Density Functional Theory Study of the Effects of NH₃ and NO₂ on Hg⁰ Oxidation over V₂O₅/TiO₂ Under Fast Selective Catalytic Reduction Conditions

Abstract

The effects of NO₂ and NH₃ on Hg⁰ oxidation under fast selective catalytic reduction conditions remain unclear. To address this problem, we performed density functional theory calculations on the V₂O₅/TiO₂(001) surface in four states: clean surface, surface covered by NH₃, surface covered by NO₂, and surface where species are co-adsorbed. On the clean surface, Hg⁰ interacts weakly with oxygen sites; the adsorption energies at terminal (O_t), bridging (O_b) and interfacial bridging (O_b′) oxygen sites are −12.31, −12.58, and −21.22 kcal mol⁻¹, respectively. Direct oxidation by lattice oxygen encounters high activation barriers, ranging from 146.817 to 184.529 kcal mol⁻¹. NH₃ hinders the capture of Hg⁰. Since NH₃ preferentially occupies the vanadium-related active sites, the interaction between Hg⁰ and the catalyst is weakened, and the adsorption energies become less negative, ranging from −11.84 to −11.04 kcal mol⁻¹. NO₂ shows the opposite effect. It strengthens Hg⁰ adsorption near V-centered regions and provides an alternative oxygen source for Hg⁰ oxidation. The NO₂-assisted oxygen-transfer pathways have lower calculated barriers, ranging from 74.581 to 108.545 kcal mol⁻¹. When NH₃ and NO₂ are co-adsorbed, the barrier increases to 113.326 kcal mol⁻¹. Even so, NO₂-derived oxygen provides a relatively lower-energy oxygen-transfer route than lattice oxygen among the pathways examined.

1. Introduction

NH₃ selective catalytic reduction (SCR) is an efficient and mature NOx reduction technology that has been widely applied in high-temperature equipment, including coal-fired boilers, cement kilns, and municipal waste incinerators. Among the two SCR routes, the fast-SCR (FSCR) pathway is considered superior to the standard SCR (SSCR) route at lower temperatures, provided that both NO and NO₂ are present in the gas mixture [1].

SSCR: 4NH₃ + 4NO + O₂ → 4N₂ + 6H₂O

(1)

FSCR: 2NH₃ + NO + NO₂ → 2N₂ + 3H₂O

(2)

In addition to NO_x, flue gas from these sources contains trace mercury species [2,3]. Elemental mercury (Hg⁰) is particularly difficult to remove because of its high volatility, chemical inertness, and poor water solubility. As a result, it is not efficiently captured by conventional particulate-control devices or wet flue-gas desulfurization systems [4]. Catalytic oxidation of Hg⁰ into more soluble and easily captured oxidized mercury species is therefore regarded as a key step for deep mercury removal [5]. As SCR devices have been widely applied in combustion systems, using the same catalyst for simultaneous NOx reduction and Hg⁰ oxidation provides a practical solution for multi-pollutant control [6]. Therefore, vanadium-based SCR catalysts, particularly V₂O₅/TiO₂ and V₂O₅–WO₃/TiO₂, have attracted widespread attention due to their ability to simultaneously support denitrification and mercury removal [6,7].

Experimental studies have long demonstrated that the Hg⁰ oxidation process over vanadium-based SCR catalysts is closely related to the surrounding SCR reaction atmosphere. Kamata et al. reported that NH₃ strongly inhibits Hg⁰ oxidation over commercial V₂O₅(WO₃)/TiO₂ catalysts, indicating significant competition between NH₃ and the surface processes involved in mercury oxidation [8]. Yang et al. subsequently discovered that the oxidation of Hg consumes surface lattice oxygen and promotes the reduction of V⁵⁺ to V⁴⁺; meanwhile, the NO conversion rate over Hg-exposed V₂O₅–WO₃/TiO₂ catalysts decreases within the 200–300 °C temperature range, further indicating a competitive relationship between Hg-related surface chemical processes and the NH₃-SCR reaction [9]. In contrast, under suitable conditions, NO-derived species can promote the conversion of Hg⁰. Liu et al. observed that NO enhances Hg⁰ oxidation over commercial V₂O₅–WO₃/TiO₂ catalysts (particularly in the presence of O₂), whereas excess NO inhibits Hg⁰ adsorption. They further proposed that NO-derived surface species, including NO₂, facilitate both the adsorption and oxidation of Hg⁰ [10]. These early studies established that NH₃ and NO_x species exert markedly different effects on the chemical behavior of Hg⁰ over vanadium-based SCR catalysts. Recent experimental studies have further confirmed and refined these patterns. Shi et al. demonstrated that the simultaneous removal of NO_x and Hg⁰ over V₂O₅/TiO₂ catalysts depends heavily on the surface density and aggregation state of supported VO_x species; specifically, the Hg⁰ oxidation rate increases as the VO_x density approaches monolayer coverage, highlighting the importance of the local vanadium species environment [11]. Park et al. subsequently showed that increasing the NH₃ concentration over V₂O₅–WO₃/TiO₂ catalysts enhances NO reduction but simultaneously reduces the Hg⁰ oxidation rate, providing direct, recent evidence of the competition between the NH₃-SCR reaction and the Hg⁰ oxidation process [12]. Notably, Jia et al. investigated the synergistic removal of NO_x and oxidation of Hg⁰ over V₂O₅–WO₃/TiO₂ catalysts under a simulated FSCR atmosphere, observing high Hg⁰ oxidation activity in the presence of NO₂. Two mechanisms have been proposed: one involving surface-mediated oxidation of adsorbed Hg⁰, and the other a homogeneous NO₂-related pathway; the latter was found to alleviate the inhibitory effect of NH₃ on Hg⁰ oxidation [7]. Collectively, these experimental studies present a consistent qualitative picture: NH₃ tends to inhibit Hg⁰ conversion through competitive surface interactions, whereas chemical processes derived from NO/NO₂ can promote Hg⁰ adsorption or oxidation under suitable conditions. However, these macroscopic observations do not directly reveal how NH₃ and NO₂ alter the local V–O structural environment, how their pre-adsorption affects subsequent Hg⁰ adsorption, or how these changes influence the elementary oxygen-transfer steps involved in Hg⁰ oxidation. These unresolved questions regarding specific active sites represent a significant gap in our understanding of the chemical behavior of Hg⁰ under FSCR conditions.

Theoretical calculations have provided crucial atomic-scale insights into the adsorption and oxidation of Hg⁰ on supported vanadium catalysts. Suarez Negreira and Wilcox investigated Hg oxidation over vanadium–titanium (V₂O₅/TiO₂) SCR catalysts under typical flue gas conditions, establishing an atomic-level theoretical framework to analyze the interactions between mercury species and the supported vanadium [13]. Subsequently, Gao and Li utilized a periodic V₂O₅/TiO₂(001) model to compare Hg⁰ oxidation pathways, revealing the significant roles of local V–O structures, lattice oxygen, and surface redox processes in determining Hg⁰ reactivity [14]. Later density functional theory (DFT) studies extended this framework to modified catalyst systems. Gao et al. investigated V₂O₅/TiO₂ catalysts modified with Cr, Fe, Co, and Nb, demonstrating that altering the local electronic and oxygen environments influences Hg⁰ oxidation behavior [15]. Wu et al. further examined WO₃-containing V₂O₅/TiO₂ catalysts, finding that WO₃ affects surface electronic structure, mercury species adsorption, and the energetics of HCl-assisted oxidation [16]. Recent computational studies have also addressed Mo/Sb- and Ru-modified V₂O₅/TiO₂ systems, similarly confirming that changes in local electronic structure, oxygen environment, and adsorption characteristics can significantly alter Hg⁰ oxidation pathways and energetics [17,18]. Despite these advances, previous theoretical research has primarily focused on the intrinsic oxidation of Hg⁰, HCl-assisted reaction pathways, or catalyst modification; relatively few studies have addressed the competitive effects of key FSCR components—specifically NH₃ and NO₂—on Hg⁰ adsorption and oxidation within the framework of the supported V₂O₅ system. This issue is particularly important because the aforementioned experimental studies indicate that NH₃ inhibits while NO₂ promotes Hg⁰ conversion, yet the atomic-scale mechanisms underlying these opposing effects remain unclear. Specifically, it remains to be elucidated how the pre-adsorption of NH₃ and NO₂ alters the local V–O environment, whether these species modify the surface sites associated with Hg⁰ adsorption or compete for them, and how their presence influences the energetics of the subsequent oxygen transfer process.

