By using the first-principles calculations in combination with the Boltzmann transport theory, we systematically study the thermoelectric properties of AlX (X = S, Se, Te) monolayers as indirect gap semiconductors. The unique electronic density of states, which consists of a rather sharp peak at the valence band maxima and an almost constant band at the conduction band minima, makes AlX (X = S, Se, Te) monolayers excellent thermoelectric materials. The optimized power factors at room temperature are 22.59, 62.59, and 6.79 mW m-1 K-2 under reasonable electronic concentration for AlS, AlSe, and AlTe monolayers, respectively. The figure of merit (zT) increases with temperature and the optimized zT values of 0.52, 0.59, and 0.26 at room temperature are achieved under moderate electronic concentration for AlS, AlSe, and AlTe monolayers, respectively, indicating that two-dimensional layered AlX (X = S, Se, Te) semiconductors, especially AlSe, can be potential candidate matrices for high-performance thermoelectric nanocomposites.
By using the first-principles calculations in combination with the Boltzmann transport theory, we systematically study the thermoelectric properties of AlX (X = S, Se, Te) monolayers as indirect gap semiconductors. The unique electronic density of states, which consists of a rather sharp peak at the valence band maxima and an almost constant band at the conduction band minima, makes AlX (X = S, Se, Te) monolayers excellent thermoelectric materials. The optimized power factors at room temperature are 22.59, 62.59, and 6.79 mW m-1 K-2 under reasonable electronic concentration for AlS, AlSe, and AlTe monolayers, respectively. The figure of merit (zT) increases with temperature and the optimized zT values of 0.52, 0.59, and 0.26 at room temperature are achieved under moderate electronic concentration for AlS, AlSe, and AlTe monolayers, respectively, indicating that two-dimensional layered AlX (X = S, Se, Te) semiconductors, especially AlSe, can be potential candidate matrices for high-performance thermoelectric nanocomposites.
With the
global energy crisis and environmental impact of fossil fuels, a compelling
need exists for thermoelectric (TE) materials which can directly convert
waste heat into electric power. Generally, the thermally driven electrical
performance of TE materials is measured by the power factor PF = S2σ, in which S and σ
are, respectively, Seebeck coefficient and electrical conductivity.
The conversion efficiency of TE materials is gauged by the dimensionless
figure of merit zT = S2σT/κ, where T is the
absolute temperature, κ is the sum of electronic (κe) and lattice (κl) thermal conductivity.
A high-performance TE material not only requires a high power factor
but also has to possess a low thermal conductivity. Unfortunately,
it is complex to efficiently achieve waste heat recovery for the interdependency
of transport parameters (S, σ, and κe). S and σ unusually interweave and
behave in an opposite trend.[1] The improvement
of TE devices thus strongly depends on the optimization of electronic
and thermal transport properties. A strategy for achieving high TE
performance is to identify materials with intrinsically high TE performance,
which can be applied as potential matrices for high-performance TE
nanocomposites. Since the arrival of graphene,[2] the large number of the two-dimensional (2D) layered materials,
such as black phosphorus (BP),[3,4] transition metal dichalcogenides
(TMDs),[5−8] and group III metal chalcogenides (GIIIAMCs),[9−13] have been synthesized and characterized.
