Unravel the Roles of the Acid Sites in Different Pore Channels of HZSM-5 Catalyst on Ethanol Conversion to Light Olefin

Abstract

Catalytic conversion of bioethanol is a promising production method for preparing light olefin. However, the role of acid sites in different pore channels of HZSM-5 catalyst is not clear. The roles of acid sites in different channels of HZSM-5 catalyst on the conversion of ethanol to ethylene and propylene was investigated by density functional theory (DFT). The results show that the conversion of ethanol to ethylene mainly occurs at the acid site of the sinusoidal channel (T11) of HZSM-5, and the conversion of ethanol to propylene mainly occurs at the acid site of the straight channel (T10) of HZSM-5 catalyst. The adsorption and diffusion behaviors of ethylene and propylene in straight and sinusoidal channels of HZSM-5 were simulated by the molecular dynamics method. The results show that for the adsorption of ethylene and propylene, the acid sites of sinusoidal channel (T11) with SiO₂/Al₂O₃ = 128 is more conducive to improving the selectivity of ethylene, and the acid sites of straight channel (T10) with SiO₂/Al₂O₃ = 128 is more conducive to improving the propylene selectivity. For the diffusion of ethylene and propylene, the acid sites in the straight channel (T10) of HZSM-5 (SiO₂/Al₂O₃ = 128) are more beneficial to improve propylene selectivity.

1. Introduction

The light olefins mainly include ethylene and propylene, which are important raw materials for chemical production. With the rapid development of the economy and industry, the demand for light olefins is increasing rapidly. In addition, some catalytic cracking (FCC) processes specially used for propylene production have been developed and industrialized [1,2]. When ZSM-5 is mixed with FCC catalyst, olefins in FCC gasoline fractions are converted into light olefins containing propylene [3,4,5]. Ali et al. [6] discovered that the distributions of acidic sites and ZSM-5 morphology significantly altered local residence time and resulted in different selectivity to aromatics and heavier hydrocarbons during gas-phase CO₂ conversion, which provided an efficient approach for direct CO₂ conversion to aromatics. Depending on the choice of the metal, metal oxide, and zeolite, light olefins, long-chain linear a-olefins, long-chain paraffins, aromatics, alcohols, and acids can be selectively produced [7]. Iwamoto et al. [8] developed a Ni-M41 catalyst for converting ethanol into ethylene, which reached the maximum ethylene yield at 673 K and 2000–3000 h⁻¹.

The adjustment of aluminum in HZSM-5 catalyst has also attracted extensive attention of researchers. Sazama et al. [9] found that for the same SiO₂/Al₂O₃ ratio and crystal size, Al pair is more conducive to oligomerization and hydrogen transfer reaction of 1-butene, while single Al is more conducive to cracking of 1-butene. Structure directing- agent (SDA) and sodium cation have a certain influence on the position of Al atom in the framework of ZSM-5 synthesis. Yokoi et al. [10] found that Al atoms on H-ZSM-5 synthesized by using tetrapropylammonium (TPA) cation structure directing-agent (SDA) were mainly located at the channel intersections in the absence of Na cations. Janda et al. [11] adjusted the distribution of aluminum in HZSM-5 by increasing the aluminum content, so that most of the aluminum was located at the intersection of channels, which was beneficial to dehydrogenation rather than cracking of n-butane. Liang et al. [12] found that HZSM-5 catalyst prepared with ethyl orthosilicate as a silicon source has more acid sites at channel intersections and higher selectivity for ethylene and aromatic hydrocarbons; HZSM-5 catalyst prepared with silica sol as a silicon source has more acid sites with straight channels and sinusoidal channels and has higher selectivity for propylene and higher olefins. Kim et al. [13] successfully prepared HZSM-5 molecular sieves with different distributions of Al atoms in the framework by adjusting the crystallization temperature. Wang et al. [14] studied the product distribution of methanol to olefins (MTO) on HZSM-5 by DFT calculation and found that the product distribution based on aromatic hydrocarbon and olefin ring was obviously different. Although people have conducted a lot of experimental research on the ETO process on HZSM-5 catalyst, the role of acid sites in different channels of HZSM-5 catalyst in the reaction of ethanol to light olefins is still unclear.

