SPAM-MQ-HETCOR: an improved method for heteronuclear correlation spectroscopy between quadrupolar and spin-1/2 nuclei in solid-state NMR

Abstract

The recently introduced concept of soft pulse added mixing (SPAM) is used in two-dimensional heteronuclear correlation (HETCOR) NMR experiments between half-integer quadrupolar and spin-1/2 nuclei. The experiments employ multiple quantum magic angle spinning (MQMAS) to remove the second order quadrupolar broadening and cross polarization (CP) or refocused INEPT for magnetization transfer. By using previously unexploited coherence pathways, the efficiency of SPAM-MQ-HETCOR NMR is increased by a factor of almost two without additional optimization. The sensitivity gain is demonstrated on a test sample, AlPO₄-14, using CP and INEPT to correlate ²⁷Al and ³¹P nuclei. SPAM-3Q-HETCOR is then applied to generate ²⁷Al–³¹P spectra of the devitrified 41Na₂O–20.5Al₂O₃–38.5P₂O₅ glass and the silicoaluminophosphate ECR-40. Finally, the method allowed the acquisition of the first high resolution solid-state correlation spectra between ²⁷Al and ²⁹Si.

---

1. Introduction

Heteronuclear correlation (HETCOR) NMR experiments can provide detailed information about complex structures in chemistry, biology and materials science by identifying atoms in local proximity to one another. Two- and three-dimensional (2D and 3D) HETCOR NMR have been well established in solution NMR^1,2 and solid-state NMR of spin-1/2 nuclei.^3,4 In solids, high spectral resolution in both dimensions is obtained by combining magic angle spinning (MAS) with various radio frequency (RF) pulse sequences, which are used to overcome the homo- and heteronuclear dipolar interactions. Coherence transfer is usually achieved through-space by the dipolar interaction (cross polarization, CP).^3,4 Through-bond coherence transfer via the scalar (J) coupling has been also reported.^5,6 In the case of quadrupolar nuclei, there is an additional challenge of suppressing the second order quadrupolar interaction, which broadens the lines in the HETCOR spectra measured under MAS alone.⁵ The techniques that were developed to overcome this broadening^7–9 proved to be suitable for implementation in the correlation schemes. The first such experiment, between spin-3/2 nuclei (²³Na) and spin-1/2 nuclei (³¹P) in Na₃P₃O₉,¹⁰ used dynamic angle spinning (DAS)⁸ to achieve the isotropic resolution for sodium. The corresponding spectrum was also reported¹¹ with the use of multiple quantum magic angle spinning (MQMAS).⁹ The MQ-HETCOR experiment was carried out by spin-locking the magnetization involved in the isotropic echo that forms as a result of MQMAS refocusing during t₁, followed by polarization transfer and acquisition of the signal from spin-1/2 nuclei in the t₂ domain. More recently, the highly resolved MQ-HETCOR spectra were obtained for spin-5/2 nuclei (²⁷Al) and ³¹P by using polarization transfer via dipolar interaction¹² and J coupling.¹³ We will refer to these methods as MQ-D-HETCOR and MQ-J-HETCOR, respectively.

While the MQMAS-based HETCOR methods offer considerable improvement in resolution, until now their applications have been restricted to samples with a relatively high abundance of ‘friendly’ spin pairs, such as ²⁷Al and ³¹P. Herein, we demonstrate that the sensitivity of the MQ-D-HETCOR and MQ-J-HETCOR methods can be enhanced by implementing the soft pulse added mixing (SPAM) scheme which uses previously unexploited coherence pathways.^14,15 We used these methods to acquire the ²⁷Al–³¹P spectra of the AlPO₄-14 molecular sieve, which has been used as a standard in our earlier reports.^12,13 In order to illustrate the effectiveness of MQ-HETCOR in structural studies of new materials, we report the ²⁷Al–³¹P 3Q-D-HETCOR spectrum of devitrified aluminophosphate glass 41Na₂O–20.5Al₂O₃–38.5P₂O₅, and the ²⁷Al–³¹P 3Q-J-HETCOR spectrum of the silicoaluminophosphate (SAPO) ECR-40. Finally, we demonstrate that these experiments are not limited to aluminium and phosphorus by reporting the ²⁷Al–²⁹Si 3Q-D-HETCOR spectrum of potassium feldspar mineral (microcline).

