Calibration of polyurethane foam (PUF) disk passive air samplers for quantitative measurement of polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs): factors influencing sampling rates

PUF disk passive air samplers are increasingly employed for monitoring of POPs in ambient air. In order to utilize them as quantitative sampling devices, a calibration experiment was conducted. Time integrated indoor air concentrations of PCBs and PBDEs were obtained from a low volume air sampler operated over a 50 d period alongside the PUF disk samplers in the same office microenvironment. Passive sampling rates for the fully-sheltered sampler design employed in our research were determined for the 51 PCB and 7 PBDE congeners detected in all calibration samples. These values varied from 0.57 to 1.55 m3 d(-1) for individual PCBs and from 1.1 to 1.9 m3 d(-1) for PBDEs. These values are appreciably lower than those reported elsewhere for different PUF disk sampler designs (e.g. partially sheltered) employed under different conditions (e.g. in outdoor air), and derived using different calibration experiment configurations. This suggests that sampling rates derived for a specific sampler configuration deployed under specific environmental conditions, should not be extrapolated to different sampler configurations. Furthermore, our observation of variable congener-specific sampling rates (consistent with other studies), implies that more research is required in order to understand fully the factors that influence sampling rates. Analysis of wipe samples taken from the inside of the sampler housing, revealed evidence that the housing surface scavenges particle bound PBDEs.     

Polyurethane foam (PUF) disks passive air samplers: wind effect on sampling rates

Different passive sampler housings were evaluated for their wind dampening ability and how this might translate to variability in sampler uptake rates. Polyurethane foam (PUF) disk samplers were used as the sampling medium and were exposed to a PCB-contaminated atmosphere in a wind tunnel. The effect of outside wind speed on PUF disk sampling rates was evaluated by exposing polyurethane foam (PUF) disks to a PCB-contaminated air stream in a wind tunnel over air velocities in the range 0 to 1.75 m s-1. PUF disk sampling rates increased gradually over the range 0-0.9 m s-1 at approximately 4.5-14.6 m3 d-1 and then increased sharply to approximately 42 m3 d-1 at approximately 1.75 m s-1 (sum of PCBs). The results indicate that for most field deployments the conventional 'flying saucer' housing adequately dampens the wind effect and will yield approximately time-weighted air concentrations.     

Polycyclic aromatic hydrocarbons in ambient air in the Philippines derived from passive sampler with polyurethane foam disk

Passive samplers with polyurethane disks (PUF) were applied in the determination of the concentration of polycyclic aromatic hydrocarbons (PAHs) in ambient air in six residential areas in the Philippines during four simultaneous sampling periods. The uptake profiles of PAHs were determined at one site during one sampling period. Most of the PAHs that were detected in air at concentrations that were significantly higher than their analytical detection limits exhibited a linear uptake trend on the PUF disk. The linear uptake profiles of some high molecular weight (HMW) PAHs were not established and this is attributed to the low concentration of the compounds in air in the gaseous phase. The retention concentrations of phenanthrene-d-10 were determined after depuration in four sampling sites during two sampling periods. The sampling rate for phenanthrene-d-10 was calculated at the linear phase of the uptake using the k_A derived from depuration experiments and the relationship of k_A and sampling rate which was established in a previous passive sampling study. The average sampling rate obtained for phenanthrene d-10 (2.94±0.69 m³ d⁻¹) was applied for derivation of the concentrations of the PAHs in the field samples.The passive sampler with PUF disk and short integration time of 42–56 days is applicable for the derivation of the concentrations of PAHs in ambient air in the Philippines. The concentrations of the organic pollutants derived from the passive sampler showed variability for the six residential areas; reflecting the influence of possible sources of emission of the pollutants at the sites at the different sampling periods. The weather conditions, including the occurrence of a tropical cyclone, increased rainfall and high-relative humidity during the rainy season, had an influence on the concentrations of PAHs derived by the passive sampler.     

