Pre-concentration of chloride in dilute water-samples for precise δ37Cl determination using a strong ion-exchange resin: Application to rainwaters

Abstract

Low chloride concentration in water samples is a limitation for chlorine stable isotope measurements using a gas source isotope ratio mass spectrometer (IRMS). Here we describe an optimized method of chloride (Cl−) pre-concentration for dilute waters using a strong anion-exchange resin (Amberlite AG 1-X4). In order to validate our method and its reliability for δ37Cl measurements we made high precision δ37Cl measurements by dual inlet (DI) IRMS, also applicable to δ37Cl measurements by continuous flow IRMS.A 3ml solid phase extraction (SPE) cartridge is filled with 1ml of Amberlite resin in hydroxide form (100–200mesh) and is rinsed with 50ml of a 1M NaOH solution. The water sample containing aqueous chlorides is then loaded through the column at a flow rate of about 1.5ml/min. Because of the high partition coefficient of Cl− between resin and water (Kd>5300) the exchange of Cl− onto the resin is very efficient. We found that a volume as low as 5ml of a 1M KNO3 solution is enough for quantitative Cl− recovery from 1ml of resin. Using an internal laboratory standard solution as well as natural samples (seawaters and hydrothermal waters) with more complex chemistries and known δ37Cl values, the efficiency of this method was validated in terms of Cl− recovery yields, as well as accuracy and reproducibility of δ37Cl measurements.For dilute standard solutions with chloride concentrations ranging from 0.02 to 0.4mmol/l (Cl amounts ranging from 20 to 80μmol), the whole procedure recovery yields (ie. including the steps of Cl− pre-concentration and conversion into CH3Cl as well as consecutive CH3Cl purification) are 101±7% (1σ, n=21) and the measured δ37Cl agrees with the expected δ37Cl value to within ±0.04‰ (1σ), our current level of external precision reached by DI-IRMS measurements. Seawaters and hydrothermal waters with known δ37Cl values have also been diluted and treated using this method. All samples show quantitative Cl− recovery (yields~100%) and accurate δ37Cl values, with one maximum shift from the expected value of +0.09‰. We also found that for water samples containing relatively high concentrations of anions with a higher affinity for the resin than chlorides, such as sulfates or nitrates, the resin capacity may be exceeded, preventing a quantitative recovery of Cl− and yielding lower δ37Cl than expected. For such samples, we recommend increasing the volume of resin adequately. This method allows the investigation of as yet poorly documented δ37Cl in dilute natural waters (e.g., soil solutions, rainwaters and river waters) with highly precise DI-IRMS measurements. In a preliminary investigation, we applied our method to 6 rainwater samples collected at the summit and lowland of the Soufrière volcano. For Cl− concentrations between 0.06 and 0.1mmol/l we found small but significant δ37Cl variations, with values ranging from −0.16 to +0.38‰ (±0.04‰, 1σ). These results demonstrate that in active volcanic contexts, where rainwater Cl− may be partly sourced from volcanic degassing, δ37Cl carries information, which deserves further investigation given its potential implication for volcanic monitoring.