Effect of milling and doping on decomposition of NH3BH 3 complex

Article abstract of DOI:10.1016/j.tca.2005.12.004

The thermal decomposition of borane-ammonia complex as well as of milled and doped samples was studied by volumetric titrations. The samples were heated at fixed temperature and the volume of the evolved gas recorded as a function of time. Milled and doped samples were prepared by mechanical and mechanochemical reactions, respectively. Samples containing 1 and 2 mol% hydrogen hexachloroplatinate hydrate were prepared. The materials were characterized by scanning electron microscopy (SEM) and X-ray diffractometry (XRD). Mechanical alloying was effective to modify the crystallinity of the complex and to change the material morphology enhancing the amount of gas evolved. Both the milled and the doped samples showed an increase of the pre-exponential factor in the Arrhenius equation. The activation energy decreased for the doped sample, and it increased for the milled sample. As a result it was found that the decomposition of 1 mol% doped sample could be provided by waste heat coming from polymer electrolyte membrane fuel cell.

Full text of DOI:10.1016/j.tca.2005.12.004

Thermochimica Acta 441 (2006) 184–190
Effect of milling and doping on decomposition of NH3BH3 complex
Sara De Benedettoa,b, Maria Carewskaa, Cinzia Centoa, Paola Gislona,∗
,
Mauro Pasqualib, Silvera Scacciaa, Pier Paolo Prosinia
a ENEA, C.R. Casaccia, S.P. 89, Via Anguillarese 301, Rome 00060, Italy
b Dipartimento ICMMPM, Facolta` di Ingegneria, Universita` di Roma “La Sapienza”, Rome, Italy
Received 12 September 2005; received in revised form 18 November 2005; accepted 4 December 2005
Abstract
The thermal decomposition of borane–ammonia complex as well as of milled and doped samples was studied by volumetric titrations. The
samples were heated at fixed temperature and the volume of the evolved gas recorded as a function of time. Milled and doped samples were
prepared by mechanical and mechanochemical reactions, respectively. Samples containing 1 and 2 mol% hydrogen hexachloroplatinate hydrate
were prepared. The materials were characterized by scanning electron microscopy (SEM) and X-ray diffractometry (XRD). Mechanical alloying
was effective to modify the crystallinity of the complex and to change the material morphology enhancing the amount of gas evolved. Both the
milled and the doped samples showed an increase of the pre-exponential factor in the Arrhenius equation. The activation energy decreased for the
doped sample, and it increased for the milled sample. As a result it was found that the decomposition of 1 mol% doped sample could be provided
by waste heat coming from polymer electrolyte membrane fuel cell.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Borane–ammonia; Ball milling; Hydrogen storage; Platinum; Activation energy
1. Introduction
hydrogen compound (about 20% in mass) prepared in 1956
by Shore and Parry [2]. The crystal structure of the complex
was studied by Lippert and Lipscomb [3]. The decomposition
of BA was first investigated by Hu et al. [4] and its applica-
tion as hydrogen source for fuel cell has already been proposed
[5]. Thermal decomposition of BA takes place in the tempera-
ture range 77–137 ◦C. It was found that the complex completely
decomposes at 1 1 2 ◦C releasing 1 mol of hydrogen per mole
of complex [5]. This temperature is higher than the working
temperature of polymer electrolyte membrane (PEM) fuel cell
that is around 90 ◦C. At higher temperatures water evaporation
strongly decreases PEM conductivity. For practical application
of BA complex as hydrogen source for PEM fuel cell the decom-
position temperature of the complex should be reduced.
The “hydrogen on demand” concept has been extensively
publicized [1]. The system is a source of hydrogen with widely
variable gas output flow rate and delivery pressure. Hydrogen
is stored in hydrogen-rich material such us lightweight metal
hydride (NaBH4) dissolved in alkaline water. The material is
storable in plastic containers at room temperature and without
pressurization. Fuel and by-product are environmentally benign,
non-flammable, and non-explosive. Hydrogen is released by
action of a platinum or other precious metals-based catalyst on
the fuel. The low cost and the simple hydrogen generation mech-
anism associated with high energy storage density, makes this
concept extremely attractive for mobile applications.
Another possibility to generate hydrogen on demand is the
thermal decomposition of materials with low decomposition
temperature. Borane–ammonia (BA) complex is a very rich
Mechanical alloying (MA), a solid-state powder processing
technique originally developed to produce oxide dispersion [6],
has now been shown capable of synthesizing a variety of equi-
librium and non-equilibrium alloy phases. The synthesized non-
equilibrium phases include supersaturated solid solutions [7,8],
metastable crystalline and quasicrystalline phases [9], nanos-
tructures [10], and amorphous alloys. MA of BA complex could
be effective to improve the reaction kinetic. Doping the mate-
rial with platinum could be also beneficial for hydrogen release.
∗
Corresponding author. Tel.: +39 0630 483 797; fax: +39 0630 486 357.
E-mail addresses: carewska@casaccia.enea.it (M. Carewska),
cinzia.cento@casaccia.enea.it (C. Cento), paola.gislon@casaccia.enea.it
(P. Gislon), Mauro.Pasquali@uniroma1.it (M. Pasquali),
silvera@casaccia.enea.it (S. Scaccia), prosini@casaccia.enea.it (P.P. Prosini).
0040-6031/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.tca.2005.12.004

