Optimally Selecting Photo- and Electrocatalysis to Facilitate CH₄ Activation on TiO₂(110) Surface: Localized Photoexcitation versus Global Electric-Field Polarization

Abstract

Photo- and electrocatalytic technologies hold great promise for activating inert chemical bonds under mild conditions, but rationally selecting a more suitable method in between to maximize the performance remains an open issue, which requires a fundamental understanding of their different catalytic mechanisms. Herein, by first-principles calculations, we systematically compare the activation mechanisms for the C–H bond of the CH₄ molecule on TiO₂(110) under the photo- and electrocatalytic modes without or with water involved. It quantitatively reveals that the activation barrier of the C–H bond decreases dramatically with a surprising 74% scale by photoexcitation relative to that in thermocatalysis (1.12 eV), while the barrier varies with a maximum promotion of only 5% even under −1 V/Å external electric field (EEF). By detailed geometric/electronic analysis, the superior photocatalytic activity is traced to the highly oxidative lattice O_br^•– radical excited by a photohole (h⁺), which motivates the homolytic C–H bond scission. However, under EEF from −1 V/Å to 1 V/Å, it gives a relatively mild charge polarization on the TiO₂(110) surface region and thus a limited promotion for breaking the weakly polar C–H bond. By contrast, in the presence of water, we find that EEF can facilitate CH₄ activation indirectly assisted by the surface radical-like OH* species from the oxidative water cleavage at high oxidative potential (>1.85 V vs SHE), which explains the high energy cost to drive electrocatalytic CH₄ conversion in experiment. Alternatively, we demonstrate that more efficient CH₄ activation could be also achieved at much lower oxidative potential when integrating the light irradiation. In such a circumstance, EEF can not only promote the h⁺ accumulation at the catalyst surface but also help H₂O deprotonation to form hydroxide, which can serve as an efficient hole-trapper to generate OH^• radical (OH^– + h⁺ → OH^•), unveiling an interesting synergistic photoelectrocatalytic effect. This work could provide a fundamental insight into the different characteristics of photo- and electrocatalysis in modulating chemical bond cleavage.

Introduction

To achieve the efficient activation of inert C–H bond of methane (CH₄) at mild temperature is highly desired for facilitating the selective and controllable conversion of CH₄ into a value-added commodity, which constitutes one of the most fundamental and challenging tasks in the chemical community.¹⁻⁴ In this regard, the transition metal oxide (TMO) catalysts have drawn enormous attention in practice with the unsaturated metallic site and lattice oxygen exposed simultaneously;⁵⁻⁷ particularly, the photo- and electrocatalytic technologies are recognized as promising approaches, with merits of altering reaction thermodynamics and kinetics under mild temperature driven by the external light and electricity energy supply. Significantly, some experiments have demonstrated the feasibility of photo- and electrocatalytic CH₄ conversion.⁸⁻¹¹ For example, Yoshida et al. reported that Pt/TiO₂ photocatalyst displayed high activity for CH₄ oxidation around room temperature (ca. 323 K).¹² The TiO₂ electrode decorated with RuO₂ and V₂O₅ was found capable of electrocatalytically converting CH₄ into methanol with a selectivity of 57% at an applied voltage of 2.25 V (vs standard hydrogen electrode, SHE).¹³

Despite these important findings, the insight into the activation of the first C–H bond in the CH₄ molecule under light irradiation and the electric-field (EF) polarization imposed by the cell potential has been not fully understood at the atomic level, which largely differs from the common thermocatalytic mode.^10,14 For example, qualitatively, the photoexcited polarons are believed to play an important role in the promotion effect with uncertainty. Wei et al.¹⁵ proposed that the hole-trapped lattice oxygen radical was the main species for CH₄ activation over the β-Ga₂O₃ photocatalyst under 254 nm light irradiation, which was also suggested on other metal oxide catalysts such as Ag-decorated ZnO, SrCO₃/SrTiO₃, and so on.¹⁶⁻¹⁸ However, the photoelectron could be also controversially responsible for the enhanced CH₄ activation at mild conditions.¹⁹⁻²¹ Thus, a quantitative description of the activation process incorporating the kinetic barrier for CH₄ activation in the photocatalytic mode remains an unresolved essential issue.

