Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Thermoelectric properties of SnS with Na-doping Binqiang Zhou, Shuai Li, Wen Li, Juan Li, Xinyue Zhang, Siqi Lin, Zhiwei Chen, and Yanzhong Pei ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b08770 • Publication Date (Web): 12 Sep 2017 Downloaded from http://pubs.acs.org on September 13, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 7
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
Thermoelectric properties of SnS with Na-doping Binqiang Zhou, Shuai Li, Wen Li, Juan Li, Xinyue Zhang, Siqi Lin, Zhiwei Chen and Yanzhong Pei* Interdisciplinary Materials Research Center, School of Materials Science and Engineering, Tongji Univ., 4800 Caoan Rd., Shanghai 201804, China. *Email:
[email protected] Abstract Tin sulfide (SnS), a low-cost compound from IV-VI semiconductors, has attracted particular attention due to its great potential for large-scale thermoelectric applications. However, pristine SnS shows a low carrier concentration, which leads to a low thermoelectric performance. In this work, sodium is utilized to substitute Sn to increase the hole concentration and thus to improve the thermoelectric power factor. The resultant Hall carrier concentration up to ~1019 cm-3, is the highest so far reported for this compound. This further leads to the highest reported so far thermoelectric figure of merit, zT of 0.65, in polycrystalline SnS. Temperature dependence Hall mobility shows a transition of carrier scattering source from a grain boundary potential below 400K to acoustic phonons at higher temperatures. The electronic transport properties can be well understood by a single parabolic band (SPB) model, enabling a quantitative guidance for maximizing the thermoelectric power factor. Using the experimental lattice thermal conductivity, a maximal zT of 0.8 at 850 K is expected when the carrier concentration is further increased to ~1×1020 cm-3, according to the SPB model. This work not only demonstrates SnS as a promising low-cost thermoelectric material, but also details the material parameters that fundamentally determine the thermoelectric properties. Keywords: Thermoelectric; SnS; Carrier concentration; SPB model; zT 1. Introduction thermal conductivity shows a good agreement with Thermoelectric materials, enabling a direct conversion experimental measurements39. These features indicate SnS between heat and electricity based on either Seebeck or as a potential thermoelectric material with cheap constituent Peltier effect, have attracted extensive interests due to the elements. energy crisis and environment issue1. A high conversion Unfortunately, peak zT>0.6 has not yet been realized efficiency of thermoelectric materials is required for large experimentally so far, in neither single- nor poly- crystalline scale applications. The performance of thermoelectric SnS39-41. This is due to the low carrier concentrations materials is determined by dimensionless figure of merit, achievable. It is known that both power factor (PF=S2/ρ) 2 zT=S T/ρ(κE+κL), where S, ρ, κE, κL and T are the Seebeck and zT can only be maximized in a certain narrow energy coefficient, electrical resistivity, electronic thermal range of Fermi level for thermoelectrics42-43, which corresponds to a certain carrier concentration range. conductivity, lattice thermal conductivity and absolute Therefore, a full assessment of a thermoelectric material and temperature, respectively2. its critical material parameters determining the Due to the strong coupling effect among S, ρ and κE, it is difficult to improve zT through an individual optimization thermoelectric performance, fundamentally requires an investigation on the transport properties in a broad range of of these parameters. Therefore, numerous efforts have been focusing on the reduction of lattice thermal conductivity, the carrier concentration. Therefore, the discovery of an effective dopant, enabling such a control in carrier only independent parameter for thermoelectric properties. This strategy has been successfully realized in various concentration, is the key not only for understanding this material as a thermoelectric but also for guiding the materials for enhancing zT through various phonon scattering sources such as nanostructures3-5, alloy defects6-13, performance enhancement. dislocations14-15, In this work, Na-doping on the Sn site is found to liquid phonons16-17, lattice 18-19 increase the hole concentration from 8×1017 in pristine SnS anharmonicity and low sound velocity20. Alternatively, band engineering approaches such as band convergence21-22 to 2×1019 cm-3 in Na0.02Sn0.98S. This is so far the largest 13, 23 and nestification range of carrier concentration achieved in this compound, , have been found to effectively decouple S, ρ and κE for an enhanced power factor by which enables a systematical assessment of the electronic increasing the band degeneracy (Nv). These strategies have and thermal transport properties. The resultant highest been demonstrated in various materials including PbTe21, carrier