Development of a passive sampling device for hydrophilic organic pesticides in aquatic environments

85 Figure 4-1: The EmporeTM disk-based passive sampling device a before assembly and b after assembly: 1 top ring, 2 middle ring and metal grid, 3 receiving phase, 4 diffusion membrane a

Development of a passive sampling device for hydrophilic organic pesticides in aquatic environments

Tran Thi Kieu Anh

Doctor of Philosophy

Department of Environmental Science

and Department of Chemistry, Materials and Forensic Science

December 2006 Unviversity of Technology, Sydney

I certify that the work in this thesis has not previously been submitted for a degree, nor has it been submitted as part of requirements for a degree except as fully acknowledged within the text

I also certify that the thesis has been written by me Any help that I have received in

my research work and the preparation of the thesis itself has been acknowledged In addition, I certify that all information sources and literature used are indicated in this thesis

Signature of Student

Tran Thi Kieu Anh

I would like to thank my Principal Supervisors, Dr Ross V Hyne and Dr Philip Doble, for your patience, guidance and enthusiasm With your help, I have gained experience in so many different areas of research

The project would not have been possible without the support of a three-year grant from The Ministry of Education and Training of Vietnam (project no 322)

Many thanks to all the staff of the Department of Environment and Conservation (DEC, NSW, Australia) and the staff and students in the Department of Chemistry, Materials and Forensic Science, particularly Roy Day, Fleur Pablo, Moreno Julli, Melissa Aistrope, Yin Latt Phyu, Hemantha Dassanayake, Reinier Mann and Catherine Choung for your kind assistance and providing the facilities for my research

I would like to thank Dr Fleur Pablo, Dr Ron Patra, Melissa Aistrope and Paul Rendell for your kind help with the field trip and Alex Oglobline (DEC, NSW, Australia) for confirmation of pesticides by HPLC-MS A special thank you to Sunderam Ramasamy for helping me with computer and software issues Valuable comments on my submitted manuscripts by anonymous referees are much appreciated

Finally, and most importancely, I would like to thank my family and my friends for supporting me during my study

Type of Publication Number Reference

[1] Tran, A T K., Hyne, R V., Pablo, F., Day, W R and Doble, P., 2006

Optimisation of the separation of herbicides by linear gradient high performance

liquid chromatography utilising Artificial Neural Networks Talanta in press

(Chapter 2)

[2] Tran, A T K., Hyne, R V and Doble, P., 2006 Determination of commonly

used polar herbicides in agricultural drainage waters in Australia by HPLC

Chemosphere in press (Chapter 3)

[3] Tran, A T K., Hyne, R V and Doble, P., 2006 Calibration of a passive

sampling device for time-integrated sampling of hydrophilic herbicides in aquatic

environments Environmental Toxicology and Chemistry in press (Chapter 4)

[4] Tran, A T K., Hyne, R V and Doble, P., 2006 Determination of

time-integrated polar herbicide concentrations using field-deployed EmporeTM disk-based

passive samplers: comparison with daily water extractions Environmental Science

& Technology submitted (Chapter 5)

[5] Tran, A T K, Pablo, F., Hyne, R V., Day, R W and Doble, P., 2004

Optimisation of the separation of herbicides by linear gradient high performance

liquid chromatography utilising Artificial Neural Networks Interact 2004,

Queensland, Australia, July (Chapter 2)

the aquatic environment Australian Society for Ecotoxicology 2005 Conference,

Melbourne, Australia, September (Chapter 4)

Certificate of Authorship/Originality ii

Acknowledgements iii

List of Publications iv

Table of Contents vi

List of Abbreviations xi

List of Figures xv

List of Tables xviii

Abstract xx

Chapter 1 General introduction 1

1.1 Pesticide use in Australia 1

1.2 Contamination of pesticides in Australian waters 3

1.3 Analysis of pesticides in water 9

1.4 Passive samplers 10

1.4.1 Principles of passive sampling 10

1.4.2 Theory of passive sampling method and calculation of TWA 11

1.4.3 Passive sampling devices 13

1.5 The choice of target pesticides 17

1.5.1 Pesticide physicochemical properties 17

1.5.2 Herbicide toxicities 23

1.5.3 Guideline values for target herbicides in the aquatic environment 24

1.6 Routine analysis methods for herbicides in water 26

1.6.1 Instrumental methods 26

1.6.2 Sample preconcentration methods 29

Chapter 2 Development of a herbicide instrumental analysis method 52

2.1 Summary 52

2.2 Introduction 52

2.3 Materials and methods 54

2.3.1 Instrumentation 54

2.3.2 Reagents and procedures 55

2.3.3 Experimental design 55

2.3.4 Artificial neural network model of response surface 58

2.3.5 Calculation of resolution (R S) 59

2.3.6 Analytical performance and limits of detection 59

2.4 Results and discussion 60

2.4.1 Optimisation of the separation using an ANN 60

2.4.2 Analytical performance and limits of detection 66

2.5 Conclusion 66

Chapter 3 Development of herbicide preconcentration methods for different water matrices 67

3.1 Summary 67

3.2 Introduction 67

3.3 Materials and methods 69

3.3.1 Materials 69

3.3.2 Preconcentration apparatus and instruments 69

3.3.3 Characteristics of water samples 70

3.3.4 LLE preconcentration 71

3.3.5 SPE preconcentration using Oasis® HLB cartridges 72

3.3.8 Analytical performance and detection limits 76

3.3.9 Statistical analysis 76

3.4 Results and discussion 76

3.4.1 LLE preconcentration 76

3.4.2 SPE preconcentration using Oasis® HLB cartridges 78

3.4.3 SPE preconcentraton using SDB-XC EmporeTM disks 79

3.4.4 Separation by HPLC 81

3.4.5 Analytical performance and the limits of detection 86

3.5 Conclusion 88

Chapter 4 Development of an EmporeTM disk-based passive sampling device 89 4.1 Summary 89

