Low-Temperature CH4 Reforming and Water Splitting with Activated NiO/CeO2 as Oxygen Carrier
Chunli Han, Akira Yoko, Yi-Ping Chang, Manuel Harder, Kakeru Ninomiya, Maiko Nishibori, Zhong Yin, Ardiansyah Taufik, Satoshi Ohara, Tadafumi Adschiri
TL;DR
This study presents a new method for efficiently producing hydrogen and syngas at low temperatures using a modified NiO/CeO2 catalyst.
Contribution
A novel NiO/cCeO2 oxygen carrier enables low-temperature CH4 reforming and water splitting with high selectivity and efficiency.
Findings
Low-temperature CH4 activation and high syngas selectivity (>98.5%) were achieved using activated NiO/cCeO2.
Nearly pure H2 was produced during the water splitting step, eliminating the need for additional purification.
Precise control over Ni site size and structural evolution improved the catalyst's performance and stability.
Abstract
Low-temperature (≤600 °C) CH4 activation and high syngas selectivity (>98.5%, H2/CO2) were simultaneously achieved using the activated NiO/cCeO2 oxygen carrier. Nearly pure H2 was produced during the water splitting step.Synergistic advantages of low operating temperature and high selectivity significantly enhance the energy efficiency of chemical looping CH4 reforming and water splitting process.Precise control over the size and density of Ni sites and activation and structural evolution of NiO/cCeO2 were systematically investigated. Low-temperature (≤600 °C) CH4 activation and high syngas selectivity (>98.5%, H2/CO2) were simultaneously achieved using the activated NiO/cCeO2 oxygen carrier. Nearly pure H2 was produced during the water splitting step. Synergistic advantages of low operating temperature and high selectivity significantly enhance the energy efficiency of chemical…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsChemical Looping and Thermochemical Processes · Catalysts for Methane Reforming · Advancements in Solid Oxide Fuel Cells
Introduction
Syngas and pure hydrogen are essential raw chemicals for various industrial processes, including oil refining, fuel cells, energy carriers, and the synthesis of various chemicals [1–5]. Steam methane reforming (SMR), the most widely used process for syngas production, is typically performed at approximately 900 °C to enhance the CH_4_ conversion rate and reaction selectivity, thereby reducing energy consumption and greenhouse gas emissions per unit of product [6, 7]. In contrast, the chemical looping partial oxidation of methane and water splitting (CL POM-WS) process can break the equilibrium limitation of the SMR process by utilizing solid oxygen carriers (OCs), offering a promising alternative for low-temperature operation [8].
As shown in Fig. 1, the CL process consists of two steps: (i) POM step (fuel reactor): CH_4_ reacts with OC to convert into syngas (H_2_/CO = 2). (ii) WS step (steam reactor): H_2_O(g) dissociates on the reduced OC, producing pure H_2_ and replenishing the lattice oxygen of OC. The CL POM-WS process avoids direct contact between CH_4_ and O_2_, eliminates the need for an air separation unit, and significantly reduces the costs associated with downstream gas separation and purification. Therefore, the CL POM-WS process is considered as a safer, cleaner, energy-efficient, and economical process for the coproduction of syngas and pure hydrogen [9–12]. The major side reactions include CH_4_ cracking (CH_4_ → C + 2H_2_) and the complete oxidation of CH_4_ to CO_2_. When the carbon by-product is transferred to the steam reactor, CO and CO_2_ are generated, compromising hydrogen purity during the WS step.Fig. 1. Schematic of chemical looping partial oxidation of methane and water splitting (CL POM-WS) process
The core of the CL POM-WS process lies in the OCs, which directly determine the operating temperature and reaction selectivity. According to an integrated process energy evaluation [6], developing efficient medium-/low-temperature (500–600 °C) OCs is crucial for the practical implementation of the CL POM-WS process. Extensive efforts have been devoted to the development of advanced OCs, including Ni-, Fe-, Cu-, Co-, Pt-, and Mn-based metal oxides, as well as perovskite oxides [12–18]. Some high-activity and coke-resistant OCs for the CL POM process have been reported [19–24]. However, the simultaneous realization of low-temperature operation and high syngas selectivity remains a significant challenge. For most reported non-precious-metal OCs, the onset and optimal operating temperatures are above 700 °C due to the high C-H bond strength of CH_4_ (434 kJ mol^−1^).
Precise manipulation of OC or catalyst microstructures has been shown to induce significant differences in reaction performance [25–30] and is generally more effective than rough compositional adjustment in achieving the desired reactivity. Among various OCs, Ni-based materials show high activity for CH_4_ dissociation but suffer from severe coke deposition and rapid OC deactivation [31–33]. For example, Guerrero-Caballero et al. reported that 50% of CH_4_ was converted to carbon over a Ni/CeO_2_ OC during CL dry reforming at 600 °C, yielding an H_2_/CO ratio of ~ 5 [34]. Zhang et al. reported that a Ni-CeO_2_/Al_2_O_3_ OC showed a syngas selectivity of ~ 48% with an H_2_/CO ratio of 4.1 at 500 °C [24]. Fundamentally, the key challenge lies in achieving a balance between the CH_4_ dissociation rate (CH_4_ → CH_3_^^ → CH_2_^^ → CH^*^ → C) and the supply rate of lattice oxygen, which significantly affects the reaction pathway and products [27, 35]. Balancing these factors is particularly intricate in composite metal oxide OCs, because both the reactivity of active sites and the lattice oxygen migration rate of the oxide support vary dynamically with the reaction temperature and the loading state of metal active sites.
