Computational and experimental studies on the effect of hydrogenation of Ni-doped TiO₂ anatase nanoparticles for the application of water splitting

Abstract

We have studied theoretically and experimentally the effect of Ni-doping in TiO₂ nanoparticles (NPs) on hydrogenation. The doped NPs can be hydrogenated readily in a much shorter time at T < 623 K under near atmospheric H₂ pressure. The hydrogenated black NP films exhibit a broad UV-Vis absorption extending well beyond 800 nm. The experimental data can be corroborated by quantum calculations. The barriers for dissociative adsorption of H₂ at the Ni and O_2c sites on the 2Ni-doped TiO₂ surface are significantly reduced from 48 kcal mol⁻¹ on the undoped surface to 17 and 12 kcal mol⁻¹, respectively. The computed densities of states of the doped TiO₂ also show new absorption peaks in the band-gaps of the hydrogenated systems which exhibit a high efficiency of solar water-splitting over those of non-hydrogenated samples based on our preliminary study. The theoretical result also indicates that Ni-doping significantly affects the enthalpies of hydrogenation for formation of 2HO(b) and H₂O(b) in the bulk from 7 and 19 kcal mol⁻¹ in the undoped TiO₂ to −76 and −69 kcal mol⁻¹ in the 2Ni–TiO₂ system, respectively, with a > 80 kcal mol⁻¹ increase in exothermicities.

---

Introduction

Hydrogenation of TiO₂ nanoparticles (NPs) has been shown to significantly enhance the efficiency of water splitting in the visible region of the solar spectrum by photo-catalysis.^1–6 Chen et al.³ carried out the hydrogenation of TiO₂ white powders at 473 K under 20 atm of H₂ for 5 days which blackened the TiO₂ powders due to surface disordering with about 0.3 wt% of H₂ incorporation. The hydrogenated TiO₂ NPs enhanced the water splitting efficiency over those of pure TiO₂ by more than 2 orders of magnitude^4,5 with a high durability using methanol as a sacrificial agent. A similar study⁴ on the hydrogenation of TiO₂ rutile nanowires (NWs) and anatase nanotubes (NTs) at 673 K under H₂ atmosphere was found to improve the performance of photo-electrochemical water splitting with 200% enhancement in photocurrent and improvement in electrical conductivity and charge transportation, which were attributed to the formation of O-vacancies in TiO₂ with enhanced UV/visible and IR absorptions. Similar surface modifications have been accomplished by Zheng et al.⁵ with the hydrogenation of protonated TiO₂ NTs in a 5% H₂ diluted in N₂ using a quartz flow tube at 773 K; the hydrogenated anatase consisted of microspheres of TiO₂ NWs with enhanced photo-catalytic activities. Sun and co-workers⁷ studied the hydrogenation of well-defined nanocrystals of anatase TiO₂ at 723 K under 70 atm H₂ pressure; they reached as much as 1.4 wt% of H-incorporation under the high H₂ pressure at a rather high temperature condition. Interestingly, the results of their XRD, TEM, and Raman spectral measurements revealed no detectable morphological and crystallographic changes by hydrogenation, contrary to the finding of Chen et al.,³ who ascribed the enhanced photo-catalytic effect to surface disordering. The nature and locations of the disordered black TiO₂ NPs have been studied in detail by Naldoni et al.⁸ who characterized the blackened samples by different surface and optical measurements to understand the band-gap narrowing mechanism. The mechanism for TiO₂ hydrogenation by H₂ dissociation and migration into its subsurface has been investigated computationally by Raghunath et al.⁹ The controlling step for the hydrogenation process was reported to be the dissociative adsorption of H₂ on the TiO₂ surface with 47.8 kcal mol⁻¹ energy barrier, to be followed by H-atom migration into the bulk with 27.8 kcal mol⁻¹ barrier. In addition, H₂ was also found to be able to migrate molecularly into TiO₂ subsurface layers with 46.2 kcal mol⁻¹ of barrier, which is competitive with the dissociative adsorption process. Most interestingly, both H and H₂ inside the cages of the crystal bulk can readily form HO-bonds, whose transformation into H₂O inside the bulk helps create O-vacancies and may result in the disordering of the surface layers as found experimentally. These theoretical results help explain the need of high temperature and high H₂ pressure for the hydrogenation process and provide a reasonable mechanism for the surface disordering process.

