HZSM-5 zeolite modification and catalytic reaction mechanism in the reaction of cyclohexene hydration

This study investigated a three-phase (liquid–liquid–solid) reaction system of cyclohexene hydration where the catalyst was hydrophilic at the bottom of the water phase. Cyclohexene conversion was low since it was difficult for it to come into contact with the oil. The HCl-OTS-HZSM-5 catalyst was prepared by acid and alkylation modification, then subsequently characterized. Acid modification enabled HZSM-5 zeolite to remove some aluminum atoms, increasing specific surface area, pore volume, and acid sites. N-octyltrimethoxysilane (OTS) was grafted onto the HZSM-5 zeolite surface, i.e., alkylation modification, to improve the contact area between immiscible reactants. Consequently, cyclohexene conversion reached 24.07%. Cyclohexene hydration was calculated using the 34T cluster model, and bridging hydroxyl and water molecule adsorption sites were explored. Simulations confirmed that the reaction energy barrier for the HCl-OTS-HZSM-5 catalyzed hydration reaction = 46.67 kJ mol−1, considerably less than that with HZSM-5 zeolite (73.78 kJ mol−1). The theoretical results reasonably explain the experiments and provide guidance to prepare catalysts with high catalytic activity.


Introduction
The Si/Al ratio is particularly critical to the catalyst acidity. The presence of aluminum atoms creates a negative charge in the framework that is compensated by hydrogen ions. 1,2 Hydrogen protons generate strong electrostatic elds enabling strong interactions with polar molecules, but zeolites' inherent hydrophilicity limits their application. [3][4][5] Zeolite alkylation modication by organic alkylation reagents can substantially increase catalyst hydrophobicity. 6 Adsorption of organic groups on HZSM-5 zeolites can improve hydrophobicity, but also signicantly reduces specic surface area and pore volume. The alkylating agent adsorbs onto the zeolite surface to form a multi-layer porous material, blocking catalyst pores. Li et al. 7 modied HZSM-5 zeolite by CLD salinization to not only passivate external surface acidity, but also reduce pore size. Increased active sites can signicantly improve olen conversion efficiency, but this was hindered for alkylation modied microporous zeolite materials. 8 Previous studies have shown that acid modication can lead to pore expansion. Chen et al. 9 modied zeolite by HCl, increasing specic surface area and pore volume, by reducing silanol groups, and hence reduced zeolite surface affinity for water. Acid modication could improve pore structure and provide more framework acid sites, providing favorable conditions for molecular reactions. 10,11 HZSM-5 zeolite has excellent catalytic performance for several reactions, including isomerization, alkylation, and aromatization. It also exhibits enhanced aromatic selectivity and organic liquid product yield compared with other catalysts. 12,13 Density functional theory (DFT) has been widely used to study factors affecting catalytic reaction mechanisms, including adsorption conguration, Gibbs reaction energy barrier, reaction transition state, etc. 14-21 Fu et al. 22 used DFT for 8T and 48T models with the mGGA-M06-L function to study mono-branched alkanes in HY and HZSM-5 adsorption energies. They concluded that pore connement was critical to adsorbate stability. Rosenbach 23 and Mullen 24 showed that carbocations occurred in the zeolite channel in the alkoxy group conguration using HZSM-5 catalyst. Carbocation was conned to near the oxygen atom, and the central carbon atom was connected with the oxygen atom on the zeolite framework by covalent bonds. 25,26 MFI zeolite has been widely to study catalysis. This paper modied HZSM-5 zeolite by acid and alkylation to obtain HCl-OTS-HZSM-5 zeolite catalyst, then subsequently investigated catalyst acid sites and hydrophobic properties to explore optimal cyclohexene conversion. The catalytic cracking mechanism for HZSM-5 zeolite was investigated using DFT, calculating elementary reaction Gibbs energy barriers through simulation. The theoretical calculation results reasonably explain experimental outcomes and provide guidance to prepare of catalysts with high catalytic activity.

Catalyst preparation
HZSM-5 zeolite was obtained from Nankai University, China. We chose HZSM-5 zeolite with Si/Al ¼ 25 because it offered high acid site density to better promote cyclohexene hydration. (Octyl)-trimethoxysilane (OTS) was obtained from MACKLIN. Methylbenzene, ethyl alcohol, and carbon tetrachloride were provided by SINOPHARM. HCl acid was obtained from Yantai San he Chemical Reagent Co., Ltd. The specic preparation steps have been described in previous work. 27,28 The best HCl-HZSM-5 was selected for 25% alkylation modication 27 (hydrochloric acid modication concentration ¼ 4 mol L À1 ). 28 The HCl-OTS-HZSM-5 catalyst was prepared by acid and alkylation modication.

