3.1 Voltammetric study

Figure 1 depicts a typical cyclic voltammogram obtained from the HOPG/0.01M CoCl₂ + 1 M NH₄Cl system at the scan rate of 20 mV s⁻¹. Note at direct scan, the presence of three slopes. The slope II and III are associated with the presence of electrochemical processes on the electrode surface (see close up in Figure 1a). The slope I is the zone of null current where not electrochemical processes are recorded, while the peak A is related to the cobalt reduction process. In the anodic zone, two current density peaks B (shoulder) and C appeared at the potentials of −0.450 V and −0.250 V respectively. These anodic peaks have been associated with either the dissolution of cobalt previously electrodeposited at direct scan,³³ or the dissolution of hydrated cobalt oxide layers.³⁵ In order to investigate if the cathodic current recorded in zone II corresponds to cobalt electrodeposited, a linear voltammetry on the HOPG/1 M NH₄Cl system was recorded at the potential range 0.600 to −1.200 V. Figure 2 shows a comparison of the experimental voltammograms obtained from the solutions with and without Co²⁺ ions. From this comparison, it is clear that only the current recorded in zone III corresponds to the cobalt electrodeposition. Therefore, the cobalt electrodeposition in the present system is beginning at potential values lower than −1.000 V.

Fig. 1 A typical cyclic voltammogram obtained at 20 mV s⁻¹ from the system HOPG/ 0.01 M CoCl₂ + 1 M NH₄Cl at pH=7. The potential scan was started at 0.600 V toward the negative direction. Arrows indicate the potential scan direction. Cathodic current density peak (A) and anodic peaks (B (shoulder) and C) are also indicated in the Figure. 

Fig. 2 A typical lineal voltammogram obtained at 20 mV s⁻¹ from the system HOPG/ 0.01 M CoCl₂ + 1 M NH₄Cl at pH=7 (solid line) and HOPG/ 1 M NH₄Cl at pH=7 (broken line). The potential scan was started at 0.600 V toward the negative direction. 

3.2 Kinetic analysis of the 3D nucleation and growth process

Figure 3 shows characteristic current density transients obtained at overpotential conditions (E ≤ −0.533 V vs Ag/AgCl) from the HOPG/ 0.01 M CoCl₂ + 1 M NH₄Cl system. Observe, that all these transients exhibit a falling current at shorter times, also the j vs t plot shows a maximum and then approaches to the limiting diffusion current. We plotted the current density, which falls after the maximum vs 1/t^½ and it was obtained a linear relationship. This result suggests that the electrodeposition process of cobalt, in the present system, is controlled by diffusion.³⁶

Fig. 3 Current transients obtained by means of the potential step technique for different potential step values (mV) indicated in the Figure. In all the cases, the initial potential was 0.600 V vs Ag/AgCl. 

In the literature it has been reported that the cobalt electrodeposition begins through progressive nucleation, which may change to instantaneous nucleation when the cobalt concentration is increased.³⁷ Also, the cobalt electrodeposition may occur through a nucleation process under a charge transfer.³⁷ In order to identify the type of nucleation and growth process of the system analyzed in the present work, we followed the methodology reported by Scharifker et al., to identify the type of nucleation and growth mechanism as progressive or instantaneous comparing the nondimensional experimental transients with their respective dimensionless plots.³⁸ These curves are plotted substituting the coordinate values of the experimental local maximum (t_m, j_m), in:³⁸

for instantaneous nucleation and in

for progressive nucleation.

A comparison of an experimental transient with the theoretical dimensionless curves derived through equations (1) and (2) is shown in Figure 4. It is clear that at t/t_m <1 the experimental curve compare favorably with the instantaneous behavior. At t/t_m >1, the experimental curve could not be classified as either progressive or instantaneous. Here, it must be reminded that the theoretical curves generated by Eqs. (1) and (2) correspond to two extreme cases of the nucleation process and in some cases a classification is not possible.³⁹ Moreover, these kinds of plots have been strongly criticized in the literature.³⁹ However, it has been reported that an additional efficacy of these plots is that if the experimental data fall within the range of validity of the theory proposed the full equation can be used to predict the overall behavior. Additionally, these plots can be used to identify the presence of an additional process to the nucleation and growth process.⁴⁰ Thus, the deviation of the current observed at t/t_m >1, indicates the presence of additional electrochemical processes, which are contributing to the total current. For the cobalt electrodeposition case, this additional contribution has been identified as a proton reduction process.⁴⁰ Therefore a classification as instantaneous or progressive process is not definitive from the observed in Figure 4.

