**The E**ff**ect of CeO<sup>2</sup> Preparation Method on the Carbon Pathways in the Dry Reforming of Methane on Ni**/**CeO<sup>2</sup> Studied by Transient Techniques**

**Constantinos M. Damaskinos <sup>1</sup> , Michalis A. Vasiliades <sup>1</sup> , Vassilis N. Stathopoulos <sup>2</sup> and Angelos M. Efstathiou 1,\***


Received: 30 May 2019; Accepted: 18 July 2019; Published: 21 July 2019

**Abstract:** The present work discusses the effect of CeO<sup>2</sup> synthesis method (thermal decomposition (TD), precipitation (PT), hydrothermal (HT), and sol-gel (SG)) on the carbon pathways of dry reforming of methane with carbon dioxide (DRM) applied at 750 ◦C over 5 wt% Ni/CeO2. In particular, specific transient and isotopic experiments (use of <sup>13</sup>CO, <sup>13</sup>CO2, and <sup>18</sup>O2) were designed and conducted in an attempt at providing insights about the effect of support's preparation method on the concentration (mg gcat −1 ), reactivity towards oxygen, and transient evolution rates (µmol gcat −1 s −1 ) of the *inactive* carbon formed under (i) CH4/He (methane decomposition), (ii) CO/He (reverse Boudouard reaction), and (iii) the copresence of the two (CH4/CO/He, use of <sup>13</sup>CO). Moreover, important information regarding the relative contribution of CH<sup>4</sup> and CO<sup>2</sup> activation routes towards carbon formation under DRM reaction conditions was derived by using isotopically labelled <sup>13</sup>CO<sup>2</sup> in the feed gas stream. Of interest was also the amount, and the transient rate, of carbon removal via the participation of support's labile active oxygen species.

**Keywords:** DRM; nickel; cerium dioxide; transient experiments; lattice oxygen; isotopes

### **1. Introduction**

Nowadays, a great academic and research interest is seen aimed at exploring the gradual replacement of conventional fossil fuels towards energy production through the utilization of alternative and renewable energy sources such as Natural Gas (NG) and Bio-Gas (BG). The driving force behind it are the findings of NG reservoirs rich in CO<sup>2</sup> (>40 vol%) [1–3] and renewable bio-gas [4,5], which can be used in the development of technologies, such as the dry reforming of methane (DRM: CH<sup>4</sup> + CO<sup>2</sup> → 2CO + 2H2, ∆H<sup>0</sup> = +261 kJ mol−<sup>1</sup> ), as more environmentally friendly processes in many aspects [6]. The latter is enforced as it uses two major greenhouse gases (CH<sup>4</sup> and CO2), while at the same time produces a favorable H2/CO gas ratio (~1) for the Fischer–Tropsch [7,8] synthesis towards liquid fuels, but also for other processes in the production of chemicals (DME, MeOH, ammonia) [9,10]. In addition, the low operational cost of DRM in comparison with the already used steam methane reforming (SMR) and partial oxidation of methane (POM) technologies, makes its use very attractive [11–14]. However, the main obstacle for the development of an industrial DRM technology is the catalyst's deactivation due to carbon accumulation, especially over Ni-supported [15] solids, which are mainly used due to their low cost and wide availability. The formation of inactive carbon, in the form of filaments, graphite, and whiskers, mainly is derived from the CH<sup>4</sup> decomposition (CH<sup>4</sup> <sup>→</sup> C-s <sup>+</sup> 2H2, <sup>∆</sup>H<sup>0</sup> = +75 kJ mol−<sup>1</sup> ) and Boudouard reaction (2CO <sup>→</sup> C-s <sup>+</sup> CO2, <sup>∆</sup>H<sup>0</sup> <sup>=</sup> <sup>−</sup>172 kJ mol−<sup>1</sup> ). Thus, the design of a suitable Ni-based catalyst

supported on reducible metal oxides emerged (e.g., use of CeO2, Zr4+-, Pr3+-, Ti4+-doped CeO2, La2O3, Nb2O5) [16–25] since the latter supports possess oxygen storage capacity (OSC), oxygen vacancies, and high oxygen mobility, leading to carbon gasification rates that significantly reduce carbon accumulation rates, but also provide high thermal stability for the supported Ni catalysts [26,27]. The Ce-based materials owe their advantages against non-reducible metal oxides to the undergoing of fast change in the Ce4<sup>+</sup> <sup>↔</sup> Ce3<sup>+</sup> oxidation state (redox behavior), leading to an oxygen release, and vice versa to an oxygen storage, in the ceria-based stable crystal structure [28,29].

Several studies reveal the effect of preparation method of CeO<sup>2</sup> nanoparticles for use in a wide variety of applications, and they argue that such solids form different surface defects by exhibiting more surface atoms than their bulk counterparts [30–36]. Such nanoparticles could have various morphological and structural differences (nanorods, nanowires, nano-cubes, etc.) with different surface area, pore volume, and mean pore diameter, thus the synthesis method seems to play an important role [37–41].

