The following chemicals were used as received from Sigma Aldrich: calcium nitrate tetrahydrate [Ca(NO₃)₂·4H₂O, 99%], chromium nitrate nonahydrate [Cr(NO₃)₃·9H₂O, 99%], manganese nitrate tetrahydrate [Mn(NO₃)₂·4H₂O, ≥ 97.0%], and nickel nitrate hexahydrate [Ni(NO₃)₂·6H₂O, 99.9%]. Acrylamide (AM, C₂H₃CONH₂, 99.9%), N,N’-methylenebisacrylamide (MBAM, C₂H₃CONHCH₂NHCOC₂H₃, 99.9%), ammonium persulfate [APS, (NH₄)₂S₂O₈, 99.9%], and N,N,N’,N’-Tetramethyl ethylenediamine (TEMED, 99.9%) were also used for gel-casting. Lanthanum nitrate hexahydrate [La(NO₃)₃·6H₂O, 99.9%] was used as received from Alfa Aesar. Milli-Q deionized water (18.2 MΩ·cm) was used during all experimental procedures.

The xNi-LCCM catalyst series was synthesized by a water-based gel-casting method adapted from a previous report^[24]. Hydrated metal nitrates were used as precursor salts instead of the oxide- and carbonate-based precursors used in earlier procedures. The catalysts were denoted according to the nominal molar amount of Ni introduced into the fixed LCCM composition; for example, 050Ni-LCCM has a nominal Ni:La:Ca:Cr:Mn molar ratio of 0.50:0.75:0.25:0.50:0.50. The nominal molar compositions are listed in Table 1.

A 50 wt% aqueous solution of each hydrated nitrate salt was prepared from La(NO₃)₃·6H₂O, Ca(NO₃)₂·4H₂O, Cr(NO₃)₃·9H₂O, Mn(NO₃)₂·4H₂O, and Ni(NO₃)₂·6H₂O. To prepare 050Ni-LCCM, the corresponding nitrate solutions were mixed and stirred for 30 min, followed by the addition of 16.54 g acrylamide (AM) and 1.10 g MBAM. After stirring for another 15 min, polymerization was initiated by sequentially adding 16.54 g of 2 wt% APS solution and 16.54 g of 2 wt% TEMED solution.

Figure 1 illustrates the nitrate-based aqueous gel-casting synthesis procedure for the xNi-LCCM catalysts. The nitrate-based aqueous route avoids organic solvents such as ethanol and isopropanol and offers a simple, scalable approach for preparing composition-tuned Ni-LCCM catalysts.

The molar ratios of metal elements in the catalysts were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, Avio 200, PerkinElmer, USA). Calibration curves were established for Ca, Cr, La, Mn, and Ni. Calibration concentrations ranged from 1-100 ppm for Ca, Cr, La, and Ni, while Mn was calibrated over 1-20 ppm. Ni was measured using the axial viewing mode, whereas all other elements were analyzed in radial viewing mode. All calibration curves showed good linearity (R² > 0.99). Phase structures of the catalysts and carbon products were analyzed by powder X-ray diffraction (XRD, Philips MPD 1880, PANalytical, Netherlands) using Cu Kα radiation (λ = 1.5406 Å) at room temperature, and the diffraction patterns were recorded over a 2θ range of 20°-80° at a scan rate of 4° min^-1. Surface chemical states were examined by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher Scientific, USA) with an Al Kα source (hν = 1,486.6 eV), operating at an accelerating voltage of 15 kV and a filament current of 10 mA. Each spectrum was accumulated over five scans to improve the signal-to-noise ratio. The pass energy was set to 30 eV for high-resolution spectra (step size: 0.05 eV), while 100 eV was used for survey spectra.

The morphology and microstructure of the catalysts and carbon products were characterized by field-emission scanning electron microscopy (FESEM, JEOL7600, JEOL, Japan), high-resolution transmission electron microscopy (HRTEM, Talos F200X, Thermo Fisher Scientific, USA), and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM, Talos F200X, Thermo Fisher Scientific, USA) equipped with energy-dispersive X-ray spectroscopy (EDS). The FESEM observations were performed at an accelerating voltage of 5 kV, while the HRTEM and HAADF-STEM analyses were conducted at 200 kV.

