A microprobe for salicylic acid detection assisted by deep learning

Salicylic acid (SA) is widely used in personal care products for its notable anti-acne and whitening effects. However, excessive levels may cause skin irritation, underscoring the need for rapid, reliable detection. In this study, a cyclotriphosphazene-curcumin fluorescent microprobe (PC-Al) was developed for SA detection. The probe provides an ultrafast response within 1 min and up to 100-fold fluorescence enhancement, with a detection limit of 0.1 μM. It showed high selectivity for SA over structural analogs and coexisting ions and was successfully applied to cosmetic samples, achieving recoveries near 100%. In addition, a deep-learning-assisted classification model was employed to validate the fluorescence-based quantification, further supporting the practical applicability of the method.