The hydroisomerization of normal long-chain paraffin is an efficient process applied in lubricant base oil production, and it can effectively improve the low-temperature physicochemical properties of lubricating oil [1-6]. Various types of catalysts have been developed to catalyze this reaction [7-12]. Among them, bifunctional catalysts prepared by loading noble metals such as Pt with a one-dimensional pore molecular sieve, such as SAPO-11 (AEL), ZSM-22 (TON), and ZSM-23 (MTT), have made a great contribution [13-16]. In particular, SAPO-11 with the AEL topology, which has a one-dimensional ten-membered ring channel structure, has been considered as a potential material for the hydroisomerization of normal paraffin due to its suitable pore structure and moderate acidity [13]. The action of bifunctional catalysts toward hydroisomerization involves two major parts: the hydro-/dehydrogenation reaction of alkane molecules occurring at the metal center, and the alkane skeleton isomerization reaction occurring at an acidic site of a molecular sieve. The acidic property plays a very important role in guaranteeing the isomerization conversion of normal paraffin, and the micropore structure properties mainly control the isomer selectivity of the reaction [17, 18].

According to the pore mouth and key–lock mechanism, the Brönsted acid sites located near the outer layer of the molecular sieve crystal were determined to contribute greatly to the isomerization of the n-paraffin carbon skeleton [19, 20]. When the supported metal content of the catalyst is sufficient to catalyze the hydro-/dehydrogenation reaction, optimizing the physicochemical properties of the molecular sieve is the key to improving the hydroisomerization performance [17, 20]. In recent years, post-treatment methods with an acid or base as the etchant have been widely used to modify the physicochemical properties of molecular sieves [21-28]. Perez-Ramirez et al. [21] found that the external surface was enhanced after treatment with an alkali solution, which increased the conversion of the reaction. Reportedly, hydrochloric acid and citric acid treatments of SAPO-11 could enhance the methylation of naphthalene, owing mainly to the number of secondary mesopores [29]. In our previous work, we reported that the hydroisomerization performance could be enhanced by alkaline post-treatment of the ZSM-22 zeolite [30]. However, the current post-treatment method mainly involves aqueous acid or alkali solutions, which produce a large amount of wastewater. Moreover, the use of etching solutions leads to a low yield of molecular sieves. In 2012, a novel solvent-free route for synthesizing zeolites was reported, which provided a new idea for zeolite synthesis [31]. However, the development of solvent-free based post-treatment for preparing zeolite catalysts remains a great challenge. Recently, our group proposed a post-processing route to synthesize SAPO-34 with butterfly patterned hierarchical pores through a melting-assisted post-synthesis route that could improve the MTO performance [32].

Herein, we propose a new melting-assisted post-synthesis route to treat a molecular sieve (SAPO-11). It involves solid oxalic acid as the etching agent. The absence of water greatly reduces environmental pollution and improves the yield of SAPO-11. This is the first time a green post-processing route has been developed to tailor the pores and acid sites of SAPO-11. In this work, SAPO-11 with a one-dimensional pore structure of the AEL topology was used as the carrier of bifunctional catalyst. The parent SAPO-11 was treated with 5%, 10%, and 15% solid oxalic acid. Thereafter, the prepared samples were characterized by XRD, SEM, BET, Py-IR, NMR, etc. Afterward, the noble metal, Pt (0.5 wt%), providing the hydro-/dehydrogenation site was loaded on the bifunctional catalysts [25, 33]. Finally, n-dodecane was chosen as a model feed to investigate the hydroisomerization performance of the prepared catalysts under different experimental conditions.

