enhancing the figure-of-merit in half-heuslers for vehicle ... · enhancing the figure-of-merit in...

Enhancing the Figure-of-Merit in Half-Heuslers for Vehicle Waste Heat Recovery

Zhifeng Ren

Boston College Chestnut Hill

Massachusetts 02467

Email: [email protected]

Zhifeng Ren · Boston College · Department of Physics

mailto:[email protected]

      

 

Outline

• Why Half-Heuslers for auto waste heat recovery? • Status of Half-Heuslers before our work • The effect of nanostructures on thermoelectric figure-of-merit • The effect of larger differences in atomic mass and size on

thermal conductivity and thermoelectric figure-of-merit

• Bonus: New promising materials with good ZT


Why Half-Heuslers?


   

Culp et al., Appl. Phys. Lett. 88, 042106 (2006)

Status of Half-Heuslers N type, Hf0.75Zr0.25NiSn0.975Sb0.025 P type, Zr0.5Hf0.5CoSb0.8Sn0.2

Culp et al., Appl. Phys. Lett. 93, 022105 (2008)

• Thermal conductivity too high! • Nanocomposite approach: reduce thermal conductivity


Phonon Engineering nanostructure in n-type half-Heusler


646 © 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL

PAPER

wileyonlinelibrary.com

www.MaterialsViews.comwww.advenergymat.de

Adv. Energy Mater. 2011, 1, 643–647

within the experimental errors, and also includes the results of an annealed nanostructured sample (run 1), which does not show any signifi cant degradation in thermoelectric prop-erties after annealing. The sample was annealed at 800 °C for 12 hours in air, which is an accelerating condition since the application temperature is expected to be at or below 700 ° C.

Here, we also show the temperature-dependent specific heat capacity and thermal diffusivity of nanostructured Hf 0.75 Zr0.25 NiSn0.99 Sb0.01 samples in comparison with theingot sample. Figure 4 clearly shows that the specifi c heatcapacity of nanostructured Hf 0.75 Zr0.25 NiSn0.99 Sb0.01 samplesis almost the same as that of the ingot sample (Figure 4 a), andthese values agree fairly well with the Dulong and Petit valueof specifi c heat capacity (solid line). However, the thermaldiffusivity of nanostructured Hf 0.75 Zr0.25 NiSn0.99 Sb0.01 sam-ples is signifi cantly lower (Figure 4b) than that of the ingotsample due to the small grain-size effect and lower electronic contribution.

Since the size of the nanoparticles is essential in reducing the thermal conductivity to achieve higher ZT values, it is pos-sible to further increase the ZT value of the n-type half-Heusler compounds by making the grains even smaller. In this report, we made grains of 200 nm and greater (Figure 2 a). It would be possible to achieve a grain size of less than 100 nm by pre-venting grain growth during hot-pressing if the correct grain growth inhibitor were found.

3. Conclusions

A cost-effective ball-milling and hot-pressing technique has been applied to n-type HHs to improve the ZT value. A peak ZT value of around 1.0 at 600 ° C–700 °C is observed in nanostructured Hf 0.75 Zr0.25 NiSn0.99 Sb0.01 samples, which is about 25% higher than the previously reported best peak ZTvalue of any n-type half-Heuslers. [25] This enhancement in ZTmainly results from reduction in thermal conductivity due to the increased phonon scattering at the grain boundaries of nanostructures, and optimization of carrier contribution leading to lower electronic thermal conductivity. Further ZTimprovement would be possible if the grains were less than 100 nm in size.

4. Experimental Section Half-Heusler phases were prepared by melting hafnium (Hf; 99.99%, Alfa Aesar), zirconium (Zr; 99.99%, Alfa Aesar) chunks, nickel (Ni; 99.99%, Alfa Aesar), tin (Sn; 99.99%, Alfa Aesar), and antimony (Sb; 99.99%, Alfa Aesar) pieces according to a composition Hf 0.75Zr 0.25NiSn 0.99Sb 0.01 using an arc-melting process. The melted ingot was then milled for 1–50 hours to get the desired nanopowders by using a commercially available ball-milling machine (SPEX 8000M Mixer/Mill). The mechanically prepared nanopowders were then pressed at temperatures of 900–1200 °C using a dc hot-press method in graphite dies with a 12.7-mm central cylindrical opening diameter to get nanostructured bulk half-Heusler samples. The samples were characterized by XRD, SEM, and TEM to study their crystallinity, homogeneity,

average grain size, and the grain-size distribution of the nanoparticles. These parameters significantly affect the thermoelectric properties of the final dense bulk samples. The volume densities of these samples were measured using an Archimedes’ kit. The nanostructured bulk samples were then cut into 2 × 2 × 12 mm bars for electrical conductivity and Seebeck coefficient measurements on commercial equipment (Ulvac, ZEM-3), 12.7-mm-diameter discs with appropriate thickness for thermal diffusivity measurements on a laser flash system (Netzsch LFA 457) from 100 °C to 700 °C, and 6-mm-diameter discs with appropriate thickness for specific heat capacity measurements on a differential scanning calorimeter (200-F3, Netzsch Instruments, Inc.) from room temperature to 600 °C (the data point at 700 °C is extrapolated). The thermal conductivity was calculated as the product of the thermal diffusivity, specific heat capacity, and volume density of the samples. To confirm the reproducibility of the sample preparation process and reliability of the measurements of nanocrystalline bulk samples, the same experimental conditions were repeated at least three to six times for each composition and it was found that the thermoelectric properties are reproducible within 5% under the same experimental conditions. The volume densities of three measured nanostructured samples (runs 1, 2, and 3) are 9.73, 9.70, and 9.65 g cm −3, respectively.

