By Marco Cantoni (EPFL)

1. FIB-Nanotomography
in Materials and Life Science
Marco Cantoni,
Graham Knott, Pierre Burdet
Ecole Polytechnique Fédérale Lausanne
CIME
Centre Interdisciplinaire de Microscopie Electronique
(EPFL-CIME)

2. Director: Prof. Cécile Hébert CIME: Centre Interdisciplinaire de Microscopie Electronique
Basic Sciences central facility for electron microscopy
Physics o 5 TEMs:
Science and Technology of Metals, alloys, ceramics, TECNAIs: Spirit, TF-20, OSIRIS
Engineering Semiconductors, CM300, JEM2200FS
nanoparticles,
Materials Science fullerenes, o 3 SEMs (2 FEI XLF-30,1 Zeiss MERLIN)
alloys, ceramics (+powder), thin films…
o 1 FIB (ZEISS NVision40)
polymers, cement/concret
biomaterials… Chemistry o Yearly ≈240 operators from 60 different labs of 4
Microengineering Catalysts faculties. 13’000-15’000 "beam hours“
micromachining electro-active coatings…
lithography o open to everybody
bio-med. eng. Mainly as a “Do it yourself” we train you... you do
yourself your observations
Life Sciences o For « small » needs, we do the investigation for
Architecture, Civil and you, feasibility studies
Environmental Eng. Conventional TEM (fixation,
staining, high-pressure freezing,
Corrosion freeze substitution…)
Wood Cryo TEM under development
Waste transforming bacteria
Facility Manager:
EM for Phys./Chem./Mat. S. : Marco Cantoni
Graham Knott: BIO-EM (since 2007)

3. Since August 2008: Zeiss NVision 40
e-beam: ZEISS Gemini, 1-30kV, 1nm @ 30kV, 2.5nm @1 kV
Ion-beam: 1-30kV, 4nm @ 30kV
EDS X-MAX (SDD) 80mm2 detector
Kleindiek micromanipulator (TEM prep)
2-3 Ga Sources / year (~5000 beam hours)
FIB Applications @ CIME
• Materials Science:
– TEM Lamellae preparation
– cross-sectioning, SE/BSE imaging, EDX
– 3D reconstruction
– 3D EDX (in collaboration with ZEISS and OXFORD
INSTRUMENTS)
– 3D reconstruction of biocompatible materials
• Life Science:
– Serial Sectioning of cells and brain tissue:
SUPER-STACKS

4. 3D FIB/SEM: volume reconstruction
WYSIWYG: What You (detector) See Is What You Get

5. outline
• low kV imaging in a SEM/FIB, the right
selection of your detector
• Applications in Materials Science, porous
samples
• Life Science, biological samples…?
• Automatic Segmentation
• (3D EDX)

6. 3D FIB/SEM: volume reconstruction
0.5 mm
Nb3Sn multifilament superconducting cable
Nb3Sn superconductor multifilament cable:
14’000 Nb3Sn filaments (diameter ~5um) in Cu matrix
Solid State BSE detector
acceleration voltage:
EDX maps
20kV, 15kV
Sn
Cu
Mechanical polishing <-> Ar ion beam polished
Nb

7. in-chamber ET-detector, SE
in-column “InLens”, SE-detector
Low kV:
acceleration voltage: 1.8 kV
No solid state BSE detector
in-column, “energy-selective” EsB, BSE-detector

8. 3D FIB/SEM: volume reconstruction
Nb3Sn multifilament superconducting cable 0.5 mm
Nb3Sn superconductor multifilament cable:
14’000 Nb3Sn filaments (diameter ~5um) in
Cu matrix
1.8kV EsB detector: Materials & orientation contrast

9. Materials & grain contrast
2048x1536x1700, (10x10x10nm voxel), 28hours

10. What is the spatial resolution of BSE electrons ?
Scatter range in Nb3Sn:
300nm 27nm 27nm
10keV-0keV 1.6keV-0keV 1.6keV-1.4keV
1.6keV
HT 10keV 1.6keV (low loss, EsB grid at 1.4kV)
BSE esc. depth 100nm 10nm 2-3nm
penetration 300nm 20nm (20nm)
Energy selective BS

