HEOPTAHHHECKHE MATEPHAMbI, 2007, moM 43, № 7, c. 853-864
UDC 548.32
A Study of the Chemistry of Isomorphous Substitution and Characterization of Al-ZSM-5 and Sc-ZSM-5 Synthesized in Fluoride Media
© 2007 r. C. T. Brigden*, C. D. Williams*, D. Apperley**
*School of Applied Sciences, University of Wolverhampton, UK e-mail: C.Brigden@wlv.ac.uk **Department of Chemistry, Durham University, UK Received 21.11.2006
Al-ZSM-5 and Sc-ZSM-5 samples have been synthesised using fluoride media at around neutral pH. The synthesis chemistry has been studied and it is shown how the lower tendency of scandium to undergo increased fluoro-complexation (compared with aluminium) coupled with its tendency to hydrolyse is conducive to its isomorphous substitution and framework incorporation into the MFI zeolite structure. Clear unit cell volume expansion, elongation of the c crystallographic axis with increased scandium content and a strong positive linear correlation between the unit cell volume expansion and the calculated unit cell framework scandium content are shown. Chemical shifts are assigned to tetrahedral and octahedral scandium from 45Sc MAS NMR analysis. Shoulders/a peak at lower wavenumber on the main internal T-O asymmetric stretches in the FT-IR spectra indicate that an interaction exists between framework-incorporated scandium and the SiO4 tetrahedra. It is concluded from the experimental evidence that isomorphous substitution of scandium into the zeolite framework has been achieved.
INTRODUCTION
Isomorphous substitution can be described as the replacement of a silicon atom in the zeolite framework by another atom whilst still retaining that particular zeolite's structure. Most often isomorphous substitution is carried out by the inclusion of a heteroatom reagent source in the synthesis gel and the incorporation takes place during hydrothermal synthesis of the zeolite structure. Barrer includes in his definition of isomorphous substitution in tectosilicates "replacements in the anionic frameworks of Si by such elements as Al, Ga, Ge, Be, B, Fe, Cr, P, Mg effected, if at all, only during synthesis" [1]. Ratnasamy [2] refers to the first synthesis involving isomorphous substitution as being the substitution of germanium in thomsonite by Goldsmith in 1952 [3].
Examples of studies of isomorphous substitution for the group 13 metals into an MFI framework using fluoride media include: Dwyer et al. [4] who studied the isomorphous substitutions of boron, aluminium and gallium into MFI type frameworks; Testa et al. [5] synthesized borosilicalite-1; Aiello et al. [6] studied the influence of alkali cations on the synthesis of ZSM-5; and the influence of alkali cations upon Ga-ZSM-5 synthesis was studied by Nigro et al. [7].
References for the synthesis of any zeolite with the group 3 element scandium are extremely limited. Bull et al. present a study of the synthesis and characterization of scandium phosphate frameworks [8]. They refer to two other recent reports of scandium-containing open frameworks by Bull et al. [9] and Riou et al. [10] but state that prior to these no organically templated
scandium-containing frameworks have been reported. They suggest that this is due to the high commercial cost of scandium. Brigden et al. report the synthesis of isomorphously substituted Sc-ZSM-5 in a fluoride medium [11].
Scandium bears some similarity in its chemistry to the group 13 elements. Nilson discovered scandium in 1879 and confirmed that it was the element eka-boron predicted by Mendeleev in 1869. Nilson stated, "conclusively it is without doubt that with scandium the element eka-boron is discovered" [12].
In the current study some of the synthesis chemistry and the factors which govern the isomorphous substitution of aluminium and scandium are explored and characterization techniques are employed to establish whether scandium has become incorporated into the MFI framework structure. Scandium was chosen as a metal to study in this context given its similar chemistry to group 13 elements, which should theoretically favour its incorporation, and hopefully therefore add to the understanding of isomorphous substitution in the fluoride medium.
EXPERIMENTAL
Reagent proportions and synthesis temperature were based on those used in various studies [13-15]. However, there were some differences from these studies: less water was used (the other studies used 330 and 400 moles H2O) and NaBr was used instead NaCl in order to keep the number of system components to a minimum.
Gels were prepared containing the following stoichiometric reagent proportions:
10SiO2 : 1NaF : 0.25NaBr : xM : 1.25TPABr : 145H2O
M - Al(OH)3 or Sc(NO3)3
For M - Al: x = 0.05, 0.3 and 0.5;
for M - Sc: x = 0.05, 0.1, 0.2, 0.3 and 0.5. In addition a gel using double the fluoride content was prepared:
10Si02 : 1NaF : 1NH4F : 0.25NaBr :
: 0.3M : 1.25TPABr : 145H2O
(because NaF has limited solubility the fluoride content of the gel was doubled by using NH4F in addition to NaF as it is more soluble). The fluoride content was doubled in order to examine the effect upon the crystallization and isomorphous substitution at the gel Si/Sc ratio of 33. For example, from their review of studies of ZSM-5 synthesis in the fluoride medium, Singh & Dut-ta [16] conclude that crystallization is proportional to F- concentration up to a maximum concentration, beyond which it declines.
