Optimization of MEMS Chips for in-situ Biasing TEM Experiments in AlN-based ReRAM Devices

Abstract

In-situ transmission electron microscopy (TEM) analysis is a pivotal technique for characterizing dynamic material behaviors under external stimuli, such as thermal and electrical biasing. Unlike conventional static analysis, this technique offers real-time insights into the operating mechanisms of advanced materials and electronic devices. The reliability of in-situ experiments is intrinsically linked to the quality of the prepared TEM lamellae. While Focused Ion Beam (FIB) milling is the standard method for sample preparation, the requisite ion-beam-induced deposition (IBID) of platinum (Pt) inevitably introduces gallium (Ga) and carbon (C) contaminants. These FIB-induced artifacts can severely degrade the electrical properties of the specimen. Furthermore, commercial MEMS chips are often cost- prohibitive, susceptible to damage during the sampling process, and lack the design flexibility required for specific experimental needs. To address these challenges, this study proposes an optimized in-situ TEM biasing MEMS chip designed to mitigate FIB-induced damage. A novel electrode configuration minimizes the gap between the lamella and the chip electrodes, significantly reducing the reliance on additional IBID processes. This approach not only minimizes Ga ion damage, thereby extending the mechanical lifespan of the chip, but also suppresses the increase in contact resistance caused by Pt deposition. Moreover, the incorporation of a high-k dielectric layer effectively reduces leakage current, ensuring precise voltage control and signal integrity. Consequently, the specimen retains a near- pristine condition, enhancing the accuracy and reliability of in-situ measurements. The streamlined fabrication process not only reduces production costs but also allows for customizable electrode materials and configurations. The performance of the developed MEMS chip is benchmarked against existing solutions to demonstrate its superiority. Finally, the practical utility of the proposed chip is validated through in-situ biasing experiments on aluminum nitride (AlN)-based Resistive Random-Access Memory (ReRAM) devices.