On the dynamics of organic-zeolite interactions: Tetramethylammonium in sodalite

On the dynamics of organic-zeolite interactions: Tetramethylammonium in sodalite

On the dynamics of organic-zeolite interactions: Tetramethylammonium sodalite C.J.J. den Ouden, K.P. Datema, Koninkli’kelShell-Laboratorium, The Ne...

831KB Sizes 4 Downloads 192 Views

On the dynamics of organic-zeolite interactions: Tetramethylammonium sodalite C.J.J. den Ouden,

K.P.

Datema,

Koninkli’kelShell-Laboratorium, The Net Izerlands

F. Visser,

Amsterdam

M. Mackay,

in

and M.F.M.

(Shell Research B.V.),

Post

Amsterdam,

The dynamics of tetramethylammonium (TMA) occluded in a @-cage of zeolite sodalite has been investigated in a comparative molecular dynamics (MD) and n.m.r. (13C, *H) study. Concerning deuterium n.m.r., this is the first time that measurements on deuterated organics that are occluded during zeolite synthesis have been performed. The main objective was to understand the dynamics of organic-zeolite interactions on a molecular level, which may enhance to the insight into the role of the organic during zeolite synthesis. MD calculations as well as solid-state deuterium and crosspolarization magic angle spinning n.m.r. results indicate that TMA occluded in the sodalite P-cage has a considerable degree of rotational freedom. MD revealed this rotational freedom to be caused mainly by the dynamics of the zeolite lattice; the distance between diametrically opposed oxygen atoms in the P-cage was found to fluctuate by as much as 0.4 A. Keywords: ‘H n.m.r.; organic material

13C n.m.r.;

molecular

dynamics;

zeolite-organic

INTRODUCTION Various organic molecules and/or (organic) cations are generally used in the synthesis of different zeolites. The role of the organics has often been, and still is, a subject of much debate. Four main roles of the organics can be distinguished from several studies. Very often only the templating or structuredirecting role as proposed by Flanigen’ back in 1973 is stressed. A clear example in this context is the role of the tetrapropylammonium (TPA) cation in ZSM-5 synthesis. It is thought that TPA stabilizes certain aluminosilicate species that replicate themselves interactions. There is, through structure-specific however, no one-to-one correlation between the molecular geometry of the organic and the macroscopic zeolite structure ultimately obtained. This indicates that the ternplating role is certainly not the only one. Lok’ reviewed the role of the organic, and he assumed that the presence of the organic has a large influence on the (reaction) gel chemistry (pH, composition, solubility, and presence of various precursor species, etc.). It was indicated that a structureAddress reprint requests to Dr. den Ouden at the Koninklijkel Shell-Laboratorium, Amsterdam, Shell Research B.V. Badhuisweg 3, 1031 CM Amsterdam, The Netherlands. Received 4 June 1990; accepted 2 July 1990 @ 1991

418

Butterworth-Heinemann

ZEOLITES,

1991, Vol 7 7, June

interactions;

zeolite

synthesis;

dynamics

of occluded

directing role is operative if the right gel chemistry is present. Recently, zeolite crystallization mechanisms have been proposed in which the crystalline phase evolved from specific double t/-ring aluminosilicate species in solution.” It was suggested that the presence of such species is relatecl to the organic usecl. Finally, the stabilizing role of the organic has been that the highlighted by Barrel-. -’ It was proposed organics act like pore fillers and, hence; prevent water to interact (in the case of high-silica zeolites) with the hydrophobic zeolitic inner surface. Obviously, the void-filling role is strongly related to the ternplating role, as has also been shown previously.’ Although the influence of organics on the formation of various zeolites has been extensively stucliecl, only little attention has been paid to the dynamic features of organic-zeolite interactions. Information that can be deduced from such stuclies might involve insight into the time scale on which organic-zeolite interactions occur and the influence of both organic dynamics and zeolite lattice clynamics. In this paper, we report the results of a comparative investigation of the clynamics of the TMA cation occluded in p-cages of zeolite sodalite. TMA is generally consiclered as the conventional template fol sodalite synthesis (see, e.g., Ref. 6). Although nothing can be deduced from this stucly concerning the gel chemistry

