Enhanced thermoelectric properties of ga and in co added cosb3 based skutterudites with optimized chemical composition and microstructure

Enhanced thermoelectric properties of Ga and In Co added CoSb3 based skutterudites with optimized chemical composition and microstructure Enhanced thermoelectric properties of Ga and In Co added CoSb3[.]

Enhanced thermoelectric properties of Ga and In Co-added CoSb3-based

skutterudites with optimized chemical composition and microstructure

Seongho Choi, Ken Kurosaki, Guanghe Li, Yuji Ohishi, Hiroaki Muta, Shinsuke Yamanaka, and Satoshi Maeshima

Citation: AIP Advances 6, 125015 (2016); doi: 10.1063/1.4971819

View online: http://dx.doi.org/10.1063/1.4971819

View Table of Contents: http://aip.scitation.org/toc/adv/6/12

Published by the American Institute of Physics

Articles you may be interested in

Vibrational mean free paths and thermal conductivity of amorphous silicon from non-equilibrium molecular dynamics simulations

AIP Advances 6, 121904121904 (2016); 10.1063/1.4968617

Role of fringing field on the electrical characteristics of metal-oxide-semiconductor capacitors with co-planar and edge-removed oxides

AIP Advances 6, 125017125017 (2016); 10.1063/1.4971845

Control of Rayleigh-like waves in thick plate Willis metamaterials

AIP Advances 6, 121707121707 (2016); 10.1063/1.4972280

Nonlinear resonance converse magnetoelectric effect modulated by voltage for the symmetrical

magnetoelectric laminates under magnetic and thermal loadings

AIP Advances 6, 125016125016 (2016); 10.1063/1.4971821

Enhanced thermoelectric properties of Ga and In Co-added CoSb3-based skutterudites with optimized chemical

composition and microstructure

Seongho Choi,1, aKen Kurosaki,1,2, bGuanghe Li,1Yuji Ohishi,1Hiroaki Muta,1 Shinsuke Yamanaka,1,3and Satoshi Maeshima4

1Graduate School of Engineering, Osaka University, Suita, Japan

2JST, PRESTO, 4 18 Honcho, Kawaguchi, Saitama, Japan

3Research Institute of Nuclear Engineering, University of Fukui, Tsuruga, Japan

4Business Unit, Panasonic Semiconductor Solutions Co., Kameoka 621 0018, Japan

(Received 14 October 2016; accepted 23 November 2016; published online 15 December 2016)

Skutterudite compounds such as Co antimonite (CoSb3) contain cage-like voids inside crystal structure, which can be completely or partially filled with various different atoms, including group 13 elements The multiple filling approach is known as an effec-tive way of reducing lattice thermal conductivity (κlat), which results in a high value

of the thermoelectric dimensionless figure of merit (zT ) In this work, enhanced zT

was achieved for the Ga and In co-added CoSb3samples with a preferable microstruc-ture and the nominal composition (Ga0.8In0.2)xCo4Sb12(x = 0.050.45) Although all

added In atoms occupied exclusively the void sites, the Ga species filled both the void and Sb sites of CoSb3 Moreover, Ga atoms added in the quantities exceeding the

sol-ubility limit precipitated as GaSb nanoparticles The sample with x = 0.45 was

charac-terized by the largest filling factions of Ga and In as well as the unique microstructure, consisting of microscale grains of the skutterudite phase and corresponding amounts

of the GaSb nanoparticles The Ga and In co-added skutterudite samples with opti-mized chemical composition and microstructure maintained high carrier mobility and sufficiently low κlatvalues, resulting in zT > 1.1, one of the best values for the skutteru-dites filled with group 13 elements © 2016 Author(s) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license ( http://creativecommons.org/licenses/by/4.0/ ) [http://dx.doi.org/10.1063/1.4971819]