Based on a V₂O₅/TiO₂(001) surface model, this study employs DFT calculations to investigate the atomic-scale effects of key components involved in the FSCR process on Hg⁰ adsorption and oxidation. Four surface environments were systematically examined: the clean surface, NH₃-precovered surface, NO₂-precovered surface, and NH₃/NO₂ co-precovered surface. By comparing optimized adsorption structures, adsorption energies, electronic properties, and characteristic oxygen-transfer pathways within a unified computational framework, we aim to elucidate how NH₃ and NO₂ modify the local V–O environment and the relative energetics of the Hg⁰ conversion process.

2. Computational Methods

2.1. Catalyst Surface Model Construction

To describe dispersed vanadia species supported on titania, a periodic V₂O₅/TiO₂(001) surface model was constructed with reference to the benchmark studies of Gao et al. and Zhang et al. [14,19]. The model adopts the widely used anatase ad-molecule representation, which is well suited for simulating low-coverage vanadia species. In this configuration, the supported vanadia is treated as an adsorbed molecular moiety on the TiO₂ substrate. Because the anatase TiO₂(001) facet has been identified as one of the most catalytically active interfaces [20], it was selected as the substrate in the present work. The substrate was modeled as a two-layer anatase TiO₂(001) slab with a 3 × 2 surface supercell, on which a vanadia dimer was anchored to represent the catalytically relevant V₂O₅ moiety. The corresponding periodic cell parameters were a = 11.328 Å, b = 7.552 Å, and c = 20.529 Å, with α = β = γ = 90°. The periodic cell contains 12 Ti, 29 O, and 2 V atoms (Ti₁₂V₂O₂₉), corresponding stoichiometrically to a Ti₁₂O₂₄ slab with a supported V₂O₅ species. This simplified V₂O₅/TiO₂(001) surface was adopted as a controlled model to isolate the effects of the selected FSCR components on the local V–O environment and Hg⁰ conversion.

Motivated by experimental studies showing that NH₃ and NO₂ in FSCR atmospheres can significantly influence Hg⁰ oxidation over vanadia-based SCR catalysts [7], the calculations were organized in a stepwise manner to clarify the individual and combined effects of these species on Hg⁰ conversion over V₂O₅/TiO₂(001). First, Hg⁰ adsorption and oxidation on the clean surface were investigated to establish a reference. Next, the adsorption behaviors of NH₃ and NO₂ were examined separately, followed by Hg⁰ adsorption and oxidation on the corresponding precovered surfaces. Finally, NH₃/NO₂ co-precovered surfaces were considered to evaluate their combined influence on Hg⁰ adsorption and oxidation. For each stage, multiple initial configurations were examined to identify representative optimized structures and possible reaction precursors. The resulting adsorption structures were subsequently used for transition-state calculations and comparison of the corresponding oxidation pathways. The computational surface model used throughout this workflow is shown in Figure 1.

Figure 1. Computational model of the V₂O₅/TiO₂(001) surface used for DFT simulations.

The oxygen atoms in the supported vanadia species can be divided into three types according to their local coordination environments, i.e., terminal oxygen (O_t), bridging oxygen in the V–O–V linkage (O_b), and interfacial bridging oxygen in the V–O–Ti linkage (O_b′). This classification serves as a useful structural framework for identifying potential adsorption sites involved in Hg⁰ oxidation under FSCR conditions.

To assess the reliability of the adopted surface model, the optimized bond lengths of the V₂O₅/TiO₂(001) structure were compared with previously reported theoretical and experimental values, as listed in Table 1. The calculated bond lengths are in good agreement with the literature data, with only small deviations, indicating that the present model is suitable for the subsequent calculations.

Table 1. Comparison of characteristic bond lengths in the optimized V₂O₅/TiO₂(001) model.

Bond Type	This Study	Theoretical Value [14,21]	Experimental Value [22]
V–Ot	1.60	1.60, 1.604	1.58
V–Ob	1.80	1.80, 1.804	1.78
V–Ob′	1.77	1.77, 1.764	1.76
Ti–Ob′	1.82	1.83, 1.833	N/A

2.2. DFT Calculation Details

First-principles calculations were performed using the DMol³ module implemented in BIOVIA Materials Studio 2017 (Dassault Systèmes, San Diego, CA, USA) [23]. To account for the open-shell nature of the gas-phase species involved in the FSCR system, spin-unrestricted calculations were employed throughout, with formal spin used as the initial state to facilitate self-consistent field (SCF) convergence. Following previous studies, exchange–correlation interactions were described using the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional, which has been widely used in previous DMol³ studies of V₂O₅/TiO₂-based SCR catalysts and Hg⁰ oxidation pathways [16,24]. A double-numerical basis set with polarization functions (DNP) was used to expand the valence electron wave functions. To account for relativistic effects, the core electrons of vanadium were treated using density functional semi-core pseudopotentials (DSPP). In addition, the Tkatchenko–Scheffler (TS) dispersion correction was included to describe non-covalent van der Waals (vdW) interactions between adsorbates and the catalyst surface. To improve the description of localized V 3d states involved in the redox chemistry of the supported vanadia species, a Hubbard U correction of 4.0 eV was applied to the V 3d orbitals throughout the calculations [25]. The real-space global orbital cutoff radius was set to 5.2 Å, and a thermal smearing of 0.01 Ha was applied to the orbital occupations.

Geometry optimization was considered converged when the changes in total energy, maximum force, and maximum displacement were below 5.0 × 10⁻⁵ Ha, 2.0 × 10⁻² Ha Å⁻¹, and 5.0 × 10⁻² Å, respectively. Brillouin-zone sampling was performed using Monkhorst–Pack k-point meshes of 3 × 3 × 3 for bulk anatase TiO₂ and 3 × 3 × 1 for the surface slab models. To avoid spurious interactions between periodic images, a vacuum layer of 15 Å was introduced along the z direction.

2.3. Adsorption and Transition-State Calculations

For adsorption calculations, the atoms in the bottom layer of the TiO₂(001) slab were fixed, whereas the upper-layer atoms, the supported V₂O₅ species, and the adsorbed molecules were fully relaxed. The adsorption strength of each gaseous species on the catalyst surface was evaluated using the adsorption energy, E_ads, defined as

E_ads = E_{adsorbate/substrate} − E_adsorbate − E_substrate

(3)

where E_{adsorbate/substrate} is the total energy of the adsorbate–surface complex, E_adsorbate is the energy of the isolated adsorbate, and E_substrate is the energy of the clean surface. A more negative E_ads value indicates stronger adsorption.

To map the elementary reaction pathways for Hg⁰ oxidation under FSCR conditions, transition-state (TS) searches were performed using the complete linear synchronous transit/quadratic synchronous transit (LST/QST) method. All optimized stationary points were further verified by vibrational frequency analysis. A stable intermediate was confirmed by the absence of imaginary frequencies, whereas a valid transition state was identified by the presence of a single imaginary frequency along the reaction coordinate. The activation barrier, Ea, for each elementary step was calculated as

E_a = E_TS − E_IM

(4)

where E_TS and E_IM denote the energies of the transition state and the corresponding adsorbed intermediate, respectively.

Since thermal and entropic corrections were not included, the calculated adsorption energies and activation barriers are interpreted primarily in terms of qualitative energetic trends among the different pathways within the present computational framework.

3. Results and Discussion

3.1. Hg⁰ Adsorption and Oxidation on V₂O₅/TiO₂(001)

3.1.1. Hg⁰ Adsorption Behavior

We first explored the adsorption and oxidation behavior of Hg on the clean V₂O₅/TiO₂ (001) surface. In this section, the configurations are named based on the following principles: for example, Hg-O_t means that Hg is initially placed above the O_t site and then geometrically optimized. The optimized structures are shown in Figure 2, and the corresponding adsorption energies, structural parameters and Mulliken charges are shown in Table 2.

Figure 2. Optimized configurations of Hg⁰ adsorbed at different oxygen sites on V₂O₅/TiO₂(001).

Table 2. Adsorption energies, Hg–O distances, and Mulliken charges for Hg⁰ adsorption on V₂O₅/TiO₂(001).

