Catalytic Ozonation of Toluene over Acidic Surface Transformed Natural Zeolite: A Dual-Site Reaction Mechanism and Kinetic Approach

Abstract

Volatile organic compounds (VOCs) are responsible for damage to health due to their carcinogenic effects. Catalytic ozonation using zeolite appears as a valuable process to eliminate VOCs from industrial emissions at room temperature. For full-scale application of this new abatement technology, an intrinsic reaction rate equation is needed for an effective process design and scale-up. Results obtained here provide a mechanistic approach during the initial stage of catalytic ozonation of toluene using an acidic surface transformed natural zeolite. In particular, the contribution of Lewis and Brønsted acid sites on the surface reaction mechanism and overall kinetic rate are identified through experimental data. The least-squares non-linear regression method allows the rate-determining step to be established, following a Langmuir–Hinshelwood surface reaction approximation. Experimental evidence suggest that ozone is adsorbed and decomposed at Lewis acid sites, forming active atomic oxygen that leads to the oxidation of adsorbed toluene at Brønsted acid sites.

1. Introduction

Volatile organic compounds (VOCs), such as benzene, toluene, ethylbenzene, and xylene (BTEX) are emitted from a wide variety of sources, such as internal combustion engines, power stations, refineries, the food processing industry, the forestry industry, wastewater treatment facilities, and the dry cleaning industry, among other industrial sectors [1,2]. VOCs are also released from indoor sources, including: construction materials, damaged furniture, aerosol sprays, deodorants, air fresheners, cleaning products, disinfectants and insecticides [3]. BTEX are considered priority pollutants due to their toxic, mutagenic, and carcinogenic effects [4], and their role in the formation of secondary atmospheric pollutants such as ozone and airborne particulate matter [5].

During the last decades, great efforts have been made to remove VOCs from both indoor and outdoor emissions, using recovery methods (e.g., absorption, adsorption, membrane separation, and condensation) or destructive methods (e.g., thermal, catalytic, or biological oxidation) [6]. However, the disposal of VOC-saturated solvents or exhausted adsorbents from recovery processes has impaired large-scale implementation. Moreover, conventional destruction methods such as thermal oxidation might generate undesirable or noxious by-products, such as dioxins and carbon monoxide, and must be applied below the lower explosive level (LEL) of target VOCs. In the case of biological oxidation, special attention should be paid to avoid microorganism deactivation [7]. Alternatively, catalytic oxidation is claimed to be the most effective and economically feasible technique for the oxidation of VOCs into less harmful compounds and, eventually, into carbon dioxide and water [8,9]. Catalytic oxidation could be applied to treat waste gas streams in a wide range of VOCs concentrations and flow rates at temperatures around 473 K. This technique can be considered as environmentally friendly due to its low energy demand and lower generation of harmful products, using appropriate catalysts [9,10,11,12].

Recently, a new catalytic process to oxidize VOCs using ozone has attracted increasing interest [13,14,15,16,17,18]. Ozone-based catalytic oxidation allows removing persistent VOCs from industrial emissions at room temperature, reducing operational cost [19,20,21,22]. Synthetic zeolites and metal-supported catalysts have been used as catalysts in the catalytic ozonation of VOCs [19,23,24,25,26,27,28]. Additionally, natural zeolites have started to be applied in combination with ozone in this abatement process [29,30]. Thus, some studies have indicated the positive contribution of modified natural zeolite on the catalytic treatment [31,32]. Even though the influence of surface acid sites of natural zeolites has been assessed in previous studies, there are still some doubts about the reaction mechanism that takes place among ozone and VOC interactions at the zeolite surface and there is also a lack of a kinetic expression to allow process scale-up.

For commercial-scale implementation of catalytic ozonation of VOCs using zeolite at room temperature, accurate kinetic rate expressions are needed to allow reactor design and optimization. In this study, the contribution of Lewis and Brønsted acid sites on the surface reaction mechanism is established and an overall kinetic model, representing the catalytic ozonation of toluene over acidic surface-transformed natural zeolite at room temperature, is proposed.

2. Results and Discussion

2.1. FTIR Evidence of the Heterogeneous Catalytic Ozonation of Toluene

Figure 1 displays FTIR spectra of uncovered acidic transformed natural zeolite sample (dashed line), upon contact with toluene (dotted line), and after catalytic ozonation experiments (solid line). An IR vibration band around 3584 cm⁻¹ is observed in the uncovered and toluene-saturated samples. Such an IR band is related to acidic bridging hydroxyl groups (Si(OH)Al) on natural zeolite, as reported elsewhere [33,34].

