High Pressure

Introduction

This year again many high-pressure experiments have been performed on a range of beamlines using a variety of techniques, more particularly diffraction, spectroscopy and inelastic scattering. This section includes reports of diffraction studies carried out on the high-pressure beamline (ID30), and on ID9 where half the beam time is used for high-pressure diffraction studies. An example of high-pressure studies by X-ray absorption spectroscopy performed on BM29 is also given.

 

High-pressure, high-temperature diffraction studies

The range of studies carried out on ID30 is mainly concerned with diffraction in extreme conditions of pressure and temperature, and this year the research programme has benefited from three important new developments: On the X-ray side, new 170 mm mirrors in a Kirkpatrick - Baez configuration deliver a monochromatic beam of about 5 µm x 10 µm. This beam has permitted the extension of angle-dispersive diffraction studies using a diamond-anvil cell (DAC) beyond 100 GPa (1,000,000 bar), and this with both resistive and laser heating of the sample. On the high-pressure side, the large-volume cell programme using the Paris-Edinburgh cell has been firmly established, and extended to diffraction studies of liquids. On the detector side, the performance of the new fast-scanning on-line image plate system has greatly speeded up data collection, with typical exposure times varying from 20 seconds to 10 minutes, and made new studies possible, in particular the kinetics and in-situ characterization of reactions. ID9 uses a 30 µm x 30 µm focused monochromatic beam with large area image plates and an off-line scanner for angle-dispersive (AD) powder diffraction (see the section on Silicon-VI below). This set-up is best suited to high-resolution studies up to 100 GPa. Beside AD diffraction, both beamlines can operate in the energy-dispersive mode. This technique is mostly used for single-crystal studies of light elements and compounds at very high-pressures.

 

The Paris-Edinburgh press

The large-volume cell currently in use at the ESRF is illustrated in Figure 73. With such a device, stable pressure and temperature conditions can be applied to a sample assembly of up to 100 mm³. The sample is typically 0.5 mm in diameter, 2 mm in length, and it is surrounded by a heater-gasket assembly up to 10 mm in diameter in the geometry shown in Figure 74. With high resistivity graphite heaters, the sample temperature is raised to 1800°C by passing a current of 100 A through the anvils, giving a maximum power of 250 W. Using high-energy monochromatic radiation and the fast-scanning image plate detector on ID30, images of the sample and its surroundings, and of the surroundings only, taken by translating the press, are obtained rapidly and subtracted from one another to extract a clean spectrum from the sample. The data obtained are of sufficiently high quality to allow structural refinement.

 

X-ray diffraction studies of liquid iron and iron-sulfur alloys

In order to better understand the structure and dynamics of planetary liquid cores, it is very important to obtain information on the structure of liquid iron, which makes up, for example, up to 90 percent of the weight of the terrestrial liquid core, and on the role played by light elements such as sulfur in these liquids. Using the large-volume press described above, high-energy monochromatic radiation (70 to 90 keV depending on the experiment), and the fast-scanning image plate system, the melting of iron and of iron - sulfur alloys has been studied on beamline ID30 over pressure and temperature ranges of up to 5 GPa and 1800°C. Images obtained just before and just after melting are shown in Figure 75. This technique gives the melting curve of iron with a high precision, and the integration of the diffraction images gives the radial distribution function from the first and second diffraction rings. In Fe-S alloys, the diffraction rings are broader and more diffuse, attesting to the lower degree of order in liquid iron in the presence of sulfur. Furthermore, these first studies were conducted for different Fe - S compositions, in particular Fe plus 27% S and Fe plus 20% S, and this technique was sensitive enough to correlate composition and degree of order.

This has demonstrated that X-ray diffraction is a very promising method for the study of the structure of metallic liquids, their evolution under static pressure, and more particularly, to correlate chemical compositions and structure in the Fe-S system.

Publication
C. Sanloup (a), P.Gillet (a), G.Fiquet (a), F. Guyot (b), I. Martinez (b), M. Mezouar (c), D. Häusermann (c), to be published.

