PARTICLE SIZE EFFECTS ON KINETICS OF

HYDROGEN SORPTION IN MAGNESIUM -

MAGNESIUM HYDRIDE SYSTEM
THESIS
Submitted in partial fulfilment of the requirements for the degree of
BACHELOR OF TECHNOLOGY in
CHEMICAL ENGINEERING

Submitted by
HIYA KAMAL GOSWAMI
(10GCH17)
Under the guidance of

Dr. Sankara Sarma V Tatiparti

Asst. Prof. Dept. of Energy Science and Engineering
I.I.T. Bombay, Mumbai

Prof. Dr. J. B. Naik

Head of Department, Dept. of Chemical Engineering
U.I.C.T. N.M.U., Jalgaon

Submitted to
DEPARTMENT OF CHEMICAL ENGINEERING
UNIVERSITY INSTITUTE OF CHEMICAL TECHNOLOGY
NORTH MAHARASHTRA UNIVERSITY, JALGAON – 425001
JULY, 2014

1
 DECLARATION

I hereby declare that the report of the Industrial Project and Training work entitled
“Particle size effects on kinetics of hydrogen sorption in magnesium-magnesium
hydride system” which is being submitted to University Institute of Chemical
Technology, Jalgaon in partial fulfilment of the requirements for the award of the
degree of Bachelor of Technology in Chemical Engineering from the Department
of Chemical Engineering, is a bonafide report of the work carried out by me. The
material contained in this report has not been submitted to any University or Institution
for the award of any degree.

HIYA KAMAL GOSWAMI

(10GCH17)

Dept. of Chemical Engineering,

U.I.C.T, N.M.U, Jalgaon

Place: I.I.T. BOMBAY, MUMBAI

Date: JULY 2014

2
 CERTIFICATE

This is to certify that the Under Graduate Project Work report entitled “Particle Size
Effects on kinetics of hydrogen sorption in magnesium-magnesium hydride”
submitted by Hiya Kamal Goswami, (Registration Number: 10GCH17) as the record
of the work carried out by her, is accepted as the under graduate project work report
submission which is a part of Industrial Project and Training required in partial
fulfilment of the requirements for the award of degree of Bachelor of Technology in
Chemical Engineering from the Department of Chemical Engineering, U.I.C.T,
N.M.U, Jalgaon.

Dr. Sankara Sarma V Tatiparti Prof. Dr. J. B. Naik

Assistant Professor Head of Department
Dept. of Energy Science and Engineering Dept. of Chemical Engineering
I.I.T. Bombay, Mumbai U.I.C.T., N.M.U, Jalgaon

3
 ACKNOWLEDGEMENT

I would like to express my sincere gratitude to Prof. Dr. Satyendra Mishra, Director,
University Institute of Chemical Technology and Prof. Dr. J. B. Naik, Head,
Department of Chemical Engineering, University Institute of Chemical Technology for
having allowed me to be working on this project.

I feel honoured to be indebted to my internal guide Prof. Dr. J. B. Naik, Head,

Department of Chemical Engineering U.I.C.T, Jalgaon and to my external guide Dr.
Sankara Sarma V Tatiparti, Assistant Professor, Dept. of Energy Science and
Engineering, I.I.T. Bombay, Mumbai for their constant and valuable support, advice,
encouragement and belief in my work.

I would also like to thank Ms Sweta Shriniwasan, PhD Student who made me feel at
ease with her geniality and steadfastness. I also thank her for the cooperation and the
helping hand she provided me during the laboratory work.

Last but not the least I would like to thank my family and my friends who stood by me
and provided me with valuable guidance and encouragement to make this endeavour a
success.

4
 CONTENTS
DECLARATION 2
CERTIFICATE 3
ACKNOWLEDGEMENT 4
LIST OF FIGURES 6
LIST OF TABLES 8
ABSTRACT 9
1. INTRODUCTION 10
2. LITERATURE REVIEW 12
2.1 Magnesium Hydride: Structure 12
2.2 Ball Milling 13
2.3 Kinetics 15
3. EXPERIMENTATION 21
3.1 Ball Milling Procedure 21
4. KINETIC ANALYSIS 29
4.1 Developing the mathematical approach 31
4.2 Reanalysing the logarithmic plot 33
4.3 Reanalysis of existing data for 44 μm particles 34
4.4 Hypothesis formulation 38
5. RESULTS AND DISCUSSION 41
5.1 Application to different Magnesium Powders with different particle sizes 41
5.1.1 Micron size particles (44 μm) 41
5.1.2Particle Size of 10-35 nm 45
5.1.3 Particle Size of 15 nm 48
5.2 Comparison of the p, q and r values 52
6. CONCLUSION 57
7. REFERENCES 58

5
 LIST OF FIGURES

Figure 1: Mechanism of Hydrogenation of Magnesium to Magnesium hydride 12

Figure 2: Magnesium and Magnesium Hydride 13
Figure 3: Fritsch Pulverisette 7 Ball Mill 15
Figure 4: SEM images of Magnesium Hydride 23
Figure 5: Particle size distribution of As Received MgH2. 24
Figure 6: Particle size distribution of 30 minutes ball milled sample 24
Figure 7: Particle size distribution of 1 hour ball milled sample 25
Figure 8: Particle size distribution of 2 hours ball milled sample 25
Figure 9: Particle size distribution of 4 hours ball milled sample 26
Figure 10: Average particle size of sample v/s milling time (hours) 26
Figure 11: X-Ray Spectra of the mechanically milled MgH2 for various time intervals and
pure as received MgH2 27
Figure 12: Crystallite size of samples v/s milling time 27
Figure 13: The schematic diagram explaining overall mechanism in sequence 29
Figure 14: Hydrogenation curves of magnesium powder sample at 210°C under 1MPa (α v/s
time) 31
Figure 15: The JMAK analysis of powders hydrogenated at 210°C under 1MPa in logarithmic
plot 32
Figure 16: JMAK (nucleation and growth) and Experimental α v/s time plots 34
Figure 17: The cross section of powder hydrogenated for different time intervals (microscopic
SEM images) 36
Figure 18: Hydrogenation curves of magnesium powder sample at 210°C under 1MPa
hydrogen pressure at each interrupted times 37
Figure 19: Expected trend of q and r from the qualitative observation of the microscopic
images 39
Figure 20: Particle size distribution of As Received Mg particles 41
Figure 21: Hydrogenation curves of magnesium powder sample at 210°C under 1MPa (44
μm) 42
Figure 22: Hydrogenation curve (figure 23) fitted with the equations for different time
segments 43
Figure 23: The trend of p, q and r with respect to time for hydrogenation of 44 μm particles 43
Figure 24: The trend of p and q with respect to time in the time interval 0-6 minutes 44

6
Figure 25: Hydrogenation curves of magnesium powder sample at 200°C under 2MPa (10-35
nm) 45
Figure 26: Hydrogenation curve (figure 25) fitted with the equations for different time
segments 46
Figure 27: The trend of p, q and r with respect to time for hydrogenation of 10-35 nm
particles 47
Figure 28: The trend of p and q with respect to time in the time interval 0-6 minutes 47
Figure 29: Hydrogenation curves of magnesium powder sample at 270°C under 15 MPa (15
nm) 49
Figure 30: Hydrogenation curve (figure 29) fitted with the equation 8 for different time
segments 50
Figure 31: The trend of p and q with respect to time for hydrogenation of 15 nm particles 50
Figure 32: p v/s t* for different particle sizes 52
Figure 33: p v/s t* for different particle sizes for 0 -5 t* values 53
Figure 34: q (chemisorption – nucleation & growth) v/s t* for different particle sizes 53
Figure 35: q (chemisorption – nucleation & growth) v/s t* for different particle sizes for 0 -5
t* values 54
Figure 36: q (nucleation & growth – diffusion) v/s t* for different particle sizes 54
Figure 37: r v/s t* for different particle sizes 55
Figure 38: α v/s t* for different particle sizes 56

7
 LIST OF TABLES

Table 1: Set of reaction models to describe the reaction kinetics from the literatures 16
Table 2: Qualitative observations from microscopic images 37
Table 3: The t1/2 values for different particle sizes 52

8
 ABSTRACT

Hydrogen is a potential future energy carrier and MgH2 is one of the most promising
hydrogen storage materials due to its high gravimetric capacity and cost. Ball milling
improves the hydrogen sorption behaviour significantly. Mechanical milling results to
particle size reduction and particle size distribution. Within the wide range of particles
the Rate Limiting Step (RLS) varies accordingly. The reaction kinetics of hydrogen
absorption by magnesium (hydrogenation) is investigated using a mathematical
approach where the presence of more than one RLS and the particle size effects on the
sorption kinetics of hydrogenation of magnesium is accounted. The approach accounts
for the dominance of the individual RLS in specific time interval considering
Chemisorption, Nucleation and Growth and Diffusion phenomenon respectively. The
extent and the presence of RLS is observed to change during the hydrogenation with
particle size of the Mg powder.

Keywords: MgH2, RLS, hydrogenation, chemisorption, nucleation and growth,

diffusion

9
 1. INTRODUCTION

With the current scenario of energy crisis and concerns over global warming, there is a
need of alternative ways of energy production, conversion and storage source of energy.
In this competition of energy sources, Hydrogen is a genuine alternative to current
carbon dependence for energy production. The continuous increase of oil prices because
of depletion in fossil fuel reserves worldwide, the growing concerns about energy
security and the environmental problems, several national and international programs
have emerged within the last decade to bring ‘hydrogen economy’ to a mature state
before the end of the 21st century [1].

