In this work, an integrative passive sampler based on a silicone membrane filled with a suspension of γ-Fe2O3 at pH 3.5 was developed. The novel device was calibrated for the measurement of microcystin concentrations in water. Laboratory calibration studies of the passive sampling devises under controlled conditions of temperature, water turbulence, and analyte concentration were conducted in order to establish how variable environmental conditions affect the novel sampler's performance. The chemical uptake of microcystin (MC)-RR, -LR, and -YR into the passive sampler remained linear and integrative throughout the 28-day exposure. The relative standard deviations of mean concentrations obtained using silicone-based sampler ranged from 1.42 to 3.74% for microcystin-LR, -RR, and -YR. The values for reproducibility from triplicate samplers ranged from 3.5 to 7.1% for microcystin-LR, -RR, and -YR. The detection limits on high performance liquid chromatography (HPLC) with PDA detection for microcystins LR, RR, and YR were 24.7, 17.2, and 23.8 μg L−1 respectively, calculated as three times the signal to noise ratio. The rate of accumulation of most of the MC compounds tested was dependent on temperature and flow velocity. Furthermore, the sample matrix, e.g. humic substances, had no significant effect on the concentration of compounds trapped in the acceptor solution and once these MC compounds were trapped in the acceptor phase they did not diffuse back during the deployment period.

INTRODUCTION

Anthropogenic activities, along with deficient water management systems, have led to the enhancement of eutrophication in water bodies (Carmichael 2007). These eutrophication processes plus specific environmental conditions such as high temperature and pH values, low turbulence and high nutrients input have in turn led to harmful blue-green algal blooms, which are characterized by excessive proliferation of cyanobacteria cells (de Figueiredo et al. 2004) resulting in the release of biotoxins. Consequently, most countries have begun to evaluate the occurrence, health effects, and susceptibility of water treatments from these algal toxins. Some studies have demonstrated that microcystins may pass through the drinking water treatment plants into the tap waters (Rapala et al. 2002). Microcystins (MCs) are the most frequently occurring class of biotoxins (hepatotoxins), of which MC-LR (most toxic of all), MC-RR and MC-YR are the most toxic and frequently detected congeners. MC-LR congener is shown in Figure 1 as an example on MCs; it is a cyclic heptapeptide containing five amino acids invariant in all MCs, and two specific amino acids, leucine and arginine, designated ‘L’ and ‘R’, respectively. Based on the toxicity data, the World Health Organization (WHO) suggested the tolerable daily intake (TDI) value for MC-LR to be 0.04 μg kg−1 of body weight, and recommended a corresponding safety guideline value of 1.0 μg L−1 for drinking waters (WHO 1998) and the United States Environmental Protection Agency (USEPA) has placed MCs on the Drinking Water Contaminant Candidate List (USEPA 2005). Some studies demonstrated that MCs may pass through the drinking water treatment plants into the tap waters (Rapala et al. 2002), and thus monitoring of these biotoxins in the aquatic environment is essential to satisfy the requirements of legislative frameworks and directives as these compounds poses threats to both human health and ecosystems. Traditional monitoring programs of MCs have been based on collection of individual samples at specific single spot and time points, then analysed for the potential pollutants in the laboratory (Huckins et al. 1993). But, information obtained from such grab environmental samples has only been about concentration levels at the time of sampling and may fail to account for episodic contaminations.
Figure 1

General molecular structure of MC molecules.

Figure 1

General molecular structure of MC molecules.

Similarly, these traditional sampling techniques have several drawbacks such as need of larger volumes to recover sufficient mass of toxin or time- and labour-consuming clean up prior to instrumental analyses. One solution to this problem is to increase the frequency of sampling or to install automatic sampling systems that can take numerous water samples over a given time period. This is costly and in many cases impractical, since a secure site and significant pre-treatment of water are required (Nyoni et al. 2011). Such systems are rarely used in widespread monitoring campaigns. Further, during traditional monitoring scheme, episodic events may be missed due to MC concentrations varying over the sampling period. Therefore, it may be difficult to formulate the time-weighted average (TWA) concentration of the contaminant, which forms a fundamental part of an ecological risk assessment process (Petty et al. 2004; Alvarez et al. 2004). Another approach that yields information on biologically relevant concentrations of MCs uses biota. A number of test species can be used, depending on the water body being investigated. These organisms can be deployed for extended periods of time, during which they passively bio-accumulate pollutants in the surrounding water. Analysis of the tissues or lipid extracts of the test organism(s) can give an indication of the equilibrium level of waterborne contamination. A number of factors can influence the results-metabolism, depuration rates, excretion, stress, viability and condition of test organism. Moreover, extraction of analytes from the tissue of animals prior to instrumental analysis is complex (Vrana et al. 2005). The proposed passive sampler's theory of extraction combined the coupled transport (Figure 2) and the one developed for the supported liquid membrane (SLM) extraction technique (Jönsson et al. 1993; Jönsson & Mathiasson 1999). The MC compounds are selectively driven across silicone membrane from the water sample by a flow of carrier molecules in the opposite direction. Once in the acceptor phase, they adsorb onto the iron oxide nanoparticles and trapped.
Figure 2

Schematic example of carrier-facilitated transport of MCs. The example shows the transport of MCs across a membrane using an ion-exchange reagent as the carrier agent.

