The redox properties of nickel (Ni) underpin its pivotal role in corrosion science, catalysis, and emerging energy technologies. Renowned for its high catalytic activity, Ni serves as a cost-effective alternative to precious metals such as Au, Pt and Pd in driving reactions that convert greenhouse gasses into value-added products and fuels [1][2][3][4][5] .
Beyond catalysis, Ni's versatility extends to energy storage [6][7][8][9] , where its unique capacity to reversibly absorb atomic hydrogen (H) into its crystal lattice positions it as a promising candidate for solid-state H storage. This capability enables H to occupy interstitial sites or form metastable hydrides (Ni-H), bypassing the energy penalties of conventional compression or liquefaction methods. Ni further excels as a critical component in heatand corrosion-resistant alloys (e.g., Ni-Fe, Ni-Al, Ni-Cr), which are indispensable for ensuring long-term structural integrity in harsh environments-from fossil fuel processing to renewable energy systems [10][11][12][13][14] . Critically, Ni's redox dynamics dictate its performance in these roles: gas-solid reactions at its surface propagate into the bulk at elevated temperatures, altering mechanical and chemical properties.
However, a critical knowledge gap persists in understanding how the structure and phase evolution of Ni respond to dynamic, mixed environments involving O2, H2, H2O vapor, and their synergistic interactions. While studies have elucidated the behavior of Ni under isolated oxidizing (O2) or reducing conditions, real-world applications, such as catalytic reactors, H2 storage systems, and industrial corrosion environments, often involve fluctuating gas mixtures [15][16][17][18][19] . For instance, the coexistence of H2O and O2 during steam reforming or alloy degradation introduces competing redox pathways that remain poorly resolved [20][21][22] . Current models fail to capture the transient microstructural transformations under such complex conditions. This gap limits the ability to predict Ni's stability, hydrogen storage capacity, or catalytic efficiency in realistic operating scenarios.
Bridging this gap requires advanced in-situ techniques to map redox-driven phase transitions in real time and quantify the interplay between environmental compositions, defect dynamics, and reaction kinetics.
This study systematically probes the oxidation and reduction behavior of polycrystalline Ni under distinct atmospheric conditions: dry O₂, H2, and O2+H₂O vapor.
Using in-situ synchrotron X-ray diffraction, we first characterize the evolution of texture in NiO films formed during the oxidation of Ni foil at 650 °C in O₂, and the subsequent H₂-
The XRD patterns obtained from the pristine poly-Ni foil (Fig. 1 The furnace-oxidized sample, shown in Fig. 1(b), is first reduced in H2, then re- However, it is important to note that the presence of the HCP-Ni phase is highly localized. This is evident by the short arc of the HCP-Ni(002) diffraction ring, which completely disappears when the sample is slightly shifted relative to the X-ray beam, thus probing a different area of the sample. As shown in Fig. 2(b), this small shift in sample position results in a slight increase in intensity of both the NiO(111) and NiO(200) rings (Fig. 2(e)), while the HCP-Ni(002) ring vanishes entirely. The increase in intensity of the NiO peaks can be attributed to the X-ray beam irradiation effect, where the initial sample position, having been subjected to more X-ray irradiation, undergoes further reduction of the oxide. In contrast, the new sample location, with less X-ray exposure, experiences less reduction, leading to a slightly increased intensity of the NiO(111) and NiO(200) rings (Fig. 2(e)). This behavior is consistent with other studies indicating that high-energy Xrays can enhance the chemical activity of gas molecules 25,26 . Over the course of 450 min at 650 °C in the H2+Ar mixture, the oxide undergoes significant reduction, with the NiO(200) ring almost completely losing its intensity, while the NiO(111) ring retains only a faint intensity (Figs. 2(c, d)). Fig. 2(e) shows the intensity evolution of the NiO(200) and NiO(111) peaks, displaying an initially rapid reduction rate, which then approaches a saturated stage.
