The competing roles of i-ZnO in Cu(ln,Ga)Se¬2 solar cells

The electrical role of the highly resistive and transparent (HRT) i-ZnO layer in Cu(In, Ga)Se2(CIGS) solar cells is investigated. By tuning the resistivity of atomic layer deposited (ALD) i-ZnO through the use of post-growth O2-plasma treatments, it is shown that low i-ZnO carrier densities (i.e. high resistivities) actually restrict the performance of CIGS solar cells by reducing the extent of band-bending of the CdS/CIGS junction (the effect of series resistance is ruled out as the origin of any observed differences). This is the first evidence that i-ZnO has a negative electrical effect in CIGS solar cells (alongside the positive effect of shunt mitigation), and based on these results, attempts to maximise resistivity of the i-ZnO (typically sought-after for this HRT layer) are not recommended. Device efficiencies of 12.5% were obtained when using low resistivity as-grown ALD i-ZnO (resistivity, ρ=0.6 Ω cm, carrier density, n=3.5·1018 cm−3, and work function, Φ=4.06 eV), but this decreased to 11.5% when using high resistivity, plasma-treated ALD i-ZnO (ρ=134 Ω cm, n=0.2·1018 cm−3, and Φ=4.21 eV). SCAPS modelling revealed the reason for the difference to be the effect that the i-ZnO work function (controlled by carrier density) has on the band-bending and built-in voltage, Vbi, of the main junction. Capacitance-voltage experiments confirmed that the Vbi is lower (∆Vbi~0.1 V) when using low carrier density, high resistivity i-ZnO. This general effect was also found when using RF-sputtered i-ZnO, whereby the inclusion of high resistivity i-ZnO similarly generated lower efficiencies (15.0%) than low resistivity i-ZnO (15.9%).