Phase Relationship of Mg2Si at High Pressures and High Temperatures and Thermoelectric Properties of Mg9Si5

Article abstract of DOI:10.1021/acs.inorgchem.1c01386

Magnesium silicide (Mg2Si) is a promising eco-friendly thermoelectric material, which has been extensively studied in recent times. However, its phase behavior at high pressures and temperatures remains unclear. To this end, in this study, in situ X-ray diffraction analysis was conducted at high pressures ranging from 0 to 11.3 GPa and high temperatures ranging from 296 to 1524 K, followed by quenching. The antifluorite-phase Mg2Si decomposed to Mg9Si5 and Mg at pressures above 3 GPa and temperatures above 970 K. The antifluorite-phase Mg2Si underwent a structural phase transition to yield a high-pressure room-temperature (HPRT) phase at pressures above 10.5 GPa and at room temperature. This HPRT phase also decomposed to Mg9Si5 and Mg when heated at ～11 GPa. When 5Mg2Si decomposed to Mg9Si5 and Mg, the volume reduced by ～6%. Mg9Si5 synthesized at high pressures and high temperatures was quenchable under ambient conditions. Thermoelectric property measurements of Mg9Si5 at temperatures ranging from 10 to 390 K revealed that it was a p-type semiconductor having a dimensionless thermoelectric figure of merit (ZT) of 3.4 × 10-4 at 283 K.

Full text of DOI:10.1021/acs.inorgchem.1c01386

pubs.acs.org/IC
Article
Phase Relationship of Mg2Si at High Pressures and High
Temperatures and Thermoelectric Properties of Mg9Si5
Motoharu Imai,* Soshi Ibuka, and Yukihiro Isoda
Cite This: Inorg. Chem. 2021, 60, 11394−11400
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Magnesium silicide (Mg2Si) is a promising eco-
friendly thermoelectric material, which has been extensively studied
in recent times. However, its phase behavior at high pressures and
temperatures remains unclear. To this end, in this study, in situ X-
ray diffraction analysis was conducted at high pressures ranging
from 0 to 11.3 GPa and high temperatures ranging from 296 to
1524 K, followed by quenching. The antifluorite-phase Mg2Si
decomposed to Mg9Si5 and Mg at pressures above 3 GPa and
temperatures above 970 K. The antifluorite-phase Mg2Si under-
went a structural phase transition to yield a high-pressure room-
temperature (HPRT) phase at pressures above 10.5 GPa and at
room temperature. This HPRT phase also decomposed to Mg9Si5
and Mg when heated at ∼11 GPa. When 5Mg2Si decomposed to
Mg9Si5 and Mg, the volume reduced by ∼6%. Mg9Si5 synthesized at high pressures and high temperatures was quenchable under
ambient conditions. Thermoelectric property measurements of Mg9Si5 at temperatures ranging from 10 to 390 K revealed that it was
a p-type semiconductor having a dimensionless thermoelectric figure of merit (ZT) of 3.4 × 10−4 at 283 K.
4) at 7.5 GPa and then to a Ni2In-type (hexagonal, space
group: P63/mmc (No. 194), Z = 2) structure at 21.3 GPa.15
Zhu reported a transition from the antifluorite-type to a
monoclinic structure at 11 GPa, whose atomic coordinates and
structure type were unsolved.16 Furthermore, no phase
transition was reported up to 15 GPa for pure Mg2Si under
hydrostatic conditions using He as the pressure medium,
whereas the antifluorite-to-anticotunnite transition occurred at
11.9 GPa in an Al-doped Mg2Si.17 First-principles calculations
suggested the antifluorite-to-anticotunnite and anticotunnite-
to-Ni2In transitions.18,19
In this study, we investigated the phase behavior of Mg2Si at
pressures ranging from 0 to 11.3 GPa and temperatures
ranging from 300 to 1530 K via in situ XRD analysis using
synchrotron radiation and elucidated its phase relationships.
We selected this pressure−temperature (p−T) window
because it includes the p−T conditions required for the
quenching experiments, wherein the structural phase tran-
sitions from the antifluorite phase were reported. To clarify the
phase relationships, the recovered sample was characterized by
INTRODUCTION
Magnesium silicide (Mg2Si), which crystallizes to an
antifluorite (anti-CaF2-type) structure (cubic, space group:
̅
Fm3m (No. 225), Z = 4), is an eco-friendly material that
■
consists of nontoxic earth-abundant elements and is the only
compound among binary Mg−Si compounds formed at
ambient pressure.1 Mg2Si is an indirect band gap semi-
conductor with a band gap energy (Eg) of 0.6−0.8 eV.2,3 The
material and its solid solutions have been extensively studied as
promising environmentally friendly thermoelectric materials
that operate at temperatures ranging from 500 to 800 K.4,5
In fact, the values of the dimensionless thermoelectric figure
of merit (ZT) are greater than 1 in ternary and quaternary
Mg2Si-based compounds, such as Mg2Si1−xGex, Mg2Si1−xSnx,
and Mg2Si1−x−yGexSny,6−13 indicating that Mg2Si and its solid
solutions are suitable for practical use. Mg2Si has also been
investigated as a material for fabricating infrared detectors.3
Although the phase stability of Mg2Si at high pressures has
been studied since the 1960s, the phase behavior remains
unclear. Cannon et al. reported the synthesis of Mg2Si with a
hexagonal structure based on quenching experiments at
pressures above 2.5 GPa and temperatures above 1173 K.14
However, this crystal structure was unsolved. Although several
researchers have conducted X-ray diffraction (XRD) analysis at
high pressures and ambient temperatures using diamond anvil
cells, the results have been unsatisfactory. Hao reported that
Mg2Si transformed to an anticotunnite (anti-PbCl2-type)
structure (orthorhombic, space group: Pnma (No. 62), Z =
Received: May 7, 2021
Published: July 19, 2021
https://doi.org/10.1021/acs.inorgchem.1c01386
© 2021 American Chemical Society
Inorg. Chem. 2021, 60, 11394−11400
11394

Inorganic Chemistry
pubs.acs.org/IC
Article
analysis using Si (National Institute for Standards and Technology,
Si640c) as the external marker.
