Studies in Surface Science and Catalysis

Advisoq EdRors: 6. DelmonandJ.T.Yates

BNL-66771
vol.49

ZEOLITES: FACTS,
FIGURES, FUTURE
Proceedings of the 8th International
The Netherlands, July 10–1 4, 1989
Zeolite Conference, Amsterdam,
RECEIVED

e
ocTt$@
o s-f I
Part A

8
IZC
Editors

P.A. Jacobs
K. U. Leuven, Leuven, Belgium

R.A. van Santen

T.U. Eindhoven, Eindhoven, The Netherlands

ELSEVIER Amsterdam — Oxford — New York — Tokyo 1989

 P.A.
JacobsandR.A vanSanten(Editors), FacLr,Figures, Future
Zeolites: 545
@ 1989Elsevier
Science
Publishers
B.V.,Amsterdam— Printed inThe Netherlands

MUL’fITECiINIQUE
ANALYSISOF THE LATTICE STRUCTURESOF HIGHLY SILICEOUS
ZEOLITES

C.A. FYFE,aH. GIES,a’bG.T. KOKOTAILOa,Y. FENGa, H. STROBL=, B. MARLERb

D.E. COXC.

a. Departmentof Chemistry,Universityof British Columbia,Vancouver, B.C.

Canada.
b. MineralogicalInstitute,CAU Kiel, FRG
c. BrookhavenNationalLaboratory,Upton, N.Y., U.S.A.

ABSTMCT
The combineduse of high-resolutionsolid state NMR techniques (both lD
: ~and},,2Cl~:
,~,itk~,,aj@@~ocron-based
powder x-ray diffractionstudies yields
‘ “dckiiledinfora@k on the lattice structuresof highly siliceous zeolites.
The two methods,are complementary,the former probing short range ordering
.W. s~rubtur!=while the latter is sensitive to long range orderings and
periodicities.

INTRODUCTIO}i

Structural investigationsof zeolites are hindered by twc) factors.

Firstly, although they are highly crystalline, they are usually micro-
crystallinewith dimensionsof the order of microns precludingthe straight-
forward applicationof single crystal diffractiontechniques. Seccmdly, the
distributionof Si and Al over the T-sites in the lattice is often disordered
in nature and only an average structure can be obtained. Although the data
from powder refinementscan be greatly improved by the use of synchrotrons
X-ray sources and Rietveld analysis techniques and single crystal
investigations of microcrystallites may also be possibleusing synchrotrons
at present combined informationfrom a variety
X-ray sources in the future,z
of techniques is necessary to deduce the lattice structures of these
materials. In this regard,high-resolutionsolid-stateNMR spectroscopyhas
emerged in recent years as an important complementary teclmique to
diffraction investigations,3reflecting as it does local environmentsand
short range orderings compared with the long range orderings and
periodicitiesto which diffractiontechniques are most sensitive. In the
present work we present several examples of the combined use of these
techniquesin the investigationof the lattice structuresof highly siliceous
analogsof zeolites.
 DISCLAIMER

This report was prepared as an account of work sponsored

by an agency of the United States Government. Neither
the United States Government nor any agency thereof, nor
any of their employees, make any warranty, express or
implied, or assumes any legal liability or responsibility for
the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or
represents that its use would not infringe privately owned
rights. Reference herein to any specific commercial
product, process, or service by trade name, trademark,
manufacturer, or otherwise does not necessarily constitute
or imply its endorsement, recommendation, or favoring by
the United States Government or any agency thereof. The
views and opinions of authors expressed herein do not
necessarily state or reflect those of the United States
Government or any agency thereof.
 DISCLAIMER

Portions of this document may be illegible

in electronic image products. Images are
produced from the best available original
document.
546

