This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights

Author's personal copy

Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

Contents lists available at ScienceDirect

Journal of Loss Prevention in the Process Industries

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

Explosion characteristics of micron- and nano-size magnesium

powders
Manju Mittal
Fire Research Laboratory, CSIR-Central Building Research Institute, Roorkee 247667, Uttarakhand, India

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 20 August 2013
Received in revised form
19 November 2013
Accepted 19 November 2013

Explosion characteristics of micron- and nano-size magnesium powders were determined using CSIRCBRI 20-L Sphere, Hartmann apparatus and Godbert-Greenwald furnace to study inuence of particle
size reduction to nano-range on these. The explosion parameters investigated are: maximum explosion
pressure (Pmax), maximum rate of pressure-rise (dP/dt)max, dust explosibility index (KSt), minimum
explosible concentration (MEC), minimum ignition energy (MIE), minimum ignition temperature (MIT),
limiting oxygen concentration (LOC) and effect of reduced oxygen level on explosion severity. Magnesium particle sizes are: 125, 74, 38, 22, 10 and 1 mm; and 400, 200, 150, 100, 50 and 30 nm. Experimental
results indicate signicant increase in explosion severity (Pmax: 7e14 bar, KSt: 98e510 bar$m/s) as particle size decreases from 125 to 1 mm, it is maximum for 400 nm (Pmax: 14.6 bar, KSt: 528 bar$m/s) and
decreases with further decrease of particle size to nano-range 200e30 nm (Pmax: 12.4e9.4 bar, KSt: 460
e262 bar$m/s) as it is affected by agglomeration of nano-particles. MEC decreases from 160 to 30 g/m3
on decreasing particle size from 125 to 1 mm, its value is 30 g/m3 for 400 and 200 nm and 20 g/m3 for
further decrease in nano-range (150e30 nm). MIE reduces from 120 to 2 mJ on decreasing the particle
size from 125 to 1 mm, its value is 1 mJ for 400, 200, 150 nm size and <1 mJ for 50 and 30 nm. Minimum
ignition temperature is 600
C for 125 mm magnesium, it varies between 570 and 450
C for sizes 38
e1 mm and 400e350
C for size range 400e30 nm. Magnesium powders in nano-range (30e200 nm)
explode less violently than micron-range powder. However, likelihood of explosion increases signicantly for nano-range magnesium. LOC is 5% for magnesium size range 125e38 mm, 4% for 22e1 mm, 3%
for 400 nm, 4% for 200, 150 and 100 nm, and 5% for 50 and 30 nm. Reduction in oxygen levels to 9%
results in decrease in Pmax and KSt by a factor of 2e3 and 4e5, respectively, for micron as well as nanosizes. The experimental data presented will be useful for industries producing or handling similar size
range micron- and nano-magnesium in order to evaluate explosibility of their magnesium powders and
propose/design adequate safety measures.
2013 Elsevier Ltd. All rights reserved.

Keywords:
Magnesium
Explosion
Micron
Nano

1. Introduction
Magnesium is being used extensively in military as well as
aeronautics, chemical and metallurgical industries. Powdered and
granulated magnesium have recently undergone rapid development to be used as metal-based fuels in propellants, explosives and
pyrotechnics, thermite and incendiary devices. Magnesium powder
manufacturing processes include cutting, milling, emulsication,
vortex milling, atomization milling, mechanical attrition milling
and airow milling producing a range from micron- to nano-size
magnesium powders. There exists high potential re and explosion risk in magnesium powder manufacturing plants and related

E-mail address: mm_s123@rediffmail.com.

0950-4230/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jlp.2013.11.001

facilities using other metalworking operations like casting,

grinding, machining and welding. High purity and ultrane powders are currently attracting more attention than in the past and an
increasing range of materials including magnesium are being produced as nano-powders composed of particles in size range from
about 1 to 100 nm for use in industrial and research elds. Information on explosion characteristics of magnesium powder is
necessary to predict the likelihood and severity of explosions and
design explosion prevention and mitigation measures. Literature
studies concerning the evaluation of ammability and explosion
risks of micron-sized powders do not enable to evaluate the re and
explosion risk probabilities and gravities of nano-powders. It is
desirable that the explosion characteristics of nano-powders be
determined using the standard apparatus and procedures already
employed for assessing dust explosion hazards. Comparison with

Author's personal copy

56

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

Khan, Dastidar, & Eckhoff, 2012; Wu, Chang, & Hsiao, 2009; Wu,
Ou, Hsiao, & Shih, 2010). Kwok et al. (2002) studied nanoaluminium explosion characteristics. Using 20-L Sphere, Vignes
et al. (2009) investigated explosion data of carbon black, carbon
nanotubes and aluminium; Wu et al. (2009, 2010) studied explosibility of aluminium, titanium and iron; and Holbrow et al. (2010)
determined explosion characteristics of different nano-materials:
iron, copper, silicon carbide, zinc oxide, iron oxide, zirconium oxide, zinc, multiwalled carbon nanotubes, carbon monobres,
aluminium, carbon black, etc. in specially designed 2-L explosion
chamber. Boilard et al. (2013) studied explosibility of micron and
nano-size titanium and Dufaud et al. (2011) studied aluminium,
zinc, carbon nanotubes and carbon black. The knowledge of explosion data of nano-powders is still limited. The recent edition of
NFPA 484 (NFPA, 2012) covering the unique hazards associated
with metal ne dusts and powders, does not include the risks of
metal dust combustibility at nano-scale.
To the authors best knowledge, there exists no data on explosion characteristics of nano-size magnesium. The reliability of
standard experimental method designed for micron-size dust explosions has been questioned due to various factors inuencing
behaviour of nano-size dust cloud during experiments (Boilard
et al., 2013; Vigens et al., 2009). In the standard 20-L sphere,
powder is dispersed using compressed air. With nano-powders,
their large surface to volume ratio means that many of them are
spontaneously ammable on contact with air, or surface oxidation
alters their properties. Hence equipment/procedure that avoids
oxidation until the point of ignition is required.
The present investigation was undertaken to measure the ignition sensitivity and explosion severity parameters of magnesium
powder in relation to particle size in both the micron- and nanosize ranges and compare the results to examine whether production, handling and use of magnesium nano-powders can present an

