Kinetics of the catalytic oxidation of toluene over Mn,Cu co-doped Fe₂O₃: Ex situ XANES and EXAFS studies to investigate mechanism

2.5.1 Power-law model

The oxidation rate as a function of toluene and oxygen concentration is described via the Power-law model:

Because C_O2 >> C_e, hence C_O2 can be assumed to be a constant, the equation can be rewritten as follows:

When concentration of O₂ is fixed, the reaction rate of toluene catalytic oxidation is calculated by linear regression fitting. The slope m represents reaction order, whereas rate constant k_p can be derived from the intercept ln(k_p) at each heating temperature point.

2.5.2 Mars-van Krevelen model

MVK model consists of two steps. The first is the reaction of oxidized catalyst sites (OCS) and toluene to form reduced catalyst sites (RCS). Another is the reaction of RCS and O₂ from the gas phase to regenerate OCS for the next cycle. The process can be elucidated by the following reaction equations:

When the combustion reaction of toluene reaches equilibrium, 1 mol of toluene needs 3 mol of oxygen to be completely oxidized (C₇H₈ + 3O₂ → 2CO₂ + 2H₂O), the reaction rate (r) can be combined into the following formula:

Since C_O2 is a constant, k_e and k_O2 can be derived from the slope of a linear regression fitting (1/(−r) versus 1/C_e) for each temperature point.

2.5.3 Langmuir – Hinshelwood model

In Langmuir Hinshelwood model, it is assumed that all catalyst components have the same activation position, where oxygen and reactant molecules are adsorbed (hereafter denoted by C₇H₈*, O*, and O₂*), then the reactions take place as follows:

Oxygen molecule adsorption and oxygen atom adsorption are explained separately:

• (1)

  Reactions involving the adsorption of molecular oxygen:

• (2)

  Reactions involving the adsorption of atomic oxygen:

When the concentration of toluene is fixed, the Eq. (13) is simplified into two linear equations:

a. Molecular oxygen (O₂) adsorption:

A plot of $\sqrt{\frac{C_{O2}}{-r}}$ versus C_O2 provides a linear regression.

b. Atomic oxygen (O) adsorption:

A plot of $\sqrt{\frac{C_{O2}}{-r}}$ versus $\sqrt{C_{O2}}$ provides a linear regression.

2.5.4 Activation energy

The activation energy (E_a, kJ mol⁻¹) was calculated using Arrhenius equation:

3.1.1 Initial concentration of toluene

It is well-known that the operating conditions for catalytic oxidation are changeable and complicated. Hence, exploring the effects of working conditions on toluene conversion is essential to get the highest conversion efficiency of the Mn,Cu-Fe₂O₃ catalyst in practical application. Fig. 1a displays the dependence of conversion rate on the initial concentration of toluene at a constant O₂ concentration (10%) and inlet flow rate (200 L min⁻¹). The conversion rate decreases when the concentration of toluene increases from 55 to 220 ppmv. Specifically, when the heating temperature is in the range of 200–270 °C, the conversion rate is inversely proportional to the concentration of feeding toluene, i.e., the higher the concentration of toluene is, the lower the conversion rate will be (Fig. S1a). This result may be due to the limit of the active surface of Mn,Cu-Fe₂O₃ catalyst for toluene adsorption before oxidation. Because the reactor is a continuous flow, toluene molecules probably compete with each other for the adsorption on Mn,Cu-Fe₂O₃ catalyst surface. In other words, the more C₇H₈ molecules enter the reactor, the higher competition for them to occupy on the catalyst surface, resulting in more C₇H₈ molecules to be kicked out. Hence, it is concluded that the reactant adsorption over the catalyst surface is a rate-determining stage.

Close

Fig. 1

(a) and (b) the effect of initial toluene concentration (C_O2 = 10%, Q = 200 mL min⁻¹) and inlet O₂ concentration (C_C7H8 = 165 ppmv, Q = 200 mL min⁻¹) on the conversion rate at different temperatures, respectively; (c) the effect of flow rate on catalytic activity (C_O2/C_C7H8 ratio = 91); (d) mass balance of C₇H₈ and CO₂ (C_C7H8 = 165 ppmv, C_O2 = 10%, Q = 200 mL min⁻¹).

Export to PPT

It is noteworthy that heating temperature has a supportive effect on the efficiency of C₇H₈ conversion. Remarkably, at 220 °C, the conversion rate of four investigated C₇H₈ concentrations is under 20%. However, when the reactor temperature increases to 270 °C, the conversion rate of C₇H₈ at 55, 110, 165, and 220 ppmv is 97, 95, 90, and 80%, respectively. At 290 °C, C₇H₈ was removed nearly completely at most concentrations. When the heating temperature reaches 300 °C, C₇H₈ was converted completely, which promotes the mass transfer process (Wu et al., 2021).

