non-stoichiometric ferrospinels display a spinel phase of magnetite (Fe ), as reported in our
3
O
4

Green Chemistry
Page 6 of 33
View Article Online
DOI: 10.1039/C5GC02139A
37,36
previous papers,
amorphous form. Table 1 reports the surface area measurements. The highest surface area is shown
, and this value declines with an increase of Co-content, which may be explained by
even though Mn-based samples exhibit a lower crystallinity or are partially in
by Mn0.6Fe2.4O
y
a decrease of the particle size in Mn-modified samples. Moreover, the particle size, morphology and
structure of the samples were investigated by means of TEM. The results indicate that the samples
prepared by co-precipitation method are almost uniform in both morphology and particle size
distribution (~2-6 nm) (Figure 2-A).
Table 1. Specific surface area for calcined samples
Sample
Surface area,
2
m /g
Co0.6Fe2.4
O
y
74±2
Co0.54Mn0.06Fe O 88±2
2.4
y
Co0.3Mn0.3Fe2.4
Mn0.6Fe2.4
O
y
138±3
157±3
O
y
Several MFe
sites. Normally, Fe
ferrospinels, MnFe O
4
2
O
4
ferrospinels exhibit inversion of the cations between the tetrahedral and octahedral
and CoFe are completely inverse spinels; however, unlike the other
is a normal spinel. In spite of this, nanostructured powders of various
3
O
4
2 4
O
43
2
synthesis and treatment methods can exhibit different degrees of inversion. In order to disclose the
information on the internal structure, i.e. bonding and distribution of substituted ions among the
octahedral (Oh) and tetrahedral (Td) sites of the materials, Raman Spectroscopy measurements
were performed. Raman features of ferrospinels are highly specific, i.e. they are depending on
44
stoichiometry, cation distribution, defects, and also on the experimental conditions. Spectra of the
inverse spinels are more complicated and show additional Raman-active modes over the five
4
4
g
Raman modes predicted by group theory: F2g(1), E , F2g(2), F2g(3) and A1g. Additional modes may
appear in vibrational spectra due to local distortions of the crystal lattice. These defects do not
affect the long-range ordering of the system, and often cannot be detected by XRD.
Figure 3 shows the Raman spectra of the calcined materials. The spectrum A - Co0.6Fe2.4
y
O reveals
6
vibration modes: 5 predicted by the literature and an additional one caused by the local distortions
-1
in the lattice. The five first order Raman modes are at about 182, 300, 471, 554 and 688 cm . The
-1
4
most intense vibration mode at 688 cm is the A1g, which can be assigned to the AO stretching,
3+
and for the inverse ferrospinels it should be the same for all the samples since the Fe cations are
occupying these Td sites. However, with an increase of the Mn-content the shape and the position

Page 7 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
-1
of the A1g at 688 cm , which corresponds to the Fe-O symmetric stretching in Td sites, changes.
-1
The band shifts to a lower frequency at about 657 cm and becomes broader (Figure 3-B,C). The
other four modes are decreasing in intensity and for Mn0.6Fe2.4 sample, they vanish completely.
O
y
2
+
3+
Mn ions are proposed to partially replace Fe ions in the Td sites (partly due to the non-
stoichiometric composition of the ferrospinels), which may cause the change of the inversion degree
43
towards the formation of the normal spinel. The resulting spectrum of Mn Fe O is in
0
.6 2.4
y
45
agreement with previous reports. The low intensity of the peaks and partial energy shifts can be
4
6
related also to a high defectivity and a low crystallinity as confirmed by XRD.
TPR measurements were performed in order to investigate the role of different metals on
reducibility of samples. The position of the temperature maxima may vary from sample to sample
depending on the particle size and other parameters such as temperature ramp rate. The substitution
2
+
of ferrous iron with another M ion affects the oxygen mobility in the solid, thus modifies the
4
7
reducibility of the material. In Figure 4, the deconvolution of reduction profiles was based on
40,47–49
reduction steps and corresponding temperature maxima, as inferred from the literature.
All M-modified non-stoichiometric ferrospinels show mainly two steps of the reduction processes:
the reduction of iron oxide to metallic iron and the reduction of the other metal oxide to metal. The
reduction of iron oxide:
Fe
2 3 3 4
O (hematite) → Fe O (magnetite) → FeO(wustite) → Fe
47,48
strongly depends on the presence of another metal oxide in M-modified ferrospinels.
In the TPR
o
profile of Co Fe O (Figure4-A), the reduction of hematite to magnetite is observed at ∼400 C,
0
.6 2.4
y
o
o
followed by the reduction of Co
3
O
4
to CoO at ∼502 C and of CoO to metallic Co at ∼576 C , and
o
finally magnetite to wustite reductive transition was observed at ∼690 C. Further reduction of
0
o
0
wustite to metallic Fe appears around 780 C, but the complete reduction of the FeO to Fe could
not be observed.
o
Following the TPR profile of Mn0.6Fe2.4
O
y
(Figure 4-D), the first peak at 200 C corresponds to the
4+
3+
3+
2+
o
reduction of Mn to Mn , while the reduction of Mn to Mn occurs at 525 C. It turned out that
0
it is difficult for MnO to be further reduced to Mn , due to its highly negative reduction potential (-
1
.18 eV), thus there was no evidence of the MnO reduction process under the present experimental
o
conditions. The 616 C peak corresponds to the reduction of magnetite to wustite, beyond which
total reduction to metallic iron is expected to occur. These results are in a good agreement with the
48
3+
2 3 3 4
report of Khan et al. The presence of Mn facilitates the reduction of Fe species [Fe O to Fe O ].
0
o
However, the reduction of FeO→Fe appears at higher temperature, ∼850 C, than for the other M-
modified non-stoichiometric ferrospinels. This effect could be primarily attributed to the formation
δ δ (2−δ)] 4
of either normal or mixed spinels of composition A B(1−δ)[A B O ].

