Talanta, Vol. 31, No. 8, pp. 789-794, 1990 Printed in Great Britain

0039-9140/90 53.00 + 0.00 Pergimon Press plc

VOLTAMMETRIC DETERMINATION OF LINURON AT A CARBON-PASTE ELECTRODE MODIFIED

WITH SEPIOLITE

PEDRO HERNANDEZ, Jo% VICENTE, MARIA GONZALEZ and LUCAS HERNANDEZ

~p~mento de Quimica, U~ve~idad Autonoma de Madrid, Madrid 28049, Spain

(Received 6 February 1989. Revised 16 February 1990. Accepted 27 February 1990)

Summary-The determination of linuron by differential pulse voltammetry with a carbon-paste electrode modified with 20% w/w sepiolite has been studied. The linuron is preconcentrated under open-circuit conditions at pH 2.0. With 0.0144 potassium nitrate at pH 1.7 in the measurement cell, a sweep rate of 30 mV/sec and a pulse amplitude of 100 mV, an oxidation wave with a peak potential of 1.2 V is obtained. Under these conditions, determination limits of 75 ng/ml have been obtained, with a relative error of +2.8% and a relative standard deviation of 8.0%. The method has been applied to the direct determination of linuron in river water with no previous separation of the pesticide. Determination in sea-water is not possible, as chloride interferes at high concentration.

The use of magnesium silicates in chromato- graphic columns or in clean-up systems for the separation and subsequent determination of different pesticides’ takes advantage of the adsorption capacity of these clay minerals for certain organic compounds.

Carbon-paste electrodes modified with clays can easily be prepared and can be used for the oxidation or reduction of electroactive groups in adsorbed compounds. Such electrodes can be used for the determination of organic compounds which are difficult to determine electrochemically with mercury, carbon or metal electrodes, on account of signals that are difficult to interpret owing to capacitive or residual currents which may mask the analytical signal.

Such modified carbon-paste electrodes have been used for the determination of various organic compounds without prior separationt-

In the present work, the use of an electrode modified with sepiolite (a magnesium silicate) for determination of the organochlorine herbi- cide linuron, 3-[3,4-(dichlorphenyl)-l-methoxy- I-methylurea], was examined.

Linuron residues have been determined by gas chromatography with electron-capture,>’ or mass spectraP” detection, for example in air,‘O and in soya beans by HPLC.”

Various polarographic techniques have been applied to the determination of Linuron.‘* It is now shown that linuron can be directly

adsorbed from aqueous medium onto a carbon- paste electrode modified with sepiolite and then measured by anodic-stripping differential-pulp voltammetry.

EXPERIMENTAL

Apparatus

The electrode was prepared with carbon paste made from spectroscopic grade graphite (particle size ~42 pm) and nujol oil (1: 1 w/w). Sepiolite previously ground to a particle size of 0.2 pm was added to this under humid con- ditions. The paste was inserted in a 5-cm length of 1.4-mm bore polyethylene tube to a depth of 0.5 cm, along with a copper wire to achieve direct contact with the mixture. A Metrohm E-506 Polarecord was used for differential pulse (DPV) voltammetry and an Amel 448/A oscillo- polarograph coupled to a Hewlett-Packard X-Y recorder for cyclic voltammetry.

Two cells were used: one for preconcen- tration of linuron from solution and the other, containing the supporting electrolyte, the refer- ence electrode (SCE) and a platinum counter- electrode, for the stripping measurement.

Reagents

Stock solutions of linuron were prepared by dissolving the commercial product (96% pure) in methanol and diluting with water.

789

790 PEDRO HERNANDEZ et al.

Procedure

The modified electrode was placed in the preconcentration cell containing 20.0 ml of a solution of linuron and left under open- circuit conditions for a preselected time. The electrode was then washed with water and trans- ferred to the measurement cell and the DP voltamperogram was recorded between + 0.8 and +1.5 V.

RESULTS AND DISCUSSION

Figure 1 shows the cyclic voltammetric behaviour of an electrode modified with 20.0% sepiolite, kept for 5 min in a 50 ,ug/ml solution of linuron at pH 2, then transferred for measurement to a cell containing O.OlM potassium nitrate adjusted to pH 1.7 with nitric acid.

In the first sweep there is a wave at + 1.3 V and in the return sweep two waves appear at (a) 0.45 V and (b) 0.22 V; these are the reduction waves for the oxidation product of linuron.