electron-probe microanalysis (EPMA) and XRD analysis and
compared with Mg9Si5 synthesized at high pressures and
temperatures. The thermoelectric properties of Mg9Si5 were
also investigated at temperatures ranging from 10 to 390 K.
Thermoelectric Property Measurements. The thermoelectric
properties, such as electrical resistivity, Seebeck coefficient, and
thermal conductivity, were measured at temperatures ranging from 10
to 390 K using the Physical Properties Measurement System
(Quantum Design Co.). The typical sample dimensions were 1 mm
× 1 mm × 6 mm.
EXPERIMENTAL METHODS
■
Synthesis of Mg2Si. Mg2Si was synthesized from Mg (99.9%
pure, Kojundo Chemical Laboratory Co. Ltd.) and Si (99.9999%
pure, Hirano Seizaemon Syouten Co. Ltd.) in a carbon boat by a
liquid−solid synthesis method. The synthesis details have been
described in a previous paper.20 The phases in the samples were
examined by powder XRD analysis using a Bragg−Brentano
diffractometer with the D/teX detector (Rigaku, RINT-TTR-III)
and Cu Kα radiation (1.5406 Å) in the 2θ range of 5.00−100.00° at
room temperature (297 K). We confirm that the synthesized sample
consisted of a single antifluorite phase. The lattice parameter was
estimated to be 6.354(1) Å, using Si (National Institute for Standards
and Technology, Si640c) as the external marker. This lattice
parameter was consistent with that previously reported.21−23 The
chemical compositions of the samples were evaluated by inductively
coupled plasma-optical emission spectroscopy (ICP-OES, Agilent,
720ES). The evaluated chemical composition was 63.48(6) wt % Mg
and 36.65(3) wt % Si, corresponding to 66.68(6) atom % Mg and
33.31(3) atom % Si, where the figures in the parentheses indicate the
standard deviations. This result indicated that the chemical
composition of our sample corresponded to the stoichiometric
composition.
RESULTS AND DISCUSSION
In Situ XRD Analysis of Mg2Si at High Pressures and
Temperatures. Figure 1 shows the XRD patterns obtained
■
In Situ XRD Analysis at High Pressures and Temperatures.
High pressures were applied using MAX80, the multianvil high-
pressure apparatus installed in the beamline of the advanced ring (AR-
NE5) at the National Laboratory for High Energy Physics. The 6−6
assembly24,25 was used with tungsten carbide (WC) anvils having
square flat surfaces with dimensions of 6 mm × 6 mm, 4 mm × 4 mm,
and 3 mm × 3 mm. The sample cell used in the 6 × 6 WC anvils is
described in the Supporting Information. Powdered Mg2Si was loaded
in the hexagonal BN (h-BN) capsule for the experiments at pressures
below 11 GPa and in the MgO capsule for the experiments at
pressures above 11 GPa. The capsule was set in the center of a
boron−epoxy pressure-transmitting medium. The sample was heated
by applying the current through a graphite tube heater set in the
sample cell. The temperature was measured using an alumel−chromel
thermocouple (Furuya Metal Co. Ltd) attached to the sample capsule.
The pressure was evaluated based on the lattice constant of the NaCl
internal pressure marker.26 The typical error in the pressure was 0.1
GPa. XRD patterns were obtained using an energy-dispersive method
via a white beam synchrotron radiation from the bending magnet. The
measurements were performed at intervals of ∼50 K, and the lattice
constants were obtained using the least-squares fitting of the indexed
patterns.
High-Pressure Synthesis. High-pressure synthesis was con-
ducted using a belt-type high-pressure apparatus. The details of the
high-pressure apparatus have been described in previous studies.27,28
The powdered material was loaded into the h-BN capsule.
Subsequently, the sample was pressurized at room temperature, at
∼297 K, under a target pressure of 5.75 GPa and then heated up to
1270 K. The errors in the pressure and temperature were 0.5 GPa and
50 K, respectively. The pressure and temperature conditions were
kept constant for 2 h; the samples were then cooled to room
temperature and returned to ambient pressure.
