J. Am. Ceram. Soc., 88 [7] 1870–1874 (2005)
DOI: 10.1111/j.1551-2916.2005.00358.x
ournal
J
Spark Plasma Sintering Effect on the Decomposition of MgH2
Jurgen Schmidt,w Rainer Niewa, Marcus Schmidt, and Yuri Grinz
¨
Max-Planck-Institut fur Chemische Physik fester Stoffe, 01187 Dresden, Germany
¨
Spark plasma sintering (SPS) is one of the advanced consoli-
dation techniques developed in the last few decades. We have
studied the decomposition behavior of MgH2 and MgH2/graph-
ite 1:1 mixtures in the SPS process. The standard SPS setup
chosen was modified by including the temperature measurement
inside the sample, so that the data can be compared with findings
from thermal analysis. The results show clearly the reduction of
the decomposition temperature measured in the SPS process if
the necessary conditions (sufficient current density by applying
the insulation) are realized.
The present paper describes the influence of the pulsed electric
current on the decomposition of magnesium hydride during the
SPS process. This particular reaction was investigated in detail,
because it was subsequently used for preparation of magnesium
11,12
compounds, i.e. magnesium diboride Mg0.96B2
or magnesi-
um silicide Mg2Si.13 A thermocouple measuring the tempera-
tures inside the sample allowed comparison with results
obtained by differential thermal analysis (DTA). The decompo-
sition of pure MgH2 and MgH2 in a 1:1 mixture with graphite
was studied in electrically conducting graphite dies as well as in
insulating alumina dies.
I. Introduction
II. Experimental Procedure
PARK plasma sintering (SPS) system is a combination of hot
Commercial powders of magnesium hydride (95 wt%, 5 wt%
excess Mg, Th. Goldschmidt AG, Essen, Germany) and graph-
ite (99.91 wt%, particle size—100 mm, Chempur GmbH,
Karlsruhe, Germany) were used as starting materials. Chemical
analysis of the magnesium hydride gave the following results (in
wt%): Mg—91.6170.15, H—6.9670.07, O—0.2770.010, N—
0.9270.08. Magnesium hydride and graphite powder (50 wt%
each) were blended using a vibrating mill with a polyethylene
capsule and SiAlON balls for 24 h. To avoid contamination with
oxygen during milling or mixing, the complete handling was
carried out in a dry argon atmosphere (argon 99.999%, Messer
Griesheim, Kirchen, Germany, H2O and O2 content o0.1 ppm).
The experiments were performed using an SPS system (Dr.
Sinters SPS 515-S, Sumitomo Coal Mining Co., Ltd., Tokyo,
Japan). The standard SPS setup was modified, and a second
thermocouple was placed inside the sample to analyze the tem-
perature difference between the wall of the die and the reaction
mixture. This thermocouple (type S, + 1.5 mm, Inconels-sheet-
ed) was plunged through the upper punch and fixed with a zir-
conia-based cement with the measuring tip 4 mm inside the
sample space (Fig. 1). The temperature was controlled by a dig-
ital programmable controller CHIO KP 1000 (Chino Corp.,
Tokyo, Japan). Temperature calibration in the range from 3001
to 9501C was carried out with three melting points of pure el-
ements (Pb, Al, Ge). All experiments were performed in dynam-
ic vacuum (PVacuumo3 Pa), and a uniaxial pressure of 14 MPa
on the die was applied during the entire process.
For reacting, magnesium hydride powder or powder mixture
was SPS heated stepwise from ambient temperature to 5001C at
different heating rates (varying from 2 to 20 K/min) and sub-
sequently annealed at this temperature for 10 min. For process
control, the temperature within the die (TD) was used. The rel-
evant parameters—die temperature TD, sample temperature TS,
gas pressure P inside the vacuum chamber, and displacement D
along the pressing direction—were recorded with the time in-
crement of 2 s. Graphite and alumina dies were used in order to
investigate the influence of the resistivity of the die material on
the decomposition process.
S
pressing and pulse current generation.1 This technique has
been developed for compaction of various metals, composites,
and, especially, ceramics.2,3 In the SPS method, current pulses
with a length of about 3 ms pass through the die and the sample
while mechanical pressure is applied. This procedure is suggest-
ed to generate sparks between powder particles due to the pulsed
current.1 However, the mechanism of the spark formation in
insulating materials is unknown.
The effects caused by SPS are summarized as follows:
(1) Destruction of the surface oxidation layers on metal
particles.4
(2) Promotion of neck growth between particles.1
(3) Cleaning of powder surfaces from adsorbed materials.
(4) Intensification of diffusion.
(5) Increase of sintering efficiency.
While effects (1) and (2) were confirmed for metal sintering,
they have not been sufficiently investigated for sintering of ce-
ramics, although several investigations were carried out in order
to explore the effect of pulse current on rapid sintering of insu-
lating materials in SPS, e.g., there are indications for an en-
hanced sintering of AlN.5 On the other hand, Tomino et al.6
measured the electric current passing through an alumina com-
pact in the SPS process and reported that the current was very
weak compared with the operating current applied on the sin-
tering tool with alumina powder. These results are in agreement
with microstructure investigations on Al2O3 bodies formed dur-
ing the similar plasma-activated sintering technique.7,8 The
densification of the compact was shown to be driven by Joule
heat transferred from the graphite die without any indication for
9
a plasma. The findings of indicate a difference between the
measured temperature of the die and the temperature inside the
sample, which depends on the material properties, especially the
electrical resistivity and the thickness of the sample.10 This tem-
perature difference complicates a straight comparison of the SPS
with other sintering techniques.
In order to determine the decomposition temperature of
MgH2, all SPS experiments were carried out at at least four dif-
ferent heating rates. Two different variables indicating the de-
composition (D and P) can be used for further analysis. The
displacement reflects the shrinkage of the sample and the ther-
mal expansion of the graphite punches, the spacers, and the
plungers. In combination with the experimental noise, this pre-
A. H. Carim—contributing editor
Manuscript No. 10823. Received January 28, 2004; approved January 18, 2005.
wPresent address: Fraunhofer Institute for Manufacturing and Advanced Materials,
Winterbergstr. 28, 01277 Dresden, Germany.
zAuthor to whom correspondence should be addressed. email: grin@cpfs.mpg.de
1870