concentration indeed enables the highest zT of ∼0.65, 24-25 , SnTe11-12, 26, GeTe27-28, Mg2Si29, half-Heusler30-31 and to be achieved in this work in polycrystalline SnS with a Te13, 23, for a significantly enhanced zT. lattice thermal conductivity as low as ~0.4 W/m-K at high Due to a strong lattice anharmonicity, single crystalline temperatures. The electronic transport properties can be well SnSe was reported to show a low lattice thermal understood by a single parabolic band model with acoustic conductivity thus a high zT 18, 32, and therefore attracts scattering at T>400 K, enabling an expectation of zT up to increasing attentions. Polycrystalline SnSe 33-34 was found to 0.8 in this material when the carrier concentration is further show a promising zT as well. As an analogue to SnSe, SnS increased. Moreover, considering the lower-lying valence crystalizes in the same Pnma structure and shows a very bands in SnS 35, an even higher zT can be expected when similar band structure35, receiving increasingly attentions for these bands are engineered to contribute to the charge its potential thermoelectric applications. However, existing transport as well. At T99.99%) at 1193 K for 4 hours, quenching in cold water and then annealing at 923 K for 3 days. Na is used to tune the carrier concentration. The obtained ingots were ground into powders for X-ray diffraction (XRD) and hot pressing. The dense pellets (>95% of theoretical density) with ~12 mm in diameter were prepared by induction heating at 823 K for 30 minutes under a uniaxial pressure of ~80 Mpa44. The electrical transport properties, including electrical resistivity, Seebeck coefficient, and Hall coefficient were simultaneously measured in the temperature range of 300-850 K under helium. The Hall coefficient and electrical resistivity were measured through the van der Pauw technique under a reversible magnetic field of 1.5 T. The Seebeck coefficient was obtained from the slope of the thermopower vs. temperature difference within 0~5 K45. Thermal diffusivity (D) was measured by a laser flash technique (Netzsch LFA457 system). The thermal conductivity was calculated via κ=dDCp, where d is the measured density through the mass and geometric volume of the pellets and Cp is the heat capacity from literature39. The measurement uncertainty for S, ρ and κ is about 5%. The microstructure was characterized by scanning electron microscope (SEM, Phenom Pro) equipped with an energy dispersive spectrometer (EDS). Infrared Fourier transform spectroscopy (FTIR, Bruker Tensor 2) with a diffuse reflectance attachment was used in this work for optical measurements. The optical band gap was estimated by the Tauc method46. 3. Results and discussion
(a)
of Pnma at room temperature (as that of SnSe). Being similar to the case of SnSe, SnS also shows a transition to a high-temperature Cmcm structure, and the phase transition temperature is 858 K 47. This leads the current work to focus on the thermoelectric properties of SnS in the low-temperature orthorhombic structure (T< 850 K). Powder XRD patterns for Sn1-xNaxS (0≤x≤0.03) are shown in Fig. 1b. All of the peaks can be well indexed to the orthorhombic structure of SnS, indicating the high purity of the obtained samples. SEM observations (Fig. 1c-1f, for Sn1-xNaxS with x≤0.02) and mapping on composition by EDS (Fig. 2, for Sn0.98Na0.02S) are carried out, which shows the highest carrier concentration. The SEM observations again confirm the single phase nature and the homogeneity of the sample obtained, which is consistent with the XRD results.
Sn0.98Na0.02S
Sn
10µm
S
Na
(b) Sn1-xNaxS Intensity (a.u.)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 7
x=0.03 x=0.02 x=0.01 x=0.005
x=0 PDF#00-014-0620
10
20 30 40
50 60 70 80
90 100
2θ (deg.)
(c)
x=0
(d)
x=0.005
10µm
(e)
x=0.01 (f)
x=0.02
Fig. 1 Crystal structure of SnS, with red and blue balls for Sn and S atoms, respectively (a). Powder X-ray diffraction patterns (b) and SEM images for Sn1-xNaxS (c-f).
Fig. 2 SEM images with the corresponding mapping on compositions by EDS for Sn0.98Na0.02S.
Temperature dependent Hall carrier concentration (nH) for Sn1-xNaxS (0≤x≤0.03) is shown in Fig. 4a. The room temperature nH is significantly increased from ∼8×1017 cm-3 in pristine SnS to ∼2×1019 cm-3 in the samples doped with Na. It should be noted that the nH achieved in this work is the highest reported so far for this compound, which is much higher than that with Ag-doping 39. This indicates that Na is an effective acceptor in SnS. Since both Na and Ag would release one hole when used to dope SnS, these dopants should offer a similar efficiency on tuning the carrier concentration. Therefore, the higher carrier concentration achieved by Na-doping can be understood by its higher solubility in SnS. It is frequently observed that Na is more soluble than Ag in many IV-VI thermoelectrics including PbTe48-49, PbSe50-51, SnSe32. The broad carrier concentration achieved enables a systematical discussion on the electronic transport properties, to fundamentally understand the material parameters.