Catalyst evaluation
The reactants were added to the reaction kettle for reaction, and the products were separated for quantitative analysis by gas chromatography aer reaction. The main product was cyclohexanol, which the by-products were small amounts of methylcyclopentene and methylcyclopentanol. The calculation formulas of cyclohexene conversion and cyclohexanol selectivity were as follows:

¼
Reaction to produce cyclohexanol mass Â 82:15 Cyclohexene conversion quality Â 100:16 Â 100% (2) The catalyst evaluation device was shown in Fig. 1. The reactants were added to the autoclave and replaced with nitrogen for 3-5 times. Aer the temperature of the reaction kettle rises to a certain temperature, the timing was started, and the reaction was completed for a period of time. The reaction kettle was immediately quenched in a cold-water bath, and the reaction solution was poured into a centrifuge tube. Solidliquid separation was performed by centrifugation, and the catalyst was recovered. The water-oil biphasic product was separated using a separating funnel, and the aqueous phase was extracted with 1,2-dichloroethane. The obtained extract phase and oil phase were added with internal standard ethanol, respectively, and analyzed by gas chromatography. Fig. 2 shows the 34T cluster model employed to represent the HZSM-5 zeolite, following previous studies. 29 ZSM-5 zeolite comprised two intersecting 10-ring channels forming sinusoidal channels moving along the crystal a-axis and straight channels along the crystal b-axis, with corresponding 5.5 Â 5.1 and 5.3 Â 5.6Å pore sizes, respectively. The 10-ring channel comprised 10 T atoms, where T represents Si or Al. Replacing T12(Si) with Al reduced steric hindrance for larger intermediates in the framework (see Fig. 2(a) and (b)). Articially cutting the Si-O bond caused the boundary atoms to produce unrealistic dangling bonds, with consequential boundary effects. The boundary effect was reduced by adding hydrogen atoms since H and Si have similar same electrical properties and the Si-H bond direction was consistent with the Si-O bond direction in the periodic model. Replacing a silicon atom by an aluminum atom cased the zeolite framework to become negatively charged, which was added to the zeolite by introducing hydrogen protons to maintain the structure's electroneutrality, and also forming a Brønsted acid site. 30 Fig. 2(c) shows the (SiO) 3 -Si-OH-Al-(OSi) 3 8T structure was relaxed to reduce the calculation overhead and prevent the structure from collapsing during geometry optimization, and the remaining part was xed on crystal coordinates.

