See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276211518

High-efficient thermoelectric materials: The case of orthorhombic IV-VI

compounds

Article in Scientific Reports · May 2015

DOI: 10.1038/srep09567 · Source: arXiv

CITATIONS READS

217 2,548

3 authors, including:

Guangqian Ding Guoying Gao

Huazhong University of Science and Technology Huazhong University of Science and Technology
65 PUBLICATIONS 2,188 CITATIONS 176 PUBLICATIONS 5,596 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Guangqian Ding on 12 June 2015.

The user has requested enhancement of the downloaded file.

 OPEN High-efficient thermoelectric materials:
SUBJECT AREAS:
The case of orthorhombic IV-VI
THERMOELECTRICS
ELECTRONIC PROPERTIES AND
compounds
MATERIALS
Guangqian Ding, Guoying Gao & Kailun Yao

Received
School of Physics & Wuhan National High Magnetic Field Center, Huazhong University of Science and Technology, Wuhan
19 November 2014 430074, China.
Accepted
11 March 2015
Improving the thermoelectric efficiency is one of the greatest challenges in materials science. The recent
Published discovery of excellent thermoelectric performance in simple orthorhombic SnSe crystal offers new promise
5 JU u ne 2015 in this prospect [Zhao et al. Nature 508, 373 (2014)]. By calculating the thermoelectric properties of
orthorhombic IV-VI compounds GeS,GeSe,SnS, and SnSe based on the first-principles combined with the
Boltzmann transport theory, we show that the Seebeck coefficient, electrical conductivity, and thermal
conductivity of orthorhombic SnSe are in agreement with the recent experiment. Importantly, GeS, GeSe,
Correspondence and and SnS exhibit comparative thermoelectric performance compared to SnSe. Especially, the Seebeck
requests for materials coefficients of GeS, GeSe, and SnS are even larger than that of SnSe under the studied carrier concentration
and temperature region. We also use the Cahill’s model to estimate the lattice thermal conductivities at the
should be addressed to
room temperature. The large Seebeck coefficients, high power factors, and low thermal conductivities make
G.G. (guoying_gao@ these four orthorhombic IV-VI compounds promising candidates for high-efficient thermoelectric
mail.hust.edu.cn) materials.

T
hermoelectric materials which can directly convert waste heat into electrical energy have been a hot spot in
research community especially under the challenges of energy crisis and global warming. The efficiency of
thermoelectric materials is given by the dimensionless figure of merit1:
S2 s
ZT~ T: ð1Þ
ke zkl
Where S, s, ke, kl, and T are the Seebeck coefficient, electrical conductivity, electrical thermal conductivity, lattice
thermal conductivity, and temperature, respectively. A large value of ZT means a higher thermoelectric conver-
sion efficiency which requires a high power factor (S2s) along with a low thermal conductivity (ke 1 kl).
Unfortunately, the thermoelectric figure of merit of conventional three-dimensional crystalline systems can
not be infinitely increased because these parameters are restricted: a large Seebeck coefficient needs low carrier
concentration, which results in low electrical conductivity; a large electrical conductivity always comes with a high
electrical thermal conductivity. Most works in the past decade focused on improving ZT by several approaches2.
For example, enhancing the power factor by band engineering (degeneration of multiple valleys, electronic
resonance states, and depressing bipolar effect, etc.)3–5 and reducing lattice thermal conductivity by all length-
scale hierarchical structuring (nanoscale precipitates and superlattice)6,7. Although these approaches may greatly
enhance ZT in theory or experiment, it is difficult to realize them in practical applications due to the complicated
process and high cost. Therefore, searching for new three-dimensional materials is still an important route to
obtain high-performance thermoelectricity.
Very recently, ultralow thermal conductivity and high thermoelectric figure of merit have been experimentally
found in three-dimensional bulk SnSe by Zhao et al.8. An unprecedented ZT of 2.6 at 973 K was found in SnSe
single crystal along the b axis owing to it’s exceptionally low thermal conductivity. Afterwards, Sassi et al.9 and
Chen et al.10 also reported high ZT values and low thermal conductivities in polycrystalline SnSe, but the ZT
values are lower than those measured in single crystal SnSe by Zhao et al.8, and there is also large discrepancy
between the thermal conductivity measured on single crystals8 and polycrystals9,10. These works gave us inspira-
tion that high thermoelectric performance can be realized in simple bulk materials without complicated crystal
structure and heavy atoms instead of complex strategies to nanostructuring1. Moreover, further studies on the
thermal conductivity of SnSe are necessary in order to understand the different measured results of thermal