In this study, highly active Ni for CH_4_ activation and cubic CeO_2_ nanoparticles (cCeO_2_ NPs) was used to construct a low-temperature OC through the precise microstructural regulation strategy. The OC microstructures were systematically manipulated by the Ni loading concentration control, activation pretreatment, and support effect. The resulting OC, surface-fused NiO NPs on cCeO_2_, was denoted as NiO/cCeO_2_. The reaction-driven activation pretreatment for boosting low-temperature activity and stability was highlighted, and the electronic structure of Ni active sites, dynamic structural evolution, the metal**–**support interaction, as well as the active sites for water splitting was studied in detail. The reaction performance of the NiO/cCeO_2_ OC in the 500–800 °C range and the OC long-term stability at 550 and 600 °C was evaluated. This study developed a highly efficient and robust low-temperature OC for CL POM-WS and proposed a refined design and regulation strategy for high-performance OCs.
Experimental Section
Materials
Nickel (II) nitrate hexahydrate (Ni(NO_3_)2·6H_2_O, 98.0%+) and ethanol (C_2_H_5_OH, 99.5%) were purchased from FUJIFILM Wako Pure Chemical Corporation (Japan). Nickel (II) oxide nanopowder (NiO, < 50 nm, 99.8%) and cerium (IV) oxide nanopowder (< 25 nm, sCeO_2_) were purchased from Sigma-Aldrich Co. (USA). cCeO_2_ NPs (average particle size: 7 nm) and ZrO_2_ NPs (average particle size: 9 nm) were prepared using continuous-flow supercritical hydrothermal method by Super Nano Design Co., Ltd. and ITEC Co., Ltd., respectively, according to our previous studies [36–38] (see Fig. S1 for transmission electron microscopy (TEM) images of cCeO_2_, sCeO_2_, and ZrO_2_ NPs). All chemicals were used as received without any further purification. Deionized water was used for all the experiments.
Synthesis of NiO/cCeO2 OCs
A surface fusion method was used for the synthesis of NiO/cCeO_2_ OCs. cCeO_2_ NPs and Ni(NO_3_)2_6H_2_O were mixed and milled uniformly. The mixture was then heated to 400 °C at a rate of 6 °C min^−1^ in air in a muffle furnace and held at 400 °C for 2 h to obtain the surface-fused xNiO/cCeO_2 OCs (x represents the molar ratio of Ni: nNi/[nNi + nCe]). For comparison, commercial CeO_2_ NPs were used as the support to synthesize NiO/sCeO_2_ OC.
Activation and Reactivity Evaluation
The activation and reactivity of OC NPs were evaluated in a continuous-flow catalyst analyzer (BELCAT II, Microtrac Inc., USA) at atmospheric pressure (Fig. S2). The fixed-bed quartz tube reactor has an inner diameter of 13 mm and a length of 160 mm. The OC NPs (approximately 30 mg) were loaded at the bottom of the reactor. The reactor was heated at a rate of 10 °C min^−1^. Before activation, the fresh OC NPs were pretreated in 3.8% O_2_/Ar with a flow rate of 52 mL min^−1^ at 400 °C for 15 min to remove the carbonaceous contaminants. The reactor was then heated to the OC activation temperature of 700 °C at a rate of 10 °C min^−1^. For activation, pre-reactions involving CL POM-WS were performed isothermally for five cycles: (i) 2% CH_4_/Ar flowed through the reactor for 20 min at a flow rate of 100 mL min^−1^, (ii) Ar flowed at 100 mL min^−1^ for 10 min to purge the line, (iii) 46.7% steam/Ar was introduced into the reactor at a total flow rate of 100 mL min^−1^ for 10 min, and (iv) Ar flowed at 100 mL min^−1^ for 10 min to purge the line. The steam was generated in a bubbler at 80 °C, and the ratio of steam in the mixed gas was calculated according to the saturated vapor pressure curve (Fig. S3) [39]. All gas lines were insulated in an air bath at 110 °C.
The reactivity of the activated OCs was evaluated with the above-mentioned procedure in the range of 500–800 °C. The gaseous products were analyzed using an online quadrupole mass spectrometer (QMS). The MS calibration curves for CH_4_ (m/z = 15), CO (m/z = 28), H_2_ (m/z = 2), and CO_2_ (m/z = 44) were measured for quantitative analysis (Fig. S4). Figure S5 shows the MS background curves obtained under the same temperature program as the CL POM-WS cycles, without loading OCs.
The long-term stability of OCs was tested over 30–40 cycles using similar procedure, with slight adjustments to the time for each step (POM: 10 min → Purge: 10 min → WS: 5 min → Purge: 5 min).
