Full text

Turn on search term navigation

ARTICLE

Received 4 Nov 2013 | Accepted 13 Mar 2014 | Published 2 May 2014

Yeseul Lee1, Shih-Han Lo2,*, Changqiang Chen2,*, Hui Sun3, Duck-Young Chung4, Thomas C. Chasapis1, Ctirad Uher3, Vinayak P. Dravid2 & Mercouri G. Kanatzidis1,4

Increasing the conversion efciency of thermoelectric materials is a key scientic driver behind a worldwide effort to enable heat to electricity power generation at competitive cost. Here we report an increased performance for antimony-doped lead selenide with a thermoelectric gure of merit of B1.5 at 800 K. This is in sharp contrast to bismuth doped lead selenide, which reaches a gure of merit of o1. Substituting antimony or bismuth for lead achieves maximum power factors between B2327 mWcm 1 K 2 at temperatures above 400 K. The addition of small amounts (B0.25 mol%) of antimony generates extensive nanoscale precipitates, whereas comparable amounts of bismuth results in very few or no precipitates. The antimony-rich precipitates are endotaxial in lead selenide, and appear remarkably effective in reducing the lattice thermal conductivity. The corresponding bismuth-containing samples exhibit smaller reduction in lattice thermal conductivity.

DOI: 10.1038/ncomms4640

Contrasting role of antimony and bismuth dopants on the thermoelectric performance of lead selenide

1 Department of Chemistry, Northwestern University, Evanston, Illinois 60208, USA. 2 Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, USA. 3 Department of Physics, University of Michigan, Ann Arbor, Michigan 48109, USA. 4 Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439, USA. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to V.P.D. (email: mailto:[email protected]

Web End [email protected] ) or to M.G.K. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640

Advances in high-performance thermoelectric materials can potentially impact electrical power generation through waste-heat recovery13. Such thermoelectric materials

demand high value of the dimensionless gure of merit (ZT) dened as ZT (S2s/k)T (T is the temperature, S is the Seebeck

coefcient, s is the electrical conductivity, and k is the total thermal conductivity (comprising kel and klat; representing contributions from charge carriers and lattice, respectively) for practical applications. Combining ZT and Carnot efciency determines the efciency of thermoelectric energy conversion. Thus, increasing such conversion efciency requires both high ZT and a large temperature gradient across the materials. One effective way to achieve high ZT is to reduce lattice thermal conductivity through embedding micro/nanostructures, which provide multiple scattering mechanisms for phonons1,2,418. Other strategies address the challenge of maximizing a power factor from existing semiconducting materials by band alignment5,6,19,20, modifying their electronic structures close to the Fermi level21,22, and increasing band degeneracy by convergence of bands23.

The rocksalt-structured lead chalcogenide semiconductors have many compelling attributes as promising thermoelectric materials. As one of the most investigated materials in this system, lead telluride (PbTe) has been extensively studied. PbTe remains the top-performing thermoelectric at mid-temperature range (600900 K). However, concerns regarding the scarcity of tellurium (Te) in the Earths crust24 are motivating the development of materials based on more abundant elements. Lead selenide (PbSe), a much lower cost analogue of PbTe, is an attractive alternative as selenium (Se) has a better long-term price stability and lower cost than Te; and is B50 times more abundant than Te24. Moreover, PbSe has much higher melting temperature (1,357 K) than PbTe (1,198 K) that may enable a higher efciency for the same value of ZT at higher operating temperatures.

Recently, high ZT values have been reported for n-type PbSe (ZT B0.9 for PbSe-In/Ga (ingot)25 and B1.2 (hot-pressed

samples)26 at 900 K, B0.8 at 700 K for PbSe-Cl27, B1.2 at 850 K for PbSe-Br28, B1.3 at 850 K for PbSe-Al29 and B1.3 at 900 K for PbSe/PbS-Cl9). Advances also have been reported for p-type PbSe (ZT B1.2 at 850 K for PbSe-Na30, B1.0 at 770 K for

PbSe-Ag31, B1.0 at 700 K for PbSe-Tl26, B1.3 at 900 K for PbSe/(Sr/Ba)Se-Na4 and B1.6 at 900 K for PbSe/CdS-Na5).

Theoretical calculations have suggested that heavily doped p-type PbSe can potentially show a ZT B2 at 1,000 K32 and n-type PbSe can be a high-performance thermoelectric as well because of large thermopower33, which offers added motivation for investigations in this system.

Here, we report contrasting high-temperature thermoelectric properties of PbSe doped with antimony (Sb) and bismuth (Bi). Sb and Bi act as n-type dopants and enable tuning of the carrier concentration by substituting nominal Sb3 /Bi3 for Pb2 in the rock-salt sublattice. We demonstrate that Sb not only has merely the simple role as a dopant for carrier generation but that, unexpectedly, it is also remarkably effective in generating a unique form of nanostructuring that dramatically reduces the lattice thermal conductivity of PbSe. The surprising fact is that even very small amounts (r0.25 mol%) of Sb in PbSe generate a highly dense array of well dispersed nanostructures. In sharp contrast to Sb, Bi dopant generates only few nanoscale precipitates, if any, and does not notably affect the lattice thermal conductivity of PbSe. We have employed the pulsed electric current sintering technique, also known as spark plasma sintering, to obtain highly dense and robust thermoelectric specimens starting from mesoscale powders. We report high ZT values that peak at different temperatures based on Sb concentration. We observe a ZT of B1.2 at 700 K for

Pb0.99875Sb0.00125Se, B1.5 at 830 K for Pb0.9975Sb0.0025Se, B1.4

at 930 K for Pb0.995Sb0.005Se and B1.2 at 930 K for samples of

Pb1-xSbxSe

6.13

6.125

Intensity (a.u.)

x = 2% 1%0.75%

a()

6.12

6.115

6.11

0.5%0.25%0.125%

30 40 50

Vegard's law

0.075%

Pb1-xSbxSe

Pb1-xBixSe Theoretical

60 70 80 2 (degree)

1 2 3 4 5

% Of Sb

0 0 0.5 1 1.5 2 Sb/Bi concentration (mol %)

Pb1-xSbxSe

Carrier conc. (1019 cm3 )

Figure 1 | PXRD and Hall charge-carrier concentration of Pb1 xSbxSe and Pb1 xBixSe. (a) PXRD patterns for samples of Pb1 xSbxSe. (b) lattice

parameters as a function of Sb concentration. (c) Typical spark plasma sintering pellet and samples used in this study. Red arrows indicate the pressing direction. Results of parallel to the pressing direction measurements are shown in the text and perpendicular direction measurements are in the supporting information. Scale bar, 1 cm. (d) Hall-effect charge-carrier concentration in Pb1 xSbxSe and Pb1 xBixSe samples. The solid line is the predicted carrier

concentration.

