Effect of Al(OH)3 on the hydrogen generation of aluminum–water system

Abstract

Many metal oxides effectively modify metallic aluminum (Al) powders to produce hydrogen in neutral water at room temperature. In this study, hydrogen is generated even more rapidly from the reacted solution when the by-product Al(OH)3 (bayerite) is present. The influence of Al(OH)3 on the hydrogen generation during the aluminum/water reaction is highly dependent on the surface area and the crystallinity of Al(OH)3. High surface area and poor crystallinity of Al(OH)3 render the system a rapid hydrogen generation rate. The self-catalytic phenomenon can be explained by the formation of AlO(OH) (boehmite) from the reaction of by-product β-Al(OH)3 and the surface passive oxide layer Al2O3. Furthermore, by taking advantage of the exothermic heat from the Al/water reaction, hydrogen yield is able to reach 70% within 30 min in a common aqueous solution (pH < 10) containing fine Al(OH)3. This hydrogen generation technology relies on common, nontoxic, noncorrosive additives and is therefore a very promising system for creating a use-on-demand hydrogen source.

Full text

Effect of Al(OH)
3
on the hydrogen generation of aluminumewater system
Hsin-Te Teng
a
, To-Ying Lee
a
, Yu-Kuang Chen
a
, Hong-Wen Wang
a
,
*
, Guozhong Cao
b
a
Department of Chemistry, Master Program of Nanotechnology, Center for Nanotechnology, Chung-Yuan Christian University, Chungli 320, Taiwan, ROC
b
Department of Materials Science and Engineering, University of Washington, Seattle, WA, USA
highlights
<Novel effect of Al(OH)
3
powders on Al/H
2
O system for hydrogen generation was demonstrated.
<Fine and poor crystalline Al(OH)
3
powder acts an effective additive to the Al/H
2
O system.
<A proposed reaction mechanism for the enhancement of hydrogen generation was given.
<The reaction is able to rapidly produce hydrogen in a speed of 70% yield within 30 min.
article info
Article history:
Received 10 May 2012
Received in revised form
14 June 2012
Accepted 16 June 2012
Available online 24 July 2012
Keywords:
Bayerite
Aluminum
Hydrogen generation
Water
abstract
Many metal oxides effectively modify metallic aluminum (Al) powders to produce hydrogen in neutral
water at room temperature. In this study, hydrogen is generated even more rapidly from the reacted
solution when the by-product Al(OH)
3
(bayerite) is present. The influence of Al(OH)
3
on the hydrogen
generation during the aluminum/water reaction is highly dependent on the surface area and the crys-
tallinity of Al(OH)
3
. High surface area and poor crystallinity of Al(OH)
3
render the system a rapid
hydrogen generation rate. The self-catalytic phenomenon can be explained by the formation of AlO(OH)
(boehmite) from the reaction of by-product
b
-Al(OH)
3
and the surface passive oxide layer Al
2
O
3
.
Furthermore, by taking advantage of the exothermic heat from the Al/water reaction, hydrogen yield is
able to reach 70% within 30 min in a common aqueous solution (pH <10) containing fine Al(OH)
3
. This
hydrogen generation technology relies on common, nontoxic, noncorrosive additives and is therefore
a very promising system for creating a use-on-demand hydrogen source.
Ó2012 Elsevier B.V. All rights reserved.
1. Introduction
The corrosion of metal Al in aqueous alkaline solutions has long
been known as a good way to generate hydrogen to produce clean
fuel [1]. Recently, there has been increasing interest in the reaction
of the Al/water system for generating hydrogen [2e13] using either
AleGa alloys [2],AleGaeIn alloys [3], KCl/NaCl [4,5] or metallic
oxide modifiers such as
g
-Al
2
O
3
and calcined boehmite AlO(OH)
[6e12]. We have reported that many metallic oxideemodified Al
exhibited far better hydrogen generation than those of pristine Al in
pure water at room temperature [13]. The key issue in the Al/H
2
O
reaction is breaking the surface’sAl
2
O
3
passive layer. Oxide modi-
fiers or catalysts mixed with Al powders are an excellent way to
produce hydrogen in neutral water at room temperature [6e13] .
However, the size of the metallic Al particles, the choice of modifier
materials, the size of the modifier particles, the mixing duration,
the reaction temperature, and the purity of water all influence the
corrosion of metal Al in water and therefore influence hydrogen
production. The optimal conditions for generating an adequate
amount of hydrogen to meet the use-on-demand system is still
challenging for the Al/H
2
O system using oxide modifiers.
One gram of Al will produce 1.36 l of H
2
when completely
reacted with water in ambient conditions. However, the direct
reaction of metallic Al with pure water is difficult due to the
surface’s passive Al
2
O
3
layer. Grinding can remove the oxide layer
and expose a fresh surface of aluminum metal. Hydrogen will then
be generated according to reaction (1):
2Al þ6H
2
O/2Al(OH)
3
þ3H
2
(1)
However, the surface will be passivated quickly by the
increasing Al(OH)
3
(bayerite) layer.
Reaction (1) generates hydrogen easily under ambient condi-
tions when modifiers are used; however, this reaction takes many
hours to reach 100% yield [7e13]. Deng et al. [7e10] confirmed the
role of the catalyst/modifier
g
-Al
2
O
3
in enhancing the hydrogen
*Corresponding author. Tel.: þ886 3 2653310; fax: þ886 3 2653399.
E-mail address: hongwen@cycu.edu.tw (H.-W. Wang).
Contents lists available at SciVerse ScienceDirect
Journal of Power Sources
journal homepage: www.elsevier.com/locate/jpowsour
0378-7753/$ esee front matter Ó2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jpowsour.2012.06.077
Journal of Power Sources 219 (2012) 16e21