Cyclohexene conversion ¼
Mass of cyclohexene before reaction À Mass of cyclohexene after reaction Mass of cyclohexene before reaction Â 100% (1) modication did not change the crystal structure. Thus, deposited alkylating agent (OTS) was highly dispersed, i.e., amorphous phase, during alkylation modication.
3.1.2 FT-IR. Fig. 4 shows FT-IR spectra for HZSM-5, HCl-HZSM-5, and HCl-OTS-HZSM-5 exhibit characteristic absorption near 450, 550, 840, 1095, 1220 and 1633 cm À1 peaks, con-rming that all three catalysts were MFI-type zeolites. The peak around 3450 cm À1 was ascribed to terminal hydroxyl group and hydrogen-bonded adjacent hydroxyl group stretching vibrations. 33 Absorption peak intensities for the HCl-HZSM-5 and HCl-OTS-HZSM-5 hydroxyl groups reduced compared with HZSM-5. Surface wettability could have changed from hydrophilic to hydrophobic and the modied catalyst characteristic peak at 3610 cm À1 intensity decreased due to Al atom removal reducing Si-OH-Al groups. The HCl-OTS-HZSM-5 sample exhibits an infrared absorption peak $3000 cm À1 , mainly due to C-H stretching vibration in the CH 3 -group. The OTS C8 hydrophobic chain was successfully adsorbed on the zeolite surfaces, forming a densely packed alkyl chain layer.
3.1.3 TPD and pyridine-IR. Fig. 5 shows NH 3 -TPD (Ammonia adsorption curve) for HZSM-5, HCl-HZSM-5, and HCl-OTS-HZSM-5 all exhibit two absorption peaks: a weak acid center in the low temperature region (about 150 C) and strong acid center in the high temperature region (about 400 C). The low temperature peak was mainly due to the weak acid center for the external skeleton Al, and the high temperature peak   mainly due to the strong acid center for the internal skeleton Al. 34 The weak acid center peak shied forward aer acid modication. It was difficult for OTS molecules to enter zeolite channel interiors due to zeolite's relatively small size; but they covered the outer surface, hence covering acid sites on the outer surface. Fig. 6 and Table 1 show zeolite B and L acid regions distinguished by Py-IR adsorption infrared spectroscopy. B and L acid centers occurred in all samples, with B acid content decreasing slightly aer acid modication, from 348 to 338 mmol g À1 ; whereas L acid content decreased signicantly, from 41 to 33 mmol g À1 . This was mainly due weak Al acidity, hence removing Al reduced weak acid content, and retained B acid enhanced olen hydration. In contrast, the OTS-HCl-HZSM-5 zeolite exhibits considerably different strong and weak acid reductions, mainly because OTS was graed on the catalyst surface, covering some acid sites and hence reducing acid content. Physical OTS adsorption bridged hydroxyl groups, causing pore blockages and hence reducing Brønsted and Lewis acid sites.
3.1.4 Nitrogen adsorption and desorption. Fig. 7 shows N 2 adsorption isotherms of all catalyst samples in the IV microporous materials. Table 2 shows structural parameters for the three catalysts. HCl-OTS-HZSM-5 zeolites increased from 264 to 280 m 2 g À1 ; HCl-OTS HZSM-5 reduced to 232 m 2 g À1 , and total     Table 2 The BET surface area, pore diameter and pore volume of HZSM-5, HCl-HZSM-5 and HCl-OTS-HZSM-5 zeolites pore volume reduced to 0.14 cm 3 g À1 . The alkylating agent (OTS) had difficulty entering the zeolite interior and connecting with hydroxyl groups through the "Si-O-Si" bond. Thus, OTS deposited on the outer catalyst surface, forming a thick hydrophobic layer, effectively reducing the porosity and increasing surface resistance to N 2 . The plugged mesoporous structure changed intergranular pore size, hence greatly increasing catalyst specic surface area and pore volume compared with previous studies. 25 3.1.5 Contact angle. Fig. 8 shows contact angle (CA), to investigate sample wettability. Unmodied samples exhibit good hydrophilicity with CA # 20 . Al atoms were removed from the zeolite framework by acid modication, destroying Al-O bonds and enhancing Si-O bonds. Since the Si-O-Al bond was more polar than Si-O-Si (the covalent bond had higher binding energy), the zeolite became less polar and more hydrophobic. Thus, HCl-OTS-HZSM-5 zeolite CA ¼ 122.6 was modied by alkylation to determine the hydrophobicity. Acid modication enlarges pore volume, exposing more hydroxyl groups and hence OTS molecules became more easily adsorbed on the zeolite surface. Therefore, we achieved signicant CA improvement compared with previous studies. 25 Acid modication increases hydrophobicity by removing aluminum atoms and reducing the polarity of molecular sieves; while alkylation modication increases hydrophobicity by adsorbing C8 hydrophobic chains on the surface of molecular sieves through "Si-O-Si" bonds. From the macroscopic observation, the modication effect of molecular sieve alkylation is more obvious.
3.1.6 TGA. Fig. 9 shows TGA proles to assess sample stability. Mass loss from 0 to 150 C was mainly attributed physiosorbed water removal. The HZSM-5 sample with lower Si/ Al ratio adsorbed more water, conrming enhanced hydrophilicity, and adsorbed up to 8 wt% water; whereas HCl-HZSM-5 and OTS-HCl-HZSM-5 samples with higher Si/Al ratio adsorbed only 6 and 5 wt% water, respectively. Thus, acid and alkylation modied samples were more hydrophobic. Catalyst mass for HZSM-5 and HCl-HZSM-5 no longer changed with increasing temperature aer physiosorbed water desorption. However, OTS-HCl-HZSM-5 suffered serious mass loss, mainly due to n-octyltrimethylsilane (OTS) detachment from the zeolite surface at high temperature.
3.1.7 XPS. Fig. 10 shows elemental spectra from C1s, O1s, Al2p, Si2s and Si2p XPS proles over binding energy ranges from 0 to 1250 eV; and Table 3 shows corresponding calculated elemental contents. Acid modication effectively removed Al. Efficient OTS coverage increased carbon signal; whereas effective organosilanes coverage strongly enhanced the carbon signal and decreased both silicon and oxygen signal intensities.   Thus, XPS results conrm n-octyltrimethoxysilane was covalently bonded to HZSM-5 zeolite surface, consistent with FT-IR outcomes (Fig. 4).

