The research team will develop a series of multivalent nanozyme materials (MOF@Arg@His) to construct a sensing platform with dual-mode colorimetric and fluorescent responses, thereby markedly improving the sensitivity and selectivity for aluminum ion detection. By incorporating machine-learning algorithms, a signal-correction model will be established to enable highly accurate quantification of aluminum ions in complex matrices. Based on a clinical observational cohort, samples of PN raw materials, compounded admixtures, and infant biospecimens will be collected to systematically quantify aluminum exposure levels in preterm infants and to assess the associated risks of relevant complications. Integrating cutting-edge technological innovation with strong clinical applicability, this project aims to build a full-chain "formulation-exposure-outcome" evaluation framework, providing an intelligent detection tool and risk early-warning evidence base to enhance medication safety in neonates.