Scalable Synthesis of Oxygen Vacancy-Rich Unsupported Iron Oxide for Efficient Thermocatalytic Conversion of Methane to Hydrogen and Carbon Nanomaterials

Thermocatalytic methane decomposition (TCMD) involving metal oxides is a more environmentally friendly and cost-effective strategy for scalable hydrogen fuel production compared to traditional methane steam reforming (MSR), as it requires less energy and produces fewer CO/CO2 emissions. However, the unsupported metal oxide catalysts (such as α-Fe2O3) that would be suited for this purpose exhibit poor performance in TCMD. To overcome this issue, a novel strategy was developed as a part of this work, whereby oxygen vacancies (OVs) were introduced into unsupported α-Fe2O3 nanoparticles (NPs). Systematic characterization of the obtained materials through analytical techniques demonstrated that mesoporous nanostructured unsupported α-Fe2O3 with abundant oxygen vacancies (OV-rich α-Fe2O3 NPs) could be obtained by direct thermal decomposition of ferric nitrate at different calcination temperatures (500, 700, 900, and 1100 °C) under ambient conditions. The thermocatalytic activity of the resulting OV-rich α-Fe2O3 NPs was assessed by evaluating the methane conversion, hydrogen formation rate, and amount of carbon deposited. The TCMD results revealed that 900 °C was the most optimal calcination temperature, as it led to the highest methane conversion (22.5%) and hydrogen formation rate (47.0 × 10−5 mol H2 g−1 min−1) after 480 min. This outstanding thermocatalytic performance of OV-rich α-Fe2O3 NPs is attributed to the presence of abundant OVs on their surfaces, thus providing effective active sites for methane decomposition. Moreover, the proposed strategy can be cost-effectively scaled up for industrial applications, whereby unsupported metal oxide NPs can be employed for energy-efficient thermocatalytic CH4 decomposition into hydrogen fuel and carbon nanomaterials.


Introduction
Owing to the growing demand for the abandonment of fossil fuels and the adoption of clean energy sources, academic research into applications involving hydrogen has increased considerably in recent years. The available evidence indicates that electricity can be efficiently obtained from hydrogen-based sources without any environmental pollution [1][2][3][4]. However, at present, hydrogen is typically attained through methane steam reforming (MSR) and partial natural gas oxidation, resulting in considerable CO 2 emissions [5][6][7]. Thus, greener methane decomposition strategies are urgently required, as methane is superior to all other hydrocarbons in terms of the hydrogen/carbon ratio. Moreover, it is

Characterization Techniques
For crystalline phase identification, the obtained fresh OV-rich α-Fe2O3 NPs and spent samples collected after use for TCMD were studied via X-ray diffraction (XRD) analysis over the 2θ = 10−70° scanning range, using the X-ray diffractometer (Rigaku International, Tokyo, Japan), Cu Kα radiation (λ = 1.5406 A). The surface morphology of the obtained fresh OV-rich α-Fe2O3 NPs and spent samples were observed using a scanning electron microscope (FESEM, QUANTA 250 FEI; Hillsboro, OR, USA), while the textural properties of the fresh samples, including surface area, pore volume, and pore radius, were measured using N2 physisorption isotherms using a Quantachrome Autosorb IQ model ASIQA3V600000-6 instrument. X-ray photoelectron spectroscopy (XPS) was further employed to investigate the near-surface chemical composition and oxidation states of the studied materials. A Thermo K Alpha spectrometer operating at 1486.6 eV with a 400 μm spot size and charge correction was utilized for this purpose. As an internal standard, all binding energy estimates were calibrated to the C 1s energy (284.5 eV). In addition, thermogravimetric analyses (TGA) were carried out in ambient air using a Netzsch proteus 70 instrument at 10 °C/min increments within the 25-1000 °C temperature range. Collected TGA data demonstrated typical mass loss with regard to temperature, and their corresponding first derivative thermogravimetric (DTG) graphs provided information on mass loss rate versus temperature.

Characterization Techniques
For crystalline phase identification, the obtained fresh OV-rich α-Fe 2 O 3 NPs and spent samples collected after use for TCMD were studied via X-ray diffraction (XRD) analysis over the 2θ = 10−70 • scanning range, using the X-ray diffractometer (Rigaku International, Tokyo, Japan), Cu Kα radiation (λ = 1.5406 A). The surface morphology of the obtained fresh OV-rich α-Fe 2 O 3 NPs and spent samples were observed using a scanning electron microscope (FESEM, QUANTA 250 FEI; Hillsboro, OR, USA), while the textural properties of the fresh samples, including surface area, pore volume, and pore radius, were measured using N 2 physisorption isotherms using a Quantachrome Autosorb IQ model ASIQA3V600000-6 instrument. X-ray photoelectron spectroscopy (XPS) was further employed to investigate the near-surface chemical composition and oxidation states of the studied materials. A Thermo K Alpha spectrometer operating at 1486.6 eV with a 400 µm spot size and charge correction was utilized for this purpose. As an internal standard, all binding energy estimates were calibrated to the C 1s energy (284.5 eV). In addition, thermogravimetric analyses (TGA) were carried out in ambient air using a Netzsch proteus 70 instrument at 10 • C/min increments within the 25-1000 • C temperature range. Collected TGA data demonstrated typical mass loss with regard to temperature, and their corresponding first derivative thermogravimetric (DTG) graphs provided information on mass loss rate versus temperature.

