8/3/2019 G. Fiquet et al- Laser Heating Under Pressure: A Brilliant Journey to the Centre of the Earth

1/3

HIGH PRES SURE WORKS HOP

ESRF - APRIL 1998

26

LASER HEATING UNDER

PRESSURE: A BRILLIANT

JOURNEY TOWARDS THE

CENTRE OF THE EARTH

G. FIQUET1

, D. ANDRAULT2

,A. DEWAELE1, T. CHARPIN2,F. VISOCEKAS2, D. HUSERMANN3,

M. KUNZ4 AND T. LEBIHAN3

1 LABORATOIRE DE SCIENCES DE LA TERRE,

ECOLE NORMALE SUPRIEURE DE LYON

2 LABORATOIRE DE MINRALOGIE-CRISTALLOGRAPHIE,

INSTITUT DE PHYSIQUE DU GLOBE, PARIS VI

3 ESRF, EXPERIMENTS DIVISION

4 LABORATORY OF CRYSTALLOGRAPHY, ETH

ZENTRUM, ZRICH

INTRODUCTION

The study of the deep Earth has beenmotivating generations of scientists whohave to take up the challenge given bythe extreme conditions existing at thecentre of the Earth: over 300 GPa(3 million bar) and 5000 Kelvin. As thestudy of the propagation of elasticwaves created by earthquakes onlyallow the determination of the densityprofile of our planet, there remainsthe all important problem of thedetermination of the chemicalcomposition and crystalline structuresexisting in the deep Earth, as thesegovern the Earth global exchanges(thermal regimes, convection drifts,plate tectonics...). Thus relationships haveto be established between chemicalcomposition, crystalline structure andspecific volume over the whole range ofpressures and temperatures existingwithin our planet, and x-ray diffractionis by far the best technique to obtainreliable structural and molar volume

data on the compounds and materials ofinterest. Consequently, as very highpressures can only be generated inextremely small samples, the highbrilliance of the ESRF, coupled withadvances in detector, optics and highpressure technologies, has stimulatedthe rapid development of techniquesfor collecting structural data ongeophysical samples in extremeconditions of P and T. We report heresome recent results from a projectconcerned with in-situ x-ray diffraction

studies of laser-heated samples underpressure [1], and the examples selectedare studies devoted to the two majorconstituents of the deep Earth: iron andMgSiO3 perovskite.

EXPERIMENTAL

DEVELOPMENTS

Before this project, structural data onmaterials in laser-heated diamond-anvilcells (DACs) were obtained using white-beam energy-dispersive diffraction, a

technique which suffers from an intrinsiclow resolution and poor crystallitestatistics (hence unreliable intensity data)due to the small window of diffractionspace sampled, but thanks to the highbrilliance of the ESRF, it is now possibleto combine monochromatic angle-dispersive diffraction and image-platedetectors to collect quality data up topressures and temperatures in excess of90 GPa and 3000 K. This is achieved byfocusing the high brilliance beamproduced by two phased 40 mm period

undulators [2] using single-electrodebimorph mirrors [3] on the HighPressure beamline (ID30). The resultingfocal spot of about 8 m x 15 m(FWHM) is compatible with the size ofthe laser-heated hot spot, and thewavelength of the monochromatic beam,selected by a water-cooled channel-cutSi(111) monochromator in the 0.4 to0.5 range, is well matched to theaperture of custom-built DACs.Combining these beam characteristicswith an experimental set-up especially

designed for the project, and consistingof TEM00 CO2 and multimode YAGlasers, optical set-ups for on-line P,Tmeasurements and alignment of the

sample and beam, large aperture DACsallowing in-situ P,T measurements andfull 4 data collection, for the first time ithas been possible to collect angle-dispersive diffraction data on image-plates during the laser-heating ofsamples only a few micron thick.

A NEW-PHASE OF IRON

Iron being the dominant constituentof the Earths core, information on itsbehavior at high P and T is fundamentalin Earth sciences, but despite numerousstudies on this subject [4] there is stillmuch uncertainty about its structure inthe P,T conditions relevant to the core.The accurate determination of the phasediagram of this element is indeed anexperimental challenge because of the

extreme conditions involved, and evenbelow 100 GPa, recent x-ray diffractionexperiments have led to conflictingresults on the structure of its -phase.Thus that region of the phase diagram,which was previously regarded assimple, is in fact complicated and clearlyin need of new experimental data.Indeed, as mentioned by Anderson [4],the final choice between the and phases for the core depends on theoutcome of future studies aiming atproving the existence of the -phase and

identifying its crystallographic structure.The iron phase diagram up to 100

GPa and 2700 K has thus been studiedusing the best diffraction technique

6 8 10 12 14 16

2-[]

Intensities(a.u.)

18 20 22 24

Iron + Corundum, P ~ 45 GPa, T ~ 2125 K

Fig. 1: Full structure refinement of a spectrum recorded at 2125 K, 44.6 GPa.

The cell is orthorhombic with space group Pbcm and iron location {0.239, 0.472,

0.25}. Observed, calculated and difference spectra are shown, with lower ticks for

iron and upper ticks for Al2O3.

8/3/2019 G. Fiquet et al- Laser Heating Under Pressure: A Brilliant Journey to the Centre of the Earth

2/3

HIGH PRES SURE WORKSHOP

ESRF - APRIL 1998

27

currently available, and these resultsintroduce strong constraints on itsstructures at high P and T. Havingsignificantly improved both the resolutionand the reliability of the data, it has beenpossible for the first time to perform fullstructural Rietveld-refinement in these

extreme conditions of P and T (Figure 1).The space group determined is Pbcm, andthe atomic topology is close to thatof the -hcp phase. The structure isalso closely related to the lower pressure,high-T polymorph, -fcc. The high-Tpolymorph appears unquenchable atmoderate pressures, but the spectra ofthe back-transformed -phase showsome anomalies which can explainambiguities reported in previous structuredeterminations [5].

To eliminate possible artifacts

introduced by the pressure transmittingmedium (corundum, which is also used asthermal insulator), the occurrence of theorthorhombic lattice in a SiO2 mediumwas also checked. This is illustrated inFigure 2 which shows results obtained upto about 100 GPa. Unfortunalely at thispressure the sample was not sufficientlyinsulated from the diamonds, and it wastoo difficult to obtain in-situ spectraduring stable laser heating. However,spectra quenched from about 2500 Kclearly show the doubling of the 100 and101 lines of the hcp-lattice, here again anevidence of a transformation at high P andT. All the experimental lines are explainedby an orthorhombic lattice similar to thatpreviously observed, and in contrast withresults obtained at moderate pressures, thestructure of the high-T polymorph is nowpreserved after the quench. At thispressure, the orthorhombic lattice is foundto be about 1% denser than -iron.

THE STRUCTURE OF MGSIO3

PEROVSKITE

The perovskite form of (Mg,Fe)SiO3being currently accepted as the dominantphase of the Earths lower mantle (700 to2900 km deep), its equation of state(EOS) plays an important role in manyfields of geophysics. It is howeverpresently impossible to choose betweenthe perovskite-pure and perovskite-magnesiowstite (Mg, Fe)O models forthe Earths lower mantle on the basis ofthe existing data, and in-situ high P and T

diffraction is certainly the only methodavailable to measure correctly its EOSand solve the stuctural problem. Previousstudies were conducted in the stabilityfield of the perovskite, but energy-

dispersive diffraction and large-volumepresses limited the performance and P,Tranges to 30 GPa and 2000 Krespectively. Using the techniquedescribed earlier, our measurements onMgSiO3 perovskite were extended to86 GPa and 2700 K. Here however thenew on-line image-plate detector (theFastscan [6]) now available on the ID30beamline was used for the data collection.

