Electroanalytical Methodology for the Direct Determination of 2 , 4-Dichlorophenoxyacetic Acid in Soil Samples Using a Graphite-Polyurethane Electrode

An electroanalytical methodology was developed for the direct determination of the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) using a graphite-polyurethane composite electrode and square wave voltammetry (SWV). 2,4-D exhibited one reduction peak with characteristics of an irreversible process at −0.54V (versus Ag/AgCl), which is controlled by the diffusion of the reagent on the electrode surface. After the experimental parameters optimization (pH 2.0, f = 50 s, a = 0.50V, and ΔE


Introduction
2,4-D (2,4-Dichlorophenoxyacetic acid) is one of the most widely used herbicides in the world and is characterized as low-cost, quite efficient even at low concentrations, and moderately hazardous (class II) according to the World Health Organization [1].Although this herbicide easily moves though the soil, its leaching into the groundwater can be minimized due to the degradation by microorganisms and also the absorption by plants.Its primary route of degradation in the environment is by microorganisms, which increases with temperature, humidity, pH, and organic matter content [2,3].In Brazil, the maximum acceptable concentration of 2,4-D in water is 4.0 g L −1 , according to resolution 357 of the Brazilian National Environment Council [4].
Several chromatographic [5][6][7][8][9] and immunoassay methods [10][11][12][13] have been used to determine 2,4-D in biological and environmental matrices.However, chromatographic methodologies are time-consuming due to the derivatization step and complex extraction procedures.Immunoassay methods use biological subtracts that may present poor stability and in some cases result in low repeatability of the analytical signals for a long analysis interval [14].
Electroanalytical procedures have been successfully applied for the determination of inorganic and organic compounds, including pesticides, in many environmental matrices, due to their good sensitivity, short time analysis, low-cost, and the possibility of sample analysis without extractions or pretreatments [15][16][17].
It is well known from the literature that 2,4-D exhibits poor electroactivity on most common electrode material surfaces, such as carbon, platinum, and mercury [2].For this reason, it is crucial to develop modified electrodes for the determination and evaluation of 2,4-D.Under this circumstance, the electrochemical behavior for 2,4-D was investigated using a biosensor based on competitive immunoassay [18], an immunochemical biosensor based on a screenprinted amperometric transducer and monoclonal antibodies [19], an electrochemical sensor based on alkaline phosphatase inhibition [20], a carbon paste electrode modified with International Journal of Electrochemistry polyaniline [21], a glassy carbon electrode modified with cobalt chloride (III) protoporphyrin IX [22], a glassy carbon electrode modified with 2,4-D-molecular imprinted film [23], and conducting polymers (polyaniline and polypyrrole) as sensor materials [24].However, despite the intensive use of this herbicide and its potential contamination of water resources, there is little information about the application of electroanalytical techniques for the analysis of 2,4-D in environmental samples, which motivates additional studies in this area.
Currently, there has been an increasing interest in the use of composite electrodes for a great variety of applications [25], and the graphite-polyurethane composite electrode (GPU), developed by Mendes and coworkers [26], is an important electrode material for the determination of many compounds of environmental significance, such as organic pesticides.This electrode has been successfully used for the determination of dopamine [27], imipramine [28], hydroquinone, paracetamol and atenolol [29], rutin [30], and indole-3-acetic acid [31].
The aim of the present work is to develop, for the first time, an electroanalytical methodology to directly quantify the herbicide 2,4-D in soil samples, using the GPU electrode and the square wave voltammetry technique (SWV).The procedure was used to evaluate the behavior of 2,4-D in undisturbed soil columns under field conditions in order to determine the concentration and distribution of the applied 2,4-D along the soil profile with time and also the residual amounts of herbicide remaining in the soil after a relatively long period of time (100 days).

