The effect of accessibility to acid sites in Y zeolites on ring opening reaction in light cycle oil hydrocracking

LCO and HDT-LCO characterization

To obtain the molecular size of aromatic hydrocarbons in LCO and HDT-LCO, hydrocarbon composition and carbon number distribution were analyzed in detail. The hydrocarbon composition is measured by standard methods ASTM D 2425 and ASTM D 8276 with Agilent 7890A/MSD5977 gas chromatography-mass spectrometry (GC-MS). The carbon number distribution of hydrocarbons is measured by JEOL JMS-T200GC-TOF gas chromatography-field ionization source time of fight mass spectrometry (GC-FI TOFMS). A 0.2 μL sample was injected onto a HP-5 MS column (30 m × 0.25 mm inner diameter × 0.25 μm film thickness), which was held at 60 °C for 2 min, ramped to 300 °C at 20 °C/min, and then held at a final temperature for 5 min. The GC-TOF MS interface temperature was at 300 °C. The field ionization emitter was at 10 kV voltage, with a 10 mA emitter current on a 10 μm emitter.

Catalyst preparation

Three hierarchical porous HY zeolites with similar Si/Al molar atomic ratios but different pore distributions, denoted as HY1, HY2, and HY3, were selected as acidic components of catalysts [Supplementary Table 1]. Their atomic molar ratios of Si to Al range from 3 to 5.

Subsequently, a composite carrier of zeolite and alumina with a dry mass ratio of 60 wt% HY zeolite and 40 wt% binder was prepared as follows: the HY zeolite mixed with pseudo-boehmite (Sasol Co., Germany) and homogenized, and then extruded using a single-screw extruder (the BONNOT Co., America) through a die with a diameter of 1.6 mm. The resulting extrudates were dried at 393 K for 3h followed by calcination at 873 K under an air atmosphere for an additional 3 h.

Finally, the NiMo/(HY-Al₂O₃) catalysts were prepared by impregnating molybdenum trioxide, nickel carbonate hydroxide, and phosphoric acid aqueous solution onto the supports with 15 wt% MoO₃, 3 wt% NiO and 3 wt% P₂O₅. Subsequently, they were dried at 393 K for 3 h and calcined at 723 K under an air atmosphere for another 3 h to obtain corresponding catalysts labeled as C1, C2, and C3 [Supplementary Table 2].

Zeolites and catalysts characterization

The crystal structures of zeolite were characterized through X-ray diffraction (XRD) using a German Siemens D5005 X-ray diffractometer, with Cu K_α radiation source and a scanning range of 2θ from 4° to 36°, at a tube voltage of 40 kV and tube current of 40 mA.

The specific surface areas and pore size distributions were determined by N₂ adsorption-desorption isotherms at 77 K, performed on an ASAP 2400 automatic adsorption analyzer produced by Micromeritics. Zeolite samples (20-40 mesh) were first vacuum-treated at 350 °C for 4 h. N₂ was used as the adsorbate, and adsorption was performed at -196 °C until reaching a state of static adsorption equilibrium. The specific surface area of the molecular sieve was calculated using the Brunauer-Emmett-Teller (BET) equation, and the pore size distribution was determined using the Barrett-Joyner-Halenda (BJH) model.

The aluminum distribution of the zeolites was obtained by using ²⁷Al magic angle spinning nuclear magnetic resonance (MAS NMR) on a Bruker Avance 500 NMR spectrometer.

A DB-FTIR using the Nicolet IS 50 FTIR spectrometer was employed to qualitatively characterize the acidic sites of the zeolites by the adsorption of alkaline probe molecules Py, 2,4,6-TTBPy and THA. The zeolite powder was pressed into a self-supporting thin wafer (1.5 cm²), which was placed in the sample beam of the DB-FTIR; the reference beam was not placed with the sample. Then, the sample was decontaminated at 400 °C under vacuum (10^-2 Pa) for 4 h. After the pretreatment, the sample was cooled down to room temperature for adsorption. The spectra of the adsorbed probe molecules were obtained as follows: firstly, excess vapor-phase probe molecules were introduced at room temperature, and then desorbed at 150 ^oC for 30 min. The spectra were obtained by subtracting the reference spectrum from the measured sample spectra. The amount of acid is calculated by the integral area of the corresponding band of the spectrum^[26].

