F-0249 Eielson PFAS TS Report_FINAL_May-2019.pdf

Bench-Scale Study for Volume Reduction of PFAS Contaminated Soils at

Eielson Air Force Base, Alaska

Contract Number: W9128F-18-D-0020 Task Order Number: W9128F-18-F-0249

Prepared For:

Omaha District 1616 Capitol Avenue

Omaha, NE 68102-4901

FINAL

May 2019

Prepared By:

Brice Engineering, LLC 3800 Centerpoint Drive, Suite 417

Anchorage, Alaska 99503

907.275.2899 PH

www.BriceEng.com http://www.briceeng.com/

Bench Scale Study – PFAS Soil Volume Reduction Eielson AFB, Alaska i

SIGNATURE PAGE

Bench-Scale Study for

Volume Reduction of PFAS Contaminated Soils at

Eielson Air Force Base, Alaska

Prepared For:

U.S. Army Corps of Engineers Omaha District

1616 Capitol Avenue Omaha, NE 68102-4901

Contract Number: W9128F-18-D-0020 Task Order Number: W9128F-18-F-0249

This document has been prepared by Brice Engineering, LLC. The material and data in this report were prepared under the supervision and direction of the undersigned.

Steven R. Becker, CEP Project Manager

Jamie Oakley, PG President

Eielson AFB, Alaska iii

TABLE OF CONTENTS

Abstract ........................................................................................................................................... vii

Introduction

Background

Materials and Methods

Soil Collection Physical Treatment Procedure Laboratory Analyses Data Analysis

Results

Soil Characterization Laboratory Analysis

PFAS in Soils PFAS in Process Water PFAS Mass Balance PFAS Leachability

Quality Assurance and Quality Control Laboratory Data Statistical Analysis

Discussion

PFAS in Soil Fractions Post-Processing Reconstituted Soil Results Process Water Leaching Removal Efficiency Data Quality

Findings and Recommendations

Findings Recommendations

Conclusion

References iv Bench Scale Study – PFAS Soil Volume Reduction Eielson AFB, Alaska

LIST OF REFERENCES

LIST OF TABLES

Table 1. Particle Size Distribution & Soil Characteristics Table 2. ADEC Regulated and Proposed PFAS Homologues in Soil Fractions (µg/kg) Table 3. ADEC Regulated and Proposed PFAS Homologues in Process water (µg/L) Table 4. Total PFAS Recovery by Soil Fraction (µg) Table 5: Leaching of ADEC Regulated and Proposed PFAS Homologues from Soil Fractions (µg/L) Table 6. Total PFAS Comparisons not Meeting 95% Confidence Threshold Table 7. Post-Processing Exceedances of Proposed ADEC Cleanup Levels Table 8. Post Processing Reconstituted Soil Results with All Fractions Table 9. Post-Processing Reconstituted Soil Results with <200# Mesh Removed Table 10. Post-Processing Reconstituted Soil Results with 20% of <200# Mesh Retained Table 11. PFAS Removal Efficiency (µg/kg)

LIST OF FIGURES

Figure 1. Particle Size Distribution and Moisture Holding Capacity Figure 2. Total Organic Carbon of Bulk Soil and Fractions Figure 3. PFAS Distribution in Bulk Soil and Fractions Figure 4. PFAS Homologues Leached by 0.1M MgSO4 Figure 5. PFAS Homologues Leached by Deionized Water Figure 6. Percent of PFAS Subject to Leaching Figure 7. PFAS Leaching Normalized to Soil Mass

APPENDICES

Appendix A: Data Summary Tables Appendix B: DoD ELAP Certification Appendix C: Level 2 Laboratory Report

Eielson AFB, Alaska v

ACRONYMS

°C degrees Celsius AAC Alaska Administrative Code ADEC Alaska Department of Environmental Conservation AFCEC Air Force Civil Engineer Center AFFF Aqueous Film-Forming Foam Brice Brice Engineering, LLC CL Cleanup Level DI Deionized Water DoD Department of Defense Eielson Eielson Air Force Base, Alaska ELAP Environmental Laboratory Accreditation Program EPA U.S. Environmental Protection Agency Geosyntec Geosyntec Consultants, Inc.

HDPE High Density Polyethylene LHA Lifetime Health Advisory MgSO4 Magnesium Sulfate ND Non-detected OM Organic Matter PFAS Poly- and perfluoroalkyl substances PFBS Perfluorobutanesulfonic acid PFHpA Perfluoroheptanoic acid PFHxS Perfluorohexanesulfonic acid PFNA Perfluorononanoic acid PFOA Perfluorooctanoic acid PFOS Perfluorooctanesulfonic acid QSM Quality Systems Manual RSL Regional Screening Level TOC Total Organic Carbon UCL Upper Confidence Limit USACE United States Army Corps of Engineers UU/UE Unlimited Use and Unrestricted Exposure vi Bench Scale Study – PFAS Soil Volume Reduction

Bench-scale study for volume reduction of PFAS contaminated soils at Eielson Air Force Base, Alaska

Steven R. Beckera*, Carl F. Bensonb, Jacques J. Smithc, Duane A. Gravesc, Michael A. Riggled, Paul M. Gedbawd, Jamie J. Oakleya, and Craig J. Jonesb.

aBrice Engineering LLC, Anchorage, Alaska USA; bBrice Environmental Services Corporation, Fairbanks, Alaska USA; cGeosyntec Consultants Inc., Knoxville, Tennessee, USA; dU.S. Army

Corps of Engineers, Omaha, Nebraska USA

*corresponding author; Brice Engineering LLC, 100 E. San Marcos Blvd., Suite 418, San

Marcos, CA 92096, USA. sbecker@briceeng.com. OrciD #0000-0001-8016-4168

This work was supported by the U.S. Air Force, Air Force Civil Engineer Center, and the

U.S. Army Corps of Engineers, Omaha District, under contract #W9128F-18-D-0020, task order #W9218F-18-F-0249.

mailto:sbecker@briceeng.com

Eielson AFB, Alaska vii

Bench-scale study for volume reduction of PFAS contaminated soils at

Eielson Air Force Base, Alaska

Abstract

Investigations conducted by the U.S. Air Force Civil Engineer Center and US Army Corps of Engineers have demonstrated that Eielson Air Force Base has a widespread occurrence of soils contaminated with poly- and perfluoroalkyl substances (PFAS) at levels exceeding 2018 Alaska Department of Environmental Conservation migration-to-groundwater cleanup levels. Construction of infrastructure to support two new squadrons of F-35A Lightning II fighter aircraft is anticipated to generate more than 140,000 cubic yards of PFAS-contaminated soils requiring treatment. Current Air Force policy requires that PFAS-contaminated waste be destructively treated to eliminate long-term liability to the Air Force.

