Role of particle sizes in hydrogen generation by the reaction of Al with water

Article abstract of DOI:10.1111/j.1551-2916.2010.03969.x

Four different-sized pure Al powders were used to react with distilled water at 55°C in a low vacuum, and their effect on the reaction dynamics and hydrogen-generation rate was investigated. It was found that the Al particle sizes not only affect the hydrogen-generation rate but also affect the induction time for the beginning of the reaction. The induction time strongly depends on the Al particle sizes, which obviously increased with the Al particle sizes. The hydrogen-generation rate increased with the decreasing Al particle sizes, because the small-sized Al powder has a high surface area. Mechanism analyses revealed that the long induction time for the large Al particles probably originates from a long time for their H diffusion to reach a saturation concentration in the bulk Al metal. The reaction rate of Al with water could be qualitatively explained by a shrinking core model.

Full text of DOI:10.1111/j.1551-2916.2010.03969.x

J. Am. Ceram. Soc., 93 [10] 2998–3001 (2010)
DOI: 10.1111/j.1551-2916.2010.03969.x
r 2010 The American Ceramic Society
ournal
J
Role of Particle Sizes in Hydrogen Generation by the
Reaction of Al with Water
Zhen-Yan Deng,w Li-Li Zhu, and Ye-Bin Tang
Energy Materials & Physics Group, Department of Physics, Shanghai University, Shanghai 200444, China
Yoshio Sakka, Jinhua Ye, and Rong-Jun Xie
National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan
Four different-sized pure Al powders were used to react with
distilled water at 551C in a low vacuum, and their effect on the
reaction dynamics and hydrogen-generation rate was investi-
gated. It was found that the Al particle sizes not only affect the
hydrogen-generation rate but also affect the induction time for
the beginning of the reaction. The induction time strongly
depends on the Al particle sizes, which obviously increased
with the Al particle sizes. The hydrogen-generation rate
increased with the decreasing Al particle sizes, because the
small-sized Al powder has a high surface area. Mechanism
analyses revealed that the long induction time for the large
Al particles probably originates from a long time for their H
diffusion to reach a saturation concentration in the bulk Al
metal. The reaction rate of Al with water could be qualitatively
explained by a shrinking core model.
hydrogen-generation materials for portable and mobile applica-
tions.
Previous works6,7 showed that metal Al particle surfaces can
be modified by fine oxide grains. The modified Al powder could
continuously react with water and generate hydrogen under
ambient condition. The reaction byproduct is chemically neu-
tral. In contrast to methanol and NaBH4, hydrogen generation
by the reaction of modified Al powder with water is simple and
cheap.8 Because particle sizes play an important role in the
reaction dynamics, in this work, Al powders with different par-
ticle sizes were used. The effect of particle sizes on hydrogen
generation by the reaction of pure Al powders with water was
investigated. At the same time, the possible reaction dynamics
mechanisms were analyzed.
II. Experimental Procedure
Four kinds of Al powders with the average sizes of 7.29, 15.08,
24.94, and 172.05 mm (99.9% purity, High Purity Chemical Co.,
Tokyo, Japan), respectively, were adopted in the present exper-
iment, where the Al particle sizes were measured using a laser
particle size analyzer (CILAS 1064, Orleans, France). Although
the modified Al powder could react with water and generate
hydrogen under an ambient condition,6 pure Al powder requires
special vacuum and temperature to react with pure water. In this
work, the hydrogen-generation experiment of pure Al powders
was carried out at a temperature of 551C. 0.2 g of pure Al pow-
der was used in each test, which was suspended in 270 mL of
distilled water in a Pyrex glass cell (Makuhari Rikagaku Garasu
Inc., Chiba, Japan), with agitating by a magnetic stirrer. At the
beginning of the experiment, the glass cell was evacuated to a
low vacuum of 0.03 bar. The gases evolved were determined
using a thermal conductivity detector gas chromatograph
(Model No. GC-8A, Shimadzu, Kyoto, Japan), which was con-
nected to the glass-made gas circulating line attached to the
Pyrex glass cell. The measurement uncertainties were about
0.05%.6 The schematic representation of the present hydro-
gen-generation equipment can be found in Hara et al.8
I. Introduction
ITH the environment pollution and the resultant global
climate warming, clean and renewable energy tech-
W
nologies have attracted much attention in the past few years.