Different pore structures and acid site distributions of HZSM-5 catalyst have significant effects on the adsorption and diffusion behavior of reactants, which in turn affects the selectivity of the products. Therefore, understanding the adsorption and diffusion behavior of ethylene and propylene on the HZSM-5 catalyst is of great significance for optimizing the catalyst design and improving the reaction selectivity. However, the Roles of the acid sites in different pore channels of HZSM-5 catalyst on ethanol conversion to light olefin have not been systematically investigated up to now. In this work, DFT was used to calculate the energy required to catalyze the conversion of ethanol to ethylene and propylene at the acid sites of the straight, sinusoidal, and intersection channels of the HZSM-5 zeolite. The reaction path of ethanol conversion to ethylene and propylene on HZSM-5 catalyst is based on our previous work [15] and the work of Yin et al. [16] (as shown in Figure 1). The adsorption and diffusion behaviors of ethylene and propylene on different pore acid sites of HZSM-5 catalyst were simulated by molecular dynamics method, and the effects of different pore acid sites on the selectivity of ethylene and propylene were discussed.

2. Results and Discussion

2.1. DFT Simulation Results

Based on our previous work [15] and the work of Yin et al. [16], the reaction path of ethanol conversion to ethylene and propylene is proposed. The specific process is as follows: Ethanol is initially adsorbed on the acid site of HZSM-5 catalyst, in which the hydroxyl group of ethanol interacts with hydrogen protons in the HZSM-5 catalyst, resulting in the production of ethylene and water. The resultant ethylene and reactant ethanol are co-adsorbed on the surface of the HZSM-5 molecular sieve, leading to ethylation reactions that give rise to 1-butyl cations. Subsequently, a proton is lost from the methyl group to form butylene. Some butylene undergoes dimerization with 1-butylcation, resulting in the formation of 3-octyl cations. Proton transfer occurs within these 3-octyl cations, leading to their conversion into 2-octyl cations, and 2-octyl cations are decomposed by β-scission to form propylene and 1-pentyl cations. 1-pentyl cations undergo proton transfer to form the 2-pentyl cations, and the 2-pentyl cations undergo β-scission to form propylene.

Eight reaction processes involve eight transition states (TSs), as shown in Figure 2. The activation barrier of a single step of each TS is calculated as the energy difference between a TS and its previous intermediate. The energy barrier differences of all TSs are shown in Figure 2. The reaction free energy barriers of three different channels were compared, and then the most favorable channel for the conversion of ethanol into ethylene and propylene was selected.

Figure 3 shows the channels that require the least energy for each step of converting ethanol into ethylene and propylene at T10, T11, and T12 acid sites of HZSM-5 catalyst.

As shown in Figure 3, the energy required to catalyze the dehydration of ethanol to ethylene on the acid site of the sinusoidal channel of HZSM-5 molecular sieve is the lowest, so the acid site of the sinusoidal channel of HZSM-5 molecular sieve is more conducive to the conversion of ethanol to ethylene. In the process of producing ethylene from ethanol, HZSM-5 catalyst with more sinusoidal pore acid sites can be selected. For the conversion of ethanol to propylene, the rate-determining step of the reaction path is β-scission. The straight channel acid site of HZSM-5 molecular sieve catalyzes the conversion of ethanol to propylene, and the energy required for β-scission is the lowest, so the straight channel acid site of HZSM-5 molecular sieve is more favorable for catalyzing the conversion of ethanol to propylene. In the process of producing propylene from ethanol, HZSM-5 catalyst with more straight channel acid sites can be selected.

The smaller the front orbital energy gap difference (∆E) (LUMO-HOMO), the easier the electron transition and the stronger the reactivity. Eight reaction steps in the reaction pathway were calculated (

Table 1. Front molecular orbital energy gap difference of reaction pathway.

Channels	∆E (kJ/mol)
	Dehydration	Ethylation	Deprotonation	Dimerization	Proton Transfer	β-Scission	Proton Transfer	β-Scission
Straight channel	499.03	474.25	456.06	478.90	469.82	463.29	470.74	456.34
Sinusoidal channel	486.05	483.46	470.54	463.20	460.45	467.68	465.16	482.78
Intersection channel	498.93	494.57	480.54	440.86	490.83	476.20	482.77	480.04

). From the energy gap value of ∆E (LUMO-HOMO), it can be seen that the energy gap value of dehydration reaction is the smallest in the acid site of the sinusoidal channel of HZSM-5 molecular sieve, indicating that the dehydration reaction is the most reactive in the acid site of the sinusoidal channel, and the dehydration reaction preferentially occurs in the acid site of the sinusoidal channel. The energy gap value of β-scission reaction is the smallest at the acid site of the straight channel, indicating that β-scission reaction is the most reactive at the acid site of the straight channel, and β-scission reaction preferentially occurs at the acid site of the straight channel. The obtained results are consistent with the above DFT calculation results. The conversion of ethanol to ethylene takes place preferentially at the acid site of sinusoidal channel, and the conversion of ethanol to propylene takes place preferentially at the acid site of straight channel.