Similar to the photocatalysis, the electrocatalytic approach is usually believed to be capable of modulating the chemical bond breaking/reforming through the EF effect,²² but to our knowledge, how largely the EF polarization effect directly affects the C–H activation on the semiconductor is less understood, despite that some important progress on the EF tuning effect has been achieved on the transition metal catalysts (e.g., Ni, Pt, Co, and so on).²³⁻²⁹ Moreover, as an extended fundamental question, one may also wonder: theoretically, which catalytic mode is more efficient in facilitating the C–H bond and other chemical bonds? Interestingly, the current research implies that the electrochemical oxidation of CH₄ appears to be kinetically slow and energy intensive.³⁰⁻³³ For example, only when the applied potential (U_app) increases to 2.25 V (vs SHE) can the efficiency of electrocatalytic CH₄ conversion increase rapidly with a selectivity toward methanol reaching 57% over a TiO₂/RuO₂/V₂O₅ electrode; with a low U_app (<2.25 V), the improvement of the EF-imposed catalytic efficiency becomes relatively limited.^13,34 Remarkably, when combining the electrocatalytic condition with the light irradiation, the activation of CH₄ can occur on pure TiO₂ catalyst at a much reduced U_app; as shown from the experimental fact, the selectivity of CH₄ steam reforming toward CO reached up to 81.9% over TiO₂ surface under −0.41 V vs SHE.^9,35 Rationalizing these questions is not straightforward, and evidently, the photo- and electrocatalytic modes exhibit distinct catalytic characters. Some key questions are as follows: How does the EF imposed by the external voltage affect the C–H activation below and above 2.25 V (vs SHE) on TiO₂ catalysts? What are the unique characteristics of photo- and electrocatalysis, respectively? How can they cooperate with each other?

To quantitatively understand the photo- and electrocatalytic characteristics on CH₄ activation and rationalize their essential differences, here we carried out systematic DFT+U calculations with the PBE functional benchmarked against the advanced HSE06 hybrid functional for selected geometries on CH₄ activation over the rutile TiO₂(110) surface, a cross-sectional oxide support in photoelectrocatalytic reaction. The results unambiguously disclose and explain the unequivalent effects of light and electric field on CH₄ activation. Light irradiation can remarkably promote the C–H bond activation directly by the localized photogenerated hole, while the direct EF polarization effect on C–H cleavage in the electrochemical mode is unexpectedly weak. Mechanistically, we found that the C–H activation undergoes different mechanisms in diverse external driving forces. To activate the C–H bond on TiO₂(110), the electrochemical mode could have to indirectly make use of the radical-like OH* intermediate from environment H₂O dissociation via the proton coupled electron transfer (PCET) mechanism. Moreover, we also demonstrated that these two distinct function mechanisms of photocatalysis and electrocatalysis will give an improved activity when integrated cooperatively.

Results and Discussion

Thermocatalytic Activation of CH₄

To begin with, the adsorption and activation of the CH₄ molecule on the TiO₂(110) surface in the thermocatalytic condition were first calculated as a benchmark. The rutile TiO₂(110) surface comprises two types of main active sites: the two-coordinated bridge oxygen and five-coordinated Ti cation (Figure 1a), denoted as O_br and Ti_5c, respectively.³⁶ Because of the geometrically high symmetry of the CH₄ molecule, it can weakly adsorb on the TiO₂(110) surface with an adsorption energy (E_ads) of −0.32 eV, wherein CH₄ interacts mainly with d_z² orbital of the surface Ti_5c atom.^37,38 As the transition state (TS1) shows in Figure 1b, when C–H bond activation of CH₄ follows the homolytic cleavage mode at the O_br site (i.e., the so-called hydrogen atom transfer (HAT) mechanism),^6,39 it gives a reaction barrier as high as 1.83 eV, which corresponds to CH₄ + Ti_5c⁴⁺ + O_br^2– → CH₃^• + Ti_5c³⁺ + O_brH^– with the CH₃^• radical formed and the released electron being trapped at Ti_5c⁴⁺ into Ti_5c³⁺ at TS1. By comparison, in response to the synergistic coupling of Ti_5c³⁺ with the CH₃^• radical, the C–H bond activation prefers to proceed via the proton coupled electron transfer (PCET) mechanism⁴⁰ (CH₄ + Ti_5c⁴⁺ + O_br^2– → Ti_5c⁴⁺–CH₃^– + O_brH^–, see TS2 in Figure 1b) at the Ti_5c and O_br dual sites with a lower barrier of 1.12 eV, which can be apparently taken as a heterolytic cleavage mechanism. Specifically, as the spin charge distribution of the transition state (TS2) illustrates in Figure 1b, the O_br site captures the dissociated hydrogen, and Ti_5c as a Lewis acidic center stabilizes the methyl group, yielding a proton and a methyl anion (corresponding to Bader charges of 1.000 and −0.376 |e|, respectively), respectively. The forming H–O_br and C–Ti_5c bonds in the TS2 are measured to be 1.200 and 2.347 Å, respectively. Nevertheless, overcoming the activation barrier of 1.12 eV is kinetically difficult at room temperature, which accords with the fact that CH₄ activation generally has to be operated at high temperature.^41,42

