Abstract

Twenty-seven triazines and metabolites were screened throughout six advanced drinking water treatment plants (DWTP) in France and their respective resources. Seven molecules were quantified in raw waters with a maximum concentration of 91 ng/L reached for desethyl-atrazine. No metabolites generated through advanced degradation pathways were quantified. Concentration profiles for five DWTP treating ground or surface waters were very similar and remained stable over time. Only one DWTP treating groundwater presented differences between sampling periods due to variations in wells' operations. As expected, most treatment units (settling, ozonation, nitrification, sand filtration, chlorination) did not allow for efficient removal of these micropollutants. Adsorption on granular or powdered activated carbon (PAC) was highlighted as the best available technology for the majority of quantified compounds. Combined PAC and ultrafiltration treatment was especially adapted for the removal of hydroxy-atrazine, one of the most refractory components evaluated during this study. Indeed, among quantified pesticides, only hydroxy-atrazine and desethyl-deisopropyl-atrazine were measured in treated water, with concentrations below 12 ng/L.

INTRODUCTION

For several decades, pesticides have represented one of the most persistent chemical families in the environment. Indeed, they are frequently detected and quantified in water, air, soil and biota (WHO 2011; Masiá et al. 2013). Nevertheless, their impacts on human health and on the environment (live beings, water resources, etc.) are not well defined. Some of them have endocrine disruption effects or are potentially carcinogenic or reprotoxic (Sinclair et al. 2010; WHO 2011; US EPA 2015). Triazines represent one of the most important families of pesticides due to their frequent application to cereal cultures in Europe starting in the 1970's. Despite being banned in Europe in 2003, they still represent one of the most frequently quantified families of pesticides in water resources (Scribner et al. 2000; Gilliom 2007; Lopez et al. 2015) and are still in use in other countries (e.g. USA).

Several metabolites can be formed by the degradation of parent compounds by chemical processes or biodegradation in water, soil or plants (Amalric et al. 2003). In the present work, the term metabolites included all the degradation products of pesticides formed during metabolization by living beings and/or physico-chemical processes, in the environment and/or in the treatment plant. Most of the time, metabolites are more hydrophilic, more polar, more mobile and more toxic than the parent compounds (Boxall et al. 2004). As such, they are also persistent in the environment, especially in water and soil. Atrazine and simazine degradation pathways are well-known (Scribner et al. 2000) but the occurrence of metabolites in water and their behaviour throughout drinking water treatments are only partially documented (Adams 2006). That is why it is interesting looking at the occurrence of metabolites formed by the degradation of desethylatrazine or desethyl-deisopropyl-atrazine, for example. Based on current knowledge, the European Drinking Water Directive (Directive 98/83/EC) has set the limit to be respected at 0.1 μg/L for each pesticide and its relevant metabolites and 0.5 μg/L for the sum of all detected pesticides and metabolites in drinking water. The limit set for raw waters is 0.1 μg/L for simple treatment lines and 2 μg/L for advanced treatment chains (French regulation decree n° 2001-1220).

It is therefore important to evaluate the treatability of these compounds regarding common and advanced treatments, such as adsorption onto activated carbon, ozonation or retention by low molecular weight cut-off membrane processes (nano filtration, reverse osmosis). Indeed, several studies show differing results on the treatability of desethyl-deisopropyl-atrazine (DeDIA) by adsorption onto activated carbon, and treatability of other metabolites is not well documented or is controversial (Nam et al. 2014; Domergues et al. 2016), especially for those metabolites formed after several steps of degradation. The issue for drinking water producers is to understand the behaviour of pesticides and their metabolites in DWTP in order to eliminate them along the treatment line. The main objective of this study was to quantify a large panel of triazines and their metabolites in resources and along DWTP. A second objective was to assess the removal efficiency of several treatment units for these micropollutants: this preliminary approach was done to evaluate the potential needs to enhance or optimize the processes operated on the studied sites.