cluring

synthesis,

observations

concerning

Organic-zeolite

the molecular fit (ternplating) and pore-filling ability of TMA in sodalite can Ix made. The two cornplementary methods for studying the (dynamic) interactions of organic material with zeolites are molecular cl namics (MD) ad solid-state n.m.r. .Y [deuterium (-H) n.m.r. and cross-pol;‘rizatic,n magicangle-spinning (CP-MAS) n.m.r.]. MD techniques applied in zeolite adsorption studies have gained much interest over the last few years. However, Ijecause . of the limited computei power and the upcoming cle\~elopnient of MD methodologies in this field of research, simulations are normally devoted to rather small systenls.‘.x To reduce the complexity and required computer time, intramoleculai-- adsorbate dynamics and zeolite lattice dynamics are usually neglected. The typical time scale of MD calculations is of the order of’ 100 ps. To 0111 knowledge, MD calculations have ne\‘er been applied before to investigate the templating effect of organic molecules used in zeolite synthesis; neither have MD studies been reported in which the zeolite lattice dynamics were taken into account. Deuterium (“H) 11.1n.r. can be conveniently used to determine the mobility of “solidlike” material on the microsecond time scale.!‘~“’ As shown before, very detailed information on molecular mobility in polymers can be obtained via this technique.’ ‘.I Recently, the technique was also introduced in the field of zeolite research to study the rotational mobility of However, as far as we know, adsorbed species. “-“’ the technique has not yet been used to monitor the dynamics of organic-zeolite interactions in situ, i.e., after the use of a deuterated organic in a zeolite synthesis. CP-MAS ‘“C n.m.r. can also be used since it independently yields additional information about molecular mobilities on the millisecond time scale.“‘.” This technique has been used by Boxhoorn et al.‘” to study the structure and position of TPA in zeolite ZSM-5. Concerning TMA trapped in various zeolite pores and cages, Hayashi et al.‘s.“” have performed very interesting ‘sC n.m.r. studies. They were able to relate the observed chemical shift to the diameter of the zeolite pore in which the TMA species are occluded. Furthermore, using various linethey investigated the dynabroadening techniques, mics of TMA trapped in various zeolite pores and cages. For TMA in the sodalite P-cage, they concluded that TMA rotates freely on the millisecond time scale. As already mentioned, thii paper is devoted to the special case of TMA occluded in the sodalite P-cages. For the 2H n.m.r. experiments, sodalite has been prepared with deuterated TMA and in situ measurements have been carried out. Furthermore, 13C n.m.r. methods have also been applied to study the mobility of TMA in sodalite p-cages. With the MD simulations, we will specifically focus on the interpretation of the n.m.r. results and give a description of the molecular fit of TMA in sodalite p-cages. For this molecular description, it was necessary to include zeolite lattice dynamics.

EXPERIMENTAL

interactions:

AND

C.J.J.

den

Ouden

et al.

COMPUTATIONAL

Synthesis Deuterated TMABr was prepared by reaction of deuterated methyl bromide (ex Merck) with ammonia gas. Methyl bromide was dissolved in a large excess of ethanol (9.6 mmol CD:sBr/ml C,H,OH). The solution was placed in a flask that was equipped with a reflux condenser and kept cool using an ice bath. The solution was stirred while ammonia gas was bubbled through the solution for 8 h. The resulting deuterated TMABr [N(CD:s) ,Br] precipitated from the solution as a white suspension (yield 13%). TMABr was filtered off, thoroughly washed, and ionexchanged with a slurry of Ag.,O in water to produce an 18% deuterated TMAOH solution in water. The was checked with highpurity of the product resolution ‘Y n.m.r. spectroscopy (data not shown). Sodalite was prepared using deuterated TMAOH from a synthesis gel having molar composition SiO~/Al.,O:~/Na~O/TMAOH/HyO = 1 I. l/l/l .25/4.6/ 2 17. This gel was placed in an autoclave and agitated under hydrothermal conditions at 190°C for 63 h. X-ray powder diffraction (XRD) revealed the crystalline product to be sodalite with a trace of zeolite omega. Another sodalite batch for the 13C n.m.r. measurements was prepared following the same recipe as pointed out above, but this time protonated TMAOH (ex Fluka) was used. The crystalline product showed the same XRD characteristics as the sodalite prepared with the cleuterated organic. The maximum an~ount of organic material that can be occluded in the sodalite void space is one TMA per Si5A101L, unit (which is equivalent to one TMA per P-cage), thus 2 1 wt70. Differential thermal analysis (d.t.a.) revealed a weight loss of 18 wt% of organic material, indicating that the samples used for our n.m.r. analyses contained occluded TMA in about 90% of the &cages. For the “H n.m.r. experiments, two n.m.r. tubes, one containing pure deuterated TMABr as a reference and another containing sodalite with deuterated TMA occluded, were filled under vacuum in a Braun MBPOOG-K glove box to ensure that the material was absolutely dry. Before filling the tubes, the samples were dried overnight in the glove box under a dry nitrogen flow. The n.m.r. tubes were subsequently sealed using a two-component epoxide resin to ensure that the cap of the tube was completely air-tight. Prior to being placed in the glove box, the samples were dried in an oven at 393 K overnight.

