Impedance of copper electrode in slightly acid Cu(II)-glycine solutions

INTRODUCTION

Glycine H₂N-CH₂-COOH is widely distributed in the nature as an integral part of many vital substances. The ability of glycine to form complexes with metal ions has found its application in various fields of human activity. For example, in recent years, aqueous glycine solutions have shown themselves to be promising environment-friendly lixiviates for leaching precious metals [1]. In addition, glycinate complexes are often used in galvanic baths, the practical implementation of which from time to time encounters certain electrochemical problems. Besides, electrochemical systems containing metal complexes are also interesting from a theoretical point of view. The abundance of various chemical stages accompanying a charge transfer step makes them an attractive object for experimental studies.

The above also applies to the  Cu|Cu (II)-glycine system, whose electrochemical properties have been investigated quite intensively until now. According to Aksu et al. [2], about 30 processes involving various species of ligands and complexes, as well as insoluble oxygen compounds, are possible in this system. The research conducted at our research institute has shown [3] that the reduction of Cu(II)-glycine complexes proceeds via two one-electron transfers, and CuL⁺ is an electrochemically active complex (EAC) on which the first electron is transferred (here and below, to shorten the entries, glycine is symbolized as LH). Further studies using a rotating disk electrode [4] confirmed the  conclusion about the  EAC composition. Besides, they showed that the nature of the charge transfer step changes in the presence of surface oxides. Then, the transfer of the second electron to the intermediate complex of Cu(I) becomes a rate-determining step. It also turned out [3] that the kinetic data depend on the nature of the foreign (supporting) ions. Therefore, a rather extensive experimental material accumulated to date should be used with some caution, taking into account the nature of the supporting electrolyte. Some aspects of this problem are discussed elsewhere [5]. Finally, we point out an interesting, in our opinion, result. It turned out [6] that the glycine system can be successfully used to simulate the electrochemical behaviour of some herbicides in the environment.

Though a  significant progress in the  study of this system has been reached, some of its electrochemical properties require further investigations. In particular, this applies to solutions with the socalled ‘hidden’ ligand deficiency. This category includes slightly acidic media, where the amount of the active deprotonated ligand L^–1 is very low, despite the apparent large excess of the total glycine.

Recently, studies of Cu(II)-glycine complexes in perchlorate media have been performed [7] employing more advanced experimental techniques, such as linear potential sweep (LPS) voltammetry, electrochemical quartz crystal microgravimetry, and appropriate methods of quantitative analysis of experimental data. LPS voltammograms obtained at pH  4 contained two current peaks. In spite of this, it turned possible to construct linear Tafel dependences in the entire region of cathodic overvoltages. Their analysis made it possible to determine the kinetic characteristics of charge transfer, which is desirable to compare with the results obtained by other methods.

In this regard, we turned to electrochemical impedance spectroscopy (EIS), since it makes possible to obtain information both on faradaic processes and on a double electric layer. Impedance characteristics of the  copper electrode, obtained for slightly acid Cu(II)-glycine solutions, are analysed in the present paper.

EXPERIMENTAL

Solutions were prepared using deionised water, Cu(ClO₄)₂ · 6 H₂O and glycine (both from Aldrich), and NaClO₄ · H₂O (Fluka) as a supporting electrolyte. Specified values of pH were adjusted by addition of HClO₄ or NaOH. Solutions were deaerated before experiments with argon stream over 0.5 h.

To prepare the  working electrodes, a  0.4  cm² platinum wire was coated with ~5 μm thick copper in the solution containing (g dm^–3) CuSO₄ · 5 H₂O  –  250, H₂SO₄  –  50. A  polycrystalline layer with well-exhibited crystallographic edges and faces was formed. The roughness factor, estimated by scanning probe microscopy measurements, was ~1.1. More data on surface morphology is given elsewhere [8]. The prepared electrodes were rinsed with water and immersed immediately into the solution under investigation.

