Welded stainless steel packages offer a number of advantages relative to those fabricated from kovar or aluminum metals and braze sealed. They are highly resistant to corrosion, especially in aqueous or elevated temperature environments, are mechanically stronger, dramatically so at temperatures above 100°C, exhibit lower outgassing rates, and have better vibration characteristics. Since welded stainless steel packages do not require over plating to facilitate braze sealing, combined material and fabrication costs are lower than kovar or aluminum packages. Also, unlike kovar, stainless steel is nonmagnetic, which is advantageous in many electronic applications.
For sealing stainless steel packages, we elected to use seam sealing rather than laser welding for two reasons. Seam sealing subjects the package contents to lower thermal excursions than laser welding because less material is melted to achieve a weld in seam sealing. Second, seam sealing confines the molten material between the cover and package, where as laser welding produces a surface filet. Consequently, there is more opportunity for molten metal to splatter or react with the atmosphere in laser welding than in seam sealing.
We successfully developed a process to weld a 2 millimeter thick cover onto a box with dimensions of 75 mm long by 50 mm wide by 15 mm deep using a seam sealer. The box walls were 1 mm thick and were penetrated by a dozen glass insulated feedthrus on one side. Both cover and box were fabricated from 316L stainless steel. A combination of analytical and finite element modeling were used in conjunction with a designed experiment to optimize the process variables of roller angle, speed and pressure, weld current, pulse shape, duration and spacing, number of weld passes, and sealing atmosphere. Weld quality and seal integrity were evaluated by leak testing before and after environmental stressing, mechanical testing and metallographic cross sectioning. The effects of component dimensions and tolerances on seal integrity were also investigated. Particular attention was paid to cover flatness, flange thickness, and tightness of fit between the cover and box. The process development was concluded by conducting a qualification experiment that used the optimized process parameters with controlled variation about their nominal values. A 100% yield of sealed boxes was obtained. These test articles were then subjected to various environmental screening tests, which were all passed with no failures.