Molecular or dissociative adsorption of water on clean and oxygen pre-covered Ni(111) surfaces

Abstract

Water adsorption and dissociation on clean and oxygen pre-covered Ni(111) surfaces have been computed systematically by using density functional theory and ab initio atomistic thermodynamics. The adsorption of H, O and OH prefers 3-fold faced centered cubic hollow sites, and that of H₂O prefers the top site. For (H₂O)_n aggregation, direct O–Ni interaction and H-bonding synergistically determine the adsorption energy, which is structure insensitive for large adsorbed clusters. At low coverage (θ ≤ 0.25 ML), OH adsorbs perpendicularly and prefers remote distribution without H-bonding, and the OH saturation coverage should be 0.625 ML on the basis of H₂O dissociative adsorption. The adsorption configuration for 4O (0.25 ML) prefers a p(2 × 2) structure, in agreement with the experiment, and the O saturation coverage should be 0.25 ML based on H₂O dissociative adsorption. On the 0.25 ML O pre-covered Ni(111), H₂O greatly prefers molecular adsorption [4O + 4H₂O(s)] over dissociative adsorption [8OH] thermodynamically, and this is in agreement with the photoelectron spectroscopy results and in disagreement with previously and recently proposed results from single crystal adsorption calorimetry of D₂O adsorption. It is noted that the computed bond energy and formation enthalpy of the supposed surface hydroxyls on the basis of the molecular adsorbed state [4O + 4H₂O(s)] are much closer to the experimentally estimated results than those computed on the basis of the dissociatively adsorbed state [8OH]. Furthermore, the computed H₂O desorption temperature on the basis of the molecular adsorbed state [4O + 4H₂O(s)] is in excellent agreement with experimental results (284 vs. 275–300 K), while that from the dissociatively adsorbed state [8OH] differs strongly (197 K). All these support H₂O molecular adsorption instead of dissociative adsorption, and this needs further experimental investigations and confirmations. The vibrational frequencies of 4O, 4O + 4H₂O and 8OH adsorption configurations have been computed to aid experimental studies.

---

Introduction

The interaction of water and solid surfaces plays an important role in many industrial processes, e.g. catalysis, corrosion, electrochemistry, environmental chemistry and materials science.^1–4 Nickel catalysts have been widely used in methanation [CO + 3H₂ = CH₄ + H₂O],^5–8 methane steam reforming [CH₄ + H₂O = CO + 3H₂],^9–18 methane dry reforming [CH₄ + CO₂ = 2CO + 2H₂]^19,20 and methane partial oxidation [2CH₄ + 3O₂ = 2CO + 4H₂O].²¹ Since water, as either a product or reactant, is involved in these reactions, it is essential to study the reaction of H₂O formation or dissociation on nickel surfaces.

Using electron stimulated desorption (ESD) ion angular distribution, low-energy electron diffraction (LEED) and temperature programmed thermal desorption (TPD), Madey et al.²² studied H₂O adsorption on clean and O-pre-covered Ni(111) surfaces from the lowest to the saturation coverage. On the clean surface, they found the desorption peak at 165–170 K, which is actually due to H₂O molecules in the first monolayer (ML) and the attractive H-bonding among these H₂O molecules, and the adsorption energy of a single molecule is 0.44 eV. On the O-pre-covered surface (≤0.05 ML) at low H₂O coverage (≤0.08 ML), there are two desorption peaks at 180–200 K and at 275–300 K, where the higher temperature peaks were supposed to be a consequence of the formation [O + H₂O(s) = 2OH] of surface OH at low temperature (≥120 K) and their subsequent disproportionation to yield surface O and gaseous H₂O at high temperature (≥200 K). Increasing the O pre-coverage from 0.05 to 0.25 ML at constant H₂O coverage (∼0.3 ML), the H₂O desorption rate and H₂O coverage are virtually unchanged in the temperature range of 250–300 K. Using TPD and ESD, Stulen and Thiel²³ studied H₂O adsorption on clean and O pre-covered Ni(111) surfaces and found that on the clean surface the desorption of the first layer chemisorbed H₂O has a first order kinetics with a desorption energy of about 0.42 eV; on the O-pre-covered surface (≈0.07 ML) the high temperature desorption (180–260 K) is due to the dissociative recombination from 2OH disproportionation.

Using TPD, angle resolved ultraviolet photoelectron spectroscopy (UPS), X-ray photoelectron spectroscopy (XPS) and LEED, Pache et al.²⁴ studied H₂O adsorption on clean as well as 0.25 ML O pre-covered Ni(111) surfaces. On the clean Ni(111) surface, they found that the saturation coverage of chemisorbed H₂O is 0.66 ± 0.05 ML by XPS and the desorption temperature of 171, 152 and 158 K comes from the chemisorbed bilayer, transition layer and condensed water, respectively; the adsorption energy of the chemisorbed layer is 0.59 ± 0.05 eV. On the O pre-covered surface (0.25 ML), an additional peak is observed in TPD of water at low coverage reaching from 180 to 260 K saturating at a coverage of 0.25 ± 0.04 ML; this desorption peak is attributed to the molecularly adsorbed H₂O on the surface and no evidence is found for H₂O dissociation and OH formation from work function and angle resolved UPS measurement. Using TPD and XPS at 10⁻⁵ mbar and low coverage, Schulze et al.²⁵ studied H₂O adsorption on clean and O pre-covered Ni(111) surfaces. On the clean surface two adsorption states were observed; the first adsorbed layer with a maximum coverage of 0.42 ML has an adsorption energy of 0.54 ± 0.03 eV as well as a desorption peak at about 163 K at low coverage, and this peak shows saturation with increasing coverage; at coverage larger than 0.42 ML, the desorption energy is about 0.40 eV. On the 0.2 ML O pre-covered surface, photoelectron spectroscopy shows that the adsorbed H₂O remains in the molecular state and does not desorb at 156 K, while it desorbs only at higher temperatures of 180 to 260 K. The UPS and XPS results^24,25 contradict H₂O formation from 2OH disproportionation as proposed by Madey et al.²² Further TPD studies of H₂O on the clean Ni(111) surface confirms desorption temperature of about 170 K for the first layer adsorbed H₂O molecules.^26–28

Recently, Zhao et al.²⁹ measured the heats of adsorption of D₂O on clean and O pre-covered Ni(111) surfaces by single crystal adsorption calorimetry. On the clean surface at 100 K, the differential heat of adsorption of D₂O starts at 0.56 eV and slightly increases up to ∼0.3 ML due to the intermolecular H-bonding; the integral heat of molecular adsorption is 0.56 eV for the most stable adlayer at 0.5 ML. On the 0.25 ML O pre-covered surface, D₂O prefers molecular adsorption at 100 K with an integral heat of 0.58 eV, which is slightly higher than that on the clean surface by 0.02 eV. At 170 K, D₂O was proposed to favor dissociative adsorption producing surface OH up to 0.23 ML D₂O coverage. The differential heat starts from ∼0.73 eV and the decreases to 0.68 eV at 0.23 ML D₂O saturation coverage. The integral heat is 0.70 eV at saturation coverage, which, in turn, gives an enthalpy of formation and bond energy of surface hydroxyls of −2.88 and 3.27 eV, respectively.

Theoretically, many studies reported H₂O adsorption and dissociation on Ni(111) and found H₂O adsorption at the top site with O–H bonds nearly parallel to the flat surface. As listed in Table 1, different models and methods gave different dissociation barriers, while coverage dependent H₂O adsorption and dissociation are missing despite the fact that H₂O prefers H-bonding and interacts with pre-covered oxygen atoms. By constructing a new nine-dimensional potential energy surface based on density functional theory points,^30,31 Jiang and Guo studied water dissociation on Ni(111) and found their results to be in better agreement with the experiment.

