|Institution:||Colorado School of Mines|
|Keywords:||Cadmium telluride; Solar cell; Thin film|
|Full text PDF:||http://hdl.handle.net/11124/170330|
Cadmium telluride-based photovoltaics are presently the leading commercialized thin film solar cell technology. A critical step during device fabrication is exposure of the as-deposited CdTe absorber to CdCl2 vapor at ~400 ºC for 10 – 30 min. This step is critical for obtaining high power conversion efficiencies and has been correlated with recrystallization and grain growth, passivation of grain boundaries, and heterojunction interdiffusion. This activation step is difficult to control and capex intensive. In addition, the use of CdCl2 presents a significant environmental concern due to its high solubility in water. In this thesis we explored molecular chlorine (Cl2) as an alternative to CdCl2 for activation of CdTe solar cells. This activation treatment could potentially serve as a direct replacement for CdCl2 processing in commercial module manufacturing. Molecular Cl2 diluted in N2/O2 mixtures was exposed to as-deposited FTO/CdS/CdTe stacks in a custom-built quartz tube furnace, enabling independent control of Cl2/Ar flowrate (1 – 5 sccm), temperature (375 – 420 ºC), oxygen concentration (0 – 25 %), and treatment time (0 – 10 min). The (111) texture coefficient was tracked during experiments to measure extents of recrystallization as parameters were explored, followed by correlations with device performance. Within the parameters explored it was found that a threshold temperature of T ≥ 400 ºC was required for recrystallization of the films. This coincides with the typical optimum temperature used for CdCl2 activation. In stark contrast to CdCl2 activation, very short treatment times (~1 min) and no oxygen ambient were needed for optimal device performance. In addition, the Cl2 process delivered significantly improved uniformity and run to run reproducibility relative to conventional CdCl2 activation. The Cl2 process also displayed a much larger process window with respect to temperature and eliminated delamination. A final optimization of 1.5 sccm of 3% Cl2/Ar, T = 415 ºC, 800 sccm N2, and 2 min resulted in an 11.6% efficient device with an open circuit voltage > 800 mV. This value surpasses the best reports in the literature (9-10 %) and exceeds that of many alternative activation agents (HCl, NaCl, etc.). Best device performance was less than the 14.5% baseline efficiency achieved with standard CdCl2 treatment. Device characterization revealed losses in both Voc (~50 mV) and Jsc (2-3 mA/cm2). Quantum efficiency showed that collection in the bulk CdTe region was 10% lower than standard CdCl2 activation, suggesting defect-based bulk recombination. Transmission electron microscopy and elemental mapping revealed that grain coalescence was incomplete under Cl2 activation, leaving regions with voids or low density. Both chlorine and oxygen accumulate at these defects, explaining the substandard performance. The optimal chlorine exposure was the lowest that could be reliably delivered, and it is recommended that further studies reduce the Cl2 concentration by an order of magnitude or more. In addition, longer… Advisors/Committee Members: Wolden, Colin Andrew (advisor), Ohno, Timothy R. (committee member), Agarwal, Sumit (committee member).