29 .
In the Sx HAS NMR spectra of highly siliceous zeolites, sharp
resonances are observed whose numbers and relative intensitiesreflect the
numbers and relative populations of the crystallographicallyinequivalent
sites in the unit cell and whose frequencies(chemicalshifts) reflect the
local geometries of the T sites.4 The amount of informationwhich can be
obtained from these spectra depends ultimately on the resolution of the
spectra and we have paid careful attention to the important factors
determiningthe spectral linewidths. By carefulhydrothermaltreatmentwhich
effects both aluminium removal and the healing of lattice defects and by
optimizing all of the spectroscopicvariables in the NMR experiment it is
29
possible to obtain ultra-high resolution Si lfASNMR spectra where the
linewidths of the resonances in some cases approach those of solution
spectra. As an example of improvementswhich can be achieved by these
29
procedures, the Si MAS NMR spectrumof zeolite ZSM-5 at ambient temperature
is shown in Figure 1. The linewidthsare -6Hz(-*07 ppm) which makes it
possible to resolve between 20 and 22 of the 24 expectedresonancesdepending
on the exact temperatureand confirmsthe monoclinicsymmetryof the crystal
5
structure. Excellent resolution has also been achievedwith samples of
zeolites Y, A, beta, mordenite, ZSM-11,ZSM-12,KZ-2 and others.

—-IiIIi

I , , t 1
- !08
- 1!8
PPM FROM TMS

Figure 1. 2gSi MAS NMR spectrumof highly crystallineZSM-5, O.OHZ

linebroadening.
 *

547

RESULTS AND DISCUSSION

This improved resolution when combined with synchrotronsx-ray powder

data gives quite detailed information on the lattice structures. For
.29
example, the Si MAS NMR spectrumof zeolite ZSM-12 (Figure”2A) shows seven
narrow resonancesof exactly equal intensity indicating seven symmetrically
6
inequivalent silicon atoms in the unit cell. This is in general agreement
7
with the structureproposed by La Pierre et al. However, synchrotrons
x-ray
data collected on exactly the same sample indicate a doubling of the
c-parameter“from 12.16A to 24.33A, the others being unaffected. Since the
NMR experiment unambiguouslyshowed that there was a maximum number of 7
symmetricallyinequivalentSi sites per unit cell, only space group C2/c
remained for the refinementof the structure. (Figure 2B) Assuming this

OBSERVEO

, jl ~ ~[

T3

T4

T2

r ,, ,., ,’

-105 -120
PP!4 FROM m.ls

Figure 2a. 2gSi MAS NMR spectrumof highly siliceousZeolite ZSM-12, O.OHZ
linebroadening.
b. Projectionof the ZSM-12 frameworkalong [0101.

space group symmetry,the structurerefined to an R factor of - 5-4% with all

bond lengths and angleswithin reasonablelimits.8 In the case of ZSM-5, the
increased resolutionmakes it possible to detect and to quantify changes in
the lattice structure induced by temperature and sorbed organic molecules
(Figure 3).9 In both cases, there is an induceciphase transition from
monoclinic (24 T-sites) to orthorhombic (12T-sites). In the case of
temperature variation there is a relatively sharP transition over a few
degrees between 355 K and 359 K. The addition of p=ylene however, causes a
548

A T-3$3K

c 35 PL WfdW D 20 PL =lYk.X10?4
i ‘1

Figure 3. Effect of temperatureand sorbed organicson the 29Si MAS NMR

spectrumof zeoliteASM-5. A. 393K B. 298K, 10 @ p-xylene/250
mg. C. 298K, 35@ pryidine/250mg D. 298K, 20pl
acetylacetone/250mg.

gradual increase in the proportionof orthorhombicform present, both phases

co-existingand highly ordered. The change is complete when approximately
1*6 molecules per unit cell have been sorbed. By ensuring that the
experimentaldata are quantitativelyreliableand by permutatingthe combined
effects of temperatureand sorbed p-~lene, concentration. A complete three
dimensionalphase diagram can be~onstructed as shown in Figure 4.9. The
lower planar area shows the relatively limited regime where the monoclinic
form is stablewhile the upper area correspondsto the stabilityrange of the
orthorhombicform for this sorbate. Similar experimentscan be carried out
with other sorbates. It is now possible to use the NMR data to choose the
exact conditions under which x-ray data will be most informative. Thus
synchrotrons
x-ray powder data have been collectedat the points on the phase
10
diagram indicated by arrows. The high temperature form with no sorbed
p-xylene refines to R= 5% in the space group Primawith appropriate bond
lengthsand angles. The other data sets are currentlyunder refinement.
 *

549

3s3

o
%
Ca3

;m
:1.

