The experimental set-up for cyclic voltammogram plots and chronoamperometry curves analysis is a Voltalab analytical Radiometer kit. It consists of a potentiostat PST050 and a polarographic stand MDE 150 comprising a measuring cell in which are placed:

1) A working electrode with an EDI 101 rotating disc having a glassy carbon tip (diameter 3 mm);

2) A silver chloride Ag/AgCl reference electrode, KClsat (potential 208 mV at 25˚C compared to the normal hydrogen electrode ENH);

3) A platinum auxiliary electrode;

4) An argon supply for solution bubbling.

The potentiostat and stand are controlled by a computer. The software used is Volta Master 4 for the execution of the commands as well as the data acquisition.

All solutions were prepared with milli-Q water. Standard solutions were prepared from mercury nitrate (purity > 99.99%, Merck), thallium sulphate (purity > 99.995%, Sigma Aldrich) dissolved in 0.1 mol·L⁻¹ perchloric acid HClO₄ (trace analysis) which was also used as the supporting electrolyte. A certified standard of thallium 1 g·L⁻¹ (Certipur, Merck) was also used for the preparation of standard solutions. The solution was bubbled through a stream of pure argon.

The surface morphology of the electrodes was observed with a HITACHI TM-1000 Tabletop Microscope. It is coupled to an Energy Dispersive X-Ray Analysis (EDX) system. The measuring device is connected to a microcomputer which allows it to be controlled via the TM-1000 software for images and Swift ED-TM for elemental chemical analysis.

The nucleation and growth kinetics of a 2D or 3D phase, during the deposition of a metal on a metallic or non-metallic substrate, is usually studied using i = f(t) transients obtained by recording the evolution of the current density as a function of time during a potentiostatic control. The transients of the deposition of Tl on Hg film were recorded at different final potentials E_f. The initial potential is maintained in all cases at E_i = −400 mV because at this potential there is no Tl deposition.

Electrochemical impedance measurements were performed with the µAUTOLABIII/FRA system (Eco Chemie, Netherlands) driven by a microcomputer using the FRA (Frequency Response Analyser) software. All experiments were performed at room temperature. Measurements were made around the open circuit potential of the system by superimposing a minimum sinusoidal perturbation of 10 mV amplitude. The frequency range explored was from 50 mHz to 50 kHz.

Among the parameters affecting electrodeposition, the nature of the metal substrate, the electrolyte and the deposition potential play an important role [1]. Thus, cyclic voltammetry in the pure diffusion regime is used to identify the regions of deposition and dissolution potentials of a metal on a foreign substrate S [18]. The first system studied was the GCE/(Hg²⁺, 10⁻³ mol·L⁻¹, supporting electrolyte HClO₄ 0.1 mol·L⁻¹) interface. The deposition of the mercury film on the GCE substrate is done with a mercuric ion solution of concentration Hg²⁺ 10⁻³ mol·L⁻¹. Figure 1 shows the cyclic voltammogram corresponding to the deposition and redissolution of mercury on the GCE in 0.1 mol·L⁻¹ HClO₄ medium at a scan rate of 40 mV/s. The potential scan is performed between 0.7 V and 0.2 V. The curve analysis shows that in the system under investigation, the formation of a monolayer (2D Hg phase) in the Under Potential Deposition (UPD) domain ( Δ E = E − E Me z + / Me > 0 ) can be excluded, indicating a relatively weak Hg deposit-substrate interaction. The hysteresis of the cathodic current in the Over Potential Deposition (OPD) domain, characterised by η = E − E Me z + / Me < 0 , as well as the cross-over of the cathodic and anodic plots at the equilibrium potential of the Hg²⁺/Hg system, E Hg 2 + / Hg ∘ = 0.44 V/ref, is related to the germination and growth of a 3D Hg phase that dissolves in the course of the anodic sweep back into the UPD domain [1] [19].

The SEM study allowed the observation of the Hg film on the surface of the glassy carbon electrode. The GCE substrate is first raised to a potential in the UPD range and then raised to a cathodic potential in the OPD range. Figure 2 is a series of images showing the morphology of the electrodeposited film on the surface of the GCE substrate according to the routine polarisation described above. Figure 2(a) shows an SEM image of the GCE substrate prior to the deposition of a mercury film. A perfectly uniform surface is observed. The EDX analysis performed does not show any metallic elements. Figure 2(b) and Figure 2(c) show the surfaces of the GCE substrate after Hg deposition from a 10⁻³ mol·L⁻¹ Hg²⁺ solution in 0.1 mol·L⁻¹ HClO₄ medium at a potential of 0.2 V at a given deposition time. The SEM image was taken after a deposition time of 2 min (Figure 2(b)) shows sparse granules on the electrode surface. The Hg deposition is in the form of random crystallites on the substrate surface, confirming the weak metal-substrate interaction. These observed 3D growth islets and the absence of 2D Hg phases indicate that the electrodeposition of Hg on the GCE can be described by the so-called “Volmer-Weber” growth mechanism or islet growth mode. At a deposition time of 4 min, the SEM image (Figure 2(c)) shows a compact film of mercury. This is in fact a surface uniformly covered by grains arranged against each other and which constitute the texture of the mercury film. This texture of mercury films has already been reported in the literature [20] [21] [22]. Energy Dispersive X-ray (EDX) analysis of a portion of this surface shows that it is well covered with Hg, the EDX analysis also revealing the presence of chlorine from the 0.1 mol·L⁻¹ HClO₄ supporting electrolyte. The experimental conditions retained for the deposition of Hg on the GCE substrate are therefore the following: deposition potential 0.2 V/ref (Ag/AgCl, KClsat), deposition time equal to 4 min.

