Environ. Sci. Technol. 2010, 44, 9419–9424
Titanium Dioxide-Based DGT Technique for In Situ Measurement of Dissolved Reactive Phosphorus in Fresh and Marine Waters JARED G. PANTHER, PETER R. TEASDALE,* WILLIAM W. BENNETT, DAVID T. WELSH, AND HUIJUN ZHAO Environmental Futures Centre, Griffith University, Gold Coast campus, QLD 4222, Australia
Received August 13, 2010. Revised manuscript received October 25, 2010. Accepted November 13, 2010.
A new diffusive gradients in a thin film (DGT) technique for measuring dissolved reactive phosphorus (DRP) in fresh and marine waters is reported. The new method, which uses a commercially available titanium dioxide based adsorbent (Metsorb), was evaluated and compared to the well-established ferrihydrite-DGT method (ferrihydrite cast within the polyacrylamide gel). DGT time-series experiments showed that the mass of DRP accumulated by Metsorb and ferrihydrite was linear with time when deployed in simple solutions. Both DGT methods showed predictable uptake across the pH (4.0-8.3) and ionic strength (0.0001-1 mol L-1 NaNO3) ranges investigated, and the total capacity of the Metsorb binding phase (∼40 000 ng P) was 2.5-5 times higher than the reported total capacity of the ferrihydrite binding phase. The measurement of DRP in synthetic freshwater and synthetic seawater by Metsorb-DGT over a 4 day deployment period showed excellent agreement with the concentration of DRP measured directly in solution, whereas the ferrihydrite-DGT method significantly underestimated (23-30%) the DRP concentration in synthetic seawater for deployment times of two days or more. Field deployments of Metsorb-DGT samplers with various diffusive layer thicknesses allowed accurate measurement of both the diffusive boundary layer thickness and DRP concentration in situ. The MetsorbDGT method performs similarly to ferrihydrite-DGT for freshwater measurements but is shown to be more accurate than the ferrihydrite method for determining DRP in seawater.
Introduction In natural waters, phosphorus (P) exists as both inorganic and organic species distributed between the dissolved, colloidal and particulate phases (1, 2). The principle form of P in many aquatic environments is the mono- and diprotonated orthophosphate species (HPO42-, H2PO4-), which are considered to be the most readily bioavailable forms of P and make up a significant portion of the commonly measured dissolved reactive phosphorus (DRP) fraction (also called soluble or filterable reactive phosphorus) (1, 2). The measurement of DRP in surface water and wastewater is critical for understanding eutrophication processes and P biogeochemical cycling, assessing water quality in drinking * Corresponding author phone: +61 7555 28358; fax: +61 7555 28067; e-mail:
[email protected]. 10.1021/es1027713
2010 American Chemical Society
Published on Web 11/23/2010
water reservoirs, monitoring wastewater discharge compliance and implementing and maintaining nutrient management strategies for environmental waters (1-3). There are a variety of methods and techniques available to measure the concentration of DRP in natural waters (2). However, many of the routinely used methods for DRP monitoring are laboratory-based and require a sample to be collected, in some cases filtered or preserved, transported and stored prior to analysis. Some of the preservation and storage protocols reported for DRP are time-consuming, inadequate for maintaining the original sample composition, show contradictory results and in some cases use preservation chemicals that can interfere with the analysis (1-5). Furthermore, processes such as precipitation, adsorption, hydrolysis, and microbial uptake and release during storage, can mean that the DRP concentrations at the time of analysis may not be representative of those within the natural water at the time of sampling (1, 5). Therefore, it is desirable to utilize in situ sampling methods, which avoid the majority of these issues, to accurately measure concentrations of DRP in natural waters. The diffusive gradients in a thin film (DGT) technique is a well-established in situ, time-integrated, passive sampling method that is designed to accumulate