11. 3D FIB/SEM: volume reconstruction
• Slice thickness (z) = image pixel size (x,y)
Z dimension ~ X or Y, typical: 10nm, possible 5nm (3nm)
• Image dimensions / data size (8-bit grey level tiff):
– 1024 x 786: 800 slices -> 640 Mb
– 2048 x 1572: 1600 slices -> 5 Gb “Leitmotiv”
– 3096 x 2358: 3000 slices -> 21 Gb
Isometric voxel size
x=y=z
• Acquisition time ~1min / slice
(40-60 slices / hour)
-> high S/N ratio, beam current (1-1.5nA), detector efficiency
• Dwell times/pixel 5- 15µsec. (detector signal -> 256 grey levels)
• High throughput: minimise overhead, no tilting, rotating, drift correction
• Z- Resolution: low kV !!!

12. InLens: SE low energy
Pb-free solder SnAgCu:
“one detector is not enough”
M. Maleki, EPFL-LMAF
EsB: Energy selective Backscattered
ETD (SE classic)

13. 10µm
EsB InLens SE
10x10x10nm voxel size, 2048x1536x2000
2 images (2x3Mb) / slice …! (DUAL Channel !)
1.6keV, EsB & InLens-SE detector
12Gb data

15. Phase 1. Dark in EsB image, White in SE-InLens
10x10x10nm voxel size, 2048x1536x2000 pixel/slices
2 images (3Mb) / slice …… 12Gb data

16. Phase 2: White in SE-InLens - Dark in EsB image
10x10x10nm voxel size, 2048x1536x2000 pixel/slices
2 images (3Mb) / slice …… 12Gb data

18. Solid Oxide Fuel Cell cathode
P. Tanasini, LENI

19. The right conditions
1.87kV, EsB detector

20. Image:
2048x1536
10nm pixel size
2200 images
36hours

21. Rendering
of dense
volume

22. Segmentation and
analysis

23. Comparison with Transmission X-ray Microscopy (TXM)
beam stop capillary condenser tomography
rotation axis optically‐coupled
Joy C. Andrews, Yijin Liu, and pin hole CCD at image plane
sample objective ZP
Piero Pianetta
Stanford Synchrotron Radiation
Lightsource
Stanford Linear Accelerator
Center
Yong S. Chu
National Synchrotron Light
Source II
Brookhaven National Laboratory
LC LC‐S LS‐ZP LZP‐CCD
George J. Nelson, William M.
Harris, Jeffrey J. Lombardo, John
R. Izzo, Jr., and Wilson K. S. Chiu*

24. YSZ
LSM FIB data down-sampled to 25nm voxel size
Pore
TXM
FIB

25. George J. Nelson, William M. Harris,
Jeffrey J. Lombardo, John R. Izzo, Jr.,
and Wilson K. S. Chiu*
Department of Mechanical Engineering,
University of Connecticut

26. FIB Nanotomography of biocompatible materials
K. Dittmar, A. Tourvielle, H. Hofmann EPFL-IMX-LTP, M.Cantoni EPFL-CIME
SEM: critical point drying, metal coating
Biocompatibility of implants (ceramic coatings)
Drug delivery from implants
How do cells attach to a surface..?

27. FIB Cross-section of a fixed, epoxy-embedded and stained sample
FIB milling of
“hollow” structure
versus
FIB milling of
massive “homogenous block”
Does this cell like the coating…?

28. Image stack: 1024x786 pixel: (10nm image pixel size)
2kV, 60um Aperture, high current, EsB detector (grid 1.5kV)
600 slices, 20nm thickness, milling current 700pA

29. Rendering of iso-surfaces
Medical steel Ceramic coating: TiO2

30. Rendering
Cell outer membrane and more…

31. Volume: 10x8x8 um, 10x10x10nm voxel

32. Biological samples….
brain tissue, resin embedded

33. Which detector…?
In-chamber SE (Everhard-Thornley)
in-Lens SE
in-Lens BSE (energy selective)

34. TEM , 100kV
thin (50nm) section
Brain tissue: synapse
vesicles (~50nm), mitochondria
SEM (FIB) , 1.4kV
“surface”, (<5nm escape depth)