data for micron-scale powders of the same materials will allow

knowledge of particle size effects to be extended into the nano-size
range.
There exist some information on explosion characteristics e
maximum explosion pressure (Pmax), maximum rate of pressure
rise ((dP/dt)max), dust explosibility index (KSt), minimum explosible
concentration (MEC), minimum ignition energy (MIE), minimum
ignition temperature (MIT), limiting oxygen concentration (LOC) e
of micron scale magnesium powders (Benson, 2012; BIA, 1997;
Cashdollar & Zlochower, 2007; Eckhoff, 1991; Jacobson, Cooper, &
Nagy, 1964; Kuai et al., 2011; Li, Yuan, Fu, Zhong, & Chen, 2009;
Li, Yuan, Zhong, & Chen, 2008; NFPA, 2012; Nifuku et al., 2007;
Roser, 1998) based on experiments using standard set-ups having
same or different volumes and congurations. The literature data
are summarized in Table 1. The explosion severity data in the work
reported by Jacobson et al. (1964) were measured in 1.2-L Hartmann apparatus while those from others work are using either 20-L
or 1 m3 chamber. The values of various ignition and explosion parameters for micron scale magnesium powder reported by different
workers are quiet different even for the closer particle size-range.
For example values of explosion severity data e Pmax and KSt for
30 mm magnesium are 10.8 bar and 400 bar$m/s (Roser, 1998) and
those for 28 mm are 17.5 bar and 508 bar$m/s (BIA, 1997; Eckhoff,
1991). The variation of values of explosion characteristics obtained by the laboratories reects the difcult nature of the magnesium dust samples testing in respect of uniform experimental
conditions.
For nano-range particles, there have been number of dust explosion studies performed since 2002 (Benson, 2012; Boilard,
Amyotte, Khan, Dastidar, & Eckhoff, 2013; Bouillard, Vignes,
Dufaud, Perkin, & Thomas, 2010; Dufaud, Vignes, Henry, Perrin, &
Bouillard, 2011; Eckhoff, 2011; Holbrow et al., 2010; Kwok et al.,
2002; Pritchard, 2004; Vignes et al., 2009; Worsfold, Amyotte,

Table 1
Explosion characteristics of micron-size magnesium powder-literature data.
Reference
Jacobson et al. (1964)
Eckhoff (1991)
BIA (1997)

Roser (1998)
Cashdollar and Zlochower (2007)
Nifuku et al. (2007)

Li et al. (2008, 2009)

Benson (2012)

Kuai et al. (2011)

NFPA (2012)

Hartmann explosion tube data.

Mean particle size, mm

MEC, g/m3

<74
28
240
240
241
400

30
500
30
500
e

30
16
0e20
20e37
37e45
45e74
74e105
105e125
125e149
149e177
6
47
104
75
e
20e60
7.5
22.4
54.5
28
240
<44
<44
16

55
90
90
120
130
270
330
500
900
e
e
e
328
e
e
20
25
40
e
500
40
30
55

MIT,
C

MIE, mJ

620

40

760
e
760
n.i.u.
850

e
e
e

e
513
530
550
563
575
578
585
625
480
520
620
560
e
e
e
e
e
e
760
620
600
e

e
4
5
12
44
82
102
194
242
>2
46e54
250e300
e
20
e
e
e
e
e
e
40
240
e

Pmax, bar
90 psig
17.5
7.0
17.5
7.0
e
10.8
8.5
e
e
e
e
e
e
e
e
8
e
e
e
e
7.8
6.6
5.8
17.5
7
e
e
7.5

(dp/dt)max, bar/s

KSt, bar$m/s

LOC, %
e

12
e

e
508
12
508
e
e

400
e

9000 psi/s

e
e
e
e
e
e
e
e
150
e
e
e
53
430
380
250

e
e
e
e
e
e
e

e
e
e

e
<8%
<8%
e
e
e
e
e
e
6.8

e
e
e
e
e
e
508
12
e
e
e

e
e
e
e
e
e
e

Author's personal copy

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

enhanced dust explosion risk compared to that of micron-size

powders. Standard dust explosion equipments were used for the
study. Some modications were incorporated in experimental
procedure to minimize the problems faced by earlier investigators.

2. Experimental procedures
The explosion characteristics measured for micron- and nanosize ranges of magnesium powders include maximum explosion
pressure, Pmax, maximum rate of pressure rise (dP/dt)max, minimum
explosible concentration, MEC, minimum ignition energy, MIE,
minimum ignition temperature, MIT, limiting oxygen concentration, LOC, and inuence of reduced oxygen level on explosion
severity data using CSIR-CBRI 20-L Sphere, Hartmann apparatus
and Godbert-Greenwald furnace (Mittal, 1996, 2012).
The magnesium sizes selected for present study in micron range
are: 125, 74, 38, 22, 10 and 1 mm. Nano-range magnesium particle
sizes are 100, 50 and 30 nm. To study explosibility behaviour of
magnesium in intermediate range particle sizes 400, 200 and
150 nm were also included. Samples for particle sizes 125, 74 and
38 mm were prepared using traditional sieve analysis. Other samples were collected from the manufacturer and particle sizes were
taken to be as reported by them. No further size analysis was performed for these sizes in laboratory.
Explosion indices-Pmax and (dP/dt)max, LOC and MEC were
measured in CSIR-CBRI 20-L Sphere (Mittal, 2012). The sphere
consists in a spherical combustion chamber, a dust storage
container connected with chamber via a dust outlet valve, a pair of
electrodes holding two pyrotechnic ignitors at sphere centre, and
two piezoelectric and one strain gauge pressure sensors to record
explosion development. Typical pressure-time curve recorded
during dust explosion experiments is shown in Fig. 1. Pressure trace
starts at partially evacuated value of 0.4 bar(a). The blast of air that
disperses dust starts at 40 ms and ends at 90 ms on pressure-time
trace. Ignitor is activated at 100 ms at a chamber pressure of
1.0 bar(a). Pex is the maximum explosion pressure. Value of Pex, for a
test at a given concentration, is highest deagration pressure (absolute) minus pressure at ignition (normally 1 bar). (dP/dt)ex is
maximum rate of pressure rise reached during the course of a single
explosion experiment.

Fig. 1. Typical pressure-verses-time trace during a dust explosion experiment in 20-L

spherical vessel.

