Chlorinated paraffin pollutants push liquid chromatography mass spectrometry to its limits

Polychlorinated alkanes, or PCAs, are the principal ingredients of technical mixtures known as chlorinated paraffins, and they may be the most analytically stubborn pollutants that environmental chemists routinely try to measure. These compounds serve as flame retardants, plasticizers and metalworking lubricants, and they follow the general formula CxH2x+2-yCly, with chlorine content typically between 30 and 70 percent by weight. Conventionally they are split by carbon chain length into short-chain PCAs of C10 to C13, medium-chain PCAs of C14 to C17, and long-chain PCAs of C18 to C30. They are persistent, chemically resistant, widely detected in environmental and food matrices, and they bioaccumulate and biomagnify through the food chain. Studies suggest long-term exposure may pose health risks including liver and kidney toxicity, endocrine disruption, neurotoxicity, and developmental and reproductive effects, and the International Agency for Research on Cancer has classified chlorinated paraffins with an average chain length of C12 and roughly 60 percent chlorination as possibly carcinogenic to humans.

Regulation has now begun to reshape the entire field. In 2017, short-chain PCAs were listed as persistent organic pollutants under the Stockholm Convention, and in 2025 the medium-chain PCAs of C14 to C17 were also added to Annex A, though with multiple exceptions. Longer-chain PCAs share several properties of concern with their shorter-chain counterparts, including bioaccumulation and biomagnification, but scientific evidence on the adverse health effects of very long chains, C18 and above, remains limited, leaving the case for further restrictions undecided. Analysis of more than 30 technical mixtures produced between the 1970s and 2022 has shown that these products are highly diverse and that regulatory measures have already driven a gradual shift from shorter to longer chains. That shift has created an analytical crisis of its own: longer-chain congeners are less volatile and cannot be characterized efficiently by conventional gas chromatography-based methods, so demand for liquid chromatography-mass spectrometry, or LC-MS, has increased substantially in recent years. A new review now provides a comprehensive overview of these LC-MS approaches, evaluating separation, ionization and detection strategies and the methodological trade-offs they entail.

The core difficulty is sheer structural complexity. Depending on chain length and chlorine content, PCAs can have hundreds to thousands of isomers per homologue group. For the short-chain PCAs alone, under the restrictive assumptions of one chlorine atom per carbon and a chlorination degree limited to five to ten chlorine atoms per molecule, over 6000 different isomers have been calculated, and the true number is expected to be higher because more than one chlorine per carbon is possible. This extreme complexity currently makes complete separation of all homologue groups impossible. Instead, recent studies have explored multiple stationary and mobile phase combinations to improve separation as far as LC allows. Octadecylsilane, or C18, columns have commonly been selected because of PCA hydrophobicity, but a broader range of stationary phase chemistries is now under evaluation, with varying degrees of success.

Alternative phases have delivered mixed results. A pentafluorophenyl, or PFP, stationary phase, which is said to be more efficient at separating halogenated compounds than conventional aliphatic phases due to enhanced dipole and pi-pi interactions, did not provide significant improvement in PCA separation compared with conventional C18 in studies of wine and other matrices. A Phenyl-Hexyl column, offering hydrophobic, pi-pi and possible hydrogen-bonding interactions, successfully analyzed the C14 to C17 homologues in municipal waste incinerator samples, although separation from other homologues that could cause interference was not investigated. The most striking result came from a comparison of C18 and cyanopropylsilane, or CN, columns for short-chain PCAs: on the C18 column, the resolution between the C10Cl6 and C13Cl6 homologues was slightly below 1, with broad peaks 2.6 to 2.8 minutes wide, whereas on the CN column the same homologues were baseline separated with a resolution close to 2 and sharp peaks of only 0.8 to 0.9 minutes. CN phases combine a lipophilic propyl spacer for hydrophobic interactions with cyanopropyl groups that introduce polar selectivity, and the multiple chlorine substituents of PCAs create localized polar regions and molecular dipole moments that may enhance retention on such phases. A comprehensive comparison of six stationary phases found that retention generally increased with chain length, and that for short-chain PCAs retention also increased with chlorination degree, but higher-chain PCAs showed non-monotonic behavior on C18 and C30 phases, with low-chlorinated homologues eluting after medium-chlorinated ones, possibly because changes in molecular shape lower the accessible hydrophobic surface area.

Even the chlorine substitution pattern matters. Experimental and modelling work shows that vicinal versus geminal chlorination and the length of unsubstituted methylene segments can shift hydrophobicity, and congeners with longer uninterrupted methylene segments and more clustered chlorination appear more hydrophobic. In a study of different C10 to C13Cl6 congeners on a C18 column, congeners with geminal chlorine atoms showed higher retention times than congeners with a more even distribution. This structural spread is consistent with the broad, hump-like signals typically observed for technical PCA mixtures, where many isomers share the same chain length and chlorination degree but elute across a wider retention window. Mobile phase composition and gradient design are equally decisive. Most methods use aqueous acetonitrile or aqueous methanol, and systematic comparison showed that methanol increased absolute PCA responses by approximately 1.5-fold compared with acetonitrile, while acetonitrile favored formation of chloride adducts and thereby simplified spectra, with deprotonated ions contributing no more than about 10 percent of total signal under acetonitrile versus about 30 percent with methanol. Gradients varied from fast UPLC programs achieving a total run time of about 6 minutes on Orbitrap MS to more gradual 25-minute programs on tandem MS instruments. Steeper gradients were associated with greater co-elution, whereas gradual organic-phase increases improved homologue separation at the cost of sensitivity, as the analyte signal spreads over broader chromatographic humps.