	Initial Sites	Hg-Ot	Hg-Ob	Hg-Ob′
Parameters
Eads (kcal mol−1)		−12.307	−12.578	−21.224
DHg-O (Å)		3.524, 3.447	3.222	3.471, 3.483
QHg (e)		0.113	0.115	0.167

During the structural relaxation, the Hg-O_b′ configuration exhibits the most negative adsorption energy (−21.22 kcal mol⁻¹), which is significantly lower than those of Hg-O_t (−12.31 kcal mol⁻¹) and Hg-O_b (−12.58 kcal mol⁻¹). This indicates that the interface O_b′ site associated with the V-O-Ti bond is the most favorable region for Hg adsorption. In contrast, the similar adsorption energies of Hg-O_t and Hg-O_b indicate that the adsorption stability of Hg⁰ at the terminal oxygen and V-O-V bridge oxygen sites is relatively weak. The optimized geometric structure reveals further insights. In the Hg-O_t model, Hg moves above the vanadium dimer and stabilizes between two adjacent oxygen atoms, with Hg-O distances of 3.524 and 3.447 Å, respectively. In the Hg-O_b configuration, Hg⁰ is located above the central bridge oxygen of the V-O-V unit, showing the shortest Hg-O distance (3.222 Å). For the Hg-O_b′ configuration, Hg migrates to the oxide-support interface and is stabilized between two adjacent interface oxygen atoms, with Hg-O distances of 3.471 and 3.483 Å, respectively. Despite these structural differences, all the Hg-O distances are still relatively long, indicating that no strong Hg–O bond is formed. Interestingly, although Hg-O_b′ does not show the shortest bond length, it shows the strongest adsorption. This indicates that the adsorption stability is affected not only by geometric factors but also by the electronic properties of the adsorption sites, especially at the oxide-support interface.

Mulliken charge analysis further supports this explanation. The charge of Hg in Hg-O_t, Hg-O_b and Hg-O_b′ configurations is 0.113 e, 0.115 e and 0.167 e, respectively. The small charge transfer in all configurations indicates that Hg basically maintains neutral characteristics after adsorption, which is consistent with the weak interaction inferred from the long Hg-O distance. However, the relatively high positive charge in the Hg-O_b′ configuration indicates a stronger electron perturbation at the interface V-O-Ti site. Additional insights come from the partial density of states (PDOS) diagram shown in Figure 3. For Hg-O_t and Hg-O_b, the PDOS of the corresponding oxygen atoms shows little change, and the peak shape, intensity and energy distribution exhibit almost no change. This weak electronic response is consistent with the small charge transfer and similar adsorption energy observed in these configurations. In contrast, the Hg-O_b′ system undergoes more significant electronic changes. After Hg adsorption, the p state of the interface O_b′ atom moves towards lower energy, reflecting a stronger local electronic structure perturbation. There is obvious hybridization between the d state of Hg and the p state of O_b′, whereas this orbital coupling is much weaker in Hg-O_t and Hg⁰-O_b. These electronic structure features explain the stronger adsorption of Hg at the O_b′ site. In other words, the higher stability is mainly due to the interfacial electron interaction, especially the Hg(d)-O_b′(p) coupling. Combined with the larger Mulliken charge on the Hg⁰ atom in the Hg-O_b′ configuration, the PDOS results show that the V-O-Ti interface region provides the most favorable local environment for the electron perturbation and adsorption stability of Hg on the V₂O₅/TiO₂(001) surface.

Figure 3. Local atom- and orbital-resolved PDOS for Hg⁰ adsorption on V₂O₅/TiO₂(001), focusing on Hg and the surface O atom directly involved in each adsorption environment.

3.1.2. Hg⁰ Oxidative Desorption Behavior

Transition-state calculations were carried out to compare Hg⁰ oxidation and oxidative desorption at the O_t, O_b, and O_b′ sites on V₂O₅/TiO₂(001). The energy profiles are shown in Figure 4, and the optimized structures along the reaction pathways are presented in Figure 5. All three pathways are strongly endothermic and exhibit high calculated activation barriers, indicating that direct oxidation of Hg⁰ by lattice oxygen is energetically demanding within the present computational framework. Among the three sites, Hg–O_t gives the lowest activation energy and reaction energy, 146.817 and 134.941 kcal mol⁻¹, respectively. Hg–Ob has the highest reaction energy, 163.136 kcal mol⁻¹, while Hg–O_b′ gives the highest activation energy, 184.529 kcal mol⁻¹. Among the three lattice-oxygen pathways examined, the Hg–O_t pathway has the lowest calculated activation barrier.

Figure 4. Reaction energy profiles for Hg⁰ oxidation and oxidative desorption on V₂O₅/TiO₂(001).

Figure 5. Optimized reactant, transition-state, and product structures for Hg⁰ oxidation on V₂O₅/TiO₂(001).

For the Hg–O_t pathway, Hg⁰ is first located above the terminal oxygen region, with an Hg–O distance of 3.524 Å. During the reaction, Hg gradually moves toward O_t, and the Hg–O distance decreases to about 2.771 Å at the transition state. In the final state, an HgO-like species forms, with an Hg–O bond length of about 2.042 Å. The pathway involves direct transfer of the terminal lattice O atom to Hg, followed by formation and detachment of an HgO-like product. A key factor behind the relatively low barrier is the structural degree of freedom of O_t: unlike the oxygen atoms locked at the bridging position, the terminal oxygen is less constrained by the surrounding skeleton, so its transfer from the surface requires less extensive structural rearrangement. The Hg–O_b reaction pathway follows a similar oxygen transfer process but entails a higher energy cost. In the reactant state, Hg⁰ is situated near the O_b atom of the V–O–V linkage; the Hg–O distance shortens from 3.222 Å in the reactant state to 2.600 Å in the transition state and further to 2.028 Å in the product state. Since the Ob atom acts as a structural bridge between two V centers, the removal of the oxygen atom disrupts the V–O–V framework and induces significant local structural distortion, consistent with the high calculated reaction energy. For the Hg–O_b′ pathway, the V–O–Ti environment at the interface undergoes more drastic restructuring. During progression toward the transition state, the terminal O_t atom migrates toward the O_b′ region while the local V–O–Ti framework rearranges. This coupled oxygen migration and interfacial restructuring gives the Hg–O_b′ pathway the highest calculated activation barrier among the three clean-surface pathways examined.

In general, on the V₂O₅/TiO₂(001) surface, the direct oxidation of Hg⁰ initiated by lattice oxygen involves overcoming relatively high energy barriers for all three pathways investigated. Among them, the Hg–O_t pathway exhibits the lowest calculated activation barrier; in contrast, the Hg–O_b and Hg–O_b′ pathways entail higher barriers, which are attributed to the disruption of the V–O–V framework and the restructuring of the V–O–Ti interfacial environment, respectively.

3.2. Effect of NH₃ on Hg⁰ Adsorption and Oxidation on V₂O₅/TiO₂(001)

3.2.1. NH₃ Adsorption on V₂O₅/TiO₂(001)

We studied the adsorption of NH₃ on V₂O₅/TiO₂ (001) to determine its optimal adsorption site and its interaction with surface vanadium–oxygen species. The adsorbed NH₃ can occupy sites that Hg may also utilize and can change the local electronic environment for the later adsorption and oxidation of Hg. Five initial configurations are considered: NH₃_N-O_t, NH₃_N-O_b, NH₃_H-O_b, NH₃_H-O_t, and NH₃_N-V, as shown in Figure 6. Their adsorption energies, main structural parameters and Mulliken charges are listed in Table 3. These models compare two adsorption modes: N-directed adsorption (N lone pair electrons interacting with the surface) and H-directed adsorption (the interaction between H atoms and surface oxygen species).

Figure 6. Optimized configurations of NH₃ adsorbed on V₂O₅/TiO₂(001).

Table 3. Adsorption energies, characteristic distances, and Mulliken charges for NH₃ adsorption on V₂O₅/TiO₂(001).

