Hydrogen production from water decomposition by redox of Fe2O3 modified with single- or double-metal additives

Article abstract of DOI:10.1016/j.jssc.2010.03.017

Iron oxide modified with single- or double-metal additives (Cr, Ni, Zr, Ag, Mo, Mo-Cr, Mo-Ni, Mo-Zr and Mo-Ag), which can store and supply pure hydrogen by reduction of iron oxide with hydrogen and subsequent oxidation of reduced iron oxide with steam (Fe₃O₄ (initial Fe₂O₃)+4H₂?3Fe+4H₂O), were prepared by impregnation. Effects of various metal additives in the samples on hydrogen production were investigated by the above-repeated redox. All the samples with Mo additive exhibited a better redox performance than those without Mo, and the Mo-Zr additive in iron oxide was the best effective one enhancing hydrogen production from water decomposition. For Fe₂O₃-Mo-Zr, the average H₂ production temperature could be significantly decreased to 276 °C, the average H₂ formation rate could be increased to 360.9-461.1 μmol min^-1 Fe-g^-1 at operating temperature of 300 °C and the average storage capacity was up to 4.73 wt% in four cycles, an amount close to the IEA target.

Full text of DOI:10.1016/j.jssc.2010.03.017

ARTICLE IN PRESS
Journal of Solid State Chemistry 183 (2010) 1075–1082
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Hydrogen production from water decomposition by redox of Fe2O3 modified
with single- or double-metal additives
Ã
Xiaojie Liu, Hui Wang
Key Laboratory of Synthetic and Natural Functional Molecule Chemistry (Ministry of Education), College of Chemistry and Materials Science, Northwest University,
Xi’an 710069, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Iron oxide modified with single- or double-metal additives (Cr, Ni, Zr, Ag, Mo, Mo–Cr, Mo–Ni, Mo–Zr
and Mo–Ag), which can store and supply pure hydrogen by reduction of iron oxide with hydrogen and
subsequent oxidation of reduced iron oxide with steam (Fe3O4 (initial Fe2O3)+4H223Fe+4H2O), were
prepared by impregnation. Effects of various metal additives in the samples on hydrogen production
were investigated by the above-repeated redox. All the samples with Mo additive exhibited a better
redox performance than those without Mo, and the Mo–Zr additive in iron oxide was the best effective
one enhancing hydrogen production from water decomposition. For Fe2O3–Mo–Zr, the average H2
production temperature could be significantly decreased to 276 1C, the average H2 formation rate could
Received 28 January 2010
Received in revised form
25 February 2010
Accepted 7 March 2010
Available online 15 March 2010
Keywords:
Hydrogen production
Fe2O3
be increased to 360.9–461.1
storage capacity was up to 4.73 wt% in four cycles, an amount close to the IEA target.
& 2010 Elsevier Inc. All rights reserved.
m
mol minꢀ1 Fe-gꢀ1 at operating temperature of 300 1C and the average
Water decomposition
Mo or Mo–Zr additives
1. Introduction
initial Fe2O3 is reduced to active Fe by H2 (hydrogen storage at
reduction step 1), and then the active Fe (reduced iron oxide) is
oxidized to Fe3O4 by H2O vapor (hydrogen production at
oxidation step 2). According to reaction 2, 1 mol of Fe reacts with
H2O can produce 1.33 mol of H2, which corresponds to the
theoretical amount of hydrogen being chemically stored (i.e.,
4.8 wt% of Fe). This value is relatively higher than hydrogen
storage capacities of the above-mentioned materials. Further-
more, the operating pressure of releasing H2 is normal pressure
and hydrogen is the only product. Despite the many advantages of
the solid material, however, the utilization of iron oxide as a
possible substitute for hydrogen storage material requires the
solution of several problems related to hydrogen production, such
as H2 production rate at a low operating temperature, reversible
cycle stability and hydrogen storage capacity.
Hydrogen is an ideal energy carrier because of its high
efficiency, zero emission and production from renewable biomass
[1,2]. Thus, hydrogen has been recognized as a clean energy
substitute for fossil fuels but has not been used yet. One of the
major problems for the large-scale use of hydrogen is the
difficulty of an efficient storage [3,4]. The present hydrogen
storage technologies are mainly physical method such as
pressurized gas and cryogenic liquid, and chemical one such as
storing in various solid and liquid materials; however, the solid
storage has been considered as the safest and most effective way
[5]. Among the various candidates of hydrogen storage solid
materials including metal hydrides, complex hydrides and carbon
materials [6–15], none is capable of meeting the US Department
of Energy (DOE) and International Energy Agency (IEA) cost and
performance targets for use in transportation vehicles that is
hydrogen storage capacity of 6.5 wt% (62 kg H2/m3) or 5 wt%
(50 kg H2/m3) [16].
