Anion Vacancy as a Killer Center in Cu2ZnSnS(Se)4

The first-principles density functional theory study of deep level non-radiative recombination centre in kesterite solar cells.
12 Jul 2018/ EURODIM 2018, Poland

Transcript

1. Anion Vacancy as a Killer Center in Cu2 ZnSnS(Se)4 Sunghyun Kim, Ji-Sang Park and Aron Walsh Dept. of Materials, Imperial College London, UK sunghyun.kim@imperial.ac.uk | frssp.github.io | frssp | 0000-0001-5072-6801 EURODIM 2018 - POLAND
2. P. 2 Acknowledgement Prof. Aron Walsh Dr. Ji-Sang Park
3. P. 3 Photovoltaic: from Diamond (IV) to Kesterite (I2-II-IV-VI4)
4. P. 4 Open Circuit Voltage Deficit in CZTS Nat. Rev. Mater. 2, 17042 (2017) 2017 Research Cell Record by NREL Max. Efficiency 12.6% at 2014
5. P. 5 Short Carrier Lifetime Is Culprit In solar cells, the carrier lifetime is limited by trap-assisted (deep level) non-radiative recombination. Nat. Rev. Mater. 50, 797 (2018) It is important to identify defects with fast non-radiative recombination.
6. P. 6 Shockley-Read-Hall Process: Activation Energy A midgap level is the most efficient recombination center. But, not every midgap level is an efficient recombination center. Physical Review 87, 835 (1952) Physical Review 87, 387 (1952) Nat. Rev. Mater. 50, 797 (2018)
7. P. 7 Shockley-Read-Hall Process: Electronic Levels and Vibrational Properties ELECTRONIC ENERGY CONFIGURATION COORDINATE ELECTRONIC + ELASTIC ENERGY Valence band Conduction band D++e−+h+ D0+h+ D+ An oscillating electronic level captures a carrier. The electron loses energy by emitting phonons during the relaxation. Proc. R. Soc. Lond. A 204, 406 (1950) Phys. Rev. B 15, 989 (1977) To understand the nonradiative recombination in microscopic level, we need to know the electronic levels and local vibrational properties
8. P. 8 Atomistic Simulation Can Be Helpful First-principles calculations within the framework of density functional theory [1] +HSE06 functional [2] + VASP [3] + Finite-size correction ion e− e− ion e− e− e− e− Many Body Theory Density Functional Theory Electron density [1] Physical Review 140, A1133 (1965), [2] J. Chem. Phys,118, 8207 (2003), [3] Phys. Rev. B 59, 1758 (1999)
9. P. 9 Equilibrium Defect Concentration Concentration of a point defect that minimizes the Gibbs free energy of a crystal. ! = #$%&' exp(−Δ./012 (3% , 56 )/9: ;) Defect concentration Lattice sites Defect formation enthalpy Atomic chemical potential: Growth & annealing condition Fermi level: Carrier concentration (n, p)
10. P. 10 Atomic Chemical Potential and Phase Diagram The complex and narrow phase diagram implies that it is difficult to synthesize high-quality single-phase CZTS without secondary phase. Even in the single phase CZTS, the native point defects exist.
11. P. 11 V S -Cu Zn VS Cu Zn ZnCu SnZn Formation energy (eV) 0 1 2 Fermi level (eV) 0 0.5 1.0 1.5 E F (300K) S-poor Growth & Annealing Condition Low formation energy of CuZn (Acceptor) Defects with low formation energies: VS, VS-CuZn and SnZn Low formation energy of ZnCu(donor)
