Annual Reports on NMR Spectroscopy, Volume 28

ANNUAL REPORTS ON NMR SPECTROSCOPY Edited by I. ANDO" and G. A. WEBBY *Department of Polymer Chemistry, Tokyo Institut...

Author: Isao Ando

21 downloads 675 Views 9MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ANNUAL REPORTS ON

NMR SPECTROSCOPY Edited by I. ANDO" and G. A. WEBBY *Department of Polymer Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo, Japan WDepartment of Chemistry, University of Surrey, Guildford, Surrey, England

VOLUME 28

ACADEMIC PRESS Harcoutt Brace & Company, Publishers London 0 San Diego New York Boston 0 Sydney 0 Tokyo Toronto

ACADEMIC PRESS LIMITED 24-28 Oval Road, LONDON NW17DX

U.S. Edition Published by ACADEMIC PRESS INC. San Diego, CA 92101 This book is printed on acid free paper

Copyright 0 1994 ACADEMIC PRESS LIMITED

AN Rights Reserved

N o part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording, or any information storage and retrieval system without permission in writing from the publisher A catalogue record for this book is available from the British Library

ISBN 0-12-505328-2 ISSN 0066-4103

Phototypesetting by Keyset Composition, Colchester, Essex Printed by Hartnolls Limited, Bodmin, Cornwall

LIST OF CONTRIBUTORS I . Ando, Department of Polymer Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo, Japan. S . Ando, NTT Interdisciplinary Research Laboratories, Midori-cho, Musashino-shi, Tokyo, Japan.

F. D . Blum, Department of Chemistry and Materials Research Center, University of Missouri-Rolla, Rolla, MO 65401, USA. H . Ernst, Universitat Leipzig, Fachbreich Physik, Linnkstr. 5 , 04103 Leiprig, Germany. S . Hayashi, National Chemical Laboratory for Industry, Tsukuba, Ibaraki 305,Japan.

H . Kawazoe, Research Laboratory of Engineering Materials, Tokyo Institute of Technology, Natatsuta, Midori-ku, Yokohama 227, Japan.

H . Kurosu, Department of Polymer Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo, Japan. H . Pfeifer, Universitat Leipzig, Fachbreich Physik, Linnkstr. 5 , 04103 Leipzig, Germany. H . Yoshimizu, Department of Materials Science and Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya, Japan.

This Page Intentionally Left Blank

PREFACE This is the first of a series of special issues of Annual Reports on NMR Spectroscopy, and it is devoted to Applications of NMR to New Materials. It is a great pleasure for me to welcome as co-editor of this issue Professor I. Ando of the Tokyo Institute of Technology. Professor Ando and I have cooperated in many areas of NMR over the past 15 years and I am especially happy to have been able to draw upon his considerable knowledge of new materials in the planning and execution of this volume. The topics covered are high-temperature superconductors, ceramics, zeolites, high-order polymer structure, and organic thin films. While not being exhaustive this coverage represents timely reports on many of the most active areas of new materials science. I wish to express my sincere thanks to all of the contributors, and my co-editor, for their very considerable help and cooperation in the production of this volume. University of Surrey Guildford, Surrey England

G. A. WEBB May 1993

This Page Intentionally Left Blank

CONTENTS List of contributors Preface .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

...

111

V

Application of NMR Spectroscopy to the Science and Technology of Glasses and Ceramics H. KAWAZOE 1. Introduction . . . . . . . . . . . . . . . . . 2. Identification and quantization of silicate anions in solid silicates by 29Si MAS-NMR and effects of bonding characteristics on the chemical shift. . . . . . . . . . . . . . . . . . 3. Distribution of silicate structures in silicate glasses by using 29Si MAS-NMR . . . . . . . . . . . . . . . . . . 4. Structure determination of “defects” in amorphous silica . . . 5. Structure determination of phosphate glasses by using 31P NMR . 6. Application of high field “B NMR to structure determination of borate and borosilicateminerals and glasses . . . . . . . 7. Application of NMR imaging to detect flaws in composite ceramics in green state . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . .

1 2

5 11 16

20 24 26

High-resolution Solid-state NMR Studies on Ceramics S. HAYASHI 1. Introduction . . . . . . . . . . 2. Experimental aspects of NMR techniques 3. Interpretation of NMR parameters. . . 4. Conventional ceramics . . . . . . . 5 . High-performance ceramics . . . . . 6. Bioceramics . . . . . . . . . . 7. Concluding remarks. . . . . . . . References. . . . . . . . . . .

, . . . . . . . . . , . . . . . . . . . , . . . . . . . . . , . . . . , . . . .

. . . . . . . . . .

. . . . .

29 30 33 36 66 83 83 84

. . . . . . . . . . . . . . . .

91 93

. . . . . .

.

. .

NMR Studies of Zeolites H. PFEIFER and H. ERNST 1. Introduction . . . . . 2. Frameworkof zeolites . . .

CONTENTS

3. Bronsted acid sites . . . . . . . . . . . . . . . . 4 . Lewis acid sitedextra-framework aluminium . . . . . . . 5. Structure of adsorbed molecules . . . . . . . . . . . 6. Molecular diffusion . . . . . . . . . . . . . . . . 7. Chemical reactions . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . .

106 126 132 138 159 176

NMR Studies of Higher-order Structures of Solid Polymers H . KUROSU. S. ANDO. H . YOSHIMIZU and I . ANDO 1. Introduction . . . . . . . . . . . . . 2. Engineering plastics and high-performance polymers 3. Polymer alloys . . . . . . . . . . . . 4 . Natural polymers . . . . . . . . . . . . 5. Conclusion. . . . . . . . . . . . . . References . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

189 190 239 251 269 269

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

277' 278 280 300 315 315 315

. . . . . . . . . . . . . . . . . . . . .

323

NMR Studies of Organic Thin Films F . D . BLUM 1. Introduction . . . 2. Background . . . 3. Polymers at interfaces 4 . Surface-active agents 5 . Conclusions . . . Acknowledgement . References . . . .

Index

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

Application of NMR Spectroscopy to the Science and Technology of Glasses and Ceramics H. KAWAZOE Research Laboratory of Engineering Materials, Tokyo Institute of Technology, Natatsuta, Midori-ku, Yokohama 227, Japan

1. Introduction 2. Identification and quantization of silicate anions in solid silicates by 29Si MAS-NMR and effects of bonding characteristics on the chemical shift 3. Distribution of silicate structures in silicate glasses by using 29SiMASNMR 4. Structure determination of “defects” in amorphous silica 5. Structure determination of phosphate glasses by using 31PNMR 6. Application of high field “B NMR to structure determination of borate and borosilicate minerals and glasses 7. Application of NMR imaging to detect flaws in composite ceramics in green state References

1

2

5 11 16 20

24 26

1. INTRODUCTION Nuclear magnetic resonance (NMR) spectroscopy has a characteristic among spectroscopic methods in that it is appropriate for identifying and quantifying structural units involved in molecules or solids rather than determining their detailed geometrical structure. This is especially useful in structure determination of amorphous materials and glasses for which the diffraction methods are less effective. Information on structure is essential to understanding the properties of solid materials. The first systematic study of glass structure by using NMR spectroscopy’ appeared shortly after the first observation of nuclear magnetic resonance phenomena in 1945.2.3In the 1960s broad-line NMR instruments were only available for solids. Therefore, the application was limited to elucidation of local structures around “B in borate and borosilicate glasses. Boron in these glasses usually has two different local structures, tetrahedral and trigonal, ANNUAL REPORTS ON NMR SPECl’ROSCOPY VOLUME zs ISBN n-12-~05328-2

Copyright 0 1994 Academic Prws Limited A / / righu of reproduction in any form reserved

2

H.KAWAZOE

depending upon local chemical fields. In the case of "B NMR, line shape and line width are markedly affected by the interaction between the nuclear quadrupole moment of "B and an electric field gradient at the position of the nucleus. The line width broadening due to dipolar interactions between the nuclei and the anisotropy of chemical shifts does not exceed the broadening due to the nuclear quadrupole effect. This enables one to analyse local structures around boron atoms in the oxide glasses without using the recently developed techniques such as magic angle spinning (MAS) .4 The structure determination contributed considerably to solving the problem of the "boric oxide anomaly", which has been important from the viewpoints of glass science and glass technology. The magic angle spinning (MAS) technique4 for solids opened new applications of high resolution NMR in the sciences and technologies of glasses and ceramics. Decrease in line width by MAS is not satisfactory in some cases for the detailed analysis of chemical structures of the materials, but is still effective in some applications. The distinction of Si04 structural units in silicates, in solids and solution^,^ which are usually abbreviated as Qo, Q1, Q2, Q3 and Q4,where Q stands for the tetrahedrally coordinated Si by four oxygens and the superscripts for the number of bridging oxygens (nearest neighbouring Si04 units), has been established. The chemical shifts in 29Si MAS NMR of the respective Q' species were given empirically by collecting the spectra of the minerals with known crystal structures. After the pioneering work on 29Si MAS-NMR, MAS or MAS-CP methods on 27A16and 31P7 were applied to a wide variety of problems in ceramics and For instance, the structure of zeolite gel was determined by a combination of 27Aland 29SiNMR." The purpose of the present article is to overview the recent advances in applications of NMR to the sciences and technologies of glasses and ceramics from the viewpoint of the materials scientist. There have been some published and excellent reviews from this standpoint' '-12 and this chapter will deal with the topics which might be epoch making in the applications. 2. IDENTIFICATION AND QUANTIZATION OF SILICATE ANIONS IN SOLID SILICATES BY 29SiMAS-NMR AND EFFECTS OF BONDING CHARACTERISTICSON THE CHEMICAL SHIFT

Identification and quantization of various types of silicate anions by reliable experiments have long been an important problem in glass science and mineralogy. 29Si MAS-NMR enables one to evaluate the distribution of silicate anions in the materials. Lippmaa ef ~ 1 . applied ~ ~ ' ~the technique to establish one by one correspondence between the isotropic chemical shift value and a particular Q' unit in solid silicates with known crystal structures.

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

/"i Q"

Silicate anions in solution

Q'

Q2

3

d

Q3

Solid silicates

-110 ppmlTMS)

t

-70

-80

-90

-100

Fig. 1. Ranges of 29Si chemical shifts of different structural units of silicate anions in solid silicates and silicate solutions.

Figure 1 shows ranges in the observed chemical shift values of 29Si in Q' species dissolved in solutions and solid silicates with respect to tetramethyl silane (TMS). The chemical shift is affected by the type of charge compensating cations. The empirical correlation can be applied to Na, K, Ca, Mg, Zn and alkylammonium salts in the case of solid silicates. The ranges in the chemical shift of the respective Q' units were well separated from each other for a particular counter cation, and this relation is extremely useful in identifying and quantizing Q' units contained in amorphous silicates. A considerable effort has been concentrated to elucidate the relation between the chemical shift values of 29Si in Q4 units and their structural characteristics. Smith and Bla~kwell'~ found that a linear relation holds between the chemical shift values and mean secant of Si-0-Si bond angles involved in several types of silica polymorphs. The relation was used in estimating a distribution in Si-0-Si bond angle in silica g1a~ses.l~ Comparing with the results of X-ray analysis, the bond angles were found to distribute from 90" to 180" consistent with X-ray analysis, but a shape of the distribution function was different. Ramdas and Klinowski16examined extensively the effect of A1 neighbours on the chemical shift of 29Si, which is important in zeolite chemistry. The number of second neighbouring A1 to a central Q4 unit ranges from 0 to 4, and the structures are differentiated by the expression Si(0-Al)n(O-Si)4 or in short Si(nA1). A Iinear correlation was found to hold between the isotropic chemical shift values of 29Siand the sum of bond lengths between a central silicon and respective second neighbouring silicon or aluminium atoms. However, a different straight line was necessary in reproducing the chemical shifts for the structures with the same n value, the number of A1 second neighbours. The above-mentioned relations were included in the more unified and extended relation found for silicates involving Q4, Q3 and Q2 species with known structures. l7 Agreement of the calculated chemical shifts with the

4

H.KAWAZOE

Experimental chemical shift ( p p rn) Fig. 2. Experimental chemical shift plotted against chemical shift calculated from sorosilicates; A, inosilicates; 0 , equations (1) and (2). X , orthosilicates;

+,

phyllosilicates;0, tectosilicates.

observed shifts was examined in Fig. 2, and was found to be satisfactory. The correlation coefficient was 0.991 over 76 reliable data. Here the chemical shifts were calculated by using a relation:

S = 650.08f - 56.06 (ppm with respect to TMS)

(1)

where ,f was obtained by , I ! =

2 Sj[(1- 3 cos2e i ) / 3 ~ :[cos ] cril(C0S

Cyi

- l)]

i

In the calculation, summation runs over four bonds and Si is the bond valence18 of the ith bond, given by Si = exp[(ro - rj)/0.37].

(3)

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

5

Si Fig. 3. Diagram showing the definitions of the angles 8 and a,and the lengths r and R. X is any second neighbour cation.

Oi, Ri, and cyi are defined as presented in Fig. 3. The success of the estimation is thought to be derived from the fact that this approach appropriately takes into consideration the geometrical characteristic of the surroundings of the central silicon and the effect of types of the second nearest cations.

3. DISTRIBUTION OF SILICATE STRUCTURES IN SILICATE GLASSES BY USING 29SiMAS-NMR This is one of the topics frequently discussed in the literature.'' In this section it will be discussed from a slightly different viewpoint, that is from the eyes of the glass scientist. The isotropic component of the chemical shift tensors of silicate anions, which are dissolved in solids, with different polymerization degree to Q4) measured from the resonance frequency of TMS, was given in Fig. l.5,'3 Each ionic species is characterized by the corresponding chemical shift and the overlapping between those for neighbouring Q's is not large. The above-mentioned empirical relation between the structures of silicate anions and the corresponding chemical shifts were successfully used in the structure determination of R20-Si02 glasses,193mwhere R means alkali cations. Distribution of different types of silicate anions in the glasses has been thought to be similar with those in the crystalline phases having the

(ao

6

H.KAWAZOE

same chemical composition. One mole of R 2 0 incorporated into Si02 is assumed to generate two non-bridging oxygens in the following reaction:

In this reaction two Q4 building units are converted to two Q3s. The reaction model is extended to the composition regions with higher alkali concentrations. Total concentration of the non-bridging oxygens in the glasses was first and directly determined by using XPS spectra of 0 ls.21 A close agreement was observed between the experimental estimations and the predictions from the chemical reaction model. However, XPS results can not identify different Q' species having different numbers of non-bridging oxygens. Figure 4 shows 29Si MAS-NMR spectra of NazO-SiOZ glasses with NazO concentration of &SO mol%.19 It is noted that in the spectrum for silica glass (NAO; Omol% Na20) only a single peak with the chemical shift value of -109 ppm with respect to TMS was observed. From a comparison with the data in Fig. 1, the resonance absorption was assigned to Q4 species. A shoulder with chemical shift of -92 ppm appeared at the tail of the Q4 peak in the spectrum of NA6 (Na20 concentration of 10mol%). The additional peak was assigned similarly to Q3. Growth of the Q3 peak and decay of Q4 resonance with increasing Na20 concentration up to 33.3 mol% were clearly seen in the spectra for NAO, NA6, NAS, NA8 and NA7. Upon the further increase in NazO content to SO mol%, the Q3 peak was similarly replaced by a new peak with chemical shift of -78ppm, which was assigned to Q2 species. Small and additional peaks were observed in both upper and lower fields of the central absorption for the most of the spectra. These are spinning sidebands (SSB), which originate from an anisotropy of chemical shift tensor. In the case of Q4 units, four oxygens coordinating to a central Si are all bridging oxygens and chemically equivalent to each other. Thus Si has an almost isotropic chemical shift tensor in Q4, which results in a weak SSB in the spectra for NAO and NA6. In the case of Q3 and Q2 units one or two oxygens among the four ligands are non-bridging oxygens and an isotropic chemical shift tensor can no longer be expected. This gives rise to much stronger SSB in the spectra of the glasses with higher concentration of Na20. The gradual replacement of Q4 with Q3 on the addition of NazO into Si02 up to 33.3 mol% Na20 and a similar change from Q3 to Q2 roughly agreed with the prediction of equation 4. However, as seen in Fig. 5,19 which exhibits a comparison of 29Si MAS-NMR spectra of sodium di- and metasilicate crystals and glasses, line widths of the crystals were far narrower than those of the corresponding glasses. The extra broadening derives from site to site structural distribution in a particular Q' species. The

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

7

NA12

NAll

-50 400 -150 CHEMICAL YllFT (PPM)

Fig. 4. 29Si spectra obtained by MAS-NMR €or a series of sodium silicate glasses: (left) 0-33.3 mol% Na20; (right) 33.3-50 mol% Na20.

absorption peak of the glass with di- or metasilicate composition, for which only Q3 or Q2 species is expected, is too broad to be assigned to a single kind of Q' species. We notice here that there is a hierarchy in the structural distribution: one is the chemical distribution, that is the coexistence of different types of Q' species or polysilicate anionic species. The chemical distribution is not limited to anionic structures. Since the polyanions constitute ligand fields for the cations, coexistence of different kind of silicate anions gives rise to coexistence of Na ions having different counter anions or chemical fields. Alkali cations in silicate glasses are well known to be responsible for a wide variety of transport phenomena such as ionic conduction, alkali diffusion, ionic exchange and internal friction. The differently coordinated alkali ions are expected to contribute differently to the properties.22

8

H.KAWAZOE

Na20.2Si02 -CRYSTAL

Na20.SiOz CRYSTAL

o

-50 -100 -150 CHEMICAL SHIFT (PPM)

Fig. 5. 29Si MAS-NMR spectra for glass and crystalline modifications of Na20-2Si02 and Na204i02. Spinning sidebands are visible in the spectra.

The second kind of structural distribution is physical distribution or topological distribution, which is the site to site structural distribution within an ensemble of a single kind of chemical species. We think about Q3 units as an example. A tremendous number of Q3 species are contained even in a small piece of the glass. Principally all Q3 units have slightly different geometrical structures reflecting variations in second nearest neighbouring structures. Magnetic resonance spectroscopy is a very appropriate method to evaluate quantitatively the structural distributions in glasses.23 This is because the line broadening by the structural distribution is large enough to evaluate quantitatively and at the same time small enough not to disturb a reliable assignment.

APPLICATION OF NMR SPECTROSCOPY TO GLASSES A N D CERAMICS

PPM

9

PPM

7-

-120.00 -100.00 -80.00 PPM

.120.00 -lOo'.OO

-80.00 -6O:OO PPM

Fig. 6. Static (upper) and MAS (lower) NMR spectra of potassium tetrasilicate (left) and potassium disilicate (right).

The coexistence of two or more types of silicate anions in the stoichiometric alkali silicate glasses with di- and metasilicate compositions was confirmed clearly by a combination of high field MAS and static measurements. Figure 6 displays NMR spectra of 29Si in potassium disilicate glass.24 The upper trace is a static spectrum obtained without spinning of the sample and the lower is the MAS spectrum obtained at 4.5 kHz rotation. Although

10

H.KAWAZOE

Table 1. Chemical shift tensor values of "Si in each of Q" units present in alkali silicate glasses and distribution of each Q" in the glasses.

Glass composition Li20.4sio2 Li20.2Si02 Li20.SiOz

Q" Q4 Q3 Q2 Q4

Q3 Q2 Q3 Q2

Q' Na20 ' 4sio2 Na20 * 2Si02

-

Na20 Si02

K 2 0 .4sio2

Q4 Q3 Q2 Q4 Q3 Q2 Q3 Q2

Q' Q4

Q3 Q2

K 2 0 . 2Si02

Q4

K 2 0 . Si02

Q3 Q2 Q3 Q2

Q' R b 20 * 2Si02

Q4 Q3 Q2

Per cent contribution 63 24 13 14.5 71 14.5 8.5 83 8.5 54

40 6 7.5 85 7.5 7.5 85 7.5 57 35 8 5.5 89 5.5

2 96 2 3 94 3

miso

*I

u2

a3

-111 - 104 - 102 -107 -93 -85 - 105 -73 -102 -109 -100 -85 - 105 -92 -85 - 105 -78 -90 - 103 -92 -85 - 103 -95 -86 -98 -88 -85 - 108 -98 -89

-111

-111

-64

-64

-111 -184 - 158 - 107 - 153 - 155 -205 -163 -186 -109 -196 - 171 -105 - 176 - 171 - 175 -148 - 170 -103 - 192 - 171 - 103 - 179 - 172 -182 - 174 - 169 -108 -184 -159

(ppm)

-46

- 107 -63 -25 -55 -8 -60 -109 -52 -26 - 105 -50 -26 -70 - 18 -50 -103 -42 -26 - 103 -53 -27 -56 -33 -43 - 108 -55 -54

- 102 - 107

-63 -75 -55 -48 -60 -109 -52 -58 - 105 -50 -58 -70 -68 -50 - 103 -42 -58 - 103 -53 -59 -56 -57 -43 - 108 -55 -54

o,chemical shift values.

line width was relatively broad, we see only a single absorption, which was assigned to Q3, in the MAS spectrum. On the other hand, the static spectrum, whose line shape was highly affected by anisotropies of the chemical shift tensors, cannot be reproduced without assuming the coexistence of Q4, Q3, and Q2. The concentrations and principal values of chemical shift tensors of the respective Q' species in several alkali silicate glasses estimated by computer simulation of the spectra are given in Table l.24 In the case of the potassium disilicate glass the simple MAS spectrum indicates that almost all of Q' species involved in the glass are Q3, whereas analysis for both the MAS and static spectra suggests that 5.5% of Q4,89% of Q3 and 5.5% of Q2 are coexisting in the stoichiometric glass. Use of the

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

11

static measurement is important in quantitative evaluation of the coexisting Q’ species in silicate glasses. It must be added that even the isotropic chemical shift value itself cannot be determined for the broad resonance absorption. 4. STRUCTURE DETERMINATION OF “DEFECTS” IN AMORPHOUS SILICA Neutron irradiation with the dose of 1 x lo2’ neutrons/cm2 to quartz crystal and silica glass induces a dramatic change in the densities of both materials.25 In the case of quartz crystal it decreased from 2.64 to 2.26. Inversely the density of silica glass increased from 2.21 to the same value with the irradiated quartz. The observation suggested a possibility at least from the macroscopic viewpoint that the irradiated materials have a similar structure. Measurements of Raman spectra on the materials before and after the irradiation confirmed that the assumption from the microscopic viewpoint is correct.26 As noted in Fig. 7 quartz crystal changes into amorphous silica upon heavy irradiation. In the case of silica glass the material remained in the amorphous state during the irradiation and the closely identical Raman spectra were obtained for both the irradiated quartz and silica g l a ~ s . ~ The ~ . ~ ’most pronounced change in the spectra of the glass before and after the irradiation is an enhancement of intensities of the peaks at 500cm-’ and 609cm-’. Since the peaks grew on irradiation, these features were ascribed to “defects”,27 and numerous authors have subsequently proposed specific structural models for their origin.2L35Most of the models assume the presence of dangling bonds or wrong bonds, because the continuous random network model frequently assumed as the structure of Si02 failed to explain the appearance of “defect” lines. However, there was no unique model which can satisfy the related data. Assignment of the Raman lines had been an open question. The “defect” lines in the Raman spectra were found to be enhanced without the irradiation under the polycondensation process of silica gels prepared by the sol-gel method.36737Drying of the gel at under 200°C resulted in enhancement of the D1 (500cm-’) line, while heating at 600°C induced the D2 line (609cm-I). The finding suggests that the structures responsible for the “defect” lines formed upon the dehydration and condensation reaction of hydrated silica gels. A planar ring consisting of three Si04 structure units was proposed to be the intermediate structure responsible for D2 line.35 This model has been experimentally illustrated by using solid state NMR.37 Figure 8 shows the changes in Raman spectra of the initially heated silica gel at 600”C, in which a strong D2 line was discernible, upon subjecting the sample to 100% relative humidity (RH) air at 25°C for the times indicated in the figure. It is

12

H.KAWAZOE l

~

~

'

f

'

l

'

i

'

l

~

Q-OUARTZ

L

I 10"'n/cm'

UNIRRAOIATEO a-

1

1200

.

4

1

1000

,

1

, Mx)

1

,

800 400 FREQUENCY Icm")

~

.

~

.

200

noticed that the intensity of the D2 line decreased during the exposure. This indicates that the structure responsible for the D2 line is attacked by moist air even at room temperature. The NMR spectra of the differently processed silica samples are shown in Fig. 9.37MAS and CP MAS spectra were obtained for the samples heated at the temperatures given in the figure. Hydrated MAS spectra were collected for the respective and heated samples after exposing to 100% RH for 24 h at 25°C. The NMR spectra were obtained at an Ho field strength of 8.45 T ("Si frequency of 71.49MHz). 29Si-'H CP MAS spectra were obtained using

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

13

b"""'/

lXxJt200

Irnmo

#x)

800 700 6(30 500 FREQUENCY tun")

400 300

200

!Oo

0

(b)

Fig. 7. Raman spectra of unirradiated and neutron irradiated a-quartz single crystals (a) and glasses (b).

contact times of 7.5 ms. The three distinct peaks at chemical shifts of about -91, -101, and -110ppm in the CP MAS spectra of 50°C and 200°C gels and in the rehydrated spectrum of the 600°C sample correspond to Q2, Q3 and Q4 structural units, respectively. The non-bridging oxygens in Q2 and Q4 species are bonded to hydrogen as shown by a pronounced enhancement of the Q2 and Q3 peaks in the CP-MAS spectra of the corresponding samples. The MAS spectrum of the 600°C sample had no resolved peaks and centred at about - 107 ppm. Statistical deconvolution of the broad band using the chemical shift values of -91 and -101 ppm for Q2 and Q3 units require the presence of a third peak at about -105 ppm. The additional peak must be due to Q4 sites with a small value of Si-0-Si bonding angle, because the chemical shift of the Q3 structure containing the OH group is

14

H.KAWAZOE

1

0

1

1

1

, 500

1000

1560

RAMAN SHIFT (CM-’) Fig. 8. Raman spectra of SiOz gel initially heated to 600°C and subsequently exposed to 100% RH water vapour at 25°C for the times indicated.

not expected to appear in this chemical shift range. This assignment is further supported by the preferential enhancement of the hydrated Q3 and Q2 resonances in the CP-MAS spectrum of the 600°C sample. A similar deconvolution of the hydrated MAS spectrum of the 600°C sample does not require the presence of the -105 ppm peak; the spectrum was reproduced from the weighted sum of the peaks with -91, -101 and -110 ppm. The NMR data thus unambiguously associate the formation of the species responsible for D2 with the presence of Q4 species with reduced value of the bonding angle. These observations uniquely identify the D2 species as a cyclic trisiloxane (three-membered ring as originally proposed by Galeener).35 Three-membered rings are absent in the gels heated at room temperature. They form in intermediate temperature regions, predomi-

APPLJCATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS 1H CPMASS

MASS

15

HYDRATED MASS

-100

-150

-50

-100

-150

-50

.roo

-150

-50

-100

-150

-50

200°C

.so

-100

-150

ppm FROM T M S Fig. 9. 29Si MAS and CP-MAS spectra of silica gels after heat treatments between 50 and 1100°C. Hydrated MAS spectra were collected after exposure of the original samples to 100% RH for 24 h at 25°C. The 'H CP-MAS spectrum of the 1100°C sample is greatly scale expanded in order to reveal the Q2 and Q3 resonances.

nantly at the surface by the condensation of silanol groups via the following react ion :

16

H.KAWAZOE

Changes in the average Si-0-Si bond angle upon the compaction of silica glass by neutron irradiations was recently demonstrated by 29Si MASNMR.38 The resonance peak with chemical shift of -112.0 & 0.5 ppm relative to TMS and of full width at half maximum height of 14.3ppm was obtained for the sample before the irradiation. In comparison, the irradiated silica had a chemical shift of -106.5 f 0.5 ppm and a line width of 17.6 ppm. Thus the change in chemical shifts from -112.0 to -106.5ppm can be unambiguously interpreted in terms of a change in the Si-0-Si bond angle distribution. The macroscopic compaction can be related to the decrease in the average bond angle of Si-0-Si.

5. STRUCTURE DETERMINATION OF PHOSPHATE GLASSES BY USING "P NMR

As in the case of silicates, PO4 building units in polyphosphates are termed as Q', where i ranges from 0 to 3. Q3 stands for the PO4 unit having three nearest neighbouring PO4 groups around the central PO4 unit, which are involved in crystalline ultraphosphates. Similarly, i denotes the number of neighbouring PO4 units; Qo, Q' and Q2 are respectively contained in crystalline ortho-, pyro-, and metaphosphates. Distribution in polyphosphate anions in a particular phosphate glass was estimated by applying chromatographic methods39 in which dissolution of the glass into aqueous solution is unavoidable. A long-standing question has been whether a distribution in polyphosphate anions with different polymerization degree in the glass is changed or not on the dissolution process. NMR has an advantage over the chromatographic methods because it is a non-destructive analysis. 31P MAS-NMR has applied to the structure determination of phosphate minerals and In the case of 31P NMR it was demonstrated that the estimation of anisotropies in the chemical shift tensor in addition to the isotropic part is important in the identification of Q' species in phosphate glasses.44 Glass samples with the composition xCaO.(l- x)P2O5, xSrO(1 - x)P205, andxBaOm(1- x)P205were obtained by melting the appropriate mixtures of starting reagents. 31P CW NMR spectra were measured at a field of 6.4T and a frequency of 109.31MHz. The spectra were accumulated up to 80 times to achieve a signal to noise ratio larger than 150. MAS spectra were recorded on a JEOL GX 270 at a frequency of 109.35MHz and spinning frequencies of 4.7-4.9 kHz. All chemical shifts have been measured relative to 85% aqueous phosphoric acid. Negative shifts correspond to higher field strength. Figure 10 denotes 31P MAS-NMR spectra of strontium phosphate glasses. The three spectra had a shoulder in the higher field side, whose intensity increased with decreasing SrO/P205ratio. Results of a deconvolution of the

APPLICATION OF N M R SPECTROSCOPY TO GLASSES AND CERAMICS

0

10

20

- experiment

30

40

50

60

17

70

- cs [PPml theory

-single-components

Fig. 10. Three typical central lines of 31PMAS spectra for xSrO-$l.- x)P205glasses. The intensity of the Q3 line decreases and the intensity of the Q hne increases with growing x .

observed spectra into two absorptions are given as dotted lines. The growth of the shoulder on going from the glass with high SrO content to that of low concentration corresponds to the expected increase in the content of Q3 species. The other resonance is assigned to Q2 groups and consequently a decrease in the intensity with respect to the shoulder. The chemical shift values and half widths were evaluated and the results

Table 2. Isotropic chemical shifts, anisotropy and asymmetry of the chemical shift tensors of 31Pin each of Q" units present in xCaO. (1 - x ) P205 glasses, and distribution of each Q" in the glasses. X

0.30 0.375 0.40 0.45

Q" content

4 ,

(PP~)

Half line width (ppm)

A6 ( P P 4

11

Q2

Q3

Q2

Q3

Q2

Q3

Q2

Q3

Q2

Q3

0.52 0.61 0.72 0.89

0.48 0.39 0.28 0.11

-36.8 -35.5 -31.3 -27.3

-48.2 -47.8 -43.5 -36.9

11.1 12.1 11.2 10.6

15.1 10.4 13.5 16.2

-260 -247 -234 -209

-193 -180 - 142

-201

0.37 0.39 0.41 0.44

0.23 0.22 0.21 0.20

The error for the Q" content is 20.02, for isotropic cs ) , 6 ( f0.05.

and half line width k0.5 ppm, for the anisotropy (As) 25 ppm, for the asymmetry (7)

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

19

31P- BL - NMR 6.4T

x(SrO)(l - x)(P206)

. -200

-100

-experiment

0

100

200

:300'

- cs [PPml theory

- single-components

Fig. 11. Three ty ical 31P CW NMR spectra for xSrO.(l -x)P205 glasses. The intensity of the Q line decreases and the intensity of the Q2 line increases with growing x .

P

are given in Table 2. Gaussian line shape was assumed in the deconvolution. A distribution in isotropic chemical shift resulted in a remarkably broad absorption, and spinning sidebands could not be observed. This disturbs a quantitative evaluation of content of the coexisting Q' groups. Figure 11 shows the CW-NMR spectra of 31P in the same glasses. Two distinct and anisotropic absorptions were seen in the spectra. The anisotropy and asymmetry of the chemical shift tensor were evaluated by the computer simulation, and given in Table 2. The calculated spectra were denoted as

20

H.KAWAZOE

1.0 -

0.2

0.4

0.3

0.5

X

-theory

B C ~ Q ~ oSr tlCaQ3 OSr

@ Q3

rBaQ2 aBaQ3

Fig. 12. Q" content versus chemical composition for xRO.(l(R = Ca, Sr, Ba).

x)P205

glasses

dotted lines in the figure. The less anisotropic and less asymmetric signal was ascribed to Q3 units. Figure 12 denotes the relative concentration of the respective Q' species in the phosphate glasses. Calculated values from chemical compositions of the glasses are shown in the solid lines. Agreement between the observed and calculated is satisfactory for high RO content, but the discrepancy is pronounced for the lower content of RO, which might be due to high hygroscopicity of the glasses. 6. APPLICATION OF HIGH FIELD IlB NMR TO STRUCTURE DETERMINATION OF BORATE AND BOROSILICATE MINERALS AND GLASSES

As stated in the introduction, "B NMR was the first example successfully applied to structure analysis in inorganic materials. The width of the central transition caused by the second order quadrupole interaction is far greater than the width induced by dipolar broadening and chemical shift anisotropy. This enables one to determine the ratio trigonal B03/tetrahedral B04. Application of "B MAS-NMR to a determination of the ratio in borate glasses and minerals is naturally expected to result in a more reliable value than those evaluated by the CW broad line technique, but this is not so simple4-' because of the overlapping of the two resonances and the broadening due to 'H-l'B dipolar interactions, the hydrogen being included accidentally or from water of hydration, Optimum conditions for obtaining well resolved "B MAS-NMR was found for borate glasses and minerals.48 It was shown that rapid sample

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

21

spinning larger than 6 kHz and high power proton decoupling are desirable for rapid acquisition of IIB NMR spectra of the borates, from which accurate trigonavtetrahedral ratios may be determined. NMR spectra were recorded on a 11.7 T spectrometer, the magnetic field corresponding to IIB NMR frequency of 160.4 MHz. All spectra were recorded using 2 ps pulse excitation. The solution 90" pulse width for boron trifluoride etherate was 9ps. Chemical shifts are reported in ppm with respect to an external standard of BF3.Etz0 and positive values correspond to low field and high frequency. Figure 13 shows static and MAS-NMR spectra of "B in three minerals; kernite (NazB406(OH)z/3HzO), inderite (MgB303(OH)5/(Hz0)4/HzO)and borax (NazB407/10HzO).48TrigonaYtetrahedral ratios of the minerals are 2/2, 1/2 and 2/2, respectively. The spectra in A and B were obtained under a static condition (without sample spinning), and those in A and B were recorded without and with proton decoupling, respectively. It is noted from a comparison of the spectra in A and B that proton decoupling contributes to reduce the width of the absorptions, this being marked in borax in which WB ratio is high, and that even the use of a high field of 11.7T is insufficient for obtaining well resolved spectra. 3.8 kHz sample spinning at the magic angle was applied in measuring the spectra shown in C and D. Again in this case, those in C are proton coupled spectra and in D are the decoupled. Although the spectra were more resolved compared with the static ones, MAS at 3.8 kHz alone was not sufficient especially for inderite and borax with high WB ratio. The width caused by 'H-"B dipole interaction was reduced in D by proton decoupling and well resolved spectra were obtained. Distinction of quadrupolar broadened absorption due to a trigonal boron which appears in large chemical shift regions from the narrow resonance of a tetrahedral boron having a chemical shift value close to 0, is possible. However, the spectra in D are still insufficient because of the presence of spinning sidebands which disturbs the accurate estimation of the trigonavtetrahedral ratios. Much higher sample spinning (6.4 kHz) and proton decoupling and the use of high frequency are the optimum conditions for the reliable estimation of the ratios. By using computer simulation of the spectra in F the theoretical ratios were well reproduced for the three samples. Quadrupole coupling constants, e2qQ/h in MHz, and isotropic chemical shift with respect to BF3.Etz0 were estimated for the trigonal and tetrahedral borons and given in Table 3. The observed quadrupole constants were in the range 2.3-2.6MHz for trigonal borons, and 0-0.5 MHz for tetrahedral ones. These are in close agreement with those previously reported. '1,42 A remark must be added here. The above-mentioned conditions are undoubtedly appropriate for accurate estimation of the ratios, but inadequate for determining asymmetry parameter q for the quadrupole coupling tensor of tetrahedral borons. In this case low field measurements are more reliable.

22

H.KAWAZOE

Kernile WB.2

H/8=5

-50

20

I

0

Borax WB.5

lnderiie

-20

50

0

-50

20 0 -2b ppm from BF3.Et20

50

20

0

-so

0

-20

Fig. 13. 11.7T "B static and MAS-NMR spectra of kernite, inderite and borax showing the effects of specimen rotation rate, and of 'H decoupling, on spectrum appearance. WB refers to the hydrogen to boron ratio. A , Static, coupled spectra; B, static, decoupled spectra; C, 3.8kHz MAS, coupled; D, 3.8kHz MAS, decoupled; E, 6.4 kHz MAS, coupled; F, 6.4 kHz MAS, decoupled. Spectra were recorded using 2 ps pulses, 219 of the 90" pulse for the solution, the recycle time was 1s, and a 10 Hz line broadening due to exponential multiplication was applied.

Table 3. "B nuclear quadrupole coupling constant and chemical shift data for a series of borates and borosilicates.

IIB NMR parameters Trigonal Sample Boracite Colemanite Danburite Datolite Inderite Inyoite Kernite Tourmaline

Formula and source

QCC

Mg3B7OI3C1,Stramfurt, Germany CaB3O4(OH),-H20, Turkey CaBzSizOs, Mexico CaB(SiO4)(OH),W. Patterson, NJ MgB303(0H)5-(H20)4*H~0, Kern co., calif. CaB303(0H)5- 4Hz0, Death Valley, Calif. NaZB406(OH)2-3H20, Kern Co., Calif. (Na, Ca)(Li, Mg, AI)(Al, Fe,

2.6 2.4

8:

Theory

Experiment

Tetrahedral QCC

sb 1.o

No. of No. of No. of No. of B O ~ B O ~ BO; BO$

d 2.4

16.0 17.0 d d 18.1

-0.3 -0.3 -0 -0 -0.2

1.4 -0.7 1.0 1.o

-

2.3

17.4

-0.2

2.4

18.5

e

-

1

6 2 1 1 2

1

6 2 1 1 2

1.5

1

2

1

2

-0.2

0.9

2

2

2

2

12.7f

d

d

1

-

1

1

2.45 d 2.5 2.5 2.4 2.5 2.3

18.2 d 17.9 18.9 19.0 12.6 16.0

1.2 -3.3 1.7 1.4 2.0 1.8

2

3 1 2

2

3 1 2 1 2 2 1

d d

d d

-0.3 0 -0.2 -0.3 -0.3 -0.5 -0.5 -0

d

1 1

1 1

0

TI

z

Lo

Mn)6(B03)3(Si6018)(OH)41

Ulexite Boron phosphate Lithium borate Potassium pentaborate Borax Pyrex (1) (2) Reedmergnerite Alkali feldspar

Newry, Mass. NaCaB506(OH),-5Hzo, Unknown BP04, Alfa Products Li2B4O7,Noah Chemical KZB10016-8Hz0, Alfa Products NazB407-10HzO,MCB Reagents NaBSi30s,Duchesne Co., Utah NaA1Si308 (B)

-0

0.2

-

2 4 1 2 2 Unknown Unknown

2 4 2 6 1

-1.9 -1.1, -2.5 ~

"Quadrupolecoupling constant, in MHz. Error is fO.1 MHz. "In ppm from BF,-Et,O. Error is f0.2ppm. 'Error is ? 10%. dNo boron of this coordination present. 'Not determined. fCentre of peak.

-

~

~~~

h)

w

24

H.KAWAZOE

7. APPLICATION OF N M R IMAGING TO DETECT FLAWS IN

COMPOSITE CERAMICS IN GREEN STATE

Composite ceramic materials, such as S i c fibre reinforced alumina, are believed to be the most promising ceramics of high fracture toughness and high reliability. There are several factors such as agglomerates of the fibres, metallic inclusions, and pores formed during the formation of the green compact, all of which can either cause fracture of the composite or contribute to the formation of even larger flaws.49-50A considerable improvement in the reliability can be achieved by a complete examination in the green state, where most of the conditions that contribute to the fracture rigi in ate.^' In this section the application of NMR imaging as a nondestructive technique to detect physical flaws in green state fibre reinforced ceramic matrix composites is disc~ssed.~' NMR imaging experiments were first reported in 197353 and several different approaches to data acquisition and processing developed The two-dimensional spin warp method57 was used. This technique uses a standard 90" pulse-delay-180" pulse-delay echo sequence in which the 90" pulse, and occasionally the 180" pulse, are selective. A field gradient is applied along the direction of the magnetic field (2) during the selective pulse (s) in order to control the width of the cross-section of the sample in which the NMR signal is excited. A second gradient, applied orthogonally to the t gradient immediately following the 90" pulse and during the acquisition of the echo signal, provides frequency encoding, while a third gradient, orthogonal to the first two, is incremented stepwise for phase encoding. These latter two gradients yield spatial information in the plane of the cross-sectional slice. The samples were fabricated by slipcasting slurries of alumina containing 20 vol% S i c fibres (diameter 8 p m and length 1mm). Total solid content was 60 wt% of the slip. Slipcast samples of cylindrical shape, 8 mm diameter and 20 mm high, were sealed in 10 mm outside diameter NMR tubes as soon as their surface was dry enough to handle. In the imaging experiment 'H is chosen as an NMR nucleus due to its high sensitivity, and a high field instrument of 9.2 T (lH frequency is 400 MHz) was used. Representative images obtained from the sample are shown in Fig. 14. The images were obtained using a 10 mm Helmholtz coil. A soft 90" pulse of 1.8ms duration and a hard 180" pulse of 54ps duration were employed in the imaging sequence. TE, defined as the time from the centre of the soft 90" pulse to mid-echo, was 8ms, and TR, the delay between repetitions, was set to 500ms. Sixteen scans were acquired at each of 256 phase encoding steps, for a total acquisition time of approximately 34 min per image. Resolution of 70 X 70 pmZ was achieved with X,Y gradients of 9.6 X 10d4T/cm and acquisition of a 256 x 256 data point matrix. Slice thickness in each image is approximately 0.8 mm.

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

25

Fig. 14. Proton NMR images of three 0.8mm-thick slices of the sample, spaced 4 mrn apart. The pixel resorution in each case is 70 x 70 pm'. The diameters of the flaws marked with arrows are (a) 300 pm, (b) 350 p m (flaw near edge) and 375 p m (flaw near centre), (c) 425 pm. Diameter of the tube is 9 mm.

The images reflect the relative amount of 'H spins across a transverse cross-section through the sample. Bright areas indicate high water content and dark areas represent pockets of air between particles or large physical flaws such as inclusions and agglomerates of fibres. Fibre reinforced composites are very porous in the green state, due to the loose packing of the ceramic grains around the fibres. The fine distribution of small spots across the images indicated the presence of water in these open porosities. The dark spots or areas indicate flaws in the composites, caused either by foreign particles or open pores and fibre bundles formed during the casting process. Dimensions of the larger flaws detected in the images are noted in the figure caption.

26

H.KAWAZOE

REFERENCES A. H. Silver and P. J. Bray, J . Chem. Phys., 1958, 29, 984. E. M. Purcell, H. C. Torrey and R. V. Pound, Phys. Rev., 1946, 69, 37. F. Bloch, W. W. Hansen and M. E. Packard, Phys. Rev., 1946, 69, 127. E. R. Andrew, Progr. Nucl. Mag. Res. Spectrosc., 1971, 8 , 1. E. Lippmaa, M. Magi, A. Samoson, G. Engelhardt and A.-R. Grimmer, J. A m . Chem. SOC.,1980, 102, 4889. 6. D. Miiller, W. Gessner, H.-J. Behrens and G. Scheler, Chern. Phys. Lett., 1981, 79, 59. 7. A.-R. Grimmer and U. Haubenreisser, Chem. Phys. Lett., 1983,99,487. 8. R. Dupree, D. Holland and D. S. Williams, Phys. Chem. Glasses, 1985, 26, 50. 9. U. Haubenreisser, Glastechn. Ber., 1986, 59, 174. 10. C. A. Fyfe, G. C. Gobbi, J. S. Hartman, J. Klinowski and J. M. Thomas, 1. Phys. Chem., 1982, 86, 1247. 11. P. E. Stallworth and P. J. Bray, in Nuclear Magnetic Resonance in Glass In Glass: Science and Technology, Vol. 4B (eds D. R. Uhlman and N. J. Kreidl), pp. 77. Academic Press, London, 1990, pp. 77-149. 12. J. F. Stebbins, MRS Bull., 1992,45. 13. M. Magi, E. Lippmaa, A. Samoson, G. Engelhardt and A.-R. Grimmer, J . Chem. Phys., 1984, 88, 1518. 14. J. V. Smith and C. S. Blackwell, Nature, 1983, 303, 223. 15. R. Dupree and R. F. Pettifer, Nature, 1984, 308,523. 16. S. Ramdas and J. Klinowski, Nature, 1984, 308,521. 17. B. L. Scherriff and H. D. Grundy, Nature, 1988,332,819. 18. D. Altermatt and I. D. Brown, Acra Crystullogr., 1985, B41, 240; I. D. Brown and D. Altermatt, ibid., 1985, B41, 244. 19. R. Dupree, D. Holland, P. W.McMillan and R. F. Pettifer, J. Non-Cryst. Solids, 1984,68, 399. 20. A.-R. Grimmer, M. Magi, M. H b e r t , H. Stade, A. Samoson, W. Wieker and E. Lippmaa, Phys. Chem. Glasses, 1984, 25, 105. 21. R. Briickner, H.-U. Chun, H. Goretzki and M. Sammet, J. Non-Cryst. Solids, 1980, 42, 49. 22. H. Kawazoe, M. Takagi, T. Kanazawa and I. Yasui, Glasrechn. Ber., 1983, 56K, 1035. 23. P. C. Taylor, J. F. Baugher and H. M. Kriz, Chem. Rev., 1975,75,203. 24. J . F. Emerson, P. E. Stallworth and P. J . Bray, J . Non-Cryst. Solids, 1989, 113, 253. 25. E. LeII, N. J. Kreidl and J. R. Hensler, in Progress in Ceramic Science, Vol. 4 (ed. J. E. Burke), p. 6. Pergamon, Oxford, 1966. 26. J. B. Bates, R. W. Hendricks and L. B. Shaffer, J . Chern. Phys., 1974, 61, 4163. 27. R. H. Stolen, J. T. Krause and C. R. Kurkjian, Disc. Faruday Soc., 1970, 50, 103. 28. F. L. Galeener, J. C. Mikkelsen, Jr and N. M. Johnson, in The Physics o f S i O z and i&s Interfaces (ed. S . T. Pantelides), p. 284. Pergamon, New York, 1978. 29. R. B. Laughlin, J. D. Joannopoulous, C. A. Murray, K. J. Hartnett and T. J. Greytak, Phys. Rev. Lett., 1978, 40,461. 30. C. A. Murray and T. J. Greytak, J . Chem. Phys., 1979, 71, 3355. 31. G. N. Greaves, J . Non-Cryst. Solids, 1979, 32, 295. 32. G. Lucovsky, Phil. Mag., 1979, 39, 513. 33. E. J. Friebele, D. L. Griscom, M. Stapelbroeck and R. A. Weeks, Phys. Rev. Lett., 1979, 42, 1346. 34. A. R. Silin and P. J. Bray, Bull. A m . Phys. SOC., 1981, 26, 218. 35. F. L. Galeener, 1. Non-Cryst. Solids, 1982,49, 53. 36. V. Gottardi, M. Guglielmi, A. Bertoluzza, C. Fagnano and M. A. Morelli, J. Non-Crysr. Solids, 1984, 63, 71. 1. 2. 3. 4. 5.

APPLICATION OF NMR SPECTROSCOPY TO GLASSES AND CERAMICS

27

37. C. J. Brinker, R. J. Kirkpatrick, D. R. Tallant, B. C. Bunker and B. Montez, J. Non-Cryst. Solids, 1988, 99,418. 38. A. C. Wright, B. Bachra, T. M. Brunier, R. N. Sinclair, L. F. Gladden and R. L. Portsmouth, J. Non-Cryst. Solids, 1992, 150, 69. 39. J . R. van Wazer, Phosphorus and its Compounds, Vol. 1, Interscience, New York, 1958. 40. U. Haubenreisser, G. Scheler and A.-R. Grimmer, Z. unorg. allg. Chem., 1986, 532, 157. 41. T. M. Duncan and D. C. Douglas, Chem. Phys., 1984, 87,339. 42. G. L. Turner, K. A. Smith, R. J. Kirkpatrick and E. Oldfield, J. Mugn. Res., 1986, 70, 408. 43. I . L. Mudrakovskii, V. P. Shmachkova, N. S. Kotsarenko and V. M. Mastikhin, J. Phys. Chem. Solids, 1986,47,335. 44. P. Losso, B. Schnabel, C. Jager, U. Sternberg, D. Stachel and D. 0. Smith, J. Non-Cryst. Solids,1992, 143, 265. 45. S. Schramm and E. Oldfield, J . Chem. SOC. Chem. Commun., 1982, 980. 46. C. A. Fyfe, G. C. Gobbi, J. S. Hartman, R. E. Lenkinski, J. H. O’Brien, E. R. Beange and M. A. R. Smith, J. Magn. Res., 1982,47, 168. 47. S. Ganapathy, S. Schramm and E. Oldfield, J. Chem. Phys., 1982,77,4360. 48. G. L. Turner, K. A. Smith, R. J. Kirkpatrick and E. Oldfield, J. Magn. Res., 1986, 67, 544. 49. P. S. Nicholson, Can. Cerum. Q., 1987, 60,26. 50. D. B. Marshall and J. E. Ritter, A m . Ceram. SOC. Bull., 1987, 66, 309. 51. R. W. McClung and D. R. Johnson, MRS Bull., 1988, 13, 34. 52. S. Karunanithy and S. Mooibroek, J. Mater. Sci., 1989, 24, 3686. 53. P. C. Lauterbur, Nature, 1973, 242, 190. 54. P. A. Bottomley, Rev. Sci. Instrum., 1982, 53, 1319. 55. S. L. Smith, Anal. Chem., l985,57,595A. 56. M. A. Foster and J. M. S. Hutchinson, 1. Biomed. Engng, 1985, 7, 171. 57. W. A. Edelstein, J. M. S. Hutchinson, G. Johnson and T. W. Redpath, Phys. Med. Biol., 1980, 25, 751.

This Page Intentionally Left Blank

High-resolutionSolid-state NMR Studies on Ceramics S. HAYASHI National Institute of Materials and Chemical Research, Tsukuba, Zbaraki 305, Japan and Department of Chemistry, University of Tsukuba, Tsukuba, lbaraki 305, Japan 29 30 30 31 32 32 33 33 34 35 35 36 36 43 49

1. Introduction 2. Experimental aspects of NMR techniques 2.1. High-resolution techniques 2.2. Signal enhancement techniques 2.3. Long spin-lattice relaxation time 2.4. Multinuclear approach 3. Interpretation of NMR parameters 3.1. Chemical shift interaction 3.2. Dipolar interaction 3.3. Quadrupole interaction 3.4. Mechanism of spin-lattice relaxation 4. Conventional ceramics 4.1. Clay minerals 4.2. Cements 4.3. Glasses 5. High-performance ceramics 5.1. Silicon carbide 5.2. Silicon nitride 5.3. Silica and silicates 5.4. Alumina, aluminates and aluminium nitride 5.5. Sialon and its analogues 5.6. Newly accessible nuclei 6. Bioceramics 7. Concluding remarks References

66

67 74 74 77 79 83 83 83 84

1. INTRODUCTION

It is difficult to define “ceramics” strictly, since new ceramics are developed day by day. Conventional ceramics include pottery, porcelain, cement, glass, and so on. Recently, an enormous number of “new ceramics” have ANNUAL REPORTS ON Nh4R SPECTROSCOPY VOLUME 28 ISBN 0-12-505328-2

Copyrighf @ 1994 Academic Press Limited A / / rights of reproduction in any form resewed

30

S.HAYASHI

been developed, which include high-performance ceramics such as silicon carbide, silicon nitride, sialon, etc. Nowadays, ceramics can include all the inorganic materials except for metals. Consequently, “ceramics” include too many materials to be covered here. Limiting the range of the materials, we describe the conventional ceramics, typical high-performance ceramics, and bioceramics. The following materials are not reviewed, irrespective of their importance; zeolites, zeolite-like aluminophosphates, hydrogenated amorphous silicon, and high-T, superconductors. They should be reviewed in separate chapters. Since the pioneering works of Pines et al.’ and Schaefer and Stejskal? high-resolution solid-state 13C NMR has been used widely in the field of organic polymer^.^ The high-resolution techniques have been applied also to silicates and aluminosilicates since the works of Lippmaa et and Klinowski et uL6 NMR is a powerful tool to investigate local structures in those solid materials. Although 29Si is the most popular nucleus in NMR studies on ceramics, other nuclei are important as well; for example, ‘H, llB, 13C, 23Na, 27Al, 31P, and so on. Multinuclear approach is important in NMR studies on inorganic solids including ceramics. Several review papers and books have been published for NMR on inorganic solids.”” In the present chapter, high-resolution solid-state NMR studies on ceramics are reviewed. Although a large number of works using the conventional NMR techniques have been published up to now, they are out of scope in this chapter. Since chemical shift is the most important parameter in high-resolution solid-state NMR, almost all the works reviewed here are concerned with local structures. 2. EXPERIMENTAL ASPECTS OF NMR TECHNIQUES 2.1. High-resolution techniques

Chemical shift parameters are the most important to identify structural units. To extract accurate values of isotropic chemical shift, several techniques are used to narrow the resonance line. In this section, highresolution techniques are summarized briefly. The techniques used depend on the properties of the observed spins. From the viewpoint of the high-resolution techniques, the spins can be classified into three groups; rare spins with a spin quantum number (S) of 1/2 (for example, I3C and 29Si), abundant spins with S = 1/2 (lH and 19F), and quadrupolar spins with S 2 1 (“B, 23Na, and 27Al). The words “rare” and “abundant” mean low and high natural abundances, respectively. In practical experiments, however, the abundant spins can become “rare” when the spin concentration is low in the samples studied. For example, although 31P has a natural abundance of loo%, techniques for the rare

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

31

S = 1/2 spins can be applied to the 31P measurements of phosphoruscontaining materials such as inorganic phosphates. The line width in the spectra of the rare S = 112 spins is dominated by chemical shift anisotropy and dipole-dipole interaction with other abundant spins such as 'H and 19F. Homonuclear dipole-dipole interaction is negligibly small. Magic angle spinning (MAS) of the sample is commonly used to reduce the line width due to dipole-dipole interaction and chemical shift anisotropy. l2?l3 The spinning rate is ordinarily lower than 10 kHz, which is insufficient to suppress the dipolar interaction completely. For the case of 'H, 'H dipolar decoupling (DD) is carried out only during the signal acquisition. Other spins such as 19F have rarely been decoupled, since the second channel in commercial spectrometers is usually tuned to the 'H frequency. When the observed nucleus belongs to the abundant S = 1/2 spins, homonuclear dipolar interaction is dominant, which is suppressed by multiple pulse sequences such as WHH-4,14 MREV-8,'5,'6 and BR-24. Combining with MAS, both the homonuclear dipolar interaction and chemical shift anisotropy are reduced, which method is called "combined rotation and multiple pulse spectroscopy'' (abbreviated as CRAMPS). Heteronuclear dipolar interaction can also be suppressed by MAS to some extent, even if it exists. For quadrupolar nuclei, quadrupole interaction should be considered additionally. The first-order quadrupole interaction spreads out the signals of satellite transitions to an extent of megahertz order. This interaction can be averaged out by MAS, and an enormous number of spinning sidebands are observed frequently. The central transition (rn = 1/2 -1/2) for half-integer spins is not affected by the first-order interaction but by the second-order interaction. The second-order quadrupole interaction is partially reduced by MAS, and the line width is reduced only by a factor of 3 or 4.1'),20The second-order interaction can be averaged out by either dynamic angle spinning (DAS)21or double rotation (DOR) techniques.22

''

f,

2.2. Signal enhancement techniques Cross-polarization (CP) from 'H to rare spins is used to enhance the signal intensity and to improve the accumulation e f f i ~ i e n c ywhen , ~ ~ the observed nucleus is in the neighbourhood of '€3. Another use of the CP is selective observation of those spins dipolarly interacting with 'H spins. The matching condition for CP from 'H is described as y I H l l = ysHls, where yI and ys are nuclear gyromagnetic ratios of 'H and the observed spin, respectively, and and Hls are magnetic field strengths of the RF pulses. For quadrupolar nuclei with a large coupling constant, the condition is yIYrHlr = (S + 1/2) ysHls. Recently, Vega has studied the spin dynamics of cross-polarization from Z = 1/2 to quadrupolar S = 3/2 nuclei.24

32

S.HAYASHI

Another enhancement method is enrichment of NMR active nuclei. I7O has a natural abundance of only 0.037%, and Oldfield's group has prepared 1 7 0 enriched samples to trace 170 spectra.25

2.3. Long spin-lattice relaxation time Ceramics have often very long spin-lattice relaxation times ( T I ) . For example, the 29Si relaxation time in kaolinite with high crystallinity is ca. 2000 seconds.26The 13C Tl of synthetic diamond powder is 3600 s, while the spectra of a natural diamond have been traced 3.5 days after the sample was set in the magnet.27 The waiting time between RF pulses should be long enough to equilibrate the spin state. If 'H spins are in the neighbourhood of the observed nucleus, crosspolarization from 'H can be used, as is the case of kaolinite.28 For synthetic samples, paramagnetic impurities can be added artificially to reduce the TI values. Grimmer et u L . ~have ~ added 0.1 mol% MnO to Na20-Si02 glasses. Maekawa et uL3' have added 0.05 0.1 wt% Fez03 to Na20-Al2O3-SiO2 glasses.

-

2.4. Multinuclear approach A multinuclear approach is necessary, since ceramics consist of a variety of elements. The chemical shift value is expressed with respect to the signal position of a standard compound for each nucleus. Experimentally, second reference compounds are used for instrumental set-up. Solid compounds with the following properties are d e ~ i r a b l e : ~ ~

(1) The line width is very sharp under MAS. (2) The compound contains the observed nuclei with high concentration. (3) The TI value is short. (4) The compound is available easily. (5) The compound is stable in air for a long time. (6) For quadrupolar nuclei, the crystal symmetry is very high, resulting in no quadrupolar broadening. Table 1 summarizes reference compounds and their chemical shifts for nuclei frequently included in ceramics. For alkali and halogen nuclei, alkali halides are good reference corn pound^.^^ Reference compounds containing hydrogens are adequate to set up cross-polarization experiments. For quadrupolar nuclei, the pulse width should be shorter than 1/2(2S+ 1) of the T pulse width calibrated for solution, if one wishes to quantitate the spectra.36

33

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

Table 1. Reference compounds. ~

Compounds

Shifdppm

Tetramethylsilane Silicone rubber (&H5)20*BF3 BPO, NaBH4 Tetramethylsilane Adamantane Glycine Hexamethylbenzene Silicone rubber Nitromethane "NH4CI 1M NaCl aqueous soh. NaCl 1 M A1 (N03)3aqueous soh. AlK(S04)z * 12H.20 AIN&(SO4)2 12H20 Sodium 3-(trimethylsilyl)-propionate-d4 Sodium 3-(trimethylsilyl)-propionate Sodium 3-(trimethylsilyl)-propane-l-sulphonate Silicone rubber 85% &Po4 (NH4)2HP04 NH4HzP04

0 0.119 0 -3.60 -42.06 0 38.520,29.472 176.46,43.67 132.07, 17.17 1.412 0 -341.168 0 7.21 0 -0.21"

Nucleus

'H I'B 13c

I5N

23Na 27A1

*'si 3'P

-

-0.54"

1.445 1.459 1.534 -22.333 0 1.33 1.00

~

Ref. 32 31 31 32 32 32 32 33 31 34 34 32 32 32 32 31 31

~

"Measured at 104.26 MHz. No correction was made for the second-order quadrupole shift.

3. INTERPRETATION OF NMR SPECTRA 3.1. Chemical shift interaction

In high-resolution solid-state NMR, the isotropic chemical shift value is the most important parameter to identify the structural units. For organic polymers, the I3C spectral assignments can be made based on solution data, since an enormous amount of spectral data have been accumulated up to now. However, for other nuclei lacking in data accumulation, various attempts have been made to correlate the chemical shift with crystal structure, especially for 29Si, 31P, and "Al nuclei. In silicates, 29Si chemical shifts correlate with the degree of condensation of the Si04 tetrahedra.4,37 Figure 1 shows 29Si chemical shift ranges of Q" units. Q" means Si04 tetrahedron linked with n Si04 tetrahedra. The range moves toward lower frequency with increase in n . Incorporation of aluminium into the network shifts the 29Si signal of the neighbouring silicon towards higher frequency.- Various attempts have been made to establish

34

S. HAYASHI

t

-60

I

-70

-80

-90

-100

-110

-120 ppm(TMS1

Fig. 1. "Si chemical shift ranges of silicates with a different degree of condensation of Si04 tetrahedra. Q" units have n bridging oxygen^.^' (Reprinted with permission from J. Phys. Chem., 1984, 88, 1518. 0 1984 American Chemical Society.)

empirical correlations between the chemical shifts and structural parameters, especially Si-0-T angles (T = Si, Al),3w3 Si-0 bond length^,^^,^^ Si-T distances,39cation-xygen bond strengths,43 and group ele~tronegativity.~~ Several works have been reported on 31P chemical shifts and their anisotropies in inorganic phosphates.4s56 PO4 tetrahedra are interconnected by sharing an oxygen atom. Lower frequency shift in the isotropic value is observed as the condensation proceeds, similarly to the Si04 units. In contrast to 29Si, 31P has a considerable magnitude of chemical shift anisotropy. The magnitudes of the anisotropy have a trend Qo < Q' < Q2. The coordination of aluminium in solids can be distinguished by 27Al chemical shift.57 Aluminium coordinated by four oxygens (A104) shows a 27Al signal at ca. 60 ppm with respect to A1(H20)z+, while A106 units at ca. Oppm. A105 units are much rarer than A104 and A106. Alemany et uf.58,59 have traced good 27Al spectra of aluminophosphate AlP04-21 using high-field and very fast MAS, in which two inequivalent A105 units have chemical shifts of 14 and 16ppm and quadrupole coupling constants of 5.1 and 7.4 MHz, respectively. Consequently, the isotropic chemical shift moves toward lower frequency with increase in the coordination number. 3.2. Dipolar interaction

MAS can average out the dipole-dipole interaction in principle, if the spinning rate is fast enough. Although the dipolar interaction with 'H is too

HIGH-RESOLUTION SOLID-STATE NMR STUDIES O N CERAMICS

35

large to be suppressed by only MAS practically, it is erased by the 'H dipolar decoupling. However, if the observed nucleus interacts with quadrupolar nuclei through the dipole-dipole interaction, MAS cannot average out the dipolar interaction completely. This phenomenon has been analysed theoretically, especially for nitrogen-containing organic compounds.60-62The same phenomenon should be encountered in ceramics containing nitrogen and aluminium. For example, 29Si spectra of silicon nitride have fine structure caused by 29Si-14N dipole-dipole i n t e r a ~ t i o n . ~ ~ 29Si-27A1 dipole-dipole interaction contributes the line broadening in 29Si spectra of aluminosilicates.28The residual broadening due to the dipolar interaction is inversely proportional to the magnetic field (if expressed in hertz). Therefore, high-field experiments are preferable to suppress this effect. 3.3. Quadrupole interaction

For 1 3 1 spins, quadrupole interaction broadens the resonance line much more than the other interactions, when the observed nucleus locates in a site with low symmetry, Signals of the satellite transitions are spread out, and only the central transition (rn = 1/2 c, -1/2) is observed for half-integer spins. The central transition is broadened by the second-order quadrupole interaction, which is reduced only by a factor of 3-4 under the MAS conditions. 19*20 The second-order quadrupole interaction gives characteristic line shapes in both the static and the MAS spectra,19920364 and thus quadrupole coupling parameters, quadrupole coupling constant and asymmetry factor, can be estimated from simulation of the line shape. It must be noted that the second-order quadrupole interaction shifts the resonance position toward lower frequency, and the intrinsic chemical shift can be extracted only after the correction for the second-order quadrupole shift ." Sites with different symmetries can be distinguished by the difference in the quadrupole coupling parameters. A good example is "B NMR study of boron coordination in b o r a t e ~ .Three-coordinated ~~ boron, B 0 3 , has a quadrupole coupling constant of ca. 2.5MHz, while the constant of B 0 4 units is less than 1MHz. 3.4. Mechanism of spin-lattice relaxation

Mechanisms of spin-lattice relaxation associated with ceramics are discussed in this section. When the material contains no mobile species in it, relaxations by paramagnetic impurities are dominant .269M9 The relaxation rate depends on the sample even if the crystal structure is the same. This

36

S.HAYASHI

relaxation mechanism has been studied t h e o r e t i ~ a l l y , 2 ~and , ~ ~can ~ ~ be classified into two mechanisms roughly: spin-diffusion limiting case and direct relaxation due to the dipole-dipole interaction with electron spins on paramagnetic impurities. The former is applied to abundant spins such as 'H and 27Al,28and the latter to rare spins such as 29Siand 13C.27,28,69 Although nuclei in the neighbourhood of paramagnetic impurity relax very fast due to the direct relaxation, their signals are too broad and too shifted to be detected. For materials including mobile species like H 2 0 and cations, motion of those species relaxes the surrounding nuclei through fluctuation of the dipole-dipole interaction and of the electric field gradient for quadrupolar O2 molecules in micropores also contribute to the relaxation .73,76 Replacement of O2 by organic molecules causes an increase in the relaxation time.77 4. CONVENTIONAL CERAMICS 4.1. Clay minerals

Pottery and porcelain are made of clays and clay minerals through a sintering process. Clay minerals are classified into layered aluminosilicates. Since there are an enormous number of NMR studies on silicates and aluminosilicates," the present chapter cannot cover the whole literature. Kaolins have been used as the starting materials widely in the world, and they have been the most intensively studied among the clay minerals. Consequently, attention is focused on kaolins. 4.1.1. Structure of kaolins Kaolins, A12Si205(OH)4, are layered aluminosilicates with a dioctahedral 1:1 layer structure consisting of an octahedral aluminium hydroxide sheet and a tetrahedral silica sheet. Kaolinite, dickite, and nacrite belong to the kaolins, and the layer stacking manner is different from each other. Figure 2 shows the structure of kaolinite. The 29Si chemical shift of kaolinite is about -91ppm from pure tetramethylsilane, which is attributed to Si04 tetrahedra linked with three S O 4 tetrahedra and having a non-bridging oxygen, denoted as Q3(OAl). Barron et ~ 1 first. detected ~ ~ two 29Sisignals with different chemical shifts in kaolinite, nacrite, and dickite, suggesting the presence of two silicon environments. Magnetic field dependence experiments have recently confirmed the presence of two crystallographically inequivalent sites experimentally.28 Figure 3 shows 29Si spectra of kaolinite, measured with two different fields. The splitting of the two peaks is unchanged (in ppm) by

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

37

Fig. 2. Projection of the structure of kaolinite from (100) direction." (Reprinted with permission from J. Phys. Chem., 01992 American Chemical Society.)

the magnetic field, indicating that the splitting is caused by the difference in the isotropic chemical shift. The presence of two silicon environments agrees with recent results of the crystal structure obtained by diffraction studi e ~ . The ~ ~origin ' of the splitting has been discussed by Thompsong2 and Thompson and B a r r ~ n . *The ~ latter work attributed the origin to hydrogen bonding at the surface of the silicate sheet. As shown in Fig. 3, the resolution of the "Si spectra becomes worse as the magnetic field decreases. This can be explained by the contribution of the dipole-dipole interaction between 29Si and 27Al spins.28 Disorder and paramagnetic impurities also play an important role in broadening the resonance Effect of disorder is easily masked by that of paramagnetic i r n p u r i t i e ~ . ~ ~ . ~ ~ Aluminium is coordinated by six oxygen atoms, and the 27Al NMR signal is observed at about 0 ppm from 1M A1(N03)3 aqueous solution.28Figure 4 shows 27AlMAS-NMR spectra at two different fields. Lower-frequency shift and broadening of the line width at the lower field clearly demonstrate that the line is broadened by the second-order quadrupole interaction. Shulepov

38

S.HAYASH1

-85

-90

-95

PPm Fig. 3. 29Si CPMAS-NMR spectra of Kanpaku kaolinite, measured at (A) 79.496MHz and (B) 39.683MH~.~~ (Reprinted with permission from J . Phys. Chem., 0 1992 American Chemical Society.)

et aLB6have estimated quadrupole coupling parameters from a low field measurement (6.0 MHz); 2 Q q = 3.12 MHz and 77 = 0.9. Simulation of the spectra at the high fields suggests the presence of two aluminium sites.28The diffraction studies also indicate the presence of two aluminium The CRAMPS technique has been successfully applied to kaolinite ,28 as shown in Fig. 5 . The static line width, 30 kHz, is narrowed to 0.6 kHz by use of the CRAMPS. The chemical shift of hydroxyl groups in kaolinite is 2.8 ppm from pure tetramethylsilane. The line width is considerably narrowed by only MAS, provided that the concentration of hydrogen is relatively OW.^*,*^ 29Si spin-lattice relaxation times in clay minerals may vary widely and be extremely long.26*84*88 Care should be taken when one wishes to discuss the . ~ signal intensity quantitatively. Barron et aLS8 and Watanabe et ~ 1 have suggested that the relaxation is caused by paramagnetic impurities. Hayashi et aL2‘ have presented relaxation curves showing exp(-(t/l;)”*) behaviour,

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

39

I

A

100

O

50

-50

-100

PPm Fig. 4. 27Al DD/MAS-NMR spectra of Kanpaku kaolinite, measured at (A) 104.263MHz and (B) 52.051 MHz. Marks ssb indicate spinning sidebands.28 (Reprinted with permission from J . Phys. Chem., @ 1992 American Chemical Society.)

as shown in Fig. 6, and they analysed it quantitatively. The 29Si spins relax by the dipole-dipole interaction with electron spins on paramagnetic impurities directly, and contribution of spin diffusion is negligible. 27Al and 'H spins in kaolinite also relax by paramagnetic Spin diffusion plays an important role in the relaxations of those spins.26 4.1.2. Thermal transformation of kaolins When kaolins are heated, transformation takes place in a stepwise manner. They are transformed to metakaolin at about 500-7OO0C, and then to mullite and crystobalite at about 1000°C. The first transformation from kaolins to metakaolin is accompanied by dehydroxylation.

-

-

A1203.2Si02 2H20 + A1203 2Si02 + 2 H 2 0

40

%HAYASHI

I

40

8

1

I

20

II

0 k Hz

4

-20

-40

0 PPm

Fig. 5. 'H NMR spectra of Kanpaku kaolinite, measured at 400.136MHz. (A) The ordinary single-pulse sequence is used for a static sample. (B) The CRAMPS spectrum measured with the BR24 pulse sequence in the quadrature detection mode.** (Reprinted with permission from J. Phys. Chem., 0 1992 American Chemical Society.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

41

100 80

60 Y

E 40

I

I

I

10

0

1

I

I

I

20 30 t v 2 (S”* )

I

Fig. 6. 29Si magnetization recovery curves for Kanpaku kaolinite, plotted as a function of square root of time, measured at 79.496MHz and at room temperature with the magic angle spinning at 3.0 kHz (0)and without spinning (A).26(Reprinted with permission from J . Phys. Chem., 0 1992 American Chemical Society.)

The second transformation has no loss of weight, being expressed by

-

3(AI2O3 2Si02) + 3AI2o3 2Si02 + 4sio2 The dehydroxylation process has been monitored by ‘H NMR spectra by Gastuche et and Otero-Arean et aL91 Dehydroxylation occurs layer by layer when the lost fraction of the hydroxyl groups is from 0% to 70%. For dehydroxylation more than 70%, hydrogens are left rather inhomogeneously ; patches of undehydroxylated regions are distributed in regions containing isolated hydroxyl groups. Environments of silicon and aluminium atoms during the thermal transformation have been monitored by high-resolution solid-state NMR of 29Si and *’A1 n ~ c l e i . ~Short-range ~-~~ structures can be discussed, although clear interpretation of the spectra is difficult because of the broad line width and the mixed composition. Rocha and Klinowski9’ have traced the spectra, carefully setting the pulse width and the recycle delay. Figure 7 shows 29Si MAS-NMR spectra measured at room temperature after thermal treatment. The parent sample has a -91.5 ppm peak without any splitting, being ascribed to Q3(OAl). At

42

S.HAYASHI

1OOoaC

85ooC

650%

. -20

-40

0100

-140

- .

-180

Ppm trom TMS Fig. 7. 29Si MAS-NMR spectra of kaolinite after thermal treatment at different temperatures given in the figure. They were measured at 79.5MHz and at room temperature. (Reproduced by permission of Springer-Verlag from ref. 97.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

43

500°C just below the onset of the dehydroxylation, two new signals are observed at -97 and -101ppm. Above 500°C the signal is broadened, whose centre of gravity is between -99 and -101 ppm. The broad line width reflects an amorphous nature with variations of Si-0-T (T = Si and Al) angles and bond lengths. Although Lambert et ~ 1 . ' ~and Rocha and K l i n ~ w s k attempted i~~ to deconvolute the spectra, it was difficult to assign each peak to a certain conformational unit. At 1000°C the signal is at ca. -110ppm, and a new peak at about -87ppm begins to emerge. The -1lOppm peak is assigned to crystobalite, Si02, and the -87ppm peak to mullite. Figure 8 shows 27A1MAS-NMR spectra after the thermal treatment. The untreated sample has only one peak at ca. Oppm, being assigned to six-coordinated Al. At about 500"C, two other signals are observed at ca. 28ppm, assigned to five-coordinated Al, and at 57ppm due to fourcoordinated Al. In the 550-900°C range (metakaolinite range) the three A1 coordinations coexist. The amounts of the four- and five-coordinated A1 increase at the expense of the six-coordinated A1 until ca. 80O0C, and then begin to decrease. The six-coordinated A1 reaches a minimum at 750 800"C, above which it rises again. The five-coordinated A1 disappears at ca. 900°C, and simultaneously the four-coordinated A1 begins to increase again. At ca. 900°C the chemical shifts of the six- and four-coordinated A1 are changed, suggesting creation of new phases, Rocha and K l i n ~ w s k i ~ ~ suggested phases of y-alumina and the precursor of mullite. On the other hand, the presence of y-alumina was excluded by Sanz et ~ 1 . ~ ~ A number of compounds, known as mineralizers, have been used to promote selectively the formation of certain species during the thermal transformation. Rocha et ~ 1 have . studied ~ ~ the effect of lithium nitrate mineralizer on the thermal transformations of kaolinite by means of 'Li, 27A1, and 29Si high-resolution solid-state NMR. The temperatures for the transformation are changed by the mineralizer, but no new phases are found. The dehydroxylation of kaolinite to metakaolinite can be completely reversed.'OOy'O1Metakaolinite is reacted with water at 150 250°C for several days, producing kaolinite, which has been monitored by 29Si and 27Al MAS-NMR.'007'o' The reaction is initiated at the edges of metakaolinite particles and is followed by diffusion of water into the bulk. When metakaolinite is treated with NaOH solution, Na-A type zeolite is formed.lo2 The solution phase might play a role in the formation of the zeolite.

-

-

4.2. Cements

CaO-Si02, Ca0-SiO2-Al2O3, and Ca0-SiOZAl2O3-Fe2O3 systems are solidified by reaction with water, and they are the main components in

44

S. HAYASHI

cements. Ordinary Portland cement typically contains 50% tricalcium silicate (3Ca0 SiO,; denoted as C3S) and 25% dicalcium silicate (2Ca0 .SiO,; p-GS). The remaining 25% consists largely of tricalcium aluminate and calcium aluminoferrite phases.lo3 C3S and p-C2S play an important role in the solidification process.

-

4.2.1. Calcium silicates and cements

The CaO-SiO, system has four compounds; Ca0.Si02 (denoted as CS), 3Ca0-2Si02(C3S,), 2Ca0.SiOz (C&, and 3Ca0-Si02 (C3S). C3S and the p phase of &S are main components in the ordinary Portland cement clinker, as described above. Figure 9 shows 29Si MAS-NMR spectra of p-&S, C3S, and two cements. The crystalline calcium silicates have very narrow signals. Table 2 summarizes 29Si chemical shift values of GS, C3S, and their hydrates. C2S and C3S have only Qo environment. C2S has only one equivalent site, showing one 29Sipeak. C3S has nine crystallographically inequivalent sites, among which eight sites are distinguished from the spectra. lociThe ordinary Portland cement is amorphous, and thus it shows a broad signal. The spectra of the two Portland cements are interpreted as the sum of the spectra of C3S and

p-(2,s. The hydration process has been monitored by high-resolution solid-state 29Si NMR.""'16 As a model reaction in the setting of the ordinary Portland cement, the hydration process of C3S and C2S has also been studied. Figure 10 shows 29Si spectra of the materials corresponding to Fig. 9, but after hydration for one week. In the initial stage of hydration of the cements, there is an induction period of several hours. After the induction period, a resonance appears at -79 ppm, being ascribed to end units in silicate chains (Q'). After longer times, a further resonance is clearly resolved at -84 ppm, which is middle units in polymer chains (0'). As shown in the figure, the hydration rate of p-C;S is slower than that of C3S. The hydration of C3S has been studied in detai1.'07-112 The crosspolarization experiments have demonstrated that hydrated monomeric silicate units (ao-H) are formed during the induction period.'" Figure 11 shows the formation of hydrated materials versus time in conjunction with a calorimetric data. After the end of the induction period, the concentration of Qo-H levels off to ca. 2%, and the Q' species (end units in silicate chains) begin to increase. 29Sispectra have also been used to monitor the hydration process of C3S Fig. 8. 27Al MAS-NMR spectra of kaolinite after thermal treatment at different temperatures given in the figure. They were measured at 104.2MHz and at room temperature. (Reproduced by permission of Springer-Verlag from ref. 97.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

1ooooc

9000c

800°C

550%

5000c

20% 200

100

0

-100 -200

45

46

S. HAYASHI

ORDINARY PORTLAND CEMENT

AALBORG WHITE PORTLAND

-3r-40

-

-

-

-

-

- 00

-110

PPm Fig. 9. 29Si MAS-NMR spectra of P-GS, C$, ordinary Portland cement, and Aalborg white Portland cement, measured at 39.76 MHz. (Reproduced by permission of Chapman and Hall Ltd. from ref. 107.)

in the presence of admixtures such as CaC12, sucrose,11o and The induction period becomes shorter for CaC12, and it becomes longer for sucrose. On the other hand, silica accelerates greatly the polymerization process without affecting the induction period. Hydration of actual cement samples has been studied by 29Si NMR. 101,11?-116 Figure 12 shows 29Sispectra of cement pastes. Three peaks are observed at -71, -79, and -84 ppm, being assigned to Qo, Q1, and Q2,

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

47

Table 2. 29Si chemical shifts in calcium silicates. ~~

Formula

Mineral

Shift/ppm

C3S-H C3S-H Hillebrandite

-70.8 -70.3 -71.5 -71.34 -71.4 -73.4 -73.66 -73.5 -72.5 -68.93 -69.04 -71.67 -72.67 -73.44 -73.66 -73.84 -74.46 -82.6 -84 -86.3

2Ca0. SiO2 2Ca0 SiOz 2Ca0. Si02 2Ca0 SiO2 2Ca0. SiO2 2Ca0. SiO2 2Ca0 Si02 2Ca0 SiOz 2Ca0. SiOz.H 2 0 3Ca0. SiOz

~~~

Type

Ref.

