Ternary Fe- or Mo-Au-Ni/GDC as Candidate Fuel Electrodes for the Internal Dry Reforming of CH₄: Physicochemical and Kinetic Investigation

Abstract

The present study deals with the physicochemical and catalytic/kinetic investigation of Fe, Au, Fe-Au, and Mo-Au modified Ni/GDC electrocatalysts towards their performance for the DRM, RWGS, and CH₄ decomposition reactions. For this purpose, Au-NiO/GDC (where Au = 1 or 3 wt.%), Fe-NiO/GDC (where Fe = 0.5 or 2 wt.%), 0.5Fe-3Au-NiO/GDC, and 0.4Mo-3Au-NiO/GDC were synthesized via deposition (co-) precipitation. There is discussion on the structural properties of the electrocatalysts on the oxidized and reduced state, as well as their use as electrolyte-supported (half) cells. A key remark after H₂-reduction is the formation of binary or ternary solid solutions. Ni/GDC was the most active for the catalytic CO₂ reforming of CH₄ and the CH₄ decomposition reactions and as a result the most prone to carbon deposition. On the other hand, the modified 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC exhibited the following properties: (i) the highest E_a,app for the non-desired RWGS reaction, (ii) high tolerance to carbon formation due to lower activity for the CH₄ decomposition, and (iii) were also less active for H₂ and CO production. Finally, 0.4Mo-3Au-Ni/GDC seems to perform the DRM reaction through a different mechanism when compared to Ni/GDC. Overall, the above three samples are proposed as potential fuel electrodes for further electrocatalytic measurements for the SOFC internal DRM process.

1. Introduction

Recycling CO₂ and CH₄ to produce syngas (H₂ + CO) through Dry Reforming of Methane (DRM) has received considerable attention and prominent research interest. CO₂ and CH₄ are greenhouse gases and the main compounds of biogas (55–65% CH₄ and 45–55% CO₂), which is widely produced by anaerobic fermentation of biomass [1]. The DRM process offers a feasible solution to abate CO₂ and CH₄, responsible for global warming and major climate changes, via production of synthesis gas (or syngas) with a molar H₂/CO ratio close to unity [2,3]. Furthermore, the syngas from DRM is suitable for further usage and synthesis of long-chain hydrocarbons in the Fischer-Tropsch industry [4,5].

Among the existing technologies for DRM energy applications, high temperature (750–1000 °C) Solid Oxide Fuel Cells (SOFCs) are the most efficient devices for electricity production [6,7,8,9]. A SOFC fed by biogas can operate either by using an external reformer, which converts the biogas to syngas, or without using an external reformer, inside the fuel cell, by utilizing advanced fuel electrodes [10,11]. The first approach represents the State of the Art (SoA) of the current technology, where the external reformer converts biogas to syngas and then this stream is fed to the fuel side of SOFC [6,11,12]. The second approach which is also known as “internal dry reforming of methane (IDRM)” uses modified improved electrodes to convert biogas to syngas through a catalytic DRM reaction and then to electricity through charge transfer reactions inside the SOFC [4,10,13,14,15]. The main advantages of this concept include simplified system design without external reformer and reduced air-flow/energy-losses because the heat released by the exothermic charge transfer reactions can also be used for the endothermic catalytic reactions (DRM, etc.) [10,14,15,16,17].

In general, when SOFCs operate at temperatures of 750–900 °C under internal dry reforming of methane conditions, several (electro)catalytic reactions may occur simultaneously on the fuel electrode (Equations (1)–(6)) [1,10,11,16,18,19]. In this respect, in the fuel side CO₂ and CH₄ are converted into H₂ and CO via the DRM reaction (Equation (1)). Due to the strongly endothermic character of DRM, high reaction temperatures (>750 °C) are required to achieve high H₂ and CO yields [3,4].

$${\mathrm{C}{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2\leftrightarrow 2\mathrm{C}\mathrm{O}+2{\mathrm{H}}_2\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{1073}^{\mathrm{o}}=260 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(1)

During operation, the produced H₂ from the DRM may be consumed through the Reverse Water Gas Shift (RWGS) reaction (Equation (2)), resulting in decreased H₂/CO ratio values (<1) [20].

$${\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\leftrightarrow \mathrm{C}\mathrm{O}+{\mathrm{H}}_2\mathrm{O}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{1073}^{\mathrm{o}}=41 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(2)

Moreover, carbon deposition on the surface of the electrocatalyst may occur through cracking of CH₄ (Equation (3)), and this is a very important issue since it is responsible for the progressive electrocatalyst deactivation [20,21]. The CH₄ decomposition reaction (Equation (3)) is thermodynamically favored at the high temperature region of >600 °C, whereas at lower temperatures (<650 °C) carbon is mainly produced through the CO disproportionation (Boudouard) reaction (2CO → C + CO₂) [20,21,22].

$$\mathrm{C}{\mathrm{H}}_4\leftrightarrow \mathrm{C}+2{\mathrm{H}}_2\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{\mathsf{\Delta }\mathrm{H}}_{1073}^{\mathrm{o}}=85 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(3)

During fuel cell operating conditions, the produced H₂ and CO, as well as the supplied CH₄ can be electrochemically oxidized in the triple phase boundary (TPB) region by oxygen ions (O²⁻) to produce power, according to Equations (4)–(6) [23,24,25,26,27].

$${\mathrm{H}}_2+{\mathrm{O}}^{2-}\leftrightarrow {\mathrm{H}}_2\mathrm{O}+2{\mathrm{e}}^{-}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{\mathsf{\Delta }\mathrm{H}}_{1073}^{\mathrm{o}}=-152 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(4)

$$\mathrm{C}\mathrm{O}+{\mathrm{O}}^{2-}\leftrightarrow \mathrm{C}{\mathrm{O}}_2+2{\mathrm{e}}^{-}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{\mathsf{\Delta }\mathrm{H}}_{1073}^{\mathrm{o}}=-186 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(5)

$$\mathrm{C}{\mathrm{H}}_4+{\mathrm{O}}^{2-}\leftrightarrow \mathrm{C}\mathrm{O}+2{\mathrm{H}}_2+2{\mathrm{e}}^{-}\hspace{1em}\hspace{1em} {\mathsf{\Delta }\mathrm{H}}_{1023}^{\mathrm{o}}=-22 \mathrm{k}\mathrm{J} {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$$

(6)

Although significant progress has been made on the development of advanced SOFC systems, the main issue that hinders the commercialization of this technology, particularly under CH₄ reforming processes, is the performance degradation of the cells and especially of the fuel electrodes. The performance of cells/fuel electrodes is affected by the thermal stresses resulting from the endothermic (Equations (1)–(3)) and the exothermic (Equations (4)–(6)) (electro)catalytic reactions, as well as the chemical degradation due to coke formation [19].