Green Chemistry
Page 8 of 33
View Article Online
DOI: 10.1039/C5GC02139A
In the case of Co/Mn-modified ferrospinel systems (Figure 4-B,C) Fe
2
3
O →Fe
3
O
4
reduction appears
o
o
0
o
at ∼400 C, Fe
3 4
O
→FeO at ∼700 C and FeO→Fe at ∼800 C. For the reduction of other metals we
0
o
o
found: Co
x
O
y
to Co (Co
3
O
4
x y
→CoO at ∼500 C, CoO→Co at ∼600 C) and Mn O to MnO
4
+
3+
o
3+
2+
o
(
Mn →Mn at ∼200 C, Mn →Mn at ∼560 C). In particular, Co/Mn systems present a
cooperative influence of both Mn and Co on each other’s reduction profiles (Table S1).
Therefore, from TPR measurements it can be concluded that incorporation of different metal cations
into magnetite structure alters its reducibility, which in turn is affected by the nature of the
incorporated metal cation. However, it is also important to evaluate the reducibility of the materials
using ethanol as reducing agent. The information about the ethanol adsorption and its further
transformation into different intermediate species was obtained using in-situ DRIFTS study.
Surface characterization: DRIFT spectroscopy
The adsorption of alcohols over metal oxides can lead to different kind of adsorbed species
depending on the surface properties of the material under study. According to the literature, ethanol
5
0–54
adsorbs on the surface mainly as an ethoxide, formed from the scission of the O-H bond.
However, the species formed with increasing temperatures are more related to the surface chemistry
3
8
of each compound. Figure 5 shows spectra of adsorbed ethanol at RT over the four samples.
-1
Ethoxy bands (dissociative adsorption) can be observed at 1057 and 1103 cm for the C-O/C-C
-1
-1
stretching modes together with the bands of the undissociated ethanol (at 1383 cm and 1274 cm
3
8,55
-1
for CH
3
δ and OHδ, respectively).
The complex absorption in the 2800-2400 cm region is
probably due to a Fermi resonance of the overtone of the OHδ vibration mode with the CH-
-1
stretching modes at 2971, 2929, and 2863 cm (CH
3
(as), CH
2
3
(as) and CH (s) modes of the ethoxide
5
2,38
-1
species).
Figures 5(A-C) show a strong band at 3252 cm that may be assigned to the OH
stretching of ethanol H-bonded to the surface, which also was observed over the stoichiometric
38
-1
2 4
CoFe O ferrospinel in our previous paper. The sharp band at 3681 cm together with a shoulder
-1
at 3635 cm is assigned to the –OH stretching vibration of chemisorbed undissociated ethanol over
n+ 56–58
surface Lewis acid sites (M ).
The shoulder at lower frequency may indicate the presence of
different active sites on the materials. However, with increasing the Mn-contain the shoulder
decreases in intensity and only for Mn0.6Fe2.4 it completely disappears. This effect was already
seen by Raman spectroscopic, described above.
Some remarkable differences between the DRIFTS spectra of samples are observed when the
O
y
−1
temperature is increased (Figure 6); acetate bands appear (at 1334, 1439, 1590 cm for the CH3 δ,
(OCOν(s), and OCOν(as) vibrations, respectively). The increasing in intensity and broadening of
-1
o
59
the band at 1590 cm (at 250 C) also indicates the formation of carbonate species. These species

Page 9 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
-1
could further produce CO
corresponds to adsorbed and/or gaseous CO
(Fig 6-D), also exhibit the bands for acetate species, formed by the
2
, which is detected as the doublet at the 2358-2338 cm region and
6
0
2
rotation vibration P and R bands. The DRIFT spectra
obtained on Mn0.6Fe2.4O
4
-1 53,61
reaction of ethoxy species with surface oxygen (1598, 1555, 1433 and 1334 cm ).
When the
o
temperature is increased (up to 400 C) the intensity of the bands corresponding to ethoxy species is
reduced, whereas the bands associated with acetate species (strongly bounded to the surface) remain
4
almost unchanged. However, the main difference between the spectra of Mn0.6Fe2.4O and of other
Co- and Co-Mn ferrospinels spectra lies in the absence of the bands associated with carbonate
species; the band of the adsorbed CO was not observed as well. Thus, the results may indicate that
2
Co and Co-Mn-modified systems were more reactive than the Mn ferrospinel. This is also followed
-1
by the presence of a broad absorbed feature in the 3000-1800 cm region, which is attributed to the
38,62
electronic absorption of the partially reduced sample.
In summary, the redox properties of M-modified ferrospinels and hence their catalytic activity in
the chemical-loop reforming of ethanol (see below) are strongly influenced by the nature of M and
n+
by the M distribution among the Oh and Td sites in the crystal lattice. The main differences
between the four non-stoichiometric spinels studied are the following: the incorporation of only Co
leads to the formation mainly of an inverse spinel, however the addition of Mn resulted in the
3+
2+
substitution of Td coordinated Fe ions with Mn and as a consequence the formation of new type
of active sites. The reduction of M-modified ferrospinels appears to be a complex process due to the
presence of various reduced forms of Fe oxides together with oxides of substituted metal ions
x y x y
(Co O and Mn O ). Besides, it was observed that incorporation of Co-Mn metal cations into
magnetite structure affects its reducibility, which directly depends on the nature of the incorporated
metal cation. In particular, the Co-Mn system presents a cooperative influence on each other’s
reduction profiles. Moreover, during the ethanol anaerobic decomposition/oxidation step over Co
and mixed Co-Mn ferrospinels, acetates and carbonates appeared to be important surface
intermediates and in the case of the only Mn containing catalyst, there was no evidence for the
carbonates formation.
Surface characterization: photoemission spectroscopy
In order to study the kinetics of the reduction and re-oxidation processes that take place on the
catalytic surface, in-situ XPS measurements were performed. In particular, by monitoring the
-6
changes occurred in the photoemission spectra of Fe2p, Co2p and Mn2p, during ethanol (5×10
-4
mbar at 450°C) and water (2×10 mbar at 450°C) exposure, it was possible to determine the role
played by each component of the Co/Mn-modified ferrospinels on the surface chemistry.

