international journal of hydrogen energy 35 (2010) 8310–8316

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

A comparative study on effect of microwave sintering and

conventional sintering on properties of Nd–Mg–Ni–Fe3O4
hydrogen storage alloy

Jie Meng a, Yan-Biao Pan a, Qun Luo a, Xue-Hui An a, Yang Liu a, Qian Li a,*,
Kuo-Chih Chou a,b
a
Shanghai Key Laboratory of Modern Metallurgy and Materials Processing, Shanghai University, Shanghai 200072, China
b
Department of Physical Chemistry, University of Science and Technology Beijing, Beijing 100083, China

article info abstract

Article history: The hydrogen storage samples of Nd–Mg–Ni–Fe3O4 alloy were prepared by microwave
Received 28 September 2009 sintering (MS) and conventional sintering (CS) methods, respectively. Their phase struc-
Received in revised form tures, morphologies, hydrogen storage properties were intensively studied by X-ray
27 November 2009 diffraction (XRD), scanning electron microscopy (SEM) and pressure–composition–
Accepted 3 December 2009 temperature (PCT). XRD and SEM analysis results show that the microwave sintered Nd–
Available online 23 December 2009 Mg–Ni–Fe3O4 alloy has multiphase structure involving Mg and homogeneous grains,
whereas the alloy prepared by CS has Mg41Nd5 phase and coarse grains. The alloy prepared
Keywords: by MS can release 85% of the saturated hydrogen capacity at 573 K in 600 s and its char-
Hydrogen storage alloy acteristic reaction time (tc) is less than 2900 s, while the alloy prepared by CS releases less
Microwave sintering than 70% of the absorbed hydrogen at 573 K within 1300 s and its tc is more than 3000 s. It is
Conventional sintering found that the alloy prepared by MS not only has high hydrogen capacity, but also better
Hydrogen storage property dehydriding kinetic property than the alloy prepared by CS.
ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction considered competitive because of their rather sluggish

absorption/desorption kinetics. In recent years, many
Since the energy shortage and environmental problems are attempts have been further made to overcome these draw-
getting more and more serious, developing a new kind of clean backs and improve the hydrogen storage properties: (i) adding
energy attracts many scientists and engineers from different alloying element or catalytic components [2], (ii) modifying
fields. Hydrogen is relevant to all of the energy sectors and is surface properties [3], (iii) employing novel preparing tech-
also the ideal means of storage, transport and conversion of niques [4,5] including mechanical alloying, smelting method,
energy for a comprehensive clean-energy concept. However, powder sintering, diffusion method, hydriding combustion
the lack of practical methods to store hydrogen has become synthesis, sputter method, physical vapor deposition and
a technical bottleneck for the use of hydrogen as a fuel, melt-spinning. Among them, smelting method is a kind of
especially for the zero-emission vehicles [1]. Among the many classical technique. However, in some cases it is impossible to
different candidate materials, Mg hydride has a high storage obtain a sample with good quality because of the low misci-
capacity by weight and is therefore favored for mobile appli- bility of magnesium with the most transition metals and its
cations. However, light metal hydrides have not been low melting point (923 K) and high vapor pressure [6].

* Corresponding author. Tel.: þ86 21 5633 4045; fax: þ86 21 5633 8065.
E-mail address: [email protected] (Q. Li).
0360-3199/$ – see front matter ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2009.12.011
 international journal of hydrogen energy 35 (2010) 8310–8316 8311

Microwave heating is a mature technique in chemistry

Table 1 – Experimental conditions.
and materials processing, moreover, it presents striking
advantages over conventional thermal treatments [7]. Samples CS-I MS-i CS-II MS-ii
Microwave sintering could influence the microstructure and Sintering methods CS MS CS MS
the tailored microstructure can improve the material prop- Holding temperature (K) 943 943 873 873
erties. A lot of exploring experimental and theoretical work Dwell time (min) 15 15 60 60
had been done on studying the microwave field and many Heating rate (K/min) 40 40 40 40
Compressed pressure 900 900 300 300
important experimental phenomena with great theoretical
(MPa)
value were achieved. In 1999, Roy et al. [8] reported that
Duration at the constant 10 10 5 5
particulate metals could be heated rapidly in a microwave temperature (min)
field. This has led to the use of microwave energy to Mixing method Ball Ball Ball Ball
consolidate a range of particulate metals and alloys. Micro- milling milling milling milling
wave energy has been used in recent years in hydrogen Cooling-down method Furnace Furnace Furnace Furnace
storage material. Mu et al. [9] employed microwave plasma cooling cooling cooling cooling