HCl-OTS-HZSM-5 catalytic activity
Previous studies evaluated catalysts under static conditions, whereas the current study considered dynamic temperature, water-ene molar ratio, reaction time, catalyst dosage, and catalyst recycling.
Forward reaction rate increased and equilibrium shied toward the forward reaction direction with increasing temperature since olen hydration is reversible exothermic. Cyclohexanol selectivity decreased signicantly as temperature was raised further, probably due to by-products caused by the high temperature. Optimum reaction temperature ¼ 130 C.
Excess water was selected to participate in the reaction due to high cyclohexene raw material prices. Cyclohexene conversion rst increased and then decreased with increasing water molar  ratio, whereas selectivity remained above 99%. Optimum water to ene molar ratio ¼ 6. The reaction tends to balance with increasing reaction time, and conversion rate was not signicantly improved by extending the reaction time. Thus, optimal reaction time ¼ 4 h. Cyclohexene hydration was acid-catalyzed, and increasing centers provided stronger catalytic activity. However, the conversion rate did not increase signicantly once sufficient catalyst was present, and excess catalyst tended to cause separation difficulty. Thus, optimum catalyst dosage ¼ 1.65 g, i.e., 10 wt% relative to water mass. Fig. 11 shows repeated catalyst cycling in the reaction kettle. Cyclohexene conversion decreased somewhat aer 5 cycles, but cyclohexanol selectivity did not change signicantly. This may be because more organic matter adhered to the catalyst surface, covering some acid centers and hence affecting catalyst activity. Thus, catalyst regeneration was required. Organic material was washed away using ethanol and suction ltration, then dried at 120 C for 12 h to restore catalyst activity. Optimal reaction conditions for cyclohexene hydration were reaction temperature ¼ 130 C, water to ene molar ratio ¼ 6, reaction time ¼ 4 h, and catalyst dosage ¼ 1.65 g (10 wt% relative to water mass). Maximum conversion rate ¼ 24.07%, and selectivity remained at 99%.

Cyclohexene and water molecule adsorption
We calculated cyclohexene reaction energy barriers and compared with kinetic parameters to verify parameter selections and calculated results rationality. Activation energy for HZSM-5 catalyzed cyclohexene hydration reaction ¼ 77.69 kJ mol À135 from estimated kinetic model parameters; whereas simulations predicted energy barrier ¼ 73.78 kJ mol À1 . Thus, the model and parameter settings were conrmed to be reliable. Fig. 12 shows stable adsorption congurations for cyclohexene and water molecules on the HZSM-5 cluster model, with corresponding atoms marked. Cyclohexene molecules mainly cluster at straight sinusoidal and sinusoidal channel intersections due to connement effects from the pores.
Tables S1 and S2 † summarize Mulliken electron layouts and atomic distances before and aer adsorption. Oxygen atom (O A ) Mulliken charge on the acidic site reduced from À1.192 to À1.237 eV aer olen molecule adsorption, with H atom charge increasing to 0.616 eV, and O A -H bond length increasing from 0.978 to 0.997Å, due to forming p-H bonds between cyclohexene molecules and HCl-HZSM-5 zeolite. The C 2 atom on the cyclohexene molecule lost electrons, and hence C 1 -C 2 bond length increased from 1.331 to 1.498Å. O A -H bond length on the zeolite increased from 0.978 to 0.997Å, and both Al-O A and Si-O A bond lengths shortened due to the C]C double bond inuence; and H + tended to be protonated at the acidic site. Some electrons transferred from hydrogen atoms to the doublebonded carbon atoms, hence double-bonded C atom electron density increased and O A -H bond length increased. A hydrogen bond formed between the cyclohexene molecule and bridging hydroxyl group, allowing electron ow from the cyclohexene molecule to the zeolite model, increasing cyclohexene molecule positive charge. 36 Fig. 13 shows electron density for cyclohexene and zeolites. Electron density was lowest at the carbon-carbon double bond for cyclohexene, and highest at the hydrogen proton for zeolite.