2.2. Dynamics Simulation Result

The adsorption behavior of single-component ethylene and propylene on HZSM-5 catalyst with different SiO₂/Al₂O₃ ratios and different channels was simulated at a temperature of 823 K and a pressure of 0–1000 kPa. The simulation results are presented in Figure 4 and Figure 5, and

Table 2. Adsorption parameters of pure components of ethylene and propylene.

SiO2/Al2O3	Channel	Langmuir Constant b (kPa−1 × 10−4)		Isosteric Heat (kcal/mol)		Average Loading (per cell)
		C3H6	C2H4	C3H6	C2H4	C3H6	C2H4
64	T10	2.78	2.55	11.10	8.86	7.65	5.73
	T11	2.80	2.63	11.34	8.51	8.19	5.78
96	T10	2.60	2.52	11.06	8.47	7.91	5.84
	T11	2.63	2.58	11.01	8.34	7.68	5.66
128	T10	2.73	2.38	10.89	8.25	8.00	5.56
	T11	2.70	2.48	10.56	8.26	7.97	5.58

. The Langmuir adsorption model (Equation (1)) was used to analyze the adsorption data of ethylene and propylene, as shown in Figure 4.

$$\mathrm{Q}={\displaystyle \frac{Q_{m}bp}{1+bp}}$$

(1)

where Q is the adsorption capacity in cm³/g, Q_m is the saturated adsorption capacity in cm³/g, b is the adsorption equilibrium constant in kPa⁻¹, and p is the partial pressure of the components in kPa.

As shown in Figure 4, the adsorption of ethylene and propylene on HZSM-5 molecular sieves was monolayer adsorption, and the adsorption followed the Langmuir adsorption model. Propylene is loaded more than ethylene in both straight and sinusoidal channels of HZSM-5 catalyst. This is mainly due to the fact that propylene (3.4 Å diameter) has a larger molecular structure with more carbon atoms than ethylene (2.8 Å diameter).

The isosteric heat of adsorption is an important thermodynamic parameter for designing gas separation systems. As shown in Figure 5, the adsorption process of ethylene and propylene on the HZSM-5 catalyst is exothermic, and the isosteric adsorption heat of propylene is obviously higher than that of ethylene, which is consistent with Pham et al. [17]. With the increase of adsorption amount, the adsorbate molecules reach saturation, the adsorption heat of adsorbate molecules basically remains unchanged, and the adsorption heat of propylene fluctuates at 11.0 kcal/mol, while the adsorption heat of ethylene is at 8.0 kcal/mol.

As shown in Table 2, with the increase of SiO₂/Al₂O₃, the Langmuir constant of ethylene in sinusoidal channel (T11) of HZSM-5 catalyst decreases, which is beneficial to the desorption of ethylene and improves the selectivity of ethylene; the Langmuir constant of propylene in the straight channel (T10) of HZSM-5 catalyst decreases, which is beneficial to the desorption of propylene and improves the selectivity of propylene. Among the three kinds of SiO₂/Al₂O₃ studied, the Langmuir constant of HZSM-5 catalyst with SiO_2/Al₂O₃ = 128 is smaller, so the acid sites of sinusoidal channels of HZSM-5 catalyst with SiO₂/Al₂O₃ = 128 are more conducive to improving ethylene selectivity; the acid sites of straight channels of HZSM-5 catalyst with SiO₂/Al₂O₃ = 128 are more conducive to improving propylene selectivity.

The adsorption behavior of binary mixtures of ethylene and propylene on HZSM-5 catalyst with straight (T10) and sinusoidal (T11) channels of HZSM-5 (SiO₂/Al₂O₃ = 64, 96, 128) at 823 K was investigated, and the results are shown in Figure 6 and Figure 7, and

Table 3. Adsorption parameters of binary mixture of ethylene and propylene.

