7)X-ray Spectrometry– HVC Capacitor, HV Ceramic Capacitor to build All kinds of X-ray machine.
EPMA may require a conventional SEM equipped with an ED
X-ray detector for determination of the elementary composition
of microsized objects; however, the intensive Bremsstrahlung
generated by the exciting electrons in the specimen body
increases drastically the minimum detection quantity of different
constituent elements. This limit can be improved if the SEM device
is equipped with an additional X-ray tube for excitation of such
particles or a microsized part of a large sample that can be used
optionally instead of the electron beam, as was published by
Bjeoumikhov and co-workers (C17 ). The researchers used a low-
power miniature tube that was a complete filter changer device
and different type of polycapillary and monocapillary optics in
order to provide afocal spot diameter from 5 to 100 ím depending
on the type of optic device. The authors demonstrated the
analytical capability of their new device with some measurements
performed on metal particles and alloys, and they outlined that
the analytical benefit of this combined method is the low detection
limit in the whole Z-range. Tanaka et al. reviewed (C18 ) a new
design of an undulator magnet developed for the synchrotron
facility of Spring8: a high-temperature superconductor magnet
is built and serves as a permanent magnet which is magnetized
in situ. The advantageous properties of the new magnet are that
the operating temperature is much higher than the liquid He
temperature and the higher magnetic field was found 4Tat
maximum. Chang and his international research group published
(C19 ) on an extreme device, a Fabry -Perrot resonator, for hard
X-rays; it was constructed from two Si crystal plates with 25-150-
ím thickness and 40-150- ím distance between them. They used
monochromatized SE with 14.4388-keV energy, and the ultra-high-
energy resolution of the four-crystal device was set at 0.36 meV.
The space between the Si plates acted as a cavity resonator for
the X-rays, the resonant frequency of the resonator was set by
tilting (along a horizontal axis) and rotating (along a vertical axis)
the resonator, and the back-reflected and forward-transmitted
beams interacted, resulting in a deviation of the X-ray intensity.
The researchers concluded that they observed for the first time
cavity resonance of X-rays in a Fabry-Perot resonator for hard
X-rays, and they predicted this effect as a new opportunity for
X-ray optics, microscopy, and spectroscopy. The in situ X-ray
spectroscopy measurements require sometimes extreme physical
conditions, e.g., high pressure for the study of the inner structure
or properties of the materials. Kunz et al. reviewed ( C20 ) a new
high-pressure cell for X-ray diffraction and spectroscopy purposes
at the Advanced Light Source with superconductive bending
magnet source between X-ray energies of 5 and 35 keV, up to a
pressure of 50 GPa and a temperature of 800 K. The beamline
has special optics, which consist of a plane parabola collimating
mirror, followed by a Kohzu monochromator vessel with Si(111)
crystals and W/B
4C multilayers, and then a toroidal focusing
mirror with a variable focusing distance. For setting the beam
position, an automated system was available and different slits,
ion chambers, and CCD area detectors are assembled as well.
Multilayers are a key device in X-ray monocromatization due to
their high reflectivity and resistivity against high X-ray flux;
however, there is a continuous demand for improved constructions
of new layer structures and materials. A Russian research group
with Artioukov (C21 ) developed a novel multilayer based on
depleted U for X-rays with 3-6-nm wavelengths. The reflectivity
of this new multilayer increased by a factor of 1.3-2 in comparison
with multilayers built of different structures of 3d elements. The
authors emphasized the importance of X-ray devices developed
for this wavelength range because it is near the C K-edge (E )
280 eV) that is the major element in biological samples and in
different organic materials, and X-ray optical devices are needed
for XANES, XRF, and EXAFS measurements below the C edge.