production rate and the effect of elevated temperatures. The
mechanism by which
g
-Al
2
O
3
emodified Al powders produce
hydrogen when reacted with water can be explained by the
uniform corrosion model [8,14]. Using modifiers such as
g
-Al
2
O
3
[7e10] or TiO
2
[13] causes the dense passive alumina layer to
transform into boehmite (AlO(OH)) according to the following
reaction:
Al
2
O
3
þH
2
O/2AlO(OH) (2)
Furthermore, the boehmite will then react with metal Al to
generate hydrogen:
2Al þ6AlO(OH) /4Al
2
O
3
þ3H
2
(3)
In fact, the combination of reaction (2) 4 and reaction (3) results
in reaction (4), which is thermodynamically unfavorable at room
temperature but is favorable between 277

C and 477

C[15].
2Al þ4H
2
O/2AlO(OH) þ3H
2
(4)
In a recent finding [16], sodium stannateedoped sodium
hydroxide solution effectively promoted hydrogen generation from
the Al/H
2
O system by weakening the oxide layer using sodium
hydroxide and oxidizing Al by reducing the SnðOHÞ
62
ions.
However, the continuous and subsequent hydrogen generation
reactions require more sodium stannate, and the accumulation of
Sn metal waste could also interfere with the practical application of
this system. A fascinating report by Rosenband [17] showed that 6-
m
m-sized Al in water could yield 100% hydrogen in less than 5 min
by their in-house activated treatment on Al powders using
a lithium-based thermo-chemical process. Sodium aluminate
(NaAlO
2
) was also demonstrated an excellent alternative way other
than NaOH to produce hydrogen in short time [18]. Especially,
when 0.01 M NaAlO
2
combined and 20 g dm
3
Al(OH)
3
suspension,
the system reached a maximum rate of 221 cm
3
min
1
and 100%
yield at pH ca. 12.1 in 77 min [19]. Thus, a use-on-demand hydrogen
source that uses the Al/water system could be realized in near
future. However, these systems relied on high pH solution
(pH >12) which might also be corrosive to the environment. Since
the dissociation of NaAlO
2
in water can lead to the formation of
Na
þ
,OH