Silicalite-1 controls were also prepared according to both gel formulas in the absence of a metal salt.
The method of preparation was as follows: NaBr, TPABr and NaF were dissolved in deionised water. The metal salt was added to colloidal silica and briefly stirred by hand to homogenise/dissolve. This was then added to the aqueous solution of the other reagents and stirred mechanically for 10 min. Gels were aged for 2 hours at room temperature (20°C). The syntheses were carried out in closed Teflon-lined stainless steel reaction vessels at autogeneous pressure and were not stirred, i.e., they were static. Gels were incubated at 170°C for 15 days. All products were washed with distilled water and dried overnight in air at 40°C.
Sc-Silicalite-1 control samples were prepared via an incipient-wetness doping technique for use in the characterization experiments and as standards for the XRF analysis. These control samples were prepared as follows: Silicalite-1 synthesized using the gel formula with the lower amount of fluoride was calcined for 5 hours in air at 500°C to remove the template. 1 g of the calcined zeolite was taken and 1 ml of scandium nitrate solution was added in 10 |l drops using a Finn pipette. The concentration of the scandium nitrate solution was such as to create samples with Si/Sc ratios of 200, 33 and 20 and the respective molarities of the solutions used were, to 1 SF, 0.08 M, 0.5 M and 0.9 M. After doping, the samples were dried overnight at 110°C. The samples were then calcined in air for 5 hours at 500°C. A portion of the sample with a Si/Sc ratio of 33 was left as-prepared and was not calcined. Scandia (Sc2O3) was also used as a control sample.
The following reagent sources were used in the study: Colloidal silica (SiO2) (Ludox TMA, Aldrich, 34 wt%), sodium fluoride (NaF) (Fluka Chemika, 99%), ammo-
nium fluoride (NH4F) (Aldrich, 98+%), sodium bromide (NaBr) (BDH, 99%), aluminium hydroxide (Al(OH)3) (Reheis F-2000, Dried Gel S.P.), scandium nitrate (Sc(NO3)3) (Aldrich, 99.9%), scandium oxide (Sc2O3) (Aldrich, 99%), tetrapropyl ammonium bromide (TPABr) (Lancaster 98+%), and deionised water (H2O).
Products were characterized using various instrumental analytical techniques. Analyses were performed on as-synthesized, template-containing samples (except for the calcined Sc-Silicalite-1 and scandia control samples).
No fluoride analysis was carried out.
The pH of the starting gel (after ageing) and the pH of the post-synthesis reaction vessel contents was taken using a Hannah Checker glass pH probe (accuracy 20°C = ±0.2 pH).
Powder X-ray diffraction was carried out using a Philips diffractometer; model PW 1710 generator equipped with a PW 1050/70 vertical goniometer. A copper target provided the Ka radiation source. The 20 scanning rate was 1.2° 20 min-1 (0.02° 20 s-1) and scan range was 3° to 50° 20. The relative crystallinities of the samples were calculated in relation to that of a sample which had been assigned 100% crystallinity. Sili-calite-1 synthesized using a gel of the lower fluoride content was chosen. Crystallinity was based on the mean of the peak counts for the same five largest peaks across each zeolite pattern. Unit cell volume calculations and peak indexing was carried using iterative refinement calculations with Phillips software PW1865 with the refinement error parameters set to 0.05. Samples with crystallinity <10% were not indexed.
The product Si/metal ratios were determined using X-ray fluorescence. Samples were analysed on a SPEC-TRO XEPOS instrument that had a palladium X-ray target. As there was no internal standard containing scandium, the calcined Sc-Silicalite-1 controls of known Si/Sc ratio were analysed. A calibration curve was generated using the intensity of the scandium Kax, which has a wavelength of 0.303 nm [12].
Thermogravimetric analysis was conducted using a Mettler TG50 thermobalance and a Mettler TA3000 processor. 10-15 mg of sample was used under flowing air with a flow rate of 20 ml/min. The temperature range was from ambient temperature (approx 25°C) to 900°C at a heating rate of 10°C/min.
The percentage product yield was calculated using the following formula:
%Yield = 100* ((Mass of product - total mass loss from TGA)/(Mass of SiO2 + Metal Oxide (equivalent, i.e., Al2O3 or Sc2O3) used in the gel).
45Sc MAS NMR was also carried out using a direct-polarisation experiment. Samples were run on a Varian Unity Inova instrument (with a XH frequency of 300 MHz and a magnetic field of 7.1 T). The 45Sc frequency used was
Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.