SiO2/Al2O3	Channel	Langmuir Constant b (kPa−1 × 10−4)		Isosteric Heat (kcal/mol)		Average Loading (per Cell)
		C3H6	C2H4	C3H6	C2H4	C3H6	C2H4
64	T10	1.61	1.14	11.24	8.49	4.21	2.75
	T11	3.21	1.49	11.19	8.53	3.98	2.71
96	T10	2.29	1.04	10.91	8.44	3.95	2.76
	T11	2.53	1.07	10.90	8.30	4.09	2.73
128	T10	1.50	1.30	10.96	8.33	3.88	2.71
	T11	1.91	0.64	10.96	8.35	4.02	2.72

.

As shown in Figure 6 and Figure 7, and Table 3, compared with the adsorption of pure components, the adsorption amount of propylene and ethylene in binary mixture is reduced. As shown in Table 3, the interaction between propylene and HZSM-5 catalyst skeleton is the weakest in the straight channel (T10) with SiO₂/Al₂O₃ = 128 and the strongest in the sinusoidal channel (T11) with SiO₂/Al₂O₃ = 64. At the sinusoidal channel (T11) with SiO₂/Al₂O₃ = 64, the interaction between ethylene and HZSM-5 catalyst skeleton is the strongest, and at the sinusoidal channel (T11) with SiO₂/Al₂O₃ = 128, the interaction between ethylene and HZSM-5 catalyst skeleton is the weakest. Therefore, propylene is more likely to desorb on the straight channel (T10) of HZSM-5 with SiO₂/Al₂O₃ = 128 in the adsorption of binary mixtures, which is more conducive to improving the selectivity of propylene. Ethylene is more likely to be desorbed from the sinusoidal channel (T11) of HZSM-5 with SiO₂/Al₂O₃ = 128, which is more conducive to improving the selectivity of ethylene.

The Mean Squared Displacement (MSD) is the deviation of the component’s position from the reference position over time. The MSD indicates the ability of the component to migrate on a specific time scale and can be used to determine the time-dependent diffusion coefficient. The diffusion coefficient is calculated using the Einstein equation [18], as shown in Equation (2).

$$\mathrm{D}={\displaystyle \frac{1}{2\mathrm{dN}}}\underset{\mathrm{t}\to \infty }{\lim }{\displaystyle \frac{\mathrm{d}}{\mathrm{dt}}}{\sum }_{\mathrm{i}=1}^{\mathrm{N}}〈{\left[\overrightarrow{{\mathrm{r}}_{\mathrm{i}}\left(\mathrm{t}\right) } - \overrightarrow{{\mathrm{r}}_{\mathrm{i}}\left(0\right)}\right]}^2〉$$

(2)

where D is the diffusion coefficient, t is the time, N is the number of target molecules in the system, and $\overrightarrow{{\mathrm{r}}_{\mathrm{i}}\left(\mathrm{t}\right)}$ and $\overrightarrow{{\mathrm{r}}_{\mathrm{i}}\left(0\right)}$ are the coordinates of the ith particle at moments t and 0, respectively.

As shown in Figure 8, the MSD curves of pure component ethylene are higher than pure component propylene, attributed to the stronger interaction of propylene with the HZSM-5 sieve framework. The diffusion coefficients of HZSM-5 catalyst with different silica–aluminum ratios and different channels for ethylene and propylene are listed in

Table 4. Diffusion coefficient of pure components of ethylene and propylene.

SiO2/Al2O3	Channel	Diffusion Coefficients × 10−8(cm2/s)
		C3H6	C2H4
64	T10	1.05	1.38
	T11	1.77	1.24
96	T10	0.87	3.27
	T11	1.32	2.30
128	T10	0.55	1.89
	T11	0.59	2.50

. Wang et al. [19] reported diffusion coefficients of 1.95 × 10⁻⁸ m²/s and 0.87 × 10⁻⁸ m²/s for pure ethylene and pure propylene, respectively, in ZSM-5 at 873 K, which is consistent with the present results. As shown in Table 4, for the diffusion of pure components, the diffusion coefficient of ethylene is larger than that of propylene, which is consistent with the adsorption simulation results, indicating that the generated ethylene diffuses more easily from the HZSM-5 catalyst channels in the ETO reaction. The reason for the higher ethylene yield than propylene in the ETO reaction was further explained from the diffusion point of view.

Radial distribution function (RDF) can be used to describe the spatial structure of a particle system, such as the average distance between particles and the degree of aggregation between particles. It is defined as the ratio of the number of particles neighboring a given particle at a distance of r to the total number of particles in the system, and its formula is shown in Equation (3).

$$\mathrm{RDF}\left(\mathrm{r}\right)={\displaystyle \frac{4\mathsf{\pi }{\mathrm{r}}^2\mathrm{n}\left(\mathrm{r}\right)}{\mathrm{N}}}$$

(3)

where n(r) is the particle number density at distance r, and N is the total number of particles in the system.