Figure 2

Schematic example of carrier-facilitated transport of MCs. The example shows the transport of MCs across a membrane using an ion-exchange reagent as the carrier agent.

The aim of this work is to develop a passive, in situ, integrative sampler for monitoring of biotoxins in water bodies based on a functionalised silicone membrane. The specific objectives of this project are as follows:

  1. To synthesise and characterise γ-Fe2O3 nanoparticles.

  2. To functionalise the silicone membrane by filling it with γ-Fe2O3 nanoparticles for the purpose of using it to sample cylindrospermopsin and MCs at pH values (3 < pH < 12), as demonstrated by Lee & Walker (2011).

  3. To devise a constant concentration flow through exposure system to allow calibration of the passive sampling devices to be made under controlled conditions of temperature, water turbulence and analyte concentration.

  4. To demonstrate the potential of modified silicone membrane for passive sampling of cylindrospermopsin and MCs-LR, -YR, and -RR in water bodies.

EXPERIMENTAL

Materials and methods

Chemicals and materials

CyanoBiotech GmbH, MC standards (MC-LR, MC-RR, and MC-YR) were purchased from Germany and supplied by Industrial Analytical (Pty) Ltd, South Africa. Other solvents and reagents used in this work were of high purity (Analytical grade and/or high performance liquid chromatography (HPLC) grade, >99%) and were purchased from Sigma-Aldrich (South Africa) and Merck (South Africa). Solid-phase extraction SPE cartridges (Oasis™ HLB cartridges, Water and ODS C-18 cartridges) and C18 Empore disks, 47 mm diameter, were purchased from Waters, Inc., USA, supplied by Microsep (Pty) Ltd, South Africa. Silicone membranes were obtained from Technical Products, Inc. (Georgia, USA).

MC mixture

With regards to the relatively large amounts of MCs needed for calibration experiments, and high costs of the toxin analytical standards, MCs for the experiments were isolated from the natural cyanobacterial biomass, collected into amber glass bottles from Hartebeespoort dam, by repeated extraction with 50% methanol and partial purification with solid-phase extraction using an ODS cartridge. The final extract (in 75% v/v methanol: water), containing dominant MC variants, was aliquoted and stored at −18 °C.

Synthesis of iron oxide (maghemite) nanoparticles

The synthesis method of monodispersed maghemite nanoparticles reported by Xu et al. (2010) was adapted. Iron(III) oxide (γ-Fe2O3) nanoparticles were prepared by thermal decomposition of nontoxic and inexpensive iron(III) chloride at temperatures between 240 °C and 320 °C in a solution of sodium oleate, 1-octadecene and oleic acid to enable good crystallinity, high control over the particle size and very narrow particle size distribution of the nanoparticles. Figure 3 shows the experimental set-up put in place. The solution of 0.5 mmol FeCl3·6H2O, 1.5 mmol Na-oleate, 1.5 mmol oleic acid and 10 mL 1-octadecene mixed in a three necked flask was heated to 130 °C for 1 hour under argon flow protection. The solution was then heated up to 317 °C and kept at this temperature for 2 hours. The reaction solution was washed by adding 20 mL hexane, 2 mL acetone, and 20 mL deionized water four times to remove NaCl and the nanoparticles (dispersed in hexane phase) were collected upon adding 20 mL acetone followed by centrifugation. The hexane dispersion of nanoparticles was vigorously stirred for 5 hours in an aqueous solution of tetramethylammonium hydroxide to displace the oleic acid from particle surface. The iron oxide nanoparticles were then collected by precipitation with acetone followed by centrifugation. The identity and purity of the nanoparticles was verified by X-ray diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscope with energy dispersive X-ray spectroscopy (SEM/EDS), surface zeta potential analyser, and N2-BET. Iron oxide solution (acceptor phase) of final maghemite concentration 2.3 g L−1 was prepared in a pH 3.5 buffer of 0.1 M potassium hydrogen phthalate and 0.1 M hydrochloric acid and stored in the dark prior to use.
Figure 3

Experimental set-up for the synthesis of iron oxide nanoparticles by thermal decomposition of iron(III) chloride.

Figure 3

Experimental set-up for the synthesis of iron oxide nanoparticles by thermal decomposition of iron(III) chloride.

