Determination of perfluorooctanoic acid (PFOA) extractable from the surface of commercial cookware under simulated cooking conditions by LC/MS/MS

Abstract

Salts of pentadecafluorooctanoic acid (PFOA) are polymerization aids used in the manufacture of fluoropolymers; one of the applications of fluoropolymers is the coating of metal cookware products. A method was developed to determine if PFOA might be present in and extracted from the surface of commercial frying pans coated with a DuPont fluoropolymer under simulated cooking conditions. Commercial grade cookware was obtained, then extracted with water and ethanol/water mixtures at 100 and 125 °C, and the resulting extracts were analyzed by liquid chromatography tandem mass spectrometry (LC/MS/MS). Detection and quantification limits as low as 100 pg cm⁻² were demonstrated. None of the fluoropolymer treated cookware samples analyzed showed detectable levels of PFOA when extracted under simulated cooking conditions.

---

Introduction

Fluoropolymer dispersions are used in some commercial processes for coating metal cookware products. Salts of pentadecafluorooctanoic acid are frequently used as polymerization aids during their manufacture.¹ Recent studies have shown that perfluorinated carboxylic acid anions can persist in the environment and may be found in the liver and blood serum of some birds and animals,^2–5 as well as in house dust.⁶ Other reports^7–10 indicated that low levels of these materials might be present in human blood serum. In recent studies,^11,12 perfluorinated carboxylic acids were detected qualitatively from thermal decomposition of PTFE at temperatures from 360 to 500 °C. Since salts of PFOA are used to suspend and emulsify some polymers during their manufacture, it is necessary to determine if it is still present after the processes used to coat non-stick cookware take place, and could be released into food under typical cooking conditions. A recent publication reported that ethanol, methanol, and water efficiently extracted PFOA from PTFE polymers¹³ both by pressurised fluid and reflux extraction methods. In this study, we performed extractions of commercially-available cookware that had been treated with DuPont fluoropolymer coatings, using water and ethanol/water at temperatures of 100 and 125 °C to simulate cooking conditions for typical and extended time periods. Due to the presence of fats and water in foods, 125 °C is the maximum temperature foods are expected to reach during the cooking process, according to current U. S. Food and Drug Administration (FDA) guidelines. Our studies were designed to be consistent with current FDA guidelines for evaluating the safety of cookware.¹⁴

Experimental

Reagents

Pentadecafluorooctanoic acid (CAS# 335-67-1) was obtained from the Sigma Aldrich Company, Milwaukee, WI, USA and the Oakwood Company, West Columbia, NC, USA. Both standards were 99% pure. The isotopically enriched di-¹³C-PFOA internal standard CF₃(CF₂)₅¹³CF₂¹³C(O)OH (96.4% chemical purity) was custom synthesized for DuPont, Haskell Laboratory (Newark, DE). Reagent grade sodium thiosulfate and ammonium acetate were obtained from Sigma Aldrich. Methanol (HPLC grade) was obtained from JT Baker Company, Phillipsburg, NJ, USA. Ethanol (Denatured, SDA Formula 1), acetonitrile (HPLC grade) and acetic acid (HPLC grade) were obtained from EM Scientific, Gibbstown, NJ, USA.

Extraction

Water extractions. Eleven frying pans (five uncoated stainless steel and six anodized aluminum pans coated with DuPont fluoropolymer material) were purchased from a local retailer. The total surface area of each pan was 420 cm⁻². A FDA procedure¹⁴ was modified as noted. The pans were first washed as recommended by the manufacturer with approximately 50 mL of 0.5% detergent solution and then rinsed with three 50 mL portions of deionized water and dried with a towel. A custom-made glass lid (23 cm diameter × 0.3 cm thickness) with a #19 female ground glass joint attached in the centre and fitted with a condenser was washed with the detergent solution, rinsed with deionized water and dried. The pan was placed on a Vulcan Model 38 L gas stove and deionized water was added to approximately 0.6 cm from the top. The lid and condenser were placed on the pan, and two 200 g weights were added to the lid to ensure a good seal. The condenser was cooled with wet ice. The frying pan with water and condenser was heated until the water refluxed continuously inside the condenser. The heat was turned down and reflux was allowed to continue for 30 min. Then the apparatus was removed from the burner and allowed to cool.