Performance evaluation

The catalyst performance evaluation was conducted on a 16-channel fixed-bed high-throughput evaluation equipment with HDT-LCO containing 1,000 ppm S and 500 ppm N as feedstock. Typically, a 2 mL catalyst (16-20 mesh) was loaded in a tubular reactor, and inert SiO₂ (40-60 mesh) was filled on both sides of the reactor constant temperature zone. The catalyst needs to be pre-sulfurized before the reaction, using kerosene containing 2.5% dimethyl disulfide (DMDS) as the sulfurization oil at 573 K and 6.4 MPa for 8 h. After sulfurization, switch to the reaction oil and adjust the reaction pressure to 5.5 MPa, and raise the temperature to reaction temperature in the range of 663 to 683 K. After stabilizing for 48 h, the gas products were quantitatively analyzed using online gas chromatography, and the liquid products were analyzed off-line.

The LCO conversion and ring-opening rate were determined by

The conversion of LCO:

The ring-opening rate of PAHs and polycyclic cycloalkanes in LCO:

where m_in and m_out refer to the mass fractions of the components in the feedstock and products with boiling point higher than 478 K, respectively. C_in and C_out refer to the total amounts of cycloalkanes and aromatic hydrocarbons containing two or more rings in the feedstock and products, respectively. It is worth noting that during the hydrocracking process, tricyclic aromatic hydrocarbons in LCO will undergo hydrosaturation, ring-opening and cracking reactions to transform into bicyclic aromatic hydrocarbons. Similarly, tricyclic cycloalkanes will also be converted into bicyclic cycloalkanes through ring-opening and cracking reactions. However, accurately distinguishing whether these bicyclic aromatic hydrocarbons and cycloalkanes originate from LCO feedstock or reaction generation remains challenging. Since they are intermediate products, a simplified calculation method is adopted in Equation (2). This method does not account for the ring-opening rate of tricyclic aromatic hydrocarbons and cycloalkanes to bicyclic aromatic hydrocarbons and cycloalkanes, which may lead to an underestimation of the ring-opening rate.

Aromatic distributions and typical molecular structures in LCO and HDT-LCO

The hydrocarbon compositions of LCO and HDT-LCO are presented in Table 1, showing that the total aromatic content in LCO is 89.0%. Among these aromatic hydrocarbons, the bicyclic aromatic hydrocarbons account for 68.4% of the total, with a content of 60.9%. Bicyclic aromatic hydrocarbons mainly include naphthalene, acenaphthene, fluorene, and their alkyl-substituted derivatives according to molecular structure, among which naphthalene and alkyl-substituted naphthalenes are the most abundant. In contrast to LCO, the remaining PAHs including bicyclic and tricyclic aromatic hydrocarbons in HDT-LCO show a significant decrease, while the levels of monocyclic aromatic hydrocarbons, cycloalkanes, and alkanes exhibit a notable increase. Among these changes, the most substantial rise is observed in the content of monocyclic aromatic hydrocarbons, particularly those tetralin or indane-like structures. As previously mentioned, certain PAHs present in LCO undergo partial hydro-saturation reaction during the hydrotreating process and are converted into tetralin-like monocyclic aromatic hydrocarbons, which serve as primary reactants for subsequent hydrocracking reactions.

To obtain the molecular structure, we conducted the carbon number distributions of bicyclic aromatic hydrocarbons in LCO and monocyclic aromatic hydrocarbons in corresponding HDT-LCO. As illustrated in [Figure 2], alkylated naphthalenes in LCO predominantly exhibit a carbon number ranging from 11 to 14, while the alkylated acenaphthenes and fluorenes mainly fall within the range of 14 to 16. This suggests that the bicyclic aromatic hydrocarbons in LCO are predominantly alkyl-substituted with 1-4 side chain carbon atoms. Similarly, monocyclic aromatic hydrocarbons in HDT-LCO mainly consist of tetralin-like compounds with a carbon number ranging from 11 to 14. This is attributed to the fact that polycyclic aromatic compounds primarily undergo hydro-saturation reactions during the hydrotreating process, while the dealkalation and cracking of alkyl side chains are almost negligible.

