The performance analysis of near-field thermophotovoltaics considering temperature dependence of indium tin oxide emitter

Near-field thermophotovoltaics (NFTPV) is of great interest because it can achieve both high power density and high efficiency at medium to high temperature. One key to efficient photovoltaic conversion is an emitter that can selectively emit spectrum. Indium tin oxide (ITO) is considered as the emitter for NFTPV because its plasma frequency is tunable by simple technical means. However, as a typical degenerate semiconductor, the temperature dependence of its dielectric constant is often neglected in studying the performance of NFTPV. Therefore, we propose a NFTPV model considering the temperature-dependent property of ITO. By changing the applied voltage V, the vacuum gap d, and the emitter temperature T1, the performance of NFTPV with and without temperature dependence is compared and shows a significant difference. We find that the heat flux is always underestimated as well as the optimal efficiency is overestimated when the temperature dependence is not considered. At d = 10 nm and T1 = 1000 K, the optimal efficiency is even overestimated by about 270%. Without considering the temperature dependence, the output power is overestimated at d < 20 nm, while the opposite is true at d >= 20 nm. We clarify the main reasons for these misjudgments at a small gap originate from the underestimation of the intensity of surface phonon polaritons (SPhPs) below the bandgap and the overestimation of the intensity of surface plasmon polaritons (SPPs) above the bandgap. The main factor affecting the two kinds of surface modes in near-field regime is that high temperatures lead to low electromagnetic local density of states (LDOS) for ITO emitter near the SPPs resonant frequency, but high LDOS in the ultralow frequencies. In a larger gap, it is attributed to the underestimation of the waveguide modes above the bandgap and the nearly-full-band propagation modes. Our work illustrates the necessity of considering temperature dependence, enriches the understanding of NFTPV, and facilitates the development of related applications.