Abstract—This work presents a novel design strategy for a micro-electro-mechanical system (MEMS) formalde hyde (CH2O) sensor based on high-entropy metal oxide semiconductors (HE-MOSs). To address the challenge of trace CH2O detection in complex indoor gas mixtures, the intrinsic high defect density of the HE-MOSs is exploited via a variable temperature (VT) mode and sensor array design, achieving an ultra-low detection of 355 ppt. With the opti mized VT parameters, the device achieves 100% accurate identification of trace CH2O (200 ppb) in high-concentration ethanol via machine learning algorithms. These designs hold significant potential for reliable, intelligent MEMS gas sensors deployed in complex environments. 

Index Terms—MEMS gas sensors, high-entropy metal oxide, formaldehyde, variable temperature