Unlike graphene, these 2D layered materials are semiconductors and
can potentially play key roles in the fabrication of next-generation
nano-electronic devices. In particular, these 2D semiconductors could
alleviate the coupling between S and σ because
of the quantum confinement effects, consequently enhancing TE performance.[14−18]Indeed, TMDs such as MoS2, MoSe2, WS2, and WSe2 exhibit thickness-dependent
TE properties[19] because of the thickness-dependent
electronic band structure[20,21] and maximum power factor
of about 340 and 150 mW m–1 K–2 for n-type monolayer MoSe2 and p-type MoS2 monolayers,[22] respectively, which are
much higher than those of 20 and 30 mW m–1 K–2 for n-type bulk MoSe2 and p-type bulk
MoS2, respectively. Experimentally, the Seebeck coefficient
up to 30 mV K–1 for single-layer MoS2 at room temperature is observed,[23] which
is significantly larger than the one observed in bulk MoS2 (∼7 mV K–1),[24] and moreover can be tuned between 0.4 and 100 mV K–1 by an external electric field using a field effect transistor.[25] As for BP, theoretical calculations in different
approaches indicate that the BP monolayer shows a strong spatial anisotropy
in electrical and thermal conductivities, which makes zT in the armchair direction larger than that in the zigzag direction,[26,27] and using the electric-double-layer transistor configuration experimentally,
the Seebeck coefficient of ion-gated BP reached 510 μV K–1 at 210 K in the hole-depleted state,[28] which is much higher than the reported 335 μV K–1 of single crystal BP at 300 K.[29]Among promising TE 2D layered semiconductors, monolayer
layered GIIIAMCs such as InSe, GaS, GaSe, or GaTe have entered the
spotlight. These monolayer layered semiconductors have unique electronic
structures, having a flat band at the valence band maxima (VBM) and
a parabolic band at the conduction band minima (CBM), thus resulting
in a rather sharp peak at VBM and an almost constant band at CBM in
the density of states (DOS).[30,31] Experimentally, the
monolayer or few layers of InSe,[11,32,33] GaS,[12,34,35] GaSe,[13,36−39] and GaTe[10] have
been synthesized, and much effort has been dedicated to reveal the
electronic and optical properties and potential applications in many
fields.[9] It was recently reported that
the carrier mobility in few-layer InSe was more than 103 cm2 V–1 s–1 at room
temperature,[40] which is much larger than
321 cm2 V–1 s–1 of
n-type InSe polycrystalline samples with the PF of about 250 μW
m–1 K–2 and zT of 0.42 at 700 K.[41] The thickness-dependent
TE properties of GaS, GaSe, InS, InSe monolayers have been theoretically
investigated based on a constant relaxation time,[42] and by more precise relaxation time calculation, the InSe
monolayer has both a large S and a large σ,
giving the maximum PF of 49 (43) mW m–1 K–2 for the p-type (n-type) InSe monolayer at room temperature in the
armchair direction.[43] Otherwise, it was
theoretically reported that monolayer InSe exhibited unusual phonon
behavior and ultralow thermal conductivities.[44] The TE performance of two-dimensional layered semiconductor InSe
was further effectively enhanced by reducing film thickness and modulating
electron density.[45] The study on the temperature-dependent
TE properties of GaS, GaSe, and GaTe monolayers by the Boltzmann transport
theory based on constant relaxation time showed the high zT of GaS, GaSe, and GaTe monolayers at low temperature.[46] Thus, it seems that monolayer GIIIAMCs such
as InSe, GaS, GaSe, or GaTe are promising matrices for high-efficient
TE nanocomposites.Recently, other group III–VI monolayers
such as AlX (X = S, Se, Te) have been proposed to have similar band
structures with InSe, GaS, GaSe, or GaTe monolayers and high mechanical
stability.[47] We thus expect that AlX (X
= S, Se, Te) monolayers could be suitable candidates for applications
in electronic and TE devices. However, relevant works about κl and TE properties of AlX (X = S, Se, Te) are lacking till
date. In this work, we present a detailed study on TE properties of
2D layered semiconducting AlX (X = S, Se, Te) by using first-principles
in combination with the Boltzmann transport theory, which enables
systematic analysis and comparison on TE performance for possible
III–VI 2D semiconducting configurations.
Results and Discussion
Crystal
and Electronic Structures
The structure of AlX (X = S, Se,
Te) monolayers is modeled based on the geometry of InSe, GaSe, and
GaS monolayers, which have already been synthesized experimentally,[12,37,40] for they all belong to group
III binary monolayers. All the AlX (X = S, Se, Te) compounds crystallize
in a honeycomb structure belonging to space group D3. Figure presents the top view and side view of the
monolayer AlX unit cell which consists of four atoms (two Al and two
X atoms). In a monolayer AlX, there exist four sublayers which stack
in the order of X–Al–Al–X with X = S, Se, and
Te. After fully relaxing the internal coordinates of the atoms, the
optimized configurations which have the lowest energy are considered. Table summarizes the calculated
structural parameters, together with available theoretical results.
