Figure 2 shows the oxidation of amitrole on BPPGE, MWCNT-BPPGE, and MTAPc-MWCNT-BPPGE (M = Fe, Co, Mn or Ni). The most important observations obtained from the graphs are: (i) there is a broad and weak peak due to oxidation of amitrole on unmodified BPPGE; (ii) the peak due to oxidation of amitrole is much larger on both BPPGE-MWCNT and BPPGE- MWCNT-poly-MTAPc and is observed at less positive potentials on the former showing that the former is a better catalyst than the MTAPc complexes for the oxidation of amitrole.

However, larger currents are observed on MTAPc compared to MWCNT-BPPGE; (iii) amongst the metal tetraaminophthalocyanine-modified electrodes, the catalytic efficiency (judged by shifting of the oxidation potential for amitrole to less positive values) towards amitrole oxidation follows the order: FeTAPc > CoTAPc > MnTAPc > NiTAPc. It is well known that metallophthalocyanine complexes containing Fe, Mn and Co as central metals are good catalysts for many reactions. However, this complex shows poor catalytic activity for NiTAPc when compared to FeTAPc, MnTAPc and CoTAPc. Since BPPGE-MWCNT-poly-FeTAPc gave the highest electrocatalytic response compared to the other electrodes, all subsequent studies were carried out with this FeTAPc-based electrode.

Figure 3 shows the voltammetric evolution of the BPPGE-MWCNT-poly-FeTAPc at scan rates ranging from 10 – 900 mVs^-1 in a pH 12.0 solution containing 10^-3 M amitrole. We observed linear variation of the peak current with the square root of scan rate (ν ^1/2) (R² = 0.9933) (not shown).

This result clearly indicates a diffusion-controlled electro-oxidative process. The plot stabilises from > 300 mVs^-1 (not shown) Also, the peak current varied linearly with the scan rate (ν) (R² = 0.9900) from 10 – 180 mVs^-1, stabilising from > 200 mVs^-1 (not shown). Since this type of process is characteristic of surface-adsorbed electrochemical process; it suggests the likelihood of amitrole or oxidation products and/or intermediate(s) also adsorbing onto the electrode surface. Interesting, the current function plot (Figure 3b) gave the characteristic shape of a coupled chemical reaction (EC_cat) for the amitrole, clearly confirming electrocatalytic activity. On the basis of the information, we suggest the following mechanism for the oxidation of amitrole. The redox process of the MWCNT-confined FeTAPc species is: (1) ( Fe 3 + TAPc ) film ⥬ ( Fe 2 + TAPc ) film

The interaction of the Fe³⁺TAPc with aqueous amitrole results to the regeneration of the Fe²⁺TAPc and the formation of amitrole oxidation intermediates: (2) ( Fe 3 + TAPc ) + Amitrole ( aq ) → ( Fe 3 + TAPc ) film ( Amitrole ) (3) ( Fe 3 + TAPc ) film ( Amitrole ) → ( Fe 2 + TAPc ) film + Intermediates + e -

The intermediate(s) also coordinate with Fe³⁺TAPc film and are further oxidized to the final product(s), via similar mediated electro-oxidation process: (4) ( Fe 3 + TAPc ) film + Intermediates → ( Fe 3 + TAPc ) film ( Intermediates ) (5) ( Fe 3 + TAPc ) film ( Intermediates ) → ( Fe 2 + TAPc ) film + Products

We are not aware of the electrooxidative product(s) of amitrole, and it is important that this be pursued in future work.

Figure 4 shows the chronoamperomegrams recorded at different concentrations of amitrole at a fixed potential E = 0.4 V vs Ag|AgCl. Since the oxidation of amitrole could lead to formation of radicals/intermediates (as already discussed in the scan rate studies above) that could readily combine with the polymer film and poison the electrode, the electrode surface was conditioned between measurements by rinsing in buffer (pH 12).

We observed that more than 96 % of the original current of each concentration of amitrole analysed is regenerated after this treatment, suggesting the viability of continuous use of this electrode once fabricated. A linear relationship between current response and amitrole concentrations was obtained (using LINEST statistical programme) as: (6) I p / μ A = ( 8.80 ± 0.44 ) [ Amitrole ] / nM + ( 12.35 ± 2.35 ) ( R 2 = 0.9904 )

The limit of detection (LoD = 3.3 s/m [12], where s is the relative standard deviation of the intercept and m the slope of the linear current versus the concentration of amitrole) was calculated as 0.493 nM. When compared to the presently available literature reports on the electrocatalytic detection of amitrole [13-16], this report gave the best analytical response in terms of: (i) less positive detection potential (0.3 V vs Ag|AgCl); (ii) low detection limit (approximately 50% lower than the regulatory limit), and (iii) high sensitivity (approximately two and half times higher than the 3.44 μA/nM reported for carbon paste electrode modified with iron (II) phthalocyanine nanoparticles [4]. The enhanced sensitivity obtained in this work is attributed to the ability of multi-walled carbon nanotubes to act as efficient conducting species for iron(II) tetra-aminophthalocyanine electrocatalyst.