In this work, we have carried out an experimental study on the effect of Ni-doping in TiO₂ NPs with different amounts of the dopant on the hydrogenation process and investigated the optical properties of the doped NPs with and without hydrogenation. In addition we studied computationally the effect of Ni-doping on the hydrogenation process by comparing the barriers for H₂ dissociation and H-atom migration into the TiO₂ subsurface with and without doping. Ni is well known for its efficacy in dissociating the H₂ molecule; for example, Ni-oxide has been employed as the key catalyst in the anode of the LSM/Ni-YSZ solid oxide fuel cell system.^10–13 The mechanism for H₂ dissociation and H-atom migration in the Ni-YSZ anode has been studied by Weng et al.^12a The results of our experimental finding on Ni-doping and its influence on hydrogenation and the quantum-mechanical calculations for H₂ dissociation, H-atom migration and band-gap properties for the Ni-doped TiO₂ are reported herein.

Experimental section

Synthesis of TiO₂ nanoparticles

All the reagent grade chemicals were used directly without further purification. Nickel nitrate was employed as the Ni source in the synthesis of TiO₂ NPs by sol gel hydrothermal process. A desired amount of titanium isopropoxide in i-propanol solution was dropped into acetic acid aqueous solution mixed with a varying amount of Ni(NO₃)₂ powders. After stirring at 353 K for 5 h, the final mixture was sealed in a Teflon container throughout 12 h of hydrothermal process at 593 K. The acid residues were removed by filtrating and washing for several times using de-ionic water and ethanol. After being dried at 423 K for 12 h, the Ni-doped TiO₂ NPs with approximately 20 nm in diameters were obtained. In Fig. 1(A), the Ni-doped TiO₂ NP powders are presented; their nominal weight ratios of Ni to Ti (Ni/Ti) are 1, 3.4, 5.0 and 7.5%. From the Fig. 1(A), one can see that the color of Ni-doped TiO₂ NP powders gradually changes from pale yellow to bright yellow with increasing Ni concentration.

Fig. 1 (A) The picture of Ni doped TiO₂ powders and (B) the UV-Vis absorption spectra of Ni doped TiO₂ thin film.

Fig. 2 (A) The picture of hydrogenated Ni–TiO₂ methanol mixture and (B) the UV-Vis absorption spectra of hydrogenated Ni–TiO₂ thin film.

Results and discussions

Characterization of TiO₂ NPs with doping and hydrogenation

The thin films of TiO₂ NPs for UV-Vis analysis were made by dissolving the powders in ethanol and DI water solution and spin-coating the mixture at 1000 rpm. UV-Vis absorption spectra of Ni–TiO₂ substrates are presented in Fig. 1(B). All Ni-doped samples show clear red-shifted tails around 400 to 500 nm with respect to TiO₂ blank. The E_g = 1239.8/λ relation was used to estimate the energy gap by extrapolating the onset to the related baseline.¹⁴ A significant absorption threshold band edge change from 3.26 (TiO₂) to 2.98 eV (5.0% Ni/TiO₂) is observed as shown in Table 1. A slight blue shift is also noted with the 7.5% Ni–TiO₂ NP sample. In Fig. S1† we have also shown a similar set of UV-Vis absorption spectra detected by diffuse reflection employing powders instead of thin-films, comparing those from 0.5 and 5% Ni-doped TiO₂ NPs with those of pure and hydrogenated TiO₂.