The basic problems to be solved in order to make hydrogen economy more viable are
1) Expensive production of hydrogen
2) Improvement in the fuel cells which use hydrogen as fuel in automobiles
3) Hydrogen storage system on board

On-board hydrogen storage for transportation applications continues to be one of the

most technically challenging barriers. The technology of storing hydrogen in solid hosts
dramatically improves security related to hydrogen storage (compared to existing
technologies: high pressure > 200 bar (gaseous storage) or liquid storage at -253 °C).
A solid hydrogen storage system is required to possess high volumetric and gravimetric
capacity, low absorption and desorption temperature compatible with the waste heat
generated by a PEM fuel cell, fast kinetics of absorption and desorption, reversibility
and low price; Addition of catalyst together with size reduction of the structural units
such as particles, grains of solid state hydrogen host using techniques such as ball mill
substantially improve the hydrogen storage properties of solid hydrides; the hydrogen
desorption was around 2 wt.% for as received particles but it was around 5-6 wt.% for
5-10 hours ball milled particles [2].

Among the various materials able to store sufficient hydrogen for economic viability,
Magnesium (Hydride) is a promising candidate in terms of abundance, cost and safety.
Magnesium hydride has the highest gravimetric capacity among metals (7.6 wt. %).
However, it suffers from very slow kinetics of hydrogen absorption and desorption.

10
Several hours are required to achieve complete hydrogenation and dehydrogenation on
bulk magnesium [3,4].
The reaction of Hydrogenation is as follows:
𝑀𝑔 + 𝐻2 → 𝑀𝑔𝐻2

Nanostructures increase the active surface area for H2 absorption/desorption and speed
up H diffusion thanks to the reduction of diffusion lengths and/or to the abundance of
fast diffusion paths, i.e. interfaces and grain boundaries [5,6]. In addition, surface
functionalization can be obtained by the addition of transition metals or their oxides,
which act as catalysts for the H2 dissociation/recombination. These features altogether
may result in metal hydride transformation order of magnitudes faster than for
conventional bulk material [6]. As the particle size changes, the surface area also
changes and it affects change in RLS (RLS); Particle sizes of 10-18 μm shows diffusion
controlled kinetics [7] but as the particle size is reduced to 10-100nm the RLS changes
to Nucleation and Growth [8].

Mintz et. al [9] have stated regarding the size of particle in a sample and analysed the
kinetics according to the single particle analysis, but there may be a possibility of
having different sizes of particles within the sample which may give a different scenario
of the kinetics. The change in RLS was termed as a mixed RLS by Chou et. al
considering uniform particle size but it did not significantly explained the physical
phenomena in the curve fit [10].

Hence it is necessary to develop an analytical approach to study the RLS in accordance

with the physical phenomenon with respect to time and size of particles. In this study
we explore the above problem with this proposed approach wherein we investigate the
presence of more than one RLS and the particle size effects on the sorption kinetics of
hydrogenation of magnesium. This approach also accounts for the dominance of the
individual RLS with time.

11
 2. LITERATURE REVIEW

2.1 Magnesium Hydride: Structure

Magnesium has a hexagonal hcp α-structure. In the presence of hydrogen gas in

standard conditions, magnesium absorbs and reacts with hydrogen. Magnesium hydride
is a tetragonal structure (β) and has ionic bonding between Mg and H. The magnesium
cation is octahedrally surrounded by H anions; the H anion has a planar triangle of
magnesium cations around it. In the ball milling procedure, some amount of the
magnesium hydride (β) transforms to an orthorhombic structure (γ) which is stable at
high pressure [1,7,10].

𝑀𝑔 + 𝐻2 → 𝑀𝑔𝐻2

This reaction seems to be simple and one step but the parameters governing the kinetics
and the thermodynamics is complex, and numerous conditions affect the reaction. A
schematic view (Figure 1) of the reaction will explain it in a more precise manner.
Magnesium is reacted with hydrogen gas molecule to yield magnesium hydride;
Hydrogen molecule is first physisorbed and then chemisorbed on the surface with Mg
atoms, later nuclei formation of MgH2 leads to Nucleation which further results to
Growth of Nuclei. After a specific hydride layer is formed, diffusion of H2 molecules
through the hydride layer takes place to react with the core Mg atoms. This whole
process is termed as hydrogen absorption.

H2
Hydrogen absorption
Mg MgH2
Physisorption/Chemisorption → Nucleation-Growth → Diffusion

Figure 1: Mechanism of Hydrogenation of Magnesium to Magnesium hydride

Magnesium is an interstitial solid (α hcp phase) and reacted with H2 gas molecules to
form MgH2 (tetragonal β phase).

12
A schematic views of the crystal structures of magnesium and magnesium Hydride are
shown in Figure 2 and give the insight of the transformation in the reaction.

Figure 2: Magnesium and Magnesium Hydride [11]

2.2 Ball Milling

Magnesium hydride has a high thermodynamic stability but, its use as a hydrogen
storage material is limited by its slow kinetics [12–15]. Slow kinetics results to slow
rate of reaction and also demands for high temperatures [14]. This problem of slow
kinetics is improved to some extent by reducing the particle size to a nano scale [2,12–
16]. To reduce the size of particle many methods are employed like mechanical milling,
gas phase condensation, scaffold methods, self-assembly techniques etc. [1]. Among
all the methods stated ball milling is one of the most effective and an easy method for
size reduction of powdered material. Increased surface area (due to defects, grain
boundaries), decreased particle and crystallite size and a porous surface structure are
other features achieved due to intensive ball milling [16–18]. The specific surface of
the material increases by ten-fold while the crystallite size typically decreases from
micrometres to nanometres [1].

The defects introduced within the surface of magnesium particles during the milling
process, promote fast hydrogen kinetics, and i.e. facilitate surface dissociation and
hydrogen diffusion because of shorter diffusion paths. Also by addition of catalyst
which reduces the activation energy, faster kinetics can be achieved and thus facilitates
hydrogenation and dehydrogenation reactions at low temperatures [19–22]. The RLS

13
of the hydrogen adsorption/desorption in nanocrystalline magnesium have been
reported i.e. the reaction is still controlled by nucleation and growth [1].

In ball milling there are different techniques by which the size of the Magnesium
particles are reduced. One of the technique is to ball mill in an inert atmosphere; the
vial of the ball mill is opened or closed in a glove box [2,23]. Another technique is
controlled reactive ball milling in which hydrogen gas is supplied while milling the Mg
particles. This method results in the hydrogenation reaction within the vial of the ball
mill when the milling process is occurring [24]. During mechanical milling two
processes happen, the formation of the γ phase i.e. β → γ transformation and also the
presence of β phase (γ→β phase thermally transition); γ is the small distortion in β
phase. Ball milling allows overcoming the activation energy barrier, helps in coming to
a metastable phase.

Ball milling magnesium particles for 15-20 hours reduces the crystallite size to 30 nm
and the particle size reduces from a few μm to 200 nm [25]. Addition of catalysts to
magnesium particles during ball milling, reduces the hydrogenation time and
temperature. The most common catalysts used for hydrogenation are Nb2O5, V2O5, Ti,
Ni etc. [12,19,26]. With the use of Nb2O5 catalyst (Mg+5%Nb2O5) and ball milling for
around 20 hours, helps in reducing the crystallite size to nanometres [27]. It can be seen
that a diffusion controlled reaction transforms to a Nucleation and Growth controlled
reaction as the milling time increases and particle size is reduced to nano level [1]. It
has also been observed that addition of Ni to Mg as a catalyst (Mg+2wt%Ni) reduces
the activation energy by 50 kJ/mol and enables faster kinetics of dehydrogenation [28].

14
 Figure 3: Fritsch Pulverisette 7 Ball Mill [29]
2.3 Kinetics

Numerous kinetic studies have been presented on gas–solid reactions of powder

metallic samples. Kinetics is a study of the possible reaction mechanisms; the rate of
the reaction is determined by the slowest process and is termed as the RLS. This
hydrogen sorption kinetics is improved by increasing the specific surface area of the
metal. Thus kinetics of Mg particles can be improved by particle size reduction to a
nanoscale using ball milling [23,30,31].

The contribution of different RLS can be inferred from the controlling process in
various segments. The controlling process located at different segments where the
reaction occurs within the Mg particle are: [32]

1) Interface process (Hydride metal boundary)

2) Transport controlled ( product film)
3) Surface controlled (gas-hydride boundary)

The above controlling process signifies different RLSs which can be

represented mathematically. The different phenomena and their mathematical
representation are as follows [10] [35].

