Engineering topics
Publications and source records attributed to Wald, F. V..
Growing crystal surrounded by gas containing carbon or oxygen. Ribbon of solid silicon, edgewise, grows from pool of molten silicon in die. Gases flowing through orifice ensure longer diffusion length and less contaminiation by carbide particles in product.
Mackintosh et al. (1978) have reported on the development of a multiple ribbon furnace based on the 'edge defined film fed growth' (EFG) process for the fabrication of silicon ribbon. It has been demonstrated that this technology can meet the requirements for a silicon substrate material to be used in the manufacture of solar panels which can meet requirements regarding a selling price of $0.70/Wp when certain goals in terms of throughput and quality are achieved. These goals for the multiple ribbon technology using 10 cm wide ribbon require simultaneous growth of 12 ribbons by one operator at average speeds of 4 to 4.5 cm/min, and 13% efficient solar cells. A description is presented of the progress made toward achieving these goals. It is concluded that the required performance levels have now been achieved. The separate aspects of technology must now be integrated into a single prototype furnace.
We report here progress in a program designed to examine the feasibility of large scale production of low-cost silicon sheet substrates for solar cells by the EFG process. A multiple-ribbon EFG station, containing five single-ribbon cartridges, has been continuously operated for a period of 15 hours at a duty cycle of 94%. The average width of the ribbon grown was 5 cm, and the average growth speed was 3.4 cm/min. The cartridge concept of ribbon growth has been extended to successfully grow 10 cm wide ribbon at speeds up to 4 cm/min. Large area (approximately 50 sq cm) solar cells with efficiencies in the range from 8% to 11% (AM1) have been made from ribbon grown in the cartridge systems.
Displaced die concepts were explored along with some initial work on buckle characterization. Convective impurity redistribution was further studied. Growth from single cartridges was continued to create a quality baseline to allow comparison of the results with those in the upcoming multiple run and to choose the most appropriate die design. Fabrication and assembly work on the actual five ribbon furnace continued. Progress was made toward the development of the video optical system for edge position and meniscus height control. In preparation for a detailed program, designed to explore the buckling problem, ribbon guidance in the machine was improved. Buckle free, full width ribbon was grown under stable conditions without a cold shoe, an achievement essential to finally arrive at quantitative correlations between growth conditions and buckle formation.
Progress made in the development of EFG ribbon growth is discussed. Specific areas covered include: (1) demonstration of multiple growth for ribbons 5 cm wide in runs of 12 and 20 hours duration; (2) a single cartridge crystal growth station was built expanding observational capacity by virtue of an anamorphic optical-video system which allows close observation of the meniscus over 7.5 cm wide, as well as video taping of the ribbon growth process; (3) growth station no.1 achieved reproducible and reliable growth of 7.5 cm wide ribbon at speeds up to 4 cm/min; (4) introduction of the 'mini cold shoe'; (5) increases in cell efficiency due to interface shaping using the 'displaced die' concept; and (6) clarification of the role of gaseous impurities in cartridge furnaces and stabilization of their destabilizing influence on growth.
The edge-defined, film-fed growth (EFG) technique has been employed to grow silicon ribbons for photovoltaic applications. Considerable progress has been made in recent years in developing the technique to the point that long lengths of silicon ribbon can be routinely grown. In order to attain the full low-cost potential of the EFG technique, several further developments such as the growth of thinner and wider ribbons, increase in ribbon growth rate, and improvements in material quality are needed. The technological problems to be solved and the approaches employed to achieve these goals are discussed.
The background and progress to date of EFG ribbon growth for solar cell applications are briefly reviewed. The design and operation of a multiple ribbon-per-operator manufacturing system are then described. Results are presented of cost studies of this system at three stages of development. In the latter stage, believed to be attainable by 1986, conversion costs for polycrystalline silicon into wafers are attained which are consistent with manufacture of photovoltaic systems at $0.50 per peak watt. Principal technological problems in the ribbon growth process are discussed.
The enhancement of diffusion length with intensity was examined using spectral response measurements in solar cells based on Silso silicon and EFG silicon ribbon. Local diffusion length variations at defect sites were investigated as a function of illumination level using a scanning electron microscope operated in the electron beam induced current mode. An increase of diffusion length was observed at defect sites as the intensity level was increased. The diffusion length improvements are explained on the basis of saturation of minority carrier traps in these materials. The trap distribution in Silso silicon is shown to be peaked near (Ec - 0.7)eV. EFG silicon ribbon shows a broad, Gaussian distribution of traps located near (Ec - 0.5)eV.