Thermocatalytic Activity Experiments
The activity of unsupported nanostructured OV-rich α-Fe 2 O 3 NPs was investigated in a quartz tubular fixed-bed reactor with a 1.5 cm internal diameter and 50 cm total length. For this purpose, 0.5 g of the previously prepared fresh catalyst sample was placed in the center of the tube between the ceramic fiber plugs preheated to 800 • C. Next, a 3:1 mixture of CH 4 and N 2 feed gas (supplied by AHG at 99.99% purity) was introduced into the reactor at a 20 mL/min flow rate, which was controlled by a mass gas flowmeter (MKS PR4000B). During data analysis, the nitrogen present in the mixture was used as the internal standard. The OV-rich α-Fe 2 O 3 NPs were not subjected to any pretreatment because methane is considered a reductant. Reaction products were examined using online Gas Chromatography, on an Agilent GC Model 7890B equipped with 19043 Restek Micro packed GC Columns (Shin carbon ST 80/100 2 M, 0.53 mm), while a thermal conductivity detector was utilized to analyze the output gas and determine the hydrogen, nitrogen, and unreacted methane content.

Characterization of Fresh Samples
The synthesis of OV-rich α-Fe 2 O 3 NPs through the calcination of pure ferric nitrate nonahydrate (Fe(NO 3 ) 3 ·9H 2 O) at different temperatures (500, 700, 900, and 1100 • C) under ambient conditions is an important aspect of this study, as it allows for the tailoring of the properties of the resulting NPs to optimize their catalytic performance in thermocatalytic methane decomposition.
Interestingly, we observed that the mass of the obtained OV-rich α-Fe 2 O 3 NPs remained constant at different calcination temperatures, as demonstrated in the inset of Figure 1. This finding suggests that the yield of the calcination process is independent of the calcination temperature.
This observation is crucial for the scalability of the synthesis process, as it indicates that the amount of OV-rich α-Fe 2 O 3 NPs that can be obtained is not limited by the calcination temperature. Therefore, this synthesis method can be easily scaled up to produce large quantities of OV-rich α-Fe 2 O 3 NPs for industrial applications, such as energy-efficient thermocatalytic methane decomposition into hydrogen fuel and carbon nanomaterials. Figure 2 depicts the TGA-DTG curves of pure Fe(NO 3 ) 3 ·9H 2 O under ambient conditions, demonstrating two distinct mass loss regions. These regions indicate that the decomposition of Fe(NO 3 ) 3 ·9H 2 O salt into Fe 2 O 3 occurs in two main steps. The first step occurring between 75 and 150 • C could be attributed to the removal of the crystalline water bound to the salt and the decomposition of the nitrates within the salt, which corresponds to the formation of ferric oxyhydroxide (FeOOH). The second step, occurring between 150 and 300 • C, indicates the formation of Fe 2 O 3 [30]. Furthermore, a slight decrease observed before the temperature of 75 °C could potentially be attributed to the loss of adsorbed water weight. The loss of weight corresponding to the removal of adsorbed water from the precursor salt typically begins between 45 and 70 °C, which agrees with our results in Figure 2. It is worth noting that the TGA-DTG analysis reveals that calcination temperatures of around 500 °C and above are effective in removing the nitrate and hydrate groups associated with the ferrous metal. This observation is consistent with the negligible change in the weight of the resulting Furthermore, a slight decrease observed before the temperature of 75 • C could potentially be attributed to the loss of adsorbed water weight. The loss of weight corresponding to the removal of adsorbed water from the precursor salt typically begins between 45 and Nanomaterials 2023, 13, 2461 5 of 19 70 • C, which agrees with our results in Figure 2. It is worth noting that the TGA-DTG analysis reveals that calcination temperatures of around 500 • C and above are effective in removing the nitrate and hydrate groups associated with the ferrous metal. This observation is consistent with the negligible change in the weight of the resulting Fe 2 O 3 when increasing the calcination temperature from 500 • C to 1100 • C at ambient conditions, as shown in Figure 1.
XRD was employed to examine the crystalline structure and determine the phase purity of the obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures, and the obtained diffraction patterns are shown in Figure 3.   [31]. The absence of any secondary phases confirms that Fe(NO3)3.9H2O has completely decomposed to the α-Fe2O3 hematite phase. Figure 3 demonstrates an increase in the intensity of the diffraction peaks upon increasing the calcination temperature to 700 °C, providing the highest degree of crystallinity for the obtained hematite, α-Fe2O3. The calcination process at 900 and 1100 °C results in a slight decrease in crystallinity of the formed hematite phase, which is attributed to the sintering of the nanoparticles at these temperatures, which may lead to the formation of defects [32].
