研究目的
Investigating the effects of dielectric spin coating (DSC) on surface variations over existing underlying topography in redistribution layer (RDL) fabrication and developing fill patterns to limit these variations.
研究成果
The proposed model accurately predicts post-cure spin coating thicknesses over arbitrary topographies, and fill patterns effectively limit surface variations. The study demonstrates that constructing RDL layouts using standard fill cells can substantially reduce surface variations after coating.
研究不足
The model under-predicts feature scale variations for interconnects with heights close to or greater than the nominal polyimide thicknesses, especially when line widths are small and spacings are large.
1：Experimental Design and Method Selection:
The study involves designing test structures representing a wide range of underlying feature widths, spacings, and heights, coating them with polyimide, and profiling their surfaces. An empirical model based on spatial filtering over underlying features is developed.
2：Sample Selection and Data Sources:
Test structures are fabricated to represent topographies common in RDLs, covering a wide range of interconnect widths and spacings, and fabricated with a range of different heights.
3：List of Experimental Equipment and Materials:
Polyimide HD-4110 and HD-4100 are used for coating. Profilometry is performed using a Dektak 150 with a 2 μm tip radius and a Bruker Dektak XT profilometer for 2D profiling.
4：Experimental Procedures and Operational Workflow:
TVs are coated with polyimide at specified thicknesses, soft baked, and cured. Surface profiles are measured and analyzed to develop and validate the model.
5：Data Analysis Methods:
The model's predictions are compared to experimental results, with RMS error calculated to validate the model's accuracy.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