MATERIAL AND METHODS

A total of 71 grab samples were collected within six French resources and DWTP over two sampling campaigns. The first was performed in September 2016 and the second in January 2017, covering the end of summer (dry, hot period) and winter (wet, cold period) seasons. Raw water for three DWTP was the Seine River: two upstream of Paris and one downstream. The Seine River was selected as it is the water supply for the majority of the 12 million people living in the Paris area, and its vast basin covers 30% of the French population. It is influenced by both agricultural and animal breeding activities. The three other DWTP use groundwater as their resource. Treatment processes at these DWTP were advanced, comprising all or most of the following units: pre-ozonation, clarification (coagulation + flocculation + settling + sand filtration), ozonation, granular activated carbon (GAC) filtration, ultrafiltration, ultraviolet (UV) treatment and/or chlorination. The description of each site, including resource and treatment line, is presented in Table 1.

Table 1

Short description and operating conditions of the six DWTP, CT = contact time

Name of sampling siteSite ASite BSite C
Type of raw water Ground water Ground water Ground water 
Water treatment line Raw water > coagulation > ozonea > GAC > chlorination Raw water > nitrificationa > ozone > GAC > chlorination Raw water > settling + PACb > nitrificationc > sand filtration or GAC > ozone > chlorination 
Water flow rate 1,000 m3/h - 1,000 m3/h 2,000 m3/h - 1,460 m3/h 3,300 m3/h - 4,900 m3/h 
Operating conditions - 1st sampling Campaign O3 : residual dose = 0.4 mg/L – CT = 7 min
GAC : speed = 22 m/h – CT = 8 min 
O3 : residual dose = 0.2 mg/L
GAC : CT = 9 min 
PAC: 2.5 mg/L
GAC : CT ?
O3 : CT = 41 min – residual dose = 0.05 mg/L 
Operating conditions - 2nd sampling Campaign O3 : residual dose = 0.2 mg/L
GAC : CT = 7 min 
CAP : 2.5 mg/L
GAC : CT ?
O3 : CT = 18 min – residual dose = 0.03 mg/L 
Name of sampling siteSite DSite ESite F
Type of raw water Surface water Surface water Surface water 
Water treatment line Raw water > pre-ozonec > settling > GAC > Ozone > GAC > UV > chlorinationd Raw water > settling > GAC > Ozone > PACe + UF > chlorination Raw water > pre-ozone > settling > sand filtrationa > ozone > GAC > chlorination 
Water flow rate 2,000 m3/h - 1,880 m3/h 556 m3/h - 880 m3/h 2,170 m3/h - 2,100 m3/h 
Operating conditions - 1st sampling Campaign Pre-O3 : rate = 0.4 mg/L
