The Ohno process was applied to produce Al-lwt.% Si alloy rods for semiconductor devices. The casting conditions which enable production of an 8mm diameter rod having a mirror surface with a highly longitudinally oriented microstmcture was established. It was demonstrated that the alloy could be drawn, without the need of intermediate recrystallization or stress relief, to produce wires of 0.14 to 0.10mm in diameter. The relationships between the casting conditions and wire breakage during fine wire drawing were also studied.Keywords : Ohno process, oriented microstmcture, fine wire drawing
1, INTRODUCTIONPrevious studies have demonstrated the technology of continuous casting metals using a procedure in which the solidliquid interface is maintained outside of the mold [1][2][3][4]. This results in a casting having a mirror surface (since it solidifies in situ without mold contact) and directionally solidified microstructure (since the heat of fusion is extracted through the solidified casting). This study was designed to investigate the feasibility of casting an A1-lwt.% Si alloy using the technique, and to establish the operating parameters that would ensure the satisfactory production of an 8 mm diameter rod. Furthermore, it is envisaged that a highly oriented, finely distribnted microstructure produced using this casting technology would be capable of being cold worked easily [1, 5-61. High quality cast rods of the Al-lwt.% Si alloy have been produced in this manner and drawn to wires as fine as 20 p.m in diameter for semiconductor devices. A second objective of this study was to establish the relationships between the casting conditions and wire drawing conditions.
EXPERIMENTAL PROCEDUREA schematic diagram of the experimental apparatus for horizontal continuous casting using a heated mold is presented in Fig. 1. The Al-lwt.%Si alloy was vacuum (10 -~ Torr) melted in a graphite crucible (20 kg capaci~) in a graphite resistance furnace after which pure argon was introduced into the chamber and held at 1 atmosphere. Chromel-alumel thermocouples were used with a PID (proportional integral derivative) controller to measure and control both the molten metal temperature and the mold temperature. The melt level was maintained using a ballast (level control bar) which ensured a uniform metal head and uniform flow of molten metal into the heated mold. The mold temperature was measured 2ram away from the mold exit and maintained within _+2~ using the PID controller.The casting process was initiated by inserting a stainless steel dummy bar into the mold so that the alloy would solidify against the dummy bar. The dummy bar was later extracted using three sets of pinch rolls. The bar was supported by a graphite guide block on which cooling water was sprayed. Preliminary experiments indicated that for a given range of water flow rate (0.2 to 0,6 l/min) and mold temperature (670 to 750"C) the CM distance (distance between the cooler and the mold) was a significant factor affecting the casting rate. Setting the water flow rat...