H₂ temperature-programmed reduction (H₂-TPR) was carried out using an Autochem II 2920 analyzer (Micromeritics, USA). The samples were pretreated at 200 °C for 0.5 h to remove moisture, followed by heating from 50 °C to 900 °C at a ramp rate of 10 °C min^-1 under a 10% H₂/90% Ar flow. Thermogravimetric analysis (TGA, Q500, TA Instruments, USA) was used to investigate the thermal behavior of the produced carbon nanomaterials. Approximately 10-15 mg of carbon or carbon-catalyst mixture was placed in the sample pan. The analysis was conducted at a heating rate of 10 °C min^-1 up to 900 °C under a mixed gas flow of 40 mL min^-1 N₂ and 60 mL min^-1 air. Prior to analysis, the pan was pre-cleaned by heating to 900 °C at 50 °C min^-1 to remove contaminants. Raman spectroscopy (inVia Raman microscope, Renishaw, UK) was used to evaluate carbon products. It was equipped with a 532 nm excitation laser, and the spectra were recorded over the range of 0-3,200 cm^-1. The obtained spectra were baseline-corrected and deconvoluted using Lorentzian functions to analyze the D, G, and 2D bands.

Catalytic decomposition of methane (CDM) experiments were carried out in a rotary-bed reactor. For each test, 2 g of catalyst was loaded at the center of a specially shaped quartz tube with an inner diameter of 60 mm, an expanded central section of 100 mm, and a total length of 1.2 m [Supplementary Figure 1]. The catalyst was used directly without prior reduction under H₂. Before the reaction, the system was evacuated with a vacuum pump to remove residual air. Methane (CH₄, 99.9%) was then introduced into the reactor at a flow rate of 100 sccm, corresponding to a space velocity of 3 L·g_cat.^-1·h^-1. The reactor (BTF-1200C-R-PECVD-AD, Anhui BEQ Equipment Technology, China) was rotated at 1 rpm during the reaction [Supplementary Figure 2]. The reactor was heated to the target temperature at a heating rate of 5 °C·min^-1. After the reaction, the CH₄ flow was stopped, and N₂ was introduced to terminate the reaction and prevent further carbon formation. The gaseous products were quantified using a gas chromatograph equipped with a thermal conductivity detector (TCD, Shimadzu Nexis GC-2030, Shimadzu, Japan). Methane conversion ($$ X_{C H_{4}}, \% $$, %) was calculated based on either the inlet and outlet methane amounts or the outlet hydrogen amount, assuming hydrogen and carbon as the only products. Carbon productivity was determined by normalizing the mass of deposited carbon to either the total catalyst mass (g_C·g_cat.^-1) or the Ni mass in the catalyst (g_C·g_Ni^-1), according to the following equations:

(1) $$ X_{C H_{4}}(\%)=\frac{n_{C H_{4}, i n}-n_{C H_{4}, o u t}}{n_{C H_{4}, i n}} \times 100 \% $$

(2) $$ X_{C H_{4}}(\%)=\frac{\frac{1}{2} n_{H_{2}, o u t}}{\frac{1}{2} n_{H_{2}, o u t}+n_{C H_{4}, o u t}} \times 100 \% $$

Where $$ n_{C H_{4}, { in }} $$, $$ n_{C H_{4}, { out }} $$, and $$ n_{H_{2}, { out }} $$ denote the amounts of methane at the inlet, methane at the outlet, and hydrogen at the outlet, respectively.

(3) $$ g_{C} \cdot g_{c a t .}{ }^{-1}=\frac{m_{t}-m_{c a t .}}{m_{c a t}} $$

(4) $$ g_{C} \cdot g_{N i}{ }^{-1}=\frac{m_{t}-m_{c a t .}}{m_{N i}} $$

Where m_t is the total mass of the carbon and catalyst mixture after reaction, m_cat. is the initial catalyst mass, and m_Ni is the nominal mass of Ni in the catalyst.

