Intelligent Temperature Monitoring for Microfluidic System Using Mo _0.5 W _0.5 S ₂ ‐Polymer Composite Based Memristive Sensory Neurons

Abstract Precise and intelligent temperature monitoring in microfluidic systems is significant for improving efficiency and reducing costs across various applications in biology, chemistry, and medicine. Biomimetic sensing technologies integrated with on‐chip pulse encoding offer a promising strategy by directly capturing temperature signals and encoding them for further processing, while providing advantages including high‐density integration and low energy consumption. The threshold‐switching memristor (TSM), configured to simulate leaky integrate‐and‐fire (LIF) neurons, is particularly appealing as it combines sensing and encoding functionalities within a single device. However, achieving high‐performance TS characteristic and temperature sensitivity in TSM‐based LIF neurons remains a challenge. Here, a temperature‐responsive TSM is presented with low operating voltage by using Mo 0.5 W 0.5 S 2 nanosheets (NSs)‐polymer composite as resistive switching materials. The Mo 0.5 W 0.5 S 2 NSs act as an “intermediate waystation” at the composite during the Ag ion migration process, facilitating the formation of conductive filaments and decreasing the operating voltage. A thermal flow angle recognition system is developed by integrating a 40 × 40 memristor array onto the microfluidic chip surface. With the assistance of a spiking neural network, seven categories of thermal flow angles can be successfully identified with a high accuracy of 99.75%, indicating the TSM's great potential in microfluidic chip temperature monitoring.