Flow through exposure system

The flow through exposure system was devised to allow calibration of the sampling devices to be made under controlled conditions of temperature, water turbulence, and analyte concentration. The system consisted of a 50 L glass tank with an overflow to waste (Figure 4). The distilled deionised water and the solution of test analytes dissolved in methanol were pumped into the exposure tank separately at known and controlled rates. Deionised water was fed to the exposure tank using a peristaltic pump set at a rate of 30 mL minute−1, allowing a complete renewal of water in the tank every 72 hours. Test chemicals dissolved in methanol (10 mg L−1) were delivered into the exposure tank at a rate of 100 μL per minute using a second peristaltic pump. A nominal analyte concentration of 50–55 ng mL−1 was maintained throughout the experiment and the amount of methanol added to water did not exceed 0.0002%.
Figure 4

(a) Experimental set-up to be used in flow through calibration of passive sampling services. (b) Photograph of the experimental set-up (by H. Nyoni).

Figure 4

(a) Experimental set-up to be used in flow through calibration of passive sampling services. (b) Photograph of the experimental set-up (by H. Nyoni).

Fabrication of passive sampling devices

Silicone-membrane/γ-Fe2O3-nanoparticle-sorbent-based passive sampler

Silicone membranes used for the optimization process were bought as long tubes and cut to appropriate lengths (48 cm × 0.1575 cm ID × 0.2413 cm OD giving a volume of 0.9349 cm3 ∼1000 μL). Eighteen silicone membranes, previously soaked in deionised water, were filled with a pH 3.5 acceptor buffer of the maghemite suspension (2.3 g L−1) using a 1000 μL micropipette. The membranes were tightened together and made in the form of a loop about 5 cm in diameter. The outside was rinsed with deionised water thoroughly to remove any buffer spills. After exposure, the samplers were taken out of the sample vessel, the outside flushed with deionised water, and contents transferred into a 1.5 mL vials. Prior to HPLC analysis, the extracts were sonicated in a Branson 5800 ultrasonic bath (Danbury, USA) for 5 minutes and then centrifuged using an Eppendorf 5430 centrifuge (Hamburg, USA) for 10 minutes at 7500 rpm, and the supernatants were filtered through 0.45-μm polyvinylidene fluoride (PVDF) membrane syringe filters. The extracts were either analysed immediately or stored in the refrigerator at −18 °C.

Polar organic chemical integrative sampler

The commercial polar organic chemical integrative sampler (POCIS) samplers (pharmaceutical configuration) for the comparison experiment were obtained from Exposmeter AB (Tavelsjo, Sweden). In-house made POCIS samplers (pharmaceutical configuration) for the comparison experiment were fabricated in the laboratory. These POCIS samplers contained 300 mg of Oasis HLB sorbent enclosed between two polyethersulphone (PES) membranes. The membrane-sorbent-membrane layers were compressed between two holder washers. The total exchange surface area of the membrane (both sides) was ∼41 cm2 and the surface area per mass of sorbent ratio was ∼300 cm2 g−1. Thereafter, exposed samplers were taken out of the sample vessel, the outside flushed with deionised water and the holder disassembled. Membranes with the enclosed sorbent were transferred into a 15 mL centrifuge tube and extracted two times with 5 mL of aqueous methanol (90% v/v acidified with 0.1% trifluoroacetic acid) for 15 minutes in an ultrasonic bath. After centrifugation (10 minutes at 7500 rpm); supernatants were pooled, evaporated to dryness at 40 °C under a stream of nitrogen and reconstituted with 500 μL of HPLC grade methanol. The extracts were either analysed immediately or stored in the refrigerator at −18 °C.

Solid phase extraction technique

Water samples (100 mL) were filtered through a 0.45 μm membrane and the filtrate divided into three 10 mL portions. The three portions were then extracted using Oasis HLB 3 cc (60 mg)−1 cartridges as outlined in Figure 5.
Figure 5

Outline of the solid phase extraction procedure used to extract MC compounds from grab water samples sampled from the flow through exposure device.

Figure 5

Outline of the solid phase extraction procedure used to extract MC compounds from grab water samples sampled from the flow through exposure device.

Optimisation of extraction parameters

Influence of maghemite suspension on the sampler performance

In order to assess the influence of maghemite suspension on the uptake of the MC compounds by the silicone tube, fifteen tubes filled with a pH 3.5 acceptor buffer without the maghemite suspension (control samplers) were exposed alongside fifteen POCIS. The exposures lasted for 28 days, during which triplicate samplers of both samplers under investigation were removed at set time intervals (0, 7, 14, 21, and 28 days). Every time a POCIS was removed for analysis it was replaced by an empty sampler body. Grab samples of water (100 mL) from the outlet of exposure tank were also taken each time the samplers were removed, and the concentration of biotoxins in the water pre-concentrated by an SPE technique prior to HPLC PDA according to the ISO 20179 method (ISO20179 2005). Figure 4 shows a depiction of the experimental set-up to measure the uptake of target analytes at different combinations of exposure times, temperatures, sample matrix and hydrodynamic conditions.