S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
185
It was recognized that powder mixtures can be mechanically
activated to induce chemical reactions, i.e., mechanochemical
reaction (MR) at room temperature [11]. BA is a mild, effi-
cient reducing agent for aldehydes and ketones [12] and it could
reduce platinum ions to form platinum metal.
In this work we studied the effect of MA and MR with hydro-
gen hexachloroplatinate on the decomposition process of BA
complex.
2. Experimental
BA complex (BH3NH3, Aldrich, tech. 90%) and hydro-
gen hexachloroplatinate(IV) hydrate (H2PtCl6 + xH2O, Aldrich,
99.9%, Pt 38–40%) were used as-received. BA complex for MA
or a mixture of the complex with the acid for MR were weighed
to form 1.5 g batches which were loaded in a stainless steel vial
together with two stainless steel balls. The vial was evacuated
and filled with hydrogen. It was mounted in a shaker mill (SPEX
8000, CertPrep, Metuchen, NJ) and milled for 2 min followed
by a 10 min pause. The sequence was repeated eight times. The
total milling time was 16 min. Two doped samples with 1.0 and
2.0 mol% platinum were prepared. No macroscopic modifica-
tions were observed for the MA sample. The colour of the doped
samples changed from white to dark grey.
The materials were characterized by X-ray diffractometry
(XRD) analysis (Philips PW 3710 diffractometer) using Cu K␣
radiation. The platinum grain size (D) was calculated using
the Scherrer formula [13]: β cos(θ) = kλ/D, where β is the full-
width-at-half-maximum length of the diffraction peak on a 2θ
scale and k is a constant here close to unity. The value of D was
computed from the (1 1 1), diffraction peak (the best resolved in
the diffractograms).
Powders morphology was studied by scanning electron
microscopy (SEM, Jeol JSM-5510LV). Energy dispersive X-ray
spectroscopy (EDS) analysis was performed with an IXRF EDS-
2000 System. The conditions were: accelerating voltage 25 kV,
spot size 21 and working distance 21 mm. The powder speci-
mens were directly deposited onto a conductive carbon double
face tape, which was previously mounted on a SEM slab.
Thermal decomposition was monitored with the home-made
measuring apparatus illustrated in Fig. 1.
Fig. 1. The experimental apparatus for thermal decomposition measurements.
A thermocouple positioned close to the sample holder was used
to record the temperature. The oven was heated at fixed temper-
ature and, when the temperature was reached, the sample and
the thermocouple were introduced into the oven. Both the sig-
nals from thermocouple and conductimetric cell were digitally
recorded.
3. Results
3.1. Material characterization
BA complex obtained from the market is a technical product
with 90% of purity. XRD spectra for the as-received sample
and the MA sample are showed in Fig. 2. All the peaks can be
indexed to BA (JCPDS card no. 13-0292) [14]. Fig. 3 shows an
enlargement of the regions where the most representative peaks
related to BA are located. The as-received sample shows narrow
peaks and a family of peaks strictly correlated to the main one.
The BM sample shows a decreasing of the satellite peaks. This
behavior can be related to a increasing symmetry in the ball
milled sample.
Before each measurement the sample (about 50 mg batches)
wasintroducedinacylindricalglassholderequippedwithaglass
piston to eliminate the dead volume. By using a four-way valve
the sample holder was alternatively connected with a vacuum
pump, a 5% argon/hydrogen gas mixture, and the measuring
apparatus. Before each measure the holder was evacuated and
filled with the argon/hydrogen mixture for three times and finally
connected with the measuring apparatus. The latter consists of
a 50 ml graduate burette filled with a 1.0 M solution of LiClO4
in propylene carbonate. To monitor the volume variations as
a function of time and temperature, the resistance of the solu-
tion was recorded between two nickel electrodes fixed inside
the burette. The resistance was evaluated measuring the voltage
drop when a very small alternating current flows into the circuit.
The volume of the displaced solution was directly read from the
burette and used to calibrate the conductimetric cell response.
Fig. 2. XRD spectra of as-received (a) and MA (b) NH3BH3 sample.

