Transforming Water Supplies in the Midwest: Two CBAT Pilots Demonstrate the Potential for Water Reuse

Abstract

Population growth, increasing water demands for data centers, and the need for more sustainable water practices are prompting advancement of next-generation water resource management strategies including water reuse. In the Midwestern United States (U.S.), where non-traditional approaches to augmenting water supply, including water reuse, are relatively new, full advanced water treatment (AWT), which includes microfiltration/ultrafiltration (MF/UF) and reverse osmosis (RO), produces a concentrate stream that is expensive to address (i.e., brine disposal). Carbon-based advanced treatment (CBAT), which combines ozonation, biofiltration, and granular activated carbon, has been shown to be a viable alternative with select advantages over the traditional RO approach, including the lack of brine generation. Two novel pilots were conducted in the Midwestern U.S., one at a large (>100 MGD) and one at a small (~10 MGD) wastewater reclamation facility (WRF), to provide proof of concept that CBAT could meet distinct regional needs. While additional demonstration data are ultimately needed for future regulatory approvals, results from the pilots showed that water quality objectives were met with treated water quality of <0.5 mg/L total Kjeldal nitrogen (TKN), <2 mg/L total organic carbon (TOC), and substantial reduction in constituents of emerging concern (CECs). If nitrate removal is required to meet drinking water standards (10 mg/L), additional treatment optimization at the source WRFs would be required. Areas for further research identified by this effort include mitigating the potential for ozonation to contribute to the formation of disinfection byproducts such as bromate and N-nitrosodimethylamine (NDMA). Along with the treatment performance demonstrated at these pilots, the study provided an opportunity to engage with key stakeholders to build trust in the AWT approach, which is critically important for regulatory and public acceptance. Based on two field-scale pilot studies in the U.S. Midwest, this paper analyzes CBAT application advantages and challenges in municipal water reuse and identifies research directions for the industry.

1. Introduction

Due to a number of emerging factors, including population growth, increased water demands for data centers, and an emphasis on more sustainable water practices, water reuse is gaining acceptance as an option in a diversified portfolio of water supply options [1]. Importantly, these trends are driving growth in water demands in specific geographies, including the Midwestern United States (U.S.), where non-traditional approaches to augmenting water supply are relatively new. The concept of addressing these water supply needs through water reuse has prompted the formation of new WaterReuse Association chapters in states including Ohio, Illinois, and Kansas. In the context of this move toward considering water reuse in integrated water supply planning, utilities are exploring advanced water treatment (AWT) for non-potable and potable reuse applications [2]. Due to the lack of previous experience with water reuse in the U.S. Midwest, there is a gap in identifying potential treatment options for augmenting water supplies. While it will be important to develop a body of evidence for this practice at additional locations and for a longer period, initial pilot testing can be a supporting mechanism for developing this regional experience.

To identify options for these inland facilities, it is important to consider that AWT for water reuse is built around multi-barrier “treatment trains” that combine physical, chemical, and biological processes to achieve required levels of pathogen and contaminant removal. The most established configuration of AWT for high-quality applications, including potable reuse, is microfiltration/ultrafiltration (MF/UF) → reverse osmosis (RO) → advanced oxidation process (with ultraviolet light and hydrogen peroxide, UV/H₂O₂), which provides high removal of pathogens, salts, and trace organics. While this treatment train has traditionally been considered as a regulatory benchmark in many regions [3], challenges do exist with this approach. Capital and operating costs, energy, and brine management challenges of this treatment train have driven growing adoption of non-RO based potable reuse approaches, especially at inland sites. Key examples of these include the F. Wayne Hill Water Resources Center, owned and operated by the Gwinnett County Department of Water Resources, GA, and the Upper Occoquan Service Authority (UOSA), VA [4].

One promising AWT approach that addresses the issues associated with RO concentrate is carbon-based advanced treatment (CBAT), which integrates ozone, biofiltration, and granular activated carbon (GAC) in a multi-barrier process to convert treated wastewater effluent into high-quality reclaimed water. This configuration is an efficient option for addressing a wide spectrum of water quality parameters (microorganisms, organic contaminants, and emerging contaminants) for a variety of reuse applications, including potable, non-potable, and industrial applications. The Hampton Roads Sanitation District (HRSD), VA’s Sustainable Water Initiative for Tomorrow (SWIFT), is another excellent example of a utility leveraging this non-RO based approach to augment drinking water supplies [5], and communities like Franklin, TN that have piloted a CBAT plan to incorporate CBAT as an AWT scheme into their integrated portfolio of water-supply treatment options [6]. Additionally, CBAT may offer economic advantages for inland utilities and has been shown to be less than half the cost of a comparable traditional Full Advanced Treatment (FAT) process when certain brine disposal options are considered [7].

Table 1. Comparison of carbon-based and membrane-based advanced treatment for potable reuse.

	Carbon-Based Advanced Treatment	Membrane-Based Advanced Treatment
Core Treatment Processes	Ozonation, Biofiltration, GAC, UV (advanced oxidation either accompanies ozone or UV)	MF/UF, NF/RO, UV/H2O2
Multiple Chemical Contaminant Barriers	Yes	Yes
Multiple Pathogen Treatment Barriers	Yes	Yes
Dissolved Solids Removal	No	Yes
Water Quality Considerations	Suitability of treatment processes to influent water quality	Chemical compatibility of treated water quality with receiving waters (aquifers or distribution system)
Operational Complexity	Less complex processes, generally requiring greater operator intervention	Highly complex but generally more automated processes
Capital and Operating Costs	Generally Lower	Generally Higher
Waste Stream Management Considerations	Spent filter/contactor media handling	Concentrate/brine handling
Regulatory Considerations	Less commonly regulated	More commonly regulated, considered benchmark technology

presents a general comparison of membrane-based advanced treatment trains and CBAT for potable reuse.

Two pilot studies were conducted as a novel approach to demonstrating a proof of concept regarding the feasibility and versatility of CBAT in the Midwest, where no regulatory framework for water reuse exists. The pilot treatment train included chemical pretreatment with coagulation, flocculation, and sedimentation (CFS), followed by ozone, two biologically activated carbon (BAC) filters in series, and two GAC filters in series. A process flow diagram is shown in Figure 1.

Additionally, pilot testing of CBAT in the Midwest is a novel approach to stakeholder engagement, where proof of concept for reducing uncertainty and establishing a credible technical foundation for water reuse can be demonstrated. This is critically important in a region where both treatment experience and regulatory structure are still emerging. With only a few full-scale CBAT systems implemented in the U.S., these pilots conducted at both large (>100 MGD) and small (10 MGD) utilities serve as critical early demonstrations of performance to inform feasibility. These pilots aimed to show how CBAT (Figure 1) performed under Midwest-specific conditions, including geographically unique water quality characteristics influenced by the land use and development of the region, which features a mix of rapid residential growth, legacy industrial use, modern high-technology investments, and large-scale agriculture. By generating high-quality, region-specific data, these efforts will help regulators, utilities, and engineers better understand treatment performance, operational complexity, and design requirements in a context that differs from regions where high-quality reuse applications are well-established [8]. For comparison with other utilities operating CBAT for water reuse,

Table 2. Comparison of select CBAT systems for water reuse in the U.S.

Utility	Two Undisclosed (This Study)	Gwinnett County, GA	HRSD, VA	Franklin, TN
Geographical Location	U.S. Midwest	U.S. Southeast	U.S. Mid-Atlantic	U.S. Upper South
Nearby Ocean for Potential Brine Disposal	No	No	Potentially	No
Upstream Wastewater Treatment	Secondary effluent from a conventional activated sludge facility	Tertiary effluent from a conventional activated sludge facility	Secondary effluent from conventional activated sludge facility	Tertiary effluent from a conventional activated sludge facility
Operational Status	Pilot scale completed	Full-scale in operation	Demonstration facility (1 MGD) in operation, full-scale under construction	Pilot scale completed, full-scale in design
Core Operational Features	Potential non-potable reuse and indirect potable reuse via groundwater recharge	Surface water augmentation for indirect potable reuse	Aquifer recharge for indirect potable reuse	Planned non-potable reuse and surface water augmentation for indirect potable reuse
State-Level Reuse Regulations	Not yet established	Guidance only	Established approach	Under development

summarizes key qualitative characteristics evaluated as part of this study, along with those of other CBAT systems used for inland potable reuse applications.

Equally important to demonstrating treatment performance, these pilots are supporting the broader institutional and social groundwork needed for the adoption of water reuse as a strategic water supply option [9]. In the absence of an established Midwest regulatory framework, these novel pilot studies provide the technical evidence required to inform discussions around future guidance and policy development. They create a shared basis of understanding among utilities, regulators, and public stakeholders, helping to build trust in CBAT as a viable alternative to more conventional approaches based on RO. By demonstrating feasibility at different scales and transparently evaluating performance, these projects support long-term operations planning and considerations for workforce readiness. Ultimately, the motivation extends beyond technology validation to enabling a diversified portfolio of options in integrated water supply strategies, positioning reuse as a locally appropriate solution to meet water demands as resource pressures continue to grow in the Midwest.

2. Materials and Methods

The AWT pilot was constructed by Intuitech, Inc. (Salt Lake City, UT, USA) and consisted of three treatment skids inside a modified shipping container. The unit treatment processes included: coagulation-flocculation-sedimentation (CFS) (Intuitech S300), ozone (Intuitech Z400), biofiltration and granular activated carbon (Intuitech F1200). Dedicated feed pumps for each skid were installed with overflows strategically placed between the treatment processes to allow for flexibility in modifying flow rates and treatment configurations. The flow rates through the pilot were approximately 5 gallons per minute (gpm) through the CFS skid, 2.5 gpm through the ozone skid, and 1 gpm through the filter skid. A programmable logic controller (PLC)-based SCADA system and human-machine interface (HMI) screens on each skid enabled both local operator control at the pilot trailer and remote access for adjustment of key operational parameters and alarm diagnostics.