Table 1 Dissociation barriers and energies (E_a and E_r, eV) and O–H breaking distances (d_O–H, Å) of the transition state of H₂O on Ni(111) with different sizes and methods (m for total slab layers and n for relaxed surface layers, mT/nR)

Size	Method	E a	E r	d O–H	Ref.
a This work.
H2O = OH + H
4 × 4 (4T/2R)	PBE-D3	0.68	−0.56	1.503
3 × 3 (3T/2R)	PBE	0.97	−0.22		32
2 × 2 (4T/2R)	PBE	0.67	−0.20		33
3 × 3 (4T/2R)	RPBE	0.90	−0.28		34
2 × 2 (3T/1R)	RPBE	0.92	−0.10		35
2 × 2 (4T/2R)	PBE	0.90	−0.24	1.586	36
2 × 2 (4T/2R)	PBE	0.69	−0.18	1.538	37
3 × 3 (4T/2R)	PBE	0.88	−0.19	1.55	38
3 × 3 (3T/1R)	PW91	0.96	−0.26	1.56	39
2 × 2 (3T/1R)	PBE	0.89	−0.16		40
2 × 2 (4T/2R)	PW91	0.80	−0.41		41
3 × 2	PBE	0.74	−0.56	1.881	42
2 × 2	PW91	0.77	−0.26	1.60	43
OH = O + H
4 × 4 (4T/2R)	PBE-D3	1.25	−0.18	1.378
3 × 3 (3T/2R)	PBE	1.19	−0.17		32
2 × 2 (4T/2R)	PBE	0.79	−0.03		33
3 × 3 (4T/2R)	RPBE	0.85	−0.20		34
2 × 2 (3T/1R)	RPBE	0.85	−0.34		35
2 × 2 (4T/2R)	PBE	1.01	−0.06	1.374	36
2 × 2 (4T/2R)	PBE	0.80	−0.03	1.340	37
3 × 3 (3T/1R)	PW91	1.03	−0.04	1.35	39
2 × 2 (3T/1R)	PBE	0.97	−0.33		40
2 × 2 (4T/2R)	PW91	0.97	−0.36		41
3 × 2	PBE	0.65	−0.29	1.465	42
3 × 3 (4T/2R)	RPBE	1.03		1.59	30
3 × 3 (4T/2R)	PW91	0.67		1.56	30
O + H2O = 2OH
4 × 4 (4T/2R)	PBE-D3	0.46	0.14	1.558
3 × 3 (3T/2R)	PBE	0.66	0.38		32
2 × 2 (4T/2R)	PW91	0.72	0.30		41
2H2O = OH + H + H2O
4 × 4 (4T/2R)	PBE-D3	0.68	−0.25	1.475
3 × 3 (3T/2R)	PBE	0.82	0.09		32
3 × 3 (4T/2R)	PBE	0.81	0.22	1.45	38

---

To study the interaction of H₂O and Ni(111) systematically, we computed H₂O adsorption and dissociation on clean and oxygen pre-covered Ni(111) surfaces. Our main concerns are coverage dependent adsorption and dissociation, surface oxygen mediated reactions and H₂O desorption temperatures on different surfaces. Systematic and comparative results are reported. To aid the experimental studies, the vibrational frequencies of 4O, 4O + 4H₂O and 8OH adsorption have been computed.

Computational methods and models

Density functional theory calculations were performed with the plane wave based pseudo-potential code in the Vienna ab initio simulation package (VASP).^44,45 The electron–ion interaction is described with the projector augmented wave method (PAW).^46,47 The exchange and correction energies are described within generalized gradient approximation using the Perdew–Burke–Ernzerhof formulation (GGA-PBE).⁴⁸ To include van der Waals interaction, which is known to play an important role in weak interaction systems,^49,50 we included the D3 correction of Grimme.⁵¹ All structures were optimized with PBE including D3 correction (PBE-D3). The plane wave cutoff energy was specified by 400 eV, and the electron smearing method with σ = 0.20 eV was used to ensure energies with errors due to smearing of less than 1 meV per unit cell. The convergence criteria for geometry optimizations of total energy and forces were 10⁻⁴ eV and 0.02 eV Å⁻¹. Spin polarization was included. The transition state was identified using the climbing image nudged elastic band (CI-NEB) method.⁵² For the bulk structure, the face centered cubic crystal structure was used and the reciprocal space is sampled with a 12 × 12 × 12 k-point grid using the Monkhorst–Pack method.⁵³ The calculated lattice constant (3.525 Å) and magnetic moment (0.64 μ_B) are close to the experimental values (3.52 Å⁵⁴ and 0.61 μ_B ⁵⁵). For Ni(111), to avoid interactions among slabs, the vacuum layer between periodically repeated slabs was set at 15 Å. The (4 × 4) slab model consists of four layers containing a total of 64 Ni atoms, where the top two layers including adsorbates were relaxed and the bottom two layers were fixed in their bulk positions. The 3 × 3 × 1 Monkhorst–Pack k-point sampling was used.

The adsorption energy (E_ads) was defined as E_ads = E(X/slab) − E(slab) − E(X), where E(X/slab) is the total energy of the slab with adsorbates on the surface in equilibrium; E(slab) is the energy of the clean Ni(111) surface; E(X) is the energy of the free adsorbed molecule in the gas phase. The more negative the E_ads, the stronger the adsorption. The desorption energy E_des is the negative of the adsorption energy, E_des = −E_ads. All relative energies include the correction of zero point energies (ZPE). The activation barrier (E_a) is defined as E_a = E_TS − E_IS, and the reaction energy (E_r) is defined as E_r = E_FS − E_IS, where E_IS, E_FS and E_TS are the energies of the corresponding initial state (IS), final state (FS) and transition state (TS). The stepwise adsorption energy was defined as ΔE_ads = E((n + 1)X/slab) − E(nX/slab) − E(X), where a position ΔE_ads for (n + 1) adsorbed X species indicates the saturation adsorption with nX species. In addition, we carried out revised PBE (RPBE⁵⁶) single-point energy calculation with dispersion correction on the basis of PBE optimized structures (RPBE-D3) and ZPE correction from PBE.

Further we used ab initio atomistic thermodynamics^57,58 to model the H₂O desorption temperature under given conditions. There are many related examples for the adsorption and desorption of H₂, H₂O, CO and other small molecules by using this method.^59–64 According to previous work, we took H₂O desorption on the Ni(111) surface: H₂O/Ni → Ni + H₂O(g) as an example; the change of Gibbs free energy (ΔG) for this reaction can be described as eqn (1):

ΔG = G[Ni(111)] + G_gas(H₂O) − G[{H₂O}/Ni(111)] (1)

In this equation, G[Ni(111)] is the Gibbs free energy of the Ni(111) surface. G[{H₂O}/Ni(111)] is the Gibbs free energy of the Ni(111) surface with the H₂O molecule. We apply the DFT calculated total energy to replace the Gibbs free energy of the solid surfaces.⁶⁵ The G_gas(H₂O) term is equal to μ(H₂O). The H₂O chemical potential can be described as

is the DFT calculated energy of the isolated H₂O molecule (including zero point energy), the term is the chemical potential at different temperatures, which can be found in thermodynamic tables; and k_BT ln(p_H2O/p⁰) is the contribution of temperature and H₂O partial pressure to the H₂O chemical potential. Therefore eqn (1) can be rewritten as

Results and discussion

(a) Adsorption of H, O, OH and H₂O

For H, O, OH and H₂O adsorption on Ni(111), the top (T), bridge (B), 3-fold hexagonal close-packed (hcp-3H) and 3-fold face centered-cubic (fcc-3F) sites have been considered with different initial configurations. The most stable adsorption configurations are shown in Fig. 1 and the related parameters are listed in Table 2.

Fig. 1 Top (a) and side (b) views of Ni(111) with adsorption sites: top (T), bridge (B), 3-fold hexagonal close-packed (hcp-3H) and 3-fold faced centered cubic (fcc-3F), as well as the most stable adsorption configuration and energy of (c) H, (d) O, (e) OH, and (f) H₂O (Ni/blue; O/red; H/white).

Table 2 Adsorption energy (E_ads, eV, PBE-D3) at the most stable adsorption site and the nearest distances between the Ni atom and the adsorbed H, O, OH, and H₂O compared with the available data (m for total slab layers and n for relaxed surface layers, mT/nR)

(mT/nR)	Method	E ads	d X–Ni	Ref.
a This work. b On the basis of gaseous H2 and O2.
H (fcc-3F)
4 × 4 (4T/2R)	PBE-D3	−0.62^b (−2.91)	1.708
4 × 4 (4T/2R)	RPBE-D3	−0.59^b (−2.83)	1.708
4 × 4 (4T/2R)	PBE	−0.53^b (−2.65)
2 × 2 (4T/2R)	PBE	−2.65		33
2 × 2 (3T/1R)	RPBE	−2.8		35
2 × 2 (4T/2R)	PBE	−2.77		36
2 × 2 (4T/2R)	PBE	−2.66	1.71	37
2 × 2 (3T/1R)	PBE	−2.78		40
2 × 2 (4T/2R)	PW91	−2.77	0.896	41
2 × 2 (3T/1R)	PBE	−2.80		66
O (fcc-3F)
4 × 4 (4T/2R)	PBE-D3	−2.44^b (−5.78)	1.840
4 × 4 (4T/2R)	RPBE-D3	−2.27^b (−5.47)	1.840
4 × 4 (4T/2R)	PBE	−2.33^b (−5.67)
2 × 2 (4T/2R)	PBE	−5.30		33
2 × 2 (3T/1R)	RPBE	−4.5		35
2 × 2 (4T/2R)	PBE	−4.81		36
2 × 2 (4T/2R)	PBE	−5.31	1.84	37
3 × 3 (3T/1R)	PW91	−2.14^b		39
2 × 2 (3T/1R)	PBE	−5.50		40
2 × 2 (4T/2R)	PW91	−5.43	1.129	41
2 × 2 (3T/1R)	PBE	−5.39		66
OH (fcc-3F)
4 × 4 (4T/2R)	PBE-D3	−3.16	1.971
4 × 4 (4T/2R)	RPBE-D3	−2.61	1.971
4 × 4 (4T/2R)	PBE	−3.07
2 × 2 (4T/2R)	PBE	−3.10		33
2 × 2 (3T/1R)	RPBE	−2.5		35
2 × 2 (4T/2R)	PBE	−3.34		36
2 × 2 (4T/2R)	PBE	−3.08	1.98	37
3 × 3 (3T/1R)	PW91	−3.12		39
2 × 2 (3T/1R)	PBE	−3.14		40
2 × 2 (4T/2R)	PW91	−3.06	1.360	41
2 × 2 (3T/1R)	PBE	−3.27		66
H2O (T)
4 × 4 (4T/2R)	PBE-D3	−0.42	2.156
4 × 4 (4T/2R)	RPBE-D3	−0.52	2.156
4 × 4 (4T/2R)	PBE	−0.25
3 × 3 (3T/2R)	PBE	−0.40		32
2 × 2 (4T/2R)	PBE	−0.19		33
2 × 2 (4T/2R)	PBE	−0.47		36
2 × 2 (4T/2R)	PBE	−0.20	2.23	37
3 × 3 (4T/2R)	PBE	−0.16	2.17	38
3 × 3 (3T/1R)	PW91	−0.29	2.19	39
2 × 2 (3T/1R)	PBE	−0.25	2.24	40
2 × 2 (4T/2R)	PW91	−0.04	2.216	41
3 × 3 (4T/2R)	RPBE	−0.03	2.81	30
3 × 3 (4T/2R)	PW91	−0.24	2.25	30
2 × 2 (3T/1R)	PBE	−0.27		66

