Extended field investigations of ozone-biofiltration advanced water treatment for potable reuse

Abstract

Recovering and reusing treated wastewater effluent is a sustainable and cost-effective practice for addressing global water sustainability. To date, most potable reuse advanced water treatment (AWT) solutions are based on reverse osmosis (RO), which generates a continuous reject stream of concentrated brine waste. Ozone-biofiltration based solutions have been investigated as a potential alternative for RO. However, implementation of ozone-biofiltration for potable reuse projects around the world has been limited. The goal of this study was to conduct an extended field investigation of ozone-biofiltration treatment to address regulatory, design, and operational hurdles that may hinder implementation in water-short areas. For 16 months, two parallel biological activated carbon (BAC) filters were operated at empty bed contact times (EBCTs) of 10 min and 20 min treating up to 60,000 and 30,000 bed volumes (BVs), respectively, of sand filtered effluent from a municipal wastewater treatment process. BAC 1 (EBCT = 10 min) and BAC 2 (EBCT = 20 min) used Calgon Filtrasorb 400 granular activated carbon (GAC) as filter media, with equal bed depths of 0.8 m. Increasing the specific ozone dose from 0.9 to 2.0 provided a muted response with respect to oxidation of contaminants of emerging concern (CECs) that are resistive to ozonation. N-Nitrosodimethylamine (NDMA) was generated during ozonation, with the average concentration of NDMA in ozonated effluent being 40.4 ng/L. In BAC 1 (EBCT = 10 min), NDMA was fully removed during the first month of study (<2000 BVs), partially removed between 2000 and 20,000 BVs, and completely removed when monitored between 57,000 and 62,000 BVs. These trends clearly reveal time-dependent interactions between carbon-based (e.g., adsorption) and non-carbon-based (e.g., biodegradation) removal mechanisms. In BAC 2 (EBCT = 20 min), almost all CECs, excluding NDMA, were removed consistently throughout the study (through ∼30,000 BVs). This indicates a somewhat different interaction between carbon-based and non-carbon-based removal in the more lightly loaded BAC 2, compared to BAC 1. After 482 days of operation, BAC 1 (EBCT = 10 min) produced effluent with lower NDMA concentration (<2 ng/L) than BAC 2 (10 ng/L), confirming prior evidence of cometabolic NDMA biodegradation pathways operable in more heavily loaded BACs. These findings emphasize the need for extended field testing (50,000 BVs or greater). BAC 1 removed TOC in effluent until it plateaued at around 6 mg/L after 60,000 BVs, whereas BAC 2 effluent plateaued at around 4 mg/L. Under plateau conditions, BAC 1 and BAC 2 with sand filter pretreatment and ozonation appear to have a gross TOC removal potential of around 0.2–0.3 kg of TOC removed per day per cubic meter of carbon media (kg/d/m3). A comparative analysis of findings from this study and results from a past ozone-BAC study in the Reno area (termed BAC 3 operated downstream of membrane filter with an EBCT of 30 min) shows that higher TOC removal was observed in BAC with shorter EBCT and upstream sand filter compared to BAC with longer EBCT and upstream membrane filter. The present study addresses the regulatory and financial concerns associated with ozone-BAC performance in potable reuse applications. Improved comprehension of ozone-BAC performance, coupled with its reduced capital and operations and maintenance (O&M) costs compared to RO, may accelerate the full-scale implementation of ozone-BAC treatment as a sustainable solution for the rapidly emerging potable reuse market.