You stare at the chromatogram. On top of that, baselines drifting. Peaks where they shouldn't be. And that nagging question: what exactly is in this vial?
A sample containing atoms of C and F was analyzed last week in our lab. In practice, the results weren't what the client expected. They rarely are.
Here's the thing about carbon-fluorine bonds — they're stubborn. Practically speaking, incredibly stable. Even so, that's why PFAS compounds persist in groundwater for decades. It's also why analyzing them requires a completely different mindset than your standard organic workup.
What Are We Actually Looking At
Carbon and fluorine. Worth adding: bond dissociation energy around 485 kJ/mol. Two elements that form some of the strongest single bonds in organic chemistry. For context, a typical C-H bond sits near 413 kJ/mol. That difference changes everything.
When a sample contains C-F bonds, you're dealing with fluorocarbons, hydrofluorocarbons, perfluorinated compounds, or the massive PFAS family. Could be a refrigerant. Could be a pharmaceutical intermediate. Could be contamination from firefighting foam at an airport site three states away.
The analytical challenge isn't detecting carbon or fluorine individually. Any decent elemental analyzer handles that. The challenge is structural elucidation* — figuring out how those atoms connect, what the chain length is, whether you have branching, ether linkages, or terminal groups that change the toxicity profile entirely.
The PFAS Elephant in the Room
Per- and polyfluoroalkyl substances deserve their own section because they've rewritten the analytical rulebook. Over 12,000 known structures. Most labs routinely quantify maybe 40-70 of them. In real terms, the rest? "Dark matter" PFAS — precursors, transformation products, isomers that don't match any library.
If your sample came from environmental water, wastewater, biosolids, or landfill leachate, you're not just analyzing a compound. You're chasing a fingerprint.
Why This Analysis Keeps Chemists Awake
Three words: matrix, volatility, and ionization.
Environmental matrices hate you. Sediment grabs analytes and won't let go. So humic acids co-extract and suppress ionization. Blood serum proteins bind PFAS so tightly that protein precipitation becomes its own optimization project.
Then there's volatility. Because of that, longer chains? Shorter-chain fluorotelomers? They'll evaporate during nitrogen blow-down if you look at them wrong. They stick to every surface — PTFE tubing, vial caps, even the septa you thought were inert.
And ionization. Electrospray negative mode works beautifully for carboxylic acids. Now, sulfonates too. But neutral PFAS? Fluorotelomer alcohols? They need APCI or chemical ionization. Maybe derivatization. Every choice adds steps, and every step adds uncertainty.
The Regulatory Hammer
EPA Method 1633. Practically speaking, eU Drinking Water Directive. State-level MCLs dropping into single-digit parts per trillion. The regulatory landscape shifts faster than method validation cycles. Consider this: a lab that validated 18 analytes in 2020 now needs 40. By next year, maybe 60.
Clients don't care about your validation timeline. They need compliant data yesterday.
How the Analysis Actually Works
Let's walk through a real workflow. Not the textbook version — the one where instruments behave and standards are pure.
Sample Prep: Where Battles Are Won or Lost
Solid phase extraction (SPE) remains the workhorse. Maybe mixed-mode for broader coverage. But here's what textbooks skip: cartridge lot variability matters. That's why weak anion exchange (WAX) cartridges for acidic PFAS. We've seen 20% recovery differences between lots from the same manufacturer.
For solids — soil, sediment, biosolids — you're looking at alkaline digestion or methanol extraction with shaking, sometimes sonication. Even so, then cleanup. In real terms, lots of cleanup. Graphitized carbon, PSA, C18. Which means each step loses analyte. Each step adds background.
Aqueous samples seem easier. Whole bottle extraction avoids subsampling error but loads more matrix onto SPE. They're not. Direct injection? Only if your instrument tolerates dirty samples and your reporting limits allow it.
LC-MS/MS: The Standard, With Caveats
Reverse-phase C18. Gradient elution with ammonium acetate/methanol or acetonitrile. Scheduled MRM transitions for each analyte. Sounds routine.
Until you realize PFHxS and PFOS elute within 0.15 minutes of each other on some columns. Until you see branched isomers splitting peaks that your integration software wants to call one compound. Until you discover that PFBA co-elutes with the solvent front unless you use a trapping column or a different stationary phase.
Column choice matters more than vendors admit. We've tested seven C18 columns on the same sample set. Retention time shifts of 0.5+ minutes. Resolution differences that change whether you report a detect or a non-detect.
And the mass spec? Source cleaning becomes a weekly ritual. Fluorinated compounds deposit on cones, lenses, the entrance aperture. Sensitivity drops 30% in two weeks if you're running high-throughput environmental work.
HRMS: When You Need the Full Picture
High-resolution mass spectrometry (Orbitrap, TOF) changes the game for non-targeted analysis. Exact mass gives you elemental composition. MS/MS fragmentation gives you structural clues. You can screen for 1,000+ PFAS in one run.
Continue exploring with our guides on which chemical powder separate hydrogen from water and periodic table of elements with atomic number.
But — and this is a big but — quantification gets messy. Matrix effects vary wildly across chemical classes. Here's the thing — no isotopically labeled standards for most suspects. You're semi-quantifying at best, and you need to be honest about that in the report.
Total Oxidizable Precursor (TOP) Assay
This one's clever. Take your sample. All those sneaky precursors — fluorotelomer alcohols, sulfonamides, polyfluoroalkyl phosphates — convert to terminal PFAAs. In real terms, oxidize it with hydroxyl radicals (persulfate + heat). Still, the difference? Measure before and after. That's your precursor load.
Doesn't tell you which* precursors. Just that they're there. Still, it's the only way to close the mass balance on "unknown PFAS" in many matrices.
Combustion Ion Chromatography (CIC) and AOF
Adsorbable organic fluorine (AOF) by CIC gives you a sum parameter. Measure fluoride by IC. Trap the HF. Burn the sample. Subtract inorganic fluoride. What's left is organofluorine.
Useful for mass balance. Frustrating because it doesn't speciate. But when your targeted list explains only 15% of the AOF signal, you know you're missing something big.
Common Mistakes That Ruin Good Data
Treating All PFAS the Same
PFBA behaves nothing like PFOS. But different matrix effects. Different branched isomer patterns. That said, different ionization. Which means different extraction efficiency. If your method validation only covers the "big 18," you have no idea how your weirdo analytes perform.
Ignoring Branched Isomers
Linear standards are cheap. Technical mixtures (what actually exists in the environment) contain branched isomers. They elute earlier. They fragment differently. If you integrate only the linear peak, you underreport. Sometimes by 30-50%.
Forgetting Field Blanks
PFAS are everywhere. Your PTFE tubing. Your waterproof notebook. The sunscreen on your field tech's hands. The Teflon-lined cap on the sample bottle. If you don't collect field blanks, equipment blanks, and trip blanks, you can't prove your detects are real.
Assuming Labeled
Assuming labeled standards correct for matrix effects is a tempting shortcut, but isotopically labeled analogues only compensate for ionization suppression or enhancement that occurs after the analyte enters the mass spectrometer. Relying solely on a single surrogate standard can therefore mask significant biases, especially when the target PFAS suite includes compounds with widely varying chain lengths, functional groups, or polarity. Consider this: they do not account for losses during extraction, cleanup, or derivatization steps, nor do they correct for differential recovery of branched versus linear isomers. A more solid approach is to employ a suite of surrogate standards that span the chemical diversity of the analytes — short‑chain, long‑chain, sulfonates, carboxylates, and precursors — and to apply individual recovery corrections where possible. When surrogates are unavailable, matrix‑matched calibration or standard‑addition experiments become essential to gauge the true extent of matrix effects.
Other frequent missteps that erode data quality include:
- Single‑point calibration or inadequate curve range – PFAS often exhibit non‑linear response at the low‑ng L⁻¹ level due to surface adsorption or detector saturation. A minimum of five calibration points spanning the expected concentration range, with weighted (1/x or 1/x²) regression, is required to maintain accuracy across orders of magnitude.
- Neglecting analyte stability – Many PFAS adsorb to polypropylene, glass, or PTFE surfaces and can degrade (e.g., fluorotelomer alcohols oxidizing to PFAAs) during storage. Samples should be kept in polypropylene or HDPE containers, chilled, and analyzed within the validated holding time; stability studies specific to each matrix and analyte class are advisable.
- Overlooking carryover – High‑concentration samples can contaminate subsequent injections, especially for sticky, long‑chain PFAS. Implementing strong wash solvents (e.g., methanol‑water with 0.1 % ammonium hydroxide) and injecting blanks after high‑level samples helps detect and quantify carryover.
- Ignoring instrument drift – Long batches (≥ 24 h) can show sensitivity shifts of 10‑20 % for both targeted and HRMS modes. Regularly interleaving calibration verification standards (every 10‑15 injections) and applying drift correction factors prevents systematic bias.
- Using inappropriate extraction solvents or pH – Acidic or basic conditions can hydrolyze certain PFAS precursors (e.g., polyfluoroalkyl phosphates) or cause esterification, altering the profile. Method development should include pH screens and solvent strength tests to make sure the extraction recovers both target and precursor species without inducing artefacts.
- Failing to report uncertainty and detection limits – PFAS data are often presented as single values without confidence intervals. Propagating uncertainties from calibration, recovery, blank subtraction, and instrument noise provides a realistic picture of data quality and facilitates comparison across studies.
By recognizing these pitfalls and integrating corrective actions — surrogate‑standard recovery checks, matrix‑matched or standard‑addition calibration, rigorous blank schemes, stability verification, and drift monitoring — analysts can move from semi‑qualitative screening to defensible, quantitative PFAS measurements. The combination of targeted LC‑MS/MS for regulated compounds, HRMS for suspect screening, and complementary sum‑parameter approaches (TOP, AOF/CIC) offers a holistic view of both known and unknown fluorinated contamination. When all is said and done, transparent reporting of methodological limitations, uncertainty budgets, and quality‑control metrics ensures that PFAS data support reliable risk assessments, regulatory decisions, and scientific progress.