N.m.r.

spectroscopy

*H n.m.r. spectra were recorded at 46.06 MHz with a Bruker MSLSOO Fourier transform spectrometer. The quadrupolar pulse sequence”.‘” with 5 ps, 90” pulses and pulse intervals of 30 ps was used. In all cases, the relaxation delay was 1 s, which was sufficient to eliminate saturation effects. The temperature of the sample was controlled by a Bruker variable temperature unit B-VT 1000. For tempera-

ZEOLITES,

1991, Vol 11, June

419

Organic-zeolite

interactions:

C. J. J. den

Ouden

et al.

tures below room. temperature, nitrogen gas, evaporated from a. Dewar containing liquid nitrogen, was used. Measurements were started at high temperature, and after each temperature decrease, the samples were allowed to equilibrate for at least 20 min. 13C n.m.r. spectra were recorded on a Brukel MSL300 instrument (carbon frequency 75.46 MHz). The spectra were obtained in a double-bearing CP-MAS probe operating with 4 ~.Is, 90” pulses. Other experimental parameters are specified in the figure captions. The MAS experiments were carried out with Spinning rates in the range 2-5 kHz.

Molecular

dynamics

All MD simulations were performed in the NVT ensemble (constant number of particles, constant volume, constant temperature)‘.’ using the Sybyl modeling software2” implemented on a Silicon Graphics 4Dl20 computer. The model consists of a single all-silica sodalite b-cage with one TMA occluded. The presence of al~~minum atoms in the zeolite lattice is neglected. The silicon atoms of the p-cage are attached to terminal dummy atoms that are fixed in space to mimic the effect of the surrounding sodalite lattice, which is not considerecl in the simulations. The Si-dummy bonds are regarded as Si-0 bonds; this is the only way the dummy atoms enter the calculation. Total simulation times varied from 75 to 100 ps with a constant time step length of 1 fs (= 10-l:’ s). The TMA-sodalite interaction parameters were obtained following Kiselev et al.-’ This class of parameters had proved to yield satisfactory results in earlier MD studies on diffusion and adsorption of hydrocarbons in zeolites.“” Because we are considering only a small subsystem out of an infinite lattice, electrostatics cannot be taken properly into account. Thus, neither the positive charge on the TMA cation nor the negative charge on the zeolite lattice due to the presence of the aluminum is considered. Therefore, only van der Waals atom-atom interactions between the TMA atoms and the sodalite oxygen atoms are considered (the influence of the small silicon atoms in the zeolite lattice is assumed to be negligible and has been omitted). This approximation in the TMA-sodalite interaction is expected to be reasonable, because it can be envisaged that the molecular mobility of TMA is mainly dominated by steric effects. Zeolite lattice potential parameters are deduced from ab initio calculations on silica clusters,“’ and the standard Tripos force fields” was used for the intramolecular degrees of freedom in TMA. A series of MD simulations was performed with an increasing level of complexity. First, the sodalite cage was kept rigid, whereas TMA was allowed to move freely within the cage. At the next level of complexity, the atoms of the sodalite cage were allowed to move freely as well. Thus, in the latter case, both intramolecular TMA and zeolite lattice deformations were taken into account.

420

ZEOLITES,

7997, Vol II, June

RESULTS

AND

DISCUSSION

In this section we will cliscuss the results as obtained by MD and n.m.r. We start off with the MD results because these stuclies represent the smallest time scale (100 ps) in which the mobility of TMA in the sodalite p-cages is monitored. Subsequently, the results for ‘H (microsecond time scale) ancl ‘:‘C n.m.r. (millisecond time scale) will be cliscussed. Initially, the approaches will be cliscussed separately, but at the end of this section, a comparison will be made.

Molecular

dynamics

(picosecond

time

scale)