I INTRODUCTION

Thermoelectric (TE) devices, which can convert waste heat into electrical power, have been considered clean and sustainable power sources.1The efficiency of TE devices is determined by the

properties of TE materials and represented by the dimensionless figure of merit, zT = S2T ρ−1κ−1,

where S is the Seebeck coefficient, T is the absolute temperature, ρ is the electrical resistivity, and κ

is the total thermal conductivity (κ= κlat+ κel, the lattice and electronic contributions, respectively) The reduction of κlatplays an important role in the enhancement of zT because S, ρ, and κelare closely related to each other as functions of the carrier concentration In order to achieve low κlatvalues, it is imperative to select materials with complex structures (such as large unit cells) containing disorders and various atomic types.2 4Many bulk materials have been investigated as possible advanced TE materials,5 10including filled skutterudites Skutterudites have the composition MX3 , where M is a metal atom (such as Co, Rh, or Ir), and X represents a pnicogen atom such as P, As, or Sb These

compounds are characterized by a body-centered cubic lattice with 32 atoms in the unit cell and

space group Im-3 The skutterudite structure contains two large voids that can be filled by other

a Electronic mail: shchoi@see.eng.osaka-u.ac.jp

b Electronic mail: kurosaki@see.eng.osaka-u.ac.jp

2158-3226/2016/6(12)/125015/9 6, 125015-1 © Author(s) 2016

125015-2 Choi et al. AIP Advances 6, 125015 (2016)

atoms When additional atoms R are incorporated into the voids, the compound formula becomes

RM4X12, corresponding to filled skutterudites The filler element R is bonded weakly to other atoms and vibrates independently of them, producing an effect called rattling Therefore, the behavior

of R introduced into the voids contributes to the phonon scattering, which leads to a reduction of

κlat In particular, CoSb3-based filled skutterudites are considered promising TE materials Alkali metals,11 , 12alkaline earth elements,13 – 15lanthanides,16 – 24electronegative elements,25 , 26and group

13 elements27 – 41are known as good candidates for filler elements Among group 13 elements, indium

(In), as a co-dopant in multiple filled skutterudites, show excellent TE properties of maximum zT

values of 1.2-1.4, in which co-dopants such as Barium, Cerium and Ytterbium easily react with air and thus the special process is required to apply for mass production.38,41–44In contrast, gallium (Ga), as a same group element with In, is nontoxic and not reactive with both air and water at room temperature Furthermore, since Ga indicates the unique behavior in CoSb3such as dual-site occupancy and the desirable electronic band engineering, it is worth studying as a co-dopant with In in CoSb3-based skutterudites.29,33,34,45Our group has previously investigated the TE properties of Ga and In co-added CoSb3-based skutterudites27and found that (1) Ga and In could be simultaneously introduced into the voids of the skutterudite structure; (2) Ga could also occupy Sb sites; and (3) Ga and In added in the amounts exceeding the filling limit precipitated as (Ga,In)Sb nanoinclusions These three phenomena contributed to the κlatreduction significantly, and thus the maximum value of zT close to unity was

obtained for the nominal composition Ga0.20In0.30Co4Sb12.27In the present study, we have attempted

to improve the zT of Ga and In co-added CoSb3by optimizing both its chemical composition and microstructure After characterizing the TE properties of (Ga0.8In0.2)xCo4Sb12(x = 0.050.45), we found that the sample corresponding to the stoichiometric composition of x = 0.45 has an optimized

carrier concentration together with a preferable microstructure, leading to the maximum enhancement

of zT equal to 14% (> 1.1).

II EXPERIMENT

Polycrystalline samples were prepared from the proper amounts of GaIn alloy (99.999%), Co (99.9%), and Sb (99.99%) chunks via direct reactions inside sealed silica tubes Since GaIn alloy is characterized by around the eutectic composition, the nominal compositions studied in this work were defined as (Ga0.8In0.2)xCo4Sb12(x = 0.05, 0.15, 0.25, 0.35, and 0.45) The sealed silica tubes were

heated slowly to 1323 K followed by quenching in a water bath and annealing at 873 K for one week The obtained ingots were roughly crushed by hand into powders and then subjected to spark plasma sintering in graphite dies under Ar flow atmosphere at a pressure of 50 MPa and temperature of 923 K for 15 min Bulk sample densities were calculated from the measured weights and dimensions The obtained samples were characterized by powder XRD using Cu Kα radiation at room temperature Lattice parameters were estimated via least-squares fitting to the indexed 2θ values (using Si as

an external standard) by employing PDXL, Rigaku’s integrated X-ray powder diffraction software Filling fractions were determined for the obtained samples from the Rietveld refinements performed

by utilizing the RietanFP program.46Samples’ microstructures were observed using field-emission scanning electron microscopy (FESEM), while their chemical compositions were determined by

