Tert-butylamine is an important raw material and pharmaceutical intermediate, and the direct amination of isobutene over zeolite catalysts to give tert-butylamine has been reported to be an environmentally benign and economical process. MFI and BEA have been found to be superior to other materials ^{[1, 2]}. Zeolites ZSM-5 and ZSM-11 have similar framework densities and pore sizes, but differences in their channels result in a higher selectivity for aromatics in the aromatization of 1-hexene over ZSM-11 compared with that over ZSM-5 ^[3]. In the case of the direct amination of isobutene, we found that ZSM-11 showed slightly higher activity than ZSM-5 when using similar crystal size, SiO₂/Al₂O₃ molar ratio and acidity.

Zeolites synthesized by conventional techniques are usually obtained in powder form. However, for practical applications, the zeolite powder should be formed into catalyst particles with a certain shape, size, and mechanical strength by using inert materials as binders ^[4]. Unfortunately, this causes the reduction of the active zeolite component and blocks part of the channel windows in the formed catalysts, which consequently results in a considerable reduction in the catalytic performance of the zeolites ^[5]. It would therefore be interesting to prepare a mechanically stable and self-supporting zeolite catalyst, in which the active components are protected, namely, a binderless zeolite catalyst ^[6].

Self-bonded ZSM-5 pellets are readily prepared at 170 ℃ under hydrothermal conditions from a gel with the following composition: a Li₂O : b Na₂O : 2b TPABr : b A1₂O₃ : 150 SiO₂: 490 H₂O, with 8 ≤ a ≤ 9 and 14 ≤ b ≤ 16 (TPABr = tetrapropylammonium bromide). If the Li₂O is replaced with Na₂O or K₂O, only nonpelleted, powdered ZSM-5 samples can be prepared, which underlines the role of the LiAlO₂ binder in the system ^[7]. Recently, shaped binderless ZSM-5 zeolites have been prepared from aluminosilicate extrudates using a dry-gel conversion (DGC) technique, in which the introduction of amines into the steam favored the formation of nanosized ZSM-5 zeolite. Of particular note, the morphology of these aluminosilicate extrudates was retained in the crystallization process ^[8]. Self-bonded pellets of ZSM-11 with high mechanical resistance could also be obtained from initial gels with a composition of 1 Li₂O : (3.25-7.5) Na₂O : (3.25-7.5) Al₂O₃ : 150 SiO₂ : (0.65-1.5) TBABr : 490 H₂O ^[9].

TPABr ^[10], tetrabutylammonium bromide (TBABr) ^[11] and 1, 6-hexanediamine (HDA) ^[12] are suitable structure directing agents (SDAs) for the synthesis of ZSM-5, ZSM-11, and ZSM-5/ZSM-11, respectively. However, ZSM-11 synthesized using TBABr and HDA as co-SDAs has not yet been reported. Here, a shaped binderless ZSM-11 catalyst was prepared using the DGC method, using TBABr and HDA as co-SDAs in the synthesis and without adding other alkaline materials. The obtained sample was then employed for the direct amination of isobutene. In addition, the performance of the binderless and binder-containing ZSM-11 catalysts (HZSM-11+Al₂O₃ and HZSM-11+SiO₂) in the amination reaction was compared.

Na-ZSM-11 zeolite powder with a framework SiO₂/Al₂O₃ molar ratio of 61.6, and containing 2.33 wt% Na₂O and 10.4 wt% TBABr was synthesized by following a previously reported procedure ^[3]. The powder was mixed with silica sol (SiO₂: 30 wt%, Na₂O: 0.31 wt%) to form a cylindrical extrudate that was 1.8-2.2 mm in diameter and 3-7 mm in length. The dried sample containing 30 wt% of silica was denoted S1. For comparison, S2, which contained 30 wt% of alumina, was prepared by using boehmite instead of silica sol as the binder.