Further studies on the uptake of persistent organic pollutants (POPs) by polyurethane foam disk passive air samplers

Passive air samplers (PAS) can be used to monitor semi-volatile organic compounds in the atmosphere. Polyurethane foam (PUF) disks are a popular sampling medium because they have a high retention capacity for such compounds. This paper reports a highly time-resolved uptake study, to derive uptake rate data under field conditions, and investigate the effects of using different foam densities on the uptake rate. PUF disks were deployed alongside an active sampler, for periods of up to 12 weeks. The uptake rates were measured for a range of gas- and particle-bound persistent organic pollutants (polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs)), of different properties, to explore whether gas–particle partitioning affected uptake rate. Uptake rates for two different densities of foam (0.021 and 0.035 g cm^?3) were not statistically significantly different from each other. Uptake rates of light PCBs averaged ～6.5 m³ day^?1, somewhat higher than in previous studies; higher wind speeds and lower temperatures in this study are the likely reason for this difference. The study showed: i) the uptake rate of the compound with lowest K_OA considered in this study (PCB-28/31) declined in the later weeks, indicating an approach to equilibrium; ii) uptake rates of lighter BDEs and heavier PCBs (compounds of intermediate K_OA in this study) remain similar throughout the study period, indicating that they are not approaching equilibrium during the 12-week-study; iii) uptake rates were typically: ～8 m³ day^?1 for PCB-52; ～9.5 m³ day^?1 for PCB-95; ～11 m³ day^?1 for BDE-28 and ～2 m³ day^?1 BDE-99. The latter compound has an important particle-bound component and this lowers the sampling rate compared to predicted uptake rates for compounds which are in the gas phase only. It is shown that knowledge of gas–particle partitioning is needed to correct for this effect, and to improve predicted uptake rates.     

Computational fluid dynamic modeling of two passive samplers

To effectively use a passive sampler for monitoring trace contaminants in the gas-phase, its sampling characteristics as a function of ambient wind conditions must be known. In this study two commonly used passive samplers were evaluated using computational fluid dynamics. Contaminant uptake by the polyurethane foam (PUF) was modeled using a species transport model. The external-internal flow interactions in the sampler were characterized, and the uptake rates of contaminant species were quantified. The simulations show that flow fields in the samplers have strong velocity gradients, and single-point velocity measurements do not capture flow interactions accurately. Sampling rates calculated for a PUF in freestream are in good agreement with sampling rates for PUFs in the passive samplers studied for the same average velocity over the PUF. The calculated sampling rates are in general agreement with those obtained experimentally by other researchers.     

Passive sampler derived air concentrations of PBDEs along an urban-rural transect: spatial and temporal trends

Passive air samplers consisting of polyurethane foam (PUF) disks housed in chambers were deployed at several sites along a approximately 75 km urban-rural gradient in Toronto and analyzed for polybrominated diphenyl ethers (PBDEs). Samplers were allowed to integrate gas-phase PBDEs over three consecutive seasons starting in the summer of 2000. PBDEs were fairly uniform along the transect with air concentrations in Toronto (10-30 pgm(-3)) about a factor of two greater than at rural sites. Lowest concentrations were observed during the winter and probably associated with reduced inputs from indoor sources of PBDEs and to a preference for PBDEs to partition to the particle-phase at colder temperatures. The composition of PBDEs in the air samples did not differ across the transect or for the different seasons.     

A comparative evaluation of passive and active samplers for measurements of gaseous semi-volatile organic compounds in the tropical atmosphere

The polyurethane foam (PUF) disk-based passive air samplers (PAS), mounted inside two aluminium bowls to buffer the air flow to the disk and to shield it from precipitation and sunlight, were used for the collection of atmospheric SVOCs in Singapore during April 2008–June 2008. Data obtained from PAS measurements are compared to those from active high-volume air sampling (AAS). Single factor ANOVA tests show that there were no significant differences in chemical distribution profiles between actively and passively collected samples (PAHs, F = 3.38 × 10^?8 < F_critical = 4.17 with p > 0.05; OCPs, F = 2.71 × 10^?8 < F_critical = 4.75 with p > 0.05). The average air-side mass transfer coefficient (k_A) for PAS, determined from the loss of depuration compounds such as 13C₆ – HCB (1000 ng), 13C₁₂ – 4,4′ DDT (1000 ng) and 13C₁₂ – PCB 101 (1000 ng)spiked on the disks prior to deployment, was 0.12 ± 0.04 m s^?1. These values are comparable to those reported previously in the literature. The average sampling rate was 3.78 ± 1.83 m³ d^?1 for the 365 cm² PUF disk. Throughout the entire sampling period (～68 d), most of the PAHs and all OCPs exhibited a linear uptake trend on PAS, while naphthalene, acenaphthylene, acenaphthene and fluorene reached the curvilinear phase after the first ～30 d exposure. Theoretically estimated times to equilibrium (t_eq) ranged from around one month for Acy to hundreds of years for DB(ah)A. Sampling rates, based on the time integrated active sampling-derived concentrations and masses collected by PUF disks during the linear uptake phase, were determined for all target compounds with the average values of 2.50 m³ d^?1 and 3.43 m³ d^?1 for PAHs and OCPs, respectively. More variations were observed as compared to those from the depuration study. These variation were most likely due to the difference of physicochemical properties of individual species. Lastly, multiple linear regression models were developed to estimate the log-transformed gaseous concentration of an individual compound in air based on the mass collection rate of the gaseous SVOCs measured using the PAS and the molecular weight (MW) of the particular compound for both PAHs and OCPs, respectively.     

