The Use of Original Structure-Directing Agents for the Synthesis of EMC-1 Zeolite

— By using a steric approach, it is possible to design templates matching closely with an inorganic framework. Herein, the design, synthesis and successful application of several di(azacrown


INTRODUCTION
Zeolites are well-defined microporous aluminosilicate materials obtained by the polymeric combination of TO 4 tetrahedra. Their unique design promotes the creation of multidimensional channels and cavities, responsible for their interesting adsorption and diffusion properties [1][2][3]. Combined with their intrinsic Bro¨nsted acidity and high thermal stability, zeolites are nowadays widely used in petrochemistry and oil refining processes [4][5][6][7][8]. Particularly, FAU-type zeolites are one of the main components of cracking catalysts at the industrial scale due to their structure, one of the most open among zeolites: its porosity is composed of supercages, with a free diameter of 11.6 Å , interconnected through circular 12-Member-Ring (MR) apertures with a diameter of 7.4 Å [9].
Conventional Y zeolite (FAU-structure type), with a silicon-to-aluminum molar ratio between 1.5 and 2.5, lacks of thermal and hydrothermal stabilities [10]. Nowadays, high silica ultra-stable Y zeolite, with silicon to aluminum molar ratio generally above 4, is used for cracking applications, allowing good thermal and hydrothermal stabilities [11,12]. Mainly, ultra-stable Y zeolite is produced from dealumination of a conventional Y zeolite by hydrothermal or chemical treatment [13][14][15]. Unfortunately, these methods require several post-treatments such as calcination under steam (steaming), acid leaching, filtration, washing and drying: this explains the strong interest of a direct "one step" synthesis of high-silica FAU-type zeolites for industrial applications [10].
High silica FAU-type zeolite, with a silicon to aluminum molar ratio between 3 and 5, can be synthesized directly by using specific Structure-Directing Agents (SDA) in a synthesis hydrogel. Thus, Guth et al. reported the crystallization of a high silica FAU-type zeolite commonly named EMC-1 (Elf Mulhouse Chemistry One) by using the 15-crown-5 ether as SDA [16][17][18]. The catalytic performances of FAU-type zeolite have been widely studied.
In order to reduce the cost of EMC-1 zeolite, attempts to search for less expensive SDA to direct the formation of high-silica FAU-type zeolites were conducted, but did not afford pure phase. This might arise from a low fit between the SDA and the zeolite structure.
Herein, using a similar steric approach new di(azacrown ether)-based templates were designed to favor the formation of EMC-1 zeolite. Following a High Throughput Experiment (HTE) design, the synthesis gel composition was optimized to obtain well-crystallized EMC-1 zeolite. The obtained products were fully characterized.

Molecular Modeling
Templates geometry optimization was performed in purely siliceous FAU frameworks using Materials Studio (MS) 5.5 software following a similar steric approach described elsewhere [21,24]. The energy minimized locations of SDA within theoritical purely siliceous FAU-type structure were obtained by molecular modeling. Energy minimization was performed using Monte-Carlo simulated annealing runtines in the program Discover (implemented in MS Modeling 5.5 by Accelerys Inc.). van der Waals interactions were calculated using the universal forcefield [26]. In this work, the "host-guest" interaction energy (bonding energy) is defined as the difference between the total energy of a system where a template is occluded inside a zeolite framework, and the sum of energies of the template in the vacuum and the zeolite in vacuum: The deformation energy is defined as the difference between the energy of a template sterically constricted inside a zeolite framework, and the energy of the template in vacuum.

Preparation of Di(Azacrown Ether) Templates (ACn-m )
During this study two templates (AC5-4, AC5-5), designed by molecular modeling, have been synthesized in order to study the spacer length between the macrocycles upon zeolite EMC-1 formation. A schematic representation of those templates is given in Figure 1. They were synthesized in two steps: the first one consists in a diacylation of the aza-crown ether by a diacyl chloride, followed by the reduction of the diamide with lithium aluminum hydride (LiAlH 4 ). The experimental conditions of all these syntheses being similar, only the preparation of AC5-4 (M w = 492 g/mol), molecular formula C 10 H 20 O 4 N-C 4 H 8 -NO 4 H 20 C 10 ), will be detailed thereafter.

