Comparing the Photocatalytic Oxidation Efficiencies of Elemental Mercury Using Metal-Oxide-Modified Titanium Dioxide under the Irradiation of Ultra-Violet Light

Abstract

Since the signing of the Minamata Convention in 2013, attempts have been primarily focused on reducing the emission of elemental mercury (Hg⁰) from coal-fired power plants (CFPPs). The most cost-effective measure for controlling the emission of mercury involves oxidizing Hg⁰ to mercury oxides, which are then removed using wet flue gas desulfurization (WFGD). Thus, novel photocatalysts with the best properties of photocatalytic ability and thermal stability need to be developed urgently. In this study, titanium dioxide (TiO₂)-based photocatalysts were synthesized through the modification of three metal oxides: CuO, CeO₂, and Bi₂O₃. All the photocatalysts were further characterized using X-ray diffraction, X-ray photoelectron spectroscopy, photoluminescence, and ultraviolet-visible spectrometry. The photocatalytic oxidation efficiencies of Hg⁰ were evaluated under an atmosphere of N₂ + Hg⁰ at 100–200 °C. The photocatalytic reactions were simulated by kinetic modeling using the Langmuir–Hinshelwood (L–H) mechanism. The results showed that Bi₂O₃/TiO₂ exhibited the best thermal stability, with the best oxidation efficiency at 200 °C and almost the same performance at 100 °C. L–H kinetic modeling indicated that photocatalytic oxidation reactions for the tested photocatalysts were predominantly physical adsorption. Additionally, the activation energy (E_a), taking into account Arrhenius Law, decreased dramatically after modification with metal oxides.

1. Introduction

According to the United Nations Environment Programme, the estimated mercury emissions from coal-fired power plants (CFPPs) exceed 30% of total anthropogenic mercury emissions [1]. Mercury (Hg) is recognized as a metallic pollutant with detrimental effects on human beings and ecosystems due to its biochemical properties of bioaccumulation and biomagnification [2]. After the signing of the Minamata Convention in 2013, the control of mercury emissions from CFPPs was treated as the primary target for reducing elemental mercury (Hg⁰) emission into the atmosphere. Currently, constructed CFPPs have to adopt the best available control technologies and the best environmental practices according to the Minamata Convention on Mercury [3].

The typical mercury species normally consists of three basic forms: elementary, oxidized, and particulate. Among these, Hg⁰ accounts for the largest proportion [4]. Particulate mercury can be removed by particle collectors such as the electrostatic precipitator (ESP) and fabric filter. Oxidized mercury can be either adsorbed on a fly ash surface and then removed by particle collectors, or dissolved in an absorbent and removed using wet flue gas desulfurization (WFGD) [5]. Hg⁰ is highly volatile in ambient air and insoluble in water. It can be adsorbed by activated carbon (AC) or carbon black or further oxidized to mercury oxides and then dissolved via WFGD [6]. Using adsorbents such as AC requires the installation of extra air pollution control devices, which require high maintenance and involve high costs [7].

As a result, previous studies proposed an economical method for reducing mercury emissions by adding oxidative catalysts in selective catalytic reduction (SCR) to progress the potential catalytic oxidation of Hg⁰ [8]. However, there is an urgent need to resolve several obstacles in the catalytic oxidation process in SCR. The optimal temperature for existing SCR catalysts is 300–400 °C, which, unfortunately, is a fatal defect in the photocatalytic oxidation of Hg⁰ as its optimal operating temperature is below 100 °C [9]. Moreover, to avoid installing extra heating devices, SCR devices are commonly installed upstream of particle collectors (i.e., ESP), which mostly contain high-concentration particulate matter causing a masking effect over the surface of catalysts and reducing their catalytic activity and lifetime [10]. Therefore, it is crucial to develop a low-temperature SCR catalyst operating at 100–200 °C to achieve the goal of moving the SCR behind the particle collector [11].

TiO₂ is the most commonly used photocatalyst due to its advantages of non-toxicity, high oxidation activity, good chemical and thermal stabilities, and low cost. However, the operating temperature of pristine TiO₂ is too low for direct use in industrial applications. Thus, numerous methods have been developed to modify TiO₂ by employing metal oxide [12], graphene [13], metal–organic frameworks [14], and zeolite [15]. Among these, the modification of metal oxide is the simplest method for effectively enhancing TiO₂ activity by retarding the recombination of photo-induced electron/hole pairs, decreasing the energy bandgap, and/or increasing light absorptivity. The metal modification can be classified as either single metal atom doping with Au, Ag, and Pt or metal oxide modification with CuO₂, MnO₂, Fe₂O₃, CeO₂, and Bi₂O₃ [16,17]. In particular, the radius of Cu²⁺ is 0.073 nm, which is close to that of Ti⁴⁺ (0.068 nm). Thus, Cu might enter the TiO₂ molecular structure, which is beneficial for separating the photo-induced electron/hole pairs [18]. Wu et al., (2015) reported that TiO₂ did not exhibit any photocatalytic activity on Hg⁰ under visible light. The modification of 1.25 weight % CuO had 57.8% photocatalytic oxidation efficiency of Hg⁰ under sunlight and 85% under UV light [19]. Another potential modifying material is CeO₂; the oxygen vacancies in CeO₂ can form active sites and capture either Hg⁰ or oxygen atoms from the gas steam. Additionally, CeO₂/TiO₂ shows strong thermal stability at high temperatures of 160–250 °C. This is mainly due to the conversion of the valence states of Ce(II)O₂ to Ce(III)₂O₃ (2 CeO₂ + Hg⁰ → Ce₂O₃ + HgO). The consumed oxygen can be recovered by Ce³⁺, which captures O₂ molecules from the gas stream, accelerating the oxidation of Hg⁰ [20]. Li et al. (2011) reported that 1.5% CeO₂/TiO₂ could reach 90% at 250 °C related to the weakly bonded oxygen and chemisorbed oxygen on Ce³⁺ [21]. TiO₂ responds solely to UV_A, thus its overall light absorptivity is relatively low because it does not respond to visible light. Therefore, adding visible-light-responsive additives to TiO₂ might potentially enhance the catalytic oxidation of Hg⁰. With its low energy band gap (2.85 eV), Bi₂O₃ is one of the potential materials for increasing the light adsorption ability [22].

In this study, we chose three metal oxides (CuO, CeO₂, and Bi₂O₃) to synthesize TiO₂-based photocatalysts with the aim of enhancing the photocatalytic oxidation of Hg⁰ at higher temperatures. This study analyzed the characteristics of self-prepared photocatalysts and conducted photocatalytic oxidation experiments to evaluate the photocatalytic oxidation efficiencies of Hg⁰ in an N₂ + O₂ atmosphere. Furthermore, the pros and cons of three metal-oxide-modified TiO₂ photocatalysts were also evaluated.

2. Results and Discussion

2.1. Characterization of Photocatalysts

Figure 1 illustrates the nitrogen adsorption–desorption isotherms of the photocatalysts. It shows that there is no overlap between the adsorption and desorption curves of all photocatalysts presenting type IV isotherm species according to the International Union of Pure and Applied Chemistry isothermal classification. The type IV isotherm, with a capillary effect resulting in the hysteresis loop, is usually observed in mesoporous materials [23]. A more detailed examination of the hysteresis loop revealed that TiO₂, CeO₂/TiO₂, and CuO/TiO₂ had a type H2 loop that contributed to the complex pore structure, with the pores in the shape of ink bottles [24]. The desorption curve of CuO/TiO₂ showed a steep desorption branch at a lower relative pressure region, indicating a smaller pore diameter of CuO/TiO₂. Further, Bi₂O₃/TiO₂ showed a H3-type hysteresis loop with a lower limit of the desorption branch located at the cavitation-induced P/P₀. This phenomenon showed that the average pore diameter of Bi₂O₃/TiO₂ was larger than those of TiO₂, CeO₂/TiO₂, and CuO/TiO₂.

Table 1. Specific surface areas and average pore diameters of photocatalysts.

Type of Photocatalyst	Specific Surface Area (m2/g)	Average Pore Diameter (nm)
TiO2	85.29	9.91
CuO/TiO2	48.16	9.65
CeO2/TiO2	50.82	10.15
Bi2O3/TiO2	62.05	16.92

summarizes the specific surface areas (SSAs) and average pore diameters of the photocatalysts. The SSAs of the photocatalysts were in the order TiO₂ > Bi₂O₃/TiO₂ > CeO₂/TiO₂ > CuO/TiO₂, while the average pore diameters were in the order Bi₂O₃/TiO₂ > CeO₂/TiO₂ > TiO₂ > CuO/TiO₂. Because Bi₂O₃/TiO₂ had the highest SSA among the three metal-oxide-modified photocatalysts, it could have more active sites over the surface of inner pores to adsorb the reactant (Hg⁰). Additionally, it had a larger pore diameter, which could more easily move the desorbed product (HgO) out of the photocatalyst through its inner pore passages. As a result, Bi₂O₃ had better photocatalytic oxidation efficiency for Hg⁰ compared to the other two photocatalysts.

The morphologies of the prepared TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ were observed using FE-SEM as illustrated in Figure 2. The results of FE-SEM show that TiO₂ and metal-oxide-modified TiO₂ presented as nanoparticle agglomerations due to calcination during the preparation process. The EDS analysis depicts the atomic partition of each atom. The results show that the actual stoichiometry was not with the design in the synthesis process. This could be attributed to the limitation of EDS, in which the detection area was relatively small [25]. Figure 3 presents the mapping analysis of the photocatalysts, which shows that various atoms were well distributed. Figure 4 presents the TEM analysis of the photocatalysts, showing the overlap between nanoparticles, creating a darker region. The crystal size of nanoparticles could be further estimated. As shown in Figure 4, the diameters of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ were 15.5, 16.98, 20.15, and 21.27 nm, respectively.

Figure 5 presents the XRD patterns of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂. TiO₂ had several characteristic peaks at 2θ = 25.39°, 3.7.84°, 48.09°, 53.98°, 62.79°, 70.34°, and 75.10° that were consistent with the Anatase TiO₂ from the American Mineralogist Crystal Structure Database [26]. The crystallite sizes of the photocatalysts were calculated using the Scherrer equation [27].

Table 2. Crystallite sizes calculated using the Debye Scherrer equation.

Photocatalysts	TiO2	CuO/TiO2	CeO2/TiO2	Bi2O3/TiO2
Crystallite size (nm)	11.8	13.5	15.01	14.2

presents the results of crystallite sizes in the order CeO₂/TiO₂ > Bi₂O₃/TiO₂ > CuO/TiO₂ > TiO₂. Additionally, Figure 5 clearly shows the relatively low or undetected characteristic peaks of CuO, CeO₂, and Bi₂O₃. Particularly, the characteristic peaks of CuO were not observed, while those of CeO₂ and Bi₂O₃ had relatively low intensity because the amounts of these metal oxides were too low to form crystals, with most of the photocatalysts in the amorphous state [28]. This might be attributed to the non-uniformly dispersed metal oxides over the Anatase TiO₂. Additionally, the atomic radius of Cu (ϕ = 0.14 nm) was smaller than that of Ti (ϕ = 0.187 nm), potentially causing two adverse effects [29]. First, excess Cu could enter the TiO₂ lattice, thus decreasing the crystallinity of TiO₂ and reducing the intensity of (101) facets of TiO₂. Second, CuO could easily block the inner pores of TiO₂, reducing the adsorptive capacity; this was consistent with the analytical BET results. Furthermore, overwhelmed metal oxides can accumulate over the surface of TiO₂, resulting in a decrease in light absorptivity and photocatalytic activity [19].