Reproducibility of the catalytic tests was evaluated using independent experiments under identical reaction conditions. For the Ni-loading-dependent catalytic tests, each catalyst composition was evaluated using three independent experiments. Unless otherwise stated, the reported values represent the mean of the independent measurements, with error bars indicating the corresponding standard deviation.

The as-produced carbon was purified by nitric acid treatment to remove residual metal species. Briefly, 10 mL of 5 M HNO₃ solution was added per gram of carbon product. The suspension was stirred and heated in an oil bath at 120 °C for 5 h. After acid treatment, the suspension was diluted with deionized water, mixed, and allowed to settle before solid-liquid separation. The acid-treated carbon was collected by filtration and dried at 120 °C for 24 h.

All calculations were performed based on DFT. The exchange-correlation functional was described using the generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE), together with the DFT-D3 dispersion correction^[25]. For geometry and lattice optimization, the plane-wave cutoff energy was set to 500 eV, and Brillouin zone integration was performed using a 4 × 4 × 1 Monkhorst-Pack k-point mesh^[26]. The convergence criteria for ionic relaxation and electronic self-consistent calculations were set to 0.05 eV Å^-1 and 10^-5 eV, respectively.

A 2 × 2 × 1 LCCM supercell exposing the (0-10) surface was constructed, with a 15 Å vacuum layer applied perpendicular to the surface. Different numbers of Ni atoms were introduced into the supercell to represent different relative Ni-loading levels and local Ni configurations. As summarized in Table 1, the yNi-LCCM models represent computational analogs of the experimental xNi-LCCM series, rather than the exact bulk stoichiometry of the catalysts. The pristine LCCM surface corresponds to y = 0. A pure Ni(111) slab with a comparable cell size was also constructed as a reference.

The adsorption energy was calculated as:

(5) $$ \Delta E= E_{\text {tot }} - E_{\text {sub }} - E_{\text {ad }} $$

where E_tot is the total energy of the adsorbate-substrate system, E_sub is the energy of the bare substrate, and E_ad is the energy of the isolated adsorbate.

The climbing-image nudged elastic band (CI-NEB) method was used to locate transition states^[27,28]. Ten images were used between the initial and final states for each elementary C-H bond cleavage step. For configurations where the initial and final states differed significantly in energy, forward and reverse activation barriers were calculated as:

(6) $$ \Delta E_{\mathrm{R}}= E_{\mathrm{TS}}- E_{\mathrm{I}} $$

(7) $$ \Delta E_{\mathrm{P}}= E_{\mathrm{TS}}- E_{\mathrm{F}} $$

Where E_TS, E_I, and E_F are the total energies of the transition state, initial state, and final state, respectively. Only the forward reaction was considered in this work; therefore, the activation energy was defined as:

(8) $$ E_{\mathrm{a}}= \Delta E_{\mathrm{R}} $$

ICP-OES analysis was used to verify the elemental compositions of the nitrate-derived xNi-LCCM catalysts. For the representative 050Ni-LCCM sample, the measured Ni:La:Ca:Cr:Mn molar ratio was close to the nominal composition, confirming that the nitrate-based gel-casting route effectively incorporated the metal precursors into the multicomponent catalyst system^[29]. The measured composition of 050Ni-LCCM is provided in the note to Table 1, and the ICP-OES-derived molar ratios for individual batches and the normalization basis are provided in the Supplementary Materials [Supplementary Table 1].

XRD patterns of the calcined xNi-LCCM catalysts confirm the formation of the LCCM perovskite phase across all compositions [Figure 2A]. The diffraction peaks are consistent with previously reported LCCM structures^[24], indicating that the nitrate-based gel-casting route successfully produces the targeted perovskite phase. In our previous work, xNi-LCCM catalysts were prepared using oxide and carbonate raw materials through a water-based gel-casting slurry route^[24]. By contrast, the present synthesis uses hydrated metal nitrates to form a homogeneous aqueous precursor solution before polymerization and self-combustion. Despite this change in precursor chemistry and gel-casting medium from slurry to solution, the final calcined catalysts exhibit the same LCCM perovskite framework, suggesting that the nitrate-derived route does not adversely affect phase formation. With increasing Ni incorporation, the perovskite reflections show a slight shift toward higher 2q values [Figure 2B], suggesting lattice modification associated with metal incorporation and defect formation^[19,29-31].

After reduction in H₂/N₂, metallic Ni reflections appear in the Ni-containing samples, while the LCCM perovskite structure remains intact [Figure 2C and D]. This confirms the reduction of Ni species^[24,32,33] and the structural stability of the LCCM under reducing conditions^[20]. The intensity of the metallic Ni reflections increases with Ni loading, consistent with the higher Ni content in the catalyst series.

XPS was employed to examine the surface chemical states of xNi-LCCM catalysts before and after reduction [Figure 3A-D]. Interpretation of the Ni 2p region is complicated by its overlap with La 3d signals^[34,35]. Nevertheless, constrained peak deconvolution can identify characteristic Ni features^[30,36,37]. After reduction, the component at ~852.3 eV is assigned to metallic Ni⁰, while the features at ~855.5 eV, together with the shake-up satellite at ~861 eV, indicate the presence of Ni²⁺ species in a NiO-like environment or in the near-surface region of the perovskite lattice^[38]. Contributions from La³⁺ are observed in the overlapping regions at ~850 eV and ~855-856 eV, consistent with La 3d emission^[39].

To further verify the Ni chemical states, the Ni 3p region was analyzed. The spectra show signatures of metallic Ni⁰ at ~66.3 eV and Ni²⁺ at ~67.6 eV, together with the corresponding shake-up satellite features [Figure 3C and D]^[40-42]. Owing to the spectral overlap between Ni 3p, Cr 3p, and Mn 3p signals, Cr- and Mn-related components were included in the deconvolution where required to obtain a reliable fit. The coexistence of Ni⁰ and Ni²⁺ after reduction indicates partial reduction of Ni species, consistent with Ni exsolution from the perovskite lattice and partial surface oxidation of the exsolved nanoparticles^[43,44]. Complete spectra for all compositions are provided in the Supplementary Materials [Supplementary Figures 3 and 4]

STEM-EDS mapping was further used to visualize the spatial redistribution of Ni in 050Ni-LCCM before and after reduction [Figure 3E and F]. In the calcined catalyst, the Ni signal is relatively homogeneous across the LCCM particles, indicating that Ni species are well dispersed within or on the oxide framework after the nitrate-based gel-casting and calcination process. After reduction in H₂/N₂, localized Ni-rich regions become more evident, suggesting the migration and enrichment of Ni species toward the particle surface. This observation is consistent with the XRD results, where metallic Ni reflections appear after reduction while the LCCM perovskite framework remains intact, confirming that Ni species are reduced without collapse of the parent oxide structure. The XPS analysis further supports this interpretation by showing the coexistence of metallic Ni⁰ and residual Ni²⁺ species after reduction, indicating partial reduction and possible surface oxidation of reduced Ni species.

Together, the XRD, XPS, and STEM-EDS results show that the nitrate-derived 050Ni-LCCM catalyst undergoes reduction-induced Ni redistribution while retaining the LCCM perovskite backbone. The emergence of Ni-rich surface regions provides microscopic evidence for reduction-induced Ni enrichment, consistent with the formation of surface-enriched/exsolved Ni species. These surface-enriched Ni species are expected to contribute to methane activation, while the stable LCCM framework helps maintain structural integrity under reducing conditions. This also confirms that replacing the previous oxide/carbonate-based slurry gel-casting route with a nitrate-based solution gel-casting route does not compromise the formation or redox stability of the LCCM-based catalyst.