Figure 1. XRD patterns obtained when heating up to 1420 K at ∼5
GPa. Symbols B and G represent diffraction peaks originating from
the h-BN sample capsule and graphite heater, respectively. Refer to
the text for information on the + symbol.
when heating Mg2Si up to 1420 K at ∼5 GPa. Symbols “B” and
“G” represent the diffraction peaks from the h-BN sample
capsule and graphite heater set around the sample capsule,
respectively. The diffraction peaks from h-BN and graphite
were assigned by comparison between the measured XRD
patterns and the XRD patterns calculated using lattice
constants at high pressures estimated from previous
reports.29,30 When the sample was compressed at room
temperature, no phase transition from the antifluorite phase
was observed up to 5.3 GPa. However, when the sample was
heated under a constant load, several new diffraction peaks
appeared at 720 K, indicating that the structural phase
transition to a high-pressure high-temperature (HPHT)
phase starts at 720 K. The phase transition completed at
1080 K. Upon heating the sample further, the HPHT phase
melted at 1420 K. At this temperature, the diffraction pattern
shows peaks originating from the h-BN capsule and graphite
heater, together with the peak indicated by the “+” symbol.
This peak has been discussed later when describing the HPHT
phase.
Figure 2 shows the XRD patterns recorded when heating at
∼11 GPa. The symbol “M” indicates the diffraction peaks from
the MgO sample capsule. The diffraction peaks from MgO
were assigned by comparison between the measured XRD
patterns and the calculated ones using lattice constants at high
pressures estimated from a previous study.31 At 11.0 GPa and
room temperature, a new diffraction peak appears (marked by
a solid circle), indicating the onset of a structural phase
transition to a high-pressure room-temperature (HPRT)
Characterization. The samples recovered from the in situ XRD
analysis and those synthesized using the belt-type apparatus were
characterized using EPMA and powder XRD (Cu Kα radiation). The
chemical compositions of the samples were determined via an
electron-probe microanalyzer (JEOL JXA-8900), which was operated
at an accelerating voltage of 15 kV and a beam current of 50 nA. MgO
and Si were used as the standard materials. For EPMA, the samples
were mounted onto a resin and polished using an oil-based diamond
slurry. The lattice parameters were obtained from powder XRD
11395
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400

Inorganic Chemistry
pubs.acs.org/IC
Article
Figure 3. BSE image of the recovered sample from the in situ XRD
measurements. The sample was subjected to 4.2 GPa and 1420 K,
whose XRD patterns are shown in Figure S3 of the Supporting
Information.
Figure 2. XRD patterns obtained when heating at ∼11 GPa. M, G,
and solid circles indicate the diffraction peaks from the MgO sample
capsule, graphite heater, and the high-pressure room-temperature
(HPRT) phase, respectively. “+M” represents the diffraction peaks
from MgO capsule that overlapped with the peak from the sample.
phase. During heating, the number of reflections from the
HPRT phase increased, with five peaks appearing at 573 K.
However, these peaks cannot be indexed to anticotunnite or
monoclinic phases reported in the other studies.15,16 We could
not determine the crystal structure of the HPRT phase because
of the small number of diffraction peaks associated with this
phase, as described in the Supporting Information. At 778 K,
the diffraction peaks corresponding to the HPHT phase
appeared, and the sample consisted of a single HPHT phase at
975 K.
Figure 4. (a) XRD patterns of the sample synthesized using the belt-
type apparatus, (b) XRD patterns obtained at 0 GPa for the sample
subjected to 4.9 GPa and 1270 K and the simulated pattern of h-BN32
capsule, and (c) simulated XRD patterns of antifluorite-phase Mg2Si
and Mg9Si5. The peaks marked by asterisks and the plus sign
correspond to reflections from the antifluorite-phase Mg2Si and MgO,
respectively.
Characterization of HPHT Phase. Initially, we confirmed
that the HPHT phase could be quenched to ambient
conditions via in situ XRD analysis, as described in the
Supporting Information. Then, we investigated the chemical
composition of the quenched HPHT phase. A backscattered
electron (BSE) image of the recovered sample subjected to 4.2
GPa and 1420 K in the in situ experiments (the same sample in
Figure S3 of the Supporting Information) is shown in Figure 3.
The sample consists of two phases: one with a light gray
contrast and another with a dark gray contrast, corresponding
to Mg9Si5 and Mg2Si, respectively, as confirmed by EPMA.
