Zeolite Synthesis - American Chemical Society

---

Chapter 42

Computational Studies of Zeolite Framework Stability

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

R. A. van Santen, G. Ooms, C. J. J. den Ouden, B. W. van Beest, and M. F. M. Post Koninklijke/Shell-Laboratorium, Shell Research Β. V., Badhuisweg 3, 1031 CM Amsterdam, Netherlands

For the purpose of determining the relative stabilities of topologically different aluminum-free tetrahedral networks, Hartree-Fock-level ab-initio calculations have been done of the relative stability of three-, four-, five- and six-unit SiO(OH rings. Very small differences per Τ unit are found for the four-, five­ -and six-rings; however, the energy per Τ unit is un­ favourable for the three-ring. Rigid ion lattice mini­ mization calculations have been performed on Al-free as well as high-Al-content zeolite systems. The results will be discussed for ZSM-5, mordenite and faujasite structures. Very small energy differences, of the order of ~ 20 kJ/mol, are again found for the Al-free net­ works. Open structures have less favourable energy than dense structures due to decreased Madelung energy. Large changes in relative energy are found upon variation of the Al/Si ratio. Medium- and small-pore zeolites are much more sensitive to an increase in aluminum content than the wide-pore material. This should be ascribed to stacking of the cations in the channels of the zeolite. The implications of these results for zeolite synthesis are discussed. 2

The t h e o r e t i c a l work to be discussed here has been i n i t i a t e d with the aim of providing support to fundamental studies of z e o l i t e syn­ thesis. Central to our discussion i s the question whether a synthe­ s i s approach for new z e o l i t e s can be developed on the basis of guidelines generated by computational design. For such an approach to be v i a b l e , c e r t a i n fundamental physicochemical knowledge about z e o l i t e s and t h e i r behaviour i s e s s e n t i a l . In p a r t i c u l a r , one needs to know: - what z e o l i t e structures are possible; - what laws govern their synthesis. Many leads on new z e o l i t e structures are available i n the open l i t e r a t u r e (1). We s h a l l therefore concentrate on the second of the 0097-6156/89/039&-0617$06.00/0 o 1989 American Chemical Society In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

618

ZEOLITE SYNTHESIS

two areas mentioned above, concerning the z e o l i t e formation mecha­ nism. We are interested i n the role of organic ions and organic mole­ cules i n z e o l i t e synthesis. The use of organic molecules i n z e o l i t e synthesis mixtures i n addition to inorganic bases i s known to r e s u l t i n the medium-pore-size low-aluminum-content z e o l i t e s , such as ZSM-5, that revolutionized several processes of importance to the o i l - r e f i n i n g and petrochemical industry. Since, during synthesis, the organic molecules and ions become incorporated into the z e o l i t e micropores i n quantities much larger than required to neutralize the l a t t i c e framework, Flanigan (2) pro­ posed that they act as a template around which the z e o l i t e precursor molecules are formed. Barrer (3) proposed that adsorption of organic molecules during synthesis s t a b i l i z e s the z e o l i t e l a t t i c e . This second idea provided the s t a r t i n g point to our t h e o r e t i c a l work. I t can be understood on the basis of the following considerations. The major c h a r a c t e r i s t i c of a z e o l i t i c material that d i s t i n ­ guishes i t from a n o n - z e o l i t i c one i s i t s microporous structure, due to the presence of interconnected channels. This implies that, whereas the i n t e r f a c i a l energy i s a n e g l i g i b l e quantity for large c r y s t a l l i t e s of non-porous materials, this i s no longer more the case f o r a microporous system. Consider, f o r instance, the interaction of the alumina-free s i l i c a framework of de-aluminated faujasite with water i n the l i q u i d phase. Per mol SiC^, 1.3 mol can be occluded. The heat of evapo­ r a t i o n of water i s 36 kJ/mol and the heat of adsorption -30 kJ/mol, so per mol S1O2 there i s an energy cost of 20 kJ. I t i s of interest to compare this figure with the heat of immersion of aluminum-rich f a u j a s i t e . For a compound with composi­ t i o n NagQAlgQSi^^2°384 calculates -122.5 kJ/mol (4), which i s mainly hydration energy of intra-channel sodium ions. So i n the c r y s t a l l i z a t i o n of microporous systems there i s a ba­ lance between the energy cost of micropore generation and the energy gain because of i n t e r f a c i a l s t a b i l i z a t i o n by occluded molecules. C l e a r l y there i s no gain for porous aluminum-free systems i n water, so they w i l l not be formed unless a t r i c k i s used. These notions can be given a more exact expression by the use of the rule of Gibbs: o n e

d

M i

- R Τ / a

. I Σ

±

Θ .ά1η(ρ ) ±

±

]

