Transcript - Mikroanalytik.de


EDX-Spectra Simulation
Optimization of Excitation Conditions and Detection Limit
Calculations in EPMA
F. Eggert, Röntgenanalytik Apparatebau GmbH, Berlin
•
Introduction
•
Theory of simulation complete spectra
•
Applications of spectra simulation
•
Determination of detection limits with spectra simulation
•
Summary
EDX – Spectra Simulation
Inroduction

Introduction
• Today the standardless evaluation of measured spectra is an established methodology in
Electron Probe Microanalysis (EPMA) with Energy Dispersive Spectrometer (EDX) in
Scanning Electron Microscope (SEM)
• New developments offer the possibility to calculate complete spectra in dependence to
analytical conditions (spectra simulation).
• The basics are:
- Exact knowledge about all X-ray lines of elements and about other atomic data
- Knowledge about absolute cross-sections of both, the Characteristic X-rays
and the Bremsstrahlung
- Calculation of excitation and absorption of X-rays in specimen and detector
(characteristic radiation and Bremsstrahlung)
- Calculation of the entire Bremsstrahlung-deviation as the main background and
simulation of other background components
- Simulation of detector-resolution influence and count-statistics to simulate
realistic spectra
The content of presentation is to show the benefits of spectra simulation to daily analytical
work with Electron Microscope and EDX.
EDX – Spectra Simulation
Basics

Theory of Simulation
• The ratio of emitted counts of characteristic X-ray quanta to the counts of emitted
Bremsstrahlung-quanta with same energy (in an specified energy region) is known as
P/B-ratio (or P/U in this equation).

Nich
S R A (1  FC  FB ) ich
( P U )i  br  cii qi
S R A bri
Ni
• Calculation of the Bremsstrahlung deviation for all spectra channels taking into account
the self-absorption Albr and detector-absorption l in specimen.
 mass absorption coefficients (µ/) = f (Z , E)
 absorption jumps of (µ/) with energies EC
2






E

E
E

E
br
o
l
o
l
U l   l Al a
b

E
E
l
l


X
Lifshin
empiric 2.parameters
Kramers
l is the index of current channel during spectra calculation
EDX – Spectra Simulation
Basics

Theory of Simulation
+
•
•
•
•
•
All line- and shell- energies
Relative emission rates of a single shell
Excitation of sub-shells
Coster-Kronig transitions
Fluoresence yields
Bremsstrahlung + Lines
Escape + Artefacts (ICC)
Count statistics (Noise)
____________________
= Simulated Spectrum
(2000 cps, 3 minutes)
EDX – Spectra Simulation
Basics

Atomic Data Library (Data Base)
To make the simulation possible, an atomic data library with fast access to all
element specific data is necessary:
The accuracy of data base is crucially
for quality of simulation!
EDX – Spectra Simulation
Application

Optimization Before the Measurement: Eo
15 keV
20 keV
25 keV
30 keV
EDX – Spectra Simulation
Application

Verification: Excitation of Lines (Eo)
Excitation of Au-L
lines (Sub-Shells !)
with different Eo
EDX – Spectra Simulation
Application

Optimization / Verification: Tilt-Angle
AuAg-Alloy
Eo: 15 keV
tilt: -30o...+30o
Simulation
Absorption-Effects:
- Irregular Surfaces
- Rough Specimen
- Particle
EDX – Spectra Simulation
Application

Optimization: Influence of Detector-Resolution
AuAg-Alloy:
125 eV vs. 165 eV
EDX – Spectra Simulation
Application

Verification of Possible Overlap-Problems
5% Pd in Pb
with/without Pd
EDX – Spectra Simulation
Application

Element-Identification (Verification of Unknown Peaks)
Si in Specimen ?
...with Escape
... without Escape
No !
EDX – Spectra Simulation
Application

Element-Identification (Comparison with Real Spectra)
Spectrum with Ba
...measured spectrum
...simulated spectrum
•
•
Additional elements ?
Improve data-base ?
Compare !
EDX – Spectra Simulation
Application

Teaching (Simulation of EDX X-Ray Acquisition Process)
15s Acq.time
2000 cps
„Acquisition“
ready ...
EDX – Spectra Simulation
Detection Limits
Calculation of Detection Limits
• The question is, whether an element in specimen with expected concentration is
detectable or not?
• If an element is detectable... How are the optimal measurement and excitation conditions
(SEM and spectrometer parameters) and how long does it take (acquisition time)?
• The signal/background-ratio is the base for calculations of detection limits
( P/B-ratio)
Counts
... determination is possible with spectra-simulation !
Counts

Probability
Significance Level NS
Probability
Detection Limit NDL
EDX – Spectra Simulation
Detection Limits

Detection-Limit of an Element with Different Specimens
MDL for Pd in Te
M
L
K
MDL for Pd in Au
EDX – Spectra Simulation
Detection Limits

Detection-Limits with Varying Conditions
Al in Cu
M
L
K
EDX – Spectra Simulation
Detection Limits

Simulation of Spectra-Acquisition Near Detection-Limits
 MDL = 0.2 %
•
•
•
Significant element
presence !
Concentration below the
detection-limit !
Is it really possible ... ?
Yes  You had luck !
Al: 0.15 %
Al: 0.3
#3 %
#2
#1
Al: 1 %
EDX – Spectra Simulation
Detection Limits

Simulation of Spectra Acquisition / Detection Limit = f (time)
5 s: MDL = 1.8%
10 s: MDL = 1.3%
20 s: MDL = 0.9 %
detectable !
50 s: MDL = 0.6 %
100 s: MDL = 0.4 %
2000 cps
1% Zr
in Sn ?
EDX – Spectra Simulation
Summary

• It is possible to calculate the entire EDX-spectrum with a standardless EPMA-correction
model taking into account all effects of specimen- and detector-interaction.
• Spectra simlation is useful for a better understanding and interpretation of measured
spectra.
• With spectra simulation all complex effects of excitation, absorption and detection are
shown very descriptive and didactically (teaching, coaching, …)
• The simulation of several excitation situations gives the possibility to optimize all
conditions even before the actual specimen maesurement and data acquisition.
• With spectra simulation the analyst is able to make estimations for detection limits.
• Effects of counting-statistics are possible to verify.
Future View:
 Application of spectra simulation for interactive qualitative analysis
(displacement of simple line-mark identification)
 Calculation of entire spectrum for a visual comparison after quantitative
evaluation (reconstruction) for an improvement of final result-reliabilities