Electrochemical oxidation of sildenafil citrate (Viagra) on carbon electrodes

Analytica Chimica Acta 501 (2004) 227–233

Electrochemical oxidation of sildenafil citrate (Viagra) on carbon electrodes Sibel A. Özkan a , Bengi Uslu a , Petr Zuman b,∗ a

Department of Analytical Chemistry, Faculty of Pharmacy, Ankara University, 06100 Ankara, Turkey b Department of Chemistry, Clarkson University, Potsdam, NY 13699-5810, USA Received 8 May 2003; received in revised form 9 September 2003; accepted 23 September 2003

Abstract In aqueous solutions containing 30% (v/v) acetonitrile the unprotonated form of the piperazine ring of sildenafil (I) is oxidized on a glassy carbon electrode between pH 2 and 8. Whereas citrate anions under conditions used are not oxidized, the oxidation pattern of sulfonamides differs from that of sildenafil (I). On the other hand, the oxidation of I resembles that of the oxidation of piperazine ring in nefazodone (III) and trazodone (IV). A more positive peak of sildenafil corresponds to an adsorption–desorption process. Measurement of the peak current enables rapid determination of sildenafil (I) in pharmaceutical preparations with sufficient accuracy. © 2003 Elsevier B.V. All rights reserved. Keywords: Sildenafil citrate; Oxidation mechanism; Drug analysis; Voltammetry

1. Introduction Viagra, sildenafil citrate (I), which is used to control erectile disfunction, is chemically 1-[{3-(6,7-dihydro-1-methyl7-oxo-3-propyl-1-H-pyrazolo[4,3-d]pyrimidin-5-yl)-4-ethoxyphenyl} sulfonyl]-methyl piperazine citrate (I).

Among methods used for determination of I belong uses of LC, for example, in acetonitrile–water mixture with phosphate buffer (detection at 230 nm) [1] or at 245 nm [2] or with ammonium acetate (detection at 240 nm) [3]. Alternative methods used were the micellar electrokinetic chro∗ Corresponding author. Tel.: +1-315-268-2389; fax: +1-315-268-6610. E-mail address: [email protected] (P. Zuman).

0003-2670/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2003.09.033

matography [4,5], spectrophotometry, based on function of ion-pairs of I with dyes which can be extracted [6], or flowinjection analysis with UV detection [7]. Commonly used separation techniques are undoubtedly superior, when I is to be determined in the presence of impurities from manufacture or of metabolites. But it can be questioned, whether such procedures present the fastest, most accurate and most sensitive analytical methods for determination of the drug I in pharmaceutical preparations. Under specific circumstances, electrochemical methods can offer optimal solution. Perhaps more than in other analytical techniques, to make the best use of electroanalytical procedures, it is essential that both the physical and chemical processes involved in the measured signal are understood. This condition was not fulfilled in the investigation [8] in which the authors proposed a highly sensitive method for determination of compound I. This procedure was based on application of square wave and adsorptive stripping SWV using a hanging mercury drop electrode. The nature of the processes manifested in current peaks observed in solutions of perchloric acid of pH 2 was not identified. Furthermore, in protic solvents where electron transfers are accompanied by proton transfers, the important role of pH was not reported or discussed. Resulting empirical analytical method may be prone to unexpected matrix effects.

228

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

The authors [8] only indicated that electrode processes involved are irreversible and adsorption controlled. The accumulation of the species for stripping analysis has been carried out at a potential more by 200 mV more positive than the observed peak current. This strongly indicates that it is the oxidized form of I which is adsorbed and thus the more positive peak on their current–potential curves most probably does not involve a Faradaic process. Hence the assumption [8] that “The mechanism of reduction could be due to the sulfonyl group bound to the piperazine ring [to yield a] sulfoxide groups” is not supported by the available experimental evidence. To demonstrate the usefulness of a solid electrode for determination of I, which may offer advantages for the use of such electrodes as sensors, the electrochemical behavior of I on a glassy carbon electrode was investigated in this contribution.

