Elimination of PCDD/Fs over Commercial Honeycomb-Like Catalyst of V₂O₅-MoO₃/TiO₂ at Low Temperature: From Laboratory Experiments to Field Study

Abstract

With the need for ultra-low emissions and the strict regulation of PCDD/Fs from MSWI plants, traditional SCR catalysts have been applied to remove PCDD/Fs. In this study, we compared one typical commercial V₂O₅-MoO₃/TiO₂ catalyst’s performance in removing PCDD/Fs under laboratory and industrial conditions. Various characterization methods like XRF, XPS, BET, and H₂-TPR were applied to analyze the catalyst’s properties. The laboratory results showed that the adsorption could significantly affect the removal at low temperatures. The RE on PCDD/Fs was 59.4% (55.0% for toxicity RE), 88.5% (90.3%), and 78.0% (76.0%) at 160 °C, 180 °C, and 200 °C, respectively, showing that 180 °C is the most suitable operation temperature for this V₂O₅-MoO₃/TiO₂ catalyst. The field study was conducted at 180 °C, and the results revealed that the competition between water vapor and the interaction of SO2 could lower the RE. However, comparisons between laboratory and field conditions showed that this V₂O₅-MoO₃/TiO₂ catalyst still showed good stability, with only a 6.8% drop.

1. Introduction

Incineration is the most efficient way to deal with dramatically increasing municipal solid waste (MSW). However, highly toxic polychlorodibenzon-p-dioxins and polychloro-dibenzofurans (PCDD/Fs) are generated during combustion, threatening people’s health. Multiple control strategies, such as the injection of inhibitors [1], activated carbon with fabric filters [2], and catalytic technologies [3], were applied to reduce the generation and emission of PCDD/Fs from MSWI plants. As we know, the injection of inhibitors can only reduce the regeneration of PCDD/Fs in the post-combustion area, and the injection of activated carbon combined with fabric filters can only absorb PCDD/Fs without any reduction. Catalytic technologies, such as selective catalytic reduction (SCR) catalysts [3], can decompose the PCDD/Fs without secondary pollution, which has become a critical choice for purifying the flue gas.

The elimination of PCDD/Fs using vanadium-based catalysts has been widely reported [4,5,6,7]. Hsu et al. [4] investigated one commercial V₂O₅-WO₃/TiO₂ catalyst, and the destruction efficiency of the PCDD/Fs could reach 84–91% at 280 °C, with a space velocity of 5000 h⁻¹. Liu et al. [6] studied the correlation between PCDD/Fs removal efficiency and temperatures, showing the highest removal efficiency of PCDD/Fs could reach 97.24% at 300 °C. On the contrary, Chang et al. [5] reported that the SCR sector of an MSWI plant using V₂O₅-WO₃/TiO₂ could remove 93% of PCDD/Fs at 230 °C; however, lower removal efficiency (52.3%) was observed at 290 °C. A comprehensive study by Zhao et al. [8] revealed that the regeneration of PCDD/Fs was observed at 220 °C using various catalysts, among which the V₂O₅-CuO/TiO₂ catalyst could significantly increase the total amount of PCDD/Fs. These studies [5,8] demonstrated that the reaction temperatures of the SCR catalysts needed to be set up at low temperatures to avoid overlapping the heterogeneous formation temperatures of the PCDD/Fs (200–400 °C) [9]. Moreover, appreciable energy would be saved if the flue gas could be maintained at low temperatures.

