HR-MAS NMR Spectroscopy in Material Science

In the early to mid-90’s, NMR studies were being published that recognized the power of magic angle spinning (MAS) to increase resolution in materials that were not strictly solids by averaging differences in magnetic susceptibility and residual dipolar coupling inherent in these samples. The method of utilizing MAS for non-solid materials to produce liquid-like NMR lines was termed High-Resolution Magic Angle Spinning (HR-MAS). A few of the first HR-MAS examples included investigation of resins for combinatorial chemistry,[1] solvent swollen polystyrene gels,[2] and lipid systems.[3] Then in 1996, Maas et al. [4] advanced the field of HR-MAS NMR by adding a magnetic field gradient along the magic angle (see Figure 1). Like high resolution solution NMR, this gradient improved sensitivity and resolution with the ability to more easily select coherence pathways and by reducing indirect dimension (t1) noise.[4]


Introduction
In the early to mid-90's, NMR studies were being published that recognized the power of magic angle spinning (MAS) to increase resolution in materials that were not strictly solids by averaging differences in magnetic susceptibility and residual dipolar coupling inherent in these samples. The method of utilizing MAS for non-solid materials to produce liquid-like NMR lines was termed High-Resolution Magic Angle Spinning (HR-MAS). A few of the first HR-MAS examples included investigation of resins for combinatorial chemistry, [1] solvent swollen polystyrene gels, [2] and lipid systems. [3] Then in 1996, Maas et al. [4] advanced the field of HR-MAS NMR by adding a magnetic field gradient along the magic angle (see Figure 1). Like high resolution solution NMR, this gradient improved sensitivity and resolution with the ability to more easily select coherence pathways and by reducing indirect dimension (t1) noise. [4] There are currently commercially available HR-MAS probes with magic angle gradients from companies like Bruker BioSpin Corporation (Billerica, MA), [5] Agilent Technologies (Santa Clara, CA), [6] JEOL USA, Inc. (Peabody, MA), [7] and Doty Scientific, Inc. (Columbia, SC). [8] In addition to magic angle gradients, many of these probes also have a deuterium ( 2 H) lock channel, allowing improved ease of shimming and long term stability. With the emergence of commercially available probes, HR-MAS NMR has become more popular in the last few years, especially in the biological and biomedical fields. This popularity is mainly due to the heterogeneous nature of tissues and cells that are well suited for HR-MAS. Multiple HR-MAS NMR studies involving different tissue biopsies, like brain, kidney, liver, and muscle tissues for metabonomics studies, as well as identification of abnormal tissues (i.e. cancerous tissues) have been reported. [9][10][11] HR-MAS NMR has also been applied to the characterization of foodstuffs, including the assignment of metabolites in tomatoes and apples, the study of biopolymers in fruit cuticles, quantification of n-3 fatty acids content in different fish species, and tracking the chemistry of coffee beans during the