15
Table 1: Set of reaction models to describe the reaction kinetics from the
literatures [33]

Mechanism Equations Phenomena

One dimensional diffusion 𝛼 2 = 𝑘𝑡 Diffusion

Three dimensional [1 − (1 − 𝛼)1/3 ]2 = 𝑘𝑡 Diffusion
diffusion
Second order Kinetics [1 − (1 − 𝛼)1/3 ] = 𝑘𝑡 Chemical Reaction
(Chou model)
Two dimensional (1 − 𝛼) ln(1 − 𝛼) + 𝛼 = 𝑘𝑡 Diffusion
diffusion
(bi dimensional particle
shape)
Three dimensional 2𝛼 2 Diffusion
1− − (1 − 𝛼)3 = 𝑘𝑡
diffusion 3

(Gisling–Braunshtein
equation)
First order kinetics − ln(1 − 𝛼) = 𝑘𝑡 Chemical Reaction
Two dimensional phase- 1 − (1 − 𝛼)1/2 = 𝑘𝑡 Chemisorption
boundary reaction
Three dimensional phase- 1 − (1 − 𝛼)1/3 = 𝑘𝑡 Chemisorption
boundary reaction
Zero order 𝛼 = 𝑘𝑡 Chemisorption
Random nucleation and [− ln(1 − 𝛼)]1/2 = 𝑘𝑡 Nucleation & Growth
growth of nuclei (Avrami –
Erofeev equation)
Random nucleation and [− ln(1 − 𝛼)]1/3 = 𝑘𝑡 Nucleation & Growth
growth of nuclei (Avrami –
Erofeev equation)
Phase boundary controlled [1 − (1 − 𝛼)(1−𝑛) ] Chemisorption
= 𝑘𝑡
reaction 1−𝑛
(n – 0,1/2 and 2/3)

16
 Three dimensional [1 − (1 − 𝛼)1/3 ]2 = 𝑘𝑙𝑛𝑡 Diffusion
diffusion
(Kroger –Ziegler equation)
Three dimensional 1 Diffusion
[ 1 − 1]2 = 𝑘𝑡
diffusion (1 − 𝛼)3
(modified Jander relation)

The slowest reaction step is considered as the RLS and thus for any general reaction
mechanism there is always a single RLS [10,33–37]. The slowest step results in the
highest resistance offered during the reaction. It is easier to assume that the rate of
hydrogen absorption is actually controlled by only one of the above mechanisms
presented in the Table 1. But for a practical hydrogen storage system, the situation is
more complex than expected [10]. The sequence of RLS can be explained as below:
diffusion of H2 through the gas layer surrounding Mg takes place fast and termed as the
surface phenomena. Hydrogen molecules get physically adsorbed by Van der Waals
forces on the surface of Mg (Physisorption). Hydrogen molecules further dissociate to
form single atoms of hydrogen and get chemisorbed onto the surface of the Mg metal
(Chemisorption). Further, the H atoms nucleate to form MgH2 molecules. Nucleation
is followed by the growth of the hydride layer into the α-Mg phase. Nucleation and
growth continues to be the dominant kinetic phenomena until the surface is completely
covered with hydride layer (β phase of MgH2) following which, diffusion is the
controlling phenomena for the hydriding of Mg [5, 6].

From the sequential steps of hydrogenation it can be possible that for a system there
might exist more than one controlling steps in the whole process. Jung et.al suggested
the initial stage of hydrogenation by chemical reaction which is later followed by
diffusion [10,37]. Therefore, the RLSs are sequential in nature, and there might be a
possibility that RLSs change with respect to the time of the reaction [37,38] [35,36].
Jung et. al suggested that the rate kinetics changed with time intervals and during the
change from one limiting step to the other there exists an intermediate interval which is
a linear combination of the two RLSs of the preceding and succeeding interval [37].
They observed that the RLS changed as the reaction proceeds. In the initial step in the
initial step in which the reacted fraction was less than 0.05, the rate controlling step was
the chemical reaction of the hydrogen atom with α phase hydride at the interface

17
between α and β hydrides, while the rate in the intermediate step was influenced by
both the chemical reaction and the diffusion of hydrogen atom through the β phase
hydride. In the later step in which the reacted fraction exceeds the range from 0.36 to
0.5, the rate control step was the diffusion of the hydrogen atom.

Chou et. al studied the possible RLSs as mentioned above in the Table 1 and
considered the rate kinetics to follow a shrinking core model and particles with a
uniform size distribution and shape. Chou et al had considered two RLS together but
the contribution from both the RLS was considered to be equal throughout the
hydrogenation time interval. They had considered diffusion phenomena with the
chemical reaction (i.e. Nucleation and Growth). When compared with their fitting
results they found that as the temperature increases the rate controlling step changes
from mixed diffusion and chemical reaction to single diffusion. They concluded that
temperature improves the kinetics and also changes the RLS [10,33,34].

Bloch et al mentioned that the RLS may be located at each of several possibilities [32]:

1. Surface related process (dissociation, surface penetration) for non-adherent

hydride film with effective constant thickness.
2. Transport (diffusion control) process through the protective hydride layer
with effective constant thickness.
3. Interface related process either by diffusion of H2 atoms through the
interface barrier.

JMAK (Johnson-Mehl-Avrami-Kolmogorov) equation for Nucleation and Growth is

used to describe the hydriding kinetics by various researchers [39–41]. The whole of
hydrogenation curve is fitted using equation (1).

𝐟 = 𝟏 − 𝒆𝒙𝒑(−𝐊𝒕𝒏 ) ………………………………………… (1)

Where, f is the volume fraction of magnesium hydride

K is the reaction constant which is = K0 exp (-Q/RT), where Q is the activation energy,
R universal gas constant, and T isothermal temperature.

K depends on nucleation rate and the growth rate.

This equation was proposed to describe the Nucleation and growth of the
hydriding layer. It has similarly been used by Tien et.al. In their work, they have fit the

18
functional based on the linear fit of JMAK as logarithmic plot (Figure 12) according to
the value of R2. The point up to which K and n value gave the R2 as 0.99 was considered
as a segment and in this way three segments were considered for the whole curve in
logarithmic plot. Each segment has its own set of K, n values. These values do not show
any physical significance and are purely empirical.

Thus JMAK alone however, cannot explain the sequential pattern of physical
phenomena. Also none of the equations mentioned above in the Table 1 explicitly
incorporates the size parameter in their existing equation. Particle size is important as
most of the well-established equations assume uniform particle size. But in any powder
sample we have particle size distribution. And so this distribution results to contributing
RLSs.

From the above literature review, the following points can be inferred:

1. Presence of particle size distribution may suggest that more than one RLS
govern the overall kinetics of the reaction.

2. Addition of catalyst also increases the reaction rate and decreases the
hydrogenation time. There may be effect of catalyst on the RLS.

3. Possible high level of structural defects introduced by ball milling treatment

may contribute to the change of RLS.

4. The RLSs can be considered to be contributing sequentially. However more

than one RLS may also act in parallel with each other.

5. The effect of particle size distributions on the RLSs in various time intervals.

With the above points inferred, we can conclude that a mathematical approach to
describe the kinetics has to be proposed. Šesták & Berggren [42,43] represented an
empirical formula where by changing the values of the random variables, different RLS
can be presented.

𝒅𝜶
= 𝒌𝜶𝒎 (𝟏 − 𝜶)𝒏 (− 𝐥𝐧(𝟏 − 𝜶))𝒑 ………………………………………. (2)
𝒅𝒕

Where, α = reacted fraction of reactant; t = time; m, n, p = random variables

The terms in the R.H.S. of the above equation show the change in trend
of the curve and the indices represent the contributions from the various terms in the

19
R.H.S. to the function in the L.H.S. The above equation is just an empirical formula
which does not give any physical meaning to the random variables m, n, p. The

With the same concept as the empirical equation (1) of Šesták [42,43],a hypothesis is
proposed accounting different RLS in a time interval and the fractional contribution of
each RLS to the overall RLS. Tien’s data [44,45] extensively used to propose the
mathematical approach.

20
 3. EXPERIMENTATION

Ball milling is the most common method in the mechanical processing of alloy powders
which reduces the size and changes the shape. Ball milling together with reducing the
size to few nanometres also increases the surface area [12–14,46]. In order to find out
how the particle size varies in a given ball milled sample, the as received MgH2 particles
were ball milled for certain time intervals. The elemental analysis was performed by
taking the microscopic images using Field Emission Gun-Scanning Electron
Microscope (FEGSEM). The X-Ray Diffraction was also performed to calculate the
crystallite size of the different samples.

3.1 Ball Milling Procedure

Magnesium Hydride (98%) was purchased from Alfa Aesar. Pure MgH2 was milled in
a Zirconia vial with zirconia balls to reduce the contamination from the milling tools.
Milling was done in a Planetary Ball Mill (Fritsch Pulverisette 7 Premium Line). 5 g of
sample was taken in a Zirconia vial of 45ml. Zirconia balls (18 nos.) of 10 mm diameter
were used with a sample of 5 grams of MgH2. Ball to powder ratio was about 10:1. The
ball mill was rotated about in 850 rpm. To avoid heating of the vial, 30 minutes
relaxation time was given after every 30 minutes of ball milling.

The sample was ball milled for different time intervals; 30 min, 1 hour, 2 hours and 4
hours. At each time interval some amount of ball milled sample was collected.
Together with the as received powder, the ball milled samples collected after different
time intervals, the elemental analysis was performed using FEGSEM JEOL JSM-
7600F. Figure 4 shows the microscopic images of as received and ball milled samples
obtained after different milling time. Using the Image J Software, the data was
compiled of the different particle mean tangent diameter of different particles within a
microscopic image. Using these different particle sizes in a sample, the particle
(volume %) v/s particle size was calculated. Sahagian et. al has described about a
technique by which the 3D particle size distributions can be obtained by 2D
observations ( microscopic images) [47]. The technique is based on knowing the
probability distribution of random cross-sections through various particles so that
‘small circle’ cross-sections can be subtracted from an observed population to provide

21
the 3D size distribution of particles. The results indicate that the most important
parameter controlling calculated size distribution is particle aspect ratio. For a
distribution of particles with a specific aspect ratio or range of aspect ratios, variations
of particle form (spherical vs. cubic; rectangular vs. elliptical) do not alter the results,
so the technique can be applied to a range of particle shapes. The particle size
distribution was determined by dividing the different mean tangent diameter into
linear classes on the basis of probability and multiplying with a shape factor. After
calculating total volume of the particles of a given sample, the particles on the basis of
their particle sizes calculated was binned to determine the distribution of particles
over the volume of the sample. The average particle size was determined from the
binned particle sizes.

The structural characterisation of the sample was carried with X-Ray Diffractometer
(Rigaku Smart Lab) using Cu-Kα radiation. The diffraction patterns were analysed by
Integrated X-Ray Powder Diffraction Software (PDXL). The Scherer equation was
used to determine the size of particles of crystals in the form of powder. The crystallite
size of the powder sample was calculated using this equation [48].