and AlO(OH) species, the generation of hydrogen is likely
due to the high pH value. The role of AlO(OH) in this case has been
overlooked. It is challenging to produce hydrogen rapidly from Al/
water system without the use of high pH alkaline solution.
In this study, we started from a TiO
2
emodified Al powders which
produced hydrogen in water easily [13].The system can continuously
generate hydrogen ina neutral water system at roomtemperature by
just adding additional batches of Al, i.e., it self-catalyticallygenerates
hydrogen at an increasingly rapid rate without requiring the high pH
alkaline solution,thereby reducing the cost of producinghydrogen as
well as the corrosive problem to the environment. The key to this
issue is the reaction kinetics of Al(OH)
3
.
2. Experimental procedures
Deionized (DI) water (pH ¼w7.0, resistivity >18 M
U
cm) and tap
water (pH ¼6.3e7.5, resistivity <4.5 k
U
cm) were used throughout
this study. Metallic Al powders with a particle size of 45
m
m (Alfa
Aesar, 325 mesh, 99.5% purity) were used for all of the experiments.
Metallic Al powder (1 g) was hand-mixed with 1 g modifier TiO
2
nanocrystals (P90, Degussa, 22 nm) or reagent TiO
2
powders (Shi-
makyu’s Pure Chemicals, 99%, 375 75
m
m) for 3 min to create the
starting materials. The reagent TiO
2
emodified Al powders were
added to 200 ml of deionized water and sealed in a conical flask.
After the completion of this initial reaction, the 1st batch of
unmodified Al (1 g) powders and later batches (also pristine Al
powders) were added consecutively to the same flask after the
completion of the previous reaction, which allowed the accumula-
tion of the by-product Al(OH
3
) from previous reactions, thereby
increasing its overall concentration in the flask. All of the experi-
ments were carried out at 25

C1

C using circulating water bath.
No magnetic stirring was employed within the reaction flask. The
hydrogen generated was measured using a silicone tube passing
through a water bath at room temperature to a precision gas flow
meter (Alicat Scientific, Inc.) that automatically recorded the output
data in a notebook computer everysecond until the reaction reached
100% yield (hydrogen) (1360 ml g
1
Al). The cooling water bath is
used to minimize the water vapor. Similar experiments were also
carried out for the P90emodified Al system in tap water.
Two commercialreagent Al(OH)
3
powders (Acros, USA, 5
m
m, 99%
purity, and NanoAmor, USA, 0.5
m
m, 98% purity) were also used to
compare their effects on the reaction of the Al/H
2
O system. The size
effect of Al(OH)
3
was tested as follows. First, two self-made Al(OH)
3
powders were obtained from as-received 5-
m
m Al(OH)
3
commercial
powders. That is, the Acros 5-
m
m Al(OH)
3
powders (2.88 g) and
deionized water (200 ml) were mixed and stirred for 24 h. The
solution was divided into the “top”100 ml and “bottom”100 m l
immediatelyafter the stirring process. Both solutions were dried,and
the powders obtained (0.13 g) were employed to test their effect on
hydrogen generation. These four dried powders (i.e., 5
m
m, “top”,
“bottom”, and 0.5
m
m) were hand-mixed with Al (1 g) and added to
50 ml deionized water to measure their ability to generate hydrogen.
The “top”Al(OH)
3
powder is smaller than that of “bottom”one after
24 h stirring, as seen in the FESEM images in Section 3.2. The BET
surface area analysis for the “Acros”5-
m
m, “top”,“bottom”, and the
“NanoAmor”0.5-
m
m, Al(OH)
3
powders were0.32 m
2
g
1
,6.65m
2
g
1
,
1.83m
2
g
1
, and 6.25 m
2
g
1
, respectively. The5-
m
m Al(OH)
3
powders
clearly broke down and their surface area increased after stirring for
24 h. The hydrogen generation results will be detailed in Section 3.2.
For the experiments of different Al/water weight ratios on the
hydrogen generation rate without temperature control, 3 g of metal Al
powders were added into a reduced amount of water to evaluate the
effect of the heat from the in-situ exothermic reaction on hydrogen
generation. The weight ratio of Al to deionized water was varied
between 1:5,1:10,1:20, and 1:30 (i.e., 3 g Al in 15 ml, 30 ml, 60 ml and
90 ml of water), where 20 g Al(OH)
3
(NanoAmor, USA, 0.5
m
m, 98%
purity) was first added right before the addition of the Al powders. All
of the experiments werecarried out at room temperature (25

C1

C)
using circulating water bath. The water displacement method was used
to measure hydrogen generation using two 2000-ml cylinder meters
(reading data error of less than 20 ml) because of the speed of the
hydrogen generation. After reacting with water, the product powders
werecharacterizedwithananalyticalPW3040/60X’Pert Pro X-ray
diffractometer using Cu K
a
radiation and a Ni filter over the angular
range of 10

e80

(2
q
)atascanningrateof4

min
1
and a step size of
0.01

. Field-emission scanning electron microscopy (FESEM, JEOL JSM
7600F) was employed to characterize the morphology of the powders.
The surface areas of all Al(OH)
3
powders were measured using nitrogen
adsorptionedesorption isotherms on an ASAP 2020 (Micromeritics
Instruments, USA) apparatus. All of the samples were degassed at
120

