H 2 Production from Catalytic Methane Decomposition Using Fe / x-ZrO 2 and Fe-Ni / (x-ZrO 2 ) (x = 0, La 2 O 3 , WO 3 ) Catalysts

: An environmentally-benign way of producing hydrogen is methane decomposition. This study focused on methane decomposition using Fe and Fe-Ni catalysts, which were dispersed over di ﬀ erent supports by the wet-impregnation method. We observed the e ﬀ ect of modifying ZrO 2 with La 2 O 3 and WO 3 in terms of H 2 yield and carbon deposits. The modiﬁcation led to a higher H 2 yield in all cases and WO 3 -modiﬁed support gave the highest yield of about 90% and was stable throughout the reaction period. The reaction conditions were at 1 atm, 800 ◦ C, and 4000 mL(hg cat ) − 1 space velocity. Adding Ni to Fe / x-ZrO 2 gave a higher H 2 yield and stability for ZrO 2 and La 2 O 3 + ZrO 2 -supported catalysts whose prior performances and stabilities were very poor. Catalyst samples were analyzed by characterization techniques like X-ray di ﬀ raction (XRD), nitrogen physisorption, temperature-programmed reduction (TPR), thermo-gravimetric analysis (TGA), and Raman spectroscopy. The phases of iron and the supports were identiﬁed using XRD while the BET revealed a signiﬁcant decrease in the speciﬁc surface areas of fresh catalysts relative to supports. A progressive change in Fe’s oxidation state from Fe 3 + to Fe 0 was observed from the H 2 -TPR results. The carbon deposits on Fe / ZrO 2 and Fe / La 2 O 3 + ZrO 2 are mainly amorphous, while Fe / WO 3 + ZrO 2 and Fe-Ni / x-ZrO 2 are characterized by graphitic carbon. supported Fe and Fe-Ni