In the second oxidative sweep two waves appear: (c) at 0.35 V and (d) 0.55 V correspond- ing to the oxidation steps for waves (a) and (b) respectively, the wave at + 1.3 V disappearing in this second cycle.

The other waves decrease in intensity in suc- cessive sweeps until a residual current level is reached. This phenomenon provides a method for regenerating the electrode without renewing the surface of the paste electrode.

This behaviour could be accounted for by analogy with what is known of the metabolism of linuron,13 that is, by assigning the oxidation wave at 1.3 V to the elimination of the methoxy group (replaced by a hydrogen atom) through the formation of formic acid. The reduction products observed (E = 0.4 V) would then be formaldehyde and methanol, which would sub- sequently be oxidized in a second sweep, and disappear from the surface of the electrode and into the solvent.

In DPV the same kind of behaviour is observed as in cyclic voltammetry (Fig. 2). In this case it is possible to regenerate the electrode with the same efficiency by keeping the electrode polarized at + 1.5 V for 2 min.

To establish the most suitable conditions for the determination of linuron, the variables that could affect each of the stages were studied by use of an electrode modified with 10% sepiolite, a 5.0 pg/ml linuron solution, a 6 min preconcentration time, O.OlM potassium nitrate as the supporting electrolyte for the

Fig. 1. Multi-sweep cyclic voltamperogram for linuron at 100 mV/sec.

Voltammetric determination of linuron 791

1 I I I I I I

0.2 0.4 0.6 0.8 1.0 1.2 1.4 E(V)

Fig. 2. Differential pulse voltamperogram for linuron, 5.0 pg/ml. (1) First, (2) second and (3) third scan: v = 30 mV/sec; BE = 100 mV.

DPV, a 30 mV/sec sweep rate and 100 mV pulse amplitude.

When the pH of the sample was varied by addition of nitric acid or potassium hydroxide, with measurement at pH 8 (O.OlM phosphate buffer) the highest peak current was reached for accumulation at pH 2 (Fig. 3a). When the pH in the measurement cell was varied, after accu- mulation at pH 2, the highest peak current was recorded for pH 1.7 (Fig. 3b). In neither of

I I

PH

Fig. 3. Influence of pH: (a) 4 in preconcentration cell; (b) ip measurement cell; (c) EP.

these modes of measurement did the peak po- tential (1.2 V) change, indicating that no pro- tons participate in the oxidation of linuron. A variation of the accumulation time (at pH 2.7 with measurement at pH 1.7) showed that for times above 9 min, ip remained constant (Fig. 4a).

With lower concentrations (0.3 pg/ml) the results pointed to the same kind of behaviour, although a constant ip was reached at 15 min (Fig. 4b). This confirms that the preconcen- tration stage is due to adsorption of the analyte onto the electrode, with saturation time increasing as the concentration of the analyte

Fig.

30-

:

520- .-o

a

10

/“/, 0 b

6 10 t$L,

4. Influence of accumulation time: (a) 5.0 pg/ml, (b) j~g/ml linuron.

0.3

792

Table 1. Sensitivity with different compositions

Table 2. ~nsiti~ty with different electrolytes at O.OlM concentration

Sepiolite, % W/W

IPT UA Electrolyte

5.0 1.8 10.0 2.3 20.0 6.9 40.0 No wave

decreases. Varying the proportion of sepiolite showed that maximum ip was attained with 20% w/w residual currents increasing for higher proportions (Table 1) and i,, decreasing for lower ones.

With the selected pH values and an accumu- lation time of 6 min, we studied the instrumental parameters with a pulse amplitude (AE) of 100 mV (Fig. 5a); iP increased with sweep rate to a maximum at 30 mV/sec. With a constant sweep rate of 30 mV/sec, ir varies linearly with AE = 100 mV (Fig. 5b).

A study of different electrolytes at a con- ~ntration of O.OlM and pH 1.7 (Table 2) showed that of those studied, potassium nitrate and perchlorate gave the largest peak heights, but with the latter, as successive measurements were made iP decreased until during the fifth measurement the oxidation wave disappeared and it was necessary to replace the electrode by another owing to interference by the perchlorate ion.

A parabolic plot of iP vs. potassium nitrate concentration was found, with maximum sensitivity at O.OlM, above which exchange of ~tassium with the sepiolite and a decrease in absorption power was evident.