The photocatalysts absorbed incident light energy to excite electrons from the valence band to the conduction band. This left a hole in the valence band and an electron in the conduction band, thus creating electron/hole pairs. Photo-induced electrons and holes can further react with O₂ and H₂O to produce reactive radicals O₂⁻ and ·OH, respectively, with high redox activity [30,31]. Therefore, the lifetime of the photo-induced electron/hole pair was the most important parameter for assessing the photocatalytic activity of the photocatalysts. Figure 6 presents the PL patterns of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂. It demonstrates that TiO₂ has the highest PL intensity, which was mainly attributed to the rapid recombination of electron/hole pairs, meaning that electrons could not be transited from Ti³⁺ to O⁻ [32]. However, as TiO₂ was modified with CuO, CeO₂, and Bi₂O₃, the PL intensity decreased gradually, suggesting that the addition of metal oxides could effectively retard the recombination of electron/hole pairs and thus extend the lifetime of active species. For all the tested photocatalysts, the emission bands were observed mainly in the range of visible light (λ = 428–604 nm). This could be attributed to the recombination of excited electrons with oxygen vacancies on the photocatalysts. On the other hand, a small peak appeared in the range of UV_A light, which is attributed to near-band emissions and the recombination of electron and hole pairs [33].

Figure 7 illustrates the XPS spectra of Ti 2p over TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂. It shows two characteristic peaks at around 458 and 464 eV in the pristine TiO₂, attributed to Ti 2p_3/2 and Ti 2p_1/2, respectively, also indicating that Ti appears in Ti⁴⁺. After modification with metal oxides, the electron cloud around Ti decreased and the oxidation state decreased, leading to a shift of the peaks to lower binding energy [34]. Figure 8 depicts the characteristic peaks of metal oxides in the photocatalysts. Cu had two characteristic peaks at 933.8 and 953.4 eV, labeled Cu 2p_3/2 and Cu 2p_1/2, respectively, with a binding energy gap of 19.6 eV, indicating that Cu was mainly in the form of Cu²⁺ instead of Cu¹⁺ and Cu⁰ [35]. Ce 3d spectra were classified into Ce 3d_3/2 and Ce 3d_5/2, labeled u and v, respectively. The peaks u¹ and v¹ corresponded to Ce⁴⁺, while the peaks u, v, u², v², u³, and v³ corresponded to Ce³⁺ [36]. The Bi 4f spectra showed that Bi was present as Bi⁰, with two characteristic peaks at 154.9 and 160.2 eV, and Bi³⁺, which exhibited two characteristic peaks at 157.2 and 162.5 eV [37]. Different valence states provide strong oxygen storage capacity, which can capture oxygen from the atmosphere and oxidize Hg⁰ to Hg²⁺. Figure 9 presents the O 1s spectra of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂. Herein, O_α at a lower binding energy is assigned to lattice oxygen, while O_β at a higher binding energy is the chemisorbed oxygen and surface hydroxyl group, which has been reported to have higher activity to conduct oxidation reactions [38,39]. The integration area showed an increased proportion of O_β with the modification of metal oxides, further increasing Hg⁰ oxidation efficiency. Additionally, the increased binding energy compared with that of TiO₂ indicated an increased oxidation state (O²⁻ to O⁻), which could enhance the photocatalytic oxidation of Hg⁰ [40].

2.2. Effects of Different Modifying Materials on Photocatalytic Efficiency of Hg⁰

Figure 10 illustrates the photocatalytic oxidation efficiencies of Hg⁰ in an N₂ + Hg⁰ atmosphere at reaction temperatures ranging from 100 to 200 °C for 2 h using TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂. The photocatalytic oxidation efficiencies of Hg⁰ for TiO₂, CuO/TiO₂, and CeO₂/TiO₂ decreased with reaction temperature. However, an opposite trend was observed for Bi₂O₃/TiO₂. The photocatalytic oxidation efficiencies of Hg⁰ for TiO₂ was 91% at 100 °C and then dropped to 36% at 200 °C, showing a dramatic decrease of nearly 60% as reaction temperatures increased from 100 to 200 °C. However, after modifying using Bi₂O₃, the photocatalytic oxidation efficiency of Hg⁰ improved. To clarify the contribution of the photocatalytic and the traditional catalytic oxidation, we conducted continuous experiments, first without UV_A to let the catalyst react as a traditional thermal catalytic reaction. The catalytic oxidation efficiency of Hg⁰ reached its maximum after 30 min, as shown in Figure 11. As a catalyst, TiO₂ had less than 5% of the catalytic oxidation efficiency of Hg⁰ without the irradiation of UV_A. By modifying TiO₂ with metal oxides, the oxidation efficiency of Hg⁰ increased significantly. CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ had the best catalytic oxidation efficiencies of Hg⁰ at 100 °C, which were 15%, 32%, and 52%, respectively. Secondly, we turn on UV_A to allow the materials to undergo the photocatalytic reaction. The photocatalytic reaction efficiencies of Hg⁰ increased dramatically for metal-oxide-modified TiO₂. The results showed that, for all the tested catalysts, photocatalysis contributed more than classical thermal catalysis. A previous study reported that TiO₂ modified with metal oxide could yield more surface oxygen [41]. These findings concurred with the XPS analytical results obtained in this study. Equations (1)–(3) show the production of surface chemisorbed oxygen (O_β) through the chemical reaction of metal oxides. Metal oxides can promote photocatalytic reactions, with their effectiveness in the order Bi₂O₃ > CeO₂ > CuO. At 200 °C, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ had performances of 47, 71, and 89%, respectively. Furthermore, Bi₂O₃ exhibited excellent thermal stability, retaining almost the same oxidation efficiency at temperatures between 100 and 200 °C.

