Anomalous transport and thermoelectric performances of CuAgSe compounds
A.J. Hong
a, L. Li
a, H.X. Zhu
a, X.H. Zhou
a, Q.Y. He
b, W.S. Liu
c, Z.B. Yan
a, J.-M. Liu ⁎
,a, Z.F. Ren
caLaboratory of Solid State Microstructure, Nanjing University, Nanjing 210093, China
bInstitute for Advanced Materials and Laboratory of Quantum Engineering and Quantum Materials, South China Normal University, 510006 Guangzhou, China
cDepartment of Physics and TcSUH, University of Houston, Houston, TX 77204, USA
a b s t r a c t a r t i c l e i n f o
Article history:
Received 22 January 2014
Received in revised form 28 February 2014 Accepted 28 March 2014
Available online 22 April 2014
Keywords:
CuAgSe
Phonon-glass superionic Thermoelectric performances
Copper silver selenide has two phases: low-temperature semimetal phase (α-CuAgSe) and high-temperature phonon-glass superionic phase (β-CuAgSe). In this work, the electric transport and thermoelectric properties of the two phases are investigated. It is revealed thatβ-CuAgSe is ap-type semiconductor and exhibits low thermal conductivity whileα-CuAgSe shows metallic conduction with dominantn-type carriers and low electri- cal resistivity. The thermoelectricfigure of meritzTof the polycrystallineβ-CuAgSe at 623 K is ~0.95, suggesting that superionic CuAgSe can be a promising thermoelectric candidate in the intermediate temperature range.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
The thermoelectric (TE) energy conversion technology has received much attention in the past decades[1–9]. For a TE device sandwiched between a cool bath with temperatureTcand a heat bath with temper- atureTh, the theoretical maximum conversion efficiency is governed by [10]:
η¼Th−Tc Th
M−1 MþTc
Th
; M¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1þ1
2z Tð hþTcÞ r
¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1þzT
p ; ð1Þ
wherezTis the dimensionless parameter known as thefigure of merit, which can be related to the physical properties of TE materials:
zT¼ S2T
ρ κð lþκeÞ; ð2Þ
whereS,T,ρ,κl, andκestand for thermopower (Seebeck coefficient), absolute temperature, electrical resistivity, lattice and electronic ther- mal conductivities, respectively. As a promising TE material, a highzT value is appealed via various approaches in terms of materials design and microstructure synthesis.
In the past decades, substantial efforts have been made in searching for promising TE materials of high performance including classical semi- conductors such as PbTe, CoSb3, and SiGe[2]. Along this line, some
superionic semiconducting compounds have been demonstrated to be promising candidates for TE applications. Recently, Liu et al.[9]reported that the superionicβ-Cu2Se has azTvalue as high as ~1.5 atT~ 1000 K. It is noted that Cu2Se has two phases: low-Tα-phase and high-Tβ-phase [11–14]. Theα-phase of Cu2Se is very complicated in structure and diverging crystallographic data can be found in literature. Nguyen et al.[15]presented a monoclinic (space groupP21/c) structure for the Cu2Se compound. According to Liu et al.[16], the Cu2Se compound has a monoclinic structure (space group: C2/c; #15). Theβ-Cu2Se follows the Fm3mlattice group withfixed Cu occupation in thefcclattice skeleton of the Se atoms.Fig. 1(a) shows a schematic of the lattice structure of the high-Tβ-phase where Se atoms keep thefcclattice but each Cu atom occupies randomly one of the four vertices and center of a tetrahedral unit inside the Sefccunit. In this case, the Cu atoms are random in occupation (occupation-disordered), forming the so-called
‘phonon-liquid’[9]superionic phase (β-Cu2Se) in the high-Trange.
The underlying physics lies in the fact that the occupation-disordered Cu atoms can easily move in the lattice under externalfields (tempera- turefield or electricfield), allowing strong scattering of lattice phonons and thus low lattice thermal conductivity (κl), while the low electric resistivity (ρ) is substantially retained.