Introduction
Hydrogen has been touted to be the energy of the future, as it burns cleanly without introducing any pollutant to the environment. It has a calorific value that is thrice that of traditional fuel (e.g., gasoline) [1]. Thus, it could as well substitute for fossil fuel in the near future. In the universe, hydrogen happens to be the lightest and most plentiful element. It accounts for 74% or thereabouts of the whole universe [2]. The green solar energy that we obtain from the sun originates from the fusion of hydrogen. By estimates, the provision of hydrogen by the sun will be enough to keep the fusion reaction going on for more than 5 × 10 9 years [3].
In spite of the abundant nature of hydrogen, separating it from its combined form is laborious. A number of technologies have been developed in the light of making hydrogen available in its elemental form. These ways of production have been compiled into three categories which are: water dissociation, fossil fuel conversion, and biological means. Among all, water dissociation is the most CH 4 → C + 2H 2 ∆H 298K = 74.9 kJ mol −1 (1) In reality, the methane decomposition process could produce a minute quantity of CO or CO 2 , as the methane could react with oxygen in the catalyst's carrier or in the active metal-oxide in the case of incomplete reduction during the activation stage [6]. A wide range of catalysts has been used for the MD process. In a finding conducted over a number of metal-supported catalysts, it was discovered that the MD activity increases in the order of Fe, Co, Ru, Rh, Ir, Pt, Pd, and Ni [7]. Simply put, Ni gave the best catalytic activity and therefore extensive MD studies have been done with Ni-based catalysts. MD reactions over Ni-supported catalysts often follow three basic steps, which are CH 4 activation, carbon nucleation, and lastly deposition of carbon, which ultimately leads to the growth of carbon materials [8]. The third step has been considered to be responsible for catalyst deactivation and the plugging of reactors. In addition, it shortens the time-on-stream (TOS) for MD reactions unlike steam reforming of CH 4 , which could go on for a longer period. The longest reported TOS for the MD process performed over Ni-supported catalysts has been 200 h [9]. Great efforts have been made by researchers in bringing the deactivated Ni-supported catalysts back to life by using steam or air or a combination of the two [10,11]. However, this process eventually produces CO or CO 2 , thereby defeating the objective of producing pure H 2 with no oxide of carbon. Many attempts have been made jointly and severally by researchers to improve the resistance of Ni-supported catalysts to coke deposits. A study was conducted using different masses of Ni/SiO 2 and it was found that 0.3 g was the optimum with the best relative stability [12]. In other studies, alloys were tested with a combination of Ni, Pd, Co, and/or Cu [13][14][15]. Furthermore, studies on the combination of Ni and Fe over alumina support have been reported. Bayat et al., investigated Ni-Fe catalysts at different proportions. Ni weight% was fixed while that of Fe was varied from 5-15%. They studied the effect of temperature and the addition of Fe as a promoter. It was observed that 50%Ni−10%Fe/Al 2 O 3 gave the highest conversion and began to lose its activity at a temperature above 700 • C [16]. Some other interesting results were obtained by researchers that used catalysts with a similar constituent, but with the addition of Cu and Co as promoters [17]. In all these, the results obtained leave room for improvement.
Hence, what is ideal for the MD process is to develop a less expensive active catalyst that will have better stability and saves the cost of regeneration. Fe has been touted to be a better candidate for MD reactions. It has an incompletely filled outer shell that is waiting to be paired with other atoms. This will foster the dissociation of CH 4 . Additionally, Fe has a higher melting point (1811 K) compared with Ni (1728 K), and owing to this, Fe can operate at a higher temperature than Ni without being sintered and this will be favorable to the MD reactions that are inherently endothermic. Tungsten (VI) oxide (WO 3 ) has been reported to have special acidity, which could improve its activity. In a study conducted by Rezgui et al. on hydrocracking of n-decane, it was revealed by the temperature-programmed desorption (TPD) analysis that the density of acid sites increased with WO 3 loading and resulted in a catalyst with the best performance. W 6+ has the ability to form Bronsted acid on reduction. For this, the special acidic properties of WO 3 will enhance its adsorption of CH 4 onto itself for conversion to H 2 by the active metal. Additionally, its chemical and thermal stability properties could be of utmost importance in reaction to the CH 4 decomposition. Furthermore, WO 3 has versatile applications and has been used in the cracking of heavy fractions in petroleum as well as dehydrogenation of alcohols [18][19][20]. Lanthanum (III) oxide (La 2 O 3 ) has been tested in MD, being used as support over Fe. It proved to be a good candidate by giving a high yield of H 2 of about 66%. However, the catalyst could not maintain its stability over the time of the study [21].
There has been no consensus on what the composition of the Fe-based catalyst should be and the optimum operating condition to be adopted for MD reactions. Thus, in this study, the authors synthesized Fe catalysts using different carriers (ZrO 2 , La 2 O 3 + ZrO 2 , and WO 3 + ZrO 2 ) to investigate what effect the carriers would have. Thereafter, we tried to create synergy with Fe and Ni combined to see the performance of their combination in terms of hydrogen yield, stability, and carbon deposits. In our previous publication, we studied the conversion of methane over these catalysts [22]. Here, we investigated their hydrogen yield and the effect of the Ni addition.

Characterization of Catalysts
2.1.1. X-ray Diffraction (XRD) Figure 1 shows the XRD diffractogram of the fresh Fe and Fe-Ni-supported calcined samples. The different phases available on the samples can be seen in the figure. All the Fe-supported samples showed virtually the same crystallographic phases at the same peak angle with the exception of Fe/WO 3 + ZrO 2 , which has some additional peaks. From the diffractogram, the crystallographic peaks for monoclinic ZrO 2 (m-ZrO 2 ) can be seen at 24 • , 28.

Textural Properties
The textural properties of the supports as well as those of the synthesized fresh catalyst samples were determined by the N2 adsorption-desorption. Table 1 contains the results of the different supports and the fresh catalyst samples, respectively. Furthermore, the isotherms are displayed in Figure 2A while Figure 2B contains the pore size distribution of the samples in question. Relating the surface areas of the supports to those of their respective synthesized catalyst samples, we observed a huge decrease in the surface areas. It can be said that the addition of metals to the supports culminated in blockage of the pores of the supports. Furthermore, the pore volume and pore diameter of fresh catalysts also decreased vis-à-vis those of their corresponding supports. The N2 adsorptiondesorption isotherms of all the catalysts in Figure 2A fall under type IV classification. This is typical of mesoporous materials. In addition, hysteresis loops can be observed in all these samples. According to IUPAC classification, the hysteresis loops are H3 type [24]. From Figure 2B, the pore sizes of the fresh catalyst samples fall within 2-50 nm. According to the IUPAC classification, this size range falls under the category of mesoporous materials.