EDX analysis under vacuum at room temperature The magnitudes of ρ and S were obtained using

a commercially available apparatus for simultaneous measurement of the Seebeck coefficient and electrical resistivity of thermoelectric materials (ULVAC, ZEM3) under He atmosphere The Hall

coefficient (RH) was measured at room temperature and applied magnetic field of 0.5 T using the van der Pauw method The Hall carrier concentration (nH) and Hall mobility (µH) were calculated from the obtained RHvalues based on the assumptions of a single-band model and a Hall factor of

1 (represented by the relations nH = 1/(e·RH) and µH = RH /ρ, where e was the elementary electric charge) The magnitudes of κ were evaluated from the thermal diffusivity (α), heat capacity (Cp), and sample density (d) values based on the relationship κ = αCp d Cpvalues were estimated via the

Dulong–Petit model, Cp = 3nR, where n was the number of atoms per formula unit, and R was the

gas constant The values of α were measured under Ar flow using a commercially available laser

FIG 1 Powder XRD patterns for the polycrystalline samples with the nominal compositions (Ga 0.8 In 0.2 )xCo 4 Sb 12(x = 0.05,

0.15, 0.25, 0.35, and 0.45).

III RESULTS AND DISCUSSION

The powder X-ray diffraction (XRD) patterns of the studied samples are shown in Figure 1 They indicate that all samples exhibit the skutterudite structure with negligible amounts of CoSb2

impurities (especially the sample with x = 0.45) The lattice parameters of the skutterudite phase and

sample densities are listed in TABLEI, which shows that the lattice parameter slightly increases with

increasing x in (Ga0.8In0.2) xCo4Sb12 Generally, the presence of Ga and In atoms inside the voids expands the skutterudite lattice,37,39while substituting the Sb sites with Ga atoms tends to contract

it.27 Thus, the increase in the lattice parameter observed in the present study can be caused by the increase in the In filling fraction (a detailed analysis of the lattice parameter will be presented later)

As a result, high-density bulk samples with densities exceeding 98% of the theoretical value were obtained

Figure2shows the temperature dependences of the Seebeck coefficient S, electrical resistivity

ρ, and power factor S2ρ−1 for the samples of (Ga0.8In0.2)xCo4Sb12(x = 0.05, 0.15, 0.25, 0.35, and

0.45) and In0.12Co4Sb12.47The In0.12Co4Sb12sample, as a reference for comparison, was taken with

consideration for In content close to the sample with x = 0.45.47The sample with nominal composition In0.12Co4Sb12, as a reference sample, was taken with consideration for In content of the sample with

x = 0.45.47The S values for all samples (except for x = 0.05) were negative across the entire temperature range between room temperature and 773 K For the samples with x = 0.150.35, the absolute values

of both S and ρ decreased with x, mainly due to increase in the carrier concentration caused by the

introduction of Ga and In species In particular, the co-adding of Ga and In effectively gave rise to

the enhancement of the absolute S compared to In single filling for CoSb3, which led to high S2ρ−1

Furthermore, the maximum values of the absolute S for these samples shifted to higher temperatures with increasing x, mainly due to the onset of a bipolar conduction caused by the increased electron

concentration.35On the other hand, when x reaches 0.45, the absolute values of S and ρ start increasing.

As a result, the samples with x = 0.35 and 0.45 exhibited similar characteristics in terms of S2ρ−1, corresponding to high values in the temperature range of 300700 K (above 4 mW m-1K-2) with a relatively flat temperature dependence (the origin of such a high power factor will be discussed later)

TABLE I Lattice parameter a, sample bulk density d, relative density %T.D measured for the polycrystalline samples with

a nominal composition of (Ga 0.8 In 0.2 )xCo 4 Sb 12(x = 0.05, 0.15, 0.25, 0.35, and 0.45) at room temperature.

