Process control strategies for the gas phase synthesis of silicon nanoparticles

In this contribution the identification of new reaction conditions for the production of nearly monodisperse silicon nanoparticles via the pyrolysis of monosilane in a hot wall reactor is considered. For this purpose a full finite volume model has been combined with a state-of-the-art trust-region optimisation algorithm for process control. Verified against experimental data, specific process conditions are determined accomplishing a versatile range of prescribed product properties. The main achievement of the optimisation is the possibility to control the different mechanisms in the particle formation process by mainly adjusting the temperature profile. Due to a successful separation of the nucleation and growth process, significantly narrower particle size distributions are obtained. Moreover, the presented optimisation framework establishes rate constants based on measured data.

► Production of nearly monodisperse silicon nanoparticles via pyrolysis of monosilane. ► Optimization of rate constants based on measured PSDs. ► Process control using a full finite volume discretisation coupled to a trust-region optimization. ► Identification of new reaction conditions for the hot wall reactor leading to predefined PSDs.