Exposure time

In order to assess the influence of exposure time on the uptake of the MC compounds, the silicone membrane-based passive samplers filled with a suspension of iron oxide nanoparticles were exposed for 0, 7, 14, 21, and 28 days. Deionised water spiked with 50 ng mL−1 mixtures of MCs was extracted at ambient temperature. In these experiments, up to fifteen silicone membranes, previously soaked in deionised water were used. After exposure, a set of three passive samplers were treated in the same way as described in the extraction procedure.

Effect of hydrodynamics on the uptake kinetics of MCs

The effect of hydrodynamics on the uptake rates of individual MC compounds into the samplers was studied at six different stirring speeds (i.e. 0, 20, 40, 60, 80, and 100 rpm). Deionised water spiked with 50 ng mL−1 mixtures of MCs was extracted for 14 days at ambient temperature. In these influence-of-hydrodynamic experiments, up to 18 silicone membranes, previously soaked in deionised water, were filled with a pH 3.5 acceptor buffer of the maghemite suspension (2.3 g L−1). A set of three passive samplers were exposed in appropriate stirring speeds. After exposure, the passive samplers were treated in the same way as described in the extraction procedure.

Effect of temperature on the uptake kinetics of MCs

The relationship between sampling rates of three MC compounds and temperature was compared at four temperatures (4, 17, 23, and 40 °C). In these influence-of-temperature experiments, up to twelve silicone membranes, previously soaked in deionised water, were filled with a pH 3.5 acceptor buffer of the maghemite suspension (2.3 g L−1). A set of three passive samplers were exposed in appropriate temperature-controlled systems. Deionized water containing ±50 ng mL−1 of a mixture of MCs was used as sample solution. Four sample vessels water were used: (1) vessel placed in a cooler box filled with ice and maintained at 4 °C, (2) vessel maintained at 17 °C, (3) vessel with a heating element set at 23 °C and (4) vessel heated and maintained at 40 °C. Before exposing the silicone-membrane/γ-Fe2O3-passive samplers, the water samples were allowed to equilibrate for at least 2 hours at appropriate temperature. Passive exposure period was for 14 days.

Effect of humic substances on the uptake kinetics of MCs

To study the effects of sample matrix on the performance of the passive sampler, deionised water contain­ing ±50 ng mL−1 of a mixture of MCs and 10 mg L−1 of humic substances was used as a sample solution. This sample solution was extracted with three silicone membranes filled with a pH 3.5 acceptor buffer of the maghemite suspension (2.3 g L−1). Exposure period was 14 days. Afterwards, the sampling rates obtained were compared with sampling rates of samplers exposed to deionised water without any humic substances.

Effects of humic substances on sampler performance

To study whether compounds trapped in the acceptor phase can diffuse back, a series of experiments were performed. A buffer solution of the maghemite suspension (2.3 g L−1) was spiked with 1.0 mg L−1 of MC compounds and filled into the three silicone membranes as before. These were deployed in deionized water spiked with 10 mg L−1 of humic substances only. The exposure period was 14 days. The contents of the acceptor solutions was analysed to check for any loss of MCs from the acceptor solution.

Chromatographic separation of MCs

Extract were analysed by the Surveyor Plus™ modular LC system and the ChromQuest™ data system, products of Thermo Fisher Scientific San Jose, on a 150 mm × 4.6 mm, 5 μm column (waters) at 30 °C with a mobile phase composition of 60% water + 0.1% trifluoroacetic acid and 40% acetonitrile + 0.1% trifluoroacetic acid at a flow rate of 1.0 mL min−1. Chromatograms at 238 nm were recorded with the Surveyor PDA Plus Detector, and MCs were identified by retention time and characteristic UV absorption spectra (200–300 nm). Quantification was based on external calibrations of MC-RR, -LR, and -YR, respectively.

Quality control

In ensuring that the concentration determined using the sampling devices reflect the true picture in the aquatic media, quality control procedures to address issues such as contamination and loss of the trapped analytes, accuracy and precision of the results were conducted. Inspection for signs of puncture, discoloration, or any malfunctioning upon retrieval to see any possible sources of contamination and/or loss of the trapped analytes (EPA 1986; Paschke et al. 2006) was performed. Procedural blanks, certified reference materials, and control samples were used for identification of the contamination from the process (EPA 1986; Paschke et al. 2006).

RESULTS AND DISCUSSION

Identity of the synthesised iron oxide (maghemite) nanoparticles

The identity and purity of the synthesised iron oxides was verified by XRD, TEM, SEM/EDS, surface zeta potential analyser, and N2-BET.

X-ray diffraction of synthesised iron oxide) nanoparticles

Figure 6 shows the XRD (XRD, Rigaku) pattern for the synthesised iron oxide nanoparticles. The sample's XRD pattern matched standard maghemite (γ-Fe2O3). The XRD peaks for the sample were similar in their corresponding ‘d’ values and lattice constants to the standard γ-Fe2O3 phase.
Figure 6

X-ray diffractogram of representative samples.

