PFAS analysis in aqueous samples using the Biotage® PrepXpert-8 system

Introduction

The Biotage® PrepXpert-8 automated solid phase extraction system was evaluated for the preparation of aqueous samples for PFAS analysis in accordance with EPA Methods 537.1, 533, and 1633A. The study demonstrates that the system is suitable for routine laboratory implementation across these methods and can deliver the analytical performance required for regulated and high-demand environmental testing workflows. The data demonstrate that the Biotage® PrepXpert-8 meets the method criteria for low system background, method detection limits, analyte reporting limits, precision, and accuracy across all three EPA PFAS methods. In addition, EPA method 1633A performance data are presented for challenging aqueous matrices containing PFAS at concentrations greater than 30,000 ng/L, demonstrating the system's capability to support high-concentration and complex water samples while maintaining acceptable analytical performance.

EPA method 537.1

Minimum Reporting Level (MRL)

Confirmation is performed by first establishing a target MRL concentration based on the intended use of the method. Whether defined by the laboratory or a regulatory agency, target MRLs must ensure that the lowest calibration standard in the initial calibration is at or below the proposed MRL. For EPA Method 537.1, reported single-laboratory lowest concentration MRLs (LCMRL) range from 0.53 to 6.3 ng/L, while analyte detection levels (DL) range from 0.53 to 2.8 ng/L. While laboratories are not required to determine their own LCMRL, they must demonstrate that their established MRL meets the applicable method LCMRL requirements. Table 1 outlines the LCMRL and DL listed in EPA Method 537.1 for the 18 PFAS target method analytes. 

Table 1.  EPA Method 537.1 detection limits (DL) and lowest concentration reporting levels (LCMRL) determined from extraction of fortified reagent water.

The laboratory’s proposed MRL must be confirmed by fortifying, extracting, and analyzing seven replicate laboratory fortified blanks at the proposed concentration. The mean measured concentration and standard deviation for each analyte are then calculated and used to determine the half range for the prediction interval of results (HPIR). The MRL is considered confirmed only when the upper prediction interval limit is less than or equal to 150% recovery and the lower prediction interval limit is greater than or equal to 50% recovery. To ensure the Biotage® PrepXpert-8 meets the EPA Method 537.1 requirements, an MRL of 1 ng/L was evaluated for the PFAS compounds listed in the EPA Method 537.1. Seven replicate laboratory fortified blanks (LFB) spiked at 1 ng/L were extracted on the system and analyzed following the method procedure. The HPIR was calculated from the results acquired from the seven LFB extractions to determine whether the system met the expected method performance at that concentration. Figure 1 shows the calculated upper and lower PIR from seven EPA 537.1 LFBs spiked at the selected MRL of 1 ng/L and extracted on the Biotage® PrepXpert-8. Results demonstrate that the system meets the method acceptance criteria for both the upper and lower prediction interval limits at the proposed MRL. 

Figure 1. Calculated upper and lower PIR from seven EPA 537.1 LFBs spiked at the selected MRL of 1 ng/L and extracted on the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

Detection limit (DL) determination

Detection limit determination for analytes in EPA Method 537.1 is optional. Method detection limit determination is performed by processing at least seven spiked LFB samples through the entire extraction and analytical procedure. For each target PFAS analyte, the sample standard deviation is calculated from the replicate spiked LFB results. The detection limit is then calculated by multiplying that standard deviation by the student's t-value (one-tailed 99th percentile) based on the number of replicate samples analyzed. Figure 2 compares the calculated detection limits from the seven EPA Method 537.1 LFBs spiked at 1 ng/L and extracted on the Biotage® PrepXpert-8. Detection limits determined from this study range from 0.09 ng/L to 0.31 ng/L for the 18 target analytes. This demonstrates that the Biotage® PrepXpert-8 can achieve detection limits lower than those referenced in the EPA Method 537.1 method.

Figure 2. Comparison of Biotage® PrepXpert-8 calculated detection limits (DL) to EPA 537.1 detection limits (DL) from the seven extracted LFBs spiked at 1 ng/L. Compounds marked with an asterisk were analyzed in salt form.

Demonstration of low system background

Confirmation must demonstrate that all method analyte concentrations are less than one-third of the MRL (1/3 MRL) and that background or interferences from the extraction media do not compromise analyte identification or quantitation. Automated extraction systems must extract a laboratory reagent blank (LRB) on each sample port to verify that all tubing, valves, and fluidic pathways are free of potential PFAS contamination. Figure 3 outlines system background levels after extraction of eight laboratory reagent blank (LRB) samples on the eight sample ports of the Biotage® PrepXpert-8. The results were compared to the 1/3 MRL based on the 1 ng/L minimum reporting limit (MRL).Figure 3. Biotage® PrepXpert-8 system background levels determined after extraction of laboratory reagent blanks (LRB) on each of the eight sample ports. 1/3 MRL corresponds to the minimum reporting level (MRL) demonstrated at 1 ng/L. Compounds marked with an asterisk were analyzed in salt form.

