Periodic Mesoporous Metal Oxides With Consistent Structure and Complementary Sensing Selectivity for Reliable Artificial Olfactory

The integration of functional materials and intelligent algorithms has endowed artificial olfactory systems with promising potential across various fields. Nevertheless, their accuracy and reliability in practical gas recognition remain constrained, primarily due to the unrefined combination of sensing materials and inadequate feature extraction from response curves. In this work, a series of periodic mesoporous metal oxides (PMMOs) with consistent structure and tunable compositions are synthesized by using polymer cubosomes as general templates. Within the periodic mesoporous channels, the target gases follow well-defined diffusion pathways, so that the dynamic features of response curves can be leveraged to build a reliable multidimensional dataset, and the consistency across repeated measurement cycles and different devices can be well maintained. By tailoring the chemical microenvironment, a sensor array is constructed from seven distinct PMMOs with complementary sensing selectivity, achieving 97.1% accuracy in discriminating seven representative hazardous gases at 1-5 ppm with neural network model assistance. In complex environment, the established system can also accurately recognize the indicative gases. As a verification, it successfully realize the screening of diabetic patients by the exhaled breath analysis, with acetone serving as the key gaseous biomarker.