Towards a structure-performance relationship for hydrogen storage in Ti-doped NaAlH4 nanoparticles

Article abstract of DOI:10.1039/c0cc02787a

Hydrogen storage properties of Ti-doped nanosized (～20 nm) NaAlH ₄ supported on carbon nanofibers were affected by the stage at which Ti was introduced. When Ti was deposited first followed by NaAlH_4, sorption properties were superior to the case where NaAlH₄ was deposited first followed by NaAlH₄. This was the result of both a smaller NaAlH₄ particle size and the more extensive catalytic action of Ti in the former material.

Full text of DOI:10.1039/c0cc02787a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 2143–2145
www.rsc.org/chemcomm
COMMUNICATION
Towards a structure-performance relationship for hydrogen storage in
Ti-doped NaAlH nanoparticlesw
4
,a Olivier Leynaud,bc Paul Barnes, Elena Pela
bc
ez-Jime
´
nez,a
Cornelis P. Balde
´
Krijn P. de Jong and Johannes H. Bitter*
´
a
a
Received 24th July 2010, Accepted 14th October 2010
DOI: 10.1039/c0cc02787a
Hydrogen storage properties of Ti-doped nanosized (B20 nm)
slow drying procedure was applied. This sample will be referred to
as ‘‘Ti–NaAlH /CNF’’. An undoped nano-NaAlH with a similar
NaAlH loading of 8 wt% was prepared using the same proce-
dure. This sample will be referred to as ‘‘NaAlH /CNF’’ and had
alanate particle sizes from 20–30 nm. For comparison, a bulk
sample was prepared by ball milling TiCl with purified NaAlH
(Ti : Al = 4.5 mol%) using a SPECS ball milling apparatus as
NaAlH
4
supported on carbon nanofibers were affected by the
4
4
stage at which Ti was introduced. When Ti was deposited first
followed by NaAlH4, sorption properties were superior to the
4
4
3
case where NaAlH
4
was deposited first followed by NaAlH
4
.
This was the result of both a smaller NaAlH
4
particle size and
3
4
the more extensive catalytic action of Ti in the former material.
26
described by Haiduc et al. This sample will be referred to as
‘‘ball milled TiCl –NaAlH ’’. Further details can be found in the
Sodium alanate has become a show-case for research on light-
1
weight complex metal hydrides for hydrogen storage. Neverthe-
3
4
ESI.w Bulk sodium alanate only starts to desorb hydrogen around
its melting point, i.e. 180 1C. Fig. 1 displays the hydrogen
desorption profiles for the (doped) nano-alanate samples.
less, hydrogen release rates and absorption rates are often impeded
2
in metal hydrides and need to be improved. Most effective in
improving the performance of metal hydrides are reducing the
3–13
particle size to the nm range (typically below 30 nm)
4
NaAlH /CNF shows two desorption rate maxima one around
or adding
90 1C and one around 160 1C. These two maxima have been
attributed before to the broad range of particle sizes present in
14–25
one or more promoters/catalysts to bulk materials.
3
The aim of the current research is to investigate the possible
the sample. Small particles (not detected in XRD; o10 nm)
desorb at the lower temperature while 20–30 nm particles desorb
advantage of adding a Ti catalyst to NaAlH particles of B20 nm
4
3
supported on carbon nanofibers (CNF). The order of Ti and
alanate deposition, to achieve an optimal alanate–Ti interaction is
discussed. The hydrogen storage characteristics of these materials is
hydrogen at 150 1C. When Ti was added only one maximum for
desorption was observed in this temperature range i.e. at 132 1C
for Ti–NaAlH /CNF and at 99 1C for NaAlH –Ti/CNF though
4 4
compared to undoped nano-NaAlH
TiCl –NaAlH (alanate crystallite size 150–200 nm). Structure
4
on CNF and ball milled
the peaks are broad possibly resulting from some heterogeneity
within the samples. Clearly the addition of Ti affects the
desorption temperature. It has to be noted here that the starting
3
4
activity relationships will be established by relating information
from Ti K-edge Extended X-ray Absorption Fine Structure
material for Ti–NaAlH
comparing the desorption characteristics of those two samples
the low temperature shoulder present for NaAlH was not
observed when ‘‘Ti’’ is present which might indicate that these
4 4
/CNF was NaAlH /CNF. When
(EXAFS) and in situ X-ray Diffraction (XRD) results with the
hydrogen absorption and desorption characteristics of the samples.