Impedance measurements were carried out using an Autolab PGSTAT302 potentiostat. Nyquist plots (relations between the  real (Z^/) and the  imaginary (Z^//) components of the  impedance) were obtained within the  frequency (f) range from 0.1 to 3 × 10⁴ Hz at the open-circuit potential. The amplitude of the imposed sinusoidal perturbation of the  electrode potential was 5  mV. The  electrode potential was measured in reference to the  Ag  |  AgCl  |  KCl (sat) electrode and was converted to the standard hydrogen scale. All experiments were performed at 20°C.

Computer programs elaborated by Boukamp [9] were used for analysing impedance spectra. When presenting the  impedance values, no corrections for surface roughness were done.

RESULTS AND DISCUSSION

A general view of the  data obtained is given in Figs.  1 and 2. At not too low frequencies (the left part of the  curves), Nyquist plots represent the arcs whose centres are below the abscissa axis. With a decrease in frequency, deviations from this dependence are observed, but the  plots do not go over to explicit lines, which usually occur in the case of simpler redox processes controlled by mass transport and charge transfer.

At least three main factors affect the impedance spectra. These include, in the first place, the pH of the solutions. With a decrease in acidity, the impedance of Cu electrodes increases in the entire frequency range (cf. Figs. 1, 2). Similar phenomena were also observed with a change in the total ligand concentration (c_L): a significant increase in impedance with c_L occurs both at pH 4 (Fig. 1) and at pH  5 (Fig.  2). Finally, the  third factor, which should be considered, is the exposure time (τ) of the freshly prepared electrode in the test solution. Its influence is especially pronounced at pH 4 and c_L = 0.04 M. In this case, an increase in impedance with exposure time is clearly visible, especially in the region of lower frequencies (upper part of Fig. 1). In other solutions studied by us, this effect is less pronounced. It is most likely that some data scatter might be qualified as insufficient measurement reproducibility.

The reasons for such behaviour seem to lie in the  equilibration processes that occur when the  Cu electrode is immersed in the  Cu(II)-glycine solution. Then, Cu(I)-containing species are generated at the Cu|solution interface by the process Cu + Cu²⁺ → 2 Cu⁺. This leads to the redistribution of components in the solution phase and to the onset of a new equilibrium. The aforementioned phenomena were quantified employing procedures based on the  material balance equations [5, 7]. It was found that the  total amount of Cu(I) species formed in slightly acidic media (c_Cu(I)) falls with both pH and c_L. It is reasonable to assume that the equilibration time will also change in accordance with the c_Cu(I) value. Therefore, such an interpretation is consistent with the  experimental phenomena noted above.

Quantitative analysis of experimental data is conveniently carried out using equivalent circuits (EC) if they adequately reflect the  basic properties of the  electrochemical system. When choosing them, we completely sustain Sluyters’ position [10], according to which ‘it is preferable to analyse the  impedance data according to the  appropriate theoretical models for which the mathematical expressions are derived’. It is well known that Cu²⁺ reduction proceeds via formation of a stable intermediate Cu⁺, which is capable of leaving the reaction zone by diffusion; this was confirmed experimentally using a rotating ring-disk electrode. Therefore, we focused on such theoretical developments.

In previous studies, a  theoretical analysis of the faradaic impedance has been performed provided that the final product is also soluble [11–13]. Next, the relationships obtained in Ref. [11] have been extended for the case when an insoluble final product (e.g. metal deposit) is formed [14]. No EC was proposed in the previous investigations [11, 12] until it was found [13] that the  general impedance expression corresponds to an equivalent circuit that contains 5 elements, including resistances (R), Warburg impedances (W) and capacitance (С). Its description code is as follows: R_∞(R₁W₁)(C₂W₂). Here, elements in series are given in square brackets, and elements in parallel are enclosed in parentheses. According to the authors [13], most elements of this EC ‘have no sensible physical meaning’. At the same time, another faradaic EC, viz. ([R₁W₁][R₂W₂]), has been suggested [15], which rigorously followed from the  same analytical expressions, but contained less elements. So, two different EC have been suggested for description of the same mechanism. Analysis has shown [16] that both ECs are indistinguishable. Upon a proper choice of elements, they display the same impedance spectrum over the entire frequency range. Our choice falls on the latter EC, because it is simpler, the elements have a clear physical meaning, and, most importantly, there are analytical expressions that relate the values of the elements to the kinetic parameters of the process (exchange current densities i₀₁ and i₀₂):