The State-Of-the-Art (SoA) electrocatalysts in SOFCs are Ni-ceramic-metal composites comprising Yttria Stabilized Zirconia (YSZ) and Gadolinia Doped Ceria (GDC), due to their excellent activity for electrochemical reactions and low cost [28,29]. In this respect, Ni/GDC is widely used as a fuel electrode, since it exhibits higher catalytic performance in CH₄ reforming reactions and resistance to coke formation when compared to Ni/YSZ electrodes [19,30,31]. The improved performance of Ni/GDC is attributed to the capacity of ceria to store and release oxygen. This property favours the oxidation and thus the removal of surface carbon species and the oxidation of CH₄ instead of its decomposition [19,32]. An alternative solution to suppress the carbon deposition and sintering tendency of nickel can be achieved by dispersing small amounts of transition noble (Ru, Rh, Pd, Pt, Au) or non-noble (Fe, Co, Cu, Mo, W) metals [1,16,32].

In the same direction, our research group has studied the effect of Au and/or Mo or Fe addition on commercial NiO/GDC powder in solid oxide applications [33,34,35,36,37]. The modifications have so far resulted in electrocatalysts with enhanced electrochemical performance and tolerance against nickel oxidation and carbon or sulphur poisoning under steam (or steam + CO₂) electrolysis [35,36,37] and under internal steam reforming of methane (ISRM) operating conditions in the presence/absence of 10 ppm H₂S [33,34]. In these studies, the modified cermets exhibited variations in their physicochemical and electrochemical properties, which were correlated with the formation of bimetallic Fe-Ni, Au-Ni, and ternary Mo–Au–Ni solid solutions during the H₂-reduction process. According to the ISRM studies, 0.4Mo-3Au-Ni/GDC was found to be the most carbon tolerant electrode when compared to Ni/GDC, 0.4Mo-Ni/GDC, and 3Au-Ni/GDC under both helium diluted or non-diluted harsh H₂O/CH₄ reaction feed conditions [34]. The improved performance of 0.4Mo-3Au-Ni/GDC was ascribed to the inhibited complete dehydrogenation of methyl species (CH_x) towards carbon on the active sites of the electrode [32]. Although Au utilization is not cost-effective compared to non-precious metals, its usage may be compensated if long-term operation and carbon tolerance is achieved.

In regards to the substitution of noble metals, iron is indicated as a promising transition metal dopant of Ni, due to its improved redox properties, its abundancy, and its lower price. The first research efforts of our group focused on the modification of Ni/GDC with various wt.% Fe with quite promising results for the solid oxide steam electrolysis reaction [37]. Notably, 0.5Fe-Ni/GDC displayed the highest performance compared to Ni/GDC and 2Fe-Ni/GDC, whereas the main problem was the poor stability that resulted in the decrease in the electrocatalytic activity. The CO₂ reforming of CH₄ has also been studied over Fe-Ni-based catalysts, focusing on the synergistic interaction of Fe-Ni towards the DRM activity and carbon formation propensity. According to some of the above research studies [38,39,40,41,42], under DRM reaction conditions FeO_x species are formed due to partial segregation of metallic Fe from the Fe-Ni alloyed particles, favouring the interaction of FeO_x with the accumulated carbon towards its gasification. A recent study over Fe-Ni/MgO catalysts [43] showed that Fe modification can inhibit carbon deposition under DRM reaction conditions, but at the same time may alter the nature of the deposited carbon towards a form that can be easily removed from the catalyst surface via gasification by CO₂.

In the present work, the performance of the following electrocatalysts (Ni/GDC, 1 wt.% Au-Ni/GDC, 3 wt.% Au-Ni/GDC, 0.5 wt.% Fe-Ni/GDC, 2 wt.% Fe-Ni/GDC, 0.5 wt.% Fe-3 wt.% Au-Ni/GDC, and 0.4 wt.% Mo-3 wt.% Au-Ni/GDC) was evaluated at Open Circuit Potential (OCP) conditions under biogas fuel operation. The samples were tested in the form of Electrolyte Supported (half) Cells (ESCs) at 750–900 °C by applying a reaction fuel mixture of CH₄/CO₂ = 1 with inlet gas flows that varied in the range between 150 to 300 cm³/min. The latter approach provided a reference profile for the catalytic performance of the candidate electrocatalysts by applying the same CH₄/CO₂ fuel feed conditions as those under IDRM operational mode but without the effect of applied current/potential. In addition, selected samples were further studied through specific kinetic measurements under differential conditions at various CH₄ and CO₂ partial pressures in an attempt to clarify the effect of modification on the intrinsic catalytic activity of Ni/GDC for the DRM reaction. The research objective of this investigation focuses on the sustainable conversion of the greenhouse CH₄ and CO₂ gases towards syngas by means of Solid Oxide Fuel cells, which is a highly efficient and environmentally friendly electrochemical source of energy/power and useful chemicals.

2. Experimental

2.1. Preparation of Powders

The modified powders were prepared via the Deposition—Precipitation (D.P.) and Deposition—Co Precipitation (D.CP.) methods by using the commercial NiO/GDC cermet (65 wt.% NiO-35 wt.% GDC, Marion Technologies) as the support. The precursors for the 1 wt.% Au-Ni/GDC, 3 wt.% Au-Ni/GDC, 0.5 wt.% Fe-Ni/GDC, 2 wt.% Fe-Ni/GDC, 0.5 wt.% Fe-3 wt.% Au-Ni/GDC, and 0.4 wt.% Mo-3 wt.% Au-Ni/GDC samples were HAuCl₄ (99.99% trace metals basis, 30 wt.% in dilute HCl), Fe(NO₃)₃x9H₂O (ACS reagent ≥ 98%), and (NH₄)₆Mo₇O₂₄ (99.98% metals basis) purchased from Sigma-Aldrich (St. Louis, MO, USA). Full details about the synthesis of the electrocatalysts can be found elsewhere [33,37]. During the process, the temperature was fixed at 70 °C and pH was adjusted by using NH₃ (1 M), at 7.0 for Au-NiO/GDC, 8.0 for Fe-NiO/GDC, and 6.0 for Fe-Au-NiO/GDC and Mo-Au-NiO/GDC samples. In the case of Au-modified electrocatalysts, after filtering the precipitate was mildly washed in order to eliminate any residual Cl⁻. Finally, in all cases, the precipitate was dried at 110 °C for 24 h and then each powder was calcined in air at 600 °C/90 min and a part of it at 1100 °C/75 min. The first batch was used for the preparation of a paste for electrode production and the last one for physicochemical characterization. Calcination at 1100 °C is considered necessary in order to examine the powders at similar calcination conditions, such as those where the cells are prepared. In the following sections, the examined samples will be reported as 1Au-NiO/GDC, 3Au-NiO/GDC, 0.5Fe-NiO/GDC, 2Fe-NiO/GDC, 0.5Fe-3Au-NiO/GDC, and 0.4Mo-3Au-NiO/GDC.

2.2. Preparation of Half Cells

The half cells were supported on a circular shaped planar 8YSZ electrolyte with 25 mm diameter and 300 μm thickness, purchased from Kerafol. As reported in previous studies [34,36], the deposition of the electrode was made by using the screen-printing method and a paste which consisted of a proper amount of calcined powder at 600 °C, terpineol as the dispersant, and PVB (polyvinylbutyral) as the binder, purchased from Sigma-Aldrich. After the paste-deposition, the cell was sintered at 1150 °C with a heating/cooling ramp rate of 2 °C/min. The loading of the examined fuel electrodes was ~6 mg cm⁻² and the active surface area was 1.8 cm².