4.2 Introduction 90

4.3 Materials and methods 94

4.3.1 Materials 94

4.3.2 Preconcentration apparatus and instruments 95

4.3.3 Passive sampler holder design 95

4.3.4 Preparation of EmporeTM disks 96

4.3.5 Preparation of diffusion membranes 97

4.3.6 Assembly and deployment of the passive samplers 97

4.3.7 Design of laboratory flow-through system 97

4.3.8 Herbicide uptake experiments 98

4.3.9 Release experiments 99

4.3.10 Extraction of analytes from water 99

4.3.11 Extraction of analytes from the device 100

4.3.14 Analytical performance and limits of detection 101

4.3.15 Statistical analysis and data-processing 101

4.4 Results and discussion 101

4.4.1 Selected herbicides 101

4.4.2 Evaluation of receiving phases 102

4.4.3 Conditioning and extraction of EmporeTM disks and diffusion membranes 103

4.4.4 Evaluation of diffusion membranes 104

4.4.5 Uptake kinetics of non-acidic target herbicides 108

4.4.6 Uptake and release kinetics of phenoxy acid herbicides 113

4.5 Conclusions 117

Chapter 5 Field validation of the passive sampling device 118

5.1 Summary 118

5.2 Introduction 118

5.3 Materials and methods 121

5.3.1 Materials 121

5.3.2 Preconcentration apparatus and instruments 121

5.3.3 Field-study sites 121

5.3.4 Preparation of the passive samplers 122

5.3.5 Performance reference compound release experiments 123

5.3.6 Field deployment of the passive samplers 124

5.3.7 Daily water collection 125

5.3.8 Daily water extraction 125

5.3.9 Extraction of analytes from the passive sampling devices 126

5.3.12 Herbicide confirmation 127

5.3.13 Concentration factors for the passive sampling devices 127

5.3.14 Statistical analysis 127

5.4 Results and discussion 128

5.4.1 Field study – water quality 128

5.4.2 Appearance of the passive sampler following deployment 129

5.4.3 Daily water extraction 129

5.4.4 Herbicide accumulation by passive samplers – effect of flow velocity 132

5.4.5 Release of monolinuron from passive samplers – field versus laboratory 135

5.4.6 Calculation of time-weighted average herbicide concentrations 138

5.4.7 Herbicides in water – comparison of daily water extractions versus passive samplers 139

5.5 Conclusion 146

Chapter 6 Overall conclusions and future work 147

6.1 Overall conclusions 147

6.2 Future work 149

Chapter 7 References 151

Acronyms/Abbreviations Meanings/Full text

AATSE Australia Academy of Technological Sciences and

Engineering ACIAR Australian Centre for International Agricultural

Research ACN Acetonitrile

ANZECC Australia and New Zealand Environment and

Conservation Council APCI Atmospheric Pressure Chemical Ionisation

ARMCANZ Agriculture and Resource Management Council of

Australia and New Zealand

C The analyte concentration in the aqueous environment

DDT Dichlorodiphenyltrichloroethane DEC Department of Environment and Conservation

EPA Evironmental Protection Agency

HPLC High-performance Liquid Chromatography

LC50 Concentration required to kill 50% of test organisms

LD50 Dose required to kill 50% of test organisms

LOD Limit of detection or detection limit

Log K ow Log octanol/water partition coefficient

LWRRDC Land and Water Resources Research and Development

Corporation

MeOH Methanol

NH4COOH Ammonium formate

NMRC National Medical Research Council

NRMMC National Resource Management Ministerial Council

PISCES Passive In Situ Concentration Extraction Sampler

pK a -Log10 acid dissociation constant

pKb -Log10 base dissociation constant

POCIS Polar Organic Chemical Integrative Sampler

SDB Styrenedivinylbenzene

SPMD Semipermeable membrane device

SRMM Stacking with Reverse-migrating Micelles

U S EPA U S Environmental Protection Agency

1

W and W 2 Peak widths

Figure 1-1: Pesticide usage in Australia (1978-1998) Data from AATSE (2002) 2 Figure 1-2: Analysis of pesticides in water 10 Figure 1-3: Schematic representation of a passive accumulation device Adapted from Stuer-Lauridsen (2005) 11 Figure 1-4: Time-dependent concentration profile of organic pollutant in passive sampling device Adapted from Stuer-Lauridsen (2005) 12 Figure 2-1: Chromatograms of the herbicides under each of the experimental conditions 61 Figure 2-2: Plot of observed retention times vs ANN predicted retention times 63 Figure 2-3: The minimum peak pair resolution response surface predicted by the ANN 64 Figure 2-4: Separation of a mixed herbicide standard at level 5-10 µg/L 65 Figure 3-1: SPE preconcentration procedure 75 Figure 3-2: HPLC chromatograph of (a) a natural blank water sample (500 mL) collected from the South Creek site after SPE preconcentration with HLB Oasis® cartridge and (b) HPLC of a natural water sample (500 mL) collected from the South Creek site spiked with mixed herbicide standards to give final concentrations of 0.5-1 µg/L after SPE preconcentration with HLB Oasis®

cartridge: (2) simazine, (3) 2,4-D, (4) MCPA, (5) triclopyr, (6) atrazine, (7) diuron, (8) clomazone (9) bensulfuron-methyl and (10) metolachlor 83 Figure 3-3: HPLC chromatograph of (a) an agricultural drainage blank water sample (1 L) collected from the Leeton area (site 1) after SPE preconcentration with SDB-XC EmporeTM disk and (b) HPLC of an agricultural drainage water sample (1 L) collected from the Leeton area (site 1) spiked with mixed herbicide standards to give final concentrations of 1-2.4 µg/L after SPE preconcentration with SDB-XC EmporeTM disk: (1) dicamba, (2) simazine, (3)

