Energy Dispersive X-ray

Spectroscopy

Materials Characterisation Techniques-Online

1 Course
 Interaction of electrons with a
specimen
Elastic interactions
Incident electrons with energy E0 pass through a sample or
are scattered without energy transfer.

Inelastic interaction
Transfer of energy from the
electron to the matter, causing
various effects, e.g., ionization.

Inelastic interaction
Secondary electrons
Auger electrons
Characteristic x-rays
Bremsstrahlung

Elastic interactions
Back scattered electrons
Transmitted electrons
Diffracted electrons
Materials Characterisation Techniques-Online
2 Course
 Electron interaction with matter
Back scattered electrons SEM
Secondary electrons

Auger electrons AES

Transmitted electrons TEM,

EELS
Diffracted electrons

Characteristic x-rays EDS

Bremsstrahlung

Materials Characterisation Techniques-Online

3 Course
 Energy Dispersive X-ray Spectroscopy
(EDS)

Materials Characterisation Techniques-Online

4 Course
 Plan of the talk
 Introduction
 How characteristic lines and Bremsstrahlung arises?.
 Features in an EDS profile

 Instrumentation
 Components of an energy dispersive x-ray spectrometer
 Typical artifacts in an EDS spectrum

 Applications
 Quantitative analysis
 EDS Line profile
 EDS elemental mapping

Conclusions

Materials Characterisation Techniques-Online

5 Course
 Characteristic lines
An inelastic collision between a K shell
electron with energy ≥ Ekab (absorption
energy or binding energy) can result
in ejection of the core electron.

The resulting vacancy in K shell can be

filled by an electron from one of the
higher shells L, M etc.

The excess energy of the electron is

liberated as x-ray

Auger electron emission –

competitive
Materials Characterisation Techniques-Online
6 Course
 M

L
Kα Kβ
K

N
M
Lα Lβ

Selection rules
Δl = ±1

7
Materials Characterisation Techniques-Online Δj = 0, ±1
Course
 The electron beam incident on the specimen should have
sufficient energy to exceed the excitation potential
(binding energy of a core electron)

The energy range of 10

keV is useful for
getting K line of
elements with z from 11
to 32.
L lines from 30 to 80
M lines from z=62
upwards

Materials Characterisation Techniques-Online

8 Course
  The optimum for the ionization of the regarded shell (thus
the characteristic x-ray) is
Uo = Eo / Ec = 3

 Ec – Ionization energy of given element A in the K-, L-,

or M-shell (keV)
 E0 - Accelerating voltage of the incident electrons

 The excitation should be so large that for all elements in

the sample a over voltage Uo of at least 2 exists.

Plot of the ionization cross-section over U

Materials Characterisation Techniques-Online
9 Course
 The production of characteristic x-
rays by electron bombardment
was first observed in 1909 by
Charles G. Barkla and C A Halder.
Nobel prize in 1917

A reproduction of a 1910 photograph showing

Moseley in the Balliol-Trinity College Laboratories
Materials Characterisation Techniques-Online
10 Course
  The energy of a x-ray line (energetic position
of the line in the spectrum) indicates, which
element it concerns (Mosely's law).

 The ' strength ' of the line depends on the

concentration of the element within the
sample.

Materials Characterisation Techniques-Online

11 Course
 Bremsstrahlung (retarding radiation)

 An electron which undergoes a

sudden deflection in path
radiates energy

 Thus a fast electron which is

deflected by an atomic nucleus
will radiate a certain amount of
energy as X-rays.

 The upper limit of

bremsstrahlung is Ve, where V
is the voltage of the electron
gun and e is the electronic
charge.
Materials Characterisation Techniques-Online
12 Course
 Real spectrum The measured spectra in
energy dispersive
spectrum consists thus of
two basic components, the
"peaks" of the
characteristic radiation
and the Bremsstrahlung
background

Materials Characterisation Techniques-Online

13 Course
 X-ray spatial resolution
X-ray spatial resolution depends on the interaction volume

 The middle penetration depth of the

electrons Zm and with this the lateral
resolution of the EDS can be
calculated from the middle loss of
energy of the electrons within the
sample:
Zm = 0.033 ( Eo2 - Ec2) /
dens (Eo und Ec in keV; dens
in g/cm )
3