(a) Laboratoire de Sciences de la Terre, ENS, Lyon (France)
(b) Institut de Physique du Globe, Paris (France)
(c) ESRF

 

 

First in-situ X-ray characterization of mercury cuprate superconductors synthesis at high-pressure-temperature

Amongst the superconducting mercury cuprate compounds with the general formula HgBa₂Ca_n-1Cu_nO_2n+2+d (n = 1 up to 6), the member with n = 3 has the highest critical temperature of 133 K. Single phase samples of the compounds with n > 3 can only be prepared by high-pressure, high-temperature synthesis, and the preparation of high-quality single phases, free from intergrowths, becomes increasingly difficult as n increases. Hence there is a need to improve the synthesis process which involves the high-pressure reaction of HgO with a precursor, a pre-reacted mixture of Ba, Ca and Cu oxides. From the work undertaken so far, it has been proposed that the formation of each phase proceeds through an intercalation starting from the lower members of the series. Thus the reaction starts with the formation of Hg-1212 and CaHgO₂, which is then converted into higher terms by incorporating CaCuO₂, and in this scheme, longer times or higher temperatures favour the formation of higher members of the series. However, there is no direct evidence of this mechanism, and the aim of the work reported here was to attempt to clarify the question.

The large-volume cell and sample arrangement described in Figures 73 and 74 were used to investigate the synthesis mechanism, but in order to reproduce the laboratory synthesis conditions, the initial mixtures of oxides were sealed in gold capsules to avoid reactions with external media. High-quality diffraction images were recorded in 80 seconds using unfocussed 77 keV wiggler radiation and the ID30 "Fastscan" image-plate detector. This fast data collection rate gave access, for the first time, to the kinetics of phase formations and chemical reactions at high-pressure and high-temperature.

The first experiments focused on the synthesis of HgBa₂CuO_4+d(Hg-1201) and HgBa₂Ca₂Cu₃O_4+d (Hg-1223) phases, investigated the influence of temperature, dwell time and pressure on the synthesis process. All the phases involved were identified, and their proportions determined by Rietveld refinement. These experiments, carried out at 2, 3 and 4 GPa, revealed that the Hg-1201 phase is formed via an intermediate phase, that the Hg-1223 phase is formed via the Hg-1212 phase, and also that pure Hg-1201 and Hg-1223 samples are obtained at 3 GPa. The formation of Hg-1223 is illustrated in Figure 76. These first experiments have not only demonstrated the feasibility of in-situ diffraction studies of complex oxide synthesis at high-pressure and high-temperature, but even more importantly, they open the way to the study of phase formation and chemical reaction kinetics at high-pressure and high-temperature.

Publication
S. Le Floch (a), P. Toulemonde (a), A. Prat (a), J.J. Capponi (a), P. Bordet (a), M. Mezouar (b), D. Häusermann (b), to be published.

(a) Laboratoire de cristallographie, CNRS, Grenoble (France)
(b) ESRF

 

 

Crystallography in extreme conditions of pressure and temperature: A structural study of MgSiO₃

The study of the deep Earth has been motivating generations of scientists who have to take up the challenge given by the extreme conditions existing at the centre of the Earth: over 300 GPa and 5000 K, and as the study of the propagation of elastic waves created by earthquakes only allows the determination of the density profile of our planet, the all-important problem of the determination of the chemical composition and crystalline structures existing in the deep Earth remains. (These govern the Earth global exchanges such as thermal regimes, convection drifts, plate tectonics). Relationships have to be established between chemical composition, crystalline structure and specific volume over the whole range of pressures and temperatures existing within the planet, and X-ray diffraction is by far the best technique to obtain reliable structural and molar volume data on the compounds and materials of interest.

Extreme conditions of static pressure and temperature are obtained by combining laser heating with diamond-anvil cells, and although high P-T conditions of 200 GPa and 6000 K have already been achieved with this technique, very few diffraction experiments have been performed in these conditions [1]. Difficulties in obtaining uniform heating, reliable temperature measurements and very high-pressures require the samples to be a few microns in thickness and a few tens of microns across, and hence require the use of micro-beams. The work reported here was carried out using a focussed monochromatic beam of about 10 µm x 15 µm on beamline ID30.