In the first calculations, the sodalite p-cage is assumed rigicl while TMA is flexible. In this case, no rotation of the organic is observed al room temperature. Only at ele\fatetl temperatures up to 973 K, some rotational mobility of TMA can he olxerved. If the sodalite lattice is linconstrainecl, however, this situation changes considerably. Color. Pltr~cv Iu, b, and c display the rotation of TMA occlucled in sodalite at T = 273, 153, ancl 123 K, respectively. The dots in these color plates represent subsequent positions of one hydrogen atom of TMA at 50 fs time intervals. Thus, the trajectories formed by the clots give direct information on the rotational movement of TMA occluded in the sodalite p-cage. From the color plates it can be deduced that the rotational movement of TMA at T = 273 K is not severely hindered given the “cloucl” of positions over the simulation time. However, at temperatures towel than or equal to 123 K, the rotational movement becomes hindered. For these low temperatures, two distinct orientations (Color flak Ic) of TMA in the sodalite p-cage are observed. It should be noted, however, that the duration of the MD simulation is only of the order of 100 ps. Thus, for longer times, say of the order of microseconds or even milliseconds, the molecular rotation may appear to be very fast. Close examination of the computational data reveals that intramolecular rotations of the CHy groups about the N-C bond in TMA do not occur. The rotation of TMA in the P-cage seems to be stimulated by the lattice deformations. At T =273 K, a spread in the average distance of two diametrically opposed oxygen atoms of 0.4 A was observed. Apparently, this fluctuation is large enough to cause a rapid rotation of TMA occluded in sodalite. The lattice deformations are localized effects because the volume of the P-cage was found to be nearly constant during the simulation; variations in the cage volume were only of the or’der of 0.5% of the total volume. Although TMA is observed to rotate very rapidly, this does not mean that the organic can have any random orientation in the sodalite P-cage. Analysis of the data reveals that the TMA species “hop” from one preferred orientation to another in the cage. The jumping frequency is directly related to the frequency at which the deformations of the sodalite cage occur. Color Plate 2 clearly visualizes this phenomenon. In this colour plate, the sodalite cage and occluded TMA are displayed in the state they were in

Organic-zeoiite

*H n.m.r. scale)

Figure 1 Deuterium n.m.r. spectra, 46.06 MHz, of deuterated TMA as a solid at 243 K (a), in sodalite at 243 K (b), and at the lowest possible temperature, 153 K (cl. The number of scans was 1354 (a), 1810 lb), and 6026 (c). The relaxation delay was 1 s, and the time between the two 90” pulses of 5 ps of the quadrupolar echo sequence was 30 p

at1= 1,2,3,... 70 ps. The view is perpendicular to one of the 6-ring windows from which the sodalite P-cage is built up. Color Plate 2 allows two main observations: First, it shows that the variance in the cage diameter is caused mainly by vibrations of zeolite oxygen atoms around their average position (Si-0-Si bond-angle vibrations); the positions of the T-sites are observed to be constant mainly in time. Second, it shows that although the TMA species rotate very rapidly the organics occupy only distinct symmetryequivalent positions. This phenomenon can be explained by assuming a perfect molecular fit of TMA in the sodalite p-cage. In this case, TMA can be accommodated in the b-cage only in a limited number of degenerated orientations due to steric hindrance. Because of symmetry, TMA is always in one of these orientations, but a “hopping” between those orientations does occur. The main conclusions to be drawn from this MD study are as follows: The distance between two diametrically opposed oxygen atoms in the p-cage fluctuates by as much as 0.4 A around the timeaveraged value of 6.6 A. As no variation in the cage volume is observed, it can be concluded that this fluctuation is responsible for the rotational freedom of occluded TMA, which is further demonstrated by the total absence of TMA rotations in the rigid cage simulations. Additionally, the TMA-socialite interaction cannot be considerecl as a static “hand-inglove” system while TMA seems to be able to rotate rapidly at T = 273 K. The rotational freedom of TMA o’ccluded in the socialite p-cage is significantly reduced at temperatures of 123 K or lower.

interactions:

spectroscopy

C.J.J.

(microsecond

den Ogden

et al.

time

In the first instance, solid deuterated TMABr is measured as a reference. The reference ‘H n.m.r. spectrum at 243 K consists of three components (Figure la): two axially symmetric powder patterns with a zero-asymmetry parameter and a quadrupolar splitting (wJ23~) of 78 and 24 kHz, respectively, and an isotropic peak, which has less than 30% of the total intensity. The fact that some of the methyl groups of TMABr are mobile may be due to the presence of residual hydration water in spite of the fact that all samples had been carefully dried beforehand. The conclusion that can be dratin from the two powder patterns in Figure la is that two inequivalent, immobile sites exist for the methyl groups in TMABr as a solid, which may be due to the presence of the BI anions or hydration water, and, for comparison with the TMA templates occluded in sodalite, that more than 70% of the TMA species in the solid is immobilized on the microsecond time scale, as would be.expected for a solid. In the sodalite p-cage, the TMA species give rise to predominately one isotropic peak. At 243 K, this component represents more than 95% of the total spectral intensity (Figure Ib). At lower temperatures down to 153 K, the low-temperature limit of our equipment, the isotropic peak still remains the majo, component, but two minor powder patterns occur with wJ2x values of 24 and 78 kHz, having a combined intensity of 5% (F;~w Zc). The spectra unambiguously indicate that 95%’ of the methyl groups of the TMA cations reorient isotropically in the sodalite P-cage. It is pointed out that if TMA species were in the intermediate dynamic range for ‘H n.m.r., thus having short T, relaxation times, they would not contribute to the echo intensity under these conditions. In this case, because of this effect, no quantitative parameters can be obtained.“.‘i’ At present, however, due to the observed isotropic peak, these conditions do not apply. Clearly, the TMA cations occluded in the sodalite b-cage are not immobilized by the surrounding zeolite lattice. This is a remarkable and unexpected effect. If the TMA would have been completely immobilized in sodalite, a wideline spectrum consisting of powder pattern(s) with nonzero asymmetry parameter would have been observed. For such entirely rigid methyl groups, ZOJPX values of 165-175 kHz have been reported... To explain the isotropic peak in the ‘H n.m.r. spectrum, both rapid rotation about the N-C bond of the CHS groups and fast librational motion of the N-C bond itself must be involved. In case of only one of these two motions, still a wideline spectrum would have been observed, albeit that this would have been an axially symmetric powder pattern with a reduced 7~42~~ value. Therefore, the model with two types of rapid motion for the methyl groups is proposed. It corresponds to a virtually isotropic reorientation of the methyl groups of the TMA cation OII the microsecond

