Dataset of surveyed PFAS in water, sediment, and soil of Fountain Creek Watershed, Colorado, USA

Per- and polyfluoroalkyl substances (PFAS) are widespread and highly persistent organic chemicals with adverse health effects. The US Environmental Protection Agency has issued health advisory limits of 70 ng/L for aqueous concentrations of PFOA + PFOS. In the Colorado Springs, Colorado (USA), metro area, the Widefield Aquifer (groundwater) and Fountain Creek Watershed (surface water) have been contaminated by PFAS from aqueous film-forming foams. Here we present the concentrations of selected linear and branched isomers of legacy PFAS found in surface water (n = 95), soil (n = 83), and sediment (n = 34) samples collected from several creeks of the Fountain Creek Watershed. Collected samples were prepared for high-performance liquid chromatography tandem mass spectrometry (LC/MS/MS) analysis via liquid/liquid extraction and/or solid phase extraction (SPE). This dataset includes the geographic locations of sampled creeks, LC/MS/MS instrumental conditions, method verification data including percent recovery to assess method accuracy and background contamination of PFAS in laboratory reagents and supplies, and determined concentrations of PFAS in water, soil, and sediment samples. These locations were surveyed monthly for a full year and provide a rich dataset to assess influence of sampling location, temporal variability in concentration, and overall contaminant persistence.


a b s t r a c t
Per-and polyfluoroalkyl substances (PFAS) are widespread and highly persistent organic chemicals with adverse health effects. The US Environmental Protection Agency has issued health advisory limits of 70 ng/L for aqueous concentrations of PFOA + PFOS. In the Colorado Springs, Colorado (USA), metro area, the Widefield Aquifer (groundwater) and Fountain Creek Watershed (surface water) have been contaminated by PFAS from aqueous film-forming foams. Here we present the concentrations of selected linear and branched isomers of legacy PFAS found in surface water (n = 95), soil (n = 83), and sediment (n = 34) samples collected from several creeks of the Fountain Creek Watershed. Collected samples were prepared for high-performance liquid chromatography tandem mass spectrometry (LC/MS/MS) analysis via liquid/liquid extraction and/or solid phase extraction (SPE). This dataset includes the geographic locations of sampled creeks, LC/MS/MS instrumental conditions, method verification data including percent recovery to assess method ac-curacy and background contamination of PFAS in laboratory reagents and supplies, and determined concentrations of PFAS in water, soil, and sediment samples. These locations were surveyed monthly for a full year and provide a rich dataset to assess influence of sampling location, temporal variability in concentration, and overall contaminant persistence. ©

Value of the Data
• These data of selected legacy PFAS in waters, soils, and sediments are crucial to understanding longer-term impact of PFAS release into an urban environment. • These data of selected legacy PFAS in water, soils, and sediments are of interest to scientific investigators, public health officials, and water utilities seeking to understand the longer-term impact of PFAS release into water resources. • These data may support further inquiries into additional surface water or groundwater resources or public health investigations to assess impact of PFAS contamination within a community.

Objective
Per-and polyfluoroalkyl substances (PFAS), which have been used widely in consumer and industrial products, have long-term environmental persistence. In our reported work, we measured PFAS in water, sediment, and soil surveyed at eight locations monthly for a full year within the Fountain Creek Watershed, Colorado Springs, CO, to assess how selected legacy PFAS varied over time and distance [1] . The PFAS concentration data to support that study are included in a Mendeley repository and described here [2] . Beginning in 2016, local news outlets in the Colorado Springs, Colorado, USA metro region reported the contamination of water resources by aqueous film-forming foams (AFFF) [3][4][5][6] , which contain PFAS. Several teams of investigators evaluated the impact of AFFF-contaminated water on drinking water utilities and community health impacts [7][8][9] . Our contribution here was to report concentrations of PFAS from the Fountain Creek Watershed, which was affected by AFFF-contamination on the Widefield Aquifer. The Widefield Aquifer is a shallow groundwater resource that contributes water to the surface water of the Fountain Creek Watershed. We sampled eight locations within the Watershed for each month over one year to assess the concentrations of selected PFAS in water, sediment, and soil. These data are useful for other investigators seeking to assess long-term impact of PFAS contamination over time and distance.

