Using a Wavelength Dispersive Spectrometer for EXAFS

1
Using a Wavelength Dispersive Spectrometer to
measure XAFS
Matt Newville, Steve Sutton, Mark Rivers, Peter
Eng GSECARS
GSECARS beamline and microprobe station,
Kirkpatrick-Baez mirrors
XAFS and x-ray fluorescence measurements
The Wavelength Dispersive Spectrometer
Comparison of WDS and solid state detectors
WDS Applications
XRF Cs sorbed onto mica
XANES Au in FeAsS
EXAFS Re in K7ReOP2W17O61.nH2O
2
The GSECARS Microprobe
The GeoSoilEnviroCARS beamline 13-IDC provides a
micro-beam facility for x-ray fluorescence (XRF)
and x-ray absorption spectroscopy (XAS) studies
in earth and environmental sciences.
sample x-y-z stage 0.1mm step sizes
Horizontal and Vertical Kirkpatrick-Baez
focusing mirrors
fluorescence detector multi-element Ge detector
(shown), Lytle Chamber, Si(Li) detector, or
Wavelength Dispersive Spectrometer
optical microscope (10x to 50x) with video system
3
Kirkpatrick-Baez focusing mirrors
The table-top Kirkpatrick-Baez mirrors use a
four-point bender and a flat, trapezoidal mirror
to dynamically form an ellipsis. They can focus
a 300x300mm monochromatic 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-EXAFS. We routinely use Rh-coated silicon
(horizontal) and fused-silica (vertical) mirrors
to produce 4x4mm beams for XRF, XANES, and EXAFS.
4
X-ray Absorption Spectroscopy XANES and EXAFS
Measure the energy-dependence of the x-ray
absorption coefficient m(E) either log(I0 /I)
or (If / I0 ) through a core-level energy of a
selected element.
XANES X-ray Absorption Near-Edge Spectroscopy
EXAFS Extended X-ray Absorption Fine-Structure
Characteristics of XANES and EXAFS
Element Specific all elements (with Z20 or so)
can be measured at APS
Low Concentration selected element can be as low
as a few ppm
Natural Samples crystallinity is not required --
samples can be
liquids, amorphous solids, soils, aggregates,
and surfaces.
Local Structure EXAFS gives atomic species,
distance, and number
of near-neighbor atoms around selected
element
Valence Probe XANES gives chemical state and
formal valence of selected element
5
X-ray Absorption Fine-Structure Spectroscopy
1. An x-ray of energy E is absorbed by an atom,
destroying a core electron 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 and species (the
energy-dependence of the scattering amplitude
depends on Z).
6
Typical GSECARS Microprobe Application XRF /
EXAFS Sr in coral
Nicola Allison, Adrian Finch (Univ of Brighton,
Univ of Hertfordshire, UK)
A common use of the microprobe is to make an
x-ray fluorescence (XRF) map and then collect
XANES or EXAFS on selected spots in that map.
The abundance of Sr in aragonite (CaCO3) formed
by corals has been used as an estimate of
seawater temperature and composition at aragonite
formation. XRF maps of a section of the coral
were made with a 5mm X 5mm beam and a 5mm step
size. The Sr and Ca fluorescence (and several
other trace elements) were measured
simultaneously at each pixel with a multi-element
Ge solid-state detector. The Sr and Ca maps show
incomplete correlation. The relative Sr
abundance therefore varies substantially on this
small length scale, although this section of
aragonite must have been formed at constant
temperature.
Ca
Sr
200mm
300mm
The Sr XAFS was measured at a spot with fairly
high Sr concentration -- above the solubility
limit of Sr in aragonite...
7
Typical GSECARS Microprobe Application XRF /
EXAFS Sr in coral
Since the Sr concentration was above its
solubility limit (1) in aragonite, it was not
known if Sr would precipitate out into
strontianite (SrCO3 a structural analog of
aragonite), or remain in the aragonite
phase. First shell EXAFS is same for both
strontianite and aragonite 9 Sr-O bonds at
2.5A, 6 Sr-C at 3.0A. Second shell EXAFS
clearly shows Sr-Ca (not Sr-Sr) dominating, as
shown at left by contrast to SrCO3 data, and by
comparison to a FEFF-simulated EXAFS spectrum of
Sr substituted into aragonite. The coral is able
to trap Sr in aragonite at a super-saturated
concentration.
8
Fluorescence XAFS measurements
XRF and XAFS in natural and heterogeneous samples
can be complicated by the presence of
fluorescence lines from other elements near the
line of interest. A detector with some
energy-resolution helps discriminate against
photons at uninteresting energies.
Fe K-edge 7.112 KeV Fe Ka line 6.403
KeV Fe Kb line 7.057 KeV Co K-edge
7.709 KeV Co Ka line 6.930 KeV
Example a dilute quantity of Co in an Fe-rich
system. The Fe will be excited by the Co K-edge
radiation. Even though the Fe Kb is 5X weaker
than the Fe Ka intensity, it may be much larger
than the Co Ka intensity.
Similar conflicts occur when two L lines
interfere with each other (the La and Lb are
about the same intensity, too), or when an L and
a K-line interfere.
Si(Li) and Ge solid-state detectors give energy
resolutions of 100 to 300 eV (with the best
resolution often limiting count rates to 1KHz),
which is sometimes not good enough. These
detectors are also limited in total count rate
(up to 100KHz, but at the worst resolution),
which can be a problem -- especially with intense
x-ray beams.
9
The Wavelength Dispersive Spectrometer (Oxford
WDX-600)
Borrowing technology developed for the
electron-microscope community, the Wavelength
Dispersive Spectrometer uses an analyzer crystal
on a Rowland circle to select a fluorescence
line. This has much better resolution (30eV)
than a solid state detector (250eV), doesnt
suffer from electronic effects like dead-time,
and can have superior peak-to-background ratios.
The solid-angle and count-rates are somewhat
lower.
Sample and x-y-z stage
Kirkpatrick-Baez focusing mirrors
Ion chamber
Table-top slits
Optical microscope
Wavelength Dispersive Spectrometer
10
210mm Rowland circle containing sample, crystal
analyzer, and detectors
11
WDX-600 detailed view
detectors 2 proportional counters (one flowing
P-10 gas, and one sealed with 2 atm Xe) in
tandem.
slits define angular acceptance and energy
resolution

