Study on degradation of filamentary Ni catalyst on hydrolysis of sodium borohydride

Article abstract of DOI:10.1016/j.jallcom.2004.02.009

An investigation of durability characteristics of a pasted filamentary Ni catalyst was performed and extensive approaches for elucidating the degradation mechanism of the catalyst were undertaken. The H₂ generation rate of pasted filamentary Ni catalyst gradually decreases with cycling and 76% of the initial H₂ generation rate remains at 200cycles. After 200cycles, SEM images show an agglomeration of catalyst particles and a film on the pasted catalyst. An agglomeration of catalyst results in a reduction of specific surface area, as indicated by BET analysis, and a film on the pasted catalyst was formed by an accumulation of potassium originating from alkaline solution, and sodium and boron, which consists of Na₂B₄O ₇·10H₂O, a potassium borate compound (KB _xO_y) and boron oxide (B₂O₃), as indicated by XRD and XPS analyses. Consequently, a deterioration of the catalyst upon hydrolysis of sodium borohydride results from agglomeration of the catalyst and the film formed on the catalyst.

Full text of DOI:10.1016/j.jallcom.2004.02.009

Journal of Alloys and Compounds 379 (2004) 222–227
Study on degradation of filamentary Ni catalyst on
hydrolysis of sodium borohydride
a,∗
a
a
a
Jin-Ho Kim , Ki-Tae Kim , Yong-Mook Kang , Hyun-Seok Kim ,
a
a
b
a
Min-Sang Song , Yong-Ju Lee , Paul S. Lee , Jai-Young Lee
a
Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology,
3
73-1 Gusong-Dong, Yusong-Gu, Daejeon, South Korea
b
KLA Tencor, San Jose, CA, USA
Received 22 December 2003; received in revised form 3 February 2004; accepted 3 February 2004
Abstract
An investigation of durability characteristics of a pasted filamentary Ni catalyst was performed and extensive approaches for elucidating
the degradation mechanism of the catalyst were undertaken. The H generation rate of pasted filamentary Ni catalyst gradually decreases with
cycling and 76% of the initial H generation rate remains at 200 cycles. After 200 cycles, SEM images show an agglomeration of catalyst
2
2
particles and a film on the pasted catalyst. An agglomeration of catalyst results in a reduction of specific surface area, as indicated by BET
analysis, and a film on the pasted catalyst was formed by an accumulation of potassium originating from alkaline solution, and sodium and
boron, which consists of Na
2
B
4
O
7
·10H
2
O, a potassium borate compound (KB
x
y
2 3
O ) and boron oxide (B O ), as indicated by XRD and XPS
analyses. Consequently, a deterioration of the catalyst upon hydrolysis of sodium borohydride results from agglomeration of the catalyst and
the film formed on the catalyst.
©
2004 Elsevier B.V. All rights reserved.
Keywords: Sodium borohydride; Hydrogen; Filamentary Ni; Durability; Degradation
1
. Introduction
pH >11 produces hydrogen gas and water-soluble sodium
borate, NaBO2, in the presence of a suitable catalyst.
Hydrogen is forecasted to become a major source of en-
NaBH4 + 2H2O → 4H2 + NaBO2 + Q (217 kJ)
(1)
ergy in the future. Recently, the use of hydrogen is applied to
various industries, while it is anticipated that it will be used
extensively if present research and development are contin-
ued. Hydrogen is produced from various energy sources, and
it is stored, transported and used in industries, homes, au-
tomobiles, aeroplanes, and factories, and used in electricity
generation. However, for man-portable applications a major
hurdle is the development of a means to generate and store
controlled amounts of pure H2 directly without restoring to
high temperature reformers with significant heat burner. Re-
cently, numerous studies of hydrogen production through a
hydrolysis of sodium borohydride (NaBH4) have been per-
formed by various research groups [2–6]. Schlesinger et al.
Suda et al. [4] and Kojima et al. [6] examined that the ef-
fect of fluorinated metal hydrides and Pt metal, respectively,
coated on a metal oxide catalyst on the hydrolysis of NaBH4
in alkaline solution. Amendola et al. [2,3] reported that high
H2 generation rates had been achieved from NaBH4 solu-
tion with Ru supported on anion exchange resins. Levy et al.
[7] and Kaufman and Sen [8] investigated cobalt and nickel
borides as catalysts for practical, controlled generation of H2
from NaBH4. However, while various catalysts with high H2
generation rates have been developed through many stud-
ies of hydrolysis of sodium borohydride, researches on the
durability of catalysts have been scarce.
In the previous work [12], we developed a filamentary
Ni mixed Co catalyst with superior performance of short
initial waiting time and fast hydrolysis of sodium boro-
hydride. In the present study, the durability characteristics
and the approaches for elucidating the degradation mech-
[
1] reported that NaBH4 in aqueous alkaline solutions of
∗
Corresponding author.
E-mail address: fighter@cais.kaist.ac.kr (J.-H. Kim).
0
925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.02.009

