Introduction to Electron Energy

Dispersive Spectroscopy
(EDS, EDX, EDXS or XEDS)

Weilie Zhou

Advanced Materials Research Institute,

University of New Orleans, LA 70148
Email:wzhou@uno.edu
 Outline
1.Signal Original of the EDS

2.How does EDS work?

3.Instrumentation (Si(Li) and Silicon

Drift Detector (SSD))

4.Important EDS parameters

5. Examples
1. Origin of X-rays

(Frenchman Peirre Auger)

(100 keV)

(Characteristic X-rays &

Bremsstrahlung X-rays)

• What are electron shells ?

• What is the critical ionization energy ?

Inelastic interaction
 Characteristic X-rays

The complete range of possible electron transitions that give rise to K, L, and M
characteristic X-rays. Note all these X-rays are detectable by EDS in the TEM.
 Spatial Resolution
Spatial resolution of X-ray microanalysis
As the smallest distance between two volumes from which independent
X-Ray microanalyses can be obtained.

(A) Monte Carlo simulations of 103 electron trajectories through a 100 nm thin
foil of Fe at 100 kV and 300 kV. Note the improved spatial resolution at higher
kV . (B) conversely, in a bulk sample the interaction volume at 30 kV is
significantly more than at 10 kV, giving poorer X-Ray spatial resolution at higher
kV.
Spatial Resolution Equation
Since the beam retains a
Gaussian intensity
distribution.

R= (b2+d2)1/2

Considering worst case (given

by the exit beam diameter)

R=(d+Rmax )/2
2. How does EDS work?

1. The detector generates a charge pulse proportional

to the X-ray energy.

2. The pulse is first converted to a voltage.

3. The signal is amplified through a field effect

transistor (FET), isolated from other pulses, further
amplified, then identified electronically as resulting
from an X-ray of specific energy.

4. Finally, a digitized signal is stored in a channel

assigned to that energy in the MCA (multi-channel
analyzer).
• The three main parts are the detector,
the processing electronics, and the
MCA display

• The detector is reversed-biased p-i-n

diode

• Silicon-lithium {Si(Li)] semiconductor

detectors
 3. Instrumentation Si (Li) Detector

EDAX
Oxford
Genesis
 Silicon Drift Detector

OmniProbe 400

Octane Elect SDDs

 Si (Li) Detector
Si about 3.8 eV
X-ray larger than 1kV

(Acceptor Impurities)

The p and n regions are called “dead

layers” and the intrinsic region in
between is referred to as the “active
layer”
 Intrinsic Germanium Detectors

Si (Li) detectors show a drop of efficiency above 20kV.

The intrinsic GE (IG) detector have a high efficiency at higher voltages, but have
very small K ionization of high energy
Why do we have to cool the Si (Li) detector
with liquid N2?
1. Thermal energy would activated electron-hole pairs,
giving a noise level that would swamp the X-ray signals
we want to detect.

2. The Li atoms would diffuse under the applied bias,

destroying the intrinsic properties of the detector.

3. The noise level in the FET would mask signals from

low-energy X-rays.
The cooling consequences is that both hydrocarbons and ice from microscope
environmental can condense on the cold detector surface, causing absorption of
low energy X-rays. There are two obvious solutions to this problem
• Isolate the detector from the microscope vacuum
• Remove hydrocarbons and water vapor from the microscope
So, detectors are sealed in a pre-pumped tube with a “window” to allow X-rays
through into the detector
Different kinds of windows

Beryllium window detectors

Using a 7 um thin sheet of beryllium which is transparent to most X-
rays, and can withstand atmospheric pressure when the stage is vented
to air ( $3M/pound). But still too thick and can’t detect B, C, N, and O.

Utrathin window (UTW) detectors

Windows that are less absorbent than Be; usually these are made from
very thin (<100 nm) films of polymer, diamond, boron nitride, or silicon
nitride, all of which are capable of withstanding atmospheric pressure
while still transmitting 192 eV boron Kα X-ray.