Figure 1.

(a) Surface structure of optimized rutile TiO₂ (110), with the corresponding O–Ti bonds defined as d₁ and d₂. (b) Two possible mechanisms for CH₄ activation in thermocatalysis incorporating the spin density distributions of initial state (IS), transition state (TS), and final state (FS) at isovalue = 0.005. Some key bond lengths and Bader charges are marked in black and blue, respectively. Bader charge unit: |e|. Red, blue, gray, and white balls represent O, Ti, C, and H atoms, respectively. These notations are used throughout the paper. (c) Energy profiles of CH₄ activation in three different external driving forces (light, electricity, and thermal energy).

Photocatalytic Activation of CH₄

Under light irradiation, the photogenerated hole/electron pairs are generated and could be separated and localized at the different regions of the TiO₂(110) surface, where the photogenerated hole tends to be trapped by surface O_br to generate O_br^•– radical (i.e., O_br^2– + h⁺ → O_br^•–), while the photogenerated electron can be localized at Ti_5c⁴⁺ to form the Ti_5c³⁺ radical (i.e., Ti_5c⁴⁺ + e^– → Ti_5c³⁺), which has been demonstrated in our previous studies and others.⁴³⁻⁴⁵ As displayed in Figure S1, with the assistance of Ti_5c³⁺, the barrier of CH₄ activation is decreased a little to 0.99 eV with the strengthened stabilization of the CH₃ intermediate by the Ti_5c³⁺ site, and the reaction behavior is the same as that in thermocatalysis with a similar TS structure (see in Figure S1). Interestingly, in the presence of O_br^•– radical, the barrier of C–H bond cleavage of the CH₄ molecule (i.e., CH₄ + O_br^•– → HO_br^– + CH₃^•) can be greatly reduced to as low as 0.29 eV (see energy profile in Figure 1c). In comparison with the thermocatalytic case, the promotion scale of the activation barrier is quantitatively estimated to be an order of 74% (0.29 eV versus 1.12 eV). One can thus anticipate that the photocatalytic conversion would be an efficient way for activating CH₄ on TiO₂(110) if there are sufficient surface O_br^•– species by increasing light absorption and decreasing the recombination of hole/electron pairs.

By analyzing the geometric/electronic structure of the resulting TS (Figure 2), one can see that the C–H cleavage mode changes evidently owing to the participation of O_br^•– species; specifically, the TS changes from the original dihapto configuration (with the CH₃ and H atom in CH₄ bonded to the Lewis acid/base site, Ti_5c/O_br, respectively) in the thermocatalytic case to the monohapto configuration, in which the H atom is captured by the O_br^•– site and the methyl radical suspends above the TiO₂ surface. The bond lengths of the CH₃···H and H···O_br in the TS are 1.227 and 1.339 Å, respectively. In other word, it exhibits a radical-like TS type, and the radical nature of CH₃ species is also confirmed by a Bader charge of 0.090 |e| and a magnetic moment of 0.43 μ_B (Table S3). Therefore, the C–H bond activation modulated by the photoexcited hole essentially follows a homolytic cleavage mechanism or the so-called hydrogen-atom transfer mechanism.⁴⁰

Figure 2.

Geometry structures and spin density distributions of IS, TS, and FS for CH₄ activation on TiO₂(110) with the assistance of photogenerated hole.