2. Experimental

2.1 SPAM-MQ-HETCOR: pulse sequences

The pulse sequences,† shown in Fig. 1a,b, are similar to the previously proposed MQ-HETCOR experiments based on CP^11,12 and INEPTR (refocused INEPT).¹³ The isotropic correlation spectra are obtained by using magnetization of the quadrupolar nuclei (I spins), formed at time Rt₁, as a source of polarization for spin-1/2 nuclei (S spins). This approach ensures that the experiment remains 2D and that the delay between scans is controlled by the spins that relax faster.

Fig. 1 Pulse sequences used in (a) MQ-D-HETCOR^11,12 and (b) MQ-J-HETCOR¹³ experiments with SPAM.† The coherence pathways for 2I = |p| and 2I > |p| are shown at the bottom.

The excitation of p-quantum coherence in I spins is carried out via a single pulse of strong magnetic RF field. The signal enhancement by SPAM is achieved by using three coherence pathways (see the bottom of Fig. 1), where PT denotes the polarization transfer and R = p[36I(I + 1) − 17p² − 10]/[36I(I + 1) − 27], instead of a single pathway . Thus, the pQ → −1Q conversion sequence consists of a strong (hard) continuous wave (CW) RF pulse followed by a period of much weaker (soft) CW irradiation. The time interval between the hard and soft pulses should be as short as possible in order to avoid the dephasing of magnetization at the non-zero quantum levels. In the standard SPAM-MQMAS experiments, the role of the soft pulse is to maximize the total transfer of coherences from p = {−1, 1, 0} to p = −1. Since most of the SPAM signal originates from the pathway 0Q → pQ → 0Q → −1Q, the efficiency is at a maximum when the hard mixing pulse is optimized for the pQ → 0Q conversion and the flip angle of the soft pulse corresponds to 90°.¹⁴ Although the 0Q → pQ → 0Q → −1Q transfer alone is slightly less efficient than in the 0Q → pQ → −1Q scheme used previously, the added contributions from routes 0Q → pQ → ±1Q → −1Q lead to significant gain in overall sensitivity. In the case when |p| = 2I (as in ²³Na 3Q-HETCOR or ²⁷Al 5Q-HETCOR experiments) this gain in signal-to-noise (S/N) is about 1.4, according to our calculations using PULSAR.¹⁶ When |p| < 2I, as in ²⁷Al 3Q-HETCOR experiments, the expected gain increases to about 1.7. This is because in the latter case the standard echo pathway follows the ‘upper route’ 0Q → 3Q → −1Q, which compares less favorably with the SPAM scheme. Since the HETCOR schemes require the use of the echo pathway alone,¹⁷ the sensitivity gains generated by SPAM are expected to carry over to MQ-D-HETCOR and MQ-J-HETCOR experiments.

The phase cycling is straightforward, as shown in Fig. 1. For the 3Q experiment with spin-5/2 nuclei the echo pathway is selected by shifting the phase ϕ₁ of the first pulse 12 times between 0 and 360° with 30° increments, while assuming ϕ₃ = ϕ_X. In addition, the phases applied to the CP (or INEPTR) pulses in the ²⁷Al channel are cycled to achieve the spin temperature inversion, which increases the number of different phase sets within the cycle to 24. The 5Q-HETCOR experiment with spin-5/2 nuclei, which uses the pathway , also requires 12 phases, while assuming ϕ₃ = ϕ_−X. The receiver phase is set at −3ϕ₁ in 3Q- and 5ϕ₁ in 5Q-HETCOR experiments. The phase sensitive detection is performed using the hypercomplex method, which involves acquisition of a second set with the phase of the CP pulse shifted by 90° or, equivalently, the phase ϕ₁ by 90°/p.