2. Experimental section
Recently, a new class of iron oxide solid material has attracted
much attention because of their potential applications in hydro-
gen storage and production processes [17–22]. Otsuka et al.
reported a new, simple, safe, inexpensive and environmentally
compatible technology for the storage and supply of hydrogen to
polymer electrolyte fuel cell (PEFC) vehicles by the redox reaction
of Fe3O4 (or initial Fe2O3)+4H223Fe+4H2O [23]. First, Fe3O4 or
2.1. Reagent and instrument
Fe2O3 (99.8%) was purchased from Bodi chemical company in
Tianjin. AgNO3 (99.8%), Cr(NO3)3 ꢁ 9H2O(99.0%) and NiCl2 ꢁ 6H2
O(98.0%) were purchased from Kerme company in Tianjin.
(NH4)6Mo7O24 ꢁ 4H2O(99.0%) and ZrOCl2 ꢁ 8H2O(98.0%) were pur-
chased from the Chemical company in Xi’an. Scanning electron
microscopy (SEM, a Quanta 400 FEG instrument operated at
25 kV, Oxford INCA 350 detector) and X-ray diffraction (XRD,
Ã
D/max-3c, CuK
a, 50 kV, 300 mA, at room temperature in air) were
Corresponding author. Fax: +86 029 88303798.
E-mail address: huiwang@nwu.edu.cn (H. Wang).
used to characterize the morphologies, particle sizes and
0022-4596/$ - see front matter & 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2010.03.017

ARTICLE IN PRESS
1076
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
compositions of the samples. The specific surface areas of the
samples were measured with a surface area analyzer (JW-K) by
the adsorption of N2 at liquid N2 temperature (BET method).
WFSM-3012 evaluating device with a gas chromatogram (GC,
SP-6890) designed by ourselves was used to evaluate the redox
performances of the samples, including H2 production tempera-
ture, H2 production rate, hydrogen storage capacity and cyclic
stability.
theoretical total hydrogen production amount:
molactualpH ðactual hydrogen mole produced at time tÞ
2
DO ¼
moltotalpH ðtotal hydrogen mole produced after complete oxidationÞ
2
2.3. Samples’ evaluation
The redox performances of these samples were evaluated by
the WFSM-3012 evaluating device. The evaluating process
comprises two steps, i.e., hydrogen storage and production, in
which the redox performances such as hydrogen production
temperature, the cyclic stability and hydrogen storage capacity
could be tested. First, the sample bed, into which the sample
(around 0.2 g) was put, was heated to 500 1C at a heating rate of
4 1C minꢀ1, and meanwhile H2 as reduced gas and Ar as carrier
gas were injected into the bed to reduce iron oxide to active Fe.
When the temperature of the bed reached 500 1C, this tempe-
rature was being kept until no consumption of hydrogen was
detected by gas chromatograph (GC). Then, the oxidation of
active Fe could be performed successively to produce H2 by the
1:4 gas mixture of H2O vapor and Ar with the flow rate of
2.2. Sample preparation
The Fe2O3 samples with single- or double-metal additives
were prepared by impregnating Fe2O3 powder with an aqueous
solution containing metal cation additives. The amount of added
single-metal cation (M) was adjusted to be 5 mol% of total metal
cations (M/(Fe+M)¼0.05), while those of double-metal cations
were adjusted to be 5 (Mo) and 2.5 (M) mol% of total metal
cations (Mo/(Fe+Mo+M)¼0.05 and M/(Fe+Mo+M)¼0.025), re-
spectively. The Fe2O3 powder was directly impregnated into
250 ml beaker with 150 ml distilled water containing the
corresponding metal cation additives (single-metal cation (M)
or double-metal cations (Mo+M)) in the above proportion. Then,
the beaker with an aqueous solution was heated on the magnetic
force mixer with thermoelectric couple to control temperature at
95 1C and the aqueous solution containing the corresponding
metal salt were stirred thoroughly until water in the solution
were steamed away. Next, the impregnated sample was dried at
90 1C for 10 h and subsequently calcined in air at 500 1C for 15 h.
Nine kinds of as-prepared Fe2O3 samples with various metal
additives were denoted as Fe2O3–Zr, Fe2O3–Ni, Fe2O3–Mo,
Fe2O3–Ag, Fe2O3–Cr, Fe2O3–Mo–Zr, Fe2O3–Mo–Ni, Fe2O3–Mo–
Ag and Fe2O3–Mo–Cr, respectively. Fe2O3 and Fe3O4 without
additives were denoted as Fe2O3-none and Fe3O4-none,
respectively.