In spite of recent efforts to develop suitable CeO2-supported Ni catalysts exhibiting high DRM catalytic activity and carbon resistance [42], fundamental understanding of the effect of support synthesis method on the contribution of the carbon deposition and removal routes has not been reported yet, to the best of our knowledge. The synthesis of CeO<sup>2</sup> via different methods [43,44] could lead to several variations in its physicochemical properties, but also to the metal surface when ceria is used as support. The latter is well demonstrated as due to the existence of strong metal support interactions (SMSI) between Ni particles and CeO<sup>2</sup> [45,46].

Transient methods (step-gas switches, use of isotopes, and temperature programmed oxidation or hydrogenation) performed over supported metal catalysts provided important information about the carbon paths in the DRM reaction, and relationships between the catalytic activity and coke formation. Furthermore, rival reaction mechanisms and rate determined steps (RDS) under DRM reaction conditions (working catalyst surface) can be elucidated. For example, Schuurman and Mirodatos [47] suggested that on Ni/SiO<sup>2</sup> catalyst the RDS is the recombination of atomic C (derived from CH<sup>4</sup> dissociation) and atomic O (derived from CO<sup>2</sup> dissociation) over the Ni surface. On the other hand, Slagtern et al. [48] observed that on Ni/La2O<sup>3</sup> catalyst, CH<sup>4</sup> is activated on Ni as opposed to CO2, which is activated on La2O<sup>3</sup> support (or metal-support interface) towards carbonate-like species formation. Advanced kinetic and mechanistic studies to elucidate the carbon paths in DRM with the use of isotopes (C18O2, <sup>13</sup>CH4, <sup>13</sup>CO2) were recently performed to a large extent by our group [17–20,46], but also in some other works [47,49–51]. In these works, the significant participation of lattice oxygen of reducible metal oxide supports (e.g., doped ceria-based materials) towards removal of carbon to form CO was proved experimentally by <sup>18</sup>O transient isotopic experiments followed by DRM reaction. Also, the quantification of origin of carbon (CH<sup>4</sup> vs. CO<sup>2</sup> activation route) was probed as a function of reaction T and catalyst composition.

The present work aims to address the effect of CeO<sup>2</sup> support synthesis method on the carbon pathways in the dry reforming of methane over 5 wt% Ni/CeO<sup>2</sup> catalysts, where this is reported for the first time to our knowledge. Of particular interest was to investigate differences on (i) the concentration of inactive carbon and its reactivity towards oxygen, (ii) the relative contribution of CH<sup>4</sup> and CO<sup>2</sup> activation routes to the total carbon formation on the catalytic surface via methane decomposition and Boudouard reactions, and (iii) the participation of labile support's lattice oxygen towards carbon removal, and to what extent. For this purpose, various transient and isotopic experiments followed by temperature programmed oxidation (TPO) were performed.

