A perovskite p-i-n architecture PV device generally adopts solution-processed organic hole transport layers (HTLs), like PTAA or self-assembled monolayers (SAMs). However, this approach can result in an inhomogeneous HTL surface coverage, especially when processed on textured substrates [1-2]. Recent reports have shown that the adoption of atomic layer deposited (ALD) nickel oxide (NiO) in combination with organic layers, such as PTAA or SAM, addresses the above-mentioned issue and leads to higher device yield, for both single junction [3] as well as tandem (in combination with c-Si or CIGS) devices [1-2]. Nevertheless, implementing NiO in devices without PTAA or SAM is seldom reported to lead to highly performing devices.

In the present contribution, we systematically assess the effect of key properties of NiO deemed relevant in literature, namely- resistivity and surface energy, on the device performance and compare the ALD NiO-based devices to those based on PTAA. To this purpose, (thermal) atomic layer deposited (ALD) NiO, Al-doped NiO, and plasma-assisted ALD (PA-ALD) NiO films are investigated as HTLs in a single junction mid-bandgap perovskite solar cell. The resistivity of Al-doped NiO and PA-ALD NiO films are 400 Ω∙cm and 80 Ω∙cm respectively and they are lower than that of thermal ALD NiO (10 kΩ∙cm). However, the devices implementing Al-doped and PA-ALD NiO HTLs exhibit only a modest V_OC gain of ~30 mV compared to thermal ALD NiO-based devices. Overall, the best-performing NiO-based devices (~14.8% PCE) still lag behind the PTAA-based devices (~17.5%) primarily due to a V_OC loss of ~100 mV. Moreover, we observe that the average grain size and the overall crystal quality of the perovskite absorber, which can impact the V_OC, is not affected by the surface energy of the different NiO HTLs. Further investigation based on the light intensity analysis of the V_OC and FF and the decrease in V_OC compared to the quasi-Fermi level splitting (QFLS), indicate that the V_OC is limited by trap-assisted recombination at the NiO/ perovskite interface and that a better charge extraction occurs when PTAA is adopted. Additionally, SCAPS simulations show that the V_OC of the NiO-based devices decreases when trap states are present at the NiO/ perovskite interface. Whilst tuning the resistivity of NiO has a negligible impact on the device performance, we also show that passivating the NiO/ perovskite interface with Me-4PACz SAM recovers this V_OC loss with an increase of ~200 mV. Our study shows that the potential positive effect of decreasing NiO bulk resistivity on the device performance is shadowed by the high recombination at the NiO/perovskite interface. Lastly, our work highlights the necessity of comparing devices based on emerging transport layers, such as NiO, with state-of-the-art transport layers-based devices, which is often neglected in the literature, in order to draw conclusion about the influence of specific material properties on the device performance.