SciELO - Scientific Electronic Library Online

 
vol.50 número2SYNTHESIS OF IRON AND IRON-MANGANESE COLLOIDS AND NANOPARTICLES USING ORGANIC SOLVENTSAN INVESTIGATION ON THE PRESENCE OF NiAl2O4 IN A STABLE Ni ON ALUMINA CATALYST FOR DRY REFORMING índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Revista

Articulo

Indicadores

Links relacionados

Compartir


Journal of the Chilean Chemical Society

versión On-line ISSN 0717-9707

J. Chil. Chem. Soc. v.50 n.2 Concepción jun. 2005

http://dx.doi.org/10.4067/S0717-97072005000200004 

 

J. Chil. Chem. Soc., 50, N 2 (2005), págs.: 461-464

 

SIMULTANEOUS SPECTROPHOTOMETRIC DETERMINATION OF ATRAZINE AND DICAMBA IN WATER BY PARTIAL LEAST SQUARES REGRESSION

 

Judith Amador-Hernández1*, Miguel Velázquez-Manzanares1, María del Rocío Gutiérrez-Ortiz1, Beatriz Hernández-Carlos1, Magaly Peral-Torres1 and Pedro Luis López-de- Alba2

1 Instituto de Ecología, Universidad del Mar, Ciudad Universitaria S/N, CP 42070 Puerto Angel, Oaxaca, México. Fax number: (52) 958 584 3078. E-mail: judith@angel.umar.mx
2 Instituto de Investigaciones Científicas, Universidad de Guanajuato, L. de Retana 5, CP 36000 Guanajuato, Guanajuato, México.


Keywords: Herbicides, atrazine, dicamba, PLS, spectrophotometry

ABSTRACT

It is presented the development of a new and easy analytical method for the simultaneous determination of atrazine and dicamba in water. Under the experimental conditions proposed, atrazine showed a linear dynamic range of 0.3 to 4.9 µg mL-1, a limit of detection of 0.1 µg mL-1 and a limit of quantification of 0.3 µg mL-1; similarly, dicamba showed a linear dynamic range of 0.8 to 8.8 µg mL-1, a limit of detection of 0.2 µg mL-1 and a limit of quantification of 0.8 µ mL-1. Partial Least Squares Regression (Type 1) was applied for the resolution of spectral interferences between analytes, obtaining recovery percentages higher than 95 % for both compounds in synthetic mixtures. Finally, quantitative determination of both herbicides in tap, well and sea water samples, was carried out successfully by the proposed method, previous preconcentration of the analytes by solid phase extraction.


INTRODUCTION

Nowadays, herbicides are used in a wide variety of crops to control pets. According to the U. S. Environmental Protection Agency, atrazine (2-chloro-4-ethylamine-6-isopropilamine-S-triazine) has a low acute toxicity and is used to control broadleaf weeds and some grassy weeds on corn, sorghum, sugarcane, wheat, macadamia nuts, pasture, conifers, woody ornamentals, among others1). In the case of dicamba (3,6-dichloro-2-methoxybenzoic acid), this is a postemergence herbicide used to control weeds, docks, bracken, brush, to mention some examples2). Both herbicides can be used in combination; therefore, some formulations with the two active ingredients are available.

In the literature, numerous chromatographic methods have been reported for the quantification of atrazine and dicamba in environmental samples3-6), but there are no reports about their simultaneous determination by UV-Vis spectrophotometry. In general terms, this detection technique is simple, rapid and of low cost, in contrast with its limited selectivity and sensitivity. Fortunately, there are some analytical strategies that can be used to compensate these limitations.

Partial Least Squares Regression Type 1 (PLS-1) is a powerful analytical tool generally used for the resolution of multicomponent systems with chemical or spectral interference drawbacks. As an example, it has been used to interpret spectrophotometric7, 8), fluorimetric9, 10), infrared11, 12), polarographic13) or laser induced spectroscopic data14). On the other hand, solid phase extraction (SPE) is extensively used for cleanup and preconcentration of trace analytes in a wide variety of samples15). In particular, the quantitative determination of pesticides in water, soil or air samples has been carried out with positive results, using SPE and different analytical techniques16-18).

In this work, the simultaneous determination of atrazine and dicamba in water is described. The quantification of both herbicides is developed by UV-Vis spectrophotometry, individually or in mixture by PLS-1. SPE is used to preconcentrate both herbicides in different water samples, spiked with the compounds of interest.

EXPERIMENTAL

Apparatus

A UV-Vis spectrophotometer model Lambda EZ210 (Perkin Elmer, USA) is used, coupled to a Pentium IV PC. Data acquisition and storage were carried out with PESSW v.1.2.E. (Perkin Elmer, USA); Pirouette v.3.10 (Infometrix Inc., USA) was used for data treatment. SPE was applied with the aid of a 12-port visiprep vacuum manifold (Supelco, USA).

Reagents

All chemicals were of analytical-reagents grade. Atrazine (ATR) and dicamba (DIC) were pestanal reagents (Riedel de Häen, Germany). Ultrapure water, purified with an EASYPureUV equipment (Barnstead, USA), was used throughout. Nitrogen of purity of 99.999% (INFRA, Mexico) was also used.

Stock solutions containing 80 mg mL-1 of ATR and DIC in methanol:water (1:1) were prepared, respectively. These solutions were stored at 4°C and protected against light; they were stable for at least one month. Working solutions were prepared daily by appropriate dilution. A buffer solution of KH2PO4/NaKHPO4 0.1 M, pH 7, adjusted with NaOH 0.1 M, was also used.

SPE was carried out with reversed phase cartridges (Supelclean ENVI-18, from SUPELCO, USA), with 500 mg of sorbent and 6 mL of tube size.

Procedure

Initially, the individual calibration was carried out for the two herbicides studied. Therefore, different solutions containing ATR (0.3 to 4.9 mg mL-1) or DIC (0.8 to 8.8 mg mL-1) were prepared. Appropriate volumes of the working solution of each herbicide were transferred to 10-mL volumetric flasks; then, 1 mL of buffer solution and methanol to complete 1 mL were added and the flasks were filled up with ultrapure water. The absorption spectra were recorded in the range of 200 to 300 nm against a blank.

For the resolution of binary mixtures by PLS-1, a calibration set of samples was prepared in the same way. Adequate volumes of the working solutions of ATR and DIC were mixed with 1 mL of the buffer solution and methanol to complete 1 mL; finally, each sample was diluted to 10 mL with ultrapure water. The absorption spectra were also recorded in the range of 200 to 300 nm against a blank. Their composition is described in Table 1. A synthetic set of samples was also prepared (Table 2), to evaluate the ability of prediction on the calibration model.


Table 1. Calibration set of samples used for the quantification of atrazine and dicamba in water by PLS-1 (concentrations are in µg mL-1).


Table 2. Synthetic mixtures used to evaluate the prediction capability of the calibration models (external validation), by PLS-1 (concentrations are in µg mL-1).

For the analysis of water, 250 mL of each sample was spiked with different amounts of the compounds of interest. Each sample was filtered with a nylon membrane (0.2 µm of pore size) and passed through a SPE cartridge, with a flow rate of 20 mL min-1. The content of the cartridge was dried with nitrogen and the analytes were eluted with 4 mL of methanol. The eluent was removed from the sample with N2. The herbicides were redissolved in 1 mL of methanol and transferred quantitatively to 10 mL volumetric flasks. Finally, the sample was prepared with the procedure described above.

RESULTS AND DISCUSSION

Figure 1 shows the absorption spectra of ATR and DIC under the experimental conditions described before, where a partial overlapping of both spectra is observed. Therefore, the simultaneous determination of these herbicides requires (a) the use of a separation technique before their detection, or (b) the application of a chemometric technique for the resolution of the binary system. The second option was chosen, owing to its simplicity, rapidity and low cost.


Fig. 1. Absorption spectra of ATR 2.2 mg mL-1 (D), DIC 5.3 µg mL-1 (' ) and their mixture in 2.6 y 5.9 mg mL-1, respectively (*), under the experimental conditions proposed.