The cooled water was then placed in each of two 500 mL wide-mouth polypropylene bottles. The lid and condenser were rinsed with deionized water, and the pan was rinsed with three 50 mL portions of deionized water. All rinse water was added to each of the two polypropylene bottles. Duplicate 40 mL aliquots of each water sample were treated with 200 µL of 250 mg mL⁻¹ sodium thiosulfate solution. Solid phase extraction cartridges (Sep Pak Vac 1 g/6 mL C18 cartridges, Waters, Milford, MA, USA) were used for concentrating the samples. The cartridges were preconditioned with 10 mL of methanol followed by 5 mL of HPLC-grade water. The aliquot of the sample was loaded onto the cartridge, which was then washed with 5 mL of 40% methanol in water, then eluted with 5 mL of 100% methanol. The eluate was collected for analysis by LC/MS/MS. The treatment results in an eight-fold concentration prior to analysis.

Ethanol/water extractions. Further studies were conducted to simulate cooking conditions other than boiling water. Ethanol/water mixtures were used to extract frying pans at a temperature of 125 °C, which is the maximum temperature food is expected to reach during cooking.¹⁴ In order to carry out solvent extractions at this elevated temperature, a pressurized fluid extraction apparatus was used (Accelerated Solvent Extractor (ASE^®) Model 200, Dionex Corporation, Sunnyvale, CA, USA). An ethanol/water mixture of 1 ∶ 9 was used to simulate watery and acidic foods, and a mixture of 19 ∶ 1 ethanol/water was used to simulate fatty or oily foods, which was consistent with FDA guidelines.¹⁵ Fifteen frying pans (five uncoated stainless steel and five aluminum pans coated with DuPont fluoropolymer material, and five anodized aluminum pans coated with DuPont fluoropolymer material) were purchased from a local retailer. Each pan was cut into multiple rectangular pieces measuring 1.5 × 7.5 cm using a band saw. Pressurized fluid extraction was performed by placing two of the rectangular pieces back to back, with the treated surfaces facing outward, in a 22 mL ASE cell. The cells were then extracted using an ASE 200 operated at 125 °C and 1000 psi. No preheating was done, and the static extraction time was 10 min. The flush volume was 5%, the purge time was 60 s and one cycle was performed.

Each extract volume was approximately 25 mL. Di-¹³C-PFOA internal standard (100 µL of a 100 ng mL⁻¹ solution in water) was added to each collection vial and mixed. The extracts in 1 ∶ 9 ethanol/water were analyzed directly by LC/MS/MS. The extracts in 19 ∶ 1 ethanol/water were prepared for analysis by taking a 2 mL aliquot and adding 0.5 mL of water. The aliquots were then evaporated to a volume of 0.5 to 0.7 mL under a stream of nitrogen, and diluted to 2 mL with water. (The evaporation and dilution were necessary to avoid adding strong solvent to the liquid chromatographic system.) The resulting solutions were then analyzed by LC/MS/MS.

Equipment

Water extractions. A Micromass Quattro Ultima tandem mass spectrometer (MS/MS) System (Waters Associates, Milford, MA, USA) coupled to a Hewlett Packard Series 1100 liquid chromatograph (Agilent, Little Falls, DE, USA) was used to analyze the eluate from the water extractions. A 4 × 10 mm Hypercarb drop-in guard cartridge (Keystone Scientific, Bellefonte, PA. USA) was attached in-line after the purge valve and before the sample injector to trap any residual contaminants that may have been in the mobile phase and/or components of the HPLC system. A 2.1 mm × 50 mm, 4 µm Genesis C8 (Jones Chromatography, Foster City, CA, USA) column was used at 35 °C. Upon injection of 15 µL of sample, a gradient consisting of Mobile Phase A (2 mM ammonium acetate) and Mobile Phase B (methanol) was initiated at a flow rate of 0.3 mL min⁻¹ using the time program shown in Table 1.