	Initial Sites	NH3_N-Ot	NH3_N-Ob	NH3_H-Ob	NH3_H-Ot	NH3_N-V
Parameters
Eads (kcal mol−1)		−32.139	−10.452	−14.118	−13.692	−32.285
D (Å)		2.237	3.088	2.095	2.090	2.232
QNH3 (e)		0.324	0.126	0.163	0.146	0.328

The N-oriented model with NH₃ close to the exposed vanadium center is obviously the most stable. The NH₃_N-O_t and NH₃_N-V configurations exhibit more negative adsorption energies of −32.139 and −32.285 kcal mol⁻¹, respectively, with characteristic adsorption distances of 2.237 and 2.232 Å and Mulliken charges of 0.324 and 0.328 e, respectively. Although the two initial configurations are different, the structural relaxation drives them toward essentially the same final geometry, indicating that the V-O_t local environment is essentially suitable for strong NH₃ coordination. In contrast, the interaction between ammonia nitrogen and bridge oxygen (NH₃_N-O_b) is much weaker: the adsorption energy is reduced to only −10.452 kcal mol⁻¹, the optimized distance is extended to 3.088 Å, and the charge transfer amount is moderate at 0.126 e. Unlike the exposed vanadium sites, the bridging oxygen in the V-O-V bond does not act as an accessible electron-pair acceptor (a Lewis acid center) and cannot strongly bind to the nitrogen lone pair electrons. When ammonia interacts with surface oxygen through its hydrogen atom, the bonding strength is between these extremes, but remains moderate. The adsorption energies of NH₃_H-O_b and NH₃_H-O_t are −14.118 and −13.692 kcal mol⁻¹, respectively. There is a similar short contact distance near 2.09 Å, and the charges are 0.163 e and 0.146 e, respectively. These values are significantly lower, i.e., substantially less negative, than those of the stable N-coordination configuration, and the nature of the interaction is better described as similar to a stable hydrogen bond, rather than a strong donor–acceptor bond.

The PDOS results in Figure 7 explain the electronic origin of this adsorption trend. As shown in Figure 7a–e, clear adsorbate–surface hybridization appears only in the NH₃_N–O_t and NH₃_N–V configurations, whereas the other three cases show no obvious hybridization features. Since NH₃_N–O_t and NH₃_N–V relax to nearly the same stable adsorption structure, we selected NH₃_N–V as the representative case for comparing the PDOS before and after adsorption, as shown in Figure 7f, to represent the stable N-coordinated state. After NH₃ adsorption, a new feature appears in the N p states, accompanied by obvious hybridization with V d states. Meanwhile, the V d states shift after adsorption, demonstrating that the exposed V center is electronically perturbed by NH₃ coordination. These PDOS results are fully consistent with the adsorption energies, optimized distances, and Mulliken charge analysis. NH₃ adsorption at NH₃_N–V is mainly governed by direct N(p)–V(d) interaction between the NH₃ lone pair and the exposed V center, rather than by simple electrostatic attraction. This orbital interaction accounts for the much higher stability of NH₃_N–V and NH₃_N–O_t compared with the O_b-centered and H-oriented configurations. Overall, NH₃ preferentially adsorbs at exposed V-related sites, especially in the V–O_t local environment, whereas adsorption near the O_b region or through H-oriented interaction is relatively weak.

Figure 7. PDOS analysis of NH₃ adsorption on V₂O₅/TiO₂(001).

3.2.2. Effect of Pre-Adsorbed NH₃ on Hg⁰ Adsorption and Oxidation

Hg⁰ Adsorption Behavior

The effect of pre-adsorbed NH₃ on Hg⁰ adsorption was examined using three representative co-adsorption configurations on V₂O₅/TiO₂(001), as shown in Figure 8. In the notation used here, NH₃–O_b+ Hg–O_t denotes a configuration in which NH₃ is first adsorbed in the O_b-related region, followed by Hg⁰ adsorption near the O_t site. The same notation is used for the other co-adsorption models. Two types of NH₃-precovered surfaces were considered. The first involves NH₃ occupation of an oxygen-related region, represented by NH₃–O_b, and mainly reflects local site blocking. The second involves N-coordinated NH₃ adsorption at an exposed V-related site, represented by NH₃_N–V, where both the coordination structure and the electronic environment of neighboring V–O units may be modified. The corresponding adsorption energies, characteristic distances, and Mulliken charges are listed in Table 4.

Figure 8. Optimized co-adsorption configurations of Hg⁰ on NH₃-precovered V₂O₅/TiO₂(001).

Table 4. Adsorption energies, characteristic distances, and Mulliken charges for Hg⁰ adsorption on NH₃-precovered V₂O₅/TiO₂(001).

	Initial Sites	NH3-Ob + Hg-Ot	NH3_N-V + Hg-Ot	NH3_N-V + Hg-Ob
Parameters
Eads (kcal mol−1)		−11.04	−11.50	−11.84
D (Å)		3.023	3.744, 3.850	3.178
QHg (e)		0.088	0.053	0.056

Compared with the clean surface, the adsorption of Hg⁰ is weakened in all NH₃ pre-adsorbed configurations. The adsorption energies of Hg-O_t and Hg-O_b on the intrinsic V₂O₅/TiO₂(001) surface are −12.31 kcal mol⁻¹ and −12.58 kcal mol⁻¹, respectively. These adsorption energies become less negative after NH₃ pre-adsorption, reaching −11.04 (NH₃-O_b + Hg-O_t), −11.50 kcal mol⁻¹ (NH₃_N-V + Hg-O_t) and −11.84 kcal mol⁻¹ (NH₃_N-V + Hg-O_b). A similar trend is observed in the Mulliken charge analysis. The charge on Hg⁰ decreases from 0.113 e and 0.115 e on the clean Hg-O_t and Hg-O_b surfaces to 0.088 e, 0.053 e and 0.056 e in the three NH₃ precovered configurations. The less negative adsorption energies and reduced charge transfer indicate that the pre-adsorbed NH₃ weakens the interaction between Hg⁰ and V₂O₅/TiO₂(001) surfaces.

For the NH₃-O_b+ Hg-O_t configuration, the optimized Hg⁰-surface distance is 3.023 Å, and the adsorption energy is −11.04 kcal mol⁻¹. Although Hg⁰ is still close to the O_t region, NH₃ located in the adjacent O_b-related region weakens its adsorption stability. This behavior is mainly related to the local occupation of the adsorption environment, which limits the entry of Hg⁰ into the surface area near the potential adsorption site. This effect becomes more pronounced when NH₃ is pre-adsorbed in the V-related region. For the NH₃_N-V + Hg-O_t configuration, the optimized Hg⁰-surface distance increases to 3.744 and 3.850 Å, the adsorption energy decreases to −11.50 kcal mol⁻¹, and the Mulliken charge on Hg⁰ decreases to 0.053 e. In the NH₃_N-V + Hg-O_b configuration, the adsorption energy is −11.84 kcal mol⁻¹, and the Mulliken charge on Hg⁰ is also small, at 0.056 e. Compared with the clean surface, the larger Hg⁰-surface distance and the lower charge transfer value indicate that the stability of Hg⁰ is weakened after NH₃ coordination, especially at the exposed V-related sites. Therefore, once NH₃ occupies the local environment around the V center, Hg⁰ is unable to interact as effectively with the adjacent V–O structure as it does on the clean surface.

Overall, NH₃ pre-adsorption weakens Hg⁰ adsorption through two related effects. The first is competitive site occupation, in which NH₃ partially blocks surface regions involved in Hg⁰ adsorption. The second is local electronic regulation, which becomes more pronounced when NH₃ coordinates directly to an exposed vanadium site. In this case, the reduced adsorption energies and smaller Hg charge transfer values reflect a less favorable local coordination and electronic environment for Hg⁰ stabilization. Therefore, NH₃ mainly suppresses Hg⁰ conversion at the initial adsorption stage by weakening the formation of stable Hg⁰ adsorption precursors. This finding provides a possible atomic-scale interpretation of the inhibitory effect of NH₃ on Hg⁰ oxidation previously observed over vanadium-based SCR catalysts [8,12].

Hg⁰ Oxidative Desorption Behavior

The subsequent oxidation and oxidative desorption of Hg on NH₃-precovered V₂O₅/TiO₂(001) surfaces were further examined by transition-state calculations. The corresponding reaction energy profiles are shown in Figure 9, and the optimized structures along the reaction pathways are presented in Figure 10. Three representative pathways were considered, namely NH₃–O_b + Hg–O_t, NH₃_N–V + Hg–O_t, and NH₃_N–V + Hg–O_b. Similar to the clean-surface routes, all three pathways are highly endothermic and require large activation barriers. Therefore, Hg⁰ oxidation on the NH₃-precovered surface remains energetically demanding within the present computational framework, while the overall oxygen-transfer/product-detachment mechanism remains similar to that on the clean surface.

Figure 9. Reaction energy profiles for Hg⁰ oxidation on NH₃-precovered V₂O₅/TiO₂(001).