The CO selectivity in the POM step is defined as:
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${S}_{\mathrm{CO}}=\frac{{n}_{\mathrm{CO}}}{{n}_{\mathrm{CO}}+{n}_{{\mathrm{CO}}_{2}}}\times 100\%$$\end{document}The H_2_ purity in the WS step is defined as:
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${P}_{{\mathrm{H}}_{2}}=\frac{{n}_{{\mathrm{H}}_{2}}}{{{n}_{{\mathrm{H}}_{2}}+n}_{\mathrm{CO}}+{n}_{{\mathrm{CO}}_{2}}}\times 100\%$$\end{document}The oxygen recovery rate of OCs in the WS step is defined as:
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{\mathrm{O}}=\frac{{n}_{{\mathrm{H}}_{2}\mathrm{-WS}}-{n}_{\mathrm{CO-WS}}-{2n}_{{\mathrm{CO}}_{2}\mathrm{-WS}}}{{n}_{\mathrm{CO-POM}}+{4n}_{{\mathrm{CO}}_{2}\mathrm{-POM}}}\times 100\%$$\end{document}Characterization
The morphology of the OC NPs and Ni dispersion states was examined using high-resolution TEM (HRTEM, JEOL JEM-ARM200F, Japan), operated at an accelerating voltage of 200 kV, and equipped with energy-dispersive spectroscopy. The samples were prepared by dropping an ethanol solution of OC NPs onto the carbon-coated copper grids three times. The diffraction patterns of OC NPs were collected using X-ray powder diffraction (XRD, Rigaku SmartLab 9MTP, Japan) with a Cu Kα radiation source (λ = 1.5418 Å) at 3° min^−1^ at a 2θ scanning range of 10°–100°. The local structures of the fresh, activated, and cycled OCs were further characterized using a microscopic confocal Raman spectroscopy (HORIBA LabRAM HR-800, Japan). Before measurements, silicon wafer was used as the reference to calibrate the Raman shift. Raman spectra were recorded in the range of 160–1800 cm^−1^ at an excitation wavelength of 532 nm and an acquisition time of 15 s. All the Raman peaks were normalized to the intensity of the F_2g_ mode (100%). X-ray photoelectron spectroscopy (XPS) data of the OCs and NiO reference sample were recorded using a scanning X-ray photoelectron spectrometer (PHI5000 VersaProbe II, Japan) with a pass energy of 58.7 eV and an emission angle of 45°. Electron neutralization was applied to avoid the charge up. The binding energy was calibrated with Ce 3d_3/2_4f^0^ peak at 916.7 eV as the reference [40, 41]. The peak area ratio of Ni 2p3/2 to Ce 3d was calculated to determine the relative Ni concentration on the OC surface. X-ray absorption spectra (XAS) of Ni L3,2 were measured at the SPECIES beamline of MAX IV Laboratory in total fluorescence yield mode with an angle of 45° and the exit slit set at 50 μm corresponding to an energy resolution of around 200 meV. The powder samples were pelletized for the XAS measurements. Energy calibration was conducted with a literature value [37], and the Ni L3,2 absorption edges were adjusted according to the literature value of NiO [42]. The specific surface area of OCs was measured via nitrogen adsorption–desorption measurements using QUADRASORB EVO4 (Quantachrome Instruments, USA). Before measurements, the samples were degassed at 140 °C for 6 h. The specific surface area was calculated using the multi-point Brunauer–Emmett–Teller method.
The oxygen storage capacity (OSC) of the activated OCs at 600 °C was measured using the catalyst analyzer (BELCAT II, Microtrac Inc., USA) by the O_2_-pulse method. After activation program, the reactor temperature was reduced to 600 °C for OSC measurement. Before each measurement, the samples were pretreated with 10% H_2_/Ar at 50 mL min^−1^ for 120 min at 600 °C. The reactor was then purged with Ar at 50 mL min^−1^ for 15 min. Subsequently, seven O_2_ pulses were introduced into the reactor, and the effluents were detected by a thermal conductivity detector (TCD). The OSC value was calculated from the first four O_2_ pulses, with the last three O_2_ pulses as the baseline.
The H_2_ temperature-programmed reduction (H_2_-TPR) of the activated OC NPs was performed in the same catalyst analyzer. The samples were first pretreated in a flow of 20% O_2_/Ar (50 mL min^−1^) at 400 °C for 30 min. Then, the reactor was cooled to 40 °C and purged with Ar at 50 mL min^−1^ for 15 min. Subsequently, 10% H_2_/Ar (50 mL min^−1^) was introduced into the reactor while ramping the reactor temperature to 900 °C at a rate of 10 °C min^−1^. The effluents were detected by TCD.
The CH_4_/H_2_O(g)-temperature-programmed reaction (CH_4_/H_2_O(g)-TPR) was performed in the same catalyst analyzer. The fresh OCs were activated, as described in Sect. 2.3. Subsequently, the reactor was cooled to 40 °C in an Ar flow (100 mL min^−1^) and then reheated to 800 °C at a rate of 10 °C min^−1^ in a 2% CH_4_/Ar flow (100 mL min^−1^), followed by cooling to 100 °C at a rate of 10 °C min^−1^ in an Ar flow (100 mL min^−1^). The flow was then changed to 46.7% steam/Ar (100 mL min^−1^), and the temperature gradually increased to 500 °C (10 °C min^−1^) to record the H_2_O(g)-TPR curve. To explore the effect of air on the activated OCs, the OC NPs were stored in air after activation and then reloaded into the catalyst analyzer for CH_4_/H_2_O(g)-TPR measurement.
Results and Discussion
Distinct CH4 Conversion Pathways on NiO/cCeO2
Initiating the CL POM-WS cycle at medium/low temperatures (≤ 600 °C) is challenging. To address this issue, highly active Ni species were deposited on the cCeO_2_ surface to enhance the low-temperature CH_4_ dissociation activity. A series of NiO/cCeO_2_ NPs was prepared and tested to determine the optimal Ni loading mode and amount on cCeO_2_. The amount of Ni loading significantly affected the CH_4_ conversion pathways.