Due to the high miscibility of PFAS in water and their suspected tendency to bind to the finer soil fraction, we hypothesized that a water-based physical separation process would be effective at reducing the volume of PFAS-contaminated soils requiring secondary treatment. Brice Engineering LLC, (Brice) has a water-based technology that utilizes particle size segregation and particle density segregation to target and separate soil fractions containing high levels of contaminants from the bulk soil (U.S. Patents #7,399,141 and #7,255,514). Brice conducted a bench-scale study to test this concept. This study confirmed that PFAS is found predominately in the finer soil fractions as well as supported the potential efficacy of the physical separation process. Based on the results of this bench-scale study, either further investigation to optimize the process for PFAS or a full-scale pilot study is recommended.

Highlights:

• Approximately 73% of recovered PFAS was partitioned in the process water, which would be treated by GAC or IX resin.

• Residual levels of PFOS and PFOA in the treated soil were below 2018 Alaska Department of Environmental Conservation migration-to-groundwater cleanup levels with no secondary treatment required.

• With the exception of the <#200 mesh fraction, residual levels of PFAS homologues were below revised migration-to-groundwater cleanup levels for all six PFAS homologues proposed by the Alaska Department of Environmental Conservation in October 2018. Removal of the <#200-mesh fraction would result in a 70% reduction in the amount of soil that would require secondary treatment under the proposed regulations.

• PFAS retention in soil fractions increases with decreasing particle size.

• Soil organic material is a strong attractant to PFAS, although there was no correlation between TOC and PFAS concentrations for these soils.

• Leachability of PFAS homologues was not affected by particle size.

Keywords: polyfluoroalkyl; perfluoroalkyl; PFAS; PFOS; PFOA; remediation; soil washing; Eielson AFB;

volume reduction; Brice.

viii Bench Scale Study – PFAS Soil Volume Reduction

This page intentionally left blank

Eielson AFB, Alaska 1

Introduction

Brice Engineering, LLC (Brice), in partnership with Geosyntec Consultants, Inc. (Geosyntec), performed a bench-scale study to assess the potential effectiveness of a patented soil washing process to reduce the volume of PFAS contaminated soil requiring destructive treatment at Eielson Air Force Base, Alaska (Eielson). This process utilizes a combination of particle size and particle density segregation to target specific soil fractions that serve as reservoirs for a given contaminant or group of contaminants. The overall objectives of the study were to make a preliminary determination of the soil fraction size or “cut point” above which the processed soil meets criteria for unlimited use and unrestricted exposure (UU/UE) and, based on these results, to predict the potential volume reduction achievable at Eielson with the process.

Background

Per- and polyfluoroalkyl substances (PFAS) have been a primary ingredient in aqueous film-forming foams (AFFF) used by the United States (U.S.) military, civilian airports, and other facilities to extinguish hydrocarbon fires for almost 50 years. Certain PFASs are highly mobile, bioaccumulative, persistent in the environment, and have been shown to pose a human health risk. Investigations conducted by the U.S. Air Force Civil Engineer Center (AFCEC) and the US Army Corps of Engineers (USACE) have determined that Eielson Air Force Base (Eielson) has a widespread occurrence of soils contaminated with PFAS at levels exceeding Alaska Department of Environmental Conservation (ADEC) migration-to-groundwater (MTG) cleanup levels (USACE, 2017; USACE 2018b; USAF, 2018; and others).

Construction of infrastructure to support two new squadrons of F-35A Lightning II fighter aircraft is anticipated to generate approximately 140,000 cubic yards of PFAS-contaminated soils above ADEC MTG cleanup levels. Current Air Force policy requires that PFAS-contaminated solid waste be destructively treated to eliminate long-term liability to the Air Force (G. Fink, personal communication, 2018). These soils are currently being stockpiled on base for future transport to, and destructive treatment at, an EPA-approved facility outside of Alaska (USACE, 2018a). The high cost of soil transport and treatment could be significantly reduced if an on-installation option for volume reduction and/or treatment was developed and approved.

Brice Engineering LLC, (Brice) has access to a soil washing process (U.S. Patents #7,399,141 and #7,255,514) that utilizes particle size segregation and particle density segregation to target soil fractions containing high levels of contaminants (Benjamin, Jones, and Brice, 2008). Due to the high miscibility of PFAS in water and their suspected tendency to bind to the finer soil fraction (Johnson et al., 2007), we hypothesized that this physical separation process would be effective at Eielson in reducing the volume of PFAS-contaminated soils requiring secondary treatment. Brice conducted a bench-scale study (proof-of-concept) to ascertain the efficacy of the physical separation process for volume reduction of these soils, and to assist decision-makers in determining whether further study to optimize the soil washing process for PFAS or full-scale pilot study is warranted.

Applicable cleanup levels for PFAS have been rapidly evolving. The U.S. Environmental Protection Agency (EPA) has recently issued Lifetime Health Advisories (LHAs) for two PFAS homologues, Perfluorooctanesulfonic acid (PFOS) and Perfluorooctanoic acid (PFOA), in drinking water (EPA, 2016) but has not promulgated cleanup levels for any PFAS homologues. The Alaska Department of Environmental Conservation (ADEC) currently has promulgated cleanup levels (CLs) for PFOS and PFOA in both soil and groundwater (ADEC, 2018a). On October 1, 2018, ADEC proposed revised CLs for PFOS and PFOA in soil and groundwater as well as new CLs for four additional PFAS homologues (ADEC, 2018b): Perfluorobutanesulfonic acid (PFBS), Perfluoroheptanoic acid (PFHpA), Perfluorohexanesulfonic acid (PFHxS), and Perfluorononanoic acid (PFNA). The comment period for the draft regulations closed

2 Bench Scale Study – PFAS Soil Volume Reduction

November 13, 2018, and ADEC is currently reviewing comments prior to adoption of the proposed cleanup levels. For purposes of this study, results were compared to both current and proposed ADEC CLs and the EPA LHAs.