Silicate perovskite MgSiO3 sampleswere synthesized from syntheticMgSiO3 enstatite crystals or synthetic

MgSiO3 glass mixed with platinumpowder, and once loaded in a largeaperture DAC, the starting materialswere transformed at high P using eitherthe CO2 or the YAG infrared lasers,

depending of the pressure transmittingmedium. The temperature wasdetermined by analyzing the thermalemission spectra recorded during thediffraction measurements and thepressure conditions were calculatedfrom the EOS of platinum, used here asinternal pressure calibrant [7].

Le Bail profile refinements wereperformed on the diffraction patterns toobtain reliable high P, high T cellparameters for the sample and thepressure calibrant up to 86 GPa and 2700

K, and the most remarkable result wasthat Rietveld structural refinements weresuccessfully carried out on selectedpatterns in these extreme conditions [8](Figure 3). This gave for the first time

1.2 1.4 1.6 1.8 2 2.2

Distance ()

Intensity(a.u.)

110102

101

002

100

Sti

Sti

StiSti

200/122

112/022

111

021

002

110

020

Sti

Sti

StiSti

hcp-iron

Ortho-iron

P = 100 GPaFig. 2:

Diffraction spectra

of hcp and

orthorhombic iron

recorded at 100 GPa

in a SiO2 medium

(labeled Sti).

The 101 reflection of-iron is truncated

for better clarity. The

top spectrum,

quenched from about

2500 K, clearly

shows the doubling

of the 100 and 101

-iron lines, evidenceof the phase

transition toward the

orthorhombic phase.

6 8 10 12 14 16

2-[]

Counts

18 20

Fig. 3: Full Rietveld structure refinement of a diffraction spectrum of MgSiO3 at

86 GPa and 2310 K integrated from an image plate exposed for 10 minutes using

a monochromatic beam focused to 10 m x 20 m. Sample and platimum (the

pressure calibrant) reflections are shown by lower and upper ticks respectively.

8/3/2019 G. Fiquet et al- Laser Heating Under Pressure: A Brilliant Journey to the Centre of the Earth

3/3

HIGH PRES SURE WORKS HOP

ESRF - APRIL 1998

28

precious structural information on thesecompounds, as for instance the firstobservation of the increase of the internaldistortion of the SiO6 octahedra withincreasing pressure in a powder sample.Furthermore, the data analysis allowed usto identify a set of thermoelastic

parameters to constrain the compositionalmodel of the Earths lower mantle.Assuming that the thermoelasticparameters obtained from this study areapplicable to perovskites with moderateiron content, then the comparison of thedensity and KT profiles calculatedfor a mixture of perovskite andmagnesiowstite with those obtainedfrom the PREM [9] model indicates thata pure perovskite lower mantle is veryunlikely. On the other hand, a very goodmatch between the PREM density

and KT profiles is obtained for amixture of 83 vol% (Mg0.93, Fe0.07)SiO3perovskite and 17 vol% (Mg0.79Fe0.21)Omagnesiowstite [8].

REFERENCES

[1] G. Fiquet, D. Andrault, J.P. Iti, Ph. Gillet,

and P. Richet (1996) High-pressure and high-temperature x-ray diffraction study of periclase

in a laser-heated diamond anvil cell. Phys.

Earth Planet. Int., 95, 1-17.

[2] J. Chavanne, P. Elleaume and P. Van

Vaerenbergh, Phasing multi-segment undulators,

ESRF Newsletter No 25, 1996, 12-14.

[3] J. Susini, Active/Adaptive optics at the

ESRF, ESRF Newsletter No 21,1994, 20-23.

[4] O.L. Anderson (1995) Mineral physics of

iron and the core. Reviews of Geophysics,

Supplement: 429-441. U.S. National report to

international union of geodesy and geophysics

1991-1994.