The pilot operated in multiple phases at each site, with different objectives described in the following sections. Water temperatures in the pilot influent ranged from an average of approximately 55 to 65 °F (approximately 13 to 18 °C) from February to June in Pilot Study 1 and from an average of approximately 68 to 73 °F (approximately 20 to 23 °C) from July to August in Pilot Study 2. Each pilot study operated with all unit processes in service except where otherwise noted.

At each of the two facilities where pilot testing was conducted, secondary clarifier effluent was pumped to the pilot trailer and conveyed through the AWT train. Both studies included coagulation-flocculation-sedimentation as the first treatment step, with the goal of lowering turbidity and reducing particle loading onto the downstream filters. The coagulant selected for each pilot study was dosed into the rapid mix basin on the CFS skid. Pilot Study 1 used ferric chloride at a dose of 50 mg/L and Pilot Study 2 used aluminum chlorohydrate (ACH) at a dose varying between 35 and 40 mg/L. The water then passed through a flocculation basin with 60 min of retention time followed by a sedimentation basin with 45 min of retention time. The sedimentation basin included plate settlers to increase the effective settling area in the small footprint of the pilot trailer.

Clarified water was pumped to the ozone skid, which had two ozone contact basins available with over-under baffles and a fine bubble diffuser. One basin was actively used during each study, providing a retention time of 50 min. Because 50 min was greater than the time required for ozone concentrations to decay to zero residual, the second contactor was not needed. Ozone is a strong chemical oxidant and disinfectant that improves the biodegradability of organic compounds [10]. When combined with downstream biofiltration, ozone improves the removal of organic compounds through the filters and can strengthen the performance of BAC filters and GAC contactors; accordingly, ozone-biofiltration is a viable and commonly utilized advanced water treatment technology [11]. The ozone dosing strategy was based on targeted ozone to total organic carbon (ozone/TOC) ratios at certain project phases and experiments, such as an ozone/TOC ratio of approximately 0.5 for transforming bulk organic matter and trace organic contaminants, and higher ratios at or above an ozone/TOC ratio of approximately 1.0 that is on the scale of that used for disinfection where a substantial ozone residual persists in the contactor. These targets have been informed by potable reuse regulations in other states [12] as well as recent research [13]. The applied ozone dose was recorded, and online measurements of ozone off-gas were used to calculate the real-time transferred ozone dose that was achieved. Additionally, ozone residual measurements were collected at 10 min contact time.

Ozonated water was pumped to the filtration skid, first passing through two BAC filter columns in series. BAC filters are typically paired immediately downstream of ozonation for polishing of organic matter, nutrients, and certain emerging contaminants by the bacterial communities present on the biofilm of the filter media [14,15]. A total of 10 to 15 min of contact time was provided in each of the two BAC filters for over 20 min of overall empty bed contact time (EBCT). Acclimated BAC media was collected from a full-scale drinking water facility running BAC as part of its normal treatment process. The donor facility used Norit GAC 816 (Norit Activated Carbon, Horsham, PA, USA) that was installed in approximately 2015. The BAC media from Pilot Study 1 was reused in Pilot Study 2. Upon optimization of the filtration process as described in the subsequent sections, the typical backwashing frequency for the BAC filter columns was every 48 to 72 h.

Following the second BAC filter, water passed through two GAC contactor columns in series. GAC media provides removal of organic compounds and emerging contaminants, such as per- and polyfluoroalkyl substances (PFAS), through the mechanism of adsorption to pores on the surface of the media [16,17]. A total of 10 to 15 min of contact time was provided in each of the two GAC contactors for over 20 min of overall EBCT. Virgin GAC (Filtrasorb 300^®, Calgon Carbon Corporation, Moon Township, PA, USA) was used in Pilot Study 1, with the same media reused in Pilot Study 2. Typical backwashing frequency of the GAC contactor columns was once per month.

Table 3. BAC and GAC media characteristics.

Parameter	BAC Filter Media	GAC Contactor Media
Iodine Number (mg/g)	900 (min)	900 (min)
Moisture by Weight	2% (max)	2% (max)
Effective Size (mm)	1.3–1.5	0.8–1.0
Uniformity Coefficient	1.4 (max)	2.1 (max)
Abrasion Number	75 (min)	78 (min)
Mesh Size
Greater than:	8 mesh (2.36 mm): 15%	8 mesh (2.36 mm): 15%
Less than:	16 mesh (1.18 mm): 4%	30 mesh (0.595 mm): 4%
Apparent Density (g/cm3)	0.51	0.56

summarizes the physical characteristics of the BAC and GAC media used for both pilots.

Throughout the duration of each pilot, comprehensive water quality monitoring was conducted across the AWT train. In addition to the robust set of conventional parameters that were monitored, disinfection byproducts (DBPs), pharmaceuticals and personal care products (PPCPs), and other organic and inorganic contaminants were also measured. Each week, over 20 water quality parameters were tested across the treatment train (Appendix A,

Table A1. Water quality parameters tested weekly at each pilot location.

Parameter	Method	Locations Tested	Frequency
Conductivity	EPA Method 10256	AWT Feed, GAC 2 Effluent	5 x/week
Turbidity	EPA Method 180.1	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	5 x/week
UVT-254	HACH Method 10243	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	5 x/week
Oxygen, Dissolved	HACH Method 10360	AWT Feed, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	5 x/week
Ozone Residual	Indigo Method, HACH Method 8311	Settled Water, Ozone Effluent	5 x/week
pH	EPA Method 10257	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	5 x/week
Iron, Total	HACH Method 10306	AWT Feed, Settled Water, GAC 2 Effluent	1 x/week
Iron, Dissolved	HACH Method 10306	AWT Feed, Settled Water, GAC 2 Effluent	1 x/week
Total Alkalinity	Colorimetric Method, HACH 10239	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Nitrate	Dimethylphenol Method, HACH 10206	AWT Feed, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Nitrite	Diazotization Method, HACH 10237	AWT Feed, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Ammonia	Salicylate Method, HACH 10205	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Total Phosphorus	Ascorbic Acid Method, HACH 10209/10210	AWT Feed, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Soluble Reactive Total Phosphorus (Orthophosphate)	Ascorbic Acid Method, HACH 10209/10210	AWT Feed, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week
Total Chemical Oxygen Demand	EPA Reactor digestion Method, HACH 8000	AWT Feed, Settled Water, Ozone Effluent, BAC 2 Effluent, GAC 2 Effluent	3 x/week
Solids, Total Dissolved	Standard Method 2540C-20	AWT Feed, Settled Water, GAC 2 Effluent	1 x/week Pilot Study 1 3 x/week Pilot Study 2
Solids, Total Suspended	Standard Method 2540D-20	AWT Feed, Settled Water, GAC 2 Effluent	1 x/week Pilot Study 1 3 x/week Pilot Study 2
Carbonaceous Biochemical Oxygen Demand	Standard Method 5210B-16	AWT Feed, Settled Water, GAC 2 Effluent	1 x/week Pilot Study 1 3 x/week Pilot Study 2
Total Kjeldahl Nitrogen	EPA Method 351.2 Rev. 2.0	AWT Feed, Settled Water, Ozone Effluent, BAC 2 Effluent, GAC 2 Effluent	1 x/week
Dissolved Organic Carbon	Standard Method 5310B-00,14	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	2 x/week
Total Organic Carbon	Standard Method 5310B-00,14	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	3 x/week

). Ultimately, over two hundred unique water quality parameters were sampled and analyzed (Appendix A,

Table A2. Larger suite of water quality parameters tested with less frequency at each pilot location. Note: PFAS were only tested in Pilot Study 1.