Textural Properties
The textural properties of the supports as well as those of the synthesized fresh catalyst samples were determined by the N 2 adsorption-desorption. Table 1 contains the results of the different supports and the fresh catalyst samples, respectively. Furthermore, the isotherms are displayed in Figure 2A while Figure 2B contains the pore size distribution of the samples in question. Relating the surface areas of the supports to those of their respective synthesized catalyst samples, we observed a huge decrease in the surface areas. It can be said that the addition of metals to the supports culminated in blockage of the pores of the supports. Furthermore, the pore volume and pore diameter of fresh catalysts also decreased vis-à-vis those of their corresponding supports. The N 2 adsorption-desorption isotherms of all the catalysts in Figure 2A fall under type IV classification. This is typical of mesoporous materials. In addition, hysteresis loops can be observed in all these samples. According to IUPAC classification, the hysteresis loops are H3 type [24]. From Figure 2B, the pore sizes of the fresh catalyst samples fall within 2-50 nm. According to the IUPAC classification, this size range falls under the category of mesoporous materials.  The reducibility, together with the degree of interaction between the metal and the different carriers, was studied using TPR for the Fe and Fe-Ni-supported fresh catalyst samples. As we have it in Figure 3, for Fe/x-ZrO 2 (x = 0, La 2 O 3 + ZrO 2 , WO 3 + ZrO 2 ), three reduction peaks at different temperatures can be seen in the TPR profile. The peaks represent the stepwise reduction of Fe 3+ to Fe 0 , i.e., Fe 2 O 3 →Fe 3 O 4 →FeO→Fe. The peak appearing at the temperature range of 310-410 • C stands for the reduction of Fe 2 O 3 to Fe 3 O 4 . Additionally, the peak within 430-700 • C can be attributed to the reduction from Fe 3 O 4 to FeO. Lastly, the final stage of the reduction process to obtain zero-valence Fe can be assigned to the peak that appeared within 710-900 • C. Similar reduction behavior has been reported in different studies, such as the work done by Bayat et al. [15]. The reduction peak of Fe/WO 3 + ZrO 2 can be seen at a higher range in comparison with the other catalysts with only Fe. This could be as a result of the strong interaction that exists between Fe and its carrier (WO 3 + ZrO 2 ).
carriers, was studied using TPR for the Fe and Fe-Ni-supported fresh catalyst samples. As we have it in Figure 3, for Fe/x-ZrO2 (x = 0, La2O3 + ZrO2, WO3 + ZrO2), three reduction peaks at different temperatures can be seen in the TPR profile. The peaks represent the stepwise reduction of Fe 3+ to Fe 0 , i.e., Fe2O3→Fe3O4→FeO→Fe. The peak appearing at the temperature range of 310-410 °C stands for the reduction of Fe2O3 to Fe3O4. Additionally, the peak within 430-700 °C can be attributed to the reduction from Fe3O4 to FeO. Lastly, the final stage of the reduction process to obtain zero-valence Fe can be assigned to the peak that appeared within 710-900 °C . Similar reduction behavior has been reported in different studies, such as the work done by Bayat et al. [15]. The reduction peak of Fe/WO3 + ZrO2 can be seen at a higher range in comparison with the other catalysts with only Fe. This could be as a result of the strong interaction that exists between Fe and its carrier (WO3 + ZrO2).
For the Fe-Ni impregnated on different carriers, the reduction profiles show almost the same trend as the single metal-supported samples (Fe/ZrO2, Fe/La2O3 + ZrO2, and Fe/WO3 + ZrO2) according to the H2-TPR analysis. However, it is obvious that the Fe-Ni-supported catalysts have higher hydrogen consumption, which translates to higher intensity peaks. This could be because of the presence of NiO. Furthermore, there is a slight shift in the reduction temperature of the bi-metallic catalysts with the third peak appearing at a lower reduction temperature. In addition, the reduction peaks of NiO appear to have coincided with that of the progressive reduction of Fe3O4. Usually NiO is used to undergo reduction at about 345, 460 and 700 °C on typical support [14,25,26].

Catalytic Test
The results of methane decomposition (MD) for the duration of 4 h over Fe and Fe-Ni catalysts supported on different carriers are shown in Figure 4. The feed's space velocity was kept at 4000 mL(hgcat) −1 while the reaction was carried out at 1 atmosphere and 800 °C . The difference in the catalysts' activities has been reported in terms of H2 yield. The results reveal the effect of adding La2O3, WO3 to ZrO2 and the role of Ni in the catalysts' performances. According to Figure 4, Fe/WO3 For the Fe-Ni impregnated on different carriers, the reduction profiles show almost the same trend as the single metal-supported samples (Fe/ZrO 2 , Fe/La 2 O 3 + ZrO 2 , and Fe/WO 3 + ZrO 2 ) according to the H 2 -TPR analysis. However, it is obvious that the Fe-Ni-supported catalysts have higher hydrogen consumption, which translates to higher intensity peaks. This could be because of the presence of NiO. Furthermore, there is a slight shift in the reduction temperature of the bi-metallic catalysts with the third peak appearing at a lower reduction temperature. In addition, the reduction peaks of NiO appear to have coincided with that of the progressive reduction of Fe 3 O 4 . Usually NiO is used to undergo reduction at about 345, 460 and 700 • C on typical support [14,25,26].

Catalytic Test
The results of methane decomposition (MD) for the duration of 4 h over Fe and Fe-Ni catalysts supported on different carriers are shown in Figure 4. The feed's space velocity was kept at 4000 mL (hg cat ) −1 while the reaction was carried out at 1 atmosphere and 800 • C. The difference in the catalysts' activities has been reported in terms of H 2 yield. The results reveal the effect of adding La 2 O 3 , WO 3 to ZrO 2 and the role of Ni in the catalysts' performances. According to Figure 4, Fe/WO 3 + ZrO 2 has the highest H 2 yield of about 90% among the monometallic-supported samples. Fe/ZrO 2 and Fe/La 2 O 3 + ZrO 2 had a high initial H 2 yield corresponding to about 58% and 80%, respectively, but plummeted earlier during the time-on-stream. This could be as a result of carbon deposits, formed from MD, enveloping the active metal sites. In all, Fe supported on the ZrO 2 -doped carrier (i.e., WO 3 + ZrO 2 and La 2 O 3 + ZrO 2 ) gave a higher H 2 yield than Fe supported on the pristine ZrO 2 . Additionally, Fe/WO 3 + ZrO 2 , with the highest yield of H 2 , showed excellent stability during the test. The performance of Fe/WO 3 + ZrO 2 can be attributed to the properties of its support: like the unique acidity of WO 3 , its chemical and thermal stability [18][19][20]. The addition of WO 3 to the pristine ZrO 2 enhanced the diffusion of carbon deposits from the MD, making the active metal sites accessible to the feed, and its reported thermal stability contributes to the inhibition of metal aggregation, sintering, and the loss of catalyst structure.

Catalyst
Carbon Yield (g C/gcat) Space Velocity (mLg −1 h −1 ) Reaction Temperature (°C ) Reference 30%Fe/MgO 5.5 5000 700 [29] 70%Ni-10%Cu-10%Fe/Al2O3 8 2867 700 [33] 50Fe-6Co-Al2O3 41.0 45,000 675 [34] 62Fe-8Ni-Al2O3 122.0 45,000 675 [34] NiO-Al2O3 88 900,000 500 [35] 20%Fe/WO3 + ZrO2 0.58 4000 800 Present work Fakeeha et al. [25], studied hydrogen and carbon yield in methane decomposition over Fe supported on MgO and TiO2. In their research, magnesia-supported Fe gave better activity at high Fe loading of about 30-40%. This shines a light on the adoption of higher Fe content in catalytic methane decomposition, as this will assist in the diffusion of carbon deposits and ultimately bring about better stability for the catalyst. For better understanding of the catalysts' performances, their CH4 conversions are shown in the Figure 5 alongside their H2 yields. For the Fe-Ni-supported catalysts, the addition of Ni had a positive impact on the stability as well as the yield of H 2 for both Fe/ZrO 2 and Fe/La 2 O 3 + ZrO 2 catalysts. The initial H 2 yield can be seen to have increased from 58% to 82% for Fe/ZrO 2 and from 80% to 90% for Fe/La 2 O 3 + ZrO 2 . The influence of Ni was not really pronounced in the activity of the Fe-Ni/WO 3 + ZrO 2 catalyst, however, it raised the percentage yield of H 2 to a value above 90% and maintained the stability. Similarly, it was reported by Anis et al. that the Ni addition to Fe/Al 2 O 3 enhanced the performance of the catalyst in terms of methane conversion and hydrogen yield [27]. Ni being an active metal with stronger reduction potential than Fe in the electrochemical series would do more in reaction than Fe. The influence of the Ni addition in catalytic methane decomposition in the past works is summarized in Table 2.
Furthermore, activities of the catalysts can be related to the individual specific surface area. High surface area could translate to more room for active metal dispersion, which would transform the feed to the product after adsorption. It has been reported that a higher specific surface area and pore volume do enhance catalyst performance [28]. In this study, Fe/ZrO 2 had the lowest specific surface area (12.74 m 2 /g) and performed the worst while Fe/WO 3 + ZrO 2 had the highest area (26.68 m 2 /g) and gave the best performance.
We compared the H 2 yield from past research works with this study. The results obtained, together with the operating conditions, are summarized in Table 2. In addition to that, the carbon yields that resulted from the methane decomposition reactions in this study as well as that of the past work are shown in Table 3 for comparison.  Fakeeha et al. [25], studied hydrogen and carbon yield in methane decomposition over Fe supported on MgO and TiO 2 . In their research, magnesia-supported Fe gave better activity at high Fe loading of about 30-40%. This shines a light on the adoption of higher Fe content in catalytic methane decomposition, as this will assist in the diffusion of carbon deposits and ultimately bring about better stability for the catalyst. For better understanding of the catalysts' performances, their CH 4 conversions are shown in the Figure 5 alongside their H 2 yields.