The optimized lattice parameters (a) for AlS, AlSe, and AlTe monolayers
are respectively 3.46, 3.70, and 4.10 Å, which compare favorably
with those reported previously.[47] The calculations
show that our approach is reliable for these monolayers, which is
the precondition to accurately predict the transport coefficients.
In AlX (X = S, Se, Te) monolayers, the distance between Al and X atoms
(dAl–X) increases as the atomic
radius of the X element increases. The optimized lattice parameters
and the distance between X and X atoms (dX–X) exhibit a similar trend as dAl–X. Otherwise, the distance between Al and Al atoms (dAl–Al) shows a downward trend and is independent
of the type of X atoms. By fitting the total energy (E) of the system to the lattice dilation of Δl/l0, the two-dimensional elastic modulus
(C2D) is listed in Table .
Figure 1
Ball-and-stick model
of the AlX (X = S, Se, Te) monolayers from the top view (a) and side
view (b).
Table 1
Optimized Lattice Constant (a), Bond
Length (dAl–Al, dX–X, dAl–X),
Bond Angle (θ), and Band Gap (Eg) of AlX (X = S, Se, Te) Monolayersa
AlX
a (Å)
dAl–Al (Å)
dX–X (Å)
dAl–X (Å)
θ (deg)
Eg (eV)
AlS
3.46 (3.57)
2.59 (2.59)
4.82 (4.73)
2.29 (2.32)
98.22 (100.4)
2.05 (2.10)
AlSe
3.70 (3.78)
2.57 (2.57)
4.95 (4.90)
2.45 (2.47)
98.31 (99.75)
2.01 (1.99)
AlTe
4.10
(4.11)
2.55 (2.58)
5.11 (5.14)
2.69 (2.70)
99.30 (99.30)
1.86
(1.84)
The values in brackets are theoretical
results in ref (47).
Table 2
Elastic
Constant (C2D), Effective Mass (m*), DP Constant (E1), Carrier
Mobility (μ), and Relaxation Time (τ) of AlX (X = S, Se,
Te) Monolayers at Room Temperature
AlX
carrier type
C2D (N/m)
m* (me)
E1 (eV)
μ (×103 cm2 V–1 s–1)
τ (fs)
AlS
electron
125.23
1.45
0.85
1.65
1357.37
hole
125.23
2.68
4.47
0.03
44.04
AlSe
electron
106.03
2.30
1.09
0.46
600.73
hole
106.03
3.74
2.65
0.02
46.52
AlTe
electron
86.59
2.72
1.39
0.15
239.22
hole
86.59
4.87
0.80
0.03
96.86
Ball-and-stick model
of the AlX (X = S, Se, Te) monolayers from the top view (a) and side
view (b).The values in brackets are theoretical
results in ref (47).Figure a–c show the band structures and total
DOS for AlX monolayers with X = S, Se, Te, indicating that all the
monolayers exhibit as indirect band gap semiconductors. The obtained
band gap for each system is given in Table , which is in good agreement with the previous
theoretical calculation.[47] The minima of
the conduction band (CBM) sits at the highly symmetric M point and the maxima of the valence band (VBM) is seated along the
Γ–K direction, and as we go down the
period, the VBM moves toward higher energy, resulting in the decrease
of the band gap from 2.05 for AlS to 1.86 eV for AlTe monolayers.
It is clear from band structures that the top of the valence band
is quite flat and the bottom of the conduction band shows a parabolic
band, and consequently resulting in an unusual combination of a rather
sharp peak at VBM and an almost constant band at CBM, which is the
character of a good TE property similar to renowned PbTe1–Se bulk TEs.[48]Figure d–f show the atom-resolved local DOS of three monolayers.
It is observed that the states below the Fermi energy mainly originate
from the X atoms and the states above the Fermi energy are composed
by admixing of X and Al atoms. With the purpose of further analyzing
the features around the CBM and VBM, the s- and p-orbitals’
projected DOS (PDOS) profiles of all the atoms are shown in Figure g–l. The analysis
of the PDOS demonstrates that VBM and CBM for AlS/AlSe/AlTe monolayers
are both formed by the S/Se/Te p-orbitals with some admixture from
Al p-orbitals. Based on the VBM and CBM, we can evaluate the effective
mass (m*) of electron and hole carriers by
Figure 2
Band structures and total DOS (a–c) and
projected
DOS (d–l) of AlX (X = S, Se, Te) monolayers.