J.-H. Kim et al. / Journal of Alloys and Compounds 379 (2004) 222–227
223
anism of a catalyst are extensively studied and reported
here.
study. In this work, all measurements for hydrogen genera-
tion rate were conducted at 30 C by a thermostat.
◦
The structure of the catalyst was studied using an X-ray
diffractometer (XRD). Powder X-ray diffraction patterns
for these catalysts were obtained on a Rigaku, D/max-IIIC
2
. Experimental
(3 kW) X-ray diffractometer using Cu K␣-source operating
At the outset, KOH alkaline solution was used as sup-
at 40 kW and 45 mA. All XRD patterns were recorded at a
◦
−1
porting solution for safe storage of NaBH4 (purchased from
Aldrich Inc.). The self-hydrolysis rate of NaBH4 solutions
depends on the pH and solution temperature. Kreevoy and
Jacobson [9] reported that this rate is empirically represented
by:
scan rate of 0.5 2θ min , and analyzed by using a MIDI
JADE5 program. In order to examine the morphology of the
pasted catalyst, scanning electron microscopy (SEM) was
employed and the hydrogen gas acquired from hydrolysis of
NaBH4 was quantitatively and qualitatively analyzed by gas
mass spectroscopy. In order to closely inspect the specific
surface area of the pasted catalyst, a volumetric adsorption
apparatus (ASAP 2010M) was used through the liquid nitro-
gen adsorption at 77.35 K by the Brunauer–Emmett–Teller
log t1/2 = pH − (0.034 T − 1.92)
(2)
(
t1/2 is the time required for one-half of the NaBH4 solution
to decompose (min) and is dependent on the solution pH and
temperature (K)). Hence, the presence of selected catalysts
increases the production rate of hydrogen, even in an alkaline
solution at normal temperature.
(BET) method. The XPS analysis technique, which was con-
ducted using Mg K␣ radiation and an Escalab MKII photo-
electron energy spectrum instrument was used to determine
the chemical state and content of the constituent elements
of the catalyst.
Filamentary Ni (INCO Inc., T-210H, 99.9%), which dis-
plays good performance with respect to the hydrolysis of
sodium borohydride because of its large specific surface
area, was used as a catalyst for hydrolysis of sodium boro-
hydride [12]. The catalyst was prepared by a paste method
as explained in our earlier work [10]. The 2 wt.% hydrox-
ypropyl methyl cellulose (HPMC) aqueous solution was
obtained by dissolving it in distilled water. This solution
has the function of controlling the viscosity of the slurry.
Styrene–butadiene rubber (SBR) emulsion, which has toler-
ance to alkaline solution and hydrophilic property, used as
a binding material was stirred vigorously in a vessel with a
propeller-type stirrer to prepare the catalyst slurry. Using a
doctor-blade technique, the slurry was applied over Ni foam
to make a catalyst sheet followed by drying and rolling to
a very thin pasted catalyst. Fig. 1 shows the schematic di-
agram of preparation procedure of pasted catalyst in this
3. Results and discussion
Ni is a promising catalyst with good performance and low
cost for the hydrolysis of sodium borohydride [11]. Among
various types of Ni powders, filamentary Ni and Raney Ni
have good performance in terms of short initial waiting time
and efficient hydrolysis of sodium borohydride due to their
high specific surface area [12]. In particular, filamentary Ni
can be handled very easily and is inexpensive as well as dis-
playing good characteristics for the hydrolysis of NaBH4. In
consequence to this, filamentary Ni has been used as a cat-
Catalyst powder HPMC solution
75 wt. % )
SBR
(25 wt. % )
100
(
8
0
0
0
0
6
4
2
Mixing of paste
Loading of paste (on Ni-foam)
Rolling (t=0.25 mm)
o
Drying (at 40 C)
0
0
50
100
150
200
Cutting (3X3 cm)
Number of cycles
Wrapping (with stainless mesh)
Fig. 2. Cyclic behavior on H2 generation rate for filamentary Ni catalyst
with SBR (20 wt.% of catalyst) over Ni foam immersed in 14 wt.% NaBH4
alkaline solution (pH = 13) at room temperature.
Fig. 1. Schematic of preparation procedure of pasted catalyst.