Determination of deployment specific chemical uptake rates for SDB-RPD Empore disk using a passive flow monitor (PFM)

The use of the adsorbent styrenedivinylbenzene-reverse phase sulfonated (SDB-RPD) Empore disk in a chemcatcher type passive sampler is routinely applied in Australia when monitoring herbicides in aquatic environments. One key challenge in the use of passive samplers is mitigating the potentially confounding effects of varying flow conditions on chemical uptake by the passive sampler. Performance reference compounds (PRCs) may be applied to correct sampling rates (R_s) for site specific changed in flow and temperature however evidence suggests the use of PRCs is unreliable when applied to adsorbent passive samplers. The use of the passive flow monitor (PFM) has been introduced for the assessment of site-specific changes in water flow. In the presented study we have demonstrated that the R_s at which both atrazine and prometryn are accumulated within the SDB-RPD-Empore disk is dependent on the flow conditions. Further, the calibration of the measured R_s for chemical uptake by the SDB-RPD-Empore disk to the mass lost from the PFM has shown that the PFM provides an accurate measure of R_s for flow velocities from 0 to 16 cm s⁻¹. Notably, for flow rates >16 cm s⁻¹, a non linear increase in the R_s of both herbicides was observed which indicates that the key resistance to uptake into the SDB-RPD Empore disk is associated with the diffusion through the overlying diffusion limiting membrane. Overall the greatest uncertainty remains at very low flow conditions, which are unlikely to often occur in surface waters. Validation of the PFM use has also been undertaken in a limited field study.     

A directional passive air sampler for monitoring polycyclic aromatic hydrocarbons (PAHs) in air mass

A passive air sampler was developed for collecting polycyclic aromatic hydrocarbons (PAHs) in air mass from various directions. The airflow velocity within the sampler was assessed for its responses to ambient wind speed and direction. The sampler was examined for trapped particles, evaluated quantitatively for influence of airflow velocity and temperature on PAH uptake, examined for PAH uptake kinetics, calibrated against active sampling, and finally tested in the field. The airflow volume passing the sampler was linearly proportional to ambient wind speed and sensitive to wind direction. The uptake rate for an individual PAH was a function of airflow velocity, temperature and the octanol-air partitioning coefficient of the PAH. For all PAHs with more than two rings, the passive sampler operated in a linear uptake phase for three weeks. Different PAH concentrations were obtained in air masses from different directions in the field test.     

Passive air monitoring of PCBs and PCNs across East Asia: a comprehensive congener evaluation for source characterization

A comprehensive congener specific evaluation of polychlorinated biphenyls (PCBs) and polychlorinated naphthalenes (PCNs) in the atmosphere was conducted across East Asia in spring 2008, applying polyurethane foam (PUF) disk passive air sampler (PAS) as monitoring device. Mean concentrations derived for Japan, China and Korea were 184 ± 24, 1100 ± 118, and 156 ± 20 pg m⁻³ for ∑₂₀₂ PCBs, and 9.5 ± 1.5, 61 ± 6, and 16 ± 2.4 pg m⁻³ for ∑₆₃ PCNs, respectively. Relative to reported data from 2004, the present results suggest that air PCBs concentrations have not changed much in Japan and Korea, while it has increased by one order of magnitude in China. From principal component analysis, combustion emerged highly culpable in contemporary emissions of both PCBs and PCNs across the East Asian sub-region. Another factor derived as important to air PCBs was re-emissions/volatilization. Signals from PCBs formulations were also picked, but their general importance was virtually consigned to the re-emissions/volatilization tendencies. On the contrary, counterpart PCNs formulations did not appear to contribute much to air PCNs.     