The perovskite form of (Mg,Fe)SiO₃ is currently accepted as the dominant phase of the Earth's lower mantle (700 to 2900 km deep), and its equation of state (EOS) plays an important role in many fields of geophysics. It is however presently impossible to choose between the perovskite-pure and perovskite-magnesiowüstite (Mg,Fe)O models for the Earth's lower mantle on the basis of the existing data, and in-situ high P and T diffraction is the only method available to measure correctly its EOS and solve the stuctural problem. Previous studies were conducted in the stability field of the perovskite, but energy-dispersive diffraction and large-volume presses limited the performance and P,T ranges to 30 GPa and 2000 K respectively. Using the technique described below, measurements on MgSiO₃ were extended to 86 GPa and 2700 K, and here again the measurements were greatly facilitated using the "Fastscan" detector.

Silicate perovskite MgSiO₃ samples were synthesized from synthetic MgSiO₃ enstatite crystals or synthetic MgSiO₃ glass mixed with platinum powder, and once loaded in a large-aperture DAC, the starting materials were transformed at high P using either CO₂ or YAG infrared lasers, depending on the pressure transmitting medium. The temperature was determined by analysing the thermal emission spectra recorded during the diffraction measurements, whilst the pressure conditions were calculated from the EOS of platinum, used here as internal pressure calibrant. LeBail profile refinements were performed on the diffraction patterns to obtain reliable high P, high T cell parameters for the sample and the pressure calibrant up to 86 GPa and 2700 K. The most remarkable result was that Rietveld structural refinements were successfully carried out on selected patterns in these extreme conditions [2], as Figure 77 shows. This gave, for the first time, precious structural information on these compounds, as for instance the first observation of the increase of the internal distortion of the SiO₆ octahedra with increasing pressure in a powder sample. Furthermore, the data analysis allowed the identification of a set of thermoelastic parameters to constrain the compositional model of the Earth's lower mantle. Assuming that the thermoelastic parameters obtained from this study are applicable to perovskites with moderate iron content, then the comparison of the density and KT profiles calculated for a mixture of perovskite and magnesiowüstite with the current accepted Earth model indicates that a pure perovskite lower mantle is very unlikely. On the other hand, a very good match between the model density and KT profiles is obtained for a mixture of 83 vol% (Mg_0.93,Fe_0.07)SiO₃ perovskite and 17 vol% (Mg_0.79Fe_0.21)O magnesiowüstite.

References
[1] D. Andrault, G. Fiquet, M. Kunz, F. Visocekas and D. Häusermann, Science, Vol. 278, pp. 831, (1997).
[2] G. Fiquet, D. Andrault, A. Dewaele, T. Charpin, M. Kunz and D. Haüsermann, Phys. Earth Planet. Int., in press.

Publication
G. Fiquet (a), A. Dewaele (a), D. Andrault (b), T. Charpin (b), F. Visocekas (b), D. Häusermann (c), M. Kunz (c, *), T. LeBihan (c), to be published.

(a) Laboratoire de Sciences de la Terre, Ecole Normale Supérieure de Lyon (France)
(b) Laboratoire de Minéralogie-Cristallographie, Institut de Physique du Globe, Paris (France)
(c) ESRF
(*) now at the Laboratory of Crystallography, ETH Zentrum, Zürich (Switzerland).