Green Chemistry
Page 10 of 33
View Article Online
DOI: 10.1039/C5GC02139A
o
Let us first concentrate on the changes in the XPS data after the sample annealing at 450 C in
y
UHV. As an example, in Figure S1 we compare the spectra of Co0.3Mn0.3Fe2.4O as received and
after annealing: a significant decrease of the FWHM in the O1s spectrum and a small shift of Fe2p
and Mn2p peaks lower binding energy (BE) values are clearly observed, which indicate a surface
modification occurred after elimination of the adsorbed hydroxyl (OH-) groups or/and water
molecules, and the reduction of highly oxidized metal species (e.g. Fe O to Fe O formation). In
2
3
3
4
any case, the spectra obtained after this UHV annealing are taken as a reference for the subsequent
chemical changes induced by the chemical treatments.
The chemical composition determined by photoemission and the spectral fingerprints observed
confirm the presence of M-modified ferrospinels whose stoichiometric ratio is in agreement with
the nominal ratio of the synthesis protocol.
The samples were exposed to an increasing amount of ethanol and the evolution of the main core
level spectra (see Figures S2, S3, S4, S5), was measured observing clear differences in the surface
composition as a function of the presence of different M.
3
+
2+
Starting from the very beginning of ethanol feeding, the reduction of iron from Fe to Fe was
observed for all tested samples, which is evidenced by the decrease of the intensity of the Fe2p peak
3
+
in conjunction with the energy shift of Fe2p3/2 satellite from 8 eV(typical for Fe ) to 5 eV , which
63,64
can be ascribed to oxides with a lower valence.
In the case of prolonged ethanol exposure, the appearance of a further component at 707.0 eV
indicates the formation of a metal component. Such a phenomenon is strongly dependent on the
actual composition of the corresponding sample, and especially on the Co:Mn ratio. As shown in
Figure 7, it was found that Co(II)-incorporation effectively improves the reducibility of the catalyst,
whereas addition of Mn has an opposite effect. This effect is once again confirmed by the
previously described TPR and DRIFTS results. Moreover, Co0.54Mn0.06Fe2.4
reduction in comparison to the sample without Mn (Co0.6Fe2.4 - Figure 7). This behaviour can be
explained by the size effect induced by the presence of Mn (with increasing of Mn-content the
specific surface area also increased). In the Co0.3Mn0.3Fe2.4 ferrospinel within 80 min of reduction
y
O showed a stronger
O
y
O
y
by ethanol, a peak corresponding to the formation of an iron metal phase was not detected and it
was seen only after 120 min of ethanol feeding (see Figure S3). This resilience towards the
reduction was observed also for Mn0.6Fe2.4
S2).
y
O , whose reduction behaviour is very similar (see Figure
Moreover, we could observe changes in the line shape of Mn2p only after the annealing in UHV at
50°C (due to the desorption of surface OH- groups and/or water molecules), whereas afterwards
spectra remained unchanged, despite the iron and cobalt underwent substantial reduction.
4

Page 11 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
It has been already mentioned that Co boosts the reduction of ferrospinels; however it is interesting
to compare the tendency toward the reduction of Co vs Fe. The photoemission peak of Co2p
overlaps with Fe Auger peak, which complicates its analysis, nevertheless some qualitative
2
+
information can be obtained. Cobalt is present as Co and therefore no more reduced species can
be formed, but the metal. As a matter of fact in the Co2p photoemission line it can be observed a
2
+
progressive disappearance of the Co 2p1/2 peaks (795.5 eV) and its satellite in favour of the
formation of a metal component (793.3 eV). Metal components of Fe and Co were observed almost
at the same time, however it can be noted that cobalt was fully reduced to metal in the case
Co Fe O and Co0.54Mn0.06Fe O whereas the reduction of iron species to metallic iron was only
0
.6 2.4
y
2.4
y
partial (Figure 7).
In the case of relatively large Mn-content (Co0.3Mn0.3Fe2.4
y
O ), Co cations are not reduced
immediately, but a relevant oxidized component can be observed in standard conditions (80 min of
ethanol feeding), whereas only after 120 min is fully reduced. It is also interesting to note that such
reaction seems to be autocatalytic; actually in the stage from 80 to 120 min there is a fast increase in
the reduced components with the formation of metallic phases. These experiments suggest that
when Mn is present in large amount there is a substantial depression of reduction. This indicates
that the synthesized ferrospinel are homogeneous and not a physical mixture of Co0.6Fe2.4
. Moreover, the chemical activity of the M-modified ferrospinels is ruled in a complex
way by the sum of the effects due to the two incorporated metals.
y
O and
Mn0.6Fe2.4O
y
To summarize the picture emerging from the XPS investigation we can propose that the reduction
process involves the fast formation of a reduced oxide with a satellite at 5 eV and afterwards the
nucleation of iron metal. This shift of the satellite can be qualitatively associated with the reduction
3+
2+
of Fe to Fe which from the structural point of view corresponds to the reduction to wustite, and
only afterwards this latter is progressively reduced to metal iron.
The formation of the metal phase seems to be a rather slow process and requires a significant
3
+
2+
induction period whereas the first reduction from Fe to Fe is quite fast in the used conditions.
Globally from photoemission we can outline a scenario where the reactivity of the ferrospinels is
strongly dependent on the type of M introduced in the lattice, which however determines a quite
complex reactivity scheme based on a synergetic effect. Manganese in large amount has the effect
of slowing down the reduction reaction and in any case it can never be reduced. The high
2
+
thermodynamic stability of Mn is the probable origin of its scarce attitude toward reduction.
Cobalt on the contrary can be reduced to the metallic state very quickly, even faster and in a larger
extent than iron.