Air concentrations of PCDD/Fs, PCBs and PCNs using active and passive air samplers

The concentrations of polychlorinated dibenzo-p-dioxins and furans (PCDD/Fs), polychlorinated biphenyls (PCBs) and polychlorinated naphthalenes (PCNs) were determined in air samples collected at four sampling sites located in two zones of Barcelona (Spain): near a municipal solid waste incinerator (MSWI) and a combined cycle power plant (3 sites), and at a background/control site. Samples were collected using high-volume active samplers. Moreover, 4 PUF passive samplers were deployed at the same sampling points during three months. For PCDD/Fs, total WHO-TEQ values were 27.3 and 10.9 fg WHO-TEQm(-3) at the urban/industrial and the background sites, respectively. The sum of 7 PCB congeners and the Sigma PCN levels were also higher at the industrial site than at the background site. In order to compare active and passive sampling, the accumulated amounts of PCDD/Fs, PCBs and PCNs in the four passive air samplers, as well as the total toxic equivalents in each sampling site were also determined. To assess the use of PUF passive samplers as a complementary tool for PCDD/F, PCB and PCN monitoring, sampling rates were calculated in accordance with the theory of passive air samplers. PUF disks allowed establishing differences among zones for the POP levels, showing that they can be a suitable method to determine POP concentrations in air in areas with various potential emission sources. Although both particle and gas phase were sorbed by the PUFs, data of gas phase congeners are more reproducible.     

Atmospheric monitoring of organochlorine pesticides across some West African countries

Most African countries have ratified the Stockholm Convention on persistent organic pollutants (POPs) and are expected to reduce emissions of POPs such as organochlorine pesticides (OCPs) to the atmosphere. Emerging evidence, however, suggests that there are contemporary sources of OCPs in African countries despite the global ban on these products. This study investigated the atmospheric contamination from OCPs in four West African countries—Togo, Benin, Nigeria, and Cameroon—to ascertain the emission levels of OCPs and the characteristic signatures of contamination. Polyurethane foam (PUF) disk passive air samplers (PAS) were deployed in each country for ca. 55 days in 2012 and analyzed for 25 OCPs. Hexachlorocyclohexanes (HCHs) and DDTs constituted the highest burden of atmospheric OCPs in the target countries, at average concentrations of 441 pg m^?3 (range 23–2718) and 403 pg m^?3 (range 91–1880), respectively. Mirex had the lowest concentration, ranged between 0.1 and 3.3 pg m^?3. The concentration of OCPs in rainy season was higher than in dry season in Cameroon, and presupposed inputs from agriculture during the rainy season. The concentrations of ∑25 OCPs in each country were in the following order: Cameroon > Nigeria > Benin > Togo. There was significant evidence, based on chemical signatures of the contamination that DDT, aldrin, chlordane, and endosulfan were recently applied at certain sites in the respective countries.     

Field validation of an active sampling cartridge as a passive sampler for long-term carbonyl monitoring

A carbonyl sampler originally designed for the active sampling method (Sep-Pak XPoSure) was used for long-term passive sampling, and its applicability as a passive sampler was examined through field experiments. The uptake rates of passive sampling were determined experimentally from collocated passive and active samplings for various sampling periods. The obtained uptake rates of formaldehyde, acetaldehyde, and acetone were 1.48, 1.23, and 1.08 mL/min, respectively. These uptake rates were consistent for a wide range of the sampling term (12 hr-2 weeks). Uptake rates of each carbonyl were proportional to the diffusion coefficients of each. Therefore, the ratios of diffusion coefficients were used to calculate the uptake rates of carbonyls for which the rates were not determined experimentally. Lower limits of determination were 2.16-17.5 microg/m3 for 2-week sampling. It was confirmed that 2-week monitoring of carbonyl concentrations up to 118-229 microg/m3 was possible. Relative standard deviations of the passive method generated from the repeatability test were 2-12.3% error for five samplings, and the recovery efficiencies were larger than 90%. Thus, the passive sampler was found to be highly suitable for long-term monitoring of carbonyl compounds.     