Acknowledgements The work is funded by the “Solid State Solar-Thermal Energy Conversion Center (S 3TEC), an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Award Number: DE-SC0001299/DE-FG02–09ER46577 (ZFR and GC).

Received: March 11, 2011 Published online: May 3, 2011

Figure 4 . a) Temperature-dependent specific heat capacity and b) thermal diffusivity of arc-melted and then ball-milled and hot-pressed Hf 0.75Zr 0.25NiSn 0.99Sb 0.01 samples, and the sample annealed at 800 °C for 12 hours in air, in comparison with the ingot sample.

[ 1 ] S. J. Poon , Recent Trends in Thermoelectric Materials Research II , Semiconductors and Semimetals 2001, 70, 37 .

[ 2 ] 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 2008, 320, 634 .

[ 3 ] W. Xie , X. Tang , Y. Yan , Q. Zhang , T. M. Tritt , J. Appl. Phys. 2009, 105, 113713 .

[ 4 ] K. F. Hsu , S. Loo , F. Guo , W. Chen , J. S. Dyck , C. Uher, T. Hogan , E. K. Polychroniadis , M. G. Kanatzidis , Science 2004, 303, 818 .

[ 5 ] J. P. Heremans , V. Jovovic , E. S. Toberer, A. Saramat , K. Kurosaki , A. Charoenphakdee , S. Yamanaka , G. J. Snyder, Science 2008, 328, 554 .

Specific heat and thermal diffusivity


ZT Improvement due to Lower Thermal Conductivity by Nanostructures in n-type


Effect of Nanostructures on p-type Xiao Yan, et al., Nano Letters 11, 556-560 (2011)


559 dx.doi.org/10.1021/nl104138t |Nano Lett. 2011, 11, 556–560

Nano Letters LETTER

By comparison with the data on ingot samples previouslyreported,8 we found that the resistivity of our ingot sample isalmost the same as the reported value, and the Seebeck coeffi-cient is higher, which leads to a higher power factor. However,the thermal conductivity of our ingot sample is proportionallyhigher than the reported value,8 which leads to the same ZT of

our ingot samples as the reported value.8 These small differencesin individual properties may be due to someminor differences onsample preparation procedures, which is very reasonable andunderstandable.Since the ingot samples are made and measured by the same

person on the same machine with the nanostrcutured samples,

Figure 3. Temperature-dependent (a) electrical conductivity, (b) Seebeck coefficient, (c) power factor, (d) total thermal conductivity, (e) lattice part ofthermal conductivity, and (f) ZT of ball-milled and hot-pressed sample in comparison with that of the ingot.

Figure 4. Temperature-dependent specific heat (a) and thermal diffusivity (b) of ball-milled and hot-pressed sample in comparison with that ofthe ingot.

Specific heat and thermal diffusivity Xiao Yan, et al., Nano Letters 11, 556-560 (2011)


Lattice thermal conductivity and nanostructures Xiao Yan, et al., Nano Letters 11, 556-560 (2011)


Improved ZT by Nanostructures in p-type Half-Heuslers Xiao Yan, et al., Nano Letters 11, 556-560 (2011)


Phonon Engineering larger differences in mass and size in half-Heuslers


How to Further Cut the Cost?

Need to reduce the usage of Hf as much as possible.


Summary for Half-Heuslers


N-type P-type

LETTERS NATUREMATERIALS DOI: 10.1038/NMAT3273

[111]

Se (111) planea

c

bSeCu

a b

SeCu

Figure 1 | Crystal structure of Cu2Se at high temperatures (�-phase) with a cubic anti-fluorite structure. a, Unit cell where only the 8c and 32f interstitialpositions are shown with Cu atoms. b, Projected plane representation of the crystal structure along the cubic [11̄0] direction. The arrows indicate that theCu ions can freely travel among the interstitial sites. There are two Cu layers between the neighbouring Se (111) planes. The structure changes to amonoclinic ↵-phase by stacking the ordered Cu ions along the cubic [111] direction when cooled to room temperature.