Synthesis of AC5-4 Compound
In a typical synthesis, 4.17 mmol (2.5 eq) of 1-aza-15crown-5 ether (molecular formula C 10 H 21 NO 4 , Alfa Aesar) and 6.72 mmol of triethylamine (Alfa Aesar) were dissolved in 11 mL of freshly distilled dichloromethane (Alfa Aesar). The mixture is first maintained at 0°C in an ice bath during dropwise addition of 0.187 mL (1.69 mmol, 1 eq) of succinyl chloride (Alfa Aesar), and then stirred at room temperature for 15 hours. The solvent is removed under reduced pressure and the residue is mixed in 15 mL of ethyl acetate (VWR Intl) and then filtered. The filtrate, containing the desired diamide and the residual 1-aza-15-crown-5 ether, is purified by column chromatography on silicagel (acetone/triethylamine elutent, 95:5 v/v). Then 1.4 mmol (1 eq) of the desired diamide and 2.8 mmol (2 eq) of LiAlH 4 (Alfa Aesar) were dissolved in 12.6 mL of freshly distilled THF (Alfa Aesar) and refluxed for 14 hours. The mixture is treated first with 0.5 g of Na 2 SO 4 .10H 2 O and then with 1.6 mL of NaOH 1 M. The crude residue was washed three times with 30 mL of ethyl acetate. The crude was then dried with MgSO 4 and concentrated under reduced pressure. The final product was identified by 1 H and 13 C NMR, which were taken from CDCl 3 at 22°C using a Bruker Instrument. For each chemical shift the number of protons for 1 H NMR and Carbon atoms for 13 C NMR and their positions (underlined) were assigned. 1 H NMR (d): (400 MHz): 1.38 (4H, NCH 2 -CH 2 in the linear chain); 2.46 (4H, NCH 2 in the linear chain); 2.72 (8H, NCH 2 in the azacrown ether); 3.61 (32H, OCH 2 in the azacrown ether). 13 C NMR (d): (100.6 MHz): 24.7 (2C, NCH 2 -CH 2 in the linear chain); 54.3 (4C, NCH 2 in the azacrown ether); 56.5 (2C, NCH 2 in the linear chain); 69.6, 69.8, 70.1, 70.6 (16C, OCH 2 in the azacrown ether).

Synthesis of AC5-5 Compound
AC5-5 (M w = 506 g/mol, molecular formula C 10 H 20 O 4 N-C 5 H 10 -NO 4 H 20 C 10 ), was synthesized following the same procedure using glutaryl chloride (Alfa Aesar) instead of succinyl chloride. The product was then identified by 1 H and 13 C NMR, which were taken from CDCl 3 at 22°C using a Bruker Instrument. For each chemical shift the number of protons for 1 H NMR and Carbon atoms for 13

Synthesis of EMC-1 Zeolite
The 15-crown-5 (C5) and 1-aza-15-crown-5 (AC5) were used as received from Alfa Aesar. A HTE design was applied to the study, aiming to afford pure FAU-type zeolites by varying Na 2 O and H 2 O molar contents in initial gel compositions. The gel preparations and hydrothermal syntheses were performed using a Tecan Freedom EVO liquid handling robot and a Top Industry high-throughput heated and stirred multi-autoclave unit (100 Monel autoclaves, 5 mL), respectively. First, an appropriate amount of template was homogenized in deionized water. Sodium hydroxide (Prolabo, 99%), then sodium aluminate (Carlo Erba, 56% Al 2 O 3 , 37% Na 2 O) were mixed under magnetic stirring until the solution became clear. Finally, Ludox AS-40 colloidal silica (Aldrich, 40% in water) was added dropwise, so that the molar gel composition for template-mediated FAU syntheses were 10 SiO 2 :Al 2 O 3 :x Na 2 O:y H 2 O:z template; with x = 2.1, 2.4 or 2.7, y = 100, 140 or 180 and z = 0.6 (di(azacrown ether) template, noted AC5-m with m = 4 or 5 and corresponding to the number of carbon atoms between two macrocycles) or 1 (crown ether template, noted C5, or azacrown ether template, noted AC5). The gel was stirred at 200 rpm during 24 hours, prior to be heated at 110°C for 15 days in static conditions. After synthesis, the product was filtrated, washed with de-ionized water and dried overnight at 100°C. The templates were finally removed by calcination in a muffle furnace at 550°C during 8 hours in air (ramp at 0.5°C/min).