57

Explosion severity measurement experiments are conducted

over a range of dust concentration in accordance with standard
procedure (ASTM, 2010). Pyrotechnic ignitors with energy 10 kJ are
used. Pmax is maximum pressure and (dP/dt)max is maximum value
for rate of pressure increase per unit time reached during the
course of explosion for optimum dust concentration and equals
maximum slope of a tangent through the point of inexion in the
rising portion of pressure vs. time curve. Deagration constant, KSt,
characterizes explosibility of the material. KSt is maximum value of
dP/dt normalized to a 1.0 m3 volume measured at optimum dust
concentration and dened by equation (1) (Eckhoff, 1991):


KSt

dP
dt


$Vs1=3

(1)

max

where,
P e Pressure, bar
T e Time, s
Vs e Vessel volume, m3
KSt e Dust explosibility constant, bar.m/s
KSt values rounded to nearest integer are used.
MEC, the lowest concentration of dust in air that will ignite on
contact with an ignition source and propagate a dust explosion, is
determined following the standard procedure (ASTM, 2007a) using
two chemical ignitors with total energy of 2 kJ. The test commences
with experiments at a high powder concentration. If this concentration of powder ignites, the dust concentration is reduced in steps
until no ignition is evident. An ignition is deemed to have occurred
if the maximum explosion pressure is at or above 2 bar(a). The test
is then repeated with a lower dust concentration which is reduced
further until a concentration is reached at which no ignition of
dust/air mixture is observed in three consecutive tests.
LOC, the highest permissible oxygen concentration below which
dispersed dust ignition is not possible at ambient temperature and
pressure, is determined in 20-L Sphere by conducting experiments
at various oxygen levels over a wide range of dust concentration as
per standard procedure (BS EN 14034-4, 2005) using two chemical
ignitors with total energy of 2 kJ. The experiments are started with
an oxygen concentration of 21% and continued at reduced oxygen
concentrations lowered in steps of 3% initially and 1% when it came
closer to LOC. The pressure-time data for each experiment is
recorded. From pressure-time curve the explosion pressure, Pex,
and rate of explosion pressure rise, (dP/dt)ex, are determined for
each dust concentration. An ignition/explosion of dust is considered to have taken place, when the measured overpressure (inuence of chemical ignitors included) relative to the initial pressure Pi
is 0.5 bar (Pex  (Pi 0.5 bar)). Optimum dust concentration, Cex,
at which the highest explosion pressure occurs at any experimental
oxygen concentration, is determined. Maximum explosion pressures, Pmax, and maximum rate of pressure rise, (dP/dt)max are
determined by varying the dust concentrations at various oxygen
levels. By increasing the ratio of inert gas to air step by step and
varying the dust concentration, the oxygen concentration is
reduced to a level at which explosions no longer occur at any dust
concentration. The highest oxygen concentration where no ignition
occurs in three consecutive tests is reported as the LOC.
Explosion severity data (Pmax and KSt), MEC and LOC for micronrange of particles and nano-sizes 400 and 200 nm were measured
as per standard procedures. However, Nitrogen was used in place of
air as dispersion medium for nano-magnesium sizes: 30, 50, 100
and 150 nm to prevent pre-ignition of the powder in the dust
storage and explosion chambers as reported in literature (Vignes
et al., 2009). Similar approach was used by Boilard et al. (2013)

Author's personal copy

58

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

for studying explosibility of nano-titanium powder. The vacuum

created in 20-L chamber was lower than usual and oxygen was
lled to create an elevated oxygen level (prior to dust dispersion).
As the dispersing nitrogen and elevated oxygen level are mixed,
pressure in explosion chamber is atmospheric. The concentration of
oxygen in the explosion chamber while using nitrogen as
dispersing medium in reservoir was ascertained by some tests done
without dusts and measuring the oxygen concentration in the 20-L
chamber using the oxygen analyser. A sugar dust with known
characteristics was tested using this method for validating the
procedure.
MIE, used to represent the ease or difculty of powder ignition
by electrical sparks and arcs or electrostatic discharge was
measured in 1.2-L Hartmann apparatus as per standard procedure
(ASTM, 2007b; Mittal, 1996). The apparatus used for this test consists of a glass tube placed over a dispersion cup and tted with two
brass electrodes (3 mm diameter) connected to a circuit which
produces an electrical spark of known energy. The spark generating
system has three units: variable capacitor bank having capacitors
with different values; variable voltage supply/continuous spark
unit; and step-up transformer. The spark is triggered by high
voltage transformer using two-electrode system connected to the
secondary winding of the transformer and tted in the Hartmann
tube. The self-inductance of the secondary coil of the trigger
transformer is 1 mH. Electrodes with rounded tips have been used
to reduce corona effects as for circuit in which high voltage is
maintained across the gap prior to spark breakdown, a signicant
fraction of the energy stored in the capacitor may drain away as
corona discharge from the sharp electrode tips prior to spark
discharge. This is important at low stored energies. The spacing
between the electrodes is varied between 2 and 6 mm while carrying tests. With variable combination of capacitance and voltage, it
is possible to obtain sparks with ignition energies in the range
0.5 mJe3.2 J using this circuit. There is precise electronic synchronization between dust dispersion and spark onset. The ignition
delay times can be selected between 20 and 200 ms with an
increment of 20 ms. Energy discharged from the capacitor is
calculated from the following formula, assuming no energy losses
in the transformer,