#### **2. Results**

#### *2.1. Catalysts Surface Texture and Structural Properties*

The BET specific surface area (SSA, m<sup>2</sup> g −1 ), mean pore diameter (dp, nm), and the specific pore volume (Vp, cm<sup>3</sup> g −1 ) of the four CeO<sup>2</sup> solid supports prepared by different methods, namely: Thermal decomposition (TD), precipitation (PT), hydrothermal (HT), and sol-gel (SG), are given in Table S1 (Electronic Supplementary Information, ESI). The SSA was found to be in the 5.6–50 m<sup>2</sup> g −1 range, the d<sup>p</sup> in the 6.7–22.5 nm range, and the V<sup>p</sup> in the 0.029–0.203 cm<sup>3</sup> g −1 range. The CeO2-HT solid exhibited the largest value of SSA and V<sup>P</sup> (50 m<sup>2</sup> g −1 , 0.203 cm<sup>3</sup> g −1 ), and a pore size of d<sup>p</sup> = 15.8 nm, as opposed to CeO2-PT (5.6 m<sup>2</sup> g −1 , 0.032 cm<sup>3</sup> g −1 ) with the largest pore size (22.5 nm). The powder XRD diffractograms of 5 wt% Ni supported on the various CeO<sup>2</sup> solid supports are given in Figure S1. The CeO<sup>2</sup> support exhibits the cubic structure [46], and after using the Scherrer equation and the CeO<sup>2</sup> (111) diffraction line, the mean primary crystal size (dC, nm) of support was estimated. Similarly, after using the NiO (111) diffraction line, the particle size (dNiO, nm) of NiO was also estimated. The latter value was then used to estimate the particle size (dNi, nm) of Ni<sup>0</sup> via Equation (3), and the obtained results are reported in Table S1. There was not any shift of the NiO (111) 2θ diffraction peak (see Figure S1B) among the different samples, however, variations of the mean Ni particle size (8.4–20.8 nm) and the mean primary crystal size of ceria support (11.5–43.1 nm) were observed (Table S1). The latter results find good support by the literature as will be discussed in Section 3.1. *Catalysts* **2019**, *9*, x FOR PEER REVIEW 3 of 23 d<sup>p</sup> in the 6.7–22.5 nm range, and the V<sup>p</sup> in the 0.029–0.203 cm<sup>3</sup> g−1 range. The CeO2-HT solid exhibited the largest value of SSA and V<sup>P</sup> (50 m<sup>2</sup> g−1, 0.203 cm<sup>3</sup> g−1), and a pore size of d<sup>p</sup> = 15.8 nm, as opposed to CeO2-PT (5.6 m<sup>2</sup> g−1 , 0.032 cm<sup>3</sup> g−1) with the largest pore size (22.5 nm). The powder XRD diffractograms of 5 wt% Ni supported on the various CeO<sup>2</sup> solid supports are given in Figure S1. The CeO<sup>2</sup> support exhibits the cubic structure [46], and after using the Scherrer equation and the CeO<sup>2</sup> (111) diffraction line, the mean primary crystal size (dC, nm) of support was estimated. Similarly, after using the NiO (111) diffraction line, the particle size (dNiO, nm) of NiO was also estimated. The latter value was then used to estimate the particle size (dNi, nm) of Ni<sup>0</sup> via Equation (3), and the obtained results are reported in Table S1. There was not any shift of the NiO (111) 2θ diffraction peak (see Figure S1B) among the different samples, however, variations of the mean Ni particle size (8.4–20.8 nm) and the mean primary crystal size of ceria support (11.5–43.1 nm) were observed (Table S1). The latter results find good support by the literature as will be discussed in Section 3.1.

#### *2.2. H<sup>2</sup> Temperature-Programmed Desorption (H2-TPD) 2.2. H<sup>2</sup> Temperature-Programmed Desorption (H2-TPD)*

Figure 1 presents H2-TPD traces of the 5 wt% Ni supported on the various CeO<sup>2</sup> solids. It is clearly seen that the CeO<sup>2</sup> preparation method resulted in drastic changes of the H<sup>2</sup> desorption kinetic features in terms of strength of hydrogen binding states (TM, peak maximum temperature) and their corresponding surface coverage (area under a given desorption peak). Thus, the different preparation method of CeO<sup>2</sup> support followed by the same Ni deposition method (wet impregnation) led to differences in the heterogeneity of the Ni surface (e.g., distribution of strength of hydrogen chemisorption sites on surface Ni). In the low-temperature range of 50–200 ◦C, the amount of H<sup>2</sup> desorbed (µmol g−<sup>1</sup> ) was significantly larger for CeO2-HT (23.8 µmol g−<sup>1</sup> ) compared to the other three ceria-supported Ni catalysts (CeO2-PT, -TD, and -SG), where similar amounts were found (8.8, 8.0, and 6.2 µmol g−<sup>1</sup> , respectively). In the high-temperature range of 200–500 <sup>o</sup>C, the amount of H<sup>2</sup> desorbed follows a different order: Ni/CeO2-TD (25.3 µmol g−<sup>1</sup> ) > Ni/CeO2-PT (17.3 µmol g−<sup>1</sup> ) > Ni/CeO2-HT (17 µmol g−<sup>1</sup> ) > Ni/CeO2-SG (8.5 µmol g−<sup>1</sup> ). Of interest is the fact that all ceria-supported Ni catalysts present three main desorption peaks. However, shoulders to these main desorption peaks appear at different temperatures. For example, the Ni/CeO2-TD presents three main desorption peaks centered at 57, 222, and 354 ◦C with shoulder at the falling part of the 3rd peak (Figure 1a), Ni/CeO2-PT at 57, 303, and 393 ◦C with clear shoulders at the falling part of the 1st peak, the rising part of 2nd peak, and the falling part of 3rd peak (Figure 1b). Ni/CeO2-HT presents the three main desorption peaks centered at 95, 258, and 362 ◦C with shoulders at the falling part of 3nd peak (Figure 1c), whereas Ni/CeO2-SG at 79, 157, and 383 ◦C with shoulders at the low-T side of 3nd peak (Figure 1d). Figure 1 presents H2-TPD traces of the 5 wt% Ni supported on the various CeO<sup>2</sup> solids. It is clearly seen that the CeO<sup>2</sup> preparation method resulted in drastic changes of the H<sup>2</sup> desorption kinetic features in terms of strength of hydrogen binding states (TM, peak maximum temperature) and their corresponding surface coverage (area under a given desorption peak). Thus, the different preparation method of CeO<sup>2</sup> support followed by the same Ni deposition method (wet impregnation) led to differences in the heterogeneity of the Ni surface (e.g., distribution of strength of hydrogen chemisorption sites on surface Ni). In the low-temperature range of 50–200 °C, the amount of H<sup>2</sup> desorbed (μmol g−1) was significantly larger for CeO2-HT (23.8 μmol g−1) compared to the other three ceria-supported Ni catalysts (CeO2-PT, -TD, and -SG), where similar amounts were found (8.8, 8.0, and 6.2 μmol g−1, respectively). In the high-temperature range of 200–500 oC, the amount of H<sup>2</sup> desorbed follows a different order: Ni/CeO2-TD (25.3 μmol g−1) > Ni/CeO2-PT (17.3 μmol g−1) > Ni/CeO2-HT (17 μmol g−1) > Ni/CeO2-SG (8.5 μmol g−1). Of interest is the fact that all ceria-supported Ni catalysts present three main desorption peaks. However, shoulders to these main desorption peaks appear at different temperatures. For example, the Ni/CeO2-TD presents three main desorption peaks centered at 57, 222, and 354 °C with shoulder at the falling part of the 3rd peak (Figure 1a), Ni/CeO2-PT at 57, 303, and 393 °C with clear shoulders at the falling part of the 1st peak, the rising part of 2nd peak, and the falling part of 3rd peak (Figure 1b). Ni/CeO2-HT presents the three main desorption peaks centered at 95, 258, and 362 °C with shoulders at the falling part of 3nd peak (Figure 1c), whereas Ni/CeO2-SG at 79, 157, and 383 °C with shoulders at the low-T side of 3nd peak (Figure 1d).

