Back to search

University of the Aegean

Βελτιστοποίηση μεθόδου ανάλυσης ζιζανιοκτόνων σε υδατικά υποστρώματα με HPLC

Abstract

dc:description

The aim of this study was to optimize an analytical method for the determination ofherbicides in water matrices. The compounds were extracted from surface water usingsolid-phase extraction and analyzed with high performance liquid chromatography. Seventeen compounds were selected which belong to the chemical classes of triazine and phenylurea herbicides. The herbicide chloridazon, a pydidazinone, was also included in the study. Firstly, the chromatographic behavior of the mixture was studied with isocratic and gradient experiments. Isocratic experiments with methanol/water and acetonitrile/ water as the mobile phase were conducted in order to test the fitting of the linear (logk=logkw-Sφ) and the quadratic retention models (logk=logkw+Bφ+A<φ²) to the experimental data. The results showed that the quadratic equation provided better fit to the retention - mobile phase composition data and this was due to the curvature of the logk-φ graphs. The S parameter is a measure of the selectivity of the chromatographic system and it is estimated using the linear retention model. logkw is used to estimate the octanol-water partition coefficient logP, it is related to the toxicity, bioconcentration and adsorption of the organic compounds and it is determined using the linear or/and the quadratic retention models. The results showed that both S and logkw depend on the range of the organic modifier in the mobile phase. So for the accurate determination of logkw, the highest possible percentage of water in the mobile phase should be used. For the estimation of logkw, methanol is a more suitable organic solvent because it resembles the properties of water better than acetonitrile. The quadratic model estimates more accurately logkw because it fits better to the data. Regarding the goodness of fit of the linear model between the values of S and logkw as an index of mixture similarity (congenerity) and system selectivity, it follows that acetonitrile is a more selective solvent compared to methanol for the mixture of the herbicides. On the contrary the two organic modifiers are equivalent in terms of selectivity for a mixture of organophosphorous insecticides. Moreover, the goodness of fit of the linear model between the estimated values of logkw and values of logP that were collected from the literature was regarded as an index for the suitability of the HPLC system for the estimation of the logP. Methanol provided a better fit for the model compared to acetonitrile. Hexazinon was an outlier and this is probably due to secondary interactions of the compound with the stationary phase or to errors in the determination of logP. The results from the gradient experiments showed that gradient steepness at roomtemperature was not efficient for the separation of the mixture of herbicides with eitheracetonitrile or methanol. On the contrary, the simultaneous optimization of gradient steepness and temperature resulted in the adequate separation of the herbicide mixture whereas some peak pairs of a mixture of organophosphourous compounds remained unresolved. This is attributed to the more diverse and polar molecular nature of the herbicide mixture. The stability of the mixture of herbicides was tested for a period of 50 days under laboratory conditions. The herbicides were stored in natural water, at 4 °C (pH=3, pH=7.5 or with 10 mg/L of HgC1₂) or on Cl8 and SDVB solid phase extraction cartridges at -10 °C. Storing the herbicides in water, at pH=7.5 and 4 °C, resulted in minimal losses for all the compounds. The herbicides remained stable on the SDVB cartridges for 18 days and on the Cl8 cartridges for 50 days but the variability of the recoveries in Cl8 is greater due to interfering compounds. The acidification of the sample resulted in the acid-catalyzed hydrolysis of cyanazine. HgC1₂ strongly interacted with the extraction of some of the triazine herbicides on the SDVB cartridges and with atraton on Cl8 cartridges. Additional experiments were conducted in order to elucidate the effect of HgC1₂ on the efficiency of solid phase and liquid/liquid extraction of the herbicides from water matrices. The results showed that whereas most of the Hg²⁺ is significantly retained on the SDVB material, almost all of it is detected in the effluent of the Cl8 cartridge. The retention on the SDVB material is probably due to complex formation with polar impurities or with the π-electrons of the aromatic rings of the resin. A small amount of Hg²⁺ is probably hold on the Cl8 cartridge with ion-exchange forces with the free silanols of the material. In the presence of Hg²⁺, the symmetric triazines except hexazinon and the pyridazinone chloridazon are strongly retained on the SDVB resin, they are not eluted with acetonitrile whereas they are eluted with acidified methanol. A possible mechanism of the retention on the SDVB resin is the formation of a complex between Hg²⁺, the resin and the herbicides. On the contrary, the substituted urea compounds are eluted with acetonitrile, indicating a reversed phase mechanism of retention. In the presence of Hg²⁺, all the herbicides except atraton are quantitatively eluted with acetonitrile from the Cl8 resin. A possible mechanism of the retention of atraton on the Cl8 material is the formation of a complex between Hg and the herbicides and the subsequent attraction of the complex on the silanols of the Cl8 material. On the contrary, Hg²⁺ is does not influence the liquid/liquid extraction efficiency. Taking into account the above observations, HgC1₂ is not a suitable preservative for the analysis of triazine herbicides in water samples. In the next set of experiments, the performance of a solid phase extraction procedure of the herbicides was tested with respect to parameters such as the solid phase extraction material and the eluting solvent. The first four polar herbicides were not sufficiently retained on the Cl8 cartridge because their water solubility is very high. Methanol, acetonitrile and a mixture of acetonitrile and dichloromethane resulted in sufficient recoveries for the rest of the compounds. Moreover, the acetonitrile/ dichloromethane mixture eluted only a small amount of humic compounds from the cartridge and it was easily evaporated. The addition of methanol or NaCl in the water sample before the extraction did not improve the efficiency of the procedure. The use of SDVB cartridges resulted in almost quantitative recoveries for all the compounds, even with 2 L of water samples. The extraction with SDVB resin, the subsequent elution with an acetonitrile/ dichloromethane mixture and the removal of the water phase over the organic solvent resulted in adequate recoveries, precision and clean extracts. Moreover the method was robust to changes in the flow rate during extraction and the time of the evaporation of the solvent. Taking into account this initial evaluation, the method was considered as promising for the analysis of the herbicides in water matrices. In the final experimental set the proposed method was validated against several criteria (calibration curve, precision, accuracy, limit of detection, limit of quantification, range, linearity, recovery and selectivity-specificity) in order to demonstrate that it is suitable for its intended use. The results of the validation experiment showed that the calibration procedure should be applied to the whole analytical method and in parts of it (e.g. HPLC analysis) in order to include the variability of all the different stages. The estimation of the limit of detection using the confidence intervals of the calibration curve for all the concentrations of the calibration experiment resulted in higher LOD values compared with those that resulted in by the method of the standard deviation of blank samples. On the contrary, the use of a reduced calibration range (the first three concentrations) resulted in comparable values with the method that uses the standard deviation of blank samples. This is due to heteroscedasticity of the data of thecalibration experiment. So in case that the limit of detection is estimated by the confidence intervals of the calibration curve, the experiment should be performed with concentrations near the supposed limit of detection. The estimation of the limit of detection using the standard deviation of the results of a small concentration it is considered unreliable because it is influenced by the choice of the concentration level. Taking into consideration the above remarks, the limit of detection of the method is determined as the average value from the confidence intervals of the reduced calibration curve (first three concentrations) and the standard deviation of the blank samples. The limits of detection of the method ranged between 0.04-0.14 /μg/L. The limits of quantification were determined as LQ=3.04LD and at the same time it was assured that at the quantification level the accuracy ranged between 80-120% and the intermediate precision was ≤10%. The limits of quantification of the method ranged between 0.11-0.41 μg/L. The limits of detection of the proposed method are comparable with values that appear in the recent literature under the same experimental conditions (solid phase extraction of triazine and urea herbicides from surface waters and analysis with HPLC/DAD). The limits of detection are considered adequate for the determination of the herbicides in the surface waters. The range of the method was considered to be between the limit of quantification and the upper concentration of the calibration experiment. For these concentrations, certain criteria regarding linearity, accuracy and precision were fulfilled. The recoveries of the method ranged between 74.4%-112.9% in accordance with the European Union recommendations. Finally, the selectivity of the method was checked by analyzing two surface and one seawater sample from different locations in Lesvos. The interference that elutes at the beginning of the chromatogram and it is due to the dissolved organic matter of the water was efficiently removed in all samples. The level of interference peaks in the other parts of the chromatogram depended on the sampling point.

Degree

thesis:*
Grantor dc:publisher
University of the Aegean
Year dc:date
2002

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Kotrikla, Anna-Maria
  • Κοτρικλά, Άννα

Subjects

dc:subject × 18

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/13544

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Kotrikla, Anna-Maria; Κοτρικλά, Άννα. Βελτιστοποίηση μεθόδου ανάλυσης ζιζανιοκτόνων σε υδατικά υποστρώματα με HPLC. University of the Aegean, 2002. http://hdl.handle.net/10442/hedi/13544