2 NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640 ARTICLE

Pb0.9925Sb0.0075Se and Pb0.99Sb0.01Se; which represent highest

reported ZT values for such systems. In contrast, all respective Bi-containing PbSe samples exhibit ZTo1. As explained below, the contrasting behaviour of Sb- and Bi-doped PbSe lies in the remarkably different microstructure evolution, primarily in the nanoscale.

ResultsStructural characterization. Figure 1a presents the Powder X-ray diffraction (PXRD) patterns of Pb1 xSbxSe samples. All

patterns were indexed in the NaCl structure (Fm-3m space group) without any noticeable second phase (within the detection limit of our X-ray diffraction equipment) for Sb contents of up to 2 mol%. Figure 1b shows the extracted lattice parameters as a function of Sb content. The lattice parameter of PbSe decreases with increasing Sb content. This is Vegards law behaviour, and it is consistent with size trends (Sb is smaller than Pb) and indicates some solubility of the Sb ions in the lattice. As we will show, however, the key feature of this system is the extensive nanostructuring created in the PbSe matrix that is detectable with only with transmission electron microscopy (TEM). Figure 1c shows a typical spark plasma sintering disk-shaped pallet cut and polished for measurements in this study. Both directions (parallel and perpendicular to pressing direction) were measured and showed no directional dependence. Only results of parallel direction measurements are shown in the text and perpendicular direction measurements are in the Supplementary Fig. 1.

Charge-carrier concentration and electron mobility. In this study, Sb and Bi were used as dopants and from simple valence counting each atom is expected to add one extra n-type carrier to PbSe. In Fig. 1d, the room temperature Hall carrier-concentrations are plotted as a function of the added amount of dopant. Table 1 also shows a summary of the physical properties of the samples. Here, the solid line represents one electron per Sb/Bi atom added. For both Sb and Bi samples, the experimental values fall well below the theoretical ones. In particular, for Sb samples from 0.75 mol%, the values of experimental Hall charge-carrier concentration almost reach saturation, while Bi samples exhibit almost linearly increasing carrier concentration. This suggests that Bi is signicantly more soluble in PbSe than Sb and the Hall measurements indicate that the carrier saturation of Sb is reached at B0.75 mol%

Sb, and that of Bi is not reached even up to 2 mol% Bi. The lower than expected carrier concentration values of Sb-containing samples are attributed to nanostructuring as explained below.

The electrical conductivity (s) of Pb1 xSbxSe samples is

depicted in Fig. 2a (corresponding results of Pb1 xBixSe samples

are shown in Fig. 3a). For all samples, a monotonic decrease in s with increasing temperature is observed, consistent with heavily doped semiconductors with metallic transport behaviour. For Sb-containing samples, the electrical conductivity increases with increase in Hall carrier concentration at any given temperature. However, the electrical conductivity at lower temperature decreases for 1.5 mol% and 2 mol% of Bi-containing samples despite the higher Hall carrier concentrations. This can be explained by the possibility of increasing impurity scattering from the higher doping (See Supplementary Note 1).

To understand the behaviour of s, we performed temperature-dependent Hall-effect studies. The Hall coefcient, RH, for the Sb-containing samples for x 0.25, 0.5, and 0.75% is plotted

in Fig. 2b which shows it is almost temperature independent. Hall-effect measurements can determine the carrier densities under the assumption of a unitary of single band transport and a Hall coefcient RH 1/nHe, where nH is the Hall charge-carrier

density and e is the electron charge. The temperature-independent Hall coefcient shows that the specimens retain the same carrier density up to 850 K. Figure 2c presents the Hall mobility, mH

RHs, as a function of temperature. The room temperature mH values decrease fromB550 cm2V 1 s 1 for 0.25% of Sb toB310 cm2 V 1 s 1 for 0.75% of Sb, because of enhanced point defect scattering with increasing doping level. At high temperature, a rapid decrease is observed and mH B70 cm2

V 1 s 1 at 850 K for all three samples.

The rapid decrease of mH is shown more clearly on a log(mH)-log(T) plot in Fig. 2d. In the temperature range from 300 K to 475 K, the slopes of the curves are B 1.2, but for T4500 K the

slopes increase to B 2.4. Table 2 summarizes the tting results.

This type of mobility exponent is characteristic of scattering mechanisms that are related to the band structure parameters. Assuming parabolic bands, electronphonon scattering due to thermal vibrations of the lattice cause mH to scale as mHBT 1.5 (ref. 34). The low-temperature exponent in Fig. 2d can be attributed to acoustic phonon scattering of electrons in agreement with previous results on highly doped PbSe samples35. In both the work of Gobrecht and Schlichting36 and our data, the exponent

Table 1 | Thermoelectric properties of Pb1 x(Sb/Bi)xSe samples.

Sample composition nH (1019 cm 3) lH (cm2 V 1 s 1) S (lV K 1) r (S cm 1) jlat (Wm 1 K 1) at 300 K

jlat (Wm 1 K 1) at 700 K

Maximum ZT (Temperature (K))

Pb1Sb0Se 0.32 878 213 455 1.81 1.18 0.49 (500)

Pb0.99925Sb0.00075Se 1.1 704 143 1,184 1.81 1.06 0.82 (600)

Pb0.99875Sb0.00125Se 1.4 664 118 1,500 1.56 0.82 1.15 (700)

Pb0.9975Sb0.0025Se 2.6 523 87 2,341 1.56 0.68 1.45 (830

Pb0.995Sb0.005Se 5.2 372 55 3,329 1.54 0.66 1.38 (930)

Pb0.9925Sb0.0075Se 7 307 41 3,542 1.51 0.6 1.21 (930)

Pb0.99Sb0.01Se 7.9 286 44 3,607 1.51 0.55 1.15 (930)

Pb0.98Sb0.02Se 8.4 260 37 3,476 1.42 0.56 1.06 (930)

Pb1Bi0Se 0.3 833 222 405 1.79 1.15 0.46 (550)

Pb0.9975Bi0.0025Se 2.4 582 91 2,272 1.81 1.17 0.83 (690)

Pb0.995Bi0.005Se 6.1 348 43 3,379 1.77 1.1 0.89 (920)

Pb0.9925Bi0.0075Se 8.8 231 36 3,266 1.76 1.13 0.45 (720)*

Pb0.99Bi0.01Se 10 209 31 3,377 1.69 1.1 0.73 (920)

Pb0.985Bi0.015Se 17 90 29 2,445 1.58 1.17 0.63 (910)

Pb0.98Bi0.02Se 21 60 26 2,018 1.59 1.16 0.32 (720)*

k , lattice thermal conductivity; n , room temperature carrier concentration; m , hole mobility; s, electrical conductivity; S, Seebeck coefcient, and ZT of Pb (Sb/Bi) Se samples. *These samples measured only up to 720 K so their actual maximum ZT and temperature may be higher than values listed here.