This result suggested that the HPHT phase was Mg9Si5. To
confirm this finding, we synthesized Mg9Si5 at 5.75 GPa and
1270 K from a 1:0.1 molar mixture of Mg2Si and Si using the
belt-type apparatus and compared it with the recovered sample
used for the in situ observation. We used Mg2Si + 0.1Si as the
starting material for the high-pressure synthesis because Mg9Si5
approximately corresponds to Mg2Si1.11. Figure 4 shows the
angle-dispersive XRD pattern of the sample synthesized using
the belt-type apparatus and the XRD patterns of the sample
subjected to 4.9 GPa and 1270 K, obtained under ambient
conditions, and the calculated pattern of h-BN32 together with
the simulated XRD patterns of the antifluorite-phase Mg2Si
and Mg9Si5, as reported previously.33 The crystal structure of
Mg9Si5 was originally reported based on the results of TEM
observation of β′ phase precipitate in Al-Mg-Si alloys.34 The
crystal structure of the β′ phase is slightly different from that of
Mg9Si5 prepared by high-pressure synthesis. The simulated
diffraction pattern of Mg9Si5 prepared using high-pressure
synthesis is more in agreement with that of the sample
synthesized using the belt-type apparatus than that of the
Mg9Si5 precipitates, as shown in the Supporting Information.
The peaks marked by asterisks in pattern (a) are the diffraction
peaks from the antifluorite-phase Mg2Si. Since pattern (b) was
obtained using the energy-dispersive method at 2θ = 6.00°, we
converted the photon energy values to 2θ values for
comparison using the relation sin θ = sin(3.00°)λCu KαE/hc,
where λCu Kα, E, h, and c are the wavelength of Cu Kα
radiation, photon energy, Planck constant, and speed of light,
respectively. Pattern (a) is consistent with pattern (b),
although pattern (b) contains several diffraction peaks
originating from the h-BN sample capsule. Herein, we discuss
the small peak marked by the + symbol. This peak appeared
simultaneously with those originating from Mg9Si5 in the
heating process, as shown in Figure 1, and during quenching to
ambient conditions, as shown in Figure S3 of the Supporting
Information. The 2θ value of the peak marked by + at 0 GPa
(Figure 4b) agrees with that of the 200 peak of MgO.35
Furthermore, the d-value of the + peak in Figure 1 is consistent
with that of the 200 peak of MgO at 4.2 GPa and 1420 K
estimated using the equation of state of MgO.31 Thus, we
11396
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400

Inorganic Chemistry
pubs.acs.org/IC
Article
concluded that the peak marked by + corresponded to the 200
peak from the MgO formed from the oxidization of Mg
resulting from the decomposition of Mg2Si. Based on the
above results, we concluded that the substance believed to be
the HPHT phase of Mg2Si was a mixture of Mg9Si5 and Mg,
indicating that the antifluorite-phase Mg2Si decomposed to
Mg9Si5 and Mg at high pressures and high temperatures. This
result is qualitatively consistent with the first-principles
calculation results, stating that Mg2Si can transform to
Mg9Si5.19 The lattice parameters a and c of Mg9Si5 were
12.410(1) and 12.345(1) Å, respectively, which are consistent
with previously reported ones (12.411(1) and 12.345(1) Å).33
Interestingly, the d-values of the hexagonal Mg2Si, synthesized
in previous studies, involving quenching experiments,14 are
consistent with those of Mg9Si5 observed in this study. Thus,
we believe that the hexagonal Mg2Si synthesized by Cannon et
al. was Mg9Si5.
temperatures can be quenched to ambient conditions,
indicating that energy is necessary for the reverse reaction,
wherein Mg9Si5 decomposes to Mg2Si. Thus, heat energy is
necessary for this decomposition, although Huan claimed that
Mg2Si can transform to Mg9Si5 without requiring energy.19
Volume of Mg2Si at High Pressures and High
Temperatures. The volume per chemical formula unit of
the antifluorite-phase Mg2Si at room temperature is plotted as
a function of pressure in Figure 6. The volume was normalized
p−T Diagram of Mg2Si. The aforementioned results are
summarized in the p−T diagram shown in Figure 5. We used a
Figure 6. Volume per chemical formula unit of the antifluorite-phase
Mg2Si at room temperature as a function of pressure. The volume was
normalized with values under ambient conditions. The solid and
dashed lines are fitted to the Murnaghan equation with fixed B0′ (4)
and variable B0′ values, respectively.
Table 1. Bulk Modulus B0 and Its Pressure Derivative, B0′,
of the Antifluorite-Phase Mg2Si
B0 (GPa)
B0′
ref
51.9(7)
46(2)
57.03(2)
58(2)
4(fix)
6.2(7)
4(fix)
4(fix)
present
present
15
Figure 5. p−T diagram of Mg2Si. “HPRT” stands for the HPRT
phase. The dashed lines are guides to the eye.
16
previously determined melting point value1 as the temperature
at 0 GPa. Mg2Si decomposed to Mg9Si5 and Mg at above 3
GPa and 970 K. The HPRT phase appeared at room
temperature and pressures over 10 GPa. The HPRT phase
also decomposed to Mg9Si5 and Mg when heated at ∼11 GPa.
The melting point decreased slightly at pressures up to 3 GPa,
and then increased at higher pressures.