---

As shown in Fig. 1, H₂O adsorption prefers the top site with an adsorption energy of −0.42 eV, and the adsorption of H, O and OH prefers the fcc-3F site with an adsorption energy of −0.62, −2.44 and −3.16 eV, respectively. It is noted that OH adsorbs vertically to the surface and H₂O adsorbs nearly parallel to the surface plane. As shown in Table 2, different methods and models give different adsorption energies for each surface species despite the same adsorption site and configuration. We found that the computed H₂O adsorption including dispersion correction agrees with the experimentally determined value (−0.42 vs. −0.44,²² −0.42,²³ −0.54²⁵ and −0.56 eV²⁹). We further computed H₂O adsorption using the RPBE method including dispersion correction with the PBE structure, and the estimated H₂O adsorption is −0.52 eV, more close to the experimental values. For H₂ adsorption, only the PBE computed adsorption energy agrees very well with the experimental values (−1.07 vs. −1.0,⁶⁷ −0.90,⁶⁸ −0.84 to −1.0,⁶⁹ and −0.92 eV⁷⁰), while the PBE-D3 and RPBE-D3 computed values (−1.24 and −1.18 eV, respectively) are slightly higher.

(b) Surface H₂O aggregation

Due to the interaction of H₂O and Ni(111) as well as H-bonding among adsorbed H₂O molecules, we used the stepwise adsorption energy to find the most stable adsorption configurations of (H₂O)_n, where one H₂O molecule was added to the previous one which is the most stable after considering both interactions. The most stable adsorption configurations of (H₂O)_n (n = 1–8) by using PBE-D3 are given in Fig. 2 and the relevant adsorption energies and bond parameters are listed in the ESI† (Table S1). The adsorption energies by using RPBE-D3 of (H₂O)_n (n = 1–8) are given in the ESI† (Fig. S1).

Fig. 2 Most stable adsorption configuration and adsorption energy (eV, PBE-D3) of (H₂O)_n clusters on Ni(111) (Ni/blue; O/red; H/white; small O atoms for shorter Ni–O distances and large O atoms for longer Ni–O distances).

For 2H₂O co-adsorption, one H₂O prefers the top site with an O–Ni distance of 2.061 Å, shorter than that of single H₂O adsorption (2.156 Å), and the second adsorbed H₂O has a very long O–Ni distance (3.031 Å). The first adsorbed H₂O provides a H atom and the second adsorbed H₂O provides an O atom for H-bonding (1.687 Å). The total adsorption energy is −1.11 eV, larger than double that of single H₂O adsorption (−0.84 eV). The average adsorption energy is −0.56 eV, and the stepwise adsorption energy is −0.69 eV. To verify the synergy effect of both interactions, we computed the gas phase single-point energies of the co-adsorbed 2H₂O as well as the two individual H₂O molecules from their co-adsorbed states. It is found that the resulting H-bonding energy is −0.19 eV, weaker than the stepwise adsorption energy by 0.50 eV. This indicates that the stepwise adsorption energy (−0.69 eV) comes from the synergy effect of both interactions.

For 3H₂O co-adsorption in a bent shape, the middle H₂O adsorbs at the top site and provides both H atoms for H-bonding. The O–Ni distance is 2.005 Å, shorter than that of 2H₂O and H₂O adsorption (2.061 and 2.156 Å, respectively), and the O–Ni distance of the other two H₂O molecules is 3.028 and 3.026 Å, respectively. The H-bonding distances are 1.727 Å, longer than that of the co-adsorption of 2H₂O (1.687 Å). The total adsorption energy is −1.77 eV, larger than triple that of single H₂O adsorption (−1.26 eV). The average adsorption energy is −0.59 eV, and the stepwise adsorption energy is −0.66 eV, close to that for the second H₂O stepwise adsorption.

For 4H₂O co-adsorption, the most stable adsorption configuration has a star-like structure, in which the middle H₂O adsorbs at the top site with an O–Ni distance of 2.101 Å and provides both H atoms for H-bonding (1.668 Å) to two H₂O molecules. In addition, the O–Ni distance of these two H₂O molecules is 2.910 and 2.924 Å, respectively. The last H₂O molecule provides a H atom for H-bonding to the central H₂O molecule (1.678 Å) and the O–Ni distance is 2.065 Å. The total adsorption energy is −2.50 eV, larger than four-fold that of single H₂O adsorption (−1.68 eV). The average adsorption energy is −0.63 eV, and the stepwise adsorption energy is −0.73 eV. This shows that two H₂O molecules, which provide H atoms for H-bonding, interact directly with surface Ni atoms, and the other two H₂O molecules, which provide O atoms for H-bonding, are stabilized by H-bonding.

On the basis of these results, we build large H₂O clusters by considering O–Ni and H-bonding interactions. For the co-adsorption of 5H₂O, the most stable adsorption configuration has a cyclic structure, where one H₂O molecule provides both H atoms for H-bonding (1.681 and 1.698 Å) with two neighboring H₂O molecules and the O–Ni distance is 1.990 Å. In turn, these two neighboring H₂O molecules interact with their next neighboring H₂O molecules via H-bonding (1.919 and 1.859 Å). The last two H₂O molecules also interact via H-bonding (1.749 Å). It is noted that the last H₂O has one O–H bond to the surface rather than the O atom. The total adsorption energy is −3.10 eV, the stepwise adsorption energy is −0.60 eV and the average adsorption energy is −0.62 eV.

For the co-adsorption of 6H₂O, we computed a hexagonal adsorption configuration with three short (1.567 Å) and three long H-bonds (1.796, 1.798 and 1.797 Å). Each H₂O molecule adsorbs at the top site, and the O–Ni distances are 2.149, 2.150, 2.151, 2.984, 2.972, and 2.978 Å, indicating that there are three strong and three weak O–Ni interactions. The total adsorption energy is −3.88 eV, the stepwise adsorption energy is −0.78 eV and the average adsorption energy is −0.65 eV. On the basis of the cyclic (H₂O)₅, we computed another (H₂O)₆ structure with an exocyclic H₂O and the total adsorption energy is −3.87 eV. This indicates that the adsorption energy is not structure sensitive for large adsorbed clusters. This is probably due to the same number of Ni–O and H-bonding interactions. Indeed, ring-like hexamer D₂O clusters on Ni(111) were observed in wide coverage ranges at 20 K by infrared reflection absorption spectroscopy.⁷¹ In addition, we computed the O–H stretching and bending frequencies for (H₂O)_n (Table S6†). It is found that the O–H stretching frequencies are shifted to lower wavenumbers with the formation of H-bonding. Taking the (H₂O)₆ ring hexamer as an example, the O–H stretching frequencies are in the range of about 3650 to 2560 cm⁻¹, and this differs strongly from the monomeric H₂O adsorption at 3700 and 3584 cm⁻¹. In contrast, the O–H bending frequencies are shifted to higher wavenumbers compared with that of single H₂O adsorption (1635–1574 vs. 1556 cm⁻¹).