etching and Pd decoration for modification of carbon nano-

tubes (CNTS) to increase the defects on the nanotube and
improve the hydrogen storage capacity of the etched CNTS. The PCT and kinetic properties were measured by a volu-
Sung et al. [10] comparatively studied dehydrogenation of metric method according to the Sievert’s law using automatic
bulky, branched and polycondensed naphthenes for apparatus from SUZUKI HOKAN.CO., LTD. in Japan. The phase
hydrogen storage in microwave and thermal modes and structure was determined by a D\max-2550 X-ray diffrac-
pointed that the microwave-assisted mode could decrease tometer with Cu Ka radiation. The X-ray intensity was
the dehydrogenation temperature and power supply. measured over a diffraction angle 10–90 with a speed of 8 per
However, up to date, although some good results have been minute. The data obtained from XRD was analyzed by the
obtained in application of microwave field in hydrogen software of MDI Jade 5.0. The morphological and microstruc-
storage materials, the mechanisms on microwave field tural characterizations of samples were examined by JSM-
which plays a role on microstructures, phase transformation 6700F scanning electron microscope with an energy dispersive
and hydrogenation properties are still unclear [11]. X-ray spectrometer (EDS).
In this work, in order to show some evidence to explain an
effect of a microwave field on hydrogen storage material,
Nd–Mg–Ni–Fe3O4 were prepared by microwave sintering
method and conventional sintering method and the hydrogen 3. Results and discussion
storage properties of the different samples were compara-
tively investigated as well as the kinetic mechanism of their 3.1. Microstructural analysis
dehydriding reaction was discussed using a new model.
Fig. 1 shows the XRD patterns of different samples before and
after hydriding activation. It illustrates that the phases in the
2. Experimental details samples prepared by CS are mainly the tetragonal Mg41Nd5
(PDF-#45-1031) phase as well as minor Fe0.872O (PDF#71-0161)
The nominal composition of the experimental alloy is phase, Fe3O4 (PDF#79-0418) phase and Fe0.946Ni0.054 (PDF#44-
Nd1.5Mg17Ni0.5(Fe3O4)0.15. The purity of all the component 1088) phases, whereas the samples prepared by MS contain
powders (Nd, Mg, Ni and Fe3O4) is at least 99.5%. The initial the hexagonal Mg (PDF-#63-3655) phase, NdO2 (PDF#46-1074),
materials were mixed for 18 h in planetary ball milling Fe3O4 phase and [Fe, Ni] (PDF#37-0474) phase. There is also
(QF-1sp, made by Nanjing University) under high purity MgO (PDF#75-0447) phase in sample MS-ii. It was reported
argon without oxidization during ball milling process. All that the application of microwave technology could decrease
operation including filling in the sample and taking it out of activation energy of chemical reaction [12–14], so there were
ball milling tanks was carried out in the glove box. The many other metallic oxides formed by some subsidiary
tankage used in this experiment was 80 cm3 and the weight reaction in microwave sintering. After hydriding activation,
ratio of grinding media to material was 8:1. The powder the phases of conventionally sintered samples change to the
samples were compressed to form pellet 5 mm in height and tetragonal MgH2 (PDF-#12-0697) phase, the tetragonal Nd2H5
15 mm in diameter by a uniaxial single-acting press. The (PDF-#36-1247) phase and [Fe, Ni] phase, whereas the
pellet samples were put into the center of microwave oven microwave sintered samples contain the tetragonal MgH2
and conventional heating oven, respectively. The samples (PDF-#12-0697) phase, Nd2O3 (PDF#65-3184) phase, some
were sintered under high purity Ar gas according to the unreacted Mg phase and [Fe, Ni] phase. The metallic oxides
processing parameters which were showed in Table 1. It is act as catalysts to improve hydrogen absorption/desorption
the same processing condition and different sintering kinetics [15]. Lattice parameters and cell volumes of main
methods between sample CS-I and sample MS-i. Finally, the phase of different samples were calculated and listed in
samples were crushed into micro-particles and activated for Table 2. It can be seen that the cell volumes of the Nd5Mg41
6 times (absorption at 573 K and under 4.0 MPaH2 for phase in both samples prepared by CS were slightly smaller
120 min, and desorption in vacuum of 0.001 MPa for 120 min than those of normal Nd5Mg41 phase. In the microwave sin-
at the same temperature). tered samples, the cell volume of Mg in sample MS-i was
8312 international journal of hydrogen energy 35 (2010) 8310–8316