35. 2048 x 1536 x 1600 Volume: 10 x 8 x 8 um voxel: 5x5x5nm
2 days of fully automated acqusition, 5 ~GB of Data
Milling current 700pA,20sec. milling , 1.2min.imaging / slice

37. Bigger volumes
• Voxel: 7.5x7.5x7.5nm
• Image 3096x2304
• 3300 slices (48hours)
• 23x17x24 um
• 9700um3
• ~7000 synapses
• 23Gb data

38. Automated segmentation of neuronal structures
Ilastik v0.5 - Fred Hamprecht, University of Heidelberg

39. Automated segmentation of neuronal structures
Ilastik v0.5 - Fred Hamprecht, University of Heidelberg
Synapse recognition - Anna Kreshuk

40. FIB Nanotomography
in life science
• Specimen preparation
(fixation, staining,
dehydration, resin
infiltration same as
for BIO-TEM)
• Image contrast and
resolution TEM quality
• Stable and reliable
automated acquisition
(less artifacts than
ultra-microtomy)

41. FIB Nanotomography
in life science
• Specimen preparation
(fixation, staining,
dehydration, resin
infiltration same as
for BIO-TEM)
• Image contrast and
resolution TEM quality
• Stable and reliable
automated acquisition
(less artifacts than
ultra-microtomy)

42. FIB-NT compared with other 3D-techniques
• isotropic voxel size ~5-10nm
• Dwell time ~5-10µsec.
• 1 slice, image / min.
• HT: 1-2kV
• Escape depth of signal (BSE) ≤ 5nm
10x10x10 nm voxel, ZnO film
8x8x8 nm voxel, malaria parasite
New possibilities in
3D-microscopy:
Combination with quantitative
analytical SEM techniques:
EBSD, EDX

43. The “SDD age”
New detectors speed up the acquisition !
dreaming of 1M counts/sec.
50-100k counts/sec. are more realistic at the moment

44. 2008 (“SDD age”), FIB @ CIME, use the full potential of the machine
o Stack of 269 EDX maps
3D-EDX, Pierre Burdet: Ph.D. Thesis
o High tension : 5kV
Goal: FIB Nano-Tomography based on EDX-elemental maps o Voxel size : 20 x 20 x 40 nm
new generation of EDX detectors (SDD) o Pixel per map : 256 x 192 (x 269)
Develop algorithms do “deconvolute” the interaction volume of o Time per slice : 4+1 minutes
characteristic X-ray o Time of acquisition : 24 hours
Ion beam
Sample: Al/Zn, Jonathan Friedli, STI-LSMX

45. evaluation of delocalisation: Model system
Intensity
800
90 %
600
Al K
400 50 %
Zn L
200
100 nm
10 %
Zn Al position nm
200 100 0 100 200
– Simulated linescan across the interface normal to y
• Signal is shifted towards Al because of the incident angle
• Positions of threshold (10 %, 50 % and 90 %) are used to compare with other geometries

46. Jonas Vannod, EPFL-CIME /LSMX
– Potential
• NiTi – stainless steel welding
– Biomedical application N. L. Abramycheva, V. Mosko, Univ.
Ser. 2: Khimiya 40 (1999) 139-143
• Complex microstructure
– Intermetallic phases
• Fracture location
– In weld close to NiTi
Laser
NiTi
SS
300µm NiTi SS
100 m Welding process
NiTi ? SS
Longitudinal cut through welded wires

47. SE image with high Fe phases
• Segmentation based on ternary x
diagram
• Green 4: Between Ni3Ti and Fe2Ti
• Red 5: Fe2Ti y
• Blue 6: -(FeNi)
Ternary diagram 4
5
6
z
z
2 m

48. x
• Small microstructure
– EDX phases used as mask
– Threshold on SE contrast
y
3
Ternary diagram 2
6b
6a
z
z
2 m

49. x
2 1
Ternary diagram 3
5
4
2 1
4
y
z
5 6
3
6 2 m
Phases visualization

50. Thank you for your attention