How HR-MAS works
Materials that are crystalline or rigid solids (for example resins, ceramics, etc.) exhibit an extremely broad NMR signal due to extensive homo-and hetero-nuclear dipolar coupling, chemical shift anisotropy (CSA), and quadrupolar interactions. A variety of solid state NMR techniques have been developed to improve the resolution and sensitivity including the use of multiple pulse sequences, cross polarization, high power decoupling, multiple quantum NMR and fast MAS NMR. For example, solid state 1 H MAS NMR has found a number of applications in material characterization. [17] Even with these advances, solution-like NMR spectra are rarely realized for solid samples.
For samples that are in the liquid/solid classification, motional averaging will partially reduce or remove many of these broadening interactions. It is also possible to swell or plasticize a material to increase the local mobility, thereby reducing the magnitude of these interactions. [18] Even for liquid environments in a heterogeneous sample, differences in the magnetic susceptibility within the material can drastically reduce the observed resolution. For these types of dynamically averaged or susceptibility broadened systems, MAS even at moderate speeds will produce high resolution NMR spectra: this is the niche of HR-MAS NMR.
This improved resolution due to MAS arises because the Hamiltonians describing the dipolar, CSA and magnetic susceptibility interactions all contain an orientational component that scales as (3cos 2 θ-1), where θ is the angle between the rotor spinning axis and the magnetic field (Figure 1). When a sample is spun about an axis that is at the "magic angle" (θ = 54.7 o ) these interactions vanish. A simplistic schematic of a HR-MAS stator is pictured in Figure 1. As mentioned before, the main difference between a standard MAS stator and an HR-MAS stator is the magic angle gradient coil that can produce a gradient along the rotor spinning axis. For samples where residual dipolar interactions or differences in magnetic susceptibility are small, HR-MAS NMR reduces the observed line widths to be similar to those observed in solution high resolution NMR spectroscopy. For liquids adsorbed into/onto materials MAS speeds between 1 and 2 kHz may result in improved resolution, while for plasticized/swollen materials MAS speeds between 4 and 10 kHz may be necessary to obtain the desired resolution. Care must also be taken to ensure that the residual spinning sidebands (due to incomplete averaging) fall outside the spectral window of interest.
As an example of the enhanced resolution that can be achieved under HR-MAS two different swollen polymer systems are shown in Figure 2. These anion exchange membranes (AEM) operate in methanol solution, and readily adsorb significant amounts of water and methanol from the solution during use. Our group is interested in identifying the different chemical environments that these solvent molecules see within the membrane, as well as directly measuring the diffusion rates for these different species. Under static conditions, the water and methanol are not resolved in the NMR spectrum (Figure 2A, green). This lack of chemical shift resolution precludes any information about the local membrane environment being obtained. The broadening is caused by the magnetic susceptibility differences between the rigid AEM and the methanol solution. In this example, minimal MAS of ~750 Hz ( Figure  2A, red) was sufficient to average the magnetic susceptibility allowing individual solvent environments to be easily observed. Increasing the spinning speed to 4 kHz removed the spinning side bands from the spectral region of interest, but did not dramatically increase the resolution of the solvent resonances. In this swollen AEM example, signals from the actual membrane are not observed at 4 kHz (only the swelling solvent is observed) indicating that the dipolar interactions present in the membrane remain much larger than the MAS speed. Since the protons from the membrane are not readily observable, it provides a unique opportunity to further study the solvent behavior in the membrane without interference. Additional discussion of this system is presented in Section 2.1.1. The second example, involves the swelling of a polyButadiene-AcryloNitrile (pBAN) polymer in CDCl3. Inspection of Figure 2B (green) shows that under static conditions this rather soft material produces a relatively unresolved NMR spectrum. Even after extensive solvent swelling, the NMR spectrum is essentially unchanged ( Figure 2B, red). With HR-MAS (4 kHz spinning) the resolution is dramatically enhanced, enabling a more detailed analysis of the pBAN 1 H NMR spectra. With the dramatic increase in resolution observed utilizing HR-MAS, the arsenal of standard solution NMR techniques can be implemented for HR-MAS experiments, including solvent suppression, gradient-assisted sequences, and multiple-dimensional experiments. These can include INEPT (Insensitive Nuclei Enhanced by Polarization Transfer), COSY (COrrelation SpectroscopY), NOESY (Nuclear Overhauser Effect SpectroscopY), TOCSY (TOtal Correlation SpectroscopY), HETCOR (HETeronuclear CORrelation), HMQC (Heteronuclear Multiple Quantum Coherence), and HMBC (Heteronuclear Multiple Bond Correlation) to name a few. As an example, a gradientassisted 2D 1 H-1 H HR-MAS COSY NMR spectra for an ionic liquid adsorbed on an aluminum oxide membrane is shown in Figure 3. These types of correlation experiments could not be realized without the resolution afforded by HR-MAS. Additional discussion of this IL material is presented in Section 2.3.1.
It is also important to note that because anisotropic interactions (dipolar, CSA) are still present within these HR-MAS samples (though greatly reduced by molecular motion), it is also possible to apply some solid state NMR techniques during HR-MAS experiments. For example, to measure the residual homonuclear dipole-dipole interaction it is possible to incorporate a dipolar recoupling sequence to re-introduce this interaction under MAS. This allows the incoherent NOE exchange process to be replaced by a coherent polarization transfer process. These types of HR-MAS recoupling experiments have been demonstrated using DQ COSY for solid state synthesis samples, [19] while our group has incorporated a radio frequency dipolar recoupling (RFDR) sequence into the standard NOESY experiment for lipid membranes, including the extension of this dipolar recoupling into a mixing period during 1 H-13 C Heteronuclear correlation experiments. [20]