Fig. 1 – XRD patterns of different samples: (A) before initial activation and (B) after hydriding activation.

slightly smaller than that of normal Mg phase and the cell decreases after hydriding activation sufficiently. From the
volume of Mg in sample MS-ii was larger than that of normal EDS, both samples CS-II and MS-ii are mainly composed of Nd,
Mg phase. Mg, Ni, Fe and O, but element content is slightly different.
Figs. 2 and 3 show the SEM morphologies and EDS From the atom amount of the elements, it reveals that sample
diagrams of different samples. It reveals that the micro- CS-II may contain Nd5Mg41 phase and sample MS-ii may
structure of microwave sintered samples is homogeneous and contain Mg phase and oxide phases, which almost agrees with
grains are distributed uniformly. Compared with the the XRD analysis.
conventionally sintered samples, the samples prepared by MS
have smaller particles, which is beneficial to the hydrogen
3.2. Thermodynamic properties
storage properties of the alloys due to uniform and volumetric
heating in microwave field. Heating method during a conven-
Fig. 4 shows that the PCT curves of different samples at
tional sintering is a thermal conduction from the heating
different temperatures and Van’t Hoff diagrams for hydrogen
elements through the oven wall and holding vessel, and from
absorption/desorption of all the samples. The hydrogen
the outside to the inside of the sample. Therefore, all heat
storage capacity of the samples prepared by CS is less than
energy must pass through its surface of sintering sample. The
4.0 wt.% under 553 K–623 K, whereas the samples prepared by
source of the heating in the microwave oven is oscillations of
MS have higher hydrogen storage capacity and wider plateau
the free electrons at high frequency microwave (w2.45 GHz) in
metal samples. Therefore, the volumetric heating during
a microwave sintering is initiated throughout the sample
instantaneously as a result of the high frequency microwave Table 2 – Lattice parameters and cell volumes of main
energy. Thus, thermal equilibrium is almost instantaneous phases for different samples.
and it takes only a few minutes to complete the microwave Samples Phase a (Å) b (Å) c (Å) V (Å3)
sintering process. However, several hours are needed for
conventional heating because of the thermal gradient from PDF(#45-1031) Nd5Mg41 14.76 14.76 10.39 2263.54
PDF(#65-3365) Mg 3.21 3.21 5.21 46.50
the surface to the center of the sintering body. This is the
PDF(#12-0697) MgH2 4.517 4.517 3.021 61.63
reason why the grain growth is generally less in microwave
CS-I Nd5Mg41 14.7667 14.7667 10.3555 2258.08
processed material than that in conventionally sintered one. CS-I-H MgH2 4.5089 4.5089 3.0149 61.29
There is a large difference in grain size between the conven- CS-II Nd5Mg41 14.7691 14.7691 10.3750 2263.05
tionally made sample and the microwave sintered one. The CS-II-H MgH2 4.5190 4.5190 3.0212 61.70
distribution of particles is fine and homogeneous in the MS-i Mg 3.2087 3.2087 5.2132 46.48
microwave sintered sample. It represents uniform distribu- MS-i-H MgH2 4.5214 4.5214 3.0256 61.85
MS-ii Mg 3.2106 3.2106 5.2130 46.54
tion with small particles in the SEM picture of sample MS-i,
MS-ii-H MgH2 4.5178 4.5178 3.0281 61.81
and the grain size in the SEM picture of sample MS-i-H further
 international journal of hydrogen energy 35 (2010) 8310–8316 8313

Fig. 2 – SEM morphologies of different samples before and after hydriding activation.

