1. Quantitative comparison of perfluorinated active substances in drinking water between tandem triple quadrupole MS/MS and high resolution mass spectrometry using orbitrap technology – knowns and unknowns
Ali Wolfgang Haghani (Eurofins Eaton Analytical)
Andy Eaton, PhD, BCES (Eurofins Eaton Analytical)

2. Perfluorinated compounds as emerging contaminants
~3000
Perfluoroalkyl substances: Compounds for which all hydrogens on all carbons (except for carbons associated with functional groups) have been replaced by fluorines e.g., perfluoroalkyl acids (e.g., PFOA, PFOS)
Polyfluoroalkyl substances: Compounds for which all hydrogens on at least one (but not all) carbons have been replaced by fluorines e.g., fluorotelomer-based compounds. Carbon-hydrogen linkages allow for biotic and abiotic degradation
The C–F bond is is considered the strongest single bond in organic chemistry with a bond enthalpy of 481 kJ/mol in CH3F, which is substantially higher than that of other bonds. This pronounced bond strength is reflected in the notorious environmental and chemical stability of PFC as POP.

3. Publications shows continued interest in PFAS.
From
2000
1000

4. Minimum Reporting Level
PFAS Occurrence – UCMR 3
2012: Six PFASs added to Unregulated Contaminant Monitoring Rule 3 (UCMR 3) list, including PFOS and PFOA using EPA 537 method. October 2015 UCMR 3 data summary
Contaminant
Minimum Reporting Level
PFOS
0.04 µg/L
PFOA
0.02 µg/L
PFNA
PFHxS
0.03 µg/L
PFHpA
0.01 µg/L
PFBS
0.09 µg/L
Six PFASs under UCMR but >100s of compounds exists.
Literature:
WWTPs exhibited a different set of PFASs as well as drastically different concentrations between industrial and municipal WWTPs. Samples from industrial WWTPs exhibited 32 (influent), 20 (effluent), and 34 (air above influent) PFASs of the 65 PFASs under investigation, whereas samples from municipal WWTPs only contained 9 (influent), 10 (effluent) and 7 (air) compounds, respectively.
2016: EPA issued a final lifetime drinking water Health Advisory for PFOA and PFOS of 0.07 μg/L

5. EPA 537 Method Used – Summary: Extraction
A 250-mL preserved water sample with Trizma is fortified with surrogates and passed through a solid phase extraction (SPE) cartridge containing Solex HRPHS in lieu of polystyrenedivinylbenzene (SDVB) to extract the method analytes and surrogates. The compounds are eluted from the solid phase with a small amount of methanol. The extract is concentrated to dryness with nitrogen in a heated water bath, and then adjusted to a 1-mL volume with 96:4% (vol/vol) methanol:water after adding the IS(s). A 5-μL in lieu of 10-μL injection is made into an LC equipped with a C18 column.
For this presentation we decided to evaluate Q-Exactive HRAM quantitation and sensitivity using EPA 537 standardized method written for triple quads with some minor cahanges to expand the scope of compounds using EPA 537.

6. EPA 537 Method M. Used – Mas Spec
Q-Exactive hybrid HRAM capable of producing MS/MS data in lieu of “low resolution triple” -MS/MS. The analytes are separated and identified by comparing the acquired mass spectra and retention times to reference spectra and retention times for calibration standards acquired under identical LC/MS/MS conditions. The concentration of each analyte is determined by using the internal standard technique. Surrogate analytes are added to all Field and QC Samples to monitor the extraction efficiency of the method analytes.
For this presentation we decided to evaluate Q-Exactive HRAM quantitation and sensitivity using EPA 537 standardized method written for triple quads with some minor cahanges to expand the scope of compounds using EPA 537.

