A multivariate approach to amperometric free chlorine monitoring: calibration, interference compensation, and application in water distribution systems.

Reliable monitoring of Free Available Chlorine (FAC) in drinking water is essential for ensuring disinfection efficiency and detecting contamination events, yet in situ amperometric measurements remain strongly affected by matrix variability and long-term instability. In this work, we present an optimized amperometric protocol for FAC determination implemented in NEMO, a low-cost, battery-powered Online Water Quality Monitoring (OWQM) device designed for continuous operation in water distribution systems. The working potential was identified through cyclic voltammetry and subsequently optimized using a multivariate Design of Experiments (DoE) approach. To address the strong dependence of the amperometric signal on pH and conductivity, two Multiple Linear Regression calibration strategies were developed: an in silico calibration model (Approach A) and a direct multivariate model (Approach B). Both approaches were built on 125 calibration samples and validated on an independent test set, showing a consistent agreement with the reference DPD spectrophotometric method. Limits of detection were evaluated through four complementary procedures, confirming suitability for typical FAC levels in drinking water. Field deployment on six devices over two months demonstrated overall stable performance, limited hysteresis, and reliable tracking of temporal FAC variations. The proposed methodology enables robust semi-quantitative chlorine monitoring in real networks and represents a scalable solution for distributed early-warning systems in water quality management.