The excellent TE performances of Cu2Se compound have stimulated substantial efforts in searching for more TE materials with similar superionic structure and preferred TE performances. An immediate candidate is CuAgSe, i.e., randomly Ag-substituted Cu2Se compounds.
Similar to Cu2Se, CuAgSe also has two phases: low-Tα-phase and high-Tβ-phase[17–22]. The low-Tα-phase has a unique layered structure consisting of alternating stacking of the Ag and CuSe layers, allowing probably high mobility of Ag atoms (Fig. 1(b)). Theβ-phase
⁎ Corresponding author.
E-mail address:[email protected](J.-M. Liu).
http://dx.doi.org/10.1016/j.ssi.2014.03.025 0167-2738/© 2014 Elsevier B.V. All rights reserved.
Contents lists available atScienceDirect
Solid State Ionics
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / s s i
is similar to the superionic structure of Cu2Se with the mobile Ag/Cu cations randomly distributing among the tetrahedral sites (Fig. 1(c)).
This may allow even lower lattice thermal conductivity thanβ-Cu2Se, due to the random half-substitution of Cu ions by Ag ions and thus higher degree of occupation disorder, and thus an investigation of the structural and TE behaviors is interested.
Also different from the low-Tphase of Cu2Se, Ishiwata et al.[18]re- ported that CuAgSe at lowTis a semimetal and a TE material. Usually, semimetals are not preferred for highzTvalue, since the thermal broad- ening of the Fermi distribution function for most semimetals almost equally activates two kinds of carriers with opposite signs. This effect re- sults in a low thermopower (S). However,first-principles calculations predict that the CuAgSe phase possesses a significant electron–hole asymmetry in the band structure near the Fermi level[18], which may benefit to thermopower enhancement. In addition, a semimetal would have lower resistivity than conventional TE semiconductors.
Unfortunately, due to the drastic difference in the structure and atom occupation, the two phases are expected to have very different electrical/thermo transports and TE properties, which are highly concerned from the point of views of TE applications of CuAgSe in com- parison with Cu2Se. These motivations certainly deserve explorations, which will be the main purpose of this work. Nevertheless, instead of investigating the effects of substitution or microstructural controls, we pay more attention to the microscopic mechanisms for the electrical/
thermo transports and TE properties associated with the phase transfor- mations for the CuAgSe compounds.
2. Experimental details
The sample preparation follows the standard procedure. Polycrystal- line CuAgSe samples were prepared by melting high-purity powder of species Cu, Ag, and Se enclosed in a fused silica tube in vacuum. The powder mixture in stoichiometric ratios was heated to 573 K at a rate
1 K/min and maintained at this temperature for 24 h, and then slowly heated up to 973 K at a rate 2 K/min and maintained at 973 K for 24 h and then slowly cooled down to 773 K in 12 h and held there for 12 h.
Then the spontaneous cooling down to room temperature yielded the pre-samples. These pre-samples were ground intofine powder using an agate jar and plunger. The powder was subjected to direct current hot-press sintering (DCHP-2000A-04, USA) around 673 K under a pres- sure of 78 MPa for 6 min, producing ingots of 12.8 mm in diameter and 10 mm in height. Each ingot was cut into rectangular bars with dimen- sions of 3 × 4 × 8 mm3for electrical transport measurements and disks with dimensions of 12.6 mm in diameter and 2.1 mm in thickness for thermal transport measurements.
The structure and crystallinity of the as-prepared samples were checked using the X-ray diffraction (XRD; Bruker Corporation) equipped with Cu-Kαradiation. The longitudinal resistivity and Hall resistivity as a function of magneticfieldBwere measured using the standard four- probe method. The electric resistivity and thermopower were mea- sured using an ULVAC ZEM-3 system (Japan) with a sealed chamber filled with a low pressure He ambient. The temperature difference across the sample was 25 K, while sufficient data were obtained by running the measurements from different initiating temperatures to guarantee the data reproducibility. The thermal diffusivity (D) and heat capacity (CP) were measured using the laser-flash thermometer (LFA457, Netzsch). The total thermal conductivity was calculated usingκ=D×CP×d, wheredis the sample density. The differential scanning calorimetry (DSC) of these samples was carried out using a Netzsch STA449F3 device, at the heating rate 10 K/min.