Scanning electron microscopy (SEM) was employed to examine the morphological features of the obtained OV-rich α-Fe2O3 NPs at different calcination temperatures, and the results are depicted in Figure 4. From Figure 4, it can be seen that the particle size changes gradually with increasing calcination temperatures from 500 to 1100 °C. Although the particle size increases slightly at 700 °C, it remains small and randomly inter-  [31]. The absence of any secondary phases confirms that Fe(NO 3 ) 3 .9H 2 O has completely decomposed to the α-Fe 2 O 3 hematite phase. Figure 3 demonstrates an increase in the intensity of the diffraction peaks upon increasing the calcination temperature to 700 • C, providing the highest degree of crystallinity for the obtained hematite, α-Fe 2 O 3 . The calcination process at 900 and 1100 • C results in a slight decrease in crystallinity of the formed hematite phase, which is attributed to the sintering of the nanoparticles at these temperatures, which may lead to the formation of defects [32].
Scanning electron microscopy (SEM) was employed to examine the morphological features of the obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures, and the results are depicted in Figure 4. From Figure 4, it can be seen that the particle size changes gradually with increasing calcination temperatures from 500 to 1100 • C. Although the particle size increases slightly at 700 • C, it remains small and randomly interconnected on a cotton-like surface. The effect of the calcination temperature between 900 and 1100 • C is clearer, as the particles are formed from quasi-spherical shapes and undefined shapes with different sizes, and their sizes increased with the increase in calcination temperatures from 900 to 1100 • C. It is known that when the calcination temperature rises, the particle size becomes larger due to particle growth and/or agglomeration [12]. The textural properties of the OV-rich α-Fe2O3 NPs were measured utilizing N2 adsorption-desorption isotherms and Brunauer-Emmett-Teller (BET)/Barrett-Joyner-Halenda (BJH) analysis, and the findings are presented in Table 1 and Figure 5. As can be seen from the data reported in Table 1, calcination temperature significantly affected the BET surface parameters of the studied materials, whereby the BET surface area declined from 16.70 to 3.90 m 2 /g as the calcination temperature increased from 500 to 1100 °C.  The textural properties of the OV-rich α-Fe 2 O 3 NPs were measured utilizing N 2 adsorption-desorption isotherms and Brunauer-Emmett-Teller (BET)/Barrett-Joyner-Halenda (BJH) analysis, and the findings are presented in Table 1 and Figure 5. As can be seen from the data reported in Table 1, calcination temperature significantly affected the BET surface parameters of the studied materials, whereby the BET surface area declined from 16.70 to 3.90 m 2 /g as the calcination temperature increased from 500 to 1100 • C.  This trend could be attributed to particle agglomeration associated with elevated temperatures. Figure 5a displays the N2 adsorption−desorption isotherms of the obtained OV-rich α-Fe2O3 NPs at different calcination temperatures and shows that they are of type IV with H3 hysteresis loop, which is typical of mesoporous hematite α-Fe2O3 [33]. Furthermore, as the hysteresis loop areas decline with increasing calcination temperature, this finding further confirms that calcination may induce the collapse of some porous materials [34]. Usually, H3-type hysteresis is produced by solids with aggregate or agglomerate particles of non-uniform size and/or shape [35]. The isotherms depicted in Figure 5b reveal the mesoporous nature of OV-rich α-Fe2O3 NPs based on the pore size distribution, which remained in the 19−22 nm range.
Thus, from the SEM images ( Figure 4), it is concluded that the particle size increased with the increase in the calcination temperature due to the sintering and/or agglomeration of the particles. This observation is consistent with the results of the surface area analysis (Table 1), as it decreased with the increase in calcination temperature.
The chemical composition and elemental states of the obtained OV-rich α-Fe2O3 NPs at different calcination temperatures were probed by X-ray photoelectron spectroscopy (XPS), and the findings are shown in Figure 6. The XPS survey spectra of the examined OV-rich α-Fe2O3 NPs depicted in Figure 6 confirm that Fe, O, and C are the only constituents, whereby the presence of carbon is mainly attributed to its absorption from the atmosphere before XPS measurements. On the other hand, the oxygen peaks indicate successful α-Fe2O3 formation. The O 1s spectra of the OV-rich α-Fe2O3 NPs samples presented in Figure 7 confirm the presence of four chemical states of the surface oxygen species. The main peak located at around 529.5 eV can be ascribed to the lattice oxygen of α-Fe2O3 (Fe 3+ -O, labeled as O2), while the signal at the 531.8 eV binding energy corresponds to the hydroxyl group (-OH, labeled O3). Interestingly, the peak resulting from the hydroxyl group shifts towards lower energy for the sample that was calcined at 900 °C, which indicates these samples are oxygen-deficient. Thus, the peak O3 at around 531.1 in the FeN-900 °C sample is attributed to the low-coordinated oxygen species adsorbed onto the oxygen vacancies, likely arising from the formation of oxygen vacancies on the NP's surface. On the other hand, the peaks that appear in the 533-536 eV energy range (labeled as O4) may be related to the surface-adsorbed H2O molecules and carbonate species [27,29,36]. Finally, the peak (O1) that appeared at 527.6 eV in FeN-500 °C, FeN-700 °C, and FeN-900 °C samples is attributed to the fact that some of the Fe 3+ on the surface can be changed to a low oxidation state such as Fe 2+ which is usually associated with the formation of OVs. The OVs are expected to play an important role in the adsorption process, as these vacancy sites can act as initiators for the adsorption of mole- This trend could be attributed to particle agglomeration associated with elevated temperatures. Figure 5a displays the N 2 adsorption−desorption isotherms of the obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures and shows that they are of type IV with H3 hysteresis loop, which is typical of mesoporous hematite α-Fe 2 O 3 [33]. Furthermore, as the hysteresis loop areas decline with increasing calcination temperature, this finding further confirms that calcination may induce the collapse of some porous materials [34]. Usually, H3-type hysteresis is produced by solids with aggregate or agglomerate particles of non-uniform size and/or shape [35]. The isotherms depicted in Figure 5b reveal the mesoporous nature of OV-rich α-Fe 2 O 3 NPs based on the pore size distribution, which remained in the 19−22 nm range.
