Research on the Transfer Learning of MOSFET Physical of Failure Model Based on LSTM

In view of the wide variety of MOSFETs, if all design of experiments is carried out for physical of failure (PoF), the cost is too high and it is not easy to operate, If all PoF models are established based on experimental results, the engineering quantity is too large and not suitable for implementation. Based on the above reasons, the paper conducts research on the transfer learning of MOSFET's PoF models. Firstly, MOSFET’s PoF model is established which the failure mechanism is negative bias temperature instability(NBTI). Then, the key materials, structures, and processes is determined that affect MOSFET’s characteristic parameters, and a mapping relationship is established between key parameters and PoF model's coefficients based on the long short term memory(LSTM), namely transfer learning model. Finally, by comparing the accuracy of the proposed method with that of the traditional method (Polynomial regression), it is found that the accuracy of the transfer learning model based on LSTM is higher than that of the traditional method, which verifies the effectiveness of the proposed method.