PFAS Removal in AWT: Using RSSCT to Accelerate GAC Design

Introduction Emerging contaminants like per- and polyfluoroalkyl substances (PFAS) have gained attention in the past 5 years due to anticipated health concerns and a rapidly changing regulatory climate. The draft EPA regulations propose MCLs of 4 ppt for PFOA and PFOS, and a hazard index of 1 for PFHxS, HFPO-DA, PFNA, and PFBS. Additionally, the upcoming fifth Unregulated Contaminant Monitoring Rule (UCMR5) is focused on sampling for 29 different PFAS with greatly improved analytical methods. It is expected that >20% of all drinking water utilities will have reportable levels of PFOA and PFOS, based on past data sets. The estimated national capital costs for drinking water utilities to adequately remove these compounds are anticipated to exceed $200M net, with similar life cycle/replacement costs incurring on an annual basis. All these factors have made it essential to act now for preventing PFAS exposure and finding effective ways to treat PFAS in water. Drinking water utilities have been stressed to rapidly install reliable and cost-effective PFAS treatment technologies. In many cases, customer demands have pressured utilities to identify and implement these solutions as quickly as possible. Problem Granular activated carbon (GAC) can be highly effective for removal of total organic carbon (TOC) and PFAS. With water-stressed communities across the U.S. shifting towards non-traditional supplies, such as potable reuse and impaired groundwaters, and the growing national concern around PFAS and other emerging contaminants, the water industry is looking for reliable, rapid implementation strategies for proven technology solutions, like GAC. Currently, data from Rapid Small Scale Column Testing (RSSCT), is used for predicting full-scale GAC performance by scaling up the adsorption data. RSSCT can enable bench testing different GAC media in a short amount of time and provide fast results using less resources (media, water and labor). This method is faster and cheaper than pilot or demonstration-scale systems. However, RSSCT only provides qualitative insights on performance and lacks the sensitivity to characterize the impacts of media size, pretreatment, and moderate source water quality fluctuations. The main objectives of this study were: (1) To determine the best GAC media for PFAS and TOC removal, (2) To determine bed volumes needed for PFAS and TOC breakthrough, (3) To calculate GAC media changeout frequency, (4) To develop a tool for predicting GAC breakthrough for full-scale GAC design Methods The bench and pilot-scale work were conducted at Anne Arundel County, MD's advanced water treatment (AWT) pilot. This AWT pilot is being used for validating a carbon-based treatment approach for managed aquifer recharge. Figure 1 shows the process flow diagram of the carbon-based pilot treatment train. This AWT pilot was installed at the Patuxent WRF in September 2022 and consists of coagulation, flocculation, sedimentation, ozone, biologically activated carbon filtration (BAF), GAC adsorption, and UV disinfection. The GAC treatment acts as a final barrier for removing contaminants that are not removed in the upstream treatment processes. RSSCT was conducted using small glass columns of varying diameters (7, 10 and 15 mm) filled with GAC media from two different vendors Calgon (Filtrasorb(R) 400) and Evoqua (UltraCarb(R) 1240LD10). The media was grounded and packed in these glass columns. Two different empty bed contact times (EBCT) were tested, 6.5 and 12 min. BAF effluent from the pilot study was used for RSSCT. PFAS and TOC samples were collected at 5,000, 10,000, 20,000, 30,000 and 40,000 bed volumes. This setup allowed to see the impact of various media types, media sizes and EBCT on the PFAS removal and breakthrough. Results and Discussion Pilot-scale data showed that only short-chain unregulated PFAS such as PFPeA, PFHxA and PFBA are being detected in GAC effluent after 6,000 BV as shown in Figure 1 and 2. The long-chain PFAS such as PFOA and PFOS that might be regulated by EPA were removed by GAC below detection even after 12,000 bed volume (BV) at an EBCT of 30 min. At 12,000 BV, TOC removal was 60%, with an effluent TOC level of 1 mg/L. Figure 3 shows the RSSCT PFAS and TOC data for one column. Short-chain PFAS such as PFPeA and PFHxA similar to the pilot study were detected in RSSCT. This data showed that GAC sites were exhausted by TOC and removal plateaued around 15,000 BV. Significance and Conclusion This data is important for GAC media changeout which can be a cost-intensive process due to cost of GAC media. This data can help in deciding if media changeout is dominated by TOC limits for potable reuse or by PFAS regulations. Next steps involve comparing RSSCT PFAS breakthrough data with full-scale GAC PFAS data. This will allow to understand the discrepancies in predicting PFAS breakthrough for RSSCT. Using this comparison, a predictive tool can be developed for full-scale GAC design. Using RSSCT scale up models, we can predict the bed volumes needed for breakthrough of PFAS and CECs to help design full-scale GAC treatment. Developing modified RSSCT scale up models for different contaminants can ensure consistent and reliable data for the industry. Optimizing GAC performance using real time RSSCT tools can revolutionize the way GAC treatment is implemented in the One Water industry.