Thus, from the SEM images (Figure 4), it is concluded that the particle size increased with the increase in the calcination temperature due to the sintering and/or agglomeration of the particles. This observation is consistent with the results of the surface area analysis (Table 1), as it decreased with the increase in calcination temperature.
The chemical composition and elemental states of the obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures were probed by X-ray photoelectron spectroscopy (XPS), and the findings are shown in Figure 6. The XPS survey spectra of the examined OV-rich α-Fe 2 O 3 NPs depicted in Figure 6 confirm that Fe, O, and C are the only constituents, whereby the presence of carbon is mainly attributed to its absorption from the atmosphere before XPS measurements. On the other hand, the oxygen peaks indicate successful α-Fe 2 O 3 formation. The O 1s spectra of the OV-rich α-Fe 2 O 3 NPs samples presented in Figure 7 confirm the presence of four chemical states of the surface oxygen species. The main peak located at around 529.5 eV can be ascribed to the lattice oxygen of α-Fe 2 O 3 (Fe 3+ -O, labeled as O 2 ), while the signal at the 531.8 eV binding energy corresponds to the hydroxyl group (-OH, labeled O 3 ). Interestingly, the peak resulting from the hydroxyl group shifts towards lower energy for the sample that was calcined at 900 • C, which indicates these samples are oxygen-deficient. Thus, the peak O 3 at around 531.1 in the FeN-900 • C sample is attributed to the low-coordinated oxygen species adsorbed onto the oxygen vacancies, likely arising from the formation of oxygen vacancies on the NP's surface. On the other hand, the peaks that appear in the 533-536 eV energy range (labeled as O 4 ) may be related to the surface-adsorbed H 2 O molecules and carbonate species [27,29,36]. Finally, the peak (O 1 ) that appeared at 527.6 eV in FeN-500 • C, FeN-700 • C, and FeN-900 • C samples is attributed to the fact that some of the Fe 3+ on the surface can be changed to a low oxidation state such as Fe 2+ which is usually associated with the formation of OVs. The OVs are expected to play an important role in the adsorption process, as these vacancy sites can act as initiators for the adsorption of molecules in many surface reactions (i.e., many species such as H 2 O and CO 2 will adsorb on the OVs from ambient air). Nanomaterials 2023, 13, x FOR PEER REVIEW 9 of 23 cules in many surface reactions (i.e., many species such as H2O and CO2 will adsorb on the OVs from ambient air).  As defect states (such as OVs) are known to appear on the surface of metal oxides at high temperatures, it is clear from Figure 7 that the number of defects increases with increasing temperature from FeN-500 °C-FeN-900 °C. Moreover, the O2 peak ascribed to the lattice oxygen in the FeN-900 °C sample is shifted to lower bending energy compared As defect states (such as OVs) are known to appear on the surface of metal oxides at high temperatures, it is clear from Figure 7 that the number of defects increases with increasing temperature from FeN-500 • C-FeN-900 • C. Moreover, the O 2 peak ascribed to the lattice oxygen in the FeN-900 • C sample is shifted to lower bending energy compared to the other samples, which implies the oxygen vacancies are present in high concentration in this sample. Figure 8 illustrates the Fe 2p 3/2 XPS spectra of the OV-rich α-Fe 2 O 3 NPs obtained in this study. The spectra exhibit five peaks, with two peaks located at 707.6 and 709.7 eV attributed to Fe 2+ and the peaks at 712.6 and~715.4 eV corresponding to the oxidation state of Fe 3+ . Additionally, a peak at~720.6 eV can be assigned to the satellite peak, as suggested by previous studies [29,37,38]. The Fe 2+ /Fe 3+ ratio was calculated to be 0. Notably, an additional small peak appeared at 706.2 eV in the FeN-900 °C sam which is attributed to Fe 0 , suggesting that some Fe 3+ was reduced into Fe 0 , a phenome commonly associated with the formation of OVs. The absence of this peak in other sa ples suggests that the FeN-900 °C sample may contain a higher concentration of O OVs may accelerate the iron oxide reduction process and lead to the formation of m active sites, thereby enhancing catalytic activity.

Thermocatalytic Activity of the Obtained OV-Rich α-Fe2O3 NPs
The obtained OV-rich α-Fe2O3 NPs at different calcination temperatures were u as catalysts for the TCMD reaction. Their thermocatalytic activities were investigated fixed operating temperature of 800 °C and a gas flow rate of (20 mL/min). It should noted that these materials did not undergo pre-reduction treatment before the react Figure 9a,b show the methane conversion percentage (%) and hydrogen formation rat a function of time on stream (TOS). From Figure 9a,b, it is clear that the prepared OVα-Fe2O3 NPs at 1100 °C showed the lowest activity, where the highest methane conv sions achieved were 9.20%, whereas the highest hydrogen formation rate was 20.80 × mol H2 g −1 min −1 . In contrast, the prepared OV-rich α-Fe2O3 NPs at 500, 700, and 900 displayed higher activity, which can be observed in the initial phase of the react Notably, an additional small peak appeared at 706.2 eV in the FeN-900 • C sample, which is attributed to Fe 0 , suggesting that some Fe 3+ was reduced into Fe 0 , a phenomenon commonly associated with the formation of OVs. The absence of this peak in other samples suggests that the FeN-900 • C sample may contain a higher concentration of OVs. OVs may accelerate the iron oxide reduction process and lead to the formation of more active sites, thereby enhancing catalytic activity.

Thermocatalytic Activity of the Obtained OV-Rich α-Fe 2 O 3 NPs
The obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures were used as catalysts for the TCMD reaction. Their thermocatalytic activities were investigated at a fixed operating temperature of 800 • C and a gas flow rate of (20 mL/min). It should be noted that these materials did not undergo pre-reduction treatment before the reaction. Figure 9a,b show the methane conversion percentage (%) and hydrogen formation rate as a function of time on stream (TOS). From Figure 9a,b, it is clear that the prepared OV-rich α-Fe 2 O 3 NPs at 1100 • C showed the lowest activity, where the highest methane conversions achieved were 9.20%, whereas the highest hydrogen formation rate was 20.80 × 10 −5 mol H 2 g −1 min −1 . In contrast, the prepared OV-rich α-Fe 2 O 3 NPs at 500, 700, and 900 • C displayed higher activity, which can be observed in the initial phase of the reaction, where the maximum methane conversions reached 37.8, 26.1, and 35.2%, respectively, while the maximum hydrogen formation rate was 74.0 × 10 −5 , 40.3 × 10 −5 and 68.2 × 10 −5 mol H 2 g −1 min −1 . It appears that these samples have a partially reducible component that forms active sites, which was followed by a rapid carbon deposit on them, subsequently suppressing their performance. Moreover, after 200 min of TOS, it is noted that the samples of OV-rich α-Fe 2 O 3 NPs showed different activity as there was a gradual increase in their activity, and the order of performance was as follows: the prepared OV-rich α-Fe 2 O 3 NPs at 900 • C > OV-rich α-Fe 2 O 3 NPs at 700 • C > OV-rich α-Fe 2 O 3 NPs at 500 • C > OV-rich α-Fe 2 O 3 NPs at 1100 • C, and the maximum methane conversions achieved were 22.5, 20.6, 18.6, and 9.3%, respectively, whereas the maximum hydrogen formation rate was 47.0 × 10 −5 , 44.0 × 10 −5 , 40.7 × 10 −5 and 20.83 mol H 2 g −1 min −1 , respectively.  Although these materials have the same phase of α-Fe2O3, the reaction was carried out under the same conditions, and the amount of catalyst and its iron content were constant, there is a variation in their activities. It appears that the surface area did not play a role in the activity of these materials, as increasing the calcination temperature led to an increase in the particle size and thus a decrease in the surface area, as seen in SEM images in Figure 4, and a sharp decrease in its surface area as shown in Table 1.
In an attempt to find the reasons for these disparate activities, it would appear to be due to the presence of defects in the iron oxide composition that arose from the effect of calcination temperatures. The results of XPS analysis (Figures 7 and 8) showed that there are OVs on the α-Fe2O3 NPs surface, which may have played a significant role in the activity of these materials. It is noteworthy that the FeN-500 °C sample at the beginning of the reaction showed a high methane conversion percentage and hydrogen formation rate compared to the FeN-900 °C sample. However, the high activity of FeN-900 °C at the end of the TCMD reaction may be due to the presence of a high concentration of OVs on its surface, as observed in the XPS results in Figures 7 and 8. The OVs on the α-Fe2O3 NPs surface play an important role in accelerating the iron oxide reduction process compared to those that contain fewer vacancies, which lead to the formation of more active sites during the TCMD, which act as effective active sites for the adsorption of methane. On the other hand, the low catalytic activity of the FeN-1100 °C catalysts can be attributed to the low concentration of the OVs, as is clear from the XPS results of these catalysts, which made them difficult to reduce during the reaction run period as these catalysts were not subjected to a prior reduction process [39].