Characterization of Spent Catalysts and Carbon Produced
Thermo-gravimetric (TG) analysis was carried out on the used catalyst samples. They were taken out of the reactor and kept in a safe container. The results of the TG analysis are presented in Figure  6. This figure shows the percentage of weight loss as a function of temperature. The weight loss is a

Thermo-Gravimetric Analysis (TGA)
Thermo-gravimetric (TG) analysis was carried out on the used catalyst samples. They were taken out of the reactor and kept in a safe container. The results of the TG analysis are presented in Figure 6. This figure shows the percentage of weight loss as a function of temperature. The weight loss is a representation of the amount of carbon deposits over the used catalysts that have been converted to CO/CO 2 while being heated in the atmosphere. Fe/WO 3 + ZrO 2 had the highest amount of carbon deposits of about 58% while the least reactive Fe/ZrO 2 had the lowest carbon deposits amounting to about 11%.

Raman Spectroscopy Analysis
Further post-reaction analysis was performed on the used catalysts to determine the degree of graphitization of the carbon deposits. The spectra obtained from the Raman analysis are displayed in Figure 7. For all the used catalysts, peaks related to the D (Defect) and G (Graphite) bands can be seen at approximately 1472 and 1535 cm −1 Raman shifts, respectively. The G band represents the ideal vibration mode of the well-ordered graphitic layers having E2g symmetry, while the D band is caused by defects in the graphitic lattice. Its source has been expatiated by the double resonant Raman scattering [36].
It has been reported that the ratio of the integral intensities of ID to IG and the bands' full width at half maximum height (FWHM) could be correlated with the microstructure of carbonaceous materials. Additionally, the intensity ratio, ID/IG, represents the degree of the crystalline order of the graphite in the carbon deposits [37,38]. The implication of this is that a higher intensity, above 1, indicates a higher structural disorder in the carbon deposits, while a lower ratio, below 1, represents lesser structural disorder [38]. From Figure 7, Fe/WO3 + ZrO2 gave carbon deposits with the highest structural defect, while Fe-Ni/La2O3 + ZrO2 had the least structurally disordered carbon deposits. This implies that the carbon deposits generated by the Fe/WO3 + ZrO2 catalyst are mostly amorphous and those of the Fe-Ni/La2O3 + ZrO2 catalyst contain higher amounts of graphite. It can be seen that the ID/IG ratios of the bimetallic catalysts are less than those of the monometallic counterparts, with the exception of Fe/ZrO2. Hence, the addition of Ni improved the structural order of the carbon deposits on Fe/WO3 + ZrO2 and Fe-Ni/La2O3 + ZrO2. For the used Fe-Ni-supported catalysts, the same trend as in Fe/x-ZrO 2 (x = 0, La 2 O 3 , WO 3 ) was observed with a difference of higher carbon deposits for both Fe-Ni/ZrO 2 and Fe-Ni/La 2 O 3 + ZrO 2 . The increase in carbon deposits can be linked to the presence of Ni, which enhanced the performance of the catalysts as shown before in Figure 4.