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640

Pb1-xSbxSe0%0.075%0.125%0.25%0.5%0.75%

4,000 0

x = 0.25% x = 0.5%

x = 0.75%

Pb1-xSbxSe

[afii9846] (S cm1)

3,000

2,000

1,000

R H(C cm3 )

0.1

0.2

200 400

Temperature (K)

0200

400

Temperature (K)

600 800

600

x = 0.25% x = 0.5% x = 0.75%

H( cm2 V

500

400

300

200

100

200 400 600 800

Log( H( cm2 V1 s 1 ))

Temperature (K) Log(T (K))

32.82.62.42.2 2

1.81.6

100

200

300

0 200 400 600 800

1 s 1 )

H( cm2 V1 s 1 )

x = 0.25% x = 0.5% x = 0.75%

Slope = 1.5 Slope = 2.3

2.4 2.5 2.6 2.7 2.8 2.9

1,000 Theoretical Pb1-xSbxSe Pb1-xBixSe

B(Ren) Ga(Ren)

In(Ren)

800

600

400

200

250

200

150

100

00 5 10 15 20

S(V K1)

5 10

15 20

200 400

Temperature (K)

600 800

nH (1019 cm3)

S(V K1)

Pisarenko Pb1-xSbxSe

Pb1-xBixSe B(Ren)

Ga(Ren) In(Ren)

PF(W cm 1 K 2 )

Figure 2 | Electrical transport properties of Pb1 xSbxSe. (a)Temperature dependence of electrical conductivity of Pb1 xSbxSe. (b) Hall coefcient as a

function of temperature for x 0.25, 0.5 and 0.75% containing Pb1

xSbxSe samples. Note that the Hall coefcient is almost temperature independent. (c) Hall mobility as a function of temperature for x 0.25, 0.5, and 0.75% containing Pb1

xSbxSe samples. (d) log(mH)log(T) plot showing two regions of linearity, one below 475 K and the other above 500 K. The high-temperature region is characterized by a much steeper slope reecting a strong electronic-scattering mechanism at play. (e) Room temperature Hall mobility for Pb1 xSbxSe and Pb1 xBixSe samples as a function of the experimental Hall

charge-carrier concentration. (f) Temperature-dependent Seebeck coefcient of Pb1 xSbxSe samples. (g) Room temperature Seebeck coefcient for

Pb1 xSbxSe and Pb1 xBixSe samples as a function of the experimental Hall charge-carrier concentration. (h)Temperature-dependent thermoelectric power

factor of Pb1 xSbxSe samples. Data for PbSe doped with B, Ga, In by Zhang et al.26 Black dashed line is due to the model described in the text.

changes and approaches 1 as concentrations increase, again in

good agreement with theory and experimental data for the effective mass. At high-temperature, strong electronic-scattering mechanisms, non-parabolicity of the bands and/or scattering from high-frequency vibrations of the lattice (optical phonons) possibly have a role to rapidly decrease the values of exponents at high temperature37,38.

The experimental transport data were analysed with solutions to the Boltzmann transport equations within the relaxation time approximation. It is assumed that electron conduction occurs within a single parabolic band considering multiple ellipsoidal valleys with degeneracy Nc 4, where in the same valley the

effective masses along different directions of the ellipsoids are different. The different effective masses give rise to band

4 NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640 ARTICLE

4,000 Pb1-xBixSe 0%0.25%0.5%0.75%

1.5%

3,000

(S cm1)

100

S(V K1)

2,000

200

1,000

300

0200 400 600Temperature (K)

800

400

200 600Temperature (K)

800

4.5 4

Pb1-xBixSe

PF(W cm1 K 2 )

3.5

K tot(W m1 K 1 )

0%0.25%0.5%0.75%

1.5%

2.5

1.5 1

0200 400 600Temperature (K)

800

200 400 600Temperature (K)

800

K lat(W m1 K 1 )

Pb1-xBixSe Pb1-xBixSe

0%0.25%0.5%0.75%

1.5%

1.8

1.6

0%0.25%0.5%0.75%

1.5%

0.5

1.4

1.2

0.8200 400 600 Temperature (K)

800

200 400 600

Temperature (K)

800

Figure 3 | Thermoelectric properties of Pb1 xBixSe. (a) Temperature-dependent electrical conductivity, (b) Seebeck coefcient, (c) thermoelectric power

factor, (d) total thermal conductivity, (e) lattice thermal conductivity and (f) thermoelectric gure of merit of Pb1 xBixSe samples.

Table 2 | Characteristic slopes of the log(lH)-log(T) for Pb0.9975Sb0.0025Se, Pb0.995Sb0.005Se and Pb0.9925Sb0.0075Se

samples for two different temperature regimes.

300 KoTo475 K 500 KoT0.25% 0.50% 0.75% 0.25% 0.50% 0.75%

1.5 1.2 1 2.6 2.3 2.4

Here, m c is the conductivity-effective mass dened as m c31=m l 2=m t 1. The Fermi integrals Fn(Z) are the Fermi

integrals dened by May et al.43 and evaluated for a specic reduced chemical potential, Z EF/kT. The energy-independent

term of the relaxation time in the case of acoustical phonon scattering is dened28

F 1=2Z

F0Z

anisotropy, which is dened as K ml*/mt* where ml* is the

effective mass associated with the longitudinal axes of the ellipsoidal energy surfaces, and mt* associated with two of the transverse axis of the surfaces. In the case of PbSe, the band anisotropy factor is K 1.7539,40.

The equations shown below are valid for a single-scattering mechanism where the energy dependence of the carrier relaxation time can be expressed by a simple power law. Thus, the relaxation time, t, has a simple power law dependence on the charge-carrier energy, e, so t(T,e) t0(T)el 1/2, here t0 is the energy-

independent term of the relaxation time, and l is the scattering parameter. For acoustic phonon scattering l 0, and it will be

assumed that this is the dominant scattering mechanism. Note that some representations may differ by the factor of 1/2