2. Experimental

Table 1 Dependence of peak currents (ip1 and ip2 ) and peak potentials (Ep1 and Ep2 ) of anodic waves of 1 × 10−4 M sildenafil citrate on glassy carbon electrode in solutions containing 30% (v/v) acetonitrile Solution or buffer

2.2. Reagents Sildenafil citrate (I) and its pharmaceutical formulation were kindly provided by Eczacibasi Pharm. Ind. (Istanbul, Turkey) and nefazodone (III) and trazodone (IV) were kindly supplied by Bristol-Myers-Squibb (Istanbul, Turkey) and Santa Farma Pharm. Ind. (Istanbul, Turkey), respectively. p-Aminobenzenesulfonamide (II) and citric acid were supplied by Merck. All chemicals for preparation of buffers and supporting electrolytes were reagent grade. Acetonitrile was chromatographic purity. Stock solutions of I, II, III and IV (1 × 10−3 M) were prepared in acetonitrile, and kept in the dark in a refrigerator. All stock solutions were protected from light and were used within 24 h after preparation to avoid decomposition.

ip1 (␮A)

Ep1 (V/SCE)

ip2 (␮A)

Ep2 (V/SCE)

0.5 M H2 SO4 0.3 M H2 SO4 0.2 M H2 SO4 0.1 M H2 SO4 0.03 M H2 SO4 b 0.01 M H2 SO4 b

0.3a 0.5a 0.7a 1.0a 1.5a 2.0a

– – – – 4.0 4.0

– – – – 1.21 1.17

10.7 10.7 11.3 10.8 12.7 12.0

1.38 1.38 1.37 1.36 1.35 1.32

Phosphate

2.0 2.5 3.0

3.3 3.5 4.0

1.18 1.17 1.13

10.0 12.0 12.0

1.34 1.33 1.30

Acetate

3.6 4.0 4.7 5.0 5.5 5.8

4.0 4.1 4.5 4.3 4.3 4.4

1.15 1.07 1.05 1.03 1.00 0.99

14.0 14.3 12.7 11.7 11.7 12.0

1.30 1.25 1.23 1.22 1.20 1.21

Phosphate

5.8 6.5 6.8 7.0 7.4 7.8

4.7 4.0 4.0 4.7 3.6 3.3

0.99 0.99 1.00 0.97 0.99 0.99

9.3 7.3 6.7 6.7 5.7 3.7

1.25 1.24 1.23 1.20 1.20 1.21

Ammonia

8.3 8.8 9.3 9.8 10.3

3.7 3.7 3.45 4.25 4.0

1.03 1.04 1.04 1.04 1.04

– – – – –

– – – – –

2.1. Apparatus The cyclic, linear sweep, DPV (differential pulse) and SWV (square wave) voltammetric experiments at a stationary solid electrode were performed using a BAS 100 W electrochemical analyzer. A three electrode cell system incorporating the glassy carbon disc electrode as working electrode, an Ag/AgCl (3 M KCl) reference electrode and a platinum-wire auxiliary electrode were also used. Before each measurement the glassy carbon electrode was polished manually with alumina (∅ = 0.01 ␮m) dispersed in bidistilled water on a smooth polishing cloth. Operating conditions for SWV were: pulse amplitude, 25 mV; frequency, 15 Hz; potential step, 4 mV and for the DPV were: pulse amplitude, 50 mV; pulse width, 50 ms; scan rate, 20 mV s−1 . The pH was measured using a pH meter Model 538 (WTW, Austria) using a glass electrode with an accuracy of ±0.05 pH.

pH

a b

Approximate pH values. Solution contains 0.1 M Na2 SO4 .

The working solutions for voltammetric investigations were prepared using double distilled water by a dilution of the stock solution by the selected supporting electrolyte so that (if not otherwise stated) the final solution contained 30% (v/v) acetonitrile, supporting electrolyte, and 0.1 mM electroactive species. The following supporting electrolytes used are given in Table 1. Current–potential curves of sample solutions recorded after 72 h after preparation did not show any appreciable change in assay values. 2.3. Tablet assay procedure Ten tablets were reduced to a homogeneous fine powder in a mortar. An aliquot of this powder corresponding to about 0.1 mM stock solution was accurately weighed and transferred into a 50 ml calibrated flask and filled up to the volume with water. The contents of the flask were sonicated for 10 min to achieve complete dissolution. Analyzed solutions were prepared by taking aliquots of the clear supernatant and diluting with the selected supporting electrolyte. Voltammograms were recorded in a similar way as for standard solutions of I.

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

229

2.4. Recovery experiments from tablets In order to exclude interferences by the excipients, known amounts of the pure drug were added to the different preanalyzed formulations of I and the mixtures were analyzed by the proposed methods. The recoveries were determined based on five parallel analyses.