The selectivity of the FeTAPc-modified BPPGE-MWCNT was investigated using the mixed solution method [17]. The concentration of the interfering species and amitrole were 10^-6 and 10^-7 M, respectively. The selectivity was checked against the pesticides atrazine, simazine, 3,6-dichloro-2-methoxybenzoic acid (dicamba) and ammonium thiocyanate (components of many herbicide formulations). The values of K_amp (where K_amp is the amperometric selectivity coefficient) were determined from Equation (7) for analysis in the presence of interfering ions: (7) K amp = ( Δ I mixture Δ I amitrole − 1 ) [ amitrole ] [ interferent ]where ΔI_mixture and ΔI_amitrole are respectively, the changes in current for the mixture containing amitrole and the interfering ion, and amitrole alone. The K_amp values are graphically represented in Figure 5.

A K_amp value of less than 10^-3 indicates non-interference, while one which falls within the order of 10^-3 suggests that the species is an interferent, but not a strong one. Of all the possible interferents studied, dicamba interfered more strongly, while ammonium thiocyanate is the least interferent. Thus, the electrode could be most conveneinetly used for the detection of amitrole in solution where ammonium thiocyanate is present.

All electrochemical experiments were performed with an Autolab PGSTAT 30 potentiostat (Eco Chemie, Utrecht, The Netherlands) driven by the General Purpose Electrochemical Systems data processing software (GPES, software version 4.9). Chronoamperograms were obtained at 0.4 V (for FeTAPc), 0.5 V (for CoTAPc), 0.5 V (for MnTAPc), and 0.65 V (for NiTAPc). A conventional three-electrode system was used. The basal plane pyrolytic graphite electrode (BPPGE) disk (d = 5 mm in Teflon) used as working electrode was fabricated in-house by the Rhodes University Chemistry Machinery Workshop. Electrical contact with the disk was obtained via an inserted copper wire held in place with conducting silver varnish L 100 (Kemo^® Electronic, Germany). The working electrode was BPPGE modified with MWCNTs (BPPGE-MWCNT) or BPPGE modified with MWCNTs and Fe, Co, Mn or Ni tetra-aminophthalocyanine by electropolymerization (designated as BPPGE-MWCNT-poly-FeTAPc, BPPGE-MWCNT-poly-CoTAPc, BPPGE-MWCNT-poly-MnTAPc and BPPGE-MWCNT-poly-NiTAPc, respectively). A Ag|AgCl wire and platinum wire were used as pseudo-reference and counter electrodes, respectively. The potential response of Ag|AgCl pseudo-reference in aqueous conditions was less than the normal Ag|AgCl (3M KCl) and SCE by 0.15 V and ∼ 0.01 V, respectively. A Wissenschaftlick-Technische Werkstätten (WTW) pH 330/set-1 (Germany) pH meter was used for pH measurements. All solutions were deaerated by bubbling nitrogen prior to each electrochemical experiment. All experiments were performed at 25±1 °C.

The MWCNTs were purified as described previously [19]. Briefly, MWCNTs (1 g) were added to 2.6 M HNO₃ (140 mL) and the mixture was refluxed for 45 hours. The carbon nanotube sediment was separated from solution and washed with distilled water. It was then sonicated in a 3:1 mixture of H₂SO₄ and HNO₃ for 24 hours. The sediment was thereafter washed with distilled water, stirred for 30 minutes in a 4:1 H₂SO₄/H₂O₂ mixture at 70°C, and washed again with distilled water. The purified MWCNT paste was then air-dried for 48 hours.

A BPPGE was prepared for use or for further modification by renewing the electrode surface with cellotape. This procedure involves polishing an old BPPGE surface on carborundum paper, pressing cellotape on the cleaned BPPGE surface and then removing the tape, along with several layers of graphite. Before use, the electrode was then rinsed in acetone to remove any adhesive. To prepare a carbon nanotube modified BPPGE (BPPGE-MWCNT) electrode, carbon nanotubes were abrasively immobilized onto the BPPGE by gently rubbing the electrode on a fine quality filter paper containing the carbon nanotubes [20]. To prepare a MWCNT-BPPGE modified with iron or cobalt or manganese or nickel tetra-amino phthalocyanine (FeTAPc, CoTAPc, MnTAPc or NiTAPc), the electropolymerization method was used. Briefly, the MWCNT-BPPGE was immersed in a ∼1 × 10^-3 M solution of FeTAPc, CoTAPc, MnTAPc or NiTAPc in pure dry dimethylsulfoxide (DMSO) containing 10^-1 M tetrabutylammonium tetrafluoroborate (TBABF₄) as the supporting electrolyte and the solution was scanned within a potential range of -1.0 V and 1.2 V (vs Ag|AgCl) using cyclic voltammetry. The electrodes (BPPGE-MWCNT-poly-FeTAPc, BPPGE-MWCNT-poly-CoTAPc, BPPGE-MWCNT-poly-MnTAPc and BPPGE-MWCNT-poly-NiTAPc) were conditioned for use by immersing them in a 0.1 M phosphate buffer solutions of pH 12, and then scanning between -1.0 V and 1.0 V (vs Ag|AgCl). Ten cycles were sufficient to obtain stable cyclic voltammograms for all electrodes.