Figure 10. Optimized reactant, transition-state, and product structures for Hg⁰ oxidation on NH₃-precovered V₂O₅/TiO₂(001).

For the NH₃-O_b + Hg-O_t pathway, Hg⁰ is initially located above the terminal oxygen region, while NH₃ is pre-adsorbed at the adjacent O_b site. During the reaction, Hg moves to the active oxygen, and the Hg-O distance decreases from 3.023 Å to about 2.500 Å at the transition state. In the final state, a HgO-like species with a Hg-O bond length of about 2.036 Å is formed, while NH₃ remains near the V-O unit. The reaction energy and activation energy are 136.587 and 143.064 kcal mol⁻¹, respectively, which are close to the values of the clean Hg-O_t pathway. This indicates that the NH₃ at the O_b site does not block the oxygen transfer from O_t to Hg, but only slightly changes the energy of the reaction. The NH₃_N-V + Hg-O_t pathway emphasizes the role of NH₃ coordination at the exposed V site. Here, NH₃ binds to the V center, and Hg⁰ begins near the O_t region. As the reaction proceeds, Hg approaches the terminal oxygen, and the Hg-O distance decreases from 3.744/3.850 Å to about 2.269 Å at the transition state. The final structure forms a HgO-like product with a Hg- O bond length of about 2.044 Å. The reaction energy and activation energy are 115.919 and 144.189 kcal mol⁻¹, respectively. The activation barrier is similar to that on the clean surface, but the reaction energy is lower. Following the formation and detachment of the HgO-like product, NH₃ shifts from the original position, indicating that the local V-related environment relaxes after the reaction. The NH₃_N-V + Hg-O_b pathway follows similar oxidation and desorption processes but is not as favorable as the O_t-related pathway. The Hg-O distance decreases from 3.178 Å to about 2.596 Å in the transition state. In the final state, the Hg-O bond length is about 2.035 Å. The reaction energy and activation energy are 150.883 and 167.627 kcal mol⁻¹, respectively. These values are lower than those of the clean Hg-O_b pathway, indicating that V-site NH₃ coordination can partially reduce the energy cost. However, the bridging oxygen pathway still exhibits a comparatively high calculated barrier because substantial distortion of the V–O–V framework is required.

Combining the transition state and adsorption analysis results, NH₃ primarily influences the Hg⁰ oxidation conversion during the initial adsorption stage. The pre-adsorbed NH₃ weakens the ability of Hg to be captured, as evidenced by the smaller negative adsorption energy and the reduced Hg charge transfer. Once Hg⁰ is already near the reactive oxygen site, the subsequent oxygen-transfer/product-detachment process is similar to that on the clean surface.

3.3. Effect of NO₂ on Hg⁰ Adsorption and Oxidation over V₂O₅/TiO₂(001)

3.3.1. NO₂ Adsorption on V₂O₅/TiO₂(001)

We studied the adsorption of NO₂ on V₂O₅/TiO₂(001) to identify its preferred adsorption configuration and its effect on the local V–O environment. As a key component in the FSCR atmosphere, NO₂ may not only participate in gas-phase reactions but also alter the redox and electronic properties of vanadium-based catalysts. Its effect depends strongly on its adsorption geometry, because different geometries disturb different V–O units and change the local distribution of reactive oxygen species. Three representative initial configurations were considered: NO₂-O_t, NO₂-O_b and NO₂-V, as shown in Figure 11. Their adsorption energies, characteristic distances and Mulliken charges are summarized in Table 5.

Figure 11. Optimized configurations of NO₂ adsorbed on V₂O₅/TiO₂(001).

Table 5. Adsorption energies, characteristic distances, and Mulliken charges for NO₂ adsorption on V₂O₅/TiO₂(001).

	Initial Sites	NO2-Ot	NO2-Ob	NO2-V
Parameters
Eads (kcal mol−1)		−2.969	−7.938	−10.028
D (Å)		1.485	2.932	2.035, 1.297
QNO2 (e)		0.126	0.081	0.085

The results of adsorption energy show that NO₂ is more inclined to adsorb in the exposed V-related region. Among the three configurations, NO₂-V has the most negative adsorption energy, at −10.028 kcal mol⁻¹, followed by NO₂-O_b, which has an adsorption energy of −7.938 kcal mol⁻¹. The stability of NO₂-O_t is poor, and its adsorption energy is only −2.969 kcal mol⁻¹. Optimizing the structure supports this trend. In NO₂-O_t, the characteristic distance is 1.485 Å, indicating that NO₂ is in close contact with the terminal oxygen region. In NO₂-O_b, NO₂ is located above the bridging oxygen region with a characteristic distance of 2.932 Å. In NO₂-V, two characteristic distances are observed: O(NO₂)-V_surface is 2.035 Å, and N(NO₂)-O_t_surface is 1.297 Å. This means that the relaxed NO₂-V structure interacts with the exposed V center and the adjacent terminal oxygen at the same time, which explains its higher stability.

Mulliken charge analysis shows that NO₂ adsorption involves limited charge transfer. The calculated charges of NO₂-O_t, NO₂-O_b and NO₂-V adsorbing NO₂ are 0.126 e, 0.081 e and 0.085 e, respectively. Although charge transfer occurs in all three cases, the magnitude is still very small. The most stable NO₂-V configuration does not exhibit the maximum charge transfer. This indicates that the adsorption strength is not solely controlled by net electron transfer. On the contrary, the local geometric and cooperative interactions between NO₂, the V center and nearby oxygen atoms play a major role in stabilizing NO₂ adsorption.

The PDOS results further clarify the electron source of NO₂ stability. As shown in Figure 12, the adsorption of NO₂ causes a significant change in the local electronic structure of the catalyst surface. In NO₂-O_b, the PDOS of O_b atoms changes significantly after adsorption, especially the peak intensity and energy distribution, which manifests as an electronic disturbance in the V-O-V bridging region. In NO₂-V, the V-related states change more strongly, especially the V d state, which indicates that there is a direct electronic interaction between NO₂ and the exposed V site. After the adsorption of NO₂-O_t and NO₂-V, the O_t-related states also change. This indicates that the terminal oxygen environment can be affected both by the direct contact of NO₂ and the adsorption of NO₂ near the V center. The atomic and orbital-resolved PDOS in Figure 13 identify the orbital coupling responsible for NO₂ stabilization. For NO₂–O_t, the N p states of NO₂ overlap with the p states of the nearby surface O_t atoms, demonstrating local electronic coupling between NO₂ and the terminal oxygen region. In NO₂–O_b, the N p states of NO₂ overlap with the p states of the O_b surface, showing a similar coupling. The strongest orbital interactions occur in the NO₂–V configuration. In this configuration, the N-related states of NO₂ couple with the adjacent O_t-related states, while the O p states of NO₂ show more pronounced overlap with the V d states on the catalyst surface. This synergistic N(NO₂)–O_t and O(NO₂)–V interaction explains the higher stability of NO₂–V.

Figure 12. PDOS comparison of selected surface atoms before and after NO₂ adsorption on V₂O₅/TiO₂(001).

Figure 13. Atom- and orbital-resolved PDOS of NO₂ adsorption systems on V₂O₅/TiO₂(001).

3.3.2. Effect of Pre-Adsorbed NO₂ on Hg⁰ Adsorption and Oxidation

Hg⁰ Adsorption Behavior

Four representative co-adsorption configurations were considered to evaluate the effect of pre-adsorbed NO₂ on Hg⁰ adsorption over V₂O₅/TiO₂(001), as shown in Figure 14. In the notation used here, NO₂–O_t + Hg–O_b denotes a structure in which the NO₂-precovered surface was first generated from the NO₂–O_t adsorption model, followed by Hg⁰ adsorption near the O_b region. The same convention is used for the other configurations. During structural relaxation, the model initially constructed from NO₂–O_t evolved into a configuration close to that obtained from the NO₂–V model. Therefore, the NO₂–O_t-derived precovered surface is used here to represent NO₂ occupation of the V-related adsorption region, whereas the NO₂–O_b model represents NO₂ pre-adsorption near an oxygen-related region. Hg⁰ was initially placed either near a surface oxygen site, including O_t or O_b, or above the O atom of pre-adsorbed NO₂, in order to compare whether Hg⁰ preferentially interacts with the catalyst surface or with the adsorbed NO₂ moiety. The corresponding adsorption energies, characteristic distances, and Mulliken charges are listed in Table 6.