HRTEM analysis of the 100Ni-LCCM catalyst was conducted as a representative case because of its clearly resolved interfacial features [Figure 4]. The low-magnification transmission electron microscopy (TEM) image shows aggregated catalyst particles [Figure 4A], while the enlarged TEM image reveals a Ni-containing region located at the LCCM matrix interface [Figure 4B]. Further magnification of the left and right dashed regions in Figure 4B shows the HRTEM images in Figure 4C and D, respectively. In these interfacial regions, lattice spacings of 0.211 nm are assigned to NiO(200), while spacings of 0.219-0.222 nm are attributed to the LCCM perovskite lattice. The coexistence of Ni-containing and LCCM lattice fringes within the same interfacial region indicates intimate contact between the Ni-containing phase and the oxide support.

This interfacial structure is consistent with the preceding XRD and XPS results. XRD shows that Ni-containing phases appear after reduction while the LCCM perovskite framework remains intact, and XPS reveals the coexistence of metallic Ni⁰ and Ni²⁺ species after reduction. The before-and-after STEM-EDS comparison in Figure 3E and F provides direct evidence for reduction-induced Ni redistribution and local Ni enrichment in 050Ni-LCCM. In contrast, Figure 4E-K shows the local elemental distribution in the selected interfacial region of 100Ni-LCCM. Together with the HRTEM observations in Figure 4A-D, these results support the formation of Ni-containing domains anchored to the LCCM matrix. The partially embedded morphology of these domains suggests a socketed particle-support interface, which is characteristic of exsolution-derived or strongly anchored nanoparticles rather than physically detached Ni particles.

Across the catalyst series [Supplementary Figures 5-8], lattice spacings of 0.218-0.222 nm and 0.276-0.278 nm are consistently assigned to the LCCM perovskite support, while spacings of 0.208-0.213 nm correspond to NiO(200). Additional spacings of 0.244-0.245 nm are associated with Ni-containing oxide domains, and in the 150Ni-LCCM sample [Supplementary Figure 8], a spacing of ~0.179 nm is consistent with metallic Ni(200), indicating a higher degree of Ni reduction at higher Ni loadings. Overall, the coexistence of Ni-containing and LCCM lattice fringes, together with the observed interfacial anchoring, supports strong metal-support integration in the xNi-LCCM catalysts.

H₂-TPR was conducted on 050Ni-LCCM and 150Ni-LCCM to further probe the reducibility of Ni species and their interaction with the LCCM framework [Figure 4L]. Both samples exhibit a main reduction peak centered at ~400 °C, together with a broad shoulder in the range of 450-550 °C. The low-temperature reduction peak becomes more intense with increasing Ni loading, indicating a larger amount of reducible NiO-like species that are more accessible to H₂. This assignment is consistent with the XRD results, where NiO reflections are observed after calcination and metallic Ni reflections appear after reduction.

The shoulder at higher temperature is attributed to Ni species that interact more strongly with the LCCM perovskite lattice or are partially incorporated near the oxide framework. Such species require higher temperatures for reduction than surface NiO-like species^[45,46]. This behavior is consistent with previous studies showing that Ni species with stronger interactions with the support or lattice incorporation exhibit higher reduction temperatures than surface Ni species^[47]. Comparable multi-step reduction features have also been reported for Ni-containing perovskites, where reduction peaks in the ~400-600 °C range are associated with the progressive reduction of lattice-incorporated Ni species^[48]. Together with the XPS results showing the coexistence of metallic Ni⁰ and residual Ni²⁺ after reduction, the before-and-after STEM-EDS comparison in Figure 3E and F, and the interfacial TEM/STEM-EDS observations in Figure 4A-K, the H₂-TPR profiles further support the presence of multiple Ni environments in Ni-LCCM catalysts, including accessible surface NiO-like species and more strongly bound lattice-associated Ni species. These different Ni environments are expected to contribute to the formation of surface-enriched/exsolved Ni sites during reduction and subsequent methane decomposition.

To correlate the Ni-loading-dependent increase in reducible Ni species observed by H₂-TPR with catalyst morphology, FESEM was used to examine the calcined xNi-LCCM catalysts after treatment in air at 1,100 °C for 5 h. The corresponding images and particle size distributions are provided in the Supplementary Materials [Supplementary Figure 9]. These morphology data provide additional context for understanding how Ni loading affects the distribution and accessibility of reducible Ni species.