Band structures and total DOS (a–c) and
projected
DOS (d–l) of AlX (X = S, Se, Te) monolayers.The obtained values of m* for the two carriers in AlX (X = S, Se, Te) monolayers
are collected in Table . Expectedly, the effective mass of electrons is smaller than that
of holes in all the three monolayers, which is mainly attributed to
a steeper CBM (for electrons) compared with a VBM (for holes). Table also summarizes the
calculated deformation potential (DP) constant (E1), obtained carrier mobility (μ), and relaxation
time (τ) at room temperature. It is obvious to find that the
carrier mobility of electrons is larger than that of holes resulting
from the unusual DOS feature seen in Figure . The temperature dependence of the relaxation
time (τ) is presented in Figure .
Figure 3
Calculated
carrier relaxation time for electrons (a) and holes (b) as a function
of temperature in AlX (X = S, Se, Te) monolayers.
Calculated
carrier relaxation time for electrons (a) and holes (b) as a function
of temperature in AlX (X = S, Se, Te) monolayers.
TE Properties
The carrier concentration (n) dependence of transport
coefficients including the Seebeck coefficient (S), electrical conductivity (σ), power factor (PF = S2σ), and electronic thermal conductivity
(κe) at 300, 500, and 700 K for AlX (X = S, Se, Te)
monolayers are presented in Figure (for n-type systems) and Figure (for p-type systems). As can be seen from Figures and 5a–c, the absolute Seebeck coefficients decrease for
n-type and p-type AlS, AlSe, and AlTe monolayers with increasing carrier
concentration, but they show an increasing trend when increasing the
temperature at constant carrier concentrations, reflecting that S is inversely proportional to n, but is
proportional to T. Moreover, the absolute S of n-type AlS and AlSe monolayers is smaller than that
of the p-type ones. For example, the absolute values of S for n-type AlS are 322.89 (at 300 K), 367.61 (at 500 K), and 397.79
μV K–1 (at 700 K) at the carrier concentration
of 1011 cm–2, while for the p-type systems,
they are 652.39 (at 300 K), 704.82 (at 500 K), and 717.92 μV
K–1 (at 700 K), which are about twice of those for
the n-type ones at the same condition. Also, the absolute S are 489.73 (at 300 K), 554.62 (at 500 K), and 578.94 μV
K–1 (at 700 K) for n-type AlSe, which are also smaller
than 641.74 (at 300 K), 680.89 (at 500 K), and 700.56 μV K–1 (at 700 K) for p-type AlSe. Otherwise, the absolute S of n-type AlTe monolayer are 350.18, 402.69, and 445.74
μV K–1 at the temperature of 300, 500, 700
K, which are larger than 316.47, 298.24, and 321.46 μV K–1 for the p-type AlTe monolayer at the same carrier
concentration and temperature. The consequent electrical conductivity
(σ) is shown in Figure d–f (for n-type systems) and Figure d–f (for p-type systems). Contrary
to the Seebeck coefficient, the electrical conductivity increases
with n in the whole concentration range. The behavior
of σ with respect to temperature, however, is completely inverse.
All the three n-type systems hold bigger σ rather than p-type
systems at the same carrier concentration, originating from the larger
carrier mobility of electrons because of σ = neμ. An ideal TE material not only requires to minimize the thermal
conductivity but also simultaneously needs a maximized power factor
(PF). However, transport coefficients are usually coupled with each
other. As can be seen from above discussion, the presence of large
effective mass is required for obtaining a high Seebeck coefficient,
but electrical conductivity is inversely proportional to the value
of m*. One should make a balance between the Seebeck
coefficient and electrical conductivity for a high power factor (PF).