186
S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
Fig. 3. Details of the BA peaks of Fig. 2; the upper three are for the MA sample, the lower ones are the corresponding angle peaks for the as-received sample.
XRD spectra for the 1 and 2 mol% platinum MR samples
are showed in Fig. 4(a). There is a strong reduction of the peak
intensities related to the BA complex and the appearance of
new very broad peaks. The peaks at 39.7◦, 46.2◦, and 67.4◦ can
be ascribed to platinum metal (JPDC no. 04-0802). Fig. 4(b)
and (c) are a zoom of Fig. 4(a) and show the [1 1 1] platinum
reflection peak. The platinum grain size (D) was about 9.39 nm
and 9.16 nm for the 1 and 2 mol% doped samples, respectively.
Two broad peaks located at 26.9◦ and 32.0◦ were unidentified,
they could belong to intermediate reaction products. It is worth
to note that other compounds containing platinum and chlorine
have peaks located in these regions.
3.3. Thermal decomposition
Fig. 6 shows the time evolution of the decomposition reac-
tion of BA complex as-received (Fig. 6(a)), after MA (Fig. 6(b))
and MR with 1 mol% Pt (Fig. 6(c)). It was assumed that the
gas evolved in the decomposition process was hydrogen and
in the figures the hydrogen weight percent with respect to the
weight of the sample is reported as a function of time. In fur-
ther analysis the decomposition is assumed to be a first order
reaction. At 95 ◦C, the decomposition rate for the as-received
sample (Fig. 6(a)) is very slow and after 4500 s about 3% hydro-
gen was desorbed. By increasing the temperature the same
amount of hydrogen is desorbed in shorter times and the final
desorbed amount slightly increased. The maximum hydrogen
amount desorbed at 140 ◦C after 4500 s was 5.2%. The melting
point of NH3BH3 (110 ◦C) lies in the middle of the explored
region. As a consequence, the data are collected from solid or
partially melted materials. Anyway, the effect of phase tran-
sition does not seem to affect the decomposition times and
rates.
The milled sample shows decomposition time of the same
order of magnitude of the as-received sample but increased
amount of desorbed hydrogen. The hydrogen released at 4500 s
and 90 ◦C was about 4% and it increased to 6.4% at 140 ◦C. In the
case of 1 mol% doped sample the decomposition process started
at temperature lower than the as-received and MA sample. At
about 73 ◦C the decomposition was already well established. By
increasing the temperature up to 83 ◦C about 3% hydrogen was
collected in 4500 s and a maximum of 7.5% was measured at
150 ◦C. The 2 mol% doped sample showed faster kinetics and it
promptly reacted at temperature as low as 70 ◦C.
3.2. Morphological investigation
The morphological features of the four samples are showed in
Fig. 5. The SEM image of the as-received BA (Fig. 5(a)) shows
a disorganized structure. The presence of some fibres, arranged
to leave some holes, could suggest reef-type morphology. When
BA is subjected to ball milling the material loses its original
size and the aggregate dimensions are reduced (Fig. 5(b)). It is
worth to note that the fibre-like morphology is still retained and
material surface looks like foam. The incorporation of 1 mol%
H2PtCl6 during ball milling induced a transformation in the
sample morphology and a more dense structure takes place
(Fig. 5(c)). The material presents several holes, some of these
having white contours are covered with a thin layer, so they look
like bubbles. The increase in the amount of H2PtCl6 (2 mol%)
increases the bubble number and dimension (Fig. 5(d)). The
map analysis of the elements of the doped samples showed
that Pt and Cl are homogeneously distributed on the samples
as showed in Fig. 5(e) for the 1 mol% doped sample. The white
spots on the material surface were identified as metal platinum
by EDS.
Fig. 7 shows the time dependence of the hydrogen evolution
for the 1% doped sample heated at 73 ◦C. Less than 0.5%
hydrogen develops at temperature lower than the equilibrium