GAC : CT = 9 min – speed = 7 m/h
post-ozone : CT = 28 min
UV : Dose = 40 mJ/cm 
O3 : dose = 0.95 mg/L – Residual dose = 0.35 mg/L
CAP : dose = 4.5 mg/L 
Pre-O3 : CT = 3 min
GAC : CT = 12 min – speed = 9 m/h
O3 : CT = 8 min – rate = 0.25 mg/L 
Operating conditions - 2nd sampling Campaign Pre-O3 : no
GAC : CT = 11 min – speed = 5.4 m/h
post-O3 : rate = 0.59 mg/L – residual dose = 0.15 mg/L
UV : Dose = 40 mJ/cm 
O3 : dose = 0.49 mg/L – Residual dose = 0.2 mg/L
PAC : dose = 3.7 mg/L 
Pre-O3 : CT = 3 min
GAC : CT = 12 min – speed = 9 m/h
O3 : CT = 8 min – rate = 0.14 mg/L 
Name of sampling siteSite ASite BSite C
Type of raw water Ground water Ground water Ground water 
Water treatment line Raw water > coagulation > ozonea > GAC > chlorination Raw water > nitrificationa > ozone > GAC > chlorination Raw water > settling + PACb > nitrificationc > sand filtration or GAC > ozone > chlorination 
Water flow rate 1,000 m3/h - 1,000 m3/h 2,000 m3/h - 1,460 m3/h 3,300 m3/h - 4,900 m3/h 
Operating conditions - 1st sampling Campaign O3 : residual dose = 0.4 mg/L – CT = 7 min
GAC : speed = 22 m/h – CT = 8 min 
O3 : residual dose = 0.2 mg/L
GAC : CT = 9 min 
PAC: 2.5 mg/L
GAC : CT ?
O3 : CT = 41 min – residual dose = 0.05 mg/L 
Operating conditions - 2nd sampling Campaign O3 : residual dose = 0.2 mg/L
GAC : CT = 7 min 
CAP : 2.5 mg/L
GAC : CT ?
O3 : CT = 18 min – residual dose = 0.03 mg/L 
Name of sampling siteSite DSite ESite F
Type of raw water Surface water Surface water Surface water 
Water treatment line Raw water > pre-ozonec > settling > GAC > Ozone > GAC > UV > chlorinationd Raw water > settling > GAC > Ozone > PACe + UF > chlorination Raw water > pre-ozone > settling > sand filtrationa > ozone > GAC > chlorination 
Water flow rate 2,000 m3/h - 1,880 m3/h 556 m3/h - 880 m3/h 2,170 m3/h - 2,100 m3/h 
Operating conditions - 1st sampling Campaign Pre-O3 : rate = 0.4 mg/L
GAC : CT = 9 min – speed = 7 m/h
post-ozone : CT = 28 min
UV : Dose = 40 mJ/cm 
O3 : dose = 0.95 mg/L – Residual dose = 0.35 mg/L
CAP : dose = 4.5 mg/L 
Pre-O3 : CT = 3 min
GAC : CT = 12 min – speed = 9 m/h
O3 : CT = 8 min – rate = 0.25 mg/L 
Operating conditions - 2nd sampling Campaign Pre-O3 : no
GAC : CT = 11 min – speed = 5.4 m/h
post-O3 : rate = 0.59 mg/L – residual dose = 0.15 mg/L
UV : Dose = 40 mJ/cm 
O3 : dose = 0.49 mg/L – Residual dose = 0.2 mg/L
PAC : dose = 3.7 mg/L 
Pre-O3 : CT = 3 min
GAC : CT = 12 min – speed = 9 m/h
O3 : CT = 8 min – rate = 0.14 mg/L 

