ArticlesSodium Environments in Dry and Hydrous Albite Glasses: Improved 23Na Solid State NMR Data and Their Implications for Water Dissolution Mechanisms
Introduction
The presence of water in silicate melts has long been recognised as an important control on the behaviour of igneous systems (Bowen 1928; Goranson 1931; Burnham 1975. Water dissolution has dramatic effects on phase equilibria (Kushiro 1969; Pichavant et al. 1992 and on the physical properties of melts (e.g., Lange 1994; Dingwell et al. 1996; Richet et al. 1996; Schulze et al. 1996) and is, therefore, crucially important in a wide range of geological processes as diverse as partial melting of the mantle and explosive silicic vulcanism. In order to rationalize the effects of water on the physical and chemical properties of melts, it is important to have an understanding of the mechanisms of dissolution of water in melts of different compositions. One approach to this goal is to perform spectroscopic and diffraction measurements on hydrous silicate glasses and melts, and many such studies have been published (e.g., Farnan et al. 1987; Okuno et al. 1987; Kohn et al. 1989b; Silver and Stolper 1989; Kümmerlen et al. 1992; Mysen 1992; McMillan et al. 1993; Nowak and Behrens 1995; Shen and Keppler 1995; Holtz et al. 1996; Zotov et al. 1996). However, despite the large literature on the subject and numerous attempts to explain phase equilibria, solubility measurements and spectroscopic data in terms of a physical mechanism for water dissolution, no model has yet found universal acceptance (McMillan 1994).
Kohn et al. 1989a proposed a new model for the dissolution mechanism of water in albite melts, which represented a major departure from all earlier models. The model was based on multinuclear NMR measurements on glasses, the principal observations being (1) only very small changes in the spectra of 29Si and 27Al were seen; (2) the 1H spectra were consistent with the OH/H2O ratio in glasses determined from infra-red by Silver and Stolper 1989; (3) major changes in 23Na spectra as a function of dissolved water concentration were seen. The NMR data were interpreted in terms of a model whereby the aluminosilicate framework of the glass is not depolymerised by rupture of T-O-T linkages, but instead a Na+ ↔ H+ cation exchange occurs, leading to aluminate tetrahedra charge balanced by H+, and an equal number of Na+OH− complexes. Such a configuration can also be thought of in terms of bridging hydroxyl groups (Al-OH(br)-Si). The model proposed was in marked contrast to all previous models in which rupture of T-O-T linkages to give T-OH was universally advocated. Additional data on other framework aluminosilicate compositions (Kohn et al. 1992) were fully consistent with the Kohn et al. 1989a model. Kohn et al. 1992 also attempted to rationalise changes in the vibrational spectra and physical properties of hydrous aluminosilicate melts. More recently Sykes and Kubicki 1993 have proposed an alternative model partially based on MO calculations of aluminosilicate clusters, but also relying heavily on the Kohn et al. 1989a 23Na quadrupole coupling constant data. Sykes and Kubicki 1993 suggested that over the range 0-30 mol% water, the dissolution mechanism involves breakage of Al-O-Al bridges to give Al-OH. Above 30 mol%, they suggested that the mechanism changes to the one proposed by Kohn et al. 1989a. The value of 30 mol%, which is central to the model of Sykes and Kubicki 1993, is based solely on the discontinuity in the 23Na quadrupole coupling constant (Cq) data presented by Kohn et al. 1989a. Other more recent contributions also cite the discontinuity in the 23Na Cq as a crucial piece of evidence in elucidating the water dissolution mechanism in albite melt (Zeng and Nekvasil 1996; Tossell and Sàghi-Szabó 1997; Sykes et al. 1997).
As changes in Cq and the isotropic chemical shift (δiso) for 23Na are central to some models of water dissolution mechanisms, obtaining detailed and accurate knowledge of changes in these parameters is very important. In this paper we will present new data using the techniques of off-resonance nutation NMR spectroscopy and fast magic angle spinning (MAS) over a range of magnetic fields. The technique of off-resonance nutation NMR spectroscopy has recently been used to determine the quadrupole coupling constant Cq (= e2qQ/h, where eq is the maximum component of the electric field gradient at the nuclear site and eQ is the nuclear electric quadrupole moment). In a study of a series of aluminosilicate glasses it was demonstrated that off-resonance nutation can be used in amorphous systems to get information about the average quadrupole parameters (Dirken et al. 1995. Due to the fact that off-resonance nutation mainly relies on line intensities to discriminate between different sites well resolved spectra can still be obtained in systems (e.g., glasses) where a distribution of quadrupolar interactions is present. Thus, the average Cq can be obtained from nutation, but there is no straightforward information about the details (i.e., width and shape) of its distribution.