(ref. 41). Within the approximations outlined above, the Hall mobility can be expressed as42:

et0 2m c

p 4Cl

p E2defm bkT3=2

Here, Cl 9.1 1010 Pa is a combination of elastic con

stants44,45, and Edef is the deformation potential, which describes the carrier-scattering strength by acoustic phonons. Ravich et al.39 listed a room temperature deformation potential of 20 eV, while Wang et al.28 gave a value of 25 eV for n-type PbSe. The effective mass in the denominator of equation (2) dened as m bm lm 2t1=3 is the mass of a single valley, which is related to

the total density of states mass m d via40

m dN2=3cm b 3

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640

For a specic Z, the Hall mobility as calculated by equation (1) can be related to the Hall charge-carrier concentration through:

nH4p

2m dkT

3=2F

1=2

rH 4

where the Hall factor, rH, is dened by43

32 F1=2Z

F 1=2Z

2F20Z

In Fig. 2e, the room temperature Hall mobility values of n-type PbSe samples as well as data reported by Ren and coworkers26 are plotted as a function of Hall charge-carrier concentration. From equation (2), t0 8.9 10 14 s was obtained and a temperature

constant Edef 29 eV, close to the literature values mentioned

above39, and a density of states effective mass of m* 0.36me,

where me is the free electron mass were used, in agreement with the effective mass obtained from the carrier dependence of the Seebeck coefcient discussed in the succeeding section. As seen from Fig. 2e, the theoretical line describes well the experimental data up to carrier densities B3 1019 cm 3, but highly Sb/Bi-

containing samples fall below the values expected by the model. We have to notice that the same tting quality was found assuming single spherical surfaces43; however, the obtained effective mass was found underestimated relative to the one obtained from the Pisarenko plot discussed below.

The disagreement between the calculated and the experimental room temperature mobility of Fig. 2e may be attributed to nonparabolicity of bands, microscopic cracks, grain boundaries and defects leading to a higher residual resistance, and consequently, a lower Hall mobility. As it is shown in the Supplementary Fig. 2 and Supplementary Note 2, the conduction band nonparabolicity was found to be the major deterrent effect for the description of the carrier dependence of Hall mobility within the framework of a single parabolic model. However, the Seebeck coefcient was found to be well described by the simple parabolic band model in the whole composition range and this is the model we have used to calculate the reduced chemical potentials and the Lorenz numbers.

Seebeck coefcient and power factor. Figure 2f shows the Seebeck coefcients (S) of Pb1 xSbxSe samples as functions of

temperature (corresponding results of Pb1 xBixSe samples are

shown in Fig. 3b). As expected, the S values are negative over the entire temperature range, indicating n-type conduction. Most samples show almost degenerate linear behaviour. The magnitudes (absolute values) of the S of the low Sb-containing samples display maxima at high temperature, where thermally activated holes reduce the thermoelectric voltage.

In Fig. 2g, the absolute Seebeck coefcients for n-type PbSe samples are plotted as a function of the Hall charge-carrier concentration (Pisarenko plot) at room temperature. Data from this study and the literature26 are compared with a theoretical calculation. The dashed black curve is calculated using a single parabolic model. In this model, dominated by acoustic phonon scattering outlined above, the Seebeck coefcient is given by42:

k e

tot(W m1 K1 )

4.5 Pb1-xSbxSe 0%

0.075%

0.125%

0.25%

0.5%

0.75%

3.5

2.5

K 1.5

0.5200 400 600

Temperature (K)

800

K lat(W m1 K1 )

1.5

0.5

200

400 600

Temperature (K)

800

Pb1-xBixSe, 300 K Pb1-xSbxSe, 300 K

Pb1-xBixSe, 700 K Pb1-xSbxSe, 700 K

1.5

0.5

The S can be related to nH through equation (4) provided the effective mass, m*, is known. The dashed line shown in Fig. 2g is generated by assuming a same effective mass of m* 0.36me that

is used in the calculation of theoretical mobility and agrees well with the experimental data for all n-type PbSe.

Figure 2h presents power factors derived from the electrical conductivities and Seebeck coefcients as a function of temperature for Pb1 xSbxSe samples (corresponding results of

Pb1 xBixSe samples are in Fig. 3c). As expected, the maximum

for the power factor is pushed to higher temperatures as the amount of dopant is increased. Whereas the x 0% sample have

its maxima at B300 K, the x 0.075 and 0.125% samples have

their maximum at B400 K, the x 0.25% sample at B500 K, the

x 0.5% sample at B650 K, and the xZ0.75% at B800 K and all

the maximum values are large enough (Z19 mW cm 1 K 2) for good thermoelectric properties. This shows that varying the dopant concentration controls the temperature of optimum thermoelectric performance.

0 0 0.5 1 1.5 2

X (%)

2F1Z

F0Z

Figure 4 | Thermal transport properties of Pb1 xSbxSe. (a) Temperature

dependence of total thermal conductivity and (b) lattice thermal conductivity of Pb1 xSbxSe samples. (c) Lattice thermal conductivity of

Pb1 xSbxSe samples (lled) and Pb1 xBixSe samples (empty) shown as a

function of x at 300 K (black) and 700 K (red). Solid and dashed lines are

guide to the eye.

6 NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640 ARTICLE

Thermal conductivity. The temperature dependence of total thermal conductivity is shown in Fig. 4a (corresponding results of Pb1 xBixSe samples are shown in Fig. 3d). The total thermal

conductivity (ktot), which consists of contributions from the charge carriers (kel) and the lattice phonons (klat), decreases with decreasing Hall charge-carrier concentration and increasing temperature. The charge-carrier thermal conductivity is directly proportional to the electrical conductivity that derives from the WiedmannFranz relation kel LsT, where L is the Lorenz

number. In the assumption of a homogeneous material with a parabolic band dominated by acoustic phonon scattering, the Lorenz number is given as42,46:

k e

23F

0 ZF2Z 4F1Z2

F0Z2

The reduced chemical potential was obtained from ts of the experimental Seebeck coefcient. Using these chemical potential values, L and kel for each sample can be calculated and are shown in Supplementary Figs 3 and 4.

The lattice thermal conductivities (klat) for Pb1 xSbxSe samples are shown in Fig. 4b (corresponding results of

Pb1 xBixSe samples are shown in Fig. 3e). The lattice thermal

conductivity decreases with increasing temperature for xZ0.5% of Sb samples but lower doping samples show upturn at high temperature (more details in the Supplementary Fig. 3). The x 0.075% sample shows similar klat values to that of pristine

sample. However, for xZ0.125% samples, klat decreases compared to pristine samples. The room temperature klat decrease from B1.81 Wm 1 K 1 for both pristine PbSe and x 0.075%

to B1.56 W m 1 K 1, B1.56 W m 1 K 1, B1.54 W m 1 K 1, B1.51 W m 1 K 1, B1.51 W m 1 K 1 and 1.42 W m 1 K 1 for x 0.125, 0.25, 0.5, 0.75, 1 and 2% of Sb, respectively (See

Table 1 and Fig. 4c). The reduction of klat is even greater at higher temperature. At 700 K, klat decrease from B1.18 W m 1 K 1,

B1.06 W m 1 K 1, B0.82 W m 1 K 1, B0.68 W m 1 K 1, B0.66 W m 1 K 1, B0.60 W m 1 K 1, B0.55 W m 1 K 1 and B0.56 W m 1 K 1 for x 0, 0.075, 0.125, 0.25, 0.5, 0.75,

1 and 2% of Sb, respectively (See Table 1 and Fig. 4c). The lowest value of B0.42 W m 1 K 1 is obtained at 930 K for x 2%

sample. Clearly, very large reductions in the Sb system are observed.

