Methods for the determination of phenolic brominated flame retardants, and by-products, formulation intermediates and decomposition products of brominated flame retardants in water

Brominated flame retardants (BFRs) are the chemicals of high importance within the REAch framework. In addition to polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD) and tetrabromobisphenol A (TBBPA), other BFRs such as bromophenols, intermediates in FR formulation like bromoanilines, and their brominated and non-brominated by-products such as bromoanisoles, bromotoluenes, bromoalkanes and 1,5,9-cyclododecatriene, respectively should be monitored and controlled because of their toxicity and their very low odour and taste thresholds, below sub-nanogram-per liter levels. In the present study several analytical methods for the simultaneous determination, i.e., combining one single sample treatment and one analysis step, of these compounds in water have been developed, optimized and evaluated. The methods involve a (pre-concentration)-extraction technique, such as liquid–liquid (LLE), solid-phase (SPE), headspace (HS) extraction or solid-phase microextraction (SPME), followed by gas chromatography (GC)–mass spectrometry (MS) analysis with either electron capture negative ionization (ECNI) or electron impact (EI) as ionization techniques. ECNI is more sensitive than EI for analytes with more than one bromine atom. HS and SPME were previously optimized by means of a multifactorial experimental design. Extraction temperature and the liquid/headspace volume ratio were the most significant factors in HS extraction. In SPME, the variables studied were the nature of the fiber, the mode of extraction and the extraction temperature. Polydimethylsiloxane (PDMS) fibers appeared to be more suitable than carboxen-polydimethylsiloxane (CAR-PDMS) for the analysis of the target compounds with more than one bromine atom. The extraction of 2,4-dibromoaniline was only achieved in a direct immersion mode, in which the optimal extraction temperature was 60 °C. The methods LLE–GC–(ECNI)MS, LLE–GC–(EI)MS, SPE–GC–(ECNI)MS, SPE–GC–(EI)MS, HS–GC–(EI)MS and SPME–GC–(EI)MS were evaluated in terms of linearity, precision, detection limits and trueness. All methods, with the exception of HS–GC–(EI)MS, were linear in a range of at least two orders of magnitude, giving recoveries above 75% and detection limits at the low ng/L level for most of the target analytes. SPE–GC–(ECNI)MS is the most sensitive and reliable method for the determination of most of the bromine compounds, whereas SPE–GC–(EI)MS is the most suitable to quantify the three isomers of 1,5,9-cyclododecatriene. Both methods together with SPME–GC–(EI)MS (for qualitative confirmation) were applied to water samples from the Western Scheldt (The Netherlands), where 2,6-dibromophenol and 2,4,6-tribromoanisole could be detected at levels higher than their respective odour thresholds.