Structural characterization of natural and engineered ... · Structural characterization of natural and engineered barrier materials: from macro to ... respect to the various potential

jeudi 9 mars 2017

Structural characterization of natural and

engineered barrier materials: from macro to

nanoscopic scales

Gaboreau Stéphane (BRGM, France)

In collaboration with C. tournassat, A. Loschetter, J. Rohmer, P. Audigane, A. Sbai and J.C. Robinet

Microstructure interest

(e.g. smectite)

(e.g. illite)

Interlayerpore volume

External basalsurfaces

Bulk pore solution

Clay edges

Solution volume influencedby basal surface charge

Swelling clay mineral(e.g. montmorillonite)

Bulk solution

Non-swelling clay mineral(e.g. illite)

Macroscale properties of finely divided materials arise for a large part

from the surface properties of their nano-sized minerals or phases

constituents and from the characteristics of associated microstructure

and pore network.

Pore network and microstructure characterization

Which technics for which scale ?

: from macroscopic scale to nanoscopic scale

1 nm 10 nm 100 nm 1 µm 10 µm 100 µm 1 mm

IUPAC

Macroscopic scale Mesoscopic Micro/Nano

In the last decades, large technologic development allowing the characterization of the

microstructure (mineralogy and porosity)

X-ray µtomography

SEM FIB/SEM

TEM/electron tomography

Among many others technics (Autoradiography, PDF, PDF tomography, SAXS ….)

3D

2D and 3D 2D

2D and 3D

Reservoir rock pore network

Macroscopic scale

Carbonate reservoir rock

Large connected pores in the macroscopic domain

Gaz transfer/permeability modelling and poroelastic behavior at pore scale

Laboratory X-ray µTomography

Pore size distribution

‘Maximal ball algorithm’

Voxel size 20 µm3

Oolith

Porous domain

Transport and or mechanical properties

Clays and cement materials have been studied as potential host-rocks or barriers

for wastes disposal facilities.

What are the transfer properties of these formations with

respect to the various potential pollutants (anions,

cations and neutral species) ?

Micrometer-sized / un-porous grains

⇒ No transfer

Mixture of nanometer-sized clay

minerals and cement phases

(negatively charged particles)

3D volume of the COx clay rock extracted form X-ray

synchrotron tomography

Organization of the porous

matrix domain ?

Transfert pathways ?

Quantifying the nano-microstructure of compacted clay materials in order to

better understand its role on pollutants transfer and model gas/water transfer

and mechanical behavior

1 cm

Natural and engineered barrier material

Concrete polished section Argilite SEM image

50 µm

Finelly divided materials with spatial heterogeneity, multiscale distribution

of mineral and pore size

Multi-technics downscaling approach to cover the multi field of view of the

microstructure from the mineralogy to the porosity

1 nm 10 nm 100 nm 1 µm 10 µm 100 µm 1 mm

IUPAC

Macroscopic scale Mesoscopic Micro/Nano

Interlayer space

Inter particules

Inter aggregates

Carbonate agregate Quartz

Pyrite

Carbonates

Bulk data – Reference Total porosity

He pycnometry

MIP intrusion - extrusion

𝜀 = 1 −𝜌𝑑

𝜌𝑔𝑟

𝜌𝑔𝑟 = 𝐴𝑖 𝑤𝑡% × 𝜌𝑔𝑟𝑖

MIP

He

jeudi 9 mars 2017 > 8

densité de

grain (pycno

He)

densité

apparente

sèche

(poromercure)

porosité

mesurée

(poromercure)

densité

apparente

séche kerdane

porosité

totale

kerdane

Porosité

calculée

teneur en

eau W

Degré de

saturation Sr

g/cm3 g/cm3 (%) g/cm3 (%) (%) (%) (%)

Ciment 2,38 1,54 15,96 1,55 38,4 35,0 23,85 96

Béton 2,56 2,26 7,62 2,25 14,2 12,2 5,44 86

Bulk data – Reference Total porosity

Grain density, apparent dry density, total porosity

Information on the size of the throat -> Possibility to detect the pores

with imaging technics

densité de

grain (pycno

He)

densité

apparente

sèche

(poromercure)

porosité

mesurée

(poromercure)

densité

apparente

séche kerdane

porosité

totale

kerdane

Porosité

calculée

teneur en

eau W

Degré de

saturation Sr

g/cm3 g/cm3 (%) g/cm3 (%) (%) (%) (%)

Ciment 2,38 1,54 15,96 1,55 38,4 35,0 23,85 96

Béton 2,56 2,26 7,62 2,25 14,2 12,2 5,44 86

Autoradiography – Spatial distribution of the total porosity

Microstructure

preservation

Impregnated samples with 14C-MMA

Impregnation with a resin

in order to preserve the

microstructure in a water-

like saturation state

All the pore even interlayer

space are impregnated

Autoradiography – Spatial distribution of the total porosity

2D spatial distribution of the Total porosity (from nano

to macropores, with a spatial resolution of 10µm)

2 cm

Non porous agregate Total porosity map (Quantification of the total porosity each 10 µm)

Autoradiography

100%

0%

e

Porosity

scale

Quantification of the

beta emission of the 14C resin

Non porous agregate

Voids filled with MMA

In grey, the hydrated porous domain with a porosity of 40-45 %

Total porosity : 12%

jeudi 9 mars 2017

X-ray µtomography

Macroporosity Unreacted anhydrous grain

Assemblage

Volume 3D (voxel 1 µm3)

1 mm

Pixel size – 1 µm

PCH CEMV

Unreacted anhydrous grain Unreacted anhydrous grain

µtomographie RX – Macroscopic field of view

jeudi 9 mars 2017

Weight % distribution of the mineralogy

Evolution of the hydration of the cement

materials

Size distribution of the phases and reactive

surface estimation

At this scale, 10 % of the porosity is probed

on the 35%

No chemical information!!