Here Warburg coefficients (σ) have their common meaning [17]. By definition, the  Warburg admittance Y_W = 1/Z_W = Y₀(jω)^1/2, where j is an imaginary unit and ω = 2πf; then σ = 1/√2 Y₀.

These equations were derived for the two one-step electron transfers in the  system containing simple metal ions [14, 15]. It seems possible to apply such relations to processes involving metal complexes if the chemical stages in a complex system are sufficiently fast. Then, concentrations of simple ions (i.e. [Cu²⁺] and [Cu⁺]) should be replaced by the  corresponding total concentrations c_Cu(II) and c_Cu(I) [18]. The above EC was supplemented by two non-faradaic elements: the solution resistance R_Ω and the impedance of the double electric layer. Since the structure of the double layer is not known in detail, its impedance is best represented by the constant phase element Q_dl, the admittance of which is expressed by the relation Y = Y₀(jω)ⁿ. This generalized EC is shown in Fig. 3.

When comparing the  EС impedance with the  experimental data, the  following procedures were carried out. First, extrapolating the  impedance to an infinite frequency, the  resistance of the  solution R_Ω was determined and eliminated from the  experimental data. Then, the  data were transformed to the  admittance spectrum. Again, the  high-frequency limit provided parameters of Q_dl, which was also subtracted from the  admittance spectrum. The rest was fitted to a sub-circuit containing faradaic elements R and W. In most cases, a  satisfactory agreement was obtained between the EC impedance and experimental data, and the frequency error did not exceed 3–4%. Such data are presented below.

Table  1 contains data obtained for pH  4. We have already noted that the impedance in this solution depends on the exposure time τ. It can be seen that this effect is caused by changes in the sub-circuit [R₁W₁]. Both the charge transfer resistance R₁ and the complex conductance of Warburg impedance Y₀ increase with exposure time. Initially, the  latter element is determined with a  sufficient degree of reliability, but later it is possible to determine only the order of conductivity (data marked with an asterisk). However, this circumstance does not prevent the determination of the exchange current density i₀₁, which also decreases with time.

Other parameters do not correlate with τ. This also applies to the CPE Q_dl, which transforms into electrical capacitance at n  =  1. The  average value of n for Q_dl is ~0.85, and this shows the capacitive nature of the  double layer impedance. Assuming that the  parameter Y₀ approximately represents the double-layer capacitance we obtain the order of С_dl ~70 μF cm^–2. Relatively high values of С_dl, typical of a Cu electrode, were observed both in solutions of simple salts and in the presence of ligands [15, 16, 19, 20].

Using the data of Table 1, we calculated the spectra of the equivalent circuit, which were compared with the corresponding experimental data. The examples of such comparison are presented in Figs. 3 and 4. A suitable agreement between the compared data is observed at f  >  1  Hz. Some discrepancies observed at the  lowest frequencies seem to arise from an insufficiently accurate representation of the imaginary part of the impedance. This is most clearly seen when considering the  phase angle ψ = arc tan(Z^///Z^/). Discrepancies in ψ are most pronounced at ~1 Hz (Fig. 4).

