5th Silicate Melt Workshop—Introduction

EISEVIER CHEMICAL GEOLOGY ,VCU~DI~O ISOTOPE OEO~CIE,VCE zyxwvutsrqpon Chemical Geology 128 (1996) l-4 5th Silicate Melt Workshop-Introduction About 60 researchers, active in the field of silicate melt/glass research, mainly from Europe, met in La Petite Pierre, Alsace, France, from 4 to 8 April 1995, for the 5th Silicate Melt Workshop (SMW). This special issue of Chemical Geology contains a representative collection of 18 papers presented at this workshop. Why should Earth Scientists be preoccupied with silicate melt research? The answer is that without a fundamental understanding of silicate melts many of the scientific disciplines dealing with the “solid” Earth (e.g., igneous petrology, geochemistry, volcanology, planetology) would have an incomplete scientific foundation. The participants of the SMWs are conscious of the fact that without an understanding of the properties of silicate melts, the efficiency and the rate of processes involving melts and volatiles taking place at surface or the interior of our planet cannot be evaluated. When addressing themselves to physicists or physical chemists in the search for answers, they discover that the physics of amorphous condensed matter is a very young scientific field, where much remains to be learned. Melt formation, transport and selective solidification serve as a chief agent of mass and heat transfer and of differentiation, throughout the Earth since its early beginning. These processes give rise to volcanicity, which is still largely unpredictable. Here we present a collection of papers dedicated uncompromisingly to the description and understanding of the liquid state of silicates. It should provide for Earth Scientists a glimpse of current trends and progress in the field. An important topic in silicate melt research is the glass transition and the glass transition temperature (7”) i.e. the temperature that this transition takes place under conventionally defined conditions. In the Daniel et al. contribution high-temperature Raman spectroscopy is used to observe the spectral changes accompanying the heating of CaAl,O, glass to T > Ta. The spectra were corrected for the effect of temperature up to 2nd order. These authors succeeded in observing for the first time a striking increase in the density of states in a certain part of the spectrum, when T became greater than Tg. Of course, it was known from the specific heat increase with T in the neighborhood of Tg that this should happen, but a nice demonstration, such as this one, has not yet been published. By means of in situ Brillouin scattering, Xu (this issue) measured the acoustic velocities (V, and V,) at high P and T in two glasses. He determined in this way how the temperature changes at which the glass transition takes place in these glasses when one varies the pressure. At the moment there are very few published reliable measurements of the pressure variation of Tg. Holtz et al. analyze the speciation of water dissolved in silicate melts and glasses, this is a continuation of previous work, i.e. Holtz et al. (1992) and Pichavant et al. (1992). Using high-T Raman spectroscopy, they observed water speciation to be temperature dependent at 120°C in quartzofeldspathic glasses. Infrared (FTIR) spectroscopy by Keppler and Bagdassarov (1993) and near-infrared spectroscopy by Nowak and Behrens (1995) demonstrated also that water speciation was temperature dependent down to T < T,. At the 5th SMW, Zotov and Behrens reported the results obtained by neutron spectroscopy of normal and deuterated water dissolved in silicate glasses. Their spectral results indicated the presence of at least two different OH species. The dissolution mechanism of water in silicate melts is of enormous importance, volcanological as well as economical; it is a recurrent topic in the physical-chemical literature. One aspect of water speciation research concerns the 0009.2541/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0009.2541(95)00158-l calibration of the spectroscopic methods used to determine the concentrations of the species. The paper by Behrens et al. in this issue deals with this matter, the authors compare and discuss the results obtained on the same samples in three different laboratories. This work evidences that high precision is only possible if careful calibrations have been performed. In situ high-pressure luminescence spectroscopy of transition elements is the subject of the contribution by Que’rel and Revnard. Traditionally this type of spectroscopy is used to further our understanding of transitionelement partitioning between crystals and coexisting melt. Que’rel and Reynard show that the pressure dependence of the luminescence spectrum gives also information on the structural relaxation in silicate glass at has attracted the attention of many spectroscopists. Evidently, high P. The structure of silicate melts/glasses the structure of an amorphous substance has little in common with that of a crystalline phase. But both structures have many structural elements in common, i.e. bond lengths, bond angles, radial distribution functions, coordination numbers, etc. The study of the structure of amorphous substances is concerned with these elements and the extent of order they show. Silicate liquids and glasses at P = IO5 Pa and T < 2000 K show short range order, i.e. the SiO, tetrahedron is a recurrent feature. But medium range order (MRO), that is to say at distances greater than to the nearest neighbor atoms, may be present to a certain degree. MRO concerns in the first place the inter tetrahedral space in amorphous silicates. Geochemically this space is of great interest because most trace elements and molecular dissolved CO, and Hz0 are found in