aSample available for campaign 2 only.

bOccasional PAC addition; dose equal to 5 mg/L of highly activated carbon (Iodine index up to 900 mg/g).

cSample available for campaign 1 only.

dMixing of different treatment lines in the DWTP.

eOn-line continuous PAC addition; dose equal to 5 mg/L of highly activated carbon.

Grab samples (two liters) were collected at each point and, once received in the laboratory, they were directly filtered on 0.45 μm glass fibre filters. In order to preserve all the targeted samples, they were stored at −20 °C if analyses could not be performed directly after filtration (Capdeville 2011). In the rare cases where precipitation was observed following freezing, samples were re-filtered at 0.45 μm before extraction. The analysis of the samples (Table 2) was performed by ultra-high performance liquid chromatography coupled with high resolution mass spectrometry (UPLC-HRMS, Orbitrap, ThermoFisher) using direct injection.

Table 2

Conditions of UPLC-MS/MS analyses

UPLC parametersHRMS conditions
Injected volume 10 μl Type of ionisation HESI 
Mobile phase flow rate 400 μl/min Polarity positive 
Type of column BEH C18 column (2.1 × 100 mm, 1.7 μm, 100 Å, Waters®Acquisition type SIM or MS2 
Column temperature 30 °C Spray voltage 3,000 V 
Solvent A Water S-lens 90 
Solvent B Methanol Sheath gas pressure 35 psi 
Gradient Time (min) A (%) B (%) Auxiliary gas pressure 10 psi 
95 95 Capillary temperature 320 °C 
0.5 Fragmentation energy 40 
16 40 60   
18 100   
20 100   
21 95   
23 95   
UPLC parametersHRMS conditions
Injected volume 10 μl Type of ionisation HESI 
Mobile phase flow rate 400 μl/min Polarity positive 
Type of column BEH C18 column (2.1 × 100 mm, 1.7 μm, 100 Å, Waters®Acquisition type SIM or MS2 
Column temperature 30 °C Spray voltage 3,000 V 
Solvent A Water S-lens 90 
Solvent B Methanol Sheath gas pressure 35 psi 
Gradient Time (min) A (%) B (%) Auxiliary gas pressure 10 psi 
95 95 Capillary temperature 320 °C 
0.5 Fragmentation energy 40 
16 40 60   
18 100   
20 100   
21 95   
23 95   

SIM, selected ion monitoring; MS2, use of mass spectrometer with collision cell; HESI, heated electrospray.

Compound detection was based on accurate exact mass using a resolution of 70,000 and/or MS2 spectra in order to correctly discriminate isomers (see Table S1 in the supplementary material, available with the online version of this paper). A calibration range was prepared with 27 standard solutions and eight internal labelled standards, ranging from 1 ng/L to 500 ng/L. Internal marked standards were added to all samples for internal calibration and samples were analyzed twice with and without standard addition at 1 μg/L in order to monitor losses during injection and possible matrix effects.

Instrument calibration was verified over the full range prior to each analytical run and quality controls were analyzed every eight sample injections. Internal standard recoveries in the samples ranged between 93% and 103%, regardless of the compound and the sample. Targeted compound recoveries in quality controls and in spiked samples were in the range 92%–108% and 86%–105% respectively. Results for each compound are presented in Table S2 of the supplementary material (available online).

Repeatability was measured between 5% and 12% depending on the compound. Limits of quantification were 5 ng/L for all compounds except hydroxy-propazine and hydroxy-atrazine whose limits were 100 ng/L and 1 ng/L respectively.

RESULTS AND DISCUSSION

Characterization of raw waters

Figure 1 presents the compounds quantified in raw waters for both sampling campaigns. The number of measured compounds was between three for site B and six for site C out of the 27 searched for in the samples.

Figure 1

Concentrations of quantified compounds in raw waters for two sampling campaigns.

Figure 1

Concentrations of quantified compounds in raw waters for two sampling campaigns.

The maximum concentration for these compounds was 90 ng/L for desethyl-atrazine (DEA) in the raw water of site B during the 2nd campaign. For the majority of DWTP, DEA was the compound with the maximum concentration, except for site B during the 1st campaign. Comparing the two campaigns, the profiles of concentrations in raw waters were similar except for site B. At this site, concentration of DEA was higher and atrazine was quantified only during the 2nd campaign. This could be explained by the variation in the use of groundwater wells which supply this plant. Two compounds were only quantified during the 1st campaign: deisopropyl-atrazine (DIA) at site A and desethyl-terbumeton at site F. Consequently, it was not possible to highlight a significant seasonal effect on pesticide and metabolite concentrations for these raw waters.

Profiles of concentrations were almost the same for sites D, E and F, which were supplied by surface water. Two of these sites were located upstream of the Paris area and the third downstream, meaning that urban activities did not have an impact on these compounds' concentrations, as was expected. The surface water's quality was therefore stable over the study period for this family of micropollutants.

The sum of concentrations in the raw water ranged from 50.5 to 252.1 ng/L for the 1st campaign and from 38.7 to 209 ng/L for the 2nd campaign. The 2 μg/L raw water E.U. regulation (Directive 98/83/EC) was respected for each sample and each compound.

Impact of granular or powdered activated carbon adsorption on pesticides

Concentrations of triazines and their metabolites throughout the DWTP from site F are presented for both sampling periods on Figure 2. As presented before, five compounds (DEA, DeDIA, hydroxy-atrazine, atrazine and desethyl-terbumeton) and three compounds (DEA, DeDIA, hydroxy-atrazine) for the two sampling periods respectively were quantified in the raw water. They were not eliminated by the first five treatment units (pre-ozonation, clarification through settling, sand filtration and post-ozonation) except for desethylterbumeton, quantified in the raw water only during the 1st campaign, which was removed by sand filtration due to probable biodegradation within filter media. Most of the compounds were totally removed by adsorption or biodegradation on GAC filters except hydroxy-atrazine, which was quantified at 2 ng/L in the treated water during the 2nd campaign and whose removal was 70% in these specific operational conditions. During the first campaign, pre-ozonation appeared to generate five hydroxylated metabolites but no parent compounds' raw water concentration could explain the resulting metabolites levels. Furthermore, the total concentration of pesticides and metabolites was higher after ozonation. This could be explained by the reactivity of pesticides adsorbed on organic matter with ozone and the transfer of the by-products in the dissolved phase. Indeed, analyses were performed only for the pesticides present in the dissolved phase. This hypothesis needs to be investigated for confirmation.

Figure 2

Concentrations of pesticides and metabolites at DWTP F during two seasons.

Figure 2

Concentrations of pesticides and metabolites at DWTP F during two seasons.

The same removal trend was underlined at the DWTP of sites E (Figure 3), A, B and D (see Table S3 in the supplementary material, available with the online version of this paper) with a total elimination of the majority of the compounds on activated carbon filters and a residual HA concentration in treated waters due to its partial removal in the different treatment units. Indeed, as an example, at site E, hydroxy-atrazine removal was between 65% and 80% in the clarification unit. Its elimination was total after GAC filter during the 1st campaign and additional removal of 65% was measured during the 2nd campaign on GAC filter. Thus, global removal of hydroxy-atrazine ranged from 82% to 100% at the different DWTP, with an efficiency of GAC filter varying between 14% (site D) and over 85% (site F) during the 1st campaign and between 39% (site C) and over 90% (site A). This high variation in the removal on GAC filters shows that operational conditions, especially the contact time, the age of the filter and the flow velocity on the GAC filters, are crucial for the elimination of polar metabolite. Note that removal calculations could not be done if concentration was below the limit of quantification.

Figure 3

Concentrations of pesticides and metabolites at DWTP E for both sampling campaigns.

Figure 3

Concentrations of pesticides and metabolites at DWTP E for both sampling campaigns.

At site B, GAC filter efficiency changed from one sampling campaign to the other: DEA, DeDIA and atrazine were only partially adsorbed on GAC filters during the 2nd campaign, while their concentrations were below the quantification limits for the 1st campaign. This could be explained by a change in well operations and the increase of pesticides concentrations in raw water. This change in concentration could induce different equilibrium states of GAC filter and a possible desorption of pesticides and/or a change of adsorption efficiency. Thus, removal of the sum of quantified triazines and metabolites varies from 95% to 40% for the 1st and the 2nd sampling campaigns respectively. Consequently, the concentrations after GAC filtration were higher when concentrations in raw water are higher.

Behaviour of DEA was exactly the same on the 2nd GAC filtration stage at site D. Note that the adsorption on the 1st GAC stage was partial for this site: elimination ranged from 10% to 40% and from 20% to 60% for 1st and 2nd campaigns respectively, depending on the compound. This was due to the variation in condition of use: contact time varied from 9 min for the 1st campaign to 11 min for the 2nd campaign, for a flowrate of 7 m/h and 5.4 m/h respectively.