Assessing seasonal and spatial trends of persistent organic pollutants (POPs) in Indian agricultural regions using PUF disk passive air samplers

The first survey of persistent organic pollutant (POP) concentrations in air across several Indian agricultural regions was conducted in 2006-2007. Passive samplers comprising polyurethane foam (PUF) disks were deployed on a quarterly basis at seven stations in agricultural regions, one urban site and one background site. The project was conducted as a sub-project of the Global Atmospheric Passive Sampling (GAPS) Network. In addition to revealing new information on air concentrations of several organochlorine pesticides (OCPs) and polychlorinated biphenyls (PCBs), the study has demonstrated the feasibility of conducting regional-scale monitoring for POPs in India using PUF disk samplers. The following analytes were detected with relatively high concentrations in air (mean for 2006 and 2007, pg/m³): α- and γ-hexachlorocyclohexane (HCH) (292 and 812, respectively); endosulfan I and II (2770 and 902, respectively); p,p′-DDE and p,p′-DDT (247 and 931, respectively); and for the sum of 48 PCBs, 12,100 (including a site with extremely high air concentrations in 2007) and 972 (when excluding data for this site).     

New passive samplers for chlorinated semivolatile organic pollutants in ambient air

Two new types of passive samplers were designed and tested on semivolatile organic compounds. The first type (a spiral-rod sampler) consists of a low-density polyethylene membrane acting as a permeation film and a silicone elastomer as the receiving material; the second (a stir-bar sampler) has the same membrane material but a polydimethylsiloxane-coated stir bar acting as the collector phase and installed radially symmetrically in the sampler. The advantages of the new samplers are their simple design, low costs, and their easy processing via thermodesorption coupled with capillary gas chromatography and mass selective detection. In both samplers, the uptake of selected analytes was integrative over exposure periods of up to 384 h. The sampling rates calculated from a laboratory calibration study using the chlorinated semivolatiles hexachlorobenzene, hexachlorocyclohexane isomers and polychlorinated biphenyls ranged from 88.1 ml h-1 for delta-hexachlorocyclohexane to 3443 ml h-1 for 2,2',5,5'-tetrachlorobiphenyl. A field trial at a hazardous waste dump near Bitterfeld, Germany, for up to 21 days combined with periodical determinations of air concentrations using low-volume sampling indicated that the new samplers can in principle be used in the field, although the sampling rates derived from the field results differed considerably from the laboratory findings. Nevertheless the preliminary results suggest that the new sampler types are promising for the long-term air monitoring of semivolatiles.     

Quality control for sampling of PCDD/PCDF emissions from open combustion sources

Both long duration (>6 h) and high temperature (up to 139 °C) sampling efforts were conducted using ambient air sampling methods to determine if either high volume throughput or higher than ambient air sampling temperatures resulted in loss of target polychlorinated dibenzodioxins/dibenzofurans (PCDDs/PCDFs) from a polyurethane foam (PUF) sorbent. Emissions from open burning of simulated military forward operating base waste were sampled using EPA Method TO-9A for 185 min duration using a filter/PUF/PUF in series combination. After a 54 m³ sample was collected, the sampler was removed from the combustion source and the second PUF was replaced with a fresh, clean PUF. An additional 6 h of ambient air sampling (171 m³) was conducted and the second PUF was analyzed to determine if the PCDD/PCDF transferred from the filter and the first PUF. Less than 4.4% of the initial PCDD/PCDF was lost to the second PUF. To assess the potential for blow off of PCDD/PCDF analytes during open air sampling, the mobility of spiked mono- to hepta-PCDD/PCDF standards across a PUF sorbent was evaluated from ambient air temperatures to 145 °C with total volumes between 600 L and 2400 L. Lower molecular weight compounds and higher flow amounts increased release of the spiked standards consistent with vapor pressure values. At 600 L total sampled volume, the release temperature for 1% of the tetra-CDD (the lowest chlorinated homologue with a toxic compound) was 87 °C; increasing the volume fourfold reduced this temperature to 73 °C.     

Performance evaluation of a tailor-made passive sampler for monitoring of tropospheric ozone

Introduction

This study presents the performance evaluation of a tailor-made passive sampler developed for the monitoring of tropospheric ozone.

Methods

The performance of the passive sampler was tested in the field conditions in terms of accuracy, precision, blank values, detection limit, effects of some parameters such as sampling site characteristics and sampling period on the field blanks, self-consistency, experimental and theoretical uptake rates, shelf life and comparison with commercial passive samplers.