In contrast, Pb1 xBixSe shows no such effects and instead

similar klat values to that of pristine PbSe are observed. The room

temperature klat is B1.79 W m 1 K 1, B1.81 W m 1 K 1,

B1.77 W m 1 K 1, B1.76 W m 1 K 1, B1.69 W m 1 K 1, B1.58 W m 1 K 1 and B1.59 W m 1 K 1 for x 0, 0.25, 0.5,

0.75, 1, 1.5 and 2% of Bi, respectively (See Table 1 and Fig. 4c). By adding Bi to PbSe matrix, the high temperature klat does not

change either. The klat at 700 K is B1.15 W m 1 K 1,

B1.17 W m 1 K 1, B1.10 W m 1 K 1, B1.13 W m 1 K 1, B1.10 W m 1 K 1, B1.17 W m 1 K 1 and B1.16 W m 1

K 1 for for x 0, 0.25, 0.5, 0.75, 1, 1.5 and 2% of Bi, respectively

(See Table 1 and Fig. 4c). Even the lowest value of Bi system is quite high with B1.00 W m 1 K 1at 920 K for x 0.5%.

Nanostructuring and TEM. The very large reductions in lattice thermal conductivity of the Sb-containing system can be explained by understanding the microstructure, particularly at the nanoscale. Typical low- and medium-magnication bright eld TEM images of Pb1 xSbxSe samples are shown in Fig. 5. The

low-magnication TEM images of 0.5% Sb-doped sample, acquired respectively along [001] (Fig. 5a) and [011] zone axes (Fig. 5b), show the dark-strain contrast from the precipitates. In the corresponding selected-area electron diffraction (SAED)

patterns (insets in Fig. 5a,b), which include both the matrix and the precipitates, only Bragg spots consistent with that of PbSe can be observed, conrming crystallographic alignment (endotaxy) of the nanoscale precipitates and matrix. Shown in Fig. 5cf are the atomic-scale high-resolution TEM images. When the electron beam is parallel to the [001] zone axis (Fig. 5c,e,f), the precipitates appear as dark straight and thick line-segments with length of a few nanometres that are orthogonal to each other in about equal proportion. The directions normal to these projected lines are either [010] or [100] as indicated. When the beam is oriented along [011] zone axis (Fig. 5d), most of the precipitates appear to have quasi-circular (oval) and elongated shape. These observations are consistent with an oval disc-shaped morphology of Sb-rich precipitates with three crystallographic variants, along three sets of {100} planes. Figure 5c shows two such obvious variants orthogonal to each other, while the third variant is within the thin foil with normal to the eld of view; which is expected to have low contrast, as indicated by the dotted arrows and the circle, respectively. Figure 5d conrms the elongated sectional view expected for [011] matrix orientation. Figure 5g shows the schematic illustration of the three variants of the oval disc-shaped precipitates parallel to {100}, and with their projection along /001S and /011S directions.

The number density of these oval disc-shaped precipitates increases with increasing Sb fraction. A high number of well-dened nanoscale precipitates is observed for the Pb0.995Sb0.005Se

samples. High-number density of precipitates is also observed for samples with xZ0.125% of Sb (Fig. 5e) that exhibit considerably reduced klat. These characteristics point to additional scattering mechanisms arising from the matrix-nanoscale precipitate interfaces, mass-contrast effects as well as accompanying spatially varying strain. In contrast, the even more dilute Pb0.99925Sb0.00075Se sample (Fig. 5f) showed few or no precipitates

and its klat is similar to that of the pristine compound, strongly suggesting that nanostructure scattering mechanism for heat-carrying phonons is not prevalent for much diluted Sb-doped samples. The existence of Sb in the nanoscale precipitates was conrmed by energy dispersive spectroscopy (EDS; Fig. 5h) for Pb0.9Sb0.1Se samples, which contain similar nanoscale precipitates and morphology; and are conducive to EDS analysis given larger amount of Sb in these samples. The main EDS spectrum shown in Fig. 5h was obtained when the beam was focused at a precipitate. As shown in the inset of Fig. 5h, the relative intensity of the Sb La1 peak is higher when the beam is focused at the precipitates (red line) compared with the matrix (black line). Unlike Sb-doped samples, Bi-doped samples do not exhibit nanostructures (Supplementary Fig. 5), suggesting that it forms solid solution with matrix PbSe, and thereby explains why Bi-doped samples show similar klatas pristine PbSe.

One of the primary goals of this investigation was to determine the evolution of klat with changing dopants (Sb versus Bi) and their amount (mol% fraction). In Fig. 4c, the lattice thermal conductivity at 300 K and 700 K is plotted as a function of x for both Pb1 xSbxSe and Pb1 xBixSe. In the Pb1 xSbxSe samples,

there is a marked decrease in klat with the addition of even fewer per cent of Sb at both 300 and 700 K. The decrease in klat is more

dramatic at high temperature. The values of klat at 300 K and

700 K for Pb1 xBixSe samples are shown for comparison. Adding

Bi seems to slightly reduce klat values at room temperature but there is almost no effect at high temperature. Thus, all klat values

at 700 K for Bi samples up to x 2% are similar to the value of

pristine PbSe. This trend suggests that the solubility limit of Sb is B0.125 mol% Sb, but that of Bi is Z2 mol% Bi. In addition to the lack of nanoscale precipitates, the lower mass contrast between Pb and Bi further reduces the prospects for additional phonon scattering mechanism.

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640

[001]

[011]

[100]

[010]

1-11

200

020

200

[001] [011]

[001] [001]

[010]

Pb 25,000

8,000

20,000

L1 Se

4,000

Precipitate

Matrix

Precipitate

Intensity (a.u.)

15,000

L1 K1

Pb L1

10,000

Cu K1K1 0

0 5Energy (keV)

Sb L1

5,000

03.0 3.5 4.0 4.5

Pb L1

10 15 20

a b

Figure 5 | TEM and EDS spectra of Pb1 xSbxSe. (a) Low-magnication TEM images of Pb0.995Sb0.005Se along the [001] and (b) [011] zone axes. The insets show the SAED patterns. (c) Medium-magnication TEM images of Pb0.995Sb0.005Se along the [001] and (d) [011] zone axes, (e) Pb0.9975

Sb0.0025Se, and (f) Pb0.99925Sb0.00075Se, both along the [001] zone axis. The arrows and circle shown in (c) indicate the oval disc and elongated

shapes of the precipitates. (g) schematic illustration of the three crystallographic variants of oval disc-shape precipitates. (h), EDS spectra of Pb0.9Sb0.1Se

sample from the precipitate The inset is the magnied Sb peak comparing between the precipitate and the matrix. Scale bar, 20 nm (a,b), 5 nm (cf).