………………………………………. (3)
Where

 τ is the mean size of the ordered (crystalline) domains, which may be smaller or
equal to the grain size (crystallite size);
 K is a dimensionless shape factor, with a value close to unity. The shape factor
has a typical value of about 0.9, but varies with the actual shape of the crystallite;
 λ is the X-ray wavelength;
 β is the line broadening at half the maximum intensity after subtracting the
instrumental line broadening, in radians. This quantity is also sometimes denoted
as Δ(2θ);
 θ is the Bragg angle

XRD analysis gave the crystallite size of the different samples while the elemental
analysis helped in finding out the average particle size of a given sample. The particle
size distribution of the samples was also analysed.

22
 (a) (b)

(c) (d)

(e)
Figure 4: SEM images of Magnesium Hydride in (a) As received condition and
after ball milling for (b) 30 min (c) 1 hr (d) 2 hr (e) 4 hr (magnified to 50X).

23
The particle size distributions for the different ball milled samples are as follows

Particle Size Distribution (Vol.%) 35

30
As Received
25 Avg. Particle Size 106 ± 41 μm
20

15

10

0
0 34 47 51 57 66 81 82 99 115 155 164 169
Particle Size (μm)

Figure 5: Particle size distribution of As Received MgH2.

20
Particle Size Distribution (Vol.%)

18
16 30 minutes ball milling
Avg Particle Size 82 ± 31μm
14
12
10
8
6
4
2
0
27 41 44 50 54 62 68 75 89 90 107 122 128
Particle Size (μm)

Figure 6: Particle size distribution of 30 minutes ball milled sample

24
 0.18

Particle Size Distribution (Vol.%) 0.16 1 hour ball milling

0.14 Avg. Particle Size 84 ± 32 μm
0.12
0.1
0.08
0.06
0.04
0.02
0
31 44 47 54 60 68 69 79 87 104 118 127 133
Particle Size (μm)

Figure 7: Particle size distribution of 1 hour ball milled sample

12
Particle Size Distribution (Vol.%)

10 2 hours ball millimg

Avg. Particle Size 42 ± 42 μm
8

0
11 14 15 16 19 21 23 25 29 33 36 42 44 51 56 65 68 81 86 109 124 126 170
Particle Size (μm)

Figure 8: Particle size distribution of 2 hours ball milled sample

25
 18

Particle Size Distribution (Vol.%) 16

4 hours ball milling
14
Avg. Particle Size 29 ± 11 μm
12
10
8
6
4
2
0
11 14 15 16 19 21 23 26 29 34 36 37 49 50
Particle size (μm)

Figure 9: Particle size distribution of 4 hours ball milled sample

The average particle sizes of the different samples from the above size distribution
can be plotted with respect to the milling time.

120

100
Avg. Particle size (μm)

80

60

40

20

0
0 1 2 3 4 5
MillingTime (hours)

Figure 10: Average particle size of sample v/s milling time (hours)

Figure 5 -9 gives the plot of particle size distribution of as received and ball milled
sample for different time intervals. There is wide range of particle sizes when sample
is milled more compared to a narrow range of particle sizes for as received sample. It
can be inferred from the figure 10 that as the milling time is increased, the average
particle size decreases.

26
 -MgH2 - Mg - MgO

Figure 11: X-Ray Spectra of the mechanically milled MgH2 for various time
intervals and pure as received MgH2

9
8
7
Crystallite Size (μm)

6
5
4
3
2
1
0
0 1 2 3 4 5
Milling Time (hours)

Figure 12: Crystallite size of samples v/s milling time

27
The X –Ray Diffraction Spectra of the MgH2 mechanically milled for various times are
shown in figure 11. The as received sample, the peaks of MgH2 are present. As the
sample is ball milled, the peak size decreases. The small peak now corresponds to Mg.
During ball milling, some amount of hydrogen is desorbed from the MgH2 which results
to the formation of Mg. Also some traces of oxygen remain in the vial, which reacts
with the Mg and hence traces of MgO are also found. There are small peaks
corresponding to MgO in the 4 hours ball milled sample.

Figure 12 gives the plot of crystallite size v/s the milling time in hours. The crystallite
size calculated using the Scherer equation using the above XRD data signifies that as
the milling time is increased the crystallite size decreases.

28
 4. KINETIC ANALYSIS
Absorption kinetics of magnesium hydride is improved considerably by the intensive
ball milling. This improved kinetics is explained by the features introduced during ball
milling like: small particle size, increased specific surface area, and increase in
nucleation site density and reduced diffusion length.

A single RLS cannot describe the whole kinetics of the hydrogenation curve where
different physical phenomena take place. This may lead to that more than one RLS may
contribute to the overall RLS.

A general mechanism can be suggested as the sequential steps of hydriding [10] ;

1) H2 in the bulk gas transfer to the metal surface.

2) H2 diffusion through the boundary layer between gas phase and solid particle
3) Physisorption of H2 on the solid surface.
4) Dissociation of H2 molecules and Chemisorption
5) Surface penetration of H2 atoms
6) Diffusion of H2 atoms through the hydride product layer to metal interface
7) Chemical reaction and nucleus formation producing hydride.

Physisorption, gas layer diffusion of H2 molecules and diffusion of H atoms through

the Mg layer are found to be instantaneous and hence not considered as RLS.
Chemisorption, Nucleation and Growth and Diffusion physical phenomena are
more dominant among all the other sequential steps taking place in the
hydrogenation of Mg. And hence they are considered as the sequential RLSs.

→ →

Chemisorption Nucleation−Growth (N−G) Diffusion

Figure 13: The schematic diagram explaining overall mechanism in sequence [44]

Chemisorption involves the chemical bonding between the diffused H atoms and Mg
atoms at the surface of the particle (Surface phenomena). The forces of attraction are

29
chemical bond between the atoms. The Nucleation and Growth involves the formation
of MgH2 nuclei and the progressive growth of the nuclei towards the core of the Mg
particle. When Growth of MgH2 results to the substantial formation of layer of hydride,
the diffusion of H atoms through this hydride layer is difficult and thus Diffusion
attributes to the transport phenomena of the H atoms diffusing through the hydride layer
and reacting with the core Mg atoms.

Relative contribution from each process to rate of product formation may alter with the
progress of reaction. Early stages of hydrogenation may be limited to adsorption/
transport of H2 through protective oxidation over layer (Physisorption/ Chemisorption)
[33]. As hydride layer grows reaction rate may be controlled by the interface process
(Nucleation and Growth) and as reaction develops furthermore, the Rate of reaction
may change to transport of H2 through the hydride film (Diffusion). Thus the
mechanism of such gas-metal reaction may be altered during course of reaction and the
RLS may change accordingly and different RLS related to certain limited stage of the
reaction [34].

This implies that one RLS may change to another during the course of reaction and with
respect to time.

In short all the above steps can be summarised as the following steps

Physisorption/Chemisorption Nucleation and Growth Diffusion

All the three RLS may occur simultaneously also to contribute fractionally to the
overall RLS. The changing RLS is considered as one of the important factor to develop
the mathematical approach. The other factor considered is the contribution of the RLSs
to the overall RLS.

30
 4.1 Developing the mathematical approach

Tien et. al carried out the hydrogenation experiment of Mg at 210 ºC and at 1 MPa of
pressure [40,44,45].

0.07

0.06
Hydrogen uptake (wt. fraction)

0.05

0.04

0.03

0.02
Experiment

0.01

0
0 50 100 150 200 250 300
Time (minutes)

Figure 14: Hydrogenation curves of magnesium powder sample at 210°C under

1MPa (α v/s time) [44]
(α is the weight fraction)

According to Tien et. al, the formation of magnesium hydride was explained by the
Nucleation and Growth model. Hence they considered Nucleation and Growth to be the
RLS and the curve was fitted using the JMAK equation.

The JMAK (Johnson-Mehl-Avrami-Komalogrov) equation; the relation between

hydride fraction and hydrogenation time is as follows:
𝐟 = 𝟏 − 𝒆𝒙𝒑(−𝐊𝒕𝒏 ) …………………………………………………………… (4)

f is the volume fraction of magnesium hydride.

31
K is the reaction constant which is = K0 exp (-Q/RT) …. (Q is the activation energy, R
ideal Gas constant, and T Isothermal temperature)

K depends on nucleation rate and the growth rate.

Figure 15: The JMAK analysis of powders hydrogenated at 210°C under 1MPa in
logarithmic plot [44]
The above equation (4) can be written as:

(1-f) = exp (-Ktn)

ln (1-f) = -Ktn

ln {ln [1/ (1-f)]} = ln K + n ln t …………………………………………………(5)

The plot of ln {ln [1/ (1-f)]} v/s ln t is shown in figure 15 is the JMAK analysed of the
curve of figure 14.

The Figure 15 divided into three linear stages including a transition region between
stage 2 and stage 3. The stages are divided on the basis of the best fit of the JMAK
equation. Tien analysed the ln {ln [1/ (1-f)]} v/s ln t plot according to the value of R2
(which is the statistical measure of how well regression approximates; R2 of 1 indicates
a perfect fit). The time range for each stage was defined based on the best fit analysis.

32
The range was selected when the R2 for each stage was equal or larger than 0.995. In
this way three stages were considered. The division of the plot into three stages do not
signify any physical phenomena.

So stage 1 may possibly be associated with the dissociation of hydrogen molecules.

Stage 2 may be associated with the growth of the magnesium hydride considering
nucleation rate constant and independent of time. Transition stage between stage 2 and
3. After the end of transition stage, impingement is complete and further growth in stage
3 may be through diffusion.