In addition, by analyzing the Bader charge of the O_br^•– center, it is found that it holds less electron than the O_br^2– site in thermocatalysis (−0.690 vs −1.082 |e|). In this sense, we can rationalize that the O_br^•– center, as a strong oxidative site with an unoccupied 2p orbital, can more easily capture electrons from the dissociated hydrogen atom than O_br^2–, thereby leading to a much decreased barrier of C–H bond activation. By contrast, in the thermocatalytic condition, the oxidizability of the close-shell O_br^2– is too weak to break the strong C–H bond in the CH₄ molecule.

Electrocatalytic Activation of CH₄

As a quantitative comparison, the adsorption and activation of CH₄ in a series of electric-field intensities (−1 V/Å ≤ F ≤ 1 V/Å) were calculated and are shown in Table 1. When the negative EEFs are applied, the CH₄ adsorption and activation progresses can be facilitated. For example, in comparison with the zero EEF, the adsorption energy of CH₄ at −1 V/Å increases from −0.32 eV to −0.41 eV, and the CH₄ activation barrier deceases from 1.12 to 1.06 eV (Figure 3a). However, to our surprise, the contribution of EEFs on CH₄ activation is relatively weak (on an order of only 5% in −1 V/Å relative to that in 0 V/Å), whereas the positive EEFs suppress this progress. Specifically, the activation barrier increases gradually with the increasing of the applied electric fields from −1 V/Å to 1 V/Å (Figure 3b). Moreover, by fitting the activation barriers (E_a) versus the EEF intensities (F), we can obtain two independent linear relations with different slopes in the negative and positive EEF regions:

Table 1. Adsorption Energies of the CH₄ Molecule on the TiO₂(110) Surface (E_ads), as well as the Activation Barriers (E_a) and Enthalpy Changes (ΔH) of CH₄ Activation, and the Optimized Average O–Ti bond lengths (d₁ and d₂), Bader Charges of O_br and Ti_5c Sites, and -ICOHP Values of C–Ti_5c and H–O_br Bonds in TSs, Obtained in Different EEFs (F) Using DFT+U.

				bond length		Bader charge		-ICOHP (eV)
EEF (V/Å)	Eads (eV)	Ea (eV)	ΔH (eV)	d1 (Å)	d2 (Å)	Obr	Ti5c	C–Ti5c	H–Obr
–1.00	–0.41	1.06	0.32	1.840	1.863	–1.048	2.432	1.516	2.452
–0.50	–0.35	1.10	0.45	1.846	1.854	–1.066	2.434	1.409	2.465
–0.35	–0.35	1.11	0.48	1.848	1.853	–1.066	2.409	1.391	2.468
0.00	–0.32	1.12	0.54	1.854	1.847	–1.082	2.404	1.341	2.468
0.35	–0.27	1.14	0.57	1.859	1.841	–1.094	2.405	1.246	2.463
0.50	–0.24	1.15	0.59	1.860	1.838	–1.114	2.402	1.216	2.471
1.00	–0.18	1.15	0.62	1.868	1.831	–1.135	2.397	1.113	2.482

Figure 3.

(a) Structures and spin density distributions of IS, TS, and FS for CH₄ activation at −1 V/Å. (b) Correlations between E_a and F, as well as the mechanism schemes for CH₄ oxidation modulated by negative (left) and positive (right) EEFs, in which red arrows indicate the promotion effect, whereas blue arrows signify the suppression effect. (c) Schematic diagram for surface charge polarization and atomic fluctuation at different EEFs. (d, e) Bond length changes of d₁ and d₂, as well as Bader charge changes of O_br and Ti_5c versus F.

Obviously, the slope in negative EEFs is more than twice than that in positive EEFs, indicating that the negative EFs possess a more sensitive influence on promoting the CH₄ activation progress. Furthermore, it is worth noting that the C–H bond dissociation mechanism under EEF is the same as that in thermocatalytic conditions, maintaining a two-site assisted TS configuration. Taking the reaction in −1 V/Å as an example, the CH₃···Ti_5c bond is 2.280 Å, which is slightly shorter than that without the EF effect (2.347 Å). Additionally, the changes of CH₃···H and O_br···H bonds are also tiny (1.440 vs 1.446 Å and 1.200 vs 1.200 Å, respectively, Figure 3a).