S1 was recrystallized using the DGC method. First, 4 g S1 was placed in a stainless-steel tube fixed and supported in a horizontal position in a stainless-steel autoclave with a Teflon lining. Then 1.2 g HDA and 4 mL deionized water were poured into the bottom of the Teflon-lined autoclave taking care that the solid sample did not come into direct contact with the liquid ^[13]. The autoclave was heated at 175 ℃ for 10.5 h, and then cooled to room temperature. The solid sample was dried at 120 ℃ for 2 h and then calcined in air at 540 ℃ for 5 h. This sample was denoted S3.

All three samples (S1, S2, and S3) were converted to the protonic form by conventional ion exchange in aqueous NH₄NO₃ solution ^[14], followed by calcination (550 ℃, 4 h, air). The resultant samples were denoted Cat-A, Cat-B, and Cat-C, respectively.

The samples were characterized by powder X-ray diffraction (XRD), N₂ adsorption-desorption, temperature-programmed desorption of NH₃ (NH₃-TPD), infrared spectrascopy of adsorbed pyridine (Py-IR), high-resolution transmission electron microscopy (HRTEM) and nuclear magnetic resonance (NMR) techniques as described elsewhere ^[15]. The side-pressure strength of the samples was tested using a ZQJ-II intelligent particle strength tester.

The direct amination of isobutene to tert-butylamine was carried out in a stainless-steel fixed-bed reactor and the mode of operation was down-flow. All the samples were pressed, crushed, and sieved to give a particle size of 0.38-0.85 mm before they were loaded into the reactor. Initially, 5 g of the catalyst was loaded into the center of the reactor and pretreated at 500 ℃ for 1 h in N₂. The reactor was cooled to 250 ℃, and then liquid NH₃ was pumped in from the top to fully fill the reactor before isobutene was introduced. Ma et al. ^{[16, 17]} have reported that this feeding sequence is favorable for the reaction. After 2 h, it was confirmed that the NH₃/i-C₄H₈ molar ratio remained stable at 4:1 by measuring the concentrations of the educts, and the products were analyzed on-line with an Agilent 7980B gas chromatograph equipped with a flame ionization detector (FID) and a PONA capillary column.

The isobutene conversion (C_isobutene), tert-butylamine selectivity (S_{tert-butylamine}) and formation rate of tert-butylamine (V_formation) were calculated according to the following equations:

(1)

(2)

(3)

where x_{tert-butylamine} is the molar percentage of tert-butylamine in the products, x_C8 is the molar percentage of C₈ hydrocarbons in the products, x_isobutene is the molar percentage of isobutene in the products, F_{M isobutene} is the molar flow rate of isobutene (μmol/h), and m_cat is the mass of the catalyst (g).

First, two binder-containing ZSM-11 samples (Cat-A and Cat-B) were prepared, and their XRD patterns are shown in Fig. 1. The relative crystallinity (RC) of the sample was calculated based on the areas of the diffraction peaks at 2θ = 7.9° ± 0.1°, 8.8° ± 0.1°, 23.0° ± 0.1°, 23.8° ± 0.1°, and 45.0° ± 0.1° using pure ZSM-11 in protonic form (HZSM-11), which was assumed to have 100% RC, as a standard. The RC of Cat-A was 69.9%, which is in good agreement with its composition, and that of Cat-B was 63.7%.

The DGC method uses steam to assist in heating the precursor and converting the amorphous structure to the crystalline form ^[18].Compared with hydrothermal synthesis, the impact of diffusion on the zeolite growth rate is minimized in the DGC method; this is attributed to the higher rate of diffusion for gases than for liquids. In the DGC process, the solid sample was not in direct contact with water; however, capillary condensation of water vapor usually means that the surface of the amorphous silica gel is covered with a thin layer of water during the steaming treatment ^[8]. In addition, seed addition can be used for the rapid synthesis of zeolites with controllable crystal size and morphology. Taking these ideas into consideration, we attempted to use the DGC method to prepare a binderless ZSM-11 zeolite catalyst.