Figure 6

X-ray diffractogram of representative samples.

Transmission electron microscopy of synthesised iron oxide nanoparticles

TEM (Joel 4000EX HRTEM) showed that maghemite was composed of a mixture of cubic and hexagonal particles with a diameter of 15–100 nm. The particle size distribution was wide. Figure 7(b) shows the selected area electron diffraction pattern. It clearly shows the dot pattern along with the diffused rings indicating thereby the presence of a small mass of ultrafine particles of γ-Fe2O3.
Figure 7

(a) TEM micrograph and (b) selected area diffraction pattern of as-prepared powder.

Figure 7

(a) TEM micrograph and (b) selected area diffraction pattern of as-prepared powder.

Surface zeta potentials and N2-BET of synthesised iron oxide nanoparticles

Surface zeta potentials of synthesised iron oxide nanoparticles were measured by a zeta potential analyser (Malvern Zeta Nanosizer, Malvern, UK) at 0.01 M NaCl. The isoelectric point of maghemite was found to be 8.0, as shown in Figure 8. Specific surface area was determined by N2-BET analysis (Micrometrics) to be 389.816m2 g−1.
Figure 8

Zeta potential of maghemite (0.5 g L−1) as a function of pH in the presence of 0.01 M NaCl.

Figure 8

Zeta potential of maghemite (0.5 g L−1) as a function of pH in the presence of 0.01 M NaCl.

Scanning electron microscopy with energy dispersive X-ray spectroscopy

Figure 9 shows the ED spectrum of synthesised iron oxide nanoparticles, showing K peaks of Fe and O elements. SEM/EDS confirmed that the synthesised material was composed mainly of iron (Fe) and oxygen (O) elements only which further confirms that the synthesised material was an iron oxide.
Figure 9

ED spectrum of synthesised iron oxide nanoparticles, showing peaks of Fe and O elements.

Figure 9

ED spectrum of synthesised iron oxide nanoparticles, showing peaks of Fe and O elements.

Flow through exposure system

A nominal analyte concentration of 50–55 μg L−1 was maintained throughout the experiments (Figure 10). Grab samples (100 mL) of water from the outlet of exposure tank were taken each time the samplers are removed, and the concentration of test analyte in the water determined by an SPE technique.
Figure 10

Nominal analyte concentration of 50–55 μg L−1 maintained in water throughout the experiments.

Figure 10

Nominal analyte concentration of 50–55 μg L−1 maintained in water throughout the experiments.

Suitability studies for integrative sampling: silicone membrane vs. POCIS

The time courses of the amounts of individual test substances into the passive sampler are shown in Figure 11. There was no chemical uptake of MC compounds into the silicone membrane without a suspension of maghemite nanoparticles. But, a linear uptake of MC compounds was observed into the POCIS samplers throughout the exposure period. However, satisfactory linear regression fits of analytes from water to sampler were obtained for test compounds when silicone membranes were filled with a pH 3.5 acceptor buffer of the maghemite suspension (2.3 g L−1). This observation agrees with what Lee & Walker (2011) demonstrated about MCs strongly adsorbing onto iron oxide nanoparticles (γ-Fe2O3).
Figure 11

Uptake of MC-RR, -LR, and -YR compounds by the (a) POCIS and (b) silicone membrane-based passive samplers without maghemite suspension. The data used represent the 23 °C exposure at 50 μg L−1 nominal water concentration of each analyte.

Figure 11

Uptake of MC-RR, -LR, and -YR compounds by the (a) POCIS and (b) silicone membrane-based passive samplers without maghemite suspension. The data used represent the 23 °C exposure at 50 μg L−1 nominal water concentration of each analyte.

Optimization of novel silicone membrane passive sampler

Extraction time

When passive samplers are deployed in the environment, they accumulate micro-pollutants. Chemicals diffuse from the bulk water through the boundary layer to the sampler surface and then partition between the sampler and the water. Depending on the sampler design, the mass of pollutant accumulated by a sampler reflects either the equilibrium or the time-averaged concentration. The use of integrative passive samplers enables direct estimation of time weighted average (TWA) concentrations provided the sampling rates are known (Petty et al. 2004; Alvarez et al. 2004; Namiesnik et al. 2005).