Parameter	Locations Tested	Frequency
Odor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Color	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Foaming Agents	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Bromate	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Corrosivity	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Sulfate	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Bromide	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Chlorite	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
MBAS	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Sulfamethoxazole	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Formaldehyde	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 2 Effluent	1 x/week during challenge testing
Acetone	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 2 Effluent	1 x/week during challenge testing
Carbamazepine	AWT Feed, Settled Water, Ozone Effluent, GAC 2 Effluent	1 x/week during challenge testing
Assimilable Organic Carbon	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	1 x/week during challenge testing
1–4 Dioxane	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	1 x/week during challenge testing
Nitrosamines	AWT Feed, Settled Water, Ozone Effluent, BAC 1 Effluent, BAC 2 Effluent, GAC 1 Effluent, GAC 2 Effluent	1 x/week during challenge testing
Haloacetic Acids (HAA5)	AWT Feed, BAC 2 Effluent, GAC 2 Effluent	Minimum 3 x/duration of pilot
Total Trihalomethanes (TTHM)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Gross Alpha	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Gross Beta	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Radium 226	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Radium 228	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Uranium	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Silicon	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfide, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sodium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sodium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Manganese, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Manganese, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Silver, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Silver, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Zinc, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Zinc, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2-Dibromo-3-Chloropropane (DBCP)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
DEET	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dicamba	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
2(2,4,5-Trichlorophenoxy)propionic acid (Silvex)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
2,3,7,8-Tetrachlorodibenzodioxin (TCDD, Dioxin)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
2,4-Dichlorophenoxyacetic acid (2,4-D)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
3-Hydroxycarbofuran	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Alachlor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Aldicarb	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Aldicarb sulfaoxide	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Aldicarb sulfone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Atrazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Benzo[a]pyrene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Butachlor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Carbaryl	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Carbofuran	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chlordane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dalapon	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Di(2-ethylhexyl)adipate	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Di(2-ethylhexyl)phthalate	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dinoseb	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Diquat	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Endothall	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Endrin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Ethylene Dibromide (EDB)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Glyphosate (Round-Up)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Heptachlor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Heptachlor epoxide	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Hexachlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Hexachlorobutadiene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Hexachlorocyclopentadiene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Lindane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Methiocarb	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Methomyl	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Methoxychlor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Metolachlor	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Metribuzin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Oxamyl (Vydate)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Pentachlorophenol	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Picloram	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Polychlorinated Biphenyls (Arochlor 1016–1260)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Simazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Toxaphene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1,1,2-Tetrachloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1,1-Trichloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1,2,2-Tetrachloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1,2-Trichloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1-Dichloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1-Dichloroethylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,1,-Dichloropropene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2,4-Trichlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2,3-Trichloropropane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2,4-Trichlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2,4-Trimethylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2-Dichlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2-Dichloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,2-Dichloropropane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,3,5-Trimethylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,3-Dichlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,3-Dichloropropane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,4-Dichlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
2,2-Dichloropropane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
2-Chlorotoluene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
4-Chlorotoluene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Benzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Carbon Tetrachloride	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
cis-1,2-Dichloroethylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
cis-1,2-Dichloroethene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
cis-1,3-Dichloropropene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dichloromethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Ethylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Monochlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Styrene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Tetrachloroethylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Toluene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
trans-1,2-Dichloroethylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
trans-1,2-Dichloroethene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
trans-1,3-Dichloropropene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
tert-Butylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Trichloroethylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Trichlorofluoromethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Vinyl Chloride	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Xylenes, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chromium, Dissolved Hexavalent	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Calcium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Calcium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Aluminum, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Aluminum, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Antimony, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Antimony, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Arsenic, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Asbestos, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Barium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Barium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Beryllium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Beryllium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Boron, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cadmium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cadmium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chloride	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chromium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chromium, Hexavalent	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chromium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cobalt, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cobalt, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Copper, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Copper, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cyanide	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cyanide, Free	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Lead, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Lead, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Magnesium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Magnesium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Molybdenum, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Fluoride, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Mercury, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Nickel, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Nickel, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Potassium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Potassium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Selenium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Selenium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Strontium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Thallium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Thallium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Tin, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Titanium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Vanadium, Dissolved	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Vanadium, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Coliform, Total	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
E. coli	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Whole Effluent Toxicity	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Microcystins	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorobutanoic acid (PFBA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorobutanesulfonic acid (PFBS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorodecanoic acid (PFDA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoro-3,6-dioxaheptanoic acid	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorododecanoic acid (PFDoA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoro (2-ethoxyethane) sulfonic acid (PFEESA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoroheptanoic acid (PFHpA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoroheptanesulfonic acid (PFHpS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorohexanoic acid (PFHxA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorohexanesulfonic acid (PFHxS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoro(4-methoxybutanoic acid)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoro-3-methoxypropanoic acid (PFMPA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorononanoic acid (PFNA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorooctanoic acid (PFOA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorooctanesulfonic acid (PFOS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoropentanoic acid (PFPeA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoropentanesulfonic acid (PFPeS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorotetradecanoic acid (PFTeDA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluorotridecanoic acid (PFTrDA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Perfluoroundecanoic acid (PFUnA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
11-Chloroeicosafluoro-3-oxaundecane-1-sulfonic acid	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorohexane sulfonic acid (4:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorooctane sulfonic acid (6:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorodecane sulfonic acid (8:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
9-Chlorohexadecafluoro-3-oxanonane-1-sulfonic acid	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
4,8-Dioxa-3H-perfluorononanoic acid (ADONA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Hexafluoropropylene Oxide Dimer Acid (HFPO-DA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
N-ethylperfluorooctanesulfonamidoacetic acid (NEtFOSAA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
N-methylperfluorooctanesulfonamidoacetic acid (NMeFOSAA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
11-Chloroeicosafluoro-3-oxaundecane-1-sulfonic acid	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorohexane sulfonic acid (4:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorooctane sulfonic acid (6:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1H,1H,2H,2H-Perfluorodecane sulfonic acid (8:2 FTS)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
9-Chlorohexadecafluoro-3-oxanonane-1-sulfonic acid	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
4,8-Dioxa-3H-perfluorononanoic acid (ADONA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Hexafluoropropylene Oxide Dimer Acid (HFPO-DA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
N-ethylperfluorooctanesulfonamidoacetic acid (NEtFOSAA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
N-methylperfluorooctanesulfonamidoacetic acid (NMeFOSAA)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Acetaminophen	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Amoxicillin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Androstenedione	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Atenolol	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Azithromycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Caffeine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Carbadox	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Carbamazepine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cimetidine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Codeine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Cotinine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dexamethasone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Diazepam	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Diltiazem	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Diphenhydramine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Epitestosterone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Erythromycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Fluoxetine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Lincomycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Meprobamate	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Metformin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Monensin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Narasin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Nicotine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Oleandomycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
1,7-Dimethylxanthine (Paraxanthine)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Phenazone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Primidone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Progesterone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Roxithromycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Salinomycin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfamethoxazole	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfadiazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfadimethoxine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfamerazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfamethazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfamethizole	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfasalazine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Sulfathiazole	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Testosterone	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Theobromine	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Trimethoprim	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Tylosin	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Virginiamycin M1	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Actinolite (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Amosite (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Anthophyllite (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chrysotile (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Crocidolite (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Naphthalene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Quinoline	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
TCEP	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
TCPP	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
TDCPP	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Tremolite (MFL)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Bromochloromethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Benzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Bromobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Bromomethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chlorobenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chloroethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Chloromethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dibromomethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Dichlorodifluoromethane	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Isopropylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
m,p-Xylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Methylene Chloride	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
Methyl-tert-butyl-ether (MTBE)	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
n-Butylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
n-Propylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
o-Xylene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
p-Isopropyltoluene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot
sec-Butylbenzene	AWT Feed, GAC 2 Effluent	Minimum 3 x/duration of pilot

). Most of the weekly testing was conducted by the pilot operations team, while some samples were sent to a local laboratory for analysis. A larger suite of analytes was tested two to three times at each location. These samples were sent out for analysis to both the local laboratory, Alloway Environmental Company (now Eurofins Scientific), and to Eurofins Scientific. Both laboratories are accredited by the National Environmental Laboratory Accreditation Program (NELAP). Information regarding chains of custody, accreditation, and quality control practices can be found at each institute’s respective website. Each pilot study included additional testing unique to the site.

Statistical analyses were performed for various water quality parameters in RStudio (2026.01.0) using R version 4.5.1 [18]. Any sample replicates (a particular parameter sampled at the same location and time) were averaged before analysis. Unless otherwise specified, statistical tests were performed using the base R statistics package stats. Excel data were imported into R with the readxl package and formatted using the tidyverse package [19,20]. For each parameter, distributional assumptions were evaluated for each sample location group using the Shapiro-Wilk test for normality, supplemented by visual inspection via quantile-quantile (Q-Q) plots, and equal variance assumptions were evaluated using both Bartlett’s and Levene’s tests. Levene’s test was performed using the car package [21]. Because violations of normality were observed across all parameters, nonparametric statistical methods were applied. Specifically, differences between sample locations were evaluated using the Kruskal-Wallis rank-sum test, followed by Dunn’s test for post hoc pairwise comparisons. Dunn’s Test was performed using the FSA package [22], with p-values adjusted using the Holm-Bonferroni method to control the family-wise error rate.

2.1. Pilot Study 1 Description

Pilot Study 1 was conducted using the secondary effluent water (the AWT feed, or feed water) of a conventional activated sludge (CAS) treatment facility with a permitted average design flow of 10 MGD. The pilot was conducted with the objective of assessing whether sufficient water quality could be met to augment drinking water supplies without producing a concentrated stream. The service area contributing flow to the facility consists primarily of residential wastewater inputs. At this pilot, ferric chloride was used as the coagulant and was added to a rapid mix chamber. Jar tests were performed to obtain a starting ferric chloride dose. Soda ash was used for alkalinity and pH adjustment necessitated by using ferric chloride (which lowers pH) as well as for maintaining an optimal pH for BAC. Soda ash was applied immediately before the ozone contactor. Additional water quality testing at this pilot study location included PFAS (persistent, man-made “forever chemicals” used in industrial and consumer products and that are resistant to natural degradation and treatment), microplastics (University of Missouri, manuscript in preparation), nitrogenous DBPs (N-DBPs, currently unregulated DBPs that form when organic nitrogen precursors are present and are of concern due to their higher toxicity than regulated DBPs, manuscript in preparation), and pathogens (manuscript in preparation).

Table 4. Operational phases in Pilot Study 1.