7. Hybrid HRAM can be used for EPA LC-MS/MS Analysis
Using Parallel Reaction Monitoring (PRM) - Targeted MS2
Serial monitoring
PRM is the closest set up to an MRM used in a triple quad used in the method.
The only difference is that all fragments are collected concerted high resolution mass analysis.
Parallel monitoring
Parallel reaction monitoring for high resolution and high mass accuracy quantitative, targeted proteomics. Peterson et al., MCP 2012, O

8. Q-Exactive tune page: the setup.
No need for optimization
Targeted + Non Targeted
Targeted
Needs RT and specific collision E.
For Full scan no optimization is needed.
For PRM retention time, selective compound formula or monoisotopic MW and collision energy is needed to be included in the inclusion list.

9. EPA 537 – FULL-MS at 70K resolution showing good peak shapes, and S/N for a 2.5ppt standard.
PFBS
PFHXA
PFHXS
PFHpA
PFOA
PFNA
PFOS
PFDA
PFuNA
As the homologous CF2 back bone increases the compounds becomes less polar exhibiting greater RT on a RP column. Sulfonates group is less ionic polar than carboxylic acid hence elute later than PFCA with equal number of carbon backbone. ie. PFOS elutes later than PFOA both having C8.
Boxed compounds from UCMR 3 study.
PFDoA
PFTrDA
PFTA

10. EPA 537M (PRM) – targeted-MS2 for PFOA at the lowest cal std 0
EPA 537M (PRM) – targeted-MS2 for PFOA at the lowest cal std 0.5ppt shown with spectra confirmation using Trace Finder 4.1
Great S/N at 0.5ppt on the 250x concentrated extract. Library and RT matching increases in confidence of identifying the correct compound. Good linearity ppt forced through zero to check for systematic contaminations..
Sample MS2 Spectra
Library MS2 Spectra

11. By adding Full-Scan to PRM workflow – 0.5ppt PFOA shown
In Fs area is 10X PRM due to no transition loss, however the S/N is less than in PRM being less selective. Isotopic pattern for confirmation can be utilized and mass accuracy <2PPM. Again a good linearity through all points.
Isotpic pattern
Isotopic pattern match

12. Determination of Minimum Reporting Limit Using LCMRL Using Regressional LOQ Calculation
EPA for drinking water uses LCMRL calculation for determination of minimum reporting limit (MRL). It has accuracy in addition to precision.

13. Target & non targeted compounds EPA 537 plus.
PRM
Full Scan
 EPA 537 PFCA’s, and PFSA’S target list
Critical level DL
LCMRL
PFBS
0.077
0.12
<0.5
0.15
0.2
PFDA
0.18
0.26
PFDoA
0.14
0.29
0.47
0.73
PFHpA
0.35
0.97
0.09
PFHxA
0.16
0.27
0.13
0.19
PFHxS
0.52
0.77
1.7
2.4
PFNA
0.11
0.17
PFOA
0.36
0.5
0.22
PFOS
0.21
PFTA
0.48
0.71
PFTrDA
0.32
0.31
0.55
PFuNA
0.72
0.38 1
PFBA
0.64
PFODA
PFDS
PFHxDA
PFPA
LCMRL equal or better than high end - mid range triple quads – background contamination is the limiting factor.
Sensitivity is not an issue for these class of compounds. Limited by the background concentrations'.
Stock standard contained other compounds not part of EPA 537 target list which were identified and quantified using Full-MS

14. A 2.5ppt standard of PFOA. Excellent quantitation and sensitivity is obtained with HRAM in comparison to QQQ analysis.
Apple to orange but to show that similar sensitivity can be achieved. Quantiva SRM ~ 2x HRAM PRM.
a) SRM Analysis-QQQ
b) HRAM Full Scan
c) HRAM PRM

15. PFOS quantification is challenging: Showing a technical grade standard
Isomer 1 2 3 4 5 6 7 8 9 10 11
MS/MS relative response factors
m/z m/z 99
100
117 97 49 39 43 78 19
m/z m/z 80
135
241
142
123
113
118
220 90
EEA
There is no perfect practical way for quantification of branched compounds due to possible different ratio and RF found in branched isomers. They can differ from calibration std linear or branched vendor dependent).
For PFOS 499/99 is in general low bias as 499/80 is high bias. Some isomers have no response and others around 200% compared to linear isomer.
Nicole Riddell etal 2009

16. Samples from different locations can have different branch ratios
Field Sample
Std PFOS Branched
Bottom right branch calibration std does not match the branch ratio’s of samples. 4% RPD up to 23% is shown. The method does call +/- 30% accuracy so that might not be as a great critical issue but if a sample is at compliance level such as in US at 70ppt that might make a big difference.
Quantitation comparison of full scan in a Q-Exactive to SRMs 499>80 and 499>99 (represented as peak area ratios). Results suggest that full scan will have less bias and be close to average of using two SRMs for quantitation.

17. Linear PFOS used for calibration
Fs scan covers all of the branches and looks to be more reliable for PFOS quantitation.
10 samples - conc ppt
Full scan close to the average of the two MRM and less prone to other factors effecting isomers RF.
Oust side US 499/99 is mostly used as EPA 537 uses 499/80. United Nations Environment Program (UNEP) has suggested to take the average of the two MRM’s when using a triple quad which makes the results closer to the Fs quantitation.
Linear PFOS used for calibration

18. Till around 1970 PFOA was also produced by Electrochemical fluorination (ECF) process creating branched isomers which can still be detected occasionally so it is important to also integrate the branch isomers.

19. Full Scan 70000 resolution PFNA PFHpA PFOA
Possible of other than PFOA to contain branched isomers.

20. FS and 413/369 compares well in real field samples having branched PFOA - ppt

21. Including branched isomers when present about 20% difference in a technical grade

22. A UCMR3 sample shown having a trace hit for non-targeted known compound: PFDS
80 ppt STD
Post run identification using build in house library for Fs. Using isotopic pattern, accurate mass and RT for confirmation. Sample shows branched isomers as expected.

23. Looking for Unknowns : Contaminated GW in a AFFF site: On-line SPE UHPLC / Fs-ddms2, top 5
Fast 9 minute run set up direct injection 250uL of a contaminated ground samples from AFFF site for identification of known/suspects/unknowns.
FTOHs are very delicate species with respect to their ESI-MS performance. If only traces of organic anions, such as formiate or acetate are present, FTOHs will form adducts such as [M + HCOO] or [M + CH3COO]. Under exclusion of salts of these ions, the deprotonated molecule is formed. It was discovered that MeOH favors its formation, whereas ACN inhibits.
Loading eluent 0.05% formic acid, Gradient 20 mM ammonium acetate, ACN. As organic
2.1X50 thermo aq gold 1.9 um.
Flow 0.3ml/min
Online HRPHS SPE

24. Data mining software using “Compound Discoverer”

25. Refining data to explore known classes of compounds
C90 H100 F100 S10 O10 P5 I3
C90 H F100 S10 O3 P5 I3
RDB
Isotopic pattern
Possible formula
Mass defect
Mass range
Etc.
Top left about 2000 individual compounds were identified, Top right after tighter more related filter about 200 compounds are shown. Further looking for known characteristics pattern suspects can be visualized.
Negative Mass defect - Fluorine has an atomic number of 9 and a relative atomic weight of u. This negative mass defect leads to substantially lower monoisotopic masses of highly fluorinated compounds than the respective nominal mass.
For PFAA the RDB = 0 because of saturated fluorinated backbone.
SO3 has good isotopic pattern mostly because of Sulfur containing element.
F lacks any isotopic pattern and is monoisotopic element – a general withdraw for identification.
Pattern MW vs RT