Figure 3-4: HPLC chromatograph of (a) an agricultural drainage blank water sample collected from the Leeton area (site 1) (3 L) after LLE preconcentration and (b)

an agricultural drainage blank water sample (3 L) collected from the Leeton area (site 1) spiked with mixed herbicide standards to give final concentrations

of 1-2.4 µg/L after LLE preconcentration: (1) dicamba, (2) simazine, (3) 2,4-D, (4) MCPA, (5) triclopyr, (6) atrazine, (7) diuron, (8) clomazone (9) bensulfuron-methyl and (10) metolachlor 85 Figure 4-1: The EmporeTM disk-based passive sampling device (a) before assembly and (b) after assembly: (1) top ring, (2) middle ring and metal grid, (3) receiving phase, (4) diffusion membrane and (5) bottom ring 96

Figure 4-2: Mean accumulation (n = 2) of the study herbicides on the SDB-XC

EmporeTM disks of passive sampling devices fitted with either a polyethersulfone or a polysulfone as the diffusion membrane The samplers were exposed for six days at high herbicide concentrations (Table 4-3) at 23 oC

in the continuous-flow system Error bars shown are standard errors of mean 105 Figure 4-3: Regressions between days of exposure and mean accumulation of simazine measured in the passive sampling devices The samplers were exposed for 21 days at the two highest herbicide concentrations used (Table 4-3) in the continuous-flow system Error bars shown are standard errors of

mean ( n = 2) 109

Figure 4-4: Regressions between days of exposure and concentration factors of the study herbicides measured in the passive sampling devices fitted with a SDB-

XC EmporeTM disk as the receiving phase and a polyethersulfone membrane as

a diffusion membrane The values shown are the mean concentration factors for the device (both the receiving phase and the diffusion membrane) from different water concentrations (Table 4-3) for a 21-day exposure period under the flow-through system Error bars shown are standard errors of mean 111

retention of triclopyr in the passive sampling devices during a 21-day deployment period in the flow-through system The values shown are mean

percentage retention of triclopyr (n =2) (Figure 4-5(a)) or natural logarithms of the percentage retention of triclopy (n =2) (Figure 4-5(b)) in the passive

sampling devices Error bars shown are standard errors of mean 116 Figure 5-1: Daily concentrations of study herbicides in 24-hour composite water samples collected from (a) site 1 and (b) from site 2 during a 13-day monitoring period 131 Figure 5-2: Regression between days of exposure and percentage retention of monolinuron in the passive samplers during a 49-day deployment period in the laboratory continuous-flow system The values shown are mean and standard

errors of mean (n = 3) 136

Figure 5-3: Regression between days of exposure and percentage retention of monolinuron in the passive samplers (a) during a 13-day deployment period at site 1 and (b) during a 13-day deployment period at site 2 The values shown

are mean and standard errors of mean (n = 2) 137

Figure 5-4: Comparison of herbicide concentrations in agricultural drainage channels derived from passive sampler concentrations for each deployment period with the corresponding cumulative mean concentrations calculated from daily extractions of the drainage water for each period The passive samplers were collected every third day from day 7 to day 13: B = bank-channel position and M = middle-channel position at site 2 and numeric code site (1 or 2) – deployment period (7, 10 or 13 days) 141 Figure 5-5: Chromatogram of (a) agricultural water (1 L) sampled at site 1 at day 6 after SPE preconcentration and (b) agricultural water at site 1 after a 13-day deployment period by the EmporeTM disk-based passive sampler: (1) simazine, (2) atrazine and (3) diuron 145

Table 1-1: Herbicides in surface waters from irrigation areas in Australia 4

Table 1-2: Herbicides in groundwater of irrigation areas in Australia 7

Table 1-3: Structures and physicochemical properties of the target herbicides 19

Table 1-4: Estimated herbicide usage in the MIA (kg active ingredient (ai) /year) 23

Table 1-5: Toxicities of herbicides for some aquatic organisms 24

Table 1-6: Guideline values for the herbicides 25

Table 1-7: Analysis methods for target herbicides 34

Table 2-1: Experimental conditions 57

Table 2-2: Retention times (min) for nine analytes at each of the experimental points 62

Table 2-3: Linearity and the detection limits of the HPLC-PDA 66

Table 3-1: Physicochemical properties of water samples 71

Table 3-2: Percentage mean recoveries of spiked herbicides in different matrices by LLE extraction 78

Table 3-3: Percentage mean recoveries of spiked herbicides in different matrices by SPE using Oasis® HLB cartridges 79

Table 3-4: Percentage mean recoveries of spiked herbicides in agricultural water samples collected from the Leeton sites by SPE using SDB-XC EmporeTM disks 80

Table 3-5: Linearity and detection limits of the HPLC – UV 86

Table 3-6: Linearity and detection limits after SPE preconcentration 87

Table 4-1: Receiving phases and diffusion membranes 95

Table 4-2: Mean percentage recovery of the study herbicides from the EmporeTM disk receiving phases 103

Table 4-4: Analysis of variance in concentration factors of the study herbicides on the SDB-XC EmporeTM disks of passive sampling devices fitted with either a polyethersulfone or a polysulfone diffusion membrane 107 Table 4-5: Analysis of variance of the non-acidic herbicide uptake model fitted through origin (Figure 4-4) 112 Table 5-1: Structure and physicochemical properties of monolinuron 121 Table 5-2: Range of water physicochemical parameters at the deployment locations and release half-life of monolinuron (PRC) from the passive sampling devices 129 Table 5-3: Accumulated herbicides in the SDB-XC EmporeTM disks and polyethersulfone membranes of the passive samplers following deployment at two agricultural sites 133 Table 5-4: Analysis of variance in the mean mass of each herbicide accumulated in the passive samplers at the bank and the middle position of the drainage channel at site 2 135 Table 5-5: Sampling rates (R ) for each herbicide at each deployment site 138 S

