Xray Microprobe for Environmental Sciences

1
X-ray Microprobe for Environmental Sciences
Matt Newville, GeoSoilEnviroCARS Consortium for
Advanced Radiation Sources University of Chicago
Steve Sutton Mark Rivers Peter Eng
X-ray Microprobe Techniques
X-ray Fluorescence
Elemental abundance and correlations
X-ray Absorption (XANES and EXAFS)
Oxidation state of selected element,
Near-neighbor distances and
coordination numbers
X-ray Diffraction
Crystallographic and surface structures
Fluorescence Tomography
3-dimensional density and elemental
abundances and correlations
Objective Chemical associations, speciation, and
structural information for heavy elements (Z gt
20) with spatial sensitivity at the 1-10 micron
scale, and few sample constraints.
2
GSECARS Beamline Optics for Microprobe
GeoSoilEnviroCARS Sector 13 Advanced Photon
Source, Argonne National Lab
Undulator Beamline High collimation allows
efficient focusing, for x-ray microprobe,
and x -ray
diffraction (small crystals, high pressure).
Very high power densities of white beam need to
be absorbed by beamline optics.
High Pressure Diffraction Station (Diamond-Anvil-C
ell, Large Volume Press)
Monochromator LN2-cooled Si (111)
Microprobe Station
Storage Ring
Large Focusing Mirrors (May, 2001)
Diffraction Station
Tomography and diffraction Station
Bending Magnet Beamline 2nd-generation source,
with fairly high energy x-rays
(up to
100KeV)
3
Advanced Photon Source Undulator Beamlines
APS UNDULATOR A Period length 3.30 cm Number of
periods 7 Length 2.47 m Kmax 2.78
(effective at minimum gap) Minimum gap 10.5
mm Energy Tuning Range 2.9 - 13.0 keV
(1st harmonic) 2.9 - 45.0 keV (3rd and 5th
harmonic) On-axis peak brilliance (at 6.5 keV)
9.6x1018 ph/s/mrad2 /mm2 /0.1bw Power
(closed gap) 6.0 kW On-axis power density
(closed gap) 167 kW/mrad2
Source Size and Divergence Vert s 21mm,
s 7mrad Horiz s 359mm, s 21mrad
4
X-ray Fluorescence Microprobe
X-ray Fluorescence Measure characteristic x-ray
emission lines from de-excitation of electronic
core levels for each atom.
Element Specific Elements with Zgt16 can be seen
at the APS, and it is usually easy to distinguish
different elements.
Quantitative precise and accurate elemental
abundances can be made. x-ray interaction with
matter well-understood.
Low Concentration concentrations down to a few
ppm can be seen.
Natural Samples samples can be in solution,
liquids, amorphous solids, soils, aggregrates,
plant roots, surfaces, etc.
Small Spot Size measurements can be made on
samples down to a few microns in size.
Combined with Other Techniques XANES, EXAFS, XRD
5
X-ray Absorption Spectroscopy XANES and EXAFS
X-ray Absorption Spectroscopy Measure
energy-dependence of the x-ray absorption
coefficient m(E) either log(I0 /I) or (If / I0
) of a core-level of a selected element
Element Specific Elements with Zgt20 can have
EXAFS measured at the APS.
Valence Probe XANES gives chemical state and
formal valence of selected element.
Local Structure Probe EXAFS gives atomic
species, distance, and number of near-neighbor
atoms around a selected element..
Low Concentration concentrations down to 10 ppm
for XANES, 100 ppm for EXAFS.
Natural Samples samples can be in solution,
liquids, amorphous solids, soils, aggregrates,
plant roots, surfaces, etc.
Small Spot Size XANES and EXAFS measurements can
be made on samples down to 5 microns in size.
XANES X-ray Absorption Near-Edge Spectroscopy
EXAFS Extended X-ray Absorption Fine-Structure
6
X-ray Absorption Fine-Structure Spectroscopy
1. An x-ray of energy E is absorbed by an atom,
destroying a core electron state with energy E0
and creating a photo-electron with energy (E-E0).
2. The probability of absorption m(E) depends on
the overlap of the core-level and photo-electron
wave-functions. Since the core-level is
localized, this overlap is determined by the
photo-electron wave-function at the center of the
absorbing atom. For an isolated atom , this is
a smooth function of energy.
3. With another atom nearby, the photo-electron
can scatter from the neighbor. The interference
of the outgoing and scattered waves alters the
photo-electron wave-function at the absorbing
atom, modulating m(E).
4. The oscillations in m(E) depend on the
near-neighbor distance, species and coordination
number.
7
GSECARS Fluorescence and XAFS Microprobe Station
Beamline13-ID-C is a world-class micro-beam
facility for x-ray fluorescence (XRF) and x-ray
absorption spectroscopy (XAS) studies
Incident Beam Monochromatic x-rays from LN2
cooled Si (111)
Sample Stage x-y-z stage, 0.1mm resolution
Fluorescence detector 16-element Ge detector