Figure 14. Optimized configurations of Hg⁰ adsorbed on NO₂-precovered V₂O₅/TiO₂(001).

Table 6. Adsorption energies, characteristic distances, and Mulliken charges for Hg⁰ adsorption on NO₂-precovered V₂O₅/TiO₂(001).

	Initial Sites	NO2-Ot + Hg-Ob	NO2-Ot + Hg-NO2_O	NO2-Ob + Hg-Ot	NO2-Ob + Hg-NO2_O
Parameters
Eads (kcal mol−1)		−19.54	−18.09	−14.52	−12.27
D (Å)		3.588, 3.449, 3.757	3.749, 3.659	2.997, 3.063, 3.401	3.019, 3.030
QHg (e)		0.065	0.046	0.240	0.132

Compared with the clean surface, pre-adsorbed NO₂ generally strengthens Hg⁰ adsorption. On pristine V₂O₅/TiO₂(001), the adsorption energies of Hg–O_t and Hg–O_b are −12.31 and −12.58 kcal mol⁻¹, respectively. After NO₂ pre-adsorption, the adsorption energies decrease to −19.54 kcal mol⁻¹ for NO₂–O_t + Hg–O_b, −18.09 kcal mol⁻¹ for NO₂–O_t + Hg–NO₂_O, −14.52 kcal mol⁻¹ for NO₂–O_b + Hg–O_t, and −12.27 kcal mol⁻¹ for NO₂–O_b + Hg–NO₂_O. Among these configurations, NO₂–O_t + Hg–O_b is the most stable, followed by NO₂–O_t + Hg–NO₂_O. Both are substantially more favorable than Hg⁰ adsorption on the clean surface. By contrast, the two NO₂–O_b-based configurations show weaker stabilization, especially NO₂–O_b + Hg–NO₂_O, whose adsorption energy is close to those of the clean-surface cases. The optimized structures support this energy trend. For NO₂–O_t + Hg–O_b, three characteristic distances are obtained: 3.588 Å, 3.449 Å, and 3.757 Å, indicating that Hg⁰ is stabilized around the NO₂ pre-adsorbed vanadium–oxygen unit through multiple weak interactions. Similar multi-contact modes appear in NO₂–O_t + Hg–NO₂_O, with distances of 3.749 and 3.659 Å, respectively. Mercury is initially located above the O atom of NO₂. Although the adsorption strength is slightly weaker than that of NO₂–O_t + Hg–O_b, the configuration remains stabilized by interactions with the NO₂-derived local environment. Nevertheless, the adsorption energy is much more negative than on a clean surface. This indicates that the O atom in the adsorbed NO₂ can be used as an additional interaction site for Hg⁰. This behavior differs from that of NH₃: NH₃ mainly blocks surface adsorption sites, whereas NO₂ can introduce additional oxygen-containing interaction sites and contribute to the stabilization of adsorbed Hg⁰. The surface based on NO₂–O_b exhibits different behavior depending on the position of Hg⁰. In NO₂–O_b + Hg–O_t, the adsorption energy is −14.52 kcal mol⁻¹, with distances of 2.997, 3.063, and 3.401 Å. In NO₂–O_b + Hg–NO₂_O, the adsorption energy is −12.27 kcal mol⁻¹, with distances of 3.019 and 3.030 Å. This is the weakest case in the NO₂-precovered model. The optimized structure in Figure 14d shows that NO₂ does not remain unchanged. Before Hg⁰ adsorption, NO₂ is located above the O_b region. After the introduction of Hg⁰, NO₂ moves to the O_t′ region. This indicates that the adsorbed NO₂ species can relocate during structural relaxation when a more favorable co-adsorption configuration is formed. The Mulliken charge results show that the adsorption of Hg⁰ is strongly dependent on the configuration. On the clean surface, the charge of Hg-O_t and Hg-O_b to Hg⁰ is 0.113 e and 0.115 e, respectively. After NO₂ pre-adsorption, the corresponding Hg⁰ charges are 0.065 e for NO₂–O_t + Hg–O_b, 0.046 e for NO₂–O_t + Hg–NO₂_O, 0.240 e for NO₂–O_b + Hg–O_t, and 0.132 e for NO₂–O_b + Hg–NO₂_O. Although the adsorption energy of the NO₂–O_t-based configuration is more negative, its Hg⁰ charge transfer value is relatively small, indicating that the enhancement of its stability mainly comes from the favorable multi-center local environment, rather than large net electron transfer. In contrast, NO₂–O_b + Hg-O_t combines stronger adsorption and a larger Hg⁰ charge, which indicates that the electron perturbation induced by Hg⁰ adsorption is more obvious.

Overall, pre-adsorbed NO₂ strengthens Hg⁰ adsorption in several of the configurations examined, particularly when the V-related region is involved in the co-adsorption structure. Unlike NH₃, which weakens Hg⁰ adsorption through site occupation and local electronic regulation, NO₂ not only modifies the original V-O environment but also provides an additional O-containing interaction site for Hg⁰.

Hg⁰ Oxidative Desorption Behavior

The oxidation and oxidative desorption of Hg⁰ on NO₂-precovered V₂O₅/TiO₂(001) surfaces were examined by transition-state calculations. The optimized structures along the reaction pathways and the corresponding energy profiles are shown in Figure 15 and Figure 16, respectively. As discussed above, pre-adsorbed NO₂ can occupy the V-related region or remain near an oxygen-related region while still providing an O-containing moiety available for interaction with Hg⁰. Accordingly, two types of oxidation routes were compared: direct O abstraction from adsorbed NO₂ and oxidation by lattice oxygen from the catalyst surface. The results show that NO₂ pre-adsorption does not simply perturb the original clean-surface pathways. Instead, it introduces NO₂-assisted oxygen-transfer routes with substantially lower calculated activation barriers than the lattice-oxygen-mediated pathways.

Figure 15. Optimized reactant, transition-state, and product structures for Hg⁰ oxidation on NO₂-precovered V₂O₅/TiO₂(001).

Figure 16. Reaction energy profiles for Hg⁰ oxidation on NO₂-precovered V₂O₅/TiO₂(001).

For the NO₂–V-precovered surface, four representative pathways were considered. In the NO₂–V + Hg–NO₂_O1 and NO₂–V + Hg–NO₂_O2 pathways, Hg⁰ reacts directly with an O atom of the adsorbed NO₂ species to form an HgO-like product. The corresponding reaction and activation energies are 84.943 and 108.545 kcal mol⁻¹, respectively. By contrast, when Hg⁰ reacts with lattice oxygen on the same NO₂–V precovered surface, the energetic cost increases markedly. For the NO₂–V + Hg–O_t′ pathway, the reaction and activation energies are 139.116 and 155.786 kcal mol⁻¹, respectively, while for the NO₂–V + Hg–O_b pathway, they rise further to 160.037 and 180.331 kcal mol⁻¹. These results show that oxygen transfer from adsorbed NO₂ has a lower calculated activation barrier than lattice-oxygen extraction within the same computational framework, with a barrier reduction of approximately 47–74 kcal mol⁻¹. A similar pattern appears on the NO₂–O_t′-precovered surface. Three pathways were examined, which involve oxygen transfer from NO₂, from a nearby O_t atom, and from an O_b atom. The NO₂–O_t′ + Hg–NO₂_O pathway has the lowest calculated activation barrier among these three pathways, with a reaction energy of 73.672 kcal mol⁻¹ and an activation barrier of 74.581 kcal mol⁻¹. The lattice-oxygen pathways require much higher energies. NO₂–O_t′ + Hg–O_t has reaction and activation energies of 138.676 and 149.585 kcal mol⁻¹, and NO₂–O_t′ + Hg–O_b requires 164.750 and 186.384 kcal mol⁻¹. The difference in activation energy further indicates that oxygen transfer from adsorbed NO₂ is energetically preferred over direct extraction of lattice oxygen among the pathways considered. For the NO₂–O_b + Hg–NO₂_O pathway, the O atom of the pre-adsorbed NO₂ species directly serves as the oxygen donor for Hg⁰ oxidation. The reaction energy is 71.346 kcal mol⁻¹, and the activation energy is 81.767 kcal mol⁻¹, both of which are considerably lower than those associated with lattice-oxygen-mediated oxidation on either the clean or NO₂-precovered surfaces. During the oxygen-transfer step, Hg⁰ approaches one O atom of the adsorbed NO₂ species. In the optimized product state, this O atom is transferred to Hg, forming a detached HgO-like moiety, whereas the N atom and the remaining O atom of the original NO₂ remain associated with the catalyst surface as an NO-like residue (denoted as NO *). Accordingly, the elementary step can be represented schematically as Hg⁰ + NO₂ * → HgO (detached) + NO *. Here, NO * denotes the immediate surface-associated residue identified in the optimized product structure; its independent adsorption stability and subsequent desorption or further reaction were not evaluated in the present study. The present pathway therefore describes the NO₂-to-Hg oxygen-transfer step rather than the complete subsequent NOx reaction cycle. In this work, “oxidative desorption” refers to the coupled oxygen-transfer and product-detachment process observed along the calculated pathway; a separate HgO desorption barrier was not calculated. This structural evolution shows that the O atom of adsorbed NO₂ can directly participate in the oxygen-transfer step, regardless of whether NO₂ is initially located near a V-related site or an oxygen-related site.