Experimental protocol and limitations
Many HR-MAS samples contain liquids or liquid-like materials. These samples require special care to prevent dehydration that can occur in a standard rotor with only a rotor cap. Various inserts are commercially available for packing of HR-MAS samples and can be seen in Figure 4. These inserts provide a tight seal to prevent dehydration or loss of solvent during MAS experiments. The inserts used generally depend on the sample size and need of the user. Examples of various insert options from Bruker BioSpin for a standard 4 mm rotor are shown in Figure 4. For ~12 μL samples, a Kel-F® bottom spacer (or a half drilled out rotor) and a top spacer with a seal screw can be utilized ( Figure 4B, 4F and 4G). The top spacer insert ( Figure 4F) contains a small hole in the top. When placed into the rotor at the properly gauged distance, any additional liquid will protrude through this hole. A Kimwipe® can then be utilized to remove this liquid, followed by using a small Kel-F® screw to seal the insert ( Figure 4B) and prevent dehydration. For ~50 μL volumes, a standard rotor can be utilized with the above top spacer and seal screw ( Figure 4B, 4F, 4L). In addition to the spacers, a 30 μL disposable insert is also available ( Figure 4I). The disposable Kel-F® inserts use a plug and screw cap to keep the sample well sealed ( Figure  4B, 4H). The inserts are efficient if multiple samples need to be run, as samples can be packed into inserts and then easily placed into and removed from rotors, without having to wash the rotors between runs, and without having to own a large number of MAS rotors. Temperature regulation is important in HR-MAS NMR for both sample purposes, and experimental reasons. The temperature of the bearing gas in many probes can be regulated to maintain a stable sample temperature; however, there is additional frictional heating that comes from spinning the rotor. To compensate for MAS related heating of the sample, the HR-MAS probe temperature should be calibrated at various MAS speeds and changes made to the temperature of the bearing gas to regulate the sample at the desired temperature. Temperature calibration of a HR-MAS probe has been discussed in literature. [21] This study utilized both methanol and glucose as NMR thermometers to determine the impact of MAS on temperature. Temperature changes measured at 2, 4, 6, and 8 kHz MAS where found to be 0.8 ± 0.1, 2.2 ± 0.1, 5.0 ± 0.1, and 7.9 ± 0.2 K in methanol and -0.3 ± 0.1, 1.2 ± 0.2, 3.0 ± 0.2, and 6.1 ± 0.1 K for glucose. [21] Like all high resolution NMR techniques where resonances can be very narrow, it is important to be able to shim a sample to achieve the best possible resolution. Unlike standard high resolution NMR probes, the HR-MAS probes have the samples spinning at the magic angle, therefore the standard shimming protocols cannot be utilized. Instead, a combination of the standard shims must be used to address inhomogeneity of the B0 field. By placing the probe so that the rotor is in the (x, z) plane in the laboratory frame, a combination of the laboratory shims can be used to express the Z shims along the magic angle. To obtain the equivalent response of a Z shim in standard shimming protocol, i.e. the Z shim along the magic angle ), the user needs to optimize a combination of the Z and X room temperature shims given by the linear relationship Likewise, the equivalent of a Z 2 shim is given by 22 2 () 22 and the equivalent Z 3 shim is defined using Spectroscopy points out that in theory shimming could be performed with just laboratory shims X, XZ,XZ 2 , Z 4 , and Z 5 . [23] In reality, additional shims can be used to compensate for any inefficiency in the shim coil. A previous study has shown that once good shims are obtained, those shims can be utilized for samples with the same detection volume and position, for example the same sample size in the same type of rotor. [24] The shims should be independent of solvent because any susceptibility differences caused by solvent will be averaged out by MAS and does not affect the shims. [24] Piotto et al. also demonstrated shimming issues involved in systems containing water. In many cases, improved shims will be detected in an increase in lock signal. Deuterated water (HDO), however, has a strong chemical shift dependence on the temperature. When a sample spins, there can be a temperature gradient across the sample which can cause a portion of the HDO resonances to shift and broaden. If the probe is shimmed again based on the lock level from the water resonance, then any other signal in the spectra will most likely be deshimmed as those signals do not have as large of a chemical shift temperature dependence. With this in mind, Piotto et al. emphasize that shimming should be performed based on line shape, not lock level and attention should be taken when working with materials containing water. [24] More recently the MAGIC SHIMMING method has been described which utilizes a conventional homospoil gradient pulse to perform gradient shimming on MAS probes. This techniques does not require a gradient along the magic angle, and will reduce some of the trial and error presently inherent with manual shimming of HR-MAS probes. [25]

HR-MAS NMR spectroscopy in materials characterization
In this section a brief review of HR-MAS NMR studies involving the characterization of materials is presented. The majority of studies involved modified surfaces or surface interactions, and demonstrates the characterization need that HR-MAS NMR can fulfill.