Figure 6 shows the stripping curves and a calibration plot obtained with an accumulation time of 9 min; it was linear, and could be described by iP @A) = 11.7 $4.12C &g/ml) with a correlation coefficient (r) of 0.999. At the

0 80 AE (mv) k

Fig. 5. Infiuence of (a) scan rate; (b) pulse amplitude, on ip.

1 ie3lmf level, measurement with 5 different electrodes (10 dete~nations with each) indi- cated a relative error of + 1.7% and a relative

KNO, 25.0 KClO, 24.6 KC1 15.6 Na2HP0, 15.0 CH,COOK 10.1

0.8 1.0 1.2 E (VI

Fig. 6. Calibration voltamperograms and graphs obtained for linuron: (I) a, 1.0; b, 2.0; c, 3.0; d, 5 pg/ml. (II) Below

0.3 /rg/ml. R = residual current.

Voltammetric determination of linuron 193

1 I I I I

0.8 1.0 1.2 E(V)

the response of the electrode to concentrations of linuron below 0.3 pg/ml, using an accumula- tion time of 15 min. The equation for the linear calibration was then 6 (CIA) = 11.6 + 27.1C &g/ml) with r = 0.988.

Under these conditions the relative error was found to be +2.8% for 0.1 pg/ml, with a relative standard deviation of 8.0%; the limit of detection was 75 ng/ml.

The method was employed for the determina- tion of linuron in river water and sea-water with no prior separation step. Samples were acidified to pH 2 and analysed according to the pro- cedure described.

The voltammetric curves are shown in Fig. 7. In this figure, the difference between the two matrices can be seen. Whereas with river water (a) there is a perfectly defined peak, with a large difference between the blank and the sample, the salt concentration of sea-water (b) causes a poorer definition of the voltammetric peak owing to adsorption of the chloride ions present in the matrix. This effect was also observed when potassium chloride was tried as the sup- porting electrolyte, which is why use of this compound was discarded and replaced by use of the phosphate buffer. The effect of the chloride concentration is to distort the signal to the extent of preventing direct measurement in sea- water.

Interferences

Endosulfan, dinocap, phenol, aniline, nitro- benzene and tetramethin were examined and all found to interfere in the linuron determination. For environmental samples some sort of clean- up or preliminary separation would be neces- sary, to avoid interferences. On the other hand, the electrode could be used for determination of other pesticides.

Acknowledgemenr-The authors would like to express their gratitude to the CICYT for support for Project PA 86/0367, under the auspices of which this work was carried out.

tJ I I I I I

OS 1.0 1.2 E (‘4

Fig. 7. Determination of linuron in water: I, river; II, sea-water; a, residual current; b, without linuron; c, with

aqldition of 2 pg/ml linuron.

standard deviation of 5.0%. The limit of detec- tion was 0.3 pg/ml. With a view to improving the performance of the method, we also studied

REFERENCES

K. Eeyermann, Organic Trace Analysis, Horwood, Chichester, 1984. L. Hemlndez, P. Hernkndez, E. Lorenzo and Z. Sosa, Analyst, 1988, 113, 621. L. Hemhndez, P. Hemlndez and Z. Sosa, Z. Anal. Chem., 1988, 331, 525. L. Hemhndez, E. Gontilez and P. Herniindez, Analyst, 1988, 113, 1715.

194 PEDRO HERNANDEZ ef al.

5. V. D. Chmil, Zh. Analit. Khim., 1981, 36, 1813. 9. D. Barcelo, Chromatographia, 1988, 25, 295. 6. L. I. Knyr and V. P. Sukhoparova, Khim. Selsk. Khoz., 10. A. Gudehn and B. Kolmodin-Hedman, .I. Chromatog.,

1982, No. 8, 60. 1987, 387, 421. 7. A. P. Primark, K. V. Krivolutskii and N. F. Biryukova, 11. E. W. Zahnow, J. Agric. Food. Chem., 1987, 35, 403.

ibid., 1983, No.2, 58. 12. R. J. Hance, Pestic. Sci., 1970, 1, 112. 8. T. Tamiri and S. Zitrin, Biomed. Environ. Mass Spec- 13. K. A. Hassall, The Chemistry of Pesticides, p. 301.

tram., 1987, 14, 39. Verlag Chemie, Weinheim, 1982.