The values of K and n change in accordance with the stage considered. (The numerical
values do not signify the presence or absence any particular physical phenomena)
Change in the numerical values of the constant K and n does not signify any change of
RLS with respect to time. There exists a transition region which does not follow the
linear trend as the other stages. This deviation from the fit may occur due to the possible
existence of more than one rate limiting phenomenon.

4.2 Reanalysing the logarithmic plot

The logarithmic plot was reanalysed by us and the JMAK analysis logarithmic plot
(Figure 15) was converted to the hydrogenation curve (α v/s time) using the values of
K and n for different stages from the above plot.

When compared with the experimental curve of α v/s time (Figure 14) we found out
that the JMAK logarithmic plot converted to α v/s time plot does not matche with the
experimental curve (Figure 16).

The discrepancy of the JMAK plot was diminished in the logarithmic plot, which
therefore gave best fit while when the same JMAK is used to plot α v/s time domain,
the discrepancy showed up. Therefore when figure 15 was plotted with the experimental
curve using the JMAK equation and using the K and n values from the different stages,
the discrepancies showed up. This can be a major disadvantage of the JMAK equation
in logarithmic form, which gave a best fit and linear trend in logarithmic domain, but
could not fit the curve in α v/s time domain. The end segments of the stages in the
logarithmic plot also deviated from the linear trend. This can also signify that a single
RLS cannot explain the whole hydrogenation curve.

33
 0.07

0.06
Hydrogen uptake (wt. fraction)

0.05

0.04

0.03
Nucleation-Growth Experiment

0.02

0.01

0
0 50 100 150 200 250 300
Time (minutes)

Figure 16: JMAK (nucleation and growth) and Experimental α v/s time plots

4.3 Reanalysis of existing data for 44 μm particles

In order to understand the incorporation of different RLS to the overall RLS, the
microscopic images at different hydrogenation time intervals were observed again. The
experimental results from Tien’s thesis were re-analysed, to explain the probable
existence of more than one RLS. The assumption of JMAK equation considering as the
single RLS does not go accordingly with the microscopic images taken in different time
intervals. The stages are not divided with accordance to the microscopic images. Even
though a mathematical equation may fit the curve to some extent but the physical
meaning of the microscopic images and its relation with the hydrogenation curve is
side-lined. The microscopic images of magnesium hydride after different times of the
hydrogenation experiment were analysed. The presence of a particular RLS or a

34
combination of RLS and the effect of the changing RLS was inferred by observing the
microscopic images again.

(a)

(b)

(c)

35
 (d)

(e)

Figure 17: The cross section of powder hydrogenated for different time intervals
(microscopic SEM images)

(a) 2.5 minutes (b) 7.5 minutes (c) 15 minutes (d) 75 minutes (e) 300 minutes [44]

36
Figure 18: Hydrogenation curves of magnesium powder sample at 210°C under
1MPa hydrogen pressure at each interrupted times [44]

Table 2: Qualitative observations from microscopic images

Time Un- Nucleation & Nucleation Impingeme Complete Probable

(min) Reacted Growth near & Growth nt coverage Rate
Surface Surface Internally Limiting
Phenomen
a
2.5 Very large A few nuclei few Very less No Nucleation-
Growth
7.5 Large A few more Extremely less No Nucleation-
nuclei is few Growth
observed
15 Lesser Growth is Extremely More No Nucleation-
higher few Growth &
Diffusion
75 Lesser More growth None large extent No Nucleation-
than 15 than previous Growth &
min Diffusion
300 Complete No growth None large extent Yes Diffusion

37
When the microscopic images were reanalysed it was found that in the later stages of
hydrogenation around 300 minutes, Diffusion is predominant rather than Nucleation
and Growth. Chemisorption is instantaneous and hence cannot be observed in the
microscopic images. It occurs very fast before Nucleation and Growth step. There is a
possibility of change in RLS. Thus we can state from the above observation that initially
Chemisorption takes place which is overtaken by Nucleation and Growth which is later
overtaken by Diffusion.

The sequential order of RLS can be summarised as:

Physisorption/Chemisorption Nucleation and Growth Diffusion

The microscopic images give us an idea that when the dominance of one rate limiting
phenomena reduces, the other rate limiting phenomena takes over the dominance.
Considering all the above conditions and the observation of the microscopic images a
hypothetical thesis ( equation 1) similar to the empirical equation proposed by Šesták
& Berggren[42,43] was developed; A mathematical approach which defines the
dominance of a particular RLS for different time segments.

4.4 Hypothesis formulation

A hypothetical theory has been proposed considering the physical phenomena observed
from the microscopic images (figure 17) and the mathematical representation of the
different phenomena clubbed together. The theory was put forward in accordance to the
changing RLS of magnesium undergoing hydrogenation reaction.

The equation is as follows:

𝑭 = 𝑲(𝑭𝑪𝒉𝒆𝒎 )𝒑 (𝑭𝑵−𝑮 )𝒒 (𝑭𝑫𝒊 𝒇𝒇 )𝒓 ……………………………………………….. (6)

Where p + q + r = 1 ………………………………………………………………. (7)

Equation 6 represents that α (fractional conversion) is a function of the contributions

from individual RLSs. Fchem, FN-G, FDiff are the individual functions representing the
physical phenomena of Chemisorption, Nucleation and Growth and Diffusion
respectively. These functions contribute fractionally to the overall function F. These
fractional contributions are represented by the variables p, q, and r for Chemisorption,

38
Nucleation and Growth and Diffusion respectively. The fractional contribution equals
to 1 (equation 7). The variables p, q, and r give extent of a phenomenon in a particular
time interval. The indices p, q, r indicate the dominance of the respective rate limiting
phenomena in a given time segment. With the change in the time segment, the values
of p, q, and r also may change if there is a change in the dominance of the rate limiting
phenomena. The change in trend of the values of p, q, and r will give the change in
dominance of the RLS.

Observing the microscopic images (figure 17) of different time intervals for 44 μm
particles, it was inferred that the order of change in RLS will be.

Physisorption/Chemisorption Nucleation and Growth Diffusion.

0.8

0.6
q,r

N-G
0.4
Diff
0.2

0
0 100 200 300 400
time (minutes)

Figure 19: Expected trend of q and r from the qualitative observation of the
microscopic images
As observed the microscopic images the Nucleation and Growth phenomena goes on
decreasing while steadily the diffusion phenomena catches up. The above expected plot
of q, r v/s time is plotted in accordance to the observation of the microscopic images of
the 44 μm hydrogenated particles at various time intervals.

When all the 3 functions are considered together, there is a mathematical limitation and
hence trivial solution exists. In order to solve and simplify the equation, function
considers only two steps at a time instead of 3 as said earlier.The function is split into
two parts considering two phenomena at a time. The necessary conditions are: the sum
of the indices of the individual functions are equal to 1 and the constant of

39
proportionality (K) tends to 1 .These two conditions indicate that only two individual
function are best conditions to the limitation imposed for indicating the overall RLS.

The equations are as follows with the conditions:

1) 𝑭𝟏 = 𝑲𝟏 (𝑭𝑪𝒉𝒆𝒎 )𝒑 (𝑭𝑵−𝑮 )𝒒 ; p+q=1 ……………………………… (8)

2) 𝑭𝟐 = 𝑲𝟐 (𝑭𝑵−𝑮 )𝒒 (𝑭𝑫𝒊𝒇𝒇 )𝒓 ; q+r=1 ………………………………. (9)

Chemisorption is a surface phenomenon and occurs instantaneously. Hence it can be

seen only in the initial stages. In the later stages when hydride layer is formed there is
no significance of Chemisorption. At later stages of hydrogenation when continuous
hydride layer is formed, Diffusion is the predominant RLS. Therefore Chemisorption
and Nucleation and Growth are considered as the two physical phenomena for the initial
stages (F1 equation 8) and Nucleation and Growth with Diffusion are considered as the
two physical phenomena for later stages (F2 equation 9). This division is in accordance
with the observation of the microscopic images of the 44 μm particles [44].

40
 5. RESULTS AND DISCUSSION

5.1 Application to different Magnesium Powders with different particle sizes

The proposed mathematical approach considers the contribution of three physical
phenomena to the overall RLS. This approach was applied to three different
hydrogenation curve having different particle sizes. The analysis are as follows.

5.1.1 Micron size particles (44 μm)

As discussed in the Tien’s Thesis, the as received magnesium hydride particles were
used to hydrogenate in 210 ºC temperature and 1 MPa pressure. A small amount of
nickel (2 wt. %) was added as a catalyst to enhance the rate of reaction of
hydrogenation. The average particle size is about 44 μm [44].

The particle size distribution of the as received magnesium hydride is as follows:

9
Particle Size Distribution (Vol. %)

8 Avg. Particle Size 44 ± 22 μm

7
6
5
4
3
2
1
0
1 12 19 30 48 77 122
Particle Size (μm)

Figure 20: Particle size distribution of As Received Mg particles

The particle size distribution gives a fair idea that with a wide range of different particle
sizes within a sample. This size distribution can be a reason for more than one RLS
contributing to overall RLS.

The hydrogenation curve gives the weight fraction of H2 absorbed with respect to time
of the reaction. The experimental plot of hydrogenation is as represented in Figure 21.

41
 0.07

0.06
Hydrogen uptake (wt. fraction)

0.05

0.04

0.03

0.02
Experiment

0.01

0
0 50 100 150 200 250 300
Time (minutes)

Figure 21: Hydrogenation curves of magnesium powder sample at 210°C under

1MPa (44 μm)
Using the equations

𝑭𝟏 = 𝑲𝟏 (𝑭𝑪𝒉𝒆𝒎 )𝒑 (𝑭𝑵−𝑮 )𝒒 …………………… (8)

and 𝑭𝟐 = 𝑲𝟐 (𝑭𝑵−𝑮 )𝒒 (𝑭𝑫𝒊𝒇𝒇 )𝒓 ……………………… (9)

The hydrogenation curve (figure 22) is fitted. Till 6 minutes the equation 8 which
involves the phenomena chemisorption and Nucleation and Growth fits the curve.
Beyond 6 minutes chemisorption phenomena diminishes and equation 9 involving
Nucleation and Growth and Diffusion phenomena fits the curve (figure 21).