For Ni loading amounts of 2.1 and 6.5 mol%, the onset temperature for CH_4_ dissociation was 600 °C, but the products in the fuel and steam reactors differed completely (Fig. 2). At 2.1NiO/cCeO_2_ OC, the primary product was syngas (H_2_:CO = 2), which was generated through the partial oxidation of CH_4_ (δ CH_4_ + MeO_x_ → δ (CO + 2H_2_) + MeO_x−δ_). In addition, pure H_2_ was produced (MeO_x-δ_ + δ H_2_O → MeO_x_ + δ H_2_) during the WS step. However, for 6.5NiO/cCeO_2_ OC, the primary reaction shifted to CH_4_ cracking (CH_4_ → C + 2H_2_) after the rapid reduction of NiO to Ni (CH_4_ + 4NiO → 4Ni + CO_2_ + 2H_2_O; CH_4_ + NiO → Ni + CO + 2H_2_). Multiwalled carbon nanotubes were formed, as evidenced by the TEM images and Raman spectrum (Fig. S6). Consequently, high concentration of CO_2_ was generated in the steam reactor (C + 2H_2_O → CO_2_ + 2H_2_). The loaded Ni species govern the CH_4_ dissociation rate. Different Ni loading amounts create Ni/NiO agglomerates of various sizes, leading to different coordination environments and electronic structures of the Ni sites (as detailed in Figs. 5 and 7). These variations result in distinct reaction pathways. The OC structure enabling the highly selective and low-temperature conversion of CH_4_ to syngas is discussed in detail in Sect. 3.4.Fig. 2. Distinct reaction pathways of CH_4_ on NiO/cCeO_2_ with varying Ni loading amounts. a Ni: 2.1 mol%, main reaction: CH_4_ to syngas, b Ni: 6.5 mol%, main reaction: CH_4_ cracking. Dark red, orange, purple, green, and blue lines refer to CH_4_, CO_2_, CO, H_2_, and furnace temperature, respectively. The intermittent sharp noise peaks represented by purple lines are primarily caused by trace amounts of residual N_2_ in the bubbler, as evidenced by the MS background curve shown in Fig. S5
Activation of NiO/cCeO2 OCs
Activation and pretreatment are essential for catalysts and OCs [43], yet detailed reports on the variations in the particle structure and activity are limited. In this study, the effect of the activation step on the reactivity of OCs was investigated in detail.
Initially, the cyclic performance of the 2.1NiO/cCeO_2_ OC at 600 °C without activation exhibited poor stability (Fig. 3a). Increasing the temperature to 700 °C induced structural evolution from the first to the second cycle (Fig. 3b). Concurrently, the reaction kinetics accelerated, the selectivity of CO slightly increased, and the H_2_:CO molar ratio approached 2 (Fig. 3e).Fig. 3. Vital role of activation treatment in realizing stable low-temperature reactivity of the NiO/cCeO_2_ OC. a Cyclic reaction performance of 2.1NiO/cCeO_2_ OC at 600 °C before activation. b Reactivity evolution of 2.1NiO/cCeO_2_ OC during activation at 700 °C. c Cyclic reaction performance of 2.1NiO/cCeO_2_ OC at 600 °C after activation. d CH_4_-TPR curve of the activated 2.1NiO/cCeO_2_ OC without contact with air after activation. e Reaction performance and gaseous product composition of the activated 2.1NiO/cCeO_2_ OC at different temperatures
Surprisingly, the 2.1NiO/cCeO_2_ OC demonstrated stable reactivity at 600 °C after being pretreated at 700 °C (Fig. 3c). This pretreatment process is referred to as the activation of the OCs. The CH_4_-TPR result (Fig. 3d) shows that the onset temperature for CH_4_ dissociation on the activated 2.1NiO/cCeO_2_ OC was 586 °C, with the maximum syngas generation rate (0.21 mL s^−1^ g_OC_^−1^) observed at 636 °C. The production of H_2_, CO, and CO_2_ at 586 °C likely resulted from the reactions between CH_4_ and surface-active oxygen species (CH_4_ + O_S_ → CO + 2H_2_; CH_4_ + 4O_S_ → CO_2_ + 2H_2_O) or the reaction between CH_4_ and NiO. The activated OCs exhibited superior low-temperature activity and a broad operating window. Across the 600–800 °C range, the reaction pathway remained constant (Fig. S7). Syngas with a H_2_:CO molar ratio near 2 was the primary product during the POM step, and high-purity H_2_ was produced in the WS step (Fig. 3e). As shown in Table 1, the reduction degree of the CeO_2_ support increased linearly with an increase in the reaction temperature. The oxygen recovery rates of the reduced OCs in the WS step were higher than 75%. Preliminary cyclic tests indicated the structural stability of the activated OCs. Therefore, a five-cycle CL POM-WS activation procedure at 700 °C was used to pretreat all samples in this study. It should be noted that not all samples require five activation cycles. It depends on the mass transfer between the OCs and the activating gases. In practice, the optimal number of activation cycles can be determined by monitoring the gas evolution profiles. Then, the activation inducing factors were studied in detail, including the roles of temperature and activating gases. First, a control sample was prepared by calcination at 700 °C in air for 4 h. As shown in Fig. S8a, this sample exhibited unstable CL POM-WS performance at 600 °C, indicating that merely increasing the calcination temperature and duration cannot effectively activate the OC particles. This result demonstrates that the activation pretreatment involves not only thermal effects. Then, H_2_ was employed to substitute CH_4_ as the reducing gas. The OC was reduced by H_2_ at 700 °C for 3 h, followed by the WS reaction (Fig. S8b). However, the reaction curve was similar to that of the first POM-WS cycle shown in Fig. 3b, as reflected by the CO evolution curve, and the key structural evolution still occurred between the first and second POM-WS cycles. It indicates that H_2_ reduction pretreatment at 700 °C also cannot induce activation. Subsequently, O_2_ was used to substitute steam for reoxidation. As shown in Fig. S8c, d, the OC activity decreased after O_2_ reoxidation. During the POM-WS cycles, the H_2_ generation amount in the POM step at 700 °C was approximately 1.3 times higher than that of POM-O_2_ cycles (Fig. S8e). These results clearly demonstrate that the proposed activation pretreatment is a reaction-driven activation process, in which both the activation temperature and the activating gases play crucial roles. The key structural reconstruction during activation is discussed further in Sect. 3.4.Table 1. Reduction degree and oxygen recovery rate of the activated 2.1NiO/cCeO_2_ OC in the POM and WS stepsT (°C)H_2_ generation amount-POM (mmol g_OC_^−1^)H_2_ generation amount-WS (mmol g_OC_^−1^)H_2_-POM/H_2_-WSReduction degree of CeO_2_-POM (%)Oxygen recovery rate-WS (%)6001.900.862.227.690.26202.040.832.468.281.46502.621.022.5710.777.77003.181.202.6413.275.88004.291.692.5417.978.6