A comparison with the clean surface further highlights the relative energetic effect of NO₂ pre-adsorption. On clean V₂O₅/TiO₂(001), Hg⁰ oxidation proceeds only through lattice-oxygen abstraction and requires high barriers of 146.817, 177.851, and 184.529 kcal mol⁻¹ for the O_t, O_b, and interfacial pathways, respectively. After NO₂ pre-adsorption, the NO₂-assisted pathways reduce the barriers to 74.581–108.545 kcal mol⁻¹, whereas lattice-oxygen oxidation on the same NO₂-precovered surfaces retains substantially higher calculated barriers (Table 7). Therefore, NO₂ pre-adsorption does not simply modify the energetics of the original lattice-oxygen pathway; it also introduces an alternative oxygen-transfer channel in which the adsorbed NO₂ species directly participates. This alternative oxygen-transfer pathway provides a possible mechanistic interpretation of the NO₂-associated promotion of Hg⁰ oxidation reported experimentally under fast-SCR conditions [7].

Table 7. Reaction energies and activation barriers for Hg⁰ oxidation on NO₂-precovered V₂O₅/TiO₂(001).

Energy (kcal mol−1)	NO2-V + Hg-NO2_O1	NO2-V + Hg-Ob	NO2-V + Hg-Ot′	NO2-V + Hg-NO2_O2	NO2-Ot′ + Hg-NO2_O	NO2-Ot′ + Hg-Ot	NO2-Ot′ + Hg-Ob	NO2-Ob + Hg-NO2_O
Reaction energy	84.943	160.037	139.116	82.416	73.672	138.676	164.75	71.346
Reaction barrier	108.545	180.331	155.786	106.588	74.581	149.585	186.384	81.767

Table 7. Reaction energies and activation barriers for Hg⁰ oxidation on NO₂-precovered V₂O₅/TiO₂(001).

Energy (kcal mol−1)	NO2-V + Hg-NO2_O1	NO2-V + Hg-Ob	NO2-V + Hg-Ot′	NO2-V + Hg-NO2_O2	NO2-Ot′ + Hg-NO2_O	NO2-Ot′ + Hg-Ot	NO2-Ot′ + Hg-Ob	NO2-Ob + Hg-NO2_O
Reaction energy	84.943	160.037	139.116	82.416	73.672	138.676	164.75	71.346
Reaction barrier	108.545	180.331	155.786	106.588	74.581	149.585	186.384	81.767

3.4. Effect of Co-Pre-Adsorbed NH₃ and NO₂ on Hg⁰ Oxidation over V₂O₅/TiO₂(001)

3.4.1. Hg⁰ Adsorption Behavior

Hg⁰ adsorption was examined on NH₃/NO₂ co-precovered V₂O₅/TiO₂(001) surfaces to clarify how the coexistence of the two adsorbates affects the initial capture of Hg⁰. As shown in Figure 17, two types of co-precovered surfaces were considered. In the first type, NH₃ and NO₂ occupy oxygen-related adsorption regions that could otherwise serve as possible Hg⁰ adsorption sites, represented by NO₂(O_t) + NH₃(O_b) + Hg and NO₂(O_b) + NH₃(O_t) + Hg. In the second type, both adsorbates are associated with the V-related region, represented by NO₂(VO_t) + NH₃(V) + Hg. This configuration was used to examine the effect of simultaneous modification of the V-centered local environment on Hg⁰ adsorption. In all models, Hg⁰ was initially placed above the NO₂-related region or a neighboring surface region before full structural optimization. The corresponding adsorption energies, representative interatomic distances, and Mulliken charges are listed in Table 8.

Figure 17. Optimized configurations of Hg⁰ adsorbed on NH₃/NO₂-co-precovered V₂O₅/TiO₂(001).

Table 8. Adsorption energies, characteristic distances, and Mulliken charges for Hg⁰ adsorption on NH₃/NO₂-co-precovered V₂O₅/TiO₂(001).

	Initial Sites	NO2(Ot) + NH3(Ob) + Hg	NO2(Ob) + NH3(Ot) + Hg	NO2(VOt) + NH3(V) + Hg
Parameters
Eads (kcal mol−1)		−12.606	−12.723	−15.469
D (Å)		3.01, 2.886	2.924, 2.937	3.683, 3.318
QHg (e)		0.145	0.136	0.037

The adsorption energy results show that the coexistence of NO₂ and NH₃ does not lead to the strong inhibition of Hg⁰ adsorption observed on NH₃-precovered surfaces. On clean V₂O₅/TiO₂(001), Hg–O_t and Hg–O_b have adsorption energies of −12.31 and −12.58 kcal mol⁻¹, respectively. After NH₃ pre-adsorption, Hg⁰ adsorption becomes weaker, with adsorption energies of −11.04 to −11.84 kcal mol⁻¹. In contrast, on the NH₃/NO₂ co-precovered surface, the adsorption energies are −12.606, −12.723, and −15.469 kcal mol⁻¹, corresponding to NO₂(O_t) + NH₃(O_b) + Hg⁰, NO₂(O_b) + NH₃(O_t) + Hg⁰, and NO₂(VO_t) + NH₃(V) + Hg⁰, respectively. The first two adsorption energies are comparable to those on the clean surface, or even slightly more negative, whereas NO₂(VO_t) + NH₃(V) + Hg⁰ shows significantly stronger stability. This indicates that NO₂ can partially counteract the inhibitory effect of NH₃ on Hg⁰ adsorption. The optimized geometries provide further evidence for this role of NO₂. For NO₂(O_t) + NH₃(O_b) + Hg⁰, the two Hg–O distances of 3.010 and 2.886 Å correspond to interactions between Hg⁰ and the two O atoms of pre-adsorbed NO₂. A similar feature is found for NO₂(O_b) + NH₃(O_t) + Hg⁰, where the optimized Hg–O distances of 2.924 and 2.937 Å also involve the O atoms of the adsorbed NO₂ species. In both configurations, Hg⁰ remains close to the NO₂-derived O-containing region rather than being displaced away from the surface. Thus, once NO₂ is present, its O atoms can act as accessible local interaction sites for Hg⁰ adsorption. For NO₂(VO_t) + NH₃(V) + Hg⁰, the optimized distances of 3.683 and 3.318 Å correspond to the separations between Hg⁰ and the O atom of pre-adsorbed NO₂, and between Hg⁰ and a neighboring surface O atom in the V-centered region, respectively. Unlike the oxygen-region co-precovered models, Hg⁰ in this configuration is not stabilized by a single isolated site. In contrast, adsorption is driven by a combined local environment that involves the adsorbed NO₂ species and the nearby surface oxygen. Therefore, the strong adsorption energy of −15.469 kcal mol⁻¹ can be attributed to the synergistic multi-center adsorption environment formed around the V-related region.

The Mulliken charge analysis further highlights the configuration dependence of Hg⁰ stability. The calculated Hg⁰ charges are 0.145 e, 0.136 e, and 0.037 e, corresponding to NO₂(O_t) + NH₃(O_b) + Hg⁰, NO₂(O_b) + NH₃(O_t) + Hg⁰, and NO₂(VO_t) + NH₃(V) + Hg⁰, respectively. The first two values are higher than those obtained on the NH₃ pre-adsorbed surface and slightly higher than those obtained on the clean surface, indicating that when NH₃ and NO₂ occupy the adjacent oxygen-related regions, Hg⁰ will still undergo significant electronic perturbation due to adsorption. In contrast, NO₂(VO_t) + NH₃(V) + Hg⁰ has the smallest Hg⁰ charge but the most favorable adsorption energy.