50 ml minꢀ1
, as the temperature of the sample bed was
increased to 600 1C from room temperature at a heating rate
of 4 1C minꢀ1 and the temperature was kept until hydrogen was
no longer analyzed by GC.
3. Results and discussions
3.1. Selection of initial material
Both Fe2O3 and Fe3O4 were able to serve as the starting material
for hydrogen storage, from which the better one should be found out
by the redox reaction. In order to achieve the goal, the redox
performances of Fe2O3 obtained by direct purchase and Fe3O4
obtained by a co-precipitation method were investigated. Fig. 1
shows that the changes of H2 formation rate vs. temperature and of
The degree of oxidation of the reduced samples (DO) was
defined as the ratio of actual hydrogen production amount to the
1000
1
Fe2O3-none
Fe2O3-none
800
0.8
1st cycle
1st cycle
2nd cycle
2nd cycle
3rd cycle
3rd cycle
600
0.6
4th cycle
4th cycle
400
200
0
0.4
0.2
0
0
100 200 300 400 500 600
0
50 100 150 200 250 300
Time /min
Temperature /°C
1
0.8
0.6
0.4
0.2
0
1000
800
600
400
200
0
Fe3O4-none
Fe3O4-none
1st cycle
2nd cycle
3rd cycle
4th cycle
1st cycle
2nd cycle
3rd cycle
4th cycle
0
100 200 300 400 500 600
0
50 100 150 200 250 300
Time /min
Temperature /°C
Fig. 1. Variations of H2 formation rate vs. temperature, and of degree of oxidation vs. time in the re-oxidation of Fe2O3-none and Fe3O4-none in four redox cycles.

ARTICLE IN PRESS
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
1077
Table 1
The redox performances of Fe2O3-none and Fe3O4-none without additives.
Sample
Cycle
Peak
H2 formation
temperature
at the rate of
250 m
mol minꢀ1 Fe-gꢀ1 (1C)
The rate of H formation
The required time at the fixed DO (min)
H2 (wt%)
2
temperature
(1C)
(m )
mol minꢀ1 Fe-gꢀ1
At peak
At 300 (1C)
At 0.5000
At 0.7500
temperature (1C)
Fe2O3-none
Fe3O4-none
1st
528
530
549
528
380
475
491
525
668.3
802.8
478.6
249.3
43.5
31.1
25.9
29.8
128
140
145
158
140
146
160
240
4.88
4.74
4.64
3.94
2nd
3rd
4th
1st
353
352
370
371
288
300
310
321
480.1
368.2
339.3
312.7
323.8
238.3
224.9
157.8
104
108
112
125
139
157
162
168
4.69
4.53
4.48
4.05
2nd
3rd
4th
DO vs. time for the Fe2O3-none and Fe3O4-none samples in four
cycles. The corresponding data including the peak temperature, the
rate of H2 formation at peak temperature and 300 1C, the
4.42, 4.64, 4.38 and 4.7 wt% in four redox cycles, respectively,
indicating that Fe2O3–Mo also had the highest storage capacity, a
near theoretical value of 4.8 wt%. Furthermore, the storage
capacities of Fe2O3–Mo from the first to forth redox cycle (4.74,
4.63, 4.64 and 4.78 wt%) also showed a good cyclic stability for the
sample. Based on the comparisons of all the modified Fe2O3
samples above, it is concluded that Mo additive had the most
remarkable catalytic effect on improving hydrogen production. In
addition, the kinetic curves in Fig. 2a–e show that the other metal
additives also had a certain influence on improving the redox
performances such as the cyclic stability. This may be due to the
fact that the metal additives dispersed in the sample could
effectively prevent the particle sintering of the sample. Therefore,
it is necessary for us to further investigate the cooperative effect
of Mo cation with other transition metal additive on improving
hydrogen production.
temperature of H2 formation at the rate of 250 m ,
mol minꢀ1 Fe-gꢀ1
hydrogen storage capacity and the required time at a fixed DO were
listed in Table 1.
As shown in Fig. 1a and b and listed in Table 1, the rates of H2
formation for Fe2O3-none and Fe3O4-none at 300 1C were in the
range of 25.9–43.5 and 157.8–323.8 m
mol minꢀ1 Fe-gꢀ1 in four
repeated cycles, respectively. It is clear that the H2 production
rate of Fe3O4-none was higher than that of Fe2O3-none at 300 1C.
