Multiphysical modeling of nanosecond laser dicing on ultra-thin silicon wafers

We propose an approach for the numerical modeling of a laser ablation (LA) process on silicon targets. The work is motivated by the increasing application of lasers in the separation of ultra-thin power semiconductors. In order to optimize the process, reduce the energy cost per laser pulse and minimize the extension of the thermally induced damage, a deeper insight into the mechanisms underlying laser dicing and a proper selection of laser settings are crucial. Numerical modeling is useful for understanding the tightly coupled physics involved in the interaction of laser with matter, as well as in the identification of the optimum laser configuration. With this aim, two numerical models have been prepared and combined. Initially, we set up a custom written one-dimensional hydrodynamic code which describes the main mechanisms triggered during LA, as vaporization and plasma formation. This first simulation allows to estimate the laser energy loss due to plasma absorption. The remaining available energy is used as input to perform a Finite Element transient thermal simulation on a three-dimensional geometry of the target. Here, an element deactivation technique is adopted to remove the vaporized elements from the computational mesh, therefore describing the geometry and the progressive formation of the ablated crater. The calculated crater geometries have been compared with experimental ones for two fluence values, showing reasonable agreement.