S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
187
Fig. 8 shows the Arrhenius plot of the kinetic constants, as
obtained from τ (k = ln(1/2)/τ); the as-received and the MA sam-
ple show similar behaviour and the points are scattered on the
same curve. Referring to the pristine sample, the 1 mol% doped
sample showed better desorption kinetics. The lower curve is
related to the further decomposition process appearing only in
the doped material.
Fig. 9 shows the amount of hydrogen released from the
1 mol% doped sample as a function of the sample temperature.
In the figure the oven temperature is also reported. The latter
was always lower than the sample one. By increasing the oven
temperature until 120 ◦C it results an increase of the desorption
temperature (that increase up to 105 ◦C). The desorption tem-
perature did not change by further increasing oven temperature.
A decomposition process was observed for temperature higher
than 100 ◦C with the decomposition temperature increasing with
the oven temperature.
4. Discussion
The hydrogen desorbed at 95 ◦C from BA complex as-
received was 3% that is less than the 50% of the theoretical
value (6.49%) as calculated from reaction (1):
BH3NH3 → −(BH2NH2)− + H2
(1)
The volumetric titrations reported in the work of Wolf [15]
showed that 1 mol of hydrogen is desorbed at 90 ◦C in few hours.
The amount of hydrogen released in our experiment is lower than
the previously reported one. This result cannot be explained tak-
ing in consideration only the material purity (90%). We observed
that part of the material is pushed up from the thermalized zone
as a consequence of the thermal decomposition, remaining non-
reacted.
The MA produced a material that showed increased hydrogen
desorption properties. At 96 ◦C the amount of hydrogen released
in 1 h was about 4% and the theoretical amount, corresponding
to one hydrogen per mol of BA complex, was released at 136 ◦C.
The changes in the morphology induced by MA could be both
responsible for the enhanced gas evolution.
The doped sample spectra showed the presence of nano-scale
platinum metal particles well distributed on the material surface.
The following chemical reaction (2) can be evoked for platinum
displacement:
2H2PtCl6 + 3BH3NH3 → 3BCl3 + 2Pt + 3NH4Cl + 5H2
(2)
Fig. 4. (a) XRD spectra for the 1 and 2 mol% platinum MR samples; (b, c)
zooms of the peaks in (a) related to [1 1 1]Pt.
Hydrogen and boron chloride are gaseous species (boiling
point for BCl3 is 12 ◦C) and gas evolution was expected as
result of MR. From Eq. (2) the ammonium chloride should be
present but this compound was not identified by XRD spec-
tra. Anyway, Eq. (2) should be regarded as one the possible
pathway leading to metallic platinum. Other intermediate reac-
tions could be involved in the process and the unidentified peaks
located at 26.9◦ and 32.0◦ could be related to intermediate
platinum–chlorine adducts.
one. About 1.5% hydrogen was collected during a fast kinetic
process at 73 ◦C and a third slower decomposition process,
observed at longer time, contributes for additional 1% hydro-
gen. The final amount of hydrogen desorbed in 6 h run was
3%. The characteristic time of the reaction (τ) was evaluated
by calculating the time corresponding to the maximum of the
volume/time derivative curves.