**Figure 1.** H<sup>2</sup> temperature-programmed desorption (H2-TPD) traces obtained over 5 wt% Ni/CeO2-<sup>δ</sup> catalysts prepared by (**a**) Thermal Decomposition (TD), (**b**) Precipitation (PT), (**c**) Hydrothermal (HT), and (**d**) Sol Gel (SG) method; FHe = 50 NmL min−1; β = 30 °C min−1; W = 0.3 g. **Figure 1.** H<sup>2</sup> temperature-programmed desorption (H<sup>2</sup> -TPD) traces obtained over 5 wt% Ni/CeO2-<sup>δ</sup> catalysts prepared by (**a**) Thermal Decomposition (TD), (**b**) Precipitation (PT), (**c**) Hydrothermal (HT), and (**d**) Sol Gel (SG) method; FHe = 50 NmL min−<sup>1</sup> ; β = 30 ◦C min−<sup>1</sup> ; W = 0.3 g.

The Ni dispersion (DNi, %) of the given solids was estimated based on the total amount of H<sup>2</sup> desorbed (Figure 1), and results are presented in Table S1 (ESI). The lowest dispersion of Ni was

The Ni dispersion (DNi, %) of the given solids was estimated based on the total amount of H<sup>2</sup> desorbed (Figure 1), and results are presented in Table S1 (ESI). The lowest dispersion of Ni was found when CeO2-SG was used as support (3.4%), followed by the CeO2-PT (6.1%), CeO2-TD (7.8%), and the CeO2-HT (9.6%, highest dispersion). Thus, the Ni particle size (dNi, nm) estimated via Equation (4) was found to be: 10.1, 12.4, 15.9, and 28.5 nm for the CeO2-HT, -TD, -PT, and -SG, respectively. The latter results were supported by those obtained from the powder XRD analyses (Section 2.1, Figure S1 and Table S1) and the HR-TEM (Section 2.3), but also with those reported previously [20]. *2.3. Transmission Electron Microscopy (TEM) Studies*  HR-TEM images obtained over the fresh 5 wt% Ni/CeO2-HT (the support was prepared by the hydrothermal method, HT) calcined in air for 4 h at 750 °C are given in Figure S2 (ESI). It is seen that dispersed Ni nanoparticles of ∼8–12 nm in size were observed, in good agreement with the H2-TPD and powder XRD results. *2.4. Scanning Electron Microscopy (SEM) Studies* 

*Catalysts* **2019**, *9*, x FOR PEER REVIEW 4 of 24

and the CeO2-HT (9.6%, highest dispersion). Thus, the Ni particle size (dNi, nm) estimated via Equation (4) was found to be: 10.1, 12.4, 15.9, and 28.5 nm for the CeO2-HT, -TD, -PT, and -SG, respectively. The latter results were supported by those obtained from the powder XRD analyses

#### *2.3. Transmission Electron Microscopy (TEM) Studies* SEM images obtained over the fresh CeO2-supported Ni solids are presented in Figure S3 (ESI).

