Article pubs.acs.org/cm
Controlling the Aluminum Distribution in the Zeolite Ferrierite via the Organic Structure Directing Agent Ana B. Pinar,*,†,‡ Luis Gómez-Hortigüela,† Lynne B. McCusker,‡ and Joaquín Pérez-Pariente† †
Institute of Catalysis and Petroleum Chemistry−CSIC, c/Marie Curie 2, ES-28049, Madrid, Spain Laboratory of Crystallography, ETH Zurich, CH-8093 Zurich, Switzerland
‡
S Supporting Information *
ABSTRACT: Analysis of the structure of the zeolite ferrierite (framework type FER) synthesized using a combination of tetramethylammonium (TMA) and pyrrolidine as organic structure-directing agents (|((CH3)4N)0.4((CH2)4NH)3.6H1.7| [Si33.9Al2.1O72]) has revealed that TMA ions reside exclusively in the [586682] cavities of the FER framework, whereas pyrrolidine species are found in both the [586682] cavities and the main 10-ring channel. A similar, but not identical, arrangement of pyrrolidine molecules was found for a sample synthesized with pyrrolidine as the only SDA (|((CH2)4NH)4.2H2.2|[Si33.8Al2.2O72])). A comparison of the two structures shows the influence of TMA on the location of pyrrolidine and the framework Al. Pyrrolidine species establish stronger interactions with the zeolitic framework than does TMA, suggesting a different ability of the amine to exert an influence on the aluminum distribution of the zeolite. These findings have implications in the catalytic performance of the samples. KEYWORDS: ferrierite, FER, SDA location, Rietveld refinement, pyrrolidine, TMA, control of acid site distribution
■
INTRODUCTION The three-dimensional frameworks of zeolites are composed of corner-sharing tetrahedral [SiO4] and [AlO4] units. The isomorphic substitution of Si(IV) by Al(III) in tetrahedral (T) sites of the framework generates a negative charge that, when it is counter-balanced by a proton, constitutes a Brönsted acid site. The distribution of Al in a zeolite framework is not necessarily random and may vary in a given zeolite depending on its preparation and postsynthesis modification. The influence of synthesis parameters on acid-site distribution has been studied for several zeolites.1−7 In particular, Dedecek et al. showed in an elegant study that the sources of Si and Al used in the synthesis can alter the Al distribution in Al−O−(Si−O)n− Al sequences in the MFI, FER, *BEA, and MWW frameworks.8−10 However, tuning the location of the acid sites remains an elusive dream in zeolite synthesis. It would allow the acid-site distribution of a zeolite catalyst to be tailored for a specific reaction, but the complexity of this task lies in the intricacy of the hydrothermal process itself, in which many variables interact in a delicate equilibrium. For example, the initial inorganic building blocks, the organic SDA species, the alkaline cations, the water molecules, the pH and the temperature all affect the end product. The crucial role played by the organic species in the distribution of Al was recognized by Sastre et al.11 as early as 2002. With this in mind, we designed a new approach to the synthesis of the zeolite ferrierite (framework type FER) with the aim of controlling the aluminum location.12,13 The framework structure of this medium-pore zeolite has 8-ring channels running parallel to the a axis and these intersect 10© XXXX American Chemical Society
ring channels running parallel to the b axis (Figure 1). Between these intersections along the 8-ring channel, a [826658] cavity that is only accessible through the 8-ring-openings is formed. By employing different combinations of SDAs, we were able to produce a series of ferrierite samples with different distributions of strong Brönsted acid sites.12−16 We proposed that the organic SDAs could indeed drive the aluminum atoms toward certain T positions provided (1) a strong interaction between the organic cation and the framework oxygen atoms surrounding Al (SDA+···OAl−) can be forged, and (2) the organic SDAs occupy well-defined positions within