The time segments is about 1 minute interval in the chemisorption and Nucleation &
Growth phenomena (equation 8), while about 15 minutes interval for the Nucleation &
Growth and Diffusion phenomena (equation 9). Fitting the equations for different time
segments, the hydrogenation curve fit is satisfactory (figure 22).

42
 0.07
44 μm
0.06
Experiment
Hydrogen uptake (wt. fraction)

0-1min
0.05 1-2 min
2-4 min
4-5 min
0.04 5-6 min
6-15min
15-30 min
0.03 30-45 min
45-60 min
60-75 min
75-90 min
0.02
90-105 min
105-120 min
120-180 min
0.01 180-240 min
240-300 min

0
0 50 100 150 200 250 300
Time (minutes)

Figure 22: Hydrogenation curve (figure 23) fitted with the equations for different
time segments

0.8

0.6
p, q, r

q NG
0.4 r
p
q chem
0.2

0
0 50 100 150 200 250 300 350
Time (minutes)

Figure 23: The trend of p, q and r with respect to time for hydrogenation of 44 μm
particles

43
 1

0.8

0.6
p, q

p q
0.4

0.2

0
0 1 2 3 4 5 6 7
Time (minutes)

Figure 24: The trend of p and q with respect to time in the time interval 0-6
minutes

The trend of p, q and r v/s time is similar to the trend of q, r v/s time (figure19) inferred
from the observation of the microscopic images (figure 17) observed from the for
hydrogenation kinetics of 44 μm Mg particles. With the change in time segments, the
value of p, q and r also changes. In the initial stage Chemisorption is dominant which
is indicated by the high value of p. As reaction proceeds the value of p decreases. This
signifies that the dominance of Chemisorption as the RLS is reduces as reaction time
proceeds. After around 6 minutes of hydrogenation, Chemisorption is almost negligible
as the surface is almost hydride (figure 26). In the meantime nucleation and growth step
catches up and dominates from around 3-4 minutes from chemisorption. The value of
q steadily increases from a negligible amount to a value higher than p as observed from
the microscopic images (figure 17 b). This indicates that the dominance of Nucleation
and Growth over Chemisorption phenomena in the kinetics. All the three physical
phenomena cannot be considered together (mathematical limitation) and hence the next
2 consecutive phenomena Nucleation & Growth and Diffusion are considered. As the
reaction time progresses q steadily decreases and r increases as observed in the
microscopic images (figure 17 c). This signifies that Diffusion takes the dominance
over Nucleation and Growth. The crossover of Nucleation and Growth to Diffusion is
around 35-40 minutes. Diffusion starts as soon as there is a significant thickness of

44
hydride layer and hence catches up early. This can be inferred from the microscopic
images (figure 17 c). This indicates the transport of hydrogen molecules through the
hydride layer. Being the slowest amongst all it takes over its dominance as RLS by
around 37 minutes.

5.1.2Particle Size of 10-35 nm

The same above analysis is also done for a particle size of about 10-35 nm. The Mg
particles was catalysed by about 0.75 wt. % Nickel powder and hydrogenated [39].

The experimental plot of hydrogenation is as follows:

0.06

0.05
Hydrogen uptake (wt. fraction)

0.04

0.03
experiment

0.02

0.01

0
0 50 100 150 200
Time (min)

Figure 25: Hydrogenation curves of magnesium powder sample at 200°C under

2MPa (10-35 nm)
Using the equations

𝑭𝟏 = 𝑲𝟏 (𝑭𝑪𝒉𝒆𝒎 )𝒑 (𝑭𝑵−𝑮 )𝒒 …………………… (8)

and 𝑭𝟐 = 𝑲𝟐 (𝑭𝑵−𝑮 )𝒒 (𝑭𝑫𝒊𝒇𝒇 )𝒓 ……………………… (9)

45
The hydrogenation curve (figure 25) is fitted. Till 4 minutes the equation 8 which
involves the phenomena chemisorption and Nucleation and Growth fits the curve.
Beyond 4 minutes chemisorption phenomena diminishes and equation 11 involving
Nucleation and Growth and Diffusion phenomena fits the curve (figure 25).

The time segments is about 0.6 minute interval in the chemisorption and Nucleation &
Growth phenomena (equation 8), while about 10 minutes interval for the Nucleation &
Growth and Diffusion phenomena (equation 9). Fitting the equations for different time
segments, the hydrogenation curve is satisfactorily fitted (figure 26)

0.06
10-35 nm

0.05

experiment
0.2-0.4min
0.04
0.4-0.5min
.5-.6min
Reacted fraction

.6-1min
1-2min
0.03
2-4min
4-5min
6-10min
0.02 10-15min
15-35min
60-85min
85-115min
0.01
120-150min
150-180min

0
0 50 100 150 200
Time (minutes)

Figure 26: Hydrogenation curve (figure 25) fitted with the equations for different
time segments

46
 1

Trend of p,q,r
0.8

q N-G
0.6
r
p,q,r

0.4 p
q chem
0.2

0
0 50 100 150 200
Time (minutes)

Figure 27: The trend of p, q and r with respect to time for hydrogenation of 10-35
nm particles

1
Trend of p,q
0.8

0.6
p
p,q

0.4 q

0.2

0
0 2 4 6 8 10
Time (minutes)

Figure 28: The trend of p and q with respect to time in the time interval 0-6
minutes
The trend of p, q and r v/s time is similar to the trend of q, r v/s time (figure19) inferred
from the observation of the microscopic images (figure 17) observed from the for
hydrogenation kinetics of 10-35 nm Mg particles. It is also similar to the trend of p, q

47
and r for 44 μm Mg particles. The dominance of the RLSs are similar to the trend
followed by the 44 μm particles. With the change in time segments, the value of p, q
and r also changes. In the initial stage Chemisorption is dominant which is indicated by
the high value of p. As reaction proceeds the value of p goes on decreasing. This
signifies that the dominance of Chemisorption as the RLS is deteriorating as reaction
time proceeds. After around 4 minutes of hydrogenation, Chemisorption is almost
negligible because the surface is almost completely hydrided (figure 33). In between
Nucleation and Growth step catches up and dominates from around 2-3 minutes. The
value of q steadily increases from negligible amount to a value higher than p. This
indicates that the dominance of Nucleation and Growth over Chemisorption phenomena
in the kinetics.

All the three physical phenomena cannot be considered together (mathematical

limitation) and hence the next 2 consecutive phenomena Nucleation & Growth and
Diffusion are considered. As the reaction time progresses q steadily decreases and r
increases. This signifies that Diffusion takes the dominance over Nucleation and
Growth. The crossover of Nucleation and Growth to Diffusion is around 10-15 minutes.
Diffusion starts as soon as there is a significant thickness of hydride layer and hence
catches up early. This indicates the transport of hydrogen molecules through the hydride
layer. Being the slowest amongst all it takes over its dominance as RLS by around 12
minutes.

5.1.3 Particle Size of 15 nm

The above analysis was repeated for a particle size of about 15 nm. The MgH2 particles
were catalysed by about 5 wt. % Nickel powder and (de)hydrogenated around 270 ºC
temperature and 15 MPa [49].

The experimental plot of hydrogenation is as follows:

48
 0.06

0.05
Hydrogen uptake (wt. fraction)

0.04

0.03
experiment

0.02

0.01

0
0 10 20 30 40 50 60 70
Time (min)

Figure 29: Hydrogenation curves of magnesium powder sample at 270°C under

15 MPa (15 nm)
The diffusion equation does not give satisfactory fit the curve, so it was concluded that
diffusion may not occur during the course of hydrogenation. The probable reason of
diffusion not acting as a Rate limiting Phenomena is discussed further. It is observed
that Nucleation and Growth gives satisfactory fit for the curve. So this indicated that
only two physical phenomena governed the kinetics of 15 nm particles.

Using the equation

𝑭𝟏 = 𝑲𝟏 (𝑭𝑪𝒉𝒆𝒎 )𝒑 (𝑭𝑵−𝑮 )𝒒 …………………… (8)

The hydrogenation curve (figure 29) is fitted. The whole of the hydrogenation curve is
fitted using the two physical phenomena Chemisorption and Nucleation & Growth
(Figure 29). Fitting equation 8 for different time segments, the hydrogenation curve is
satisfactorily fitted (figure 30).

49
 0.06

15 nm
experiment
0.05 0-0.39min
0.39-0.5min
0.5-1min
0.04 1-2min
hydrogen wt fraction

2-3min
3-5min
0.03
5-10min
10-15 min
15-30 min
0.02
30-35 min
45-50 min
50-55 min
0.01
55-60min
35-40min
40-45min
0
0 10 20 30 40 50 60 70
time (minutes)

Figure 30: Hydrogenation curve (figure 29) fitted with the equation 8 for different
time segments

0.8 Trend of p,q

0.6 p
p,q

0.4 q

0.2

0
0 10 20 30 40 50 60 70
Time (minutes)

Figure 31: The trend of p and q with respect to time for hydrogenation of 15 nm
particles

50
The expected trend of p and q with respect to time (figure 19) is similar to the trend of
p and q (figure 39) for hydrogenation kinetics of 15 nm Mg particles. It is also similar
to the trend of p and q for 44 μm Mg particles. With the change in time segments, the
value of p and q also changes. As reaction proceeds the value of p goes on decreasing.
This signifies that the dominance of Chemisorption, as the RLS deteriorates and
diminishes as reaction time proceeds. After around 2.5 minutes of hydrogenation,
Chemisorption is almost negligible (figure 33). As dominance of Chemisorption falls,
Nucleation and Growth step catches up and dominates from around 2-3 minutes. The
value of q steadily increases from negligible amount to a value higher than p. This
indicates that the dominance of Nucleation and Growth over Chemisorption phenomena
in the kinetics.