Table 1 HPLC gradient used for water extractions at 100 °C

Time/min	A (%)	B (%)
0.0	60	40
0.4	60	40
1.0	10	90
7.0	10	90
7.5	0	100
9.0	0	100
9.5	60	40
15.0	60	40

---

The electrospray interface was operated in the negative ion mode. The parent to daughter ion transition at 413 → 369 Da was monitored.

Ethanol/water extractions. A Micromass Quattro II tandem mass spectrometer (MS/MS) System (Waters Associates) coupled to a Hewlett Packard Series 1100 liquid chromatograph (Agilent) was used to analyze the samples from the ethanol/water extractions. A 30 mm × 4.6 mm id, 3 µm Luna C-8(2) column (Phenomenex, Torrance, CA) was attached in-line after the purge valve and before the sample injector to trap any residual contaminants that may have been in the mobile phase and/or components of the HPLC system. A 15 cm × 2.1 mm id, 5 µm Zorbax, Rx-C8 (Agilent Technologies, Little Falls, DE) column was used at 30 °C. Upon injection of 50 µL of sample, a gradient consisting of Mobile Phase A (0.15% acetic acid in water) and Mobile Phase B (acetonitrile) was initiated at a flow rate of 0.4 mL min⁻¹ using the time program shown in Table 2.

Table 2 HPLC gradient used for ethanol/water extractions at 125 °C

Time/min	A (%)	B (%)
0.0	95	5
0.5	20	80
9.0	20	80
9.1	95	5
14.0	95	5

---

The HPLC column effluent was split 1 ∶ 10 to reduce the flow into the mass spectrometer. The electrospray interface was operated in the negative ion mode. The parent to daughter ion transition at 413 → 369 Da was monitored for PFOA and the transition at 415 → 370 Da was monitored for the dual ¹³C PFOA internal standard.

Results and discussion

Recovery experiments

Experiments were conducted to demonstrate that PFOA is extracted and recovered using boiling water as an extraction solvent. Fortifications corresponding to a level of 400 pg cm⁻² were made directly to the surface of pans previously shown not to contain extractable PFOA by adding the fortifying solution in acetonitrile and allowing the solvent to evaporate at room temperature. The extraction with boiling water was then carried out, and the di-¹³C internal standard was added to aliquots of the extract before analysis. The internal standard was not added to the sample prior to extraction so that any losses in the extraction would be more apparent. The mean recovery level for four replicate fortifications was 88% with a relative standard deviation of 3%.

A quality control program was developed in order to determine the background level expected in the ethanol/water analysis and to demonstrate recovery of any PFOA that might have been present. The PFOA levels found on the stainless steel (control) pans were similar to those found in the ethanol and water solvents when they were directly injected. Therefore, this level was taken as the background level, and corresponds to a level of 40 pg mL⁻¹ in the final extract, or approximately 50 pg cm⁻² of pan surface area. The limit of detection, LOD, was set at twice this value, or 100 pg cm⁻². Control pan samples, injections of diluents and instrumental background checks were run with every analytical set, and the background level remained consistent.

Fortifications were made at levels corresponding to 100 and 1000 pg cm⁻². Fortifications were made in three different ways. First, PFOA solution was applied directly to the surface of both treated and untreated coupons, to demonstrate recovery of any material that may have been present. In this case, the fortification was made using a small volume (22.5 or 225 µL) of acetonitrile, which was allowed to evaporate prior to extraction. Second, PFOA solution was added to aliquots of the extract from control pans, in order to demonstrate the absence of matrix effects. Finally, PFOA was fortified to sand packed into ASE cells, to demonstrate that PFOA was neither adsorbed on the extraction apparatus nor converted to its ethyl ester under extraction conditions. The results of the fortification experiments are shown in Table 3. The internal standard was always added to the collection vial after extraction, so that any losses during extraction would be more apparent and could be addressed if necessary.

