The Extraction of Perfluorinated Alkyl Acids

(PFAAs) in Drinking Water using Automated

Cartridge SPE

Key Words

PFOS, PFOA, perfluorinated alkyl acids, PFAA, drinking water, EPA 537,

ISO 25101:2009

Russ Wolff and Craig Caselton, Northern Lake Service, Inc., Crandon, WI, USA; Alicia Cannon and Michael Ebitson, Horizon Technology, Inc., Salem, NH, USA

Introduction

Perfluorinated chemicals (PFCs), also known as PFAAs, can be found in commercial and industrial uses such as firefighting foams, package

material coatings, nonstick cookware, waterproofing and stain proof fabrics. PFCs are all manmade and have unique properties such as

repelling water, oil and heat. They are very difficult to break down and persist in the environment. Small amounts of PFCs can dissolve in

water and is a cause for concern for human exposure, especially in drinking water sources such as large public water systems and private

wells.

PFC contamination poses risks to the developmental, immune, metabolic, and endocrine health of consumers.1 The largest

potential source of human exposure is through drinking water. To better understand PFC occurrence levels throughout the US

the US EPA included six of these chemicals into the third Unregulated Contaminant Monitoring Rule (UCMR 3) in 2012. UCMR 3

required monitoring for 30 contaminants (28 chemicals and two viruses) between 2013 and 2015 using analytical methods

developed by EPA, consensus organizations or both. This monitoring provides a basis for future regulatory actions to protect

public health. The data summary from the extensive study found 0.9% of the public water supplies studied showed

concentrations of PFOS greater than the reference level (0.07 µg/L).2 Of the public water supplies studied 0.3% showed levels

of PFOA above the reference level (0.07 µg/L). As a result of this study, a Health Advisory has been issued by US EPA. Water

utilities should notify customers if greater than 70 ppt (0.07 µg/L) PFOS or PFOA or a total for the two combined are detected in

the water supply.3

The US is not the only country concerned about exposure to PFCs.4 In Europe, the twelfth meeting of the United Nations

Persistent Organic Pollutants Review Committee was held in Rome in September 2016 to move the consideration of PFCs

forward for further regulatory consideration.5 Rules developed under this framework will have global impact.

PFCs have become ubiquitous in the environment and in the laboratory. As very nonreactive compounds they are included in

equipment parts, tubing and other commonly found laboratory supplies. Therefore, it is a challenge to prepare samples and

measure low concentrations without contamination and interferences. This work will demonstrate the capability of the

SmartPrep® Cartridge Extraction II System for the extraction of six PFCs in compliance with US EPA 537 or ISO 25101:2009.6,7

Low concentrations will be measured with strict contamination control and reproducible performance demonstrated through

the entire analytical process.

Page 2

The perfluorinated compounds include:

perfluorooctanesulfonic acid (PFOS)

perfluorooctanoic acid (PFOA)

perfluorononanoic acid (PFNA)

perfluorohexanesulfonic acid (PFHxS)

perfluoroheptanoic acid (PFHpA)

perfluorobutanesulfonic acid (PFBS)

These six compounds were selected in the third UCMR although other similar

compounds will likely be captured with the same methodology.

Experimental

Solid phase extraction (SPE) is used to remove and concentrate the PFCs from

water samples. In US EPA method 537, a 0.5 g, 6-mL SPE cartridge containing styrene

divinylbenzene (SDVB) sorbent phase is specified. The analysis is done using HPLC/

MS/MS for a sensitive detection step.

Automated conditioning, application of sample and elution was done using the SmartPrep® Cartridge Extractor II (Horizon

Technology, Inc.) configured to handle 6-mL cartridges with 20-mL vessel collection tray. The original sample bottle was

rinsed automatically with the following procedure:

The configuration for the rinse kit was enabled in the software.

Sample lines 1-6 (how sample is loaded) were configured normally to load the samples onto the cartridge.