Characterization
Liquid 1 H NMR (400 MHz) and 13 C (100.6 MHz) spectra were recorded with a Bruker Advance 400 spectrometer, in CDCl 3 with Me 4 Si as internal standard. Powder X-ray diffraction patterns were collected on a Bruker D4 Endeavor diffractometer equiped with CuK a1 monochromatic radiation source (40 kV, 40 mA). Diffraction patterns in the 2-40°region were recorded with a 0.02°s tep size (step time = 1 s). After background subtraction, peak areas in the 19-32°2h range (hkl) = 440, 533, 642 and 555 were used to determine the crystallization rate of the as-synthesized powder, by comparison to the peak areas of the most crystalline EMC-1 sample prepared in this work in the presence of C5. The accuracy is ±5%. Scanning Electron Microscopy (SEM) was performed using a Jeol JSM-6340F, under an acceleration voltage of 1 kV. Samples were prepared by depositing a drop of solution containing the sample dispersed in ethanol on an aluminum mounting. Statistical distributions of crystal size were carried on populations of 100 crystals. ThermoGravimetric Analysis (TGA), Differential Thermal Analysis (DTA) were carried out using a TA Instrument SDT 2960 thermoanalyzer between the room temperature and 800°C, at a heating rate of 5°C/ min. Reconstituted air (80% N 2 and 20% O 2 ) with a flow rate of 1 L/h was used during measurement. Nitrogen sorption measurements were performed at À196°C on a Micromeretics ASAP 2010 apparatus. Samples were degassed under primary vacuum at 350°C during 10 hours prior analysis. Surface area was calculated according to the BET method [27]. Micropore volume and external crystal surface were determined by using the t-plot method [28].

Rational Design of Di(Azacrown Ether) Templates by Molecular Modeling
It has been shown in previous works that potential new templates for a specified zeolite framework should at least have similar "host-guest" interaction energies to the usual structure-directing agent of the framework, as a lack of "host-guest" interactions might hinder the inclusion of the template or direct the formation of other frameworks [21]. These potential templates should also have the lowest possible deformation energy, meaning that their geometry closely fits to the inner porosity [24]. Regarding this, Figure 2 presents the geometry optimization results of several di(azacrown ether) molecules (AC5-m) in the FAU framework, compared to the conventional structure-directing agent of EMC-1 zeolite, the 15-crown-5 ether (C5), and the 1-aza-15-crown-5 ether (AC5). As expected, the "host-guest" interaction energy is favorable for all the evaluated AC5-m, due to the presence of hydrocarbon spacers linking two structure-directing functions. This, in turn, increases hydrogen bonds between the template and the framework. Deformation energies of the AC5-m templates with spacers composed by three to five carbons are comparable and below 5 kcal/mol, which is close to the C5 deformation energy, making them an optimal fit for FAU framework. Below and beyond three to five carbons, the deformation energy is clearly unfavorable, meaning that templates can hardly accommodate or fit into the framework. From molecular modeling, it appears that an optimal fit is obtained when a four-carbons spacer length links two structure-directing functions, noted AC5-4, even if it's deformation energy is higher than their crown ether and azacrown ether counterpart. A second AC5-m, with a spacer composed by five carbons, noted AC5-5, gives also similar results and will also be used for zeolite synthesis to evaluate the impact of the spacer length on the crystallinity of the material.

