Calibration and deployment of POCIS and Chemcatcher for the monitoring of explosives on lake-bottom sediments
Between 1920 and 1967, approximatively 8000 tons of ammunition waste was dumped into some Swiss lakes. Previous studies based on grab sampling revealed small traces of cyclotetramethylenetetranitramine (HMX), cyclotrimethylenetrinitramine (RDX) and pentaerythritoltetranitrate (PETN) in the water column of two studied lakes; however these researches showed that the ammunition waste at the lake bottom was likely not involved in that contamination. On the contrary, external sources (tributaries contaminated by hot spots such as ammunition destruction locations or ammunition factories) might be the main cause. The ultimate goal of this study is to assess whether explosive monitoring by passive sampling is technically feasible in the affected lakes and rivers, and whether integrative sampling confirm former results obtained by grab sampling. As a first step, POCIS (HLB powder sandwiched by two PES membranes) and Chemcatcher (SDB-RPS disk covered by a PES membrane) were calibrated in a channel system supplied with continuously refreshed lake water spiked with two nitramines (HMX, RDX), one nitrate ester (PETN) and seven nitroaromatics (2,4,6-trinitrotoluene (TNT) and three of its degradation products, 1,3-dinitrobenzene, 2,4- and 2,6-dinitrotoluene). Exposure parameters in the system were kept as close as possible to the ones expected in the first meters above sediments of the studied lakes. Sixteen POCIS and Chemcatcher were simultaneously deployed and removed in duplicates at 8 different intervals over 21 days whereas water analysis (SPE extraction) was carried out daily. PES membranes (pore size of 0.1μm) and sorbent phases were separately extracted and analyzed by UPLC-MS/MS. LDPE strips spiked with PRCs were also deployed in the channel system to obtain dissipation data in the calibration conditions. Results of the calibration showed that the accumulation of HMX and RDX (log Kow <1) in sorbent materials was higher or similar to the accumulation in PES membranes. The transfer of these two compounds from water to sorbents was quick (lag-phase < 1day). For PETN (log Kow about 2), the accumulation was slightly higher in PES membranes than in sorbent materials, leading to lag-phase of about 3 (POCIS) to 5 (Chemcatcher) days. The affinity of nitro aromatic compounds for PES membranes was shown to be very high, strongly delaying the transfer of these compounds to sorbents (lag-phase up to 9 days). Sampling rate (Rs) of HMX and RDX were in the range of 0.1 L/d with POCIS and 0.05 L/d with Chemcatcher. Because of the long lag-phases, Rs of nitroaromatics were up to 10 times lower, which can potentially be an issue in terms of sensitivity (expected concentrations in the lakes in the sub-ng/L range). As a second step, the samplers were deployed on lake-bottom sediments (200m depth) using a Remotely Operated Vehicle (ROV) to unhook and re-hook the rope that was used to descend and retrieve the sampler holders. The holders were equipped with a temperature probe. In addition of estimating water current velocity by measuring particle velocity with the ROV camera, LDPE strips spiked with PRCs were co-deployed to compare the dissipation data with the ones from the calibration experiment. Extraction and analysis are currently being carried out and will be discussed along with the first tests done in rivers.