Table 1 Band gaps of Ni doped TiO₂

Sample	Absorption edge/nm	Band gap/eV
TiO2	380.4	3.26
1.0% NiTiO2	392.0	3.16
3.4% NiTiO2	406.1	3.05
5.0% NiTiO2	416.6	2.98
7.5% NiTiO2	410.0	3.02

---

To further reduce the TiO₂ band-gap and increase the photo-absorption efficiency in the visible range, hydrogenation of the Ni-doped TiO₂ NPs has been performed and tested. In a homemade glass tube furnace, the yellow Ni-doped TiO₂ NP powders were placed in an open quartz boat. The hydrogenation process was carried out at 573 K for 3 h under 780 Torr H₂ gas. The system as well as the complete procedure of hydrogenation is presented in the ESI† section.¹⁵ After the reaction, the color of Ni-doped TiO₂ powders changed dramatically as shown in Fig. 2(A). The color varies from pale grey to dark grey with increasing Ni contents. The UV-Vis absorption spectra of the hydrogenated samples are presented in Fig. 2(B) which shows a pronounced enhancement in the absorption spectra throughout the entire visible range. Under the same hydrogenation condition, the white color of the undoped TiO₂ powder remains unchanged after the treatment. In Fig. 2(A), the pale yellow powders (1% Ni) changes to light grey and the more concentrated Ni–TiO₂ powders (7.5% Ni) changes to black while the undoped TiO₂ powders remain white. This result indicates that Ni-doping significantly enhances the hydrogenation process and helps reduce the TiO₂ band-gap as supported by the results of the quantum-chemical calculation presented below. Our preliminary testing using the 0.5 and 5% Ni-doped TiO₂ NPs with 10% ethanol as sacrificial agent also show a significant enhancement in its water-splitting efficiency over those of non-hydrogenated TiO₂ with or without Ni-doping. Fig. 3 shows the measured hydrogen evolution data obtained from the photo-catalytic dissociation of water irradiated with about 2 W power from a 1 kW Xe lamp; the result gives the relative efficiencies for the pure TiO₂, hydrogenated TiO₂, 5% Ni-doped TiO₂ and hydrogenated 5% Ni-doped TiO₂, 1 : 1 : 4 : 18. Table S1 in the ESI†¹⁵ presents the corresponding H₂ evolution rate in units of mmol g⁻¹ hr⁻¹.

Fig. 3 Hydrogen production rates under 1 h illumination of NPs with about 2 W Xe-lamp output using 10% ethanol as sacrificial agent (light on: time 20–80 min). The relative efficiencies of pure TiO₂, hydrogenated TiO₂, 5% Ni-doped TiO₂ and hydrogenated Ni-doped TiO₂ with 5% Ni-doping are 1 : 1 : 4 : 18.

Computational study

The adsorption and dissociation of H₂ on Ni-doped TiO₂(101) surface by the first-principles calculations were carried out with the Vienna ab initio simulation package (VASP)¹⁶ using the DFT + U method.¹⁷ For the total energy prediction, the exchange-correlation function treated by the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof formulation (PBE)¹⁸ has been applied with spin-polarization throughout the system. Twenty four [TiO₂] units doped with one and two Ni atoms were modeled by (2 × 2 × 3) supercell slabs separated perpendicularly by a 15.0 Å vacuum space. The predicted lattice constants at the PBE level for the anatase crystal bulk are a = 3.828 Å and c = 9.677 Å, in good agreement with the experimental values of a = 3.782 Å and c = 9.502 Å.¹⁹ In our calculation, the surface area of the anatase (101) surface was 11.097 × 7.655 Å extended along the 〈1̄11〉 and 〈010〉 direction. A 2 × 3 × 1 Monkhorst–Pack k-point sampling was used in the calculations. The predicted structure is presented in Fig. S2 of the ESI.†¹⁵ To locate the transition states of the dissociative adsorption the climbing-image nudged-elastic band (CINEB) method was applied.²⁰ To correct the strong on-site Coulomb repulsion of Ti and Ni 3d states, the value of U was taken to be 4 eV, consistent with that employed in previous reports.^9,21,22

H₂ adsorption and dissociation on Ni-doped TiO₂(101)