Experimental
2.1.Chemical and Reagents.All reagents were of analytical grade (Sigma) and were used without prior purification.All solutions were prepared with purified water from a Milli-Q system (Millipore).Britton-Robinson (BR) buffer solution (0.1 mol L −1 ) was used as supporting electrolyte.The pH of the solutions was adjusted with 1.0 mol L −1 NaOH.Stock standard solution of 2,4-D herbicide (221 mg L −1 ) was prepared in methanol, kept under refrigeration, and protected from light.
2,4-D commercial formulation (DMA 806 BR-Dow AgroSciences Industrial LTDA) was obtained in the local market with a declared amount of 670 g L −1 .A stock solution of the commercial formulation (403 mg L −1 ) was prepared in aqueous media, kept under refrigeration, and protected from light as well.The 2,4-D commercial formulation was applied to the soil surface using a watering container at the maximum amount specified by the manufacturer (50.3 mg L −1 ).For a uniform distribution of the herbicide, both soils were carefully irrigated with 6 liters of water, just after the herbicide application.

Instrumentation.
Voltammetric measurements were performed using a potentiostat/galvanostat AUTOLAB PGSTAT 30 in an electrochemical cell made from borosilicate glass with a Teflon cover, which contains orifices to insert the working (GPU), reference (Ag/AgCl-KCl 3.0 mol L −1 ), and auxiliary (platinum wire) electrodes and for nitrogen purging.
The GPU electrode was constructed by mixing graphite and polyurethane resin at the ratio 60 : 40 (w/w) [26].There is no special condition for electrode storage.The easy and fast regeneration of the GPU electrode surface was achieved by polishing the electrode disk surface with a 2,000 grit sand paper before each measurement to get a clean and reproducible electrode surface.

Electroanalytical Methodology.
Analytical curves were obtained through the external standard calibration procedure.The sensitivity was checked by the limit of detection (DL = 3/) and quantification (QL = 10/), where  is the standard deviation of ten voltammograms registered for the blank and  is the slope of the analytical curve [32] for an average of six curves.
The precision was calculated by the relative standard deviation (RSD) and checked by an experiment performed in one day ( = 10, intra-assay precision/repeatability) as well as by experiments in different days ( = 6, interassay precision/intermediate precision), with a 0.33 mg L −1 2,4-D solution.
The accuracy was evaluated in terms of relative error (bias %) obtained from 2,4-D recovering experiments for two soils (sandy and clayey).The experiments were performed ( = 3) by adding 100 mL of water to 5.0 g of dried soil.The mixture was fortified with a 0.33 mg L −1 2,4-D solution (i.e., spiked with 150 L from a 221 mg L −1 2,4-D stock solution), agitated for 24 h in a horizontal shaker, and centrifuged for 20 min at 13,500 rpm (Sorvall SL-50T Rotor, diameter = 22.2 cm).The pH of the supernatant was adjusted to 2.0 with concentrated phosphoric acid and the concentration of 2,4-D was determined with the previously developed electroanalytical method by means of a standard addition procedure, where 50 L of a 221 mg L −1 2,4-D stock solution was used to obtain concentrations of 0.88, 1.43, and 1.97 mg L −1 of 2,4-D.The 2,4-D estimated concentration was directly obtained through extrapolation, in the -axis, from the linear regression of the   versus  2,4-D curves.
The robustness of the methodology was checked during the optimization of the electrolyte pH.After selecting the best pH value, a small pH variation study was done near this region in order to measure the methodology capacity to remain unaffected by short-term pH fluctuations.A statistical analysis was performed (-test with a 95% confidence interval) to estimate the effect of small pH variation on the methodology performance.

Soil Columns
Experiments.The commercial herbicide formulation was applied to the surface of two types of soil, a sandy textured (AM1) and a medium clayey textured soil (AM2).
Soil samples were chemically (pH in CaCl 2 , organic matter-OM and cation exchange capacity-CEC) and physically (clay, silt, and sand percentages) characterized in the soil laboratory PIRASOLO, Piracicaba, São Paulo, Brazil.Undisturbed soil cores were packed in cement cylindrical containers with a total volume of 1.0 m 3 capacity (Figure 1).After the herbicide application, the soil containers were kept uncovered to sunlight and rainfall to provide the herbicide natural transport and degradation along the soil profile.Water percolation was monitored at depths of 25, 50, and 75 cm for three months.Soil solutions were obtained using porous cup extractors made of silicon carbide (SiC), which were installed in each cement container, at the same aforementioned depths.After 100 days from the beginning of the experiment, soil samples were collected with a Dutch-type auger at the same sampling depths and in all the cement containers.These soil samples were used to determine the residual amounts of the herbicide remaining in both soils.Soil solutions were also collected from the three depths before application of the herbicide.Soil containers with no herbicide addition (blank) were also prepared, sampled, and analyzed for comparison.Figure 1 presents the field experimental setup with three soil solution extractors and the vacuum pump used to collect the soil solutions in three depths.
The collected soil solutions were buffered and directly analyzed using the developed electroanalytical procedure.All the disturbed soil samples extracted (5 g) at the end of the experiment were treated with 1.0 mL of acidified water pH 1.0 (H 2 SO 4 9.0 mol L −1 ) and 15 mL of dichloromethane.The herbicide was extracted by an ultrasound system for 1 h.The extracts were combined and filtered through quantitative paper.The final extract was reduced in a rotary evaporator up to 1.0 mL.To certify that all the dichloromethane was evaporated, a flux of N 2 was applied over the extract for 2 min.The final extract was then diluted in 20 mL of BR buffer solution and analyzed with the developed electroanalytical procedure.

Results and Discussion
The voltammetric response of 2,4-D (10.5 mg L −1 ) was initially evaluated and compared using different carbon working electrodes, such as glassy carbon (GC), carbon paste (CP) electrode, and GPU. Figure 2 illustrates the square wave voltammograms of each working electrode in 0.1 mol L −1 BR buffer solution (pH 2.0).2,4-D presents a very well-defined reduction peak on the GPU electrode at −0.54 V (versus Ag/AgCl), whereas on the GC and on the CP electrodes the reduction process exhibits broadened and poor voltammetric profiles.The best response of the GPU electrode suggests a favorable interaction between the polyurethane and the herbicide that results in a more effective electron transfer process.

Cyclic Voltammetric Studies. 2,4-D voltammetric curves
were registered in the range of 10-500 mV s −1 (Figure 3) to investigate the degree of reversibility of the electrode process and also to determine the predominate type of mass transport.The irreversibility nature of the reaction is suggested by the absence of anodic peak in the reverse scan and confirmed by the linear displacement of the peak potential (  ) with the increase of the scan rate [33].From the linearity of   versus V 1/2 graph (inset of Figure 3,   (A) = 0.99 + 0.44V 1/2 (mV s −1 ) 1/2 ,  2 = 0.995), it is possible to conclude that the ratelimiting step of the 2,4-D electrochemical reaction at GPU electrode is diffusion-controlled [34].A plot of log   versus log V gave a straight line (log   (A) = −0.11+ 0.42 log V (mV s −1 ),  2 = 0.995) with a slope of 0.42, very close to the theoretical value of 0.50, which is expected for a diffusioncontrolled process [34].

Electroanalytical Determination of 2,4-D.
SWV was chosen from other electroanalytical techniques due to its high sensitivity and fast voltammograms acquisition.In order to obtain the best analytical signal for 2,4-D, some experiments were performed to optimize the electrolyte pH and the acquisition parameters of the SWV technique, namely, frequency (), pulse amplitude (), scan increment (Δ  ), accumulation time ( acc ), and accumulation potential ( acc ).
The effect of the pH was studied in the range of 2.0 to 8.0.The cathodic peak potential shifted linearly (dE/dpH = −0.064V/pH) to more negative potentials for pH ranging from 2.0 to 3.5, indicating that H + is involved in the electrode reaction.The best analytical signal was obtained in pH 2.0, which was then selected for the analytical applications.
The slope of −0.064 mV/pH is close to the Nerstian −0.059 mV/pH and consequently indicates that equal numbers of electrons and protons were involved in the electrochemical reduction of 2,4-D.From the literature, it is suggested that 4 electrons and 4 protons participated in the reduction of the acid carboxylic group to the correspondent alcohol [35].
The frequency was optimized in the range of 10 s −1 to 100 s −1 and for analytical purposes, a frequency value of 50 s −1 was selected.The pulse amplitude (0.10 to 0.50 V) and the scan increment (0.01 to 0.05 V) were also varied and the best values for these parameters were 0.50 V and 0.03 V, respectively.
Although the reaction is controlled by diffusion, adsorption plays an important role in the electroanalytical signal of 2,4-D on the GPU electrode.For this reason, the accumulation potential and time were also studied.It was observed that accumulation times higher than 10 s did not significantly increase the analytical signal, due to the saturation of the GPU electrode surface.The accumulation potential ( acc ) also affects the analytical signal with the highest peak current obtained for  acc = −0.40V.This behavior suggests that only the reagent adsorbs on the electrode surface, since the electrode reaction occurs at −0.54 V. Therefore, the accumulation potential and time were selected as −0.40 V and 10 sec for the subsequent analytical studies.
After optimizing the voltammetric acquisition parameters, analytical curves were obtained in the range of 0.66 to 2.62 mg L −1 in BR buffer solution (pH 2.0).The linear regression equation for the analytical curves was   (A) = (0.86 ± 0.09) + (4.47 ± 0.05)  2,4-D (mg L −1 ), with a determination coefficient  2 = 0.999.A detection limit (DL) of 17.6 g L −1 and a quantification limit (QL) of 58.6 g L −1 were obtained according to the procedure described in Section 2.3.The precision was checked in terms of intra-assay precision (repeatability) and interassay precision (intermediate precision).The RSD values for these statistical parameters were 2.5% and 3.9%, respectively.
Average recovered values of 2,4-D (concentration level: 0.33 mg L −1 ) in both soils were 93.6% ± 1.2% for the sandy soil (AM1) and 94.6% ± 4.2% for the clayey soil (AM2).The accuracy was evaluated in terms of the relative errors (bias %), which describes the deviation from the expected values in fortified samples.The obtained bias values were 6.8% for soil AM1 and 5.8% for soil AM2, which are within acceptable limits.The robustness, checked by small variations on the electrolyte pH (i.e., from 1.8 to 2.2) was satisfactory, since there is a 95% probability that the true population parameter (peak current) is between 5.09 A and 6.98 A.
In order to check the applicability of the developed electroanalytical procedure for soil solutions without previous treatment, analytical curves for 2,4-D were also obtained in the same concentration range as previously discussed for the supporting electrolyte in clayey and sandy soil solutions.Figure 4 shows square wave voltammograms for varying concentrations of 2,4-D in the clayey soil solution.
Table 1 shows the comparison between the proposed electroanalytical methodology (GPU/SWV) and other ones already published in the literature for the determination of 2,4-D.The detection limit of the GPU/SWV methodology is lower than the others previously reported [2,35,36].Moreover, high degrees of recovery were obtained in soil samples with standard deviations lower than 5%, suggesting the feasibility of the proposed methodology.In general, GPU electrode is an environmental friendly option comparing to the use of mercury electrodes and it is simple, fast regeneration and good stability open the possibility of its use for the analysis of other environmental contaminants.

