3.1.1. X-ray analyses
The Cu@PtRu/C structure is confirmed by XRD which shows partial core-shell formation for Cu@PtRu/C-16 and Cu@PtRu/C-18, and full core-shell formation for Cu@PtRu/rGO-16 crystallites. These structures are a consequence of the substitution of atoms on the surface of Cu nanoparticles by Pt and Ru atoms.
Figure 1 shows the XRD pattern for Pt/C-30, PtRu/C-30, Cu@PtRu/C-16, Cu@PtRu/C-18, and Cu@PtRu/rGO-16. For all PtRu-containing catalysts there is a slight shift in relation to the Pt/C-30 diffraction peak in the (111) plane. These shifts to values higher than 2Ɵ are indicative of the formation of PtRu alloy, either directly on the support, or in the Cu crystallites. The formation of PtRu can also be inferred from the values calculated for the lattice parameters of the PtRu catalysts that present lower values in relation to the monatomic Pt alloy. The first peak located at about 2Ɵ = 25º for Pt/C-30, PtRu/C-30, Cu@PtRu/C-16, and Cu@PtRu/C-18 is associated with the face (002) of the crystal structure of the carbon. The absence of this peak is noticed in Cu/rGO and Cu@PtRu/rGO-16 catalysts, and is associated with the efficiency in the process of exfoliation of the graphite sheets by chemical routes. Exfoliation increases the distance among the graphite planes as the oxidation process proceeds, which according to Scherrer equation (d=kλ/βcosƟ) implies in smaller angles. The other three peaks at 40°, 47° and 68° are associated with the (111), (200) and (220) planes, characteristic of face-centered cubic crystal alloys (fcc) of Pt [
23,
24,
25,
26].
As can be seen in
Figure 1, the crystallinity of Cu@PtRu/C is high and the presence of peaks characteristic of the Pt structure suggests that the layer covering the core is formed, or specifically, in this case, implies that the average size of the PtRu alloy on the surface of Cu nanoparticles is higher than 1 nm [
27,
28,
29,
30,
31,
32,
33]. The reason for the statement that the Cu@PtRu/rGO-16 nanoparticles (NPs) forms a perfect core-shell system is the absence of the standard diffraction peaks characteristic of the indication of Cu atoms and their oxides at 2Ɵ equal to 36º and 43º. The presence of these peaks for the Cu@PtRu/C-16 and Cu@PtRu/C-18 alloys suggests that the coating of Cu NPs in these alloys is only partial. The average size of the crystallites is 3.6 nm (Cu/rGO), 3.8 nm (Pt/C -30), 3.4 nm (PtRu/C-30), 4.9 nm (Cu@PtRu/C-16), 4.4 nm (Cu@PtRu/C-18) and 4.5 nm (Cu@PtRu/rGO-16)
3.1.2. HRTEM Results
The core-shell structures have been also characterized by HRTEM in order to compare with the XRD results and visualize changes in the distances between the crystallographic planes due to the substitution of atoms on the NPs surfaces that form the nuclei.
Figure 2A–E display representative images obtained using HRTEM. From these images, the spherical shape is assumed for the Cu/rGO and Cu@PtRu/rGO-16 NPs and a regular dispersion of these NPs on the rGO support is observed. The average diameter measured for the Cu/rGO NPs is 3.7 nm and for the Cu@PtRu/rGO-16 NPs is 4.9 nm, suggesting that a layer with a size above 1nm of PtRu is deposited on the surface of the Cu particles. The growth of Pt alloys on the surface of copper NPs can be understood as an expansion of the copper crystal lattice, where in general the outer layer is favored by two factors: the higher oxidation potential of constituent metals of the core in relation to that of the metals concerning the covering layer, and because this layer has the same crystalline arrangement of the core atoms [
34,
35,
36,
37,
38].
The PtRu formation on Cu NPs is the result of charge transfer at the Cu/solution interface, and the increase in size observed by HRTEM of Cu/rGO NPs is attributed to the reduction of PtCl
42- and Ru
+4 species. This reduction is favored over the Cu core because the interaction between the metal ions and the copper atoms is stronger than the interaction between them and the carbon support. Another factor favorable to the growth of the PtRu layer on the copper core is the crystalline arrangement of Pt and Cu, both FCC. From
Figure 2A,B, it is possible to observe, in darker contrast, the formation of agglomerates of Cu/rGO particles resulting from the employed synthesis method. Regularly distributed regions are represented by NPs in less dark contrast.
