Electropolymerization and electrochemical behavior of nickel Schiff base complexes with different groups between imine linkages

Abstract

Nickel Schiff base complexes Ni(salen), Ni(salphen) and Ni(saldmp) are synthesized and electropolymerized on multiwalled carbon nanotube electrodes. The structures of the three monomers are similar except for the groups between the imine linkages, so the difference in electrochemical behavior can be related to the influence of the groups. Polymerization parameters such as the consumed charge, the apparent surface coverage and the doping level are investigated to elucidate the effects of groups between imine linkages. The results show that poly[Ni(salen)] has higher consumed charge and apparent surface coverage than others, which means that poly[Ni(salen)] can be deposited more easily on the electrodes. While poly[Ni(salphen)] has the highest doping level, there are more electrons transferred per monomer unit, indicating a better capacitance for energy storage. The electrochemical characteristics are also evaluated and the peak potential in cyclic voltammetry plots is about 0.9 V for Ni(salen) and Ni(salphen), and about 0.7 V for Ni(saldmp). The different peak potentials indicate the redox potential will be related to the various groups. Meanwhile the galvanostatic charge/discharge curves display a specific capacitance of about 200 F g⁻¹ for poly[Ni(salphen)], and about 150 F g⁻¹ for poly[Ni(salen)] and poly[Ni(saldmp)]. The variation in electrochemical behavior is mainly caused by the different molecular structure and the groups between imine linkages are the unique differences in structure. So we propose a new electronic transmission mechanism that the electrons will transmit via the Ph–C=N–Y–N=C–Ph path (Y represents groups between imine linkages), and these groups can provide an electronic transmission path as imine bridges and then influence the electrochemical behavior.

---

1 Introduction

Transition metal Schiff base electroactive polymers derived from salicylaldehyde, such as Ni(salen),^1,2 Co(salen)^3,4 and Cu(saltMe),⁵ have been extensively studied due to their potential application as electrocatalysts, chemical sensors or optical devices. This type of transition metal complex can be anodically polymerized to generate electroactive films in solvents with low donor numbers such as acetonitrile (AN). The polymer can reserve and release charge repeatedly by redox switching of electroactive films in the oxidation/reduction state in different potential regions,^6,7 and it could be used for charge storage such as for supercapacitor materials.

Transition metal Schiff base electroactive polymers are generally considered to have both redox conductivity and electron conductivity by the metal ion centers and the conjugated system in the ligand.⁸ The polymerization of Ni(salY) will be following a ligand–radical coupling polymerization mode, which results from the generation of the C–C bonds between the phenyl rings of the ligands in the para-position of the phenol moieties to form individual polymer chain fragments as chain structure. During the polymerization process, electron transmission occurs via the Ph–O–metal–O–Ph bridge (Ph means phenyl), and the charge conduction will occur through the metal atom.^9,10 Meanwhile, the counter ions in the electrolyte will insert into the polymer to provide electroneutrality. So the polymerization is ultimately a ligand-based process with the transmission of electrons via the metal bridge. But it's necessary to evaluate the influence of the imine bridges on polymerization process and electrochemical behaviors for the Schiff base polymers when the groups between imine linkages for monomers are different. The difference of molecules structure may influence the electron response and the ions transfer, and the transmission of electrons can be influenced by the electronic transmission path within molecules.

Schiff base polymers, Ni(salen), Ni(salphen) and Ni(saldmp), are synthesised by salicylaldehyde and ethylenediamine,⁸ o-phenylenediaminethe, dimethyl propanediamine,¹¹ respectively, labeled as Ni(salY), where Y represents bridges between imine linkages, i.e. en for salen, phen for salphen, and dmp for saldmp, shown in Scheme 1. For the molecular structure, Ni(salen) has the smaller bridge ethyl, and Ni(salphen) and Ni(saldmp) have the larger imine bridges phenyl and dimethyl propyl. Moreover, both the imine bridge and the ligand group in Ni(salphen) contain a phenyl. During the polymerization process, the monomers will form the chain structure via ligand-based process, and the polymers with different groups may have different electrochemical characteristics. In this paper, we use cyclic voltammetry (CV) method to prepare poly[Ni(salY)]/multiwalled carbon nanotubes (MWCNTs) composites, and evaluate the anodic polymerization process and electrons transmission path, and electrochemical performance of composites is also presented.

