Reverse osmosis membranes are intended to constitute a complete physical barrier against nanometricsized pathogens such as enteric viruses. Literature describes low-pressure reverse osmosis achieves high viral removal rates (above 5 log), surpassing those of ultrafiltration (1 to 3 log). However, these studies often used individual viruses and high feed viral concentrations (above 10 9 virus L -1 ), greater than typical viral concentrations present in the environment like groundwater, to promote virus detection in the permeate. These high concentrations can promote viral aggregation, potentially affecting the observed retention. This work evaluates the simultaneous elimination of three viruses during the production of drinking water by low-pressure reverse osmosis: two enteric viruses (adenovirus 41 and coxsackievirus-B5) and bacteriophage MS2, a widely used virus surrogate in the literature. The permeates produced by low-pressure reverse osmosis were concentrated to allow virus detection in permeate at lower feed concentrations (10 6 virus L -1 ) while staying above the limits of detection and quantification. Experiments were carried out on two pilot plants of different scales (laboratory and semi-industrial) to assess the potential effect of the number of membranes and O-rings on virus retention. The effect of the volume concentration factor on low-pressure reverse osmosis efficiency was evaluated for each scale. Results indicate an average viral reduction of 6 log (up to 7 log), regardless of the size of the virus or the scale of LPRO pilot. For the semi-industrial scale, better retention was observed as the volume concentration factor increased. However, viruses were still present in the permeate for each scale (even if close to the detection limit), indicating that retention was not complete. At the same feed viral concentrations, the number of viruses recovered in the semi-industrial scale permeates was higher than in the laboratory scale. A 24-fold greater the number of membranes and O-rings used for the semi-industrial scale showed that micro-leaks through O-rings could be responsible for the passage of viruses into the permeate.
Human enteric viruses are important etiological agents of waterborne diseases. Environmental waters are usually contaminated with low virus concentration requiring large concentration factors for effective detection by (RT)-qPCR. Low pressure reverse osmosis is often used to remove water contaminants, but very few studies focused on the effective virus removal of reverse osmosis treatment with feed concentrations as close as possible to environmental concentrations, and principally relied on theoretical virus removal. The very low viral concentrations usually reported in the permeates (i.e. at least 5 log of removal rate) mean that very large volumes of water need to be analysed to have sufficient sensitivity and assess the process efficiency. This study evaluates two methods for the concentration of adenoviruses, enteroviruses and MS2 bacteriophages at different viral concentrations in large (<200 L) and very large (>200 L) volumes. The first method is composed of two ultrafiltration membranes with low molecular weight cut-offs while the second method primarily relies on adsorption and elution phases using electropositive-charged filters. The recovery rates were assessed for both methods. For the ultrafiltration-based protocol, recovery rates were similar for each virus studied: 80 % on average at high virus concentrations (10 6 -10 7 viruses L -1 ) and 50 % at low virus concentrations (10 3 -10 4 viruses L - 1 ). For the electropositive-charged filter-based method, the average recoveries obtained were about 36 % for ADV 41, 57 % for CV-B5 and 1.6 % for MS2. The ultrafiltration-based method was then used to evaluate the performance of a low-pressure reverse osmosis lab-scale pilot plant. The retentions by reverse osmosis were similar for all studied viruses and the validated recovery rates applied to the system confirmed the reliability of the concentration method. This method was effective in concentrating all three viruses over a wide range of viral concentrations. Moreover, the second concentration method using electropositive-charged filters was studied, allowing the filtration of larger volumes of permeate from a semi-industrial low-pressure reverse osmosis pilot plant. This reference method was used because of the inability of the UF method to filter volumes on the order of one cubic metre.