At the same time, the temperature of H2 formation for Fe3O4-none
at the rate of 250 m
mol minꢀ1 Fe-gꢀ1 was much lower than that
for Fe2O3-none. In addition, the average required time of Fe3O4-
none at a fixed DO in four cycles was shorter than that of Fe2O3-
none (see Fig. 1a0 and b0 and Table 1). Accordingly, it seems that
Fe3O4-none should be preferred as the initial material for
hydrogen storage compared with Fe2O3-none. However, hydrogen
storage capacity, another important factor, for Fe2O3-none
(4.69 wt%) was higher than that for Fe3O4-none (4.44 wt%)
(Table 1), and Fe2O3 has other commercial advantages over
Fe3O4, such as the direct purchase, abundant source and low cost.
Therefore we chose the Fe2O3 powder as the starting material in
the present work.
As described above, Mo metal additive in iron oxide had the
most effective influence on enhancing hydrogen production
compared with other metals. So the cooperative effect of Mo
metal with each one of the other four metals (Zr, Cr, Ag and Ni
metal cations) as double-metal additives (Mo–Zr, Mo–Cr, Mo–Ag
and Mo–Ni) in the sample on hydrogen production was also
investigated. Comparing the kinetic curves and data in Fig. 2a0–e0
and Table 3 with those in Fig. 2a–e and Table 2, it is obvious that
the double-metal additives (Mo–M) in the samples could improve
hydrogen production more remarkably than the corresponding
single-metal additives (M) except Mo additive. E.g., the average
temperatures of H2 formation for the Fe2O3–Mo–Zr, Fe2O3–Mo–
Cr, Fe2O3–Mo–Ag and Fe2O3–Mo–Ni samples were 276, 275, 293
and 271 1C in four cycles, respectively, indicating that the
temperatures were indeed lower than those for the corresponding
samples with single-metal additives; the rates of H2 formation
at 300 1C for Fe2O3–Mo–Zr, Fe2O3–Mo–Cr, Fe2O3–Mo–Ag and
Fe2O3–Mo–Ni were 360.9–461.1, 340.9–366.2, 168.5–343.4 and
3.2. Effect of single- double-metal additives on hydrogen storage
Fig. 2 shows that the changes of H2 formation rate vs.
temperature for the samples with single-metal (Zr, Cr, Ag, Ni
and Mo) and double-metal (Mo–Zr, Mo–Cr, Mo–Ag and Mo–Ni)
additives, and the data related to hydrogen production were listed
in Tables 2 and 3.
As shown in Fig. 2a–e and listed in Table 2, no matter what
type of single metal added into Fe2O3 was, the redox perfor-
mances of the samples were notably elevated. However, the
effect of different metal on hydrogen production was different
(Fig. 2a–e). E.g., the average temperatures of H2 formation were
418, 396, 443, 501 and 276 1C for the Fe2O3–Zr, Fe2O3–Cr, Fe2O3–
Ag, Fe2O3–Ni and Fe2O3–Mo samples in four repeated cycles,
respectively. Obviously, the H2 formation temperature of Fe2O3–
Mo was much lower than that of the other samples. At a relatively
low temperature of 300 1C, the H2 formation rate of Fe2O3–Mo
516.2–369.8 m
mol minꢀ1 Fe-gꢀ1 from the first to forth cycle,
respectively, each of which is much higher than that of the
samples with the corresponding single metal additive. This also
indicates that these double-metal additives in the samples can
enhance hydrogen production significantly. Furthermore, the
average hydrogen storage capacities of Fe2O3–Mo–Zr, Fe2O3–
Mo–Cr, Fe2O3–Mo–Ag and Fe2O3–Mo–Ni were 4.73, 4.61, 4.66
and 4.50 wt% in four cycles, respectively. It is apparent that
the hydrogen storage capacity of Fe2O3–Mo–Zr (4.73 wt%), a close
theoretical amount of 4.8 wt%, was the highest among all the
samples tested in this work. It can be seen from Fig. 2a0–e0 that
the kinetic curves for Fe2O3–Mo–Zr were overlapped more
wonderfully than those for the others, showing that the Mo–Zr-
modified sample had an excellent ability to preserve catalytic
(487.5–453.1
the samples with single metal, increased by about ten times