The ‘‘Ti first’’ sample was prepared by first a pore volume
impregnation with 10 mol% Ti(OBu) (compared to NaAlH ) in
4
smallest particles have reacted with the Ti(OBu)
4
during doping.
4
4
diethyl ether. After removal of the diethyl ether, 8 wt% NaAlH
was impregnated using THF followed by slow drying. The sample
will be referred to as ‘‘NaAlH –Ti/CNF’’. All sample handling
4
4
was performed in an inert atmosphere either using Schlenk
techniques or using a glove box. The ‘‘alanate first’’ sample of
identical composition was prepared by reversing the impregnation
4 4
order of NaAlH and Ti(OBu) ; however, after both steps the
a
b
c
Inorganic Chemistry and Catalysis, Utrecht Univeristy,
Sorbonnelaan 16, 3096 TB Utrecht, The Netherlands
School of Crystallography, Birkbeck College (University of London),
Malet Street, London, WC1E 7HX, UK
Department of Chemistry, Univ. College London, 20 Gordon Street,
London, WC1H 0AJ, UK
À1
Fig. 1 Desorption profiles for various. (2 1C min , normalized
NaAlH
4
loading (> 90% of the theoretical amount of hydrogen
w Electronic supplementary information (ESI) available: Additional details:
Figures S1 and S4, and Tables S1 and S2. See DOI: 10.1039/c0cc02787a
was desorbed for each sample).
This journal is c The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 2143–2145 2143

View Article Online
Please note that also for reloading, in a Rubotherm high
2
Literature shows that the reduction of particle size of metal
7,28
pressure magnetic suspension balance,
the addition of Ti
hydrides can significantly increase H
3–13
temperatures.
2
desorption rates at lower
to nanosized NaAlH was beneficial (ESI, Fig. S1w). However,
Thus, the particle size reduction of the NaAlH4,
4
the storage capacity of the samples decreased with the number
of absorption/desorption cycles (Table S1, ESIw) indicating
irreversible deactivation.
induced by the presence of Ti, in NaAlH –Ti/CNF explained, at
4
least partially, the lowest Tmax in the TPD profile of Fig. 1. How-
ever, this might not be the sole explanation of the observed kinetic
improvements, as the Ti species can be active as a catalyst as well.
Whether this is the case was investigated by an in situ XRD
study. The evolution of crystalline alanate phases during hydrogen
desorption were evaluated in the temperature range from room
From Fig. 1 it is concluded that the order of Ti addition to
nanosized NaAlH
4
had a significant influence on the hydrogen
–Ti/CNF
desorption properties of the materials of which NaAlH
4
showed the highest desorption rate maximum at the lowest
temperature. To investigate the possible reason for that, the
samples were characterized by XRD (synchrotron radiation,
SRS Daresbury station 6.2) and EXAFS. XRD diffractograms
of the prepared samples are shown and discussed in the ESI
temperature to 200 1C for NaAlH
The integrated areas of the broadened NaAlH
Na AlH (2y = 31.641), and Al (2y = 34.791) diffraction lines are
4
/CNF and NaAlH
4
–Ti/CNF.