HR-TEM images obtained over the fresh 5 wt% Ni/CeO2-HT (the support was prepared by the hydrothermal method, HT) calcined in air for 4 h at 750 ◦C are given in Figure S2 (ESI). It is seen that dispersed Ni nanoparticles of ~8–12 nm in size were observed, in good agreement with the H2-TPD and powder XRD results. The secondary particle size (agglomerates) of the catalyst's support was in the range of 10–50 nm, where different porous structures were derived after using the four different CeO2 synthesis methods. *2.5. Catalytic Performance Studies in DRM* 

#### *2.4. Scanning Electron Microscopy (SEM) Studies* Figure 2A presents catalytic performance results in terms of specific integral rate (mol gcat−1 min−1)

SEM images obtained over the fresh CeO2-supported Ni solids are presented in Figure S3 (ESI). The secondary particle size (agglomerates) of the catalyst's support was in the range of 10–50 nm, where different porous structures were derived after using the four different CeO<sup>2</sup> synthesis methods. of CH4 conversion and H2/CO gas product ratio obtained after 30 min of DRM at 750 °C over the 5 wt% Ni supported on differently prepared CeO2 solids. The Ni/CeO2-HT presented the highest catalytic activity (5.5 mmol gcat−1 min−1), while Ni/CeO2-SG the lowest activity (2.0 mmol gcat−1 min−1), thus a significant difference by a factor of ∼2.9 existed between these two catalysts. On the other hand,

#### *2.5. Catalytic Performance Studies in DRM* the H2/CO gas product ratio did not follow the activity differences shown in Figure 2A (compare

Figure 2A presents catalytic performance results in terms of specific integral rate (mol gcat <sup>−</sup><sup>1</sup> min−<sup>1</sup> ) of CH<sup>4</sup> conversion and H2/CO gas product ratio obtained after 30 min of DRM at 750 ◦C over the 5 wt% Ni supported on differently prepared CeO<sup>2</sup> solids. The Ni/CeO2-HT presented the highest catalytic activity (5.5 mmol gcat <sup>−</sup><sup>1</sup> min−<sup>1</sup> ), while Ni/CeO2-SG the lowest activity (2.0 mmol gcat <sup>−</sup><sup>1</sup> min−<sup>1</sup> ), thus a significant difference by a factor of ~2.9 existed between these two catalysts. On the other hand, the H2/CO gas product ratio did not follow the activity differences shown in Figure 2A (compare Figure 2A,B) since the most active Ni/CeO2-HT exhibited a value of ~1.2 (similar to -TD and -PT) and the least active Ni/CeO2-SG presented a value of ~1.1. These results clearly demonstrated that the series of four catalysts presented different orders in terms of H<sup>2</sup> and CO reaction selectivity. It should be noted that all four catalytic systems presented XCH4 (%) and XCO2 (%) larger than 80% (81–94%), with H2-yields larger than 45% (48–59%) and H2/CO gas product ratio close to the desired value of ~1, tested at the same GHSV (30,000 h−<sup>1</sup> ) (see Table S2). Figure 2A,B) since the most active Ni/CeO2-HT exhibited a value of ~1.2 (similar to -TD and -PT) and the least active Ni/CeO2-SG presented a value of ~1.1. These results clearly demonstrated that the series of four catalysts presented different orders in terms of H2 and CO reaction selectivity. It should be noted that all four catalytic systems presented XCH4 (%) and XCO2 (%) larger than 80% (81–94%), with H2-yields larger than 45% (48–59%) and H2/CO gas product ratio close to the desired value of ~1, tested at the same GHSV (30,000 h−1) (see Table S2). Figure 2B presents the stability test (up to 50 h on TOS) for the 5 wt% Ni/CeO2-PT catalyst, which exhibited the least amount of accumulated carbon (mg gcat−1) after 12 h on TOS among the series of catalysts. It was clearly seen that after up to ∼12 h on TOS, the catalyst's activity remained practically constant, while a drop by ∼17.5% in the integral rate of methane conversion occurred after 50 h on TOS (see also Table S3). Similar results were also observed for the other three catalyst compositions (not reported). The comparative activity behavior based on 30-min on TOS shown in Figure 2A is thus very representative for the true effect of ceria support synthesis method.

**Figure 2.** (**A**) Specific integral rates of CH4 conversion (mmol gcat<sup>−</sup>1 min<sup>−</sup>1) and H2/CO gas product ratio obtained after 30 min of DRM at 750 °C (GHSV ∼30,000 h<sup>−</sup>1) over the four catalysts; (**B**) Stability test **Figure 2.** (**A**) Specific integral rates of CH<sup>4</sup> conversion (mmol gcat <sup>−</sup><sup>1</sup> min−<sup>1</sup> ) and H<sup>2</sup> /CO gas product ratio obtained after 30 min of DRM at 750 ◦C (GHSV ~30,000 h−<sup>1</sup> ) over the four catalysts; (**B**) Stability test in terms of integral rate of CH<sup>4</sup> conversion conducted over 50 h of TOS on 5 wt% Ni/CeO<sup>2</sup> -PT catalyst; GHSV ~30,000 h−<sup>1</sup> .