Demonstration of precision and accuracy

The initial demonstration of precision is established by preparing, extracting, and analyzing a minimum of four replicate laboratory fortified blank (LFB) samples spiked near the midpoint of the initial calibration range in accordance with the EPA Method 537.1. For this study, five calibration standards were used to cover a sample concentration range of 0.1 to 20 ng/L. This allowed for low-level precision and accuracy evaluation of the Biotage® PrepXpert-8 at 1 ng/L, since this concentration is near the midpoint of the calibration range. Acceptable system performance for replicate laboratory fortified blanks is demonstrated by a relative standard deviation (RSD) of less than 20% and average recovery within +/-30% of the true value. To demonstrate precision and accuracy at low levels, ten replicate laboratory fortified blanks were spiked at 1 ng/L and processed on the Biotage® PrepXpert-8 following the method procedure. Figure 4 shows the average recovery of four replicate LFB samples at concentrations of 1 ng/L recovering within the EPA Method 537.1 target analyte recovery range of 70-130%. Most target analytes showed average recoveries of90-110%, exceeding method requirements. Figure 5 demonstrates precision between the same four replicate LFB samples spiked at 1 ng/L, with all target analytes below 12% relative standard deviation (RSD), which exceeds the EPA Method 537.1 requirement of <20% RSD.

Figure 4. Calculated average percent recovery of four replicate laboratory fortified blank (LFB) samples spiked at a concentration of 1 ng/L extracted on the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

Figure 5. Calculated percent relative standard deviation (%RSD) for four replicate laboratory fortified blank (LFB) samples spiked at a concentration of 1 ng/L extracted with the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

EPA method 533

Minimum Reporting Level (MRL)

Confirmation is performed by first selecting a target MRL concentration based on the intended use of the method and ensuring that the lowest calibration standard, including the low-level continuing calibration check, is at or below that level. For EPA Method 533, the analyte calibration range is approximately 0.50 ng/mL to 25 ng/mL for the extract concentration. A 9-point calibration ranging from 0.20 ng/mL to 100 ng/mL for the extract concentration was used for all 25 target PFAS compounds to evaluate the Biotage® PrepXpert-8 against EPA Method 533.

Ten replicate laboratory fortified blanks (LFB) at a proposed MRL of 2 ng/L were extracted using the Biotage® PrepXpert-8. To confirm that the proposed MRL meets the method criteria, the mean and standard deviation were calculated for each analyte. Using these data, the half range prediction interval of results (HPIR) was calculated to determine whether the system met the expected method performance at that concentration. The MRL is considered confirmed when the upper prediction interval result (PIR) is less than or equal to 150% and the lower prediction interval result (PIR) is greater than or equal to 50%. Figure 6 demonstrates that the calculated upper and lower PIR for the ten replicate LFBs extracted on the Biotage® PrepXpert-8 meet EPA Method 533 criteria at the proposed MRL of 2 ng/L. Figure 7 shows that  the calculated MDLs from ten replicate laboratory fortified blanks spiked at 2 ng/L were well below the calculated MRL of 2 ng/L, meeting the acceptance criteria of EPA Method 533.

Figure 6. Calculated upper and lower PIR from ten EPA 533 LFBs spiked at the selected MRL of 2 ng/L and extracted on the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

Figure 7. Calculated method detection limit (MDL) from ten replicate EPA 533 LFBs spiked at a concentration of 2 ng/L and extracted on the Biotage® PrepXpert-8 compared to the minimum reporting limit (MRL) calculated from the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

Demonstration of low system background

Confirmation is performed by analyzing a laboratory reagent blank immediately after injection of the highest calibration to verify that the extraction system is free from potential contamination. EPA Method 533 does not specifically list required MRL values, noting that the MRL concentration is based on the intended use of the method. The method does define the lowest concentration minimum reporting level (LCMRL) as the lowest concentration at which future recovery is predicted, with 99% confidence, to fall between 50% and 150%. Table 2 compares the LCMRLs defined in EPA Method 533 with the LCMRL demonstrated with the Biotage® PrepXpert-8 using the half range prediction interval of results (HPIR) at 2 ng/L. Most LCMRL values demonstrated on the Biotage® PrepXpert-8 are well below the LCMRL values noted in the method, which confirms that a target MRL of 2 ng/L will provide an appropriate benchmark to demonstrate low system background.

Table 2. EPA Method 533 lowest concentration reporting levels (LCMRL) determined from extraction of fortified reagent water.