Optimization of physico-chemical variables

The influence of pH in absorption spectra of ATR and DIC was studied initially. A solution of 250 mL containing 5 µg mL-1 of each analyte was prepared in KCl 0.1 M, which pH was varied in the range of 1.0 to 13.5 using HCl or KOH. The results obtained for DIC are presented in Figure 2, where can be seen small variations in absorption spectra as a function of pH. High absorbance values at pH > 11 were observed owing to the presence of KOH, not to hyperchromic effects in the spectrum of DIC. The results to ATR showed a similar tendency. A pH of 7 was selected in further experiments. A buffer solution of KH2PO4/NaKHPO4 0.1 M at pH 7 was used, since it presented absorbance values lower than 0.25 AU (l < 203 nm) in a 1:10 ratio with the sample. The buffer concentration was evaluated by adding volumes of it between 0.5 to 2.0 mL to the flasks; a volume of 1 mL was selected as optimum. Likewise, methanol was used in a 1:10 ratio with the sample, taking into account the limited solubility of atrazine in water and that this is an eluent common in SPE.


Fig. 2. Influence of the pH in the absorption spectra of DIC (5 µg mL-1 in KCl 0.1 M).

One component calibration

The analytical figures of merit were established for each single-component determination at the wavelengths corresponding to the maximum absorption of the herbicides. According to the Beer´s law, ATR exhibit a linear working range from 0.3 to 4.9 µg mL-1 at 222.2 nm and DIC from 0.8 to 8.8 µg mL-1 at 203.4 nm. Calibration functions and correlation coefficients (R) are presented in Table 3. The limits of detection (LD=3s/m) and quantification (LC = 10s/m) were calculated by means of the standard deviation of the analytical signals (s), at 222.2 nm for ATR or 203.4 nm for DIC, in a set of 10 blank samples; m represents the slope of each calibration curve. For the estimation of RSD, sets of 10 samples of ATR (2.9 µg L-1) and DIC (3.5 µg L-1) were also prepared (Table 3).


Table 3. Figures of merit of the spectrophotometric methods proposed to quantify atrazine and dicamba in water.

Resolution of the binary system by PLS-1

Table 1 describes the composition of the calibration set of samples used for this purpose, with the analytes present in different ratios. The design obeys to topics previously discussed, which has been applied for the resolution of several multicomponent systems7-9). Working conditions for PLS-1 were: (a) mean center as preprocessing strategy for spectral data, (b) cross internal validation leaving out one sample by iteration, (c) a maximum of 9 factors for the construction of the calibration model, and (d) a spectral range from 200 to 240 nm (201 independent variables).

PRESS (Prediction Error Sum of Squares) as a function of the number of factors was estimated, to identify the factors required in the construction of the calibration model; the criterions of the F-test and the first local minimum8-10), as well as cumulative variance 19), were taking into account during the optimization. Two factors were selected as optimum for ATR and three for DIC; the later required an additional factor probably because the presence of the absorption band of the blank in the same spectral region that this herbicide. The statistical parameters of R (the correlation coefficient between theoretical and estimated concentration values), SEC (standard error of calibration), RMSD (the average error index in the analysis) and REP (the error average percentage in the set) were calculated to evaluate the prediction capability of the calibration model by PLS-1; a summary of the results are in Table 4.


Table 4. Statistical parameters estimated during the internal validation of the calibration models proposed for ATR and DIC by PLS-1.

Then, the optimized calibration models proposed for ATR and DIC by PLS-1 were used to estimate the concentration of both compounds in synthetic mixtures (Table 2). The mean recovery percentages, SEP (the standard error of prediction) and REP (%) obtained for this set of samples are shown in Table 5, with good results in all cases. Comparison between SEC and SEP allows to identify an over- or sub-fitting calibration model, with more or less factors than strictly necessary. In this case, the magnitudes of SEC and SEP are similar, which confirm the adequate selection of factors in both cases. The REP (%) for calibration samples was higher than that for synthetic mixtures; the first set includes samples without one of the compounds, which highlight the differences between theoretical and estimated concentrations; even so, both values are of the same order for the two herbicides.


Table 5. Statistical parameters estimated during the external validation of the calibration models proposed for ATR and DIC by PLS-1.

Determination of ART and DIC in water

Samples of 500 mL of tap, well and sea water were stored in borosilicate containers at 4°C during less than seven days before the analyses. Aliquots of 250 mL were spiked with different amounts of the compounds of interest and analyzed by the proposed method. Some tap water samples was directly analyzed by spectrophotometry (without SPE), to test the prediction capability of the PLS-1 models with real samples (Figure 3). The rest of the set of tap, well and sea water samples were treated by SPE before the detection step (Figure 4). Comparison between expected and calculated concentrations is satisfactory in all cases.