The second system studied was the interface GCE/Hg film/(Tl⁺, concentration C, supporting electrolyte HClO₄ 0.1 mol·L⁻¹), C is the concentration of Tl⁺ ions to be determined in an aqueous matrix. The investigations were first carried out for a Tl⁺ ion concentration C equal to 10⁻³ mol·L⁻¹. Figure 3 is a cyclic voltammogram corresponding to the deposition and redissolution of Tl on the mercury film at a scan rate of 2 mV/s. The hysteresis of the cathodic current is obtained at a cathodic potential below −0.9 V/ref and a cross-over of the cathodic and anodic plots at this potential. In contrast to the deposition of Hg on the GCE (Figure 2), the cross-over of the cathodic and anodic plots does not occur at the equilibrium potential of the Tl⁺/Tl system. The equilibrium potential of the Tl⁺/Tl system, E Tl + / Tl ( Hg ) ∘ = − 0.75 V/ref, was measured in the open circuit cell. This value is identical to the potential at the intersection of the curve with the zero current axis. In a first approximation, the hysteresis of the cathodic current and the cross-over at the potential of −0.9 V/ref could indicate an increase in the density of 3D Tl nuclei and the growth of these nuclei in a potential range below −0.9 V/ref.

In agreement with Nernst's law, the potential at which the cathodic and anodic plots cross is shifted towards the negative potentials when the concentration of Tl⁺ ions increases from 10⁻⁶ to 10⁻³ mol·L⁻¹. Analysis of the curve does not reveal the formation of a 2D Tl phase. Before the cross-over potential, the electrochemical reaction corresponding to the incorporation of Tl⁺ ions into the mercury amalgam is:

Tl + + e − ( Hg ) ⇌ Tl ( Hg ) (2)

This reaction is controlled by the diffusion of Tl⁺ ions into the mercury amalgam. As mercury has no growth sites (imperfections, surface irregularities, etc.), the nucleation and growth processes of 3D Tl phases are inhibited in a first step. Subsequently, as the Tl film becomes compact, the surface is characterized by a higher density of imperfections or growth sites. The increase of the density of growth sites is significant after the cross-over potential. The nucleation of 3D Tl nuclei, followed by the growth of these nuclei, is then easier on the surface [23].

The influence of the scan rate (v) during Tl deposition was studied in the potential range −0.2 to −1 V, with the scan rate varying between 2 and 100 mV/s. The results are shown in Figure 4. It is important to note that at high scan speeds, the cross-over is no longer visible on the voltammograms.

The normalized voltammograms obtained by plotting i.v^−1/2 = f(E) show that in reduction all curves are superimposed. The reduction peak varies with the square root of the scan rate (v^1/2), which confirms the diffusion of Tl⁺ ions in the mercury amalgam [24]. On the other hand, the oxidation does not vary with v^1/2, nor with v. The oxidation therefore goes through a limiting kinetic stage.

A series of Tl deposits on the GCE/Hg substrate were carried out at different potentials and the corresponding SEM images are shown in Figure 5. The GCE/Hg substrate is first raised to a potential in the range where Tl deposition does not take place, and then raised to a cathodic potential corresponding to the electrochemical reaction. At a potential of −0.7 V (Figure 5(a)), a few scattered particles are observed. EDX analysis, carried out on this surface, did not detect Tl (84.9% Hg, 0% Tl and 15.1% Cl). At this potential, the Tl(Hg) amalgam consists mainly of mercury. The SEM image in Figure 5(b) corresponds to a Tl deposit at a potential of −1 V with a deposition time of 2 min. Considering that the increase in density of growth sites occurs below −0.9 V, it is expected to observe a larger population of Tl grains, stacked against each other, some of them being stick-shaped. At this potential, Tl deposition takes place on a surface consisting mainly of Hg grains. By extending the deposition time to 4 min at a potential of −1 V, we find that the surface is almost covered with a Tl film, indicating the formation of tree-like branches (Figure 5(c)). Indeed, the EDX analysis carried out on this surface reveals a massive Tl deposit (100%). At a potential of −1.2 V for 4 min, a surface more or less covered by sheets can be observed on the SEM image of Figure 5(d), still indicating tree-like ramifications. However, this surface is much less covered than at the −1 V potential. This result is confirmed by the EDX analysis, which shows a lower deposition than at −1 V, i.e. 80.7% Tl. This result can be explained if we consider that below −1 V, the reduction of protons starts to be important and therefore decreases the quantity of Tl deposited on the surface [25].