Scheme 1 Molecular structure of Ni(salen), Ni(salphen) and Ni(saldmp).

2 Experimental

The Ni(salen), Ni(salphen) and Ni(saldmp) monomers were synthesized following the procedure,¹² in which an ethanolic solution of a diamine was added to a rapidly stirred ethanolic solution of salicylaldehyde, then solids formed was added to ethanolic nickel nitrate; after cooling, microcrystalline monomers could be filtered off. The monomers were recrystallized from AN (A.R. Grade, from Xilong Chemical Co., Ltd). Reagents for synthesis were obtained from Sinopharm Chemical Reagent Co., Ltd. Tetrabutylammonium perchlorate (TBAP, C.P. grade) and triethylmethylammonium tetrafluoroborate (Et₃MeNBF₄, C.P. grade) were purchased from Zhong Sheng Hua Teng Co., Ltd. All were used as received. MWCNTs were purchased from Shenzhen NanotechPort Co., Ltd. with an average diameter of 10–20 nm, an average length of 5–10 μm. The MWCNTs were functionalized under reflux with concentrated HNO₃ and H₂SO₄ (volume ratio = 1/3), stirred for 3 h at 60 °C, washed with distilled water until the pH value reached neutral, and dried in a vacuum oven, as described in the literature.¹¹

Anodic polymerization were performed in AN solution containing 1.0 mmol L⁻¹ Ni(salphen) monomer and 10 mmol L⁻¹ TBAP using a VMP2 electrochemical workstation with EC-Lab software (version 10.02) made by Princeton. A closed standard three-electrode cell was used with a Ti sheet coating MWCNTs as working electrode, an activated carbon sheet as counter electrode, with an Ag/AgCl reference electrode. The working electrode was prepared by mixing MWCNTs in N-methyl-2-pyrrolidone (NMP) until homogeneous slurry. Then the slurry was coated on a Ti sheet (current collector, 1 × 1 cm²) with the coating mass of 0.3 mg using a micropipet. CV method was used to prepare polymer electrodes, with the scan rate 20 mV s⁻¹, the potential range from 0 V to 1.3 V, and 10 recycle numbers.

Fourier transform infrared spectroscopy (FTIR) (Shimadzu, FTIR-8400S, solid sample with KBr), X-ray Photoelectron Spectroscopy (XPS) (AXIS Ultra^DLD) and field emission scanning electron microscope (FESEM) (Zeiss Supra™55 microscope) were used to confirm the elemental composition and surface morphologies of the composites. CV, electrochemical impedance spectroscopy (EIS), galvanostatic charge/discharge (GCD) tests were used to evaluate the electrochemical performance of the composites for each electrodes in 1.0 mol L⁻¹ Et₃MeNBF₄/AN.

3 Results and discussion

3.1 Electropolymerization process

Doping level (n) can be used to investigate the electrochemical activity of conducting polymer, which means the relative charge stored per monomer unit inside polymer. For a single scan polymerization, a double coulometric assay can be chosen to determine the degree of oxidation for each monomer unit, and in the previous work on Ni(salen) based polymer, it's widely recognized that three processes contribute to the overall charge passed: polymerization, ligand-based redox processes, and metal-based redox processes.^5,13–15 The number of electrons expended during electropolymerization contains three parts: 2 electrons for polymerization, 2y electrons for ligand-based redox switching (y electrons per phenyl ring), and z electrons for metal-based redox switching (z = 0 or 1). The ratio of polymerization charge (Q_poly) and redox charge (Q_redox) is Q_poly/Q_redox = (2 + 2y + z)/(2y + z). The doping level (n) can be estimated from the sum of 2y and z, i.e. n = 2y + z, which means the relative charge stored per monomer unit inside conducting polymer.