C prior to the BrunauereEmmetteTeller (BET) measurements.
3. Results and discussion
3.1. Enhanced hydrogen generation from the addition of
consecutive Al batches
Fig. 1 shows the amount of hydrogen generation from the
reaction of TiO
2
(reagent)emodified Al powders and deionized
H.-T. Teng et al. / Journal of Power Sources 219 (2012) 16e21 17

water (200 ml) at 25

C1

C (Al:H
2
O¼1:200). The initial batch of
modified Al (1 g) powder corroded away 50% approximately 12 h
after adding it to the deionized water. For the 1st batch of
unmodified Al powder (1 g), the 50% yield of hydrogen occurred at
approximately 3 h. For the 2nd and 3rd batches of pristine Al
powders, the 50% yield of hydrogen occurred at approximately 2 h
and reacted increasingly faster than those of the initial and 1st
batch. The 4th, 5th, and 6th batches of unmodified Al powders
yielded 50% hydrogen at 2 h, 1.5 h, and 1.5 h, respectively. The 7th,
8th and 9th batches of unmodified Al powders had almost identical
generation curves over time and yielded 50% hydrogen in less than
1.5 h. Thus, hydrogen generation was accelerated as the consecutive
batches of unmodified Al powders were added to the solution and
the by-product Al(OH)
3
accumulated. The hydrogen generation rate
reached a maximum after the 7th consecutive batch, resulting in
8.5 ml min
1
g
1
Al calculated from the 50% yield.
Fig. 1 also shows the increasing pH values during hydrogen
generation from the reaction of reagent TiO
2
emodified Al powders
and the subsequent unmodified Al batches in deionized water. The
pH values for the solution after the termination of reaction were
recorded for each Al batch. The pH value increased slightly from the
initial pH of 6.92 to 9.52. The slight increase of the pH value was
considered to be due to the limited dissociation of the by-product
Al(OH)
3
in water. However, the pH value ultimately reaches equi-
librium, and the generation rate of hydrogen no longer increases. To
elucidate the effect of the pH value alone on the Al/H
2
O reaction,
a clear deionized water (200 ml) at pH 9.60 was prepared using
NaOH (SigmaeAldrich). However, this high pH did not assist
hydrogen generation (see curve “pH ¼9.60”in Fig. 1).
Fig. 2 shows that the reaction of TiO
2
(P90)emodified Al
powders in neutral tap water (200 ml) at 25

C1

C
(Al:H
2
O¼1:200) generates hydrogen in a manner similar to the
systems in Fig. 1. The initial modified batch of Al (1 g) powder had
corroded 50% approximately 10 h after adding it to the tap water.
However, for the 1st batch of unmodified Al powder, the 50% yield
of hydrogen occurred at approximately 2.5 h, which was the same
for the 2nd batch; both were much faster than that of the initial
batch. The 3rd, 4th, 5th and 6th batches of unmodified Al powders
yielded 50% hydrogen at 1.5 h, 1 h, 1 h and 1 h, respectively. The
generation curves over time were almost identical for the 7the12th
batches of unmodified Al powders, resulting in 50% yield in less
Fig. 1. TiO
2
(reagent)emodified Al (1 g, initial) and consecutive batches of Al (1 g, 1st
and after) in deionized water at room temperature. The batch “pH ¼9.60”was carried
out for comparison purpose only, which is simply a basic solution.
Fig. 2. TiO
2
(P90)emodified Al (1 g, initial) and consecutive batches of pristine Al (1 g,
1st and after) in tap water at room temperature. The batch “pH ¼9.60”was carried out
for comparison purpose only, which is purely aqueous alkaline solution.
Fig. 3. Effect of filtered reaction solution and two different Al(OH)
3
powders on the
hydrogen generation of Al/H
2
O system.
Fig. 4. Effect of different sized commercial Al(OH)
3
powders on the hydrogen gener-
ation on Al/H
2
O system.
H.-T. Teng et al. / Journal of Power Sources 219 (2012) 16e2118