Polymers
For rigid polymer materials the dipolar interactions are very large, leading to broadening of the NMR resonances, and can be characterized using standard solid state MAS NMR techniques. As noted before, these rigid polymers are not readily observed under HR-MAS conditions. For solvent swollen polymers the increased mobility of the polymer segments leads to a semi-solid regime where the moderate spinning speeds of HR-MAS are sufficient to produce high resolution NMR spectra. For example see the improved resolution observed for CDCl3 swollen pBAN ( Figure 2B). This ability to obtain spectra for mobile domains or components is the basis for the majority of HR-MAS polymer studies. A variety of different solvents are employed in combinatorial and solid phase synthesis, many of these with multiple 1 H resonances. In addition, there may be reasons to avoid having to dry the sample prior to introducing a deuterated solvent. In these cases, complicated solvent suppression techniques may be required. [26] It has been demonstrated that diffusion-filter HR-MAS spectra (see PFG discussion, Section 3.1) can provide suppression of solvent resonances based on differential diffusion behavior. [27] One example of using HR-MAS NMR for characterizing polymer materials is the demonstration of the cyclic polyamide receptor threading onto the highly flexible polyethylene glycol (PEG) polymer chain attached to a polystyrene bead. Distinct PEG resonances for the threaded and non-threaded complexes could be easily resolved under HR-MAS conditions, while 2D NOESY spectra showed cross-peaks between the aromatic protons (δ ~ +8.87 ppm) of the rotaxane and the methylene protons (δ ~ +3.55 ppm) of the PEG polymer chain. [28] In another study, the complexation of Zn and Ru metalloporphyrins to beads functionalized with pyridyl ligands revealed that supermolecular interactions and changes in the dynamics are directly probed by HR-MAS techniques. [29] The vulcanization of butadiene rubber (BR) with different curing systems has also been monitored by 1 H and 13 C HR-MAS NMR. This rubber is a very mobile system, with the increased resolution afforded by HR-MAS allowing the chemical identity of cross links to be determined, and revealing that α-substitution or addition depended on the disulfide crosslinkers employed. [30,31] Other HR-MAS studies observed the impact of cross-linking in solvent swollen poly(amidoamine) polymers, the nature of water interactions in these same class of polymers, [32,33] or the cross-linking performance in silicon-containing soybean-oil copolymers. [34] HR-MAS NMR has also been used to study the functionalization of poly hydroxyethyl methacrylate (HEMA) cryogels, [35] cyclomaltoheptaose polymers, [36] the synthesis of hyper-branched bis (hydroxymethyl) propionic acid (bis-MPA) polymers as a function of catalysts, [37] and polymers for nanoparticle stabilization. [38] HR-MAS NMR has found use in investigation of porous polyalkylvinyl ether polymer particles being used for stationary phases in chromatographic applications. These measurements were performed using the same solvent as HPLC and allowed details about the polymer structure and mobility to be evaluated; these properties are expected to impact the chromatographic process. [39] Interactions between solution molecules and the polymer component of HPLC stationary phases (C18, C30 and PEAA) [40] or the interaction with molecularly imprinted polymers [41] have also been investigated using HR-MAS NMR. These studies used saturation transfer difference pulse sequences to identify those molecules that were involved in binding to the larger stationary phase.
The NMR spectroscopic characterization of polymer degradation is commonly directed towards analysis of the small degradation fragments that are solvent soluble. By using HR-MAS NMR the hydrolytic degradation of the biodegradable photo-initiated cross linked poly(DL-lactide)-dimethacrylate (PDLLA) polymer network was directly monitored. Swelling these polymers in the solvent dimethyl sulfoxide (DMSO) and a combination of 1D and 2D (COSY, NOESY) experiments allowed the different signals in the PDLLA chain to be identified. Degradation in a base was shown to occur through hydrolysis of the ester bonds within the poly(lactide) segment. [42] HR-MAS has also been used to study the role of partial hydrolysis in controlling the composition of hydrophobic polyacrylamide gels. [43]

Example: HR-MAS NMR investigations of anion exchange membranes
The increased resolution achieved with 1 H HR-MAS on the solvent swollen anion exchange membrane (AEM) was demonstrated in Figure 2. Due to differences in the magnetic susceptibility between the swelling solution and the polymer membrane, the water and methanol resonances were not resolved under static conditions. Under HR-MAS four distinct NMR resonances were observed for the solvent species; two water resonances and two methanol resonances. Based on line widths and PFG HR-MAS diffusion measurements (Section 3.2.1) the higher ppm shifted water and methanol were assigned to bulk (or free) methanol or water within the membrane, while the lower ppm shifted water and methanol resonances exhibit slower diffusion rates and were assigned to water or methanol associated (bound) within the membrane. The increased resolution obtained under HR-MAS was further exploited to explore the interchange of water and methanol between the different binding environments within the AEM using 2D HR-MAS NMR NOESY exchange experiments ( Figure 5). At short mixing times (1 ms) no correlations were observed between any of the resonances. With increasing mixing time (> 10 ms), cross peaks were observed between free water and membrane-associated water, plus cross peaks between free methanol and membrane-associated methanol, resulting from physical exchange of water (red dashed lines) or methanol (red dashed lines) between the free and unbound environments. With long mixing times (> 200 ms), NOE correlation between methanol and water were observed (green dashed line), and result from through space NOE magnetization exchange (not physical spatial exchange). Interestingly, these NOE correlations were only observed between associated water and associated methanol, as well as free water and free methanol, supporting the assignment and indicating that the spatial contact of water and methanol is maintained in these associate membrane environments. Additional exchange studies are ongoing. The increased spectral resolution obtained for the swelling solvent made it possible to detect exchange between free and associated water and between free and associated methanol (red dashed lines), as well as NOE magnetization exchange between free water and methanol and associated water and methanol (green dashed lines).