α-Cu2Se

β-Cu2Se

α-Cu1.98Se

β-Cu1.98Se

S (µ

V K

¬1)

200 400 600 800 1,000T (K)

200 400 600 800 1,000T (K)

0

0.2

0.4

0.6

0.8

1.0

(10

¬4Ω

m)

ρ

0

100

200

300

400a b

Figure 2 | Thermoelectric properties of the low-temperature (↵) and high-temperature (�) phases in Cu2�x

Se. a–d, Temperature dependences ofelectrical resistivity ⇢ (a), thermopower S (b), thermal conductivity (c) and dimensionless figure-of-merit zT (d).

considering the large bandgap inCu2�xSe (ref. 22).Wemeasure holemobilities as high as 14–20 cm2 V�1 s�1 (Supplementary Fig. S4)at the carrier density on the order of 1020 cm�3 in our Cu2�xSe,comparable to other typical TE materials, such as skutterudites3.Based on the measured ⇢ and S, the calculated power factors(PF = S2/⇢) at room temperature are around 6–8 µWcm�1 K�1

for the ↵-phase and (7–12 µWcm�1 K�1) for the �-phase (420K–1,000K), again comparable to other typical TE materials. It isquite interesting that a compound such as Cu2�xSe, with theelectronegativity difference (0.65) between Cu and Se much higherthan the value in the good TEmaterial (usually less than 0.5; ref. 3),attains such a high value of the power factor.

Cu2�xSe has very low thermal conductivity values (around or lessthan 1Wm�1 K�1 for Cu2Se), see Fig. 2c. Although the transport ofholes across the phase boundary seems unimpeded, heat transportis very sensitive to the exact configuration of Cu ions, and a markeddiscontinuity is observed at the ↵–� transition (Fig. 2c). The latticethermal conductivity values, L, are around 0.4–0.6Wm�1 K�1 athigh temperatures (Supplementary Fig. S4), indicating the phononmean free path is quite small in this binary material. It is verysurprising that such very low values of L are realized in a compoundwith a simple chemical formula with a small unit cell and lightelements, whereas usually low L is reported in complex compoundswith heavy elements6–9.

The zT values are shown in Fig. 2d. Because of the very lowL and moderate values of the power factor, the zT reaches 1.5 at

1,000K in Cu2Se. This is very high value, and comparable to the bestof the novel bulk thermoelectric materials2. As the material shownhere is neither doped nor optimized, further fine tuning of Cu2�xSecould possiblymake it evenmore competitive.Wewish to stress thatthe transport data are fully reproducible on temperature cycling toat least 700K (see Supplementary Information).

The origin of very low values of L in Cu2�xSe can be associatedwith the fluid arrangement of the Cu atoms. The high temperature�-phase of Cu2�xSe is known to be a superionic conductor owingto the high mobility of its Cu ions25. Superionic conductors aresolid materials which have ionic conductivities that are as highas those found in molten salts, about 1� cm�1 (ref. 26). This isdue to high ionic diffusivities, of the order 10�5 cm2 s�1, the valuefor liquid water at room temperature. Thus a simplified model�-Cu2�xSe is one where there exists a solid, crystalline sublatticeof Se atoms around which a ‘liquid-like’ charged fluid of Cu ionsdiffuses (Fig. 1b). The ionic conduction of Cu in �-Cu2�xSe, witha diffusion coefficient of Cu similar to or larger than that of liquidwater, makes it one of the best superionic conductors25,27. Cu2�xSe,like many superionic conductors, undergoes a structural phasetransition when it becomes superionic to the liquid-like state, whichcan be considered a molten Cu sublattice26. Whereas the entropyof melting in alkali halides is 24 J K�1 mol�1, superionic conductorstypically split this entropy change between the superionic transitionand actual melting; Cu2�xSe increases its entropy by 17 J K�1 mol�1

when it becomes superionic across the ↵–� transition.

2 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials

LETTERSMethodsPolycrystalline Cu2 xSe samples were prepared by melting the elements Cu (shot,99.999%, Alfa Aesar) and Se (shot, 99.999%, Alfa Aesar) in a pyrolitic boron nitridecrucible enclosed in a fused silica tube at 1,423K for 12 h in vacuum, and thenslowly cooled down to 1,073K in 24 h and held there for seven days. Finally, thetubes were furnace cooled to room temperature. The resulting ingots were groundinto a fine powder by hand using an agate jar and plunger and subjected to sparkplasma sintering (Sumitomo SPS-2040) around 710K under a pressure of 65MPa.The actual composition is determined by electron microprobe analysis (EPMA,Shimadzu 8705QH2), showing almost the same values as the nominal startingcomposition. The high-temperature thermopower and electrical resistivity, and thecycling test were undertaken using an Ulvac ZEM-3 system under a sealed chamberwith a small amount of helium gas. The thermal diffusivity (D) and heat capacity(CP) from 300K to 1,000K were measured using the laser flash method (Netzsch,LFA427) and differential scanning calorimetry (Netzsch DSC 404F3), respectively.We measured Cp twice and the averaged value is shown in Fig. 4. The density (dd)was measured by Archimedes method. The thermal conductivity was calculatedusing =D⇥CP⇥dd . Hall effect measurements were performed using a cryostatequipped with a 5.5 T magnet. The thermal expansion coefficient was measuredusing a Linseis L75 system. TEM examination was performed on a JEM-2100Ffield-emission transmission electron microscope. The in situ TEM was undertakenusing a Gatan Model 652 heating specimen holder. The estimated measurementaccuracies are listed below for the commercial equipment used: 5% for electricalresistivity, 7% for thermopower, 10% for specific heat, 5% for thermal diffusivityand 1% for density. The data precision (reproducibility) is smaller than the accuracy(Supplementary Fig. S5), leading to zT values within the range of±0.2.

