1714
Advanced Process Control of Nickel Electrodeposition for Packaging in Semiconductor Industry
The inherently low current efficiency of Ni electrodeposition leads to pitting due to hydrogen bubbles. Tools with horizontal wafer orientation can usually spin off bubbles by fast wafer rotation. Tools with vertical wafer orientation typically require the use of organic surfactants to reduce surface tension and promote the release of bubbles from the surface. Surfactant is removed from the process by drag-out and foaming which requires tight control. Not enough surfactant increases the potential for pitting, while too much can cause inclusion of organics into the deposit which can result in defects and reliability failures by a burn out during heat cycling [2]. One process control option is to measure surface tension. However, this property by itself is not highly selective. The same concentration of surfactant can cause different surface tensions depending on the concentrations of other components. Authors has developed a highly selective and accurate potentiometric titration method to address this process demand. Details of the method will be elaborated in the presentation.
Type I Brighteners are typically used for nickel electro-deposition to fine-tune the stress level by incorporating sulfur-bearing compounds into the deposit. The process window for these chemicals is narrower than for surfactants with the potential for out-of- spec related to stress level, deposit purity, and reliability under heat cycling. Traditional metrology for brighteners employs long and hazardous methods such as polarography and HPLC.. A non-reagent, real-time, spectral analytical method was developed by ECI for this application. Sulfamate ion is known to decompose at the anode producing azodisulfonate which itself behaves as a “Type I Brightener” [3]. However, the generation of this compound depends on the type of anodes and whether a membrane is used to separate anode and cathode compartments. Figure 1 shows spectra of the same process solution after operation in different process tools. Acceptable deposition was obtained only in the process tool where azodisulfonate products were detected in the solution. It can be argued that use of these products as an additive is better than lengthy burn-ins to generate these products.
The truly unique challenge of nickel deposition for semiconductor packaging is related to the presence of a lithographic photoresist mask on the surface. Ideal positive type photoresist should be fully resistant to solutions in unexposed form but quickly dissolves in developer following UV exposure. In reality, unexposed photoresist is slowly attacked by the hot nickel solution. The presence of photoresist has detrimental effects on the performance of Ni baths. The concentration of leached photoresist is a main limiting factor for electrolyte lifetime. Thus, its metrology becomes crucial to bath dump and/or bleed and feed adjustment. A non-reagent spectroscopy method has been developed by ECI to monitor the accumulation of photoresist (Figure 2). The presentation will show some examples of this method.
In conclusion, a comprehensive array of metrology techniques was developed to monitor surfactant, stress controlling compound, and leached photoresist in the nickel electrodeposition processes used in the semiconductor industry.
References:
- M.Paunovic, P.J.Bailey, R.G. Schad, D.A. Smith J.Electrochem.Soc, 141 (7), 1843 (1994).
- L.M.Wang. J.Electrochem.Soc, 156 (6), D204 (2009)
- O.J. Klingenmaier, Plating, 52, 1138 (1965)