of hydrogen absorption/desorption. Especially, the hydrogen a certain time to the saturated hydrogen capacity). The DH and
storage capacity of sample MS-i is more than 4.5 wt.% at DS of the hydrogen absorption/desorption reactions in all
553 K–623 K, which is higher than that of sample CS-I prepared samples can be calculated by the Van’t Hoff equation [16]
with the same processing condition as the sample MS-i. Table using the slopes and intercepts of the lines and the relation-
3 presents the result of the hydrogen absorption/desorption ship between equilibrium plateau pressure and temperature
properties and thermodynamic data of different samples. It in the hydrogen absorption/desorption procedure is then
reveals that all samples can release more than 94% of the derived. The DH and DS of hydrogen absorption of microwave
absorbed hydrogen among 553 K–623 K, which keep high sintered samples were lower than those of the conventionally
desorption ratios (the amount of hydrogen desorption within sintered samples, which is benefit to hydrogen absorption. It

Fig. 3 – SEM morphologies and EDS diagrams of samples CS-II and MS-ii.
8314 international journal of hydrogen energy 35 (2010) 8310–8316

Fig. 4 – PCT curves of the different samples at different temperatures and Van’t Hoff diagrams for hydrogen absorption/
desorption of the alloys.

is generally recognized that the larger cell volume of phase, the practical application. Fig. 5 presents hydrogen desorption
the more stability of the hydride [17]. The microwave sintered kinetic curves of different samples at temperature from 553 to
samples are not worse than the conventionally sintered 623 K. As is known, at low temperature dehydriding kinetic
samples in thermodynamic data. properties of Mg-based hydrogen storage alloy are not satis-
factory. At 553 K, the hydrogen desorption ratio of different
3.3. Kinetic properties samples reaches 34.4% (CS-I), 42.3% (CS-II), 83.1% (MS-i) and
91.0% (MS-ii) within 2000 s, respectively. The dehydriding
Compared with the hydriding reaction rate, the desorption kinetic properties were improved at 573 K, within 600 s, the
rate is more important property of hydrogen storage alloy for hydrogen desorption ratios of four samples reached 54.0%

Table 3 – The hydrogen absorption/desorption properties and thermodynamic data of different samples.
Sample T Pab. Pde. Cab. Cde. Drate Absorption Desorption
(K) (MPa) (MPa) (wt.%H) (wt.%H) (%)

MS-i 553 0.118 0.077 4.483 4.393 98.0 DH ¼ 71.9  2.6 kJ/mol DH ¼ 83.5  8.8 kJ/mol
573 0.211 0.173 4.602 4.423 96.1 DS ¼ 112.2  4.5 J/(mol K) DS ¼ 130.5  15.2 J/(mol K)
623 0.676 0.659 4.602 4.519 98.2 ln(P MPa) ¼ 8652.2/T þ 13.5 ln(P MPa) ¼ 10040.8/T þ 15.7

MS-ii 553 0.158 0.062 3.848 3.619 94.0 DH ¼ 72.7  10.4 kJ/mol DH ¼ 99.6  7.6 kJ/mol
573 0.209 0.157 4.167 4.010 96.2 DS ¼ 115.6  18.0 J/(mol K) DS ¼ 158.0  13.0 J/(mol K)
623 0.861 0.652 3.610 3.454 95.7 ln(P MPa) ¼ 8764.2/T þ 13.9 ln(P MPa) ¼ 11975.0/T þ 19.0

CS-I 553 0.112 0.092 4.051 3.998 98.7 DH ¼ 74.5  0.4 kJ/mol DH ¼ 80.2  1.3 kJ/mol
573 0.190 0.156 3.736 3.580 95.8 DS ¼ 116.4  0.7 J/(mol K) DS ¼ 125.5  2.3 J/(mol K)
623 0.670 0.628 3.823 3.670 96.0 ln(P MPa) ¼ 8956.0/T þ 14.0 ln(P MPa) ¼ 9642.1/T þ 15.1

CS-II 553 0.118 0.083 3.572 3.373 94.4 DH ¼ 74.6  6.8 kJ/mol DH ¼ 83.3  1.5 kJ/mol
573 0.172 0.152 3.416 3.343 97.9 DS ¼ 116.4  11.6 J/(mol K) DS ¼ 130.5  2.6 J/(mol K)
623 0.684 0.621 3.615 3.487 96.5 ln(P MPa) ¼ 8975.0/T þ 14.0 ln(P MPa) ¼ 10023.1/T þ 15.7

