32 PFAS in Baby Food by LC-MS/MS: Dr. Ehrenstorfer Reference Standards in a Shimadzu Application

Why is PFAS analysis in food matrices difficult?

In this study, several PFAS were quantified at an LOQ as low as 0.01 µg/kg — while the analytical system itself is one of the contamination sources Shimadzu points out.

Shimadzu states that one of the main challenges is reducing background contamination from PFAS residues inside the LC system itself, as well as during sampling and sample preparation. PFAS are present in many common laboratory materials, so cross-contamination comes not from the sample but from the labware and the flow path.

The second challenge is the matrix. Shimadzu addressed it by building matrix-matched calibration curves and back-calculating spiked-sample concentrations against those curves, rather than using clean solvent.

The third is the concentration level. The lowest limit of quantitation in the study is 0.01 µg/kg, applied to 27 of the 32 analytes. Four analytes use 0.05 µg/kg and PFBA uses 0.25 µg/kg.

How did Shimadzu build the method?

This section summarises Application News 06-SAIP-LC-077-EN, published openly by Shimadzu on shimadzu.com, first edition October 2024.

The whole workflow follows a short path:

1. Weighing. 5.0 g of infant formula.
2. Hydration and extraction. Add 15 mL water, vortex 2 minutes; add 5.0 mL acetonitrile, shake and vortex 2 minutes.
3. QuEChERS partitioning. Add 4.0 g MgSO₄ and 1.0 g sodium acetate, vortex 2 minutes, centrifuge 10 minutes at 7,800 rpm.
4. dSPE clean-up. Collect the supernatant, add dSPE sorbent, vortex 1 minute, centrifuge 5 minutes at 7,800 rpm.
5. Concentration and reconstitution. Take 2.0 mL of the supernatant, dry under nitrogen, reconstitute to 1.0 mL with MeOH:H₂O 80:20 (v/v).
6. Analysis. LC-MS/MS in MRM mode on an LCMS-8060NX coupled to a Nexera X3.

Calibration was external with 1/C² weighted regression over 0.005–10 µg/L; three compounds used 0.05–10 µg/L and PFBA used 0.25–10 µg/L. Spiked samples were prepared in six replicates at three levels: 0.01 · 0.05 · 0.25 µg/kg.

Instrument configuration

32 PFAS quantified simultaneously

The list spans five structural classes:

Retention times range from 3.1 to 15.7 minutes; Shimadzu separated every peak within twenty minutes.

The reference standard that was used

In the Material and Method section, Shimadzu writes:

Each lot is supplied with a Reference Material Certificate stating the certified concentration and expanded uncertainty of every component, individual lot numbers per component, purity, reference retention times and expiry date. Certified values are established gravimetrically and volumetrically, with confirmation by analytical technique.

What does the method data tell us?

Validation followed the SANTE guidance. Below are the ranges Shimadzu reports across all 32 analytes:

A closer look at a few numbers

In the dataset Shimadzu published, N-Me-FOSAA (63.07 %) and N-Et-FOSAA (70.50 %) show the lowest recoveries. The document describes “most of the compounds” as falling within 65–135 %, which means some fall outside that window.

The lowest R² belongs to PFBS (0.9673); the remaining 31 analytes are all above 0.98. PFBA was assigned a higher LOQ than the rest of the analytes in this method (0.25 µg/kg versus 0.01 µg/kg).

In other words, most analytes in the list meet the criteria with a wide margin, while a small group sits close to the threshold. In routine use, that group is the one to monitor separately in quality control.

Peak data for six representative analytes

One detail worth attention: for PFBA and PFPeA, the document leaves the qualifier transition column empty. A laboratory adopting the method should clarify the identification criteria for these two analytes against the regulation it works under, since many confirmation guidelines require two transitions.

Calibration equations

The document publishes calibration curves for six analytes at the LOQ level:

The slopes differ by more than a factor of four between PFNA (1714) and 10:2 FTS (320). Signal response varies considerably from analyte to analyte, so each one needs its own calibration curve — exactly how Shimadzu built the calibration in this study.

How should you choose standards when developing a PFAS method?

Ordering the right catalogue number is the last step. Seven questions come before it, in order.

A concrete example from this very study

Mixture DRE-A50000151MW certifies 24 components, while the method quantifies 32 analytes. Eight analytes fall outside the mixture and must be sourced from other standards:

This is the easiest thing to overlook. A mixture named after a reference method does not necessarily cover the full analyte list your laboratory needs. Build the component-by-component comparison table before ordering, rather than discovering the gap after the first batch has been run.

Reference documents