Electron density distribution and LUMO and HOMO
The cyclohexene molecule will be repelled due to electrostatic repulsion when its high electron density region was close to the zeolite framework. Stabilizing effects from channel connement were also weak at this time, hence the zeolite framework separated from the adsorbate. Only there was a certain distance between the zeolite framework and the adsorbate, the carbon-carbon double bond of cyclohexene was close to the hydrogen proton. The low electron density region occupied the channel, and the stabilization effect produced by the channel connement effect was the strongest. This made the smaller pore size HZSM-5 zeolite, which could stabilize the cyclohexene molecules in the pores by electrostatic action. 37 Highest occupied orbital (HOMO) and lowest unoccupied orbital (LUMO) for the reactant molecule are key to determining the chemical reaction system. Zeolite catalyst HOMO orbitals were mainly distributed on the oxygen atoms adjacent to the Al phase, whereas cyclohexene LUMO orbitals were mainly distributed on the double-bonded carbon atoms (Fig. 12). Thus, catalyst hydrogen proton HOMO matched cyclohexane LUMO symmetry, enabling electron transfer from the HOMO to symmetry-matched LUMO orbitals (see number of electrons in Table 3). Therefore, high energy HOMO orbitals in the zeolite channel indicates strong electron donation in the reaction, and H + was preferentially transferred to the double-bonded carbon atoms.

Molecular adsorption site
4.3.1 Bridging hydroxyl protonation sites. Fig. 14 shows the bridged hydroxyl group protonation position comprises two reaction paths: cyclohexene protonation produced carbocations, and water molecule protonation formed hydronium ions. Fig. 14(a) and (b) show reaction product and optimized product stability, with calculated reaction energy barrier ¼ 73.78 kJ mol À1 . The transition state search was performed on the reactants and produced to explore the highest reaction energy barrier. Fig. 14(c) shows optimal hydronium ion reaction production, decomposing hydronium ions into water molecules and hydrogen ions once the reaction completes, and transferring hydrogen ions to oxygen atoms to form bridging hydroxyl groups, as shown in Fig. 14(d). Thus, hydronium ions were unstable when alkenes occurred and hydrogen protons transferred to the zeolite framework to reform B acid sites. Joshi et al. 38 conrmed hydronium ion instability under low water loading by RMD simulation, showing difficulty to form water clusters, hence proton transfer stalled and the protons eventually returned to the zeolite catalyst, retaining zeolite acidity. Liu et al. 39 showed that hydronium ions exhibited higher barrier to alkene protonation than bridging hydroxyl protonation, which preferentially underwent bridging hydroxyl protonation. Thus, bridging hydroxyl groups preferentially protonated the alkene during cyclohexene hydration at low water loading (Fig. 14(a)  and (b)).

Water molecule adsorption sites.
Bridging hydroxyl groups in the zeolite framework were strongly acidic and protons (H + ) were exchanged on adjacent oxygen atoms in HZSM-5 to maintain charge neutrality. 40,41 Water molecules are strongly adsorbed near these acidic centers to form protonated water clusters at higher water loadings; whereas water molecules formed hydrogen-bonds with HCl-HZSM-5 zeolite at low water loading. Fig. 15 shows that the reaction region can be divided into a, b, c and d regions due to the presence of water molecules.
Consequently, we investigated interaction between water molecules and Si-OH-Al sites during hydration by adding   a single water molecule to identied regions, and calculating O A -H bond lengths, as shown in Fig. 16. Distance between framework oxygen and hydrogen atoms ¼ 0.995Å in the absence of adsorbed water molecules; whereas O A -H bond lengths ¼ 0.997, 0.998, and 0.991 in the a, b, and c regions, respectively, when water molecules were adsorbed on the zeolite framework.
Acidic zeolite catalytic sites changed with the presence of water molecules, and changing O A -H bond length changed Brønsted acid site properties. Bridging hydroxyl groups were stable under low water loading, and O A -H bond lengths lengthened in the a and b regions due to O A -H bond interaction with hydrogen bonding of water molecules. Rapid proton exchange between bridging hydroxyl groups and water molecules occurred when water molecules were in the a and b regions, increasing zeolite acidity. However O A -H distance and proton mobility did increase signicantly when a single water molecule was in the c region. Fig. 17 shows calculated energy barriers for cyclohexene hydration reactions with water molecules in the three regions >88.78, 75.85, and 73.78 kJ mol À1 for regions c, b, and a, respectively. This ordering may be due to O A -H bond length was only 0.991Å in the c region, and weak B acid strength made proton transfer difficult. Thus, the rst step energy barrier for acid catalysis reaction was high (88.78 kJ mol À1 , Fig. 17(c)). The reaction energy barrier for dehydrogenation in the b region was  as high as 192.20 kJ mol À1 , which may be due to steric hindrance making dehydrogenation difficult, as shown in Fig. 18(c) and (d). Wang et al. 18 showed that 29 Si resonance in analyzed 29 SiNMR spectra produced larger downeld shi when interacting with protons, resulting in less shielding relative to trivalent Al ions displacing nearby Si atoms, and hence making it easier to adsorb near Al atoms. This result is consistent with the theoretical calculation data: reaction energy barrier for water molecules was smallest in the a region (73.78 kJ mol À1 ), and hence the reaction proceeded easily. Fig. 19 shows the cyclohexene hydration reaction route olen hydration mechanism can be divided into three stages: Fig. 19 The mechanism of cyclohexene hydration reaction.  1. The proton on the acidic catalyst was added to the doublebonded carbon atom, generating intermediate carbocation;