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 1

 www.nature.com/scientificreports

conductivities between crystals8 and polycrystals9,10. SnSe undergoes structures of GeS, GeSe, SnS, and SnSe have been studied detailed
a reversible phase transition from low-temperature phase (Pnma in previous works11,26. In this work, we will mainly concern the
space group) to high-temperature phase (Cmcm space group) with thermoelectric properties in the moderate temperature region
transition temperature around 800 K. The Pnma phase contains (Pnma phase) because the measured temperatures of phase
eight atoms and two adjacent double layers in the primitive cell. transition in these compounds are around or above 800 K26.
One atom in a single layer is collected with three nearest neighbors,
which forms zigzag chains11. Since the low thermal conductivity and Results and discussion
high ZT have been found in SnSe, this kind of special layered and The optimized lattice constants and calculated band gaps are shown
anisotropic zigzag structure is expected to exhibit novel thermoelec- in Table 1, which are in good agreement with the experimental
tric properties. Thus, it is necessary and interesting to explore the values11. Obviously, GeS shows the largest gap of 1.25 eV while
thermoelectric performance of new materials with the same or sim- SnSe has the lowest gap of 0.69 eV, and a little gap difference between
ilar crystal structure of SnSe. GeSe and SnS is found. These gaps are much higher than those of
Experimentally, the IV-VI compounds GeS, GeSe, and SnS all traditional thermoelectric materials (0.105 eV for Bi2Te3 and
crystallize in GeS-type structure which belongs to the orthorhombic 0.14 eV for Sb2Te3)7, which would make great difference in ther-
structure with space group Pnma11. However, studies on the ther- moelectric properties. A large gap may cause lower carrier concen-
moelectric properties of these orthorhombic compounds except for tration around the Fermi level. As transport coefficients mainly
SnSe are few11–13, and there is not a comparison of thermoelectric depends on the electronic properties. It would result in large
performance among them. Recently, Tan et al. experimentally found Seebeck coefficients, and also, large power factors can be obtained
a modest ZT of 0.6 at 873 K in polycrystalline SnS14, which is com- by reasonable doping.
parable to those measured in polycrystalline SnSe9,10. Here, in order We firstly calculate the thermoelectric properties of orthorhombic
to search for more high-performance thermoelectric materials, we SnSe in order to compare our calculated results with the recent
investigate and compare the thermoelectric properties of the ortho- experiment8. The electrical conductivities and Seebeck coefficients
rhombic IV–VI compounds by using the density function theory of SnSe along different crystallographic directions as a function of
combined with the Boltzmann transport equation. We reveal that temperature are shown in Figs. 1(a) and 1(b). Temperatures from
GeS, GeSe, and SnS show similar thermoelectric properties to that of 300 K to 750 K are considered in our calculations, which are below
SnSe: large Seebeck coefficient, high power factor, and low thermal the phase transition temperature around 800 K26. We have assumed
conductivity. Importantly, the Seebeck coefficients of GeS,GeSe, and a series of carrier concentrations ranging from 1 3 1017/cm3 to 1 3
SnS are even larger than that of SnSe under the studied carrier con- 1018/cm3 according to the experimental Hall measurements8. We
centration and temperature region. Moreover, the Seebeck coeffi- finally adopt the value of 6 3 1017/cm3 because the obtained ther-
