Analysis of chlordecone and its transformation products in environmental waters by a new SPME-GC-MS method and...

Among the numerous organochlorines (OCs) applied in the French West Indies (FWI), chlordecone (hydrated form C₁₀Cl₁₀O₂H₂; CLD) still causes major environmental pollution nowadays. A recent report revealed the unexpected presence in FWI environment of transformation products (TPs) of CLD not routinely monitored due to a lack of commercial standards. Here, we present a method for surface waters and groundwaters to analyze CLD, its main TPs (hydroCLDs, chlordecol (CLDOH), 10-monohydroCLDOH and polychloroindenes) and other OCs. We developed an SPME-GC-SIM/MS method with a PDMS-DVB fiber. Since CLDOH-d commonly used as internal standard (IS) proved unsuitable, we synthesized several IS candidates, and finally identified 10-monohydro-5-methyl-chlordecol as a satisfactory IS for CLDOH and 10-monohydroCLDOH avoiding the use of ¹³C-labelled analogue. LODs for CLD and its TPs varied from 0.3 to 10 ng/L, equal to or below LODs of the two laboratories, BRGM (the French geological survey) and LDA26 (one of the French Departmental Analytical Laboratories), requested in FWI pollution monitoring that used liquid-liquid extractions and advanced facilities (LLE-GC-MS/MS and LLE-LC-MS/MS methods, respectively). Then, we extended the multi-residue method to 30 OCs (CLD and its TPs, mirex, β-HCH, lindane, dieldrin, aldrin, HCB, hexachlorobutadiene, TCE, PCE) and applied it to 30 surface and ground waters from FWI. While CLD, 8- and 10-monohydroCLD, CLDOH, 10-monohydroCLDOH, dieldrin, and β-HCH were detected and quantified, pentachloroindene, another CLD TP, was sporadically found in trace levels. A comparison with BRGM and LDA26 confirmed the interest of the SPME method. Results suggested an underestimation of CLDOH and an overestimation of high CLD concentrations with one of the currently used routine protocol. In light of these findings, previous temporal monitoring of environmental waters in FWI were re-examined and revealed some atypical values, which may indeed be due to analytical bias. These discrepancies call for intensified efforts to reliably quantify CLD and its TPs.