The experimental aspects of this thesis deal with the development of a passive sampling device for measuring hydrophilic organic herbicides in the aquatic environment The target herbicides included five non-acidic herbicides (simazine, atrazine, diuron, clomazone, and metolachlor) and five acidic herbicides (dicamba, (2,4-dichlorophenoxy)acetic acid [2,4-D], (4-chloro-2-methylphenoxy)acetic acid [MCPA], triclopyr and bensulfuron-methyl) with log octanol/water partition

coefficient (log Kow) values less than three in water

Chapter 1 of the thesis presents a literature review This includes information on the

usage of the target hydrophilic herbicides in Australia and details on the routine analysis methods for determination of these herbicides in aqueous environments

Chapter 2 focuses on the optimisation of the separation of the ten target herbicides

by high-performance liquid chromatography (HPLC) with a photo diode array (PDA) or an ultra-violet (UV) detector An artificial neural network (ANN) was employed to model the chromatographic response surface for the linear gradient separation of the ten target herbicides The ANN was trained using the pH of the mobile phase and the slope of the acetonitrile/water gradient as input variables

Nine experiments were required to generate sufficient data to train the ANN to accurately describe the retention times of each of the herbicides within a defined experimental space of mobile-phase pH range 3.0 to 4.8 and linear gradient slope 1

to 4% acetonitrile/min The modelled chromatographic response surface was then used to determine the optimum separation within the experimental space This approach allowed the rapid determination of experimental conditions for baseline resolution of all ten herbicides

Chapter 3 focuses on the development of sample preconcentration methods for the

target compounds in the aquatic environment Liquid-liquid extraction (LLE) with

efficiency of these herbicides in different water matrices, especially water samples from contaminated agricultural drainage water containing high concentrations of particulate matter Herbicides were then separated and quantified by HPLC-UV

Chapter 4 presents the results of experiments carried out to develop an EmporeTM

disk-based passive sampling device for the target hydrophilic herbicides The herbicide uptake or sampling rate of passive sampling devices depended on the configuration of the passive sampler, the properties of the receiving phase, the overlying diffusion membrane and the properties of the compound being sampled

Two types of EmporeTM solid-phase materials, a styrenedivinylbenzene copolymer sorbent (embedded in a SDB-XC EmporeTM disk), and a styrenedivinylbenzene copolymer sorbent that was modified with sulfonic-acid functional groups (embedded in a SDB-RPS EmporeTM disk) were compared as a receiving phase in a device for uptake of the target herbicides The SDB-XC EmporeTM disk proved to be the superior receiving phase and was then further evaluated with either a polysulfone

or polyethersulfone diffusion-limiting membrane

Uptake of the target herbicides was generally higher into a device constructed of a SDB-XC EmporeTM disk covered with a polyethersulfone membrane This device showed linear uptake of the non-acidic (simazine, atrazine, diuron, clomazone, and metolachlor) for up to 21 days Daily sampling rates of the herbicides from water in

a laboratory flow-through system were determined Uptake of the phenoxy acid herbicides (2,4-D, MCPA, and triclopyr) obeyed first-order kinetics and rapidly reached equilibrium in the passive sampler after approximately 12 days of exposure The EmporeTM disk-based passive sampler displayed isotropic kinetics, with a release half-life for triclopyr of approximately six days

Chapter 5 details the evaluation of the developed device for monitoring the

non-acidic target herbicides in the field The herbcide uptake depended on the configuration of the passive sampler, the properties of the receiving phase, the

water turbulence Calculations of time-weighted average water concentrations of the herbicides during the deployment period were based on the sampling rates of the passive sampling devices determined in the laboratory The field sampling rates were adjusted by the application of a correction factor determined from the release

of monolinuron as a performance reference compound (PRC)

The field study was undertaken in the Murrumbidgee Irrigation Area (MIA) (Leeton, NSW, Australia) Daily water extractions of 24-hour composite water samples collected at two sites were performed by SPE using SDB-XC EmporeTM disks with Filter Aid 400 EmporeTM disk-based passive samplers, consisting of a SDB-XC EmporeTM disk as a receiving phase and a polyethersulfone membrane as a diffusion membrane, were deployed at the same sites

Over a two-week study period, the daily aqueous concentrations of herbicides from 24-hour composite samples detected at two sites increased after run-off from a storm event and were in the range of: 0.1 to 17.8 µg/L, < 0.1 to 0.9 µg/L and 0.2 to 17.8 µg/L at site 1; < 0.1 to 3.5 µg/L, < 0.1 to 0.2 µg/Land < 0.2 to 3.2 µg/Lat site 2 for simazine, atrazine and diuron, respectively The release half-life for monolinuron from the device in the laboratory was 106 days compared to 34 days at site 1 and 31 days at site 2 in the field The differences in the release rates of monolinuron were used to compensate for variations in the environmental conditions The time-weighted average (TWA) herbicide concentrations determined from the recovered passive sampling devices after 7, 10 or 13 days were generally within twofold of the mean of the daily drainage-water concentrations of simazine, atrazine and diuron for each corresponding period

1.1 Pesticide use in Australia

Pesticides are chemicals which can be natural or synthetic substances, are usually toxic, and are used to protect crops from insects, weeds or diseases, in both dry land and irrigated agriculture (AATSE, 2002; Cabras, 2003; ACIAR, 2005) The principal categories of pesticides are insecticides, herbicides, fungicides and growth regulators (AATSE, 2002)