Figure 2B presents the stability test (up to 50 h on TOS) for the 5 wt% Ni/CeO2-PT catalyst, which exhibited the least amount of accumulated carbon (mg gcat −1 ) after 12 h on TOS among the series of catalysts. It was clearly seen that after up to ~12 h on TOS, the catalyst's activity remained practically GHSV∼30,000 h<sup>−</sup>1.

constant, while a drop by ~17.5% in the integral rate of methane conversion occurred after 50 h on TOS (see also Table S3). Similar results were also observed for the other three catalyst compositions (not reported). The comparative activity behavior based on 30-min on TOS shown in Figure 2A is thus very representative for the true effect of ceria support synthesis method. that obtained over other CeO2-supported Ni catalysts [17,50], and is mainly attributed to the effect of reverse water-gas shift (RWGS) side reaction. It will be shown in the following Section 2.6, that the four catalysts, for their activity performance depicted in Figure 2A, also exhibited significantly different amounts of carbon accumulation due to their different CeO2 support preparation method.

*Catalysts* **2019**, *9*, x FOR PEER REVIEW 5 of 24

in terms of integral rate of CH4 conversion conducted over 50 h of TOS on 5 wt% Ni/CeO2-PT catalyst;

that of CH4 for all catalytic systems except in the case of Ni/CeO2-SG. The latter result was similar to

It's worth mentioning that regarding the CO<sup>2</sup> conversion (%), this was found to be lower than that of CH<sup>4</sup> for all catalytic systems except in the case of Ni/CeO2-SG. The latter result was similar to that obtained over other CeO2-supported Ni catalysts [17,50], and is mainly attributed to the effect of reverse water-gas shift (RWGS) side reaction. It will be shown in the following Section 2.6, that the four catalysts, for their activity performance depicted in Figure 2A, also exhibited significantly different amounts of carbon accumulation due to their different CeO<sup>2</sup> support preparation method. *2.6. Characterization of Carbon Formed under Different Reaction Conditions*  2.6.1. Dry Reforming of Methane (12CO2/12CH4) at Steady-State Reaction Conditions Transient response curves of CO2 obtained during temperature-programmed oxidation (TPO) of carbon deposited over the four Ni/CeO2 catalysts after 12 h in DRM (20 vol% CH4/20 vol% CO2/60

#### *2.6. Characterization of Carbon Formed under Di*ff*erent Reaction Conditions* vol% He) at 750 °C are presented in Figure 3A. The Ni/CeO2-PT led to a lower carbon accumulation, ca. ~3.8 times (30.7 vs. 116.1 mg C gcat−1) compared to the Ni/CeO2-HT catalyst, with the other two

#### 2.6.1. Dry Reforming of Methane (12CO2/ <sup>12</sup>CH4) at Steady-State Reaction Conditions catalysts, Ni/CeO2-TD and Ni/CeO2-SG showing a decrease by 1.8 and 1.4 times (66.2 and 80.4 mg C gcat−1), respectively. In the case of Ni/CeO2-TD and Ni/CeO2-PT catalysts, a main peak starting at 450

Transient response curves of CO<sup>2</sup> obtained during temperature-programmed oxidation (TPO) of carbon deposited over the four Ni/CeO<sup>2</sup> catalysts after 12 h in DRM (20 vol% CH4/20 vol% CO2/60 vol% He) at 750 ◦C are presented in Figure 3A. The Ni/CeO2-PT led to a lower carbon accumulation, ca. ~3.8 times (30.7 vs. 116.1 mg C gcat −1 ) compared to the Ni/CeO2-HT catalyst, with the other two catalysts, Ni/CeO2-TD and Ni/CeO2-SG showing a decrease by 1.8 and 1.4 times (66.2 and 80.4 mg C gcat −1 ), respectively. In the case of Ni/CeO2-TD and Ni/CeO2-PT catalysts, a main peak starting at 450 ◦C and ending at 750 ◦C with peak maximum at ~630 ◦C was observed, whereas in the case of Ni/CeO2-HT, a wider main peak was observed, which was centered at ~670 ◦C. As opposed to the latter behavior, the Ni/CeO2-SG (Figure 3Ad) presented likely several types of carbon, since it started reacting with oxygen at ~500 ◦C with a shoulder at 600 ◦C and a main peak at 700 ◦C, but a clear sharp peak at 750 ◦C was also observed; the latter might have also been the result of a hot spot in the catalytic bed formed at these high temperatures given the large exotherm of carbon oxidation to CO2. °C and ending at 750 °C with peak maximum at ~630 °C was observed, whereas in the case of Ni/CeO2-HT, a wider main peak was observed, which was centered at ~ 670 °C. As opposed to the latter behavior, the Ni/CeO2-SG (Figure 3Ad) presented likely several types of carbon, since it started reacting with oxygen at ∼500 °C with a shoulder at 600 °C and a main peak at 700 °C, but a clear sharp peak at 750 °C was also observed; the latter might have also been the result of a hot spot in the catalytic bed formed at these high temperatures given the large exotherm of carbon oxidation to CO2. The Ni/CeO2-PT catalyst, which led to the lowest amount of carbon deposition, was also tested for longer time-on-stream (ca. 50 h, see Figure 2B, Table S3), and the TPO trace recorded is presented in Figure 3B. The amount of carbon deposition was increased when the TOS increased from 12 h to 50 h, ca. 147.1 vs. 30.7 mg C g−1cat (see also Table S3). These results will be discussed below in relation to a synergy observed for carbon accumulation between CH4 and CO presence in the same gas mixture compared to the CH4 decomposition and Boudouard reaction contribution alone.