Figure 2D depicts a better distribution in the support for the Cu@PtRu/rGO-16 NPs, suggesting that the second step of the synthesis is a stage of reconfiguration of the particles distribution on the support.
The brightness contrast between the core and shell regions observed in some Cu@PtRu/rGO NPs in
Figure 2D is commonly attributed to the separation between the Cu core, the darkest region, and the PtRu shell, the brighter region. However, the boundary between the core and the shell is not well defined in the images and the separation between the phases is better defined by measuring the interplanar distance of the crystal constituting the core and subsequently the formed core-shell. The lattice planes of the Cu/rGO-30 NPs in
Figure 2C exhibit an interplanar spacing of 0.2206 nm, corresponding to the (111) face of the copper alloy. In this case, a single particle is chosen arbitrarily. In
Figure 2E, the distance measured between the two closest rows of atoms for the Cu@PtRu/rGO-16 NPs is 0.2317 nm, referring to the orientation of the (111) crystalline plane of the FCC structure of platinum. According to the XRD results that show the absence of standard peaks characteristic of the copper crystal structure, it can be concluded that the distance between the planes in
Figure 2E refers to the PtRu alloy.
Eighty particles are randomly measured to obtain the average particle size distribution of Cu@PtRu/rGO-16. The HRTEM results suggest that, in some cases, discrete Pt and PtRu particles are formed, however, this statement refers to a few measurements of particles having sizes below the measured values for most Cu@PtRu/rGO-16 NPs obtained either by HRTEM or by the Scherer equation.
3.1.3. Cyclic voltammetry
The electrochemical performance of Pt/C-30, PtRu/C-30 Cu@PtRu/C-16, Cu@PtRu/C-18 and Cu@PtRu/rGO-16 catalysts have been investigated by cyclic voltammetry in sulfuric acid, 0.5 mol.L
-1H
2SO
4, at 20 mV/s. The cyclic voltammograms of the catalysts are shown in
Figure 3. Each voltammogram is obtained after 22 cycles under N
2 atmosphere, in order to obtain a more stable response from the electrode surface.
The currents of the cyclic voltammetry and chronoamperometry curves of all electrocatalysts have been normalized using the oxidation method of a monolayer of CO adsorbed on the electrodes surface. This method is used to compare the electrocatalytic activity of different electrodes, taking into account the number of surface active sites, i.e., it is an evaluation of the electrocatalytic activity intrinsic to the real area of these electrodes. The basis of this method is the fact that carbon monoxide molecules occupy, in principle, all active sites available for electrocatalytic reactions. The charge value of the conversion of CO
ads to CO
2(g) [Q
CO(C)] in microColoumbs (µC) divided by 420 µC is numerically equal to the electrochemically active area (EAA), defined by equation (I). The EAA values obtained for each electrode are used as normalization factors for the currents that are now expressed in A.cm
-2 units. The CO amount is obtained by integrating the voltammetric peak (shaded area), as shown in
Figure 4 [
39].
The adsorption and oxidation of methanol occurs in few materials, in acidic medium, only in platinum and platinum alloys. In electrochemical systems, interest is the oxidation of methanol on the anodic substrate. This process goes through different stages depending on the applied potential values [
40,
41,
42].
Despite containing a single carbon atom in its molecule, methanol has electro-oxidation in the presence of PtRu electrocatalysts in a series of steps which chemical equations below show only the additional reactions suggested by Hamnett [
43]. Other similar proposals can be seen in other works dealing with the methanol oxidation mechanism [
44,
45].
There are two more relevant items regarding the mechanism of methanol oxidation. The first is that methanol adsorption takes place on energetically favored sites on the surface. The second is the occurrence of sequential loss of protons to give rise to a sequence of multiply bonded hydrogenated intermediates that convert to linearly adsorbed CO. This evidence comes from the results obtained by DEMS (Differential Electrochemical Mass Spectrometry) and FTIR (Fourier transform infrared) [
46,
47].
