Atom-by-Atom Imaging and Analysis

Ondrej L. Krivanek Nion Co., www.nion.com

in collaboration with

Niklas Dellby, Neil Bacon, George Corbin, Petr Hrncirik, Nathan Kurz, Tracy Lovejoy, Matt Murfitt, Gwyn Skone and Zoltan Szilagyi,

Nion Co., Kirkland, WA (www.nion.com)

and

Phil Batson, Andrew Bleloch, Mick Brown, Matt Chisholm, Christian Colliex, Juan Carlos Idrobo, Vladimir Kolarik, Lena Fitting Kourkoutis, David Muller, Valeria Nicolosi, Steve Pennycook, Tim

Pennycook, Quentin Ramasse, John Silcox, Kazu Suenaga, Wu Zhou, and many others

February 2012

Main topics

• Scanning Transmission Electron Microscopy (STEM): - basic principles - a little history

• Single atom imaging and spectroscopy

• Summary

C and O in BN Si in graphene Si3N4

An electron probe with ~1010 electron per second that’s smaller than an atom is formed and scanned across the sample. Many types of fast electron – single atom interactions can be detected, typically in parallel.

1) Elastic scattering from the

atomic nucleus (Rutherford scattering): high angle ADF

4) Inelastic scattering from the nucleus: high resolution EELS

2) Inelastic scattering from electrons: electron energy loss spectrometer (EELS)

3) e- wavefront reconstruction (holography): 2D camera

3

STEM - an instrument for imaging and analyzing atoms

2

1

4Key primary signals and detectors:

An electron probe with ~1010 electron per second that’s smaller than an atom is formed and scanned across the sample. Many types of fast electron – single atom interactions can be detected, typically in parallel.

3’) Auger electrons arising from de-excitation of inner shell hole: low-energy electron detector

STEM - an instrument for imaging and analyzing atoms

There are many signals, and this why the STEM approach is very powerful.

2’) X-rays arising from de-excitation of inner shell hole: X-ray spectrometer (EDXS, WDS)

Key secondary signals and detectors:

1’) Secondary electrons (SE) arising from various scattering processes: low-energy electron detector

4’) Optical, infrared + UV photons arising from various de-excitiation processes: cathodoluminescence (CL) detector

X-ray detector

CL detector

SE detectorwith energy filtering

1’

2’

4’

3’

2

1

4

Albert Crewe showing single U atoms in a Z- contrast image of stained DNA (1970)

The father of modern STEM: Albert Crewe

Chicago 40 kV STEM

Washington state, USA: 1st EM outside of Europe…

Washington State EM history continued:

1998: Nion Co. started. It makes correctors for VG STEM microscopes.

2007: Nion starts delivering complete STEMs.

… and now the home of a revolutionary new STEM

Members of the Orsay STEM group (Christian Colliex, Odile Spehan, Katia March, Marcel Tence) and Nion’s Niklas Dellby with Orsay’s new 200 kV UltraSTEM

Nion’s first 200 keV, 0.53 Å resolution STEM is shipped to CNRS Paris-Sud (in Orsay).

Nion UltraSTEM™ 200

Described in: Krivanek et al. Ultramicroscopy 108 (2008) 179-195 and Dellby et al. EPJAP 2011. More info at www.nion.com.

Fully modular and thus very flexible.

Operating voltage range 20-200 kV.

UHV at the sample (<10-9 torr; <10-7 Pa).

Ultra-stable, friction-free sample stage

Efficiently coupled EELS

Other major Nion firsts

2000: first commercial aberration corrector in the world delivered

2001: sub-Å electron probe

2007: atomic-resolution EELS elemental mapping

2009: atomic-resolution images of graphene and monolayer BN

2011: EELS fine structure from single light atoms

2012: X-ray spectrum from a single atom

C-K

Si-K

C and O in BN

EELS of one Si atom EDXS of one Si atom

STEM probe size in the aberration-corrected era

Ic = coherent probe current (~0.1-0.5 nA for CFEG)

Graph shows probe size for probe current Ip = 0.25 Ic

uncorrected STEM,Cs = 1 mm

Resolution reached in the Nion 200 keV column (and illustrated in this talk)

dprobe(Cc) ~ (Cc δE) 1/2 / E*o

3/4

dprobe(C7,8) ~ C7,81/8 / E*

o1/2

Area of great current interest, by Matt Chisholm, Juan Carlos Idrobo, David Muller, Quentin Ramase, Kazu Suenaga, Wu Zhou, Jannik Meyer, Ute Kaiser, David Bell and others.

STEM probe size in the aberration-corrected era

Ic = coherent probe current (~0.1-1 nA for CFEG)

Graph shows probe size for probe current Ip = 0.25 Ic

uncorrected STEM,Cs = 1 mm

Resolution reached in the Nion 200 keV column (and illustrated in this talk)

dprobe(Cc) ~ (Cc δE) 1/2 / E*o

3/4

dprobe(C7,8) ~ C7,81/8 / E*

o1/2

Area of great current interest: work by Kazu Suenaga, Jannik Meyer, Ute Kaiser, David Bell and others.