shown, Si(Li) detector, Lytle Detector, or
Wavelength Dispersive Spectrometer at 90o to
incident beam
Optical Microscope (5x to 50x) with external
video system
Data Collection Flexible software for x-y
mapping, traditional XAFS scans, XAFS scans vs.
sample position.
Focusing Horizontal and Vertical
Kirkpatrick-Baez mirrors
8
Double Focussing With Elliptical Mirrors
Kirkpatrick-Baez Focusing Mirrors Use x-ray
reflection from an elliptical shaped mirror to
focus the beam with a large demagnification.
9
Kirkpatrick-Baez Focusing Mirrors
The table-top Kirkpatrick-Baez mirrors use
four-point benders and flat, trapezoidal mirrors
to dynamically form an ellipsis. They can focus
a 300x300mm beam to 1x1mm - a flux density gain
of 105. With a typical working distance of
100mm, and an energy-independent focal distance
and spot size, they are ideal for micro-XRF and
micro-EXAFS. We use Rh-coated silicon for
horizontal and vertical mirrors to routinely
produce 3x3mm beams for XRF, XANES, and EXAFS.
10
X-ray Fluorescence Detector
Solid-State Multi-Element Ge Detector for X-Ray
Fluorescence detection
Ge solid-state detectors have energy resolutions
of 250 eV, which separates most fluorescence
lines from different elements. They allow a full
XRF spectrum (or the windowed signal from several
lines) to be collected in seconds Ge detectors
are limited in total count rate (to 100KHz), so
multiple elements (10 to 30) are used in parallel
to make one large detector. We use a detector
with 16-elements. Detection limits are at the
ppm level for XRF. XANES and EXAFS measurements
of dilute species (10ppm) in heterogeneous
environments can be measured.
11
Cr Redox in Boreholes below Hanford Waste Tanks
Sam Traina, Chia Chen, Isao Yamakawa (Ohio State
Univ.), Gordon Brown, Jeff Warner, Jeff
Catalano (Stanford Univ.)
Cr has been found at concentrations gt10,000 ppm
in the vadose zone beneath the high-level nuclear
waste tanks at Hanford, WA. The very alkaline
solution in the tanks favors Cr6 in chromate
(CrO42-), which is highly mobile in groundwater,
acutely toxic, teratogenic, and carcinogenic.
Cr3 is relatively immobile in groundwater, and
far less toxic. Therefore, the environmental
impact of Cr from tank leachates in the vadose
zone is highly dependent on the chromium
speciation. Determining the extent of Cr6 ?Cr3
reduction within the soils, and identifying
possible reduction pathways (Fe- and
Mn-(oxy)hydroxides, bacterial) are vital.
Borehole samples were collected from the vadose
zone under leaking waste tanks. Radioactive (10
mCi/g Cs-137) soil sections were embedded in
epoxy, and sent for synchrotron analysis to APS
and SSRL during January, 2001.
12
Cr under Leaking Hanford Waste Tanks XRF Maps
These x-ray fluorescence maps of elemental
concentration were made with incident x-ray beam
5x5mm, E 7200eV (just above the Fe K-edge).
The sample was scanned through this beam in 5mm
steps. (8hr collection). Each map shows the
integrated fluorescence of the characteristic
line (Cr and Fe Ka, Ba La).
micro-XRF mapping and Cr XANES were measured at
GSECARS, bulk Cr XAS SSRL beamline 11-2, and Cs
and Cr XANES at PNC-CAT (APS 20-ID). Elemental
maps (below) show several Cr hot-spots, but
also a fairly high background level. Cr
correlations with Fe and Mn (the proposed abiotic
reducing agents for Cr6 ) varied considerably
between 4 different borehole samples measured.
Cr XANES measurements were made on selected
areas of high/low Cr, with/without Fe, and so on.
13
Hanford Tank Cr Oxidation State with XANES
XANES can easily distinguish Cr6 from Cr3 from
the height of the pre-edge peak.
Prelminary analysis of the Cr XANES from
different core samples shows high variability of
the Cr6 / Cr3 ratio, with fairly significant
reduction of Cr6 to Cr3 and little obvious
correlation with the presence of Fe. Cr EXAFS
measurements are still being processed (but seem
to be dominated by typical Cr3-oxide)
Hanford Cr Oxidation State
14
Metals Distribution at Root/soil Interface
Andreas Scheinost, Ruben Kretzschmar (ETH Zurich)
Arrow mark regions where Zn m-XAFS spectra were
collected.
Metal oxide dust was introduced to a forest
topsoil resulting in 5000 ppm Zn and 2500 ppm Cu
in the soil. Undisturbed soil samples will be
taken a various times up to 5 years to follow the
distribution pathways of the toxic metals. mXRF
maps show the distribution of Zn, Cu, Fe and Mn
near a barley root growing in the contaminated
soil. Close to a Zn oxide particle, the root is
strongly enriched in Zn.
stele
Mn-rich
Zn-rich
cortex
Fe-rich
Cu-rich
Optical microscope image of root
15