 

 

 

Modulated phases and proton centering in ice up to 170 GPa

Structural studies of ice under pressure have mostly been the domain of neutron measurements since these can 'see' protons and deuterons. Such measurements have recently been extended to 20 GPa using cells similar to that shown in Figure 73. The predicted symmetric form of ice, known as ice X, in which the proton lies midway between the two oxygen atoms remains, however, far out of neutron reach. Using the bright wiggler beams of beamlines ID9 and ID30, accurate structural data have been obtained using the single-crystal diffraction technique described in reference [1], and this work shows that the non-molecular symmetric state of ice should be stable above 150 GPa. The (111) diffraction line assigned to the hydrogen atoms has been measured up to 170 GPa, thus confirming that X-rays can also "see" hydrogen atoms at the ESRF! The change with pressure of the intensity of the (111) line presents a sequence of transitions which agrees with the findings of a recent quantum simulation. Ice VII, the stable room temperature phase of ice above 2.2 GPa, was found to be unexpectedly ordered in a sequence of spatially modulated phases between 2.2 GPa and 25 GPa. This being similar to the phase diagram of the Ising model, its incommensurate phases are suggested [2].

Single-crystals of Ice VII were grown from a single germ by optically looking at the solid-fluid equilibrium on its melting curve. The single crystals were embedded in a quasi-hydrostatic cushion (a gold ring, as helium cannot be used here since it forms clathrates) to reduce its fragmentation up to the 100 GPa range. Good hydrostatic conditions could be achieved up to 20 GPa, and the equation of state (EOS) was measured up to 170 GPa, despite a significant deterioration in hydrostaticity at very high-pressures.

Single-crystal diffraction spectra associated with the (222) reflection are shown in Figure 78, and they also contain two other reflections. The (111) reflection has been observed in neutron measurements, and fleetingly in powder data obtained on beamline ID9 in the past, but the superlattice reflection in the 10 to 20 keV region was unexpected. Its corresponding d-spacing value differs slightly from the value expected from a doubling of the unit cell, which indicates a periodic modulation of the structure along the (111) direction. The evolution of this reflection, labelled S1/21/21/2, with pressure was studied on three large single crystals in good hydrostatic conditions.

The S_{1/2, 1/2, 1/2} peak could be measured in all the four (111) directions with identical d-spacings, within the experimental error bars, and very similar intensity. No distortion from the cubic structure was observed. The periodicity of the modulation (d_{1/2, 1/2, 1/2}) / (4 x d₂₂₂) versus pressure is plotted in Figure 79. The evolution is continuous and very reproducible in the three samples up to 8 GPa, and is independent of the orientation of the crystal with the load axis or the thickness of the gold ring. Also, the S_{1/2, 1/2, 1/2} peak was observed in each run at the starting pressure of the growth of the single crystal from the melt. These two last observations rule out the possibility that this new peak might be provoked by uniaxial stress, and hence it is likely that the modulated phases are intrinsic properties of ice VII. Between 10 and 20 GPa the modulation shows flat steps, as predicted by theories of modulated systems, and it is interesting to note that this corresponds to the pressure range where the S_{1/2, 1/2, 1/2} peak has a maximum intensity (Figure 78). As seen in the inset of Figure 78, the (111) peak could be followed up to 170 GPa. Structure factor calculations show that its intensity can be considered as entirely due to the hydrogen atoms since the contribution of the oxygen atoms, when taking into account possible oxygen disorder, is at most 1% of the total intensity. The (111)/ (222) intensity ration, I₁₁₁/I₂₂₂, should be quite accurate in the current single-crystal measurements.

Recent quantum ab-initio simulations have shown dramatic quantum effects of the proton in ice that lead to a sequence of phase transitions. The same sequence of phase transitions was observed in this work, and the expected corresponding shape of the proton density distribution along the (111) direction is drawn in Figure 79, namely: (a) orientationally-disordered ice VII, with a proton site 50% occupied, (b) translationally-disordered asymmetric ice VII, with the proton tunnelling between the two sites, (c) proton-disordered symmetric ice, in which the zero-point fluctuations lead to the proton distribution function peaked at the bond midpoint, and (d) symmetric ice (ice X), with a single-well potential and the proton increasingly localized.

A detailed discussion of these results is beyond the scope of the present report (see reference [2]), but one can say that reality is between the simple linear extrapolation and the sophisticated quantum simulation. Indeed, there is still some way to go before the quantum effects of proton-transfer systems are quantitatively understood.