time

scale.

ZEOLITES,

1991,

Vol

71, June

421

Organic-zeolite

1

1 90

interactions:

I

1 70

C.J.J.

1

den Ouden

I 50

I

et al.

I 30

I

I 10 pm

Figure 2 Room-temperature 75.48 MHz high-resolution solidstate 13C CP-MAS n.m.r. spectrum of TMA occluded in the sodalite P-cage. The sample spinning speed was 3.04 kHz, and a &-field for proton-decoupling corresponding to a 4 kts 90” pulse was employed. The relaxation delay was 6 s, and the contact time 5 ms. The full width of the TMA methyl group peak at 58.4 ppm at half-height was 32 Hz

At 153 K, a small amount of the TMA has become immobile, yielding a wide line spectrum consisting of two axially symmetric powder patterns similar to that observed for TMA as a solid. From this observation, the offset of at least one of the two methyl group motions (rotation about the N-C bond, librational motion of the N-C bond) is concluded. This may very well be due to a decreased mobility of the surrouncling 0 atoms in the zeolite lattice that are part of the sodalite P-cage (see Molecular dynamics section above).

“C n.m.r. scale)

spectroscopy

(millisecond

time

In addition to ‘H n.m.r., “’ C n.m.r. line-narrowing experiments have been carried out to stucly the TMA (nondeuterated) mobility in socialite. information about the TMA mobility can be obtained by determining which interactions cause the line broadening observed in the “‘C spectrum. Normally, if one wishes to obtain a well-resolvecl ‘“C n.m.r. spectrum of a solidlike sample, application of MAS and proton-decoupling techniques is required.‘“.” In the resulting high-resolution spectrum, the peak positions contain the information of interest. For the present case of TMA in sodalite (F@Lw 2), the methyl TMA peak is positioned at 58.J ppm, corresponding to an average pore size of 7 A for the sodalite P-cage, in . agrement with earlie observations.*” However, from the static “‘C n.m.r. spectrum of TMA in sodalite (Fig?r~ 3~), it is observed that the peak of the TMA methyl groups is already quite narrow without the application of line-narrowing techniques: 1.03 kHz. In general, the line-broadening mechanism in “‘C n.m.r. spectrum can be separated in mechanisms that leacl to homogeneous broadening arising from molecular motion in the solid and mechanisms that lead to inhomogeneous broadening appearing as a chemical-

422

ZEOLITES,

7997, Vol 17, June

shift dispersion.“” Even if a strong templating action is involved, an organic occluclecl in a zeolite cage is not entirely fixed and mobility will occur to some extent. In the case of TMA in sodalite, it provicles a relaxation pathway foi- the “‘C nia$netization via dipolar ‘“C-“N interaction with the ” N (quaclrupolar) spin”“*“” or modulates the resonance frequency via the anisotropic chemical shift.‘” The clominant inhomogeneous interactions are anisotropy in the chemical shift and scalar ’ H-‘“C and clipolar ’ H-“‘C coupling.‘i” In the absence of any internal TMA methyl group motion in the socialite p-cage, the broadening due to dipolar interaction with neighboring hydrogen ancl nitrogen nuclei wouIcl have been about 20 kHz. Therefore, a substantial spontaneous averaging of these dipolar interactions, and the chemical shift dispersion along with it clue to the same motion, is concluded. A relatively large residual linewidth of the TMA peak, not yet averaged out by the internal motion, manifests itself in the presence of high-power proton decoupling, 655 Hz (F&J-P 3b). High-power proton decoupling averages out all dipolar proton couplings. Consequently, the line broadening of 633 Hz is stemmilig from the dipolar “‘C-“‘N interaction and anisotropy in the chemical shift. Therefore, since the linewidth in the high-i-esolution spectrum (I;igz/~ 2) is only 32 Hz, a broadening of at least 623 Hz originates