Eo ...primary electron energy

Ec ...energy of the excited electron
shell
dens ... density of the sample
Example:
Detection of Fe (iron, Ec = 7.12 keV) in
a sample with dens = 8 g/cm3:
Materials Characterisation Techniques-Online
14 Course Eo 15 keV 20 keV 30 keV
Zm 0.7 um 1.5 um 3.5 um
  Interaction volume depends on the accelerating
voltage

green: Pb-L / red: Pb-M

Materials Characterisation Techniques-Online

15 Course
 Interaction volume depends Average atomic number
of the specimen

Comparison of electron paths (top) and sites of X-ray excitation

(bottom) in targets of aluminum, copper, and gold at 20 keV,
simulated in a Monte Carlo procedure
Materials Characterisation Techniques-Online
16 Course
 Shape of the excited volume
depends on..
 Average atomic number of the specimen
 Instrument operating voltage
 Angle of incidence of the electron beam

Materials Characterisation Techniques-Online

17 Course
 Instrumentation
Materials Characterisation Techniques-Online
18 Course
 Essentials..
Window
Crystal
FET Collimator
assembly

Materials Characterisation Techniques-Online

19 Course
 Window

 The window provides a barrier to maintain vacuum within the

detector whilst being as transparent as possible to low energy x-
rays
 Be is usually used a window material (~7.5 micron thick)
 Although this window is transparent to most of the X- radiations,
its transmission falls rapidly around 1keV.
 Transmission of 1.04 keV Na Kα line is 60%.
 That of 0.52 keV O Kα is 2%.
 Hence normal EDS system is useful for elements above Na
(Z=11).
 It is in fact, the detectors blindness to carbon X- rays which
makes it possible to employ that material for charge dispersing
coating on an insulating SEM specimen.

Materials Characterisation Techniques-Online

20 Course
 Windowless or thin window detector

 If the microscope vacuum is sufficiently ‘clean’ i.e. free from condensable

vapors such as water and pumping fluids, detectors can be used without
the protection of a window.

 Certain polymers and sub micrometer thick boron compounds are less
absorbing and they are termed as thin window detectors.

 They must be vacuum tight and strong enough to withstand 1 bar pressure
over 10-30 mm2 sensitive area when the chamber is opened for specimen
change.

Materials Characterisation Techniques-Online

21 Course
  Detectors incorporating windows, such as
the atmosphere (supporting) thin window
Electron traps (ATW) and super ATW (capable of
transmitting X-rays below 1 keV), are
generally fitted with a device called an
electron trap. The purpose of the device
is to prevent electrons entering the
crystal and contributing to the spectrum
background or striking window materials
and generating spurious X-rays. The
traditional Be window, which used to be
the most common entrance window,
efficiently absorbs electrons below about
20 keV, and so with this type of window,
an electron trap is rarely required.
 The trap contains two small magnets
which are mounted in front of the window
but behind the collimator. These produce
a strong magnetic field which deflects
the path of incident electrons to the
sides of the trap. In this way the
22 electrons are prevented from entering
Materials Characterisation Techniques-Online
Course
the crystal.
 Detector
 Uses a semiconductor detector for detecting x-rays.

 Usually the detector will be a single crystal disc of

drifted Silicon Si(Li) or high purity Germanium (HPGe).

 Thickness~3-5 mm.
 Area~ 10-30 mm2.

 The detector converts energy of the incident x-ray

photons into pulses of current proportional to the photon
energy.

Materials Characterisation Techniques-Online

23 Course
 The device is constructed
from silicon, a material which
normally contains p-type
impurities. To compensate
for these, lithium atoms were
drifted into the lattice under
an applied electric field and a
reverse biased p-i-n diode is
created with its central
active region consisting of
high resistance silicon fully
compensated with lithium.
When an x-ray deposits its
energy in the active region
e-h pairs are created and the
electrons liberated travels to
the rear electrode.
Materials Characterisation Techniques-Online
24 Course
  No. of e-h pairs produced, n=E/=E/3.65 eV
E is the energy of the incident x-ray
 is the energy required to create an e-h pair.

 Total charge collected at the positive contact is Q=ne

n is the no. of e-h pairs and ‘e’ is the charge of an electron

 Since Eg is small (~1.1 eV) at RT thermally generated charge carriers cross this barrier
easily and become conductive even with no x-ray striking the detector. This causes a
high noise level.