More recent studies [10,11,12] focused on the performance of catalysts at low temperatures. Yu et al. [10] investigated several transient metal oxide catalysts’ removal efficiency on PCDD/Fs, among which the V₂O₅/TiO₂ catalyst showed a more than 75% removal efficiency of toxic concentration at 200 °C. Wang et al. [12] found that the removal efficiency of the V₂O₅/TiO₂-CNTs catalyst could reach 99.9% at 150 °C, and the adsorption was regarded as the primary mechanism. Our previous field study [11] reported that the granular V₂O₅/TiO₂ catalyst showed excellent synergistic removal of the PCDD/Fs at low temperatures; i.e., the removal efficiency of PCDD/Fs can reach 94.59%, 90.57%, and 87.48% at 200 °C, 180 °C, and 160 °C, respectively. However, further study [13] showed the serious poisoning issues of the granular V₂O₅/TiO₂ catalyst after the 1 year operation. Characteristic results illustrated that the oxidative and acid sites were covered by sulfate species, hindering the adsorption of PCDD/F molecules and further redox progress. The sulfate poisoning problem became more prominent due to the decreasing temperatures [14,15,16]. Various transient metal oxides, such as MoO₃, WO₃, CoO₃, and CeO₂, were introduced to solve this problem and enhance catalytic performance. The V₂O₅-WO₃/TiO₂ [4,17,18] and V₂O₅-MoO₃/TiO₂ catalysts [17,19,20] in particular, had a significant market share in the SCR units of MSW incineration plants, due to their excellent performance in poison resistance [21], high removal efficiency [4], and wide reaction temperature window [22]. The V₂O₅-WO₃/TiO₂ catalysts were regarded as the most suitable for coupled de-NOx and PCDD/F destruction systems when the reaction temperature window [7], working life [21], and mercury conversion [4] were considered. However, further studies [17,23,24] discovered more advantages of V₂O₅-MoO₃/TiO₂ at low temperatures in recent years. Huang et al. [23] reported that the V₂O₅-MoO₃/TiO₂ catalyst’s conversion efficiency of chlorobenzene reached 50% at 185 °C, while V₂O₅-WO₃/TiO₂ reached the same efficiency at 210 °C. Chai et al. [24] found that HSO_4⁻ was easier to be resolved into SO₃ and H₂O, enhancing the sulfur resistance of the MoO₃/TiO₂ catalyst. Peng et al. [17] compared the conduction band gap of O 2p of the V₂O₅-MoO₃/TiO₂ and V₂O₅-WO₃/TiO₂ catalysts; the band gap of AsWTi was more significant than that of AsMoTi, which indicated the MoO₃ had a better arsenic resistance effect than WO₃. Debecker et al. [20] evaluated the oxidation performance overPCDD/Fs using one MoO₃-promoted V₂O₅/TiO₂ catalysts; the removal efficiency of PCDD/Fs could achieve over 85% at 200 °C. These newfound properties of the V₂O₅-MoO₃/TiO₂ catalyst enhanced its application to PCDD/Fs at low temperatures.

The catalytic performance of V₂O₅-MoO₃/TiO₂ has been previously investigated [10,19,23]. However, it was mainly based on indicators like chlorotoluene [19] and chlorobenzene [23]. Some researchers pointed out the shaky reliability of this method [25]. In addition, catalytic destruction performance and the mechanism of the V₂O₅-MoO₃/TiO₂ catalyst over PCDD/Fs were less reported in the laboratory and industrial conditions, compared to V₂O₅-WO₃/TiO₂ catalysts. The catalysts mentioned [18,19,20,23] were mainly in powder form due to the multiple preparation methods, such as impregnation and sol-gel methods. Monolithic catalysts, such as honeycomb-like forms, were more popular in the industry due to their excellent structural and thermal stabilities, low-pressure drop, and high plugging tolerance [26]. Also, industrial conditions are much more complicated. Multiple components, such as fly ash, SO₂, steam, and NOx, could influence the catalytic degradation of PCDD/Fs [12,18,27]. These differences distinguish the laboratory results from industrial applications.

This study conducted a series of experiments to evaluate the performance of one commercial honeycomb-like V₂O₅-MoO₃/TiO₂ catalyst under laboratory and field conditions. Under laboratory conditions, the catalyst’s adsorption and destruction efficiency were dissected to investigate the influence of adsorption properties on removing PCDD/Fs. Then, full-scale experiments were conducted to explore how the influence of complicated flue gas conditions compromise the catalyst’s oxidation properties. These results could provide a combined perspective of V₂O₅-MoO₃/TiO₂ on PCDD/F removal under laboratory and industrial operation conditions. Considering PCDD/Fs have multiple congeners, a total of 134 congeners and 17 toxic congeners were analyzed to better reveal the regularity of PCDD/Fs destruction.