1  2
C Vi  Vf2
2

(2)

where,
W e Discharge energy, J
C e Total capacitance of discharge circuit, F
Vi e Initial voltage of charged capacitor, V
Vf e Final voltage of charged capacitor, V

from the charged capacitor. The stored energy discharged into the
spark and the occurrence or non-occurrence of ame is recorded by
visual observation of ame propagation away from the spark gap. If
ame propagation is observed, the energy of the spark is reduced
until no ame propagation is seen for ten consecutive tests. The
minimum ignition energy is sought by varying the dust concentration, the spark discharge energy and the ignition delay time.
When changing the ignition delay time, only points of non-ignition
are tested. Experiment is started with a dust loading corresponding
to dust concentration twice as high as the experimentally determined MEC or 250e500 g/m3 and go/no-go spark energy is estimated. Once a limit is found for a particular dust concentration, the
procedure is repeated for higher and lower dust concentrations
until a roughly parabolic curve is obtained for ignition energy
verses dust concentration. Ten tests are conduced at most ignitable
dust concentration. MIE is determined from the MIE vs. concentration curve.
The MIT is the lowest temperature of a hot surface that will
cause a dust cloud to ignite and propagate ame. GodbertGreenwald furnace (heated section: 3.9 cm diameter, 23 cm long,
0.27 l volume) was used to measure MIT of magnesium samples as
per standard procedure (ASTM, 2006; Mittal, 1996). The initial tests
are carried out at dust concentrations at which the maximum
pressure, Pmax, and the maximum normalized rate of pressure-rise,
KSt, are found. The dust is placed in a dust holder at the top of
temperature controlled furnace with an open bottom. The temperature of the furnace is set at a predetermined value. The dust is
dispersed by the compressed air downwards past the hot surface of
the furnace to see if ignition occurs and ame is produced. Ignition
is dened as the ame exiting from the furnace chamber. Furnace
temperature, as well as dust concentrations are varied to ensure
that the auto ignition temperature is measured at the most easily
ignitable concentration. At least ve tests over a range of concentrations are run at the highest temperature at which ignition is not
observed. At the temperature at which ignition is observed, dust
concentration is varied to determine the range over which ignition
occurs. To assure repeatability of data, at least 3 ignitions are
observed over a range of concentrations at the lowest temperature
at which ignition occurs. The MIT is usually relatively independent
of dust concentration.
Adequate precautions were taken in execution of experiments
and handling the magnesium powder throughout the work. Magnesium may react with moisture to evolve hydrogen which is a
highly dangerous re and explosion hazard. At ambient temperatures, micron-size magnesium does not pose a hazard in this regard; however, nano-size magnesium can react with moisture to
create hydrogen gas. The nano-samples were stored under dry nitrogen atmosphere in this study.
3. Results & discussions

A spark energy measurement system has been integrated in the

spark generation circuit and net spark energies generated for
various combinations of capacitance and voltages are determined
in the conventional way by measuring current and voltage across
spark gap as a function of time and integrating the power-versestime curve. Measured spark energies were typically 90e95% of
the theoretical energies computed using equation (2).
To establish MIEs, the tests are done by starting at relatively high
spark energy levels, i.e. a relatively large discharge capacitance is
used. The capacitance is then reduced in steps until no ignition
occurs for ten ignition trials. A weighed sample of dust is placed in
the dispersion cup. A blast of compressed air is used to disperse the
dust in the glass cylinder where it is ignited by a spark between two
electrodes. Ignition trials of this dust-air mixture are then
attempted, after a specic ignition delay time, by a spark discharge

3.1. Explosion severity data

Experiments for explosion severity measurement were conducted for dust concentration range 20e2000 g/m3 for nano-size
particles (30, 50, 100, 150, 200 and 400 nm) and for 30e2500 g/
m3 for micron-size particles (1, 10, 22, 38, 74 and 125 mm). Pressuretime curves were recorded for all the experiments. The experimental curves were analysed and values of Pex, (dP/dt)ex and (dP/
dt)ex$V1/3were estimated for each experiment. The values of Pmax
and KSt corresponding to optimum dust concentration were
determined for each particle size.
The experimental data on explosion severity of nano-powders
from the work of Vignes et al. (2009) are based on experiments
in 20-L Sphere using chemical ignitors as per standard procedure.

Author's personal copy

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

The reason for low value of explosion severity data for nanopowders in comparison to micron size particles was explained as
uncontrolled pre-ignition of the dust cloud by friction/impact
during injection/dispersion of the powder. As mentioned under
experimental procedure, nitrogen was used for dispersing nanosize magnesium dust to eliminate chances of ignition and explosion in the external dispersion reservoir and explosion chamber
during dust dispersion. The explosion severity tests were started
with 30 nm magnesium. It was observed that as the nanomagnesium-nitrogen mixture was dispersed forming the oxygen
concentration in the explosion chamber as 21%, there was immediate ignition of the mixture. The recorded pressure-time trace was
different from that shown in Fig. 1. There was an overlap of the dust
dispersion and explosion steps; with the ignitors ring after the
dust had exploded. Further experiments for 30, 50, 100 and 150 nm
were conducted without using chemical ignitors. Ignition followed
by explosion occurred in all these tests.
The sudden explosion during the dispersion process is referred
to as the pre-ignition of nano-size magnesium. The resultant
pressure change is recorded as a pressure-time curve. Fig. 2 shows a
typical pressure-time trace recorded using strain gauge pressure
transducer for nano-range of magnesium for the size 150 nm and
dust concentration 30 g/m3 without chemical ignitor. For all the
experiments involving nm magnesium in size range 30e150 nm
results similar to those in Fig. 2 were obtained. Since the explosion
is not occurring at the proper conditions (time or proper dispersion), the results are different from those for micron size magnesium. The explosion is violent and causes complete combustion
even before the time specied for activation of ignitor. Typical
procedure used for determining maximum explosion pressure and
rate of pressure rise using Fig. 1 cannot be applied to particle size
range 30e150 nm. Maximum explosion pressure and maximum
rate of pressure rise were estimated by manual evaluation of
pressure-time curve.
Figs. 3 and 4 present experimental results showing variation of
explosion pressure, Pex, and normalized rate of pressure rise, (dP/
dt)ex$V1/3, with dust concentration, respectively, for micron- and
nano-size magnesium particles. Explosion severity data for the
particle sizes 125, 74, 38, 22, 10 and 1 mm, 400 nm and 200 nm are
based on standard experimental procedure and 150, 100, 50 and
30 nm involve ignition without chemical ignitors due to frictional
or static sparking during dispersal sequence using nitrogen as
dispersing gas in the dispersion chamber. The optimum dust concentrations, Cex, at which maximum values of explosion
severity data occur, are in the range 1000e1500 g/m3 for micron

Fig. 2. Typical pressure-verses-time trace for nano-magnesium explosion without

chemical ignitor.