Figure 2. (A) Carbon number distribution of bicyclic aromatic hydrocarbons in LCO and (B) monocyclic aromatic hydrocarbons in HDT-LCO. LCO: Light cycle oil; HDT-LCO: hydrotreated LCO.

GC-MS is utilized to further identify the predominant bicyclic aromatic hydrocarbons in LCO for calculating their molecule sizes, with a specific focus on alkyl-substituted naphthalenes with carbon atom numbers of 12 and 13. As illustrated in Figure 3, alkyl-substituted naphthalenes with C₁₂ predominantly consist of various isomers including dimethylnaphthalene, along with a minor presence of ethylnaphthalene. Similarly, alkyl-substituted naphthalenes with C₁₃ primarily consist of diverse isomers such as trimethylnaphthalene, accompanied by a small quantity of methylethyl naphthalene and propylnaphthalene. These findings indicate that the PAHs in LCO mainly comprise polysubstituted alkyl-aromatic compounds featuring short side chains, which undergo transformation into the corresponding tetralin-like monocyclic aromatic hydrocarbons in HDT-LCO during the hydrotreating process.

Figure 3. Main molecular structure of naphthalenes in LCO. LCO: Light cycle oil.

Table 2 presents the three-dimensional dimensions of the predominant alkyl-naphthalenes and corresponding alkyltetralins in C₁₂ and C₁₃. It is evident that these molecules are significantly larger than the micropores of HY zeolite (0.74 nm × 0.74 nm), rendering them unable to penetrate the pores of the zeolite.

Table 2

The three-dimensional dimensions of typical C₁₂ and C₁₃ naphthalenes and tetralins

Molecule	Structure	a × b × c / nm × nm × nm
2,7-dimethylnaphthalene		1.10 × 0.72 × 0.40
1,7-dimethylnaphthalene		1.01 × 0.83 × 0.41
1,3,6-trimethylnaphthalene		1.09 × 0.85 × 0.40
1,3,5-trimethylnaphthalene		0.99 × 0.87 × 0.40
2,7-dimethyltetralin		1.14 × 0.72 × 0.53
1,7-dimethyltetralin		1.03 × 0.83 × 0.53
1,3,6-trimethyltetralin		1.12 × 0.85 × 0.52
1,3,5-trimethyltetralin		1.02 × 0.89 × 0.53

Quantitative analysis of physicochemical property and BAS accessibility in HY zeolites

XRD characterization of the three HY zeolite samples was conducted and the results were depicted in Figure 4A. All the samples display typical diffraction peaks of the FAU topology structure zeolite. No significant additional peaks are observable.

Figure 4. (A) XRD pattern, (B) ²⁷Al MAS NMR spectra, (C) pore size distribution of HY zeolites. XRD: X-ray diffraction; MAS NMR: magic angle spinning nuclear magnetic resonance.

The aluminum coordination states on the three HY zeolites are illustrated in Figure 4B. The peak with a chemical shift (δ) at approximately 60 ppm corresponds to the tetrahedrally coordinated framework aluminum (FAl), while the peak with δ at around 0 ppm belongs to the octahedrally coordinated Extra-FAl (EFAl). The assignment of the peaks with chemical shifts between 30 and 50 ppm is still under debate, but they are generally considered as distorted tetrahedral or penta-coordinated EFAl. As depicted in Figure 4B, all the HY zeolites predominantly contain FAl, with a minor presence of EFAl. Among them, HY2 exhibits the lowest content of EFAl.

Table 3 and Supplementary Figure 1 show the textural properties of the HY zeolites. The specific surface areas of these zeolites range from 651 to 700 m²/g, and the pore volumes vary from 0.355 to 0.400 cm³/g. The mesoporous specific surface area and mesoporous pore volume increase gradually from HY1 to HY3. It can be seen from Figure 4C that there is a gradual increase in the proportion of mesopores (primarily 10-40 nm) in zeolites from HY1 to HY3, which could influence the accessibility of acidic centers.