2 CuO → Cu₂O + O_β

(1)

2 CeO₂ → Ce₂O₃ + O_β

(2)

Bi₂O₃ → 2 Bi⁰ + 3 O_β

(3)

2.3. Kinetic Modeling of the Photocatalytic Oxidation of Hg⁰

The L–H kinetic mechanism was applied to simulate the photocatalytic oxidation efficiency of Hg⁰ [42]. In the context of a heterogeneous photocatalytic reaction, a decrease in K_Hg⁰ with the rise in the reaction temperature signified a chemical reaction characterized as physical adsorption. In contrast, an increase in K_Hg⁰ with reaction temperature characterized the chemical reaction as chemisorption, where adsorbed molecules form chemical bonds and adhere to the surface of the adsorbents [43].

As depicted in

Table 3. Langmuir–Hinshelwood kinetic parameters for different photocatalysts.

Types of Photocatalysts	Reaction Temperatures (°C)	KHg0 (m3 μg−1)	kr (μg m2 min−1)	Ea (kcal mol−1)
TiO2	100	9.264	0.341	9.24
	150	0.007	1.038
	200	0.005	1.071
CuO/TiO2	100	9.049	0.692	2.54
	150	0.268	0.940
	200	0.234	1.52
CeO2/TiO2	100	11.671	0.647	1.49
	150	4.843	0.894
	200	0.778	1.331
Bi2O3/TiO2	100	9.854	0.142	1.35
	150	1.069	0.429
	200	0.192	1.238

, an increase in reaction temperature led to an elevation in the kr and a decrease in the value of K_Hg⁰. This trend suggests that at higher reaction temperatures, the adsorption efficiency of Hg⁰ on the surface of photocatalysts diminishes. Based on the aforementioned outcomes, in this study, we inferred that physical adsorption predominantly governed the photocatalytic reactions of Hg⁰ for all the photocatalysts prepared. We further calculated the activation energy (E_a) according to Arrhenius Law. The respective activation energies of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ were 9.24, 2.54, 1.49, and 1.53 kcal/mol. These results indicate that the modification of metal oxides on TiO₂ could effectively reduce the reaction barrier and improve the photocatalytic oxidation efficiency of Hg⁰ [44].

3. Materials and Methods

3.1. Chemicals

The following chemical reagents were used for synthesizing metal-oxide-modified TiO₂-based photocatalysts with no further purification, included titanium isopropoxide (ACROS ORGANICS, Waltham, MA, USA, Ti(OCH(CH₃)₂)₄, 98%), copper nitrate (Alfa Aesar, Haverhill, MAs, USA, Cu(NO₃)₂, 99.5%), cerium nitrate (Alfa Aesar, Ce(NO₃)₃·6H₂O, 99.5%), bismuth nitrate (Alfa Aesar, Bi(NO₃)₃·5H₂O, 99.5%), and isopropanol (Shimakyu’s Pure Chemicals, Osaka, Japan).

