Determination of Malachite Green in Aquaculture Water by Adsorptive Stripping Voltammetry

ABSTRACT An adsorptive stripping voltammetric method for the determination of malachite green in aquaculture water has been developed. Initial studies were made using the cyclic voltammetry of malachite green at a glassy carbon electrode in 0.1 M phosphate buffer from pH 2 to 10. The redox behavior observed for malachite green was verified by the characterization of malachite green and its reduction product, leucomalachite green. Furthermore, leucomalachite green was found not to interfere with the determination of malachite green at pH 7.4, the optimum pH for malachite green determination. As a result, further studies were performed using adsorptive stripping voltammetry for the determination of malachite green in aquaculture water. The voltammetric waveform, accumulation potential, and accumulation time were optimized. The calibration plot was linear from 0.2 µM to 1.2 µM for malachite green using differential pulse voltammetry with a sensitivity of 0.8311 µA/µM. Using the method of multiple standard addition, aquaculture water fortified with 0.5 µM and 0.75 µM malachite green provided mean recoveries of 78.79% and 87.20% with coefficients of variation of 2.07% and 1.45%. Therefore, analytical figures of merit suggest that this method provides rapid, simple, economical, and precise determination of malachite green in aquaculture water.


Introduction
Malachite green (C 23 H 25 N 2 Cl) is a member of the triphenylmethane class of dyes and is known to exhibit both antimicrobial and antiparasitic properties. It has been extensively used as a biocide in the aquaculture industry worldwide (Aiderman 1985). However, malachite green and its reduced form, leucomalachite green, have been shown to have toxic effects on the human immune system, reproductive system, and to be potential carcinogens (Meyer and Jorgenson 1983;Gouranchat 2000;Srivastava, Sinha, and Roy 2004;Sudova et al. Netherlands). The voltammetric cell contained a graphite rod counter electrode, a silver/ silver chloride reference electrode (Ag/AgCl, with 3 M KCl solution) (Chen Hua Instruments, Shanghai, China), and a 3 mm diameter glassy carbon electrode as the working electrode (Chen Hua Instruments, Shanghai, China). Cyclic voltammograms were initially recorded in 0.1 M phosphate buffer, then in the buffer containing malachite green. For initial studies, a starting potential of 0.0 V with a switching potential of þ1.0 V was employed. Differential pulse voltammetry was performed using a starting potential of 0.0 V and a final potential of þ1.0 V with a step height of 10 mV, pulse repetition time of 0.2 s, pulse amplitude of 100 mV, and pulse duration of 50 ms.
All chemicals were supplied by Fisher Scientific (China), unless stated otherwise. Malachite green was obtained from Tanmo Quality Inspection, Beijing, China. Leucomalachite green was obtained from Dr. Ehrensorfer, Germany and was analytical grade. A 10 mM malachite green stock solution was prepared by dissolving the appropriate mass in ultrapure water. Working standards were prepared by dilution of the stock solution with ultrapure water. Ultrapure water was obtained from a Milli-Q Academic System (Millipore, USA). Solutions of disodium, trisodium, sodium o-phosphate, and o-phosphoric acid (Sinopharm Chemical Reagent Company, China) were made at a concentration of 0.2 M by dissolving the appropriate mass in ultrapure water. These were then combined to provide the desired pH in 0.9% NaCl. An appropriate volume was then added directly to the voltammetric cell and diluted with sufficient ultrapure water to give an overall phosphate concentration of 0.1 M.

Water sample collection and pretreatment
The water sample was obtained from Badu River Channel, Jumahe River Basin, Fangshan District, Beijing, 14 August, 2014 at a depth of 0.5 m. A 2.5 L glass hydrophore was used to collect the water which was transferred directly into sample bottles. Samples were refrigerated and sent to the laboratory as soon as possible for analysis. These were then passed through a microporous membrane filter (water-system, 0.22 µm) prior to voltammetric analysis.