2.3. Physicochemical Characterization

The powders, in their oxidized and H₂-reduced form, were characterized with XRF (X-ray Fluorescence), BET (Brunauer-Emmett-Teller), XRD (X-ray Diffraction), and TGA (Thermogravimetric Analysis) in the presence of CH₄. Specifically, the elemental wt.% concentration in each sample was determined by means of the non-destructive X-ray fluorescence (XRF) analysis using a portable Bruker Tracer III SD set with a beam diameter of 3 mm. The BET Specific Surface Area values (SSAs) were measured with a Micromeritics TriStar 3000 apparatus, employing nitrogen physisorption at the temperature of liquid nitrogen (77 K). In each measurement, the sample was pre-heated and outgased under dynamic vacuum at 250 °C for 2 h.

XRD measurements were performed by using a Bruker D8 Advance instrument equipped with a nickel-filtered Cu Kα (0.15418 nm) radiation source. The step size and the time per step were respectively fixed at 0.02° and 5 s, in the range of 20° $\le$ 2θ $\le$ 60°. The patterns were analyzed via the DiffracPlus-EVA software. The peaks were identified by using the standard reference crystallographic patterns of NiO: 044-1159, Ni: 004-0850, GDC: 046-0508, Au: 065-2870, and Fe₂O₃: 039-1346. The primary crystallite size of the main detected nanocrystals was estimated by means of Scherrer’s formula:

$$\mathrm{d}=\frac{0.9\mathsf{\lambda }}{\mathrm{B}\mathrm{c}\mathrm{o}\mathrm{s}\mathsf{\theta }}$$

(7)

where $\mathsf{\lambda }$ is the X-ray wavelength corresponding to Cu Kα radiation (0.15418 nm), $\mathsf{\theta }$ is the diffraction angle, and $\mathrm{B}$ is the line broadening (in radians) at half maximum. For the estimation of the $\mathrm{d}$ values, the diffraction peaks that were used are located at 2$\mathsf{\theta }$ equal to 43.4° for NiO (0 1 2), 44.4° for Ni (1 1 1), 28.6° for GDC (2 2 2), 38.1° for Au (1 1 1), and 35.6° for Fe₂O₃ (3 1 1).

TGA measurements were performed in a TA Q50 instrument isothermally at 800 °C under 10 vol.% CH₄/Ar with a total flow rate of 100 cm³ min⁻¹. The weight of each sample was ~25 mg and before the measurement the powder was reduced in-situ with 80 vol.% H₂/Ar at 800 °C for 100 min. After the H₂-reduction period, the reaction mixture was switched and CH₄/Ar was added in the feed.

2.4. Catalytic/Kinetic Measurements

The electrolyte-supported half cells were attached on a YSZ tube-reactor and were sealed airtight by using a glass sealing material. The half cells were investigated catalytically at OCP conditions under biogas fuel operation. The experiments were performed in the presence of a Ni mesh at 750–900 °C under a fuel mixture with CH₄/CO₂ = 1 without dilution in a carrier gas. Specifically, four experimental cycles were performed per temperature. Each cycle consisted of a total inlet gas flow ranging between 150–300 cm³/min with a step of 50 cm³/min, while the operating temperature was fixed. By changing the cycle, the experimental procedure was the same, but the temperature varied from 750 to 900 °C with a step of 50 °C. The kinetic measurements were carried out under differential conditions at various CH₄ and CO₂ partial pressures. Reactants and products were detected by using an on-line gas chromatograph (Varian CP-3800) with a thermal conductivity detector. A Porapak Q column (80–100 mesh, 1.8 m × 1/8 in. × 2 mm) was used for the analysis of H₂O at 150 °C while a Carbosieve S-11 column (80–100 mesh, 2 m × 1/8 in. × 2 mm) was used for the analysis of H₂, CO, CH₄, and CO₂ (in parallel with the Porapak Q).

The catalytic rates of reactants and products were calculated through Equation (8) and the carbon formation rates were determined by using the measured production rates of H₂, H₂O, and CO in the mass balance equation of carbon (Equation (9)).

$${\mathrm{r}}_{\mathrm{i}}\left[\frac{\mathrm{m}\mathrm{o}\mathrm{l}}{\mathrm{s}}\right]=\frac{\mathrm{F}\left[\frac{\mathrm{c}{\mathrm{m}}^3}{\mathrm{m}\mathrm{i}\mathrm{n}}\right]\ast \left({\mathrm{C}}_{\mathrm{i},\mathrm{i}\mathrm{n}}-{\mathrm{C}}_{\mathrm{i},\mathrm{o}\mathrm{u}\mathrm{t}}\right)}{{\mathrm{V}}_{\mathrm{m}}\ast 60\left[\frac{\mathrm{s}}{\mathrm{m}\mathrm{i}\mathrm{n}} \right]}$$

(8)

where ${\mathrm{r}}_{\mathrm{i}}$ is the consumption/production rate for H₂O, H₂, CO, CH₄ and CO₂, $\mathrm{F}$ is the total volumetric flow, ${\mathrm{V}}_{\mathrm{m}}$ is the molecular volume of ideal gases (24,451 $\frac{\mathrm{c}{\mathrm{m}}^3}{\mathrm{m}\mathrm{o}\mathrm{l}}$ at 25 °C and 1 atm), ${\mathrm{C}}_{\mathrm{i},\mathrm{i}\mathrm{n}}$ and ${\mathrm{C}}_{\mathrm{i},\mathrm{o}\mathrm{u}\mathrm{t}}$ $\left[\frac{\mathrm{c}{\mathrm{m}}^3}{\mathrm{m}\mathrm{i}\mathrm{n}}\right]$ are the reactor inlet/outlet concentrations of each compound, respectively.

$${\mathrm{r}}_{\mathrm{C}}\left[\frac{\mathrm{m}\mathrm{o}\mathrm{l}}{\mathrm{s}}\right]=\frac{{\mathrm{r}}_{{\mathrm{H}}_2}+2{\mathrm{r}}_{{\mathrm{H}}_2\mathrm{O}}-{\mathrm{r}}_{\mathrm{C}\mathrm{O}}}{2}$$

(9)

The corresponding % conversions of the reactants (CH₄, CO₂) were calculated according to Equations (10) and (11).

$${\mathrm{\%}\mathrm{C}\mathrm{H}}_4 \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}=\frac{{\mathrm{C}\mathrm{H}}_{4,\mathrm{i}\mathrm{n}}-{\mathrm{C}\mathrm{H}}_{4,\mathrm{o}\mathrm{u}\mathrm{t}}}{{\mathrm{C}\mathrm{H}}_{4,\mathrm{i}\mathrm{n}}}$$

(10)

$$\mathrm{\%}{\mathrm{C}\mathrm{O}}_2 \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}=\frac{{\mathrm{C}\mathrm{O}}_{2,\mathrm{i}\mathrm{n}}-{\mathrm{C}\mathrm{O}}_{2,\mathrm{o}\mathrm{u}\mathrm{t}}}{{\mathrm{C}\mathrm{O}}_{2,\mathrm{i}\mathrm{n}}}$$

(11)

3. Results and Discussion

Physicochemical Characterization

The wt.% amount of each dopant (Mo, Fe, or Au) in the total mass of the oxidized NiO/GDC powder was investigated by means of XRF analysis and the results confirmed that the calculated wt.% loadings are close to the nominal values which are presented in

Table 1. XRF analysis of the Mo, Au, and Fe wt.% concentration on the examined oxidized powders after calcination at 1100 °C.