Hydrogen is an ideal fuel, because its oxidation product (water)
is environmentally benign. However, storage and transportation
of hydrogen is a difficult problem, due to its low boiling point
(ꢀ2521C) and very poor compressibility.1,2 To solve this prob-
lem, scientists have endeavored much, e.g. to develop hydrogen-
storage materials and hydrogen-generation materials.1–3
Recently, considerable effort has been devoted to small fuel
cells and the kW-grade fuel cells as the energy sources for field-
work or military use due to their high-energy density.4,5 Because
of the small volume of these fuel cells, hydrogen-generation
materials are more suitable for them. Nowadays, there are two
main hydrogen-generation materials that are tentatively used in
small and kW-grade fuel cells: methanol and NaBH4. However,
both of them have some shortcomings in the application. For
example, reforming of methanol not only generates hydrogen
but also CO; CO has a serious poisoning effect on the anode of
the fuel cells.5 For NaBH4, it requires a definite amount of
NaOH as a stabilizing agent, and the hydrogen-generation
reaction of NaBH4 with water needs the aid of special catalysts.
Moreover, the high price of NaBH4 (BUS$55 kgꢀ1) also atten-
uates its competitiveness for commercial applications.3 There-
fore, it is of significance to develop other cost-effective
An X-ray diffractometry (XRD; Model No. RINT2000, Rig-
aku Co., Tokyo, Japan) was used to analyze the phases in pure
Al powders and those after reaction with water. Scanning elec-
tron microscopy (SEM) was used to observe the morphologies
of as-received and the reacted pure Al powders.
III. Results and Discussion
J. Ferreira—contributing editor
(1) Al Particle Morphologies and Hydrogen-Generation Tests
Figure 1 shows the SEM morphologies of four kinds of Al
powders with different average particle sizes. It can be seen that
the Al particle shape is not exactly spherical and some particles
Manuscript No. 27792. Received April 7, 2010; approved June 1, 2010.
wAuthor to whom correspondence should be addressed. e-mail: zydeng@shu.edu.cn
2998

October 2010
Rapid Communications of the American Ceramic Society
2999
1μm
10μm
(a)
(b)
20μm
100μm
(c)
(d)
Fig. 1. Scanning electron micrographs of as-received Al powders with the average sizes of (a) 7.29 mm, (b) 15.08 mm, (c) 24.94 mm, and (d) 172.05 mm,
respectively.
are elongated; there is a particle size distribution for each Al
powder. Figure 2 shows the gas evolution from the distilled
water containing as-received Al powders at 551C under an initial
vacuum of 0.03 bar. In addition to hydrogen, no other gases
were detected during the reaction. It is clear that there is an
induction time for the beginning of the reaction, and the initial
induction stage for different Al powers is very different. The
induction time noticeably increases with the increasing Al
particle sizes, as listed in Table I. After the beginning of the
reaction, the hydrogen-generation rate decreases with the
increasing Al particle sizes, mainly due to the larger sized Al
powder with a lower surface area. It takes B10.8 h to consume
87.4% of the metal Al at 551C for 7.29 mm Al powder. Table I
gives the maximum hydrogen-generation rate (MHGR) for
different-sized Al powders. 7.29 mm Al powder has the highest
MHGR, which is 9.3 cm3H2/minꢁ g-Al. Although this MHGR
value is lower than those (450 cm3H2/minꢁ g-Al) of the alloying
Al by doping B20 wt% Ga, In, etc. in water9 and the Al
in alkaline solution,10 however, the MHGR of the pure Al pow-
der could be increased up to 50–100 cm3H2/min ꢁ g-Al if the Al
particle sizes can be decreased to 1–2 mm (Z. Y. Deng et al.,
unpublished data).
(2) Induction Time
Figure 2 indicated that the induction time is an important time
interval for pure Al powders to completely react with water. In
fact, when Al particles are put into water, H2 accumulates and
forms uniformly distributed H2 bubbles at the Al:Al2O3 inter-
face by the reaction of metal Al with water in the front of the
hydration oxide film on Al particle surfaces, where the water is
produced by a reverse process of hydrolysis.7 It takes some time
for the gas pressure in the H2 bubbles at Al:Al2O3 interface to
reach a critical gas pressure, over which the hydrated oxide film
on Al particle surfaces would be broken.