(1)

In t h i s expression i s the i n t e r f a c i a l chemical p o t e n t i a l , a the average surface area, the f r a c t i o n of the surface covered by the adsorbed molecule i , and p^ the corresponding p a r t i a l pressure. Ex­ pression (1) can be p a r t i a l l y integrated, assuming Langmuir adsorp­ t i o n for the adsorbed molecules, to give: μ

±

- RT / f . [ l n d - E ^ i ) ]

(2)

In expression (2), i s the f r a c t i o n of c a v i t i e s present i n the structure per mol S i 0 . For instance, i n s i l i c a l i t e f equals 24. Of course, c r y s t a l l i z a t i o n of a microporous system completes with the formation of a dense system unless: 2

< „dense _

zeo.lat.

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

( 3 )

42. VANSANTENETAL.

Zeolite Framework Stability

619

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

In order to quantify this r e l a t i o n one needs information on the l a t t i c e energies of z e o l i t e s as a function of l a t t i c e topology and composition. In the next sections calculations of the l a t t i c e energies of z e o l i t e s v i a several d i f f e r e n t computational approaches w i l l be r e ­ ported. Each approach i s approximate; each has i t s own advantage(s). To begin with, we s h a l l discuss computations of the l a t t i c e energies of aluminum-free z e o l i t e s . We s h a l l then go on to discuss the l a t t i c e energies of aluminum-containing z e o l i t e s . We s h a l l con­ clude t h i s paper with a discussion of the relevance of the results obtained to z e o l i t e synthesis. The S t a b i l i t y of Z e o l i t i c S i l i c a Lattices Methods. I n i t i a l calculations were done using the semi-empirical Extended Huckel method (5). The purpose of these quantum chemical calculations was to explore the main electronic c h a r a c t e r i s t i c s of chemical bonding i n s i l i c a tetrahedral networks. The calculations showed that bonding can be considered covalent and can be considered to be due to l o c a l i z e d SiOSi units. Very small differences i n bond strength between d i f f e r e n t s i l i ­ ca polymorphs were found. Since the Extended Huckel Method i s too approximate to calculate r e l i a b l y the small diffferences i n energy between low-density material, containing micropores, and high-densi­ ty material without micropores, work was i n i t i a t e d to study the same problem but now with two rigorous techniques that are currently con­ sidered to be state of the a r t . Hartree-Fock-level a b - i n i t i o calculations can provide r e l i a b l e p o t e n t i a l energy diagrams for small clusters. Such calculations can be applied to the z e o l i t e l a t t i c e i f the clusters are chosen care­ f u l l y and use i s made of the property that bonding i s highly l o c a l i ­ zed i n these materials. Calculations were done using the GAMESS abi n i t i o package. Since t h i s approach does not account for long-range e l e c t r o ­ s t a t i c p o t e n t i a l s present i n the extended material, the second approach chosen was the r i g i d - i o n l a t t i c e energy minimization tech­ nique, widely used i n s o l i d - s t a t e chemistry. This method i s based on the use of e l e c t r o s t a t i c potentials, as well as Born repulsion and bond-bending potentials parametrized such that computed atom-atom distances and angles and other material properties, such as, f o r i n ­ stance, e l a s t i c constants, are well reproduced for related mate­ r i a l s . In our case, parameters were chosen to f i t α-quartz. Results of Hartree-Fock Calculations. Using a ST03G basis set, c a l ­ culations were performed on the open dimer (Figure 1) and 3-, 4-, 5and 6-rings of C symmetry (Figure 2), i n which η stands for the number of (Si0)(0H>2 units, the dangling bonds being saturated with Η atoms. Figure 3 shows the equi-energy contours of the dimer as a function of bridging SiO length and SiOSi angle. The r e s u l t f u l l y concurs with results from similar work obtained by Gibbs et a l . ( 6 ) . I t can be seen that there i s a steep increase i n energy i f the SiO distance s t a r t s to d i f f e r from i t s equilibrium value, but that the energy changes involved i n angle variations are very small. The SiO bond strength i s of the order of 450 kJ, whereas the change i n energy with angle v a r i a t i o n i s only of the order of 20 kJ. The n v

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

620

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

ZEOLITE SYNTHESIS

Figure 2. Ring structures; C

n v

symmetry assumed (n-3, 4, 5, 6).