The sample lines that rinse the container (samples lines 7-12) were connected directly to the shower head, omitting the extra tubing from the rinse kit.

The extraction procedure implemented with the SmartPrep is summarized in Figure 1. The details of the method necessary to enter into the software are included in the Appendix.

A Strata® SDBL 100 µm Styrene

divinylbenzene, 6-mL cartridge was used

(Phenomenex).

The extracts were evaporated using an N

-Evap 112 (Organomation) to dryness

and then reconstituted in 1 mL of 96:4

methanol/water.

The HPLC system used was a Prominence

System (Shimadzu) with Atlantis® dc18, 5

µm, 2.1 x 150 mm HPLC column

(Waters). The mass spectrometer used

was an API4000 LC/MS/MS (SCIEX) for

the analysis step.

SmartPrep Automated Extraction

System II

Figure 1. Method Summary for the SmartPrep Extractor II.

Reagents

Reagent Water

Methanol

Method Summary

1. Prior to extraction the 250 mL samples must be preserved with a dechlorinating reagent.

2. Do not transfer the sample to another bottle as some of the PFAAs may adsorb to surfaces.

3. Add all appropriate Standards and Surrogates, then cap and invert sample bottle to mix.

4. Load each sample into a position on the SmartPrep extractor with sample sip tube and rinse cap.

5. Load the sample method and sample name in the sequence for each sample to be extracted.

6. Once the sequence is initiated, all the appropriate conditioning, air-dry, rinsing and elution steps started will be fully

automated.

7. Once the extracts are collected, they must be concentrated to dryness under a gentle stream of nitrogen to remove the

water/methanol mixture.

8. Once the extracts are evaporated to dryness,

9. Then add Internal Standard and the appropriate amount of methanol: water solution (96:4%) to bring up to a final volume of

1.0 mL.

10. Follow appropriate techniques (EPA method 537, in this case) for the storage of extracts.

11. Analyze by LC/MS/MS.

Results and Discussions

Prior to running field samples, an Initial

Demonstration of Capability (IDC) is required. Within

the results, five of the minimum acceptable QC criteria

are demonstrated with the SmartPrep Cartridge

Extractor System.

Table 1 shows a Laboratory Fortified Blank (LFB)

extracted with the SmartPrep Extractor System. The

SmartPrep II has FEP tubing as standard and this will

minimize any contamination. All of the sample, rinse

and solvent lines pass through the same valves and

tubing within the instrument. This table demonstrates

one instrument valve position (as per method 537)

and all of the SmartPrep’s internal components,

solvents and the SPE cartridge to demonstrate low

background contribution. In addition the surrogate compounds are within the range specified of 70-130 % recovery.

Figure 2, on the following page, shows a chromatogram of the LFB. It contains the Surrogates and Internal standards as well as a

non-detectable PFOS peak. The Final Concentration of PFOS is well below the DL (0.0037 µg/L, determined by DL in Table 3) and is

not reportable.

Page 3

Table 1. Initial Demonstration of Low System Background

Compound Area INST ( ng/L) Final Conc. (ng/L)

PFBS 0.00 0 0

PFHpA 0.00 0 0

PFHxS 0.00 0 0

PFOA 0.00 0 0

PFNA 0.00 0 0

PFOS 1177 71.6 0.290

SUR C13-PFHxA 176007 10.1 µg/L 101%

SUR C13-PFDA 404135 9.42 µg/L 94.2%

C13-PFOS-(ISTD) 166900 10.0 µg/L 0.0400 µg/L

C13-PFOA-(ISTD) 561125 10.0 µg/L 0.0400 µg/L

Table 2 confirms that the upper and lower limits for the prediction interval of result (PIR) met the recovery criteria for the six PFCs.

Each compound is within the limits and shows consistency throughout the seven runs. The upper criteria’s are set at <= 150% and

the lower PIR is >= 50%.