the void volume of the framework structure. Then the Al distribution could be expected to be determined by the location of the SDA species within the zeolite framework and their tendency to attract Al atoms during the synthesis. Much effort has been devoted to the complex topic of analyzing the siting and distribution of Al atoms in the framework of Si-rich zeolites. Dedecek et al.17 recently published a comprehensive review, in which they describe the use of several techniques, including UV−vis spectroscopy of ion-exchanged Co2+-zeolites,23Na MAS NMR of Na-zeolites and amine adsorption monitored by FTIR. 27Al MAS NMR is also widely applied because the chemical shifts are related to the average Si−O−Al angles in the structure.18 However, the shifts can be affected by other factors, and assigning NMR Received: June 4, 2013 Revised: July 21, 2013
A
dx.doi.org/10.1021/cm4018024 | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
washed with ethanol and water, and dried at room temperature overnight. Given the amount of aluminum in the framework, some of the pyrrolidine molecules are expected to be protonated. Hereafter, the abbreviation pyrr should be understood to refer to either pyrrolidine molecules or pyrrolidinium ions. The nature of the crystalline phases obtained was determined from X-ray diffraction data (PANalytical X́ Pert PRO−MPD diffractometer, CuKα radiation). The organic content was characterized by thermogravimetric analysis (TGA) (Perkin−Elmer TGA7 instrument, heating rate 20 °C/min, air flow 30 mL/min, temperature range 30− 900 °C) and chemical CHN elemental analysis (Perkin−Elmer 2400 CHN analyzer). The crystal morphology was determined by scanning electron microscopy (JEOL JSM 6400 Philips XL30 electron microscope, operating at 20 kV). The composition of the samples was determined by ICP−AES in an ICP Winlab Optima 3300 DV Perkin−Elmer spectrometer. 19F MAS NMR spectra were collected with a Bruker AV 400 WB spectrometer, using a BL2.5 probe, using pulses of π/2 rad of 4.5 μs and delays of 80 s between two consecutive pulses, while spinning the sample at 20 kHz. For the structure analysis of FER−PYRR−TMA, powder diffraction data were collected on the Materials Science Beamline at the Swiss Light Source (SLS) synchrotron facility in Villigen, Switzerland,25 and for FER−PYRR, with a laboratory diffractometer (Stoe STADI P, CuKα1 radiation). Data collection details are given in Table 1. The
Table 1. X-ray Powder Diffraction Data Collection Parameters for FER−PYRR−TMA and FER−PYRR
Figure 1. Framework structure of ferrierite viewed down the [010] direction (top) with side views (bottom) of the 10-ring channels (red) and the ferrierite cavity (green). Bridging oxygen atoms have been omitted for clarity.
synchrotron facility beamline diffraction geometry detector monochromator wavelength sample nominal step size detector positions time per pattern 2θ range
signals to Al in particular T sites remains difficult, despite the progress made with multiquantum MAS experiments.19−24 Generally, diffraction techniques cannot distinguish between Si and Al, because the scattering factors of these two elements are too similar. The longer T−O distances and smaller T−O− T angles for Al can sometimes be used to deduce its location, but the distances and angles for a given T site will be proportional to the relative Si and Al occupancies (in general, T sites are not pure Al), so these differences can be subtle. However, indirect evidence regarding the siting of the Al atoms can be obtained from the positions of the SDA species. In this case, diffraction techniques can be powerful, because they are one of very few characterization tools that can be used to obtain information about the location of the organic species within the void volume of a zeolite, provided that these species are ordered. Here we report the application of this concept to two ferrierite samples prepared following our strategy to tune Al siting.