m
mol minꢀ1 Fe-gꢀ1) was the highest among all of
compared with that of Fe2O3-none (43.5–29.8 m
mol minꢀ1
Fe-gꢀ1). Similarly, the average hydrogen storage capacities of
Fe2O3–Zr, Fe2O3–Cr, Fe2O3–Ag, Fe2O3–Ni and Fe2O3–Mo were 4.54,

ARTICLE IN PRESS
1078
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
1000
800
600
400
200
0
1000
Fe2O3 (Zr)
1st cycle
2nd cycle
3rd cycle
4th cycle
Fe2O3 (Mo+Zr)
800
1st cycle
2nd cycle
3rd cycle
4th cycle
600
400
200
0
Fe2O3 (Cr)
Fe2O3 (Mo+Cr)
800
600
800
600
400
1st cycle
2nd cycle
3rd cycle
4th cycle
1st cycle
2nd cycle
3rd cycle
4th cycle
400
200
0
200
0
Fe2O3 (Mo+Ag)
Fe2O3 (Ag)
800
800
600
400
200
0
1st cycle
2nd cycle
3rd cycle
4th cycle
1st cycle
2nd cycle
3rd cycle
4th cycle
600
400
200
0
Fe2O3 (Ni)
Fe2O3 (Mo+Ni)
800
600
400
200
0
800
600
400
200
0
1st cycle
2nd cycle
3rd cycle
4thcycle
1st cycle
2nd cycle
3rd cycle
4th cycle
Fe2O3 (Mo+Zr)
Fe2O3 (Mo)
800
600
400
200
0
800
600
400
200
0
1st cycle
2nd cycle
3rd cycle
4th cycle
5th cycle
6th cycle
7th cycle
8th cycle
9th cycle
10th cycle
1st cycle
2nd cycle
3rd cycle
4th cycle
0
100
200
300
400
500
600
0
100
200
300
400
500
600
Temperature /°C
Temperature/°C
Fig. 2. Variations of H2 formation rate vs. temperature for Fe2O3–Zr, –Cr, –Ag, –Ni, Mo and Fe2O3–Mo–M (Zr, Cr, Ag, Ni) in four redox cycles, and Fe2O3–Mo–Zr in ten cycles.
activity and cyclic stability. In the present work, the cycle times of
Fe2O3–Mo–Zr were prolonged to ten in order to further observe
the catalytic activity and cyclic stability. Obviously, the kinetic
curves of Fe2O3–Mo–Zr in Fig. 2e0 were superposed perfectly in
ten cycles, manifesting that the sample was not deactivated even
after 10 cycles. According to the comparisons of all the Fe2O3
samples modified with single- and double-metal additives, it is
concluded that all the samples with Mo additive show the good
redox performances, and the Mo–Zr additive in the sample had
the most remarkable role on improving hydrogen production.
3.3. Characterization of the samples before and after redox
To further understand the reason that the single- and double-
metal additives in the modified Fe2O3 samples can enhance
hydrogen production, SEM images, X-ray diffraction patterns and
BET were used to characterize the morphologies, particle sizes
and compositions of the samples before and after the redox. Figs. 3
and 4 depict the SEM images of the Fe2O3 modified with single-
and double-metal additives before and after redox, respectively.
From the SEM images of the samples in Figs. 3 and 4, it can be seen

ARTICLE IN PRESS
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
1079
Table 2
The redox performances of Fe2O3 modified with single metal (Zr, Cr, Ag, Ni and Mo).
Sample
Cycle
Peak temperature (1C)
H2 formation temperature
at 250 m
mol minꢀ1 Fe-gꢀ1
The rate of H formation (
m
mol minꢀ1 Fe-gꢀ1
)
H2 (wt%)
2
At peak temperature (1C)
At 300 (1C)
At 354 (1C)
Fe2O3–Zr
1st
510
508
495
510
441
444
419
370
796.8
932.3
822.9
640.4
78.4
98.7
165.2
218.7
234.9
216.2
4.45
4.43
4.75
4.52
2nd
3rd
4th
130.4
102.1
Fe2O3–Cr
Fe2O3–Ag
Fe2O3–Ni
Fe2O3–Mo
1st
453
464
466
488
394
406
376
410
518.8
543.5
627.1
616.4
159.1
125.0
126.6
137.5
285.2
204.1
242.6
203.1
4.33
4.34
4.45
4.55
2nd
3rd
4th
1st
517
523
544
506
432
449
455
435
963.6
911.5
758.3
659.8
41.6
42.8
61.5
61.9
99.7
121.4
126.1
114.1
4.76
4.44
4.70
4.68
2nd
3rd
4th
1st
336
403
385
497
411
560
522
511
354.1
386.7
312.3
353.4
100.0
60.78
87.5
277.8
170.2
207.8
135.4
4.51
4.70
4.67
3.63
2nd
3rd
4th
77.1
1st
333
333
333
323
270
279
279
277
670.8
704.2
654.9
453.1
487.5
416.6
435.2
453.1
616.6
500.0
385.5
395.8
4.74
4.63
4.64
4.78
2nd
3rd
4th
Table 3
The redox performances of Fe2O3 modified with double metals Mo+M (Zr, Cr, Ag and Ni).