Windowless detectors
Use windowless detectors in UHV AEM. Take great care to eliminate
hydrocarbons from your specimen and keep the partial pressure of water
vapor below 10-8 Pa best detection is Be (110eV) Kα X-rays
Energy Efficiency of Different Detectors
 Silicon Drift Detector
Silicon drift detectors (SDDs) are X-ray radiation detectors used in x-ray
spectrometry (EDS) and electron microscopy (EDX). Their chief
characteristics compared with other X-ray detectors are:
• high count rates, faster, and lower noise (1 pair/ 3.8eV X-ray)
• Comparatively high energy resolution
(Mn Kα energy resolution down to 121 eV,
compared Si(Li)) 140eV)
• Peltier cooling (no LN2)

Working principle:

In the SDD, this material is high purity silicon with a very low leakage current. The
high purity allows for the use of Peltier cooling instead of the traditional liquid
nitrogen. The major distinguishing feature of an SDD is the transversal field
generated by a series of ring electrodes that causes charge carriers to 'drift' to a
small collection electrode. The 'drift' concept of the SDD (which was imported from
particle physics) allows significantly higher count rates coupled with a very low
capacitance of the detector.
 Advantages:
• In older detector designs, the collection electrode is centrally located with
an external FET (field effect transistor) to convert the current into a voltage
and thus represents the first stage of amplification. Newer designs
integrate the FET directly into the chip, which greatly improves energy
resolution and throughput. This is due to the reduction of capacitance
between anode and FET, which reduces electronic noise.

• Other designs move the anode and FET outside of the irradiated area. This
causes a slightly longer response time, which leads to a slightly lower
throughput (750,000 counts per second instead of 1,000,000). However,
due to the smaller anode size, this leads to better energy resolutions (down
to 121 eV for M Kα wavelength). Combined with improved or adapted
signal processing, it is possible to maintain the Silicon drift detector's
energy resolution up to 100,000 counts per second.
4. Important EDS parameters

1. Time constant and dead time

2. Resolution of the detector

3. What you should know about your EDS?

4. Operating factors
Pulse Processing

The electronic components attached to the detector convert the

charge pulse created by the incoming X-ray into a voltage pulse,
which can be stored in the appropriate energy channel of the MCA.

A single isolated X-rays enters the detector and creates a pulse of

electron at the back of the Si(Li) crystal.

•The charge pulse enters the FET, which acts as a preamplifier and
converts the charge into a voltage pulse.

•This voltage pulse is amplified several thousand times by a pulse

processor, and shaped so that an analog-to-digital converter can
recognize the pulse as coming from an X-ray of specific energy.

•The computer assigns it to the appropriate channel in the MCA

display.
1. Time constant and dead time

Time constant is the time (about 10-50us) allowed for the pulse
processor to evaluate the magnitude of the pulse .

• If you select a longer constant, the system is better able to

assign an energy to the incoming pulse, but few counts can be
processed in a given analysis time.

• The shortest time constant will allow you to process more

counts per second (cps) but with a greater error in the
assignment of a specific energy to the pulse, and so the
energy resolution will be poorer.

• A longer time constant will give better resolution but the count
rate will be lower
Dead time

When the electronic circuitry detects the arrival of a pulse, it takes less than a
microsecond before the detector is effectively switched off for the period of
time called the dead time while the pulse processor analyzes that pulse.

The dead time will increase as more X-rays try to enter the detector, which
closes down more often. The dead time can be defined in several ways.

Dead time in %=(1-Rout/Rin) x100%

(Rout output count rate, Rin input count rate)

Or Dead time in%=(clock time-live time)/live time x 100%

If collect a spectrum for a “live time” of 100s, this means that the detector
must be “live” and receiving X-rays for this amount of time. If it is actually
dead for 20s while it is processing the X-rays, then the dead time will be
20%, and it will take 120s of “clock time” to accumulate a spectrum. If Dead
time excesses 30%, that means that the detector is being swamped with X-
rays and collection becomes increasingly inefficient. It is better to turn down
the beam current or move to a thin area of the specimen to lower the count.
2. Resolution of the detector

We can define the energy resolution R of the detector as follows:

R2 = P2+I2+X2

P is a measure of the quality of the associated electronics

I is the intrinsic line width of the detector which is controlled by
fluctuations in the number of electron-hole pairs created by a given X-ray
X is the FWHM (the full width at half maximum) equivalent attributable of
a randomized electronic pulse generator.