The negligible contribution of Diffusion as a RLS can be due to the following reasons:

1) The particle size is greatly reduced and the surface area is increased
considerably. Therefore the conversion of Magnesium to Magnesium hydride is
almost complete by Chemisorption and Nucleation and Growth, as the diffusion
path lengths may be reduced considerably. Diffusion acts mainly when there is
some significant hydride layer thickness (path length).
2) Among all the cases discussed 15 nm particles are catalysed by 5 wt.% of Ni,
whereas the 44 μm particles were catalysed by 2 wt. % Ni and 10-35 nm
particles were catalysed by 0.75 wt.% Ni. The highest concentration of catalyst
in 15 nm may be facilitating the fastest kinetics.
3) Also, catalyst helps in altering the path of reaction, and hence Chemisorption
Step lasts longer in 15 nm particles.
4) This analysis gives the fair idea that if the particle size is reduced considerably
to nanometres together with considerable amount of catalyst, the contributing
RLSs can be reduced from the overall RLS.

51
5.2 Comparison of the p, q and r values
In order to compare the values of p, q and r for the different particle size, the
hydrogenation time was converted to t*.The maximum theoretical hydrogen uptake (wt.
fraction) is 0.076. In the following discussion, it is convenient to utilise the reduced
time t* defined as

𝑡
𝑡∗ = With t1/2 is the time required for 50% conversion; α1/2 =0.038.
𝑡1/2

Table 3: The t1/2 values for different particle sizes

Particle sizes t1/2 min

44 μm 22.74
10-35 nm 30.4
15 nm 3.89

The following figures give the comparison of p, q and r for different particle sizes with
respect to time.

0.8
p
0.7

0.6

0.5
35 nm
0.4
p

44 micron
0.3 15 nm

0.2

0.1

0
0 2 4 6 8 10 12 14 16 18
t*

Figure 32: p v/s t* for different particle sizes

52
 0.8
p
0.7

0.6

0.5
35 nm
0.4
p

44 micron

0.3 15 nm

0.2

0.1

0
0 1 2 t* 3 4 5

Figure 33: p v/s t* for different particle sizes for 0 -5 t* values

1.2

q Chem-NG
1

0.8
35 nm
0.6 44 micron
q

15 nm
0.4

0.2

0
0 2 4 6 8 t* 10 12 14 16 18

Figure 34: q (chemisorption – nucleation & growth) v/s t* for different particle
sizes

53
 1.2

1
q Chem-NG
0.8
35 nm
0.6
q

44 micron
15 nm
0.4

0.2

0
0 1 2 3 4 5
t*

Figure 35: q (chemisorption – nucleation & growth) v/s t* for different particle
sizes for 0 -5 t* values

0.8
q NG-Diff
0.7

0.6

0.5 35 nm
44 micron
0.4
q

0.3

0.2

0.1

0
0 2 4 6 t* 8 10 12 14

Figure 36: q (nucleation & growth – diffusion) v/s t* for different particle sizes

54
 1.2

r
1

0.8

0.6
r

35 nm
44 micron
0.4

0.2

0
0 2 4 6 t* 8 10 12 14

Figure 37: r v/s t* for different particle sizes

The following points can be inferred from the above results:

1) Almost similar trend is observed in different particle sizes when the p, q and r
are plotted with respect to t*.
2) The chemisorption step dominates as RLS in all the three cases for around 4
minutes and then gradually decreases. This is because it is a surface
phenomenon and the surface gets hydrided within the first few minutes.
3) The 15 nm particles due to reduced size and increased surface area follow
chemisorption and Nucleation and Growth phenomenon. The reaction is almost
complete with these two physical phenomena. Diffusion may not occur as the
particle is small and therefore the particle gets hydrided with Nucleation and
Growth phenomenon.
4) Also the amount of catalyst added to the Mg particles may alter the kinetics an
extent. The faster kinetics is observed when catalyst is added in optimum
quantity. [20].
5) p, q and r are the weighted fraction of the amount of hydride formed from a
particular RLS.
6) The amount of catalyst was about 0.75% for 10-35 nm, 2 % for 44 μm and about
5% for 15 nm. Therefore faster kinetics was observed in 15 nm particles
followed by 44 μm and 10-35 nm particles.

55
 7) Nano sizing of the particles with optimum addition of catalyst like Ni will
therefore help in increasing the kinetics of absorption and desorption of H2 in
magnesium hydride.
8) The kinetics of hydrogenation is a combination of different RLS which acts
simultaneously.

The following figure show the plot of hydrogen uptake (wt. fraction) α v/s t

0.07

0.06

0.05

0.04
Mg-2%Ni(44µm)
α

0.03
Mg-0.75%Nano
0.02 Ni(10-35nm)

Mg-5%Nano
0.01 Ni(15nm)

0
0 2 4 6 8 10 12 14 16 18
t*

Figure 38: α v/s t* for different particle sizes

It can be observed that α v/s t* for different particle size, the trend is similar, the amount
of hydrogen uptake depends on the particle size and the amount of catalyst added. With
no catalyst the uptake may be higher but the kinetics is slow. When catalyst is added
the reaction speeds up by altering the path of reaction mechanism. The addition of
catalyst may reduce the hydrogen uptake to some extent but not completely.

The hydrogen curves of all the three different particle sizes was satisfactorily fitted
using the new mathematical approach, where the contribution of different physical
phenomena are taken into account.

56
 6. CONCLUSION

A single RLS does not satisfactorily explain the whole of hydrogenation curve as there
is change of RLSs in the kinetics of hydrogenation. This is in accordance with the
literature. This problem has been attempted to be resolved using the new proposed
mathematical approach. Through this approach, we investigated the presence of more
than one RLS and the particle size effects on the sorption kinetics of hydrogenation of
magnesium. This approach accounted for the change dominance of the individual RLS
with the change in the specific time interval considered.

The ball milling of the MgH2 for different time intervals contributed to the particle size
distribution, which results in different RLSs as a result of non-uniformity in sizes of
particles in the samples considered. The trend of the contributing indices p, q and r for
Chemisorption, Nucleation & Growth and Diffusion respectively was almost similar.
All the three phenomenon was not studied together due to mathematical limitation. So
the region where these three phenomenon together may occur can be termed as a grey
area. Also in this mathematical approach the particle size was not incorporated. The
amount of catalyst also effects the kinetics.

It can be concluded that the overall RLS is a combined RLS where the fractional
contributions of different RLS is accounted. The particle size reduced by mechanical
ball milling contributes to particle size distribution which results in the different RLS.
The addition of a catalyst also helps in reducing the hydrogenation.

57
 7. REFERENCES

[1] K.-F. Aguey-Zinsou, J.-R. Ares-Fernández, Hydrogen in magnesium: new

perspectives toward functional stores, Energy Environ. Sci. 3 (2010) 526.
doi:10.1039/b921645f.

[2] R.A. Varin, T. Czujko, Z.S. Wronski, Z. Zaranski, NANOMATERIALS FOR

HYDROGEN STORAGE PRODUCED BY BALL MILLING, Can. Metall. Q.
48 (2009) 11–26.

[3] D.K. Ross, Hydrogen storage : The major technological barrier to the
development of hydrogen fuel cell cars, Vacuum. 80 (2006) 1084–1089.
doi:10.1016/j.vacuum.2006.03.030.

[4] I.P. Jain, C. Lal, A. Jain, Hydrogen storage in Mg : A most promising material,
Int. J. Hydrogen Energy. 35 (2010) 5133–5144.
doi:10.1016/j.ijhydene.2009.08.088.

[5] N.E. Tran, S.G. Lambrakos, M.A. Imam, Analyses of hydrogen sorption
kinetics and thermodynamics of magnesium – misch metal alloys, J. Alloys
Compd. 407 (2006) 240–248. doi:10.1016/j.jallcom.2005.07.007.

[6] E. Callini, L. Pasquini, T.R. Jensen, E. Bonetti, Hydrogen storage properties of

Mg–Ni nanoparticles, Int. J. Hydrogen Energy. 38 (2013) 12207–12212.
doi:10.1016/j.ijhydene.2013.05.139.

[7] M. Pourabdoli, S. Raygan, H. Abdizadeh, Determination of kinetic parameters

and hydrogen desorption characteristics of MgH 2 -10 wt % ( 9Ni e 2Mg e Y )
nano-composite, Int. J. Hydrogen Energy. 38 (2013) 11910–113919.
doi:10.1016/j.ijhydene.2013.06.133.

[8] S. Løken, J.K. Solberg, J.P. Maehlen, R. V Denys, M. V Lototsky, B.P.

Tarasov, et al., Nanostructured Mg – Mm – Ni hydrogen storage alloy :
Structure – properties relationship, J. Alloys Compd. 446-447 (2007) 114–120.
doi:10.1016/j.jallcom.2006.11.200.

[9] M.H. Mintz, Y. Zeiri, Hydriding kinetics of powders, J. Alloys Compd. 216
(1995) 159–175.

[10] G. Wu, J. Zhang, Q. Li, K. Chou, A new model to describe absorption kinetics
of Mg-based hydrogen storage alloys, Int. J. Hydrogen Energy. 36 (2011)
12923–12931. doi:10.1016/j.ijhydene.2011.07.026.