Data description
Selected legacy per-and polyfluoroalkyl substances (PFAS), including linear and branched isomers [10] , were evaluated in water, sediment, and soil samples collected each month for a full year from eight locations within the Fountain Creek Watershed, Colorado Springs, Colorado (USA). Table 1 includes a description of each sample location, its abbreviated name included in following table data sets, name of the sampled creek, and geographic coordinates. Samples of water (n = 95), soil (n = 83), and sediment (n = 34) were collected at each location each month ( Table 2 ) and were prepared for liquid chromatography tandem mass spectrometry (LC/MS/MS) analyses following sample preparation, which included a solid phase extraction (SPE) clean up. PFAS analytes were separated on a Waters XBridge C18 column and analysed in multiple reaction monitoring (MRM) mode ( Table 3 ). Our method used no internal standard or surrogate standard to account for analyte loss in sample preparation. As such, Table 4 includes determined concentrations via external calibration curves and the pooled recovery data to assess method accuracy for analyte spikes made at 0.400 ng/mL (final concentration after SPE clean up) to all sample types, including water, sediment, and soil. In Table 5, the determined concentrations using external calibration curves for linear and branched PFAS compounds in soils (n = 83) and sediments (n = 34) are reported for each location and each month collected and analysed. In addition to including the determined concentration in the SPE sample extract as ng/mL, the concentration of PFAS in the original soil or sediment is reported as ng PFAS/g soil or sediment (dry weight). Table 6 includes the determined concentrations, by external calibration curves, in A indicates water, soil, and sediment were all collected and analyzed for the given sampling day. B indicates that water and soil were analyzed for the given sampling day. D indicates that only water was analyzed for the given sampling day. E indicates that only water and sediment were analyzed for a given day. NC indicates not collected. Note: in many cases, sediment samples were collected but not analyzed owing to limited resources.
replicate analyses of water (n = 95) samples collected from the study locations for linear and branched isomers of selected PFAS. The sample locations, collection date, and creek name are included in this table as well. The determined concentrations, as ng/mL, in the SPE sample extracts are reported along with the original concentration (as ng/L) in the original water sample. Lastly, given the ubiquitous nature of PFAS, the concentrations of PFAS in the LC/MS/MS system (reported as 'system blanks') and laboratory reagents used in SPE sample extraction (reported as 'method blanks') are included in Table 7. Lastly, Table 8 is a master spreadsheet that includes all analytical standards, laboratory and method blanks, field samples, and spiked samples with a list of where previous concentration data (Tables 4 through 7) may be found. A file folder with

Sample locations and sample collection protocols
Eight sample locations within the Fountain Creek Watershed were chosen and their names, abbreviations, associated creek, and coordinates are included in Table 1 . Table 2 describes the types of samples that were collected at each location at every monthly survey of the field sites. Water and sediment samples were collected in 50 mL Falcon R polypropylene centrifuge tubes that had been triple field rinsed. Water samples were stored in a freezer until analysis, at which point they were defrosted prior to SPE. Sediment samples were collected from the soils and gravel samples of the creek bed. These sediment samples were collected from the top 5 to 7 cm of the creek bed in 50 mL Falcon R polypropylene centrifuge tubes that had been triple field rinsed with the creek water. Soil samples were collected in 15 mL Falcon R polypropylene centrifuge tubes. Soils were collected approximately 6 ft away from the edge of the creek and approximately one foot below the surface to avoid weathering effects or other PFAS contamination by surface features. Sediment and soil samples were prepared for analysis by sifting out large rocks and/or organic matter through a metal screen. Approximately 25 g of sifted soil or sediment were dried in aluminum boats at 80 °C for 2 d. The dried samples were transferred to new 15 mL Falcon R tubes and stored at ambient temperature (25 °C) until prepared for instrumental analysis.

Preparation of analytical standards
An intermediate stock solution of 0.020 ng/μL was prepared by adding 10 μL of the Native PFAS Stock (see 2.1 ) to 1 mL of 18 M water. This intermediate stock was stored in a 1.5 mL polypropylene microcentrifuge tube at 4 °C and re-made every two weeks.
Calibration standards of 0.04 ng/mL, 0.10 ng/mL, 0.20 ng/mL, 0.39 ng/mL, 0.77 ng/mL, and 1.82 ng/mL were prepared from the intermediate stock solution (0.020 ng/μL) by adding 2, 5, 10, 20, 40, and 100 μL, respectively, to six 1-mL aliquots of 18 M DI water. Standards were vortex mixed. Only Teflon-free Hamilton 700 series syringes (Hamilton Company, Reno, NV) were utilized in the preparation of the intermediate stock and calibration standards. Three hundred μL aliquots of these analytical standards were dispensed into Wheaton 12 × 32 mm polypropylene vials (Fisher Scientific) and capped with a polyethylene olefin snap cap (Agilent Technologies, Santa Clara, CA). Calibration standards were stored for a maximum of two weeks at 4 °C.