cients and power factors can be greatly enhanced by both n-type moelectric properties with it are better consistent with experiment
and p-type doping. results. As shown in Fig. 1(b), the calculated Seebeck coefficients
along b and c axis are in good agreement with experimental data8,
Computational details. Our calculations are based on the density while the behavior of a axis is lower than the experimental value and
function theory combined with the Boltzmann transport equation. decreases rapidly above 450 K and finally down to negative value
Firstly, structure optimization and electron energy calculations are when the temperature exceeds 600 K. We attribute this rapid
performed by the full-potential linearized augmented plane-wave decrease and sign modification along the a axis to the increase of
method (WIEN2K code)15, which is a powerful tool to deal with thermally excited negative electron. We note that that the recent
the electronic structure of various solid systems16,17. In this experimental Seebeck coefficients on polycrystalline SnSe showed
method, the exchange-correlation functional is adopted with the comparable values9, however, unlike the result of Zhao et al.8, the
generalized gradient approximation in the form of Perdew-Bueke- Seebeck coefficient doesn’t show constant but further decreases
Ernzerhof. The plane-wave cutoff is set as RmtKmax 5 9 and the total above the transition temperature9, which remains to be understood.
energy is converged to within 1025 Ry. Then, based on the calculated Fig. 1(a) shows the anisotropic electrical conductivities in the tem-
electronic energy, we obtain the thermoelectric transport coefficients perature range considered. Under the relaxation-time approximation
by solving the linearized Boltzmann equation in relaxation time in Boltzmann transport calculations, the Seebeck coefficient(S) is
approximation within Boltztrap code18. In order to obtain reliable independent of the relaxation time t, whereas the electrical conduc-
results, the k points in the irreducible Brillouin zone are tested to use tivity (s) depend linearly on t. Furthermore, the electrical thermal
at least 60000 points. Boltzmann transport calculations have long conductivity (ke) is obtained by the Wiedemann-Franz equation (ke
been used for bulk semiconductors and show reliable results19,20. The 5 LsT, L: Lorentz factor), thus, ke is also relevant to t. Usually, it is
phonon dispersion relations are obtained from the phonopy code21, difficult to determine relaxation time in a bulk material because of
which realizes the Parlinski-Li-Kawazoe method, conducted in the complex scattering mechanism (phonon scattering, carrier scat-
supercell approach with the finite displacement method22. In order tering, defect scattering, and boundary scattering). Here, we fix the
to obtain reliable phonon spectrum, a 2 3 4 3 4 supercell is relaxation time at 1 3 10214 s in our calculations by comparing the
employed for force constant calculations and at least 1000 k points calculated results with experiment. Yabuuchi et al.27 also adopt this
are used. Here, we evaluate the room temperature minimum lattice value in their calculations for its rationality. Electrical conductivities
thermal conductivities of these compounds by the Cahill’s23 method, show weakly decrease in the 300–500 K range but increase rapidly
which borrowed from Debye model and has been successfully used for more carriers are excited as temperature shifts to higher value,
in predicting the lattice thermal conductivity24,25. The crystal which corresponds to the slowly increase and quickly decrease of

