EDS in the TEM

Contents
• Why EDS in the TEM?
• EDS setup in the TEM
• Recapitulation of the important exp. Parameters
• X ray detector and detection in the TEM?
• Artefacts in the TEM
- above sample
- below
- inside (electron channeling: useful or annoying?)
• Quantification
- Basics, Cliff-Lorimer factors
- Corrections
• Special cases
- Light element analysis
- Trace element analysis (minimum mass fraction)
• How to get best spatial resolution
- beam broadening
- electron beam shaping
- contamination and solutions
Why EDS in the TEM
A different point of view than bulk analysis
Example: EBID Rh small particle analysis
1
6 50 Electron
7
42
probe 3
12 10 11
13 14 8 9
16 15 21 27 17 20 18 Total surface of small
23
24 19 22 30
26 28 3129 25 33 particles: 6700nm2
32
388 36 34 35 37
39 42 43 40
41
Total surface 33000nm2
44
45 49 46 4748
51 52 50 5 -> total Rh coverage
54 56 57 5558 59
62 60 61 20%
63 64 65
68 67
69 70
107 7166 But in Nanoprobe
73 74 72 75
7677 79
78 80
81 analysis Rh
85 84 82 86
83 87 concentration 100%
88 89 90
93 9694
99
9195
97 Or Rh grains no pure
98 92 101 100 102
1033 Rh grains?
106 104 105
109 108 110
1111114113 112 115
 Why EDS in the TEM

• A different point of view to bulk analysis

<-> ppm concentrations become 100% locally
 Probe size <-> spatial resolution
Example
Si/CoSi2
interface
Si Si/CoSi2 Si How « sharp » are
interfaces
chemically?
Use chemical
analysis to
complement i.e.
high resolution
methods
 Why EDS in the TEM

• A different point of view to bulk analysis

<-> ppm concentrations become 100% locally
• Probe size <-> spatial resolution
• Easy control of sample temperature,
contamination - > C analysis
• During conventional image analysis chemical
information essential -> complementarity of
EDS and imaging
 Electron Channelling

• Show example of electron channelling

• I.e. electron channeling as a fction of
sample thickness
 Why EDS in the TEM

• A different point of view to bulk analysis

<-> ppm concentrations become 100% locally
• Probe size <-> spatial resolution
• Easy control of sample temperature, contamination
- > C analysis
• During conventional image analysis chemical
information essential -> complementarity of EDS
and imaging
• Electron channelling
 EDS setup in the TEM

• Nanometer probe focussed on thin sample

• Space in TEM limited
• Take off angle fixed in TEM
• Detector close to sample -> artefacts
 Sample illumination

• TEM illumination typical probe size: 10nm

• STEM illumination typical probe size: 0.1-1nm
Recapitulation of important exp. parameters

• Electron intensity high <-> countrate not

too high -> Deadtime 25%
• Protect X ray detector from high countrates
-> shutters, no apertures inserted
• Overvoltag
OvervoltageeUin =TEM
E
>> 1/ E
Electron Ionisation

• K lines fluorescence yield high -> choose K lines

• Take off angle fixed in TEM
Recapitulation of important exp. parameters

• Electron intensity high <-> countrate not too high

-> Deadtime 25%
• Overvoltage in TEM >> 1
• K lines fluorescence yield high -> choose K lines
• Take off angle fixed in TEM
• Pulse processor:
- Sum peaks
- Quantitative analysis -> long sampling
- Chemical mapping -> high countrate->short sampling
Choice of right electron beam current

For 1%
statistical
error count
at least
intensity of
105 counts
i.e. about 1min
Choice of the Dead time

• Low dead time:

not enough cts
• High dead time:
too many counts
suppressed
• High dead time:
sum peaks too
intense
The detector crystal
3mm crystal
TEM permits
detection of X
rays at >30kV
-> Ge crystal
first choice in
TEM
Count rate increase by beam blanking

electron beam
blanked as
soon as X ray
detected in
EDS detector
-> 4 fold
increase of
countrate
 How increase countrate in TEM

• Increase of beam current (spot size)

– > larger electron probe
• Larger of condensor aperture
– > larger electron probe
• Choice of sample thickness
– > multiple scattering modifies probe diameter
– > X ray absorption must be considered
How EDS detection should be
How EDS detection is: Artefacts
Backscattered electrons
- pole pieces
- BSE absorbed by detector
Bremsstrahlungs cone
- fluorescence in the sample
- problem when sample is
tilted
X ray fluoresence
- X rays from above the
column hit sample and pole
pieces
Electrons scattered at high
angle
- electrons hit pole pieces
 Hole count
Beam on sample
X rays
propating
through
Hole count columns
Thin C2 aperture
fluoresce
-> hole count
Hole count signal
Thick C2 aperture
Prevent Hole Count

EDS aperture

1) Use thick C2
aperture
2) Use additional
EDS aperture
above sample
 Post specimen artefacts
• High angle
scattering
-> electrons generate X rays
far away from beam
centre
• Synchroton
radiation
-> X rays fluoresce far
away from beam centre
-> Minimize by
keeping sample
horizontal
-> possible solution: use sample holder, grids surrounding in Be
 C Contamination
Stop contamination!!
- limit organics in sample
preparation
Contamination spot
- O Plasma cleaner
- Ultrasonic bath….
- Clean sample holder
- ……
 Peak to background ratio P/B

Peak to Background ratio measure for analytical

quality of microscope
P/B ratio increases with acceleration voltage
Dedicated AAEM

There are individual

solutions to each
artefact, but only
one solution to
make it right
-> build AAEM
(Zaluzec, Argonne
Nat. Lab)
Spectrum Acquisition

• Peaks well
defined
• Good P/B ratio
for concentrations
above 0.1%
 Quantification
In the TEM:
No absorption correction I ~c
No fluorescence correction

with kAB Cliff Lorimer factor (1975)

 Cliff Lorimer factors
Q: cross section
ω: fluorescence rate
A: molar weight
ε: detector sensitivity

Detector sensitivity

Be window Au contact Si deadlayer Si intrinsic

Experimental and Theoretical K factors

KAFe factors • Rather good

agreement
between
theory and
Theory
experiment
• Discreapancy
Experiment from different
cross section
models
Absorption Correction

Mass absorption
coefficient
Example: absorption correction
 Example: Lineprofile
Ni/Ti multilayers with V-O-N impurities
Ni
Ti

-> N in interface region between

Ni and Ti layers
 Spatial resolution
Initial probe diameter:d= 0.1-1nm
But muliple scattering: Beam broadening b

Total
diameter
Average
diameter

-> Nanoprobe doubles diameter in 50nm Ni sample

Ultralight element analysis I

Detector
sensitivity
for ultralight
elements low
 Ultralight element analysis II

Example O, N, B
1) strong X ray in sample
2) strong X ray in detector
window
3) Weak detector sensitivity
But: Analysis still possible:
Example: Si/Si2 Ti interface
Enhancement of spatial resolution

Either get Angstrom

probe
Or
Deconvolve

Deconvolution may
yield subnanometer
resolution
 Minimum mass fraction I
Trace elements often hidden in unsufficient
experimental conditions

Si 0.2%Fe

-> choose acquisition

time high enough
Minimum mass fraction II
ALCHEMI