Fig. 1a–c show the continuous electropolymerization of Ni(salen), Ni(salphen) and Ni(saldmp) occurred on MWCNTs coated Ti electrode from 1^st to 10^th sequential scans, respectively. For all monomers, well-fined anode and cathode peaks are obtained in 10 cycles, but peak current of Ni(salen) reaches 1.8 mA at 10^th scan, about twice than Ni(salphen) and Ni(saldmp). Before polymerization, the concentration of all monomers are 1 mmol L⁻¹, when the applied potential is closed to about 0.6 V for Ni(salen) and Ni(salphen), and about 0.4 V for Ni(saldmp), the consumed charge (Q) increase by a linear variation, indicating that all monomers undergo polymerization, but the polymerization rate of Ni(salen) is faster than Ni(salphen) and Ni(saldmp) in the 10 cycles, as shown in Fig. 1d. The consumed charge of monomers in the electropolymerization process can be used to measure the total mass of the deposited polymer by the double coulometric assay. Charge depleted for Ni(salen) polymerization is about twice than that of Ni(salphen) and Ni(saldmp), with the value of 110 mC for Ni(salen) and 40 mC, 50 mC for Ni(salphen) and Ni(saldmp), respectively. This phenomenon suggests a more considerable charge trapping over the timescale of the CV experiment for Ni(salen), illustrating an easier polymerization of Ni(salen), meanwhile the polymerization rate of Ni(salphen) is very slow.

Fig. 1 CV plots for anodic polymerization of (a) Ni(salen), (b) Ni(salphen) and (c) Ni(saldmp) on MWCNTs-coated Ti electrode. Coulometric assay of (d) consumed charge (Q), (e) surface coverage (Γ) and (f) doping level (n) in 10 cycles.

The respective polymer electroactive surface coverage, Γ (mol cm⁻²), in accordance with consumed charge, is observed in Fig. 1e. Chronoamperometric experiments are performed on polymer films prepared by 10 polymerization cycles, correspondingly, Γ increases from 5.5 × 10⁻⁵ to 58.0 × 10⁻⁵ mol cm⁻² for Ni(salen), from 3.2 × 10⁻⁵ to 18.5 × 10⁻⁵ mol cm⁻² for Ni(salphen), and from 3.2 × 10⁻⁵ to 25.5 × 10⁻⁵ mol cm⁻² for Ni(saldmp). Γ is found to increase linearly, and the mass of the deposited polymers are also increase linearly by the formula m = MΓA, where M means relative molecular mass, and A means the area of electrode.

So with the different groups between imine linkages, the monomers exhibit a different deposition rate in the same condition, and Ni(salen) monomers show the fastest polymerization rate. It is because the steric hindrance of substituent is one of the crucial factors to affect the yields during the electropolymerization process. Steric hindrance, a steric effect for the interatomic interactions in organic compounds, will occur when the large size of groups within a molecule prevents chemical reactions that are observed in related molecules with smaller groups. Hence for Ni(salY) monomers with different substituents of groups between imine linkages, steric hindrance will have various influence. Ni(salen) has the smaller imine bridge ethyl, and Ni(salphen) and Ni(saldmp) have the larger imine bridges, so steric hindrance will occurs obviously in the larger size molecules such as Ni(salphen) and Ni(saldmp) monomers, and too much steric hindrance will reduce the electroactivity and then prevent the chemical reactions, while the smaller monomers Ni(salen) will have less steric hindrance with a faster reaction rate in the same condition.

3.2 Electronic diffusion path

Fig. 1f shows the doping levels with the different variation. For salen-type polymer, n = 2y + z. When n < 1, z = 0 (z = 0 or 1), y = n/2, so the polymerisation and redox switching are associated with ligand-based process, and the role of the Ni is purely structural. As for Ni(salphen), the highest doping levels are observed. The value n = 1.15 for Ni(salphen) while all other values are less than 1 at the 1^st scan, and n = 0.66 for Ni(salphen), while n = 0.50 and 0.46 for Ni(salen) and Ni(saldmp) respectively at 10^th cycle. The phenomenon means the Ni(salphen) monomers possess higher charge storage ability, and the value y of Ni(salphen) is more than y of Ni(salen) and Ni(saldmp) (when y = n/2), meanwhile Ni(salen) and Ni(saldmp) almost have the same doping level. If the electron transmission occurs via the metal ion, the electronic transmission path would be same following Ph–O–Ni–O–Ph path, and the value n (or y) should be almost the same with all polymers.

The structure of groups between imine linkages may strongly influence the kinetics of the electronic transfer and lead to a different polymerization rate. In the classical analyses, it's widely considered that the electron diffusion occurs via the Ph–O–Ni–O–Ph bridge in the conjugated polymer backbone during film oxidation.^1,16 But in our previous work,⁹ we have notice there is a small redox peak current in the CV plot for poly[H₂(salen)], in which the O–Ni–O bond in Ni(salen) is replaced by two OH bonds. So we can infer that the electron diffusion will not only occur via the metal ion which acts as a bridge between ligand moieties, but also occur in the bridges via the Ph–C=N–Y–N=C–Ph transmission path. With different bridges between imine linkages, the electronic transmission path is different, and the ethyl has the shortest molecular chain, leading to a fastest electron transport rate. The differences in electrochemical behaviors can be attributed to the imine bridge groups, which establish a bridge between the phenyl moieties and provide electronic transmission path for poly[Ni(salY)] composites.

So in this paper, we hold the view that the electron transmission will also occur in the bridges between imine linkages. For Ni(salphen), n = 3y + z, for three phenyl rings in the polymer. And for Ni(salen) and Ni(saldmp), n = 2y + z, because of two phenyl rings in the polymer. And the values of y are 0.36, 0.38, 0.40 at the 1^st cycle, and 0.25, 0.22, 0.23 at the 10^th cycle for Ni(salen), Ni(salphen) and Ni(saldmp), respectively. Although the doping levels are different and Ni(salphen) shows the maximum values, the electrons through per phenyl ring are almost the same per cycle. With the cycles increasing, the doping levels and the electrons per phenyl ring will decrease. It is because the increasing of deposited mass will improve the cross-linking extent, then leading to a passivation of the electroactive composites. By the way, when n = 1.15, there is still no electron for metal-based redox switching (z = 0), consistent with the variation of subsequent cycles. So during the redox process, the groups between imine linkages can not only act as bridges between ligand moieties in structure, but also provide electronic transmission path and then influence the electrochemical behaviors.

3.3 Physical characteristics

In Fig. 2a, it's noticed that the C=N imine stretching band is observed at around 1630 cm⁻¹ for the all composites in the FTIR spectra, and the band at 2360 cm⁻¹ is nitrile vibration from acetonitrile.^10,16 XPS analyses for the polymers are shown in Fig. 2b–e are detailed Ni 2p spectra for Ni(salen), Ni(salphen) and Ni(saldmp) respectively when applying the potential to 1.2 V by CV. Curve fitting of the Ni 2p_3/2 XPS signals of the polymer film, gives generally a well resolved peaks at 855.5 eV ± 0.3 of binding energy for the three polymers, assigned to the 2p_3/2 binding energy of Ni(II),^17–19 and the results are consistent with the typical of diamagnetic nickel centers in Ni(salen)-type complexes.²⁰ Hence, nickel exists as Ni(II) during the redox process, i.e. z = 0. Besides, in Fig. 2b the peak at 399 eV in the N region is attributed to the nitrogen atom of the C=N bond and the peak of F at 687 eV is BF₄⁻ from the solution.