In general, the role of NO₂ in the NH₃/NO₂ co-precovered system is significantly different from that of NH₃ alone. In the case of NH₃ alone, the pre-adsorbed NH₃ mainly occupies the catalyst-related adsorption region, which reduces the accessibility of Hg adsorption sites. However, in the co-precovered system, the O atom of adsorbed NO₂ provides additional interaction sites for Hg⁰. Therefore, although NH₃ still occupies part of the surface adsorption environment, the presence of NO₂ introduces an oxygen-containing local environment that can still be in contact with Hg, which helps to maintain or even enhance the initial adsorption of Hg⁰ on V₂O₅/TiO₂(001).

3.4.2. Hg⁰ Oxidative Desorption Behavior

Three representative oxidation pathways were calculated on the NH₃/NO₂ co-covered V₂O₅/TiO₂(001) surface to evaluate how the coexistence of NH₃/NO₂ affects the subsequent Hg⁰ conversion. The optimized structures along the reaction coordinates are shown in Figure 18, and the corresponding reaction energy profiles are presented in Figure 19. In the first two configurations, NH₃ and NO₂ occupy neighboring oxygen-related regions, represented by NO₂(O_t) + NH₃(O_b) + Hg⁰ and NO₂(O_b) + NH₃(O_t) + Hg⁰. In the third configuration, both adsorbates are associated with the V-centered local environment, represented by NO₂(VO_t) + NH₃(V) + Hg⁰. For these surfaces, two types of oxidation routes were examined: oxygen transfer from the adsorbed NO₂ species and oxidation by lattice oxygen from the catalyst surface. The calculated results show that, even under NH₃/NO₂ co-precovered conditions, oxidation through pre-adsorbed NO₂ exhibits a lower calculated activation barrier than lattice-oxygen abstraction within the present computational framework.

Figure 18. Optimized reactant, transition-state, and product structures for Hg⁰ oxidation on NH₃/NO₂-co-precovered V₂O₅/TiO₂(001).

Figure 19. Reaction energy profiles for Hg⁰ oxidation on NH₃/NO₂-co-precovered V₂O₅/TiO₂(001).

For NO₂(O_t) + NH₃(O_b) + Hg and NO₂(O_b) + NH₃(O_t) + Hg, Hg⁰ is oxidized by an O atom of the adsorbed NO₂ species, as shown in Figure 18a,b. The corresponding reaction energies and activation barriers are 71.930/81.371 kcal mol⁻¹ and 72.727/83.208 kcal mol⁻¹, respectively. The two pathways are close in energy, showing that when NH₃ and NO₂ occupy neighboring oxygen-related regions, the initial location of NO₂ near O_t or O_b has only a limited influence on the overall oxidation energetics. In both cases, Hg⁰ approaches the O atom of pre-adsorbed NO₂, followed by oxygen transfer and formation of a detached HgO-like product, while the N atom and the remaining O atom of the original NO₂ remain associated with the surface as an NO-like residue (NO *). The subsequent fate of this NO * species was not considered in the present elementary-step analysis. These barriers remain close to those of the most favorable NO₂-only precovered routes, which means that NH₃ does not significantly suppress the NO₂-assisted oxidation channel when it is located away from the V-centered adsorption environment. A different behavior is observed for the NO₂(VO_t) + NH₃(V) + Hg co-precovered surface, where both adsorbates modify the V-centered local environment. Two oxidation pathways were examined, as shown in Figure 18(c1,c2). In NO₂(VO_t) + NH₃(V) + Hg_NO₂_O, Hg⁰ is oxidized by the O atom of pre-adsorbed NO₂, giving a reaction energy of 79.558 kcal mol⁻¹ and an activation barrier of 113.326 kcal mol⁻¹. In contrast, when Hg⁰ reacts with the lattice O_b atom in NO₂(VO_t) + NH₃(V) + Hg_Ob, the reaction energy and activation barrier increase sharply to 149.186 and 170.893 kcal mol⁻¹, respectively. Thus, even when NH₃ and NO₂ are both concentrated around the V-centered region, oxygen transfer from adsorbed NO₂ remains energetically preferred over the corresponding lattice-oxygen pathway. The barrier difference between the NO₂-assisted and lattice-oxygen routes exceeds 57 kcal mol⁻¹, indicating a relative energetic preference for oxygen transfer from pre-adsorbed NO₂ over the lattice-oxygen pathway.

Compared to the clean surface and NH₃ precovered surface, the NO₂-assisted pathway under NH₃/NO₂ co-coverage maintains a lower energy barrier. In particular, the two oxygen-region co-precovered pathways in Figure 18a,b exhibit calculated barriers of approximately 81–83 kcal mol⁻¹, which are lower than the corresponding lattice-oxygen-mediated barriers within the present computational framework. However, when NH₃ and NO₂ simultaneously occupy the V-centered local environment, the barrier of the NO₂-assisted pathway rises to 113.326 kcal mol⁻¹. This value is higher than those of the most favorable NO₂-only pathways, indicating that NH₃ increases the energetic cost of the NO₂-assisted pathway when both adsorbates modify the same local V–O environment.

Overall, the co-precovered systems show a competitive but asymmetric effect between NH₃ and NO₂. NH₃ increases the calculated barrier of the NO₂-assisted pathway when both adsorbates perturb the V-centered region, but NO₂-derived oxygen transfer remains energetically preferred over the corresponding lattice-oxygen pathway. Among the pathways examined, the lower-energy oxidation route still involves the O atom of pre-adsorbed NO₂ rather than lattice oxygen. Therefore, within the present computational model, NO₂-derived oxygen provides the relatively lower-energy oxidation route, whereas NH₃ mainly modifies the local adsorption environment and can increase the calculated barrier when both species perturb the same V-centered region. These configuration-dependent effects are qualitatively consistent with previous experimental observations of NH₃ inhibition and NO₂-associated promotion of Hg⁰ oxidation [7,12].

4. Conclusions

In this work, DFT calculations were performed using a V₂O₅/TiO₂(001) model to clarify the effects of representative NH₃- and NO₂-derived surface environments on Hg⁰ adsorption and oxidation relevant to FSCR conditions. On the clean surface, Hg⁰ adsorption was generally weak. Although the O_b′ site showed the strongest adsorption energy of −21.22 kcal mol⁻¹, it did not provide the most favorable oxidation route. Oxidation by lattice oxygen remained energetically demanding, with calculated barriers of 146.817, 177.851, and 184.529 kcal mol⁻¹ for the O_t, O_b, and O_b′ pathways, respectively. NH₃ preferentially adsorbs at exposed V-related sites through N coordination, particularly in the V–O_t local environment. Its pre-adsorption weakens Hg⁰ capture, as reflected by the less negative adsorption energies, whereas the subsequent lattice-oxygen-mediated oxidation pathways remain energetically similar to those on the clean surface. Therefore, NH₃ mainly affects Hg⁰ conversion at the initial adsorption stage by modifying or occupying the local adsorption environment. NO₂ pre-adsorption produces a different energetic response. In addition to modifying Hg⁰ adsorption, pre-adsorbed NO₂ introduces alternative oxygen-transfer pathways with calculated activation barriers of 74.581–108.545 kcal mol⁻¹, substantially lower than those of direct lattice-oxygen abstraction within the same computational framework. Under NH₃/NO₂ co-precoverage, oxygen transfer from NO₂ remains the relatively lower-energy route among the pathways examined. When NH₃ and NO₂ occupy neighboring oxygen-related regions, the calculated barriers are approximately 81–83 kcal mol⁻¹, whereas simultaneous perturbation of the V-centered local environment increases the NO₂-assisted barrier to 113.326 kcal mol⁻¹. These results indicate that NH₃ can modify the relative energetics of the NO₂-assisted pathway. Overall, the present calculations indicate that NH₃ and NO₂ exert different effects on Hg⁰ conversion on V₂O₅/TiO₂(001): NH₃ mainly weakens the formation of favorable Hg⁰ adsorption precursors, whereas pre-adsorbed NO₂ introduces an alternative oxygen-transfer channel that is relatively lower in energy than direct lattice-oxygen oxidation among the pathways examined in the present model. It should be noted that, since these calculations are based on a simplified surface model without thermal and entropic corrections, the results should be interpreted as relative energetic trends rather than quantitative predictions of reaction kinetics under practical FSCR conditions.