The calculated PF for n-type and p-type AlS, AlSe, and AlTe monolayers
is presented in Figures and 5g–i. The curves first increase
with carrier concentration to a maximum and then go down with further
increasing in n. For the n-type AlS monolayer, the
power factor can be optimized to be as high as 22.59 (300 K), 22.45
(500 K), and 23.39 mW m–1 K–2 (700
K), which are all larger than that for the p-type system. This indicates
the superior TE behavior of the n-type system with respect to the
p-type system. The largest power factor (62.59 mW m–1 K–2) among the three systems is obtained in the
n-type AlSe monolayer at the carrier concentration of 5.5 × 1012 cm–2 at T = 300 K because
of the largest electrical conductivity. As discussed above, the largest
electrical conductivity is mainly attributed to the highest carrier
mobility of electrons in the AlSe monolayer (in Table ).
Figure 4
Seebeck coefficient
(S) (a–c), electrical conductivity (σ)
(d–f),
power factor (PF) (g–i), and electronic thermal conductivity
(κe) (j–l) as a function of carrier concentration
for n-type AlS, AlSe, and AlTe monolayers at 300, 500, and 700 K.
Figure 5
Seebeck
coefficient (S) (a–c), electrical conductivity
(σ) (d–f), power factor (PF) (g–i), and electronic
thermal conductivity (κe) (j–l) as a function
of carrier concentration for p-type AlS, AlSe, and AlTe monolayers
at 300, 500, and 700 K.
Seebeck coefficient
(S) (a–c), electrical conductivity (σ)
(d–f),
power factor (PF) (g–i), and electronic thermal conductivity
(κe) (j–l) as a function of carrier concentration
for n-type AlS, AlSe, and AlTe monolayers at 300, 500, and 700 K.Seebeck
coefficient (S) (a–c), electrical conductivity
(σ) (d–f), power factor (PF) (g–i), and electronic
thermal conductivity (κe) (j–l) as a function
of carrier concentration for p-type AlS, AlSe, and AlTe monolayers
at 300, 500, and 700 K.We then turn to explore the phonon
transport properties of AlX (X = S, Se, Te) monolayers, as the thermal
transport in semiconductors is mainly contributed by the phonon. Figure a–c shows
the dispersion of phonon modes of AlX (X = S, Se, Te) monolayers,
no imaginary frequencies in all phonon branches confirm the dynamical
stability of the AlX (X = S, Se, Te) monolayers, which are consistent
with the previous theoretical work.[47] The
calculated phonon dispersions for the three systems are all composed
of three acoustic modes and nine optical ones. The three acoustic
modes contains two linear modes [longitudinal acoustic (LA) and transverse
acoustic (TA)] for in-plane vibrations and a flexural one (ZA) for
out-of-plane vibrations. The dispersion of all the three acoustic
modes is quite stronger away from the zone center, thus giving higher
velocity modes than typical optical phonons and possesses most of
the heat. Besides, it is found that the acoustic modes are more dispersive
from AlS to AlSe to AlTe monolayers, leading to a higher group velocity
and thus an enhanced lattice thermal conductivity.[49−51] The group velocity of ZA, TA, and LA phonons
near the Γ point for the AlX (X = S, Se, Te) monolayers is calculated
and listed in Table . For comparison, the corresponding results of MoS2, GaSe,
and InSe monolayers are also presented.[49,52] As a fact,
the ZA branch plays a crucial role in κl of 2D materials.[53,54] For instance, 75% κl is derived from the ZA mode
in graphene.[55] In this study, the larger
phonon group velocities for the ZA branch in the AlX (X = S, Se, Te)
monolayers compared with those of MoS2 monolayer will contribute
to the lower κl in the MoS2 monolayer.
However, the calculated κl of the MoS2 monolayer is much larger than that of AlX (X = S, Se, Te) monolayers,
and thus suggesting lower Debye temperature in the AlX (X = S, Se,
Te) monolayer. In Slack’s model,[56] κl has cubic dependence on Debye temperature, and
accordingly, a higher Debye temperature corresponds a higher κl. A comparison of κl for three selenide-based
monolayers, AlSe, GaSe, and InSe, is illuminating. The atomic mass
of Al (27) is much smaller than that of Ga (70) and In (115), and
thus, one might expect a higher κl for the AlSe monolayer.