The XRD pattern of the Cat-C sample obtained from 4 g of a mixture of ZSM-11 and SiO₂, 1.2 g HDA, and 4 mL H₂O after reaction at 175 ℃ for 10.5 h is shown in Fig. 1. It contains the typical diffraction pattern of HZSM-11 zeolite, and the RC of Cat-C was 101%, which is similar to that of HZSM-11. The detailed crystallization mechanism will be explored in another manuscript. It is worth noting that no other amines were added to the system during the synthesis except for HDA and TBABr. Generally, amines such as ethylamine and n-butylamine should be added into the DGC process; these amines have two functions: enhancing the alkalinity and acting as SDAs ^[8].

To confirm the coordination state of the Si species and the percentages of these species contained in the samples, ²⁹Si MAS NMR measurements were carried out and the spectra of various samples are given in Fig. 2. Deconvolution of the spectra of the Cat-A, Cat-C, and HZSM-11 samples showed that they each contained three peaks with chemical shifts centered at approximately δ = −114, −106, and −100. These chemical shifts correspond to Si(4Si), Si(3Si, 1Al), and Si(OH), respectively ^{[19, 20]}. Quantitative information regarding the percentages of the different Si species was also obtained from these spectra. The percentages of Si(4Si), Si(3Si, 1Al), and Si(OH) in HZSM-11 were 85.0%, 13.0%, and 2.0%, respectively. After the DGC process, the percentage of Si(3Si, 1Al) changed from 13.0% (Cat-A) to 10.7% (Cat-C), whereas that of Si(4Si) increased only slightly (from 86.1% to 87.6%). The percentage of Si(OH) was between 1.0% and 2.0%. These results suggest that the amorphous SiO₂ in Cat-A was transformed into ZSM-11 zeolite in Cat-C, and this resulted in the similar RCs for Cat-C (101%) and HZSM-11 (100%).

HRTEM images (Fig. 3) show that HZSM-11 has relatively regular crystal edges and does not contain any clearly visible amorphous materials. A large amount of amorphous SiO₂ was observed around the ZSM-11 in the Cat-A sample, but this phenomenon was hardly observed in Cat-C. The Cat-C sample contained crystal particles with relatively clear and regular shapes and no evidence of amorphous SiO₂. The NMR and HRTEM results are thus consistent with the XRD results. Based on the results described above, it could be inferred that the amorphous SiO₂ binder in the extrudate was transformed into ZSM-11 in the DGC process, and the shaped binderless zeolite catalyst was therefore synthesized successfully.

The initial cylindrical shape of the catalyst was retained during the DGC process; Fig. 3 shows a photograph of the Cat-C sample. The side pressure strengths of Cat-A, Cat-B, and Cat-C were 78, 110, and 94 N/cm, respectively, which indicates that Cat-C retained its high mechanical strength after recrystallization. This may be attributed to the formation of a matrix or bridge structure, as reported in Ref. ^[21]. In this previous study MFI with a low SiO₂/A1₂O₃ molar ratio of 80 was mixed with amorphous SiO₂ to give small crystals of MFI with a high SiO₂/A1₂O₃ molar ratio of 900 using the DGC method. The second type of zeolite particles bound to the first type of zeolite particles by adhering to their surface, thereby forming a matrix or bridge structure. This resulted in zeolite particles with a high mechanical strength.

N₂ adsorption-desorption isotherms with a steep rise at P/P₀ < 0.01 are characteristic of microporous zeolites, and a hysteresis loop at 0.44 < P/P₀ < 1 corresponds to the existence of dissimilar mesopores ^[22]. As shown in Fig. 4, all of the catalysts showed typical type I + IV isotherms. Data on the textural properties of the catalysts are summarized in Table 1.

In comparison with HZSM-11, Cat-A and Cat-B had lower micropore surface areas and micropore volumes; this was a result of the presence of the amorphous binder (silica or alumina). In contrast, HZSM-11 and Cat-C had similar micropore surface areas and micropore volumes as a result of their similar crystallinities.