Fine Adjustment of the Density of Ni Active Sites
Next, the local structure and density of the Ni active sites were finely adjusted around 2.1 mol% to maximize the activity of NiO/cCeO_2_ OC. The activation processes of NiO/cCeO_2_ OCs with varying Ni loading amounts are shown in Fig. S9, exhibiting similar structural evolution to that of the 2.1NiO/cCeO_2_ OC. As shown in Fig. 4a, b, the amount of H_2_ generated in the POM and WS steps significantly increased with an increase in the Ni loading amount from 1.1 to 2.5 mol%. However, a further increase in the Ni concentration did not improve the OC activity substantially. With Ni loading amounts of 3.0 and 3.5 mol%, the H_2_ purity (WS step) at 600 °C slightly decreased (Fig. 4d), indicating the onset of coke deposition. The OSC value of the activated 2.5NiO/cCeO_2_ OC at 600 °C was 175.3 μmol-O g^−1^. The optimal Ni loading of 2.5 mol% well balanced CH_4_ dissociation and oxygen migration rates. The specific surface areas of the fresh and activated 2.5NiO/cCeO_2_ OC were 38.0 and 3.5 m^2^ g^−1^, respectively, and the Ni loading density on the activated OC was approximately 25 Ni atoms nm^−2^.Fig. 4. Fine optimization of Ni loading amount on NiO/cCeO_2_ OCs. a, b H_2_ generation amount in the POM and WS steps with varying Ni loading amounts. 2.5 mol% Ni-sCeO_2_ represents that commercial CeO_2_ NPs are used as support. c, d Gaseous product quality with varying Ni loading amounts. e Variations of reduction degree of the CeO_2_ support with reaction temperature in the POM step. f CH_4_ cracking ratios within the total CH_4_ conversion during the POM step with varying Ni loading amounts. g, h Long-term stability of 2.5NiO/cCeO_2_ OC at 600 °C. Note that reaction time becomes half during the long-term test (POM: 10 min, WS: 5 min). i Comparison of the reaction performances of 1 wt%Pt/cCeO_2_ and 2.5NiO/cCeO_2_ OCs
The reduction degrees of the 2.5NiO/cCeO_2_ OC at 600, 700, and 800 °C were 8.9%, 15.9%, and 21.0%, respectively, which were higher than those of 2.1NiO/cCeO_2_ OC, as shown in Fig. 4e. The Raman spectra of fresh and activated 2.5NiO/cCeO_2_ OCs are shown in Fig. S10a. Carbon-related Raman peaks were not observed, indicating the absence of coke deposition. Figure S10b, c shows the CH_4_/H_2_O(g)-TPR curves of the 2.5NiO/cCeO_2_ OC with and without atmospheric exposure after activation. The onset and peak temperatures of the POM and WS reactions were consistent. Consequently, no significant change was observed in the reaction performance of the OC NPs after contact with air. Therefore, ex situ characterizations (HRTEM, XRD, Raman, and XPS) of the OC structure were conducted without special vacuum protection. The H_2_ peak position during the WS reaction appeared at 145 °C on the reduced 2.5NiO/cCeO_2_ OC, and the operating temperature of the CL POM-WS process was determined by the CH_4_ activation temperature. Notably, some discrepancies remained between the onset temperatures for CH_4_ dissociation observed during the CH_4_-TPR test (579 °C) and those in the actual reaction process (approximately 500 °C, as shown in Fig. S11 and Table S1). This difference primarily stemmed from minor structural variations that occurred during cooling between the activation pretreatment and CH_4_-TPR measurement.
Additionally, with 2.5NiO/sCeO_2_ as the OC, the H_2_ purity in the WS step at 600 °C was only 84.3%. The coke deposition amount and CH_4_ cracking ratio during the POM step were quantitatively estimated based on the CO and CO_2_ generation observed in the subsequent WS step. As shown in Fig. S12, at 600 °C, the coke deposition on 6.5NiO/cCeO_2_ and 2.5NiO/sCeO_2_ OCs was approximately 645 and 29 times higher than that on 2.5NiO/cCeO_2_ OC, respectively. Correspondingly, the CH_4_ cracking ratios within the total CH_4_ conversion at 600 °C, i.e., the selectivity of coke formation reaction, were 0.6% for 2.5NiO/cCeO_2_, 15.5% for 2.5NiO/sCeO_2_, and 81.2% for 6.5NiO/cCeO_2_ (Fig. 4f). These results further demonstrate the superior coking resistance of the 2.5NiO/cCeO_2_ OC at low temperatures. Detailed discussions of the structural differences between these OCs are provided in Sect. 3.4.
The long-term stability of the 2.5NiO/cCeO_2_ OC was evaluated at 600 °C with POM and WS reaction times of 10 and 5 min, respectively. As shown in Fig. 4g, after the initial stabilization, the amount of H_2_ generated and the product quality in the POM and WS steps remained constant during 40 cycles (time on stream: approximately 20 h). The average CO selectivity and H_2_:CO molar ratio in the POM step were 96.5% and 2.1, respectively (Fig. 4h). During the cyclic tests, the WS step stably yielded nearly pure H_2_. At 550 °C, 2.5NiO/cCeO_2_ OC also demonstrated remarkable long-term stability (Fig. S13). The specific surface area of 2.5NiO/cCeO_2_ OC after 40 cycles at 600 °C was 7.3 m^2^ g^−1^. Figure S14 shows the Raman spectra of 2.5NiO/cCeO_2_ OC after long-term tests at 550 and 600 °C. No carbon-related peaks were found, indicating strong coke resistance of the 2.5NiO/cCeO_2_ OC. The Raman characteristic peak position (F_2g_) of CeO_2_ did not shift, demonstrating the integrity of the CeO_2_ lattice structure.