Ionization strategy is the other central design choice. Because PCAs do not ionize efficiently in either electrospray ionization, ESI, or atmospheric-pressure chemical ionization, APCI, without suitable additives, LC-MS methods commonly use mobile-phase or post-column additives to promote adduct formation. The dominant approach uses chlorine-containing additives such as ammonium chloride, typically added at low concentrations around 25 micromolar, or dichloromethane, which can be introduced directly into the mobile phase or post-column, to promote chloride adducts, [M + Cl]−. Comparisons have shown that ESI consistently produced higher responses than APCI regardless of additive, and that deprotonated ions of PCAs result from the dissociation of anionic adducts, which then lose HCl. Recently introduced additives have delivered meaningful gains: tetraphenylphosphonium chloride at 10 micromolar yielded instrumental detection limits of 0.01 to 0.03 picograms per microliter for short-chain PCAs, while tetramethylammonium chloride at 50 micromolar dissociated directly into ions in the spray process, providing several-fold higher responses and cleaner spectra at much lower additive levels than dichloromethane. Acetate additives promote weaker acetate adducts but produce more characteristic fragmentation, first losing acetic acid and then HCl, which makes them valuable for LC-MS/MS workflows. Bromide adduct formation using bromoform has also been explored, achieving near-exclusive formation of stable [M + Br]− ions, though the higher toxicity of bromoform raises concerns.

Quantification faces a further obstacle: instrumental response is not constant across PCA homologue groups but depends strongly on carbon chain length and chlorination degree. Across four studies using the same commercial standards but different LC-HRMS configurations, the total response of short-chain PCAs increased approximately three- to fourfold between standards containing 51.5 and 63 percent chlorine, and the response of medium-chain PCAs increased approximately three- to fivefold between 42 and 57 percent chlorine standards, with one study reporting a 40-fold lower response for the less chlorinated mixture. Direct liquid injection APCI-ToF-MS at lower flow rates observed an approximately 50-fold difference in response between the least and most chlorinated short-chain mixtures, containing 49 and 70 percent chlorine respectively, yet the increase does not continue indefinitely: a long-chain mixture at 70 percent chlorine produced a response approximately 30 percent lower than the corresponding 49 percent mixture, consistent with homologue-level observations that response increases up to roughly ten chlorine atoms and then decreases. Response also depends on congener structure, with stronger responses for congeners with chlorine distributed along the chain and poorer responses for low chlorination and extensive terminal substitution.

Mass spectra of PCAs are intrinsically dense and interference-prone. While typical organic molecules produce only two to three isotopologues above 10 percent relative abundance, PCAs yield several because both stable chlorine isotopes, 35Cl and 37Cl, are sufficiently abundant: the chloride adduct of C10H18Cl4 produces five isotopologues above this threshold, and C17H28Cl8 produces nine. Co-formation of deprotonated ions, off-target adducts from residual additives such as formate adducts traced to residual formic acid in the LC system, and transformation products all compound the problem. Chlorinated olefins and diolefins, known as CP-enes, can form during ionization and interfere with spectra, with chloride enhancement apparently increasing olefin signals because deprotonated PCA ions overlap with chloride adducts of the olefin products. However, CP-enes are also present in technical mixtures from the manufacturing process itself, and mono- and diolefinic PCAs remain detectable even after acidified silica cleanup. Vinyl chloride oligomers, often present in PVC items, fall within the compositional space of PCAs and can contribute to measured signals, though they can be distinguished by their limited polymer-repeat series and narrower chromatographic peaks compared with the broad, unresolved humps of technical PCA mixtures.

Because complete chromatographic separation is not feasible, high resolving power becomes critical. Orbitrap MS at a resolution of 140,000 was sufficient to resolve most key overlaps, including separation of chloride-adduct signals from interfering deprotonated ions across much of the relevant mass range, whereas the mass difference between a chloride adduct and the deprotonated ion of a homologue with one additional chlorine is only 0.019, achievable only at the highest resolution settings. At lower resolutions of 10,000 or 35,000, such interference can shift apparent peak masses and cause peaks to disappear during data processing if they drift outside applied mass tolerance windows, requiring wider tolerances and deconvolution procedures. Direct-injection methods, which skip chromatographic separation entirely, offer faster analysis and greater flexibility but introduce the full extent of unresolved spectral complexity, making them convincing only when coupled to ultra-high-resolution instruments or robust deconvolution. LC-MS/MS offers an alternative for narrowly defined PCA ranges, but the number of multiple reaction monitoring transitions required is a practical limitation: even the relatively narrow scope of short-chain PCAs with five to eight chlorine atoms requires 32 transitions, while expanding to C10 through C20 with four to ten chlorine atoms would exceed 100, lengthening cycle times and reducing data points across each peak. The review concludes that no universal LC-MS method for PCAs yet exists; the most suitable approach depends on the analytical purpose, the PCA range of interest, the required selectivity, and whether broad profiling or targeted quantification is the primary goal.

Subject of Research: Analytical chemistry

Article Title: Liquid chromatography-mass spectrometry approaches for analysis of polychlorinated alkanes: a review

Article References: Lazdiņa, L., & Pērkons, I. (2026). Liquid chromatography-mass spectrometry approaches for analysis of polychlorinated alkanes: a review. Environmental Science and Pollution Research. https://doi.org/10.1007/s11356-026-38341-3

Image Credits: AI Generated

DOI: 10.1007/s11356-026-38341-3

Keywords: polychlorinated alkanes, chlorinated paraffins, liquid chromatography, mass spectrometry, persistent organic pollutants, high-resolution mass spectrometry, environmental analysis, Stockholm Convention, Liquid, chromatography-mass, spectrometry, approaches