 The detector is cooled by liquid nitrogen to about 100 K to prevent lithium diffusion
under the applied voltage and to minimize noise.

 For effective use the detector should subtend a large solid angle at the specimen. The
largest angle currently in use is 0.3 sterdian, which means that about 4% of X-rays are
detected.

Materials Characterisation Techniques-Online

25 Course
 FET AMPLIFIER
 FET is positioned just behind the detector.
 FET converts the charge produced in the
crystal to a voltage whose amplitude is
proportional to energy of the x-rays.

Materials Characterisation Techniques-Online

26 Course
 MULTI CHANNEL PULSE HEIGHT
PROCESSER
 The pulses are amplified and fed to a multi channel analyzer
which stores them in a location appropriate to the pulse
height. The memory have 1000 channels each 10,20,40 and
80 eV wide so that the total energy range covered would be
10, 20 ,40 and 80 keV respectively.

Materials Characterisation Techniques-Online

27 Course
 Materials Characterisation Techniques-Online
28 Course
 Resolution of the instrument
Energy resolution of a spectrometer is defined by its ability to
separate two adjacent peaks in the energy spectrum.
In EDS it is determined by the detector.
The resolution  of the detector can be calculated in accordance with
 = (P2+I2)1/2
 = FWHM (eV)
P = Electronic noise contribution
I = intrinsic line width of the detector
I = 2.35(FE)1/2
F = Fano factor
 = mean energy necessary to create an electron –hole pair in the
detector
E = energy of the x-ray photon.
Assume electronic noise P = 0
Then the theoretical limit for a Si(Li) detector can be estimated
Taking F = 0.11
 = 3.8 eV
E = 5.9 KeV (Mn Kα)
I = 111 eV practically it is about 130 eV due to electronic noise
29 Materials Characterisation Techniques-Online
Course
 Materials Characterisation Techniques-Online
30 Course
 Artifacts in EDS spectrum

Materials Characterisation Techniques-Online

31 Course
 Escape peaks.
 Escape peaks in a Si(Li) detector occurs when the
absorption of an incoming x-ray results in the excitation of
a Si Ka x-ray which subsequently escapes from the
detector.
 So the spectrum exhibits additional peaks a fraction of
the height of the parent peak and 1.74 keV lower in
energy.
 In the case of HPGe the escape peak is 9.885 keV below
the parent peak.

Materials Characterisation Techniques-Online

32 Course
 Sum peaks
 Sum peaks occurs when two high intensity
peaks arrives too close in time that the signal
processing electronics cannot separate them.

 A single peak is registered at an energy that

is the sum of the two peaks.

Materials Characterisation Techniques-Online

33 Course
 Applications
Materials Characterisation Techniques-Online
34 Course
 Quantitative analysis
 Specimen intensities are ratioed to
standard intensities: (Castaing’s
approach )

 Where K is the “K ratio” for element

i, I is the X-ray intensity of the unk
phase and subscript i is the element. I
Ki  i
std
 With the counts acquired on BOTH
unknowns and standards on the same
Ii
instrument, under the same operating
conditions, many physical parameters
of the machine that would be needed
in a rigorous physical model cancel
each other out (same in numerator
and denominator).
Materials Characterisation Techniques-Online
35 Course
 Castaing’s First Approximation
Castaing’s “first approximation”
follows this approach. The
composition C of element i of
the unknown is the K ratio
times the composition of the
standard. In the simple case
where the standard is the unk Iiunk std std
pure element, then, the C
i  std Ci  KiCi
fraction K is roughly equal to Ii
the fraction of the element in
the unknown.

The x-ray intensity generated

is linear with the
concentration

Materials Characterisation Techniques-Online

36 Course
  Note that the Fe-Mn alloys plot along
a 1:1 line, and so is a good reference.

 The Fe-Ni alloys plot above the 1:1

line (have apparently higher Fe than
they really do), because the Ni atoms
present produce X-rays of 7.278
keV, which is greater than the Fe K
edge of 7.111 [Link], additional Fe
Ka are produced by this secondary
fluorescence

 The Fe-Cr alloys plot below the 1:1

line (have apparently lower Fe than
they really do), because the Fe atoms
present produce X-rays of 6.404
keV, which is greater than the Cr K
edge of 5.989 keV. Thus, Cr Ka is
increased, with Fe Ka are “used up” in
this secondary fluorescence process.

