2nd TUG Meeting

1
2nd TUG Meeting

• August 19, 2002

2
Presentations
3
Presentations
4
Agenda

• results of the 1st TUG Meeting
• presentation on XEDS (F. Ernst)
• arrangements for plasma cleaner demonstration
• demand for further action

5
Results of 1st TUG Meeting

• what we achieved
• newsgroup (thanks Leah Lucas)
• simulator (thanks Yeonsop Yu)
• XEDS detector geometry (thanks Yeonsop Yu)
• Gatan Duo Mill (thanks Lichun Zhang)
• demo of plasma cleaner by Fischione tomorrow!
• schedule of presentations
• what we have not (yet) achieved
• decision about purchase of a wire saw
• acquisition of tripods

6
XEDS

• X-ray Energy-Dispersive Spectroscopy

7
Inelastic Scattering
8
Which Elements?

• ratio between X-ray emission and emission of
  Auger electrons fluorescence yield ?
• XEDS is not particularly suited to detect light
  elements
• ? EELS

9
Detector
10
Detector Efficiency
11
Pulse Processing

• ideally high counting rate, without shifting or
  distorting the spectrum
• adjustable parameters
• time constant ?
• dead-time ratio
• time constant
• time available for processing the last pulse
• typically 10 - 50µs
• larger time constant ? better accuracy, but
  smaller counting rate

• dead-time ratio
• percentage of time the detector needs to switch
  off for pulse processing
• closely related to ?
• the more X-ray photons arrive at the detector per
  second, the larger the dead-time ratio
• the dead-time ratio can be expressed as
• Ri input pulse rate Re output pulse rate (rate
  of counts).

12
Dead-Time Ratio

• a high dead-time ratio (larger than 25) means
• the detector receives too many pulses per second
• the detection of X-ray photons tends to be
  inefficient
• reduce the beam current!
• reduce spot size!
• remove objective aperture before inserting the
  detector!

13
Energy Resolution

• definition of the energy resolution, R
• R energy resolution P parameter
  characterizing the quality of he associated
  electronics X effect of leakage current and
  incomplete charge collection I intrinsic line
  width of the detector.
• intrinsic line width of the detector
• F Fano factor (Poisson statistics) ? energy
  to create an electron hole pair in the detector
  E energy of the X-ray photons.
• ? energy resolution can only be specified for
  specific analysis conditions

• standard (IEEE) Mn-K? line, generated by a
  well-defined source (Fe55 with 103 cps and 8 µs
  pulse processor time constant)
• typical energy resolution
• Si detectors  140 eV Ge detectors 114 eV
  (record)
• Tecnai F30 CWRU 129.9 eV (at Mn-K?)

14
Artifacts in XEDS Spectra

• Escape-Peaks
• phantom peaks, caused by fluorescence
• a small fraction of the X-ray photons absorbed by
  the detector is not converted to electron-hole
  pairs
• these photons cause the detector material to
  fluorescence (emit photons that are
  characteristic of the detector material)

• case of Si detectors
• Si-K? photon escapes from the detector
• energy of Si-K? photons 1.74eV
• ? detector does not register energy E, but
  E-1.74 keV
• ? additional peak (escape peak), 1.74 keV below
  the actual peak at E

15
Escape Peaks
16
Sum Peaks

• Sum Peaks
• phenomenon instead of evaluating a single X-ray
  photon, the detector simultaneously evaluates two
  photons
• reason for this problem electronics cannot
  switch-off the detector fast enough
• ? strong peaks may be accompanied by a phantom
  peak at twice the energy

17
Sum Peaks Example
18
XEDSTEM Interface
19
Detector Geometry Tecnai F30

• window
• Super Ultra Thin type
• 0.3 µm thick
• polymer type over Al support grid
• detector
• dead layer thickness 0.085 µm.
• Al thickness 0.04 µm
• area 30mm2

• geometry
• detector distance to the sample 10.72 mm
• corresponds to scale setting of 10.3 mm
• elevation angle 0
• azimuth 0

20
Acceptance Angle

• ? angle between plane of specimen (untilted) and
  line to center of detector
• minimize absorption of X-rays in the specimen ?
  maximize acceptance angle ?
• problem increasing ? requires to decrease ?
• ideal acceptance angle of ?  90 position the
  detector above upper pole piece

• ? ? small, detector needs to view specimen
  through pole piece hole
• if detector below the upper pole piece, ? is
  limited to typically 20
• for ideally thin TEM specimens absorption does
  not cause problems
• thicker specimens
• increase acceptance angle by tilting specimen
  towards detector
• Tecnai F30 13

21
Optimum Geometry

• reduce absorption
• use adequately thin specimens
• appropriately position the specimen with respect
  to the detector
• region of interest should not be on the same side
  of the specimen hole as the detector, but on the
  opposite side

• rotation holders facilitate this setup
• when using standard specimen holders, load
  specimen with the region of interest rotated to
  the appropriate position

22
Optimum Geometry?
23
Optimum Geometry for Interfaces
24
Spurious X-Rays
25
Qualitative XEDS

• optimize the conditions for recording spectra
• identify all element-characteristic peaks
• identify artifacts
• identify all relevant elements in the specimen

26
Peak Significance

• significance of peaks in XEDS spectra
• intensity IA of the peak above the background
• integrate signal and background over the same
  interval of N channels
• the peak is significant at the 99  level of
  confidence if

• if necessary, improve sensitivity
• not by inadequate increase of the intensity (beam
  current, specimen thickness  this reduces the
  energy resolution and favors the formation of sum
  peaks)
• instead extend recording time!

27
Example Increased Recording Time
28
Quantitative XEDS

• quantitative analysis of an alloy AB
• what are the relative parts of A and B?
• distinguish
• mole fractions ?Ax, ?Bx dimension-less
• mass fractions XAx, XBx in wt  (weight
  percent should be mass percent!)

• ? determine intensities IA and IB of the
  corresponding peaks, above the background of the
  spectrum
• assumption absorption and fluorescence may be
  neglected
• Ansatz (CliffLorimer)
• k factor kAB
• conversion from relative peak intensities to
  relative amounts of elements A and B at x

29
Background Bremsstrahlung
30
Background Subtraction

• two windows of equal width, before and one after
  peak(s)
• window width should roughly correspond to peak
  width (FWHM)
• integrate counts over both windows N
• ? average the results to obtain an estimate for
  the background below the peak(s)
• ? this method works well only for slowly varying
  background

31
Example of an XEDS Linescan