3.1.2 Inlet concentration of O₂

In terms of reactants, oxygen concentration also plays an indispensable role in the heterogeneous reaction of catalytic system. As illustrated in Fig. 1b, the conversion rate increases significantly when O₂ concentration is in the range of 1–5%, but varies slightly when O₂ concentration is in the range of 5–30%. Apparently, reaction rate changes obviously when oxygen concentration is within 5% but keeps constant when oxygen concentration is in the range of 5–30%. This result suggests that to achieve complete C₇H₈ oxidation, the inlet O₂ should be maintained at a concentration greater than 5% (Fig. S1b).

When the O₂ concentration increase from 1 to 5%, the probability of oxygen adsorption on the metal oxide surface increases to replenish the oxygen vacancies made by consumed lattice oxygen, resulting in enhancement of C₇H₈ oxidation (Hu, 2011). Therefore, once oxygen is supplied with a higher concentration, the oxidation will be accelerated. Previous studies also pointed out that metal oxides still can catalyze the oxidation of C₇H₈ at low concentration or even without oxygen, which is assigned to the effect of lattice oxygen or OH groups on the catalyst surface (Su et al., 2020; Halim et al., 2007). When the supplied O₂ concentration is over 5%, the catalytic oxidation of toluene reaches a certain limit and no longer increases, leading to the unchanged performance of toluene oxidation.

The relationship between reaction rate and reactant concentration is shown in Table S1. Typically, the reaction rate increases when the concentration of O₂ and C₇H₈ increases. According to the principle of L-H model, when the C₇H₈ and O₂ are supplied to reactor with low level, increasing the concentration of C₇H₈ and O₂ will also increase the reaction rate. This phenomenon takes place if the adsorption positions of reactants are different, the reaction rate will increase. In contrast, the same adsorption positions may lead to competitive adsorption among reactants, resulting in the reduction of the reaction rate. In terms of the MVK model, increasing O₂ concentration improves the reaction rate because more activation sites with oxygen vacancies are supplied on the metal oxide surface for toluene coupling and oxidation (Zhao et al., 2019).

3.1.3 Flow rate

The change in the catalytic reaction efficiency under different flow rates was tested at different temperatures, while the ratio between C₇H₈ and O₂ were kept constant. Fig. 1c shows that the reaction efficiency decreases when the flow rate increases. This result can be assigned to the overload of limited active catalyst surface due to the higher content of toluene in the reactor caused by the higher flow rate. Obviously, the catalytic oxidation of toluene over Mn,Cu-Fe₂O₃ is the surface-supported reaction. Hence, the limited surface of Mn,Cu-Fe₂O₃ catalyst cannot accommodate all toluene to be adsorbed on the Mn,Cu-Fe₂O₃ catalyst surface, resulting in efficiency decrease. Moreover, at the flow rate of 200, 300, 400, and 500 mL min⁻¹, the empty-bed residence time (EBRT), a fraction of catalyst volume (12 cm³) and flow rate (cm³/min), is calculated as 3.6, 2.4, 1.8, and 1.4 s, respectively. Decreasing EBRT from 3.6 s to 2.4 s caused a decrease in toluene removal efficiency by 13.2%, demonstrating that shorter residence time is not favorable for toluene to be adsorbed and decomposed on the Mn,Cu-Fe₂O₃ catalyst surface, thereby reducing the reaction efficiency (Brandt et al., 2016). In overall, it is concluded that flow rate is conversely proportional to the conversion rate.

3.1.4 Mass balance

In order to determine the catalytic oxidation conversion of toluene over Mn,Cu-Fe₂O₃ catalyst in the reaction process, the change in concentration of both toluene and CO₂ was measured simultaneously through GC/FID. A converting system was used to confirm if other intermediates are generated during the catalytic reaction. As presented in Fig. 1d, toluene is totally converted to CO₂ without the generation of any intermediate products. Furthermore, the concentration of C₇H₈ and CO₂ are divided by the total carbon content (summary of C₇H₈ and CO₂) to find the individual conversion rate. Fig. S2 illustrates that the conversion of C₇H₈ and production of CO₂ approximately intersects at 50%, indicating that C₇H₈ is completely oxidized to CO₂. Therefore, the relationship between the reactant (C₇H₈) and product (CO₂) suggests that the possible steps affecting the reaction rate are in the following order: C₇H₈ adsorption, O₂ adsorption, C₇H₈ oxidation reaction, and product desorption.

3.2.1 Power-law model

As displayed in Table S2, under constant oxygen concentration (10%), the reaction rate of toluene oxidation over Mn,Cu-Fe₂O₃ catalyst increases as the reaction temperature increases. The linear regression fittings for the P-L model are shown in Fig. 2a, in which the slope represents the reaction order, whereas the intercepts symbolize the rate constant k_p. The Arrhenius fitting graph is exhibited in Fig. 2b with R² = 0.943. The activation energy is calculated to be 164.8 kJ mol⁻¹, and the frequency factor is 11.1 × 10¹² s⁻¹. Since the R² value is low, the PL model is not suggested to be feasible for the catalytic oxidation of C₇H₈ over Mn,Cu-Fe₂O₃ catalyst. In general, the Power-law model mainly assumes a heterogeneous catalytic reaction mode based on homogeneous mode (Pandhare et al., 2018). Therefore, it is less applicable when complex reactions, such as adsorption or multi-molecular reactions, are involved. According to the catalytic reaction data results, the low catalytic reaction is more complicated and may include the mechanism of the reactant or product adsorption.

Close

Fig. 2

(a) and (b) linear fitting of C₇H₈ oxidation for the PL model and corresponded Arrhenius equation, respectively; (c) and (d) linear fitting of C₇H₈ oxidation for the MVK model and corresponded Arrhenius equation, respectively.

Export to PPT

3.2.2 Mars-van Krevelen model

Generally, the mechanism of MVK is explained that the adsorption of one molecule occurs on top of the previous adsorbed one. Specifically, MVK model clarifies a redox reaction cycle of surface lattice oxygen, reactants, and oxygen, focusing on surface redox reactions, i.e., the source of oxygen required for toluene oxidation is lattice oxygen. Fig. 2c illustrates the linearity data fitted by the MVK model at each temperature point. The reaction rate constants derived from the slope for temperature points were then collected and plotted linearity using Arrhenius equation. As shown in Fig. 2d, the results show the R² value is 0.993 and the activation energy E_a was obtained as 59.1 kJ/mol and frequency factor A of 0.41 × 10⁶ s⁻¹. Based on chemical mechanism, the MVK consists of the two stages occurred in the catalytic reaction between the catalyst and C₇H₈ (Wang et al., 2018). The first stage is the interaction of C₇H₈ with the lattice oxygen on the surface of the metal oxide, resulting in the formation of CO₂ and H₂O, which was proved in mass balance discussion part. The followed stage relates to the supplement of lattice oxygen by adsorbed oxygen in the gas phase for the next cycle of reactions.

3.2.3 Langmuir Hinshelwood model

From the result of the MVK model, the oxygen concentration is assumed to have no effect on the catalytic reactions; it can therefore be a constant. However, the L-H model shows that the oxygen concentration has a higher correlation with the catalytic reaction rate than the C₇H₈ concentration. Fig. 3 illustrated the linearity curves with high R² values fitted by L-H model. The K′ and $K^{″}$ values for each line can be derived from the slope and intercept and shown in Table S3. Even though the O₂ concentration is greater than C₇H₈ concentration, its influence on the catalytic oxidation of C₇H₈ should be put in a deep consideration to get more understanding of the oxidation kinetics and mechanism as well. Typically, the interaction of reactants (toluene) and O₂ on the catalyst surface can be described by L-H model. If we assume that C₇H₈ and O₂ adsorb at the same activation position, the progress of the catalytic reaction should then be discussed. In other words, both lattice oxygen and adsorbed oxygen participate in the C₇H₈ oxidation reaction, which also agrees with previous studies (Tang et al., 2018; Gao et al., 2018). The reaction rate in the case of fixed C₇H₈ concentration and altered O₂ concentration is shown in Table S4.

Close

Fig. 3

Linear fitting of C₇H₈ oxidation for the L-H model.

Export to PPT

The result data are then substituted to Eq. (14) and plotted the linear regression at various heating temperatures (Fig. 4a) to show the dependence of reaction rate on the oxygen molecules (O₂). Subsequently, K₁′ and K₁″ at each temperature point can be derived from the intercept and slope of the linear equation (Table S5). When temperature increases, both K₁′ and K₁″ decrease. The obtained data of K', K“, K₁′, and K₁″ are combined with the known values of C_e and C_O2 to compute the value of k, K_e, K_O2 for the O₂ adsorption. Fig. 4b exhibits the rate constant k fitted by the Arrhenius plot and the obtained activation energy of 61.6 kJ mol⁻¹, which is within the range for the catalytic reaction (Tseng and Chu, 2001). Previous studies have also proved that the adsorption of oxygen molecules on the catalyst surface is a vital step during the catalytic process of pollutants oxidation (Tseng and Chu, 2001; Chu et al., 2001).

Close

Fig. 4

(a) and (b) kinetic linear fitting of C₇H₈ oxidation in the case of K₁′, K₁″ by the L-H model and Arrhenius equation, respectively; (c) and (d) kinetic linear fitting of C₇H₈ oxidation in the case of K₂′, K₂″ by the L-H model and Arrhenius equation, respectively.

