Environ. Sci. Technol. 2009, 43, 1443–1448
Time Integrative Passive Sampling: How Well Do Chemcatchers Integrate Fluctuating Pollutant Concentrations? M E L A N I E S H A W * ,†,‡ A N D JOCHEN F. MUELLER† National Research Centre for Environmental Toxicology (EnTox), University of Queensland, 39 Kessels Road, Coopers Plains, 4108 Queensland, Australia, and Catchment to Reef Cooperative Research Centre (CRC)
Received July 31, 2008. Revised manuscript received November 25, 2008. Accepted December 1, 2008.
Many environments are subject to periodic fluctuations in pollutant concentrations. Passive samplers, which can provide time weighted average concentrations integrated over the period of deployment, are ideally suited as monitoring tools for these environments. However, the potential for fluctuating concentrations to limit the integrative period of sampling needs to be investigated before sampling data from these environments can be interpreted with confidence. In this study, Chemcatchers using SDB-RPS Empore disks as the sorbent phase were exposed to herbicides for 28 days in a calibration chamber. A pulsed event of 10-fold greater concentrations was introduced on day 5 and returned to background concentrations over a period of 3 days. Observed uptake was compared with that predicted by a first order uptake model and by the reduced form of this model describing a strictly integrative response for samplers deployed with two surfaces exposed (two-sided naked), with one surface exposed (one-sided naked) and with a polyethersulfone membrane. Membrane covered samplers predicted time weighted average water concentrations within a factor 0.7-1.2 after 28 days exposure, while one- and two-sided naked samplers under predicted the average by a factor 1.9-2.2 and 2.4-3.2, respectively. First order modeling predicted uptake in membrane covered and one-sided naked samplers and was therefore applied to predict sampler response to several fluctuating concentration event scenarios.
Introduction Among the greatest advantages provided by the use of passive sampling techniques for pollution monitoring over traditional grab sampling is the potential to derive time weighted average concentrations rather than point in time measurements. Knowledge of time weighted average concentrations is important for ecological risk assessment and allows representative values to be considered in decision making. The ability to derive average concentration values is particularly attractive in environments subject to periodic peak pollution events such as industrial discharges, stormwater runoff or floods (1-4). Flood events, for example, often result in a * Corresponding author phone: +61-7-32749147; fax: +61-732749003; e-mail;
[email protected]. † University of Queensland. ‡ Catchment to Reef Cooperative Research Centre (CRC). 10.1021/es8021446 CCC: $40.75
Published on Web 02/04/2009
2009 American Chemical Society
pulse of increased herbicide concentrations in waterways (1, 5, 6). Characterization of these events poses a challenge to environmental monitoring, and a sampling tool capable of providing integrated concentration estimates for herbicides throughout flood events is highly desirable. Monitoring of fluctuating pollutant concentrations using grab sampling requires extensive repetitive sampling which can be both expensive and logistically difficult to achieve, while integrative passive samplers are ideally suited to monitoring such events. SDB-RPS Empore disks contain a modified copolymer that imparts a high affinity for polar compounds, and the suitability of these disks as passive samplers for compounds such as triazine and phenylurea herbicides has been demonstrated (7). Uptake of polar compounds (log10KOW 1.79-4.0) has been shown to remain integrative for a period of 30 days in samplers deployed with a diffusion limiting membrane and for periods of 5-10 days in samplers without a membrane under constant water concentrations (8). The uptake of pollutants from the environment into a sampler can be described mathematically by the following first order differential equation (9):
(
CS dCS koA C ) dt VS W KSW
)
(1)
where CS is the pollutant concentration in the sampler, CW is the ambient water concentration, KSW represents the sampler to water partition coefficient and VS is the volume of the sampler. Together, ko, the overall mass transfer coefficient and A, the sampler surface area, describe the rate of analyte accumulation (L day-1) into the sampler (RS ) koA). This equation describes the full first order uptake curve which incorporates a linear uptake phase, a curvilinear uptake phase and finally an equilibrium phase. Derivation of time weighted average environmental concentrations from the levels sorbed by a sampler assumes a response to all environmental fluctuations and negligible clearance of pollutants following pulsed events. This assumption holds true in the initial, linear phase of uptake. Samplers remain in the linear (or integrative) phase of uptake when CW . CS/KSW, in which case eq 1 can be reduced to dCS RS ) CW dt VS
(2)
When the sampler has been deployed for an extended period and CW ≈ CS/KSW, equilibrium is reached and there is no change in the sampler concentration with time providing the environmental concentration and temperature remain constant. Passive sampling techniques are widely accepted as monitoring tools capable of taking into account extreme variations in pollutant concentrations (e.g., refs 10-12). The characteristic half-life of a compound in a sampler is commonly used to predict the period of time that a sampler will remain in the linear phase of uptake (13, 14) and samplers deployed for less than this time are considered to be sampling integratively. However, this calculation assumes relatively constant water concentrations without dramatic drops which bring the sampler concentration closer to the equilibrium value. Gale (14) conducted a modeling exercise in which the integrative sampling ability of semipermeable membrane devices (SPMDs) exposed to fluctuating pollutant concentrations was investigated and suggested that errors of