2.2 NMR measurements

²⁷Al–³¹P spectra of aluminophosphate AlPO₄-14 and aluminophosphate devitrified glass were acquired at 9.4 T on a Bruker Avance-400 spectrometer equipped with a 4 mm triple tuned (HXY) MAS probe. AlPO₄-14 was also studied at 14.1 T on a Varian InfinityPlus spectrometer, using a 3.2 mm HXY MAS probe. The ²⁷Al–³¹P correlations in SAPO ECR-40 and ²⁷Al–²⁹Si correlations in the mineral microcline were observed on a Chemagnetics Infinity spectrometer operating at 9.4 T using a 5 mm HXY MAS probe. The essential experimental parameters are given in figure captions using the following notation: ν_R—the sample rotation rate, ν^X_RF—the magnitude of radiofrequency field applied to X nuclei and τ_CP—the contact time during cross polarization.

All experiments were performed with rotor synchronization during t₁.¹⁸ For spin-1/2 nuclei, the spectra are plotted using the chemical shift values δ_X (X = Si and P), which are referenced to tetramethylsilane (²⁹Si) and aqueous 85% H₃PO₄ (³¹P), both at 0 ppm. The indirect dimension (²⁷Al) is referenced using the isotropic shifts δ_Al,ISO¹⁹ to aqueous solution of Al(NO₃)₃, also at 0 ppm.

3. Results

To illustrate the performance of SPAM-MQ-HETCOR, we chose a well-studied AlPO₄-14 aluminophosphate and three previously untested samples, devitrified 41Na₂O–20.5Al₂O₃–38.5P₂O₅ glass, SAPO ECR-40 and microcline silicate.

3.1 AlPO₄-14

The sample of AlPO₄-14 was templated using isopropylamine and studied in the as-synthesized state. This AFN-type material forms a 3D channel system, made of alternating AlO₄ and PO₄ tetrahedra, with 8-ring pores containing four P and four Al sites.²⁰ The aluminium sites, one of which is five-coordinated (Al₁, with the quadrupole coupling constant C_q = 5.6 MHz), two are tetrahedral (Al₂, Al₃, with C_qs of 4.1 and 1.8 MHz, respectively), and one is six-coordinated (Al₄, with C_q = 2.6 MHz), can be easily resolved by MQMAS.²¹ All four phosphorus sites have been resolved as well, using simultaneous ²⁷Al and ¹H decoupling²² and in the MQ-J-HETCOR spectrum,¹³and are at −20.6 ppm (P₁), −5.8 ppm (P₂), −24.3 ppm (P₃) and −20.1 ppm (P₄). Each phosphorus is connected to four aluminium atoms as follows: P₁ (1Al₁, 2Al₂, 1Al₃), P₂ (1Al₁, 1Al₂, 2Al₄), P₃ (1Al₁, 2Al₃, 1Al₄) and P₄ (1Al₁, 1Al₂, 1Al₃, 1Al₄).

The ²⁷Al–³¹P correlations between Al₁, Al₂, Al₃ and phosphorus observed at 9.4 T using the SPAM-3Q-D-HETCOR scheme are shown in Fig. 2. Also shown, at the top of the figure, is the ³¹P projection of an analogous spectrum acquired without SPAM. As expected, the correlations P₂–Al₃ and P₃–Al₂ were not observed. The sensitivity gain afforded by SPAM was between 1.6 and 1.7, which is in good agreement with our theoretical predictions.

Fig. 2 The correlations of tetrahedral and five-coordinated aluminium sites with phosphorus in the 3Q-D-HETCOR spectrum of aluminophosphate AlPO₄-14 recorded at 9.4 T. On top of the spectrum are shown the ³¹P projections (in skyline mode) observed (a) without and (b) with SPAM, under otherwise equivalent conditions. Spectra were acquired using ν_R = 12 kHz, ν^Al_RF = 150 kHz during hard pulses, ν^Al_RF = 20 kHz during the soft pulse, ν^Al_RF = 7 kHz during CP, ν^P_RF = 9 kHz during CP and τ_CP = 3.5 ms. Total acquisition time was 20 h per spectrum.