Phase	Description	Duration
Phase 1: CFS Optimization	CFS optimization included varying the coagulant (ferric chloride) dose, rapid mixing and flocculation speeds, and sedimentation blowdown intervals to optimize floc formation and particle removal.	1 month
Phase 2: BAC Optimization	Air and water backwash rates and intervals were varied to optimize filter runtimes, along with pH adjustment with soda ash for biofiltration optimization.	2 weeks
Phase 3: Steady State	The pilot was run with minimal changes to operating conditions. Flow rates and chemical doses were generally constant, and 21 min of EBCT were provided through BAC (10 min in BAC column 1 followed by 11 min in BAC column 2) and 28 min through GAC (13 min in GAC column 1 followed by 15 min in GAC column 2). Ozone doses were varied based on a transferred ozone/TOC ratio between 0.75 and 1.0.	3 weeks
Phase 4: Challenge Testing	This phase included testing the performance of the ozone-biofiltration processes by varying the transferred ozone/TOC ratio between 0.5 and 1.2 while reducing BAC EBCT to a total of 10 min (by using one BAC column only).	1 month

shows the operational phases performed during the duration of Pilot Study 1.

2.2. Pilot Study 2 Description

The second pilot was conducted using the secondary effluent water of a major conventional activated sludge WRF that receives residential and industrial waste streams and flow from combined sewers, with a permitted average design flow of 114 MGD and a peak hydraulic capacity of 330 MGD. The objective of the pilot was to determine whether the system could provide water quality suitable for industrial reuse as well as potentially groundwater recharge for indirect potable reuse.

Table 5. Operational phases in Pilot Study 2.

Phase	Description	Duration
Phase 1: Startup and Commissioning	Startup and commissioning involved maintenance, troubleshooting, and jar testing of various coagulants. This phase also included external process and power connections.	1 week
Phase 2: BAC Optimization and Maintenance	Air and water backwash rates and intervals were varied to optimize filter runtimes and establish ideal conditions for biofiltration. During this period, a failure of the ozone mass flow controllers resulted in the ozone skid being offline for approximately two weeks while a manual bypass was prepared.	3 weeks
Phase 3: CFS Optimization	CFS optimization included varying the coagulant (ACH) dose, rapid mixing and flocculation speeds, and sedimentation blowdown interval to optimize floc formation and particle removal. During this period, there was a failure of the air conditioning system for the pilot enclosure, resulting in the pilot being offline for one week. Ozone, BAC, and GAC skids were shut down during this time, leaving only the CFS running.	2 weeks
Phase 4: Challenge Testing	Challenge testing involved assessing the limits and performance of the treatment processes by changing coagulant dose and varying the transferred ozone/TOC ratio between 0.5 and 1.0 while maintaining a constant EBCT through BAC and GAC. The same EBCTs were used as in Pilot Study 1, a total of 21 min through BAC and a total of 28 min through GAC.	3 weeks

shows the breakdown of the operational phases of Pilot Study 2.

Jar tests were performed with ferric chloride, polyaluminum chloride (PACL), and ACH to determine the optimal coagulant for the water source. Ultimately, ACH was selected as it showed strong settleability and did not necessitate pH adjustment downstream.

Additional parameters tested at Pilot Study 2 included microplastics (University of Missouri, manuscript in preparation), toxicity bioassays (INDIGO Biosciences, manuscript in preparation), and pathogen analysis (H₂O Molecular, manuscript in preparation). Due to the relatively high total dissolved solids (TDS) encountered at this facility, with expectations for increased TDS in the future due to new industry discharges, bench-scale hollow-fiber nanofiltration (HFNF) was investigated as an auxiliary treatment process option for selective ion removal to lower TDS. Bench-scale HFNF testing was performed using an NX Filtration Mexplorer Pilot Unit with a MP025-PVC dNF80 (800 Dalton cutoff) membrane module (NX Filtration, Enschede, The Netherlands). HFNF can potentially offer important advantages over RO for some inland water reuse applications. The HFNF membrane configuration provides exceptionally high packing densities (~500–3000 m²/m³), yielding substantially greater membrane surface area per unit volume than spiral-wound membranes and enabling higher throughput with a smaller footprint. HFNF membranes exhibit superior fouling resistance and support routine hydraulic backwashing, allowing effective performance recovery and improved tolerance to variable and high-foulant feed waters, capabilities that are not available in spiral-wound NF or RO systems. In addition, HFNF can achieve high water-recovery rates (>95% in certain applications), significantly reducing the volume of concentrate produced. Compared with RO, HFNF typically requires lower energy input for applications targeting selective TDS reduction (e.g., divalent ion removal), while emerging membrane materials also offer enhanced chemical stability for more aggressive cleaning. Collectively, these attributes make HFNF a compact, energy-efficient, and operationally robust alternative to traditional RO and NF membranes, especially in inland reuse settings where concentrate management is a key constraint.

Settled water and BAC effluent samples were tested at an effective 80% recovery (achieved by concentrating the influent water to 20% of its original volume using the dNF80 with a recirculating concentrate line and the permeate line to waste) in order to mimic the recovery of typical full-scale systems. The crossflow velocity was approximately 0.3 m/s with a membrane flux ranging from 44 to 50 L/m²-h (LMH). Projection models were also run to compare bench-scale versus modeled HFNF and to compare HFNF to other TDS removal membranes. This modeling included an NX Filtration dNF80 membrane (like used at bench-scale), an NX Filtration dNF40 membrane (400 Dalton cutoff), a DuPont NF270 spiral wound membrane (DuPont, Edina, MN, USA), and a DuPont BW30 RO membrane (DuPont, Edina, MN, USA). Modeling was performed using NX Filtration’s projection tool (NX Filtration, Hengelo, The Netherlands) and DuPont’s WAVE, projection tool, v1.83, Database 63 (DuPont, Wilmington, DE, USA).

3. Results

3.1. Pilot Study 1

Comprehensive water quality monitoring for Pilot Study 1 was conducted across the advanced treatment train, including conventional water quality parameters, organic contaminants, inorganic contaminants, and emerging contaminants. Results of the water quality monitoring demonstrated substantial removal of key constituents. This was exemplified in Figure 2, which shows a consistent trend in TOC reduction across the treatment train. Overall, there was a substantial reduction in TOC (85%) from approximately 11 mg/L to 1.7 mg/L, which is on the order of typical drinking water concentrations.

The boxes in Figure 2, as with subsequent box plots in other figures throughout this manuscript, encompass the interquartile range (IQR, 25th to 75th percentiles), with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in TOC concentrations across the treatment train (χ² = 91.03, df = 3, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment confirmed that every treatment stage achieved a statistically significant reduction in TOC from the preceding stage (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = b; Ozone-BAC Effluent = c; GAC Effluent = d).

Notably, the acclimated BAC demonstrated biomass stabilization between 2000 and 3000 bed volumes, following CFS optimization, when BAC effluent TOC concentrations were maintained at <4 mg/L. Similarly, GAC columns showed substantial initial adsorption capacity during CFS optimization with TOC stabilizing to under 2 mg/L during steady state (Figure 3).

UV Transmittance (UVT, specifically UVT₂₅₄), another surrogate for tracking natural organic matter (NOM), also showed substantial improvement throughout treatment from approximately 65% in the AWT feed, to 79% after ozone, and to 96% after GAC (Figure 4). The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in UVT concentrations across the treatment train (χ² = 148.63, df = 4, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment confirmed that every treatment stage achieved a statistically significant increase in UVT from the preceding stage (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = b; Ozone Effluent = c; BAC effluent =d, GAC Effluent = e).

Total Kjeldahl Nitrogen (TKN), which is the sum of organic nitrogen and ammonia, is well removed through ozone-biofiltration, where ozone increases biodegradability and the biofilter biologically degrades bioavailable nitrogen. The TKN concentrations (Figure 5) were reduced from just under 2 mg/L in the pilot feed to below the MDL of 0.5 mg/L, indicating high rates of treatment. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in TKN concentrations across the treatment train (χ² = 23.00, df = 3, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment was used to evaluate TKN changes between stages (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = ab; Ozone-BAC Effluent = b; GAC Effluent = b).

Except for nitrate, all GAC effluent parameters measured during Pilot Study 1 were below their respective U.S. Environmental Protection Agency (EPA) Drinking Water Maximum Contaminant Levels (MCLs). Nitrate through the pilot with time is shown in Figure 6, with an average concentration of 13.6 mg/L over the duration of the study, compared to the MCL of 10 mg/L.

The average pilot feed and GAC effluent concentrations for conventional water quality parameters, along with percent change in the concentrations, for Pilot Study 1 are presented in

Table 6. Pilot Study 1 conventional water quality parameters and percent change. Data include samples collected during steady state and challenge testing, following CFS and BAC optimization.

Parameter	Units	Average AWT Feed Value	Standard Deviation (Sample)	Average GAC Effluent Value	Standard Deviation (Sample)	% Change
Conductivity	mS/cm	0.87	0.05	1.00	0.05	+16%
Turbidity	NTU	2.71	0.86	0.15	0.05	−94%
UVT-254	%	65	0.01	97	0.01	+47%
DO	mg/L	7.26	0.39	11.53	1.39	+59%
pH	N/A	6.58	0.14	6.70	0.12	−2%
Total Alkalinity	mg/L-CaCO3	101	24.90	107.7	14.89	+7%
Nitrate	mg/L-N	13.9	3.06	15.7	2.53	+14%
Nitrite	mg/L-N	0.547	0.05	0.012	0.01	−98%
Ammonia	mg/L-N	0.109	0.05	0.015	0.00	−86%
Total Phosphorus	mg/L-P	0.24	0.05	0.01	0.02	−94%
Soluble Reactive Phosphorus	mg/L-PO4	0.21	0.15	<MDL	0.00	−100%
Total COD	mg/L	38.3	3.22	4.57	2.18	−88%
TDS	mg/L	540	15.0	581	26.2	+8%
TSS	mg/L	5.88	1.91	0.80	0.75	−86%
CBOD	mg/L	4.37	0.88	2.00	0.00	−54%
TKN	mg/L	1.75	0.34	0.5	0.00	−71%
DOC	mg/L	10.1	0.95	2.1	0.72	−80%
TOC	mg/L	10.9	1.10	1.7	0.31	−85%

. Data include occasional replicate sampling that was conducted (sampling at the same location and time for a particular parameter). Results equal to the MDL are reported as the MDL in this table.