than 1 h. Thus, hydrogen generation was accelerated as the
consecutive batches of unmodified Al powders were added to the
solution, and the rate reached its maximum w12.3 ml min
1
g
1
Al
calculated from 50% yield of hydrogen for the 7the12th batches.
Fig. 2 also shows the pH changes during the hydrogen genera-
tion from the reaction of TiO
2
(P90)emodified Al powders and the
subsequent batches of unmodified Al powders in neutral tap water.
The pH values of the solutions after the termination of reaction
were recorded for each Al batch. The pH value increased slightly
from its initial pH of 6.84 to 9.42. Again, the slight increase of pH
value was considered to be due tothe limited dissociation of the by-
product Al(OH)
3
in water. However, the pH value ultimately reaches
equilibrium, and the generation rate of hydrogen no longer
increases. Clear tap water with pH at 9.60 was also prepared using
NaOH (SigmaeAldrich) to determine the effect of pH value alone on
the reaction. However, pH at 9.60 does not assist the hydrogen
generation, as shown by the curve “pH ¼9.60”in Fig. 2.
From literature [1,6], to assist hydrogen generation from Al/H
2
O
reaction, alkaline solution at pH greater than 11 is desired. pH lower
than 10 shows little effect on hydrogen generation as the curve
“pH ¼9.60”shown above. Obviously, the by-product Al(OH)
3
plays
an important role on the consecutive runs of corrosion reactions.
However, the effect of Al(OH)
3
on the reaction of Al/water for the
enhanced hydrogen generation is clearly not due to the increased
pH of the solution; the alternative factors affecting hydrogen
production deserved further studies. Although the initial batch of
Al was modified by TiO
2
, the later experiments show that TiO
2
is
not the major contributor to the consecutive enhanced hydrogen
generation. More TiO
2
actually deteriorates the hydrogen genera-
tion rate of the later consecutive reactions. In fact, the initial batch
of Al could be modified by using commercial Al(OH)
3
powders and
the results as Figs. 1 and 2 are still valid.
3.2. The role of Al(OH)
3
in the solution for the subsequent batches
Two Al(OH)
3
aqueous solutions were prepared by adding the
commercial as-received Al(OH)
3
powders (Acros, 5
m
m, 2.88 g,
gibbsite phase) into a 200 ml deionized water and stirred for 1 h to
Fig. 5. FESEM for Al(OH)
3
powders, (a) commercial as-received 5
m
m, (b) top from stirred 5
m
m for 24 h, (c) bottom from stirred 5
m
m for 24 h, (d) commercial as-received 0.5
m
m,
and (e) rod-like Al(OH)
3
powders on Al surface after reaction.
H.-T. Teng et al. / Journal of Power Sources 219 (2012) 16e21 19