Raman Spectroscopy Analysis
Further post-reaction analysis was performed on the used catalysts to determine the degree of graphitization of the carbon deposits. The spectra obtained from the Raman analysis are displayed in Figure 7. For all the used catalysts, peaks related to the D (Defect) and G (Graphite) bands can be seen at approximately 1472 and 1535 cm −1 Raman shifts, respectively. The G band represents the ideal vibration mode of the well-ordered graphitic layers having E 2g symmetry, while the D band is caused by defects in the graphitic lattice. Its source has been expatiated by the double resonant Raman scattering [36].
It has been reported that the ratio of the integral intensities of I D to I G and the bands' full width at half maximum height (FWHM) could be correlated with the microstructure of carbonaceous materials. Additionally, the intensity ratio, I D /I G , represents the degree of the crystalline order of the graphite in the carbon deposits [37,38]. The implication of this is that a higher intensity, above 1, indicates a higher structural disorder in the carbon deposits, while a lower ratio, below 1, represents lesser structural disorder [38]. From Figure 7, Fe/WO 3 + ZrO 2 gave carbon deposits with the highest structural defect, while Fe-Ni/La 2 O 3 + ZrO 2 had the least structurally disordered carbon deposits. This implies that the carbon deposits generated by the Fe/WO 3 + ZrO 2 catalyst are mostly amorphous and those of the Fe-Ni/La 2 O 3 + ZrO 2 catalyst contain higher amounts of graphite. It can be seen that the I D /I G ratios of the bimetallic catalysts are less than those of the monometallic counterparts, with the exception of Fe/ZrO 2 . Hence, the addition of Ni improved the structural order of the carbon deposits on Fe/WO 3 + ZrO 2 and Fe-Ni/La 2 O 3 + ZrO 2 .

Catalyst Preparation
The supported Fe monometallic catalyst and Fe-Ni bi-metallic catalysts were synthesized by an incipient wetness-impregnation technique. Briefly, a stoichiometric amount of the investigated supports was added to an aqueous solution containing an appropriate amount of iron precursor or a mixture of iron and nickel precursors. Fe and Fe-Ni contents were maintained at 20 wt%, i.e., for the monometallic samples, Fe loading was 20% while both Fe and Ni had 20% loading each for the bimetallic samples. The mixture was stirred for 4 h before drying at 120 °C overnight. After that, the resultant samples were calcined in the air by heating from the ambient temperature to 800 °C over a 7-h period and keeping it at this temperature for 3 h. The resulting calcined samples are designated as Fe/x-ZrO2 and Fe-Ni/x-ZrO2, where (x = 0, La2O3, ZrO2).

Catalyst Activity
Methane decomposition reaction on the surface of Fe and Fe-Ni-supported catalysts was carried out in the PID Eng & Tech microactivity reaction system. The experiments were performed by using 300 mg of the prepared catalysts sandwiched between two layers of glass wool in the fixed bed reactor. Before starting the reaction, the synthesized catalyst was activated in situ at 800 °C in a hydrogen atmosphere for 1.5 h. The reactants' gas mixture, consisting of 65% CH4 with N2 as a balance, was introduced into the reactor at a total flow rate of 20 mL/min. The decomposition of the CH4 reaction was conducted at atmospheric pressure and 800 °C reaction temperature. The analysis  ] were obtained from Daiichi Kigenso (Osaka, Japan). All chemicals were used as-received without further purification.

Catalyst Preparation
The supported Fe monometallic catalyst and Fe-Ni bi-metallic catalysts were synthesized by an incipient wetness-impregnation technique. Briefly, a stoichiometric amount of the investigated supports was added to an aqueous solution containing an appropriate amount of iron precursor or a mixture of iron and nickel precursors. Fe and Fe-Ni contents were maintained at 20 wt%, i.e., for the monometallic samples, Fe loading was 20% while both Fe and Ni had 20% loading each for the bimetallic samples. The mixture was stirred for 4 h before drying at 120 • C overnight. After that, the resultant samples were calcined in the air by heating from the ambient temperature to 800 • C over a 7-h period and keeping it at this temperature for 3 h. The resulting calcined samples are designated as Fe/x-ZrO 2 and Fe-Ni/x-ZrO 2 , where (x = 0, La 2 O 3 , ZrO 2 ).

