Synergistic defect engineering for improving n-type NbFeSb thermoelectric performance through high-throughput computations

Abstract

The p-type NbFeSb half-Heusler compound has been proved to be a promising high-temperature thermoelectric material, and many works have been devoted to improving its properties. However, its corresponding n-type compound shows a fairly low ZT, which greatly hampers its realistic applications due to the asymmetrical performance in thermoelectric devices. In this work, we compute in a high-throughput manner a large amount of (intrinsic and extrinsic) defects in NbFeSb, and systematically investigate their effects on the n-type thermoelectric performance (electrical and thermal properties). For electrical properties, the intrinsic defects of Fe_Nb antisites and Fe_i interstitials have low formation energies, and their donor behavior is favorable for n-type conductivity. The extrinsic defects (such as Co_Nb and Pt_i), on the other hand, introducing strong resonant states slightly above the conduction band maximum, can greatly enhance the power factor (PF) of n-type: noticeably 2–3 times higher than that of the intrinsic NbFeSb compound. For thermal properties, the low-energy extrinsic defects of F/Cl/P/S substitution at Sb sites can significantly reduce the lattice thermal conductivity (lowered to 1.25 W m⁻¹ K⁻¹ at 300 K) due to the large-disorder scatterings. Unfortunately, none of the defects simultaneously satisfy the three criteria for promising n-type NbFeSb thermoelectrics (having low formation energy, showing strong resonant states and depressing the lattice thermal conductivity). Therefore, we propose a synergistic defect strategy for those low-energy defects: some (Co_Nb or Pt_i) boosting the PF and others (F_Sb, Cl_Sb, S_Sb or P_Sb) suppressing the thermal conductivity. Our work benefits future experimental defect engineering of NbFeSb-based materials and paves the way to narrow the gap between the n- and p-type thermoelectric properties.