References
[1] P. Loubeyre et al. X-ray diffraction and equation of state of hydrogen at megabar pressures. Nature 383, 702-704 (1996).
[2] P. Loubeyre, R. LeToullec, E. Wolanin, M. Hanfland and D. Häusermann, Nature, in print.

Publication
P. Loubeyre (a), R. LeToullec (c), E. Wolanin (a), M. Hanfland (b), D. Häusermann (b), to be published.

(a) CNRS and Université Pierre et Marie Curie, Paris (France)
(b) ESRF

 

 

 

High-resolution powder diffraction at high pressures: crystal structure determination of silicon-VI

The high-pressure phases of silicon have been studied extensively, both experimentally and theoretically (for an overview see [1]). At least ten different modifications are known for Si. All high-pressure phases which are stable above 12 GPa are metals. An unsolved problem is the crystal structure of Si-VI, one of the superconducting phases which occurs near 40 GPa, intermediate between the primitive hexagonal (ph) and hexagonal closed-packed (hcp) phase [2]. The structure of Si-VI was first assumed to be double-hexagonal close-packed; early theoretical studies suggested a restacking of hcp layers, and orthorhombic structures with four atoms per unit cell were also proposed. A more recent synchrotron diffraction study [3] showed that all structural assignments were incorrect, but failed to find a satisfactory structural solution.

Previous X-ray powder diffraction studies [2, 3] were performed using the energy-dispersive (EDX) mode. Its disadvantages (low resolution and intensity information not suitable for determining atomic positions) are tolerable when high-pressure phases with small unit cells and atoms on symmetry-fixed positions are investigated, as for instance ph-Si (Z = 1 atom/unit-cell) or hcp-Si (Z = 2). For solving structures with larger unit cells, low symmetry, and/or atoms on variable positions, better resolution and reliable intensity information is needed for a full structure determination. The structure of the so-called Imma-phase (Z = 4), another high-pressure phase of Si which is stable near 14 GPa, was, for example, solved by angle-dispersive (ADX) powder diffraction [4]. Resolution and intensity determination are considerably improved in ADX compared to EDX experiments due to the use of monochromatic radiation, focusing optics and image-plate detectors, a combination well adapted to diffraction experiments on the small samples of typically 100 µm mounted in diamond anvil high-pressure cells (DAC).

The angle-dispersive powder diffraction set-up at the ID9 beamline of the ESRF delivers ~ 10¹¹ photons/s into a 30 x 30 µm² spot at 27 keV. Part of this immense gain in flux compared to first and second generation synchrotron sources is utilized to improve the resolution by placing large image plates (area: 350 x 430 mm²) further away from the sample. At a sample-to-image-plate distance of 450 mm, an angular resolution of less than 0.04° in 2-Theta can be achieved.

Using the superb experimental conditions at ID9 we have performed a diffraction study of Si in the pressure range from 30 to 50 GPa in order to determine the crystal structure of Si-VI. Si powder was loaded into the gasket hole of a DAC using condensed argon as a pressure transmitting medium, which provides nearly hydrostatic pressure conditions. A diffraction pattern of Si-VI measured at 42.5 GPa is shown in Figure 80. At this pressure weak diffraction lines from hcp-Si were also observed. The Si-VI pattern can be indexed assuming an orthorhombic unit cell containing 16 atoms. Lattice parameters are a = 7.9686(8) Å, b = 4.7759(5) Å and c = 4.7546(5) Å resulting in an atomic volume of 11.308(3) ų. Earlier studies [2, 3] failed to determine the correct unit cell, because the weak (200) reflection (Figure 80, upper part) as well as the (020) and (002) reflections (Figure 80, lower part) essential for finding the unit cell were not observed. The (112), (220) and (202) reflections contributing to the strongest diffraction peak, another hint of a large orthorhombic unit cell, were not resolved.

Based on the powder diffraction diagrams of Si-VI, it was possible to determine the full crystal structure. In short, systematic extinctions of diffraction lines indicated space group Cmca, where Si atoms occupy two different Wyckoff positions with 10- and 11-fold coordination, respectively. The three free positional parameters were determined from full profile refinements. The crystal structure is shown in Figure 81.