Catalyst Activity
Methane decomposition reaction on the surface of Fe and Fe-Ni-supported catalysts was carried out in the PID Eng & Tech microactivity reaction system. The experiments were performed by using Catalysts 2020, 10, 793 11 of 14 300 mg of the prepared catalysts sandwiched between two layers of glass wool in the fixed bed reactor. Before starting the reaction, the synthesized catalyst was activated in situ at 800 • C in a hydrogen atmosphere for 1.5 h. The reactants' gas mixture, consisting of 65% CH 4 with N 2 as a balance, was introduced into the reactor at a total flow rate of 20 mL/min. The decomposition of the CH 4 reaction was conducted at atmospheric pressure and 800 • C reaction temperature. The analysis was started 20 min after the beginning of the decomposition reaction and contained for 240 min for each of the synthesized catalysts. The decomposition products, as well as the unconverted reactants, were analyzed by an online Shimadzu GC 2004 gas chromatography.
The yield of H 2 was calculated based on the following equation:

Catalyst Characterization
Powder X-ray diffraction (XRD) patterns of prepared monometallic (Fe) as well as the bimetallic (Fe-Ni) catalysts were conducted using the Miniflex Rigaku diffractometer operated at 40 kV and 40 mA, equipped with Cu Kα X-ray radiation. The isotherms of nitrogen sorption were determined in the Micromeritics Tristar II 3020 surface area and porosity analyzer at −196 • C after the samples were outgassed at 200 • C for 3 h to remove any adsorbed gases or vapors. The distributions of pore size of the samples were analyzed from the adsorption profiles of isotherms using the Barrett-Joyner-Halenda (BJH) model. The H 2 -TPR of the synthesized fresh catalysts as well as the spent samples was performed on the Micromeritics Auto Chem II 2920. The analyses were conducted over a temperature range of 50-800 • C and in a mixture of 40 mL/min flow of 10% H 2 /Ar in the TPR analysis. The coke formation and the amount of carbon deposit on the surface of tested catalysts were measured by the thermo-gravimetric analyzer (Shimadzu-TGA). The samples were heated up to 1000 • C while the temperature was ramped at 10 • C/min. The deposited carbons were burned in an air atmosphere by heating the samples up to 1000 • C at a rate of 10 • C/min and the weight losses were recorded. A laser Raman spectroscopy NMR-4500 was used to study the extent of graphitization of the carbon deposited over the used catalysts. The measurement was performed using a 5x magnification objective lens and beam excitation wavelength of 532 nm. The adopted range for spectra recording was 1350-1600 cm −1 .

Conclusions
This research focused on hydrogen yield from catalytic methane decomposition using Fe and Fe-Ni supported on different carriers. The catalysts were synthesized by the wet-impregnation method and tested in the reaction under atmospheric conditions, at 800 • C, and at 4000 mL/(hg cat ) −1 space velocity.
Among the monometallic catalysts, Fe/WO 3 + ZrO 2 had the highest H 2 yield of about 90%. Its excellent performance was attributed to the special properties of its support such as dense acid sites, the ability to form Bronsted acid on reduction, stable chemical and thermal properties, and its wide use in the cracking of heavy crude oil fractions. When Ni was added to Fe/x-ZrO 2 (x = 0, La 2 O 3 , WO 3 ), a significant increase in the activities of Fe/ZrO 2 and Fe/La 2 O 3 + ZrO 2 was observed. These two catalysts had a relatively low conversion and poor stability before the addition of Ni.
Both fresh and used catalysts were analyzed by some characterization techniques. The X-ray diffractogram of the fresh catalyst samples showed the existing phases. Monoclinic and tetragonal zirconia were identified on all the samples at their typical angle of diffraction. The presence of La 2 O 3 resulted into peaks with higher intensity. Crystal phases of iron oxide such as hematite and maghemite were found on all samples. However, magnetite was identified only on the bi-metallic samples. Additionally, NiO phases were seen on the Ni-promoted samples. The nitrogen physisorption analysis revealed that the surface areas of the catalysts decreased relative to their corresponding supports. This could be as a result of the active metals blocking the pores of the support. The TPR revealed the progression in the reduction of Fe to its zero-valent form and no interaction of iron oxide with the support was observed. The reduction profiles were virtually the same for Fe and Fe-Ni catalysts except that the intensity of the peaks was higher for the latter. Reduction peaks of NiO appeared to have coincided with those of Fe 3 O 4 , as reported reduction temperatures of NiO fall within range of those of iron oxide. Fe and Fe-Ni supported on WO 3 + ZrO 2 have the highest amount of carbon deposits compared to their counterparts according to the TG analysis. The Raman analysis revealed that the carbon deposited on the used catalysts has a higher structural defect, however, the structural disorder reduced with the addition of NiO.