Origins behind the Differences between Photo- and Electro-catalysis

Intuitively, the electrocatalytic approach is similar to photocatalysis, except for the energy source which comes from EEFs.⁴⁶ However, why are the EEFs far less efficient than light irradiation on CH₄ activation, and how do the EEFs affect the CH₄ activation progress? To solve these puzzles and unveil the inherent mechanisms of EEF and light on CH₄ activation, the geometric and electronic structure analyses on TiO₂(110) catalyst under different external driving forces were carried out.

First, the structural responses and charge states of O_br and Ti_5c sites, as the main reactive centers on TiO₂(110), were investigated. In photocatalysis, the hole activated O_br and electron activated Ti_5c are featured with an obvious extension of O_br–Ti and Ti_5c–O bonds (denoted as d₁ and d₂ in Figure 1a, respectively) by 0.196 and 0.205 Å, respectively (Figures 1b and 2).^36,47 Interestingly, under different EEFs from 1 V/Å to −1 V/Å, we found that the d₂ increased, whereas the d₁ decreased gradually (see Table 1 and Figure 3d), with the increment (decrement) magnitude of d₂ (d₁) being only 0.03 Å. Noteworthily, this finding seems to be contrary to our general cognition that, with the negative EEF imposed (corresponding to the oxidative potential), the surface is polarized to be electron-deficient (Figure 3c), which is favorable to the activation of O_br, and d₁ should be elongated compared to the inactivated ones.

To shed light on these different behaviors, we conducted the detailed electronic structure analysis, including plane-averaged electrostatic potential (PAEP) distribution and Bader charge analysis. As illustrated in Figure S2 for the PAEPs of the TiO₂(110) surface under different EEFs, the main difference of electrostatic potential appears at the surface region and above, indicating the relatively stronger polarization of EEF on the catalyst surface than in the bulk. Nevertheless, by Bader charge analysis on surface atoms (Figure 3e and Table 1), we found that the most negative EEF (−1 V/Å) only led to a small fraction of charge depletion for both O_br and Ti_5c atoms, which existed as O_br^–2+α and Ti_5c^+4+β (0 < α, β ≪ 1), respectively. Thus, we can deduce that the electronic field line interacts weakly with groups of positively charged Ti_5c and negatively charged O_br atoms, respectively, and relaxes themselves along the electronic field line to minimize the potential energy, which contributes to the weak fluctuation of surface atoms, i.e., the decrement/increment of d₁/d₂ (see the illustration in Figure 3c). This is different from the photocatalytic mechanism: accompanying the localization of a photoexcited hole on O_br, the interaction between this specific hole-trapped O_br and substrate is strongly weakened, and the charge changes a lot from −1.082 |e| to −0.690 |e|, which is much larger than the effect of EEF.

Second, with the above basic understandings, we are now at the position to discuss why the barrier of CH₄ activation in positive EEF increases much slower than that in negative EEF, as Figure 3b shows. The periodic natural bond orbital (NBO) analysis^48,49 was performed to show the bonding nature of two-site assisted TS complex in thermocatalysis. As can be seen from the H–O_br NBOs in Figure 4a (with details labeled), the H–O_br σ bond is largely composed of the hydrogen s orbital and oxygen p orbitals, which corresponds to the characteristic peak at the energy level of −5.70 eV from the PDOS of TS (Figure 4b). Noticeably, a tiny characteristic peak of the Ti_5c–CH₃ σ bond appears below the Fermi level at the energy level of −5.20 eV with an extremely small occupied area of only 0.008 |e|, which implies a weak C–Ti_5c bond. One can thus speculate that the activation of CH₄ depends mainly on the reactivity of the O_br site, whereas the Ti_5c site contributes secondarily. The higher value of integrated crystal orbital Hamilton population (-ICOHP)⁵⁰ on the H–O_br (2.468 eV/bond) than that on the C–Ti_5c bond (1.341 eV/bond) confirms this speculation (see Table 1). Thus, the observation that a smaller slope exists in positive EEFs can be explained. When the positive EEF is exposed, the surface is in an electron-rich state, which reduces the reactivity of the O_br site for accepting H; simultaneously, the Ti_5c site, as a secondary reaction site, is activated and facilitates binding with the CH₃ group, thus leading to an overall slow increment of reaction barrier with the increment of positive EEFs (see inserted mechanism scheme in Figure 3b). By contrast, the O_br is activated while Ti_5c is inactivated in the negative EFF, and a relatively larger variation of the reaction barrier can be expected.