On the basis of the hexagonal structure of (H₂O)₆, we added next exocyclic H₂O molecules. The two (H₂O)₇ structures have close adsorption energies (−4.57 and −4.46 eV, respectively), the stepwise adsorption energy is −0.69 and −0.58 eV, respectively, and the average adsorption energy is −0.65 and −0.64 eV, respectively. For the monolayer (H₂O)₈, on the basis of the more stable (H₂O)₇ with another exocyclic H₂O molecule, the total adsorption energy is −5.26 eV, the stepwise adsorption energy is −0.69 eV and the average adsorption energy is −0.66 eV.

On the basis of these results, one can expect that H₂O adsorption on Ni(111) can have simple structures at very low coverage and very complex structures at high coverage; both direct O–Ni and H-bonding synergistically determine the total adsorption energy. It is also interesting to note that the stepwise adsorption energy has an average value of −0.69 eV.

(c) Surface H₂O dissociation

As listed in Table 1, there are many computational studies about H₂O dissociative adsorption on Ni(111). Although all these studies gave the same qualitative trend, they differ quantitatively from different methods and models. For example, some studies showed that the first step H–O dissociation has a lower barrier than the second step, while some studies gave disordered results. In addition, the computed dissociation energies also differ strongly. The optimized geometries of the initial, transition and final states are shown in Fig. 3. Relevant structural parameters are listed in Table 2. The reaction barriers, reaction energies and relevant structural parameters of the transition state are listed in the ESI† (Table S3). The potential energy surface is shown in Fig. 4.

Fig. 3 Adsorption configuration and energy (eV, PBE-D3) of H₂O dissociation (Ni/blue; O/red; H/white).

Fig. 4 Potential energy surfaces for H₂O and (H₂O)₂ dissociative adsorption on Ni(111) using PBE-D3 (the barrier of elementary steps in parentheses, s for surface species, g for gas phase species).

Starting with H₂O adsorbed at the top site, the first H–O dissociation has a barrier of 0.68 eV and is exothermic by 0.56 eV. In the transition state, the breaking O–H distance is 1.544 Å. In the final state, both OH and H are located at the fcc-3F site. Starting with the co-adsorbed OH + H, the second H–O dissociation has a barrier of 1.25 eV and is exothermic by 0.18 eV. This indicates that the second H–O dissociation has a higher barrier than the first one and is less exothermic than the first one. The total reaction is exothermic by 1.16 eV on the basis of gaseous H₂O or exothermic by 0.74 eV on the basis of adsorbed H₂O.

By consideration of gaseous H₂O (Fig. 4), both steps have very close apparent barriers (0.26 and 0.27 eV, respectively). On the basis of the adsorbed H₂O and O + 2H, gaseous H₂ evolution is endothermic by 0.27 and 1.01 eV, respectively. However, gaseous H₂ evolution becomes exothermic by 0.15 eV on the basis gaseous H₂O. Experimentally, it is found that D₂O forms adsorbed clusters at first at 80 K and then dissociates at 165 K; the dissociation temperature becomes high with an increase in coverage.⁷¹

Due to the H-bonding in co-adsorbed (H₂O)₂, we computed the H-bonding mediated H₂O dissociation. The optimized structures of the initial, transition and final states are shown in the ESI† (Fig. S2). Relevant structural parameters are listed in Table S2.† The reaction barriers and reaction energies as well as relevant structural parameters of the transition state are listed in the ESI† (Table S3). The potential energy surface is shown in Fig. 4.

At first, we computed the dissociation of the H₂O molecule with the longer Ni–O bond distance. The first H–O dissociation has a barrier of 0.68 eV and is exothermic by 0.25 eV. In the transition state, the breaking O–H distance is 1.475 Å. For the first H–O dissociation of the H₂O molecule with the shorter Ni–O bond distance, the barrier is 0.95 eV and the reaction is exothermic by 0.26 eV. The breaking O–H distance in the transition state is 1.348 Å.

In the second H–O dissociation, we considered two competitive reactions. The reaction OH + H + H₂O = O + 2H + H₂O has a barrier of 0.84 eV and is exothermic by 0.40 eV. In the transition state, the breaking O–H distance is 1.478 Å, and the forming H–Ni distance is 1.528 Å. The reaction OH + H + H₂O = 2OH + 2H has a barrier of 0.69 eV and is exothermic by 0.22 eV. In the transition state, the breaking O–H distance is 1.512 Å. As shown in Fig. 4, the co-adsorbed H₂O molecule did not significantly affect the first step dissociation barrier of the H₂O molecule, but obviously boosted the dissociation barrier of OH. This is due to the H-bonding (1.687 Å) in the transition states. Therefore, we computed the surface O promoted H₂O dissociation.

(d) Surface O assisted H₂O dissociation

To study the effect of surface O on H₂O dissociation, the H₂O dissociative adsorption on the O pre-covered Ni(111) surface is computed with a sequential increase of the H₂O molecule and the evolution of gaseous H₂, nO + H₂O(g) = (n + 1)O + H₂(g) (n = 1–4). The optimized configurations for the stationary points of the initial, transition and finial states (Fig. S3–S6†), as well as the structural parameters of the initial and final states are listed in the ESI† (Table S4). The energy barriers, reaction energies and related critical bond distances of the transition states are listed in the ESI† (Table S5). The whole potential energy surface is shown in Fig. 5 and the ESI† (Fig. S7).

Fig. 5 H₂O dissociation on nO (n = 1–2) pre-covered Ni(111) using PBE-D3 (the barrier of elementary steps in parentheses, s for surface species, g for gas phase species).

At first, we computed the H₂O dissociative adsorption on one oxygen pre-covered Ni(111) surface (0.0625 ML), which is close to the oxygen coverage used in the experiment (θ ≤ 0.05 ML).²² Starting with the co-adsorption of one adsorbed surface O atom and one gaseous H₂O molecule, the adsorption energy of H₂O is −0.65 eV, higher than that on the clean surface (−0.42 eV); this increase of adsorption is due to the enhanced H-bonding (1.908 Å). For H₂O dissociation [O + H₂O = 2OH], the barrier is 0.46 eV and the reaction is endothermic by 0.14 eV. In the transition state, the breaking O–H distance is 1.558 Å. For the reverse reaction [2OH = H₂O + O], the barrier is 0.32 eV and the reaction is exothermic by 0.14 eV. Similar results were also reported by Wang et al.,⁴²i.e., the reaction has a barrier of 0.65 eV and is endothermic by 0.29 eV. This indicates that H₂O adsorption on the O pre-covered surface does not prefer dissociation; applying the energy difference of 0.14 eV gives a ratio much higher than 99.99% in preferring molecular H₂O, and this agrees with the XPS results of Pache et al.²⁴ and Schulze et al.,²⁵ where the adsorbed H₂O does not dissociate and remains in its molecular state, while it disagrees with the proposal of Madey et al.,²² where the adsorbed dissociates into surface OH.

For the reaction [2OH = O + H + OH], the barrier is 0.99 eV and the reaction is exothermic by 0.15 eV. The breaking O–H distance in the transition state is 1.490 Å. For the next dissociation [O + H + OH = 2O + 2H], the barrier is 1.16 eV and the reaction is nearly thermal neutral (−0.04 eV). In the transition state, the breaking O–H distance is 1.486 Å. Finally, we computed the H₂ desorption energy, which is exothermic by 0.16 eV on the basis of gaseous H₂O [O + H₂O(g)], and endothermic by 0.49 eV on the basis of adsorbed H₂O [O + H₂O(s)]. On the basis of co-adsorbed O and H [2O + 2H], the computed H₂ desorption energy is 0.54 eV.

For H₂O dissociation on the nO pre-covered Ni(111) surface (n = 2–4), the potential energy surface is similar to that of O assisted H₂O dissociation (Fig. S3–S6†). The H₂O adsorption energies are in the range of −0.65 to −0.67 eV, very close to that of n = 1 (−0.65 eV). For the first step dissociation, nO + H₂O = (n − 1)O + 2OH, the barrier is 0.37, 0.45 and 0.46 eV, respectively, and the reaction is endothermic by 0.08, 0.18 and 0.14 eV, respectively, in favor of molecular adsorption instead of dissociative adsorption of H₂O. In the transition state, the breaking O–H distance is 1.530, 1.698 and 1.507 Å, respectively. For the second step, (n − 1)O + 2OH = nO + H + OH, the barriers are around 0.90–0.99 eV, and the reaction is nearly thermal neutral (−0.06 eV) for n = 2, and slightly endothermic (0.18 and 0.09 eV, respectively) for n = 3 and 4. For the last step dissociation, nO + H + OH = (n + 1)O + 2H, the barrier is 1.19, 1.29 and 1.19 eV, respectively, and the reaction is endothermic by 0.06, 0.12 and 0.21 eV, respectively. All these show that higher O pre-coverage does not promote H₂O dissociative adsorption, and the adsorbed H₂O remains in the molecular state.