59

Fig. 3. Variation of explosion pressure, Pex, with dust concentration for micron- and
nano-size magnesium powder.

range 1e125 mm, and 750 g/m3 for particle size range 30e400 nm.
This indicates that optimum dust concentrations at which
maximum values of explosion severity data, Pmax and KSt, occur are
higher for micron-size particles than that for nano-size particles as

Fig. 4. Variation of normalized rate of explosion pressure-rise, (dP/dt)ex$V1/3, with dust

concentration for micron- and nano-size magnesium powder.

Author's personal copy

60

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

Table 2
Explosion characteristics for micron and nano-size magnesium powder e experimental results e CSIR-CBRI study.
Particle size

MEC,
g/m3

MIE, mJ

MIT,
C

Pmax,
bar

KSt,
bar$m/s

LOC, %

125 mm
74 mm
38 mm
22 mm
10 mm
1 mm
400 nm
200 nm
150 nm
100 nm
50 nm
30 nm

160
90
60
50
40
30
30
30
20
20
20
20

120
50
10
4
3
2
1
1
1
<1
<1
<1

600
570
530
510
480
450
400
380
350
350
350
350

7
8.8
10.8
12.4
13.2
14
14.6
12.4
11.0
10.6
10
9.4

98
202
362
450
482
510
528
460
400
360
332
262

5
5
5
4
4
4
3
4
4
4
5
5

for micron-size magnesium, larger amount of dust are needed to

achieve the same effective surface area as with a smaller mass of
nano-size magnesium and there is a high reactivity at low concentrations for nano-size magnesium.
The Pmax and KSt values at optimum dust concentration for all
the particle sizes of magnesium are presented in Table 2 and their
variations with particle size are shown in Fig. 5 which indicates
Pmax and Kst both increase signicantly with decreasing size from
125 to 1 mm in the micron range as expected. The data for 400 nm
size show further increase in explosion severity. However, the data
for 200 nm powder show decrease in explosion severity. This is

probably due to the degree of passivity of the magnesium, that

could be different for these two ranges of particles and degree of
agglomeration which is more important for nm-particles. Explosion severity data set for size range 150e30 nm (ignition without
chemical ignitors) also show further decrease of explosion severity
with decrease in particle size.
Comparison of above results with literature data on explosion
severity of micron range magnesium having sizes closer to that
involved in present study indicates values of Pmax and KSt for 22 mm
were found higher than those for 30 mm magnesium (Roser, 1998)
and lower than those for 28 mm (BIA, 1997; Eckhoff, 1991).
3.2. Minimum explosible concentration (MEC)
MEC determination for micron-size (125, 74, 38, 22, 10 and
1 mm), 400 nm and 200 nm magnesium samples in CSIR-CBRI 20-L
chamber was done as per standard procedure using chemical
ignitor with 2 kJ ignition energy. However, MEC of nano-size (150,
100, 50 and 30 nm) was determined considering the pre-ignition
aspect of the dust during dispersion sequence. For 150 nm powder experiments were started with a dust concentration 10 g/m3
using nitrogen dispersion and chemical ignitors with 2 kJ ignition
energy. There was no pre-ignition and no explosion was recorded.
Similar observations were made at a dust concentration of 20 g/m3.
Experiments at a dust concentration of 30 g/m3, using nitrogen
dispersion and ignitor, showed pre-ignition occurred and the
maximum recorded pressure was 2.4 bar(a). The experiment at
30 g/m3 dust concentration was then done by nitrogen dispersion
without chemical ignitor. The recorded explosion pressure was
again 2.4 bar(a). As in case of explosion severity testing, for 150,
100, 50 and 30 nm magnesium powder MEC experiments were
conducted without chemical ignitors. Experimental MEC value for
all the magnesium particles are given in Table 2. The MEC of
magnesium powder is 20 g/m3 for nano size range 30e150 nm,
30 g/m3 for particle sizes 200 nm, 400 nm and 1 mm, and increased
with further increase in particle size. Similar data for micron size
particles have been reported in literature (Eckhoff, 1991; Kuai et al.,
2011; NFPA, 2012). Metal particles melt, evaporate and burn as
discrete entities as described by Eckhoff (1991), thus dust clouds
with smaller particles ignite and explode more easily.
3.3. Minimum ignition energy (MIE)

Fig. 5. Effect of moving from micron- to nano-size particles on explosion severity data,
Pmax and KSt, for magnesium powder.

The MIEs of all the magnesium samples were measured over a

range of dust concentration following the procedure explained
under Section 2. Fig. 6 presents typical MIE test results showing
ignition energy verses dust concentration for particle sizes 22 mm
and 150 nm, for optimum ignition delays. Experimental points
showing ignition and non-ignition at ignition energies higher and
lower than MIE, respectively, are also shown. This gure indicates
MIE for 22 mm as 4 mJ and 150 nm as 1 mJ. Similar curves were
obtained for all the sizes of magnesium particles and MIE was
determined using these.
Experimental results on MIE measurement for all sizes of
magnesium are given in Table 2 indicating that ignition sensitivity
increases (MIE decreases) when moving from micron-particle dust
clouds to nano-particle clouds. Once within the nano-range, the
minimum ignition energy is quite small (<1 mJ) for 30, 50 and
100 nm sizes. Nano-particle agglomeration does not seem to affect
the minimum ignition energy as particle agglomerates may remain
agglomerates in ame propagation or may disintegrate into primary nano-particles when they approach the extremely hot spark
channel and a small zone of genuine nano-particles may be
generated in the vicinity of the spark and facilitate ignition of dust

Author's personal copy

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

61

Fig. 7. Typical detailed data for determining minimum ignition temperature for
micron and nano-size magnesium powder.