,.*....*O+.O
.. ....0....*...O .....4y.*
.. .... .. “...
.. ..y..*.
“-.. ..
‘*.:”:*:+++. .:*.::.+ ... ... ... . ... ...:&:....
...
=0:.+. .+. ..+ ... hr.”+: .+:... .+.....: +. .:*
.....+.+. +.. ++y 2.*
. ..:,
,,,
....:*.+..:.+:-+. .....=:,..+.+..
:.”*::..”.:.:+
a
...
...+.+
-.
1‘“4i..-;:*...”.*..+...2*’:+..::>
: “,”....7.-:. -
.
...
..
.+.
.
.. ..*;...:>*;.
. ..
. .-.... .....
“*..J8
.:* . ...0.:.. ... . . q: . . y.. ,m.
...
‘--..&:,.~.~:*.:4 ...+...+....’-.....”*.::.+
.. .. ... ... ~
.: .

:. ...?- -’&._:>
> ....0..
..%..
;
“

.-.
● ‘: /- !-.0

0 U2M04!OJ3U12,.4 Is 1.8 20

SORBATE

Figure 4. A three-dimensionalsurface of the crystallographic phase

relationshipbetween the concentrationof p-xylene (increasing
left to right) and temperature (increasingfront to back). The
vertical axis gives sites in the unit cell) phase present. The
solid line going left to right shows the effect of increasing
P-KYlene concentrationupon the crystal symmetryat ambient
temperature. The near vertical solid line at 355 K shows the
effect of temperature on an unloaded ZSM-5 sample. The solid
circlesare selected experimentalresults from the MAS W data,
and the dotted curves joining them show variationwith p-~lene
concentrationat constant temperature(leftto right) and
variationwith temperature at constant p-xyleneconcentration
(frontto back). (B) A two-dimensionalprojectionof part A as
indicatedshowing the relationshipof the c,rthorhombic and
monocliniccrystallographicphases with varyingtemperatureand
concentrationof p-~lene.
550

Further information on the nature of the interaction and improved

diffraction data can be obtained from NMR investigationsof the effect of
substitutingthe methyl groups in p-xylene with chlorine atoms which are
known from physical organic chemistry to have identical steric interactions
but differ in all other aspects. Figure 5 shows that the 2gSi MAS NMR
spectra of ZSM-5 loadedwith p-xylene,p-chlorotolueneand p-dichlorobenzene
are almost identical indicating that, at least for hydrocarbons, the
11
substrate-latticeinteractionsare based on size and shape alone.
Confirmation of this conclusion comes from the fact that the
three-dimensionalphase diagrams for p-xylene and p-dichlorobenzene are
identical when the concentrations are expressed in molecules/unit cell.
These &ta may now be used as a justificationfor using the p-dichlorobenzene
system as an exact model for the p-xylene/ZSM-5structure and synchrotrons
x-ray data have been collectedon this system. The advantage of the former
system is that the p-dichlorobenzenesorbatemakes a much larger contribution

B a

1 i 1 I 1 # 8 I t 1 4 1 I 1
-105 -120
P FM / TMS

Figure 5. 29Si MAS NMR spectra obtainedat 79.6 MHz of highly siliceous
ZSM-5 with A p-xylene (lO@/250 mg zeolite,B p-chlorotoluene
(1OIW25O mg zeolite);C.p-dichlorobenzene(lOmg/250 gm
zeolite.
 I

551

to the overall scattering due to the substantial electron density on the

chlorine atoms. This refinement is in progress and when complete should
provide the most detailed structural information on the nature of the
h,~st/guest interactions which induce the phase-transition of the
10
lattice.
Other structural investigations may be complicated by temperature
induced phase transitions. Thus, Figure 6 shows the variable temperature
2<3
Si MAS NMR spectra of a pure sample of zeolite Zm-11 synthesizedfree of
12
any intergrowths of ZSM-5 in our laboratory. Narrow resonances are
ol>servedin all of the spectra indicating a high degree of crystallinityand
local order. ‘The postulated lattice structure space group has seven
inequivalentsites of relative proportions 1:1:2:2:2:2:2.13