Ceramics, zeolites, catalysts surfaces and composites
Due to the rigid or crystalline nature of ceramics, zeolites, catalysts or inorganic/organic composites, solid state MAS NMR spectroscopy remains the dominant NMR characterization method, and can involve a range of nuclei and multi-pulse NMR techniques. In contrast, the limited number of HR-MAS NMR studies is directed towards the investigation of gas/solid adsorption or intercalation within these materials. For example, the degradation of the pesticide diethyl parathion and the chemical warfare agent (CWA) simulant diisopropyl fluorophosphates in functionalized montmorillonite clay using 31 P HR-MAS NMR has been reported. The HR-MAS removed susceptibility effects present from the clay material, and allowed a direct measure of the decomposition kinetics. The adsorption of organosilanes on the MgCl2 in Ziegler-Natta catalysts has been investigated using 1 H HR-MAS, and demonstrates differential binding (differential mobility) to the Mg surface with changing surface coverage and as a function of the degree of saturation in the Mg coordination sites. [44]The polymer mobility in the condensed sol-gels produced from organosilanes and phenyl siloxanes was also evaluated. [45]

Ionic liquids and liquid crystals
HR-MAS NMR should be an ideal tool for the characterization of ionic liquids (ILs) and liquid crystals (LC) due to the high viscosities commonly encountered in ILs, and the thermal-or concentration-induced ordering transitions of LCs. Ionic liquids are interesting compounds, and continue to be used for a wide range of material science applications. ILs are used as liquid electrolytes in energy storage and production devices, solvents for CO2 capture and green chemistry, solutions for biomass processing, nanostructured synthesis and solvents for catalysis reactions. For many of these applications room temperature ionic liquids (RTIL) are utilized as neat solutions without another solvent. Strong attractions between the IL cation and anion component may lead to the formation of important structural motifs that could be missed or changed with the introduction of a solvent. For NMR characterization, solvent-IL interactions may also mask subtle changes in the chemical shift produced by other molecular interactions of interest. In the case of ILs, it is now recognized that HR-MAS NMR provides a powerful tool for the characterization of these systems.
One of the earliest reports involving HR-MAS NMR of ILs studied silica-immobilized ILs suspended in DMSO, [46] which produced nearly liquid-like resolution. In this same study, they were also able to obtain 2D 1 H-13 C HR-MAS NMR HMQC spectra of these attached ILs.
Another study explored the use of ILs as a solvent or chemical reaction media. The standard solution NMR of compounds dissolved in these ILs suffered from low resolution, produced by the high viscosities of the ILs, dynamic range issues due to the lack of deuterated ILs, and concerns about chemical shift referencing due to the high magnetic susceptibility of ILs. To demonstrate the capabilities of HR-MAS NMR, Rencurosi and co-workers [47] dissolved para-methoxy benzyl acetate and a glucopyranoside in a series of different ILs. Even at moderate spinning speeds there was significant improvement in resolution, and the chemical shift referencing became consistent with that observed in solution NMR using traditional deuterated solvents. In addition, they were able to follow acetylation of paramethoxybenzyl alcohol directly in the ILs. The increased resolution inherent in this technique was also important for investigations of CO2 interactions with imidazolium based ILs using 13 C HR-MAS NMR. [48] 1 H HR-MAS NMR has also been used to probe the interaction of the RTIL N-methylimidazolium chloride ([Hmim] + Cl -) with the silica surface of Aerosil. This adsorption was monitored through changes in chemical shift with increased IL loading and Clsalt concentration. Based on the magnitude of the chemical shift variations, the interaction was found to preferentially involve the H(2) position of Hmim + , and was a physisorption versus a chemisorption process. [49] In another study, the dynamics of IL's confined in monolithic silica ionogels were followed by 1 H HR-MAS NMR and relaxation experiments, and demonstrated that the IL maintained liquid like behavior with very little reduction in motions even as the pore size diminished from 12 nm to 1.5 nm. [50] Surprisingly, only a single 1 H HR-MAS NMR diffusion study of a non-biological LC material confined in nanopores has been reported. [51]

Example: HR-MAS NMR investigation of ionic liquids on surfaces
While many RTILs are "liquids" at room temperature, they can be highly viscous, making standard solution NMR analysis on the neat ILs challenging (as noted above). In our laboratory, a novel series of quaternary ammonium and cyclic pyrrolidinium RTIL have been studied using different NMR techniques. This includes pulse field gradient (PFG) NMR to measure diffusion and 14 N NMR relaxation experiments to determine IL molecular reorientation times. [52] More recently, we have explored using 1 H HR-MAS NMR to characterize the interactions between IL and metal oxide surfaces. Figure 6A shows    Figure 3. This correlation experiment immediately confirms the chemical shift assignments for the neat IL sample, and clarifies differences seen for the IL dissolved in a solvent. The 2D 1 H HR-MAS NOESY NMR spectra for the IL [MBPyrr] + [TFSI]adsorbed into the ANAPORE membrane is shown in Figure 7. Interestingly there are numerous cross peaks observed even at moderate mixing times, reminiscent of the liquidordered phase observed in lipid bilayers (For example see Ref. [20,53,54]). These throughspace correlations are stronger than those observed in the neat IL (not shown), and suggest that an increase in the dipolar-dipolar interactions is occurring for the surface adsorbed species. This argues that local motions have become reduced for the adsorbed species. There are also some missing or weak correlations (dashed circle) implying distinct conformations present for the n-butane chain on the adsorbed species. Also of interest is the appearance of new chemical environments for the adsorbed IL. This includes a weak resonance at δ = +4.8 ppm which is assigned to residual water, but shows no correlations with the IL. There are two additional 1 H environments at δ = +5.2 and +5.3 ppm, which are attributed to additional surface water species on the aluminum oxide surface. The environment at δ = +5.2 ppm shows a through space NOE correlation with a new resonance near 2 ppm. This small resonance is buried under the shoulder of the dominant methylene (H-2) resonance and has not been presently assigned. These types of correlation experiments demonstrate the capabilities and information that can be obtained from HR-MAS studies of ILs, and suggest future efforts along these lines are warranted.