Preparation of HPLC mobile phases and gradient profile
Mobile phase A was 2.5 mM ammonium formate in organic-free 18 M DDI water modified with 200 μL ammonium hydroxide to pH 8.5 and mobile phase B was 2.5 mM ammonium formate in LC/MS Optima-grade methanol modified with 200 μL ammonium hydroxide to pH 8.5. Mobile phases were prepared as needed and the pH was stable in this buffered system with no impact on chromatographic peak shape. Forty μL injections were made onto a Waters XBridge C18 column (50 × 2.1 mm i.d., 3 μm particle). The gradient conditions were as follows: 35% B at 0 min with a ramp to 95% B at 6.0 min (hold 1.0 min) with a return to 35% B at 7.1 min (hold for 4.9 min) for a total run time of 12 min. The flow rate was held at 0.400 mL/min. Retention times of analyzed PFAS are included in Table 3 .

Shimadzu LCMS-8030 operating conditions
The triple quadrupole mass spectrometer was operated using an electrospray ionization source in negative-ion mode at 3.5 kV. Settings were as follows: nebulizing gas at 1.5 mL/min, drying gas at 15 mL/min, desolvation line temperature at 250 °C, and heat block temperature at 400 °C. Argon was utilized as the collision gas and was maintained at 230 kPa. The mass spectrometer was operated in multiple reaction monitoring (MRM) mode with unit resolution for Q1 and Q3. The instrument was optimized to find the best precursor and product ion(s) for each PFAS along with optimal voltages for Q1, Q3, and collision energy ( Table 3 ). Dwell times for each ion transition were automatically calculated and ranged from 38 to 122 ms. Acquisition windows for closely eluting compounds were established to enhance the sensitivity of the method.

Water
Collected water samples (Table 6) were prepared for LC/MS/MS analysis using Waters Oasis HLB columns (3 cc, 60 mg sorbent per cartridge, 30 μm particle size; Milford, MA), which were conditioned using one column volume of methanol followed by one column volume of 18 M water. Twenty-five mL of the collected water sample were loaded onto the conditioned column using a volumetric pipette into a 70 mL polypropylene SPE reservoir (Restek, Bellefonte, PA) attached via a Restek SPE cartridge/reservoir adapter. The water sample was eluted from the conditioned column at a rate of 2-3 drops per second. After the entire 25 mL sample had been eluted through the conditioned SPE column, the column was dried under vacuum (10 in Hg) for approximately 10 minutes. PFAS were eluted from the SPE sorbent bed and collected into a 15 mL polypropylene centrifuge tube by adding 1 column volume of methanol ( ∼ 3 mL) followed by 1 mL ethyl acetate. These solvent eluents containing the analytes were collected into a 15 mL Falcon tube to which three hundred μL of 18 M water were added as a "keeper" solvent, which is a low-volatility solvent added during drying procedures to minimize analyte loss. The methanol/ethyl acetate solvent system was removed by drying the sample to a final volume of ∼400 μL using a gentle stream of nitrogen and mild heat (60 °C). The final sample volume was adjusted to exactly 1 mL with the addition of 18 M water by drawing up into a Hamilton syringe and mixing well. A 300 μL aliquot of the sample was transferred to a Wheaton 12 × 32 mm polypropylene autosampler vial and capped with a polyethylene olefin cap to await analysis by LC/MS/MS. According to our QC plan (see 2.6 ), field water samples were analyzed in replicate analyses to assess precision (as relative percent difference).

Sediments and soils
One gram of the sieved and dried sediment or soil sample (Table 5), with exact mass recorded on a Mettler XS64 balance, was transferred to a 15 mL Falcon R polypropylene centrifuge tube. To each gram of soil/sediment, 3 mL of 75/25 (v/v) methanol/18 M water were added. The tube contents were mixed by shaking prior to sonication in a water bath sonicator for 15 min at 55 °C. Following sonication, samples were centrifuged for 5 min at 40 0 0 rpm (or 3005 × g ) using a Thermo Scientific Sorvall ST16 centrifuge. The supernatant was decanted into a clean 15 mL polypropylene centrifuge tube and saved. Soil/sediment samples were extracted a second time with 3 mL 50/50 methanol/18 M water (v/v) and 15 μL ammonium hydroxide so that the sample pH was ∼9 as measured by pH test strips. Soil/sediment samples were shaken by hand, sonicated for 15 min in a water bath sonicator at 55 °C, and centrifuged for 5 min at 40 0 0 rpm (30 05 × g ) as previously described. The supernatant from the second centrifugation was decanted and combined with the first supernatant. The samples were evaporated under under a gentle stream of nitrogen and mild heat (60 °C) to ∼2 mL. The remaining ∼2 mL supernatant fraction was transferred to a clean microcentrifuge tube and centrifuged at 170 0 0 × g for 20 min using a Thermo Scientific Legend Micro 17 centrifuge. The supernatant was decanted and added to 25 mL 18 M water and cleaned up via SPE as described in 2.5.1 . The methanol/ethyl acetate solution was evaporated to ∼400 μL using a gentle stream of nitrogen and mild heat (60 °C). The final sample volume was adjusted to exactly 1 mL with the addition of 18 M water using a Hamilton syringe and mixed well. The 1 mL of reconstituted sample was transferred to a 1.5 mL microcentrifuge tube and centrifuged at 170 0 0 × g for 20 min. The supernatant was filtered through a Minisart Regenerated Cellulose (RC) hydrophilic syringe filter (15 mm; Fisher Scientific) and 300 μL transferred to a Wheaton 12 × 32 mm polypropylene autosampler vial and capped with a polyethylene olefin cap. According to our QC plan (see 2.6 ), field samples of soils/sediments were analyzed in replicates.