[11] WebElements Periodic Table of the Elements | Magnesium | magnesium

dihydride, (n.d.).

[12] A. Zaluska, L. Zaluski, Synergy of hydrogen sorption in ball-milled hydrides of

Mg and Mg 2 Ni ¨, J. Alloys Compd. 289 (1999) 197–206.

58
[13] L. Zaluski, A. Zaluska, J.O.S. Olsen, Hydrogen absorption in nanocrystalline
Mg2Ni formed by mechanical alloying, J. Alloys Compd. 217 (1995) 245–249.

[14] L. Zaluski, A. Zaluska, Nanocrystalline metal hydrides ¨, J. Alloys Compd. 254

(1997) 70–79.

[15] B. Chen, C. Kuo, J. Ku, P. Yan, C. Huang, M. Jeng, Highly improved with
hydrogen storage capacity and fast kinetics in Mg-based nanocomposites by
CNTs, J. Alloys Compd. 568 (2013) 78–83.

[16] M. Khrussanova, E. Grigorova, J. Bobet, M. Khristov, P. Peshev, Hydrogen

sorption properties of the nanocomposites Mg – Mg 2 Ni 1 − x Co x obtained
by mechanical alloying, J. Alloys Compd. 365 (2004) 308–313.
doi:10.1016/S0925-8388(03)00684-4.

[17] O. Friedrichs, Hydrogen sorption improvement of nanocrystalline MgH 2 by

Nb 2 O 5 nanoparticles, Scr. Mater. 54 (2006) 1293–1297.
doi:10.1016/j.scriptamat.2005.12.011.

[18] J. Huot, G. Liang, R. Schulz, M agnesium-based nanocomposites chemical

hydrides, J. Alloys Compd. 353 (2003) 12–15. doi:10.1016/S0925-
8388(02)01306-3.

[19] G. Liang, J. Huot, S. Boily, R. Schulz, Hydrogen desorption kinetics of a

mechanically milled MgH 2 1 5at.%V nanocomposite, J. Alloys Compd. 305
(2000) 239–245.

[20] L. Xie, Y. Liu, X. Zhang, J. Qu, Y. Wang, X. Li, Catalytic effect of Ni

nanoparticles on the desorption kinetics of MgH2 nanoparticles, J. Alloys
Compd. 482 (2009) 388–392. doi:10.1016/j.jallcom.2009.04.028.

[21] J. Huot, G. Liang, R. Schulz, Mechanically alloyed metal hydride systems,

Appl. Phys. 195 (2001) 187–195.

[22] H. Imamura, K. Masanari, M. Kusuhara, H. Katsumoto, T. Sumi, Y. Sakata,

High hydrogen storage capacity of nanosized magnesium synthesized by high
energy ball-milling, J. Alloys Compd. 386 (2005) 211–216.
doi:10.1016/j.jallcom.2004.04.145.

[23] J. Huot, G. Liang, S. Boily, A. Van Neste, R. Schulz, Structural study and
hydrogen sorption kinetics of ball-milled magnesium hydride, J. Alloys
Compd. 295 (1999) 495–500.

[24] Y. Yuan, W. Heng, H.U. Lian-xi, S.U.N. Hong-fei, F. Wen-bin, Hydriding and
microstructure nanocrystallization of ZK60 Mg alloy by reaction milling in
hydrogen, Trans. Nonferrous Met. Soc. China. 19 (2009) s363–s367.
doi:10.1016/S1003-6326(10)60070-2.

[25] G. Liang, J. Huot, S. Boily, A. Van Neste, R. Schulz, Catalytic effect of

transition metals on hydrogen sorption in nanocrystalline ball milled MgH 2 –

59
 Tm (Tm 5 Ti, V, Mn, Fe and Ni) systems, J. Alloys Compd. 292 (1999) 247–
252.

[26] J.R.A. Fernandez, T. Klassen, R. Bormann, Effect of Nb 2 O 5 on MgH 2

properties during mechanical milling, Int. J. Hydrogen Energy. 32 (2007)
2400–2407. doi:10.1016/j.ijhydene.2006.10.068.

[27] K.-F. Aguey-Zinsou, T. Nicolaisen, J.R. Ares Fernandez, T. Klassen, R.

Bormann, Effect of nanosized oxides on MgH2 (de)hydriding kinetics, J.
Alloys Compd. 434-435 (2007) 738–742. doi:10.1016/j.jallcom.2006.08.137.

[28] T.R. Jensen, A. Andreasen, T. Vegge, J.W. Andreasen, K. St, A.S. Pedersen, et
al., Dehydrogenation kinetics of pure and nickel-doped magnesium hydride
investigated by in situ time-resolved powder X-ray diffraction, Int. J. Hydrogen
Energy. 31 (2006) 2052–2062. doi:10.1016/j.ijhydene.2006.02.004.

[29] C. Gerhardt » Lab Supplies » FRITSCH Planetary Micro Mill “premium line” -
PULVERISETTE 7, (n.d.).

[30] M. Polanski, J. Bystrzycki, T. Plocinski, The effect of milling conditions on

microstructure and hydrogen absorption/desorption properties of magnesium
hydride (MgH2) without and with Cr2O3 nanoparticles, Int. J. Hydrogen
Energy. 33 (2008) 1859–1867. doi:10.1016/j.ijhydene.2008.01.043.

[31] J. Bobet, E. Grigorova, M. Khrussanova, M. Khristov, P. Stefanov, P. Peshev,

et al., Hydrogen sorption properties of graphite-modified magnesium
nanocomposites prepared by ball-milling, J. Alloys Compd. 366 (2004) 298–
302. doi:10.1016/S0925-8388(03)00746-1.

[32] J. Bloch, M.H. Mintz, Kinetics and mechanisms of metal hydrides formation—
a review, J. Alloys Compd. 253-254 (1997) 529–541.

[33] K. Chou, K. Xu, A new model for hydriding and dehydriding reactions in
intermetallics, Intermetallics. 15 (2007) 767–777.
doi:10.1016/j.intermet.2006.10.004.

[34] K. Chou, Q. Li, Q. Lin, L. Jiang, K. Xu, Kinetics of absorption and desorption
of hydrogen in alloy powder, Int. J. Hydrogen Energy. 30 (2005) 301–309.
doi:10.1016/j.ijhydene.2004.04.006.

[35] F. Schweppe, M. Martin, E. Fromm, Model on hydride formation describing

surface control , diffusion control and transition regions, J. Alloys Compd. 261
(1997) 254–258.

[36] M. Martin, C. Gommel, C. Borkhart, E. Fromm, Absorption and desorption

kinetics of hydrogen storage alloys, J. Alloys Compd. 238 (1996) 193–201.
doi:10.1016/0925-8388(96)02217-7.

[37] W.B. Jung, W.Y. Lee, THE REACTION KINETICS OF HYDROGEN

STORAGE IN Mg2Ni, Int. J. Hydrogen Energy. 15 (1990) 641–648.

60
[38] M.G. Chourashiya, Y. Kim, C. Park, C. Park, Hydrogenation and
microstructural properties of hydriding combustion synthesized Mg-Ni-C
composite ball-milled with NbF 5 catalyst, J. Alloys Compd. 584 (2014) 47–
55.

[39] A. Denis, E. Sellier, C. Aymonier, J. Bobet, Hydrogen sorption properties of

magnesium particles decorated with metallic nanoparticles as catalyst, J. Alloys
Compd. 476 (2009) 152–159. doi:10.1016/j.jallcom.2008.09.016.

[40] H. Tien, M. Tanniru, C. Wu, F. Ebrahimi, Effect of hydride nucleation rate on

the hydrogen capacity of Mg, Int. J. Hydrogen Energy. 34 (2009) 6343–6349.
doi:10.1016/j.ijhydene.2009.06.008.

[41] R.R. Shahi, A.P. Tiwari, M.A. Shaz, O.N. Srivastava, Studies on de /
rehydrogenation characteristics of nanocrystalline MgH 2 co-catalyzed with Ti
, Fe and Ni, 8 (2013) 4–10. doi:10.1016/j.ijhydene.2012.11.073.

[42] J. Sestak, Study of the Kinetics of the mechanism of solid state reactions at
increasing temperatures, Thermochim. Acta. 3 (1971).

[43] V.M. Gorbachev, Some Aspects of Sestak’s Generalised kinetic eqation in

thermal analysis, J. Therm. Anal. Calorim. (1980) 3–7.

[44] H. Tien, Microstructural-based Characterization of Hydrogenation and

Dehydrogenation Mechanisms in Mg Powders, University of FLorida, 2010.

[45] M. Tanniru, H. Tien, F. Ebrahimi, Study of the dehydrogenation behavior of

magnesium hydride, Scr. Mater. 63 (2010) 58–60.
doi:10.1016/j.scriptamat.2010.03.019.

[46] H. Chen, Z. Wang, H. Zhou, C. Ni, J. Deng, Q. Yao, Hydrogen storage

properties and thermal stability of amorphous, J. Alloys Compd. 563 (2013) 1–
5. doi:10.1016/j.jallcom.2013.01.159.

[47] D.L. Sahagian, A. a. Proussevitch, 3D particle size distributions from 2D

observations: stereology for natural applications, J. Volcanol. Geotherm. Res.
84 (1998) 173–196. doi:10.1016/S0377-0273(98)00043-2.

[48] B.D. Cullity, Elements of X-Ray Diffraction, 1956.

[49] R.R. Shahi, A.P. Tiwari, M.A. Shaz, O.N. Srivastava, Studies on de /
rehydrogenation characteristics of nanocrystalline MgH 2 co-catalyzed with Ti
, Fe and Ni, Int. J. Hydrogen Energy. 38 (2013) 2778–2784.

61
62