It is worth mentioning that the activity of the prepared unsupported α-Fe2O3 cata- Although these materials have the same phase of α-Fe 2 O 3 , the reaction was carried out under the same conditions, and the amount of catalyst and its iron content were constant, there is a variation in their activities. It appears that the surface area did not play a role in the activity of these materials, as increasing the calcination temperature led to an increase in the particle size and thus a decrease in the surface area, as seen in SEM images in Figure 4, and a sharp decrease in its surface area as shown in Table 1.
In an attempt to find the reasons for these disparate activities, it would appear to be due to the presence of defects in the iron oxide composition that arose from the effect of calcination temperatures. The results of XPS analysis (Figures 7 and 8) showed that there are OVs on the α-Fe 2 O 3 NPs surface, which may have played a significant role in the activity of these materials. It is noteworthy that the FeN-500 • C sample at the beginning of the reaction showed a high methane conversion percentage and hydrogen formation rate compared to the FeN-900 • C sample. However, the high activity of FeN-900 • C at the end of the TCMD reaction may be due to the presence of a high concentration of OVs on its surface, as observed in the XPS results in Figures 7 and 8. The OVs on the α-Fe 2 O 3 NPs surface play an important role in accelerating the iron oxide reduction process compared to those that contain fewer vacancies, which lead to the formation of more active sites during the TCMD, which act as effective active sites for the adsorption of methane. On the other hand, the low catalytic activity of the FeN-1100 • C catalysts can be attributed to the low concentration of the OVs, as is clear from the XPS results of these catalysts, which made them difficult to reduce during the reaction run period as these catalysts were not subjected to a prior reduction process [39].
It is worth mentioning that the activity of the prepared unsupported α-Fe 2 O 3 catalysts by our scalable and cost-effective strategy is comparable with previously reported Fe-based catalysts for TCMD, as presented in Table 2, where WHSV is weight hourly space velocity.  Figure 10 shows the XRD patterns produced by spent catalysts obtained after methane decomposition over OV-rich α-Fe 2 O 3 catalysts at 800 • C, revealing that the previously observed XRD peaks in Figure 3 have disappeared and OV-rich α-Fe 2 O 3 catalysts have almost lost their original structure. It is worth noting that the spent catalysts are denoted as SFeN-500 • C, SFeN-700 • C, SFeN-900 • C, and SFeN-1100 • C. From Figure 10, it can be seen that the phase of OV-rich α-Fe 2 O 3 catalysts has changed to Fe 0 , Fe 3 C, Fe 2 O 3, and graphite. Two clear peaks can be seen at 2θ = 26 • and 2θ = 45 • suggesting the formation of graphite carbon and metallic iron, respectively.

Characterization of Spent Catalysts after the TCMD
Furthermore, the presence of iron carbide (Fe 3 C, albeit in small amounts) is confirmed by the peak situated in the 2θ = 38-50 • range, and it was suggested that Fe 3 C species have a catalytic activity for methane decomposition [44,45]. In addition, the appearance of very small peaks of Fe 2 O 3 phases indicates that the catalysts were not completely reduced due to the carbon deposited on iron, explaining the continuous TCMD up to 480 min as observed in Figure 9 [46]. Generally, it can be suggested that the existence of metallic Fe and Fe 3 C phases in the catalyst component plays a role in the activity of the catalyst [46]. Interestingly, the diffraction peaks in the SFeN-900 • C catalyst showed higher intensity compared to other spent catalysts, suggesting its improved crystallinity degree.
The morphological characteristics of the deposited nanocarbon in spent catalysts were examined by performing SEM analyses, and the images are shown in Figure 11. The SEM images of all the spent catalysts, as a result of TCMD, demonstrated different morphologies of deposited nanocarbon. A negligible amount of carbon nanofilaments in the catalysts was produced, along with multilayered graphene or graphite nanosheets encapsulating the catalyst. It was reported that the scarcity of the appearance of filamentous carbon in the catalysts is attributed to the fact that unsupported metal oxide catalysts are unable to form filamentous carbon during the TCMD. This is due to the catalyst particles undergoing rapid accumulation with each other and forming large particles as a result of harsh reaction conditions [47]. While the appearance of many graphene sheets may be attributed to the high diffusion coefficient of carbon atoms through metallic Fe at 800 • C, which may lower the formation of metal-encapsulated carbon nano-chunks and further expedite the formation of graphene sheets [44], furthermore, it has been observed that irregularly shaped carbon nano-chunks have appeared in abundance in the SFeN-700 • C and SFeN-900 • C samples.
thane decomposition over OV-rich α-Fe2O3 catalysts at 800 °C, revealing that the previously observed XRD peaks in Figure 3 have disappeared and OV-rich α-Fe2O3 catalysts have almost lost their original structure. It is worth noting that the spent catalysts are denoted as SFeN-500 °C, SFeN-700 °C, SFeN-900 °C, and SFeN-1100 °C. From Figure 10, it can be seen that the phase of OV-rich α-Fe2O3 catalysts has changed to Fe 0 , Fe3C, Fe2O3, and graphite. Two clear peaks can be seen at 2θ = 26° and 2θ = 45° suggesting the formation of graphite carbon and metallic iron, respectively.  The XPS spectra of the spent catalysts shown in Figures 12 and 13 indicate that all spent catalysts are rich in carbon. Moreover, a weak O 1s peak can be seen in the survey spectrum shown in Figure 12 while the Fe 2p peak is absent, suggesting that the α-Fe 2 O 3 reduction, as well as carbon deposition, highlights the successful TCMD, which confirms XRD and SEM findings.