The structure of Si-VI is isostructural to that of Cs-V, a high-pressure modification of cesium, whose structure was solved only very recently from diffraction patterns also measured in the angle-dispersive mode at ID9 [5]. Thus, the Cmca structure with 16 atoms per unit cell, which is a new structure type for elemental solids, is not unique to Cs-V with a predominantly d-like character of the conduction electrons, but also occurs in Si-VI, which is essentially a sp-band nearly-free electron metal. Our preliminary diffraction studies of Ge above 100 GPa (done in collaboration with K. Takemura) indicate that the Cmca structure is at least consistent with the observed diffraction data of the phase Ge-VI.

In summary, the high resolution and improved sensitivity resulting from the use of high-intensity monochromatic synchrotron radiation from a third generation source is indispensable for a reliable determination of high-pressure crystal structures. The results presented here not only illustrate the important progress made recently in this direction, but also lead to new insights in crystal structure systematics. The Cmca (Z = 16) structure can be viewed as a distorted face-centered cubic structure. First-principle band structure calculations are in progress in order to understand the energetics of the phase stability of the Cmca (Z = 16) structure and details of the chemical bonding.

References
[1] R. J. Needs and A. Mujica, Phys. Rev. B 51, 9652 (1995).
[2] H. Olijnyk, S. K. Sikka, and W. B. Holzapfel, Phys. Lett. 103A, 137 1984).
[3] S. J. Duclos, Y. K. Vohra, and A. L. Ruoff, Phys. Rev. B 41, 12021 (1990).
[4] M. I. McMahon and R. J. Nelmes, Phys. Rev. B 47, 8337 (1993).
[5] U. Schwarz, K. Takemura, M. Hanfland, and K. Syassen, Phys. Rev. Lett. 81, 2711 (1998).

Publication
M. Hanfland (a), U. Schwarz (b) and K. Syassen (b), to be published.

(a) ESRF
(b) MPI für Festkörperforschung, Stuttgart (Germany).

 

 

 

The physics of fluid iodine under extreme conditions

It is well known that diatomic molecular systems under high-pressure exhibit a transition to a metallic atomic phase when the distances between non-bonded atoms become comparable with the molecular bond length. This behavior has been observed in various molecular crystals and particularly in halogens such as I₂, Br₂ and IBr where the molecular dissociation occurs at relatively "low pressure'' (i.e. in the pressure range between 20 and 90 GPa). Moreover, before reaching the atomic phase, these systems exhibit a semiconductor to semi-metal transition.

Similar transitions have been observed recently in fluid iodine and hydrogen. In particular the metallization transition, as detected by the abrupt increase of the electrical conductivity (about 3 orders of magnitude), has been observed for both I₂ and H₂ around 3 GPa and 140 GPa respectively. Moreover, in the case of iodine, a second transition, ascribed to molecular dissociation, has also been observed at 4 GPa.

It should be noted that these transitions occur at pressures much lower in the fluid phase than in the solid phase (i.e. 3-4 GPa versus 16-21 GPa for iodine and 140 GPa versus somewhere above 340 GPa for hydrogen). A lower metallization pressure in the fluid phase is surprising and an explanation of this effect is not available at present; there is also a substantial lack of experimental data.

The availability of high-brilliance high-energy X-ray absorption spectrometers has opened new opportunities for studies of condensed matter under extreme conditions and, in particular, high-pressure X-ray absorption spectroscopy investigations at the I K-edge have become feasible with the availability of the ESRF source. X-ray absorption spectroscopy (XAS) provides a rather powerful experimental probe for molecular bonding properties through the analysis of the X-ray absorption fine structure (XAFS), and, in addition, simultaneous information on the density of unoccupied states can be achieved by the analysis of the edge shape.

The possibility of measuring accurately the I2 bond length distribution and revealing subtle differences among various condensed phases was demonstrated first [1]. Such a sensitivity was exploited in various subsequent experiments in different regions of the phase diagram. The I₂ bond length in the liquid phase along the liquid-vapour coexistence curve was accurately measured [2] using a quartz coaxial cylinder cell contained in an autoclave system to compensate for the internal iodine pressure.