this region. Cormier et al. describe the spectroscopic methods to study MRO and the results obtained with these methods. This type of research is still in an active stage of development and simplicity is not one of its hallmarks. Proper interpretation of the spectra obtained with these methods requires frequently taking care of anharmonicity of the periodic motions of the atoms in silicate liquids/glasses. This topic is discussed by Farges and Brown. The contributions by Adamkocicoca et al. and De Ligny et al. are the only two on chemical thermodynamics. The first group of authors is very well known for their high-quality calorimetry; they report here new relative enthalpy measurements obtained by high-T drop and solution calorimetry on melts in the system Ca,Mg,Si,O,-Casio,. Such data augment our knowledge of silicate melts and are badly needed to verify proposed thermodynamic models of melts. The great variety of naturally occurring silicate melts necessitates the application of thermodynamic models, such as developed by Ghiorso ( 1985), in order to understand the genesis and evolution of such melts. De Ligny and Richet et al. together with Westrum (a pioneer of adiabatic calorimetry) measure the low-temperature (T > 5 K) enthalpy of several silicate compositions. The importance of such measurements has been spelled out in Richet (1984). De Ligny et al. used adiabatic calorimetry. This may well be the last time that this method is used on material of geochemical and petrological interest. Slowly but steadily adiabatic calorimeters are dying out; they are suffering from a lack of love. When physicists rediscovered the glass transition about 15 years ago, they became aware of the boson peak. This peak is a positive low-frequency deviation, from the Debye model of the density of vibrational states, shown by amorphous materials. The presence of the boson peak is also evident in the low-T (T < 50 K) specific heat of silicate glasses. De Ligny et al. remind us that crystalline silicate compounds may also have abnormal low-T specific heats. They interpret their data in terms of MRO in silicate glasses. The origin of the boson peak is still poorly understood and a subject of hot debate among physical-chemists and physicists. Low-T heat capacity measurements are needed to evaluate the residual entropy, i.e. entropy at 0 K, of silicate glasses. This information is needed to verify quantitatively entropy mixing models for silicate melts, moreover it is a key feature in the configurational entropy theory of Adam and Gibbs (19651, see below. The following series of papers deal with transport properties of silicate melts and glasses. These properties are related to relaxation processes. During the last decade, with the development of Mode Coupling Theory (MCT), structural relaxation in condensed amorphous phases has been intensively studied and the subject of numerous congresses. MCT is a theory about the movements of atoms in liquids. Recent reviews of this theory are by Giitze (1991) and Giitze and Sjiigren (1992). In MCT two types of transport processes have been recognized. namely the (Y and p processes. At high temperature, T > T,, the two processes are indistinguishable. T, is known as the crossover temperature, and is larger than the conventional glass transition temperature. Introduction/Chemical Geology 128 (1996) 1-4 3 At T < T,, the CYprocess is slower than the /3 process; viscous flow is an (Y process while self-diffusion of network-modifying cations is a p process, see Dingwell (1990). MCT has been successful in predicting that at T > T, the viscosity of liquids varies with temperature according to a power law. At T I++T,, the low viscosity of silicate liquids behaves in this way (Bottinga et al., 1995). But at T < c, the complexity of the MCT equations does not permit the formulation of simple relationships expressing the viscosity in terms of T, or P, or composition. Adam and Gibbs (1965) proposed a different way, based on statistical thermodynamics to study the relaxation of glass-forming liquids. They obtained a simple expression for the structural relaxation time, which is inversely proportional to the viscosity, see Bottinga and Rich& (this issue). The Adam and Gibbs (A-G) method is valid at T < T,. In the contribution by Riissler and Sokolor:, MCT is discussed in the light of the dynamics of particles in glass-forming liquids at Tp < T < T,. Most petrological and volcanological processes take place in this temperature interval. Riissler et al. (1996) have shown that the T dependencies of the viscosities of different types of liquids, in the temperature interval T, < T < T,, can be expressed in the same way if this is done in terms of the values of T, and Tg of each liquid. Hess et al. demonstrate that the viscosity of aluminosilicates at different temperatures behaves as described in Rijssler et al. (1996). Perhaps, this work shows a new way to tackle the old problem of the composition dependence of the viscosity of silicate liquids. The need for a better way to evaluate the viscosity of silicate liquids than was proposed by Bottinga and Weill (1972) and Shaw (1972), should be obvious to everybody who has witnessed traffic tie-ups due to the closing of one lane of a four-lane highway when the traffic is dense. In this case we deal with a reduction of the road capacity by 25%, which is equivalent to a 25% increase of the viscosity of the fluid consisting of “car particles” . The precision of the Bottinga-Weill and Shaw methods are not better than a factor two, hence they are inadequate to permit the detailed evaluation of the flow of lava in the volcanic vent during an eruption. Bottinga and Richet summarize briefly the evidence in favor of the A-G theory. Lack of sufficient data