ARTICLE pubs.acs.org/est
Method for the in Situ Calibration of a Passive Phosphate Sampler in Estuarine and Marine Waters Dominique S. O’Brien,*,† Kees Booij,‡ Darryl W. Hawker,§ and Jochen F. Mueller† †
The University of Queensland, National Research Centre for Environmental Toxicology (Entox), 39 Kessels Road, Coopers Plains, Queensland 4108, Australia ‡ Royal Netherlands Institute for Sea Research, P.O. Box 59, 1790 AB Texel, The Netherlands § School of Environment, Griffith University, Nathan, Queensland 4111, Australia
bS Supporting Information ABSTRACT: Passive samplers for phosphate were calibrated in the laboratory over a range of flow velocities (0-27 cm s-1) and ionic strengths (0-0.62 mol kg-1). The observed sampling rates were between 0.006 and 0.20 L d-1. An empirical model allowed the estimation of these sampling rates with a precision of 8.5%. Passive flow monitors (PFMs), based on gypsum dissolution rates, were calibrated for the same range of flow velocities and ionic strength. Mass loss rates of the PFMs increased with increasing ionic strength. We demonstrate that this increase is quantitatively accounted for by the increased gypsum solubility at higher ionic strengths. We provide a calculation scheme for these solubilities for an environmentally relevant range of temperatures and salinities. The results imply that co-deployed PFMs can be used for estimating the flow effect on the in situ sampling rates of the phosphate samplers, and we expect that the same may hold for other passive samplers.
’ INTRODUCTION Time integrated passive samplers have become important tools for the cost-effective monitoring of organic pollutants1-7 and metals.8-10 The use of ferrihydrite, a ferric oxyhydroxide mineral, as a sequestering phase has facilitated the development of a number of passive sampling methods for the measurement of phosphate concentration in water.11,12 The method employed by Zhang et al.12 employs ferrihydrite within a diffusive gradient in thin-film (DGT) gel. The P-trap developed by M€uller et al.11 contains a ferrihydrite suspension behind a commercially available filter membrane. Alternatively, commercially available resins have been used in the construction of the SorbiCell sampler13 which has been employed for the assessment of nitrate and phosphate concentrations in fresh waters.14 Phosphate concentrations in environmental waters (Cw) can be determined from the mass (Ms) of phosphate accumulated within the deployed passive sampler, assuming that the sampler is operating in the linear or integrative mode Cw ¼
Ms Rs t
ð1Þ
where t is the deployment time, and Rs is the sampling rate. Values of Rs may depend on exposure conditions such as flow rate, ionic strength, and temperature. When using a DGT sampler, the thickness of the gel diffusion barrier is designed to be sufficiently large so that any change in the external water boundary layer (WBL) will not influence the overall flux to the sampler. Conversely, with the P-trap, the ambient flow and temperature will influence the flux of molecules into the sampler as the thickness of the WBL is changed. The work undertaken by M€uller et al.11 showed that phosphate concentrations obtained from grab sampling and from exposed P-traps were similar. These authors concluded that flow had relatively r 2011 American Chemical Society
little influence on the performance of the P-trap. However, calibration experiments undertaken by O’Brien et al.15 using a P-sampler, a passive sampler of configuration similar to that of the P-trap showed that Rs (L d-1) did increase with increasing water flow velocity. Calibration of passive samplers may be achieved by measuring sampling rates in the laboratory for the range of exposure conditions that are encountered in the field (flow, temperature, salinity). However, the determination of the parameters that are needed to calculate in situ sampling rates may not always be possible. The cost of deploying current meters and data loggers at multiple sites may be prohibitive, and the same holds for the manual recording of these parameters during multiple visits to the exposure sites. Alternatively, in situ sampling rates have been calculated from the dissipation rates of performance reference compounds (PRCs), that are spiked into the sampler before exposure.16,17 The PRC method is successful for samplers in which partitioning is the predominant sorption mechanism (e.g., semipermeable membrane devices, nonpolar Chemcatcher).7,18 By contrast, the PRC method generally fails for adsorption based passive samplers.19,20 An example of such a sampler is the P-trap, for which it is proposed that complexes of nonprotonated bidentate binuclear species of tFe2PO4 form at the iron-water interface.21 O’Brien et al.15 introduced the passive flow monitor (PFM) to assess the ambient flow at a deployment site. The PFM devices are constructed using gypsum or calcium sulfate dihydrate [CaSO4 3 2H2O], from which mass is lost through a constant exposed surface area. The basis of the approach is that dissolution rates for the PFM devices are limited by transport through the WBL.22 Received: May 13, 2010 Accepted: January 13, 2011 Revised: November 10, 2010 Published: March 03, 2011 2871
dx.doi.org/10.1021/es101645z | Environ. Sci. Technol. 2011, 45, 2871–2877
Environmental Science & Technology It has been shown for fresh water that flow rates can be calculated from the mass loss rates of the PFM (rPFM, g d-1) and that these flow rates in turn can be used to calculate the sampling rates for the phosphate samplers.15 It can be expected that these relationships are different for estuarine and marine waters as the solubility of CaSO4 3 2H2O increases with increasing ionic strength of the solution.23-25 These studies show that at 298 K, the molal solubility of CaSO4 3 2H2O more than doubles in solutions with NaCl concentrations typical of seawater compared to solutions with an initial ionic strength of zero. Additionally, the higher background concentrations of calcium and sulfate in marine and estuarine waters cause the gypsum solubility to decrease. Therefore, the use of PFM devices in estuarine or marine systems requires further investigation. Ionic strength has also been shown to influence both the retention of heavy metals within ferric iron oxyhydroxide (ferrihydrite) nanoparticles26 and the diffusion of molecules within thin film gels.9,27,28 A decrease in the uptake and retention of heavy metals by nanoparticulate ferrihydrite occurred following aggregation of the nanoparticles when exposed to media with an increased ionic strength or pH.26 This suggests that changes in the structure and binding characteristics of the receiving phase may also contribute to the changes in diffusion behavior. While the salinity (and therefore ionic strength) varies between marine and estuarine water, there can also be considerable variation with inland and surface waters. In Australia, the salinity of such water can vary from