3. Results and discussion Sildenafil (I) is manifested on current–voltage curves recorded by cyclic voltammetry (CV) on a glassy carbon electrode by two anodic peaks (ip1 and ip2 ) and one cathodic peak (ipc ) (Fig. 1). Only anodic processes will be discussed in this contribution. The peak potentials (Ep1 and Ep2 ) and the peak currents (ip1 and ip2 ) were determined in supporting electrolytes containing 30% (v/v) acetonitrile to maintain solubility (Table 1). The less positive peak ip1 corresponds to an electrooxidation of I and is diffusion controlled. This is confirmed by a linear dependence of peak current ip1 on concentration of I and on its linear dependence on ν1/2 (Fig. 2a). The height of peak ip1 remains practically pH-independent up to pH 10.3. No oxidation current was observed in phosphate buffers pH > 10.5. The peak potential of ip1 remains practically pHindependent at pH higher than about 5.5. At lower pH values potentials Ep1 are shifted to more positive values with decreasing pH (Figs. 3 and 4). It is a general rule that a conjugate base is oxidized at less positive potentials than the corresponding acid form. Thus the observed pH-dependence indicates that the electroactive grouping responsible for the oxidation process in ip1 is in acid–base equilibrium with pKa about 5.5. At pH > pKa the conjugate base predominates in the bulk of the solution. As no dissociation occurs before the electron transfer, the oxidation potential remains pH-independent. At pH < pKa , the conjugate base must be formed by a rapid dissociation of the protonated form. The use of additional energy is manifested by the shift to a more positive potential. The increase in the slope at pH < 5.5 indicates presence of an antecedent acid base equilibrium with pKa about 5.5. This corresponds to the second pKa value of unsubstituted piperazine (5.68) indicating that the electroactive form contains the monoprotonated form of the piperazine ring. The deviation from the linear Ep–pH plot with a slope of 0.065 V/pH at pH < 1 indicates that in these strongly acidic solution the dissociation might not be fast enough. Complex behavior in borate buffers pH 8.3–10.3 was considered outside the scope of the present study. To support the working hypothesis that it is the piperazine ring in I that undergoes oxidation, the behavior of anodic peaks ipa of I was compared with that of some model compounds. There are three main groupings present in the structure of I which might be considered as under-

Fig. 1. Cyclic voltammograms of 0.1 mM solution of sildenafil citrate (I) using a glassy carbon electrode. Current–voltage curves recorded from 0.0 V (Ag/AgCl electrode) to +1.6 V and back to −1.6 V. Scan rate: 100 mV/s. (a) In acetate buffer of pH 4.7 containing 30% MeCN and (b) in phosphate buffer of pH 6.8 containing 30% MeCN.

going electro-oxidation: the citrate anion, the p-alkoxybenzenesulfonamide grouping and the piperazine ring. In the presence of citrate anions no anodic waves were obtained in buffers at pH 2–10 and hence the oxidation of the anions of citric acid cannot be responsible for the observed anodic waves of I. As a model for the sulfonamido grouping the p-aminobenzenesulfonamide (II) was used. When comparing the structure of II with that of I, it is observed that the p-alkoxy group in I is replaced in II by a p-amino group. Such replacement—apart from introduction of another acid–base system—should not have a pronounced effect on the mechanism of the oxidation involved. Sulfonamide (II) is oxidized in two steps: the peak current of the more positive step, ip1 (Table 2), is comparable

230

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

Fig. 4. Comparison of pH-dependences of peak potentials of 0.1 mM solutions of model compounds. p-Amino benzene sulfonic acid (䊐), sildenafil citrate (䊊), nefazodone () and trazodone (䉫). Scan rate: 100 mV/s.

Fig. 2. Dependences of peak currents of 0.1 mM solutions of sildenafil citrate (I) on scan rate. Dependence of peak current (a) ip1 at 1.20 V on square root of scan rate from 5 to 400 mV/s in acetate buffer of pH 5.8 containing 30% MeCN and (b) ip2 at 0.98 V on scan rate from 10 to 150 mV/s.

to that of compound I and most probably corresponds to a transfer of the same number of electrons. The potential range, in which the sulfonamide (II) is oxidized is comparable to that observed for oxidation in peak ip1 of I. But

Fig. 3. Dependence of peak potentials obtained in a solution of 0.1 mM sildenafil (I) on pH. Empty symbols: peak ip1 ; full symbols: peak ip2 . Supporting electrolyte: (䊐) sulfuric acid, (䊊) phosphate buffers and () acetate buffers. Scan rate: 100 mV/s.