The complete ²⁷Al–³¹P SPAM-3Q-J-HETCOR spectrum observed at 14.1 T is shown in Fig. 3. Again, all expected Al–O–P connectivities are revealed, including the previously undetected resonance between sites Al₁ and P₂.¹³ The distinction between sites P₁ and P₄ is not apparent in Fig. 3, but can be made upon closer inspection. In this experiment, the S/N ratio has increased by a factor of 2.1 (±0.1) upon the introduction of SPAM.

Fig. 3 ²⁷Al–³¹P 3Q-J-HETCOR spectra of aluminophosphate AlPO₄-14 recorded at 14.1 T (a) without and (b) with SPAM under equivalent conditions. The standard MAS spectrum of aluminium is also shown on the right. Spectra were acquired using ν_R = 15 kHz, ν^Al_RF = 210 kHz during hard pulses, ν^Al_RF = 5 kHz during the soft pulse and INEPTR, ν^P_RF = 64 kHz and a delay of 0.2 s between scans. Total acquisition time was 12 h per spectrum.

3.2 Devitrified aluminophosphate glass

Phosphate based glasses have applications as sealants for electronic applications, as biomaterials,²³ or as encapsulating host matrices for radioactive waste.²⁴ These applications often benefit from the process of devitrification, which involves the precipitation of crystals after a thermal treatment. The 41Na₂O–20.5Al₂O₃–38.5P₂O₅ glass used in this study was synthesized from reagent grade NaPO₃, Al(OH)₃ and Na₂CO₃, and devitrified at 612 °C for 6 h. Since devitrification often leads to the formation of structures with low crystallinity, solid-state NMR can become useful in studying the resulting structures.

The ²⁷Al resonances in the devitrified sample cannot be resolved under MAS (see Fig. 4a, top projection), however no less than six octahedral Al sites can be identified by 3QMAS. Numerous correlations between these sites and phosphorus are observed in the SPAM-3Q-D-HETCOR spectrum (Fig. 4b). The three peaks between ²⁷Al site at δ_Al,ISO = −4 ppm and ³¹P sites at −28.7, −15.9 and −12.5 ppm are characteristic of the crystalline phosphate Na₇(AlP₂O₇)₄PO₄. Although the complete spectral assignment will require additional experiments involving homonuclear correlation methods, such as ³¹P double-quantum, RFDR and INADEQUATE, we can already infer the presence of aluminophosphate network, rather than a sodium phosphate network in this sample. The presence of aluminophosphate phases is important for the chemical durability of the matrix and its use as a host for radioactive materials.

Fig. 4 (a) ²⁷Al SPAM-3QMAS and (b) ²⁷Al–³¹P SPAM-3Q-D-HETCOR spectra of 41Na₂O–20.5Al₂O₃–38.5P₂O₅ glass. Spectrum (b) was acquired in 23 h under the conditions described in the caption of Fig. 2.

3.3 SAPO ECR-40

The microporous SAPO ECR-40 was prepared using methyltriethanol ammonium as a template.²⁵ This material represents an MEI-type structure, similar to ZSM-18,²⁶ with a space group P6₃/m and unit cell composition of [H₄Si₆Al₁₆P₁₂O₆₈]. The MEI framework is made from 34⁶5⁶ building units, consisting of two T₁, six T₂, six T₃, and three T₄ sites occupied by Al, P, Al and Si, respectively.²⁵ P and Si atoms are connected to Al via oxygen atoms as follows: T₁ (1T₁,3T₂), T₂ (1T₁,3T₃), T₃ (3T₂,1T₄), and T₄ (2T₃,2T₄). Thus each P atom is surrounded by four Al atoms, whereas Al atoms have three P atoms and one Si or Al atom as neighbors. Finally, each Si atom is connected to two Si and two Al atoms.

