Improving low-temperature plasma jet sintering using data-driven Bayesian optimization

To meet the growing need for adaptable devices such as sensors, wearables, and touch screens, it is essential to devise innovative manufacturing techniques for devices made from inks that incorporate nanoparticles. Following the deposition of the nanoink, an important stage in the manufacturing process typically involves sintering at elevated temperatures, which is untenable for heat-sensitive substrates. Non-thermal plasma jet sintering using a dielectric barrier discharge (DBD) plasma jet enables sintering to occur at or close to room temperature and under atmospheric pressure conditions. It allows the consolidation of printed nanoinks into uniform films without causing harm to the underlying material or the film's surface.



Nevertheless, the task of determining the optimal conditions for plasma jet sintering is challenging due to multiple variables exhibiting intricate interdependencies. In this work, we employed Bayesian optimization (BO) to identify the optimal experimental condition for seven input variables, using indium tin oxide (ITO) nanoink as a model material. Following five experimental cycles, the optimization process resulted in a 99.2% enhancement in the measured electrical conductivity of plasma jet-sintered ITO thin films. In addition, the optimal sintering condition resulted in an electrical conductivity that was 81.4% of the conductivity achieved with traditional furnace sintering at a temperature of 300 °C. Notably, the plasma jet sintering process exhibited a threefold increase in speed, and the peak substrate temperature remained below 47 °C. These findings demonstrate the potential of utilizing BO to enhance procedures for plasma jet processes in industrial applications.