the slope of the pH-dependence of Ep of II is different from that of Ep1 of I (Fig. 4). Most importantly, the pHindependent range of the plot of Ep = f (pH), which for piperazine derivatives was observed at pH > 4, is missing in the plot for II. That indicates pKa of the acid dissociation, preceding the electron transfer, is for II larger than 7. The decrease in the slope dEp/dpH at pH < 3.5 indicates that at lower pH the rate of dissociation of the protonated form is no more extremely fast. These results indicate that the electro-oxidation of I most probably does not involve changes in the sulfonamido grouping. As model substances for oxidation of the piperazine ring, were used two drugs, nefazodone (III) and trazodone (IV). The electro-oxidation of the piperazine moiety of IV has already been reported [9]. Compound III is oxidized over most of the pH range in a single wave ip1 , at potentials by about 0.2 V less positive than those of peaks ip1 of I (Fig. 4). The heights of these peaks of II remain practically unchanged up to pH 7. The peak potentials Ep1 are shifted with increasing pH to less positive potentials similarly as Ep1 of I, and became pH-independent at pH higher than about 6. So again, the monoprotonated form of the piperazine ring, rapidly generated from the dication, is oxidized at pH < 6. The pKa of the dissociation of the diprotonated form for II is at about 6.0, not too different from the pKa 5.5 observed for I. Only at pH 5 and 5.5 another wave was observed at more positive potentials. The nature of this electrode process was not further investigated. Trazodone (IV) is oxidized in two steps ipa and ip1 . The less positive peak ipa is higher and corresponds to an adsorption of the oxidized form of IV. Peaks ip1 of IV are comparable to peaks ip1 of I and III as they show two linear segments on the Ep1 = f (pH) plot with slopes dEp/dpH equal to 0.06 V/pH at pH < 4.7 and 0.00 V/pH at higher pH. The intercept at pH 4.7 corresponds to the pKa of the dissociation of the diprotonated piperazine ring into a monoprotonated form. The pH independence of Ep1 at pH < 3 indi-

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

231

Table 2 Dependence of anodic peak currents ip1 and ip2 and peak potentials Ep1 and Ep2 of waves of 1 × 10−4 Ma Buffer

pH

Compound II

Compound III

Compound IV

ip1 (␮A)

Ep1 (V/SCE)

ip2 (␮A)

Ep2 (V/SCE)

ip1 (␮A)

Ep1 (V/SCE)

ip2 (␮A)

Ep2 (V/SCE)

ip1 (␮A)

Ep1 (V/SCE)

ip2 (␮A)

Ep2 (V/SCE)

Phosphate

1.66 2.00 2.50 3.03

3.75 4.2 4.3 4.2

1.10 1.08 1.08 1.07

– – – –

– – – –

5.4 6.7 6.7 6.7

0.96 0.95 0.93 0.92

– – – –

– – – –

7.5 8.0 8.0 8.0

0.98 0.97 0.95 0.94

5.8 6.0 5.7 6.0

1.20 1.19 1.20 1.20

Acetate

3.51 4.00 4.70 5.50 5.80 6.50 6.80

4.3 3.7 4.0 3.7 4.0 4.0 3.3

1.05 1.04 1.00 0.96 0.94 0.91 0.90

– – 7.3 6.7 8.0 8.0 7.3

– – 1.38 1.36 1.42 1.44 1.39

4.0 4.7 5.0 5.5 6.2 5.7 6.0

0.94 0.88 0.85 0.83 0.84 0.82 0.80

– – 3.75 3.75 – – –

– – 0.96 0.93 – – –

5.7 6.3 6.25 5.75 6.70 6.25 6.0

0.93 0.90 0.82 0.84 0.82 0.81 0.80

5.25 5.7 5.75 5.75 6.0 5.9 5.3

1.16 1.15 1.10 1.10 1.13 1.13 1.12

a

p-Aminobenzene sulfonamide (II), nefazodone (III) and trazodone (IV) in buffers containing 30% (v/v) acetonitrile.