Sample
Cycle
Peak temperature (1C)
H2 formation temperature
at 250 m
mol minꢀ1 Fe-gꢀ1
The rate of H formation (
m
mol minꢀ1 Fe-gꢀ1
)
H2 (wt%)
2
At peak temperature (1C)
At 300 (1C)
At 354 (1C)
Fe2O3–Mo+Cr
1st
338
340
316
312
275
277
278
272
358.7
441.7
458.2
488.1
340.9
331.6
342.7
366.2
356.0
312.5
323.1
334.7
4.66
4.62
4.72
4.44
2nd
3rd
4th
Fe2O3–Mo+Ag
Fe2O3–Mo+Ni
Fe2O3–Mo+Zr
1st
396
332
337
336
305
285
290
293
687.2
756.8
848.9
634.4
168.5
350.3
307.1
343.4
448.0
523.0
490.0
558.0
4.61
4.65
4.71
4.66
2nd
3rd
4th
1st
332
291
344
344
262
274
274
275
936.4
565.2
518.6
577.3
516.2
562.5
408.1
369.8
402.5
427.1
494.7
542.1
4.68
4.53
4.35
4.45
2nd
3rd
4th
1st
335
338
327
313
339
322
326
326
338
338
279
276
273
278
276
278
282
277
282
278
555.4
875.3
868.9
818.5
828.1
700.0
802.1
875.1
733.3
604.1
360.9
396.1
429.1
461.1
433.4
434.6
420.8
404.3
457.3
443.1
518.0
439.0
488.0
435.0
405.0
484.0
423.0
411.0
559.0
490.0
4.67
4.74
4.79
4.70
4.67
4.72
4.68
4.56
4.62
4.64
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
that the particles of all the samples after the redox existed sintering
in a certain degree due to repeated cycle times at relatively high
reaction temperature, but there were no obvious differences in
particle size between the samples with double-metal additives
before and after the redox except the Fe2O3–Mo–Zr sample.
Furthermore, it can be seen from the SEM images in Fig. 3b and
b0 that the particle size of the Fe2O3–Cr sample seems unchanged
before and after the redox, but the sample still didn’t show the
good redox performances because of the low average H2
observed from Fig. 4a and a0 that the particles of the Fe2O3–Mo–
Zr sample were sintered severely after 10 redox cycles, but the
sample displayed the most excellent redox performances due to
the lowest H2 production temperature (2761C). Accordingly, it is
suggested that the change of particle size, i.e., whether the sample
sintering or not, was not the key factor to decide the redox
performances of the sample, but the type of metal additive in the
sample, such as Mo metal, may be more important one to promote
the production of hydrogen effectively.
production rate (137.05
m
mol minꢀ1 Fe-gꢀ1) at 300 1C and the
From the conclusions obtained above, it is certain that Mo
additive in the sample acted an important role on improving
high average H2 production temperature (3971C). It is also

ARTICLE IN PRESS
1080
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
Fe2O3-Mo-Zr before
Fe2O3-Mo-Zr after
Fe2O3-Mo-Cr after
Fe2O3-Mo-Ag after
Fe2O3-Mo-Ni after
Fe2O3-Zr after
Fe2O3-Zr before
Fe2O3-Mo-Cr before
Fe2O3-Cr before
Fe2O3-Cr after
Fe2O3-Mo-Ag before
Fe2O3-Ag before
Fe2O3-Ag after
Fe2O3-Mo after
Fe2O3-Ni after
Fe2O3-Mo before
Fe2O3-Mo-Ni before
2 µm
Fe2O3-Ni before
Fig. 4. SEM images of the Fe2O3 samples with double-metal additives before and
after the redox.
hydrogen production, thus more attention were paid to the
characterization of the samples with Mo additive. The reason for
the Mo additive decreasing the H2 formation temperature in a
large degree might be just as assumed that the Mo cation took
part in the redox reaction and further resulted in accelerating the
reaction between active Fe and H2O vapor, i.e., water decomposi-
tion reaction. This conclusion is the same as our previous
experiment result [19], and will be further confirmed by the
XRD results below.
2µm
Fig. 3. SEM images of the Fe2O3 samples with single-metal additive before and
after the redox.