26. At higher resolutions more trace isotopic patterns can be used for MS scans.
70K Res.
34S
140K Res.
13C8

27. Ms2 Data Shows CmF2m+1- “9 series” and CnF2nSO3- radical “0 series” as expected.
Fragmentation of classic PFCs, such as PFOS and PFOA has been studied very thoroughly. The PFSA are mainly known to produce sulfur-containing ions, such as SO3 (m/z 80) and FSO3 (m/z 99). However, this is only one part of the story. Besides the two sulfur-containing ions, product ions of linear PFOS comprise – albeit to a low extent – perfluoroalkyl carbanions CmF2m+1 (the so-called “9-series,” since the m/z values end with “9”) as well as CnF2nSO3 radical anions (the so-called “0-series”). The latter ones are suspected to be generated by initial radical cleavage of a C–C-bond within the perfluoroalkyl chain and subsequent losses of tetrafluoroethene. The
perfluoroalkyl carbanions are supposed to derive from initial loss of SO3 and subsequent loss of perfluoroalkenes such as tetrafluoroethene, hexafluoropropene, and so forth.
SO3 is the first fragment energetically favorable, hence the most intense compared to the other ones.

28. The structure of the selected compound can be drawn in “Custom Explanation” using Mass Frontier to check against MS and MS2 collected data
C2 and C3 are less known from PFSA’s and seen less, but the expectation is to see these more in future due to Stewardship program for reducing production of longer chains such as PFOS.

29. Summary of all PFAS’s found for the studied group.
C2HF5O3S
Less studied. No commercial std’s available
C2 and C3 are less known from PFSA’s and seen less, but the expectation is to see these more in future due to Stewardship program for reducing production of longer chains such as PFOS.
C2 TO C8 PFSA’s
Study sample / area
Std column / area

30. A few identified suspects using prescribed workflow which were missing from the built in library. Some need additional confirmations. C8
C C10
C C8
hexadecafluorononanoic acid
3:2 FTOH
difluoro{[(3E)-1,1,2,2,3,4,5,5,5-nonafluoropent-3-en-1-yl]oxy}methanesulfonic acid
C C8
C C8
C C8
octafluorobutane-1-sulfonic acid
3-[amino(hydroxy)methyl]-2-[(2,2,2-trifluoroethyl)amino]pentanoic acid
nonafluorobutane
Nomenclatures are not added due to lengthy IUPAC names for fluorinated backbone chains.
3:2 FTOH, and 3:2 FTA . From these two 3:2 FTA had only MS data no MS2 due to lower intensity than the method was set for Fs-ddms2 set up for this study. These two can potentially be of interest since FTOH are known to produce PFCAA during biotransformation and FTA is an intermediator compound in the process. Which we were able to find the PFCAA according to wang et al.
sulfonamine known precursor for PFHXS and PFPS and others depending on chain length in aerobic biotransformation.
Non-target screening by HPLC coupled to high-resolution mass spectrometry (HRMS) revealed the presence of a whole homologous series of 2H-PFCAs in several industrial WWTP samples, which underlines the importance of expanding the target PFAS list in monitoring campaigns. 2H-PFCAs are known persistent biotransformation products of fluorotelomer-based compounds.
C C8
C C8
C C8
undecafluorohex-4-ene-1-sulfonic acid
3:2 FTA
heptafluoropropane-1-sulfonamide

31. Conclusions
Q-Exactive HRAM instrumentation in the PRM scan mode can be used for quantitation with performance like a triple quadrupole in SRM mode with added specificity, selectivity and comparable sensitivity.
Full scan HRAM can likely produce more accurate quantitative data for compounds that contain branched isomers such as PFOS.
Routine quantitative workflows and non-target analysis can be performed in a single analysis.
HRAM data processing using Thermo Fisher Scientific Compound Discoverer software can simplify complex data reduction/save time.
Other techniques may be necessary for further confirmation of suspects/unknowns structures such as MSn, 13C and 19F NMR, when standards are not commercially available.

32. Any Questions?
Ali Wolfgang Haghani (Eurofins Eaton Analytical) Andy Eaton, PhD, BCES (Eurofins Eaton Analytical) Collaborators: Richard Jack, PhD (ThermoFisher) Ed George, PhD (ThermoFisher) Charles Yang T., PhD (ThermoFisher)