Sample	wt.% Concentration
	Mo	Au	Fe
1Au-NiO/GDC	−	0.8	−
3Au-NiO/GDC	−	2.7	−
0.5Fe-NiO/GDC	−	−	0.5
2Fe-NiO/GDC	−	−	2.1
0.5Fe-3Au-NiO/GDC	−	2.5	0.6
0.4Mo-3Au-NiO/GDC	0.7	2.6	−

.

The Specific Surface Area (SSA) values of modified cermets, in their oxidized and reduced form, are presented in

Table 2. Specific Surface Area (SSA) values determined with the BET method of the calcined powders in their oxidized and reduced form. Error/accuracy = ±0.2 m² g⁻¹.

Sample	SSA (m2 g−1)
	T = 1100 °C, (Oxidized)	After H2-Reduction at T = 900 °C/2 h
NiO/GDC	4.2	2.5
1Au-NiO/GDC	4.3	2.2
3Au-NiO/GDC	2.7	2.1
0.5Fe-NiO/GDC	3.4	3.1
2Fe-NiO/GDC	4.0	3.4
0.5Fe-3Au-NiO/GDC	3.9	2.1
0.4Mo-3Au-NiO/GDC	2.8	2.3

.

Concerning the effect of the dopants, and by considering the error/accuracy limit of BET measurements on these materials (±0.2 m² g⁻¹), the addition of 3 wt.% Au, with or without 0.4 wt.% Mo, caused a ~33% decrease in the SSA of the oxidized NiO/GDC, whereas in the H₂-reduced form there was no difference in the SSA of Ni/GDC. Interestingly and opposite to the case of Mo/Au, the addition of Fe caused an increase in the SSA value. Specifically, this was ~24% and ~36% for 0.5Fe-Ni/GDC and 2Fe-Ni/GDC, respectively. The beneficial effect of Fe addition in inhibiting the decrease in SSA_Ni/GDC upon H₂-reduction was also reported in previous studies [36,37]. Another remark refers to the SSA of the oxidized ternary 0.5Fe-3Au-NiO/GDC sample, which exhibited higher value than that of 0.5Fe-NiO/GDC and 3Au-NiO/GDC. This can be realized via a synergistic interaction between the co-deposited Fe and Au dopants with NiO/GDC, which seems to retain the SSA of the specific sample. However, upon H₂-reduction the SSA decreased to a similar value like that in the majority of the examined samples.

The effect of modifiers on the bulk phase of NiO/GDC powder was further studied by means of XRD analysis. Figure 1 and Figure S1A in the Supplementary Materials present the XRD patterns of cermets calcined at 1100 °C. Typical diffraction peaks of NiO (JCPDS 044-1159) and GDC (Gd_0.6Ce_0.4)₂O_3.2 (JCPDS 046-0508) phases were detected in all samples. It is worth mentioning that the diffraction peaks of NiO and GDC appeared at the same 2θ position for all samples and no displacement was observed, as it can be seen in the magnification of the main peaks for NiO (0 1 2) and GDC (2 2 2) in the regions (42.8° ≤ 2θ ≤ 43.6°) and (28° ≤ 2θ ≤ 28.8°), respectively (Figure S1B,C in the Supplementary Materials).

Iron was detected in the form of maghemite-C, syn-Fe₂O₃ (JCPDS 039-1346), only for the 2Fe-NiO/GDC sample with three peaks at 2θ = 30.2° for the plane (2 2 0), 35.6° for (3 1 1), and 57.3° for (5 1 1). The absence of Fe₂O₃ diffraction peaks in the 0.5Fe-NiO/GDC and 0.5Fe-3Au-NiO/GDC samples can be attributed to the low wt.% content of iron which is below the detection limit of the XRD technique.

Moreover, gold was detected in the form of metallic Au (JCPDS 065-2870) with two peaks at 2θ = 38.1° for the plane (1 1 1) and 44.3° for (2 0 0). Concerning molybdenum, the XRD pattern of 0.4Mo-3Au-NiO/GDC highlighted its possible absence as MoO_x bulk particles. Complementary results from previous XRD and “quasi” in-situ (H₂ or H₂O) XPS analysis on Au or/and Mo and Mo-Au-NiO/GDC samples [33,35,44] confirm the above XRD analysis, but also highlight the presence of molybdate species on the surface of the oxidized 0.4Mo-3Au-NiO/GDC sample, which interact with Ni (or Ni-Au) upon H₂-reduction towards the formation of a Mo-Au-Ni solid solution.

The corresponding XRD patterns of the H₂-reduced powders at 900 °C for 2 h are presented in Figure 2A and Figure S1D in the Supplementary Materials. Two magnified areas of the main diffraction peaks of Ni (syn-Ni, JCPDS 004-0850) with the plane (1 1 1) and GDC (JCPDS 046-0508) with the plane (2 2 2) are also presented in Figure 2B and Figure 2C, respectively. It is observed that after H₂-reduction at 900 °C, both iron and gold were not detected in the XRD patterns of the samples, suggesting their possible absence in the bulk phase of the samples, as well as a re-arrangement in the structure of the reduced materials.

More specifically, the absence of bulk iron particles in 2Fe-Ni/GDC is confirmed by a shift of the Ni diffraction peaks to lower 2θ values, indicating the formation of a Ni-Fe solid solution during H₂-reduction, which is further confirmed by previous studies in the literature [37,45,46,47,48] (Figure 2B). This shift in the present study can be realized through the detection of a shoulder at 44.3° of the main diffraction peak Ni (1 1 1), which at this point is ascribed to Ni-2Fe solid solution and is currently under further evaluation. The latter effect is also detectable in the XRD pattern of 0.5Fe-Ni/GDC, where Ni (1 1 1) was slightly shifted.

Similarly, an absence of bulk gold particles was observed in 1Au-Ni/GDC, 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC, which was further verified by a shift of the Ni diffraction peaks to lower 2θ values, indicating the formation of Ni-Au or Ni-Fe-Au and Ni-Mo-Au solid solutions during the H₂-reduction process. Particularly, Ni-Au and Ni-Mo-Au solid solutions in Ni/YSZ and Ni/GDC cermets have been thoroughly reported in previous studies of our research group [33,35].

Concerning the Ni-Fe-Au solid solution, this is the first time that it is studied by means of XRD measurements. Specifically, the incorporation of the larger atomic radius ions of Fe (1.56 Å) or/and Au (1.74 Å) or/and Mo (1.90 Å) into the Ni (1.49 Å) [49] lattice has been reported to cause an increase in the crystal lattice parameter value of the Ni-Fe, Ni-Au, Ni-Fe-Au, and Ni-Mo-Au phases [33,45,50,51]. The lattice parameters for the above solid solution phases were calculated by the (1 1 1) diffraction plane of Ni (Figure 2B) via the DiffracPlus-EVA software. In particular, the lattice parameter values were estimated as equal to 3.530 Å for Ni, 3.532 Å for 1Au-Ni, 3.539 Å for 3Au-Ni, 3.532 Å for 0.5Fe-Ni, and 3.540 Å for 2Fe-Ni, 0.5Fe-3Au-Ni, and 0.4Mo-3Au-Ni. On the other hand, it should be mentioned that no shift was detected for the GDC diffraction peaks, as it can be seen in the magnification of the main peak of GDC (2 2 2) (Figure 2C).