The influence of water vapor on the oxidation process is examined by oxidizing a fresh polycrystalline Ni foil sample at 650 °C under two conditions with varying H2O vapor content. Fig. 3 presents the time-sequence XRD patterns obtained during oxidation under a low H2O content condition, with an H2O/O2 pressure ratio of 1:100,000 in a 5.0´10 -6 Torr H2O + 0.5 Torr O2 mixture. As shown in Fig. 3(a), the as-loaded Ni foil has a thin layer of native NiO layer, as indicated by the weak diffraction rings for NiO(111) and NiO(200). In Figs. 3(b-d), oxidation under the lower H2O content results in the appearance of additional peaks, beyond those corresponding to NiO and FCC-Ni. These new peaks, observed at 2θ = 17.53°, 18.11°, 18.67°, and 20.09°, correspond to the HCP-Ni(010), γ-NiOOH(104), HCP-Ni(002), and HCP-Ni(011), respectively. The d-spacings associated with these phases-HCP-Ni and γ-NiOOH-are consistent with those reported in the literature 27 . Fig. 3(e) shows the intensity evolution of these peaks, demonstrating an initially rapid oxide growth rate, which then slows down and reaches a saturation stage of the oxidation process. oxidation; (e) Time-dependent plot of ring intensities, showing the relative intensities of the labeled diffraction rings. The absence of diffraction intensity during the time periods from ~70 min to ~100 min and from ~210 min to ~310 min is due to the loss of the X-ray beam during the oxidation. The bottom panels in (a-d) display line profiles across the respective ring patterns, showing the diffraction intensity as a function of the Bragg angle. (f-i) Evolution of the normalized peak intensity of NiO(111), NiO(200), HCP Ni(010), and NiOOH(104) as a function of oxidation time, and corresponding fits (in red) to the JMAK equation, with the extracted Avrami constant n and rate constant k. (j) Schematic illustration of 1D growth of NiO, NiOOH and HCP-Ni during the oxidation of a Ni surface initially covered by a pre-existing native NiO layer. The kinetics of the oxidation process, shown in Fig. 3(e), are evaluated by fitting the experimental data to the Johnson-Mehl-Avrami-Kolmogorov (JMAK) equation, ( ) = 1 -exp( -) 28,29 . As shown in Figs. 3(f, g), the JMAK fitting yields an Avrami exponent n » 1 for both the NiO(111) and NiO(200) peaks, indicating a 1D oxide growth mechanism. This behavior arises because the Ni surface is pre-covered with an oxide layer, eliminating the need for additional nucleation events. Instead, oxidation proceeds primarily along the thickness (out-of-plane) direction, rather than laterally (in-plane), as depicted in the schematic in Fig. 3(j). Lateral oxide growth is restricted as adjacent oxide grains physically constrain one another, forcing vertical growth. The consistency of the rate-constant k values for both the NiO(111) and (200) peaks-within the experimental uncertainty-further supports the vertical oxide growth mechanism, driven by the preexisting oxide layer. The Avrami exponent extracted from the NiOOH peak is also close to 1, consistent with 1D vertical growth. Since the NiOOH formation occurs via the direct transformation of the existing NiO lattice, driven by the inward dissolution of H from dissociative adsorption of H2O molecules at the surface, the near-unity Avrami exponent suggests that nucleation is saturated at the onset of the transformation, with growth being primarily driven by unidirectional H diffusion. In contrast, the Avrami exponent derived from the HCP-Ni peak is n < 1 (Fig. 3(h)), indicative of diffusion-limited kinetics. In this case, the long-distance inward diffusion of H through the NiO layer to reach the Ni substrate creates a kinetic bottleneck that restricts the nucleation and growth of the HCP-Ni within the Ni lattice. While the rate constants k for HCP Ni and NiOOH appear higher than those for intensities of the labeled diffraction rings. (f-i) Evolution of the normalized peak intensities of NiO(111), NiO(200), HCP Ni(002), and NiOOH(104) as a function of oxidation time, and corresponding fits (in red) to the JMAK equation, with the extracted Avrami constant n and rate constant k. (j) Schematic illustration of the nucleation and growth of 3D NiO islands along with resulting 3D growth of NiOOH and HCP Ni during the oxidation of a pristine Ni surface. The oxidation of the pristine Ni surface results in the formation of a more welldeveloped, less defective NiO oxide layer, as evidenced by the sharp NiO(111) and NiO(200) rings shown in Figs. 4(b). The reduced defect density in the oxide layer makes it less reactive toward dissociative adsorption of H2O molecules, which in turn delays the formation of the H-dissolution-induced phases. This observation is consistent with our DFT modeling, shown later in Fig. 7 which demonstrates that the presence of atomic defects (i.e., O and Ni vacancies) on the NiO(100) surface can significantly enhance the dissociative adsorption of H2O molecules. This process leads to the generation of OH and H on the surface, where OH can further decompose into O and H. The resulting H can easily penetrate the NiO lattice, overcoming only small energy barriers of ~0.35 eV 30,31 . The presence of a hydroxylated surface