The time courses of the amounts of individual test substances into the passive sampler are shown in Figure 12. The chemical uptake of MC (MCYST)-RR, -LR, and -YR into the passive sampler remained linear and integrative throughout the 28 days' exposure. This means that the extraction can go on until all the analytes in the sample are extracted. Such linear relationships have been observed in other passive samplers working in the kinetic regime, such as the POCIS (Kohoutek et al. 2010). The relative standard deviations of mean concentrations obtained using silicone based sampler were 1.42%, 2.50%, and 3.74% (from triplicate determinations) for MC- LR, RR, and YR, respectively. The use of an experimental tank allowed a degree of control over some environmental conditions, e.g. constant analyte concentration, turbulence (rotation speed) and temperature. This may be the reason for the greater precision obtained in this study compared with what could have been in field trials. An average standard error of 0.13 was obtained for the linear regression giving us an indication that the mean concentration is relatively close to the true mean of the overall population.
Figure 12

(a) Uptake of MC-RR, -LR, and -YR compounds by the silicone membrane-based passive samplers. The data used represent the 23 °C exposure at 50 μg L−1 nominal water concentration of each analyte. (b) Typical chromatogram (HPLC–PDA) obtained after passive extraction of deionised water (50 L) spiked with 50 μg L−1 mixture of MC-RR, -LR, and -YR compounds.

Figure 12

(a) Uptake of MC-RR, -LR, and -YR compounds by the silicone membrane-based passive samplers. The data used represent the 23 °C exposure at 50 μg L−1 nominal water concentration of each analyte. (b) Typical chromatogram (HPLC–PDA) obtained after passive extraction of deionised water (50 L) spiked with 50 μg L−1 mixture of MC-RR, -LR, and -YR compounds.

The limits of detection for MCs LR, RR, and YR were 24.7 μg L−1, 17.2 μg L−1, and 23.8 μg L−1, respectively, calculated as three times the signal to noise ratio. Comparing these data with published data for POCIS (Kohoutek et al. 2010); the sensitivity of the silicone based sampler is lower than POCIS. This could be due to the fact that hydrophilic MCs have large molecular mass (∼1000 Da) and thus do not easily pass across a silicone membrane into the receiving phase. However, mass transfer processes of these biotoxins across a silicone membrane are dependent on the physical form of the biotoxin molecules and the temperature. Geinoz et al. (2002), reported that exposure of silicone material to solvents may result in some swelling of the membrane, thus allowing the permeation of even larger compounds through the material.

Effect of temperature

The relationship between sampling rates of three MC compounds and temperature was compared at four temperatures (4, 17, 23, and 40 °C). The typical dependence of analyte concentration on temperature is shown in Figure 13. In general, the analyte concentration increased with the increasing exposure temperature. The amounts quantified in the silicone membrane based passive samplers had relative standard deviations mostly between 11 and 19% (from triplicate determinations) and did in no case exceed 24%. From a kinetic point of view, it is quite clear that temperature of the environmental media can influence the uptake rates in a sampler. A study on the effect of water temperature over a range of 4°C to 20°C has been reported by Kingston et al. (2000).
Figure 13

Effect of temperature on the analyte concentration. The data represent exposures of three MC compounds, performed at various temperatures (4, 17, 23, and 40 °C).

Figure 13

Effect of temperature on the analyte concentration. The data represent exposures of three MC compounds, performed at various temperatures (4, 17, 23, and 40 °C).

At elevated temperature of 40°C, the cyclic heptapeptide MCs are de-natured and this is seen by the decline of the analyte concentration trapped in the sampler. Therefore, from the results obtained it is quite clear that increased temperature of the environmental media can to some extent enhance mass transfer of MC compounds into the silicone membrane based passive sampler. This is in agreement with what Harada et al. (1996) reported about MCs slowly breaking down at high temperature (40 °C), and at either very low (<1) or high (>9) pH. Tsuji et al. (1995) also reported that MCs’ half-life at pH 1 and 40 °C is 3 weeks, but this can stretch to 10 weeks at typical ambient conditions. In full sunlight, especially when water-soluble pigments are present, these toxins also break down slowly. Although MCs and can be broken down by some bacterial proteases, in many circumstances these bacteria are not present in the cooler, dark, natural water bodies (Jones et al. 1995; Rapala et al. 2005). Thus, these toxins may persist for months or even years once released in such areas. MCs can still persist after boiling, indicating that cooking is not sufficient to destroy them (WHO 1999).

Figure 12(b) is a typical chromatogram (HPLC–PDA) obtained after passive extraction of deionised water (50 L) spiked with 50 μg L−1 mixture of MC-RR, -LR, and -YR compounds.

In this study, one sampler consisted of a polyethersulphone limiting membrane while the other consisted of polyethylene. Both of these samplers used the same 47 mm C18 Empore disk as receiving phase. In both cases, an increase in water temperature led to an increase in sampling rate. The effects of temperature on sampling rates have also been observed in semi-permeable membrane devices (SPMDs) (Yusà et al. 2005) and in membrane enclosed sorptive coating (MESCO) sampler (Vrana et al. 2001). For practical purposes, it is therefore necessary to determine the effects of temperature in the laboratory for each compound of interest and also measure the temperature during field deployment.

