Here, we report on the first copper-doped nickel oxide films deposited by ALD and the application of ALD-NiO as an electron-blocker in perovskite solar cells. First, a nickel amidinate precursor (Bis(N,N'-di-t-butylacetamidinato)nickel(II)) was tested with various oxidizers at temperatures from 150 to 200°C. Only ozone was shown to produce near-stoichiometric NiO, with hydroxide content decreasing as process temperature increases. Observed growth rates varied, but were in the range of 0.4 – 1.0 Å/cycle. A copper oxide (CuO) ALD process, based on Cu(dmap)2 (Bis(dimethylamino-2-propoxy)copper(II)) and ozone, was also developed, and co-ALD of the two precursors was used to deposit CuxNi1-xO. UV-ozone-treated ALD-NiO films were also fabricated. 20-30 nm films were analyzed in detail by X-ray photoelectron spectroscopy and X-ray diffraction. At low concentrations (on the order of 5-10%), the copper is incorporated into the polycrystalline NiO lattice as Cu2+, as shown in Figure 1a, while at high concentrations (30%) the CuO precipitates out, as shown in Figure 1b. Temperature-dependent electrical measurements (Figure 1c) showed that copper doping decreases the lateral resistivity, from 5.2⋅104 Ohm-cm at room temperature in undoped NiO to 5.4⋅103 Ohm-cm for Cu0.1Ni0.9O. Activation energy differences between doped (0.34 eV) and undoped (0.28 - 0.29 eV) samples imply that either the conduction mechanisms are different or that copper inclusion significantly reduces dipole screening in the ALD films (see Austin & Mott (2001), “Polarons in crystalline and non-crystalline materials,” Advances in Physics). Ultraviolet photoelectron spectroscopy (UPS) and optical spectroscopy revealed that copper doping increased the NiO:Cu work function by 170 meV while decreasing the optical bandgap and leaving the ionization energy unchanged. Ozone treatment of the NiO film reduced the resistivity to 2.4⋅103 Ohm-cm. It also deepened the work function by 700 meV and the ionization energy by approximately 130 meV.
ALD-NiO deposited at 200°C with ozone as oxidizer was also tested as an electron-blocking layer in ITO/NiO/perovskite/PCBM/BCP solar cells. Compared to a similar device made with solution-deposited NiO, the ALD-NiO device showed an approximately 20 meV VOC improvement, suggesting reduced recombination and/or better electron blocking. The fill-factor and short-circuit-current were both slightly smaller, as shown in Figure 1d, most likely in part because of the high vertical resistivity of the ALD-NiO film or because of surface layers inhibiting charge transfer or current flow. Future work will focus on incorporating ALD-NiO:Cu into PV designs, optimizing film processing conditions, and investigating activation and conduction mechanism in the ALD-NiO:Cu films themselves.
Fig. 1 Caption: a) Cu2p XPS Spectrum for a 25%-Cu-by-cycle ALD-NiO:Cu film, and reference CuO spectrum. The close fit indicates the copper is being incorporated as Cu2+. b) GIXRD data of NiO and NiO:Cu films. The NiO has a polycrystalline structure, with no CuO peaks visible at lower concentrations. At 30%-Cu NiO:Cu, the CuO precipitates out and CuO peaks are visible. c) Temperature-dependent conductivities for NiO, NiO:Cu, and UV-ozone-treated NiO, along with Arrhenius fits. Both UV-ozone treatment and copper doping increase the conductivity, but the differing activation energies point to some difference in transport mechanism or film dielectric environment. d) Experimental JV curves for ITO/NiO/CH3NH3PbI3/PCBM/BCP/Ag solar cells, using both ALD-NiO and solution-deposited NiO.