JJ-
320
M___ J’.-’L- 3e3
—

222 M 35a
J-JL-
318 3s3
J-’i---
313
N 348
JJ’’L--
20s
J4.- 243

ao3 J’Ji J“JL 32e

Figure 6. 2gSi MAS NMR spectra (79.6 MHz) of zeolite ZSM-11recorded at the
temperaturesindicated. Each spectrum is the result of 720 scans
wit~ a delay time of 10S.
552

Inspectionof Figure 6 shows that there are too many resonancesin the
room temperature(293K)spectrum and in fact the systemis in an intermediate
state between two limiting structuresat ambient. Using the data of Figure 6
as our guide, synchrotrons
powder x-ray data have been obtainedin the high
and low temperaturelimits. Data obtained at 373K refinedsmoothly (R=4.4%)
in the tetragonal space group I~m2 yielding the postulatedstructurewith
13
reasonable values for all bond lengths and angles. Refinementof the low
9
temperature limiting structure is currently in progress. The relative
intensitiesof the limiting high temperaturespectrumare in exact agreement
with the postulated structure while those of the low temperature form
indicate a lowering of the symmetry of the lattice with a basic structural
unit which still contains a total of twelve T-sites.
In high-resolution NMR studies in solution, two dimensional NMR
spectroscopyhas made it possible to detect two dimensionalcomectivities
14
betweenbonded atoms from the through-bondinteractionsbetweenthe nuclei.
In principle,at least, it should be possible to carry out such measurements
in the solid-stateand probe the bonding within a zeolite lattice. We have
recently extended our NMR investigationsof these systemsby showing for the
first time that three-dimensionallattice connectivitiescan be established
15
using two-dimensionalsolid-stateNMR techniques.
In order to test the validity and reliabilityof the experiments,a
system of known structure (zeoliteZSM-39) was chosen and synthesizedin the
completelysiliceousform from Si02 enriched in 2gSi to approximately80%.
The zSM-39 high temperature structure (spacegroup Fd3) contains three
crystallographically inequivalent sites Tl, T2, and T3 with relative
16,17
propor.rions8:32:96 and is shown schematically in Figure 7. At room

Figure 7. Schematicrepresentationof the latticestructureof ZSM-39. The

three T-sites T1 T2 and T3 are shown circledand the nature of
the attachedT a~oms indicated.
 ?

553

temperature, the symmetry is lower and the degeneracy of the T3 site is

lif”:ed,giving three resonancesof intensities 32:32:32. The solid state
29
Si MAS NMR spectrum shown at the top of Figure 8 clearly reflects the
asymmetric unit in the lattice, the assignments of the resonances being
unambiguous from the known site occupancies. In the present work we will
interpret the interactionsas being, to a first approx~~tion, between the

three sites T ~, T2, andT3. From the structureshown in Figure 7,”it can be

seen that there are direct Si-O-Si connectionsbetween T,, and Tqand also
1. L
between T2 and T3 but none between Tl and T3.

I
T2 T3

TI

, -m -lM -110 -il!l **.

Pm
\
+,, +,* .~
L-
-* -,-a, “

Figure 8, 2gSi MAS NMR spectrumof the 2gSi enrichedsample of ZSM-39 used
in the 2D NMR experiments.