2. Materials and Methods

2.1. Catalyst and Characterization

The V₂O₅-MoO₃/TiO₂ catalyst was provided by the MSW incineration plant, which was also applied in the SCR system. Multiple characterization methods were applied to evaluate the characteristics of the catalyst. X-ray fluorescence (XRF) spectrometer (Thermo Fisher Scientific, ADVANT’X 4200, Waltham, MA, USA) was used to determine the main contents of the catalyst. Specific surface area and pore size distribution of the catalysts were gained by N₂-adsorption/desorption and Brunner–Emmet–Teller (BET) measurements (TRISTAR 3020, Micromeritics Instrument Corporation, Norcross, GA, USA). The H₂-temperature-programmed reduction (H₂-TPR, Atochem II 2920, Micromeritics Instrument Corporation, Norcross, GA, USA) experiment was carried out to analyze the reducibility of catalysts, X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher Scientific Inc., Waltham, MA, USA) was applied to determine the chemical states of the elements on the catalysts’ surface, and X-ray diffraction (XRD, X’Pert PRO, PANalytical B.V., Almelo, Overijssel, The Netherlands) was applied to analyze crystal morphology with scan scope from 10° to 90°.

2.2. Laboratory Experimental Design

The laboratory study was conducted on a PCDD/F generation system, which can generate consistent and stable vapor PCDD/Fs. The vapor PCDD/Fs evaporated into the gas phase, heated by the furnaces. As shown in Figure 1, the system consists of a PCDD/F solution nebulizer, a preheater with a temperature controller and quartz tube, and gas flow rate controllers controlling the gas flow rate of the ratio of N₂ and O₂. The PCDD/F solution was extracted from a medical waste incineration plant’s fly ash, and the purification and condensation process were described in our previous study [28]. A gas mixture of N₂ and O₂ (the volume of O₂ was 11%) was introduced as carrier gas with a total flow volume of 500 mL/min, and a honeycomb-like catalyst was cut into a 3.75 mL cuboid with 4 square holes at the bottom (to ensure GHSV was 8000 h⁻¹). The catalyst was placed in the quartz tube, and the reaction temperatures were controlled at 160 °C, 180 °C, and 200 °C, respectively. The catalytic reaction time of each run was 1 h. 3 initial parallel samples without catalyst were collected to determine the initial concentration of PCDD/Fs, and 2 parallel samples with catalyst were collected at each temperature. The initial parallel samples were used to determine the initial concentration of PCDD/Fs and the instrument’s stability. The instrument had been running for 8 h before the evaluation tests of the catalysts. The toluene solution absorbed the exhaust gas. The catalyst was collected to analyze the catalyst’s adsorption amount of PCDD/Fs.

2.3. Full-Scale Experiments Designed in the MSWI Plant

The experiments were conducted in a full-scale (720 t/d) mechanical-grate MSW incinerator with a flow gas volume of 34.05 m/s (based on daily average) in the SCR sector. The catalyst volume was about 36 m³, and they were put into operation for 2 weeks before sampling. Two sampling sites were set at the inlet and outlet of SCR shown in Figure 2. Before the SCR sector, the semi-dry scrubber-activated carbon contacting chamber and the baghouse filter were placed to remove fly ash, SO₂, and PCDD/Fs. The operating temperature of exhaust gas was 180 °C. During each test, the samples of PCDD/Fs were simultaneously collected at two sampling sites, and each sample collected more than 2 Nm³ of flue gas. The gas-phase PCDD/Fs were collected in XAD-2 polymeric resin and condensed water, while the particulate-phase PCDD/Fs were collected by filter cartridge, and they were pretreated separately.

2.4. Statistical Analysis

For the laboratory study, the removal efficiency can be calculated by Equation (1); destruction efficiency and adsorption efficiency are denoted as “DE” and “AE”, respectively, and are calculated as follows:

$$\mathrm{DE}(\%)=\left(C_{initial}-C_{exhaust}-C_{adsorption}\right)÷ C_{initial}$$

(1)

$$\mathrm{AE}(\%)=C_{adsorption}÷C_{initial}$$

(2)

where $C_{initial}$ indicates the initial concentration of PCDD/Fs; $C_{exhaust}$ represents the concentration of PCDD/Fs collected by toluene solution in the exhaust gas; and $C_{adsorption}$ denotes the adsorption concentration of PCDD/Fs on the catalyst’s surface.

For the field study, 136 congeners were calculated as mass concentration, and 17 toxic congeners were calculated as mass concentration and I-TEQ (International Toxicity Equivalence Quantity) [29]. The removal efficiency can be defined as follows:

$$\mathrm{RE}(\%)=\left(C_{inlet}-C_{outlet}\right)÷C_{inlet}$$

(3)

where $C_{inlet}$ and $C_{outlet}$ indicate the concentration of PCDD/Fs in the inlet and outlet flue gas, respectively. To better know the distribution of gas- and particulate-phase PCDD/Fs, the $C_{inlet}$ and $C_{outlet}$ can also be indicated as:

$$C_{inlet}=C_{inlet,gas}+C_{inlet,particulate}$$

(4)

$$C_{outlet}=C_{outlet,gas}+C_{outlet,particulate}$$

(5)

where $C_{inlet,gas}$ and $C_{outlet,gas}$ are the gas-phase PCDD/Fs collected by XAD-2 polymeric resin and condensed water in the inlet sampling site and outlet sampling site, respectively; $C_{inlet,particulate }$ and $C_{outlet,particulate}$ are the particulate-phase PCDD/Fs captured by filter cartilage in the inlet sampling site and outlet sampling site respectively.

To simplify the analysis process, 134 congeners were classified into 10 homologues in the figures below.

3. Results and Discussion

3.1. Laboratory Research on the PCDD/Fs Removal

The 10-homologue distributions are exhibited in

Table 1. The initial table concentration and the RE of different homologues at different temperatures (ng/Nm³, ng I-TEQ/Nm³).

	160 °C			180 °C			200 °C
Homologues	Initial	After	RE	Initial	After	RE	Initial	After	RE
TCDD	4.28	1.99	53.5%	8.99	1.24	86.2%	5.57	2.01	63.9%
PeCDD	47.17	19.28	59.1%	91.25	10.34	88.7%	62.97	15.71	75.0%
HxCDD	19.63	8.43	57.1%	39.48	3.78	90.4%	26.24	6.16	76.5%
HpCDD	28.57	13.28	53.5%	65.93	7.19	89.1%	72.63	12.72	82.5%
OCDD	12.62	5.49	56.5%	27.92	3.24	88.4%	49.85	10.00	79.9%
TCDF	5.58	2.57	54.0%	9.14	1.37	85.0%	7.10	2.38	66.5%
PeCDF	2.33	1.04	55.6%	7.81	1.56	80.1%	3.75	0.96	74.3%
HxCDF	1.27	0.63	50.8%	2.81	0.40	85.7%	1.55	0.53	65.6%
HpCDF	0.39	0.31	22.5%	1.04	0.24	77.3%	0.92	0.24	73.8%
OCDF	0.23	0.11	53.0%	0.57	0.09	84.9%	0.82	0.22	73.5%
PCDD	112.28	48.48	56.8%	233.57	25.79	89.0%	217.27	46.61	78.5%
PCDF	9.81	4.64	52.7%	21.37	3.65	82.9%	14.14	4.34	69.3%
PCDD/Fs	122.09	53.12	56.5%	254.94	29.44	88.5%	231.41	50.94	78.0%
I-TEQ	1.55	0.70	55.0%	3.39	0.36	89.3%	2.34	0.56	76.0%

. The average inlet concentration was 122.10, 254.94, and 231.41 ng/Nm³ at 160 °C, 180 °C, and 200 °C, respectively (1.55, 3.39, 2.34 ng I-TEQ/Nm³, respectively). The initial concentration varied because the PCDD/Fs had different adsorption performances due to the various experiment temperatures. These samples appeared to have good consistency in congener distribution, among which the PeCDD, HxCDD, HpCDD, and OCDD devoted most mass concentration in initial samples, and 1,2,3,4,7,8-HxCDD, 1,2,3,7,8,9-HxCDD, 1,2,3,4,6,7,8-HpCDD, OCDD, and 1,2,3,4,7,8-HxCDF devoted most TEQ concentration in initial samples.

Figure 3 depicts the RE of the PCDD/Fs, and the RE reached 56.5%, 88.5%, and 78.0% at 160 °C, 180 °C, and 200 °C, respectively (54.6%, 89.3% and 76.0% for toxicity, respectively). Our team had conducted the same experiment using several commercial honeycomb-like V₂O₅-WO₃/TiO₂ catalysts under the 8000 h⁻¹ 180 °C condition, and the RE was 89.7% [28] and 91.2% [28], respectively. Compared with the V₂O₅-WO₃/TiO₂ catalysts, this V₂O₅-MoO₃/TiO₂ catalyst has comparable catalytic performance on PCDD/Fs under the same laboratory conditions. The RE dropped from 88.5% to 78.0% when the temperature increased from 180 °C to 200 °C, because the adsorption ability decreased with the increasing saturated vapor pressure of the PCDD/Fs [30]. The DE of total PCDD/Fs increased from 25.6% to 65.4%, with temperature rising from 160 °C to 200 °C. The oxidation of PCDD/Fs obeyed the Mars–Van Krevelen mechanism [31], which depended on V⁵⁺ species. The V⁵⁺ species provided V=O and V–O support sites, the PCDD/F’s molecule adhered at the site, and the Cl-C was attacked. The cyclic structure became unstable with the loss of Cl, and later, the surface-absorbed oxygen attacked the cyclic structure and broke the benzene ring. The V⁵⁺ species were reduced into V⁴⁺ species, and the O₂ re-oxidized V⁴⁺ into V⁵⁺. The higher temperature enhanced the circulation rate, resulting in higher destruction efficiency.

However, the DE of toxicity dropped by 4.64% when the temperature increased from 180 °C to 200 °C. The DE was combined and determined by adsorption ability and oxidation ability. The adsorption ability of the PCDD/Fs decreased with increasing temperature, and the physical adsorption was regarded as the first step of the PCDD/Fs’ destruction [32,33]. Generally, the low-chlorinated PCDD/Fs with higher saturated vapor pressures are more toxic, and sensitive to temperature change [30]. In this study, the catalytic performance was limited by oxidation ability at 160 °C and 180 °C. When the temperature increased to 200 °C, the adsorption ability became the main hindrance to destruction performance, rather than the oxidation ability. The low-chlorinated PCDD/Fs failed to be captured by the catalyst, causing the decreasing DE calculated as toxic concentration. A transition of dynamics to diffusion limitation was observed at 300 °C using the V₂O₅/TiO₂ catalyst [34], and at 220 °C using the V₂O₅-WO₃/TiO₂ catalyst [28], proving that the V₂O₅-MoO₃/TiO₂ catalyst showed better oxidation ability at low temperatures.

Also, the de-chlorination effect happened during the destruction process [35]. Figure 4a shows that the V₂O₅-MoO₃/TiO₂ catalyst had only slight catalytic abilities on OCDF (2.37%), and performed much better on degrading the low-chlorinated PCDD/Fs with higher toxicity, such as TCDD (81.13%), PeCDD (83.97%), HxCDD (78.65%), TCDF (48.34%), and PeCD (58.67%), at 180 °C. When the temperature increased to 200 °C, the degradation over OCDF (52.81%), HpCDF (51.10%), and OCDD (65.40%) increased conspicuously, while the DE of TCDD (52.82%), PeCDD (64.35%), HxCDD (63.44%), TCDF (47.78%) and PeCDF (55.89%) decreased because of the de-chlorination effect. The RE of the 17 toxic congeners in Figure 4b exhibited the recession of DE on the more toxic congeners, such as 2,3,7,8-TCDD (72.87% at 180 °C and −5.2% at 200 °C), 1,2,3,7,8-PeCDD (87.13% at 180 °C and 50.92% at 200 °C), 1,2,3,4,7,8-HxCDD (78.11% at 180 °C and 52.14% at 200 °C), 1,2,3,6,7,8,-HxCDD (77.37% at 180 °C and 63.24% at 200 °C), 2,3,7,8-TCDF(53.99% at 180 °C and 9.35% at 200 °C), 1,2,3,7,8-PeCDF(48.80% at 180 °C and 31.01% at 200 °C), and 1,2,3,6,7,8-HxCDF (24.37% at 180 °C and −5.10% at 200 °C).

3.2. Field Investigation on the PCDD/F Removal

Table 2. The PCDD/Fs’ concentrations in different phases. (ng/Nm³, ng I-TEQ/Nm³).

	Inlet		Outlet		RE
	Gas Phase	Particulate Phase	Gas Phase	Particulate Phase	Gas Phase	Particulate Phase
TCDD	7.04	0.04	0.49	0.06	93.0%	−26.4%
PeCDD	2.14	0.01	0.24	0.02	88.6%	−87.6%
HxCDD	0.41	0.02	0.08	0.03	79.6%	−56.7%
HpCDD	0.14	0.02	0.07	0.01	50.2%	55.1%
OCDD	1.83	1.37	0.58	0.04	68.3%	96.8%
TCDF	16.39	0.05	1.17	0.08	92.9%	−55.6%
PeCDF	2.01	0.02	0.20	0.04	90.0%	−69.7%
HxCDF	0.44	0.02	0.16	0.03	64.6%	−50.9%
HpCDF	0.14	0.01	0.09	0.01	37.2%	−21.1%
OCDF	0.05	0.01	0.04	0.01	26.5%	−24.6%
PCDD	11.55	1.46	1.47	0.16	87.3%	89.1%
PCDF	19.03	0.11	1.65	0.16	91.3%	−51.6%
PCDD/Fs	32.15	3.44	3.12	0.32	89.8%	79.4%
	32.15		3.44		89.3%
Toxicity of PCDD	0.040	0.0018	0.0083	0.0010	79.4%	50.0%
Toxicity of PCDF	0.097	0.0011	0.0150	0.0016	84.7%	−45.5%
Total toxicity	0.14	0.0029	0.0230	0.0025	83.1%	13.6%
	0.141		0.026		81.7%

exhibited the average PCDD/F distribution of the gas and particulate phase, and Figure 5 shows the PCDD/F congener distribution. The PCDD/F concentration in the inlet was 32.15 ng/Nm³ (0.141 ng I-TEQ/Nm³) on average. The concentration of the particulate-phase PCDD/Fs was much lower than that of the gas phase. In the gas-phase PCDD/Fs, the TCDD and TCDF with high toxicity are the most prominent congeners of PCDD/Fs, while the low-toxicity OCDD contributed most in the particulate phase. The phase distribution showed a similar regularity in various waste incinerators in previous studies [36,37]. That is because the PCDD/Fs with higher chlorination degree levels were easier to be adsorbed onto particulate matter and be captured by activated carbon and fly ash before the bag filters due to the low saturated vapor pressure [37]. When the flue gas came through the bag filter, most particulate phase PCDD/Fs were removed, while the remaining gas-phase PCDD/Fs were still in the flue gas [30].

The concentration of the 10 homologous PCDD/Fs in the outlet was 3.44 ng/Nm³ (0.026 ng I-TEQ/Nm³), and the RE was 89.2% (81.7% for toxic concentration). The V₂O₅-MoO₃/TiO₂ catalyst used in this study showed better low-temperature performance than the V₂O₅/TiO₂ catalyst (The RE was 39% at 200 °C) reported by Li et al. [38] and the V₂O₅-WO₃/TiO₂ catalyst (The RE was 75.5% at 350 °C) tested by Wang et al. [39]. According to the gas PCDD/Fs, the RE of all congeners varied from 26.54% to 93.01%. The total concentration RE of the PCDFs (91.32% for total mass concentration, 84.70% for toxic concentration) is higher than the PCDDs (87.28% for total mass concentration, 79.40% for toxic concentration). This can be attributed to the large proportion of the less thermodynamically stable 2,3,4,7,8-PeCDF (0.060 TEQ ng I-TEQ/Nm³). Because the PCDF with chlorinated substituents on the 2, 4 and 8 chlorinated positions are easier to be destructed [40].