Received 10 October 2011; accepted 8 February 2012;published online 11 March 2012

References1. Bell, L. E. Cooling, heating, generating power, and recovering waste heat with

thermoelectric systems. Science 321, 1457–1461 (2008).2. Snyder, G. J. & Toberer, E. S. Complex thermoelectric materials. Nature Mater.

7, 105–114 (2008).3. Slack, G. A. in CRC Handbook of Thermoelectrics (ed. Rowe, D. M.) 407–440

(CRC, 1995).4. Sales, B. C., Mandrus, D. & Williams, R. K. Filled skutterudite antimonides: A

new class of thermoelectric materials. Science 272, 1325–1328 (1996).5. Christensen, M. et al. Avoided crossing of rattler modes in thermoelectric

materials. Nature Mater. 7, 811–815 (2008).6. Poudel, B. et al. High-thermoelectric performance of nanostructured bismuth

antimony telluride bulk alloys. Science 320, 634–638 (2008).7. Hsu, K. F. et al. Cubic AgPbmSbTe2+m: Bulk thermoelectric materials with high

figure of merit. Science 303, 818–821 (2004).8. Heremans, J. P. et al. Enhancement of thermoelectric efficiency in PbTe by

distortion of the electronic density of states. Science 321, 554–557 (2008).9. Pei, Y. et al. Convergence of electronic bands for high performance bulk

thermoelectrics. Nature 473, 66–69 (2011).10. Venkatasubramanian, R., Siivola, E., Colpitts, T. & O’Quinn, B. Thin film

thermoelectric devices with high room-temperature figures of merit. Nature413, 597–602 (2001).

11. Vineis, C. J., Shakouri, A., Majumdar, A. & Kanatzidis, M. G. Nanostructuredthermoelectrics: Big efficiency gains from small features. Adv. Mater. 22,3970–3980 (2010).

12. Rhyee, J. et al. Peierls distortion as a route to high thermoelectric performancein In4Se3 � crystals. Nature 459, 965–968 (2009).

13. Pilgrim, W. C. & Morkel, C. State dependent particle dynamics in liquid alkalimetals. J. Phys. Condens. Matter 18, R585–R633 (2006).

14. Frenkel, J. in Kinetic Theory of Liquids (eds Fowler, R. H., Kapitza, P. &Mott, N. F.) 188–249 (Oxford Univ. Press, 1947).

15. Trachenko, K. Heat capacity of liquids: An approach from the solid phase.Phys. Rev. B 78, 104021 (2008).

16. Chrissafis, K., Paraskevopoulos, K. M. & Manolikas, C. Studying Cu2 xSephase transformation through DSC examination. J. Therm. Anal. Cal. 84,195–199 (2006).

17. Skomorokhov, A. N., Trots, D. M., Knappb, M., Bickulova, N. N. & Fuess, H.Structural behaviour of �-Cu2 �Se (� = 0, 0.15, 0.25) in dependence ontemperature studied by synchrotron powder diffraction. J. Alloys Compounds421, 64–71 (2006).

18. Heyding, R. D. & Murry, R. M. The crystal structures of Cu1.8Se,Cu3Se2, ↵- and CuSe, CuSe2, and CuSe2II. Can. J. Chem. 54,841–848 (1976).

19. Oliveria, M., McMullan, R. K. & Wuensch, B. J. Single crystal neutrondiffraction analysis of the cation distribution in the high-temperaturephases ↵ Cu2 xS, ↵ Cu2 xSe, and ↵ Ag2Se. Solid State Ion. 28-30,1332–1337 (1988).

20. Borchert, W. Gitterumwandlungen im System Cu2 xSe. Z. Kristallogr. 106,5–24 (1945).

21. Stevels, A. L. N. & Jellinek, F. Phase transformations in copperchalcogenides: I. The copper–selenium system. Recl. Trav. Chim. 90,273–283 (1971).

22. Sorokin, G. P., Papshev, Yu. M. & Oush, P. T. Photoconductivity of Cu2S,Cu2Se, and Cu2Te. Sov. Phys. Solid State 7, 1810–1811 (1966).

23. Hampl, E. F. Jr & Grove, C. Thermoelectric composition, US patent 3,853,632(1974).

24. Ohtani, T. et al. Physical properties and phase transitions of �Cu2 xSe(0.20 x 0.25). J. Alloys Compounds 279, 136–141 (1998).

25. Chatov, V. A., Iorga, T. P. & Inglizyan, P. N. Ionic conduction anddiffusion of copper in copper selenide. Fiz. Tekh. Poluprovodn. 14,807–809 (1980).

26. Boyce, J. B. & Huberman, B. A. Superionic conductors: Transitions, structures,dynamics. Phys. Rep. 51, 189–265 (1979).

27. Takahashi, T., Yamamoto, O., Matsuyama, F. & Noda, Y. Ionic conductivityand coulometric titration of copper selenide. J. Solid State Chem. 16,35–39 (1976).

28. Kashida, S. & Akai, J. X-ray diffraction and electron microscopystudies of the room-temperature structure of Cu2Se. J. Phys. C 21,5329–5336 (1988).