Furthermore, the role of the CeO_2_ support was investigated by substituting it with ZrO_2_. With the ZrO_2_ NPs as the support, the Ni loading amount varied from 1.8 to 4.0 mol%. Figure S15 compares the reaction performances of NiO/cCeO_2_ and NiO/ZrO_2_. Changing the support significantly affected the activity of Ni sites and the CH_4_ conversion pathways. One mole of CH_4_ can produce two moles of H_2_ via the POM pathway (δ CH_4_ + MeO_x_ → δ (CO + 2H_2_) + MeO_x−δ_) or CH_4_ cracking (CH_4_ → C + 2H_2_). With the same Ni loading amount, the H_2_ generated in the CH_4_ conversion step on NiO/ZrO_2_ was substantially lower than that generated with NiO/cCeO_2_, indicating that the absolute activity of Ni sites for CH_4_ activation decreased substantially. The intrinsic property differences between CeO_2_ and ZrO_2_ lead to different electron transfer behaviors between Ni and the respective supports, which significantly influence the activity of the Ni sites. With ZrO_2_ as the support, the reaction proceeded mainly through CH_4_ cracking (Fig. S15b) due to the poor redox property of ZrO_2_. According to the above analysis, the cCeO_2_ support not only provides lattice oxygen but also modulates the electronic structure and activity of the Ni sites.
Table S2 compares the reaction performance of the 2.5NiO/cCeO_2_ OC with that of several representative OCs for the CL methane reforming. The 2.5NiO/cCeO_2_ OC exhibited excellent low-temperature activity, high syngas selectivity, and a broad operable temperature window. Additionally, our previously developed superior OC, 1 wt%Pt/cCeO_2_ for CL CH_4_ reforming [14], was analyzed with identical CL POM-WS reaction procedures. Noted that the 1 wt%Pt/cCeO_2_ OC was not pretreated by the aforementioned activation method, as such treatment would decrease its low-temperature activity. Figure S16 exhibits gas generation curves using 1 wt%Pt/cCeO_2_ as OC. 1 wt%Pt/cCeO_2_ OC and 2.5NiO/cCeO_2_ OC exhibited comparable CL POM-WS reactivities at 600 and 700 °C, as shown in Fig. 4i. In contrast, the 2.5NiO/cCeO_2_ OC appeared to be an economical and promising OC for the robust, efficient, and low-temperature CL POM-WS process for the coproduction of syngas and pure hydrogen. Owing to the flexible operating temperature, the as-synthesized OC could be used in various energy and heat recycling systems.
Local Structural Effects on OC Reactivity
The above results demonstrate the remarkable effects of the fine microstructure regulation (Ni loading concentration, activation, and support) on the reaction performance of the NiO/cCeO_2_ OC. To gain insight into the structure**–**activity relationship, the loading state and local structure of the Ni species were examined in detail.
For the 2.5NiO/cCeO_2_ OC, before activation, the Ni species were homogeneously dispersed in the cCeO_2_ support, as shown in Fig. 5. After activation, the Ni species existed in two forms: highly dispersed Ni atoms and NiO NPs of approximately 10–20 nm in size. The XPS results show that the Ni surface molar ratio increased from 1.4 to 2.9 mol% during activation for the 2.5NiO/cCeO_2_ OC (Fig. 6a) and from 0.9 to 1.4 mol% for 1.1NiO/cCeO_2_ OC. The changes in surface composition indicate that the as-prepared samples included Ni species not only on the surface but also inside of the CeO_2_ lattice, and the activation treatment promoted Ni migration and enrichment on the OC surface owing to the strong interactions between the Ni species and CH_4_. As shown in Fig. 6b, with 2.5NiO/cCeO_2_ as OC, the H_2_ generation peak rate in the POM step, i.e., maximum H_2_ generation rate, gradually increased from the first to the fifth cycle, and the total H_2_ generation amount (10 min) increased significantly during the first three cycles and then stabilized, indicating a notable increase in the density of effective Ni active sites in the early cycles, followed by minor structural evolution. The XRD patterns of the 2.5NiO/cCeO_2_ OC (Fig. 6c) show that all diffraction peaks were attributed to cubic fluorite CeO_2_ (Fm 3^-^ m, 225, ICSD 72155). However, for the 6.5NiO/cCeO_2_ OC, phase segregation was observed even in the fresh OC, resulting in large NiO agglomeration in the size range of 100–200 nm (Fig. 5e). The XRD patterns (Fig. 6d) confirmed the existence of the NiO phase in the fresh 6.5NiO/cCeO_2_ OC and the resulting Ni metal phase after reaction with CH_4_ at 600 °C. The larger Ni NPs in the 6.5NiO/cCeO_2_ OC substantially increased CH_4_ dehydrogenation rate and CH_4_ cracking ratio, and these Ni NPs could not be re-oxidized to NiO by steam due to their more stable coordination structure.Fig. 5. High-angle annular dark-field scanning transmission electron microscopy images and energy-dispersive spectroscopy mappings (HAADF-STEM-EDS) of a, b fresh and activated 2.5NiO/cCeO_2_ OCs, c, d fresh and activated 2.5NiO/sCeO_2_ OC, and e fresh 6.5NiO/cCeO_2_ OCFig. 6Structure characterization of NiO/cCeO_2_ OCs and key active site study responsible for the WS reaction. a Ni 2p3/2 and Ce 3d XPS spectra of fresh and activated 2.5NiO/cCeO_2_ OCs. b Activity variation of the 2.5NiO/cCeO_2_ OC during the activation cycles. c, d XRD patterns of 2.5NiO/cCeO_2_ and 6.5NiO/cCeO_2_. e H_2_ MS signals for 2.5NiO/cCeO_2_ and cCeO_2_ OCs after H_2_ pretreatment for 5–180 min, followed by water splitting. f Correlation between H_2_ reduction time and H_2_ generation amount during the WS step