8 NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640 ARTICLE

1,000

800

600

Pb1-xSbxSe

Pb1-xBixSe

H(cm2 V1 s1 )

400

200

0 0.5 1 1.5 x (%)

1.6 Pb1-xSbxSe

1.4

1.2

0.8

0.6

0.075%0.125%0.25%0.5%0.75%

0.4

0.2

200 400 600 800Temperature (K)

Figure 6 | Contrast between Pb1 xSbxSe and Pb1 xBixSe and ZT of

Pb1 xSbxSe. (a) Mobility of Pb1 xSbxSe samples (black) and Pb1 xBixSe

samples (red) shown as a function of x at room temperature. Solid lines are guide to the eye. (b), the thermoelectric gure of merit as a function of temperature for Pb1 xSbxSe samples.

In Fig. 6a, the Hall mobility at 300 K is presented as a function of x for both Pb1 xSbxSe and Pb1 xBixSe. The Sb-rich endotaxial

nanocrystals in the PbSe matrix do not appear to affect the electron scattering. Pb1 xSbxSe samples have even better mobility

than Pb xBixSe samples that lack nanostructuring. A possible

reason for this is the endotaxy of Sb-rich precipitates in PbSe that leaves fewer point defects of Sb dissolved in the Pb sublattice. By comparison, the more soluble Bi creates more point defects directly in the Pb sublattice and raises the probability of electron scattering. This conuence of endotaxially arranged Sb-rich nanocrystals effectively inhibiting heat ow in the system, but not electron ow, allows for the large power factor to be achieved.

Thermoelectric gure of merit. Figure 6b depicts the temperature dependence of ZT for Pb1 xSbxSe samples (corre

sponding results of Pb1 xBixSe samples are shown in Fig. 3f).

Combining high-power factors and low-thermal conductivity gives high ZT values. The highest ZT of B1.5 at 830 K is observed for the Pb0.9975Sb0.0025Se sample. This system outperforms other

bulk n-type PbSe systems and might be able to control the different optimal temperatures with varying dopant concentration. This value matches the best n-type PbTe-based materials. We nd high ZT values over the large temperature range as the content of Sb is tuned. Thus, ZT values of B0.8 at 600 K for

Pb0.99925Sb0.00075Se, B1.2 at 700 K for Pb0.99875Sb0.00125Se, B1.5

at 830 K for Pb0.9975Sb0.0025Se, B1.4 at 930 K for Pb0.995Sb0.005Se,

and B1.2 at 930 K for Pb0.9925Sb0.0075Se and Pb0.99Sb0.01Se

samples are observed. In comparison, ZT values for all Bi samples are lower than 1 (B0.83 at 690 K for Pb0.99925Bi0.00075Se, B0.89

at 920 K for Pb0.995Bi0.005Se, B0.73 at 920 K for Pb0.99Bi0.01Se and

B0.63 at 910 K for Pb0.985Bi0.015Se).

Counter to expectations, Sb in PbSe does not behave as an ideal dopant. A very small amount of Sb causes a surprisingly extensive nanostructuring in PbSe, resulting in dramatic increase in thermoelectric performance. It is remarkable that only0.25 mol% of Sb can generate such an extensive array of well-dened endotaxial nanoscale precipitates that do not reduce power factor and function predominantly as strong phonon scatterers, particularly at high temperature. As a result, ZT of B1.5 is achieved for this system at 830 K, the highest ever reported for PbSe-based systems. This essentially matches the highest ZT value of any n-type PbTe-based materials. The addition of Bi instead seems to lead to ideal dopant behaviour, that is, solid solution formation and carrier generation but has no other benecial effect on thermal conductivity reduction compared with pristine PbSe. From this work, it is apparent that PbSe-based materials are now emerging as promising, less expensive alternatives to PbTe that have the potential for enabling energy conversion devices with good thermoelectric performance at higher operating temperatures, exceeding the traditional PbTe-based materials. The results reported here for n-type PbSe samples coupled with the previous achievement of high ZT B1.6 in p-type PbSe samples, propel the PbSe system to the higher echelons of the group of top-performing thermo-electric materials composed of more earth abundant elements. We believe these results open new vistas for broad-based viable high-temperature thermoelectric power generation applications and high-temperature waste-heat utilization.

Methods

Synthesis. Ingots (B20 g) with nominal compositions Pb1 xSbxSe (x 0,

0.00075, 0.00125, 0.0025, 0.005, 0.0075, 0.01 and 0.02) and Pb1 xBixSe (x 0.0025,

0.005, 0.0075, 0.01, 0.015 and 0.02) were prepared by mixing appropriate ratios of reagents in quartz tubes. The tubes were sealed under vacuum (B10 4 Torr) and heated up to 1,150 C over a period of 12 h, soaked at that temperature for 5 h and rapidly cooled to room temperature over 3 h. The obtained ingots were cleaned and ground to a powder using a mechanical grinder to reduce the particle sizes to o53 mm. These powders were densied at 650 C for 5 min under an axial compressive pressure of 40 MPa in an argon atmosphere. The relative densities were achieved in the range of 97.8100% of theoretical values (see Supplementary Information).

PXRD. PXRD patterns for all samples were collected using Cu Ka radiation on an INEL diffractometer operating at 40 kV and 20 mA equipped with a position sensitive detector.

Electrical properties. The samples for charge transport properties measurement were cut and polished into a parallelepiped with dimensions of B2 mm

3 mm 10 mm. The electrical conductivity and Seebeck coefcient were measured

simultaneously under a helium atmosphere (B0.1 atm) from room temperature to B923 K using an ULVAC-RIKO ZEM-3 system. Samples measured to 923 K were coated with boron nitride to protect the instrument and the sample against evaporation of elements. Multiple cycles and different specimens produced similar properties.

Hall measurement. The charge transport specimens were subsequently used in room temperatures Hall-effect measurements to determine their carrier densities on the assumption of a unitary Hall factor, which gives a Hall coefcient, RH 1/

nHe, where nH is the Hall charge-carrier density and e is the electronic charge. Hall coefcients were measured in a home-built system in a magnetic eld ranging from0.5 to 1.25 T, utilizing a simple four-contact Hall-bar geometry, in both negative and positive polarity of the magnetic eld to eliminate Joule-resistive errors. The high-temperature Hall coefcient was measured for selected Sb-containing samples (x 0.25, 0.5, 0.75%) in a homemade high-temperature apparatus under argon

atmosphere, which provides a working range from 300 to 850 K. The charge transport specimens polished to thin slab samples with dimensions of B1 mm 2 mm 8 mm were used. The Hall resistance was recorded with a

Linear Research ac resistance bridge (model LR-700), which was operated at 17 Hz. The magnetic eld was provided by an air-bore Oxford superconducting magnet and the data were taken from 0.5 to 0.5 T. The carrier concentration is calculated

from the same assumption of a unitary Hall factor. The values of carrier

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640

concentration at room temperature are similar to those from room temperature Hall-effect measurement within 5% error. The mobility values, m, at room temperature were calculated from the relation mH s/nHe.