Table 3 Fortification data obtained during ethanol water extractions

Fortification type	Fortification level/pg cm^−2 (n^a)	Recovery (%) (RSD^b (%))
a Number of replicates.b Relative standard deviation.c Results combined for 1 ∶ 9 and 19 ∶ 1 ethanol/water extractions.
To surface of aluminum pan^c	100 (4)	96 (15)
	1000 (4)	87 (3)
To surface of anodized aluminum pan^c	100 (4)	100 (15)
	1000 (4)	86 (11)
To aliquot of stainless steel pan extract^c	100 (4)	96 (15)
	1000 (2)	97, 95
To sand in ASE cell^c	1000 (4)	98 (8)

---

It is evident that acceptable recovery was obtained at levels equivalent to both 100 and 1000 pg cm⁻², when fortifications were made directly to the surface of the pans. Since acceptable recovery data were obtained at 100 pg cm⁻² (overall average = 98%, RSD = 14%), we are defining that level as our LOQ as well as our LOD. Fortifications made to aliquots of control pan extracts demonstrated that matrix suppression, which was noticed through examination of the internal standard responses in the samples and standards, was compensated by use of the stable isotope internal standard. Finally, the fortifications made to sand in the ASE cells demonstrated that esterification did not occur and that PFOA was not lost in the apparatus. The presence of water and lack of a suitable catalyst most likely prevented this.

Analyses of commercial samples

Fluoropolymer-coated and uncoated pans were obtained commercially and subjected to the same extraction procedure. The results from analyses of the water extractions are shown in Table 4. All water from the fluoropolymer-coated frying pans showed no quantifiable PFOA present. With a limit of quantification (LOQ) of 50 ppt and a limit of detection (LOD) of approximately 10 ppt per aliquot, this number corresponds to 60–130 pg cm⁻² LOQ or 10–30 pg cm⁻² LOD, depending on the volume of water collected.

Table 4 Results of analysis of water extractions at 100 °C

Pan type	n^a	PFOA/ng L^−1
a Number of replicates.b Not detected (LOD = 10 ppt).
Stainless steel	5	nd^b
Coated anodized aluminum	6	nd^b

---

One of the stainless steel frying pans (with no coating) gave a positive result for one aliquot and a non-detect for the second aliquot. Since PFOA is easily adsorbed onto analytical apparatus, or can be a component of fluoropolymeric parts, background levels can be problematic. PFOA can be found in analytical solvents as well as transfer lines in chromatographic systems. The background levels are usually around 10 ppt. The US EPA Guideline¹⁶ recommends that any measurements obtained that are less than five times the blank level should be reported as detected but not quantified (NQ), because the quantity reported cannot reliably be discriminated between the sample and the background. Due to background levels and the possibility for carryover or laboratory cross-contamination, the limit of quantification was set at 50 ppt in the extract, which corresponds to 200 pg cm⁻² in the sample.

The results of the analysis of ethanol/water extractions are summarized in Table 5. None of the pans in the coated aluminum or anodized aluminum groups showed detectable levels of PFOA (<100 pg cm⁻²). When coupons were extracted three successive times for 10 min, no PFOA was detected. Similarly, when the extraction time was increased to 90 min, no PFOA was detected. Finally, one pair of coupons from each set was scored multiple times with a utility knife, and no PFOA was detected during subsequent extractions.

Table 5 Results from analysis of ethanol/water extractions at 125 °C

Cookware type, conditions	PFOA detected/pg cm^−2
	Acid/watery food simulation (n^a)	Fatty food simulation
a Number of replicates.b Not detected (LOD = 100 pg cm^−2).
Stainless steel, 10 min extraction	nd^b (5)	nd (5)
Aluminum, 10 min extraction	nd (5)	nd (5)
Anodized aluminum, 10 min extraction	nd (5)	nd (5)
Aluminum, abraided	nd (1)	nd (1)
Anodized aluminum, abraided	nd (1)	nd (1)
Aluminum, 3 × 10 min extraction	nd (1)	nd (1)
Anodized aluminum, 3 × 10 min extraction	nd (1)	nd (1)
Aluminum, 3 × 90 min extraction	nd (1)	nd (1)
Anodized aluminum, 3 × 90 min extraction	nd (1)	nd (1)

---

Individual results for all quality control samples as well as all cookware tested can be found in the supplementary information to this paper.