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

42.

VANSANTENETAL.

Zeolite Framework Stability

621

Figure 3. Total energy as a function of Si-0 bond length and Si-O-Si bond angle. Contour lines i n units of kJ/mol.

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

622

ZEOLITE SYNTHESIS

l a t t e r measurement agrees well with the very low rocking and t o r s i o ­ nal frequencies around the SiOSi bond measured by infrared experi­ ments on z e o l i t e s (7). This implies that l a t t i c e deformation can take place at l i t t l e energy cost as long as no SiO bonds are broken. Tables I and II summarize the results found f o r the s i l i c a t e r i n g c l u s t e r s . Table I,A presents the computed t o t a l energies per Si0(0H)2 u n i t . Bond lengths and bond angles i n the rings were o p t i ­ mized within the constraint of C symmetry. As can be seen i n Table I,A the energy difference between the 3-ring and the other rings i s considerable. However, the energy differences per unit Si0(0H)2 of the 4-, 5- and 6-rings are computed to be within 1 kJ. Computed SiO distances and SiOSi angles are presented i n Table I I . The reason for the i n s t a b i l i t y of the 3-ring can be r e a d i l y seen. Within the tetrahedra the OSiO angles d i f f e r from the optimum tetrahedral angles; i n addition, the SiO distances appear enlarged. A c l e a r trend i n SiO distance as well as SiOSi angle i s observed as a function of r i n g s i z e : the SiO distance shortens, whereas the SiOSi angle increases. The r e l a t i o n between SiO distance and SiOSi angle i s f u l l y consistent with the p o t e n t i a l energy diagram presen­ ted i n Figure 1 and i s due to changes i n h y b r i d i z a t i o n of the e l e c ­ trons on the 0 atom (5). I t i s of interest to compare the extent to which geometry de­ pends on r i n g size with the average bond lengths and angles observed i n alumina-free z e o l i t e s with d i f f e r e n t numbers of rings. This com­ parison i s given i n Table I I I . A trend i n angle and distance v a r i a ­ t i o n with increased amount of 5-rings compared to 4-rings s i m i l a r to that observed i n the calculations can be distinguished. The d i s t a n ­ ces compare well; the average angle measured for mordenite and ZSM-5 appears somewhat large, though. This may be due to the d i f f i c u l t y i n determining oxygen positions accurately.

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

n v

Table I. Ring Structures A. Total Energies η (Ring Size)

e/n (a.u.)

3 4 5 6

-508. .41206 -508. .41547 -508, .41578 -508, .41580

B. Energy Differences i n kJ/mol η 3 4 5 6

3 0..000 -8..944 -9..765 -9,.806

4

0,.000 -0,.820 -0,.861

5

0..000 -0..041

6

0.000

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

42. VANSANTENETAL.

Zeolite Framework Stability

623

Table I I . Optimized Geometries for Ring Structures Distances i n Angstrom, angles i n degrees 3 Si-O Si-Si

1..662 2,.942

Si-O-Si O-Si-O Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

4

130, .1 105, .7

5

6 1,.592 3,.067

1..595 3..063

1..602 3..042

147, .5 108, .4

143, .5 108, .3

148, .8 108 .1

Table I I I . Comparison of Average Distances (Angstrom) and Angles (Degrees) Calculated for a Few Zeolites with Different Ratios of η-Rings (nR) 4R(%)

5R(%)

6R(%)

70 5 5

0 95 85

30 0 10

Faujasite(26) Mordenite(22) ZSM-5(20)