The IR spectrum after toluene adsorption shows an IR band at 1470 cm⁻¹ that is linked to adsorbed toluene molecules. In particular, the IR bands around 1450–1540 cm⁻¹ (considering a maximum at 1477) have been associated with the stretching vibrations of C=C bonds of the aromatic ring of adsorbed toluene on zeolites [35]. After catalytic ozonation of toluene, chemical surface intermediaries are identified in the collected spectrum (solid line). A typical absorption band at 1384 cm⁻¹ is registered. This IR band is associated with active oxygen species formed at the zeolite surface [36,37,38,39]. Therefore, it can be related to ozone adsorption and decomposition at Lewis acid sites. It has been suggested that ozone is adsorbed and decomposed into atomic oxygen (O^•) at Lewis acid sites, as represented by Equation (1) [37,38,39,40,41,42,43,44].

$$O_3{}_{ (g)}+s_L\rightleftharpoons O^{•}-s_L+O_2{}_{(g)}$$

(1)

The IR bands at 1057 and 1027 cm⁻¹, as well as the bands around 800 cm⁻¹ and 1168 cm⁻¹, have been ascribed to molecular ozone adsorption, ozonide species (O₃⁻), and superoxide (O₂⁻), respectively [33]. However, no significant differences on these adsorption bands (before and after ozone exposure) can be observed here, probably due to a rapid generation of atomic oxygen (1384 cm⁻¹) at strong Lewis surface acid sites.

Additionally, an IR band at 1743 cm⁻¹ is registered after catalytic ozonation (see solid line in Figure 1). IR bands around 1718–1752 cm⁻¹ have been associated with adsorbed formic acid [45,46]. Additionally, some other distinctive peaks can be seen at 2851, 2920, and 2957 cm⁻¹ after catalytic ozonation. IR bands around 2899 and 2935 cm⁻¹ have been related to the (C–H) stretching vibrations of formate species [47]. Thus, the observed IR bands could be due to the shift of such IR bands associated with adsorbed toluene oxidation by-products. Complementary results obtained using HPLC analyses indicate the presence of formic acid, acetaldehyde, benzoic acid, and benzaldehyde as the main intermediate surface compounds after catalytic ozonation of toluene [48]. Furthermore, different studies of catalytic ozonation of toluene using different catalysts have shown that ozone is decomposed into active oxygen species, interacting with adsorbed toluene [24,49,50]. Therefore, it has been indicated that toluene is oxidized via methyl interaction. The lower dissociation energy of the methyl C–H bond (3.7 eV) could support that statement. This value is lower than the dissociation energies of other C–H bonds of toluene molecule (C–H bond of aromatic ring 4.3 eV; C–C bond of methyl 4.4 eV; C–C bond of aromatic ring 5.0–5.3 eV; C=C bond of the ring 5.5 eV), which justifies the schematic reaction route displayed in Figure 2. Thus, benzaldehyde, benzoic acid, and formic acid have been identified as primary reaction intermediates of catalytic ozonation of toluene in agreement with the results obtained in this work [15,48,51,52].

IR evidence obtained here suggests that a surface reaction occurs between adsorbed toluene at Brønsted acid sites and active oxygen species generated after ozone adsorption and decomposition at Lewis acid sites. Such a surface reaction mechanism is in agreement with previous reports, that claim the adsorption of toluene molecules at Brønsted acid sites and the adsorption of ozone molecules at Lewis acid sites followed by its transformation into active oxygen species [32,48,55,56,57]. After that, it has been indicated that active oxygen species interact with adsorbed toluene molecules, leading to the formation of toluene oxidation by-products, as reported elsewhere [58,59]. Previously published works confirm the surface reaction mechanism proposed here, using not only natural zeolite mainly composed of clinoptilolite and mordenite and modified natural zeolite [32,48]; but also using synthetic zeolites such as faujasite, mordenite, and ZSM-5 [55,56,57] under the same experimental conditions.

2.2. Kinetic Approaches of Catalytic Ozonation of Toluene

In the absence of mass transfer limitations, the heterogeneous catalytic ozonation of toluene using acidic transformed natural zeolite can be represented by a general and simplified reaction, as follows [58,59]:

$$C_7{H}_8+6O_3\underset{}{\stackrel{cat}{\to }}7C{O}_2+4H_{2}O$$

(2)

Firstly, a power law kinetic model is used here as an initial approximation for the estimation of the overall kinetic rate expression of catalytic ozonation of toluene, $r_{Tol}$ (μmol·dm⁻³ s⁻¹):

$$-r_{Tol}=k_{app}{C}_{{}_{C_7{H}_8}}^{\alpha }\cdot C_{O_3}^{\beta }$$

(3)

where $k_{app}$ is the apparent reaction rate constant ((dm³∙μmol)^1+α+β∙s⁻¹); $C_{{}_{C_7{H}_8}}^{}$ and $C_{O_3}^{}$ stand for toluene and ozone concentrations (µmol·dm⁻³), respectively; α and β are orders of reaction related to toluene and ozone, respectively.