The other technique we have used is to measure changes in the centre of gravity of fast MAS NMR spectra as a function of the Larmor frequency to extract estimates of both the isotropic chemical shift (δiso) and Cq. This is essentially an improved version of the technique used by Kohn et al. 1989a which was an early example of a multiple field study of a quadrupolar nucleus in an amorphous solid. In the intervening years there have been great improvements in both the understanding of how the NMR interactions are related to the spectral features in amorphous materials and in the experimental technology available, allowing better, more unambiguous extraction of the NMR interaction parameters.
For MAS NMR resonances from quadrupole nuclei where there is a distribution of interaction parameters, both their width and position can be strong functions of applied magnetic field and MAS rate. For a nucleus such as 23Na that experiences a substantial quadrupolar interaction it is usually only the central (1/2, −1/2) transition which is observed and this experiences second-order broadening. This has two important implications for the study of such nuclei in solids; (1) the width (in Hz) is inversely proportional to the applied magnetic field and (2) the maximum narrowing which can be attained using MAS at 54.7° is a factor of about 3.4-4, and to achieve this factor the MAS rate must exceed the residual second-order quadrupolar width. For a glass, where there will be a distribution of environments and hence quadrupolar interactions, slow MAS speed will only narrow those environments with the smallest values of Cq, as these will have the smallest residual width. As the spinning speed is increased more sites narrow but these have larger residual linewidths so that the MAS linewidth increases. This has been observed in other studies of quadrupolar nuclei (e.g., 27Al) in glasses (Sato et al. 1991. To observe the true lineshape it is important that the spinning speed is fast enough to exceed the residual width of the largest Cq-components, and this turned out not to be the case in the original 4.7 T data of Kohn et al. 1989a. Calculations of the MAS NMR lineshapes of quadrupolar nuclei in glasses have shown that the characteristic tail to low frequency which is often observed, results from the distribution of quadrupolar interactions (Phillips et al. 1988, Jäger et al. 1993). The relationship of the peak position to the centre of gravity depends on the size and nature of the distribution, and for such asymmetric lineshapes it is important to use the centre of gravity rather than the peak maximum (which was used by Kohn et al. 1989a) of the resonance to evaluate the mean interaction parameters from the multiple magnetic field data.
Section snippets
Off-Resonance Quadrupole Nutation NMR
The basic idea of the nutation NMR experiment is to examine the response of the spin system to an r.f. pulse. This response is a function of the ratio of the quadrupole frequency νq (= 3Cq/2I(2I-1) where I is the nuclear spin quantum number) and the precession rate of the magnetisation in the applied rf-field ν1 (= γB1/2π, where γ is the gyromagnetic ratio of the nucleus and B1 the rf field strength) (Samoson and Lippmaa 1983). This was refined into a two-dimensional NMR experiment which means
Experimental Details
Nine samples have been investigated: crystalline Amelia albite (from the Natural History Museum, specimen number BM 1962, 439) which was used as a model compound for which δiso and Cq are known; two dry albite glasses, AB5 which was used previously by Kohn et al. 1989a, and AbB4 which was a portion of AB5 glass remelted at 1200°C for this study; four of the hydrous samples of Kohn et al. 1989a AB7, AB4B, AB4A, and AB5G containing 29, 40, 50, and 60 mol% H2O (on an 8 oxygen basis), respectively;
Results
Nutation spectroscopy was used to obtain the quadrupole interaction in these albite glasses. Initially a well characterised sample of crystalline Amelia albite was studied by both MAS and 2D nutation NMR. The MAS NMR spectrum at 7.05 T (Fig. 1a upper) shows a well defined second-order quadrupolar powder pattern which could be closely simulated (Fig. 1a lower) using the parameters previously published in the literature of Cq = 2.6 MHz and δiso = −8 ppm (Phillips et al. 1988). The full
Discussion
The previous 23Na work of Kohn et al. 1989a made three major observations relating to sodium, namely (1) Cq remained constant at around 0.7 MHz up to 30 mol% and then rose sharply to ∼ 1.8 MHz at 60 mol%, (2) there was a continuous change in δiso of +15 ppm with increasing water content, and (3) the dry glass had a considerable distribution of both quadrupolar and chemical shift interactions whereas the 60 mol% H2O glass has a very narrow distribution of Cq and δiso. These observations were
Conclusions
Remeasurement of the 23Na MAS NMR spectra of hydrous albite glasses reveals some modification of the trends presented by Kohn et al. 1989a. We find that the mean isotropic chemical shift changes from −13.4 ppm in the dry glass to −4.4 ppm in the glass containing 56 mol% water. These values represent a small modification from the earlier analysis of a more limited data set, but the broad features of the original and revised analyses are the same, i.e., there is a continuous change in the mean
Acknowledgements
We acknowledge NERC and EPSRC for financial support, in particular for a NERC advanced fellowship to SCK and access to the ultrahigh field NMR facility at the University of Edinburgh (GR/K43667). MES thanks the University of Kent for supporting NMR. We thank the Natural History Museum for providing the specimen of Amelia albite (BM 1962, 439) and Dr P.F. Schofield for electron probe analyses of the sample. Mrs G. Nachtegaal, Mr J. van Os and Mr H. Janssen of the National HF-NMR Facility at
References (47)
- et al.