The first organic pesticide introduced to the market by Geigy in 1939 was dichlorodiphenyltrichloroethane (DDT), as a result of systematic research on its insect killing activity by the Swiss entomologist Paul Muller (Cabras, 2003) Before being used in agriculture, DDT was applied extensively against the carriers of diseases during and after the Second World War DDT is the historic predecessor of synthetic pesticides and organochlorine compounds It was followed in fast succession by other molecules belonging to the same chemical family, such as lindane (1942), aldrin (1948), dieldrin (1949) and endrin (1951) (Cabras, 2003)

DDT and organochlorine compounds are harmful both to agriculture and to human health Their high persistence, their lipophilicity and their tendency to accumulate in the food chain directly or indirectly affected the fertility and reproduction of many wild species For this reason, since the mid-1980s, the use of DDT and organochlorine compounds has been banned in agriculture in all countries of the world (Cabras, 2003) In recent times, pesticide use is trending towards polar pesticides such as N-methylcarbamates, triazines, chlorophenoxy acids and so-called modern herbicides, for example, sulfonylureas and imidazolinones, with favourable properties such as a low dose rate of application and a high degree of biodegradation (Hogendoorn and van Zoonen, 2000; Cabras, 2003)

Chemical pesticides have been significantly applied globally in agriculture and their use has increased substantially since the late 1940s The estimated United States production for 1987 was 748 million pounds for herbicides, compared to 304 million and 100 million pounds for insecticides and regulators, respectively (Stevens and Sumner, 1991) Herbicides have the largest market share In past years, herbicide use in United States has increased significantly from 34.8% in 1970 to 51.9% in

2001 (Cabras, 2003) In Australia, herbicides are also the largest group of pesticide products with annual sales in 1998 and 1999 of just over $800 million (Figure 1-1) (AATSE, 2002) They have been used to kill weeds in broadacre cropping (wheat, oats, barley, sorghum, canola, peanuts), cotton, sugar cane, horticulture (fruit, vegetables) and other farm crops

Figure 1-1: Pesticide usage in Australia (1978-1998) Data from AATSE (2002)

Over the four decades 1950-1990, the value of Australian farm output increased by

250 per cent Moreover, the productivity growth rate achieved in Australian agriculture over this period was substantially higher than that achieved in the rest of the Australian economy and in the agricultural sectors of other developed countries taken as a whole (AATSE, 2002) Despite the world-wide health benefits arising from the use of pesticides, ensuring an improved and relatively stable food supply, there continue to be concerns regarding aspects of their use There are concerns regarding exposure to workers or families in agricultural areas due to spray drift, exposure of aquatic organisms to contamination or contamination of food by

0 100

residues from direct pesticide application or from environmental exposure Polar pesticides can be more hazardous when, through leaching or other processes, they move away from the farm where they were applied and contaminate the adjacent aquatic system

1.2 Contamination of pesticides in Australian waters

Pesticides contaminate waterways generally through spray drift, runoff, direct overspray and leaching Polar herbicides such as atrazine, diuron and simazine are often detected in surface waters of irrigation areas in Australia (Table 1-1) Atrazine

or triazine herbicides have a low ability in binding to soils and are therefore relatively mobile They have often been found in groundwater in rural regions of Australia (Table 1-2) Atrazine has been described as one of the most widely used herbicides in Australian agriculture, with high potential to contaminate ground and surface water, and with narrow safety margins for aquatic organisms Simazine may also occur in groundwater but is not as mobile as atrazine (AATSE, 2002) Concentrations of these herbicides have been found to be low in groundwater or drinking water, higher in surface waters, and sometimes very high in runoff water from agricultural areas

Australian surface waters

Herbicides Environmental limits Location/Time Notes

Cotton-growing areas in western Australia

Atrazine < LOD a to 10 µg/L, some

isolated peaks up to 20 µg/Lduring storm events

1999-2000 Exceeded the guideline level

for 99% ecosystem protection (0.07 µg/L) in 8% of samples Diuron 0.2 to 3 µg/L Murray-Darling Basin

(1997-1998)

Exceeded the irrigation water guidelines (2 µg/L: ANZECC b

1992) for 15% of samples Atrazine Up to 2.25 µg/L Cox’s Creek in the

Liverpool Plains of NSW,

1995

High concentration detected during a 24-hour storm event

Diuron Up to 24 µg/L The Gwydir River, 1997 High concentration detected

during a storm event Irrigation areas in

south-western NSW

Atrazine and 2,4-D

0.08 and 0.2 µg/L for atrazine; 0.5 µg/L for 2,4 D

Supply water Detected in few channels, data

Australian surface waters

Herbicides Environmental limits Location/Time Notes

Atrazine 3.6 µg/L Supply water mixed with

MIA c drainage water

Detected in few channels, data from a report in 1995

Victorian waterways Atrazine,

simazine and diuron

Up to 4.9 µg/L for atrazine and 1.4 µg/L for simazine and 4.8 µg l-1 for diuron

Surface water used in vegetable production (in Gippsland area), 1994

Exceeded the ANZECC 1992

at only two of 21 sites

simazine

0.14 to 3.2 µg/L for atrazine and 0.61 µg l-1 for simazine

Rosebud and Bairnsdale areas

Found in a few stream samples

Tasmanian waterways

Atrazine A median of µg/L on the day

of spraying to a median of 0.3 µg/L around 13-15 months later

Atrazine From trace levels up to 2 400

µg/L

The Condamine-Balonne catchment

Data from a report in 1996

Australian surface waters

Herbicides Environmental limits Location/Time Notes

Atrazine 0.4 to 14.4 µg/L The Johnstone and

Daintree rivers

Data from a report in 1996

Western Australian waterways

Atrazine 0.8 to 38 µg/L Data from a report in 1990

Data from AATSE (2002) Abbreviations: a LOD: Limit of detection, b ANZECC: Australian and New Zealand Environment and Conservation Council, c MIA: Murrumbidgee Irrigation Area