It should be mentioned that the synthesizing conditions for the samples are optimized and the data to be presented below can be well reproduced from sample to sample.
3. Results and discussion
3.1. Structure and composition analysis
X-ray power diffraction patterns are shown inFig. 2. At room tem- perature, all the main peaks can be indexed by the CuAgSe characteristic peaks (PDF#10-0451) with an orthorhombic structure. There is one dif- fraction peak of CuAgSe with a tetragonal structure. CuAgSe was initially reported to have a tetragonal structure[23]. Later studies reveal that CuAgSe has an orthorhombic structure[21,24]. In this work, CuAgSe has the above two structures and the orthorhombic structure is primary.
The existence of the tetragonal structure may be due to inadequate response.
3.2. Electrical transport
The typical data on electrical resistivity (ρ) as a function ofTare plot- ted inFig. 3(a). It is clearly seen that theT-dependence is evidenced with two distinct regions separating atT=Tαβ~ 470 K. In the low-Tregion, a metallic conduction behavior is shown, consistent with earlier report [18], while a semiconducting conduction is observed in the high-Tregion.
It is clear that the change of the electrical transport is associated with the Fig. 1.Schematic drawing of lattice structures of compoundsβ-Cu2Se (a),α-CuAgSe (b),
andβ-CuAgSe (c).
Fig. 2.XRD pattern for CuAgSe.
A.J. Hong et al. / Solid State Ionics 261 (2014) 21–25
α–βstructural transformations which are evidenced with the DSC heat flow as a function ofTshown inFig. 3(b). The sharp valley does appear atTαβ~ 470 K.
If one compares the measuredρvalues inside the two phase regions with those for polycrystalline Cu2Se, it is seen that the CuAgSe has an electrical resistivity comparable with that of the Cu2Se. Typically, Cu2Se has theρ~ 10μΩ⋅m and 30μΩ⋅m atT~ 323 K and 600 K,[9]re- spectively, while CuAgSe hasρ~ 8μΩ⋅m and ~58μΩ⋅m at the two tem- peratures. Here the lowρvalues ofα-CuAgSe are ascribed to the fact that theα-CuAgSe phase is a semimetal and has high carrier mobility, to be confirmed below. The slightly higher electrical resistivity of the β-CuAgSe phase is due to the random mixing of Cu and Ag atoms, which brings additional scattering to electron transport with respect to theβ-Cu2Se phase, which is why the electrical resistivity ofβ-CuAgSe is higher than that ofβ-Cu2Se.
3.3. Two types of carriers
In accompanying with theα–βtransformations atTαβ, onefinds the substantial change of the thermopower (Seebeck coefficientS) as a function ofT, as plotted inFig. 3(c). TheS(T) increases rapidly from
−95μV/K atT~ 323 K to nearly zero atT=TS~ 400 K, and then to
~ 130μV/K atT~Tαβ. In theβ-CuAgSe phase theS(T) becomes nearly unchanged and the saturatedSvalue is ~ 190μV/K. We are interested in the sign reversal ofSupon increasingT. This sign reversal from nega- tive value at lowTto positive value at highTwas reported earlier[18], with the reversal temperatureTS~ 400 K. This feature is not observed in Cu2Se and it reflects the existence of two types of carriers in the α-phase, as claimed in earlier work[18]. The electrons as carriers are dominant in the low-Trange while the holes are dominant in the high-Trange. For a simple two-carrier system, the thermopower is written asS= (σpSp+σnSn) / (σp+σn), whereσp,Sp,σn, andSnare thep-type carrier electrical conductivity,p-type carrier thermopower, n-type carrier electrical conductivity, andn-type carrier thermopower,
respectively. SinceSpN0 andSnb0, the sign reversal ofSwith increasing Tis reasonably understood. Towards the high-Tβ-CuAgSe phase, the dominant carriers are thep-type and a positiveSis expected in this phase.