dissolved labile species in aquatic environments (6-9). Analyte species diffuse through a polyacrylamide hydrogel (diffusive layer) of known thickness and are immobilized by an adsorbent (8). After deployment (typically 1-5 days), the accumulated analyte is eluted from the adsorbent and the mass of analyte in the eluent is determined using sensitive instrumental methods. The DGT-measured concentration is calculated using the DGT equation (see Experimental Section) which is derived from Fick’s First Law of Diffusion (7, 8). The first DRP DGT application, reported by Zhang and co-workers in 1998 (10), used an iron hydroxide (ferrihydrite) adsorbent. This adsorbent, when incorporated within DGT samplers, was shown to accumulate DRP in a predictable manner from both simple laboratory solutions and natural freshwaters (10-12), and has also been used to provide sediment porewater profiles for DRP (13, 14), estimate available P in soils (15-17) and monitor DRP aquaculture emissions (18). Ferrihydrite has also been used as a DGT adsorbent to accumulate anionic species of As, V, Se, and Mo (11, 14, 19-22). Although there has been a substantial body of work undertaken to evaluate the ferrihydrite-DGT method for measuring DRP in natural waters (10-12, 15), a detailed study of its value as a monitoring tool for DRP in marine and estuarine waters is yet to be reported. Recently, a new titanium dioxide based adsorbent (Metsorb, http://www.gravertech.com) was identified by Bennett and co-workers as an alternative binding phase for the DGT measurement of inorganic arsenic and selenium in water (23). The Metsorb-DGT technique measured AsIII, AsV, and SeIV quantitatively over the pH and ionic strength range encountered in most natural waters and was also used to measure total inorganic arsenic in situ. The present study focuses on development of a new DGT method, based on the Metsorb binding layer, for the measurement of DRP in fresh and marine waters. A series of laboratory and field experiments were carried out to validate the new MetsorbDGT method and compare it to the well-established ferrihydrite method.
Experimental Section Reagents, Materials, and Solutions. Deionized water (Milli-Q Element) was used to prepare all solutions. Details of reagents VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9419
for solution preparation and procedures for preparation of P stock solutions, polyacrylamide diffusive gels and Metsorb and ferrihydrite binding gels (ferrihydrite was mixed with the gel solution prior to casting), are given in the Supporting Information (SI). All Metsorb, ferrihydrite, and diffusive gels were prepared in house. Plastic containers used for experimental work and for the preparation and storage of solutions and all DGT components (including materials used to prepare DGT gels) were acid-cleaned in 10% (v/v) HNO3 (AR grade, Merck) for at least 24 h and rinsed thoroughly with deionized water prior to use. All salts used to prepare solutions were AR grade or higher. Sample Analysis. P measurements were performed spectrophotometrically by a discrete nutrient analyzer (EasyChem, Systea Scientific Analytical Technologies) utilizing the standard molybdenum-blue method. Analytical standards were prepared in the concentration range 0-250 µg L-1 P (prepared in 2% v/v HCl (AR grade, Merck)) from a 1 mg L-1 P working standard. Natural water certified reference materials (Queensland Health Scientific Services) were used for quality assurance purposes; typical recoveries are given in the SI. For all measurements, the matrix of the standards was matched to the samples. DGT Assembly and Deployment. Solution DGT sampling devices were supplied by DGT Research Ltd. and were assembled as described previously (8). A 0.45 µm cellulose nitrate membrane filter (Millipore) of 0.010 cm thickness was used. The combined thickness of the diffusive gel and membrane filter was used for all DGT calculations. Cleaning procedure for the membrane filters is given in the SI. DGT samplers were assembled in a Class 100 laminar flow cabinet within a Class 1000 clean room. Preparation of experimental P solutions, deployment of DGT samplers, and