2
24
J.-H. Kim et al. / Journal of Alloys and Compounds 379 (2004) 222–227
alyst so as to evaluate the cyclic property of the catalyst on
the hydrolysis of NaBH4 in an alkaline solution. Fig. 2 plots
the cyclic behavior on H2 generation rate for the pasted fila-
mentary Ni catalyst with SBR (20 wt.% of catalyst loading)
over Ni foam in 14 wt.% NaBH4 alkaline solution (pH = 13)
at room temperature. It is observed that the H2 generation
rate gradually decreases with cycling and only 76% of the
initial H2 generation rate remains at 200 cycles. The degrada-
tion mechanism of the catalyst on the hydrolysis of NaBH4
in an alkaline solution is still not clear, and hence attempts
have been made to elucidate the degradation mechanism for
the pasted filamentary Ni catalyst through SEM, XPS, XRD
and BET analyses.
has not so far been reported, and hence an attempt has been
made to test the influence of the film on the degradation of
the catalyst.
Fig. 4 plots the cyclic behavior on H2 generation rate for
the pasted filamentary Ni catalyst with and without rinsing
with distilled water every cycle. This figure indicates that
the pasted filamentary Ni catalyst with rinsing shows bet-
ter durability than that without rinsing. R/Rmax value for
the pasted catalyst with rinsing and without rinsing was ob-
tained as 93 and 88%, respectively, at the end of 80 cycles.
SEM photograph in Fig. 5 shows that even at 80 cycles the
film formed on the catalyst as mentioned above was not ob-
served on it due to rinsing. There is a clear improvement
on the surface structure of the catalyst in the absence of the
film, as compared with the results depicted in Fig. 3. These
results imply that the deterioration of the catalyst could be
attributed to the film formed on the catalyst. In order to fur-
ther elucidate the degradation mechanism of the catalyst, the
composition of the film was examined.
3
.1. Formation of film on catalyst
Fig. 3 shows SEM images (10,000×) of the pasted fila-
mentary Ni catalyst before, 100 and 200 cycles, respectively.
Peculiarly, with cycling, an unknown film that was not ob-
served prior to cycling was formed and gradually spread over
the whole surface of the pasted catalyst. The film formation
From EDX analysis, Na and K compositions, which were
not detected before cycling, were found to exist together
Fig. 3. SEM images (10,000×) of the pasted filamentary Ni catalyst: (a) before, (b) after 100, and (c) after 200 cycles.

