Top Banner
This article was downloaded by: [93.108.184.95] On: 08 November 2011, At: 02:40 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Applied Spectroscopy Reviews Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/laps20 Applications of Solid-State NMR to the Study of Organic/ Inorganic Multicomponent Materials Marco Geppi a , Silvia Borsacchi a , Giulia Mollica a & Carlo Alberto Veracini a a Dipartimento di Chimica e Chimica Industriale, Università di Pisa, Pisa, Italys Available online: 04 Dec 2008 To cite this article: Marco Geppi, Silvia Borsacchi, Giulia Mollica & Carlo Alberto Veracini (2008): Applications of Solid-State NMR to the Study of Organic/Inorganic Multicomponent Materials, Applied Spectroscopy Reviews, 44:1, 1-89 To link to this article: http://dx.doi.org/10.1080/05704920802352564 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.tandfonline.com/page/terms- and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages
91

Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

Oct 08, 2014

Download

Documents

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

This article was downloaded by: [93.108.184.95]On: 08 November 2011, At: 02:40Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

Applied Spectroscopy ReviewsPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/laps20

Applications of Solid-StateNMR to the Study of Organic/Inorganic MulticomponentMaterialsMarco Geppi a , Silvia Borsacchi a , Giulia Mollica a &Carlo Alberto Veracini aa Dipartimento di Chimica e Chimica Industriale,Università di Pisa, Pisa, Italys

Available online: 04 Dec 2008

To cite this article: Marco Geppi, Silvia Borsacchi, Giulia Mollica & Carlo AlbertoVeracini (2008): Applications of Solid-State NMR to the Study of Organic/InorganicMulticomponent Materials, Applied Spectroscopy Reviews, 44:1, 1-89

To link to this article: http://dx.doi.org/10.1080/05704920802352564

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes.Any substantial or systematic reproduction, redistribution, reselling, loan,sub-licensing, systematic supply, or distribution in any form to anyone isexpressly forbidden.

The publisher does not give any warranty express or implied or make anyrepresentation that the contents will be complete or accurate or up todate. The accuracy of any instructions, formulae, and drug doses should beindependently verified with primary sources. The publisher shall not be liablefor any loss, actions, claims, proceedings, demand, or costs or damages

Page 2: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

whatsoever or howsoever caused arising directly or indirectly in connectionwith or arising out of the use of this material.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 3: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

Applied Spectroscopy Reviews, 44: 1–89, 2009Copyright © Taylor & Francis Group, LLCISSN: 0570-4928 print / 1520-569X onlineDOI: 10.1080/05704920802352564

Applications of Solid-State NMR to the Studyof Organic/Inorganic Multicomponent Materials

Marco Geppi, Silvia Borsacchi, Giulia Mollica,and Carlo Alberto Veracini

Dipartimento di Chimica e Chimica Industriale, Universita di Pisa, Pisa, Italy

Abstract: The characterization of a variety of organic/inorganic multicomponent ma-terials (OIMM) through solid-state NMR (SSNMR) spectroscopy will be reviewed.Many examples of applications to OIMM will be described, based on the observation ofdifferent nuclei and the use of various SSNMR methods, such as 1D and 2D techniques,measurements on relaxation and spin diffusion processes. OIMM are a very generalcategory of systems differing, for example, by chemical nature and relative amount oforganic and inorganic components, shape and size of the domains, and type of organic-inorganic interface. Some of the most investigated classes of OIMM are organicallymodified silicates, polymer/clay composites, polymer/inorganic filler systems, polymerelectrolytes, stationary chromatographic phases, zeolites, and mesoporous silicas in-cluding small organic molecules. The aspects most efficiently investigated by SSNMRand discussed in this review include physical and/or chemical interactions occurring atthe organic-inorganic interface, structural and dynamic behavior of the organic compo-nents, and dimensions and dispersion of organic and inorganic domains.

Keywords: Nanocomposites, hybrids, microcomposites, clays, zeolites, silicates, poly-mers, chromatographic stationary phases, sol-gel, polymer electrolytes, interface,CP/MAS, line shape and FID analysis, relaxation times, spin diffusion, structure, con-formations, domain dimensions, dynamics

INTRODUCTION

Under the quite general definition of organic/inorganic multicomponent ma-terials (OIMM), many different classes of systems can be included, currentlyattracting an extraordinary interest in both academic research and applications.A codified definition of OIMM is not present in the scientific literature, sowe will use this term referring to those materials constituted by both organic

Address correspondence to Dr. Marco Geppi, Dipartimento di Chimica e Chim-ica Industriale, Universita di Pisa, v. Risorgimento 35, 56126, Pisa, Italy. E-mail:[email protected]

1

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 4: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

2 M. Geppi et al.

and inorganic components, where the average domain sizes of either or bothcomponents range from nano- to micrometers and both components exhibit afunctional role in determining the final properties of the whole material. Thisdefinition excludes, for instance, materials like organometallic compounds,where organic and inorganic moieties are mixed at a subnanometer (molecular)level, as well as materials where the coexistence of organic and inorganiccomponents does not result in significantly different properties with respect tothose of at least one of the components.

Typical OIMM include a huge variety of systems depending on manyfactors; for instance:

1. chemical nature of both organic and inorganic components;2. relative amounts of organic and inorganic components;3. shape and size of organic and inorganic domains;4. kind of interface between organic and inorganic domains.

By suitably combining these factors, materials with very different finalpurposes can be designed. Despite this huge variety, all the OIMM have thecommon feature that the properties of each component are to some extent af-fected by the presence of the others, the combination of the various componentsbeing essential to obtain a final composite material with new and improved prop-erties with respect to the original organic and inorganic systems. For instance,organic polymers are often modified by dispersing in them an inorganic fillerin order to improve several chemico-physical and mechanical properties (e.g.,barrier toward gases, solvent and heat resistance, stiffness, etc.) with respectto the pristine polymer. The inorganic filler can vary for its chemical nature(salts, silica, clays, etc.) and dimensions (from nano- to micrometer domains).One of the most important types of fillers is clays; in this case, not only candifferent types of clays be employed but they can be dispersed within the poly-mer matrix forming either micro- or nanodomains, depending on clay sheetsexfoliation, and possibly leading to polymer intercalation between adjacentclay layers. Often the compatibility between the polymeric matrix and the claycan be increased by treating the surface of the filler with organic modifiers; forexample, the hydrophilic character of clays can be changed into organophilicby exchanging the sodium cations present on the clay sheet surface with organicones, usually long-chain alkyl ammonium cations.

OIMM with very different organic/inorganic relative ratios exist, corre-sponding to different classes of materials. An organic polymer matrix is presentnot only in the above-described polymer/inorganic filler systems but also, forinstance, in polymer electrolytes and catalysts made of polymer-supportedmetal particles. In the last two cases, in spite of its predominant amount, therole of the organic phase is essentially to support and modulate the main actionplayed by the inorganic component (transport of positive charges or catalyticactivity of the metal particles). OIMM where the inorganic component is

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 5: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 3

quantitatively predominant include: (a) organically modified silicates or glassesand (b) zeolites or porous silica including organic compounds. In the firstcase, common uses are as fillers for polymers or stationary chromatographicphases. In the second case, they are, for instance, employed as drug-deliverysystems or to study catalytic processes, size-selective molecular separations,and confinement effects on fluids.

The final macroscopic properties of the OIMM, and therefore their applica-tions, are strongly dependent on their microscopic properties, and in particularon the nature and dimensions of organic-inorganic interfaces, mechanisms ofinteraction between organic and inorganic components, and structural and dy-namic properties of either the organic or inorganic phases. However, it mustbe outlined that the inorganic component is often constituted by crystalline orglassy phases, which do not experience strong modifications in passing fromtheir pure state to the OIMM, apart from the regions at the interface with theorganic domains. On the contrary, when the organic components are processedto give the OIMM, often both their interfacial and bulk properties exhibit no-ticeable changes, which deserve to be investigated.1 Having a deep knowledgeof the above-mentioned microscopic properties and understanding their rela-tionships with OIMM macroscopic behavior are crucial steps for new materialsfor more and more specific purposes to be designed and tailored.

In this context, the role played by solid-state NMR (SSNMR) is wellrecognized, and it is in continuous growth. In general, this is due to the greatdetail that SSNMR can give at a microscopic level, also arising from thepossibility of studying many different nuclear spins and, for each of them, avariety of properties. For OIMM, some very important aspects are, in particular:

1. Different NMR-active nuclei are usually present in the organic and inor-ganic components, thus allowing their selective investigation directly in theOIMM, without the need of physically separating them.

2. The structural properties of either the single components or their interfacecan be studied over a broad spatial range (1–1000 A), giving the possibilityof investigating aspects ranging from the chemical structure and nature ofthe interface, to the phase structure within the single domains, to the relativedistribution of the heterodomains within the OIMM.

3. Specific dynamic processes can be studied over a broad time range(10−11 − 102 s), thus including a variety of molecular motions ranging, forinstance, from very slow chain fluctuations in glassy or crystalline polymersto very fast reorientations of either chains in rubbery polymers or molecularmoieties in low-molecular weight compounds.

1A somewhat special case is that of composites obtained by sol-gel processes, wherethe organic and inorganic domains are simultaneously built from suitable monomersrather than derived from preformed organic and/or inorganic components.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 6: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

4 M. Geppi et al.

Despite the importance of SSNMR for the study of OIMM, only a fewreviews have been published on this subject, all of them concerning SSNMRapplications to specific OIMM classes: polymer/clay nanocomposites (1, 2),sol-gel–derived materials (3, 4), guest molecules confined in mesoporous silicamaterials (5), chromatographic stationary phases (6, 7), bones (8), and silicon-based polymers (9).

We attempt here to review the applications of SSNMR to the general cate-gory of OIMM through many examples reported in the literature, concerning alarge variety of materials, which we believe representative and/or explicative ofthe various SSNMR techniques and approaches. Rather than on the type of ma-terial, the main focus will be on the SSNMR techniques and on the informationthey can give. The review is divided into four sections, reporting applicationsfor the study of different aspects. The first section deals with the structural prop-erties of the organic-inorganic interfaces, including the chemical and physicalinteractions between the two components. The second and third sections con-cern structural and dynamic behavior of the organic component, constituted bylow-molecular weight (low-MW) compounds and polymers, respectively, andin particular the changes experienced in passing from the pure organic systemsto the OIMM. The final section deals with organic and/or inorganic domaindimensions and their reciprocal arrangement within the OIMM. Within eachsection, a further division is based on the different SSNMR techniques thathave been employed, often corresponding to specific information attainable onthe systems investigated. In spite of this division, made necessary for the sakeof clarity, it must be outlined that in most cases a satisfactory characterizationof OIMM requires a synergic approach. This consists not only in the investi-gation of the different components but in the use of many different SSNMRexperiments, often to be combined with various microscopic, spectroscopic,diffractometric, and calorimetric techniques.

This review will not cover the basic aspects of NMR or, more precisely,of SSNMR, for which the reader can refer to several textbooks (10–15) andreference works (16, 17). In this review the NMR techniques will be treatedconcisely; for a more complete discussion and/or for more technical aspects,the original literature should be consulted. For a concise, albeit nonexhaustiveintroduction to SSNMR basics and commonly used techniques, the reader canalso refer to our review on SSNMR of pharmaceuticals, recently appearing inthis journal (18).

ORGANIC-INORGANIC INTERFACES

The organic-inorganic interface is a very important part of OIMM since itscharacteristics have a crucial role in determining the properties of the finalmaterials. Taking as a very general definition of interface that of the “place”where organic and inorganic components come in some “contact,” in the wide

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 7: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 5

variety of OIMM such definition includes many different cases. There aresystems in which the organic-inorganic interface can be quite easily identifiedand distinguished from the individual components. It is the case, for example, offillers obtained by coating inorganic particles with organic modifiers covalentlybonded onto the particles’ surface: the interface can be identified with thesubmolecular layer where chemical bonds between the inorganic particles andthe organic modifiers occur. On the other hand, there are systems, such asmany hybrid networks obtained by sol-gel processes, in which the organic andinorganic components are intimately mixed and it is more difficult to clearlydistinguish organic and inorganic domains; in these cases, the interface canbe considered as spread over the entire material. Another important variableis the nature of the interactions occurring at the organic-inorganic interface:covalent or hydrogen-bonds, ionic or weak Van der Waals interactions can beindeed established. Moreover, the interface can be “static,” when the organicand inorganic components are covalently bonded, or “dynamic,” if a sort ofequilibrium among different situations is present.

Characterizing either the interactions occurring at the interface or its chem-ical structure is a very important task for obtaining a deep molecular knowledgeof OIMM. In fact, this is not only fundamental to elucidate the results of thepreparation of new OIMM but also for trying to understand the molecularmechanisms at the basis of the observed functional properties of the materi-als (improvement of the mechanical properties of a polymer in polymer/fillercomposites, catalytic activity of zeolites, etc.). As already stated, SSNMR isone of the most powerful techniques available for accessing such a kind ofinformation, and in fact many studies can be found in the literature in whichorganic-inorganic interfaces in a variety of OIMM have been investigated bymeans of this technique.

This section is divided in three subsections reporting examples of studieswhere the organic-inorganic interface in OIMM has been investigated by meansof: (i) mono-dimensional (1D) spectra of spin-1/2 nuclei; (ii) bi-dimensional(2D) spectra of spin-1/2 nuclei; (iii) experiments on quadrupolar nuclei. Thisdivision reflects the intrinsic differences in NMR behavior between nuclei withspin I = 1/2 and quadrupolar ones, with spin I > 1/2. Moreover, as far asspin-1/2 nuclei are concerned, the two big categories of 1D and 2D spectra arerepresentative of almost all the studies reported in the literature.

1D Spectra of Spin-1/2 Nuclei

Despite the wide variety of 1D SSNMR experiments available, most of theapplications to the characterization of organic-inorganic interfaces in OIMMpresent in the literature are based on the acquisition and analysis of simple 1Dhigh-resolution spectra, in which just the dependence of the isotropic chemicalshift on the local environment of the nuclei is exploited. With respect to NMR

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 8: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

6 M. Geppi et al.

Table 1. Some relevant properties of spin-1/2 nuclei, taken from official IUPAC rec-ommendations (© 2001 IUPAC) by Harris et al. (181)

Natural abundance Frequency ratioa Relative receptivityb

Isotope (x/%) (�/%) (DP )

1H 99.9885 100.000000 1.00029Si 4.6832 19.867187 3.68 × 10−4

13C 1.07 25.145020 1.70 × 10−4

31P 100 40.480742 6.65 × 10−2

19F 100 94.094011 0.83415N 0.368 10.136767 3.84 × 10−6

aRatio of the resonance frequency of the reference to that of the protons of TMS atinfinite dilution (in practice at volume fraction φ = 1%) in CDCl3. Reference compoundsfor each nucleus can be found in Harris et al. (181).

bReceptivity relative to that of 1H; receptivity of a nucleus in natural abundance,which influences the NMR signal strength, is proportional to [γ 3xI (I + 1)], where γ isthe magnetogyric ratio of the nucleus and I the nuclear spin quantum number.

in solution, in the solid state the isotropic chemical shift is also sensitive tocrystalline and molecular conformational properties. While this dependence iswidely exploited for characterizing the organic component, as it will be de-scribed and exemplified in the next two sections, in most of the reported studiesconcerning the organic-inorganic interface, the dependence of the isotropicchemical shift on chemical modifications or physical interactions occurring inthe nuclear environment is especially exploited. 29Si, 1H, and 13C are the mostcommonly observed nuclei but several studies concerning 15N, 31P, 19F havealso been reported; some useful properties of these nuclear spins are reportedin Table 1. Apart from 1H (and in some cases 19F), which deserves special con-siderations that will be discussed later, the others are “rare”2 nuclei, for whichthe two main causes of poor spectral resolution occurring in SSNMR spectraare chemical shift anisotropy and dipolar couplings with protons. These canusually be removed, and therefore high-resolution 1D spectra can be obtained,by combining magic angle spinning (MAS) and high power decoupling (HPD)from protons.

1D spectra of rare nuclei can be acquired either by direct excitation (DE)or by cross polarization (CP), where the signal of the rare nuclei is built upby magnetization transfer from dipolarly coupled abundant nuclei, typicallyprotons. Provided that long enough recycle delays between two consecutive

2Here we will refer to rare nuclei as those either with a low natural abundance (e.g.,13C) or belonging to those elements usually present at low concentrations in a molecule(e.g., 31P or 19F, except in perfluorinated systems); on the contrary, 1H is usually anabundant nucleus.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 9: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 7

radio-frequency pulses are used, DE spectra give quantitative responses; thatis, the areas underlying the signals are proportional to the amounts of the cor-responding resonating nuclei. On the contrary, unless a quite time-consumingcalibration procedure is applied, CP spectra are usually not quantitative. On theother hand, quantitative DE spectra with a sufficiently good signal-to-noise ratiooften require long experimental times, while CP significantly improves the rarenuclei sensitivity. Therefore, the choice of the technique is strongly dependenton the system under investigation and on the information to be obtained.

In the following, examples of studies of the organic-inorganic interface inOIMM based on 1D SSNMR spectra are reviewed, organized on the basis ofthe observed nucleus.

Silicon-29

29Si is probably the most exploited nucleus for the investigation of organic-inorganic interfaces in OIMM. This is due to both its massive presence at theinterface and its good SSNMR accessibility. As reported in Table 1, its relativereceptivity is about twice as much as that of 13C. Moreover, its chemical shiftanisotropy is usually moderate, and it can be almost completely averaged outat MAS frequencies accessible on most commercial spectrometers. The mostcommon use of 29Si 1D spectra in the investigation of organic-inorganic inter-faces is the qualitative characterization of their chemical structure, but quantita-tive information can also be obtained, as it will be exemplified in the following.

29Si isotropic chemical shift is very sensitive to silicon chemical environ-ment and, in particular, to its connectivities with oxygen and carbon nuclei. Inorder to readily identify silicon sites differing for their connectivities, the Q, T ,D, M notation has been introduced and it is commonly used. Slightly differentdefinitions of this notation can be found in the literature, often depending on thespecific class of chemical system (silicates, polysiloxanes, alkoxysilanes, etc.)(see, for instance, references 7, and 19–21). Nonetheless, in general, siliconatoms forming four, three, two, and one Si O bonds can be indicated with thesymbols Q, T , D, and M , respectively. Often a superscript n is also indicated;when only silicons, carbons or protons are present in the second coordinationsphere, n indicates the number of silicon atoms further bonded to that of interest(Si) through Si O Si bonds. n can assume integer values in the range 0–4,0–3, 0–2, and 0–1, for Q, T , D, and M silicon atoms, respectively. In most ofthe cases encountered in OIMM, T , D, and M silicon atoms also form one, two,and three Si C bonds, respectively. Q, T , D, and M silicon nuclei resonatein different spectral regions and they can be easily distinguished from theirisotropic chemical shift, which regularly increases from Q to M . Moreover, foreach type of silicon atom, the isotropic chemical shift increases with decreasingthe silicon cross-linking (number of Si O Si bonds, n). Even if chemical shiftvalues depend on the particular system considered, rough chemical shift rangesfor Q, T , D, and M silicon nuclei are −120 ÷ −85 ppm, −70 ÷ −45 ppm,

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 10: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

8 M. Geppi et al.

Figure 1. 29Si CP/MAS spectra of organic-inorganic nanocomposite films. The assign-ment of the signals to the different silicon sites (Q, T , D, M) is reported, together withthe corresponding molecular structures. Reused with permission from Brus et al. (23).Copyright (2004) American Chemical Society.

−25 ÷ −5, and 0 ÷ 15 ppm, respectively. In Figure 1, 29Si CP/MAS spectra ofnanocomposite films are reported, in which signals arising from Q, T , D, andM silicon sites can be observed and the corresponding chemical structures areshown. From the above, it is quite clear that a 29Si 1D high-resolution SSNMRspectrum can be a very direct and almost unparalleled source of information

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 11: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 9

for characterizing the chemical structure of the organic-inorganic interface,as demonstrated by the huge amount of relevant papers in the literature. Theobtained information is often very useful for characterizing an innovative mate-rial, for elucidating the results of a preparation method, and, when possible, fortrying to establish correlations between molecular properties of the interfaceand observable “macroscopic” properties of the materials.

Some examples of applications to many different kinds of OIMM arelisted in the following. 29Si 1D spectra have been used for characterizing theinterface by Brus (22) in organically modified polysiloxane networks pre-pared by sol-gel co-condensation; by Brus et al. (23) in silica/epoxy films; byMarini et al. (24) and Geppi et al. (25) in polyethylene-b-poly(ethylene gly-col)/silica hybrid materials prepared by sol-gel processes; by Apperley et al.(26) in silica/dimethylsiloxane hybrids; by Deng et al. (27) in perfluorosul-fonic acid (PFSA)/organically modified silicon oxide (ORMOSIL) hybrids;by Framery and Mutin (28) in silica wet gels and xerogels obtained by hy-drolysis of tetraethoxysilane; by Joseph et al. (29) in 3-(trimethoxysilyl)propylmethacrylate-modified silica and silica/poly(methyl methacrylate) composites;by Li et al. (30) in sol-gel–derived hybrid materials based on triethoxysilylateddiethylenetriamine and tetramethoxysilane; by Wang et al. (31) in poly(amide-imide-silica) hybrids; by Comotti et al. (32) and Kao et al. (33) in orderedmesoporous organosilicas; by Bauer et al. (34) in organically modified nano-sized silica; by Herrera et al. (35, 36), Wheeler et al. (37), and Borsacchi et al.(38) in organically modified laponites; by Borsacchi et al. (39) in silica-coatedbarium sulphate submicronic particles surface-modified with stearic acid; byLynch et al. (40) and Pursch et al. (41–43) in chromatographic organic-bondedsilica stationary phases; by Lindner et al. (44) in polysiloxane-bound ether-phosphines and ruthenium complexes; and by Armelao et al. (45) in Zr- andHf-containing silica-based hybrid materials. Applications to systems where el-ements different from carbon, proton, and silicon are in the second coordinationsphere of silicon atoms were also reported. For instance, Templin et al. (46)observed that the formation of Si O Al bonds for T silicon atoms was asso-ciated to a chemical shift increase with respect to the corresponding Si O Sisites. A chemical shift increase was also observed by Alonso et al. (47) forT silicon atoms forming Si O Ge bonds, while a decrease was observed byHoebbel et al. (48) associated to the formation of Si O Ti bonds.

One example by Borsacchi et al. (49) of both qualitative and quantitativestudy of an organic-inorganic interface by means of 29Si 1D SSNMR spectra isdescribed in more detail in the following, concerning an organically modifiedsilica used as filler for polyethylene films. The filler was prepared by reactingsilica gel with 3-(trimethoxysilyl)propyl methacrylate (TSPM; see Figure 2),through hydrolysis of the TSPM alkoxysilane groups and condensation of thesilanols, arising from both TSPM and silica. In Figure 2 the quantitative 29SiDE/MAS spectra of the silica before and after the reaction with TSPM arereported. In addition to the Q signals between −112 and −93 ppm (Q4, Q3,

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 12: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

10 M. Geppi et al.

Figure 2. 29Si DE/MAS spectra of (a) silica and (b) silica-TSPM. The TSPM structureand a schematic picture of the silica-TSPM structure as obtained from NMR results arealso shown.

Q2), also present in the spectrum of silica, T 3 and T 2 signals were observedin the spectrum of the modified silica (silica-TSPM) at −68 and −59 ppm,respectively, ascribable to reacted TSPM silanol groups. The presence of T 3

and T 2 signals confirmed the occurrence of the condensation reaction betweensilica and TSPM, and the absence of T 0 signal indicated that unreacted TSPMwas not present. The analysis of the peak areas, obtained by a spectral deconvo-lution and suitably normalized, allowed detailed information to be obtained onthe condensation reaction. By comparing the number of reacted silica hydroxylgroups with the maximum number of TSPM potential grafting sites with silica,it could be inferred that the condensation reaction occurred not only betweenTSPM and silica but also among different TSPM molecules, leading to theformation of TSPM aggregates, grafted to the silica surface mainly throughmono-dental anchoring and, only to a minor extent, through bi-dental anchor-ing. A picture of the silica-TSPM structure as obtained from NMR results,in agreement with what was previously proposed by Bauer et al. (34, 50), isreported in Figure 2.

