Defect screening of complex oxides for high ZT thermoelectrics

Contributed presentation at the Materials Research Society Fall 2017 Conference.

Transcript

1. E-mail: alex.ganose.10@ucl.ac.uk @alexganose Defect screening of complex oxides for high ZT thermoelectrics Alex Ganose, Winnie Leung, Adam Jackson, Robert Palgrave and David Scanlon Department of Chemistry, University College London Diamond Light Source Ltd. &
2. Convert waste heat to useable energy: α = Seebeck coefficient σ = conductivity T = temperature = lattice thermal conductivity = electrical thermal conductivity Thermoelectric Materials Snyder and Toberer, Nat. Mater. (2008), 7, 105 ZT = 2 +
3. ZT Optimisation ZT optimisation is difficult & lengthy process Kanatzidis and coworkers, Chem. Rev. (2016), 116, 12123–12149
4. Oxide Thermoelectrics Pros: – Cheap, earth-abundant, non-toxic, high stability Limitations: – High lattice thermal conductivity limits efficiencies E.g. ZnO:
5. Chemical Intuition What makes ZnO, SnO2 , In2 O3 etc such great TCOs? Overlap of metal s states with O 2p Can we find alternatives? Post transition metal (V) oxides: – Cheaper than In & Sn – Large EAs & small me * – Heavy elements help lower κ latt
6. Candidate Material: BaBi2 O6 BaBi2 O6 structure: – Lead antimonite PbSb2 O6 -type structure (Pത 31m) – Layers of edge-sharing BiO6 octahedra – Layered nature and heavy Ba and Bi should reduce thermal conductivity
7. BaBi2 O6 Electronic Structure Band gap 2.80 eV (HSE06), experiment ~2.6 eV Dispersive CBM (Bi s states) gives low effective masses: – 0.37 me in layers, 0.67 me across layers Ideal electronic structure for n-type material
8. BaBi2 O6 Vibrational Properties Phonons indicate dynamic stability κ latt very low compared to other n-type TEs: – ZnO (15 Wm–1K–1) & SrTiO3 (11 Wm–1K–1) Lowest in z-direction due to layered-nature
9. BaBi2 O6 Defect Chemistry Dopability is essential!
10. BaBi2 O6 Thermoelectric Performance Reasonable ZTs: – ZT ~ 0.1 at 450 K – ZT ~ 0.2 at 600 K – Doping levels are realistic using La as dopant Better performance than state-of-the art SrTiO3 (0.15 at 600 K) Limited by relatively small temperature stability range Rosseinsky and coworkers, Ener. Environ. Sci. (2017), 10, 1917–1922
11. Candidate Material: ZnSb2 O6 ZnSb2 O6 structure: – Ordered trirutile structure (P42 /mnm) – Proposed as TCO in 2005 (Hosono group) and 2014 (Hautier group) – Sb5+ more resistant to reduction than Bi5+ – Stable to 1400 K!
12. ZnSb2 O6 Electronic Structure Band gap 3.52 eV (PBE0), experiment ~3.5 eV Small electron effective masses: 0.25 me – Due to CBM made from O p with Zn s and Sb s Another ideal electronic structure for n-type material
13. ZnSb2 O6 Vibrational Properties Phonons – confirm dynamic stability κ latt remains low despite 3D connectivity! Smaller than ZnO but larger than BaBi2 O6 Analysis reveals nano-structuring will be effective
14. Nominally undoped ZnSb2 O6 not degenerate: ZnSb2 O6 Defect Chemistry Carrier concentration: 5 x 1016 cm–3
15. Fluorine doped: ZnSb2 O6 Defect Chemistry Carrier concentration: 1 x 1018 cm–3
16. Aluminium doped: ZnSb2 O6 Defect Chemistry Carrier concentration: 2 x 1020 cm–3
17. ZnSb2 O6 Thermoelectric Performance High ZTs: – ZT ~ 0.26 at 1000 K – ZT ~ 0.65 at 1400 K Doping levels realistic: 4 x 1019 cm–3 Better performance than any n-type TE: – In2 O3 :Ge ~ 0.48 at 1300 K – Zn0.98 Al0.02 O ~ 0.30 at 1300 K Maignan and coworkers, J. Phys. Condens. Matter (2016), 28, 013001
18. Conclusions Bi5+ and Sb5+ oxides produce efficient n-type materials: – Good dispersion similar to other (n –1)d10ns0np0 oxides – Large fundamental & optical band gaps BaBi2 O6 : – Low κ latt due to layered nature; dopable with La – ZT comparable with SrTiO3 at low temperature ZnSb2 O6 : – κ latt remains low & Al enables high doping concentration – Highest predicted ZT of any n-type oxide TE
19. Acknowledgements People: • Dr Adam Jackson, Winnie Leung (SMTG) • Dr Robert Palgrave, Dr David Scanlon (UCL) • Dr Jon Alaria (UoL) Compute Resources: • Legion & Grace (UCL) • Archer (EPSRC)