Towards an Improved Ceramic Water Filter Paradigm: From Design Optimization to Participatory Implementation

Abstract

Ceramic water filters (CWFs) are home-based drinking water treatment technologies promoted in rural and resource-restricted contexts due to their simplicity in terms of use, maintenance, and manufacture. However, cost, flowrate, breakage, and general misuse/disuse in the field are barriers to proliferation, adoption, and associated impact. This research addresses these limitations by evaluating technological innovations and participatory implementation in Tanzania. CWFs are produced by mixing clay, sawdust, water, and silver nanoparticles (AgNPs), pressing the mixture into a pot shape, and firing it in a kiln. AgNPs are added to enhance disinfection but are among the most expensive of filter elements. Cost reduction is thus explored through investigation of AgNP replacement and/or supplementation with zinc oxide (ZnO). Metals are challenged in isolation and combination against E. coli within batch and filter phases and disinfection and elution are assessed across time and varying water qualities. Combined AgNP-ZnO treatment proved synergistic under all conditions, consistently outperforming either metal alone. Supplementing silver with zinc in CWFs can therefore simultaneously reduce cost while improving bactericidal efficacy. Sawdust size and proportion within CWFs also relate to flowrate, strength, and bacteria removal, yet the nature of these relationships is unclear. Modelling these performance measures is consequently challenging. CWF disks with varied material characteristics are therefore evaluated for flow, strength, and disinfection efficacy. Nested multiple regression analysis with possibility-based design optimization is then investigated as a novel approach to predicting filter performance with generalized porous media characteristics. Optimum material parameters that maximize flowrate while maintaining sufficient disinfection and strength are identified. Optimization results show flowrate may increase to >8 L/hr without compromising water treatment capacity or strength if is porosity is <48%, dry density is >1.2 g/cm3, and intrinsic permeability is <29 x 10-9 cm2. Finally, multi-year community engagement and participatory program development processes in Longido, Tanzania are described. Three education and CWF provision structures are subsequently evaluated. Filter adoption, breakage, and associated health are monitored over 27 months. Multi-week and repeated knowledge communication proved critical to promoting CWF usage, maintenance, and protection. Long-term participant engagement and support is necessary to promote CWF longevity and improve related diarrheal health.