3.4. High mobility of n-type carriers in theα-CuAgSe phase
Before going to the TE properties of the high-Tβ-CuAgSe phase, we come to investigate the electrical transport of theα-CuAgSe phase.
Given the existence of two types of carriers in theα-phase, we measure the Hall resistivity to evaluate the density and mobility of the dominant n-type carriers at room temperatureT= 300 K. To properly estimate the carrier concentrationnand mobilityμ, we assume that the magne- toresistance effect is dominated by the high-mobility carriers so that the conductivity tensor can be analyzed using the two-carrier model.
In this case,μB≪1 is assumed for the low-mobility carriers, whereB is the magneticfield. In the standard geometry for the Hall effect, the conductivity tensors can befitted by the following equations[25]:
σxxð Þ¼enB xxμxx 1 1þðμxx BÞ2þCxx
σxyð Þ¼B enxyμ2xxB 1
1þðμxy BÞ2þCxy
h i; ð3Þ
whereσxx(σxy),nxx(nxy),μxx(μxy), andCxx(Cxy), denote the conductivity tensors, the carrier concentration components, the carrier mobility components, and the low-mobility components for the conductivity tensor, respectively. The conductivity tensors are related toρxxandρxy:
σxx¼ρxx=ðρ2xxþρ2xyÞ
σxy¼−ρxy=ðρ2xxþρ2xyÞ; ð4Þ
whereρxxandρxyare the longitudinal resistivity and Hall resistivity, respectively.
TheB-dependencies ofρxxandρxyat room temperature are plotted inFig. 4(a) and (b), respectively. Both theρxxandρxyincrease with increasingB, as well known. These data arefitted using Eqs.(3) and (4)and thefitted data forσxxandσxyare presented inFig. 4(c) and (d), respectively, where the solid lines represent thefitted curves.
The as-evaluated carrier concentrationsnxx= 3.9587 × 1018cm−3 (nxy= 3.0996 × 1018cm−3) and carrier mobilityμxx= 2018 cm2/V⋅s (μxy = 2204 cm2/V⋅s). It is seen that the Hall mobility of low-T α-phase is far higher than that ofα-Cu2Se (μxx~ 20 cm2/V⋅s)[9]while the carrier density of the low-Tα-phase is much lower than that of α-Cu2Se (nxx= 3.2 × 1020cm−3)[9]. Surely, this high carrier mobility has its structural origin, noting that theα-CuAgSe phase has a layer- like structure, as shown inFig. 1(b), consistent with the semi-metallic band structure. There is the experimental evidence for the semi- Fig. 3.MeasuredT-dependencies of electrical resistivityρ(a), heatflow (b), and
thermopowerS(c). The negative heatflow indicates that the sample absorbs the thermal power.
Fig. 4.Measured longitudinal electrical resistivityρxx(a) and Hall resistivityρxy(b) as a function of magneticfieldBat room temperature. The evaluated longitudinal conductivity σxx(c) and Hall conductivityσxy(d) as a function ofB. The solid curves are thefitting data using Eqs.(3).
metallic band structure with the light-mass electron carriers which can result in high carrier mobility.
3.5. Thermal transport
We then turn to the thermal transport behavior which is critical for the TE performances. The total thermal conductivity was calculated usingκ=D×CP×d, wheredis the sample density. The sample density is 7.858 g/cm3. The specific heat at constant pressureCPis showed in Fig. 5(a). The measured thermal conductivityκas a function ofTis plotted inFig. 5(b), noting that this data set is consistent from different runs in different institutes. Basically, the thermal conductivity is high in theα-phase region and low in theβ-phase region, with a gradual tran- sition across theα–βphase transformations aroundTαβ. Regarding the T-dependencies ofκin the two regions, one sees a gradual but weak in- crease ofκwith increase inTin theα-phase region while no remarkable T-dependence in theβ-phase region is observed. In particular, theκ value of theβ-phase is as low as ~0.47 W m−1K−1, much lower than that ofβ-Cu2Se in thisT-range (0.9 W m−1K−1atT~ 600 K)[9]. Such a lowκvalue is highly preferable for thermoelectric applications.