Distance

Angle

1.61 1.58 1.59

142 157 156

The results of these calculations are consistent with our ear­ l i e r conclusions. As long as no 3-rings are present i n the aluminafree material, differences i n covalent energy are very small. The a b - i n i t i o calculations indicate that these differences do not exceed 1 kJ/mol. R i g i d Ion L a t t i c e Energy Minimization Calculations. Table IV pre­ sents l a t t i c e energies derived from f u l l y converged r i g i d ion l a t ­ t i c e energy minimization calculations. Again one notices the small difference i n energy between dense and open structures. The depen­ dence on density (8) i s presented i n Figure 4. The lower energy of the more open structures relates to the decrease i n Madelung energy. However, as Figure 2 shows, l o c a l topo­ l o g i c a l effects also play a role. See, for instance, the difference i n energy calculated between z e o l i t e A and f a u j a s i t e . Table IV. Calculated Rigid Ion L a t t i c e Energies of Aluminum-Free Zeolite L a t t i c e s

Zeolite

Lattice Energy, kJ/mol

Faujasite Mordenite Zeolite A Silicalite Sodalite α-Quartz

-11914 -11931 -11931 -11945 -11949 -11959

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

ZEOLITE SYNTHESIS

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

624

Discussion of Alumina-Free-Lattice Calculations. Two main conclu­ sions emerge from the results presented so f a r . F i r s t l y , the energy differences between tetrahedral networks with d i f f e r e n t r i n g systems are very small, except when the networks contain 3-rings not found i n nature. This i s a very s i g n i f i c a n t conclusion because i t i s widely be­ l i e v e d (9) that, i n order to synthesize systems with 5-rings, such as ZSM-5, the only requirement i s to synthesize systems low i n a l u ­ mina. Both our quantum-chemical and our e l e c t r o s t a t i c l a t t i c e calcu­ l a t i o n s contradict t h i s theory. The calculations show, for example, that sodalite, which contains only 4- and 6-rings and no alumina, i s more stable than ZSM-5, i n which 5-rings predominate. This agrees with recent experimental work r e l a t i n g to the synthesis of highs i l i c a sodalite (10). Secondly, the decrease i n l a t t i c e energy with increasing micro­ porous volume reaches a maximum at 45 kJ/mol, the difference between α-quartz and f a u j a s i t e , f a u j a s i t e being the z e o l i t e with the largest micropore volume. Expressions 2 and 3 show that, i n order to overcome this energy difference, the micropore c a v i t i e s should be largely f i l l e d with ad­ sorbed molecules. As mentioned e a r l i e r , i n cases where low-aluminacontent materials have been d i r e c t l y synthesized, high values f o r are invariably found. This confirms Barrer's postulation. The domi­ nant i n t e r a c t i o n that governs narrow- and medium-pore z e o l i t e syn­ thesis i s the strong i n t e r a c t i o n of the occluded organic molecule with the micropore wall. Of interest with respect to this hypothesis i s the s i g n i f i c a n t difference i n heat of p a r a f f i n adsorption between the medium-pore z e o l i t e s i l i c a l i t e and large-pore, de-aluminated f a u j a s i t e . The heat of p a r a f f i n adsorption i s much smaller i n the case of the de­ aluminated f a u j a s i t e , which has so f a r had to be prepared by an i n d i r e c t route, than for s i l i c a l i t e , which can be synthesized d i r e c t i n the presence of an organic molecule. The difference, which i n ­ creases with chain length, i s of the order of 5 kJ/mol per CH2 unit, and may be ascribed to the optimum f i t of hydrocarbon and channel i n the case of the medium-pore z e o l i t e (11, 12). ^ C s o l i d - s t a t e NMR studies on occluded organic ions by our­ selves (13) and others (14) have provided further evidence f o r the strong interactions between occluded organic molecule and micropore wall mentioned above. We have also reported l a t t i c e s t a b i l i z a t i o n of s i l i c a l i t e (ZSM-5) by occlusion of tetrapropylammonium ions (15.) . I t i s therefore very u n l i k e l y that a synthesis procedure can be defined for the preparation of highly s i l i c e o u s forms of large-pore (12-ring or larger) z e o l i t e s such as faujasite d i r e c t . V a r i a t i o n of Zeolite L a t t i c e

Composition

The changes i n heat of formation as a function of alumina content were studied v i a two d i f f e r e n t approaches. One approach (4) i s to calculate, for a c e r t a i n z e o l i t e struc­ ture, the Madelung and p o l a r i z a t i o n energies for fixed l a t t i c e p o s i ­ tions. The heat of formation due to ionic bonding i s calculated both for the z e o l i t i c aluminosilicate with varying amount of aluminum and