Table 1 | Optimized lattice constants and calculated band gaps of GeS, GeSe, SnS, and SnSe
Composition a(Å) b(Å) c(Å) gap(eV)
GeS 10.635 3.668 4.738 1.25
GeSe 10.954 3.828 4.492 0.87
SnS 11.392 4.043 4.353 0.91
SnSe 11.754 4.203 4.441 0.69

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 2

 www.nature.com/scientificreports

Figure 1 | Thermoelectric parameters as a function of temperature for SnSe: (a) Anisotropic electrical conductivity, (b) Anisotropic Seebeck
coefficient, (c) Electrical conductivity, (d) Seebeck coefficient, (e) Electrical thermal conductivity, and (f) Thermoelectric figure of merit (only electrical
thermal conductivity concerned).

Seebeck coefficients in the temperature region. Figs. 1(c) and 1(d) excited electrons will lead to the increase of electrical thermal con-
give the total electrical conductivities and Seebeck coefficients, which ductivity, while the lattice contribution is decreased because of the
equal to the average values of three crystal directions. The largest increasing ratio of phonon scattering resulting from severe lattice
Seebeck coefficient of 570 mV/K is found at 450 K, and electrical vibration. Therefore, the ratio ZTe is a good estimate of ZT as tem-
conductivity has an almost constant value near 5 Scm21 in the room perature increase to high value. As shown in Fig. 1(f), a ZTe platform
temperature region. Fig. 1(e) presents the low electrical thermal about 0.92 at 300 K to 450 K is found, and it rapidly decrease down
conductivity of SnSe, which ranges from 0.04 to 0.3 Wm21K21 as to near zero in higher temperature because of the rapidly increasing
temperature increase. electrical thermal conductivity. It is evident that the Pnma phase of
When towards the thermoelectric figure of merit, the relaxation SnSe may reach its upper limit ZT in the moderate temperature.
time t will also be an undetermined parameter because of the relaxa- We adopt the same carrier concentration of 6 3 1017/cm3 to cal-
tion-time approximation. Here, we rewrite the equation (1) as, culate the transport coefficients of GeS, GeSe, and SnS for better
comparison with SnSe. Remarkably, the Seebeck coefficients of
S2 sT ke GeS, GeSe, and SnS (Fig. 2(b)) are larger than that of SnSe
ZT~ : ð2Þ
ke ke zkl (Fig. 1(d)) in the whole temperature range. In Fig. 2(b), GeS shows
the largest Seebeck coefficient while SnS has the lowest one in the
Thus, the ratio (ZTe 5 S2sT/ke) is independent of the relaxation time 300–600 K range. However, the Seebeck coefficient of GeSe
t and is an upper limit to the thermoelectric figure of merit, which decreases faster compared to that of SnS when the temperature
only ignores the thermal conductivity from lattice contribution. The exceeds 600 K. Obviously, the maximum values of Seebeck coeffi-
value of ZTe approaches ZT if the lattice contribution to the thermal cients for GeSe, SnS, and GeS (706 mV/K at 560 K for GeSe, 668 mV/
conductivity is insignificant compared with the electronic term. In K at 650 K for SnS, and 769 mV/K at 760 K for GeS, respectively)
very low temperature, a very few electrons are excited, which leads to shift to higher temperature compared with SnSe (570 mV/K at
low electrical thermal conductivity. Thus, the thermal conductivity 450 K). In Fig. 2(c), electrical thermal conductivities of GeS, GeSe,
may dominated by lattice term and estimating ZT needs to better and SnS are independent of temperature and much lower than that of
consider the lattice thermal conductivity. As temperature increases SnSe (see Fig. 1(e)) in 300–750 K range, and the values keep nearly
especially when higher than room temperature, a large number of under 0.1 Wm21K21 until 750 K, which indicate more excellent

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 3

 www.nature.com/scientificreports

Figure 2 | Thermoelectric parameters as a function of temperature for GeS, GeSe, and SnS: (a) Electrical conductivity, (b) Seebeck coefficient, (c)
Electrical thermal conductivity, and (d) Thermoelectric figure of merit (only electrical thermal conductivity concerned).