Materials and Methods

The bench-scale study was designed to replicate the operation of the Brice volume reduction process in the field. To ensure that the results were applicable to Eielson, PFAS contaminated soil from an existing stockpile on site were used in this study. Three independent trials of were performed for statistical analysis.

Soil Collection

A representative volume of soil was collected by Brice from a PFAS-contaminated soil stockpile resulting from the construction of USACE project EIE 385, F-35A OSS/Weapons/Intel Facility on Eielson. The EIE 385 site contained the highest reported levels of PFAS based on the results of USACE Supplemental Soil Investigations for the F-35A construction effort (USACE, 2017; USACE, 2018b). Collection of the representative volume was based on a field sampling method commonly used by the U.S. Natural Resources Conservation Service (NRCS, 1996):

• Numerous aliquots were randomly collected with a shovel across one exposed face of the stockpile.

Each shovel aliquot ranged in depth from one to two feet.

• Aliquots were placed on a large poly tarp located immediately adjacent to the stockpile.

• All aliquots were homogenized by roll-mixing on the tarp several times in each direction to create a windrow.

• Homogenized aliquots were by placing a shovel flat on the tarp and pushing it into the windrow at random locations to remove aliquots,

• The aliquots were then placed in two (2) new five-gallon high density polyethylene (HDPE) buckets.

The two buckets of soil were sealed and shipped under chain of custody to the Geosyntec material testing facility in Knoxville, TN. After receipt, the soil from the two buckets was re-homogenized by Geosyntec into one large composite in order to address any settling that occurred during shipping. Three bulk soil aliquots (Bulk A, Bulk B, and Bulk C) were then taken from this homogenized composite. For quality control purposes, the Bulk A sample was then split and labeled as Bulk AS. Each of the four discrete bulk soil aliquots (Bulk A, Bulk AS, Bulk B, and Bulk C) were sampled pre-processing, and submitted to TestAmerica Laboratory, West Sacramento, California (TestAmerica Sacramento) to be analyzed for PFAS and Total Organic Carbon (TOC) as described below. PFAS and TOC levels were used to provide a baseline for mass balance determinations as well as to provide indices of process removal efficiency. Bulk soil PFAS results are presented in Table 10.

Physical Treatment Procedure

To replicate the particle size and particle density segregation process, a four-kilogram grab sample of each primary aliquot of contaminated soil (A, B, and C) was collected and processed in accordance with the following procedure:

• Sample was placed in a new five-gallon HDPE bucket with four (4) liters of PFAS-free water (supplied by TestAmerica Sacramento) and sealed with a lid.

Eielson AFB, Alaska 3

• Sample was tumbled for 15 minutes to deagglomerate the soil (disperse clods).

• The entire contents of the bucket (deagglomerated soil and water) was poured onto a sequential series of brass sieves (Humboldt Manufacturing, Inc., Elgin, Indiana) that were placed over a new 5-gallon HDPE bucket to collect the process water. The series of sieves were at the following screen sizes:

o ¾-inch (19mm) o ½-inch (12.5 mm) o #4 mesh (4.75 mm, 0.187 inches) o #10 mesh (2.00 mm, 0.066 inches) o #50 mesh (0.300 mm, 0.012 inches) o #100 mesh (0.150 mm, 0.0059 inches) o #200 mesh (0.075 mm, 0.0029 inches)

• An additional 12L of PFAS-free water was used to wash the deagglomerated soil through the sieves in accordance with ASTM Method D-2487. This wash water was collected in the same container and combined with the process water from deagglomeration. The soil fractions collected on each sieve were allowed to drain to remove free water and achieve field water-holding capacity (i.e., the point at which water is drained from all large soil pores). This water was also collected and combined as part of the process water.

The three discrete and independent trials of the soil washing methodology are designated as trials A, B, and C. The <0.5-inch - #4 mesh and <#4 - #10 mesh fractions were combined in each trial to produce sufficient material for sampling and analysis.

The following samples were collected for PFAS analysis from each of the three independent trials:

• A discrete soil sample of remaining solids from each of the seven sieves (seven samples per trial).

• A leachate sample resulting from combining all remaining solids from each of the seven sieves and leaching the solids overnight by gentle tumbling in either deionized water (DI) or a 0.1M magnesium sulfate (MgSO4) extraction solution (Acros Inc., Geel, Belgium). The resulting leachate sample was clarified by centrifugation (a total of two samples per trial)

• A sample of coarse organic matter (OM), such as small roots, leaf fragments, and other detritus, that were manually removed and combined into a composite from each of the sieves (one sample per trial).

• A sample of the process water was clarified using Superfloc N1986 (Cytec Industries, West Patterson, New Jersey) at a rate of 70 mL of 0.5 percent solution per liter of water; however, some turbidity remained so the process water sample was filtered and analyzed as a filtered sample (one sample per trial).

• The mass of recovered solids on the filter was insufficient to allow for analysis; therefore, some PFAS may have escaped detection in this small fraction of non-settling fines.

The <#50-#100, <#100-#200, and <#200 fractions of the “A” trial were split for quality control.

All samples were handled with stainless steel or brass equipment rinsed with PFAS-free water and contained in new HDPE or polypropylene buckets and bottles. Equipment rinsate, filters used to collect fines, and PFAS-free water were tested to confirm the lack of PFAS contamination originating from laboratory activities.

4 Bench Scale Study – PFAS Soil Volume Reduction

Laboratory Analyses

Gravimetric measurements were made on Ohaus and Mettler Toledo top loading analytical balances. Soil moisture content was measured by drying the samples in tared containers at 105°C per EPA Method 160.3.

TOC was analyzed by EPA Method 9060A.

Samples were analyzed for 21 PFAS homologues using Liquid Chromatography and Tandem Mass Spectroscopy (LC/MS/MS) compliant with Table B-15 of U.S. Department of Defense (DoD) Quality Systems Manual (QSM) Version 5.1. Laboratory analyses were performed by TestAmerica, West Sacramento, California, which is accredited for this method under the DoD Environmental Laboratory Accreditation Program (ELAP) for both soil and aqueous matrices (Appendix B). Samples were shipped to the lab under chain of custody and on ice by overnight courier. The Level 2 laboratory report is provided in Appendix C.

Unused bulk sample material and individual samples not consumed by testing were returned to Eielson for disposal.

