potentials using a Nion UltraSTEM Jordan A. Hachtel Andrew R. - - PowerPoint PPT Presentation

ORNL is managed by UT-Battelle, LLC for the US Department of Energy

Multidimensional mapping of vibrations, plasmons, and electrostatic potentials using a Nion UltraSTEM

Jordan A. Hachtel

Andrew R. Lupini, Miaofang Chi, Juan Carlos Idrobo, Jingsong Huang, Jacek Jakowski, Ilja Popovs, Santa Jansone-Popova, Shin Hum Cho, Ankit Agrawal, Delia J. Miliron, Tracy C. Lovejoy, Niklas Dellby, Ondrej L. Krivanek

Part of this work was performed at CNMS, a DOE Office of Science User Facility that provides free access if the intent is to publish. (cnms.ornl.gov)

2

1-Dimension

L-Alanine: C3H7NO2

O O C C C N

H H H H H H H

Vibrational Spectroscopy of Organic Molecules

3

Vibrational Spectroscopy in the Electron Microscope

FTIR Spectrum Needs:

• Enhanced resolution
• Reduced background
• Monochromated EELS

FWHM 100 50 20

hBN

4

Energy Resolution Matters

• Old Spectrometer

– 15 meV Res.

• New Spectrometer

– 5 meV

• Vibrational Fine

Structure

• Peaks @ 21 meV

L-Alanine Vibrational EELS

5

Dark Reference and ZLP Movement

Acquire 300 Acquier 300 dark

6

Identifying Isotope-Labelled Amino Acids in Real Space

L-Alanine: C3H7NO2

12C 12C 12C 13C 13C 13C

2-Dimensions

7

Measuring Mass with Electrons

• Isotopic-labeling

influences vibrations

• Isotopic changes

– 160-180 meV Peaks – 200 meV Peak

• Shift of ~200 meV

peak for isotopic identification

8

Nature of 200 meV Peak

200 meV: C-O Stretch

9

Spatial Isotopic Resolution: Sample Preparation

12C L-Alanine Sample 12C L-Alanine Sample (After 13C Loading)

10

Spatial Isotopic Resolution: Damage Free Linescan

11

Spatial Isotopic Resolution

12

Mapping Plasmon Hybridization in Doped Semiconductors

Flourine-doped Indium Oxide

3-Dimensions

13

Infrared Optical Response

• Monochromated EELS detects IR

plasmons

• Geometric dependence observable

Intensity (au)

100 nm

14

Non-negative Matrix Factorization Deconvolution

• NMF for

deconvolution

• Available in

Hyperspy

• Finds true

localizations

• Multiple Peaks in

Side Mode

Intensity (au)

100 nm

100 nm 100 nm

SI.decomposition(algorithm='nmf',output_dimension=2)

15

Effect of Component Number on Deconvolution

• 2 Comp Fit:

Not enough

• Different Corner

Modes

• Different Side

Modes

• Plasmon

Hybridization

• 5 Comp fit:

Too much

16

4-Dimensions

Sub-Ångstrom Electric Field Mapping

17

Detectors for Four-Dimensional STEM

Segmented Detectors Advantages ➢ Fast Acquisition ➢ Direct Signal Readout Pixelated Detectors Advantages ➢ 4D Datasets ➢ Phase Reconstruction Ronchigram Camera Advantages ➢ Convenient and Versatile ➢ Surprisingly Powerful

Shibata et al., Nat. Comm. (2017) Tate et al., Microsc. Microanal. (2016)

18

Ronchi Camera BF Detector ABF Detector HAADF Detector

e-

Ronchigram

2D Image 9 x 9 4D Dataset 9 x 9 x 128 x 128

4D-STEM Imaging on a Nion 2020 Camera

1 ms per Ronchigram Acquisition

19

4D-STEM in a Nion: What do you need?

• 1. Dipole in the Projector Lens Layer
• 2. Dipole in DQCM
• 3. Projector Lens

4D-STEM in a Nion: What do you do?

• 1. Switch to EELS Mode (Make a copy)
• 2. Adjust size of BF disk with Projector Lens(I go for 120 pixels)
• 3. Crop (and bin if you are in a hurry)
• 4. Align BF Disk to HAADF detector (DQCM dipole)
• 5. Align Beam to Camera (Projector Lens dipole)
• 6. Acqurie 4D-STEM

Good News: Easy Bad News: Slow

20

Automatically Finding Ronchigram Center

• 4D-STEM requires accurate

knowledge of Ronchigram center

• Center Finding Routine:

– Sum, Blur, Cutoff

• Sub-pixel accuracy

21

Image Reconstruction from a 4D Dataset

Bright Field Annular Dark Field Annular Bright Field < 30 mrad > 45 mrad 15-30 mrad

22

Differential Phase Contrast on a Universal Detector

Electric Field Surrounding Atom Ronchi Camera

23

Weighting Ronchigrams by the Center of Mass

24

Atomic Potential from the Electric Field

25

5-Dimensions

Not yet, but inevitable

• 1. x-Scan Direction
• 2. y-Scan Direction
• 3. u-Momentum Direction
• 4. v-Momentum Direction
• 5. Defocus
• 6. Time
• 7. Temperature
• 8. Electrical Bias
• 9. Dose
• 10. Accelerating Voltage

11.Orbital Angular Momentum

String Theory in the Electron Microscope?

26

Dreams for the Future:

• Live DPC
• The triple-whammy bottleneck
• Swift writes 4D-dataset to buffer
• Buffer writes to library
• Export from library to external hard drive
• ORCA on SuperScan
• Read DPC like other detector signals
• Tune defocus condition
• Alleviate hard drives, direct experiments
• Interface between iPython Notebooks and Swift
• Swift: Explore Newly Acquired Dataset
• Notebook: Stored Analysis Pathway
• Arbitrary Acquisition From a Reference Image
• Linescans
• Aloof Spectrum Imaging?

27

Measuring and Mapping Octahedral Tilts

θ Polar Angle Azimuthal Angle φ

Predicted: 25.7o / 154.3o Measured: 26.0o / 152.5o

28

: 0.73 Å : 1.38 Å Revealing Light Elements with STEM-DPC

29

Infrared Plasmonics

• Noble Metals

– Normally visible frequencies – Infrared with large dimensions

• Doped Semiconductors

– High free-carrier densities – Mid-Infrared wavelengths – Tunable by:

• Geometry
• Material
• Doping Concentration

30

Tilted Spectrum Image

Distal Face Proximal Face Distal Corner Distal Edge & Face Proximal Corner Proximal Edge & Face

100 nm

• Observe

substrate induced effects

• Measure and

map infrared plasmons

31

Phonons and Plasmons in the Same Dataset

• Plasmons and phonons

share infrared regime

• Examine both in same SI
• Future: Spatially-Resolved

Plasmon-Phonon Interactions

32

Sub-Ångstrom Resolution in DPC Imaging

➢Denoise and Interpolate for Subpixel Analysis ➢Dy/Sc sites identifiable from Z-contrast ➢Dy-pair resolvable in DPC : 0.72 Å

33

ICoM vs. DPC: Field Mapping

1 nm 1 nm 1 nm 1 nm ICoM DPC

34

Atom-by-Atom Imaging of the Electric Field

HAADF Detector