In semiconductor industries, keeping competitive requires to develop efficient and high yielding manufacturing facilities.
With Advanced Process Control base on run-to-run control we can improve the yield and get the fast throughput.
One way to incr...
In semiconductor industries, keeping competitive requires to develop efficient and high yielding manufacturing facilities.
With Advanced Process Control base on run-to-run control we can improve the yield and get the fast throughput.
One way to increase yield is to reduce process variability, but semiconductor processes often drift due to equipment aging, depletion of chemicals, fluctuation in ambient conditions, machine replacement or maintenance.
With algorithms such as exponentially weighted moving average(EWMA), Neural Network, we can get proper input values that might keep output uniformity.
In this paper, I mainly focused to get proper input values on Photo stepper equipment with an artificial neural network which is based on the mean field theory and performs very good results on nonlinear function such as a white noise characteristic.
The new ANN algorithm is tested and demonstrated with a very remarkable results not only increasing throughput with reduced Rework ratio but also enhancing Cp/Cpk.