Data Analysis

The PFAS analytical data contained many non-detected (ND) results. Of the 21 PFAS homologues in the analytical list, 10 were not detected in any sample and five were detected only in the process water. Only perfluorohexanesulfonic acid (PFHxS) was detected in all soil fractions and process water. In total, 87 percent of the results for the 21 homologues in the various soil fractions were non-detect. Extreme left censoring of the data (a large number of data points below a certain value) complicated analysis and, in this case, the non-detects were valuable observations which need to be included. There are no generally accepted methods for handling data sets with such a high percentage of non-detects; therefore, a stepped approach was adopted based on detections associated with specific PFAS homologues. The rationale used to analyze the PFAS data was as follows:

• Several PFAS homologues were never detected in any solid or liquid sample. These homologues were assumed to be absent from the Eielson soil and were assigned the value of zero.

• All soil fractions were analyzed in triplicate and some fractions also had one sample analyzed as a split sample; therefore, each fraction had three to four values for each PFAS homologue. Some of these sets had a mix of detections and non-detections. In these cases, the Kaplan-Meier method was used to estimate values for the non-detections. The use of Kaplan-Meier on very small data sets is not ideal, but it still provides an estimated value for non-detects although the potential error with the estimate is large, that is, the estimated values have very wide confidence limits.

• For cases where all the soil samples were non-detect for a homologue, but that homologue was detected in the process water, the homologue must have also been present in the bulk soil and presumably in some or all the soil fractions but below the method detection limit. For these cases, the percentage of PFAS in each soil fraction for homologues that were detected was used to define the typical distribution of PFAS in the various fractions. This distribution (percent PFAS in each fraction) was used to estimate likely concentrations for non-detect values in cases where PFAS was detected in the process water but not in the soil fractions.

• In some cases, all samples for a homologue in each soil fraction were detected. In these cases, the detected values were averaged, and a standard deviation and 95 percent confidence interval were calculated.

For mass balance estimates of PFAS contained in each soil fraction compared to the initial bulk soil sample, the results, which were reported as micrograms (µg) per sample, µg/kg, and µg/L, were converted to µg per bulk soil sample based on the mean bulk soil sample size of 3.344 kg dry weight.

Eielson AFB, Alaska 5

Results

Soil Characterization

The bulk soil was characterized by the amount of soil in each particle size fraction and by the water holding (field) capacity of the fractions, as shown in Figure 1 and in Table 1 below.

Table 1. Particle Size Distribution & Soil Characteristics

Soil Fraction Approximate Percentage of Total Soil TOC (mg/kg) Field Capacity

EIE 385 Soils in situ* Eielson Test Sample (Mean Values)

>0.75” 13 13 0 6.3

<0.75” – 0.5” 9 5 0 0.9

<0.5” - #10 Mesh 29 9 0 2.9

<#10 - #50 Mesh 15 3 5,367 15

<#50 - #100 Mesh 11 12 1,933 23

<#100 - #200 Mesh 4 28 1,900 26

< #200 Mesh 19 30 8,200 40 *EIE 385 Geotechnical Data Report (R&M Consultants, 2017)

The sample obtained for this study had a higher content of fines than is typical for the site based on the 2017 geotechnical study of EIE 385. The dominant soil fractions in the sample were >0.5-inch gravels and fines less than #100 mesh. The TOC content of the soil indicates a relatively low organic carbon content of less than one half percent for the bulk soil. Individual fractions with enough mass for analysis are also shown. The small sample size and sample to sample variability results in wide 95 percent confidence limits for the mean (points on the diamonds in Figure 2) and standard deviations (whiskers). The moisture holding capacity generally increased as the particle size decreased. A detailed data summary of the above can be found in Appendix A, Tables A-1, A-2, and A-3.

Laboratory Analysis

PFAS in Soils

Table 2 summarizes the partitioning by soil fraction of the two regulated and four potentially-regulated PFAS homologues potentially regulated by ADEC.

Table 2. ADEC Regulated and Proposed PFAS Homologues in Soil Fractions (µg/kg)

Soil Fraction PFBS PFHpA PFHxS PFNA PFOS PFOA

𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ±

Bulk Soil 0.13 0.008 ND - 4.175 0.324 ND - 2.325 0.217 3.25 0.204

Coarse O.M. ND - ND - 4 * ND - 4.9 * 3.5 *

>0.75” ND - ND - 0.0011 0.0008 ND - 0.0007 * 0.001 *

<0.75”-0.5” ND - ND - 0.0018 0.0009 ND - ND - 0.0014 0.0005

<0.5”-#10 ND - ND - 0.0045 0.0011 ND - 0.0045 0.0014 0.0046 0.0003

<#10-#50 ND - ND - 0.26 0.045 ND - 0.46 0.052 0.21 0.062

<#50-#100 ND - ND - 0.215 0.038 ND - ND - ND -

6 Bench Scale Study – PFAS Soil Volume Reduction

Soil Fraction PFBS PFHpA PFHxS PFNA PFOS PFOA

𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ± 𝑿𝑿 ±

<#100-#200 ND - ND - 0.245 0.023 ND - ND - 0.433 *

<#200 ND - ND - 1.3 0.080 ND - 1.3 0.080 0.835 0.043

Current CL1 - - - - - - - - 3.0 1.7

Proposed CL2 1400 0.24 0.29 0.41 0.53 0.29 *Single detection, unable to calculate BOLD Underline = 95% Upper Confidence Limit (UCL) exceeds current ADEC Cleanup Level Italics = 95% UCL exceeds Proposed ADEC Cleanup Level 1ADEC, 2018a, Table B-1 Method 2 Soil Cleanup Levels 2ADEC, 2018b, Table B-1 Method 2 Soil Cleanup Levels

Figure 3 illustrates the partitioning of all 21 PFAS homologues by soil fraction. The compiled results for each sample from each fraction of the soil produced during the washing exercise are provided in Appendix A, Table A-4. Coarse OM was analyzed separately with results shown in Appendix A, Table A-5.

PFAS in Process Water

Table 3 describes the partitioning of the two regulated and four potentially regulated PFAS homologues in the process water (combined water from sample deagglomeration, fractional sieving, and post wash fractional free water draining). The partitioning of the 21 PFAS homologues in process water for each sample run can be found in Appendix A, Table A-6.