We have investigated the molecular adsorption and dissociative adsorption of H₂ on the 1 and 2 Ni-doped anatase surfaces, denoted by 1Ni–TiO₂ and 2Ni–TiO₂, respectively. In the absence of a dopant four sites have been identified on the TiO₂ surface as labeled in Fig. S2.† Among the four, there are 2 active sites on the surface which may interact with molecular hydrogen; these are five-fold coordinated titanium (Ti_5c) and two fold bridging oxygen (O_2c). The remaining two sites are three-fold coordinated oxygen (O_3c) and six-fold coordinated titanium (Ti_6c); they are not active as the other two. In the 1-Ni doped case, one Ni atom replaces one of the Ti_5c atoms on the surface, labeled as D1, and in the 2-Ni doped case 1Ni substitutes a Ti_5c on the surface, also labeled as D1, and the other Ni atom replaces a Ti_6c atom inside the bulk which is labeled as D2 (see Fig. S2†). In terms of the Ni concentration, the Ni/Ti ratio in the 1- and 2-Ni doped TiO₂ is approximately equivalent to the experimental concentration of 4 and 8%, respectively.

The interaction of H₂ with clean and doped TiO₂ surfaces was found to be rather weak; on the undoped surface H₂ can physisorb on an O_2c site with 0.3 kcal mol⁻¹ binding energy,⁹ whereas on the Ni-doped surface the interaction becomes weakly repulsive at both Ni and O_2c sites. A Bader charge analysis gives small net charges of ∼0.02e to the H₂ on the 1Ni–TiO₂ and 2Ni–TiO₂ surface giving rise to the small repulsive energy as shown in Table S2.† For the dissociative adsorption of H₂, on the other hand, the barriers for H₂ dissociation at the Ni- and O_2c-sites of the 1Ni–TiO₂ surface are predicted to be significantly reduced to as low as 12.1 and 6.1 kcal mol⁻¹, respectively, from that on the undoped TiO₂ surface, 47.8 kcal mol⁻¹ (ref. 9) (see Fig. S3 (a) and (b)).† Most notably, the predicted enthalpy changes for the dissociative adsorption processes producing H-O_3c,H-O_2c-1Ni-TiO₂(a) and 2H-O_2c-1Ni-TiO₂(a) from the initial interaction on the Ni and O_2c sites are significantly greater than that on the undoped surface, ΔH = −18.3 kcal mol⁻¹ producing 2H-O_2c-TiO₂(a),⁹ by as much as −77 and −80 kcal mol⁻¹, respectively, as shown in Fig. S3 (a) and (b).† The dissociative adsorption energy barriers, 17.4 and 12.4 kcal mol⁻¹ on the Ni and O_2c sites, respectively, on the 2Ni–TiO₂ surface are depicted in Fig. S3 (c) and (d).† The corresponding geometries are shown in Fig. S4 and S5.† For the dissociative adsorption on the latter, more reactive O_2c site, its detailed PES including the H₂ dissociation on the surface, H-atom migration into the bulk and the ultimate H₂O formation inside the bulk, to be discussed below, is presented in Fig. 4 for comparison with the analogous steps predicted for the undoped TiO₂ surface.⁹ The corresponding geometries of H₂ dissociation and migration on undoped TiO₂ are shown in Fig. S6.† The energetics for the key intermediates, transition states and products of these two systems are also summarized in Table S2† for a closer examination. The most significant finding revealed by the theoretical calculation for the initial dissociative adsorption of H₂ on the Ni-doped TiO₂ surface is the enormous catalytic effect of Ni on the reduction of the H₂-dissociation barriers and the very large exothermicities associated with the formation of surface hydroxyl species (2 HO_2c(a) (i.e., 2H–O_2c–2Ni–TiO₂(a)) or the mixed HO_2c(a) and HO_3c(a), H–O_3c–2Ni–TiO₂(a)). It should also be mentioned that on account of the considerably lower barriers for the dissociative adsorption processes, our extensive searches for the direct H₂-molecular migration from the 1Ni- or 2Ni–TiO₂ surface into its subsurface failed to locate its transition state (which was found to be 46.2 kcal mol⁻¹ for the undoped surface⁹ as alluded to in the introduction); the searches always converged to those of the dissociative adsorption processes. This is also an interesting finding which evidently suggests that the hydrogenation of the Ni-doped TiO₂ NPs occurs exclusively by H-atom formation and diffusion reactions on the surface and inside the bulk of the systems as discussed below.