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

42. VAN SANTEN ET AL.

Zeolite Framework Stability

625

sodium ions and for the z e o l i t i c s i l i c a , with the same framework distances as i n the aluminum-containing material. The difference between the two heats of formation i s assumed to represent the c o n t r i bution of the t o t a l heat of formation due to ionic bonding stemming from the presence of aluminum ions and cations i n a p a r t i c u l a r z e o l i t e structure. In these calculations averaged charges on the intra-tetrahedral l a t t i c e cation positions were used. The difference between the two heats of formation due to ionic bonding i s added to the heat of f o r mation due to covalent bonding r e s u l t i n g from the simple Extended Huckel Method for z e o l i t i c s i l i c a s i n order to arrive at the t o t a l heat of formation of the z e o l i t e structure as a function of the amount of aluminum. For the calculations we used the method and computer program described by Piken and Van Gool (16,17). Adsorption of water into the pores of a z e o l i t e structure may contribute considerably to the t o t a l heat of formation because of hydration of the cations present i n the pores. Using empirical data on hydration with water we have estimated the magnitude of t h i s e f f e c t (4). The second approach involves the use of the r i g i d ion l a t t i c e energy minimization method discussed e a r l i e r . This technique proved e s p e c i a l l y useful for determining the s t a b i l i t y of structures f o l lowing c a l c u l a t i o n of their energy by the method discussed i n the previous paragraph. The charges of aluminum and s i l i c o n were again averaged. Results of Madelung and Potential Energy Calculations. Figure 5 gives the heat of formation due to ionic bonding for faujasite, mordenite and ZSM-5 i n the presence and i n the absence of A l and Na ions. L a t t i c e s with the same l a t t i c e constant are compared. SiO and AlO distances used are values extrapolated from low- and high-Alcontent z e o l i t e crystallographic data. The results f o r the t o t a l heat of formation computed i n the way explained e a r l i e r are given i n Figure 6. A decrease i n the heat of formation with increasing aluminum content i s observed. In the f i n a l step of the c a l c u l a t i o n the e f f e c t of hydration of the cations was taken into account. The r e s u l t s are presented i n Figure 7. Changing the aluminum content has a s i g n i f i c a n t e f f e c t on the r e l a t i v e s t a b i l i t y of z e o l i t e structures with very d i f f e r e n t topology. While the heat of formation of the wide-pore z e o l i t e i s a f f e c ted very l i t t l e , the heats of formation of the medium-pore z e o l i t e s f a l l significantly. +

Results of L a t t i c e Energy Minimization Calculations. Relative l a t t i c e energies of faujasite, mordenite, s i l i c a l i t e and sodalite were compared. For the framework and cation positions of faujasite and sodalite the same data were used as before, from Hseu (18) and Olson (19), and Baerlocher (20) and Chao (21), respectively. For mordenite and sodalite the data of Meier (22) and Mortier (23.) (on mordenite) and Baerlocher and Meier (24) (on sodalite) were used. The s t a r t i n g unit c e l l s f o r faujasite and mordenite have the chemical composition N a [ A l S i ^ g _ 0 g g ] . For sodaite and ZSM-5 we used Si^2°24 * x[ ^x ^96-x°192l * ^ f k i cation positions were allowed to relax under constant pressure. Parameter values used and d e t a i l s of the c a l c u l a t i o n can be found i n (25). x

a n c

iia

A

s

x

x

T

e

r a m e w o r

a n <

In Zeolite Synthesis; Occelli, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

ZEOLITE SYNTHESIS

LATTICE ENERGY, kJ/moL χ 10 -11 8 8 r

-11.91 -

Downloaded by UNIV OF MISSOURI COLUMBIA on April 16, 2013 | http://pubs.acs.org Publication Date: July 31, 1989 | doi: 10.1021/bk-1989-0398.ch042

O FAUJASITE

-11.94 -

Ο MORDENITE

Ο LINDE TYPE A

^FERRIERITE ZSM-50 Ο SODALITE

QUARTZ Ο

-11.97 -

JL 15.20

-12 12

1840

^.60

24.80

J 28

DENSITY, Τ SITES PER A χ 1C? Figure 4. Lattice energy as a function of density.

AL / S i RATIO 0

Ο. 2

0.1 —Τ­

0.3

0.4

0.5

ι

1

1

-2500 ZSM-5

Δ -3500

MORDENITE