at 650 °C is unexpected, as previous studies have shown that Ni hydroxides and hydrides typically lose thermal stability and decompose at temperatures above approximately 300 °C32 . However, in the present study, the in-situ nature of the experiment reveals that the continued dissociative adsorption of H2O not only dynamically replenishes the Ni hydroxide and hydride that are thermally degrading but also promotes their further growth. This is evidenced by the increasing intensity of the γ-NiOOH and HCP-Ni peaks over time, as shown in Fig. 4(e). XRD results obtained from oxidation in 5.0´10 -3 Torr H2O + 0.5 Torr O2 at 650 °C for 460 min (Fig. 4(d)) show the peaks associated with the FCC-Ni, NiO, HCP-Ni and γ-NiOOH phases, similar to those observed with the oxidation under the lower H2O content (Fig. 3(d)). In addition to those commonly observed peaks, two new peaks at 2θ = 17.30 and 21.11 (with corresponding d-spacings of 2.34 Å and 1.92 Å, respectively) appear after oxidation with the higher H2O content (Fig. 4(b-d)). These additional peaks do not match any known phases of NiO or any other hydroxylated forms of NiO or hydride structures. Therefore, they are believed to arise from the long-range ordering of atomic vacancies in Factors such as variations in crystallinity, texture, and differences in experimental conditions compared to those used to derive the RIR values can all introduce error into the phase quantification 46,47 . Nonetheless, the in-situ XRD detection of HCP-Ni under O₂ + H₂O conditions is consistent with prior transmission electron microscopy (TEM) studies of NiAl oxidation in water vapor 48 , where HCP Ni formation was attributed to H uptake and stabilization under proton-rich, oxidizing environments. To more precisely resolve the spatial distribution and depth profiles of the NiOOH, HCP-Ni and NiO phases, future studies incorporating surface-sensitive or cross-sectional techniques-such as sputtering-based X-ray photoelectron spectroscopy (XPS) or TEM-would be highly valuable. 4% H2 gas for 3 h. (d) The same sample in (c) following oxidation in 4 Torr O2 at 650 °C for 2 h. (e) Cross-sectional image showing the growth of a thick oxide after oxidation of a clean polycrystalline Ni foil in 4 Torr H2O + O2 at 650 °C for 2 h, revealing the formation of multiple oxide sublayers. (f) Cross-section revealing the formation of a NiO overlayer after oxidation of a clean Ni foil in 0.5 Torr O2 at 650 °C for 2 h.
In a separate experiment, oxidation in a humidified environment (4 Torr H₂O + O₂ at 650 °C for 2 h, similar to the conditions in Figs. 3 and 4, but at a higher total pressure) leads to the formation of a multilayered NiO structure with a significantly greater total thickness of ~1.48 µm (Fig. 5(e)). This pronounced growth is attributed to the catalytic effect of water vapor, which promotes the formation of proton-related defects and vacancies, thereby enhancing diffusion through the oxide scale 31,52 . The rapid oxide growth in this environment also generates internal stresses that promote buckling and interfacial delamination, as evidenced by the uneven surface and interrupted interface in DFT modeling To gain a better understanding of the proton-induced phase and structure dynamics during oxidation involving H2O vapor, DFT calculations are performed to evaluate the energetics of H incorporation into the NiO and Ni lattices. For the intact NiO lattice, the adsorption energy of H is -0.94 eV, with a diffusion barrier of 0.62 eV (Fig. 6(a)). In contrast, for the NiO lattice containing an O vacancy, H occupies the vacancy site, with the adsorption energy of -3.22 eV (Fig. 6(b)). This significantly stronger binding (-3.22 eV vs. -0.94 eV) highlights the critical role of O vacancies in stabilizing H within NiO and enhancing proton retention. For the FCC Ni lattice, the H adsorption energy is -4.50 eV, with a diffusion barrier of 0.55 eV (Fig. 6(c)). The lower diffusion barrier (0.55 eV vs. 0.62 eV in intact NiO) and more negative adsorption energy (-4.50 eV vs. -0.94 eV in intact NiO, and -3.22 eV in O-deficient NiO) in FCC Ni indicate a thermodynamic and DFT calculations are then used to investigate why hydride phases (NiOOH, HCP Ni) form more slowly on a well-developed NiO overlayer than on a defective native NiO film, despite higher H2O vapor exposure (Fig. 4). As shown in Fig. 3 Additional DFT calculations reveal that H2O dissociation at Ni sites is also strongly influenced by nearby vacancies (Figs. 7(d-f)). On perfect NiO(100), the adsorption energy is -0.38 eV, with a dissociation barrier of 0.43 eV and a system energy increase of +0.35 eV (Fig. 7(d)). With an adjacent O vacancy adjacent, adsorption strengthens (-0.46 eV), the barrier increases to 0.61 eV, and the reaction becomes less endothermic (+0.19 eV, Fig. 7(e)). By contrast, a nearby Ni vacancy lowers the barrier significantly to 0.14 eV, with stronger adsorption (-0.49 eV) and near-thermoneutral dissociation (0.08 eV, Fig. 7(f)). These calculations demonstrate that surface vacancies reduce dissociation barriers and stabilize H and OH species, shifting H2O dissociation from endothermic toward thermoneutral or exothermic on defective NiO surfaces.
