Electron Beam Analysis
(EPMA, SEM-EDS)
Warren Straszheim, PhD
EPMA, Ames Lab, 227 Wilhelm
SEM-EDS, MARL, 23 Town Engineering
wesaia@iastate.edu 515-294-8187
With acknowledgements to John Donovan of the University of Oregon
Instrumental Techniques
• Excite
• measure characteristic response
• quantify by comparison to standards
Bulk or microanalysis
• Can excitation be focused?
• Can detector be focused?
Electron beam microanalysis
Excitation: focused electron beam
Sample interactions
 secondary electrons
 backscattered electrons
 auger electrons
 cathodoluminescence
 absorbed current
 X-rays
Electron-Sample
Interactions
•Precise x-ray intensities
•High spectral resolution
•Sub-micron spatial resolution
•Matrix/standard independent
•Accurate quantitative chemistry
X-rays
• characteristic emissions
• Be and heavier elements
• background (bremsstrahlung)
X-ray Lines - K, L, M
Ka X-ray is produced due to
removal of K shell electron,
with L shell electron taking
its place. Kb occurs in the
case where K shell electron is
replaced by electron from
the M shell.
La X-ray is produced due to
removal of L shell electron,
replaced by M shell electron.
Ma X-ray is produced due to
removal of M shell electron,
replaced by N shell electron.
Ranges and
interaction
volumes
It is useful to have an
understanding of the
distance traveled by the
beam electrons, or the
depth of X-ray
generation, i.e. specific
ranges. For example: If
you had a 1 um thick
layer of compound AB
atop substrate BC, is
EPMA of AB possible?
Differences between SEM and EPMA
Many shared components
Resulting from intent - imaging vs. analysis
Stability (higher for EPMA)
Current capability (higher for EPMA)
Spatial resolution (higher for SEM) via smaller
spot and limited aberration correction
attached analyzer (EDS vs. WDS)
EDS vs. WDS
• technology – solid state crystal vs. wavelength
spectrometer
• Resolution~126 eV vs 20eV
• P/B ratio
• Detection limit
• count rate limitations
500 kcps in total vs. 70 kcps/element
• parallel vs. serial operation
Spectral Resolution
WDS provides roughly an order of
magnitude higher spectral resolution
(sharper peaks) compared with EDS.
Plotted here are resolutions of the 3
commonly used crystals, with the xaxis being the characteristic energy
of detectable elements.
Note that for elements that are
detectable by two spectrometers
(e.g., Y La by TAP and PET, V Ka by
PET and LIF), one of the two crystals
will have superior resolution (but
lower count rate).
Reed, 1995, Fig 13.11, in Williams,
Goldstein and Newbury (Fiori volume)
Spectrometer Efficiency
The intensity of a WDS spectrometer
is a function of the solid angle
subtended by the crystal, reflection
efficiency, and detector efficiency.
Reed (right) compared empirically
the efficiency of various crystals vs
EDS. However, the curves represent
generation efficiency (recall
overvoltage) and detection
efficiency.
Reed suggests that the WDS spectrometer has ~10% the
collection efficiency relative to the EDS detector.
Reed, 1996, Fig 4.19, p. 63
How to explain the curvature of each crystal’s intensity function? At high Z,
the overvoltage is presumably minimized (assuming Reed is using 15 or 20
keV). Low Z equates larger wavelength, and thus higher sinq, and thus the
crystal is further away from the sample, with a smaller solid angle.
25kV 5um
Effect of voltage
• Excitation volume
goes as V1.7
• Available X-ray lines
5kV 0.4um
10kV 1.3um
15kV 2.5um
Typical steel spectrum, 15 kV
Lines available at low kV
Note overlap of V, Cr, Mn, and Fe. Also, O has its line at 0.53 keV.
Effect of current
spatial resolution reduced with high currents
greater sensitivity with high currents
• detectability
• precision/repeatability
Overlap considerations
• Smaller issue for WDS – effects background
choices
• Deconvolution option for EDS if statistics
permit
• Statistics become problematic if trace element
on major element background
EDS Overlap: S, Mo, Hg
HgS std
Line Type
Wt%
Wt% Sigma
Atomic %
S
K series
13.38
0.14
49.15
Hg
M series
86.62
0.14
50.85
Total
100.00
100.00
Stoichiometry is onthe-mark - in this case.
WDS “overlap”: HgS, PbS, Mo
Note that signal drops to
background in between
most peaks.
Mo tail interferes with S.
Rare earths by EDS and WDS
Tb
Dy
Er
Pr peak fits between
Ce La and Lb peaks.
EDS Atomic fraction
Compound
Fe
Y
D5 Y2Fe17
88.49
11.51
B4 Ce2Fe17
89.05
B5 Pr2Fe17
88.39
C1 Nd2Fe17
88.81
C2 Gd2Fe17
90.12
C5 Tb2Fe17
86.21
D1 Dy2Fe17
88.43
D2 Ho2Fe17
87.59
D3 Er2Fe17
84.69
E2 Lu2Fe17
89.77
Ce
Pr
Nd
Gd
Tb
Dy
Ho
Er
Lu
10.95
11.61
11.19
9.88
13.79
11.57
12.41
15.31
10.23
2/19 = 10.53%
Suitable samples
•
•
•
•
solid/rigid
stable under beam
conductive (while under beam)
nonconductive samples can be coated with C
or metal (e.g., Au, Pt, Ir)
(coating obscures features and elements but
only a little)
Samples include
•
•
•
•
Metals
Geologic samples
Ceramics
Polymers
• Experimental materials
Quantitative Considerations
• Homogeneous (within excitation volume)
• Thick (enclosing interaction volume);
therefore, problems with layered samples
• Known geometry (preferably “flat” compared
to excitation volume; thus, polished);
therefore problems with rough samples
• Be smart with construction (e.g., glass vs. Si)
• Standards collected each time vs.