Percolation of 2,4-D in Soil
Columns.After evaluation of the applicability of the GPU electrode for determining 2,4-D in soil solutions without previous treatment, the electroanalytical procedure was also applied to study transport and percolation of 2,4-D in two soils (sandy and a clayey textured), which were previously submitted to commercial herbicide solution treatment, according to the experimental procedure described in Section 2.4.
Table 2 presents some physicochemical properties of the soils evaluated.Soil AM1 is a sandy soil and AM2 is a medium texture soil with 40% clay (designated here as clayey soil).The pH values of the two soils are similar (pH around 5) and did not change significantly along the soil profiles.The amount of organic matter (OM) and cation exchange capacity (CEC) contents are higher for the clayey soil and decreased with depths as expected (higher concentration of roots and more decomposed materials close to the soil surface).
Table 3 shows the measured concentrations of 2,4-D in the extracted soil solutions, using the developed electroanalytical methodology, at different depths for soils AM1 International Journal of Electrochemistry  and AM2, as a function of time.In both soils, 2,4-D was detected in the deeper layer (75 cm depth) 24 hours after its application on the soil surface.The highest 2,4-D concentration values were detected in the first three days after its application.After one week, practically no 2,4-D was detected in the extracted soil solutions for both soils, according to the measured concentrations and their standard deviations.The inserted graphics show the concentration decay as a function of time.
2,4-D concentrations measured in both soils were very similar in terms of spatial dissipation.There were no significant differences in the herbicide spatial distribution for the sandy and clayey soils as measured in the extracted soil solutions.This was probably caused by the anionic behavior of the 2,4-D, exhibiting low sorption on clay minerals (negatively charged) and a negative correlation with pH and to some extent a positive correlation with OM.These results (Table 3) show a very low persistence of 2,4-D in the studied soils (around 7-14 days), which is in accordance with the reported half-life for field dissipation of about 10 days reported for the 2,4-D in soils [37].
Herbicides derived from weak acids such as the 2,4-D (pKa 2.8) lose protons and are predominantly in the anionic form in soils (pH 5 to 8) and since soil pH is higher than the herbicide pKa, the ionized species are predominant.However, a reduction in herbicide sorption due to the negative nature of soil net charge is expected [38,39].For this reason, 2,4-D sorption in soils is generally lower than that of cationic or basic herbicides and, therefore, such herbicide is more susceptible to photodegradation and microbiological degradation [40].
After 100 days from the beginning of the experiment, soil samples were collected for residual analysis of 2,4-D in the soils profile.During this period of time the total precipitation was 23.6 mm and the average temperature and relative humidity were 24.5 ∘ C and 80%, respectively.
The concentrations of 2,4-D determined at depths 0-25 cm, 25-50 cm, and 50-75 cm are shown in Table 4.A significant difference in the herbicide concentration along the profile and between the two soils can be verified.The average concentrations obtained in the soil profile (0 to 75 cm of depth) were 0.69 mg L −1 and 0.38 mg L −1 for soils AM1 (sandy) and AM2 (clayey), respectively.These results indicate a higher accumulation of 2,4-D in the sandy soil 100 days after the application on the soil surfaces.This behavior can be attributed to the higher amount of organic matter content in the clayey soil (AM2) compared to the sandy soil (AM1), which provides a favorable environment for herbicide microbiological degradation.Differences in concentration along the soil profile between soils can be explained by higher macroporosity and hydraulic conductivity of the sandy soil (AM1) when compared to the clayey soil AM2.

Conclusions
The developed electroanalytical procedure using the GPU electrode allowed the direct determination of the herbicide 2,4-D in soil solutions without previous extraction or pretreatment.The obtained limit of detection was adequate in order to study the percolation and sorption behavior of 2,4-D in soils under field conditions.The effect of matrices or soil solution interferences on the electroanalytical responses of the two soils was minimal, as verified in the analytical curves, providing consistent measurements in both porous cup extracted soil solution and solutions extracted from soil samples collected with an auger.The GPU electrode has proved to be an alternative for the direct electroanalytical determination of 2,4-D in soil samples to the conventionally used mercury electrode.

Table 1 :
Comparison between the proposed electroanalytical methodology and other ones already published in the literature for the detection of 2,4-D.

Table 2 :
Some physicochemical parameters of the soil studied.

Table 3 :
Concentration of 2,4-D in the soil solutions extracted with the porous cup extractor system in three depths for soils AM1 and AM2.Sample without herbicide (SH); b blank sample (BR); c concentration below the quantification limit of the technique (ND); d no sample-volume extracted was not enough for the analysis (SA). a

Table 4 :
Concentration of 2,4-D in the soil solutions extracted from the soil samples collected with an auger after 100 days of the herbicide application.