Table 1. Power law kinetic parameters for catalytic ozonation of toluene.

Catalysts	Reaction Temperature (K)	$k_{app}$ × 103 (dm3∙μmol)1+α+β∙s−1)	EA app (kJ·mol−1)	α	β	R2	Reference
AZ	293	5.07	17.21	0.66	0.82	0.97	This study
MnO2/graphene	295	3.25	29.3	0.63	0.55	0.91	[62]
MnOx/γ-alumina	343	n.r	31	−1	2	n.r	[63]
MnOx/γ-alumina	298	n.r	33	0.18	0.56	0.98	[64]

n.r: non-reported.

lists such kinetic parameters obtained after applying the NLR technique, using the tools of the Microsoft Excel Solver add-in [60]. Results are compared with others reported in the literature for catalytic ozonation of toluene using different catalysts.

As can be seen, the order of reaction obtained for toluene is lower than the order of reaction for ozone, which is evidence of the higher ozone dependence during catalytic ozonation of toluene promoted by acidic transformed natural zeolite. Moreover, the obtained apparent activation energy (17.21 kJ·mol⁻¹) is lower than the reported activation energy in the toluene–ozone homogenous reaction (55.5 kJ·mol⁻¹) [61]. Such results clearly indicate that active surface sites of acidic transformed natural zeolite play a significant role during the heterogeneous reaction.

Secondly, a kinetic expression that takes into account the contribution of acidic surface sites of transformed natural zeolite during the catalytic ozonation of toluene is postulated following the dual-site Langmuir–Hinshelwood (dsL-H) surface reaction approximation and is validated using the experimental data collected during this study. It is assumed that all chemical species react after their adsorption at acidic surface sites, and each active site is occupied by only one reactive specie [65,66]. Thus, previous experimental results can be taken as shreds of evidence of the existence of two main active acidic surface sites where selective adsorption of toluene and ozone occur. Toluene has proved its affinity toward Brønsted acid sites and ozone for Lewis acid sites [32,67,68]. Hence, the surface reaction mechanism could be described by three elementary steps, including a dual-site surface reaction as the rate-determining step (RDS), as follows:

1. Ozone adsorption and decomposition at Lewis acid sites (S_L) could be described by a simplified adsorption-desorption rate equation, $r_1$(μmol·g⁻¹·s⁻¹), as follows:

$$\begin{gathered} O_3{}_{ (g)}+s_L\underset{k_{-1}}{\overset{k_1}{\rightleftharpoons }}{O}^{•}-s_L+O_2{}_{(g)} \\ {r}_1=k_1{C}_{O_3}\left[s_L\right]-k_{-1}\left[O^{•}-s_L\right]C_{O_2} \end{gathered}$$

(4)

where $\left[s_L\right]$ and $\left[O^{•}-s_L\right]$ represent the total vacant Lewis acid sites of acidic transformed natural zeolite or the total surface concentration of Lewis acid sites (μmol·g⁻¹), and active atomic oxygen specie adsorbed at Lewis acid sites (μmol·g⁻¹), respectively. k₁ and k₋₁ represent forward and reverse rate constants of active atomic oxygen surface specie formation (dm³·s⁻¹·μmol⁻¹), $C_{O_3}$ is the concentration of ozone (μmol·dm⁻³) and $C_{O_2}$ is the concentration of oxygen (μmol·dm⁻³).

2. Toluene adsorption at Brønsted acid sites (S_B) could be represented by a simplified adsorption-desorption rate equation, $r_2$(μmol·g⁻¹·s⁻¹), as follows:

$$\begin{gathered} C_7{H}_8{}_{(g)}+s_B\underset{k_{-2}}{\overset{k_2}{\rightleftharpoons }}{C}_7{H}_8-s_B \\ {r}_2=k_2{C}_{C_7{H}_8}\left[s_B\right]-k_{-2}\left[C_7{H}_8-s_B\right] \end{gathered}$$

(5)

where $\left[s_B\right]$ and $\left[C_7{H}_8-s_B\right]$ represent the total vacant Brønsted acid sites of acidic transformed natural zeolite or the total surface concentration of Brønsted acid sites (μmol·g⁻¹), and toluene-adsorbed specie at Brønsted acid sites (μmol·g⁻¹), respectively. k₂ represents forward rate constant of toluene surface complex formation (dm³·s⁻¹·μmol⁻¹), k₋₂ represents the reverse rate constant of toluene surface complex formation (s⁻¹), $C_{C_7{H}_8}$ is the concentration of toluene (μmol·dm⁻³).