188
S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
Fig. 5. SEM images of as-received BA (a), ball milled BA (b), with 1 mol% H2PtCl6 (c, e), with 2 mol% H2PtCl6 (d).
The presence of bubbles as observed by SEM indicates gas
metal Eq. (3) or other gaseous species as described by the work
of Wolf et al. [15].
evolution and this could be related to gaseous product forma-
tion during MR (hydrogen and/or boron chloride). Evolution of
gaseous hydrogen, produced during MR and trapped inside the
material, could be correlated to the first decomposition step as
observed by volumetric titration at 70 ◦C (see Fig. 7). The sec-
ond step is the decomposition of BA as described in Eq. (1).
Finally the third step could be related to hydrogen evolution
coming from the decomposition product catalyzed by the noble
Pt
(BH2NH2) −→ (BHNH) + H2
(3)
Considering only the second step, the total amount of hydro-
gen released from the doped sample at 103 ◦C is about 3.5%.
This value is lower than the one showed by the milled sample
(4.2%). This result can be related to hydrogen consumption: (a)

S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
189
Fig. 6. Time evolution of the decomposition reaction of BA complex as-received (a), after MA (b) and MR with 1 mol% Pt (c).
for platinum reduction as illustrated in Eq. (2); (b) for water
molecules reduction as described in Eq. (4). Water adsorbed
on hydrogen hexachloroplatinate acid (about 1:1 mol) can react
with the BA complex decreasing the hydrogen content.
The 2 mol% doped sample showed higher decomposition
rates. The use of larger amount of expensive platinum was not
reputed convenient with respect both economic as well as tech-
nical point of view (decomposition temperatures lower than fuel
cell working temperature were achieved).
3
9
2
2BH3NH3 + H2O → B2O3 + 2NH3 + H2
(4)
2
Fig. 7. Time evolution of hydrogen desorption for the 1% doped sample heated
at 73 ◦C.
Fig. 8. Arrhenius plot of the kinetic constants.

190
S.D. Benedetto et al. / Thermochimica Acta 441 (2006) 184–190
evaluate hydrogen purity and clarify the effect of the catalyst on
the decomposition reaction. Further research activities are on
going to study the effect of less noble metals as catalyst on the
decomposition reaction of BA complex.
Acknowledgment
We thank the Ministero Italiano per la Ricerca Scientifica e
Tecnologica (MIUR) for financial support.
References
[1] http://www.millenniumcell.com.
[2] S.G. Shore, W. Parry, J. Am. Chem. Soc. 77 (1955) 6084.
[3] E.L. Lippert, W.N. Lipscomb, J. Am. Chem. Soc. 78 (1956) 503.
[4] M.G. Hu, R.A. Geanangel, W.W. Wendlandt, Thermochim. Acta 23
(1978) 249.
[5] F. Baitalow, J. Baumann, G. Wolf, K. Jaenicke-Ro¨ßler, G. Leitner, Ther-
mochim. Acta 391 (2002) 159.
Fig. 9. The amount of hydrogen released from the 1 mol% doped sample as a
function of the sample temperature. The nominal temperature is also reported.
[6] Y.L. Chen, A.R. Yones, U. Miller, Metall. Mater. Trans. A 33A (2002)
2713.
[7] F. Zhou, Y.T. Chou, E.J. Lavernia, J. Mater. Res. 17 (12) (2002) 3230.
[8] M.A. Rao, S. Bhargava, D. Deva, Matall. Mater. Trans. A 36A (2005)
3195.
5. Conclusions
It was found that nano-sized platinum metal is formed as
result of MR of BA complex. The decomposition temperature
of the 1% mol Pt doped sample match the working temperature
of PEM fuel cell. As result the thermal decomposition of doped
BA could be an efficient way to produce hydrogen on demand at
convenient and safe low pressure. Evolved gas detection as well
as solid decomposition product analysis should be conducted to
[9] M. Oehring, T. Klass, R. Bormann, J. Mater. Res. 8 (11) (1996) 2819.
[10] M. Pardavi-Horvath, L. Takacs, J. Appl. Phys. 75 (1994) 5864.
[11] P.G. McCormick, Mater. Trans. Jpn. Inst. Met. 36 (1995) 161.
[12] G.C. Andrews, T.C. Crawford, Tetrahedron Lett. 21 (1980) 693.
[13] J.L. Langford, A.J.C. Wilson, J. Appl. Cryst. 11 (1978) 102.
[14] Hughes, J. Am. Chem. Soc. 78 (503) (1956).
[15] G. Wolf, J. Baumann, F. Baitalow, F.P. Hoffmann, Thermochim. Acta
343 (2000) 19.