125015-4 Choi et al. AIP Advances 6, 125015 (2016)

FIG 2 Temperature dependences of the (a) Seebeck coefficient S, (b) electrical resistivity ρ, and (c) power factor S2 ρ −1 for the samples with the nominal compositions (Ga 0.8 In 0.2 )xCo 4 Sb 12(x = 0.05, 0.15, 0.25, 0.35, and 0.45) and In0.12 Co 4 Sb 12 47

Since the samples with x = 0.35 and 0.45 produced the best values of S2ρ−1, the magnitudes of κ for these two samples were investigated

Figure 3 shows the temperature dependences of the κ, κlat, and zT for the samples of (Ga0.8In0.2)xCo4Sb12 (x = 0.35 and 0.45) and In0.12Co4Sb12.47 The value of κlat was obtained by subtracting κel(= LT ρ−1, where L was the Lorentz number = 2.45×10-8W Ω K-2) from the measured

κ, resulting in the formula κlat= κ−LT ρ−1, which indicated that κlatcontained a bipolar contribution

It was clearly confirmed that the κlatvalues for the samples with x = 0.35 and 0.45 were systematically

lower than those obtained at In single filled CoSb3across the entire temperature range, the co-adding

of Ga and In for CoSb3 effectively scatters heat carrying phonon Additionally, compared with

x = 0.35, the lower κlat values of x = 0.45 suggest that the increase in the added amount of Ga and In led to effective phonon scattering Owing to the combination of a high S2ρ−1magnitude and significantly reduced κlat, the sample with x = 0.45 exhibited the maximum of zT > 1.1 at 675 K, corre-sponding to the enhancement of the best value obtained in the previous study (equal to approximately 14%).27

To confirm the origin of high zT achieved at x = 0.45, Ga and In site occupancies in the skutterudite

phase, the existing states of Ga and In atoms added in the amounts over the filling limits, and sample’s microstructure were investigated in detail

Figure4 shows the observed and calculated powder XRD patterns and the difference profile

obtained for the samples with x = 0.35 and 0.45 The detailed results of the Rietveld refinement

are provided in TABLE II CoSb2 peaks corresponding to the secondary phase of the XRD pat-tern were excluded during the refinement The Rietveld refinement was performed in accordance with the procedure described in our previous study.27 The site occupancy refinement was con-ducted for each element to check the degree of disorder As shown in TABLEII, the x =0.35 and

0.45 samples had the skutterudite phase with an actual composition of Ga0.11In0.06Co4Sb11.94Ga0.06

FIG 4 Powder XRD patterns for the samples with (a) x = 0.35 and (b) x = 0.45 recorded at room temperature, which contain

the observed, calculated, difference, and background curves The expected peak positions are marked with vertical ticks.

and Ga0.05In0.09Co4Sb11.98Ga0.02, respectively For the x = 0.45 sample, by comparing this refined composition with the nominal composition Ga0.36In0.09Co4Sb12, it was concluded that all In atoms filled the voids, while the Ga species added over the solubility limit would precipitate as GaSb Based on the actual sample composition determined by the Rietveld analysis, the lattice parameter of the sample was estimated by using the Vegard rule.38 The lattice parameter a of

Ga0.05In0.09Co4Sb11.98Ga0.02can be written as:

a of Ga0.05InxCo4Sb11.98Ga0.02= a of Co4Sb12+ ∆a1 + ∆a2, (1)

where a of Co4Sb12is 0.9034 nm,40∆a1= a of Ga0.05Co4Sb11.98Ga0.02 a of Co4Sb12, and ∆a2= a of

InxCo4Sb12 a of Co4Sb12 The values of ∆a1and ∆a2can be calculated from the changes in lattice parameters of Ga and In single-filled single-filled CoSb3, respectively.21 , 33By fitting the calculated

lattice parameter to the experimental value, we obtained x ≈ 0.09, which was in good agreement with x = 0.09 of the actual composition determined by the Rietveld analysis This result implies that

the assumptions utilized for the Rietveld analysis are valid; in other words, in the case of Ga and In co-coped CoSb3, In easier occupies the void sites as compared to Ga, which occupies not only the void sites, but also the Sb sites at a ratio of 2:1

Figure5shows the scanning electron microscopy (SEM) images obtained for the x = 0.45 sample

(the image for the Ga0.2In0.15Co4Sb12sample27is shown for comparison) The sample with x = 0.45

was composed of dense grains with sizes over approximately 30 µm and nanoscale precipitates with

TABLE II Rietveld refinement results for the (Ga 0.8 In 0.2 )xCo 4 Sb 12samples with x = 0.35 and 0.45 Atomic positions: Ga/In, 2a (0, 0, 0); Co, 8c (0.25, 0.25, 0.25); Sb, 24g (0, y, z).

