Machine learning in electron beam lithography to boost photoresist formulation design for high-resolution patterning

Thermal radiation analysis of a broadband solar energy-capturing absorber using Ti and GaAs

This study employed a time-domain finite-difference (FDTD) approach to design an efficient solar energy-capturing absorber consisting of a high melting point metal (Ti) and a semiconductor (GaAs). The structure generated cavity resonance (CR) and surface plasmon resonance (SPR), leading to extremely high absorption across different wavelength bands. The structure exhibited >90% absorption over a wide wavelength range (280-3000 nm). It achieved an average absorption efficiency of 97.0% in the wavelength range from 280 nm to 3000 nm at an air mass of 1.5 (AM1.5). The structure showed insensitivity to the angle of incidence, maintaining stable absorption of over 94% for angles of incidence ranging from 0° to 60°. The structure could also operate at 1400 K, with thermal radiation efficiencies of up to 98.2%. As the operating temperature increased from 600 K to 1400 K (with a temperature gradient of 200 K), the thermal radiation efficiency of the structure remained above 98% at all times. Based on the excellent radiation and absorption properties of this absorber, it holds promising application in the fields of solar energy absorption and thermal emission.

Convolutional Neural Network-Assisted Photoresist Formulation Discriminator Design of a Contact Layer for Electron Beam Lithography

The photoresist formulation is closely related to the material properties, and its composition content determines the lithography imaging quality. To satisfy the process requirements, imaging verification of extensive formulations is required through lithography experiments. Identifying photoresist formulations with a high imaging performance has become a challenge. Herein, we develop a formulation discriminator of a metal oxide nanoparticle photoresist for a contact layer applied to electron beam lithography. This discriminator consists of convolutional neural network photoresist imaging and formulation classification models. A photoresist imaging model is adopted to predict the contact width of variable formulations, while a formulation classification model is used to classify formulations according to relative local critical dimension uniformity (RLCDU). The verification results indicate that the discriminator can accurately recognize photoresist formulations that simultaneously meet the conditions of contact width and RLCDU, and its feasibility has been demonstrated, providing a valuable reference for the preparation of photoresist materials.

Machine Learning Applied to Electron Beam Lithography to Accelerate Process Optimization of a Contact Hole Layer

Determining the lithographic process conditions with high-resolution patterning plays a crucial role in accelerating chip manufacturing. However, lithography imaging is an extremely complex nonlinear system, and obtaining suitable process conditions requires extensive experimental attempts. This severely creates a bottleneck in optimizing and controlling the lithographic process conditions. Herein, we report a process optimization solution for a contact layer of metal oxide nanoparticle photoresists by combining electron beam lithography (EBL) experiments with machine learning. In this solution, a long short-term memory (LSTM) network and a support vector machine (SVM) model are used to establish the contact hole imaging and process condition classification models, respectively. By combining SVM with the LSTM network, the process conditions that simultaneously satisfy the requirements of the contact hole width and local critical dimension uniformity tolerance can be screened. The verification results demonstrate that the horizontal and vertical contact widths predicted by the LSTM network are highly consistent with the EBL experimental results, and the classification model shows good accuracy, providing a reference for process optimization of a contact layer.