25 Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants

Full text

443
25
Analytical Tools for Exploring
Metal Accumulation and
Tolerance in Plants
Katarina Vogel-Mikuš, Iztok Arčon, Peter Kump,
Primož Pelicon, Marijan Nečemer, PrimožVavpetič,
Špela Koren, and Marjana Regvar
CONTENTS
25.1 Introduction ..........................................................................................................................444
25.2 Tools for Determination of Metal Uptake in Plant Tissues ..................................................447
25.2.1 X-ray Fluorescence ...................................................................................................447
25.2.1.1 Basic Principles ..........................................................................................447
25.2.1.2 X-ray Fluorescence Process .......................................................................448
25.2.2 Standard Energy-Dispersive X-ray Fluorescence Analysis ...................................... 451
25.2.2.1 XRF Instrumentation ................................................................................. 451
25.2.2.2 Excitation Sources ..................................................................................... 452
25.2.2.3 Detectors .................................................................................................... 453
25.2.2.4 Sampling and Sample Preparation for EDXRF ......................................... 453
25.2.3 Quantitative Elemental Analysis by EDX-ray Fluorescence Spectroscopy ............. 455
25.2.3.1 Basic Principles .......................................................................................... 455
25.2.3.2 Starting the Quantitative XRF-Analysis: Principal Problems and
Necessary Assumptions ............................................................................. 457
25.2.3.3 Calibrations ................................................................................................ 458
25.2.3.4 Detection Limits ........................................................................................ 458
25.2.4 Total Reection X-ray Fluorescence ......................................................................... 459
25.2.4.1 TXRF-Excitation Module .......................................................................... 459
25.2.4.2 Sample Preparation and Quantication .....................................................460
25.3 Exploring Metal-Localization and Distribution in Plant Tissues and Cells .........................460
25.3.1 Micro-Proton-Induced X-ray Emission Spectroscopy .............................................. 461
25.3.1.1 Sample Preparation ....................................................................................465
25.3.1.2 Data Processing and Evaluation ................................................................465
25.3.1.3 Instrumentation and Examples ..................................................................466
25.3.2 Synchrotron Micro-X-ray Fluorescence Spectroscopy .............................................467
25.3.2.1 Basic Principles ..........................................................................................468
25.3.2.2 General X-ray Microprobe Characteristics and Set-up ..............................468
25.3.2.3 Use of SR X-ray Microprobes in Plant Science and Quantitative
Element Distribution Mapping .................................................................. 470
25.4 Exploring Ligand Environment of Metals in Plant Tissues.................................................. 471
25.4.1 X-ray Absorption Spectroscopy ................................................................................ 472
25.4.1.1 Basic Principles .......................................................................................... 472
K13484_C025.indd 443 8/9/2012 8:31:39 PM

444 Phytotechnologies: Remediation of Environmental Contaminants
25.1 INTRODUCTION
The origin of naturally occurring metals is in rocks, soils and sediments, where they are primarily
trapped in stable, insoluble forms. Yet, through natural processes small amounts of metals can be
mobilized and allowed to circulate in water, soil and air, through biogeochemical cycles that keep
distribution of any given metal within an ecosystem at relatively constant concentrations over time.
Their constant cycling is especially important for the biosphere, since certain amounts of metals
are essential for organisms, but can be highly toxic when present in excess. With mining of ores, in
addition to smelting and other purication practices, metals have been rapidly released from their
more stable insoluble forms to less stable, soluble forms and released into the environment. So now
days metal contents released into the environment by anthropogenic activities highly overwhelm
natural biogeochemical cycles (Singh 2005).
Humans have been introducing trace metals into the environment since they rst gained knowl-
edge on their numerous useful properties. Recent archaeological ndings have revealed that humans
have been using metals since the Neolithic period. Mining and processing of metals began in around
8,000 BCE, 10,000 years ago. The rst metals that were used by humans were copper and gold.
This is understandable, as both are easy to see and easy to process using simple tools. Since then,
humans have used various means to extract these and also other metal compounds from the earth’s
crust. Although the use of metals has brought numerous benets to human society, we have had to
embrace the harsh consequences of metal pollution (Singh 2005). Early on, metal pollution only
affected human populations in the nearest vicinity of the source of metal mining and smelting
activities; at the turn of 19th century, with the start of Industrial Revolution, however, metal pollut-
ants were distributed over wide areas by means of air and water causing visible detrimental effects
to numerous ecosystems (Singh 2005).
Increased metal concentrations present in the environment pose threat to the living organisms
from microorganisms and plants to animals and humans, because they interfere with vital bio-
logical processes, including photosynthesis and respiration. By replacement of essential minerals in
structural and functional organic molecules, especially enzymes, binding to free thiol and amino
groups of proteins (especially Cd, Hg, Pb, Zn have high afnity for binding to free thiol or amino
groups), or induction of free radical formation (Cu, Fe) (Figure 25.1), they cause malfunctioning of
important biological molecules and oxidative stress.
Due to pollution and degradation of ecosystems worldwide, there was a growing need to develop
powerful analytical tools for monitoring trace element concentrations in the biosphere and its abi-
otic environment with the common goal to be able to assess metal bioavailability, toxicity and risk
relationships, and to revert to the environmentally friendly technologies, namely phytoremedia-
tion, which use suitable plant species for cleaning and remediation of metal polluted and degraded
ecosystems (Thompson-Eagle and Frankenberger 1990; Wenzel et al. 1999; Lombi et al. 2000;
Mulligan et al. 2001; Barceló and Poschenrieder 2003; Kidd et al. 2009).
Plants have been related to human exploitation connected with metal mining already from
the 5th century , when a connection between vegetation and the minerals located underground
was rst noticed in China. There were particular plants that thrived on and indicated areas rich
in copper, nickel, zinc, and possibly gold, though the latter has never been conrmed. Also in the
25.4.1.2 Extended X-ray Absorption Fine Structure ............................................... 474
25.4.1.3 X-ray Absorption Near Edge Structure ..................................................... 479
25.4.1.4 Systematic Errors in XANES Analysis .....................................................483
25.4.1.5 X-ray Absorption Spectroscopy Experiments ...........................................485
25.4.1.6 EXAFS and XANES in Practice ...............................................................486
25.5 Conclusions ...........................................................................................................................488
Acknowledgments ..........................................................................................................................489
References ......................................................................................................................................489
K13484_C025.indd 444 8/9/2012 8:31:39 PM

445Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
West the early starts of geobotany, which depends solely on visual observation of the vegetation
cover, dates back at least to Roman times (Brooks 1998). However the famous mediaeval metal-
lurgist and miner Georgius Agricola who published his celebrated De Re Metallica (Of Metallic
Matters) in 1556, describing vegetation cover that can be found over concealed ore bodies could
be acknowledged as the world’s rst practical geobotanist. In the same century Thalius (1588)
determined Minuartia verna (Figure 25.2) as an ore indicator in Harz Mountains in Germany
(adapted after Brooks 1998).
In contrast to geobotany, biogeochemistry that depends on advances in analytical chemistry
before it could be developed, dates back only to 1938 when the Soviet scientist S.M. Tkalich observed
that iron content of tundra plants reected the concentration of this element in the soil (adapted after
Brooks 1998). After the Second World War, when rapid analytical and data handling techniques
became available, geochemical and biogeochemical investigations were undertaken in many parts
of the world mainly for mineral exploration (Cole 1965; Cole and Smith 1984). These have identied
an increasing number of mineral indicator and accumulator plant species, and nd out that some
plants, so called metallophytes, are endemic to particular naturally metal enriched regions. The
examples of communities of indicator plants occupying naturally toxic ground over copper, lead-
zinc and nickel deposits in Africa, Australia, Europe and the USA are in detail described by Cole
and Smith (1984), and Brooks (1998). Besides revealing mineralization in sub-surface bedrock,
including economic ore-bodies, indicator plants and anomalous vegetation communities can also be
found in areas contaminated by man’s activities.
Soon it has become clear that plants occurring on sites with elevated metal concentra-
tions present in soils have evolved peculiar physiological mechanisms which enable them to
Membrane metal
transport proteins
Cation
transporters
Cell wall
Plasma membrane
Metallo-enzymes, other functional
and structural molecules
Cytosol
Essential metals
Nonessential metals
1
Oxidative stress
R•, -OH•, O
2
–
SH
|
|
-NH
S
-N
Blocking functional
groups
2
3
FIGURE 25.1 Schematic presentation of the main mechanisms of metal toxicity in plant cells. Non-essential
metals as well as essential metals when present in excess can: 1) Replace essential metals in metallo-enzymes
and other structural and functional molecules; 2) Interact with functional groups (-SH, -NH, -OH, -COOH…)
of proteins and enzymes and block them; 3) Cause oxidative stress in the cells by the formation of free
radicals.
K13484_C025.indd 445 8/9/2012 8:31:39 PM

446 Phytotechnologies: Remediation of Environmental Contaminants
survive in such hostile environments. The majority of plant species growing at the sites with
elevated metal concentrations exclude metals from their tissues or retain metals in their roots
(so called excluders) while a small proportion of plants known as (hyper)accumulators are
known to (hyper)accumulate extremely high concentrations of different metal in their shoots
(Baker 1981). Brooks and his colleagues rst coined the term “hyperaccumulation” in 1977,
however, the pioneer scientist dealing with hyperaccumulating plants was A. Bauman, who
already in 1885 reported over 1% of zinc in dry shoots of Viola calaminaria and Thlaspi cala-
minare. Later, in the 1930s, O. A. Beath and his coworkers discovered hyperaccumulation of
selenium in Astragalus plants from western USA, although horse hoofs loosing disease con-
nected with feeding on poisonous plants (selenium accumulators) was already described in
the 13th century by Marco Pollo. Finally, credit must be given to Italian scientists O. Vergano
Gambi from University of Florence, and C. Minguzzi who in 1948 discovered the unusual
hyperaccumulation of nickel by the Tuscan serpentine plant Alyssum bertolonii. Nevertheless
few decades has to pass before the applicability of metal hyperaccumulators was introduced to
the “green” movements in late 1980s and 1990s of the past century (adapted after Brooks 1998).
Since then, the outstanding ability of particular plant species to (hyper)accumulate and toler-
ate extremely high concentration of metals in their shoots without being affected by the metal
toxicity has drawn attention of many plant physiologists, and in the past 15 years the scientists
have been challenged to reveal underlying physiological mechanisms of metal (hyper)accumu-
lation and tolerance from organismal down to tissue, cellular and molecular levels. In depth,
understanding of the uptake, transport, localization and speciation of metals in plants is critical
for understanding of metal metabolic pathways within plants. This can be provided by the use of
complementary analytical techniques spanning from “bulk” analyses of metal concentrations
in plant organs to those enabling imaging of metal distribution in plant tissues and cells, and
techniques enabling determination of speciation and local ligand environment. The techniques
addressed in this chapter therefore include X-ray uorescence and X-ray absorption techniques:
energy dispersive X-ray uorescence spectrometry (EDXRF), micro-proton induced X-ray
emission (PIXE), synchrotron micro-X-ray uorescence spectroscopy (SR-micro-XRF) and
X-ray absorption spectroscopy (extended X-ray absorption ne structure [EXAFS], and X-ray
absorption near-edge structure [XANES]) using synchrotron light.
FIGURE 25.2 (See color insert.) Minuartia verna, a Pb and Zn ore indicator from the river bed Gailitz,
Arnoldstein, Austria.
K13484_C025.indd 446 8/9/2012 8:31:40 PM

447Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
25.2 TOOLS FOR DETERMINATION OF METAL UPTAKE IN PLANT TISSUES
25.2.1 X-ray Fluorescence
Determination of metal concentrations in plant organs (roots, stems, leaves, owers, seeds) followed
by assigning metal uptake strategies (exclusion, accumulation) is a prerequisite to further study metal
uptake and tolerance mechanisms at tissue, cellular and molecular levels. In view of the growing needs
of global environmental protection and also to minimize the relevant research costs, it is important
that the analytical procedures for determination of “bulk” elemental concentrations in soil, water and
biological materials are accurate, reliable and reproducible. In addition, analytical techniques have to
be accessible (cheap), with simple sample preparation procedures, enabling analysis of many replicates
in reasonable time, therefore providing more accurate estimation of the processes going on in soil-
plant systems. The rst part of this chapter focuses on the main characteristics and sample prepara-
tion protocols of X-ray uorescence-based analytical techniques for “bulk” sample analyses, namely
energy dispersive X-ray uorescence spectrometry (EDXRF) and total reection X-ray uorescence
spectrometry (TXRF). At present, EDXRF and TXRF may be far less frequently applied for analy-
ses of element concentrations in soil, water, air and biological materials than, for example, atomic
absorption spectroscopy (AAS) and/or inductively coupled plasma atomic-emission spectroscopy
(ICP-AES). However, with the rapid development of detecting and signal processing systems they can
offer low cost, fast, sensitive and accurate element analysis which is particularly advantageous from
the economic and environmental protection points of view (Nečemer et al. 2008, 2011).
Besides standard energy dispersive X-ray uorescence analysis (EDXRF) enabling “bulk” elemental
analyses in different materials, the process of X-ray uorescence also represents fundaments of highly
sophisticated particle induced (micro-PIXE) and synchrotron-based micro-X-ray uorescence (micro-
XRF) spectroscopy techniques that enable visualization of spatial localization and distribution of par-
ticular elements in biological samples (Vogel-Mikuš et al. 2007, 2008a, b, 2010; Kaulich et al. 2009).
Therefore an overview of basic principles of X-ray uorescence process and spectroscopy is provided
in order to better understand the working principles of analytical tools described in the present chapter.
25.2.1.1 Basic Principles
X-ray Radiation—X-rays are electromagnetic waves with a spectrum wavelengths spanning from
about 80 nm (about 15 eV) down to about 0.001 nm (about 1.2 MeV), overlapping to some extent the
region of γ-rays. Electromagnetic radiation (usually above 1 MeV) generated by nuclear processes
is called γ-radiation while the radiation below 80 nm wavelength generated by electrons slowed
down in the outer eld of an atomic nucleus or by transitions between bound states of electrons in
the electronic shells of an atom is called X-radiation.
The history of X-ray uorescence (XRF) dates back to the accidental discovery of X-rays in 1895
by the German physicist Wilhelm Conrad Roentgen. While studying cathode rays in a high-voltage,
gaseous-discharge tube, Roentgen observed that even though the experimental tube was encased in a
black cardboard box, a barium-platino-cyanide screen, which was lying adjacent to the experiment, emit-
ted uorescent light whenever the tube was in operation. This was possible because the energies of X-ray
photons are of the same order of magnitude as the binding energies of inner-shell electrons (K, L, M, N,
…), and therefore they can be used to excite and/ or probe these atomic levels (Janssens 2004).
Interaction of X-rays with Matter -When an X-ray beam passes through matter, some photons are
absorbed inside the material or scattered away from the original direction (Figure 25.3).
The intensity (I
0
) of an X-ray beam passing through a layer of thickness (d) and density (ρ) is
reduced to intensity (I) according to the Lambert-Beer law (Equation 25.1)
I = I
0
e
−μρd
(25.1)
The number of photons per second (the intensity) is reduced, but their energy remains gener-
ally unchanged. The term (μ) is called the mass attenuation coefcient and has the dimensions of
K13484_C025.indd 447 8/9/2012 8:31:40 PM

448 Phytotechnologies: Remediation of Environmental Contaminants
cm
2
g
−1
. The product (μ
L
= μρ) is called linear absorption coefcient and is expressed in cm
−1
. μ(E)
is sometimes also called the total cross-section for X-ray absorption at energy (E).
25.2.1.2 X-ray Fluorescence Process
X-ray uorescence occurs after the excitation (ionization) of atoms in the tightly bound inner K and
L atomic shells with energies that must exceed the binding energies of the K and L electrons through
I
0
I = I
0
e
–µρd
d
Scattered and
fluorescence
X-rays
Ψ
1
Ψ
2
Ω
1
Ω
2
Incident
beam
Detector
FIGURE 25.3 Interactions of X-rays with matter. (Adapted from Nečemer, M., et al., Use X-ray uorescence-
based analytical techniques in phytoremediation. In Handbook of Phytoremediation, (Environmental Science,
Engineering and Technology), 331–358, 2011, New York: Nova Science Publishers.)
X-ray
Photoelectron
Auger electron
Accelerated particle
Photoelectron
Fluorescence X-ray
X-ray
Photoelectron
Fluorescence X-ray
(a) (b)
(c)
FIGURE 25.4 Interaction of an atom with X-rays (a) photoelectric effect; (b) Auger effect; or charged par-
ticles c) electromagnetic Coulomb interaction. (Adapted from Nečemer, M., et al., Use X-ray uorescence-
based analytical techniques in phytoremediation. In Handbook of Phytoremediation, (Environmental Science,
Engineering and Technology), 331–358, 2011, New York: Nova Science Publishers.)
K13484_C025.indd 448 8/9/2012 8:31:40 PM

449Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
the process called the photoelectric effect (Markowicz 1993) (Figure 25.4a). Beside the excitation of
atoms by photoelectric process with X-ray radiation emitted from different X-ray sources (e.g., X-ray
tube, radioisotope source, synchrotron radiation source) the atoms can be also excited by accelerated
charged particles via the Coulomb interaction (electrons, protons, α particles) (Figure 25.4c).
In the photoelectric absorption process, an X-ray photon from the X-ray source is completely
absorbed by an atom and an electron (called a photoelectron, Figure 25.4a) is ejected from one
of the inner (K or L) shells (Figure 25.5). In the excitation process a part of the photon energy is
used to overcome the binding energy (Φ) of the electron, and the rest is left to the electron in the
form of the kinetic energy. After the interaction with an X-ray photon, the atom (in fact the ion) is
left in an excited state with a vacancy created in one of its inner (K or L) shells. This atom almost
immediately returns to a more stable electronic conguration in the process called relaxation or
de-excitation, in which the electrons from outer shells (L, M or N) ll the vacancy created in the K
or L shells (Figure 25.5). Since during relaxation the electrons pass from a higher to a lower energy
state, the difference in energies of respective energy states can be emitted in the form of charac-
teristic X-ray photons (K, L, or M) (Figure 25.4a; Figure 25.5), or this energy can be absorbed by
an weakly bound electron in one of the outer shells (L, M, or N) (Figure 25.4b; Figure 25.5). This
electron is then ejected from the atom as an Auger electron (Figure 25.4b). Characteristic K, L and
M X-ray photons are, according to their energy, typical for a particular element, and this is the
basis for elemental characterization by X-ray spectrometry (Markowicz 1993). The ratio between
thenumber of emitted characteristic X-rays and the total number of inner shell vacancies in a par-
ticular atomic shell that gave rise to them, is called the uorescence yield of that shell (e.g., ω
K
). For
light elements (Z < 20) predominately Auger electrons are produced during relaxation after K-shell
ionization (ω
K
< 0.2) (Figure 25.4b), while medium and high Z elements preferentially relax in a
radiative manner (0.2 < ω
K
< 1.0) (Figure 25.4a).
When atoms are excited by accelerated charged particles (Figure 25.4c), these charged particles
tear off the electrons from the inner atom shells through the electromagnetic Coulomb interaction,
resulting, similarly to the photoelectric effect, in the electron vacancies in particular inner electron
shells. The vacancies are then lled with electrons from outer shells and the difference in kinetic
energy is emitted in a form of characteristic K, L or M lines.
N-shell
M-shell
L-shell
K-shell
K
α
L
α
K
β
L
β
Nucleus
FIGURE 25.5 Electronic transitions in an excited rubidium atom, where one of the electrons in the K shell
is missing.
K13484_C025.indd 449 8/9/2012 8:31:41 PM

450 Phytotechnologies: Remediation of Environmental Contaminants
Photoelectric absorption can only occur if the energy of the photon (E) is equal or higher than
the binding energy (Φ) of an electron. For example, the X-ray photon with the energy of 22.16 keV
(Ag-Kα from Cd-109 radioisotopic source) can eject a K-electron (Φ
K
= 12.66 keV) or an L
3
-electron
(Φ
L3
= 1.433 keV) from the selenium (Se) atom, but the 5.9 keV photon (Mn-Kα from the Fe-55
radioisotope source) can only eject an electron from the L-shell of the Se atom. Since photoelectric
absorption can occur at each of the (excitable) energy levels of the atom, the total photoelectric
cross-section σ
i
is the sum of the (sub)shell-specic contributions (Equation 25.2).
σ σ σ σ
σ σ σ σ σ
i i K i L i M
i K i L i L i L
= + + +
= + + + +
, , ,
, , , ,
( ) (

1 2 3 ii M i M i M, , ,
)
1 2 5
+ + + +σ σ 
(25.2)
In the case for example of Mo (Figure 25.6) at high energy, e.g., >50 keV, the probability of ejecting a
K-electron is rather low and that of ejecting an L
3
-electron is even lower. As the energy of the X-ray pho-
ton decreases, the cross-section increases, i.e., more vacancies are created. At a binding energy Φ
KMo
=
19.99 keV, there is an abrupt decrease in the cross-section, because X-rays with lower energy can no lon-
ger eject electrons from the K-shell. However, these photons continue to interact with the (more weakly
bound) electrons in the L- and M-shells. The discontinuities of the photoelectric cross-section are called
absorption edges. The ratio of the cross-section just above and just below the absorption edge is called
the jump ratio (r). As XRF is the result of selective absorption of radiation, followed by spontaneous
emission, an efcient absorption process is required. An element can therefore be determined with high
sensitivity by means of XRF when the exciting radiation has its maximum intensity at energy just above
the K or L-edge (for heavier elements) of that element (Janssens 2004; Nečemer et al. 2011).
Soon after the discovery of X-rays, in 1913, Henry Moseley established the relation between the
atomic number (Z) and the specic X-ray wavelength (λ) of an element (Equation 25.3),
1/λ = K(Z − s)
2
(25.3)
where, K and s are constants; s is the shielding constant and takes value close to one, and K has
a different value for each of the line series considered (e.g., the K
α
-lines, the L
α
-lines,…) (Figure
25.5). Each unique atom has a number of available electrons that can take part in the energy transfer
and since millions of atoms are typically involved in the excitation of a given specimen, all possible
de-excitation routes are taken.
10000
Mo L-edge
σ
M
+ σ
L
Mo K-edge
σ
M
+ σ
L
+ σ
K
1000
100
10
1
0.1
1 10
Photon energy (keV)
Cross-section (cm
2
g
–1
)
100
FIGURE 25.6 Variation of photon cross-section (σ in Mo) as a function of X-ray photon energy. The K, L1, L2
and L3 absorption edges are clearly visible. (Adapted from Janssens, K., X-ray based methods of analysis. In Non-
destructive microanalysis of cultural heritage materials, 129–226, 2004, Amsterdam: Elsevier. Nečemer, M.,
etal., Use X-ray uorescence-based analytical techniques in phytoremediation. In Handbook of Phytoremediation,
(Environmental Science, Engineering and Technology), 331–358, 2011, New York: Nova Science Publishers.)
K13484_C025.indd 450 8/9/2012 8:31:41 PM

451Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
25.2.2 standard energy-dispersive X-ray Fluorescence analysis
After discoveries of Charles G. Barkla who in 1909 found a connection between X-rays radiating
from a sample and the atomic weight of the sample and already mentioned observations of Moseley
in 1913, which helped to count the elements with the use of X-rays, the potential of the X-ray
uorescence technique was quickly realized, with half of the Nobel Prizes in Physics given to the
development in X-rays physics from 1914 to 1924. Originally X-ray spectroscopy used accelerated
electrons as an excitation source, but the requirements such as a high vacuum, electrically conduct-
ing specimens, and volatility of the sample posed major roadblocks. To overcome these problems an
X-ray source was used to promote the uorescent emission in the sample. Excitation of the sample
by this method introduced roadblocks of its own, by lowering the efciency of photon excitation and
requiring instrumentation with complex detection components. Despite these disadvantages, the
uorescent emission of X-rays would provide the most powerful tool for the analyst using commer-
cial instruments. From the 1950s to 1960s nearly all the X-ray spectrometers were wavelength dis-
persive spectrometers. In a wavelength dispersive spectrometer, a crystal separates the wavelengths
of the uorescence from the sample, similar to grating spectrometers for visible light. Although the
earliest commercial XRF devices used simple air path conditions, machines were soon developed
utilizing helium or vacuum paths, permitting the detection of lighter elements. In the 1960’s, XRF
devices began to use lithium uoride crystals for diffraction and chromium or rhodium target X-ray
tubes to excite longer wavelengths. This development was quickly followed by that of multichannel
spectrometers for the simultaneous energy measurement of many elements (Beckhoff et al. 2006).
The other X-ray uorescence spectrometer available at that time was the electron microprobe,
which uses a focused electron beam to excite X-rays in a solid sample as small as 10
−12
cm
3
with an
energy dispersive proportional detector. The rst microprobe was built by R. Castaing in 1951 and
became commercially available in 1958 (Beckhoff et al. 2006).
In the early 1970s, energy dispersive spectrometers became available and used Li-drifted silicon
or germanium detectors. The advantage of these instruments brought the ability to measure the entire
X-ray uorescence spectrum simultaneously. With the help of computers, deconvolution methods
were developed to extract the net intensities of overlapping individual X-rays lines (Beckhoff et al.
2006). An XRF device was even included on the Apollo 15 and 16 missions (LearnXRF 2011) and
Mars Pathnder mission in 1996–1997 (Pantazis et al. 2010). While in 1970s the XRF spectrom-
etry applications in environmental sciences demonstrated some amateurism, by the end of the 20th
century the mentioned applications were the most published X-ray analysis results in the scientic
literature. X-ray uorescence spectrometry proved to be very efcient tool also in life sciences for
analysis of major and minor mineral constituents of organisms with its nondestructiveness and
simple sample preparation that allows fast analysis of large number of samples (Nečemer et al.
2008, 2009, 2011).
25.2.2.1 XRF Instrumentation
Conventional energy-dispersive X-ray uorescence (EDXRF) spectrometers consist of only two
basic units—the excitation source and the spectrometer or detection system (Figure 25.7).
In this case, the resolution of the EDXRF system depends directly on the resolution of the detec-
tor. Typically, a semiconductor detector of high intrinsic resolution is employed [Si(Li)]. The use of
this type of detector allows one to record an electronic signal (voltage pulse) processed by the pre-
amplier and amplier, which is proportional to the energy of the detected photon dissipated within
the sensitive volume of the detector. An analogue to digital converter (ADC) and a multi-channel
analyzer can be then used to sort, integrate, store, and display the detected pulses into an X-ray
spectrum (Figure 25.8). In recently built systems, digital pulse acquiring, shaping and processing
is used instead of analogue, which is much faster with greatly reduced dead times from transient
overload signals (Warburton et al. 2000). Using this conguration, all of the X-rays emitted by the
sample are collected at a very high rate irrespective of their size. In addition, this conguration also
K13484_C025.indd 451 8/9/2012 8:31:41 PM

452 Phytotechnologies: Remediation of Environmental Contaminants
enables high speed acquisition and display of spectral data. EDXRF systems are therefore classied
according to the type of excitation source, the geometry of excitation and the type of energy disper-
sive detector installed (Margui et al. 2009).
25.2.2.2 Excitation Sources
Excitation of the elements in the sample can be performed using almost monochromatic radio-
isotope sources, or partially monochromatic and polychromatic radiation from an X-ray tube in a
secondary target or direct mode of operation. Table 25.1 lists radioisotope sources typically used in
EDXRF analysis.
The most commonly used sources include Fe-55, Co-57, Cd-109, and Am-241. Each of these
emits radiation at specic energy levels and therefore efciently excites elements within a specic
atomic number range. As a result, no single radioisotope source is sufcient for exciting the entire
range of elements of interest in environmental analysis, and many instruments use two or three
sources to maximize element range. The half-life of a source is important, especially for Fe-55,
Co-57, and Cd-109 sources. With half-lives as short as 270 days some sources may have to be
replaced after a few years when their intensity decreases to a level too low to provide adequate exci-
tation of the elements of interest (Kalnicky and Singhvi 2001). Alternatively, an X-ray tube can be
used to irradiate the sample with characteristic and continuum X-rays. X-ray tubes can be air cooled
X-ray
source
Detector
Amplifier
ADC
MCA
PC
Preamplifier
Sample
FIGURE 25.7 Diagram of an XRF system composed of X-ray source (X-ray tube or radioisotopic source)
energy dispersive detector, preamplier, amplier, multi-channel analyzer and PC.
0
0
500
1000
1500
2000
2500
3000
100 200 300 400 500 600
Channel
Intensity (total counts)
700 800 900 1000
Zn-Kα
Rb-Kα
Ag-Kα
Scattering
Ca-Kα
Fe-Kα
K-Kα
Zn-Kβ
FIGURE 25.8 X-ray uorescence spectrum of dried leaves of Zn hyperaccumulating pennycress Thlaspi
caerulescens recorded after excitation with Cd-109 radioisotope source.
K13484_C025.indd 452 8/9/2012 8:31:42 PM

453Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
(low power, 3–50 W) or water cooled (high power, 2 kW), with different anodes, such as Ag (22.1
keV), Rh (20.2 keV), Mo (17.4 keV), and Cr (5.4 keV) (Nečemer et al. 2011). XRF measurements
may be performed in different geometries. When thick specimens are excited by the continuous
X-ray spectrum from an X-ray tube, the sensitivity of the method is lowered because of the relatively
high background from the scattered continuous radiation from the sample, or its substrate, in the
spectral region of the uorescent radiation. A secondary target irradiation geometry using different
metal targets as Mo, Rh, Cr, etc. can be used to partly monochromatize the X-ray tube radiation,
and thus to decrease this background scattering (Jaklevic and Giauque 1993; Nečemer et al. 2011).
25.2.2.3 Detectors
The X-ray detector converts the energies of the X-ray uorescence photons into voltage pulses that
can be counted to provide a measurement of the total X-ray ux. X-ray detectors are typically “pro-
portional” devices where the absorbed energy of the incident X-ray photon in a sensitive volume of the
detector determines the size of the output voltage. A polychromatic uorescence beam of radiation
incident upon the detector produces a spectrum, with a pulse height distribution proportional to the
energy distribution of the incident uorescence radiation beam. A multichannel analyser collects the
spectrum and enables discrimination between characteristic uorescence lines of different elements,
depending on the resolution of the detector (Kalnicky and Singhvi 2001). The three most common
types of detectors are: the gas ow proportional detector, the scintillation detector, and semiconductor
detectors (Si(Li), SiPIN, SiDrift and Hyperpure Ge detectors). These detectors differ in resolution and
intrinsic efciency. Resolution is the ability of the detector to separate X-rays of different energies, and
is important for minimizing spectral interferences and overlapping, while the efciency depends on
the absorption of X-rays in the sensitive volume of the detector. Semiconductor detectors have the best
resolution and are preferred for EDXRF instruments. These detectors may require liquid nitrogen as a
coolant or employ electric cooling via built-in Peltier elements (Nečemer et al. 2011).
The selection of the detector is very important. In cheaper spectrometers, a radioactive source and a
proportional detector (gas) can be used. However, one shortcoming of such a device is the poor energy
resolution of the detector (800–1.000 eV at 5.9 keV), making the quantication of the insufciently
resolved characteristic X-ray lines in the measured spectrum quite difcult. This problem can be over-
come by using semiconductor detectors such as Si(Li), Si PIN or Si drift (SDD) detectors, which have
energy resolutions of around 120 eV to 140 eV at 5.9 keV (Nečemer et al. 2011).
25.2.2.4 Sampling and Sample Preparation for EDXRF
Sampling and sample preparation represent two the most critical steps in the analysis of environmen-
tal samples, regardless of the analytical method applied. To accurately characterize site conditions,
TABLE 25.1
Commonly Used Radioisotope Sources Used in EDXRF Analysis
Isotope Half-Life Useful Radiation Energy (keV) X-rays Excited Efficiency
Fe-55 2.7 years Mn-K X-rays 5.9 Al-Cr
Co-57 270 days Fe-K X-rays 6.4 <Cr
γ-rays
14.4;122;136
Cd-109 1.3 years Ag-K X-rays 21.1 K-Tc
Am-241 470 years Np-L X-rays 14, 21 Sn-Tm
γ-rays
59
Source: Adapted from Nečemer, M., P. Kump, and K. Vogel-Mikuš, Handbook of Phytoremediation, (Environmental
Science, Engineering and Technology), 331–358, New York: Nova Science Publishers, 2011.
K13484_C025.indd 453 8/9/2012 8:31:42 PM

454 Phytotechnologies: Remediation of Environmental Contaminants
the samples collected must be representative of the site or area under investigation (Margui et al.
2009). Representative soil or vegetal sampling ensures that a sample or group of samples accurately
reects the concentration of the contaminant(s) of concern at a given time and location. Analytical
results from representative samples reect the variation in contaminant presence and concentration
range throughout a site. Parameters affecting the variability of the results of representative sam-
ples include: (1) geologic and plant material variability, (2) contaminant concentration variability,
(3)collection and preparation variability, and (4) analytical variability (Nečemer et al. 2011).
Use of analytical techniques such as AAS or ICP-AES usually requires sample-preparation
procedures involving total destruction of the matrix by chemical treatment. Sample dissolution
is usually a demanding, time-consuming step that sometimes limits application of the analytical
procedures in environmental studies and quality-control processes. Commonly, dry ashing (involv-
ing combustion of the sample) and wet digestion (involving digestion with strong acids) have been
used to destroy the organic matter and dissolve the analytes in such matrices (Margui et al. 2009).
Compared to dry-ashing methods, wet-mineralization procedures using acid digestion present a
wide range of options, depending on the choice of reagents and their mixtures as well as the devices
used for the procedure (Margui et al. 2009). Especially demanding is wet digestion of soil and plant
samples containing silicon, as in such cases, the use of hydrouoric acid is essential to achieve total
destruction of the matrix and thus determination of the total concentrations of analytes. In the last
15 years, classical open systems (digestions at atmospheric pressure) using conventional sources
of heating (e.g., sand baths and hot plates) were gradually replaced by digestion procedures using
microwave ovens and closed vessels, since in open systems there were problems with losses of
volatile elements, such as Hg, Cd and others. In addition, microwave assisted procedures can also
shorten the digestion procedure and reduce the amount of reagents employed, as well as avoid ana-
lyte losses and contamination from other samples or from the surroundings (Margui et al. 2009).
To sum up, the choice of the best decomposition procedure for soil and vegetal samples should be
preceded by verication of the procedure for each specic matrix and analyte under study. This
becomes quite difcult in environmental studies where several or many soil and plant samples are
used as pollution indicators for different metals, and in such cases, the application of techniques that
obviate matrix destruction become even more attractive. For this reason, study of the suitability of
other methods for direct and multi-elemental analysis of soil and vegetal samples has increased in
recent years, from instrumental neutron activation analysis (INAA) to simple X-ray uorescence-
based techniques. INAA is based on measuring the radioactivity produced by neutron reactions on
naturally-occurring nuclides (Nečemer et al. 2008), but the most serious shortcomings of INAA are
its high costs, the availability of a nuclear reactor for irradiation and the rather long time of analy-
sis imposed by the cooling periods for the decay of interfering activated short-lived radionuclides,
although the method is very sensitive and accurate. In addition, INAA does not allow determination
of some environmentally-important elements (e.g., Pb) (Margui et al. 2009). When summing up
the advantages and disadvantages of INAA, the method is not very suitable for routine application
in environmental studies. Most X-ray uorescence (XRF) techniques, on the other hand, comply
with the desirable features for analysis of soil and vegetal specimens, including: i) the possibility
of performing analysis directly on solid samples; ii) simultaneous multi-element capability; iii) the
possibility of performing qualitative, semi-quantitative and quantitative determinations; iv) a wide
dynamic range; v) high throughput, and vi) low cost per determination (Nečemer et al. 2011).
The main drawbacks of XRF instrumentation restricting its more frequent use for environmental
purposes have been its limited sensitivity for some important pollutant elements (e.g., Cd, Pb and
Hg) and a somewhat poorer precision and accuracy compared to atomic spectroscopic techniques
(Nečemer et al. 2008, 2011). The main source of uncertainty in quantitative XRF analysis origi-
nates from the inhomogeneity of the sample in a rather small surface part of the sample (due to the
absorption of excitation and uorescence X-rays; it is usually just few mg cm
−2
), which contrib-
utes to the measured uorescence intensities. Nevertheless, there have recently been improvements
in XRF instrumentation (e.g., development of spectrometers using digital signal processing and
K13484_C025.indd 454 8/9/2012 8:31:42 PM

455Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
enhancement of X-ray production with better designs for excitation-detection geometry), which
have added the advantage of increased instrumental sensitivity, thus allowing improvements in both
precision and productivity. These improvements have therefore increased the possibility of the use
of XRF spectroscopy as a technique in the environmental and life science eld (Margui et al. 2009;
Nečemer et al. 2011).
Both solid and liquid samples can be analysed by EDXRF. In the case of solid samples, no
special chemical treatment of the sample is necessary. Determination of the composition of solid
samples, without any sample preparation, is possible for samples that are homogeneous in all three
dimensions and with a at surface. This is the case for direct analysis of metals and alloys and
that has been one of the main applications of XRF. However, most environmental solid materials
(e.g., soil and biological materials) require sample pre-treatment to make them homogeneous and
to ensure the quality and the reproducibility of measurements (Margui et al. 2009; Nečemer et al.
2011). Commonly, this procedure is based on crushing or grinding the materials into ne powder
followed by pelletization at high pressure, which is the most frequent method of preparing soil and
vegetal samples for analysis by XRF techniques (Margui et al. 2009). Prior to grinding, soil or
vegetal samples are usually oven dried at 60–100ºC, or in the case of vegetal samples freeze dried
to remove their water content. In reducing the soil or vegetal material to a ne powder, grinding
or milling is usually employed with concomitant problems and contamination arising from the
grinding matrix. Particularly for trace elements, precautions should be taken by choosing suitable
materials (e.g., agate, silicon carbide, boron carbide, and tungsten carbide). Agate grinding may
introduce signicant contamination into biological material by Ti, V, Cr, Mn, Fe and Pb (Margui et
al. 2009), but this is the case in all analytical procedures, because for AAS and ICP-AES analysis
the samples should also be well ground and homogenized prior to acid digestion. For EDXRF analy-
sis approximately 100–200 mg of solid, well ground and homogenized (soil or vegetal) material is
sufcient; however any inhomogeneity of the pulverized solid sample can inuence the accuracy of
the measurement, especially in the case of lighter elements (Nečemer et al. 2008). In addition care
should be taken during pellet preparation to assure a uniform thickness in order to avoid bias due
to inhomogeneity.
For liquid samples, 100–1000 ml of solution is required, and the elements are concentrated
fromthe sample by precipitation. Several precipitation agents are available for this purpose. For
instance, the reagent ammonium pyrrolidine dithiocarbamate (APDC) can be used for the precipi-
tation of Cu, Fe, Ni and Pb. Note that any particular precipitating agent can selectively precipitate
only certain specic elements. Therefore, only these specic elements can be determined in liquid
samples using this method of sample preparation. The precipitated elements are separated from the
liquid phase by ltration, and the precipitate that is deposited on the lter is measured directly by
the EDXRF system. Due to the pre-concentration of the precipitated elements, the limits of detec-
tion for these elements decrease to a few 10 μg l
−1
, which cannot be achieved in the analysis of
solid samples. This approach is especially suitable for monitoring contaminating elements in water
(Nečemer et al. 2011).
25.2.3 Quantitative elemental analysis by edX-ray Fluorescence spectroscopy
25.2.3.1 Basic Principles
As previously described, X-ray uorescence spectroscopy is based on X-ray excitation of atoms in
the sample material by the photo-effect process, followed by radiative decay of the excited atoms,
i.e., by emission of characteristic X-rays of the particular atoms. In the relaxation process the radia-
tive and Auger transitions compete and the uorescent yield determines the probability of radia-
tive transition. The uorescent yield favours the radiative decay of heavier atoms. The intensities
of radiative transitions of atoms in the sample, measured by the X-ray spectrometer, are then used
in qualitative and quantitative analyses of the elemental composition of the sample. The measured
intensities of characteristic X-rays depend on the mode of excitation (radioisotope, X-ray tube or
K13484_C025.indd 455 8/9/2012 8:31:42 PM

456 Phytotechnologies: Remediation of Environmental Contaminants
synchrotron excitation), on the fundamental physical constants which determine the probabilities of
photo-effect excitation of atoms, the probability of radiative decay (uorescent yield), the absorption
of excitation radiation and that emitted within the sample while penetrating toward the detector, and
nally on the detector efciency and the geometry of the excitation-detection experiment. The pro-
cess of X-ray uorescence is well understood and presented in many books and papers (Tertian and
Claisse 1982; Rousseau 1984; He and Van Espen 1991; Van Dyck et al. 1986; Nečemer et al. 2011).
In order to better understand quantitative X-ray uorescence analysis basic principles, adapted after
Nečemer et al. (2011), are summarized below.
The relation between the measured characteristic intensities and concentrations of respective
atoms in the sample is established using the above mentioned fundamental parameters and experi-
mental conditions. In the case of monochromatic excitation by energy E
1
in the K shell of atoms, the
respective relation is as follows (Equation 25.4).
I
i
= S
i
c
i
T
i
(c
1
, c
2
,…c
n
)Hi(c
1
, c
2
,…c
n
), (25.4)
Where I
i
: measured uorescent intensity; S
i
: elemental sensitivity (slope of calibration curve in the case of
a thin or diluted sample); T
i
: absorption correction factor (depends on sample composition); H
i
: enhance-
ment correction factor (depends on sample composition) and c
i
: is the concentration of element i;
S
i
= GK
i
(25.5)
G = A
0 1 2 1
Ω Ω cosecψ
(25.6)
K E
J
f E
i i
ph
k
i
i
k
i
K
rel i
= −






σ ω ε
α
( ) ( )
1
1
1
(25.7)
In the above equations (5–7), A
0
: activity of the excitation source; Ω
1
,Ω
2
: solid angles at the sample
from the source and detector, respectively;
σ
i
ph
E( )
1
: photo effect cross-section at energy E
1
in ele-
ment i; 1
1
−






J
k
i
: relative probability for excitation of K-shell of element i;
ω
i
k
: uorescent yield
for K-shell of element i;
f
i
K
α
: relative transition probability for K
a
X-ray of element i; ε
rel
(E
i
): relative
detector efciency for characteristic X-rays of energy E
i
of element i;
The combined absorption of primary and uorescence X-rays in the sample is determined as
follows (Equation 25.8):
a E E
i s s s i,
( ) ( )= +µ ψ µ ψ
1 1 2
cosec cosec (25.8)
μ
s
(E
1
): absorption cross section in the total sample at excitation energy E
1
; μ
s
(E
i
): absorption cross
section in the total sample at characteristic energy E
i
of element i;
The expressions for the absorption and enhancement correction factors T
i
and H
i
are as follows:
T c c
a d
a
i n
i s
i s
( ,... )
exp( )
,
,
1
1
=
− − ρ
(25.9)
H
i
(c
1
, c
2
,…c
n
) = 1 + Σρ
i,k
(c
1
,…c
i
,…c
k
,…c
n
)c
k
(25.10)
In the case of a polychromatic excitation dened by the distribution w(E
j
) of primary X-rays,
which is usually calculated (Pella et al. 1985), the excitation and absorption or enhancement of
uorescent radiation must be treated together. Factorization of the basic equation is impossible,
K13484_C025.indd 456 8/9/2012 8:31:43 PM

457Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
and the evaluation of particular concentrations becomes more complicated. The basic equation
becomes:
I GK E w E T E E E c
i i j i
ph
j j i j k i k j k
= +
( )


∑
σ ρ
( ) ( ) ( , ) ( )
,1
1




∑
(25.11)
Where ρ
i,k
are contributions to the enhanced intensity of element “i” by excitation of the uorescent
radiation of elements “k”. The summation is performed over all the elements in the sample which
could enhance element “i”. This factor depends on the composition of the sample through absorp-
tion in all elements of the sample. In this case the constant K
i
is no longer expressed as above but
is dened as:
K
J
f E
i
k
i
i
k
i
K
rel i
= −






1
1
ω ε
α
( )
(25.12)
Exact expressions for the above mentioned correction factors, as well as for the expressions of K
i
for L-series X-rays can be found in Tertian and Claisse (1982), Rousseau (1984), He and Van Espen
(1991) and Van Dyck et al. (1986).
25.2.3.2 Starting the Quantitative XRF-Analysis: Principal
Problems and Necessary Assumptions
The theoretical background of the X-ray uorescence process is well known and supported by the
above equations. The measured uorescent intensities represent starting data for the quantication
procedure. But there is a basic problem, namely, when using the above equations the concentration
of the element can be determined from the measured intensity only, if the composition of the sample
is known. Namely the uorescent intensity depends not only on the concentration of the respective
element but also on concentrations of all other elements in the sample, which attenuate the excita-
tion and uorescent radiation in the sample before it excites the atoms within the sample and when
the emitted uorescent radiation penetrates toward the surface of the sample in the direction of the
detector. In this case, but only if all the elements in the sample respond by a uorescent signal in
the spectrum, the concentrations can be obtained from the respective set of equations (4–12) by
iteration (the number of unknowns c
i
in this case equals the number of equations). But in almost
all other cases a problem exists due to the unknown part of the sample, namely that part of the
sample which does not give a response in the uorescence spectrum (light elements like H, C, O,
F, and sometimes also Na, Mg, Al, Si, etc., depending on the excitation, low uorescence yield, and
detector efciency). These elements, which comprise the so-called residual or dark or low-Z matrix,
additionally attenuate the excitation and the measured uorescent radiation (Nečemer et al. 2011).
Different approaches to quantication are therefore applied to solve this problem:
i) In the case of known composition of the residual or dark or low Z part of the sample
matrix (i.e., oxides, cellulose, alumo-silicates, etc.), the concentrations can be calculated by
iteration of a set of equations considering additional absorption in the preselected residual
matrix (Nečemer et al. 2011);
ii) Use of measured intensities of the scattered excitation radiation in the spectrum enables
assessment of the composition of the sample matrix (Van Dyck and Van Grieken 1980;
Nečemer et al. 2011);
iii) Additional measurement of absorption performed on the sample by the transmission-
emission method (Markowicz and Van Grieken 1993; Nečemer et al. 2011).
The rst approach leads to semi-quantitative analysis. Namely in most cases the selected com-
position of the residual matrix is only a more or less good guess or approximation and therefore
K13484_C025.indd 457 8/9/2012 8:31:43 PM

458 Phytotechnologies: Remediation of Environmental Contaminants
it is in principle not possible to say that the result is quantitative (determined only from measured
quantities and fundamental constants) (Nečemer et al. 2011).
The second approach uses the scattered primary radiation from the sample, which is usually
measured together with the uorescence. Scattering is in principle a rather complicated physical
process, dependent on the geometry of the experiment, on the thickness and on the average atomic
number of the atoms in the sample. This correction is usually applied to rather thin samples, for
which the absorption corrections are rather small and therefore uncertainty of these corrections
only little affect the uncertainty of the results (Nečemer et al. 2011).
On the other hand, the absorption process is a straightforward process and yields good experimental
results. In two of the models of our approach to quantitative analysis we utilized absorption measure-
ments in the sample by the transmission-emission method at a single energy, usually at 8.04 keV or
17.44 keV, corresponding to application of Cu or Mo radiators (Markowicz and Van Grieken 1993;
Nečemer et al. 2011). By the iteration of the set of equations (4–12) for the measured elements and
including the absorption in a selected residual matrix, the concentrations of measured elements could
be obtained. The absorption in this particular sample at energy of 8.04 or 17.44 keV is calculated and
compared with the measured value. If the values do not coincide, further iteration, selecting gradually
larger absorption in the selected residual matrix then leads toward the nal values of concentrations of
the measured elements and also to a correct residual matrix, so that the measured absorption at a par-
ticular energy in the sample coincides with the calculated one (Nečemer et al. 2011).
25.2.3.3 Calibrations
In any model or approach to quantitative analysis it is rst necessary to calibrate the XRF system,
in order to evaluate the geometrical constant G as dened in equation (6). For this purpose a set
of thin samples (Van Espen and Adams 1981) or thick pure metals or samples of stable chemical
compounds (Yap et al. 1987) are measured and the sensitivities S
i
are calculated using equations
(5) to (9), employing the known compositions of selected samples. It should be mentioned that the
uncertainty of a such calibration can greatly inuence the quantication of the results. To evalu-
ate the uncertainty of the calibration, we preceded in a somehow different way from that of other
authors Markowicz et al. (1992), and evaluated the geometrical constant G by equations (5–7) from
the calculated sensitivities (Nečemer et al. 2011). The evaluated constant G by denition should be
the same for all calibrated elements. Therefore the calculated standard deviation of the average value
of G in principle determines the uncertainty of the calibration procedure, which includes the uncer-
tainty of the measured intensities, uncertainties of fundamental parameters in constant K
i
, as well as
uncertainties in the absorption coefcients used in the calculations (McMaster et al. 1986; Nečemer
et al. 2011). This uncertainty represents a part of the total uncertainty of the complete quantication
procedure. In most cases this method of calibration yields an average geometrical constant G with an
uncertainty of 2% to 5%. In the case of the polychromatic excitation, the uncertainty can also reach
5% to 10%, due to the uncertainty of the calculated distribution of the excitation X-rays w(E
j
) (Pella et
al. 1985). In the quantication procedure the experimental geometrical constants for particular cali-
brated elements are usually used rather than the average geometrical constant (Nečemer et al. 2011).
25.2.3.4 Detection Limits
In quantitative XRF analysis, it is important to determine the limit of detection (LOD). It is usually
accepted that the minimal detectable intensity of a spectral line must exceed by a factor 3 the stan-
dard deviation of the integrated background under the spectral line. According to this denition the
minimal detectable mass or LOD in grams for a denite element is expressed as:
LOD[ ]g
B
t
S
=






3
(25.13)
K13484_C025.indd 458 8/9/2012 8:31:43 PM

459Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
Sensitivity S is expressed as the signal count rate per gram of sample, B is the integrated back-
ground in counts under the spectral line, and t is the time of measurement. It is important to stress
that the sensitivity and therefore the LOD depend very much on the absorption in the sample matrix.
But the explicit dependence of the LOD on the time of measurement is not quite appropriate (Kump
1997; Nečemer et al. 2011). Therefore the above relation can be written in somewhat different form,
which shows that the accepted LOD as such is quite arbitrary and also requires some additional data
to really correctly determine or assess the limit of detection:
LOD[ ]g
I
S
B
B
B
=












3
(25.14)
From this expression it is evident that the LOD depends on the relative standard deviation of the
background under the spectral line and not only on the measuring time. But since the sensitivity Sis
also dened as the signal count rate of the spectral line corresponding to the sample mass m
0
(I
0
/m
0
),
the LOD at 33% relative standard deviation of the background can be expressed as:
LOD[ ]g
I
I
m
B
=






0
0
(25.15)
This conrms that the LOD depends only on the ratio of background to signal count rates (Kump
1997; Nečemer et al. 2011).
25.2.4 total reFlection X-ray Fluorescence
25.2.4.1 TXRF-Excitation Module
The basic fundamentals of total reection XRF spectrometry (TXRF) are similar to those of
EDXRF, although they have quite different excitation modes. In TXRF systems, extremely small
amounts of the sample (just few μl) are rst deposited as a liquid solution on the optically smooth
substrate, which is usually quartz and then dried. The dried residues are then excited by a well-
collimated X-ray beam at an angle smaller than the angle of total reection for the substrate (<1.8
mrad for quartz) (Klockenkämper 1997; Kump et al. 1997; Neč emer et al. 2011). In this case, the
majority of the incident X-ray radiation is totally reected from the quartz surface, and only a minor
part of it is absorbed by the deposited sample to excite uorescence. The penetration of the incident
X-ray beam into the reecting material is drastically reduced under these conditions, and the scat-
tered and uorescence radiation contributed by the carrier in this geometry is therefore negligible.
Consequently, the background radiation due to scattering on a small amount of sample is very low,
signicantly increasing the sensitivity of TXRF when compared to standard XRF spectrometry
(Schwenke and Knoth 1993; Kump et al. 1997).
The sensitivity level, however, still strongly depends on the atomic number of the element,
although it does extend down to a few 10 ppb (10 μg kg
−1
) dry weight. To achieve the described
excitation conditions, a special total reection module is required to shape the excitation beam
from the X-ray tube into a suitable form that will excite a small amount of the dried sample residue
placed on the quartz substrate. Although there are many expensive commercially available TXRF
spectrometers, there are also several cheaper laboratory-built systems that exist worldwide. EDXRF
systems, which usually have a ne-focus Mo anode X-ray tube, an X-ray generator, a semiconduc-
tor X-ray detector and spectroscopy electronics, may be equipped with a total reection module
provided by the Atom Institute (Vienna). In this way, a cheap alternative to the commercial TXRF
system can be built and used for multi-element analyses of different environmental samples (Kump
1997; Nečemer et al. 2011).
K13484_C025.indd 459 8/9/2012 8:31:43 PM

460 Phytotechnologies: Remediation of Environmental Contaminants
25.2.4.2 Sample Preparation and Quantification
For the TXRF analyses the sample material in form of solution must be prepared. A solid ground
sample thus requires destructive treatment using wet or dry digestion procedures. This process
usually utilizes a decomposition procedure with a small amount of ground material (0.1–0.2 g),
and involves the application of a mixture of mineral acids followed by microwave digestion. The
resulting solution can be analysed by the TXRF after the addition of an internal standard, which is
usually a Ga, V or Y in the form of a standard AAS solution. A small amount of this decomposed
sample solution (10 μl) is then pipetted onto a quartz substrate, dried in a desiccator, and measured
(Nečemer et al. 2011). As in the case of EDXRF, the TXRF method enables multi-element analysis.
Usually, with a Mo anode excitation tube, elements from Z = 16 (S) to Z = 92 (U) can be determined
in the concentration range of a few percent to a few mg kg
−1
. The determination of lighter elements
like Na, Al and Mg is possible in a vacuum and with the application of a Si drift detector with thin
beryllium or polymer window. For analysis of silicon the substrate should be optically at Plexiglas
instead of quartz.
Since only a very small amount of dry sample is analysed by TXRF, the relative sensitivity is
about one to two orders of magnitude better than for EDXRF, although the absolute sensitivity of
TXRF is very good and reaches few pictograms, in comparison to a few micrograms for EDXRF.
The main advantage of TXRF over EDXRF is the possibility of rapidly analysing a larger number
of liquid samples (i.e., waters, different soil extracts for determining the soluble fractions of metals
extracted by CaCl
2
or NH
4
-acetate) and samples that may be prepared by a simple procedure (i.e.,
by the dilution of soluble samples, like bee honey in water or juices, milk and vines) (Nečemer et
al. 2009, 2011). TXRF is also very suitable for the analysis of very small amounts of biological
samples, like plant xylem sap, where only a small amount of material is available. In the latter case
the TXRF is actually the only method which can provide multi-element analysis of a very small
amount of the sample (Nečemer et al. 2011).
25.3 EXPLORING METAL-LOCALIZATION AND
DISTRIBUTION IN PLANT TISSUES AND CELLS
Metal pollution is frequently resulting in changes of uptake and distribution of several mineral
nutrients and metals in plant tissues and cells, hence affecting their primary functions. To lower
their effects, the mechanisms of ion homeostasis in plants seem to exert a tendency for compart-
mentalization of excess metals in tissues and cell parts with the lowest metabolic activity. Bulk
elemental analyses do not allow for specic analysis of mechanisms contributing to metal tolerance
and detoxication on the cellular and sub-cellular levels, therefore multi-elemental imaging tech-
niques have to be applied when their accumulation and distribution patterns within plant tissues are
questioned and the mechanisms underlying metal detoxication and compartmentalization on the
cellular, tissue and organ levels are being described. Due to the complex morphology and highly
heterogeneous chemical composition of biological systems, discriminating qualitatively and quan-
titatively metal composition and their local distribution, and correlating them to the cellular and
sub-cellular biological structures, continues to be a challenge (Kaulich et al. 2009).
Determination of the distribution and chemical state of elemental constituents within biological
systems at sub-cellular level down to trace level concentrations is of growing importance for gaining
new insights about the highly complex functions of elements within the particular biological tissues
or cells. Different analytical techniques have been developed in the past years, which are comple-
mentary in terms of lateral resolution, chemical sensitivity, quantitative analysis, depth proling or
bulk sensitivity and detection of elemental isotopes (Kaulich et al. 2009). Among those techniques,
energy dispersive X-ray analysis (EDX) in a transmission electron microscope provides the highest
lateral resolution (<10 nm), but low to moderate chemical sensitivity (0.01–0.1 wt %) and requires
the specimen to be analysed in vacuum and to be sectioned to (ultra)thin slices. Secondary ion
K13484_C025.indd 460 8/9/2012 8:31:44 PM

461Analytical Tools for Exploring Metal Accumulation and Tolerance in Plants
milling spectroscopy has very high chemical sensitivity (ppb–ppm range) and high spatial resolu-
tion (<100 nm) and allows detection of elemental isotopes, but quantication of data is very difcult.
Electron energy-loss spectroscopy has very high lateral resolution (<10 nm) and chemical sensitivity
in the ppm range, but requires sectioning of the specimen, and like secondary ion milling spectros-
copy and EDX is only surface sensitive and does not provide bulk information (Mills et al. 2005;
Kaulich et al. 2009). A distinct limitation of the electron-based techniques is also that they require
conducting surfaces and high vacuum environment (Kaulich et al. 2009).
X-ray spectrometric techniques by means of excitation of an inner shell electrons and emission
of quanta of characteristic uorescence X-rays was rst proposed in 1928 and was well established
during the 20th century (Jenkins 1999; Kaulich et al. 2009). Nowadays micro-proton-induced X-ray
emission (PIXE) based on excitation of the atoms in the sa</