Laser heating in the diamond anvil cell: The basics Denis ANDRAULT - - PowerPoint PPT Presentation

Laser heating in the diamond anvil cell: The «basics»

Denis ANDRAULT Laboratoire Magmas et Volcans Université Blaise Pascal First scientific articles based on LH-DAC (WoS):

In canberra, Australia: Lin- Photo: Pierru et al. Gun JIU (1974) discovery of the (Mg,Fe)SiO3 brigmanite, the major mineral on Earth. He then wrote ~10 other major articles based on LH-DAC. In Washington-DC: By: T. YAGI, H.K. MAO, P.M. BELL (1978) Bridgmanite In Hawaï (?), USA: L.C. MING and M.H. MANGHNANA (1979) Phase transition in MgF2 In Paris: A. LACAM, M. MADON, J.P. POIRIER (1980) Upper-Lower mantle discontinuity on Earth

First article using X-ray diffraction coupled with LH-DAC (WoS):

In canberra, Australia: Lin-Gun JIU (1974) discovery of the (Mg,Fe)SiO3 brigmanite.

First articles using synchrotron radiation coupled with LH-DAC (WoS):

In Washington-DC: Y. KUDOH, C.T. PREWITT, L.W. FINGER (1990). Bridgmanite EoS In Mainz, Germany: R. BOEHLER, N. VONBARGEN, A. CHOPELAS (1990). Melting of Fe

First article using ESRF and LH-DAC (WoS):

S.K. Saxena, L.S. Dubrovinsky, P. Lazor et al. (1996) Breakdown of perovskite (MgSiO3) in the Earth's mantle

• G. Fiquet, D. Andrault, A. Dewaele et al. (1998) P-V-T equation of state of MgSiO3 perovskite

D C B A A = Piston B= Cylinder C= Cap D= WC seats Membrane He-loader

Infrared laser Infrared laser Re Gasket Diamond Ca

At the first glance, one could think that LH-DAC techniques did not evolved much in recent years In fact, there is a continual, and critical, evolution in the control of experimental conditions for a large part thanks to the use of synchrotron radiation

librant => Pressure Radiation => Temperature

Chervin-type DAC

What is the limitation in pressure generation when using laser heating ?

Culet of e.g. 250 μm

Hemley et al., 1997

Ideal laser heating up to ~60 GPa Bevel of e.g. 100/300 μm Laser heating is 1 bar 1 bar High P High P still fine! Max 200-300 GPa

No laser heating is possible witho The maximum pressure is ~300 GPa today for temperatures of several 1000 K ut some space (microns) between the hot sample and the cold diamonds

Bevel 10/300 μm

What is the temperature limitation in the LH-DAC ? For measurement at high T, use reflective objectives ! I do not known a limitation today Temperature measurement at ID27 Raw pattern Corrected for system response Two-color temperature

• Int. / Sys-Resp

Intensity (a.u.)

400 600 800 1000 1.8 1.6 1.4 1/wavelength (106 m-1) Wavelength(nm)

6140 K

Wavelength(nm) 1.2 650 700 750 800 850 4000 6000 7000 5000

600 650 700 750

Wavelength (nm) Intensity / system-response (a.u.)(nm)

800 850 900

T = 2900 K T = 3800 K T = 5200 K T = 6250 K T = 1250 K

Use of reflective objectives...

As we need the light intensity as a fonction of any chromatic abheration disables the temperature measurement

600 650 700 750

Wavelength (nm) Intensity / system-response (a.u.)(nm)

800 850 900

T = 2900 K T = 3800 K T = 5200 K T = 6250 K T = 1250 K Spectrometer entrance pinhole Hot sample

How to be sure that the sample properties are measured in the laser hot spot ? Misaligned : 1500 K Aligned : 1500 K 2500 K Spectrometer entrance Alignement of the laser spot Visualisation and alignement of the X-ray beam The use of pico or piezo motors is very convenient for the Warning: The good optical alignement must be checked carrefully at high laser power !

Large sample Small sample Energy deposited Energy

No energy deposited 1 μm away from the sample

deposited No, in most cases, it makes the radial gradient even worse ! Pressure medium thermal conductivity Can we resolve the radial temperature gradient using a very small sample ? Temperature Temperature low k Temperature High k

Cold diamonds

High k FWHM Diamond: ~500K Glass < 1000K Melt > 4000K 20-30 μm FWHM << sample thickness

Take away message: Minimizing the axial temperature gradient is often the major How critical is the radial temperature gradient ?