Note: Pab. and Pde. are hydrogen absorption/desorption plateau pressure; Cab. and Cde. are capacities of hydrogen absorption/desorption; Drate is
the percent of desorbed hydrogen.
 international journal of hydrogen energy 35 (2010) 8310–8316 8315

The dehydriding kinetic of the samples prepared by two

methods was studied by using Chou model [18] at 553 K, 573 K
and 623 K. The results showed that the rate-controlling step of
the dehydriding reactions is the diffusion of hydrogen in the
hydride and the curves fitted very well by Chou model in Fig. 5.
The tc of the samples prepared by MS was shorter than that of
the samples prepared by CS. For instance, at 573 K and 623 K
the tc of MS-i was 2836.7 s and 687.8 s, but the tc of CS-I was
3092.6 s and 973.5 s. At 573 K and 623 K the tc of MS-ii was
2457.0 s and 709.8 s, but the tc of CS-II was 4194.2 s and 874.6 s.
It was well known that the smaller the characteristic reaction
time, the faster the reaction rate. The comparison mentioned
above suggests that the samples prepared by MS are superior
to the samples prepared by CS in respect of hydrogen
desorption kinetic. There are several reasons for explanation:
(i) The oxides of Nd element in the samples prepared by MS
before and after hydriding activation may act as catalyst to
improve hydrogen absorption/desorption kinetics. (ii) The
samples prepared by CS have coarse grains than samples
prepared by MS. The distribution of particles is fine and
homogeneous in the microwave sintered sample. (iii) In
microwave field heating is a consequence of the energy
exchange of the electromagnetic field with the dielectric
system. It occurs by rotation and distortion of induced and
permanent molecular dipoles, since their typical relaxation
times are of the same magnitude of microwave frequencies.
The microwave energy absorbed by the material is converted
into molecular kinetic energy and dissipated by each molecule
against the inertial, elastic and frictional forces of the
surroundings. Because of these microscopic mechanisms, the
microwave field generates volumetric, fast and selective
heating in the material. The samples are heated all over and
sintered uniformly, and then grains of the samples are fine
and homogeneous, which are in favor of improving the
properties of hydrogen storage.

4. Conclusions

The Nd–Mg–Ni–Fe3O4 hydrogen storage alloys were prepared

by microwave sintering and conventional sintering, respec-
tively. Compared with the conventionally sintered samples in
structure and properties of hydrogen storage, the microwave
sintered Nd–Mg–Ni–Fe3O4 alloy has fine and homogeneous
crystal grains, and a large specific surface area, leading to
better properties of hydrogen absorption/desorption. In
respect of thermodynamics the microwave sintered Nd–Mg–
Fig. 5 – Hydrogen desorption kinetic curves and the fitted Ni–Fe3O4 alloy keeps a good capacity of hydrogen storage (at
curves calculated by Chou model at different temperatures 623 K, 4.6 wt.% H), and in respect of kinetics the alloy has
for different samples. better dehydriding kinetic properties, which desorbed more
than 80% of the absorbed hydrogen at 623 K and cost less than
200 s to finish desorbing. The alloy prepared by MS could
(CS-I), 58.1% (CS-II), 87.5% (MS-i) and 86.1% (MS-ii). At 623 K, release 85% of the saturated hydrogen capacity at 573 K in
the dehydriding kinetic properties of all samples were further 600 s and its tc is less than 2900 s, while the alloy prepared by
improved. The time needed to arrive the largest hydrogen CS released less than 70% of the absorbed hydrogen at 573 K
desorption ratio is 320 s (CS-I), 280 s (CS-II), 200 s (MS-i) and within 1300 s and its tc is more than 3000 s. Generally
160 s (MS-ii), respectively. Obviously, the dehydriding kinetic speaking, the Nd–Mg–Ni–Fe3O4 hydrogen storage alloy
property of microwave sintered samples is better than that of prepared by microwave sintering has better overall properties
the samples prepared by CS. of hydrogen storage.
8316 international journal of hydrogen energy 35 (2010) 8310–8316

sintering of WC/Co composites. Mater Sci Eng A 2005;391:

Acknowledgements 285–95.
[8] Roy R, Agrawal D, Cheng JP, Gedevanishvili S. Full sintering of
This work was financially sponsored by the National High powdered-metal bodies in a microwave field. Nature 1999;399:
Technology Research and Development Program of China 668–70.
[9] Mu SC, Tang HL, Qian SH, Pan M, Yuan RZ. Hydrogen storage
(2007AA0FZ118), the National Natural Science Foundation of
in carbon nanotubes modified by microwave plasma etching
China (50804029), a Foundation for the Author of National
and Pd decoration. Carbon 2006;44:762–7.
Excellent Doctoral Dissertation of PR China (200746), and [10] Sung JS, Choo KY, Kim TH, Tarasov AL, Soldatov AV,
Changjiang Scholars and Innovative Research Team in Kustov LM. Comparative study on dehydrogenation of bulky,
University (IRT0739). branched and polycondensed naphthenes for hydrogen
storage in microwave and thermal modes. Int J Hydrogen
Energy 2008;33:4116–21.
references [11] Liu Y, Li Q, Lin GW, Chou KC, Xu KD. Properties of Mg2NiH4
prepared by microwave-assisted activation synthesis from
micro-particles. J Alloys Compd 2009;468:455–61.
[1] Elam CC, Padro CEG, Sandrock G, Luzzi A, Lindblad P, [12] Rao KJ, Ramesh PD. Use of microwaves for the synthesis
Hagen EF. Realizing the hydrogen future: the international and processing of materials. Bull Mater Sci 1995;18(4):
energy agency’s efforts to advance hydrogen energy 447–65.
technologies. Int J Hydrogen Energy 2003;28(6):601–7. [13] Freeman SA, Booske JH, Cooper RF. Microwave field
[2] Bobet J–L, Akiba E, Nakamura Y, Darriet B. Study of Mg–M enhancement of charge transport in sodium chloride. Phys
(M¼Co, Ni and Fe) mixture elaborated by reactive mechanical Rev Lett 1995;74:2042–5.
alloying: hydrogen sorption properties. Int J Hydrogen Energy [14] Rybakov KI, Semennov VE, Freeman SA, Booske JH,
2001;26:493–501. Cooper RF. Dynamics of microwave-induced currents in
[3] Feng F, Northwood DO. Effect of surface modification on the ionic crystals. Phys Rev B 1997;55:3559–67.
performance of negative electrodes in Ni/MH batteries. Int J [15] Wang Y, Gao XP, Lu ZW, Hu WK, Zhou Z, Qu JQ, et al. Effects
Hydrogen Energy 2004;29(9):955–60. of metal oxides on electrochemical hydrogen storage of
[4] Huang LJ, Liang GY, Sun ZB, Zhou YF. Nanocrystallization nanocrystalline LaMg12-Ni composites. Electrochim Acta
and hydriding properties of amorphous melt–spun 2005;50:2187–91.
Mg65Cu25Nd10 alloy. J Alloys Compd 2007;432:172–6. [16] Schlapbach L, Züttel A. Hydrogen-storage materials for
[5] Sakintuna B, Lamari–Darkrim F, Hirscher M. Metal hydride mobile applications. Nature 2001;414(6861):353–8.
materials for solid hydrogen storage: a review. Int J Hydrogen [17] Mendelsohn MH, Gruen DM, Dwight AE. The effect of
Energy 2007;32:1121–40. aluminum additions on the structural and hydrogen
[6] Shang CX, Bououdina M, Song Y, Guo ZX. Mechanical absorption properties of AB5 alloys with particular reference
alloying and electronic simulations of (MgH2 þ M) systems to the LaNi5-xAlx ternary alloy system. J Less-Common Met
(M¼Al, Ti, Fe, Ni, Cu and Nb) for hydrogen storage. Int J 1979;63:193–207.
Hydrogen Energy 2004;29:73–80. [18] Chou KC, Li Q, Lin Q, Jiang LJ, Xu KD. Kinetics of absorption
[7] Breval E, Cheng JP, Agrawal DK, Gigl P, Dennis M, Roy R, and desorption of hydrogen in alloy powder. Int J Hydrogen
et al. Comparison between microwave and conventional Energy 2005;30:301–9.