Fig. 4 Predicted potential energy diagrams for (a) H₂ dissociation on the TiO₂(101) surface and (b) H₂ dissociation on the 2Ni–TiO₂(101) surface with the DFT + U method. The corresponding geometries are shown in Fig. S5 and S6.† The hashed lines show omission of several small barriers in both energy profiles.

H-atom migration

In the undoped anatase system,⁹ the migration of the H atoms into the bulk played a key role in the hydrogenation of TiO₂ NPs which led to enhanced water-splitting efficiencies.^4,5,23 The barrier for H-migration into a subsurface layer was predicted to be 26.4 kcal mol⁻¹.⁹ We have investigated the effect of Ni-doping on the migration of an H atom adsorbed on the surface, HO_2c(a), into the bulk and compared the transition state barriers for each step along the migration path: HO_2c(a) → HO_3c(a) → H_BD1(b) → H_BD2(b), where H_BD1(b) and H_BD2(b) denote the H atom undergoes bulk diffusion to site 1 and site 2, respectively (see Fig. S3 (e) and (f) and S7 and Table S3†). Interestingly, the barriers for H-migration into the bulk for the doped and undoped⁹ systems are very similar to each other, suggesting that in the 1Ni-doped case, the migration kinetics should be quite similar. Notably, the full PESs for the migration of H atoms after the dissociative adsorption of H₂(g) on the undoped and doped 2Ni–TiO₂ surfaces, up to the formation of H₂O inside the bulk as shown in Fig. 4, are also very similar in their barrier heights for the individual steps along their reaction paths. The major differences, again, lie in the initial dissociative adsorption barriers and the overall exothermicities for the formation of 2HO_2c(a) on the surface, and 2HO(b) and H₂O(b) inside the bulk; for the latter two products, the enthalpy differences without and with doping are as much as −83.6 and −87.7 kcal mol⁻¹, revealing a much greater stability of hydrogen inside the bulk of TiO₂ with Ni-doping.