90

70

10 Pm

Figure 3 Room-temperature 75.48 MHz 13C n.m.r. linenarrowing spectra of TMA in sodalite under various conditions: the static spectrum, i.e., no proton decoupling and no MAS; full-width at half-height was 1.03 kHz (a). The spectrum recorded with high-power proton decoupling, but no MAS, 655 Hz (b). The spectrum recorded with MAS, but no proton-decoupling, 173 Hz (c). The experimental parameters are specified in Figure 4

Organic-zeolite interactions: from these cot~trihittiotts. The IlliljOl~ ptt’~ of tl1e linewidth is assumed to be due to strong coupling with the ’ ’ N quadrttpolar spin, and ottly a titit~or part is assumed to be due to the residual anisotropy in the chemical shift. .I‘he peak position in the highresolution ‘Y: t1.m.r. specfra of orgatiics occlt’ded itt the pores of dif’ferent zeolites has been correlated empirically to the pore diameters.” From this correlation and the calculated pore-diameter flt’ctit;ttioti as computecl I>), our MD stud?:, it can t-eatlil), be calculated that a maximum contrtbittioti of 1 ppm (= 75 Hz) to the line I~roatlet~it~g due to dispersion in the chemical shift should be obser\~ed in the static proton decoitplecl ‘:‘C n.m.r. specfrttni (I;igrrt~ 36). Co’iseqttently, the remaining cotirributioti to the total line broadening of 655 Hz can be entirely ascribecl to hotiiogetieotts broadening ii;’ rhe dipolar “t<:-’ ‘N quaciri’poIar coupling. In this respecf, it shottltl be noted that the above-tiieti~iotiecl cotitril~ittiotts causing this line broacleniti~ r are already substatitiall!~ averaged 01’1 by the high ThlA mobiliry. Therefore. the 623 Hz provicles only a lower limit to the dipolat ‘Y-’ ‘N coupling and the atiisotrop)’ in the chemical shift. The 32 Hz linewidth obser\~etl i’t the hjghresolution spectrum (Figr,t.c# 2) is c;it~setl by rhenttc~~lshift dispersion and coupling with the ’ ‘N spin. Hence, it follows that the dispersion it1 the chemical shift is maximally about 0.4 ppm. Using the empirical relation bet\veen chemical shict and pore cliatiieter.“~ a maximum fluctuation of 1 A in the sodalite p-cage diameter is nieasitred. Finally, it should be noted that the TMA motion in socialite is not entirely isotropic (i.e., Ut~ownian motion as in liquids) ancl that some residual, anisotropic broadening (chemical shift and clip&t“‘C“‘N coupling) is left on the millisecond time scale since MAS is still capable of averaging the static linewidth further, i.e., down to 173 Hz (Fiq/rcJ 3~). This indicates a hindered rotation of TMA in the sodalite p-cage, which can be envisageci as a “hopping” motion from one orientation to another. However, the hopping frequency is very fast on the millisecond time scale. confirm In general, the “‘C n.m.r. experiments that TMA in sodalite has a considerable motional freedom on the millisecond time scale, as expected on the basis of the ‘H n.m.r. results. Spontaneous TMA motion already averages out the major part of the line-broadening interactions.

Comparison computational

of spectroscopic results

and

It can be concluded from the present n.m.r. results that TMA templates occluded in the sodalite P-cage have a considerable mobility. At least two types of TMA motion, each with its own specific time scale, have to be invoked to account for all the data. From ‘H n.nl.r. measurements it becomes cleat that CHs rotation about the N-C bonci and/or a librational motion of the N-C bond occurs on the

C.J.J.

den Ouden

et al.

microsecond time scale fill- TMA methyl groups occluded in zeolites. In contrast, MD simulations on the same system indicate that there is no rapid rotation of CH:+ groups ahut the N-C bond because of severe obstruction by the sodalite &cage atoms. Since MD deals with a much shorter time scale (100 ps), the currently observed rotation about the N-C bond must be slow on the MD time scale but fast on the ‘H 11.m.r. time scale (microseconds). The librational motion of the N-C bonds in TMA is observed to be fast on the MD time scale, and thus these movements are also traced with ‘H n.tn.r. It is worth noting that MD predicts a freezing of librational motion of TMA at temperatures around 123 K. This temperature is in reasonable agreement with the temperature at which immobility starts to show up in our ‘H 11.m.r. measurements. “‘C t1.m.r. also reveals TMA occluded in sodalite given the narrow peak in the static Cl’ “‘C n.m.r. spectrum (l;iguu~ 30). With MD, a rather large fluctuation in the distance between two diametrically opposed osygen atoms in the p-cage is observeci: 0.4 A. Such large fluctttations cati also he cleclitcecl from the high-resoltttiotI (:I’-i\/lAS spectrum (F;~qt~r~ -3) in which a linewicith of (1.4 ppm is observed. This linewidth $ati be correlated with a maximtt~~~ fluctiw tion of 1 A of the sodalite fl-cage diameter.