29 18
Figure 9 shows the result of a Si 2D NMR spin diffusion experiment
carried out using the parametersgiven in the figure caption. The attraction
of this experiment,analogousto the NOESY techniquein solution lies in the
strong dependence of the interaction on the through-space internuclear
d~stance (1/r6 if there is motion and l/r3 if there is not). Because of the
large differences in internuclear distance between “directly” connected
29 Si-o_29SiVs _ ~ ~ ~ for
silicons and others further removed (-3 L for —“
29~i-O-Si-O-2g Si) it was hopedthat they would be clearly differentiated.
Figure 8 shows that the expected connectivitiesTlT2 and T2T3 are clearlY
observed while that betweenT and T is not seen. Although this result is
1 3
encouraging and certainlymerits further investigations,we feel that some
caution should be usea in the interpretationof data of this type as the
theoreticaldescr~ptionof the interactionsinvolvesadditionalcontributions
from non-fiis~ance dependent parameters, namely the chemical shift
s.nisotropiesand the isotropicshift differencesof the interactingnuclei
554

19
whose importanceis difficultto estimate. Furtherexperimentson systems
of known structurewill be needed to properly establishwhether experiments
of this type are reliableprobes of latticeconnectivitiesin these systems.

T2 T3
TI

U4

ra

ii$

lie

M4

ils

MO

:00

$Cd

PPM

Figure 9. Contourplot of 2gSi 2D-spindiffusionexperimenton ZSM-39 with

128 experiments,8 scans in each experiment,sweepwidthof 5 ICHZ
and 256 data points collectedduring acquisition. The fixed delay
during which spin diffusionoccurswas 10 sec., the spinningrate
2 kHz and the total~tie for the experimentapproximately6 hours.
Sine bell squaredanodizationwas used and the plot has been
symmetrized.

An alternative and more reliable and unambiguousmethod is the COSY

20
technique which relies on the scalar couplings between the interacting
29
nuclei which operate through the bonding electrons, Si-O-2gSi
connectivities within the reference molecule Q8M8 have recently been
21
“demonstrated by Berm et al . The resultsof a 2D COSY experimenton the
ZSM-39 sample are presentedin Figure 10. The sequencewas identicalto that
used in solution studies except that the initial preparation step was a
 555

22
cross-polarization sequence and the experimentwas carried out using the
conditions given in the figure caption. Again, the connectivitiesT1T2 and
T2T3 are exactly those expected and in this case the result is quite
.-
13
unambiguous.

A -i24
T3
-122

-120

@
@ -iie

-1 is

ccl
00 -f 14

T2 T3 +22

-110

0
-104
0

TI T2 -:02

* Fu
-104 - 10E -i 10 -i 12 -%:4 +16 -1 i2 -120 -122

Figure 10. Contour and stacked plots of a 2D COSY experiment on ZSM-39 at

373K. 128 Experiments,64 scans in each experiment,5 kHz
sweepwidth,256 data points on acquisitionand a fixed delay of 5
m.sec. Sine bell anodizationwas used and the data are presented
without symmetrizationor smoothing. The total experimentaltime
was approximately23 hours.
*

556

Although the two dimensionalexperimentsdescribed above are very time

consumingand demandingin terms of both sample preparationand spectroscopy,
they greatly extend the potential of the technique in the investigationof
zeolite lattice structures and provide informationnot obtainable by ocher
means. Furtherwork in this area is in progress including the developmentof
an experimentalprotocolto obtain such data on natural abundance samples.
Thus, considerable insight into the structures of zeolites can be
obtained by the combineduse of solid state NMR and diffractiontechniques.
The two are complementary in nature and should always be used together
preferablyon the same sample to obtain the maximum information.

ACKNOWLEDGEMENTS

The authors acknowledge the financial assistance of the Natural

Sciencesand EngineeringResearch Council of Canada in the form of Operating
and EquipmentGrants (CAF) and a Graduate Scholarship (HS). J.G. and G.T.K.
acknowledgethe financialsupport of the Alexander von Humboldt Foundation.
C.A.F. acknowledgesthe award of a Killam Research Fellowship by the Canada
Council. The work was supportedby the Division of Material Sciences,U.S.
Departmentof Energy, ContractNo. DEAC02-76CHOO016.