The XRD results for the samples with Mo, Zr and Mo+Zr before
and after the redox were shown in Fig. 5 and the standard XRD

ARTICLE IN PRESS
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
1081
Fe2O3+Zr+Mo before
Fe2O3+Zr before
Fe2O3+Zr+Mo after
Fe2O3+Zr after
Fe2O3+Mo after
Fe2O3+Mo before
Pure Fe2O3
Pure Fe3O4
20
30
40
50
60
70
20
30
40
50
60
70
2θ/degree
2θ/degree
Fig. 5. XRD spectra of Fe2O3–Mo, Fe2O3–Zr and Fe2O3–Mo–Zr before (a) and after (b) the redox; pure Fe2O3 and Fe3O4 as reference standards:
and ., ZrO2.
r
, Fe2(MoO4)3 ~, Fe2Mo3O8
Table 4
BET surface areas of Fe2O3-none, Fe2O3–Mo and Fe2O3–Mo–Zr before and after the redox.
Sample
BET surface area of the
fresh sample (m2/g)
BET surface area of the sample
after 4 cycles (m2/g)
BET surface area of the sample
after 10 cycles (m2/g)
Ratio of BET surface area decrease
after 4 cycles (m2/g)
Fe2O3-none
Fe2O3–Mo
3.96
3.54
0.16
2.87
8.24
–
0.96
0.19
0.40
–
Fe2O3–Mo–Zr
13.67
2.64
spectra of pure Fe2O3 and Fe3O4 were used for comparison.
Besides the main characteristic diffraction peaks assigned to
Fe2O3 and Fe3O4, some other diffraction peaks due neither to
Fe2O3 nor to Fe3O4 could also be observed for the Fe2O3–Mo,
Fe2O3–Zr and Fe2O3–Mo–Zr samples before and after the redox
cycle in Fig. 5a and b. E.g., for the samples before and after the
redox cycle, two obvious diffraction peaks ascribed to Fe2(MoO4)3
(JCPDS 35-0183) could be observed in Fig. 5a for Fe2O3–Mo–Zr
and Fe2O3–Mo, while a few obvious peaks for Fe2O3–Mo and
slight peaks for Fe2O3–Mo–Zr nearly in accordance with Fe2Mo3O8
(JCPDS 74-1429) could be found in Fig. 5b. Two slight diffraction
peaks ascribed to ZrO2 (JCPDS 65-1022) could be found in Fe2O3–
Zr before and after the cycle but did not exist in Fe2O3–Mo–Zr,
which may be that the amount of Zr added in Fe2O3 was too small
to be detected due to only amount of 2.5 mol% Zr cation in Fe2O3–
Mo–Zr. Accordingly, it is concluded that the oxidation number of
Mo that existed in the form of molybdate species varied from
Mo6+ before the redox to Mo4+ after the redox, and the Zr
additive in the form of ZrO2 presented in its sample before and
after the redox cycle. It is the valence change of Mo that prevented
the particles of the samples from sintering. Meanwhile, based on
our previous experimental result of Mo3+ existing in the reduced
fresh sample and the present experiment, we consider that the
following ionic reactions may take place: 2Mo6+ +3H2-
2Mo3+ +6H+ (only for the reduction step of the fresh sample)
and the subsequent reversible reaction of 2Mo3+ +H2O2
2Mo4+ +H2+O2ꢀ (for the oxidation step of the reduced iron
oxide). It is the reversible change of Mo3+2Mo4+ that
accelerated the water decomposition and further resulted in the
decrease of H2 production temperature, which here was called
as the catalytic effect of Mo catalysis. In addition to the Mo
catalysis in Fe2O3–Mo–Zr, another reason for the Fe2O3–Mo–Zr
sample having the best performances to produce hydrogen may
be the assisted catalysis of Zr in Fe2O3–Mo–Zr due to the
unchanged valence of Zr cation in the sample before and after
the redox.
The specific surface areas of the samples were measured with a
surface area analyzer by the adsorption of N2 at liquid N2
temperature (BET method). The detailed data of Fe2O3-none,
Fe2O3–Mo, Fe2O3–Mo–Zr before and after the redox cycle were
listed in Table 4. The data in Table 4 showed that BET specific
surface areas for the cation-modified samples decreased a little
compared with the none-modified Fe2O3 sample after four cycles,
proving that the addition of foreign metals to Fe2O3 could
effectively prevent the sintering of the particles. Whereas,
with the cycle times rising to 10, the specific surface area of
the Fe2O3–Mo–Zr sample changed from 8.24 after 4 cycles to
2.64 m2/g after 10 cycles. With increasing cycle time, a large
change of specific surface area did not cause the decrease of redox
performances of Fe2O3–Mo–Zr, which also confirmed that the
type of additive may be more important for improving the redox
performances of the modified sample.