Nevertheless, the calculated κl of the AlSe monolayer
is the smallest among the three monolayers. To explain the unusually
low κl of the AlSe monolayer, we consider the mechanical
properties of the three selenide-based monolayers. Young’s
modulus (E) of AlSe, GaSe, and InSe monolayers are,
respectively, 26.71, 28.62, and 27.96 GPa. Poisson’s ratio
(σ) of AlSe, GaSe, and InSe monolayers is listed as 0.24, 0.24,
and 0.29.[47] The AlSe monolayer has very
small E and σ, indicating lower vibrational
strength.[57] From the phonon dispersion
of the AlSe, GaSe, and InSe monolayers shown in other literature,[47] the ZA modes are more and more flat from AlSe
to GaSe to InSe monolayers. Therefore, we conclude that the AlSe monolayer
has the smallest Debye temperature among the three selenide-based
monolayers as discussed above. Therefore, a small Debye temperature
together with low Poisson’s ratio always means a weak interatomic
bonding, which will decrease the κl of the AlSe monolayer.[58] As a result, the AlSe monolayer has a relative
κl compared to GaSe and InSe monolayers. The calculated
lattice thermal conductivity (κl) of the AlX (X =
S, Se, Te) monolayers as a function of temperature in the range from
300 to 700 K is shown in Figure d–f. One can clearly see that the lattice thermal
conductivity of all the three AlS, AlSe, and AlTe monolayers decreases
as the temperature increases. The values of κl for
AlS, AlSe, and AlTe monolayers are respectively 3.00, 4.21, and 4.45
W m–1 K–1 at room temperature.
When taking the thickness into consideration, the thermal sheet conductances
of two-dimensional AlS, AlSe, and AlTe are, respectively, 2.39, 3.61,
and 3.86 nW K–1 (300 K), which are comparable to
those of high zT bulk materials such as PbTe[48] but much smaller than those of Mo- and W-based
dichalcogenides[59−63] and GIIIAMCs
such as InSe[64] at the same temperature.
The low thermal conductivity of AlS, AlSe, and AlTe monolayers suggests
that these three monolayers could have favorable TE performance.
Figure 6
Phonon dispersion relations
(a–c) and
lattice thermal conductivity (κl) as a function of
temperature (d–f) for AlX (X = S, Se, Te) monolayers.
Table 3
Calculated
Group Velocities (vg) (km/s) of ZA, TA,
and LA Phonons Near the Γ Point for the AlX (X = S, Se, Te)
Monolayers with MoS2, GaSe, and InSe Monolayers for Comparison
monolayers
ZA
TA
LA
AlS
1.726
1.745
2.967
AlSe
2.092
2.518
4.525
AlTe
3.036
3.811
4.882
GaSe
2.487
4.162
InSe
1.853
3.272
MoS2
1.40
3.96
6.47
Phonon dispersion relations
(a–c) and
lattice thermal conductivity (κl) as a function of
temperature (d–f) for AlX (X = S, Se, Te) monolayers.Based on the obtained TE transport parameters (S, σ, and κ) of AlX (X = S, Se, Te) monolayers, the TE
performance can be evaluated. Figure plots the zT values as a function
of carrier concentration and temperature for n-type and p-type monolayers.