Fig. 2 (a) FTIR spectra, (b) XPS analysis for the polymers. XPS detailed Ni 2p spectra for (c) Ni(salen), (d) Ni(salphen) and (e) Ni(saldmp) when applying the potential to 1.2 V by CV.

Structure of the electrode and SEM images are shown in Fig. 3. The monomers are deposited on the surface of MWCNTs which is coated on a Ti sheet with the coating mass of 0.3 mg. Morphology of MWCNTs shows the thin tubular intertwist structure with larger gaps, as Fig. 3a. And for composites, the tubes seem much thick and some gaps have been filled in Fig. 3b–d. It can be seen that all morphologies of composites have the similar structure, and poly[Ni(salen)]/MWCNTs composites have more thicker tubes than others, along with some agglomerative phenomenon, as shown in Fig. 3b. The phenomenon is consistent with the mass of deposited polymers that Ni(salen) monomers have fastest polymerization rate than others.

Fig. 3 Structure of the electrode fabricated by Ti–MWCNTs–poly[Ni(salY)] ternary sample and SEM images for (a) MWCNTs, (b) poly[Ni(salen)]/MWCNTs composite, (c) poly[Ni(salphen)]/MWCNTs composite and (d) poly[Ni(saldmp)]/MWCNTs composites.

3.4 Electrochemical behaviors

To evaluate the electrochemical performances of the composites, CV and EIS measurements are performed shown in Fig. 4. CV curves of the three composites measured at the scan rate of 20 mV s⁻¹ and 100 mV s⁻¹ show a reversible redox peak, and this might be aroused by the faradaic reaction of the ligands with reversible redox transition at about 0.6–1.0 V. The faradaic reaction occurred on the surface of electrode can be expressed as follows:

poly[Ni(salY)] + nBF₄⁻ ⇔ poly[Ni(salY)](BF₄)_n + ne⁻

where the redox-active cites are ‘salY’. The CV curves are close to a symmetrical rectangle-like shape at 0–0.6 V, indicating polymer films are in the reduced state with a EDLC-like behavior, with no faradaic process taking place. We can also see that the poly[Ni(salen)] exhibits much higher peak current than others, and it's caused by the more deposited mass of Ni(salen).

Fig. 4 CV plots for the composites at (a) 20 mV s⁻¹, (b) 100 mV s⁻¹, (c) plots of j_pa, j_pc vs. v^1/2 of composites, and (d) EIS plots for composites at 1.2 V.

Another obvious phenomenon is that peak potential (E_p) is different. E_pa for Ni(salen) and Ni(salphen) is about 0.9 V, almost the same, but E_pa for Ni(saldmp) is about 0.7 V, and E_pc for Ni(salen) and Ni(salphen) is about 0.8 V, while E_pc for Ni(saldmp) is about 0.6 V. The shift in E_p means that the ability of gain or loss electrons is different and the polymers have different redox potentials. For poly[Ni(salY)], the difference may be related to the groups between imine linkages which are the unique difference in structure. With the different electronic transmission path, the polymer will have different electronic transmission rate, then the intercalation and deintercalation of the electrolytic ions will be influence by the different molecular structure. When the polymer gains or loses electrons and therefore becomes oxidized or reduced correspondingly, it takes on either a positive or negative charge and the electronic transmission path will be influenced by the groups between imine linkages. So it can be further concluded that the electron transmission will not only occur via the metal ion, but also occur in the imine groups. And the different bridge structure will finally influence the ability of gain or loss electrons, leading to a different redox potential.

Under a semi-infinite diffusion solvent system, the peak current (j_p) in CV plots is given by the Randles–Sevick equation,⁸ j_p = 0.4463nFC(nF/RT)^1/2D^1/2v^1/2, which can be used to estimate the charge transport diffusion coefficient (D). And from Fig. 4c, j_p vs. v^1/2 at room temperature presents an approximately linear dependence. The values n³D are 3.4 × 10⁻⁸ cm² s⁻¹, 1.6 × 10⁻⁸ cm² s⁻¹ and 1.5 × 10⁻⁸ cm² s⁻¹ for poly[Ni(salen)], poly[Ni(salphen)] and poly[Ni(saldmp)], respectively. So the electron transport rate of poly[Ni(salen)] is better than others, leading to the most deposited mass with the thicker polymer layer. However, as for charge transfer resistance (R_ct) shown in Fig. 4d, it's clearly that R_ct is about 0.6, 0.3 and 0.5 ohm for the poly[Ni(salen)], poly[Ni(salphen)] and poly[Ni(saldmp)] electrodes, respectively. It is observed that poly[Ni(salphen)] electrode exhibits lower R_ct among the electrodes and therefore it provides less resistance for the electrolytic ions (mainly BF₄⁻) to interact with the electrode material, which further substantiates the better electrochemical behavior.

The GCD analyses are performed to examine the supercapacitive behaviors of MWCNTs and polymers. Fig. 5a shows the variation of the E vs. t at 1 mA cm⁻². For comparison, MWCNTs are also evaluated and have symmetric change for charge/discharge course, and the specific capacitance is about 10 F g⁻¹ as shown in Fig. 5b. The polymers possess longer charge/discharge time than MWCNTs, demonstrating that the polymers contribute the pseudocapacitance by the reversible redox switching. However, the charge storage abilities are different for the polymers when the groups between imine linkages of the polymers are different. From Fig. 5b, it is obtained that the specific capacitance is about 200 F g⁻¹ for poly[Ni(salphen)], and about 150 F g⁻¹ for poly[Ni(salen)] and poly[Ni(saldmp)]. The Ni(salphen) has the highest doping levels and a faster electronic transmission rate via the bridge than others, and the R_ct is lower among the electrodes, resulting a better specific capacitance. And the phenomenon further indicates that with different groups between imine linkages, the polymers will have different energy storage abilities because the number of electrochemical activity site (i.e. ‘salY’) is different within the molecules. Correspondingly, the poly[Ni(salphen)] has more effective electronic transmission path while the electrons will transmit via the phenyl-imine bridge, and resulted to a higher doping level with a smaller charge transfer resistance, then leading to a better electrochemical performance.

Fig. 5 (a) GCD curves of MWCNTs and the composites at a current density of 1.0 mA cm⁻² under the voltage range of 0 V to 1.2 V, (b) comparison of the specific capacitances at current densities from 0.5 mA cm⁻² to 10 mA cm⁻².

4 Conclusions

In this paper, we first investigate the electropolymerization process of three Schiff base complex Ni(salen), Ni(salphen) and Ni(saldmp), and the polymers have similar chemical structure and morphologies obtained from FTIR, XPS, and SEM. Poly[Ni(salen)] has higher consumed charge and apparent surface coverage in 10 cycles with more deposited mass, and it's mainly because it has smaller steric hindrance by the groups ethyl, which will then lead to a faster polymerization rate. Poly[Ni(salphen)] shows the highest doping level and n = 0.66, while n = 0.50 and 0.46 for Ni(salen) and Ni(saldmp) respectively at 10^th cycle. When we consider the bridges into the electronic transmission path, electrons during the redox switching are 0.25, 0.22, 0.23 per phenyl ring at the 10^th cycle for Ni(salen), Ni(salphen) and Ni(saldmp) respectively, almost the same number of the electrons per phenyl. Then from the electrochemical performance, the peak potential in CV plot is about 0.9 V for Ni(salen) and Ni(salphen), but about 0.7 V for Ni(saldmp). It means the ability of gain or loss electrons is different for the monomers with different groups and the monomers have different redox potential.

