Assessing the viability of hydrogen-based perovskites for optoelectronic and thermoelectric applications via first principle modeling

In this study, we delve into the physical characteristics of six hydride perovskites of ABH₃-type materials (CsCaH₃, CsSrH₃, KMgH₃, LiBaH₃, NaBeH₃, and RbCaH₃). Our investigation primarily focuses on assessing their structural stability by determining the enthalpy of formation and examining the dispersion of phonons. Using band structure calculations, we discern the characteristics of semiconductors, observing a direct bandgap in all four perovskites except NaBeH₃ and KMgH₃, which exhibit indirect gaps. Among these, NaBeH₃ possesses the narrowest gap at 1.91 eV, while the widest gap is observed in the perovskite RbCaH₃, measuring 4.56 eV. Furthermore, we conduct a thorough analysis of their optical properties, including parameters such as the real and imaginary dielectric function, absorption coefficient, and refractive index within an energy range of 0 to 14 eV. The results of our study are highly encouraging, suggesting that these materials hold significant promise for utilization in photovoltaic cells. This is primarily attributed to their remarkable ability to absorb light across both the ultraviolet (UV) and visible spectra. Additionally, we conducted an assessment of the thermoelectric properties of the six perovskite materials. RbMgH₃ exhibits a maximum Seebeck coefficient (S_max) of 1.5 mV/K, whereas KMgH₃ achieves a figure of merit reaching unity. These findings present promising opportunities for utilizing these compounds in thermoelectric devices.

In this study, all self-consistent field (SCF) calculations were performed using density functional theory (DFT), employing the FP-LAPW + lo method as implemented in the Wien2k code. The Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation, the modified Becke-Johnson (mBJ) methods, and the HSE06 hybrid functional were employed to characterize the exchange–correlation interactions. Thermoelectric parameters were extracted using the BoltzTraP software.