Zn EXAFS and Speciation at Rhizosphere
The Zn-rich and Cu-rich areas consist of Zn
oxide. The remaining areas lack a strong
second shell, and show tetrahedral Zn-O
coordination, suggesting that Zn dissolved from
the oxide is sorbed by the root cortex and by Fe
and Mn hydroxides.
16
Oxidation state maps Mn redox at plant roots
D. Schulze (Purdue University)
Manganese is an essential nutrient for plants,
needed for photosynthesis and response to stress
and pathogens. Reduced Mn2 is soluble and
bio-available in soils but Mn4 will precipitate
(along with Mn3) as insoluble Mn oxides. The
redox chemistry of Mn in soil is complex, with
both reduction and oxidation catalyzed by
microorganisms. Spatially-resolved m-XANES is
well-suited for mapping Mn oxidation state in
live plant rhizospheres to understand the role of
Mn redox reactions in a plants ability to uptake
trace elements.
Collecting Mn fluorescence with the incident be
at a few well-chosen energies around the Mn
K-edge, we make 3-d (X-Y-Energy) maps that give
the spatial distribution of Mn oxidation states.
XRF image of total Mn concentration (left) of
soil traversed by a sunflower root (dashed line)
showing the heterogeneous Mn and enrichment near
the root. The Mn oxidation state map (right)
shows both Mn2 and Mn4 in the Mn-rich sites,
with a tendency for the reduction near the root.
17
Cu speciation in Hydrothermal Fluid Inclusions
John Mavrogenes, Andrew Berry (Australian
National University)
Understanding the metal complexes trapped in
hydrothermal solutions in minerals is key to
understanding the formation of ore deposits. mXRF
and mXAFS are important tools for studying the
chemical speciation and form of these fluid
inclusions.
Natural Cu and Fe-rich brine fluid inclusions in
quartz from Cu ore deposits were examined at room
temperature and elevated temperatures by XRF
mapping and EXAFS. Initial Expectation
chalcopyrite (CuFeS2) would be precipitated out
of solution at low temperature, and would
dissolve into solution at high temperature. We
would study the dissolved solution at temperature.
65mm
XRF mapping showed that the initial expectation
was wrong, and that a uniform solution at room
temperature was becoming less uniform at
temperature. This was reversible.
18
Cu speciation in Hydrothermal Fluid Inclusions
John Mavrogenes, Andrew Berry (Australian
National University)
XAFS measurements at low and high temperature
were also very different, with a very noticeable
differences in the XANES indicating a change in
speciation Low temp Cu2 High temp Cu1
Low temp (?) High temp (?)
Preliminary fits to the EXAFS of the high
temperature phase (below) is also consistent with
Fulton et al Cu1 with Cl (or S) at 2.09Å, and
possibly some O at 1.96Å.
These results are consistent with Fulton et al
Chem Phys Lett. 330, p300 (2000) study of Cu
solutions near critical conditions Cu2 solution
at low temperature, and Cu1 associated with Cl
at high temperatures.
19
High Resolution X-ray Fluorescence and EXAFS
A complication in measuring fluorescence and
EXAFS in many natural samples is the presence of
fluorescence lines from other elements near the
line of interest The resolution of a
solid-state fluorescence detector (150eV) is
sometimes not good enough to resolve nearby
fluorescence lines
The Wavelength Dispersive Spectrometer has much
better resolution (20eV) than a solid-state
detector, and a much smaller solid angle. It uses
a Rowland circle, not electronics, to select
energies of interest. It needs the brightness
of an undulator, but complements the Ge
detectors, and allows XRF and even EXAFS on
systems with overlapping fluorescence lines.
20
Using the WDS for XRF Mapping Cs on Micas
J. McKinley, J. Zachara, S. M. Heald (PNNL)
137Cs cannot be wholly extracted from
contaminated soils, even using harsh chemical
treatments or cation exchangers. McKinley and
Zachara exposed natural mica, similar to that
found near PNNL, to a Cs-rich solution, embedded
the mica in epoxy resin and cut cross-sections
through the mica. Cs sorbs strongly to micas at
selective edge sites and interlayer binding
sites.
1000 x 200mm image of the Cs La line in biotite
with a 5x5mm beam, 5mm steps and a 2s dwelltime
at each point. The incident x-ray energy was
7KeV.
Detecting the Cs La fluorescence line is
complicated by the nearby Ti Ka line. The high
resolution fluorescence detector can make these
measurements much easier.
21
Using the WDS for XRF Mapping Cs on Micas
J. McKinley, J. Zachara, S. M. Heald (PNNL)
Cs La map of muscovite cross section with WDS
1300 x 150 mm, 5 mm pixels, 1 sec dwell time.

• Initial Results
• Cs concentrated on mica edges
• 67 of total Cs is located at edges
• Highest Cs concentration 300 ppm, average Cs
  concentration 10 ppm, detection limit 1 ppm,
• Non-edge regions contain one-third of the Cs
  with average concentration of 5 ppm.
• Cs-bearing interior regions appear associated
  with flake partings rather than interlayer
  lattice sites.

Optical
22
Sector Zoning of Rare Earth Elements in Apatites
John Rakovan (Miami University)
.
Apatites have a high affinity for Rare Earth
Elements (REE), and are often used to study
petrogenesis. Heterogeneities in crystal surface
structure during apatite growth can strongly
alter REE incorporation. Most REE show
sectoral zoning in apatite based on ionic size.
Ions larger than Ca2 (La3) preferring growth
along the 001 face, and those smaller than Ca2
(Sm3) preferring the 011 face
Eu is the only REE showing no zonation, but it
has two valence states and two ionic sizes that
straddle the size of Ca2. Is there a
partitioning of Eu based on valence state/ionic
size?
23
Sector Zoning of Rare Earth Elements in Apatites
Since Eu has two valence states with different
ionic sizes (Eu2 / 1.2 Å, Eu3 / 1.3 Å),
it was suggested that there may be a
valence/ionic size variation in different growth
zones. The bad news There is far too much Mn
in the apatite to separate from the Eu
fluorescence line with a solid state detector.
.
Using the high resolution WDS and the microprobe,
we measured the Eu XANES on several spots in the
different sectors, and across a lt011gt / lt001gt
boundary.
X-ray counts
Result We see almost no change at all in Eu2 /
Eu3 across the zone boundary the ratio is 17
Eu2 throughout the apatite.
Energy (keV)
24
Using the WDS for XANES 1000ppm Au in FeAsS
Louis Cabri (NRC Canada), Robert Gordon, Daryl
Crozier (Simon Fraser), PNC-CAT
1000ppm Au in FeAsS (arsenopyrite) The
understanding of the chemical and physical state
of Au in arsenopyrite ore deposits is complicated
by the proximity of the Au LIII and As K edges
and their fluorescence lines. At the Au
LIII-edge, As will also be excited, and fluoresce
near the Au La line. Even using the WDS, the
tail of the As Ka line persists down to the Au La
line, and is still comparable to it in intensity.

250x250mm image of the Au La line in arsenopyrite
with a 6x6mm beam, 5mm steps and a 2s dwell time
at each point. The x-ray energy was 12KeV.
25
Using the WDS for XANES 1000ppm Au in FeAsS
Louis Cabri (NRC Canada), Robert Gordon, Daryl
Crozier (Simon Fraser), PNC-CAT
CANADIAN MINERALOGIST 38, pp1265-1281 (2000)
The tail of the As Ka line is still strong at the
Au La energy, so using a Ge detector gave the Au
LIII edge-step as about the same size as the As K
edge-step, and the Au XANES was mixed with the As
EXAFS. With the WDS, the As edge was visible,
but much smaller, and so the Au XANES was clearer.
The Au LIII edge of two different natural
samples of FeAsS with the WDS. Both samples had
1000ppm of Au. We see clear evidence for
metallic and oxidized Au in these ore deposits.
As K-edge 11.868 KeV As Ka line
10.543 KeV Au LIII-edge 11.918 KeV Au La line
9.711 KeV
26
Surface Diffraction Surface of Hydrated Alumina
Peter Eng (CARS), Tom Trainor, Gordon Brown, Jr
(Stanford Univ), Glenn Waychunas (Lawrence
Berkeley Lab) Science 288, pp1029-1033 (2000)
The interaction of water with natural surfaces is
one of the most fundamental chemical reactions in
nature. Processes such as mineral dissolution
and sorption/ desorption reactions at
mineral-water interfaces play major roles in
weathering, contamination of groundwater,
environmental restoration, and biogeochemical
cycling of elements.
a-Al2O3 is an important model system for
understanding the reactivity of naturally
abundant phases of Al-containing (hydr)oxides
such as gibbsite or hydrous aluminosilicate
clays. The Al in these phases have similar
coordination chemistry. The reaction of water
with the a-Al2O3 (0001) has received a lot of
experimental and theoretical attention. Surface
diffraction measurements of the mineral-water
interface is an important first step to
understanding these interactions and the
atomic-level reactivity of the Hydrated mineral
surface.
27
Al2O3 Surface Scattering Methods
SAMPLE single crystal wafer of (0001) a-Al2O3,
0.5mm thick, 50mm in diameter, highly polished (1
Å rms roughness), and fully hydrated (see sample
cell). In situ liquid cell with thin membrane
mounts on diffractometer, and traps either liquid
or (in this case) humid air for scattering,
reflectivity, and fluorescence. This is enough to
fully hydrate the surface of Al2O3.
Measurement Technique Crystal Truncation Rods
A surface disrupts the infinite 3D lattice that
make Bragg diffraction spots, and moves
diffraction intensity to lines between the Bragg
points. The q-dependence and shape of these
rods is sensitive to the roughness and atomic
arrangement at the crystal surface.
This may not appear to be a normal microprobe
experiment, but it does require a bright and
highly-collimated x-ray beam 10 to 20mm high.
And this is a very large, synthetic crystal!
28
Surface Structure of Al2O3 Results
Analysis of the truncation rod data shows two
important results for the surface structure of
hydrated (0001) a-Al2O3