The ²⁷Al 3QMAS spectrum of the as-synthesized sample (Fig. 5a) showed four resonances, located at δ_Al,ISO = 52, 47, 41 and −10 ppm. To assign these resonances, we also performed ²⁷Al MAS and 3QMAS on a sample calcined overnight under vacuum at 400 °C (spectra not shown). In the 3QMAS spectrum, the resonances at −10 and 47 ppm vanished, the peaks at 52 and 41 ppm remained unchanged, and a new resonance was observed at 56 ppm. The peak at 41 ppm comprised about 25% of total intensity in the MAS spectrum, which suggests that it represents T₁ site while the peaks at 52 and 56 ppm (52 and 47 ppm in as-synthesized sample) correspond to T₃ sites. The resonance at δ_Al,ISO = −10 ppm observed in Fig. 5 is characteristic of octahedral Al sites associated with phosphorus,²⁷ which result from (partial) hydration of the T₃ sites. In agreement with the earlier report,²⁵ the ³¹P MAS spectrum (shown on the top of Fig. 5b) consisted of a strong resonance at −24.5 ppm and a smaller peak at −17 ppm.

Fig. 5 (a) ²⁷Al 3QMAS and (b) ²⁷Al–³¹P SPAM-3Q-J-HETCOR spectra of as-synthesized SAPO ECR-40 recorded at 14.1 T. The ²⁷Al and ³¹P MAS spectra are shown for comparison on top of images (a) and (b), respectively. Spectra were acquired using ν_R = 16 kHz, ν^Al_RF = 210 kHz during hard pulses, ν^Al_RF = 5 kHz during the soft pulse and INEPTR, ν^P_RF = 64 kHz, and a delay of 0.1 s in 2D spectra. Spectrum (b) was acquired in 7 h.

The result of the SPAM-3Q-J-HETCOR experiment on the as-synthesized sample is shown in Fig. 5b. The spectrum clearly shows that all Al sites detected by MQMAS are chemically bound to phosphorus resonating at −24.5 ppm, which indicates that it occupies the T₂ site. The ³¹P resonance at −17 ppm must be ascribed to structural impurities, as it was not detected in the HETCOR spectrum. Again, the efficiency was essentially doubled (the S/N gain was measured at 1.9 ± 0.1) by SPAM, which reduced the acquisition time to 7 h.

3.4 ²⁷Al–²⁹Si MQ-D-HETCOR of microcline

Microcline is a potassium feldspar mineral with formula KAlSi₃O₈, which often has intergrowths of plagioclase feldspars inside the crystal.^28–30 The sample studied here, which was purchased from Alfa Aesar, contains 30% of low albite intergrowths. Both minerals, microcline and low albite, are triclinic crystals having three distinct Si sites (in positions T₂O, T₁m and T₂m) and one Al site (in position T₁O).^30–32 The ²⁹Si chemical shifts reported for both components are listed in Table 1.³⁰ The 3QMAS spectrum (not shown) consists of two ²⁷Al resonances at 60 and 64 ppm (δ_Al,ISO). Shown in Fig. 6a is the ²⁷Al–²⁹Si HETCOR spectrum acquired under RAPT-CPMAS,³³ where the resolution in the ²⁷Al dimension is noticeably lacking due to the second order quadrupolar broadening. This is in stark contrast with the corresponding 3Q-D-HETCOR version (Fig. 6b), where all six Al–Si cross-peaks are easily resolved. Based on the ²⁹Si chemical shift data from Table 1, the ²⁷Al resonances at 60 and 64 ppm can be assigned to microcline and low albite, respectively. It should be noted that the ²⁹Si and ²⁷Al projections of the 3Q-D-HETCOR spectrum are in good agreement with ²⁹Si MAS and the isotropic projection of ²⁷Al 3QMAS spectra, respectively. Also noteworthy is the fact that despite the use of RAPT in the CPMAS experiment, which increased the population inversion across the central transition by a factor of about 1.7, spectra (a) and (b) required a comparable amount of acquisition time in order to achieve a similar S/N ratio. To our knowledge, this is the first ²⁷Al–²⁹Si HETCOR spectrum recorded under isotropic resolution in both dimensions. Previous reports involving these nuclei used 1D ²⁷Al–²⁹Si CPMAS experiments^30,34 or 2D experiments performed under the resolution provided by MAS alone.³⁵