~ Constant temperature

Pressure medium thermal conductivity

Low k

Two end-members situations arise

axial temperature gradients

Melting of the Earth’s mantle Fiquet et al., 2010

Radial temperature gradient

Laser absorption at the sample surface: (I promised Thermal conductivity of pressure medium Need 2 sides heating

• Oxide heated by CO2-laser

Laser absorption in the bulk of the sample (at =1μm):

• Oxide mixed with Pt-absorber power
• (Mg,Fe)-minerals

Do I need double sided laser heating for my sample ?

=> =>

Sample Sample

What is the sample pressure in the laser spot ?

We were successful to synthesize a HP polymorph based on thermal pressure FInite element model Dewaele et al., 1998

=> It is not possible to measure a PVT EoS without an internal pressure calibrant (e.g. Pt for MgSiO3)

30 28 24 26 22 20 18 16 14 12 1473 1873 2273

Temperature (K )

Mg2SiO4

Olivine

L

Ringwoodite

Exp

PV + PER

Wadsleyite

Pressure (GPa)

2673 3073

Energy deposited

• n sample

Temperature Relaxation of the pressure gradient Theoretical thermal pressure Pth=KT True pressure increase from 0.2 to 0.8 Pth

Any potential problem with chemical migration in the laser spot ? Yes, major problems !

We need fast measurements => ESRF-EBS

Chemical analyses of olivine (Mg,Fe)2SiO4 recovered after laser scan of the laser over the entire sample surface

For this reason, the LH-DAC cannot provide constrains on the Equation of state of (Mg,Fe)-minerals; the interdiffusion is too easy.

Fe Si

Any potential problem of chemical pollution of the sample ?

=> For high-T studies, always load

(Nomura et al., 2014) Reported solidus for various mantle compositions is a wet-solidus...

Infrared absorption of the recovered sample => 1510 ppm water Boehler et al. Fiquet et al. Andrault et al. Nomura et al.

Any potential problem with chemical polution of the sample Yes, major problems !

We need fast measurements => ESRF-EBS

At very high temperature, the diamond anvils to metallic sample. The Fe-C melting curve is much lower than that of pure Fe Only fast measurements can solve the melting curve

• f pure Fe

Anzellini et al., 2013 Boehler et al. (several)

How critical is the size of the X-ray probe ?

Probe size << hot spot => Good configuration: Small radial temperature gradient on sample Probe size >> hot spot The laser can be scanned several minutes on the sample for partial homogeneization

WE NEED ESRF-EBS Probe size

Then, the data set can be deconvoluted: Mossbauer spectrum registered at ID18 Deconvolution of a mixture of Fe2+ (LS & HS) in ferropericlase Fe2+ (LS and IS) and Fe3+ (LS) in bridgmanite

How critical is the size of the X-ray probe ?

Probe size << hot spot + short acquisition time => This allows the mapping of the sample properties

Melting of a basalt at P=120 GPa Melting of pyrolite at P=78 GPa

Electron microprobe Chemical mapping X-ray difgraction X-ray fuorescence Fe-content Mineralogical mapping in situ X-ray difgraction

Position (a.u.) Fe-XRF Intensity

0.5 0.6 0.7 0.8 0.9 1.0 10 20 30 40

1.0 0.9 0.8 0.7 0.6 0.5 0.4 36000 32000 28000 24000 20000 16000 12000 8000 4000 0.3 0.2 0.1 0.0

Ext Pv Liq

Fe Mg-Pv

10 μm

Pv Ext Liq Pv Liq

b a c

Is any type of measurement available in the LH-DAC at ESRF ? will make the LH-DAC system even more suited To date, these types of measurements were performed:

• X-ray absorption (XANES)
• Inelastic scattering (phonons)
• Mössbauer spectroscopy
• X-ray emission spectroscopy
• X-ray raman (?)
• and maybe others ?

For some techniques, some limitations remain in :

• Acquisition time
• Size of the beam
• Absorption of diamond window

CO2 or YAG laser Re gasket Diamond Sample Argon medium Ruby

Thanks for your attention! Any comments on my «questions» ? Any additional question that «we» would comment ?