The addition of powdered activated carbon (PAC) allowed for a better elimination of some metabolites. During both campaigns, concentrations of atrazine, simazine and DIA were below the quantification limits at site C after this treatment unit. Concentrations of DEA, DeDIA and HA decreased significantly for both sampling periods at site C following PAC addition. These values corresponded to removal efficiencies of the settling unit of ∼90%, ∼80% and ∼85% for DEA, DeDIA and HA respectively for both campaigns, showing the impact of PAC addition in this treatment. All the other treatment units of this DWTP did not have any impact on pesticides and metabolites' concentrations (Figure 4). The decrease of concentrations for DeDIA at site B (2nd campaign) and at site C (both campaigns) and for DEA at site C (both campaigns) and at sites B and D for the 2nd campaign could be induced by the chlorination step. The reactivity of chlorine with these compounds needs to be investigated further to be confirmed.

Figure 4

Concentrations of pesticides and metabolites in DWTP C for both seasons (sand filtration and GAC filtration are in parallel).

Figure 4

Concentrations of pesticides and metabolites in DWTP C for both seasons (sand filtration and GAC filtration are in parallel).

Quality of treated waters

Table 3 presents concentrations measured in treated waters. Two metabolites of atrazine were quantified with a maximum concentration of 12.1 ng/L for hydroxy-atrazine and 7.1 ng/L for DeDIA during the 2nd campaign.

Table 3

Concentration of quantified metabolites in treated waters for both sampling periods

Desethyl-deisopropyl-atrazine concentration (ng/L)Hydroxy-atrazine concentration (ng/L)
1st campaign Site A <5 <1 
Site B <5 2.7 
Site C <5 5.2 
Site D <5 1.7 
Site E <5 <1 
Site F <5 <1 
2nd campaign Site A <5 <1 
Site B 5.1 12.1 
Site C 7.1 3.8 
Site D <5 1.4 
Site E <5 <1 
Site F <5 2.1 
Desethyl-deisopropyl-atrazine concentration (ng/L)Hydroxy-atrazine concentration (ng/L)
1st campaign Site A <5 <1 
Site B <5 2.7 
Site C <5 5.2 
Site D <5 1.7 
Site E <5 <1 
Site F <5 <1 
2nd campaign Site A <5 <1 
Site B 5.1 12.1 
Site C 7.1 3.8 
Site D <5 1.4 
Site E <5 <1 
Site F <5 2.1 

DeDIA was only quantified at site B and C during the 2nd campaign, which corresponded to a removal of 75% on the treatment line of site B and 90% at site C. Hydroxy-atrazine was quantified at a frequency of 50% in treated waters. It was not totally removed by GAC adsorption: its global decrease within treatment lines reached 85% to 100% depending on the DWTP, except for site B (no decrease during the 2nd campaign). Nevertheless, if its adsorption on GAC was partial, the use of ultrafiltration in site E allowed a decrease of its concentration below 1 ng/L. Therefore, hydroxy-atrazine elimination was not observed in ozonation treatment units (sites B, D, E and F) or with the addition of PAC (in site C for example) in the settling unit. The only case of elimination with ozone was observed at site A during the 1st campaign. Overall, all the treated waters complied with the standard drinking water regulation limit of 0.1 μg/L for each metabolite.

Results overview

Sixteen compounds were never quantified in the samples. Table 4 presents the frequency of quantification for the compounds measured in at least one sample. Intermediate points referred to all the samples collected along the DWTP, after each simple unit. Two parent pesticides (atrazine and simazine) and nine metabolites whose parent compounds are atrazine, simazine, terbutylazine or terbumeton were quantified at least once in a sample.