Further investigation of the local structure of the Ni species revealed significant differences. Figure 7a shows the Ni 2p3/2 XPS spectra of the activated and cycled 2.5NiO/cCeO_2_ OC as well as the fresh 6.5NiO/cCeO_2_ OC. The Ni 2p3/2 spectra were fitted with four components: the main line from the bulk NiO at 853.8 eV, surface NiO species at 855.4 eV arising from the absence of apical oxygens, Ni-O-Ce structure at 856.9 eV [44, 45], and the charge-transfer satellite peak at 860.7 eV [46–48]. The peak position and width were fixed, and only the intensity was fitted. The peak area ratio of the surface to bulk (AS/AB) represents the density of surface NiO with a low-coordination structure. The AS/AB of the activated 2.5NiO/cCeO_2_ OC and fresh 6.5NiO/cCeO_2_ OC was 2.3 and 1.1, respectively, which were attributed to the much smaller size of the NiO NPs in the activated 2.5NiO/cCeO_2_ OC. Even after 40 POM-WS cycles at 600 °C, the structure of 2.5NiO/cCeO_2_ OC kept stable, retaining highly dispersed surface NiO species. Figure S17 exhibits the C 1s XPS spectra of fresh, activated, and long-term used 2.5NiO/cCeO_2_ OCs, verifying the strong coking resistance of the OCs.Fig. 7. Correlation between the local structure of Ni active sites and reaction performance. a Ni 2p3/2 XPS spectra fitted results, with bulk NiO nanoparticles as reference. b Ni L3,2 edge XAS spectra of the fresh and activated 2.5NiO/cCeO_2_ OCs, and reference spectra of different Ni chemical states [49]
To further elucidate the effects of activation treatment on the Ni species, XAS was applied in addition to XPS analysis. Figure 7b shows the Ni L3,2 edge XAS spectra of the fresh and activated 2.5NiO/cCeO_2_ OCs, along with reference spectra from the literature [49]. By activation, a slight shift of the Ni L3 edge peak toward lower energy was observed, accompanied by changes in the peak shapes of both the L3 and L2 edges. According to the reference spectra as shown in Fig. 7b, these peak shape changes indicate the different electronic states of Ni species, i.e., the fresh sample has a low-spin Ni^2+^ state, whereas the activated sample has a high-spin Ni^2+^ state. The difference in chemical states corresponds to the structural evolution of the Ni species during activation, namely from atomically dispersed states (possibly including doping into the CeO_2_ lattice) to the formation of small NiO nanoparticles on the CeO_2_ surface, as discussed in Fig. 5.
Up until now, there have been various discussions about the high- and low-spin states of Ni^2+^ based on the viewpoint of structural chemistry. In the case of NiO, high-spin state is usually favored with six coordination local environment, but low-spin state is unusual for NiO. A plausible explanation based on the previous experimental and computational studies is the local structure changes, i.e., square-planar coordination corresponds to low-spin state, and octahedral coordination corresponds to high-spin state [50, 51]. In fresh samples, Ni species exist atomically dispersed state in CeO_2_ NPs, leading to more planar like coordination with low-spin state. On the contrary, Ni species take more ideal octahedral coordination because of Ni clustering on CeO_2_ after activation treatment, which is related to high-spin state. As discussed here, the observation by TEM (Fig. 5) and XAS (Fig. 7) can be conclusively understood from the structure evolution although further extended study is expected in future.
Table 2 summarizes the properties of the NiO/CeO_2_ OCs with varying Ni loadings and the corresponding CH_4_ reaction pathways at 600 °C. The formation of Ni/NiO NPs in a limited size range (< 20 nm) contributed to the high performance of 2.5NiO/cCeO_2_ OC, with the cCeO_2_ support playing a crucial role in restricting Ni site growth. In contrast, when commercial sCeO_2_ was used as support, the Ni/NiO NPs agglomerated to sizes larger than 20 nm (Fig. 5). The specific surface areas of the fresh 2.5NiO/sCeO_2_ and 2.5NiO/cCeO_2_ OCs were 36.9 and 38.0 m^2^ g^−1^, respectively, indicating that the observed difference in Ni site size was not due to surface area of oxide support, but rather to the different confinement effects of sCeO_2_ and cCeO_2_ during Ni migration and agglomeration. As shown in Fig. S1, the commercial CeO_2_ nanoparticles exhibit polyhedral morphology. According to our previous work [52], polyhedral CeO_2_ exposes multiple facets, predominantly the {111} facets. The {001} facets are inherently polar and exhibit lower surface atom density, whereas the {111} facets are nonpolar. Consequently, the (100) facet is more liable to accept Ni atoms and enables a stronger Ni-CeO_2_ interaction. Moreover, Table 2 and Fig. S18 exhibit the O 1* s* XPS spectra and Ce^3+^ and surface non-lattice oxygen ratios in the activated OCs with varying Ni loadings. No direct relationship was found between surface Ce^3+^/non-lattice oxygen ratios and reaction performance. Additionally, the activated 2.5NiO/sCeO_2_ and 2.5NiO/cCeO_2_ OCs exhibited comparable crystallite size (25.6 and 31 nm, respectively), and their H_2_-TPR profiles showed similar oxygen release temperatures (Fig. S19). However, the OSC of the activated 2.5NiO/sCeO_2_ OC at 600 °C (118.2 μmol-O g^−1^) was lower than that of the activated 2.5NiO/cCeO_2_ OC (175.3 μmol-O g^−1^). According to the above analyses, the differences in Ni active site size and OSC@600 °C collectively account for the syngas selectivity difference between 2.5NiO/cCeO_2_ and 2.5NiO/sCeO_2_ OCs at 600 °C, as summarized in Table 2. In 2.5NiO/sCeO_2_, the larger Ni sites accelerated CH_4_ dissociation rate, which exceeded the lattice oxygen supply rate at lower temperatures, resulting in increased CH_4_ cracking and reduced syngas selectivity.Table 2. Properties of the NiO/CeO_2_ OCs with varying Ni loadings and the corresponding CH_4_ reaction pathways at 600 °CCe^3+^ (%)O_β_ (%)OSC@600 °C (μmol-O g^−1^)SBET (m^2^ g^−1^)(Fresh)Crystallite size (nm)(Activated)Ni site size (nm)CH_4_ reaction pathway[CO + 2H_2_][C + H_2_]1.1NiO/cCeO_2_19.417.8————99.3%0.7%2.5NiO/cCeO_2_19.617.0175.338.031.010–2099.4%0.6%2.5NiO/sCeO_2_19.216.5118.236.925.620–10084.5%15.5%3.5NiO/cCeO_2_19.720.0——25.0—92.0%8.0%6.5NiO/cCeO_2_18.930.1——15.1100–20018.8%81.2%
