Abstract
1. Background
Climate change and rapid population growth continue to stress drinking water sources and water reuse is being considered a sustainable option to meet growing water needs, especially in water scarce regions. Some technologies have received interest over the past few decades (e.g., Ozone/ granular activated carbon (GAC)) while others have become technically and economically viable options in recent years (e.g. RO). Previously, reuse was limited to non-potable applications. However, drastic population growth and surge in water-demand has led to applicability in potable-reuse applications as well. Presence of disinfection by-products (DBPs) such as trihalomethanes (THMs), haloacetic acids (HAAs) and N-nitrosamines in discharge water is of great concern due to their toxic nature and impact on human health. Moreover, RO generate effluents with TOC levels < 0.5 mg/L and ozone-BAC effluent TOC levels ranging between 2.5 – 3.5 mg/L. However, risk of DBP formation in ozone-BAC effluent downstream potable water treatment and distribution is unknown. Therefore, additional investigation to better understand effect of higher effluent TOC on THMs formation and overall DBPs loads during potable reuse is warranted. As BAC relies on microbial remediation of organics in filter matrix, understanding microbial-diversity in BAC filters is an aspect of 1) developing rational design and operational criteria for BAC, 2) understanding why various BAC units have produced varying results (such as for NDMA) and 3) monitoring health of BAC process and taking corrective actions as may be necessary to maintain the performance. This study focused on following objectives: - Establish baseline relationship between effluent quality and DBPs for full-scale systems using RO membranes and ozone-BAC. - Pilot-scale optimization of BAC to achieve nitrosamines and CECs (including flame retardants) removal - Microbial characterization of the systems using molecular sequencing tools.
2. Material and methods
System performances of RO and BAC were compared by sampling from 6 BAC (BAC 1, BAC 2, BAC 3A, BAC 3B, BAC 3C and BAC 3D) and 2 RO full-scale systems across US for 1 year. Various operational parameters (media type and depth, operational mode (down or upflow), influent flow rate, filter run length, headloss, Empty Bed Contact Time (EBCT), loading rate, upstream/downstream treatment processes and their configuration, media age, backwash practice etc.) were recorded. Inorganics characterization and organic carbon characterization of water were conducted. DBPs (THM, HAA, bromate, nitrosamines) were monitored for all influent and effluent. Microbial community analysis was conducted from biofilms extracted from sampled BAC filter media (full and pilot-scale) (Pre/post backwash, pre/post ozonation, climatic condition, backwash liquid, tertiary effluent). Genomic DNA was extracted, purified, and quantified using Qubit (Thermofisher, US). PCR amplified products were checked on a bioanalyzer (Shimadzu, US). V3-V4 hypervariable region of the 16S rRNA gene was amplified and sequenced using 515F (5'-GTGCCAGC(A/C)GCCGCGGTAA) and 806R (5'-GGACTAC(A/C/T)(A/C/G)GGGT(A/T)TCTAAT) (Caporaso et al. 2011). Downstream processing of raw data was conducted using Quantitative Insights into Microbial Ecology (QIIME2) (Bolyen et al. 2019). The pilot-scale study was conducted at South Truckee Meadows Water Reclamation Facility, Nevada to optimize BAC for removal of CECs, NDMA, CECs, TOC. Several ratios of ozone:TOC were tested for steady-state operation. Samples were collected from influent and effluent for water quality and microbial analysis.
3. Results and discussion
3.1. Full scale (Ozone BAC and RO) Concentration of TOC and AOC in the system's effluents were found to be within normal ranges for the type of treatment plants under study (TOC – BAC: 2.0 – 5.0 mg/L, RO: 0.1 – 0.8 mg/L; AOC – BAC: 93 – 851 μg/L, RO: 50 – 103 μg/L) with an average TOC in the effluent of 3.4 mg/L for the BAC systems. TTHM was <5 µg/L and HAAs were not detectable in BAC systems but were marginally higher in RO effluents (TTHM: 3 to 13 µg/L and HAA5: ~1 µg/L). Levels of NMOR in BAC systems (3.8-10.1 ng/L) were generally higher than RO effluents (2-4 ng/L). BAC 2 was the only location where bromate levels in ozonated effluents were above 5 μg/L (range 21 μg/L -77.4 μg/L). No bromate was detected in non-ozonated sample whereas ozonated sample (8.5 mg/L applied ozone dose; O3:TOC ratio of 1.7) yielded 21.5 μg/L of bromate.