cates that in this pH range the dissociation of the dication becomes slower. This occurs in a pH range one or more pKa units higher than the pKa value of I. These results strongly indicate that in I the monoprotonated form of the piperazine ring is oxidized. It remains to comment on the second (ip2 ), more positive peak of I. The sharp form of this peak (Fig. 5) already indicates that adsorption is involved. This has been confirmed by the non-linear, limiting shape of the plot of ip2 = f [Sildenafil] as well by the linear dependence of ip2 on the scan rate (ν) at higher concentration of I (Fig. 2b). As the peak ip2 occurs at potentials more positive than that of ip1 , it is strongly indicated that it is the oxidized form of I which is adsorbed. The peak current ip2 remains practically constant up to pH about 5.5 and shows a gradual decrease at higher pH values. The plot of ip2 = f (pH) has a shape of a dissociation curve with an inflexion point at about pH 7 (Fig. 6). A conjugate base is always more easily oxidized than the corresponding acid form, the anodic; current of an oxida-

tion accompanied by an antecedent acid–base equilibrium therefore always increases with increasing pH. The observed decrease of the anodic current with increasing pH thus cannot be attributed to an acid–base reaction occurring before the electron uptake. The observed dependence of ip2 on pH (Fig. 6) can be thus attributed to a difference between the adsorption of the protonated form of the oxidation product of I and its conjugate base. The conjugate acid of the oxidation product is more strongly adsorbed. This oxidation product has at the electrode surface a pKa of about 7, somewhat higher than the pKa of the reduced form of the piperazine derivative I, which is about 5.5.

4. Analytical applications Based on the above study, the best condition for analytical applications proved to be either a phosphate buffer of pH 2.0 or an acetate buffer of pH 3.5. For the determination

Table 3 Regression data of the calibration lines for quantitative determination of SIL in pH 2.0 phosphate buffer and pH 3.5 acetate buffer using CV, DPV and SWV Phosphate buffer (pH 2.0)

Measured potential (V) vs. Ag/AgCl Linearity range (M) Slope of calibration graph (␮A M−1 ) Intercept (␮A) Correlation coefficient (r) RSD of slope RSD of intercept Number of data points LOD LOQ Repeatability of peak current (RSD%) Repeatability of peak potential (RSD%) Reproducibility of peak current (RSD%) Reproducibility of peak potential (RSD%)

Acetate buffer (pH 3.5)

CV

DPV

SWV

CV

DPV

SWV

1.36 4 × 10−5 to 3 × 10−4 5.42 × 104 3.433 0.998 1.04 0.99 6 5.69 × 10−6 1.90 × 10−5 0.62 0.14 0.97 0.30

1.30 6 × 10−6 to 3 × 10−4 3.91 × 104 0.453 0.999 1.51 0.95 10 6.92 × 10−7 2.31 × 10−6 1.16 0.17 1.58 0.40

1.34 6 × 10−6 to 3 × 10−4 4.79 × 104 0.350 0.999 1.33 1.13 10 6.28 × 10−7 2.09 × 10−6 0.44 0.21 1.25 0.44

1.30 8 × 10−6 to 2 × 10−4 3.26 × 104 0.225 0.997 1.70 1.02 8 1.62 × 10−6 5.40 × 10−6 1.00 0.17 1.29 0.22

1.25 4 × 10−6 to 3 × 10−4 1.59 × 104 0.290 0.998 1.53 0.91 11 1.05 × 10−6 3.51 × 10−6 1.02 0.43 1.55 0.61

1.28 4 × 10−6 to 3 × 10−4 1.89 × 104 0.192 0.999 1.05 1.05 11 1.11 × 10−6 3.71 × 10−6 0.67 0.17 1.33 0.22

232

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

Fig. 5. Cyclic voltammograms of a 0.1 mM solution of sildenafil (I) in: (a) phosphate buffer of pH 3.03 containing 30% MeCN; (b) acetate buffer of pH 3.51 containing 30% MeCN; (c) acetate buffer of pH 4.7 containing 30% MeCN; (d) acetate buffer of pH 5.8 containing 30% MeCN. (1) curve of the supporting electrolyte; (2) curve in the presence of sildenafil. Scan rate: 100 mV/s.

using glassy carbon electrode the current–potential curves were recorded using either CV, differential pulse voltammetry (DPV) or square wave voltammetry (SWV). Calibration curves were used for evaluation of concentration within ranges given in Table 3. The detection limit (LOD) was obtained using equation LOD = 3s/m, and the quantification limits (LOQ) using LOQ = 10s/m, where s is the standard deviation obtained in five runs and m the slope of the dependence of measured current on concentration [10,11]. The repeatability (on the same day) and interday reproducibility of both peak potentials and peak currents reported in Table 3 are based each on five experiments using 0.1 mM solution of I. Behavior in both supporting electrolytes was tested using all these techniques (Table 3). Repetition of

sample analysis after 72-h period did not show any significant change in results of analyses. Measurements of peak currents under above mentioned conditions were successfully applied the determination of I in spiked Viagra® tablets and no interference by excipients were observed. Viagra® tablets contain in addition to the active ingredient, sildenafil citrate, also the following inactive ingredients: microcrystalline cellulose, anhydrous dibasic calcium phosphate, croscarmellose sodium, magnesium stearate, hydoxypropyl methylcellulose, titanium dioxide, lactose, triacetin and FD&C Blue #2 aluminum lake. There is no official method in any pharmacopoeia (e.g. USP, BP or EP) related to tablet dosage forms or bulk drugs of I. To prove the absence of interference by excipients, recovery