The conventional hydrothermal method was used to prepare the SAPO-11 raw powder. Tetraethyl orthosilicate, pseudo-boehmite, and orthophosphoric acid were used as the Si source, Al source, and P source, respectively. n-Butanol was used as a co-solvent and di-n-propylamine was used as a template. The synthesis steps are as follows: first, 6.79 g of pseudo-boehmite was slowly added to the phosphoric acid solution (9.01 g of phosphoric acid dissolved in 21.16 g of H₂O) and stirred for 4 h; subsequently, 1.63 g of tetraethyl silicate was dissolved in 6.37 g of n-butanol, added dropwise, and stirred for 1 h. Next, 5.54 g of di-n-propylamine was slowly added and stirred for 1 h; finally, the gel obtained with the general composition of 1 Al₂O₃: 1 P₂O₅: 0.2 SiO₂: 1.4 DPA: 2.3 n-butanol: 35 H₂O was transferred to a stainless steel reaction vessel with a polytetrafluoroethylene liner and reacted in an oven at 210 ℃ for 24 h. The reaction product was washed with deionized water to neutrality, dried overnight in an oven at 80 ℃, and calcined in a muffle furnace at 500 ℃ for 6 h to remove the organic templating agent. The prepared sample was named SAPO-11-P.

Oxalic acid was selected to treat the SAPO-11-P sample according to the following procedures: firstly, a certain weight of the sample and solid oxalic acid were thoroughly mixed in a solid pulverizer (the mass ratios of solid oxalic acid to SAPO-11-P were selected as 5%, 10%, and 15%); thereafter, the mixed powder was transferred into a stainless steel reaction vessel with a polytetrafluoroethylene liner and reacted at 100 ℃ for 6 h. The reacted SAPO-11 samples were named SAPO-11-5%, SAPO-11-10%, and SAPO-11-15%.

The bi-functional catalysts were loaded with 0.5 wt% Pt with an H₂PtCl₆·6H₂O aqueous solution by the incipient wetness impregnation method and named as Pt-SAPO-11-P, Pt-SAPO-11-5%, Pt-SAPO-11-10%, and Pt-SAPO-11-15%.

The catalyst powder was compressed into tablets under a pressure of 20 MPa using a tablet pressing machine. After crushing in an agate mortar, the particles were passed through a stainless-steel sieve with 20–40 nm mesh sizes. The catalyst particle was packed in a fixed bed reactor and reduced in a hydrogen atmosphere at 350 ℃ for 6 h under atmospheric pressure. After the catalyst was reduced, the fixed bed reactor was adjusted to the reaction conditions. n-Dodecane was added to the fixed bed reactor using a pump. The liquid product was collected by a condensate tank, and the liquid product was analyzed by a GC-2014C gas chromatograph. The conversion and selectivity of n-dodecane was calculated as follows: conversion of n-dodecane = (moles of converted n-dodecane) ÷ (moles of the initial n-dodecane) ÷ 100%; selectivity of i-dodecane = (moles of i-dodecane) ÷ (moles of the initial n-dodecane) × 100%.

The crystal structures of the samples were characterized using an X-ray diffractometer (XRD, Rigaku, Ultima IV). The crystal morphologies of the samples were observed using a scanning electron microscope (SEM, Zeiss SUPRA 55 SAPPHIRE). The surface area and pore volume of the samples were determined by the Brunauer-Emmett-Teller (BET) and t-plot methods, and pore sizes in the ranges of 0–2 nm and 2–50 nm were calculated by the Horvath-Kawazoe (HK) method and the Barre-Joyner-Halenda (BJH) method, respectively. The elemental compositions of the samples were analyzed by X-ray fluorescence spectroscopy (XRF, Bruker S4PIONEER). The types and number of acid sites were estimated based on the Py-IR spectra (EQUINOX 70, Bruker, Germany). Solid-state MAS NMR experiments were performed at 5 kHz on a Bruker AVANCE 400 spectrometer operating at frequencies of 79.5 and 104.22 MHz for ²⁹Si and ²⁷Al, respectively.