**Figure 3.** Transient response curves of CO2 concentration obtained during TPO of carbon formed after (**A**) 12 h of DRM (20% CH4/20% CO2/He; 50 NmL min<sup>−</sup>1; GHSV ∼ 30,000 h<sup>−</sup>1) at 750 °C over 5 wt% Ni/CeO2 prepared by (a) Thermal decomposition (TD), (b) Precipitation (PT), (c) Hydrothermal (HT), and (d) Sol-Gel (SG) method; (**B**) TPO trace of carbon formed after 50 h in DRM over the 5 wt% Ni/CeO2-PT catalyst. **Figure 3.** Transient response curves of CO<sup>2</sup> concentration obtained during TPO of carbon formed after (**A**) 12 h of DRM (20% CH<sup>4</sup> /20% CO<sup>2</sup> /He; 50 NmL min−<sup>1</sup> ; GHSV ~30,000 h−<sup>1</sup> ) at 750 ◦C over 5 wt% Ni/CeO<sup>2</sup> prepared by (a) Thermal decomposition (TD), (b) Precipitation (PT), (c) Hydrothermal (HT), and (d) Sol-Gel (SG) method; (**B**) TPO trace of carbon formed after 50 h in DRM over the 5 wt% Ni/CeO<sup>2</sup> -PT catalyst.

2.6.2. Isotopically Labelled Dry Reforming of Methane (13CO2/12CH4) Figure 4 presents 13CO2 and 12CO2 transient response curves recorded during TPO of the carbon formed after 30 min in isotopically labelled DRM (5 vol% 13CO2/5 vol% 12CH4/45 vol% Ar/45 vol% He) at 750 °C over the Ni/CeO2-TD, Ni/CeO2-HT, and Ni/CeO2-SG catalysts. It's worth mentioning that The Ni/CeO2-PT catalyst, which led to the lowest amount of carbon deposition, was also tested for longer time-on-stream (ca. 50 h, see Figure 2B, Table S3), and the TPO trace recorded is presented in Figure 3B. The amount of carbon deposition was increased when the TOS increased from 12 h to 50 h, ca. 147.1 vs. 30.7 mg C g−<sup>1</sup> cat (see also Table S3). These results will be discussed below in relation to a synergy observed for carbon accumulation between CH<sup>4</sup> and CO presence in the same gas mixture compared to the CH<sup>4</sup> decomposition and Boudouard reaction contribution alone.