Australian groundwater

Herbicides Environment levels Location/Time Notes

New South Wales

Atrazine and diuron Up to 0.3 µg/Lfor

atrazine and < 2 µg/L for diuron

The Wakool-Cadell, 1995 and Denimein-Berriquin, 1996

Simazine 0.01 µg/L The Alstonville Plateau, 1999 Detected in only one bore Atrazine 0.1 to 5.8 µg/L The lower Namoi Valley Data from a report in 1996 Victoria Atrazine Up to 0.04 µg/L for

atrazine and up to 0.45 µg/L

The Cobram area, 1993

Queensland Atrazine < 0.1 µg/L but one site

has between 1.3 and 1.4 µg/L

The Lower Burdekin basin, 1992 and 1993

Atrazine ≤ 0.1 µg/L The Border Rivers catchments of

southern Queensland and northern NSW, 1994 and 1995

South Australia Atrazine 0.06 and 0.65 µg/L The southern Mount Lofty Ranges,

1994 and 1995 Atrazine detected in only two out of 129 bores

Governments have responded by establishing agencies and regulatory authorities to ensure that human health is not compromised due to exposure to pesticides Increasingly restrictive regulations have been enacted for the banning of some dangerous pesticides and lowering the maximum admissible concentrations of pesticides in drinking water, surface waters, air, soils or foodstuffs

European regulations on drinking-water quality set a maximum concentration of 0.1 µg/L for individual pesticides and some of their degradation products, and 0.5 µg/L for total pesticides present in a sample (Di Corcia et al., 2000; Jeannot et al., 2000);

or in the 0.1 µg/L range for drinking-water samples and closer to 0.5 µg/L for surface water samples (Pichon et al., 1996)

The Australian water-quality guidelines for fresh and marine waters were released in

1992 and were revised in 2000 (ANZECC and ARMCANZ, 2000) These guidelines were published jointly by the Australian and New Zealand Environment and Conservation Council (ANZECC) and the Agricultural and Resource Management Council of Australia and New Zealand (ARMCANZ) The ANZECC has developed

a series of risk-based decision frameworks to refine the guidelines to suit local environment conditions

These frameworks are based on guideline “trigger values” which are concentrations

of a chemical or nutrient that have the potential to cause a problem if exceeded The guidelines recommend up to six ecosystem types (upland rivers and streams, lowland rivers, freshwater lakes and reservoirs, wetlands, and coastal and marine estuaries) and three types of ecosystem conditions with their own level of protection: high conservation/ecological value systems; slightly to moderately disturbed systems; and highly disturbed systems (ANZECC and ARMCANZ, 2000) Since the formulation of these guidelines, water samples have been regularly analysed to monitor water quality against the ANZECC trigger values (ANZECC and ARMCANZ, 2000)

1.3 Analysis of pesticides in water

The most common method for monitoring pesticides in the aquatic environment is grab or spot sampling of the water body followed by a laboratory-based analysis of the water Gas chromatography (GC) and high-performance liquid chromatography (HPLC) are the main instrumental techniques for the determination of herbicides in water Due to the extremely low environmental levels (ppb or µg/L) of pesticides in water, these techniques require extensive sample preparation prior to analysis, usually liquid-liquid extraction (LLE) and/or solid-phase extraction (SPE) (Figure 1-2)

Sampling and sample preparation are the most expensive and time-consuming steps According to various estimates, these steps typically account for 70-90% of analysis time (Górecki and Namieśnik, 2002)

The most important step of any analysis procedure is correct and representative sampling Errors committed at this stage cannot be corrected later during the analysis In environments where the contaminant concentrations may vary over time,

it is often desirable to expand the monitoring time and increase the number of samples This approach adds to costs and can miss transient events of herbicide release Passive samplers can overcome these difficulties by combining sampling, analyte isolation, and preconcentration into a single step Passive samplers also allow time-integrated sampling of compounds during deployment (Vrana et al., 2005a) Their low cost of deployment makes them suitable for pesticide monitoring programmes Moreover, a passive sampling device for chemicals is an object that collects chemical compounds without provision of energy from an external source (Stuer-Lauridsen, 2005) Therefore, passive sampling can be applied in remote areas when lack of an energy source is an important issue

Figure 1-2: Analysis of pesticides in water

Samples of sediment or biota are the usual choice to represent time-integrated waterborne contamination in the long term (years) and short term (weeks to months), respectively (Burton, 1992; Verweij et al., 2004) There are, however, drawbacks attached to the evaluation of aquatic quality based on sediment and biota data for organic micropollutants It can be very difficult or impossible to assess the influence

of sediment bioturbation and resuspension events, sediment sorbent quality, degradation and elimination rates, or the actual and recent conditions of the sampled biota, all of which may have substantial influence on the observed concentrations (Kot et al., 2000)

1.4 Passive samplers

1.4.1 Principles of passive sampling

Passive sampling is based on the free flow of analyte molecules from the sampled medium to a collecting medium as a result of a difference in chemical potentials (Vrana et al., 2005a) Passive samplers continuously take up organic compounds that are potentially bioavailable, generally through a diffusion membrane into a receiving phase (Figure 1-3) The net flow of analyte molecules from one medium to the other continues until equilibrium is established in the system, or until the sampling session

is terminated by the user (Górecki and Namieśnik, 2002) When sampling continues

Sampling(collect water samples)

Sample preparation(preconcentration, clean-up)

Sample analysis(GC or HPLC)

Sampling(collect water samples)