Table 3 shows the Method Detection Limit (MDL) determined from seven samples run on the SmartPrep over a three day period.

The MDL values for this method with this sample size and equipment are extremely low. They will be able to measure water

samples at the US Health Advisory level of 0.07 µg/L for PFOS or PFOA with MDLs of 0.004 and 0.002 respectively. The limit of

quantitation (LOQ) is 0.013 µg/L for PFOS and 0.0076 µg/L for PFOA, which is acceptable for very low level determinations.

Page 4

Table 2. Minimum Reporting Limit (MRL) Confirmation (ng/L)

Table 3. MDL Study

Compound 1 2 3 4 5 6 7 AVG SD HR for PIR

Upper PIR Limit (%)

Lower PIR Limit (%) CONC

PFBS 64.7 73.6 64.9 70.5 70.9 72.9 68.3 69.4 3.58 14.2 92.9 61.4 90.0

PFHpA 7.94 9.14 7.99 9.16 9.09 8.76 8.44 8.65 0.532 2.11 107 65.4 10.0

PFHxS 21.8 25.2 22.0 25.6 26.3 25.6 25.0 24.5 1.80 7.15 116 63.4 27.3

PFOA 16.1 18.2 16.3 18.3 18.5 17.9 16.9 17.5 0.987 3.91 107 67.7 20.0

PFNA 16.8 18.2 17.0 19.8 19.7 19.0 18.7 18.5 1.19 4.70 116 68.8 20.0

PFOS 27.9 30.6 28.2 33.2 32.1 32.6 32.5 31.0 2.19 8.67 107 60.2 37.2

Compound MDL 1 MDL 2 MDL 3 MDL 4 MDL 5 MDL 6 MDL 7 MDL 8 AVG STD CONC. MDL

(µg/L) LOQ

(µg/L) CONC/MDL

PFBS 0.036 0.039 0.029 0.039 0.041 0.037 0.041 0.039 0.0375 0.0037 0.0450 0.0111 0.0371 4.05

PFHpA 0.004 0.004 0.003 0.004 0.004 0.004 0.004 0.004 0.0041 0.0003 0.0050 0.00099 0.00331 5.04

PFHxS 0.013 0.013 0.010 0.013 0.013 0.012 0.012 0.013 0.0124 0.0013 0.0137 0.00385 0.0128 3.55

PFOA 0.008 0.009 0.007 0.009 0.009 0.008 0.009 0.008 0.0084 0.0008 0.0100 0.00229 0.00764 4.37

PFNA 0.009 0.010 0.007 0.009 0.010 0.009 0.009 0.009 0.0090 0.0008 0.0100 0.00231 0.00772 4.32

PFOS 0.016 0.016 0.013 0.016 0.017 0.016 0.017 0.016 0.0157 0.0013 0.0186 0.00376 0.0126 4.94

Figure 2. Chromatogram of Low System Background on the SmartPrep Extractor II

Table 4. IDC Study, Measure of Precision and Accuracy (Spikes are in ng/L)

Table 5. Laboratory Control Spike

Page 5

Compound Result (ng/L)

Spike Amount (ng/L) Spike Recovery (%)

Acceptable Range (%)

PFBS 99.6 90.0 111 (50-150)

PFHpA 10.6 10.0 106 (50-150)

PFHxS 30.9 27.3 113 (50-150)

PFOA 20.5 20.0 103 (50-150)

PFNA 22.6 20.0 113 (50-150)

PFOS 39.3 37.2 106 (50-150)

Table 4 demonstrates four samples that were extracted using the SmartPrep Extractor II. This data can determine the Initial

Demonstration of Precision (IDP) and Accuracy (IDA). Each need a minimum of four mid-level range replicates. The IDP must meet a

RSD of less than 20% and the IDA must meet an average recovery of ± 30% of the compounds true value.