4. Conclusions
The redox performances of the modified Fe2O3 samples with
single- or double-metal additives (Zr, Cr, Ag Mo and Ni; Mo–Zr,
Mo–Cr, Mo–Ag and Mo–Ni) were investigated. Among all the
additives in Fe2O3, the catalytic effect of Mo–Zr on hydrogen
production was the most effective. The reason for it is that the
addition of Mo to Fe2O3, on the one hand, could prevent the
sintering of the particles effectively, and on the other hand, could
promote the decomposition of water due to the valence change
of Mo cation (Mo3+2Mo4+) before and after the redox. E.g.,
Fe2O3–Mo–Zr exhibited the lowest H2 production temperature
(276 1C) at the rate of H2 formation of 250
highest rate of H2 formation (360.9–461.1
m
mol minꢀ1 Fe-gꢀ1, the
mol minꢀ1 Fe-gꢀ1
m
)
from the first to fourth cycle at 300 1C, the best cyclic stability of
hydrogen storage/release at operating temperature (no deactiva-
tion phenomenon after 10 cycles) and the high average capacity of
hydrogen storage in four cycles (4.73 wt%).

ARTICLE IN PRESS
X. Liu, H. Wang / Journal of Solid State Chemistry 183 (2010) 1075–1082
[4] D.K. Ross, Vacuum 80 (2006) 1084.
1082
Acknowledgments
[5] B. Sakintuna, F. Lamari-Darkrim, M. Hirscher, Int. J. Hydrogen Energy 32
(2007) 1121.
The authors acknowledge the financial supports of the National
Hi-Tech Research and Development Program (863) of China (no.
2007AA05Z116), the National Natural Science Foundation of China
(nos. 20873099, 20673082), the Scientific Research Foundation for
ROCS, SEM (no. 2006331), the Key Project of Science and
Technology of Shaanxi Province (no. 2005k07-G2) and the Natural
Science Foundation of Shaanxi Education Committee (no. 06JK167)
the National Science Foundation for Fostering Talents in Basic
Research (no. J0830417). The authors are also grateful to the
State Key Laboratory of Continental Dynamics for the SEM
measurements.
[6] J.M. Yan, X.B. Zhang, S. Han, H. Shioyama, Q. Xu, Angew. Chem. Int. Ed. 47
(2008) 2287.
[7] Z. Xiong, G. Wu, J. Hu, Adv. Mater. 16 (2004) 1522.
[8] G. Mpourmpakis, E. Tylianakis, G.E. Froudakis, Nano Lett. 7 (2007) 1893.
[9] H.M. Cheng, Q.H. Yang, C. Liu, Carbon 39 (2001) 1447.
[10] H. Shao, H. Xu, Y. Wang, X. Li, J. Solid State Chem. 177 (2004) 3626.
[11] A.M. Seayad, D.M. Antonelli, Adv. Mater. 16 (2004) 765.
[12] L. Zhou, Renew. Sust. Energy Rev. 9 (2005) 395.
[13] M. Sahlberg, C. Zlotea, P. Moretto, Y. Andersson, J. Solid State Chem. 182
(2009) 1833.
[14] H. Shao, Y. Wang, H. Xu, X. Li, J. Solid State Chem. 178 (2005) 2211.
[15] L. Zhou, Y. Zhou, Y. Sun, Int. J. Hydrogen Energy 29 (2004) 319.
[16] A. Zu¨ttel, Mater. Today 6 (2003) 24.
[17] K. Otsuka, C. Yamada, T. Kaburagi, S. Takenaka, Int. J. Hydrogen Energy 28
(2003) 335.
[18] H. Wang, S. Takenaka, K. Otsuka, Int. J. Hydrogen Energy 31 (2006) 1732.
[19] H. Wang, G. Wang, X. Wang, J. Bai, J. Phys. Chem. C 112 (2008) 5679.
[20] S. Takenaka, T. Kaburagi, C. Yamada, K. Nomura, K. Otsuka, J. Catal. 228 (2004) 66.
[21] V. Hacker, R. Vallant, M. Thaler, Ind. Eng. Chem. Res. 46 (2007) 8993.
[22] J.C. Ryu, D.H. Lee, K.S. Kang, C.S. Park, J.W. Kim, Y.H. Kim, J. Ind. Eng. Chem. 14
(2008) 252.
References
[1] L. Barreto, A. Makihira, K. Riahi, Int. J. Hydrogen Energy 28 (2003) 267.
[2] I. Dincer, Int. J. Hydrogen Energy 27 (2002) 265.
[3] M. Felderhoff, C. Weidenthaler, R. Helmolt, U. Eberle, Phys. Chem. Chem. Phys.
9 (2007) 2643.
[23] K. Otsuka, T. Kaburagi, C. Yamada, S. Takenaka, J. Power Sources 122 (2003) 111.