Thermal conductivity. A NETZSCH LFA 457 MicroFlash instrument was used to determine the thermal diffusivity of samples cut and polished to rectangular parallelepipeds for perpendicular direction measurement and disk-shape for parallel direction measurement (Fig. 1c) coated with graphite. Thermal conductivity is calculated from k DCpd, where D is the thermal diffusivity, Cp is the heat

capacity, and d is the density. The specic heat per unit mass of the PbSe was used literature values47 of the heat capacity of PbSe:

Cp 0:170778 2:648764 10 5 T K

J g 1K 1

Supplementary Fig. 6 shows the heat capacities obtained in the thermal diffusivity measurement using a pyroceram 9606 standard. The literature heat capacity is in good agreement with these data. The mass-density values used here were calculated using the geometrical dimensions of the specimens measured with a digital caliper and mass measured on a four-digit balance; corresponding densities were Z97.8% theoretical density. The details of D and d are shown in

Supplementary Fig. 7, and Supplementary Table 1.

TEM. The samples were characterized under the JEOL 2100F S/TEM operated at 200 kV. Thin TEM specimens were prepared by conventional methods and include cutting, grinding, dimpling, polishing and low-energy Ar-ion milling equipped with a liquid nitrogen cooling stage.

References

1. Kanatzidis, M. G. Nanostructured thermoelectrics: the new paradigm?w. Chem. Mater. 22, 648659 (2010).

2. Sootsman, J. R., Chung, D. Y. & Kanatzidis, M. G. New and old concepts in thermoelectric materials. Angew. Chem. Int. Ed. Engl. 48, 86168639 (2009).

3. Bell, L. E. Cooling, heating, generation power, and recovering waste heat with thermoelectric systems. Science 321, 14571461 (2008).

4. Lee, Y. et al. High-performance tellurium-free thermoelectrics: all-scale hierarchical structuring of p-type pbse-mse systems (M Ca, Sr, Ba). J. Am.

Chem. Soc. 135, 5152 (2013).5. Zhao, L.-D. et al. High thermoelectric performance via hierarchical compositionally alloyed nanostructures. J. Am. Chem. Soc. 135, 7364 (2013).

6. Zhao, L.-D. et al. Thermoelectrics with earth abundant elements: high performance p-type PbS nanostructured with SrS and CaS. J. Am. Chem. Soc. 134, 79027912 (2012).

7. Zhao, L.-D. et al. High performance thermoelectrics from earth-abundant materials: enhanced gure of merit in PbS by second phase nanostructures.J. Am. Chem. Soc. 133, 2047620487 (2011).8. Johnsen, S. et al. Nanostructures boost the thermoelectric performance of PbS.J. Am. Chem. Soc. 133, 34603470 (2011).9. Androulakis, J. et al. Thermoelectrics from abundant chemical elements: high performance nanostructured PbSe-PbS. J. Am. Chem. Soc. 133, 10920 (2011).

10. Vineis, C. J., Shakouri, A., Majumdar, A. & Kanatzidis, M. G. Nanostructured thermoelectrics: big efciency gains from small features. Adv. Mater. 22, 39703980 (2010).

11. Androulakis, J. et al. Nanostructuring and high thermoelectric efciency in p-Type Ag(Pb1 ySny)mSbTe2 m. Adv. Mater. 18, 11701173 (2006).

12. Sootsman, J. R., Pcionek, R. J., Kong, H., Uher, C. & Kanatzidis, M. G. Strong reduction of thermal conductivity in nanostructured pbte prepared by matrix encapsulation. Chem. Mater. 18, 4993 (2006).

13. Hsu, K. F. et al. Cubic AgPbmSbTe2 m: bulk thermoelectric materials with

high gure of merit. Science 303, 818821 (2004).14. Bilc, D. et al. Resonant states in the electronic structure of the high performance thermoelectrics AgPbmSbTe2m: the role of Ag-Sb microstructures. Phys. Rev.

Lett. 93, 146403 (2004).15. Poudeu, P. F. P. et al. High thermoelectric gure of merit and nanostructuring in bulk p-type Na1-xPbmSbyTem2. Angew. Chem. Int. Ed. Engl. 45, 38353839 (2006).

16. Poudel, B. et al. High-thermoelectric performance of nanostructured bismuth antimony telluride bulk alloys. Science 320, 634638 (2008).

17. Narducci, D., Selezneva, E., Cerofolini, G., Frabboni, S. & Ottaviani, G. Impact of energy ltering and carrier localization on the thermoelectric properties of granular semiconductors. Solid State Chem. Mater. Sci. Thermoelect. Mater. 193, 19 (2012).

18. Perlt, S. et al. Microstructure analyses and thermoelectric properties of AgPb18SbTe20. J. Solid State Chem. 193, 58 (2012).

19. Biswas, K. et al. Strained endotaxial nanostructures with high thermoelectric gure of merit. Nat. Chem. 3, 160166 (2011).

20. Zhao, L.-D. et al. Raising the thermoelectric performance of p-type PbS with endotaxial nanostructuring and valence-band offset engineering using CdS and ZnS. J. Am. Chem. Soc. 134, 16327 (2012).

21. Heremans, J. P., Wiendlocha, B. & Chamoire, A. M. Resonant levels in bulk thermoelectric semiconductors. Energy Environ. Sci. 5, 5510 (2012).

22. Heremans, J. P. et al. Enhancement of thermoelectric efciency in PbTe by distortion of the electronic density of states. Science 321, 554557 (2008).

23. Liu, W. et al. Convergence of conduction bands as a means of enhancing thermoelectric performance of n-Type Mg2Si1 xSnx solid solutions. Phys. Rev.

Lett. 108, 166601 (2012).24. Hu, Z. & Gao, S. Upper crustal abundances of trace elements. Chem. Geol. 253, 205221 (2008).

25. Androulakis, J., Lee, Y., Todorov, I., Chung, D. Y. & Kanatzidis, M. G. High-temperature thermoelectric properties of n-type PbSe doped with Ga, In, and Pb. Phys. Rev. B 83, 195209 (2011).

26. Zhang, Q. et al. Study of the thermoelectric properties of lead selenide doped with boron, gallium, indium, or thallium. J. Am. Chem. Soc. 134, 17731 (2012).