On the contrary, most particulate phase PCDD/Fs increased instead. As exhibited in Figure 5, all particulate PCDD/Fs increased prominently, except HpCDD and OCDD. Some studies ascribed this to the memory effect. Zhong et al. [36] reported that the adsorptive memory effect happened, and PCDD/Fs were released from the surface of the catalysts after the SCR had been run for a certain period. Cai et al. [41] observed memory effect in other MSWI plants, and attributed this to the release of accumulation particulate matter, with PCDD/Fs in the bypass pipeline and inner wall of the apparatus.

3.3. Characterization of Catalyst

The main active compounds were analyzed by XRF, which showed the contents of MoO₃, V₂O₅, and TiO₂ were 6.76%, 3.63% and 78.90%, respectively. It is necessary to mention that an appreciable amount of SiO₂ (4.08%) was found in this catalyst, which was introduced to improve the mechanical and catalytic properties, and enhance the thermal stability [42]. The BET results showed that the catalyst BET surface area was 59.65 m²/g, among which the mesoporous surface took up 55.36 m²/g. The pore volume was 0.25 cm³/g and the average pore diameter was 16.80 nm. This BET surface area is relatively common compared to the reported V₂O₅/TiO₂ and V₂O₅-WO₃/TiO₂ catalysts [11,27]. Nevertheless, the pore volume is relatively higher, meaning PCDD/Fs molecules can easily be retained on the surface [43].

The XPS results revealed the valence of molybdenum and vanadium, as well as 2 species of oxygen: (1) The surface lattice oxygen, noted as O_S-L. (2) The surface absorbed oxygen, noted as O_S-A [44,45,46,47]. As exhibited in Figure 6a, almost all molybdenum exists as Mo⁶⁺ rather than Mo⁴⁺ (97.56% and 2.44%, respectively). The V 2p are presented in Figure 6b; the vanadium species are mainly separated into V³⁺, V⁴⁺, and V⁵⁺, identified by binding energies of 515.29 eV, 515.84 eV, and 517.26 eV [44]. The contents of O_S-L, and O_S-A were presented in Figure 6c, taking up 83.71%, 15.62%, respectively. It is widely reported that the contents of V⁵⁺ (33.28%)provide V=O bonds and V–O support bonds to oxidize pollutants with the interaction of surface oxygen and then reduce to V⁴⁺ (42.77%) species. The V⁴⁺ can be re-oxidized by O₂. The rich content of O_S-A has acted in a critical role in accelerating the transformation of V⁴⁺ and V⁵⁺ [48,49].

The XRD results in Figure 7a exhibit prominent peaks ascribed to anatase TiO₂ (PDF#21-1272). No peaks of V₂O₅ were found, meaning the vanadium dispersed equably, and no amorphous or microcrystalline forms existed. The peak at 45.1° was attributed to silica crystals. The results of H₂-TPR are shown in Figure 7b. Though a considerable loading amount of vanadium (3.63%) was observed in VMT, the redox peaks at 420.2 °C, 466.5 °C, and 713.5 °C corresponding to the formation of V₆O₁₃, VO₂, and later V₂O₃, respectively [49], did not appear. This could be attributed to the relatively high content of V⁴⁺, taking up 42.77%, while V⁵⁺ is at the low level of 33.28%. Also, the redox process of Mo⁶⁺ to Mo⁴⁺ happened at temperatures between 450 °C and 650 °C [50]. The redox peak of Mo⁶⁺ was so intense that it made V⁵⁺ show as a shoulder peak, which means the interaction between MoO₃ and V₂O₅ affected the redox properties of the catalyst conspicuously. All these factors gave good evidence to prove the good redox behavior of the V₂O₅-MoO₃/TiO₂ catalyst in this study.