Similar phenomena were discovered during the  study of the  Cu|Cu(II), glycolic acid system [19]. It was established that the matching could be significantly improved if the  Gerischer impedance G [21] was introduced into the [R₁W₁] sub-circuit. This element imitates the kinetics of the first-order chemical step, which is not so easy to find in a complex system involving various chemical interactions. The  formation of Cu(II)-glycine complexes was found to be rather fast [22]. Some ambiguity concerns the  mobility of the  NH₂-group protonization in the glycine molecule. Glycine in aqueous solutions promotes formation of a zwitterion ⁺H₃N–CH₂–COO^–, which together with its protonated form ⁺H₃N–CH₂–COOH dominates in acid media. Both species are capable of splitting protons. During the reduction of Cu (II) complexes, they behave as mobile proton donors [7]. However, zwitterions do not fall into this category when protonated species support the electrochemical evolution of hydrogen [23]. The data obtained do not allow an unambiguous choice in this dilemma, because the differences between the compared spectra are not significant. At the same time, there is a reason to believe that a preliminary analysis of impedance is possible without a quantitative assessment of the system lability.

A similar behaviour is shown by other solutions. Some differences in their impedance characteristics have already been discussed above. Below, we present the data obtained for the 0.01 M Cu(II) solution containing 0.1 M glycine at pH 5. Unlike pH 4, lower double layer admittance is observed in this case (see Table 2). Besides, there is a tendency to decrease the capacitance of the double layer with time, this being indicative of the adsorption of glycine or glycine-containing species. As in the previous case, a  satisfactory agreement is observed between the  impedance spectrum of the  equivalent circuit and the  experimental data (Fig.  5). Qualitatively, the impedance spectra of the systems studied are similar. So, for example, the position of the maximum on the ψ spectra falls at a frequency of ~400  Hz and weakly depends on the  solution composition (cf. Figs. 4, 5).

Tables  1 and 2 contain the  values of the  exchange current densities calculated according to Eqs.  (1) and (2) with the  following parameters: D  =  6.6  ×  10^–6  cm²  s^–1 [7], c_Cu(II)  =  10^–5  mol  cm^–3, c_Cu(I)  =  2  ×  10^–8  mol  cm^–3 (pH  4) and c_Cu(I)  =  10^–8  mol  cm^–3 (pH  5) [5]. Since a  number of assumptions were made for the  validity of this approach (see above), an estimate of i₀ should be considered approximate, determining the  order of these values. The condition i₀₁ << i₀₂ is characteristic of the Cu(II) reduction. Note that similar values of i₀₁ and i₀₂ were obtained for Cu(II) solutions containing gluconic acid [20, 24]. In addition, the i₀₂values defined in this article are in line with the  current density obtained from voltammetric data for similar solutions [7].

CONCLUSIONS

In weakly acidic solutions (pH 4 and 5), the impedance of Cu|Cu(II)-glycine system increases with ligand concentration (c_L) and solution pH. The third factor affecting the impedance is the exposure time of Cu electrode in the test solution. Its effect is especially noticeable at pH 4 and c_L = 0.04 M when a rather slow equilibration of the system is observed.

Nyquist plots of the  Cu electrode present the arcs centred below the abscissa axis. They are typical of electrochemical reactions proceeding via two consecutive one-electron transfers. An equivalent circuit with two parallel sub-circuits, each of which contains the charge transfer resistance and the  Warburg impedance in series, is suitable for the quantitative estimation of impedance spectra.

Two exchange current densities determined from the impedance spectra differ by 2 orders of magnitude. Their average values (~1 μA cm^–2 and ~0.1 mA cm^–2) are in line with respective literature data.

Received 25 January 2020

Accepted 9 March 2020

References

1. B.  C.  Tanda, Ph.  D.  Thesis, Western Australian School of Mines (2017).
2. S. Aksu, F. M. Doyle, J. Electrochem. Soc., 148, B51 (2001).
3. A. Survila, Electrode Processes in Systems of Labile Complexes of Metals, Mokslas, Vilnius (1989) [in Russian].
4. A.  Survila, V.  Uksienė, Electrochim. Acta, 37, 745 (1992).
5. K. Mikulskis, S. Kanapeckaitė, A. Survila, Chemija, 30, 212 (2019).
6. A.  Survila, S.  Kanapeckaitė, J. Electroanal. Chem. (2020) [in press].
7. A.  Survila, S.  Kanapeckaitė, L.  Staišiūnas, Electro­chim. Acta, 259, 1045 (2018).
8. A. Survila, S. Kanapeckaitė, Electrochim. Acta, 168, 1 (2015).
9. B.  B.  Boukamp, Equivalent Circuit (equivcrt.pas), User’s Manual, University of Twente (1989).
10. J. H. Sluyters, Russ. J. Electrochem., 45, 25 (2009).
11. A.  R.  Despić, D.  R.  Jovanović, S.  P.  Bingulac, Electrochim. Acta, 15, 459 (1970).
12. R.  D.  Armstrong, R.  E.  Firman, J. Electroanal. Chem., 45, 3 (1973).
13. M.  Rueda, M.  Sluyters-Rehbach, J.  H.  Sluyters, J. Electroanal. Chem., 222, 45 (1987).
14. A. A. Survila, A. A. Katkus, Sov. Electrochem., 16, 1483 (1980).
15. A. Survila, V. Baliukienė, Chemija, 12, 195 (2001).
16. A. Survila, Bulg. Chem. Comm. Special Issue B, 48, 18 (2016).
17. A. Lasia, in: B. E. Conway, J. Bockris, R. E. White (eds.), Modern Aspects of Electrochemistry, Book No.  32, p.  143–248, Kluwer Academic/Plenum Publishers, New York (1999).
18. A.  Survila, Electrochemistry of Metal Complexes. Applications from Electroplating to Oxide Layer Formation, Wiley-VCH, Weinheim (2015).
19. J.  Būdienė, A.  Survilienė, A.  Survila, Russ. J. Electrochem., 43, 1294 (2007).
20. J. Pileckienė, A. Grigucevičienė, A. Survila, Chemija, 22, 85 (2011).
21. H. Gerischer, Z. Elektrochem., 55, 98 (1951).
22. A. F. Pearlmutter, J. E. Stuehr, J. Amer. Chem. Soc., 90, 858 (1968).
23. K. Mikulskis, S. Kanapeckaitė, A. Survila, J. Electro­chem. Soc., 165, J3186 (2018).
24. J. Pileckienė, A. Grigucevičienė, A. Survila, Chemija, 22, 85 (2011).

A. Survila, S. Kanapeckaitė, L. Gudavičiūtė, O. Girčienė

VARIO ELEKTRODO IMPEDANSAS SILPNAI RŪGŠČIUOSE Cu(II)-GLICINO TIRPALUOSE

Santrauka

Cu(II) redukcijos kinetikos tyrimams silpnai rūgščiuose (pH 4 ir 5) Cu(II)-glicino tirpaluose panaudota elektrocheminio impedanso spektroskopija. Nustatyta, kad Cu elektrodo impedansas auga didinant ligando koncentraciją ir tirpalo pH. Trečiasis veiksnys, turintis įtakos impedanso spektrams, yra Cu elektrodo ekspozicijos trukmė tiriamajame tirpale.

Gautos lankų pavidalo Naikvisto kreivės, kurių centras yra žemiau abscisės ašies. Jos yra būdingos elektrocheminėms reakcijoms, vykstančioms nuosekliai pernešant kiekvieną elektroną. Impedanso spektrams kiekybiškai įvertinti panaudota adekvati ekvivalentinė grandinė. Ji susideda iš dviejų lygiagrečių subgrandinių, kuriose krūvio pernašos varža yra nuosekliai sujungta su Varburgo impedansu.

Nustatyti dviejų elektronų nuoseklios pernašos mainų srovės tankiai, kurie skiriasi ~100 kartų. Jų vidutinės vertės (~1 μA cm^–2 ir ~0,1 mA cm^–2) sutampa su atitinkamais literatūros šaltiniuose pateikiamais duomenimis.