Fig. 6 (a) ²⁷Al–²⁹Si D-HETCOR and (b) ²⁷Al–²⁹Si SPAM-3Q-D-HETCOR spectra of microcline recorded at 9.4 T. The 1D ²⁹Si MAS spectrum is shown for comparison on top of image (a). Data were acquired using ν_R = 10 kHz, ν^Al_RF = 70 kHz during hard pulses, ν^Al_RF = 5 kHz during the soft pulse, ν^Al_RF = 1 kHz during CP, ν^Si_RF = 11.5 ± 1 kHz (ramped), τ_CP = 100 ms and a pulse delay of 10 s. Spectra (a) and (b) were obtained in 80 and 72 h, respectively.

Table 1 ²⁹Si chemical shifts in microcline and albite³⁰

	Site	δ CS	Connectivities
Microcline	T2m	−94.0	2Si, 2Al
	T2O	−96.4	3Si, 1Al
	T1m	−99.5	3Si, 1Al
Albite	T2m	−91.8	2Si, 2Al
	T2O	−96.1	3Si, 1Al
	T1m	−103.9	3Si, 1Al

---

4. Discussion

The results presented in Figs. 2–6 demonstrate that SPAM offers a considerable sensitivity gain in MQ-HETCOR experiments. The proposed SPAM-MQ-HETCOR scheme is robust and easy to set up, because the optimization of MQMAS, polarization transfer, and decoupling strategies can be carried out separately, as detailed in earlier reports.^11–13 We have not detected any significant changes of the relative peak intensities in the MQMAS spectra upon the introduction of SPAM.

With regard to overall sensitivity, further enhancement is possible by increasing the intensity of the central transition prior to the excitation 0Q → ±3Q. This can be accomplished regardless of SPAM by double frequency sweep (DFS),³⁶ rotor assisted population transfer (RAPT),³³ fast amplitude modulation (FAM)³⁷ or hyperbolic secant (HS) pulses.³⁸

A question arises as to whether the demonstrated enhancement of the transfer ±3Q → −1Q, generated by SPAM with a CW nutation pulse, could be achieved by incorporating the alternative conversion methods, such as FAM, DFS, HS or RIACT,³⁹ into the HETCOR scheme. An earlier study has shown⁴⁰ that in the case of spin-3/2 nuclei FAM was more efficient and less sensitive to the magnitude of the quadrupole coupling than CW nutation and RIACT, even in schemes that did not involve the 0 → −3 → −1 (echo) pathway favored by RIACT. Therefore, it is possible that FAM and DFS could be competitive with SPAM, or even used jointly to further improve sensitivity of HETCOR. Such schemes are more likely to be competitive for spin-3/2 nuclei, where they work best (or exclusively, as in the case of RIACT) and where SPAM offers a lower gain. Indeed, our attempt to use the SPAM scheme with FAM conversion in ²⁷Al–³¹P MQ-HETCOR did not result in further increase in sensitivity.

The utility of MQ-HETCOR experiments is critically dependent on the efficiency of polarization transfer. It is well known that the CP transfer involving quadrupolar nuclei is often demanding due to the complex spin dynamics involved in spin-locking and the capriciousness of the Hartmann–Hahn condition.^41–43 This may lead to considerable spectral distortions, which add to the quantitative inaccuracies inherent in MQMAS. In principle, the difficulties associated with the CP transfer could be circumvented by using techniques that do not require spin-locking, such as TEDOR⁴⁴ or PRESTO.⁴⁵ While we previously demonstrated the feasibility of ¹H–²⁷Al TEDOR-based MQMAS,⁴⁶ this method proved to be considerably less sensitive than CP, especially in the presence of strong quadrupolar interactions and fast transverse relaxation. As to the INEPT-based methods, in samples with favorable relaxation rates their sensitivity compares well with standard solid-state NMR methods.¹³ For example, in AlPO₄-14 and SAPO ECR-40, INEPTR yielded better sensitivity per unit of experimental time than direct polarization of ³¹P nuclei. The observed sensitivity was comparable with cross polarization, which showed that transverse relaxation was not of major concern in these samples.