The high-resolution C 1s XPS spectra of the spent catalysts shown in Figure 13 can be deconvoluted into five peaks located at 283.3 eV, 284.4 eV, 285.2-286.0 eV, 286.8-288.0 eV, and~291 eV, which can be indexed to carbide carbon (Fe-C), graphitic carbon (C-C, C-H), C-O bonds, C=O bonds, and O-C=O bonds, respectively, demonstrating the existence of the graphitic carbon [35,48]. Because Fe has a much higher atomic sensitivity factor compared to C and O, the absence of Fe peaks indicates that α-Fe 2 O 3 was uniformly coated by thin carbon sheets. Furthermore, the presence of Fe-C agrees with the XRD results of spent catalysts and implies the formation of Fe 3 C, which has catalytic activity for methane decomposition. Figure 14 illustrates the plot of the TGA-DTG curves for the spent catalysts after TCMD. It depicts the weight loss of the spent catalysts when exposed to high temperatures in the presence of air.
Irrespective of the spent catalyst type, there is an initial increase in weight up to 500 • C, which may be related to the oxidation of the metallic iron. As temperatures increase beyond 500 to 700 • C, the weight loss in spent catalysts is remarkably different due to the oxidation of carbon nanomaterials.
It is observed from the TGA curves that the weight loss takes place in the range of 500-700 • C due to the combustion of graphitic carbon [43], which increases from SFeN-500 to SFeN-900 • C spent catalysts, as indicated in Figure 14. The difference in weight loss among the spent catalysts reflects the catalytic activity of the fresh catalysts; low catalytic activity shows less weight loss, and vice versa. These results agree with the methane conversion and hydrogen formation rate results, which indicated that the OVrich α-Fe 2 O 3 NPs prepared at 900 • C have the highest methane conversion (22.5%) and hydrogen formation rate (47.0 × 10 −5 mol H 2 g −1 min −1 ) after 480 min. Thus, the OV-rich α-Fe 2 O 3 NPs prepared at 900 • C are the most active sample for methane decomposition, and this is due to their possession of abundant oxygen vacancies, which serve as effective active sites for adsorption methane. The XPS spectra of the spent catalysts shown in Figures 12 and 13 indicate that all spent catalysts are rich in carbon. Moreover, a weak O 1s peak can be seen in the survey spectrum shown in Figure 12 while the Fe 2p peak is absent, suggesting that the α-Fe2O3 reduction, as well as carbon deposition, highlights the successful TCMD, which confirms XRD and SEM findings. The high-resolution C 1s XPS spectra of the spent catalysts shown in Figure 13 can be deconvoluted into five peaks located at 283.3 eV, 284.4 eV, 285.2-286.0 eV, 286.8-288.0 eV, and ~291 eV, which can be indexed to carbide carbon (Fe-C), graphitic carbon (C-C, C-H), C-O bonds, C=O bonds, and O-C=O bonds, respectively, demonstrating the existence of the graphitic carbon [35,48]. Because Fe has a much higher atomic sensitivity factor compared to C and O, the absence of Fe peaks indicates that α-Fe2O3 was uniformly coated by thin carbon sheets. Furthermore, the presence of Fe-C agrees with the XRD results of spent catalysts and implies the formation of Fe3C, which has catalytic activity for methane decomposition.  Figure 14 illustrates the plot of the TGA-DTG curves for the spent catalysts after TCMD. It depicts the weight loss of the spent catalysts when exposed to high temperatures in the presence of air.  Irrespective of the spent catalyst type, there is an initial increase in weight up to 500 °C, which may be related to the oxidation of the metallic iron. As temperatures increase beyond 500 to 700 °C, the weight loss in spent catalysts is remarkably different due to the oxidation of carbon nanomaterials.
It is observed from the TGA curves that the weight loss takes place in the range of 500-700 °C due to the combustion of graphitic carbon [43], which increases from SFeN-500 to SFeN-900 °C spent catalysts, as indicated in Figure 14. The difference in weight loss among the spent catalysts reflects the catalytic activity of the fresh catalysts; low catalytic activity shows less weight loss, and vice versa. These results agree with the

Proposed Reaction Mechanism
Available evidence from previous studies indicates that the surface features of catalysts are the main drivers of catalytic methane decomposition. The XPS analysis performed as a part of this work revealed that the obtained OV-rich α-Fe 2 O 3 NPs at calcination temperatures such as 900 • C possess abundant oxygen vacancies, which would, in turn, increase the number of stable and active sites for the adsorption of methane. It has also been established that the introduction of OVs into the bulk α-Fe 2 O 3 material can weaken the Fe-O bond strength and improve the lattice oxygen mobility, thus catalyzing the reduction of iron oxide to form the active phase for methane decomposition. Therefore, it is reasonable to assume that methane adsorbs on the surface-active sites of the catalyst and reacts with the reducible species, e.g., free Fe 2 O 3 clusters (adjacent Fe 3+ ), surface oxygen species, and defects such as OVs, at the 800 • C reaction temperature. As a result, the C-H bonds on the catalyst surface would be gradually decomposed to generate H 2 and nanocarbon materials. Thus, FeN-900 • C, which contains abundant oxygen vacancies, could play an important role in the enhancement of catalytic activity in the thermocatalytic decomposition of methane into hydrogen and nanocarbon materials. A schematic diagram depicting the proposed reaction mechanism of the OV-rich α-Fe 2 O 3 NPs that are used for the thermocatalytic decomposition of methane is illustrated in Figure 15.