Table 4

Frequency of quantification in different types of samples for both sampling campaigns for all DWTP (n = number of samples)

Campaign 1Campaign 2Campaign 1Campaign 2Campaign 1Campaign 2
Raw water (6)Raw water (6)Intermediate water (23)Intermediate water (24)Treated water (6)Treated water (6)
Simazine 17% 17% 4% 0% 0% 0% 
Hydroxy-simazine 0% 0% 4% 0% 0% 0% 
Atrazine 83% 83% 30% 25% 0% 0% 
Desethyl-deisopropyl-atrazine (DeDIA) 100% 100% 70% 67% 0% 33% 
Deisopropyl-hydroxy-atrazine (DIHA) 0% 0% 4% 0% 0% 0% 
Desethyl-hydroxy-atrazine (DeHA) 0% 0% 4% 0% 0% 0% 
Deisopropyl-atrazine (DIA) 33% 17% 4% 0% 0% 0% 
Desethyl-atrazine (DEA) 100% 100% 65% 75% 0% 0% 
Hydroxy-atrazine (HA) 100% 100% 74% 88% 50% 67% 
Desethyl-terbumeton 17% 0% 4% 0% 0% 0% 
Hydroxy-terbuthylazine 0% 0% 4% 0% 0% 0% 
Campaign 1Campaign 2Campaign 1Campaign 2Campaign 1Campaign 2
Raw water (6)Raw water (6)Intermediate water (23)Intermediate water (24)Treated water (6)Treated water (6)
Simazine 17% 17% 4% 0% 0% 0% 
Hydroxy-simazine 0% 0% 4% 0% 0% 0% 
Atrazine 83% 83% 30% 25% 0% 0% 
Desethyl-deisopropyl-atrazine (DeDIA) 100% 100% 70% 67% 0% 33% 
Deisopropyl-hydroxy-atrazine (DIHA) 0% 0% 4% 0% 0% 0% 
Desethyl-hydroxy-atrazine (DeHA) 0% 0% 4% 0% 0% 0% 
Deisopropyl-atrazine (DIA) 33% 17% 4% 0% 0% 0% 
Desethyl-atrazine (DEA) 100% 100% 65% 75% 0% 0% 
Hydroxy-atrazine (HA) 100% 100% 74% 88% 50% 67% 
Desethyl-terbumeton 17% 0% 4% 0% 0% 0% 
Hydroxy-terbuthylazine 0% 0% 4% 0% 0% 0% 

Hydroxy-atrazine was quantified in all types of water and remained present in treated waters for 50% and 67% of sampling sites for the 1st and 2nd campaigns respectively. During the 2nd campaign, only DeDIA was quantified with hydroxy-atrazine in treated water. Generally, DWTP were able to eliminate the compounds present in raw water. DEA, hydroxy-atrazine (HA) and DeDIA were always quantified in raw water, whatever the type of resource and the season, highlighting that metabolites of atrazine are still present in the environment despite the ban of these pesticides' use more than 15 years ago (Decision from European Union 2004/141/CE).

Among these quantified compounds, four metabolites were only measured following pre-ozonation treatment at DWTP F: deisopropyl-hydroxy-atrazine, desethyl-hydroxy-atrazine, hydroxy-simazine and hydroxy-terbutylazine. Parent-compounds were apparently degraded to their hydroxylated forms by oxidation during the ozonation step. These degraded forms were all eliminated by GAC filtration.

CONCLUSIONS

The rate of quantification for the 27 triazines and their metabolites screened within DWTP was low and most of the metabolites were never quantified in the samples. Two parent pesticides out of 14 and five of their metabolites were quantified at least once in a raw water. Concentrations in the raw waters ranged from 1 ng/L to 91 ng/L, with the maximum concentration measured for DEA. No seasonal effect was observed during this study and concentrations remained very stable whatever the molecule, the season and the site. Most of the time, removal of quantified micropollutants was observed only after GAC filtration units. Hydroxy-atrazine was the most refractory molecule to GAC adsorption. It was treated effectively with ultrafiltration coupled with PAC addition (5 mg/L of highly activated carbon) in one of the DWTP. Concentrations of hydroxy-atrazine and DeDIA in treated waters were below 12.1 ng/L and 7.1 ng/L respectively. Consequently, despite these quantifications, all treated waters complied with European regulations on drinking water.