Materials Characterisation Techniques-Online

37 Course
 ZAF Correction factor
The observed x-ray intensities must be
corrected for atomic number (Z), absorption
(A) and Flouresence (F)

Materials Characterisation Techniques-Online

38 Course
 Z Factor
Stopping power correction (S)

 Incident electrons lose energy in inelastic

interactions with the inner shell electrons
of the target. The “stopping power”
(Stopping power is the ability of a
material to reduce the energy of an
incident electron by inelastic scattering)
is not constant but drops with increasing
Stopping power of pure elements for
Z. A higher number of X-rays will be
20 keV electrons
produced in higher Z targets.

Backscatter correction (R)

 The fraction of high energy incident electrons

that are backscattered (η) increases with
atomic number. There then will be relatively less
incident electrons penetrating into higher Z
specimens, resulting in a smaller number of X-
39
rays. Materials Characterisation Techniques-Online
Course
 Z correction

The total atomic number correction is formed

by multiplication of the R and S of the unknown
and standard :
Z = Rstd/Runk * Sunk/Sstd

Overall the backscatter and the stopping

power corrections tend to cancel each
other out. But if there is a (small)
correction, it is usually in the direction of
the backscatter correction.

Materials Characterisation Techniques-Online

40 Course
 Absorption

x-rays produced have to penetrate the specimen layers above

to fly finally to the x-ray detector. In this way it is weakened within
the sample (self absorption) and within the x-ray entrance windows of
the detector


  = mass absorption coefficient
I   
 exp  x    = density of absorber, g/cm3
I0     x= thickness of absorber, cm
 I0= generated x-ray intensity
 I= intensity following absorption
Materials Characterisation Techniques-Online
41 Course
 Absorption correction
 The absorption (A) correction is
then defined as
 A= f(Х)std / f(Х)sample

 Where f(Х) is the absorption

factor

f ( )  emergent intensity
generated intensity

Materials Characterisation Techniques-Online

42 Course
 Fluorescence Correction
The X-rays produced within a specimen have the potential for
producing a second generation of X-rays: this is secondary
fluorescence, generally termed as fluorescence. This occurs when
the characteristic X-ray has an energy greater than the absorption
edge energy of another element present in the specimen.
 The form of the correction F is
1
F
If
1
Ip
 where If/Ip is the ratio of emitted X-rays from fluorescence,
compared to the X-ray intensity from inner shell ionization. In a
compound, this term is summed overall all the elements that could
fluorescence the element of interest.
Materials Characterisation Techniques-Online
43 Course
  The absorption and fluorescence corrections for TEM
specimens are very much reduced or negligible because of
the thin-ness of the specimen.

Materials Characterisation Techniques-Online

44 Course
 Procedure for quantitative
analysis
 Background removal (linear or non linear curve
fitting)
 Peak deconvolution (removal of peak overlaps)
 Counting net intensities of the characteristic
x-ray peaks
 Composition is then obtained using the equation

unk unk
unk I ZAFi std
C i  i
std std Ci
I i ZAFi

Materials Characterisation Techniques-Online

45 Course
 Single line scan

 The electron probe is

scanned slowly in a straight
line across a chosen region
of the specimen.

 The out put obtained in the

computer will be a plot of
distance versus counts.
EDS profile of O, Fe and Co concentrations
obtained from the line spectra scanning
procedure.

Materials Characterisation Techniques-Online

46 Course
 Elemental distribution map
 A elemental distribution map
is obtained by scanning a
whole field with the
spectrometer tuned to a
chosen spectral line or
energy.

 Colour display enables the

distribution of several
elements to be seen on a
single map.

Materials Characterisation Techniques-Online

47 Course
 Conclusions
 Characteristic X-rays produced as a result of electron
interaction with matter can be usefully manipulated for
elemental analysis.

 EDS provides easily detection of elements in the range

Z = 11 to 92 and Z = 4 to 92 for windowless detectors.

 If properly analyzed, this technique can be used for

finding chemical composition.

 Concentrations less than or equal to 1000 ppm can be

easily detected.

Materials Characterisation Techniques-Online

48 Course