DOS of H₂ and H in Ni-doped TiO₂

To corroborate the experimental observation of UV-Vis spectra of hydrogenated Ni-doped TiO₂ NPs, we have investigated further to understand the Ni effect on band-gap reduction by incorporation of hydrogen. The density of states (DOS) and projected DOS (PDOS) on the Ni-doped TiO₂ surface are calculated by using the DFT + U method with the value of U = 4 eV for Ti and Ni as mentioned before. The results are presented in Fig. 5; panel (5a) shows the DOS of undoped TiO₂(101). The top of the valence band (VB) mainly consisting of the 2p states of oxygen was found to lie within 4.5 and 0 eV with the domination of Ti 3d states at the lower part of the conduction band (CB), consistent with previous reports.^9,24 The calculated DOS for the 2Ni-doped TiO₂ is shown in panel (5b) which indicates that the band gap is reduced to 2.79 eV (cf. the predicted undoped TiO₂, 2.90 eV). Here, we observe new localized states appearing between VB and CB at 1.23 eV above the VB. These new peaks in the band gap of the Ni-doped TiO₂ derive mainly from the 3d states of Ni and 2p states of O with a minor contribution from Ti. Also, our experimental results show the band-gap reduction from 3.26 eV for TiO₂ to 2.98 eV for 5% Ni doped TiO₂ (see Table 1). In panel (5c) for an H atom adsorbed on an O_2c site and another H bonding with an O in the sublayer, giving H-O_sub2,H-O_2c-2Ni(a), the Fermi level is located at the 0.2 eV above the VB and the band gap is reduced by 0.2 eV compared to the un-doped TiO₂. The DOS indicates that the impurity states produced the minor peaks near the top edge of the VB mostly contributed by Ni and O. However, two new prominent localized peaks appear at 0.7–1.0 eV above the Fermi level; the PDOS of the Ni, O and H shows that the new states result mostly from Ni 3d and O 2p orbitals with a negligible contribution from the H atoms. To further elucidate the role of the H atom on the surface vs. that inside the bulk, we have carried out the DOS calculations independently for an H atom on the surface attached to O_2c and in the bulk attached to O_sub3; the results shown in Fig. S8† indicate that new impurity peaks appear at 1.0 and 1.5 eV, respectively, above the Fermi level and the band gaps have been reduced by ∼0.2 eV compared to that of the undoped TiO₂. In the H-O_sub3-2Ni(b) case, another new peak appears at 0.3 eV above the VB. The result of the DOS calculation for 2H atoms inside the bulk of the 2Ni-doped TiO₂, 2HO–2Ni(b), is shown in panel (5d). The two H atoms get incorporated interstitially producing two OH groups at the O_sub1 and O_sub3 sites. In this case, the Fermi level appears at ∼0.5 eV above the VB and the band gap is reduced to 2.80 eV (cf. the anatase gap 2.90 eV). The two new impurity peaks appearing at the edge of the VB lying below the Fermi level at ∼0.2 and 0.06 eV, may be attributed to a stronger interaction between the Ni 3d and O 2p orbitals. Also, another new peak appears at 0.7 eV above the Fermi level may be attributed to Ni 3d and O 2p orbitals. As shown in panel (5e), we have also calculated the DOS for the H₂O formed in the subsurface of the 2Ni–TiO₂ system. Two new localized states appear at 0.7 and 1.5 eV above the VB. The band gap has been reduced to 2.64 eV (cf. 2.90 eV for clean TiO₂ and 2.78 eV for 2Ni–TiO₂). The band gap reduction with the two new localized states appearing between the VB and CB derive mainly from the Ni and its neighboring O atoms with a negligible contribution from Ti 3d.

Fig. 5 Density of states (DOS) for (a) clean TiO₂, (b) two Ni doped TiO₂, (c) one H on the surface and one H inside the bulk of 2Ni doped TiO₂, (d) 2H inside the bulk of 2Ni doped TiO₂ and (e) H₂O formed inside the bulk of 2Ni doped TiO₂ calculated at the DFT + U level (U = 4.0 eV for Ni and Ti) (for clarity, H and Ni PDOS peaks are magnified). Their optimized geometries are shown in Fig. S4 and S5.† The dashed vertical line represents the position of the Fermi level.

In order to understand the reduction process in the Ni-doped TiO₂ system, we have carried out Bader charge analyses using the DFT + U method as shown in Fig. S9.† According to the analysis, the charge of O and Ti atoms in the undoped TiO₂ are −0.92e and 2.01e, respectively, where e is the magnitude of the charge on an electron (see Fig. S9 (a)†). In the 2-Ni doped case, Ni (at the D1 site) bonding with one of the neighboring O_2c oxygens, the charge of the oxygen is predicted to be −0.69e, changing noticeably from −0.92e in the clean TiO₂ surface. The charge of the Ni (D1) is 1.37e and that of the O_3c on the surface is ∼−0.96e. The bond length between Ni and O at the O_2c site is observed to be shortened to 1.793 Å, comparing with other Ni–O_3c bonds, 1.924 Å. The charge of the second Ni located in the subsurface is predicted to be 1.38e. The charges of surrounding oxygens around Ni are noted to be reduced by 0.1e–0.28e comparing with that in the clean TiO₂ (Fig. S9 (b)†). In the H₂ adsorption on the surface, as shown in Fig. S9 (c) and (d),† there is a negligible charge transfer between the H₂ and the surface of the Ni-doped TiO₂. After the H₂ dissociative adsorption at an O_2c site of the 2Ni–TiO₂ surface producing 2H–O_2c–2Ni–TiO₂(a), the charge of the O atom is predicted to be −1.20e (see Fig. S9 (e)†), a significant change from −0.69e cited above. It is noteworthy that the charge of the other neighboring O_2c is also observed to be increased by −0.3e from that in 2Ni–TiO₂. The bond length between Ni and the O_2c of the H–O_2c increases to 1.961 Å from 1.793 Å in 2Ni–TiO₂. The large charge transfer and geometrical changes on the surface may be associated with the release of the large amount of energy from the formation of the stable complex 2H–O_2c–2Ni–TiO₂(a), 94.8 kcal mol⁻¹, in the H₂(g) + 2Ni–TiO₂ reaction.