12. P. 12 Atomic Structure of VS Sn(IV): 5s05p0 (CZTS, SnS2) Sn(II): 5s25p0 (SnS) 1+ 2+ 4+ 1+ Conduction band e− Sn(III): 5s15p0 (?) e− ACS Energy Lett. 3, 496 (2018) 1+ 2+ 2+ 1+
13. P. 13 Electronic Structure of VS V S 2+ V S 0 Sn(II) 5s S 3p Sn(IV) 5s V S 0 V S 2+ CBM VBM (a) (b) (c) LUMO HOMO ACS Energy Lett. 3, 496 (2018) Strong e-ph coupling forms (bi)polaron.
14. P. 14 Deep VS State Induced by Double Sn Reduction VS is electrically benign neutral-scattering center ACS Energy Lett. 3, 496 (2018) Formation Energy Fermi level V" #$ V" % Bipolaronic V" % lattice relaxation Valence band Conduction band Donor level
15. P. 15 Optical Process: Photoexcited VS 2+ ACS Energy Lett. 3, 496 (2018) (a) 10 0 10 20 30 Q (amu½Å) 0 1 2 Energy (eV) (b) V1+ S V1+ S +h++e− V2 S ++e− E b n E abs V2+ S V1+ S V1+ S ħω CB VB Energy level diagram Configuration Coordinate Diagram ! = 3.9 × 10)*+cm. Sn(III)+h++e− Sn(IV)+e− Sn(III)
16. P. 16 VS Activated by Forming the VS -CuZn Complex Energy (eV) Energy (eV) 15 10 5 0 5 10 15 20 25 Q (amu1/2 Å) 0 1 2 (a) V S (+1/+2) (c) Sn (b) V S -Cu Zn (+1/+2) (d) Sn 0 1 2 15 10 5 0 5 10 15 20 25 Q (amu1/2Å) 0 1 2 0 1 2 Sn(III) Sn(IV)+e− Sn(III)+e−+h+ Sn(III) Sn(IV)+e− Sn(III)+e−+h+ Because of the Coulomb attraction between Sn and CuZn−, 1. The ground state configuration is Sn(III) rather than Sn(II). • Thermal excitation is not necessary. 2. The electron capture barrier decreases. • Optical excitation is not required. V S -Cu Zn VS Cu Zn ZnCu SnZn Formation energy (eV) 0 1 2 Fermi level (eV) 0 0.5 1.0 1.5 E F (300K)
17. P. 17 SnZn under S-rich Condition Similar to VS, the reduction and oxidation of Sn will trigger the recombination. V S -Cu Zn VS Cu Zn ZnCu Sn Zn Fermi level (eV) 0 0.5 1.0 1.5 E F (300K) V S 2+ V S 2+ Cu Zn 1- Sn Zn 1+ (V S -Cu Zn )1+ Sn Zn 1+ (b) (c)
18. P. 18 Minority carrier (Electron) Capture Cross Section The capture cross sections are way bigger than their atomic sizes. Concentration (cm−3) Capture barrier (meV) Capture cross section (cm2) (VS-CuZn)0 1017 ~ 1012 9 meV ~ 10−12 SnZn+2 1018 ~ 1014 54 meV ~ 10−13 SnZn+1 1012 ~ 108 184 meV ~ 10−14
19. P. 19 Experimental Evidence Deep Level Transient Spectroscopy DLTS shows the similar signature to that of SnZn(1+/0) while more measurements are required. 2016 IEEE 43rd PVSC, 2195-2199
20. P. 20 Experimental Evidence Phys. Status Solidi (c) 7, 1486 (2010) Electron Paramagnetic Resonance Landé factor g=2.006 EPR spectra of CZTS EPR of Sn3+ in Cubic ZnS Phys. Status Solidi B 111, 117 (1982) Sn(III) exists in CZTS in the form of VS, VS-CuZn or SnZn.
21. P. 21 Conclusion In quaternary I2-II-IV-VI4 materials, a precise control of atomic chemical potentials are required. • To suppress VS and VS-CuZn, a sulfurization process is essential. • To suppress SnZn, Zn-rich and Sn-poor condition is beneficial. The multivalency of Sn is the culprit to the short carrier lifetime in CZTS. ACS Energy Lett. 3, 496 (2018) https://speakerdeck.com/sunghyunkim/