5. Conclusion

MQ-HETCOR spectra are easily acquired for several classes of materials using a SPAM scheme with polarization transfer via dipolar interaction and J coupling. We have also demonstrated the correlations between ²⁷Al and ²⁹Si under isotropic resolution. We anticipate that the increasing availability of higher magnetic fields will make these methods increasingly useful for a variety of spin-3/2 and spin-5/2 nuclei.

Acknowledgements

The authors would like to thank Bruker BioSpin, Region Nord/Pas de Calais, Europe (FEDER), CNRS, French Ministry of Science, USTL and ENSCL for funding. This research was also supported at Ames Laboratory by the US Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences, under contract W-7405-Eng-82. The authors are indebted to Dr B. C. Gerstein for valuable comments. The sample of SAPO ECR-40 was kindly provided by Drs G. J. Kennedy and K. G. Strohmaier from ExxonMobil Research and Engineering.

References

1. A. A. Maudsley and R. R. Ernst, Chem. Phys. Lett., 1977, 50, 368–372.
2. R. R. Ernst, G. Bodenhausen and A. Wokaun, Principles of Nuclear Magnetic Resonance in One and Two Dimensions, Clarendon Press, Oxford, 1991.
3. P. Caravatti, G. Bodenhausen and R. R. Ernst, Chem. Phys. Lett., 1982, 89, 363–367.
4. D. P. Burum and A. Bielecki, J. Magn. Reson., 1991, 94, 645–652.
5. C. A. Fyfe, H. Meyer zu Altenschildesche, K. C. Wong-Moon, H. Grondey and J. M. Chezeau, Solid State Nucl. Magn. Reson., 1997, 9, 97–106.
6. A. Lesage, D. Sakellariou, S. Steuernagel and L. Emsley, J. Am. Chem. Soc., 1998, 120, 13194–13201.
7. A. Samoson, E. Lippmaa and A. Pines, Mol. Phys., 1988, 65, 1013–1018.
8. K. T. Mueller, B. Q. Sun, G. C. Chingas, J. W. Zwanziger, T. Terao and A. Pines, J. Magn. Reson., 1990, 86, 470–487.
9. L. Frydman and J. S. Harwood, J. Am. Chem. Soc., 1995, 117, 5367–5368.
10. T. P. Jarvie, R. M. Wenslow and K. T. Mueller, J. Am. Chem. Soc., 1995, 117, 570–571.
11. S. H. Wang, S. M. De Paul and L. M. Bull, J. Magn. Reson., 1997, 125, 364–368.
12. C. Fernandez, C. Morais, J. Rocha and M. Pruski, Solid State Nucl. Magn. Reson., 2002, 21, 61–70.
13. J. W. Wiench and M. Pruski, Solid State Nucl. Magn. Reson., 2004, 26, 51–55.
14. Z. Gan and H.-T. Kwak, J. Magn. Reson., 2004, 168, 346–351.
15. J. P. Amoureux, L. Delevoye, S. Steuernagel, Z. Gan, S. Ganapathy and L. Montagne, J. Magn. Reson., 2005, 172, 268–278.
16. J. P. Amoureux, C. Fernandez and Y. Dumazy, J. Chim. Phys., 1995, 92, 1939–1942.
17. C. Fernandez, C. Morais and M. Pruski, Phys. Chem. Chem. Phys., 2001, 3, 2552–2553.
18. D. Massiot, J. Magn. Reson., Ser. A, 1996, 122, 240–244.
