Bayesian Optimization\nof Low-Temperature Nonthermal\nPlasma Jet Sintering of Nanoinks

In response to the escalating demand for flexible devices\nin applications\nsuch as wearables, sensors, and touch panels, there is a need for\ninnovative fabrication approaches for devices made from nanomaterial-based\ninks. Subsequent to ink deposition, a pivotal stage in device manufacturing\ntypically involves high-temperature sintering, posing challenges for\nheat-sensitive substrates. Nonthermal plasma jet sintering utilizing\nan atmospheric pressure dielectric barrier discharge (DBD) plasma\njet enables sintering at room temperature and standard pressure, facilitating\nthe sintering of printed nanoparticle films without compromising substrate\nor film surface integrity. However, determining optimal plasma jet\nsintering conditions can be challenging due to multiple processing\nvariables with intricate interrelationships. This work employed Bayesian\noptimization (BO) and machine learning (ML) to identify optimal values\nfor seven primary plasma jet sintering variables. Optimization yielded\na 99.2% increase in the measured electrical conductivity for plasma\njet-sintered indium tin oxide (ITO) films after five rounds of experiments.\nMoreover, the optimal sintering conditions achieved an electrical\nconductivity that was 81.4% of conventional furnace sintering at 300\n°C, but was three times faster and with a peak substrate temperature\nbelow 47 °C. This result demonstrates the prospect of applying\nBO to optimize processing techniques for emerging low-temperature\nrequirements.