Transcript Slide 1

X-Ray Fluorescence (XRF)
Becoming the most widely used method for elemental analysis of
solids
ADVANTAGES AND DISADVANTAGES
Advantages of X-Ray Spectrometry
* Simple spectra
* Spectral positions are almost independent of the chemical state
of the analyte
* Minimal sample preparation
* It is non-destructive
* Applicable over a wide range of concentrations
* Good precision and accuracy
Disadvantages of X-Ray Spectrometry
* X-ray penetration of the sample is limited to the top 0.01 - 0.1
mm layer
* Light elements (below 22Ti) have very limited sensitivity
although C is possible on new instruments
* Inter element (MATRIX) effects may be substantial and
require computer correction
* Limits of detection are only modest
* Instrumentation is fairly expensive
NOMENCLATURE
Simplified spectral lines observed in x-ray spectra (each energy shell
actually comprises several energy levels, thus transitions are more
numerous than shown).
PRINCIPLES OF X-RAY FLUORESCENCE
X-Ray Excitation
AUGER ELECTRON EMISSION (internal photoionisation)
The Auger effect is more common in
elements of low Z because their atomic
electrons are more loosely bound and their
characteristic X-rays more readily
absorbed.
X-RAY FLUORESCENCE YIELD
The yield of X-ray photons is reduced by the Auger effect.
The fluorescence yield (ω) is the ratio of X-ray photons
emitted from a given shell to the number of vacancies
created in that shell.
Since production of Auger electrons is the only other
competing reaction the ratio of Auger electrons to
vacancies must be 1- ω.
X-RAY MASS ABSORPTION COEFFICIENTS
A plot of mass absorption coefficient vs energy of the X-ray photon
for 82Pb.
Abrupt changes are observed corresponding to absorption edges for
K, L and M electrons. At the energy (wavelength) of the edge, the
photons first become sufficiently energetic to eject K, L and M
photoelectrons.
X-RAY SOURCES
The X-ray tube is energised by a high-voltage power supply with an
output of 0.5 to 50 kV. The head of the vacuum tube consists of a target
(anode), which is often made of tungsten and chromium. As accelerated
electrons strike the target, X-rays are emitted. The tube also forms the
X-rays into a beam through a Beryllium window.
In the target the 74W is used to excite L lines of higher Z atoms and the
24Cr is used to excite atoms of 22Ti and below.
Spectral Output
X-ray spectrum produced by electron bombardment of a tungsten
target:
Continuum Spectrum: The continuum results from deceleration
of electrons by the atoms in the target. The German term
"bremsstrahlung" means "braking radiation."
Characteristic Spectra: Electron bombardment also produces
characteristic peaks provided the accelerating potentials are
sufficiently high.
DETECTORS
X-ray photons (as well as other energetic particles) can be
measured using the following types of detectors:
* gas-filled detectors register a current pulse from the
collection of electron-ion pairs formed;
* a semiconductor detector register a current pulse from the
formation of electron-hole pairs;
* a scintillation detector counts light pulses created when an
X-rays passes through a phosphor;
* a photographic plate.
Gas-Proportional Counters
A gas-proportional counter is filled with P90 gas (90% argon, 10%
methane). X-rays ionise the gas, leaving electrons that migrate to the
anode and positive ions that move to the case.
Proportional counters use gas amplification: the detector voltage is raised
to 500 to 700 V so that the primary electrons and ions, first formed, are
accelerated to produce secondary electrons and ions when they collide
with gas atoms. This yields a greatly increased signal which is,
nevertheless, proportional to the energy of the original x-ray.
Solid-State Detectors
When an electron enters the crystal it ejects a high-energy photoelectron which
ultimately dissipates its energy in multiple interactions which promote valence band
electrons to the conduction band, leaving holes in the valence band. The electronhole pairs are then collected by biasing the detector at -1000 V, giving rise to a
current pulse for each x-ray entering the detector. Charge collection is much more
efficient than in a gas.
Lithium-Drifted Si(Li) Detectors
A lithium-drifted Si(Li) detector is manufactured from high-purity p-type silica.
However, p-type silica of sufficiently high purity is difficult to fabricate. Most Si
crystals contain extrinsic holes, caused by impurities, which allow significant
"leakage" of current at the required bias voltage. In order to compensate for
these extrinsic holes, lithium, an n-type dopant, is diffused into the material at
350 - 450ºC under an electrical gradient. The lithium atoms compensate for the
extrinsic charge-carriers in the p-type silicon and provide a wide "intrinsic"
region of high resistance.
Si(Li) detectors are operated at 77 K with a liquid N2 cryostat to prevent further
diffusion and to reduce the level of random noise due to the thermal motion of
charge carriers.
Silicon Drift Detectors
Like other solid state X-ray detectors, silicon drift detectors measure the energy
of an incoming photon by the amount of ionization it produces in the detector
material. 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.
WAVELENGTH DISPERSIVE XRF SPECTROSCOPY
Instrumentation
In wavelength dispersive spectrometers, the several x-ray lines emitted
from the sample are dispersed spatially by crystal diffraction on the basis
of wavelength. The detector then receives only one wavelength at a time.