The depleted U coatings were found to be smooth enough and
forming a stable interface in different multilayer forms: U- C,
U2C3-C. Erko and co-workers described (C22 ) the beamline
capability at BESSY dedicated for XRS, XRF, diffraction, and small-
angle scattering in the micrometer and submicrometer spatial
resolution in the X-ray energy range of 2 -30 keV, combining all
these measuring techniques in identically the same beamline end
station. To arrange for a microsize focal spot for X-ray beams,
different optical elements were considered: a Bragg -Fresnel zone
lens that had a diffraction efficiency of 26%, monocapillary and
polycapillary lenses, and mirrors. An international research group
of Beyer published (C23 ) a newly constructed crystal spectrometer
for very accurate spectroscopy based on Si(211) and Ge crystals
in focusing compensated asymmetric Laue (FOCAL) geometry,
i.e., in transmission mode through a curved Laue crystal. The
spectrometer can be applied in the energy range of 30-120 keV;
the monochromatic X-rays were reflected from a limited area on
the crystal and were focused on a Rowland circle. The authors
described in detail the general geometrical scheme of the crystal
spectrometer, background suppression and shielding, the poly-
chromatic focusing technique, and the position sensitive X-ray
detector, and they dealt with the detection efficiency. The mono-
and polycapillary focusing devices are optimal optical elements
for analysis with micrometer-sized spatial resolution. A Chinese
research group published ( C24 ) a report of a tabletop spectrom-
eter device using a plane crystal WD method with a position-
sensitive proportional counter and monolithic polycapillary X-ray
focusing lens. The smallest size of the focal spot was 50 ím for
the Cu-K R line, and the energy resolution was found to be 4.4 eV
at the Ti-K R line. The X-ray source of the spectrometer was a
rotating Cu anode with a 0.3 0.3 mm
2
spot in the direction of a
6° takeoff angle. The X-rays emitted by the tube are focused by
the polycapillary optics on the sample surface, and the secondary
X-ray beam emitted by the sample elements impinges onto the
surface of the analyzing crystal that reflects to a position-sensitive
detector, i.e., the wavelength-dependent intensity distribution is
transformed into a position-dependent intensity distribution. The
spectrometer is suitable especially for the determination of rare
earth elements on the basis of their L line intensities.
A modularly designed and compact X-ray tube was developed
by Bjeoumikhov et al. (C25 ), for both XRF and XRD applications,
equipped with a changeable polycapillary lens and filters. The
device can be assigned to a position-sensitive sensor for detection
of the XRD pattern, a semiconductor ED detector for acquisition
of secondary X-rays, and a CCD camera for supervising optically
the sample analysis. The tube is a metal-ceramic designed item
with optional Cr, Co, Cu, Mo, Ag, Pd, or W anodes, and it can
work in a wide voltage interval of 0 -55 keV with 0 -1-mA current,
30-W maximum anode load, and 50-ím anode spot diameter. Low-
power X-ray tubes have a decreased area of the anode spot that
drastically decreases the brightness of the emitted X-rays, which
effect can be compensated by the use of polycapillary lenses to
focus the primary beam to the irradiated spot on the sample
surface. The authors demonstrated the spectrographical capabili-
ties of this new device with XRD and XRF experiments. Longoni
et al. (C26 ) described a newly designed compact ED X-ray
spectrometer based on a monolithic ring-shaped array SDD
consisting of 12 elements with a hole in its center, a polycapillary
lens for focusing the primary X-ray beam emitted by a microfocus
X-ray generator, andx-y sample moving setup. The maximum
high voltage of the W-anode X-ray tube was 50 kV with 500- íA
current, and the anode spot was 50ím in diameter. The gain
factor of the polycapillary lens was 2500 at 10-keV energy, and
the diameter of the focal spot of the primary X-rays was between
45 and 70 ím at 25-keV X-ray energy. The spectrometer was
principally developed for elementary mapping analysis of archaeo-
logical and biological samples with 100 -250- ím pixel sizes on
several 10- ím to 1-mm areas. A portable XRF spectrometer report
was published as well by Zarkadas and Karydas ( C27 ) using an
end-window X-ray tube operated by a battery having voltage of
40 kV and 30íA with a Au anode and a Peltier-cooled Si-PIN
detector. The minimum detection levels for Ca, Ti, Mn, and Cu
were found to be 548, 419, 219, and 248 pg, respectively. The
spectrometer was principally developed for the analysis of
archaeological samples, and the article reported on an application
identifying the soldering technique that was applied in a pair of
Hellenistic Au earrings. A novel electron-impacted X-ray source
was reviewed by Hemberg and his research group (C28 ); it uses
a high-speed liquid metal jet anode based on Sn63 Pb37 solder
alloy as anode material in melted form. The concept of electron
impact type X-ray sources has not changed principally since the
discovery of X-rays. The main technical problem is that 99% of
the impacted energy into the anode is lost by heating the material.