The decomposition temperature was nearly the same at
pressures ranging from 5 to 10 GPa, i.e., ∼700 K. In general,
when the first-order phase transition occurs thermodynami-
cally, the pressure and temperature satisfy the Clausius−
Clapeyron relationship: dP/dT = ΔS/ΔV, where ΔS and ΔV
are the differences in the entropy and volume between the
initial and final states, respectively. In this case, we considered
its inverse, i.e., dT/dP, because temperatures are plotted as a
function of pressure in Figure 5. Since the volume difference
during the decompositions is ∼6%, as described later, dT/dP is
expected to have a nonzero value. However, the observed dT/
dP value is negligible because the decomposition occurs at
approximately the same temperature. This discrepancy
suggests that the decomposition occurs kinetically, and heat
energy is necessary for the decomposition to occur.
Furthermore, Mg9Si5 synthesized at high pressures and high
with the volume under ambient conditions. Table 1 lists the
values of the bulk modulus B0 and its pressure derivative B0′,
obtained by fitting the data to the Murnaghan equation.36 The
B0 value obtained in this study is slightly lower than previously
reported ones, probably because a solid pressure medium was
used in this study, whereas a liquid pressure medium was used
in previous studies.15,16
We melted most of the samples at high pressures in our in
situ observations. When the melted samples were cooled to
room temperature, they changed from a powdery to a granular
form. Consequently, we found it difficult to obtain XRD
patterns suitable for indexing and determining the lattice
parameters of Mg9Si5 at high pressures. Therefore, it is difficult
to obtain the B0 value of Mg9Si5 by fitting the volume data to
the Murnaghan equation, as in the case of antifluorite-phase
Mg2Si. Instead, we estimated the B0 of Mg9Si5 using the
diffraction patterns shown in Figure S3 of the Supporting
Information and the following relationship: B0 ≈ [(1/V0)(ΔV/
ΔP)]−1, which is the approximate expression of B0, B0 = [(1/
V0)(dV/dP)]−1. The B0 value of Mg9Si5 was estimated to be
75.6 GPa from the unit cell volumes of 1633(4) Å3 at 0.0 GPa
and 1556(2) Å3 at 3.1 GPa, which is larger than that of
antifluorite-phase Mg2Si.
11397
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400

Inorganic Chemistry
pubs.acs.org/IC
Article
Figure 7 presents the volumes of 5Mg2Si and Mg9Si5 + Mg
at various pressures as functions of temperatures. It should be
Figure 7. Volumes of 5Mg2Si and Mg9Si5 + Mg at various pressures as
functions of temperatures. The dashed lines are guides to the eye.
Figure 8. (a) Electrical resistivity ρ, (b) Seebeck coefficient S, and (c)
thermal conductivity κ as functions of temperatures. The inset of (a)
is a plot of electrical conductivity, σ, as a function of reciprocal
temperature 1/T. The y-axis of the inset is a logarithmic scale.
noted that 5Mg2Si and Mg9Si5 + Mg contain 10 mol Mg and 5
mol Si. The volume per chemical formula unit of Mg2Si at 0
GPa was calculated using 6.354(1) Å as the lattice parameter.
The thermal expansion data of Mg2Si at 0 GPa were obtained
from a previous study.21 The atomic volume of Mg at a given
pressure and temperature was calculated using the equations of
states of Mg.37 The data at 11 GPa are not connected by the
dashed line because the crystal structure of the HPRT phase is
unknown.
At the transition stage, the volume reduced by ∼6% (6.4,
6.4, 6.7, and 6.0% at ∼4, 5, 6, and 8 GPa, respectively). Thus,
Mg9Si5 is denser than Mg2Si, which is consistent with the fact
that Mg9Si5 has a higher B0 value than Mg2Si, indicating that
Mg9Si5 is harder than Mg2Si.
Thermoelectric Properties of Mg9Si5. Mg9Si5 was
synthesized from a 1:0.1 molar mixture of Mg2Si and Si
using the belt-type high-pressure apparatus at 5.75 GPa and
1270 K. The electrical resistivity, ρ, is plotted as a function of
temperature in Figure 8a. The ρ value exhibits semiconducting
temperature dependence, i.e., it decreases with increasing
temperature. The inset in Figure 8a is a plot of the electrical
conductivity, σ, which is the inverse of ρ, as a function of the
reciprocal temperature (1/T). The values of σ ranging from
223 to 388 K can be fitted to the equation σ = σ0 exp(−Ea/
kBT); the Ea value is 44 meV. Further, the Seebeck coefficient,
S, is plotted as a function of temperature in Figure 8b. S has
positive values at all measured temperatures, indicating that the
dominant carriers are holes. S has a maximum value of 201.8
mV/K at 190 K. The thermal conductivity, κ, is relatively low.
The electronic and lattice thermal conductivity values are
provided in the Supporting Information. Figure 9 shows the
dimensionless thermoelectric figure of merit (ZT), defined as
ZT = S2/κρ. ZT has a maximum value of 3.4 × 10−4 at 283 K.
This low ZT value is responsible for the high ρ value. These
results indicate that reducing ρ is necessary for enhancing the
ZT.
Figure 9. Dimensionless thermoelectric figure of merit, ZT, as a
function of temperature.