J.-H. Kim et al. / Journal of Alloys and Compounds 379 (2004) 222–227
225
Ni
100
New small peak
After 150 cycles
80
No rinsing
Rinsing with distilled water
Before cycling
60
0
20
40
60
80
Number of cycles
20
30
40
50
60
70
80
Fig. 4. Cyclic behavior on H2 generation rate for the pasted filamentary
Ni catalyst with and without rinsing with distilled water every cycle.
2 - theta
Fig. 6. XRD patterns of filamentary Ni catalyst electrodes before cycling
and after 150 cycles.
with Ni on the surface of the catalyst after 200 cycles.
Amendola et al. [2] estimated that the precipitation of
NaBO2, a by-product on catalyst, would deteriorate a hy-
drolysis of sodium borohydride, and Kojima et al. [6], using
X-ray diffraction, observed NaBO2·2H2O as a by-product
on the hydrolysis of NaBH4 by Pt-TiO2 catalyst. The film
on the catalyst is thus likely to be a compound consisting of
Na, K, and other elements. In order to clearly elucidate the
unknown film, we have conducted the X-ray diffraction and
XPS analysis. From the XRD profile as shown in Fig. 6, Ni
is only detected before cycling, while new and small peaks
corresponding to NaBxOy·nH2O and KBxOy are observed
after 150 cycles. However, accurate determination of any
new phase by XRD analysis is very difficult because XRD
peaks of many compounds in hydrated sodium borate or
potassium borate group are nearly overlapped with each
other. Moreover, XRD analysis is not an effective tool to
examine the surface properties. Conversely, XPS analysis
is a powerful technique for determining surface states, and
hence we have performed XPS analysis to study the chem-
ical changes in the surface of the pasted filamentary Ni cat-
alysts before and after cycling. Fig. 7(a) and (b) display the
boron, B, core level spectra of the catalyst before cycling and
after 200 cycles, respectively. The peak of elemental boron
is assigned to 189.4 eV; this was not observed in Fig. 7(a).
The peak around 193 eV indicated in Fig. 7(b) is ascribed
to Na2B4O7·10H2O or B2O3, which exhibits a chemical
shift to high binding energy due to oxidation. Therefore, it
is likely that with cycling the hydrated sodium metaborate
(Na2B4O7·10H2O) stacks upon the pasted catalyst and the
Fig. 5. SEM images (10,000×) of the pasted filamentary Ni catalyst with rinsing after 80 cycles.

2
26
J.-H. Kim et al. / Journal of Alloys and Compounds 379 (2004) 222–227
(b)
After 200 cycles
After 200 cycles
(b)
(a)
Before cycling
Before cycling
(a)
1
96
194
192
190
188
186
184
298
296
294
292
290
Binding Energy / eV
Binding Energy / eV
Fig. 7. The boron, B, core level spectra of the catalyst: (a) before cycling
and (b) after 200 cycles.
Fig. 9. The potassium, K, core level spectra of the catalyst: (a) before
cycling and (b) after 200 cycles.
surface of the catalyst oxidizes, resulting finally in degrada-
tion of the catalyst. The sodium, Na, spectra of the pasted
filamentary Ni catalyst before cycling and after 200 cycles
are displayed in Fig. 8(a) and (b), respectively. The peak
of sodium element is assigned to 1072.1 eV; this peak also
cannot be observed in Fig. 8(a). Unfortunately, the new
peak around 1074 eV in Fig. 8(b) could not be identified
what the phase is, however, this strong peak at high bind-
ing energy is interpreted as the oxidation of the surface of
the pasted filamentary Ni catalyst. Fig. 9(a) and (b) show
the potassium, K, spectra of the pasted filamentary Ni cat-
alyst before cycling and after 200 cycles. The peak around
EDX, XRD, and XPS analyses, it is concluded that with cy-
cling, potassium, K, originated in the alkaline solution, and
sodium, Na, and boron, B, derived from NaBH contact to
4
the surface of catalyst. Therefore, the film which consists of
Na B O ·10H O, unidentified potassium borate (KB O )
2
4
7
2
x
y
and boron oxide (B O ), was formed, resulting in the degra-
2
3
dation of the pasted filamentary Ni catalyst.
3.2. Decrease in specific surface area of catalyst
Fig. 10 shows the low ((a) and (b)) and high ((c) and
(d)) magnification of SEM images for the pasted filamen-
2
94 eV in Fig. 9(b), shifted to a high binding energy in com-
tary Ni catalyst before cycling and after 200 cycles. As seen
in Fig. 10(a) and (b), an agglomeration of the catalyst is
observed with cycling and it is anticipated that this would
diminish the specific surface area of the catalyst. Fig. 10(c)
and (d), with a higher magnification, show more precise
observations of the agglomeration of the catalyst including
coarsening of particles with cycling. Also, the BET anal-
ysis of the pasted filamentary Ni catalyst before and after
200 cycles, as presented in Table 1, indicates that the specific
surface area is decreased by 36.6% after 200 cycles, indicat-
ing a reduction of reactive sites on hydrolysis. Therefore, it
could be concluded that the second degradation mechanism
of the catalyst on the hydrolysis with cycling can be as-
cribed to the reduction of reactive sites, which results from a
decrease of specific surface area induced by the agglomera-
tion of catalysts.
parison to elemental potassium, i.e., 292.9 eV, indicates of
any kind of potassium borate compound (KBxOy), and also
implies the oxidation of the surface of the catalyst. From
(
b)
a)
After 200 cycles
Before cycling
(
Table 1
BET analysis on specific surface area of filamentary Ni catalyst electrode
before cycling and after 200 cycles
1080 1078 1076 1074 1072 1070 1068 1066
2
BET surface area (m /g)
Binding Energy / eV
Before cycling
After 200 cycles
5.06
3.21
Fig. 8. The sodium, Na, core level spectra of the catalyst: (a) before
cycling and (b) after 200 cycles.

