Abstract
Diffusion is a critical doping process in silicon-based strain gauge fabrication, significantly influencing device performance. Traditional tube diffusion relies on empirical methods and repeated trials, limiting cost reduction and making it difficult to determine process parameters from strain gauge design requirements. Moreover, the cross effects between diffusion parameters are not fully understood. This paper develops a predictive model for tube diffusion using a solid-state boron nitride (BN) source, based on the response surface method (RSM). The model allows rapid determination of diffusion temperature, time, and protective argon flow rate to achieve target sheet resistance. Response surface analysis reveals that increasing temperature and time lowers sheet resistance and enhances uniformity, while time and flow rate show negligible cross effects. Multi-physical field simulations analyze the cross effects between temperature and flow rate, indicating that flow rates above 550 sccm increase temperature differences across wafers, reducing diffusion flux and uniformity. Experimental validation demonstrates a mean error of only 2.9% between predicted and actual sheet resistance, confirming the model’s accuracy. Using the optimized diffusion parameters, a semiconductor strain gauge with a gauge factor (GF) of 69.62 and a temperature coefficient of gauge factor (TCGF) of 0.53% F.S./°C was successfully fabricated. This predictive approach provides an efficient strategy for optimizing diffusion processes in strain gauge manufacturing, offering both performance control and cost reduction.
Funding Information
  • Chinese Aeronautical Establishment (20240012063002)
  • Chinese Aeronautical Establishment (20230012063001)
  • Science and Technology Program of Liaoning Province (2023JH1/10400044)

This publication has 33 references indexed in Scilit: