|Date:||16-19 September 2010|
|Beamline:||I22 at Diamond|
|Team on site:||Simon Cooil, Rudi Winter (Aber)|
|Marc Malfois, Tobias Richter, Nick Terrill (I22)|
|Instrument:||X-ray energy: variable, 16.7-18.2keV|
|Detectors: Hotsaxs 2D,|
Vortex energy-resolving point detector
|Sample env.:||Ystwyth dipper - dip coater with hot-plate and furnace|
|Samples:||Silicon wafers coated in zirconia or YSZ gel|
Scattering patterns from chemically complex multi-phase materials are notoriously difficult to interpret because of the variety of electron density contrast of adjacent phases present in the sample. In previous experiments, we have used anomalous scattering to introduce chemical contrast to "label" the individual contributions, which makes model fitting lot less ambiguous.
Anomalous scattering is based on the fact that the scattering factor is modified near an absorption edge, giving a sharp resonance at the edge. Experiments are typically conducted at well below the edge and at two energies within the resonance. This can be problematic for in-situ studies of chemical processes because the exact position of the energy can shift, leaving the exact contrast undefined in the course of the reaction.
In this experiment, we have tried an alternative approach to chemical contrast: Instead of taking scattering patterns at various energies below the edge, we take one on either side of the edge at some distance in energy. This way, a small chemical edge shift will cause only a negligible change in contrast, and the difference in absorptivity above and below the edge will cause a large contrast for features involving the edge element.
The downside of this approach is that absorbed photons cause fluorescence. This results in a significant angle-independent background to the scattering patterns taken above the edge. However, since fluorescent x-rays have a lower energy than incoming ones, the extent of the background can be measured with an energy-sensitive point detector and corrected for. This experiment is meant to prove the concept.
In this experiment, we used the Ystwyth Dipper, our home-grown dip coating instrument, which cycles a sample through a loop of dipping, firing and SAXS experiment. We set the beamline to take 2 second exposures continuously. The figure shows the integrated SAXS detector count, showing how the scattered intensity increases as subsequent layers of coating are added. The contrast is reduced each time a new dip is applied and before it is fired, presumably because the liquid fills some of the porosity of the previously calcined film. The empty-beam background can be seen while the sample is on its way to the furnace and back. Periods with a zero reading are during the dipping part of the cycle, when the sample support is in the beam.
The fluctuations while the sample is in the same state aren't due to noise but largely due to the varying contrast as the x-ray energy is changed. Since the dipper and the monochromator loops aren't synchronised, the order in which the energies occur isn't the same in each group of measurements.
This two-dimensional fluorescence spectrum shows how the fluorescence builds up as the energy of the incident x-rays increases (top to bottom) across the Zr K edge. The two spectral lines visible are the Kα and Kβ emission lines, while the cutoff on the right reflects the energy of the incident photons. The fluorescence detector is placed 720mm away from the sample in 52o backscatter geometry. The colour scale is logarithmic.
Since the SAXS detector is not energy sensitive, we can use the relative increase in the fluorescence line as measured by the energy-resolving point detector as a measure for the fluorescent background in the scattering patterns.
The figure below shows five uncorrected SAXS patterns around the Zr K edge, with the sample in the same state for all of them. The increase of the background due to fluorescence as a function of energy is evident. The colour scales are logarithmic.
An anomalous SAXS experiment would use the first three patterns to determine the chemical contrast in the system. In this case, we need the first and last one, after correcting for fluorescence, and use their absorption contrast instead.
The experiment demonstrates the principle of absorption-contrast scattering. In a follow-up experiment, we need to bring the fluorescence detector closer to the sample and ideally at 90o from the through beam. We also need to synchronise the data acquisition with the dipper cycle to ensure we have equivalent data for all energies for a given sample condition.