Qo

105 37 105 106 37 105 106 37 4 106

Q0

Qo Qo

Q0

Qo Qo Q0 Q0

Qo Q0 Q0

Q0

Q0

Q0 Q0 Q0

Q' Q'

Q2

4 104 4

respectively. The fraction of the sum of Q 1 and Q2 increases with curing time and temperature. The compressive strength increases with curing time and temperature (21-55°C). At 80°C the strength is lower because of the non-uniform distribution of the hydration products, although the degree of hydration is higher.

4.2.2. Calcium aluminates "Al NMR is useful to study A1 coordination^.^^^"^ Six-coordinated A1 (A106) gives a signal at ca. Oppm, while four-coordinated A1 (A104) is at 55-80 ppm. The ordinary Portland cement contains calcium aluminate phase. Table 3 summarizes 27Al chemical shifts of various calcium aluminates. Tricalcium aluminate, 3Ca0 AI2O3(denoted as C,A), is an important component in Portland cements. Skibsted et a1."' have analysed the spectra of C3A in detail, estimating isotropic chemical shifts and quadrupole coupling parameters for the two crystallographically inequivalent sites. Mueller et al."' have monitored the hydration process of monocalcium aluminate by means of 27Al MAS-NMR, as shown in Fig. 13. The hydrated species having six-coordinated A1 increase at the expense of the unhydrated compounds with four-coordinated Al. The conversion of the A1 coordination is divided into three periods; an induction period, a conversion period,

-

48

S.HAYASH1

A A LB0RG WHITE PORTLAND

V'

. , - - J w w

-33

-40

-5,

-60

-?O

PPm

-80

-90

-100

- 1 10

Fig. 10. 29Si MAS-NMR spectra of p-C2S, C3S, ordinary Portland cement, and

Aalborg white Portland cement, after hydration for 1 week. They were measured at 39.76 MHz. (Reproduced by permission of Chapman and Hall Ltd, from ref. 107.)

and an after-conversion period. Hjorth ec ~ 1 . "have ~ observed the conversion of the A1 coordination in a commercial cement upon hydration. Four-coordinated A1 is converted to six-coordinated Al. ' studied A1 coordination in calcium silicate hydrate Stade et ~ 1 . ' ~have containing a small amount of A1203by means of 27AlMAS-NMR. Specimens with CaO/Si02 = 1 contain almost entirely four-coordinated Al, and those with CaO/Si02 = 1.5 only six-coordinated Al. With increasing CaO/SiOz

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

49

I '

/.

/' /"'

End units in silicate chains

/'

.i /*

/

/

Hydrated monomeric silicate units

Time (mins)

Fig. 11. Graph showing the formation of hydrate material as a percentage of the total silicate material present versus time, together with calorimetric data. (Reproduced by permission of the American Ceramic Society from ref. 110.)

ratio the amount of four-coordinated A1 decreases, while that of sixcoordinated A1 increases. 4.3. Glasses

Conventional inorganic glasses consist of network-forming units and network modifiers. Networks of oxide glasses are ordinarily formed by silicate, phosphate, and borate anions, and network modifiers are cations. Many new glasses have recently been developed, most of which were designed for specific applications such as optical fibre, ionic conductor, photochromism, laser, light filter, and so on. NMR is a suitable method to study the short-range order in glasses. In this section, we discuss the structure of glasses studied by means of NMR. 4.3.1. Silicate glasses

Silicate glass with the simplest composition is vitreous silica. Si04 tetrahedra are interconnected without symmetry. The 29Si MAS-NMR spectrum has a relatively broad peak at ca. -110ppm with a full width at half-maximum (FWHM) of about 13ppm,lz2 which is assigned to Q4. The line shape is asymmetric, reflecting distributions of Si-0-Si bond angles, Si-0 distances,

50

S. HAYASHI

Table 3. "A1 chemical shifts in calcium aluminates.

Compounds CaO * 6A1203 CaO .2A1203

-

4Ca0 * 3A1203 3Ca0 2A1203 12Ca0.7Al2o3

-

3Ca0 A1203 6H20 2Ca0 * AIzO3. 8H20 CaO .A1203* 10HzO

Chemical shift/ppm

Ref.

6 6 4 4

9 f0.3 16 k 0.3 65 0.3 78 f 1 60 f 20 76" 80.5 f0.5 83.3 k 0.5 80.3 f 1 71 85 f 8 79f 1 85f3 79.50 f0.50 78.25 k 0.50 12" 9" 3"

119

4 4

CaO ~ 1 ~ 0 ~ CaO A1203

3Ca0. A1203 3Ca0. A1203

A14 coordination

Ib rIc

4 4 4 4 4 4 4 4 4 6 6 6

*

119

118 119 119 117 119 119 120 118 57 57

"A correction of second-orderquadruple shift is not included. bQuadrupole coupling constant (QCC) is 8.69 k 0.05 MHz, and asymmetry factor ( 7 ) is 0.32 +_ 0.02. 'QCC = 9.30 k 0.05 MHz, and q = 0.54 k 0.02.

and Si-Si distances. Dupree and Pettifer'22 have estimated the distribution of the Si-0-Si bond angles by analysing the line shape. The addition of alkali and alkaline earth metals to silica glass introduces bond breakings between SiO4 tetrahedra. Consequently, Q" units with n less than 4 are formed. 29Si MAS-NMR has been used to study the Q" distribution in the binary g l a s ~ e s . ~ ~ * 'Figure ~ " ' ~ ~14 shows 29Si spectra of Li20-Si02 glasses. The spectra were deconvoluted into three and four Q" components, each of which was assumed to have a Gaussian line shape. Since the spectral resolution is not good, the deconvolution procedures might produce considerable errors. Grimmer et al.29 have attempted to reduce the deconvolution errors in the study of Na20-Si02 glasses by use of a high magnetic field and a high spinning rate. In the M20-Si02 (M = alkali

Fig. 12. 29Si MAS-NMR spectra of cement paste specimens of waterlcement ratio 0.45 at 21,35,45,55, and 80°C (A, B, C, D, and E, respectively) for 3,7, 14, and 31 days (a, b, c, and d, respectively). They were measured at 53.54MHz and at room temperature. (Reproduced by permission of the American Ceramic Society from ref. 116.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

5

PPM

51

52

S. HAYASHI

I

100

I

0

I

-100 ppm

Fig. 13. 27Al MAS-NMR spectra of hydrated monocalcium aluminate samples obtained after selected reaction times at 70°C curing temperature. They were measured at 70.4MHz and at room temperature. (Reproduced by permission of

Academic Press, Inc. from ref. 118.)

metal) glasses, the observed Q" contents are not explained by the random statistical distribution of the five Q" units. The distribution is rather constrained, but with a small randomness. In the case of alkaline-earth silicate glasses, the overlap between the Q" peaks is too severe to deconvolute the spectra.13' Other binary glasses have also been studied by 29Si NMR. Lippmaa et ~ 1 . ' ~have ' studied PbO-Si02 glasses. Only one broad signal was observed at -85ppm for a glass with a composition Pb0.Si02, at -80ppm for 2Pb0.Si02, and at -76 ppm for 4Pb0.Si02. No clear interpretation was presented for the network structure. Coordination of aluminium is an important problem in the structure of

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

26Li20-74Si02

53

n

& calculated

Fig. 14. 29Si MAS-NMR spectra of three lithium silicate glasses, calculated spectra, and Gaussians used to construct calculated spectra. The resonance frequency was 39.7 MHz.Iz3 (Reprinted with permission from J . Am. Chem. Soc., 1984, 106, 4396. 0 1984 American Chemical Society.)

54

S.HAYASHI

1

ZOO

.

I

100

-

I

0

PPY

'

-

I

100

.

I

-200

Fig. 15. *'A1 MAS-NMR spectra. of SiOrAIOz glasses. SQ, super quench; NQ, normal quench. Sample compositions in mol% SiOz (S) and A1203 (A) are given for each spectrum. Spectra were obtained at 8.45T (93.83MHz) at spin speeds near 15 ~ H z . (Reprinted '~~ with permission from J . Phys. Chem., 1991, 95, 4483. 01991 American Chemical Society.)

aluminosilicate glasses. Risbud et ~ 1 . and ' ~ ~Sat0 et ~ 1 . lhave ~ ~ studied the structure of Si02-A1203 glasses by means of 27Al and 29Si MAS-NMR. Aluminium is present in four-, five-, and six-coordinations in rapidly quenched glasses, while samples with slower quench rate show only fourand six-coordinations, as shown in Fig. 15. Different local structures are formed by the quench rate. In contrast, only four-coordinated silicon is detected in "Si NMR spectra.

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

55

In aluminosilicate glasses, ternary glass systems of M20-AI2O3-SiO2 and MO-AI2O3-SiO2, where M is an alkali or alkaline-earth metal, have been studied e x t e n ~ i v e l y . ~Hallas ~ ~ ' ~ et~ 'al. ~ 134 ~ have studied A1 coordination in Na20-A1203-Si02 glasses, and have observed only four-coordinated A1 with moderately high symmetry. Engelhardt et ~ 1 . lhave ~ ~ studied network structure in Ca0-AI2O3-SiO2 glasses with widely varying compositions. 29Si spectra show only one peak, and 27Alspectra show the presence of four- and six-coordinated Al. They have reached several important conclusions: (1) Up to three Q"(mA1) units are coexisting, which differ by only one Si-0-T (T = Si, Al) bond. (2) A1 is preferentially bound to the most polymerized Q"(mA1) units. (3) With CaO = A1203S 0.5Si02, the network is fully polymerized. (4) With CaO 2 A1203S 0.5Si02, A1 is four-coordinated, being completely incorporated in the network. ( 5 ) With CaO0.5Si02 (and 2 C a 0 S S i 0 2 ) , A1 is present as neutral "extra-network'' species and does not act as a network modifier. (6) If CaO is present in great excess over A1203 and S O 2 , calcium aluminate is formed. Oestrike et ~ 1 . have ' ~ ~ deduced several correlations between chemical shifts of 2ySi, 27Al, and 23Na and atomic compositions of the ternary systems. have studied the network structure in Na20-AI2O3-SiO2 Maekawa et glasses with N a a Al. With the addition of A1203, the resolution between the Q2 and Q3 peaks in 29Sispectra becomes worse, as shown in Fig. 16. The network is fully polymerized in their compositions, and all A1 atoms are four-coordinated. The amount of non-bridging oxygen atoms is determined by an excess amount of Na over Al. The sol-gel process has increasingly been used in recent years as a method of preparing g 1 a ~ s e s . lThe ~ ~ structure of sol-gel derived glasses and the sol-gel process have been studied by high-resolution solid-state NMR.13s146 Dupree et ~ 1 . and ' ~ ~Wies et ~ 1 . have ' ~ ~ studied Li20-SiOz sol-gel glasses. 29Si MAS-NMR spectra are shown in Fig. 17. The -1lOppm peak is assigned to Q4, and the peak at -101ppm to Q3-H. The Q3-H peak is enhanced in the CP spectra compared with the Q4peak. The glasses contain ca. 20mol% H 2 0 before the heat treatment. The heat treatment removes H 2 0 and allows the formation of Si-O-Li+ units. Crystallization begins above 575"C, which is revealed by narrow crystalline lines at -74.8 ppm, -92.4ppm, -102ppm, and -114ppm. The structures of AI2O3-SiO2 gel and glass have been studied by 27Aland 29Si NMR.'*'" 27Alspectra are shown in Fig. 18. Three A1 coordinations are observed; 1,29, and 52 ppm for six-, five-, and four-coordinated Al. The amount of the five-coordinated A1 increases with the thermal treatment. The amounts of each species depend on the preparation method of the gel.

56

S.HAYASH1

PPm

+

Fig. 16. 29Si MAS-NMR spectra for glasses with 3Na20-4SiO2 NaA102, measured at 3 9 . 7 6 M H ~ ~ (Reprinted ' with permission from J . Phys. Chem., 1991, 95, 6822. @ 1991 American Chemical Society.)

When the gels were prepared from separate alkoxides, the amount of five-coordinated A1 is much smaller. The six-coordinated A1 decreases with addition of Na20. In contrast, the 29Si signal consists of only one broad line, having limited information. Wies et a/.144have studied SiOrTi02-Zr02 sol-gel glasses by means of 29Si and 'H MAS-NMR. Q4 units increase at the expense of Q3 and Q2 by thermal treatment. Selvaraj et have studied the sol-gel process in the synthesis of cordierite. Six-coordinated A1 is converted to four-coordinated A1 by heat treatment, and at the same time five-coordinated A1 is formed. The structural role of H 2 0 in sodium silicate glasses has been studied by Kuemmerlen et ~ 2 1 Q2 . ~ and ~ ~ Q3 silicon environments increases at the expense of Q4 with H 2 0 addition. H 2 0 depolymerizes the silica network. 'H CRAMPS spectra shown in Fig. 19 show two signals at 0 ppm and 7.5 ppm

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

I

I

-50

. . I

-1 50

PPm

I

I .

. . I . . . . I ,

-50

57

. . I

-1 50

PPm

Fig. 17. 29Si MAS-NMR spectra of Li@-SiOz sol-gel glasses after heat treatment. (a) single-pulse and (b) CP spectra. They were obtained at 71.535 MHz and at room temperature.13’ (Reprinted by permission from “Multinuclear magnetic resonance study of Li20-SiOz sol-gel glasses”, R. Dupree ef al. @ 1990 John Wiley & Sons, Ltd.)

with respect to H20. They concluded that both Si-OH (7.5ppm) and molecular water (0 ppm) are present in hydrous sodium silicate glasses. 4.3.2. Phosphate glasses

Tetrahedral PO4 units are condensed in inorganic phosphate glasses by sharing an oxygen atom. The PO4 condensation can be identified by 31P NMR, although the line width in glasses is broader than the corresponding crystalline compounds.56~’47-151 Isolated PO4 units end units (Q’), and middle units in the chain (Q’) are observed separately in the 31P spectra.

(a’),

58

S.HAYASH1

0

100

PPm

-1 00

Fig. 18. "Al MAS-NMR spectra of A1203--Si02 gels derived from di-s-butoxyaluminoxytriethoxysilaneat various stages of thermal treatment. (a) 40T, 6 h; (b) 150°C, 6 h; (c) 450°C, 2 h; (d) 800°C, 2 h. They were traced at 78.2 MHz and at room temperature. (Reproduced by permission of Chapman and Hall Ltd, from ref. 141.)

Figure 20 shows 31P MAS-NMR spectra of AgI-Ag2O-P2O5 glasses, which are known as superionic conducting glasses. In this system, 31P signals are observed at 23 21, 4 2, and -15 -22 ppm for the Qo, Q', and Q2 units, respectively. Average lengths of the chain can be estimated from the integrated peak areas including spinning sidebands. The NMR results confirmed that the chain length is determined by the ratio of AgZO to PZO5. AgI does not modify the network. Hayashi and Hayamkd6 and Brow et ~ 2 1 . 'have ~ ~ presented 31P spectra of metaphosphate glasses of alkali and alkaline-earth metals. Many spinning sidebands are observed, from which the magnitudes of chemical shift anisotropy have been estimated as well as isotropic chemical shifts. Prabhakar ef al. lS1 have studied several phosphate glasses; LiP03, AgPO3, Zn2P207,PbO-P205, M003-P205, W03-P205, V20rP205, K20-

-

-

-

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

' t i CRAMPS

59

MREV-8

hydrous Na Si40g

30

20

10

0 ppm

Fig. 19. 'H CRAMPS spectra of sample I1 (Na2Si40+l.8%H20) and sample I11 (Na2Si4Or9.1%H20), measured at 300.1 MHz. The chemical shift is with respect to H20.'& (Reprinted with permission from J. Phys. Chem., 1992, 96, 6405. @ 1992 American Chemical Society.)

Mo03-P205, Pb0-Mo03-P205, and NaP03-NaV03. They suggested that the network consists of metaphosphate units. Mueller et ~ 1 . have l ~ studied ~ A1 coordination in CaO-A1203-P205 glasses. Three resolved lines are observed; at -21 ppm (isolated A106 octahedra linked with phosphorus atoms), 4 ppm (A106 partially connected via Al-O-A1 bridging bonds), and 37 ppm (A104 tetrahedra incorporated in the phosphate network). The number of A104 groups increases for decreasing P205/A1203ratios. 4.3.3. Borate glasses Boron atoms are three- and four-coordinated in borate glasses. Bray's group has studied boron coordination in various borate glasses by means of "B continuous-wave (CW) NMR of static ~ a r n p l e s . ' ~ ~B' ~0*3 units have a

60

S.HAYASHI

No.4

I

No. 13

* I

* +

-

A

*

+

~ ~ ~ ' " ' " ~ ' " " ' ' ' " " ' ~ ' ' ' '

100

0

-100

ppm

Fig. 20. 31P MAS-NMR spectra of AgI-Ag20-P205 glasses, measured at 80.76 MHz. Compositions (in mol%) are (No. 4) 54.9%AgI-30.1%Ag2015.O%P2O5 and (No. 13) 60.0%AgI-25.O%Ag20-15.O%P~O~. (Reproduced by permission of Academic Press, Inc. from ref. 149.)

quadrupole coupling constant of ca. 2.5MHz, while B 0 4 units have a constant much less than 1MHz. Consequently, the line width of the B 0 3 units is much broader than that of the B 0 4 units, which is mainly broadened by the second-order quadrupole interaction. Typical spectra for static samples are shown in Figs 21A and 21B. The central peak in Fig. 21A is ascribed to B 0 4 units, whereas the broad line with the outer two peaks is

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

61

-

5kHz

+ J

+

+

+

I

I

I

1

200

100

0

-100

I

1

I

100

50

0

+

+ 1

I

I

-50

L

-200 ppm

-100 ppm

A

d

I

1000

I

0

I

-1000 ppm

Fig. 21. "B NMR spectra of a glass sample whose composition is 0.2Ag14.3Ag200.5BzO3.The spectra were measured (A) at 28.88MHz in a static state, (B) at 128.38MHz in a static state, and (C) at 64.03MHz with MAS (spinning rate: 4.5 kHz). (Reproduced by permission of Elsevier Science Publishers B.V. from ref. 159.)

62

S.HAYASH1

ascribed to B 0 3 units. Since the line width due to the second-order quadrupole interaction is inversely proportional to the magnetic field (in hertz), the broad B 0 3 line is collapsed into the B 0 4 line at high magnetic fields, as shown in Fig. 21B. The use of low magnetic fields is important to divide the two species for static samples. B2O3 consists of B 0 3 units. In alkali borate glasses, both B 0 4 and B 0 3 units coexist. The fraction of the B 0 4 units increases with addition of alkali oxides, until it takes the maximum value of about 0.5 at M20/B203= 0.5 (M = alkali metal). In the range M20/B203>0.5, B03 units with nonbridging oxygens are formed, and the fraction of B 0 4 decreases as the amount of alkali oxide increases, The magic angle spinning can reduce the second-order quadrupolar broadening by a factor of three or four. Figure 21C shows a ''B MAS-NMR spectrum, in which the B 0 4 line at ca. 0 ppm overlaps partially with the B 0 3 line spreading over 20 to -50ppm. Although the B03 units have isotropic chemical shifts of ca. 20ppm, the centre of gravity shifts towards lower frequency due to the second-order quadrupole shift. The use of high magnetic fields can reduce the overlap of the signals between the B 0 4 and B 0 3 units. Figure 22 shows "B MAS-NMR spectra of borate minerals and a borosilicate glass. By the use of the high magnetic field (11.7T), the signal of the B 0 3 units spreads over 20 to 0 ppm, and the overlap between the two species is considerably suppressed. Bunker et aZ.lm have studied MO-B203-A1203 glasses (M = alkaline earth). The glasses contain B03, B04, A104, A105, and A106. The presence of three-coordinated oxygen and non-bridging oxygens is inferred. The structure is complex because of the large numbers of local structures. 4.3.4. Mixed and other oxide glasses Networks of silicate, phosphate, and borate can be mixed, resulting in complexity of the structure. Weeding et ~ 1 . ' ~have ' studied silicon coordination in SO2-P205 glasses, and found that the silicon is four-coordinated in the glasses. The fourcoordinated Si is transformed to six-coordinated Si by devitrification. Yang and Kirkpatrick'62 have studied P205-doped alkaline-earth metasilicate With glasses. Phosphorus is present as monomeric structural units increasing P2O5 content the silicon network becomes more polymerized, as indicated by the lower frequency shift of the 29Si signal. Bunker et ~ 1 .have l ~ studied ~ Na2O-B2O34iO2 glasses by means of 29Si, "B, 1 7 0 , and 23Na NMR. For boron-rich glasses, phase separation into sodium borate and silicate-rich phases takes place. The alkali-rich glasses consist of a borosilicate network containing B 0 4 units and non-bridging oxygens. They have also studied structural changes of the glasses during ' ~ ~ traced 29Si leaching in water at pH1, 9, and 12.1M Martin et ~ 1 . have

(a').

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

20

0

63

-20

ppm from BF3*Ef20 Fig. 22. "B MAS-NMR spectra of (A) kernite, (C) Pyrex glass, and (E) ulexite, together with their sirnilations (B, D, F). The spectra were measured at 160.4MHz. (Reproduced by permission of Academic Press, Inc. from ref. 65.)

MAS-NMR spectra of Li20-B203-Si02 glasses. The spectra consist of only one line, whose position and line width depend on the composition. From the analysis of the 29Si NMR results, they suggested that a proportional sharing of the alkali between the borate and silicate systems begins at the initial stage of the alkali addition. On the other hand, a "B NMR study'% has concluded that at the initial stage of the alkali addition the borate groups consume all the added oxygen to form B 0 4 units. The discrepancy of the above two results has not been solved yet. ' studied PO4 condensation in M2O-B2O3-P2O5 glasses Villa er ~ f . ' ~ have (M = Li, Ag) by means of 31P NMR. In addition to the units observed in the Ag20-P2OS systems, three units are newly observed in the Ag20-B203Pz05 systems; P-[O-P*02-01-B at - 18 to -16 ppm, (B-AgP04-B)-[OP*02-01-B at -2 ppm, and (B-Ag2P04) at +12 ppm, The deconvolution of the 31P spectra of the Li20-B203-P20s glass is difficult because of their broad and featureless nature. Calcium aluminate glass contains none of the traditional glass formers such as Si, P, and B. Almost all of A1 in the glasses with compositions 63Ca0 - 37Al2O3and 48Ca0 - 15CaF2.37A1203 are four-coordinated.'68 It is

64

S.HAYASH1

speculated that fluorine enters as a non-bridging species in the form of an A103F tetrahedron. 4.3.5. Non-oxide glasses

Instead of the oxide network, sulphide, selenide, and telluride can form a glass network. Eckert's group has studied various non-oxide glasses such as P-Se,16s172 P-S,173 LiI-Li2S-SiS2,174 Ag2S-P2SS,175 Li2S-P2S5, Li2S-B2S3, Li2S-P2SrB2S3,176 and CaGeP2177by means of NMR. In the P-Se glasses, various phosphorus conformations have been observed by 31P MAS-NMR,'69 as are partly shown in Fig. 23. The 31P resonances at 63 and -75 -85ppm are ascribed to molecular P4Se3, a broad background signal not narrowed by MAS to a more phosphorus-rich environment, the ca. 135 ppm peak to PSe3,2 units, the ca. 125 ppm peak to Se2/2P-PSe2n units, and the 10ppm peak to Se=PSe3/2units. The peak at 148.5 ppm cannot be assigned. Concentrations of the above units depend on the composition. The microstructure in the glasses has been studied quantitatively also by spin echo double and spin echo experiments. 172 It is concluded that the formation of P-Se bond is preferred to those of P-P and Se-Se bonds. In the P-S system, S=PS3,* units and molecular P4S9and P4S10 are identified by 31PMAS-NMR.173 The local structure in LiI-Li2SSiS2 glasses has been investigated by 29Si, 6Li, and 7Li MAS-NMR.174 In analogy to crystalline silicates, the 29Si resonance in crystalline silicon sulphides is shifted towards higher frequency with decreasing the SiS4 condensation , with an overall chemical shift range of ca. 30ppm. In contrast, the 29Si resonance of the glasses appears invariant over the entire region of glass formation. This can be interpreted by assuming that the sulphide introduced with LiZS is shared by more than two SiS4 tetrahedra, and that the coordination of the sulphide undergoes systematic changes. Lithium iodide does not change the Q" distribution of the SiS4tetrahedra, and the formation of LiI microdomains is not suggested. Ag2S-P2S5 glasses show two 31P resonance lines centred at 97 98ppm (assigned to P2S$-: Q') and at 86 87ppm (Q2),17' as shown in Fig. 24. The fraction of Q' increases with increasing Ag2S content, which is similar to the oxide analogue. The Li2S-P2S5system shows similar b e h a ~ i 0 u r . In l~~ the Li2S-B2S3 system,176the fraction of BS4 increases with addition of Li2S at the initial stage. After it takes the maximum at a certain composition, the BS4 fraction decreases with addition of Li2S. This behaviour is similar to the oxide analogue again. Ternary chalcogenide glasses with compositions (Li2S)0.67(B2S3)1 - ,,(P2S5),, have also been studied.'76 However, no final structure has yet been deduced. The structure of CdGeP2 has been studied by '13Cd MAS and spin echo and 31P-113Cd spin echo double resonance NMR.'77 In contrast to the

-

-

-

HIGH-RESOLUTIONSOLID-STATE NMR STUDIES ON CERAMICS

65

1 " " 1 " " 1 " " l " " 1

300

200

100

-100

0

-200

-300

PPM

Fig. 23. 31P MAS-NMR of P-Se glasses. Top: 121.65MHz spectrum of a glass containing 66.6 atom% P. Bottom: 121.46MHz spectrum of a lass containing 50 atom% P (top trace) and of crystalline P4Se4 (bottom trace).lg9 (Reprinted with permission from J . Phys. Chern., 1989, 93, 7895. 0 1989 American Chemical Society.)

66

S.HAYASH1

4

x = 0.40 1

T

,

200

I

T

I

130

I " " 0

-100

PPM

Fig. 24. 31PMAS-NMR s ectra of glasses (Ag2S),(P2S5)1-,. The spectra were measured at 121.65MHz.'' (Reprinted with permission from J . Am. Chem. Soc., 1992, 114, 5775. @ 1992 American Chemical Society.)

crystalline analogue, glassy CdGePz contains a substantial fraction of P-P bonds, and the number of Cd-P bonds is significantly reduced.

5. HIGH-PERFORMANCE CERAMICS Several new ceramics have been developed in recent years. High-resolution solid-state NMR is a powerful tool to investigate the local environment of the observed nuclei. In this section, we discuss microstructures of highly refractory materials, most of which contain silicon (silicon carbide, silicon nitride, silica, etc.) and aluminium (alumina and related materials).

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

67

Table 4. Population of sites in silicon carbide. Polytype

Stacking sequence

Population of site ~

A 2H 3c 4H 6H 15R

ab abc abcb abcacb abcbacabacbcacb

B

C

D 2

3

2 3

~~

2 2 6

2 2 6

Although many works in this section are concerned with 29Siand "Al NMR, we describe works of newly accessible nuclei as well. 5.1. Silicon carbide

5.1.I . Structure of polymorphs Silicon carbide is a highly refractory material. Roughly speaking, two groups of structure are present; a and p phases. The a phase has many polytypes; 2H, 4H, 6H, and 15R are typical ones. H is hexagonal, and R is rhombohedral. The p phase has the ZnS-type structure, being designated as 3C, where C means a cubic structure. In general, there are two manners for the closest packing; face-centred cubic and hexagonal closest packed structures. The packing sequences for the above two structures can be described as abc and ab, respectively. Many polytypes are present in Sic, which differ in their stacking sequence, as are listed in Table 4. Crystallographically inequivalent sites are also listed in Table 4, together with their populations. The four types of sites are schematically expressed in Fig. 25. The environments of silicon and carbon have been studied by highresolution solid-state 29Si and 13C NMR.l7&la Typical spectra are shown in Fig. 26. The number of peaks corresponds to the number of inequivalent sites. Spectral assignment has been carried out on the basis of peak intensity, comparison with other polytypes, and contribution of the second and higher neighbour atoms. 29Si and 13C chemical shifts are summarized in Tables 5 and 6, respectively. The 13C signal of the 3C polytype has not been detected in some work^.'^^+'^ The spin-lattice relaxation time might be too long for pure samples. The estimated TI values are 35 +- 4 min for 6H("Si), 120 k 20 s for 4H(29Si), and 130k 30 s for 4H('3C).178 The recycle delay should be long enough to analyse the spectra quantitatively. Elemental boron is added as a sintering aid to silicon carbide, and the

68

S. HAYASHI

\

Si2

c3 I

6H:A

4H:B

6H:B -

-

6H:C

4H: C

2H:D -

Fig. 25. Local environments for silicon in SIC polytype in a plane containing the crystallographic c axis. The atoms Sil, C1 (the subscripts referring to nearestneighbour silicon and carbon atoms), and C2 (second-neighbour atoms) define four different silicon environments A, B, C, and D. The more remote Siz and C3 atoms allow further distinctions to be made. (Reproduced by permission of the American Ceramic Society from ref. 178.)

I

~

-10

I

-10

PPm

I

- 30

L

PPm

- 30

+

I

30

I

ppm

I

- 10

I

PPm

*10

I

-30

Fig. 26. 29Si and 13C MAS-NMR s ectra of silicon carbide. (a) 6H polytype, ?%; (b) 6H polytype, I3C; (c) cubic P-SiC, 29Si; (d) mixture of polytype, SI. (Reproduced by permission of the Royal Society of Chemistry from ref. 179.)

F.

$

70

S.HAYASH1

Table 5. 29Si chemical shifts in silicon carbide."

PoIytype

29Si chemical shift from Th4S (ppm) ~

2H 3c

4H 6H

15R

Ref. ~

~~

-20.0 (D) -18.4(A) -18.3(A) - 16.1(A) -18.4(A) -17.2(A) -20 (A) -16 (A) -22.5(B) -19.7(C) -14.7(A) -25.4(B) -20.9(C) -13.9(A) -20.2; -24.5(B;C) -14.3(A) -24.9(B) -20.4(C) -14.4; -20.5; -25(A; B; C) -14.9(A) -20.8; -24.4(B; C) -14.6(A) -24.1(B) -20.5(C) -15.2; -20.2; -24(A; B; C)

"Each signal is assigned to one of the sites A

178 178 180 181 181 182 182 183 178 178 180 181 183 180 181 183

- D.

Table 6. 13C chemical shifts in silicon carbide." ~

~

Polytype 2H 3c

4H 6H 15R

13Cchemical shift from T M S (ppm) 15.O(D) No signal No signal 23.7( A) 18.4(A) 24.7(A)

13.5(B)b 20.9(C)b 23.O(A) 15.2(B)b 20. 3(C)b 15.2(A)b 20.2; 23.2(B; C)b 15.6; 20.6; 23.8 16.0(A)b 20.7; 22.7(B; C)b

"Each signal is assigned to one of the sites A Ventative assignment.

Ref. 178 178 180 182 182 183 178 178 180 183 180

- D.

coordination of boron has been investigated by "B NMR.lWWhen sintered in argon, boron penetrates the grain boundaries and is incorporated into the bulk in a tetrahedral form. In contrast, when sintered in nitrogen atmosphere, boron nitride is formed on the intergranular surface, where boron is in a trigonal form.

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

71

5.1.2. Curing pracess of polycurbosilune Polycarbosilane is a precursor for ceramic materials such as silicon carbide fibres, The curing process of polycarbosilane has been studied by 29Si and I3C NMR.18s189 Figure 27 shows 29Si MAS-NMR spectra of polycarbosilanes. Polycarbosilanes were synthesized by thermal decomposition and condensation of tetramethylsilane and of polydimethylsilane.190 Uncured polycarbosilanes show two intense 29Si resonance lines at -0.5 and -17.5ppm (see Fig. 27a), which are assigned to Sic4 and SiC3-H, respectively. The presence of the Si-H bond for the -17.5ppm peak is confirmed by the dipolar dephasing experiment, as shown in Fig. 27b. The third weak peak is present at -38.5ppm, corresponding to Si*C3Si. Only one 13Cline is observed at 5 ppm, being assigned to CH, units (n = 1, 2, 3). Thermal conversion in argon atmosphere begins at 600°C and finishes at 700°C. 187 At 600°C the Sic3-H peak diminishes, suggesting the breaking of the SCH bond. At 700°C both 29Si and 13C lines become broad, reflecting the amorphous nature of the network. Further heating produces amorphous Sic. When the curing is carried out in oxygen at 145 195"C, three new lines appear in 29Si spectra at 9.5, -17, and -53 ppm.'85,189The 9.5 pprn peak is ascribed to Si*C3-OSi, the -17ppm peak to Si*C3-OCH2-Si, and the -53 ppm peak to Si*&(0CH2)2. Dipolar dephasing experiments can enhance the lines at -17 and -53 ppm compared with the Sic3-H peak. Nitridation of polycarbosilane is carried out by thermal treatment in NH3 gas flow, and the process has been monitored by 29Siand 13CNMR.188 The nitridation begins at 500"C, and is completed at 700°C. After breaking Si-H and Si-C bonds, Si-N bonds are created. Finally amorphous Si3N4 is formed, which gives a 29Siline at -45 ppm. No new peaks have been observed in 29Siand I3C spectra, when polycarbosilanes are cured by electron irradiation.186Si-H and C-H bonds are broken, and Si-C and S i S i bonds are newly created. Si-Ti-C fibre can be prepared by the pyrolysis of polytitanocarbosilane, which process has been monitored by 29Si NMR. 19' Polytitanocarbosilane has a 29Sisignal at 10 ppm in addition to the -0.8 and -18.1 ppm lines. The latter two lines are similar to those in polycarbosilane, and are ascribed to Sic4 and Sic3-H, respectively. The 10ppm peak is assigned to Sic3-0 units. Pyrolysis process up to 1500°C in argon atmosphere has been followed by 29Si NMR measurements at room temperature after heat treatment,'" as shown in Fig. 28. In the first stage up to lOOO"C, Si-C bonds are broken, and SiC4-xOx units are formed. Three new lines are observed at -35 pprn (%GO2), -60 -78 ppm (SiCO3), and -95 -111 ppm -15 (Si04). Above 10o0"C the number of Si-0 bonds decreases. At 1500°C the product is a mixture of crystalline SIC and Tic. Polycarbosilane modified with aluminium alkoxide is based on AI(OH)6

-

-

-

-

72

S. HAYASHI

Sic,

(

b

SiC,H

Dipolar P151-F D ephasing

(

CP/MAS

(C)

CP/MAS

(dl

I

I

I

+50 0 -50 29Si Chemical Shift (relative toTMS) ppm

Fig. 27. 29Si MAS-NMR spectra of polycarbosilane, measured at 11.9 MHz. (a,b) Uncured polycarbosilane fibres, (c) polycarbosilane fibres cured by electron irradiation of 7.5 MGy and (d) of 15 MGy, and (e) polycarbosilane fibres cured by thermal oxidation. (Reproduced by permission of Chapman and Hall Ltd, from ref. 186.)

HIGH-RESOLUTION SOLID-STATE N M R STUDIES ON CERAMICS

73

1500°C

1400°C

1m0c

PTC I

'

'

'

'

' 0

* ' a

1

'

'

'

- 100

1

' ' 1

(PPm 1 Fig. 28. Evolution of 29Si MAS-NMR spectra of polytitanocarbosilane during the pyrolysis process. The spectra were obtained at 59.6 MHz and at room temperature. (Reproduced by permission of Chapman and Hall Ltd, from ref. 191.)

74

S.HAYASHI

particles dispersed in polycarbosilane chains, which has been suggested by 29Si, 27Al,and 13C NMR. 192 The polyaluminocarbosilane is pyrolysed to produce Sic 2H polytype at 1500°C. 5.2. Silicon nitride

There are two crystalline forms of silicon nitride, a and p. In both structures silicon is tetrahedrally coordinated by four nitrogens, and nitrogen is trigonally bound to three silicons. The a phase contains two crystallographically inequivalent Si sites, while the p phase contains only one Si site. Figures 29 and 30 show 29Si MAS-NMR spectra of a-Si3N4 and p-Si3N4, respectively. Roughly speaking, the a phase has two 29Si lines at -46.8 and -48.9ppm, while the p phase has one line at -48.7ppm.’9s196 This corresponds to the number of inequivalent sites. The chemical shift assignment has been based on the idea of “sphere of i n f l ~ e n c e ” . ’The ~ ~ Sicrl and the Sip have eight Si neighbours within a 3.8 A sphere, while the Sia2 has ten Si neighbours. The Sial and Sip are expected to have the same chemical shift, since they both have eight Si neighbours. Consequently, the -48.9ppm peak is ascribed to the Sial, and the -46.8ppm peak to the Sia2. The 29Si spectra shown in Figs 29 and 30 show fine structure, which is dependent on the magnetic field strength. The MAS cannot average out the dipole-dipole interaction when the observed nuclei interact with quadrupolar nuclei. In Si3N4 a 29Si spin interacts with four 14N spins which nuclear spin is 1. Olivieri and Hatfield63 have analysed the spectra theoretically, estimating that isotropic scalar coupling constant Ji,(29Si-’4N) = 15 Hz,and 14N quadrupole coupling constant = -2.1 MHz. The dipolar coupling constant between 29Si and I4N is estimated from the interatomic distance rsrN (1.73 A), which is 335 Hz. Carduner et have recommended the use of high magnetic fields, if crystallographically inequivalent sites are to be observed. Various impurity phases in Si3N4powders have been distinguished by 29Si NMR. 194~195The chemical shifts of those phases are -46.4 ppm (amorphous Si3N4), -57 and -60 ppm (non-stoichiometric phases of silicon oxynitride), -63 ppm (Si2N20), elemental silicon (-81.1 ppm) , and - 110 ppm (Si02). Although it is difficult to trace I5N spectra of samples with natural abundance 15N, spectra with good S/N ratio can be obtained for 15N enriched silicon nitride.7 The chemical shift values are -306 and -317 ppm with respect to CH3N02for p-Si3N4 and Si2N20, respectively. 5.3. Silica and silicates

It is well known that silica is both in amorphous state and in numerous crystalline polymorphs. 197 Table 7 summarizes 29Si chemical shifts in various

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

75

Fig. 29. 29Si MAS-NMR spectra of a-Si3N4 obtained at (A) 39.5, (B) 59.5, and (C) 79.5 M H z . ' ~(Reprinted ~ with permission from J . Am. Chem. Soc., 1990, 112, 4676. 0 1990 American Chemical Society.)

forms of silica. Ordinarily silicon is coordinated by four oxygens in silica, giving 29Si signals in the range from -107 to -120 ppm. On the other hand, silicon is six-coordinated in stishovite, showing a 29Si chemical shift of -191 ppm. The chemical shifts are correlated with Si-Si distances and with Si-0-Si Cordierite is a material with a low expansion coefficient. The ideal composition is Mg2Al3(Si5A1)OI8.Two polymorphs are present; the hightemperature form with hexagonal r tructure and the low-temperature form

76

S. HAYASHI

p - Si3N,

l

~

-44

l

'

-45

l

-46

'

l

-47

'

-40

l

'

-49

l

-50

~

-51

l

~

l

~

-52 ppm

Fig. 30. *'Si MAS-NMR spectra of /3-Si3N4 obtained at (A) 59.5 and (B) 79.5 M H z . ' ~(Reprinted ~ with permission from J . Am. Chem. SOC., 1990, 112, 4676. 0 1990 American Chemical Society.)

with orthorhombic structure. Fyfe et dZo1 have studied the Si,Al ordering process by 29Si NMR. Figure 31 shows 29Si MAS spectra, On heating the glass at 1185"C, crystallization is complete within less than 1min. The amorphous structure reflected in Fig. 31A changes to the hexagonal structure, where the arrangement of Si and Al are disordered, as shown in Fig. 31B. After prolonged annealing, the spectra consist of two peaks at -79.2 and -100.2ppm with intensities of 0.144 and 0.698 (see Fig. 31J), and the crystal structure is orthorhombic. The two main peaks are assigned to Si(4Al) (SO4 tetrahedron.coordinated by four A1 atoms) for the chains and Si(3Al) for the rings. The number of Al-O-A1 bonds decreases during the ordering process.

l

~

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

77

Table 7. 29Sichemical shifts in silica.

Polymorph

29Si chemical shifts/ppm

Ref.

a-Quartz Quartz Quartz Coesite Low crystobalite Cristobalite Tridymite Silicalite Holdstite Stishovite Stishovite

-107.4 -107.1 - 108 -108.1, -113.9 -109.9 -108.5 -109.3 -114.0 -109.9- -117.0 -108.9, -115.0, -119.4 -191.1 -191.3

4 39

-

44 39 4 39 39 198,199 39 39 200

Senegas et aL202 have studied the Si,Al distribution in K-substituted cordierite, KxMg2A14+.Si5-,01s ( O C x S l), by means of 29Si and 27Al NMR. Aluminium excess compensating for the potassium insertion is accommodated in the ring part.

5.4. Alumina, aluminates, and aluminium nitride Alumina has several polymorphs, a, p, y, and so on. A1 coordination in alumina can be identified by 27Alchemical shift, as listed in Table 8. Table 8 also lists chemical shift values of alkali and alkaline-earth aluminates except for calcium aluminates (for calcium aluminates, see Table 3). Magnesium aluminate has a spinel structure. Ideally, Mg and A1 occupy tetrahedral and octahedral sites, respectively. Figure 32 shows "Al MASNMR spectra of synthetic and natural MgA1204 spinels. Natural samples show almost complete A1 ordering, since they are cooled through geometric times. However, the aluminium ordering is incomplete in synthetic samples, and both six- and four-coordinated A1 are observed in the 27Alspectra.204In yttrium aluminium garnet Y3Al5OI2,aluminium is known to occupy both octahedral and tetrahedral sites. Massiot et aL2& have analysed the spectra theoretically, and have obtained accurate populations; three octahedral for two tetrahedral sites. Sodium aluminate NaA102 is formed by dehydration of Na20 A1203 3H20. The dehydration process has been studied by 23Na and 27AlMAS-NMR.208For untreated Na20 -A1203.3H20, two major types of sodium sites are present, which have pseudo-octahedral structure. Aluminium is four-coordinated in the form A1*02(0A1)2. For the dehydrated sample, the sodium site has pseudotrigonal-bipyrimidal structure, and aluminium is four-coordinated, A104.

-

-

78

S.HAYASHI

J

-do

-do

-100

w m irom TYS

-110

-60

-90

-160

-iio

pom irom TMS

Fig. 31. 29Si MAS-NMR spectra of synthetic cordierite, Mg2AI4Si5Ol8.The crystalline cordierites were prepared from glass of the same composition (shown in (A)), by annealing at 1185°C for (B) 2 min, (C) 6.5 min, (D) 20 min, (E) 6 h, (F) 23.5 h, (G) 48.5h, (H) 96h, (I) 408h, and (J) ca. 2000h. The spectra were obtained at 79.5MHz and at room temperature.”l (Reprinted with permission from J . Am. Chem. Soc., 1 9 8 6 , 1 0 8 , 3 2 1 8 . ~1986 American Chemical Society.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

79

Table 8. "Al chemical shifts." Compounds

Al-0 coordination

Chemical shift/ppm

Ref.

6 6 6

5 3 11.5 6479 64,12 68, 0 6692 66, 8 77 76 73 77 22 0 69.7, 9.8 69 74, 0.8 76, 0.8' 9.4, 0.8 113 109

57 117 205 57 117 203 117 205 57 117 57 57 57 117 204 57 205 206 205

4,6 4, 6 4, 6 4,6 4,6 4

4 4 4 6 6

4,6

4 4, 6 4, 6 6 4 4

205

207

A ' correction for the second-orderquadrupole shift is not included, unless otherwise stated. 'A correction for the second-orderquadruple shift is included:

Aluminium nitride has a high heat-resisting property and large thermal conductivity. Ultrafine powder of aluminium nitride is easily reacted with water, producing AI(OH)3, as shown in Fig. 33.207,209AIN shows a 27Al signal at 103ppm, which corresponds to aluminium coordinated by four nitrogens. The hydrated sample has a new "Al signal at -5ppm, which is assigned to A1 coordinated by six oxygens in A1(OH)3. 5.5. Sialon and its analogues

Sialons are phases in Si-Al-0-N and M-Si-A1-0-N systems. They are constructed of mainly (Si, Al)(O,N)4 tetrahedra, and sometimes contain six-coordinated Al. The local coordination of A1 in the Si-AI-0-N system has been Sialon X-phase, of approximate composition Si3Al6OI2N2,contains both A106 and A104 units. A mixture of sialon polytypoids shows A106 and AIN4 units. Local environments of Si have been studied by 29Si NMR for Y-Si-0N,196,212 La-Si-0-N, and La-Si-AI-0-N systems.213The 29Sichemical shift

00 0

octahedral

I

c

A

B i=Q.36fo.03

tetrahedral

i=0.12fo.06

I

100

$0

0

-50

100

50

0

-50

ppm from AI(H20)63+ Fig. 32. "Al MAS-NMR spectra of (A) MgA1204spinel synthesized at 1400°C and (B) natural spinel from Thailand. The spectra were obtained at 104.2 MHz. (Reproduced by permission of the Chemical Society of Japan from ref. 204.)

HIGH-RESOLUTION SOLID-STATE! NMR STUDIES ON CERAMICS

81

3

I

I

200

I

I

I

100

8

I

1

I

0 -100 /PPm

1

I

-200

Fig. 33. 27Al MAS-NMR spectra of (a) ultrafine powder of AlN and (b) its hydrated sample. The spectra were obtained at 52.00 MHz. (Reproduced by permission of the Chemical Society of Japan from ref. 207.)

ranges for the tetrahedral units SiOxN4-x (x = 0 to 4) are summarized in Fig, 34. p'-Sialon is produced from a nanocomposite between dodecylammoniumexchanged montmorillonite and polyacryonitrile through carbothermal reduction in N2 atmosphere. The process has been monitored by 27Aland 29Si MAS-NMR.214A106 units change to AlN4 units via Al(N,0)4. In contrast, S i 0 4 units are converted to SiN4 by 1200°C without passing intermediate Si(N,O), environments. By 1600°C the Sic4 environment becomes dominant.

w Tclrahcdron

L" SiO,

SiOJN

-

SIOZNz

SON,

SiN4 -20

-30

-40

.-SO

-60

-70

-80

-90

-100

-110

-120

Chemical shil1 in ppm from TMS

Fig.

34. 29Si isotropic chemical shift ranges for SiOxN4-x (0 S x S 4) tetrahedra from the (Y, LaMi-Al-0-N systems.213 (Reprinted with permission from J. Am. Chem. Soc., 1989,111, 5125. @ 1989 American Chemical Society.)

HIGH-RESOLUTION SOLID-STATE NMR STUDIES O N CERAMICS

83

5.6. Newly accessible nuclei In this section, we discuss other nuclei newly accessible, such as 170, 25Mg, 89Y, 91Zr, and '39La. 1 7 0 has a low natural abundance (0.037%), resulting in very low sensitivity. Oldfield's group has traced 170NMR spectra for 1 7 0 enriched ~ a m p l e s . ~In~ contrast, . ~ ~ ~ ~Bastow '~ and Stuart have traced the spectra for natural-abundance samples with relatively small quadrupole interactions.218 Mueller et ul. have applied DAS and DOR techniques to 170-enriched silicate 25Mg spectra of simple inorganic solids have been presented by Dupree and Smith.220No enrichment was carried out, since the natural abundance is 10%. MgO has a chemical shift of 26 ppm with respect to 3 M MgS04 solution. "Y spectra have been traced for Y2O3, various yttrium aluminates, various yttrium silicates, Y2Sn207,221 Y2Ti207, and Y2 -,Ln,M2O7 (Ln = Ce, Pr, Nd, Sm, Eu, Yb; M = Sn, Ti).222 91Zr spectra of oxide materials have been published by Hartman et 13'La spectra have been measured for LaA103.213 6. BIOCERAMICS

Bioceramics are classified into bioinert and bioactive materials. Alumina and zirconia belong to the former group. On the other hand, calcium phosphates have a good affinity to a living body, being classified in the latter group. Calcium phosphates have been studied by means of 31P NMR.224-226 Table 9 summarizes 31P chemical shifts of various calcium phosphates. The isotropic chemical shifts tend to move to lower frequency upon protonation of the phosphate. The spectra of non-stoichiometric calcium hydroxyapatite resemble those of stoichiometric hydroxyapatite, indicating that no additional compounds are formed by the Ca deficits. Human dental enamel shows spectra similar to hydroxyapatite. Amorphous calcium phosphate is an important compound, since it is a precursor of hydroxyapatite. 31P NMR r e s ~ l t s support ~ ~ ~ * a~ model ~ ~ in which amorphous calcium phosphate consists of Ca9(P04)6 clusters. Water resides in the interstices between clusters. 7. CONCLUDING REMARKS Novel ceramics are being created day by day, while new techniques are being developed in the field of solid-state NMR. At the same time, the number of nuclei accessible by solid-state NMR is increasing. Correlation between performance of materials and atomic-level chemistry will become

84

S. HAYASHI

Table 9. "P chemical shifts in calcium phosphates.

Compounds

Form

Monocalcium phosphate monohydrate Monocalcium phosphate Dicalcium phosphate dihydrate (brushite) Dicalcium phosphate (monetite) Octacalcium phosphate P-Tricalcium phosphate Hydroxyapatite Fluoroapatite Amorphous calcium phosphate

ShiWppm

Ref.

-0.1, -4.6 -0.6, -5.0 0.5, -0.5 1.7 1.0 0.0, -1.5 -0.7, -1.8 3.4, -0.1 2.6, 1.2, -0.8 9.2, 5.2, 4.2 2.8 2.8 1.7 2.6

224 226 224 224 226 224 226 224 226 226 224 224 225 226

much more important to improve the performance and to create novel materials with high performance. NMR is a powerful method to investigate atomic-level local structures in principle, and is an indispensable method for material researches. Although both materials and NMR techniques are limited in the present chapter, application of solid-state NMR to material researches will be enlarged greatly.

REFERENCES 1. A. Pines, M. G. Gibby and J. S . Waugh, 1. Chem. Phys., 1973,59, 569. 2. J. Schaefer and E. 0. Stejskal, J . Am. Chem. SOC.,1976, 98, 1031. 3. H. Saito and I. Ando, Ann. Rep. NMR Spectrosc., 1989, 21, 209. 4. E. Lippmaa, M.Maegi, A. Samoson, G. Engelhardt and A.-R. Grimmer, 1. Am. Chem. SOC.,1980, 102, 4889. 5. E. Lippmaa, M. Maegi, A. Samoson, M. Tarmak and G. Engelhardt, J . Am. Chem. SOC., 1981, 103, 4992. 6. J . Klinowski, J. M. Thomas, C. A. Fyfe and J. S. Hartman, J. Phys. Chem., 1981, 85, 2590. 7. G. L. Turner, R. J. Kirkpatrick, S. H. Risbud and E. Oldfield, Am. Ceram. SOC.Bull., 1987,66,656. 8. E. Oldfield and R. J. Kirkpatrick, Science, 1985, 227, 1537. 9. G. R. Hatfield and K. R. Garduner, 1. Muter. Sci., 1989, 24, 4209. 10. R. E. Walstedt and W. W. Warren, Jr, Science, 1990, 248, 1082. 11. G. Engelhardt and D. Michel, High-resolution Solid-state NMR of Silicates and Zeolites, John Wiley & Sons, Chichester, 1987. 12. E. R. Andrew, Prog. Nucl. Magn. Reson. Spectrosc., 1972, 8 , 1. 13. I. J. Lowe, Phys. Rev. Left., 1959, 2, 285. 14. J. S. Waugh, L. M. Huber and U. Haeberlen, Phys. Rev, Left., 1968, 20, 180.

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

85

P. Mansfield, J . Phys. C, Solid State Phys., 1971, 4, 1444. W.-K. Rim, D. D. Elleman and R. W. Vaughan, J . Chem. Phys., 1973,59, 3740. D. P. Burum and W.-K. Rhim, J . Chem. Phys., 1979,59, 944. L. M. Ryan, R. E. Taylor, A. J. Paff and B. C. Gerstein, J. Chem. Phys., 1980,72, 508. E. Kundla, A. Samoson and E. Lippmaa, Chem. Phys. Lett., 1981, 83, 229. H.-J. Behrens and B. Schnabel, Physica, 1982, 114B, 185. A. Llor and J. Virlet, Chem. Phys. Letr., 1988, 152, 248. A. Samoson, E. Lippmaa and A. Pines, Mol. Phys., 1988, 65, 1013. S. R. Hartmann and E. L. Hahn, Phys. Rev., 1962, 128, 2042. A. J. Vega, Solid Stare Nucl. Magn. Reson., 1992, 1, 17. S. Schramm, R. J. Kirkpatrick and E. Oldfield, J. A m . Chem. Soc., 1983, 105, 2483. S. Hayashi, T. Ueda, K. Hayamizu and E. Akiba, J. Phys. Chem., 1992, 95, 10928. P. M. Hendrichs, M. L. Cofield, R. H. Young and J. M. Hewitt, J . Magn. Reson., 1984, 58, 85. 28. S. Hayashi, T. Ueda, K. Hayamizu and E. Akiba, J. Phys. Chem., 1992.96, 10922. 29. A.-R. Grimmer, B. Thomas, G. Seifert and P. Saw, Magn. Reson. Chem., 1990,28, S97. 30. H. Maekawa, T. Maekawa, K. Kawamura and T. Yokokawa, J. Phys. Chem., 1991, 95, 6822. 31. S. Hayashi and K. Hayamizu, Bull. Chem. SOC. Jpn, 1989, 62, 2429. 32. S. Hayashi and K. Hayamizu, Bulf. Chem. SOC. Jpn, 1991,64, 685. 33. S. Hayashi and K. Hayamizu, Bull. Chem. SOC.Jpn, 1991, 64, 688. 34. S. Hayashi, unpublished results. 35. S. Hayashi and K. Hayamizu, Bull. Chem. SOC.Jpn, 1990, 63, 913. 36. D. Fenzke, D. Freude, T. Froehlich and J. Haase, Chem. Phys. Lett., 1984, 111, 171. 37. M. Maegi, E. Lippmaa, A. Samoson, G. Engelhardt and A.-R. Grimmer, J . Phys. Chem., 1984,88, 1518. 38. J. M. Thomas, J. Kennedy, S. Ramdas, B. K. Hunter, D. T. Tennakoon, Chem. Phys. Lett., 1983, 102, 158. 39. J. V. Smith and C . S. Blackwell, Nature (Lond.), 1983, 303, 223. 40. S. Ramdas and J. Klinowski, Nature (Lond.), 1984, 308, 521. 41. G. Engelhardt and R. Radeglia, Chem. Phys. Letr., 1984, 108, 271. 42. R. Radeglia and G. Engelhardt, Chem. Phys. Len., 1985, 114, 28. 43. R. J. Kirkpatrick, K. A. Smith, S. Schramm, G. Turner and W.-H. Yang, Ann. Rev. Earrh Planet. Sci., 1985, 13, 29. 44. K. A. Smith, R. J. Kirkpatrick, E. Oldfield and D. M. Henderson, A m , Mineral., 1983, 68, 1206. 45. A.-R. Grimmer and R. Radeglia, Chem. Phys. Lett., 1984, 106, 263. 46. A.-R. Grimmer, Chem. Phys. Lett., 1985, 119, 416. 47. N. Janes and E. Oldfield, J . A m . Chem. SOC.,1985, 107,6769. 48. E. R. Andrew, D. J. Bryant, E. M. Cashell and B. A. Dunell, Chem. Phys. Lett., 1981, 77, 614. 49. U. Haubenreisser, G. Scheler and A.-R. Grimmer, Z . Anorg. Allg. Chem., 1986, 532, 157. 50. N. E. Burlinson, B. A. Dunell and J. A. Ripmeester, J . M a p . Reson., 1986, 67, 217. 51. A.-R. Grimmer and U. Haubenreisser, Chem. Phys. Lett., 1983,99,487. 5 2 . T. M. Duncan and D. C. Douglass, Chem. Phys., 1984, 87, 339. 53. I. L. Mudrakovskii, V. P. Shmachkova, N. S. Kotdarenko and V. M. Mastikhin, J . Phys. Chem. Solids, 1986,47, 335. 54. G. L. Turner, K. A. Smith, R. J. Kirkpatrick and E. Oldfield, J . Magn. Reson., 1986.70, 408. 55. S. Prabhakar, K. J. Rao and C. N. R. Rao, Chem. Phys. Lett., 1987, 139, 96. 56. S. Hayashi and K. Hayamizu, Bull. Chem. SOC. Jpn, 1989, 62, 3061. 57. D. Mueller, W. Gessner, H.4. Behrens and G. Scheler, Chem. Phys. Leu., 1981, 79, 59. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

86

S. HAYASHI

58. L. B. Alemany and G. W. Kirker, J . Am. Chem. SOC.,1986, 108, 6158. 59. L. B. Alemany, H. K. C. Timken and I. D. Johnson, J. Magn. Reson., 1988, 80,427. 60. S. J. Opella, J. G. Hexem, M. H. Frey and T. A. Cross, Phil. Trans. R. SOC.London A , 1981,299,665. 61. N. Zumbulyadis, P. M. Henrichs and R. H. Young, J . Chem. Phys., 1981, 75, 1603. 62. A. Naito, S. Ganapathy and C. A. McDowell, J . Chem. Phys., 1981, 74, 5393. 63. A. C. Olivieri and G. R. Hatfield, J . Magn. Reson., 1991,94, 535. 64. S. Ganapathy, S. Schramm and E. Oldfield, J . Chem. Phys., 1982, 77, 4360. 65. G. L. Turner, K. A. Smith, R. J. Kirkpatrick and E. Oldfield, J. Magn. Reson., 1986,67, 544. 66. P. F. Barron, R. L. Frost and J. 0. Skjemstad, J. Chem. SOC., Chem. Commun., 1983, 581. 67, T. Watanabe, H. Shimizu, A. Masuda and H. Saito, Chem. Led., 1983, 1293. 68. J. Rocha and J. Klinowski, Phys. Chem. Minerals, 1990, 17, 179. 69. J. S. Hartman and B. L. Sherrif, J. Phys. Chem., 1991, 95, 7575. 70. W. E. Blumberg, Phys. Rev., 1960, 119, 79. 71. D. Tse and S . R. Hartmann, Phys. Rev. Let!., 1968, 21,511. 72. M. R. McHenry, B. G. Silbernagel and J. H. Wernick, Phys. Rev. B, 1972, 5 , 2958. 73. J. Klinowski, T. A. Carpenter and J. M. Thomas, J. Chem. SOC.,Chem. Commun., 1986, 956. 74. J. Haase, H. Pfifer, W. Oehme and J. Klinowski, Chem. Phys. Lett., 1988, 150, 189. 75. J. Haase, K. D. Park, K. Guo,H. K. C. Timken and E. Oldfield, J . Phys. Chem., 1991, 95, 6996. 76. D. J. Cookson and B. E. Smith, J . Magn. Reson., 1985,63, 217. 77. L. J. M. van de Ven, J. G. Post, J. H. C. van Hooff and J. W. de Haan, J. Chem. SOC., Chem. Commun., 1985, 214. 78. P . F. Barron, R. L. Frost, J. 0. Skjemstad and A. J. Koppi, Nature (Lond.), 1983, 302, 49. 79. R. A. Young and A. W. Hewat, Clays Clay Miner., 1988, 36, 225. 80. D. L. Bish and R. B. Von Dreele, Clays Clay Miner., 1989,37,289. 81, E. Akiba, H. Hayakawa, H. Asano, F. Izumi, R. Miyawaki, S. Tomura and Y. Shibasaki, Clays Clay Miner., submitted for publication. 82. J. G. Thompson, Clays Clay Miner., 1984, 32, 233. 83. J. G. Thompson and P. F. Barron, Clays Clay Miner., 1987, 35, 38. 84. T. Watanabe, H. Shimizu,A. Masuda and H. Saito, Chem. Lett., 1983, 1293. 85. S. Komarneni, C. A. Fyfe and G. J. Kennedy, Clay Miner., 1985, 20, 327. 86. Y. V. Shulepov, A. S. Litovchenko, A. A. Melnikov, V. Y. Proshko and V. V. Kulik, J. Magn. Reson., 1983,53, 178. 87. J. P. Yesinowski, H. Eckert and G. R. Rossman, 1. Am. Chem. SOC.,1988, 110, 1367. 88. P. F. Barron, R. L. Frost and J. 0. Skjemstad, 1. Chem. SOC., Chem. Commun., 1983, 581. 89. W. E. E. Stone and R.-M. Torres-Sanchez, J . Chem. SOC., Faraday Trans. I, 1988, 84, 117. 90. M. C. Gastuche, F. Toussaint and J. J. Fripiat, Clay Miner. Bull., 1963,5, 227. 91. C. Otero-Arean, M. Letellier, B. C. Gerstein and J. J. Fripiat, Proceedings of the International Clay Conference, Bologna and Pavia, 1981 (eds H. Van Olphen and F. Veniale), p. 73. Elsevier, Amsterdam, 1982. 92. R. H. Meinhold, K. J. D. MacKenzie and I. W. M. Brown, J. Muter. Sci. Lett., 1985, 4, 163. 93. T. Watanabe, H. Shimizu,K. Nagasawa, A. Masuda and H. Saito, Cluy Miner., 1987, 22, 37. 94. J.-P. Gilson, G. C. Edwards, A. W. Peters, K. Rajagopalan, R. F. Wormsbecher, T. G. Roberie and M. P. Shatlock, J . Chem. SOC., Chem. Commun., 1987,91.

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

87

95. J. Sam, A. Madani, J. M. Serratosa, J. S. Moya and S . Aza, J. Am. Ceram. SOC., 1988, 71, C-418. 96. J. F. Lambert, W. S. Millman and J. J. Fripiat, J . Am. Chem. SOC., 1989, 111, 3517. 97. J. Rocha and J. Klinowski, Phys. Chem. Miner., 1990, 17, 179. 98. J. Rocha and J. Klinowski, Angew. Chem. Znt. Ed. Engl., 1990, 29, 553. 99. J. Rocha, J. Klinowski and J. M. Adams, J . Mater. Sci., 1991, 26, 3009. 100. J. Rocha, J. M. A d a m and J. Klinowski, J. Solid State Chem., 1990, 89, 260. 101. J. Rocha and J. Klinowski, J . Chem. SOC., Chem. Commun., 1991, 582. 102. J. Rocha, J. Klinowski and J. M. Adams, J. Chem. SOC., Faraduy Trans., 1991, 87, 3091. 103. G. C. Bye, Portland Cement, Pergamon, Oxford, 1983. 104. A.-R. Grimmer, R. Peter, E. Fechner and G. Molgedey, Chem. Phys. Lett., 1981, 77, 331. 105. A.-R. Grimmer, F. von Lampe, M. Maegi and E. Lippmaa, Cement Concrete Res., 1985, 15, 467. 106. J. Skibsted, J. Hjorth and H. J. Jakobsen, Chem. Phys. Lett., 1990, 172,279. 107. J. R. Barnes, A. D. H. Clague, N. J. Clayden, C. M. Dobson, C. J. Hayes, G. W. Groves and S. A. Rodger, J . Mater. Sci. Lett., 1985, 4, 1293. 108. E. Lippmaa, M. Maegi, M. Tarmak, W. Wieker and A.-R. Grimmer, Cement Concrete Res., 1982, 12, 597. 109. N. J. Clayden, C. M. Dobson, C. J. Hayes and S. A. Rodger, J. Chem. Soc., Chem. Commun., 1984, 1396. 110. S. A. Rodger, G. W. Groves, N. J. Clayden and C. M. Dobson, J. Am. Ceram. SOC., 1988, 71, 91. 111. J . F. Young, J. Am. Ceram. SOC., 1988, 71, C-118. 112. C. M. Dobson. D. G. C. Goberdhan, J. D. F. Ramsay and S. A. Rodger, J. Ma&. Sci., 1988, 23, 4108. 113. J. Hjorth, J. Skibsted and H. J. Jakobsen, Cement Concrete Res., 1988, 18, 789. 114. G. Parry-Jones, A. J. Al-Tayyib and A. I. Al-Mana, Cement Concrete Res., 1988, 18,229. 115. G. Parry-Jones, A. J. Al-Tayyib, S. U. Al-Dulaijan and A. I. Al-Mana, Cement Concrete Res., 1989, 19, 228. 116. S. U. Al-Dulaijan, G. Parry-Jones, A.-H. J. AI-Tayyib and A. I. Al-Mana, J. Am. Ceram. SOC., 1990, 73, 736. 117. B. H. W. S. de Jong, C. M. Schramm and V. E. Parziale, Geochim. Cosmochim. Acta, 1983, 47, 1223. 118. D. Mueller, A. Rettel, W. Gessner and G. Scheler, 1. Magn. Reson., 1984, 57, 152. 119. D. Mueller, W. Gessner, A. Samoson, E. Lippmaa and G. Scheler, Polyhedron, 1986, 5 , 779. 120. J. Skibsted, H. Bildsoe and H. J. Jakobsen, J. Mugn. Reson., 1991, 92, 669. 121. H. Stade, D. Mueller and G. Scheler, 2.Anorg. Allg. Chem., 1984, 510, 16. 122. E. Dupree and R. F. Pettifer, Nature (Lond.), 1984,308, 523. 123. C. M. Schramm, B. H. W. S. de Jong and V. E. Parziale, J. Am. Chem. Soc., 1984, 106, 4396. 124. A.-R, Grimmer, M. Maegi, M. Haehnert, H. Stade, A. Samoson, W. Wieker and E. Lippmaa, Phys. Chem. Glasses, 1984, 25, 105. 125. R. Dupree, D. Holland, P. W.McMillan and R. F. Pettifer, 1. Non-Crysf. Solids, 1984, 68,399. 126. U. Selvaray, K. J. Rao, C. N. R. Rao, J. Klinowski and J. M. Thomas, Chem. Phys. Lett., 1985, 114, 24. 127. J. B. Murdoch, J. F. Stebbins and I. S. E. Carmichael, Am. Mineral., 1985, 70, 332. 128. C. N. R. Rao, J. M. Thomas, J. Klinowski, U. Selvaraj, K. J. Rao, G. R. Millward and S. Ramdas, Angew. Chem. Int. Ed. Engl., 1985, 24, 61. 129. A.-R. Grimmer and W. Mueller, Monutsh. Chem., 1986,117,799. 130. E. Schneider, J. F. Stebbins and A. Pines, J. Non-Cryst. Solids, 1987, 89, 371.

88

S. HAYASHI

131. E. Lippmaa, A. Samoson, M. Maegi, R. Teeaeaer, J. Schraml and Goetz, J . Non-Cryst. Solids, 1982, 50, 215. 132. S . H. Risbud, R. .I. Kirkpatrick, A. P. Taglialavore and B. Montez, J . Am. Cerum. Soc., 1987, 70, C-10. 133. R. K. Sato, P. F. McMillan, P. Dennison and R. Dupree, J. Phys. Chem., 1991,95,4483. 134. Von E. Hallas, U. Haubenreisser, M. Haehnert and D. Mueller, Glastechn. Ber., 1983, 56, S63. 135. G. Engelhardt, M. Noh, K. Forkel, F. G. Wihsmann, M. Maegi, A. Samoson and E. Lippmaa, Phys. Chem. Glasses, 1985, 26, 157. 136. R. Oestrike, W.-H. Yang, R. J. Kirkpatrick, R. L. Hervig, A. Navrotsky and B. Montez, Geochim. Cosmochim. Acta, 1987, 51, 2199. 137. S. Sakka, Treatise on Mutenah Science and Technology (eds M. Tornozawa and R. H. Doremus), p. 129. Academic Press, New York, 1982. 138. R. Dupree, D. Holland, M. G. Mortuza, M. E. Smith, A. Chen and P. F. James, Mugn. Reson. Chem., 1990,28, S89. 139. Ch. Wies, K. Meise-Gresch, W. Mueller-Warmuth, W. Beier, U. Wellbrock and G. H. Frischat, J. Non-Cryst. Solids, 1990, 116, 161. 140. D. MueUer, P. Starke, M. Jank, K.-P. Wendlandt, H. Bremer and G. Scheler, Z. Anorg. A&. Chem.. 1984, 517, 167. 141. A. D. Irwin, J. S. Holmgren and J. Jonas, J. Muter. Sci., 1988, 23,2908. 142, A. Yasumori, M. Iwasaki, H. Kawazoe, M. Yamane and Y. Nakamura, Phys. Chem. Glasses, 1990, 31, 1. 143. J. Sam, I. Sobrados, A. L. Cavalieri, P. Pena, S. de Aza and J. S. Moya, J. Am. Ceram. SOC., 1991, 74, 2398. 144. Ch. Wies, K. Meise-Gresch, W.Mueller-Warmuth, W. Beier, A. A. Goektas and G. H. Frischat, Phys. Chem. Glasses, 1990, 31, 138. 145. U. Selvaraj, S. Komarneni and R. Roy, J . Am. Cerum. SOC., 1990, 73, 3663. 146. J. Kuemmerlen, L. H. Merwin, A. Sebald and H. Keppler, J . Phys. Chem., 1992, 96, 6405. 147. M. Villa, G. Chiodelli and M. Scagliotti, Solid State Ionics, 1986, 18&19, 382. 148. M. Villa, K. R. Carduner and G. Chiodelli, Phys. Chem. Glasses, 1987, 28, 131. 149. S. Hayashi and K. Hayamizu, J. Solid Stute Chem., 1989,80, 195. 150. R. K. Brow, C. C. Phifer, G. L. Turner and R. J. Kirkpatrick, 1. Am. Cerum. SOC., 1991, 74, 1287. 151. S. Prabhakar, K. J. Rao and C. N. R. Rao, Chem. Phys. Lett., 1987, 139, 96. 152. D. Mueller, G. Berger, I. Grunze, G. Ladwig, E. Hallas and U . Haubenreisser, Phys. Chem. Glusses, 1983, 24, 37. 153. P. J. Bray and G. OKeefe, Phys. Chem. Glasses, 1963, 4, 37. 154. P. J. Bray, M. Leventhal and H. 0. Hooper, Phys. Chem. Glasses, 1963,4, 47. 155. S. Greenblatt and P. J. Bray, Phys. Chem. Glasses, 1967, 8,213. 156. K. S. Kim and P. J. Bray, J . Nonmetals, 1974, 2 , 95. 157. Y.H. Yun and P. J. Bray, J . Non-Cryst. Solids, 1981, 44,227. 158. P. J. Bray, A. E. Geissberger, F. Bucholtz and I. A. Harris, J . Non-Cryst. Solids, 1982, 52, 45. 159. S. Hayashi and K. Hayamizu, J . Non-Cryst. Solids, 1989, 111, 214. 160. B. C. Bunker, R. J. Kirkpatrick, R. K. Brow, G. L. Turner and C. Nelson, J . Am. Cerum. SOC., 1991, 74, 1430. 161, T. L. Weeding, B. H. W. S. de Jong, W. S. Veeman and B. G. Aitken, Nature (Lond.), 1985, 318, 352. 162. W.-H. Yang and R. J. Kirkpatrick, 1. Am. Cerum. SOC., 1986, 69, C-222. 163. B. C. Bunker, D. R. Tallant, R. J. Kirkpatrick and G. L. Turner, Phys. Chem. Classes, 1990, 31, 30.