29. Slack, G. A. The thermal conductivity of nonmetallic crystals. Solid State Phys.34, 1–71 (1979).

30. Capps, J., Drymiotis, F., Lindsey, S. & Tritt, T. M. Significant enhancementof the dimensionless thermoelectric figure of merit of the binary Ag2Te.Phil. Mag. Lett. 90, 677–681 (2010).

31. Gascoin, F. & Maignan, A. Order–disorder transition in AgCrSe2: A new routeto efficient thermoelectrics. Chem. Mater. 23, 2510–2513 (2011).

32. Chalfin, E., Lu, H. & Dieckmann, R. Cation tracer diffusion in thethermoelectric materials Cu3Mo6Se8 and ‘�-Zn4Sb3’. Solid State Ion. 178,447–456 (2007).

33. Snyder, G. J., Christensen, M., Nishibori, E., Caillat, T. & Iversen, B. B.Disordered zinc in Zn4Sb3 with phonon-glass and electron-crystalthermoelectric properties. Nature Mater. 3, 458–463 (2004).

AcknowledgementsThis work is in part supported by National Natural Science Foundation of China (NSFC)Grants (51121064 and 50825205), Shanghai Science and Technology Commission(Pujiang Program with No. 11PJ1410200 and Program of Shanghai Subject ChiefScientist with No. 09XD1404400), and CAS/SAFEA International Partnership Programfor Creative Research Teams. F.X. wishes to acknowledge the support of the NationalBasic Research Program of China (973 program) under Project 2009CB939904and NSFC Grants (60936001). C.U. wishes to acknowledge the support of theCenter for Solar and Thermal Energy Conversion Research Center funded by theDepartment of Energy (DOE) under No. DE-SC00000957. G.J.S. acknowledgessupport from AFOSR-MURI. Q.L. acknowledges support from the US DOE, Officeof Basic Energy Science, Materials Sciences and Engineering Division, under contractno. DEAC0298CH10886.

Author contributionsH.L. and X.S. prepared the samples and measured the thermoelectric properties. F.X.,W.Z., L.C., Q.L., G.J.S. and C.U. provided discussion on the experimental data. F.X.and L.Z. performed TEM measurements and analysis. T.D. and G.J.S. performedroom-temperature speed of sound measurements. H.L., X.S., L.C., Q.L., C.U. and G.J.S.wrote and edited the manuscript.

Additional informationThe authors declare no competing financial interests. Supplementary informationaccompanies this paper on www.nature.com/naturematerials. Reprints and permissionsinformation is available online at www.nature.com/reprints. Correspondence andrequests for materials should be addressed to X.S. or L.C.

4 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials

LETTERSPUBLISHED ONLINE: 11 MARCH 2012 | DOI: 10.1038/NMAT3273

Advanced thermoelectric technology offers a potential forconverting waste industrial heat into useful electricity, andan emission-free method for solid state cooling1,2. Worldwideefforts to find materials with thermoelectric figure of merit,zT values significantly above unity, are frequently focusedon crystalline semiconductors with low thermal conductivity2.Herewe report onCu2 xSe,which reaches a zT of 1.5 at 1,000K,among the highest values for any bulk materials. Whereasthe Se atoms in Cu2 xSe form a rigid face-centred cubiclattice, providing a crystalline pathway for semiconductingelectrons (or more precisely holes), the copper ions are highlydisordered around the Se sublattice and are superionic withliquid-like mobility. This extraordinary ‘liquid-like’ behaviourof copper ions around a crystalline sublattice of Se in Cu2 xSeresults in an intrinsically very low lattice thermal conductivitywhich enables high zT in this otherwise simple semiconductor.This unusual combination of properties leads to an idealthermoelectric material. The results indicate a new strategyand direction for high-efficiency thermoelectric materials byexploring systems where there exists a crystalline sublatticefor electronic conduction surrounded by liquid-like ions.

Solid-state thermoelectric (TE) technology uses electrons orholes as the working fluid for heat pumping and power generationand offers the prospect for novel thermal-to-electrical energyconversion technology that could lead to significant energy savingsby generating electricity from waste industrial heat1. The key tothe development of advanced TE technologies is to find highlyefficient TE materials. The conversion efficiency is governed by thedimensionless thermoelectric figure of merit zT = S2T/⇢(L +c),where S is the thermopower (Seebeck coefficient), ⇢ is theelectrical resistivity, T is the absolute temperature, L is the latticethermal conductivity and c is the carrier thermal conductivity.In current commercial materials2, the zT values are limited toaround unity. Recently, several concepts have been proposed toenhance the efficiency of TE materials and laboratory resultssuggest that high zT values can be realized in several families ofmaterials. Generally, the successful strategies follow the phonon-glass electron-crystal concept, where a crystalline semiconductorstructure is desirable for the electronic properties (electron-crystal) but a disordered atomic arrangement is desirable for lowlattice thermal conductivity (phonon-glass)3. Examples includesemiconductors with caged structures, such as skutterudites4and clathrates5, or the use of fine microstructure, such asnano-inclusions in bulk matrices to form bulk nano-compositematerials6,7. Other enhancements are found with the modificationof the electron states with resonant impurities8 or engineering band

1CAS Key Laboratory of Energy-conversion Materials, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China, 2GraduateUniversity of Chinese Academy of Sciences, Beijing 100049, China, 3State Key Laboratory of High Performance Ceramics and Superfine Microstructure,Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China, 4Condensed Matter Physics and Materials Science Department,Brookhaven National Laboratory, Upton, New York 11973, USA, 5Department of Physics, University of Michigan, Ann Arbor, Michigan 48109, USA,6Department of Materials Science, California Institute of Technology, Pasadena, California 91125, USA. *e-mail: [email protected]; [email protected].

convergence9, nanostructures10,11, and strong electron–phononcoupling by charge density waves12.

