MOF-regulated aggregation-induced electrochemiluminescence for machine learning-assisted aptasensing of cancer marker.

Electrochemiluminescence (ECL) biosensing requires luminophores with high emission efficiency and favorable interfacial reaction kinetics. Herein, a nanoconfined aggregation-induced emission luminophore, TH-UiO-66-NH2, is constructed by coordinating tetracarboxyl-functionalized tetraphenylethylene (TCTPE) within UiO-66-NH2. The resulting TH-UiO-66-NH2 shows an 8.3-fold enhancement in ECL intensity compared with pristine TCTPE. Mechanistic studies reveal that the rigid UiO-66-NH2 not only enhances the interfacial enrichment and electrochemical activation of K2S2O8, but also regulates the aggregation state of TCTPE within a spatially confined coordination environment, thereby suppressing nonradiative relaxation of the AIE luminophore. Taking advantage of the excellent ECL performance of TH-UiO-66-NH2, a signal "on-off-on" biosensing platform is further developed for carcinoembryonic antigen (CEA) detection. The proposed biosensor shows a linear relationship ranging from 10-13 to 10-7 g mL-1 with detection limit down to 2.138 × 10-14 g mL-1. Moreover, a random forest-assisted machine-learning model further improves the detection accuracy, demonstrating the potential of this intelligent ECL system for highly sensitive CEA detection. This work offers a reliable and intelligent analytical strategy for the precision screening and auxiliary diagnosis of cancer markers.