The line focusing and rotating anode constructions were the only
fundamental techniques developed to increase the fraction of the
impacted energy for generating X-rays through the improvement
of the heat capacity of the anode. The authors demonstrated that
their new design for a compact X-ray source provides a 100 times
higher brightness (in photons mm
-2
sr
-1
s
-1
) in which the electron
beam, focused on the liquid jet of solder material (its temperature
was 183 ° C, the diameter of the alloy beam was 100í m) was
pumped into vacuum with a speed of 50 m/s. On the basis of the
experiments of the authors, the brightness of the first version
device was estimated at 10
13
photons mm
-2
mrad
-2
s
-1
for the
Sn-KR line. That value is remarkable in comparison with the
brightness of bending magnet radiation (10
15
photons mm
-2
mrad
– 2
s
– 1
per 0.1% bandwidth). The new construction of this
compact X-ray source is potentially a candidate in medium-scale
laboratory applications as in medical utilizations, protein crystal-
lography, phase imaging, and any other use where a high-flux
X-ray beam is a basic requirement.
Finally, in this section, we cite a review paper published by
David et al. (C29 ) about the technological and measuring aspects
of the applications of refractive and diffractive optical elements
for X-ray microanalysis. The researchers systematically character-
ized the main group of X-ray lenses: Fresnel zone plates for soft
X-ray region, linear Si zone plates for hard X-rays, Si planar
refractive lenses for hard X-rays, and diamond planar refractive
lenses for fourth-generation X-ray sources. These optical devices
cover the photon energy range from 250 eV up to 50 keV with
26-65% transmission efficiency and with 100-nm focal spot for
Fresnel zone plates. The authors outlined the extremely advanta-
geous spectral properties of the diamond refractive lenses, such
as high thermal conductivity, low thermal expansion coefficient,
and high thermal stability, predestinating them as the most
suitable optical devices for free electron laser X-ray sources; these
future X-ray sources will provide ultrashort X-ray pulses with very
high brilliance up to 10 orders of magnitude above today’s third-
generation synchrotron facilities.
QUANTIFICATION AND FUNDAMENTAL DATA
The state of the art of the quantification models for various
XRF applications offering standardless determination of elemen-
tary composition or other physical parameters of the analyzed
materials and the fundamental atomic data used by quantification
procedures are always of great importance. Therefore, in this field
of XRS, several publications were issued in the literature during
the review period. We intend to select some of them to character-
ize the main trends in this area. An Argentinean research group
(D1) determined the transition rates for radiative decays to the L
shell for Yb, Hf, and Ta by means of measurements of fluorescence
lines and accurate fitting of the obtained EDXRF spectra generated
with SR. To increase the accuracy of the calculated rates, they
considered the continuum and characteristic radiation artifacts
in the spectra as well and included them in the spectra fitting
procedure. Their results for transition rates were in good agree-
ment with data published in the literature.