Each of the compounds pass the criteria for the IDP by demonstrating a range from 0.94%-3.41% RSD. The IDA is within the method

requirement of ± 30% for determining accuracy of the spiked amounts for each of the six PFCs.

Compound 1 2 3 4 AVG CONC. % Rec STD DEV -30% 30.0% RSD

PFBS 640 671 661 629 650 720 90.3 19.1 504 936 2.94

PFHpA 79.4 81.7 78.6 76.5 79.1 80.0 98.8 2.15 56.0 104 2.72

PFHxS 208 218 210 206 210 219 96.2 5.11 153 284 2.43

PFOA 152 155 153 153 153 160 95.9 1.44 112 208 0.94

PFNA 160 170 164 158 163 160 102 5.56 112 208 3.41

PFOS 278 287 274 269 277 297 93.3 7.68 208 386 2.77

SUR C13-PFHxA 34.8 36.3 36.8 34.6 35.6 40.0 89.0 1.09 28.0 52.0 3.06

SUR C13-PFDA 41.6 42.3 43.1 40.9 42.0 40.0 105 0.920 28.0 52.0 2.19

Table 5 and Figure 3 show the results for the Laboratory Control Sample (LCS). The sample is representative of a drinking water

matrix and spiked at a mid-range concentration. The LCS spike recoveries are well within the acceptable range specified by the

method of 50-150%.

Figure 3. Laboratory Control Spike

Table 6. Laboratory Fortified Sample Matrix and Duplicate (LFSM and LFSMD)

Compound Spike Amount

(ng/L) Field Sample Result LFSM (ng/L)

LFSMD (ng/L)

LFSM Recovery (%)

LFSMD Recovery (%) RPD

PFBS 90.0 ND 102 116 113 128 12.7

PFHpA 10.0 ND 8.98 10.5 89.8 105 15.5

PFHxS 27.3 ND 28.5 33.7 104 123 16.8

PFOA 20.0 ND 18.6 20.9 93.3 105 11.8

PFNA 20.0 ND 20.5 21.8 102 109 6.21

PFOS 37.2 ND 37.1 38.9 99.9 105 4.77

Page 6

Table 6 shows the results for a sample matrix and duplicate spiked with the analytes of interest. The agreement of the duplicates

is well within the < 30% criteria for concentrations spiked near native concentrations.

References:

1. Kyle Steenland, Tony Fletcher, and David A. Savitz, Epidemiologic Evidence on the Health Effects of Perfluorooctanoic Acid (PFOA), Environ Health Perspect. 2010 Aug; 118(8): 1100–1108. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2920088/.

2. US UCMR-3 Data Summary, https://www.epa.gov/sites/production/files/2016-05/documents/ucmr3-data-summary-april-2016.pdf, (accessed October 14, 2016).

3. Jennifer Morrison, Chronic exposure limit set for PFOA in drinking water http://cen.acs.org/articles/94/web/2016/05/Chronic-exposure-limit-set-PFOA-drinking-water.html (accessed October 14, 2016).

4. Cheryl Hogue, PFOA moves closer to global restrictions, C&EN, October 3, 2016 (http://cen.acs.org/articles/94/i39/PFOA-moves-closer-global-restrictions.html )

5. Twelfth meeting of the Persistent Organic Pollutants Review Committee (POPRC.12), http://chm.pops.int/TheConvention/POPsReviewCommittee/Meetings/POPRC12/Overview/tabid/5171/Default.aspx (accessed October 14, 2016).

6. USEPA Method 537, https://cfpub.epa.gov/si/si_public_record_report.cfm?dirEntryId=198984&simpleSearch=1&searchAll=EPA%252F600%252FR-08%252F092+ (accessed October 14, 2016).

7. ISO 25101:2009, http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=42742

www.horizontechinc.com 16 Northwestern Drive, Salem, NH 03079 USA ▪ Tel: (603) 893 -3663 ▪ Email: Support-Service@horizontechinc.com

AN1131610_01

Conclusion

PFC compounds are becoming an increasingly important concern as research into health effects uncovers other related illnesses.