evaluate the role of Al(OH)
3
on the hydrogen generation. The upper
100 ml aqueous solution of the stirred Al(OH)
3
solution was taken
immediately and 30 min after settlement using pipet. Fig. 3 shows
the distinct differences in these two Al(OH)
3
solutions on the
hydrogen generation from the reaction of unmodified Al (1 g) and
the solution. The Al(OH)
3
aqueous solution taken after 30 min
settlement did not show significant hydrogen generation even after
20 h reaction time. The curve “reaction solution”in Fig. 3 shows the
hydrogen generation behavior from the reaction of 1 g pristine Al
with a previously reacted solution (200 ml) that was filtered and
had no visible Al(OH)
3
residuals. Neither the pH value (9.33) nor the
ions in the solution assisted in generating hydrogen. Thus, the
Al(OH)
3
residuals must help generate hydrogen. The solubility of
Al(OH)
3
in the two solutions should be the same, and the clear
solution without the solid Al(OH)
3
particles exerted little effect on
the Al powders to generate hydrogen. Fig. 3 clearly shows the small
white Al(OH)
3
particles exhibit significant effects on the Al surface
and greatly enhance the hydrogen generation reaction.
Fig. 4 shows the effect of the two stirred Al(OH)
3
powders (“top”
and “bottom”, stirred for 24 h, see Section 2) and the two
commercial as-received Al(OH)
3
powders on hydrogen generation.
Clearly, the as-received 5-
m
m Al(OH)
3
powders exerted little effect
on hydrogen generation from the Al/water system. However, when
the 5-
m
m Al(OH)
3
powders were stirred in water and divided into
“top”and “bottom”parts, “top”part exerted significant effects on
the metal Al surface and improved the Al/water reaction. The “top”
Al(OH)
3
powder is a breakdown species from as-received 5-
m
m. As
stated in Section 2, Experimental procedure, that the surface area of
“top”was much higher than that of “bottom”and original as-
received 5-
m
m after 24 h stirring. Their images were shown in
the following FESEM images, Fig. 5 and will be discussed shortly.
The commercial as-received 0.5
m
m Al(OH)
3
powders also showed
a typical three-stage curve of hydrogen generation [11]. i.e. the
“induction period”,“fast reaction”, and “slow reaction”. The effect
of as-received 0.5
m
m Al(OH)
3
seems less effective than that of “top”
sample, though their surface areas were similar, 6.25 m
2
g
1
vs.
6.65 m
2
g
1
, respectively. The “induction period”of “top”sample in
Fig. 3 was very short. Similarly, the “induction period”of consec-
utive batches in Figs. 1 and 2 were almost not present. There must
be a good reason for this behavior that it happened right away
when pristine Al contacted Al(OH)
3
aqueous solution.
Fig. 5 (a)e(e) shows all of the Al(OH)
3
powders obtained in this
study. Fig. 5 (a), (b) and (c) shows the FESEM images for the as-
received 5-
m
m Al(OH)
3
powders, and their “top”and “bottom”
counterparts. The powder from the “top”of the solution is much
smaller than that from the “bottom”, which consisted of very large
particles. Fig. 5 (d) and (e) shows the as-received 0.5-
m
m Al(OH)
3
powders and the fine, rod-like Al(OH)
3
product on the surface of
the Al powder after reaction, respectively. These images are con-
sisted with those of gibbsite and bayerite reported in the literature
[20]. XRD was performed to evaluate the differences in crystallinity
of these powders. Fig. 6 (a)e(f) shows the crystalline phase of
commercial as-received Al(OH)
3
powders and the by-product from
Al/water reaction. It turned out that the by-product from Al/water
reaction was
b
-Al(OH)
3
, bayerite and the commercial as-received
Al(OH)
3
powders were all
a
phase, gibbsite. As shown in Fig. 6,
the 5-
m
m Al(OH)
3
powders and “bottom”powders exhibited strong
crystallinity. The “top”, 0.5-
m
m Al(OH)
3
and
b
-Al(OH)
3
from Al/
water reaction exhibited relatively poor crystallinity. After stirring
for 24 h, the smaller-sized particles in the 5-
m
m Al(OH)
3
powders
settled down more slowly that the larger ones. Fig. 4 demonstrated
that the smaller-sized Al(OH)
3
powder with higher surface area (as
the “top”sample shown) was important for enhancing the
hydrogen generation of the Al/water system. Figs. 1 and 2 showed
that the by-product
b
-Al(OH)
3
from the Al/H
2
O reaction effectively
promotes hydrogen generation. The BET surface area of the by-
product
b
-Al(OH)
3
from the Al/water reaction was measured to
29.92 m
2
g
1
. The summary data of all Al(OH)
3
powders used in this
study is shown in Table 1. From these results, it is clear that the solid
Al(OH)
3
material with poor crystallinity as well as largesurface area
are important for catalyzing the reaction rate of Al/water system.
3.3. Catalytic mechanism of Al(OH)
3
The key to improving the rate of hydrogen generation is dis-
rupting the protective passive Al
2
O
3
layer on the surface of metal Al
powder, whether that happens through a phase change or whether
it is broken by the modifiers. The uniform corrosion mechanism
states that the entire native oxide film is hydrated at room
temperature in the initial stage [8,14] when modifiers are added,
according to reaction (2). AlO(OH) (boehmite) greatly facilitates the
production of hydrogen in the Al/H
2
O system at room temperature,
while the 5-
m
m Al(OH)
3
did not [13]. Once the AlO(OH) (boehmite)
was formed, the front of it reacted with the Al to produce Al
2
O
3
and
hydrogen in the contact area (reaction (3)). Reactions (2) and (3)
continued until all of the Al metal was completely consumed, as
long as the boehmite was still being produced from reaction (2).
The above reactions occurred when the initial 1 g of modified Al
was added to water. However, when the 1st 1 g of unmodified Al
was put in the same flask after the completion of the initial batch,
the hydrogen generation was greatly enhanced due to the fine by-
product
b
-Al(OH)
3
, as shown in Figs. 1 and 2. The hydrogen
generation reaction is increasingly rapid as the later batches of Al
are added to the same flask. This result contradicts the report by
Soler et al. [19], where the consecutive experiments performed
after the first run showed lower rates. They claimed that the slower
hydrogen production rates of the later runs were smaller because of
the slower diffusion of species in a more viscous media due to the
accumulated Al(OH)
3
mud in the 75 ml solution. We thought that
the contradiction may come from the water quantity (200 ml) used
in our study, where rapid diffusion was possible.
We propose a mechanism for these rapid hydrogen generations
as shown in the following reaction (5):
Al
2
O
3
þAlðOHÞ
3
/3AlOðOHÞ
D
G¼39:4 kJ mol
1
(5)
Fig. 6. XRD for (a) by-product of Al/water after reaction using as-received 0.5
m
m
Al(OH)
3
, (b) by-product of Al/water after reaction, (c) the filtrated residuals from
“bottom”sample, (d) the filtrated residuals from “top”sample, (e) as-received 5
m
m
Al(OH)
3
, and (f) as-received 0.5
m
m Al(OH)
3
.
H.-T. Teng et al. / Journal of Power Sources 219 (2012) 16e2120