The growth kinetics were studied by potentiostatic transient methods. These transients characterize the deposition of Tl on Hg. Figure 6 shows these transients in a representation −i = f(t).

The transients at −700 mV and −725 mV are monotonic decreasing, indicating diffusion controlled kinetics. They correspond to the Tl(Hg) amalgam formation reaction and no 3D nuclei formation can be envisaged at these potentials. The transients pass through a maximum, starting at a potential of −800 mV. The time to reach this maximum decreases as the potential becomes more negative. Thus, at a potential of −1000 mV, there is a well-defined maximum characteristic of a nucleation and three-dimensional growth mechanism [23] [26] [27].

A theoretical model taking into account the overlap of diffusion zones has been proposed by Sharifker et al. [23]. Figure 7 shows the theoretical model curves (i/i_m)² versus (t/t_m) for 3D progressive and instantaneous nucleation and an experimental curve (i/i_m)² versus (t/t_m). The experimental curve in Figure 7 was obtained from the transient at the −1000 mV potential. The experimental curve fits perfectly with the theoretical model corresponding to progressive nucleation. It can therefore be concluded that at this potential of −1000 mV, the nucleation of 3D Tl nuclei is progressive. This result is in agreement with the microscopic study which showed that at this potential, we have a quasi-covered surface of a thallium film with a high density of growth sites.

Impedance measurements were carried out to characterize the interface GCE/Hg film/Tl film/(Tl⁺, concentration C, supporting electrolyte HClO₄ 0.1 mol·L⁻¹). This system represents the electrochemical sensor that will be used for the quantitative determination of Tl⁺ ions in aqueous solution. The bare GCE and the GCE covered with a mercury film had to be characterized first. All the curves obtained are shown in Figure 8.

The impedance diagram of the bare GCE is recorded in the 0.1 mol·L⁻¹ HClO₄ supporting electrolyte and in a solution of the supporting electrolyte containing the mercuric ions Hg²⁺ 10⁻³ mol·L⁻¹. The impedance diagrams (in Nyquist representation) obtained are identical and overlap perfectly (Figure 8(a)). The charge transfer resistance, measured by evaluating the diameter of the semi-circle corresponding to the transfer loop, is 1.53 × 10⁷ Ω (Ohm). This high resistance obtained for the bare GCE reflects that when no electroactive redox couple is present, there is no charge transfer reaction at the electrode. In contrast, when a mercury film is electrodeposited on the GCE at a potential of 0.2 V for 4 min and the impedance diagram is recorded in the 0.1 mol·L⁻¹ HClO₄ supporting electrolyte solution containing 10⁻³ mol·L⁻¹ Hg²⁺ ions (Figure 8(b)), we obtain a decreasing charge transfer resistance as expected: it goes from 1.53 × 107 Ω for the GCE/(Hg²⁺ 10⁻³ mol·L⁻¹, HClO₄ 0.1 mol·L⁻¹) system to 4.66 × 10⁶ Ω for the GCE/Hg film/(Hg²⁺ 10⁻³ mol·L⁻¹, HClO₄ 0.1 mol·L⁻¹) system. This decrease is related to the charge transfer reaction at the electrode [22] [27] [28].

The deposition of thallium on the ECV/mercury film is carried out for 4 min at a potential of −1 V. The impedance diagram is then recorded for the ECV/Hg film/Tl film system (Tl⁺ 10⁻³ mol·L⁻¹, HClO₄ 0.1 mol·L⁻¹) (Figure 8(c)). In pure diffusion regime, in the high frequency range, the capacitive loop is not clearly visible and this is due to the fact that the electron transfer is very fast. The Warburg line is mainly observed at medium and low frequencies. The Warburg impedance (W) has a value of 7.4 × 10⁻⁴ Ω·s^−1/2. The resistance of the solution is 55 Ω. The charge transfer resistance decreases significantly from 4.66 × 10⁶ Ω for the GCE/Hg film/(Hg²⁺ 10⁻³ mol/L, HClO₄ 0.1 mol·L⁻¹) system to a value of 2290 Ω for the GCE/Hg film/Tl/(Tl⁺ 10⁻³ mol·L⁻¹, HClO₄ 0.1 mol·L⁻¹) system. These results, compiled in Table 1, can be easily explained if one considers that the charge transfer is made easier on the Tl film [29] [30] [31]; this was also shown in the SEM analysis.