The mean primary size $\left(\mathrm{d}\right)$ of the detected nanocrystals in the XRD patterns was estimated from the main peaks of NiO (0 1 2), Ni (1 1 1), GDC (2 2 2), Fe₂O₃ (3 1 1), and Au (1 1 1), according to the Scherrer equation (Equation (7)). The calculated values for the oxidized and reduced samples are presented in

Table 3. Primary, mean, crystallite size (nm) of NiO, Ni, GDC, Fe₂O₃, and Au, estimated from XRD line broadening and by using the Scherrer equation (Equation (7)), for Mo-Au-Fe-NiO/GDC electrocatalysts in their oxidized and reduced form.

Primary, Mean, Crystallite Size, d, (nm)
	NiO/GDC	1Au	3Au	0.5Fe	2Fe	0.5Fe-3Au	0.4Mo-3Au
	Oxidized, calcined at T = 1100 °C
NiO	98	116	119	113	106	111	134
GDC	67	67	70	69	65	71	69
Au	–	56	75	–	–	83	72
Fe2O3	–	–	–	–	67	–	–
	After H2-reduction at T = 900 °C
Ni	132	116	91	100	79	100	73
GDC	61	63	64	58	59	60	61

. More specifically, the crystal size of GDC did not change after calcination or H₂-reduction or after Fe, Au, and Mo modification, with a mean diameter of ~65 nm. Moreover, the crystallite size of NiO showed minor deviation (~15% increase) upon modification with Fe or/and Au, whereas a ~37% increase was observed upon modification with 0.4 wt.% Mo and 3 wt.% Au in the oxidized samples. Finally, by increasing the loading of Au from 1 to 3 wt.%, the Au crystal size increased by 34% and this value did not change upon addition of 0.5 wt.% Fe or 0.4 wt.% Mo. On the other hand, a decrease in the Ni crystallite size was detected on the H₂-reduced samples. Specifically, the primary size of Ni decreased by 12% and 31% in the cases of 1Au-Ni/GDC and 3Au-Ni/GDC, by 24% and 40% in 0.5Fe-Ni/GDC and 2Fe-Ni/GDC, and by 24% and 45% in 0.5Fe-3Au-Ni/GDC and 0.4Mo-3Au-Ni/GDC.

Overall, concerning the physicochemical properties of the oxidized powders, XRF analysis revealed that the wt.% concentration of each dopant was close to the nominal, whereas XRD verified the detection of iron in the form of syn-Fe₂O₃ and gold as metallic Au. After H₂-reduction, there was formation of Ni-Fe, Ni-Au, Ni-Fe-Au, and Ni-Mo-Au solid solutions.

4. Catalytic—Kinetic Measurements

4.1. Carbon Deposition from the Catalytic CH₄ Dissociation Reaction—TGA Measurements

The carbon tolerance of the modified samples was studied isothermally at 800 °C by means of thermogravimetric (TG) analysis under 10 vol.% CH₄/Ar flow. The TG profiles (Figure 3) depict the change in weight (Δwt.%), as a function of time due to carbon deposition.

Specifically, the presented TG measurements investigated the activity of the examined samples for the catalytic CH₄ dissociation reaction and the concomitant carbon formation/deposition. In this reaction scheme, CH₄ is adsorbed and dissociates on the Ni surface sites (*) towards adsorbed H* and CH_x species, according to Equations (12)–(16) [4,52].

CH₄ + 2* ↔ CH₃* + H*

(12)

CH₄ + * ↔ CH₄*

(13)

CH₃* + * ↔ CH₂* + H*

(14)

CH₂* + * ↔ CH* + H*

(15)

CH* + * ↔ C* + H*

(16)

The resulting adsorbed carbon (C*) species cover the surface of Ni, with subsequent diffusion and coverage of the bulk phase. The basic remark in this scheme is that the strongest CH₄ interaction with the Ni sites results in a higher carbon formation/deposition rate.

The TG profiles (Figure 3) show that the modified samples exhibited lower activity for the catalytic CH₄ decomposition, and as a result sufficient tolerance to carbon formation/deposition when compared to Ni/GDC. Concerning the iron-modified samples, 2Fe-Ni/GDC and to a lesser extent 0.5Fe-Ni/GDC exhibited immediate and rapid weight increase due to carbon deposition, whereas after the first minute the wt.% increase was less acute. After 10 min under 10 vol.% CH₄/Ar exposure, the Fe-modified samples exhibited a slightly higher tolerance against coking than Ni/GDC. On the other hand, modification of Ni/GDC with 3 wt.% Au resulted in the highest carbon resistance when compared to the other samples. In regards to the ternary samples 0.5Fe-3Au-Ni/GDC and 0.4Mo-3Au-Ni/GDC, they exhibited adequate tolerance against carbon formation. Thus, the presence of 3 wt.% Au, with or without Fe/Mo, makes Ni/GDC less prone to carbon deposition.

The presented TGA results on Au and Mo-Au-Ni/GDC samples confirm previous findings from our research group [33]. Moreover, studies in the literature verify that Au can modify the catalytic activity of a Ni-based catalyst by inhibiting the dissociative adsorption of CH₄ or at least the dehydrogenation reaction steps that lead to carbon formation (Equations (12)–(16)) [52,53]. In particular, Fan et al. [53] employed DFT calculations to study the synergistic interaction between Ni and transition metal atoms (M = Cu, Ru, Rh, Pd, Ag, Pt, and Au) on M/Ni (1 1 1) surface alloys for CH₄ dissociation. In accordance with our results, it was found that the binding strength of CH_x was lower over Au-Ni (1 1 1) compared to Ni₃ sites. Additionally, the adsorption energy of C on Ni-Au sites was found to be less negative than that on the Ni (1 1 1) surface, indicating the inhibition of coke formation over Au/Ni catalysts.

4.2. Catalytic—Kinetic Investigation for the Dry CH₄ Reforming Reaction

Catalytic experiments at OCP mode on electrolyte-supported half cells were performed in order to have a reference on the performance and carbon tolerance of each electrocatalyst under CO₂ reforming of CH₄ reaction conditions, without the effect of the applied current (and consequently of the O²⁻ flux). Regarding the homogenous catalytic reaction, no activity was observed. In addition, since the specific measurements are going to be used as a reference catalytic profile for further full cell electrocatalytic measurements, it was decided to examine the effect of the applied current collector for the fuel electrode. In this respect, comparative measurements of Ni/GDC with and without the presence of Ni mesh (Figure S2 in the Supplementary Materials) showed that there is no direct catalytic contribution of the Ni mesh on the electrocatalysts’ activity.