CONCLUSIONS An experimental and coml”tf~ttion;tl approach has been followed to describe and explain the dynamics of a TMA cation occluded in a p-cage in zeolite sodalite. Both the 11.111.1‘.techniques (solid-state “H n.m.r. and CP-MA.5 “‘C t1.m.r) and the com~~‘tt;ttion~t~ results from an MD simulatiott inclirate that the TMA rotational cation is able to perform a restricted movement at temperatures above some I‘)5 _. K . Moreover, internal rotation of the TM A <:H:t groups about the C-N axis was observed in the n.m.r. experiments but not in the MD simulations. This might indicate that these rotations are slow on the MD time scale. From MD, it followed that the tumbling of TMA is stimulated by the fluctuations of the distance between two diametrically opposeci oxygen atoms of the sodalite p-cage. Furthermore, MD simulations clearly show the four distinct and equivalent sites in the sodalite P-cage whet-e the TMA CH:, groups have the highest probability to reside. The simulations suggest that the TMA-sodalite system is not a static one, but that TMA very rapidly “hops” from one equilibrium orientation to another equivalent one. Deuterium n.m.r. reveals a smaller rotational freedom of the TMA cation in the bromide salt as in the P-cage. Clearly, the interaction between the TMA cations and the zeolite are much weaker than those between TMA and the bromine ions. This observation is consistent with earlier reported inelastic neutron scattering studies.37 The close agreement between the results obtained from well-established and reliable spectroscopic tech-

ZEOLITES,

1991, Vol 11, June

423

Organic-zeolite

interactions:

C.J.J.

den

Ouden

et al.

niques and from an MD simulation indicates that computational approaches in zeolite research have gradually de ve 1oped to valuable tools. These tools can be successfully applied for the interpretation and prediction of the interaction and dynamics of t-elatively large organic molecul& in zeolites. It should be mentioned, however, that agreement with experimental (spectroscopic) results is only obtained if the dynamics of the zeolite structure is taken into account. This observation also indicates that MD simulations of the diffusivity of bulky molecules in zeolites have only little value if rigid zeolite structures are used. A point that finally needs to be addressed relates to the templatihg action of the TMA cation in the b-cage during the fc%-mationlstabilization of zeolite sodalite. The classical concept of a template is that of an organic molecule or cation that tightly fits within a porous zeolite (“hand in glove”). The high-energy gain due to enhanced interfacial energy is supposed to be responsible for the structure-directing effect of the template during the nucleation step and for the stabilizing effect on the zeolite crystals formed during the subsequent crystallization. As a result of the right fit of the template and the intracrystalline space, a ternplating material is generally considered to be frozen in the zeolite. The results of the present study show that the classical idea of templating might need some revision. Apparently, the perfect fit of a TMA entity in a P-cage does not rule out the occurrence of rapid rotational hopping-type movements of a template. However, the (time-averaged) orientation of the TMA cation is well defined. From a thermodynamic point of view, the observed dynamics of TMA might even contribute to the ternplating effect, since entropic terms associated with the intracrystalline movements of the organic will end up as a favorable contribution to the minimization of the chemical potential of the crystallized material. It should be noted, however, that the rapid internal motion as observed here might be specific to the TMA@-cage system and cannot be generally extended to other systems such as TPA cations in zeolite MFI.

REFERENCES

ZEOLITES,

10

11 12 13 14 15 16 17 18 19 20 21

22

23 24 25 26 27 28 29 30 31 32 33 34

1 Flanigen, E.M. Adv. Chem. Ser. 1973, 121, 119 2 Lok, B.M., Zeolites 1983, 3, 282 3 Boxhoorn, G., Sudmeyer, 0. and Van Kasteren, P.H.G. J. Chem. Sot., Chem. Commun. 1983, 1416; Groenen, E.J.J., Kortbeek, A.G.T.G., Mackay, M. and 0. Sudmeyer Zeolites 1986, 6, 403; Van den Berg, J.P., De Jong-Versloot, P.C., Keijsper, J.J. and Post, M.F.M., Stud. Surf. Sci. Catal. 1988,