As shown in Fig. S9 (f),† one H atom adsorbed on an O_2c site and another H atom bonding with an O_sub2 site in the bulk, the charge of 2Ni atoms have been reduced by 0.1e while those of O_2c and O_sub2 are increased by 0.24e comparing to those 2Ni–TiO₂. The charges of both H atoms are observed to be 0.6e.

As shown in Fig. S9 (g) and S9 (h),† the two H atoms get incorporated interstitially producing two OH groups at O_sub1 and O_sub3 sites in different regions of the sub surface of 2Ni–TiO₂. We observe significant atomic charge changes in the neighboring oxygen atoms inside the bulk. The charge of the sub surface oxygen is −1.03e, whereas the charges of the two oxygens associating with 2H become −1.26e and −1.20e as shown in Fig. S9 (g)† and −1.14e and −1.21e in Fig. S9 (h).† The charges of H atoms in both cases are around 0.7e and 0.6e. Here we have not observed a significant change in the charges of neighboring Ti atoms. The charges of one H atom at a surface O_2c and one in the bulk of 2Ni–TiO₂ are shown in Fig. S9 (j) and (k).† We have also carried out the Bader charge analysis for the O being reduced to H₂O in 2Ni–TiO₂ (Fig. S9 (i)†); the result shows that the charge of the O in H₂O inside the bulk is −1.30e, while those of the 2H are 0.64e and 0.70e. Here we also did not note any significant change in the charges of Ni's neighboring Ti atoms. These results clearly show that the new localized states appearing between the VB and CB due to the H₂O formation in the subsurface, H₂O–2Ni(b), derive primarily from the contributions of the Ni and its neighboring O atoms. The contribution from Ti 3d is significantly lower comparing with those of Ni and O.

Conclusions

To summarize, we have studied the effect of Ni-doping in TiO₂ nanoparticles (NPs) experimentally and computationally on the hydrogenation process. Experimentally, we have found that TiO₂ NPs of approximately 20 nm in diameter doping with a small amount of Ni prepared by a sol gel method can be readily hydrogenated at T < 623 K under near atmospheric pressure condition. The hydrogenated black NP films exhibit a very broad UV-Vis absorption extending well beyond 800 nm. The black NPs also show a high efficiency for photo-catalytic water splitting (using 10% ethanol as a sacrificial agent) according to our preliminary test. The mechanism for the hydrogenation process on Ni-doped TiO₂ has also been investigated by quantum-chemical calculations. The result indicates that the presence of 1Ni atom on the surface of a 24-unit [TiO₂] cell, representing approximately 4% dopant, can reduce the barrier for the dissociative adsorption of H₂ by as much as ∼40 kcal mol⁻¹ with a significant enhancement in H-atom generation and migration into the bulk. The predicted density of states of the hydrogenated Ni–TiO₂ NPs also reveal the presence of prominent impurity states in the band-gap when H or its reaction product, H₂O, is present in the bulk. Furthermore, the theoretical result also indicates that the enthalpies of hydrogenation reactions producing the reduced H- and H₂O-species on the surface or inside the bulk increase by as much as 80 kcal mol⁻¹ because of the presence of the Ni-dopant.

Acknowledgements

The authors acknowledge the supports from the ATU Plan of the Ministry of Education, Taiwan, and also from the National Center for High-performance Computing for providing the computer time. MCL thanks the Ministry of Education of Taiwan for the distinguished visiting professorship at NCTU.