Sample preparation(preconcentration, clean-up)

Sample analysis(GC or HPLC)

until the sampling session is terminated by the user, the amount of analytes collected

by the sampler depends on both its concentration in the sampled medium and the exposure time If the relationship between the sampling rate and analyte concentration is known, time-weighted average (TWA) analyte concentration can be easily determined Analytes in samplers can be determined by extractions of analytes from the samplers and analysed by GC or HPLC Sampling rates can be determined under laboratory conditions

Figure 1-3: Schematic representation of a passive accumulation device Adapted from

Stuer-Lauridsen (2005)

1.4.2 Theory of passive sampling method and calculation of TWA

Accumulation of pollutants or pesticides from water into the passive sampling device generally follows the pattern shown in Figure 1-4

Receiving phase (solvent

or sorbent)

or sorbent)

Diffusion membrane

Housing

AnalytesWater

Figure 1-4: Time-dependent concentration profile of organic pollutant in passive sampling

device Adapted from Stuer-Lauridsen (2005)

The exchange kinetics between a passive sampler and water phase can be described

by a first-order, one-compartment mathematical model:

)1()

where )C S (t is the concentration of the analyte in the sampler at exposure time t,

W

C is the analyte concentration in the aqueous environment, and k and 1 k are the 2

uptake and offload rate constants, respectively

In equilibrium sampling, the exposure time is sufficiently long to permit the establishment of thermodynamic equilibrium between the water and the device In this situation, Equation (1-1) reduces to:

K C k

k C t

t k C t

C S()= W 1 (Equation 1-3)

Equation (1-3) can be rearranged to an equivalent relationship:

t R C t

When R is known, S C (TWA concentration of a pollutant in the water phase over W

the exposure period) may be calculated from the sampling rate R , exposure time S

)

(t and the amount (M S (t)) of the analyte accumulated by the device

For most devices operating in the kinetic mode, sampling rates are dependent on both intrinsic and extrinsic factors Intrinsic factors include physicochemical properties of analytes and sampler designs Extrinsic factors include water flow or turbulence, temperature and biofouling Therefore, passive sampling devices have been usually calibrated in the laboratory at known exposure concentrations and conditions to calculate sampling rates Extensive calibration studies may be necessary to examine the effects of extrinsic factors on sampling rates prior to application of the device in the field

1.4.3 Passive sampling devices

Passive samplers for monitoring pollutants in water have been described in many review papers (Kot et al., 2000; Górecki and Namieśnik, 2002; Lu et al., 2002; Stuer-Lauridsen, 2005; Vrana et al., 2005a) The following sub-sections detail several samplers which can be used practically for monitoring organic contaminants

in the aquatic environment

1.4.3.1 Passive samplers filled with solvent

The first passive sampling method for the aquatic environment was developed by Södergren (1987) It was a simple device consisting of a tube of dialysis membrane (regenerated cellulose) with 3 ml of organic solvent inside, typically hexane

Hydrophobic organic pollutants diffused through the membrane from the water and accumulated in the organic phase The method is based on the partitioning process similar to the bioconcentration of hydrophobic organic contaminants in fish and invertebrates Because hexane is used as the receiving medium, laboratory analysis

of the sample is more rapid and less expensive than conventional water, sediment or tissue analysis (Sabaliūnas and Södergren, 1996) However, hexane is moderately soluble in water and tended to migrate into the surrounding water These devices have not been extensively or routinely deployed and are now characterised as the prototype of the more recent developments (Stuer-Lauridsen, 2005)

1.4.3.2 Semipermeable membrane devices (SPMDs)

The SPMD was first published in 1990 (Huckins et al., 1990) and a substantial number of studies have followed (>100) (Stuer-Lauridsen, 2005) The SPMD is a low-density polyethylene tube (lay flat style) filled with approximately 1 ml of triolein and sealed at both ends The hydrophobic compounds such as polychlorinated biphenyls (PCBs) or polycyclic aromatic hydrocarbons (PAHs) with

log octanol/water partition coefficient (log K ow) values greater than four in water diffused through the membrane and accumulated in the triolein Compounds that accumulated in the SPMD were then extracted from the triolein receiving medium and determined by GC after sample clean-up SPMD has been reviewed for background and environmental applications, analytical chemical issues and quality control (Petty et al., 2000; Lu et al., 2002) A guidance document has been developed detailing the procedures in the field and laboratory (Huckins et al., 2000) The SPMD has also been used for sampling hydrophobic compounds in air (Petty et al., 1993; Müller et al., 1998; Bartkow et al., 2004; Bartkow et al., 2005) and sediment (Rantalainen et al., 1998; Rantalainen et al., 2000; Williamson et al., 2002; Verweij et al., 2004)

1.4.3.3 Trimethylpentane passive sampler (TRIMP)

Zabik et al (1992) first tested polymeric tubing with tri-methyl pentane (iso-octane) for the accumulation of pesticides in soil They tested several receiving phases and membranes and found the combination of polyethylene and trimethylpentane superior for pesticides Peterson et al (1995) tested a similar polyethylene-trimethylpentene device for the uptake of organochlorine pesticides in aquatic environments It was shown that the sampling of chlorinated pesticides was linear and proportional to the concentration in the surroundings for at least two to sevent weeks The method has been further developed and applied for monitoring hydrophobic pesticides in river water ( Leonard et al., 1999; Leonard et al., 2002; Hyne et al., 2004) Loss of the solvent from the devices was reduced by deploying the device in a mesh bag The pesticides from the device can be analysed directly by

GC without extraction procedures ( Leonard et al., 1999; Leonard et al., 2002; Hyne

et al., 2004)