The crystal and detector are made to synchronously rotate through
angles of θ and 2θ respectively.
ENERGY DISPERSIVE XRF SPECTROSCOPY
The primary X-ray beam excites several spectral lines from the
sample. In energy dispersive XRF all wavelengths enter the detector at
once. The detector registers an electric current having a height
proportional to the photon energy. These pulses are then separated
electronically, using a pulse analyser.
WAVELENGTH AND ENERGY DISPERSION COMPARED
Advantages of Energy Dispersion:
* simplicity of instrumentation - no moving parts
* simultaneous accumulation of the entire X-ray spectrum
* qualitative analysis can be performed in 30 s, or so
* a range of alternative excitation sources can be used in place of
high-power x-ray tubes with their large, heavy, expensive and
power-consuming supplies
* alternative sources include, low power x-ray tubes, secondary
monochromatic radiators, radioisotopes and ion beams.
Advantages of Wavelength Dispersion:
* resolution is better at wavelengths longer than 0.08 nm
* higher individual intensities can be measured because
only a small portion of the spectrum is admitted to the
detector
* with multichannel analysers sensitivity for weak lines in
the presence of strong lines is limited because the
strongest line determines the counting time
* lower detection limits are possible
SAMPLE PREPARATION
Reproducible sample preparation methods are essential. Samples must be
in a form that are similar to available standards in terms of matrix, density
and particle size.
* Solids, generally solids must be polished as surface roughness may give
erratic results.
* Powders and pellets, powdered samples are often pressed into pellets,
suspensions may also be analysed
* Fusions, with potassium pyrophosphate (K2P2O7) or a tetraborate
(Na2B4O7 or Li2B4O7) present a homogenised sample which can often be
analysed directly
* Liquids and solutions, a x-ray transparent cover and sample cup must be
provided to prevent volatility under vacuum conditions
* support media, such as filter paper, millipore filters, ion-exchange
membranes
MATRIX EFFECTS
Types of Matrix Effects
In XRF absorption-enhancement effects arise from the following
phenomena:
1. The matrix absorbs primary x-rays (primary-absorption effect); it may
have a larger or smaller absorption coefficient than the analyte for primary
source x-rays
2. The matrix absorbs the secondary analyte x-rays (secondary-absorption
effect); it may have a larger or smaller absorption coefficient for the analyteline radiation
3. The matrix elements emit their own characteristic lines, which may lie on
the short wavelength side of the analyte absorption edge, thereby exciting
the analyte to emit additional radiation to that excited by the primary source
of X-rays alone (enhancement)
QUANTITATIVE ANALYSIS
1. Calibration-Standard Methods. The analyte-line intensity from samples is compared
with that from standards having the same form as the samples and, nearly as
possible, the same matrix.
2. Internal Standardisation. The calibration-standard method is improved by
quantitative addition to all samples of an internal standard element having excitation,
absorption and enhancement characteristics similar to those of the analyte in the
particular matrix. The calibration function involves measuring the intensity ratio of the
analyte and internal standard lines.
3. Matrix-Dilution Methods. The matrix of all samples is diluted to a composition such
that the effect of the matrix is determined by the diluent rather than the matrix.
4. Thin-Film Methods. The samples are made so thin that absorption-enhancement
effects substantially disappear.
5. Mathematical Corrections. Absorption-enhancement effects are corrected
mathematically by the use of influence coefficients for each element present (these
are derived experimentally from reference samples). The basic approach is that the
XRF intensity at a particular wavelength will in some way be affected by each element
in the sample.
XRF applications
During the last two decades, the development in X-ray detectors has established
the XRF method as a powerful technique in a number application fields, including:
Ecology and environmental management: measurement of heavy metals in soils,
sediments, water and aerosols
Geology and mineralogy: qualitative and quantitative analysis of soils, minerals,
rocks etc.
Metallurgy and chemical industry: quality control of raw materials, production
processes and final products
Paint industry: analysis of lead-based paints
Jewelry: measurement of precious metals concentrations
Fuel industry: monitoring the amount of contaminants in fuels
Food chemistry: determination of toxic metals in foodstuffs
Agriculture: trace metals analysis in soils and agricultural products
Art Sciences: study of paintings, sculptures etc. in order to make an expertise
EDX Microanalysis –
As the name suggests, this refers to the analysis of a
sample on a microscopic scale, resulting in structural,
compositional and chemical information about the
sample.
There exists a whole host of analytical techniques that
exploit the many signals which may be generated within
the sample. X-ray microanalysis specifically gives
information about the elemental composition of the
specimen, in terms of both quantity and distribution.
On entering a sample, the energetic incident electrons undergo a number
inelastic and elastic scattering events resulting in a zig-zag path into the
sample until they either come to rest or are backscattered out of the
surface.
The distribution of trajectories is contained within the so called
‘interaction volume’, the shape and dimensions of which are strongly
affected by both:
- the atomic number and the incident energy of the electrons.
http://www.
matter.org.u
k/tem/defaul
t.htm
At any point along a given trajectory, characteristic X-rays can be
produced provided that the energy of the electron or indeed X-ray is
greater than the absorption edge associated with that characteristic
emission line.