100
7.29μm
80
15.08μm
60
Usually, pure Al powder does not react with water and gen-
erate hydrogen under normal atmospheric pressure even at a
24.94μm
40
Table I. Induction Time and Maximum Hydrogen-Generation
Rate (MHGR) for Different-Sized Al Powders
20
172.05μm
Al particle sizes (mm)
Induction time (h)
MHGR (cm3H2/minꢁ g-Al)
0
7.29
1.22
2.00
3.85
5.00
9.3
4.4
2.0
0.4
0
2
4
6
8
10
12
14
15.08
24.94
172.05
Reaction Time (hours)
Fig. 2. H2 evolution from distilled water using four different-sized Al
powders at 551C under an initial vacuum of 0.03 bar, where each test
used 0.2 g of Al powder.
wMHGR is obtained by the generated H2 volume (normalized in the standard
condition) divided by the time interval and the Al weight.

3000
Rapid Communications of the American Ceramic Society
Vol. 93, No. 10
temperature of 1001C.2 However, in vacuum condition, the gas
pressure in the glass reaction cell is much lower than the normal
atmospheric pressure so that the critical gas pressure in H2 bub-
bles decreases considerably. In this case, the reaction equilib-
rium gas pressure in H2 bubbles could exceed the critical gas
pressure such that the passive oxide film on Al particle surfaces
could be broken. This is why pure Al powder could continu-
ously react with pure water and generate hydrogen in vacuum.7
Before the H2 molecules accumulate to form H2 bubbles at
the Al:Al2O3 interface, H would diffuse into the bulk Al metal.
Of course, hydrogen permeability through the passive oxide film
is very slow and could be negligible.11 If a constant H2 accu-
0.6
0.5
0.4
0.3
0.2
0.1
0.0
7.29μm
15.08μm
24.94μm
mulation rate, SH , is assumed at the Al:Al2O3 interface, and the
2
172.05μm
saturation concentration of H in the bulk Al metal is set to Cs,
for an infinite diffusion speed, the saturation time, ts, of H in the
metal Al particle with a radius, r, can be written
0
2
4
6
8
10
Reaction Time (hours)
4
3
2 ꢂ pr3Cs
H2
8
Fig. 3. Plot of 1ꢀ(1ꢀR)1/3 versus the observed reaction time for four
different-sized Al powders at 551C under an initial vacuum of 0.03 bar,
where the dots and solid lines represent the experimental data and fitting
curves, respectively.
ts ¼
¼
rCs=SH
(1)
2
pr2S
3
Because the saturation of H in metal Al particles is the most
important process in the induction stage of the reaction, Eq. (1)
clearly indicates that the induction time is proportional to the
radius of metal Al particles, which is qualitatively consistent
with the results in Fig. 2 and Table I. In the practical situation,
the H diffusion observes the Fick’s laws and the diffusion speed
is limited. There is an H concentration gradient from the
Al:Al2O3 interface to the center of metal Al particle. Even so,
the tendency predicted by Eq. (1) is right. Moreover, the atoms/
molecules on the surface of large particles are in a lower energy
than those on the surface of small particles, and the hydration
reaction in the passive oxide film on large Al particles probably
is slower than that in small Al particles, leading to the longer
induction time of the larger sized Al powder.
(3) Possible Reaction-Controlling Mechanism
The reaction of pure Al powders with water can be categorized
as
A ðliquidÞ þ bB ðsolidÞ ! cC ðsolidÞ þ dD ðgasÞ
(2)
Because water in glass reaction cell was agitated by a mag-
netic stirrer and the H2 bubbles coming out made water move,
the reaction of Al particles with water is a fluid–solid particle
reaction. The conventional fluid–solid particle reaction has been
studied for a long time, which could be described by the
renowned shrinking core model (SCM).12–14 Detailed derivation
2μm
(a)
(b)
1μm
1μm
10μm
(c)
(d)
Fig. 4. Scanning electron micrographs of the pure Al powders with the average sizes of (a) 7.29 mm after reaction for 10.8 h, (b) 15.08 mm after reaction
for 11.8 h, (c) 24.94 mm after reaction for 11.7 h, and (d) 172.05 mm after reaction for 23.4 h, respectively, where the reaction temperature is 551C and the
initial vacuum is 0.03 bar.