Provided that sufficiently long recycle delays are used between two con-secutive excitation pulses, quantitative DE spectra can be straightforwardly

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 13: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 11

obtained; the drawback of this approach is that 29Si nuclei with spin-latticerelaxation times of tens of seconds are commonly encountered, requiring recy-cle delays of several minutes and then long experimental times for collectinga spectrum with a good signal-to-noise ratio. On the contrary, good CP spec-tra can be obtained in significantly shorter times, since the recycle delay isdetermined by the usually much shorter 1H T1. However, unless suitably cal-ibrated, CP spectra do not give quantitative information since the efficiencyof 1H →29Si magnetization transfer strongly depends on the strength of theheteronuclear dipolar couplings, which can be very different for different 29Sinuclei. In spite of this, it is not uncommon to find in the literature studies inwhich 29Si CP spectra have been mistakenly used for obtaining quantitativeresults. On the other hand, CP spectra can be correctly used for quantitativepurposes, provided that CP dynamics curves3 are built. Brus (22) presented anextensive quantitative 29Si study, in which both DE and CP 29Si spectra wereused and compared. From 29Si DE/MAS spectra the author could obtain severalquantitative parameters concerning the composition and condensation degreeof several organically modified polysiloxane networks. Moreover, 1H-29Si CPdynamics curves were built for each distinguishable signal. 29Si CP dynamicscurves for one of the systems investigated are reported in Figure 3. By fittingthese curves with a suitable equation, the amounts of the different silicon nucleicould be obtained. In particular, the curves were fitted with equation (1), validunder the common assumptions that TSiH<T H

1ρ and TSiH << TSi1ρ :

I (tC) = I ∗

1 − λ

(1 − exp

[−{1 − λ} tC

TSiH

])exp

(− tC

T H1ρ

)(1)

where I ∗ is the ideal full cross-polarized 29Si magnetization, tC is the contacttime and λ = TSiH/T H

1ρ . TSiH is the cross-polarization time, which is shorterfor silicon nuclei experiencing stronger dipolar interactions with protons (i.e.,major density of spatially close protons and minor motional averaging). T H

and T Si1ρ are the spin-lattice relaxation times in the rotating frame of 1H and 29Si

nuclei, respectively. By comparing the results obtained from DE spectra andCP dynamics curves, the author found that the amount of the fully condensedQ4 silicon nuclei as determined from CP was smaller than that obtained fromDE. This clearly indicated that a fraction of Q4 silicon nuclei, straightforwardlyquantified, was too far from protons for the CP process to be effective.

The dependence of TSiH on the strength of the dipolar couplings withprotons allows CP spectra acquired at different contact time values to also beused for investigating spatial proximities between silicon and hydrogen atoms,

3CP dynamics curves are plots of rare nucleus signal intensity vs. contact time, i.e.,the experimental time during which the magnetization transfer from protons occurs.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 14: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

12 M. Geppi et al.

Figure 3. 29Si CP dynamics curves for the signals of Q4, Q3, and D2 silicon nucleiin an organically modified polysiloxane network. In the figure, τ is the contact time,indicated as tc in the text. Reused from Brus (22) with kind permission from SpringerScience+Business Media, Copyright (2002) Kluwer Academic Publishers.

which can be especially useful for characterizing the structure of an interface.An example of an advanced application of this use of CP is that reportedby Baccile et al. (51), who exploited a 1H-29Si-1H double CP experiment toinvestigate the interface between a templated silica and its organic templatingagent, the surfactant cetyltrimethylammonium bromide. In this experiment,the magnetization is first transferred from protons to dipolarly coupled siliconnuclei, during a contact time tC1, and then transferred back from the siliconnuclei to the nearby protons, during a second contact time tC2; at the end aproton spectrum is acquired. The authors chose to use a tC1 value at which theintensity of the Q3 silica silicons signal was optimized and then spectra wereacquired at increasing values of tC2. At very short tC2 the spectra showed onlysignals of protons strongly coupled with the silicon nuclei, while at increasingvalues, signals from protons more weakly coupled, progressively farther fromthe silica surface, started to appear. In Figure 4, 1H-29Si-1H double CP spectraacquired for two silicas, containing the same templating agent, but preparedunder either acidic (CTAB/HCl) or basic (CTAB/OH) conditions, are shown.In the CTAB/HCl spectra an evident signal was observed due to silica silanolprotons (around 7 ppm), while only a weak and broad corresponding signal(ranging from 5 to 15 ppm) was detected in the CTAB/OH spectra, assignedto the scarcely present silanols hydrogen bonded to a siloxy group SiO−.Moreover, in the CTAB/OH spectra the signal at 3.3 ppm, due to the protons ofthe ammonium polar head of the surfactant, appeared at tC2 = 100 µs, whilein the case of CTAB/HCl it was observed only for tC2 ≥ 300 µs, suggesting a

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 15: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 13

Figure 4. 1H-29Si-1H double CP spectra acquired for CTAB/HCl and CTAB/OH fora series of tC2 (indicated as tCP 2 in the figure) values, ranging from 100 µs to 15 ms.Reprinted with permission from Baccile et al. (51). Copyright (2007) American Chem-ical Society.

stronger interaction with the polar head of the surfactant for the silica preparedunder basic than acidic conditions, as also confirmed by other experiments.

Hydrogen-1

Despite their very extensive use in liquid-state NMR, the obtainment of high-resolution 1H 1D spectra is not straightforward in the solid state. In fact, thepresence of strong 1H-1H dipolar couplings (typically tens of kHz), which canbe only partially averaged out at the commonly available MAS frequencies, candramatically reduce the spectral resolution obtainable. This problem has beenpartially overcome by the development of advanced multiple-pulse sequencesapplied in combination with MAS (CRAMPS) (52), but the future achievementof higher spinning rates in MAS probes still looks very appealing. Neverthe-less, there are cases in which standard 1H MAS spectra show a good resolution.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 16: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

14 M. Geppi et al.

Several papers reporting 1H MAS spectra of OIMM are present in the litera-ture, where useful information for the characterization of the organic-inorganicinterface has been obtained. In particular, 1H MAS spectra can be very usefulfor investigating the hydrogen-bonding interactions often occurring at the in-terfaces. Indeed, the isotropic chemical shift of protons involved in hydrogenbonds is extremely sensitive to the strength of the interaction, a stronger hydro-gen bond determining a larger deshielding of the proton nuclei and thereforean increase of their chemical shift. Moreover, the signal line width is oftendirectly related to the heterogeneity degree of the chemical environment of thehydrogen-bonded protons. For example, it is known that in silica gel three maintypes of protons exist, all distinguishable on the basis of their isotropic chemicalshift: “isolated” non-hydrogen-bonded silanol protons, resonating at about 1.2–1.7 ppm, physisorbed hydrogen-bonded water protons, resonating at about 3–5ppm, and hydrogen-bonded silanol protons, which usually give rise to a broadpeak, ranging from 2 to 8 ppm (22, 53). In the studies of OIMM, the most com-mon case in which interfacial hydrogen bonds have been studied by 1H MAS isindeed that of silanol protons, which are often present at the organic-inorganicinterface (for instance, in organically modified silicas, silicates, zeolites, hybridmaterials obtained by alkoxysilanes sol-gel reactions, etc.).

An example is given in Figure 5, taken from a paper by Shenderovich et al.(54), where the 1H MAS spectra of an ordered mesoporous silica (MCM-41)before and after treatment with an excess of deuterated pyridine are shown. Thesilanol protons signal, observed at about 1.8 ppm as due to isolated groups, wasshifted, in the presence of pyridine, at about 10 ppm, indicating the occurrenceof a strong hydrogen bonding between silica silanols and pyridine.

Other examples are those reported by Brus (22) and Brus et al. (55).In both the papers the authors drew information concerning the structure of

Figure 5. 1H MAS spectra of the ordered mesoporous silica MCM-41 (a) before and(b) after treatment with an excess of deuterated pyridine-d5. Reused with permissionfrom Shenderovich et al. (54). Copyright (2003) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 17: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 15

organically modified polysiloxane networks, in which the organic-inorganicinterface could be considered as spread over the whole material, by acquiring1H MAS spectra either by DE or Hahn-echo pulse sequence, which allowsthe dipolar coupling strength to be estimated. In particular, the dependenceof the silanol and hydrogen-bonded water proton signals on dehydration andchemical composition of the materials was followed. In Brus (22), the isotropicproton chemical shifts were also used to estimate the hydrogen bond lengthsfor hydrogen-bonded water and silanols. Moreover, in Brus and Skrdlantova(55), an interesting comparison between the results of mechanical tests andthe line width of signals of the strongly hydrogen-bonded water moleculeswas reported; it was observed that a higher mechanical strength correspondedto a larger line width, suggesting that hydrogen-bonding interactions had asignificant effect onto the mechanical properties of the materials.

Hybrid networks obtained by sol-gel processes were also studied by Geppiet al. (25), who observed that both the chemical shift and line width of hydrogen-bonded silanol protons resulted to be very sensitive to the type of curingtreatment to which PE-PEG/silica hybrid materials were subjected.

Schenkel et al. (56) exploited the proton chemical shift dependence on thehydrogen-bonding strength for investigating the adsorption of short-chain n-alcohols on Na-X zeolite. In particular, they observed the occurrence of zeolite-alcohol hydrogen bonding from the chemical shift increase of both hydroxyland alkylic protons of the alcohols (with respect to the values observed forthe same alcohols highly diluted in CDCl3, where hydrogen bonds could beconsidered absent) and studied the dependence of the hydrogen-bonding on thealcohol chain length.

Interactions between guest molecules and host matrices for silica matricesdoped with pH indicators, and mesostructured hybrid silicates templated bysurfactant molecules, have been investigated by Camus et al. (57) by means of1H MAS spectra acquired at high field (14 T) and high spinning rates (30 kHz).

As a last example, Bohlmann and Michel (58) applied 1H MAS to the studyof the interface between catalytically activated zeolites and organic moleculesadsorbed on them. By comparing the spectra acquired before and during theabsorption of 2-butene and acetone on several activated zeolites, as well asafter their desorption, the authors could obtain detailed information concerningthe adsorption interactions and the catalytic conversions. Hydrogen-bondinginteractions between acetonitrile and several zeolites were also investigated byproton isotropic chemical shift calculations, as reported by Simperler et al. (59).

Carbon-13

13C is one of the most exploited nuclei in SSNMR of OIMM; this is due toits large presence in the organic components and to the fact that usually 13C1D high-resolution spectra can be easily obtained by combining MAS andHPD. Premising that in many cases it is difficult to clearly mark the interface

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 18: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

16 M. Geppi et al.

boundaries, in most of the studies concerning OIMM 13C is mainly exploitedfor the characterization of the structural and dynamic properties of the organiccomponents, and for this reason 13C studies will be extensively reviewed inthe next two sections. Nevertheless, there are cases in which information morespecifically concerning the organic-inorganic interface can be obtained. In thefollowing we will mention some applications to different systems.

Often 13C spectra can be very useful as support for the synthesis of newmaterials. For example, Kao et al. (33) reported the synthesis and SSNMRcharacterization of an ordered vinyl-functionalized mesoporous silica; fromthe presence of the vinyl group signals at 129 and 137 ppm in the 13C CP/MASspectra (see Figure 6) and the strong decrease, after suitable extraction, of thetemplate signals at 70–76 ppm, the authors could confirm both the successfulorganic modification of the silica and the almost complete template removal.

Armelao et al. (45) used 13C CP/MAS spectra to characterize theorganic part of a nanostructured hybrid material in which methacry-late moieties could be considered as bridging groups between silica andmetal oxoclusters. Such hybrid materials were indeed obtained by copoly-merization of methacrylate-modified metal oxoclusters with prehydrolyzedmethacryloxymethyltriethoxysilane (MAPTMS). In MAPTMS the Si(OR)3

Figure 6. 13C CP/MAS spectra of an ordered vinyl-functionalized mesoporous silica(a) before and (b) after template extraction. Reused from Kao et al. (33). Copyright(2006), with permission from Elsevier Inc.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 19: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 17

functionality acted as silica network former, while the methacrylate group al-lowed the metal oxoclusters to be covalently bonded to the silica matrix. From13C CP/MAS spectra the authors could confirm the presence of the oxoclustersby the observation of their carbonyl group signals; moreover, by looking at thesignals arising from non-polymerized double bonds, they could observe a mi-nor efficiency in the silane polymerization and silane/cluster copolymerizationin correspondence of a larger amount of incorporated oxoclusters. Moreover,by acquiring spectra at increasing temperature, they could observe that thepresence of the oxoclusters both induced an early cleavage of the Si C bondsand catalyzed the thermal decomposition of the propyl groups.

The interactions occurring at the interface between inorganic surfaces andorganic compounds (surface modifiers, molecules included in pores, etc.) couldbe studied in many cases through 13C 1D high-resolution spectra. Boiadjiev etal. (60) could confirm the grafting of n-alkanethiols on dimethylzinc-modifiedsilica surfaces, occurring through the formation of Zn-S chemical bonds, byobserving the absence in the 13C CP/MAS spectra of signals due to Zn-bondedmethyl carbons, as well as chemical shift and line broadening effects on signalsof the carbons in α and β positions from sulfur.

The presence in 13C CP/MAS spectra of a structured signal for the car-boxylic carbons was observed by Pawsey et al. (61) for self-assembled mono-layers of alkanoic acids on zirconium oxide (see Figure 7) and by Borsacchi

Figure 7. 13C CP/MAS spectrum of a 13C-labeled octadecanoate monolayer on zirco-nium oxide, CH3(CH2)16

13CO2/ZrO2. The signals assignment and the molecular struc-ture of octadecanoate are also reported. Asterisks denote spinning sidebands. Reusedwith permission from Pawsey et al. (61). Copyright (2000) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 20: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

18 M. Geppi et al.

et al. (39) for silica-coated BaSO4 submicron particles, modified on the surfaceby treatment with stearic acid. This was ascribed to a considerable hetero-geneity of the chemical environment experienced by the carboxylic head ofthe organic modifier, due to either the presence of different types of bond-ing sites on the inorganic surface or the establishment of different types ofinteraction.

A detailed 13C study concerning the interactions occurring between acetoneand silica gel surface was reported by Pan et al. (62); from a large bodyof data, also including relaxation measurements, the authors could validate amodel in which acetone was in a rapid equilibrium between a state in which itwas hydrogen-bonded with silica silanol groups and one in which it was justphysisorbed. By analyzing the chemical shift difference of the acetone carbonsignals between the values measured in silica and those in liquid acetone, as afunction of temperature and acetone loading level, the authors could determineKeq , H0 and S0 of the process.

As already described for 29Si, quantitative data concerning the organic-inorganic interface can be obtained also from 13C spectra. For example, Liet al. (30), by building up and analyzing 13C CP dynamics curves, obtaineddetailed quantitative data for sol-gel–derived hybrid materials based on tri-ethoxysilylated diethylenetriamine and tetramethoxysilane.

Other Nuclei

Even if to a minor extent, 1D high-resolution spectra of 31P, 19F, and 15N nucleihave been also exploited for investigating the organic-inorganic interface inOIMM. In the following, one example for each nucleus will be briefly described.

The observation of 31P is favored by its 100% isotopic natural abundanceand quite high receptivity; moreover, high-resolution 31P 1D spectra can beeasily obtained by combining HPD and MAS. Gao et al. (63) exploited 31PCP/MAS spectra for obtaining information concerning the interfacial interac-tions between ZrO2, TiO2, and zirconated silica powders and self-assembledoctadecylphosphonic acid monolayers adsorbed on them. In particular, the au-thors could relate the different 31P isotropic chemical shifts and line widths,observed in the presence of different metal oxide substrates, with differentinteractions occurring between the acid head group and the substrates.

19F has a 100% natural abundance and almost the same receptivity asproton. Due to the possible presence of 19F-19F homonuclear dipolar couplings,the obtainment of spectra with a sufficiently high resolution can require the useof a high MAS frequency. Zhang et al. (64) reported the 19F MAS spectra ofnanocomposites of the perfluorinated Nafion resin and an organically modifiedmontmorillonite (m-MMT), acquired at a 25 kHz MAS frequency and forsamples at different contents of m-MMT (see Figure 8). The authors observedthat the line width of the signals of the -CF3, -OCF2, and -SCF2 groups ofthe Nafion side chains increased with increasing the m-MMT content, while

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 21: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 19

Figure 8. 19F MAS spectra of (a) Nafion and (b)–(e) Nafion/organically modifiedmontmorillonite (m-MMT) nanocomposites with different m-MMT contents (b): 3 wt%,(c): 5 wt%, (d): 8 wt%, (e): 10 wt%). CF(S) and CF(B) stand for side chain and backboneCF group, respectively; (CF2)n indicates backbone CF2 groups; SCF2 indicates the CF2

group next to the SO−3 group. Reused from Zhang et al. (64). Copyright (2007), with

permission from Elsevier Inc.

this effect was not observed for the other signals. This line width increase wasascribed to a decrease in the molecular mobility of the Nafion side chains dueto the occurrence of interactions with m-MMT, also in agreement with 19Frelaxation times measurements.

The observation of 15N nuclei is dramatically hindered by their extremelylow natural abundance (0.368%), which determines very long and often unfea-sible experimental times for collecting a spectrum with a good signal-to-noiseratio. Shenderovich et al. (54) reported a study based on several SSNMR ex-periments, in which they characterized the hydrogen-bonding interactions ofpyridine with silanol groups of two ordered mesoporous silicas and used theNMR results for developing molecular models for the inner surface structureof the silicas. In particular, 15N spectra were used for obtaining direct infor-mation on the hydrogen bonds between the 15N isotopically enriched pyridineand the silica silanols, as well as on the reorientational properties of pyridine.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 22: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

20 M. Geppi et al.

Figure 9. 15N DE spectra acquired in both static and MAS conditions for 15N-pyridine inMCM-41 and SBA-15. Reused with permission from Shenderovich et al. (54). Copyright(2003) American Chemical Society.

In Figure 9 the 15N proton decoupled DE spectra acquired in both static andMAS conditions for pyridine in the two different silicas (MCM-41 and SBA-15,the former with smaller pore diameters than the latter) are shown. In both theMAS spectra a single narrow signal at 253–254 ppm was observed, assignedto pyridine hydrogen-bonded with silica silanol groups. The static spectra,mainly determined by the modulation of 15N chemical shift anisotropy per-formed by molecular motions, were very different for the two samples. In thecase of pyridine/MCM-41, a broad powder pattern was observed, while thestatic signal in pyridine/SBA-15 had an isotropic shape. From these resultsthe authors could infer that pyridine experienced fast anisotropic and isotropicreorientations in MCM-41 and SBA-15, respectively.

2D Spectra of Spin-1/2 Nuclei

Several 2D SSNMR techniques based on the observation of spin-1/2 nucleihave been applied to the study of organic-inorganic interfaces in OIMM. Ingeneral, 2D NMR spectra are obtained using pulse sequences in which thesignal is made dependent on two time variables, usually followed by doubleFourier transformation. The introduction of a second dimension is aimed atincreasing the information obtainable from the spectrum. One of the mostcommonly exploited interactions in 2D experiments applied in the investigationof organic-inorganic interfaces is the dipolar interaction, whose strength isproportional to the inverse of the sixth power of the distance between the

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 23: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 21

dipolarly coupled nuclei. Therefore, 2D SSNMR experiments exploiting homo-or heteronuclear dipolar interactions are very useful in providing structuralinformation, especially revealing pairs of nuclei in reciprocal spatial proximity.The 2D technique most used in the characterization of OIMM is HETCOR (65–68); in the corresponding 2D map the 1D high-resolution spectra of differentnuclei (e.g., 1H and 13C) are present in the two dimensions, and cross-peaksconnect the isotropic signals of the dipolarly coupled pairs. The most usedHETCOR experiments are based on dipolar interactions between protons andeither 13C or 29Si nuclei.

Comotti et al. (32) applied 1H-29Si and 1H-13C HETCOR experiments toinvestigate the interfacial interactions between a periodic mesoporous hybridp-phenylenesilica and the surfactant octadecyltrimethylammonium bromide(ODTMA) filling the silica nanochannels space. In Figure 10, the 1H-29Si and1H-13C 2D HETCOR spectra of the ODTMA-filled organosilica, in which thearomatic moieties were 2H-labeled, are reported. The spectra revealed corre-lations between the methyl protons of the surfactant head groups ((CH3)3N)

Figure 10. 1H-29Si and 1H-13C 2D HETCOR spectra of p-phenylene-d4-silica (host)/ODTMA (guest) nanocomposite. The structure of the nanocomposite interface, showingthe interaction between the p-phenylene-d4-silica building blocks and the polar heads ofthe surfactant molecules, is also reported. Reused with permission from Comotti et al.(32). Copyright (2007) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 24: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

22 M. Geppi et al.

and both the T 3 and T 2 silicon species, which constituted the silica frame-work, indicating that the condensed inorganic groups faced the channels andinteracted with the surfactant. On the contrary, no correlations were observedbetween the silicon species and the methylene protons of the long surfactantchains, indicating that they stayed apart from the silica surface. In the 1H-13C HETCOR spectrum, apart from the intramolecular correlations expectedbetween protons and carbons of the surfactant, an interesting cross-peak wasobserved between the methyl protons of the surfactant head groups and thearomatic carbons of the organosilica, indicating the occurrence of host-guestinteractions, mainly involving the surfactant polar heads. The authors proposedthat the bulky polar heads, interacting with silica, prevented the long surfac-tant hydrocarbon tails to access the interface and schematically represented thesilica-surfactant interface as reported in Figure 10.

A detailed characterization of the structure of the organic-inorganic inter-face in the mesoporous silica MCM-41 functionalized with different allylsi-lanes (69) and chloroalkylsilanes (70) was obtained by Pruski and coworkersby means of 1H-13C and 1H-29Si HETCOR experiments. In these cases, theauthors performed the experiments at a very high MAS frequency (40 kHz),which was sufficient to average out the proton homonuclear dipolar couplingswithout having recourse to the more commonly used multiple pulse homonu-clear decoupling sequences.

Rawal et al. (71) exploited 1H-31P HETCOR spectra for demonstrating theclose proximity of a significant phosphate fraction to polyamide in phosphate-glass/polyamide-6 hybrids. In the 2D HETCOR spectra the authors clearlydetected correlations between the 31P signal of the phosphate-glass and thealiphatic polyamide protons and also observed peculiar features of the signalof the interfacial phosphate groups with respect to those of the bulk.

Rottstegge et al. (72) reported an SSNMR study on cement pastestreated with several organic components aimed at influencing their hydrationand hardening. Beside other techniques, the authors applied two 1H-1H 2Dtechniques, i.e., 1H 2D double quantum (73–75) and 1H 2D (NOESY type)exchange (76) experiments, both based on the homonuclear dipolar couplinginteraction, very effective in revealing proton-proton spatial proximities. Boththese techniques exploit the dependence of the homonuclear dipolar couplingon the internuclear distance; the 2D double quantum experiment was usedfor revealing protons at short reciprocal distances of about 0.2–0.5 nm, while2D exchange was applied for detecting proton pairs at longer distances (about10 nm). By observing 1H-1H correlations between the cement protons andthose of methyl cellulose for one sample and those of poly(vinyl alcohol covinyl acetate) for another, the incorporation of either organic modifiers in theinorganic cement matrix was revealed.