Green Chemistry
Page 12 of 33
View Article Online
DOI: 10.1039/C5GC02139A
In order to get information on the second step of the chemical loop we have investigated the
exposure of ethanol reduced Co0.6Fe2.4 to water at 450°C (see Figures S3
y y
O and Co0.3Mn0.3Fe2.4O
and S5). In both cases Fe2p spectra indicate a fast oxidation process that leads to an almost
complete suppression of the metal component. However it must be noted that the final Fe2p
spectrum is rather broad with a smeared fine structure, which can be due to the co-presence of many
different phases: not only ferrospinels but also iron suboxides and hydroxides. The presence of
these latter is supported by the larger FWHM and increased intensity on the high BE side of the O
1
speak after water dosing. Similarly in the Co2p spectra, the quick formation of oxidized species
2
+
i.e. Co can be observed immediately after water exposure. However, whereas in the pure cobalt
ferrospinel the oxidation of cobalt is almost quantitative, in the case of the Co/Mn sample such
process is partial and a large metal cobalt component is visible even after extended water dosing (60
min). Such difference can be probably traced back to the presence of manganese, which somehow
hinders the reaction with water.
Reactivity tests
Figure 8 shows conversion and yield for the main products (yield >5%), obtained during the first
o
step of the chemical loop, where the material is being reduced by ethanol at 450 C for 20 min. The
cycling time was chosen according to the initial results achieved over pristine Fe
3
O
4
and
37,39
2 4 2 4
stoichiometric ferrospinels - CoFe O , NiFe O
. We reported that in approximately 20 min of
ethanol feeding, the material was constantly reduced until it reached its `steady state` with the stable
composition under these reaction conditions. Therefore, we were able to avoid a further coke
deposition that occurred over the strongly reduced spinel.
y
The conversion of ethanol over Co0.6Fe2.4O (Figure 8-A) reached 100% starting from the first
minutes of the reaction, and it was stable during 20 min of ethanol feeding. One of the main
products of ethanol conversion was hydrogen, with an initial yield of about 50%, which increased
up to 82% while the extent of spinel reduction increased. Also, CO and CO showed similar
2
increasing trends as for H
2
, thus suggesting that the main reactions during spinel reduction involved
the partial oxidation of ethanol:
▪
▪
C H OH + MeO → 2 CO +3 H + MeO
(1)
2
5
x
2
(x-1)
C
2
H
5
OH + 3 MeO
x
→ 2 CO
2
2
+ 3 H + 3 MeO(x-1) (2)
The formation of acetone, which declined with time, can be a result of an aldol condensation of the
intermediately formed acetaldehyde, with further isomerization of aldol into hydroxyketone
followed by a reverse reaction to form acetone and formaldehyde (converted into CO and H
formation of water (with yield ~11%) involves several possible reactions:
2
). The

Page 13 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
▪
▪
▪
C
2
H
5
OH + 6 MeO→ 2 CO
+ MeO → H
O + Me
OH → 2 C + 2 H
+ H O
2
2
+ 3 H
2
O + 6 Me
(3)
(4)
(5)
H
2
2
C
2
H
5
2
Coke and heavy compounds (mainly aromatics) also formed starting from the very beginning of the
reaction time (~20%), with a progressive increase in yield. Under these reaction conditions it was
impossible to avoid the formation of carbonaceous species. The generation of coke was likely due
to dehydrogenation and total deoxygenation of ethanol:
C
2
H
5
OH → C + CO + 3 H
C H OH → 2 C + H O + 2 H
2
2
(6)
(7)
2
5
2
Moreover, disproportionation of CO (Boudouard reaction) can result in coke deposition:
CO→CO
+ C (8)
2
2
Results obtained with the other samples (Figures 8–B,C,D) demonstrate similar trends in the
products distribution. However, some significant differences can be seen: with increasing Mn-
y
content the conversion of ethanol was slightly decreasing. The Mn0.6Fe2.4O sample (Figure 8-D) at
first showed 100% of conversion followed by its decrease down to 80%. The main reason for such
behavior can be an accumulation of MnO phase in the catalyst, which has a high thermodynamic
stability and as a consequence its scarce attitude towards the reduction to metallic component. This
result also has been confirmed by TPR, DRIFTS and in-situ XPS measurements.
As discussed above, we could not completely avoid coke formation, but despite this, we were able
to significantly reduce its amount by changing the composition of the catalyst, as shown in Figure
9
, plotting the overall yield to C residues integrated over the 20 min reduction time. Indeed, the
presence of Mn had the beneficial effect of decreasing the coke deposition, which is an important
issue in order to produce C-free hydrogen. The integrated yields for other products formed during
2
0 min of reduction with ethanol are plotted in Figure S6. In particular, Co-containing ferrospinels
) have shown high yields to CO , CO and H O and enhanced
reactivity in the anaerobic decomposition/oxidation of ethanol. On the other hand, Mn Fe O
y
(Co0.6Fe2.4
y
O and Co0.54Mn0.06Fe2.4
O
y
2
2
0
.6 2.4
appeared to be more active in the formation of acetone, which is the result of the reverse aldol
condensation reaction.
Figure S7 shows the degree of reduction for the four catalysts with different ratio Co/Mn as a
39
function of reaction time (20 min), as calculated from the oxygen balance . Even though the
highest surface area was shown by Mn0.6Fe2.4 catalyst, and it declined with an increase of Co-
content, the samples with high content of Co (Co Fe O and Co0.54Mn0.06Fe O ) were strongly
O
4
0
.6
2.4
y
2.4
y
reduced (almost 100 %) in comparison to only Mn-modified ferrospinel with just ca 20% of
reduction extent. This suggests that the difference in the reducibility of the four M-modified