As countries around the world move to regulate exposure through drinking water, reliable and accurate measurement methods

become critical.

Methods US EPA 537 and ISO 25101:2009 specify concentration with SPE and analysis using HPLC/MS/MS. Automation of the

cartridge SPE step is important to ensure the best reproducibility with the least chance of contamination. It also ensures the

efficient use of technician time through walk-away operation.

Demonstration of compliance with all requirements of the method was shown. Method detection limits are excellent and

sufficiently low to provide adequate results at the level of interest in the current Health Advisory and in future regulations. The

matrix spike and matrix spike duplicate show excellent spike recoveries at low concentrations. The difference between duplicate

spikes was less than 20% in all cases. The combination of automated SPE for sample concentration and LC/MS/MS for

determination was shown to be effective and contamination free for the low-level determination of PFCs.

Appendix

RA

TE 1

R

ATE

2

LIQ

UID

PU

RG

E TI

ME

Sip

or

Sip

or

SEN

SE

MIN

. EX

PEC

TED

SO

AK

[D

RY

, Rin

se]

V

OLU

ME

De

live

ry

De

live

ry

Tru

e (

ON

) V

OLU

ME

VO

LUM

E TI

ME

N2

Pu

rge

STEP

O

PER

ATI

ON

R

eag

en

t, V

en

t,

or

Sam

ple

(m

L)

(mL/

min

) (m

L/m

in)

Fals

e (

OFF

) (m

L)

(mL)

(s

ec)

(s

ec)

1

Co

nd

itio

n C

artr

idge

M

eth

ano

l 1

5

10

0

FA

LSE

0

0

1

0

0

2

Co

nd

itio

n C

artr

idge

R

eage

nt

Wat

er

18

1

0

0

FALS

E

0

0

0

0

3

Load

Lar

ge V

olu

me

Sam

ple

Sa

mp

le b

ott

le

0

75

1

0

FALS

E

30

0

10

00

0

0

4

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

3

5

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

3

6

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

3

7

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

3

8

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

3

9

Sam

ple

Bo

ttle

Rin

se

Rea

gen

t W

ater

2

.5

30

0

FA

LSE

0

0

0

2

.5

10

Lo

ad L

arge

Vo

lum

e Sa

mp

le

Sam

ple

bo

ttle

0

1

5

10

FA

LSE

1

5

20

0

0

11

N

2 P

urg

e Ti

mer

Ste

p

0

0

0

0

FALS

E

0

0

0

5

12

Sa

mp

le B

ott

le R

inse

M

eth

ano

l 2

.5

30

0

FA

LSE

0

0

0

3

13

Sa

mp

le B

ott

le R

inse

M

eth

ano

l 2

.5

30

0

FA

LSE

0

0

0

3

14

Sa

mp

le B

ott

le R

inse

M

eth

ano

l 2

.5

30

0

FA

LSE

0

0

0

3

15

Sa

mp

le B

ott

le R

inse

M

eth

ano

l 2

.5

30

0

FA

LSE

0

0

0

3

16

Lo

ad S

amp

le F

or

Cle

anu

p

Sam

ple

bo

ttle

2

0

20

8

FA

LSE

0

2

0

1

5

17

C

lean

Sam

ple

Lin

e

Met

han

ol

6

10

0

FA

LSE

0

0

0

0

18

C

lean

Sam

ple

Lin

e

Rea

gen

t W

ater

6

1

0

0

FALS

E

0

0

0

10

19

C

lean

Plu

nge

r M

eth

ano

l 6

1

0

0

FALS

E

0

0

0

0

20

C

lean

Plu

nge

r R

eage

nt

Wat

er

6

10

0

FA

LSE

0

0

0

1

0

Detailed SmartPrep Extractor II Method for PFCs.