In addition to conventional parameters, 212 unique constituents were evaluated in Pilot Study 1, including DBPs, PPCPs, PFAS, and other organic and inorganic constituents. Of these, 174 parameters were not detected (were less than the method detection limit, MDL) in the AWT feed water throughout sampling. The non-detected AWT feed water constituents included all tested polychlorobiphenyls (PCBs, a group of legacy chemicals used for their thermal resistance that, while banned, are persistent in the environment), volatile organic compounds (VOCs), and microcystins (a group of toxic cyanotoxins produced by certain blue-green algae). Of the measured constituents, 38 were detected above their MDLs, including certain halogenated DBPs, synthetic organic compounds (SOCs), pesticides, PFAS, PPCPs, and N-DBPs.

Table 7 summarizes the average, standard deviation, and number of samples for constituent detections in AWT feed water. Following treatment, 13 parameters were detected in the final GAC effluent:

• Low levels of 1,4-dioxane were detected in the effluent near the MDL of 0.07 ug/L, reduced from an average of 0.083 ug/L in the AWT feed water. 1,4-dioxane is a synthetic industrial chemical primarily used as a solvent and a stabilizer for chlorinated solvents. It is used as a treatment indicator in water reuse because it is exceptionally difficult to remove using standard treatment methods.
• Low levels of tris(2-chloroethyl) phosphate (TCEP) were detected in the effluent near the MDL of 0.01 ug/L, reduced from an average of 0.064 ug/L in the AWT feed water. TCEP is a flame retardant and plasticizer frequently assessed in water reuse because it is ubiquitous and highly persistent in the environment.
• Low concentrations of N,N-Diethyl-meta-toluamide (DEET), at less than 1 ug/L, were detected in the AWT feed water well under the Minnesota guidance value of 200 ug/L [23]. DEET is an active ingredient in personal insect repellents, and an indicator of human impact in water reuse. DEET was mostly removed by ozone and was typically under the MDL of 0.01 ug/L after GAC, as shown in Figure 7. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in DEET concentrations across the treatment train (χ² = 25.47, df = 4, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment was used to evaluate DEET changes between stages (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = a; Ozone Effluent = ab; BAC effluent = ab; GAC Effluent = b).
• Three shorter chain PFAS compounds were detected in the final GAC effluent, each at increasing concentrations as the pilot testing continued. Perfluorobutanoic acid (PFBA) was detected at a maximum concentration of 7.55 ng/L, Perfluorohexanoic acid (PFHxA) at a maximum concentration of 8.40 ng/L, and Perfluoropentanoic acid (PFPeA) at a maximum concentration of 9.90 ng/L.
• Detections of four PPCP compounds were observed throughout the pilot, including primidone at an average of 0.0087 ug/L, caffeine at an average of 0.0117 ug/L, metformin at an average of 1.66 ug/L, and cotinine at an average of 0.0056 ug/L. These compounds do not have federal or state guidelines and are not included on the proposed Contaminant Candidate List 6 (CCL 6) [24].
• Two N-DBPs were detected in the final effluent. N-nitrosodiphenylamine (NDPA) was detected on one occasion at 23 ng/L, and N-nitrosomorpholine (NMOR) was detected several times near the MDL of 2.00 ng/L. These nitrosamines are not currently regulated compounds; they may form during ozonation in the presence of certain precursors and are important in potable reuse because many are potent carcinogens.
• Gross Beta particles were detected in the effluent once at 10 pCi/L, which is below the EPA screening trigger level of 50 pCi/L [25].

Table 7. Constituent detections above their MDLs in the AWT feed of Pilot Study 1, excluding conventional parameters. Note that some MDL values varied for samples taken at different sampling periods due to the contract laboratory calibration protocols.

Detected in AWT Feed	Category	Average Result	Standard Deviation (Sample)	Units	No. of Samples	MDL
1,4-Dioxane	SOCs	0.083	0.010	ug/L	7	0.07
Androstenedione	PPCPs	0.005	0.000	ug/L	7	0.01
Atenolol	PPCPs	0.202	0.101	ug/L	7	0.01
Atrazine	Pesticides	0.093	0.041	ug/L	3	0.07
Azithromycin	PPCPs	0.700	0.383	ug/L	7	0.01 or 0.10
Caffeine	PPCPs	0.023	0.0159	ug/L	7	0.01
Carbamazepine	PPCPs	0.103	0.047	ug/L	7	0.01
Chloroform	DBPs	0.583	0.098	ug/L	6	0.50
Cimetidine	PPCPs	0.067	0.032	ug/L	7	0.01
Codeine	PPCPs	0.019	0.008	ug/L	7	0.01
Cotinine	PPCPs	0.021	0.005	ug/L	7	0.01
DCAA	DBPs	1.033	0.058	ug/L	3	1.00
DEET	Pesticides	0.567	0.254	ug/L	7	0.01
Diltiazem	PPCPs	0.114	0.049	ug/L	7	0.01
Diphenhydramine	PPCPs	0.335	0.203	ug/L	7	0.01 or 0.05
Erythromycin	PPCPs	0.023	0.005	ug/L	7	0.02
Fluoxetine	PPCPs	0.104	0.046	ug/L	7	0.01
Gross Beta	Radiological	11.220	12.134	pCi/L	2	4.00
Meprobamate	PPCPs	0.015	0.007	ug/L	7	0.01
Metformin	PPCPs	2.543	0.637	ug/L	7	0.02 or 0.20
Metolachlor	Pesticides	0.250	0.087	ug/L	3	0.20 or 0.19
NDMA	N-DBPs	2.471	0.544	ng/L	7	2.00
NMOR	N-DBPs	7.871	1.555	ng/L	7	2.00
Paraxanthine	PPCPs	0.040	0.026	ug/L	7	0.02
PFBA	PFAS	5.825	0.699	ng/L	4	2.00 or 1.90
PFBS	PFAS	3.850	0.887	ng/L	6	2.00 or 1.90
PFHxA	PFAS	12.417	4.348	ng/L	6	1.90 or 2.00
PFOA	PFAS	6.071	3.545	ng/L	7	2.00 or 1.90 or 2.10
PFOS	PFAS	1.971	0.076	ng/L	7	2.00 or 1.90 or 2.10
PFPeA	PFAS	11.825	2.862	ng/L	4	2.00 or 1.90
Primidone	PPCPs	0.243	0.089	ug/L	7	0.01
SMX	PPCPs	0.167	0.080	ug/L	7	0.01
Sulfasalazine	PPCPs	0.062	0.023	ug/L	7	0.05
TCAA	DBPs	1.033	0.058	ug/L	3	1.00
TCEP	SOCs	0.064	0.028	ug/L	7	0.01
TCPP	SOCs	0.570	0.171	ug/L	7	0.40
Trimethoprim	PPCPs	0.249	0.119	ug/L	7	0.01
Tylosin	PPCPs	0.006	0.001	ug/L	7	0.01

Table 7. Constituent detections above their MDLs in the AWT feed of Pilot Study 1, excluding conventional parameters. Note that some MDL values varied for samples taken at different sampling periods due to the contract laboratory calibration protocols.

Detected in AWT Feed	Category	Average Result	Standard Deviation (Sample)	Units	No. of Samples	MDL
1,4-Dioxane	SOCs	0.083	0.010	ug/L	7	0.07
Androstenedione	PPCPs	0.005	0.000	ug/L	7	0.01
Atenolol	PPCPs	0.202	0.101	ug/L	7	0.01
Atrazine	Pesticides	0.093	0.041	ug/L	3	0.07
Azithromycin	PPCPs	0.700	0.383	ug/L	7	0.01 or 0.10
Caffeine	PPCPs	0.023	0.0159	ug/L	7	0.01
Carbamazepine	PPCPs	0.103	0.047	ug/L	7	0.01
Chloroform	DBPs	0.583	0.098	ug/L	6	0.50
Cimetidine	PPCPs	0.067	0.032	ug/L	7	0.01
Codeine	PPCPs	0.019	0.008	ug/L	7	0.01
Cotinine	PPCPs	0.021	0.005	ug/L	7	0.01
DCAA	DBPs	1.033	0.058	ug/L	3	1.00
DEET	Pesticides	0.567	0.254	ug/L	7	0.01
Diltiazem	PPCPs	0.114	0.049	ug/L	7	0.01
Diphenhydramine	PPCPs	0.335	0.203	ug/L	7	0.01 or 0.05
Erythromycin	PPCPs	0.023	0.005	ug/L	7	0.02
Fluoxetine	PPCPs	0.104	0.046	ug/L	7	0.01
Gross Beta	Radiological	11.220	12.134	pCi/L	2	4.00
Meprobamate	PPCPs	0.015	0.007	ug/L	7	0.01
Metformin	PPCPs	2.543	0.637	ug/L	7	0.02 or 0.20
Metolachlor	Pesticides	0.250	0.087	ug/L	3	0.20 or 0.19
NDMA	N-DBPs	2.471	0.544	ng/L	7	2.00
NMOR	N-DBPs	7.871	1.555	ng/L	7	2.00
Paraxanthine	PPCPs	0.040	0.026	ug/L	7	0.02
PFBA	PFAS	5.825	0.699	ng/L	4	2.00 or 1.90
PFBS	PFAS	3.850	0.887	ng/L	6	2.00 or 1.90
PFHxA	PFAS	12.417	4.348	ng/L	6	1.90 or 2.00
PFOA	PFAS	6.071	3.545	ng/L	7	2.00 or 1.90 or 2.10
PFOS	PFAS	1.971	0.076	ng/L	7	2.00 or 1.90 or 2.10
PFPeA	PFAS	11.825	2.862	ng/L	4	2.00 or 1.90
Primidone	PPCPs	0.243	0.089	ug/L	7	0.01
SMX	PPCPs	0.167	0.080	ug/L	7	0.01
Sulfasalazine	PPCPs	0.062	0.023	ug/L	7	0.05
TCAA	DBPs	1.033	0.058	ug/L	3	1.00
TCEP	SOCs	0.064	0.028	ug/L	7	0.01
TCPP	SOCs	0.570	0.171	ug/L	7	0.40
Trimethoprim	PPCPs	0.249	0.119	ug/L	7	0.01
Tylosin	PPCPs	0.006	0.001	ug/L	7	0.01

