Computational Study of Hybrid Nanomaterial/Insulator/Silicon Solar Cells

Abstract

We present a computational study on hybrid nanomaterial-insulator-silicon solar cells where single-walled carbon nanotube or graphene forms the emitter as well as top conducting electrode on n-type crystalline silicon having a thin interfacial tunnel oxide. The effects of nanomaterial doping and tunnel oxide thickness on cell characteristics are modeled. Similar to bulk emitters, cell efficiency could be increased by chemical doping (p-type) of the nanomaterial. Unlike bulk, nanomaterial could get electrostatically doped (n-type) due to its low quantum capacitance, by the surface charge density in silicon. For chemically undoped graphene on lightly ( $10^{{16}}$ /cm $^{{3}}$ ) doped silicon, efficiency loss due to the electrostatic doping effect is $\sim 11$ %. A moderate p-type chemical doping (0.2 eV shift in Fermi level) of graphene reduces the aforementioned loss to $\sim 2$ %. The electrostatic doping effect in carbon nanotube-based cells is relatively small and independent to nanotube’s chemical doping. For tunnel oxide thickness $\ge 2$ nm, photogenerated carrier accumulation at silicon/oxide interface considerably enhances the electrostatic doping effect. The effect of tunnel oxide thickness variation on fill factor and open circuit voltage is shown to be qualitatively similar to standard bulk metal–insulator–silicon solar cells. Our model predicts an optimal oxide thickness of $\sim 1$ nm which confirms the experimental reports.