Results

There was an agreement (R ²?=?0.84) between the responses of passive sampler and the continuous automatic analyser. The accuracy of the sampler, expressed as percent relative error, was obtained lower than 15%. Method precision in terms of coefficient of variance for three simultaneously applied passive samplers was 12%. Sampler detection limit was 2.42???g?m^?3 for an exposure period of 1?week, and the sampler can be stored safely for a period of up to 8?weeks before exposure. Satisfactory self-consistency results showed that extended periods gave the same integrated response as a series of short-term samplers run side by side. The uptake rate of ozone was found to be 10.21?mL?min^?1 in a very good agreement with the theoretical uptake rate (10.32?mL?min^?1). The results of the comparison study conducted against a commercially available diffusion tube (Gradko diffusion tube) showed a good linear relationship (R ²?=?0.93) between two passive samplers.

Conclusions

The sampler seems suitable to be used in large-scale measurements of ozone where no data are available or the number of existing automated monitors is not sufficient.     

Passive air sampling theory for semivolatile organic compounds

The mathematical modelling underlying passive air sampling theory can be based on mass transfer coefficients or rate constants. Generally, these models have not been inter-related. Starting with basic models, the exchange of chemicals between the gaseous phase and the sampler is developed using mass transfer coefficients and rate constants. Importantly, the inter-relationships between the approaches are demonstrated by relating uptake rate constants and loss rate constants to mass transfer coefficients when either sampler-side or air-side resistance is dominating chemical exchange. The influence of sampler area and sampler volume on chemical exchange is discussed in general terms and as they relate to frequently used parameters such as sampling rates and time to equilibrium. Where air-side or sampler-side resistance dominates, an increase in the surface area of the sampler will increase sampling rates. Sampling rates are not related to the sampler/air partition coefficient (K(SV)) when air-side resistance dominates and increase with K(SV) when sampler-side resistance dominates.     

A multi-variate statistical model integrating passive sampler and meteorology data to predict the frequency distributions of hourly ambient ozone (O3) concentrations

A multi-variate, non-linear statistical model is described to simulate passive O3 sampler data to mimic the hourly frequency distributions of continuous measurements using climatologic O3 indicators and passive sampler measurements. The main meteorological parameters identified by the model were, air temperature, relative humidity, solar radiation and wind speed, although other parameters were also considered. Together, air temperature, relative humidity and passive sampler data by themselves could explain 62.5-67.5% (R(2)) of the corresponding variability of the continuously measured O3 data. The final correlation coefficients (r) between the predicted hourly O3 concentrations from the passive sampler data and the true, continuous measurements were 0.819-0.854, with an accuracy of 92-94% for the predictive capability. With the addition of soil moisture data, the model can lead to the first order approximation of atmospheric O3 flux and plant stomatal uptake. Additionally, if such data are coupled to multi-point plant response measurements, meaningful cause-effect relationships can be derived in the future.     

Calibration and field application of a solvent-based cellulose membrane passive sampling device for the monitoring of polar herbicides

A passive sampler device selective for hydrophilic analytes was constructed from cellulose membrane (40microm thickness) pre-stained with ruthenium red for 96-168h to impede degradation of the cellulose. The sampling device consisted of pre-stained cellulose membrane tubing containing a binary mixture of the solvents 1-dodecanol and 2,2,4-trimethylpentane as the sequestering medium. A laboratory flow-through system was used to investigate the rates of uptake of herbicides into the solvent mixture of the device and their release. The target herbicides were diuron, atrazine, metolachlor and molinate. Uptake of the herbicides into the solvent mixture of the cellulose membrane device was linear for up to 22 days, and daily sampling rates were determined. Release half-lives from the solvent mixture of the sampling device varied from 14 days for diuron, 15 days for atrazine, 84 days for metolachlor and 28 days for molinate. A field study was undertaken to determine if herbicide concentrations in agricultural drainage water derived from the passive sampler devices deployed for periods from 7 to 22 days, using the laboratory-derived sampling rates, would compare closely with time-weighted average herbicide concentrations determined from extractions of daily composite water samples. The concentrations of diuron, atrazine, metolachlor and molinate determined using the cellulose membrane devices were within twofold of the cumulated mean of the daily drainage water extractions.