■
diffractometer radiation sample step size 2θ range counting time 3−27.6°2θ 27.6−50°2θ 50−90°2θ
EXPERIMENTAL AND COMPUTATIONAL METHODS
FER−PYRR−TMA SLS Material Science Debye−Scherrer Mythen II Si microstrip Si 111 1.000 Å rotating 0.5 mm capillary 0.004 °2θ 5 10 s 4.5−70 °2θ FER−PYRR Stoe STADI P CuKα1 rotating 0.3 mm capillary 0.01 °2θ 3.0−90 °2θ 1 s/step 5 s/step 10 s/step
Rietveld refinement was performed using the program package XRS82,26 and the structure drawings were produced using the program CrystalMaker.27 The plots with observed, calculated, and differences patterns were made using the program ppp13.28 Modeling of the structure directing effect of pyrrolidine and TMA was performed using molecular mechanics simulations. The molecular structures of the SDAs and their interaction energies with the framework were described using the CVFF forcefield29 under periodic boundary conditions. The geometry of the ferrierite framework was kept fixed during the calculations. A single unit cell of the ferrierite framework system with one TMA or pyrrolidinium cation loaded in the required position was used. To balance the negative charge created by the aluminum atoms in the framework, some of the pyrrolidine molecules were expected to be protonated, so protonated pyrrolidinium ions were studied. The atomic charge distribution for the two SDAs was calculated using the charge-equilibration method,
Ferrierite samples were synthesized in fluoride medium from gels with the following molar composition: 0.94 SiO2:0.03 Al2O3:x TMA: (0.54−x) pyrrolidine: 0.48 HF:4.6 H2O, with x = 0.06 (sample FER− PYRR−TMA), and x = 0 (sample FER−PYRR). Tetraethylorthosilicate (TEOS, Merck, 98 wt %) and aluminum isopropoxide (Fluka, 97 wt %) were added to an aqueous solution of the organic SDAs, pyrrolidine (Aldrich, 99 wt %) and tetramethylammonium hydroxide (Sigma−Aldrich, 25 wt %). The mixture was stirred until all the ethanol produced by the hydrolysis of TEOS and the excess of water had evaporated. Hydrofluoric acid (Panreac, 48 wt %) was then added, and the resulting thick gel was homogenized manually (pH ca. 8.6) and transferred to 20 mL Teflon-lined stainless steel autoclaves, which were then heated under static conditions at 150 °C and autogenous pressure for selected periods of time. The solid products were filtered, B
dx.doi.org/10.1021/cm4018024 | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
setting the total net molecular charge for each SDA to +1; this net SDA charge was balanced by using a modified version of the uniform charge background method for the framework,30 where the atomic charge for every silicon atom is reduced until charge neutrality is reached. The atomic charge for oxygen was fixed at −1.2. The location of the TMA and pyrrolidinium ions with the ferrierite framework in the two different locations (within the cavity and in the 10-ring channels) were obtained by geometry optimization, and the resulting interaction energies were calculated by subtracting the energy of the isolated ions optimized in vacuo from the total energy of the system. All the interaction energies are expressed in kcal/mol.
Rietveld method combined with difference Fourier techniques was applied. Structure Analysis. Sample FER−PYRR−TMA. Refinement of the structure of FER−PYRR−TMA (obtained after 10 days of hydrothermal treatment) was performed in the orthorhombic space group Immm, which is the highest one possible for the FER framework type. The structure of the natural aluminosilicate form from single-crystal data was first determined in this space group,34 and it was subsequently confirmed by other studies.35,36 However, Yokomori et al. used I222,37 and several other subgroups have also been reported for different ferrierite samples.32,38−42 A thorough examination of the diffraction data for FER−PYRR−TMA did not reveal any violations of the body-centering condition; nevertheless, refinement in Pnnm was also tried to be sure. No deviation from Immm symmetry was observed, and hence Immm was used thereafter, even though one of the oxygens lies on an inversion center in this space group, causing the associated Si−O−Si angle to be 180°. Structure refinement was initiated using the atomic coordinates of the framework atoms reported by Morris et al.40 in Pnnm. The coordinates were symmetrized to Immm and then the geometry was optimized using a distance-least-squares procedure. All the atoms in tetrahedral coordination were modeled as Si. The first difference Fourier map revealed the presence of electron density clouds in the middle of the ferrierite cavity and along the 10-ring channel. A certain amount of chemical interpretation was required in order to fit the organic molecules into these electron density clouds, since the organics do not follow the high symmetry of the FER framework, and as a consequence, TMA and pyrr are disordered. This means that not all of the atoms of these molecules can occupy all symmetry equivalent positions simultaneously, so their occupancies are necessarily fractional (see Figure 2). Furthermore, as already mentioned, both TMA and pyrrolidine can be accommodated at the different sites.
■
RESULTS Synthesis. Pure ferrierite is obtained at 7, 10, or 20 days of hydrothermal treatment of the gel containing either pyrrolidine and TMA or pyrrolidine only as SDA. Both FER−PYRR−TMA and FER−PYRR have an organic content of ca. 12 wt % and a very low water content (