3.2. Preparation of Metal-Oxide-Modified TiO₂ Photocatalysts

Anatase TiO₂ was synthesized using the hydrothermal method [45]. First, 20 mL of titanium isopropoxide and 40 mL isopropanol were stirred for 1 h to form a uniform solution. Next, the solution was transferred to a Teflon container in an autoclave and kept in a furnace at 200 °C for 24 h. After the hydrothermal reactions, the solution was cleaned three times with DI water to remove organic compounds by centrifuging at 7000× rpm for 10 min. The solution was then dried at 80 °C to remove the residual organic compounds and the dry precipitate was further calcined at 450 °C for 4 h to obtain TiO₂.

The modification of CuO and CeO₂ to TiO₂ was further performed using an impregnation method [46]. The molar ratios of CuO/TiO₂ and CeO₂/TiO₂ were 5% and 7%, respectively. Initially, certain amounts of metal oxide precursors were added to the solution of self-prepared TiO_2, and DI water, and magnetically stirred for 12 h at room temperature. The solution was centrifuged to remove metal ions and the precipitate was then dried at 80 °C and calcined at 450 °C for 2 h. In this study, the molar ratio of Bi₂O₃ and TiO₂ was 3%. Bi₂O₃/TiO₂ was synthesized using the hydrothermal method [47]. Bismuth nitrate was added to the solution of prepared TiO₂ and DI water and magnetically stirred for 1 h at room temperature. The solution was then transferred to the autoclave for a hydrothermal reaction at 180 °C for 24 h and then centrifuged to remove metal ions. The precipitate was further dried at 80 °C and then calcined at 450 °C for 4 h.

3.3. Characterization Analysis of Photocatalysts

Surface characterization analysis of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ was performed using various physicochemical analytical instruments. The morphology and element dispersion of photocatalysts were observed using a field emission-scanning electron microscope (FE-SEM, Zeiss, Oberkochen, Germany, Gemini 450), transmission electron microscope (TEM, Philips, Amsterdam, The Netherland, CM-200 TWIN), and energy-dispersive X-ray spectroscopy (EDS, Zeiss, Germany, Oberkochen, Gemini 450). The specific surface area (SSA) of each photocatalyst was measured using a specific surface area analyzer (SSAA, Micromeritics, Norcross, GA, USA, ASAP2000). The crystallographic structures of the photocatalysts were analyzed using an X-ray diffractometer (XRD, Bruker, Billerica, MA, USA, D8 DISCOVER). The chemical compositions and valence states were analyzed using X-ray photoelectron spectroscopy (XPS, Philips, Amsterdam, The Netherland Hybrid Quantera). The recombination times of photo-induced electron/hole pairs were estimated using photoluminescence (PL, Horiba, Kyoto, Japan, HR800). The band gap of each photocatalyst was analyzed using an ultraviolet-visible spectrometer (UV-Vis, Perkin-Elmer Precisely, Waltham, MA, US, Lambda 850).

3.4. Photocatalytic Activity of Self-Prepared Photocatalysts

For this particular study, a continuous flow photocatalytic reaction system comprising a standard gas generator, a mass flow controller, a mixing chamber, a photocatalytic reaction tube, and a real-time mercury monitor (NIC, EMP-2, measurement range = 0–1000 µg/m³, resolution = 0.1 µg/m³, and response time = 1 s) was established. A standard gas generator with a Hg⁰ permeation tube released the desired concentration of Hg⁰, which was heated to 100 °C in an inert gas (N₂) and further diluted with N₂ and 6% O₂ in the mixing chamber. The mixed gas was allowed to flow through the photocatalytic reactor to react with the photocatalysts coated on the surface of glass beads with 2 mm diameters. The photocatalytic reactor consisted of a photocatalytic reaction tube, with a black light of 365 nm wavelength (Sankyo Denki, BB-15W) in the middle of the reactor. The photocatalysts were placed between the reaction tube and the blacklight lamp. The blacklight lamps provided near-UV light of 15 W intensity and 365 nm wavelength. Additionally, the photocatalytic reaction tube was surrounded by a heating tape to maintain the photocatalytic reaction temperatures of 100, 150, and 200 °C.

