ASAXS - In Depth -Part 3 - setting up an in-situ

high temperature ASAXS experiment.

This section is more of a recipe page rather than a descriptive page. What I shall do is explain how I conducted my high temperature in-situ ASAXS experiments on samples of alumina-zirconia-silicate nanoceramics, and also some of my rational behind each step. Some people might consider the steps I take to be a slight “overkill”, however it has been my experience in the past that it is better to be over cautious than reckless, especially when trying to conduct extremely tricky experiments such as this.

The experiment I shall use as my example was conducted on beamline 7T-MPW-SAXS at BESSY, Berlin. This beamline uses synchrotron radiation from a 7T multipole wiggler with 13 poles. At the beginning of the optics hutch the beam dimensions are 25m x 1.2mm. The x-ray beam is then passed through a mask and a pre-filter to provide a beam fan of 0.3 x 3 mrad2 . This is then passed through a set of four separate water cooled Cu blades which act as slits to define the beam profile as it hits the first mirror. The first mirror is a collimating mirror, and is a large rhodium coated Si crystal (1200mm x 160mm x 60 mm2) with an active length of 1100mm, which produces a vertically parallel beam.

The beam is then passed to a double Si(111) crystal fixed-exit monochromator. The first crystal is water cooled with a Frahm type bending mechanism to iron out any thermal effects. The second crystal has an ESRF type sagittal bending mechanism allowing for horizontal focusing. The monochromator provides photons in the 3-18keV range using mirrors, and up to 30keV without mirrors (and as such is vertically unfocused). The beam then passes through another slit system which defines the cross-section of the beam hitting the second mirror and eliminates radiation scattered by the upstream components. The beam is then passed to a downward deflecting focusing mirror and then onto a final set of slits which reduces the background from the last mirror. Scattering patterns are recorded using a two dimensional position sensitive gas detector which has dimensions of 200 x 200mm2 and a resolution of 200m. The gas chamber is a 2D multi-wire gas position sensitive detector with 1024 x 1024 pixels, filled with 1.6 bar xenon/ethane mixture. It is possible to change the sample to detector distance to allow some variance of the possible scattering vector range.

PIC LEFT: The on-line furnace mount with the sample holder taken out to show the heating elements.

PIC RIGHT: The on-line furnace with the sample holder inserted and the water cooling pipes attached.

In both cases x-ray are incident on the sample from the right of the picture and exit in transmission to a 2D gas detector at the end of an evacuated flight tube to the left.

It is important to consider what exactly you want from your experiment when deciding the type of furnace to use. Ramp rate is important for experiments requiring instantaneous heating, whilst location of thermocouples and temperature stability is important for slow heating experiments. I have found that the furnace used on 7T-MPW-SAXS is a particularly good furnace, designed by the university of Rostock, providing a near infinite ramping rate, whilst also providing excellent temperature stability at high temperatures, (this furnace has been specifically designed for high temperature work and as such the temperature stability at low temperatures is expectedly shakey).

Since we are using a furnace, one should use the furnace mounted in place, with the sample holder in for all background measurements. For completeness I would also recommend taking background measurements at the working temperatures of the experiment, just to check for any issues with evaporation off the heating elements from previous users of the furnace, (eg sodium deposits on the heating elements).

Dark currents should be taken in the normal manner, and the optics set-up to maximise the beam intensity at the chosen edge energy - in our case we used the Zr K-edge which is at 18keV, (17998eV for Zr-foil).

The choice of which energies to conduct the ASAXS experiment now rests largely on the beamline optics and the sample under investigation. It is necessary to use three energies, two close (but below) to the edge position, and one well below. The energy of the “well below” should be sufficiently far away from the edge position that it is in the flat portion of the f ` curve. How far away you can go is largely based on the beamline optics, in particular the monochromators rocking angles and stability. I would have a very good chat with the beamline scientist about this before choosing the energy.

As for the two near (but below) the edge. This is where I have found problems in the past. It has been common practice to choose the energies based on equal increments of f ` based on the theoretical values provided by Cromer and Lieberman. However, unless you are dealing with metals there will always be a chemical shift between the sample and the Zr-foil values, and also it is highly likely that there will be a change in the f ` resonance minimum width.

PIC LEFT: This is the f ` minimum provided after various corrections (please contact me for details) and a Kramers-Kronig transformation from transmission measurements of the sample.

One can see from the diagram that there is a significant reduction in the width of the f ` resonance minimum between the sample (green) and the theoretical values (red).

I choose to go as close to the resonance minimum position (the edge) as the beam spectrum will let me go (delta E), this is typically 1eV. Then I choose halfway up the resonance minimum and then the final energy is the one far away. This then gives me the biggest spread in f ` and f `` values.

It is then important to understand how this edge shape at position changes during the evolution of the sample given some evolution parameter (temperature, time, or both). If you have enough sample I would suggest conducting an experiment, concentrating on taking edge scans (look at sample transmission as a function of energy across the edge) at values of the evolution parameter which are of interest in the final experiment. I would then plot the edge position and shape as a function of the evolution parameter. This will then let me know what energies I should use at each point during the experiment to provide the maximum ASAXS contrast. If you are really lucky neither will change, but I have seen the edge shift as much as 8eV in the past!

Transmission measurements at each of the energies used to take scattering patterns must be included in the experimental ASAXS loop in order to correct for the changes in sample transmission during the evolution of the sample. This measurement is then used to correct the transmitted scattering intensity to a value comparable to the incident beam intensity.

During the final analysis of the scattering patterns it is important to correct for the changes in intensity due to the energies used. This can be done by using a known scatterer, such as glassy carbon, such that the intensity differences can be normalised. The full amount of corrections are - Dark current, background scattering, transmission, beam current, scattering intensity.

The experimental ASAXS loop should include the transmission measurement, and at least three ASAXS energies. This experimental ASAXS loop should be done for each point of the evolution parameter under observation, e.g every X degC, or Y minutes.