(e) High coverage adsorption of OH and O on Ni(111)

The most stable OH adsorption configurations at different coverages on Ni(111) are shown in Fig. 6a. For 2OH, we computed two adsorption configurations (Fig. S8†): a remote one and an adjacent one, and both have OH perpendicularly adsorbed at the fcc-3F sites without H-bonding. In addition, the remote one is more stable than the adjacent one by 0.17 eV. For 4OH (0.25 ML), we computed five combined adsorption configurations (Fig. S9†) and found the remote regularly distributed p(2 × 2) configuration with all OH perpendicularly adsorbed at the fcc-3F sites without H-bonding to be the most stable, and the linear one with all OH at the bridge sites with H-bonding to be the least stable; the energy difference between these two adsorption configurations is 0.41 eV, indicating the strong lateral repulsive interaction despite the H-bonding in the linear one. The energies of the other configurations are in between.

Fig. 6 The most stable geometries and adsorption energies (eV, PBE-D3) of different coverages of (a) OH and (b) O on the Ni(111) surface (Ni/blue; O/red; H/white).

For 6OH, we computed eight adsorption configurations by considering the remote adsorption and H-bonding (Fig. S10†). The most stable one has 4OH at the bridge sites that form a line with H-bonding (1.640, 1.637 and 1.637 Å) and the other two OH are remote at two fcc-3F sites. Several other configurations are less stable by less than 0.1 eV. This might reveal the flexibility of the adsorption configurations at high OH coverage, and random adsorption under given conditions becomes possible.

Next, we computed six adsorption configurations for 8OH (0.5 ML, Fig. S11†). In contrast to the expected regular structures, the most stable one has all OH at the bridge sites and each OH interacts with another OH via H-bonding. The least stable one with one 4OH line and one 3OH line at the bridge site via H-bonding with another one OH at the fcc-3F site is less stable by 0.40 eV. Despite the repulsive interaction among the adsorbed OH groups at high coverage, we computed the adsorption configuration for 10OH with two parallel 4OH lines and one 2OH line in between, for 12OH (0.75 ML) with three parallel 4OH lines and 16OH (1 ML) with four parallel 4OH lines, where the H-bonding is between two OH lines instead within one line (Fig. S12†).

The average adsorption energy decreases with the increase of OH coverage from −3.16 eV (OH), −3.23 eV (2OH) to −2.30 (16OH). On the basis of gaseous H₂O [nH₂O(g) = nOH(s) + (n/2)H₂(g)], the adsorption of 8OH (0.5 ML) and 10OH (0.625 ML) is exothermic by 2.10 eV and 0.69 eV, respectively, while the adsorption of 12OH (0.75 ML) and 16OH becomes endothermic by 0.79 and 8.37 eV, respectively. This indicates that the saturation coverage of OH adsorption should be 0.625 ML on the basis of H₂O dissociative adsorption, and it is not possible to get 1 ML OH coverage.

The most stable adsorption configurations of O atoms with different coverages on Ni(111) are shown in Fig. 6b; all O atoms are adsorbed at fcc-3F sites. The adsorption energy of one O atom is −2.44 eV, and that of the remote 2O is higher than that of the adjacent 2O (−4.89 vs. −4.55 eV), indicating the lateral repulsive interaction between two adjacent O atoms by 0.34 eV, and this energy difference is much higher than that of 2OH adsorption (0.17 eV). It is noted that the adsorption energy of the remote 2O is double that of single O adsorption (−4.89 eV); therefore we optimized high coverage O adsorption with remote O atoms. For 4O adsorption (0.25 ML), the adsorption energy (−9.79 eV) is four-fold that of single O adsorption (−9.74 eV). This shows that there is no lateral repulsive interaction among these 4O atoms at 0.25 ML, and the adsorption configuration has a p(2 × 2) structure, in agreement with the experimental observation.⁷² On the basis of the 4O adsorption configuration, we computed 6O adsorption and the adsorption energy is −12.97 eV, which is lower than that of six-fold of single O adsorption (−14.64 eV), indicating the lateral repulsive interaction by 1.67 eV. Since high coverage O adsorption will result in surface oxidation and reconstruction,⁷² we did not consider even higher coverage O adsorption. For comparison we also computed the most stable configuration for n = 8, 12, 16 in the ESI† (Fig. S13); the repulsive interaction is 3.81, 11.44 and 23.96 eV, respectively.

On the basis of gaseous H₂O [nH₂O(g) = nO(s) + nH₂(g)], the adsorption is exothermic by 0.14, 0.31 and 0.61 eV for nO (n = 1, 2, 4), respectively, while endothermic by 0.80 eV for n = 6. Therefore, the oxygen saturation coverage on the basis of gaseous H₂O should be 4O (0.25 ML).

(f) H₂O desorption on Ni(111)

On the basis of the most stable adsorption structures, we computed the H₂O desorption temperature under certain pressure (1 × 10⁻⁵ mbar²⁵ for H₂O in the experiment) by using ab initio atomistic thermodynamics. Table 3 lists the experimentally estimated H₂O desorption temperature, and the computed H₂O desorption temperatures are shown in Table 4.

Table 3 H₂O desorption temperature in experiments

mbar	T des (K)	θ O (ML)	Ref.
	165–170	0	22
1.3 × 10^−10	168–171	0	23
2 × 10^−10	152–171	0	24
10^−7–10^−5	155–175	0	25
	180–200; 275–300	0.05	22
1.3 × 10^−10	180–260	0.07	23
2 × 10^−10	180–260	0.25	24
10^−7–10^−5	180–260	0.25	25

---

Table 4 Computed H₂O desorption temperature (T_des, K; 10⁻⁵ mbar)

Reaction	PBE-D3	RPBE-D3
H2O(s) = H2O(g)	140	174
2OH(s) = O(s) + H2O(g)	191	60
(H2O)n(s) = nH2O(g) (n = 1–8)	184–211	196–214
nO + H2O(s) = nO + H2O(g) (n = 1–4)	212–219	247–259
4O + 4H2O(s) = 4O + 4H2O(g)	245	284
4O + 4H2O(s) = 4O + 3H2O(s) + H2O(g)	237	283
4O + 3H2O(s) = 4O + 2H2O(s) + H2O(g)	236	279
4O + 2H2O(s) = 4O + H2O(s) + H2O(g)	254	291
4O + H2O(s) = 4O + H2O(g)	254	282
8OH = 4O + 4H2O(g)	129	197

---

For the desorption of an adsorbed single H₂O molecule [H₂O(s) = H₂O(g)], the calculated desorption temperature is 140 K by using PBE-D3 adsorption energy and 174 K by using RPBE-D3 adsorption energy; the latter is closer to the detected 163,²⁵ 164,²⁴ 165,²² and 168 K²³ at low coverage. In addition, we computed the desorption temperature of the adsorbed (H₂O)_n clusters, which is in the range of 184–211 K by using PBE-D3 and 196–214 K by using RPBE-D3.

For H₂O desorption from OH disproportionation [2OH → O(s) + H₂O(g)], the calculated desorption temperature is 191 K by using PBE-D3 and 60 K by using RPBE-D3; both are far away from the detected range of 275–300 K. This shows that H₂O does not come from OH disproportionation. Indeed, UPS and XPS studies did not support the supposed formation of surface OH from H₂O adsorption on the O pre-covered surface. Using the adsorption energy of H₂O on the O pre-covered surface [O + H₂O(s)], however, the computed H₂O desorption temperature is 212 K by using PBE-D3 and 247 K by using RPBE-D3; the latter is closer to the experimental value than the former.

Since the experimental study used the 0.25 ML O pre-covered p(2 × 2) surface and the adsorbed H₂O has 0.25 ML, we optimized the proposed structure of Pache et al.,²⁴ where each surface O interacts with two adjacent H₂O molecules via H-bonding in a zigzag manner, and H₂O molecules adsorbed at hollow sites (Fig. 7a). The computed average adsorption energy is −0.35 and −0.39 eV by using PBE-D3 and RPBE-D3, respectively, and the corresponding desorption temperature is 123 and 135 K, respectively. They disagree with the experiment because H₂O molecules have very weak interaction with the surface.

Fig. 7 H₂O adsorption (0.25 ML) on 0.25 ML O pre-covered Ni(111): (a) H₂O at the hollow site; (b) H₂O at the top site.

By considering both interaction of H₂O with the surface and H-bonding with surface O, we designed a new structure, where H₂O molecules are adsorbed at top sites and have H-bonding with surface O atoms (Fig. 7b). The computed average adsorption energy is −0.76 and −0.90 eV by using PBE-D3 and PRBE-D3, respectively. In turn, the computed corresponding desorption temperature is 245 and 284 K, respectively, which are much closer to the experimental values,²⁴ in particular, the one by using RPBE-D3. Our new structure with adsorbed H₂O at the top sites is supported by X-ray diffraction and infrared reflection absorption spectroscopy.⁷³

In addition to the concerted desorption of all 4H₂O on the 4O pre-covered surface, we computed H₂O stepwise desorption, and the desorption temperature is 254, 254, 236, and 237 K by using PBE-D3 and 282, 291, 279, and 283 K by using RPBE-D3, indicating that it is possible for H₂O desorption to proceed in either a stepwise or a concerted way, since the desorption temperature is 275–300 K on the 0.25 ML O pre-covered p(2 × 2) surface and the 0.25 ML adsorbed H₂O molecule.²⁴

Recently, Zhao et al.²⁹ reported a single crystal adsorption calorimetry study of D₂O adsorption on clean and 0.25 ML O pre-covered Ni(111) surfaces. This is very related to our study of H₂O adsorption, particularly, with the co-adsorption of 0.25 ML O (4O) and 0.25 ML H₂O (4H₂O) on a p(4 × 4) slab model. This provides a chance for direct comparison between experimental and DFT computation.