Finally, the concentration of Hg⁰ at the inlet and outlet of the photocatalytic reactor was measured at a frequency of 1 plot/s. Calculating the concentration gradient between the inlet and outlet, the photocatalytic oxidation efficiency of Hg⁰ (η_Hg⁰) was further derived as shown in Equation (4). The mean and standard deviation (Mean ± SD) were further calculated to describe the variation in the experimental data when the data reached the steady state.

$${\eta }_{Hg0}=\frac{{\Delta Hg}^0}{{Hg}_{in}^0}=\frac{{Hg}_{in}^0-{Hg}_{out}^0}{{Hg}_{in}^0}\times 100\%$$

(4)

3.5. Langmuir–Hinshelwood (L–H) Kinetic Model

In this investigation, a Langmuir–Hinshelwood (L–H) kinetic model was applied to assess the correlation between the photocatalytic oxidation efficiency and the reaction rate of Hg0 across its various inlet concentrations. The adsorption equilibrium of Hg0 on the surface of photothermal catalysts was characterized by employing a Langmuir isotherm [48,49]. By considering the distribution of adsorbates on the surface of photocatalysts, the L–H kinetic model exhibited the reaction kinetics of the oxidation of Hg0. This model helped elucidate the adsorption of Hg0 over the surface of photocatalysts and the activation energy of photocatalytic oxidation of Hg0. By substituting inlet and outlet Hg0 concentrations into Equations (5) and (6), a linear correlation was derived, subsequently determining its intercept (Equation (7)) and slope (Equation (8)). This process allows us to deduce the reaction rate constant (kr) and the equilibrium constant (KHg0) of the photocatalytic reaction of Hg0 using a straight line plot of r−1 (herein, r represents reaction rate) versus CHg0−1.

$$r=\frac{k_r\left(K_{H{g}^0}{C}_{H{g}^0}\right)}{1+K_{H{g}^0}{C}_{H{g}^0}}$$

(5)

$$\frac{1}{r}=\frac{1}{k_r}+(\frac{1}{{k_{r}K}_{H{g}^0}})(\frac{1}{C_{H{g}^0}})$$

(6)

$$\frac{1}{k_r}=intercept$$

(7)

$$\frac{1}{{k_{r}K}_{H{g}^0}}=slope$$

(8)

4. Conclusions

In this study, three metal oxides (CuO₂, CeO₂, and Bi₂O₃) were employed to modify TiO₂ in order to enhance the photocatalytic oxidation efficiency of Hg⁰. While the modification of metal oxides might have blocked the pore structure of anatase TiO₂, it significantly reduced the PL intensity, which benefitted photocatalytic oxidation reactions. Moreover, the catalysts under modification exhibited a higher partition of chemisorbed oxygen. At 200 °C, these two enhancements could effectively improve the photocatalytic oxidation efficiency of Hg⁰ in the following order: Bi₂O₃/TiO₂ > CeO₂/TiO₂ > CuO/TiO₂ > TiO₂. Based on the comparison of catalytic (light off) and photocatalytic (light on) oxidation experiments of Hg⁰, we revealed that Bi₂O₃/TiO₂ was the best composite to photocatalytically oxidize Hg⁰ at 100–200 °C compared with CeO₂/TiO₂ and CuO/TiO₂. The innovative findings obtained in this study were mainly attributed to the fact that Bi₂O₃/TiO₂ had a higher specific surface area, which could enrich more chemisorbed oxygen on its surface, compared with CuO/TiO₂ and CeO₂/TiO₂. The simulation of the L–H kinetic model revealed that the overall Hg⁰ photocatalytic reaction was dominated by physical adsorption. Furthermore, another novelty of this study was to inter-compare the activation energy (E_a) based on the Arrhenius Law. It showed that the activation energies of TiO₂, CuO/TiO₂, CeO₂/TiO₂, and Bi₂O₃/TiO₂ were 9.24, 2.54, 1.49, and 1.35 kcal/mole, respectively, and the energy barriers were reduced significantly by metal oxide modification. We thus concluded that Bi₂O₃/TiO₂ was the best photocatalyst because it had the highest photocatalytic oxidation of Hg⁰.