MBC (98% m/m) and 2-AB (>97% m/m) were purchased from Sigma-Aldrich Chemical Company (USA) and were used without further purification. Spectroscopic grade solvent including dichloromethane (DCM), acetoni- trile and dimethyl sulfoxide (DMSO) were obtained from Sigma Aldrich (France). Ethanol was purchased from Prolabo (France), and borate buffer solutions of different pH values (4, 7 and 10) were from Cluzeau Info Labo. The chemical structure of benzimidazoles is presented in Figure 1.

All spectral measurements were performed at room temperature with a Varian Cary Eclipse Fluorescence Spectrophotometer interfaced with a microcomputer and processed by software FL-WinLab. Standard quartz fluorescence with a 1-cm path lengths (Labo Moderne, France) were used for measurements and a Pipetman of 10 - 1000 µL (Gilson, France) was used for dilutions. The non-purgeable organic carbon (NPOC) was measured on Total Organic Analyzer Shimadzu (TOC-VCSH/CSN). The pH measurements were performed with a Consort pH-meter Model P107. The conductivity was recorded on CDM 92, 2-pole Conductivity Cells Radiometer Analytical.

The solvent effect on fluorescence spectra is of particular importance. A change in solvent is accompanied by a change in polarity of the surrounding medium. We have studied the fluorescence emission spectra of MBC and 2-AB in several organic solvents of different polarities and checked the pH effect in aqueous solutions.

We investigated the fluorescence behaviour of MBC and its metabolite 2-AB in several solvents of strongly different polarities, including, dichloromethane, ethanol, acetonitrile, DMSO. The results concerning the MBC and 2-AB spectral and chemical properties (λ_ex, λ_em and I_F) in these different media are reported in Table 1. As can be seen, these spectral and chemical parameters were found to vary more or less importantly with the medium.

The excitation maximum wavelengths presented two main bands in the regions 218 - 260 and 275 - 290 nm, whereas the emission maximum wavelengths occurred in the range 300 - 323 nm, according to the solvent and the compound (Figure 2 and Table 1). Therefore, taking into account the important changes of solvent polarity, we can consider that the λ_ex and λ_em, values only weakly varied with the medium polarity, except the case of water (pH = 7) for which an important blue-shift of λ_ex over than 56 nm is observed.

The emission band of MBC shows minor variations with increasing solvent polarity; the maximum bathochromic shift being 12 nm. The emission band for 2-AB is more sensitive to solvents and presents a maximum red-shift value of 23 nm for the different solvents. The fluorescence intensity is much higher in DMSO and acetonitrile than the others media (Table 1). For analytical purposes, DMSO and acetonitrile were chosen as optimal solvents.

On the other hand, the variation of the fluorescence intensity of MBC and 2-AB as a function of the pH (curve not shown) was characterized by well defined maximum values of pH obtained at acid medium for MBC and neutral for 2-AB. The effect of pH has allowed us to know that the analysis of MBC and 2-AB by fluorescence in water is most appropriate in a strongly acid (pH = 4) and neutral (pH = 7) medium, respectively, for MBC and 2-AB.

In order to select the most interesting conditions from the analytical standpoint, we optimized our spectrofluori- metric method for MBC and 2-AB determination by comparing the I_F values in the various media under study. Relative to the fungicides fluorescence signal in DCM, which was used as a reference medium, the highest I_F

^aWavelength of fluorescence excitation maxima (nm); ^bWavelength of fluorescence emission maxima (nm); ^cRelative fluorescence intensity, normalized to the fluorescence intensity in DCM.

values were found in DMSO, acetonitrile and water at pH = 4 and 7. Therefore, for comparison purposes, we decided to evaluate the analytical figures of merit of the fungicide MBC and its metabolite in these selected media, namely DMSO, acetonitrile and in water at pH = 4 and 7 (Table 2). Linear plots of fluorescence intensity vs. MBC or 2-AB concentration I_f = f (c) were obtained by fluorescence method. The limits of detection (LOD) values were determined on the acetonitrile, DMSO and water (pH = 4 and 7) from a MBC and 2-AB concentration giving a signal-to-noise (S/N) ratio of 3 (IUPAC criterion). The limit of quantification (LOQ), is determined for a signal-to-noise (S/N) ratio of 10.

The data summarized in Table 2 indicates that linear calibration curves of fluorescence intensity vs. MBC and 2-AB concentration were obtained with large linear dynamic range (LDR) values of about two orders of magnitude, with r² values very close to unity showing a good linearity of the analytical curves. The concentration effect on fluorescence spectra of the MBC and 2-AB in acetonitrile are presented in Figure 3.

The reproducibility of measurements, evaluated with freshly-prepared MBC and 2-AB solutions, was satisfactory, as indicated by the small relative standard deviation (RSD) values less than 1%. The LODs were significantly low, ranging from 0.01 to 0.06 ng/mL for MBC and 0.002 to 0.01 ng/mL for 2-AB. The corresponding limit of quantification (LOQ) were 0.03 - 0.2 ng/mL for MBC and 0.006 - 0.03 ng/mL for 2-AB (Table 2).

^aLimit of detection was defined as the amount of analyte giving a signal-to-noise ratio of 3; ^bLimit of quantification defined as the amount of analyte giving a signal-to-noise ratio of 10; ^cLinear dynamic range; ^dRelative standard deviation (n = 6); ^eCorrelation coefficient.

