Multi-variable Optimization to Improve Temperature Uniformity in Burn-in Applications Subjected to Significant Thermal Shadowing

Abstract

Burn-in test typically employs voltage and/or temperature to accelerate the appearance of latent reliability defects in semiconductor devices. As a result, burn-in becomes a critical step in the parts screening process and is the primary technique to eliminate defective parts in the early phase of the product life. Thermal control in burn-in processes is generally achieved by either active or passive thermal controlling. In an active thermal control, each device has an individual temperature monitoring system to maintain a specific temperature in which the cooling/heating of each socket is controlled by an individual fan or heater. Whereas in passive thermal control, a single fan provides convective heat transfer to an array of devices. Recent research regarding burn-in process suggests that passively controlled sockets are impacted more by temperature and airflow variation in burn-in ovens when compared to actively controlled sockets. In our study, a fan and two sockets were placed linearly and the airflow characteristics were examined past the first-row socket. An air deflector was designed on the ceiling of the oven between the two sockets and was optimized for its location, size, and orientation to improve the heat transfer characteristic of the second-row socket.For analysis and optimization, socket and burn-in chamber design specification were obtained from Plastronics Inc. 3D CAD model was designed in Solidworks and was solved numerically considering appropriate boundary conditions in Solidworks native Flow Simulation. Finally, Multi-variable design optimization and response surface methodology were performed to propose an optimum design for air deflector to increase the thermal performance of socket in the burn-n application.