The adsorbed CO molecules (2) resulting from the partial oxidation of methanol are undesirable, since they adsorb on the surface of the platinum atoms where fuel molecules must oxidize; this poisoning of the catalytic surface results in a decrease in the activity of the electrocatalyst [
48].
To circumvent the problem of poisoning the catalyst by CO, two effects can be considered from the formation of the metallic alloy of Pt with a second metal, in particular with Ru: the bifunctional mechanism and the electronic effect act together, leading the CO to oxidize at low potentials [
13]. In addition, there is the free energy that originates at the interface between the core and the layer that covers it. The core-shell morphology presents an anisotropy of forces at its interface, this inequality of forces on the atoms at the phase boundary is caused by the different electronic arrangements of the atoms in that region, by the difference in crystalline structures and, in the case of the same structure, difference in the lattice parameter of the crystalline arrangement adopted by NP. The sum of these factors causes the chemical potentials to be different in the different phases of the core-shell structure, which will cause this system to release energy at appropriate potentials, in order to reach a lower energy configuration. The synergy of these effects are probably the main factors contributing to more CO tolerant PtRu catalysts [
49,
50,
51,
52].
In
Figure 5, the curves obtained for the anodic scans referring to the oxidation of methanol on Pt/C-30, PtRu/C-30, Cu@PtRu/C-16, Cu@PtRu/C-18, and Cu@PtRu/rGO-16 electrocatalysts are superimposed. These curves have been obtained in solutions containing 0.5 M CH
3OH plus 0.5 M H
2SO
4, at 20 mV/s, in the potential range between -0.2 and 0.8 V vs. Ag/AgCl.
The anodic scans shown in
Figure 5 display that for the region of technological interest (0.1 to 0.4 V), the Cu@PtRu/rGO-16 electrocatalyst depicts higher current density value than the others. Similarly, its cyclic voltammograms in 0.5 M H
2SO
4 solution show higher current density values. The second best is Cu@PtRu/C-16, followed by Cu@PtRu/C-18, both with core-shell structure. Although the Cu@PtRu/C-16 and Cu@PtRu/C-18 catalysts show lower current densities than that of Cu@PtRu/rGO-16, their current densities are higher than those of Pt/C-30 and PtRu/C-30.
Starting oxidation potential (SOP) is an important criterion used to evaluate and compare the catalytic activities of electrodes. The values measured for the SOP of each catalyst are shown in
Table 1. The oxidation-reduction peaks of the Cu@PtRu/rGO-16 catalyst are lower than all others. Cu@PtRu/C-16, Cu@PtRu/C-18 have lower SOP values than those of Pt/C-30 and PtRu/C-30. The PtRu/C electrocatalyst, although it has lower SOP value than that of Pt/C, has a higher value than the core-shell morphology electrocatalysts. Among the core-shells, Cu@PtRu/C-16 displays the lowest value in relation to Cu@PtRu/C-18. Among all tested electrocatalysts, Cu@PtRu/rGO-16 exhibits the lowest SOP value, 0.250 V. Thus, the order of catalytic activities, by the SOP value criterion, is: Pt/C < PtRu/C-30< Cu@PtRu/C-18 < Cu@PtRu/C-16 < Cu@PtRu/rGO-16.
The better performance of Cu@PtRu/rGO-16 electrocatalyst in relation to the others for methanol oxidation can be attributed to its high active surface, the higher number of oxygen-containing groups attached to the ruthenium atom (carboxyl, hydroxyl, etc.), the bifunctional mechanism and the electronic effect, and to its efficient exfoliation process or the higher thermal and mechanical stability of its carbon support [
53,
54].
The ratio between forward (i
F) and reverse anodic (i
R) peak current densities can be used to describe the catalyst's tolerance to catalytic poison accumulation [
55,
56]. A high i
F/i
R value indicates more effective removal of species that poison the catalyst surface.
Table 1 shows the calculated values for i
F, i
R and i
F/i
R.
The Pt/C-30 (1.80 A/cm2) and PtRu/C-30 (1.96 A/cm2) electrocatalysts demonstrate the lowest iF values in relation to the Cu@PtRu/C-16 (2.79 A/ cm2) and Cu@PtRu/C-18 (2.89 A/cm2). The iF for Cu@PtRu/rGO-16 electrocatalyst (2.68 A/cm2) is lower than that of the other core-shell systems, however, it possesses the lowest iR (1.85 A/cm2) compared to the others, except for PtRu/C-30 (1.78 A/cm2). These results show that the Cu@PtRu/rGO-16 electrocatalyst has the highest iF/iR value (1.45 A/ cm2) among all the studied ones herein. In general, the iF/iR ratio is a measure of the efficiency of the bifunctional mechanism and the electronic effect on the electrode kinetics.
3.1.4. Chronoamperometry measuments for methanol electrooxidation
Chronoamperometric experiments have been carried out in solution of 0.5 M methanol plus 0.5 M H
2SO
4 to observe the stability of the catalysts over time at constant potential. The results are shown in
Figure 6.
The results obtained by chronoamperometry demonstrate that the PtRu/C-30, Cu@PtRu/C-16, Cu@PtRu/C-18 and Cu@PtRu/rGO-16 systems exhibit higher current density values than that of the standard Pt/C, after 20 minutes of operation in a potential value of 500 mV. This fact is probably due to the bi-functional mechanism and/or the electronic effect resulting from the presence of ruthenium in the alloy. However, the core-shell electrocatalysts show higher stability and higher current density values than those of PtRu/C-30 and the conventional Pt/C-30, again indicating the beneficial effect of the core-shell structure for the electrocatalysts performance.
The current density value of Cu@PtRu/rGO-16 catalyst is about 4 times higher than that of Cu@PtRu/C-16 and Cu@PtRu/C-18 catalysts. In addition, it is about 5 times higher than that of PtRu/C-30 catalyst and around 6 times higher than that Pt/C-30. These results suggest that Cu@PtRu/rGO-16 exhibits high long-term bulk activity for methanol electrooxidation, high stability, and high tolerance to CO poisoning. The lower instability of Cu@PtRu/C-16 and Cu@PtRu/C-18 catalysts may be related to the pseudo core-shell morphology that allows the presence of copper and its oxides on the surface of the NPs, which according to the XRD results, does not take place in Cu@PtRu/rGO-16. The chronoamperometric results are further supported by the HRTEM images that display a regular distribution of their NPs over the rGO support.
In Cu@PtRu/C-16 and Cu@PtRu/C-18 electrocatalysts, the irregular formation of the layers that cover the core suggests that PtRu is reduced on the support in the form of a bimetallic alloy or in the form of monatomic Pt and Ru alloys. In the latter case, the absence of the bifunctional mechanism and the electronic effect on dispersed Pt NPs significantly affect the performance of these catalysts.
The following order of catalytic activities can be assigned approximately after 20 minutes of reaction: Pt/C < PtRu/C-30< Cu@PtRu/C-18 < Cu@PtRu/C-16 < Cu@PtRu/rGO-16. The results obtained by chronoamperometry agree with those obtained by cyclic voltammetry in 0.5 M H2SO4 solution, in the sense that the best electrocatalyst is Cu@PtRu/rGO-16 and the worst is Pt/C.
The higher current density values for Cu@PtRu/rGO-16 in relation to the other electrocatalysts, suggest that the surface tension and other characteristic properties of the interface may be the cause of its better performance. This energy originating at the core-shell interface may, as well as the electronic effect, be associated with the weakening of the Pt-COads bonds on the surfaces of the core-shell NPs. Thus, the free energy released at the phase boundary inside NPs with core-shell morphology, added to the electronic effect and the bi-functional mechanism, may be the cause of the greater tolerance to CO, and the consequent increase in the electrocatalytic activity of electrocatalysts with core-shell morphology. The preponderance of one of these effects depends on the potential range considered. In fact, the increased tolerance of the PtRu alloy to CO has the effect of the preponderant ligand at potentials below 0.3V, while the bifunctional mechanism at potentials above 0.3V [
57,
58,
59,
60].
The higher values of electrical current densities for Cu@PtRu/rGO-16 must be associated with the higher conductivity of rGO compared to Vulcan XC72R, thus, rGO support is crucial for the best performance of the Cu@PtRu/rGO-16 electrocatalyst.