Our direct fluorimetric method developed, has limits of detection lower than those reported by authors using various methods, such as dispersive liquid-liquid microextraction (DLLME) combined with UV-Vis spectrophotometry (LOD = 2.1 ng/mL) [9] , fluorescence method (0.125 ng/mL) [10] , DLLME-HPLC method with fluorimetric detection, (LOD = 0.5 ng/mL) [11] , spectrofluorimetric method in hydroxypropyl-β-CD solutions (LOD = 17.4 ng/mL) [12] , SLM-MMLLE-HPLC with UV detection (LOD = 0.1 ng/mL) [13] . Consequently, the LOD and LOQ values obtained in this work were very low, suggesting that direct fluorescence can be considered as a suitable and sensitive method for determining MBC and 2-AB in natural waters.

Solvent effect has shown that MBC and 2-AB have high fluorescence signal in DMSO, acetonitrile and water (pH = 4 and 7). Because of its viscosity, DMSO can neither be used to pack the column nor to elute MBC and 2-AB. The elution of the two benzimidazoles with water (pH = 4 and 7) gives also poor percentages of recovery with spectra deformation. As a result, we used acetonitrile as packing and elution solvent to extract the fungicide MBC and 2-AB by SPE.

We studied the effect of the volume of the elution solvent on the extraction efficiency for different volumes of the solvent acetonitrile (0.5, 1, 2, 3, 4, 5, 5.5, 6 mL) and the recovery percentage was determined at the same time. Figure 4 illustrates the variation of the extraction recovery versus the volume of the elution solvent. The percentage of recovery was increased with the increase of the volume of elution and became constant when the elution volume of acetonitrile exceeded 4 mL. The maximum elution volume of 4 mL was chosen to elute the MBC fungicide and its metabolite 2-AB separately.

To verify the applicability of the proposed direct spectrofluorimetric method to authentic samples, MBC and 2-AB were determined in natural water samples after a Solid-Phase extraction (SPE) procedure. For recovery studies, 5 mL of water samples were spiked at three different concentration levels of each pesticide. We used acetonitrile as a packing solvent and as a solvent of elution to extract the fungicide MBC and 2-AB. Acetonitrile gave good percentage recovery. The results obtained in the different water samples are satisfactory because the recovery rate varies between 88.5% and 119.2% (Table 3).

On the other hand, the physicochemical parameters of the natural water samples including the dissolved organic carbon NPOC (mg/L), the conductivity s (ms/cm) and the pH (Table 4) seems to play some effects on the MBC and 2-AB recovery. Indeed, after analysis we have noticed that the draining and well waters (d = 1.5 m) have high recoveries rate relative to the others water samples. The more important recoveries rate in draining water is probably due to the excess of dissolved organic matters (NPOC = 103.5 mg/L) which is at least ten or three times more high. The increase of recovery percentage with the amount of dissolved organic matters has been already noted by several authors [14] [15] .

Since, the amount of dissolved organic matters in the natural water samples plays a certain role on the rate of recovery of both benzimidazoles, we decided to study the effect of foreign species and of some inorganic ions on the fluorescence intensity of the MBC in water (pH = 4) and of 2-AB in water (pH = 7).

The effect of some pesticides such as benomyl, thiabendazole (commonly-used in Senegal) and several inorganic ions including Na⁺, Ca²⁺, K⁺, Cl⁻, , , (likely to exist in natural waters) on the fluorescence intensity was investigated for possible interference on the determination of MBC and 2-AB. We fixed the concentration of MBC or 2-AB, and we evaluated the effect of increase amounts of the interfering species on the fluorescence intensity in appropriate ranges. The tolerance limit of the interfering species was defined as its concentration of which the percentage of fluorescence signal variation did not exceed ± 5% in the determination of MBC and 2-AB. We calculated the percentage of variation of the fluorescence signal, using the following equation for each concentration of interfering species.

d = depth of the well water.

^aNPOC (mg/L) = dissolved organic carbon, corresponding to the non-purgeable organic carbon present in natural waters. Calculated from the equation: NPOC = TOC − POC, with NPOC = non-purgeable organic carbon, TOC = total organic carbon and POC = purgeable organic carbon; ^bσ (mS/cm) = conductivity of the water samples.

The tolerance limit obtained for the selected foreign species are presented in Table 5. Fixed concentrations of 2 ng/mL of MBC and 0.6 ng/mL of 2-AB were used to evaluate the effects of interference for the concentration ranges of selected foreign species surrounding the concentrations of MBC and of 2-AB.

An increase or decrease in intensity was observed during the addition of pesticides or inorganic ions in the studied concentration range. The observed phenomenon may be due to direct interactions between foreign species (pesticides or inorganic ions) with MBC or 2-AB. Our results showed that all tested foreign species have interference on the MBC and the 2-AB. The most important effect of interference on the MBC was observed with benomyl (tolerance limit = 0.05 ng/mL) and 2-AB (tolerance limit = 0.07 ng/mL). For the 2-AB, the effect of the higher interference was observed with benomyl (tolerance limit = 0.06 ng/mL) and MBC (tolerance limit = 0.06 ng/mL). These limits correspond to concentrations about 40 times lower than that of MBC and 10 times

lower than that of 2-AB. These results can be explained by the fact that benomyl is a compound that is metabolized to MBC during ordinary analytical procedures in organic solvents and water [2] [16] [17] . The 2-AB is also a metabolite of MBC, and the three benzimidazoles emit at the same wavelength region. These relatively high interference phenomena require a significant improvement in selectivity when applying the fluorescence method to determine the MBC and 2-AB in the natural water samples containing the interfering species (pesticides or organic ions). To reduce the impact of the effects of interference and improve the selectivity of the fluorescence method, we believe it would be possible to use spectral method including synchronous fluorescence [18] derivative fluorescence [19] or partial least-square (PLS) method [20] .