The crystal structure of SnS is shown in Fig. 1a, which crystalizes in an orthorhombic structure with a space group
ACS Paragon Plus Environment
11.6
(b) 11.4 Lattice parameter ( Å)
(a) 1E21
-3
nH(cm )
at ect Exp
ion
1E20
nH
11.2
10.8 4.4 4.3
a
4.1
c
4.0 3.9
nNa (cm )
100 200
4.2
150
1E21
-3
250
11.0
1E19 1E20
150
(c) b
ρ (mΩ cm)
0.00
0.01
x
0.02
0.03
1E19
-3
nH (cm )
50 2E19
Fig. 3. Na concentration dependent the expected carrier concentration assuming each Na releases one hole (a), composition dependent lattice parameters (b) and carrier concentration dependent Seebeck coefficient and resistivity (c) for Sn1-xNaxS at room temperature.
Assuming each Na releases one hole, when Na substitutes Sn in SnS, the expected carrier concentration is found to be much higher than those actually obtained (Fig. 3a). This indicates that the carrier concentration easily saturates at a very low value (400 K. The details about the SPB model can be found elsewhere42, 60 According to this model, the density of state effective mass (m*) for Sn1-xNaxS (x≤0.03) are estimated and its temperature dependence is shown in Fig. 4d. m* increases slightly from ~0.9me at 400 K to ~1.2 at 850 K. Similar temperature dependence on the effective mass has been observed in lead chalcogenides61-62 and CuGaTe260 Using the average experimental m*, the SPB model further enables predictions on the Hall carrier concentration dependent Seebeck coefficient and Hall mobility at different temperatures for Sn1-xNaxS (0≤x≤0.03), which agree well with the measurements as shown in Fig. 4e and 4f at both 500 K and 700 K. It is known from the band structure calculations35 that the first valence band maximum is located along the Γ-Z direction in the Brillouin zone, therefore, the band degeneracy Nv is 2. Other valence bands (along Γ-Y direction and Z point) are at least 0.2 eV (8kBT at 300 K) lower in energy, and therefore, are not sufficiently involved to contribute to the transport of charge carriers since the samples obtained here all show a weakly degenerated conduction (i.e. a high Seebeck coefficient). Using the band degeneracy Nv of 2, the SPB model enables an estimation of band effective mass mb* via m*=Nv2/3mb* at any temperature. Further assuming the band to be isotropic, the
ACS Paragon Plus Environment
ACS Applied Materials & Interfaces
Sn1-xNaxS 100 300
(c)
400
500
600
700
800
300
900
T (K)
4
(d)
Sn1-xNaxS
5
x=0
3
400
500
600
700
x=0.01 x=0.02
1
0.9
Sn
5
Ag
0.0
5
S
x=0.03
850K
4
m*=1.2me
3
700K m*=1.1me
2 500K
0 400
500
600
T (K)
700
800
900
Prediction This work Expt.Data Tan et.al.
m*=1.0me
1
300
800
T (K)
2
PF (µ W/cm-K )
x=0.005
2
6
900
0 18 10
19
10
nH(cm-3)
20
10
Fig. 5 Temperature dependent Seebeck coefficient (a), electrical resistivity (b) and powder factor (c) for Sn1-xNaxS (0≤x≤0.03), with a comparison to those of Sn0.95Ag0.05S 39, and Hall carrier concentration dependent power factor at different temperatures with a comparison to the model prediction (d).
The measured Seebeck coefficient and electrical resistivity for Sn1-xNaxS (x≤0.03), as a function of temperature, are shown in Fig. 5a and 5b, respectively. The Seebeck coefficient for all the samples is positive, indicating a p-type conduction. The Seebeck coefficient and electrical conductivity significantly decrease with the increase of nH due to Na-doping. It is found that the much higher nH achieved in Na-doped samples leads the resistivity to be much lower than that of Ag-doped ones 39. The temperature dependent power factor (PF) for Sn1-xNaxS (0≤x≤0.03) is shown in Fig. 5c. Comparing to pristine SnS, Na-doping significantly enhances PF in the entire temperature range, due to the increased carrier concentration(Fig. 4a). Moreover, the nominal content (≥0.5%) of Na in this work is much higher than that is soluble (~0.1%) at room temperature, therefore the rest dopant in the material likely continues to be dissolved and leads to a further increase in carrier concentrations at high temperatures (~700 K in this work). Such a solubility increase for an increase in carrier concentration at high temperatures is frequently observed in thermoelectrics such as PbTe:Ag63, PbTe:Na64 and PbSe:Cu65. It should be noted that intrinsic excitation of minority carriers is unlikely strong in SnS, because its band gap of ~1.1 eV is as large as 15 kBT even at the highest measured temperature in this work while the SPB model estimated Fermi level at 650 K is only about 2 kBT above the valence band edge. Otherwise, comparable power factors for both doped and undoped materials would be expected, which is not supported either by this work with Na-doping or by literature work with Ag-doping39. In addition, the continuous decrease in
(a)
(b)
Sn1-xNaxS
1.2
1.0
1.0
x=0 x=0.005 x=0.01 x=0.02 x=0.03
0.8 0.6
Sn
0 .9 5
Ag
0.0 5
κ L (W/m-K)
200
T -1
0.8
Sn 0.9 5A
0.6 min
S
kL
g 0.0 5S
0.4
0.4 300
400
500
0.8
600
T (K)
700
800
300
900
1.0
Sn1-xNaxS
(c)
(d)
0.8
0.6 x=0 x=0.005 x=0.01 x=0.02 x=0.03
0.4
400
500
600
T (K)
700
800
900
Prediction This work Expt.Data Tan et.al.
850K
0.6
m*=1.2me
zT
2
10
S
S Ag 0.05 Sn 0.95
300
Sn0.95Ag0.05S
0 .05
400
ρ (mΩ cm)
S (µ V/K)
500
Ag
(b)
x=0.02 x=0.03
0.4
0.9 5
10
x=0 x=0.005 x=0.01
600
Sn
3
700
κ (W/m-K)
(a)
temperature dependent lattice thermal conductivity (Fig. 6b) suggests a weak bipolar conduction in this work as well, since intrinsic excitation of minority carriers usually increases thermal conductivity. Using the average m* and Edef (Fig. 4d), the SPB model further enables a prediction on the carrier concentration dependent power factor at any temperatures. Shown in Fig. 5d are the cases for 500, 700 and 850 K. It is seen that both power factor and the carrier concentration required for maximizing power factor, increase with increasing temperature, which is typical observed for thermoelectrics66. Temperature dependent total thermal conductivity and its lattice component for Sn1-xNaxS (x≤0.03) are shown in Fig. 6a and 6b, respectively. The lattice thermal conductivity is obtained by subtracting the electronic thermal conductivity (κE) from the total thermal conductivity via the Wiedemann-Franz Law, κE=LT/ρ, where L is the Lorenz factor determined by SPB model. The lattice thermal conductivity is comparable for all the samples due to the low concentration of impurities. It is found that Umklapp processes dominate scattering of phonons in all the samples, since κL decreases with increasing temperature via a T-1. It is interesting to note that SnS intrinsically shows a κL as low as its amorphous limit estimated by the Cahill model67, when the literature sound velocities (3368 m/s for longitudinal and 1537, 2368 m/s for transverse branches) 68 are used. Such an intrinsically low κL has been reported in both SnS35 and its analogue compound SnSe, with the same origin from a strong lattice anharmonicity18, 38, 69.
zT
deformation potential coefficient, Edef, measuring the strength of carrier scattering by acoustic phonons, can be determined form the measured mobility. Edef is found to decrease with increasing temperature as shown in Fig. 4d. It should be noted that Edef at temperatures slightly higher than 400 K can be overestimated due to the existence of carrier scattering by the boundary potential (a lower mobility as compared to that purely by acoustic scattering) as discussed above.
PF (µ W/cm-K2)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 7
0.2
700K
m*=1.1me
0.2
500K
0.0 300
400
500
600
T (K)
700
800
900
0.0 18 10
m*=1.0me 19
10
-3
nH(cm )
20
10
Fig. 6 Temperature dependent total thermal conductivity (a), lattice thermal conductivity (b), with a comparison to those of Sn0.95Ag0.05S39, thermoelectric figure of merit (zT) (c) and the predicted zT versus Hall carrier concentration and temperature (d) for Sn1-xNaxS (x≤0.03).
The temperature dependent figure of merit (zT) for Sn1-xNaxS (x≤0.03) is given in Fig. 6c. zT of all the samples increases with increasing temperature. A significantly enhanced zT up to 0.65 is obtained in Na-doped SnS in this work, which is a result of the synergy of both PF-enhancement and the intrinsic low κL. The SPB model further enables a prediction in carrier concentration dependent zT, where an average measured lattice thermal conductivity is used. As shown in Fig. 6d, the predicted zT agrees well with the experimental results. It is further expected, according to the model, that a peak zT up to 0.8
ACS Paragon Plus Environment
Page 5 of 7
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
can be achieved at 850 K, once the carrier concentration is further increased to ~1020 cm-3 through doping by more soluble dopants, strongly indicating SnS as a promising low-cost thermoelectric material.
Fig. 7. Temperature dependent Seebeck coefficient (a), electrical resistivity (b), thermal conductivity (c) and zT (d) for Sn1-xNaxS during heating and cooling, where Sn0.98Na0.02S is measured twice.
In order to investigate the thermal stability of Sn1-xNaxS, thermoelectric properties are measured during both heating and cooling for all the samples and Sn0.98Na0.02S is measured twice. The results are shown in Fig. 7. Although Seebeck coefficient, resistivity and thermal conductivity show hysteresis particularly at T