Fig. 4 Comparison of an experimental current transient obtained at −1.080 V and their theoretical non-dimensional curves corresponding to a 3D instantaneous nucleation (equation 1) and 3D progressive nucleation (equation 2). 

Thus, considering the above mentioned, the density current transients depicted in Figure 3, should be explained by the following equation.

where j_a-d(t)) is the current density for a Langmuir-type adsorption–desorption process given by:⁴¹

while j_3D-R (t) is the current associated with the cobalt nucleation and growth process, j_3D(t), involving a proton reduction process, j_R(t):⁴²

where

z_PR is the number of electrons transferred during the proton reduction reaction, while k_PR corresponds to the rate constant of the proton reduction process. The nucleation rate, the number of active nucleation sites and the diffusion coefficent are represented as A, N₀ and D, respectively. All the other parameters used in equations (4) to (8) have their electrochemical conventional meanings. The evaluation of the kinetic parameters associated with the transients depicted in Figure 3, we carried out through a non-linear fitting of the experimental data to equation (3)). Figure 5 shows a comparison of an experimental current transient obtained at – 1.080 V with the theoretical current transient generated by Eq. (3). Note that the proposed model describes adequately the experimental current transient behavior.

Fig. 5 Comparison of an experimental current density transient (—) recorded during the Co electrodeposition process at −1.080 V and a theoretical transient (o o o) obtained by nonlinear fitting of Eq. (3) to the experimental data. The contribution to the total current of the adsorption, 3D nucleation and growth and a proton reduction processes are indicated in the Figure. 

The kinetic parameters obtained in the present work are reported in Table 1. The average diffusion coefficient value is 7×10⁻⁵ cm² s⁻¹, see Table 1. Note that the coefficient value obtained in the present work is larger than the reported for electrolytic solutions containing ammoniacal sulfate (2.65×10⁻⁶ cm²s⁻¹),⁴³ sodium sulfate (0.9×10⁻⁵cm²s⁻¹),⁴⁰ citrate (2.0×10⁻⁵cm²s⁻¹)⁴⁴, aqueous chloride (0.5×10⁻⁵cm²s⁻¹),⁴⁵ or acid ammoniacal (1.3×10⁻⁵cm²s⁻¹)⁴⁶ solutions. Even though, similar cobalt concentrations were employed in such electrolytic baths, the pH value of these electrolytic media is acid. In this sense, it is important to mention that the dissociation of the media will tend to reduce the overall mobility of the solute component.⁴⁷ Thus, at acid pH values a lower diffusion coefficient is expected in comparison with neutral electrolytic baths. On the other hand, the A and N₀ values increase as the applied potential becomes more negative. The nucleation rates obtained in the present work are similar to those obtained onto HOPG substrates from sulfate systems.^40,34 On the other hand, the number of active nucleation sites is larger that the obtained from sulfate systems onto HOPG electrodes.^40,34 Also, it can be observed an increment in k_PR values at lower applied potentials, which indicates that the reduction proton process is favored at bigger overpotentials. The k_PR values in the present work are lower that the obtained from acid sulfate solutions which suggest a less competition by H⁺ ions with the Co²⁺ ions by the active sites on the electrode surface, under neutral pH values.

