Baixe Anatomy of an XAFS Measurement e outras Notas de estudo em PDF para Engenharia Elétrica, somente na Docsity! Anatomy of an XAFS Measurement Matt Newville Consortium for Advanced Radiation Sources University of Chicago Experiment Design Transmission v. Fluorescence modes X-ray detectors Data Collection Strategies Sample Preparation Problems to Avoid (version 1.1.1, 21-June-2004) X-ray Absorption Measurements XAS measures the energy dependence of the x-ray absorption coefficient µ(E) at and above the absorption edge of a selected element. µ(E) can be measured in two ways: Transmission: The absorption is measured directly by measuring what is transmitted through the sample: I = I0e −µ(E)t µ(E)t = −ln(I/I0) Fluorescence: The re-filling the deep core hole is detected. Typically the fluorescent x-ray is measured, but sometimes emitted electrons are mea- sured. Either way, µ(E) ∝ If/I0 Monochromator: Harmonic Rejection It is important to remove the harmonics (energies of 2E, 3E, . . . ) from the x-ray beam before it gets to the experiment. Two ways of to reject harmonics: Detuning: The angular width of a Bragg reflection is finite (a few µrad), and decreases with increasing energy: the harmonics have a nar- rower angular width than the fundamental energy. Making the two crystals slightly non-parallel rejects most of the harmonics, an preserves most of the fundamental. A piezo-electric crystal on the second crystal adjusts the parallel- ness, and so intensity, and harmonic content. Rule of thumb: Adjust the piezo until the total intensity is about half the maximum intensity. Harmonic Rejection Mirror: An x-ray mirror will not pass energies above a critical energy set by the pitch of the mirror (in mrad). This works very effectively for harmonic rejection. Use one or both of these methods! Monochromator: Energy Calibration and Reproducibility The absolute monochromator angle is usually not known to great precision (rela- tive changes are much more precise). We need to calibrate the energy for a particular edge. Typically, a metal foil is used and an arbitrary position on the edge (say, maximum of dµ/dE) is set to the tabulated edge energy. Many monochromators drift in angle from scan to scan (or over time). It’s good to measure a energy reference sample, such as a metal foil. The energy reference can be measured periodically or, in some cases, at the same time as the sample: Either of these approaches can be used to calibrate the edge energy during a normal XAFS scan. An XAFS Beamline End Station A typical XAFS station for Transmission and Fluorescence XAFS: Slits: To define beam size (just out of the picture). I0 Ion Chamber: To measure of the incident x-ray intensity I Ion Chamber: To measure of the transmitted x-ray intensity Fluorescence Detector: To measure fluorescence signal. Motorized Sample Stage: To move sample into beam. X-ray Absorption Measurements: The Experiment Energy Scanning The monochromator gives an energy-tunable x-ray source, so we can scan energy across the absorption edge. We’ll scan from ∼ 200 eV below to ∼ 800 eV above the selected edge energy E0, like this: Region Starting Energy (eV) Ending Energy (eV) Step Size (eV) Pre-edge E0 − 200 E0 − 20 5.0 – 10 XANES E0 − 20 E0 + 30 0.25 – 1.0 EXAFS E0 + 30 E0+ ∼ 800 0.05 Å−1 • In the EXAFS region, it’s common to step in k rather than energy. • Typical count times are 1 to 15 seconds per point, so that a spectrum is collected in 10 minutes to several hours (dilute samples take longer than concentrated samples!). Multiple sweeps is common. • Very fast measurements (1 second for the whole spectra) can be made at specialized beamlines. X-ray Absorption Measurements: Transmission For concentrated samples, XAFS is best measured in transmission. But we need enough transmission through the sample to get a decent sig- nal for I . The sample thickness t should be chosen so that µ(E)t ≈ 2.5 above the absorption edge and/or the edge step ∆µ(E)t ≈ 1. The sample must be uniform, and free of pinholes. For a powder, the grain size should not be bigger that an absorption length. If a transmission measurement can be made, it is easy and gives excellent data. It’s usually appropriate for compounds with element concentrations > 10%. X-ray Transmission Thicknesses Typical X-ray Absorption Thicknesses (to give t∆µ = 1): material edge 1/∆µ ( µm) Fe foil Fe K 3.6 Fe2O3 Fe K 6.8 Pb foil Pb LIII 5.5 An absorption length is very small. We need a uniform sample of these thicknesses, and free of pinholes. This is not always easy! Calculating absorption lengths: Absorption cross-sections σ(E) are tabulated in the McMaster tables (online or book form) in barns/atom (1barn = 10−24cm2). From a material’s chemical for- mula and density ρ (in gr/cm3), an absorption length (in cm) is given by: t = 1/µ = 1.66 P i niMi ρ P i niσi(E) sums are over the elements i in the chemical formula, ni the elemental stoichiom- etry , Mi is the atomic mass (in amu), and σi(E) the cross-section. Getting both t∆µ > 0.1 and tµ < 4 are important! Fluorescence Measurement Considerations There are two main considerations for fluorescence measurements: Energy Discrimination either physically or electronically filter the un- wanted portions of the fluorescence/scatter spectra. Solid Angle fluorescence is emitted isotropically in all directions, so we collect as much of the solid angle as possible. Fluorescence measurements are better when the samples are: Uniform not as stringent as for transmission, but a uniform sample gives better data. Thick, Dilute Samples The element of interest should be below ∼ 10 wt. % of the sample. Measurements are possible for more concentrated samples, but the data may need corrections. Very Thin Sample Alternatively, the sample can be much