crystals LiF (200), LiF(220), LiF(420), and
PET, on a six crystal turret. Crystal size 45 x
15 mm
By using a Johannson geometry Rowland circle, a
point source focuses to a point at the detector
slit. Aberrations are minimized, and the
signal-to-noise ratio is improved.
12
Comparisons of the WDS and solid-state detectors
Steve Sutton and Mark Rivers, data collected at
NSLS X-26A.
Heres part of the XRF spectra for a synthetic
glass containing several rare-earth elements
using both a Si(Li) detector and the WDS.

13
Comparisons of the WDS and solid-state detectors
Typical values for the WDS and a Ge solid-state
detector
Ge Solid-State
WDS
energy resolution 30eV 100eV to
300eV
depending on
shaping time active area 500mm2
100mm2 (per detector, often 13X)
(varies with angle) working
distance 180mm 100mm max total count
rate none 100KHz
(per detector, often 13X)
14
Issues using the WDS
Alignment The WDS weighs 30kg, and needs to
be aligned fairly well 1 mm vertical 1
mm in/out-board
10mm up/down-stream For our initial
run, we adjusted the height by hand, and had a
motorized in/out-board motion. For the
up/down-stream position, we brought the
sample to the spectrometer, which limits the
focusing ability of the microprobe.
Tunability The WDS selects one energy at a
time, and looking at different
energies requires a mechanical scan. So, unlike
a solid-state detector, the
WDS does not simultaneously measure multiple
energies --- it does not have an
MCA. So XRF maps of
multiple elements (like the Sr/Ca example) are
not practical with the WDS.
15
Using the WDS for XRF Cs on biotite
S Sutton, J McKinley, J Zachara (PNNL)
Biotitie is a mica that contains trace amounts of
many transition metals, a few percent Ti, and
major components of Ca and Fe. To study how Cs
would bind to the surface and layers of the
biotite, McKinley and Zachara exposed a
cross-cleavage plane of biotite to a Sr-rich
solution. With a solid-state detector, the Cs
La line (at 4.286KeV) was a small shoulder on
the Ti Ka line (at 4.510KeV), making a map of Cs
concentration from the La intensity was
impossible. Mapping with the Cs K-edge was not
useful either (x-rays too penetrating into bulk
mica, and too much inelastic scattering). The
map at right shows the Cs concentration as
measured with the WDS on the Cs La line.
100 x 100mm image, with a 5 x 5mm beam, taking
3mm steps, with a 30s dwelltime at each point.
The incident x-ray energy was 10KeV.
16
Using the WDS for XANES 1000ppm Au in FeAsS
(arsenopyrite)
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.
As K-edge 11.868 KeV As Ka line 10.543
KeV Au LIII-edge 11.918 KeV Au La line
9.711 KeV
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.
17
Using the WDS for XANES 1000ppm Au in FeAsS
(arsenopyrite)
Louis Cabri (NRC Canada), Robert Gordon, Daryl
Crozier (Simon Fraser), PNC-CAT
With a 13-element Ge detector (at PNC-CAT
ID-20), the tail of the As Ka line was still
strong at the Au La energy, so the Au LIII
edge-step was 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, so the Au XANES was clearer.
Measuring two different natural samples of FeAsS,
both with 1000ppm of Au, we see evidence for
both metallic and oxidized Au.
18
Using the WDS for EXAFS Re in
K7ReOP2W17O61.nH2O
Mark Antonio (ANL)
The inorganic molecule a-P2W17O61 is a candidate
for stabilizing transition and rare-earth metal
ions. It can lose a WO ligand and replace it
with several valence states of Re (a nice, safe
chemical analog of Tc).
W LIII-edge 10.204 KeV W La line
8.396 KeV Re LIII-edge 10.534 KeV Re La line
8.651 KeV
The proximity of the Re and W LIII-edges, and
their La lines, and the relative concentrations
of Re and W (161) in this sample makes EXAFS
measurements using a solid-state detector nearly
impossible.
Venturelli, et al, J. Chem. Soc., Dalton Trans.,
p 301 (1999)
19
Using the WDS for EXAFS Re in
K7ReOP2W17O61.nH2O
Here are m(E), the EXAFS kc(k), and the Fourier
transform of the EXAFS c(R) for data collected
with the WDS. The data is the average of 3
scans, each having an integration time of 5
seconds per point. The data quality is
acceptable up to 12A-1, and initial analysis
supports a first shell with 4 oxygens at 1.8A.
20
Using a Wavelength Dispersive Spectrometer to
measure XAFS
The Wavelength Dispersive Spectrometer can be
used for XANES and EXAFS measurements. In some
cases it is sometimes the only detector capable
of such measurements. In many cases, the WDS
compares favorably with solid state detectors.
In some cases, the WDS is superior to
solid-state detectors, and is the only detector
capable of XRF, XANES, and EXAFS measurements.