3.3 Analysis of the kinetic parameters

A distinctive feature of electrochemical nucleation and growth is the size of a critical nucleus, a nucleus that has a 50 % chance of becoming a stable entity, rather than dissolving back into the electrolyte or becoming a part of nuclei growing in different places on the electrode surface.^48,49 From the nucleation rate values reported in Table 1, it is possible to calculate the number of atoms that form this critical nucleus by employing the following equation:^48,49

In the last equation k_B, T, Z, and e₀ are the Boltzmann constant, the absolute temperature, the number of transferred electrons and the elementary electric charge, respectively. The value of the slope dlnA(M1)/dη≈45.6 was obtained from the linear experimental dependence of ln A vs η (see Figure 6). If one substitutes α= 0.5 and the value of the slope dlnA(M1)/dη in Eq. (9) the result is n_c = 0. This value means that each active site on the HOPG electrode surface acts as a critical nucleus.^49,50

Fig. 6 In A vs η plot, used to calculate the critical size of Co nucleus according to Eq. (9). The broken straight line corresponds to the linear fit of the experimental data. 

On the other hand, the kℝ values were analyzed by using the well known Buttler-Volmer equation,⁵¹

Note that this equation can be linearized if we take the natural logarithm of both sides of the equation. Then, from the slope of the ln kℝ vs E plot, it is possible to calculate the value of αℝ as 0.35. This value compares very well with the calculated for glassy carbon electrodes.⁴²

On the other hand, in a diffusion controlled nucleation and growth process, the growing nuclei is able to inhibits the nucleation process, because the diffusion fields associated with already existing growing nuclei, and their overlapping of these zones, may interrupt the nucleation and thus lead to a saturation nucleus density (N_S). Thus, it is possible to calculate the efficiency of the use of the HOPG surface available nucleation sites as N_S/N₀, where N_S is the saturation number of nuclei. N_S can be calculated through the equation (11)),⁵²

where

In Table 2 are reported the N_S and N_S /N₀ values obtained in the present work. Note that these values increase with the applied potential. Also, it is interesting to observe that the values of the N_S/N₀ are low, which may be caused by the occupation of the active sites of the HOPG electrode surface by NH₄⁺ and H⁺ ions.²⁹ However, these adsorbed ions may stop the diffusion of Co ad-atoms towards the growing nuclei, which may induce the growth of small clusters on the electrode surface.²⁹ Moreover the ratio N_S/N₀ is lower than the reported for glassy carbon electrodes (GCE). Probably this is due to the existence of less structural defects on HOPG electrodes, in comparison with GCE electrodes.

3.4 Morphological study

AFM images of cobalt electrodeposited on HOPG electrode at −1.040, −1.080 and −1.100 V are epicted in Figure 7. Figure 7a, show the AFM image obtained at −1.040 V, in this image it is possible to observe the formation of small clusters of 50-100 nm in diameter and 80-100 nm in height. Similar results were obtained at −1.080 V, see Figure 7b. Observe that the cobalt clusters obtained at −1.100 V are bigger than those obtained at −1.040 and −1.080 V, see Figure 7c. This is due to the overlapping of small cobalt clusters of 50 nm in diameter approximately. Also, it is interesting to observe that the bigger cobalt clusters have similar height and width at the different applied potentials. Thus, in the present system to apply lower potentials increase the number of nuclei on the electrode surface, but not modify substantially their maximum size. However it is clear the formation of cobalt cluster with different size, which may be indicative of a progressive nucleation process.

Fig. 7 AFM topographic images of the Co electrodeposition from the system HOPG/ 0.01 M CoCl₂ + 1 M NH₄Cl at pH=7 at a) −1.040 V, b)−1.080 V and c) −1.100 V. 

Figure 8 shows the magnetic images associated with Fig. 7. Fig. 8a corresponds to the magnetic images for clusters obtained at −1.040 V, observe the presence of disperse magnetic clusters, which can be directly related to the topographic features observed in Figure 7a. In this image, the color contrast within the overlapped nuclei suggests different magnetic domain regions. Since, bright colors correspond to repulsive interactions while the dark ones represent attractive interactions. A similar behavior is exhibited by the clusters shown in Figure 8b. From Figure 8c, it is clear the formation of aggregates with a magnetic coupling between them.

Fig. 8 MFM images of topographic images obtained at a) −1.040 V, b)−1.080 V and c) −1.100 V.