For the full expressions describing the above curves, see Krivanek et al.’s chapter in the just-published Pennycook-Nellist STEM volume (Springer).

HAADF imaging of ß-Si3N4

0.94 Å

Nitrogen columns, separated by only 0.94 Å from Si columns, are clearly visible. Nion UltraSTEM200, 200 kV. Courtesy Tim Pennycok, ORNL.

HAADF imaging of gold particles at 40 and 200 keV

The image was acquired in the so-called “second zone” OL mode, with 2 beam crossovers in the objective lens.

This lowered Cc and gave better than the regular imaging mode.

Image recorded by N. Dellby.Dellby et al, EPJAP (2011), DOI: 10.1051/epjap/2011100429

40 keV: 1.23 Å lattice planes well resolved

(Nion UltraSTEM200, Orsay, France)

HAADF imaging of gold particles at 40 and 200 keV

200 keV: 0.53 Å information transfer that’s independent

of the scan direction(Nion UltraSTEM200)

regular scan scan rotated by 90°

40 keV: 1.23 Å lattice planes well resolved

(Nion UltraSTEM200, Orsay, France)

Single-wall carbon nanotube imaged at 60 keV

Microscope is housed in a soft steel box, shown here with one of its side doors open. The box makes the microscope relatively insensitive to external disturbances. It also serves as a bake-out enclosure.

MADF image of single wall carbon nanotube,Nion UltraSTEM100.

Masking a set of reflections in the FFT allows the front and the back of the nanotube to be visualized separately.

Image courtesy Matt Chisholm, ORNL.

MAADF images of graphene taken 2 minutes apart

Medium angle annular dark field (MAADF) STEM images of a graphene edge, recorded 2 minutes apart. Nion UltraSTEM, 60 keV primary energy.

Configuration changes at the edge are nicely documented, a single heavier adatom (probably Si) is seen.

Recorded in July 2009.

EELS atomic-resolution chemical mapping (2007)

EELS chemical maps of La0.7Sr0.3MnO3/SrTiO3 multilayer structure

40 mr illum. half-angle0.4 nA beam current~1.2 Å probe>80% efficient EELS coupling

64x64x1340 voxel spectrum-image7 msec per pixel, i.e. 29 sec total acquisition time 10 sec additional processing time

i.e., <1 min total time

Nion UltraSTEM100, 100 keV

Muller et al., Science 319, 1073–1076 (2008)

Imaging different chemical species separately

Courtesy Maria Varela and Steve Pennycook, ORNL.

EELS chemical mapping: imaging of oxygen and other sub-lattices due to specific chemical elements in LaMnO3.

Octahedral rotations in the O sub-lattice are clearly seen. Nion Ultra-STEM100, Gatan Enfina EELS, 100 keV.

BN monolayer with impurities imaged by MAADF

Result of DFT calculation overlaid on an experimental image

Cx6

Na adatom

O

N

Longer bonds

C ring is deformed

BC

C

O

Si substituting for C in monolayer graphene

Medium angle annular dark field (MAADF) images.Nion UltraSTEM100, 60 kV. Image courtesy Matt Chisholm, ORNL, sample courtesy Venna Krisnan and Gerd Duscher, U. of Tennessee.

Si at and near topological defects

Si in topologically correct graphene

Si at graphene’sedge

Si

Si

N

Si

Si

Si

Si

Si

Si

2 Å

Si in defective, but less strained graphene is more stable. (15 images added together, no other processing, courtesy Wu Zhou and Juan-Carlos Idrobo)

Si substituting for C: 2 structures are possible

Si in defect-free graphene strains (and buckles) the foil. (courtesy Matt Chisholm)

Si

2 Å

Binding of a single Si atom in a stable defect structure

Si-L edge EELS from single Si atom

C N SiNion UltraSTEM100, 60 keV. Courtesy Juan-Carlos Idrobo and Wu Zhou, ORNL

Exp.: adding together the signal of the pixels corresponding to the Si atom in the graphene spectrum-image

Simultaneous EELS and EDXS from a single Si atom

Nion UltraSTEM100, 60 keV, Daresbury UK. Gatan Enfina EELS, Bruker SDD EDXS. Q. Ramasse, T.C. Lovejoy, O.L. Krivanek et al., to be published.

ADF image of 2-3 graphene layers recorded after spectra were acquired. Arrow points to a tracked impurity atom.

EDXS of single Si atom

C-K Si-K

E (keV)

EELS of single Si atomon graphene

EELS and EDXS data recorded simultaneously.

Ip = 100 pA, 90 s acquisition.

Summary

• The ability to image and analyze matter atom-by-atom was always inherent to the nature of the electron-matter interaction, and it’s now finally available.

• We are able to perform atom-by-atom analysis because we have: ultra-bright electron guns aberration-corrected electron optics ultra-stable electron microscopes ultra-high vacuum at the sample

• The ability to analyze matter atom-by-atom has arrived just in time: atom-by- atom is how we now make the smallest devices.

• Being small and nimble is an advantage when it comes to creating revolutions.

C-K

Si-K

EELS of one Si atom

EDXS of one Si atom

Si in graphene