HIGH-RESOLUTION SOLID-STATE NMR STUDIES ON CERAMICS

89

164. B. C. Bunker, D. R. Tallant, T. J. Headley, G. L. Turner and R. J. Kirkpatrick, Phys. Chem. Glasses, 1988, 29, 106. 165. S. W. Martin, D. Bain, K. Budhwani and S. Feller, J. Am. Cerum. SOC., 1992,75, 1117. 166. J. Zhong, X.Wu, M. L. Liu and P. J. Bray, J . Non-Cryst. Solids, 1988, 107, 81. 167. M. Villa, K. R. Carduner and G. Chiodelli, J . Solid Stare Chem., 1987, 69, 19. 168. .I. E. Shelby, C. M. Shaw and M. S. Spess, J. Appl. Phys., 1989, 66, 1149. 169. D. Lathrop and H. Eckert, J. Phys. Chem., 1989, 93, 7895. 170. D. Lathrop and H.Eckert, J . Am. Chem. SOC., 1989, 111,3536. 171. D. Lathrop and H. Eckert, J. Am. Chem. SOC., 1990, 112,9017. 172. D. Lathrop and H.Eckert, Phys. Rev. B , 1991, 43, 7279. 173. M. Tullius, D. Lathrop and H. Eckert, J . Phys. Chem., 1990, 94, 2145. 174. H. Eckert, Z. Zhang and J. H. Kennedy, J. Non-Cryst. Solids, 1989, 107, 271. 175. Z. Zhang, J. H. Kennedy and H. Eckert, J. Am. Chem. SOC., 1992, 114, 5775. 176. Z. Zhang, J. H. Kennedy, J. Thompson, S. Anderson, D. A. Lathrop and H . Eckert, Appl. Phys. A , 1989, 49, 41. 177. D. Franke, R. Maxwell, D. Lathrop and H. Eckert, J. Am. Chem. SOC.,1991, 113, 4822. 178. D. C. Apperley, R. K. Harris, G. L. Marshall and D. P. Thompson, J. Am. Cerum. SOC., 1991, 74, 777. 179. G. R. Finlay, J. S. Hartman, M. F. Richardson and B. L. Williams, J. Chem. SOC.,Chem. Commun., 1985, 159. 180. J. S . Hartman, M. F. Richardson, B. L. Sherriff and B. G. Winsborrow, J. Am. Chem. SOC., 1987, 109, 6059. 181. J. R. Guth and W. T. Petuskey, J. Phys. Chem., 1987, 91, 5361. 182. G. W. Wagner, B.-K. Na and M. A. Vannice, J. Phys. Chem., 1989, 93, 5061. 183. N. R. Dando and M. A. Tadayyori, J. Am. Cerum. SOC., 1990,73, 2242. 184. D. J. O’Donnell, T. T. P. Cheung, G. F. Schuette and M. Sardashti, J. Am. Cerum. SOC., 1991, 74, 2025. 185. T. Taki, S. Maeda, K. Okamura, M. Sat0 and T. Matsuzawa, J. Muter. Sci. Lett., 1987,6, 826. 186. T. Taki, K. Okamura, M. Sato, T. Seguchi and S. Kawanishi, J. Muter. Sci. Lett., 1988,7, 209. 187. T. Taki, M. Inui, K. Okamura and M. Sato, J . Muter. Sci. Letr., 1989, 8, 918. 188. T. Taki, M. Inui, K. Okamura and M. Sato, J. Muter. Sci. Lett., 1989, 8, 1119. 189. T. Taki, K. Okamura and M. Sato, J. Muter. Sci., 1989, 24,1263. 190. Y. Hasegawa and K. Okamura, J. Mater. Sci., 1986, 21, 321. 191. F. Babonneau, G. D. Soraru and J. D. Mackenzie, J. Muter. Sci., 1990, 25, 3664. 192. F. Babonneau, G. D. Soraru, K. J. Thorne and J. D. Mackenzie, J. Am. Cerum. SOC., 1991, 74, 1725. 193. K. R. Carduner, C. S. Blackwell, W. B. Hammond, F. Reidinger and G. R. Hatfield, J. Am. Chem. SOC., 1990, 112,4676. 194 R. Dupree, M. H. Lewis, G . Leng-Ward and D. S. Williams, J. Muter. Sci. Len., 1985.4, 393. 195. K. R. Carduner, R. 0.Carter 111, M. E. Milberg and G. M. Crosbie, Anal. Chem., 1987, 59, 2794. 196. K. R. Carduner, R. 0. Carter 111, M. J. Rokosz, C. Peters, G. M. Crosbie and E. D. Stiles, Chem. Muter., 1989, 1, 302. 197. N. R. Keskar and J. R. Chelikowsky, Phys. Rev. B , 1992, 46, 1. 198. C. A. Fyfe, G. C. Gobbi, J. Klonowski, J. M. Thomas and S. Ramdas, Nuture (Lond.), 1982, 296, 530. 199. C. A. Fyfe, J. H. O’Brien and H. Strobl, Nuture (Lond.), 1987, 326,281. 200. J. M. Thomas, J. M. Gonzalez-Calbert, C. A. Fyfe, G. C. Gobbi and M. Nicol, Geophys. Res. Lett., 1983, 10, 91.

90

S.HAYASHI

201. C. A. Fyfe, G. C. Gobbi and A. Putnis, J. Am. Chem. Soc., 1986,108,3218. 202. J. Senegas, A. R. Grimmer, D. Muller, P. Thomas, D. Mercurio and B. Frit, J . Muter. Sci., 1991, 26, 5053. 203. H. D. Moms and P. D. Ellis, 1. Am. Chem. SOC., 1989, 111, 6045. 204. G. C. Gobbi, R. Christoffersen, M. T. Otten, B. Miner, P. R. Buseck, G. J. Kennedy and C . A. Fyfe, Chem. Lerr., 1985,771. 205. R. Dupree, M. H. Lewis and M. E. Smith, J. Appl. Crystullogr., 1988, 27, 109. 206. D. Massiot, C. Bessada, J. P. Coutures and F. Taulelle, J. Mugn. Reson., 1990, 90, 231. 207. S. Hayashi, K. Hayamizu and 0. Yamamoto, Bull. Chem. SOC. Jpn, 1987, 60,761. 208. S. F. Dec, G. E. Made1 and J. J. Fitzgerald, J. Am. Chem. SOC., 1990, 112, 9069. 209. S. Hayashi, K. Hayamizu and 0. Yamamoto, Jpn. J . Appl. Phys., 1987, 26,682. 210. N. D. Butler, R. Dupree and M. H. Lewis, J . Muter. Sci. Lett., 1984, 3, 469. 211. J. Klinowski, J. M. Thomas, D. P. Thompson, P. Korgul, K. H. Jack, C. A. Fyfe and G. C. Gobbi, Polyhedron, 1984, 11, 1267. 212. R. Dupree, M. H.Lewis and M. E. Smith, J . Am. Chem. SOC., 1988, 110, 1083. 213. R. Dupree, M. H. Lewis and M. E. Smith, J. Am. Chem. SOC., 1989, 111, 5125. 214. T. Bastow, S. G. Hardin and T. W. Turney, J. Muter. Sci., 1991,26, 1443. 215. S. Schramm and E. Oldfield, J. Am. Chem. SOC., 1984,106,2502. 216. G. L. Turner, S. E. Chung and E. Oldfield, J. Mugn. Reson., 1985, 64, 316. 217. E. Oldfield, C. Coretsopoulos, S. Yang, L. Reven, H. C. Lee, J. Shore, 0. H. Han and E. Ramli, Phys. Rev. €3, 1989, 40,6832. 218. T. J. Bastow and S. N. Stuart, Chem. Phys., 1990, 143, 459. 219. K. T. Mueller, Y. Wu, B. F. Chmelka, J. Stebbins and A. Pines, J. Am. Chem. SOC., 1991, 113, 32. 220. R. Dupree and M. E. Smith, J. Chem. SOC., Chem. Commun., 1988, 1483. 221. R. Dupree and M. E. Smith, Chem. Phys. Lett., 1988, 148.41. 222. C. P. Grey, M. K. Smith, A. K. Greetham, C. M. Dobson and R. Dupree, J. Am. Chem. SOC.,1990, 112,4670. 223. J. S. Hartman, F. P. Koffyberg and J . A. Ripmeester, J. Mugn. Reson., 1991.91, 400. 224. W. P. Rothwell, J. S, Waugh and J. P. Yesinowski, 1. Am. Chem. SOC., 1980, 102, 2637. 225. J. Tropp, N. C. Blumenthal and J. S. Waugh, 1. Am. Chem. SOC., 1983, 105,22. 226. P. S. Belton, R. K. Harris and P. J. Wilkes, J. Phys. Chem. Solih, 1988,49, 21.

NMR Studies of Zeolites H. PFEIFER and H. ERNST Universitat Leipzig, Fachbreich Physik, Linnkstr. 5, 04103 Leipzig, Germany 1. Introduction 2. Framework of zeolites 2.1. High-resolution solid-state NMR of nuclei in zeolite lattices-general aspects 2.2. 29Si MAS-NMR studies 2.3. 27AIMAS-NMR studies 3. Bronsted acid sites 3.1. Unloaded (evacuated) zeolites 3.2. Accessibility of Bronsted acid sites, hydrated zeolites 4. Lewis acid sitedextra-framework aluminium 4.1. ”Al NMR spectroscopy 4.2. Use of probe molecules 5. Structure of adsorbed molecules 5.1. High-resolution NMR of adsorbed molecules-general aspects 5.2. 13C and ‘H MAS-NMR studies on sealed samples 6. Molecular diffusion 6.1. Basic principles 6.2. Intracrystalline diffusion of hydrocarbons 6.3. Anisotropic molecular diffusion 6.4. NMR tracer desorption technique 7. Chemical reactions 7.1. Experimental methods for in situ MAS-NMR reaction studies on sealed samples 7.2. Conversion of methanol 7.3. Synthesis of methylamines References

91 93 93 97 101 106 106 123 126 127 129 132 132 133 138 138 147 152 154 159 159 164 166 176

1. INTRODUCTION

Zeolites’4 are porous inorganic crystallites built from TO4 tetrahedra. In the case of the original zeolites T represents silicon or aluminium, for the ALPOs (aluminophosphates) aluminium or phosphorus, for the SAPOs (silicoaluminophosphates) silicon, aluminium or phosphorus while in the so-called MeAPOs and MeAPSOs additional metal atoms Me such as Mg, Mn, Fe, Co, or Zn are incorporated. The word ZEOLITE, which means “boiling stone” in Greek, was coined by the Swedish scientist Cronstedt’ in ANNUAL REPORTS ON NMR SPECTROSCOPY VOLUME 28 ISBN 0-12-505328-2

Copyright 0 1994 Academic Press Limited AN rights of reproduction in any form reserved

92

H. PFEIFER AND H. ERNST

1756 to describe the behaviour of the newly discovered mineral stilbite which, when heated, rapidly loses water and thus seems to boil. The original zeolites can be represented by the general formula (Si02),,(AlOz)-M+ where n 3 1 denotes the silicon-to-aluminium ratio of the zeolite framework (Si02),,(A102)-. This framework is built from corner-sharing SiOj- and A102- tetrahedra and contains regular systems of intracrystalline cavities and channels of molecular dimensions. The inner surface is formed nearly entirely by the oxygen atoms due to their relatively large ionic radius (the ionic radii of 02-,A13+, and Si4+ are 0.132 nm, 0.051 nm, and 0.042 nm, respectively). The negative charge of the framework and hence the concentration of the exchangeable cations M+ which are located on extra-framework sites6 can be made to zero by enhancing the silicon-toaluminium ratio so that in this limiting case zeolites exhibit a perfect homogeneous solid surface formed by oxygen atoms. The cations M+ neutralizing the electric charge of the framework can be exchanged for other cations of corresponding net charge (e.g. 1/2 M2+). An exchange with ammonium ions followed by a heat treatment leads to the formation of bridging OH groups (SiOHAl), which are known to be able to protonate adsorbed molecules (Bronsted acid sites). In the last 40 years more than 30 natural zeolites and more than 100 synthetic zeolites were identified. Several classifications of zeolite structures, based on the framework topology, have been proposed. For practical reasons it is useful to distinguish between large-pore zeolites with 12membered oxygen rings (e.g. zeolites of the faujasite type: X,Y, ZSM-20, see Table l), medium-pore zeolites with 10-membered rings (e.g. zeolites of the pentasil group: ZSMJ or silicalite, see Table 1) or small-pore zeolites with eight-membered rings (e.g. A-zeolites or ZK-4, see Table 1). The spatial arrangements of cavities and channels of the zeolites characterized in Table 1 are shown in Fig. 1 where each corner is occupied by a T-atom which is connected to four others by oxygen bridges (indicated by the straight lines). Zeolites are largely used in industry for a wide range of applications including adsorption and separation of gases and of hydrocarbons (“molecular sieves”), drying, ion exchange and catalysis. Catalysts based on zeolite Y containing bridging OH groups have a cracking activity which is orders of magnitude greater than that of conventional silica-alumina catalysts and have displaced them now almost completely. The synthetic zeolite ZSM-5 is very useful for many applications since its high silica content (n is typically not less than about 15) gives it high thermal stability, while the channel diameter leads to quite striking effects of shape selectivity including the ability to synthesize gasoline from methanol in a single step.

NMR STUDIES OF ZEOLITES

93

Table 1. Characteristic data of typical large-pore, small-pore and medium-pore zeolites. Faujasite group: zeolites X,Y, ZSM-20 Unit cell

Large cavities/ channels

Zeolite A, ZK-4

Cubic a = b = c = 2.47nm

Cubic a = b = c = 2.46nm

8 cubooctahedra

8 cubooctahedra +24 oxygen bridges 384 0; 192 T Inner diam.: 1.14nm window (8 O-ring): 0.41 nm (NaA) 0.50 nm (CaA)

+16 oxygen bridges 384 0; 192 T Inner diam.: 1.16nm window (12 O-ring): 0.74 nm

Cubooctahedra (small cavities)

Inner diam.: 0.66 nm; window (6 O-rings): 0.25 nm

Number of T-atoms

24 per large cavity

Number of large cavitiedintersections per g zeolite

4.18 X

Id"( n

= m)

3.53 X I d n ( n = 1)

Pentasil group: ZSM-5, silicalite Orthorhombic a = 2.01 nm b = 1.99nm c = 1.34nm 4 intersections 129 0;96 T Straight channels: 0.53 x 0.56 nm2 zig-zag channels: 0.51 x 0.55 nm2 (10 O-rings)

24 per intersection 4.18 X 10" (n = 0 0 )

Summarized, zeolites share the following five properties: (1) well-defined crystalline s t r ~ c t u r e , ~ (2) large internal surface areas (a 600 m2/g), (3) uniform pores, (4) good thermal stability, and ( 5 ) Bronsted acidity of adjustable concentration and strength, that make them attractive not only for practical application (see above) but also for basic research studies and especially to NMR spectroscopists since virtually every chemical element they contain can be studied by N M R with a sufficiently large signal-to-noise ratio due to the large internal surface areas. 2. FRAMEWORK OF ZEOLITES 2.1. High-resolution solid-state NMR of nuclei in zeolite lattices-general aspects

The field of N M R has developed very rapidly during the last 30 years especially due to the invention of techniques that allow a measurement of

94

H. PFEIFER AND H. ERNST

12 RING

ZEOLITE

PENTASIL ZEOLITE

@ 8 RING

A ZEOLITE

Fig. 1. Framework structures of typical large-pore, medium-pore, and small-pore zeolites (cf. Table 1).

highly resolved NMR spectra in solids. As is well known, the Hamiltonian which determines the properties of a spin system can be written as

HO = H z + H D + HCsA+ H j + H o + HRF

(1)

where Hz is the Zeeman interaction including the isotropic chemical shift between the nucleus and the applied static magnetic field Bo. HD denotes the magnetic dipolar interaction, HCsAthe interaction due to chemical shift

NMR STUDIES OF ZEOLITES

95

anisotropy, HJthe indirect electron-coupled nuclear spin interaction, HQ the electric quadrupolar interaction which is only present for nuclei with spin Z> 1 and HRF the interaction with alternating magnetic fields, mostly applied in the form of short rf pulses. In non-viscous liquids HD and Ho average to zero, which permits the much weaker isotropic value of the chemical shift and the J coupling to be studied. In the solid state, however, the homonuclear dipolar, the heteronuclear dipolar and the quadrupolar interactions as well as the chemical shift anisotropy are dominant, which leads to such a broadening of the NMR lines that in general the valuable information contained in the chemical shift and J-coupling (fingerprint system of liquid state NMR7) is hidden. To obtain highly resolved NMR spectra in solids, a selective removal or alteration of each of the broadening interactions is necessary. Generally, manipulations of spatial or spin variables of the Hamiltonian given in equation (1) can be performed. The new techniques of solid-state NMR, namely multiple pulse sequences,g2o high-power dipolar decoupling,21 cross-polarization (CP) in combination with high-power d e c ~ u p l i n g , ~ ~ ~ ~ magic angle spinning (MAS) of the sample,25726combined rotation and multiple pulse sequence (CRAMPS) ,27-30 variable angle spinning (VAS),31"4 dynamic angle spinning (DAS)35-37and double rotation technique (DOR)38-41eliminate or at least reduce to a certain degree the influence of HD,HQ and the chemical shift anisotropy HCSA, thus making possible the observation of one- and two-dimensional highly resolved NMR spectra in solids. The benefits of strong magnetic fields lie at least in three factors: (1) the sensitivity is roughly proportional to Bo to the power 3/2; (2) for quadrupolar nuclei with half-integer spin, lines tend to become narrower at higher fields, since the second-order quadrupolar broadening is inversely proportional to Bo, and (3) since the chemical-shift dispersion is proportional to Bo, potentially more chemical-shift information becomes available. More details and reviews of the present state of the high resolution NMR in solid^^^,^^ and applications to zeolites and related molecular sieves and catalysts4357 can be found in the literature. In ref. 55 an instructive and concise survey on the applicability of multinuclear solid-state NMR in structural studies of silicates, aluminosilicates and zeolites is given. From the point of view of NMR, each of the four basic atomic constituents of the framework of zeolites including aluminophosphatesoxygen, silicon, aluminium and phosphorus-are amenable to measurements by their naturally occurring isotopes 1 7 0 , "Al, 29Siand 31P (see Table 2). However, the 1 7 0 isotope has a very low natural abundance and a nuclear quadrupole moment giving rise to a large line broadening. Moreover, the isotopic "0enrichment is difficult. Therefore, 1 7 0 NMR on zeolites has been little used till now. The 27Al isotope has a natural abundance of 100% but a quadrupole moment ( I = 5/2), which may also cause a

96

H. PFFJFER AND H. ERNST

Table 2. N M R properties of some selected nuclei which may be used in NMR studies of zeolites. The relative sensitivity is defined as the product of natural abundance and NMR relative sensitivity to protons. The resonance frequency is given for a magnetic field of ca. 11.74T.

Isotope

Spin

Natural abundance (%)

Relative sensitivity

NMR frequency (MW

~~

'H 7 ~ i I*B 170

23Na 2 7 ~ 1

29~i 3*P 69Ga 'lGa

73Ge

112 312 312 512 312 512 112 112 312 312 912

99.985 92.58 80.42 0.037 100.00 100.00 4.70 100.00 60.4 39.6 7.76

1.om 0.27 0.13 1.08x 10-5 9.25 x lo-' 0.21 3.69 x 10-4 6.63 x lo-' 4.17 x 5.62 X lo-' 1.08 x 10-4

500 194.3 160.4 67.8 132.3 130.3 99.3 202.4 120.0 152.5 17.4

strong broadening of the NMR signals. The 29Si isotope is 4.7% abundant and isotopic enrichment is also difficult. But in contrast to 1 7 0 and 27Al, the 29Sinucleus has a spin 1/2, so that Ho = 0 in equation (1) and one may get narrow resonance lines. Compared with the 27Al isotope the longitudinal relaxation time of 29Si is often very long. In these cases the NMR experiments are time-consuming. Due to its natural abundance of 100% and the absence of an electric quadrupole moment, 31P solid-state NMR is the most widely used analytical method for zeolites of type A L P 0 and SAPO. Several other elements that replace silicon or aluminium in the zeolitic framework (e.g. B, Ga, Ge) or are present as charge-compensating cations (e.g. Li+, Na+) can be studied by solid-state NMR of their respective NMR-active isotopes ("B, 69Ga, 71Ga, 73Ge, 'Li, 23Na). With respect to the 'H MAS-NMR see the sections on Bronsted acid sites and Lewis acid sites/extraframework aluminium. Since the first application of solid-state NMR to zeolites, important information about the zeolitic framework has been retrieved5* concerning especially (1) the coordination and local atomic environment of silicon, aluminium and phosphorus; (2) the number of distinct tetrahedral sites and of the framework composition (silicon-to-aluminium ratio) from an analysis of the 27Si MAS-NMR spectra; (3) the de- and re-alumination, and de- and re-gallination, and last but not least (4) the characterization of Bronsted acid and Lewis acid sites. In the following, typical examples for the information of type (1)-(3) derived from 29Si, "Al and 1 7 0 MAS-NMR studies shall be given.

NMR STUDIES OF ZEOLITES

.................................................. ................................................... J - C

S1(4AI)-SOD

Si(4AI)

f SI I 2A I)

I

- 80 I

97

I

- 90 I

I

- 100 I

1

I

-110

1

-120

dsi/ P P ~ Fig. 2. Ranges of %i chemical shifts of Si(nA1) units in zeolites. Si(4Al)-SOD denotes Si(4Al) units in ~odalites.~'

2.2. 29Si MAS-NMR studies

The 29Si chemical shifts of silicates and ahminosilicates depend sensitively on the number and type of T-atoms (T = Si, A1 or other tetrahedral framework atoms) connected with a given Si04 tetrahedron. That means, distinct signals appear in the "Si MAS-NMR spectra for the five different Si(nA1) environments (n = 04).The typical shift ranges established from a large number of data measured in various types of zeolites and other framework silicates, are shown in Fig. 2.47The largest range observed for a Si(nA1) unit is that of Si(4A1) in aluminosilicate sodalites. As usual, all values of the chemical shift of the 29Sisignals are with respect to TMS. For an interpretation of the spectra obtained it is important to note that the relative signal intensities of the 29Sisignals are directly related to the relative concentrations of these units in the zeolite framework. Figure 3 shows the 29SiMAS-NMR spectra of a series of zeolites X and Y with different Si/Al ratios.59 With increasing SYA1 ratio, the decrease of the signal intensities of aluminium-rich units and a corresponding increase of the aluminium-poor Si(nA1) signals can be seen. Provided that no AlOAl linkages are present in the zeolite (i.e. Loewenstein's rule appliesa), the

98

H.PFEIFER AND H.ERNST

L

1.02

1.60

2.82 3 ,

2 1

3

4 l d . 4 h I

.

.

-80 -100 -120

-80 -100 -120 1

2.02

1.17

2

4

I

-80 -100 -120

-80 -100 -120

4

I

I

I

,

I

I

I

-80 -100 -120

2

,

-80 -100 -120

:.k

,

- 80

,

I

-160 -1'O ;

.

,

-80 -100 -120

Fig. 3. "Si MAS-NMR spectra of zeolites X and Y. The SilAl ratio is indicated for each spectrum, the peak assignments are given b the number n of the corresponding Si(nA1) unit. J

N M R STUDIES OF ZEOLITES

99

framework Si/AI ratio can be directly calculated from the 29Si MAS-NMR spectra according to the equation (Si/AI)F = 4.1I,,/Cn.I,

(2)

where n = 0-4 and I, are the intensities of the Si(nA1) peaks.59 Equation (2) is independent of the specific structure of the zeolite, but excludes the existence of silanol groups. Figure 4 shows the 29Si MAS-NMR spectra of some typical zeolites with relatively low SUAI ratio together with the corresponding highly siliceous forms obtained after a hydrothermal dealumination.61 In all highly siliceous materials, the resonances observed are only due to Si(0AI) units and are extremely narrow. Because of their high resolution, these spectra can be used in various ways to get subtle information regarding zeolite structures which are not easily obtainable by other techniques. The residual line broadening in the 29Si MAS-NMR spectra of highly siliceous zeolites is due to the chemical shift distribution arising from the distribution of the residual aluminium atoms and imperfections of the structure.61 The effect of a careful optimization of all factors affecting the spectral resolution is shown in Fig. 5 for a zeolite ZSM-5.62 From the optimized spectrum, 21 or 22 of the 24 postulated signals can be clearly observed depending also on the temperature of the measurement. In NMR studies on liquids, the application of two-dimensional (2D) techniques has provided a wealth of information on the connectivities between atoms within molecular structures. In principle, 2D MAS-NMR experiments can also be used to establish connectivities in the solid state. As typical examples, in Fig. 6 a 2D COSY and in Fig. 7 a 2D INADEQUATE 29Si MAS-NMR spectrum of a zeolite ZSM-12 is shown.49 The lattice framework of the zeolite ZSM-12 is represented schematically in Fig. 8.63In this figure, the seven crystallographically inequivalent tetrahedral lattice sites are indicated, and it can be seen that their relative concentrations are equal. Accordingly the 29Si MAS-NMR spectrum shows seven clearly resolved resonances with equal intensities for six of them. The fact that one of the signals exhibits a lower intensity was used by the authors to assign it to the T5 site which is the only site not accessible by adsorbed oxygen. Therefore the longitudinal relaxation rate and hence also the signal intensity of the 29Si nuclei occupying sites T5 must be reduced with regard to the others. Starting from this assignment of T5 the authors could proceed to assign the connectivities yielding the labelling of the cross-peaks given in Fig. 6. All of the connectivities found and assigned in the 2D COSY experiment are confirmed in the 2D INADEQUATE experiment shown in Fig. 7. This result includes the connectivity between the sites T4 and T6, which is clearly resolved in the 2D INADEQUATE experiment but was ambiguous in the 2D COSY experiment due to the close proximity of the cross-peaks to the diagonal.

100

H. PFEIFER AND H. ERNST

Si(1AI) Si(0AI)

Si(2AI) Si(2AI)

1M Si(0AI)

Si(1AI)

-90

-100

Si(0AI)

-110

-90

-1201

-110

-100

-120

Si(1AI) Si(2Al) Si(0Al)

-90 -100 -110 -120

-90 -100 -110 -120 SsllPPm

Fig. 4. 29SiMAS-NMR spectra of some typical zeolites with a relatively small Si/A ratio together with the corresponding highly siliceous forms: (A) zeolite Y, (B) mordenite, (C) offretite, (D) zeolite omega.6'

N M R STUDIES OF ZEOLITES

n

A

101

5 Hr

B

C

-

I

I

I

I

108

I

I

1

d5;/ PPm

I

I

I

-

I

118

Fig. 5. 29Si MAS-NMR spectra of a highly siliceous zeolite ZSM-5: (A) without special pretreatment, (B) after optimization of all experimental variables, (C) the simulated spectrum.62

2.3. "Al MAS-NMR studies

Compared with 29Si the nuclei of the other most abundant atoms of the zeolite framework, "Al and 1 7 0 , show a more complex NMR behaviour, since even the 112 to - 112 transition which yields the smallest line widths is distorted and shifted by the second-order quadrupolar interaction. The shift of the centre of gravity of this signal is given by:52

+

w c c - 0 ~=

-(1/30).0&/%.[I(I+

1)- (3/4)](1 +v2/3)

(3)

102

H. PFEIFER AND H.ERNST

T3

T4

T6

7-2

~ ~ " ' " " " " " " ' ~ ' " ~ ' ' . ' - . ' ~ . . -108.0

-109.0

-110.0 -111.0 -112.0 -113.0 &l/PPm

Fig. 6. Contour plot of the 2D COSY "Si MAS-NMR spectrum of a zeolite ZSM-12.

The projection in the F2 dimension (1D "Si MAS-NMR spectrum) is shown above.49

103

N M R STUDIES OF ZEOLITES

r

- S6

'TqT6

s5 s4

s3

s2 SI

L

-108.0

-109.0

-110.0

-111.0

WPPm

-112.0

-113.0

Fig. 7. Contour of the 2D INADEQUATE *'Si MAS-NMR spectrum of the same zeolite ZSM-12 as in Fig. 6. The projection in the F dimension (1D "Si MAS-NMR spectrum) is shown above.492

with oQ = 32qQ/(21(21- 1)fi). Here eq denotes the z-component of the electric field-gradient tensor, eQ the quadrupole moment of the nucleus, 7 the asymmetry parameter, and y the resonance frequency for Ho = 0. Therefore it is of advantage to measure at high fields Bo. This is illustrated in Fig. 9 which shows the "A1 MAS-NMR spectra of a zeolite Y at 23.5 and 1 0 4 . 2 M H ~ As . ~ ~usual, all 27Al NMR shifts are given with respect to the signal of an aqueous solution of aluminium chloride. Since the framework aluminium nuclei of hydrated zeolites give rise to a relatively narrow line in the 27Al MAS-NMR spectra, the higher NMR sensitivity of "Al compared with 29Si (cf. Table 2) allows a quantitative

104

H. PFEIFER AND H. ERNST

Fig. 8. Schematic representation of the lattice framework of a zeolite ZSM-12. The seven crystallographically inequivalent tetrahedral lattice sites are indicated by TI (=Sil) to T, (=Si7).63Ts is the only site which is not located on the surface of a channel wall.

determination of small concentrations of framework aluminium also in those cases where it cannot be seen in the 29SiMAS-NMR spectra. In contrast, for the non-framework aluminium species the 27Al MAS-NMR signals may be very broad which reduces dramatically the sensitivity of this method. Moreover, complications may arise if bmad signals of aluminium sites subject to different quadrupolar interactions overlap in the spectra. The separation of these lines can be achieved by application of the twodimensional quadrupole nutation MAS-NMR technique As an example, Fig. 10 shows "Al MAS-NMR spectra of framework dealuminated Y zeolites.65 The different lines could be resolved by 27Al nutation MASNMR. It has been concluded that the 27Al MAS-NMR spectra consist of three superimposed lines which correspond to framework ( A F ) , non-

N M R STUDIES OF ZEOLITES

200

100

0

105

-100

101.22 MHz

Fig. 9. 27A1MAS-NMR spectra of a zeolite Y obtained at (a) 23.45MHz and (b)

104.22M H z . ~ ~

framework octahedral (AINFo) and non-framework tetrahedral (AINFT) aluminium. Details of the dealumination process are described in a recent paper from a theoretical point of view on the basis of ab initio SCF MO calculations.% The results given in this study may contribute to a better understanding of the nature of non-framework aluminium species in hydrothermally treated zeolites. In Table 3 a survey is given of solid-state NMR studies performed on zeolites including aluminophosphates (zeolites of type ALPO, SAP0 etc.) since 1987. With regard to references before 1987 the reader is referred to ref. 43.

106

H. PFEIFER AND H. ERNST AIF

AINm

60

40

AtNFo

1

SiIAI

2

SilAl

3

io

0

6dPm

Fig. 10. "A1 MAS-NMR spectra of framework dealuminated zeolites Y. Integral intensities (in %) of the lines have been determined by the use of line sha es measured in a two-dimensional 27AI nutation experiment. A f , AINm, and AlgF0 denote tetrahedral framework Al, tetrahedral non-framework Al, and octahedral non-framework Al, re~pectively.~~

3. BRONSTED ACID SITES

3.1. Unloaded (evacuated) zeolites The elementary step of a catalytic reaction effected by a so-called Bronsted acid site is the proton transfer from an acidic surface hydroxyl group ZOH to the adsorbed molecule M:

ZOH + M 2 ZO-

+ MH+

(4)

N M R STUDIES OF ZEOLITES

107

Therefore, apart from a possible rate-determining influence of diffusion, the efficiency of a catalyst will be determined by three independent parameters for each sort of acidic surface hydroxyl groups (Bronsted acid sites): (1) the strength of acidity as defined qualitatively by the ability to protonate an adsorbed molecule; (2) the concentration; and (3) the accessibility of the respective OH gtoups. Apparently the ability to protonate an adsorbed molecule will depend both on the properties of the acidic site ZOH and of the molecule M. In order to define a quantity which characterizes the protonation ability of the ZOH group but does not depend on the particular molecule, the reaction described by equation (4)is decomposed into two processes:

ZOH T)c ZO-

+ H+

H+ + M jc MH+

(5) (6)

Denoting the standard Gibbs free energy change of the first process (equation 5) by AGgp we define the strength of acidity S, of the ZOH group in vacuum (the so-called gas phase acidity261)by S, = 1iAGgp

(7)

so that decreasing values for the standard Gibbs free energy of the deprotonation of ZOH correspond to increasing values of the strength of acidity as it should be. To compare the strength of acidity with the deprotonation energy of a ZOH group, AEDp,a quantity which follows from quantum chemical calculations, one must take into consideration that the standard Gibbs free energy change AGgp is the sum of the deprotonation energy AEDP,of the zero-point energy change AEgp and of the Gibbs free energy change which results from the conversion of the three vibrational degrees of freedom of the proton as part of the ZOH group into its three translational degrees of freedom after leaving this group. Assuming that the zero-point energy change AEgp is a constant and that the the reciprocal value of the deprocontribution of A G & F is tonation energy AEDpcan be taken approximately as a measure for the strength of acidity. With respect to nuclear magnetic resonance spectroscopy there are at least four arguments for the statement that the position of the 'H MAS-NMR signal of an OH group is related to its strength of acidity: (1) A qualitative argument. An increase of the so-called chemical shift S, of an OH group which is defined by &= -I UTMS-UH

(8)

108

H.PFEIFER AND H. ERNST

Table 3. Survey of solid-state NMR measurements which were performed on zeolites (including alurninophosphates)since 1987.

Zeolite Nucleus

Faujasitetype

ZSM-type, mordenite

Aluminophosphates

Others

”A1

65,’ 87: 88,’ 90, 56,75,’ 76,390, 110,’ 126,’ 130, 104,3117, 124, 131; 153,’ 165,’ 125, 126; 130, 183,4200,’ 131; 185,226: 231, 233,’ 234,’ 227,4*5229,4 237,3 238, 241,’ 243; 245 247

67,68, 78,3 80, 83,389,90, 98,3 81: 93, 107,6 104,3153: 217, 155, 156,3157, 245 160,161, 162,6 173,6 176,3180,6 198,205, 222, 223,3 228, 240 253, 256: 2575

29Si

82,’ 84,’ 101, 102,112,152, 154; 170, 171, 179,’ 189; 197, 211,’225,’249’

70,77,‘79,105; 106,3108, 118, 121,135, 136, 138,5139, 140,5 142-149,’ 150: 152,163: 172, 187,192; 239; 246, 258,’ 259’

127; 181, 194, 203

70, 97,3 100,119, 120, 122, 136, 137, 138; 140,5 141,’ 184, 194, 195, 215, 216, 230, 244

27A1

73,74,’86,91,’ 92,109: 115,3 116,3133,’ 158,’ 159; 164,’ 166,3 168, 169, 174, 188,’ 190, 193, 199,’ 204,* 212, 220,224; 232, 235,’ 242,2 252’

92,94,’95,’96, 98; 99,’ 114, 115,3123,133,’ 134,151: 186; 190,193,209, 210,213, 214,3 236,251’

69,85,92, 111,3 113; 132,167, 177,’ 193,201,’ 202; 218: 219: 232,250,’ 254,’ 2553

71,372, 103,’ 128,129: 151; 167,175, 178, 182,191,196; 206-208, 220, 221, 248

232

98, 99,238, 2604 67,78,80, 81,6 194 85,92, 111, 113, 132, 155-157, 160,161,167, 177,194, 198, 201,202,218, 219,222,223, 232,240,250, 254,255, 256,4 2575

+

29Si

31P

NMR STUDIES OF ZEOLITES

109

Table 3-contd.

"B

152

Other

86,116,130,131, 75,92,94-96,98, 92, 132, 181,254 83, 122, 128, 129, 151, 184, 208, 217, 133, 183, 18899,118, 123, 124, 190,245 130, 131, 133, 230,245, 248 151, 172, 185, 190,213,214, 226,231

152, 186

92,93

83,97

'insertion of nuclei in the framework of the zeolites; 'dealurnination of the zeolite framework; 3synthesisof zeolites; 4CP (cross-polarization) MAS-NMR experiments; '2D (two-dimensional) MAS-NMR experiments; 6DOR (double-rotation) MAS-NMR experiments.

with uTMS and uH as the chemical shielding of the 'H nucleus in tetramethysilane and the OH group, respectively, corresponds to an increase of the net atomic charge of the hydrogen atom. On the other hand, an enhancement of the net atomic charge will be accompanied by a reduction of the deprotonation energy. Hence, the chemical shift of OH groups should increase with increasing strength of acidity. (2) Experimental results for molecules in the gaseous state. In Fig. 11 values of S, which were measured by Chauvel and for various molecules in the gaseous state are plotted as a function of experimental values of AGL, taken from ref. 264. As can be seen, in agreement with the qualitative argument, the chemical shift increases with increasing strength of acidity (cf. equation 7). (3) Experimental results for zeolites of varying Si/AI ratio. In Fig. 12 values for Sanderson's intermediate electronegativity S, which were computed for zeolites of composition HA102(Si02), according to the formula265

with the following values of the atomic electronegativitie~:~~~ SH = 3.55, SAl = 2.22, So = 5.21, Ssi = 2.84 are plotted together with experimental , in dependence on the silicon-to-aluminium ratio n. The nearly values for 6 identical functional dependence of both quantities provides ample evidence for the usefulness and sensitivity of 8, as a measure for the strength of acidity. In this connection it should be mentioned that due to the use of methane as an inner standard2& the absolute values of 8~ plotted in Fig. 12 are smaller by about 0.3 ppm with respect to a former paper.267

110

H. P F E m R AND H. ERNST

.4t E

4-

Q P

*=

3-

2-

't

i-PrOH

0

1550

-AG&

1500

1450

1400

/ kJ rno1-l

Fig. 11. Experimental values for the chemical shift S, (ref. 263) and for the standard Gibbs free energy of deprotonation AGODp of the OH group of various molecules in the gaseous state.

4.4

'4.3

H-Y

OH-2

OH-Z

0HwM

H-M

-4.2 E

v, -4.1

I

1

4.0

nFig. 12. Values for the intermediate electronegativity S,, and the chemical shift S, of the accessible bridging OH groups (i.e. those OH groups which give rise to line b, corresponding to the HF-band in IR spectroscopy) in dependence on the silicon-toaluminium ratio ~tfor various zeolites.268

111

NMR STUDIES OF ZEOLITES

t3 5 30 -

E

Q Q \

bx

25 /

/

20 r

I

Hi0

/

+

I

I

I

I

800

1000

1200

1400

I

1600

A E , , ~ / kJ rnol‘’

I

1800 c

Fig. 13. Quantum chemically calculated values for the chemical shielding aH and for the deprotonation energy AEDp of various OH groups.262

(4) Quantum chemical calculations. In Fig. 13 the results of non-empirical quantum chemical calculations for the shielding constant oH and the deprotonation energy AEDPare plotted for OH groups of various species.262 In agreement with the qualitative argument presented above, the chemical shielding increases with increasing values of the deprotonation energy. Summarized one can state that all four arguments are in favour of a direct correlation between the chemical shift and the strength of acidity of OH groups. However, it is necessary to add two comments: (1) The OH groups must be isolated which requires the absence of

adsorbed molecules (all measurements have to be performed in vacuum: gas phase acidity) and of additional electrostatic interactions with the framework as caused, for example, by formation of hydrogen bonds to neighbouring oxygen atoms. The latter interaction leads to different values of the chemical shift for the OH groups pointing into large and small cavities of zeolites (lines b and c, see below). (2) The OH groups must be of similar type as e.g. those compared in Figs 11-13. In a recent paper however, Sauer et ~ 1 were. able~to show ~ that the direct correlation between 6~ and S, is valid for all sorts of surface OH groups of zeolites and related catalysts.

~

112

H.PFELFER AND H. ERNST

- 1

6,l PPm

I

I

I

7

4.5

2

I

Fig. 14. Enhancement of resolution for a typical zeolite catalyst (90 H-Y 300 SB) by spinning of the evacuated powder samples which are contained in sealed glass ampoules about the magic angle. Top: usual 'H NMR spectrum. Bottom: 'H MAS-NMR spectrum with a spinning rate of 2.5 kHz. 270 MHz resonance frequency, room temperature."'

Hence, however, the chemical shift of the 'H MAS-NMR signal can be used as a reliable and sensitive measure for the strength of acidity S, of isolated Bronsted acid. The high sensitivity follows from an inspection of Figs 11 and 13, which yield a slope of ca. 35-100Id/mol ppm, and from the fact that the position of the 'H MAS-NMR signals can be determined with an accuracy of better than 0.1ppm. Hence, it should be possible to measure a change of the deprotonation energy of as low as ca. 5 kJ/mol. The first highly resolved 'H NMR spectrum of a catalyst which could be achieved by a fast rotation of the sample about the so-called magic angle ('H MAS-NMR technique) is shown in Fig. 14. The problem in these experiments is the fast rotation of evacuated powder samples which requires a high degree of axial symmetry of the sealed glass ampoules. In the 'H MAS-NMR spectra of evacuated zeolites containing only oxygen, silicon and aluminium in the framework, five lines can be separated in general which have been denoted266,269.272 as lines a, b, c, d, and e.

NMR STUDIES OF ZEOLITES

113

Line a which appears in the interval between ca. 1.8 and 2.3ppm is caused by non-acidic (silanol) groups. In contrast to an older ~ a p e ? ~ it' seems necessary to state that a distinction between single (SiOH) and geminal (Si(OH)2) hydroxyl groups is not possible by this technique since the difference between the corresponding values of S, is less than 0.1 ppm.268In 29SiMAS-NMR spectra, however, well-separated signals due to single and geminal silanols appear at about -100ppm and about -90 ppm, re~pectively.~~ Unfortunately, the possibility of confirming these two species is limited to adsorbents built up only by silicon and oxygen since the insertion of other metal atoms, as e.g. aluminium, leads to overlapping signals in the 29SiMAS-NMR spectra. Line b at 3 . 8 4 . 4 ppm is ascribed to acidic OH groups which are known to be of bridging type (SiOHAl). The value of S, increases with increasing silicon-to-aluminium ratio of the zeolite (cf. Fig. 12). Line c at 4.8-5.6 ppm is also ascribed to acidic OH groups of the bridging type but under the influence of an additional electrostatic interaction, as e.g. in the case of formation of hydrogen bonds to neighbouring oxygen atoms ("non-isolated" OH groups). The same effect has been found for the stretching vibration of OH groups giving rise to the so-called LF band (3540 cm-I) in addition to the H F band (3650 cm-1).271 In a former ~ a p e ? ' ~line c was ascribed tentatively to OH groups located in the large cavities which are known to be responsible for the HF band in the IR spectrum. In a later paper,267 however, it could be shown unambiguously by the use of deuterated pyridine as a probe molecule that the correlation line b line c

f,

HF band

LF band

is the correct one. The fact that for the bridging OH groups pointing into the large and into the small cavities separate lines appear in the 'H MAS-NMR spectra (lines b and c, respectively) excludes the possibility of a fast proton exchange among the four oxygens around an aluminium atom of the zeolite framework. Line d at 6.5-7.0 ppm is due to residual ammonium ions. Line e at 2.5-3.6 ppm represents hydroxyl groups associated with extraframework aluminium species. Due to the limited space available for these OH groups their S, value will be affected by additional electrostatic interactions. Accordingly, for isolated AlOH groups the chemical shift is much smaller with values in the interval between -0.5 and +l ppm. In Fig. 15 measured values for the 'H NMR chemical shift of isolated (black) and interacting (hatched area) OH groups appearing in zeolite catalysts of various type and composition are presented.274

114 AlOH

Y

CaOH

e

I

MgOH

I

BOH

- a

SiOH POH

SiOHAl

m c

-

&,I

-

H. PFEIFER AND H.ERNST

I

PPm 5.0

m

b

I

I

I

I

I

I

4.0

3.0

2.0

1.0

0

-1 .o

Fig. 15. Measured values for the 'H NMR chemical shift of isolated (black) and interacting (hatched area) OH groups observed in zeolite catalysts.274

As has been shown by a systematic study of the influence of the susceptibility of the zeolite crystallites upon the measured value of the 'H NMR signal, errors of the order of 0.5-1.0ppm must be taken into account.275Hence, the use of an inner standard is strongly recommended if one wants to determine absolute values for the 'H NMR shift of OH groups in zeolites. In this respect, methane has been proved to be suitable and convenient to handle.'% The ultimate resolution of the 'H MAS-NMR spectra of zeolites is a problem of basic interest since the line widths of the signals apparently determine the lower limit of the strength of acidity which can be measured by this method. Line-broadening mechanisms which must be considered are:276 magnetic field inhomogeneity, misadjustment of the magic angle, anisotropy of the magnetic susceptibility, influence of the homo- and heteronuclear magnetic dipole interaction and of the anisotropy of the chemical shift, heteronuclear magnetic dipole interaction with quadrupolar nuclei (27Al)and thermal motion. The dominating mechanisms are the magnetic dipole interaction with protons of neighbouring O H groups (homonuclear magnetic intera~tion)'~~ and with aluminium nuclei. The former line broadening mechanism can be reduced by partial deuteration of the hydroxyl see e.g. Fig. 16, or by application of combined rotation and multipulse sequences (CRAMPS te~hnique)~'or by an enhancement of the spinning rate. An example for the latter possibility is shown in Fig. 17. The line broadening effect of the magnetic dipole interaction of the protons with quadrupolar nuclei (neighbouring 27Al)

NMR STUDIES OF ZEOLITES

10

115

1

5

0

6,l PPm

Fig. 16. 'H MAS-NMR spectra of a zeolite H-Y with a relatively high concentration of OH groups silicon-to-aluminium ratio of 2.6): (A) non-deuterated, (B) deuterated, measure&77 at room temperature and at a resonance frequency of 300MHz with a spinning rate of 3 kHz. * denotes spinning sidebands.

decreases with increasing resonance frequency.278 The enhancement of resolution which can be achieved by a transition of the resonance frequency from 300 MHz to 500 MHz is also demonstrated in Fig. 17, and one can see that the line width for the signal of the bridging OH groups is reduced to less than 0.5 ppm or 250 Hz. However, it is not clear at present whether this value can be reduced further by an increase of the spinning rate above 11 kHz. In contrast, for highly siliceous zeolites like H-ZSM-5, due to the small concentration of bridging O H groups the ultimate resolution has been achieved at relatively small spinning rates (3 kHz). The observed line width of ca. 0.8ppm for these zeolites cannot be further reduced by the NMR technique, it is the natural line width which is given by the distribution width of the bridging O H groups in zeolites H-ZSMJ. of the strength of With respect to a measurement of the concentration of hydroxyl groups (Bronsted acid sites and non-acidic OH groups) nuclear magnetic resonance

116

H. PFEIFER AND H. ERNST

90H-Y 400 SB

b

I

I

500 MHz, 11.O kHz

500 MHz, 3.4 kHz C

4

10

0

-10

10

0

-10

J,lPPm Fig. 17. 'H MAS-NMR spectra of a zeolite H-Y with a relatively high concentration of OH groups (silicon-to-aluminium ratio of 2.6) measured at resonance frequencies of 300 MHz and 500 MHz and with a spinning rate of ca. 3.0 kHz and 11.0 kHz. * denotes spinning sidebands.

spectroscopy has an extremely important advantage compared with infrared spectroscopy since the area of a 'H MAS-NMR signal is directly proportional to the concentration of the hydrogen nuclei contributing to this signal irrespective of their bonding state, so that any compound with a

NMR STUDIES OF ZEOLITES

117

known concentration of hydrogen atoms can be used as a reference (mostly water). In Fig. 18 'H MAS-NMR (Bruker MSL 300) and infrared stretching vibration spectra (Digilab FTS-20) are shown268for two differently synthesized specimens of SAPO-5. While the positions of the various signals in the NMR and IR spectra correspond to each other quite well and are in agreement with IR results published by other authors,279there are dramatic differences in the relative intensities. From the IR spectrum of the first specimen one would erroneously conclude that the concentrations of bridging OH groups of type b and c are approximately equal, and from the IR spectrum of the second specimen that the concentration of POH groups is about three times larger than that of the bridging OH groups of type b. Therefore, even the relative intensity of an OH stretching vibration band cannot be taken as a measure for the concentration of the respective hydroxyl groups in contrast to the intensity of an NMR signal. On the other hand, there is a surprisingly good correlation between the positions of the various signals in the IR and 'H MAS-NMR spectra. Experimental results which were available to uszxoare collected in Table 4 and plotted in Fig. 19. With the exception of the cationic OH groups there is a nearly linear interdependence between the wave number vOH of the IR band and the chemical shift S, of the 'H MAS-NMR signal. The straight line approximating the experimental results is given by:280

-

vOH cm-' = 3870 - 67.8 SH ppm

(10)

Hence, the above-mentioned minimum line width of less than ca. 0.5 ppm for the 'H MAS-NMR signal of bridging OH groups which is tantamount to a resolution of less than ca. 0.1-0.05ppm, corresponds to an equivalent resolution of less than ca. 6.8-3.4cm-' for the OH stretching vibration band. This value has to be compared with the ultimate resolution of 2 cm-' achieved in ref. 281 which leads to the remarkable statement that for both methods the ultimate resolution is of the same order of magnitude. As has been shown above, the resonance position of the 'H MAS-NMR lines, which is given by the isotropic value of the chemical shift, can be used as a reliable and sensitive measure for the strength of acidity is isolated Bronsted acid sites are compared. On the other hand, for not too high spinning rates of the sample, rotational sidebands of the various lines appear which contain information about the anistropy of the chemical shift and the magnetic dipole interaction between the 'H and the neighbouring 27Al nuclei, or in other words, about their distance. In order to analyse these sideband patterns for each of the various signals it is necessary to separate them carefully. This can be accomplished by the following experiment which is a typical example of two-dimensional NMR spectroscopy.282At first a 7r/2

118

H.PFEIFER AND H.ERNST SAPO-5/1

SAPO-5/2 POH

IR

c

b

1

3630 vOH

'H MAS NMR

J cm-'

I C

b

I I POH

H,o

3.8

4.8

1.8

A SiOH

SiOH

6.8

3630

4

6, J P P ~

I

1

1

1

3.7 1.1

1.9 0.2

Fig. 18. IR stretching vibration (Digilab FTS-20) and 'H MAS-NMR s ectra (Bruker MSL 300) of two differently synthesized specimens of SAPOJ.2 8

N M R STUDIES OF ZEOLITES

I 3500

I

I

3600

3700

uo

119

I 3800

/cm-

Fig. 19. Experimental values (cf. Table 4) for the wave number VOH of the IR band the 'H MAS-NMR signal of the stretching vibration and of the chemical shift of OH groups in zeolites.

pulse and then after a time interval tl a T pulse is applied to the sample. Under the condition that the following equation is fulfilled

where vrot denotes the spinning rate, the nuclear magnetic resonance signal is measured at time t2 after the T pulse. Hence, this signal which we denote by F(t,,t,) depends both on tl and 1,. A twofold Fourier transformation of F(t,,t,) then gives the two-dimensional 'H MAS-NMR spectrum S(ol, 0,). A typical example is shown in Fig. 20. The important fact is that S(wl = const., 02) yields the ordinary 'H MAS-NMR spectrum, i.e. the central lines together with all sideband patterns, while S ( q , ~2 = constant) gives the 'H MAS-NMR spectrum without the sidebands. Therefore the sidebands for each line can be analysed separately. In ref. 283 a procedure is described which allows a determination of both the distance rHAl between the 'H and the neighbouring 27Al nuclei and of the chemical-shift anisotropy ACTH for the bridging OH groups (lines b and c in the 'H MAS-NMR spectra) of zeolites. In Table 5 results283for the distance rHAl of bridging OH groups which give rise to lines b and c of zeolites H-Y are collected and compared with theoretical data found by a computer

120

H. PFEIFER AND H. ERNST

Table 4. Experimental values for the chemical shift SH and for the wave number vOH of the various OH groups of zeolites.280 ~

Zeolite H-Y H-ZSM-5 H-ZSMJ (dealuminated) H-SABO SAPO-5

SAPO-11

Mg-Y* * Ca-Y**

OH group SiOH SiOHAl (HF) SiOHAl (LF) SiOH SiOHAl SiOH AlOH (extra framework) SiOHAl SiOH SiOHB SiOHAl SiOH* POH* SiOHAI(1) SiOHAI(I1) SOH* POH* SiOHAI(1) SiOHAl(I1) M~:OHCa OH-

~

~~~~~~~

SH ppm

vOHcm-'

2.0 4.1 5.0 2.1 4.3 2.1 3.0 4.3 2.1 2.5 4.3 1.7'

3742 3636 3543 3734 3604 3738 3654

3.8 4.8 1.9' 3.9 4.9 0.0 0.0

Data point in Fig. 19

3740 3720 3604 3739 3674 3623 3523 3741 3671 3615 3538 3620 3612

*assignment uncertain; **SOH and SiOHAl as in H-Y;' these NMR signals are the sum of two or more components.

~ i m u l a t i o n ~and ~ with results of neutron powder diffraction mea~urements.**~ It can be seen that the values for the distances between the hydrogen and aluminium nuclei derived from the neutron powder diffraction measurements are much smaller and moreover that the ratio of the distances for the two sorts of bridging OH groups is at variance with the result of the two other papers. Presumably this discrepancy is due to the fact that in ref. 285 the AI-0 and Si-0 distances have been assumed to be equal. Table 6 contains experimental results for rHA1 and the chemical-shift anisotropy AuH which have been found by the 'H MAS-NMR spinning sideband analysis283for zeolites of type H-Yand SAPO-5. For the bridging OH groups which point into the small cages (line c) the H-A1 distance has been found to be 2 3 7 k 4 p m for the zeolite H-Y and 2 3 4 k 4 p m for the zeolite SAPO-5. In contrast, for the bridging OH groups in the large cavities the corresponding distances are equal and distinctly larger, viz. 248 f.4 pm. Within the limits of error, the values for the anisotropy of the chemical shift are equal (1922ppm) except for line b of the zeolite SAPO-5 which exhibits a smaller value (14.5 k 2 ppm).

NMR STUDIES OF ZEOLITES

121

Fig. 20. Two-dimensional 'H MAS-NMR spectrum of a zeolite S A P O J . The pulse

sequence applied is shown in the inset. The measurements were performed at room temperature, at a resonance frequency of 300 MHz with a spinning rate of 2 kHz, with 64 consecutive values of tl and 100 accumulations per free induction decay F(tl,t2). The resonance frequencies w1 and are related to that of the reference tetramethylsilane (TMS) by (w1mS - wlb)/wlmS = 3.8 ppm and (wIms - w& alms = 4.8ppm.

Summarizing these results one can state that neither the isotropic chemical shift of the 'H NMR signals nor their anisotropy seems to be simply related to the 1H-27Al distance of the bridging OH groups. It should be the goal of forthcoming quantum chemical calculations to throw some light upon the interdependence between the geometrical (rHAI) and electronic (chemical shift) parameters of the bridging OH groups in zeolites and related catalysts. Sometimes it is claimed that a crucial disadvantage of NMR spectroscopy is its relatively small signal-to-noise ratio. However, with the present state of the art allowing measurements at high resonance frequencies, a linenarrowing by magic angle spinning of the sample and accumulation of

122

H.PFEIFER AND H. ERNST

Table 5. Results for the distance rHAl of bridging OH groups which give rise to lines b and c of zeolites H-Y derived from an analysis of the 'H MAS-NMR spinning sideband pattern283 compared with theoretical data found by a corn uter simulation284and with results of neutron powder diffraction measurements.$5

Line b HF band 01H

Line c LF band 03H

Ref. 283 Ref. 284 Ref. 285

248 k 4 pm 238.6 pm 213.2 pm

237 f4 pm 233.2 pm 219 pm

Ref. 283 Ref. 284 Ref. 285

-

-

169.4 pm 167.7 pm

169.7 pm 165.4 pm ~~

rAlO

-

Ref. 283 Ref. 284 Ref. 285

191.O pm 167.7 pm

~

-

~~

193.0 pm 165.4 pm

Table 6. Experimental resultsB3 for the isotropic value S, of the chemical shift and for the anisotropy AUH, as well as for the distance r ~ A between l the 'H and 27Al nuclei of bridging OH groups in zeolites H-Y and SAPOJ.

'H MAS-NMR signal

Line b

Line c

4.0 f 0.1 ppm 18.3 +_ 1.5ppm 248 k 4 pm

5.0 k 0.1 ppm 20.2 +_ 1.5 ppm 237 +_ 4 pm

3.8 f 0.1 ppm 14.5 -+ 1.5 ppm 248 f4 pm

4.8 +_ 0.1 ppm 19.5 2 1.5 ppm 234 f4 pm

Zeolite H-Y 6 ,

AUH

~HAI

Zeolite SAPOJ 6H

AUH

rHA1

signals, the signal-to-noise ratio seems to be quite sufficient for a study of OH groups in zeolites. For the minimum number of hydrogen nuclei Nmin detectable by 'H MAS-NMR one may write:286

Nmin (

*

(TI * Av/ T,)ln

where T denotes the temperature, vo the resonance frequency, Av the line width, TI the longitudinal relaxation time and T, the time of measurement.

NMR STUDIES OF ZEOLITES

123

The factor of proportionality depends on the quality and the filling factor of the rf coil, the noise of the electronic system and similar other parameters. For a signal-to-noise ratio of 10, a measuring time of 10 minutes, a resonance frequency of 300 MHz, room temperature, a longitudinal relaxation time of ca. 1 second and a line width of 500 Hz which are typical values for 'H MAS-NMR measurements of OH groups in zeolites, Nminis of the order of 10" hydrogen nuclei per sample. Hence, for a sample of 0.2 g zeolite with typically 4 X lo2' cavities per g this value corresponds to 0.01 O H groups per cavity which can be measured by 'H MAS-NMR spectroscopy. In a recent paper287solid-state deuterium NMR has been used to study hydroxyl groups in deuterium-exchanged H-Y zeolites. Through a line-shape analysis it was possible to separate the signals from Bronsted acid sites and non-acidic silanols, and to determine the quadrupolar coupling constant of the former sites as 234 k 2 kHz. However, the resolution of this method is not sufficient to separate the lines from the two sorts of Bronsted acid sites (lines b and c in the 'H MAS-NMR spectra), of residual ammonium ions (line d), and of OH groups at extra-framework aluminium species (line e), so that at present *H NMR may not compete with 'H MAS-NMR in characterizing Bronsted acidity of zeolites and related catalysts. 3.2. Accessibility of Bronsted acid sites, hydrated zeolites

The accessibility of hydroxyl groups can be easily determined through a study of the 'H MAS-NMR spectra after loading the adsorbent with a suitable molecule which, however, must be fully deuterated in order to avoid an unwanted additional 'H NMR signal. Using deuterated pyridine, the concentrations of accessible and non-accessible silanol groups of silica since the formation of a hydrogen bond between could be pyridine and the silanol group shifts the 'H MAS-NMR signal of the latter by ca. 8 ppm to higher values. This effect is demonstrated in Fig. 21A where a zeolite H-ZSMJ with a high concentration of non-acidic silanols (3.7 x lo2' SiOWg) has been loaded with about the same concentration (4 x lo2') of deuterated pyridine molecules. In the case of acidic OH groups (Bronsted acid sites), however, the adsorption of pyridine leads to a protonation, i.e. to a formation of pyridinium ions with a larger shift of ca. 12ppm to higher values. The spectra of a zeolite H-ZSMJ with 4.9 x lo2' Bronsted acid sites per g zeolite and only a negligible concentration of non-acidic silanols (this can be achieved e.g. by a synthesis of the zeolite ZSM-5 without template2") unloaded and after loading with a concentration of 6 X lo2' deuterated pyridine molecules are shown in Fig. 21B. With the same probe molecule it was also possible to show unambiguously that line b in the 'H MAS-NMR spectra of zeolites H-Y is due to bridging OH groups which are easily

124

H. PFEIFER AND H.ERNST

B

A

_sL a

4.1OZ0 Pyr

d

A* SiOHAL I g

Q

6.1Oto Pyr I g

1

I

I

1

10

2

16

4.3 2

6,

I PPm

1

Fig. 21. 'H MAS-NMR spectra of zeolites H-ZSMJ unloaded and after loading with deuterated pyridine: (A) Prevailing concentration of non-acidic OH groups (zeolite synthesized with template289):formation of a hydrogen bond between the OH group and the adsorbed pyridine molecule. (B) Prevailing concentration of Bronsted acid sites (zeolite synthesized without template): formation of pyridinium ions.

accessible by pyridineZ6' and that line c is caused by bridging O H groups pointing into the small cavities (LF band in infrared spectroscopy). In Fig. 22 resultsza are shown for a shallow-bed (400°C) pretreated zeolite SAPO-5 unloaded (A), and after keeping it loaded with deuterated n-hexane for 1hour at 50°C (B). There is no doubt that the 3.9 ppm signal is caused by bridging OH groups which are easily accessible to n-hexane in contrast to the 4.9ppm signal. 'H MAS-NMR studies of hydrated zeolites may be complicated by a superposition of three effects: (1) Hirschler-Plank mechanism, i.e. the adsorption and dissociation of water molecules on extra-framework multivalent cations like Ca2' with a formation of cationic and bridging OH groups.2w The 'H MAS-NMR signal of the former hydroxyls appear at S, values of ca. Oppm and 2.8ppm for calcium ions in the large and small cavities of zeolites Y,re~pectively,'~~ and an inspection of Fig. 19 shows that equation (10) is not fulfilled for those OH groups. Up till now a simple interpretation of this experimental result cannot be given. (2) Formation of hydroxonium ions (chemical shift ah+) at Bronsted acid sites which, however, take part in a fast proton exchange with physically adsorbed water molecules (chemical shift 6,) and with accessible bridging OH groups (chemical shift SOH). Physically adsorbed water molecules may

NMR STUDIES OF ZEOLITES

125

B

4.9 3.9

1.8

4.9

1.8

Fig. 22. ‘H MAS-NMR spectram of a shallow-bed (400°C) pretreated zeolite SAPOJ: (A) unloaded, (B) after keeping the sample loaded with deuterated n-hexane for 1 hour at 50°C.

include water molecules adsorbed on extra-framework cations (adsorption energy 791 kJ/mol, 335 kJ/mol, and 117 kJ/mol for A13+, Mg2+, and Na+, respectively292i293)and water molecules hydrogen bonded to bridging O H groups, to other water molecules and to silanol groups (58.4kJ/mol, 20.1 kJ/mol, and 16.4 kJ/mol, r e s p e c t i ~ e l ~ ~ ~Hence, * ~ ” ) . the resulting ‘H MAS-NMR shift 6 can be described quantitatively by

where cOH,ch+/3, and c, denote the concentration of the ‘H nuclei in the accessible bridging OH groups, in the hydroxonium ions, and in the physically adsorbed water molecules, respectively. With aOH= 4.3 ppm, 6, = 3.2-4.8ppm (the upper limit is for hydrogen bonded water molecules) and &+ = 13ppm2w it is possible to determine quantitatively the concentration of hydroxonium ions from the position 6 of the ‘H MASNMR signal. In shallow-bed (400°C)treated zeolites H-Y the probability of finding a water molecule in the state of a hydroxonium ion is ca. 0.2-0.3 for a rehydration corresponding to one water molecule per accessible bridging OH group (line b).2” (3) Adsorption of water molecules on Lewis acid sites giving rise to a narrow tine at S,= 6.5ppm. In the case of hydrothermally pretreated zeolites H-Y (540°C; 20 h; 4 kPa water vapour pressure) a concentration of 2k0.5 Lewis acid sites of this type per unit cell could be found. Surprisingly, the MAS sideband pattern of the signal at 6.Sppm could be

126

H. PFEIFER AND H. ERNST

explained quantitatively only if these Lewis acid sites are not connected with aluminium. Therefore it was concluded that the signal at 6.5 ppm is caused by water molecules adsorbed on threefold coordinated and positively charged silicon atoms of the zeolite framework,294i.e. on sites which were proposed in order to explain an infrared band at 4035cm-’ for hydrogen adsorbed on a zeolite H-Y activated at 400°C under deep-bed conditions, see refs 295, 296. Further experimental and theoretical work seems necessary to prove whether threefold coordinated and positively charged silicon atoms do exist in the framework of zeolites. Another method to study the formation of hydroxonium ions in hydrated zeolites has been introduced by J. Fraissard et a1.297Through a measurement at very low temperatures (4 K) the fast proton exchange among the various species can be excluded and a line-shape analysis should yield quantitative information about the concentration of the various species. Unfortunately, however, the line widths of the respective ‘H NMR signals are so large that they cannot be distinguished by their chemical shift. Instead it is necessary to make use of the different line shape for the ‘H NMR signal of rigid one-, two-, and three-spin systems. Assuming that there are only isolated hydroxyl groups (one-spin systems), isolated water molecules (two-spin systems), water molecules hydrogen bonded to OH groups (three-spin systems with the configuration of an isosceles triangle) and hydroxonium ions (three-spin systems with the configuration of an equilateral triangle) the shape of the observed ‘H NMR broad line is decomposed into contributions of these species by a fitting procedure where, however, the magnetic dipole interaction among these species must be considered by introducing a further (“line-broadening”) parameter. Since the concentration of hydroxonium ions ch+ thus determined will depend both on the concentration of the acidic OH groups c,, and on their strength of acidity S, the ratio ch+/c, should be a reasonable measure for S,. Apparently this method fails if the zeolite catalyst contains various sorts of OH groups differing in their concentration and strength of acidity. With respect to the ‘H MAS-NMR technique it should be of interest whether the sensitivity of the broad line method is sufficient to confirm, for example, the dependence of the strength of acidity on the electronegativity of zeolite catalysts (Fig. 12).

4. LEWIS ACID SITESEXTRA-FRAMEWORK ALUMINIUM Apart from relatively weak signals in the IR spectra of adsorbed hydrogen295and carbon monoxide296which are ascribed to an interaction of these molecules with threefold aluminium and silicon atoms of the zeolite framework affected by, for example, a hydrothermal treatment, it is generally accepted that the Lewis acidity of zeolites298and related catalysts (silica-alumina, y-alumina) is connected with the presence of aluminium

NMR STUDIES OF ZEOLITES

127

species on the surface. Therefore two possibilities should exist in principle to study the Lewis acidity of zeolites: (1) an analysis of highly resolved 27Al NMR spectra similar to the ‘H

MAS-NMR method described in Section 3, and (2) the use of probe molecules as e.g. ~ y r i d i n or e ~the ~ ~above-mentioned molecules in IR spectroscopy. 4.1.

27AIN M R spectroscopy

In well-crystallized zeolites exhibiting no Lewis acidity, aluminium is tetrahedrally coordinated with an isotropical chemical shift between 55 and 6 5 p ~ m with ~ ~respect to an aqueous dilute solution of A13+. After a hydrothermal treatment and/or dealumination, however, extra-framework aluminium species are formed (Lewis acid sites) and the 27Al MAS-NMR spectrum becomes more complicated. The initial signal at ca. 60ppm changes its intensity due to ejection of aluminium from the framework into the intracrystalline space and due to 27Al signals from tetrahedrally coordinated extra framework aluminium, which must exist at least partly as Al(OH), species as could be shown by ‘H-27Al cross-polarization (CP MAS-NMR) experiment^."^ The other extra framework aluminium species give rise to distinct resonance lines in the interval between -15 and +4 ppm which could be assigned to polymeric aluminium species,3o1and to a line near 30 ppm which is interpreted as being due to either penta~oordinated”~or tetrahedrally c o ~ r d i n a t e d ’ extra ~ ~ ~framework ~ ~ ~ . ~ ~aluminium. ~ In addition, a very broad hump extending from ca. -180ppm to ca. +230ppm (for an external magnetic field strength of 7.1 T) appears below these more distinct resonances. As has been shown recently,lg3this broad hump merges into the lines near 0, 30, and 60 ppm for ultrahigh (518 kHz) MAS speeds. In hydrated samples a narrow line at Oppm is often observed, suggesting the presence of A13+ cations balancing lattice charges. A survey on these experimental results is given in Table 7. 27Alnutation experirnend5 performed on hydrothermally treated zeolites Y suggest that the quadrupolar interaction of the extra framework aluminium species is relatively large corresponding to a value of at least ca. 1MHz for the quadrupole frequency vQ defined by equation (15) (see below). Large values of vQ give rise to a large second-order quadrupolar shift:52

where vcG and vL denote the centre of gravity of the signal and the Larmor frequency (signal for vQ = 0) in Hz, respectively, Z is the spin quantum

128

H. PFEIFER AND H. ERNST

Table 7. Experimental results for the chemical shift 8, observed on zeolites.

of 27Al NMR signals

Species 0 4..

Octah. coord. extra framework aluminium (A13+)53

. -15

Extra framework polymeric aluminium speciesM’

~

~

~~

30

Penta-~oordinated”~ extra framework aluminium

30 (50?,ref. 159)

Tetrahedrally coordinated extra framework aluminium: AlOOH associated with two framework oxygenslW aluminium in amorphous sili~a-alumina”~

-~~~~~~

-180. . . +230 (Bo = 7.1 T)

Extra framework aluminium of low symmetry (“NMR invisible”). The very broad hump merges into the lines at 0, 30, and 60 ppm for ultrahigh spinning speeds ( a 1 8 kHz)ls3

60

Framework aluminium183and tetrah. coord. extra framework aluminium (CP MAS-NMR183)

number ( I = 512 for 27Al), r) the asymmetry parameter ( 0 s 7 s 1) and uQ the quadrupole frequency VQ

= 3e2qQ/(21(2Z- 1)h)

(15)

with eq, eQ and h denoting the electric field gradient, the electric quadrupole moment and Planck’s constant, respectively. In addition to this effect which must be taken into consideration if one wants to determine the true chemical shift of a ”A1 NMR signal, a strong quadrupole interaction leads to a dramatic broadening of the lines. The line widths in ppm for a static sample (A), for a sample under conditions of magic angle spinning (AMAS) and under conditions of double rotation (ADoR) are given by:40,303

A = (u ~ u* (25 ~ +) 229 ~ + q2)* 106/18 AMAS = h13.64 if ur 3 ps ADOR

= Ac

if ui = 5 . ~ *OR ~ 5

where u, is the spinning rate of the rotor in the case of MAS, and q and vo denote the corresponding rates for the inner and outer rotor of DOR, respectively. The critical values for the rates are given by

NMR STUDIES OF ZEOLITES

129

Table 8. Values for the nuclear magnetic resonance frequency % of *'Al, for the line and width AMAs under magic angle conditions and for the critical spinning rates &'OR of MAS and DOR experiments, respectively, if the signals of extra-framework condensed (Al-A1 distance ca. 300pm) aluminium species of low symmetry (YQ = 3 MHz) should be resolved.

eAS

7.05

11.7

78.2 130.3

562 202

25 25

44 26 ~~

~

Bo denotes the strength of the external magnetic field.

Ac denotes the line width in ppm due to a distribution of the chemical and/or quadrupole shift and AD the line width in ppm due to homonuclear magnetic dipole interaction. It is the influence of these two latter quantities which determines the ultimate resolution of 27Al MAS and of 27Al DOR-NMR spectra. In Table 8 numerical data are given for vQ = 3 MHz (aluminium species of low symmetry), 11 = 0 and for an assumed mean distance of ca. 300pm between neighbouring 27Al nuclei in condensed extra framework species. The latter value corresponds to AD = 5 kHz. An inspection of this table shows that the conditions of equations (17) and (18) cannot be fulfilled experimentally so that we must state that neither MAS nor DOR NMR spectroscopy of 27Al will be able to detect extra framework condensed aluminium species of low symmetry. 4.2. Use of probe molecules

In analogy to infrared spectroscopy, probe molecules may be used to study Lewis acid sites. 15NCP MAS-NMR spectra of adsorbed pyridine have been studied by Maciel et aL3" (pyridine on silica-alumina), Ripmeester305 (pyridine on y-alumina, mordenite), and by Majors and Ellis3& (pyridine on y-alumina). Values for the resonance shifts relative to solid pyridine are collected in Table 9. The major drawbacks of these experiments are that (1) an absolute determination of concentrations is connected with large errors due to the strong dependence of the line intensities on quantities controlling the efficiency of cross-polarization, see e.g. ref. 307, and (2) often, caused by molecular exchange, some or all of the lines merge into a single averaged line so that an unambiguous separation into its components, namely the concentrations of physisorbed pyridine, pyridinium ions and pyridine adsorbed on Lewis acid sites is not possible. Compared with pyridine, phosphines are roughly three orders of magnitude stronger bases. 31P CP MAS-NMR spectra of trimethylphosphine

130

H.PFEIFER AND H.ERNST

Table 9. Values for the "N NMR shift of pyridine-* in the liquid state, physisorbed, and adsorbed on Bronsted and Lewis acid sites relative to the resonance of solid pyridine. ~~

~

s,PPm

Species Solid pyridine (- 105OC) Liquid pyridine Physisorbed pyridine Pyridinium ion (Bronsted acid sites) Pyridine on tetrah. A13' (Lewis) Pyridine on octah. A13+(Lewis)

0 -26 -10 f 10 88f2 22.5 f 3.5 46.5 k 7.7

Table 10. Values for the 31P NMR shift of trimethylphosphine physisorbed and adsorbed on Bronsted and Lewis acid sites, relative to the liquid state.3087309

4 PPm

Species ~~

Liquid trimethylphosphine Physisorbed trimethylphosphine Protonated trimethylphosphine (Bronsted acid sites) Trimethylphosphine on Lewis acid sites

0 0.7 f 6

59f2 12 f 10

adsorbed on zeolite H-Y and on y-alumina were investigated by Lunsford et al. ,308 and of various trialkylphospines adsorbed on silica-alumina and on y-alumina by Maciel et al.309 Values for the resonance shift relative to liquid trimethylphosphine are shown in Table 10. For this probe molecule the same drawbacks hold as mentioned above for pyridine, although due to the larger signal-to-noise ratio of the 31PNMR signal compared with I5N, cross-polarization must not be applied. The sensitivity of trimethylphosphine to distinguish between different sorts of Lewis acid sites seems to be less than that of pyridine since for y-alumina only one sort of Lewis acid sites could be found (see Table 10). In principle also water can be used as a probe molecule since the chemical shift of water molecules adsorbed on Lewis acid sites (6.5 ppm, see ref. 294) is different from that of physisorbed water (less than 4.8 ppm) and of water molecules adsorbed on Bronsted acid sites (hydroxonium ions: 13 ppm). As has been shown in section 3.2, however, the 'H MAS-NMR spectrum of hydrated zeolites is complicated by a superposition of various effects so that water is not very suitable to probe Lewis acidity of zeolites and related catalysts. Another way to study quantitatively Lewis acidity results from the fact that probe molecules could be found for which the resonance shift caused by Lewis acid sites is much larger than that caused by Bronsted acid sites or physisorption. Typical examples are carbon monoxide31c312 and dinitrogen

NMR STUDIES OF ZEOLITES

131

Table 11. Values for the 13C NMR shift of carbon monoxide physisorbed and

adsorbed on Bronsted and Lewis acid sites, relative to the gaseous state.31s312 Sc PPm

Species ~~~

~

Carbon monoxide gas Physisorbed carbon monoxide Carbon monoxide on Bronsted acid sites Carbon monoxide on Lewis acid sites

~

~~~

0 -3+2 -3f2

300 . . . 400 (ref. 310, 311) 590 f 60 (ref. 312)

Table 12. Values for the "N NMR shift of dinitrogen oxide physisorbed and

adsorbed on Bronsted and Lewis acid sites relative to the gaseous state.313

Species Dinitrogen oxide gas Physisorbed dinitrogen oxide Dinitrogen oxide on Bronsted acid sites Dinitrogen oxide on Lewis acid sites

s,PPm 0 4.5 k 3 4.5 k 3 50 k 20

oxide.313 Since at room temperature and above these molecules exchange rapidly between the various adsorption sites, only a single averaged line for the I3C NMR of carbon monoxide and for the 15N NMR of the terminal nitrogen of dinitrogen oxide appears, and it becomes necessary to measure the position of this line as a function of loading to determine the resonance shift of the molecules adsorbed on the Lewis acid sites and their concentration. The procedure developed by Borovkov and described in detail in ref. 275 makes use of an extrapolation of the resonance shift of the single averaged line to zero loading which is connected with relatively large errors due to the decreasing signal-to-noise ratio. Results for the resonance shift of the probe molecules carbon monoxide and dinitrogen oxide physisorbed and adsorbed on Bronsted and Lewis acid sites are collected in Tables 11 and 12, respectively. At present the accuracy is far from the goal to take the values of the resonance shift for the molecules adsorbed on Lewis acid sites as a measure for their strength of acidity. Assuming an average value for this resonance shift, however, the method seems suitable to determine the concentration of Lewis acid sites. Summarizing the results obtained up till now with probe molecules, the following statements can be made: (1) At room temperature and above no separate lines for water, carbon monoxide and dinitrogen oxide adsorbed on Lewis acid sites can be observed due to fast molecular (or proton) exchange. Hence, a measurement of coverage dependence becomes necessary, yielding

132

H. PFEIFER AND H. ERNST

large errors for the chemical shift (strength of acidity) and for the concentration of the Lewis acid sites. (2) In those cases where due to a poor signal-to-noise ratio crosspolarization has been applied (pyridine, trimethylphosphine) only a semiquantitative determination of the concentration of Lewis acid sites is possible. (3) Steric effects connected with larger probe molecules may lead to large errors. As an example the result presented in a paper by Lunsford et d 3 1 4 shall be mentioned: with trimethylphosphine as a probe molecule, 37 protonated species per unit cell have been found for a zeolite H-Y in contrast to the real value of 54 Bronsted acid sites (framework aluminium atoms) per unit cell.