Surface modified nanoparticles
HR-MAS NMR studies of surface modified nanoparticles (NP) are closely related to those of the polymer resin studies mentioned in the previous sections, with accurate characterization of the surface-attached ligands being the primary objective. For NPs the solvent plays the role of suspending or dispersing the material, in contrast to the role of swelling that the solvent plays in analysis of resins. For HR-MAS NMR studies of NPs, the optimal solvent provides both high solubility, and good dispersion (prevents aggregation) of the NPs. HR-MAS NMR studies of surface modified NPs include the 1D and 2D 1 H NMR investigation of modified gold (Au) NPs using a wide range of high resolution correlation experiments: COSY, TOCSY and HMQC. [55] This revealed that the relative signal sensitivity (intensity) depends on the distance between the detected 1 H and the surface of the Au NP. This distance correlation is either a function of reduced local dynamics, or spin-spin T2 relaxation effects governed by the Au surface. Another study involved the active molecular compounds Aloin A and Aloesin extracted from the leaf of Cape Aloe, and was able to show how these compounds selectively stabilize Au NPs preferentially through the glucose component. [56] HR-MAS NMR has also been used to identify the binding motif of peptides on the Au NP surface. [57] While HR-MAS NMR is typically employed to overcome anisotropy of magnetic susceptibility or residual dipolar interactions, Poito and coworkers [58,59] have reported an interesting set of experiments that demonstrated HR-MAS can be employed to overcome paramagnetic effects present in iron oxide NPs. Through a careful analysis of these iron oxide systems, the surface structure and ligand binding (chelation) configuration was determined, and in several cases were quite different from the standard single point attachment proposed by others. [58] 1 H HR-MAS NMR has also been used to look at the surface modification in polymer based NPs, including the monitoring of multiple synthetic steps employed during the peptidesurface modification of poly(vinylidene fluoride) (PVDF) nanoparticles, [60] or the surface modification of Dendron based NPs. [61] Polystyrene embedded silver clusters produced by a thermolysis reactions have also been characterized using HR-MAS NMR. [62]

Surface immobilized linkers and catalysts and chiral stationary phases
HR-MAS NMR spectroscopy has also been extensively used in investigating the mobility of linkers and catalysts attached to other materials, including 1 H and 31 P HR-MAS NMR studies of phosphine, bisphosphinoamine linkers and corresponding metal catalysts. [63][64][65][66] These investigations demonstrate very nicely that the surface mobility as probed by HR-MAS is strongly dependent on the swelling solvent employed in the studies. [67] The grafting of organotin catalysts to polystyrene for the tranesterification reaction between ethyl acetate and n-octanol or the ring-opening polymerization of -caprolactone has been followed by both 1D and 2D 1 H, 13 C and 119 Sn HR-MAS NMR. [68][69][70] While the catalytic activity of the attached tin was unchanged following numerous cycles, the NMR revealed there was actually a change in the liquid-solid interface with a reduction in the mobility of the undecyltin trichloride catalyst at the resin surface. For experiments with rapid catalytic turnover rate there was a reduction in performance. This reduction in activity was reversible with solvent extraction, and was confirmed by HR-MAS to involve the change of the Sn environment back to the original state. [69] The use of 119 Sn HR-MAS NMR has also been demonstrated to provide a quantitative measure of the tin loading in supporting catalysts. [71] This HR-MAS NMR study was also the first to incorporate the ERETIC (electronic reference to access in vivo concentration) method.
HR-MAS NMR techniques have been used to characterize the surfaces of silica and polystyrene particles modified with chiral agents that allow identification of different enantiomers. In one study the NMR was able to probe regiochemistry of the surface chemistry, along with monitoring the stability against decomposition of the polystyrene resins. [72] Another investigation used 2D transfer NOESY NMR experiments to evaluate the stereoselective binding of molecules to the chiral stationary phase based on the negative cross-peaks associated with the transfer NOE effect. [73]