A reasonable explanation for the low κ value is the specific superionic (liquid-like) lattice structure of theβ-phase in which the mixed Cu and Ag atoms distributed randomly over tetrahedral sites, which substantially weakens the thermal transport. For the low-T α-phase, the layer lattice structure however allows relatively easy pho- non and electron thermal transport. Furthermore, we look at the lattice phonon and carrier contributions to the thermal conductivity. One esti- mates the carrier thermal conductivityκefrom the Wiedemann–Franz lawκe=LσTwhere the Lorenz numberL= 2.1 × 10−8V−2K−2is taken. The evaluated lattice thermal conductivityκl=κ−κeas a function ofTis plotted inFig. 5(b). Several features regarding the thermal conductivity mechanism can be seen. First, in the low-T α-phase, the highκevalues are expected to be associated with the semimetal behavior, while theκeof the high-Tβ-phase is much lower. Second, in the high-Tβ-phase, theκlcan be as low as ~ 0.27 W m−1 K−1 which are much lower than the values ofβ-Cu2Se
(0.4–0.6 W m−1K−1)[9]. Local atomic jumps and rearrangement of the‘liquid’β-CuAgSe inhibit the propagation of transverse waves, thereby leading to a significant reduction in the lattice thermal conduc- tivity[9].
3.6. Thermoelectric performance
Finally, we present thefigure-of-meritzTas a function ofT, as shown inFig. 6. ThezTvalue atT= 323 K is ~0.28. In between 323 K and 400 K (bTαβ), the very lowzTvalue is attributed to the existence of two types of carriers which lead to the sign reversal of thermopowerS. However, in theβ-phase region, thezTvalue increases with increasingTand reaches 0.95 atT= 623 K, while thezTvalue of Cu2Se is ~0.6 at 650 K [9].
In addition, the phase stability of this CuAgSe compound was checked from two aspects. First, the same sample was submitted to four consecutive TE measuring runs using the ULVAC ZEM-3 system (Japan) with a sealed chamberfilled with a low pressure He ambient.
The measured electrical resistivity and thermopower for three runs as a function ofTfrom 300 K to 650 K are plotted inFig. 7(a) and (b), re-
Fig. 5.MeasuredT-dependence of the specific heat at constant pressureCP(a) and the thermal conductivity and extracted lattice thermal conductivity (b).
Fig. 6.Thermoelectricfigure of meritzTas a function ofT.
Fig. 7.Electrical resistivityρ(a) and thermopowerS(b) as a function ofT, as evaluated from consecutive runs (run Nos. 1, 3, & 4) of the TE measurements from 300 K to 650 K.
A.J. Hong et al. / Solid State Ionics 261 (2014) 21–25
spectively, for comparisons. No remarkable differences among these data are seen.
4. Conclusion
In conclusion, we have investigated the structural, electrical, and thermal transport properties of CuAgSe compound. Theα–βstructural transformation happens atT~ 470 K, accompanied with remarkable dif- ference in the electrical and thermal transports. It has been revealed that the low-Tα-phase accommodates dominantn-type carriers with minorp-type carriers, while the high-Tβ-phase is dominated with p-type carriers. The high-Tβ-phase exhibits good thermoelectric prop- erties, in particularly very low thermal conductivity, which allow the superionic CuAgSe to be a promising candidate for potentially interme- diate temperature thermoelectric applications.
Acknowledgment
This work was supported by the National 973 Projects of China (Grant No. 2011CB922101), the National Natural Science Foundation of China (Grants Nos. 11234005, 11374147, and 51372092), and the Priority Academic Program Development of Jiangsu Higher Education Institutions, China.
References
[1] Y.L. Pei, H.J. Wu, J.H. Sui, J. Li, D. Berardan, C. Barreteau, L. Pan, N. Dragoe, W.S. Liu, J.Q.
He, L.D. Zhao, Energy Environ. Sci. 6 (2013) 1750.
[2] G.J. Snyder, E.S. Toberer, Nat. Mater. 7 (2008) 105.