Nominal composition (Ga 0.8 In 0.2 ) 0.35 Co 4 Sb 12 (Ga 0.8 In 0.2 ) 0.45 Co 4 Sb 12

Actual composition Ga 0.17 In 0.06 Co 4 Sb 11.94 Ga 0.07 In 0.09 Co 4 Sb 11.98

125015-6 Choi et al. AIP Advances 6, 125015 (2016)

FIG 5 FESEM images of the polycrystalline samples Panels (a), (b), and (c) correspond to the sample with x = 0.45, and

panel (d) represents the sample with the nominal composition Ga 0.20 In 0.15 Co 4 Sb 12 21The square area depicted in Fig 5(d) has been described in the previous study.21

sizes of below 100 nm, which were formed inside the grains and at the boundaries In the present case, the precipitates correspond to the GaSb phase, because Ga and In added in the quantities exceeding the filling limit react with Sb and form GaSb and InSb species, respectively.15 , 41Furthermore, Figures5(a) and (d) reveal that the sample with x = 0.45 contains larger grains and smaller amounts of the

nanoscale precipitates as compared with those observed for the Ga0.2In0.15Co4Sb12 sample The nanoscale precipitates were widely dispersed in the present sample in contrast to the agglomerations formed inside the previously studied sample,27which could originate from the differences in nominal compositions and synthesis conditions It should be also noted that the Ga0.2In0.15Co4Sb12sample was prepared by hot-pressing from ball-milled fine powders,27while the x = 0.45 sample was obtained

by spark plasma sintering from roughly crushed powders

Figures6(a)and(b)show the dependences of the reduced effective mass m*/m0and Hall mobility µ

FIG 6 Relationships between (a) the reduced effective mass m*/m0 and Hall carrier concentration nH and (b) the Hall mobility µ H and Hall carrier concentration nH for Yb and Ga co-added CoSb 3 , 29 Yb single filled CoSb 3 , 17 , 29 three band model, 17 and the samples with nominal compositions of (Ga 0.8 In 0.2 )xCo 4 Sb 12 (x = 0.15, 0.25, 0.35,

and 0.45), (Ga VF ) 0.06 Co 4 Sb 11.97 (Ga Sb ) 0.03 , 34 (Ga VF ) 0.10 Co 4 Sb 11.95 (Ga Sb ) 0.05 , 34 (Ga VF ) 0.15 Co 4 Sb 11.925 (Ga Sb ) 0.075 , 34 and

Ga 0.2 InxCo 4 Sb 12(x = 0.15, 0.20, 0.25, and 0.30)27 obtained at room temperature.

we can express S and the true carrier concentration n as follows:48

s=k e

e

(

η −2F1(η)

F0(η)

)

n=

r 2 π

m ∗ k B T

~2

!3/2

F i(η)=

 ∞ 0

x i dx

where kB is the Boltzmann constant, F i is the Fermi integral of order i, η = EF /kBT is the reduced chemical potential, m* is the effective carrier mass, and ~ is the reduced Planck constant The measured Hall carrier concentration nH is connected to n via nH = n/rHwith the Hall scattering factor determined

by rH = (1.5F0.5 F-0.5)/2F0 As shown in Figure6(a), the m*/m0values gradually increased for all

samples with increasing nH, which could be explained by the band convergence at high carrier concentration region Further increasing in the m*/m0was observed for Ga and Yb co-filled CoSb3at

high nHvalues, which could be explained by the change in the band structure observed in the vicinity

of the Fermi level through the first-principles calculations.29Furthermore, m*/m0is much larger for

the x = 0.45 sample than for the Ga and Yb co-filled CoSb3samples, which is related to the combined effect of both the degeneracy of the band structure and the energy filtering caused by the presence of evenly dispersed GaSb nanoscale precipitates According to Ref.23, the nanoparticles of p-type GaSb

in the n-type host phase form an interfacial energy barrier, which impedes the diffusion of electrons with low energies, thus contributing to the enhancement of S On the other hand, Figure6(b)shows that the µHvalues for the samples with x ≥ 0.15 are systematically higher than those obtained for the