Typically, Si (Li) detectors have a resolution of ~140 eV at Mn Kα with the

best being 127 eV. The best reported IG resolution is 114eV. SSD detector
is about 121 eV
3. What you should know about your EDS?
1. Detector Variables
The detector resolution
• Damage to the intrinsic region by high-energy fluxes of radiation
• Bubbling in the liquid-N2 dewar due to ice crystals building up

Incomplete Charge Collection (ICC)

• Because of the inevitable presence of the dead time.
Usually , the peak will have a low-energy tail, because some energy
will be deposited in the dead layer and will not create detectable
electron-hole pairs.

Detector Contamination
• Over a period of time, ice and hydrocarbon will eventually build up on
the cold detector surface or on the window.
The NiKα/NiLα ratio will rise with time if contamination or ice is building
up on the detector and selectively absorbing the lower-energy L line

•Your detector resolution at the Mn or CrKα line (typically 140-150 eV for Si (Li)
and 120-130 eV for IG)
•The ICC defined by the FWTH/FWHM ratio of CrKα line (ideally 1.82)
•The ice build-up reflected in the Ni (or Cr) Kα/Lα ratio.
 2. Processing Variables

• First, check the calibration of the energy range of the spectrum

• Second, check the dead-time correction circuitry
• Third, check the maximum output count rate

Collect a spectrum from a material which generates a pair of K X-ray lines

separated by about the width of the display range. (e.g. Al-Cu for 0-10 KeV).
If the computer markers are correctly positioned at the peak centroid (e.g. Al
Kα at 1.54 KeV and Cu Kα at 8.04 keV and marker are more than 1 channel
(10eV) apart, then you should re-calibrate your display using the software
routine supplied by the manufacturer.
4. Operating Factors

Count Rate
For a good quality spectrum (ie. Good resolution and fewest artifacts) you
should use the 40 us time constant (pulse processing time) and 500 to 2000
cps. These are good numbers if the sample consists largely of high energy
peaks (>1 keV), but if the spectrum is dominated by low energy peaks (1
keV) then a count rate of 500-1000 cps is better.

Accelerating Voltage
The over voltage is a ratio of accelerating voltage used to the critical
excitation energy of a given line for an element. Typically, the over voltage
should be at least 2 for the highest energy line and no more that 10 to 20
times the lowest energy line of interest.

Take-Off Angle
Typical take-off angle will range from 25 to 40 degrees. This angle is a
combination of the detector angle, its position, sample working distance and
sample tilt. The sensitivity for very low energy X rays and/or signals
characterized by high absorption can be enhanced by increasing the take-off
angle.
 5. Operation EDS and Peak Identification
Quantitative X-ray Microanalysis

Ci Concentration of an element i in the specimen

C (i) Known standard of composition for element i
Ii Measured intensity emerging from the specimen
I (I) Measured intensity emerging from the standard

Ci / C (i) =[K] Ii / I (I)

Where K is a sensitivity factor that takes into account the difference between the
generated and measured X-ray intensifies for both the standard and unknown
specimen. The contribution to K come from three effects:

Z The atomic number

A The absorption of X-rays within the specimen
F The fluorescence of X-rays within the specimen

The correction procedure in bulk microanalysis is often referred to as the ZAF

correction.
 Peak Identifications

X-ray Artifacts

1. Escape Peaks (signal–detection artifacts)

2. Sum Peaks (signal-processing artifacts)
3. The internal-fluorescence peak
(Stray Radiation, A Warming Detector, etc.)

The input count rate is high.

Fluoresces a Si Kα X-ray The dead times are in excess of about 60%.
There are major characteristic peaks in the spectrum.
 Element Silicon Drift Detectors (SDDs)

Figure shows a 5 kV spectrum of a cross section of a

printed circuit board showing a low energy Sn M peak in
between C K and O K peaks of an epoxy.

The Element SDDs deliver powerful analytical capability in a compact package,

maximizing performance and flexibility, while providing streamlined operation to
guarantee fast results and ease of use. The Element SDDs provide fast and
efficient results for industrial analysis needs.
 Octane Elect SDDs

SiO2

Spectra acquired from a silicon dioxide sample at 10 kV. The

comparison of the scaled spectra to the Si K peak clearly shows the
increased oxygen and carbon peak intensities achieved with a
Si3N4 window.