Table 3. ADEC Regulated and Proposed PFAS Homologues in Process water (µg/L) PFBS PFHpA PFHxS PFNA PFOS PFOA

Process water 0.0225 0.007 0.680 0.0039 0.328 0.51

Current CL1 - - - - 0.4 0.4

EPA LHA2 - - - - 0.070* 0.070*

Proposed CL3 400 0.070* 0.070* 0.070* 0.070* 0.070*

% Contribution - 0.5% 44.5% 0.3% 21.4% 33.3% *Limit of 0.070 µg/L is a cumulative limit for the listed PFAS homologues BOLD Underline = exceeds Current ADEC Cleanup Level and LHA Italics Underline = exceeds Proposed ADEC Cleanup Level and LHA Italics = exceeds Proposed ADEC Cleanup Level 1ADEC, 2018a, Table C Groundwater Cleanup Levels

2EPA, 2016

3ADEC, 2018b, Table C Groundwater Cleanup Levels

PFAS Mass Balance

The total amount of PFAS in each fraction was determined based on the laboratory results and were compared to the results from the original bulk soils. The mass per fraction results were added back together and compared to the values for the bulk soil. As shown in Table 4 below, we achieved approximately 87 percent conservation of mass, which is well within accepted standards for mass balance studies (AOAC, 1998). Table A-7 in Appendix A provides a breakdown of percent recovery by PFAS homologue.

Eielson AFB, Alaska 7

Table 4. Total PFAS Recovery by Soil Fraction (µg)

Bulk Soil >0.75” <0.75”- 0.5”

<0.5”- #10

<#10- #50

<#50- #100

<#100- #200 <#200 Process

Water Total

Total PFAS 34.7 0.031 0.017 0.080 0.110 0.136 0.827 3.71 25.2 30.1

% Total PFAS in Fraction <0.1% <0.1% 0.2% 0.3% 0.4% 2.4% 10.7% 72.6% 86.8%

*% recovery figures rounded to nearest 0.1%

Two small fractions are missing from the compilations above. The dispersed (non-settling solids) in the process water and the coarse OM (roots, detritus, etc.) are not included in the mass balance calculations.

As mentioned above, the dispersed solids removed by filtration did not produce enough mass of soil to analyze. Given the high concentrations of PFAS in the coarse OM and the increasing PFAS concentrations by particle size, we expect that these fractions contain a substantial amount of the unrecovered PFAS.

PFAS Leachability

The retention of PFAS in the soil fraction was tested by splitting the fractions and leaching one portion with a 0.1 M magnesium sulfate (MgSO4) solution and one portion with deionized water. In both cases (shown in Figures 4 and 5 respectively), particle size did not seem to appreciably affect which PFAS homologues were leached (note the matching colors in each stacked column). Table 5 summarizes the leaching results for the two regulated and four potentially-regulated homologues.

Table 5: Leaching of ADEC Regulated and Proposed PFAS Homologues from Soil Fractions (µg/L)

Soil Fraction PFBS PFHpA PFHxS PFNA PFOS PFOA

DI MgSO4 DI MgSO4 DI MgSO4 DI MgSO4 DI MgSO4 DI MgSO4

>0.75” ND ND ND ND 0.0039 0.0029 ND ND 0.0017 0.0034 0.0028 0.0034

<0.75”-0.5” ND ND ND ND 0.0067 0.0047 ND ND 0.0022 0.011 0.0053 0.0033

<0.5”-#10 0.0005 ND 0.006 ND 0.018 0.007 0.007 0.0007 0.017 0.0088 0.016 0.0068

<#10-#50 ND ND .0008 0.001 0.012 0.013 0.008 0.0009 0.035 0.028 0.012 0.011

<#50-#100 ND ND ND ND 0.0093 0.0083 ND ND 0.015 0.012 0.0076 0.0062

<#100-#200 ND ND ND ND 0.0096 0.011 ND ND 0.017 0.012 0.007 0.0063

<#200 0.0039 0.0043 0.0016 0.0014 0.14 0.14 0.0016 0.0012 0.14 0.1 0.1 0.1

Current CL1 - - - - 0.4 0.4

EPA LHA2 - - - - 0.070* 0.070*

Proposed CL3 400 0.070* 0.070* 0.070* 0.070* 0.070* *Limit of 0.070 µg/L is a cumulative limit for the listed PFAS homologues BOLD Underline = exceeds Current ADEC Cleanup Level and LHA Italics Underline = exceeds Proposed ADEC Cleanup Level and LHA Italics = exceeds Proposed ADEC Cleanup Level 1ADEC, 2018a, Table C Groundwater Cleanup Levels

2EPA, 2016

3ADEC, 2018b, Table C Groundwater Cleanup Levels

Leaching ranged from approximately 10 to 70 percent of the total PFAS present in the soil fractions. Only the leachate from the <#200 mesh fraction exceeded the LHA and proposed ADEC CLs for groundwater;

however, it should be noted that the <#50-#100 mesh and the <#10-#50 mesh were both within 10 percent of the proposed ADEC CLs.

8 Bench Scale Study – PFAS Soil Volume Reduction

While the <#200 mesh fraction leached the greatest amount of total PFAS, there were several other fractions that leached a higher percentage of their total PFAS (Figure 6). Leachate results of all 21 PFAS homologues tested can be found in Appendix A, Tables A-8 (MgSO4) and A-9 (DI).

Quality Assurance and Quality Control

Laboratory Data

The results presented in this section were generated using industry standard of care and in accordance with the DoD QSM Version 5.1. Although some results were qualified as estimated, no data were rejected and no findings affected data usability (Geosyntec, 2019). Quality control analysis of PFAS-free water, equipment rinsate, and filter blank samples demonstrated that the materials and equipment used were free of detectable levels of PFAS (Appendix A, Tables A-10, A-11, and A-12, respectively).

Statistical Analysis

The results for PFAS in soil fractions did not exhibit a normal distribution (Shapiro-Wilk, p = <0.0001 to 0.0006), and median values per fraction were statistically different (Kruskal-Wallis, p < 0.0001). Individual comparisons by Wilcoxon-Mann-Whitney showed statistical significance (p = <0.0001 to 0.0386) for the majority of comparisons of the bulk soil, the mass balance (recovered) total, and process water total PFAS values with the results of the individual fractions, as well as between some of the individual fractions.

However, the small scale of this bench-scale study, the resulting small number of samples, and the very low concentrations of analytes meant that the experiment lacked the statistical power to determine significance between some of the individual fractions at the 95% confidence level (p < 0.05). The following individual comparisons were not statistically significant:

Table 6. Total PFAS Comparisons not Meeting 95% Confidence Threshold

Comparisons p Value Implication of No Significant Difference

Bulk soil Mass balance 0.7676 Near full accounting of PFAS in mass balance.