FTIR is usually employed to qualitatively characterize the acidic sites of the zeolites. Many basic molecules, such as small alkaline probe molecules (ammonia and Py) and large probe molecules (multi-alkylated Py or other organic amine molecules), can be used to characterize the acidity of samples. The selection of different alkaline probe molecules can provide diverse information on acidity. Therefore, it is desirable and highly challenging to choose the appropriate alkaline probe molecule based on the chemical properties and molecular size of reactants.

Three alkaline probe molecules with different molecular sizes and chemical properties were selected to assess the accessibility of acid sites in HY zeolites, aiming to investigate their impacts on the performance of LCO hydrocracking for light aromatic hydrocarbon production. The adsorption energies on the external surface BAS were calculated by DFT (B3LYP/6-31 + G^**). Table 4 shows the total acid amount and external surface Brønsted acid amount of the three HY zeolites. Considering its molecular size (~0.57 nm), Py can approach almost all the acidic sites of the zeolites. Thus, the acid amount measured by Py is defined as the total acid amount. Additionally, 2,4,6-TTBPy (~1.1 nm) and THA (~1.4 nm) were chosen as probe molecules, resembling the aromatics in LCO and HDT-LCO [Table 2], which can only approach the acidic sites on the external surface of the zeolites due to their large molecular size. The acid amount measured by these molecules is defined as external surface acidity.

As shown in Figure 5, 1,540 cm^-1 of Py adsorption, 3,377 cm^-1 of 2,4,6-TTBPy adsorption, and 1,450 cm^-1 of THA adsorption on zeolites are attributed to the presence of BAS. Table 4 shows that the total Brønsted acid amount measured by Py gradually decreases from HY1 to HY3, while the external surface Brønsted acid amount measured by THA shows a gradual increase. The external surface Brønsted acid amount measured by 2,4,6-TTBPy is exceedingly low and also exhibits a gradual decrease. The three alkaline probe molecules provide totally different information on acidity of these HY zeolites.

Figure 5. Acid amount on BAS measured by (A) Py and external surface BAS measured by (B) 2,4,6-TTBPy and (C) THA of the three HY zeolites. BAS: Brønsted acid sites; Py: pyridine; 2,4,6-TTBPy: 2,4,6-tri-tert-butylpyridine; THA: trihexylamine.

The calculation results from Figure 6 and Table 5 indicate that Py exhibits strong ion adsorption with BAS. Among the probe molecules, THA possesses the largest molecular size, which hinders its entry into zeolite channels. However, due to its quasi-planar structure, it exhibits relatively small steric hindrance with the zeolite framework, resulting in relatively strong ion adsorption on the surface BAS. In contrast, 2,4,6-TTBPy has a smaller molecular size with THA, but its increased steric hindrance between the three tert-butyl branches and the zeolite framework prevents nitrogen on the Py ring from interacting with the BAS and leads to the reduced adsorption energy. Therefore, the Brønsted acid amount measured by 2,4,6-TTBPy would be underestimated.

Figure 6. Structures of Py, 2,4,6-TTBPy and THA adsorbed on BAS of HY zeolite. Py: Pyridine; 2,4,6-TTBPy: 2,4,6-tri-tert-butylpyridine; THA: trihexylamine; BAS: Brønsted acid sites.

LCO hydrocracking process and the effects of BAS accessibility on catalytic performance

As stated in Figure 1, monocyclic aromatic hydrocarbons in HDT-LCO undergo isomerization, ring-opening, and cracking reactions on the hydrocracking catalysts to yield desirable BTX products. Additionally, side reactions such as over-cracking, excessive hydrosaturation, and polymerization coking are also facilitated. Based on their distinct boiling points, the products are classified into C₁-C₄ gases, naphtha ranging from C₅ to 478 K, and unconverted diesel above 478 K. The naphtha products are further divided into C₅-C₁₂ paraffins, C₅-C₁₂ cycloalkanes, and C₆-C₁₂ aromatic hydrocarbons according to carbon number and molecular structure.