3. Dual-site surface reaction (rate-determining step) $r_3$(μmol·g⁻¹·s⁻¹), between adsorbed toluene molecules at Brønsted acid sites and active oxygen species adsorbed at Lewis acid sites, leading to the formation of toluene oxidation by-products, as follows:

$$\begin{gathered} C_7{H}_8-s_B+O^{•}-s_L\stackrel{k_3}{\to }\hspace{0.17em}{C}_7{H}_{6}O-s_B+s_L \\ -r_3=k_3\left[O^{•}-s_L\right]\left[C_7{H}_8-s_B\right] \end{gathered}$$

(6)

where $\left[C_7{H}_8-s_B\right]$ and $\left[O^{•}-s_L\right]$ stand for the amount of toluene adsorbed at Brønsted acid sites (μmol·g⁻¹), and the amount of active oxygen species adsorbed at Lewis acid sites (μmol·g⁻¹), respectively. k₃ is the disappearance rate constant of the adsorbed toluene surface complex (g·s⁻¹·μmol⁻¹).

The unknown surface concentrations of vacant sites of Lewis [S_L] and Brønsted [S_B] acidic sites can be calculated from site balances, as follows:

$$C_{L_t}=\left[s_L\right]+\left[O^{•}-s_L\right]$$

(7)

$$C_{B_t}=\left[s_B\right]+\left[C_7{H}_8-s_B\right]$$

(8)

where C_Lt and C_Bt are the total site concentration of Lewis acid sites (μmol·g⁻¹) and Brønsted acid sites (μmol·g⁻¹), respectively; $\left[O^{•}-s_L\right]$ and $\left[C_7{H}_8-s_B\right]$ represent adsorbed active oxygen specie (μmol·g⁻¹) and adsorbed toluene molecule (μmol·g⁻¹) at Lewis and Brønsted acid sites, respectively.

Here, use is made of the steady-state approximation for the rate of ozone adsorption and its decomposition into active atomic oxygen and for the rate of toluene adsorption. This assumes that $r_1\simeq 0$ and $r_2\simeq 0$ or the right-hand sides of Equations (4) and (5) are in a pseudo-equilibrium or stationary state.

Thus, the unknown concentration of active atomic oxygen specie adsorbed at Lewis acid sites can be estimated from Equation (4), assuming $r_1\simeq 0$, as follows:

$$\left[O^{•}-s_L\right]=\frac{k_1{C}_{O_3}\left[s_L\right]}{k_{-1}{C}_{O_2}}=K_1\frac{C_{O_3}\left[s_L\right]}{C_{O_2}}$$

(9)

Since the inlet concentration of oxygen in all experiments was kept in excess, the oxygen concentration can be considered to remain constant. Thus, a new constant can be set as indicated by Equation (10):

$$K_1^{’}=\frac{K_1}{C_{O_2}}$$

(10)

Then, the unknown concentration of adsorbed active oxygen species can be finally obtained, as follows:

$$\left[O^{•}-s_L\right]=K_1^{’}{C}_{O_3}\left[s_L\right]$$

(11)

Substituting Equation (11) into Equation (7) yields Equation (13):

$$C_{L_t}=\left[s_L\right]+K_1^{’}{C}_{O_3}\left[s_L\right]=\left[s_L\right]\left(1+K_1^{’}{C}_{O_3}\right)$$

(12)

$$\left[s_L\right]=\frac{C_{L_t}}{(1+K_1^{’}{C}_{O_3})}$$

(13)

In the case of the unknown concentration of adsorbed toluene molecules at Brønsted acid sites, it can be estimated from Equation (5) at a pseudo-equilibrium condition $\left(r_2\simeq 0\right)$, leading to:

$$\left[C_7{H}_8-s_B\right]=\frac{k_2}{k_{-2}}{C}_{C_7{H}_8}\left[s_B\right]=K_2{C}_{C_7{H}_8}\left[s_B\right]$$

(14)

Then, Equation (14) can be combined with Equation (8) to obtain Equation (16):

$$C_{B_t}=\left[s_B\right]+K_2{C}_{C_7{H}_8}\left[s_B\right]=\left[s_B\right]\left(1+K_2{C}_{C_7{H}_8}\right)$$

(15)

$$\left[s_B\right]=\frac{C_{B_t}}{\left(1+K_2{C}_{C_7{H}_8}\right)}$$

(16)

After substituting Equations (11) and (14) in Equation (6) and then incorporating Equations (13) and (16), this yields Equation (18):

$$-r_3=k_3\left[C_7{H}_8-s_B\right]\left[O^{•}-s_L\right]=k_3{K}_2{C}_{C_7{H}_8}\left[s_B\right]K_1^{’}{C}_{O_3}\left[s_L\right]$$

(17)

$$-r_3=K_1^{’}{K}_2{k}_3{C}_{L_t}{C}_{B_t}\frac{C_{O_3}{C}_{C_7{H}_8}}{\left(1+K_1^{’}{C}_{O_3}\right)\left(1+K_2{C}_{C_7{H}_8}\right)}$$

(18)

In Equation (18) the values of $K_2$ and $C_{B_t}$ can be obtained from adsorption experiments since $K_2$ is the Langmuir adsorption equilibrium constant of toluene surface complex formation at Brønsted acid sites (dm³·μmol⁻¹) and $C_{B_t}$ is the maximum toluene uptake, corresponding to the saturation of Brønsted acid sites or monolayer coverage (μmol·g⁻¹). Such equilibrium parameters come by substituting Equation (8) into Equation (14), resulting in Equation (19), as follows:

$$\left[C_7{H}_8-s_B\right]=\frac{K_2{C}_{C_7{H}_8}{C}_{B_t}}{1+K_2{C}_{C_7{H}_8}}$$

(19)

where $C_{C_7{H}_8}$ is the concentration of toluene at the equilibrium (μmol·dm⁻³), $\left[C_7{H}_8-s_B\right]$ is the amount of toluene adsorbed at Brønsted acid sites (μmol·g⁻¹) at equilibrium, $C_{B_t}$ is the maximum toluene uptake, corresponding to the saturation of Brønsted acid sites or monolayer coverage (μmol·g⁻¹). Thus, equilibrium parameters are gathered experimentally from a set of breakthrough curves of q-concentration data obtained using Equation (26).

Figure 3 depicts the adsorption equilibrium data of toluene at 293 K over acidic transformed natural zeolite, computed from breakthrough curves using Equation (26). As can be seen, the Langmuir adsorption isotherm adjusts very well to the experimental data (R² = 0.9959). The maximum toluene uptake by Brønsted acid sites, $C_{B_t}$ (0.23429 μmol·g⁻¹), and the adsorption equilibrium constant, $K_2$ (3.81 dm³·μmol⁻¹) were obtained from the intercept and the slope of the linearized form of Langmuir isotherm expression, respectively. Such equilibrium parameters agree with values found in previous works using natural and modified natural zeolite mainly composed by clinoptilolite and mordenite [39].

If an apparent constant $k_{app}^{’}$ is defined as:

$$k_{app}^{’}=K_1^{’}{K}_2{k}_3{C}_{L_t}{C}_{B_t}$$

(20)

Then, the overall kinetic expression can be rewritten, as follows:

$$-r_3=k_{app}^{’}\frac{C_{O_3}{C}_{C_7{H}_8}}{\left(1+K_1^{’}{C}_{O_3}\right)\left(1+K_2{C}_{C_7{H}_8}\right)}$$

(21)

Hence, substituting the obtained experimental equilibrium constant of toluene adsorption into Equation (21), gives:

$$-r_3=k_{app}^{’}\frac{C_{O_3}{C}_{C_7{H}_8}}{\left(1+K_1^{’}{C}_{O_3}\right)\left(1+3.81C_{C_7{H}_8}\right)}$$

(22)

Table 2. Experimental and calculated reaction rate values.

Runs	Inlet Concentration (µmol·dm−3)		Conversion of Toluene (%)	−r3 (µmol·g−1·s−1)
	Toluene	Ozone		Experimental	Estimated by the Power Law Model	Estimated by the dsL-H Model
1	42.36	3.57	11.0	0.207	0.1769	0.1810
2	42.36	6.48	15.7	0.295	0.2876	0.2871
3	42.36	6.48	14.4	0.272	0.2876	0.2871
4	42.36	1.92	5.6	0.106	0.1054	0.1060
5	42.36	1.12	3.0	0.057	0.0676	0.0646
6	42.36	1.12	3.2	0.061	0.0676	0.0646
7	42.86	1.92	4.3	0.081	0.1063	0.1068
8	33.99	3.57	10.6	0.160	0.1519	0.1562
9	33.85	6.43	16.8	0.253	0.2461	0.2456
10	33.96	1.92	5.4	0.081	0.0910	0.0914
11	50.99	3.57	9.3	0.212	0.1990	0.2032
12	42.71	3.57	10.4	0.198	0.1768	0.1820
13	42.36	1.79	4.3	0.081	0.0995	0.0995
14	42.56	7.59	16.4	0.311	0.3287	0.3218

summarizes experimental and estimated reaction rate values of catalytic ozonation of toluene using acidic transformed natural zeolite. During a set of experimental runs, the inlet concentration ratio of toluene/ozone was varied from 5.3 to 37.8. The NLR technique was conducted to obtain model parameters of Equation (22), using the tools of the Microsoft Excel Solver add-in [60].

Finally, the proposed overall kinetic rate expression of catalytic ozonation of toluene at room temperature, using acidic transformed natural zeolite, is represented by Equation (23):

$$-r_3=0.233\frac{C_{O_3}{C}_{C_7{H}_8}}{\left(1+0.063C_{O_3}\right)\left(1+3.81C_{C_7{H}_8}\right)}$$

(23)

where the constants $k_{app}^{’}$, $K_1^{’}$, and $K_2$ take the values of 0.233 (dm³)²∙µmol⁻¹∙g⁻¹∙s⁻¹, 0.063 dm³·µmol⁻¹, and 3.81 dm³·μmol⁻¹, respectively.

Figure 4 shows a visual comparison of the obtained experimental and estimated reaction rate values of catalytic ozonation of toluene. An excellent coefficient of determination (R² = 0.97) is obtained. A value of R² = 0.91 was reported previously by Hu et al. [62] using the dsL-H reaction approach in the catalytic ozonation of toluene over MnO₂/grapheme, suggesting that two active sites are involved in the reaction mechanism.

On the one hand, results obtained in this study provide evidence that the dsL-H reaction model shows a higher dependence of ozone toward Lewis acid sites during the catalytic ozonation of toluene. Such a result is in agreement with that obtained with the applied power law kinetic model. Our previous work [32], reported the influence of Lewis acid sites on the toluene ozonation reaction rate, using natural and modified zeolites. Experimental results evidenced a linear dependency (slope = 0.05) of the registered reaction rate and Lewis acid sites. Consequently, a higher surface concentration of atomic oxygen enhances catalytic ozonation of toluene, in concordance with experimental findings reported before [36,37,39,42,43,44]. On the other hand, the influence of Brønsted strong acid sites density over toluene ozonation reaction rate also showed a linear dependency, resulting in a lower contribution (slope = 0.01). Former studies reported the affinity of toluene for Brønsted acid sites [69,70]. However, this interaction might lower the reaction rate due to strong toluene adsorption at Brønsted acid sites. Therefore, surface acidic sites of the modified natural zeolite framework reveal a remarkable contribution to toluene ozonation at room temperature.

3. Materials and Methods

3.1. Materials

Chilean natural zeolite mainly composed of clinoptilolite (53%), mordenite (40%) and quartz (7%) was acquired from Minera FORMAS™, Santiago, RM, Chile. As-received zeolite sample was ground and sieved. Particle sizes in the range 0.3–0.425 mm were used in the whole study and they were washed with de-ionized water, oven dried at 398 K and stored in a desiccator. This zeolite exhibits an apparent density, 2.3 g·cm⁻³; BET surface area, 205 m² g⁻¹; total pore volume, 0.11 cm³ g⁻¹; cation exchange capacity, 2.05 meq·g⁻¹; with a Si/Al of 5.34; and it is mainly made up of SiO₂ (75.25%) and Al₂O₃ (14.1%) followed by CaO (4.57%), Fe₂O₃ (2.31%), Na₂O (1.89%), K₂O (0.74%), MgO (0.66%), TiO₂ (0.42%), and MnO (0.05%); with a concentration of Brønsted and Lewis acid sites of 0.4 μmol·g⁻¹ and 31.1 μmol·g⁻¹, respectively [32]. Natural zeolite was used as a parent material to produce a zeolite sample with a higher content of acidic sites.

Ozone was generated in situ from instrumental dry air provided by AGA using an AZCOZON A-4 ozone generator (Vancouver, BC, Canada). A contaminated stream made of toluene vapors was used here as a target VOC. The stream enriched with toluene vapors was produced by continuously bubbling dry argon into a pure liquid toluene (confined in a flask at a controlled temperature of 261 K). Toluene was supplied by Merck (Darmstadt, Hessen, Germany) with a purity of 99.8%.

3.2. Transformation of As-Received Natural Zeolite into a Modified Zeolite Rich in Acidic Sites

Natural zeolite was modified by ion-exchange treatment using an ammonium sulphate solution (0.1 mol·dm⁻³) with a volume/solid ratio of 10/1 at 363 K during 3 h. Then, it was rinsed with ultra-pure water for 4 h, replacing water after 2 and 3 h of contact time. After that, a second ion-exchange using the ammonium sulphate solution was applied using the same procedure and steps as described above. This modification method decreases the content of compensating cations, being ion-exchanged by ammonium cations. Then, ammonium-exchanged zeolite sample was oven-dried at 378 K for 24 h and stored in a desiccator until further use. Finally, a modified zeolite sample with a higher content of acid sites was produced just before the catalytic ozonation experiments by thermal outgassing at 823 K and it is identified here as AZ.