Bulk soil Process water 0.6695 Majority of PFAS in bulk soil was recovered in process water.

Mass balance Process water 0.6695 Majority of recovered PFAS was in process water.

> 0.75” <0.75”-0.5” 0.2244

PFAS concentrations between fractions are similar.

> 0.75” <0.5”-#10 mesh 0.2505

> 0.75” <#10-#50 mesh 0.2244

> 0.75” <#50-#100 mesh 0.2505

<0.75”-0.5” <0.5”-#10 mesh 0.1228

*<0.75”-0.5” <#10-#50 mesh 0.0710 Significance at the 90% confidence level indicates that these fractions are likely different.

<0.75”-0.5” <#50-#100 mesh 0.1077

PFAS concentrations between fractions are similar.

<0.5”-#10 mesh <#10-#50 mesh 0.5767

<0.5”-#10 mesh <#50-#100 mesh 0.6695

<#10-#50 mesh <#50-#100 mesh 0.6224

*<#10-#50 mesh <#100-200 mesh 0.0527 Significance at the 90% confidence level indicates that these fractions are likely different. *<#100-#200 mesh <#200 mesh 0.0613

*Significant at p < 0.1 (90% confidence)

Eielson AFB, Alaska 9

Summary tables for the above analyses can be found in Appendix A, Table A-13.

Discussion

PFAS in Soil Fractions

In general, PFAS concentrations increased with decreasing particle size. Following the physical separation process, approximately 11 percent of total PFAS were contained in the <#200 mesh fraction.

Despite the bulk soils substantially exceeding 2018 ADEC CLs for PFOS and PFOA, none of the soil fractions exceeded current regulatory limits post-processing, and only the <#200 mesh fraction approached those levels. Two soil fractions exceeded the proposed ADEC CLs. The <#10-#50 mesh fraction showed an exceedance for PFHxS, and the <#100-#200 mesh fraction had one exceedance for PFOA, as shown in Table 7.

Table 7. Post-Processing Exceedances of Proposed ADEC Cleanup Levels.

Soil Fraction PFAS Homologue Value (µg/kg) Proposed CL (µg/kg) % of Total Soil

<#10-#50 mesh PFHxS 0.26 ± 0.045 0.29 3

<#100-#200 mesh PFOA 0.433* 0.29 28

< #200 mesh

PFHxS 1.3 ± 0.080 0.29

30 PFOS 1.3 ± 0.080 0.53

PFOA 0.835 ± 0.043 0.29

*Single detection

Our study also indicates that coarse OM can sorb high levels of PFAS and may be a substantial reservoir for PFAS in soils. In the test results above, coarse OM contained regulated PFAS in concentrations more than 3 times that of the <#200 mesh fraction. However, no correlation could be determined between TOC in the soil fractions and total PFAS, likely due in part to our limited sample size.

Post-Processing Reconstituted Soil Results

A key factor in this study was determining the soil gradation and the distribution of non-soil constituents such as organic matter in soils from the site. This is a crucial determination because the distribution plays a key role in the requirements for process implementation at the site. By precisely deriving fractional PFAS concentrations and weights this also allows for weight averaging to derive post processing residual PFAS concentrations for the reconstituted soil matrix with a high degree of confidence.

Table 8 below depicts the fractional weights and soil gradation percentages for all soil fractions evaluated.

Also shown are the average post treatment PFAS results by fraction. If all fractions were recombined following processing the reconstituted soil mass meets current ADEC CLs with no volume reduction or secondary treatment required.

Table 8. Post Processing Reconstituted Soil Results with All Fractions

Soil Fraction Dry

Weight (grams)

Soil Gradation

95% UCL PFAS Homologue Concentration in Fraction (µg/kg)

PFBS PFHpA PFHxS PFNA PFOS PFOA

>3/4" 435 13 0 0 0.0019 0 0.007 0.001

<3/4" - 1/2" 172 5 0 0 0.0027 0 0 0.0019

10 Bench Scale Study – PFAS Soil Volume Reduction

Soil Fraction Dry

Weight (grams)

Soil Gradation

95% UCL PFAS Homologue Concentration in Fraction (µg/kg)

PFBS PFHpA PFHxS PFNA PFOS PFOA

<1/2" - #10 mesh 302 9 0 0 0.0056 0 0.0059 0.0049

<#10 - #50 mesh 11 <1 0 0 0.305 0 0.512 0.272

<#50 - #100 mesh 391 12 0 0 0.253 0 0 0

<#100 - #200 mesh 944 29 0 0 0.268 0 0 0.433

<#200 mesh 1015 31 0 0 1.380 0 1.380 0.878

RECONSTITUTED SOIL PFAS

CONCENTRATIONS* (µg/kg)

0 0 0.5379 0 0.4307 0.3991

Current CL1 (µg/kg) 3.0 1.7

Proposed CL2 (µg/kg) 1400 0.24 0.29 0.41 0.53 0.29

BOLD Underline = 95% Upper Confidence Limit (UCL) exceeds current ADEC Cleanup Level Italics = 95% UCL exceeds Proposed ADEC Cleanup Level *Mass balanced using 95% UCL values if available 1ADEC, 2018a, Table B-1 Method 2 Soil Cleanup Levels 2ADEC, 2018b, Table B-1 Method 2 Soil Cleanup Levels

If the proposed ADEC CLs are promulgated, the results in Table 8 above for the reconstituted soil indicate that PFHxS and PFOA, with results of 0.5213 and 0.4113 µg/kg respectively, would exceed the proposed CLs of 0.29 µg/kg for each compound.

Should proposed ADEC CLs be promulgated, all or a portion of the <#200 mesh fraction requires removal and treatment in order to reduce the PFHxS and PFOA concentrations in the reconstituted soil. Table 9 below presents the reconstituted soil results following complete removal of the < #200 mesh fraction, which comprised approximately 30% of the bulk soil. Removal of this fraction changes the distribution and resulting PFAS contribution from the remaining fractions, as well as the reconstituted residual soil PFAS totals as shown below.