The low perfluorooctanesulfonic acid (PFOS) concentrations (<MDL) in the AWT feed water remained below their EPA drinking water MCLs throughout the treatment train. The detected perfluorooctanoic acid (PFOA) concentrations in the AWT feed water were relatively low, except for a brief period where there may have been an industrial contribution that passed through the plant into the effluent, but this was effectively removed through the BAC and GAC (Figure 8). While PFOA was measured in the pilot influent, it was effectively reduced through the treatment train even during the PFOA spike that occurred during testing (Figure 8). The Kruskal-Wallis test indicated a statistically significant difference in PFOA concentration between the AWT Feed and GAC effluent (p = 0.008). Since the figure includes limited sampling for settled water, ozone effluent, and BAC effluent (number of samples, n = 1 at each location), statistical analysis was performed for the AWT feed and GAC groups only. Lowercase letters in parentheses indicate statistically unique groups (p < 0.05; AWT Feed = a; GAC Effluent = b).

3.2. Pilot Study 2

Similar comprehensive water quality monitoring was conducted for Pilot Study 2. Additional testing for Pilot Study 2 included microplastics, toxicity bioassays, pathogen analysis, and bench-scale HFNF. The results showed similar treatment performance to Pilot Study 1 for TOC, UVT, TKN, and nitrate (Figure 9, Figure 10, Figure 11 and Figure 12, respectively), with unique insights from Pilot Study 2 for TDS management using HFNF and other membrane projections.

TOC was reduced substantially by 72% across the pilot train to approximately 1.8 mg/L, as shown in Figure 9, with the largest reduction occurring through ozone-BAC. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in TOC concentrations across the treatment train (χ² = 60.44, df = 3, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment confirmed that every treatment stage achieved a statistically significant reduction in TOC from the preceding stage (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = b; Ozone-BAC Effluent = c; GAC Effluent = d).

UVT also improved, as shown in Figure 10, which shows an increase from approximately 74% in the AWT feed water to 94% after ozone and BAC and to 96% in the final effluent. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in UVT concentrations across the treatment train (χ² = 73.60, df = 3, p < 0.0001). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment confirmed that every treatment stage achieved a statistically significant increase in UVT from the preceding stage (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = b; Ozone Effluent = c; BAC effluent = d, GAC Effluent = e).

TKN was reduced to the MDL of 0.5 mg/L in the GAC contactor effluent, with CFS and ozone-biofiltration both being major drivers behind TKN removal. The box encompasses the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles; points beyond the whiskers represent outliers. The Kruskal-Wallis test indicated a significant difference in TKN concentrations across the treatment train (χ² = 13.25, df = 3, p < 0.0041). Post hoc pairwise analysis using Dunn’s Test with a Holm adjustment was used to evaluate TKN changes between stages (p < 0.05). Lowercase letters in parentheses indicate statistically unique groups based on pairwise comparisons; treatment stages sharing no common letters are significantly different (p < 0.05; AWT Feed = a; Settled Water = ab; Ozone-BAC Effluent = b; GAC Effluent = b).

As in Pilot Study 1, nitrate (NO3-N) removal was insignificant during Pilot Study 2, as expected with CBAT processes, and ranged from 6 to 15 mg/L throughout the pilot, compared to the MCL of 10 mg/L (Figure 12).

The average pilot feed and GAC effluent concentrations for conventional water quality parameters, along with the percent change in the concentrations, for Pilot Study 2 are presented in

Table 8. Pilot Study 2 conventional water quality parameters and percent reductions. Data includes samples collected during challenge testing, following CFS and BAC optimization.

Parameter	Units	Average AWT Feed Value	Standard Deviation (Sample)	Average GAC Effluent Value	Standard Deviation (Sample)	% Change
Conductivity	mS/cm	0.925	0.03	0.918	0.03	−1%
Turbidity	NTU	0.86	0.31	0.09	0.01	−90%
UVT-254	%	74	0.01	96	0.01	+30%
DO	mg/L	5.64	1.16	12.92	2.30	+129%
pH	N/A	7.02	0.08	6.89	0.14	−2%
Total Alkalinity	mg/L-CaCO3	130	11.53	127	15.54	−2%
Nitrate	mg/L-N	12.0	2.89	12.1	2.13	+1%
Nitrite	mg/L-N	0.054	0.02	0.000	0.00	−100%
Ammonia	mg/L-N	0.100	0.05	0.015	0.00	−85%
Total Phosphorus	mg/L-PO4	0.801	0.20	0.183	0.06	−77%
Soluble Reactive Phosphorus	mg/L-PO4	0.541	0.21	0.193	0.06	−64%
Total COD	mg/L	20.3	2.50	5.8	1.79	−72%
TDS	mg/L	565	48.6	554	41.7	−2%
TSS	mg/L	4.0	0.00	4.0	0.00	0%
CBOD	mg/L	2.3	0.55	2.0	0.00	−12%
TKN	mg/L	1.20	0.34	0.5	0.00	−58%
DOC	mg/L	6.0	0.40	1.9	0.18	−69%
TOC	mg/L	6.3	0.38	1.8	0.18	−72%

. Data include occasional replicate sampling that was conducted (sampling at the same location and time for a particular parameter). Results equal to the MDL are reported as the MDL in this table.

In addition to conventional parameters, 192 unique constituents were screened in Pilot Study 2 to assess whether the treatment performance was similar to Pilot Study 1, and the testing included a similar list of parameters to those in Pilot Study 1. Throughout testing and sampling, 169 of the measured parameters were not detected (<MDL) in the AWT feed water. The non-detected contaminants included all tested total PCBs, VOCs, and microcystins, as with Pilot Study 1. Of the measured parameters, 23 were above their MDLs in the AWT feed water and included certain SOCs, PPCPs, halogenated DBPs, N-DBPs, radiological particles, and pesticides and herbicides.

Table 9. Constituent detections above their MDLs in the screening study for the AWT Feed of Pilot Study 2, excluding conventional parameters. Note that some MDL values varied for samples taken at different time points due to outside laboratory calibration protocols.

Detections in AWT Feed Water	Category	Average Result	Units	No. of Samples	MDL
1,4-Dioxane	SOCs	1.210	ug/L	2	0.07
Atenolol	PPCPs	0.070	ug/L	2	0.01
Atrazine	Pesticides and Herbicides	0.405	ug/L	2	0.07
Azithromycin	PPCPs	0.365	ug/L	2	0.01
Carbamazepine	PPCPs	0.114	ug/L	2	0.01
Chloroform	Halogenated DBPs	0.665	ug/L	4	0.50
Cimetidine	PPCPs	0.093	ug/L	2	0.01
Diltiazem	PPCPs	0.092	ug/L	2	0.01
Diphenhydramine	PPCPs	0.089	ug/L	2	0.01
Erythromycin	PPCPs	0.029	ug/L	2	0.02
Fluoxetine	PPCPs	0.067	ug/L	2	0.01
Gross Beta	Radiological	11.220	pCi/L	2	4.00
Meprobamate	PPCPs	0.022	ug/L	2	0.01
Metformin	PPCPs	0.138	ug/L	2	0.02
Metolachlor	Pesticides and Herbicides	0.315	ug/L	2	0.20 or 0.19
NDMA	N-DBPs	6.725	ng/L	4	2.00
NMOR	N-DBPs	9.500	ng/L	4	2.00
Primidone	PPCPs	0.195	ug/L	2	0.01
Simazine	Pesticides and Herbicides	0.054	ug/L	2	0.05
Sulfamethoxazole	PPCPs	0.440	ug/L	2	0.01
TCEP	SOCs	0.073	ug/L	2	0.01
TCPP	SOCs	0.880	ug/L	2	0.40
Trimethoprim	PPCPs	0.190	ug/L	2	0.01

includes the average, standard deviation, and numbers of samples for each of the AWT feed detections. Notably, statistical analyses could not be performed on the parameters where only a small number of samples (<3) were collected in this screening evaluation. Eight parameters were detected in the final treated water, including the following:

• Halogenated DBPs: Bromate (detailed later) and very low levels of chloroform (CHCl₃).
• SOCs: Low levels of 1,4-dioxane that were mostly removed by ozone.
• PPCPs: Acetaminophen and trimethoprim were found near their MDL values, and metformin was also detected.
• N-DBPs: NMOR was detected in the final effluent at approximately 7 mg/L.
• Radiological: Gross Beta particles were detected once at 7 pCi/L, under the EPA screening trigger level of 50 pCi/L.