Singh et al. calculated the band structure of Mg9Si5 using the
linearized augmented plane-wave method.38 They used a
modified Becke−Johnson functional for the exchange
correlation. The calculation suggested that Mg9Si5 was a
narrow band gap semiconductor having a band gap of 170
meV. We obtained an Ea value of 44 meV during the σ
measurements. When electric conduction is assumed to occur
in the intrinsic region, Ea is equal to half the band gap energy,
Eg39 (88 meV in this case). Thus, Mg9Si5 is a narrow band gap
semiconductor as suggested by the calculation.
CONCLUSIONS
■
The phase behavior of Mg2Si was investigated at high pressures
ranging from 0 to 11.3 GPa and high temperatures ranging
from 296 to 1524 K, using a combination of in situ XRD
analysis and quenching experiments. The antifluorite-phase
Mg2Si decomposed to Mg9Si5 and Mg at pressures over 3 GPa
and temperatures above 970 K. Additionally, the antifluorite-
11398
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400

Inorganic Chemistry
pubs.acs.org/IC
Article
Funding
phase Mg2Si underwent a structural phase transition to a
HPRT phase at room temperature and pressures above 10.5
GPa. This HPRT phase decomposed to Mg9Si5 when heated
under a pressure of ∼11 GPa. The thermoelectric property
measurements of Mg9Si5 synthesized at 5.3 GPa and 1273 K
revealed that it was a p-type semiconductor having a
dimensionless thermoelectric figure of merit (ZT) of 3.4 ×
10−4 at 283 K.
This work was partially supported by the Japan Society for the
Promotion of Science through its “Funding Program for
World-Leading Innovative R&D on Science and Technology
(FIRST Program)”.
Notes
The authors declare no competing financial interest.
The results obtained herein clearly demonstrate that two
condensed states exist: one is a mixture of Mg9Si5 and Mg,
while the other is the HPRT phase, at high pressures ranging
from 0 to 11.3 GPa and high temperatures ranging from 296 to
1524 K. We also revealed that the hexagonal Mg2Si reported in
the studies during the 1960s14 was Mg9Si5.
However, we could obtain only limited information
regarding the HPRT phase. Further investigations are
necessary to determine the characteristics of the HPRT
phase. Powder XRD analysis, using diamond anvil cells along
with a heating system and synchrotron radiation, is an effective
approach for clarifying the properties of the HPRT phase
because it is necessary to heat Mg2Si at temperatures ranging
from 370 to 570 K to accelerate the sluggish phase transition
occurring at ∼11 GPa.
ACKNOWLEDGMENTS
■
The authors thank Akira Ishitoya of the National Institute for
Material Science (NIMS) for conducting inductively coupled
plasma-optical emission spectroscopy, Mitsuaki Nishio of
NIMS for the EPMA measurements, Hiroaki Mamiya of
NIMS for allowing the use of Physical Properties Measurement
System, Masashi Miyakawa of NIMS for his support during
high-pressure synthesis, and Takashi Taniguchi of NIMS for
allowing the use of the belt-type high-pressure apparatus and
excellent advice on high-pressure synthesis. The authors also
thank Takumi Kikegawa of the National Laboratory for High
Energy Physics for his support with the in situ XRD analysis
using synchrotron radiation, and Norimasa Nishiyama of
Ehime University (currently Sumitomo Electric) for his
excellent advice on how to use the 6−6 assembly. This work
was performed under the approval of the Photon Factory
Program Advisory Committee (Proposal Nos. 2011G094 and
2013G074).
This study should enable a better understanding of the
structure and properties of Mg2Si and its related materials and
increase the application scope of Mg2Si-based thermoelectric
devices and infrared detectors.
REFERENCES
■
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.inorgchem.1c01386.
■
(1) Kevorkov, D.; Schmid-Fetzer, R.; Zhang, F. Phase Equilibria and
Thermodynamics of the Mg-Si-Li System and Remodeling of the Mg-
Si System. J. Phase Equilib. Diffus. 2004, 25, 140−151.
(2) Borisenko, V. Semiconducting Silicides: Basics, Formation,
Properties; Springer: Berlin, Heidelberg, 2000.
(3) Udono, H.; Tajima, H.; Uchikoshi, M.; Itakura, M. Crystal
Growth and Characterization of Mg2Si for IR-Detectors and
Thermoelectric Applications. Jpn. J. Appl. Phys. 2015, 54, No. 07JB06.
(4) Zaitsev, V. K. Thermoelectrics on the Base of Solid Solution of
Mg2BIV Compounds (BIV = Si, Ge, Sn). In Thermoelectric Handbook
Macro to Nano; Rowe, D. M., Ed.; CRC Taylor & Francis, 2006.
(5) Nozariasbmarz, A.; Agarwal, A.; Coutant, Z. A.; Hall, M. J.; Liu,
sı
Sample cells for anvils with square flat surfaces of 6 mm
× 6 mm, discussion on HPRT phase, powder XRD
patterns of Mg2Si in cooling and depressurization
process, powder XRD patterns of Mg9Si5, and thermal
conductivity of Mg9Si5 (PDF)
AUTHOR INFORMATION
Corresponding Author
Motoharu Imai − National Institute for Materials Science,
Research Center for Functional Materials, Tsukuba 305-
0047, Japan; orcid.org/0000-0002-5848-113X;
Email: IMAI.Motoharu@nims.go.jp
■
̈
J.; Liu, R.; Malhotra, A.; Norouzzadeh, P.; Ozturk, M. C.; Ramesh, V.