The volume of the sample from which X-rays are produced is known as
the X-ray production volume or X-ray generation volume, the size
and dimensions of which depends on the X-ray line being excited
and the density of the material.
For example, in the case of lead, the sample volume producing the
higher energy L series X-rays will be smaller and nearer to the surface
than the volume from which M series X-ray lines are generated.
J(rz)
Depth
(microns)
The typical features of the j(rz) curve are shown in the
adjacent figure. Near the surface, X-ray production is greater
than for a thin unsupported film because of scattered electrons
travelling up from below being able to generate X-rays and
therefore resulting in the value of j(rz) being greater than one.
After initially rising to a maximum, the curve decreases due to
scattering and deceleration of the electrons, eventually falling
to zero.
Geometry
The position of the front end
of the detector in relation to
the surface of the sample is
important in order to optimise
the collection of X-rays.
Detector
Working
distance
Entrance
angle
Sample at ideal
working distance
for X-ray
microanalysis
Sample at incorrect
working distance for
X-ray microanalysis
EDX-Mapping
The X-ray spectrum detected by EDX can be used to construct a true color
response that would be obtained if the human visual sensitivity to the
electromagnetic spectrum could be offset to the X-ray wavelength region.
This color input is then used to augment a conventional electron image.
In this way the detail of the original electron image is retained whilst
portraying the underlying elemental composition because the spectrum
from each compound gives it a characteristic color. Thus topographic and
compositional information from all elements is compressed into a single
view giving the analyst a useful 'first look' to guide further microanalysis.
EDX mapping on nanoscale scale
EDX mapping on micron scale
SEM image
100 nm
Mix
100 nm
C map
Au map
100 nm
100 nm
Point and ID
Types of EDX analysis
Line scans
Mn doped Ge NW
Simple EDX
Point and ID
Line scans
EDX mapping
Images
Quantitative microanalysis
Quantitative analysis of elements in any sample, requires an accurate
measure of the intensity of peaks, before the concentration of elements in
a sample can be calculated. In determining peak areas in spectra, two
problems arise,
1) a typical spectrum contains characteristic peaks, which are
superimposed on a slowly varying background, which is 'noisy' because of
statistical variations. This background contribution needs to be carefully
subtracted from the spectrum.
2) The energy resolution of the detector imposes a limit on the separation
of peaks. Identification of peaks is generally not a problem, but
overlapping peaks require deconvolution, before being able to extract the
true peak intensities relevant to the elements present in the sample.
Once these intensities have been determined, a comparison is then made
with standards of known composition, followed by application of matrix
corrections, before the concentration of each element can be determined.
Background subtraction
The simple method of linear interpolation of the background beneath a peak
is not appropriate, since the background is non linear, both locally in the
vicinity of peaks, and over the entire energy range. There are various schools
of thought as how to best remove this background and to separate peaks
from each other:
1) One such method relies on the fitting of a theoretical background to the
spectrum background, and then using a least squares fitting technique to
obtain peak intensities. However, construction of a model to fit the
background requires an accurate knowledge of a number of physical
parameters, which are often difficult to determine precisely.
2) A second method suppresses the background, using a filtering method
which avoids any specific shape calculation. The peak intensities are then
obtained by using a least squares fit of standard peaks, in which the
background has also been suppressed.
Peak deconvolution
The energy resolution of an EDS imposes a limit on the separation of
peaks. There are several examples of peak overlap which commonly
occur: the Ka peaks for Be, B, C, O, N and F, are sufficiently close for the
tails of one peak to overlap into the neighbouring peak. Overlaps between
the Kb line of one element and the Ka line of another, often overlap.
Examples include the V Kb and Cr Ka (15eV apart) and overlaps between
lines from different shells e.g. between Mo La and S Ka, can often occur.
Fluorescence correction
Fluorescence arises as a result of the ionization of atom shells by X-rays
rather than electrons. The result is the emission of characteristic X-rays.
This contribution to the spectrum will be in addition to those X-rays which
have been produced directly by electrons. Fluorescence can only occur if
the energy of the incident X-ray is greater than the critical excitation
energy. There will be a fluorescence contribution from both the
continuum and characteristic X-rays but the contribution from
characteristic X-rays is the most dominant.
Electron
Beam
FeKa
CrKa
Resolution and count rate
Energy resolution is the primary test of detector performance, and the
main specification for an EDX detector is the resolution at Mn. The
benefits of improved resolution, are improved detection limits, because
a narrower peak is higher above the background. Well defined peak
shapes make peak ID faster and more reliable, and in addition,
overlapped peaks are better resolved, leading to significantly improved
detection limits, and accuracy of routines used in quantitative analysis.
The acquisition rate into the spectrum is important and this is related
to the input count rate, via the deadtime and the selected processing
time.
Input rate: This shows the approximate rate of photons striking the
detector.
Acquisition rate: This shows how fast the system is accumulating
spectrum counts.
Deadtime (%): is the percentage time for which the pulse processor is
unavailable for further counting. (see acquisition rate).