S.A. Özkan et al. / Analytica Chimica Acta 501 (2004) 227–233

233

Table 4 Tablet assay results and mean recoveries obtained for five determinations of SIL in spiked Viagra® tablets Phosphate buffer (pH 2.0)

Labeled claim (mg) Amount found (mg)a RSD% 95% Confidence limit Added (mg) Found (mg) Recovered % RSD% of recovery a

Acetate buffer (pH 3.5)

CV

DPV

SWV

CV

DPV

SWV

50.00 47.82 1.11 0.65 5.00 4.966 99.33 1.23

50.00 48.93 0.96 0.58 5.00 4.970 99.40 0.60

50.00 49.45 1.13 0.69 5.00 4.976 99.52 0.34

50.00 49.70 1.15 0.71 5.00 4.970 99.38 0.87

50.00 49.40 0.57 0.35 5.00 4.930 98.50 0.80

50.00 49.33 0.68 0.42 5.00 4.920 98.30 1.04

Each value is the mean of five experiments.

of peak currents allows determination of the drug in tablet dosage forms. The analytical procedure has been validated (Tables 3 and 4). Depending on the voltammetric method used, the analyzed solution of the sample may contain between 0.004 and 0.3 mM electroactive species. The proposed methods are simple, fast and low cost suitable for I analysis in pharmaceutical formulations. The recovery results prove that the proposed procedures are sufficiently accurate and precise and can be applied to pharmaceuticals.

References

Fig. 6. Dependence of the peak current ip2 on pH in 0.1 mM solutions of sildenafil (I). Scan rate: 100 mV/s.

studies were carried out using the standard addition method. Recovery tests were carried out after known amounts of the pure drug were added various pre-analyzed formulations of I (Table 4). The results (Table 4) demonstrate the validity of the proposed method for the determination of I in commercial tablet dosage forms. No significant difference was observed between the performance of individual proposed methods with regard to accuracy and precision. 5. Conclusions The developed electroanalytical method for determination of sildenafil citrate (Viagra) is based on voltammetric oxidation of the piperazine ring on a carbon anode. Measurement

[1] J.D.H. Cooper, D.C. Muirhead, J.F. Taylor, P.R. Baker, J. Chromatogr. B 701 (1997) 87. [2] N.D. Dinesh, B.K. Vishukumar, P. Nagaraja, N.M. Made Gowda, K.S. Rangappa, J. Pharm. Biomed. Anal. 29 (2002) 743. [3] N. Daraghmeh, M. Al-Omar, A.A. Badawan, A.M.Y. Jaber, J. Pharm. Biomed. Anal. 25 (2001) 483. [4] J.J. Berzas Nevado, J. Rodriguez Flores, G. Castaneda Penalvo, N. Rodriguez Farinas, J. Chromatogr. A 953 (2002) 279. [5] J.J. Berzas Nevado, J. Rodriguez Flores, G. Castaneda Penalvo, N. Rodriguez Farinas, Electrophoresis 22 (2001) 2004. [6] N.D. Dinesh, P. Nagaraja, N.M. Made Gowda, K.S. Rangappa, Talanta 57 (2002) 757. [7] G. Altiokka, Z. Atkosar, E. Sener, M. Tuncel, J. Pharm. Biomed. Anal. 25 (2001) 339. [8] J.J. Berzas, J. Rodriguez, G. Castaneda, M.J. Villasenor, Anal. Chim. Acta 417 (2000) 143. [9] J.-M. Kauffmann, J.-C. Vire, G.J. Patriarche, L.J. Nunez-Vergara, J.A. Squella, Electrochim. Acta 32 (1987) 1139. [10] C.M. Riley, T.W. Rosanske, Development and Validation of Analytical Methods, Elsevier, New York, 1996. [11] M.E. Swartz, I.S. Krull, Analytical Method Development and Validation, Marcel Dekker, New York, 1997.