Fig. 3. Determination of ATR and DIC by PLS-1 in tap, well and sea water, without SPE. Concentrations are in µg mL-1: ATR added () and founded (); DIC added ( ) and founded ( ).


Fig. 4. Quantification of ATR and DIC by PLS-1 in tap, well and sea water, previous SPE. Concentrations are in µg mL-1: ATR added ( ) and founded ( ); DIC added () and founded ( ).

CONCLUSIONS

The proposed method showed to be a rapid, easy and low cost alternative for the quantification of ATR and DIC in water by SPE and UV-Vis spectrophotometry. It could be used for the screening of these herbicides in water (e.g., in situ analyses) or as a quantification method in cases where the chromatographic ones cannot be implemented owing to cost limitations, lack of analytical instrumentation, etc. On the other hand, chemical or spectral interferences could be overcame successfully with the aid of PLS-1.

ACKNOWLEDGEMENTS

The Consejo Nacional de Ciencia y Tecnología (CONACyT, México) is thanked for financial support (project J13743E).

 

REFERENCES

1. U.S. Environmental Protection Agency. "Summary of atrazine risk assessment. http://www.epa.gov/oppsrrd1/reregistration/atrazine/ . Accessed 2004 Jan 15. (2002).         [ Links ]

2. J.T. Stevens, D.D. Summer. "Handbook of Pesticide Toxicology", Vol. 3. J. Wayland, JR. Hayes, R. Edward and JR. Laws (Eds.). Academic Press, San Diego (1991).         [ Links ]

3. A.C. Hogenboom, M.P. Hofman, D.A. Jolly, W.M.A. Niessen, U.A.Th. Brinkman. J. Chromatogr. A, 885, 377 (2000).         [ Links ]

4. R. Jeannot, H. Sabik, E. Sauvard, E. Genin. J. Chromatogr. A, 879, 51 (2000).         [ Links ]

5. V. Pichon, M. Charpak, M.C. Hennion. J. Chromatogr. A, 795, 83 (1998).         [ Links ]

6. S. Weigel, K. Bester, H.J. Hühnerfuss. Chromatogr. A, 912, 151 (2001).         [ Links ]

7. P.L. López-de-Alba, L. López-Martínez, K Wrobel-Kaczmarczyk, K. Wrobel-Zasada, J. Amador-Hernández. Anal. Chim. Acta, 330, 19 (1996).         [ Links ]

8. P.L. López-de-Alba, L. López-Martínez, K Wrobel-Kaczmarczyk, K. Wrobel-Zasada, J. Amador-Hernández. J. Radioanal. Nucl. Chem. Letters, 220, 167 (1997).         [ Links ]

9. J. Amador-Hernández, A. Cladera, J.M. Estela, P.L. López-de-Alba, V. Cerdá, Analyst, 123, 2235 (1998).         [ Links ]

10. J. Amador-Hernández, J.M. Fernández-Romero, M.D. Luque de Castro, Anal. Chim. Acta, 448, 61 (2001).         [ Links ]

11. J.R. Harmer, T.G. Callcott, M. Maeder, B.E. Smith. Fuel, 80, 1341 (2001).         [ Links ]

12. M. Blanco, J. Coello, H. Iturriaga, S. Maspoch, C. Pérez-Maseda. Anal. Chim. Acta, 407, 247 (2000).         [ Links ]

13. Y. Ni, L. Wang, S. Kokot. Anal. Chim. Acta, 431, 101 (2001).         [ Links ]

14. J. Amador-Hernández, J.M. Fernández-Romero, M.D. Luque de Castro. J. Anal. At. Spectrom., 15, 587 (2000).         [ Links ]

15. A.J. Handley, R.D. McDowall. "Extraction Methods in Organic Analysis", A.J. Handley (Ed.), CRC Press, Sheffield (1999).         [ Links ]

16. J. Haib, I. Hofer, J.M. Renaud. J. Chromatogr. A, 1020, 173 (2003).         [ Links ]

17. A. Sze'ka'cs, H.T.M. Le, F. Szurdoki, B.D. Hammock. Anal. Chim. Acta, 487, 15 (2003).         [ Links ]

18. R. Rial, B. Cancho, J. Simal. J. Chromatogr. A, 992, 121 (2003).         [ Links ]

"Multivariate Data Analysis for Windows 95/98 y NT, v. 3.10. User's Manual", Infometrix, Inc. (2002).         [ Links ]

Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons