The Effect of Adding CeO₂ Nanoparticles to Cu–Ni–Al Alloy for High Temperatures Applications

Abstract

This work presents the effect of CeO₂ nanoparticles (CeO₂–NPs) on Cu–50Ni–5Al alloys on morphological, microstructural, degradation, and electrochemical behavior at high temperatures. The samples obtained by mechanical alloying and spark plasma sintering were exposed to a molten eutectic mixture of Li₂CO₃–K₂CO₃ for 504 h. The degradation of the materials was analyzed using gravimetry measurements and electrochemical impedance spectroscopy. Different characterization techniques, such as X-ray diffraction and scanning electron microscopy, were used to investigate the phase composition, parameter lattice, and microstructure of Cu–Ni–Al alloys reinforced with CeO₂–NPs. The hardness of the composite was also examined using the Vickers hardness test. Gravimetry measurements revealed that the sample with 1 wt.% CeO₂–NPs presented the best response to degradation with a less drastic mass variation. Impedance analysis also revealed that by adding 1 wt.% CeO₂–NPs, the impedance modulus increased, which is related to a lower porosity of the oxide film or a thicker oxide layer. The microhardness also significantly increased, incorporating 1 wt.% CeO₂–NPs, which reduced with higher CeO₂–NPs content, which is possibly associated with a more uniform distribution using 1 wt.% CeO₂–NPs in the Cu–Ni–Al matrix that avoided the aggregation phenomenon.

1. Introduction

Fuel cells (FC) are considered a promising technology for being an alternative source of electric power [1]. These devices convert chemical energy into electricity [2,3] and can be classified according to the work temperature [4]. For example, a molten carbonate fuel cell (MCFC), the operational temperature of which is 650 °C, is one of the most efficient FCs, and is contemplated as a carbon capture and storage (CSS) technology because it can capture and convert CO₂ [5]. The high-temperature fuel cell (600–1000 °C) uses nickel or other non–precious catalytic materials to decrease the electrode cost [6]. The electrode used as the cathode is a porous nickel oxide, where O₂ and CO₂ are injected, and on the anode side, nickel aluminum alloy is employed, generally Ni5Al, and supplies H₂. According to the report by Lee et al. [7], a charge transfer process controls the electrochemical reactions in a slow reaction system and a mass transfer process in a rapid reaction system. Even though the MCFC has been working for decades, there are still problems to resolve due to the high operational temperature. It is essential to consider that the corrosion of the electrodes and equipment that operate at high temperatures can decrease the FC’s service life, as Hacker and Mitsushima proposed [8]. More specifically, few works have focused on the corrosion phenomenon on the anode of MCFC. In this context, Accardo et al. [9] studied the addition of copper [10,11] and cerium oxide nanoparticles (CeO₂–NPs) to improve the mechanical and electrochemical behavior of the Ni5–Al commercial anode. In this sense, as copper has good electrical and thermal conductivity and mechanical resistance [11,12], the Cu–Ni–Al alloy can be an option for the MCFC anode. On the other hand, some authors reported that incorporating CeO₂–NPs in the Cu–Ni alloy increased the catalyst performance on the anode for the H₂ oxidation reaction [13], and incorporating it in the Ni–5Al alloy reduced the creep strain because it can keep the pore structure stable [9]. However, no available data evaluate the micro–macrostructural and corrosion behavior of a Cu–Ni–Al alloy reinforced with CeO₂ nanoparticles used for the anodes in MCFC.

Therefore, the effect of nanoparticle addition on the degradation of Cu–50Ni–5Al in the electrolyte Li₂CO₃–K₂CO₃ (62–38 mol.%) is analyzed in this paper. The microstructural and morphological changes during this process provide valuable insights into the potential use of nanoparticles in the anode.

2. Materials and Methods

2.1. Sample Obtention

Mechanical alloying was performed using pure powders: Nickel (<10 μm, 99+, Merck, Darmstadt, Germany), Copper (<63 µm, >99%, Sigma-Aldrich, Darmstadt, Germany), and Aluminum (<60 μm, 99.9%, Good Fellow, Hamburg, Germany), which were mechanically alloyed (MA) in a Planetary mill PQ4 Across International, obtaining powder compositions of Cu–50Ni–5Al (wt.%). The milling conditions included a ball–a–powder ratio (BPR) of 10:1 and 2 wt.% stearic acid as a control agent under an inert Ar atmosphere. The milling time used was 100 h effective and there was an on/off cycle of 30/15 min at a speed of 350 r.p.m. Subsequently, 1, 3, and 5 wt.% of the nanoparticles CeO₂ (CeO₂–NPs) (<25 nm, >99.9%, Sigma-Aldrich, Hamburg, Germany) were added to the alloy using Mixer Y–type Astecma for 1 h.

Table 1. Chemical composition of the Cu–50Ni–5Al + xCeO₂ (wt.%) alloys.

Sample	Cu	Ni	Al	CeO2
0 wt.% CeO2–NPs	Bal.	50	5	0
1 wt.% CeO2–NPs	Bal.	50	5	1
3 wt.% CeO2–NPs	Bal.	50	5	3
5 wt.% CeO2–NPs	Bal.	50	5	5

shows the chemical composition of Cu–50Ni–5Al + xCeO₂ (wt.%) alloys.

Cu–50Ni–5Al without and with the CeO₂–NPs samples were consolidated by Spark Plasma Sintering (SPS) using a Fuji Electronic Industrial Co model DR. SINTER^® SPS1050. The disks were 10 mm in diameter and 7 mm in thickness, and were produced using a high–density graphite die. The samples were heated from room temperature to 800 °C at a heating rate of 100 °C min⁻¹, applying a pressure of 50 MPa simultaneously during the heating and holding time of 5 min at the sintering temperature. The entire SPS process was kept under a vacuum of approximately 20 Pa. Finally, the samples were free–cooled to room temperature at a cooling rate of roughly 10 °C s⁻¹ in the SPS chamber. Figure 1 shows the sample preparation (powders) diagram by mechanical alloying and mechanical mixing and a schematic representation of the consolidation process by the SPS system.

The metal samples were polished using sandpaper from #800 to #4000 and then with colloidal silica suspension to reveal their microstructure. The polished samples were rinsed with ethanol for 10 min in a bath cleaning sonicator. Afterward, the samples were cleaned with distilled water and dried at room temperature.

2.2. Gravimetric Measurements

The Cu–50Ni–5Al samples were immersed in molten eutectic Li₂CO₃–K₂CO₃ (62–38 mol.%) [14,15] for 504 h (21 days) in an aerated atmosphere or not–controlled medium at 550 ± 5 °C. The gravimetric measurements of samples were carried out as described previously by Arcos et al. [16]. The samples were removed and cleaned to eliminate the deposits and corrosion products on the surface. The bulk samples were submerged in hot distilled water (~100 °C) in a sonicator bath (Elma D–78224 Singen/Htw) for 30 min. Subsequently, the samples were dried with hot air and weighed until they reached a constant value, as reported by the ASTM G1–03 [17]. The average mass (%) was calculated using Equation (1).

$$\frac{{\mathrm{m}}_{\mathrm{i}}-{\mathrm{m}}_{\mathrm{f}}}{{\mathrm{m}}_{\mathrm{i}}}\times 100$$

(1)

where ${\mathrm{m}}_{\mathrm{i}}$ and ${\mathrm{m}}_{\mathrm{f}}$ are the initial and final sample masses at different exposure times.

2.3. Morphological and Chemical Characterization

The porosity was determined through Archimedes’ method, according to the Standard Test ASTM C373–88 [18]. To reveal the microstructure, an etching was employed: 5 g of Fe₃Cl, 10 mL of HCl, and 100 mL of distilled water for 8 s. A field–emission scanning electron microscope (FE-SEM), QUANTA FEG 250, was used to obtain the samples’ images.

X-ray diffraction (XRD) was implemented using Bruker D2 PHASER with Cu–Kα radiation to characterize the material’s structure. The diffraction patterns were recorded from 2θ between 40° and 100° with a 0.02° step and counting time of 1 s/step.

2.4. Mechanical Characterization

The microhardness of the samples was calculated by a micro–Vickers durometer, Wilson^® VH1150 Macro Vickers Hardness Tester, under 0.3 kgf of force before the gravimetric measurements.

2.5. Electrochemical Measurements

The electrochemical behavior of the Cu–50Ni–5Al samples was studied using open circuit potential and electrochemical impedance spectroscopy (EIS) measurements in the molten Li₂CO₃–K₂CO₃ (62:38 mol.%) as an electrolyte at 550 ± 5 °C under an aerated atmosphere. The electric contact for the working electrode, the Cu–50Ni–5Al samples, was performed using conductive silver printing ink (resistivity 5–6 µΩ cm) around the sample and copper wire of 25 cm in length. In addition, a Pt wire that was 25 cm in length was used as a counter electrode, and an Ag wire that was 25 cm in length and placed inside a quartz glass tube with a porous plug in the tip was used as the reference electrode. The electrochemical measurements were carried out with a Potentiostat Solartron Analytical.

3. Results and Discussion

3.1. Gravimetric Measurements

Figure 2 shows the effect of CeO₂–NPs on the weight gain of the Cu–50Ni–5Al samples after exposure to molten carbonates, revealing a reduction in the weight by adding CeO₂–NPs. In addition, Figure 2 shows that during the initial stage, the weight increased rapidly for all the Cu–50Ni–5Al samples, which was lower for Cu–50Ni–5Al + 1 wt.% CeO₂–NPs, reaching a maximum weight gain of only 1.2% at 504 h of exposure. For a longer exposure time, the weight gain reaches a stationary state associated with a passive oxide layer [16] formed after 80 h for Cu–50Ni–5Al + 3 wt.% CeO₂–NPs, and after 160 h for Cu–50Ni–5Al + 0 wt.% CeO₂–NPs. It should be noted that with 5 wt.% CeO₂–NPs, the weight gain did not reach a plateau of up to 504 h of exposure, like the sample with 3 wt.% CeO₂–NPs. Therefore, the sample that suffered the least degradation at high temperatures was 1 wt.% CeO₂–NPs, which could be attributed to the homogeneous distribution of the CeO₂–NPs.

3.2. Microstructural and Chemical Characterization

Figure 3 shows the morphology of the samples before and after 504 h of exposure to molten carbonate (Li₂CO₃–K₂CO₃ 62–38 mol.%) in the aerated atmosphere. Before exposure, porosity can be observed for all the samples, agreeing with the analysis performed. For example, Cu–50Ni–5Al had 16 ± 0.8% porosity, which reduced the addition of CeO₂–NPs. The Cu–50Ni–5Al + 1 wt.% CeO₂–NPs samples had 1.0 ± 0.3% porosity, Cu–50Ni–5Al + 3 wt.% CeO₂–NPs had 2.0 ± 0.2% porosity, presenting some spots more lightly over the surface, and Cu–50Ni–5Al + 5 wt.% CeO₂–NPs had 1.0 ± 0.1% porosity. One reason for this phenomenon is that nanoparticles can easily remain in the pores and voids of the nanocomposite matrix due to their small size [19]. After exposure, a strong surface modification was observed for all the samples, possibly due to the corrosion product formation, which could be a passive film, as suggested by the gravimetric measurements. Ren et al. [20] studied a Cu–35Ni–10Al alloy in molten carbonate (Li₂CO₃–K₂CO₃ 62–38 mol.%) for 48 h in an aerated atmosphere, reporting the formation of porous corrosion products composed mainly of Al₂O₃ and Ni–Al oxides, which is in agreement with the SEM images of Cu–50Ni–5Al (see Figure 3a), which present a porous surface after exposure.

Figure 4 presents the EDS results before and after exposure to analyze the chemical composition of the sample’s surface. Before exposure, the alloy’s surface is very similar for all the samples, revealing a homogeneous distribution of all the elements (Cu, Ni, Al, and O). However, Cu–50Ni–5Al has some Al spots, and Cu–50Ni–5Al + 3 wt.% CeO₂–NPs have some zones not identified by the mapping, which could be Li because it has deficient energy and is difficult to detect. Cu–50Ni–5Al+ 5 wt.% CeO₂–NPs present some nanoparticles agglomeration (CeO₂) corresponding to the element Ce. Frattini et al. [21] observed the same effect when adding small amounts of ZrO₂–NPs in the Ni–Al alloy, although, with the increase in the amount to 10% ZrO₂–NPs, the distribution of the NPs becomes homogeneous. After exposure, the quantity of oxygen (O) increased significantly for all the samples, which can be attributed to the oxide formation on the surface, as proposed above. Nevertheless, potassium (K) was also found on the surface, a component of the molten carbonates (Li₂CO₃–K₂CO₃), revealing the possible formation of a deposit. According to Gonzalez–Rodriguez et al. [22], the Ni–50Al alloy was immersed in Li₂CO₃–K₂CO₃ (62:38 mol%) for 100 h in static air at 650 °C. They reported that the main corrosion products were Ni, Al, and K, such as NiO, Al₂O₃, LiAlO₂, and LiKCO₂. Also, Ren et al. [20] described mainly Al₂O₃ and Cu₂O as the corrosion products of Cu–35Ni–10Al exposure to Li₂CO₃–K₂CO₃ (62:38 mol%) during 1 h in air at 650 °C. Accardo et al. [9] also proposed that CeO₂–NPs could migrate from the alloy to the electrolyte. EDS analysis revealed a slight decrease in the Ce content for the Cu–50Ni–5Al + 5 wt.% CeO₂–NPs sample. For the other samples, the Ce content increases, which can be associated with diffusion from the bulk to the surface of the molten carbonates.

Figure 5 compares the XRD patterns recorded before and after exposure to Li₂CO₃–K₂CO₃ (62:38 mol%). Before exposure, the reflections were identified as corresponding to typical fcc structures (Fm-3m). No peaks were associated with Ni or Al, indicating that the solid solution Cu–Ni–Al obtained by mechanical alloying is maintained post–sintering by SPS. The samples reinforced with CeO₂–NPs show low-intensity reflections associated with CeO₂ (Fm-3m; JCPDS 010750076). The lattice parameter of the Cu-Ni–Al alloys without CeO₂–NPs is 0.358 nm, which remains constant when incorporating the different CeO₂–NPs. This indicates that the CeO₂–NPs do not react with the Cu–Ni–Al matrix in the consolidation process because the SPS technique is a fast method for sintering [23]. After 21 days of exposure to the Li₂CO₃–K₂CO₃, peaks associated with NiO (Fm-3m; JCPDS 010731519), Cu₂O (Pn-3m; JCPDS 010751531), and Al₂O₃ (R-3c; JCPDS 010772135) can be seen in all the samples. In addition, the intensities of the reflections associated with CeO₂–NPs can be seen, which can be attributed to the fact that the CeO₂–NPs migrate to the surface, as reported by Accardo et al. [9]. Note that the increase in the intensity of the reflections associated with CeO₂ is much lower for the sample with 5 wt.% of CeO₂–NPs than in the samples with 1 wt.% and 3 wt.% of CeO₂–NPs, possibly because the Ce content decreased on the sample surface, confirming what was previously mentioned in Figure 4.

3.3. Mechanical Properties

Figure 6 shows the hardness of the Cu–50Ni–5Al + x wt.% CeO₂–NPs samples before exposure. The results indicate that the sample with 0% CeO₂–NPs presents the lowest hardness value, corresponding to 205 ± 21 HV_0.3. This can be attributed to the higher porosity of the sample, which reaches 16%, by incorporating different amounts of CeO₂–NPs in the sample. Cu–50Ni–5Al +1 wt.% CeO₂–NPs perform better due to the CeO₂–NPs allowing a decrease in the porosity and a homogenous distribution of CeO₂–NPs in the matrix, as mentioned above. However, if the concentration of CeO₂–NPs exceeds 1 wt.%, they agglomerate at grain boundaries, reducing the hardness value, as seen in Figure 4. Zawrah et al. [24] concluded that adding Al₂O₃–NPs to pure Cu improves hardness due to their uniform distribution. The improved hardness can be attributed to the relative contribution of the Orowan strengthening effect, mainly when the reinforcement size is less than 100 nm [25]. The CeO₂–NPs are very small and hard, impeding the movement of dislocations in the Cu–50Ni–5Al matrix, leading to an improvement in the hardness of the microstructure.

3.4. Electrochemical Measurements

Figure 7 shows the effect of the addition of 1 wt.% CeO₂–NPs to the open circuit potential (E_OC) of Cu–50Ni–5Al after exposure to Li₂CO₃–K₂CO₃ at 550 °C and an aerated atmosphere. After a shorter exposure time, the E_OC was shifted to more negative values by incorporating CeO₂–NPs, suggesting an activation of the corrosion phenomena. However, for a longer exposure time, the E_OC reached similar values to the sample without CeO₂–NPs, which can be associated with a stable oxide layer formed on the metal surface. Meléndez-Ceballos et al. [26] studied a Ni porous sample coated by CeO₂–NPs using the atomic layer deposition in Li₂CO₃–K₂CO₃ at 650 °C in a CO₂/air 30/70 vol.% atmosphere. The authors determined an initial potential close to -0.76 V vs. Ag/Ag⁺; the reference electrode is a silver wire submerged in Ag₂SO₄ (10⁻¹ mol kg⁻¹) saturated in Li₂CO₃–K₂CO₃, which was shifted to more positive values as a function of exposure time, attributed to a delay in the Ni oxidation process due to the presence of CeO₂–NPs film.

Figure 8 shows the Nyquist diagrams of Cu–50Ni–5Al after 1 h of exposure to Li₂CO₃–K₂CO₃ in an aerated atmosphere at E = E_OC and 550 °C, revealing a significant increase in the impedance modulus due to the incorporation of 1 wt.% CeO₂–NPs, which could be related to the formation of a passive oxide layer on the alloy, as previously mentioned. As can be seen at E = E_OC, the impedance responses reveal two time constants at high and low frequency ranges (HF and LF), which can be associated with the cathodic current, not only involving the capacitance of the electric double layer (C_dl) and the oxygen reduction reaction, but also the formation of an oxide layer due to the alloy dissolution. Different equivalent circuits have been proposed to represent the physical model, which can be composed of capacitors, resistances, and constant phase elements (CPE) related to the heterogeneity of the surface [27].

Figure 9 shows the Bode plots of Cu–50Ni–5Al + 1 wt.% CeO₂–NPs after 1 h of exposure to Li₂CO₃–K₂CO₃ in an aerated atmosphere at E = E_OC and 550 °C, as a representative example of the study system. Figure 9 reveals a capacitive response with two or three time constants at all frequency ranges, possibly associated with the formation of an oxide film and oxygen reduction reaction. Additionally, the Bode plots revealed a higher impedance modulus at the LF range when CeO₂–NPs were added, which can be related to the polarization resistance of the system, suggesting an enhancement of the corrosion resistance due to the incorporation of the CeO₂–NPs to the Cu–50Ni–5Al matrix [27,28,29]. Moreover, Figure 9 shows the non–corrected () and corrected Bode plots by electrolyte resistance (), revealing that this effect is mainly in the high frequency range.

Figure 9c shows the variation of the imaginary part of the impedance of Cu–50Ni–5Al + 1 wt.% CeO₂–NPs as a function of frequency after 1 h of exposure to Li₂CO₃–K₂CO₃ in an aerated atmosphere at E = E_OC and 550 °C, revealing a constant phase element (CPE) behavior in the MF range, with a negative slope (α) that varied between −0.33 ± 0.01 for Cu–50Ni–5Al and -0.65 ± 0.004 for Cu–50Ni–5Al + 1% wt.% CeO₂–NPs, which can be related to the oxide film formed on the metal surface and described by the following relation, as reported by Orazem and Tribollet [30], Tribollet et al. [31], and Hirschorm et al. [32].

$$Z_{oxide}={\int }_0^{\delta }\frac{\rho \left(\gamma \right)}{1+j\omega \rho \left(\gamma \right)\epsilon \left(\gamma \right){\epsilon }_0}d\gamma$$

(2)

The authors proposed the power-law model (PLM) to analyze the film properties using Equation (3):

$$Z\left(\omega \right)=g\frac{\delta {\rho }_{\delta }^{1-\alpha }}{{\left({\rho }_0^{-1}+j\omega \epsilon {\epsilon }_0\right)}^{\alpha }}$$

(3)

In this case, α is the slope in the Log Z_Imag vs. Log f plots, ε represents the dielectric constant of the oxide layer formed on the metal alloy, ε₀ is the vacuum permittivity that is equal to 8.85 × 10⁻¹⁴ F·cm⁻¹, and $g$ is a numerical coefficient close to 1 when α is 1, which can be estimated using the following equation:

$$g=1+2.88{(1-a)}^{2.375}$$

(4)

In addition, ρ₀ and ρ_δ represent the lower and upper limits in the frequency range where CPE behavior is observed. The Q value corresponds to a CPE parameter that can be determined using the following equation:

$$Q=\frac{{\left(\epsilon {\epsilon }_0\right)}^{\alpha }}{g\delta {\rho }_{\delta }^{1-\alpha }}$$

(5)

The graphical method of the impedance data allowed us to estimate the CPE parameters for the Cu–50Ni–5Al Q coefficient of 1.62 × 10⁻³ F·cm⁻²·s^–(1−α) and |α| value of 0.60, and for the Cu–50Ni–5Al + 1% wt.% CeO₂–NPs, the Q coefficient of 6.95 × 10⁻³ F·cm⁻²·s^–(1−α) and |α| value of 0.67. Furthermore, electrolyte resistance (Re) was determined by the graphical method, finding values close to 2.7 Ω cm² for Cu–50Ni–5Al and 0.84 Ω cm² for the Cu–50Ni–5Al + 1% wt.% CeO₂–NPs samples. Those parameters reveal that the addition of CeO₂–NPs to Cu–50Ni–5Al improves the corrosion resistance, possibly due to a lower porosity of the oxide film or a thicker oxide layer.

4. Conclusions

The Cu–50Ni–5Al alloys were obtained by mechanical alloying and the SPS process, which allowed the addition of CeO₂–NPs into the metal matrix. The samples were exposed to molten carbonate (Li₂CO₃–K₂CO₃), consequently forming corrosion products over the surface, which were analyzed by SEM-EDS and corroborated by X-ray diffraction. The addition of CeO₂–NPs in the Cu–50Ni–5Al matrix reduced the mass variation over time, mainly using 1% wt. CeO₂, reaching a maximum weight gain of only 1.2% at 504 h of exposure. The corrosion products were composed of nickel oxide, aluminum oxide, and copper oxide in all the alloys, regardless of the amount of NPs incorporated.

Moreover, the microhardness significantly increased for the alloy containing 1 wt.% of CeO₂–NPs, reaching a hardness value of 340 HV_0.3. Furthermore, the impedance analysis revealed that the samples with 1 wt.% of CeO₂–NPs in the molten carbonates in an aerated atmosphere had a higher impedance modulus, possibly due to the lower porosity of the oxide film or a thicker oxide layer.

Therefore, the alloys that showed better mechanical behavior and higher corrosion resistance were Cu–50Ni–5Al + 1 wt.% CeO₂–NPs, which have a promissory use at high temperatures.