While synchrotron radiation can induce radiolysis of water vapor, potentially producing reactive species that may influence surface reactions 54,55 , its effect on oxidation kinetics in our in-situ XRD measurements is considered minimal (Figs. 3, 4). A slight enhancement in NiO reduction under H2 is observed under X-ray exposure (Fig. 2), but oxidation at 650 °C is dominated by thermally driven Ni diffusion, as H₂O readily dissociates on NiO at such high temperatures. This is supported by ex-situ furnace oxidation without X-ray, which shows thicker NiO scales in O₂ + H₂O than in dry O₂ (Fig. 5). These results, consistent with in-situ data, confirm that the promoting role of H2O in intrinsic to the oxidation process-not a radiation-induced artifact-and that the observed kinetics reflect the true oxidation behavior under humid conditions.
The XRD measurements demonstrate that furnace oxidation of polycrystalline Ni Ni oxidation in H2O vapor produces a highly defective oxide 31 , as confirmed from our SEM observations (Fig. 5(e)) and diffraction rings linked to ordered vacancies in the NiO lattice (Fig. 4). These vacancies, promoted by dissolved protons 31 , enhance lattice diffusion and accelerate oxide growth in the presence of H2O vapor (Fig. 5). O2, H2, and H2O vapor were introduced in the chamber via a leak valve allowing for the controlled flow of gases into the chamber.
The samples treated in the tube furnace were examined using a Hitachi 4800S SEM to observe their surface morphology, with an accelerating voltage of 10 kV and a beam current of 10 µA. To better visualize the thickness of the nickel oxide layer, crosssectional imaging was carried out using a Helios G5 dual-beam SEM/FIB system. A ~1.5 µm-thick Pt protective layer was first deposited on the target area to protect the surface during subsequent ion beam milling. The region was then milled using the "regular cross section" mode at 30 kV and 9.9 nA, followed by fine milling using the "cleaning cross section" mode at 2.6 nA to obtain high-quality cross-sectional images.
The Hubbard U corrected density-functional theory (DFT+U) calculations were performed using the Vienna ab-initio simulation package (VASP) 61,62 , with the Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA) and projector augmented wave (PAW) potentials 63 . For the Ni d orbitals, the Hubbard U correction is applied with parameters of U = 5.0 and J = 0.1, while anti-ferromagnetic spin ordering is integrated into the NiO rock salt structure's calculations for calibration. Spin-polarized calculations are performed for all structures. A plane-wave cutoff energy of 500 eV was used to calculate adsorption energies. Brillouin zone integration was conducted with a (2×4×1) k-point mesh for slabs and a (8×8×8) mesh for bulk, based on Monkhorst-Pack grids 64 . NiO(100) surfaces were constructed by cleaving supercells derived from the bulk structure, comprising successive slabs with five atomic layers, separated by a 15 Å vacuum region. The positions of atoms in the two bottom layers were fixed, while the top three layers were allowed to relax, ensuring energy convergence below ~ 10 -5 eV and all force components on each atom less than 0.015 eV/Å. Adsorption energies of H2O species on the NiO(100) surfaces with various types of defects were investigated, along with structure evolution in each calculation. Energy barriers for given dissociation pathways were computed using the NEB method. The resulting atomic structures were visualized using the Visualization for Electronic and Structure Analysis (VESTA) software package.