Deposition Window Research Articles

Challenges and progress in the scale up of CuInSe2 thin film photovoltaic technology

The development of high-efficiency, low-cost thin film CuInSe2 (CIS) photovoltaic technology is progressing in laboratory research and engineering development efforts. In research laboratories, higher deposition temperatures, improved junction and window layers, and I–III–VI2 multinary absorbers have resulted in cell efficiencies approaching 15%. In engineering development efforts, low-cost processing techniques and durable commercial module packages have been demonstrated and large area (0.4 m2) prototype modules have been fabricated with power outputs that exceed 40 W and aperture area efficiencies that exceed 10%. The challenge facing engineering development now is to improve the processing yields on large area modules to a level which justifies a transition to pilot manufacturing. Recent work has focused on characterizing and improving the uniformity of CIS junctions and the reliability of monolithic circuit interconnects. Diagnostic techniques used to characterize the adhesion and uniformity of the individual cell layers and to spatially resolve the quality of devices and interconnects are discussed. Optical beam induced current measurements and scanning electron microscopy have been used to identify four unique defect types which stem from either defects in glass substrates or particulate contamination during fabrication.

Low pressure chemical vapor deposition of silicon dioxide below 500 °C by the pyrolysis of diethylsilane in oxygen

Low pressure chemical vapor deposition (LPCVD) of SiO2 in a horizontal LPCVD furnace using liquid diethylsilane and oxygen has been studied. A temperature deposition window ranging from 425–500 °C was observed resulting in a maximum deposition rate of ∼275 Å/min. The pressure dependence of the deposition rate revealed a threshold of ≳950 mTorr for gas phase reactions at a deposition temperature of 450 °C. Analysis of the films by Rutherford backscattering spectroscopy has indicated that as-deposited films are stoichiometric SiO2 for deposition temperatures ≤450 °C. Best case across wafer uniformity was ±5% for a caged boat. Wet chemical and reactive ion etch rates were found to be comparable to those of thermal oxides after annealing. Cross-sectional scanning electron microscopy images of the SiO2 films deposited on 2 μm deep 1 μm wide silicon trenches revealed a conformality of ∼80%. The electrical properties of films deposited at 450 °C were studied. The electrical properties of the films were studied as-deposited and after annealing the films in a cold-wall rapid thermal annealing (RTA) system. RTAs were performed at temperatures ranging from 950 to 1100 °C in Ar, N2, or O2 ambients. Current–voltage, current–temperature, and capacitance–voltage measurements were performed for the electrical characterization. Catastrophic breakdown field measurements have shown electric field strengths of 9.5 MV/cm for as-deposited 500 Å films. A study of the leakage current conduction mechanisms has indicated that as-deposited films exhibit trap conduction mechanisms at high electric fields and temperatures. However, if deposition is followed by a RTA in Ar or O2, the leakage current follows closely the Fowler–Nordheim mechanism and yields a leakage-current electric field dependence comparable to a thermal oxide. Results have shown that values as low as 6 × 1010/cm2 for fixed charge density can be obtained if oxide deposition is followed by a RTA in Ar or N2.

Molybdenum and tungsten–10% titanium deposition characterization

Refractory metal interconnects are of interest for very large scale integration to obtain reliable electromigration resistant conductors. GE has developed and characterized deposition processes for these films in a Varian 3180 cassette-to-cassette sputtering system. The base metal is tungsten containing 10 wt. % titanium. The bulk resistivity varies slightly (7%) with deposition pressure. The bulk resistivity minimum value has a deposition rate dependence as well. The functional form of RB seems to be independent of deposition temperature, while the absolute value depends very strongly on temperature. Bulk resistivity is a monotonically decreasing function with increasing temperature. All RB values were taken in the as-deposited condition and remeasured after a 400 °C forming gas anneal. Step coverage appears to be independent of deposition conditions in the range of operating parameters accessible to the 3180. Contact resistance is also a function of deposition temperature, roughly tracking the bulk resistivity. Stress in the TiW was in the mid 109 dyn/cm2 range, and was compressive in all cases. Stress versus deposition rate and temperature was measured. The stress appears to be independent of deposition pressure. Molybdenum deposition has been characterized as a function of deposition temperature. The bulk resistivity decreases strongly with increasing temperature, but shows almost no variation with pressure outside the measurement uncertainty. Stress increases with increasing deposition pressure. The bulk resistivity is apparently not a function of deposition rate. Step coverage of moly is equal to or slightly worse than aluminum step coverage. The optimum deposition window for aluminum step coverage is much narrower than that for moly. Any deposition condition we had access to in the 3180 had no effect on moly step coverage.