The Octane Elect SDDs offer both increased functionality and high resolution at
an optimal value. The Octane Elect SDDs provide remarkable improvements in
low energy sensitivity for light element detection and low kV microanalysis.
Octane Elite SDDs

Cr-N

Spectra acquired from a chromium nitride sample at 10 kV. The

comparison of the scaled spectra to the Cr K peak clearly shows the
increased nitrogen and oxygen peak intensities achieved with a
Si3N4 window.

This line of detectors incorporates a new silicon nitride (Si3N4) window, which offers
remarkable improvements in low energy sensitivity, light element detection, and low kV
microanalysis. The Octane Elite SDDs also use state of the art electronics, which yield high
speed X-ray data processing within a smaller and fully vacuum encapsulated detector device.
 Octane SDD Series for the
Transmission Electron Microscope (TEM)

SiO2

Spectra of SiO2 collected with the Octane T Plus detector (shown in red) compared to a SiLi detector (shown in blue). The
display of these spectra has been normalized to the highest energy part of the spectrum and shows that the Octane T Plus
is much more sensitive to low energy peaks. The Si K series peak intensity is improved by 30% and the O K by 150% with
the Octane T Plus detector.

EDAX’s Octane SDD Series for the TEM are the world's first SDDs for the TEM that are
fully integrated. Data acquisition and signal processing electronics are fully integrated into
the detector. The integrated detector presents an elegant design that improves
performance, facilitates installation and offers easy remote access via Ethernet from
virtually any computer. For more information, please visit
http://www.edax.com/products/eds/detectors/silicon-drift-detector-TEM.aspx
EDS Mapping
EDS Spectrum
 ZnO/ZnS Core/shell Nanowire

Structural characterization of a ZnO/ZnS core/shell nanowire showing low magnification

top view, FESEM images of (a) an as-grown, ZnO nanowire array and (b) a ZnO/ZnS
core/shell nanowire array, (c) an HRTEM image of a single ZnO/ZnS core/shell nanowire,
and (d) the corresponding energy dispersive spectroscopy (EDS) lateral line scan
depicting elemental peaks characteristic to core/shell nanowire structure.

Satish C. Rai, et al., ACS NANO 9, 6, 6419–6427 (2015), 10.1021/acsnano.5b02081

(a) XRD patterns of Co(CO3)0.5(OH)∙0.11H2O and CoP nanowire arrays. (b) SAED pattern of CoP
(orthorhombic) nanowire. (c and d) SEM images of Co(CO3)0.5(OH)∙0.11H2O and CoP nanowire
arrays. (e) TEM image, insert shows the HRTEM lattice image of (111) plane. (f) HAADF-STEM
and corresponding EDS elemental mapping images of P, O, Co and P+O. (g and h) dark field
image and EELS mapping (Co2+), showing 8 nm cobalt oxide outside the NW.
Z. Zheng et al, ACS Appl. Mater. Interfaces 2017, 9, 16986,39
DOI: 10.1021/acsami.7b01109
 FeP/PEDOT
(a) TEM image, (b) ED
pattern, and (c) HRTEM image
of FeP/PEDOT nanorod, (d
and e) EELS spectrum and (f)
EDS spectrum of FeP/PEDOT
nanorods, and (g) HAADF-
STEM image and the
corresponding EDS elemental
mapping images of Fe, P, O,
C, S and Fe + S, respectively.
(h) EDS line profiles extracted
from each map along with the
white arrow (from A to B; 200
nm in distance) represented
on the Fe + S map. Each
profile is normalized at
maximum intensity.

Z Zheng, et al. Chem Comm. 2018, 54, 794 DOI: 10.1039/c7cc09163j

Mapping of Atoms
 Summary
Energy Dispersive X-ray Microanalysis is an analytical technique
used for the elemental analysis or chemical characterization of a
sample. Si(Li) and Silicone Drift Detector (SDD) are commonly used in
element detection. Liquid Nitrogen cooling is needed for Si(Li) detector
and no need for SDD. With different detectors, the energy resolution,
speed, and noise are different. SDD detector has better performance
with higher count rates, faster, lower noise (1 pair/ 3.8eV X-ray) and
comparatively higher energy resolution (Mn Kα energy resolution down
to 121 eV). Using different windows would also improve counts reading.
In addition to the quantification of elements, the spectrum can be used
for element mapping down to atomic level.