Method verification protocols and quantitation
All samples, including water, sediments, and soils, were analyzed in analytical batches that commenced with analysis of a system blank (see 2.6.1 ), followed by six analytical standards (see 2.3 ), method blanks, and individual samples. Additionally, quality control (QC) samples included the requisite method blank samples, a field duplicate, and laboratory fortified sample matrix replicate as part of the batch. Calibration check standards were analyzed after every six field samples. Each sample batch concluded with the analysis of at least three analytical standards. All analytical standards were included in the generated calibration curves. If more than three of these values did not meet bias criteria of 100% ± 30%, the entire batch was re-analyzed.

Determination of background PFAS
Laboratory method blanks were included in every batch to assess potential carryover and/or PFAS contamination. The laboratory method blanks were included at the start of the sample extractions. For water samples, laboratory method blanks included 25 mL 18 M water extracted via SPE and prepared for analysis as outlined in 2.5.1. For soils and sediments, laboratory method blanks included the 75/25 (v/v) methanol/water and 50/50 (v/v) methanol/water solutions used for the soil/sediment extractions of 2.5.2. All laboratory blanks were treated as actual samples and carried through the extraction protocols in their entirety prior to LC/MS/MS analysis (Table 7) and are called 'Method Blanks'. Concentrations of linear and branched isomers are reported in all blanks analyzed.
In addition to these 'method blanks' that were processed throughout the entire sample preparation procedure, blank reagent water from the 18 M water system was added to autosampler vials without being cleaned up via SPE. These latter 'system blanks' assessed the background level of PFAS within the LC/MS/MS system. These data are also included in Table 7.

Method Verification of Water
To assess accuracy of the SPE protocol, 25 mL laboratory reagent water samples were spiked with 20 μL of the intermediate stock solution (0.020 ng/μL PFAS) and prepared for analysis as described in 2.5.1 . The percent recovery of the spiked PFAS from laboratory reagent water was determined relative to the levels of background PFAS in laboratory method blanks (Table 7).
% Recov ery of P F AS f rom W ater = 100 × C spiked sample − C method blank 0 . 400 ng/mL

Method Verification of Soils & Sediments
To determine the accuracy of the extraction protocols, two 1-g soils or sediments per batch were spiked with 20 μL of the intermediate stock solution (0.020 ng/μL PFAS), where a batch included six samples, two spiked samples, and a laboratory blank. These spikes with a theoretical PFAS concentration of 0.400 ng/mL, along with the six samples and laboratory solvent blank, were then extracted using the entire protocol of 2.5.2 . Given that the soil and sediment samples themselves contained measurable levels of PFAS, the spiked samples must be a duplicate of a sample without the spike. The percent recovery of the spiked PFAS were then determined using the equation below: % Recov ery of P F AS f rom Soils = 100 × C spiked sample − C unspiked sample 0 . 400 ng/mL To assess the accuracy and precision of the method, the first soil or sediment sample of each batch was analyzed in triplicate (sample 1A, sample 1B, sample 1S) with precision calculated by determining relative percent difference (RPD, %) for samples 1A and 1B, while % recovery (accuracy) was determined for the spiked sample 1S. Additionally, the recovery data for all nine (excluding branched isomers) PFAS were pooled to determine method accuracy for all spiked samples for all matrices throughout the study (Table 4).

Quantitation
The quantitation of the PFAS included in this study were completed for each batch, where every batch included at least n = 9 analytical standards with six standards (0.04 ng/mL to 1.82 ng/mL) were included at the start of every batch and three analytical standards (high, medium, and low concentrations) were included at the conclusion of the batch. Calibration check standards were analyzed every six field samples and included in the generated curve for each analyte. All standards were included in external calibration curves, which were fit with a linear regression and 1/ x weighting. All curves included in this study had R 2 values of 0.97 or better. Concentrations of branched isomers were determined using the linear isomer contained with the PFAS stock solution (see 2.1 ) [10] .