To demonstrate the low system background requirement for EPA Method 533, automated extraction systems must extract laboratory reagent blanks on each sample port. Acceptable system background is demonstrated when all method analytes in the blanks are measured at concentrations below one-third of the minimum reporting level (1/3 MRL). Figure 8 shows Biotage® PrepXpert-8 system background for all target analytes, with results showing no recovery or recovery well below the 1/3 MRL based on the 2 ng/L minimum reporting limit (MRL). This evaluation confirms that background contribution from the analytical system or extraction platform is sufficiently controlled for low-level quantitative analysis. 

Figure 8. Biotage® PrepXpert-8 system background levels determined after extraction of laboratory reagent blanks (LRB) on each of the eight sample ports. 1/3 MRL corresponds to the minimum reporting level (MRL) demonstrated at 2 ng/L. Compounds marked with an asterisk were analyzed in salt form.

Demonstration of precision and accuracy

Precision and accuracy are established by preparing, extracting, and analyzing a minimum of seven replicate laboratory fortified blanks spiked near the midpoint of the initial calibration range. Precision is demonstrated when the percent relative standard deviation of the measured concentrations for all method analytes is less than 20%. Accuracy is evaluated by using the same replicate data and calculating the average percent recovery for each analyte. Acceptable accuracy is achieved when the average recovery for each analyte falls within 70%-130%. Figure 9 shows the average recoveries of seven replicate LFB samples spiked near the midpoint of the calibration range, with recovering within the EPA Method 533 target analyte recovery range of 70-130%. Most target analytes showed average recoveries  of 80-100%, exceeding method requirements. Figure 10 demonstrates precision between the same seven replicate LFB samples spiked at a concentration near the midpoint of the calibration range with all target analytes below 10% relative standard deviation (RSD), which exceeds the EPA Method 533 requirement of <20% RSD. Together, these results confirm that the method can produce consistent and accurate measurements under the defined test conditions.

Figure 9. Calculated average percent recovery of seven replicate laboratory fortified blank (LFB) samples spiked at a concentration near the midpoint of the calibration range extracted on the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

Figure 10. Calculated percent relative standard deviation (%RSD) for seven replicate laboratory fortified blank (LFB) samples spiked at a concentration near the midpoint of the calibration range extracted with the Biotage® PrepXpert-8. Compounds marked with an asterisk were analyzed in salt form.

EPA method 1633A

The Biotage® PrepXpert-8 automated SPE system was evaluated at a commercial environmental testing laboratory to assess its performance for the preparation of aqueous samples in support of EPA Method 1633A. The study was designed to determine the system's ability to perform in a laboratory already routinely performing PFAS testing for surface water, groundwater, and wastewater samples. The goal was to evaluate system robustness in a real-world setting by examining key performance characteristics such as extraction consistency, reproducibility, and overall workflow suitability in a production laboratory environment.

Minimum reporting limit

The EPA Method 1633A uses the terms Minimum Level (ML), reporting limit (RL), quantitation limit (QL), and limit of quantitation (LOQ) interchangeably. EPA Method 1633A references ranges for aqueous LOQs from 1 ng/L to 100 ng/L, with ranges varying by individual target PFAS compounds. Table 3 compares the laboratory’s specific LOQs with the minimum and maximum LOQs referenced for the 40 target PFAS analytes listed in the method.

Table 3. Comparison of the laboratory’s specified limit of quantitation (LOQ) in ng/L and ranges allowed per EPA Method 1633A for the 40 PFAS target analytes.

Method blank acceptance criteria

If any target PFAS is present in the method blank at a concentration greater than the highest of the analyte LOQ, sample analysis shall be suspended until the contamination source has been identified and corrected. Seven 500 mL method blanks were extracted on the Biotage® PrepXpert-8 following the EPA Method 1633A extraction procedure to demonstrate that the system meets method background level requirements. Table 4 outlines results from the background evaluation for the 40 target PFAS analytes, with most compounds below the laboratory’s detection limit (DL) and reported as non-detect (N.D.). PFOA was detected in method blank 5 at 0.67 ng/L, however, this was still below the laboratory’s LOQ (2 ng/L) for the target compound. In addition, the fact that this compound was not detected in the other six method blanks suggests that it could have come from another source within the laboratory. 

Table 4. Determined Biotage® PrepXpert-8 system background levels after extraction of seven replicate method blanks (MB) are all below the specific limit of quantitation (LOQ) set by the laboratory.

Initial demonstration of capability

For initial precision and recovery, the laboratory must demonstrate acceptable extraction performance. The laboratory must extract and analyze a minimum of four 500 mL reagent water blanks spiked with native standard solution containing the 40 target compounds at mid- to high-level concentration along with isotopically labeled extracted internal standards (EIS). For each target analyte, the laboratory must compare percent recovery and relative standard deviation to the applicable acceptance criteria in the method tables. System performance is considered acceptable only if all target analytes and isotopically labeled EIS analogs meet the required precision and recovery limits.