J.-H. Kim et al. / Journal of Alloys and Compounds 379 (2004) 222–227
227
Fig. 10. Low (a) and (b) and high (c) and (d) magnification of SEM images for the pasted filamentary Ni catalyst before cycling and after 200 cycles.
4
. Conclusions
References
H2 generation rate of filamentary Ni catalyst electrode
[1] H.I. Schlesinger, H.C. Brown, A.B. Finholt, J.R. Gilbreath, H.R.
Hockstra, E.K. Hydo, J. Am. Chem. Soc. 75 (1953) 215.
gently decreases with cycling and 76% of the initial H2 gen-
eration rate remains at 200 cycles. From SEM and BET anal-
yses, the agglomeration of the catalyst was observed and
brought about the diminishing of specific surface area of the
catalyst electrode with cycling. Therefore, the deterioration
of the catalyst with cycling is ascribed to the reduction of re-
active sites by a decrease of specific surface area. Also, with
cycling, a film not observed prior to cycling, was formed and
gradually spreads over the whole surface of the pasted cat-
alyst, eventually covering the reactive sites. From XRD and
XPS analyses, it was observed that with cycling, potassium
originated from alkaline solution, and sodium and boron ac-
cumulated on the surface of pasted catalyst, and finally a film
consisting of Na2B4O7·10H2O, potassium borate (KBxOy)
and boron oxide (B2O3) was formed. Consequently, a de-
terioration of the catalyst results from an agglomeration of
the catalyst particles and the film formed on the catalyst.
[
2] S.C. Amendola, P. Onnerud, M.T. Kelly, P.J. Petillo, S.L.
Sharp-Goldman, M. Binder, J. Power Source 85 (2000) 186.
3] S.C. Amendola, J.M. Janjua, N.C. Spencer, M.T. Kelly, P.J. Petillo,
S.L. Sharp-Goldman, M. Binder, Int. J. Hydrogen Energy 25 (2000)
969.
[
[4] S. Suda, Y.M. Sun, B.H. Liu, Y. Zhou, S. Morimitsu, K. Aral, N.
Tskamoto, M. Uchida, Y. Caudra, Z.P. Li, Appl. Phys. A 72 (2001)
2
09.
5] I.I. Korovov, N.G. Mozgina, L.N. Blinova, Kinet. Catal. 36 (3) (1995)
80.
[
3
[6] Y. Kojima, K.I. Suzuki, K. Fukumoto, M. Sakai, T. Yamamoto, Y.
Kawai, H. Hayashi, Int. J. Hydrogen Energy 27 (2002) 1029.
[7] A. Levy, J.B. Brown, C.J. Lyons, Ind. Eng. Chem. 52 (1960) 211.
[8] C.M. Kaufman, B. Sen, J. Chem. Soc. Dalton Trans. (1985) 307.
[9] Kreevoy, M.M. Jacobson, Ventron Alembic, 1979, pp. 2–3.
[
10] D.M. Kim, H. Lee, K.J. Jang, J.Y. Lee, J. Electrochem. Soc. 145 (10)
(1998) 3387.
[
[
11] H.C. Brown, C.A. Brown, J. Am. Chem. Soc. 84 (1962) 1493.
12] J.H. Kim, H. Lee, S.C. Han, H.S. Kim, M.S. Song, J.Y. Lee, Int. J.
Hydrogen Energy 29 (2004) 263–267.