HZSM-5 zeolite reaction path
2. Intermediate carbocations reacted with water molecules to form protonated alcohols; 3. Protonated alcohol lost protons to form cyclohexanol. Olen molecule adsorption on HZSM-5 zeolite generated van der Waals forces within the zeolite framework, the olen molecule had p electrons, and the B acid centers generated electron induction. Thus, olen molecules and B acid centers formed a p-H bond. Thus, cyclohexene molecules were protonated by Brønsted acid centers before cleavage reaction, transferring protons to the cyclohexene double-bonded carbon atom. This reaction is consistent with previously observed free fatty acid protonation, 23 as shown in Fig. 20 and 21. Cyclohexene molecules were protonated by acidic H + to form adsorption intermediates (carbocations), and the TS-1 transition state (Fig. 21), where C 1 -H and C 2 -H distances were transformed to (2.694, 1.968) and (2.085, 1.149)Å, respectively. O A -H bond distance extended from 0.977 to 1.769Å, whereas C 1 -C 2 bond length increased to (1.331, 1.393)Å. Thus, acidic protons in zeolite moved towards the cyclohexene C 1 -C 2 bond, forming carbocations (C 1 -H-C 2 ). Fig. 22 shows Gibbs activation energy ¼ 73.78 kJ mol À1 for cyclohexene protonation from simulation.
The adsorbed carbocation was unstable and highly reactive, 15,16 and hence can decompose with only small activation energy. Thus it could difficult to locate this intermediate experimentally. Fig. 22 shows that lower activation energies were required for carbocations to generate protonated alcohols and cyclohexanol. Water molecules participated throughout the reaction (Fig. 20(b)-(d)). First, water molecules combined with carbocations to form protonated alcohols with low reaction energy barrier ¼ 33.03 kJ mol À1 (Fig. 22). C 1 -H 2 O bond length reduced to 1.529Å during this process, and C 1 -C 2 and C 2 -H distances reduced to (1.331, 1.498) and (1.149, 1.085)Å, respectively. Thus, water molecules were adsorbed on the carbocation to form protonated alcohols (Fig. 20(b) and (c)). Protonated alcohols were extremely unstable in the unsaturated cluster model and tended to undergo hydrogen transfer. Experimental results also conrmed activation energy for proton transfer reaction was very small ¼ 23.24 kJ mol À1 (Fig. 21, TS-3). C 1 -OH bond length gradually decreased from 1.529 to 1.453Å, and free protons transferred to the 34T O A atom as O A -H bond gradually reduced from 1.632 to 1.033Å (Fig. 20(c) and (d)). Thus, the proton transfer mechanism can be understood through structural parameter variations.    Fig. 23 shows the zeolite alkylation modication route. The hydroxyl group is cleaved on the surface of the zeolite, then the methyl group (CH 3 -) in n-octyltrimethoxysilane (OTS) is cleaved, and the two molecules are subsequently connected by Si-O-Si to complete the alkylation modication, producing HCl-OTS-HZSM-5 zeolite. DPT calculations shows the alkylation modication reaction energy barrier ¼ 189.27 kJ mol À1 (Fig. 24).
4.5.2 HCl-OTS-HZSM-5 reaction path. Fig. 25 shows the cyclohexene molecule is protonated by the Brønsted acid center, transferring the proton to the cyclohexene double bond C atom.
Adsorbed cyclohexene molecules are protonated by acidic H + to form adsorption intermediates (carbocations) (transition state TS-1, Fig. 25). C 1 -H and C 2 -H bond lengths change to (2.482, 1.947) and (1.871, 1.179)Å, respectively, O A -H bond length increases from 1.006 to 1.619Å, and C 1 -C 2 bond length increases to (1.334, 1.385)Å. Thus, acidic protons in zeolite moved towards the cyclohexene C 1 -C 2 bond, forming the C 1 -H-C 2 carbocation. Fig. 26 shows DPT simulation with reaction energy barrier ¼ 46.47 kJ mol À1 (Fig. 26). Fig. 26 shows adsorbed carbocations combine with water molecules in the zeolite channels to form protonated alcohols ( Fig. 25(b) and (c)), with low reaction energy barrier ¼ 22.62 kJ mol À1 , C 1 -H 2 O bond reduces to 1.556Å, and C 1 -C 2 and C 2 -H bond lengths change to (1.385, 1.493) and (1.179, 1.086)Å. Thus, water molecules are adsorbed on the carbocation to form protonated alcohols, which are extremely unstable in the unsaturated cluster model. Dehydrogenation subsequently occurs with Gibbs energy barrier for proton transfer ¼ 32.5 kJ mol À1 , C 1 -OH bond length gradually decreases from 1.556 to 1.454Å, and free protons transfer to the HCl-OTS-HZSM-5 zeolite O A atom as O A -H bond length reduces to 1.030Å (Fig. 25(c) and (d)).
Gibbs activation energy for cyclohexene protonation ¼ 73.78 kJ mol À1 during unmodied HZSM-5 catalytic reaction (Fig. 21); whereas Gibbs activation energy for protonated cyclohexene ¼ 46.47 kJ mol À1 aer alkylation modication during the HCl-OTS-HZSM-5 catalytic reaction (Fig. 26). Thus, the reaction energy barrier was greatly reduced aer alkylation modication. Fig. 27 shows O A -H bond lengths ¼ 0.997 and 1.006Å; whereas C 2 -H bond length ¼ 2.360 and 1.871Å before and aer adsorption, respectively. Thus, two-phase reactant contact was macroscopically enhanced aer alkylation modication.  Microscopically, water molecules were more likely to activate hydrogen protons, increasing O A -H bond length and B acid strength, and hence reducing the reaction energy barrier to more easily form carbocations. This is consistent with experimental outcomes, and hence the reaction conversion rate was greatly improved aer alkylation modication.