Complete resolution of Si-O-Si and Si-C-Al fragments in an aluminosilicate glass by 17O multiple quantum magic angle spinning NMR spectroscopy
Chem. Phys. Lett.
(1997) - et al.
A study of the structural role of water in hydrous silica glass using cross-polarisation magic angle spinning NMR
Geochim. Cosmochim. Acta
(1987) - et al.
High-temperature Raman-spectroscopy of silicate and aluminosilicate hydrous glasses - Implications for water speciation
Chem. Geol.
(1996) - et al.
Determination of the distribution of the quadrupolar interaction in amorphous solids by 27Al satellite transition spectroscopy
Solid State NMR
(1993) Off-resonance nutation NMR spectroscopy of half-integer quadrupolar nuclei
J. Magn. Reson. A
(1993)- et al.
Two-dimensional nutation NMR of half-integer quadrupolar nuclei
J. Magn. Res.
(1987) - et al.
A multinuclear magnetic resonance study of the structure of hydrous albite glasses
Geochim. Cosmochim. Acta
(1989) - et al.
The interaction between water and aluminosilicate magmas
Chem. Geol.
(1992) - et al.
Comment on “A model for H2O solubility mechanisms in albite melts from infrared spectroscopy and molecular orbital calculations.” by D. Sykes and J. D. Kubicki
Geochim. Cosmochim. Acta
(1994) - et al.
The speciation of water in haplogranitic glasses and melts determined by in situ near-infrared spectroscopy
Geochim. Cosmochim. Acta
(1995)
Phase relations and compositional dependence of H2O solubility in quartz-feldspar melts
Chem. Geol.
Water and the viscosity of andesite melts
Chem. Geol.
Satellite transition high resolution NMR of quadrupolar nuclei in powders
Chem. Phys. Lett.
2D NMR nutation spectroscopy in solids
J. Magn. Res.
One- and two-dimensional 1H magic-angle spinning experiments on hydrous silicate glasses
Solid State NMR
A model for H2O solubility mechanisms in albite melts from infrared spectroscopy and molecular orbital calculations
Geochim. Cosmochim. Acta
Aluminosilicate and borosilicate single 4-ringsEffects of counterions and water on structure, stability, and spectra
Geochim. Cosmochim. Acta
An associated solution model for albite-water melts
Geochim. Cosmochim. Acta
The effect of water on the structure of silicate-glasses - A neutron-diffraction study
J. Non-Cryst Solids
The Evolution of the Igneous Rocks
Water and magmas, a mixing model
Geochim. Cosmochim. Acta
Relaxation in silicate melts
Eur. J. Mineral.
The effect of water on the viscosity of a haplogranitic melt under P-T-X conditions relevant to silicic volcanism
Contrib. Mineral. Petrol.
Cited by (99)
The influence of H<inf>2</inf>O and SiO<inf>2</inf> on the structure of silicoborate glasses
2019, Journal of Non-Crystalline SolidsQuadrupolar nutation NMR to discriminate central and satellite transitions: Spectral assignments for a Ziegler-Natta catalyst
2017, Journal of Magnetic ResonanceNew insight into atmospheric alteration of alkali-lime silicate glasses
2017, Corrosion ScienceCitation Excerpt :The significant decrease of the quadrupolar coupling constant with alteration, from 4.3 MHz to 1.6 MHz, is reflective of a higher symmetrical environment for the sodium in the hydrated glass, when sodium ions are surrounded by water molecules and/or silanol groups. This was already reported in previous alteration experiments for aluminosilicate glasses [44,45]. The decrease of the isotropic chemical shifts from 3 ppm to −1 ppm suggests an increase of the average Na-O distance [46], possibly as a result of more species in the Na+ surrounding when hydrated.
Applications of Solid-State <sup>43</sup>Ca Nuclear Magnetic Resonance: Superconductors, Glasses, Biomaterials, and NMR Crystallography
2017, Annual Reports on NMR Spectroscopy