27. Androulakis, J. et al. High-temperature charge and thermal transport properties of the n-type thermoelectric material PbSe. Phys. Rev. B 84, 155207 (2011).

28. Wang, H., Pei, Y., LaLonde, A. D. & Snyder, G. J. Weak electronphonon coupling contributing to high thermoelectric performance in n-type PbSe. Proc. Natl Acad. Sci. USA. 109, 9705 (2012).

29. Zhang, Q. et al. Enhancement of thermoelectric gure-of-merit by resonant states of aluminum doping in lead selenide. Energy Environ. Sci. 5, 5246 (2012).

30. Wang, H., Pei, Y., LaLonde, A. D. & Snyder, G. J. Heavily doped p-type PbSe with high thermoelectric performance: an alternative for PbTe. Adv. Mater. 23, 13661370 (2011).

31. Wang, S. et al. Exploring the doping effects of Ag in p-type PbSe compounds with enhanced thermoelectric performance. J. Phys. D. Appl. Phys. 44, 475304 (2011).

32. Parker, D. & Singh, D. High-temperature thermoelectric performance of heavily doped PbSe. Phys. Rev. B 82, 035204 (2010).

33. Parker, D., Singh, D. J., Zhang, Q. & Ren, Z. Thermoelectric properties of n-type PbSe revisited. J. Appl. Phys. 111, 123701 (2012).

34. Schockley, W. Electrons and Holes in Semiconductors (Van Nostrand, 1950).35. Schlichting, U. & Gobrecht, K. H. The mobility of free carriers in PbSe crystals.J. Phys. Chem. Solids 34, 753758 (1973).36. Gobrecht, K. H. & Schlichting, U. Some electrical properties of lead selenide at low temperatures (0.3K300K). Solid State Commun. 8, 197199 (1970).

37. Batista, J., Mandelis, A. & Shaughnessy, D. Temperature dependence of carrier mobility in Si wafers measured by infrared photocarrier radiometry. Appl. Phys. Lett. 82, 4077 (2003).

38. Ravich, Y. I., Emova, B. A. & Tamarchenko, V. I. Scattering of current carriers and transport phenomena in lead chalcogenides ii. experiment. Physica. Status Solidi. B 43, 453 (1971).

39. Ravich, B. A., Emova, B. A. & Smirnov, I. A. Semiconducting Lead Chalcogenides (Plenum, 1970).

40. Wang, H., Pei, Y., LaLonde, A. D. & Snyder, G. J. in Thermoelectric Nanomaterials. (eds Koumoto, K. & Mori, T.) (Spring Berlin Heidelberg, 2013).

41. Lundstrom, M. & Jeong, C. Near-Equilibrium Transport (World Scientic Publishing Company, 2013).

42. May, A. F. & Snyder, G. J. in Thermoelectrics and its Energy Harvesting Vol. 1. (ed. Rowe, D. M.) (CRC Press, 2012).

43. May, A., Toberer, E., Saramat, A. & Snyder, G. Characterization and analysis of thermoelectric transport in n-type Ba8Ga16 xGe30x. Phys. Rev. B 80, 125205

(2009).44. Landolt, H. & M Bornstein, R. Landolt-Brnstein: Numerical Data and Functional Relationships in Science and Technology (Springer, Berlin, 1983).

45. Zhang, Y., Ke, X., Chen, C., Yang, J. & Kent, P. R. C. Thermodynamic properties of PbTe, PbSe, and PbS: rst-principles study. Phys. Rev. B 80, 024304 (2009).

46. Fistul, V. I. Heavily Doped Smiconductors Vol. 1 (Plenum Press, 1969).47. Blachnik, R. & Igel, R. Thermodynamische Eigenschaften von IV-VIVerbindungen: Bleichalkogenide. Zeitschrift fur Naturforschung B 29, 625629 (1974).

Acknowledgements

This material is based upon work supported as part of the Revolutionary Materials for Solid State Energy Conversion, an Energy Frontier Research Center funded by the U.S. Department of Energy, Ofce of Science, Ofce of Basic Energy Sciences, under Award No. ED-SC 0001054. This work was also supported by the U.S Department of Energy, Ofce of Science, under Contract No. DE-AC02-06CH11357. TEM work was performed in the EPIC/NIFTI/Keck-II facility of the NUANCE Center at Northwestern University. The NUANCE Center is supported by NSF-NSEC, NSF-MRSEC, Keck Foundation, the State of Illinois, and Northwestern University.

Author contributions

Y.L. and M.G.K. prepared the samples, designed, carried out thermoelectric experiments and analysed the electrical and thermal transport data. S.-L., C.C. and V.P.D. carried out the TEM experiment and analysed the TEM data. H.S. and C.U. carried out the Hall

10 NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4640 ARTICLE

measurements. D.C. helped in some electrical transport measurement. Y.L. and T.C.C. gave physical explanations. Y.L., S.L., T.C.C., V.P.D. and M.G.K. wrote the manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Lee, Y. et al. Contrasting role of antimony and bismuth dopants on the thermoelectric performance of lead selenide. Nat. Commun. 5:3640 doi: 10.1038/ ncomms4640 (2014).

NATURE COMMUNICATIONS | 5:3640 | DOI: 10.1038/ncomms4640 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

Word count: 8254

Show less

Abstract

Translate

Increasing the conversion efficiency of thermoelectric materials is a key scientific driver behind a worldwide effort to enable heat to electricity power generation at competitive cost. Here we report an increased performance for antimony-doped lead selenide with a thermoelectric figure of merit of ~1.5 at 800 K. This is in sharp contrast to bismuth doped lead selenide, which reaches a figure of merit of <1. Substituting antimony or bismuth for lead achieves maximum power factors between ~23-27 μW cm^-1 K^-2 at temperatures above 400 K. The addition of small amounts (~0.25 mol%) of antimony generates extensive nanoscale precipitates, whereas comparable amounts of bismuth results in very few or no precipitates. The antimony-rich precipitates are endotaxial in lead selenide, and appear remarkably effective in reducing the lattice thermal conductivity. The corresponding bismuth-containing samples exhibit smaller reduction in lattice thermal conductivity.

Details

Title

Contrasting role of antimony and bismuth dopants on the thermoelectric performance of lead selenide

Author

Lee, Yeseul; Lo, Shih-han; Chen, Changqiang; Sun, Hui; Chung, Duck-young; Chasapis, Thomas C; Uher, Ctirad; Dravid, Vinayak P; Kanatzidis, Mercouri G

Pages

3640

Publication year

2014

Publication date

May 2014

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms4640

ProQuest document ID

1520399885

Contrasting role of antimony and bismuth dopants on the thermoelectric performance of lead selenide

Jump to:

Full text

Abstract

Details

Suggested sources