The value of the mass absorption cross sections forms the most
substantial data set for all theoretical or semiempirical calculations
for the determination of characteristic material properties by using
X-rays. Therefore, research efforts that intend to determine
experimentally or calculate theoretically these fundamental pa-
rameter functions via X-ray energy and atomic numbers are of
great importance. Chitralekha et al. published ( D2) their results
about the experimental determination of mass attenuation coef-
ficients on mono- and disaccharides at photon energies 5.947,
6.460, and 14.413 keV, using
55
Fe and
57
Co point sources with 0.74
and 0.37 MBq activity, respectively. To test their experimental
setup, the absorption coefficients of some pure metals were
measured and compared to the theoretical values; the agreement
was within 1%. Instead of interpolating tabulated theoretical data,
the authors used for calculation the exact theoretical values of
attenuation WinXCom software. Many scientific, engineering, and
medical applications require absorption data of X-ray in a wide
energy range that cannot be satisfied by a conventionally tabulated
discrete data set; however, the prompt and unlimited calculation
of this fundamental parameter is necessary in the continuous
energy range from 1 keV to 1 MeV for pure elements and
compounds. For this reason, the XCOM and WinXCom software
were developed and its main features published by Gerward and
co-workers (D3). The software provides the total cross sections,
attenuation coefficients, cross sections for incoherent and coherent
scattering effects, photoelectron absorption, and pair production
as well. The huge data set calculated by WinXCom is available in
graphical form in logarithmic scale, and it offers a possibility for
exportation of data to Excel for further evaluation processes. A
surprising result was published by Kulshrehth et al. (D4)onthe
chemical effect for the K R/Kâ intensity ratio of different Ag
compounds (Ag
2CO3,Ag
2SO
4, AgNO
3
, AgCl, AgBr, Ag) using 59.6-
keV ç-rays from an annual 100-mCi
241
Am source. The applied
Si(Li) detector had 30-mm
2
active area and 170-eV energy
resolution at 5.9-keV energy. They found the K R/Kâ intensity ratio
for pure Ag metal to be 0.206 ( 0.003 and observed a wide
variation for the Ag compounds, between 0.190 and 0.207. The
authors explain their result as the consequence of the difference
in the electron density in the K shell affected by the change in
the distribution of the valence electrons, due to the different
chemical status in the investigated compounds. The cross-sectional
data for K and L shell ionization have great importance for
developing more reliable theoretical models for describing the
fundamental inner shell ionization processes. For the K shells,
the available data sets of fluorescence yield and ionization cross
sections are very accurate, but for L shells, the fundamental data
are very incomplete due to the experimental difficulties and the
existence of the subshells. The vacancies on the L shells can be
redistributed among the three subsells by the fast Coster-Kronig
transition process. Mandal et al. published (D5) their results on
the experimental determination of L X-ray fluorescence cross
sections of elements with atomic numbers in the range of 62-70
at 17.8-, 22.6-, and 25.8-keV X-ray excitation energy, using Mo,
Ag, and Sn anode tubes. The authors compared the obtained
fundamental cross-section values to theoretically calculated data
sets of fluorescence yields and Coster -Kronig transition prob-
abilities published by other researcher, and they found a reason-
able agreement. Similar experimentally determined data were
published by Bonzi and Barrea ( D6) for L X-ray cross sections in
the Z-range between 45 and 50 for a 7-keV synchrotron generated
linearly polarized monoenergetic X-ray beam. The authors deter-
mined Lé,LR,Lâ1,Lâ2,Lç1, and L ç2
lines and compared with well-
known tabulated theoretical data from the literature and found
the experimentally determined values to be higher by 7-10%; in
some particular cases, this difference was up to 40%. A standard-
less analytical procedure was developed by Sitko and Zawisza ( D7)
for WDXRF using a scintillation detector or gas proportional
counter based on the fundamental parameter method (FPM). The
basic calculations of the analyte concentrations are based on the
mathematical description of Sherman et al.; that theory also takes
into account the secondary enhancement effects causing the
excitation of analyte characteristic lines by the radiation of other
elements in the sample. The authors developed a calculation
procedure for the determination of the energy-dependent ef-
ficiency of the applied detectors on the basis of measured intensity
of the sample elements. The greatest problem of the FPM
application in WDXRF is that the efficiency of the reflection
crystals is unknown; therefore, these functions must be deter-
mined experimentally using standard samples. For calibration of
the whole WDXRF device, synthetic samples were used, in pellet
form, of K
2CO3, CaCO
3, TiO
2
,Cr 2O3
,Co 2O3
,Ga 2O3, SeO
2,Y2O3
,
ZrO 2, AgNO
3, and CdCO
3
mixed homogeneously with boric acid.