The spacing distributions of ethylene and propylene molecules with acid site H protons were statistically analyzed using RDF, and the calculated results are shown in Figure 9. The RDF plots of ethylene and propylene on the structure of HZSM-5 catalyst with different SiO₂/Al₂O₃ ratios and different channels exhibit a similar trend, with a high probability of finding ethylene or propylene molecules at a distance of 2.70 Å, and a low probability of finding ethylene or propylene molecules at distances exceeding 7 Å. Thus, the adsorption location is close to the zeolite backbone.

As shown in

Table 5. Diffusion coefficient of ethylene and propylene in binary mixture.

SiO2/Al2O3	Channel	Diffusion Coefficient × 10−8(cm2/s)
		C3H6	C2H4
64	T10	0.69	2.51
	T11	0.85	2.48
96	T10	0.47	1.55
	T11	1.00	2.69
128	T10	1.36	1.12
	T11	1.58	1.58

, for the diffusion of the mixture of ethylene and propylene, the diffusion coefficient of ethylene is greater than that of propylene, so there is more interaction between propylene molecules and HZSM-5 catalyst. In HZSM-5 (SiO₂/Al₂O₃ = 128), the highest diffusion coefficient was found for propylene and the lowest for ethylene, which is beneficial to propylene production.

Diffusion selectivity refers to the separation of different components due to different diffusion rates on the adsorbent surface during the adsorption separation of a mixture. The results of adsorption selectivity (α) and diffusion selectivity (S) [20,21] of ethylene/propylene binary gas mixtures in HZSM-5 catalyst framework were calculated by Equations (4) and (5) and the results are shown in

Table 6. The diffusion selectivity (S) and adsorption selectivity (α) for binary mixture in the studied structures.

SiO2/Al2O3	Channel	${\mathbf{S}}_{{\mathbf{C}}_{\mathbf{3}}{\mathbf{H}}_{\mathbf{6}}\mathrm{/}{\mathbf{C}}_{\mathbf{2}}{\mathbf{H}}_{\mathbf{4}}}$	${\mathsf{\alpha }}_{{\mathbf{C}}_{\mathbf{3}}{\mathbf{H}}_{\mathbf{6}}\mathbf{/}{\mathbf{C}}_{\mathbf{2}}{\mathbf{H}}_{\mathbf{4}}}$
64	T10	0.27	1.53
	T11	0.34	1.47
96	T10	0.30	1.43
	T11	0.37	1.50
128	T10	1.21	1.43
	T11	1.00	1.48

.

$$\mathrm{S}={ \mathrm{D}}_{{\mathrm{C}}_3{\mathrm{H}}_6}/{\mathrm{D}}_{{\mathrm{C}}_2{\mathrm{H}}_4}$$

(4)

$$\mathsf{\alpha }={\displaystyle \frac{{\mathrm{y}}_{{\mathrm{C}}_3{\mathrm{H}}_6}/{\mathrm{y}}_{{\mathrm{C}}_2{\mathrm{H}}_4}}{{\mathrm{x}}_{{\mathrm{C}}_3{\mathrm{H}}_6}/{\mathrm{x}}_{{\mathrm{C}}_2{\mathrm{H}}_4}}}$$

(5)

where ${\mathrm{D}}_{{\mathrm{C}}_2{\mathrm{H}}_4}$, ${\mathrm{D}}_{{\mathrm{C}}_3{\mathrm{H}}_6}$ are the diffusion coefficients of ethylene and propylene, ${\mathrm{y}}_{{\mathrm{C}}_2{\mathrm{H}}_4}$, ${\mathrm{y}}_{{\mathrm{C}}_3{\mathrm{H}}_6}$ are the adsorption loadings of ethylene and propylene, and ${\mathrm{x}}_{{\mathrm{C}}_2{\mathrm{H}}_4}$, ${\mathrm{x}}_{{\mathrm{C}}_3{\mathrm{H}}_6}$ are the molar fractions of ethylene and propylene.

As shown in Table 6, the diffusion selectivity of propylene to ethylene increased with the increase of SiO₂/Al₂O₃ ratio, and the maximum diffusion selectivity to propylene in HZSM-5 (SiO₂/Al₂O₃ = 128) straight channel (T10) was 1.21, which is beneficial to the separation of propylene and improves the selectivity of propylene. In HZSM-5 (SiO₂/Al₂O₃ = 128) straight channel (T10), propylene has less selectivity for ethylene adsorption. Therefore, in the straight channel (T10) of HZSM-5 (SiO₂/Al₂O₃ = 128), it is more beneficial to improve the selectivity of propylene.