̈
P.; Sargolzaeiaval, Y.; Suarez, F.; Vashaee, D. Thermoelectric Silicides:
A Review. Jpn. J. Appl. Phys. 2017, 56, No. 05DA04.
(6) Zaitsev, V. K.; Fedorov, M. I.; Gurieva, E. A.; Eremin, I. S.;
Konstantinov, P. P.; Samunin, A. Y.; Vedernikov, M. V. Highly
Effective Mg2Si1‑xSnx Thermoelectrics. Phys. Rev. B 2006, 74,
No. 045207.
Authors
(7) Isoda, Y.; Nagai, T.; Fujiu, H.; Imai, Y.; Shinohara, Y.
Thermoelectric Properties of Sb-Doped Mg2Si0.5Sn0.5, 2006 25th
International Conference on Thermoelectrics; IEEE, 2006; pp 406−
410.
(8) Zhang, Q.; He, J.; Zhu, T. J.; Zhang, S. N.; Zhao, X. B.; Tritt, T.
M. High Figures of Merit and Natural Nanostructures in Mg2Si0.4Sn0.6
based thermoelectric materials. Appl. Phys. Lett. 2008, 93,
No. 102109.
Soshi Ibuka − National Institute for Materials Science,
Research Center for Functional Materials, Tsukuba 305-
0047, Japan
Yukihiro Isoda − National Institute for Materials Science,
Center for Green Research on Energy and Environmental
Materials, Tsukuba 305-0047, Japan
(9) Liu, W.; Tang, X. F.; Sharp, J. Low-Temperature Solid State
Reaction Synthesis and Thermoelectric Properties of High-Perform-
ance and Low-Cost Sb-Doped Mg2Si0.6Sn0.4. J. Phys. D: Appl. Phys.
2010, 43, No. 085406.
(10) Nolas, G. S.; Wang, D.; Beekman, M. Transport Properties of
Polycrystalline Mg2Si1‑ySby (0 ≤ y <0.4). Phys. Rev. B 2007, 76,
No. 235204.
(11) Gao, H. L.; Zhu, T. J.; Liu, X. X.; Chen, L. X.; Zhao, X. B. Flux
Synthesis and Thermoelectric Properties of Eco-Friendly Sb Doped
Mg2Si0.5Sn0.5 Solid Solutions for Energy Harvesting. J. Mater. Chem.
2011, 21, 5933−5937.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.inorgchem.1c01386
Author Contributions
M.I.: conceptualization, supervision, project administration,
funding acquisitions, formal analysis, validation, data curation,
visualization, writingoriginal draft, writingreview and
editing; S.I.: investigation, writingreview and editing; and
Y.I.: resources, formal analysis. All authors have given their
approval for the final version of the manuscript.
11399
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400

Inorganic Chemistry
pubs.acs.org/IC
Article
(12) Noda, Y.; Kon, H.; Furukawa, Y.; Nishida, I. A.; Masumoto, K.
Temperature Dependence of Thermoelectric Properties of
Mg2Si0.6Ge0.4. Mater. Trans., JIM 1992, 33, 851−855.
(13) Khan, A. U.; Vlachos, N. V.; Hatzikraniotis, E.; Polymeris, G.
S.; Lioutas, C. B.; Stefanaki, E. C.; Paraskevopoulos, K. M.;
Giapintzakis, I.; Kyratsi, T. Thermoelectric Properties of Highly
Efficient Bi-Doped Mg2Si1‑x‑ySnxGey Materials. Acta Mater. 2014, 77,
43−53.
(33) Ji, S. D.; Imai, M.; Zhu, H. K.; Yamanaka, S. Structural
Characterization of Magnesium-Based Compounds Mg9Si5 and
Mg4Si3Al (Superconductor) Synthesized under High Pressure and
High Temperature Conditions. Inorg. Chem. 2013, 52, 3953−3961.
(34) Vissers, R.; van Huis, M. A.; Jansen, J.; Zandbergen, H. W.;
Marioara, C. D.; Andersen, S. J. The Crystal structure of the β′ phase
in Al-Mg-Si Alloys. Acta Mater. 2007, 55, 3815−3823.
(35) Alper, A. M.; McNally, R. N.; Ribbe, P. H. K.; Doman, R. C.
The System MgO-MgAl2O4. J. Am. Ceram. Soc. 1962, 45, 263−268.
(36) Murnaghan, F. D. Elastic Deformation of an Elastic Solids;
Dover: New York, 1967.
(37) Courac, A.; Le Godec, Y.; Solozhenko, V. L.; Guignot, N.;
Crichton, W. A. Thermoelastic Equation of State and Melting of Mg
Metal at High Pressure and High Temperature. J. Appl. Phys. 2020,
127, No. 055903.
(38) Singh, V.; Pulikkotil, J. J.; Auluck, S. Mg9Si5: A Potential Non-
Toxic Thermoelectric Material for Mid-Temperature Applications.