Ga0.20InxCo4Sb12(x = 0.15, 0.20, 0.25, 0.30) samples studied previously.27As shown in Figure5,

the x = 0.45 sample contains larger grains as compared to those for the previously studied one, which

could lead to higher µHmagnitudes

Figure 7(a) and (b) show the temperature dependences of κlat and zT, respectively, for

the samples with the nominal compositions Ga0.20Co4Sb12,39 (GaVF)0.10Co4Sb11.95(GaSb)0.05,34 Ga0.20In0.15Co4Sb12,27and (Ga0.8In0.2)xCo4Sb12(the x = 0.45 sample), while Figure7(c)displays the

maximum zT values for these samples As indicated by Figure7(a), the κlatvalue for the sample with

x = 0.45 is much lower than those for the other samples, despite its larger grain sizes The introduction

of both Ga and In into the CoSb3 lattice leads to the effective phonon scattering caused by the fol-lowing reasons: (1) rattling of both Ga and In atoms inside the voids, (2) the replacement of Sb with a small amount of Ga, (3) the formation of dispersed nanoscale precipitates, and (4) CoSb2compounds were formed to preserve the mass balance In the Ga single-filled system,34,39Ga species added over

125015-8 Choi et al. AIP Advances 6, 125015 (2016)

FIG 7 Temperature dependences of the (a) thermal conductivity κ and (b) dimensionless figure of merit zT for the samples

with nominal compositions of Ga 0.20 Co 4 Sb 12 ,38(Ga VF ) 0.10 Co 4 Sb 11.95 (Ga Sb ) 0.05 ,34and Ga 0.20 In 0.30 Co 4 Sb 1227as well as for

the sample with x = 0.45 (c) Maximum zT values and schematic microstructural views for these samples.

the filling limit normally exist as microscale Ga metal precipitates, which maintain low values of

zT.39However, the optimization of the chemical composition in terms of the void filling and charge compensation results in a large decrease in the κlat magnitude for (GaVF)0.10Co4Sb11.95(GaSb)0.05,

which significantly enhances zT.34The obtained results suggest that the dual-site occupancy of Ga

is an effective way for enhancing the zT of Ga single-filled CoSb3 In addition, further decreases in

κlatand the related zT enhancement are achieved for Ga0.20In0.15Co4Sb12by introducing In and Ga atoms into the CoSb3lattice.27In the present study, the x = 0.45 sample exhibited lower κlatthan that

of Ga0.20In0.15Co4Sb12, mainly due to the presence of large Ga contents not only inside the voids, but

also at the Sb sites of the skutterudite structure The actual compositions of the x = 0.45 sample and the

Ga0.20In0.15Co4Sb12 sample were Ga0.05In0.09Co4Sb11.98Ga0.02and Ga0.02In0.11Co4Sb11.99Ga0.01,27

respectively Further zT enhancement was achieved for the x = 0.45 sample, which was

approxi-mately 14% higher than that obtained for the Ga0.20In0.30Co4Sb12sample due to the optimization of both the chemical composition and microstructure The studied sample contained moderately large grains with small amounts of nanoscale precipitates, which maintained high carrier mobility while preserving low values of κlat

IV SUMMARY

In this work, polycrystalline samples with nominal compositions of (Ga0.8In0.2)xCo4Sb12(x =

0.05, 0.15, 0.25, 0.35, and 0.45) were synthesized, and their TE properties were investigated The obtained TE data were compared with those for other skutterudites filled with Ga and/or In atoms

The sample with x = 0.45 was mainly composed of two phases: a skutterudite phase with a

compo-sition of Ga0.05In0.09Co4Sb11.98Ga0.02(in which Ga atoms occupied both the voids and Sb sites) and

GaSb nanoscale precipitates The x = 0.45 sample contained larger grains with smaller amounts of

nanoscale precipitates as compared with other previously investigated skutterudite samples A sig-nificant increase in the effective mass was observed, which resulted from the combined effect of both the change in electronic band structure and scattering of low-energy electrons due to the formation of GaSb nanoscale precipitates Owing to the described electronic and structural features, very high mag-nitudes of the power factor and significantly reduced κlatvalues were achieved simultaneously, leading

to zT values greater than 1.1 (at a temperature of 675 K), which corresponded to the enhancement of

the previously obtained data for Ga and In co-added skutterudites by approximately 14%