4. Discussion

The laboratory results revealed that the adsorption and oxidation properties of the catalyst, also showed the de-chlorination effects could significantly influence the RE of toxic PCDD/F. The mass concentration RE of the PCDD/Fs in laboratory conditions at 180 °C (88.5%) was close to that at MSWI plant conditions (89.3%), but the RE of toxic concentration was opposite (89.3% in laboratory and 81.7% in field). Zhan et al. [27] observed that the RE and DE dropped variously in field conditions using several V₂O₅-WO₃/TiO₂ catalysts, due to the negative effect of the high content of water vapor and SO₂. The water vapor was reported to compete for the active vanadium sites with PCDD/F molecules and create a diffusion block of cluster-forming water molecules [18,51]. Moreover, SO₂ could react with the NH₃, forming NH₄HSO₄ and blocking the surface of the catalyst, which was illustrated by Kwon et al. [52]. Chai et al. [37] revealed that HSO₄⁻ adheres to the MoO₃ surface, and could be transformed into SO₃ and H₂O, protecting the acid sites and mitigating the adverse effect of SO₂. The V₂O₅-MoO₃/TiO₂ catalyst used in this study showed good stability in the field condition.

The V₂O₅-MoO₃/TiO₂ catalyst showed different performances in removing PCDD and PCDF under laboratory and field conditions. In the laboratory conditions, the mass concentration RE of PCDD was 89.0% (90.3% for toxic concentration), which was better than that of PCDF (82.9% for mass concentration, and 84.4% for toxic concentration). On the contrary, the RE of PCDD (87.3% for mass concentration, and 79.4% for toxic concentration) was lower than that of PCDF (91.3% for mass concentration, and 84.7% for toxic concentration) under field conditions. This could be attributed to the influence of the complicated components in the flue gas, especially the water vapor. Yu et al. [18,34] reported the inhibition effect of water vapor on PCDD/Fs, among which the PCDDs were more sensitive to the water vapor than PCDF.

5. Conclusions

The low-temperature catalytic performance of a commercial honeycomb-like V₂O₅-MoO₃/TiO₂ catalyst on removing PCDD/Fs was investigated under both industrial and laboratory conditions. Various methods detected the characteristic properties of the catalyst. The study showed that:

(1) The higher pore volume (0.25 cm³/g), rich amount of V⁵⁺, V⁴⁺ species, and surface ensured the PCDD/Fs could be retained on the surface of the V₂O₅-MoO₃/TiO₂ catalyst and later be oxidized.

(2) In laboratory conditions, the mass concentration RE was 59.4% (55.0% for toxicity RE), 88.5% (90.3%), and 78.0% (76.0%) at 160 °C, 180 °C, and 200 °C, respectively. The RE decreased as the temperature increased from 180 °C to 200 °C, due to the de-chlorination effect of highly chlorinated congeners and the receded adsorption ability of low-chlorinated congeners. Its has better performance in removing PCDD/Fs at low temperatures than the V₂O₅/TiO₂ catalyst and V₂O₅-WO₃/TiO₂ catalysts mentioned in the references.

(3) The close RE of toxic concentration results between laboratory (88.5%) and field (81.7%) conditions at 180 °C showed that this V₂O₅-MoO₃/TiO₂ catalyst has good adaptability for MSWI flue gas conditions. In MSWI flue gas, most PCDD/Fs existed in the gas phase. The particulate-phase PCDD/Fs were hard to destroy, and even increased, due to the memory effect. The reasons that the RE decreased to 81.7% in field conditions could be attributed to the competition between water vapor and the interaction of SO₂.

(4) The V₂O₅-MoO₃/TiO₂ catalyst showed better performance in PCDD in laboratory conditions, and favored PCDF removal in the field study. Because the water vapor jeopardized removing PCDD more conspicuously.

More studies will be carried out in the laboratory to further reveal the influence of other compounds in the flue gas, such as NOx, SO₂, and HCl. The kinetic analysis and catalytic mechanism are still on the way.