Green Chemistry
Page 14 of 33
View Article Online
DOI: 10.1039/C5GC02139A
ferrospinels is related to the nature of the incorporated M cation and its intrinsic reactivity, which is
also confirmed by in-situ XPS results.
o
After the reduction with ethanol, the materials were re-oxidized with steam at 450 C, which
2 2
corresponds to the second step of the chemical loop. Figure S8(A-D) reports the yield of H , CO
and CO versus time (yields have been calculated based on the H or O content in products, with
respect to the water fed), whereas Figure 10 reports the overall yields to H , CO and CO , integrated
2
2
over the 20 min oxidation time. High hydrogen yields were obtained over Co0.6Fe2.4
y
O ,
Co0.54Mn0.06Fe2.4
y y
O and Co0.3Mn0.3Fe2.4O . The common problem of all samples was the formation
of CO and CO during the re-oxidation step (Figs.10 and S8), because of the gasification of
2
carbonaceous residues, previously formed during the reduction step. The main difference was in the
amount of produced CO
selective to hydrogen formation, as shown in Table 2, reporting the ratio between the overall yield
to H and the overall yield to CO+CO
In order to quantify the real amount of coke formed on the tested materials an elemental analysis
x
. According to the obtained results, Mn-containing catalysts were more
2
2
.
(
CHNS) was carried out. Table 3 shows the obtained results: Mn-containing materials (Mn Fe O
0.6 2.4 y
y
and Co0.3Mn0.3Fe2.4O ) showed the lowest amount of C residues accumulated during the first
reduction step. However, with all materials a part of the coke still remained on the surface after one
o
complete cycle, being not completely re-oxidized with water at that temperature (450 C). As a
consequence, these materials underwent some morphological changes significantly affected by
accumulation of coke, sintering (particle size increased to ~90-100 nm) and segregation (see
Figures 2-B,C; these images are of samples which underwent several redox cycles).
However, the behavior shown by Co0.3Mn0.3Fe2.4
O
y
can be considered the most promising one,
st
nd
since it forms less coke during the 1 step of ethanol reduction and less CO
x
during the 2 step of
re-oxidation with steam, nevertheless it can provide high hydrogen yield. On the other hand,
Mn0.6Fe2.4
O
y
and Mn0.3Co0.3Fe2.4
O
y
were less efficient in the removal of C residues during the re-
nd
oxidation with steam; this means that the low amount of CO and CO formed during the 2 step are
2
also due to the fact that coke is only in minimal part re-oxidised by steam. This may be due to the
st
lower degree of reduction achieved during the 1 step, since the reduced metal may catalyze the
gasification of coke by reaction with steam. As already pointed out, a third step of re-oxidation with
air is necessary in order to completely remove the C residues formed during the reduction with
22,65
ethanol.

Page 15 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
Table 2. Ratio H
2
/(CO+CO
2
) obtained during the re-oxidation step with water
2 2
H /CO+CO
Mn Fe O
y
28.0
12.2
5.2
0
.6
2.4
Mn0.3Co0.3Fe2.4
Mn0.06Co0.54Fe2.4
Co0.6Fe2.4
O
y
O
y
O
y
5.0
Table 3. Amount of coke (C) in weight % obtained over the four samples investigated
C content(%) after 20 min C content(%) after 20 min
of reduction (ethanol)
reduction (ethanol) + 20 min re-
2
oxidation (H O)
Mn0.6Fe2.4
O
y
2.2
7.9
11.0
9.7
2.1
6.5
6.1
6.2
Co Mn Fe O
0
.3
0.3
2.4
y
Co0.54Mn0.06Fe2.4
Co0.6Fe2.4
O
y
O
y
Conclusions
The chemical-loop reforming of ethanol for H
2
production was studied by using M-modified non-
stoichiometric ferrospinels as ionic oxygen carrier materials. The redox properties of ferrospinels
were found to depend strongly on the nature of incorporated cation in addition to its distribution
inside the crystal lattice. The four catalysts with different ratio between two supplemented metals
2+
2+
(
Co/Mn) were investigated. Substitution of Fe with Co ions led to the formation mainly of an
3+
inverse spinel, however the addition of Mn resulted in the substitution of Td coordinated Fe ions
2
+
with Mn and as a consequence the formation of new type of active sites. The reduction of M-
modified ferrospinels appeared to be a complex process due to the presence of various reduced
forms of Fe oxides together with Co and Mn oxides. Besides, it was observed that incorporation of
Mn-Co metal cations into magnetite (Fe O ) crystal structure affects its reducibility, depending on
3
4
the nature of the incorporated metal cation. In particular, Co/Mn systems showed a cooperative
influence on each other’s reduction profiles. Moreover, in-situ DRIFTS study was performed in
order to follow the stepwise reduction of the catalyst surface, which allowed us to detect different
adsorbed species and their transformation into the products. Acetates and carbonates appeared to be
important surface intermediates over Co- and mixed Co-Mn ferrospinels and in the case of the only
Mn-containing catalyst, there was no evidence of the carbonates formation. Furthermore, according

