Abstract

Semiconductor foundries have entered an era of 12-inch wafers and over five hundred production processes involved in manufacturing and measurement. In order to stabilize equipment and reduce variations in processes, an effective equipment maintenance model is required. This study thus aimed to establish a model to maintain the key equipment in a semiconductor factory, based on its history of preventative maintenance (PM) and applying Monte Carlo simulation to predict the probability of the next PM time-point. Focusing on a semiconductor foundry producing RAM, this study found that Chamber2 in the diffusion zone is the key equipment in the FMEA (failure mode and effect analysis), categorized the historical data from various phases to be applied to Monte Carlo simulation, and, with 10,000 simulations and computations, obtained the maintenance time-point probability and the appropriate date for Chamber2 to be the subject of a management review.

Key words: preventive maintenance, Monte Carlo simulation, failure mode and effect analysis (FMEA).

Abbreviation

 PM, Preventative maintenance; FMEA, failure mode and effect analysis; IC, integrated circuits; CD, critical dimensions; RPN, risk priority number; CB, crystal ball.