424

9

1991, Vol 11, June

35 36 37

37,85; Keijsper, J.J. and Post, M.F.M. ACS Symp. Ser. 398, Am. Chem. Sot., Washington, DC, 1989, p. 28 Barrer, R.M. Stud. Surf. Sci. Catal. 1985, 24, 1 Van Santen, !?.A., Keijsper, J.J., Ooms, G. and Kortbeek, A.G.T.G. Stud. Surf. Sci. Catal. 1986, 28, 169 Baerlocher, C. and Meier, W.M., He/v. Chim. Acta 1969, 52, 1853 Leherte, L., Lie, G.C., Swamy, K.N., Clementi, E., Derouane, E.G. and Andre, J.M. Chem. fhys. Lett. 1988, 145(3), 237 Pickett, S.D., Nowak, A.K., Thomas, J.M., Swift, J., Peterson, B.K., Cheetham, A.K., Den Ouden, C.J.J., Smit, B. and Post, M.F.M. J. Phys. Chem. 1990.94, 1233 Davis, J.H., Jeffrey, K.P., Bloom, M., Valic, M.E. and Higgs, T.P. Chem. Phys. Lett. 1976, 42, 390 Datema, K.P., Van Boxtel, B.J.H. and Hemminga, M.A., J. Magn. Reson. 1988, 77, 372; Datema, K.P., De Jager, P.A. andHemminga, M.A., J. Magn, Reson. 1989,81, 559 Soies. H.W. Colloid Polvm. Sci. 1983. 261. 193 Shies] H.W. Adv. Polym. SC;. 1985, Sk. 23. Eckman, R. J. Chem. Phys. 1982, 76, 2767 Kustanovich, I., Fraenkel, D., Luz, 2.. Vega, S. and Zimmermann, H. J. fhys. Chem. 1988, 92, 4134 Vega, A.J. and Lutz, Z. Zeolites 1988, 6, 19 Eckman, R.R. and Vega, A.J. J. Am. Chem. Sot. 1983, 105, 4841 Vega, A.J. and Lutz, Z. J. Phys. Chem. 1987, 91, 365 Lutz, Z. and Vega, A.J. J. fhys. Chem. 1987, 91, 374 Lutz, Z. and Vega, A.J. J. Phys. Chem. 1986, 90, 4903 Mehring, M. Principles of High Resolution NMR in So/ids, Springer-Verlag, New York, 1983 Ernst, R.R. Bodenhausen. G. and Wokaun. A. Principles of Nuclear Magnetic Reson&ce in One and Two Dimensions, Clarendon Press, Oxford, UK, 1987 Boxhoorn, G., Van Saten, R.A., Van Erp, W.A., Hays, G.R., Huis, R. and Clague, D. J. Chem. Sot., Chem. Commun. 1982,264 Hayashi, S., Suzuki, K., Shin, S., Hayamizu, K. and Yamamoto, 0. Chem. fhys. Lett. 1985, 113, 368 Hayashi, S., Suzuki, K. and Hayamizu, K., J. Chem. Sot., Faraday Trans. 11989, 85(g), 2973 Allen, M.P. and Tildesley, D.J. Computer Simulation of Liquids, Clarendon Press, Oxford, UK, 1987 Sybyl Molecular Modelling Software Version 5.2, Tripos Associates, 1989 Kiselev, A.V., Lopatkin, A.A. and Schulga, A.A. Zeolites 1985, 5, 261 Den Ouden, C.J.J., Smit, B., Wielers, A.F.H., Jackson, R.A. and Nowak, A.K. Mol. Simul. 1989,4, 121 Van Beest, B.J., Verbeek, J. and Van Santen, R.A. Catal. Left. 1988, 1, 147 Sybyl Theory Manual Version 5.2, Tripos Associates, 1989 Spies, H.W. and Sillesur, H. J. Magn. Resort. 1981, 42, 381 Vega, A.J. and Lutz, Z. J. Phys. Chem. 1987, 66, 1803 Ikeda, R., Kubo, A. and McDowell, C.A. J. Phys. Chem. 1989, 93, 7315 Van der Hart, D.L., Earl, W.L. and Garoway, A.N. J. Magn. Reson. 1981,44,361 Hexem, J.G., Frey, M.H. and Opella, S.J. J. Chem. fhys. 1982, 77, 3847 Abragam, A. The Principles of Nuclear Magnetism, Clarendon Press, Oxford, UK, 1961 Brun, T.O., Curtiss, L.A., Itou, L.E., Kleb, R., Newsam, J.M., Beyerlein, R.A. and Vaughan, D.E.W. J. Am. Chem. Sot. 1987,109,4118