The free energy change was calculated based on the data of
g
-
alumina, bayerite, and boehmite from Ref. [21]. When the fine
b
-
Al(OH)
3
powders in water contact the Al
2
O
3
surface layer of
metallic Al, these two reactants form AlO(OH) according to reaction
(5), where
b
-Al(OH)
3
may simply lose its OH

and H
þ
and become
one AlO(OH). The species of OH

and H
þ
were highly mobile and
migrated into the Al
2
O
3
layer and formed the other two AlO(OH),
whose reaction formula could also be described in reaction (2).Itis
also possible that the fine and poor crystalline Al(OH)
3
particles
enhance the ionic migration of H
þ
ions from the bulk H
2
O to the Al
surface. It’s known from Bunker’s report [14] that the hydroxide
ions (or water molecules) are the mobile species in the films (rather
than H
þ
,O
2
,orAl
3þ
) and that transport rates are sufficiently rapid
to account for the kinetics of Al corrosion and pitting processes.
Coarse Al(OH)
3
exhibited little effect on the hydrogen production
due to its low reactivity. The coarse crystalline Al(OH)
3
particles
require at least 1 h to be broken down in the water and become
effective for generating hydrogen from the Al/water system. The
by-product
b
-Al(OH)
3
from the reaction of Al/water is poorly
crystalline and has a large surface area; thus, this by-product is
quite effective for allowing reaction (5) to proceed. To compare the
effect between gibbsite (
a
) and bayerite (
b
) is trivial, since their
crystal structure differ only in the way that the double layers of
coordinated OH