Meanwhile the GCD curves display a specific capacitance of about 200 F g⁻¹ for poly[Ni(salphen)], and about 150 F g⁻¹ for poly[Ni(salen)] and poly[Ni(saldmp)]. The charge transfer resistance R_ct shows the similar regulation, about 0.3 ohm for the poly[Ni(salphen)], 0.6 and 0.5 ohm for poly[Ni(salen)] and poly[Ni(saldmp)] respectively. Hence the groups between imine linkages have a crucial influence to the electrochemical behaviors. It can be concluded that the electronic transmission path in nickel Schiff base complex will not only occur via the metal bridge, but also occur in the bridge. As a result, this Ni(salY)-type Schiff base complex can be best described as both poly-phenylene-type and poly-imine-type polymers. The groups between imine linkages can act as bridges between ligand moieties and provide electronic transmission path via the Ph–C=N–Y–N=C–Ph transmission process.

Acknowledgements

This work was financially supported by National Natural Science Foundation of China (No. 51372021), National Science and Technology Support Program (2015BAG01B01), National Natural Science Foundation of China (No. 51572024).

References

1. P. Guo, T. W. Hui, K. C. Cheung, K. Y. Wong and K. K. Shiu, J. Electroanal. Chem., 2001, 498, 142–151.
2. J. Tedim, F. Goncalves, M. Pereira, J. Figueiredo, C. Moura, C. Freire and A. Hillman, Electrochim. Acta, 2008, 53, 6722–6731.
3. O. Buriez, L. M. Moretto and P. Ugo, Electrochim. Acta, 2006, 52, 958–964.
4. D. A. Annis and E. N. Jacobsen, J. Am. Chem. Soc., 1999, 121, 4147–4154.
5. M. Martins, M. V. Boas, B. de Castro, A. R. Hillman and C. Freire, Electrochim. Acta, 2005, 51, 304–314.
6. P. Audebert, P. Hapiot, P. Capdevielle and M. Maumy, J. Electroanal. Chem., 1992, 338, 269–278.
7. C. E. Dahm, D. G. Peters and J. Simonet, J. Electroanal. Chem., 1996, 410, 163–171.
8. F. Gao, J. Li, F. Kang, Y. Zhang, X. Wang, F. Ye and J. Yang, J. Phys. Chem. C, 2011, 115, 11822–11829.
9. G. Yan, J. Li, Y. Zhang, F. Gao and F. Kang, J. Phys. Chem. C, 2014, 118, 9911–9917.
10. M. Vilas-Boas, C. Freire, B. de Castro, P. A. Christensen and A. R. Hillman, Chem.–Eur. J., 2001, 7, 139–150.
11. F. Gao, J. Li, Y. Zhang, X. Wang and F. Kang, Electrochim. Acta, 2010, 55, 6101–6108.
12. C. Freire and B. de Castro, J. Chem. Soc., Dalton Trans., 1998, 1491–1498.
13. M. Vilas-Boas, C. Freire, B. De Castro and A. Hillman, J. Phys. Chem. B, 1998, 102, 8533–8540.
14. M. Vilas-Boas, I. C. Santos, M. J. Henderson, C. Freire, A. R. Hillman and E. Vieil, Langmuir, 2003, 19, 7460–7468.
15. J. Tedim, A. Carneiro, R. Bessada, S. Patricio, A. Magalhaes, C. Freire, S. Gurman and A. Hillman, J. Electroanal. Chem., 2007, 610, 46–56.
16. M. Vilas-Boas, C. Freire, B. de Castro, P. A. Christensen and A. R. Hillman, Inorg. Chem., 1997, 36, 4919–4929.
17. X. Li and W. x. Zhang, Langmuir, 2006, 22, 4638–4642.
18. R. Shalvoy, P. Reucroft and B. Davis, J. Catal., 1979, 56, 336–348.
19. I. G. Casella and M. Contursi, J. Electroanal. Chem., 2013, 692, 80–86.
20. S. Patrício, A. I. Cruz, K. Biernacki, J. Ventura, P. Eaton, A. L. Magalhaes, C. Moura, A. R. Hillman and C. Freire, Langmuir, 2010, 26, 10842–10853.

---