Fig. 6. Typical experimental results for determining minimum ignition energy, MIE, of
micron- and nano-size magnesium powder.

or explosion risks are inherently higher. The MIT value for micron
size magnesium falls in a range closer to the values reported by
Jacobson et al. (1964), Nifuku et al. (2007) and Li et al. (2008).
3.5. Limiting oxygen concentration

cloud. The MIE for micron size magnesium is closer to that published by Jacobson et al. (1964) and Nifuku et al. (2007).
3.4. Minimum ignition temperature (MIT) of dust clouds
The MIT for each particle size of magnesium powder was
determined in Godbert-Greenwald furnace starting with dust
concentrations at which the maximum pressure Pmax, and the
maximum normalized rate of pressure-rise, KSt, were found in
present study as per standard procedure. Typical detailed experimental data for determining MIT of magnesium dust cloud are
given in Fig. 7 for 22 mm and 150 nm magnesium powder. The ignitions/explosions are shown as the solid data symbols and the
non-ignitions/non-explosions as the open data symbols. This gure
shows MIT for 22 mm as 510
C and 150 nm as 350
C.
Experimental results on MIT measurement for all sizes of
magnesium are given in Table 2. Minimum ignition temperature
was found to decrease with decreasing particle size, increasing the
probability of explosions with nano-sized particles over larger
particles. 125 mm particles ignited in the G-G furnace at 600
C.
Smaller micron-size magnesium particles (74, 38, 22, 10, 3, 1 mm)
ignited at a lower furnace temperature in the range 570e450
C.
The minimum ignition temperature of the nano-size magnesium
(400, 200, 150, 100, 50 and 30 nm) was in the range of 400e350
C.
Lower values of MITs for nano-size than those required to ignite the
micron-size magnesium, indicates that nano-size magnesium has
enhanced potential for ignition by hot surfaces and inammation

Systematic experiments were conducted to determine LOC and

study the effect of reduced oxygen on explosion violence data for all
magnesium samples at oxygen levels 18, 15, 12, 9, 6, 3, 4 and
5% for dust concentration range 60e2500 g/m3 for micron- and
60e2000 g/m3 for nano-size particles. Each experiment was
repeated thrice. The pressure-time curves were recorded for all the
experiments. The experimental curves were analysed and values of
Pex and (dP/dt)ex were estimated for each experiment. Pmax and (dP/
dt)max and KSt were estimated at each oxygen concentration.
Experiments for magnesium samples with sizes 125, 74, 38, 22,
10 and 1 mm, 400 nm and 200 nm were performed as per standard
procedure and those for nano-sizes 150, 100, 50 and 30 nm using
nitrogen as dispersing media. The dust concentration range was
same as for explosion severity tests at 21% oxygen level for each
sample. The experiments, in which recorded explosion pressure
was 0.5 bar, were considered as indicating ignition/explosion. The
boundary between oxygen concentration that support combustion
and that do not support combustion is the limiting oxygen concentration. Typical experimental results showing the variation of
maximum explosion pressure, Pex, and normalized rate of pressure
rise (dp/dt)ex$V1/3 with dust concentration for each oxygen level are
presented in Fig. 8 for 22 mm magnesium dust. In order to compare
the inerting effect the corresponding values for tests at 21% O2 level
are also illustrated. This gure indicates that the explosion severity
decreases with a decrease in oxygen concentration. Explosible
behaviour was observed and ame propagation were recorded for

Author's personal copy

62

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

Fig. 8. Effect of oxygen concentration on explosion severity data e typical experimental results for 22 mm magnesium powder.

Fig. 9. Effect of oxygen concentration on explosion severity data e typical experimental results for 150 nm magnesium powder.

oxygen concentrations down to 5% and there is no explosion at 4%

oxygen for this size of particle as Pex is < 0.5 bar at all the dust
concentrations for 4 and 3% oxygen. Based on these results, LOC for
22 mm magnesium powder was considered 4%. The maximum explosion data at various oxygen levels occur at a dust concentration
1250 g/m3. The results also conrm the existence of rich limits,
rarely observable at higher oxygen concentrations, and made
evident at reduced oxygen concentration (5% curves in Fig. 8). The
maximum explosion pressure, Pmax, and KSt (i.e. (dP/dt)max$V1/3)
varied between 1.8 to 12.4 bar and 28e450 bar$m/s, respectively,
when oxygen concentration was increased from 5 to 21%. Similar
data were collected for particle sizes e 125, 74, 38, 10 and 1 mm,
400 nm and 200 nm. The resultant values of LOC for these sizes are
given in Table 2.
For nano-sizes 150, 100, 50 and 30 nm magnesium experiments
were conducted without chemical ignitors for oxygen concentrations 18 and 15% as pre-ignition followed by explosion was
observed at these oxygen concentrations. However, when the oxygen concentration was reduced to 12%, there was no pre-ignition
of dust and the test series were completed using chemical ignitors. Typical experimental results showing the variation of
maximum explosion pressure, Pex, and normalized rate of pressure
rise (dP/dt)ex$V1/3 with dust concentration for each oxygen level for
150 nm magnesium are presented in Fig. 9. The data were analysed
in similar manner as explained above. The LOC for 150 nm magnesium was found 4%. Pmax and KSt varied between 1.4 to 11 bar and
24e400 bar$m/s, respectively, at an optimum dust concentration of
750 g/m3 when oxygen concentration was increased from 5 to 21%.
Similar data were collected for particle sizes 100, 50 and 30 nm.
Resultant values of LOC for these sizes are given in Table 2. For
micron-size magnesium LOC is 5% for the size range 125e38 mm, 4%
for 22e1 mm, 3% for 400 nm and 4% for 200e100 nm. As expected
LOC was again 5% for 50 and 30 nm magnesium due to

agglomeration effect in nano-powders. It has been observed that

dust particle size has a comparatively small inuence on the
limiting oxygen concentration as LOC was found to vary from 3 to
5% for the particle size range 125 mm to 30 nm. Literature values of
LOC are 6.8% for 47 mm magnesium (Li et al., 2009) and <8% for
37 mm magnesium (Nifuku et al., 2007).
Figs. 10 and 11 show variation of maximum explosion pressure,
Pmax, and KSt with oxygen concentration for all the experimental
results presented in Figs. 8 and 9 for 22 mm and 150 nm, respectively. The maximum explosion pressure in Fig. 10 varies almost
linearly with oxygen concentration upto certain oxygen level (9%).
However, this linear relationship changes as the LOC is
approached. Near the LOC, the pressure would decrease very
rapidly with decreasing oxygen concentration, until mixture
would no longer be explosible. As expected, explosion overpressure decreases with decreasing oxygen concentration. Fig. 11
indicates that the KSt value i.e. the rate of pressure rise varies
almost exponentially with oxygen concentration on this semilogarithmic plot upto certain oxygen level (9 or 12%) and this
linear relationship changes as the LOC is approached. Explosion
violence and ignition sensitivity of dust clouds decrease with
decreasing oxygen content of the gas in which the dust is suspended. It is worth mentioning that for 22 mm and 150 nm
magnesium a reduction of the oxygen content from that of air to
9% caused a reduction of KSt by a factor of w4, whereas Pmax was
reduced by a factor of w2. Similar data were obtained for other
sizes of magnesium and reduction in oxygen levels to 9% resulted
in decrease in Pmax and KSt by a factor of 2e3 and 4e5, respectively, for micron as well as nano-sizes. This illustrates the strong
inuence of the oxygen content on the kinetics of the combustion
process. The reduction of the maximum pressure is approximately
proportional to the reduction of the oxygen content, as would be
expected from thermodynamic considerations, whereas the

Author's personal copy

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

63

Fig. 10. Variation of maximum explosion pressure, Pmax, with oxygen concentration for
micron- and nano-size magnesium powder.

