Impact of NaF during chloride treatment of CdTe solar cells

Sodium Fluoride Doping Approach to CdTe Solar Cells

“…Despite the large spread of values common for laboratory made solar cells with various processing steps, it is clear that the addition of NaF is not causing a drastic reduction in device efficiency as seen previously, 14,18 while improving the quality of the back contact. This suggests that the excessive recrystallization which proved detrimental to device performance in previous attempts at sodium incorporation in CdTe devices is avoided by depositing NaF after the chlorine treatment.…”

“…The EQE curves shown in Figure a are normalized to maximum collection efficiency for ease of comparison in Figure b and show very little difference in the CdS shoulder region at short wavelength. In previous attempts which have incorporated sodium into CdS/CdTe devices prior to the chlorine treatment, there has been significant recrystallization at the interface, ,, consuming the CdS layer which manifests itself as an increase of EQE in the short-wavelength region. Since there is no such effect observed here, it is likely that the CdS layer remains intact and therefore can create a suitable junction with CdTe which can effectively separate charge carriers.…”

“…The evaporation of a thin layer of NaF prior to contacting CdTe solar cells has been shown to be an effective strategy for intentionally incorporating sodium into the device structure without the adverse structural effects observed by previous authors. ,− However, the improvement in device performance has been modest even for optimized treatment conditions. This is likely due to several overlapping processes occurring during the evaporation of NaF and subsequent annealing of devices, such as oxidation of the back contact and sodium accumulation in the CdS layer.…”

“…Here we report on the incorporation of sodium into CdS/CdTe devices by thermally evaporating a thin layer of NaF prior to contacting. By depositing NaF after the chloride treatment, the excessive recrystallization which has been problematic in previously reported attempts is avoided. − In this work, this has improved the back contact as shown by the absence of “rollover” in current density–voltage ( J – V ) curves, implying a more highly doped region localized at the back surface. Annealing these devices in air at temperatures greater than 300 °C causes sodium to diffuse into the bulk CdTe region, whereby a small increase in acceptor concentration is measured.…”

“…While an increased doping density was measured to ∼10 15 cm –3 upon sodium inclusion, this also has deleterious effects on film morphology. This includes agglomeration of the CdS window layer, leading to a poor diode response, and widening of CdTe grain boundaries, thereby shorting contacts, which has so far limited study at a device level. − …”