ACKNOWLEDGEMENTS

The authors wish to thank Mrs Veronique Lahoussine from the French Water Agency of Seine-Normandy Basin for its financial support.

REFERENCES

REFERENCES
Amalric
L.
,
Baran
N.
,
Jeannot
R.
,
Martin
J. C.
&
Mouvert
C.
2003
Les mécanismes de transfert des produits phytosanitaires du sol vers les nappes et les méthodes d'analyse des produits phytosanitaires dans les eaux (Mechanism of Phytosanitary Products Transfer From Soil to Groundwater and Analytical Methods of Phytosanitary Products in Water)
.
Report BRGM/RP-51590-FR
,
Bureaux de Recherches Géologiques et Minières, BRGM
,
Orléans
,
France
.
Boxall
A. B. A.
,
Sinclair
C. J.
,
Fenner
K.
,
Kolpin
D.
&
Maund
S. J.
2004
When synthetic chemicals degrade in the environment
.
Environmental Science and Technology
38
(
19
),
368
375
.
Capdeville
M. J.
2011
Etude des cycles biogéochimiques des contaminants organiques dits ‘émergents’ dans les systems aquatiques (Study of the Biogeochemical Cycles of Organic Contaminants Called « Emerging » in Aquatic Systems)
.
PhD thesis
,
University of Bordeaux
,
France
.
Directive 98/83/EC – quality of water intended for human consumption
1998
.
Domergues
L.
,
Cimetière
N.
&
Giraudet
S.
2016
Adsorption de pesticides et de leurs métabolites en eau naturelle sur lit de charbon actif (adsorption of pesticides and their metabolites from natural water on activated carbon bed)
. In:
Proceedings, Congress JIE
,
Poitiers, France
.
Gilliom
R. J.
2007
Pesticides in U.S. streams and groundwater
.
Environmental Science and Technology
41
(
10
),
3408
3414
.
Lopez
B.
,
Ollivier
P.
,
Togola
A.
,
Baran
N.
&
Ghestem
J. P.
2015
Screening of French groundwater for regulated and emerging contaminants
.
Science of the Total Environment
518–519
,
562
573
.
Masiá
A.
,
Campo
J.
,
Vázquez-Roig
P.
,
Blasco
C.
&
Picó
Y.
2013
Screening of currently used pesticides in water, sediments and biota of the Guadalquivir River Basin (Spain)
.
Journal of Hazardous Materials
263
(
1
),
95
104
.
Nam
S. W.
,
Choi
D. J.
,
Kim
S. K.
,
Her
N.
&
Zoh
K. D.
2014
Adsorption characteristics of selected hydrophilic and hydrophobic micro-pollutants in water using activated carbon
.
Journal of Hazardous Materials
270
,
144
152
.
Scribner
E. A.
,
Thurman
E. M.
&
Zimmerman
L. R.
2000
Analysis of selected herbicide metabolites in surface and ground water of the United States
.
Science of Total Environment
248
(
2–3
),
157
167
.
Sinclair
C.
,
Van Beinum
W.
,
Adams
C.
,
Bevan
R.
,
Levy
L.
,
Parsons
S.
,
Goslan
E.
&
Baumann
G.
2010
A Desk Study on Pesticide Metabolites, Degradation and Reaction Products to Inform the Inspectorate's Position on Monitoring Requirements
.
Report DWI-70-2-232
,
The Food and Environment Research Agency
,
York
,
United Kingdom
.
U.S. EPA
2015
Chemicals Evaluated for Carcinogenic Potential
.
Annual Cancer Report. http://npic.orst.edu/chemicals_evaluated.pdf (accessed 12 July 2017)
.
WHO
2011
Atrazine and Its Metabolites in Drinking-Water – Background Document for Development of WHO Guidelines for Drinking-Water Quality
.
Report WHO/HSE/WSH/10.01/11/Rev/1
,
World Health Organization
,
Geneva
,
Switzerland
.

Supplementary data