Interpreting Images from a Transmission Electron Microscope

1
Interpreting Images from a Transmission Electron
Microscope

• C. Zak Jost
• University of MO St. Louis

2
Background

• The pre-solar grain being imaged came from the
  meteorite Murchison
• Murchison hit the Earth on September 28, 1969 in
  Australia1

3
Oak Ridge National Lab

• In the summer of 2006, Dr. Phil Fraundorf and
  Eric Mandell went to Oak Ridge to take Scanning
  Transmission Electron Microscope (STEM) images
• The resolution of the Oak Ridge STEM is in the
  sub-Angstrom (10-10 m) range, and it may be
  possible to resolve single, heavy atoms

4
Knowledge Gathered

• Using intensities of the gray values on the
  images, data concerning the shape and thickness
  of the specimen can be inferred
• It may be possible to approximate the Z-value of
  single, heavy atoms

5
Why is this Knowledge Desirable?

• By locating the respective positions of heavy
  atoms, one could gather more data concerning how
  the material formed and what the environment it
  came from was like
• If the heavy atoms could be identified and it
  could be concluded that they were not due to
  contamination, this could be a quick and
  non-destructive way to measure abundances

6
Microscope Detectors

• The two experimental images analyzed for this
  project came from the Bright Field (BF) and Dark
  Field (DF) detectors
• The BF gives information about the amount of
  un-scattered electrons
• The DF gives information about scattered
  electrons

7
Microscope Detectors (cont.)

• The electron probe scans rows of the specimen
• At the same point in time, each detector records
  their respective currents into images, resulting
  in two simultaneous experiments

8
Dark Field Image

• Since the brightness of a pixel increases with
  the number of protons encountered on the specimen
  at that point, a graph of intensity should give a
  qualitative idea of relative thickness

9
Dark Field Intensity Plot
10
Simulated DF Images

• To test this hypothesis, I generated an atom
  position list
• I then used a program by Kirkland2 to simulate BF
  and DF images from these atom coordinates

11
Intensity Plots of Simulated Images

• Using MatLab again, I plotted intensity versus
  position as in the Experimental DF Image
• Notice the correlation of thickness and
  intensity, with the peaks representing the
  position of the heavy atoms

12
Comparison of Atom Position and Intensity Plots
13
Profile Plot of Simulated Image

• A more quantitative way to measure relative
  thickness is to plot the profile of the region of
  interest
• Notice the same V shape as seen before

14
Calculating Absolute Thickness

• One possibility for getting approximations of
  absolute thickness is using an equation involving
  the Mean Free Path I Ioe-t / ?
• Since the image contains a region with no
  specimen, the intensity of this region in the BF
  image is related to the incident electrons, Io.
• Using the simulated image, I calculated ? since
  the thickness, t, was known

15
Bright Field Image
16
Applying MFP Equation to BF Image

• Getting a list of intensity versus position, I
  solved the equation for t, and plotted the results

17
Comparison between BF thickness and DF intensity
plots

• Notice the correlation between the intensity plot
  of the DF image, which has shown to be a good
  measure of relative thickness, to the absolute
  thickness plot made from the BF image

18
Z Value Approximations

• Though this is a work in progress, some
  simulation work has been done to show the
  relationship between scattering and Z-value

19
Scattering versus Z

• Though the simulations only used three different
  heavy atoms, the trend line uses the power of
  about 1.7, which has been shown to be a
  reasonable relation from previous efforts
• By getting the total scattering due to a heavy
  atom from subtracting the background, one could
  use this relationship to get approximations of
  relative Z values

20
Z-value and Thickness

• If one could get an accurate relationship between
  Z-value and scattering, the thickness could be
  calculated by solving for the number of particles
  (N) in the following equationIntensity
  NconstZ1.7
• The constant would be determined by simulation or
  other means

21
Future Explorations

• Work is currently being done to use a
  mean-free-path equation to get an independent
  Z-value approximation
• More simulations are being ran that test whether
  the BF aperture size affects the mean free path

22
Summary

• Intensity plots of Dark Field images provide a
  good qualitative understanding of relative
  thickness and heavy atom positions
• A mean-free-path equation applied to the Bright
  Field image gives an approximate value of
  absolute thickness if there is a region in the
  image with no specimen
• It might be possible to infer approximations of
  Z-value of single heavy atoms from the intensity
  in Dark Field images
• The Z-value approximation may independently give
  values to thickness, which can be compared to the
  mean-free-path method