Export to PPT

Similarly, by substituting the achieved data to Eq. (15), a linear regression at various heating temperatures can be plotted (Fig. 4c) to show the dependence of reaction rate on the oxygen atoms (O). Herein, K₂′ and K₂″ at each temperature point can be derived from the intercept and slope of the linear equation (Table S6). The values of K', K“, K₂′, and K₂″ are combined with the known values of C_e and C_O2, the value of k, K_e, K_O2 for the reaction involved with O adsorption can also be estimated from their definition. Fig. 4d exhibits the Arrhenius equation fitting of rate constant and activation energy of 28.3 kJ mol⁻¹. In both cases, the plot for calculating K′ and K″ shows that the catalytic reaction has a lower correlation with the adsorption of reactants, while K₁′, K₁″ and K₂′, K₂″ plotting show high linearity of the attributable oxygen concentration. Therefore, the L-H model indicates that the main influencing factors of the catalytic reaction have a low correlation with the reactant adsorption but a high correlation with the redox properties of metal oxides. This redox ability may come from the metal itself or be affected by the desorption of the product on the surface.

3.3.1 XANES

Fig. 5a shows that there is almost no change in the valence and structure of the spectrum of Fe before and after the reaction. Therefore, it can be inferred that Fe₂O₃ plays a role as an electron balance, transfer, and structure stability in the catalytic reaction of the toluene oxidation process.

Close

Fig. 5

XANES (a, b,c) and EXAFS (d, e, f) of Fe, Mn, and Cu before and after catalytic oxidation, respectively.

Export to PPT

In Fig. 5b, the XANES spectrum of Mn reduced slightly, from 14.891 eV before the reaction to 13.996 eV after the reaction (set E₀ = 6542 eV), indicating that part of Mn might be reduced from Mn⁴⁺ to Mn³⁺ state (Manceau et al., 2012). This result suggests the establishment of oxygen vacancies to compensate for the charge loss. In Fig. 5c, the spectra of Cu before and after the catalytic reaction are slightly different, indicating that the structure and the valence did not change. Hence, it can be estimated that Cu may also serve as an active center, but the effect is not as good as Mn.

In summary, the XANES results show the reduction of Fe, Mn, and Cu. The activation of Mn⁴⁺ was documented to be more active than noble metal oxides and gives better catalytic performance (Wu et al., 2013). It was also reported that a synergistic combination of copper and manganese provides more active sites than individual ones (Aguilera et al., 2011). Einaga et al. demonstrated the incorporation of Cu into Mn oxides can build up the average oxidation state of Mn on the surface sites higher than bulk sites, resulting in higher activity compared to single-metal oxides (Einaga et al., 2014).

3.3.2 EXAFS

As shown in Fig. 5d, there is almost no difference in Fe structure between before and after the reaction. Therefore, it can be determined that Fe may act as an electron balance or transfer or a carrier in the catalytic reaction. The formation of Fe/CuO_x composite reduces the phase shift of Mn oxide. In Mn K-edge EXAFS (Fig. 5e), the 1–2 Å Mn-O in the first layer is slightly shifted, and the second layer Mn-O-Mn has more obvious changes, indicating that Mn may be the main reaction activation center of the catalyst (Kou et al., 1998). Hence, it can be concluded that the activation center of Mn could be the formation of a complex structure from Mn³⁺ and Mn⁴⁺. The results show that it may be part of the partial reduction process of Mn⁴⁺ to Mn³⁺. The change in Mn-O-Mn structure of the second layer about 2–3 Å might come from the adsorption of reactants and products on the catalyst surface (Yang et al., 2020). This interference signal is generated by Mn-O-C, which superimposes with the original layer.

Accordingly, Mn-O-Mn will cause the change and compensation of the signal and become significant changes after 3 Å. The results of the EXAFS analysis of Cu in Fig. 5f show a slight change in the structure of the catalyst after the catalytic reaction. Hence, the activation center is formed by co-doping of Mn and Cu the composite structure of Mn³⁺ and Mn⁴⁺, resulting in higher catalytic reaction activity. Therefore, the formation of part of the Mn and Cu structures in Mn,Cu-Fe₂O₃ can also improve the reactivity. Li et al. have demonstrated the existence of Cu²⁺/Mn³⁺ ↔ Cu⁺/Mn⁴⁺ in the reaction system (Li et al., 2021). The authors also stated that in the metal oxide catalyst, the reaction is based on the redox of the metals and the interaction between the lattice O and oxygen. The effect of Cu²⁺/Mn³⁺ ↔ Cu⁺/Mn⁴⁺ can reduce the hindrance of the aforementioned reaction, resulting in the improvement of the catalytic efficiency. Besides, the mixed structure of Mn³⁺ and Mn⁴⁺ accelerates the electron transfer and speeds up the catalytic reaction. Therefore, the formation of Mn/Cu structure in Mn,Cu-Fe₂O₃ composite can also improve the reactivity.