The absence of PFOA from the coated cookware was not surprising. Fluoropolymers are typically manufactured near ambient temperature. Fabrication processes, such as those used to coat cookware, require temperatures of greater than 300 °C (typically 350 to 450 °C). These temperatures, and the large surface area of the coating material, would easily vaporize any PFOA (boiling point 189 °C¹⁷) that may have originally been present. Also, Krusic and Roe¹⁸ conducted gas-phase NMR studies where the ammonium salt of PFOA was observed to completely decompose within minutes at 234 °C.

Under common cooking conditions and using food stimulants (water and water/ethanol), no PFOA is detected from either coated or uncoated cookware with detection limits as low as 100 pg cm⁻².

Acknowledgements

The authors thank Bert Sooy, Dale E. Lenker, William W. Crook and Robert E. Borton of DuPont Fluoroproducts, and Steven W. George of DuPont Haskell Laboratories for technical assistance and John Flaherty of Exygen Research (State College, PA, USA) for providing the analytical results from the water extractions.

References

1. Guide to the Safe Handling of Fluoropolymer Dispersions, The Society of the Plastics Industry, Inc., 2001, p. 7.
2. J. W. Martin, M. M. Smithwick, B. M. Braune, P. F. Hoekstra, D. C. G. Muir and S. A. Mabury, Environ. Sci. Technol., 2004, 38, 373–380.
3. K. Kannan, J. Newsted, R. S. Halbrook and J. P. Giesy, Environ. Sci. Technol., 2002, 36, 2566–2571.
4. M. K. So, S. Taniyasu, N. Yamashita, J. P. Giesy, J. Zheng, Z. Fang, S. H. Im and P. K. S. Lam, Environ. Sci. Technol., 2004, 38, 4056–4063.
5. J. P. Giesy and K. Kannan, Environ. Sci. Technol., 2002, 36, 146A.
6. H. Moriwaki, Y. Takata and R. Arakawa, J. Environ. Monit., 2002, 5, 753–757.
7. K. J. Hansen, L. A. Clemen, M. E. Ellefson and H. O. Johnson, Environ. Sci. Technol, 2001, 35, 766–770.
8. G. W. Olsen, T. R. Church, J. P. Miller, J. M. Burris, K. J. Hansen, J. K. Lundberg, J. B. Armitage, R. M. Herron, Z. Medhdizadehkashi, J. B. Nobiletti, E. M. O'Neill, J. H. Mandel and L. R. Zobel, Environ. Health Perspect., 2003, 111, 1892–1901.
9. K. Kannan, S. Corsolini, J. Falandysz, G. Fillmann, K. S. Kumar, B. G. Loganthan, M. A. Mohd, J. Olivero, N. van Wouwe, J. H. Yang and K. M. Aldous, Environ. Sci. Technol., 2004, 38, 4489–4495.
10. S. Taniyasu, K. Kannan, Y. Horii and N. Yamashita, Environ. Sci. Technol., 2003, 37, 2634–2639.
11. D. A. Ellis, S. A. Mabury, J. W. Martin and D. C. G. Muir, Nature, 2001, 412, 321–324.
12. D. A. Ellis, J. W. Martin, D. C. G. Muir and S. A. Mabury, Analyst, 2003, 128, 756–764.
13. B. S. Larsen, M. A. Kaiser, M. Botelho, G. R. Wooler and L. William Buxton, Analyst, 2004, 130, 59–62.
14. US FDA, see 21 CFR 175.300(d), table 2 condition of use B.
15. US FDA “Guidance for Industry – Preparation of Food Contact Notifications and Food Additive Petitions for Food Contact Substances: Chemistry Recommendations” Final Guidance, April 2002, Appendix I.
16. “Data Validation Standard Operating Procedure for Contract Laboratory Program Routine Analytical Services”, US EPA Region IV, Athens, Georgia, Revision 2.1, July 1997.
17. E. A. Kauck and A. R. Diesslin, Ind. Eng. Chem., 1951, 43, 2332–2334.
18. P. J. Krusic and D. C. Roe, Anal. Chem., 2004, 76, 3800–3803.

---

Footnote

† Electronic supplementary information (ESI) available: Supplementary tables of information. See http://dx.doi.org/10.1039/b505377c

---