Figure 4 and Figure 5 show the catalytic performance of each sample for the CO₂ reforming of CH₄ at 750–900 °C (CH₄/CO₂ = 50/50, F_tot,in = 150 cm³ min⁻¹). Specifically, Figure 4 depicts the consumption/production rates of CH₄, CO₂, H₂, CO, H₂O, and the formed carbon, whereas Figure 5 shows the % conversions of CH₄ and CO₂. The measurements were performed on electrolyte supported (half) cells that comprised only the fuel electrocatalyst and Ni mesh.

In respect to the catalytic activity of Ni/GDC, it was found to be the most active sample for the CO₂ reforming of CH₄, yielding the highest consumption/production rates (Figure 4) and % conversions (Figure 5). However, it also exhibited measurable carbon formation rates at high temperatures (≥850 °C). The modified electrocatalysts were less active in terms of H₂ and CO production, but at the same time they were less prone to carbon formation (Figure 4). In particular, the majority of them did not show measurable carbon rates within the examined temperature region.

More specifically, 1Au-Ni/GDC and 2Fe-Ni/GDC were less active than Ni/GDC, but more active compared to the other modified samples (Figure 4). However, carbon formation was measured for both of these electrocatalysts at temperatures higher than 875 °C (Figure 4), which is a drawback for their use as potential fuel electrodes for the IDRM process. Although the above samples exhibited adequate tolerance to carbon formation under the catalytic decomposition of 10 vol.% CH₄/Ar at 800 °C (Figure 3), this did not apply for the DRM catalytic measurements under 50 vol.% CH₄−50 vol.% CO₂ (Figure 4). Furthermore, the fact that 1Au-Ni/GDC was prone to carbon formation (Figure 4) suggests that the 1 wt.% content of Au is probably too low to inhibit the coke formation under DRM reaction conditions.

On the other hand, 3Au-Ni/GDC and 0.5Fe-Ni/GDC showed sufficient carbon tolerance, but slightly lower rates than 1Au-Ni/GDC and 2Fe-Ni/GDC (Figure 4). Thus, it seems that the modification of Ni/GDC with 3 wt.% Au or 0.5 wt.% Fe inhibits carbon formation under DRM conditions and this observation concurs with the results from the TG analysis (Figure 3). Finally, the ternary samples 0.5Fe-3Au-Ni/GDC and 0.4Mo-3Au-Ni/GDC exhibited the lowest catalytic activity (Figure 4), but at the same time zero carbon deposits, which is also in agreement with the TGA results (Figure 3).

Regarding the % conversions of CO₂ and CH₄ at 750–900 °C (Figure 5), a general observation is that these values were lower than the thermodynamic equilibrium values at the corresponding temperatures under a mixture of CH₄/CO₂ = 1 at 1 atm (Table 4).

In addition, the % conversion of CO₂ at each temperature was always higher than the % conversion of CH₄. This is ascribed to the contribution of the RWGS reaction, which consumes CO₂ and H₂, resulting in a H₂/CO ratio less than unity [55,56,57,58,59]. Finally, the % conversions (Figure 5) and the consumption/production rates (Figure 4) increased by increasing the reaction temperature, due to the endothermic character of the DRM reaction [56,60].

Table 4. Equilibrium % conversions of CH₄ and CO₂ for stoichiometric carbon dioxide reforming of methane (at 1 atm) as a function of temperature [60,61].

Temperature (°C)	% CH4	% CO2
750	89.5	93.5
800	95.0	97.3
850	96.7	98.3
875	97.5	98.8
900	98.3	99.2

Table 4. Equilibrium % conversions of CH₄ and CO₂ for stoichiometric carbon dioxide reforming of methane (at 1 atm) as a function of temperature [60,61].

Temperature (°C)	% CH4	% CO2
750	89.5	93.5
800	95.0	97.3
850	96.7	98.3
875	97.5	98.8
900	98.3	99.2

The “used” electrocatalysts, after the applied DRM conditions, were examined by means of scanning electron microscopy (SEM) with a HR − SEM (Zeiss SUPRA 35VP) (Figure S4 in the Supplementary Materials). Further analysis of the SEM images with the Gwyddion 2.49 software enabled the calculation of the particle size (Table S2 in the Supplementary Materials). As a general remark, the observed particles consist of both Ni and GDC and the mean diameter of their size was calculated in the range of 200 nm (std. error ± 10 nm), indicating similar agglomeration after the DRM measurement. Moreover, 0.5Fe-3Au-Ni/GDC and 0.4Mo-3Au-Ni/GDC exhibited larger “void space” when compared to Ni/GDC which is an indication of larger macro-porosity/tortuosity [35].

According to the literature [4,32,33,53,61,62,63], Au-Ni surface alloys exhibit less activity for methane reforming, but stronger resistance to carbon formation, even with small amounts of Au doping. Indicatively, DFT studies from Besenbacher et al. [62] on Au-Ni catalysts led to the suggestion that Au-Ni is less active in methane reforming, but more stable because the effect of Au on the atomic carbon adsorption was found to be stronger than on CH₄ activation. The low catalytic activity of Au-Ni-based catalysts is in accordance with the d-band center theory that was proposed by Nørskov et al. [62,63] dealing with the dissociation of CH₄ on the Ni surface doped with Au. According to these studies, the rate of CH₄ dissociation and consequently the amount of carbon formation (Equations (12)–(16)) can be inhibited on Au-modified surfaces, due to the fact that the center of the d-band density of states is displaced to lower energies below the Fermi level of Ni. The same trend can be considered for the case where Au, Mo, and Ni coexist through the formation of the ternary solid solution [33]. The latter remark has been confirmed in previous studies [32,33,34] for the catalytic and electrocatalytic internal H₂O reforming of CH₄ (also in the presence of 10 ppm H₂S) where the 0.4Mo-3Au-Ni/GDC electrode was the least prone to carbon deposition (and sulfur poisoning) and the most active electrocatalytically, compared to both the non-doped and the Au-doped electrodes.

Other studies proposed that carbon formation in reforming processes depends on the size of metal particles [64,65]. Thus, smaller Ni particles have stronger ability to inhibit coke deposition, suggesting high stability of the catalytic activity [66]. In the present study, the particle size of Ni was smaller in all modified samples compared to Ni/GDC. Another remark especially for the case of the CO₂ reforming of CH₄ is that the addition of a small amount of a noble metal (i.e., Au in our case) favors CO₂ dissociation, resulting in oxygen formation that can assist the coke removal from the catalyst surface [4].

Furthermore, in regards to the Fe-Ni surface alloys/solid solutions, these have been proposed to partially promote carbon gasification [39,40,41,42,43,67]. Specifically, S.M. Kim et al. reported [67] that Fe in Ni-Fe alloys is partially oxidized by CO₂ to FeO (CO₂ + Fe → CO + FeO), which in turn provides lattice oxygen to the deposited carbon on the surface, leading to its partial oxidation to CO (C + FeO → CO + Fe), and therefore results in materials with improved coke tolerance. In the present study, only the 0.5Fe-Ni/GDC sample was found to be tolerant towards coke formation, whereas 2Fe-Ni/GDC was prone to carbon deposition. Finally, the 0.5Fe-3Au-Ni/GDC is a new proposed electrocatalyst that combines the improved catalytic characteristics of Au and Fe towards elimination of carbon deposits, although it exhibited lower catalytic performance compared to Ni/GDC.