5. STRUCTURE OF ADSORBED MOLECULES 5.1. High-resolution NMR of adsorbed molecules-general aspects

The surface properties of zeolites have received much attention since the pioneering work of B a t ~ e r . ~The ’ ~ application of high-resolution NMR to study adsorbed molecules is of interest from a standpoint both of basic and applied re~earch.~’,~’~ The systems under study are closely related to such important processes in chemical industry as selective adsorption or heterogeneous catalysis. In general, the mobility of adsorbed molecules is intermediate between the solid and the liquid state, which gives rise to a more or less strong broadening of the NMR signals. In those cases where a sufficient resolution can be achieved, an analysis of the spectra may yield valuable information about the existence of adsorption sites and their interaction with adsorbed molecules. In this connection the influence of other interactions, as, for example, the van der Waals interaction between the molecules and the surface or intermolecular interactions, susceptibility effects and exchange processes of molecules between different adsorption sites must be taken into con~ideration.~~*~~~ In 1972 the first highly resolved ‘’C N M R spectra of adsorbed hydrocarbon molecules could be m e a ~ u r e d . ~ ~As ~ ~ a’ ”typical example, the study18 of adsorption complexes formed between but-1-ene and exchangeable cations in zeolites X and Y shall be mentioned. The carbon atoms in this molecule are denoted as follows:

133

NMR STUDIES OF ZEOLITES

Table 13. Experimental values for the difference ($9- ~ 9 ) o~ f the ' 13C chemical shielding in ppm of but-1-ene adsorbed on zeolites TI-X and NaAg-X (60% Ag+) and dissolved in AgN03 H20.

+

c1

c2

c3

c4

0 (113.5)

0 (140.5)

0 (27.4)

0 (13.4)

NaAg-X

-1.1 11.5

-1.6 -1.5

0 -0.6

0 -0.6

Dissolved in AgN03 + H2O

12.2

Neat liquid ~~

~

TI-X

0.9

0.5

0.2

Values in parenthesis are referred to Th4S. The loading of the zeolites corresponds to ca. 4 molecules per large cavity."'

and in Table 13 experimental values for the difference of the chemical shielding in the neat liquid and for the adsorption complexes are given. For comparison also values are included for the complex which is formed in an aqueous solution of silver nitrate. From these results it must be concluded that in NaAg-X zeolites silverlbut-1-ene complexes are formed of the same structure as in solution while in TI-X zeolites the but-1-ene is differently bonded to the thallium ion. By use of the MAS (magic angle spinning) technique a significant enhancement of the resolution and hence of the sensitivity can be achieved. In the first proton MAS-NMR experiments, performed on ethanol adsorbed on diatomaceous earth, the static line width could be reduced by more than two orders of r n ag n i t ~ d e . ~This ' ~ reduction can be ascribed to the fact that the broadening caused by the magnetic inhomogeneity of the adsorbateadsorbent system and by the anisotropy of the motion of the adsorbed molecules is controlled by the same angular factor as in the case of magnetic dipole interaction in solids. Furthermore, in a recent paper320 it could be shown both experimentally and theoretically that the line width of the 'H MAS-NMR signals approximates zero if (1) the molecule performs rotational motion only around axes fixed in space and (2) the intermolecular homonuclear dipole interaction can be neglected.

5.2. I3C and 'H MAS-NMR studies on sealed samples MAS-NMR experiments (cf. for example refs 321-325) performed on sealed samples containing zeolites loaded with different hydrocarbons have shown that the enhancement in spectral resolution brought about by the use of MAS allows a detailed assignment of NMR spectral lines to individual molecular species including those which are strongly bound to the surface.

134

H. PFEIFER AND H. ERNST

- CH,

‘c = o

0

I

230

I

.

I

.

I

I

220

.

PPH

I

.

I

21 0

I

. ” {

’

l . . . . I . . , . I . , . ,

40

30

PPH

20

Fig. 23. I3C NMR spectra (resonance frequency 74.5 MHz) of acetone molecules adsorbed on a Na-X zeolite pretreated at 400°C with a loading of 4 molecules per large cavity: (a) static NMR spectrum; (b) MAS-NMR spectrum with proton decoupling (spinning rate 2 W ) ; (c) MAS-NMR spectrum without proton decoupling.330

N M R STUDIES OF ZEOLITES

135

I

0 Na'

Fig. 24. Geometry of possible arrangements between an acetone molecule and a sodium ion.32g

As an example in Fig. 23 the I3C static NMR spectrum (a) as well as the 13C MAS-NMR spectra (b, c) with (b) and without (c) proton decoupling of acetone adsorbed on a Na-X zeolite is shown for a resonance frequency of 74.5 MHz. The values of the chemical shifts of the 13C MAS-NMR signals could be determined with an error of less than 0.1 to be 29.0ppm for the CH3 groups and 213.5ppm for the CO group compared with 30.6 and 206.0 ppm in the liquid state, respectively. In order to understand the formation of adsorption complexes between an acetone molecule and an adsorption site in the large cavities of the zeolites semiempirical and ab initio quantum chemical calculations have been carried o ~ t . In~ Fig. ~ ~24 , two ~ possible ~ ~ arrangements between an acetone molecule and a sodium ion are shown. The result of ab initio calculations of the energy as well as of the chemical shifts lead to the conclusion that position I has to be preferred where the sodium ion interacts with the lone pair electrons of the oxygen atoms.328 Anderson et aZ.'90,329have used 'H MAS-NMR spectroscopy to study the adsorption of methanol on zeolites H-ZSM-5, Na-ZSM-5 and several Si,AIand Si,AI,P-based zeolites. Due to the relatively high resolution of the spectra (cf. Fig. 25) and a comparison of the 'H MAS-NMR results after loading the zeolites with CH30H, CH30D, CD30H, and CD30D under identical conditions, the authors were able to conclude that in the case of acidic zeolites at low coverages each methanol molecule is bonded to a Briinsted acid site and forms a methoxonium ion (CH30H;, cf. Fig. 26a). At high coverages, however, charged clusters of methanol molecules associated with each Bronsted acid site are formed (cf. Fig. 26b). In the case of non-acidic zeolites (Na-ZSM-5) methanol is coordinatively adsorbed on the cation.

136

H. PFEIFER AND H. ERNST

4.1 ppm

(b) H.Y

5.1 ppm

\

9.1 ppm

~

15

-~

10

5

-0

-5

Fig. 25. 'H MAS-NhfR spectra (resonance frequency 400 MHz, spinning rate up to 3 kHz) of methanol molecules adsorbed on zeolites: (a) Na-ZSM-5, (b) H-Y, (c) H-ZSM-5. The loading corresponds to six molecules of methanol per Bronsted acid

site .329

N M R STUDIES OF ZEOLITES

SI

At

137

Si

Fig. 26. Structures for methanol molecules hydrogen bonded to a Bronsted acid site:329(a) one methanol molecule forms a methoxonium ion (CH,OH,+); (b) at high coverages protonated clusters are formed.

The advantage of magic angle spinning is clearly demonstrated by Figs 27 and 28, where static 13C and 'H NMR43 as well as MAS-NMR330spectra are compared which were taken on the same sample of but-1-ene molecules (in the case of Fig. 28 I3C enriched in the =CH2 position) adsorbed on Na-Y zeolites. Besides the much higher accuracy of the 13C MAS-NMR measurements, the liquid-like high-resolution 'H MAS-NMR spectrum confirms the 13C: NMR results for b ~ t - l - e n eand ~ ~ other olefinic hydrocarbons331 adsorbed on Na-Y zeolites: with respect to the liquid, the change of the 13C chemical shift (4.1 ppm) as well as of the 'H chemical shift (0.38 ppm) of the =CH- group (to lower magnetic field) is much higher than the corresponding values for the other C and H atoms in the but-1-ene molecule.

138

H. PFEIFER AND H. ERNST

ZCH-

CD ,,

=CH, 1

Fig. 27. Static NMR spectra of but-1-ene molecules adsorbed on a Na-Y zeolite43 with a loading of 3.3 molecules per large cavity: (a) 13C NMR spectrum (resonance frequency 22.6 MHz); (b) 'H NMR spectrum (resonance frequency 90.0 MHz).

In Table 14 recent results of 13C MAS-NMR studies on adsorbateadsorbent systems are given. With regard to results of 13C NMR measurements the reader is referred to ref. 43. 6. MOLECULAR DIFFUSION

6.1. Basic principles Since in general zeolites are only available as crystallites with a size of the order of a few micrometers, measurements of molecular diffusion have to be carried out with assemblages of zeolite crystallites (powder samples). The conventional way of measuring molecular diffusion in such systems is to follow the rate of mass change of the sample after changing the pressure of the surrounding a t m o ~ p h e r eHowever, .~~ the interpretation of such sorption or uptake experiments is not always unambiguous (cf. below) and allows only an indirect determination of the intracrystalline self-diffusion coefficients. A completely different way to study molecular migration is provided by the pulsed field gradient NMR ~ p e c t r o s c o (PFG-NMR). p~~~~

N M R STUDIES OF ZEOLITES

139

Although this method is generally combined with the formation of a so-called spin echo in order to eliminate the influence of an inhomogeneity of the constant magnetic field Bo, its essential features can be explained by a consideration of the free induction decay, i.e. the NMR signal following a short intense rf pulse. During the decay time of this signal ( t > O ) , two succeeding magnetic field gradient pulses of different sign are applied so that the intensity of the external magnetic field is given by

where g denotes the intensity and S the duration of the magnetic field gradient pulses. Assuming that S is much smaller than the time interval A between the two magnetic field gradient pulses, the ratio J, of the amplitudes of the free induction decay at time t = 28 + A and t = 0 can be shown to be :348

Here T2 denotes the transverse nuclear magnetic relaxation time of the adsorbed molecules and y the magnetogyric ratio of the resonating nuclei. p(z’ - z , A) is the mean propagator given by

p(z’-z,A) =

I

p(z)*p(z,z’,A)*dz

with the a priori probability density p ( z ) to find a molecule at z , and the conditional probability density, the ‘propagator’, p ( z , z ’ ,A) to find a molecule at time A at z f if it has been at z at time 0. Assuming T2

>> A

(24)

the first factor in equation (22) can be neglected and the thus simplified equation offers two different possibilities to evaluate the experimentally determined data, namely the ratio J, as a function of g or of S: (1) By a Fourier transformation of J, the mean propagator can be directly determined:

I

p ( ~-‘ 2 , A ) = ( 2 ~ ) ~I,/.J*cos[(z~ ’ -~)~gS]*d(ygS)

(25)

H. PFEIFER AND H. ERNST

140

=CH2

115

O

.

I

140

.

i

.

120

*

.

100

I

.

80

I

110

.

60

’

.

40

1

.

1

20

Sc/PPm

Fig. 28. MAS-NMR spectra of but-1-ene molecules adsorbed on a Na-Y zeolite with a loading of 3 molecules per large cavity (unpublished): (a) 13CMAS-NMR spectrum with proton decoupling (resonance frequency 74.5 M H z , spinning rate 3 kHz); \b) ”C MAS-NMR spectrum of the =CH* group without proton decoupling; (c) H MAS-NMR spectrum (resonance frequency 300 M H Z ) .

(2) The initial slope of a plot of In+ as a function of (yg8)’ yields the mean square displacement ((2’ - 2)’) of the adsorbed molecules: (dln $/d(ygS)2),=o = ((2’ - 2)’)/2

(26)

Since for an unrestricted molecular diffusion the following equation holds:

where D denotes the self-diffusion coefficient, we introduce for the general case an apparent self-diffusion coefficient Dappby the equation

Dapp= ( ( 2 ‘ - z ) ~ ) / ~ = A (dln +/d(ygti)’),,,/A

(28)

141

NMR STUDIES OF ZEOLITES

-CY

-CH2I

=CH, 1

=CH4

L 1

.

1

7.0

.

l

6.0

.

~

5.0

.

4.0

~

.

3.0

l

.

8.0

l

.

l

1.0

.

,

0.0

.

l

.

,

-1.0

6, I ppm

Fig. S o n t d .

The limits of the PFG-NMR method are determined by the magnetogyric ratio y of the resonating nuclei and the maximum achievable values of g, A, and S for which we assume

D,,,

= T2; ,,g

20 T/m; S,,

=

s

(29)

where in principle the value of a,,, can be enhanced further as long as the condition S,, << A,, is fulfilled. Inserting equation (27) into equation (26) it can be easily shown that the minimum value of the self-diffusion coefficient Dminwhich is accessible by the PFG-NMR method follows from the condition

With the value of the magnetogyric ratio of the proton yprot= 2.675 x 10' T-l s-' and the maximum values given by equations (29), this condition may be rewritten as

D ~ ,=,3 X

(yp,ot/y)2/T2m2/s

(31)

142

H. PFEIFER AND H. ERNST

Table 14. Results of 13CMAS-NMR measurements of the chemical shift for various adsorbate-adsorbent systems.

Molecule

Zeolite

13Cchemical shift (ppm)

Temperature, if different from r.t.; remarks Gaseous CH4in the sample

Ref.

CH4 (-2.3g)

H-ZSMJ

5 0

332 332

co

H-ZSM-5

184

321, 333

H-ZSMJ

126

333

Na-X H-ZSMJ

49.7 50.8 50.0 51 50 49.6 49.7 49.8 50 50

343 190, 321 344 333 334 323 323 323 190, 335, 336 69, 190

(182. l g )

co2 (125.7g) CH3OH (49.9')

Na-ZSM-5 Na-Mor 20HNa-Mor SAPO-5 SAPO-34 CH30CH3 (59.7')

Various SilAl ratios 173 K

Na-X H-ZSM-5

59.8 60.0 60.5 62 60

Na-ZSM-5 Na-Mor SAPOJ

60 59.6 60

CHZ=CHZ (122.8')

H-ZSMd

122.8 120.1

CH=CH

H-Y

ca. 70

342

1: 112.7 2: 147.5 3: 27.5 4: 12.8 1: 112.9 2: 144.6 3: 26.5 4: 12.5

330

1 2 3 4 CH2=CHCH2CH3Na-X (1: 113.5', 2: 140.51, 3: 27.4', 4: 13.4') Na-Y

173 K 173 K

223 K Free ethene in the sample

343 344 190,321 333 334 334 323 190 339 339

330

N M R STUDIES OF ZEOLITES

143

Table 14-contd. Molecule

1 2 (CH&C=O (1: 30.6' 2: 206.d)

Zeolite

Na-X H-Y H-ZSM-5

CH3NH2 (28.0')

H-ZSM-5

13C chemical shift (ppm)

Temperature, if different from r.t.; remarks

1: 29.2 2: 214.2 1: 29.7 2: 217 1: 30 2: 210 1: 29 2: 213.5 26.4 26.1 31

Ref.

341 340 340 34 1 Only few molecules are adsorbed Gaseous amine in the sample

324 324

Na-ZSM-5 Na-Mor 20HNa-Mor H-Rho

27.0 26.6 26.0 24.4

324 323 323 323 345

(CH3)2NH (38.2 )

H-ZSM-5 Na-ZSM-5 Na-Mor 20HNa-Mor H-Rho

36.0 35.8 37.2 32.9 34.1

324 323 323 323 345

(Cq3N (47.1 )

H-ZSMJ Na-ZSM-5 Na-Mor 2OHNa-Mor H-Rho

45.9 46.2 46.7 44.2 46.4

324 323 323 323 345

In parenthesis values of the chemical shift for the gaseous (g) or liquid (1) phase are given. All values are referred to TMS.

with T2 in seconds. For hydrocarbons adsorbed in zeolites Na-X, the transverse nuclear ('H) magnetic relaxation time T2 is of the order of lO-'s, so that the minimum value of the intracrystalline self-diffusion coefficient m2/s. Similarly from the condition which can be measured is ca. 3 X

144

H. PFEJFER AND H. ERNST

Table 15. Comparison of the intracrystalline self-diffusion coefficients D obtained by the PFG-NMR method with the corresponding values derived from classical uptake e~perirnents.~’~

PFG-NMR MethaneNaCa-A (23°C) EthaneNaCa-A (23°C) PropaneNaCa-A (23°C) n-HeptaneNa-X (164°C) c-HexaneNa-X (164°C) Benzenema-X (164°C)

2x 2x 5x sx 4x 2x

10-9

Classical uptake

s x 10-14

10-’0 10-12

1 x 10-14 3 x lo-’’

10-9 10-9 10-10

3 x 10-13 4 x 10-13 1 x 10-14

which is completely equivalent to equation (30), the minimum value of the r.m.s. displacements follows as

i.e. a value of 0.3 pm for the ‘H PFG-NMR method. In the first applications of the PFG-NMR method to molecules adsorbed on zeolites (see e.g. the review articles3497350 and references cited therein) two surprising results have been found: (1) The apparent self-diffusion coefficient exhibits a plateau if D,, is plotted as a function of the reciprocal value 1/T of the absolute temperature (cf. Fig. 29) and (2) the values of Dapp are larger by up to five orders of magnitude than the values determined by conventional sorption experiments (cf . Table 15). The interpretation of the temperature dependence which has been verified by numerous additional experimentsM9 is straightforward: at low temperatures, i.e. below the plateau, due to their low translational mobility, the majority of adsorbed molecules do not encounter the external surfaces of the crystallites so that D,, coincides with the intracrystalline self-diffusion coefficient D. In the temperature interval of the plateau the majority of molecules still remains adsorbed in the intracrystalline space so that their translational diffusion is confined to the size of the crystallite. Consequently the mean square displacement and hence D,, of the adsorbed molecules are determined by the dimension of the crystallites. If the crystallites can be approximated by spheres of radius R, the apparent self-diffusion coefficient is given by the relation353

Dapp= R2/5A

(34)

For higher temperatures an increasing number of molecules has left the

NMR STUDIES OF ZEOLITES

1

145

10-9

2

4

6

- 1 0 3 ~ 1 ~ Fig. 29. Temperature dependence of the apparent self-diffusion coefficient D,, of n-butane adsorbed on Na-X zeolite crystallites with a mean diameter of 9pm. h e pore filling factor, i.e. the ratio of the concentration of adsorbed molecules in the intracrystalline pore system to its maximum value, is 0.67 at room temperature. The PFG-NMR measurements were performed with a distance of A = 4 ms between the two magnetic field gradient pulses.35’

crystallites and the apparent self-diffusion coefficient is given by the relation348 Dapp

= Pintere Dinter

(35)

where pinterand Dinter denote the relative number and self-diffusion coefficient respectively, of molecules outside the crystallites, i.e. in the intercrystalline space. According to equation (34) the value of the apparent self-diffusion coefficient of the plateau can be used for an estimation of the mean radius of the crystallites. For the example presented in Fig. 29 Dapp= 1.5 x lo-’ m2/s and A = 4 x loF3s so that a mean radius of ca. 5.5 pm follows which is in satisfactory agreement with the mean radius of about 4.5 pm determined by a microphotographic analysis of the crystallites.

146

H. PFEIFER AND H. ERNST

t /

0

/

10- 13

/'

R Ipm Fig. 30. Intracrystalline self-diffusion coefficient of methane adsorbed on chabazite at 0°C as a function of the mean radius R of the crystallites determined by the PFG-NMR method (true values) and by classical uptake experiments (apparent

values) .352

Results for one of the crucial experiments352by which the self-consistency of the PFG-NMR diffusion data could be proved is shown in Fig. 30. Chabazite, a natural zeolite available as relatively large single crystals, has been ground and sieved so that samples with zeolite crystallites of different diameter could be prepared. As it should be, the intracrystalline selfdiffusion coefficient of methane determined by the PFG-NMR method is independent of the radius of the crystallites. In contrast, an unreasonable radius dependence has been found for the values measured by classical uptake experiments: while for the smaller crystallites the values for the diffusivity determined classically are nearly four orders of magnitude smaller, they approach the result of the PFG-NMR method with increasing radius of the crystallites. This result and similar observations led to a critical re-examination of the earlier uptake data which revealed that the effects of external heat and mass-transfer resistances in limiting the adsorption rates were often far greater than originally assumed (see e.g. refs 354, 355). Additional arguments in favour of the PFG-NMR data are revealed by

NMR STUDIES OF ZEOLITES

147

0 0

0 0

10-l2

b 0

2

4

6

C,H, per cavity Fig. 31. Intracrystalline self-diffusion coefficient at 185°C for benzene adsorbed on zeolite Na-X as a function of the concentration determined by PFG-NMR ( o ) ? ~ ~ , ~ ~ by incoherent quasielastic neutron scattering (A),”’ and by the zero length column method (o).361The results of uptake measurements with different specimens359are indicated by the hatched areas.

incoherent quasielastic neutron scattering experiment^.^'"^'^ A typical example is shown in Fig. 31. In spite of these results there exist a few systems (e.g. propanehilicalite, xylenesma-X) where it seems unlikely that the remaining discrepancies between the microscopic (PFG-NMR and incoherent quasielastic neutron scattering) and macroscopic (classical uptake, frequency response, zero length column, etc.) methods can be explained by mere shortcomings of the individual techniques and it was suggested362 that a slow equilibration between mobile and immobile intracrystalline states may take place in these systems. 6.2. Intracrystalline diffusion of hydrocarbons

As an example for a systematic study of molecular diffusion in zeolites, the intracrystalline diffusion of methane, ethane and propane adsorbed in Na-X, silicalite and NaCa-A shall be treated.363These zeolites are typical examples of large-, medium-, and small-pore zeolites (see Table 1). Experimental

148

H. PFEIFER AND H. ERNST

results for the intracrystalline self-diffusion coefficient measured at a temperature of 300K are plotted in Fig. 32 as a function of the concentration (coverage of the zeolite) expressed in molecules per large cavity (Na-X, NaCa-A) and per channel intersection (silicalite), respectively. The main characteristic results are: (1) D decreases with increasing chain length. (2) For small concentration of the adsorbed molecules D decreases with decreasing diameter of the pores. (3) For Na-X and silicalite D decreases and for NaCa-A it increases with increasing concentration.

While the first two results are self-explanatory, the different concentration dependencies require a more detailed discussion. As is well known, the self-diffusion coefficient may be related to the mean square jump length ( P ) and the mean residence time 7, between two succeeding translational jumps by

D

= (P)16~,

so that the question arises whether ( P ) or 7, is the quantity controlling the dependence of D or whether a variation of both quantities must be taken into consideration. In the case of NaCa-A where large cavities are connected via windows with a small diameter (see Table 1) it is reasonable to assume a constant r.m.s. jump length given by the distance of neighbouring large cavities. Therefore, Eyring's absolute rate theory can be applied364yielding 7,

a

za/zf

(37)

where Za and Z" denote the partition function of the molecule in the adsorbed and in the activated state, respectively. The former state corresponds to the residence of the molecule in the large cavity and the latter to its transient state in the window. Using the proportionality between the pressure p and the concentration a of the adsorbed molecules

where Zg is the partition function of the molecules in the gaseous state, it follows by combining equations (36)-(38):

Da (Z*/Zg).pla

(39)

Since the first term is a constant and since the adsorption isotherm is of Langmuir type, i.e. the ratio pla increases with increasing values of a , the

NMR STUDIES OF ZEOLITES

0

5

coverage

149

10

-+

.d

1

m

N

E

\

0

0

5 coverage

-

10

Fig. 32. lntracrystalline self-diffusion coefficient at 27°C for methane, ethane, and propane adsorbed on Na-X ( 0 ) . silicalite (m), and NaCa-A (0)as a function of the concenti.ation expressed as number of molecules per large cavity (Na-X, NaCa-A) or per channel intersection (silicalite).363

self-diffusion coefficient should increase with increasing concentration of the adsorbed molecules in agreement with the experimental result. In contrast, for the zeolite Na-X the diameter of the windows is not much smaller than the diameter of the large cavities (see Table 1).Therefore it is not justified to assume a constant jump length, and the behaviour of the adsorbed

150

H. PFEIFER AND H. ERNST

T/K 400 IIII I

4

2

O2 4

1

I

I

200

100

I

1

1% A0 to lo

Q, A

Fig. 33. Temperature dependence of the longitudinal proton magnetic relaxation time of propane adsorbed on Na-X for a concentration of 6.5 (A),4 (e), and 2 ( 0 )

molecules per large molecules should resemble that of molecules in a liquid. The same holds for the silicalite where the intracrystalline space is formed by intersecting channels so that the diameter of the “windows” (channels) is even equal to that of the “cavities” (channel intersections). Therefore, on the right-hand side of equation (36) both quantities may change with an increase of the concentration. The problem can be solved quantitatively by a direct measurement of the mean residence time T, between two succeeding translational jumps. Except for zeolites synthesized in the laboratory under conditions of extreme purity, the longitudinal proton magnetic relaxation time TI of adsorbed molecules is determined by the magnetic dipole interaction with Fe3+ impurities of the zeolite framework.365 This interaction is modulated by the translational motion of the molecules so that the correlation time of the nuclear magnetic relaxation is given by 7,. On the other hand TI exhibits a minimum if:365

where w denotes the NMR frequency, so that the value of T, can be determined from the position of the minimum. In Figs 33 and 34 the temperature dependence of the longitudinal proton magnetic relaxation time TI of propane adsorbed on Na-X and silicalite is plotted.363 The measure-

NMR STUDIES OF ZEOLITES

151

temperature/'C

75 1

0 -50 1

-150

-100 I

1

1

lo20 0 0

42-

210' -

Ah

A-

22

3

4

5

6

7

8

9

1 0 3 ~ 1 ~ Fig. 34. Temperature dependence of the longitudinal proton magnetic relaxation time of propane adsorbed on silicalite for a concentration of 3 ( o ) , 2 (o), and 1 ( A ) molecules per channel intersection.363

ments were performed at a resonance frequency of 90 MHz. As can be seen, for propane molecules adsorbed on Na-X the mean residence time between two succeeding jumps is constant, in contrast to silicalite, where re increases with increasing concentration. Inserting the value of re which follows from equation (40) for the temperature of the minimum of TI into equation (36) the concentration dependence of the r.m.s. jump length can be determined. Results are presented in Table 16. While in silicalite the r.m.s. jump length is nearly independent of the concentration and approximately equal to the mean distance between two neighbouring channel intersections, in Na-X the decrease of translational mobility with increasing concentration is due to a reduction of the r.m.s. jump length. By a modification of the free-volume theory3& it was possible to interpret quantitatively the decrease of the r.m.s. jump length with increasing concentration of alkane molecules adsorbed on Na-X zeolites.

152

H. PFEIFER AND H. ERNST

Table 16. Concentration dependence of the r.m.s. jum$/ength self-diffusion in Na-X and silicalite. Number of propane molecules

(f)of

propane

Temperature (minimum of Tl) in K

D/mzs-l

(12)"2/nm

Per large cavity of Na-X 1.5 2.0 4.0 5.5 6.5

135 135 135 135 135

5.8 X lo-" 4.0 X lo-" 2.5 x lo-" 0.6 X lo-" 0.1 x 10-l1

0.78 0.65 0.51 0.25 0.1

Per channel intersection of silicalite 1.0 1.5 2.0 2.5 3.0

150 175 210 270 315

13 X lo-" 12 x 10-11 13 x lo-" 14 X lo-" 6 X lo-"

1.2 1.1 1.2 1.2 0.8

6.3. Anisotropic molecular diffusion In zeolites of non-cubic symmetry as, for example, silicalite (see Table l), molecular diffusion should be anisotropic. Assuming that condition (24) is fulfilled and that the mean diameter of the zeolite crystallite under study is much larger than the r.m.s. displacement of the molecules during the time interval A between the two magnetic field gradient pulses it follows from equation (22): JI = exp{

-+ .$- 6'- (oxcos2

rp sin26

+ D~sin2cp sin26

+ D,cos' 6)A}

(41)

where Dx, Dy,I), are the principal values of the self-diffusion tensor and cp and 6 determine the direction of the magnetic field gradients in the principal axes system. For an assemblage of zeolite crystallites (powder sample) one has to average over all orientations, and it follows

JI = (47r)-' with

I I

Z(6) =

exp{-+-2. S2. (0, cos' 6 + D,,sin26)}-Z(6) * d(cos 6) (42)

-

exp{ -?.g'-6'- (sin 6 . ~ 0 p)' s (Ox- Dy) - A } -dcp

(43)

In the case of silicalite the intracrystalline space consists of intersecting linear and zig-zag channels. Assuming that on passing a certain channel

N M R STUDIES OF ZEOLITES

I

0

I

2

I

4

’\

I

I

I

6

8

10

153

Fig. 35. Theoretical dependence of the signal decay q9 for anisotropic molecular diffusion in silicalite for various values of the ratio D,/D,. The broken line represents the result for an isotropic self-diffusion. (D)denotes the mean self-diffusion coefficient defined by equation (45). The experimental data were measured at 193 K ( o ) , 223 K (a), 273 K (O), and 298 K ( A ) for methane adsorbed on silicalite with a concentration of 3 molecules per channel interse~tion.’~’

intersection a molecule will take i &sfurther way independent of the direction where it has come from, the principal values of the self-diffusion tensor are related by367

&D, = a2/D,+ b2/D,

(44)

with a, b and c denoting the unit cell lengths in x , y and z direction. Inserting this relation into equation (42) and introducing the mean selfdiffusion coefficient

( D ) = ( 1 / 3 ) * ( D x +D y + D,)

(45)

it is possible to calculate $ numerically as a function of y 2 . 2 - 8 2 - ( D for ) given values of the ratio DJD,. Due to the structure of the channel system in silicalite, D y < D, can be excluded so that in Fig. 35 curves are plotted only for D,lDx2 1. The experimental data presented also in Fig. 35 were measured at various temperatures for methane adsorbed on silicalite at a sorbate concentration of three molecules per channel intersection.368

154

H.PFEIFER AND H. ERNST

Another possibility to study the anisotropy of self-diffusion of methane adsorbed on silicalite has been realized by introducing the crystallites into a system of parallel capillaries. Since the longer axis of the silicalite crystallites is along the z-direction, the z-axis of the channel system is preferentially oriented along the direction of the capillaries.369 Therefore, with the magnetic field gradients applied in the direction of the capillary axes, the measured self-diffusion coefficient will be

and for the orientation perpendicular to the axes

where S denotes the order parameter

s = (1/2)(3 C O S ~8 - 1)

(48)

with 8 as the angle between the z-axis of the crystallites and the direction of the capillaries. By the use of capillaries with an inner diameter of (200 -+ 20) pm and crystallites of typical size 300 X 60 X 60 pm3, values between 0.63 and 0.73 could be realized for S. For a concentration of three methane molecules per channel intersection, the measurements performed at room temperature yielded a value of 1.6 X m2 s-l for D, and of m2 s-l for the quantity Dxy = (Ox Dy)/2. It can be seen from 7.2 x Fig. 36 that these values are in good agreement with the results of the measurements on powder samples (full symbols). For comparison, in Fig. 36 also the results of molecular dynamics calculations are included which were published by three independent While the averages of these results are in excellent agreement with the principal values of the selfdiffusion tensor determined experimentally by the PFG-NMR method, it is apparent that in ref. 370 the anisotropy viz. the ratio Dy/Dx is overestimated.

+

6.4. NMR tracer desorption technique

If one approximates the assemblage of zeolite crystallites loaded with one sort of molecules by a “two-region diffusion model” which means that the behaviour of the system can be described by only four quantities, viz. the relative number pa of molecules in region a, their mean residence time 7, in this region and the self-diffusion coefficients D, and D bin regions a and b, respectively, an analytical expression for Ji can be derived (see e.g. ref. 373).

N M R STUDIES OF ZEOLITES

2

155

lo

3.0

-

3.5

4.0

4.5

5.0

10%

/T

5.5

Fig. 36. Temperature dependence of the principal values of the self-diffusiontensor for methane adsorbed on silicalite as determined from PFG-NMR measurements on powder samples (full symbols),368and comparison with the results of oriented samples (1)369 and of molecular dynamics calculations (2),370(3),”’ and (4).372

Neglecting the influence of the transverse nuclear magnetic relaxation (T2>> A) and assuming Pb

the general

= 1-pa

(49)

<< 1 and Da<
simplifies to

The practical value of this equation lies in the fact that the suppositions (49) are met if region a is related to molecules adsorbed on the intracrystalline space and b to molecules outside (intercrystalline space), so that we may write with the denotation used in the other sections of this chapter: D,

D ; Db

Dinter; Pa

1-p.inter;

7 ,

7.

(51)

A plot of In $ as a function of $.g2- 8’ yields then for large values of j ? - g 2 -a2 a straight line ln$ = - A / r - $ . 8 - S 2 - A * D

(52)

156

H. PFEIFER AND H. ERNST

Table 17. Experimental results for the time constant ? (see equation 53). and the corresponding real residence time T of a methane molecule in zeolite crystallites NaCa-A with a mean radius of 2.5 km. Sample

f'

ms

7

ms

TIP

~~~

NaCa-A SB NaCa-A DB NaCa-A DB compacted NaCa-A DB granulated

0.3 f0.1 0.3 k 0.1

0.3 f0.1 2.5 f0.5

1 8.3

0.3 f 0.1

2.3 f0.5

7.7

0.3 f 0.1

5.6f 1

19

The measurements were performed at 20°C and with a concentration of 6 methane molecules per large cavity. The type of pretreatment described in the text is indicated by SB (shallow bed) and DB (deep bed).

so that the intracrystalline self-diffusion coefficient (D), and the mean residence time in a crystallite (7)can be determined independently. The latter information is equivalent to that obtained in conventional tracer exchange experiments, and therefore this method of analysing PFG-NMR data has been called NMR tracer desorption technique374or-in view of the very short observation times in the NMR experiments (milliseconds up to seconds)-fast tracer d e ~ o r p t i o n . ~ ~ For spherical crystallites of radius R the mean intracrystalline residence time 7 must be determined only by the intracrystalline self-diffusion coefficient and the radius if there is no additional transport resistance at the outer surface. It can be shown easily that in this case 7 must be equal to

Numerical results for a deviation from the spherical shape of the crystallites there must be an can be found in ref. 375. If, however, 7 is larger than additional transport resistance at the external surface, and the ratio T/Pcan be taken as measure of the strength of this resistance. In Table 17 experimental results are collected for a NaCa-A zeolite pretreated under shallow bed conditions (a thin layer of zeolite powder is heated in vucuo with a slow rate of 10 Wh to the final temperature of 400°C) and under deep bed conditions (a thick layer of zeolite powder is heated in air with a high rate of 10Owh to the final temperature of 40O0C), for the powder compacted with a pressure of 20MPa and for powder granulated with a binder, as usual, if zeolites are used as commercial catalysts. It can be seen that neither the deep bed pretreatment nor the compaction or the granulation does lead to a change of the intracrystalline self-diffusion coefficient which means that the intracrystalline pore system is unaffected.

NMR STUDIES OF ZEOLITES

157

0.3 0

2

6

-..+

6

8

16

24

coking time/ h

Fig. 37. Mean intracrystalline residence time r ( A , A ) and the time constant ? (0,m) which is related to the intracrystalline self-diffusion coefficient (see equation 53) for methane at 2% K and a concentrationof 3 molecules per channel intersection in ZSM-5 coked by n-hexane (full symbols) and mesit lene (open symbols) in dependence on the time on stream.Y76

In contrast, the external surface of the crystallites is deteriorated both by the deep bed procedure and the granulation while the compaction is of no influence within the limits of accuracy. With this method it is also possible to answer the question whether the deactivation of a zeolite catalyst by coking is due to a blocking of the entrance of molecules through the external surface or by a blocking of the active sites in the intracrystalline space. As an example the coking of ZSM-5 by n-hexane and mesitylene shall be discussed.376In Fig. 37 values for the mean intracrystalline residence time T and the time constant T" defined by equation (53) are plotted as a function of the coking time. Apparently, the two compounds used lead to completely different dependences: for ncoincide over a large range of coking times, whereas for hexane T and rnesitylene only T is found to increase while the intracrystalline self-diffusion coefficient represented by T" remains unaffected. Hence, one must conclude that during rnesitylene coking, the carbonaceous compounds are deposited exclusively on the external surfaces of the crystallites which should be expected since the mesitylene molecules are too large to penetrate the intracrystalline channel system. For n-hexane, two stages of the coke deposition become visible: at shorter coking times n-hexane is mainly

158

H. PFEIFER AND H. ERNST

deposited in the intracrystalline space, thus simultaneously affecting a retardation of intracrystalline diffusion and of the mean intracrystalline residence time. In a second stage, similar to the behaviour observed with mesitylene, coke is predominantly deposited at the external surfaces of the crystallites. The observed enhancement of the ratio r / p can be explained by two limiting cases: (1) A certain amount of pores at the external surfaces of the crystallites are blocked totally. (2) The mean diameter of all pores at the external surfaces of the crystallites has been reduced to a certain degree. should While in the former case the apparent activation energies of T and be equal, they must be different if the latter case applies to the system under study. Instead of measuring the temperature dependence in order to determine the apparent activation energy it is also possible to study the dynamic behaviour of suitably selected molecules: if the ratio r / p for the diffusing molecules does not depend on their diameter, the surface barrier is caused by a few totally blocked pores at the outer surfaces of the crystallites while in the other case it is due to a reduced mean diameter of all external pores. Table 18 shows results of NMR tracer desorption measurements for different short-chain paraffins in NaCa-A zeolites. In agreement with the findings represented in Table 17, for all molecules the values of the mean intracrystalline residence time are significantly enhanced by the hydrothermal pretreatment, where the second procedure (sample mode 2) leads to a still larger effect. Using methane and butane as test molecules it was confirmed by measuring the intracrystalline diffusivities that in analogy to the results of Table 17 the intracrystalline mobility remained nearly unaffected by the pretreatment and that in the starting material (sample mode 0) r and p coincide.377The ratio between the mean intracrystalline residence time in the treated sample and in the starting material (~/7(0))as presented in Table 18 may be assumed, therefore, to coincide with T/P. From the fact that this ratio depends significantly on the applied probe molecules one has to conclude that the surface barrier is brought about by a reduction of the width of essentially all pores in the surface layer (case (2)), rather than by a complete obstruction of only a certain fraction (case(1)). From this finding one has to conclude that the phenomenon of a surface barrier must only be discussed in connection with a given adsorbate and for a certain temperature range since the activation energies of T and #' will be different in general. Therefore it is quite possible that PFG-NMR measurements with a certain probe molecule over a certain temperature interval yield significant surface barriers, while with other molecules (or at different temperatures) no indications of surface resistances and hence of a deterioration of the external surface of the zeolite crystallites are observed.

NMR STUDIES OF ZEOLITES

159

Table 18. Values for the mean intracrystalline residence time r of various short-chain paraffins adsorbed on NaCa-A zeolite crystallites after different modes of hydrothermal pretreatment. ~~

~

~

Probe molecules

Molecule per cavity

TK

Sample mode

Methane

6

193

0 1 2

0.5 k 0.2 40 k 10

60f20

1 80 120

~~~~

7

ms

r h (0)

~

4

343

0 1 2

1+. 0.3 1525 35 k 10

1 15 35

Propane

3

483

0 1 2

28f9 360 f 110 600f200

1 13 20

n-Butane

3

483

0 1 2

42f6 320 f 110 540 f 180

1 8 13

Ethane

~~

~

0, no hydrothermal pretreatment; 1, self-steaming (4 h, 600°C) with a bed height of 2-3 mm; 2, self-steaming (4 h, 600°C) with a bed height of 20 mm and subsequent steaming in saturated water vapour (8 h, 180"C).377

7. CHEMICAL REACTIONS 7.1. Experimental methods for in situ MAS-NMR reaction studies on sealed samples

During the last decades NMR spectroscopy has found widespread application in the field of heterogeneous catalysis. Its success is a consequence of the worldwide introduction of zeolites as catalysts and selective adsorbents since their well-defined structure as porous crystallites corresponding to specific surface areas of up to 1000m2/g allows the application of various NMR methods378in spite of the relatively poor signal-to-noise ratio of NMR signals compared with infrared spectroscopy or particle beam methods. Using the chemical shift of the various lines of an NMR spectrum as a fingerprint, the appearance of organic compounds can be observed quantitatively during a catalytic reaction. With respect to an application of NMR methods three different possibilities exist378to study heterogeneous catalytic systems contained in sealed samples, namely an investigation of the system before the onset of the catalytic reaction (Fig. 38a), during the catalytic reaction (Fig. 38b) or after

160

H. PFELFER AND H. ERNST

the system has reached its equilibrium state (Fig. 38c). The gradual transition of the system from the initial (Fig. 38a) to the final state (Fig. 38c), i.e. the catalytic reaction itself, can be studied either by a cyclic heating of the sample to increasing temperatures (“pretreatment temperature”) and measurement of the NMR spectra at such a low temperature, mostly room temperature, where the reaction is frozen (method A, see refs 43, 275, 318, 321 and Table 21) or by in situ variable temperature measurements (method B, see refs 322, 379-381 and Table 21). As a typical example for an application of method A, the first investigation of a catalytic process in zeolites which was devoted to the double bond isomerization of but-1-ene adsorbed on CaNa-Y zeolites shall be mentioned3I8 (see Fig. 39). For more details and a survey of other reactions studied later on, the reader is referred to refs 43 and 275. The introduction of the magic angle spinning (MAS) technique to 13C NMR spectroscopy of heterogeneous catalytic reactions has led to two important improvements: (1) the resolution of the spectra and hence the sensitivity to ascertain reaction products could be enhanced significantly; and (2) it has become possible to observe also the carbon-13 signals of strongly bound reaction products including coke (cf. Fig. 38c). From such measurements quantitative information about the amount and chemical composition of the coke can be derived.79~123~338~382-385 With regard to the location of the coke which may be deposited in the intracrystalline space and/or on the outer surface of the zeolite crystallites, however, probe molecules have to be used. The so-called NMR tracer desorption technique386described in the section on molecular diffusion and the 129Xe NMR of adsorbed xenon387shall be mentioned here. 13C MAS-NMR studies of catalytic reactions taking place in sealed samples have been performed by cyclic heating of the sample (method A: refs 51, 69, 190, 321, 329, 335, 336, 338, 340, 343, 388-390, 393), or by in situ variable temperature measurements (method B: refs 322, 333, 334, 339, 342, 391, 392). A survey is given in Table 21. As examples for the method B, in Fig. 40 proton decoupled 13C MAS-NMR spectra of an activated CaNa-Y (20% Ca2’) zeolite loaded with but-1-ene during the double bond isomerization process and in Fig. 41 two-dimensional heteronuclear ‘H-13C chemical shift correlated MAS-NMR spectra394of the same sealed sample are shown.393The measurements were performed on a commercial NMR spectrometer (Bruker MSL 300) at 74.5 MHz with a spinning rate of ca. 1.5 kHz in the temperature range from 25°C to 70°C. The but-1-ene was enriched to 99% in 13C in the H2C= position. From Figs 40 and 41 one can see that both the resolution as well as the sensitivity are sufficient for (1) a detailed discussion of the isomerization process including a better examination of spectral assignments obtained from 1D spectra and (2) a study of adsorbed molecules even with natural abundance in 13C.

NMR STUDIES OF ZEOLITES

161

0

reactant(s)

active site(s) catalyst

framework

reactant

0

(b)

I

H catalyst

-0-

\

carbo-

@cation

-o--

ingress of the reactant. accessibility

proton transfer

-o--

9

products

/

f l %

-0&coke

decomposition

egress of the products. coking

products (c) active sites

damage

\ cake

catalyst

Fig. 38. Different possibilities to study heterogeneous catalytic systems with NMR method^:^" (a) reactant(s) and catalyst before the onset of the reaction; (b) schematic representation of the kinetics of a typical heterogeneous catalytic reaction (Bronsted acid catalysed cracking reaction); (c) products and (partially) deactivated catalyst after the reaction.

162

H. PFEIFER AND H. ERNST 131 1 1 1

ILI

161

but-1 -ene

H,C=CH -CH2-CH,

trans-but-2-ene

H,C-HC=CH-CH,

cis -but-2 - ene

H,C-CH =CH -CH,

(51

(21

161 1 2 1

I21

I2J

(51

161

final

stote

368 K

t

353 K

start

Fig. 39.

Proton decoupled 13C NMR spectra during the double bond isomerization of but-1-ene adsorbed on a CaNa-Y (72% Ca2+) zeolite.318

In Fig. 42 13C CP MAS-NMR spectra (CP = cross-polarizationZ1) are shown which were obtained after loading an H-Y zeolite and an ultrastable H-Y zeolite with propene enriched to 99% in I3C in the =CH2 position.388 The samples were kept at room temperature and measured at a resonance frequency of 25 MHz and at spinning rates of 2-3 kHz. With this paper, the authors have shown that 13C MAS-NMR spectroscopy can be used to identify both the static isopropyl carbocation (the peak at ca. 250 ppm which is formed on acidic sites immediately after adsorption) and dynamic

NMR STUDIES OF ZEOLITES

160

140

120

100

PPM

80

60

40

163

20

Fig. 40. Proton decoupled 13C MAS-NMR spectra of a CaNa-Y (20% Ca2+) zeolite loaded with but-1-ene. The but-1-ene was enriched to 99% in I3C in the H2C= position. The pore filling factor is 0.8, the NMR resonance frequency 74.5 MHz and the spinning rate ca. 1kHz. The sealed sample was measured in dependence on the temperature (method B).393

carbocations (the peak at ca. 160ppm which is due to a rapid exchange of two free alkyl cations) in zeolites. In their first application of the in sifu variable temperature 13C CP MAS-NMR method (method B) Haw et ul.322have studied the oligomerization reactions which take place in H-Y zeolites loaded with propene enriched t o 99% in 13C in the H3C-, -CH=, and =CH2 positions,

164

H. PFEIFER AND H. ERNST

respectively. In a recent publication391this technique was used to monitor the cracking of propene oligomers adsorbed on H-Y zeolites in a temperature range from room temperature to ca. 250°C. The experiments were performed at 75.4 MHz and at spinning rates of 3.5-4 kHz. It was found that at the upper temperature the oligomers cracked to form branched butanes, pentanes, and other alkanes and that the driving force for this reaction was the formation of highly aromatic coke. 7.2. Conversion of methanol

Applying method A, Anderson and Klinowski51.69,190,321,329,335,336,389,390 have studied the catalytic conversion of methanol to hydrocarbons adsorbed on H-ZSMJ zeolites. By magic angle spinning of the sealed samples a considerable gain in resolution of the 13C. NMR spectra could be achieved with regard to earlier work. 337,382939s397 The experiments were performed at room temperature at a resonance frequency of 75.4 MHz and spinning rates up to 3 kHz. In these experiments it was possible to identify 29 different organic species in the adsorbed phase. The 13C MAS-NMR spectrum of an H-ZSM-5 zeolite loaded with methanol enriched to 30% in 13C is shown in Fig. 43(a). After heating the sample to 150°C for 20min, the spectrum shown in Fig. 43(b) is composed of signals at 50.5 and 60.5ppm, which correspond to methanol (MeOH) and dimethyl ether (DME), respectively. Figure 44 shows the spectrum of the sample after heating it to 300°C for 35 min. Methanol and dimethyl ether have been completely converted to a mixture of aliphatics and aromatics. Although some signals (13 in the aliphatic and 19 in the aromatic region) overlap, particularly those from methyl groups attached to aromatic rings, it was possible to assign all of them.51,389In Fig. 45 the corresponding heteronuclear two-dimensional J-resolved I3C MAS-NMR ~ p e c t r u m is ~ shown, ~ ~ ~which ~ has been used to confirm the spectral assignments obtained from one-dimensional 13CMASNMR experiments. With this NMR method it is possible to determine the connectivity of carbons and the number of protons attached to each carbon atom in the various organics and the details of 13C-'H couplings. For example, the resonance at -10.7ppm in Fig. 45 is split into five components, which confirms that it must be due to adsorbed methane. Comparing the distribution of the species observed in the sealed samples by 13CNMR and the distribution of the reaction productions by gas chromatography it was possible to observe different kinds of shape selectivity in the zeolites. Fig. 41. Two-dimensional heteronuclear 'H-I3C chemical shift correlated MASN M R spectra (method B) of a CaNa-Y (20% Ca2+)zeolite loaded with but-1-ene at measuring temperatures of 70°C (top) and 24°C (bottom). The experimental conditions are the same as in Fig. 40.393

NMR STUDIES OF ZEOLITES

I

160

.

,

140

.

,

120

1

80

100

PPM

PPM

60

40

.

8

20

165

166

H. PFEIFER AND H. ERNST

I

L

300

200

I

1

L

100

0

-100

6 IPPm Fig. 42. I3C CP MAS-NMR spectra (method A) of propene adsorbed 011 an H-Y zeolite (a), and on an ultrastable H-Y zeolite (b). As usual, * denotes spinning sidebands. The propene was enriched to 99% in I3C in the =CH2 position.398

However, in such a comparison not only shape selectivity but also the influence of strongly adsorbed species upon the equilibrium distribution in sealed samples must be taken into consideration. This important fact has been derived both experimentally and theoretically in a recent on methylamine synthesis. 7.3. Synthesis of methylamines

The catalytic conversion of methanol and ammonia to monomethylamine (MMA), dimethylamine (DMA), and trimethylamine (TMA) adsorbed on various catalysts was investigated by I3C MAS-NMR spectroscopy applying method A.323The following catalysts have been studied:

K1: sodium mordenite ( S U N = 5 ) , K2: sodium ZSM-5 (SUAI = 15), K3: 20% proton-exchanged sodium mordenite K1, K4 : amorphous alumina, K5: 65% K1+ 35% K4 as binder, K6 : 65% K3 + 35% K4 as binder.

N M R STUDIES OF ZEOLITES

167

MoOH

80

70

60

50

40

30

6 IPPm Fig. 43. Proton decoupled 13C MAS-NMR spectra (method A) of methanol adsorbed on an H-ZSM-5 zeolite, measured at room temperature: (a) without thermal pretreatment, (b) after heating the sealed Sam le to 150°C for 20min. The methanol was enriched to 30% in p3C.51.3sy

168

H. PFEIFER AND H. ERNST

I

50

40

I

30

I

I

I

20

10

0

-10

- 20

d JPPm Fig. 44. Proton decoupled I3C MAS-NMR spectrum (aliphatic region) of the same samples as in Fig. 43, but after heating it to 300°C for 35 min.51,389

A11 samples were prepared under the conditions of shallow-bed pretreatment: ca. 50mg of a catalyst were pumped and heated with a temperature rate of 10 K/h to the activation temperature of 400°C where the catalyst was kept under a pressure below lo-* Pa for 24 h. Then the catalyst was cooled, loaded under vacuum at first with 70 2 10 pmol methanol enriched to 99% in 13C and after that with 50+ 10 pmol ammonia and sealed. The concentration of ammonia corresponds to an NH3/AI ratio of about 1/3, 1, 1, and 1/20 for K1, K2, K3, and K4, respectively. All 13C MAS-NMR measurements, combined in most cases with the cross-polarization (CP) technique to enhance the sensitivity, were performed on a commercial NMR spectrometer (Bruker MSL 300) at a resonance frequency of 74.5 MHz and at a spinning rate of ca. 2 kHz. Figure 46 shows the 13C CP-NMR spectra of sodium mordenite (catalyst K1) loaded with methanol and ammonia pretreated at 275°C for 1h: (a) for the static sample and (b) under MAS conditions (spinning rate ca. 2 kHz). Spectrum (b) demonstrates clearly the enhancement of resolution achieved by the use of the MAS technique. The assignment of the five

* 300 *

200

’ 100

- 0

Hz

.-lo0

.-200

- -300

26

Fig.

24

22

4

18

17 16 WPPm

-

15

0

-11

-12

45. Two-dimensional J-resolved I3C MAS-NMR spectrum of the same sample as in Fig. 43 but after heating it to 300°C for 30 min.s17340

s

\o

170

H. PFEIFER AND H. ERNST

52 \

I . ( . I

60

PPm

DME

1

L

b

13C

4' MeOH

1

120

.

1

100

.

1

80

.

1

60

.

1

40

.

1

.

20

MAS NMR

1

0 -6lppm

.

1

-20

Fig. 46.13C CP-NMR spectra of a sodium mordenite, loaded with methanol and ammonia after heating the sample to 275°C for 1 h: (a) for the static sample; (b) under MAS conditions (spinning rate ca. 2 kHz); (c) the 13C MAS-NMR signal of the

tetramethylammoniumions ((CH&N+) without proton d e c o ~ p l i n g . ~ * ~

significant lines is: MMA (26.6 ppm), DMA (37.2 ppm), TMA (46.7 ppm), (CH3)4Nf (tetramethylammonium ions, 55.8 ppm), and DME (dimethyl ether, 59.6ppm) in the order of increasing values of the chemical shift. Between the signals of TMA and (CH3)4N+ a small line due to residual methanol (49.7ppm) can be seen. The assignment of the fourth line to (CH3)4N+follows not only from its position in the spectrum but also from the splitting of this line into a quartet if the proton decoupling is released (cf. Fig. 46c). In principle, it is possible to derive the same information also from heteronuclear two-dimensional J-resolved MAS-NMR spectra (similar to Fig. 45). In Fig. 47 13C CP MAS-NMR spectra are shown for the same system as a function of the pretreatment temperature at which the sample

N M R STUDIES OF ZEOLITES

(CH3)4N* I

171

DMA

I

DME

Fig. 47. 13CCP MAS-NMR spectra of a sodium rnordenite, loaded with methanol and ammonia, as a function of the pretreatment temperature at which the sample was kept for 30min: (a) 275"C, (b) 30O0C, (c) 3SO0C, (d) W 0 C ,(e) 450°C. As usual, * denotes spinning sidebands.323

was kept for 30 min. From these spectra it is possible to derive the values for the relative concentration of the methylamines (MMA, DMA and TMA) depending on the pretreatment temperature, because the same relative intensities have been measured using spectra without cross-polarization enhancement (see below). At about 450°C the decomposition of the

172

H. PFEIFER AND H. ERNST

methylamines starts, giving rise to new lines in the aliphatic region of the 13C NMR spectrum, and with increasing pretreatment time, also in the aromatic region due to coking of the catalyst. From the I3C MAS-NMR spectra information on (1) the activity of the catalyst, (2) the concentration of the reaction products (selectivity), and (3) the deactivation of the catalyst can be derived. (1) Activity of the catalyst. The temperature of the onset of the catalytic reaction can be taken as a measure of the activity of the catalyst.318For the system under study the reaction starts in the case of the acidic catalysts K3, K4, K5, K6 at a temperature of about 250"C, and in the case of the non-acidic catalysts K1 and K2 at about 300°C. This difference can be explained by the existence of Bronsted acid sites in the proton-exchanged mordenites as well as in alumina, where these sites are formed by the interaction of water and Lewis acid sites. (2) Relative concentration of the reaction products. From the intensities of the lines, values for the relative concentration of the corresponding reaction products can be derived. However, this is only true if the cross-polarization technique does not mask the real relative intensities3" which holds for methanol and the amines (MMA, DMA, and TMA), in all samples under study. Only for the tetramethylammonium ions a significant deviation has been observed which may be ascribed to the spherical shape of these ions allowing a fast isotropic reorientation. Values for the relative concentration of the amines (the ratio of the signal intensity for a given amine to the sum of MMA, DMA and TMA) derived from the I3C CP MAS-NMR spectra for the non-acidic catalysts K1 and K2 at a pretreatment temperature of 400°C as well as computed for the thermodynamic equilibrium398are given in Table 19. Within the limits of accuracy of the NMR measurements a reasonable agreement can be observed. In contrast, for the acidic catalysts K3, K4, K5, and K6 pretreated at the same temperature, the relative concentration of MMA is strongly enhanced (cf. Table 20). This result can be explained by the formation of strong adsorption complexes between MMA and the Bronsted acid sites of the catalysts. The tripod structure of these complexes399is based on hydrogen bond formation between the Bronsted acid site and the lone electron pair of MMA as well as between the N-Hs and suitably located nearby bridging oxygens of the catalyst framework. For DMA and TMA only bipod and monopod structures result with a corresponding smaller stability. The essential point here is that due to the formation of adsorption complexes (by at least one of the reaction products), the equilibrium of the methylamine synthesis is shifted to the right so that the values for the concentration of the reaction products measured with sealed samples (c,) must be different from the values (ce) calculated under the assumption of a thermodynamic equilibrium between methanol, ammonia and the reaction products in the

NMR STUDIES OF ZEOLITES

173

Table 19. Values for the relative concentrations of amines MMA, DMA and " M A derived from the 13C CP MAS-NMR spectra for the non-acidic catalysts K l and K2 at a pretreatment temperature of 400°C as well as those computed for thermodynamic eq~ilibrium.'~'

Relative concentration (%) Meth ylamine

from NMR323 catalyst

MMA DMA TMA

K1

K2

40 20 36

25 23 49

for thermodynamic equili b r i ~ m ~ ~ ~

22 32

45

Table 20. Values for the relative concentrations of amines MMA, DMA and TMA (1) derived from the 13C CP MAS-NMR spectra for the acidic catalysts K3, K4,K5, K6 at a pretreatment temperature of 400°C; (2) computed for the thermodynamic equilibrium and (3) computed for the case of total consumption of CH30H and ammonia.

Relative concentration (%) Methylamine

MMA DMA TMA

(1) from NMR3= catalyst

(2) for thermod namic equilibriumt98

K3

K4

K5

K6

67 18 15

71 17 12

66 20 14

60 23 . 17

22 32

45

(3) value for total consumption of methanol323 5 0 . . . 75 50 . . . 0 0 . ..25

gaseous state. It is well known that, due to shape selectivity the latter values (ce) will be also different in general from the values (cf) measured under flowing gas conditions. (3) Amine decomposition and catalyst deactivation. The onset of the amine decomposition and the nature of the decomposition products (carbonaceous deposits400,401) characterize the deactivation of the catalyst. Figure 47 shows that for non-acidic catalysts the amine decomposition starts only at temperatures above 350°C, while the latter temperature is quite sufficient for the signal of methanol to vanish completely and for the signals of the amines MMA, DMA, and TMA to appear (Fig. 47c, d). At 450°C the amines start to be decomposed as can be seen from the signals in the

174

H. PFEFER AND H. ERNST

aliphatic region of the 13C CP MAS-NMR spectra (Fig. 47e). With increasing pretreatment time, additional signals appear in the aromatic region due to coking (deactivation) of the catalyst (not shown). Table 21 presents results of 13C and 'H MAS-NMR studies on catalytic reactions of adsorbate-zeolite systems. With regard to results of static 13C and 'H NMR measurements the reader is referred to ref. 43.

Table 21. 13C and 'H MAS-NMR studies of catalytic reactions using zeolites in sealed samples. Molecule (reaction)

Zeolite

Temperature interval of measurement

Nucleus Remarks MHz)

Ref.

(YO

(K) H-Y

r.t., 493A

H-Y

163-313'

H-Y

r .t ,423'

13C (75.4)

Coke formation

391

CH2=CHCH2CH3 CaNa-Y (isomerization)

r. t . - W B

I3C (75.4)

Two-dimensional heteronuclear 'HI3C MAS-NMR

393

CHz=CHCH=CH, H-Y, (oligomerization) H-ZSM-5

143-r.t.'

13C (25)

No evidence of carbocations

392

CHZ=CH, (polymerization, isornerization, cracking, coking)

r.t.-7W

13C (75.4)

Evidence of carbocations, lowand hightemperature coke

338

CHz=CHz H-ZSM-5, (oligomerization) H-Y, H-Mor

223-303'

13C (75.4)

Evidence of alkoxide species

339

CH-CH

r.t.-503B

I3C (75.4)

Evidence of alkoxy intermediates

342

173-423B

13C

CHZ=CHCH3 (oligomerization)

CHz=CHCHp (cracking of oligomers) ~~~~~~~~

13C (25)

388

Evidence of carbocations

13C Study of the (25,75.4) reaction steps, evidence of carbocations

322

~~

~~

H-Mor

H-Y, H-ZSM-5 ~~~~~

CH30CH3 (MTG-process)

~~~

H-ZSM-5, Na-ZSM-5

~

(50)

~~

~~~

~~~~~

Formation of trimethyloxonium

334

NMR STUDIES OF ZEOLITES

175

Table 21-contd. Molecule (reaction)

Zeolite

Temperature interval of measurement

Nucleus Remarks MHz)

Ref.

(vg

(K) CH30H (h4TG-process)

H-ZSM-5

29tF-523B

I3C (75.4)

No evidence of CO (cf. ref. 321)

333

H-ZSMJ

r. t .-643A

13C Observation of CO (100.6) as an intermediate (cf. ref. 333), observation of shape selectivity

321

H-ZSMJ

r .t.-643A

'3C (100.6)

3 89

'H

(@o)

H-ZSMJ

CH:IOH

593A

I3C Heteronuclear two(100.6) dimensional J-resolved I3C spectroscopy

390

No evidence of CO

69

SAPO-34

r.t.4~43~

H-ZSMJ, SAPO-34

r. t .-643A

13C cf. refs 69, 321 (100.6)

H-ZSM-5, SAPO-5

r. t .-573A

1 3 c Evidence of 335 (100.6) methoxy groups in the case of SAPO-5 'H (400)

H-ZSM-5, S APO-34

r.~ - 5 7 3 ~

H-ZSMS, Na-ZSM-5, H-Y, Na-Y, Na-A, H-L, silicalite, SAPOJ, -11,-34

r.t.A

(100.6)

Review (cf. refs 69, '3c (100.6) 190,329,335,336, 'H 389, 390), twodimensional (400) heteronuclear J-resolved spin diffusion I3C MASNMR 'H (400)

Evidence of methoxonium ions

190

51

329

176

H. PFEIFER AND H. ERNST

Table Il-contd. Molecule (reaction)

Zeolite

Temperature interval of measurement

Nucleus Remarks (v" MHz)

Ref.

(K) CH30H

H-, Na-, K-ZSMJ

r.t.A

'H (400)

FTIRa nd' H MAS-NMR spectroscopy, cf.

336

ref. 329

CH3COCH3

~~~~~

H-ZSM-5, H-Y

r.t., 453A

Na-X

r.t., 523, 673A

(50.3)

FT IR spectroscopy, 340 protonization of acetone

I3C (75.4)

Review

I3C

~

CH,OH

+ H2S

343

'H (300) ~~

~

CH30H + NH3 (aminization)

Na-ZSM-5, Na-Mor, 20HNa-Mor

r.t.-723*

"C (75.4)

Activity of catalysts, 323 selectivity with regard to mono-, diand trimethylamine, evidence of tetramethylammonium ions, coking

AMethod A: cyclic heating of the sample to increasing temperatures and measurement at such a low temperature where the reaction is frozen. BMethod B: in situ variable temperature measurements.

REFERENCES 1. D. W. Breck: Zeolite Molecular Sieves, Structure, Chemistry and Use. John Wiley & Sons, London, 1974

2. R. M. Barrer: Zeolite and Clay Minerals as Sorbenrs and Molecular Sieves. Academic Press, London, 1978. 3. W. M. Meier and D. H. Olson: Atlas of Zeolite Structure Types, 2nd rev. edn. Butterworth, London, 1987. 4. Studies in Surface Science and Catalysis 58,65,69. Elsevier, Amsterdam, 1991. 5. A. Cronstedt, Kongl. Svenska Vetenkaps Acad. Handlingar, 1756, 17, 120. 6. W. J. Mortier: Compilation of Extra Framework Sites in Zeolites. Butterworth, London, 1982. 7. R. K. Hams and B. E. Mann, NMR and the Periodic Table. Academic Press, New York, 1978. 8. J. S. Waugh, L. M. Huber and U. Haeberlen, Phys. Rev. Lett., 1968,20, 180. 9. U. Haeberlen and J. S. Waugh, Phys. Rev., 1968, 175, 453.

NMR STUDIES O F ZEOLITES

177

P. Mansfield, J. Phys. C , 1971, 4, 1444. P. Mansfield and D. Ware, Phys. Lett., 1966, 22, 133. P. Mansfield and D. Ware, Phys. Rev., 1968, 168, 318. W. K. Rhim, D. D. Elleman and R. W. Vaughan, J . Chem. Phys., 1973,58, 1772. W. K. Rhim, D. D. Elleman, L. B. Schreiber and R. W. Vaughan, 1. Chem. Phys., 1974, 60, 4595. 15. H. Ernst, D. Fenzke and W. Heink, Phys. Status Solidi B , 1973, 57, K103. 16. N. Q. Lan, H. Pfeifer and H. Schmiedel, Wiss. Z. Univ. Leipzig, Math.-Naturwiss. R., 1974, 23,498. 17. H. Ernst, D. Fenzke and W. Heink, Wiss.Z. Univ. Leipzig, Math.-Naturwiss. R., 1974, 23, 545. 18. D. Fenzke and H. Schmiedel, Wiss. 2.Univ. Leipzig, Math.-Nuturwiss. R . , 1974, 23, 519. 19. I>. Burum and W. K. Rhim,J. Magn. Reson., 1979, 34, 241. 20. D. Burum and W. K. Rhim, 1. Chem. Phys., 1979, 71,944. 21. M Mehring, High Resolution NMR Spectroscopy in Solids, 2nd ed. Springer-Verlag, Berlin, 1983. 22. A. Pines, M. G. Gibby and J. S. Waugh, J. Chem. Phys., 1972, 56, 1776. 23. A. Pines, M. G. Gibby and J. S. Waugh, J. Chem. Phys., 1973, 59, 569. 24. H. Ernst, D. Fenzke and H. Pfeifer, Ann. Phys. (Leipzig), 1981, 38, 257. 25. E. R. Andrew, A. Bradbury and R. G. Eades, Nature (Lond.), 1958, 182, 1659. 26. I. J. Lowe, Phys. Rev. Lett., 1959, 2, 285. 27. B. C. Gerstein, Adv. Colloid Interface Sci., 1985, 23, 45. 28. R. E. Taylor, R. G. Pempleton, L. M. Ryan and B. C. Gerstein, J. Chem. Phys., 1979, 71, 4541. 29. L. M. Ryan, R. E. Taylor, A. J. Paff and B. C. Gerstein, J. Chem. Phys., 1980, 72, 508. 30. G. Scheler, U. Haubenreisser and H. Rosenberger, J. Magn. Reson., 1981,44, 134. 31. S. Ganapathy, S. Schramm and E . Oldfield, J. Chem. Phys., 1982, 77,4360. 32. E. Oldfield, S. Schramm, M. D. Meadows, K. A. Smith, K. A. Kinsey and J. Ackerman, 1.Am. Chem. Soc., 1982, 104, 919. 33. E. Oldfield, R. A. Kinsey, B. Montez, T. Ray and K. A. Smith, J . Chem. SOC., Chem. Commun., 1982, 254. 34. S. Schramm and E. Oldfield, J . Chem. SOC., Chem. Commun., 1982,980. 35. B. F. Chmelka, K. T. Mueller, A. Pines, J. Stebbins, Y.Wu and J. W. Zwanziger, Nature (Lond.), 1989, 42, 339. 36. K. T. Mueller, B. Q. Sun, G. C. Chingas, T. Terao and A. Pines, J . Magn. Reson., 1990, 86,470. 37. K. T. Mueller, E. W. Wooten and A. Pines, J. Mugn. Reson., 1991, 92,620. 38. A. Samoson, E. Lippmaa and A. Pines, Mol. Phys., 1988, 65, 1013. 39. A. Samoson and E. Lippmaa, J. Magn. Reson., 1989,84,410. 40. A. Samoson and A. Pines, Rev. Sci. Instr., 1989, 60,3239. 41. Y. Wu, B. 0.Sun, A. Pines, A. Samoson and E. Lippmaa, 1. Magn. Reson., 1990, 89, 297. 42. C. A. Fyfe, Solid State NMR for Chemists, CFC Press, Guelph, Ontario, 1983. 43. G. Engelhardt and D. Michel, High-Resolution Solid-State NMR of Silicates and Zeolites. John Wiley & Sons, Chichester, 1987. 44. J. Klinowski, Ann. Rev. Muter. Sci., 1988, 18, 189. 45. J. B. Nagy and E . G. Derouane, NMR Spectroscopy and Zeolite Chemistry, ACS Sympos. Ser., 1988, 368,2. 46. A. D. H. Clague and N. C. M. Alma, in Analytical NMR (eds L. D. Field and S. Sternhell), p. 115. John Wiley & Sons, New York, 1989. 47. G. Engelhardt, Stud. Surf. Sci. Catal., Elsevier, 1991, 58, 258. 48. J. Klinowski, Colloids and Surfaces, 1989, 36, 133. 10. 11. 12. 13. 14.

178

H. PFEIFER AND H. ERNST

49. C. A. Fyfe, Y. Feng, H. Grondey, G. T. Kokotailo and H. Gies, Chem. Rev., 1991, 91, 1525. 50. H. Pfeifer, Nato ASZ Series B: Physics, Vol. 265, p. 151. Plenum Press, New York and London, 1991. 51. J. Klinowski, Chem. Rev., 1991,91, 1459. 52. D. Freude and J. Haase, NMR - Basic Principles and Progress, 1993, 29, 1. 53. J. W. Akitt, Prog. Nucl. M a p . Reson. Spectrosc., 1989, 21, 1. 54. D. R. Corbin, B. F. Burgess, A. J. Vega and R. D. Farlee, Anal. Chem., 1987,59,2722. 55. G. Engelhardt, Trend Anal. Chem., 1989, 8 , 343. 56. D. Freude, Stud. Surf. Sci. Catal., 1989, 52, 169. 57. J. M. Thomas and D. E. W. Vaughan, J. Phys. Chem. Sol., 1989,50,449. 58. J. M. Thomas, Stud. Surf. Sci. Catal., Elsevier, 1989, 49, 3. 59. G. Engelhardt, U. Lohse, E. Lippmaa, M. Tarmak and M. Magi, Z . Anorg. Allg. Chem., 1981,482, 49. 60. W. Loewenstein, Am, Mineral., 1954,39,92. 61. C. A. Fyfe, G. C. Gobby, W. J. Murphy, R. S. Ozubko and D. A. Slack, J . A m . Chem. SOC.,1984, 106, 4435. 62. H. Strobl, C. A. Fyfe, G. T. Kokotailo, C. T. Pasztor and D. M. Bibby, 1. Am. Chem. SOC.,1987, 109, 7433. 63. R. B. La Pierre, A. C. Rohrman, J. L. Schlenker, J. D. Wood, M. K. Rubin and W. J. Rohrbaugh, Zeolites, 1985, 5 , 346. 64. A. Sarnoson and E. Lippmaa, Chem. Phys. Lett., 1983, 100,205. 65. A. Samoson, E. Lippmaa, G. Engelhardt, U. Lohse and H. G. Jerschkewitz, Chem. Phys. Lett., 1987, 134, 589. 66. A. G. Pelmenschikov, E. A. Paukshtis, M. 0. Edisherahvili and G. M. Zhidomirov, J . Phys. Chem., 1992, %, 7051. 67. D. Akporiaye and M. Stocker, Zeolites, 1992, 12, 351. 68. L. B. Alemany, H. K. C. Tirnken and I. D. Johnson, J. Magn. Res., 1988,80,427. 69. M. W. Anderson, B. Sulikowski, P. J. Barrie and J. Klinowski, J. Phys. Chem., 1990,94, 2730. 70. M. W. Anderson, Magn. Reson. Chem., 1992,30,898. 71. K. J. Andries, B. De Wit, P. J. Grobet and H. J. Bosmans, Zeolites, 1991, 11, 116. 72. K. J. Andries, H. J. Bosmans and P. J. Grobet, Zeolites, 1991, 11, 124. 73. M. J. Annen, D. Young, J. P. Arhancet, M. E. Davis and S. Schramm, Zeolites, 1991,11, 98. 74. L. Aouali, J. Jeanjean, A. Dereigne, P. Tougne and D. Delafosse, Zeolites, 1988, 8 , 517. 75. S. A. Axon and J. Klinowski, Stud. Surf. Sci. Catal., 1989, 52, 113. 76. S. A. Axon and J. Klinowski, J . Chem. SOC.,Faraday Trans., 1990, 86, 3679. 77. S. A. Axon, K. Huddersman and J. Klinowski, Chem. Phys. Len., 1990,172, 398. 78. I. Balakrishnan and S. Prasad, Appl. Catal., 1990,62, L7. 79. M. C. Barrage, F. Bauer, H. Emst, J. Fraissard, D. Freude and H. Pfeifer, Cataf. Lett., 1990, 6, 201. 80. P. J. Barrie and J. Klinowski, 1.Phys. Chem., 1989, 93,5972. 81. P. J. Barrie, M. E. Smith and J. Klinowski, Chem. Phys, Lett., 1991, 180, 6. 82. P. J. Barrie, L. F. Gladden and J. Klinowski, J. Chem. Soc., Chem. Commun., 1991,592. 83. G. BeIlussi, A. Millini, A. Carati, G. Maddinelli and A. Gervasini, Zeolites, 1990, 10, 642. 84. R. D. Bezman, J . Chem. SOC., Chem. Commun., 1987, 1562. 85. C. S. Blackwell and R. L. Patton, J. Phys. Chem., 1988, 92, 3965. 86. G. Borbely, H. Beyer, L. Radin, P. Sandor and H. G. Karge, Zeolites, 1989, 9, 428. 87. V. Bosacek and V. M. Mastikhin, J. Phys. Chem., 1987, 91,260. 88. V. Bosacek and D. Freude, Stud. Surf. Sci. Catal., 1988, 37, 231.