3. Computational Methods

3.1. DFT Simulation Calculation

The GGA-PBE functional and high-precision DNP basis set were employed for calculation using the All Electron method with a real space Orbital Cutoff of 4.5 Å. The convergence standard was set to Fine with allowable deviations of total energy, gradient, and structure displacement at 10⁻⁵ Ha, 0.002 Ha·Å⁻¹, and 0.005 Å, respectively. Transition state search and reaction energy barrier calculations were performed using the Complete LST/QST method with a Medium accuracy setting allowing for deviations in total energy (2 × 10⁻⁵ Ha), gradient (0.004 Ha·Å⁻¹), and structure displacement (0.005 Å). The acid sites in the straight channel, sinusoidal channel and intersection channel of zeolite HZSM-5 are simulated by strategically placing an aluminum atom at the T10, T11, and T12 positions, respectively [22,23] (Figure 10).

Equation (6) below calculates the energy required for the reactants to the transition state on the HZSM-5 catalyst:

$$\text{∆}\mathrm{E}=\left({\mathrm{E}}_{\mathrm{TS}} - {\mathrm{E}}_{\mathrm{R}}\right)·2625.50184$$

(6)

where ΔE represents the energy required for the reactant to the transition state, kJ·mol⁻¹, ΔE_R represents the total energy of the reactants adsorbed on the catalyst, and Ha; ΔE_TS represents the total energy of the transition state adsorbed on the catalyst, Ha.

3.2. Molecular Dynamic Simulation

The original cell of ZSM-5 molecular sieve was obtained from the International Zeolite Association (IZA), and the 2 × 2 × 2 (a = 40.180 Å, b = 39.476 Å, and c = 26.284 Å) cell was modeled in Materials Studio 2018 software. Replacing the silicon atoms at the T10 and T11 sites with aluminum atoms and adjusting the charge of each atom of the molecular sieve, the charge of all Al atoms is 1.4 Å, the charge of all O atoms is −1.2 Å, and the charge of all Si atoms is 2.4 Å. Six molecular sieve model structures with SiO₂/Al₂O₃ = 64, 96, and 128 were constructed, respectively (Figure 11).

The Sorption module in Materials Studio 2018 was used for adsorption calculations. The Compass force field was selected for the calculation. Charges were managed using the Forcefield Assigned method. Van der Waals potential was calculated using the Atom Based method, and Ewald method was employed for electrostatic interactions with an accuracy of 0.042 J/mol. The simulation consisted of 10⁶ steps, with the first 10⁵ steps dedicated to achieving equilibrium.

Molecular dynamic calculations for diffusion were performed in the Forcite module of the Materials Studio 2018, with Dynamics selected for the computational task, Medium accuracy, a value of 12.5 Å for the cutoff distance, a time step of 1.0 fs, and a total simulation time of 2500 ps. The NVE system was used, and the Nose method was employed for the temperature control.

4. Conclusions

The catalytic behavior of T10, T11, and T12 acid sites (representing straight channel, sinusoidal channel, and intersection channel, respectively) of HZSM-5 catalyst on the reaction network of ethanol conversion to ethylene and propylene was studied by DFT. The results show that the conversion of ethanol to ethylene occurs preferentially at the acid site of the sinusoidal channel (T11); the conversion of ethanol to propylene takes place preferentially at the acid site of the straight channel (T10). The adsorption and diffusion behavior of ethylene and propylene in the straight and sinusoidal channels of HZSM-5 catalyst was simulated by molecular dynamics method. The results show that for the adsorption of pure components and binary mixtures, the acid sites of sinusoidal channel (T11) with SiO₂/Al₂O₃ = 128 are more conducive to improving the selectivity of ethylene, and the acid sites of straight channel (T10) with SiO₂/Al₂O₃ = 128 are more conducive to improving the selectivity of propylene. For the diffusion of pure components and binary mixtures, the diffusion coefficient of ethylene is greater than that of propylene. The diffusion selectivity of propylene to ethylene increases with the increase of SiO₂/Al₂O₃ ratio. Therefore, in the straight channel (T10) of HZSM-5 (SiO₂/Al₂O₃ = 128), it is more beneficial to improve propylene selectivity.