At first, the computed average H₂O adsorption energy at 0.5 ML coverage on the clean Ni(111) surface (−0.66 and −0.65 eV by using PBE-D3 and RPBE-D3, respectively) is slightly larger than the determined value (−0.56 eV) for D₂O adsorption at 0.5 ML coverage and at 100 K by about 0.1 eV.

Next, we computed the OH average bond energy at 0.5 ML OH coverage (8OH), which is 3.09 and 3.06 eV using PBE-D3 and RPBE-D3, respectively, lower than the estimated value for D₂O at 170 K (3.26 eV) by 0.17–0.20 eV. Further, we estimated the surface OH enthalpy of formation on the basis of 4H₂(g) + 4O₂(g) = 8OH, and the computed value is −5.11 and −5.09 eV, respectively, using PBE-D3 and RPBE-D3, also lower than the estimated −5.77 eV. All these show a difference of 0.66–0.68 eV.

Since the computed OH bond energy and enthalpy of formation differ from the experimental values on the basis of the supposed OH formation, we are wondering if the adsorbed H₂O prefers molecular adsorption over dissociative adsorption. On the basis of the molecular adsorbed state [4O(s) + 4H₂O(s)], the dissociation of surface adsorbed H₂O [4O(s) + 4H₂O(s) = 8OH(s)] is endothermic by 1.57 and 1.19 eV, respectively, using PBE-D3 and RPBE-D3, and the corresponding stoichiometric value is endothermic by 0.39 and 0.30 eV, respectively. This indicates that H₂O dissociative adsorption is unlikely, and the expected equilibrium greatly favors H₂O molecular adsorption (≫99.99%) over dissociation adsorption. This agrees with the UPS results of Pache et al.²⁴ and XPS results of Schulze et al.,²⁵ where the adsorbed H₂O does not dissociate and remains in its molecular state, while it disagrees with the proposal of Madey et al.,²² where the adsorbed dissociates into surface OH.

Taking the co-adsorbed 4O(s) + 4H₂O(s) as a reference, the supposed surface hydroxyls have an enthalpy of formation from 4H₂(g) + 4O₂(g) of −5.51 and −5.38 eV, respectively, using PBE-D3 and RPBE-D3, which are closer to the estimated −5.77 eV. Further taking the adsorbed 4O(s) + 4H₂O(s) as a reference, the supposed surface hydroxyls have a bond energy of 3.28 and 3.21 eV, respectively, using PBE-D3 and RPBE-D3, which are also closer to the estimated value of 3.26 eV at 170 K.

Since the computed bond energy and enthalpy of formation of surface hydroxyls support the molecular and non-dissociative adsorption of H₂O on the O pre-covered surface, we computed the desorption temperature of the molecularly adsorbed H₂O [4O(s) + 4H₂O(s) = 4O(s) + 4H₂O(g)] (Table 4). Indeed, the computed H₂O desorption temperature agrees excellently with the experimentally determined range (284 vs. 275–300 K). In contrast, the calculated desorption temperature of H₂O from surface hydroxyl disproportionation [8OH = 4O(s) + 4H₂O(g)] is much lower (129 and 197 K). These results indicate once again the molecular state of the adsorbed H₂O on the 0.25 ML O pre-covered surface.

All these results prefer H₂O molecular adsorption on the 0.25 ML O pre-covered Ni(111) surface, and this is in agreement with the UPS and XPS studies,^24,25 while in disagreement with the recently supposed H₂O dissociative adsorption from the single crystal adsorption calorimetry study²⁹ and previous results.^22,23 This disagreement encourages further experimental investigations and confirmations.

On the basis of the disagreement in dissociative and molecular adsorption of H₂O on the 0.25 ML O pre-covered surface, we computed vibrational frequencies to aid further experimental investigation (Table S6†). On the surface with the increase of OH coverage up to 0.25 ML, where all adsorbed OH species have vertical configurations, the O–H vibrational frequencies are about 3700 cm⁻¹. For 0.5 ML OH coverage (8OH), where the adsorbed OH species interact via H-bonding, the O–H vibrational frequencies are shifted to lower wavenumbers, i.e., in the range of 3642–3422 cm⁻¹, lower than those in vertical adsorption configurations. For the co-adsorbed 4O + 4H₂O, the O–H vibrational frequencies are in the range of 3552–3440 cm⁻¹, although somewhat lower than those of the OH that interact via H-bonding in the high wavenumber region, they also exhibit an overlap in the low wavenumber region. The characteristic feature of molecular H₂O adsorption can come from the O–H bending frequencies in the range of 1551–1537 cm⁻¹. All these need experimental investigation and confirmation.

Conclusion

In this work, we investigated the dissociative adsorption of H₂O on clean and oxygen pre-covered Ni(111) surfaces on the basis of density functional theory and ab initio atomistic thermodynamics systematically. Due to the importance of water chemistry in Ni-based catalysts, there are intensive experimental and theoretical investigations. However, there are discrepancies among the experimental results and also differences among computational results, where different computational models and methods have different results. In addition, there are deficits in understanding H₂O dissociative adsorption regarding hydrogen-bonding interaction at high coverage.

On the clean Ni(111) surface, the most stable adsorption configuration of H, O, and OH prefers the 3-fold face centered-cubic site and that of H₂O prefers the top site. The computed adsorption energies of H₂O by using PBE and RPBE including dispersion are close to the reported experimental results, while that of H₂ using only PBE is closer to the experimental value than the corresponding PBE and PPBE values including dispersion.

For the adsorption of (H₂O)_n clusters (n = 2–8), both direct O–Ni and H-bonding synergistically determine the total adsorption energy. H₂O adsorption on Ni(111) have simple structures at very low coverage, and very complex structures at high coverage, and the adsorption energy is structurally insensitive for large adsorbed clusters. The stepwise adsorption energy has an average value of −0.69 eV.

At low coverage (θ ≤ 0.25 ML), OH adsorbs perpendicularly and prefers remote distribution without H-bonding among OH groups, the OH saturation coverage is 0.625 ML on the basis of H₂O dissociative adsorption and H₂ evolution [nH₂O(g) = nOH(s) + (n/2)H₂(g)] and it is not possible to get 1 ML OH coverage. The adsorption configuration for 4O (0.25 ML) prefers a p(2 × 2) structure, in agreement with the experiment, and the O saturation coverage is 0.25 ML on the basis of H₂O dissociative adsorption and H₂ evolution [nH₂O(g) = nO(s) + nH₂(g)].

For one H₂O dissociative adsorption, the first H–O dissociation is favored both thermodynamically and kinetically, but the second H–O dissociation has a higher barrier and is less exothermic than the first one. On the basis of the gaseous H₂O, both steps have close apparent barriers (0.26 and 0.27 eV, respectively). For (H₂O)₂ dissociation, the co-adsorbed H₂O molecules do not significantly affect the first step dissociation barrier of the H₂O molecule, but obviously boost the dissociation barrier of OH due to the H-bonding in the transition states.

The co-adsorption of O and H₂O each at 0.25 ML prefers a p(2 × 2) structure, in which top site H₂O interacts with two hollow site oxygen atoms via H-bonding, which is energetically more stable than the previously proposed one by 0.51 eV; in the latter, the adsorbed H₂O molecules are proposed to be over the hollow sites.

On the 0.25 ML O pre-covered Ni(111), H₂O greatly prefers molecular adsorption [4O + 4H₂O(g) → 4O + 4H₂O(s)] over dissociative adsorption [4O + 4H₂O(g) → 8OH] thermodynamically by 0.30 eV, and this is in agreement with the UPS and XPS results and in disagreement with previous results as well as recent results from the single crystal adsorption calorimetry study of D₂O.

It is noted that the computed bond energy (3.06 eV) and formation enthalpy (−5.09 eV) of the supposed surface hydroxyls on the basis of the dissociatively adsorbed state [8OH] differ from the experimentally estimated values (3.26 and −5.77 eV, respectively), while those (3.21 and −5.38 eV, respectively) on the basis of the molecular adsorbed state [4O + 4H₂O(s)] are closer to the experimentally obtained values. This further supports H₂O molecular adsorption.