4
(2y = 26.851),
3
6
shown in Fig. 2 as a function of temperature. We realize that
amorphous phases or nanocrystalline phases cannot be observed
by XRD. It is, however, tentatively assumed that the conclusions
drawn from XRD are also relevant for the non crystalline
fractions of the samples in view of the broad but single-peak
(
Fig. S2).w We first focus here on the comparison between
Ti-NaAlH /CNF and NaAlH /CNF. The average NaAlH
4
4
4
crystallite sizes of the samples were about 20 nm. The peak
intensities of the NaAlH diffraction lines for Ti-NaAlH /CNF
and NaAlH /CNF were also comparable, thus the content of
B20 nm NaAlH particles was similar. Therefore, it is concluded
that the particle size distribution, for the XRD detectible part of
the samples, did not differ significantly in Ti–NaAlH /CNF and
NaAlH /CNF after synthesis. This implies that the decrease in
temperature at which the maximum desorption rate occurred
4
4
4
desorptions of H
2
in the TPD experiments (see Fig. 1). For
line intensity decreased with
4
NaAlH –Ti/CNF, the NaAlH
4
4
increasing temperature starting at 20 1C and became zero at
115 1C. The Al line grew steadily from 40 1C and stabilized when
the temperature reached 130 1C. Weak Na AlH diffraction lines
3 6
4
4
were detected at temperatures between 105 and 150 1C.
For NaAlH /CNF, the NaAlH diffraction started to
(Tmax) for Ti–NaAlH /CNF, is the result of the catalytic action
4
4
4
of the Ti-species present. In order to elucidate the structure of the
decrease from 20 1C and became zero at 150 1C. In the same
low temperature range, the Al diffraction intensity was
constant. However above 150 1C, the Al diffraction rapidly
Ti-nanoparticles in Ti–NaAlH /CNF, their local structure was
4
investigated with EXAFS prior to hydrogen desorption. EXAFS
fit parameters (Table S2, ESIw) reveal that Ti is surrounded by O,
C and Ti atoms. The absence of Na or Al atoms in the local
increased and stabilized at 190 1C. Weak Na
were only observed at high temperature (150 to 180 1C) i.e.,
after all NaAlH had decomposed.
Since the FWHM of the NaAlH
NaAlH /CNF and NaAlH –Ti/CNF, i.e., the crystalline part of
the NaAlH had a similar crystallite size (Fig. S2, ESIw), thus
the role of the ‘‘Ti’’ on the NaAlH decomposition could
3 6
AlH diffractions
structure indicates that Ti did not interact with the NaAlH
4
to a
4
level detectable by EXAFS (10 at%). Since the local structure
comprised the same atoms as the original Ti-precursor
4
diffraction was similar in
4
4
(
Ti(OBu)
4
), it was concluded that the majority of ‘‘Ti’’ was
, or a decomposition
4
present as nanoparticles of Ti(OBu)
4
4
product thereof, in the as-prepared Ti–NaAlH /CNF.
4
therefore be independently investigated by monitoring the
evolution of the diffraction patterns. It was observed that the
Apparently, after synthesis the Ti(OBu)4 had not fully
reacted with the alanate in Ti–NaAlH /CNF. This incomplete
4
NaAlH diffraction peaks reached zero intensity at a consider-
4
reaction of the catalyst-precursor with the NaAlH has been
4
ably lower temperature when ‘‘Ti’’ was present (NaAlH –Ti/
4
reported previously for NaAlH
4
ball-milled with TiF
3
30
CNF) than when ‘‘Ti’’ was absent (NaAlH
‘‘Ti’’ accelerated decomposition of the NaAlH
NaAlH –Ti/CNF, thus acting as a catalyst.
From Fig. 2 it also becomes clear that crystalline Na
did not show up as long as the nano-sized NaAlH was present.
In contrast, for non supported Ti(OBu) catalyzed NaAlH , an
in situ diffraction study by Gross et al. indicated that Na AlH
4
/CNF). Apparently,
2
9
(
micro-meter sized) and Ti(OBu)
For Ti(OBu)
from an in situ EXAFS study that the Ti species were reduced
to the catalytically active TiAl phase during hydrogen extrac-
tion in the first desorption step. In accordance, we assume that
the nanoparticles of Ti(OBu) , or a decomposition product
4
ball-milled with NaAlH
4
.