Karydas published ( D8) a remarkable study and calculations on
the self-element secondary enhancement effect; that means the
K and L lines of an element may excite its own L or M lines,
respectively. His results show that the self-element excitation
fluorescence contribution in the case of a monochromatic X-ray
beam and pure element target must be taken into consideration.
The author proposed to estimate the self-element secondary effect
by measuring the ratio of the corresponding fluorescence intensi-
ties emitted by a thin and an infinitely thick target. On the basis
of the published results, it can be predicted that this correction
calculation will improve the accuracy of quantitative XRF analysis
by means of FPM.
The coherently and incoherently scattered X-rays may have
increased the first-order excited X-ray fluorescence intensity
through a second-order enhancement effect, namely, when the
X-ray that is scattered in a sample bulk excites sample elements
before exiting the sample body. Huan et al. published (D9)
theoretical calculations and experimental results of this secondary
excitation in the direction of the detector, considering in their
calculations both the coherently and incoherently scattered X-rays
in a light matrix. The controlling experiments were carried out
on synthesized fused samples consisting of Li
2B4O7, LiBO
2, SiO
2
,
CaO, V2O5
,Fe 2O3, NiO, ZnO, WO
3
, and PbO. They found that
the scattering contribution increases the fluorescence intensi-
ties, mostly for high-Z analytes and low concentrations of fluo-
rescencing. The author proposed to introduce this correction
method into the FPM calculations. The research group of Alvarez
(D10 ) also published on the role of X-ray scattering in quantitative
XRF analysis. The coherently and incoherently scattered radiation
is used in several FPM-based algorithms. The author performed
Monte Carlo simulation calculations to trace the radiation transport
in the sample, collimator, and detector to estimate the scattering
contribution for Rayleigh and Compton peaks in the case of
annular radioisotope excitation. The results of the simulations were
applied in the IAEA-QXAS software using the standardless FPM
approach. In the case of polychromatic X-ray fluorescence excita-
tion and using the FPM approach, exact knowledge of the energy
distribution of the excitation X-ray beam is required. Applying low-
power X-ray tubes, the local primary flux can be increased
significantly by application of polycapillary lenses. However, in
this case, direct measurement of spectral distribution of the
excitation beam is impossible due to the very high output flux
(10
9
photons s
-1
cm
-2
) of the capillary lens. Padilla et al.
published (D11) their results for solving this measurement
problem by detecting the scattered radiation from a thin layer
instead of measuring the direct X-ray beam emerging from the
polycapillary lens. The authors calculate the coherently and
incoherently scattered intensities theoretically and simulated the
scattered spectra. These measurements were the basis of their
semiempirical approach to find the primary energy distribution
of a microfocus tube. The FPM can be used for determining the
thicknesses of single and multiple layers, as Nygård et al. (D12)
have shown theoretically and experimentally. The authors used
a white beam from a Mo X-ray tube for excitation of Au-Ni-Cu
multilayer elements detecting the secondary X-ray photons by a
Peltier-cooled SDD. The experimentally determined thicknesses
concerning single layers showed an excellent agreement with the
nominal values; however, the agreement in case of multiple layers
was poorer due to the overestimation of the theoretical intensity
of the secondary fluorescence in multiple layers. The authors
recommend carefulness when the FPM is applied for quantitative
thickness determinations. Two empirical coefficient models were
proposed by Sitko (D13) for calculating simultaneously the sample
composition and intermediate thickness. The first model can be
used when the sample is deposited on a infinite thickness
substrate or without substrate. However, in this case, the mass
per unit area has to be known. The second model estimates the
sample thickness on the basis of the scattered X-ray intensity.