NMR STUDIES OF ZEOLITES

179

89. E. Bourgeat-Lami, P. Massiani, F. Di Renzo, F. Fajula and T. Des Courieres, CutuZ. Lett., 1990, 5 , 265. 90. E. Bourgeat-Lami, P. Massiani, F. Di Renzo, P. Espiau, F. Fajula and T. Des Courieres, Appl. Catal., 1991, 72, 139. 91. M. Brien-Faure, 0. Cornu, D. Delafosse, R. Monque and M. J. Peltre, Appl. Catul., 1988, 38, 71. 92. E. Brunner, H. Ernst, D. Freude, M. Hunger and H. Pfeifer, Stud. Surf. Sci. Catal., 1988, 37, 155. 93. E . Brunner, D. Freude, M. Hunger, H. Pfeifer, W. Reschetilowski and B. Unger, Chem. Phys. Lett., 1988, 148, 226. 94. E. Brunner, H. Ernst, D. Freude, M. Hunger, C. B. Krause, D. Prager, W. Reschetilowski, W. Schwieger and K. H. Bergk, Zeolites, 1989, 9,282. 95. E. Brunner, H. Ernst, D. Freude, T. Frohlich, M. Hunger and H. Pfeifer, Stud. Surf. Sci. Catal., 1989, 49, 623. 96. E. Brunner, H. Ernst, D. Freude, T. Frohlich, M. Hunger and H. Pfeifer, J . Catal., 1991, 127, 34. 97. J . C. Buhl, G. Engelhardt and J. Felsche, Zeolites, 1989, 9, 40. 98. J . Caro, M. Bulow, M. Derewinski, M. Hunger, J. Karger, U. Kiirschner, H. Pfeifer, W. Storek and B. Zibrowius, Stud. Surf. Sci. Catal., 1989,52,295. 99. J . Caro, M. Biilow, M. Derewinski, J. Haber, M. Hunger, J. Karger, H. Pfeifer, W. Storek and B. Zibrowius, 1. Caral., 1990, 124, 367. 100. R. Challoner, R. K. Harris, K. J. Packer and M. J. Taylor, Zeolites, 1990, 10, 539. 101. K. J. Chao and J. Y. Chern, J. Phys. Chem., 1989,93, 1401. 102. K. J. Chao, J. Y. Chern, S. H. Chen and D. S. Shy, J. Chem. SOC.,Faraday Trans., 1990, 86, 3167. 103. R . Chauvin, P. Massiani, R. Dutartre. F. Figueras, F. Fajula and T. Des Courieres, Zeolites, 1990, 10, 174. 104. Q. J. Chen, J. L. Guth, A. Seive, P. Caullet and J. Fraissard, Zeolites, 1991, 11, 798. 105. J . M. Chezeau, L. Delmotte, J. L. Guth and M. Soulard, zeolites, 1989, 9, 78. 106. J. M. Chezeau, L. Delmotte, J. L. Guth and Z. Gabelica, Zeolites, 1991, 11, 598. 107. B. F. Chmelka, Y.Wu, R. Jelinek, M. E. Davis and A. Pines, Stud. SurJ Sci. Catal., 1991, 69, 435. 108. W. C. Comer, R. Vincent, P. Man and J. Fraissard, Catal. Lett., 1990, 4, 75. 109. A. Corma, Stud. Surf. Sci. Catal., 1989, 49, 49. 110. J. M. Cruz, A. Corma and V. Fornes, Appl. Catal., 1989, 50, 287. 111. M. E. Davis, C. Montes, P. E. Hathaway and J. M. Garces, Stud. Su$. Sci. Catal., 1989, 49,199. 112. A. de Carvalho, Q. Wang, G. Giannetto, D. Cardoso, M. B. de Carvalho, F. R. Ribeiro, J. B. Nagy, J. El-Hage-Ai-Asswad and E. Derouane, 1. Chim. Phys., 1990, 87, 271. 113. E. C. Derouane, L. Maistriau, Z. Gabelica, A. Tuel, J. B. Nagy and R. von Ballmoos, Appl. Catal., 1989,51, L13. 114. R. M. Dessau, K. D. Schmitt, G. T. Kerr, G. L. Woolery and L. B. Alemany, J . Catal., 1987, 104, 484. 115. J . Dwyer, D. Millward, P. O’Malley, A. Araya, A. Corma, V. Fornes and A. Martinez, J . Chem. Soc., Faraday Trans., 1990,86, 1001. 116. J . Dwyer and K. Karim, J . Chem. SOC., Chem. Commun., 1991, 14,905. 117. G. V. Echevskii, V. M. Nekipelov, K. G. Ione and K. I. Zamaraev, React. Kinet. Caral. Lett., 1987,33, 233. 118. A. Endoh, K. Nishimiya, K. Tsutsumi and T. Takaishi, Stud. Surf. Sci. Caral., 1989, 46, 779. 119. G. Engelhardt, S. Luger, J. C. Buhl and J. Felsche, Zeolites, 1989, 9, 182. 120. G. Engelhardt, Stud. Surf. Sci. Catal., 1989, 52, 151.

180

H. PFEIFER AND H. ERNST

G. Engelhardt and H. Van Koningsveld, Zeolites, 1990, 10, 650. G. Engelhardt, J. Felsche and P. Sieger, J . Am. Chem. SOC., 1992,114,1173. H. Ernst and H. Pfeifer, Z. Phys. Chem. (Leipzig), 1990, 271, 1145. H. Ernst, D. Freude, M. Hunger and H. Pfeifer, Stud. Surf. Sci. Catal., 1991, 65, 397. G. V. Eshevskii, V. M. Nekipelov, K. G. Ione and K. I. Zamaraev, React. Kinet. Catal. Left., 1987, 33, 233. 126. C. Fernandez, F. Lefebvre, J. B. Nagy and E. G. Derouane, Stud. Surf. Sci. Catal., 1988, 37, 223. 127. G. Finger and J. Kornatowski, Zeolites, 1990, 10, 615. 128. R. X. Fischer, W. H. Baur, R. D Shannon, R. H. Staley, L. Abrams and A. J. Vega, Acta Crystallogr., 1988, B44, 321. 129. D. Freude, J. Haase and J. Klinowski, Chem. Phys. Lett., 1987, 133,491. 130. D. Freude, M. Hunger and H. Pfeifer, Z. Phys. Chem. (Neue Folge), 1987, 152, 171. 131. D. Freude, E. Brunner, H. Pfeifer, D. Prager, H. G. Jerschkewitz, U. Lohse and G. ONmann, Chem. Phys. Lea., 1987, 139, 325. 132. D. Freude, H. Ernst, M. Hunger, H. Pfeifer and E. Jahn, Chem. Phys. Lett., 1988, 143, 477. 133. D. Freude, J. Klinowski and H. Hamdan, Chem. Phys. Lett., 1988, 149, 355. 134. V. Fdop, G. Borbely, H. K. Beyer, S. Ernst and J. Weitkamp, J . Chem. SOC..Faraday Trans. I, 1989, 85, 2127. 135. C. A. Fyfe, H. Strobl, G. T. Kokotailo, G. J. Kenned and G . E. Barlow, J. Am. Chem. SOC., 1988, 110, 3373. 136. C . A. Fyfe, H. Strobl, G. T. Kokotailo, C. T. Pasztor, G. E. Barlow and S. Bradley, Zeolites, 1988, 8, 132. 137. C. A. Fyfe, G. T. Kokotailo, H. Strobl, H. Gies, G. J. Kennedy, C. T. Pasztor and G. E. Barlow, Stud. Surf. Sci. Catal., 1989, 46, 827. 138. C. A. Fyfe, H. Gies, G. T. Kokotailo, Y. Feng, H. Strobl, B. Marler and D. E. Cox, Stud. Surf. Sci. Catal., 1989, 49, 545. 139. C. A. Fyfe, H. Gies, G. T. Kokotailo, C. Pasztor, H. Strobl and D. E. Cox, J. Am. Chem. SOC., 1989, 111,2470. 140. C. A. Fyfe, G. T. Kokotailo, H. Strobl, C. S. Pasztor, G. Barlow and S. Bradley, Zeolites, 1989, 9, 531. 141. C. A. Fyfe, H. Gies and Y. Feng, J. Chem. SOC.,Chem. Commun., 1989,1240. 142. C. A. Fyfe, H. Gies and Y. Feng, J. Am. Chem. Soc., 1989,111,7702. 143. C. A. Fyfe, H. Grondey, Y. Feng and G. T. Kokotailo, Chem. Phys. Lett., 1990, 173, 211. 144. C. A. Fyfe, H. Grondey, Y. Feng and G. T. Kokotailo, J. Am. Chem. SOC.,1990, 112, 8812. 145. C. A. Fyfe, H. Gies, Y. Feng and H. Grondey, Zeolites, 1990, 10, 278. 146. C. A. Fyfe, Y. Feng, H. Gies, H. Grondey and G. T. Kokotailo, 1. Am. Chem. SOC., 1990, 112,3264. 147. C. A. Fyfe, H. Gies, G. T. Kokotailo, B. Marler and D. E. Cox, J . Phys. Chem., 1990, 94, 3718. 148. C . A. Fyfe, Y.Feng, H. Grondey, G. T. Kokotailo and A. Mar, J. Phys. Chem., 1991, 95,3747. 149. C. A. Fyfe, H. Grondey, Y.Feng, G. T. Kokotailo, S. Ernst and J. Weitkamp, Zeolites, 1992, 12, 50. 150. Z. Gabelica and J. L. Guth, Stud. Surf. Sci. Catal., 1989,49,421. 151. Z. Gabelica, N. Dewaele, L. Maistriau, J. B. Nagy and E. G. Derouane, ACS Symp. Ser., 1989, 398, 518. 152. T. R. Gaffney, R. Pierantozzi and M. R. Seger, ACS Symp. Ser., 1989,398, 374. 153. J. P. Gilson, G. C. Edwards, A. W. Peters, K. Rajagopatan, R. F. Wormsbecher, T. G. Roberie and M. P. Shatliock, J. Chem. SOC., Chem. Commun., 1987,91.

121. 122. 123. 124. 125.

NMR STUDIES OF ZEOLITES

181

154. D. M. Ginter, C. J. Radke and A. T. Bell, Stud. Surf. Sci. Cutul., 1989,49, 161. 155. M. Goepper, F. Guth, L. Delmotte, J. L. Guth and H. Kessler, Stud. Surf. Sci. Curd., 1989,49,857. 156. M. Goepper and J. L. Guth, Zeolites, 1991, 11, 477. 157. G. Goldfarb, Zeolites, 1989,9,509. 158. P. J. Grobet, H. Geerts, J. A. Martens and P. A. Jacobs, J . Chem. Soc., Chem. Commun., 1987, 1688. 159. P. J. Grobet, H. Geerts, M. Tielen, J. A. Martens and P. A. Jacobs, Stud. Surf. Sci. card., 1989,46,721. 160. P. J. Grobet, H. Geerts, J. A. Martens and P. A. Jacobs, Stud. Surf.Sci. Catul., 1989,52, 193. 161. P.J. Grobet, J. A. Martens, I. Balakrishnan, M. Mertens and P. A. Jacobs, Appl. Cutul., 1989,56,L21. 162. P. J. Grobet, A. Samoson, H. Geerts, J. A. Martens and P. A. Jacobs, 1. Phys. Chem., 1991,95,9620. 163. H.Hamdan and J. Klinowski, Chem. Phys. Lett., 1987, 139,576. 164. H.Hamdan, B. Sulikowski and J. Klinowski, J . Phys. Chem., 1989,93,350. 165. H. Hamdan and J. Klinowski, J. Chem. SOC.,Chem. Commun., 1989,240. 166. H.Hamdan and J. Klinowski, ACS Symp. Ser., 1989,398,448. 167. D. Hasha, L. S. de Saldarriaga, C . Saldarriaga, P. E. Hathaway, D. F. Cox and M. E. Davis, J. Am. Chem. SOC.,1988,110, 2127. 168. S. Hayashi, K. Hayamizu and 0. Yamamoto, Bull. Chem. SOC.Jpn., 1987,60, 105. 169. Y. He, C. Li and R. Min, Stud. Surf Sci. Cutal., 1989,49,189. 170. C.P.Herrero, L. Utrera and R. Ramirez, Chem. Phys. Lett., 1991, 183, 199. 171. C.P. Herrero, J. Phys. Chem., 1991,95,3282. 172. M. Hunger, D. Freude, H. Pfeifer and W. Schwieger, Chem. Phys. Lett., 1990, 167,21. 173. R. Jelinek, B. F. Chmelka, Y. Wu, P. J. Grandinetti, A. Pines, P. J. Barrie and J. Klinowski, J . Am. Chem. SOC.,1991,113,4097. 174. P.S. Iyer, J. Scherzer and Z. C. Mester. ACS Symp. Ser., 1988,368,48. 175. H. S. Jacobsen, P. Norby, H. Bildsoe and H. J. Jacobsen, Zeolites, 1989,9,491. 176. E.Jahn, D. Muller, W. Wiecker and J. Richter-Mendau, Zeolites, 1989,9,177. 177. E. J a b , D. Miiller and K. Becker, Zeolites, 1990, 10, 151. 178. R. Janssen, G. A. H. Tijink, W. S. Veeman, T. L. M. Maesen and J. F. van Lent, J . Phys. Chem., 1988,93,899. 179. J. Jeanjean, L. Aouali, D. Delafosse and A. Dereigne, J . Chem. SOC., Furuduy Trans. 1 , 1989,85, 2771. 180. R. Jelinek, B. F. Chmelka, Y. Wu, P. J. Grandinetti, A. Pines, P. J. Barrie and J. Klinowski, J. Am. Chem. SOC., 1991, 113,4097. 181. L. Jian, T. Chang-yu and D. Quing-nian, J. Cutul., 1990, 11, 462. 182. P. N.Joshi, A. Thangaraj and V. P. Shiralkar, Zeolites, 1991, 11, 164. 183. L. Kellberg, M.Linsten and H. J. Jakobson, Chem. Phys. Lett., 1991,182, 120. 184. P. B. Kempa, G. Engelhardt, J.-C. Buhl, J. Felsche, G. Harvey and C. Baerlocher, Zeolites, 1991,11, 558. 185. A. P. M. Kentgens, C. R. Bayense, J. H. C. Vanhooff, J. W. Dehaan and L. J. M. Vandeven, Chem. Phys. Lett., 1991,176,399. 186. H. Kessler, J. M. Chezeau, J. L. Guth, H. Strub and G. Coudurier, Zeolites, 1987,7,360. 187. J. Klinowski, T. A. Carpenter and L. F. Gladden, Zeolites, 1987,7 , 73. 188. J. Klinowski, Srud. Surf. Sci. Curul., 1989,52,39. 189. J. Klinowski, H. Hamdan, A. Corma, V. Fornes, M. Hunger and D. Freude, Card Left., 1989,3, 263. 190. J. Kliiowski and M. W. Anderson, Mugn. Reson. Chem., 1990,28, 68. 191. G.T. Kokotailo, C. A. Fyfe, G. J. Kennedy, G. C. Gobbi, H. Strobl, C. T. Pasztor, G. E. Barlow and S . Bradley, Pure Appl. Chem., 1987,58, 1367.

182

H. PFEIFER AND H. ERNST

192. W. Kolodziejski, P. J. Barrie, H. Y. He and J. Klinowski, J . Chem. Soc., Chem. Commun., 1991,961. 193. B. Kraushaar-Czametzki, thesis, Eindhoven, 1989. 194. G. H. Kuehl and K. D. Schmidt, Zeolites, 1990, 10, 2. 195, K. P. Lillerud, Zeolites, 1987, 7, 14. 196. K. P. Lillerud and M. Stockner, Stud. Surf. Sci. Cutuf., 1989, 46,769. 197. W. Lutz, U. Lohse, B. Fahlke and T. I. Titova, Cryst. Res. Technol., 1988, 23, 925. 198. L. Maistriau, Z. Gabelica, E. G. Derouane, E. T. C. Vogt and J. Vanoene, Zeolites, 1991, 11, 583. 199. P. P. Man and J. Klinowski, Chem. Phys. Lett., 1988, 147, 581. 200. P. P. Man and J. Klinowski, J. Chem. SOC., Chem. Commun., 1988,1291. 201. P. P. Man, M. Briend, M. J. Peltre, A. Larny, P. Beaunier and D. Barthomeuf, Zeolites, 1991, 11, 563. 202. J. A. Martens, C. Jansens, P. J. Grobet, H. K. Beyer and P. A. Jacobs, Stud. Surf. Sci. Cutul., 1989, 49, 215. 203. J. A. Martens, P. J. Grobet and P. A. Jacobs, J. Cutul., 1990, 126, 299. 204. J. A. Martens, H. Geerts, P. J. Grobet and P. A. Jacobs, J. Chem. Soc., Chem. Commun., 1990, 1418. 205. J. A. Martens, E. Feijen, J. L. Lievens, P. J. Grobet and P. A. Jacobs, J. Phys. Chem., 1991,95, 10025. 206. P. Massiani, B. Chauvin, F. Fajula and F. Figueras, Appl. Cutul., 1988, 42, 105. 207. P. Massiani, F. Fajula, F. Figueras and J. Sanz, Zeolites, 1988, 8, 332. 208. V. M. Mastikhin, N. V. Klyueva, K. G. Ione and K. I. Zarnaraev, React. Kinet. Catuf. Lett., 1987,33,465. 209. R. H. Meinhold and D. M. Bibby, Zeolites, 1990, 10, 146. 210. R. H. Meinhold and D. M. Bibby, Zeolites, 1990, 10, 74. 211. M. T. Melchior and J. M. Jansen, Stud. Surf. Sci. Cutul., 1989, 49, 805. 212. B. L. Meyer, T. H. Fleisch, G. J. Roy, J. T. Miller and J. B. Hall, J. Carul., 1988, 110, 82. 213. J. B. Nagy, P. Bodart, H. Collette, A. A. J. El-Hagej, Z. Gabelica, R. Aiello, A. Nastro and C. Pellegrino, Zeolites, 1988, 8, 209. 214. J. B. Nagy, P. Bodart, H. Collette, C. Femandes, Z. Gabelica, A. Nastro, and R. Aiello, J. Chem. SOC., Furuduy Trans. 1 , 1989, 85, 2749. 215. J. M. Newsam, J. Phys. Chem., 1987, 91, 1259. 216. J. M. Newsarn, J. Chem. Soc., Chem. Commun., 1987, 123. 217. N. C. Nielsen, H. Bildsoe, H. J. Jakobsen and P. Norby, Zeolites, 1991, 11, 622. 218. A. F. Ojo, J. Dwyer, J. Dewing and K. Karim, 1. Chem. SOC.,Faruday Trans., 1991,87, 2679. 219. A. F. Ojo, J. Dwyer, J. Dewing, P. J. O’Malley and A. Nabhan, 1. Chem. Soc., Furuduy Trans., 1992, 88, 105. 220. R. J. Pellet, C. S. Blackwell and J. A. Rabo, J. Cutul., 1988, 114, 71. 221. J. Perez-Pariente, J. Sanz, V. Fornes and A. Corrna, 1. Cutul., 1990, 124, 217. 222. S. Prabakar, K. J. Rao and C. N. R. Rao, Muter. Res. Bull., 1991, 26, 805. 223. S. L. Qiu, W. T. Tian, W. Q. Pang, T. Sun and D. Z. Jiang, Zeolires, 1991, 11, 371. 224. G. J. Ray, B. L. Meyers and C. L. Marshall, Zeolites, 1987,7,307. 225. G. J. Ray, A. G. Nerheirn and J. A. Donohue, Zeolites, 1988, 8,458. 226. W. Reschetilowski, W.D. Einicke, M. Jusek, R. Schollner, D. Freude, M. Hunger and J. Klinowski, Appl. Cutul., 1989, 56, L15. 227. J. Rocha, S. W. Carr and J. Klinowski, Chem. Phys. Lett., 1991, 187, 401. 228. J. Rocha, X. S. Liu and J. Klinowski, Chem. Phys. Lett., 1991, 182, 531. 229. J. Rocha and J. Klinowski, J . Chem. SOC., Chem. Commun., 1991, 1121. 230. F. Rossner, K. H. Steinberg, D. Freude, M. Hunger and H. Pfeifer, Stud. Surf. Sci. Cutaf., 1989, 46,421.

NMR STUDIES OF ZEOLITES

183

231. T. Romotowski, J. Komorek, V. M. Mastikhin and A. V. Nosov, Zeolites, 1991, 11, 491. 232. L. S. Saldamaga, C. Saldamaga and M. E. Davis, J . Am. Chem. Soc., 1987, 109,2686. 233. T. Sano, K. Suzuki, H. Shoji, S. Ikai, K. Okabe, T. Murakami, S. Shin, H. Hagiwara and H. Takaya, Chem. Lett., 1987.7, 1421. 234. T. Sano, K. Suzuki, H. Okado, K. Fujisawa, K. Kawamura, S . Ikai, H. Hagiwara and H. Takaya, Stud. Sud. Sci. Catal., 1987, 34, 613. 235. J. Sanz, V. Fornes and A. Coma, J. Chem. Soc., Faraday Trans. 1 , 1988,84, 3113. 236. M.Sawa, M. Niwa and Y. Murakami, Zeolites, 1990, 10,532. 237. W. Schwieger, K. H. Bergk, D. Freude, M. Hunger and H. Pfeifer, ACS Symp. Ser., 1989, 398,275. 238. G. Seo and R. Ryoo, J. Catal., 1990, 124,224. 239. M. Stocker, S. Ernst, H. G. Karge and J. Weitkamp, Acta Chem. Scand., 1990,44, 519. 240. M. Stoecker, D. Akporiaye and K.-P. Lillerud, Appl. Catal., 1991, 69, L7. 241. B. Sulikowski, J. Rakoczy, H. Hamdan and J. Klinowski, J. Chem. Soc., Chem. Commun., 1987, 1542. 242. B. Sulikowski and J. Klinowski, J. Chem. Soc., Faraday Trans., 1990, 86, 199. 243. B. Sulikowski, G. Borbely, H. K. Beyer, H. G. Karge and I. W. Mishin, J. Phys. Chem., 1989,93,3240. 244, T. Takaishi, J . Chem. SOC., Faraday Trans. 1 , 1988, 84, 2967. 245. H. K. C. Timken and E. Oldfield, J. Am. Chem. SOC., 1987, 109,7669. 246. B. H. Toby, M. M. Eddy, C. A. Fyfe, G. T. Kokotailo, H. Strobl and D. E. Cox, J. Mater. Res., 1988, 3, 563. 247. B. Unger, W. Reschetilowski, K. P. Wendlandt and D. Freude, Z. Chemie, 1988,28, 194. 248. A. J. Vega and Z. Luz, J. Phys. Chem., 1987, 91, 365. 249. Q. L. Wang, G. Giannetto, M. Torrealba, G. Perot, C. Kappenstein and M. Guisnet, J. Caul., 1991, 130, 459. 250. X. Wang, X. Liu, T. Song, J. Hu and J. Qiu, Chem. Phys. Let?., 1989, 157, 87. 251. T. Yashima, K. Yamagishi, S. Namba, S. Nakata and S. Asaoka, Stud. Surf. Sci. Catal., 1988, 37, 175. 252. Z , Zhang, X. Liu, Y. Xu and R. Xu, Zeolites, 1991, 11, 233. 253. B. Zibrowius, U. Lohse and J. Richter-Mendau, J. Chem. Soc., Faraday Trans., 1991,87, 1433. 254. B. Zibrowius, E. Loffler and M. Hunger, Zeolites, 1992, 12, 167. 255. H.-L. Zubowa, E. Alsdorf, R. Fricke, F. Neissendorfer, J. Richter-Mendau, E. Schreier, D. Zeigan and B. Zibrowius, J. Chem. SOC.,Faraday Trans., 1990, 86, 2307. 256. C. A. Fyfe, H. Grondey, K. T. Mueller, K. C. Wong-Moon and T. Markus, J. Am. Chem. Sac., 1992, 114,5876. 257. C. A. Fyfe, K. T. Mueller, H. Grondey and K. C. Wong-Moon, Chem. Phys. Lett., 1992, 199, 198. 258. C . A. Fyfe, H. Gies, Y. Feng and G. T. Kokotailo, Nature (Lond.), 1989, 341, 223. 259. W. Reschetilowski, W.-D. Einicke, B. Meier, E. Brunner and H. Ernst, Stud. Surf Sci. Catal., 1991, 65, 529. 260. W. Reschetilowski, W.-D. Einicke, B. Meier, E. Brunner and H. Ernst, Srud. Surf. Sci. Catal., 1991, 69, 119. 261. J. E. Bartmess, J. A. Scott and R. T. McIver, J. Am. Chem. Soc., 1979, 101, 6046. 262. J. Sauer, J. Mol. Cat., 1989, 54, 312. 262a. U. Fleischer, W. Kutzelnigg, A. Bleiber and J. Sauer, J. Am. Chem. Soc., 1993, 115, 7833. 263. J. P. Chauvel and N. S. True, Chem. Phys., 1985, 95, 435. 264. R. W. Taft, I. A. Koppel, R. D. Topsom and F. Anvia, J. Am. Chem. Soc., 1990, 112, 2047. 265. R. T. Sanderson, Chemical Bonds and Bond Energy, Academic Press, New York, 1976. 266. H. Pfeifer, Colloids and Surfaces, 1989, 36, 312.

184 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306.

H. PFEIFER AND H. ERNST

D. Freude, M. Hunger, H. Pfeifer and W. Schwieger, Chem. Phys. Left., 1986, 128, 62. H. Pfeifer, D. Freude and J. Karger, Stud. Surf. Sci. Catal., 1991, 65, 89. H. Emst, Z. Phys. Chemie (Leipzig), 1988,269, 1073. E. Lippmaa, A. V. Samoson, V. V. Brey and Y. I. Gorlov, Doklady Akad. Nauk, 1981, 259,403. P. A. Jacobs and W. J. Mortier, Zeolites, 1982, 2, 226. H. Pfeifer, D. Freude and M. Hunger, Zeolites, 1985, 5, 274. V. M. Mastikhin, I. L. Mudrakovsky and A. V . Nosov, Bruker Report, 1989, 2, 18. M. Hunger, Thesis (Prom. B), University of Leipzig, 1991. H. Pfeifer, W. Meiler and D. Deininger, in Annual Reports on NMR Spectroscopy, vol. 21 (ed. G. A. Webb), p. 291. Academic Press, London, 1983. E. Brunner, J. Chem. SOC.,Faraday Trans., 1990,86,3957. E. Brunner, D. Freude, B. C. Gerstein and H. Pfeifer, J. Magn. Reson., 1990, 90, 90. J. Bohm, D. Fenzke and H. Pfeifer, J. Magn. Reson., 1983,55, 197. S. Halik, J. A. Lercher and H. Mayer, J. Chem. SOC. Faraday Trans. I , 1988, 84, 4457. E. Brunner, H.-G. Karge and H. Pfeifer, Z. Phys. Chemie, 1992, 176, 173. J. Dwyer, J. Dewing, N. E. Thompson, P. J. O’MaUey and K.Karim, 1. Chem. SOC., Chem. Commun., 1989,843. R. R. Emst, G. Bodenhausen and A. Wokaun, Principles of Nuclear Magnetic Resonance in One and Two Dimensions. Clarendon Press, Oxford, 1986. D. Fenzke, M. Hunger and H. Pfeifer, J. Magn. Reson., 1991, 95, 477. K.-P. Schroder, J. Sauer, M. Leslie, C. R. A. Catlow and J. M. Thomas, Chem. Phys. Len., 1992, 188,320. M. Czjzek, H. Jobic, A. N. Fitch and T. Vogt, J. Phys. Chem., 1992.96, 1535. A. Abragam, The Principles of Nuclear Magnetism. University Press, Oxford, 1961. T. J. Gluszak, D. T. Chen, S. B. Sharma, J. A. Dumesic and T. W. Root, Chem. Phys. Lett., 1992, 190, 36. H. Rosenberger, H. Emst, G. Scheler, I. Junger and R. Sonnenberger, 2. phys. Chemie (Leipzig), 1982, 263, 864. M. Hunger, D. Freude, T. Frohlich, H. Pfeifer and W. Schwieger, zeolites, 1987,7, 108. A. E. Hirschler, J. Catal., 1%3,2,428; C. J. Plank, Proceedings of the Third International Congress on Catalysis, Amsterdam, 1964, 1,568. M. Hunger, D. Freude, H. Pfeifer, D. Prager and W. Reschetilowski, Chem. Phys. Lett., 1989,163,221. J. Sauer, Proc. Int. Symp. Zeolite Catal., Siofok (Hungary), 1985, 19. J. Sauer and P. Hobza, Theor. Chim. Acta (Berlin), 1984, 65, 279 and 291. M. Hunger, D. Freude and H. Pfeifer, J. Chem. SOC.,Faraday, Trans., 1991.87, 657. L. M. Kustov and V. B. Kazanski, J. Chem. SOC., Faraday Trans., 1991, 87, 2675. L. M. Kustov, V. B. Kazanski, S. Beran, L. Kubelkova and P. Jim, I . Phys. Chem., 1987, 91, 5247. P. Batamack, C. Doremieu-Morin, J. Fraissard and D. Freude, J. Phys. Chem., 1991, 95, 3790. J. Dwyer, Stud. Surf. Sci. Catal., 1988, 37, 333. H. G. Karge, Stud. Surf. Sci. Catal., 1991, 65, 133. E. Lippmaa, A. Samoson and J. Magi, J. Am. Chem. SOC.,1986, 108, 1730. D. Miiller, W. Gessner, H. J. Behrens and G. Scheler, Chem. Phys. Lett., 1981, 79, 59. J. Klinowski, C. A. Fyfe and G. C. Gobbi, J . Chem. Soc., Faraday Trans. I , 1985, 81, 3003. D. E. Woessner and H. K. C. Timken, 1. Magn. Reson., 1990, 90, 411. G. E. Maciel, J. F. Haw, I. Chuang, B. L. Hawkins, T. A. Early, D. R. McKay and L. Petrakis, J. Am. Chem. SOC.,1983, 105, 5529. J. A. Ripmeester, 1. Am. Chem. SOC.,1983, 105, 2925. P. D. Majors and P. D. EUis, J. Am. Chem. SOC.,1987, 109, 1648.

NMR STUDIES OF ZEOLITES

185

307. I). Schulze, H. Ernst, D. Fenzke, W. Meiler and H. Pfeifer, J. Phys. Chem., 1990, 94, 3499. 308. J. H. Lunsford, W. P. Rothwell and V. Shen, J. Am. Chem. SOC., 1985, 107, 1540. 309. L. Baltusis, J. S. Frye and G. E. Maciel, J . Am. Chem. Soc., 1987, 109, 40. 310. A. Michael, W. Meiler, D. Michel and H. Pfeifer, Chem. Phys. Lett., 1981, 84, 30. 311. A. Michael, W. Meiler, D. Michel, H. Pfeifer, D. Hoppach and J. Delmau, J. Chem. Soc., Faraday Trans. I , 1986, 82, 3053. 312. E. Brunner, H. Pfeifer, T. Wutsacherk and D. Zscherpel, Z . Phys. Chemie, 1992, 178, 173. 313. V. M. Mastikhin, J. L. Mudrakowsky and S. V. Filimonova, Chern. Phys. Lett., 1988, 149. 175. 314. J. H. Lunsford, P. N. Tutunjian, P. Chu, E. B. Yeh and D. J. Zalewski, J . Phys. Chem., 1989, 93, 2590. 315. R. M. Barrer, Endeavour, 1964, 23, 122. 316. D. Geschke, Z. Phys. Chem., Leipzig, 1972, 249, 125. 317. D. Michel, Z. Phys. Chem., Leipzig, 1973, 252, 263. 318. D. Michel, W. Meiler and H. P€eifer, J. Mof. Car., 1975/76, 1, 85. 319. D. Doskocilova and B. Schneider, Chem. Phys. Lett., 1970, 6 , 381. 320. S. Sekine, A. Kubo and H. Sano, Chem. Phys. Lett., 1990,171,155. 321. M. W. Anderson and J. Klinowski, Nurure (Lond.), 1989, 339, 200. 322. J. F. Haw, B. R.Richardson, J. S. Oshiro, N. D. Lazo and J. A. Speed, J . Am. Chem. Soc., 1989, I l l , 2052. 323. H. Ernst and H. Pfeifer, J . Catal., 1992, 136, 202. 324. R. H. Meinhold, L. M. Parker and D. M. Bibby, Zeolites, 1986, 6 , 491. 325. I. Kustanovich, D. Fraenkel, 2. Luz and S. Vega, J. Phys. Chem., 1988, 92, 4134. 326. W. Meiler and T. Weller, Z. Chemie, 1982, 22, 382. 327. T. Weller, H.-J. Kohler, R. Lochmann and W. Meiler, J. Mol. Struct./lhermochem., 1982, 90, 81. 328. W. Meiler and T. Wutscherk, Isotopenpraxis, 1989, 25, 41. 329. M. W. Anderson, P. J. Barrie and J. Klinowski, J. Phys. Chem., 1991, 95, 235. 330. Unpublished 331. H. Herden, W. Meiler and W. Robien, Chemical Monthly, 1988, 119, 913. 332. K. I. Zamaraev and V. M. Mastikhin, Coll. Surf., 1984, 12, 401. 333. E. J. Munson, N. D. Lazo, M. E. Moellenhoff and J. F. Haw, J. Am. Chem. Soc., 1991, 113, 2783. 334. E. J. Munson and J. F. Haw, J . Am. Chem. Soc., 1991, 113, 6303. 335. M. W. Anderson and J. Klinowski, J. Chem. SOC., Chem. Commun., 1990, 918. 336. G . Mirth, J. A. Lercher, M. W. Anderson and J. Klinowski, J. Chem. Soc., Faraday Trans., 1990, 86,3039. 337. E. G. Derouane and J. B. Nagy, ACS Sympos. Ser., 1984,248, 101. 338. 3.-P. Lange, A. Gutsze, J. Allgeier and H. G. Karge, Appl. Catal., 1988.45, 345. 339. K. P. Datema, A. K. Nowak, J. van Braam-Houckgeest and A. F. H. Wielers, Catal. Lerr., 1991, 11, 267. 340. V. Bosacek and L. Kubelkova, Zeolites, 1990, 10, 64. 341. M. Kanowski, Diploma thesis, Leipzig, 1990. 342. N. D. Lazo, J. L. White, E. J. Munson, M. Lambregts and J. F. Haw, J. Am. Chem. SOC., 1990, 112,4050. 343. V. M. Mastikhin, I. L. Mudrakovsky and A. V. Nosov, Progr. in NMR Spectroscopy, 1991, 23,259. 344. C. Tsiao, D. R. Corbin and C. Dybowski, J. Am. Chem. Soc., 1990, 112, 7140. 345. A. J. Vega and Z. Luz, Zeolires, 1988, 8 , 19. 346. D. M. Ruthven: Principles of Adsorption and Adsorption Processes. John Wiley & Sons, New York. 1984.

186 347. 348. 349. 350. 351. 352. 353. 354. 355. 356. 357. 358. 359. 360. 361. 362. 363. 364. 365. 366. 367. 368. 369. 370. 371. 372. 373. 374. 375. 376. 377. 378. 379. 380. 381. 382. 383. 384. 385. 386.

H. PFEIFER AND H. ERNST E. 0. Stejskal and J. E. Tanner, J . Chem. Phys., 1965, 42, 288. J. Karger, H. Pfeifer and W. Heink, Adv. M a p . Reson., 1988, 12, 1. J. Karger and H. Pfeifer, Zeolites, 1987, 7, 90. J. Karger and H. Pfeifer, in A. Pines and A. Bell (eds): NMR and Catalysis. Marcel Dekker, Inc., New York, in press. 3. Karger, H. Pfeifer and M. Biilow, Z. Chemie, 1976, 16, 85. J. Karger and J. Caro, J. Chem. SOC., Faraday Tram. 1 , 1977, 73, 1363. J. E. Tanner and E. 0. Stejskal, 1. Chem. Phys., 1968,49, 1769. D. M. Ruthven and L.-K. Lee, AlChE J., 1981, 27,654. H.-J. Doelle and L. Riekert, ACS Symp. Ser., 1977,40, 401. H. Jobic, M. Bee, J. Caro, M. Biilow and J. Karger, J . Chem. SOC.,Faraday Trans. 1 , 1989, 85, 4201. H. Jobic, M. Bee, J. Karger, H. Pfeifer and J. Caro, J. Chem. SOC.,Chem. Commun., 1990, 341. H. Jobic, M. BBe, J. Caro, M. Biilow, J. Karger and H. Pfeifer, Stud. Surf. Sci. Catal., 1991, 65, 445. J. Karger and D. M. Ruthven, J. Chem. SOC.,Faraday Trans. I , 1981, 77, 1485. A. Germanus, J. Karger, H. Pfeifer, N. N. Samulevic and S . P. Zdanov, Zeolites, 1985,5, 91. M. Eic, N. V. Goddard and D. M. Ruthven, Zeolites, 1988,8, 327. J. Karger and D. M. Ruthven, Zeolites, 1989, 9, 267. J. Caro, M. Biilow, W. Schinner, J. Karger, W. Heink, H. Pfeifer and S. P. Zdanov, J . Chem. SOC.,Faraday Trans. 1 , 1985,81,2541. J. Karger, H. Pfeifer and R. Haberlandt, J. Chem. SOC., Faraday Tram. I , 1989,76, 717. H. Pfeifer, Phys. Rep., 1976, 26C, 293. J. Karger, H. Pfeifer, M. Rauscher and A. Walter, J . Chem. SOC., Faraday Trans. 1 , 1980, 76, 717. J. Karger and H. Pfeifer, Zeolites, 1992, 12, 872. U. Hong, J. Karger, H. Pfeifer, U. Miiller and K. K. Unger, Z. Phys. Chem., 1991, 173, 225. U. Hong, J. Karger, R. Krarner, H. Pfeifer, G. Seiffert, U. Muller, K. K. Unger, H.-B. Luck and T. Ito, Zeolites, 1991, 11, 816. P. Demontis, E. S. Fois, G.-B. Sufritti and S . Quartieri, 1. Phys. Chem., 1990, 94,4329. S. J. Goodbody, K. Watanabe, D. MacGowan, J. P. B. Walton and N. Quirke, 1. Chem. SOC., Faraday Tram., 1991, 87, 1951. R. L. June, A. T. Bell and D. N. Theodorou, 1. Phys. Chem., 1990,94,8232. H. Pfeifer, NMR Basic Principles and Progress, 1972, 7, 53. J. Karger, AIChE J., 1982, 28,417. Chr. Forste, J. Karger, H. Pfeifer, L. Riekert, M. Biilow and A. Zikanova, J. Chem. Soc., Faraday Trans., 1990, 86, 881. J. Karger and H. Pfeifer, ACS Syrnp. Ser., 1988, 368,376. J. Karger, H. Pfeifer, F. Stallmach, M. Biilow, P. Stnwe, R. Entner. H. Spindler and R. Seidel, AIChE J., 1990, 36, 1500. H. Pfeifer, Coll. Surf.,1990, 45, 1. J. R. Lyerla, C. S. Yannoni and C. A. Fyfe, Acc. Chem. Res., 1982, 15, 208. J. F. Haw, Anal. Chem. 1988, 60,559A. J. F. Haw and J. A. Speed, J. M a p . Reson., 1988, 78, 344. E . G. Derouane, J. P. Gilson and J. B. Nagy, Zeolites, 1982, 2, 42. J. Weitkamp and S . Meixner, Zeolites, 1987, 7, 6. R. H. Meinhold and D. M. Bibbi, Zeolites, 1990, 10, 121. B. R. Richardson and J. F. Haw, Anal. Chem., 1989, 61, 1821. J. Karger, H. Pfeifer, J. Caro, M. Biilow, H. Schlodder, R. Mostowicz and J. Volter, Appl. Catal., 1987, 29, 21.

NMR STUDIES OF ZEOLITES 387. 388. 389. 390. 391. 392. 393. 394. 395. 396. 397. 398. 399. 400. 401.

187

J. Fraissard and T. Ito, Zeolites, 1988, 8, 350. M. Zardkoohi, J. F. Haw and J. H. Lunsford, J . Am. Chem. Soc., 1987, 109,5278. M. W. Anderson and J. Klinowski, J. Am. Chem. SOC., 1990, 112, 10. M. W. Anderson and J. Klinowski, Chem. Phys. Lett., 1990, 172, 275. J. L. White, N. D. Lazo, B. R. Richardson and J. F. Haw, J . Catal., 1990, 125, 260. B. R. Richardson, N. D. Lazo, P. D. Schettler, J. L. White and J. F. Haw, J. Am. Chem. SOC., 1990, 112, 2886. R. Meusinger, W. Meiler and H. Ernst, 13. Diskussionstagung der Gesellsch. dt. Chemiker, Fachgruppe Magnetische Resonanzspektroskopie, Todtmoos, 1991. A. Bax, Two-dimensional Nuclear Mugnetic Resonance in Liquids, Thesis, Delft, 1981. E. G. Derouane, J. B. Nagy, P. Dejaifeve, 3. H. C. van Hooff, B. P. Spekman, J. C. V6drine and C. Naccache, J . Catal., 1978, 53, 40. J. B. Nagy, J. P. Gilson and E. G. Derouane, J. Mol. Catal., 1979, 5 , 393. C. E. Bronnimann and G. E. Maciel, J . Am. Chem. SOC., 1980, 108,7154. F. Fetting and U. Dingerdissen, Statusseminar im Fritz-Haber-Institut der Max-PlanckGesellschaft, Berlin-Dahlem, 1990, 61. I. Kustanovich, Z. Luz, S. Vega and A. J. Vega, J. Phys. Chem., 1990, 94, 3138. E. G. Derouane, Catalysis by Acids and Bases (eds B. Imelik et al.). Elsevier, Amsterdam, 1985. M. Guisnet and P. Magnoux, Appl. Catal., 1989, 54, 1.

This Page Intentionally Left Blank

NMR Studies of Higher-order Structures of Solid Polymers H. KUROSU,* S. ANDO,' H. YOSHIMIZU* and I. ANDO* *Department of Polymer Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo, Japan TNTT Interdisciplinary Research Laboratories, Midori-cho, Musashino-shi, Tokyo, Japan +Department of Materials Science and Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya, Japan

Introduction Engineering plastics and high-performance polymers Polymer alloys Natural polymers 4.1. Fibrous proteins 4.2. Membrane proteins 4.3. Miscellaneous biopolymers 5. Conclusion References

1. 2. 3. 4.

189 190 239 251 252 261 268 269 269

1. INTRODUCTION

Polymers are one of the most important discoveries made during the twentieth century. At present we have many kinds of polymer materials with various physical properties and chemical properties. These are due to the vitality of polymer research and development of a diversity of interests on polymer materials. In order to develop new polymer materials, polymer design has been done on the basis of advanced polymer science and technology. The properties of polymers are closely related to their structures. For this, the establishment of methods for determining the structures is very important for obtaining reliable polymer design and for developing new advanced polymers. Since the first observation of a high-resolution 'H NMR spectrum of uncured Heva rubber in CS2 solution in 1957 by Gutowsky et d.,' high-resolution NMR spectroscopy has developed to become the most ANNUAL REPORTS ON Nh4R SPECTROSCOPY VOLUME 28 ISBN 0-12-505328-2

Copyright 0 1994 Academic Press Limifed A// righrs of reproduction in any form reserved

190

H.KUROSU,S . ANDO, H.YOSHIMIZU A N D I. A N D O

powerful method available for characterizing structures of polymers in the solution and solid states and for analysing dynamic aspects of their structures.2 In this chapter, the most recent high-resolution NMR studies of solid polymers, including synthetic and natural polymers, are reviewed, with the emphasis being placed on revealing higher-order structures. The present review complements previously published reviews2 and provides a new dimension in polymer structures and in addition is concerned with engineering plastics, high-performance polymers, liquid crystalline polymers, polymer alloys, and natural polymers. 2. ENGINEERING PLASTICS AND HIGH-PERFORMANCE POLYMERS Engineering plastics and high-performance polymers are the most popular polymers in polymer materials. Their physical properties and functions are closely related to the structure and dynamics in the solid state. From such a situation, solid-state high-resolution NMR spectroscopy has provided the potential perspective of becoming a powerful means of determining the higher-order structure and dynamics of the polymers in the solid state, associated with physical properties and functions. Polyethylene (PE) is one of the typical engineering plastics. The polymer takes a trans and two gauche conformations, in which the conformational energy difference is small (about 500 cal/mol) and so takes various structures under appropriate conditions. The structure of PE has been studied by several kinds of spectroscopic methods such as NMR, X-ray diffraction, electron diffraction, IR and neutron diffraction. PE takes two kinds of crystal forms. One of them is the orthorhombic form which occurs under normal conditions; the other is the monoclinic form which occurs under high pressure or drawn conditions. In both the crystal forms the conformation takes the all-trans zig-zag conformation, but the chain arrangements are different from each other. The all-trans zig-zag planes in the orthorhombic form are perpendicular to each other, and in the monoclinic form they are parallel to each other as shown in Fig. 1. There also exist non-crystalline and interfacial phases. Their existence strongly affects the physical properties of polymers. The fraction of these phases depends on the preparation conditions. Therefore, it is very important to analyse the detailed structure of PE in the solid state to understand the physical properties of PE. High-resolution solid-state NMR is a powerful tool for investigating these phases separately. Figure 2 shows I3C cross-polarization (CP) magic angle spinning (MAS) NMR spectra of a single crystal (SC) PE,3 melt-quenched (MQ) PE3 and drawn (DR) PE.4 Only one sharp peak at 33.0ppm appears in the SCPE spectrum. On the other hand, MQPE and DRPE have two and three peaks, respectively. This means that MQPE and DRPE have at least two or three

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

191

a)

o HYDROGENATOM Q

P

9

(b) o HYDROGEN ATOM

Fig. 1. Polyethylene structure model: (a) orthorhombic form and (b) monoclinic form.

192

H. KUROSU, S. ANDO, H . YOSHIMIZU AND I. ANDO

Fig. 2. 13CCP MAS-NMR spectra of polyethylene. (a), SC; (b), MQ; (c), DR; PE.

kinds of magnetically inequivalent carbon atoms. VanderHart et al. have assigned these peaks by spin-lattice relaxation time measurements; peak A (31 ppm) comes from the amorphous region which is in a mobile state and peak I (33ppm) comes from the crystalline region which is in an immobile state.’ Further, the high frequency peak of DRPE (35.0 ppm) is assigned to the methylene carbons in the monoclinic crystal region. These assignments were justified by quantum chemical shielding calculations with the tightbinding (TB) sum-over-states (SOS) Kitamaru er al. have reported that peak A comes from the amorphous region and the interfacial region which exists between the crystal and amorphous regions.’ The crystal structure of PE has been proposed from some models such as sharp-fold, switch-board and loose-loops models (Fig. 3).97’0 In order to obtain detailed information on fold structure, Ando et al. have studied SC and MQPE by high-resolution solid-state 13C NMR.3 This study reveals that SC and MQPE take the “adjacent re-entry” type of macroconformation in addition to loose and long loops in the fold surface as shown in Fig. 4. The number of carbon atoms in the trans zig-zag chain from one fold to the next is estimated to be approximately 100, and thus this leads to an estimate of the stem length being about 125A. This magnitude is consistent with the crystal thickness (120-150 A) measured directly for polyethylene single crystals by electron microscopy. Nakai et al. have determined the orientation of the chemical shift tensor of PE by measuring the heteronuclear dipolarkhemical shift two-dimensional

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

193

Fig. 3. Schematic drawing of the conformation models of polyethylene single crystal: (a) sharp-fold model; (b) switch-board model; (c) loose-loop model.

powder pattern." Figure 5 shows experimental and theoretical 13C chemical shift/I3C-'H dipolar two-dimensional (2D) powder patterns of PE. The o1 and q axes represent the 13C--'H dipolar interaction and the 13C chemical shift, respectively. We can determine the C-H distance, angle (LH-C-H) and the directions of the principal values for the chemical shift tensor from this spectrum by fitting with the observed and calculated dipolar/chemical shift 2D powder pattern. These measurements show that the directions of the principal values of the chemical shift tensor have deviations from the ideal orientation of the chemical shift tensor, and that the C-H length and LH-C-H angle have been obtained to be 1.12 k 0.002 8, and 107" k 2", respectively. The physical properties of polymers in the solid state are strongly affected by temperature. Variable-temperature (VT) NMR techniques should provide much useful data on the structural and dynamic aspects of polymers in

194

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

Fig. 4. A schematic illustration of the folded chain conformation for melt-quenched polyethylene: (a) trans zig-zag, (b) sharp fold and (c) loose and long loops.

the solid state.l2.l3 The advantage of VT solid-state NMR techniques is that the temperature dependence of the conformation and molecular motion can be studied by observing the chemical shift. Akiyama et al. have investigated the structure of ultrahigh molecular weight (UHMW) PE in the solid state by VT 13C CP MAS-NMR.14 Figure 6 shows the 13C NMR spectra of UHMW PE measured as a function of temperature, and the determined chemical shift values are listed in Table 1. The 13C CP-MAS spectrum at room temperature has two peaks (the low and high frequency peaks are designated by A and I, respectively). Peak A appears to a low frequency by about 2 ppm from peak I. The 13CNMR chemical shifts of peaks I and A for the UHMW PE agree with those for the melt-crystallized PE. Peaks I and A have been assigned to the crystalline component with the trans zig-zag conformation and to the non-crystalline component respectively. It can be seen from this figure that as the temperature is decreased, peak A shifts to low frequency but peak I shifts to high frequency. The 13C chemical shift behaviour for peak A for the UHMW PE is similar to that of the melt-quenched PE.15 At -108"C, the molecular motion is frozen and the chemical shift for peak A is about 32ppm. Therefore, peak A comes from the methylene carbons in the trans conformation in the non-crystalline component because in the frozen state a methylene carbon in the gauche

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

195

1

2

0

-1 40

30

35

40

35

30

o2Chemical Shifl

t 1

g -1

40

30

35 PPm

Fig. 5. ' C chemical shift/13C-'H dipolar 2D powder patterns of polyethylene: (a) experimental; (b) and (c) calculated.

conformation should appear at about 27 ppm by the ly gauche effect. Such behaviour in the UHMW PE is very similar to that in melt-crystallized PE. Further, the temperature dependence of the I3C chemical shift of peak I is considered. The low frequency shift of peak I is opposite that observed for the case of the melt-crystallized PE, where the chemical shifts are independent of temperature. At -108"C, the chemical shift is about 32 ppm, and peaks I and A coalesce. It is hard to consider that the causes for the low frequency shift of peak I is a change of structure in going from the orthorhombic form to a different crystal structure. It is known that the 13C

196

H.KUROSU, S . ANDO, H. YOSHIMIZU AND I. ANDO

l

40

'

"

'

"

35

'

'

~

'

"

'

30

~

'

'

'

~

.25

ChemicalShiftlppm

Fig. 6. I3C CP MAS-NMR spectra of ultrahigh molecular weight polyethylene as a

function of temperature.

chemical shifts of the methylene carbons with the trans zig-zag structure in orthorhombic, triclinic and monoclinic forms are about 33, 34 and 35 ppm, respectively.'6 From these results, the 13C chemical shift of 32 ppm at -108°C indicates that the crystalline structure is different from that of any of the three kinds of crystal structures found at room temperature. The other possibility for the low frequency shift is the distortion of the orthorhombic form. The a, b and c axis lengths of UHMW PE, determined

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

197

Table 1. 13C chemical shifts of UHMW and MC polyethylenes as a function of

temperature.

13Cchemical shift S (ppm)"

Temperature ("C) UHMW polyethylene" 23 -20 -50 -70 -90 - 108 MC polyethylened 90 60 25 -50 -90 - 120

+ 1rIb

Peak I

Peak A

32.8 (1.O)e 32.4 (1.1) 32.3 (1.2) 32.1 (1.5) 32.0 (1.3) 31.9 (2.0)

31.1 (1.6) 31.1 (2.5) 30.9 (2.5)

32.9 (2.5) 33.2 (0.7) 33.2 (0.5) 33.2 (1.2) 33.2 (1.2) 33.2 (1.5)

31.0 (1.0) 30.9 (0.9) 31.4 (1.7)

0.25 0.24

32.4 (1.0) 32.1 (1.0)

0.30 0.27

IAI(IA

0.67 0.65 0.31

-

31.9 (2.0)

-

0.44

-

"50.2 ppm from TMS. *I,: intensity of peak x . 'This work. dRef. 13. 'Values in parentheses indicate half-height width for deconvoluted peaks.

by X-ray diffraction, are listed in Table 2 together with those of the melt-crystallized PE at room temperature and -88°C.'' The largest difference between UHMW PE and melt-crystallized PE is the chan e of the c axis length. The c axis length of UHMW PE changes from 2.55 to 2.44 A in going from room temperature to -88"C, while that of melt-crystallized PE does not change. Therefore, it is suggested that the I3C chemical shift for the UHMW PE may be influenced by a change in the c axis length. In order to elucidate the origin of the low frequency shift, quantum chemical calculations were performed. The change in c axis may come from a change in the C-C-C angle and the dihedral angle. Therefore, the I3C chemical shift calculations were carried out, using n-decane, by the FPT INDO method as functions of the bond angle C-C-C and the dihedral angle. Figure 7 shows the calculated 13C chemical shift contour map. It has already been reported that the dihedral angle and bond angle for the crystalline PE are 180" and 112", respectively, as deduced from X-ray diffraction. This position is indicated by the open circle in the contour map. When the dihedral angle is varied from 180" to 150", the methylene carbon signal shifts to low frequency with a decrease in the dihedral angle as seen from Fig. 7. This agrees with the experimental finding. From these results, it is suggested

1

198

H. KUROSU, S. ANDO, H. YOSHUlIZU AND I. ANDO

Table 2. Lattice constants of UHMW and MC polyethylene samples as a function of temperature.

Lattice constant (A) Temperature ("C)

a

b

C

7.43 7.31

4.95 4.93

2.55 2.44

7.44 7.25

4.96 4.90

2.53 2.53

~

UHMW polyethylene 20 -88 MC polyethylene 18 -100

that a distortion of the bond angle and dihedral angle from the staggered trans conformation in the UHMW PE, which contains a large amount of non-crystalline component , occurs in going from room temperature to -108°C because of the influence on the crystalline structure produced by the reorientation of the non-crystalline structure through transfer from the gauche to the trans conformation and freezing of molecular motion. The non-crystalline region of '3C-labelled PEs crystallized under different conditions were studied by VT I3C CP MAS-NMR.18 The dynamics of the

dihedral angle/degree Fig. 7. 13C chemical shielding contour map of polyethylene calculated by the

FPT-INDO method.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

199

PPM

l ' ' ' ' l ' ' ' ' l ~ ' ' ' i ' ' ' ' l ' ' ' ' l 40 35 30 25 20

45

Fig. 8. Typical I3C CP MAS-NMR spectra of samples of PESL and MQPESL at

room temperature.

non-crystalline region were discussed by measuring 13C spin-lattice relaxation times (Tl) and dipolar-dephasing relaxation times ( TDD) over a wide temperature range, from -120 to 44°C. Two types of 13C-labelled PE samples were prepared. One is single *3C-labelled(polymerized using 90% single 13C-enriched ethylene) PE and the other is double labelled (polymerized using 90% double 13C-enriched ethylene) PE. Typical 13C CP MAS-NMR spectra of the single labelled solution-crystallized PE (PESL; the crystallinity is 95%) and single labelled melt-quenched PE (MQPESL; the crystallinity is 66%) are shown in Fig. 8. Each of these spectra consists of three peaks corresponding to an orthorhombic crystalline peak, 0, at 33.0 ppm, monoclinic crystalline peak, M, at 34.4 ppm (a small shoulder on the left side of the orthorhombic peak), and a non-crystalline peak, N , which appears at 30.8-31.3ppm. The 13C TI data of samples of PESL and MQPESL over a wide range of temperatures, obtained using the inversionrecovery method with the PST (pulse saturation transfer) pulse sequence, are given. The PST pulse sequence enhances the intensity of mobile methylene carbons. The resulting Tl values for samples PESL and MQPESL are plotted against the inverse of the absolute temperature (l/T) in Fig. 9. It had been suggested previouslylS2' that local molecular motion in the non-crystalline region of PE is independent of the degree of crystallinity, higher-order structures or morphologies. However, these suggestions are not supported by the experimental results because: (1) the Tls of the non-crystalline region of samples PESL and MQPESL at room temperature

200

H. KUROSU, S . ANDO, H. YOSHIMIZU AND I. ANDO

h

fn

Y

F

0.1

'

1

1

0.004

0.005

(K-') Fig. 9. Plots of non-crystalline I3C spin-lattice relaxation times: TI of samples PESL (0)and MQPESL (0) versus the reciprocal absolute temperature.

are 570ms and 440ms, respectively, and the difference of 130ms between them is beyond experimental error; and (2) the TI minimum values for the two samples (Fig. 9) are at different temperatures, the T I minimum of the sample of PESL appearing at -10.5"C and that of the sample of MQPESL at -32°C. These facts show that the local molecular motions in the non-crystalline regions of samples PESL are more constrained than that of MQPESL. In order to study whether the dynamic behaviour of the two kinds of non-crystalline region is also different on the T2 time scale, TD, values of samples of PESL and MQPESL were measured over a wide range of temperatures. The relative intensity of the non-crystalline peak obtained from the computer simulation was plotted against the delay time T in Fig. 10. It can be clearly seen that the non-crystalline peak of the sample of PESL relaxes more quickly than that of the sample of MQPESL. The dipolar-dephasing time, TDD, usually depends on molecular motion, carbon-proton dipolar interactions, MAS rate and spin diffusion.22 Fundamentally, it can be said that the dipolar-dephasing time in the noncrystalline region becomes a measure of molecular motion because of the high mobility. Therefore, the longer TDD value of the sample of MQPESL, compared with that of the sample of PESL, obviously suggests that the carbon-proton dipolar interaction is partially averaged by molecular motion on the T2 time scale.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

lo

201

1 0

100

200

300

400

500

Fig. 10, Intensity of non-crystalline peaks of samples of PESL and MQPESL versus delay time 7. The peak intensity was obtained from computer simulation of the 13C partially relaxed dipolar-dephasingNMR spectra.

Further, Chen et al.23 have measured the 13C CP MAS spectra of a '3C-labelled solution crystallized PE at temperatures from - 120 to 144°C. The measurements (Fig. 11) were taken in order to study changes of structure and molecular motion of the polymer with temperature variation for the crystalline and non-crystalline regions. As the crystalline and non-crystalline signals are incompletely resolved in the 13C CP MAS spectrum of the PE sample, computer-fitting of the spectra is performed, and the determined I3C chemical shifts of the crystalline and non-crystalline signals are shown in Fig. 12. It is shown that the 13C chemical shift of the non-crystalline signal decreases with an increase in temperature, whereas the 13C chemical shift of the crystalline signal does not change greatly with temperature before the melting point. This shows that an increase of the fractional population of the trans conformer leads to the observed high frequency shift of the non-crystalline signal with a decrease of temperature. At temperatures below -80°C the I3C chemical shift of the non-crystalline signal does not change with temperature. Such a result indicates that the molecular motion in the non-crystalline region of PESL is completely frozen below -80°C on the NMR time scale. The half-height width (half width) of the crystalline and non-crystalline 13C signals are plotted against temperature in Fig. 13. The half width of the non-crystalline region becomes a

202

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

144'c 132.8% 121.6c 110.4%

992% 88% &.SC

Fig. 11. 13C CP MAS-NMR spectra of the PESL sample as a function of temperature: (a) -120°C to room temperature; (b) 65.6 to 144°C.

maximum at -30°C. This means that molecular motion of the noncrystalline region occurs at the frequency corresponding to the amplitude of the proton decoupling field (about 55 kHz in this case). Contrary to the results of the non-crystalline signal, the half width of the crystalline signal is almost constant at temperatures from 0 to -120°C. The half width decreases from 0.9 to 0.7ppm as the temperature is increased from 0°C to room temperature, because of the molecular-motional narrowing effect. However, the half width of the crystalline signal increases as the temperature is increased from 65.6"Cand becomes a maximum at 110.4"C. This result may suggest that there is an a-transition, T,, in the crystalline region of PE and thus the temperature (110.4"C) at which the maximum of the half width is observed may be correlated with T,. Polypropylene and polystyrene are very important polymers in engineering plastics. For this, many NMR studies have been performed as reviewed in previous review articles. Therefore, these polymers are not reviewed here.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

31

t

nn --140

O

203

0 0 0000 0 (

-100

-60

-20

20

60

100

140

T("C1 Fig. 12. Plots of the I3C chemical shifts of the crystalline (0)and non-crystalline (0) signals of the PESL sample against temperature.

t - 1 4 0 -100

d -60

-20

20

60

100

140

V"C) Fig. 13. Plots of the width at half the maximum peak height of the crystalline (0)

and non-crystalline (0) signals of the PESL sample against temperature.

204

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

1 100

1

I

I

80

I 60

I

,

40

d 20

PPM FROM Me,Si

Fig. 14. 13C CP MAS spectra of poly(viny1 alcohoi) at room tern erature: (A) PVA-A; (B) PVA-B; (C) PVA-C. Line spectra represent the spectra in Me2S04-d6solution at 353 K.

'&

One of the high performance fibres is poly(viny1 alcohol) (PVA) which has received much attention. This polymer has a high function as a barrier material against oxygen. It is important to get detailed information about the structure and molecular motion for developing and designing these materials. The moiecular structure of PVA has been investigated by high-resolution solid-state 13C NMR. Terao et al. have measured the 13C CP MAS spectra of PVA at room t e m p e r a t ~ r e The . ~ ~ methine carbon resonance splits into three peaks in the CP MAS spectrum (Fig. 14). In the case of solution state NMR, the three methine carbon resonances and relative intensities correspond to the triad tacticity. However, the chemical shifts of the three methine carbons in the solid state are about 77, 71 and 65ppm and the relative intensities are not consistent with the triad tacticity in solution. The chemical shifts of two high frequency peaks move significantly to higher frequencies in the solid state. Terao et al. have assigned these peaks on the basis of the formation of intramolecular hydrogen bonds as follows; the highest frequency peak of the methine carbon is assigned to the mm triad

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

205

Table 3. I3C spin-lattice relaxation times of the respective resonance lines of different dry PVA samples, measured at room temperature.

I

Sample S-PVA A-PVA I-PVA

74.2, 13.2, n.m. 60.0, 12.1, 1.1 51.3, 10.4, n.m.

I1 58.3, 8.7, n.m. 65.0, 14.6, 1.2 39.7, 7.3, n.m.

I11 89.5, 12.3, n.m. 62.0, 12.3, 2.2 32.9, 4.6, n.m.

CH2 71.1, 9.2, n.m. 65.0, 14.6, 1.3 43.0, 7.8, n.m.

n.m., not measured.

with two intramolecular hydrogen bonds, the second highest frequency peak of the methine carbon is assigned to the mm and mr triads with one intramolecular hydrogen bond, and the other peak is assigned to the m m , mr and rr triads with no intramolecular hydrogen bonds. The barrier properties of PVA film decreases upon the addition of water. There is a significant change in the 13C CP MAS spectrum of isotactic PVA which is left in the air. The intensity of the highest frequency peak of the methine carbon increases and that of the lowest frequency peak decreases. This behaviour can be understood in terms of water molecules, which occupy positions between molecules, breaking intermolecular hydrogen bonds, and new intramolecular hydrogen bonds are formed. PVA films with different tacticities have been studied.25 The TIC measurements, using the Tochia pulse sequence, were performed relaxation times of three CH carbons. Table 3 lists the TI, values of the respective resonance lines of the CH carbons of syndiotactic (S-PVA), isotactic (I-PVA) and atactic (A-PVA) PVA samples. The 13C spin-lattice relaxation time analysis has revealed that three components exist for each of the samples, which are assigned to the crystalline, less-disordered non-crystalline and amorphous regions. Using the TIC difference, the spectra of the crystalline and non-crystalline components are separately recorded as shown in Fig. 15. The exact relative intensity for the crystalline component was corrected according to equation (1).

On the other hand, the exact relative intensity for the non-crystalline component was calculated by multiplying the fraction of line j by the factor gj, which is expressed as:

and then normalizing those values, where M y and MjNCt(t)are the peak intensities of line j for the spectra shown in Figs 15(a) and 15(c), and f j is the

206

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

.

80

60

40

20 ppm

from

MeqSi

Fig. 15. 13C CP MAS-NMR spectra of different components of PVA-A: (a) total;

(b) crystalline; (c) non-crystalline ((a) -0.29(b)).

peak intensity ratio for the crystalline and non-crystalline components of line j, which is given by:

The corrected relative intensities of the triplets of the CH resonance lines are not in accord with the contents of the triad sequences as listed in Table 4, suggesting the formation of intramolecular and intermolecular hydrogen bonds in the meso sequences of almost equal probability. Nylon is one of the important polymers with high performance such as moderate hygroscopicity, dyeing property and high strength. Therefore, nylon is widely used as an engineering plastic and a fibre. In order to elucidate the molecular structure and dynamics of nylon in the solid state, many investigations have been done by solid-state NMR. The 13C CP MAS-NMR spectrum and assignments, of a nylon 6 single crystal, is shown in Fig. 16.26 The PST MAS with a short pulse repetition time (5 s), which emphasizes the carbons with high mobilities, as well as the CP MAS technique, were employed for the single crystal and melt-quenched

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

207

Table 4. Integrated fractions, chemical shifts and linewidths of lines I, I1 and I11 of CH carbons for the crystalline and non-crystalline spectra of different dry PVA samples.

Integrated fraction I

Sample Crystalline S-PVA observed corrected" A-PVA observed corrected" I-PVA observed corrected" Non-crystalline A-PVA observed correctedb

I1

Chemical shift (PPm) I

I11

0.080 0.349 0.571 0.064 0.350 0.586 0.105 0.467 0.428 0.109 0.457 0.434 0.481 0,329 0.190 0.386 0.351 0.263 0.175 0.477 0.348 0.170 0.493 0.337

Line width (Hz) I1

111

76.7 71.1 64.9

193 201

156

77.0 71.2 64.8

156 214

197

75.4 69.8 64.7

141 237

211

76.5 70.9 65.5

203 211

209

I1

I

111

"Corrected by equation (1). bCorrectedby equation (2).

NYLON 6

II 1'1

GO

~

200

"

~

.

I

I50

,

~

.

' I00

I

'

.

.

so

'

I

"

"

PPm

Fig. 16. Typical I3C MAS-NMR spectrum of nylon 6 single crystals at room temperature.

~

208

H.KUROSU, S. ANDO, H.YOSHIMIZU A N D I. ANDO NYLON 6

+NH-cH~-cH~-CH~-CH~-CH~-C~S, I

o

50

i

40

I y

l p

30

l a

20

50

40

30

20

PPm

Fig. 17. 13C CP MAS and PST MAS-NMR spectra of nylon 6 (a) single crystal; (b) melt-quenched sample.

samples (Fig. 17). The detailed comparison of these spectra revealed that the amorphous peaks of the w-CH2 and (6+ y)-CH2 carbons of the melt-quenched sample are increased compared with those of the single crystal sample. These spectra show that the amorphous peaks appear at lower frequency than the crystalline peaks. The chemical shift behaviour of these peaks can be understood on the basis of the y-gauche effect. The crystalline state has a trans zig-zag structure and the amorphous carbons are undergoing rapid transitions between the trans and gauche conformations. Not only the 13C NMR spectra but also the 15N NMR spectra have been observed in the solid ~tate.''-'~ The "N NMR spectrum of I5N-enriched nylon 6 and computer-fitting with theoretical line shape are shown in Fig. 18. This spectrum shows that there are the crystalline and non-crystalline regions and supports the results of the 13C NMR spectra. The spin-lattice relaxation time analysis of 15N nuclei shows that the rigid crystalline region has a much longer TINrelaxation time than the more mobile amorphous region. The TIN data of the crystalline region are 111-416s and two non-crystalline regions are observed in the relaxation of the amorphous peak. The faster component has TIN values of 1-3s and the slower

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

I

loo

I

I

1

I

1

I

95

90

u

80

75

7o

209

PPM Fig. 18. 15N CP MAS line shape analysis of nylon 6.

component has values for TIN of 19-29s. Two non-crystalline regions are assigned to the amorphous and non-crystalline interface regions. The 15N chemical shielding anisotropy (CSA) pattern of nylon 6 is observed as a function of temperature (Fig. 19). The CSA powder patterns show the growth of an amorphous signal at elevated temperatures with a chemical shift near the isotropic value obtained with MAS. The u33component, which lies along the N-H bond of the amide group, becomes less prominent with increasing temperature and finally disappears above 115°C. Other nylon samples such as nylon 1130 or nylon 6631are also investigated by high-resolution solid-state 13C NMR. Polyoxymethylene (POM) is the first member of the polyether series

210

H. KUROSU, S . ANDO, H. YOSHIMIZU AND I. ANDO

15oOc

115°C

100°C

7 7°C

27°C 200

150

100

50

0

-50

PPM

Fig. 19. Static 'N NMR spectrum of annealed nylon 6 sample obtained with CP and

high power decoupling at different temperatures.

expressed by the general formula [(CH2),-O-],. This polymer has the property of abrasion resistance. From an X-ray diffraction study, POM usually takes the trigonal form with a 915 helical (all gauche) c ~ n f o r r n a t i o n . ~ ~ Veeman et al. have measured the 13C CP and the 90" pulses (without MAS) powder pattern, as shown in Fig. 20.'3 Fig. 20B was observed by the 90" pulses with proton-decoupling and the pulse repetition time is 1s. This spectrum shows only a mobile region since TIC of the immobile region is longer than the mobile region in the solid state. Therefore, Fig. 20B shows the mobile (non-crystalline) region and Fig. 20A shows both the noncrystalline and crystalline regions with different Tics. The spin-lattice

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

211

3

ppm

150

100

50

0

Fig. 20. Proton-decoupled I3C spectra of a non-spinning poly(oxymethy1ene) (Delrin) sample. Spectrum A results from 10ms CP, while for spectrum B the carbon magnetization is prepared via a 90" pulse.

relaxation time in the rotating frame ( T I P measurement ) of POM also shows that POM has two different kinds of regions with different mobilities. Kurosu et al. have measured the 13C CP MAS and powder pattern of melt-quenched POM in the solid state.34 It was found that the I3C NMR chemical shifts of the crystal and non-crystalline structures have different chemical shifts. The 13C NMR isotropic chemical shift for the crystalline region appears further towards high frequency by about 2 ppm than that for the non-crystalline region. Further, the principal values of the chemical shielding tensor of POM were determined for the crystalline and noncrystalline regions. These results show that the chemical shielding anisotropy

212

H. KUROSU, S . ANDO, H. YOSHIMIZU AND I. ANDO

Fig. 21. Pure absorption-mode 2D exchange spectra of isotropic POM: (a) temperature T = 360K and mixing time rm = 2s; (b) theoretical spectrum of (a) with x = 2w, tm = 2.5.

Au (= u33- ull) of the non-crystalline region is smaller than that of the crystalline region. They also calculated I3C shieldings of POM using crystalline and non-crystalline models. This calculation shows that the Aa value of the non-crystalline region is smaller than that of the crystalline region and the isotropic chemical shift of the non-crystalline region is displaced to high frequency as compared with that of the crystalline region. Recently, two-dimensional solid-state NMR has been developed. The molecular motions of POM have been investigated by this method. The observed (T = 360 K, mixing time t, = 2 s) and the calculated twodimensional exchange spectra of POM are shown in Fig. 21.35 The spectrum observed at 252 K shows only diagonal signals and this means that there is n o molecular motion during the mixing time. On the other hand, if molecular motion occurred during the mixing time, the spectrum would show steric effects as shown in Fig. 22. The calculations are based on the model of helical jump motions assuming a one-dimensional random walk in continuous time with an elementary process of 200" jumps. Thus at T = 360K, the subspectra for the discrete steps of a given CH2 group of POM experiences, as the helix rotates, can be identified. Kobayashi et al. have investigated the 13C NMR spectra of a needle-like single crystal of trigonal POM (t-POM).36The 13Csignal in the single crystal

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

213

Fig. 22. Pure absorption-mode 2D exchange spectrum at 75.47MHz of isotropic POM with temperature T = 252 K and mixing time tm = 1 s.

splits into a doublet at 88.5 and 87.7ppm in contrast to the singlet at 88.4 ppm in the semicrystalline sample. The resolved pattern is ascribed to the non-equivalent monomeric units in the crystal field of the t-POM lattice. The corresponding I3C signal of a plate-shaped single crystal of orthorhombic POM (0-POM) appears at 82.0 ppm as a singlet as anticipated from the space group of o-POM. The large chemical shift difference (6 ppm) between t-POM and o-POM is interpreted in terms of the intramolecular y-gauche shielding and the intermolecular packing effect. Since the discovery of polyacetylene film, which becomes electrically conductive by doping, a number of organic conjugated polymers have been synthesized and their electric properties have been studied. The NMR studies on polyacetylene and other conducting polymers have already been reviewed.*" In this section, the most recent studies on polypyrrole will be introduced. Polypyrrole is one of a series of heterocyclic polymers which has attracted much attention due to its characteristic electric and electronic properties. However, there are some problems relating to the physical and material properties which are associated with its structure. The fundamental structural formulae shown in Fig. 23 have been generally proposed for the structures of dedoped and doped polypyrroles, where the aromatic form corresponds to the dedoped state and the quinoid form corresponds to the doped However, the actual structure appears to be more complicated. At present the exact structure is not known because the polymer is amorphous and insoluble. Consequently, various structures have been proposed for p o l y p y r r ~ l e . ~ ~

214

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

High-resolution solid-state NMR spectroscopy provides useful information about the structures of synthetic polymers in the solid state, which sometimes cannot be determined by X-ray diffraction. NMR chemical shifts vary depending on the structure in the solid state and the separable resonance lines lead to an exact structural analysis. The structure of polypyrrole in the solid state has been studied by means of high-resolution solid-state 13C NMR spectroscopy.a However, the structure is insufficiently analysed because of the complexity of the unresolvable broad aromatic 13C signal. This is due to the fact that there are several magnetically non-equivalent aromatic carbons as shown in Fig. 23.

Fig. 23. Aromatic (a) and quinoid (b) structures for polypyrrole.

The structure of polypyrrole, prepared electrochemically, has been analysed by using high-resolution solid-state 15N NMR spectroscopy. The sample used is 15N-labelled in order to obtain 15N spectra with a high signal-to-noise ratio, as attempted by Wehrle et ~ 1However, . ~ ~they could not carry out a successful analysis of the I5N CP MAS spectra of the polymers in the solid-state because of insufficient resolution. As expected from Fig. 23, "N NMR spectroscopy will provide a simpler spectral pattern, when compared with 13C NMR spectroscopy, because a 15N resonance line may correspond to a given structure. Therefore, the structure of doped and dedoped '5N-labelled polypyrrole films can be successfully studied by high-resolution solid state NMR.42 Doped and dedoped samples were prepared by electrochemical polymeri~ation~~ using 20--30% '5N-labelled pyrrole. In order to obtain a dedoped sample the electrodes were inverted after the doping experiment and the same voltage was applied to them. The observed I5N CP MAS-NMR spectrum of polypyrrole (sample c; electrical conductivity = 4 X Skm) is shown in Fig. 24(a). It can be seen that the "N signals of the polypyrroles considered here consist of two major peaks and two minor peaks, which are designated by a,p, y and 6 with decreasing shielding. The four peaks were decomposed by computer fitting, and their 15N chemical shift values are approximately 90, 113, 129 and 145 ppm. This chemical shift range is very large and the results indicate that the polymer has at least four kinds of structure corresponding to the four peaks. The determined 15N chemical shifts, half widths and relative S/cm), b peak intensities of samples a (electrical conductivity = 7 x Ycm) and c are summarized in Table 5. (electrical conductivity = 2 X

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

-

215

(8)

W ~ 8 p E C T A u (

- tMEORETICAL SPECTRUM -- OECO)(POSEO ZIPECIRUM

170

150

130

110

90

70

50

(PPm)

OBSERYEO SPECTRUM WORETICAL OECOWOSEO SPECTRUM

(PPm)

Fig. 24. A 50.55MHz I5N CP MAS-NMR spectrum and a simulated spectrum of polypyrrole (sample c) in the solid state. The four peaks were decomposed by computer fitting: (a) contact time (CT) = 800 ps and (b) CT = 100 p s .

216

H. KUROSU, S. ANDO, H. YOSHIMIZU AND 1. ANDO

Table 5. Observed I5N chemical shifts, half widths and relative peak intensities of doped and dedoped polypyrrole samples.' Sample a

Peak

15N chemical shift ( P P 4

Half width (PPd

Relative peak intensity (%)

a

91.0 113.6 129.1 145.5 91.0 113.6 129.1

30.0 23.0 20.6 19.0 30.0 23.0 20.6 19.0 30.0 23.0 18.6 19.0

5.1 28.8 53.9 12.2 5.5 28.7 53.7 12.2 8.0 30.2 49.4 12.4

P b

Y 6

ff

B C

Y 6 a

P Y 6

145.5

87.5 112.8 128.2 142.5

"Determined by computer fitting.

As shown in this table, the relative intensities of peaks CY and p increase from 5.5 to 8.0% and from 28.7 to 30.2%, respectively, on going from sample b (doped) to sample c (dedoped). However, the relative intensity of peak y decreases from 53.7 to 49.4% by dedoping. Hence, the relative intensities of peaks a and p increase with a reduction in conductivity, but that of peak y decreases. In addition, the relative intensity of peak 6 does not change with the increase in conductivity. When the 15N CP MAS experiment is performed using a contact time of 100ps, the intensities of the peaks a and @ are relatively enhanced as shown in Fig. 24(b), and the chemical shifts and half widths of the observed shoulder peaks are determined exactly. Furthermore, the difference of the intensity enhancement between peak a,@ and peaks y, 6 shows the difference of the magnetic environments, i.e. a difference in TNHvalues between "N and 'H and in Tb, between the peaks a, p and the peaks y , 6. In order to get more detailed information about the I5N NMR chemical shift behaviour and physical properties of polypyrrole in the solid state, quantum chemical calculations have been performed. As suggested above, it is thought that polypyrrole predominantly takes the aromatic form and the quinoid form (Fig. 25). The isotropic "N NMR chemical shifts calculated by the finite perturbation theory (FPT)-INDO method for the aromatic forms are listed in Table 6, and those for the quinoid forms are listed in Table 7. The calculated values are shielding constants, and so the negative sign means deshielding. Since the observed values are the relative chemical shifts and the positive sign corresponds to deshielding, only the relative difference in the calculated 15N NMR shielding constants (u)should be compared with

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

217

Fig. 25. The structure of models used in the FPT-INDO calculation.

the observed data (8). As seen from these Tables, the calculated "N NMR chemical shift for the quinoid form appears towards a high frequency with respect to that for the aromatic one. Polypyrrole in the solid state is in an amorphous state.38 Therefore, some local structures are assumed. The calculations for these structures show that the N ' NMR chemical shift moves considerably to high frequency when a hydrogen atom bonded to a nitrogen atom (N2 in Fig. 25(b) and N4 in Fig. 25(d)) is very close to a hydrogen atom bonded to a carbon atom of the other ring. From these calculated results and the experimental findings that

218

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

Table 6. Calculated 15N shielding constants for the aromatic models' by the FPT-INDO method. Model

Nitrogen species

Calculated I5N shielding constant aiso(ppm)

N1 N2 N3 N4 N5 N1 N2 N3 N4 N5

-318.81 -317.33 -317.34 -317.23 -319.49 -319.18 -379.48 -319.54 -312.89 -320.56

"See Figs 25(a) and (b).

Table 7. Calculated "N Model

shielding constants for the quinoid models' by the FPT-INDO method.

Nitrogen species

Calculated I5N shielding constant aim(ppm)

N1 N2 N3 N4 N5 N1 N2 N3 N4 N5

-322.67 -325.73 -325.51 -325.72 -322.67 -324.41 -329.29 -315.82 -367.05 -324.00

"See Figs 25(c) and (d).

the observed I5N N M R chemical shift for the peak y appears towards a high frequency with respect to that for the peak p, and the fact that the intensity of the peak y for doped polypyrrole is larger than that for dedoped polypyrrole, it can be concluded that the major peak y, at about 129 ppm, is assigned to the nitrogen atoms in the quinoid form. The other major peak p, at about 113ppm, is assigned to the aromatic form, and the minor peak S at about 145ppm probably comes from the nitrogen atoms which are bonded to hydrogen atoms approaching other hydrogen atoms bonded to different atoms. From the calculations performed, the other minor peak a, at about 90 ppm, cannot be assigned.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

219

In order to get information about the I5N chemical shifts and electronic energy band structures of an infinite polypyrrole chain with aromatic or quinoid forms, calculations were carried out by the tight binding (TB) INDO/S method. As listed in Table 8, the calculated "N NMR chemical shift for the quinoid form appears at high frequency compared with that for the aromatic form. This agrees with the results calculated by the FPT-INDO method. The calculated band structures for both the aromatic and quinoid forms are shown in Fig. 26. The band gap is an important factor in determining electrical properties such as electric conductivity, where the band gap is the energy difference between the highest occupied band and Table 8. Calculated 15N shielding constants and band gaps for the aromatic and quinoid polypyrrole models by the tight-binding INDO/S-SOS method. ~~

Structure Aromatic form Quinoid form

~~

Calculated shielding constant a i , (PPm)

Band gap ( 4

-223.50 -232.51

5.12 2.86

the lowest unoccupied band. Therefore, if this value becomes smaller, the electric conductivity increases. The band gaps for the aromatic and quinoid forms are 5.1 and 2.9eV, respectively. This result shows that the electric conductivity for the quinoid form is larger than that for the aromatic form. Therefore, it can be expected that if the amount of the quinoid form is increased, polypyrrole with a higher electric conductivity can be obtained. Fully aromatic polymers that have no aliphatic groups in the main chain make up the upper class of engineering plastics. The representative ones are poly(pheny1ene oxide), poly(pheny1ene sulphide), polysulphone, poly(ether sulphone), poly(ether ketone), poly(ether ether ketone), polyamides (aramide), and polyimides. Their high thermal stability, high strength properties, chemical resistance, and electrical properties have attracted attention to their structures. Poly(pheny1ene oxide) (PPO) , poly(pheny1ene sulphide) (PPS), and poly(2,6-dimethyl-l ,Cphenylene oxide) (PDMPO) adopt the same crystalbetween line conformation illustrated in Fig. 27. The dihedral angles, protonated carbons (C,) and quarternary carbons (C,) are 45" and 45". The 13C CP MAS spectrum of solid PDMPO has been reported, where two resonances are observed for the C P . They ~ ~ attributed this splitting to the non-equivalent environment of the Cp produced by the non-linear C-0-C bonds because the Cp become equivalent only when the phenyl rings rotated 90" out of plane of the oxygen atoms or when the phenyl rings are able to rotate rapidly. One can also expect to see a doubling of the Cp

+,

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

221

$J,= 45"

cp2 = 45" Fig. 27. Schematic drawing of the crystalline conformation of poly(pheny1ene sulphide) (PPS).

signals in the I3C CP MAS-NMR spectrum. However, the I3C CP MAS spectra of PPS exhibit single resonances for the Cps in PPS. Gomez and T o r ~ e l l tried i ~ ~ to resolve this question by measuring the solid-state NMR spectra of diphenyl sulphide (DPS). Figure 28(c) shows a difference spectrum which shows only the resonances of the protonated carbons Cp. The rn and p resonances are singlets, but the 0-Cp carbon manifests a resonance doublet centred at 135.5 ppm with a 2 : l ratio of intensities and a singlet resonance ca. 6 ppm upfield at 129.9ppm corresponding in intensity to a single 0-Cp carbon. It is shown that the triplet of 0-Cp carbons is consistent with the conformation, in which the dihedral angles of four CPs are OD, M O O , 30-40°, and 14G15Oo, respectively. The least shielded resonances would correspond to the 2 o-Cps with 140-150" and 180"; the resonance shielded by ca. 1ppm would correspond to the single 0-Cp with 30". The reason why no split was observed in the Cp resonance of PPS was the 1ppm difference of chemical shift between the resonances with 3040"

222

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

a

A,,

, , I ,

145 140

135

130 125

120

PPm Fig. 28. (a) MAS-DD 13C NMR spectrum of diphenylsulphide (DPS) recorded at -60°C with a 420s delay between decoupling pulses. (b) CP MAS-DD 13C NMR spectrum of DPS recorded at -60°C with a 100 ps delay (without spin-locking) in the 'H channel after the Hartmann-Hahn match. (c) The difference spectrum [(a) - (b)] showing only protonated carbon resonances.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

1.4

5

-0.3

223

-2.0 5

ppm

Fig. 29. 300MHz 'H CRAMPS spectrum of poly(pheny1ene sulphide) (PPS) obtained using the BR-24 pulse sequence; 64 transients; relaxation delay 8 s.

and 140-150" which are smaller than the ca. 2ppm line width of the Cp resonance. This implies that the short C-0 bonds (ca. 1.4A) lead to a greater conformational sensitivity of 13C chemical shifts than the longer C-S bonds (ca. 1.8 A) in PPS. The spin-lattice relaxation time, TI, measured for PPS and DPS, of the order of several minutes, indicates the absence of rapid 180" ring flips. have observed the 'H CRAMPS Most recently Zumbulyadis et (combined rotation and multiple-pulse spectroscopy) spectrum of semicrystalline PPS (Fig. 29). Two well-resolved resonances, separated by 1.4 ppm, are observed together with a broad shoulder on the high frequency resonance. They interpret this split as reflecting the chemical shift nonequivalence of the two protons expected in the crystalline phase. Polyimides are usually generated by thermal imidization of poly(amic acid). Water and evaporated solvent bring about voids in the polyimide resins during the reaction. The problem has been solved mainly with the use of acetylene-terminated polyimide prepolymeric resins (Fig. 30) in which homopolymerization at elevated temperatures and pressures occurs without the formation of volatile products. The mechanism of cure is considered to be aromatization of the acetylenic end groups, but this has been impossible to verify because of the intractability of the cross-linked resin. The 13C CP MAS-NMR spectra of the resin and the cured polyimide polymermeasured4749 changes in the line shape of the aromatic carbon region are found as well as a diminution of intensity in the acetylene carbon region

Fig.

30. Structure of acetylene-terminated polyimide.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

225

(Fig. 31). Reactions involving the terminal acetylenic groups of the polyimide resin are involved in the polymerization process, but less than 30% of these acertylenes can undergo cyclotrimerization or other condensation reactions. The remainder appears to be consumed by addition reactions. To make the reaction clear, Swanson et aL5' have studied the solid-state chemical reactions in a '3C-labelled polyimide which had been prepared using starting materials labelled at sites that are thought to be involved in the cure reaction. The acetylene-terminated isoimide structure and '3C-labelled monomers are shown in Figs 32 and 33. Figure 34 shows the 13C CP MAS-NMR difference spectra of acetylene-terminated oligomer labelled at the carboxyl (isoimidefimide) positions. The smaller peak at 148.0ppm in the uncured oligomer, which is due to the C=N isoimide carbon, disappears, and a single resonance is observed in the spectrum of the cured polymer. This is consistent with the thermal isomerization of the isoimide to the imide. These carbons are not involved in any reactions with the ethynyl (acetylenic) carbons. Similarly, the diary1 carbonyl is not involved in the "cure" chemistry. On the other hand, the difference spectra for the samples which had been selectively 13C-labelled at the C-1 (non-protonated carbon) and C-2 (protonated carbon) ethynyl carbons are shown in Figs 35 and 36. In the delayed decoupling spectra, only the non-protonated carbons are observed. The spectrum of the C-1 labelled cured sample exhibits an intense resonance at 139.6 ppm (non-protonated), a smaller resonance at 129.6 ppm (protonated) in the aromatic region, and a fairly broad resonance at 52ppm (protonated) in the aliphatic region. The spectrum of the cured C-2 labelled sample shows a distinct peak at 128.7ppm (protonated), a broad resonance between 58 and 70 ppm (nonprotonated), and a shoulder on the high frequency peak at about 138ppm (non-protonated). The major products of the ethynyl carbons are aromatic groups as evidenced by the largest C-1 and C-2 peaks in the aromatic region. Another reaction product class is the condensed polycyclic aromatic structures as indicated by deprotonation of the C-2 and protonation of the C-1 carbons in the spectra. However, none of the cure products accounts for the resonances in the 50-70ppm range. The possibility of the further reaction of the Friedel-Crafts backbone addition, which forms a bridged structure, is proposed. Bismaleimides are also used as void-free high-temperature matrix resins for fibre-reinforced composite materials. The high strength and high modulus properties of these resins persists after hot-wet or hot-dry exposures to 250°C. Bismaleimides by themselves tend to form highly cross-linked networks which produce brittle resins. To alleviate this problem, chain extenders are introduced into the resins to lower the cross-link density and thereby increase the fracture toughness of the cured composite. Fry and Lind" have measured the 13C CP MAS-NMR spectra of 1,1'-(methylenedi-4,1-phenylene)-bismaleimide (BMI) as received, after

31. Difference spectrum (a - b) obtained by subtraction of the polyimide resin spectrum (with lines artificially broadened, b) from that of the cured polyimide polymer (a). Small deviations in the difference spectrum should be ignored since line broadening across the cured polymer spectrum may not be uniform.

Fig.

(v

)+

U a 4:

N

I

N

I

xN

m

a 4

t-

n m

4

:

t

N

I

B

227

228

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

C-1 Ethynyl

C-2 Ethynyl

'%

'%

BTDA

Carboxyl (ireimlde/lmide)

Diary1 CarbonyJ '%

'%

Fig. 33. 13C-labelled monomers.

200

100

0

(PPd

Fig. 34. I3C NMR difference of acetylene-terminated oligomer labelled at the carboxyl(isoimide/imide) positions.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

SSB 200

229

Uncured

100

0

(PPd Fig. 35. 13C NMR difference spectra of acetylene-terminated oligomer labelled at the C-1 ethynyl position.

quenching, and after curing for 4 h at 493 K (Fig. 37). The cured BMI spectrum clearly shows a new carbonyl resonance at 175ppm. A decrease is observed in the intensity of the resonance at 132ppm, and a new resonance occurs at about 45ppm. Two primary reactions, shown in Fig. 38, appear to take the chain place during curing: (Rl), the cross-linking reaction, and (W), extension reaction. In the (R2) reaction, the chain extender is methylene dianiline (MDA). Figure 39 shows the assignments for the reactants and products. As the maleimide rings react to form substituted succinimides (the products of R1 and R2), the carbonyl concentration at 169 ppm shifts to provides a good measure of the 175ppm. The ratio of a175/(a175+u169) extent of cure as shown in Fig. 40 (ax is the area of the resonance at xppm estimated from curve fitting). On the other hand, since the 45ppm and the 52ppm resonances occur only as the products in (Rl) and in (R2), respectively, the ratios involving u45and a52 suggest the preference of these

230

H. KUROSU,S. ANDO, H. YOSHIMIZU AND I. ANDO

:'I.-

Oelayed Decoupling

I

Uncured

s SB I

200

JL I

I

100

I

I

0

(PPd Fig. 36. "C NMR difference spectra of acetylene-terminated oligomer labelled at the C-2 ethynyl position.

reactions under a certain cure condition. Figure 41 thus clearly shows that the chain extension reaction (R2) is favoured at 418K, whereas the cross-linking reaction (Rl)occurs to a greater extent at 493 K (the 39 ppm resonance should be constant). Other examples of the application of 13C CP MAS-NMR to engineering plastics can be found in the l i t e r a t ~ r e ; ~ - ~including ~ studies on polycarbonate ," poly(ethy1ene t e r e ~ h t h a l a t e ) , ~ ~sulphonesS3 and polyirnide~.~~ In addition, recent studies are found for p ~ l y c a r b o n a t e , ~ ~ poly(ethy1ene t e r e ~ h t h a l a t e ) ,poly(buty1ene ~~ terephthalate),m poly(ether ether ketone)6' and polyamide.62 Above any specified temperature, bulk polymers sometimes form a liquid crystalline phase before going from solid to liquid. For example, in a series of poly (L-g1utamate)s with long n-alkyl side or n-alkenyl side chains,& the side chains form a crystalline phase at low temperature, but at

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

I

200

I

1

100

0

231

PPM Fig. 37. 13C CP MAS spectra of BMI as received (bottom), after quenching (middle), and after curing for 4 h at 493K (top). Spinning sidebands are labelled SSB.

higher temperatures the side chains melt like liquid n-alkanes or liquid alkenes and the polymers have a liquid crystalline character. The thermotropic behaviour of poly(L-g1utamate)s with long n-alkyl side chains or with oleyl side chains with double bonds have been investigated by VT CP MAS-NMR and pulse NMR. Figure 42 shows the 13CCP MAS-NMR spectrum of poly(y-n-octadecy1-Lglutamate) at room temperature together with the assignment of peaks. It has been demonstrated that 13Cchemical shift values of CO (amide) and C, carbons and interior CH2 carbons depend on the conformation and crystal structure. At room temperature, 13C chemical shift values of the CO (amide) and the C, carbons are 176.0 and 57.6ppm, respectively. These

232

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

v*

Y

x

HNQ-

M-

N

V

CHrX

G"

0

Fig. 38. Reactions [Rl]and [El.

values show that the main chain takes an a-helix conformation. On the other hand, the interior CH2 signal splits into two peaks at 33.4 and 30.6ppm indicated by I and A, respectively. From the reference data of n-alkanes and polyethylene, it is seen that peak I arises from the CH2 carbons in the all-trans zig-zag conformation in the crystalline state and that peak A arises from the CH2 carbons in the non-crystalline phase or the liquid phase. Figure 43 shows the VT 13C CP MAS-NMR spectra of the polymer at temperatures ranging from room temperature to 100"C.a (It is demonstrated that the polymer forms a liquid crystalline phase above 40°C.) The n-alkyl peaks change observably as the temperature is increased. Peak I

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

233

c,e, w, x, d

F,F,F" SSB

2.

M'

I

I

1

200

100

0

PPM Fig. 39. I3C TOSS spectrum of (a) pure BMI cured 4 h at 498 K and (b) 1.5 :1 BMI :MDA cured 1 h at 418 K. Assignments are listed in Table 2.

disappears above 35°C and the intensity of peak A increases noticeably. This is due to the melting of side chain crystallites. On the other hand, as for the CO (amide) and C, carbons, a progressive broadening of these carbons is observed. At about 40°C, the peaks are broadened to the point of disappearing from the spectrum. At higher temperature the peaks appear at the same chemical shift values again. This means that the main chain takes an a-helix conformation. The broadening can be explained on the basis that the reorientation rate of the main chain becomes insufficient to average dipolar interactions with protons. This reduces the efficiency of the radiofrequency decoupling and leads to a maximum line width of the carbons when the molecular motion occurs at the frequency corresponding to the amplitude of the proton decoupling field (about 60kHz for the experiment). Therefore, it can be said that the main chain with the a-helix form is undergoing reorientation at a frequency of about 60kHz in the liquid crystalline phase. The VT 13C CP MAS experiments on poly(y-oleyl-L-glutamate)66with unsaturated long side chains show that the main chain takes the a-helix form

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

234 1.o

0.9

+

0.8

41

:

0.7

0.6 0 .-

0.5

0.4

+

0

0.3 A

0.2

0.1

0

10

100

1000

10000

Fig. 40. The 13C NMR carbonyl concentration ratios a175/(a175 + a169)for the 1:0 BMI :MDA resin cured at 493 K for up to 10 OOO min are shown by solid points. The square point represents a 1:0 sample that was first dissolved in chloroform and then held at 418 K for 2 h. The triangle point represents a 1:0 sample that was quickly melted and then quenched before curing at 418 K for 2 h.

within the temperature range from -40 to 80”C, while the long side chains are in a mobile state above -40°C. Such a situation is very similar to the case of poly(y-n-octadecyl-L-glutamate) . A detailed discussion of the dynamics of these two polymers is based on proton relaxation times and proton line widths measured over a wide range of temperatures from -150 to 1 2 0 ~ ~ ~ Long n-alkanes, on going from the crystalline state to the liquid state, take the hexagonal form in the narrow temperature range prior to the melting point, in which the n-alkane chains are rotating about their long axes. The hexagonal form is the so-called “rotator phase” like the liquid crystalline phase. Ishikawa et aLm have investigated this problem by VT 13C CP MAS-NMR experiments over a wide range of temperatures. It is revealed that the 13C chemical shift value of the interior CH, carbons in n-G4HS0moves from 34.2 to 33.3ppm on going from the triclinic form to the rotator phase, and the 13C chemical shift of the interior CH, carbons of n-CI9Ha and n-C3,1-& move from 32.8 to 33.3ppm on going from the orthorhombic form to the rotator phase. From these results, it can be said that 13C chemical shift values of n-alkanes in the rotator phase are the same

2.0

8

1.8

0 a511a39 - 4 1 8 K m m 0 a451a39 -418Kcure

1.6

a521a39 - 4 9 3 K c u m /a39 493 K cum

-

1 hat418K+ 2hat493K

1

1.4 1.2

2.-

-

0

0

0

1.0

0 0

0

0.8

0

0

8

0.6

0 0

0

0.4

0

0

B

0.2 0

1

10

0

I-

0 0

0

8 hat 418 K +

0

Cure time (min)

100

1000

41. Ratios of 13C CP MAS resonance areas for the chain-extension methylene at 53ppm (circles) and the cross-linking methylene at 45 ppm (squares) to the diphenyl methylene at 39 ppm are shown for cures performed at 418 K (open points) and at 493 K (solid points) for the 2.5 :1 BMI :MDA system.

Fig.

N

w ul

236

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

CO(amide) CO(ester)

1

1

,

200

,

1

~

.

~

1

150

1

100

1

1

~

1

50

.

.

1

~

1

.

1

0

S/PPm

Fig. 42. 13C CP MAS-NMR spectrum of poly(y-n-octadecyl-L-glutamate) at room temperature. Peaks in the vicinity of 30 ppm are expanded. Assignment of each peak is done by reference to the data for poly( y-benzyl-L-glutamate) and n-alkanes in the solid state.

irrespective of chain length and therefore every n-alkane in the rotator phase has the same structural aspects. Further, in order to study the dynamic features of n-alkanes in the rotator in the crystalline phase, rotator phase phase, the 13C Tl values of and liquid phase were measured by Torchia’s pulse sequence method and The the inversion-recovery pulse sequence method as shown in Fig. order of the magnitudes of the TI values for the CH2 carbons is (w-CH2< &CH2 < interior CH2 in the crystalline phase, in contrast to that in the melting phase. According to BBP theory,” as the correlation time for molecular motion increases, Tl first decreases and then increases again, passing through a minimum. Therefore, it can be said that the molecular motion of the CH2 carbons of n-C32H66in the crystalline state at 60°C is in the slow motion region and that the molecular motion of the CH2carbons in

.

~

~

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

237

Glpprn

Fig. 43. I3C CP MAS-NMR spectra of poly(y-n-octadecyl-L-glutamate)as a function of temperature. The main chain carbon peaks are expanded.

238

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

I

I

I

1

1

I

100

a

E

!ii1 0 I

melting

1 CH3

a-CH2

PCHz

y-CHz

inLCHz

Fig. 44. I3C spin-lattice relaxation times NT, for n-C32H66 chain at various temperatures. N indicates the number of protons bonded to each carbon considered.

the liquid phase at 80°C is in the extreme narrowing region. However, for TI values in the rotator phase at 70°C a-CH2 = @CH2 = y-CH2 = interior CH2. This means that every CH2 carbon in the rotator phase is undergoing molecular motion with the same mode. Therefore, it can be said that all-trans zig-zag chains in the rotator phase rotate about their long axes. Mathias71 has examined the structural behaviour of long n-alkyl side chains of some comb polymers with the following structures which form a liquid crystalline phase above about 4 0 T , through the observation of the VT I3C CP MAS-NMR spectra. A I3C chemical shift of about 3 ppm to low frequency is observed for the methylene carbons as the side chain crystallite is melted. The CP efficiency for the main chain carbons is decreased. These behaviours are similar to that of poly( y-n-alkyl-L-glutamate) mentioned above. Cooling the liquid crystals below the melting transition associated with the liquid crystalline transition results in broadened lines. The relaxation times of the liquid crystalline polymer, shown below, which has a liquid crystalline transition at 44"C, show that below the melting transition the mesogenic unit and the spacer carbons alpha to the mesogen

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

239

are quite rigid.72 The spacer has a substantial degree of mobility and the backbone has a much higher relative mobility compared to the flexible free substituent methyl carbons. 13C CP MAS spectra for the liquid crystal as a function of thermal history are shown in Fig. 45. Heating to 100°C and cooling cause excessive broadening of all the lines indicative of an amorphous polymer. Heating to 70°C, the top of the liquid crystalline transition, narrows the lines and this is maintained when cooled below the glass transition. ~ ~ indicated ~ ' that the orientational order is maintained Spiess er ~ 1 . have when cooled below the glass transition for side chain liquid crystals. Positional order is also required to crystallize. Further, the other liquid Crystalline polymers such as polyester, side chain polysiloxane liquid crystals, polyamide, etc., have been investigated by solid-state NMR.7s83 The structural change of lipids, dimyristoylphosphatidylcholine and distearoylphosphatidylcholine, in going from the crystalline phase to the liquid crystalline phase was investigated by VT 13C CP MAS-NMR at temperatures from -90 to 140"CM The experimental results show that the conformational change of the n-alkyl chains of DMPC and DSPC in the liquid crystalline state is very similar to that of n-alkanes in the liquid state near the melting point, and below the crystalline-liquid crystalline transition temperature the n-alkyl chains are still reorientating with some degree of disorder. 3. POLYMER ALLOYS

The physical blending of two or more polymers provides new polymer materials with suitable physical proper tie^.^^ Such polymer blends are often referred to as "polymer alloys" to use a term from metallurgy. The

240

H.KUROSU,S. ANDO, H.YOSHIMIZU AND I. ANDO

Fig. 45. The effect of thermal history on the 13C CP MAS-NMR spectra of the liquid crystalline polymer in scheme (below). (a) 20"C, (b) 40°C, (c) 70"C, (d) 20°C after cooling from lOO"C, (e) 70°C after heating d, ( f ) 20°C after cooling from 70"C, and (g) 20°C after cooling from 100°C. k 1

H H4-0 H W

f

e

d

n=7

m

c

b

a

HHHTH

H H O-C-(C)n-C-C4-Si-C-H H H H H H l H

*

x = 80

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

241

thermodynamics of polymer-polymer mixtures plays an important role in the molecular state of dispersion, leading to an understanding of the morphology of two phase mixtures and the interface between phases. In the polymer blends, it is very important to understand the size of the phases or domains on a molecular or segmental scale. In this section, most recent NMR studies will be introduced on these subjects. In polymer blends, spin diffusion between individual polymers can occur when they are intimately mixed. From such a point of view, miscibility in polymer-polymer mixtures has been the subject of considerable discussion through the 2D NMR and CP MAS-NMR results. The first 2D 'H NMR experiments on a mixture of polystyrene and poly(viny1 methyl ether) (PVME), one of the typical polymer alloys, was carried out by Ernst et a1.& The cross-peaks between signals coming from polymers, when cast from toluene solution, were successfully observed but not when cast from chloroform solution. In the former blend well-mixed domains exist, but in the latter blend they do not. Ernst et al. have suggested that through the observation of the quantitative determination of cross-peak intensities and spin diffusion rates, the fraction of mixed domains and domain size can be estimated. I3C CP MAS spectra of the mechanically mixed deuterated polystyrene (d-PS)/poly(vinyl methyl ether) sample were measured*' as a function of contact time at -33"C, since poly(viny1 methyl ether) has a proton TIP minimum at room temperature which makes cross-polarization ineffective. The spectrum is dominated by the signals of PVME. However, at longer contact times, a signal coming from the aromatic carbons of d-PS is observed. This comes from domains where interfacial mixing has occurred between the PVME, which is above its glass temperature when mixed, and the d-PS. 13C CP MAS-NMR spectra of the same sample, heated to 130°C for 30 min and quenched, were measured under identical conditions as before, as shown in Fig. 46. In addition to the PVME signals, a very intense signal due to the phenyl carbons of the d-PS and resonances due to the d-PS methylene and methine backbone carbons appear. The signal intensities decrease at long contact times due to PVME proton Tb relaxation. Further, the I3C CP MAS-NMR experiments on two blends of d-PS and poly(methy1 methacrylate) (PMMA) were performed but no signal due to d-PS was observed. The above technique demonstrates that d-PS and PVME are compatible, but d-PS and PMMA are not compatible. As seen from these experiments, the polarization transfer has an r-6 spatial dependence** which generally limits effective transfer to distances of less than about 20 A. Parmer et al.89 determined polymer-polymer miscibility by I3C CP MAS-NMR in blends of deuterated and protonated polymers, d-PSIPVME, deuterated PMMA (d-PMMA)/PS and deuterated PMMA/poly(vinyl chloride) (PVC). I3C CP MAS experiments on d-PSIPVME and d-PMMA/PS were done and similar results found. In Fig. 47 the 13C CP MAS-NMR

25ms ( e )

t

I

I

I

1

I

250

200

150

100

50

O

ppm from TMS

1

-50

-1

t

1

I

1

I

I

I

-100

250

200

150

100

50

0

-50

1

-100

ppm from TMS

Fig. 46. 50.17MHz I3C CP MAS-NMR spectra of a 50:50 mechanical mixture by weight of deuterated polystyrene/poly(vinyl methyl ether) (d-PS/PVME) blends as a function of contact time at -33°C. Cross-polarization and decoupling fields are ca. 55 kHz. The number of accumulations is 128. (a)-(d) are of the mechanical mixture before heating. (e)-(h) are of the same sample after heating to 130°C for 2 X 30 min and quenching after each heating cycle.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

2 00

100

243

do

PPM

Fig. 47. 50 MHz 13C CP MAS-NMR spectra o f (a) poly(viny1 chloride) (PVC); (b) poly(methy1 methacrylate) (PMMA); (c) 95% deuterated PMMA/S% PVC and (d) a spectrum of (c) obtained with a dipolar decoupling delay of 40 ps.

spectra of PVC (a), PMMA (b) and a blend of 95% of d-PMMA and 5% of PVC (c and d) are shown. In the 13C CP MAS spectrum of the blend the PVC resonances are suppressed by a delayed dipolar decoupling sequence. Even though Fig. 47d shows the spectrum from less than half of the transients used to obtain Fig. 47c, the signals are three to four times more intense. This shows not only miscibility at the molecular level but also the applicability of the technique to blends containing only a few per cent of protonated polymer components. The miscibility enhancement of PS/PMMA blendsg0through ionic interactions was studied by the intermolecular cross-polarization method as mentioned above. When PS is slightly sulphonated (3.3 mol% of -S03H) and PMMA is copolymerized with 2.3 mol% 4-vinylpyridine (4VP), the miscibility of the blend is enhanced, due to proton transfer from the -S03H to the 4VP moiety with the formation of ionically interacting sites on the d-PS and the PMMA. By the introduction of about 9.5 mol% of interacting groups (-S03H and 4VP), the blend shows a much greater enhancement in miscibility. When the S 0 3 H on the d-PS is neutralized with N(CH&OH and the 4VP is quaternized with CH31, the blends exhibit similar or even better miscibility compared with the proton-transfer blends. This is attributed to direct ion-ion interactions in the blends.

244

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

Chu et aL91 have studied thermally induced phase separation in PS/PVME blends through the observation of the proton spin-lattice relaxation, in both the laboratory and the rotating frame, for the entire range of blend composition. Under conditions in which the blends are compatible, 'H TIP results obtained at -5°C showed microheterogeneity at a 10 A scale using the relation9' given by (L') = (Tl/T2)(?) where (15') is the molecular contacts over a coherence scale L, (Tl/T2)is a ratio of spin-lattice relaxation time to spin-spin relaxation time and r is the average distance between protons. 'H TIP values at room temperature are closer to the longer relaxation time of PS than that expected from a simple weighted average of the relaxation times of the constituent homopolymers. This shows incomplete averaging by spin diffusion and a restraining effect of PS on the segmental motions of PVME. These blends are heated to cause phase separation. From the biphase decay of 13C magnetization, the compositions of the separated phases give a lower critical solution temperature phase diagram. NMR relaxation in PVME blends with PS molecular weights of 9, 100, and 900 K are compared. By the observation of the temperature dependence of the 13C line width, for the CH carbon of PVME in a PS/PVME blend, the effects of blending on molecular motion of the individual polymer components have been i n ~ e s t i g a t e dThis . ~ ~ shows that the molecular motion averages the distribution of isotropic chemical shifts in the glassy state and the interference between local anisotropic motion and high power proton decoupling is also noted. Solid-state phase behaviour and molecular level mixing phenomena in a strongly interacting poly(ethy1ene oxide) (PEO) and resorcinol mixture have been investigated through I3C CP MAS-NMR and 'H CRAMPS" experiment^.'^ One-dimensional (1D) 'H CRAMPS spectra of the polymer blend are shown in Fig. 48. The aromatic proton signal of resorcinol in the beta phase is l p p m to high frequency of its counterpart in the undiluted crystalline state of the small molecule. The hydroxyl proton signal of resorcinol in the weaker hydrogen-bonded phase beta is shifted slightly to low frequency relative to that characteristic of phase gamma. The above results are consistent with the F T IR results. These experiments provide substantial evidence that PEO and resorcinol in the molecular complex (phase beta) are intimately mixed on a molecular level. Order-of-magnitude estimates of interproton separations based on a spin-diffusion coefficient of 5 x lo-'' cm2/s suggest that dipolar distances of 1, 10 and 100 8, should produce proton spin diffusion on a time scale of 2 0 p , 2ms and 200ms, respectively. In this respect, 'H CRAMPS spectra are capable of distinguishing between proton spin diffusion in intimately mixed and phaseseparated blends. The duration of the mixing period necessary to detect spin diffusion depends on the solid-state morphology. Figures 49 and 50 show the 2D proton spin-diffusion spectrum of the

-R

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS Aromatics

I\

Resorcinol

0

15 '

,-. &d-J

10

-5

I

r----r-----r' 15

245

1

5

.

I

0

-5

PPm

Fig. 48. 'H CRAMPS spectra of undiluted crystalline resorcinol (upper spectrum) and the molecular complex of poly(ethy1ene oxide) and resorcinol (lower spectrum). The concentration of resorcinol in the molecular complex (phase beta) is 33 mol%.

PEOhesorcinol blend. This allows us to distinguish the signal contours of the methylene protons of PEO from the aromatic and hydroxyl proton signal contours of resorcinol. In these figures, the signal contours on the main diagonal from lower left to upper right are assigned, respectively, to the hydroxyl protons of resorcinol, the aromatic protons of resorcinol, and methylene protons of PEO. After 100 ps of proton dipolar interaction, spin diffusion between PEO and resorcinol can be observed in Fig. 49, through off-diagonal contours indicated by the arrows. This shows that proton

246

H. KUROSU,S. ANDO, H. YOSHIMIZU AND I. ANDO

t t

?

Fig. 49. 2D 'H spin diffusion spectrum of the molecular complex of poly(ethylene oxide) and resorcinol. The mixing period persists for 200 ps in the presence of proton dipolar interactions. The horizontal and vertical arrows identify off-diagonal contours generated via proton dipolar communication (aromatic-hydroxyl) within the resorcinol molecule. The slanted arrows identify off-diagonal contours generated via intermolecular dipolar couplings between poly(ethy1ene oxide) and resorcinol.

magnetization transport between PEO and resorcinol in phase beta is operative on the 100 ps time scale, suggesting that intermolecular proton distances are in the range of 2-5A. Figure 49 shows that spin diffusion between PEO and resorcinol is absent when homonuclear 'H dipole-dipole interactions are applied for 40 ys. Four of the six off-diagonal contours (Fig. 49) are absent in Fig. 50. Dipole-dipole couplings between the methylene protons of PEO and the hydroxyl protons of resorcinol, and the methylene protons of PEO and the aromatic CH protons of resorcinol do not produce off-diagonal contours in the 2D 'H NMR experiment on a time scale of 4Ops. This is a consequence of dipolar couplings between protons that are separated by more than 2 hi. Miscible blends of polybenzimidazole (PBI) with an aromatic polyimide

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

Resorcinol Hydroxyl OH

L

4

Resorcinol Aromatic CII

247

a

PEO CH

Fig. 50. 2D 'H spin-diffusion spectrum of the molecular complex of poly(ethytene oxide) and resorcinol (phase beta). Dipolar couplings are exerted for 4Ops during the mixing period. Only the aromatic-hydroxyl off-diagonal proton contours, generated via dipolar communication within resorcinol, can be observed at this shorter spin-diffusion mixing time.

and a polyetherimide were investigated by 13C CP MAS-NMR.96 The I3C CP MAS experiments show that specific hydrogen bonds exist between the phthalimide carbonyl of polyimide and the imidazolic amine bond of PBI. This evidence comes from a change in the carbonyl signal of polyimide. Addition of PBI broadens this signal, principally in the high frequency direction, and a high frequency carbonyl chemical shift shows the existence of a fraction of the carbonyl to amine hydrogen bonds. Further, miscibility in the blends has also been confirmed by a study of the proton rotating frame spin-lattice relaxation behaviour. Charge-transfer interactions in polymer blends containing the electron donor (N-ethylcarbazol-3-y1)methyl methacrylate (NEC-MM) and the electron acceptor 2-[(3,5-dinitrobenzoyl)oxy]ethyl methacrylate (DNBEM) moieties were investigated by 13C CP MAS-NMR and differential scanning calorimetry (DSC).97The number of inter- vs intramolecular charge-transfer interactions was varied by preparing blends of polydonor with acceptor, as

248

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. A N D 0

well as blends containing a homopolymer and an acceptor-donor copolymer. The blends were one phase for NECMM contents greater than 35 mol%. At lower donor contents, two phases with different amounts of charge transfer complexation were shown to exist. NMR, via 'H Tb and DSC, via Tg (glass transition temperature) indicate that charge-transfer interactions result in restricted mobility and reduced free volume and interatomic distances in the bulk polymer blends. Intermolecular charge-transfer complexes are decomplexed on heating above 185"C, and two phases, composed of only one blend component, are formed. A chemical shift observation by charge transfer suggests that the complex has an asymmetrical structure. Phase behaviour of blend systems containing nitrile copolymers (acrylonitrile/methyl acrylate/butadiene terpolymer) (B210) with poly(ethy1ene-comaleic anhydride) (PEMA) and poly(oxycarbonyloxyl-1,4-phenyleneisopropylylidene-1,Cphenylene) (PC) was investigated by I3C CP MAS-NMR and DSC.98Spin-lattice relaxation times of the protons in the rotating frame for blends and pure components have been measured by monitoring the CP MAS generated carbon signal intensities as a function of a variable proton spin-lock time. Since single-component NMR relaxation behaviour was observed over the entire range of compositions, phase homogeneity was demonstrated for the B210/PEMA system. These monophase blends exhibit single-component rotating frame spin-lattice relaxation times, intermediate in value as compared to the pure components. The DSC examinations also show that B210/PEMA blends have a single composition-dependent glass transition temperature. In contrast, the CP MAS-NMR experiment of B211PC blends shows that they display multicomponent relaxation behaviour indicating the presence of a mixed phase along with pure phases of each individual polymer. The modulus and transition temperature of polymer materials can be controlled through the addition of low molecular weight diluents. From this point of view, diluent motion in a glassy blend such as a 50/50 blend of poly(pheny1ene oxide) (PXE) and PS plus a phosphate ester diluent, tetraxylyl hydroquinone diphosphate (HQDP), was investigated through 2D 31P NMR.w 2D 31P NMR spectra of the polymer blends are shown as a function of mixing time at 71 and 83°C (Fig. 51). The first static 2D spectrum (a) is measured at 71°C and a mixing time of 1s, at which a mechanical loss peak is observed. The spectrum shows only intensity along the diagonal corresponding to the axially asymmetric chemical shift anisotropy line shape. The absence of off-diagonal peaks eliminates the presence of 31P spin diffusion on this time scale. The 2D spectrum (e) at 83°C and a mixing time of 1s, at which the dynamic mechanical spectrum shows an increased loss, shows significant diagonal peaks. This can arise from molecular reorientation with a time constant less than, or equal to, the mixing time of the pulse sequence, The rate and amplitude of the motion can be further character-

N M R STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

249

0

Simulation

400 300 200 100 PPm

0 -100 -200

PPm

Fig. 51. (a) 31P2D line-shape pattern at 71°C and a mixing time 1 s. (b-e) 2D line-shape pattern at 83°C and mixing times of 1, 10, 100, 1000 ms. ( f ) Calculated 2D line-shape pattern for comparison with the pattern obtained at 83°C and a 100ms mixing time.

250

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

ized by acquiring 2D patterns at several times, 1, 10, 100 and 1000ms. From these spectral patterns a variety of specific descriptions of molecular motion may be presented. Using the similar 2D 13C NMR method, a blend of PS and 13C-labelled poly(2,6-dimethylphenylene oxide) was investigated as a function of temperature, and mixing time, in the vicinity of the thermal glass transition.lW Chain motion of the '3C-labelled polymer on a ms time scale commenced at temperatures of approximately 10°C below the thermal glass transition, in contrast to a single-component polymeric glass which only shows such motion at temperatures above the glass transition. The miscibility and separation of solution-cast blends of poly(viny1idene fluoride) ( P W ) and PMMA, with ageing, for a range of composition was examined by I3C CP MAS-NMR.lo1 One amorphous phase and intimate mixing of the polymer chains in this phase existed for all compositions of the blends, even after 2 months of ageing at room temperature as determined by the 'H Tlp and the TCH.The 'H TQ shows the presence of phases or domains in the amorphous component of the blend larger than approximately 19 A. The TCH values increase with ageing for all the carbons of PMMA. This shows that a subtle separation between polymer chains is occurring on the scale of 4-5 A and that transfer of polarization is not as efficient with ageing. A partially miscible blend of d-PMWsolution chlorinated PE (dPMMNSCPE) has been examined through the CP time TcH obtained by 13C CP MAS-NMR.lo2 The mean internuclear distance between SCPE protons and d-PMMA carbons was estimated by use of the TCH data of the well-defined SCPE methylene carbon. Proton to carbon intramolecular Tcp values are proportional to r6, where r is the internuclear distance, and also proportional to n-', where n is the number of protons attached to the carbon. Thus, the two protons on the SCPE methylene at a distance of 1.09A give rise to a TCHof 25ps. It is not possible to know how many SCPE protons are involved in interchain CP to the d-PMMA carbons, but an approximation can be made that on average four SCPE protons are near any one d-PMMA carbon. Therefore, the value obtained for Tcp of 11ms for the d-PMMA methyl carbon is calculated to result from interactions with four protons (n = 4) at a mean distance of 3.4 A. The scale of miscibility of blend films of cellulose (CELL) with poly(viny1 alcohol) (PVA) , polyacrylonitrile (PAN), poly(ecaprolactane) (PCL) and nylon 6 was estimated by solid-state NMR relaxation measurement^.'"^ The miscibility of the synthetic polymers with CELL, as obtained from NMR relaxation measurements, follows the order PAN = PVA > PCL > nylon 6. This order agrees well with one deduced from the comparison of the dynamic mechanical analysis of the different blends. By combining NMR and dynamic mechanical analysis results it is possible to estimate more precisely the domain size produced upon blending two polymers; hence,

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

251

blends can be looked at on a scale of 1nm (CP MAS), 4 nm ( T I P ) 15 , nm (dynamic mechanical analysis), and 25-40 nm ( T I ) . The elastomeric components of several tyre sections were examined by multiple-slice, three-dimensional 'H NMR imaging at medium resolution (1W200 pm) and short echo times (0.5-2 ms).'04 The various rubber layers and cords were readily distinguished for a non-steel-belted tyre section at 200 pm resolution, presumably based on T2 differences among the different components. 13C MAS-NMR spectra of samples from each layer provided elastomer composition, which supported the origin of the intensity differences seen by NMR imaging. Experimental tyre tread sections with good and poor carbon black dispersions produced substantially different images at 150 pm in-plane resolution. Hikichi et af.'05 have investigated the miscibility and domain structure of PVNpoly(vinylpyrro1idone) (PVP) blends through the 13C CP MAS spectra and intermolecular cross-polarization. It was demonstrated that a hydrogenbonding interaction between the two polymers occurs. The 'H TI and Tlp results show that the blends are miscible for all components on a scale of 200-300A. On a scale of 20-30& however, the miscibility of the blends depends significantly on the composition. When the PVA composition is more than 46%, the blends are composed of two phases, an amorphous miscible phase of PVA and PVP and a pure PVA phase. The crystallinity of the PVA phase decreases rapidly with decreasing PVA composition. When the PVA composition is less than 46%, the blend is completely miscible. The composition of the PVA phase in the blends was inferred from the results of 'H Tb. Further, they have investigated the miscibility of PVNpoly(methacry1ic acid),lO6 PVA/poly(acrylic acid),"' polycarbonate/ PMMA'OB and poly(viny1 phenol)/poly(methyl acrylate)'@ blends using similar methods. Spevacek et ~ 1 . ' ~ 'have investigated the mechanism of stereocomplex formation for a mixture of isotactic and syndiotactic PMMAs by 13C CP MAS-NMR.

4.

NATURAL POLYMERS

To clarify the structures and functions of proteins is very important for understanding life processes and provides an index with respect to the design of artificial biomaterials. Therefore, new methods, which can exactly determine the protein structure, have been urgently required. Solid-state NMR is a powerful tool for the determination of protein structure and satisfies this requirement. The advantages of this method have been described in many monographs and review^.'''-'^^ In this section, through recent studies on fibrous, membrane, globular, and conjugated proteins, we

252

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

show that the solid-state NMR methods are useful for the structural analysis of proteins. 4.1. Fibrous proteins

Since fibrous proteins generally have periodical amino acid sequence and higher-order structure, the clarification of their fine structure in the solid state becomes very important not only when discussing the physical and chemical properties, but when obtaining information about the molecular design of synthetic polypeptides. Actually, the comparison of some fibrous proteins and their model synthetic polypeptides has been carried out. The conformation-dependent 13C CP MAS-NMR chemical ~ h i f t s ~ ' @are ' ~ ~particularly useful for the determination of the conformational features of fibrous proteins. Because silk fibroin is one of the fibrous proteins having a simple amino acid composition and the isotopic enrichment can be easily performed, the peak assignments of its NMR spectra are almost a c ~ o m p l i s h e d ' ~ ' -and ~~ more detailed analyses for static or dynamic structures have been made by solid-state NMR. 113~147-156 Amino acid composition of silk fibroin varies with the species of silkworms. In fact, fibroin from Phylosamia Cynthia ricini contains alanine (Ala) (48.4%), glycine (Gly) (32.2%), serine (Ser) (5.5%), and tyrosine (Tyr) (4.5%), while fibroin from Bombyx mori contains Gly (42.9%), Ala (30.0%), Ser (12.2%), and Tyr (4.8%).15' In reflecting the higher Ala content of P.c. ricini fibroin it is shown that the 13C NMR chemical shifts of the Ala residues of a dried sample, taken from the silk gland of P.c. ricini, are identical to those of the @-helical (Ala),. By contrast, the cocoon of P.c. ricini is found to have the anti-parallel /?-sheet form, as determined from the I3C NMR chemical shift of the Ala and Ser residues. On the other hand, the 13C peaks of the Ala and Gly residues of the silk I1 form of Bombyx mori silk fibroin are in good agreement with those of (Ala-Gly), I1 as well as (Gly), I1 and (Ala), of the p-sheet form. The I3C NMR chemical shifts of samples taking the silk I form are significantly displaced from those of the silk I1 form. Models for the silk I structure of B. mori silk fibroin have been proposed on the basis of 13C NMR data, X-ray diffraction data, and conformational energy calculations.15* Recently, on the basis of this information the solvent- and mechanical-treatment-induced conformational transitions of silk fibroins were studied by means of 13C CP MAS-NMR.152In addition, the dynamic features of side chains of silk fibroins in the solid state were also studied by 'H NMR and 13C CP MAS method^.^^','^^ For the Ser residue of B. mori silk fibroin, especially, it was suggested that the hydroxyl groups might be hydrogen-bonded to the C=O or NH groups in intra- or interchain bonding. Most recently, 15Nsolid-state NMR techniques were applied to elucidate the

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

253

atomic resolution details of the silk I1 conformation of silk f i b r o i n ~ . ~ ~ ~ , ~ ~ ' The "N NMR chemical shift tensors of the sites, ["NIGly, ["NIAla, and [l'N]Tyr were determined for the "N-labelled powder samples. 15N CP NMR spectra of the oriented silk fibroin block were observed by setting it parallel and perpendicular with respect to the applied magnetic field. The static spectra were simulated with 15N chemical shift tensors as functions of several Euler angles and distribution of fibre axis (Fig. 52). Then, the torsion angles (4, $) of the site-specific amino acid residues of silk fibroins were determined (Table 9), using the results of a I3C-l5N double-labelled model compound. Collagen has a rod-like shape, a dimension of about 3000 x 15 8, and is composed of a triple-stranded helix which is assembled into cylindrical fibrils having a diameter of 50-20008,.159 The individual triple chains are composed of a repeating pattern (Gly-X-Y), , where X and Y are frequently occupied by proline (Pro) and hydroxyproline (Hyp) residues, respectively. Torchia et al. have reported substantial works on the molecular dynamics of collagens by means of multinuclear solid-state NMR spectroscopy, using isotopically labelled samples which were obtained by cultivation technique^.'^'^'^^^^ Fujiwara and Kuboi have reported the 13C and 31P CP MAS-NMR spectra of dentin collagen.176Saito et al. have found that most of the 13C signals of collagen fibrils from bovine tendon and skin are in good agreement with those of (Pro-Gly-Pro), , although some peaks of collagen are slightly displaced from those of the model polypeptides. 177 Most recently, they attempted to assign all of the 13C CP MAS-NMR peaks of collagen fibrils on the basis of computer simulation by utilizing amino acid composition and chemical shift data from both the solid state and solution, and confirmed that some unassigned peaks were not ascribable to a denatured portion but to the minor amino acid residues.17* Furthermore, they found that the 13C spin-lattice relaxation times (TI) of both the C, and C, carbons of Pro and Hyp in fibrils are substantially reduced as compared with those of some crystalline oligopeptides. It was shown that the presence of rapid ring puckering motion in these residues results in a reduction of the NTl values, where N stands for the number of protons attached to the carbon under consideration. The I3C CP MAS-NMR signals of the soluble collagens in the anhydrous state are generally broader than those of the insoluble collagen and they are substantially narrowed upon hydration (Fig. 53). In particular, it is noteworthy that the carbonyl signal is split into an asymmetrical doublet or three signals upon hydration, which can be considered as characteristic of the presence of the native collagen structure. Collagens with various degrees of cross-linking were also studied by high-resolution liquid and solid-state 13C NMR methods, especially by means of relaxation time measurements. 179 13C CP MAS-NMR spectra of tropomyosin, one of muscle proteins, were measured in the solid state, in order to elucidate the higher-order structure

254

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. AND0

~.-.'-.-.

I . . . . I . . . . I

200

100

...

r I . . , . L

0

PP'n Fig. 52. "N solid-state NMR spectra of ['5N]Gly-labelled silk fibroin. Solid line: observed; broken line: calculated.

Table 9. I5N chemical shift tensors, Euler angles and torsion angles of B. mori silk fibroin.

Chemical shift tensors (ppm) c 1 1 a22

a33

Euler angles (deg) QF

PF

Torsion angles (deg)

*@

22 54 186

33 56 200

32 54 194

24 k 10 72+2

2_+10 70+2

22 10 72t2

-141 _+5 147 _+ 5

-139+5 146k5

*

-139 _+ 5 147 +_ 5

NMH STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

255

L

150

50

100

0

PPm

Fig. 53. 75.46MHz 13C CP MAS-TOSS-NMR spectra of insoluble bovine tendon (sigma) collagen (A and B) and acid-soluble collagen from calf skin (C and D). A and C, anhydrous sample; B and D, hydrated sample after humidification in a desiccator at 96% RH for 12 h.

of the protein through the observation of the I3C NMR chemical shifts of the amino acid residues and their mobility. 1m-182The higher-order structure of tropomyosin is a right-handed a-helix coiled-coil structure, which contains two different sites, which are characterized as the internal hydrophobic site and the external hydrophilic site (Fig. 54A).lS3A typical 13C CP MAS-NMR spectrum for tropomyosin in the solid state is shown in Fig. 54B. Since the Ala residue is one of the major components, the *3C

External hydrophilic site

C

L-AlaCo

Internal hydrophobic site

-

Side chain oliphatic carbons

1 c=o

I

I

150

.

.

”

l

.

100

.

~

‘

I

’

50

”

’

I....I’”’I’’..r.~..

l

0

21

20

19

18

8 ‘ .

17

*’1.-..1’...1

16

15

1L

13 6/PPM

‘/PPM Fig. 54. (A) A schematic picture of the cross-section of the coiled-coil structure in tropomyosin. (B) The typical 13C CP MAS-NMR spectrum of tropomyosin at room temperature. (C) The expanded 13C NMR spectra of the Ala C, carbon atoms of tropomyosin, measured using the inversion-recovery method at room temperature. Recovery time, t,: (a) 10 ms; (b) 500 ms; (c) 3500 ms.

N M R STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

257

signal which appears at 15-17ppm can be easily assigned to the Ala C, carbons, and it can be seen that this signal consists of two peaks. Figure 54C shows the expanded Ala C, signal, measured using the inversion-recovery method. From these spectra, there are at least two kinds of Ala C, carbons. Further, the carbons contributing to the peak at 15.8ppm have a longer TI value than those at 16.7ppm. This indicates that the former carbons are more mobile than the latter carbons, because the Ala C, carbons in tropomyosin have correlation times with values appearing in an extremely narrow region at toom temperature. The distance between the two a-helical axes in the coiled-coil structure is shorter than the coiled-coil helicles which are packed in parallel in native muscle from X-ray studies.ls4 Therefore, it can be said that the mobility of the Ala C, carbons in the internal site is expected to be more restricted than that in the external site. From such a situation, the two peaks at 15.8 and 16.7ppm are assigned to the Ala C, carbons in the external and internal sites of the coiled-coil structure, respectively. On the other hand, high-resolution solid-state I3C NMR studies of wool keratin proteins have been systematically carried out.185-189The native wool fibre consists of intermediate filaments (termed “microfibrils”) composed of low-sulphur proteins which are embedded in a non-filamentous matrix. The non-filamentous matrix usually contains two classes of proteins; one is a high-sulphur protein and the other is a protein containing Gly and Tyr residues. 190 Wool keratin can be divided into three main fractions after reducing disulphide bonds and protection of the resulting thiol groups with iodoacetic acid to form S-carboxymethyl kerateine (SCMK).’91.’92The I3C CP MAS-TOSS-NMR spectra of wool and four kinds of SCMKs extracted from wool [low-sulphur fractions (SCMKA), helix-rich fragments (SCMKAhf) , high-sulphur fractions (SCMKB), and high-Gly-Tyr fractions (HGT)] in the solid state are shown in Fig. 55, and the observed 13C chemical shift values in these samples are summarized in Table 10 together with the assignments which are made using their amino acid compositions and the 13C chemical shift data120-141with respect to homopolypeptides in the solid state. For characterization of the main chain conformation of polypeptides and proteins in the solid state, it is very useful to use the 13C chemical shift value of the main chain carbonyl carbon, because it is strongly influenced by the conformation of the main chain but not by varieties of amino acids Therefore, it can be said that and/or a specific amino acid the line shape of the 13C signals in the carbonyl region for proteins is varied with the conformations. The carbonyl 13C signal can be decomposed to four peaks by computer-fitting (Fig. 56) and the results for the main chain carbonyl carbons are summarized in Table 11. The relative intensity of the high frequency peak at ca. 176ppm corresponds to the proportions of the a-helix component because this peak comes from the main chain carbonyl carbons in the a-helix form. The proportion of the a-helix component

258

H.

Wool

SCMKA

SCMKA-hf

SCMKB

HGT

1

200

1

1

1

1

$0

1

1

1

I

lh0

I

I

I

(

I

PPM t

1

50

Fig. 55. I3C CP MAS-TOSS-NMR spectra of (a) wool, (b) SCMKA, (c) SCMKA-hf, (d) SCMKB, and (e) HGT and their expanded spectra for the carbonyl carbon region.

0

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

259

Table 10. Observed 13C NMR chemical shifts for wool, SCMKA, SCMKA-hf, SCMKB, and HGT in the solid state. Chemical shifts" (ppm) SCMKA SCMKA-hf

Wool ~-

~

~

SCMKB

HGT

~~~

175.3 173.2 156.2

176.0

176.3

155.1

157.2

128.5

129.1

128.2

72.2 68.6 65.2

172.8 154.8

71.8 67.8 64.4 60.6

56.6

56.9

56.4

54.8

40.5

40.3

40.3

40.1

36.7

36.6

28.9 25.2

30.8 25.4

(23.7) 21.0 (21.1)

(16.5) 16.2(16.5) 16.1 (16.5) 14.9 (14.8) 11.9(11.7) 12.1 12.4(11.9) 12.1

~-

Tyr C, and Phe C, d

Mainly CH, in lipid

30.2 24.6 (23.1) 20.6 (20.8)

20.3 (20.7)

~

Carbonyt carbons (ah) 172.3 Carbonyl carbons (ps) 156.0(156.4) Tyr C, and Arg C, 136.9 (137.3) Phe C, 128.9 (128.9) Phe CS.Eltand Tyr C, 116.3 TYr c, Thr c, (13s) Ser C, (Ps) Val C, and Ile C, (ah) Thr C, (ps) and Pro C, d 60.2 54.6 C, methine carbons 42.5 Gly CQ 38.4

36.3 35.6 30.1 (30.0) 28.8 25.0 25.4 (23.1) 20.3 (20.3) 20.7 (20.9)

Assignment" ~

d

d

Leu CB Mainly Val C, Mainly Thr C, (as) Mainly AIa C, (ah) Mainly Ile C,. Ile Cs

"The numbers in the parentheses are chemical shifts of the peaks observed in I3C CP MAS-DDph spectra. ?he assignment was made by reference data of homopolypeptides in the solid state (refs 120-136). a h and Bs in the parentheses mean the a,-helix and f3-sheet forms, respectively. 'Observed as the shoulder peak. dunassigned at this stage.

increases in the order of wool, SCMKA and SCMKA-hf. On the other hand, the low frequency peak at 172 ppm comes appreciably from the main chain carbonyl carbons in the P-sheet form. From the conformational characterization on the basis of the above assignment, it can be said that the P-sheet form is rich in SCMKB and HGT. These findings are also confirmed from the peaks in other carbon regions. Further, it is suggested that the coiled-coil structure exists in wool, SCMKA, and SCMKA-hf, because the 13Cchemical shift values of the Ala C, carbons in these samples coincide in experimental error with that of the Ala residue located in the internal site of the coiled-coil structure in tropomyosin (Table 11). This conformational

260

H. KUROSU, S. ANDO, H. YOSHNIZU AND I. ANDO

~~~~

~

SCMKA film cast from aqu. sol.

Fig. 56. I3C NMR spectra for the carbonyl carbon region in SCMKA deconvoluted by computer fitting with Gaussian functions. The minor peak appearing at about 166 ppm comes from the NMR rotor.

Table 11. Observed 13C NMR chemical shifts, half widths, and relative peak intensities of the main chain carbonyl carbons in wool, SCMKA, SCMKA-hf, SCMKB, and HGT."

Wool

SCMKA SCMKA-hf SCMKB HGT

13C chemical shift (ppm)

Half width (PPm)

Relative peak intensity (%)

176.3 172.2 176.2 172.5 176.4 173.0 176.0 172.5 176.6 172.2

4.0 4.8 4.0 4.8 3.6 4.5 3.7 4.6 4.0 5.6

42 58 56 44 65 35 25 75 8 92

"Determined by computer fitting.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

261

analysis, using the conformation-dependent 13C CP MAS-NMR chemical shifts, has been applied to estimate the conformational transition of SCMKA by stretching, heating, or ~ t e a m - t r e a t i n g . ' ~The ~ ' ~ P-sheet ~ form appears and the proportion of the a-helix form decreases upon stretching and steam-treating. For the SCMKA heated at 200°C for 3 h in vacuo, the 13C MAS-NMR spectrum shows that each peak is broader than those of other treated specimens. This indicates the existence of various conformations and/or different microenvironments in the heated SCMKA. Thus, it can be said that the random coil form appears by heating. From the X-ray diffraction, the a-helix form completely vanishes in the SCMKA heated under the same condition^.'^^ The result obtained by 13C CP MAS-NMR spectroscopy, however, indicates that the a-helix form remains to an appreciable extent in the heated sample, although the random coil form also appears. The difference between the results from X-ray diffraction and NMR spectroscopy suggests that only the packing of the ordered structure (the a-helix form) in the SCMKA is disrupted by heating, but that the secondary structure is retained. On the other hand, in the case of the structural transition for the same protein, the change of the line shape for the C, methine carbons can be mainly attributed to the conformational change. Figure 57 shows the observed and simulated 13C CP MAS-NMR spectra for the C, methine carbon region in SCMKA films. Spectrum simulation was performed by taking into account the experimental fact that the 13C chemical shift value of the C, methine carbon is influenced by both the main chain conformation and the variety of amino acid residues in contrast to the 13C chemicaI shift behaviour of the carbonyl carbons.120*121 As shown in Fig. 57, there is a tendency for the intensity of the peak appearing at about 52-53 ppm to increase as the stretching ratio is increased. This can be interpreted by means of the conformation dependence of the I3C chemical shifts. I3C CP MAS-NMR spectra of horse hoof, horse hair, parrot feather, and human hair were reported by Kricheldorf and Muller.'" The dynamics and structure of mouse keratin intermediate filaments were investigated by solid-state I3C and *H NMR spectroscopy, using isotopically labelled samples. 195 4.2. Membrane proteins

The structural and dynamic analyses of membrane proteins are well carried out by solid-state NMR. The reasons are as follows: first, biological membranes greatly restrict the motion of embedded proteins and consequently are difficult to study by solution NMR methods. Second, only a few membrane proteins have been crystallized and studied by X-ray diffraction methods. Third, if the membrane protein is uniaxially oriented in the lipid,

262

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

n Simulated

a:p/%

1oo:o

Observed

200% stretched 7

70

50

PPM

Fig. 57. Observed and simulated 13C CP MAS-NMR spectra for the C, methine carbon region in SCMKA films prepared from an aqueous solution. The Gaussian function for a peak with half width 3.5ppm is assumed, and the chemical shift positions of the amino acid residues constituting SCMKA determined from reference data of homopolypeptides are used. Further, the corresponding peak heights are estimated from the amino acid composition of SCMKA. Under these assumptions, the spectrum simulation is carried out as follows: at first, two reference spectra are simulated, one for the case in which all the amino acid residues take the a-helix form, and the other for the case in which all the amino acid residues take the p-sheet form. Next, using the two reference spectra, some spectra with a given specified ratio of the a-helix to &sheet form content are simulated, where this ratio is determined from the carbonyl carbon signal (see Fig. 55).

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

263

there is the advantage of being able to determine the structure at the atomic level by NMR as well as by X-ray crystallography. Fourth, since phospholipids in many cases coexist, the NMR method is still suitable because of its ability to observe selectively by means of isotopic labelling, multinuclear NMR, and by utilizing the differences of relaxation times. Therefore, phospholipids are well studied by solid-state NMR methods including 31P ~ ~ ~ . 1 9 6 . 1 9 7

Gramicidin A (GA) is a pentadecapeptide consisting of 15 alternating and D-amino acids and has the following chemical formula:'98

L-

This sequence allows the peptide to fold into a unique secondary structure, the P-helix. When GA is incorporated into a lipid environment, it forms a helical channel that transports monovalent cations. On the basis of a variety of spectroscopic data and model building, a structure has been proposed for the channel conformation that is an amino terminus to amino terminus dimer (the head-to-head dimer of single helices).199This structure has been confirmed by solid-state NMR Until recently, it was generally accepted that the sense of the helix was left-handed. 199?213 However, it has now been shown by solid-state NMR studies that in a fully hydrated lipid environment the helix is right-handed.206*207,212 13C CP MAS spectra of l3C-1abelled GA were measured to determine directly the structure of this peptide in a lipid The seven GA analogs, each labelled in a single carbonyl group of Gly2, ~ - A l a ~ , D-Leu4, ~ - V a l ~D-Leulo, , leu^^, or leu^^ were synthesized by the solid-phase method. These 13C-labelledGAS were incorporated into aligned multilayers of dimyristoylphosphatidylcholine (DMPC) or diether lipid bearing 14- or 16-carbon chains, at a 1:15 peptide :lipid molar ratio. 13CCP NMR spectra were obtained at several temperatures and over a range of sample orientations with respect to the spectrometer magnetic field. The results are summarized in Table 12. The observed chemical shielding anisotropies indicate that all of the labelled carbonyl bonds are oriented almost parallel to the molecular long axis and perpendicular to the lipid bilayer plane. These orientations are consistent with GA forming a p 6.3 single-strand helix that is oriented parallel to the methylene chains of the lipid molecules. Comparison of the line widths from labelled residues that are in the innermost turn of the helix (Gly2, ~ - A l a and ~ , leu^), in the centre of the molecule (L-Val,), and in the turn nearest the lipid bilayer surface (D-Leulo, leu,^, and leu^^) suggests that although the peptide behaves largely as a right-handed barrel, segments of the peptide close to the membrane surface possess greater motional freedom.

264

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

Table 12. Reduced chemical shift anisotropies (CSA) and line widths from the labelled carbonyl sites in gramicidin A analogues, and CSAs for lipid carbonyl and methylene resonances.

Label position

Lipid

Gly'

Gramicidin CSA (PP4

Gramicidin line width

Lipid CSAs

(Hi)

Carbonyl

Methylene

11f2

800f100

27+1

17 -+_ 1

16k1

800k100

28fl

15 k 1

DMPC ~

1

~

3

DTPC DHPC

I

D-Leu4

DMPC

12k2

800k100

2S&1

14f 1

Val'

DMPC

16k 1

800k100

30k 1

15-l-1

D-LeU'O

DHPC

DMPC1

9+2

400 k 70

25 f 1

16k1

D-Leu"

DMPC

11k1

400 2 70

25 k 1

16k 1

D-Leu14

DMPcl

13 +_ 1

27 k 1

14f 1

DHPC

Measurements were performed on samples containing a 1:15 peptidehpid molar ratio at 307 K in DMPC and DTPC, and at 325 K in DHPC.

Recently, an analytical method for the determination of torsion angles of the GA backbone from solid-state "N NMR spectroscopic data has been Advantage is taken of the "N-lH and 1sN-'3C dipolar interactions as well as the I5N chemical shift interaction in the oriented samples. The torsion angles for the L-Ala3 site are determined by obtaining the NMR data for both the Gly2-~-Ala3and ~ - A l a ~ - ~ - Lpeptide eu~ Fig. 58. Method for simulation of spectra obtained from oriented samples at temperatures below the phase transition. (a) The PAS (all, a=,4 representation of the "N chemical shift tensor is known in the defined initial position (4 = 0) with respect to the laboratory reference frame (X, Y, 2).The axis of local motion, e, is fixed relative to both frames of reference. The line defined by the intersection of planes C and R is placed coincident with Y, where B is the a l l a 3 3 plane and R is the plane perpendicular to e and passing through the origin. Since the chosen position in the X Y plane is arbitrary in the NMR experiment, when Bo is the Zlab direction, subsequent rotation of the PAS through 0 about e results in a new orientation that can be fully described by 4, the angle between e and the X Y plane, and 0. (b) Continuous line: as the PAS is rotated about e, the 2-component of the chemical shift tensor, gob,is plotted as a function of 0. Dotted line: a Gaussian probability distribution is assumed, characterized by the mean orientation, 0, and the amplitude, 6 (half width at half height). (c) The chemical shift spectrum resulting from applying the assumed probability distribution to the calculated trajectory of the chemical shift is shown.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

Y

I

I

I00

I

I

I

-100

0 U*@l ( P P m

1

265

266

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

linkages.208Two possible sets of torsion angles for the ~ - A l asite ~ are obtained (4,+ = -129", 153" and -129", 122"), both of which are consistent with a right-handed a-helix. In addition, the dynamics of the backbone of GA in DMPC bilayers have been investigated using solid-state "N NMR.209 "N CP NMR spectra of single-site "N-labelled GA at 8°C are analysed to yield a spatial model for local backbone motion. This model includes the axis of motion, the mean orientation, and the maximum amplitude of displacement for individual peptide planes (Fig. 58). Specific sites in the first turn of the amino terminus were investigated, with emphasis on the L-Ala3 and D-Leu4 linkages, for which the orientation of the "N chemical shift tensor with respect to the molecular frame has been determined. Figure 59 shows the simulation of spectra obtained from oriented bilayer samples containing specific-site "N-labelled GA at 8°C. Samples were placed in the magnet such that the bilayer normal was parallel to the direction of the field. The effect of bilayer defect structure, and of parabolic focal conics214 is included in the spectral simulation. From these simulations, it is found that only relatively small amplitude motions are possible at the two sites, with amplitudes of not more than f8" and f15" for the ~ - A l and a ~ D-Leu4 sites, respectively. Bacteriorhodopsin and rhodopsin are integral membrane proteins that contain vitamin A aldehyde retinal as a photoreactive chromophore. Many works on the structure and environment of the retinal chromophore in these proteins by means of solid-state NMR have been reported, using mainly specifically 13C-labelled retinals. 115,215-225 So far, it can be said that the structure of retinal in bacteriorhodopsin and rhodopsin has been determined only by solid-state NMR. Therefore, solid-state NMR is the only experimental procedure for the clarification of the sense of sight. Most recently, the 13C CP MAS-NMR spectra of 13C-labelled bacteriorhodopsin in purple membranes were measured.226 On the basis of the I3C CP MAS-NMR experiments, a discussion has been presented of the change of the manner of mutual orientation between a-helices as induced by hydratioddehydration. Fig. 59. Simulations of spectra obtained from oriented samples in the gel phase that superimpose the effects of local disorder due to backbone dynamics and global channel axis disorder due to the PFC (parabolic focal conics) defect structure. The chemical shift scales are relative to miso = Oppm. (a) Experimental spectrum recorded for the D-Leu4 site at 8°C. (b) Simulation for the D-Leu4 site using a local motion defined by 4 = 5", 0 = 262" and 6 = 15", and a slow global lateral diffusion through the PFC defect structure (10ppm line broadening applied). (c) Same as in (b), with 2 ppm line broadening applied. (d) Experimental spectrum recorded for the L-Alas site at 8°C. (e) Simulation for the L-Ala3 site using a local motion defined by 4 = 19.5", 0 = 82", and 6 = So, and a slow global lateral diffusion through the PFC defect structure (10 ppm line broadening applied). (f) Same as in (e), with 2 pprn line broadening applied.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

(a

l

I50

I00

50

0

-50

-I00

0

-so

-I00

p p m

(f

l I so

I00

SO

p p m

267

268

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

4.3. Miscellaneous biopolymers

The solid-state NMR method complements and extends other structural methods such as X-ray diffraction and solution NMR. Soluble proteins such as enzymes and conjugated proteins such as glycoproteins have been investigated by solid-state NMR.227-242 Ellis and coworkers have characterized the 13Cd chemical shielding tensor in Cd complexes, particularly those containing oxygen ligands, and then studied the metal binding site in lyophilized metalloprotein samples by solution and solid-state '13Cd NMR spectroscopy.22s233 Principal elements of the '13Cd chemical shielding tensor have been determined by spinning sideband analysis of CP MAS spectra for the calcium-binding sites in parvalbumin and for the manganese (Sl) and calcium (S2) sites in concanavalin A. Similarities in isotropic chemical shifts and in its tensor parameters for the parvalbumin calcium sites and for the S2 site in concanavalin A reflect similar metal-coordination environments in the two proteins. Santos et al. have also reported the solid-state '13Cd and '19Hg spectra of some model complexes for biologically occurring [M(SC y ~ ) ~ ( H i scentres ) ~ ] of metal lop rote in^.^^^*^^^ 13C CP MAS-NMR spectroscopy has been used to characterize covalent conjugates of alachlor, an a-chloroacetamide hapten, with glutathione (GSH) and bovine serum albumin (BSA).236The solid-state NMR method demonstrates definitively the covalent nature of these conjugates and can also be used to characterize the sites of 13C-labelled hapten attachment to proteins. Three different sites of alachlor binding are observed in the BSA system. Accurate quantitation of the amount of hapten covalently bound to GSH and BSA is possible. The solid-state 13C NMR technique can easily be generalized to study other small molecule/protein conjugates and can be used to assist the development and refinement of synthetic methods needed for the successful formation of such protein alkylation products. Similar methods have been applied to characterize enzyme-substrate interaction^^^^.^^' and aromatic cross-links in insect cuticle,239respectively. On the other hand, for proteins which are in more complex systems, molecular motion is mainly characterized by solid-state NMR. The motional dynamics of lens cytoplasmic proteins present in calf lens homogenates have been investigated by solid-state NMR techniques in order to define further the organizational differences between the cortex and nucleus.240Mainly from the results of the spin-locking experiments at several temperatures, it has been established that both mobile and solid-like protein phases are present in calf lens nuclear homogenate. However, for the native cortical homogenate, within the detection limits of NMR, the protein phase is mobile, except at low temperature where a small fraction of solid-like protein phases is present. The dynamics of intact lime cuticle and its two components, cutin and wax, have been similarly investigated.241For wheat

NMR STUDIES O F HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

269

gluten, four different subsets are identified by solid state and solution 13C NMR methods.242

5. CONCLUSION The present chapter shows that solid-state high-resolution NMR spectroscopy provides great perspectives for elucidating the high-order structure of polymers such an engineering plastics, high-performance polymers, polymer alloys and natural polymers in bulk. It is clear that solid-state highresolution NMR is an indispensable technique for the routine analysis as well as for the study of the structures and dynamics of polymers in bulk. Another important future role for solid-state high-resolution NMR might be an extension to NMR imaging as a tool for materials research, although this is outside the scope of this chapter.