on the residual entropy of silicate glasses is one of the reasons that no one has yet tried to solve the problem of the composition dependence of silicate melt viscosities in the frame of the A-G theory. Webb and Knoche have shown by a very careful analysis of data on the specific volume, enthalpy and shear viscosity in silicate melts, that all these physical properties are relaxed by the same mechanism. Hence the relaxation of these properties is characterized by the same relaxation time, namely the characteristic time for structural relaxation. The structural relaxation time is determined by the state of the structure of the liquid. By means of the A-G theory one can calculate this structural relaxation time, see Bottinga and Richet in this issue. The effect of dissolved water on the structural relaxation time, and thus on the viscosity, of an andesite has been measured. Richer et al. report the variation of the viscosity of this lava with water content. The importance of these measurements for the understanding of volcanological processes is evident. Lisku et al. have measured how the viscosity of Na,Si,O, melts are affected by additions of up to 10 oxide mole% TiO,. The authors concluded that their observations agreed with the A-G model. The pressure dependence of the diffusion of Ar and Kr in a jadeite melt at 800°C was determined by Roselieb et al. Diffusion at T < T,, of Ar and Kr is a /3 process. It is well known that such a process has an Arrhenian temperature dependence. The measurements of Roselieb et al. indicate that also the pressure dependence of the Ar and Kr diffusion constants is Arrhenian. Jaupart discussed how volcanic eruptive processes are affected by the physical properties and their P and T dependencies of lavas. He stresses the need for measurements of the solubility of water at low pressures because the amount of exsolved gases determine the eruption regime. The kinetics of degassing is very important because it is unlikely that during volcanic eruptions equilibrium conditions prevail. The last paper in this collection is by Knoche and Luth. The title of this contribution speaks for itself. It is about the possibilities and pitfalls of the sink/float method at high P to determine silicate liquid density. Obviously, to be a successful expetimentalist one has to be tenacious. Accounts of these silicate melt workshops are published, because we feel that Earth Scientists are not sufficiently acquainted with the properties of silicate melts. This is well illustrated by the fact that in frequently 4 Introduction / Chemical Crology I28 (1YY6) l-4 used computer models of volcanic eruptions no realistic evaluation of the viscosity of the phenocryst and bubble carrying magma is attempted. Either it is assumed that the viscosity is constant or that it varies with temperature in Arrhenian fashion. No one seems to be aware that both hypotheses are unwarrantable. Besides it is highly questionable if the viscosity values measured under steady-state conditions are of great value when one is dealing with eruptive processes, which do not give rise to a steady state. A sufficiently viscous silicate liquid put under stress shows important viscosity transients, which decay during a time interval much longer than the time it takes for the magma to traverse the distance between magma chamber and volcano crater. The organizers would like to thank all the participants for their interesting contributions and their willingness to take part in frank discussions. We have appreciated very much how the reviewers of the different papers presented in this issue accomplished this anonymous and selfless task. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON References Adam, G. and Gibbs, J.H.. 1965. On the temperature dependence of cooperative relaxation properties in glass-forming liquids. J. Chem. Phys., 43: 139-146. Bottinga, Y. and Weill, D.F., 1972. The viscosity of magmatic silicate liquids: A model for calculation. Am. J. Sci., 272: 438-475. Bottinga, Y., Richet, P. and Sipp, A.. 1995. Viscosity regimes of homogeneous silicate melts. Am. Mineral., 80: 305-318. Dingwell, D.B., 1990. Effects of structural relaxation on cationic tracer diffusion in silicate melts. Chem. Geol., 82: 209-2 16. Ghiorso, MS., 1985. Chemical mass transfer in magmatic systems, I. Thermodynamic relations and numerical algorithms. Contrib. Mineral. Petrol., 90: IO7- 120. G&e, W., 1991. Aspects of structural glass transitions. In: J.P. Hansen, D. Levsque and J. Zin-Justin (Editors), Liquids. Freezing and Glass Transition, Part I. Elsevier, Amsterdam, pp. 292-503. G&e, W. and SjBgren, L., 1992. Relaxation processes in supercooled liquids. Rep. Progr. Phya.. 55: 241-376. Holtz, F., Behrens, H., Dingwell, D.B. and Taylor, R.P., 1992. Water solubility in aluminosilicate melts of haplogranite composition at 2 kbar. Chem. Geol.. 96: 289-302. Keppler, H. and Bagdasaarov, N.S., 1993. High-temperature FIIR spectra of Hz0 in rhyolite melt to 1300°C. Am. Mineral., 78: 1324-1327. Nowak, M. and Behrens, H., 1995. The speciation of water in haplogranitic glasses and melts determined by in situ near-infrared spectroscopy. Geochim. Cosmochim. Acta, 59: 3445-3450. Pichavant, M., Holtz, F. and McMillan, P.F., 1992. Phase relations and compositional dependence of H,O solubility in quartz-feldspar melts. Chem. Geol., 96: 303-320. Richet, P., 1984. Viscosity and configurational entropy of silicate melts. Geochim. Cosmochim. Acta, 48: 47 l-483. Riissler, E.. Hess, K.-U. and Novikov, V.N., 1996. The concept of fragility revisited. J. Non-Cryst. Solids (submitted). Shaw, H., 1972. Viscosities of magmatic silicate liquids: An empirical model of prediction. Am. J. Sci., 272: 870-893. Y. BOTTINGA. D.B. DINGWELL and P. RICHET (Guest-Editors)