To demonstrate that the Biotage® PrepXpert-8 meets EPA Method 1633A precision and recovery requirements, eight replicate 500 mL reagent water blanks were spiked with the 40 native PFAS targets at 22.2–625 ng/L representing mid-range concentrations with respect to the method calibration range. Figure 11 shows that results from the eight replicate extractions meet method recovery criteria, with all 40 target analytes recovered within 70 – 130% of their spiked concentration.  The calculated precision (RSD) from the eight replicate extractions also meet EPA Method 1633A criteria with 39 target compounds falling below 20% RSD (Figure 12). NEtFOSAA had an RSD of 22.6% and was the only target that had an RSD greater than 20%, however, this was still below the EPA method requirement of 28%. The Biotage® PrepXpert-8 demonstrated excellent overall extraction performance with average EIS recoveries falling within 70–130% for the eight mid-level spikes (Figure 13). 

Figure 11. Calculated average percent recovery of eight replicate laboratory fortified blank (LFB) samples spiked at concentrations of 22.2-625 ng/L extracted on the Biotage® PrepXpert-8.

Figure 12. Calculated percent relative standard deviation (%RSD) for seven replicate laboratory fortified blank (LFB) samples spiked at concentrations of 22.2-625 ng/L extracted with the Biotage® PrepXpert-8.

Figure 13. Calculated average percent recovery of EIS in the eight replicate laboratory fortified blank (LFB) samples spiked at concentrations of 22.2-625 ng/L extracted on the Biotage® PrepXpert-8.

Groundwater extraction performance

Four groundwater samples (GW1, GW2, GW3, GW4) containing a wide range of native PFAS levels were extracted on the Biotage® PrepXpert-8. The system was able to process all four 500 mL samples without any issues. These samples contained high levels of PFAS with detected concentrations as high as 8,610 ng/L. Figure 14 shows the results from these extractions. The y-axis max for Figure 14 was set at 500 ng/L to show detail at lower concentrations with detections above 500 ng/L noted for PFHxA, PFOA, and PFHxS.  The extraction performance was determined using the EIS recovery criteria found in Table 6 of EPA Method 1633A. Results for EIS recoveries were within EPA acceptance criteria, with the exception of 13C2-4:2FTS for GW1 (209%) and GW2 (217%). However, when compared to 13C2-4:2FTS acceptance criteria for leachate, these samples would have passed the criteria of <220% recovery (Figure 15).

Figure 14. Concentrations of PFAS natives recovered from ground water samples extracted with the Biotage® PrepXpert-8. EIS recoveries were within the passing limits of EPA Method 1633 except for two samples, GW1 (209%) and GW2 (217%), where ¹³C₂-4:2FTS was out of range. If it were a leachate, this EIS would have passed since the upper limit is 220%.

Figure 15. Calculated average percent recovery of EIS in the ground water samples extracted on the Biotage® PrepXpert-8. EIS recoveries are within the passing limits of EPA Method 1633 except for two samples, GW1 (209%) and GW2 (217%), where ¹³C₂-4:2FTS was out of range. If it were a leachate, this EIS would have passed since the upper limit is 220%.

Manual and automated extraction comparison

Extraction performance of the Biotage® PrepXpert-8 was compared with a manual vacuum manifold extraction from a ground water sample taken from the same location. Although these samples were not collected at the same time, the laboratory was confident that they would provide a good comparison with historical data from this sample location. Results from this study demonstrated consistent detection of 12 PFAS analytes at high concentrations ranging from 237 ng/L-30,700 ng/L for both manual and automated extraction techniques (Figure 16). Although most of the detected concentrations were well above the highest calibration point for EPA Method 1633A, results were still consistent between techniques.  In addition, a 10X dilution was required to process the sample using the manual vacuum manifold, whereas the Biotage® PrepXpert-8 was able to process the undiluted sample. This allowed for detection of 4:2FTS at 18.8 ng/L, which would have been below the laboratories LOQ with a 10X dilution factor applied.

Figure 16. Comparison of manual extraction with a vacuum manifold and automated extraction with the Biotage® PrepXpert-8 for a ground water sample taken from the same location at different times. Results are consistent with both techniques since each detected high concentrations of PFAS analytes ranging from 237 ng/L -– 30,,700 ng/L.

Biotage® PrepXpert-8 extraction methods

Table 5. EPA 537.1 method.

Table 6. EPA 533 method.

Table 7. EPA 1633A – aqueous method.

Literature number: AN1036

Ordering information