Fig. 1 shows the XRD patterns of the SAPO-11 samples before and after oxalic acid treatment. All the crystal samples have diffraction peaks at 2θ of 8.1°, 9.4°, 13.1°, 15.6°, 20.3°, 21.2°, and 22.1°–23.2°, which are attributed to the typical SAPO-11 phase with the AEL topology. However, the peak intensities of different peaks change to a certain extent. The relative crystallinities of the samples were calculated using the diffraction peak area method in a specific diffraction angle range (2θ = 5°–50°). Compared to that of the parent sample, SAPO-11-P (100%), the relative crystallinities of the treated samples were 98.1%, 95.3%, and 92.1% for SAPO-11-5%, SAPO-11-10%, and SAPO-11-15%, respectively. The elemental analysis results in Table 1 also indicate that the molar compositions of the prepared samples changed slightly after treatment. With increasing the oxalic acid content, the relative crystallinity decreased gradually, although no impurity phase peaks were observed even with 15% solid oxalic acid, indicating that the pure SAPO-11 crystal was well-persevered after the treatment. Three crystal faces, (200), (020), and (002), were marked. Based on the peak area of the three characteristic peaks, the peak intensity ratios of the three crystal faces in the samples were calculated, as shown in Fig. 2. The results indicated that more (002) crystal faces were exposed, implying that more pores of the crystal could be exposed; thus, more acid sites were provided for the reaction [34]. As indicated by the pore mouth and key–lock mechanism, changing this crystal face may improve the hydroisomerization performance.

The SEM images of all the samples are shown in Fig. 3. SAPO-11-P (Fig. 3(a)) has spherical particles with sizes in the range of 4–6 μm with a compact packing structure formed by the accumulation of flaky crystals around a fixed center. Expectedly, it is difficult for the n-dodecane molecules to interact with the crystals inside the particles due to the limitation of the diffusion. For SAPO-11-5% (Fig. 3(b)), SAPO-11-10% (Fig. 3(c)), and SAPO-11-15% (Fig. 3(d)), SAPO-11-P was reacted with different amounts of solid oxalic acid at a microscopic scale. The three samples exhibited no obvious change in morphology according to the SEM images, which also indicate that the primary crystal structure was retained after the mild solid oxalic acid treatment.

N₂ adsorption and desorption isotherms for different SAPO-11 samples are presented in Fig. 4(a). All the samples exhibit the type IV isotherm with an H4-type hysteresis loop. The sharp increase in adsorption at a relative pressure of 10^–6 < P/P₀ < 0.01 is attributed to micropore filling, indicating the presence of microporous structures in the samples [35]. Simultaneously, the adsorption amount is significantly increased and the hysteresis loop appears near the saturation pressure (0.4 < P/P₀ < 1.0), which proves the existence of mesoporous and macroporous structures [36]. Generally, the intracrystalline mesopores in the samples are possibly generated by the lamination of the layered structure demonstrated by SEM images (Fig. 3).

Table 1 lists the BET data results. All the samples contain microporous and mesoporous structures, although mesoporous structures are dominant. The micropore and mesopore distribution curves of the samples are shown in Figs. 4(b) and 4(c). The micropore size is concentrated at about 0.5 nm for all the samples, and the mesopore size of SAPO-11-P is concentrated at about 3.8 nm and increases to about 4.0 nm for the samples after acid treatment. Meanwhile, the micropore and mesopore volumes of the samples changed accordingly, as shown in Table 1. Interestingly, there is no obvious difference between the pore volume and surface area of the samples after oxalic acid etching, due to the mild post-treatment condition. However, with the increase in the amount of oxalic acid, the samples possess relatively high ratios of the micropore surface area to the total specific surface area (R_a), and the micropore volume to the total pore volume (R_b), as shown in Table 1. The SAPO-11-P sample has a compact packing structure (Fig. 3), which causes the pores at the mutual contact faces of the crystals to be clogged. The pore volume and specific surface area of the contact surface cannot be effectively measured. When oxalic acid is etched to these contact faces, more micropores are released, which is consistent with the data in Table 1. The BET surface area, total pore volume, and micropore volume of SAPO-11-10% are 210 m²/g, 0.153 cm³/g, and 0.076 cm³/g, respectively; moreover, SAPO-11-10% has the largest R_a and R_b values among the samples.

Py-IR was used to evaluate the acidity of the SAPO-11 samples [37]. Figs. 5(a) and 5(b) show the infrared absorption spectra of pyridine at 200 and 350 ℃ of the samples, separately. The wave number regions of 1540–1548 and 1445–1460 cm^–1 are the infrared absorption peaks of pyridine at the Brönsted acid sites and Lewis acid sites, respectively [38]. Table 2 shows the intensity distributions of the number of acidic sites with the samples; the total Brönsted and Lewis acid sites were measured at 200 ℃, and the medium and strong Brönsted and Lewis acid sites were measured at 350 ℃. According to the results, the samples contained mainly medium and strong acid sites. In addition, with the increase in oxalic acid content, the total number of Brönsted acid sites increased first and then decreased. This is because, at a relatively low oxalic acid content, the acidic sites of the overlapping portions of the flaky crystals inside the SAPO-11 sample particles are exposed, resulting in an increase in the number of acid sites. At a relatively high oxalic acid content, the crystal structure is partially destroyed, resulting in an overall decrease in the number of acid sites. The SAPO-11-10% sample has the maximum number of total Brönsted acid sites and the minimum proportion (66.5%) of medium and strong Brönsted acid sites when compared to those of the other SAPO-11 samples. Fig. 6(a)and 6(c) presents the ²⁷Al and ²⁹Si MAS NMR spectra of SAPO-11-P and SAPO-11-10%. Fig. 6 shows that both samples have a tetra-coordinated framework of Al atoms, which was indicated by the resonance peak at about 54.5 × 10^‒6 [39]. The pentacoordinated Al resonance peak at about 21 × 10^‒6 resulted from the coordination of the template to the Al framework [40, 41]. The hexacoordinated Al resonance peak at about 0 × 10^‒6 is associated with the interaction between the tetra-coordinated Al species and water molecules [41]. Meanwhile, the signals of hexacoordinated Al of SAPO-11-10% is stronger than that of SAPO-11-P, which indicates that the acid treatment induces dealumination of the silicoaluminophosphate framework of SAPO-11. Fig. 6(c) and 6(d) exhibit a broad band ranging between –80 and –120 × 10^‒6. The values, –87 (peak 1), –94 (peak 3), –102 (peak 4), –107 (peak 5), and –112 × 10^‒6 (peak 6), are indicative of Si(4Al), Si(3Al, 1Si), Si(2Al, 2Si), Si(1Al, 3Si), and Si(4Si), respectively, in the SA domain, and –91 × 10^‒6 (peak 2) represents Si(4Al) in the SAPO domain [38]. The original data were deconvolved into six peaks using Gaussian curves; the results are provided in Table 3. The decrease in the Si(4Al) peaks in the SAPO and SA domains demonstrates the destruction of the SAPO-11 crystal structure, as shown in Fig. 1, which also implies that the Al species are very sensitive to acid treatment. The skeletal isomerization of hydrocarbons was promoted by the medium and strong Brönsted acid site distributions in Si(1Al) and Si(2Al) in the SA domains, and Si(3Al) located at the interface between the SA and SAPO domains [42]. Thus, the SAPO-11-10% sample with a high Si(nAl) (0 < n < 4) has more medium and strong Brönsted acid sites than those of the SAPO-11-P sample, which is in accordance with the Py-IR characterization results (Table 2).