Figure 4.

(a) NBOs and (b) projected density of state (PDOS) of the TS for CH₄ activation on TiO₂(110) in thermal conditions. (c, d) PDOSs of O_br-2p and Ti_5c-3d bands of the bare surface in different EEFs and light, respectively. (e) Correlation between CH₄ activation barrier (E_a) and p-band center (ε_p) of the O_br site in EEFs (or energy level of characteristic peak of the hole localization, ε_h, in light).

Third, we did the orbital-projected densities of state (PDOS) of surface O_br-2p and Ti_5c-3d bands, which determined the reactivity of the catalyst, to uncover the reactivity trends of photo/electro- versus thermocatalysis. As illustrated in Figure 4c, with the decrement of EEFs from 1 V/Å to −1 V/Å, the energy level of the O_br-2p band (constituting the valence band maximum, VBM) up-shifts gradually, which indicates a better oxidizability of the O_br site to accept the H atom by forming the polar covalent H–O_br bond. Quantitatively, we calculated the occupied p-band center (ε_p) of O_br site, an important indicator to describe the reactivity of the reaction center. The ε_p of the catalyst in different EEFs (+1, +0.5, 0, −0.5, and −1 V/Å) were calculated to be −1.04, −1.01, −0.93, −0.58, and −0.54 eV, respectively, confirming that negative EEFs result in the upshifting of the p-band center of the O_br site. Remarkably, a linear correlation between the activation barriers (E_a) and ε_p exists (Figure S3), implying that the upshifting of the EEF-induced p-band mainly accounts for the decrement of CH₄ activation barrier on TiO₂(110) surface.

In the presence of the photoinduced hole, by analyzing the PDOS in Figure 4d, we found that the formation of O_br^•– radical introduced a localized paramagnetic state (1.75 eV above the Fermi level) in the band gap, manifesting itself in the large upward shift relative to the VBM. The localized paramagnetic hole state serves as a springboard to accommodate the released electron from CH₄ dissociation (i.e., CH₄ + O_br^•– → HO_br^– + CH₃^•), which is characterized with an extremely strong oxidative ability to facilitate CH₄ activation. Interestingly, the decreased E_a with the high-energy-level O_br^•– characteristic peak (1.75 eV) can be also well fitted in the linear correlation between E_a and ε_p in EEFs (Figure 4e).

Based on these computational results, a picture for CH₄ activation on the TiO₂(110) surface in different external fields can be proposed. Upon introducing negative/positive EEFs, the 2p-bands of all the surface O_br sites are polarized to a higher/lower energy level, which enhances/depresses the activation of CH₄. Similarly, for CH₄ activation in light illumination, the unoccupied 2p hole state is the key for surging the catalytic activity. More specifically, the origins behind the superior performance of light than electricity could be understood. In photocatalysis, photogenerated carriers are quantized. These holes activate surface individual lattice oxygen to form O_br^•– radicals, which have a strong oxidizability, whereas for electrocatalysis, groups of surface lattice oxygens are influenced gently by the electronic field line owing to the strong Ti–O chemical bond, which contributes to upshift of 2p-band and thereby prompts oxidizability of the O_br site on CH₄ activation.

Since the EEFs have a weak contribution to the direct CH₄ activation by the TiO₂(110) catalyst, one may naturally question the two experimental results mentioned above that (i) the kinetics of electrocatalytic CH₄ steam reforming can be accelerated over the TiO₂-based electrode, when the potential is above 2.25 V (vs SHE),¹³ and (ii) when combining the electrocatalytic condition with the light irradiation, the activation of CH₄ can occur even on pure TiO₂ catalyst with a much reduced voltage (−0.41 V vs SHE).³⁵ Considering that the H₂O molecule, as the main medium in electrocatalysis and photoelectrocatalysis systems, may have some indispensable roles on CH₄ conversion via the dissociated reactive oxygen species, we examined the electrocatalytic H₂O dissociation and CH₄ activation to answer these two questions.