thermal stability of GeS, GeSe, and SnS than SnSe. It can be readily and SnS are much larger than that of SnSe which has been obtained
seen that rapid increase of electrical thermal conductivities occurs by recent experiment8. We summarize in Table 2 the peaks of power
above 750 K. This behavior corresponds to the increase of electrical factors and corresponding carrier concentrations for both n-type and
conductivities shown in Fig. 2(a) due to the proportional relationship p-type doping at 300 K and 600 K of these compounds. The carrier
between electrical conductivity and electrical thermal conductivity. It concentrations are calculated from the Hall coefficients, based on the
is shown in Fig. 1(d) that the trends of ZTe of GeS, GeSe, and SnS are following formula29,
similar to that of SnSe. However, there are two differences we have to
make clear: (i) the temperature regions of the ZTe platforms are wider 1
N~ : ð3Þ
than SnSe especially for GeS (from 300 K to 650 K); (ii) it can be seen RH e
that the ZTe platforms of GeS, GeSe, and SnS are all higher than that
of SnSe, and the highest value approaching to 1 is found in SnS. Here, N and RH are the carrier concentration and Hall coefficient,
Comparatively speaking, GeS, GeSe, and SnS may show better ther- respectively. These carrier concentrations are in the reasonable dop-
moelectric performance than SnSe. ing level30, and it is desirable to dope these compounds in the range of
In Fig. 3, the calculated Seebeck coefficients, power factors, and 1020–1021 cm23 to achieve high power factors. However, high power
ZTe under different temperatures (300 K and 600 K) as a function of factor doesn’t mean corresponding high thermoelectric figure of
chemical potential are presented in order to optimize the ther- merit because of the high electrical thermal conductivity. As shown
moelectric performance of these compounds. Within the rigid-band in Fig. 3, the ZTe of these compounds reach their upper limit in
picture, the chemical potential indicates the doping level of the com- smaller jmj compared with power factors. Higher ZTe at 300 K for
pound. As for n-type doping, the Fermi level shifts up, which corre- these compounds are found, and with the highest value of 1.15 in
sponds to positive m. Instead, the Fermi level shifts down for p-type GeSe, which further indicate the better performance in moderate
doping and the corresponding m is negative28. Surprisingly, S is temperature for the Pnma phase of these IV–VI compounds.
greatly enhanced in a narrow region around m 5 0, which indicates Finally, in order to give a theoretical prediction of the lattice ther-
that a quite remarkable value of S can be achieved through small n- mal conductivity, we calculate the phonon dispersion relations of
type or p-type doping. It is evident that the maximum jSj move GeS, GeSe, SnS, and SnSe along several high symmetry lines as shown
towards the gap edges from SnSe to GeSe, SnS, and GeS because their in Fig. 4 (C-X Brillouin zone direction represents the a crystal axis, C-
band gaps increase. At 300 K, GeS almost reaches the largest value of Y the b axis, and C-Z the c axis, respectively). There is little difference
2000 mV/K while SnSe shows lowest value of 930 mV/K, GeSe and SnS in these dispersion relations except for the maximum frequency of
show almost the same maximum jSj because of their small difference vibrational modes. Since the primitive cells of these compounds
in gaps (see Table 1). However, these values almost halved at 600 K. contain eight atoms, twenty four vibrational modes can be found,
It is worth noting that jSj at 300 K comes closer to S 5 0 than that at including three acoustic branches and twenty one optical branches.
600 K at higher jmj, which shows that higher power factor at 600 K in The acoustic branches describe the vibration of the center of mass,
these chemical potential regions would be found. Obviously, large jSj and the optical branches describe the relative vibration between two
leads to large power factor for both n-type and p-type doping, and n- atoms. In most cases, the frequency of the optical modes approx-
type doping performs better results than p-type doping. On the imate to constant within the range of wave vector, leading to low
contrary, the peaks of power factors under 600 K are much higher group velocity in these phonons. Thus, the optical modes will not be
than those under 300 K due to the higher jSj in higher chemical interested in some cases because their contribution to thermal con-
potential regions. These optimized power factors for GeS, GeSe, ductivity is limited. Considering Cahill, whose theoretical model also

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 4

 www.nature.com/scientificreports

Figure 3 | Seebeck coefficients, power factors (PF), and thermoelectric figure of merit (only electrical thermal conductivity concerned) as a function of
chemical potential for GeS, GeSe, SnS, and SnSe under different temperatures (left part for 300 K and right part for 600 K).

ignores the contribution of optical branches, and is thus an approx- Where kB is the Boltzmann constant, and n is the number density of
imate value approaching to the minimum lattice thermal conduc- atoms. The sum is taken over the three acoustic modes (two trans-
tivity23. However, unlike half-Heusler alloys31, these compounds verse and one longitudinal). vi is the phonon velocity, and hi is the
contain a large number of optical modes, which may cause an obvi- debye temperature, hi 5 vi( /kB)(6p2n)M. The phonon velocities are
ous underestimation of lattice thermal conductivity in Cahill’s the slope of the acoustic phonon dispersions around the C point,
model. Even so, we can also obtain a predictive result. Here, we use which is given by,
Cahill’s model to theoretically predict the minimum lattice thermal Lvi (k)
conductivities at room temperature 300 K. The formula of Cahill’s ni ~ : ð5Þ
model is given by Ref. 23, Lk
ð hi=T Where i represents LA, TA, and TA9. The calculated results are given
2 X
3
p 1 T x 3 ex in Table 3. The phonon velocities for each acoustic mode are the
kmin ~( ) =3 kB n =3 vi ( )2 2 dx: ð4Þ
6 i~1
hi x
0 (e {1) average of three crystal directions. As for SnSe, the calculated room

Table 2 | The peaks of power factors (PF) and corresponding carrier concentrations (N) per cubic centimeter for both n-type and p-type
doping of GeS, GeSe, SnS, and SnSe at 300 K and 600 K
300 K 600 K

n-type p-type n-type p-type

Composition PF N PF N PF N PF N
GeS 30.25 26.8 3 1020 15.03 7.3 3 1020 69.35 24.0 3 1020 44.91 1.5 3 1021
GeSe 35.03 21.2 3 1020 15.02 1.4 3 1020 80.47 21.4 3 1020 41.94 8.0 3 1020
SnS 21.25 27.0 3 1020 14.70 1.3 3 1020 63.41 29.0 3 1020 36.01 3.0 3 1020
SnSe 20.07 26.7 3 1020 23.48 6.8 3 1021 64.64 25.8 3 1020 52.52 7.1 3 1021

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 5

 www.nature.com/scientificreports

Figure 4 | Phonon dispersion relations for GeS, GeSe, SnS and SnSe with Pnma phase along several high symmetry lines.

Table 3 | Calculated phonon velocities (two transverse and one longitudinal) and theoretical minimum lattice thermal conductivities at 300 K
of GeS, GeSe, SnS, and SnSe
n (1022 cm23) vTA (ms21) vTA9 (ms21) vLA (ms21) kmin (Wm21K21)
GeS 4.6 1647 2447 3412 0.52
GeSe 4.1 1138 2124 2820 0.39
SnS 3.9 1537 2343 3368 0.45
SnSe 3.6 1121 1568 2412 0.32