October 2010
Rapid Communications of the American Ceramic Society
3001
different H2 bubble density at Al:Al2O3 interface for different-
sized Al particles, an issue required for further study.
Figure 5 shows the X-ray patterns of pure Al powder and that
after reaction with water. It can be seen that bayerite (Al(OH)3)
is the reaction byproduct of pure Al powders with water at 551C.
Therefore, the reaction equation can be written
Al
Bayerite
b
a
2Al þ 6H2O ! 2AlðOHÞ3 þ 3H2 "
(4)
This reaction produces 3.7 wt% hydrogen (weight ratio of gen-
erated H2 to Al and H2O).
IV. Conclusions
20
40
60
2θ (degree)
80
In this work, the effect of Al particle sizes on hydrogen gener-
ation by the reaction of pure Al powders with pure water was
investigated. The results showed that the induction time for the
beginning of the reaction increased with the increasing Al par-
ticle sizes, because the large-sized Al particles take a long time
for H to diffuse in saturation in the bulk Al metal. The hydro-
gen-generation rate decreased with the increasing Al particle
sizes, due to the large-sized Al powder with a low surface area.
The surface chemical reaction is the possible rate-limiting step
for the Al particle–water reaction system. SEM morphologies
of the reacted Al powders confirmed the fact that the breakage
of the H2 bubbles at Al:Al2O3 leads to the continuous reaction
of Al particles with water and generation of hydrogen.
Fig. 5. X-ray patterns for (a) as-received 24.94 mm Al powder, (b) those
in (a) after reaction with water at 551C for 11.8 h, where the initial vac-
uum is 0.03 bar.
of the SCM can be found in Levenspiel.13 Usually, the form of
the rate equation is determined by the control regime or the rate-
limiting step, which can be one of the following three: (1) diffu-
sion through the liquid film surrounding the solid particle, (2)
diffusion through the dense byproduct solid layer, and (3) chem-
ical reaction at the surface of the unreacted core.14 Because the
liquid in the present reaction is water only, and diffusion of the
reacting matter through the liquid film surrounding the solid
particle does not exist. At the same time, no dense oxide layer
(Fig. 4) existed on the Al particle surfaces after the reaction with
water. Therefore, the most possible rate-limiting step for the Al
particle–water system is the chemical reaction at the surface of
the unreacted Al core.
Acknowledgments
Zhen-Yan Deng would like to thank the financial supports of the Key Project
of STCSM (10JC1405300), the Research Fund for the Doctoral Program of
Higher Education of China (20093108110003), Shanghai Leading Academic
Discipline Project (S30105), Pujiang Talent Project (07pj14040), ‘‘211 Project’’
and Innovation Fund of Shanghai University.
If the solid particle is spherical and it reacts with the fluid
isothermally, when the surface chemical reaction is the rate-con-
trolling step, the rate equation can be written
References
1L. Schlapbach and A. Zuttel, ‘‘Hydrogen-Storage Materials for Mobile Appli-
1=3
1 ꢀ ð1 ꢀ RÞ ¼ kt
(3)
¨
cations,’’ Nature, 414 [6861] 353–8 (2001).
2Z. Y. Deng, J. M. F. Ferreira, and Y. Sakka, ‘‘Hydrogen-Generation Materials
for Portable Applications,’’ J. Am. Ceram. Soc., 91 [12] 3825–34 (2008).
3J. H. Wee, ‘‘A Comparison of Sodium Borohydride as a Fuel for Proton
Exchange Membrane Fuel Cells and for Direct Borohydride Fuel Cells,’’ J. Pow-
der Sources, 155 [2] 329–39 (2006).
where R is the fraction of the solid consumed in the reaction; k is
a constant related to the particle radius, solid density, and
concentration of A in bulk of fluid.14
Figure 3 shows the experimental data and linear fitting curves
of 1ꢀ(1ꢀR)1/3 versus the observed reaction time (excluding
induction time), t, for the reaction of pure Al powders with
water. In general, the theoretical linear fittings are in qualitative
agreement with the experimental data, especially for large-sized
Al powders. However, for small-sized Al powders, the experi-
mental data slightly deviate from the linear fitting. The possible
reason is that the present reaction experiment was performed in
a closed glass cell system, and its gas pressure was not a constant
and increased with the hydrogen evolution. For the small-sized
Al powders, the gas pressure change in the closed glass cell is
more obvious than that for large-sized Al powders for the same
reaction time, because there is more hydrogen generation for the
small-sized Al powders due to their high surface area.