First, the oxidative water cleavage via the proton coupled electron transfer (PCET) process was tested, including (i) H₂O + * → OH* + H⁺ + e^– and (ii) OH* → O* + H⁺ + e^–, where * denotes Ti_5c.⁵¹ Following this path, the formed surface OH* and O* intermediate would be radical-like, evidenced with Bader charges of −0.011 |e| and −0.014 |e| on TiO₂(110) surface, respectively. Remarkably, our calculations uncover that the OH* and O* species can facilitate CH₄ activation greatly with much lower barriers of 0.36 and 0.45 eV, respectively (Figure 5a). However, this PCET process is thermodynamically difficult to occur, which has to be driven by high oxidative potential. Specifically, the thermodynamic stability of the surface OH*/O* as a function of the applied potential U at pH = 7 is presented in Figure 5b (see details in Table S4). We can see that only when the potential increases to 1.85 V, the OH* intermediate becomes thermodynamically favored to be formed; when the potential continues to increase to 2.18 eV, the O* begins to be formed. This accords with the electrocatalytic CH₄ steam reforming experiment over TiO₂/RuO₂/V₂O₅ electrode that CH₄ can be activated when the applied potential is above 2.25 V (vs SHE)^13,34 and is also arguably in line with the observation that hydroxide and oxygen adatoms are produced on Ti_5c sites under a voltage pulse of 2.4 or 2.8 V on the TiO₂(110) surface.⁵²

Figure 5.

(a) Optimized TSs and barriers for CH₄ activation with the assistance of OH^•/O^• species from H₂O dissociation via PCET and OH^– intermediate via PT progress, respectively. (b) Thermodynamic phase diagram of H₂O conversion on TiO₂(110) as a function of the potential at the pH = 7, including (i) H₂O + * → OH* + H⁺ + e^– (red), (ii) OH* → O* + H⁺ + e^– (blue), and (iii) H₂O + * → O* + 2H⁺ + 2e^– (black). (c) Geometries of PT progress for H₂O dissociation with the proton released into the liquid phase.

Thus, we can comprehend that the EEF effect at high oxidative potential can hardly activate CH₄ directly but could alternatively activate CH₄ (when H₂O exists) by resorting to the surface active species (i.e., OH*) from the oxidative water cleavage indirectly. In addition, it may be worth noting that the applied potential is better when it is not higher than 2.18 V, as the reactive O* species can not only facilitate CH₄ activation (E_a = 0.45 eV, Figure 5a) but also simultaneously promote oxygen evolution via O–O coupling.⁴⁷ This is in line with the experimental observation over the TiO₂/RuO₂/V₂O₅ electrode, that the CH₄ oxidation process is prompted below 2.25 V (vs SHE) but is inhibited at a more positive potential with simultaneous oxygen evolution.¹³

Second, the dissociation of water via the deprotonation progress (i.e., proton-transfer mechanism, PT) into the solution was comparatively calculated at the H₂O/TiO₂(110) interface by the ab initio molecular dynamics (AIMD) simulation⁵³ (see method details in Supporting Information), which corresponds to Ti_5c–H₂O → Ti_5c–OH^– + H⁺ (Figure 5c). In comparison with the PCET process, this PT process occurs relatively easily with a low barrier of 0.39 eV and reaction enthalpy of 0.21 eV at the thermal condition. Furthermore, our calculations verified that the potential-induced EF effect would efficiently affect these proton transfer kinetics to form OH^– species. Specifically, the negative EFs (corresponding to the oxidative potential) lead to an evident promotional effect on H₂O dissociation (Table S5), resulting from the dipole moment and large polarizability of the O–H bond;⁵⁴ for example, there is a much decreased barrier of 0.08 eV at −1 V/Å. It implies that a low potential could be strong enough to facilitate the hydroxide OH^– formation. However, different from the radical-like OH* species in PCET progress, the formed OH^– can hardly facilitate CH₄ activation with a high barrier of 0.93 eV, which is comparable with the pristine inert O_br (1.12 eV, Figure 5a). Thus, the charge state of the OH intermediate on TiO₂(110) is critical for modulating its reactivity.