Nitrogen adsorption-desorption isotherms at 77 K, X-ray powder diffraction (XRD), X-ray fluorescence (XRF), scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), and transmission electron microscopy (TEM) were conducted to characterize the transformed natural zeolite sample. Surface area (S_BET) was obtained from nitrogen adsorption-desorption isotherms measured in a Micromeritics Gemini 3175 sorptometer (Norcross, GA, USA). The crystalline structure was registered by XRD analysis in a Bruker AXS Model D4 ENDEAVOR diffractometer (Billerica, MA, USA); the presence of main compensating cations was determined by X-ray fluorescence, using a RIGAKU Model 3072 spectrometer (Akishima, Tokyo Metropolis, Japan). SEM studies were carried out in a JEOL (Akishima, Tokyo Metropolis, Japan) JSM-6380 microscope, operating at 20 kV. Samples were ground to 150 µm and a small amount was deposited in the apparatus. Chemical composition was measured by EDS coupled to the SEM microscope. TEM images were obtained with a JEOL (Akishima, Tokyo Metropolis, Japan), JEM 1200 EX-II device equipped with a Gatan 782 camera for Electron Microscope Erlangshen ES 500. Samples with a particle size of 150 µm were deposited over a mesh previously covered by carbon. Analyses were carried out using an accelerating voltage of 120 kV. Additionally, the nature and strength of acidic sites were investigated by in situ Fourier transform infrared spectroscopy (FTIR), using pyridine as a probe molecule in adsorption and desorption assays followed by FTIR analysis (Py-FTIR), according to a procedure reported previously [32]. More details about characterization procedures can be found in other publications [67,71,72,73,74].

The XRD diffraction pattern of transformed natural zeolite showed no significant changes with respect to as-received natural zeolite, keeping the same crystalline structure after the applied ion-exchange treatments with ammonium sulphate (see Figure 5). Characteristic peaks of clinoptilolite (C), mordenite (M) and quartz (Q) are noticed with equal intensity in both difractograms and were identified according to JCPDS 39–183, JCPDS 29–1257 and JCPDS 461045, respectively. However, XRF analysis revealed a reduction in the amount of compensating cations without any effect on the Si/Al ratio (5.34). Thus, the modified zeolite sample is mainly composed by SiO₂ (79.26%) and Al₂O₃ (14.85%) followed by CaO (1.82%), Fe₂O₃ (2.53%), Na₂O (0.26%), K₂O (0.39%), MgO (0.37%), TiO₂ (0.47%), and MnO (0.05%).

Table 3. Textural and acidic content of acidic transformed natural zeolite.

SBET a (m2∙g−1)	Brønsted b Acid Sites (µmol∙g−1)	Lewis b Acid Sites (µmol∙g−1)	Total Acidity (µmol∙g−1)
261	179.8	282.8	462.6

^a Obtained from nitrogen adsorption-desorption isotherm. ^b Determined by in situ Py-FTIR analyses, quantifying the remaining adsorbed pyridine after heating the saturated-samples up to 823 K.

lists the main physical-chemical characteristics of the acidic transformed natural zeolite sample. As expected, the surface area and the content of acidic sites in the form of both Brønsted and Lewis acid sites were increased with the applied modification treatment.

SEM-EDS analyses (see Figure 6) corroborated the results obtained by XRD and XRF. A reduction in the content of compensating cations could be noticed in the as-received zeolite sample after the applied ion-exchange modification procedures.

EDS results shown in Figure 6 indicate a 49.1% and a 40.1% of decrease in the amount of Ca and Fe cations, respectively. Also, the Si/Al ratio obtained with this technique remained almost constant, confirming that the zeolite structure was not significantly changed. Based on previous experimental results [32,67,68,71], a lower content of compensating cations has been related to a better ozone and toluene diffusion and accessibility to active surface sites of the zeolite framework, suggesting a better performance of acidic transformed natural zeolite during catalytic ozonation of VOCs.

TEM images of natural zeolite and acidic transformed natural zeolite samples are presented in Figure 7A,B, respectively. As can be seen, no substantial changes in the structure and morphology of zeolite samples are observed.

3.3. Experimental Description of Catalytic Ozonation Studies

Catalytic ozonation experiments were conducted at 293 K, using a U-type quartz packed-bed reactor (see Figure 8). Experimental conditions were set after preliminary experiments, following the Madon–Boudart criterion to assure the absence of mass transfer limitations [66]. Briefly, a mass of 150 mg of zeolite sample mixed with quartz particles of the same size with a mass ratio of 1:4 (d_p ≈ 0.36 mm) was loaded and thermally outgassed (before reaction) at a heating rate of 1 °C·min⁻¹ under argon flow (100 cm³·min⁻¹) until reaching a temperature of 823 K. Then, isothermal conditions were maintained for two hours, before being cooled down to room temperature. After that, at a temperature of 293 K, a total gas flow of 100 cm³·min⁻¹ containing toluene and ozone was supplied over the reactor. Toluene and ozone concentrations were varied during the kinetic study, within ranges limited by physical chemical equilibria after dilutions of saturated streams and inert fresh streams. The concentrations of toluene and ozone were monitored on-line using a Perkin Elmer Clarus 500 chromatograph (Perkin Elmer, Waltham, MA, USA) and a BMT 964 ozone analyzer (BMT Messtechnik GmbH, Stahnsdorf, Germany), respectively. More details about experimental conditions and procedures can be found elsewhere [32]. All trials were performed in the absence of mass transfer limitations.

