An Enhanced Charge Carrier Separation in a Heterojunction Solar Cell with a Metal Oxide

Abstract

A model of charge carrier transport in heterojunction solar cells composed of a solar light‐absorbing semiconductor and a wide‐bandgap semiconducting metal oxide is proposed. It describes an electric field in the semiconductor originating from the difference in the electron work functions between the two contacting materials, an enhanced hole separation by this field, and subsequent trap‐assisted tunneling of holes through the semiconducting oxide. The model predicts a dramatic influence of the donor concentration in the semiconductor, trap parameters in the oxide, and ideality factor of the semiconductor/oxide heterojunction on the performance of the cell. The efficiency of the solar energy harvesting by MoO x /n‐Si solar cells is calculated to reach 23% of the donor concentration in Si of 1018 cm−3 at a limited hole current density. It is predicted to be reduced to 16−18% when donor concentration is in the range of 1015−1016 cm−3. The numerical predictions agree well with experimental data, at least for donor concentrations in Si below 1018 cm−3, confirming suitability of the model for heterojunction solar cells composed of other semiconductors and semiconducting metal oxides.

Hole transport through an isotype amorphous-crystalline Ge2Sb2Te5 heterojunction under forward bias

“…The first exception occurs across the QNR of the a-GST layer; though it is subtle in figure 7(b), there is a slight curvature in the simulated E Fp across the a-GST QNR. Consequently, the hole current density becomes increasingly negative with increasing distance from the c-GST/a-GST interface, i.e., from −40.6 mA cm −2 (at x = 20 nm) to −41.6 mA cm −2 (at x = 150 nm) at 0.15 V and from −1278.2 mA cm −2 (at x = 20 nm) to −1331.8 mA cm −2 (at x = 150 nm) at 0.4 V. Using equations (13a) and (14), this implies that the net recombination rate is non-negligible. Hence, the approximate model does not capture the effect of the considerable recombination on the E Fp within the QNR at larger biases, given the assumption made previously in section 3.3 that the net recombination rate within the QNR can be neglected.…”

“…Therefore, the a-GST/c-GST heterojunction could find application in similar devices with these characteristic properties. After all, even heterojunctions between different material systems such as MoO x /c-Si [13,14] and a-Si/SiO 2 /c-Si [15], which are both amorphous-crystalline and isotype, exhibit both passivation of defects at the amorphous-crystalline interface, as well as selectivity of carriers, for application in heterojunction solar cells with enhanced efficiency.…”