Macroporosity

Hydrated phases

Unreacted anhydrous phases

Porous domain

Non porous domain Unreacted anhydrous phases

Unreacted anhydrous phases

Voxel size – 1 µm3

Quantitative chemical map obtained from EPMA. Na, K, Ca, Si, Al, Fe, Mg, S, Cl Maps of 512 x 512 pixels with a resolution 2µm/pixel BSE image associated to the chemical maps 262 144 pixels x 9 elements

N pixels

N p

ixe

ls

In each pixels

Sum of atomic Wt%

Concrete BSE

Concrete Ca X-ray map

Quantitative mineralogy

Si

Ca Al3 – Fe – Mg Ca Al3 – Fe – Mg

Si

Concrete Cement paste

Calcite

Portlandite

Quartz

Silica Fume

Dolomite

Slag

Stratlïngite Hydrogarnet

Solid solution

Katoite Hydrotalcite

AFt – AFm

solid solution

1.6

0.6

zeolite C-(A-)S-H

Solid solution

Successive ternary scatterplots procedure to threshold the mineralogy

Concrete Cement paste

Phases Formula (Ati)wt%

Ca Si O* H2O

(not analyzed) S (oxide) wt%

C2S 2CaOSiO2 46 16 38 100

Portlandite Ca(OH)2 54 0 22 24 76

Quantitative mineralogy – Image segmentation

Quantitative mineralogy – Image segmentation

Quantitative mineral map

512

512

2 µm/pixel

Macroporosity (5 %)

Fe oxides (0.2 %)

Aluminates (0.4 %)

Aluminates (5 %)

Dolomite (1.5%)

Slag (3%)

Calcite (25%)

S bearing (1.6%)

SiO2 (7%)

C-A-S-H (33%)

Mg-rich C-A-S-H (7%)

Al-rich C-A-S-H (6.5%)

C2S (2.5%)

Hydrogrenat (0.5%)

Hydrogrenat (0.4%)

With the associated chemical composition for each

phase/mineral

X Ca wt%; y Al wt%, Z Si wt%, alkalis wt%

Weight % distribution of the

mineralogy

Porosity map – S of atomic weight %

𝜀 = 100

1+𝑚 (𝑤𝑡 %)

100−𝑚 (𝑤𝑡 %)×

𝜌𝑟𝜌𝑚

with 𝑚 𝑤𝑡% = (𝑜𝑥𝑖𝑑𝑒𝑠𝑚𝑒𝑎𝑠 )𝑤𝑡%

(𝑜𝑥𝑖𝑑𝑒𝑠𝑡ℎ𝑒𝑜 )𝑤𝑡%× 100

Intrinsic porosity for

each phases

carbonate

hydrates

Non hydrated phases

100%

50%

0%

Ø 75%

25%

Porosity

scale

Image BSE

Mineralogical map

Porosity map

Intrinsic porosity of hydrated phases ~

45-50 %

In agreement with the bulk data and the

autoradiography with respect to the

proportion of porous and non porous

phases

Downscaling 3D FIB-nt images – mesoscopic scale

Slice and view techniques (Holzer et al., 2004)

Reconstruction of a two-dimensional images into a three-dimensional representation

Successive milling/imaging process

SEM images (low energy)

Downscaling 3D FIB-nt images – image analysis

In order to detect the pores and the throat to threshold a

connected pore network

Isolated pores

Large connected pore network

Downscaling 3D FIB-nt images – image analysis

To provide the Pore Size Distribution

(PSD) and large connected pore network

Support for diffusive model on real pore

network

Classic method Improved method

Pore network Carbonates

Spatial distribution of the porosity from macro to mesoscopic scale

Downscaling 3D FIB-nt images

Argilite

Reactive surface area of some

detritic grains 2 µm

Diffusion simulation

Time domain diffusion algorithm

Particle tracking method (Delay et

al., 2002)

Determination of (i) the water diffusion coefficient (ii) diffusion anisotropy and

geometrical factor

HTO being considered as a non-reactive species, its

diffusion behavior is controlled by its mobility through

the pore network, consequently our modelling results

can be directly compared against HTO diffusion

experiments

Comparison with HTO diffusion experiments

Preliminary test on compacted clay material

Impact of the microstructure on the anion distribution Sensitivity of anion exclusion prediction to

microstructural parameters

Pore Size Distribution

Shape description

Anion accessible

porosity calculation

Downscaling TEM

Dark field HAADF

1 nm 10 nm 100 nm 1 µm 10 µm 100 µm 1 mm

IUPAC

Macroscopic scale Mesoscopic Micro/Nano

Interlayer space

Inter particules

Inter aggregates

We are here, following the next episode

Argilite

TEM foil prepared from fully impregnated samples

100 nm thick.

Argilite

Microstructure and mechanical behavior

Swelling pressure measurement according to the solute

Microstructure evolution

Crystalline pressure

vs.

Osmotic pressure

In situ, time resolved

X-ray µtomography

> Quantify the mineralogy to support the

thermodynamic and reactive transport model

> In the case of cement materials, follow the

hydration reaction to identify the reaction

pathway

> Calculate the reactive surface area

> Support the poro elastic model done at the pore

scale

> Display real pore network from macroscopic to

nanoscopic scale for transfer simulation

> Improve the knowledge in the mechanical

behavior of materials

> ….. Thank you for your attention

Why we are interested in Microstructure