4
particles in
4
ball milled with NaAlH
4
, it has been concluded
4
3
0
AlH
3 6
x
4
4
4
4
3
6
thereof, in the as-prepared Ti-NaAlH /CNF will react with
4
the TiAl catalyst in Ti–NaAlH /CNF during the first desorption
4
x
step. Therefore, although the Ti-alanate interaction was limited,
still catalytic effects on desorption are apparent (Fig. 1).
Hereafter, we focus on the properties of the sample that
4
showed the lowest Tmax in the TPD profile (NaAlH –Ti/CNF
in Fig. 1). The crystallite size of the alanate was also in this
sample B20 nm. However, the intensity of the alanate diffrac-
tion was considerably reduced for this sample indicating that a
larger fraction of NaAlH
4
is amorphous or nanocrystalline
(
Fig. S2, ESIw). Thus the average size and/or crystallinity of
the NaAlH4 has been decreased in NaAlH –Ti/CNF com-
Fig.
2
4 3 6
Integrated areas of NaAlH (2y = 26.851), Na AlH
4
(2y = 31.641) and Al ((2y = 34.791) diffractions during the in situ
desorption (5 1C min , normalized to CNF peak).
À1
pared to NaAlH /CNF during the preparation.
4
2
144 Chem. Commun., 2011, 47, 2143–2145
This journal is c The Royal Society of Chemistry 2011

View Article Online
3
1
crystallized in co-presence of NaAlH
presence of Na AlH and NaAlH during decomposition was
also reported in an in situ XRD study of TiCl -catalyzed bulk
4
.
The simultaneous
Notes and references
3
6
4
1
2 S. Orimo, Y. Nakamori, J. R. Eliseo, A. Zuttel and C. M. Jensen,
¨
L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353.
¨
3
3
2
NaAlH4. Thus, the decomposition mechanism of nano-
Chem. Rev., 2007, 107(10), 4111.
C. P. Balde, B. P. C. Hereijgers, J. H. Bitter and K. P. de Jong,
J. Am. Chem. Soc., 2008, 130, 6761.
3
4
NaAlH4 differs noticeably from that of bulk-NaAlH . We
4
hypothesize that nano-NaAlH4 decomposed directly to H2,
NaH and Al, thus largely bypassing the formation of crystalline
A. Gutowska, L. Li, Y. Shin, C. M. Wang, X. S. Li, J. C. Linehan,
R. S. Smith, B. D. Kay, B. Schmid, W. Shaw, M. Gutowski and
T. Autrey, Angew. Chem., Int. Ed., 2005, 44, 3578.
A. Feaver, S. Sepehri, P. Shamberger, A. Stowe, T. Autrey and
G. Cao, J. Phys. Chem. B, 2007, 111(26), 7469–7472.
P. E. de Jongh, R. W. P. Wagemans, T. M. Eggenhuisen,
B. S. Dauviller, P. B. Radstake, J. D. Meeldijk, J. W. Geus and
K. P. de Jong, Chem. Mater., 2007, 19(24), 6052.
3 6
Na AlH . Thermodynamically this is feasible (Fig. S3, ESIw),
also for bulk materials. The thermodynamic properties of
5
6
Ti-doped nano alanate particles could be different than those for
9,14,33
bulk materials as shown for undoped nanosized systems.
The structural properties of the ‘‘Ti’’ species in NaAlH
CNF as inferred from EXAFS indicated that Ti(OBu) reacts with
4
–Ti/
7
8
J. J. Vajo and G. L. Olson, Scr. Mater., 2007, 56(10), 829.
B. Bogdanovic and M. Schwickardi, J. Alloys Compd., 1997,
253–254, 1.
