Water Hardness from Photos of Dried Drop Stains: Reporter Salts Enable Low-Concentration Detection.
Journal:
Environmental science & technology
Published Date:
Jul 8, 2025
Abstract
Water hardness is critical in domestic and industrial systems, typically measured by titration, atomic absorption spectroscopy, or inductively coupled plasma mass spectrometry. Here, we present a novel, photo-based method for quantifying water hardness using deposit patterns formed during the evaporation of hard water mixed with highly concentrated solutions of the reporter salts NaHCO and NaCl. This indirect approach enables the detection of hardness levels as low as 25 ppm, an improvement by a factor of 1,400 over earlier stain-based studies. Our method leverages high-resolution imaging, automated sample preparation, and advanced image analysis, including machine learning, to detect subtle morphological changes in the deposit patterns. A robotic drop imager (RODI) generated over 16,000 images, from which 47 geometric and textural metrics were extracted to train high-dimensional linear regression models and multilayer perceptron (MLP) networks. Using synthetic water samples with hardness levels of up to 250 ppm, we achieved prediction accuracies exceeding 91% for sodium bicarbonate-based patterns. Extending this methodology, we validated the methodology on real-world tap water, demonstrating its potential as a cost-effective alternative to traditional techniques and a versatile tool for water quality monitoring.