Effect of the Synthesis Gel Composition
Preliminary experiments showed that a variation of the SiO 2 /Al 2 O 3 molar ratio in the presence of C5 promotes the crystallization of P (Gismondine type (GIS-type) structure) and A Linde Type A (LTA) zeolites. Thus, the SiO 2 /Al 2 O 3 molar ratio should be kept constant at the typical value of 10 during this study to avoid the formation of these undesired phases. The molar content of template was also modified until no further increase of the sample crystallinity was observed.
Results of the applied experimental design are summarized in Figure 3. At high Na 2 O molar content, co-crystallization of P zeolite (GIS) generally occurs. At the opposite, no crystallized material is detected after 15 days at low Na 2 O molar content: lowering the concentration of mineralizing agent (OH À ) usually induces a lower sursaturation, resulting in a less reactive hydrogel. As expected, C5 shows the most specific structuredirecting effect amongst all the tested templates. It leads to the formation of well-crystallized EMC-1 zeolite, from a wide range of gel compositions. This is in accordance with previous works [16][17][18].
The use of AC5 allows to obtain a FAU-type zeolite from several gel compositions, but P zeolite also co-crystallizes in a wide range of gel composition, which might be interpreted as a lower structure-directing effect of AC5 compared to C5. Previous studies concluded that the structure-directing agent of EMC-1 zeolite was not the crown ethers alone but rather the association of crown ethers with sodium cations as a complex with a planar conformation [29][30][31]. The AC5/sodium cation complexe is expected to adopt a curved conformation [32]. Then, the lack of structure selectivity observed might arise from the presence of nitrogen elements in the templates.
Observations were sensibly identical when using di(azacrown ethers), except that the proportion of amorphous material obtained is more important at low Na 2 O molar content. Moreover, as it will be seen below (thermal analysis and N 2 adsorption measurements) for the AC5-5 sample, amorphous material is always present besides the crystalline phase. Therefore, for this sample it is better to talk of crystallization rate rather than relative crystallinity.
The X-ray diffraction patterns of FAU-type zeolites synthesized in the presence of the different templates and selected for further characterizations are presented, in Figure 4. Every template possesses a relatively good structure-directing effect, leading specifically to FAUtype zeolite within close synthesis gel compositions. Particularly, the relative crystallinity of AC5 (85%) match to the crown ethers C5 (100%, reference). The crystallization rate of the sample synthesized in the presence of AC5-5, reaching 60% (Fig. 4d), is clearly lower than the crystallinity of the sample synthesized with AC5-4, up to 95% (Fig. 4c), indicating that the 4-carbons spacer fits better to the EMC-1 structure than a 5-carbons spacer as initially observed by molecular modeling.
Templates integrity was verified by 1 H liquid NMR, after dissolution of the inorganic framework in a HF solution, by comparaison of the NMR spectrum with the spectrum of the same template before synthesis. These crystallized materials were further characterized.   Figure 5 presents SEM images of the as-synthesized samples prepared with C5, AC5 and AC5-m as structuredirecting agents. A typical octahedral shape, attributed to FAU-type crystals, is observed for each one. One may notice that in the presence of AC5-m, the synthesized crystals were sensibly smaller. This observation was further verified by studying the statistical distributions of crystal sizes, summarized in Table 1. Indeed, in comparison with the crystals synthesized in the presence of C5 (Fig. 5a) a 10% size reduction is observed when using AC5-4, with a slightly more uniform distribution (Fig. 5c), and reaches the 32% when using AC5-5 (Fig. 5d ). However, the crystallization rate of the latter (60%) indicates that this size reduction could be related to an incomplete growth.

Characterization of the Crystallized Samples
Thermal analyses of FAU-type materials are presented in Figure 6. Below 200°C, a first weight loss corresponding to water molecules desorption and characterized by an endothermic peak on the DTA curve (not shown) is observed. Between 200 and 650°C, several   Fig. 4d). Figure 7 presents the nitrogen adsorption-desorption isotherms of the different materials. A typical type I adsorption isotherm is obtained in each case, indicating that the materials were microporous. Additional H4 hysteresis can be observed on most of the isotherms. This generally occurs in the presence of narrow slit-like mesopores, responsible for capillary condensation. Table 2 resumes the textural properties of the chosen crystallized materials. A very similar microporosity was observed in each case, apart for AC5-5 confirming thus the presence of amorphous material in this sample.

CONCLUSION
Molecular modeling was used to design new templates favorable to the synthesis of EMC-1 zeolite. An optimal fit was obtained when a four-carbons spacer links two azacrown ether molecules. A high throughput experiment design was applied to optimize the synthesis gel composition in the presence of the mono or di(azacrown ether) templates. EMC-1 zeolite was obtained in restricted gel composition domains. This work paves the way for the further design of complex templates with the ability to finely tailor both structural and textural properties of FAU-type zeolites.  Nitrogen adsorption/desorption isotherms at À196°C of the calcined FAU-type zeolites.