Mechanical modeling of metal thin films on elastomers for femtosecond laser patterned interconnects

In this work, we present the mechanical analysis of thin Au/Ti film deposited on elastomer substrate, e.g. Polydimethylsiloxane (PDMS) comparing to samples on Kapton®. After deposition, the metal was patterned by ultrashort (270 femtosecond) pulse laser ablation. Minimizing the thermal damage to the substrate, reducing Heat Affected Zone (HAZ) compared to previously demonstrated laser patterning with longer pulses. One of the major issues in the metallization of thick elastomers is the formation of wrinkles followed by brittle cracks in the metallic thin film during the process, deteriorating the metal lines performance and reliability. Such cracking has not been observed on polyimide. The crack formation was investigated using stress analysis of the metallic-elastomeric system, assuming that cracking appears when the stress in the metal film is exceeding a critical value. A general mechanical model and analysis of a multilayer and multi-material laminated system is presented and applied to the specific case of Au/Ti on PDMS. Using this model, the root cause for the stress development in the metallic film has been investigated. A series of samples had been prepared using various elastomer substrates made of Sylgard 184, Dow Corning PDMS. A series of samples was prepared using different monomer/cross-linker ratios in the range of 1:2 to 1:19 monomer to cross-linker weight percentage ratio. Using the thin film stress analysis, the inner stress between and inside the different layers was calculated. Based on the calculations, we could estimate the specific stress in the metallic film. Using this value, we could choose the right specific materials that would not surpass their critical failure stress. This hypothesis was tested and verified by the failure analysis conducted by optical microscope and scanning electrons microscopy (SEM) secondary electrons (SE) images. We report here on a good correlation between the mechanical analysis predictions and the various samples integrity. Therefore, using that model, we could define a “safe zone”, i.e. for what elastic modulus and Coefficient of Thermal Expansion (CTE) there are no “mud-crack” defects.