Another set of experiments was performed in the 2-50 kbar pressure range [3] using a novel high-pressure X-ray absorption technique based on the use of the Paris-Edinburgh press, developed first at LURE [4] and successively at the ESRF. The BM29 set-up was improved with the invaluable technical help of R. Weigel to combine X-ray absorption and X-ray diffraction detection, for pressure calibration, and to achieve an automatic pressure and temperature control.

Several samples have been measured following different pressure and temperature cycles (as shown in Figure 82). It was possible to follow the melting and recrystallization of I₂ through the disappearance/reappearance of Bragg peaks up to 20 kbar. From these measurements it has been possible to determine the bond length expansion both in the solid and liquid phases, as a function of pressure (Figure 83). The comparison between bond length variations shows that it is stronger in the liquid than in the solid, which explains why the pressure needed to induce the dissociation is smaller for liquid iodine.

The main result of these experiments has been the observation of the anomalous bond length expansion with density in liquid iodine with respect to the solid phase [3]. The X-ray absorption results, together with optical absorption measurements [5], show that the density-dependence of both electronic and structural properties of the molecule is much stronger in the liquid phase than in the solid, at least in the high density fluid. The density appears to be the "critical'' thermodynamic parameter even in the fluid phase. The role of temperature seems to be mainly confined to that of making the system disordered thus allowing local configurations, forbidden in the solid phase, with lower binding energy and larger overlaps of the electronic clouds [2].

The experimental results suggest that, in the transition region, fluid I₂ follows a path similar to that calculated for fluid H₂ at much higher temperatures and pressures. The first step of the transition is a large, density-driven increase in molecular ''ionization'', induced by the strong intermolecular interactions brought about by the disorder of the fluid phase, which produces a rapid increase of the electrical conductivity. A further increase in density quickly drives the system to molecular dissociation.

The general validity of this mechanism for metallization in diatomic fluids will be confirmed by the extension of these studies to other halogens (like Br₂ and IBr) in the high-pressure, high-temperature region. It should be emphasized that an accurate bond length determination such as those obtained by XAFS can hardly be achieved using X-ray diffraction. In fact, in the solid phase, reliable intensities of the diffraction peaks are needed, while for the liquid phase, the required measurement of the iodine structure factor up to k = 15-20 Å^-1 is not really feasible even for standard conditions.

All these results demonstrate the potential of X-ray absorption spectroscopy (XAS) to probe the changes of the intra-molecular bond length distribution as a function of pressure and temperature and as a probe for the inter-molecular interactions in molecular fluids.

References
[1] U. Buontempo, A. Di Cicco, A. Filipponi, M. Nardone, and P. Postorino, J. Chem. Phys. 107, 5720 (1997).
[2] U. Buontempo, A. Filipponi, P. Postorino, and R. Zaccari, J. Chem. Phys., 108, 4131 (1998).
[3] U. Buontempo, A. Filipponi, D. Martinez-Garcia, P. Postorino, M. Mezouar, and J. P. Itiè, Phys. Rev. Lett. 80, 1912 (1998).
[4] Y. Katayama, M. Mezouar, J. P. Itiè, J. M. Besson, G. Syfosse, P. Le Fevre, and A. Di Cicco, J. Phys. (Paris) IV, 7, C2-1011 (1997).
[5] U. Buontempo, E. Degiorgi, and P. Postorino, Nuovo Cimento D, 20, (1998).

Publication
U. Buontempo (a), A. Filipponi (a), J. P. Itiè (b, c), D. Martinez-Garcia (b, d) and P. Postorino (e, f), to be published.

(a) Dipartimento di Fisica, Università dell' Aquila (Italy)
(b) LURE, Université Paris- Sud (France)
(c) LPMC, Université Paris VI (France)
(d) ICMUV, Ed. Investigación, Valencia (Spain)
(e) Dipartimento di Fisica, Università di Roma (Italy)
(f) INFM, Roma (Italy).