[5] D. Andrault, G. Fiquet, M. Kunz,

F. Visocekas and D. Husermann (1997) The

orthorhombic structure of iron: an in-situ study

at high-T and high-P, Science, Vol. 278, pp. 831.

[6] A new image plate detector system for

diffraction experiments at high pressure, ESRF

Highlights 1997-1998, 93-94.

[7]D. Andrault, G. Fiquet, J.P Iti, P. Richet,

P Gillet, D. Hasermann and M. Hanfland.

(1997) Pressure in the laser-heated diamond

anvil cell: implication for the study of the

Earths interior. Eur. J. Mineral., in press.

[8] G Fiquet, D. Andrault, A. Dewaele,

T. Charpin, M. Kunz and D. Hasermann.

(1997) P-V-T equation of state of MgSiO3perovskite. Phys. Earth Planet. Int., in press.

[9] A.M. Dziewonski and D.L. Anderson

(1981), Preliminary reference Earth model,

Phys. Earth Planet. Int., 25: 297-356.

DENSITY

MEASUREMENTS OF

LIQUID IRON ALLOYS AT

HIGH PRESSURES:

TOWARDS A BETTER

UNDERSTANDING OFTHE PLANETS

C. SANLOUP1, I. MARTINEZ2, F. GUYOT3,P. GILLET1, G. FIQUET1, M. MEZOUAR4,

D. HUSERMANN4 AND T. LEBIHAN4

1 LABORATOIRE DE SCIENCES DE LA TERRE,

ECOLE NORMALE SUPRIEURE DE LYON

2 LABORATOIRE ISOTOPES STABLES, INSTITUT DE

PHYSIQUE DU GLOBE, PARIS

3 LABORATOIRE DE MINRALOGIE-CRISTALLOGRAPHIE,

INSTITUT DE PHYSIQUE DU GLOBE, PARIS

4 ESRF HIGH PRESSURE GROUP, GRENOBLE

Physical properties of iron-basedliquids are of much interest to betterunderstand both the current state ofplanetary cores and their formationduring the differentiation of planets.Here we present the first experimentsperformed on metallic liquids in the Fe-Ni-S system, which might be relevant atleast to the terrestrial outer-core and themartian core. Using a large-volumepress apparatus (a Paris-Edinburgh

press), the P-T range of 0-4 GPa and20-1250 C was explored by measuringthe absorption profiles, hence density,of samples using high-energy x-rays.Equations of state of liquid ironalloys are therefore on the way to bedetermined, along with accuratemelting-phase diagrams as a function ofpressure and temperature relevant togeophysical conditions.

GEOPHYSICAL INTERESTS

Density measurements of Fe-basedliquids at pressure and temperaturerelevant to planetary cores are essential tomodel accurately the core compositionand convection. This should help resolvetwo important geophysical issues: thegeneration of the Earth's magnetic fieldand the thermal history of the planet.

Also relevant to these measurementsis the differentiation of planets, i.e. at firstorder, the individualization of a metalliccore towards the center of the planet. Allthese phenomena refer to the liquid state

of core materials, which concerns atleast the outer terrestrial core, but alsoprobably Mars, Venus and some Galileansatellites such as Ganymede for example,as a substantial magnetic field (roughly a

ACKNOWLEDGEMENTS

We acknowledge the use ofexperimental facilities provided by

the Extreme Conditions Consortium(ECC, a collaboration between theUniversities of Paris VII, Lausanne,Uppsala, the Ecole NormaleSuprieure de Lyon and the ESRFHigh Pressure Group). We aregrateful for the support and efforts ofS. Bauchau and M. Hanfland from theESRF, J.P. Iti from Paris VI / LURE,F. Guyot and P. Richet from Paris VII,and P. Gillet from the ENS Lyon. TheFastscan detector project is acollaboration between the ESRF and

the University of Erlangen, Germany.We thank M. Thoms for all his workand A. Winnaker for supporting thisproject.