Standardless and normalization
Matrix effects
Z-A-F or Phi-Rho-Z
corrections
accounting for
penetration depth,
absorption, secondary
fluorescence
Accuracy depends on
well known curvature.
Alternatively, need
standard in region for
better results.
Range of Quantitation
100% down to
0.05% (500 ppm) EDS,
0.001% (10s of ppm) WDS
Limited by statistics,
differentiation from background
More counts help!
Mapping and Line-scans
Point analysis are most sensitive to
concentration differences (30s/point)
Line scans are next (500 ms/pixel)
Mapping is least sensitive (12 ms/pixel)
Graphics convey much information quickly
(i.e., a picture is worth a thousand words)
Digital image showing regions
of analysis and line-scan
Mg portion of overlapped peak
Ge portion of overlapped peak
Line-scan using typical windows
Ge-Mg overlap causes problems
Line-scan using deconvolution
Ge contribution is stripped from Mg profile
Mapping using deconvolution
EDS-WDS comparison
Characteristic
EDS
WDS
Geometric collection efficiency
(solid angle)
<3%
<0.2%
Spectral resolution (FWHM)
<130 eV
2-10 eV
Instantaneous X-ray detection
entire spectrum (0.2 keV thru E0)
single wavelength (a few eV)
Maximum count rate
100s of thousands cps
over entire spectrum
tens of thousands cps
(single wavelength)
Artifacts
sum peaks, Si escape peaks, Si fluor. peak
higher order peaks,
Ar escape peaks
Low-Z limit = Be
With thin window detector
With synthetic "crystals"
Detection Limits
0.05 wt% (500 ppm)
0.001 wt% (10 ppm)
Bottom Line
Cheaper, quicker but some elements are too
close to resolve,
e.g., S-Ka, Mo-La, Pb-Ma
Slower, more expensive, but with
better resolution and higher
peak/bkgd ratios giving lower
detection limits
“Harper’s Index” of EPMA
1 nA of beam electrons = 10-9 coulomb/sec
1 electron’s charge = 1.6x 10-19 coulomb
ergo, 1 nA = 1010 electrons/sec
Probability that an electron will cause an ionization: 1 in 1000 to 1 in 10,000
ergo, 1 nA of electrons in one second will yield 106 ionizations/sec
Probability that ionization will yield characteristic X-ray (not Auger electron):
1 in 10 to 4 in 10.
ergo, our 1 nA of electrons in 1 second will yield 105 x-rays.
Probability of detection: for EDS, solid angle < 0.03 (1 in 30). WDS, <0.001
ergo 3000 x-rays/sec detected by EDS, and 100 by WDS. These are for pure
elements. For EDS, 10 wt% = 300 X-rays; 1 wt% = 30 x-rays; 0.1 wt % = 3 x-ray/sec.
ergo, counting statistics are very important, and we need to get as high count rates
as possible within good operating practices.
From Lehigh Microscopy Summer School
Raw data needs correction
This plot of Fe Ka Xray intensity data
demonstrates why we
must correct for matrix
effects. Here 3 Fe alloys
show distinct variations.
Consider the 3 alloys at
40% Fe. X-ray intensity
of the Fe-Ni alloy is
~5% higher than for the
Fe-Mn, and the Fe-Cr is
~5% lower than the FeMn. Thus, we cannot
use the raw X-ray
intensity to determine
the compositions of the
Fe-Ni and Fe-Cr alloys.
(Note the hyperbolic functionality of the upper and lower curves)
One slide Schrödinger Model of the Atom
n = principal quantum number and indicates the electron shell or orbit (n=1=K, n=2=L, n=3=M, n=4=N) of the
Bohr model. Number of electrons per shell = 2n2
l = orbital quantum number of each shell, or orbital angular momentum, values from 0 to n –1
Electrons have spin denoted by the letter s, angular momentum axis spin, restricted to +/- ½ due to magnetic
coupling between spin and orbital angular momentum, the total angular momentum is described by j = l + s
In a magnetic field the angular momentum takes on specific directions denoted by the quantum number m <= ABS(j)
or m = -l… -2, -1, 0, 1, 2 … +l
Rules for Allowable Combinations of Quantum Numbers:
The three quantum numbers (n, l, and m) that describe an orbital must be integers.
"n" cannot be zero. "n" = 1, 2, 3, 4...
"l" can be any integer between zero and (n-1), e.g. If n = 4, l can be 0, 1, 2, or 3.
"m" can be any integer between -l and +l. e.g. If l = 2, m can be -2, -1, 0, 1, or 2.
"s" is arbitrarily assigned as +1/2 or –1/2, but for any one subshell (n, l, m combination), there can only
be one of each. (1 photon = 1 unit of angular momentum and must be conserved, that is no ½ units,
hence “forbidden transitions)
n
l
s
m
j
number of
Sub shell
X-ray
electrons
No two electrons in an atom can have
the same exact set of quantum
numbers and therefore the same
energy. (Of course if they did, we
couldn’t observably differentiate them
but that’s how the model works.)
notation
1
0
½
0
½
2
1s
K
2
2
2
0
1
1
½
0
2
2s
½
-1, 0, 1
½
½
½
½
6
2p
LI
LII
LIII
3
3
3
0
1
1
½
½
0
-1, 0, 1
2
6
3s
3p
MI
MII
MIII
3
3
2
2
½
-2, -1, 0,
1, 2
½
½
½
½
½
½
½
½
½
10
3d
MIV
MV
Origin of X-ray Lines for K and L Transitions