temperature kmin is 0.32 Wm21K21, which is underestimated com- these IV–VI compounds, and will stimulate experimental and theor-
pared with Zhao’s experimental measurement8 and Carrete’s calcula- etical studies of high-efficient thermoelectric materials.
tion32 due to the ignorance of optical modes contributions in the
Cahill’s model. What’s more, we find the detailed study on the lattice
1. Levi, B. G. Simple compound manifests record-high thermoelectric performance.
thermal conductivity presented by Carrete et al.32 somewhat contra- Phys. Today 67, 14–16 (2014).
dicts to Zhao’s measurement8. Moreover, both Sassi et al.9 and Chen 2. He, J., Kanatzidis, M. G. & Dravid, V. P. High performance bulk thermoelectrics
et al.10 shown that the thermal conductivity measured by Zhao et al.8 via a panoscopic approach. Mater. Today 16, 166–176 (2013).
is underestimated and doubt the achieved high ZT. These differences 3. Pei, Y. et al. Convergence of electronic bands for high performance bulk
thermoelectrics. Nature 473, 66–69 (2011).
need further theoretical and experimental investigations. The kmin of 4. Heremans, J. P. et al. Cooling, heating, generating power, and recovering waste
GeS, GeSe, and SnS are still much low but all higher than that of SnSe. heat with thermoelectric systems. Science 321, 554–557 (2008).
While the kmin is underestimated, these compounds are also believed 5. Tan, G. et al. High Thermoelectric Performance of p-Type SnTe via a Synergistic
to show lower lattice thermal conductivity than other thermoelectric Band Engineering and Nanostructuring Approach. J. Am. Chem. Soc. 136,
7006–7017 (2014).
materials such as Bi2Te3 and TiNiSn33,34.
6. Garg, J. & Chen, G. Minimum thermal conductivity in superlattices: A first-
principles formalism. Phys. Rev. B 87, 140302 (2013).
Conclusion 7. Hinsche, N. F. et al. Thermoelectric transport in Bi2Te3/Sb2Te3 superlattices.
Phys. Rev. B 86, 085323 (2012).
In summary, motivated by the recent experiment that the unpreced- 8. Zhao, L.-D. et al. Ultralow thermal conductivity and high thermoelectric figure of
ented ultralow thermal conductivity and high thermoelectric figure merit in SnSe crystals. Nature 508, 373–377 (2014).
of merit in SnSe single crystal, we have used the first-principles 9. Sassi, S. et al. Assessment of the thermoelectric performance of polycrystalline p-
combined with the Boltzmann transport theory to investigate the type SnSe. Appl. Phys. Lett. 104, 212105 (2014).
10. Chen, C.-L., Wang, H., Chen, Y.-Y., Day, T. & Snyder, G. J. Thermoelectric
thermoelectric properties of four orthorhombic IV–VI compounds. properties of p-type polycrystalline SnSe doped with Ag. J. Mater. Chem. A. 2,
It is found that GeS, GeSe, and SnS show comparative thermoelectric 11171–11176 (2014).
properties compared with SnSe. Large Seebeck coefficients, high 11. Taniguchi, M., Johnson, R. L., Ghijsen, R. L. & Cardona, M. Core excitons and
power factors, and low thermal conductivities can be achieved in conduction-band structures in orthorhombic GeS, GeSe, SnS, and SnSe single
crystals. Phys. Rev. B 42, 3634–3643 (1990).
these compounds, which make them promising candidates for high 12. Nassary, M. M. Temperature dependence of the electrical conductivity, Hall effect
efficient thermoelectric materials. The present work provides a and thermoelectric power of SnS single crystals. J. Alloys. Compd. 398, 21–25
detailed understanding of the excellent thermoelectric properties of (2005).