RSC Adv. 2016, 6, 62445−62450.
(39) Seeger, K. Semiconductor Physics, 5th ed.; Springer-Verlag:
Berlin, 1991.
(14) Cannon, P.; Conlin, E. T. Magnesium Compounds: New
Dense Phases. Science 1964, 149, 487−489.
(15) Hao, J.; Zou, B.; Zhu, P. W.; Gao, C. X.; Li, Y. W.; Liu, D.;
Wang, K.; Lei, W. W.; Cui, Q. L.; Zou, G. T. In Situ X-ray
Observation of Phase Transitions in Mg2Si under High Pressure. Solid
State Commun. 2009, 149, 689−692.
(16) Zhu, F.; Wu, X.; Qin, S.; Liu, J. A Re-Investigation on Pressure-
Induced Phase Transition of Mg2Si. Solid State Commun. 2012, 152,
2160−2164.
(17) Zhao, J. B.; Liu, Z. X.; Gordon, R. A.; Takarabe, K.; Reid, J.;
Tse, J. S. Pressure-Induced Phase Transition and Electrical Properties
of Thermoelectric Al-Doped Mg2Si. J. Appl. Phys. 2015, 118,
No. 145902.
(18) Yu, F.; Sun, J.-X.; Yang, W.; Tian, R.-G.; Ji, G.-F. A Study of the
Phase Transitions, Electronic Structures and Optical Properties of
Mg2Si under High Pressure. Solid State Commun. 2010, 150, 620−
624.
(19) Huan, T. D.; Tuoc, V. N.; Le, N. B.; Minh, N. V.; Woods, L. M.
High-Pressure Phases of Mg2Si from First Principles. Phys. Rev. B
2016, 93, No. 094109.
(20) Isoda, Y.; Tada, S.; Shioda, N.; Shinohara, Y. Si-Particle-Size
Dependence of Mg2Si Powders Directly Produced by Liquid−Solid
Synthesis Method. J. Alloys Compd. 2016, 656, 598−603.
(21) Imai, M.; Isoda, Y.; Udono, H. Thermal Expansion of
Semiconducting Silicides β-FeSi2 and Mg2Si. Intermetallics 2015, 67,
75−80.
(22) Labotz, R. J.; Mason, D. R.; O’Kane, D. F. The Thermoelectric
Properties of Mixed Crystals of Mg2GexSi1‑x. J. Electrochem. Soc. 1963,
110, 127−134.
(23) Wu, H.; Zhou, W.; Udovic, T. J.; Rush, J. J. Structure and
Hydrogenation Properties of the Ternary Alloys Ca2‑xMgxSi (0 ≤ x ≤
1). J. Alloys Compd. 2007, 446−447, 101−105.
(24) Nishiyama, N.; Wang, Y.; Sanehira, T.; Irifune, T.; Rivers, M. L.
Development of the Multi-Anvil Assembly 6-6 for DIA and D-DIA
Type High-Pressure Apparatuses. High Pressure Res. 2008, 28, 307−
314.
(25) Kawazoe, T.; Nishiyama, N.; Nishihara, Y.; Irifune, T. Pressure
Generation to 25GPa Using a Cubic Anvil Apparatus with a Multi-
Anvil 6−6 Assembly. High Press. Res. 2010, 30, 167−174.
(26) Brown, J. M. The NaCl Pressure Standard. J. Appl. Phys. 1999,
86, 5801−5808.
(27) Yamaoka, S.; Akashi, M.; Kanda, H.; Osawa, T.; Taniguchi, T.;
Sei, H.; Fukunaga, O. Development of Belt Type High Pressure
Apparatus for Material Synthesis at ∼8 GPa. J. High Press. Inst. Jpn.
1992, 30, 249−258.
(28) Kanke, Y.; Akaishi, M.; Yamaoka, S.; Taniguchi, T. Heater Cell
for Materials Synthesis and Crystal Growth in the Large Volume High
Pressure Apparatus at 10 GPa. Rev. Sci. Instrum. 2002, 73, 3268−
3270.
(29) Fujinaga, Y.; Shono, Y. Pressure calibration by Measuring
Lattice Constants of Hexagonal Boron Nitride. J. Jpn. Inst. Metals
1995, 59, 519−524. (in Japanese).
(30) Zhao, Y. X.; Spain, I. L. X-ray Diffraction Data for Graphite to
20 GPa. Phys. Rev. B 1989, 40, 993−997.
(31) Tange, Y.; Nishihara, Y.; Tsuchiya, T. Unified Analyses for P-V-
T Equation of State of MgO: A Solution for Pressure-Scale Problems
in High P-T Experiments. J. Geophys. Res. 2009, 114, No. B03208.
(32) Villars, P.; Cenzual, K. Pearson’s Crystal Data: Crystal Structure
Database for Inorganic Compounds, Release 2020/21, No. 1250502;
ASM International: Materials Park, Ohio, USA, 2007.
11400
https://doi.org/10.1021/acs.inorgchem.1c01386
Inorg. Chem. 2021, 60, 11394−11400