Available evidence from previous studies indicates that the surface features of catalysts are the main drivers of catalytic methane decomposition. The XPS analysis performed as a part of this work revealed that the obtained OV-rich α-Fe2O3 NPs at calcination temperatures such as 900 °C possess abundant oxygen vacancies, which would, in turn, increase the number of stable and active sites for the adsorption of methane. It has also been established that the introduction of OVs into the bulk α-Fe2O3 material can weaken the Fe-O bond strength and improve the lattice oxygen mobility, thus catalyzing the reduction of iron oxide to form the active phase for methane decomposition. Therefore, it is reasonable to assume that methane adsorbs on the surface-active sites of the catalyst and reacts with the reducible species, e.g., free Fe2O3 clusters (adjacent Fe 3+ ), surface oxygen species, and defects such as OVs, at the 800 °C reaction temperature. As a result, the C-H bonds on the catalyst surface would be gradually decomposed to generate H2 and nanocarbon materials. Thus, FeN-900 °C, which contains abundant oxygen vacancies, could play an important role in the enhancement of catalytic activity in the thermocatalytic decomposition of methane into hydrogen and nanocarbon materials. A schematic diagram depicting the proposed reaction mechanism of the OV-rich α-Fe2O3 NPs that are used for the thermocatalytic decomposition of methane is illustrated in Figure 15. Figure 15. A schematic diagram depicting the proposed reaction mechanism of TCMD by the unsupported OV-rich α-Fe2O3 NPs. Figure 15. A schematic diagram depicting the proposed reaction mechanism of TCMD by the unsupported OV-rich α-Fe 2 O 3 NPs.
Furthermore, based on the results reported in extant studies [9,18,19], the reduction of α-Fe 2 O 3 catalyst by CH 4 occurs in three steps, namely Fe 2 O 3 → Fe 3 O 4 → FeO → Fe 0 . Once Fe 0 is formed, it decomposes CH 4 , which results in the formation of Fe 3 C, a crucial step in the TCMD process for the formation of nanocarbon materials. The XRD, SEM, and XPS findings obtained in the current investigation indicate that the reduction of the OV-rich α-Fe 2 O 3 catalysts follows the global mechanism (Fe 2 O 3 → Fe 0 ), whereby Fe 3 C is formed as a result of the reaction of Fe 0 with the produced carbon from CH 4 decomposition. Therefore, it is clear that Fe 0 is an active phase for initializing the TCMD reaction at high temperatures (e.g., 800 • C), which suggests that the OV-rich α-Fe 2 O 3 catalysts possess an effective catalytic nature for TCMD for hydrogen and nanocarbon production.

Conclusions
In this study, a scalable and cost-effective one-step approach was developed for obtaining efficient unsupported OV-rich α-Fe 2 O 3 NPs with abundant oxygen vacancies on the surface via the thermal decomposition of ferric nitrate at different temperatures (500, 700, 900, and 1100 • C). The obtained OV-rich α-Fe 2 O 3 NPs at different calcination temperatures were employed as catalysts for the thermocatalytic decomposition of methane to produce H 2 and nanocarbon. Further investigations confirmed that the introduction of OVs on the asprepared OV-rich α-Fe 2 O 3 NPs affected the characteristics and thermocatalytic performance of these catalysts. Based on the XPS analysis, the OVs on the OV-rich α-Fe 2 O 3 surface were responsible for the increase in the number of active sites, thus aiding methane absorption. The morphological characterization results revealed that the most commonly obtained spent catalysts were nanosheets, which were the predominant carbon form deposited on the spent catalysts. According to the XRD results, the TCMD reduction follows the Fe 2 O 3 → Fe 0 global mechanism. By evaluating the thermocatalytic activity of the obtained OV-rich α-Fe 2 O 3 NPs in terms of TCMD for hydrogen production and nanocarbon formation, it was determined that 900 • C calcination temperature is optimal for producing highly efficient OV-rich α-Fe 2 O 3 NPs, which showed the highest methane conversion (22.5%) and hydrogen production rate (47.0 × 10 −5 mol H 2 g −1 min −1 ) after 480 min. Such outstanding thermocatalytic performance of OV-rich α-Fe 2 O 3 NPs is attributed to the presence of abundant OVs on their surfaces, which serve as effective methane adsorption sites.