and Al
þ3
are stacked to form the three-
dimensional crystal units [20]. It is the surface area of Al(OH)
3
that highly dominates the hydrogen generation of Al/water system.
By utilizing the in-situ self-heating exothermic reaction with a large
quantity of initial Al powder and a minimum water content (3 g
Al:15 g H
2
O) containing fine
b
-Al(OH)
3
, we are able to produce
more than 70% yield hydrogen within 30 min without using alka-
line solution (pH <10). The setup is able to produce hydrogen
rapidly and yet no corrosive problem to the environment, making it
an ideal candidate of hydrogen source for a use-on-demand system.
4. Conclusions
We have demonstrated the novel effect of Al(OH)
3
powders on
the Al/H
2
O system for generating hydrogen. Fine, poorly crystalline
Al(OH)
3
powder is an effective additive to the Al/H
2
O system. We
also proposed a reaction mechanism to explain the enhancement of
hydrogen generation in this system. By using high surface area,
poor crystalline Al(OH)
3
powders in aqueous solution, Al
2
O
3
surface layer of metallic Al powders are readily reacted, making
rapid hydrogen generation possible without using highly corrosive
alkaline solution. Further promotion of hydrogen generation rate
could be achieved by taking advantage of its in-situ exothermic
reaction as well as special designed Al(OH)
3
powders having high
surface area.
Acknowledgments
This work was supported financially by the National Science
Council (NSC99-2113-M-033-003-MY3) and the CYCU Distinctive
Research Area project (grant CYCU-98-CR-CH).
References
[1] D. Belitskus, J. Electrochem. Soc. 117 (1970) 1097e1099.
[2] J.M. Woodall, US 0056986, US 0063597 (2008).
[3] A.V. Parmuzina, O.V. Kravchenko, Int. J. Hydrogen Energy 33 (2008)
3073e3076.
[4] E. Czech, T. Troczynski, Int. J. Hydrogen Energy 35 (2010) 1029e1037.
[5] B. Alinejad, K. Mahmoodi, Int. J. Hydrogen Energy 34 (2009) 7934e7938.
[6] A. Chaklader, US 0048548 A1, 6440385 B1, 6582676 B2, WO 2002014213
(2002).
[7] Z.Y. Deng, Y.F. Liu, Y. Tanaka, J. Ye, Y. Sakka, J. Am. Ceram. Soc. 88 (2005)
977e979.
[8] Z.Y. Deng, J.M.F. Ferreira, Y. Tanaka, J. Ye, J. Am. Ceram. Soc. 90 (2007)
1521e1526.
[9] Z.Y. Deng, J.M.F. Ferreira, Y. Sakka, J. Am. Ceram. Soc. 91 (2008) 3825e3834.
[10] Z.Y. Deng, Y.B. Tang, L.L. Zhu, Y. Sakka, J.H. Ye, Int. J. Hydrogen Energy 35
(2010) 9561e9568.
[11] P. Dupiano, D. Stamatis, E.L. Dreizin, Int. J. Hydrogen Energy 36 (2011)
4781e4791.
[12] K. Mahmoodi, B. Alinejad, Int. J. Hydrogen Energy 35 (2010) 5227e5232.
[13] H.W. Wang, H.W. Chung, H.T. Teng, Guozhong Cao, Int. J. Hydrogen Energy 36
(2011) 15136e15144.
[14] B.C. Bunker, G.C. Nelson, K.R. Zavadil, J.C. Barbour, F.D. Wall, J.P. Sullivan,
C.F. Windisch Jr., M.H. Engelhardt, D.R. Baer, J. Phys. Chem. B 106 (2002)
4705e4713.
[15] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, E. Artacho, J. Phys.
Chem. B 106 (2002) 5155e5162.
[16] H.B. Dai, G.L. Ma, H.J. Xia, P. Wang, Energy Environ. Sci. 4 (2011) 2206e2212.
[17] V. Rosenband, A. Gany, Int. J. Hydrogen Energy 35 (2010) 10898e10904.
[18] L. Soler, A.M. Candela, J. Macanas, M. Munoz, J. Casado, J. Power Sources 192
(2009) 21e26.
[19] L. Soler, A.M. Candela, J. Macanas, M. Munoz, J. Casado, Int. J. Hydrogen Energy
34 (2009) 8511e8518.
[20] S. Goldberg, J.A. Davis, J.D. Hem, in: Garrison Sposito (Ed.), The Environment
Chemistry of Aluminum, Lewis Publishing, 1996, pp. 271e331 (Chapter 7).
[21] G. Lefèvre, M. Due, P. Lepeut, R. Caplain, M. Fédoroff, Langmuir 18 (2002)
7530e7537.
Table 1
Summary of Al(OH)
3
powders used in this study.
ID Commercial as-received
5
m
m (Acros)
“Top”of 5
m
m (Acros) “Bottom”of
5
m
m (Acros)
Commercial as-received
0.5
m
m (NanoAmor)
By-product after
Al/water reaction
Surface area (m
2
g
1
) 0.32 6.65 1.83 6.25 29.92
Crystalline phase Gibbsite Gibbsite Gibbsite Gibbsite Bayerite
Assigned name
a
-Al(OH)
3
a
-Al(OH)
3
a
-Al(OH)
3
a
-Al(OH)
3
b
-Al(OH)
3
Crystallinity Very strong Poor Very strong Poor Very poor
H.-T. Teng et al. / Journal of Power Sources 219 (2012) 16e21 21