The CBAT pilot results from the screening study in Pilot Study 2 indicated similar treatment performance to Pilot Study 1 with effective removal of various compounds, including low levels of 1,4-dioxane, a commonly monitored constituent in AWT trains. 1,4-dioxane detected in the AWT feed water was reduced by ozone to below the EPA health advisory level of 0.35 ug/L [26], as shown in Figure 13.

Data shown in Figure 13 compares pre-ozone sampling (AWT feed water or settled water, n = 4) with post-ozone sampling (ozone effluent, n = 3) for the duration of the pilot, with the transferred ozone/TOC ratio ranging from 0.5 to 1.2. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles. The Kruskal-Wallis test indicated no statistically significant difference in 1,4-dioxane concentrations from pre-ozone to post-ozone (p = 0.074). While a substantial physical reduction was observed following ozone treatment, the small sample size (n = 3 paired dates) limited the statistical power of the analysis.

Additionally, N-nitrosodimethylamine (NDMA), a N-DBP relevant in water reuse applications, was found to increase after ozone to an average of 29 ng/L, then was effectively reduced by BAC to below the MDL and below the California notification level of 10.0 ng/L and the California Public Health Goal of 3.0 ng/L [27], as shown in Figure 14. Data shown in Figure 14 are for the duration of the pilot, with the transferred ozone/TOC ratio ranging from 0.5 to 1.2. The boxes shown in the figure encompass the IQR, with the central horizontal line indicating the median and “x” denoting the average. Whiskers extend to the furthest data within 1.5 times the IQR of the quartiles. The Kruskal-Wallis test indicated a statistically significant difference in NDMA concentrations across the treatment train (χ²= 12.14; df = 4; p = 0.016). However, individual pairwise comparisons could not be distinguished (p > 0.05) due to the conservative nature of the Holm correction factor. This lack of distinct grouping is likely due to a low sample size, as unadjusted comparisons indicated a significant drop in the NDMA concentration following ozone treatment (p = 0.0064).

With exception of nitrate and bromate, all parameters during Pilot Study 2 were measured in the GAC effluent below their respective U.S. EPA Drinking Water MCLs. Bromate was formed by ozone above the EPA drinking water MCL to approximately 30 µg/L average, and with a high event of 65 ug/L (manuscript in preparation).

Bench-scale exploratory HFNF experiments were performed to provide insights for TDS management, which may be needed for certain treatment goals or end-uses, including meeting the secondary EPA MCL (SMCL, which is a non-enforceable drinking water guideline for contaminants that mainly affect the aesthetic quality of drinking water rather than direct health impacts) of 500 mg/L. During the pilot, TDS typically ranged from 500 to 650 mg/L throughout testing (manuscript in preparation). The bench-scale HFNF experiment using the NX Filtration dNF80 was performed with pilot settled water and an effective recovery of 80%. The settled water TDS was reduced from 576 mg/L to 490 mg/L (a 15% reduction) in the effluent permeate. The total hardness was also reduced, as well as various ions that were tested (manuscript in preparation). Sulfate, a larger ion, was reduced from 110 mg/L in the pilot settled water to 43 mg/L after bench-scale HFNF, as shown in Figure 15. The concentrate from the HFNF showed a large increase in ions such as sulfate, as expected.

In addition to bench-scale HFNF testing, membrane performance modeling was conducted for three different NF products and one RO product. For the 800 Dalton HFNF (NX filtration dNF80), the bench-scale versus projected results for permeate water values were consistent for TDS, total hardness, and sulfate, as shown in Figure 16. The bench-scale permeate alkalinity value was higher than the influent (pilot settled water) value and the projected value, and additional testing would be warranted to assess full-scale impacts on treated water.

Projections to assess the potential performance with other membrane sizes and products (not used during bench-scale testing) showed various reductions in TDS, total hardness, and sulfate, as shown in Figure 17. Projections show greater reduction with tighter membrane sizes, as expected, with higher to lower permeate concentrations occurring in order from the dNF80, the dNF40 (NX Filtration HFNF with a 400 Dalton cutoff size), NF270 (DuPont NF270 spiral wound membrane), and the BW30 (DuPont RO membrane). While the modeled dNF80 reduced TDS from approximately 576 to 498 mg/L (consistent with bench-scale dNF80 results of 490 mg/L in the permeate), the tighter HFNF membrane, dNF40, projected further TDS reduction to 430 mg/L, which is under the SMCL by a margin of 70 mg/L. The NF270 projected effluent TDS is 380 mg/L. The projections using the BW30 RO membrane showed a very large reduction in TDS, as expected, with an effluent concentration of 26 mg/L. This would notably produce a high concentrate, or brine stream.

3.3. Stakeholder Engagement

For both pilot studies, multiple site visits were conducted with a broad range of stakeholders including regulators, operators, engineers, and other community members. The pilot served as a mechanism for building trust, transparency, and shared ownership of the project. The site visits included tours led by the project team to establish technical credibility, operational realism, regulatory openness, and a tangible public experience. For these Midwest CBAT pilots especially, the events positioned the pilot as a practical, regionally appropriate solution to water resource challenges, not just a research exercise.

Consistent with other pilot studies, results from the site visits at both pilots were positive with narrative objectives being met. Feedback from stakeholders, while anecdotal, show that the efforts were successful in building regulator confidence, increasing utility leadership support, introducing non-potable and potable reuse concepts to the visitors, demonstrating pilot performance, and building momentum for future funding/permitting outcomes. From a stakeholder perspective, CBAT must demonstrate consistent, reliable removal of contaminants while remaining cost-effective and practical to operate within a robust, multiple-barrier treatment system. Notably, regulators prioritize public health protection and validated performance, while operators focus on operational reliability and lifecycle costs, and the public is primarily concerned with drinking water safety, affordability, and transparency in monitoring that builds confidence in water reuse.

Further, water reuse projects in regulation-deficient areas of the U.S. Midwest are hindered by unclear or lacking regulations, limited funding, aging infrastructure, and public skepticism, all of which increase implementation risk and discourage investment. Regional uptake can be enhanced through state reuse policies, coordinated permitting, financial incentives, pilot demonstration projects, and transparent science communication that emphasizes water quality, public health protections, and the economic and environmental benefits of reuse.

4. Discussion

4.1. Operational Performance Analysis of Pilot Studies

Consistent with other studies [13,28,29], results from these pilot studies demonstrated that the secondary effluents from both WRFs are potentially viable source waters for reuse when treated through CBAT. Conducting the pilot studies at WRFs with significant differences in treatment plant capacity and wastewater collection system characteristics was a unique opportunity to demonstrate the flexibility of CBAT for water reuse applications. While both WRFs utilize conventional activated sludge secondary treatment, the WRF in Pilot Study 1 has a permitted average design flow of 10 MGD and a collections system serving primarily residential areas. Conversely, the WRF in Pilot Study 2 has a permitted average design flow of 114 MGD and a peak hydraulic capacity of 330 MGD, with a combined sewer collections system serving both residential and industrial customers.

All parameters during both Pilot Study 1 and 2 were measured in the GAC effluent below their respective U.S. EPA Drinking Water MCLs, except for nitrate, which is discussed further in Section 4.2, and bromate in Pilot Study 2. During Pilot Study 2, bromate formation was higher due to substantial levels of bromide detected, typically 100 to 150 µg/L (manuscript in preparation), as bromate formation during ozonation increases with the bromide concentration [30]. Elevated bromate formation has been observed at bromide concentrations on the order of tens of µg/L in drinking water [31,32].

The similarities in treated GAC contactor effluent water quality between the two pilots underscore the effectiveness of CBAT for potential users with a wide range of community characteristics. In both pilot studies, GAC effluent TOC concentrations were consistently below 2.0 mg/L and GAC effluent UVT averaged 96%. TKN concentrations were effectively reduced to the MDL of 0.5 mg/L in both pilot studies as well. The equipment issues experienced during Pilot Study 2 with the ozone mass flow controllers and air conditioning system may have limited the full performance potential of BAC during Challenge Testing due to the shorter period to acclimate BAC to Pilot Study 2 ozonated water. During the ozone outage, the BAC filters continued operating using settled water as the feed. After the ozone outage, BAC was fed ozonated water for approximately one week during the CFS Optimization phase, prior to the start of the Challenge Testing phase. The treatment effectiveness for most parameters, including TOC, UVT, and TKN, was assessed during the Challenge Testing phase for Pilot Study 2.

On an individual treatment process level, in both pilot studies the CFS step was a significant driver behind the reduction in solids and organic compounds, such as carbonaceous biochemical oxygen demand (cBOD) (manuscript in preparation) and TOC. Ozone-biofiltration was also a significant driver behind the reduction in organic compounds such as TOC, ammonia conversion, and TKN reduction, and certain emerging contaminants including 1,4-dioxane. TKN concentrations in the BAC effluent from these two pilots (averaging 0.55 mg/L for each pilot) were between the BAC effluent TKN concentrations observed at SWIFT (approximately 0.75 mg/L, [29]) and Franklin, TN (approximately 0.15 mg/L, [33]). The GAC process provided further organics removal, as indicated by the increase in UVT between the BAC effluent and the GAC effluent, with an additional focus on reduction in emerging contaminants such as PFAS and PPCPs. TOC concentrations in the GAC effluent from these two pilots were lower than the GAC effluent TOC concentrations observed at SWIFT (approximately 3.0 mg/L, [29]) and Franklin, TN (approximately 3.2 mg/L, [33]), and similar to those observed at Gwinnett County (less than 2.0 mg/L, [34]). Results of PFAS removals through this process in Pilot Study 1 importantly highlight the potential variability in contaminant loads to the WRF, recognizing that PFAS are not well removed through conventional wastewater treatment processes [35,36]. Regardless, the overall results are similar to those reported in other research indicating that, when GAC is part of a CBAT process, excellent PFAS removal can be achieved [37]. The results of PFAS from pilot testing can be used to project the expected GAC lifetime in a full-scale application with respect to the required media regeneration frequency to meet continuous PFOA and PFOS control [38,39,40,41].