Figure 4 shows the SEM morphologies of pure Al powders
with different particle sizes after the reaction with water at 551C,
under an initial vacuum of 0.03 bar. Compared with the
morphologies of pure Al powders in Fig. 1, the piece-like
morphologies on Al particle surfaces in Fig. 4 clearly show the
breakage of the Al particle surface layers after the reaction with
water. This gives direct evidence that there are small H2 gas
bubbles formed at the Al:Al2O3 interface when Al particles are
placed in the water; the continuous reaction of Al particles with
water is due to the breakage of these small H2 bubbles. Figure 4
also shows that the broken sizes of the passive surface layer on
Al particles are different for different-sized Al powders, which
increase with the Al particle sizes. This probably is related to the
4A. S. Patil, T. G. Dubois, N. Sifer, E. Bostic, K. Gardner, M. Quah, and C.
Bolton, ‘‘Portable Fuel Cell Systems for America’s Army: Technology Transition
to the Field,’’ J. Power Sources, 136 [2] 220–5 (2004).
5Q. F. Li, R. H. He, J. O. Jensen, and N. J. Bjerrum, ‘‘Approaches and Recent
Development of Polymer Electrolyte Membranes for Fuel Cells Operating Above
1001C,’’ Chem. Mater., 15 [26] 4896–915 (2003).
6Z. Y. Deng, Y. F. Liu, Y. Tanaka, J. H. Ye, and Y. Sakka, ‘‘Modification of Al
Particle Surfaces by g-Al2O3 and Its Effect on the Corrosion Behavior of Al,’’
J. Am. Ceram. Soc., 88 [4] 977–9 (2005).
7Z. Y. Deng, J. M. F. Ferreira, Y. Tanaka, and J. H. Ye, ‘‘Physicochemical
Mechanism for the Continuous Reaction of g-Al2O3 Modified Al Powder with
Water,’’ J. Am. Ceram. Soc., 90 [5] 1521–6 (2007).
8K. Hara, K. Sayama, and H. Arakawa, ‘‘UV Photoinduced Reduction of
Water to Hydrogen in Na2S, Na2SO3, and Na2S2O4 Aqueous solutions,’’
J. Photochem. Photobiol. A: Chem., 128, 27–31 (1999).
9O. V. Kravchenko, K. N. Semenenko, B. M. Bulychev, and K. B. Kalmykov,
‘‘Activation of Aluminum Metal and Its Reaction with Water,’’ J. Alloys Compd.,
397 [1–2] 58–62 (2005).
10L. Soler, A. M. Candela, J. Macana
´
s, M. Munoz, and J. Casado, ‘‘In Situ
Generation of Hydrogen from Water by Aluminum Corrosion in Solutions of
Sodium Aluminate,’’ J. Power Sources, 192 [1] 21–6 (2009).
11W. H. Song, J. J. Du, Y. L. Xu, and B. Long, ‘‘A Study of Hydrogen Per-
meation in Aluminum Alloy Treated by Various Oxidation Process,’’ J. Nucl.
Mater., 246 [2–3] 139–43 (1997).
12S. Yagi and D. Kunii, ‘‘Studies on Combustion of Carbon Particles in Flames
and Fluidized Beds’’; pp. 231–44 in Fifth Symposium (International) on Combus-
tion. Reinhold, New York, 1955.
13O. Levenspiel, Chemical Reaction Engineering, 3rd edition, Wiley, New York,
1999.
14P. K. Gbor and C. Q. Jia, ‘‘Critical Evaluation of Coupling Particle Size
Distribution with the Shrinking Core Model,’’ Chem. Eng. Sci., 59, 1979–87
(2004).
&