A High-throughput Coaxial DBD Reactor for Water Purification through Synergistic Thin-film and Bubble-flow Dual-interface Design

Abstract Non-thermal plasma (NTP) enables reagent-free in situ generation of reactive oxygen and nitrogen species (RONS) for water treatment, yet practical reactor designs face a persistent trade-off between stable discharge ignition and efficient gas–liquid mass transfer. This study presents a continuous-flow coaxial dielectric barrier discharge (DBD) reactor engineered with a dual-interface configuration that deliberately couples an annular thin liquid film with forced air bubbling to stabilize microdischarges and intensify plasma–liquid–gas interfacial transport. Two environmentally relevant water matrices—municipal utility backwash effluent and contaminated river water from Sorocaba, São Paulo (Brazil)—were treated at 120 L·h − 1 with ambient air injection (~ 40 L·min − 1 ) in a treatment train comprising pre-filtration, plasma contact, and a downstream mixed-media polishing filter. Performance was assessed using ISO/IEC 17,025-accredited analytical methods across five independent experimental runs. The integrated system achieved complete inactivation of total coliforms and Escherichia coli (non-detect in 5/5 replicates; Fisher’s exact test, p = 0.0079), reduced ammonia from 14.18 ± 0.28 mg·L − 1 to below the limit of quantification (< 0.01 mg·L − 1 ; >99.9%), and decreased dissolved aluminum and iron by ≥ 99%. The plasma-specific energy input was 2.4 kJ·L − 1 (80 W discharge power determined by Lissajous analysis), corresponding to an ammonia removal efficiency of 21.3 g-NH 3 ·kWh − 1 and an electrical energy per order (EE O ) of 0.37 kWh·m − 3 ·order − 1 (total system power basis). Microbiological indicators (total coliforms, E. coli , HPC) and dissolved metals (Al, Fe) were evaluated in the utility backwash matrix, while the remaining physicochemical endpoints were evaluated in river water; comparisons across analytes should therefore be interpreted with this matrix assignment in mind. The present results demonstrate the effectiveness of the integrated treatment train under the stated operating conditions and an electrical behavior consistent with stable filamentary microdischarge activity, but do not yet isolate the incremental contribution of the thin-film flow, the bubble injection, or their combination; the study is therefore positioned as a concept validation of the integrated dual-interface reactor rather than a component-by-component deconvolution.