Both models require calculating the empirical coefficients by
measurements of multielement standard samples whose thick-
nesses are known, and the concentration ranges of the set of these
reference samples must cover the concentration ranges of each
element in the unknown sample. The author rigorously tested
experimentally these proposed models by WDXRF analysis of
synthetic samples made of Ga
2O3
,As
2O3, SeO
2, ZrO
2, and SrCO
3
mixed with boric acid, in the form of 4-cm-diameter pellets, and
he used, for the R coefficient determination, granite, black shale,
greisen, feldspathic sand, and basalt standards.
The secondary target arrangement is useful for enhancement
of the selected elementary excitation and to obtain “clear”
monoenergetic spectra for sample excitation; however, the FPM
calculation needs the development of an appropriate theory and
algorithm. Zarkadas and Karydas developed this special model
(D14) in order to examine the accuracy of this FPM algorithm
with comparison of theoretical and experimentally determined
results. For the experiments, they used a 3-kW X-ray generator
with a four-window Mo anode tube having a fine focus of 12
0.4 mm
2
and a point focus of 1.2 0.4 mm
2
, a Be window with a
thickness of 400 ím, and a Si(Li) detector with 30-mm
2
active
area. The materials for the secondary targets were Y
2O3
, Mo,
Sb
2O3, and BaCO
3
. The authors concluded from their work that
this algorithm can provide accurate information on various
combinations of secondary target-based X-ray spectrometers and
helps in the optimization process of the geometrical parameters.
One of the most problematic steps in the application of
calculation models based on the FPM is to estimate the dark
matrix composition taking into consideration the absorption of
this undetectable part of the sample. Because the dark matrix
consists of mainly of H, C, O, and N, Bamford and his research
group (D15) used Rutherford backscattering (RBS) for determin-
ing the low-Z element matrix in biological samples. They found,
on the basis of RBS measurements, that the matrixes can be
described very similarly by the stoichiometry of C
7 H 10O 5
.
Therefore, this “virtual” composition can be used as input
parameter for the dark matrix composition for the QXAS software,
calculating the quantitative composition of the samples for other
elements having a higherZ. The authors tested their method by
analysis of the IAEA-336 standard and found a good agreement
for most of the elements of the standard. They deducted from
this result that this composition of dark matrix, determined by
RBS analysis, can be applied as general input data for FPM
calculations of biological samples without repetition of RBS
measurements.
Bottigli et al. (D16) published a Monte Carlo algorithm for
the simulation of X-ray images or spectroscopic experiments
carried out on heterogeneous samples used SR or X-ray tubes.
The novel method is based on a 3D regular grid division of the
sample body, and it is capable of determining the X-ray fluores-
cence signal emitted each voxel. The core of the simulation code
considers the basic interactions between X-ray photons and atoms
of the sample material, such as photoelectric absorption, fluores-
cent emission, and elastic and inelastic scattering processes. The
authors pointed out that the voxel-based simulation mode is much
faster and less expensive in terms of computational cost. The
typical calculation time can be characterized by the following
data: 300 300 95 sets of voxels, each with 2 2 2 ím
3
volume were used for simulation, the input data were obtained
from a 512 512 pixel array detector, the CPU time was for one
projection 3 s using a Pentium IV processor with 1800 MHz.
A special geometrical arrangement for XRF analysis requires
the optimization of the performance of the measuring set up as
published by Rao et al. (D17), who estimated theoretically the
geometrical efficiency and solid angle for detection, the Compton
scattered contribution to the detected spectra, and the monochro-
macy in a triaxial system equipped with X-ray tube.