From laboratory to field: a critical review of multi-modal sensing frameworks for emerging contaminant monitoring.

Emerging pollutants are now widely detected in water at trace concentrations (ng/L to µg/L). Conventional monitoring relies on laboratory methods such as LC-MS/MS and GC-MS, which offer exceptional sensitivity (0.5-50 ng/L) and unambiguous compound identification but remain slow, expensive, and confined to centralized facilities. Sample preparation alone consumes over 70% of analysis time, and grab sampling misses transient pollution events. Field-deployable sensors offer real-time data but lack the sensitivity, selectivity, and long-term stability required for routine monitoring. Therefore, this review examines three sensor classes, metal-oxide semiconductors, surface-enhanced Raman spectroscopy, and quartz crystal microbalance, as potential components of multi-modal sensing networks. Their analytical performance, reproducibility, and technology readiness levels (TRL 3-7) are compared. The analysis finds that direct evidence showing multi-modal systems outperform single sensors is surprisingly sparse; most improved platforms remain at TRL 3-5, with few field validations. Practical deployment barriers, calibration drift, biofouling, matrix effects, machine-learning overfitting, nanomaterial reproducibility, and the lack of regulatory standards are discussed. The proposed Environmental Health Monitoring (EHM) framework is therefore presented as a transformative framework, as complete integration of laboratory reference methods, field sensors, artificial intelligence, and cloud-based decision systems has yet to be demonstrated experimentally. Sensor stability, field validation, and standardization remain critical priorities for future work.