Notes and references

1. T. Leshuk, R. Parviz, P. Everett, H. Krishnakumar, R. A. Varin and F. Gu, ACS Appl. Mater. Interfaces, 2013, 5, 1892–1895.
2. Y. Yan, M. Han, A. Konkin, T. Koppe, D. Wang, T. Andreu, G. Chen, U. Vetter, J. R. Morante and P. Schaaf, J. Mater. Chem. A, 2014, 2, 12708–12716.
3. X. Chen, L. Liu, P. Y. Yu and S. S. Mao, Science, 2011, 331, 746–750.
4. G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 3026–3033.
5. Z. Zheng, B. Huang, J. Lu, Z. Wang, X. Qin, X. Zhang, Y. Dai and M. H. Whangbo, Chem. Commun., 2012, 48, 5733–5735.
6. Y. H. Hu, Angew. Chem., Int. Ed., 2012, 51, 12410–12412.
7. C. Sun, Y. Jia, X. H. Yang, H. G. Yang, X. Yao, G. Q. Lu, A. Selloni and S. C. Smith, J. Phys. Chem. C, 2011, 115, 25590–25594.
8. A. Naldoni, M. Allieta, S. Santangelo, M. Marelli, F. Fabbri, S. Cappelli, C. L. Bianchi, R. Psaro and V. D. Santo, J. Am. Chem. Soc., 2012, 134, 7600–7603.
9. P. Raghunath, W. F. Huang and M. C. Lin, J. Chem. Phys., 2013, 138, 154705.
10. S. Primdahl and M. Mogensen, J. Electrochem. Soc., 1997, 144, 3409–3419.
11. P. Holtappels, I. C. Vinke, L. G. J. de Haart and U. Stimming, J. Electrochem. Soc., 1999, 146, 1620–1625.
12. (a) M. H. Weng, H. T. Chen, Y. C. Wang, S. P. Ju, J. G. Chang and M. C. Lin, Langmuir, 2012, 28, 5596–5605; (b) Z. Cheng, J. H. Wang, Y. M. Choi, L. Yang, M. C. Lin and M. Liu, Energy Environ. Sci., 2011, 4, 4380–4409.
13. A. Faes, A. Hessler-Wyser, A. Zryd and J. V. Herle, Membranes, 2012, 2, 585–664.
14. (a) M. Z. Selcuk, M. S. Boroglu and I. Boz, React. Kinet., Mech. Catal., 2012, 106, 313–324; (b) T. Sun, J. Fan, E. Liu, L. Liu, Y. Wang, H. Dai, Y. Yang, W. Hou, X. Hu and Z. Jiang, Powder Technol., 2012, 228, 210–218.
15. See ESI†.
16. (a) G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169; (b) G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 47, 558.
17. S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys and A. P. Sutton, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 1505–1509.
18. (a) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785; (b) J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
19. J. K. Burdett, T. Hughbanks, G. J. Miller, J. W. Richardson and J. V. Smith, J. Am. Chem. Soc., 1987, 109, 3639–3646.
20. G. Henkelman, B. P. Uberuaga and H. Jónsson, J. Chem. Phys., 2000, 113, 9901.
21. E. Finazzi, C. Di Valentin, G. Pacchioni and A. Selloni, J. Chem. Phys., 2008, 129, 154113.
22. M. J. Han, T. Ozaki and J. Yu, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 045110.
23. (a) D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, J. M. Friedrich and F. C. Walsh, J. Phys. Chem. B, 2005, 109, 19422–19427; (b) S. H. Lim, J. Luo, Z. Zhong, W. Ji and J. Lin, Inorg. Chem., 2005, 44, 4124–4126.
24. (a) M. M. Islam, M. Calatayud and G. Pacchioni, J. Phys. Chem. C, 2011, 115, 6809–6814; (b) F. Lin, G. Zhou, Z. Li, J. Li, J. Wu and W. Duan, Chem. Phys. Lett., 2009, 475, 82–85.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra16119c

---