1.4.3.4 The Empore TM disk-based passive sampler

EmporeTM disks are filters coated with standard absorption materials and have high surface-to-volume rates (Thurman and Mills, 1998) They are commercially available and conventionally used for solid-phase extraction of organic contaminants

in water samplers (Ruberu et al., 2000; Senseman et al., 2003; Öezhan et al., 2005; Riley et al., 2005) These EmporeTM disks have been used coupled with a diffusion-membrane and used as a receiving phase for an EmporeTM disk-based passive sampling device Uptake of organic compounds has been found to be dependent on the physicochemical properties of individual target analytes (Kingston et al., 2000) The authors demonstrated that a C18 EmporeTM disk employed as the receiving phase had a linear uptake phase of 14 days when a rate-limiting membrane was present (polysulfone for polar compounds or polyethylene for non-polar compounds) Sampling rates varied with water temperature and turbulence There was a reasonable agreement between the concentrations estimated using the passive samplers and those based on a series of spot samples for diuron at two marine sites

1.4.3.5 The polar organic chemical integrative sampler (POCIS)

The POCIS was constructed by forming a membrane-sorbent-membrane sandwich consisting of a solid-phase extraction sorbent or a triphasic admixture of SPE sorbents as a receiving phase and with an overlying polyethersulfone diffusion membrane (Alvarez et al., 2000; Alvarez et al., 2004) The POCIS sampler is comparable to the EmporeTM disk sampler in the way that it can be loaded with different receiving phases and diffusion membranes Using this sampler, linear

uptake of selected herbicides and pharmaceuticals with log K ow < 3 was observed for

up to 56 days The POCIS is directed towards polar compounds, and in a field study

in New Jersey stream, many monitored polar pesticides including atrazine and its derivatives were detected (Alvarez et al., 2005)

1.4.3.6 Other passive sampling devices

In addition to the passive sampling devices described above, a number of different methods have been published which may all be applied in the aquatic environment For example, a method using active carbon with an acrylic polymer lid with holes providing a non-turbulent diffusion zone (DiGiano et al., 1988) was tested in the

laboratory on p-xylene and atrazine The diffusion pathways of the sampler design

limit the amount of analyte adsorbed The device was exposed for 5-50 days After exposure, the sorbent (carbon) was extracted with carbon disulphide for xylene and methylene chloride for atrazine, dried with sodium sulphate and analysed by GC

Another device is the passive in situ concentration extraction sampler (PISES), which has been used for PCB analysis and consists of a brass tube filled with hexane and enclosed by a polyethylene membrane (Litten et al., 1993) Sampling occurs by molecular diffusion of organic contaminants from the water through the membranes into the collecting solvents It is simple and cheap to produce, but has a poor surface area-to-volume ratio and is subject to solvent loss

LeBlanc et al (2003) described a passive sampling device based on thin-layer chromatographic plates (C2 or C18) which has been tested with two organophosphate pesticides, diazinon and chlorpyrifos Linear uptake was observed

up to seven days exposure under laboratory conditions It is inexpensive to produce

and was investigated as a screening method for determining the presence of organic compounds in water

The rapid laboratory extraction method of solid-phase microextraction (SPME) has been used as an equilibrium sampling device (Verbruggen et al., 2000; Mayer et al., 2003) SPME can be considered to be between passive and dynamic sampling methods That is, the analytes are sampled from the medium in a controlled manner similar to passive sampling, yet the solid-phase fibre is in direct contact with the analytes (Kot et al., 2000) The main driving force is the difference in concentration The method is simple and does not require complicated extraction and clean-up procedures However, the sampling time to reach equilibrium is in the order of several minutes to several hours for hydrophobic compounds (Ramos et al., 1998; Mayer et al., 2000; Ter Laak et al., 2005) Therefore, this type of sampling is not suitable for long-term monitoring (Kot et al., 2000)

1.5 The choice of target pesticides

1.5.1 Pesticide physicochemical properties

Polar herbicides that are widely used for control of broad-leaved weeds and other vegetation in the Murrumbidgee Irrigation Area (MIA) (NSW, Australia) (Simpson and Haydon, 1999; AATSE, 2002) were selected as target analytes Table 1-3 details these herbicides’ structures and physicochemical properties

The Murrumbidgee Irrigation Area is the gazetted area defined by all the lands in which Murrumbidgee Irrigation Limited (MIL) supplies water from the Murrumbidgee River to approximately two thousand commercial farms producing rice, wheat, barley, corn, soybeans and horticultural crops A pesticide monitoring program has been established by MIL since 1997 to determine improvements in the management of contaminated irrigation wastewater from the farms (MIL, 2001) These herbicides are relatively inexpensive and very potent, even at low concentrations (AATSE, 2002)

The most prolific of these are the triazine herbicides, simazine and atrazine Over

3000 tones of each of these are applied in Australia annually In addition, over 1000 tones of other classes of chemicals including phenoxy acids, benzoic acids and pyridines are also used every year There are no records of the amounts of pesticides actually used on farms Accurate estimates of pesticide use on irrigated pastures are also unavailable (AATSE, 2002) Information of pesticide usage within the MIA was obtained from experienced people working in the major cropping industries (Simpson and Haydon, 1999) Table 1-4 shows the estimated herbicide usage in the MIA The diuron (urea herbicide), phenoxy acid herbicides, 2,4-D [(2,4-dichlorophenoxy)acetic acid], along with MCPA [(4-chloro-2-methylphenoxy)acetic acid], are the major herbicides within this group (Simpson and Haydon, 1999; AATSE, 2002) Other target herbicides that were suspected to be present include simazine, atrazine, bensulfuron methyl (sulfonyl urea herbicide), metolachlor (miscellaneous herbicide) and clomazone (unclassified herbicide) (AATSE, 2002) It

is also noted that these data were reported in 1999 and amounts of the target herbicide usage may change every year