Figure 1. (a) Schematic diagram of the TLM test structure; (b) Dependence of the ρ c on ZnTiO3 thickness and annealing temperature after metallization; (c) ρ cof ZnTiO3-based contacts and other well-known Ohmic contacts as a function of annealing temperature; (d) Schematic of the c-Si solar cell. SiO2/ZnTiO3/Al is deposited as the electron-selective contacts; (e-h)V OC, J SC, FF and PCE of the c-Si solar cells with different thickness of ZnTiO3 under different annealing temperatures in air ambient for 30 min; (i) The normalized PCE of different dopant-free electron-selective contacts as a function of annealing temperature.
Figure 1(i) shows the thermal stability of c-Si solar cells with various dopant-free electron-selective contacts, including the ZnTiO3-based contacts in this work. The PCE at each temperature step are normalized by their initial value (before annealing). The solar cells with TiO2/Al,a-Si/LiFx/Al or a-Si/LiFx/Ti/Al as electron-selective contacts suffers relatively severe degradation.30 The cells with Al-doped TiO2(ATO)/LiFx/Al,17a-Si/CaAcac/Al,28 TiN/Al23 or ZnSe41 contacts present greatly improved stability. Uniquely, the solar cells with ZnTiO3-based contacts show increased PCE with increasing post-metallization annealing temperature up to 300 °C. Hence unlike other dopant-free contacts, the ZnTiO3-based contacts embrace thermal treatment, which is usually unavoidable in device fabrication. Moreover, it is worth pointing out that the annealing time at each temperature is generally within 5 minutes for those works in literature, but the annealing time is 30 minutes in this work. This means that the ZnTiO3-based electron-selective contacts is compatible with the silicon heterojunction (SHJ) production lines (the curing temperature after screen-printing electrodes is between 200-300 °C and the time is 20-30 minutes).
Besides J -V characteristics, the EQE spectra were also examined at each annealing step, as shown in Figure 2a. The results show that the EQE gradually increases at long wavelengths (700-1100 nm) from 25-300 °C, which soundly explains the improvement of J SC. In addition, considering that the improvement of EQE only happens in long wavelength and the fact that the optical reflectance is unchanged (see Figure S1), we can conclude that the electron selectivity of ZnTiO3-based contacts is enhanced with increasing post-metallization annealing temperature (within 300 °C). 𝜌c and interface passivation are the two factors affecting carrier selectivity. Because 𝜌c is nearly unchanged, at least which is true for the ZnTiO3thickness not greater than 3nm, thus the enhanced performance (V OC, J SC and PCE) of the solar cells mainly originates from the improved interface passivation at the back side. This conclusion is further examined by PL mapping andτ eff characterization. Figure 2b compares the PL mappings of an as-fabricated solar cell and a 300 °C post-annealed solar cell. The PL intensity of the annealed solar cell is much stronger than that of the as-fabricated one, indicating the greatly reduced nonradiative carrier recombination. Surface passivation can also be evaluated by τ eff, which is often obtained by utilizing photoconductance decay (PCD) method. However, this method is unsuitable to characterize our samples because Al capping layer plays an important role in the ZnTiO3-based electron-selective contacts. Recently, Liang also found that fully metalized samples are not allowed the minority carrier lifetime to be measured with the PCD technique.43 Therefore, here we characterizeτ eff in final device level by Suns-V oc measurement mode.44 Figure 2c shows that theτ eff increases with post-annealing temperature (within 300 °C), again demonstrating that the passivation effect is improved by post-annealing treatment. Benefiting from the enhanced passivation effect in rear side, the optimal ZnTiO3/c-Si heterojunction solar cell, which is annealed at 300 °C, possesses a PCE of 22.0% with a V OC of 662 mV,J SC of 40.1 mA cm−2 andFF of 82.7%, as shown in Figure 2d. The PCE is impressive and is comparable with that of conventional c-Si solar cells in current PV industry. Moreover, it is worth mentioning that capital intensive equipment such as PECVD that used for growing a-Si:H films can be avoided.