1 T M Tritt and M A Subramanian, MRS Bull.31, 188 (2006).

2 G J Snyder, M Christensen, E Nishibori, T Caillat, and B B Iversen, Nat Mater.3, 458–463 (2004).

3 K Biswas, J He, I D Blum, C.-I Wu, T P Hogan, D N Seidman, V P Dravid, and M G Kanatzidis, Nature489, 414–418

5 Y Gelbstein, J Tunbridge, R Dixon, M J Reece, H Ning, R Gilchrist, R Summers, I Agote, M A Lagos, K Simpson,

C Rouaud, P Feulner, S Rivera, R Torrecillas, M Husband, J Crossley, and I Robinson, J Electron Mater.43, 1703 (2014).

6 O Appel, T Zilber, S Kalabukhov, O Beeri, and Y Gelbstein, J Mater Chem C3, 11653 (2015).

7 S A Yamini, D R G Mitchell, Z M Gibbs, R Santos, V Patterson, S Li, Y Z Pei, S X Dou, and G Jeffrey Snyder,

Adv Energy Mater.5, 1501047 (2015).

8 S I Kim, K H Lee, H A Mun, H S Kim, S W Hwang, J W Roh, D J Yang, W H Shin, X S Li, Y H Lee, G J Snyder, and S W Kim, Science348, 109 (2015).

9 D Fuks, G Komisarchik, M Kaller, and Y Gelbstein, J Solid State Chem.240, 91 (2016).

10 M K Jana, K Pal, U V Waghmare, and K Biswas, Angew Chem Int Ed.55, 7792 (2016).

11 K Kurosaki, G Li, Y Ohishi, H Muta, and S Yamanaka, Front Chem.2, 84 (2014).

12 Y Z Pei, J Yang, L D Chen, W Zhang, and J R Salvador, Appl Phys Lett.95, 042101 (2009).

13 L D Chen, X F Tang, T Goto, and T Hirai, J Mater Res.15, 2276–2279 (2000).

14 D R Thompson, C Liu, J Yang, J R Salvador, D B Haddad, N D Ellison, R A Waldo, and J Yang, Acta Mater.92, 152–162 (2015).

15 X Su, H Li, Y Yan, H Chi, X Tang, Q Zhang, and C Uher, J Mater Chem.22, 15628–15634 (2012).

16 Y Tang, R Hanus, S W Chen, and G J Snyder, Nat Commun.6, 7584 (2015).

17 Y Tang, Z M Gibbs, L A Agapito, G Li, H.-S Kim, M B Nardelli, S Curtarolo, and G J Snyder, Nat Mater.14, 1223 (2015).

18 S Wang, J Yang, L Wu, P Wei, W Zhang, and J Yang, Adv Funct Mater.25, 6660 (2015).

19 R Liu, J Y Cho, J Yang, W Zhang, and L Chen, J Mater Sci Technol.30, 1134–1140 (2014).

20 G Rogl, A Grytsiv, P Rogl, E Bauer, M Hochenhofer, R Anbalagan, R C Mallik, and E Schafler, Acta Mater.76, 434–448 (2014).

21 J Leszczynski, V D Ros, B Lenoir, A Dauscher, C Candolfi, P Masschelein, J Hejtmanek, K Kutorasinski, J Tobola,

R I Smith, C Stiewe, and E M¨uller, J Phys D: Appl Phys.46, 495106 (2013).

22 G Rogl, A Grytsiv, M Falmbigl, E Bauer, P Rogl, M Zehetbauer, and Y Gelbstein, J Alloys Compd.537, 242–249 (2012).

23 Z Xiong, X Chen, X Huang, S Bai, and L Chen, Acta Mater.58, 3995–4002 (2010).

24 S Wang, J R Salvador, J Yang, P Wei, B Duan, and J Yang, NPG Asia Mater.8, e285 (2016).

25 B Duan, J Yang, J R Salvador, Y He, B Zhao, S Wang, P Wei, F S Ohuchi, W Zhang, R P Hermann, O Gourdon,

S X Mao, Y Cheng, C Wang, J Liu, P Zhai, X Tang, Q Zhang, and J Yang, Energy Environ Sci.9, 2090 (2016).