The reactor was operated in a differential mode. Thus, from a series of experimental runs, a set of rate-concentration data was obtained using the plug flow performance equation [65], as follows:

$$\frac{W}{F_{To{l}_{in}}}=\frac{X_{Tol}}{{\left(-r_{Tol}\right)}_{exp}}$$

(24)

where $F_{To{l}_{in}}$ represents the feed rate of toluene (µmol·s⁻¹), ${\left(-r_{Tol}\right)}_{exp}$ is the rate of catalytic oxidation reaction of toluene by ozone (µmol·g⁻¹·s⁻¹), W is the mass of zeolite sample (g), and $X_{Tol}$ is the fraction of toluene converted.

In order to obtain a kinetic reaction model that represents the catalytic ozonation of toluene promoted by transformed natural zeolite at room temperature, firstly, a power law kinetic model was assessed as an initial approximation. Secondly, a set of several elementary steps were proposed with individual rates and an overall kinetic rate expression was postulated, considering the rate-determining step (RDS) by taking into account the experimental findings obtained in this study and those reported previously [32]. In both cases, the proposed overall rate expressions were validated with the experimental data using the least-squares non-linear regression method (NLR). Parameters of the proposed kinetic expressions were gathered by minimizing the sum of the squares (σ²) of the difference of obtained reaction rates for each experimental run, using the solver tool of Microsoft Excel [60] implemented through Equation (25).

$$s^2={\displaystyle \sum _{i=1}^{14}\left[{\left({\left(-r_{Tol}\right)}_{exp}-{\left(-r_{Tol}\right)}_{mod}\right)}_i^2\right]}$$

(25)

where ${\left(-r_{Tol}\right)}_{exp}$ and ${\left(-r_{Tol}\right)}_{mod}$ stand for the rate of catalytic ozonation of toluene (µmol·g⁻¹·s⁻¹), obtained experimentally and using the proposed kinetic model, respectively.

Additionally, Langmuir adsorption equilibrium parameters were obtained from dynamic adsorption experiments conducted at 293 K, using the same experimental set-up without the inlet of ozone. Prior to adsorption experiments, zeolite samples were thermally outgassed as described before. The adsorbed amounts of toluene over acidic transformed natural zeolite at equilibrium (μmol·g⁻¹) for different inlet concentrations of toluene, were calculated from breakthrough curves using Equation (26), as reported elsewhere [68].

$$q=\hspace{0.17em}\frac{1\hspace{0.17em}}{\hspace{0.17em}W\hspace{0.17em}}{\displaystyle \underset{0}{\overset{t_s}{\int }}\left(\hspace{0.17em}{F}_{To{l}_{in}}-F_{To{l}_t}\right)dt}$$

(26)

where q is the adsorbed amount of toluene over acidic transformed natural zeolite at equilibrium (μmol·g⁻¹). W is the zeolite mass (g), and t_s is the time needed to reach the zeolite saturation (s). $F_{To{l}_{in}}$ and $F_{To{l}_t}$ are the feed rate of toluene (μmol·s⁻¹), at the adsorber inlet and outlet streams as a function of time, respectively.

3.4. In Situ FTIR Study of Surface Interactions between Zeolite Sample and Toluene Oxidation By-Products

Oxidation by-products that remained adsorbed after catalytic ozonation of toluene were identified by FTIR spectroscopy in a Bruker Mod Tensor 27 spectrometer (Billerica, MA, USA). IR spectra were collected at an average of 100 scans with a resolution of 4 cm⁻¹ using a self-pressed disc. Additionally, IR spectra before and after toluene adsorption were also recorded.

4. Conclusions

Results obtained in this study provide new insights on the surface reaction mechanism during the catalytic ozonation of toluene using acidic transformed natural zeolite. Experimental evidence confirmed that surface interactions take place between adsorbed toluene at Brønsted acid sites with active atomic oxygen, coming from ozone adsorption and decomposition at Lewis acid sites, leading to toluene oxidation. Experimental data proved that catalytic ozonation of toluene using acidic transformed natural zeolite is very well represented by the dual-site Langmuir–Hinshelwood (dsL-H) kinetic approximation where the rate-determining step results to be the surface reaction between adsorbed species on adjacent active sites. This kinetic expression can be used in the design and optimization of this new process technology. Acidic transformed natural zeolite could be applied as an effective and low-cost catalyst for the catalytic oxidation of VOCs using ozone at room temperature in a novel air pollution control technology. Further research using metal-supported modified natural zeolites in toluene oxidation is under way to complement information provided here in order to facilitate large scale implementation.