RIEE.RENCES

la. NcCusker,L.B.; Baerlocher,C. Proc, Sixth Int. Conf. Zeolites, Reno,

1983.
b. Rietveld,H.M.. J. ADD1. Crvst. 1969. 2. 65.
2. Eisenberg,P.;”Newma~,J.B.-;Leonowicz,-M.E.;Vaughan, D.E.W, —~
Nature
1984,309, 45.
3. Englehardt,G.; Michel, D., High ResolutionSolid State NMR of Zeolites
and Related Systems,John Wiley L Sons, (1987).
4a. Fyfe, C.A.; Gobbi, G.C.; Murphy,W.J.; Ozubko, R.S., and Slack, D.A.,
Chem. Lett. 1983, 1547.
b. Fyfe, C.A.; Gobbi, G.C., Murphy,W.J., Ozubko, R.S., and slack, D.A., AJ
Am. Chem. Sot. 19+4, 106, 4435. —
5. Fyfe, C.A., O’Brzen,J.H.; and Strobl,H., Nature, 1987, ~ 6~10.
6. Fyfe, C.A.; Strobl,H.; Kokotailo,G.T.; Pasztor, C.T.; Barlow, G.E.;
Bradley,S.; Zeolites,1988, Q, 132.
7. LaPierre,R.B. et al Zeolites 1985, ~, 346.
8. Gies, H.; Marler, B; Fyfe, C.A.; Kokotailo,G.T., COX, 1).E. unpublished
results.
9. Fyfe, C.A.; Strobl,H., Kokotailo,G.T.; Kennedy, G.J.; and Barlow,
G.E., J. ~. Chem. Sot. 1988,~, 3373.
10. Gies, J.; Fyfe, C.A.; Strobl,H.; Kokotailo,G.T.; Feng, y.; COX, D.~.;
Unpublishedresults.
11. Fyfe, C.A.; Gies, H.; Kokotailo,G.T.; pasztor,C.; Strobl, H.; COX,
D.E., Can. J. Chem. (1988),~, 1942.
 557

12a. Fyfe, C.A.; Kokotailo, G:T.; Kennedy,G.J.; and De Schutter,C.T., AJ

Chin.Sot. Chem. Cormnun.1985, 306.
b. Fyfe, C.A.; Gies, H.; Kokotailo~G.T.; Pasztor, C.; Strobl,H.; Cox,
D.E., J. Amer. Chem. SOC. (In press).
Kokotailo, G.T.; Chu, P.; Lawton, S.L., and Meier, W.M., —~
lJature 1978
~, 119.
14a. Berm, R. and Gunther, J., “ModernPulse Methods in High ResolutionNMR
Spectroscopy”,Angew. Chem. Int. Ed. Engl. 22, 250 (1983).
b. Bax, A., “Two Dimensional Nuclear MagneticResonance in Liquids”,Delft
University Press (1982).
c. Delorme, A.E., “Modem NMR Techniquesfor ChemistryResearch”,Pergamon
Press (1987).
d. Sanders, J. And Hunter, B., “ModernNMR Spectroscopy,A guide for
Chemists”,Oxford UniversityPress (1987).
15. Fyfe, C.A.;.Gies, H.; Geng, Y.; submitted.
16. Schlenker,J.L.. Dwyer, F.G., Jenkins,E.E., Rohrbaugh,W.J. Kokotailo,
G.T.; and Meier, W.M., Nature 1981,~, 340.
17. Gies, H.; Liebau, F.; G~~.; Angew. Chem., 1982,~, 214. Gies, H.;
Z. Kirstallogr., 1984,.~, 73.
18. Szeverenyi,N.lf.;Sulllvan,M.J.; Maciel, G.E.; J. Msg. Res., 1982,47,
462.
19a. Vanderhart, D.L.; J. Msg. Res., 1987,~, 13.
b. Kubo, A.; McDowell, C.A.; J. Chem. Sot. Farad. Trans I, 1988,84, (In
press).
20. Muller, L.; Kumar, A.; Ernst, R.R.; J. Chem. Phys., 1975,Q, 5490.
21. Berm, R.; Grandey, H.; Brevard,C.; Pagelot,A.; J. Chem. Sot. Chem.
Commun., 1988, 102.
22. Pines, A.; Gibby, M.G.: Waugh, J.S.; Chem. Phys. Lett., 1972,~, 273.