Integrated microfluidic sensor array based on mixed-valence cerium metal organic framework for machine...

Accurate identification of antioxidants is crucial for health assessment and food safety, yet conventional analytical methods often suffer from limited portability and operational efficiency. Herein, we propose an integrated microfluidic colorimetric sensor array based on a mixed-valence cerium metal-organic framework (MVCM) for the sensitive and patterned identification of six key antioxidant biomarkers. The MVCM exhibits exceptional oxidase-like activity, with density functional theory (DFT) calculations revealing that its catalytic prowess stems from a narrow bandgap (0.63 eV) and optimized adsorption energy for H₂O₂ and O₂. The obtained MVCM can oxidate the 3,3',5,5'-tetramethylbenzidine (TMB) into oxTMB with blue color. By exploiting the differential inhibitory effects of biothiols and phenolic molecules on the MVCM-TMB system, a high-dimensional Red-Green-Blue (RGB) colorimetric fingerprint database was established. Integrated with machine learning-assisted principal component analysis (PCA), the sensor array achieved 100% classification accuracy for the six antioxidants within a linear range of 4-32 μM and maintained robust discrimination performance in complex binary and ternary mixture systems. This platform was further translated into a portable polydimethylsiloxane (PDMS)-glass microfluidic chip combined with smartphone imaging, enabling reliable assessment of antioxidant profiles in real samples. Overall, this work provides a low-cost, instrument-free, and non-invasive paradigm for point-of-care testing (POCT) and personalized oxidative stress monitoring.