The definitive support of H₂O molecular adsorption [4O + 4H₂O(s)] comes from the computed H₂O desorption temperature, which is in excellent agreement with the experimental result (284 vs. 275–300 K), while that from the dissociatively adsorbed state [8OH] differs strongly (197 K).

All these support H₂O molecular adsorption instead of dissociative adsorption, which needs further experimental investigations and confirmations. Despite these results, the vibrational frequencies of the adsorbed OH species for dissociative adsorption as well as molecularly adsorbed H₂O on the 0.25 ML O pre-covered surface have been computed to aid experimental investigations. The characteristic difference between the molecular and dissociative H₂O adsorption comes from the O–H bending of molecularly adsorbed H₂O in the range of 1551–1537 cm⁻¹. Our systematic study provides insights into the water-involved reactions catalyzed by nickel particularly and broadens our fundamental understanding of water interaction with metal surfaces generally.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (no. 21673268) and the Chinese Academy of Sciences and Synfuels CHINA. Co., Ltd. We also acknowledge general financial support from the BMBF and the state of Mecklenburg-Vorpommern.

References

1. J. Carrasco, A. Hodgson and A. Michaelides, A Molecular Perspective of Water at Metal Interfaces, Nat. Mater., 2012, 11, 667–674.
2. K.-R. Hwang, C.-B. Lee and J.-S. Park, Advanced Nickel Metal Catalyst for Water–Gas Shift Reaction, J. Power Sources, 2011, 196, 1349–1352.
3. A. Hodgson and S. Haq, Water Adsorption and the Wetting of Metal Surfaces, Surf. Sci. Rep., 2009, 64, 381–451.
4. J. R. Rostrup-Nielsen, Activity of Nickel Catalysts for Steam Reforming of Hydrocarbons, J. Catal., 1973, 31, 173–199.
5. P. Sabatier and J. Senderens, New synthesis of methane, C. R. Hebd. Seances Acad. Sci., 1902, 134, 514–516.
6. P. Sabatier and J. Senderens, Direct Hydrogenation of Oxides of Carbon in Presence of Various Finely Divided Metals, C.R. Acad. Sci., 1902, 134, 689–691.
7. D. W. Goodman, R. D. Kelley, T. E. Madey and J. T. Yates, Kinetics of the Hydrogenation of CO over a Single Crystal Nickel Catalyst, J. Catal., 1980, 63, 226–234.
8. J. Sehested, S. Dahl, J. Jacobsen and J. R. Rostrup-Nielsen, Methanation of CO over Nickel: Mechanism and Kinetics at High H₂/CO Ratios, J. Phys. Chem. B, 2005, 109, 2432–2438.
9. J. Sehested, Four Challenges for Nickel Steam-Reforming Catalysts, Catal. Today, 2006, 111, 103–110.
10. H. S. Bengaard, J. K. Nørskov, J. Sehested, B. S. Clausen, L. P. Nielsen, A. M. Molenbroek and J. R. Rostrup-Nielsen, Steam Reforming and Graphite Formation on Ni Catalysts, J. Catal., 2002, 209, 365–384.
11. J. G. Xu and G. F. Froment, Methane Steam Reforming, Methanation and Water-Gas Shift: I. Intrinsic Kinetics, AIChE J., 1989, 35, 88–96.
12. K. H. Hou and R. Hughes, The Kinetics of Methane Steam Reforming over a Ni/α-Al₂O catalyst, Chem. Eng. J., 2001, 82, 311–328.
13. H. S. Bengaard, J. K. Nørskov, J. Sehested, B. S. Clausen, L. P. Nielsen, A. M. Molenbroek and J. R. Rostrup-Nielsen, Steam Reforming and Graphite Formation on Ni Catalysts, J. Catal., 2002, 209, 365–384.
14. C. Pistonesi, A. Juan, B. Irigoyen and N. Amadeo, Theoretical and Experimental Study of Methane Steam Reforming Reactions over Nickel Catalyst, Appl. Surf. Sci., 2007, 253, 4427–4437.
15. G. Jones, J. Jakobsen, S. Shim, J. Kleis, M. Andersson, J. Rossmeisl, F. Abildpedersen, T. Bligaard, S. Helveg and B. Hinnemann, First Principles Calculations and Experimental Insight into Methane Steam Reforming over Transition Metal Catalysts, J. Catal., 2008, 259, 147–160.
16. D. W. Blaylock, T. Ogura, W. H. Green and G. J. O. Beran, Computational Investigation of Thermochemistry and Kinetics of Steam Methane Reforming on Ni(111) under Realistic Conditions, J. Phys. Chem. C, 2009, 113, 4898–4908.
17. N. Eranda, S. Johannes and L. Suljo, Comparative Study of the Kinetics of Methane Steam Reforming on Supported Ni and Sn/Ni Alloy Catalysts: The Impact of the Formation of Ni Alloy on Chemistry, J. Catal., 2009, 263, 220–227.
18. D. W. Blaylock, Y. A. Zhu and W. H. Green, Computational Investigation of the Thermochemistry and Kinetics of Steam Methane Reforming over a Multi-Faceted Nickel Catalyst, Top. Catal., 2011, 54, 828–844.
19. J. J. Guo, H. Lou, H. Zhao, D. F. Chai and X. M. Zheng, Dry Reforming of Methane over Nickel Catalysts Supported on Magnesium Aluminate Spinels, Appl. Catal., A, 2004, 273, 75–82.
20. Y. A. Zhu, D. Chen, X. G. Zhou and W. K. Yuan, DFT Studies of Dry Reforming of Methane on Ni Catalyst, Catal. Today, 2009, 148, 260–267.
21. Y. H. Hu and E. Ruckenstein, Catalytic Conversion of Methane to Synthesis Gas by Partial Oxidation and CO₂ Reforming, Adv. Catal., 2004, 48, 297–345.
22. T. E. Madey and F. P. Netzer, The Adsorption of H₂O on Ni(111); Influence of Preadsorbed Oxygen on Azimuthal Ordering, Surf. Sci., 1982, 117, 549–560.
23. R. H. Stulen and P. A. Thiel, Electron-Stimulated Desorption and Thermal Desorption Spectrometry of H, H₂O on Nickel(111), Surf. Sci., 1985, 157, 99–118.
24. T. Pache, H. P. Steinrück, W. Huber and D. Menzel, The Adsorption of H₂O on Clean and Oxygen Precovered Ni(111) Studied by ARUPS and TPD, Surf. Sci., 1989, 224, 195–214.
25. M. Schulze, R. Reißner, K. Bolwin and W. Kuch, Interaction of Water with Clean and Oxygen Precovered Nickel Surfaces, Fresenius' J. Anal. Chem., 1995, 353, 661–665.
26. M. E. Gallagher, S. Haq, A. Omer and A. Hodgson, Water Monolayer and Multilayer Adsorption on Ni(111), Surf. Sci., 2007, 601, 268–273.
27. J. Shan, J. F. M. Aarts, A. W. Kleyn and L. B. F. Juurlink, The Interaction of Water with Ni(111) and H/Ni(111) Studied by TPD and HREELS, Phys. Chem. Chem. Phys., 2008, 10, 2227–2232.
28. J. Shan, A. W. Kleyn and L. B. F. Juurlink, Identification of Hydroxyl on Ni(111), ChemPhysChem, 2009, 10, 270–275.
29. W. Zhao, S. J. Carey, Z. Mao and C. T. Campbell, Adsorbed Hydroxyl and Water on Ni(111): Heat of Formation by Calorimetry, ACS Catal., 2018, 8, 1485–1489.
30. B. Jiang and H. Guo, Towards An Accurate Specific Reaction Parameter Density Functional for Water Dissociation on Ni(111): RPBE Versus PW91, Phys. Chem. Chem. Phys., 2016, 18, 21817–21824.
31. B. Jiang and H. Guo, Dynamics of Water Dissociative Chemisorption on Ni(111): Effects of Impact Sites and Incident Angles, Phys. Rev. Lett., 2015, 114, 166101.
32. Z. Y. Du, Y. X. Ran, Y. P. Guo, J. Feng and W. Y. Li, A Theoretical Study on the Role of Water and its Derivatives in Acetic Acid Steam Reforming on Ni(111), Appl. Surf. Sci., 2017, 419, 114–125.
33. A. Mohsenzadeh, T. Richards and K. Bolton, DFT Study of the Water Gas Shift Reaction on Ni(111), Ni(100) and Ni(110) surfaces, Surf. Sci., 2016, 644, 53–63.
34. D. W. Blaylock, Y. A. Zhu and W. H. Green, Computational Investigation of the Thermochemistry and Kinetics of Steam Methane Reforming over A Multi-Faceted Nickel Catalyst, Top. Catal., 2011, 54, 828–844.
35. D. W. Blaylock, T. Ogura, W. H. Green and G. J. O. Beran, Computational Investigation of Thermochemistry and Kinetics of Steam Methane Reforming on Ni(111) under Realistic Conditions, J. Phys. Chem. C, 2009, 113, 4898–4908.
