Dave Walker1, S. Clark2, and R.L. Jones3
1Lamont-Doherty Earth Observatory, Columbia University, Palisades NY 10964, USA
2Advanced Light Source, Ernest Orlando Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720, USA
3CLRC Daresbury Laboratory, Daresbury, Warrington WA4 4AD
dwalker@ldeo.columbia.edu, smclark@lbl.gov, r.l.jones@dl.ac.uk
ALS
INTRODUCTION
Chlorides in the series between NaCl (halite) and KCl (sylvite) show a stable solvus with consolute temperature of ~500 ºC and a consolute composition near X(NaCl) = 0.65 at 1 bar pressure [1,2,3]. Recent P-V-T equation of state work in this series performed in the large-volume press of station 16.4 of the CLRC Daresbury Synchrotron Radiation Source has shown the solvus to expand markedly, about 80 ºC in the first 18 kbar of applied pressure [4]. Parameterized values of the thermal expansion recovered by fitting the high-pressure data from Daresbury across the compositional series showed an unexpected maximum at intermediate compositions near the consolute composition [5]. Confirmation of this maximum recovered by parameterized fitting at high pressure was sought by direct measurement at low pressure of the thermal expansions across this compositional series. This experiment was more convenient to perform at the ALS than at DL. The thermal expansions observed at ALS confirm the parameterizations found at DL. Furthermore the new data allow a favorable comparison of the observed expansion of the solvus with pressure with that predicted from low-pressure volumes.
EXPERIMENTAL PROCEDURES
Mixed K-Na crystalline chlorides were synthesized from reagent chemicals by melting and quenching. These mixed chlorides exsolve on quench into 2-phase mixtures at intermediate compositions within the solvus. Single-phase chlorides were recovered for unit cell measurements by heating them within a heating stage [borrowed from the DL-MRL installation] mounted behind the monochromator on ALS station 7.3.3 and about 184 mm in front of the image plate. The samples were suspended in the heating stage in 0.2-0.3 mm glass capillary tubes and illuminated with X-rays of wavelength 1.1169 Å. Sample homogenization with temperature could be monitored directly by X-ray diffraction. Data collection was usually initiated at 650 ºC and XRD spectra were taken at 50 ºC intervals down temperature until exsolution was encountered. XRD spectra collection times of 3-5 minutes were sufficient to enable least squares refinement of the cell parameters of the single-phase chlorides to almost 4 significant figures from observation of about 5 peaks. To make this precision meaningful it was necessary to admix an internal standard [MgO] to correct for variations of the sample to image plate distance. Tabulated vales of MgO volume with temperature were taken from [6]. The heating stage performed well at the high-temperature end of the data collection range but unacceptably large temperature fluctuations and gradients developed at temperatures below 350 ºC.
RESULTS
Cell volumes for the compositional end-members and 9 intermediate compositions are given as a function of temperature in Figure 1. Figure 2 gives the thermal expansions. Figures 1 and 2 Thermal expansion for the end-members halite and sylvite is in good agreement with the values in [6]. The volumes of the intermediate compositions at room temperature are the metastable values given by [2]. Thermal expansion computed for the intermediate compositions from the high temperature ALS data alone are greater than values computed from the average of volume changes from room temperature to, for example, 500 ºC. This is to be expected for thermal expansions that increase with temperature and is consistent with the negative curvature shown for the volume curves in figure 1. The fact that the end-member curves for halite and sylvite are lumpy and linear is a reflection of the poor temperature control at low temperature. Even so the average low-temperature thermal expansion is indistinguishable from accepted values in [6]. The parameterized thermal expansion derived from e.o.s. fitting of high P-T Daresbury data show a maximum in thermal expansion 'o extrapolated to room temperature. This unexpected maximum was the proximal cause for our investigation of room-pressure thermal expansion directly. It is clear that the parameterized maximum in 'o is even more striking in the higher temperature thermal expansion. The average of the expansion over the T interval from the room temperature data of [2] to the ALS data gives a lower average thermal expansion than just the high-temperature ALS data alone. No matter which way the data is sliced, the maximum in thermal expansion at intermediate compositions is confirmed. It becomes more prominent at high temperature. The use of the ALS volume data at high temperature across this series allows a prediction of the rate of increase of the consolute temperature with 18 kbar of pressure of about 45 ºC which is of the same order as the value of 80 ºC observed by [4]. By contrast, use of the volume values at low temperature and pressure (an incorrect but necessary procedure in the absence of high temperature data) predicts a temperature rise in the consolute point of only 8 ºC over the same 18 kbar pressure interval. We are unaware of previous tests of the theory of [7] on the pressure dependence of consolute phenomena.
ACKNOWLEDGMENTS
Thanks to Lachlan Cranswick of CCP14, Pramod Verma of Delhi University, and Stephan Buhre of Frankfurt University for their role in the DL stages of this project. Support from the ALS, CLRC, and the NSF are gratefully acknowledged.
REFERENCES