4
NaAlH during the synthesis (see ESIw). Since our samples have a
4
total Ti–Al coordination number of 4.5 and a Ti–Ti coordination
9 J. Gao, P. Adelhelm, M. H. W. Verkuijlen, C. Rongeat, M. Herrich,
P. J. M. van Bentum, O. Gutfleisch, A. P. M. Kentgens, K. P. de
Jong and P. E. de Jongh, J. Phys. Chem. C, 2010, 114.
number of 0.7 the total Ti : Al ratio in our particles is B1 : 6.4.
Thus the TiAl species have an average stoichiometry of TiAl6.4.
x
1
0 P. Adelhelm, J. Gao, M. H. W. Verkuijlen, C. Rongeat,
M. Herrich, P. J. M. M. van Bentum, O. Gutfleisch, A. P. M.
Kentgens, K. P. de Jong and P. E. de Jongh, Chem. Mater., 2010,
In the previous part, the structure of the Ti catalyst and its
ability to decrease the particle size of the NaAlH
CNF has been discussed. Now, we will elaborate on an explana-
tion of how the ‘‘Ti’’ addition dispersed the NaAlH during its
preparation. The first step in the preparation was the impregna-
tion of liquid Ti(OBu) dissolved diethylether to the CNF.
Subsequent drying deposited the liquid Ti(OBu) on the fibers.
4 4
in NaAlH –Ti/
2
2, 2233.
1
1 M. H. W. Verkuijlen, J. Gao, P. Adelhelm, P. J. M. van Bentum,
P. E. de Jongh and A. P. M. Kentgens, J. Phys. Chem. C, 2010,
114, 4683.
4
1
2 R. K. Bhakta, J. L. Herberg, B. Jacobs, A. Highley, R. Behrens Jr,
N. W. Ockwig, J. A. Greathouse and M. D. Allendorf, J. Am.
Chem. Soc., 2009, 131, 13198.
4
4
In a second impregnation, NaAlH dissolved in THF was added.
4
13 H. Wu, ChemPhysChem, 2008, 9, 2157.
14 W. Lohstroh, A. Roth, H. Hahn and M. Fichtner,
Ti(OBu) is expected to dissolve in THF, thus could react with
4
ChemPhysChem, 2010, 11, 789; A. Zu
P. Sudan, P. Mauron and C. Emmenegger, J. Power Sources, 2003,
18, 1.
¨
ttel, P. Wenger, S. Rentsch,
NaAlH forming a homogenously-dispersed TiAl species. This
4
x
species is insoluble and precipitated on the CNF forming finely
dispersed anchoring sites on the CNF. When the impregnated
1
15 X. Yao, C. Wu, A. Du, J. Zou, Z. Zhu, P. Wang, H. Cheng,
S. Smith and F. Lu, J. Am. Chem. Soc., 2007, 129, 15650.
material was dried, the NaAlH
4
deposited on the finely dispersed
during
1
1
1
1
6 B. Bogdanovic, M. Felderhoff, A. Pommerin, F. Schu
N. Spielkamp, Adv. Mater., 2006, 18, 1198.
¨
th and
anchoring sites, which increased the dispersion of NaAlH
4
the preparation. Thus, the enhanced desorption rates at low
temperature can be explained by two phenomena (a) the forma-
tion of a well mixed Ti–Al catalyst and (b) the formation of a
fraction of smaller (o20 nm) alanate particles.
7 M. Felderhoff, C. Weidenthaler, R. von Helmolt and U. Eberle,
Phys. Chem. Chem. Phys., 2007, 9, 2643.
8 B. Sakintuna, F. Lamari-Darkrim and M. Hirscher, Int. J.
Hydrogen Energy, 2007, 32, 1121.