Soil and humic materials
While standard solid state MAS NMR techniques have become common in the investigations of humic, soil and coal materials, the implementation of HR-MAS techniques for these materials has been more limited. The ability of HR-MAS to identify different motional regimes has proven powerful in the elucidation of the chemistry at water/soil interfaces. [74][75][76][77][78][79] These NMR studies show that at this interface fatty acids, aliphatic esters and alcohols are the prominent species, and that aromatic functional groups are protected by hydrophobic regions and are not directly accessible to the penetrating water, [74] while surface polymethylene groups may control the sorption properties of organo-clay complexes. [76,80] These types of studies have been extended to three-dimensional (3D) HMQC-TOCSY to further increase the resolution of the highly overlapping spectra from humic materials. [81]

PFG HR-MAS NMR to measure diffusion in materials characterization
Modern HR-MAS probes include a gradient coil that can produce a magnetic gradient along the long axis of the MAS rotor which is set at the magic angle (θ = 54.7 o ). Standard MAS probes have also been combined with micro-imaging gradient systems in which gradient coils wrapped around the stator were not employed, but instead rely on imaging gradients that were external to the probe. [82] This is not the common configuration, and has been replaced by significant development efforts from the instrumental vendors involving the integration of gradients directly into the HR-MAS probes. With the gradient coil along the magic angle pulsed field gradient (PFG) experiments can be performed under MAS conditions. Fortuitously, enhanced T2 times are generally observed under MAS allowing PFG experiments with longer diffusion times to be implemented than would have been accessible with static conditions.
During the PFG diffusion experiments, the application of a gradient "tags" a spin with a phase that is related to its spatial position. Figure 8A provides a pictorial representation of the dephasing of spins around the magic angle caused by the magic angle gradient. If the position of the spin does not change during the diffusion period (Δ), this dephasing is refocused and the original signal intensity (S0) is recovered. If on the other hand the spin changes spatial position (diffuses) during Δ, the dephasing for that spin is not refocused, and the signal intensity decreases. The loss in signal intensity with increasing gradient strength is related to the self-diffusion rate with the classic Stejskal-Tanner equation: where S is the experimental amplitude of the signal,  is the gyro magnetic ratio, g is the gradient strength,  is the gradient pulse length, Δ is the diffusion time, and the D is the diffusion constant. [83] By fitting this decay the diffusion constant can be determined. Figure  8B shows an example of the signal intensity loss observed during diffusion experiments for the different two water environments present in swollen AEM with increasing gradient strength.
PFG diffusion HR-MAS NMR experiments can also be used to obtain diffusion-filtered NMR spectra through the separation of different motional regimes present in complex mixtures. This filtering is accomplished by selecting Δ times where the signal intensity (Eqn. 1) for the fast diffusing components has been highly attenuated, while the slow diffusing components have significant signal intensity remaining. Practical aspects of diffusion measurements using HR-MAS have been previously discussed by Viel et al. [84] A significant finding from this study was that sample volume played a role in the reliability of diffusion rates measured. It was shown that small volumes (~12μL) exhibit reproducible diffusion rates, while larger volumes (~50μL, a full 4 mm rotor) produced inconsistent or unreliable data. Three commonly used PFG diffusion pulse sequences are shown in Figure 9. The basic spinecho diffusion sequences is depicted in Figure 9A, but is limited by loss of signal intensity due to spin-spin T2 relaxation during the diffusion period Δ. Two variations of the stimulated echo (STE) sequence are shown in Figure 9B and 9C, and in this case spin-lattice T1 relaxation is occurring during the diffusion Δ period. For most materials T1 values are longer than T1 making these STE sequences the preferred choice for material analysis. It should be noted that all of the gradient pulses in these sequences are trapezoidal shaped to compensate for the inability of instrumentation to generate perfect rectangular gradients. Shaped pulses, like sine or trapezoidal shapes, are used to produce experimentally reproducible gradient pulses. The PFG stimulated echo with dipolar gradients and spoil gradient, depicted in Figure 9B, is beneficial for the use with many HR-MAS samples which exhibit differences in magnetic susceptibility across the sample. [85,86] The PFG stimulated echo in Figure 9C has an additional delay to the PFG stimulated echo in Figure 9B that is utilized to address eddy currents within the sample. [87] [85], and C) PFG Stimulated Echo with dipolar gradients and spoil gradient with an additional eddy current delay. GMAS indicates that the gradient is applied along the magic angle.