• The surface is oxygen-terminated, with a slight
  relaxation. This differs from in vacuum
  measurements, which show Al termination. The
  relaxed structure is similar to the basal plane
  of gibbsite g-Al(OH)3
• There is good evidence for a disordered water or
  hydroxyl layer

Disordered Oxygen Overlayer
Oxygen Terminated Al2O3 Surface
Hydrated Al2O3 Surface
29
X-ray Fluorescence Tomography Overview
X-ray computed microtomography (CMT) gives 3D
images of the x-ray attenuation coefficient
within a sample. At each angle, a 2D absorption
image is collected. The angle is rotated around
w in 1o steps through 180o, and the 3D image is
reconstructed with software. Element-specific
imaging can be done by acquiring tomograms with
incident energies above and below an absorption
edge.
Microscope objective
broad x-ray beam
Sample
Phosphor
CCD camera
x-rays
w
Visible light
rotation stage
Transmission detector
Fluorescence x-ray tomography is done with a
pencil-beam scanned across the sample. The
sample is rotated around w and translated in x.
Tranmission x-rays are can be measured as well to
give an overall density tomograph.
Sample
thin x-ray beam
fluoresced x-rays
transmitted x-rays
w
Fluorescence detector
rotation stage

• can collect multiple fluorescense lines.

x

• data collection is relatively slow.

translation stage

• can be complicated by self-absorption.

30
Fluorescence Tomography Experimental Setup
Optical microscope, KB mirrors
Fluorescence detector multi-element Ge detector
Sample mounted on silica fiber
Sample stage x-y-z-q
31
Fluorescence Tomography Sinograms
The Raw fluorescence tomography data consists of
elemental fluorescence (uncorrected for
self-absorption) as a function of position and
angle a sinogram. This data is reconstructed
as a virtual slice through the sample by a
coordinate transformation of (x,w) ? (x, y). The
process can be repeated at different z positions
to give three-dimensional information.
Fluorescence Sinograms for Zn, Fe, and As
collected simultaneously for a section of
contaminated root (photo, right) x 300mm
in 5mm steps w 180? in 3? steps
w
As
Zn
Fe
x
32
3D Distributions of Heavy Metals in Roots
S. Fendorf, C. Hansel (Stanford) Toxic Metal
around Root-borne Carbonate Nodules
The role of root-borne carbonate nodules in the
attenuation of contaminant metals in aquatic
plants is investigated with EXAFS, SEM and X-Ray
fluorescence tomography. These images of a 300
mm root cross-section (Phalaris arundinacea) show
Fe and Pb are uniformly distributed in the root
epidermis while Zn and Mn are correlated with
nodules. Arsenic is poorly correlated with the
epidermis, suggesting a non-precipitation
incorporation.
Slicing the root would cause enough damage that
2D elemental maps would be compromised.
Such information about the distribution of
elements in the interior of roots is nearly
impossible to get from x-y mapping alone
photograph of root section and reconstructed
slices root from fluorescent x-ray CT.
33
X-ray Microprobe for Environmental Sciences
Synchrotron x-ray microprobes have wide
applicability in the earth, planetary, soil and
environmental sciences.
Greatest current demand is for microbeam
applications of XANES and EXAFS.
Expect demand to growth areas to be
fluorescence microtomography studies of fragile
materials study of in-situ biogeochemical
processes surface diffraction and
spectroscopy high-pressure, high-temperature
(extreme condition) studies
Thanks to Steve Sutton, Mark Rivers, Peter
Eng GeoSoilEnviroCARS, University of
Chicago http//cars.uchicago.edu/gsecars/