9 J. Wang, A. D. Ebner, T. Prozorov, R. Zidan and J. A. Ritter,
J. Alloys Compd., 2005, 395, 252.
20 S. Zheng, F. Fang, G. Zhou, G. Chen, M. Ouyang, M. Zhu and
D. Sun, Chem. Mater., 2008, 20, 3954.
1 A. Marashdeh, R. A. Olsen, O. M. Løvvik and G.-J. Kroes,
In conclusion, Ti(OBu) -modified NaAlH nanoparticles have
4
4
been synthesized on carbon nanofibers (CNF) by impregnation
and drying techniques. In the cases where the NaAlH4 was
2
impregnated first and Ti(OBu) second, the particle size of the
4
J. Phys. Chem. C, 2008, 112, 15759.
22 K. Bai, P. Shan, E. Yeo and Ping Wu, Chem. Mater., 2008, 20, 7539.
NaAlH
NaAlH
4
did not change compared to a non-doped nano-
. The temperature at maximum desorption rate
2
3 G. Krishna, P. Dathar and D. S. Mainardi, J. Phys. Chem. C, 2010,
14, 8026.
4 J. W. Kim, J.-H. Shim, S. C. Kim, A. Remhof, A. Borgschulte,
O. Friedrichs, R. Gremaud, F. Pendolino, A. Zuttel, Y. W. Cho
and K. H. Oh, J. Power Sources, 2009, 192, 582.
5 Y. Song, J. H. Dai, C. G. Li and R. Yang, J. Phys. Chem. C, 2009,
13, 10215.
4
1
decreased from above 160 1C to 132 1C and is ascribed to the
catalytic role of the Ti in that sample. Samples prepared by
2
¨
impregnating the Ti(OBu)
desorption maximum at 99 1C in Ar and absorbed hydrogen
from 10 bar H pressure at 115 1C after hydrogen extraction. The
4 4
first, and NaAlH second showed an
2
2
H
2
1
2
6 A. G. Haiduc, H. A. Stil, M. A. Schwarz, P. Paulus and
J. J. C. Geerlings, J. Alloys Compd., 2005, 393(1–2), 252–263.
outstanding hydrogen sorption properties were ascribed to two
roles of the ‘‘Ti’’. First, the NaAlH particle size in NaAlH –Ti/
4
27 F. Dreisbach, R. Seif and H. W. Lo
003, 71, 73.
8 H. W. Losch, R. Kleinrahm and W. Wagner, Neue Magnetschwe-
bewaagen fu¨r gravimertrische Messungen in der Verfahrentechnik,
VDI-Verslag, Dusseldorf, 1994.
¨
sch, J. Therm. Anal. Calorim.,
4
2
CNF was smaller (o20 nm) than in all other samples as a result
of Ti nuclei facilitating alanate dispersion. Secondly, a highly
dispersed ‘‘TiAl6.4’’ phase was formed by the reaction between
2
¨
¨
2
9 F. Fang, J. Zhang, J. Zhu, G. Chen, D. Sun, B. He, Z. Wei and
S. Wei, J. Phys. Chem. C, 2007, 111, 3476.
0 C. P. Balde, H. A. Stil, A. M. J. van der Eerden, K. P. de Jong and
´
J. H. Bitter, J. Phys. Chem. C, 2007, 111, 2797.
Ti-precursor and NaAlH
thus restricting growth of Al crystallites during the desorption of
NaAlH . The Ti-doped nanoalanate decomposed without the
formation of crystalline Na AlH most likely as result of the
4
which acted as an anchoring point
3
4
31 K. J. Gross, S. Guthrie, S. Takara and G. Thomas, J. Alloys
Compd., 2000, 297, 270.
3
6
modified thermodynamics of these small particles.
Financial support by NWO/ACTS systainable hydrogen,
Hasylab Hamburg (station E4) and SRS Daresbury (station 6.2)
are gratefully acknowledged.
3
2 K. J. Gross, G. Sandrock and G. J. Thomas, J. Alloys Compd.,
002, 330–332, 691.
33 R. D. Stephens, A. F. Gross, S. L. van Atta, J. J. Vajo and
2
F. E. Pinkerton, Nanotechnology, 2009, 20, 204018.
This journal is c The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 2143–2145 2145