The catalytic performance of the catalysts toward the hydroisomerization of n-dodecane was investigated. The reaction pressure was 2.0 MPa; the temperature range was 280–360 ℃, and the GHSV was 1.2 h^–1. Fig. 7(a) (temperature-conversion diagram) clearly illustrates that the conversion of n-dodecane increases with increasing temperature. Additionally, the conversion of n-dodecane increases in the following order: Pt-SAPO-11-P < Pt-SAPO-11-5% < Pt-SAPO-11-15% < Pt-SAPO-11-10%. Over the entire reaction temperature range, Pt-SAPO-11-10% exhibits a much higher conversion than those of the other samples. In addition, Fig. 8(a) shows that the reaction conversion of the catalyst increases as the number of Brönsted acid sites in the SAPO-11 samples increases. The mechanism of hydroisomerization of long-chain alkanes over bifunctional catalysts indicates that under conditions of sufficient hydrogenation sites, the reaction rate-determining step is a rearrangement of carbonium at the Brönsted acid sites, including the skeletal rearrangement or carbon–carbon bond rupture [17]. The Brönsted acid sites catalyze the alkane carbon skeletal isomerization reaction; the higher the Brönsted acid content, the higher the reaction conversion rate, which is consistent with the experimental result shown in Fig. 8(a). As indicated in Table 2, the SAPO-11-10% sample (161.6 μmol/g) has the highest number of Brönsted acid sites compared to those of the other SAPO-11 samples. Fig. 7(b) shows the changes in selectivity as a function of temperature. Normally, the selectivity decreases as the temperature increases due to the production of more cracking products at a relatively high temperature, which is consistent with the previous reports [28, 30, 32]. More interestingly, we found that the Ra value is positively correlated with the reaction selectivity indicated by the curves in Fig. 8(b). The pore mouth and key–lock mechanism of the alkane hydroisomerization reaction indicated that the main part of the alkane carbon skeletal isomerization occurs near the pores of the SAPO-11 molecular sieve. Even if the alkane reaches the depth of the pore by diffusion, the reaction would still be difficult to occur owing to the presence of space resistance, which causes the alkane carbon skeletal to crack easily. Therefore, the more micropore specific surface areas and (002) crystal faces of the SAPO-11 molecular sieve crystal are exposed, the more efficient reaction sites are exposed, which improve the selectivity of the hydroisomerization reaction [43]. Fig. 7(c) shows the isomer yield of the catalysts at different temperatures, and the best isomer yield performance is recorded at 320 ℃. The total yield of the isomers displays a volcanic type profile, and the highest yield of 80.1% appears for Pt-SAPO-11-10%, which is much higher than that of the parent Pt-SAPO-11-P sample (55.34%). Moreover, an excellent hydroisomerization performance requires not only high conversion and high isomer yield, but also high mono-branched selectivity to maintain its high viscosity index [30]. Table 4 lists the isomer product selectivity of the catalyst at 300 ℃; the selectivity of the mono-branched isomer increases after the post-treatment; however, the selectivity of multi-branched isomers decreases simultaneously. Consequently, the S_Mo/S_Multi ratio increases significantly (more than 3 times) after the oxalic acid treatment. Among all the prepared catalysts, Pt-SAPO-11-10%, having a higher n-dodecane conversion, higher isomer yield, and higher selectivity of mono-branched isomers, is the best catalyst. Based on the above experimental results, the hydroisomerization performance could be enhanced significantly over the bifunctional catalysts after subjection to facile solid oxalic acid treatment.

In this work, a facile solvent-free post-treatment process was developed for SAPO-11 based on a novel melting-assisted post-synthesis using solid oxalic acid. After loading Pt (0.5 wt%), the performance of the prepared bifunctional catalysts toward the hydroisomerization of n-dodecane was investigated. The treated SAPO-11 samples had more Brönsted acid sites, which significantly contributed to improved reaction conversion, higher micropore specific surface area ratio, and increased (002) crystal face ratio that enhanced the reaction selectivity, simultaneously. The isomeric product yield over Pt-SAPO-11-10% reached 80.1% compared to that over Pt–SAPO-11 (55.3%). Based on these results, the melting-assisted solvent-free synthesis method not only reduces wastewater but also optimizes the physicochemical properties of the catalyst; thus, an efficient and environmentally friendly method for enhancing hydroisomerization performance is provided.