Effect of hydrodynamics on sampler's uptake kinetics

Turbulence of the environmental media is one factor which is difficult to control and may affect the amount of compounds trapped in the receiving phase and therefore the quality of the results. The extent to which turbulence affects uptake kinetics depends on factors such as sampler material, hydrophobicity of the compound and environmental flow rates (Vrana et al. 2001). For membrane based passive samplers, permeation of compounds through the membrane is seen as the rate-limiting step and is more pronounced for polar compounds (membrane/permeation controlled samplers). On the other hand, for non-polar compounds, diffusion through the unstirred layer and sampler controls the mass transfer (donor/diffusional controlled (Mergesa et al. 2001)). Once turbulence occurs in the environmental media, the unstirred layer becomes thin and therefore enhances the uptake of non-polar compounds by the sampler. Depending on the extent of the turbulence, at very high flow rates, poor dissolution of polar compounds into the membrane can result in decreased uptake rates. In this work, the analyte concentrations obtained for individual MC compounds at six different stirring speeds were compared (i.e. 0, 20, 40, 60, 80, and 100 rpm). A significant increase of analyte concentration accumulated in the passive samplers with change of hydrodynamic conditions from static to high turbulence was observed for all compounds under investigation (Figure 14). At low turbulence, the diffusion through the aqueous layer limits the mass transfer for more hydrophobic compounds, resulting in higher accumulated analyte concentrations at high turbulence. The extent to which the mass transfer process is influenced by turbulence is dependent on the polarity of the compound and trapping in the acceptor phase. For the relatively hydrophobic compound, an increase in turbulence is accompanied by an increase in the sampling rate. The variation of the analyte concentrations with turbulence for compounds with varying polarity has also been discussed by Nyoni et al. (2011) in the theoretical treatment of SLM technique.
Figure 14

Effect of hydrodynamics on the accumulated analyte concentration. Data are presented from the exposure experiments conducted at 23 °C and stirred and unstirred conditions.

Figure 14

Effect of hydrodynamics on the accumulated analyte concentration. Data are presented from the exposure experiments conducted at 23 °C and stirred and unstirred conditions.

Effects of humic substances on sampler's performance

Figure 15(a) shows the accumulated amounts of MC-LR, -RR, and -YR at two different humic substances concentrations. The amount of biotoxins accumulated in the passive sampler decreased significantly with increasing initial concentration humic substances from 0 to 10 mg L−1. The amounts of MC-LR, -RR, and -YR accumulated in the silicone membrane samplers had relative standard deviations mostly between 1.6 and 2.4% (from triplicate determinations) and did not exceed 4.3%. This negative influence of humic substances on the performance of silicone membrane passive sampler can attributed to humic substances forming weak or strong complexes with target compounds making it difficult for target compounds to move across the membrane. On the other hand, Figure 15(b) shows that humic substances have no significant effect on the concentration of compounds trapped in the acceptor solution. Once these MC compounds are trapped in the acceptor phase, the water sample matrix has no influence on the already trapped.
Figure 15

Influence of sample matrix (humic substances) on the silicone-based passive sampler's performance for MC compounds.

Figure 15

Influence of sample matrix (humic substances) on the silicone-based passive sampler's performance for MC compounds.

CONCLUSIONS

An alternative and more informative, cost-effective approach has been developed and calibrated in the laboratory to obtain a time-weighted average (TWA) concentration of MCs, which forms a fundamental part of an ecological risk assessment process. The potential of the silicone membrane functionalised with γ-Fe2O3 nanoparticles for passive sampling of MC-LR, -YR, and -RR in water bodies has been demonstrated. Variable environmental conditions (e.g. effects of temperature, sample matrix, and hydrodynamics) were found to affect the sampler performance. Furthermore, the sample matrix, e.g. humic substances, do not have an effect in the concentration of compounds trapped in the acceptor solution. More research is necessary to provide an understanding of the effect of bio-fouling on the sampler performance. Other future research will focus on incorporating the performance reference compounds (PRC) concept into the sampler configurations and bioassays in the trapping media as well as field application to test the sampler performance alongside spot sampling and commercially available passive samplers.

ACKNOWLEDGEMENTS

The authors would like to thank the financial support from the Nano-Science Centre, Department of Applied Chemistry, University of Johannesburg as well as their valuable inputs, seminars and guidance.