A comparative analysis was conducted to investigate the hydrocracking process of HDT-LCO, focusing on changes of hydrocarbons in unconverted diesel and distributions of product (Catalyst 1, C1). As illustrated in Figure 7A, there is a significant reduction in levels of monocyclic aromatic hydrocarbons, such as alkylbenzenes, indanes or tetralins, and indenes compared to the HDT-LCO feedstock. Additionally, an increase in reaction temperature leads to reduced paraffin contents along with cycloalkanes, alkylbenzenes and indanes or tetralins. However, the contents of bicyclic and tricyclic aromatic hydrocarbons initially decrease before subsequently rising. This phenomenon may be attributed to thermodynamic equilibrium limitations on hydrosaturation reaction for aromatic hydrocarbons at elevated temperatures.

Figure 7. (A) Changes of hydrocarbons in unconverted diesel and (B) distributions of product on C1 catalyst.

Under the experimental conditions [Figure 7B], C₆-C₁₂ aromatic hydrocarbons are identified as the main products. With rising reaction temperature, there is a continuous increase in the yield of C₆-C₁₂ aromatic hydrocarbons, a gradual increase in the yield of C₁-C₄ gaseous products, and a slight decrease in the yield of C₆-C₁₂ cycloalkanes. The yield of C₆-C₁₂ paraffins initially increases before exhibiting a decreasing trend.

These observations indicate that, within the hydrocracking reaction zone, minimal amounts of cycloalkanes and paraffins in HDT-LCO undergo cracking reactions, and ring-opening reactions of tetralin-like monocyclic aromatic hydrocarbons and cracking reaction of corresponding ring-opening products predominantly occur, leading to the formation of C₆-C₁₂ aromatic hydrocarbons along with minor amount of gaseous and light saturated hydrocarbons.

A comprehensive comparison of product distribution over the three catalysts is presented in Table 6, demonstrating a gradual increase in the conversion of HDT-LCO from C1 to C3. Moreover, the conversion was enhanced with rising reaction temperature. The hydrocarbon changes in unconverted diesel and product distribution at 683 K on the three catalysts are further illustrated in Figure 8. As shown in Figure 8A, the aromatic hydrocarbons in unconverted diesel on C3 catalyst, including monocyclic, bicyclic, and tricyclic compounds, are the lowest, followed by C2 and C1. Consequently, as depicted in Figure 8B, the yields of C₁-C₄ gases, C₅-C₁₂ alkanes, C₅-C₁₂ cycloalkanes and C₆-C₁₂ aromatic hydrocarbons produced by the C3 catalyst are the highest among all catalysts examined. Additionally, C2 catalyst exhibits slightly higher C₅-C₁₂ cycloalkanes and C₆-C₁₂ aromatic hydrocarbons than those on C1.

Figure 8. Influences of catalysts on (A) changes of hydrocarbons in unconverted diesel and (B) distributions of product.

Although the LCO conversions and product yields differ among the three catalysts, there is a notable increase in the LCO conversion as the ring-opening rate of PAHs and polycyclic cycloalkanes in LCO increases [Figure 9]. A strong linear correlation is evident, further affirming that the ring-opening reaction plays a crucial role in determining the reaction activity during LCO hydrocracking processing.

Figure 9. Relationship between the ring-opening rate of PAHs and LCO conversion. PAHs: Polycyclic aromatic hydrocarbons; LCO: light cycle oil.

Figure 10 revealed a strong positive correlation between the amount of accessibility Brønsted acid on HY zeolites determined by THA and the ring-opening rate for PAHs and polycyclic cycloalkanes, as well as the yield of C₆-C₁₂ benzenes and BTX. As a heterogeneous catalytic process, the initial and key step of the ring-opening reaction involves diffusion and adsorption of reactant molecules onto acidic sites. While both THA and 2,4,6-TTBPy are unable to enter HY zeolite micropores, it is worth noting that THA molecules exhibit smaller steric hindrance compared to 2,4,6-TTBPy. This makes it more similar to the adsorption behavior of aromatics in LCO and HDT-LCO (the C–H and O–H bond strengths of 2,7-dimethylteralin adsorbed on HY zeolites are 2.08 and 0.99 Å, respectively). Therefore, THA is better suited for characterizing the accessibility of aromatics in LCO and HDT-LCO to the acidic sites over HY zeolite hydrocracking catalysts.

Figure 10. Relationship between ring-opening rate, light aromatic hydrocarbons yield and acid amount of accessibility BAS. BAS: Brønsted acid sites.