Fig. 11. Variation of dust explosibility constant, KSt, with oxygen concentration for
micron and nano-size magnesium powder.

maximum rate of pressure rise falls much more sharply. There

appears to be a correlation between the two variables. This is in
agreement with the idea of the oxygen being the limiting reactant
at optimum conditions. In fact, measurements of the residual oxygen in the after gases shows that oxygen is almost totally
consumed in the conditions of maximum reactivity for each series
of Figs. 8 and 9, down to 9%. In the experiments conducted at
initial oxygen content lower than 9%, measurements of residual
oxygen in the after gases indicate that complete combustion is
never achieved, although considerable overpressures are
measured.
The trend of curves for variation of Pmax and (dP/dt)ex$V1/3 in
Figs. 8 and 9 is similar to those obtained by Li et al. (2009) in their
investigation of explosion violence of 47 mm magnesium at reduced
oxygen levels.

values than those of dust clouds of micron-size particles. The delay

between the moment of formation of a well-dispersed nano-size
particle dust cloud and ignition decides the level of coagulation. In
laboratory scale determination of MIE and KSt typical delays between dust cloud generation and ignition are of the order of 20e60
ms and coagulation is in its initial stage. Laboratory test results
overestimate ignition sensitivity and explosions severity of real
industrial nano-magnesium clouds as expected ignition delays in
accidental explosion scenarios are signicantly longer than in laboratory tests. It is feasible to use dust explosion safety measures
venting, suppression and isolation as per standards (NFPA, 2007,
2008; VDI, 2002) for dust clouds of nano-size particles as with
micron-size.
The MEC, MIE and MIT values are lower for clouds of nano-size
magnesium than for clouds of micron-size primary particles. Low
ignition energies may allow nano-size magnesium powders to
ignite due to electric/electrostatic sparks, collision, or mechanical
friction/impact during various operations like mixing, grinding,
drilling, sanding, cleaning. Careful precautions have to be taken to
prevent ignition of clouds of nano-size magnesium powders in air
by these ignition sources. Moreover, particles below about one
micron can be considered to remain airborne for long periods of
time, with minimum fall out of the dust, posing higher risk of explosion. Sparks with 2 mJ energy are more likely to be generated
during production, such as with static electricity discharge, mechanical friction or impact. As MIE is less than 10 mJ for nanomagnesium and for particle size upto 38 mm, preventing explosions solely by avoiding ignition sources becomes impossible. For
this reason, inerting technology has been widely recommended in
handbooks, standards and guidelines as the preferred means of
preventing magnesium dust explosions (NFPA, 2007, 2012).

3.6. Interpretation of results and safety measures

The explosibility data for magnesium powders having particle
sizes from 125 mm down to 400 nm, encourage to imagine that dust
clouds from combustible nano-size (<200 nm) particles will have
extreme explosion severity. However, laboratory scale experimental data from the current work have shown that nano-size
powder clouds of magnesium do not have any larger KSt evalue
than clouds of micron-size powders. In fact, in actual industrial
scenario, formation of clouds of nano-size particles is extremely
difcult due to comparatively very strong inter-particle cohesion
forces and in case such clouds are formed, coagulation would
transform the primary nano-size particles into much larger agglomerates. Thus dust clouds consisting of nano-size primary particles in actual practice are not expected to exhibit more severe KSt

Author's personal copy

64

M. Mittal / Journal of Loss Prevention in the Process Industries 27 (2014) 55e64

4. Conclusion
The experimental results presented in this paper indicate that
explosion severity of magnesium nano-powders (size < 200 nm) is
lower than that of micron-powders, but the likelihood of explosion
is quite high due to very low ignition energies of magnesium nanopowders. Relatively low quantities and very low spark energies can
cause nano-magnesium to ignite leading to severe explosions. The
magnesium dust with nano-size range can self-ignite under the
appropriate conditions during laboratory testing and handling. It
was not possible to conduct experiments using standard experimental procedures in 20-L Sphere and some modications were
incorporated during the experiments in present study. Explosion
severity of nano-magnesium may be affected by pre-ignition phenomenon and agglomeration level of the nano-powders during
testing. For loss prevention purposes, with the explosion severity
not being very different between the nano-and micron-scale metal
powders, mitigation techniques used to reduce the magnitude of
the explosion for micron-scale metal dusts may be applicable to the
nano-scale materials as well. However, mitigation techniques in
NFPA 484 (NFPA, 2012) and other such standards which focus on
ignition source avoidance will have to be modied for nano-scale
metal dusts. The explosion data for micro-scale powders cannot
be extrapolated to nano-powders with any degree of condence. In
order to assess the explosion hazards of nano-powders experimental data for a representative range of materials, at both scales,
are required for designing explosion safety measures for facilities
handling the powders. The experimental data obtained in present
investigation will be useful for industries producing or handling
similar size range micron- and nano-magnesium in order to evaluate explosibility of their magnesium powders and propose/design
adequate safety measures.
Symbols used
C [F]
Cex [g/m3]
KSt [bar$m/s]
P [bar]
Pex [bar]
Pmax [bar]
t [s]
Vs [m3]
Vi, [V]
Vf [V]
W [J]

Total capacitance of discharge circuit

Optimum dust concentration
Dust explosibility constant
Pressure
Maximum explosion pressure for the dust
concentration tested
Maximum explosion pressure
Time
Vessel volume
Initial voltage of charged capacitor
Final voltage of charged capacitor
Discharge energy

References
ASTM. (2006). Standard test method for minimum autoignition temperature of dust
clouds, E1491. West Conshohocken, PA: ASTM International.
ASTM. (2007a). Standard test method for minimum explosible concentration of
combustible dusts, E1515. West Conshohocken, PA: ASTM International.
ASTM. (2007b). Standard test method for minimum ignition energy of a dust cloud in
air, E 2019-03. West Conshohocken, PA: ASTM International.
ASTM. (2010). Standard test method for explosibility of dust clouds, E1226. West
Conshohocken, PA: ASTM International.
Benson, J. M. (2012). Safety considerations when handling metal powders. The
Journal of Southern African Institute of Mining and Metallurgy, 7A, 563e575.