Results and discussion
Figures 1-4 show the cyclic voltammograms obtained for a 1.0 mM solution of malachite green in 0.1 M pH phosphate buffer at pH values 2.0, 4.0, 7.4, and 10 at scan rates between 20 mV/s and 200 mV/s. Previous reports have shown that at pH 2.0 the electrochemical oxidation of malachite green leads to the formation of the oxidized form of N,N,N,Ntetramethylbenzidine([1,1-biphenyl]-4,4-diamine) characterized by quasi-reversible diffusion controlled behavior (Ngamukot et al. 2006;Ma et al. 2008) 2e -, 2H þ redox couple (Galus and Adams 1962). In the present study, the cyclic voltammograms at pH 2.0 ( Figure 1) showed similar behavior, with a single oxidation peak on the initial positive scan and single reduction peak on the return negative scan. Plots of peak current (i p ) vs. the square root of scan rate (v ½ ) showed a linear dependence, showing the electrochemical process to be diffusion controlled. Consequently, malachite green provided similar redox behavior at this pH to the results provided in the literature. Figure 2 shows a cyclic voltammogram obtained at pH 4.0. The oxidation at pH 2.0 is still present, but the reduction peak on the return negative going scan is much broader and less well-defined. Several other redox processes were also present because the pH was close to the pK a value of malachite green. These conditions result in two or more forms of malachite green being present producing the additional peaks. However, at pH 7.4 ( Figure 3) and at values between 6.0 to 8.0 along with this same oxidation peak at þ0.90 V, an additional oxidation peak is seen (E p ¼ þ 0.5 V). The peak current values for this oxidation peak were linearly related to scan rate (v). Further investigations of current function (i p /v ½ ) vs. the square root of scan rate (v ½ ) (Nicholson and Shain 1964;Wopschall and Sharin 1967) demonstrated this process to be reactant adsorption in nature. On the return negative scan, a single reduction peak was present at þ0.4 V. In addition, there was little   It is believed that the oxidation peak at þ0.5 V resulted from the oxidation of the carbinol form of malachite green. At pH values higher than 6.9 (Cuong, Ishizaka, and Kitamura 2012), malachite green was chemically reduced to its carbinol form, which may undergo a two electron oxidation to provide malachite green and hydroxyl ion. The carbinol form is more nonpolar and hence would explain the reactant adsorptive behavior observed. The more positive oxidation at þ0.90 V is believed to result from oxidation of the amine nitrogen lone electron pair to form a radical species, as shown in Equation (1). This  phenomenon has also been reported previously (Masui, Sayo, and Tsuda 1968). Figure 4 shows that at pH 10, one oxidation peak was recorded across the potential range that was caused by the oxidation of the amine. The high pH and high concentration of OHmakes the oxidation to be unfavorable. In some scans at this pH, an additional oxidation peak was observed that was attributed to the instability of malachite green at this high pH.
To further investigate the electrochemical mechanism for malachite green, the cyclic voltammetry of leucomalachite green was also investigated under the same conditions. The same oxidation peak for malachite green was also recorded for leucomalachite green ( Figure 5). Moreover, the solution around the working electrode was seen to change from clear to green when the potential was scanned from 0.0 V to þ0.8 V at pH 2.0. It was concluded that this green color resulted from electrochemically generated malachite green formed by the oxidation of leucomalachite green to malachite green. This phenomenon was not observed when the supporting electrolyte pH was increased. Further cyclic voltammetric investigation of leucomalachite green showed a second peak that became larger as the pH was changed to 4.0. However, the peak potential remained constant ( Figure 6). This result may be due to the oxidation of the carbinol form of leucomalachite green as supported by the results. The voltammetric peaks of leucomalachite green decreased in magnitude with increasing pH, as it was more difficult to oxidize to malachite green under alkaline conditions.
Additional studies were made into the origin of the reduction peak recorded at þ0.4 V for malachite green. A series of cyclic voltammetric investigations were undertaken using a switching potential of þ0.6 V, before the beginning of the second oxidation. Consequently, it was concluded that this reduction was the result of the species formed from the oxidation at þ0.5 V. An additional two oxidation peaks were found when extending the switching potential to the more positive potential of þ2.0 V, which was believed to result from the polymerization of malachite green (De-Lin et al. 1989;Raoof, Ojani, and Baghayeri 2013) at its amine group (Chen, Chen, and Thangamuthu 2007). The mechanism of the voltammetric dimerization of aromatic amines has been investigated by Hart, Smyth, and Smyth (1981). Several mechanisms for the polymerization of malachite green have been proposed (De-Lin et al. 1989;Raoof, Ojani, and Baghayeri 2013). However, we believe that the tail-to-tail dimerization (Hart, Smyth, and Smyth 1981;Honeychurch, Hart, and Kirsch 2004) shown in Scheme 1 explains the voltammetric behavior observed in this study. Malachite green (i) was first oxidized to a cation radical (ii). This radical (ii) may form the resonance structure (iii). The presence of (iii) may cause the electrophilic tail-to-tail dimerization of two radical species to provide the dimer (iv) that may be converted to (v) by a two electron, two hydrogen ion oxidation.
The reactant adsorption of the malachite green at neutral pH values is analytically a very useful finding as it allows for the development of an adsorptive stripping voltammetric method for the determination of malachite green. Consequently, this approach was investigated in detail. Figure 7 shows the effect of accumulation potential on the resulting peak current of both oxidation peaks using an accumulation time of 15 s. The peak current increased from þ0.6 V to a maximum value between þ0.4 V and À 0.2 V vs. Ag/AgCl which was found to decrease at more negative potentials. Consequently, further studies were carried out using an accumulation potential of 0.0 V vs. Ag/AgCl. When using longer accumulation times or more negative accumulation potentials, two oxidation peaks were recorded at peak values of þ0.54 V and þ0.63 V. The more positive peak is believed to result from the oxidation of a monolayer of malachite green on the glassy carbon electrode surface, whereas the more negative peak results from the oxidation of a multilayer of malachite green deposited on the glassy carbon electrode. This more negative peak was only seen at extended accumulation times, as it was only formed once the monolayer had been established (Honeychurch, Hart, and Cowell 2000). Figure 8 shows the effect of increasing accumulation time at an applied potential of 0.0 V using 2.0 µM malachite green. The first oxidation stripping peak was found to increase with  increased accumulation time and reached a maximum value at ca. 100 s. The second oxidation peak was also found to increase with accumulation time over the time frame studied.