REFERENCES

REFERENCES
Alvarez
D. A.
Petty
J. D.
Huckins
J. N.
Jones-Lepp
T. L.
Goddard
J. P.
Manahan
S. E.
2004
Development of a passive, in situ, integrative sampler for hydrophilic organic contaminants in aquatic environments
.
Environ. Toxicol. Chem.
23
,
1640
1648
.
Carmichael
W.
2007
A world overview — one-hundred twenty-seven years of research on toxic cyanobacteria — where do we go from here?
In:
Proceedings of the interagency, international symposium on cyanobacterial harmful algal blooms
(
Kenneth Hudnell
H.
, ed.).
Advances in Experimental Medicine and Biology
,
Dayton, OH
, pp.
95
115
.
de Figueiredo
D. R.
Azeiteiro
U. M.
Esteves
S. M.
Goncalves
F. J. M.
Pereira
M. J.
2004
Microcystin-producing blooms-a serious global public health issue
.
Ecotox. Environ. Safe
59
,
151
163
.
Environmental Protection Agency (EPA)
1986
EPA Method 8310, Polynuclear Aromatic Hydrocarbons, EPA, Washington, DC, USA, 8310-1-8310-13
.
Geinoz
S.
Rey
S.
Boss
G.
Bunge
A. L.
Guy
R. H.
Carrupt
P. A.
Reist
M.
Testa
B.
2002
Quantitative structure-permeation relationships for solute transport across silicone membranes
.
Pharm. Res.
19
(
11
),
1622
1629
.
Harada
K. I.
Tsuji
K.
Watanabe
M. F.
1996
Stability of microcystins from cyanobacteria—III. Effect of pH and temperature
.
Phycologia 35
6
,
83
88
.
Huckins
J. N.
Manuweera
G. K.
Petty
J. D.
Mackay
D.
Lebo
J. A.
1993
Lipid-containing semipermeable membrane devices for monitoring organic contaminants in water
.
Environ. Sci. Technol.
27
,
2489
2496
.
ISO20179
.
2005
Water quality: determination of microcystin-method by solid phase extraction (SPE) and high performance liquid chromatography (HPLC) with ultraviolet (UV) detection. ISO, Geneva, Switzerland.
Jones
G.
Falconer
I. R.
Wilkins
R. M.
1995
Persistence of cyclic peptide toxins in dried Microcystis aeruginosa crusts from Lake Mokoan, Australia
.
Environ. Toxicol. Water Qual.
10
(
1
),
19
24
.
Jönsson
J. Å.
Mathiasson
L.
1999
Liquid membrane extraction in analytical sample preparation: II. Applications
.
Trend. Anal. Chem.
18
,
325
334
.
Namiesnik
J.
Zabiega
B.
Kot-Wasik
A.
Partyka
M.
Wasik
A.
2005
Passive sampling and/or extraction techniques in environmental analysis: a review
.
Anal. Bioanal. Chem.
381
,
279
301
.
Paschke
A.
Schwab
K.
Brümmer
J.
Schüürmann
G.
Paschke
H.
Popp
P.
2006
Rapid semi-continuous calibration and field test of membrane enclosed silicone collector as passive water sampler
.
J. Chromatogr. A
1124
,
187
195
.
Petty
J. D.
Huckins
J. N.
Alvarez
D. A.
Brumbaugh
W. G.
Cranor
W. L.
Gale
R. W.
Rastall
A. C.
Jones-Lepp
T. L.
Leiker
T. J.
Rostad
C. E.
Furlong
E. T.
2004
A holistic passive integrative sampling approach for assessing the presence and potential impacts of waterborne environmental contaminants
.
Chemosphere
54
,
695
705
.
Rapala
J.
Lahti
K.
Rasanen
L. A.
Esala
A. L.
Niemela
S. I.
Sivonen
K.
2002
Endotoxins associated with cyanobacteria and their removal during drinking water treatment
.
Water Res.
36
,
2627
2635
.
Rapala
J.
Berg
K. A.
Lyra
C.
Niemi
R. M.
Manz
W.
Suomalainen
S.
Paulin
L.
Lahti
K.
2005
Paucibacter toxinivorans gen. nov., sp. nov., a bacterium that degrades cyclic cyanobacterial hepatotoxins microcystins and nodularin
.
Int. J. Syst. Evol. Microbiol. 55
4
,
1563
1568
.
Tsuji
K.
Watanuki
T.
Kondo
F.
Watanabe
M. F.
Suzuki
S.
Nakazawa
H.
Suzuki
M.
Uchida
H.
Harada
K. I.
1995
Stability of microcystins from cyanobacteria – II. Effect of UV light on decomposition and isomerization
.
Toxicon
33
(
12
),
1619
1631
.
USEPA
2005
Fact sheet: The drinking water contaminant candidate list – The source of priority contaminants for the drinking water program, EPA 815-F-05–001. Office of Water, United States Environmental Protection Agency, Washington, DC.
Vrana
B.
Mills
G. A.
Allan
I. J.
Dominiak
E.
Svensson
K.
Knutsson
J.
Morrison
G.
Greenwood
R.
2005
Passive sampling techniques for monitoring pollutants in water
.
Trends Anal. Chem.
24
,
845
848
.
WHO
1998
Guidelines for drinking water quality
.
World Health Organization
,
Geneva
.
WHO
1999
Toxic Cyanobacteria in Water: A guide to their public health consequences, monitoring and management
.
Routledge
,
London and New York
.