BIA. (1997). Combustion and explosion characteristics of dusts. Report 13/97 from BIA
(Professional Associations Occupational Safety Institute), published by Germany: HVBG (Federation of the Statutory Accident Insurance Institution of the
Industrial Sector).
Boilard, P. S., Amyotte, P. R., Khan, F. I., Dastidar, A. G., & Eckhoff, R. K. (2013).
Explosibility of micron and nano-size titanium powders. Journal of Loss Prevention in the Process Industries. Available online June 27, 2013.
Bouillard, J., Vignes, A., Dufaud, O., Perkin, L., & Thomas, D. (2010). Ignition and
explosion risks of nanopowders. Journal of Hazardous Materials, 181, 873e880.
BS EN 14034e4. (2005). Determination of explosion characteristics of dust clouds-Part
4: Determination of the limiting oxygen concentration LOC of dust clouds. British
Standards Institution.
Cashdollar, K. L., & Zlochower, I. A. (2007). Explosion temperatures and pressures of
metals and other elemental dust clouds. Journal of Loss Prevention in the Process
Industries, 20, 337e348.
Dufaud, O., Vignes, A., Henry, F., Perrin, L., & Bouillard, J. (2011). Ignition and explosion of nanopowders: something new under the dust. In , Journal of Physics:
Conference series: 304. Nanosafe 2010: International conference on safe production
and use of nanomaterials (p. 012076). IOP Publishing.
Eckhoff, R. K. (1991). Dust explosions in the process industries. Oxford: Butterworth
Heinemann.
Eckhoff, R. K. (2011). Are enhanced dust explosion hazards to be foreseen in
production, processing and handling of powders consisting of nano-sized
particles. In , Journal of Physics: Conference series: 304. Nanosafe 2010: International conference on safe production and use of nanomaterials (p. 012075).
IOP Publishing.
Holbrow, P., Wall, M., Sanderson, E., Bennett, D., Rattigan, W., Bettis, R., et al. (2010).
Fire and explosion properties of nanopowders. RR782. UK Health and Safety
Executive.
Jacobson, M., Cooper, A. R., & Nagy, J. (1964). Explosibility of metal powders. Report of
investigation 6516. Pittsburg: U.S. Bureau of Mines.
Kuai, N., Li, J., Chen, Z., Huang, W., Yuan, J., & Xu, W. (2011). Experiment-based
investigations of magnesium dust explosion characteristics. Journal of Loss
Prevention in the Process Industries, 24(4), 302e313.
Kwok, Q. S. M., Fouchard, R. C., Turcotte, A., Lightfoot, P. D., Bowes, R., &
Jones, D. E. G. (2002). Characterization of aluminum nanopowder composition.
Propellants, Explosives, Pyrotechnic, 27, 229e240.
Li, G., Yuan, C. M., Fu, Y., Zhong, Y. P., & Chen, B. Z. (2009). Inerting of magnesium
dust cloud with Ar, N2 and CO2. Journal of Hazardous Materials, 170(1), 180e183.
Li, G., Yuan, C. M., Zhong, Y. P., & Chen, B. Z. (2008). Experiment-based re and
explosion risk analysis for powdered magnesium production methods. Journal
of Loss Prevention in the Process Industries, 21, 461e465.
Mittal, M. (1996). An analysis of hazards in chemical industries. Indian Chemical
Engineer, 38(2), 69e80.
Mittal, M. (2012). Dust explosion violence measurement with 20-L spherical vessel.
Chemical Products Finder, 31(5), 34e36.
NFPA. (2007). Standard on explosion protection by deagration venting, NFPA 68.
Quincy, MA: National Fire Protection Association.
NFPA. (2008). Standard on explosion prevention systems, NFPA 69. Quincy, MA: National Fire Protection Association.
NFPA. (2012). Standard for metal dust explosions, NFPA 484. Quincy, MA: National
Fire Protection Association.
Nifuku, M., Koyanaka, S., Ohya, H., Barre, C., Hatori, M., Fujiwara, S., et al. (2007).
Ignitability characteristics of aluminum and magnesium dusts that are generated during the shredding of post-consumer wastes. Journal of Loss Prevention in
the Process Industries, 20, 322e329.
Pritchard, D. K. (2004). Literature review-explosion hazard associated with nanopowders. UK Health and Safety Executive. HSL/2004/12.
Roser, M. (1998). Investigation of dust explosion phenomena in interconnected process
vessels (Ph.D. thesis). UK: Loughborough University.
VDI- 3673. (2002). Part 1: Pressure venting of dust explosions. Dusseldorf, Germany:
Verein Deutschr Ingeniere.
Vignes, A., Traore, M., Dufaud, O., Perrin, L., Bouillard, J., & Thomas, D. (2009).
Assessing explosion severity of nanopowders with the 20-L sphere. In Paper
presented at the 8th World Congress of chemical engineering (August 23e27, 2009,
Montreal, Canada).
Worsfold, S. M., Amyotte, P. R., Khan, F. I., Dastidar, A. G., & Eckhoff, R. K. (2012).
Review of the explosibility of nontraditional dusts. Industrial & Engineering
Chemistry Research, 51(22), 7651e7655.
Wu, H. C., Chang, R. C., & Hsiao, H. C. (2009). Research of minimum ignition energy
for nano titanium powder and nano iron powder. Journal of Loss Prevention in
the Process Industries, 2, 21e24.
Wu, H. C., Ou, H. J., Hsiao, H. C., & Shih, T. S. (2010). Explosion characteristics of
aluminum nanopowders. Aerosol and Air Quality Research, 10(1), 38e42.