REFERENCES 1. H. S. Gutowsky, A. Saika, M. Takeda and D. E. Woessner, J . Chem. Phys., 1957, 27, 537. 2. For example, (a) J. Schaefer and E. 0. Stejskal, Topics in Carbon-I3 NMR Spectroscopy (ed. G. C . Levy), Vol. 3, p. 283. Wiley, New York, 1979; (b) R. A. Komoroski (ed.), High Resolution NMR of Synthetic Polymers in Bulk. VCH Publishers, Florida, 1986; (c) I. Ando. T. Yamanobe and T. Asakura, Progress in NMR Spectroscopy (eds J. W. Emsley, J. Feeney and L. H. Sutcliffe), Vol. 22, p. 349. Pergamon Press, Oxford, 1990. 3. I. Ando, T. Sorita, T. Yamanobe, T. Komoto, H. Sato, K. Deguchi and M. Imanari, Polymer, 1985, 26, 1864. 4. D. L. VanderHart and F. Khoury, Polymer, 1984, 25, 1589. 5. W. L. Earl and W. L. VanderHart, Macromolecules, 1979, 12, 762. 6. T. Yamanobe, T. Sorita, T. Komoto and I. Ando, J. Mol. Struct., 1985, 31, 267. 7. H. Kurosu, T. Yamanobe and I. Ando, J . Mol. Strucf. (Theochern.), 1989, 201, 239. 8. T. Kitamaru, F. Horii and K. Murayama, Macromolecules, 1986, 19, 636. 9. H. Tadokoro, Kobunshi, 1983, 32, 202. 10. I. Ando and T. Yamanobe, Seni to kogyo, 1985,41, 255. 11. T. Nakai, J. Asada and T. Terao, J . Chem. Phys., 1988, 88, 6049. 12. I. Ando, T. Yamanobe, S. Akiyama, T. Komoto, H. Sato, T. Fujito, K. Deguchi and M. Imanari, Solid State Commun., 1987, 62, 785. 13. T. Yamanobe, M. Tsukahara, T. Komoto, J. Watanabe, I. Ando, I. Uematsu, K. Deguchi, T. Fujito and I. Imanari, Macromolecules, 1988, 21, 48. 14. S. Akiyama, T. Komoto, I. Ando and H. Sato, J. Polym. Sci.: Part B: Polym. Phys., 1990,28,587. 15. I. Ando, T. Yamanobe, S. Akiyama, T. Komoto, H. Sato, T. Fujito, K. Deguchi and M. Imanari, Solid State Commun., 1987, 62, 785. 16. I. Ando and G. A. Webb, Theory of NMR Parameters. Academic Press, London, 1983. 17. N. Kasai and M. Kakubo, IUPAC International Symposium on Macromolecular Chemistry, Prague, 1965, p. 491.

270

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

18. Q. Chen, T. Yamada, H. Kurosu, I. Ando, T. Shiono and Y. Doi, J . Polym. Sci.: Part B: Polym. Phys., 1992, 30,591. 19. J. J. Dechter, R. A. Komoroski, D. E. Axelson and L. Mandelkern, I . Polym. Sci. Polym. Phys., 1981, 19, 631. 20. B. Schroter and A. Posern, Makromol. Chem., 1981, 182,675. 21. R. Kitamaru, F. Horii and K. Murayama, Polym. Bull., 1982, 7, 583. 22. D. E. Axelson, High Resolution NMR Spectroscopy of Synthetic Polymers in Bulk (ed. R. A. Komoroski), Ch. 5, p. 197. VCH Publishers, Florida, 1986. 23. Q. Chen, T. Yamada, H. Kurosu, I. Ando, T. Shiono and Y. Doi, I. Mol. Srruct., 1991, 263, 319. 24. T. Terao, S. Maeda and A. Saika, Macromolecules, 1983, 16, 1535. 25. F. Horii, S. Hu, T. Ito, H. Odani, R. Kitamaru, S. Matsuzawa and K. Yamaura, Polymer, 1992,33, 2299. 26. (a) K. Kubo, T. Yamanobe, T. Komoto, I. Ando and T. Shiibashi, J. Polym. Sci. Pan B: Polym. Phys., 1989, 27, 929; (b) K. Kubo, I. Ando, T. Shiibashi, T. Yamanobe and T. Komoto, J. Polym. Sci.: Part B: Polym. Phys., 1991, 29, 57. 27. D. G. Powell and L. J. Mathias, J. Am. Chem. SOC., 1990, 112, 669. 28. G. R. Hatfield, J. H. Glans and W. B. Hammond, Macromolecules, 1990, 23, 1654. 29. L. J. Matias, D. G. Powell, J. Autran and R. S. Porter, Muter. Sci. Eng., 1990, A126, 253. 30. P. Holstein, J. Spevacek, D. Geschke and V. Thiele, Progr. Colloid Polym. Sci., 1991, 85,60. 31. S. Yamazaki, T. Kimura, K. Neki and N. Tamura, Anal. Sci., 1991, 7, 1613. 32, G. Carazzolo and M. M a d , 1. Polym. Sci., Part A , 1964, 1,965. 33. W. S. Veeman, E. M. Menger, W. Ritchey and E. de Boer, 1979, 12,924. 34. H. Kurosu, T. Komoto and I. Ando, J. Mol. Structures, 1988, 176, 279. 35. A. Hagemeyer, K. Schmidt-Rohr, H. W. Spiess, Adv. Magn. Reson., 1989, 13, 85. 36. M. Kobayashi, M. Murano and A. Kaji, Macromolecules, 1990, 23, 3051. 37. J. L. Bredas and G. B. Street, Acc. Chem. Res., 1985, 18, 309. 38. M. G. Kanatzldis, Chem. Eng. News, 1990,3,36. 39. J. L. Bredas, B. Themans, J. G. Fripiat and J. M. Andre, Phys. Rev. B, 1984, 29, 6761. 40. F. Devreux, G. Bidan, A. A. Syed and C. Tsintavis, J . Phys.., 1985,46, 1595. 41. B. Wehrle, H. H. Limbach, J. Mortensen and J. Heinze, Synth. Meth., 1990,38, 293. 42. M. Kikuchi, H. Kurosu and I. Ando, J. Mol. Struct., 1992, 269, 183. 43. N. Ogata, Doudenseikoubunshi (Conducting Polymers). Koudansha, Tokyo, 1990. 44. J. Schaefer and E. 0. Stejskal, Topics in Carbon-I3 NMR Spectroscopy (ed. G. C . Levy), Vol. 4, p. 283. Wiley-Interscience,New York, 1979. 45. M. A. Gomez and A. E. Tonelli, Polymer, 1991,32,796. 46. N. Zumbulyadis, J. M. O’Reilly and D. M. Teegarden, Macromolecules, 1992, 25, 3317. 47. J. Schaefer and E. 0. Stejskal, J . Am. Chem. Soc., 1976, 98, 1031. 48. J. Schaefer and E. 0. Stejskal, Macromolecules, 1977, 10, 384. 49. M. D. Sefcik, E. 0. Stejskal, R. A. McKay and J. Schaefer, Macromolecules, 1979, 12, 423. 50. S. A. Swanson, W. W. Fleming and D. C. Hofer, Macromolecules, 1992, 25, 582. 51. C. F.Fry and A. C. Lind, New Polym. Muter., 1990, 2, 235. 52. J. Schaefer and E. 0. Stejskal, Macromolecules, 1977, 10, 384. 53. J. R. Havens and J. L. Koenig, Appl. Spectrosc., 1983,37,226. 54. J. R. Lyerla, in High Resolution NMR Spectroscopy of Synthetic Polymers in Bulk (ed. R. A. Komoroski), Ch. 3. VCH Publishers, Florida, 1986. 55. J. Schaefer, E. 0. Stejskal, R. A. McKay and W. T. Dixon, Macromolecules, 1984, 17, 1479. 56. M. D. Sefcik, J. Schaefer, E. 0. Stejskal and R. A. McKay, Macromolecufes, 1980, 13, 1132.

NMR STUDIES O F HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

271

57. J. R. Havens, H. Ishida and J. L. Koenig, Macromolecules, 1981, 14, 1327. 58. J. Schaefer, B. Hansen, B. Bluemich and H. W. Spiess, J. Non-Cryst. Solids, 1991, 131-133, 777. 59. D. L. Tzou, P. Desai, A. S. Abhiraman and T. H. Huang, Polym. Prep. (Am. Chem. SOC.), 1990, 31, 157. 60. M. A. Gomez, M. H. Cozine and A. T. Tonelli, Macromolecules, 1988, 21, 388. 61. M. D. Poliks and J. Schaefer, Polym. Prep. (Am. Chem. SOC.), 1990, 31, 109. 62. S. A. Curran, C. P. LaClair and S. M. Aharoni, Macromolecules, 1991, 24, 5903. 63. M. Tsukahara, T. Yamanobe, T. Komoto, J. Watanabe and I. Ando, J. Mol. Struct., 1987, 159,345. 64. T. Yamanobe, M. Tsukahara, T. Komoto, J. Watanabe, I. Ando, I. Uematsu, K. Deguchi, T. Fujito and M. Imanari, Macromolecules, 1988, 21, 48. 65. T. Yamanobe, H. Tsukamoto, Y. Uematsu, I. Ando and I. Uematsu, J. Mol. Struct., 1993, 295, 25. 66. B. Mohanty, T. Komoto, J. Watanabe, I. Ando and T. Shiibashi, Macromolecules, 1989, 22, 4451. 67. B. Mohanty, J. Watanabe, 1. Ando and K. Sato, Macromolecules, 1990, 23, 4908. 68. S. Ishikawa, H. Kurosu and I. Ando, J. Mol. Struct., 1991,248,361. 69. S . Ishikawa and I. Ando, J. Mol. Struct., 1992, 273, 227. 70. N. Bloembergen, E. M. Purcell and R. V. Pound, Phys. Rev., 1948, 73,679. 71. L. J. Mathias, Polym. Commun., 1988, 29, 352. 72. B. C. Perry and J. L. Koenig, 1. Appl. Polym. Sci., Appl. Polym. Symp., 1989,43, 165. 73. H. W. Spiess, Pure Appl. Chem., 1987, 57, 1717. 74. H. W. Spiess, R. Ebelhauser and T. Fahmy, Makromol. Chem., Rapid Commun., 1984, 5, 333. 75. H. W. Spiess and R. Ebelhauser, Makromol. Chem., Rapid Commun., 1984, 5, 4038. 76. H. W. Spiess, H.Sillescu and E. Rossler, Polymer, 1985, 26, 203. 77. U. Pschorn, H. W. Spiess, B. Hisgen and H. Ringsdorf, Makromol. Chem., 1986, 187, 271 1. 78. F. Laupretre, C. Noel, W. N. Jenkins and W. N. Williams, Faraday Discws. Chem. SOC., 1985,79, 191. 79. T. Kato, G. M. A. Kabir and T. Uryu,J. Polym. Sci., A, Polym. Chem., 1989, 27, 1447. 80. J. Hong and G. S. Harbison, Colloidr Surf., 1990,45,313. 81. U. Wiesner, K. Schmidt-Rohr, C. Boeffel, U. Pawelzik and H. W. Spiess, Adv. Muter., 1990, 2, 484. 82. S. A. Curran, C. P. LaCIair and S. M. Aharoni, Macromolecules, 1991, 24, 5903. 83. D. Geschke and G. Fleischer, Prog. Colloid Polym. Sci., 1991, 85, 127. 84. S. Ishikawa and I. Ando, J. Mol. Structure, 1992, 271, 57. 85. D. R. Paul and S. Newman (eds), Polymer Blends, Academic Press, New York, 1978. 86. P. Caravatti, P. Neuenschwander and R. P. Ernst, Macromolecules, 1985, 18, 119. 87. G. C. Gobbi, R. Silvestri, T. P. Russell, J. R. Lyerla, W. W. Fleming and T. Nishi, I. Polym. Sci., Polym. Lett., 1987, 25, 61. 88. S. R. Hartmann and E. L. Hahn, Phys. Rev., 1962, 128, 2040. 89. J. F. Parmer, L. C. Dickinson, J. C. W. Chien and R. S. Porter, Macromolecules, 1987, 20,2308. 90. X. Zhang, A. Natansohn and A. Eisenberg, Macromolecules, 1990, 23,412. 91. C. W. Chu, L. C. Dickinson and J. C. W. Chien, J. Appl. Polym. Sci., 1990,41, 2311. 92. V. J. McBrierty and D. C. Douglas, J. Polym. Sci., Macromol. Rev., 1981, 16, 295. 93. K. Takegoshi and K. Hikichi, J. Chem. Phys., 1991, 94, 3200. 94. R. K. Harris, P. Jackson, L. H. Merwin, B. J. Say and G. Haegele, 1. Chem. SOC., Faraday Trans. I , 1988, 84, 3649. 95. L. A. Belfiore, T. J. Litz, C. Cheng and C. E. Bronnimann, J. Polym. Sci., B, Polym. Phys., 1990,28,1261.

272

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

96. J. Grobelny, D. M. Rice, F. E. Karasz and W. J. MacKnight, Macromolecules, 1990, 23, 2139. 97. A. Simmons and A. Natansohn, Macromolecules, 1991, 24, 3651. 98. S. Percec and T. Hammond, Polymer, 1991, 32, 1252. 99. C. Zhang, P. Wang, A. A. Jones, P. T. Inglefield and R. P. Kambour, Macromolecules, 1991, 24,338. 100. Y. H. Chin, C. Zhang, P. Wang, P. T. Inglefield, A. A. Jones, R. P. Kambour, J. T. Bendler and D. M. White, Macromolecules, 1992, 25, 3031. 101. R. A. Grinsted and J. Koenig, J. Polym. Sci., B, Polym. Phys., 1990, 28, 177. 102. R. Ibbett, D. Bucknali and H. S. Higgins, Polymer, 1992,33,423. 103. J. F. Masson and R. St. J. Manley, Macromolecules, 1992, 25, 589. 104. S. N. Sarkar and R. A. Komoroski, Macromolecules, 1992, 25, 1420. 105. X. Zhang, K. Takegoshi and K. Hikichi, Polymer, 1992, 33, 712. 106. X. Zhang, K. Takegoshi and K. Hikichi, Polym. J., 1991,23,79. 107. X. Zhang, K. Takegoshi and K. Hikichi, Polym. J . , 1991,23, 83. 108. A. Asano, K. Takehoshi and K. Hikichi, Polym. J . , 1992,24,473. 109. X. Zhang, K. Takegoshi and K. Hikichi, Polym. J . , 1992,24,1403. 110. B. Schneider, J. Spevacek, J. Straka and J. Stokr, Makromol. Chem., Macromol. Symp., 1990, 34,213. 111. D. A. Torchia and D. L. VanderHart, in Topics in Carbon-I3 NMR Spectroscopy (ed. G. C. Levey), Vol. 3, Ch. 5. Wdey, New York, 1979. 112. C. A. Fyfe, Solid State NMR for Chemists. C.F.C. Press, Guelph, Canada, 1983. 113. H. Saito, T. Tabeta, A. Shoji, T. Ozaki, I. Ando and T. Asakura, in Magnetic Resonance in Biology and Medicine (eds G. Govil, C. L. Kheterapal and A. Saran), p. 195. Tata McGraw-Hill, New Delhi, 1985. 114. G. A. Hatfield and G. E. Maciel, Methods Protein Anal., 1988, 214. 115. S. 0. Smith and R. G. Griffin, Ann. Rev. Phys. Chem., 1988, 39, 511. 116. (a) S. J. Opella and P. L. Stewart, Methods Enzymol., 1989, 176, 242; (b) K. V. Ramanathan and S . J. Opella, Trans. Am. Crystallogr. Assoc., 1989, 24, 145; (c) S . J. Opelia, Biol. Magn. Reson., 1990,9,177; (d) L. E. Chirlian and S. J. Opella, New Polym. Muter., 1990, 2, 278. 117. M. A. Keniry, Methods Enzymol., 1989, 176,376. 118. M. A. Hemminga and T. W. Poile, Spec. Publ. R. SOC. Chem., 1989, 74, 74. 119. A. E. Derome and S . Bowden, Chem. Rev., 1991, 91, 1307. 120. H. Saito and 1. Ando, Ann. Rep. NMR Specfrosc., 1989,21,210. 121. A. Shoji, S. Ando, S. Kuroki, I. Ando and G. A. Webb, Ann. Rep. NMR Spectrosc., 1993, 26, 55. 122. T. Taki, S. Yamashita, M. Satoh, A. Shibata, T. Yamashita, R. Tabeta and H. Saito, Chem. Lett., 1981, 1803. 123. H. Saito, R. Tabeta, A. Shoji, T. Ozaki and I. Ando, Macromolecules, 1983, 16, 1050. 124. H. Saito, R. Tabeta, I. Ando, T. Ozaki and A. Shoji, Chem. Lett., 1983, 1437. 125. I. Ando, H. Saito, R. Tabeta, A. Shoji and T. Ozaki, Macromolecules, 1984, 17, 457. 126. A. Shoji, T. Ozaki, H. Saito, R.Tabeta and I. Ando, Macromolecules, 1984, 17, 1472. 127. A. Shoji, T. Ozaki, H. Saito, R. Tabeta and I. Ando, Makromol. Chem., Rapid Commun., 1984, 5, 799. 128. T. Yamanobe, I. Ando, H. Saito, R. Tabeta, A. Shoji and T. Ozaki, Bull. Chem. SOC. Jpn, 1985, 58, 23. 129. T. Yamanobe, I. Ando, H. Saito, R. Tabeta. A. Shoji and T. Ozaki, Chem. Phys., 1985, 99, 259. 130. S . Ando, T. Yamanobe, I. Ando, A. Shoji, T. Ozaki, R. Tabeta and H. Saito, J . Am. Chem. SOC., 1985, 107, 7648. 131. S . Tuzi, T. Komoto, I. Ando, H. Saito, A. Shoji and T. Ozaki, Biopolymers, 1987, 26, 1983.

NMR STUDIES OF HIGHER-ORDER STRUCTURES OF SOLID POLYMERS

273

132. H. Miyamoto, R. Takezaki, T. Komoto and I. Ando, J . Mol. Struct., 1988, 172, 395. 133. H. Saito, R. Tabeta, F. Formaggio, M. Crisma and C. Toniolo, Biopolymers, 1988, 27, 1607. 134. S. Ando, I. Ando, A. Shoji and T. Ozaki, J . Mol. Struct., 1989, 192, 153. 135. S. Ando, K. Matsumoto, I. Ando, A. Shoji and T. Ozaki, J. Mol. Struct., 1989, 212, 123. 136. H. Miyamoto, T. Komoto, H. Kurosu, I. Ando, T. Ozaki and T. Shoji, J. Mol. Struct., 1990,220,251. 137. H. R. Kricheldorf, M. Mutter, F. Mazer, D. Muller and D. Forster, Biopolymers, 1983, 22, 1357. 138. H. R. Kricheldorf and D. Muller, Macromolecules, 1983, 16, 615. 139. H. R. Kricheldorf, D. Muller and K. Ziegler, Polym. Bull., 1983, 9, 284. 140. H. R. Kricheldorf and D. Muller, Znt. J. Bwl. Macromolecules, 1984, 16, 145. 141. D. Muller and R. H. Kricheldorf, Polym. Bull., 1981, 6 , 101. 142. T. Asakura, Suzuki and Y. Watanabe, Macromolecules, 1983, 16, 1024. 143. T. Asakura, Y. Watanabe, A. Uchida and H. Minagawa, Macromolecules, 1984, 17, 1075. 144. T. Asakura, Y. Watanabe and T. Itoh, Macromolecules, 1984, 17, 2421. 145. T. Asakura and T. Murakami, Macromolecules, 1985, 18, 2614. 146. T. Asakura, H. Kashiba and H. Yoshimizu, Macromolecules, 1988, 21, 644. 147. H. Saito, Y. Iwanaga, R. Tabeta and T. Asakura, Chem. Left., 1983,427. 148. H. Saito, R. Tabeta, T. Asakura, Y. Iwanaga, A. Soji, T. Ozaki and I. Ando, Macromolecules, 1984, 17, 1405. 149. T. Asakura, A. Kuzuhara, R. Tabeta and H. Saito, Macromolecules, 1985, 18, 1841. 150. H. Saito, R. Tabeta, A. Kuzuhara and T. Asakura, Bull. Chem. SOC. Jpn, 1986,59, 3383. 151. H. Saito, M. Ishida, M. Yokoi and T. Asakura, Macromolecules, 1990, 23, 83. 152. M. Ishida, T. Asakura, M. Yokoi and H. Saito, Macromolecules, 1990, 23, 88. 153. T. Asakura, M. Demura, Y. Watanabe and K. Sato, J. Polym. Sci., Polym. Phys., 1992, 30, 693. 154. T. Asakura, T. Konakazawa, M. Demura, K. Komatsu and T. Itoh, 31st Meeting on NMR, Himeji, Hyogo, 1992,285. 155. L. K. Nicholson, Biopolymers, submitted. 156. X. Han, J. Guo, J. Hu, T. Ji, X. Zhang, J. Hu and J. Qui, Bopurue Zazhi, 1990, 8, 245. 157. R. D. Fraser and T. P. MacRae, Conformation in Fibrous Proteins. Academic Press, New York, 1973. 158. T. Asakura and T. Yamaguchi, J . Seric. Sci. Jpn, 1987,56, 300. 159. W. Traub and K. A. Piez, Adv. Protein Chem., 1971, 25, 243. 160. D. A. Torchia and D. L. VanderHart, J . Mol. Biol., 1976, 104, 315. 161. L. W. Jelinski and D. A. Torchia, 1. Mol. Biol., 1979, 133, 45. 162. L. W. Jelinski and D. A. Torchia, J. Mol. Biol., 1980, 138, 255. 163. L. W. Jelinski and D. A. Torchia, Frontiers in Protein Chemistry, p. 89. Elsevier-North Holland, Amsterdam, 1990. 164. L. W. Jelinski, C. E. Sullivan, L. S. Batchelder and D. A. Torchia, Biophys. J., 1980, 32, 515. 165. L. S. Barchelder, C. E. Sullivan, W. Jelinski and D. A. Torchia, Proc. Natl Acad. Sci. USA, 1982, 79,386. 166. D. A. Torchia, Methods Enzymol., 1983, 82, 174. 167. S. K. Sarkar, C. E. Sullivan and D. A. Torchia, J. Biol. Chem., 1983, 258, 9762. 168. S. K. Sarkar, D. A. Torchia, K. D. Kopple and D. L. VanderHart, J . Am. Chem. SOC., 1984, 106, 3328. 169. D. A. Torchia, Annu. Rev. Biophys. Bioeng., 1984, 13, 125. 170. S. K. Sarkar, C. E. Sullivan and D. A. Torchia, Biochembfry, 1985,24, 2348. 171. D. A. Torchia, Y. Hiyama, S. K. Sarkar, C. E. Sullivan and P. E. Young, Biopolymers, 1985,24,65. 172. S. K. Sarkar, P. E. Young and D. A. Torchia, J . Am. Chem. SOC., 1986, 108, 6459.

274

H. KUROSU, S. ANDO, H. YOSHIMIZU AND I. ANDO

173. Y. Hiyama, J. V. Silverton, D. A. Torchia, J. T. Gerig and S. J. Hammond, J . Am. Chem. SOC., 1986, 108, 2715. 174. S . K. Sarkar, Y. Hiyama, C. H. Niu, P. E. Young, J. T. Gerig and D. A. Torchia, Biochemistry, 1987, 26, 6793. 175. Y. Hiyama and D. A. Torchia, Basic Life Sci., 1990, 56, 273. 176. R. Fujiwara and Y. Kuboi, Biochem. Biophys. Res. Commun., 1990, 167, 761. 177. H. Saito, R. Tabeta, A. Shoji, T. Ozaki, I. Ando and T. Miyata, Biopolymers, 1984, 23, 2279. 178. H. Saito and M. Yokoi, J. Biochem., 1992, 111,376. 179. M. Bennet, L. Foucat, J. Dupuis and J. P. Renou, J. Chim. Phys. Phys.-Chim. Biol., 1992, 89, 227. 180. S. Tuzi and I. Ando, 1. Mol. Struct., 1989, 1%, 317. 181. S. Tuzi, S. Sakamaki and I. Ando, J. Mol. Struct., 1990, 221, 289. 182. S. Tuzi, PhD Thesis, Tokyo Institute of Technology, 1990. 183. P. L. Privalov, Adv. Protein Chem., 1982, 35, 1. 184. C. Cohn and K. C. Holes, J. Mol. Biol., 1%3,6, 423. 2908. 185. H. Yoshimizu and I. Ando, Macromolecules, 199O,W, 186. H. Yoshimizu, H. Mmura and I. Ando, Macromolecules, 1991, 24, 862. 187. H. Yoshimizu, H. Mimura and I. Ando, J . Mol. Struct., 1991, 246, 367. 188. H. Yoshimizu, PhD Thesis, Tokyo Institute of Technology, 1991. 189. H. Mimura, H. Yoshimizu and I. Ando, 30th Meeting on NMR, Tokyo, 1991, p. 291. 190. For example, (a) R. D. Fraser, T. P. MacRae and G. E. Rogers, Keratins: Their Composition, Structure and Biosynthesis. Charles C. Thomas, Springfield, IL, 1972; (b) J. H. Bradbury, Adv. Protein Chem., 1973,27, 111; (c) R. D. Fraser and T. P. MacRae, Conformationin Fibrous Proteins, Ch. 16. Academic Press, New York, 1973. 191. I. J. O’Donnel and E. 0.P. Thompson, Aust. 1. Biol. Sci., 1964, 17, 973. 192. L. M. Dowling and W. G. Crewther, Prep. Biochem., 1974, 4, 203. 193. H. Sakabe, T. Miyamoto and H. Inagaki, Sen-i Gukkuishi, 1981, 37, 273. 194. H. R. Kricheldorf and D. Muller, Colloid Polym. Sci., 1984, 262, 856. 195. J. W. Mack, D. A. Torchia and M. Steinert, Biochemistry, 1988, 27, 5418. 196. H. Tournois, J. L. Bijvelt, C. W. M. Haest, J. Ier and B. Kruijff, Biochemistry, 1987, 26, 6613. 197. T. Odahara, H. Akutsu and Y. Kyogoku, Biochemistry, 1990,29,5968. 198. R. Sages and B. Witkop, Biochemistry, 1%5,4, 2491. 199. D. W. Urry, Proc. Nati Acad. Sci., USA, 1971, 68, 672. 200. R. Smith and B. A. Cornell, Biophys. J., 1986,49, 117. 201. B. A. Cornell, F. Separovic, A. J. Baldassi and R. Smith, Biophys. J . , 1988, 53, 67. 202. R. Smith, D. E. Thomas, F. Separovic, A. R. Atkins and B. A. Cornell, Biophys. J., 1989, 56, 307. 203. F. Separovic, K. Hayamizu, R. Smith and B. A. Cornell, Chem. Phys. Lett., 1991, 181, 157. 204. L. K. Nicholson, F. Moll, T. E. Mixon, P. V. LoGrasso, J. C. Lay and T. A. Cross, Biochemistry, 1987, 26, 6621. 205. L. K. Nicholson. P. V. LoGrasso and T. A. Cross, J. Am. Chem. SOC., 1989, 111, 400. 206. P. V. LoGrasso, L. K. Nicholson and T. A. Cross, J. Am. Chem. Soc., 1989, 111, 1910. 207. L. K. Nicholson and T. A. Cross, Biochemistry, 1989,28, 9379. 208. Q. Teng, L. K. Nicholson and T. A. Cross, J. Mol. Biol., 1991, 218, 607. 209. L. K. Nicholson, Q. Teng and T. A. Cross, J. Mol. Biol., 1991, 218, 621. 210. S. W. Chiu, L. K. Nicholson, M. T. Brenneman, S . Subramaniam, Q. Teng, J. A. McCammon, T. A. Cross and E. Jakobsson, Biophys. J., 1991, 60, 974. 211. J. H. Davis, Biochemistry, 1988,27,428. 212. R. S. Prosser, J. H. Davis, F. W.Dahlquist and M. A. Lindorfer, Biochemistry, 1991,30, 4687.

NMR STUDIES OF HIGHER-ORDER STRUCTURES O F SOLID POLYMERS

275

213. D. W. Urry, J. T. Walker and T. L. Trapane, J. Membr. Biol., 1982, 69, 225. 214. (a) M. Kleman, C. E. Williams, M. J. Costello and T. Gulik-Krywicki, Phifos. Mug., 1977, 35, 33; (b) C. S. Rosenblantt, R. Pindak, N. A. Clark and R. B. Meyer, J . Phys.., 1977, 38, 1105; (c) S. A. Asher and P. S. Pershan, Biophys. J., 1979, 27, 393. 215. G. S. Harbison, J. Herzfeld and R. G. Griftin, Biochemistry, 1983,22, 1. 216. G . S. Harbison, S. 0. Smith, J. A. Pardoen, P. P. J. Mulder, J. Lugtenburg, J. Herzfeld and R. G. Griffin, Biochemistry, 1984,23,2662. 217. G. S. Harbison, S. 0. Smith, J. A. Pardoen, C. Winkel, J. Lugtenburg, J. Herzfeld, R. A. Mathies and R. G. Griffin, Proc. Natl Acad. Sci. USA, 1984, 81, 1706. 218. G. S. Harbison, S. 0. Smith, J. A. Pardoen, J. M. L. Courtin, J. Lugtenburg, J. Herzfeld, R. A. Mathies and R. G. Griffin, Biochemistry, 1985,24, 6955. 219. G. S. Harbison, P. P. J. Mulder, J. A. Pardoen, J. A. Lugtenburg, J. Herzfeld and R. G. Griffin, J. Am. Chem. SOC., 1985, 107,4809. 220. A. Schmidt, S. 0. Smith, D. P. Raleigh, J. E. Roberts, R. G. Griffin and S. Vega, J, Chem. Phys., 1986,85,4248. 221. S. 0. Smith, I. Palings, V. Copie, D. P. Raleigh, J. Courtin, J. A. Pardoen, J. Lugtenburg, R. A. Mathies and R. G. Griffin, Biochemistry, 1987, 26, 1606. 222. H. J. M. De Groot, V. Copie, S. 0. Smith, P. J. Allen, C. Winkel, J. Lugtenburg, J. Hertzfeld and R. G. Griffin, J. Magn. Reson., 1988,77,251. 223. S. 0. Smith, J. Couritin, E. van den Berg, C. Winkel, J. Lugtenburg, J. Herzfeld and R. G . Griffin, Biochemistry, 1989, 27, 237. 224. S. 0. Smith, H. J. M. De Groot, R. Gebhard, J. M. L. Courtin, J. Lugtenburg, J. Herzfeld and R. G. Griffin, Biochemistry, 1989,28,8897. 225. L. P. J. Mollevanger, A. M. Kentgens, J. Pardoen, J. L. Courtin, W. Veeman, J. Lugtenburg and W. J. de Grip, Eur. J . Biochem., 1987,163,9. 226. S. Tuzi, A. Naito and H. Saito, 3Jsr Meeting on NMR, Himeji, Hyogo, 1992, p. 51. 227. V. D. Fedotov, N. P. Obuchov, J. Spevacek and J. Straka, Makromol. Chem. Macromol. Symp., 1990,34,249. 228. M. A. Hemminga, P. A. deJager, J. Kruse and R. M. J. N. Lamerichs, 1. Mugn. Reson., 1987, 71, 446. 229. R. S . Honkonen and P. D. Ellis, J. Am. Chem. Soc., 1984, 106,5488. 230. P. S. Marchetti, P. D. Ellis and R. G. Bryant, J . Am. Chem. SOC., 1985, 107, 8191. 231. R. S. Honkonen, P. S. Marchetti and P. D. Ellis, J . Am. Chem. SOC., 1986, 108, 912. 232. P. S. Marchetti, R. S. Honkonen and P. D. Ellis, J. Mugn. Reson., 1987, 71, 294. 233. P. S. Marchetti, L. Bhattacharyya, P. D. Ellis and C. F. Brewer, J. Magn. Reson., 1988, 80,417. 234. R. A. Santos, E. S. Gruff, S. A. Koch and G. S. Harbison, J. Am. Chem. SOC.,1990,112, 9257. 235. R. A. Santos, E. S. Gruff, S. A. Koch and G. S. Harbison, J. Am. Chem. Soc., 1991,113, 469. 236. J. R. Garbow, H. Fujiwara, C. R. Sharp and E. W. Logusch, Biochemistry, 1991, 30, 7057. 237. C. M. Dobson and L. Y.Lian, FEBS Lett., 1987, 225, 183. 238. A. I. Scott, J . Mol. Catal., 1988, 47, 139. 239. J. Schaefer, K. J. Kramer, J. R. Garbow, G. S. Jacob, E. 0. Stejskal, T. L. Hopkins and R. D. Speirs, Science, 1987, US,1200. 240. C. F. Morgan, R. Scheich, H. Caines and N. Famsworth, Biochemistry, 1989, 28,5065. 241. J. R, Garbow and R. E. Stark, Macromolecules, 1990, 23,2814. 242. P. S. Belton, S. Duce and A. S. Tatham, Int. J . Biol. Macromol., 1987, 9 , 357.

This Page Intentionally Left Blank

N M R Studies of Organic Thin Films F. D. BLUM Department of Chemistry and Materials Research Center, University of MissouriRolla, Rolla, MO 65401, USA 1. Introduction 2. Background 2.1. Prior reviews 2.2. General NMR considerations 3. Polymers at interfaces 3.1. Solid-liquid systems 3.2. Elastomer-solid systems 3.3. Solid-gas interface 3.4. Solid-solid systems 4. Surface-active agents 4.1. Bonded phases 4.2. Non-bonded phases 5. Conclusions Acknowledgements References

277 278 278 279 280 283 290 295 299 300 300 313 315 315 315

1. INTRODUCTION In many respects, the most exciting and important developments in science and technology often occur at interfaces. For technological advances this is because of the different backgrounds and views of the scientists involved. In the materials area, different species can be combined in ways which produce new materials whose properties can be significantly different from any component. In general, the properties of materials containing multiple phases are not necessarily the superposition of the properties of the components. The influence of the interfacial layer or layers plays a significant role in the determination of the bulk properties in a wide range of materials. The purpose of the present work is to review NMR studies of thin organic films-particularly those at the interface with solid substrates. These typically have relevance in composite materials used for structural applications, although other applications of these materials in catalysis, separations, and other surface modifications may also be important. For the purpose of ANNUAL REPORTS ON NMR SPECIROSCOPY VOLUME 28 ISBN 0-12-505328-2

Copyright 0 1994 Academic Press Limited All rights of reproduction in any form reserved

278

F. D. BLUM

this review, the term “film” implies something that is both semi-permanent and semi-continuous. Thus, this review will be restricted to organic materials strongly bound to a substrate, including those which are permanently attached to a surface. Adsorbed polymers and non-covalently bound surfactants are included; however , other small molecules that are simply “adsorbed” reversibly and discontinuously are not. For example, this excludes materials used as catalysts such as zeolites. A general discussion of NMR studies focusing on polymer blends, interpenetrating networks, and multiphase (crystalline/amorphous) polymer systems has also been excluded. The major goal of this work is to review how NMR can yield specific information on the structure and dynamics of interfacial material and how this might be different from that in the bulk phase. In some respects, the number of papers in this area is not yet large-there are many more papers that deal with the interaction of small molecule adsorbates. The lack of papers is probably because of the inherent insensitivity of the NMR technique and the limited ways in which one can distinguish between bulk and interfacial materials. However, for appropriate systems, the information obtained from NMR experiments is of exceptional importance in the understanding of microscopic properties of interfacial thin films.

2. BACKGROUND 2.1. Prior reviews

Several authors have previously reviewed NMR studies of systems related to those included in this report. Resing’ reviewed the relaxation effects of adsorbed chemical species and Pfeifer,2 the basic principles of NMR of surface adsorbed species focusing on how their adsorption might affect common NMR parameters. Special emphasis was placed on phenomena such as exchange and its effects on the line shapes, diffusion, and relaxation behaviour. Examples of these included (1) silicdbenzene; (2) surface hydroxyl groups interacting with adsorbates; and (3) NaX-zeolite/water systems. An extensive literature survey was also included as a table. Later Pfeifer et d 3extended this work to cover the NMR studies of molecules adsorbed on porous solids, with special emphasis on zeolites. Advances in cross-polarization (CP) and magic angle spinning (MAS) have significantly increased the knowledge of the properties of both the adsorbate and substrate in these systems. Duncan and Dybowski4 have reviewed catalytic and chemically adsorbed species. All three of these reviews have some overlap with the current topics, but none of them centres on them. More recently, a few reviews have been partly targeted at the characterization of surface-bound alkylsilanes5-’0 and/or polymer^'^'^ adsorbed on

NMR STUDIES OF ORGANIC THIN FILMS

279

solid surfaces. Zaper and Koenig' have reviewed the applications of high-resolution solid-state NMR to both coupling agents and catalysts. For the coupling agents, coverage of 13C and 29Si chemical shifts, crosspolarization and relaxation rates provided an initial picture of the structure and dynamics of these species. Boddenberg6 has reported a variety of deuterium NMR studies of labelled chemical and physisorbed materials including a silane, but most of the material covered was from non-bonded small organic molecules. Blum7 has focused on the use of isotopes in surface-bound species such as silane coupling agents. The NMR-active isotopes provided the contrast and sensitivity necessary to probe the interface, even in composite materials. Albert and Bayerg have reported on the applications relative to bonded phases in chromatography, focusing on the characterization and modifications of silica gel and silica. Legrand et aL9 have a limited review of solid-state NMR of materials. Choli" has reported on the use of multinuclear NMR in industrial applications of polymeric materials. He covered a few select applications of coupling agents and also polymers and at interfaces. The NMR behaviour of polymers at interfaces has received less attention. Grandjean" has recently reviewed studies of interfacial phenomena with a focus on interfacial molecules, ions at the interface, and interphase components. This report included a few of the studies of polymers adsorbed at interfaces. Cosgrove et u L . ' ~ ~ ' have ~ described NMR studies of interfacial polymers with particular emphasis on the work done at Bristol. Blum14 has also reviewed experiments on polymers at interfaces in colloidal suspensions, composites, and surface coatings. 2.2. General NMR considerations

The fundamental reason for using NMR to study thin films is to gain information on either structure or dynamics, or both. Structural studies are used to determine the composition and bonding of the interfacial species (perhaps across the interface). NMR is also one of the few ways to determine the molecular motion which occurs in this very thin layer. The relationship between the physical properties of polymers and molecular motion is established, but it is difficult to probe the physical properties of interfacial layers that may be as thin as say 10-100pm. Consequently, it is hoped that the measurement of the mobility of the interfacial dynamics will allow the prediction of its physical properties. There are several basic advantages that NMR has for the study of interfacial or interphase phenomena. One of the most significant is that the optical clarity of the sample is unimportant. This is contrasted with optical studies, where this is not the case. Because of this, NMR has the ability to probe the inside of composite materials for most non-ferrous materials. For

280

F. D. BLUM

these heterogeneous materials, magnetic susceptibility effects usually do not create major complications in obtaining spectra. A wide variety of different NMR experiments are available to the experimenter, although the choice of “solid-like” or “liquid-like” experiments are often dictated by the properties of the system. Advances in the application of solids NMR techniques’’ have significantly broadened the potential use of NMR to organic films. In general, most organichnterface systems offer a variety of nuclei to probe. Generally, different nuclei in the same system, even in the same molecule, provide complementary information. Different experiments on the same nucleus can be made to elucidate different surface features. The measurement of NMR parameters including isotropic chemical shift and chemical shift tensor, relaxation and crosspolarization parameters, plus line shape analyses provide information on different aspects of the system. In many cases, contrast in the NMR experiment can be achieved through the creative use of different pulse sequences. For example, the surface of the 29Si atoms in silica may be probed selectively through the use of cross-polarization. The transfer of magnetization through the protons is only efficient near the surface because that is where the protons reside. Isotope enrichment can also provide the necessary contrast between bulk and interfacial material. Common nuclei for this include 2H, 13C, and possibly ”N.’ Since these isotopes have low isotopic abundance, background from naturally abundant isotopes does not usually interfere in spectra with labelled species. For example, this contrast is needed to distinguish the carbons in a silane coupling agent on a silica surface from the carbon used in the matrix resin. There are some serious limitations to NMR studies of interfacial species. The chief problem is sensitivity. The inherent insensitivity of NMR is such that high surface area materials need to be used as substrates. While many of those surfaces studied are indicative of the behaviour in the actual application, it would be preferable to study the actual system. For example, studies of the polymer coating on an aluminium container may have to be simulated by studies of the polymer coating on high surface area alumina.

3. POLYMERS AT INTERFACES The adsorption of polymers at interfaces is rather different from that of small molecules. Polymers may associate with the surface through covalent or ionic bonds, polar, and even van der Waals forces. The range of these direct interactions with surface groups is quite short and of the order of a few Angstroms. If this were the whole effect, the adsorption and dynamics of attached polymers would be similar to small molecules. Since the polymer segments attached to the interface are covalently bound to other segments,

NMR STUDIES OF ORGANIC THIN FILMS

281

the surface binding may impact the entire chain. In fact, it may even impact neighbouring chains. The two obvious effects of this interaction are restrictions of (1) overall motions of the adsorbed chain and (2) conformational states and internal motions of the chain segments. The importance of the adsorption of polymers at interfaces is such that a number of reviews exist on the structure, kinetics, and dynamics of adsorbed It is convenient to think of adsorbed polymers as macromolecules. either randomly or terminally attached. Polymers which randomly attach are usually homopolymers or random copolymers without “adsorbing” end groups. These adsorb to solid surfaces with certain segments in direct contact with the surface and others not in direct contact. However, even though a particular segment may not be directly attached, its covalently bonded neighbours may be, so that it may have more in common with directly attached segments that the other unattached segments. An example of this is shown in Fig. 1 where portions of the chain are labelled trains, loops, and tails in accordance with current convention. Trains are polymer segments either directly attached or close to segments directly attached. The local motions of these segments are expected to be relatively restricted. Loops and tails are segments further away from the surface which end in either two or one train, respectively. It is generally accepted that these segments are more mobile than those in trains. However, they are not necessarily less mobile than those in the corresponding bulk or solution systems (see below). Terminally attached polymers may be attached to the surface through a variety of mechanisms. Bonding with single (or at least few) points of attachment is possible through covalent bonding, ionic, zwitterionic, or polar groups normally attached at the end of the polymer chain. The use of anionic synthesis where the end groups of the polymer are controlled is convenient for making these polymers. Larger attachment points are possible through the use of block polymers. In most systems, the thermodynamics of mixing is such that one component selectively adsorbs at the surface. For the adsorbed blocks, the distributions of segments into trains, loops, and tails depend on the specifics of the system. The attached block typically ends in an unattached block in the form of a tail. The conformation of the tail is a function of its length, interaction with the surroundings, and distance between neighbouring terminally attached chains. Following the accepted terminology,22 these segments are shown for terminally attached polymers in Fig. 1 as “mushrooms” and “brushes”. Generally mushrooms are expected for dilute systems while brushes are expected when the chains are tightly packed. Finally, in order to get an appropriate picture of polymers adsorbed to interfaces, it is necessary to compare their behaviour to those in the parent systems. To this end, detailed knowledge from NMR is available and 14916-23

282

F.D.BLUM

-- -

Brushes

I-

/Tail

Mushrooms

Fig. 1. Idealized representation of polymers attached to a solid surface: (A) randomly attached showing trains, loops and tails; and (B) terminally attached chains

showing the two extremes of brushes and mushrooms.

summarized in several reviews. Cheng4 and Axelson and Russellz5 have reviewed solids and liquids characterization of polymers. Yu and G u o ~and ~ Volke12’ have reviewed I3C solids studies of polymers. Roy and Inglefield” have reviewed studies of local motion in solids and J e l i n ~ k has i ~ ~reported ’ Spiess3* the use of deuterium in polymer chain studies. Blumich et ~ 1 . ~and have reviewed solid-state applications with special emphasis on multidimensional studies. Heatleg2 has reviewed relaxation studies of polymers in solution. The relationship between NMR parameters and mobility has been summarized by McCalP3 in terms of relaxation maps from various forms of spectroscopy. Of significance to the present work is the observation that the NMR Tg is typically on the order of 30-50°C higher than that observed by differential scanning calorimetry (DSCbreflecting the higher frequency of the NMR experiment. Several book^'^,^^,^' and collected also have good reference to polymer studies. These provide good starting points for understanding bulk properties so that the differences observed in adsorbed polymers can be understood.

NMR STUDIES OF ORGANIC THIN FILMS

283

3.1. Solid-liquid systems

Polymers adsorbed on solid surfaces in contact with a liquid are important for structural coatings, dispersion stabilizers and adhesives. The conformation and dynamics of the macromolecule at the interface are affected by the polymer's interaction with the surface and the solvent(s) present. A variety of magnetic resonance techniques have been used to characterize these systems. If the polymer segments are extended into the solvent region or there is significant solvent penetration into the polymer layer, NMR solution techniques may be appropriate to characterize the material. If the polymer is tightly bound to the surface with little solvent penetration, this part of the system will be solid-like. There are a variety of ways to characterize an adsorbed polymer.20 Generally, knowledge of the amount adsorbed, layer thickness, and density profile are useful, but not always easily determined. Spectroscopic methods for determining the bound fraction (trains) are important, but different techniques give somewhat different estimates. Infra-red studies rely on changes in vibrational frequencies due to the interaction of a polymer moiety (e.g. carbonyl) with the surface. Deconvolution of the IR resonances give an accurate estimate of the fraction of bound groups, but this technique counts only those directly bound. In contrast, NMR, ESR and microcalorimetry, for example, are sensitive to those groups directly bound, plus those groups close enough to the directly bound groups to be similarly reduced in mobility. Thus IR and the other techniques provide complementary information. Miyamoto and Cantow3' were early users of NMR to study polymer adsorption for poly(methy1 methacrylate) (PMMA) on silica adsorbed from deuterochloroform. They found that at low coverages, no high-resolution proton signals from the polymer were observed, but they did become apparent at higher coverages. With increasing polymer coverages, the surface polymer resonances became narrower. This is consistent with the polymer at low coverages being strongly attached and solid-like. Thus they were too broad to appear in a liquid high-resolution spectrum. Higher coverages produced more mobile material. They also found that the isotactic polymer is more motionally restricted than the atactic one at the interface. At low coverages, where the surface-PMMA signal in deuterochloroform was not observable, the addition of deuteromethanol had the effect of producing relatively narrow polymer resonances. This was consistent with a more expanded and mobile polymer layer due to the interaction with deuteromethanol. A number of ESR studies of surface attached polymers have been reported. While not the main focus of this report, these are of relevance because they employ very similar techniques and analysis. Central to the use of ESR is the incorporation of a stable unpaired electron, normally in the

284

F. D. BLUM

form of a nitroxide spin label. Use of spin labels in polymer systems has been re~iewed.~’ The main advantages are that the sensitivity of the electron is significantly greater than that of all NMR nuclei and that a similar apparatus can be used for many systems regardless of the mobility of the labelled polymer. The disadvantage is that the spin label is not native to the chain and must be added in some fashion. Consequently, the potential perturbation of the polymer motion by the label, and the relationship between the motion reported by the label and that of the polymer must be considered. A comparison of ESR and NMR studies of the mobility of poly(viny1 acetate) (PVAc) has been made in the presence of several diluenkm The correlation time estimates from both experiments were not equal, but showed similar trends. Consequently, it was concluded that these experiments were sensitive to the same types of motion. Robb and c o - w o r k e r ~have ~ ~ ~used ~ ESR to study both the interfacial dynamics of the polymer and as well as the absorbed fraction. For the adsorption of poly(viny1 pyrrolidone) (PVP) on carbon black (and also silica) in chloroform, the fraction of segments in trains decreased with increased surface coverage.43 PMMA was found to be relatively flat under similar conditions. A comparison of calculations of train fractions from ESR and IR was made. Caution was given regarding the use of ESR spin labels which themselves may bind to the substrate. The segments of PVP in trains on silica also decreased with increasing thermodynamic quality of the solvent.42Miller and coworker^^,^^ have used ESR to probe how the type of surface, amount absorbed, and solvent used affect the dynamics of the adsorbed polymer. A deconvolution of the line shapes allow the separation of mobile (tails and loops) and somewhat immobile (trains) components. PVAc on alumina or titanium dioxide had little conformational freedom compared to that on silica regardless of the solvent quality.44b For polystyrene (PS) on mica in the presence of cyclohexane, weak adsorption, even desorption, was found, in contrast to PMMA where a highly immobile polymer at the interface exists.45 The adsorption and dynamics of poly(ethy1ene oxide) (PEO) on silica have been studied using both ESR and NMR by Legrand, Hommel et PEO chains were end grafted on to silica by direct esterification of surface silanol groups. Analysis of the spectra based on division into mobile and immobile components allowed the estimation of the segments that were adsorbed or free. In benzene, it was found4’ that oligomers behave like brushes and polymers like partially collapsed coils. An increase in the thermodynamic solvent quality results in greater swelling of the grafted PE0.49 In mixed solvent systems, the swelling increases with the fraction of good solvent.50 At high grafting levels, the hyperfine splitting was used to identify solvent partitioning in the adsorbed laye?l and changes were found with grafting High-resolution 13C and ‘H NMR relaxation time measurements of

NMR STUDIES O F ORGANIC THIN FILMS

285

V Flg. 2. {‘H} 13C inversion recovery spectra of 2000g/m0l PEO grafted on silica in benzene. (Reprinted with permission from reference 54, 0 1984, American

Chemical Society.) grafted PEO in benzene also showed two-component behaviour in relaxation studies.54 An example of 13C relaxation behaviour is shown in Fig. 2, Both components show anisotropic motion with that of the less mobile component much more restricted in the direction perpendicular to the surface. The ‘H relaxation times were also believed to be sensitive to the local concentration of segments. NMR, ESR and DSC were all used to probe the effect of the surface on a small molecule nematic liquid crystal.55 It was found that the grafted PEO layer caused the formation of a disordered region near the surface. In order to more fully exploit and characterize the mobility differences between attached and free polymer segments, Cosgrove and Barnett56 developed the “driven equilibrium” method. This technique employs both a solid and liquid echo train to distinguish the proton spin density of the two species. Two examples are shown in Fig. 3. Applications of this technique to PVP on silica in the presence of D 2 0 showed a bound fraction which decreased with increasing surface ~overage.~’ The results at low coverages were consistent with a relatively flat conformation of PVP on the surface. The bound fractions determined were also in good agreement with earlier experiments based on ESR studies. It was again noted that N M R and ESR

286

F. D. BLUM

yielded similar results for the bound fraction while IR measured a fraction which was significantly lower. The results were consistent with magnetic resonance experiments being sensitive to mobility changes while IR is sensitive only to the carbonyls directly bonded to the silica. For comparison, terminally bonded PS on carbon black with CC14 showed a much lower fraction of bound segment^.^' The driven equilibrium method has been applied to a number of systems and used in conjunction with neutron and light scattering to provide a fuller picture of the surface-adsorbed polymer .5M3 For random ethylenelvinyl acetate copolymer^,^^ the data suggest preferential adsorption of vinyl acetate groups on both methyl and cyclohexyl surface modified silica. There was reasonable agreement between the results from neutron scattering and NMR for the bound fraction. Further studies have identified preferential adsorption in polydisperse system^.'^ Polystyrene latex particles have also been used as the surface. Adsorbed poly(viny1 alcohol) (PVA) in water had a small, but measurable value of the bound fraction while no bound material was found for terminally attached PE0.61762This was consistent with PEO having a highly extended conformation. An alternative approach, combining the use of neutron scattering and NMR, has been proposed by Cosgrove and Ryan.-6 It was found that the solution T2 of PEO in water varied linearly with PEO concentration. The segment density profile from the neutron scattering experiments was then used to reproduce the T2 profile for the terminally attached PEO. The comparison was based on the assumption that the T2 values would be the same for the solution and surface-bound polymer of the same local segment concentration. The good agreement between the predicted and experimentally measured T2 profile suggests the appropriateness of the assumption. The relaxation time approach was extended to poly(styrene sulphonate) (PSS) adsorbed on silica.67Since this is a non-absorbing system, the results were found to be useful in determining the thickness of the depletion layer. The use of a solvent (water) as a probe of polymer adsorption and desorption at interfaces for PEO and PVP on silicam was also examined. The proton relaxation rate in water was enhanced (shorter T I ) when polymer was adsorbed on a silica surface. The relative independence of the relaxation rate with polymer molecular weight was given as evidence that the solvent is mainly sensitive to the amount of polymer segments in trains. In addition, this method was also capable of measuring a critical displacer concentration in good agreement with other techniques. In contrast,69 the same technique applied to water in the poly(styrene sulphonate) (PSS) on a polystyrene latex yielded somewhat different results. The formation of a relatively flat PSS layer at the interface excludes solvent and the solvent is sensitive to only the segments in chains and loops. The calculation of bound fraction by this method was in agreement with ESR studies of a similar system.'O

287

NMR STUDIES OF ORGANIC THIN FILMS 1 .oo

0.75

z 0 0.50 (3

a

f

0.25

0.00

0.

51.2

102.4

153.6

204.8

1536

48

TIME (/MICROSECONDS)

1 .oo

0.75

g

z

E2

go,"-9Wy-

90~y-180"y-180",

0.50

0.25

0.00

I

512

1024 TIME(/MICROSECONDS)

Fig. 3. Solid and liquid echo spectra for (a) PVP on Aerosil in D20 and (b) terminally attached polystyrene on Philblack "0". The intensities for the B regions are representative of the liquid-like resonances while A is representative of both liquid- and solid-like spins. (Reprinted with permission from reference 57, 0 1981, American Chemical Society.)

288

F.D. BLUM

In some of the previous studies, measured relaxation times were related to the local segmental concentration of polymers adsorbed on to a silica who studied the relaxation surface. This was exploited by Blum et al.*4771 behaviour of deuterated styrene segments in adsorbed copolymers of 2-vinylpyridene (2-VP) and styrene (S) . Two specifically deuterated polymers (WDSS and VPSDS) were prepared with the deuteron on different halves of the styrene part. This is shown below: (CH&H),-(

CH=CD),-(

CH&H)z

(CH2-CH),-(

CHz-CH),-(

CHAD)*

@ b @6 6 6 VPDSS

VPSDS

An example of the high-resolution 13C spectra of the block copolymer in toluene and adsorbed on silica, swollen with toluene, is shown in Fig. 4. In solution, the resonances from all of the carbons are observed, while the spectrum from the surface-bound material shows only high-resolution resonances for the styrene. This is consistent with the 2-VP being rigidly adsorbed on the surface, while the styrene segments are mobile. Comparison of the deuterium relaxation times for the surface-bound and solution polymers allowed an estimate of the local concentration and thus the extension of the styrene segments to be made. It was found that the styrene segments were extended to four times their normal radius of gyration. This was in agreement with surface force studies. A comparison of the relaxation times for these block copolymers is given in Table 1. A surprising result was that when the relaxation times were compared at similar TI values, the styrene surface segments had a Tl/T2 ratio closer to unity than in solution. These data are consistent with the notion that the styrene surface segments are locally more mobile than those in solution. This observation requires that some rethinking of how the Table 1. Comparison of deuterium relaxation times for W D S S and VPSDS polymers in the presence of toluene.71

VPDSS Surface Solution VPSDS Surface Solution

3.63 3.63

3.38 2.75

1.07 1.32

3.94 3.94

3.60 3.25

1.09 1.21

NMR STUDIES OF ORGANIC THIN FILMS

289

Poly(styrene-deuterostyrene-2-vinylpyridine)

on silica in toluene-d8at 80°C

Poly(styren&co-2-vinylpyridine) in toluene-da at 80°C

Tand b

h

I

h

I

f T = Toluene

TMS

20

0

Fig. 4. { 'H} 13C spectra of poly(Zviny1 pyridine-co-styrene) in toluene solution (top) and on silica swollen with toluene (bottom). The absence of liquid-like signals for 2-vinyl pyridine on the silica surface implied that it was firmly held, while the styrene segments are quite mobile. (Reprinted with permission from reference 71, 0 1990, American Chemical Society.)

surface packing of chains affects the local segmental motions be done. This phenomenon was only observed in toluene, a good solvent for polystyrene. In poorer solvents, more reduced mobility was in the poly(viny1 Relaxation times were also measured by Parker et pyrro1idone)-co-(vinyl acetate) (PVP-co-VAc) alumina system. Sedimentation studies showed that dispersion stability was increased with the amount of VAc present. Dispersion stability was also enhanced in the presence o f

290

F. D. BLUM

iso-propanol (good solvent for VP, poor for VAc) compared to toluene (good solvent for VAc, poor for VP). On alumina, the carbonyl T , values for VP were longer in the presence of iso-propanol than toluene. For the VAc carbonyl, the solvent quality did not affect the TI much. This was consistent with the notion that the solvent does not displace VAc which is mainly associated with the surface. However, the solvent quality does seem to affect the layer (mostly VP) thickness in agreement with the dispersion stability. 3.2. Elastomer-solid systems

Most elastomers used for applications from balloons to tyres are made from material which is filled with a solid species. Typical fillers include carbon black or silica. The interaction of elastomers with solid particles is reasonably well suited for NMR studies. Elastomers at room temperature show behaviour which is in between solids and liquids. NMR studies on the bulk polymer systems have identified the motions responsible for the different behaviour observed. Much of the earlier proton work prior to 1980 was reviewed by Douglass and McBrierty7' for a number of different systems with interfacial effects. A more recent introduction to some of the ~ of elastomers have previous work is given by Kenny et ~ 1and. N M~ R studies been reviewed by K i n ~ e y The . ~ ~ latter contains a small section on filled systems. Schaefer first identified the anisotropic motions of cis-1,bpolyisoprene (&-PIP) above its glass transition temperature for filled and unfilled system^.^' He found that the presence of a carbon black filler produced an effect on the line widths, but not the Tls. Subsequent studies79 employing low speed magic angle spinning permitted the resolution of the methylene resonance and identified the inhomogeneous nature of the broadening. Dybowski and Vaughan" used multiple pulse proton NMR to remove dipolar broadening on similar cis-PIP samples. They were able to properly separate the relaxation and chemical shift effects, and the latter was determined to be the major contribution to the residual line widths in the filled rubber. Spin-spin relaxation rates were found to be an order of magnitude faster than in the corresponding unfilled rubber. It probably can be concluded that molecular motion is homogeneous on the TI time scale, but inhomogeneous on the T2 time scale.76 More recently, a number of studies have more fully characterized bulk elastomers. English and Dybowskis1,82have used a combination of techniques to show that the room temperature line width (a few hundred Hertz) of cis-polybutadiene (cis-PB) was caused by rapid anisotropic motion, restricted by chain constraints. Cohen-Addad et ul.83,84have shown how some of these restrictions may relate to chain entanglements or cross-links.

NMR STUDIES OF ORGANIC THIN FILMS

291

pol y (cis-l,4- is0 pr e ne)

140

130

120

410

I

I

30

20

Fig. 5. Heteronuclear 2D J spectrum of carbon black-filled &PIP with magic angle spinning and high power proton decoupling during detection. (Reprinted with permission from reference 85, 01987, American Chemical Society.)

Interestingly, there is enough motion in filled rubber systems that a combination of NMR techniques can ultimately provide quite narrow resonances for filled elastomers. The remaining widths are small enough that ‘H-13C J-couplings were resolved by Kentgens et aLS5for carbon black-filled natural rubber (cis-PIP). An example of the heteronuclear 2D J spectrum is shown in Fig. 5 where the coupling is resolved. The carbon spectra under a variety of different conditions are shown in Fig. 6. It is observed that the line widths can be reduced by an order of magnitude and the J-coupling can be resolved with the use of multiple pulse techniques. Various studies have attempted to characterize the physical inhomogeneity in filled e l a s t ~ r n e r s . ~Kaufman ~ ~ ~ ” ~ ~et found that filled ethylenepropylene-diene rubber (EDPM) and cis-PB showed a two-component T2 behaviour. The T2s varied by over an order of magnitude. The more immobile region was assigned to the rubber bound to the carbon black, The relatively free rubber was more motionally constrained than that in the bulk

292

F. D.BLUM

NATURAL RUBBER I

A

e

C

I

D

E

Fig. 6. Carbon-13 spectra of filled cis-PIP under a variety of conditions: (A) no decoupling, no MAS; (B) high power proton decoupling, no MAS; (C) MAS, no decoupling; (D) MAS, multiple pulse decoupling; (E) MAS, high power proton decoupling. (Reprinted in part with permission from reference 85, 0 1987, American Chemical Society.)

NMR STUDIES OF ORGANIC THIN FILMS

293

n

L

Fig. 7. Diagram of regions different mobilities in filled elastomers. The boundaries of tightly bound chains are given by the dashed lines. A, Physical attachment; B, chemical attachment; C, cross-linked rubber molecule; D, loose fold; E, tight fold; F, multiple adsorptive attachments; G, interparticle linkage molecules; H, cilia molecules. (Reprinted with permission from reference 87, @ 1976, American Chemical Society.)

polymers. The authors also proposed that the motion of the bound polymers have measured T l , T,, was somewhat less than isotropic. O’Brien et and Tlp values and identified at least three different regions of polymer segments with differing mobilities. These are shown in Fig. 7 and assigned to tightly bound, more loosely bound, and bulk-like rubber. Material balances suggested domain thicknesses of 6.6 nm for bound material 1.4 nm of which was tightly bound ~ n a t e r i a l . ~ ~ . ~ ~ Refinements of the above model have been made76 and it was concluded that: (1) there is motional coupling between the tightly and loosely bound rubber molecules; (2) a range of adhesion energies are indicated; and (3) removal of bound polymer has Arrhenius behaviour. These studies also implied that the glass transition temperature of the bound polymer was significantly greater than the bulk polymer. Similar results on motional differences have been found by Ito et dgga and Legrand et for silica-filled PIB and PIP and styrene-butadiene rubber (SBR). The effects of

294

F.D.BLUM

heat treatment on the materials showed that the fraction of bound rubber decreased irreversibly with heating.88a This was believed to be due to the degradation of the loosely bound rubber. When the silica was treated with organosilanes,’*” the amount of bound rubber decreased with surface coverage by the silane. These effects were not as pronounced with carbon black. Simon et ~ 1 . ’have ~ interpreted differences in relaxation time in terms of phantom network properties. Brown er aLW have used both CW and I T NMR to measure cross-link densities in filled and unfilled rubbers. Several studies have been made on filled poly(dimethy1 siloxane) (PDMS).84,91,92Cohen-Addad and Viallats4 have used relaxation experiments on solvent-swollen systems to characterize the network structure of the system. The silica concentration also was observed to control the superposition of relaxation data with a time scale shift factor.” Litvinov and have used deuterium-labelled PDMS to observe the behaviour of various fillers on the interface. Deuterium powder patterns observed above the glass transition temperature were associated with the bound elastomer. Imaging techniques may be sensitive to the interphase region in filled composites. K ~ h and n ~K ~~ e n i ghave ~ ~ reviewed ,~~ some of the applications in materials science. Chang and K o m ~ r o s k ihave ~ ~ demonstrated that contrast for elastomers can be obtained so that a 0.07 mm sheet of paper can be detected in a sandwich of carbon black-filled cis-polybutadiene. In fact, upon removal of the paper, the feature is still observed where the two polymer layers are put together. It is possible that some of the contrast observed was due to the differences in T2 from the differences in mobility of the surface and bulk material. It does not appear that a detailed study of this effect has been made. Along similar lines, Sakar and Komoroski9’ have imaged tyre cords and found differences between different rubbery regions. Differences in the T2s of carbon blacurubber differentiated good and poor dispersions. The reduced T2 value in a poorly dispersed sample may have been caused by the presence of “bound” rubber, though this was not observed directly. T2 weighted images made it possible to differentiate between “certain morphological and defect structures”. In elastomers filled with silica reinforcements, Garrido et aL9’ demonstrated the presence of an outer layer that appeared to have lower mobility, presumably due to both a lower concentration and mobility of the polymer in this layer. Subsequent density measurements confirmed a high density for this layer, consistent with a higher concentration of the more dense silica. Even more convincing evidence of an interfacial layer with different mobility was identified by Bliimler and Blumichw who studied the ageing and phase separation in elastomers. Upon ageing for 2 hours, a ring on the outside of the polymer cylinder was formed with a thickness of about 70 pm. With increased time, the width of the ring increased and was assigned to the aged fraction of the material. The contrast was proposed to be due to the decreased mobility of the polymer in the aged region. Interestingly, the

NMR STUDIES OF ORGANIC THIN FILMS

295

sample did not show any physical signs of ageing until after 8 hours, while the image showed noticeable effects after just 2 hours. Additional contrast for the interface was observed because the mobility of the polymer in the region next to the aged material had higher mobility than the unaged material. This material was identified with chains which had been broken or possibly due to larger local concentrations of plasticizer excluded from the aged region. 3.3. Solid-gas interface The coating of solid materials with polymers can be used to improve the surface properties of materials. In addition, these surfaces may serve as model systems for composites (with the matrix removed). Typically, these might be used in applications where the interfacial polymer layer would be in contact with a gaseous layer, usually air. These studies are often difficult to perform because polymers at an interface are typically rigid and consequently have broad resonances, although coherent averaging techniques can be used. They are also much more dilute than bulk systems so that sensitivity is of primary importance. An example of this for PVAc on silica is given in reference 14. There have not been a large number of NMR studies of polymer films at the solid-air interface. ESR studies of adsorbed PS in the dry state at the air interface showed decreased mobility compared to that of the bulk materiaL4 The adsorbed polymer can be considered to have a higher glass transition temperature than that of the bulk polymer. ESR studies have the advantage of increased sensitivity over NMR. The study of poly (iso-propyl acrylate) (PIPA) on silica has been performed by Blum et al. with 13CNMR.'4,100-'mPIPA is a flexible polymer that is below its Tgat room temperature. With CP MAS-NMR of PIPA in bulk, it was possible that under the proper conditions no high-resolution resonances for the polymer backbone were observed. This is shown in Fig. 8. When adsorbed on the surface, the CP MAS spectra showed resolved resonances for all of the carbons.'" This behaviour was similar to that of a glassy polymer where the rigid nature of the polymer backbone allows MAS and high power proton decoupling to sufficiently narrow the resonances. As a function of increasing surface coverage, the resonances were observed to broaden. This was consistent with a view that there was increased mobility of the polymer with an increase in distance from the surface. A monolayer is roughly the amount of coverage that distinguishes surface-modified from bulk-like in terms of dynamics as observed from NMR. The use of a deuterium label allowed *H NMR to be used for the surface-adsorbed labelled polymer, although the sensitivity of these experiments was not good. '03

296

F. D. BLUM

For PDMS chains adsorbed on modified silica gel, Van AlstenlW has shown that the Tls of the polymer increase with surface coverage. At a surface coverage of about 1.1g PDMS/g silica, the T1 is comparable to that in the bulk polymer. Chu and Murphy"' have used I5N-labelled polyimides (PI) to study PI adhesion to alumina. For thin layers of 30-150A (2-10 monolayers), solid-state 15N chemical shifts provided evidence for amine production and PI backbone cleavage at 150-255°C. Adhesion appears to be correlated with a higher temperature reaction which implies imidelalumina bonding. Treatment of the alumina with a silane coupling agent "passivates" the alumina towards this reaction. Conformational changes of polypeptides on solid surfaces have been ' ~ chemical shifts were consistent studied by Fernandez et ~ 1 . Carbon-13 with more extended structures (vs &-helices) on silica and hydroxyapatite than in the bulk material for poly-L-lysine and poly-L-glutamic acid. Deuterium NMR experiments were used to characterize the reduction of the mobility of the side chains on the surface. Spin-locking experiments also identified reduced mobilities of both the backbone and side chain upon adsorption. Solid-state NMR studies have been used with some difficulty in the characterization of plasma polymer^.^^^*'^^ These are often obtained as tenacious, thin films on low surface area substrates (e.g. glass slides). Characterization with electron spectroscopy for chemical analysis (ESCA) has had good success with fluorine, but for carbon separation into sp2 and sp3 carbons is only achieved with deconvolution techniques. Carbon-13 studies by Kaplan and Dilks on hydrocarbonslWl1' and fluorocarbons' l2 have shown the superiority of NMR in terms of resolution. Since the deposition rates were slow, polymer had to be collected using several runs between which the polymer was removed. Identification of specific chemical species such as methyls, methylenes, methines, and quaternary carbons was possible. Dipolar dephasing allowed non-protonated and weakly coupled (methyl) species to be separated from strongly coupled (rigid, protonated) species. The use of 13C-labelledprecursors also allowed the pathway to be ~ similar techniques to follow sp2/sp3ratios followed. Gambogi et ~ 1 . " used from ethane and methane plasmas. Dipolar dephasing allowed the identification of non-protonated aromatic or aliphatic carbons which could be precursors to diamond- or graphite-like regions. Both I3C and I5N NMR were used on plasma polymers made from acrylonitrile.' l4 Acrylonitrile/nitrogen plasmas produced a variety of structures with C-N bonds. The type of material formed was dependent on the Fig. 8. Spectra of PIPA (from bottom to top) in bulk, at 1/4 monolayer, at 1/2 monolayer, and at monolayer coverage on silica. (Reprinted from reference 101 with the permission of Plenum Publishing.)

NMR STUDIES OF ORGANIC THIN FILMS d

-kCH,-

I

c

CH%

I

e

O = C-O-CH-(CH,), a a

b

e b

c d

I Bulk

297

298

F. D. BLUM

location in the cell. A variety of amide species, including heterocyclic rings, were found with "N NMR. A different approach has been used by Lock et al. '15 who used dynamic nuclear polarization (DNP) to transfer magnetization from unpaired electrons to 13C resonances. This technique offers the potential for performing experiments with substantially lower amounts of material. Unfortunately the authors reported the use of an 83 mg sample for 13C analysis which is about an order of magnitude more than a typical single plasma run produces. For the "diamond" films produced, the DNP derived spectra showed only one resonance at 36 ppm corresponding to "diamondlike" material. For this sample, a useful 13C spectrum could not be produced without DNP. The DNP technique also offers the advantage of avoiding the long relaxation times of the diamond-like carbons. It may ultimately be possible to study fairly thin organic films using NMR imaging. Applications of NMR imaging to materials science have been Most commonly (as in medical imaging), the NMR signals from mobile components protons are responsible for the more intense signals while non-mobile species with short T2sprovide little signal. Contrast is then due to differences in mobility. In many studies, this mobility difference is due to the presence or absence of relatively small molecules. These can be observed using liquid-like imaging techniques. Early imaging studies have been applied to composite systems such as glass/epoxy by Rothwell et ~ 1 . ' ~ where ' water was used to provide contrast. This approach was extended to glass/polyester and glass/nylon systems' l8 where the uptake of water was also of prime interest. These studies demonstrated that the imaging technique could not only be sensitive to the amount of water uptake, but also to voids and imperfections in the composite. Extensions of these studies to polymer/solvent systems demonstrated that solvent fronts and inhomogeneities could be observed and related to both the amount of solvent present, but also its The sharpness of the images of the front was sufficient to demonstrate case I1 diffusion behavi~ur."~Contrast between various regions in these materials was obtained via Tl, T2,or diffusion coefficient weighted images.12' While the mobility of the polymer was not directly imaged, the properties of the solvent are likely to be directly related to those of the polymer.12* Thin slices of poly(viny1 alcohol) and poly(viny1 acetate) were also imagable as glue lines between wood and ceramic tiles, though no attempt at identifying an interfacial layer was made.'23 Cracks and voids which allow water penetration are also observable in ceramic tiles joined with styrene-acrylate adhesives.124 The previously mentioned imaging studies were all performed on materials which had liquid-like resonances where more or less standard (medicallike) imaging techniques could be used. For typical composite materials, this is not possible because the samples can be very solid-like. There have been some attempts at using combinations of multipulse and cross-polarization to

NMR STUDIES OF ORGANIC THIN FILMS

299

the imaging of true These appear to be promising, but to date no detailed studies of relevance to thin organic films have been published. 3.4. Solid-solid systems

The solid-solid interface is of importance to composite materials in applications where a rigid matrix is required. In some cases, the presence of solid surfaces changes the behaviour throughout much of the system (see Section 3.2 and below). In others, the surface only appears to affect a thin layer. In that case, some technique which contrasts the interfacial material from that of the matrix is required. Isotope enrichment of surface-active species such as coupling agent^','^','^^ and magnetization t r a n ~ f e r ~ ~are ”~~’ both effective in providing this contrast. The former will be discussed in Section 4. The use of 13C NMR to probe composite materials containing solid polymers was realized about 10 years ago by Resing et ~ 1 . ’ ~ ~ Unfortunately, none of the applications reported directly addressed the behaviour of an interfacial polymer layer. For composites made from poly(viny1 alcohol) and silica, Zumbulyadis and O’Reilly1293’30 found that the cross-polarization time constant, TsiHfor ‘H to 29Si was shorter for the composite than the silica alone. This was interpreted as being due to the PVA making the surface hydroxyls more rigid and hence more effective at cross-polarizing silica. Deuteration of the exchangeable protons (surface and alcohol hydroxyls) resulted in a slower, but still measurable, T S i ~This . showed that even the backbone PVA protons could effectively cross-polarize the surface silicons. A rough distance estimate was also made from these data. The authors concluded that this was a “tightly coupled spin system”. Composites made from nylon 6 filled with glass were examined by Weedling et al.132,133using 13C chemical shift data. A comparison of bulk and glass-filled (including surface-treated) composites showed that a greater amount of y-crystalline form was produced when the composite material was processed as compared with the original bulk polymer. In the absence of glass spheres, the y-form reverted back to the a - f ~ r m . ’In~ ~follow-up studies, the effect of various processing and compositional variations was probed. 133 An amorphous polymer layer was found to be intimately associated with the glass. Parker et ~ 1 .have l ~ probed ~ the effect of the surface treatment of alumina on the properties of green tapes made with a polystyrene binder. Through measurements of the 13C cross-polarization times and ‘H Tlps it was found that the mobility of the polystyrene varied as a function of surface treatment. The presence of n-octylsilane in the system increased the mobility of the plasticized tapes. The NMR measurements correlated with the expected densities and moduli of the tapes.