1H 2D double quantum techniques were also applied by Arrachart et al.(77) in organosilicas obtained from silyl derivatives of nucleic acid base pairs,revealing the occurrence of adenine-thymine pairing.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 25: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 23

A technique that exploits the heteronuclear dipolar coupling for quantita-tively measuring internuclear distances is the rotational echo double resonance,REDOR, experiment (78). Jaeger et al. (79) reported an application of the 13C-31P REDOR experiment to the investigation of the organic-inorganic interfaceoccurring in bone, a natural complex composite material with an organic matrix,mainly collagen, and inorganic crystals (a complex calcium phosphate phasedominated by hydroxy-apatite) deposited on it. In particular, they observed thatthe strongest dipolar coupling occurred between inorganic phosphorous nucleiand glutamate carboxylate carbons; from the data analysis, a corresponding13C-31P distance of the order of 0.4–0.5 nm was derived. This result was inter-preted with the occurrence of a strong interaction between glutamate-containingproteins, present in the non-collageneous part of the organic matrix, and Ca2+

ions present on the surface of the inorganic phase.Several 2D experiments can also be modified in order to exploit the spin

diffusion process; even if spin diffusion will be more extensively describedin the last section of the review, here it is worth mentioning that it is animportant process in SSNMR, consisting in a spatial diffusion of nuclearmagnetization occurring without diffusion of matter, realized through a po-larization exchange between dipolarly coupled homonuclear spins. Hou et al.(80) reported a thorough SSNMR study in which, by means of several 2Dtechniques modified for taking into account spin diffusion, they investigatednanocomposites of poly(styrene-ethylene oxide) block copolymer (PS-b-PEO)and hectorite (HCT), a smectite clay mineral. Based on a preliminary moretechnical study (81), in which the SSNMR experiments were tested on thepristine clay and on nanocomposites with poly(ethylene oxide), the authorscould obtain detailed information on the extent of intercalation of the polymerin the clay by applying 1H-29Si and 1H-13C HETCOR, 1H-1H correlation, and1H-29Si WISE techniques, all modified by inserting in the pulse sequencesa time window during which proton spin diffusion was allowed to proceed.In the 1H-29Si HETCOR experiment, in absence of spin diffusion, only thecross-peaks connecting the signals of spatially very close 29Si and 1H nucleiare detected, while, with increasing spin diffusion mixing time, correlationsbetween pairs of nuclei at progressively increasing distances (approximativelyup to 10 nm) can be observed. In Figures 11a and 11b, the 1H-29Si HETCORspectra with 1 ms of 1H spin diffusion for two nanocomposites obtained withcopolymers differing for the molecular weight of the PS block (3.6 and 29.8kDa for PS4EO and PS30EO, respectively) are shown. Beside the expectedcross-peak correlating the clay OH protons signal (at 0.35 ppm) with that ofthe clay silicons, in both the spectra another cross-peak was observed, due tothe PEO protons at an approximate distance of 0.3–2 nm from the clay surface,much more evident for the PS4EO-HCT sample. In Figures 11c and d, the crosssections taken at the center of the Si peak in the 2D HETCOR spectra of the twonanocomposites, at different values of spin diffusion mixing time, are shown.In PS4EO-HCT spectra the PEO protons signal was observed at even very

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 26: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

24 M. Geppi et al.

Figure 11. 1H-29Si 2D HETCOR spectra with 1 ms of 1H spin diffusion of (a) PS30EO-HCT and (b) PS4EO-HCT. On the right of each spectrum the homonuclear decoupled1H spectrum along the ω1 dimension, taken at the center of the Si signal (resonating atzero frequency due to refocused on-resonance detection), as marked by the arrows, isreported. Cross sections taken at the center of the Si peak in the HETCOR spectra of(c) PS4EO-HCT, (d) PS30EO-HCT, at different values of spin diffusion mixing time.Reused with permission from Hou et al. (80). Copyright (2003) American ChemicalSociety.

short mixing times, while it clearly appeared only after 10 ms of mixing timefor PS30EO-HCT. The authors interpreted the fast spin diffusion occurring inPS4EO-HCT as indication of a significant PEO intercalation between the clayplatelets, which on the contrary seemed to take place to a much minor extent inPS30EO-HCT. Moreover, the fact that for both the samples no signals due to PS

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 27: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 25

protons (expected at around 7 ppm) were observed for mixing times less than30 ms indicated that the PS blocks were not intercalated. The authors couldconfirm these results by acquiring suitable 1H-1H 2D correlation spectra withspin diffusion, in which they could observe the signals arising from protonsinvolved in spin diffusion with the clay OH protons but with a higher sensi-tivity with respect to the HETCOR experiment. Moreover, the application of a1H-13C HETCOR experiment with spin diffusion allowed the clay intercalatedspecies to be identified by observing the 13C signals correlated with the OHproton ones. A correlation with the 13C PEO signal at 70 ppm was observedfor PS4EO-HCT, while no correlations with PS signals were detected. At last,information concerning the mobility of the intercalated polymer was obtainedby the comparison between 1H-29Si HETCOR and 1H-29Si WISE experiments,both with spin diffusion. In the 2D WIdeline SEparation (WISE) experiment(82) the proton signals obtained without homonuclear decoupling can be sep-arated on the basis of the isotropic chemical shift of the X nuclei to which theprotons are dipolarly coupled. Protons in rigid environments give rise to verybroad signals, sometimes unobservable, while molecular mobility can signif-icantly narrow the line width. For PS4EO-HCT the authors could observe, atthe same spin diffusion time (1 ms), a strong PEO proton peak in the HETCORspectrum, but not in the corresponding 2D WISE one, indicating a noticeablerigidity of the PEO chains intercalated in the clay.

Another example of application of 1H-29Si HETCOR with spin diffusion isthat reported by Bose et al. (83), who could characterize the interface betweensilica and a siloxane polymer (OV-225) physisorbed on it, revealing in particularthe occurrence of hydrogen-bonding between the silica surface and cyanogroups and oxygen atoms of the polymer.

Other 2D techniques including spin diffusion, more specifically applied toa quantitative analysis of the spin diffusion process, and hence to the determi-nation of domain sizes, will be described in the last section of the review.

Experiments on Quadrupolar Nuclei

Nuclei with spin I > 1/2, called quadrupolar nuclei, are characterized by anonspherical distribution of the nuclear charge. For these nuclei the NMRproperties in the solid state are strongly affected by the quadrupolar interactionbetween the nuclear quadrupole moment (peculiar property of the nucleus)and the electric field gradient (EFG) due to the electronic environment in thevicinity of the nucleus. With the important exception of 2H, some applicationsof which will be reviewed in the next two sections, most of the exploitedquadrupolar nuclei have a half-integer spin. For these nuclei, the “centraltransition” between the spin levels corresponding to 1/2 and −1/2 values of thequantum number mI , is usually the easiest to be observed, and it is then usually

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 28: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

26 M. Geppi et al.

selectively excited in acquiring an NMR spectrum.4 The central transition is notaffected by the quadrupolar interaction at the first order, but it can be stronglyinfluenced by second-order quadrupolar effects. Static lines for quadrupolarnuclei can be very broad, and MAS can drastically improve spectral resolution.However, it must be noticed that second-order quadrupolar effects cannotbe completely averaged out by MAS, and therefore the MAS signals ofquadrupolar nuclei can still show a significant line width and a peculiar shape.From the analysis of either the static or the MAS line shape, chemical shift,and quadrupolar interaction parameters, useful for characterizing the particularnuclear site under investigation, can be obtained. The quadrupolar interactionis characterized by two parameters, Cq (quadrupolar coupling constant) andη (asymmetry), very sensitive to the local environment of the quadrupolarnucleus and, in particular, to the molecular geometry.

In the following, some examples of studies of quadrupolar nuclei forthe characterization of organic-inorganic interfaces in OIMM are reported,organized on the basis of the type of nucleus. Some relevant properties of thenuclei considered in the following are reported in Table 2.

Aluminum-27

27Al is one of the most present and studied quadrupolar nuclei in OIMM. It isa spin-5/2 nucleus with a natural abundance of 100% (see Table 2).

In general, 27Al MAS spectra recorded at standard spinning frequenciesand magnetic fields allow signals from differently coordinated aluminum sitesto be resolved. An example is shown in Templin et al. (46) (see Figure 12),who could distinguish tetrahedrally, octahedrally, and fivefold coordinated alu-minum sites present in hybrids obtained by aluminum sec-butoxide and (3-glycidyloxypropyl) trimethoxysilane through sol-gel process.

Haouas et al. (84) exploited 27Al MAS spectra to investigate the interactionbetween the organic polymer and the clay in poly(ε-caprolactone)/maghnitenanocomposites. From the analysis of 27Al MAS spectra the authors coulddistinguish three signals assigned to two different types of tetrahedral andone octahedral aluminum sites and observed that the incorporation of the clayinto the nanocomposites led to a decrease in the population of one type oftetrahedral site and a corresponding increase in that of the octahedral ones.This was ascribed to a selective grafting of the organic polymer onto thosetetrahedral aluminum sites, which then transformed in hexa-coordinated ones.

Bertmer et al. (85) used 27Al for investigating the possible interactionsoccurring at the interface between the organic polymer and the clay plateletsin nylon-6/montmorillonite nanocomposites. In both the 27Al MAS spectra of

4All the other single-quantum transitions but the central one (satellite transitions)usually give rise to very broad spectral patterns and are difficult to be observed, apartfrom a few cases. In the following discussion we will not consider satellite transitions.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 29: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 27

Table 2. Some relevant properties of quadrupolar nuclei, taken from official IUPACrecommendations (© 2001 IUPAC) by Harris et al. (181)

Natural Frequency Quadrupole momentIsotope Spin abundance (x/%) ratioa (�/%) (Q/fm2)

27Al 5/2 100 26.056859 14.6617O 5/2 0.038 13.556457 −2.55891Zr 5/2 11.22 9.296298 −17.623Na 3/2 100 26.451900 10.47Li 3/2 92.41 38.863797 −4.012H 1 0.0115 15.350609 0.2860

aSee definition given in Table 1.

pure montmorillonite and of a nanocomposite with nylon-6 two signals wereobserved: a very intense one, at around 0 ppm, assigned to the hexa-coordinatedaluminum sites of the central clay layer, and a weaker one, at 67 ppm, assignedto the tetra-coordinated aluminum sites present on the superficial layers. A smallshoulder was observed at about 50 ppm in the spectrum of the nanocomposite,ascribed to a second type of tetra-coordinated Al site, possibly interactingwith the organic polymer. More detailed information could be obtained by theacquisition of 27Al 2D MQMAS spectra (86, 87) of both the pure clay andthe nanocomposite, shown in Figure 13. In an MQMAS spectrum the firstdimension (f2) is given by the 1D MAS spectrum (containing both isotropicchemical shift and second order quadrupolar effects), while in the seconddimension (f1) a purely isotropic spectrum is obtained (without contributionsfrom quadrupolar interactions). For each signal, the combined analysis of the f1

and f2 frequencies allows the isotropic chemical shift and the SOQE (second-order quadrupolar effect) parameter (a combination of quadrupolar couplingconstant and asymmetry parameter) to be determined. The authors could clearlydetect three separate signals in the MQMAS spectrum of the nanocomposite,also determining both isotropic chemical shift and SOQE for each of them, thusconfirming the presence of a second tetrahedral site ascribable to the interactionwith the polymer.

Also for quadrupolar nuclei, the dipolar interactions can be exploited forinvestigating spatial proximities among nuclei at the interface. For example,Janicke et al. (88) applied the 1H-27Al HETCOR experiment in the characteri-zation of an aluminosilicate MCM-41 containing the surfactant cetyltrimethy-lammonium bromide as structure directing agent. By the analysis of cross-peaks, they could detect the occurrence of a strong interaction between thetetrahedrally coordinated Al sites of the aluminosilicate and the protons of thetrimethylammonium head-group moieties of the surfactant.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 30: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

28 M. Geppi et al.

Figure 12. 27Al MAS spectrum of an hybrid with a (3-glycidyloxypropyl) trimethoxysi-lane to aluminum sec-butoxide ratio of 90:10 mol%. The signals assignment to differ-ently coordinated aluminum sites is shown. Reproduced with permission from Templinet al. (46). Copyright (1997), Wiley-VCH Verlag GmbH & Co. KGaA.

Oxygen-17

17O is a spin-5/2 nucleus with a natural abundance of 0.038%. Oxygen wouldbe a very important molecular probe in investigating structural properties ofOIMM; unfortunately, the very low natural abundance of its NMR-active iso-tope hampers its extensive application, often making necessary a preliminaryisotopic enrichment.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 31: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 29

Figure 13. 27Al 2D MQMAS spectra of (a) pure montmorillonite and (b) a nanocom-posite with nylon-6. Reprinted with permission from Bertmer et al. (85). Copyright(2007) American Chemical Society.

Lafond et al. (89), by the analysis of 17O MAS spectra of selectivelyisotopically enriched titanium oxide/phenylphosphonate hybrids and relatedmodel compounds, could distinguish and characterize the chemical shift andquadrupolar parameters of the different oxygen nuclei present and determinethat the binding mode of phosphonate groups to titania was mainly tridentate.

Moreover, Julian et al. (90) reported a 17O SSNMR characterization of aseries of polydimethylsiloxane-metal oxide (PDMS-MxOy) OIMM in which17O signals due to oxygen atoms bonded to a variety of metals (M = Ge, Ti,Zr, Nb, Ta) were detected and assigned.

Zirconium-91

91Zr is a spin-5/2 nucleus with a natural abundance of 11.22%. Armelao et al.(45) reported the 91Zr 1D static spectrum of the already mentioned hybridscontaining Zr-oxoclusters. Such a spectrum was recorded using the QCPMGpulse sequence (91, 92), which, for each nuclear site, produces a series ofspikelets that mimic the powder pattern of the stationary sample. This method istypically used to increase the sensitivity of quadrupolar nuclei by concentratingin the spikelets the signal intensity, which otherwise would be spread over abroad spectral region. From the bad signal-to-noise ratio obtained, and fromthe comparison with the spectrum of ZrO2, the authors concluded that only thesample crystalline fraction was observed, and that in this fraction zirconiumwas in a low-symmetry tetragonal environment.

Sodium-23

23Na is a spin-3/2 nucleus with a 100% natural abundance. As in the case of7Li, most SSNMR studies on 23Na at the organic-inorganic interface in OIMM

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 32: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

30 M. Geppi et al.

concern solid polymer electrolytes (SPEs), which are formed by either complex-ing or dissolving alkali metal salts in ion-conducting/coordinating polymers.

Bartolotta et al. (93) used 23Na techniques to study sodium thio-cyanate/poly(ethylene oxide) electrolytes ((PEO)nNaSCN, where n is the num-ber of oxygen atoms per atom of sodium) at different values of n. The dissolutionof Na+ cations in the PEO amorphous phase or their presence in the crystallinecomplex (PEO)3NaSCN could be investigated by means of 1D MAS spectra,measurement of spin-lattice relaxation times, and 2D nutation spectra (94, 95).In particular, in 2D nutation spectra, the 1D MAS spectrum is present in thefirst dimension, while in the second dimension signals can be present at thepulse radiofrequency (ωrf) or at 2ωrf , depending if the quadrupolar interactionis smaller or larger than ωrf . In Figure 14, 1D MAS and 2D nutation spectraof some investigated samples are reported. At low salt concentrations (n = 49)sodium dissolved only in the amorphous regions of PEO, as indicated by boththe 1D MAS line shape, and the presence of the sole signal at ωrf in the 2D nu-tation spectrum, in agreement with a strongly reduced quadrupolar interactiondue to fast motions occurring in amorphous PEO. On the contrary, the pres-ence of sodium cations in a crystalline phase could be detected by the typical

Figure 14. (a)–(c) 23Na MAS spectra of (PEO)49NaSCN (a), (PEO)19NaSCN (b),and (PEO)3NaSCN (c); (d)–(g) 23Na 2D nutation spectra of (PEO)49NaSCN (d),(PEO)19NaSCN (e), (PEO)4.5NaSCN (f), and (PEO)3NaSCN (g). In the nutation ex-periments, ωrf = 75 kHz. Reused from Bartolotta et al. (93). Copyright (1997), withpermission from Elsevier B.V.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 33: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 31

quadrupolar line shape in 1D MAS spectra for n = 19 and n = 3. For n = 3sodium cations resulted to be present in the sole crystalline phase, as observedfrom the 2D nutation spectrum (signal at 2ωrf only), in agreement with thesupposed stoichiometry of the PEO/NaSCN crystalline complex. A distribu-tion of sodium cations between the crystalline phase and amorphous PEO wasinstead observed for 3 < n < 49 (presence of both ωrf and 2ωrf signals in the2D nutation spectra).

A very detailed study of the dynamics of sodium cations of NaClO4 andNaCF3SO3 salts complexed in amorphous poly(propylene oxide) (PPO) was re-ported by Chung et al. (96). The authors quantitatively analyzed and interpretedT2 and T1 relaxation data as a function of temperature, sodium concentration,and polymer molecular weight, obtaining information about the hopping mo-tion of sodium cations.

Lithium-7

7Li is a spin-3/2 nucleus with a 92.41% natural abundance. With respect tothe quadrupolar nuclei discussed so far, it is characterized by much smallerquadrupolar interactions. Similarly to 23Na, 7Li has been extensively employedin the characterization of SPEs. The study of dynamic behavior of the mobileLi cations in SPEs is crucial to understand the conductivity properties of thematerials, and can be usually performed by means of different 7Li NMR ap-proaches, for instance by analyzing the trends of line width or relaxation timesvs. temperature.

In a series of recent papers, Kao et al. (97–99) and Bonagamba and cowork-ers (100, 101) investigated solid organic-inorganic hybrid electrolytes calledormolytes (organically modified electrolytes), obtained via in situ formation ofthe inorganic component within the polymer matrix. Depending on the condi-tions, the organic and inorganic components can be either merely intimatelymixed (type-I hybrids) or chemically bonded (type-II hybrids).

Kao et al. (98) studied type-II ormolytes based on poly(propylene glycol)-block-poly(ethylene glycol)-block-poly(propylene glycol) bis(2-aminopropylether) complexed with LiClO4 via the co-condensation of an epoxy trialkoxysi-lane and tetraethoxysilane by means of variable temperature 7Li experiments.Two different 7Li chemical environments were detected at low temperaturesfrom the presence of two partially overlapped peaks in 1H-decoupled 7Li MASspectra (see Figure 15). They were ascribed by the authors to lithium cations inthe polyether domains and at the polymer/silica interface or located in silica-rich domains. With increasing the temperature the two peaks merged into aunique, much narrower peak. The 7Li static line width resulted to be mainlydetermined by residual dipolar couplings with surrounding protons. Its temper-ature dependence revealed a rapid transition between two different motionalregimes, occurring just above the polymer glass transition (see Figure 15),indicating that Li ions could experience a fast dynamic situation (motional

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 34: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

32 M. Geppi et al.

Figure 15. (Left) Variable-temperature 1H-decoupled 7Li MAS spectra of a type-IIormolyte: two different peaks, indicated with I and II, are clearly distinguished atlow temperature. (Right) 7Li static line widths measured at different temperatures (a)without and (b) with proton decoupling for two type-II ormolytes (A-B) characterizedby a different alkoxysilane/polymer ratio. Reused with permission from Kao et al. (98).Copyright (2006) American Chemical Society.

frequencies larger than about 6 kHz) only in the presence of an activated seg-mental motion of the polymer chain, regardless of the alkoxysilane/polymerratio. To get information about the dynamics of Li ions in the nanosecondstimescale, the authors also performed 7Li T1 measurements. In the whole tem-perature range explored, corresponding to the slowest side of the T1 curve,the sample with the lowest Li content exhibited the shortest relaxation times,indicating a faster dynamics, in agreement with conductivity measurements.5

The authors also applied the pulsed gradient spin echo (PGSE) technique (102)to directly measure self-diffusion coefficients of Li ions as a function of tem-perature, which allowed a clear correlation between the mobility of the Li ionsand the ionic conductivity to be established; a higher diffusivity was measured

5 7Li T1 in ormolytes will be further described in the next section, in a synergicanalysis with 13C T1.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 35: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 33

for the samples with the lower Li content, in agreement with 7Li line width andT1 measurement results.

LOW-MW ORGANIC COMPONENTS

Low-molecular weight (low-MW) organic species play a functional role inOIMM, where they can interact with an inorganic matrix being either includedin inorganic pores, intercalated between inorganic layers, or chemically bondedto inorganic surfaces or networks. The role of the low-MW organic speciescan be the modification of the properties of the inorganic matrix, such as inorganically modified layered silicates, where they can confer an organophiliccharacter to the clays and/or act as compatibilizers with organic polymersfor the preparation of micro- or nanocomposites. In other cases, the mainactive component of the OIMM is indeed the low-MW organic species, as,for instance, in chromatographic stationary phases, where usually the organiccomponent, chemically bonded to an inorganic support, is responsible for theseparation process. A somehow peculiar class of OIMM is represented byzeolites or other porous materials either in the presence of organic structure-directing agents, or in which small organic molecules, typically benzenoids,have been adsorbed. These systems are widely studied as intermediates intemplate syntheses or in catalytic organic reactions, as well as to investigatethe confinement effects on organic molecules.

For these reasons, the characterization of the structural and dynamic be-havior of low-MW organic species in OIMM is particularly important, andthe relevance of SSNMR in this field is demonstrated by the large number ofstudies reported in the literature. In the following, examples of these studieswill be presented considering five categories of experiments, namely: 1D high-resolution spectra, CP dynamics experiments, 2H static spectra, relaxation timemeasurements, and 13C 2D exchange techniques.

1D High-Resolution Spectra

A significant portion of the SSNMR studies reported in the literature on thecharacterization of low-MW organic components in OIMM makes use of 1Dspectra of rare nuclei acquired in high-resolution conditions. As already pointedout in the previous section, the NMR observable exploited in these studiesis the isotropic chemical shift. Its dependence on conformational properties,phase structure, and dynamic behavior of the organic molecules is particularlyimportant for the characterization of the organic components. In the majorityof cases, 13C 1D DE or CP/MAS spectra on natural abundance samples areexploited, although it is not rare to run into studies in which 1D high-resolution

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 36: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

34 M. Geppi et al.

spectra of other nuclei, such as 15N, 19F, or 31P, are used. Moreover, despite thealready pointed out difficulties in the achievement of high-resolution spectraof abundant nuclei, many studies have been reported in the literature includingthe use of 1H 1D MAS spectra. In the following, examples of applications of1D spectra to the characterization of low-MW organic components in OIMMare reported, organized on the basis of the observed nucleus.

Carbon-13

The study of the low-MW organic component in OIMM by means of 13C 1Dspectra is particularly relevant in the case of two classes of systems, namelychromatographic stationary phases and clays exchanged with alkylic-chaincations. In both cases, the organic component is often a linear Cn chain withn ≤ 30, possibly functionalized at one edge with a silanic, ammonium, or phos-phate group. In polyethylene (PE), two signals are usually detectable: one atabout 31 ppm, corresponding to chain segments experiencing fast interconfor-mational jumps between gauche (G) and trans (T ) conformations, and the otherat about 33 ppm, due to the rigid all-trans arrangement. This can be largelyextended to sufficiently long linear alkylic chains of low-MW compounds.

Many significant studies exploiting SSNMR for the characterization ofinverse chromatographic phases are those of Albert, Pursch, and coworkers. Inmost of the stationary phases investigated, the organic selectors, constituted bya linear alkylic chain Cn, were chemically bonded to silica particles acting assupport. In trying to understand the mechanisms of the chromatographic sep-aration, studies were performed at different bonding density, chain length andtemperature. Pursch et al. (42) observed the occurrence of the all-trans arrange-ment in C18 phases, previously observed only for longer chains (see Figure 16).Pursch et al. (43) also observed that in C18 phases the CH2 signal moved towardhigher frequencies with increasing the bonding density, corresponding to anincreasing fraction of T conformations. In the attempt of evaluating the effectsof alkylic chain length on the behavior of Cn phases (41, 103), Pursch et al.showed that the T /G ratio, measured from the intensity of the correspondingsignals, increased with increasing chain length, suggesting the presence ofa higher order degree for longer chains in the presence of similar bondingdensities. The behavior of the bonded phase with temperature has also beenobject of investigation, since it is well known that a strong dependence of shapeselectivity on column temperature exists. In this regard, Pursch et al. reportedseveral studies (see for instance references 21, 41, and 104), which demon-strated that T conformers mainly existed at low temperatures, while the fractionof G conformations increased with increasing temperature. These studies alsoshowed that the chain order associated with the presence of T conformationsresisted in a wider temperature range for phases with longer alkylic chains.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 37: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 35

Figure 16. 13C CP/MAS spectrum of a C18 chromatographic stationary phase. Thelabeling refers to the structure reported in the inset. Reused with permission fromPursch et al. (42). Copyright (1996) American Chemical Society.

In a similar way, 13C 1D spectra have been extensively and successfullyemployed to gain insights into the conformational properties of surfactantsintercalated in clays and, in particular, to study their dependence on temperature,surfactant properties (number and length of the alkyl chains), as well as on clayfeatures (platelets size, interlayer charge, distance, etc.).

Borsacchi et al. (38) investigated the behavior of a double-chain surfactant(2C18) intercalated in the synthetic layered silicate laponite, before and after thegrafting of an organosilane (3-(trimethoxysilyl)propyl methacrylate; TSPM)onto the clay platelets. The authors observed, in passing from laponite-2C18 tothe TSPM-grafted sample, an increase in the ratio between the 13C signals at31 ppm (due to surfactant chains experiencing fast G-T inter-conformationaljumps) and 33 ppm (ascribed to the ordered all-trans chains). By combiningthese results with those of 1H and 29Si SSNMR and X-ray diffraction, theauthors could conclude that TSPM produced a quite disordered arrangement ofthe clay disks, leading a noticeable fraction of surfactant molecules to be lessconstrained in the inter-platelet spaces.

Khimyak and Klinowski (105) studied the conformational properties of theC16-trimethylammonium (C16-TMA) surfactant occluded in several hexagonal(Hex) or lamellar (L) aluminophosphates by means of variable-temperature13C DE/MAS spectra. Monitoring the conformational changes occurring inthe occluded surfactant with temperature, the authors defined as transition

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 38: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

36 M. Geppi et al.

temperature that corresponding to the 1:1 ratio between the intensities of thesignals at about 33 and 31 ppm. 13C DE/MAS spectra recorded as a function oftemperature for the surfactant occluded in the lamellar phase L3 are shown inFigure 17 as an example. A higher transition temperature was observed for the

Figure 17. 13C DE/MAS spectra of C16-TMA cations in the aluminophosphate lamellarphase L3 measured at different temperatures. The spectra show the occurrence of a T ⇔G transition at ca. 40◦C (see text for details). Reproduced from Khimyak and Klinowski(105) by permission of the PCCP Owner Societies (http://dx.doi.org/10.1039/b007473j).

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 39: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 37

lamellar phases L1 and L3 with respect to two Hex phases, therefore associatedto a lower mobility of the aliphatic chains.

Examples of similar studies have also been reported by Muller et al.(106) on surfactants in poly(ε-caprolactone)/clay nanocomposites; by Kubieset al. (107) on surfactant molecules intercalated in laponite; by Grandjeanet al. (108), Wen et al. (109), and Zhu et al. (110) on surfactants inter-calated in montmorillonites; by Osman et al. (111) on dialkyldimethylam-monium cations electrostatically bound to mica platelets; and by Simonuttiet al. (112) on surfactant molecules adsorbed on MCM-41 mesoporoussilicas.

Beside the studies described so far, 13C 1D spectra have been exploitedfor the characterization of the dynamic properties of guest molecules adsorbedon zeolites or other inorganic frameworks. Similarly to a previous study byAzaıs et al. (113), Tang et al. (114) exploited the changes observed in 13C MASspectra of ibuprofen molecules confined in titania nanotubes (TiNT) acquired atdifferent temperatures to study the correlation between the molecular dynamicsand melting transition of the guest molecule with respect to the bulk. Theoccurrence of melting in both bulk and confined ibuprofen was revealed bysudden changes of both DE and CP/MAS spectral features (see Figure 18). Themelting temperature was observed to decrease by 12 K in passing from the bulkto the TiNT-confined ibuprofen. Moreover, contrary to the bulk ibuprofen, theCP spectrum of the confined ibuprofen did not completely disappear above themelting point: a set of weaker peaks rather emerged, suggesting that the confinedmaterial could retain small effective 1H-13C dipolar couplings, possibly due togeometrical constraints arising from the nanotube structure. The occurrenceof specific motional processes in ibuprofen could also be monitored from13C CP/MAS spectra. The coalescence of the four peaks at about 130 ppm,associated to the non-quaternary aromatic carbons in solid ibuprofen (115),indicative of the activation of the ibuprofen ring-flip motion about its para-axis,occurred well below the melting point in both the bulk and confined ibuprofen,even though it was slightly faster in the latter. On the contrary, the activation ofthe rotation of the isobuthyl group about its axis (as seen from the coalescence ofthe two methyl peaks at about 23 ppm) was concurrent with the phase transitionin both the cases, thus representing a signature of the ibuprofen melting.

Other examples of studies on the low-MW organic components in OIMM,based on 13C 1D high-resolution spectra, are reported in the following.

Gambogi and Blum (116), from the mere observation of differences inthe number or the chemical shift of the peaks, could confirm the bonding ofcoupling agents onto silicas.

Marini et al. (24), from quantitative 13C DE/MAS spectra, obtained infor-mation about the conformational properties of the organic component and itsdistribution in the core and shell of PE-PEG/silica nanoparticles.

Thangaraj et al. (117) could detect a restricted mobility of pyrrolidinemolecules occluded in zeolites by observing linebroadening effects.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 40: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

38 M. Geppi et al.

Figure 18. 13C MAS spectra of ibuprofen bulk (a)–(b) and confined in titania nanotubes(c)–(d) acquired from 297 to 355 K by means of CP (a), (c) and DE (b), (d) experiments;the relaxation time used in the DE experiments was chosen to saturate signals fromsolid ibuprofen. Reused from Tang et al. (114). Copyright (2008), with permission fromElsevier B.V.

Satozawa et al. (118) characterized the dynamics of benzene and p-xyleneadsorbed on zeolites and the mesoporous material FSM-16, on the basis of thesecond moments evaluated from the spinning sidebands intensities.

In some cases, 13C chemical shifts of low-MW molecules in OIMM havebeen studied using a computational approach. For instance, Simperler et al.(119) studied the adsorbtion of toluene molecules on zeolites, while Zhenget al. (120) predicted the chemical shift values of various molecules adsorbedon H-ZSM-5 zeolite.

Other Rare Nuclei

Also the study of rare nuclei other than 13C has been exploited to obtain indirectinformation about the location of the organic molecules inside the inorganicmatrix.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 41: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 39

Gunther et al. (121) reported a study in which 13C and 19F MAS spectrawere used to monitor the degree of absorption of benzenoid aromatic com-pounds on porous silica by observing, for the methyl groups or fluorine sub-stituents signals, small chemical shift differences between the microcrystallinebulk material and the adsorbed species. Similarly, O’Brien et al. (122) charac-terized the binding of triethylphosphate (TEP) to montmorillonite via 13C and31P MAS spectra. From 31P DE and CP/MAS spectra, they were able to revealthe presence of two different situations for the TEP inside the clay, correspond-ing to two distinct signals. On one hand, the DE spectra acquired in quantitativeconditions showed that the high-frequency signal, absent in the CP spectrum,increased linearly with TEP loading; on the other hand, the low-frequency peakhad nearly constant intensity at all loadings in both the spectra, thus suggestingthat it corresponded to a bound site.

To model the process of adsorption/desorption of an organic pollutanton the surface of soils, Ukrainczyk and Smith (123) used 15N-isotopicallyenriched pyridine adsorbed onto K-, Ca-, Mg-, and Al-hectorite. Intensitiesand/or chemical shift differences between 15N DE/MAS and CP/MAS spectraof pyridine adsorbed onto the different hectorites were interpreted in terms ofdistinct locations of the organic molecule within the clay galleries and revealedthe presence of hydrogen bonding involving nitrogen as the main bindingmechanism.

Hydrogen-1

Even though, as previously stated, the obtainment of high-resolution 1H spectrais often prevented by the presence of strong homonuclear dipolar couplings, inthe case of low-MW organic molecules 1H MAS spectra can show improvedresolution due to the intrinsically quite simple molecular structure and can bequalitatively employed for specific purposes. In particular, in a number of stud-ies the 1H spectral line width in MAS spectra has been used to get informationon the dynamic behavior of low-MW organic components in OIMM.

In several studies on chromatographic phases, 1H MAS spectra were usedto support the results obtained from 13C experiments (103, 124). In the inves-tigation of Cn chromatographic phases with different chain lengths and similarsurface coverage, Pursch et al. (41) observed a general increase in the 1H linewidth with increasing chain length, ascribable to a progressive decrease ofchain mobility (see Figure 19). Pursch et al. (43) also compared the 1H MASspectra of several C18 phases with different surface coverage, observing a linearrelationship between the 1H line width and bonding density. Differences in 1Hline widths were also observed by the authors for two C18 phases with similarcoverage but different chromatographic performances; the most efficient phaseshowed the larger 1H line width, indicating a more rigid environment and re-sulting in stronger dipolar couplings: this was interpreted as different ligand

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 42: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

40 M. Geppi et al.

Figure 19. High-speed 1H MAS spectra of alkyl-bonded phases with chain lengthsranging from C18 to C34 and similar bonding densities (indicated above each spectrum).Reprinted from Pursch et al. (41). Copyright (1997), with permission from Elsevier B.V.

spacing on a local level, possibly justifying the observed chromatographicbehavior.

In a study by Borsacchi et al. (125), two diastereoisomeric chiral chro-matographic phases with a complex chemical structure were investigated. Inthis case, differences in 1H spectral features, combined with 1H FID analysis

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 43: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 41

results, were interpreted in terms of different structural arrangement of thetwo diastereoisomeric phases with respect to the silica surface and tentativelyrelated with their different chromatographic performances.

CP Dynamics

As already discussed in the previous section, variable contact time CP experi-ments allow the extraction of 1H T1ρ and TXH parameters. These are potentiallyinformative about molecular dynamics as well as structural features of low-MWorganic components in OIMM. TXH depends on both internuclear distances andmolecular dynamics, in particular monotonically depending on motional rates(the faster the motions the higher TXH ). 1H T1ρ exhibits a more complex de-pendence on motional rates; moreover, its use for getting dynamic informationis often prevented or drastically limited by the effects of spin diffusion. Forthese reasons, TXH is most exploited in the majority of CP dynamics studiesreported in the literature.

Mielczarski et al. (126) studied the behavior of oleate molecules adsorbedon apatite, comparing the results obtained from CP dynamics experimentson samples with different oleate adsorption densities. Composites with lowadsorption density showed longer TCH values than in bulk oleate samples, sug-gesting a greater mobility for the adsorbed molecules. Except for the carboxylcarbon, all the different parts of the chain had the same mobility, indicating thepresence of an overall motion of the chain, contrary to what was observed inbulk oleate samples, characterized by a gradient of mobility along the chain.With increasing the loading up to the value corresponding to the statisticalmonolayer, the TCH values decreased, indicating a substantial stiffening of thechains, which was particularly pronounced for the carboxyl terminal fragmentsand the C=C groups. On the basis of these observations the authors proposedthe presence of a brush-like structure of the adsorbed layer with some lateralinteractions between parts of the chain.

Recently, Fyfe et al. (127) developed a complex method for determiningthe three-dimensional structures of sorbed small organic molecules in molec-ular sieve frameworks based on the analysis of the cross polarization ratesbetween 1H nuclei located on the guest molecule and 29Si nuclei. To validatethe method, the authors chose a model system constituted by the orthorombicform of zeolite ZSM-5 containing 8 molecules of p-xylene per unit cell, forwhich accurate internuclear distances were available from X-ray data. TSiH val-ues were measured for the different 29Si-1H pairs, exploiting the 29Si spectralresolution. A linear relationship between 1/TSiH and the heteronuclear secondmoment, calculated from the known structure and taking into account inde-pendently characterized motional processes, was found (see Figure 20). Thislinear relationship was proposed by the authors as the main criterion for find-ing unknown structures. This protocol was applied to determine the location of

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 44: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

42 M. Geppi et al.

p-xylene molecules inside ZSM in a low-loaded form. In subsequent extensionsof this work, Fyfe et al. investigated low-loaded p-dichlorobenzene/ZSM-5(128) and o-xylene/ZSM-5 (129) complexes, taking into account the issuesconnected with the experimental method for the measurement of TSiH and theuse of the algorithm for determining the sorbate molecule location. In partic-ular, noticing that a large number of solutions were usually obtained in theapplication of this protocol, the authors extensively discussed the criteria usedin the selection of the acceptable solutions, stating that the method was rel-atively independent of the exact atomic coordinates of the zeolite frameworkand appeared to be relatively insensitive to well-defined motions of the guestmolecules.

Figure 20. (a) Intensities of the 29Si CP/MAS signals for ZSM-5 loaded with eightmolecules of p-xylene-d6 per unit crystal at 273 K as functions of the contact time,along with fitting curves to extract TSiH values; (b) plot of experimental 1/TSiH valuesvs. the calculated second moments. Reprinted with permission from Fyfe et al. (127).Copyright (2005) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 45: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 43

2H Static Spectra

Deuterium is a spin-1 nucleus with a very low natural abundance (0.015%;see Table 2). Even though its use almost always requires a selective isotopicenrichment, it is a very exploited nucleus, since it possesses a small quadrupolecoupling constant (usually between 150 and 200 kHz), rendering possible theobservation of the whole powder spectrum by resorting to standard quadrupoleecho techniques on average magnetic fields. Contrary to what was previouslyseen for half-spin quadrupole nuclei, for which the transition commonly ob-served in the spectrum is the central one, in this case the spectrum is givenby the superposition of the signals arising from the two transitions 1 → 0 and0 → −1. The 2H static line shape is very sensitive to the presence and type (rateand geometry) of dynamic processes, occurring in the “intermediate” regime(characteristic frequencies in the range 103–106 Hz), involving the deuteratedmolecular fragment. In fact, in the absence of motions or for very slow ones,the 2H line shape is a typical Pake doublet; when isotropic motions with acharacteristic frequency much higher than the static line width are present,the shape reduces to a single peak; eventually, the presence of motions in theintermediate regime and/or with an anisotropic character causes distortions inthe static Pake doublet that reflect the characteristics of the motion. By suitablysimulating the observed pattern it is possible to detect very fine details aboutthe dynamic processes, as distinguishing between site-jump or diffusional mo-tions, revealing the presence of different dynamic processes and determiningprecise motional rates.

A large part of the literature concerning the use of 2H line shape anal-ysis in the characterization of the low-MW organic component in OIMMhas been devoted to the detection of the dynamic behavior of small guestmolecules (especially, benzenoids) adsorbed on zeolites or porous silicas;examples have been reported by Komori and Hayashi on p-nitroaniline (p-NA) adsorbed onto different functionalized mesoporous silicas (130) and ze-olites (131), by Shenderovich et al. (54) on pyridine absorbed onto variousmesoporous silicas, by Kustanovich et al. (132) on p-xylene absorbed ontozeolites, and by Sato et al. (133) on benzene, cyclohexane, and n-hexanein zeolites. Nonetheless, examples of studies reporting the dynamic inves-tigation of selectively labeled chains of surfactants in inorganic phases orchromatographic organic selectors, as well as coupling or structure-directingagents, have also been reported by Soderlind and Blum (134), Keluskiand Fyfe (135), Gambogi and Blum (116), and Shantz and Lobo (136),respectively.

Komori and Hayashi (131) investigated the dynamics of p-nitroaniline(p-NA) in zeolite ZSM-5 including Na+ (NaZSM-5). The authors comparedthe behavior of p-NA in the micropores of the hydrated and dehydratedzeolite by means of 2H line shape analysis of selectively deuterated p-NA

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 46: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

44 M. Geppi et al.

(p-NA-d) as a function of temperature, in order to evaluate possible effects dueto charge, pore dimensions, and water. Variable-temperature spectra recordedfor the dehydrated sample, together with an example of line shape simulation,are reported in Figure 21. At 149 K the spectrum of the dehydrated sample isconstituted by a typical static Pake doublet pattern, unaveraged by motions. Onthe other hand, at 380 K the 2H line shape could be reproduced by consideringa single, fast 180◦ flip-flop ring motion. At intermediate temperatures, two si-multaneous 180◦ flip-flop processes, occurring at different frequencies, shouldbe taken into account to give a satisfactory reproduction of the experimentalspectra. These two processes were ascribed to p-NA either in the microp-ores of NaZSM-5 (slower motion), or on the outer surface (faster motion), bycomparison with previous results on ZSM-5 by the same authors (137). Theanisotropic character of the motion on the outer surface suggested that p-NAshould interact with specific adsorption sites such as Na+. For the hydratedsample a different powder pattern was observed, ascribed to p-NA moleculesundergoing slow motions inside the micropores and fast isotropic motions onthe outer zeolite surface.

Figure 21. (Right) Temperature dependence of 2H static spectra of dehydrated NaZSM-5/p-NA-d . (Left) Comparison between the experimental and calculated 2H spectrumof dehydrated NaZSM/p-NA-d at 299 K: the calculated spectra corresponding to thetwo components of the 180◦ flip-flop motion used in the simulation are also shown.Reused from Komori and Hayashi (131) by permission of the Chemical Society ofJapan, Copyright (2004).

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 47: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 45

Zeigler and Maciel (138) used variable temperature 2H line shape analysisto investigate the dynamics of C18 chains bound to the surface of silica in atypical chromatographic stationary phase. 2H static spectra were acquired fordifferently deuterated samples, which could be reproduced only consideringdifferent dynamic environments at the silica surface. Constrained motions weredetected for a low-loading sample, suggesting that an interaction between theC18 chains and the silica surface occurred. As the density of the organic chainson the surface increased, the motion became less constrained because of thedisplacement of the C18 chains from the silica surface.

Relaxation Times

Generally speaking, the processes through which the spins go back to the equi-librium after the end of the application of an exciting radio-frequency pulse arereferred to as relaxation. These processes usually have an exponential natureand are characterized by time constants indicated as relaxation times: differentrelaxation times lead to distinct information, and their investigation is poten-tially informative about dynamic processes occurring in solids over a very broadrange of characteristic frequencies (104–1011 Hz). Spin-lattice relaxation timesin the laboratory (T1) and rotating frame (T1ρ) are sensitive to fast (MHz-GHz)and intermediate (mid-kHz) dynamic regimes, respectively; spin-spin relax-ation times (T2) are instead determined by dynamic processes occurring in aslow to intermediate regime. In principle, the possibility of measuring differentrelaxation times for all the NMR-active nuclei in a material allows one to ac-cess very detailed and site-specific information about a variety of motional pro-cesses. On the other hand, a complete and fully reliable characterization of thedynamics would require, for each system investigated, the measurement of re-laxation times of different nuclei as a function of temperature and frequency, andthe simultaneous analysis of the relaxation curves so obtained by means of suit-able dynamic models (see for instance, references 139–141). This is often un-feasible, and most of the applications reported in the literature are based on sim-pler comparisons between relaxation times measured for pure components andOIMM or for similar OIMM, based on the implicit assumption that the differentsystems have similar relaxation curves. This assumption is often acceptable,but in some cases biased dynamic results could be obtained if the contributionof important relaxation sinks (such as paramagnetic centers) were neglected.

In the following, some examples in which relaxation times have been usedto investigate the dynamic properties of the low-MW organic components ofOIMM are reported.

1H Spin-Spin Relaxation

As already introduced, because of the scarce resolution commonly achievablein 1H spectra of solid samples even in fast-MAS conditions, low-resolution

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 48: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

46 M. Geppi et al.

Figure 22. Temperature dependence of the two T2 components observed for naphtha-lene in silica with pores of 10 nm. The liquid signal intensity, shown as a solid line,revealed two distinct phase transitions for naphtalene inside and outside the pores andit was used as a reference for interpreting the T2 results. Reprinted from Mitchell andStrange (143) by permission of Taylor & Francis Ltd (http://www.informaworld.com).

techniques are often employed, where differences in isotropic chemical shiftsare rather minimized by using static conditions and low magnetic fields. Theseexperiments are usually performed “on-resonance” (removing the effects ofchemical shift) and directly analyzing the time-domain signal (free inductiondecay; FID), without resorting to Fourier transform. 1H relaxation times can beextremely useful especially in the presence of multiple dynamic processes, typi-cally occurring in heterophasic materials, since they can be revealed as separatecontributions to the observed relaxation process. Nevertheless, 1H spin-latticerelaxation times are strongly affected by the spin diffusion process, which oftenlimits their interpretation in purely dynamic terms. On the contrary, 1H spin-spinrelaxation times, not affected by spin diffusion, can be more easily used to getinformation on molecular motional processes. T2 (inversely proportional to thespectral line width) is mostly determined by residual (i.e., not averaged by themotional processes) homonuclear dipolar interactions: motions occurring witha frequency equal to or exceeding the “static” line width cause an increase in themeasured T2.

In a series of papers, Strange and coworkers (142, 143, and referencestherein) investigated the molecular mobility of cyclohexane and naphthalene ad-sorbed on different porous silicas by means of 1H T1, T1ρ , and T2 measurements

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 49: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 47

as a function of temperature. In one of these studies, Mitchell and Strange (143)discussed the molecular mobility of naphtalene molecules within silicas withpores of 4, 6, 10, 20, and 50 nm. Except for the case of the silicas withthe smallest and largest pores, which showed a peculiar behavior, for all thesamples two T2 components could be detected, as shown in Figure 22. Bycomparison with previous observations by Strange et al. (144) on the meltingof naphthalene inside and outside the pores, the two components were ascribedto a crystalline phase representing the core naphthalene within pores (shortT2), and to a plastic crystal phase, not existing in the bulk, forming a layerbetween the rigid core and the silica surface (long T2). The T2 of the plas-tic phase increased by decreasing the pore diameter, indicating an increaseddisorder and mobility of the layers as the confining pore size approached themolecular diameter. The ratio between the relative amplitudes of the two T2

components allowed the authors to estimate a surface layer approximately twomolecular diameters in thickness for the samples with pores of 6, 10, and20 nm. This behavior was not observed for the 4 nm diameter pores, wherea single T2 component of the order of milliseconds, expected for a viscousliquid, was detected. On the contrary, when naphthalene was confined withinthe largest pores, its behavior was virtually indistinguishable from that of thebulk.

13C Spin-Lattice Relaxation

In spite of their low sensitivity, 13C nuclei show some advantages with respect to1H nuclei in terms of the possibility of exploring molecular dynamics. Indeed,on one side 13C relaxation times can be measured exploiting high-resolutiontechniques; on the other side the contribution of spin diffusion to spin-latticerelaxation times can be usually neglected. These features allow site-specificrelaxation times, containing purely dynamic information, to be measured.

Komori et al. (130) investigated the dynamics of the alkyl chain of a C8

silylating agent used as surface modifier of a FSM-type mesoporous silica toget information about its possible effects on the behavior of p-nitroanilinemolecules absorbed on the inorganic framework. For the different sites of thechain, 13C T1 values increased with increasing the distance from the zeolitesurface, indicating the presence of a faster motion for the tail. This motionrevealed to be in the fast dynamic regime, with a correlation time shorter than 1ns above room temperature. Similar results were obtained by Lindner et al. (44),who observed the presence of an increased mobility along the alkylic chain ofan ether-phosphine ligand bound to silica in a sol-gel–synthetized material.

Khimyak and Klinowski (105) reported 13C T1ρ measurements for thecharacterization of the dynamics of the surfactant C16-trimethylammonium(TMA) intercalated in alumina with various cavity geometries (namely, lamel-lar [L] and hexagonal [Hex]). For all the C16-TMA inner-chain carbons, T1ρ

decreased going from L to Hex, indicating a larger mobility of the organic

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 50: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

48 M. Geppi et al.

Figure 23. 2H spin-lattice relaxation times vs. temperature of p-NA-d in the mesoporesof (a) FSM and (b)–(d) FSM modified with various organic groups. Solid lines indicatefittings to the BPP model. Reprinted from Komori and Hayashi (130). Copyright (2003),with permission from Elsevier Inc.

chains in the hexagonal mesocomposites, in agreement with the larger fractionof gauche conformations in the latter, already discussed in the subsection 1DHigh-Resolution Spectra/Carbon-13. The T1ρ relaxation decay for the TMAcarbons always sensibly deviated from a mono-exponential trend, indicatinga substantial dynamic hetereogeneity for this group, which was found to belarger in the lamellar phases L1 and L3.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 51: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 49

Figure 24. 31P spin-lattice relaxation times vs. temperature of polysiloxane-boundether-phosphine ligands either free (left) or Ru-complexed (right). The different sym-bols refer to different organic ligands (left) or Ru-coordinations (right). Reused withpermission from Lindner et al. (44). Copyright (1994) American Chemical Society.

Spin-Lattice Relaxation of Other Nuclei

In addition to 1H and 13C, relaxation times of other nuclei can be exploited forobtaining dynamic information.

Spin-lattice relaxation times of deuterium nuclei have been used to get in-formation on fast dynamics of selectively deuterated low-MW organic compo-nents in OIMM. For instance, Shantz and Lobo (136) used 2H T1 to characterizemultiple relaxation processes occurring in the structure-directing agent N,N,N-trimethylcyclohexylammonium, used in the synthesis of tectosilicate nonasil.Komori and Hayashi (130) found different relaxation behavior associated to thetwo components of the 2H line shape, associated to deuterated p-nitroaniline(p-NA-d) molecules within the micropores or on the outer surface of the zeoliteNaZSM-5 (see subsection 2H Static Spectra). The same authors (130) alsostudied the dynamic behavior of p-NA-d molecules in FSM-16 mesoporoussilicas, either unmodified or modified with different organic groups, by variabletemperature measurements of 2H T1. The relaxation time trends vs. temperaturewere reproduced taking into account a single motional process described by theBloembergen-Purcell-Pound (BPP) model, assuming an Arrhenius behaviorfor the correlation times (see Figure 23). Apparent activation energies couldbe determined, which resulted larger for the modified silicas with respect toFSM-16. A peculiar case was that of octyl-FSM-16 (Figure 23c), for whichtwo motional processes were needed to describe the experimental trend.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 52: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

50 M. Geppi et al.

31P spin-lattice relaxation times can also be exploited to get informationabout the dynamics of specific molecular moieties, since they are substantiallyunaffected by spin diffusion. For example, Lindner et al. (44) employed 31P T1 tocharacterize the dynamics of ether-phosphine ligands either free or complexedto Ru in a polysiloxane network prepared by sol-gel. An increase in T1 wasobserved in passing from the free to the bound ether-phosphine, which wasascribed to the expected reduced mobility of the polysiloxane-bound ligandsupon coordination to the ruthenium atom. Moreover, an increase in T1 (decreaseof mobility) with increasing the dimensions of the ligand was also observed.As shown in Figure 24, the decrease of T1 with temperature was larger in all thefree samples with respect to the polysiloxane-bound Ru complexes, indicatinga higher activation of the phosphine ligand motions for the former. On the otherhand, a sudden drop in T1 trend was observed for the bound samples at about330 K, which was ascribed by the authors to a thermal change in the materialpossibly related to a phase transition.

2D Exchange Techniques

Geometrical and timescale information about slow motional processes (char-acteristic frequencies in the range 10−2–103 Hz) can be obtained from 2Dexchange experiments (14). In solids, the NMR frequency of a given nuclearsite is in general dependent on the orientation of the corresponding molecu-lar fragment with respect to the direction of the static magnetic field. In 2Dexchange experiments, changes in the NMR frequency occurring after a timeinterval tm give rise to off-diagonal peaks; the specific shapes obtained can betherefore related to the reorientational angle spanned by the fragment duringtm and hence to the geometry and frequency of the motion.

Schaefer et al. (145) applied static 13C 2D exchange techniques to thestudy of the geometry of benzene hopping motions among the adsorption sitesin Ca-LSX zeolites; the timescale of these motions was obtained model-freeusing the 1D exchange induced sidebands (EIS) technique (146). Examples ofthree-dimensional views obtained from the 2D exchange technique are givenin Figure 25; it is evident that little or no exchange of benzene moleculesamong adsorption sites with different orientations occurred on a 1-ms timescale(absence of off-diagonal intensity), while this process was clearly seen on a300-ms timescale. Distributions of jump angles about the two mean values of0◦ and 109◦ were found. Moreover, after correcting the 1D EIS results for thecontribution of 13C spin diffusion, the authors determined reorientational ratecoefficients of about 5 ÷ 25 s−1 at room temperature and apparent activationenergies of 66 kJ/mol for the benzene site-hopping motions.

Another example of the use of 13C 2D exchange techniques to probe dy-namics of p-xylene in ZSM-5 zeolite was reported by Fyfe and Diaz (147).By means of 2D CP NOESY (148) experiments at different temperatures, theauthors investigated the slow motion of [13CH3] p-xylene molecules within

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 53: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 51

Figure 25. Experimental 13C 2D exchange spectra for benzene adsorbed in the su-percages of Ca-LSX zeolite, acquired at 298 K and with mixing times of: (a) 1 ms, (b)300 ms. Reused with permission from Schaefer et al. (145). Copyright (1997) AmericanChemical Society.

the zeolite channels. From previous X-ray studies, p-xylene molecules wereknown to occupy two different locations within ZSM-5, namely the sinusoidal(zigzag) channel and the (straight) intersection of the channels of ZSM-5. 2DCP NOESY experiments performed at increasing mixing times (see Figure 26)revealed the presence of two different exchange processes occurring at 297K. No exchange was detected at short mixing times (20 ms), as shown by theabsence of off-diagonal peaks. At a mixing time of 40 ms the activation ofintramolecular exchange between the two methyl groups of the molecule bothin the zigzag and in the straight channels was observed. At longer mixing times(100 ms) an additional, very slow intermolecular exchange process was de-tected, interchanging methyl group environments between molecules in zigzagand straight channels. Below room temperature no exchange was observed ateven long mixing times, suggesting that it was possible to quench the motionsby decreasing the temperature.

POLYMERIC ORGANIC COMPONENTS

In a large number of OIMM the organic component is a polymer, to whichan inorganic component is added, usually with the aim of improving severalmechanical and/or physical properties of the polymer itself.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 54: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

52 M. Geppi et al.

Figure 26. Contour plots of 13C 2D CP NOESY experiments on the complex of eightmolecules of [13CH3]p-xylene in ZSM-5 carried out at 297 K at mixing times of (a)20 ms, (b) 40 ms, and (c) 100 ms. The 1D CP/MAS spectrum is shown above. Thespectra show the three regimes of exchange observed in this system at the differentmixing times: (a) no exchange, (b) intramolecular exchange, and (c) intermolecular andintramolecular exchange. S and Z indicate signals assigned to methyl carbons in straightand zigzag channels, respectively. Reprinted with permission from Fyfe and Diaz (147).Copyright (2002) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 55: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 53

SSNMR has been extensively employed for the study of polymeric materi-als, by means of low- and high-resolution techniques. Several works have beenpublished before the birth of high-resolution SSNMR techniques, concerningthe investigation of the dynamic properties of polymers through the analysisof 2H line shapes or the low-resolution measurement of 1H relaxation times.Moreover, after the advent of high-resolution techniques and spectrometers inthe late 1970s, most of the applications of SSNMR in the 1980s and 1990sconcerned polymers (13, 14). Later, this experience could be used for the studyof the polymeric component in OIMM, in particular for what concerns thechanges on either the structural or dynamic properties of the polymer inducedby the presence of, and/or the interactions with, the inorganic component.

To this aim, several NMR techniques can be exploited, often giving com-plementary information. In the following, we tried to review some examplesof SSNMR studies to characterize the properties of the polymeric fractionof different classes of OIMM. These applications are classified on the basisof the NMR technique and included in the following subsections: 1D High-Resolution Spectra of Rare Nuclei and 13C Double Quantum (DQ) Experiments(both mostly giving structural information); 1H Free Induction Decay (FID)Analysis, 2D WISE Experiments, 1H MAS Spectra, 2H Static Spectra, Mea-surement of Dipolar Couplings, Exchange Techniques, and Relaxation TimeMeasurements (all of them being used to obtain dynamic information, albeiton different timescales).

1D High-Resolution Spectra of Rare Nuclei

As far as the experimental techniques for the obtainment of high-resolution1D spectra of rare nuclei and the isotropic chemical shift dependence on thephysical properties are concerned, what was stated in the previous sections holdstrue also for polymers. In particular, 1D spectra can easily give informationabout the phase structure of the polymer and/or its conformational propertiesin OIMM.

As an example, VanderHart et al. (149) investigated the effects of thepresence of exfoliated clays on the crystalline form (either α or γ ) of nylon-6. Indeed, the α and γ forms of nylon-6 gave rise to unique peaks in the13C high-resolution spectra at 26.3 and 34.0 ppm, respectively, allowing theirstraightforward identification. The authors observed that, while the α-form wasexclusively found in pure nylon-6 under a wide range of crystallization con-ditions, the development of γ -form crystallites was usually promoted by thepresence of nanodispersed montmorillonite and laponite clays (see Figure 27).Similar conclusions could also be drawn on different nylon-6/montmorillonitenanocomposites by Mathias et al. (150) from 15N CP/MAS spectra, since inthis case the α and γ forms gave rise to separate signals at 84.9 and 89.3 ppm,respectively, superimposed to a much broader line, ascribable to the amorphous

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 56: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

54 M. Geppi et al.

Figure 27. Expansions of the aliphatic regions of the “crystalline” 13C spectra ofnylon-6 (N6-a) and a nylon-6/laponite nanocomposite (N6C-B-L1-b). These spectrawere obtained from a linear combination of the 13C CP/MAS spectra recorded with 0-and 7-ms proton spin-locking times prior to the 1-ms CP time. Reused with permissionfrom VanderHart et al. (149). Copyright (2001) American Chemical Society.

phase (Figure 28). Moreover, Bertmer et al. (85), looking at the relative intensi-ties of the two 15N crystalline signals (obtained after suitable spectral deconvo-lution) in a series of nylon-6/montmorillonite nanocomposites, observed thatthe amount of γ -crystalline phase increased upon increasing the clay content.

Park et al. (151) investigated the crystalline structure of poly(vinylidenefluoride) (PVDF) in binary composites with silica and ternary composites withsilica and poly(methyl methacrylate) (PMMA) by means of 19F MAS spectra.In this case, different resonances could be attributed to amorphous domains(−88 ppm), regioirregular structures (−112.4 and −110.4 ppm), α-crystalline(−93.7 and −79.6 ppm) and γ -crystalline domains (−101.3 and −84.2 ppm),also on the basis of the 19F T1ρ values corresponding to the different signals.The authors could establish that only α-crystalline and amorphous domainswere present in the pure polymer and in its binary composites with silica, whilethe γ -crystalline form was also present in the ternary composites.

On the other hand, Panziera et al. (152) detected differences in the 13C, 15N,and 31P high-resolution spectra in passing from pure polydimethylphosphazene(PDMP) to the composite systems resulting from deposition of Pd nanoparti-cles, obtained by metal vapor synthesis, on PDMP (PDMP/Pd). In particular,four 31P signals were detected in PDMP, which, thanks to the application ofseveral selective techniques, were assigned to one mobile amorphous phase and

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 57: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 55

Figure 28. 15N CP/MAS spectra of (a) nylon-6, (b) nylon-6/clay nanocomposite, (c) re-precipitated nylon-6/clay nanocomposite, (d) annealed nylon-6, and (e) annealed nylon-6/clay nanocomposite. Asterisks denote spinning sidebands. Reprinted with permissionfrom Mathias et al. (150). Copyright (1999) American Chemical Society.

three rigid ones (presumably two crystalline and one interfacial) of the poly-mer (see Figure 29). These signals reduced to two (one amorphous and onecrystalline) in the 31P spectrum of PDMP/Pd. Moreover, the shifts experiencedby all the 31P signals suggested that both polymeric amorphous and crystallinephases experienced a structural rearrangement after Pd deposition. On the otherhand, from both 31P and 13C quantitative spectra a mobile/rigid ratio of about45/55, substantially equal in PDMP and PDMP/Pd, was calculated.

Geppi et al. (25), by acquiring 13C selective spectra on PE-b-PEG/silicahybrid materials (see Figure 30), could detect the simultaneous presence ofcrystalline and amorphous components for both PE and PEG blocks. In the caseof PEG the two components could be identified by the presence of two signalsresonating at very similar chemical shifts but characterized by very different linewidth, the larger one being unusually associated with the crystalline fraction.

13C Double Quantum (DQ) Experiments

Conformational studies are particularly important in polymer-containingnanocomposites, where these properties can be strongly affected by theinteractions of the polymer chains with the inorganic surface and/or by theirspatial confinement. Apart from the analysis of isotropic chemical shifts, moreadvanced SSNMR techniques can be applied in some cases to get very detailed

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 58: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

56 M. Geppi et al.

Figure 29. 31P spectra of PDMP (a)–(c) and PDMP/Pd (d)–(f). (a) and (d) are quan-titative DE/MAS spectra. (b) and (e) are CP/MAS spectra, enhancing the signals fromrigid phases. (c) and (f) are 13C Delayed CP/MAS spectra, where the signals of carbonscoupled to protons with long T2 (i.e., in rigid domains) were suppressed. Reused fromPanziera et al. (152). Copyright (2007), with permission from Elsevier Inc.

information on the conformational properties of polymers, such as precise tor-sion angles. Among these, the 13C 2D double quantum (DQ) experiment (153)detects and combines anisotropic 13C chemical shifts and 13C-13C dipolar cou-plings, giving 2D patterns that depend on the relative orientations of molecularsegments to which the coupled 13C nuclei belong, eventually allowing torsionangles to be obtained by simulation procedures.

Harris et al. (154) applied this technique to systems where poly(ethyleneoxide) (PEO) was intercalated in layered clays and MoS2, investigating the

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 59: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 57

Figure 30. 13C spectra of a PE-b-PEG/silica hybrid material: (a) 13C Delayed CP/MASspectrum recorded inserting a 100 µs delay between the proton 90◦ pulse and the contacttime (500 µs), in order to select mobile regions; (b) 13C CP/MAS spectrum recordedwith a contact time of 50 µs, in order to select rigid regions. (c) 13C CP/MAS spectrumrecorded with a contact time of 5 ms (presence of both mobile and rigid regions).Reprinted from Geppi et al. (25).

conformations of the OCH2 CH2O bonds of PEO. In order to detect the13C-13C dipolar couplings, the PEO used contained 13% of 13C-13C labeledrepeat units. The aim of this study was the detection of possible changesin PEO conformations in passing from the pure polymer (where the chainsforming the crystalline phase are arranged in a 72 helical structure, with gaucheconformations having torsion angles ranging from 49◦ to 92◦) to the composites,ascribable to polymer intercalation in the inorganic structure. The simulation ofthe experimental 2D patterns (see Figure 31 as an example) allowed the authorsto determine a ratio between gauche and trans units of 92:8 and 88:12 inclay nanocomposites and PEO/MoS2 compounds, respectively, with a gauche

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 60: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

58 M. Geppi et al.

torsion angle of 70 ± 15◦, thus ruling out the possible occurrence of a planarall-trans conformation.

1H Free Induction Decay (FID) Analysis

Proton static spectra or, equivalently, free induction decays (FIDs) of solids,recorded under low-resolution conditions (i.e., without MAS), are mostly de-termined by the homonuclear 1H-1H dipolar couplings left unaveraged bymolecular motions. In heterogeneous solids and, in particular, in OIMM, theon-resonance FID is usually the combination of several decaying functions,each characterized by a specific analytical expression (Gaussian, exponential,Weibullian, etc.) and by a characteristic time T2. For this reason, from FIDanalysis it is quite straightforward to identify fractions of samples exhibitingdifferent dynamics (from the number of components needed to describe theFID), to get information about the “rate” of the motion (from T2) and to quan-tify the number of protons belonging to the different fractions (from the weightof the different FID components). In spite of the absence of any distinctionamong protons in different chemical environments, this experiment is there-fore particularly powerful in getting dynamic information on heterogeneousmaterials, such as OIMM.

For example, Litvinov et al. (155) applied the 1H FID analysis to thestudy of poly(dimethylsiloxane) (PDMS) grafted onto silica surfaces. The two

Figure 31. 13C 2D double-quantum spectrum of PEO/MoS2: (a) experimental, (b)simulated. The simulation was carried out using a trans content of 12%. The trans andgauche angles were distributed within ±10◦ around 180 and 70◦, respectively. Reprintedwith permission from Harris et al. (154). Copyright (1999) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 61: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 59

components found in the 1H FIDs of these materials, with T2 of about 80 µsand 1 ms, were ascribed to semi-rigid PDMS at the silica interface and tomobile chain portions outside the interface, respectively. A third component,with a much longer T2 (20–100 ms), was detected in samples swollen in C2Cl4,assigned to highly mobile chain portions of long chain loops located abovedensely grafted PDMS layers or in areas of low grafting density. From theweights of the different FID components the fractions of semi-rigid, mobile,and highly mobile fractions could be determined, allowing the thickness of the“interface” (the PDMS semi-rigid layer adjacent to the silica surface) to beestimated about 1 nm, roughly corresponding to four siloxane units.

In a similar study, carried out on natural rubbers filled with silica particles,either functionalized or not with various coupling agents, ten Brinke et al.(156) could determine for the different materials the fraction of immobilizedrubber chains present on the silica surface. The largest amount was found forpure silica (15%), while lower amounts were found in the presence of couplingagents (from 13% for 3-mercaptopropyltriethoxysilane-grafted silica to 1% forpropyltriethoxysilane-grafted silica), decreasing with the decreasing affinity ofthe coupling agent functional groups with the natural rubber C=C bonds.

As a last example, Borsacchi et al. (49), by 1H FID analysis, observedchanges in the dynamics of LDPE, induced by the presence of silica as a filler.Even though a good description of the FID required at least four components,rendering a detailed analysis of the dynamic behavior quite complex, the in-teraction of LDPE with the silica surface resulted in a clear stiffening of thepolymeric chains, as detected by both larger weights of the most rigid fractionsand shorter T2 of the various components observed in the composites respectto neat LDPE.

2D WISE Experiments

1H FIDs (or 1H 1D static spectra) contains information on residual homonu-clear dipolar couplings arising from all the protons of a sample. When thedifferent protons are coupled to 13C nuclei experiencing different chemicalenvironments, it is possible to resort to the 2D WIdeline SEparation (WISE)experiment, from which individual contributions to the 1H 1D static spectracan be extracted corresponding to different 13C isotropic chemical shifts. Inprinciple, this allows specific dynamic information at different molecular sitesand/or in different components to be obtained.

An example is given by De Paul et al. (157), who studied composites pre-pared from block copolymers poly(isoprene-b-ethylenoxide) (PI-b-PEO), (3-glycidyl-oxypropyl) trimethoxysilane (GLYMO), and aluminum sec-butoxidevia a sol-gel process. The 2D WISE experiment (see Figure 32) showed that theslices extracted for 13C resonances ascribable to the PI block contained narrow

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 62: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

60 M. Geppi et al.

(≈10 kHz) 1H signals, indicating a mobile environment, while broad protonsignals (≈50 kHz) were found in correspondence of PEO and GLYMO 13C res-onances, ascribable to very rigid environments. These results were in agreementwith an arrangement of the different components of the material in alternatePI and aluminosilicate/PEO homogeneous layers, which could be confirmedalso with the aid of other SSNMR techniques (see subsection Dimensions andDispersion of Organic and Inorganic Domains/Spin Diffusion Experiments).

1H MAS Spectra

As already pointed out, when MAS is applied in recording proton spectra,a partial reduction of the line width is usually obtained, with an improvedresolution, strictly depending on the system under study and on the MAS rate.On the other hand, when motions with characteristic frequencies (ωmot ) of theorder of the spinning rate (ωr ) are present, the proton line width in 1H MASspectra shows a characteristic profile vs. ωr , the analysis of which can allow

Figure 32. (a) Projection along the 13C dimension of the 13C-1H 2D WISE experi-ment performed on PI-b-PEO/aluminosilicate lamellar composite. Proton traces cor-responding to: (b) vinyl protons in PI, (c) -CH2O- protons in PEO and GLYMO, (d)aliphatic protons not near oxygen or silicon atoms, and (e) -CH2-Si- protons in GLYMO.Reused with permission from De Paul et al. (157). Copyright (1999) American ChemicalSociety.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 63: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 61

the motional process to be characterized, since relations exist linking ωr andωmot .

For example, Tajouri et al. (158 and references therein) applied this ap-proach to the study of PEO grafted on silica. In this case, all the PEO protons(far from chain ends) contributed to the same signal. The line width of thissignal, measured as a function of ωr , was interpreted by the authors in terms oftwo different dynamic processes: a faster segmental motion of the PEO chainsand a slower reorientation of the whole chains relative to each other. The corre-lation times of these motions were extracted by fitting the experimental curveto simplified theoretical models at different temperatures, PEO grafting ratios,and preparation conditions.

2H Static Spectra

As pointed out in the previous section, 2H static spectra provide very detailedinformation on dynamic processes occurring in an intermediate frequency range(tens of kHz), the main drawback being the need of molecules selectivelydeuterated on the sites of interest.

As an example, in a series of papers Blum and coworkers (159–162)have investigated the segmental dynamics of poly(methyl acrylate), selectivelydeuterated on the methyl groups (PMA-d3), in composites with silica. Theroom-temperature 2H spectrum of bulk PMA-d3 was constituted by a reducedPake pattern, as a result of the fast continuous rotation of the methyl groupabout its ternary symmetry axis, which scaled the full Pake pattern by a factorof 3. By increasing the temperature, this pattern gradually transformed intoa single resonance, due to the reduction of the quadrupole couplings causedby the backbone segmental mobility associated with the glass transition. Theauthors showed that the pattern could be reproduced by taking into account asingle-component line shape, and a complex motion including isotropic rota-tional diffusion and discrete jumps, each described by a single correlation time.In spite of the complexity of the motion, a homogeneous dynamic behavior wastherefore found for all the polymeric segments. The situation was quite differentfor samples where PMA-d3 was adsorbed on the surface of silica; in this case,the deuterium spectra could be reproduced only taking into account at leasttwo or three motionally distinct line shape components, indicating the occur-rence of dynamic heterogeneity, in agreement with a broader glass transitionwidth observed by calorimetric techniques. The comparison among samplesof PMA-d3 adsorbed on silica and samples also containing overlayers of un-deuterated polymers (PMA, polystyrene) allowed the authors to explain thisdynamic heterogeneity as a motional gradient on the silica surface, ranging frommore mobile polymeric segments at the polymer-air interface (absent when a

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 64: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

62 M. Geppi et al.

polymeric overlayer was present) to more rigid segments at the polymer-silicainterface (161) (see Figure 33).

Measurement of Dipolar Couplings

Information about polymer molecular dynamics can also be obtained by mea-suring residual 1H-13C dipolar couplings. Different methods can be used de-pending on the rate of the motions investigated. Two applications are reported inthe following, the first concerning small fluctuations in rigid polymeric chains,the second dealing with fast chain reorientations in rubbery polymers.

Methods exist allowing site-specific measurement of one-bond 1H-13Cdipolar coupling in CH or CH2 groups belonging to polymer backbones. In-deed, in this case, being the one-bond C H distance approximately constant,the dipolar coupling directly reflects the dynamics of the polymeric backbone.Moreover, systems at natural isotopic abundance can be directly investigated,provided that site-specific dipolar couplings can be measured for the differentcarbon nuclei. This can be in principle achieved by exploiting the 13C spectralresolution in MAS spectra. However, in most cases encountered in polymer-inorganic composites, the spectral resolution may not be very good becauseof the heterophasicity of the polymeric systems. In these cases, a preventiveselection of signals arising from different phase domains is needed in order toindividually characterize their dynamic behavior. In this sense, an interestingapproach has been recently reported by Brus et al. (163), who introduced someselective amplitude-modulated 2D PISEMA experiments (164), able to indi-vidually measure dipolar couplings for amorphous and crystalline polymericphases. The dipolar couplings DCH are proportional to the dipolar splittingsν directly observed in the traces of the relevant 2D experiments. The dynamicbehavior of a fragment could be then quantified by means of an order parameterS, defined as the ratio between the measured ν and its value for a completelyrigid fragment, estimated to be 12.0 kHz. In turn, S was related to a root meansquare (rms) angular fluctuation of the C-H bond vector (

√〈θ2〉), assuming that

the segmental motion was axially symmetric and small in amplitude, throughthe equation:

S = 1 − 3

2〈θ2〉 (2)

This approach was applied by the authors to nanocomposites constituted bynylon-6 and clay platelets and, in some cases, also by ethylene-methyl acrylatecopolymer (EMA). The order parameter was determined in both pure nylon-6and nanocomposite systems for each methylene group of the backbone andin both crystalline and amorphous phases. The order parameter describing therestricted librations above discussed was found to change within the polymer

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 65: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 63

Figure 33. Deuterium static spectra recorded at various temperatures (indicated inthe figures in ◦C) for PMA-d3 (“bulk PMA”), PMA-d3 adsorbed on silica (“Surfacesample”), and PMA-d3 adsorbed on silica and containing an overlayer of polystyrene(“PS composite”). A reduction of the central component of the spectrum, ascribableto mobile polymeric segments, is clearly seen at 44 and 52◦C in passing from thesurface sample to the PS composite. Reused with permission from Lin and Blum (161).Copyright (2001) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 66: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

64 M. Geppi et al.

repeat unit slightly decreasing from NH and CO groups to the central parteither in the amorphous phase (rms from 18 to 29◦ at 308 K, see Figure 34) orin α- and γ -crystalline ones (rms from 12 to 17◦ at 308 K). Librations weretherefore characterized by higher amplitudes in the amorphous phase. Besidethese “constrained” domains, responsible for the spectral dipolar doublets, also“free” domains were found to be present in the amorphous phase, characterizedby fast trans-gauche two-site tetrahedral jumps, contributing to the centralsignals in the dipolar profiles (see Figure 34). Comparing the dynamic behaviorof neat nylon-6 with that of two- and three-component nanocomposites, theauthors found an increased amplitude of polymer backbone librations occurringin the amorphous phase as a result of the presence of the clay platelets (see inparticular the signals C1 and C5 in Figure 34), further slightly increased by thepresence of EMA.

Figure 34. Traces of 13C-1H 2D PISEMA experiments at 308 K, where the amorphousregions were selected, for neat nylon-6 (PA6), and its nanocomposites with the claywithout (NC) and with EMA (NC/EMA). Traces for different nylon-6 carbon nuclei,labeled as indicated in the structure above the spectra, are reported. The number aboveeach spectrum is the dipolar splitting in kHz, while that next to each spectrum is thecalculated rms angular fluctuation. Reused with permission from Brus et al. (163).Copyright (2006) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 67: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 65

When more mobile polymeric systems, like elastomers, are considered,residual proton dipolar couplings can still be used to investigate dynamics. Inthis case, it is possible to resort to 1H multiple quantum (MQ) methods, whichhave been extensively reviewed by Saalwachter (165). The MQ experiment iscarried out at different effective evolution times te. The trend of the intensityof the normalized MQ signal vs. te can be described by means of suitablemodels, from which equations containing the residual dipolar coupling 〈�d〉 asa parameter are derived.6 Maxwell et al. (166) applied this method to the studyof DC745, a commercial silicone elastomer containing quartz and fumed silicafillers, aged by exposition to ionizing γ radiation. The MQ build-up curvesobtained for three samples exposed to γ radiation for different times wereclearly different, indicating that a longer exposition resulted in a higher MQintensity, corresponding to a more restricted molecular mobility (see Figure 35).Moreover, the curves could be described only considering a bimodal distributionof dipolar couplings, indicating the presence of two distinct network domainsin the material. These two domains were ascribed by the authors to generalnetwork polymer chains (lower residual dipolar coupling, higher mobility) andto polymer chains either physically or chemically interacting with the silicafiller surface (higher residual dipolar coupling, lower mobility). By increasingthe dose of radiation, both the residual dipolar coupling in each domain and theamount of the low-mobility domain increased, suggesting occurrence of bothcross-linking in the polymer network and increasing thickness of the polymerlayer interacting with the silica surface.

Exchange Techniques

Information about chain dynamics in the slow-motion regime (characteristicfrequencies in the 1–1000 Hz range) can be obtained by means of either 1D or2D exchange techniques. These experiments exploit the dependence of nuclearinteractions, such as chemical shift and quadrupolar ones, on the molecularorientation respect to the magnetic field. The exchange pulse sequences includea mixing time during which slow molecular reorientations cause a change inthese interactions. This situation can be monitored by looking at intensitychanges in 1D experiments or at typical patterns in 2D experiments.

Bonagamba and coworkers (167, 168) applied both 13C 1D and 2D ex-change techniques to study the polymer dynamics in type-I and -II ormolytesconstituted by Li+-doped siloxane/poly(ethylene oxide) (PEO) materials. Fromthe analysis of 1D PUREX experiments, the authors found a nonexponentialbehavior of the correlation functions describing the PEO chain motions, which

6For a precise definition of the terms used and a more detailed description of theexperimental and data analysis procedures, the reader can refer to the original literature.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 68: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

66 M. Geppi et al.

Figure 35. Experimental 1H MQ buildup curves for three DC745 samples irradi-ated with cumulative doses of 5, 50, and 250 kGray. Lines are simulated growthcurves assuming either single dipolar couplings or their bimodal distribution. Reprintedwith permission from Maxwell et al. (166). Copyright (2005) American ChemicalSociety.

could be described in more detail by means of 2D exchange experiments. For in-stance, the 2D exchange spectra of a type-II ormolyte at different temperaturesaround its Tg were simulated using a model combining small-angle isotropicrotational diffusion and random large-angle jumps, with a log-Gaussian dis-tribution of correlation times (see Figure 36). More in general, large-anglePEO motions occurred in regions far from the silica clusters, while a restricteddynamics was experienced by PEO chains interacting with the inorganic do-mains. The authors found that dynamic heterogeneities in the slow regime weremore pronounced in type-II than in type-I ormolytes. Moreover, Li+ ions werecomplexed by PEO, inducing cross-linking of its chains: at lower doping levelsPEO chains involved or not in Li+ complexation were both present, while athigher doping levels most of the polymer segments were cross-linked, resultingin a more homogeneous dynamic behavior.

Similar exchange techniques can be applied to nuclei other than 13C, andin particular to 2H. An example is represented by the study carried out by

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 69: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 67

Figure 36. Simulated and experimental 13C 2D exchange spectra for a type-II ormolyteat −35◦C (Tg), −25◦C (Tg+ 10◦C), and −15◦C (Tg+ 20◦C). tm indicates the mixing timeand τD the mean correlation time of the diffusive process. Reprinted with permissionfrom Bathista et al. (167). Copyright (2007) American Chemical Society.

Yang and Zax (169) on perdeuterated PEO intercalated into fluorohectorite. Bysimulating deuterium 2D exchange spectra recorded at different temperatures,the authors found that, when PEO was nanoconfined within the clay layers,its dynamic behavior was strongly modified with respect to the bulk state. Inparticular, below Tg , while polymer backbone reorientation was negligible inbulk PEO, small-angle reorientations were present in the nanocomposite. Onthe contrary, above Tg , large-angle chain reorientations were observed in bulkPEO but not in the nanocomposite.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 70: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

68 M. Geppi et al.

Relaxation Time Measurements

What was stated in the previous section about relaxation times mostly holdstrue for the study of the polymeric component in OIMM. In the following, afew examples of applications will be briefly discussed.

Mello et al. (101) measured 13C, 1H, and 7Li T1 as a function of temper-ature for type-I and type-II ormolytes obtained by dissolution of LiClO4 intosilica/poly(ethylene glycol) (PEG) matrices. Similar trends were obtained for13C and 1H relaxation times in both type-I and type-II ormolytes, with a T1

minimum occurring about 40◦C below the 7Li T1 minimum (see Figure 37).These observations led the authors to hypothesize that the cation mobility wasassisted by the segmental motion of the polymer, the rapid cation motions oc-curring only when the motions of the polymer segments were fast enough toassist the cationic jumps.

Forte et al. (170) studied the polymer dynamics in complexes of hec-torite and methyl methacrylate (MMA)/2-(N-methyl-N,N-diethylammoniumiodide)ethyl acrylate (MDEA) copolymers by measuring 1H T1ρ and 13C T1

and T1ρ at room temperature for a variety of samples differing for MMA/MDEAmolar ratio (n) and preparation procedure (either copolymerization of MMAwith the clay prefunctionalized with MDEA-pol, or direct interaction of a pre-formed MMA/MDEA copolymer with hectorite-ins). The sample 6ins showedrelaxation times very similar to the pure copolymer with the same MMA/MDEAmolar ratio, indicating the absence of detectable variations in the kHz and MHzdynamics following insertion in the clay. On the other hand, sample 2ins showed

Figure 37. Temperature dependence of 1H, 7Li, and 13C line widths and spin-latticerelaxation rates for a type-I ormolyte. Reprinted with permission from Mello et al. (101).Copyright (2000) American Chemical Society.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 71: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 69

similar MHz dynamics, but more restricted kHz motions, as suggested by thereduction of both 13C and 1H T1ρ , in agreement with the higher number of poly-mer links (ammonium ions of the MDEA units) to the clay. A peculiar dynamicbehavior in the kHz regime for all the chain groups, and in the MHz regime forthe sole MMA methyl group, was instead observed for the 8pol sample (abouttwo-time longer 1H and 13C T1ρ), which, in agreement with layer spacing dataobtained from X-ray powder diffraction spectra, was ascribed to the occurrenceof the MMA polymerization mostly outside the clay layers.

As a last example, in a series of papers Grandjean et al. (171 and referencestherein) studied the dynamics of poly(ε-caprolactone) (PCL) in nanocompositeswith synthetic saponites by measuring 13C spin-lattice relaxation times. Theauthors reported that for PCL carbons 13C T1ρ was in the slow-motional side ofthe relaxation curve, thus decreasing with increasing mobility. In all samples T1ρ

was found to decrease (and consequently dynamics in the kHz regime increased)within the PCL chain going from C6 to C2/C5 to C3/C4 within each monomericunit (1 is the carboxylic carbon and the numbering is progressive along the alkylchain, 6 being the carbon bonded to the oxygen). Moreover, all the measured 13CT1ρ values decreased passing from pure PCL to the nanocomposites, indicatingthat intercalation in saponite involved an increased polymer mobility in thekHz regime, probably associated to an observed decrease in PCL crystallinity.A corresponding decrease in 13C T1 was also observed, but these relaxationtimes remained very long (30–90 s) even in the nanocomposites, indicatingvery inefficient motions in the MHz regime.

DIMENSIONS AND DISPERSION OF ORGANICAND INORGANIC DOMAINS

SSNMR is a very important tool to get information concerning spatial distancesand domain dimensions. This can be done by measuring dipolar couplings,which strongly depend on internuclear distances, and by following the protonspin diffusion process. The first approach is more often applied to the mea-surement of specific distances between pairs of nuclei and therefore usuallyrequires either very simple or suitably labeled molecular systems. Nonetheless,from the “nuclear” standpoint, OIMM represent somewhat simplified systems,since the organic and inorganic components contain different nuclear species.Indeed, the HARDSHIP technique, based on the detection of heteronucleardipolar interactions, will be described in this section as a tool to measure thethickness of inorganic domains in nanocomposites. On the other hand, spindiffusion represents the most used way to get information on average dimen-sions of domains in OIMM. This process, originated by the flip-flop mechanismoccurring by homonuclear dipolar couplings, tends to cancel out proton mag-netization gradients throughout a sample. Spin diffusion can be described bymeans of the traditional diffusional equations, which, in many cases, are difficult

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 72: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

70 M. Geppi et al.

to analytically solve. Resorting to the use of simplified geometrical models forthe single domains, considering uniform their distribution, and assuming thediffusion coefficient as spatially independent, a solution is obtained of the type:

〈x2〉 = f Dτ (3)

where D is the diffusion coefficient, 〈x2〉 is the magnetization mean squaredisplacement resulting from spin diffusion during time τ , and f is a numericfactor depending on the geometrical model chosen. The measurement of acharacteristic time, depending on the technique used, can allow average lineardimensions of the domains to be obtained from equation (3).

Three distinct groups of techniques that can give information on domaindimensions will be described in the following: (i) measurement of protonspin-lattice relaxation times; (ii) spin diffusion experiments; (iii) HARDSHIPexperiment. We also recall here that a variety of 2D techniques including aspin diffusion evolution time have been used to give information on spatialproximities at the organic-inorganic interface, as described in the subsectionOrganic-Inorganic Interfaces/2D Spectra of Spin-1/2 Nuclei.

Measurement of 1H Spin-Lattice Relaxation Times

Proton spin-lattice relaxation times are strongly affected by spin diffusion.Indeed, the different intrinsic relaxation times characterizing the different 1Hnuclei in a sample cause magnetization gradients, which tends to be averagedout by spin diffusion. This results in the tendency to average the measuredspin-lattice relaxation times to a single value for all the protons belonging to adipolarly coupled network. This average can be complete or not, depending onthe type of relaxation time (either T1 or T1ρ) and on the heterogeneity (domaindimensions) of a sample. The most straightforward application consists inmeasuring the relaxation times of different domains (for instance, exploitingthe spectral resolution of a rare nucleus via CP techniques) and simply verifyingwhether they are the same or not. If a single relaxation time is measured for thedifferent domains, this roughly indicates that the system is homogeneous on a100–200 A (T1) or 10–20 A (T1ρ) scale, as derived by using typical diffusionalcoefficients and geometrical factors in equation (3).

Several cases have been reported in the literature. As an example, Apperleyet al. (26) measured 1H T1 and T1ρ values by 1H-29Si CP techniques for silica-dimethylsiloxane (DMS) hybrids synthesized by sol-gel processes startingfrom SiCl4, tetraethoxysilane (TEOS), and hexamethylcyclotrisiloxane (D3).While slight differences were present in systems containing a D3 percentage≤10%, equal T1 and T1ρ values were measured within the experimental er-ror for protons belonging to DMS and silica for higher D3 percentages. The

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 73: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 71

authors could conclude that the systems investigated were highly homoge-neous, the siloxane segments being uniformly distributed throughout the silicanetwork.

More detailed and quantitative information can be obtained in materialswhere paramagnetic centers are present in the inorganic domains, such asnatural montmorillonite clay layers, containing a sensible amount of Fe3+ ions.In this case, the paramagnetic centers act as powerful relaxation sinks, stronglyreducing the spin-lattice relaxation times of the protons present in the proximityof the clays surface and, consequently, those of the remaining protons of theorganic phase by spin diffusion. Obviously, the influence of the Fe3+ ions on theT1 of the organic phase is stronger when the inorganic phase is well dispersed(e.g., fully exfoliated clay). This implies that proton T1 measurements cangive information on the quality of the inorganic domains dispersion within anorganic matrix. If a single T1 is measured, this will contain two contributions:one arising from the dynamics of the organic phase, the other from the effectsof paramagnetic centers:

(1

T1

)exp

=(

1

T1

)dyn

+(

1

T1

)par

(4)

where the contribution of the paramagnetic centers to the relaxation rate willbe greater when the inorganic domains are well dispersed, being proportionalto their surface-to-volume ratio.

A semiempirical approach was used by Bertmer et al. (85) and by Calberget al. (172) in the study of montmorillonite nanocomposites with nylon-6 andpoly(ε-caprolactone), respectively. In particular, Bertmer and coworkers took(1/T1)dyn equal to the value measured for the pure organic phase (implicitlyassuming that it did not experience dramatic dynamic changes in passing tothe organic-inorganic complex), so they could derive (1/T1)par from equation(4) by measuring 1H T1. The surface-to-volume ratio could be calculated from(1/T1)par as a function of the clay content, assuming different clay dispersions(from perfect exfoliation to groups of several clay platelets stuck together),and using known values for composition of the nanocomposites, densities, andclay platelets dimensions. The experimental data resulted in good agreementwith the values calculated in the hypothesis of two platelets stuck together(see Figure 38). For each clay content, average distances between the differentcouples of platelets could be estimated, which, however, were significantlylarger than those obtained by TEM.

VanderHart and coworkers performed proton T1 measurements on mont-morillonite nanocomposites with nylon-6 (173), polystyrene (PS) (174), andstyrene-acrylonitrile copolymer (175). More sophisticated models for the anal-ysis of these data are in particular present in Bourbigot et al. (174, 175). Themain assumptions contained in these models were (a) the intrinsic proton T1

in the nanocomposite was equal to the T1 measured for the pure polymer; (b)

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 74: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

72 M. Geppi et al.

Figure 38. Contribution to the proton spin-lattice relaxation rate arising from the effectsof paramagnetic centers (left axis) and surface-to-volume ratio (right axis) as a functionof clay content for a nylon-6/montmorillonite nanocomposite. The lines refer to thesurface-to-volume ratios in the case of perfect exfoliation (solid line) and of two (dashedline) and three platelets (dotted line) stuck together. Reprinted with permission fromBertmer et al. (85). Copyright (2007) American Chemical Society.

the clay platelets were parallel to one other and equally spaced; and (c) theinfluence of the paramagnetic centers was mimicked by a fast-relaxing layerof polymer. In a heterogeneous system where a very efficient relaxation sinkis present (the paramagnetic centers), and where the protons experience spindiffusion, the relaxation profile is in principle multiexponential. The T1 decay(or recovery) curves should be therefore described by multiple T1. Importantparameters concerning clay exfoliation or dispersion, as well as mean spacingand homogeneity of clay-polymer interfaces, could be obtained by the authorsfrom the analysis of the relaxation curves. The initial slope of the trend of themagnetization vs. square root of time (corrected for the contribution arisingfrom the intrinsic T1) was about proportional to the total polymer-clay interfa-cial area. This was compared with the maximum amount of interfacial area thatcould have formed in the case of perfect clay exfoliation, to give the fraction ofeffective polymer-clay interfacial area (f ). Ten different PS/montmorillonitenanocomposites were studied by the authors, who found f values ranging from0.08 to 1, revealing in most cases the occurrence of clay tactoids (multilayerstacks of montmorillonite platelets). On the other hand, the longest relaxationtime could be related to the apparent mean spacing between clay-polymer in-terfaces (app). This value could also be compared with f , the spacing inan ideally layered structure having the surface area derived from f , to obtainthe parameter ε = f /app, which represent a measure of the homogeneityof the clay distribution (ε = 1 in case of perfect homogeneity). For the tenPS/montmorillonite nanocomposites above mentioned, values of app ranging

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 75: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 73

from 350 to 600 nm were obtained, corresponding to ε values of 0–0.7, indicat-ing in most cases a considerable amount of dispersion for the polymer-clay in-terfaces. Some more applicative aspects of this methodology, as applied to poly-mer/paramagnetic clays nanocomposites, have been discussed by Gilman et al.(176).

Spin Diffusion Experiments

Experiments can be performed to directly monitor the spin diffusion process.The first experiment of this type was proposed by Goldman and Shen (177)and consisted of three steps: (a) selection of proton magnetization in mobiledomains, exploiting differences in T2; (b) mixing time (tm) allowing spindiffusion to proceed towards rigid domains; and (c) detection of the protonsignal in the rigid (and/or mobile) domains. The spin diffusion curves builtby reporting the intensity of the magnetization as a function of tm couldbe interpreted in terms of diffusional equations to get average domain sizes(see equation (3)). Several variants of this sequence have been devised; forinstance, modifying the selection criterium in step (a) or the way of detectingproton magnetization in step (c) (the most notable case is after CP to carbonnuclei, in order to exploit the 13C spectral resolution).

The Goldman-Shen pulse sequence was applied by Brus and Dybal (178),with a phase cycling eliminating undesired T1 effects, to hybrid siloxanenetworks obtained by copolymerization of tetraethoxysilane (TEOS) anddimethyl(diethoxy)silane (DMDEOS). The initial step of the sequence selectedthe magnetization of silanol protons, located in the TEOS particles, and its dif-fusion to the DMDEOS/TEOS copolymer phase was monitored. From theanalysis of the signal intensity vs.

√tm curves, the authors could estimate aver-

age linear domain sizes of 1.3, 2.1, and 0.5 nm for TEOS homopolymer phase,copolymer phase, and interface, respectively.

A similar study was presented by Bertmer et al. (179) for short-chainPDMS grafted on pyrogenic silica. In this case, the model consisted of atwo-phase structure, made of rigid (close to the interface) and mobile PDMSdomains: a double-quantum dipolar filter was used in the pulse sequence toselect proton magnetization from the rigid fraction and to subsequently ob-serve its diffusion to the mobile fraction. Domain dimensions for the mobileand rigid fractions could be estimated at different PDMS chain lengths andtemperatures.

De Paul et al. (157) applied this approach to obtain domain sizes in thenanocomposite already described in the subsection Polymeric Organic Compo-nents/2D WISE Experiments, consisting of PI-b-PEO and an aluminosilicate.In particular, they employed a dipolar filter to select the proton magnetizationof the mobile PI block, and the signal was acquired both directly and after

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 76: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

74 M. Geppi et al.

Figure 39. Intensity of the signals of (a) Si-CH2 carbons and (b) PI protons as a functionof

√tm, acquired through 1H-13C CP and directly on protons, respectively. Reprinted

with permission from De Paul (157). Copyright (1999) American Chemical Society.

1H-13C CP. The trends of signal intensities, suitably corrected to take into ac-count T1 relaxation occurring during the evolution time tm, as a function of√

tm (see Figure 39) could be interpreted in terms of domain sizes, by usingdiffusional coefficients estimated from measured 1H T2. In particular, the factthat the initial buildup of the curve relative to Si-CH2 signal intensity extrap-olated to

√tm = 0 indicated that no significant (>1 nm thick) PEO interphase

was present between the PI block and the rigid inorganic phase incorporat-ing PEO. Indeed, in the presence of a PEO interphase, the PI magnetizationshould first diffuse to the interphase and only later reach the GLYMO protons,

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 77: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 75

giving rise to a√

tm-intercept greater than zero. Moreover, the total thick-ness of the two lamellae formed by mobile PI and rigid PEO+aluminosilicatecould be estimated to be about 23 nm, in very good agreement with SAXSobservations.

Spin diffusion buildup curves were obtained by Brus et al. (23) for sil-ica/epoxy films using 2D 1H-1H CRAMPS and 1H-13C HETCOR experiments,modified to include a spin diffusion mixing time (see also subsection Organic-Inorganic Interfaces/2D Spectra of Spin-1/2 Nuclei). The systems investigatedwere hybrids made from [3-(glycidyloxy)propyl] trimethoxysilane (GTMS) ordiethoxy[3-(glycidyloxy)propyl] methylsilane (GMDES), polyoxypropylene,and colloidal silica particles. For instance, by looking at the HETCOR cross-peak between GTMS or GMDES CH2-Si carbons (9–14 ppm) and CH2-Oprotons (3.5 ppm), present in both GTMS or GMDES and polyoxypropylene(see Figure 40a), the authors obtained the buildup curves shown in Figures 40band 40c. A two-step process was observed for GTMS-silica (Figure 40b), thefaster corresponding to magnetization transfer within GTMS monomer units,and the slower (occurring on a spatial scale of about 0.8 nm) reflecting thedistance between protons in oxypropylene units and those on the surface ofsiloxane clusters, suggesting a significant spatial separation of GTMS residuesand oxypropylene chains. On the contrary, a single-step process (Figure 40c)was observed for GMDES-silica, reflecting a complete homogeneity, the in-dividual segments of various monomer units being mixed within the 0.45 nmrange.

HARDSHIP Experiments

An alternative approach to the measurement of nanoparticles thicknesses in therange 1–10 nm has been recently proposed by the group of Schmidt-Rohr (71,180). Contrary to the previous techniques, rather than on proton spin diffusion,this approach is based on the investigation of heteronuclear dipolar couplingsbetween protons in the organic phase and X nuclei (31P, 29Si, 13C, etc.) inthe inorganic phase, exploiting the strong dependence of the strength of thedipolar coupling on the internuclear distance rXH (∝ 1/r3

XH ). This is based ona REDOR-type pulse sequence (HARDSHIP: heteronuclear recoupling withdephasing by strong homonuclear interactions of protons), where the dephas-ing contribution to the X REDOR signals from protons in the inorganic phaseis suppressed, exploiting their usually longer T2. Similarly to REDOR exper-iments, decay curves are built reporting signal intensities vs. an experimentaltime, which is multiple of the rotor period (see Figure 41). These curves can bereproduced by suitable models, based on the calculation of dephasing curves forX nuclei at various depths from the organic-inorganic interface within the inor-ganic domains and taking into account all the molecular orientations respect to

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 78: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

76 M. Geppi et al.

Figure 40. (a) 1H-13C 2D HETCOR spectra obtained for the GTMS-silica systemwith a 800 µs spin diffusion mixing time; the spectra on the right are 13C slices forCH3(0.8 ppm) and CH2-O (3.5 ppm) proton signals. (b), (c) Spin diffusion built-upcurves of cross-peaks in 1H-13C HETCOR spectra correlating: (b) the GTMS carbonsignal at 9 ppm and the CH2-O proton signal at 3.5 ppm, in GTMS-silica; (c) the GMDEScarbon signal at 14 ppm and the CH2-O proton signal at 3.5 ppm, in GMDES-silica.Reused with permission from Brus et al. (23). Copyright (2004) American ChemicalSociety.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 79: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 77

Figure 41. 29Si-1H HARDSHIP experimental data (circles) of hectorite clay dispersedin poly (vinyl alcohol) and corresponding best-fitting curve (line). Reused with permis-sion from Schmidt-Rohr et al. (180). Copyright (2007) American Institute of Physics.

the rotor axis. The inorganic domain dimensions can be obtained as best-fittingparameters of the decay curves provided that some quantities are independentlyknown and inserted in the relevant model equations (e.g., shape of the nanopar-ticles and proton density in the organic phase). The authors have shown appli-cations of this technique to several materials, including a hectorite-poly(vinylalcohol) composite, bone, and phosphate-glass/polyamide-6 hybrids. A typicalerror on the particle thickness determination, as obtained by HARDSHIP, wasreported by the authors to be about 20%.

REFERENCES

1. Grandjean, J. (2006) Solid-state NMR study of modified clays and polymer/claynanocomposities. Clay Mat., 41 (2): 567–586.

2. Olejniczak, S., Kazmierski, S., Pallathadka, P.K., and Potrzebowski, M. (2007)A review on advances of high-resolution solid state NMR spectroscopy instructural studies of polymer/clay nanocomposites. Polimery, 52 (10): 713–721.

3. Babonneau, F., Bonhomme, C., Gervais, C., and Maquet, J. (2004) Advances incharacterization methods for sol-gel derived materials: High resolution solid statenuclear magnetic resonance. J. Sol-Gel Sci. Tech., 31 (1–3): 9–17.

4. Bonhomme, C., Coelho, C., Baccile, N., Gervais, C., Azais, T., and Babonneau,F. (2007) Advanced solid state NMR techniques for the characterization of sol-gel-derived materials. Accounts Chem. Res., 40 (9): 738–746.

5. Buntkowsky, G., Breitzke, H., Adamezyk, A., Roelofs, F., Emmler, T., Gedat,E., Grunberg, B., Xu, Y., Limbach, H.H., Shenderovich, J., Vyalikh, A., andFindenegg, G. (2007) Structural and dynamical properties of guest molecules

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 80: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

78 M. Geppi et al.

confined in mesoporous silica materials revealed by NMR. Phys. Chem. Chem.Phys., 9 (35): 4843–4853.

6. Albert, K. (2003) NMR investigations of stationary phases. J. Sep. Sci., 26 (3–4):215–224.

7. Albert, K. and Bayer, E. (1991) Characterization of bonded phases by solid-stateNMR spectroscopy. J. Chrom., 544 (1–2): 345–370.

8. Kolodziejski, W. (2005) Solid-state NMR studies of bone. Topics in CurrentChemistry, 246: 235–270.

9. Kuroki, S., Kimura, H., and Ando, I. (2004) Structural characterization of Si-basedpolymer materials by solid-state NMR spectroscopy. Ann. Rep. NMR Spectros.,52: 201–243.

10. Levitt, M.H. (2001) Spin Dynamics. John Wiley & Sons: Chichester, UK.11. Keeler, J. (2005) Understanding NMR Spectroscopy. John Wiley & Sons: Chich-

ester, UK.12. Duer, M.J. (ed.) (2002) Solid-State NMR Spectroscopy—Principles and Applica-

tions. Blackwell Science: Oxford.13. McBrierty, V.J. and Packer, K.J. (1993) Nuclear Magnetic Resonance in Solid

Polymers. Cambridge University Press: Cambridge.14. Schmidt-Rohr, K. and Spiess, H.W. (1994) Multidimensional Solid-State NMR

and Polymers. Academic Press: London.15. Saito, H., Ando, I., and Naito, A. (2006) Solid State NMR Spectroscopy for

Biopolymers: Principles and Applications. Springer: Dordrecht, The Netherlands.16. Harris, R.K. and Wasylishen, R.E., Eds. Encyclopedia of Magnetic Reso-

nance [online]. John Wiley & Sons. www.mrw.interscience.wiley.com/emrw/9780470034590/home/

17. Webb, G.A. (ed.) (2006) Modern Magnetic Resonance. Springer: Dordrecht, TheNetherlands.

18. Geppi, M., Mollica, G., Borsacchi, S., and Veracini, C.A. (2008) Solid-state NMRstudies of pharmaceutical systems. Appl. Spectros. Rev., 43 (3): 202–302.

19. Engelhardt, G. Silicon-29 NMR of solid silicates. In Encyclopedia of MagneticResonance [online], Harris, R.K. and Wasylishen, R.E., Eds. John Wiley & Sons.DOI: 10.1002/9780470034590.emrstm0506

20. Engelhardt, G. and Jancke, H. (1981) Structure investigation of organosiliconpolymers by silicon-29 NMR. Polymer Bull., 5 (11–12): 577–584.

21. Pursch, M., Sander, L.C., and Albert, K. (1999) Understanding reversed-phaseLC with solid-state NMR. Analytical Chemistry News & Features, 733A–741A.

22. Brus, J. (2002) Solid-state NMR study of phase separation and order of watermolecules and silanol groups in polysiloxane networks. J. Sol-Gel Sci. Tech., 25(1): 17–28.

23. Brus, J., Spırkova, M., Hlavata, D., and Strachota, A. (2004) Self-organization,structure, dynamic properties, and surface morphology of silica/epoxy films asseen by solid-state NMR, SAXS, and AFM. Macromolecules, 37 (4): 1346–1357.

24. Marini, M., Toselli, M., Borsacchi, S., Mollica, G., Geppi, M., and Pilati, F.(2008) Facile synthesis of core-shell organic-inorganic hybrid nanoparticles withamphiphilic polymer shell by one-step sol-gel reactions. J. Polymer Sci. Part A:Polymer Chem., 46 (5): 1699–1709.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 81: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 79

25. Geppi, M., Mollica, G., Borsacchi, S., Marini, M., Toselli, M., and Pilati, F. (2007)Solid-state NMR characterization of PE-PEG/silica hybrid materials prepared bymicrowave-assisted sol-gel process. J. Mater. Res., 22 (12): 3516–3525.

26. Apperley, D., Hay, J.N., and Raval, H.M. (2002) Silica-dimethylsiloxane hybrids-non-hydrolytic sol-gel synthesis and characterization by NMR spectroscopy.Chem. Mater., 14 (3): 983–988.

27. Deng, Q., Jarrett, W., Moore, R.B., and Mauritz, K.A. (1996) NovelNafion R© /ORMOSIL hybrids via in situ sol-gel-reactions: 2. Probe of ORMOSILphase nanostructure by 29Si solid state NMR spectroscopy. J. Sol-Gel Sci. Tech.,7 (3): 177–190.

28. Framery, E. and Mutin, P.H. (2002) 29Si MAS-NMR study of silica gels andxerogels: influence of the catalyst. J. Sol-Gel Sci. Tech., 24 (3): 191–195.

29. Joseph, R., Zhang, S.M., and Ford, W.T. (1996) Structure and dynamics of acolloidal silica-poly(methyl methacrylate) composite by 13C and 29Si MAS NMRspectroscopy. Macromolecules, 29 (4): 1305–1312.

30. Li, C.H., Glass, T., and Wilkes, G.L. (1999) NMR studies of sol-gel derived hybridmaterials based on triethoxysilylated diethylenetriamine and tetramethoxysilane.J. Inorg. Organomet. Polymer., 9 (2): 79–106.

31. Wang, G.P., Chang, T.C., Hong, Y.S., and Chiu, Y.S. (2002) Dynamics of novelhydrogen-bonded acidic fluorinated poly (amide-imide-silica) hybrids studied bysolid-state NMR. Polymer, 43 (8): 2191–2200.

32. Comotti, A., Bracco, S., Valsesia, P., Ferretti, L., and Sozzani, P. (2007) 2D mult-inuclear NMR, hyperpolarized xenon and gas storage in organosilica nanochan-nels with crystalline order in the walls. J. Am. Chem. Soc., 129 (27): 8566–8576.

33. Kao, H.M., Chang, P.C., Wu, J.D., Chiang, A.S.T., and Lee, C.H. (2006) Directsynthesis, characterization and solid-state NMR spectroscopy of large-pore vinyl-functionalized cubic mesoporous silica FDU-12. Microporous and MesoporousMaterials, 97 (1–3): 9–20.

34. Bauer, F., Ernst, H., Decker, U., Findeisen, M., Glasel, H.J., Langguth, H.,Hartmann, E., Mehnert, R., and Peuker, C. (2000) Preparation of scratch andabrasion resistant polymeric nanocomposites by monomer grafting onto nanopar-ticles, 1. FTIR and multi-nuclear NMR spectroscopy to the characterization ofmethacryl grafting. Macromol. Chem. Phys., 201 (18): 2654–2659.

35. Herrera, N.N., Letoffe, J.-M., Putaux, J.-L., David, L., and Bourgeat-Lami, E.(2004) Aqueous dispersions of silane-functionalized laponite clay platelets. Afirst step toward the elaboration of water-based polymer/clay nanocomposites.Langmuir, 20 (5): 1564–1571.

36. Herrera, N.N., Letoffe, J.-M., Reymond, J.-P., and Bourgeat-Lami, E. (2005)Silylation of laponite clay particles with monofunctional and trifunctional vinylalkoxysilanes. J. Mater. Chem., 15 (8): 863–871.

37. Wheeler, P.A., Wang, J., Baker, J., and Mathias, L.J. (2005) Synthesis and char-acterization of covalently functionalized laponite clay. Chem. Mater., 17 (11):3012–3018.

38. Borsacchi, S., Geppi, M., Ricci, L., Ruggeri, G., and Veracini, C.A. (2007) Interac-tions at the surface of organophilic-modified laponites: A multinuclear solid-stateNMR study. Langmuir, 23 (7): 3953–3960.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 82: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

80 M. Geppi et al.

39. Borsacchi, S., Geppi, M., Veracini, C.A., Lazzeri, A., Di Cuia, F., and Geloni, C.(2008) A multinuclear solid-state magnetic resonance study of the interactionsbetween the inorganic and organic coatings of BaSO4 submicronic particles.Magn. Reson. Chem., 46 (1): 52–57.

40. Lynch, B., Glennon, J.D., Troltzsch, C., Menyes, U., Pursch, M., and Albert, K.(1997) A (-)-menthyl bonded silica phase for chiral separations: Synthesis andsolid state NMR characterization. Anal. Chem., 69 (9): 1756–1762.

41. Pursch, M., Brindle, R., Ellwanger, A., Sander, L.C., Bell, C.M., Handel, H., et al.(1997) Stationary interphases with extended alkyl chains: A comparative study onchain order by solid-state NMR spectroscopy. Solid State Nucl. Magn. Reson., 9(2–4): 191–201.

42. Pursch, M., Jager, A., Schneller, T., Brindle, R., Albert, K., and Lindner, E.(1996) The sol-gel method: A new way to reversed phase materials. Synthesis andcharacterization by solid-state NMR spectroscopy. Chem. Mater., 8 (6): 1245–1249.

43. Pursch, M., Sander, L.C., and Albert, K. (1996) Chain order and mobility of high-density C18 phases by solid-state NMR spectroscopy and liquid chromatography.Anal. Chem., 68 (23): 4107–4113.

44. Lindner, E., Kemmler, M., Mayer, H.A., and Wegner, P. (1994) Polysiloxane-bound ether phosphines and ruthenium complexes. A characterization by solid-state NMR-spectroscopy and catalysis. J. Am. Chem. Soc., 116 (1): 348–361.

45. Armelao, L., Gross, S., Muller, K., Pace, G., Tondello, E., Tsetsgee, O., and Zattin,A. (2006) Structural evolution upon thermal heating of nanostructured inorganic-organic hybrid materials to binary oxides MO2-SiO2 (M = Hf, Zr) as evalu-ated by solid-state NMR and FTIR spectroscopy. Chem. Mater., 18 (25): 6019–6030.

46. Templin, M., Wiesner, U., and Spiess, H.W. (1997) Multinuclear solid-state-NMRstudies of hybrid organic-inorganic materials. Adv. Mater., 9 (10): 814–817.

47. Alonso, B., Massiot, D., Babonneau, F., Brusatin, G., Della Giustina, G., Kid-chob, T., and Innocenzi, P. (2005) Structural control in germania hybrid organic-inorganic materials. Chem. Mater., 17 (12): 3172–3180.

48. Hoebbel, D., Nacken, M., and Schmidt, H. (1998) A NMR study on the hy-drolysis, condensation and epoxide ring-opening reaction in sols and gels of thesystem glycidoxypropyltrimethoxysilane-water-titaniumtetraethoxide. J. Sol-GelSci. Tech., 12 (3): 169–179.

49. Borsacchi, S., Geppi, M., Veracini, C.A., Fallani, F., Ricci, L., and Ruggeri,G. (2006) Improving compatibility in LDPE-silica dispersions by photo-graftingreaction. Preparation and solid state NMR investigation. J. Mater. Chem., 16 (47):4581–4591.

50. Bauer, F., Glasel, H.J., Decker, U., Ernst, H., Freyer, A., Hartmann, E., Sauerland,V., and Mehnert, R. (2003) Trialkoxysilane grafting onto nanoparticles for thepreparation of clear coat polyacrylate systems with excellent scratch performance.Progr. Org. Coating., 47 (2): 147–153.

51. Baccile, N., Laurent, G., Bonhomme, C., Innocenzi, P., and Babonneau, F. (2007)Solid-state NMR characterization of the surfactant-silica interface in templatedsilicas: Acidic versus basic conditions. Chem. Mater., 19 (6): 1343–1354.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 83: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 81

52. Gerstein, B.C., Pembleton, R.G., Wilson, L.C., and Ryan, L.M. (1977) High-resolution NMR in randomly oriented solids with homonuclear dipolar broaden-ing: Combined multiple pulse NMR and magic angle spinning. J. Chem. Phys.,66 (1): 361–362.

53. Bronnimann, C.E., Zeigler, R.C., and Maciel, G.E. (1988) Proton NMR study ofdehydration of the silica gel surface. J. Am. Chem. Soc., 110 (7): 2023–2026.

54. Shenderovich, I.G., Buntkowsky, G., Schreiber, A., Gedat, E., Sharif, S., Albrecht,J., Golubev, N.S., Findenegg, N.H., and Limbach, H.-H. (2003) Pyridine-15N—Amobile NMR sensor for surface acidity and surface defects of mesoporous silica.J. Phys. Chem. B, 107 (43): 11924–11939.

55. Brus, J. and Skrdlantova, M. (2001) 1H MAS NMR study of structure of hybridsiloxane-based networks and the interaction with quartz filler. J. Non-CrystallineSolids, 281 (1–3): 61–71.

56. Schenkel, R., Jentys, A., Parker, S.F., and Lercher, J.A. (2004) INS and IR andNMR spectroscopic study of C1-C4 alcohols adsorbed on alkali metal-exchangedzeolite X. J. Phys. Chem. B, 108 (39): 15013–15026.

57. Camus, L., Goletto, V., Maquet, J., Gervais, C., Bonhomme, C., and Babonneau,F. (2003) 1H MAS-NMR at high spinning rate: An interesting characterizationtool for hybrid materials. J. Sol-Gel Sci. Tech., 26 (1–3): 311–314.

58. Bohlmann, W. and Michel, D. (2001) 1H MAS NMR studies of molecules ad-sorbed on activated zeolites. J. Catal., 202 (2): 421–426.

59. Simperler, A., Bell, R.G., and Anderson, M.W. (2004) Probing the acid strengthof Brønsted acidic zeolites with acetonitrile: Quantum chemical calculation of 1H,15N, and 13C NMR shift parameters. J. Phys. Chem. B, 108 (22): 7142–7151.

60. Boiadjiev, V., Blumenfeld, A., Gutow, J., and Tysoe, W.T. (2000) Infrared andNMR spectroscopic studies of n-alkanethiols chemically grafted on dimethylzinc-modified silica surfaces. Chem. Mater., 12 (9): 2604–2613.

61. Pawsey, S., Yach, K., Halla, J., and Reven, L. (2000) Self-assembled monolayersof alkanoic acids: A solid-state NMR study. Langmuir, 16 (7): 3294–3303.

62. Pan, V.H., Tao, T., Zhou, J.W., and Maciel, G.E. (1999) Hydrogen bonding betweenacetone and silica gel, as studied by NMR. J. Phys. Chem. B, 103 (33): 6930–6943.

63. Gao, W., Dickinson, L., Grozinger, C., Morin, F.G., and Reven, L. (1996) Self-assembled monolayers of alkylphosphonic acids on metal oxides. Langmuir, 12(26): 6429–6435.

64. Zhang, L.M., Xu, J., Hou, G.J., Tang, H.R., and Deng, F. (2007) Interactionsbetween Nafion resin and protonated dodecylamine modified montmorillonite: Asolid state NMR study. J. Colloid Interface Sci., 311 (1): 38–44.

65. Burum, D.P. and Bielecki, A. (1991) An improved experiment for heteronuclear-correlation 2D NMR in solids. J. Magn. Reson., 94 (3): 645–652.

66. Caravatti, P., Bodenhausen, G., and Ernst, R.R. (1982) Heteronuclear solid-statecorrelation spectroscopy. Chem. Phys. Lett., 89 (5): 363–367.

67. Caravatti, P., Braunschweiler, L., and Ernst, R.R. (1983) Heteronuclear correlationspectroscopy in rotating solids. Chem. Phys. Lett., 100 (4): 305–310.

68. van Rossum, B.-J., Forster, H., and de Groot, H.J.M. (1997) High-field and high-speed CP-MAS 13C NMR heteronuclear dipolar-correlation spectroscopy of solidswith frequency-switched Lee-Goldburg homonuclear decoupling. J. Magn. Re-son., 124 (2): 516–519.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 84: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

82 M. Geppi et al.

69. Trebosc, J., Wiench, J.W., Huh, S., Lin, V.S.Y., and Pruski, M. (2005) Studiesof organically functionalized mesoporous silicas using heteronuclear solid-statecorrelation NMR spectroscopy under fast magic angle spinning. J. Am. Chem.Soc., 127 (20): 7587–7593.

70. Wiench, J.W., Avadhut, Y.S., Maity, N., Bhaduri, S., Lahiri, G.K., Pruski, M.,and Ganapathy, S. (2007) Characterization of covalent linkages in organicallyfunctionalized MCM-41 mesoporous materials by solid-state NMR and theoreticalcalculations. J. Phys. Chem. B, 111 (15): 3877–3885.

71. Rawal, A., Urman, K., Otaigbe, J.U., and Schmidt-Rohr, K. (2006) Detection ofnanometer-scale mixing in phosphate-glass/polyamide-6 hybrids by 1H-31P NMR.Chem. Mater., 18 (26): 6333–6338.

72. Rottstegge, J., Wilhelm, M., and Spiess, H.W. (2006) Solid state NMR investiga-tions on the role of organic admixtures on the hydration of cement pastes. CementConcr. Compos., 28 (5): 417–426.

73. Schnell, I., Brown, S.P., Low, H.Y., Ishida, H., and Spiess, H.W. (1998) Aninvestigation of hydrogen bonding in benzoxazine dimers by fast magic-anglespinning and double-quantum 1H NMR spectroscopy. J. Am. Chem. Soc., 120(45): 11784–11795.

74. Feike, M., Demco, D.E., Graf, R., Gottwald, J., Hafner, S., and Spiess, H.W.(1996) Broadband multiple-quantum NMR spectroscopy. J. Magn. Reson. A, 122(2): 214–221.

75. Brown, S.P. and Spiess, H.W. (2001) Advanced solid-state NMR methods forthe elucidation of structure and dynamics of molecular, macromolecular, andsupramolecular systems. Chem. Rev., 101 (12): 4125–4155.

76. Heffner, S.A. and Mirau, P.A. (1994) Identification of intermolecular interactionsin 1,2-polybutadiene/polyisoprene blends. Macromolecules, 27 (25): 7283–7286.

77. Arrachart, G., Carcel, C., Moreau, J.J.E., Hartmeyer, G., Alonso, B., Massiot, D.,Creff, G., Bantignies, J.-L., Dieudonne, P., Man, M.W.C., Althoff, G., Babonneau,F., and Bonhomme, C. (2008) Organosilicas based on purine-pyrimidine base pairassemblies: A solid state NMR point of view. J. Mater. Chem., 18 (4): 392–399.

78. Gullion, R. and Schaefer, J. (1989) Rotational-echo double-resonance NMR. J.Magn. Reson., 81 (1): 196–200.

79. Jaeger, C., Groom, N.S., Bowe, E.A., Horner, A., Davies, M.E., Murray, R.C.,and Duer, M.J. (2005) Investigation of the nature of the protein-mineral interfacein bone by solid-state NMR. Chem. Mater., 17 (12): 3059–3061.

80. Hou, S.S., Bonagamba, T.J., Beyer, F.L., Madison, P.H., and Schmidt-Rohr, K.(2003) Clay intercalation of poly(styrene-ethylene oxide) block copolymers stud-ied by two-dimensional solid-state NMR. Macromolecules, 36 (8): 2769–2776.

81. Hou, S.S., Beyer, F.L., and Schmidt-Rohr, K. (2002) High-sensitivity multinuclearNMR spectroscopy of a smectite clay and of clay-intercalated polymer. Solid StateNucl. Magn. Reson., 22 (2–3): 110–127.

82. Schmidt-Rohr, K., Clauss, J., and Spiess, H.W. (1992) Correlation of structure,mobility, and morphological information in heterogeneous polymer materials bytwo-dimensional wideline-separation NMR spectroscopy. Macromolecules, 25(12): 3273–3277.

83. Bose, A.B., Gangoda, M., Jaroniec, M., Gilpin, R.K., and Bose, R.N. (2006) Two-dimensional solid state NMR characterization of physisorbed siloxane polymer(OV-225) on silica. Surf. Sci., 600 (1): 143–154.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 85: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 83

84. Haouas, M., Harrane, A., Belbachir, M., and Taulelle, F. (2007) Solid state NMRcharacterization of formation of poly(epsilon-caprolactone)/maghnite nanocom-posites by in situ polymerization. J. Polymer Sci. Part B: Polymer Phys., 45 (22):3060–3068.

85. Bertmer, M., Wang, M.F., Kruger, M., Blumich, B., Litvinov, V.M., and van Es, M.(2007) Structural changes from the pure components to nylon 6-montmorillonitenanocomposites observed by solid-state NMR. Chem. Mater., 19 (5): 1089–1097.

86. Medek, A., Harwood, J.S., and Frydman, L. (1995) Multiple-quantum magic-angle spinning NMR: A new method for the study of quadrupolar nuclei in solids.J. Am. Chem. Soc., 117 (51): 12779–12787.

87. Fernandez, C., Amoreux, J.P., Chezeau, J.M., Delmotte, L., and Kessler, H. (1996)27Al MAS NMR characterization of AlPO4-14 enhanced resolution and informa-tion by MQMAS. Microporous Mater., 6 (5–6): 331–340.

88. Janicke, M.T., Landry, C.C., Christiansen, S.C., Birtalan, S., Stucky, G. D., andChmelka, B.F. (1999) Low silica MCM-41 composites and mesoporous solids.Chem. Mater., 11 (5): 1342–1351.

89. Lafond, V., Gervais, C., Maquet, J., Prochnow, D., Babonneau, F., and Mutin,P.H. (2003) 17O MAS NMR study of the bonding mode of phosphonate couplingmolecules in a titanium oxo-alkoxo-phosphonate and in titania-based hybrid ma-terials. Chem. Mater., 15 (21): 4098–4103.

90. Julian, B., Gervais, C., Rager, M.-N., Maquet, J., Cordoncillo, E., Escribano, P.,Babonneau, F., and Sanchez, C. (2004) Solid-state 17O NMR characterization ofPDMS-MxOy (M = Ge(IV), Ti(IV), Zr(IV), Nb(V), and Ta(V)) organic-inorganicnanocomposites. Chem. Mater., 16 (3): 521–529.

91. Larsen, F.H., Jakobsen, H.J., Ellis, P.D., and Nielsen, N.C. (1997) Sensitivity-enhanced quadrupolar-echo NMR of half-integer quadrupolar nuclei. Magnitudesand relative orientation of chemical shielding and quadrupolar coupling tensors.J. Phys. Chem. A, 101 (46): 8597–8606.

92. Hung, I. and Schurko, R.W. (2004) Solid-state 91Zr NMR of bis(cyclopentadienyl)dichlorozirconium(IV). J. Phys. Chem. B, 108 (26): 9060–9069.

93. Bartolotta, A., Forte, C., Geppi, M., Minniti, D., and Visalli, G. (1997) Solid stateNMR study of sodium thiocyanate/poly(ethylene oxide) electrolytes. Solid StateNucl. Magn. Reson., 8 (4): 231–239.

94. Samoson, A. and Lippmaa, E. (1983) Central transition NMR excitation spectraof half-integer quadrupole nuclei. Chem. Phys. Lett., 100 (3): 205–208.

95. Samoson, A. and Lippmaa, E. (1983) Excitation phenomena and line intensitiesin high-resolution NMR powder spectra of half-integer quadrupolar nuclei. Phys.Rev. B, 28 (11): 6567–6570.

96. Chung, S.H., Jeffrey, K.R., and Stevens, J.R. (1998) Dynamics of sodium ions inNaClO4 complexed in poly(propylene-oxide): A 23Na nuclear magnetic resonancestudy. J. Chem. Phys., 108 (8): 3360–3372.

97. Kao, H.M., Hung, T.T., and Fey, G.T.K. (2007) Multinuclear solid-state NMRcharacterization, ion dissociation, and dynamic properties of lithium-dopedorganic-inorganic hybrid electrolytes based on ureasils. Macromolecules, 40 (24):8673–8683.

98. Kao, H.M., Chao, S.W., and Chang, P.C. (2006) Multinuclear solid-stateNMR, self-diffusion coefficients, differential scanning calorimetry, and ionic

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 86: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

84 M. Geppi et al.

conductivity of solid organic-inorganic hybrid electrolytes based on PPG-PEG-PPG diamine, siloxane, and lithium perchlorate. Macromolecules, 39 (3): 1029–1040.

99. Kao, H.M., Chang, P.C., Chao, S.W., and Lee, C.H. (2006) 7Li NMR, ionic con-ductivity and self-diffusion coefficients of lithium ion and solvent of plasticizedorganic-inorganic hybrid electrolyte based on PPG-PEG-PPG diamine andalkoxysilanes. Electrochim. Acta, 52 (3): 1015–1027.

100. de Souza, P.H., Bianchi, R.F., Dahmouche, K., Judeinstein, P., Faria, R.M., andBonagamba, T.J. (2001) Solid-state NMR, ionic conductivity, and thermal studiesof lithium-doped siloxane-poly(propylene glycol) organic-inorganic nanocom-posites. Chem. Mater., 13 (10): 3685–3692.

101. Mello, N.C., Bonagamba, T.J., Panepucci, H., Dahmouche, K., Judeinstein, P., andAegerter, M.A. (2000) NMR study of ion-conducting organic-inorganic nanocom-posites poly(ethylene glycol)-silica-LiClO4. Macromolecules, 33 (4): 1280–1288.

102. Stejskal, E.O. and Tanner, J.E. (1965) Spin diffusion measurements: Spin echoesin the presence of a time-dependent field gradient. J. Chem. Phys., 42 (1): 288–292.

103. Raitza, M., Herold, M., Ellwanger, A., Gauglitz, G., and Albert, K. (2000) Solid-state NMR and ellipsometric investigations of C-30 chains bonded to SiO2 sur-faces. Macromol. Chem. Phys., 201 (7): 825–829.

104. Albert, K., Lacker, T., Raitza, M., Pursch, M., Egelhaaf, H.-J., and Oelkrug, D.(1998) Investigating the selectivity of triacontyl interphases. Angew. Chem. Int.Ed., 37 (6): 778–780.

105. Khimyak, Y.Z. and Klinowski, J. (2001) Solid-state NMR studies of the organictemplate in mesostructured aluminophosphates. Phys. Chem. Chem. Phys., 3 (4):616–626.

106. Muller, R., Hrobarikova, J., Calberg, C., Jerome, R., and Grandjean, J. (2004)Structure and dynamics of cationic surfactants intercalated in clays. Langmuir, 20(7): 2982–2985.

107. Kubies, D., Jerome, R., and Grandjean, J. (2002) Surfactant molecules intercalatedin Laponite as studied by 13C and 29Si MAS NMR. Langmuir, 18 (16): 6159–6163.

108. Grandjean, J., Bujdak, J., and Komadel, P. (2003) NMR study of moleculesintercalated in montmorillonite and in silylated montmorillonite. Clay Miner.,38 (3): 367–373.

109. Wen, X.Y., He, H.P., Zhu, J.X., Jun, Y., Ye, C.H., and Deng, F. (2006) Arrange-ment, conformation, and mobility of surfactant molecules intercalated in mont-morillonite prepared at different pillaring reagent concentrations as studied bysolid-state NMR spectroscopy. J. Colloid Interface Sci., 299 (2): 754–760.

110. Zhu, J.X., He, H.P., Zhu, L.Z., Wen, X.Y., and Deng, F. (2005) Characterizationof organic phases in the interlayer of montmorillonite using FTIR and 13C NMR.J. Colloid Interface Sci., 286 (1): 239–244.

111. Osman, M.A., Ernst, M., Meier, B.H., and Suter, U.W. (2002) Structure andmolecular dynamics of alkane monolayers self-assembled on mica-platelets. J.Phys. Chem. B, 106 (3): 653–662.

112. Simonutti, R., Comotti, A., Bracco, S., and Sozzani, P. (2001) Surfactant organi-zation in MCM-41 mesoporous materials as studied by 13C and 29Si solid-stateNMR. Chem. Mater., 13 (3): 771–777.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 87: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 85

113. Azaıs, T., Tourne-Peteilh, C., Aussenac, F., Baccile, N., Coelho, C., Devoisselle,J.-M., and Babonneau, F. (2006) Solid-state NMR study of Ibuprofen confined inMCM-41 material. Chem. Mater., 18 (26): 6382–6390.

114. Tang, X.-P., Cheong Ng, C., Nguyen, H., Mogilevsky, G., and Wu, Y. (2008)The molecular dynamics and melting transition of confined ibuprofen in titaniananotube studied by NMR. Chem. Phys. Lett., 452 (4–6): 289–295.

115. Geppi, M., Guccione, S., Mollica, G., Pignatello, R., and Veracini, C.A. (2005)Molecular properties of Ibuprofen and its solid dispersions with Eudragit RL100studied by solid-state nuclear magnetic resonance. Pharmaceut. Res., 22 (9):1544–1555.

116. Gambogi, J.E. and Blum, F.D. (1992) Molecular mobility of the interface in amodel polymer composite—A NMR study. Macromolecules, 25 (18): 4526–4534.

117. Thangaraj, A., Kumar, R., and Ratnasamy, P. (1991) NMR study of the location andconfiguration of pyrrolidine in ZSM-5, ZSM-23, ZSM-48, and ZSM-51 zeolites.Zeolites, 11 (6): 573–576.

118. Satozawa, M., Kunimori, K., and Hayashi, S. (1997) 13C and 1H MAS NMRstudy of benzene and p-xylene in zeolites and a mesoporous material FSM-16.Bull. Chem. Soc. Jpn., 70 (1): 97–105.

119. Simperler, A., Bell, R.G., Philippou, A., and Anderson, M.W. (2002) Theoret-ical study of toluene adsorbed on zeolites X and Y: Calculation of 13C NMRparameters. J. Phys. Chem. B, 106 (42): 10944–10954.

120. Zheng, A.M., Chen, L., Yang, J., Yue, Y., Ye, C.H., Xin, L.B., et al. (2005)Prediction of the 13C NMR chemical shifts of organic species adsorbed on H-ZSM-5 zeolite by the ONIOM-GIAO method. Chem. Commun. (19): 2474–2476.

121. Gunther, H., Oepen, S., Ebener, M., and Francke, V. (1999) NMR study of adsorp-tion techniques for organic compounds on silica surfaces. Magn. Reson. Chem.,37 (13): S142–S146.

122. O’Brien, P., Williamson, C.J., and Groombridge, C.J. (1991) Multinuclear solid-state MAS and CP-MAS NMR-study of the binding of triethyl phosphate to amontmorillonite. Chem. Mater., 3 (2): 276–280.

123. Ukrainczyk, L. and Smith, K.A. (1996) Solid state 15N NMR study of pyridineadsorption on clay minerals. Environ. Sci. Tech., 30 (11): 3167–3176.

124. Brindle, R., Pursch, M., and Albert, K. (1996) 1H MAS NMR spectroscopy ofchemically modified silica gels: A fast method to characterize stationary inter-phases for chromatography. Solid State Nucl. Magn. Reson., 6 (3): 251–266.

125. Borsacchi, S., Geppi, M., Iuliano, A., and Veracini, C.A. (2005) Solid-state NMRcharacterization of diastereoisomeric chiral stationary phases and their solublemodels. Solid State Nucl. Magn. Reson., 28 (2–4): 193–203.

126. Mielczarski, J.A., Cases, J.M., Tekely, P., and Canet, D. (1993) Nature and struc-ture of adsorption layer on apatite contacted with oleate solutions. 2. In-situ andex-situ Fourier-transform infrared, NMR, and X-ray photoelectron-spectroscopystudies. Langmuir, 9 (12): 3357–3370.

127. Fyfe, C.A., Diaz, A.C., Grondey, H., Lewis, A.R., and Forster, H. (2005) Solid stateNMR method for the determination of 3D zeolite framework/sorbate structures:1H/29Si CP MAS NMR study of the high-loaded form of p-xylene in ZSM-5 anddetermination of the unknown structure of the low-loaded form. J. Am. Chem.Soc., 127 (20): 7543–7558.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 88: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

86 M. Geppi et al.

128. Fyfe, C.A. and Brouwer, D.H. (2006) Optimization, standardization, and testingof a new NMR method for the determination of zeolite host-organic guest crystalstructures. J. Am. Chem. Soc., 128 (36): 11860–11871.

129. Fyfe, C.A. and Lee, J.S.J. (2008) Solid-state NMR determination of the zeoliteZSM-5/ortho-xylene host-guest crystal structure. J. Phys. Chem. C, 112 (2): 500–513.

130. Komori, Y. and Hayashi, S. (2004) Dynamics of p-nitroaniline molecules in FSM-type mesoporous silicas studied by solid-state NMR. Microporous and Meso-porous Materials, 68 (1–3): 111–118.

131. Komori, Y. and Hayashi, S. (2004) Effects of Na+ on dynamics of p-nitroanilinemolecules in zeolite ZSM-5 studied by solid-state NMR. Bull. Chem. Soc. Jpn.,77 (4): 673–679.

132. Kustanovich, I., Fraenkel, D., Luz, Z., and Vega, S. (1988) Dynamic propertiesof p-xylene adsorbed on Na-ZSM-5 zeolite by deuterium and proton magic-anglespinning NMR. J. Phys. Chem., 92 (11): 4134–4141.

133. Sato, T., Kunimori, K., and Hayashi, S. (1999) Dynamics of benzene, cyclohexaneand n-hexane in KL zeolite studied by 2H NMR. Phys. Chem. Chem. Phys., 1 (16):3839–3843.

134. Soderlind, E. and Blum, F.D. (1993) Structure of sodium 4-(1′-heptylnonyl) ben-zenesulfonate adsorbed on alumina using 2H NMR. J. Colloid Interface Sci., 157(1): 172–179.

135. Kelusky, E.C. and Fyfe, C.A. (1986) Molecular motions of alkoxysilanes immo-bilized on silica surfaces: A deuterium NMR study. J. Am. Chem. Soc., 108 (8):1746–1749.

136. Shantz, D. F. and Lobo, R. F. (1998) Solid-state deuterium NMR studies of organicmolecules in the tectosilicate nonasil. J. Phys. Chem. B, 102 (13): 2339–2349.

137. Komori, Y. and Hayashi, S. (2003) Dynamics of p-nitroaniline molecules insiliceous ZSM-5 studied by solid-state NMR. Phys. Chem. Chem. Phys., 5 (17):3777–3783.

138. Zeigler, R.C. and Maciel, G.E. (1991) A study of the structure and dynamicsof dimethyloctadecysilyl-modified silica using wide-line 2H-NMR techniques. J.Am. Chem. Soc., 113 (17): 6349–6358.

139. Moe, N.E., Qiu, X.H., and Ediger, M.D. (2000) 13C NMR study of segmentaldynamics in atactic polypropylene. Macromolecules, 33 (6): 2145–2152.

140. Qiu, X.H., Moe, N.E., Ediger, M.D., and Fetters, L.J. (2000) Local and globaldynamics of atactic polypropylene melts by multiple field 13C NMR. J. Chem.Phys., 113 (7): 2918–2926.

141. Forte, C., Geppi, M., Malvaldi, M., and Mattoli, V. (2004) Dynamics of anamorphous polymer by an improved NMR approach based on the simultane-ous analysis of 1H and 13C relaxation times. J. Phys. Chem. B, 108 (30): 10832–10837.

142. Booth, H.F. and Strange, J.H. (1998) Organic nanocrystals: an NMR study ofcyclohexane in porous silica. Mol. Phys., 93 (2): 263–269.

143. Mitchell, J. and Strange, J.H. (2004) An NMR investigation of naphthalene nanos-tructures. Mol. Phys., 102 (19–20): 1997–2005.

144. Strange, J.H., Rahman, M., and Smith, E.G. (1993) Characterization of poroussolids by NMR. Phys. Rev. Lett., 71 (21): 3589–3591.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 89: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 87

145. Schaefer, D.J., Favre, D.E., Wilhelm, M., Weigel, S.J., and Chmelka, B.F. (1997)Site-hopping dynamics of benzene adsorbed on Ca-LSX zeolite studied by solid-state exchange 13C NMR. J. Am. Chem. Soc., 119 (39): 9252–9267.

146. Yang, Y., Schuster, M., Blumich, B., and Spiess, H.W. (1987) Dynamic magic-angle spinning NMR spectroscopy: Exchange-induced sidebands. Chem. Phys.Lett., 139 (3–4): 239–243.

147. Fyfe, C.A. and Diaz, A.C. (2002) Investigation of slow molecular motions andchemical exchange in the high loaded form of p-xylene in ZSM-5 by two-dimensional 13C solid-state NMR exchange experiments. J. Phys. Chem. B, 106(9): 2261–2268.

148. Szeverenyi, N.M., Sullivan, M.J., and Maciel, G.E. (1982) Observation of spinexchange by two-dimensional Fourier transform 13C cross polarization/magic-angle spinning. J. Magn. Reson., 47 (3): 462–475.

149. VanderHart, D.L., Asano, A., and Gilman, J.W. (2001) Solid-state NMR investiga-tion of paramagnetic nylon-6 clay nanocomposites. 1. Crystallinity, morphology,and the direct influence of Fe3+ on nuclear spins. Chem. Mater., 13 (10): 3781–3795.

150. Mathias, L.J., Davis, R.D., and Jarrett, W.L. (1999) Observation of α and γ crystalforms and amorphous regions of nylon 6-clay nanocomposites using solid-state15N nuclear magnetic resonance. Macromolecules, 32 (23): 7958–7960.

151. Park, J.W., Seo, Y.A., Kim, I., Ha, C.S., Aimi, K., and Ando, S. (2004) Investigat-ing the crystalline structure of poly(vinylidene fluoride) (PVDF) in PVDF/silicabinary and PVDF/poly(methyl methacrylate)/silica ternary hybrid composites us-ing FTIR and solid-state 19F MAS NMR spectroscopy. Macromolecules, 37 (2):429–436.

152. Panziera, N., Pertici, P., Barazzone, L., Caporusso, A.M., Vitulli, G., Salvadori,P., Borsacchi, S., Geppi, M., Veracini, C.A., Martra, G., and Bertinetti, L. (2007)MVS-derived palladium nanoparticles deposited on polydimethylphosphazene asrecyclable catalysts for Heck-type reactions: Preparation, structural study, andcatalytic activity. J. Catal., 246 (2): 351–361.

153. Schmidt-Rohr, K. (1996) A double-quantum solid-state NMR technique fordetermining torsion angles in polymers. Macromolecules, 29 (11): 3975–3981.

154. Harris, D.J., Bonagamba, T.J., and Schmidt-Rohr, K. (1999) Conformation ofpoly(ethylene oxide) intercalated in clay and MoS2 studied by two-dimensionaldouble-quantum NMR. Macromolecules, 32 (20): 6718–6724.

155. Litvinov, V.M., Barthel, H., and Weis, J. (2002) Structure of a PDMS layergrafted onto a silica surface studied by means of DSC and solid-state NMR.Macromolecules, 35 (11): 4356–4364.

156. ten Brinke, J.W., Litvinov, V.M., Wijnhoven, J.E.G.J., and Noordermeer, J.W.M.(2002) Interactions of Stober silica with natural rubber under the influence ofcoupling agents, studied by 1H NMR T2 relaxation analysis. Macromolecules, 35(27): 10026–10037.

157. De Paul, S.M., Zwanziger, J.W., Ulrich, R., Wiesner, U., and Spiess, H.W. (1999)Structure, mobility, and interface characterization of self-organized organic-inorganic hybrid materials by solid-state NMR. J. Am. Chem. Soc., 121 (24):5727–5736.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 90: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

88 M. Geppi et al.

158. Tajouri, T. and Hommel, H. (2007) Dynamics and conformations of PEO chainschemically bonded on silica: Comparison between 1H and 2H NMR. Magn. Reson.Chem., 45 (6): 488–495.

159. Lin, W.Y. and Blum, F.D. (1997) Segmental dynamics of bulk and adsorbedpoly(methyl acrylate)-d3 by deuterium NMR: Effect of adsorbed amount. Macro-molecules, 30 (18): 5331–5338.

160. Lin, W.Y. and Blum, F.D. (1998) Segmental dynamics of bulk and silica-adsorbedpoly(methyl acrylate)-d3 by deuterium NMR: the effect of molecular weight.Macromolecules, 31 (13): 4135–4142.

161. Lin, W.Y. and Blum, F.D. (2001) Segmental dynamics of interfacial poly(methylacrylate)-d3 in composites by deuterium NMR spectroscopy. J. Am. Chem. Soc.,123 (9): 2032–2037.

162. Metin, B. and Blum, F.D. (2006) Segmental dynamics in poly(methyl acrylate)on silica: molecular-mass effects. J. Chem. Phys., 125 (5): 054707 1–10.

163. Brus, J., Urbanova, M., Kelnar, I., and Kotek, J. (2006) A solid-state NMRstudy of structure and segmental dynamics of semicrystalline elastomer-toughenednanocomposites. Macromolecules, 39 (16): 5400–5409.

164. Brus, J. and Urbanova, M. (2005) Selective measurement of heteronuclear 1H-13Cdipolar couplings in motionally heterogeneous semicrystalline polymer systems.J. Phys. Chem. A, 109(23): 5050–5054.

165. Saalwachter, K. (2007) Proton multiple-quantum NMR for the study of chain dy-namics and structural constraints in polymeric soft materials. Progr. Nucl. Magn.Reson. Spectros., 5 (1): 1–35.

166. Maxwell, R.S., Chinn, S.C., Solyom, D., and Cohenour, R. (2005) Radiation-induced cross-linking in a silica-filled silicone elastomer as investigated by mul-tiple quantum 1H NMR. Macromolecules, 38 (16): 7026–7032.

167. Bathista, A.L.B.S., deAzevedo, E.R., Bloise, A.C., Dahmouche, K., Judeinstein,P., and Bonagamba, T.J. (2007) Effects of lithium doping on the polymer chaindynamics in siloxane/poly(ethylene oxide) ormolyte nanocomposites: A 13C and7Li exchange solid-state NMR study. Chem. Mater., 19 (7): 1780–1789.

168. deAzevedo, E.R., Reichert, D., Vidoto, E.L.G., Dahmouche, K., Judeinstein, P.,and Bonagamba, T.J. (2003) Motional heterogeneities in siloxane/poly(ethyleneglycol) ormolyte nanocomposites studied by 13C solid-state exchange NMR.Chem. Mater., 15 (10): 2070–2078.

169. Yang, D.-K. and Zax, D.B. (2006) Multidimensional 2H NMR study of dynamicalheterogeneity in polymer nanocomposites. Solid State Nucl. Magn. Reson., 29(1–3): 153–162.

170. Forte, C., Geppi, M., Giamberini, S., Ruggeri, G., Veracini, C.A., and Mendez, B.(1998) Structure determination of clay methyl methacrylate copolymer interlayercomplexes by means of 13C solid state NMR. Polymer, 39 (12): 2651–2656.

171. Urbanczyk, L., Hrobarikova, J., Calberg, C., Jerome, R., and Grandjean, J. (2006)Motional heterogeneity of intercalated species in modified clays and poly(ε-caprolactone)/clay nanocomposites. Langmuir, 22 (10): 4818–4824.

172. Calberg, C., Jerome, R., and Grandjean, J. (2004) Solid-state NMR study ofpoly(ε-caprolactone)/clay nanocomposites. Langmuir, 20 (5): 2039–2041.

173. Vanderhart, D.L., Asano, A., and Gilman, J.W. (2001) Solid-state NMR inves-tigation of paramagnetic nylon-6 clay nanocomposites. 2. Measurement of clay

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011

Page 91: Geppi et al, Applied Spectroscopy Reviews, 44, 1–89, 2009

SSNMR of Organic/Inorganic Materials 89

dispersion, crystal stratification, and stability of organic modifiers. Chem. Mater.,13 (10): 3796–3809.

174. Bourbigot, S., Vanderhart, D.L., Gilman, J.W., Awad, W.H., Davis, R.D., Morgan,A.B., and Wilkie, C.A. (2003) Investigation of nanodispersion in polystyrene-montmorillonite nanocomposites by solid-state NMR. J. Polymer Sci. Part B:Polymer Phys., 41 (24): 3188–3213.

175. Bourbigot, S., Vanderhart, D.L., Gilman, J.W., Bellayer, S., Stretz, H., and Paul,D.R. (2004) Solid state NMR characterization and flammability of styrene-acrylonitrile copolymer montmorillonite nanocomposite. Polymer, 45 (22): 7627–7638.

176. Gilman, J.W., Bourbigot, S., Shields, J.R., Nyden, M., Kashiwagi, T., Davis, R.D.,VanderHart, D.L., Demory, W., Wilkie, C.A., Morgan, A.B., Harris, J., and Lyon,R.E. (2003) High throughput methods for polymer nanocomposites research:Extrusion, NMR characterization and flammability property screening. J. Mater.Sci., 38 (22): 4451–4460.

177. Goldman, M. and Shen, L. (1966) Spin-spin relaxation in LaF3. Phys. Rev., 144(1): 321–331.

178. Brus, J. and Dybal, J. (2000) Solid-state NMR study of structure, size and dynam-ics of domains in hybrid siloxane networks. Polymer, 41 (14): 5269–5282.

179. Bertmer, M., Demco, D.E., Wang, M., Melian, C., Marcean-Chelcea, R.I., Fechete,R., Baias, M., and Blumich, B. (2006) Segmental dynamic heterogeneity of short-chain grafted-poly(dimethylsiloxane) by 1H spin-diffusion NMR. Chem. Phys.Lett., 431 (4–6): 404–409.

180. Schmidt-Rohr, K., Rawal, A., and Fang, X.W. (2007) A new NMR method fordetermining the particle thickness in nanocomposites, using T2,H-selective X{1H}recoupling. J. Chem. Phys., 126 (5): 054701 1–16.

181. Harris, R.K., Becker, E.D., Cabral de Menezes, S., Goodfellow, R., and Granger, P.(2001) NMR nomenclature. Nuclear spin properties and conventions for chemicalshifts (IUPAC Recommendations 2001). Pure Appl. Chem., 73 (11): 1795–1818.

Dow

nloa

ded

by [

93.1

08.1

84.9

5] a

t 02:

40 0

8 N

ovem

ber

2011