Therefore, in an attempt to further investigate the observed differences on the catalytic activity, kinetic measurements were performed by applying fuel reaction inlet flows in the range of 150–300 cm³/min. The objective of the applied flows was to keep the CH₄ and CO₂ conversions in the region between 5–20%, thus assuring operation under differential conditions. It should be noted that in the above fuel inlet range the reaction rates were found to remain practically constant, which indicates the absence of mass transfer limitations.

Specifically, Ni/GDC was the most catalytically active sample, yielding high % CH₄ and CO₂ conversions that were beyond the differential conditions region. In order to achieve this experimental parameter, it was necessary to decrease the mass loading of the sample from 5 mg/cm² to 2.5 mg/cm² and to perform a new series of kinetic measurements (Figure S3 in the Supplementary Materials). Similarly, in the case of the modified samples, the F_total was fixed in the range of 150–300 cm³/min with a mass loading of ~6 mg/cm² (Table S1 in the Supplementary Materials). In addition, for reliable comparison reasons, a Ni/GDC electrode with a similar loading (5 mg/cm²) as the modified samples was studied kinetically in the range of 250–300 cm³/min (Table S1 in the Supplementary Materials).

Figure 6 presents the Arrhenius plots for the production rates of CO, H₂, and H₂O and the consumption rates of CH₄ and CO₂ and

Table 5. Apparent Activation Energies (E_a,app, kJ mol⁻¹) of modified electrodes under differential conditions for H₂O, H₂, and CO production and CH₄, CO₂ consumption derived from Figure 6 and calculated from the Arrhenius equation.

Sample	Ea,app * (kJ mol−1) per Product or Reactant
	H2O	H2	CO	CH4	CO2
Ni/GDC_1*	139	167	125	64	85
Ni/GDC_2*	142	147	126	63	73
1Au	108	146	112	67	78
3Au	179	184	157	95	113
0.5Fe	104	143	107	71	76
2Fe	102	108	93	72	79
0.5Fe-3Au	204	193	167	138	146
0.4Mo-3Au	191	178	154	88	105

*: Arrhenius eq. r = A exp($-\frac{{\mathrm{E}}_{\mathrm{a},\mathrm{a}\mathrm{p}\mathrm{p}}}{\mathrm{R}·\mathrm{T}}$). Ni/GDC with a mass of 1*: 2.5 mg cm⁻² and 2*: 5 mg cm⁻².

presents the calculated apparent activation energies (E_a,app) for each electrocatalyst under differential conditions. The E_{a, app} for the production of H₂O (Figure 6C) corresponds to the apparent activation energy for the RWGS reaction, which consumes the valuable H₂ for the fuel cell and is considered as a non-desired side reaction for the SOFC IDRM process. Modification of Ni/GDC with iron seems to favor the RWGS reaction, since E_a,app values of 0.5Fe-Ni/GDC and 2Fe-Ni/GDC were quite low for H₂O production. The same observation also applies for the 1Au-Ni/GDC electrode. Moreover, the above samples exhibited quite similar E_a,app values like Ni/GDC for CH₄ and CO₂ consumption. Concerning H₂ and CO production, 0.5Fe-Ni/GDC, 1Au-Ni/GDC, and mainly 2Fe-Ni/GDC exhibited lower E_a,app values than that of Ni/GDC.

On the other hand, modification of Ni/GDC with 3 wt.% Au seems to hinder the activity for the RWGS reaction, since the E_a,app values of 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC were higher than those of the other examined samples. In addition, the specific electrocatalysts were less prone to carbon formation and less active for H₂ and CO production with relatively high E_a,app values when compared to Ni/GDC and the Fe-modified electrodes. Finally, the fact that 2Fe-Ni/GDC and Ni/GDC were the samples where carbon formation was detected is a point that needs to be also taken into account.

According to the literature, CH₄ decomposition (Equation (12)) is the rate-determining step at high temperatures for both DRM and decomposition reactions [65,68,69]. Experimental kinetic studies [68,70] coincide with theoretical studies [65,69] for the reaction mechanism, which suggests that CH₄ decomposition demands higher energy than further decomposition of CH_x (x ≤ 3) (Equations (14)–(16)). Thus, the available research data focus on activation energy values for CH₄ activation/consumption. Moreover, it is generally accepted that CH₄ is decomposed on metal surface sites (e.g., Ni), whereas CO₂ is mainly activated on support sites (e.g., GDC in the case of the examined electrocatalysts) on the vicinity of dispersed metal particles or/and on the metallic sites [4,52]. As reported in the literature, the calculated activation energies for Ni-based catalysts in the DRM process vary within a wide range between 29 and 360 kJ/mol, which depends on the nature of the support, the presence of additives, and the catalytic conditions [58,70,71]. The most frequently reported E_a,app value for Ni-based catalysts is ~60 kJ/mol, which concurs with our calculated E_a,app for Ni/GDC and coincides with the E_a,app for CH₄ dissociation on Ni (1 1 0) and Ni (1 1 1) [72]. It should be noted that very low E_a,app values may reflect mass transfer limitation phenomena [72,73].

The modification of Ni/GDC with 3 wt.% Au and 0.4 wt.% Mo-3 wt.% Au, as well as with 0.5 wt.% Fe-3 wt.% Au, increased the E_a,app for CH₄ dissociation, suggesting that the catalytic activity of the surface Ni active sites has been inhibited for this reaction. Furthermore, the fact that Ni-Au modification increased the E_a,app for CH₄ reforming reactions has been also observed in other studies [62,74]. Overall concerning the selection of the proper electrocatalyst for the IDRM reaction, it is suggested that this should exhibit (a) the lowest E_a,app for H₂ and CO production, (b) the highest E_a,app for H₂O production, and (c) negligible carbon formation rates.

The kinetic investigation dealt also with the dependency of the H₂, CO, and H₂O production rates on various CH₄ and CO₂ molar fractions (y_CH4, y_CO2). These measurements also took place under differential conditions, by varying the applied gas flows between 150 and 300 cm³/min. The examined electrocatalysts (in the form of half cells) were Ni/GDC and 0.4Mo-3Au-Ni/GDC with mass loadings of 2.5 and 7.5 mg/cm², respectively. The effect of y_CH4 and y_CO2 on the production rates for Ni/GDC is shown in Figure 7 and Figure 8, whereas for 0.4Mo-3Au-Ni/GDC it is shown in Figure 9 and Figure 10.

Concerning Ni/GDC, it is observed that the production rates of H₂ and CO exhibited a positive order dependence either on y_CH4 (Figure 7A,B) or on y_CO2 (Figure 8A,B) at 850 and 900 °C. On the other hand, the production rate of H₂O showed an initially positive and then zero order dependence on y_CH4 (Figure 7C), whereas it was positive on y_CO2 (Figure 8C) at the same temperatures (850 and 900 °C). At 750 °C the general remark is the inhibition of the catalytic activity towards the production of H₂, CO, and H₂O. Moreover, at each temperature and fuel feed, the CO production rates were higher than that of H₂, due to the fact that CO is produced both through the DRM (Equation (1)) and the RWGS (Equation (2)) reactions.