#### 2.6.2. Isotopically Labelled Dry Reforming of Methane (13CO2/ <sup>12</sup>CH4)

Figure 4 presents <sup>13</sup>CO<sup>2</sup> and <sup>12</sup>CO<sup>2</sup> transient response curves recorded during TPO of the carbon formed after 30 min in isotopically labelled DRM (5 vol% <sup>13</sup>CO2/5 vol% <sup>12</sup>CH4/45 vol% Ar/45 vol% He) at 750 ◦C over the Ni/CeO2-TD, Ni/CeO2-HT, and Ni/CeO2-SG catalysts. It's worth mentioning that the Ni/CeO2-PT, where the support was prepared by the precipitation method (CeO2-PT), exhibited non-measurable amounts of carbon, and neither <sup>12</sup>CO nor <sup>13</sup>CO signals were recorded in the MS. The TPO traces of <sup>13</sup>CO<sup>2</sup> and <sup>12</sup>CO<sup>2</sup> were different in shape among the three catalytic systems, and this was largely attributed to the different carbon oxidation kinetics influenced by the type of carbon deposited, and its reactivity towards oxygen. The <sup>13</sup>CO2-TPO trace originated from the <sup>13</sup>CO<sup>2</sup> activation route during DRM, while that of <sup>12</sup>CO2-TPO from the <sup>12</sup>CH<sup>4</sup> activation route. Furthermore, the three catalysts presented different amounts of carbon formed via the two activation routes but also a different total amount of carbon, which was estimated by integrating the respective TPO traces. The contribution of each reactant (CH<sup>4</sup> vs. CO2) to the carbon formation under DRM reaction conditions was estimated based on the ratio of <sup>12</sup>CO2/ <sup>13</sup>CO<sup>2</sup> (TPO traces). It was shown that in all three catalytic systems, CH<sup>4</sup> decomposition is the dominant route, but to a different extent. More precisely, the Ni/CeO2-TD (Figure 4A) and Ni/CeO2-HT (Figure 4B) presented <sup>12</sup>C/ <sup>13</sup>C = 1.6 and 1.8, respectively, as opposed to the Ni/CeO2-SG catalyst (Figure 4C), where CH<sup>4</sup> decomposition contributed in a significantly higher extent (12C/ <sup>13</sup>C = 4.7). In addition, the total amount of carbon was found to be larger in the case of Ni/CeO2-HT (29.5 µmol g−<sup>1</sup> ) compared to Ni/CeO2-SG and Ni/CeO2-TD (28.1 and 11.1 µmol g−<sup>1</sup> , respectively). The latter results agree with those presented in Section 2.6.1, where the feed gas stream (5 vs. 20 vol% of reactants) and the TOS (30 min vs. 12 h) were much different. *Catalysts* **2019**, *9*, x FOR PEER REVIEW 6 of 24 the Ni/CeO2-PT, where the support was prepared by the precipitation method (CeO2-PT), exhibited non-measurable amounts of carbon, and neither 12CO nor 13CO signals were recorded in the MS. The TPO traces of 13CO2 and 12CO2 were different in shape among the three catalytic systems, and this was largely attributed to the different carbon oxidation kinetics influenced by the type of carbon deposited, and its reactivity towards oxygen. The 13CO2-TPO trace originated from the 13CO2 activation route during DRM, while that of 12CO2-TPO from the 12CH4 activation route. Furthermore, the three catalysts presented different amounts of carbon formed via the two activation routes but also a different total amount of carbon, which was estimated by integrating the respective TPO traces. The contribution of each reactant (CH4 vs. CO2) to the carbon formation under DRM reaction conditions was estimated based on the ratio of 12CO2/13CO2 (TPO traces). It was shown that in all three catalytic systems, CH4 decomposition is the dominant route, but to a different extent. More precisely, the Ni/CeO2-TD (Figure 4A) and Ni/CeO2-HT (Figure 4B) presented 12C/13C = 1.6 and 1.8, respectively, as opposed to the Ni/CeO2-SG catalyst (Figure 4C), where CH4 decomposition contributed in a significantly higher extent (12C/13C = 4.7). In addition, the total amount of carbon was found to be larger in the case of Ni/CeO2-HT (29.5 μmol g−1) compared to Ni/CeO2-SG and Ni/CeO2-TD (28.1 and 11.1 μmol g−1, respectively). The latter results agree with those presented in Section 2.6.1, where the

feed gas stream (5 vs. 20 vol% of reactants) and the TOS (30 min vs. 12 h) were much different.

**Figure 4.** Temperature-programmed oxidation (TPO) of carbon to 12CO2 and 13CO2 formed after 30 min in 5 vol% 13CO2/5 vol% 12CH4/45 vol% Ar/45 vol% He (50 NmL min−1; GHSV ∼30,000 h<sup>−</sup>1) at 750 °C over (**A**) 5 wt% Ni/CeO2-TD, (**B**) 5 wt% Ni/CeO2-HT, and (**C**) 5 wt% Ni/CeO2-SG. **Figure 4.** Temperature-programmed oxidation (TPO) of carbon to <sup>12</sup>CO<sup>2</sup> and <sup>13</sup>CO<sup>2</sup> formed after 30 min in 5 vol% <sup>13</sup>CO<sup>2</sup> /5 vol% <sup>12</sup>CH<sup>4</sup> /45 vol% Ar/45 vol% He (50 NmL min−<sup>1</sup> ; GHSV ~30,000 h−<sup>1</sup> ) at 750 ◦C over (**A**) 5 wt% Ni/CeO<sup>2</sup> -TD, (**B**) 5 wt% Ni/CeO<sup>2</sup> -HT, and (**C**) 5 wt% Ni/CeO<sup>2</sup> -SG.

2.6.3. Transient Methane Decomposition (CH4/He) Reaction

decomposition, over each of the four catalytic surfaces presented in Figure 5A, led also to different H2 transient formation rates (Figure 5B), similar in shape with those of CH4 consumption (Figure 5A).

Figure 5 shows transient evolution rates of CH4 consumption and H2 and CO gas formation (the only gaseous reaction products observed), obtained during the step gas switch He → 20 vol% CH4/1% Ar/He (30 min) made at 750 °C over the four 5 wt% Ni supported on CeO2 carriers prepared