DFT calculations were performed to examine elementary C-H bond-cleavage steps, adsorption behavior, and activation barriers (E_a) associated with methane decomposition. To rationalize the experimentally observed non-monotonic Ni-normalized carbon productivity, calculations were carried out on LCCM, Ni(111), and representative yNi-LCCM surface models with different relative Ni-loading levels and local Ni configurations. As shown in Figure 7A and B, Ni(111) and the LCCM(0-10) surface were first constructed as reference models, representing metallic Ni and the parent LCCM surface, respectively. These two reference structures provide the baseline for evaluating the effects of progressively introduced Ni species and their local configurations in the yNi-LCCM models. The LCCM surface model was constructed based on XRD refinement [Supplementary Figure 10], and the corresponding surface configuration is shown in Supplementary Figure 11. Ni atoms were introduced into the LCCM surface model to represent increasing relative Ni-loading levels, followed by structural optimization [Figure 7C-J]. As defined in Table 1, the yNi-LCCM models serve as representative computational analogs of the experimental xNi-LCCM series, rather than exact bulk stoichiometric replicas. The correlation between the experimental Ni-normalized carbon productivity and the DFT-derived rate-limiting barriers is summarized in Table 2. Ni(111) was included as a reference surface because it is a stable and commonly used model for metallic Ni in methane decomposition studies^[49].

The stepwise C-H bond cleavage pathway of methane was modeled through four elementary dehydrogenation steps from surface-bound CH₄ to surface carbon species. Transition-state calculations were used to determine the activation energy for each elementary C-H bond cleavage step. The rate-limiting step was defined as the elementary step with the highest activation barrier for each surface. The elementary C-H bond cleavage steps are listed in Supplementary Table 4, and the calculated activation energies are summarized in Supplementary Table 5. The rate-limiting barriers extracted from Supplementary Table 5 are compared with the experimental Ni-normalized carbon productivities in Table 2. A schematic illustration of the methane decomposition pathway and surface intermediates is shown in Supplementary Figure 12. Hydrogen recombination and desorption were treated in a simplified manner and were not used to define the rate-limiting C-H activation barriers.

Reaction energy profiles are shown in Figure 8, and the corresponding rate-limiting activation energies are summarized in Table 2 and Supplementary Table 5. Comparison across different Ni loadings reveals a non-monotonic variation in the rate-limiting activation energy. Among the yNi-LCCM models, 2Ni-LCCM and 6Ni-LCCM exhibit lower rate-limiting barriers than neighboring compositions, indicating more favorable C-H bond cleavage kinetics. These two models serve as representative computational analogs of the experimental 050Ni-LCCM and 150Ni-LCCM compositions, respectively. As summarized in Table 2, the lower calculated barriers for 2Ni-LCCM and 6Ni-LCCM are consistent with the experimentally observed local maxima in Ni-normalized carbon productivity. This comparison suggests that the non-monotonic carbon productivity is closely related to the local reaction energetics of composition-tuned Ni sites.

Further insight is obtained from the optimized surface configurations [Figure 7]. At lower Ni loadings, Ni atoms are primarily located in the surface layer of the LCCM model, whereas additional Ni atoms occupy second-layer configurations at higher loadings. In the present models, up to four Ni atoms can be accommodated near the surface before excess Ni preferentially occupies subsurface or second-layer positions. This transition alters the local coordination environment and Ni-Ni ensemble structure of the active sites.

The calculated low-barrier compositions, particularly 050Ni-LCCM and 150Ni-LCCM, correspond to favorable surface Ni configurations before extensive second-layer Ni occupation. With further increases in Ni loading, additional Ni atoms increasingly contribute to Ni-Ni ensemble or bulk-like behavior rather than improving the intrinsic activity of each Ni site. This provides a possible explanation for why increasing Ni content enhances overall methane conversion and catalyst-mass-normalized carbon yield but does not lead to a proportional increase in Ni-normalized carbon productivity.