Interestingly, at the high temperature of 900 °C the production rates of H₂, CO, and H₂O approached a maximum at y_CH4 ≥ 0.4 by keeping constant y_CO2 = 0.5 (Figure 7). On the contrary, at the same temperature, r_H2, r_CO, and r_H2O increased within the whole y_CO2 range, while y_CH4 = 0.5 (Figure 8). The positive reaction order of the production rates versus y_CH4 and y_CO2 and the appearance of a maximum at y_CH4 ≥ 0.4 at 900 °C indicate that most probably the dissociative adsorption of CH₄ can be the limiting step of the DRM reaction on Ni/GDC, which is in agreement with other studies [65,68,69,70]. Finally, it should be noted that under low y_CO2 (<0.25) (Figure 8C) the production of H₂O, through the RWGS reaction, was negligible and as a result the produced CO is derived solely from the DRM reaction. In this case and in the region of y_CO2 < 0.25, the produced r_H2 should be identical with the produced r_CO, which was confirmed (Figure 8A,B).

In regards to the 0.4Mo-3Au-Ni/GDC sample at T < 800 °C (i.e., 750 °C), the overall catalytic activity was inhibited, similar to Ni/GDC. At ≥ 800 °C, a positive order dependence was also observed for the production rates of H₂ and CO by varying y_CH4 (Figure 9A,B). However, by varying the y_CO2 and at temperatures > 800 °C (Figure 10A,B), the trend of r_H2 and r_CO was different when compared to Ni/GDC. Specifically, in the non-modified sample the production rate of H₂ and CO increased within the whole range of the applied y_CO2, while y_CH4 = 0.5 (Figure 8A,B). On the other hand, in the case of the Mo-Au-modified sample, at T ≥ 800 °C the r_H2 and r_CO reached to a maximum for y_CO2 ≥ 0.25 (Figure 10A,B). This maximum indicates that the DRM reaction mechanism for 0.4Mo-3Au-Ni/GDC might be different when compared to Ni/GDC. In particular, it can be suggested that on the active sites of the Mo-Au-Ni-modified electrocatalyst there is a point where the dissociative adsorption of CO₂ reached faster to equilibrium and to a stable surface coverage. As a result, the production of CO and H₂ did not increase any further by increasing the y_CO2 when compared to Ni/GDC.

Concerning the production rate of H₂O on the 0.4Mo-3Au-Ni/GDC, it exhibited the same trend like on the non-modified sample. Specifically, r_H2O initially showed a positive order dependence on y_CH4 (Figure 9C), which altered to zero at y_CH4 ≥ 0.25. Moreover, the dependence of r_H2O on the applied y_CO2 (Figure 10C) was positive.

Therefore, the observation that in the presented DRM reactions scheme H₂O is produced only through the RWGS reaction, in combination with the similar dependence of r_H2O on y_CH4 and y_CO2, suggest that both of the Ni/GDC and 0.4Mo-3Au-Ni/GDC electrocatalysts perform the RWGS reaction through the same mechanism. On the other hand, by considering that CO is produced both through DRM and RWGS, it can be assumed that the mechanism of the main DRM reaction may be different for 0.4Mo-3Au-Ni/GDC than that of Ni/GDC, which is a remark that needs to be further studied. Furthermore, the calculated E_a,app from the Arrhenius plots and the formation rates of carbon highlight that the modified 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC samples exhibited (i) the highest E_a,app for the non-desired RWGS reaction, (ii) high tolerance to carbon formation, and (iii) they were less active for H₂ and CO production, as well as for the CH₄ decomposition. Therefore, and for the purposes of a following separate manuscript, there will be selection among these three samples for electro-catalytic measurements and comparison with the SoA Ni/GDC in the form of full cells for the internal DRM process.

5. Conclusions

The present study dealt with the physicochemical and kinetic investigation of Fe, Au, Fe-Au, and Mo-Au modified Ni/GDC electrocatalysts towards their performance for the DRM, RWGS, and CH₄ decomposition reactions, as well as their tolerance to carbon formation. The research objective of this investigation focuses on the sustainable conversion of the greenhouse CH₄ and CO₂ gases towards syngas by means of Solid Oxide Fuel cells, which is a highly efficient and environmentally friendly electrochemical source of energy/power and useful chemicals. In this respect, the catalytic-kinetic investigation, in combination with the detailed physicochemical characterization, of a variety of new candidate fuel electrodes is the first key step to understand the underlying catalytic profile and to elucidate the occurring reactions during the IDRM process.

Regarding the physicochemical properties of the oxidized powders, XRF analysis revealed that the wt.% concentration of each dopant was close to the nominal, whereas XRD verified the detection of iron in the form of syn-Fe₂O₃ and gold as metallic Au. After H₂-reduction, there was formation of Ni-Fe, Ni-Au, Ni-Fe-Au, and Ni-Mo-Au solid solutions. TG analysis at 800 °C, under 10 vol.% CH₄/Ar, showed that the modified samples exhibited lower activity for the catalytic CH₄ decomposition and were less prone to carbon deposition when compared to Ni/GDC. Specifically, modification with 3 wt.% Au, with or without Fe/Mo, resulted in the highest carbon tolerance.

In respect to the catalytic performance of electrolyte-supported half cells, Ni/GDC was found to be the most active sample for the CO₂ reforming of CH₄. However, it exhibited carbon formation rates at high temperatures (≥850 °C). On the other hand, 3Au-Ni/GDC, 0.5Fe-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC were less active catalytically, but at the same time exhibited higher tolerance to carbon deposition. Concerning the selection of the proper electrocatalyst for the IDRM reaction, it is suggested that this should preferably exhibit (a) the lowest E_a,app for H₂ and CO production, (b) the highest E_a,app for H₂O production which results mainly from the RWGS and is considered as an undesired side reaction for the SOFC IDRM process, and (c) negligible carbon formation. These kinetic pre-conditions seem to be followed by the modified 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC electrocatalysts.

Further kinetic analysis on Ni/GDC and 0.4Mo-3Au-Ni/GDC showed similar dependence of r_H2O on y_CH4 and y_CO2, suggesting that the RWGS side reaction may be performed through the same mechanism on both samples. On the other hand, it was observed that the mechanism of the main DRM reaction may be different for the 0.4Mo-3Au-Ni/GDC when compared to Ni/GDC. This is because the Mo-Au-modified sample (i) was less active for the catalytic CH₄ decomposition reaction and thus less prone to carbon formation, and (ii) at high temperature (≥800 °C) and constant y_CH4 the dissociative adsorption of CO₂ seems to reach faster to a stable surface coverage when compared to Ni/GDC.

Overall, the present study is a thorough investigation on the development of modified Ni/GDC electrocatalysts with enhanced efficiency and carbon tolerance when compared to the SoA. The selected samples, namely 3Au-Ni/GDC, 0.5Fe-3Au-Ni/GDC, and 0.4Mo-3Au-Ni/GDC, meet the experimentally defined kinetic prerequisites for the SOFC IDRM reaction. Therefore, the reported findings are considered as a significant contribution for the practical application of this technology with a valuable impact in the research field of sustainable energy. These samples are currently under further investigation as electrodes for the purposes of a separate manuscript in full cell SOFC measurements and long-term stability operation.