Conclusion
This paper prepared HCl-OTS-HZSM-5 zeolite catalyst by acid and alkylation modication and investigated olen hydration reaction effects. Acid modication removes some HZSM-5 zeolite aluminum atoms, increasing specic surface area, pore volume, and acidic active sites. Acid-modied catalyst exhibited good hydrothermal stability and congurational stability for hydrothermal processes, enhancing cyclohexene hydration reactions. We graed n-octyltrimethoxysilane (OTS) onto the HZSM-5 catalyst surface to improve interfacial mass transfer limitations for the two-phase reaction, forming Pickering emulsion which improved contact between immiscible reactants. Consequently, cyclohexene conversion reached 24.07%, with relatively stable selectivity close to 100%.
We constructed a 34T cluster model using the mGGA-M06-L function for DFT calculations to represent the zeolite channel structure, and investigated reaction activation energy for cyclohexene hydration in HZSM-5 and HCl-OTS-HZSM-5 zeolite channels. Simulations conrmed that bridging hydroxyl groups preferentially protonated alkenes at low water loadings, and water molecules were more easily adsorbed near less shielded Al atoms. The highest reaction energy barrier for HCl-OTS-HZSM-5 zeolite catalyzed by cyclohexene hydration reaction was only 46.67 kJ mol À1 , which greatly reduced the reaction energy barrier. Simulations also conrmed alkylation modication feasibility.

Conflicts of interest
There are no conicts to declare.