For p-type AlS, the optimal value of zT at room temperature
is 0.33 at the carrier concentration of 1.9 × 1013 cm–2. At a high temperature of 700 K, a maximal zT of 0.62 appears at the carrier concentration of 2.4 ×
1013 cm–2. The largest zT can be further improved to 0.52 (at T = 300 K and n = 1.5 × 1011 cm–2) and
0.67 (at T = 700 K and n = 1.1 ×
1011 cm–2) for n-type AlS. As such, the
n-type AlS monolayer exhibits a more favorable TE performance rather
than the n-type system. Similar to the AlS case, the n-type AlSe monolayer
has the largest zT of 0.59, 0.69, and 0.74 at the
temperature of 300, 500, and 700 K because of the extremely large
power factor. Notably, even at a lower temperature of 300 K, zT can also compete with that at higher temperatures. Thus,
the n-type AlSe monolayer can exhibit promising TE performance even
at relatively low temperatures. Similar to AlS and AlSe monolayers,
n-type AlTe can be used as a promising TE material. The optimal zT at 300 and 700 K for the n-type AlTe monolayer are, respectively,
0.26 and 0.73 at the carrier concentration of 5.4 × 1011 and 2.7 × 1012 cm–2. Including
the results of AlS, it can be concluded that the n-type AlSe monolayer
exhibits the largest TE figure of merit among the studied AlX (X =
S, Se, Te) monolayers because of its intrinsic band structure. Although
the DP theory has been widely used to calculate the relaxation time
for two-dimensional systems, it still may be unstable for the DP model
which fully ignores electron–optical–phonon interaction,
polar scattering, and other scattering mechanisms. In order to make
a comparison to the previous work easier, the zT values
of AlX (X = S, Se, Te) monolayers at 300 K with a relaxation time
range centered around the calculated ones from the DP model are plotted
in Figure . The solid
black lines stand for zT using the relaxation time
obtained from the DP model. The zT values obtained
on the basis of estimated relaxation time from the DP model listed
in Table are nearly
equal to the largest values for both n- and p-type AlX (X = S, Se,
Te) monolayers. Assuming the same relaxation time, our obtained values
of zT are even larger than those of the previously
promising TE material such as the Pd2Se3 monolayer.[65] Therefore, our study indicates the high performance
of AlX (X = S, Se, Te) monolayers, especially the AlSe monolayer when
used as potential candidate matrices for TE nanocomposites.
Figure 7
Figure
of merit (zT) as a function of carrier concentration
for n-type and p-type AlS (a–b), AlSe (c–d), and AlTe
(e–f) monolayers at 300, 500, and 700 K.
Figure 8
Figure of merit
(zT) values for n-type and p-type AlS (a–b),
AlSe
(c–d), and AlTe (e–f) monolayers at 300 K with a relaxation
time range centered around the calculated ones from the DP model.
Figure
of merit (zT) as a function of carrier concentration
for n-type and p-type AlS (a–b), AlSe (c–d), and AlTe
(e–f) monolayers at 300, 500, and 700 K.Figure of merit
(zT) values for n-type and p-type AlS (a–b),
AlSe
(c–d), and AlTe (e–f) monolayers at 300 K with a relaxation
time range centered around the calculated ones from the DP model.
Computational Method and Process
In
the present work, the first-principles calculations on AlX (X = S,
Se, Te) monolayers are conducted within the VASP code[66−68] using the projector-augmented
wave method.[69] The combination of Perdew–Burke–Ernzerhof
functional scheme and generalized gradient approximation[70] describes the electronic exchange–correlation
potential. After a strict convergence test, cutoff energy of the plane
wave is chosen as 600 eV and the Monkhorst–Pack uniform k-point sampling[71] is selected
as 24 × 24 × 1 in the whole Brillouin zone. The energy and
force convergence criteria are, respectively, taken to be 10–6 eV and 0.01 eV/Å. By conducting the energy minimization calculations
on the basis of the conjugate gradient method, we determine the optimized
configuration of the systems. A vacuum slab of 20 Å is enough
to eliminate the interactions between the adjacent AlX (X = S, Se,
Te) monolayers.Based on the calculated accurate electronic
structures and carrier relaxation time, the electronic transport properties
are then calculated from the semiclassical Boltzmann theory.[72] The obtained DFT results are used as an input
of BoltzTraP to estimate TE performance of AlX (X = S, Se, Te) monolayers.