Significantly, when integrating with the light irradiation (see schematic in Figure 6), the hydroxide OH^– generated under low potential (or low EEF) could bring about an evident synergy effect for CH₄ activation. Under a pure photocatalytic condition, it was reported to be only a H₂O dissociation probability of about 1.7% and 3.7% on TiO₂(110) at the 400 and 355 nm light irradiation, respectively,⁵² largely restricted by the high recombination rate of the photoexcited electron–hole (namely, the low concentration of the surface hole)^20,55 and relatively low kinetics of H₂O photocatalytic oxidation progress (i.e., H₂O + h⁺ → OH^• + H⁺).^36,47 However, with the cooperation of photo- and electrocatalysis, the photohole can readily trap the OH^– species (i.e., OH^– + h⁺ → OH^•) generated from electrocatalysis-accelerated H₂O deprotonation progress with a hole-trapping capacity of −0.77 eV,³⁶ to form OH^• radical for better CH₄ activation. In this sense, one can rationalize that, experimentally, just a low oxidative potential is required to accomplish the efficient CH₄ activation when combining with the photocatalysis.³⁵ In addition, our calculations identify that the imposed negative EEF under the oxidative potential can also enhance the direct distribution of hole/electron along/against the electric-field line, with the hole polaron being preferentially accumulated at the catalyst surface (see details in the Supporting Information, Figures S4 and S5).⁵⁶ Accordingly, the synergistic effect of photoelectrocatalalysis can be further enhanced for motivating the formation of reactive OH^• radical. It may be worth noting that this finding could also explain the experimental phenomenon that the amount of O₂ increases with the increasing of EEFs in photoelectrocatalalysis.³⁵

Figure 6.

Schematic diagram for OH^• formation via PCET and PT+ET progresses. PCET: H₂O* → OH^• + H⁺ + e^–; PT: H₂O* → OH^– + H⁺; and ET: OH^– → OH^• + e^–. Note: 1.64 eV for ET progress; see details in Figure S6.

In brief, the above discussions indicate that the electrocatalytic CH₄ conversion eventually needs to indirectly resort to the reactive oxygen species (e.g., OH^• radical). As Figure 6 illustrates, the formation of OH^• radical can be accomplished from the oxidative cleavage of H₂O via the PCET mechanism; however, the required applied oxidative potential is rather high. Alternatively, OH^• could be formed via the sequential PT+ET processes, during which the PT process occurs relatively easily kinetically on TiO₂(110) and can be efficiently modulated by the low EEF, whereas the ET process is energy-intensive with an energy cost over 1.64 eV (see details in Figure S6). In this circumstance, the light irradiation could be a competent approach to surmount this high energy demand and drive this ET process (OH^– + h⁺ → OH^•). In this regard, the integrated photoelectrochemical strategy should be promising for achieving efficient CH₄ activation.

Conclusions

In summary, the present study provides a comprehensive comparison between the photo- and electrocatalysis and uncovers their critical roles and unique mechanisms for modulating CH₄ oxidation by TiO₂(110) in the absence or presence of water. It turns out that a positively polarized surface is helpful for CH₄ activation, whereas the contribution is largely inferior to a photogenerated hole in photocatalysis. The origins behind it were revealed: the negative EEFs contribute to the O_br-2p band upshift slightly toward the Fermi level, indicating the more reactivity of lattice oxygen to make a bond with H atom. For comparison, the photocatalysis reaction proceeds with the assistance of a localized gap state as a springboard, whose oxidizing ability is largely greater than that of the electro-activated lattice oxygen. Differently, the essential role of EEFs on CH₄ conversion on TiO₂(110) can be strengthened in the presence of water. EEF can trigger the oxidative water cleavage via the PCET process under high oxidative potential, resulting in the formation of OH* radical-like species, which are reactive for CH₄ activation. Alternatively, EEF can facilitate H₂O deprotonation via the PT mechanism to form the surface hydroxide OH^–, which can easily trap the photoexcited hole to generate OH^•. In addition, it can also will promote the carrier spatial separation by stabilizing the hole–polaron configuration along the electric-field line in photoelectrocatalalysis, rationalizing the synergistic photoelectrocatalytic effect for boosting CH₄ activation on TiO₂(110) at low applied potential.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.1c00466.

• DFT setting details; calculations of free energy; promotion effect of EEF on the separation of carriers; test on the CH₄ activation barrier via DFT+U versus HSE06 functional; reaction barrier of H₂O deprotonation under different EEFs; optimized geometries and spin density plots for CH₄ activation assisted photogenerated electron; and the plane-averaged electrostatic potential of TiO₂(110) under different EEFs (PDF)

The authors declare no competing financial interest.