36. R. C. Catapan, A. A. M. Oliveira, Y. Chen and D. G. Vlachos, DFT Study of the Water−Gas Shift Reaction and Coke Formation on Ni(111) and Ni(211) Surfaces, J. Phys. Chem. C, 2012, 116, 20281–20291.
37. A. Mohsenzadeh, K. Bolton and T. Richards, DFT Study of the Adsorption and Dissociation of Water on Ni(111), Ni(110) and Ni(100) surfaces, Surf. Sci., 2014, 627, 1–10.
38. Y. C. Huang, C. Y. Ling, M. Jin, J. Y. Du, T. Zhou and T. S. Wang, Water Adsorption and Dissociation on Ni surface: Effects of Steps, Dopants, Coverage and Self-Aggregation, Phys. Chem. Chem. Phys., 2013, 15, 17804.
39. A. A. Phatak, W. N. Delgass, F. H. Ribeiro and W. F. Schneider, Density Functional Theory Comparison of Water Dissociation Steps on Cu, Au, Ni, Pd, and Pt, J. Phys. Chem. C, 2009, 113, 7269–7276.
40. M. Pozzo, G. Carlini, R. Rosei and D. Alfè, Comparative Study of Water Dissociation on Rh(111) and Ni(111) Studied with First Principles Calculations, J. Chem. Phys., 2007, 126, 164706.
41. W. Wang and G. Wang, A Theoretical Study of Water Adsorption and Dissociation on Ni(111) Surface During Oxidative Steam Reforming and Water Gas Shift Processes, J. Energy Inst., 2015, 88, 112–117.
42. G.-C. Wang, S.-X. Tao and X.-H. Bu, A Systematic Theoretical Study of Water Dissociation on Clean and Oxygen-Preadsorbed Transition Metals, J. Catal., 2006, 244, 10–16.
43. P. M. Hundt, B. Jiang, M. E. van Reijzen, H. Guo and R. D. Beck, Vibrationally Promoted Dissociation of Water on Ni(111), Science, 2014, 344, 504–507.
44. G. Kresse and J. Furthmüller, Efficiency of Ab-initio Total Energy Calculations for Metals and Semiconductors Using A Plane-Wave Basis Set, Comput. Mater. Sci., 1996, 6, 15–50.
45. G. Kresse and J. Furthmüller, Efficient Iterative Schemes for Ab Initio Total-energy Calculations Using A Plane-Wave Basis Set, Phys. Rev. B: Condens. Matter, 1996, 54, 11169–11186.
46. P. E. Blöchl, Projector Augmented-Wave Method, Phys. Rev. B, 1994, 50, 17953–17979.
47. G. Kresse and D. Joubert, From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775.
48. J. P. Perdew, K. Burke and M. Ernzerhof, Generalized Gradient Approximation Made Simple, Phys. Rev. Lett., 1996, 77, 3865–3868.
49. J. L. F. Da Silva, C. Stampfl and M. Scheffler, Adsorption of Xe Atoms on Metal Surfaces: New Insights from First-Principles Calculations, Phys. Rev. Lett., 2003, 90, 066104.
50. J. L. F. Da Silva, C. Stampfl and M. Scheffler, Xe Adsorption on Metal Surfaces: First-principles Investigations, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 075424.
51. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 elements H-Pu, J. Chem. Phys., 2010, 132, 154104.
52. G. Henkelman, B. P. Uberuaga and H. Jónsson, A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths, J. Chem. Phys., 2000, 113, 9901–9904.
53. H. J. Monkhorst and J. D. Pack, Special Points for Brillouin-Zone Integrations, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
54. A. Taylor, Lattice Parameters of Binary Nickel Cobalt Alloys, J. Jpn. Inst. Met., 1950, 77, 585–594.
55. C. Kitte, Introduction to Solid State Physics, Wiley, New York, 7th edn, 1996.
56. B. Hammer, L. B. Hansen and J. K. Nørskov, Improved Adsorption Energetics within Density-Functional Theory Using Revised Perdew–Burke–Ernzerhof Functionals, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 7413–7421.
57. K. Reuter and M. Scheffler, Composition, Structure, and Stability of RuO₂(110) as a Function of Oxygen Pressure, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 65, 035406.
58. K. Reuter and M. Scheffler, Composition and Structure of the RuO₂(110) Surface in an O₂ and CO Environment: Implications for the Catalytic Formation of CO₂, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 68, 045407.
59. T. Wang, X. W. Liu, S. G. Wang, C. F. Huo, Y. W. Li, J. G. Wang and H. J. Jiao, Stability of β-Mo₂C Facets from Ab Initio Atomistic Thermodynamics, J. Phys. Chem. C, 2011, 115, 22360–22368.
60. F. Zasada, W. Piskorz, S. Cristol, J. F. Paul, A. Kotarba and Z. Sojka, Periodic Density Functional Theory and Atomistic Thermodynamic Studies of Cobalt Spinel Nanocrystals in Wet Environment: Molecular Interpretation of Water Adsorption Equilibria, J. Phys. Chem. C, 2010, 114, 22245–22253.
61. T. Wang, S. G. Wang, Y. W. Li, J. Wang and H. J. Jiao, Adsorption Equilibria of CO Coverage on β-Mo₂C Surfaces, J. Phys. Chem. C, 2012, 116, 6340–6348.
62. W. X. Li, C. Stampfl and M. Scheffler, Insights into the Function of Silver as an Oxidation Catalyst by Ab Initio Atomistic Thermodynamics, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 68, 165412.
63. W. Piskorz, J. Gryboś, F. Zasada, P. Zapała, S. Cristol, J. F. Paul and Z. Sojka, Periodic DFT Study of the Tetragonal ZrO₂ Nanocrystals: Equilibrium Morphology Modeling and Atomistic Surface Hydration Thermodynamics, J. Phys. Chem. C, 2012, 116, 19307–19320.
64. S. L. Liu, Y. W. Li, J. G. Wang and H. J. Jiao, Mechanisms of H₂O and CO₂ Formation from Surface Oxygen Reduction on Co(0001), J. Phys. Chem. C, 2016, 120, 19265–19270.
65. T. Zeng, X. D. Wen, Y.-W. Li and H. J. Jiao, Density Functional Theory Study of Triangular Molybdenum Sulfide Nanocluster and CO Adsorption on It, J. Phys. Chem. B, 2005, 109, 13704–13710.
66. M. Zhou and B. Liu, First-Principles Investigation of Adsorbate−Adsorbate Interactions on Ni(111), Ni(211), and Ni(100) Surfaces, Ind. Eng. Chem. Res., 2017, 56, 5813–5820.
67. K. Christmann, O. Schober, G. Ertl and M. Neumann, Adsorption of Hydrogen on Nickel Single Crystal Surfaces, J. Chem. Phys., 1974, 60, 4528–4540.
68. K. Christmann, R. J. Behm, G. Ertl, M. A. Van Hove and W. H. Weinberg, Chemisorption Geometry of Hydrogen on Ni(111): Order and Disorder, J. Chem. Phys., 1979, 70, 4168–4184.
69. A. Winkler and K. D. Rendulic, Adsorption Kinetics for Hydrogen Adsorption on Nickel and Coadsorption of Hydrogen and Oxygen, Surf. Sci., 1982, 118, 19–31.
70. Y. Hong and J. L. Whitten, Dissociative Adsorption of H₂ on Ni(111), J. Chem. Phys., 1993, 98, 5039–5049.
71. M. Nakamura and M. Ito, Ring Hexamer Like Cluster Molecules of Water Formed on A Ni(111) Surface, Chem. Phys. Lett., 2004, 384, 256–261.
72. A. R. Kortan and R. L. Park, Phase Diagram of Oxygen Chemisorbed on Nickel (111), Phys. Rev. B: Condens. Matter Mater. Phys., 1981, 23, 6340–6347.
73. M. Nakamura, M. Tanaka, M. Ito and S. Sakata, Water Adsorption on A p(2×2)-Ni(111)-O Surface Studied by Surface X-ray Diffraction and Infrared Reflection Absorption Spectroscopy at 25 and 140 K, J. Chem. Phys., 2005, 122, 224703.

---

Footnote

† Electronic supplementary information (ESI) available: Adsorption energies and structural parameters of (H₂O)_n adsorption (Table S1, Fig. S1); dissociation energetics of H₂O and (H₂O)₂ on the clean surface (Tables S2 and S3; Fig. S2 and S3) as well as on the oxygen pre-covered surface (Tables S4 and S5; Fig. S4–S6); stretching and bending frequencies of 4O, (OH)_n, 4O + 4H₂O and (H₂O)_n on Ni(111) (Table S6); potential energy surface of H₂O dissociative adsorption on 3O and 4O pre-covered surfaces (Fig. S7); high coverage OH and O adsorption configurations and energies (Fig. S8–S12; Fig. S13). See DOI: 10.1039/c8cy02198h

---