The acidity of the catalysts was investigated using the NH₃-TPD technique, and the spectra are shown in Fig. 5. Two NH₃ desorption peaks were observed in the tested temperature range. The desorption peak temperatures for Cat-B were centered at approximately 250 and 423 ℃, and correspond to weak and strong acid sites, respectively. Those for the other catalysts (HZSM-11, Cat-A, and Cat-C) were centered at approximately 239 and 442 ℃. The total concentration of acid sites followed the order: Cat-B (486 μmol/g) > HZSM-11 (453 μmol/g) > Cat-C (395 μmol/g) > Cat-A (362 μmol/g). The acidity of the sites was then further investigated with the Py-IR technique. The bands at approximately 1540 and 1450 cm⁻¹ were integrated to determine the respective concentrations of Brönsted and Lewis acid sites ^[23]. Conventionally, the total acidity is determined when pyridine is desorbed at 150 ℃ (Fig. 6(a)), whereas medium and strong (M-S) acidity is obtained when pyridine is desorbed at 300 ℃ (Fig. 6(b)). To determine the concentration of Brönsted and Lewis acid sites (marked C_B and C_L, respectively), we used the formulae C_B = 1.88 IA (B) R²/W and C_L = 1.42 IA (L) R²/W ^[24]. The different strength C_B and C_L values for each sample are listed in Table 1. Cat-A showed the lowest total and M-S C_B; this is ascribed to the diluting effect of SiO₂, which reduces the acidity. Owing to the presence of Al₂O₃, Cat-B displayed the highest C_L, whereas HZSM-11 and Cat-C possessed a similar C_B. In comparison with that of Cat-A, the C_B of Cat-C was increased; this may be the result of the small amount of additional Al being converted into framework Al and replacing Si-OH in HZSM-11.

The performance of the Cat-A, Cat-B, Cat-C, and HZSM-11 catalysts in the direct amination of isobutene was compared, and the isobutene conversion and tert-butylamine selectivity of the four samples are summarized in Table 2.

Under reaction conditions of 250 ℃, 5.0 MPa, weight hourly space velocity (WHSV) of isobutene of 0.5 h⁻¹, NH₃/i-C₄H₈ molar ration of 4/1 and time-on-stream (TOS) of 6 h, the isobutene conversion was in the range of 10.41%-13.04% over the catalysts employed. Cat-A displayed higher isobutene conversion than Cat-B (12.16% vs 10.41%), and the conversions obtained for the reactions over Cat-C (13.04%) and HZSM-11 (12.93%) were similar. As shown in Table 2, the major product was tert-butylamine (> 99.9%), and the remaining trace components were C₈ hydrocarbons. A small amount of oligomerization and isomerization products were observed, but the selectivities for both of these were less than 0.1%. This result is in good agreement with those obtained with H-FAU and H-MOR zeolites. Reaction over these zeolites showed a tert-butylamine selectivity of more than 99% below 280 ℃ with a NH₃/i-C₄H₈ molar ratio of 2 ^[25]. The H-Y catalyst, under conditions of 260 ℃, 5.2 MPa, and an NH₃/i-C₄H₈ molar ratio of 1, gave an isobutene conversion of 8.9% with a tert-butylamine selectivity of 98% ^{[25, 26]}. Furthermore, BASF have commercialized the BEA zeolite catalyst for the direct amination of isobutene; however, this process requires relatively harsh conditions of 270 ℃, 28.0 MPa, and a NH₃/i-C₄H₈ molar ratio of 1.5. In addition the conversion of isobutene was only 22.67% ^[27].

As shown in Fig. 7, the formation rate of tert-butylamine changed very little over the 6 h reaction time. A similar phenomenon was observed for the MFI zeolites ^[2]. Furthermore, as shown in Fig. 7, the formation rate of tert-butylamine at TOS = 6 h over the different zeolite samples decreased in the following order: Cat-C (1162 μmol h⁻¹ g⁻¹_cat) ≈ HZSM-11 (1152 μmol h⁻¹ g⁻¹_cat) > Cat-A (1083 μmol h⁻¹ g⁻¹_cat) > Cat-B (927 μmol h⁻¹ g⁻¹_cat). HZSM-11 and Cat-C showed similar performance in the isobutene amination reaction, but HZSM-11 is a powder whereas Cat-C is shaped and has high mechanical strength (94 N/cm). Cat-C therefore has the potential for use in industrial applications. Academic and fundamental research studies have been neglected in this field ^[4] because the design of shaped formulations is generally considered to be a technical process. However, a solid understanding of the relationship between catalytic performance and catalyst structure and composition, including formulation shape, is necessary for the development of practical zeolite catalysts ^[28].

The textural properties and acidity of ZSM-11 may influence its catalytic performance. First, we will discuss the effect of the textural properties on catalytic performance. As shown in Table 1, A_BET decreased in the order HZSM-11 > Cat-C > Cat-B > Cat-A, whereas A_micro decreased in the order Cat-C ≈ HZSM-11 > Cat-B > Cat-A. V_micro followed the sequence HZSM-11 = Cat-C > Cat-B > Cat-A, and V_meso decreased in the order Cat-B > HZSM-11 > Cat-A > Cat-C. Thus, the formation rate of tert-butylamine does not appear to show a clear relationship with the textural properties of ZSM-11. Li ^[29] reported that hierarchical nano-ZSM-5 zeolites with a high surface area showed little improvement in catalytic performance compared with that of nano-ZSM-5 for the direct amination of isobutene to tert-butylamine. This demonstrates that molecular diffusion is not the rate determining step.

Second, we illustrate the effect of acidity on catalytic performance. In our study, the total acid site concentration (determined using NH₃-TPD) decreased in the order Cat-B > HZSM-11 > Cat-C > Cat-A. As shown in Table 1, C_B decreased in the order HZSM-11 ≈ Cat-C > Cat-B > Cat-A, whereas the order for C_L was Cat-B > HZSM-11 > Cat-A > Cat-C, irrespective of the total and M-S acid site concentrations. It could thus be speculated that the formation rate of tert-butylamine is closely related to C_B ^[29-31].However, C_L should also be considered. Even if a catalyst has a suitable C_B, a high C_L may inhibit the formation of the ammonium cation, which reacts with gaseous isobutene. The resulting cationic intermediate, the tert-butyl cation, then reacts with NH₃, either adsorbed on the catalyst surface or in the gas phase, and produces chemisorbed tert-butylamine ^{[2, 16]}.

ZSM-5 is a common catalyst for the direct amination of isobutene, and is considerably affected by the number and strength of Brönsted acid sites ^[30]. As a result, Ce-modified HZSM-5 catalyst showed improved activity in the amination reaction ^[31]. Chang et al. ^[32] reported that the catalytic performance was related to the concentration of medium strength acid sites detected by NH₃-TPD.

The relationship between the number of Brönsted acid sites and the SiO₂/A1₂O₃ molar ratio was investigated by Mizuno et al. ^[30]. These authors reported that the amination performance was not simply related to the number of Brönsted acid sites. In our study, the ²⁹Si MAS NMR technique was used to determine the SiO₂/Al₂O₃ molar ratios in the framework of each sample, and these ratios were 61.6, 61.6, 61.8, and 73.0 for HZSM-11, Cat-A, Cat-B, and Cat-C, respectively. Using this information and the results in Table 1, the relationship between the turnover frequency per Brönsted acid site and the SiO₂/A1₂O₃ molar ratio of the catalyst was investigated. It was clear that the total turnover frequency could not be expressed by a linear relationship and did not increase on increasing the SiO₂/Al₂O₃ molar ratio of the catalyst framework. This result is not consistent with that reported in Ref. ^[30].

A shaped binderless ZSM-11 zeolite catalyst was successfully prepared using the DGC method in its protonic form (Cat-C). When using suitable synthesis conditions, the concentration of Brönsted acid sites in Cat-C can approach the concentration in HZM-11. Compared with the binder-containing catalysts (Cat-A and Cat-B), Cat-C showed a higher formation rate of tert-butylamine in the direct amination of isobutene. This was a result of the higher concentration of Brönsted acid sites and lower concentration of Lewis acid sites. Because of its high crystallinity and good mechanical strength, the binderless ZSM-11 zeolite catalyst has potential for application in the direct amination of isobutene.