Overall, the DFT results help rationalize the experimental observation that Ni utilization in xNi-LCCM is not controlled by total Ni content alone, but is strongly influenced by local Ni configuration, Ni reducibility, and site accessibility. This interpretation is consistent with the structural and surface analyses discussed above: XRD confirms the stability of the LCCM framework after reduction, XPS shows the coexistence of Ni⁰ and Ni²⁺ species, STEM-EDS reveals reduction-induced Ni redistribution, and HRTEM indicates interfacial anchoring of Ni-containing domains on the LCCM matrix. These findings suggest that composition-tuned, exsolution-related Ni configurations play an important role in determining methane decomposition activity and intrinsic carbon productivity.

Figure 9A presents the XRD patterns of the spent LCCM and xNi-LCCM catalysts after methane decomposition at 750 °C for 8 h. For the Ni-containing samples, diffraction peaks corresponding to metallic Ni are observed, which can be attributed to the in situ reduction of Ni species under the methane reaction environment. Additional diffraction features at 2θ ≈ 26.3-26.5° are assigned to graphitic carbon formed during CDM. In contrast, no diffraction peaks associated with either metallic Ni or graphitic carbon are observed for the Ni-free LCCM sample, indicating that the LCCM support remains structurally stable and shows negligible activity toward methane decomposition under the investigated conditions.

The carbon-catalyst mixture collected after methane decomposition was subsequently treated with 5 M HNO₃ to separate the carbon product from the catalyst. The purified carbon was analyzed by TGA to estimate the remaining inorganic content. As shown in Figure 9B and C, the mass loss below 200 °C is attributed to moisture removal, while the major mass loss between 200 and 800 °C corresponds to carbon oxidation. The residual mass at higher temperatures was assigned to inorganic residue. After acid treatment, the residual masses were 4.20 % and 1.41 % for the carbon products from 050Ni-LCCM and 150Ni-LCCM, respectively, corresponding to TGA-based carbon contents of approximately 95.8% and 98.6%. This indicates effective removal of most inorganic species from the carbon product. Notably, this purification was achieved using relatively mild nitric acid treatment, compared with the stronger acid conditions typically required for chemically stable oxide supports such as TiO₂ and SiO₂^[50].

Figure 10 shows representative TEM images of the carbon-catalyst mixture before and after acid treatment. Before acid washing, dark contrast regions are observed and are attributed to catalyst particles. After treatment with 5 M HNO₃, these regions are largely removed, indicating effective separation of the catalyst from the carbon product. The overall carbon morphology is largely retained after acid treatment, suggesting that the acid washing step mainly removes the inorganic catalyst component without causing obvious structural damage to the carbon framework. TEM analysis also shows that the carbon product contains mixed nanostructures, including carbon nanotubes, carbon nanofibers, and carbon nano-onions.

Raman spectroscopy was performed to evaluate the structural ordering of the carbon products before and after acid treatment. The spectra were baseline-corrected and deconvoluted using Lorentzian functions to analyze the D, G, and 2D bands. The D band at ~1,350 cm^-1 is associated with defect-induced vibrational modes, while the G band at ~1,580 cm^-1 corresponds to the in-plane stretching vibration of sp²-hybridized carbon^[38]. The D/G intensity ratio (I_D/I_G) was used to assess the structural ordering of the carbon products, with lower values indicating fewer defects and higher graphitic order.

The mean I_D/I_G values of the carbon products ranged from approximately 0.74 to 0.90 before acid washing and from approximately 0.62 to 0.75 after acid treatment [Figure 11]. The decrease in I_D/I_G after purification suggests improved apparent graphitic ordering, which is mainly attributed to the removal of disordered carbon species and residual catalyst-related impurities rather than direct structural transformation of the carbon framework. Together with the relatively high oxidation temperature observed in TGA and the tubular carbon morphologies observed by TEM, the well-defined D, G, and 2D bands indicate that the carbon products are predominantly nanocrystalline graphitic carbon rather than amorphous carbon.