experimental procedures are given in the SI. Gel Elution. The Metsorb and ferrihydrite binding gels were eluted in 1.5 mL of 1 mol L-1 NaOH for 24 h; an elution efficiency of 92 ( 5% was used for both adsorbents. For synthetic seawater experiments and both field deployments, the binding gels were rinsed in 5 mL of deionized water for 1 h (to remove unbound salts) prior to elution with either 1 or 1.5 mL of 1 mol L-1 NaOH (elution efficiencies were unaffected by eluent volume). This modified elution procedure resulted in slightly lower average elution efficiencies (85 ( 5%), although both agree within experimental uncertainty. The small decrease in elution efficiency may be due to removal of weakly bound P, or loss of adsorbent containing a small amount of P, from the binding gel during the washing step. Failure to include the washing step resulted in significantly lower elution efficiencies (see SI). The DGT-measured concentration (CDGT) of DRP in solution was calculated using eq 1 (8): CDGT )
M∆g DtA
(1)
where M is the mass of P bound to the adsorbent (ng), ∆g is the diffusive layer thickness (cm), D is the diffusion coefficient of DRP through the diffusive layer (cm2 s-1), t is the deployment time (s), and A is the area of the sampling window exposed to solution (cm2). The reported diffusion coefficient of DRP (6.05 × 10-6 cm2 s-1 at 25 °C) (10) was corrected for temperature using the Stokes-Einstein equation (8), and for seawater and high ionic strength (1 mol L-1 NaNO3) deployments a value of 0.9 × D was used (6, 24). Laboratory Evaluation. Uptake and elution efficiencies were measured by immersing Metsorb and ferrihydrite gel discs (4.91 cm2) in 5 mL solutions (prepared in 0.01 mol L-1 NaNO3 at pH 6.5) ranging in concentration from 20 to 300 µg P L-1; this was carried out in triplicate for each binding gel at each P concentration. After 24 h the gels were removed, 9420
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010
eluted and the uptake solutions were analyzed (after acidification) to determine the mass of P remaining in solution. Binding gels were eluted for 24 h in 1 mL of either 1 mol L-1 HCl, 1 mol L-1 H2SO4, or 1.5 mL of 1 mol L-1 NaOH. Average elution efficiencies are reported and the uncertainty is (1 standard deviation of the mean over the mass range studied. The accumulation of DRP over time by Metsorb-DGT and ferrihydrite-DGT was evaluated by deploying DGT samplers in a 100 µg L-1 P solution for times ranging from 5 to 55 h. The effect of pH and ionic strength on the accumulation of DRP by both the Metsorb- and ferrihydrite-DGT methods was investigated by deploying DGT samplers in P solutions (50-100 µg L-1) of varying pH (4.0-8.3; 0.01 mol L-1 NaNO3) and ionic strengths (0.0001-1 mol L-1 NaNO3; pH 6.50 ( 0.15), for 8-10 h. The P capacity of the Metsorb binding gel was estimated by deploying five Metsorb-DGT samplers in a 25 mg L-1 P solution for 24 h (pH 7.35 ( 0.05; 0.01 mol L-1 NaNO3). DGT samplers containing Metsorb or ferrihydrite were deployed in synthetic freshwater and synthetic seawater for four days to evaluate the performance of the binding gels. The composition of the synthetic freshwater was similar to that described by Langmuir (25), and the synthetic seawater was prepared according to Grasshoff et al. (26). Composition of both waters is detailed in the SI. Both the synthetic freshwater and synthetic seawater were spiked with 60 µg L-1 P (this P concentration is within the range found in many natural waters of environmental interest refs 2, 18, 27-29). Field Evaluation. DGT samplers were deployed at two field sites. The first site was a small, urban freshwater stream and the second site was a large marina located on the Gold Coast Broadwater, a coastal lagoon. At each site, 12 Metsorb DGT samplers containing varying diffusive layer thicknesses (∆g ) 0.01 cm (membrane filter only), 0.05 cm, 0.09 cm, 0.13 cm; all deployed in triplicate) and three ferrihydrite-DGT samplers (∆g ) 0.09 cm), were deployed 30 cm beneath the water surface for four days. Temperature, conductivity/ salinity and pH was measured daily on-site, and 0.45 µm filtered grab samples were collected daily and stored at