Moreover, the OC structure experiences dynamic evolution during the POM and WS cycles. As shown in Fig. 8, the Ni and NiO NPs transformed into each other during successive POM-WS cycles and the O/Ce atomic ratios of the reduced and oxidized OCs are 1.13 and 1.89, respectively, which clearly demonstrates the reversible redox behavior of the NiO/cCeO_2_ OC.Fig. 8. Phase evolution of Ni active sites during the chemical looping process. HAADF-STEM images and EDS mappings of 2.5NiO/cCeO_2_ OC: a after the fifth POM cycle and b after the fifth WS cycle
During the POM reaction, Ni serves as the principal active site for CH_4_ activation. As the reaction progresses, the concentration of oxygen vacancies increases, altering the interaction between Ni and the support and modifying the electronic structure of the Ni sites, resulting in reduced CH_4_ dissociation activity (Fig. 3b). The active sites responsible for the WS reaction were also identified. Figure 6e compares the WS reactivity of 2.5NiO/cCeO_2_ and cCeO_2_ OCs after H_2_ pretreatment for varying durations (5–180 min). Both OCs exhibited nearly identical H_2_ production, which was directly proportional to the H_2_ pretreatment time (Fig. 6f). These results indicate that oxygen vacancies, rather than Ni sites, serve as the primary active sites for water splitting in this system. As oxygen vacancies are replenished with oxygen atoms, the WS activity of the OCs gradually declined.
A schematic illustration of the OC activation, structural evolution, and key active sites responsible for the POM and WS reactions is provided in Fig. 9. Activation pretreatment is essential for establishing stable low-temperature performance. During activation, the formation of surface-fused Ni/NiO NPs occurred in two stages: (i) Ni migration and surface enrichment and (ii) confined aggregation of Ni atoms, which significantly accelerated the reaction kinetics (from the first to the second cycle, Fig. 3b). These isolated Ni/NiO NPs have a unique Ni electronic structure and moderate CH_4_ dissociation activities compared to the large Ni/NiO particles, leading to distinct CH_4_ conversion pathways. Precise control of the Ni active site size to 10–20 nm and the well-matched lattice oxygen supply capacity of cCeO_2_ are the key to the high CH_4_-to-syngas selectivity, as they effectively balance the CH_4_ activation and oxygen supply rates. Meanwhile, the intrinsic chemical properties of the support significantly influence the electron transfer between Ni and the support, thereby markedly changing the activity of the Ni sites.Fig. 9. Schematic illustration of the OC activation, structural evolution, and key active sites responsible for the POM and WS reactions
Conclusions
In this study, a high-performance NiO/cCeO_2_ oxygen carrier (OC) for low-temperature chemical looping partial oxidation of methane and water splitting (CL POM-WS) was successfully developed. The reaction temperature was significantly reduced from conventional 800–900 to 550–600 °C. A reaction-driven activation strategy was proposed, which markedly enhanced the low-temperature activity and reaction kinetics of the OCs. Upon activation, Ni species migrated and became enriched on the OC surface from initially uniform dispersion within the cCeO_2_ support, forming 10–20 nm Ni/NiO NPs, and XAS analyses revealed a chemical state transition from low-spin Ni^2+^ to high-spin Ni^2+^. The cubic CeO_2_ support exerted a confinement effect on Ni migration compared with the commercial CeO_2_ nanoparticles. The small Ni/NiO NPs and well-matched lattice oxygen supply capacity of cCeO_2_ contributed to the selective POM while suppressing CH_4_ cracking. With an optimal Ni loading amount of 2.5 mol%, the CH_4_ conversion remained CH_4_-to-syngas pathway across the 500–800 °C range. In the POM step, the CO selectivity exceeded 98.5%, producing a desirable syngas (H_2_/CO ≈2), and nearly pure H_2_ was produced in the WS step. The 2.5NiO/cCeO_2_ OC exhibited superior long-term stability and resistance to coke deposition at 600 °C over 40 cycles (approximately 20 h). The successful development of low-temperature and selective OC makes the CL POM-WS process highly competitive as an industrial alternative to steam methane reforming in terms of energy consumption, process safety, and gas purification. Instead of rough compositional adjustments, precise control over the local structures of OCs and deeper understanding of the structure–activity relationship are critical for developing more excellent low-temperature OCs. In future work, we aim to develop continuous-flow hydrothermal synthesis routes for NiO/cCeO_2_ OCs to enable reproducible and scalable production of OC particles, laying foundation for the practical application of newly developed low-temperature OCs.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 10246 KB)Supplementary file2 (DOCX 13 KB)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1N. I. o. S. a. Technology, NIST Chemistry Web Book. (2023). https://webbook.nist.gov/chemistry/