26 B R Ortiz, C M Crawford, R W McKinney, P A Parilla, and E S Toberer, J Mater Chem A4, 8444 (2016).

27 S Choi, K Kurosaki, A Harnwunggmoung, Y Miyazaki, Y Ohishi, H Muta, and S Yamanaka, Jpn J Appl Phys.54,

111801 (2015).

28 L Xi, Y Qiu, S Zheng, X Shi, J Yang, L Chen, D J Singh, J Yang, and W Zhang, Acta Mater.85, 112–121 (2015).

29 X Shi, J Yang, L Wu, J R Salvador, C Zhang, W L Villaire, D Haddad, J Yang, Y Zhu, and Q Li, Sci Rep.5, 14641 (2015).

30 W Zhao, P Wei, Q Zhang, H Peng, W Zhu, D Tang, J Yu, H Zhou, Z Liu, X Mu, D He, J Li, C Wang, X Tang, and

J Yang, Nat Commun.6, 6197 (2015).

31 A Sesselmann, B Klobes, T Dasgupta, O Gourdon, R Hermann, and E Mueller, Phys Status Solidi A213, 766 (2015).

32 Y Tang, Y Qiu, L Xi, X Shi, W Zhang, L Chen, S M Tseng, S W Chen, and G J Snyder, Energy Environ Sci.7, 812–819 (2014).

33 Y Qiu, J Xing, X Gao, L Xi, X Shi, H Gu, and L Chen, J Mater Chem A2, 10952–10959 (2014).

34 Y Qiu, L Xi, X Shi, P Qiu, W Zhang, L Chen, J R Salvador, J Y Cho, J Yang, Y C Chien, S W Chen, Y Tang, and

G J Snyder, Adv Funct Mater.23, 3194–3203 (2013).

35 D Kim, K Kurosaki, Y Ohishi, H Muta, and S Yamanaka, APL Mater.1, 032115 (2013).

36 A Grytsiv, P Rogl, H Michor, E Bauer, and G Giester, J Electron Mater.42, 2940–2952 (2013).

37 G Li, K Kurosaki, Y Ohishi, H Muta, and S Yamanaka, J Electron Mater.42, 1463 (2013).

38 A Harnwunggmoung, K Kurosaki, A Kosuga, M Ishimaru, T Plirdpring, R Yimnirun, J Jutimoosik, S Rujirawat,

Y Ohishi, H Muta, and S Yamanaka, J Appl Phys.112, 043509 (2012).

39 A Harnwunggmoung, K Kurosaki, T Plirdpring, T Sugahara, Y Ohishi, H Muta, and S Yamanaka, J Appl Phys.110,

013521 (2011).

40 A Harnwunggmoung, K Kurosaki, H Muta, and S Yamanaka, Appl Phys Lett.96, 202107 (2010).

41 H Li, X Tang, Q Zhang, and C Uher, Appl Phys Lett.94, 102114 (2009).

42 J Peng, J He, Z Su, P N Alboni, S Zhu, and T M Tritt, J Appl Phys.105, 084907 (2009).

43 W Zhao, P Wei, Q Zhang, C Dong, L Liu, and X Tang, J Am Chem Soc.131, 3713 (2009).

44 J Yu, W.-Y Zhao, P Wei, W.-T Zhu, H.-Y Zhou, Z.-Y Liu, D.-G Tang, B Lei, and Q.-J Zhang, Appl Phys Lett.104,

142104 (2014).

45 L Xi, Y Qiu, X Shi, W Zhang, L Chen, D J Singh, and J Yang, Chem Commun.51, 10823 (2015).

46 F Izumi and K Momma, Solid State Phenom.130, 15 (2007).

47 E Visnow, C P Heinrich, A Schmitz, J de Boor, P Leidich, B Klobes, R P Hermann, W E M¨uller, and W Tremel,

Inorg Chem.54, 7818 (2015).

48 X Liu, T Zhu, H Wang, L Hu, H Xie, G Jiang, G J Snyder, and X Zhao, Adv Energy Mater.3, 1238–1244 (2013).