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 6

 www.nature.com/scientificreports

13. Solanki, G. K., Deshpande, M. P., Agarwal, M. K. & Patel, P. D. Thermoelectric 29. Nariya, B. B., Dasadia, A. K., Bhayani, M. K., Patel, A. J. & Jani, A. R. Electrical
power factor measurements in GeSe single crystals grown using different Transport Properties of SnS and SnSe Single Crystals Grown By Direct Vapour
transporting agents. J. Mater. Sci. Lett. 22, 985–987 (2003). Transport Technique. Chalcogenide. Lett. 6, 549–554 (2009).
14. Tan, Q. et al. Thermoelectrics with earth abundant elements: low thermal 30. Rosi, F. D. Thermoelectricity and thermoelectric power generation. Solid. State.
conductivity and high thermopower in doped SnS. J. Mater. Chem. A. 2, Electron. 11, 833–848 (1968).
17302–13706 (2014). 31. Wee, D., Kozinsky, B., Pavan, B. & Fornari, M. Quasiharmonic vibrational
15. Blaha, P., Schwarz, K., Madsen, G. K. H., Kvasnicka, D. & Luitz, J. Wien2k, Vienna properties of TiNiSn from ab initio phonons. J. Electron. Mater. 41, 977 (2011).
University of Technology, 2002. Improved and updatedUnix version of the 32. Carrete, J., Mingo, N. & Curtarolo, S. Low thermal conductivity and triaxial
original copyrighted Wiencode, which was published by P. Blaha, K. Schwarz, P. phononic anisotropy of SnSe. Appl. Phys. Lett. 105, 101907 (2014).
Sorantin, and S. B. Trickey. Comput. Phys. Commun. 59, 399–415 (1990). 33. Chiritescu, C., Mortensen, C., Cahill, D. G., Johnson, D. & Zschack, P. Lower limit
16. Gao, G. Y. & Yao, K.-L. Antiferromagnetic half-metals, gapless half-metals, and to the lattice thermal conductivity of nanostructured Bi2Te3-based materials.
spin gapless semiconductors: The D03-type Heusler alloys. Appl. Phys. Lett. 103, J. Appl. Phys. 106, 073503 (2009).
232409 (2013). 34. Bhattacharya, S. et al. Grain structure effects on the lattice thermal conductivity of
17. Gao, G. Y., Yao, K. L. & Li, N. Preserving the half-metallicity at the surfaces of Ti-based half-Heusler alloys. Appl. Phys. Lett. 81, 43–45 (2002).
rocksalt CaN and SrN and the interfaces of the CaN/InN and SrN/GaP: a density
functional study. J. Phys.: Condens. Matter. 23, 075501 (2011).
18. Madsen, G. K. H. & Singh, D. J. S. BoltzTraP. A code for calculating band-structure Acknowledgments
dependent quantities. Comput. Phys. Commun. 175, 67–71 (2006). This work was supported by the National Natural Science Foundation of China under Grant
19. Ding, G., Gao, G. Y. & Yao, K. L. Thermoelectric performance of half-Heusler Nos. 11474113, 11004066 and 11274130, and by the Fundamental Research Funds for the
compounds MYSb (M 5 Ni, Pd, Pt). J. Phys. D: Appl. Phys. 47, 385305 (2014). Central Universities under Grant No. HUST: 2013QN014. We also acknowledge the helpful
20. Saeed, Y., Singh, N. & Schwingenschlogl, U. Superior thermoelectric response in discussion with Prof. Ruqian Wu.
the 3R phases of hydrated NaxRhO2. Sci. Rep. 4, 4390 (2014).
21. Togo, A., Oba, F. & Tanaka, I. First-principles calculations of the ferroelastic
transition between rutile-type and CaCl2-type SiO2 at high pressures. Phys. Rev. B Author contributions
78, 134106 (2008). G.D. and G.G. designed the outline of the manuscript and wrote the manuscript. G.G.
22. Parlinski, K., Li, Z. Q. & Kawazoe, Y. First-principles determination of the soft supervised the whole work. G.D. carried out the numerical calculations. K.Y. contributed
mode in cubic ZrO2. Phys. Rev. Lett. 78, 4063–4066 (1997). discussions. All authors reviewed the manuscript.
23. Cahill, D. G., Watson, S. K. & Pohl, R. O. Lower limit to the thermal conductivity
of disordered crystals. Phys. Rev. B 46, 6131–6140 (1992).
24. Kosuga, A., Umekage, K., Matsuzawa, M., Sakamoto, Y. & Yamada, I. Room-
Additional information
Competing financial interests: The authors declare no competing financial interests.
temperature pressure-induced nanostructural CuInTe2 thermoelectric material
with low thermal conductivity. Inorg. Chem. 53, 6844–6849 (2014). How to cite this article: Ding, G., Gao, G. & Yao, K. High-efficient thermoelectric materials:
25. Siegert, K. S. et al. Impact of vacancy ordering on thermal transport in crystalline The case of orthorhombic IV-VI compounds. Sci. Rep. 5, 9567; DOI:10.1038/srep09567
phase-change Materials. Rep. Prog. Phys. 78, 013001 (2015). (2015).
26. Wiedemeier, H. & Schnering, H. G. Refinement of the structures of GeS, GeSe, SnS
and SnSe. Z. Kristallogr. 148, 295–303 (1978). This work is licensed under a Creative Commons Attribution 4.0 International
27. Yabuuchi, S., Okamoto, M., Nishide, A., Kurosaki, Y. & Hayakawa, J. Large License. The images or other third party material in this article are included in the
Seebeck Coefficients of Fe2TiSn and Fe2TiSi: First-Principles Study. Appl. Phys. article’s Creative Commons license, unless indicated otherwise in the credit line; if
Express. 6, 025504 (2013). the material is not included under the Creative Commons license, users will need
28. Wang, L. L. et al. Thermoelectric performance of half-heusler compounds TiNiSn to obtain permission from the license holder in order to reproduce the material. To
and TiCoSb. J. Appl. Phys. 105, 013709 (2009). view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS | 5 : 9567 | DOI: 10.1038/srep09567 7

View publication stats