[3] K. Biswas, J.Q. He, I.D. Blum, C.I. Wu, T.P. Hogan, D.N. Seidman, V.P. Dravid, M.G.
Kanatzidis, Nature 489 (2012) 414.
[4] Y.Z. Pei, X.Y. Shi, A. LaLonde, H. Wang, L.D. Chen, G.J. Snyder, Nature 473 (2011) 66.
[5] M.S. Dresselhaus, G. Chen, M.Y. Tang, R.G. Yang, H. Lee, D.Z. Wang, Z.F. Ren, J.-P.
Fleurial, P. Gogna, Adv. Mater. 19 (2007) 1043.
[6] K.F. Hsu, S. Loo, F. Guo, W. Chen, J.S. Dyck, C. Uher, T. Hogan, E.K. Polychroniadis, M.
G. Kanatzidis, Science 303 (2004) 818.
[7] B. Poudel, Q. Hao, Y. Ma, Y.C. Lan, A. Minnich, B. Yu, X. Yan, D.Z. Wang, A. Muto, D.
Vashaee, X.Y. Chen, J.-M. Liu, M.S. Dresselhaus, G. Chen, Z.F. Ren, Science 320 (2008) 634.
[8] J.P. Heremans, V. Jovovic, E.S. Toberer, A. Saramat, K. Kurosaki, A. Charoenphakdee, S.
Yamanaka, G.J. Snyder, Science 321 (2008) 554.
[9] H.L. Liu, X. Shi, F.F. Xu, L.L. Zhang, W.Q. Zhang, L.D. Chen, Q. Li, C. Uher, T. Day, G.J.
Snyder, Nat. Mater. 11 (2012) 422.
[10]A.F. Loff, Semiconductor Thermoelements and Thermoelectric Cooling, Infosearch Publishers, London, 1957.
[11] R.D. Heyding, R.M. Murray, Can. J. Chem. 54 (1976) 841.
[12] M. Oliveria, R.K. McMullan, B.J. Wuensch, Solid State Ion. 28 (1988) 1332.
[13] W. Borchert, Z. Kristallogr. 106 (1945) 5.
[14] A.L.N. Stevels, F. Jellinek, Recl. Trav. Chim. 90 (1971) 273.
[15] M.C. Nguyen, J.H. Choi, X. Zhao, C.Z. Wang, Z.Y. Zhang, K.M. Ho, PRL 111 (2013) 165502.
[16] H.L. Liu, X. Yuan, P. Lu, X. Shi, F.F. Xu, Y. He, Y.S. Tang, S.Q. Bai, W.Q. Zhang, L.D.
Chen, Y. Lin, L. Shi, H. Lin, X.Y. Gao, X.M. Zhang, H. Chi, C. Uher, Adv. Mater. 25 (2013) 6607.
[17] Sh.K. Kiazimov, K.M. Jafarov, Yu.G. Asadov, Phase Trans. 21 (1990) 11.
[18] S. Ishiwata, Y. Shiomi, J.S. Lee, M.S. Bahramy, T. Suzuki, M. Uchida, R. Arita, Y.
Taguchi, Y. Tokura, Nat. Mater. 12 (2013) 512.
[19] R.F. Kadrgulov, A.I. Livshits, R.F. Yorkshibaev, Sov. Phys. Solid State 34 (1992) 1146.
[20] K. Chrissafis, N. Vouroutzis, K.M. Paraskevopoulos, N. Frangis, C. Manolikas, J. Alloys Compd. 385 (2004) 169.
[21] A.J. Frueh, G.K. Czamanke, C.H. Knight, Z. Kristallogr. 108 (1957) 389.
[22] S. Miyatani, J. Phys. Soc. Jpn. 34 (1973) 423.
[23] J.W. Earley, Am. Mineral. 35 (1950) 337.
[24] R.B. Baikulov, Y.G. Asadov, Inorg. Mater. 41 (2005) 338.
[25] H. Takahashi, R. Okazaki, Y. Yasui, I. Terasaki, Phys. Rev. B 84 (2011) 205215