Diffusion in zeolites, nanoparticles and liquid crystals
As noted for other material studies, heterogeneities in the magnetic susceptibility or restricted molecular motions within zeolite crystallites lead to broadening of the NMR signal, such that resolution of individual species in mixtures becomes difficult. HR-MAS NMR resolves this issue, and has led to the utilization of PFG NMR diffusion experiments on organic mixtures in zeolites. This technique has been used to study the diffusion of nbutane in silicalite-1, [88] ethane, water and benzene mixtures adsorbed to the zeolite NaX, [89] acetone-n-alkane (C6 to C9) mixtures in nanoporous silica, [90] or mixtures of nbutane and iso-butane adsorbed in MFI zeolite. The increased resolution afforded by these PFG HR-MAS studies on mixtures reveal the obstructive influence of the isopropyl molecules or bulky benzene on the diffusion of other molecular species, [89,91] and that the creation of acetone-alkane complexes greatly impacts the observed diffusion properties. [90] PFG HR-MAS NMR has also been used to obtain 2D DOSY (Diffusion Ordered SpectroscopY) spectra of surface modified iron oxide NPs. These results were able to distinguish between NP-bound and free ligands in these materials. [58] These types of PFG HR-MAS NMR experiments should prove useful in understanding the surface-ligand dynamics present in modified NPs.
Diffusion experiments using HR-MAS NMR has also been found useful in the analysis of transport properties in lipid membranes. [92][93][94] Due to the anisotropic molecular reorientation of these liquid crystalline (LC) systems, significant dipolar coupling remains, leading to broad lines and short relaxation times. However dipolar coupling can be reduced through the use of MAS. Surprisingly, the use of PFG HR-MAS NMR for non-biological LCs is more limited, with the single investigation of local molecular dynamics of the thermotropic LC 4'-pentyl-4-cyanobiphenyl (5CB) confined in Bioran glasses with pore diameters of 30 nm and 200 nm being reported. By utilizing PFG techniques it was possible to measure the diffusion constants as a function of temperature through the isotropization temperature of the liquid crystal, thus demonstrating that for this case there is only a minor reduction in the diffusion rates with molecular confinement. [51]

Diffusion in polymers
There is extensive literature on the use of PFG NMR to measure diffusion of polymer solutions and melts, along with PFG diffusion measurements of different species adsorbed into polymers, including gases, water, organic solvents and electrolytes. There have been a limited number of examples where the improved resolution afforded by HR-MAS NMR was coupled with PFG. This work includes the development of diffusion filtered HR-MAS NMR techniques to study the gelation process of super-molecular gels, [95] along with a combination diffusion-filtering and a spin-echo enhanced (T2 -filtered) experiment on DMFswollen resins. [96] These techniques allowed the identification of free and surface bound molecules, while eliminating the signal from the immobile bulk resin matrix. In complex mixtures there may be future avenues for HR-MAS to resolve subtle differences in the local chemical environments as demonstrated in the following example.

Example of HR-MAS diffusion in anion exchange membranes
As discussed in Section 2.1.1, four distinct resonances were observed in the 1D 1 H HR-MAS NMR spectra of AEM polymers swollen in a 1N methanol solution. In Figure 10A, NMR spectra for three AEM polymers with different ion exchange capacities (IEC) are shown. Two resonances were observed for both water and methanol, and were assigned to free (F) and membrane-associated (A) environments. From this HR-MAS data it is easy to see that there is a correlation between chemical shift of the associated species and the IEC of the membrane, with both the associated water and methanol resonance shifting to lower ppm with decreasing IEC. This decrease in chemical shift is most likely due to a change in hydrogen bonding between the solvent components and the membrane, reflecting how strongly the solvent molecules are associated with the membrane. Recall that the resolution of these individual environments was not observable in the static NMR spectra (Figure 2A). Using 1 H HR-MAS PFG diffusion experiments, the self-diffusion constants were obtained for each of these four different environments, which were not accessible from the static data.
Because the 1 H signal for the AEM membrane is not readily observable under HR-MAS conditions, the diffusion rates obtained for the resolved solvent resonances were not biased by the polymer membrane. Figure 10B shows the signal decay for the associate methanol environment as a function of gradient strength for the three different IEC levels. The magic angle gradients were used to perform diffusion measurements utilizing a PFG Stimulated Echo with dipolar gradients and spoil gradient with a Δ=100ms ( Figure 9B). The signal decay shows that there is a correlation between diffusion rate and IEC, exhibiting a faster diffusion rate with increasing IEC values. A more detailed analysis of this work is forth coming, but this example demonstrates the power of combining HR-MAS and PFG diffusion experiments.

Conclusions
The application of HR-MAS NMR to the characterization of materials or material interfaces that exist in the semi-solid range has been demonstrated. A wide variety of different material systems have been explored, showing that this technique can provide resolution and dynamic information where standard solution or solid state NMR techniques were unsuccessful. HR-MAS NMR is a powerful tool for the detailed characterization of modified surfaces and surface adsorbed species. This technique also provides a direct probe of differences in local mobility as reflected by line width variations. Through the combination of the enhanced resolution afforded by HR-MAS with pulse field gradient (PFG) capabilities, selective filtering and diffusion measurements of complex heterogeneous materials can also be realized. The ability to resolve and obtain diffusion rates for multiple environments in materials will prove beneficial for understanding the diffusion process in mixed chemical systems. While HR-MAS NMR is considered a mature, relatively routine technique, the application to the materials field is expected to continue being an active area of development. It is hoped that this review will encourage researchers to explore the application of HR-MAS NMR techniques to their different material systems.