Intelligent algorithms enable electrohydrodynamic jet printing for micro-LEDs defect repair

Electrohydrodynamic jet printing enables the generation of femtoliter-scale droplets, offering advantages such as additive manufacturing and high process efficiency. With these advantages, electrohydrodynamic jet printing becomes the optimal choice for Micro-LEDs pixel repair processes. Traditional repair methods cannot balance yield rates and costs effectively. This paper proposes an intelligent manufacture scheme that integrates supervised learning algorithms and genetic algorithms to address repair issues, where droplet volume control and path planning represent the greatest challenges in this process. For droplet volume control, we propose a supervised learning algorithm that uses volume deviation. The datasets include volume deviation, current process parameters, and parameter changes. We use an MLP supervised algorithm with feedforward control to recommend printing parameters. For path planning, we optimize both shortest path and minimal parameter adjustments using a dual-chain encoding genetic algorithm. Sequential encoding handles defect locations while integer encoding manages parameter adjustments, achieving optimal path planning. We propose a printing system that integrates these two functions, and we conducted Micro-LEDs repair experiments on 156 defect locations. The experimental results showed that the standard volume printing accuracy of the entire panel repair area reached 98.1%, and the printing planning time was five times faster than the traditional method.