300

F.D.BLUM

Proton Tlp studies of polyethylene (PE) have been studied by N a t a n s ~ h n ’ ~in~composites ,~~~ with calcium carbonate and cellulose. For the CaC03 composites, a “good” and “bad” sample had different relaxation behaviours. The differences are believed to be due to physical differences in the samples as the CaC03 was magnetically transparent. Yu and G u o have ~ ~ also studied this system and found from ‘H TIPvalues that the PE was more rigid due to the presence of the filler. To close this section, it is appropriate to note that the use of imaging to characterize surface species in composites would be an important accomplishment and is a major goal. New techniques based on multipulse NMR experiment^'^'.'^^ have been applied to solid polymers. However, their resolution is not yet good enough to yield significant results on filled polymers. Some work has been done on composites used in propellent^,'^^ green ceramic^,'^' and polymer ~andwiches,’~~ but these have been limited to providing information on rather large structures. 4. SURFACE-ACTIVE AGENTS Modifications of solid materials with small molecule surface-active agents are used to improve adhesion, dispersibility, wetting, and flocculation of solids. These materials may be added to the system such that they react at the surface or are just adsorbed. In the former case, the materials may form polymers or similar structures at the interface; however, these will be discussed here because they are typically applied as monomeric species. For the purpose of this review, discussion of chemically bonded species are separated from non-bonded ones. 4.1. Bonded phases

The use of bonded surface species, particularly as used in composites and chromatography, is well documented. In composites, these materials can form a bridge between an organic matrix and an inorganic filler. In chromatography, they modify the surface properties so specific binding may occur. The most common of these are based on silanes and have been described by P l ~ e d d e m a n n ’and ~ ~ others. 143 A variety of physical-chemical characterization tools have been applied to these species and were reviewed by IshidaIu in 1984. Since then a number of reviews have focused on the use of NMR to characterize these interfacially active species.5-10,14,26,101,102,145 Silane coupling agents may be used in small amounts to treat glass and inorganic fibres and fillers. The major advantage of their use in polymer composites is the enhancement in physical properties such as strength when the composite is subjected to conditions which might degrade the interface.

NMR STUDIES OF ORGANIC THIN FILMS

301

A common example of this is their use in glass/epoxy composites where good physical properties are maintained in the presence of water. While many aspects about how these interfacial species work are unclear, NMR has achieved good success at characterizing them. In chromatography, the modification of surface properties is critical to the interactions responsible for chemical ~ e p a r a t i 0 n . IThe ~ ~ chemical structure and mobility of the agents at the surface are important for the effectiveness of the surface selectivity. In addition, the organofunctional end of the molecule provides a site for the attachment of other functional groups. This allows for the tailoring of the packing material for the specific separation required. Silane coupling agents are normally obtained as chloro- or alkyl esters (typically methyl or ethyl). For example, 3-aminopropyltrimethoxysilane(I) (APMS) is shown below. Under anhydrous conditions these compounds can be relatively unreactive at room temperature. Addition of water to the sample causes the hydrolysis to a reactive silanol (II). Condensation of the silanols to form oligomers (111), where X may be another silane, occurs in competition with the formation of a surface-Si-0-SiR bond (IV). These are, of course, simplified structures. In general each of the functional groups reacts at its own rate and, in general, they may react with different species. Several workers have used NMR to identify the reaction products and rate^,^^^,^^^,^^^ but this is beyond the scope of the present work. (CH30)3SiCH2CH2CH2NH2 I

(H0)$iCH2CH2CH2NH2

I1

I (X)3SiCH2CH2CH2NH2 IV

III

At low concentrations, most of the hydrolysed coupling agents may be found as individual monomers, but oligomers form at higher concentrations. This formation does not appear to affect the adhesion, h 0 w e ~ e r . lThe ~~ extent of interaction with the surface is often dependent on the functionality of the silane groups. For example, when a trifunctional silane is used, condensation of the silanol groups occurs in addition to the surface reactions. 4.1.1. 29Sistudies

The reaction of the silanes on the surface is conveniently followed by 29Si NMR. In spite of the broad chemical shift range of silicon, the chemical shifts of different species in silica differ by only a small

302

F.D.BLUM

Nevertheless, it is possible to identify silicon atoms with different substituents as: (Si-O),Si-(OH),

(Si-O),-Si-(OH)

Qz

(Si-O)4-Si Q4

Q3

The chemical shifts for these are -91, -100, and -109 ppm,1503151 respectively, for silica gel. The use of cross-polarization from the surface hydroxyls makes this a surface-sensitive technique. Quantification of the various resonances is difficult because the TsiHsare different with the Q4s having the longe~t'~'CP time (15-20 ms). The types of species formed in silica gel were found to be dependent on the thermal and hydration history of the ~amp1e.l~'Pfleiderer et al.i52 examined the "Si relaxation and crosspolarization behaviour of various silicas and found that various types of microdomains existed in the silica. Combined rotation and multiple pulse spectroscopy (CRAMPS) has also been used for surface proton characterization.'53,154 These studies have identified hydrogen bonded silanols, isolated silanols, and adsorbed water. Various heat treatments were used to eliminate different species. Tuel ef ~ 1 . 'have ~ ~ used 29Si to differentiate between pyrogenic and precipitated silicas. The surface studies have not been limited to silica. Higgins and Ellis156have characterized high surface area alumina. At room temperature, aluminium atoms in the first two layers were not observed in solid-state 27Al NMR spectra due to their large electric field gradients. The reaction of silica with surface-modifying alkylsilanes is conveniently followed with 29Si NMR. Loosely following the nomenclature of Albert and Bayer,' trifunctional, T, difunctional, D, and monofunctional, M, silanes can be distinguished. These are shown below where the subscripts denote how many of the silicons bonds are to 0-Si groups. (Si-0)-Si-R3 M1

(Si-O)(HO)TSi-R T1

(Si-0)( HO)-Si-R2

D1 (Si-O)2( H0)-Si-R

TZ

(Si-O)+i-R2

DZ (Si-O)3-Si-R T3

where the alkyl (R) groups need not be homogeneous and often contain functional moieties, and the silicons in parentheses may be either from the surface or from other silanes. Substitution of OR groups with OH groups is also possible. The alkyl substituent on the silicon conveniently shields the silicon resonances away from the surface hydroxyls. The 13C and 29Si spectra of four modified silica gels were studied by Maciel et These were typical of stationary phases used in chromatogra-

NMR STUDIES OF ORGANIC THIN FILMS

303

phy. The 29Si spectra showed the presence of Q3 and Q4 species plus resonances due to the organosilanes. For trimethylchlorosilane reacted on to silica, the alkylsilane has a 29Siresonance at 15 ppm, well separated from the silica gel silanes. Well-resolved resonances were also found for the 13C resonances. The values of chemical shifts found in various studies will not be reproduced here. Several literature studies have been summarized.' Sindorf and M a ~ i e l ' have ~ ~ examined the details of the quantification of the structural features of trimethylchlorosilane on silica gels. Shown in Fig. 9 is the CP MAS spectrum of silica gel and its reaction product with trimethylchlorosilane. Upon reaction it can be observed that the surface silanol sites Q3 (b in Fig. 9) are converted to Q4 (a in Fig. 9) sites. The species labelled c in the figure was assigned to geminal hydroxyls (Qz). The presence of the Q2 resonance was questioned for certain samples,159although later spectra by the same authors show its presence.8 In addition, the M1 (d in Fig. 9) resonance due to the reacted alkylsilane is clearly observed. Similar results were found for trimethylchlorosilane reactions with different silica gels. An alternative to CP MAS studies was suggested by Fyfe et al. 160 who proposed the use of direct polarization MAS. Relaxation times, TI, for the Q4 silica gel resonances were about 20 and 25 s for a silica gel and high surface area glass, respectively. Shorter relaxation times were found for silicons with hydroxyls. Silylation significantly increased the Tls of the support. Monodentate (M-type) silanes have one reaction site, but D and T have multiple ones. This adds to the complexity of the systems as the multifunctional silanes can react with each other in addition to silica, water and alcohols. Sindorf and Maciel161 have done a detailed examination of the behaviour of multifunctional silanes. Dimethyldichlorosilane, methyltrichlorosilane, dimethyldiethoxysilane, and methyltriethoxysilane were adsorbed on high surface area (750m2/g) silica gel. It was found that a variety of different structures were formed with time and air exposure. For example, for methyltrichlorosilane under anhydrous conditions a large Tl (single surface linkage) resonance was found. Upon exposure to air, loss of HCl and formation and condensation of silanols occurred. The reactions of the ethoxysilanes were similar to those of the chlorosilanes. A series of mono- and trifunctional silanes on a number of different silicas were studied by Bayer et for a variety of chain lengths and functional groups. Both 13C and 29Si were deemed to be useful characterization tools for laboratory and commercial modified silicas. They compiled a 13C and 29Sichemical shift listing for a variety of silanes with different functionalities including commercial reversed phase systems. A similar multinuclear approach was also followed by Sudholter et af.162for 3-chloropropyl- and 3-aminopropylsilane. Both nuclei provided evidence for different bonding of the reactive silanol groups. For CI8 di- and trifunctional silanes on silica gel, Pfleiderer et ~ 1 . have l ~ found ~ that under dry conditions, monodentate and

304

F . D . BLUM

Me I

Me-Si -Me

I””

a+a

Me-Si-Me

L b+b

d

C

B

-‘

1

50

‘

I

0

’

.

.

.

-50

,

i

‘

.

-

-100

PPM

Fig. 9. ”Si CP MAS spectra of (A) silica gel and (B) the product of its reaction with

trimethylchlorosilane. (Reprinted with permission from reference 158, American Chemical Society.)

0 1982,

NMR STUDIES OF ORGANIC THIN FILMS

305

bidentate surface species are formed, respectively. Higher-dentate ligands ' ~ are formed when a dry inert atmosphere is not used. Tuel et ~ 1 . focused their study of modified silicas to determine the amounts of reactive sites. Akapo and Simpson16' examined the relationships between 29Si and I3C characterization and column performance. Reactions similar to those mentioned above also seem to occur for various silanes on the surface of zeolites. 166 A few detailed studies focusing on the 29Si characterization of 3aminipropylsilane (APS) on silica have been performed. 167-170 The influence of heat treatment and reaction conditions were probed for APS and 3-methacryloxypropyltrimethoxysilane(MPS) by de Hann et al. lfi7For APS on both silica gel and Cab-0-Sil surfaces, reaction from dry toluene produces mostly monodentate structures, while heat treatment or the presence of water produces bidentate or tridentate linkages. For MPS, the spectra show the presence of bidentate and higher species, even under dry conditions. 29Si spectra taken from APS and 4-aminobutylsilane (ABS) on Cab-0-Sil show significantly poorer resolution than on silica ge11679'68 though they are sufficient to identify bonding. This is probably at least partially due to the lower surface area and perhaps even surface heterohave made a detailed 29Si and 13C geneity of Cab-0-Sil. Caravajal et difi9 study on A$S on silica gel. Although the situation varied with the exact conditions, it was generally observed that APS at room temperature was adsorbed mainly as mono- ( T I ) or bidentate (T2) species. With heating, higher amounts of the T3 species were formed. Hoh et ~ l . ' ~ have ' observed changes in the cross-polarization times for surface and bulk polymerized APS species. Shorter TsiH values for the surface species were believed to be indicative of lower mobility of the surface values were somewhat less species. In mixtures with epoxy resins, the T s i ~ conclusive. Gambogi and Blum171 observed that the 29Si resonance from ABS got sharper with the addition of water to the interface of ABS/Cab-0l ~ ~ that Sil and in a composite with bismaleimide. Vrancken et ~ 1 .determined humidity could cause the partial hydrolysis of the surface APS layer. Previous studies by the Eindhoven group have been extended to and trimethoxymethoxytrimethylsilane, dimethoxydimethylsilane, methyl~ilane.'~~ They found that thionyl chloride was a more effective dehydration agent than heating. Some of the methanol produced in the hydrolysis reaction was adsorbed. Results similar to those in their earlier s tu d ie ~ "were ~ also found. Several studies have also followed the condensation reactions and adsorption with 29Si NMR.148,'74-'76Detailed studies of chromatography supports are greatly aided by this technique. 177~178 4.1.2. 13C studies

Considering that the major differences between the reactive silanes are the organo groups, 13C studies have been crucial to the identification and

306

F.D.BLUM

characterization of the bonded surface species. The “business” end of the molecule is typically away from the silane. The functional group may be responsible for preferential binding, as in chromatography, or chemical bonding, as in adhesion promoters. In solventless situations, short chain alkyl groups are solid-like, while the longer alkyl chains become more liquid-like with increasing chain length. For the solid-like chains, CP MAS is an effective way to probe the interface. Reviews of this application on silanes e ~ i s t . ~ , ~ , ~ J ~ , ~ ~ ~ One of the earliest reported I3C studies of silanes is that of Chang et ~ 1 . ’ ~ ~ who reported the CP spectra of three different alkyl silanes on chrysotile asbestos. The importance of this work probably lies in the demonstration that natural abundance I3C studies were possible on the surface. In the early 1980s a variety of CP MAS 13C studies appeared. Additional work on the asbestos system’31 showed that the addition of MAS could significantly enhance the obtainable resolution. The identification of the carbons present in the reacted material was an important test to determine if various functional groups could survive the treatments. Over the next several years a number of different surface-bound silanes were characterized. Most of the systems studied were of relevance to chromatography. The aim was to identify the chemical shifts and any specific conformational or complexation effects present. The initial studies of Leyden et u1.180,181and Maciel et ~ 1 . ’ ~demonstrated ’ the usefulness of the technique and probably stimulated work on a number of systems. An example of the resolution obtainable is shown in Fig. 10. Hayes et ~ 1 . ” ~ reported the surface spectra of 12 different monofunctional silanes. The Cls material was noted to give broader resonances due to its mobility. A wide variety of other systems have been studied with 13C in conjunction with 29Si NMR.8,’2,161,162,165.167 In addition, Jinno has characterized a number of commercial silica gel reverse phase materials with I3C.ls3 The work of Chiang er ~ 1 . ’was ~ more oriented to the materials/ composites applications. They compared the chemical shift values for APS and N-2-aminoethyl-3-aminopropyltrimethoxysilane (AAPS) on Cab-0-Sil silica, as a bulk polymerized polymer, and as a neat liquid. The chemical shifts were interpreted in terms of structural changes believed to take place in the different systems. Zaper and Koenig18’ have extended these studies to mercapto-, amino-, epoxy-, methacrylate-, and vinyl-functional silanes. These were all either derived from the triethoxy- or thimethoxyester . Comparisons were made between Cab-0-Sil surface and bulk polymerized materials. Differences in line widths and chemical shifts were found. ’ ~ ~detailed studies on APS to determine how the Carajaval et ~ 1 . did reaction conditions affected the structures formed. A few studies have used 13C to probe the dynamics of surface-bound species. Sindorf and Maciells6 have measured TCH, T l p , and TI values for dimethyloctadecylchlorosilane (DMODCS, CI8) and dimethyloctylchloro-

NMR STUDIES OF ORGANIC THIN FILMS

Yo,

307

,o’

Z O - S , CH, CHzCHzNH CH,CH, f

e

c

c

NHCOCH, COCH,

b

c

c

o

d

PPM o 205 b 170

c

d e I

b

,

,

,

250

,

,

,

30.1 22.9 11.5

i

P

l

508

,

,

200

,

,

,

1

,

150

,

,

,

,

100

,

~

,

~

50

,

,

,

0

,

,

,

,

, , , I ,,

-so

,,

,

,

-100

PPPl

Fig. 10. I3CCP MAS spectrum of a trifunctional silane on silica gel. (Reprinted with permission from reference 181, 1982, American Chemical Society.)

silane on silica gel. The cross-polarization rates for carbons decrease away from the attached end, but level off after about eight carbons. The rate of molecular motion increases away from the surface. Later studies on DMODCS on silica gel’87 identified differences in intensities between CP and single pulse spectra. At low coverages, the rate of reorientation of the silane groups was slow, but this rate increased with the addition of also probed CI8 mobility on the surface “wetting” solvents. Albert et di8* in the presence of solvents. It was observed that both the loadings and solvents influence the mobility of the chains. For shorter chains of importance in composites, Nishiyama el ~ 1 . ~found ~ ’ that hydrolysis, adsorption, and condensation could be followed with solution state 13C, as well as 29Si NMR. Earlier st~dies’”’~’ had previously identified the restricted mobility of APS on silica gel in D 2 0 with line width and relaxation time measurements. In order to characterize the behaviour in solvent-containing, modified

308

F. D. BLUM

systems, Gilpin and Gangoda'46.192-197have performed experiments on a variety of 13C-labelled silanes. In their initial studies, the labelling allowed them to observe the resonances from the labelled species at the interface. For C9 and Clo labelled materials,'92 the labels were far enough away from the silane end that they were mobile enough to be almost "liquid-like". Labels positioned nearer to the silicon atom were quite broad and very difficult to characterize with liquid techniques. Nevertheless, the motional gradient could be identified. For methyl end-labelled alkylsilanes on silica,195the Tl relaxation times were more a function of loading than chain length. Later studies by the same group demonstrated that the Tl values of an end-labelled dodecyldimethylchlorosilane (C12, monodentate attachment) varied linearly with the inverse of the solvent viscosity.'% Two distinct types of solvent behaviour were found and ascribed to differences in solvent penetration into the C12chains. The apparent energies of activation for the Tl process were also measured for the dry (ca. 13kJ/mol) samples and compared to those soaked with ethylene glycol (15-19 kJ/mol), and carbon tetrachloride containing materials. 197 The solvent-wet materials also showed greater sensitivity to the position of the 13Clabel on the silane. These results suggested to the authors that the silanes were not orientationally ordered in the presence of solvents than in their absence. It is believed that the mobility of the chain is related to the behaviour as separation m e d i ~ m . ' ~ ~ , ' ~ ~ Van AlstenlW has also used 13C-methyl-labelled dodecyldimethylchlorosilane adsorbed onto silica gel in the presence of various solvents. The T2s of the label suggested heterogeneous broadening from different environments for the alkylsilanes. The Tl also decreased with increasing solvent viscosity. Poly(dimethy1 siloxane) (PDMS) in deuterocyclohexane was added to the labelled coupling agent samples. At low coverages of PDMS, the Tl of the bound label increased with increased coverages, indicating greater motional freedom. This is possibly due to the conformational changes of the alkylsilane in response to the presence of the polymer. At higher polymer concentrations, the Tl decreased, went through a minimum and then increased with further addition of polymer. This is consistent with the environment of the alkylsilane becoming more motionally restricted as more polymer was added. In contrast, the mobility of the adsorbed polymer (and its Tl) increased as the amount of adsorbed polymer increased. It appears that at low coverages, the polymer was in a relatively flat conformation on the surface. As the coverage increased, the number of segments in the more mobile loops and tails also increased. Changing the polymer molecular weight199had little effect on the silane mobility, but some on that of the polymer. I3C labelling has also been used to provide contrast between an interfacial APS layer on glass and a nylon 6 matrix.'27 The position next to the amine was labelled. Even though relatively large spheres (5 mm) were used, there

NMR STUDIES OF ORGANIC THIN FILMS

309

was still sufficient sensitivity to observe the coupling agent. CP MAS spectra of the coupling agent on the glass beads alone, in a composite with nylon-6, and after extraction of the polymer were taken. At the surface-air interface, the resonance due to the labelled species at 43 ppm was observed. At least two components of different mobility contributed to the signal intensity. Upon mixing with nylon 6 the C, carbon of the nylon overlapped with the coupling agent resonance, but it was still possible to deconvolute the contributions to the relaxation behaviour of each to some degree. After extraction of the free nylon in the composite, the resulting I3C spectrum showed the definite contribution of the coupling agent at 43 ppm plus nylon resonances due to material which had not been extracted. The fact that a considerable amount of polymer was not removed from the glass beads suggested a strong interaction between the polymer and the treated glass surface. Relaxation measurements on the coupling agent's labelled carbon suggested that the presence of the polymer reduced its mobility. In a somewhat different but related study, Tuel et ~ 1 have. studied ~ ~the behaviour of a-o diols chemically bonded to silica. Measurements of relaxation behaviour and intensities suggested to the authors that the conformation of the chains on the surface was relatively flat. In addition, there was an evedodd effect believed to be due to the conformation of the molecules on the surface. 4.1.3. ' H studies Deuterium labelling has been used in a number of cases to provide contrast and also a spectrum which can be interpreted in terms of motional dynamics. Kelusky and FyfeZo1have reacted perdeuteroalkoxydimethoxychlorosilanes with silica gel to probe mobility in the presence and absence of solvents. The coupling agents studied had the general formula: CH3

I

--Q-Si-O-R

V and included R = -CD3, -CD2CD3, -CD(CD3)2, -(CD2),CD3, and -(CD2)&D3. For the dry, treated surfaces, the deuterium line shapes were superpositions of sometimes unresolved resonances. The mobility of the alkyl CDs increased with distance away from the surface. Differences in mobility with differing solvents were also found. Because of the overlap of resonances, quantifying the dynamics was impossible.

310

F.D.BLUM

To overcome the problems associated with overlapping resonances, ~ e d . ~ ~The ~ ~ selectively deuterated silanes have been ~ structure of these was similar to V with an R in place of the OR. The labelled compounds are designated by C,(Dy), where x is the total length of the R group and y is the position of the deuterium label with carbon number 1 next to the silicon. Gangoda et al.202,203 have prepared C+, C10, and CI2 dimethylalkylsilanes and bonded them to silica gel. Another material without the capability of surface bonding was used for comparison. The bonded silanes had more restricted mobility with the labels further away from the surface being more mobile. Spin-lattice relaxation times were in accord with that expected from the line shapes. Deuterium-labelled CI2(D11) was used to probe the surface mobility of the bound silane in the presence of various diluents including simple and binary solvent systems, surfactant solutions and nematic liquid crystals. In methanol and hexane, the silane moved faster than in the dry state. In a watedmethanol mixture, the mobility was similar to that of the dry state. In a dispersion containing surfactant, the silane moved slower than in the dry state and a further reduction was found in the presence of the nematic liquid crystal. Boddenberg et aL205 have studied the behaviour of the deuterated trimethyl species and found that the methyl rotation was fast while the 0-Si motion was much slower. Several deuterated (218 species were studied by Zeigler and M a ~ i e l . ~ ’ ~ Specifically C18(D9,10) and C,,(Dl) were studied as a function of temperature, loading and solvents. Anti-gauche jumps and more complicated motions were required to simulate the experimental data. At “high” loadings the mobility of the silanes was decreased compared to that at low loadings. The expected motional gradient extending out from the surface was found. Cyclohexane and acetonitrile significantly increased the segmental mobility, while water increased it only marginally. Quantitative rates for reorientation were provided with the use of a computerized spectral simulation program. In the author’s laboratory studies have been made on deuterated coupling agents, specifically deuteroaminopropylsilane (VI, DAPS) and deuteroaminobutylsilane (VII, DABS) because of their use in composites. Their structure is shown below and their synthesis, hydrolysis, and exchange described: 147

VI

VII

where the X groups depend on the reactions which occurred. The and the adsorption isotherms for the APS on silica have been dete~mined’~’ coupling agent adsorbed and reacted on Cab-0-Sil silica surfaces at

31 1

NMR STUDIES OF ORGANIC THIN FILMS

Experimental

-260

-,bo

6

Frequency (kHz)

Simulated

1

100

1

r

200 -200

I

-100

i

0

1

100

I

200

Frequency (kHz)

Fig. 11. Deuterium NMR spectra and simulations of deuterium-labelled C3(D3) arninopropylsilane on silica. The material was deposited with solution concentrations of (A) 0.5%, (B) 1.0%, (C) 2.0%, (D) 3.5% and (E) 10%; 2% corresponds to approximately monolayer coverage. (Reprinted with permission from reference 168, @ 1991, American Chemical Society.)

submonolayer and multilayer levels. The deuterium quadrupole spectra for At aminopropylsilane at various coverages are shown in Fig. ll.lm submonolayer coverages, the mobility of the surface-bound coupling agent was greater than that at monolayer coverage. At higher coverages, the mobility of the material starts to resemble that of the bulk polymer made by

312

F.D.BLUM

Fig. 12. Deuterium NMR spectra of deuterium-labelled Cd(D3) aminopropylsilane on silica, (A) treated surface and (B) treated surface overpolymerized with bismaleimide. (Reprinted in part with permission from reference 128, 0 1992,

American Chemical Society.)

condensing the coupling agent alone. The spectra were consistent with a large amplitude, anisotropic, rotational motion plus conformational jumps. Interestingly, the behaviour of DABS (butyl group) is similar to that of DAPS (propyl group) on silica at low coverages. At higher coverages the behaviour of the coupling agent layer approaches the behaviour of the bulk condensed polymer. Thus, at higher coverages the DABS is more mobile than DAPS. This may be why ABS is preferred over APS for glass/epoxy applications. The deuterium label can provide the contrast necessary to probe the mobility of the bonded layer in the presence of an overlayer of polymer. Shown in Fig. 12 are spectra of DABES on silica at roughly monolayer coverage in air and after reaction with bismaleimide (BMI).i28*'45It was found that the overpolymerization with BMI inhibited the motion of the coupling agent at the interface and also changed the way it moved. The spectra were consistent with the freezing out of the rotational motion chain

NMR STUDIES OF ORGANIC THIN FILMS

313

in the composite-leaving only conformational jumps. The effect of water on the deuterium spectra of the labelled coupling agent at the air interface, and also in the composite, was consistent with the coupling agent moving three to four times faster than in the dry state.’71 Consequently, the water affects the interface in these composites. Upon drying, the interface returns to its original state. This is believed to be important in how coupling agents provide enhanced performance in wet composites. 4.1.4, I5N studies

Solution and solid-state experiments have been made on I5N enriched APS 1Based . on~ comparison ~ with other known chemical shifts, they by Chu et ~ determined that different structures are present depending on conditions. For example, on alumina, from iso-propanol or n-methylpyrrolidone, the amine, “closed form”, and “unique form” are found. From water at slightly basic pH (10-11), only the “unique form” and amine cation are found. The unique nitrogen form found is non-amine-like and possibly bonded to a heteroatom. The closed form is a ring structure suggested by P1~eddemann.I~~ On silica, the amine, amine cation, and closed form are found. The authors suggest that a “flip” mechanism likely occurs when reaction of the amine and polymer occurs. 4.1.5. ‘ H studies

Proton studies are valuable to study the kinetics of silane h y d r ~ l y s i s , etc., ’~~ and the surface hydroxyl characterization. 153,154 However, there appear to be few reports of their use in the characterization of the surface-bound species. Gerasimowicz et d 2 0 7 have reported the use of multiple-quantum ( M a ) proton experiments to count the number of spins in the “spin system”. For the trimethylsilyl system the MQ experiment had a plateau at about 10 spins. This is roughly in agreement with the spin system being confined to a single molecule (nine protons). For a C8 chain, a spin system of about 25 is observed consistent with the protons on a single molecule. 4.2. Non-bonded phases

There appear to be only a few reports of NMR of thin organic films that are not permanently bound to the substrate. In the last 2 years, three groups have reported on surfactant groups at interface^.^'^^'^ Interestingly, the majority involved the use of deuterium NMR of specifically labelled surfactants. Soderlind and StilbsZo8have labelled dodecyltrimethylammonium bromide (DoTAB) and decyltrimethylammomium bromide (DTAB) with deuterons in various locations and probed the behaviour of aqueous

314

F . D . BLUM

dispersions with Cab-O-Sil using deuterium NMR line shapes and relaxation times. The spectra were consistent with fast exchange between the surfactant in solution and that on the surface. The surfactant on the surface exhibited a motional gradient with the groups near the ionic head group being more motionally restricted than the tails. This was consistent with the head group being most associated with the surface. No resolved quadrupolar splitting was observed and the mobility of the surfactant was greater than that of chemically attached chains. Extensions of these studies showed that the rate of exchange depended on the packaging of the particles.209 Based on multi-field relaxation studies, the surfactant motions were decomposed into slower isotropic (overall) motions, plus faster anisotropic ones (which are slower than those of corresponding micelles). Macdonald and coworkers21s212 have prepared three choline-methyldeuterated surfactants with different alkyl chains to probe the interaction of the surfactants with polystyrene lattices in aqueous dispersions. For dodecyl(DDPC), tetradecyl- (TDPC), and hexadecyl- (HDPC) phosphocholine-d6, the spectra of surfactant/latex dispersions consisted of “free” and “bound” surfactant resonances.210 The largest quadrupole splitting was found with HDPC. The interaction of the surfactant with the latex was ascribed to the hydrophobic effect and hence the association was believed to be through the surfactant tails. Quadrupolar splittings were observable only at higher surface concentrations. It was concluded that the “bound” surfactant was in slow exchange with the solution phase in spite of the small quadrupolar splitting observed ( 4 0 0 Hz). The quadrupolar splitting of HDPC was also sensitive to the ionic environment of the latex surface.”’ This suggested that this probe might be an effective “molecular voltmeter” for the surface. Relaxation time measurements211were used to infer that the motional state of the head group was similar to that found in the gel state. Soderlind and Blum213 studied the adsorption of head group deuterated sodium 4-( 1’-heptylnony1)benzenesulphonate (SHBS) on alumina. At saturation they found that the inner (surface-bound) layer had dynamics similar to that found in the lamellar liquid crysta1214including quadrupole splitting of the order of about 10-12kHz. Consequently, this layer was believed to be in slow exchange with the other layer or aqueous surfactant. The outer layer, which was unsplit, was believed to be in fast exchange with the aqueous surfactant. At lower coverages, a monolayer of surfactant was formed and was in fast exchange with the aqueous surfactant. Thus the presence of the outer surfactant layer impeded the exchange of the inner (surface) layer. Finally, it needs to be noted that it is possible to probe the surface structure (and dynamics) with 13C NMR as well in surfactant systems. Soderlind and Stilbs215have adsorbed dodecylpyridinium chloride (DPCL), three alkylammonium bromides (DTAB, DoTAB and CTAB) and sodium octylbenzenesulphonate (SOBS) on silica and aluminium oxide. The chemi-

NMR STUDIES OF ORGANIC THIN FILMS

315

cal shift changes observed were consistent with a higher population of trans conformers in the surface-bound surfactant. This is similar to that found in micellar surfactants, although the line widths of the surfactants in the presence of the solids were much greater than those in solution or micelles. 5. CONCLUSION

NMR was found to be a very powerful technique for examining the structure and mobility of thin organic films. It usually requires the use of high surface area materials because the effects of the surface are not usually very long range. Isotope substitution can sometimes provide the necessary increases in sensitivity to obtain useful surface spectra. It also can provide the contrast necessary in determining the properties of the interface, even in the presence of similar material, e .g. organic composites. Proton, deuterium, carbon and silicon experiments are to date the most informative. NMR characterization of bonded interfacial species can be used to identify the species that have been reacted and to determine the type of surface bonding. The mobility of the surface species may also affect the behaviour of the material in use in applications like chromatography and composites. The use of NMR to characterize chromatographic gels should continue to be strong, while that related to composites and small nonbonded molecules should continue to increase. ACKNOWLEDGEMENT The author gratefully acknowledges the financial support of the Office of Naval Research for support of the experiments reported from his laboratory.

REFERENCES 1. H. A. Resing, Adv. Molec. Relax. Proc., 1968, 1, 109. 2. H. Pfeifer, NMR Basic Principles and Progress (eds P. Diehl and R. Kosfeld), Vol. 7, p. 53. Springer-Verlag, New York, 1972. 3. H. Pfeifer, W. Meiler and D. Deininger, Annual Reports on NMR Spectroscopy (ed. G. A. Webb), Vol. 15, p. 291. Academic Press, London, 1983. 4. ‘r. Duncan and C. Dybowski, Surf. Sci. Rep., 1981, 1, 157. 5 . A. M. Zaper and J. L. Koenig, Adv. Colloid Interface Sci., 1985, 22, 113. 6. B. Boddenberg, in Lectures on Surface Science (eds G . R. Castro and M. Cardona), p. 226. Springer-Verlag, Berlin, 1987. 7. F. D. Blum, in Characterization of Composite Materials (ed. H. Ishida), Vol. 12 in Practical Guides for the Surface, Interfacial and Micro Analysis of Materials (ed. R. Brundle and C. Evans). Manning Publications, Greenwich, CT,in press.

316

F. D.BLUM

8. K. Albert and E. Bayer, J. Chromatogr., 1991, 544, 345. 9. A. P. Legrand, H. Hommel, H. Taibi, J. L. Miquel and P. Tougne, Colloids Surf., 1990, 45,39 1. 10. A. L. Choli, Spectroscopy, 1992, 7(3), 24. 11. J. Grandjean, Annual Reports on N M R Spectroscopy (ed. G. W. Webb), Vol. 24, p. 181. Academic Press, London, 1992. 12. T. Cosgrove, B. Vincent, and M. Cohen-Stuart, Adv. Colloid Sci., 1985, 25, 81. 13. T. Cosgrove and P. C. Griffiths, Adv. Colloid Interface Sci., 1992, 42, 175. 14. F. D. Blum, Colloids Surf., 1990, 45, 361. 15. C. A. Fyfe, Solid State N M R for Chemists. CFC Press, Guelph, Ontario, 1983. 16. B. Vincent, Adv. Colloid Interface Sci., 1974, 4, 193. 17. G. J. Fleer and J. M. H. M. Scheutjens, Adv. Colloid Interface Sci., 1982, 16, 341. 18. A. Takahashi and M. Kawaguchi, Adv. Polym. Sci., 1982,46, 1. 19. D. H. Napper, Polymeric Stability of Colloidal Dispersions. Academic Press, New York, 1983. 20. M. A. Cohen-Stuart, T. Cosgrove and B. Vincent, Adv. Colloid Interface Sci., 1986, 24, 143. 21. T. Cosgrove and B. Vincent, Fluid Interface Phenomena (ed. C . A. Croxon), p. 607. Wiley Interscience, New York, 1986. 22. P. G. de Gennes, Adv. Colloid Interface Sci., 1987, 27, 189. 23. S. T. Milner, Science, 1991, 251, 905. 24. H. N. Cheng, Modern Methods of Polymer Characrerization (eds. H. G. Barth and J. W. Mays), Ch. 11, p. 409. John Wiley and Sons, New York, 1991. 25. D. E. Axelson and K. E. Russell, Prog. Polym. Sci., 1985, 11, 221. 26. T. Yu and M. Guo, Prog. Polym. Sci., 1990, 15, 825. 27. R. Voelkel, Angew. Chem. Int. Ed. Engl., 1988, 27, 1468. 28. A. K. Roy and P. T. Inglefield, Prog. N M R Spectrosc., 1990, 22, 569. 29. L. W. Jelinksi, Ann. Rev. Muter. Sci., 1985, 15, 359. 30. B. Bliimich, A. Hagemeyer, D. Schaefer, K. Schmidt-Rohr and H. Spiess, Adv. Muter., 1990, 2, 72. 31. H. Spiess, Chem. Rev., 1991, 91, 1321. 32. F. Heatley, Prog. N M R Spectrosc., 1979, 13, 47. 33. D. W. McCall, Nut. Bur. Stand. (US), Spec. Publ., 1969, 301, 475. Acc. Chem. Res., 1971, 4, 223. 34. F. A. Bovey, Chain Structure and Conformation of Macromolecules. Academic Press, New York, 1982. 35. J. L. Koenig, Spectroscopy of Polymers. American Chemical Society, Washington, DC, 1992. 36. R. A. Komoroski (ed.), High Resolution N M R Spectroscopy of Synthetic Polymers in Bulk. VCH Publishers, Florida, 1986. 37. L. Mathias (ed.), Solid State N M R of Polymers. Plenum Publishing, New York, 1991. 38. T. Miyamoto and H. J. Cantow, Makromol. Chem., 1972, 162, 43. 39. W. G. Miller, Spin Labelling ZI (ed. L. Bedinner), Chapter 4. Academic Press, New York, 1979. 40. F. D. Blurn, J. E. Dickson and W. G. Miller, J. Polym. Sci., Polym. Phys. Ed., 1984,22, 211. 41. K. K. Fox, I. D. Robb and R. Smith, J . Chem. SOC., Furaday Trans. I , 1974, 70, 1186. 42. A. T. Clark, I. D. Robb and R. Smith, J. Chem. SOC., Faraday Trans. I , 1976, 72, 1489. 43. I. D. Robb and R. Smith, Polymer, 1977, 18, 500. 44a. T. M. Liang, P. N. Dickson and W. G. Miller, in Polymerization Characterization by N M R and ESR (eds A. E . Woodward and F. A. Bovey), ACS Symp. Ser., 142, p. 1. American Chemical Society, Washington, DC, 1980. 44b. W. G. Miller, W. T. Rudolf, Z. Vekksli, D. L. Coon, C. C. Wu and T. M. Liang, in

NMR STUDIES OF ORGANIC THIN FILMS

317

Molecular Motion in Polymers by ESR (eds R. F. Boyer and S. E. Keinath). Harwood Academic, Cooper Station, NY, 1980. 45. G. Thambo and W. G. Miller, Macromolecules, 1990, 23, 4397. 46. H. Hommel, L. Facchini, A. P. Legrand and J. Lecourtier, Eur. Polym. J., 1979, 14, 803. 47. H. Hommel, A. P. Legrand, J. Lecourtier and J. Desbarres, Eur. Polym. J., 1979, 15, 993. 48. H. Hommel, A. P. Legrand, H. Balard and E. Papirer, Polymer, 1983, 24, 959. 49. H. Hommel, A. P. Legrand, H. Balard and E. Papirer, Polymer, 1984, 25, 1297. 50. H. B. Ouada, H. Hommel, A. P. Legrand, H. Balard and E. Papirer, J. Chem. SOC., Faraday Trans. I , 1988, 84, 3865. 51. H. Hommel, A. P. Legrand, H. Balard and E. Papirer, Colloids Surf., 1989, 36, 263. 52. H. Hommel, A. P. Legrand, P. Tougne, H. Balard and E. Papirer, Macromolecules, 1984, 17, 1578. 53. H. B. Ouada, H. Hommel, A. P. Legrand, H. Balard and E. Papirer, J. Colloid Interface Sci., 1988, 122, 441. 54. L. Facchini and A. P. Legrand, Macromolecules, 1984, 17, 2405. 55. H. B. Ouada, H. Hommel, A. P. Legrand, H. Balard and E. Papirer, J. Phys., 1986,47, 909. 56. T. Cosgrove and K. G. Barnett, J. Magn. Reson., 1981, 43, 15. 57. K. G. Barnett, T. Cosgrove, B. Vincent, D. S. Sissons and M. Cohen-Stuart, Macromolecules, 1981, 14, 1018. 58. T. Cosgrove, N. Finch, B. Vincent and J. Webster, Colloids Surf., 1988, 31, 33. 59. T. Cosgrove, N. Finch and J. Webster, Colloids Surf., 1990, 45, 377. 60. T. Cosgrove and J. W. Fergie-Woods, Colloids Surf., 1987, 25, 91. 61. K. G. Barnett, T. Cosgrove, B. Vincent, A. N. Burgess, T. L. Crowley, T. King, J. D. Turner and T. F. Tadros, Polym. Commun., 1981, 22,283. 62. T. Cosgrove, T. L. Crowley, B. Vincent, K. G. Barnett and T. F. Tadros, 3. Chem. Soc., Faraday Symp., 1981, 16, 101. 63. T. Cosgrove, T. L. Crowley and B. Vincent, in Adsorption from Solution (eds R. H. Ottenwill, C. H. Rochester and A. L. Smith), p. 287. Academic Press, New York, 1983. 64. T. Cosgrove and K. Ryan, J. Chem. SOC.,Chem. Commun., 1988, 1424. 65. T. Cosgrove and K. Ryan, Langmuir, 1990, 6 , 136. 66. T. Cosgrove, T. G. Heath, K. Ryan and B. van Lent, Polym. Commun., 1987, 28,64. 67. T. Cosgrove, T. M. Obey and K. Ryan, Colloids Surf., 1992, 65, 1 . 68. G. P. van der Beek, M. Cohen Stuart and T. Cosgrove, Langmuir, 1991, 7, 327. 69. T. Cosgrove, T. M. Obey and M. Taylor, Colloids Surf., 1992, 64, 311. 70. J. Meadows, P. A. Williams, M. J . Garvey, R. Harrop and G. 0. Phillips, J . Colloid Interface Sci., 1989, 132, 319. 71. F. D. Blum, B. R. Sinha and F. C. Schwab, Macromolecules, 1990, 23, 3592. 72. F. D. Blum, B. R. Sinha and F. C. Schwab, Polym. Preprints, 1991, 32(1), 271. 73. B. R. Sinha, F. D. Blum and F. C. Schwab, Macromolecules, 1993, 26,7053. 74. A. A. Parker, G. H. Armstrong and D. P. Hedrick, J . Appl. Polym. Sci., 1993,47, 1999. 75. D. C. Douglass and V. J. McBrierty, Polym. Eng. Sci., 1979, 19, 1054. 76. J. C. Kenny, V. J. McBrierty, 2. Rigbi and D. C. Douglass, Macromolecules, 1991, 24, 436. 77. R. A. Kinsey, Rubber Chem. Technol., 1990, 63,407. 78. J. Schaefer, Macromolecules, 1972, 5 , 427. 79. J . Schaefer, S. H. Chin and S . I. Weissman, Macromolecules, 1972, 5 , 798. 80. C. R. Dybowski and R. W. Vaughan, Macromolecules, 1975.8.50. 81. A. D. English and C. R. Dybowski, Macromolecules, 1984, 17, 446. 82. A. D. English, Macromolecules, 1985, 18, 178. 83. J . P. Cohen-Addad and R. Dupeyre, Macromolecules, 1985, 18, 1612. 84. J . P. Cohen-Addad and A. ViaUat, Polymer, 1986, 27, 1855.

318

F. D.BLUM

85. A. P. M. Kentgens, W. S. Veernan and J. van Bree, Macromolecules, 1987, 20, 1234. 86. S. Kaufman, W.P. Schlicter and D. D. Davis, J. Polym. Sci., A2, 1971, 9, 829. 87. J. O’Brien, E. M. Cashell, G. E. Wardell and V. J. McBrierty, Macromolecules, 1976, 9, 653. 88a. M. Ito, T. Nakamura and K. Tanaka, J. Appl. Polym. Sci., 1985, 30,3493. 88b. A. P. Legrand, N. Lecomte, B. Haidar and E. Papirer, J. Appl. Polym. Sci., 1992, 46, 2223. 89. G. Simon, B. Gotschmann, D. Matzen and H. Schneider, Polym. Bull., 1989, 21, 475. 90. P. S. Brown, M. J. R. Loadman and A. J. Tinker, Rubber Chem. Technol., 1992,65,744. 91. J. P. Cohen-Addad, A. Viallat and P. Huchot, Macromolecules, 1987, 20, 2146. 92. V. M. Litvinov and H. W. Spiess, Makromol. Chem., 1992, 193, 1181. 93. W. Kuhn, Angew. Chem. Int. Ed. Engl., 1990, 29, 1. 94. J. L. Koenig, in Controlled Interphase in Composite Materials (ed. H. Ishida), p. 3. Elsevier, New York, 1990. 95. J. L. Koenig, in Magnetic Resonance Microscopy (eds B. Bliimich and M. Kuhn), p. 188. VCH Publishers, Florida, 1992. 96. C. Chang and R. A. Komoroski, Macromolecules, 1989,22,600. 97. S . N. Sarkar and R. A. Komoroski, Macromolecules, 1992, 25, 1420. 98. L. Garrido, J. E. Mark, C. C. Sun, J. A. Ackerman and C. Chang, Macromolecules, 1991, 24, 4067. 99. P. Bliimler and B. Bliimich, Macromolecules, 1991,24, 2183. 100. F. D. Blum, R. B. Funchess and W. Meesiri, in Interfaces in Polymer, Ceramic and Metal Matrix Composites (ed. H. Ishida), p. 205. Elsevier, New York, 1988. 101. F. D. Blum, R. B. Funchess and W. Meesiri, in Solid State NMR of Polymers (ed. L. Mathias), p. 271. Plenum Press, New York, 1991. 102. F. D. Blum and R. B. Funchess, Polym. Preprints, 1988, 29(1), 54. 103. F. D. Blum and R. B. Funchess, Polym. Preprinfs, 1990, 31(1), 544. 104. J. Van Alsten, Macromolecules, 1991, 24, 5320. 105. C. W. Chu and P. D. Murphy, J. Adhesion Sci. Technol., 1992, 6, 1119. 106. V. L. Femandez, J. A. Reimer and M. M. Denn, J. Am. Chem. SOC., 1992, 114, 9634. 107. H. Yasuda, Plasma Polymerization. Academic Press, New York, 1985, 108. N. Morosoff, in Plasma Deposition, Trearmenr, and Etching of Polymers (ed. R. d’Agostino), Chapter 1. Academic Press, New York, 1990. 109. S. Kaplan and A. Dilks, Thin Solid Film, 1981, 84, 419. 110. A. Dilks, S. Kaplan and A. van Laeken, J. Polym. Sci., Polym. Chem. Ed., 1981, 19, 2987. 111. S. Kaplan and A. Dilks, J . Polym. Sci., Polym. Chem. Ed., 1983, 21, 1819. 112. S. Kaplan and A. Dilks, J. Appl. Polym. Sci., Appl. Polym. Symp., 1984,38, 105. 113. R. J. Gambogi, F. D. Blum, H. Cho and H. Yasuda, J. Polym. Sci., Polym. Chem. Ed., 1991, 29, 1801. 114. I. Tajima, T. Suda, M. Yamamoto, K. Satta and H. Morimoyo, Polym. J., 1988,20, 919. 115. H. Lock, G. E. Maciel and C. E. Johnson, J. Muter. Res., 1992, 7, 2791. 116. J. M. Listerud, S. W. Sinton and G. P. Drobny, Anal. Chem., 1989, 61, 23A. 117. W. P. Rothwell, D. R. Holecek and J. A. Kershaw, J. Polym. Sci., Polym. Lett. Ed., 1984, 22,241. 118. K.-P. Hoh, B. Perry, G. Rotter, H. Ishida and J. L. Koenig, J. Adhesion, 1989, 27, 245. 119. L. A. Weisenberger and J. L. Koenig, J. Polym. Sci., Polym. Lett. Ed., 1989, 27, 55. 120. L. A. Weisenberger and J. L. Koenig, Macromolecules, 1990, 23, 2445. 121. L. A. Weisenberger and J. L. Koenig, Macromolecules, 1990, 23, 2454. 122. F. D. Blum, A. S. Padmanabhan and B. Durairaj, J. Polym. Sci., Polym. Phys. Ed., 1986, 24, 493. 123. A. K. Nieminen, J. Liu and J. L. Koenig, J. Adhesion Sci. Technol., 1989, 3 , 455. 124. A. K. Nieminen, M. Evans and J. L. Koenig, New Polymeric Muter., 1990, 2, 197.

NMR STUDIES OF ORGANIC THIN FILMS

319

125. C. G. Chingas, J. B. Miller and A. N. Garroway, J. Mugn. Reson., 1986, 66, 530. 126. D. G. Cory, J. W. M. van 0 s and W. S. Veeman, J . Mugn. Reson., 1988, 76, 543. 127. T. P. Huijgen, H. Angad Gaur, T. L. Weeding, L. W. Jenneskens, H. E. C. Schuurs, W. G. B. Huysmans and W. S. Veeman, Macromolecules, 1990, 23, 3063. 128. J. E. Gambogi and F. D. Blum, Macromolecules, 1992, 25, 4526. 129. N. Zumbulyadis and J. M. OReilly, Macromolecules, 1991, 24, 5294. 130. N. Zumbulyadis and J. M. O’Reilly, J. Mugn. Reson., 1989.82, 613. 131. H. A. Resing, A. N. Garroway, D. C. Weber, J. Ferrars and D. Slotfeldt-Ellingsen, Pure Appl. Chem., 1982, 54, 595. 132. T. L. Weeding, W. S. Veeman, H. Angad Gaur and W. G. B. Huysmans, Mucromolecules, 1988, 21, 2028. 133. T. L. Weeding, W.S. Veeman, L. W. Jenneskens, H. Angad Gaur, H. E. C. Schuurs and W. G. B. Huysmans, Macromolecules, 1989, 22, 706. 134. A. A. Parker, M.-Y. Tsai, G. Biresaw, T. T. Stanzione, G. H. Armstrong and J. J. Marcinko, Muter. Res. SOC.Symp. Proc., 1992, 249, 273. 135. A. Natansohn, Chem. Muter., 1992, 4, 182. 136. A. Natansohn, Polym. Eng. Sci., 1992, 32, 1711. 137. J. B. Miller and A. N. Garroway, J . Mugn. Reson., 1989, 82, 529. 138. D. G. Cory, J. B. Miller, R. Turner and A. N. Garroway, Mol. Phys., 1990, 70, 331. 139. S. W. Sinton, J. H. Iwamiya, B. Ewing and G. P. Drobny, Spectroscopy, 1991, 6(3), 42. 140. L. Garrido, J. L. Ackerman and W. A. Ellingson, J . Mugn. Reson., 1990, 88, 340. 141. C. G. Fry, A. C. Lind, M. F. Davis, D. W. Duff and G. E. Maciel, J. Mugn. Reson., 1989, 83, 656. 142. E. D. Pleuddemann, Silune Coupling Agents, 2nd edn. Plenum Press, New York, 1991. 143. K. Mittal (ed.), Silunes and Other Coupling Agents. VSP, Utrecht, The Netherlands, 1992. This book contains articles reprinted from the Journul of Adhesion Science and Technology based on the “Pleuddemann Symposium” of 1991. 144. H. Ishida, Polym. Composites, 1984, 5, 101. 145. F. D. Blum, W. Meesiri, H.-J. Kang and J. E. Gambogi, J. Adhesion Sci. Technol., 1991, 5,479. 146. K. K. Gilpin and M. E. Gangoda, J. Chomutogr. Sci., 1983, 21, 352. 147. H.-J. Kang, W. Meesiri and F. D. Blum, Muter. Sci. Eng., 1990, A126, 265. 148. C. J. Lund and P. D. Murphy, 1. Adhesion Sci. Technol., 1992, 6 , 33. 149. B. Coleman, in NMR of Newly Accessible Nuclei (ed. P. Laszlo), vol. 2, p. 197. Academic Press, New York, 1983. 150. G. E. Maciel and D. W. Sindorf, J . Am. Chem. SOC.,1980, 102, 7606. 151. D. W. Sindorf and G. E. Maciel, J . A m . Chem. SOC, 1983, 105, 1487. 152. B. Pfleiderer, K. Albert, E. Bayer, L. van de Ven, J. de Haan and C. Cramers, 1. Phys. Chem., 1990, 94, 4189. 153. C. E. Bronnemann, I.-S. Chuang, B. L. Hawkins and G. E. Maciel, J . A m . Chem. SOC., 1987, 109, 1562. 154. C. E. Bronnimann, R. C. Zeigler and G. E. Maciel, J . Am. Chem. SOC,1988, 110,2023. 155. A. Tuel, H. Homrnel, A. P. Legrand, Y. Chevallier and J. C. Morawski, Colloids Surf., 1990, 45, 413. 156. B. A. Huggins and P. D. Ellis, J. A m . Chem. SOC,1992, 114, 2098. 157. G. E. Maciel, D. W. Sindorf and V. Bartuska, J. Chromutogr., 1981, 205, 438. 158. D. W. Sindorf and G. E. Maciel, J. Phys. Chem., 1982, 86, 5208. 159. E. Bayer, K. Albert, J. Reiners, M. Nieder and D. Muller, J. Chromutogr., 1983, 264, 197. 160. C. A. Fyfe, G. C. Gobbi and G. J. Kennedy, J. Phys. Chem., 1985, 89, 277. 161. D. W. Sindorf and G. E. Maciel, J. A m . Chem. SOC.,1983, 105, 3767. 162. E. J. Sudholter, R. Huis, G. R. Hays and N. C. Alma, 1.Colloid Interface Sci., 1985, 103, 554.

320

F.D.BLUM

B. Pfleiderer, K. Albert and E. Bayer, J . Chromatogr., 1990, 506, 343. A. Tuel, H.Hommel, A. P. Legrand and E. Kovats, hngmuir, 1990, 6, 770. S. 0. Akapo and C. F. Simpson, J. Chromatogr. Sci., 1990, 28, 186. T. Bein, R. F. Carver, R. D. Farlee and G. D. Stucky, J. Am. Chem. SOC., 1988, 110, 4546. 167. J. W. de Hann, H. M. Van Den Bogaert, J. J. Ponjee and L. J. M. Van De Ven, J. Colloid Interface Sci., 1986, 110, 591. 168. H.-J. Kang and F. D. Blum, J. Phys. Chem., 1991, 95,9391. 169. G. S. Caravajal, D. E. Leyden, G. R. Quinting and G. E. Maciel, Anal. Chem., 1988,60, 1776. 170. K.-P. Hoh, H. Ishida and J. L. Koenig, Polym. Composites, 1990, 11, 121. 171. J. E. Gambogi and F. D. Blum, Muter. Sci. Eng., 1993, A162,249. 172. K. C. Vrancken, P. van der Voot, I. Gillis-D’Hamers, E. F. Vasnat and P. Grobet, J . Chem. SOC., Faraday Trans., 1992, 88, 3197. 173. J. M. Van Kan, J. J. Ponjee, J. W. DeHaan and L. J. van de Ven, J. Colloid Interface Sci., 1988, 126, 604. 174. N. Nishiyama, K. Hone and T. Asakura, in Interfaces in Polymer, Ceramic and Metal Marrix Composites (ed. H. Ishida), p. 279. Elsevier, New York, 1988. 175. N. Nishiyama, K. Hone and T. Asakura, J. Colloid Znterface Sci., 1989, 129, 113. 176. E. R. Pohl and C. S. Blackwell, in Controlled Interphase in Composite Materials (ed. H. Ishida), p. 37. Elsevier, New York, 1990. 177. M. L. Miller, R. W. Linton, G. E. Maciel and B. L. Hawkins, J. Chromatogr., 1985, 319, 9. 178. J. Kohler, D. B. Chase, R. D. Farlee, A. J. Vega and J. J. Kirkland, J. Chromatogr., 1986, 352, 275. 179. J. J. Chang, A. Pines, J. J. Fripiat and H. A. Resing, Surf. Sci., 1975, 47, 661. 180. D. E. Leyden, D. S. Kendall and T. G. Waddell, Anal. Chim. Acfa, 1981, 126, 207. 181. D. E. Leyden, D. S. Kendall, L. W. Burggraf, F. J. Pern and M. DeBello, Anal. Chem., 1982,54, 101. 182. G . R . Hays, A. D. H. Clague, R. Huis and G. van der Velden, Appl. Surf. Sci., 1982,10, 247. 183. K. Jinno, J. Chromatogr. Sci., 1989, 27, 729. 184. C.-H. Chiang, N.4. Lui and J. L. Koenig, J. Colloid Interface Sci., 1982, 86, 26. 185. A. M. Zaper and J. L. Koenig, Polym. Composites, 1985, 6, 156. 186. D. W. Sindorf and G. E. Maciel, J. Am. Chem. SOC.,1983, 105,1848. 187. R. C. Zeigler and G. E. Maciel, J . Phys. Chem., 1991, 95, 7345. 188. K. Albert, B. Evers and E. Bayer, J . Magn. Reson., 1985, 62, 428. 189. N. Nishiyama, T. Asakura and K. Horie, J. Colloid Interface Sci., 1988, 124, 14. 190. K. Tanaka, S.Shinoda and Y . Saito, Chem. Lett. (Japan), 1979, 179. 191. S. Shinoda and Y. Saito, 1. Colloid Znterface Sci., 1982, 89, 293. 192. M. E. Gangoda and R. K. Gilpin, J . Magn. Reson., 1983, 53, 140. 193. R. K. Gilpin, J . Chromatogr. Sci., 1984, 22, 371. 194. R. K. Gilpin and M. E. Gangoda, Anal. Chem., 1984, 56, 1470. 195. R. K. Gilpin, Anal. Chem., 1985,57, 1465A. 196. R. K. Gilpin and M. E. Gangoda, 1. Magn. Reson., 1985, 64, 408. 197. M. E. Gangoda, R. K. Gilpin and B. M. Fung, J . Magn. Reson., 1987, 74, 134. 198. B. Pfleiderer, K. Albert, K. D. Lork, K. K. Unger, H. Briickner and E. Bayer, Angew. Chem. Int. Ed. Engl., 1989,28,327. 199. J. Van Alsten, Polym. Commun., 1991, 32, 549. 200. A. Tuel, H. Hommel, A. P. Legrand, H. Balard, M. Sidqi and E. Papirer, Colloids Surf., 1991, 58, 17. 201. E. C. Kelusky and C. A. Fyfe, J. Am. Chem. SOC,1986,108, 1746. 202. M. Gangoda, R. K. Gilpin and J. Figueirinhas, J. Phys. Chem., 1989, 93, 4815.

163. 164. 165. 166.

NMR STUDIES OF ORGANIC THIN FILMS 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215.

321

M. Gangoda and R. K. Gilpin, Langmuir, 1990, 6 , 941. R. C. Zeigler and G. E. Maciel, J. Am. Chem. SOC., 1991, 113, 6349. B. Boddenberg, R. Grosse and U. Breuninger, Surf. Sci., 1986, 173, L655. C. W.Chu, D. P. Kirby and P. D. Murphy, J. Adhesion Sci. Technol., 1993, 7, 417. W.V. Gerasimowicz, A. N. Garroway, J. B. Miller and L. C. Sander, J. Phys. Chem., 1992, 96, 3658. E. Soderlind and P. Stilbs, J. Colloid Interface Sci., 1991, 143, 586. E. Soderlind and P. Stilbs, Langmuir, 1993, 9 , 2024. P. M. Macdonald, Y. Yue and J. R. Rydall, Langmuir, 1992,8, 164. Y. Yue, J . R. Rydall and P. M. Macdonald, Langmuir, 1992,8, 390. S. C. Kuebler and P. M. Macdonald, Langmuir, 1992, 8, 397. E. Soderlind and F. D. Blum, 1. Colloid Interface Sci., 1993, 157, 172. E. Soderlind and P. Stilbs, Langmuir, 1993, 9 , 1678. J. R. Duke, R. B. Funchess and F. D. Blum, Langmuir, 1991,7, 1909.

This Page Intentionally Left Blank

Index Acrylonitrilehitrogen plasmas 296 Adsorbed molecules see Zeolites Ag2S-P& glasses 64 27Al MAS-NMR, monocalcium aluminate 47 27A1MAS-NMR spectrum, zeolites 127-9 AI2O3-SiO2gels 55 Alachlor 268 n-alkanes 231,234,236 n-alkyl side chains 230, 231, 237 Alumina 77-9,83 S i c fibre reinforced 24-5 Aluminates 77-9 27Alchemical shift 77 Aluminium nitride 77-9 Aluminosilicate glasses, 27AlMAS-NMR spectra 52-5 3-Aminopropylsilane (APS) 305-8 4-Aminobutylsilane (ABS) 305 N-2-Aminoethyl-3-aminopropyltrimethoxylsilane (AAPS) 306 3-Aminopropyltrimethoxysilane (APMS) 301 Amporphous silica, structure determination of “defects” in 1116 “B MAS-NMR borate glasses and minerals 20-1 borosilicate minerals and glasses 20-1 Bacteriorhodopsin 266 Bioceramics 30,83 Biopolymers 268-9 Bismaleimides (BMI) 225-30,312 Bombyx mori 252 Bonding effect on chemical shift 2-5 Borate glasses 59-62 and minerals, I’B MAS-NMR 20-1 Borates, boron coordination in 35 Borax, I’B MAS-NMR spectra 21 Boric oxide anomaly 2 Boron coordination in borates 35 in glasses 1-2

Boron chemical shift 21 quadrupole coupling constants 21, 35 Borosilicate minerals and glasses, “B MAS-NMR 20-1 Bovine serum albumin (BSA) 268 BR-24 31 Bronsted acid sites 106-26 accessibility of 123-6 methanol molecules hydrogen bonded to 137 unloaded (evacuated) zeolites 106-23 13CMAS-NMR studies of catalytic reactions in sealed samples 160 Calcium aluminates 47-9, 63 27A1chemical shifts 50 Calcium phosphates 31P chemical shifts 83 31PNMR 83 Calcium silicates 44-7 h dration process 44 ”&i chemical shifts 47 CaO-Si02 system 43,44 Ca0-SiO2-AI2O3system 43 Ca0-Si02-Al203-Fe2O3 system 43 Carbonyl carbons 259 S-Carboxymethyl kerateine (SCMK) 257-61 Catalysts 279 CdGePl glasses 64,66 Cellulose (CELL) 250 Cements 43-9 hydration 46 Ceramics 1-27 conventional 29-30, 36-66 high-performance 30,66-83 high-resolution solid-state NMR studies 2%90 multinuclear approach 32 new 29-30 newly accessible nuclei 83 spin-lattice relaxation mechanisms associated with 35-6 spin-lattice relaxation times 32 see also Composite ceramics

324

INDEX

Chabazite, PFG-NMR diffusion data 146 Charge-transfer interactions 247 Chemical shielding anisotropy (CSA) 209 Chemical shift parameters 30 Chemical shifts aluminates 77 anisotropy 31, 34 bonding effect on 2-5 borons 21 calcium aluminates 50 calcium phosphates 83 calcium silicates 47 interaction 3 M kaolinite 36 polyethylene (PE) 192 poly(viny1 alcohol) (PVA) 207 reference compounds 32 silica 77 silicon carbide 70, 74 zeolites 97, 113, 142 Clay minerals 36-43 spin-lattice relaxation times 38 Collagen 253 Combined rotation and multiple pulse spectroscopy (CRAMPS) 31,38, 56, 114,223,244,302 Composite ceramics, flaw detection in green state 24-5 Cordierite 75 Coupling agents 279 CP MAS spectra of silica gels 12-16 CRAMPS See Combined rotation and multiple pulse spectroscopy (CRAMPS) Cross-polarization (CP) 31, 190, 278,299 2D COSY experiment 99 2D INADEQUATE experiment 99 Decyltrimethylammonium bromide (DTAB) 313 Deuteroaminobutylsilane (DABS) 310, 312 Deuteroaminopropylsilane (DAPS) 310, 312 Differential scanning calorimetry (DSC) 282 Dimethylamine (DMA) 166,170-3

Dimethyloctadecylchlorosilane (DMODCS) 306-7 Dim yristoylphosphatid ylcholine (DMPC) 263

Diphenyl sulphide (DPS) 221, 222 Dipolar interaction 34-5, 39 Dodecyl-phosphocholine-d6(DDPC) 3 14 Dodecyldimethylchlorosilane 308 Dodecyltrimethylammonium bormide (DoTAB) 313 Double rotation (DOR) techniques 31, 83, 128-9 Dynamic angle spinning (DAS) 31,83 Dynamic nuclear polarization (DNP) 298

Electron spectroscopy for chemical analysis (ESCA) 296 Engineering plastics 19C-239 ESR studies 283-4,295 Ethylene-propylene-diene rubber (EDPM) 291

Fibrous proteins 252-61 Finite perturbation theory (FPT) 216

Gas phase acidity 107 Glasdepoxy systems 298, 301 Glasdnylon systems 298 Glass/polyester systems 298 Glasses 1-27,4966 boron in 1-2 early studies 1 mixedtoxide 62-4 non-oxide 64-6 Glutathione (GSH) 268 Gramicidin A (GA) 263,264,266

Heterogeneous catalysis 159 Heteronuclear dipolar interaction 31 Hexadecyl-phosphocholine-d6(HDPC) 314 High-Gly-Tyr fractions (HGT) 257-60 High-resolution NMR techniques 30-3 Hirschler-Plank mechanism 124 Homonuclear dipolar interaction 31 Hydrocarbons, intracrystalline diffusion in zeolites 147-51

Inderite, "B MAS-NMR spectra 21

INDEX

J-coupling 95,291 Kaolinite "A1 MAS-NMR spectra 37,43 29Si chemical shift 36 29Simagnetization recovery curves 41 29Si MAS-NMR spectra 41-3 29Si spectra of 36 CRAMPS technique 38 'H NMR spectra 40 structure of 36 Kaolins structure of 36-9 thermal transformation 3943 Kernite, "B MAS-NMR spectra 21 Lewis acid sites 125, 126-32 use of probe molecules 129-32 Li20-BzO3-Si02 glasses 63 Li20-Si02 glasses, 29Si MAS-NMR spectra 50 Li20-Si02 sol-gel glasses 55 LiI-LizS-SiS2glasses 64 Magic angle spinning (MAS) 2, 9-10,31, 160.278 Membrane proteins 2 6 1 4 3-Methacryloxypropyltrimethoxysilane (MPS) 305 Methanol, conversion of 164-6 Methanol molecules hydrogen bonded to Bronsted acid site 137 Methylamines, synthesis of 166-74 Methylene dianiline (MDA) 229 M20-B203-P205glasses 63 Molecular diffusion See Zeolites Monodentate (M-type) silanes 303 Monomethylamine (MMA) 166,170-3 MREV-8 31 Multiple-quantum (MQ) proton NMR experiments 313 Na20-A1203-Si02glasses 55 Na20-B203-Si02glasses 62 Nitrile copolymers 248 NMR spectra, interpretation of 3 3 4 NMR tracer desorption technique 154-8 Nylon 2069,250,299

325

"0enriched samples 32 31P CW NMR spectra, strontium phosphate glasses 19 31PMAS-NMR spectra, strontium phosphate glasses 16 31PNMR, phosphate glasses 16-20 PbO-Si02 glasses 52 Phosphate glasses 57-9 structure determination 16-20 Phylosamia Cynthia ricini 252 Polyacrylonitrile (PAN) 250 Poly( y-n-alkyl-L-glutamate) 237 Polyamide 230 Polybenzimidazole (PBI) 246-7 cis-Polybutadiene (cis-PB) 290 Poly(buty1ene terephthalate) 230 Polycarbonate 230 Pol carbosilanes BSi MAS-NMR spectra 71 curing process 71-4 modified with aluminuim alkoxide 71-4 Poly(2,6-dimethyl- 1,4-phenylene oxide) (PDMPO) 219 Poly(dimethy1 siloxane) (PDMS) 294, 296,308 Poly(E-caprolactone) (PCL) 250 Poly(ether ether ketone) 230 Poly(ethy1ene oxide) (PEO) 244-6, 284-6 Pol ethylene (PE) 190, 300 'C CP MAS-NMR spectra 190-2 chemical shift tensor 192 crystal forms 190 crystal structure 192 drawn (DR) 190 melt-crystallized 197 melt-quenched (MQ) 190, 192 single crystal (SC) 190 single labelled melt-quenched (MQPESL) 199-200 single labelled solution-crytallized (PESL) 199-203 solution-crystallized (PESL) 199 structure model 190 UHMW 194-8 VT I3C CP MAS-NMR 194, 198 Poly(ethy1ene-comaleic anhydride) (PEMA) 248 Poly(ethy1ene terephthalate) 230 Poly(L-glutamate) 231 Polyimides (PI) 223-5,230,296

326

INDEX

Polyisoprene (PIP) 293 cb-1,4-Polyisoprene (&-PIP) 290,291

Polymer alloys 239-51 Polymer/solvent systems 298 Polymers deuterated 288 high-performance 190-239 higher-order structures 189-275 interfacial 279-300 liquid crystalline 234-9 natural 251-69 terminally attached 281 Poly(methy1 methacrylate) (PMMA) 241,243,250,283 Poly( y-n-octadecyl-L-glutamate) 23 1, 234 Poly( y-oleyl-L-glutamate) 233 Poly(oxycarbonyloxyl-1,4phenyleneisopropylylidene-1,4phenylene) (PC) 248 Polyoxymethylene (POM) 209-11 Polypeptides 257 Poly(pheny1ene oxide) 219,248 Poly(pheny1ene sulphide) (PPS) 219-23 Poly(is0-propyl acrylate) (PIPA) 295 Polypropylene 202 Polypyrrole 213-19 amorphous state 217 ‘’N CP MAS-NMR spectrum 214 structure of 214 Polystyrene (PS) 202,241,243,244, 286, 289 Poly(styrene sulphonate) (PSS) 286 Poly(viny1 acetate) (PVAc) 284,295 Poly(viny1 alcohol) (PVA) 204,250, 251, 286,299 films 205 integrated fractions, chemical shifts and linewidths 207 molecular structure 204 Poly(viny1 chloride) (PVC) 241 Poly(viny1 methyl ether) (PVME) 241 Poly(viny1 pyrrolidone) (PVP) 284-6 Poly(viny1 pyrro1idone)-co-(vinyl acetate) (PVP-co-VAc) alumina system 289 Poly(viny1idene fluoride) (PVF2) 250 Poly(vinylpyrro1idone) (PVP) 251 Proteins 257 P-Se glasses 64 Pulsed field gradient NMR spectroscopy (PFG-NMR), molecular migration 138-47

Quadrupole coupling constant, boron 21 Quadrupole interaction 31, 35 Quartz crystals, Raman spectra of irradiated 11 Raman spectra, “defect” lines 11 Reference compounds, chemical shifts 32 Resorcinol244,246 Rhodopsin 266 Rotator phase 234,236 S-carboxymethyl kerateine (SCMK) 257-61 29SiMAS-NMR distribution of silicate structures in silicate glasses 5-1 1 identification and quantization of silicate anions in solid silicates 2-5 Sialons 79-81 Signal enhancement techniques 31-2 Silane coupling agents 300-1 Silanes 279 monofunctional 303 multifunctional 303 Silanols 301 condensation of 15 CRAMPS 302 Silica 74-7,295, 299 29Sichemical shifts 77 Silica gels, CP MAS spectra of 12-16 Silica glasses, Raman spectra of irradiated 11 Silicalite methane adsorbed on 154,155 propane adsorbed on 151 propane diffusion in 152 Silicate anions in solid silicates 2-5 Silicate glasses 49-57 29SiMAS-NMR spectrum 49 silicate structures in 5-11 Silicate structures in silicate glasses 5-1 1 Silicates 74-7 Silicon carbide I3C chemical shifts 70 29Sichemical shifts 70, 74 fibre reinforced alumina 24-5 structure of polymorphs 67-70 Silicon nitride. 29SiMAS-NMR svectra 74

INDEX

Si02-P205glasses 62 Si02-Ti02-Zr02sol-gel glasses 56 Sodium aluminate 77 Sodium 4-( 1'heptylnonyl) benzenesulphonate (SHBS) 314 Sodium mordenite 168 Sodium silicate glasses 6, 56 Sol-gel process 55 Spin-lattice relaxation mechanism 3 5 4 Spin-lattice relaxation times ceramics 32 clay minerals 38 Spinning sidebands (SSB) 6 Strontium phosphate glasses 31PCW NMR spectra 19 3'P MAS-NMR spectra 16 Styrene 288 Styrene-butadiene rubber (SBR) 293 Sulphones 230 Surface-active agents 300-15 bonded phases 300-13 I3C studies 305-9 'H studies 313 2H studies 309-13 "N studies 313 29Si studies 301-5 non-bonded phases 313-15 Surface-bound alkylsilanes 278

Tetradecyl-phosphocholine-d6 (TDPC) 314 Tetramethyl silane (TMS) 3, 5 Tetraxylyl hydroquinone diphosphate (HQDP) 248 Thin organic films 277-321 elastomer-solid systems 290-5 general NMR considerations 279-80 prior reviews 278-9 solid-gas interface 295-9 solid-liquid polymer systems 283-90 solid-solid systems 299-300 Trimethylamine (TMA) 166,170-3 Tropomyosin 253, 255, 257

van der Waals forces 280 2-Vinylpyridene (2-VP) 288 VPDSS 288 VPSDS 288

327

WHH-4 31 Wool keratin 257

Zeolites 91-187, 278 A (ZK-4) 92 adsorbed molecules 132-8 Al-based 135 27A1MAS-NMR studies 101-5, 127-9 anisotropic molecular diffusion 152-4 as catalysts 159 13C MAS-NMR studies adsorbate-adsorbent systems 138 sealed samples 133-8 CaNa-Y 160, 162 catalytic conversion of methanol and ammonia to methylamines 166-74 catalytic conversion of methanol to hydrocarbons 164 characteristic data 93 chemical reactions 159-74 chemical shift for adsorbate-adsorbent systems 142 framework of 92-106 eneral formula 92 H ' MAS-NMR studies 113, 114, 116-19, 122-5 catalytic reactions in sealed samples 174 sealed samples 133-8 'H NMR chemical shifts 113 "C MAS-NMR studies, catalytic reactions in sealed samples 174 H-Y 119-20, 122,162-4 H-ZSM-5 123,135, 164 high-resolution NMR of adsorbed molecules 132-3 high-resolution solid-state NMR 93-6 hydrated 1, 123-6 in situ MAS-NMR studies on sealed samples 159-64 intracrystalline diffusion of hydrocarbons 147-51 large-pore 92, 94 Lewis acid sites 126-32 medium-pore 92,94 molecular diffusion 138-58 NaCa-A 156, 158, 159 Na-X 145, 149-52 Na-Y 137, 140 Na-ZSM-5 135

TTTTT Zeolites - contd. NMR tracer desorption technique 154-8 nuclei used in NMR studies 96 original 91 P-based 135 properties 93 SAPO-5 120, 122, 124 2ySi chemical shifts 97 29Si MAS-NMR studies 97-9, 113 Si-based 135 small-pore 92, 94

solid-state NMR measurements 108 surface properties 132 synthetic 92 unloaded (evacuated) 106-23 use of term 91-2 uses 92 X 92,97 Y 92,97, 105, 106 ZSM-5 101, 157 ZSM-12 99, 102-4 ZSM-20 92 Zirconia 83