Table 9. Post-Processing Reconstituted Soil Results with <200# Mesh Removed

Soil Fraction Dry

Weight (grams)

Soil Gradation

95% UCL PFAS Homologue Concentration in Fraction (µg/kg)

PFBS PFHpA PFHxS PFNA PFOS PFOA

>3/4" 435 19 0 0 0.0019 0 0.007 0.001

<3/4" - 1/2" 172 8 0 0 0.0027 0 0 0.0019

<1/2" - #10 mesh 302 13 0 0 0.0056 0 0.0059 0.0049

<#10 - #50 mesh 11 <1 0 0 0.305 0 0.512 0.272

<#50 - #100 mesh 391 17 0 0 0.253 0 0 0

<#100 - #200 mesh 944 42 0 0 0.268 0 0 0.433

RECONSTITUTED SOIL PFAS

CONCENTRATIONS* (µg/kg)

0 0 0.1589 0 0.0034 0.1836

Current CL1 (µg/kg) 3.0 1.7

Proposed CL2 (µg/kg) 1400 0.24 0.29 0.41 0.53 0.29

BOLD Underline = 95% Upper Confidence Limit (UCL) exceeds current ADEC Cleanup Level Italics = 95% UCL exceeds Proposed ADEC Cleanup Level *Mass balanced using 95% UCL values if available 1ADEC, 2018a, Table B-1 Method 2 Soil Cleanup Levels 2ADEC, 2018b, Table B-1 Method 2 Soil Cleanup Levels

Eielson AFB, Alaska 11

By removing the <#200 mesh fraction the reconstituted soil meets all proposed ADEC CLs. The physical separation process could be expected to result in an approximate 70 percent reduction in the soil volume requiring secondary treatment based on a single wash process. This represents the worst case for volume reduction.

In order to examine the best-case scenario for volume reduction from a single process wash, we explored the partial removal of the <#200 mesh soil fraction from the reconstituted soil. As demonstrated in Table 10 below, approximately one quarter of the <#200 mesh fraction could be included in the reconstituted soil mass and still meet the proposed ADEC CLs. A best-case scenario for volume reduction would be approximately 75 percent if this approach were approved by ADEC.

Table 10. Post-Processing Reconstituted Soil Results with 20% of <200# Mesh Retained

Soil Fraction Dry

Weight (grams)

Soil Gradation

95% UCL PFAS Homologue Concentration in Fraction (µg/kg)

PFBS PFHpA PFHxS PFNA PFOS PFOA

>3/4" 435 18 0 0 0.0019 0 0.007 0.001

<3/4" - 1/2" 172 7 0 0 0.0027 0 0 0.0019

<1/2" - #10 mesh 302 12 0 0 0.0056 0 0.0059 0.0049

<#10 - #50 mesh 11 <1 0 0 0.305 0 0.512 0.272

<#50 - #100 mesh 391 16 0 0 0.253 0 0 0

<#100 - #200 mesh 944 38 0 0 0.268 0 0 0.433

<#200 mesh 200 8 0 0 1.380 0 1.380 0.878

RECONSTITUTED SOIL PFAS

CONCENTRATIONS* (µg/kg)

0 0 0.2584 0 0.1156 0.2402

Current CL1 (µg/kg) 3.0 1.7

Proposed CL2 (µg/kg) 1400 0.24 0.29 0.41 0.53 0.29

BOLD Underline = 95% Upper Confidence Limit (UCL) exceeds current ADEC Cleanup Level Italics = 95% UCL exceeds Proposed ADEC Cleanup Level *Mass balanced using 95% UCL values if available 1ADEC, 2018a, Table B-1 Method 2 Soil Cleanup Levels 2ADEC, 2018b, Table B-1 Method 2 Soil Cleanup Levels

Process Water

Almost 73 percent of the PFAS partitioned into the process water. The process water exceeded the current ADEC Groundwater CLs for PFOA as well as the cumulative EPA LHA for both PFOA and PFOS. While the process water did not exceed the proposed ADEC Groundwater CL for PFBS, it did exceed the proposed cumulative 0.070 µg/L CL for the other five PFAS homologues.

Leaching

The <#200 mesh fraction, which contained the most PFAS, also leached the most PFAS. However, that fraction neither leached the greatest percentage of PFAS nor did it leach the lowest percentage. Figure 6 suggests that for this soil, particle size has little effect on residual PFAS leachability.

The choice of leaching solution did not appreciably affect PFAS leaching. With the exception of the <0.5 inch-#10 mesh fraction, the PFAS leached by water versus magnesium sulfate solution were within a few percentage points of each other with water usually leaching slightly more than magnesium sulfate solution.

Leaching potential did not appear to be correlated with particle size when leaching was further evaluated

12 Bench Scale Study – PFAS Soil Volume Reduction on the basis of µg leached per kg of soil. Figure 7 illustrates (i) the amount of PFAS per kg for each soil fraction, (ii) the amount leached by either deionized water or 0.1 M magnesium sulfate and (iii) the percent of the total PFAS in each fraction that was subject to leaching. These results suggest that a secondary wash of the soil fractions could yield a measurably lower PFAS concentration in the soil with the negative consequence of additional process water to treat.

Removal Efficiency

The removal efficiency using soil washing and physical separation can be approximated by taking the PFAS concentrations for the as-received Bulk A, B, and C samples (including the Bulk AS split), averaging those results, and comparing those concentrations to the reconstituted soil PFAS concentrations inclusive of all soil fractions (Table 8). These removal efficiencies are presented in Table 11. It should be noted that soil organic matter PFAS results were not factored into the removal efficiency calculations due to the small mass of material in this study.

Table 11. PFAS Removal Efficiency (µg/kg)

SOURCE PFBS PFHpA PFHxS PFNA PFOS PFOA

Bulk A 0.12 0.00 4.20 0.00 2.40 3.20

Bulk AS 0.13 0.00 4.10 0.00 2.40 3.30

Bulk B 0.14 0.00 3.80 0.00 2.00 3.00

Bulk C 0.13 0.00 4.60 0.00 2.50 3.50

Mean PFAS Concentration in Bulk Soil 0.13 0.00 4.18 0.00 2.33 3.25

Post Processing Reconstituted Soil (Table 8) 0.00 0.00 0.49 0.00 0.42 0.27

PFAS Remaining in Reconstituted Soil (%) 0% 0% 12% 0% 21% 13%

Removal Efficiency (% removed) 100% 0% 88% 0% 79% 87%

Data Quality

A review of the laboratory data by Geosyntec did not find any substantive issues with the quality of the laboratory data (Geosyntec, 2019). The bench scale study recovered 87 percent of the starting mass of PFAS in soil (Table 4), which suggests that the results are reliable (AOAC, 1998) and representative of how the process would work under pilot or full-scale conditions where mass balance may be easier to control.