In this approach, the temperature (T) and chemical
potential (μ) dependence of the Seebeck coefficient (Sαβ), electrical conductivity (σαβ), and electronic thermal conductivity (καβe)
are expressed as[73,74]in which
subscripts α and β represent
the two axis directions in the momentum space (or according to real
space). Here, Ω is the volume of the unit cell and fμ stands for the Fermi–Dirac distribution
function. The σαβ(ε) as a function
of ε is in the form ofIn this formula, i and donate the band index and wave
vector, respectively. Parameters N, e, and τ express the number of k-points samples,
electron charge, and carrier relaxation time, respectively. The carrier
group velocity vα(vβ) along the α(β) direction can be calculated
fromThe electrical
conductivity (σ) and electronic thermal conductivity (κe) are proportional to the relaxation time (τ). Therefore,
the accurate treatment of τ is of great importance. Many earlier
theoretical calculations on TE properties were performed based on
a constant relaxation time[42,75,76] and the value is generally overestimated.[27,43,77] In this study, the relaxation time is evaluated
by adopting the DP theory,[78] which is widely
used to calculate the relaxation time for two-dimensional systems.[26,79−83] Accordingly,
the relaxation time is calculated byin which
the effective mass (m*) is obtained from the accurate
band structure. For a 2D crystal, the carrier mobility (μ) can
be expressed asIn the formula, the two-dimensional elastic
modulus (C2D) is calculated fromwhere l0 is the optimized lattice parameter of the unit cell,
Δl = l – l0 is the lattice variation when the compressed and expanded
systems are compared with the optimized system, and E is the corresponding energy of the compressed and expanded systems. md is the average density-of-states effective
mass dominated by . E1 is the DP constant
defined asThe anisotropic in-plane lattice thermal conductivity
under the relaxation time approximation can be calculated as the sum
of the contributions of all phonon model λ with different wave
vectors and branch indexeswhere V is the crystal volume, Cλ is the specific heat per mode, vλα and τλ are the velocity
components along the α direction and the relaxation time of
the phonon mode λ. The lattice thermal conductivity (κl) can be obtained by solving the phonon Boltzmann transformation
related to the harmonic and anharmonic interatomic force constants
(IFCs) as performed by the ShengBTE code.[84−86] The inputs for the ShengBTE are the second-
and third-order IFCs. For this, a 4 × 4 × 4 supercell with
5 × 5 × 1 k-point sampling is used to calculate
the second-order IFCs and phonon frequencies by the Phonopy package.[87] The third-order IFCs are obtained by using the
thirdorder.py module using a 4 × 4 × 4 supercell and Γ-point
only calculations. Note that we use eq to calculate the electronic thermal conductivity (κe) because the Wiedemann–Franz law is only suitable
for the system where the scattering of electrons in the material is
dominated by elastic collision.[88] Here,
the effective van der Waals thickness of AlX (X = S, Se, Te) monolayers
is 7.97, 8.57, and 8.68 Å, which is defined as the summation
of the buckling distance and two van der Waals radii of the outermost
surface atoms of structures.[89]
Conclusions
To summarize, we have
presented the electronic, phonon, and TE properties of AlX (X = S,
Se, Te) monolayers by the density functional theory in combination
with the Boltzmann transport theory. All the AlX (X = S, Se, Te) monolayers
are indirect band gap semiconductors with unique electronic structures,
having a flat band at VBM and a parabolic band at CBM, and consequently,
a rather sharp peak at VBM and an almost constant band at CBM in DOS.
No imaginary frequencies in all phonon branches verify the dynamical
stability of the AlX (X = S, Se, Te) monolayers. A detailed study
of TE properties as a function of carrier concentration and temperature
is carried out. The largest power factors are around 22.59, 62.59,
and 6.79 mW m–1 K–2 at room temperature
and 23.40, 54.36, 10.42 mW m–1 K–2 at 700 K under reasonable electronic concentration for AlS, AlSe,
and AlTe monolayers, respectively. The figure of merit (zT) increases with temperature, and the optimized zT values reach 0.52, 0.59, and 0.26 at room temperature and 0.67,
0.74, and 0.73 at 700 K under reasonable electronic concentration
for AlS, AlSe, and AlTe monolayers, respectively, thus suggesting
that the n-type AlX (X = S, Se, Te) monolayers, especially the AlSe
monolayer can be used as potential candidate matrices for high-performance
TE nanocomposites.
Authors: K S Novoselov; A K Geim; S V Morozov; D Jiang; Y Zhang; S V Dubonos; I V Grigorieva; A A Firsov Journal: Science Date: 2004-10-22 Impact factor: 47.728
Authors: Garry W Mudd; Simon A Svatek; Tianhang Ren; Amalia Patanè; Oleg Makarovsky; Laurence Eaves; Peter H Beton; Zakhar D Kovalyuk; George V Lashkarev; Zakhar R Kudrynskyi; Alexandr I Dmitriev Journal: Adv Mater Date: 2013-08-21 Impact factor: 30.849