3.2. Pilot scale demonstration After 132 days of operation, influent TOC increased to around 8 mg/L and BAC 1 effluent TOC plateau increased to around 6 mg/L while that BAC 2 had effluent TOC concentrations around 2 mg/L (Figure 1). Average 1,4 dioxane concentration in ozone effluent was 0.72 µg/L. Negligible removal of 1,4 dioxane was observed across BAC 1 except initially. Ozonation reduced atenolol concentrations by 44%, sucralose by 20%, primidone by 38% and Lopressor by 55%. TCEP, TDCPP, and TCPP results showed a decline in removal efficiencies through the testing period in BAC 1. NDMA removal in BAC 2 was consistently better than in BAC 1 during the initial phase of operation when carbon-based mechanisms were more prevalent.
3.3. Microbial ecology In full-scale BAC 1, Proteobacteria was the most abundant phylum in all samples with relative abundance ranging from 45% to 92%, followed by Chloroflexi, Actinobacteria, Planctomycetes and Bacteroidetes. A trend of decrease in relative abundance, especially of Proteobacteria, was observed between before and after backwash (Before backwash: 70.2 ± 19 %; after backwash: 57.1 ± 10.2 %). Like BAC 1, Proteobacteria was the most abundant phylum with relative abundance ranging from 22% (before backwash collected during fall) to 55% (before backwash collected during winter) in BAC 2. Acidobacteria was next most abundant phylum average of 11.4 % of bacterial reads. Planctomycetes (2.7 ± 1.3 %), Bacteroidetes (2.3 ± 1.2 %), Chloroflexi (2 to 3 %) and Actinobacteria (2.9 ± 0.8 %) were next most abundant phyla. Similarly, average relative abundance of Proteobacteria in BAC 3A and B was 39.2 ± 17 % while BAC 3C and BAC 3D were 54 % and 99% respectively. of Planctomycetes, Acidobacteria, Bacteroidetes, Nitrospirae and Actinobacteria were next most abundant phyla. Pilot scale showed similar presence of Protoebacteria as the most abundant phylum with an average relative abundance of 38.5 ± 10.2 % and 44 ± 14.4 % in BAC 1 and BAC 2 respectively.
Conclusions - Full-scale BAC systems typically produced effluents of higher TOC levels than RO systems. - Ozone-BAC formed less NDMA than RO systems that use chloramine for biofouling control. - Effluent NDMA concentrations were relatively independent of EBCT within the ranges and loads studied in the Pilot systems. - Protoebacteria was the major phylum identified in majority of the sample from Pilot and full-scale BAC followed by Actinobacteria and Bacteroidetes. - The microbial ecology in different filters will be compared relative to each other and their respective operational conditions to provide insights for future optimization.
The current study was conducted to establish relationship between effluent quality and disinfection by products for full scale systems using RO membranes and ozone-BAC. Pilot scale optimization of BAC was conducted to achieve nitrosamines and CECs (including flame retardants) removal and characterization of organic constituents in the effluents. Microbial characterization of pilot and full scale BAC systems was conducted for optimization of operational parameters.
Author(s)Sunayna Dasgupta1; Zia Bukhari2; Ruth Marfil-Vega3; Vijay Sundaram4
Author affiliation(s)American Water Works Service Company Inc, USA1; American Water Works Service Company Inc, USA2;Shimadzu Scientific Instruments, USA 3; AECOM, USA4
SourceProceedings of the Water Environment Federation
Document typeConference Paper
Print publication date Oct 2022
DOI10.2175/193864718825158532
Volume / Issue
Content sourceWEFTEC
Copyright2022
Word count18