While the study lacked the statistical power to differentiate each of the fractions in terms of their total PFAS retention, the statistical significance (p < 0.05) between the bulk soil and all soil fractions is strong support for the physical separation concept. Additionally, the <#200 mesh fraction is statistically significant at the 90% confidence level (p < 0.1), indicating that this fraction is the primary reservoir of soil PFAS. The lack of any statistical significance between the bulk soil and the process water (p = 0.6695) confirms that the vast majority of PFAS ends up in the process water rather than remaining in the soil.

Findings and Recommendations

The significant differences that exist in PFAS concentrations between the as received soil and post processing reconstituted soil demonstrates that the principles underlying the use of water-based physical

Eielson AFB, Alaska 13 separation are sound. This bench-scale study of the physical separation process successfully produced a reconstituted soil with PFAS concentrations much lower than the as-received soil.

Findings

Key findings of this study are:

• Based on current ADEC CLs, the reconstituted soil meets those cleanup levels and volume reduction is not required.

• If new ADEC regulations are promulgated, a 70 to 75 percent reduction in soil volumes requiring secondary treatment is attainable.

• The soil fraction leaching test further indicates that additional PFAS can be removed from the soil through either additional contact time during deagglomeration or with a second wash cycle.

Several important aspects regarding the stockpiled soil used in this study and findings include the following:

• The soil evaluated in this study was from a stockpile containing the highest known PFAS concentrations in current stockpiles at Eielson, and as such may represent the worst-case scenario for residual post-processing reconstituted soil PFAS concentrations. Reconstituted soil results attained in this study would be lower with soils containing lower initial PFAS concentrations.

• The soil evaluated in this study may be atypical with respect to soil gradation and contain a higher percentage of fines (<#200 mesh) than the EIE 385 site (Table 1) or other soil stockpiles. This fraction contains the highest post-processing fractional PFAS concentrations. Thus, should other stockpiled soils at Eielson have an overall lower fines content, then one would expect post-processing reconstituted soil results to be lower as well.

• Soil OM can serve as a “sink” for many contaminants, including PFAS. This material, consisting of leaves, small sticks, roots, grass, and their decomposition products, can be recovered as a separate product with the field process. Although the total amount of soil OM is relatively low, maximizing removal of the organic matter with particle separation or other methods may reduce residual PFAS concentrations in the processed soil and decrease the incidence of analytical “spikes” when testing processed soil samples for PFAS.

The first two bullet points above highlight that there is a direct relationship between heterogeneous soil PFAS concentrations and soil gradation with respect to residual concentrations following processing. If other stockpiled soils at Eielson contain lower PFAS contamination and/or a coarser soil gradation, results indicate that complete soil treatment may be attainable at current ADEC CLs in some or most cases, or a higher volume reduction percentage than seen in this study based on the ADEC proposed CLs. Only two PFAS homologues (PFHxS and PFOA) exceeded the ADEC proposed CLs for the most contaminated PFAS soil stockpile at Eielson, and that in a soil sample with a gradation potentially containing more <#200 mesh particles than is typical for the balance of soil stockpiled at Eielson.

With 73 percent of soil PFAS ending up in the process water, the management and treatment of process water will be a key factor in upscaling this bench study to the field. There are multiple options for the treatment of PFAS contaminated water on an industrial scale, including treatment by granular activated carbon (GAC) or ion exchange resins (EPA, 2018). One remaining question is what concentration of PFAS in the recirculating process water would result in cross- or re-contamination of soils within the physical separation process. This will be a key factor in determining how many iterations the process water could be re-used prior to subsequent treatment or new water added to the system.

14 Bench Scale Study – PFAS Soil Volume Reduction

Recommendations

The results of this bench-scale study are a strong indicator of both the potential efficacy of the physical separation process for volume reduction of PFAS-contaminated soils at Eielson and of the benefit of conducting a process optimization study to determine the operational parameters and economic feasibility of the process for deployment at Eielson AFB. This study would need to be conducted with the following considerations:

• Floatation or other means of separating coarse organic materials

• Optimization of contact time for physical separation and PFAS removal

• Optimization of process water volumes and re-use between GAC or IX resin treatments

• Options for field lab screening for PFAS concentrations in soil and water

• Efficacy for soils impacted with both petroleum and PFAS

• Quantification of expected day to day volume reduction and/or complete treatment through gradation sampling of other stockpiles

• Design and cost estimate for a pilot demonstration

Conclusion

Studies were initiated to determine the characteristics of a soil sample representing the highest PFAS contamination for various soils stockpiled at Eielson with respect to soil gradation, TOC, and PFAS concentrations. Particle separation and washing were imposed on the soil to simulate processing with an actual plant. Proof of concept results have shown that the water-based process can be effective for reducing soil PFAS concentrations to meet current ADEC cleanup levels for the entire soil, or as a volume reduction process for between 70 and 75 percent of the soil mass at proposed ADEC cleanup levels. Based on these results, it is recommended that the Air Force proceed with additional studies to optimize the process and demonstrate the technology in a field environment.

Eielson AFB, Alaska 15

References

ADEC, 2018a. 18 AAC 75 Oil and Other Hazardous Substances Pollution Control, as amended. Office of the Commissioner, Alaska Department of Environmental Conservation. September 29, 2018.

ADEC, 2018b. 18 AAC 75 Oil and Other Hazardous Substances Pollution Control, Public Comment Draft.

Office of the Commissioner, Alaska Department of Environmental Conservation. October 1, 2018.

AOAC, 1998. Peer-Verified Methods Program, Manual on Policies and Procedures. AOAC International, Rockville, MD.

Benjamin, T.A., C.J. Jones, and L.A. Brice, 2008. Method and System for Removing Contaminants from Soil. United States Patents #7,399,141, July 2008, and #7,255,514, July 2004.

EPA, 2019. Resident Risk-Based Regional Screening Levels for Soil. PFOS and PFOA with Default Resident Equation for Soil. RSL Calculator, https://epa-prgs.ornl.gov/cgi-bin/chemicals/csl_search.

Accessed April 4, 2019.

EPA, 2018. Reducing PFAS in Drinking Water with Treatment Technologies. U.S. Environmental Protection Agency. https://www.epa.gov/sciencematters/reducing-pfas-drinking-water-treatment-technologies/. August 23, 2018. Accessed January 18, 2019.

EPA, 2016. Lifetime Health…