Effect of electrochemical measurement waveform
In order to improve the detection limit, differential pulse voltammetry was investigated for the determination of malachite green. Differential pulse voltammetry provides improved signal-to-noise ratios and low limits of detection (Honeychurch, Hart, and Cowell 2000). Initial studies were performed under the optimized conditions by differential pulse voltammetry. The resulting oxidation peak of malachite green was greatly improved and additional smaller peaks were also observed ( Figure 9). Consequently, differential pulse voltammetry was used for subsequent investigation.

Analytical figures of merit
A calibration study was carried out using differential pulse voltammetry in a 0.1 M pH 7.4 phosphate buffer from 0.2 µM to 2 µM. The plot was linear up to 1.2 µM with a slope of 0.8311 µA/µM and an R 2 value of 0.990. The limit of detection based on a signal-to-noise ratio of 3 was 0.12 µM. Beyond 1.2 µM, the response was found to be quasi-linear up to at least 2 µM. These performance characteristics are able to meet the detection limits required for the determination of malachite green in environmental water samples (Burchmore and Wilkinson 1993). It should be mentioned that by simple extension of the relatively short accumulation times investigated in this study it would be possible to improve these analytical figures of merit. Table 1 provides a summary of previously reported methods for the determination of malachite green in various water samples. The spectrographic and chromatographic approaches require a separate concentration and clean-up step to obtain the reported detection limits. In this method, the preconcentration and measurement are performed in the same solution, eliminating problems such as losses in sample preparation. This approach uses an unmodified carbon electrode that has the added advantage of being more stable compared to techniques based on modified electrodes.

Interference studies
As mentioned previously, cyclic voltammetric investigations showed that under the optimized conditions, equal molar concentrations of leucomalachite green did not interfere with the determination of malachite green. To further validate the performance of the method, a number of possible interferences were studied. Thirteen metal and acidic ions previously reported to be common interferences were evaluated (Lin et al. 2013;Y. Sun et al. 2015) for the determination of 1 µM malachite green. No interferences were observed for greater than 500-fold excess concentrations of K þ , Na þ , Ca 2þ , Fe 2þ , Mg 2þ , Zn 2þ , Cl À , I À , SO 4 2À ; NO 3 À ; CO 3 2À ; PO 4 3À , and CH 3 COO À .

Analysis of aquaculture water
The analytical procedure was evaluated by the determination of malachite green in aquaculture water with and without fortification of malachite green standards. The samples were diluted one-to-one in 0.2 M pH 7.4 phosphate buffer, and the concentration of malachite green was determined using the method of standard addition. Table 2 shows the precision and recovery data obtained for replicate analysis of a single aquaculture water sample fortified with 0.50 µM and 0.75 µM malachite green. These results demonstrate that this method has promise for the determination of malachite green in water samples. Further analysis of other environmental water samples showed similar good recoveries and precision.

Conclusions
The redox behavior of malachite green was investigated at a glassy carbon electrode and found that well-defined peaks were obtained in 0.1 M pH 7.4 phosphate buffer using cyclic voltammetry and differential pulse voltammetry. This is the first report to investigate the voltammetric behavior of malachite green over an extended pH range. It is also the first report to exploit the oxidation process seen at pH 7.4 at an unmodified glassy carbon electrode for the determination of malachite green in aquaculture water. Unlike previously reported methods, it was shown that no elaborate extraction or separation procedures were required as the method of multiple standard additions was shown to be both precise and accurate (Table 2). It should be possible to improve the performance characteristics of the method by the application of longer accumulation times.