Crystalline semiconductors usually possess high heat conduc-tivity because the phonon mean free path is long in a periodicstructure. Disrupting the periodicity or adding defects reduces thephonon mean free path (scattering phonons) to lower L, butsuch a reduction is limited to the of a glass. Whereas a solidglass propagates some heat through transverse, shear vibrations, aliquid does not propagate shear vibrations13–15. Here we describea strategy to decrease L below that of a glass by reducing notonly the mean free path of lattice phonons but also eliminatingsome of the vibrational modes completely. The idea of using theliquid-like behaviour of superionic conductors may be consideredan extension of the phonon-glass electron-crystal concept andsuch materials could be considered phonon-liquid electron-crystal(PLEC) thermoelectrics.

Copper chalcogenides Cu2 xX (X = S, Se or Te), in spite of theirsimple chemical formula, have quite complex atomic arrangements.The Cu-deficient region (Cu2 xSe) in the phase diagram of theCu–Se system shown in Supplementary Fig. S1 exhibits two phases,a low-temperature ↵-phase and a high-temperature -phase, with asignificant deficiency of Cu allowed in the chemical stoichiometryof Cu2 xSe. The high-temperature -phase has Se atoms in a simpleface-centred cubic (fcc) structure with the space group Fm3̄m(Fig. 1 and Supplementary Fig. S2; refs 16,17), but the superionicCu ions are kinetically disordered throughout the structure18,19. Thelow-temperature ↵-phase of stoichiometric Cu2Se is stable up toabout 400K and has a lower symmetry crystal structure18,20,21 wherethe Cu atoms are localized and not superionic. According to simpleelectron counting rules, the exactly stoichiometric composition(Cu2Se) is expected to be an intrinsic semiconductor. Indeed,optical studies indicate a bandgap of 1.23 eV (ref. 22). Withnon-stoichiometry, x > 0, Cu2 xSe is a surprisingly good p-typeelectrical conductor with a high thermopower23 and its conductivityincreases with increasing deficiency of Cu. It can be assumed thatthe Cu deficiency results in an equivalent concentration of holes, assupported by transport measurements24.

The transport properties of our polycrystalline Cu2 xSe samplesare shown in Fig. 2 and cover the behaviour of both low- and high-temperature phases. Thematerial has low ⇢ and high S values in thewhole temperature range (Fig. 2a and b). The resistivity is on theorder of 10 4–10 5 �m.These values are comparable to other state-of-the-art TE materials2. The Seebeck coefficient in the ↵-phase(300–350K) is between 60 and 100 µVK 1 and in the -phase rangeof temperatures from 420 to 1,000K reaches values between 80 and300 µVK 1. Such values are feasible in extrinsically doped material

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 1

A New Potentially Interesting TE Material: Cu2Se

Copper ion liquid-like thermoelectrics Huili Liu1,2, Xun Shi1,3*, Fangfang Xu3, Linlin Zhang3, Wenqing Zhang3, Lidong Chen1*, Qiang Li4, Ctirad Uher5, Tristan Day6 and G. Jeffrey Snyder6

NATURE MATERIALS DOI: 10.1038/NMAT3273

(W m

¬1 K

¬1 )

zT

κ

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0

0.4

0.8

1.2

1.6

c d

200 400 600 800 1,000 200 400 600 800 1,000 T (K) T (K)


journal homepage: www.elsevier.com/locate/nanoenergy

Available online at www.sciencedirect.com

RAPID COMMUNICATION

aDepartment of Physics, Boston College, Chestnut Hill, MA 02467, USAbDepartment of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

Received 23 February 2012; accepted 25 February 2012

KEYWORDSCopper selenide;Thermoelectrics;Ordered anddisordered layers;Natural superlattice;Specific heat

AbstractThermoelectric figure-of-merit (ZT) of !1.6 at 700 1C is achieved in b-phase copper selenide(Cu2Se) made by ball milling and hot pressing. The b-phase of such material possesses a naturalsuperlattice-like structure that combines ordered selenium (Se) and disordered copper (Cu)layers in its unit cell, resulting in a low lattice thermal conductivity of 0.4–0.5 W m"1 K"1. Al-shaped specific heat peak indicates a phase transition from cubic b-phase to tetragonala-phase at around 140 1C upon cooling and vice versa. An abnormal decrease in specific heat withincreasing temperature was also observed due to the increased random motion of disordered Cuatoms. These features indicate that b-phase Cu2Se could be potentially an interesting thermo-electric material, competing with other conventional thermoelectric materials.& 2012 Elsevier Ltd. All rights reserved.

Introduction

A good thermoelectric material has high dimensionless figure-of-merit ZT: defined as (S2s/k)T, where the S, s, k, and T arethe Seebeck coefficient, electrical conductivity, thermalconductivity, and absolute temperature, respectively. Nor-mally, k is composed of three components, i.e., electroniccontribution (kcar), lattice contribution (klat), and bipolar

contribution (kbipolar). As reducing klat was proven to be theeasiest and most straightforward way to improve ZT, numerousefforts have been devoted in the last two decades in order toincrease the ZT value from the longstanding 1.0 in thermo-electric bulk materials to higher values by minimizing klatthrough the concept of nanocomposite [1–4]. The key idea ofnanocomposite is to create grains or inclusions that scatter thephonons without deteriorating the electron transport. Moregenerally, a good thermoelectric material should behave as a‘‘phonon-glass-electron-crystal’’ with a high charge carriermobility and a low lattice thermal conductivity [5]. Structureswith strong phonon scattering centers, such as skutterudites[6] and clathrates [7] were first demonstrated to have‘‘phonon-glass-electron-cystal’’ type of behavior resulting in

2211-2855/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.doi:10.1016/j.nanoen.2012.02.010

nCorresponding authors.E-mail addresses: [email protected] (G. Chen),

[email protected] (Z. Ren).1These authors contributed equally to this work.

Nano Energy (]]]]) ], ]]]–]]]

Please cite this article as: B. Yu, et al., Thermoelectric properties of copper selenide with ordered selenium layer and disordered copperlayer, Nano Energy (2012), doi:10.1016/j.nanoen.2012.02.010



RAPID COMMUNICATION

Thermoelectric properties of copper selenide withordered selenium layer and disordered copper layer

Bo Yua,1, Weishu Liua,1, Shuo Chena, Hui Wanga, Hengzhi Wanga,Gang Chenb,n, Zhifeng Rena,n





Introduction






Nano Energy (]]]]) ], ]]]–]]]

Please cite this article as: B. Yu, et al., Thermoelectric of copper selenide with ordered selenium layer and disordered copperlayer,

above. This observation suggests the phase change andpossible disordering of Cu sites at elevated temperature,which agrees well with our aforementioned XRD results.Figure 2(c) shows that the grains are typically within a fewmicrometers.

Thermoelectric transport properties

Figure 3 shows the temperature dependent thermoelectricproperties of Cu2Se1.01 samples that were hot pressed atdifferent temperatures. The extra Se was used to compen-sate the potential Se loss during hot pressing due to its

high vapor pressure. Electrical resistivity (Figure 3(a)) ofa typical Cu2Se1.01 sample is around 6–8 mO m at roomtemperature and increases quickly to about 55 mO m at700 1C. With Seebeck coefficient (Figure 3(b)) increasingfrom 75 to 250 mV K!1, these samples have moderate powerfactor (Figure 3(c)) of 750–950 mW m!1 K!1 at room tem-perature and peak around 1125–1250 mW m!1 K!1 at 600 1C.The temperature dependent Cp data is shown in Figure 3(d)where we could see a l-shape peak at around 140 1C as asymbol of the phase transition discussed above. Besides thisfeature, one may also notice that the Cp value is slightlydecreasing with the temperature at above 200 1C. Those b-phase Cu2Se samples also possess low total and lattice

Figure 1 Crystal structure of cubic b-phase Cu2Se. (a) View of (1 1 1) plane. (b) FCC unit cell with proposed possible Cu distributionat (0.25, 0.25, 0.25), (0.315, 0.315, 0.315), and (0.5, 0.5, 0.5) along /1 1 1S direction. (c) Room temperature XRD patterns of Cu2Senanopowders and bulk samples hot pressed at 400 1C (HP400 bulk), 500 1C (HP500 bulk), 600 1C (HP600 bulk), and 700 1C (HP700bulk). (d) Temperature dependent XRD patterns of hot pressed (700 1C) Cu2Se bulk sample measured at 25, 200, 400, and 600 1C.

Figure 2 Microstructure images of as-prepared (pressed at 700 1C) Cu2Se1.01 sample. (a) and (b) HRTEM images at roomtemperature, and 200 1C, respectively. (c) Typical SEM image taken from the same sample to show the grain size.

Thermoelectric properties of Cu2Se with ordered Se and disordered Cu layers 3










Thermoelectric properties of copper selenide with ordered selenium layer and disordered copper layer

Bo Yua,1, Weishu Liua,1, Shuo Chena, Hui Wanga, Hengzhi Wanga, Gang Chenb,n, Zhifeng Rena,n

properties Nano Energy (2012), doi:10.1016/j.nanoen.2012.02.010




















RAPID COMMUNICATION







Introduction






Nano Energy (]]]]) ], ]]]–]]]


Effect of pressing temp. on structure of Cu2Se properties

Nano Energy (2012), doi:10.1016/j.nanoen.2012.02.010












RAPID COMMUNICATION







Introduction






Nano Energy (]]]]) ], ]]]–]]]


Structure change observed by TEM properties



thermal conductivities (Figure 3(e)) as we initially expectedfrom their natural superlattice-like structures, where thelattice thermal conductivity is around 0.4 W m!1 K!1 and atotal thermal conductivity of less than 1 W m!1 K!1 at 700 1C.Overall, ZTs of "1.6 at 700 1C were obtained (Figure 3(f)),which are competitive to other traditional medium tempera-ture thermoelectric materials, such as skutterudites [19]and PbTe [20–22], but using abundant and environment-friendly elements. Due to the aforementioned phasetransition between tetragonal a-phase and cubicb-phase, a clear change could be observed in all the curvesof Figure 3 and this sudden change was found to happenat around 140 1C which is slightly higher than the reportedvalue of 130 1C [16]. Furthermore, one may also concludefrom Figure 3 that the thermoelectric properties of as-prepared Cu2Se1.01 are not sensitive to hot pressing tempera-ture (400–700 1C).

We also studied the composition effect on the thermo-electric properties of Cu2Se1+x by changing the amount of Se inthe initial compositions (Figure 4). For all these samples, weused 700 1C for hot pressing to make sure that the samples aremechanically strong as the thermoelectric properties areinsensitive to the hot pressing temperatures as shown inFigure 3. At any given temperature, both electrical resistivityand Seebeck coefficient (Figure 4(a) and (b)) decrease withhigher selenium content indicating an increased hole concen-tration due to more Cu vacancies. Accordingly, the CuSe1.02sample has the highest power factor (Figure 4(c)) due to itslowest electrical resistivity. Figure 4(d) shows the data oftemperature dependent thermal conductivity k while thelattice thermal conductivity klatt (Figure 4(e)) was calculatedby subtracting the carrier contribution kcarr from the total k.The Cu2Se1.02 sample shows not only the highest kcarr due tohigh carrier concentration, but also the highest klatt in this

Figure 3 Temperature dependent thermoelectric properties of Cu2Se1.01 bulk samples prepared with different hot pressingtemperatures. (a) Electrical resistivity. (b) Seebeck coefficient. (c) Power factor. (d) Specific heat (Cp), and thermal diffusivity(HP700 bulk). (e) Total thermal conductivity (filled symbols) and lattice thermal conductivity (open symbols). (f) Figure-of-merit, ZT.

B. Yu et al.4





B. Yu et al.4





B. Yu et al.4




RAPID COMMUNICATION







Introduction






Nano Energy (]]]]) ], ]]]–]]]


Thermoelectric properties of Cu2Se properties Nano Energy (2012), doi:10.1016/j.nanoen.2012.02.010





B. Yu et al.4


series. As for the ZT (Figure 4(f)), the Cu2Se1.01 sample showsthe highest peak value of !1.6 at 700 1C, while the highestaverage ZT appears in Cu2Se benefiting from its lowest latticethermal conductivity. From practical application point of view,Cu2Se is the preferred composition because of the higheraverage ZT.

Discussion

The good thermoelectric performance of b-phase Cu2Se is adirect result of its unique crystal structure as it possesses lowlattice thermal conductivity and good power factor at thesame time. The disordered Cu atoms at multiple latticepositions in the high temperature b-phase would be a highly

efficient phonon scattering mechanism, which is similar tothe role of Zn in Zn4Sb3 [23]. On the other hand, themonoatomic Se ordered layer may also introduce disturbanceto the phonon propagation. Besides the structure disorder,the abnormal decreasing Cp value at above 200 1C is alsoworth thinking. Normally, the Cp should approach a constantat high temperatures according to Dulong–Petit law or slightincrease with temperature due to the thermal expansion ofthe materials [24]; however, what we observed in ourexperiments is different: a slightly decreasing Cp withtemperature (Figure 3(d)), where similar phenomena werealso reported in Ag2S [25] and AgCrSe2 [26]. Anharmonicphonons usually lead to increasing specific heat withincreasing temperature although theoretically they can alsoreduce the specific heat [27,28]. In the other extreme, many

Figure 4 Temperature dependent thermoelectric properties of Cu2Se1+x with varying selenium content. (a) Electrical resistivity.(b) Seebeck coefficient. (c) Power factor. (d) Thermal conductivity. (e) Lattice thermal conductivity. (f) Figure-of-merit, ZT.





RAPID COMMUNICATION







Introduction






Nano Energy (]]]]) ], ]]]–]]]


ZT vs. temperature of Cu2Se properties



 

 

 

Summary for Cu2Se

• Good ZT may happen in non traditional thermoelectric materials

• Structure with ordered layer for charger carrier and disordered layer for phonon scattering is probably a good way to get high ZT

• Search of ZT higher than 2 should be in a lot of exotic materials


 

 

 

 

 

Acknowledgment • Thank you for your attention!

• DOE’s support, John Fairbanks and Carl Maronde

• Dr. Xiao Yan, Dr. Shuo Chen, Mr. Bo Yu, and Dr. Weishu Liu at BC

• Dr. Giri Joshi, Dr. Bed Poudel, Dr. Chris Caylor, and Dr. Jonathan D’Angelo at GMZ

• Prof. Gang Chen at MIT


enhancing the figure-of-merit in half-heuslers for vehicle ... · enhancing the figure-of-merit in...

Documents