Green Chemistry
Page 16 of 33
View Article Online
DOI: 10.1039/C5GC02139A
to the results obtained by means of catalytic tests and CHNS elemental analysis, the addition of Mn
hindered carbon deposition, which also caused a decrease of the amount of CO generated by
oxidation with water steam, which is an important issue to produce clean H . However, the Mn
x
2
containing ferrospinels showed a lower efficiency in the removal of the C residues accumulated
during the first step. This may lead to the need of a third step of coke removal by air oxidation, if a
clean material is needed to perform a following cycle. In an alternative approach, as already pointed
out in the literature, the unreactive coke may be maintained as an inert material, provided (that)
2
2,65
during successive cycles the further accumulation of coke occurs at a limited extentonly
Co-Mn mixed ferrospinels, with high levels of Co, enhanced reactivity in the anaerobic
decomposition/oxidation of ethanol. Thus, the behavior shown by Co0.3Mn0.3Fe2.4 can be
.
O
y
considered to be the most promising one, as this type of M-modified ferrospinel formed less coke
during the reduction step (as a result less CO generated during the re-oxidation step with water),
x
and still high hydrogen yield could be obtained.
Therefore, one of the key points to improve the proposed process and make it industrially attractive
is reducing the amount of carbonaceous species which in principle can be done by changing the
catalyst composition and thus tuning its redox properties in order to optimize the ratio between the
degree of reduction and the amount of coke deposition.
Hence, the use of bulk M-modified spinel oxides as oxygen carrier materials for the chemical-loop
reforming of ethanol shows a positive aspect which allows to achieve a good recyclability of the
starting material due to the formation of thermodynamically stable spinel oxide phase during
cycling, using the same mild temperatures for both reduction and re-oxidation steps. This may an
advantage compared to other approaches reported in literature, such as the use of high
66,67
temperatures
to limit coke formation and structured materials with morphologies aimed to
6
8–70
confine the negative effects of sintering
.
Acknowledgements
The EU is acknowledged for the grant to OV (SINCHEM, The European Doctoral Program on
Sustainable Industrial Chemistry). We thank Mr. L. Calore (University of Padova) for the CHNS
elemental analyses. This work has been also funded by the Italian Ministry of Instruction,
University and Research (MIUR) through the FIRB Project RBAP115AYN Oxides at the
nanoscale: multifunctionality and applications.

Page 17 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
Figures captions
Scheme 1. The chemical-loop reforming of ethanol over modified ferrospinels
o
Figure 1. XRD patterns of ferrites calcined at 450 C: (A) Co0.6Fe2.4
O
y y
; (B) Co0.54Mn0.06Fe2.4O ; (C)
Co0.3Mn0.3Fe2.4
O
y
; (D) Mn0.6Fe2.4
O
y
o
Figure 2. TEM images of Co0.3Mn0.3Fe2.4
O
y
: (A) fresh material calcined at 450 C; (B) after 4
cycles•20 min+ 1red •20 min; (C) after 20 min•5 cycles
o
Figure 3. Raman spectra of the ferrites calcined at 450 C: (A) Co0.6Fe2.4
O
y
; (B) Co0.54Mn0.06Fe2.4
O
y
;
4 y
(C) Co0.3Mn0.3Fe2.4Oy ; (D) Mn0.6Fe2.4O
Figure 4. TPR profile Co/Mn-modified non-stoichiometric ferrospinels. (A) Co Fe O ; (B)
0
.6 2.4
y
Co0.54Mn0.06Fe2.4
y 4 y
O ; (C) Co0.3Mn0.3Fe2.4Oy ; (D) Mn0.6Fe2.4O
Figure 5. DRIFTS spectra after adsorption of ethanol at 30°C: (A) Co0.6Fe2.4
Co0.54Mn0.06Fe O , (C) Co Mn Fe O , (D) Mn Fe O
y
O , (B)
2.4
y
0.3
0.3
2.4
y
0.6
2.4
y
y
Figure 6. DRIFTS spectra recorded at increasing temperatures: (A) Co0.6Fe2.4O , (B)
Co0.54Mn0.06Fe2.4
y y y.
O , (C) Co0.3Mn0.3Fe2.4O , (D) Mn0.6Fe2.4O
Figure 7: (top)Fe 2p and (bottom) Co 2p photoemission spectra after the exposure to ethanol at
50°C for 80 minutes
4
Figure 8.Conversion and yields for the main products during 20 min reduction with ethanol at
o
4
50 C over: (A) Co0.6Fe2.4
O
y
; (B) Co0.54Mn0.06Fe2.4
O
y
; (C) Co0.3Mn0.3Fe2.4
O
y
; (D) Mn0.6Fe2.4
O
y
.
Figure 9. Integrated Yields for the coke and heavy compounds during 20 min reduction with ethanol
o
at 450 C.
2 2 2
Figure 10 .Integrated values of yields to H ,CO and CO during the re-oxidation with H O.