thinner than one absorption length. The x-rays from a synchrotron are polarized in the horizontal plane. This means the elastic scattering is zero at 90◦ to the incident beam and in the horizontal plane. Putting the fluorescence detector there minimizes the “scatter peak”. Fluorescence Measurements: Z-1 Filter A simple method of energy discrimination uses a filter of “Z-1” from the element of interest. For Fe, a Mn filter absorbs the scatter and Fe Kβ line while passing the Fe Kα line. This method can be used with a detector without any energy resolution, such as a fluorescence ion chamber (aka Stern-Heald chamber or Lytle chamber). Edge / Line Energy (eV) Edge/Line Energy (eV) Fe Kα 6405 Fe Kβ 7059 Mn K edge 6539 Fe K edge 7112 Fluorescence Ion Chamber, Z-1 Filters, Soller Slits A typical setup with a ’Z-1’ filter uses an ion chamber which has no en- ergy resolution, but high count rate and linearity. Because the filter absorbs the scat- tered beam, it can it- self re-radiate!! A set of Soller slits can be used to see the sample, but absorb most of the re- radiate scatter from the filter. This arrangement can be very effective especially when the signal is dominated by scatter , and when the concentration is at per cent levels. Fluorescence Measurements If = I0 ∆Ω 4π µχ(E) ˘ 1− e−[µtot(E)/ sin θ+µtot(Ef )/ sin φ]t ¯ µtot(E)/ sin θ + µtot(Ef )/ sin φ Thin Sample Limit: (µt 1). The 1 − e−µt term becomes ≈ [µtot(E)/ sin θ + µtot(Ef )/ sin φ] t (by a Taylor series expansion), which cancels the denominator, so that If ≈ I0 ∆Ω 4π µχ(E)t Thick Sample Limit: (µt 1), the exponential term goes to 0, and If = I0 ∆Ω 4π µχ(E) µtot(E)/ sin θ + µtot(Ef )/ sin φ Thick, Dilute Limit: (µχ µother , µχ µtot), and we can then also ignore the energy dependence of µtot, so that If ∼ I0µχ(E) These two limits (very thin or thick, dilute samples) are the best cases for fluores- cence measurements. Fluorescence Measurements: Self-Absorption Thick Sample Limit: (µt 1): If = I0 ∆Ω 4π µχ(E) µtot(E)/ sin θ + µtot(Ef )/ sin φ Thick, Concentrated Limit: (µχ ∼ µother , µχ ∼ µtot). We cannot ignore the energy dependence of µtot, and must correct for the oscillations in µtot(E). This is generally called self-absorption can dampen the XAFS – even completely wiping it out for highly concentrated elements. For very concentrated samples, it’s best to avoid this problem and measure trans- mission. For moderately concentrated samples, the self-absorption effects can be corrected. Grazing Exit Limit: (φ → 0). One way to reduce the self-absorption effects is to rotate the sample so that it is normal to the incident beam. With φ very small, so that µtot(Ef )/ sin φ µtot(E)/ sin θ, and If ≈ I0 ∆Ω 4π µχ(E) µtot(Ef )/ sin φ which gets rid of the energy dependence of the denominator. Should I measure Transmission or Fluorescence? The choice of Transmission v. Fluorescence depends solely on the sample, and particularly the concentration of the element of interest. Transmission Samples Concentrated samples – element of inter- est is above ∼10 wt. %, and a thin enough sample can be made. Sample preparation is more stringent for transmission mea- surements than fluorescence measurements. Fluorescence Samples Dilute samples – element of interest is be- low∼10 wt. %. For concentrated samples that cannot be made thin enough for transmission, use fluorescence, but pay attention to self- absorption effects and consider using grazing-exit geometry. Also consider Electron Yield: Like fluorescence, but measures electron current emitted from the sample surface. This works best for metallic samples. Because the electrons are emitted from a few 100Å from the surface, this is very surface sensitive, but is always in the Thin Sample limit, so there are no self- absorption effects. Transmission XAFS Sample Preparation First, this does not require great laboratory skills, but does require a mor- tar and pestle and a lot of tape! XAFS requires scotch tape, double-sided scotch tape, and Kapton R© tape (the “duct tape of the synchrotron”). This plastic (polyimide) is strong and unusually radiation resistant. It comes in plain film (7 µm to 75 µm or thicker) and tape form. Sample Preparation: Diluted Powder We start with a blank sample holder with known dimensions, and cover one side with tape. We calculate the necessary mass of material (from absorption thickness, sample holder cross-section, and sample density). The powder should be ground to give very fine particle size (< 1/∆µ!!). Many people use a sieve to separate particle sizes. A 400-mesh sieve has wires ∼ 37 µm apart. That’s a good start, but not fine enough for most transmission XAFS samples. Sample Preparation: Diluted Powder We dilute with a low-Z material (sugar, BN, graphite, B(OH)3, duco ce- ment, . . . ) so the remaining volume of the sample holder is filled. The sample and matrix are then ground together in a mortar and pestle and mixed thoroughly. Sample Preparation: Powder on Tape (2) • Place a piece of Scotch Tape sticky-side up on the bench, and tape it down, covering all four sides of the exposed tape. • Paint the powder on the sticky tape. The smallest grains will stick to the tape, which is what you want – try to remove as much excess as you can, so that there are few large particles. Sample Preparation: Powder on Tape (3) • finish painting the sample on the tape • cut tape into small pieces. Sample Preparation: Powder on Tape (4) • place another piece of tape on the bench, sticky-side up. • move small pieces of painted sample onto tape, stacking 4 to 8 layers. • cover with another piece of tape This method works great for metal oxides and other powders needed for EXAFS and XANES standards. It can be done in a glove box, but samples in scotch tape can still react with air.