19. J. P. Amoureux and C. Fernandez, Solid State Nucl. Magn. Reson., 1998, 10, 211–223.
20. M. Helliwell, V. Kaucic, G. M. T. Cheetham, M. M. Harding, B. M. Kariuki and P. J. Rizkallah, Acta Crystallogr., Sect. B, 1993, 49, 413–420.
21. C. Fernandez, J. P. Amoureux, J. M. Chezeau, L. Delmotte and H. Kessler, Microporous Mater., 1996, 6, 331–340.
22. L. Delevoye, C. Fernandez, C. M. Morais, J.-P. Amoureux, V. Montouillout and J. Rocha, Solid State Nucl. Magn. Reson., 2002, 22, 501–512.
23. P. Hartmann, J. Vogel, U. Friedrich and C. Jager, J. Non-Cryst. Solids, 2000, 263&264, 94–100.
24. B. L. Metcalfe and I. W. Donald, J. Non-Cryst. Solids, 2004, 348, 225–229.
25. M. Afeworki, D. L. Dorset, G. J. Kennedy and K. G. Strohmaier, Stud. Surf. Sci. Catal., 2004, 154B, 1274–1281.
26. S. L. Lawton and W. J. Rohrbaugh, Science, 1990, 247, 1319–1322.
27. L. Zhang and H. Eckert, Solid State Nucl. Magn. Reson., 2004, 26, 132–146.
28. P. H. Ribbe, Am. Mineral., 1979, 64, 402–408.
29. J. V. Smith and W. L. Brown, Feldspar Minerals, Sprinder-Verlag, Berlin, 2nd edn, 1988.
30. S. M. De Paul, M. Ernst, J. S. Shore, J. F. Stebbins and A. Pines, J. Phys. Chem. B, 1997, 101, 3240–3249.
31. S. W. Bailey, Am. Mineral., 1969, 54, 1540–1545.
32. G. E. Harlow and G. E. Brown Jr., Am. Mineral., 1980, 65, 985–994.
33. Z. Yao, H. T. Kwak, D. Sakellariou, L. Emsley and P. J. Grandinetti, Chem. Phys. Lett., 2000, 327, 85–90.
34. C. A. Fyfe, K. C. Wong-Moon, Y. Huang, H. Grondey and K. T. Mueller, J. Phys. Chem., 1995, 99, 8707–8716.
35. M. Edén, J. Grins, Z. Shen and Z. Weng, J. Magn. Reson., 2004, 169, 279–283.
36. A. P. M. Kentgens and R. Verhagen, Chem. Phys. Lett., 1999, 300, 435–443.
37. P. K. Madhu, A. Goldbourt, L. Frydman and S. Vega, Chem. Phys. Lett., 1999, 307, 41–47.
38. R. Siegel, T. T. Nakashima and R. E. Wasylishen, Chem. Phys. Lett., 2005, 403, 353–358.
39. G. Wu, D. Rovnyak and R. G. Griffin, J. Am. Chem. Soc., 1996, 118, 9326–9332.
40. M. Pruski, J. W. Wiench and J. P. Amoureux, J. Magn. Reson., 2000, 147, 286–295.
41. A. J. Vega, J. Magn. Reson., 1992, 96, 50–68.
42. A. J. Vega, Solid State Nucl. Magn. Reson., 1992, 1, 17–32.
43. J. P. Amoureux and M. Pruski, Mol. Phys., 2002, 100, 1595–1613.
44. A. W. Hing, S. Vega and J. Schaefer, J. Magn. Reson., 1992, 96, 205–209.
45. X. Zhao, W. Hoffbauer, J. Schmedt auf der Günne and M. H. Levitt, Solid State Nucl. Magn. Reson., 2004, 26, 57–64.
46. J.-P. Amoureux and M. Pruski, in Encyclopedia NMR, John Wiley & Sons, Chichester, 1st edn, 2002, vol. 9, pp. 226–251.

---

Footnote

† Electronic supplementary information (ESI) available: Program codes used during NMR measurements. See DOI: 10.1039/b512246e

---