In CBAT, as with any other advanced treatment scheme, a single treatment process is not expected to remove every compound; individual treatment processes complement one another to provide a multi-barrier approach combining physical, chemical, and biological processes. The combination of treatment processes operating together demonstrated the ability to remove constituents which the WRFs were not designed to remove, and to further decrease constituent concentrations to levels which would not typically be expected through traditional secondary wastewater treatment. These results support those presented by Liu et al. [42] in exhibiting the ability of ozone followed by BAC to reduce concentrations of emerging contaminants, particularly PPCPs, and Vaidya et al. [29] in confirming the nutrient removal capabilities of CBAT for reuse applications. Utilizing CBAT, the two pilot studies demonstrated that secondary clarifier effluent from both WRFs can be treated to meet water quality objectives for non-potable reuse and indirect potable reuse.

4.2. Engineering Challenges and Future Research

Notably, NDMA formation following ozone was observed in Pilot Study 2, which is a complex reaction pathway. Ozone can either reduce precursor compounds or transform precursors into more reactive compounds that later form NDMA [43]. Here, NDMA appeared to form with ozonation and was effectively removed by the downstream biofiltration process. It is notable that NDMA itself is difficult to degrade because it is a very small, highly soluble, and relatively biologically recalcitrant compound, although some specialized microorganisms can biodegrade NDMA under certain conditions. However, rates are often slow, and removal can be inconsistent, with treatment performance being sensitive to temperature and acclimation. Because of this, careful consideration regarding the design of biofiltration as a primary NDMA barrier is important for full-scale applications [44]. Specific to Pilot Study 1, two N-DBPs were detected in the GAC contactor effluent: NMOR and NDPA. Future research on management strategies for these compounds to satisfy EPA drinking water MCL goals is warranted if treated water is used for future potable reuse applications.

Specific to Pilot Study 2, bromate was detected in the ozone effluent at concentrations greater than the EPA drinking water MCL of 10 µg/L due to higher bromide levels reacting with ozone, whereas the AWT feed water had bromate concentrations below the EPA drinking water MCL. Future research at this WRF could focus on strategies for limiting bromate formation, such as utilizing side-stream injection, dosing hydrogen peroxide, or operating at a lower pH. Additionally, the duration of operation for Pilot Study 2 did not capture a substantial wet weather event, which would be of particular interest due to the combined sewer collections system which contributes flow to this WRF. The impacts of wet weather flow and upstream treatment at the WRF should be investigated to further understand the secondary effluent water quality and treatment performance of the CBAT train.

Findings from both pilot studies inform areas for future research focus at both WRFs. For example, while ammonia conversion was successful, nitrate was not removed through the CBAT train; thus, this process approach fits best in advanced treatment trains where upstream wastewater treatment provides denitrification to below the EPA drinking water MCL of 10 mg/L as N, or where water is blended with other sources to reduce nitrate [13]. The addition of an internal mixed liquor recycled at the upstream WRFs could reduce the nitrate concentration if determined to be necessary as part of a future reuse scheme. Should a blending approach be utilized, a simple mass balance calculation such as the one demonstrated below can be performed to determine the appropriate ratio of flows from the various sources of water, depending on their respective nitrate concentrations.

$$\left(Q_1\times C_1\right)+\left(Q_2\times C_2\right)=\left(Q_1+Q_2\right)\times C_3$$

where

• Q₁ = Flow from source 1;
• C₁ = Concentration in source 1;
• Q₂ = Flow from source 2;
• C₂ = Concentration in source 2;
• C₃ = Concentration in combined flow stream.

Using Pilot Study 2 as an example, let the nitrate concentration in the GAC effluent equal C₁ (12.1 mg/L). Assume a target nitrate concentration of 8 mg/L (C₃) in the combined flow stream to provide a buffer below the drinking water MCL of 10 mg/L. Further assume that a source of water for blending has been identified with a nitrate concentration of 2 mg/L. The mass balance equation above can be rearranged to solve for the ratio of flows from the two sources, Q₁/Q₂, as shown:

$${\displaystyle \frac{Q_1}{Q_2}}={\displaystyle \frac{C_3-C_2}{C_1-C_3}}={\displaystyle \frac{8-2}{12.1-8}}=1.46$$

TDS is another parameter which saw no significant change in concentration across either pilot, which is to be expected since CBAT does not include a barrier for TDS removal. HFNF, which is generally less costly to operate over time than traditional RO or standard NF for several reasons, was tested at bench-scale with comparative projections as an exploratory effort to address this issue. Notably, HFNF has lower energy requirements due to the slightly larger pore sizes than RO membranes, allowing water to pass through at significantly lower pressures, directly reducing electricity consumption, resulting in 30–50% lower energy compared to standard RO [45]. Pretreatment requirements are lower for HFNF, which only reduces multi-valent ions (like calcium and magnesium) but selectively allows some monovalent ions to pass, resulting in less membrane scaling; this also lowers the need for expensive anti-scalant chemicals [45]. HFNF also has higher water recovery rates—generally in the range of 80–95%, meaning less “reject water” is lost. Finally, HFNF has longer membrane life due to less strain on the membranes from lower operating pressures—combined with reduced chemical scaling—which leads to less frequent membrane fouling and replacement costs over the life of the system. However, the suitability of HFNF depends entirely on the specific water treatment goals, as it does not strip out TDS as completely as RO [45].

The bench scale HFNF test conducted during Pilot Study 2 with a dNF80 membrane demonstrated 15 percent TDS reduction from the AWT feed water, with greater reduction for larger ions such as sulfate (61 percent). These results were supported by follow-up nanofiltration and RO modeling, which also indicated that additional reduction could be achieved with a tighter HFNF membrane. While the results were promising, the extent of the bench-scale testing at one facility is insufficient to establish the operational adaptability of the technology and warrants further research. If greater TDS removal is required to achieve water quality objectives, there are a suite of solutions to remove ions. Various HFNF membranes could be used to lower TDS by selective ion removal; however, if more substantial reductions in TDS are required, it may be necessary to address this through high pressure membranes such as RO, which has limited concentrate disposal options in the Midwest U.S.

Finally, with the positive outcomes from the stakeholder engagements, both utilities where pilot studies were conducted are moving forward with additional investigations into water reuse. This is critical because population growth, increasing water demands for data centers, and need for more sustainable water practices are driving a clear need for water reuse. This study also demonstrates that it is possible, both technically and from a stakeholder perspective, to leverage this non-traditional CBAT train for water reuse at both large and small WRFs in the U.S. Midwest.

5. Conclusions

In conclusion, the two field pilots were successfully operated to demonstrate preliminary data to serve as proof of concept for justification of additional investment in CBAT for water reuse in the Midwestern U.S. While the projects were successful from the perspective of meeting most drinking water MCLs, nitrate is an inherent limitation in the CBAT process which must be addressed outside of the CBAT train through denitrification of the source water or blending with a low nitrate source. Thus, while this preliminary pilot testing was useful as a tool for developing data and for serving as a tool in stakeholder engagement, additional research is needed to fully assess CBAT for application in the U.S. Midwest, including as part of future regulatory efforts. Topics that need additional geographically specific data are centered around several key areas:

• Ozone demand and dose optimization are needed, given the variability of water quality and flows in this region, particularly in combined sewer communities. This includes optimization of reagent dosages, long-term operational safety and operational adaptability, which should be studied in subsequent specialized tests.
• Confirmation of appropriate ozone/TOC ratios for nitrified systems, given that typical ozone/TOC ratios ranging from 0.5 to 1.0 are common starting points.
• Future research should assess whether higher ozone dosing, necessary in wastewater effluent, can be conducted without exceeding bromate standards in the US EPA drinking water regulations. Additionally, shifts in DBPs should be documented, considering that bromate formation has a higher probability in the absence of ammonia in fully nitrified effluents. This should be paired with investigations into low-cost mitigation approaches to ozonation by-products for high-bromide inland water, which should at a minimum include source control, given that industrial discharges are potentially key contributors to bromide in wastewaters.
• Conversely, in wastewater effluent, which can be high in organic nitrogen precursors, ozonation can lead to an increase in NDMA. Therefore, research must determine whether BAC can consistently biodegrade this DBP and investigate the importance of this and other NDBPs in the context of public health risks.
• Additional investigations are needed to assess GAC adsorption cycles and process stability under seasonal water temperature fluctuations.
• Further testing on the HFNF approach to ion removal should be conducted if TDS is a constituent of concern, depending on the ultimate end use of the reclaimed water, to establish the scalability and adaptability of the treatment technology.
• Final disinfection was not evaluated during these pilot studies. However, the effectiveness of UV, ozone and chlorine disinfection is well established for drinking water, wastewater, and water reuse applications, and disinfection would be included in any future full-scale AWT design. Incorporation of disinfection may also enhance demonstration projects that include public stakeholder engagement, such as tasting stations, that are intended to support public outreach and acceptance [46].
• Ongoing research and future demonstration testing should also include opportunities for ongoing stakeholder engagement to help utilities as this innovative approach to water reuse in the U.S. Midwest is increasingly examined as an element in a diversified water supply portfolio.