Search

In this work, we present a machine learning neural network model to predict the impact of realistic metal grain granularity (MGG) variability on the threshold voltage V T h and on the I D − V G characteristics of a silicon-based 12 nm gate length nanosheet FET. This model is based on the multi-layer perceptron (MLP) machine learning architecture. As realistic MGG maps consist of the distribution of grains on the gate with different work-function values, it is relevant to apply algorithms such as the principal component analysis to reduce these features to the most representative ones. Once the realistic MGG features are correctly reduced without losing information, we train two different neural networks with the neurons in the output layer as the only difference, to predict the V T h and the I D − V G characteristics, respectively. The comparison between TCAD results and the model, shows excellent agreement for the mean and standard deviation of V T h distributions for different average grain sizes values (from 3 nm to 10 nm) demonstrating the accuracy of the machine learning model. Also, we study the amount of data needed to accurately train the MLPs, leading to results that allow us to drastically reduce the computational time required to perform variability studies for state-of-art nano FET devices. • Machine learning model to predict the impact of variability on state-of-the-art semiconductor devices. • Prediction of the impact of metal grain granularity on the Vth and on the I-V characteristics of a 12 nm gate length Si-based nanosheet FET. • Model that reduces the computational cost of variability statistical studies. • Comparison of the model performance against TCAD simulations. [ABSTRACT FROM AUTHOR]

We present a novel Pelgrom-based predictive (PBP) model to estimate the impact of variability on the on-current of different state-of-the-art semiconductor devices. In this work, we focus on two of the most problematic sources of variability, the metal grain granularity (MGG) and the line edge roughness (LER). This model allows us to make an accurate prediction of the on-current standard deviation σ I o n , being the relative error of the predicted data lower than 8% in 92% of the studied cases. The PBP model entails an immense reduction in the computational cost since once it is calibrated for an architecture, the prediction of the impact of a variability on devices with any given dimension can be made without any further simulations. This model could be useful for predicting the effect of variability on future technology nodes. • Model to predict the impact of variability on state-of-the-art semiconductor devices. • Predicts metal grain granularity and line edge roughness on-current variability. • Model that reduces the computational cost of variability statistical studies. [ABSTRACT FROM AUTHOR]