Green Chemistry
Page 18 of 33
View Article Online
DOI: 10.1039/C5GC02139A
References
1
R. M. Navarro, M. C. Sánchez-Sánchez, M. C. Alvarez-Galvan, F. Del Valle and J. L. G.
Fierro, Energy Environ. Sci., 2009, 2, 35–54.
2
3
4
G. Tsekouras, D. Neagu and J. T. S. Irvine, Energy Environ. Sci., 2013, 6, 256–266.
B. K. Boggs, R. L. King and G. G. Botte, Chem. Commun. (Camb)., 2009, 4859–4861.
T. Lazarides, M. Delor, I. V Sazanovich, T. M. McCormick, I. Georgakaki, G.
Charalambidis, J. A. Weinstein and A. G. Coutsolelos, Chem. Commun. (Camb)., 2014, 50,
5
21–523.
5
J. W. Jurss, R. S. Khnayzer, J. A. Panetier, K. A. El Roz, E. M. Nichols, M. Head-Gordon, J.
R. Long, F. N. Castellano and C. J. Chang, Chem. Sci., 2015, 6, 4954–4972.
6
7
M. Ni, D. Y. C. Leung and M. K. H. Leung, Int. J. Hydrogen Energy, 2007, 32, 3238–3247.
L. Chen, C. K. S. Choong, Z. Zhong, L. Huang, T. P. Ang, L. Hong and J. Lin, J. Catal.,
2
010, 276, 197–200.
8
F. Fresno, T. Yoshida, N. Gokon, R. Fernández-Saavedra and T. Kodama, Int. J. Hydrogen
Energy, 2010, 35, 8503–8510.
9
1
P. Charvin, S. Abanades, G. Flamant and F. Lemort, Energy, 2007, 32, 1124–1133.
N. Hanada, S. Hino, T. Ichikawa, H. Suzuki, K. Takai and Y. Kojima, Chem. Commun.
0
(Camb)., 2010, 46, 7775–7.
1
1
2
P. Gupta, L. G. Velazquez-Vargas and L. Fan, Energy and Fuels, 2007, 21, 2900–2908.
V. Hacker, R. Fankhauser, G. Faleschini, H. Fuchs, K. Friedrich, M. Muhr and K. Kordesch,
J. Power Sources, 2000, 86, 531–535.
1
1
3
V. Hacker, G. Faleschini, H. Fuchs, R. Fankhauser, G. Simader, M. Ghaemi, B. Spreitz and
K. Friedrich, J. Power Sources, 1998, 71, 226–230.
14
15
16
17
18
19
20
R. Sime, J. Kuehni, L. D. Souza, E. Elizondo and S. Biollaz, 2003, 28, 491–498.
V. Hacker, J. Power Sources, 2003, 118, 311–314.
K. Go, S. Son, S. Kim, K. Kang and C. Park, Int. J. Hydrogen Energy, 2009, 34, 1301–1309.
K. Go, S. Son and S. Kim, Int. J. Hydrogen Energy, 2008, 33, 5986–5995.
K. Otsuka, A. Mito, S. Takenaka and I. Yamanaka, 2001, 26, 191–194.
V. V. Galvita, H. Poelman and G. B. Marin, Top. Catal., 2011, 54, 907–913.
M. Bleeker, S. Gorter, S. Kersten, L. van der Ham, H. van den Berg and H. Veringa, Clean
Technol. Environ. Policy, 2009, 12, 125–135.
2
1
M. F. Bleeker, S. R. a. Kersten and H. J. Veringa, Catal. Today, 2007, 127, 278–290.

Page 19 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
2
2
R. Campo, P. Durán, J. Plou, J. Herguido and J. a. Peña, J. Power Sources, 2013, 242, 520–
26.
5
23
24
25
J. A. Peña, E. Lorente, E. Romero and J. Herguido, Catal. Today, 2006, 116, 439–444.
E. Lorente, J. A. Peña and J. Herguido, J. Power Sources, 2009, 192, 224–229.
K.-S. Cha, H.-S. Kim, B.-K. Yoo, Y.-S. Lee, K.-S. Kang, C.-S. Park and Y.-H. Kim, Int. J.
Hydrogen Energy, 2009, 34, 1801–1808.
2
6
7
K.-S. Kang, C.-H. Kim, W.-C. Cho, K.-K. Bae, S.-W. Woo and C.-S. Park, Int. J. Hydrogen
Energy, 2008, 33, 4560–4568.
2
D. Yamaguchi, L. Tang, L. Wong, N. Burke, D. Trimm, K. Nguyen and K. Chiang, Int. J.
Hydrogen Energy, 2011, 36, 6646–6656.
2
8
9
S. Takenaka, N. Hanaizumi, V. Son and K. Otsuka, J. Catal., 2004, 228, 405–416.
K.-S. Kang, C.-H. Kim, K.-K. Bae, W.-C. Cho, W.-J. Kim, Y.-H. Kim, S.-H. Kim and C.-S.
Park, Int. J. Hydrogen Energy, 2010, 35, 568–576.
2
3
0
1
T. Kodama, T. Shimizu, T. Satoh and M. Nakata, 2003, 73, 363–374.
K. Otsuka, T. Kaburagi, C. Yamada and S. Takenaka, J. Power Sources, 2003, 122, 111–
3
1
21.
3
2
3
E. Lorente, J. a. Peña and J. Herguido, Int. J. Hydrogen Energy, 2011, 36, 7043–7050.
S. Takenaka, T. Kaburagi, C. Yamada, K. Nomura and K. Otsuka, J. Catal., 2004, 228, 66–
3
7
4.
S. Takenaka, K. Nomura, N. Hanaizumi and K. Otsuka, Appl. Catal. A Gen., 2005, 282, 333–
41.
34
35
36
37
38
39
3
V. Crocella, F. Cavani, G. Cerrato, S. Cocchi, M. Comito, G. Magnacca and C. Morterra, J.
Phys. Chem. C, 2012, 116, 14998–15009.
S. Cocchi, M. Mari, F. Cavani and J.-M. M. Millet, Appl. Catal. B Environ., 2014, 152-153,
2
50–261.
C. Trevisanut, F. Bosselet, F. Cavani and J. M. M. Millet, Catal. Sci. Technol., 2015, 5,
280–1289.
J. V. Ochoa, C. Trevisanut, J.-M. M. Millet, G. Busca and F. Cavani, J. Phys. Chem. C,
013, 117, 23908–23918.
C. Trevisanut, M. Mari, J.-M. M. Millet and F. Cavani, Int. J. Hydrogen Energy, 2015, 40,
264–5271.
1
2
5
4
0
1
A. Khan, P. Chen, P. Boolchand and P. Smirniotis, J. Catal., 2008, 253, 91–104.
4
J. C. Ryu, D. H. Lee, K. S. Kang, C. S. Park, J. W. Kim and Y. H. Kim, J. Ind. Eng. Chem.,
2
008, 14, 252–260.

Green Chemistry
Page 20 of 33
View Article Online
DOI: 10.1039/C5GC02139A
42
43
44
K. Otsuka, C. Yamada, T. Kaburagi and S. Takenaka, Int. J. Hydrogen Energy, 2003, 28,
35–342.
3
N. Z. Lazarevic, Z, Jovalekic, C, Milutinovic, A, Romcevic, M.J, Romcevic, Acta Phys. Pol.
A, 2012, 121, 682–686.
C. S. S. R. Kumar, Ed., Raman Spectroscopy for Nanomaterials Characterization, Springer
Berlin Heidelberg, Berlin, Heidelberg, 2012.
4
5
6
J. J. Graves, P.R., Johnston, C., Campaniello, Mater. Res. Bull., 1988, 23, 1651–1660.
P. Chandramohan, M. P. Srinivasan, S. Velmurugan and S. V. Narasimhan, J. Solid State
Chem., 2011, 184, 89–96.
4
4
4
4
5
7
8
9
0
G. K. Reddy, P. Boolchand and P. G. Smirniotis, J. Phys. Chem. C, 2012, 116, 11019–11031.
A. Khan and P. G. Smirniotis, J. Mol. Catal. A Chem., 2008, 280, 43–51.
L. Zhang and Y. Wu, Hindawi Publ. Corp. J. Nanomater., 2013, 6.
C. Resini, T. Montanari, L. Barattini, G. Ramis, G. Busca, S. Presto, P. Riani, R. Marazza,
M. Sisani, F. Marmottini and U. Costantino, Appl. Catal. A Gen., 2009, 355, 83–93.
S. S. Ata-Allah, J. Magn. Magn. Mater., 2004, 284, 227–238.
5
5
5
1
2
3
M. Domok, M. Toth, J. Rasko and a Erdohelyi, Appl. Catal. B Environ., 2007, 69, 262–272.
K. Rintramee, K. Föttinger, G. Rupprechter and J. Wittayakun, Appl. Catal. B Environ.,
2
012, 115-116, 225–235.
5
4
5
G. Busca, U. Costantino, T. Montanari, G. Ramis, C. Resini and M. Sisani, Int. J. Hydrogen
Energy, 2010, 35, 5356–5366.
5
M. Schmal, D. V. Cesar, M. M. V. M. Souza and C. E. Guarido, Can. J. Chem. Eng., 2011,
8
9, 1166–1175.
5
6
7
T. W. Birky, J. T. Kozlowski and R. J. Davis, J. Catal., 2013, 298, 130–137.
T. Mathew, B. B. Tope, N. R. Shiju, S. G. Hegde, B. S. Rao and C. S. Gopinath, Phys. Chem.
Chem. Phys., 2002.
5
58
59
60
G.Busca, Phys. Chem. Chem. Phys., 1999, 1, 723–736.
L. Mattos and F. Noronha, J. Catal., 2005, 233, 453–463.
A. Boudjemaa, C. Daniel, C. Mirodatos, M. Trari, A. Auroux and R. Bouarab, Comptes
Rendus Chim., 2011, 14, 534–538.
61
62
63
L. V. Mattos and F. B. Noronha, J. Power Sources, 2005, 152, 50–59.
G. Busca, Catal. Today, 1996, 27, 457–496.
M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M. Lau, A. R. Gerson and R. S. C.
Smart, Appl. Surf. Sci., 2011, 257, 2717–2730.

Page 21 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A
64
65
66
67
A. P. Grosvenor, B. A. Kobe, M. C. Biesinger and N. S. McIntyre, Surf. Interface Anal.,
004, 36, 1564–1574.
E. Hormilleja, P. Durán, J. Plou, J. Herguido and J. a. Peña, Int. J. Hydrogen Energy, 2014,
9, 5267–5273.
2
3
J. R. Scheffe, M. D. Allendorf, E. N. Coker, B. W. Jacobs, A. H. McDaniel and A. W.
Weimer, Chem. Mater., 2011, 23, 2030–2038.
K. S. Kang, C. H. Kim, K. K. Bae, W. C. Cho, S. H. Kim and C. S. Park, Int. J. Hydrogen
Energy, 2010, 35, 12246–12254.
6
8
9
L. Neal, A. Shafiefarhood and F. Li, Appl. Energy, 2015.
6
B. M. Corbella, L. F. De Diego, F. García-Labiano, J. Adánez and J. M. Palacios, Energy &
Fuels, 2005, 19, 433–441.
7
0
Z. Sarshar, Z. Sun, D. Zhao and S. Kaliaguine, Energy & Fuels, 2012, 26, 3091–3102.

Green Chemistry
Page 22 of 33
View Article Online
DOI: 10.1039/C5GC02139A

Page 23 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A

Green Chemistry
Page 24 of 33
View Article Online
DOI: 10.1039/C5GC02139A

Page 25 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A

Green Chemistry
Page 26 of 33
View Article Online
DOI: 10.1039/C5GC02139A

Page 27 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A

Green Chemistry
Page 28 of 33
View Article Online
DOI: 10.1039/C5GC02139A

Page 29 of 33
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02139A