Why isn't my flat-field flat?

The main reasons of a non flat flat-field can be:

• The scattering from the glass rod is not uniform over the angular range. In this case you should take the flat field dynamically i.e. scanning the detector in front of the cylinder with the small window, as we do at the SLS. In this case when you shift the detector, the shape of the illumination remains in the same angular position (and shifts in channel number). Of course it depends a lot on the energy and on the geometry of the flat field acquisition.

  Figure 3.5: Example of a very bad flat field data set with highlights of some of the reasons which can cause the non-flat behavior for the MYTHEN detector. Similar effects can be visible also in 2D.

  

• The entrance window for the X-rays is deformed (we also have this problem at the SLS). In this case when you move the detector the "mountain" moves with it in angle (And remains still in channel number). However this should correct without problems with the flat field correction, even in case of fluorescent emission. Should appear at all energies.
• Differences of efficiency between the modules i.e. mainly bad energy calibration. You normally see really steps at the transition between modules. Sometimes you have some groups of strips withing a module that are not properly trimmed and look as smallish peaks or valleys in the flat field. When you move the detector, these steps or peaks move in angle and remain still in channel number. These differences can slightly change as a function of the energy (probably more evident at lower energies) but should normally always be there for the same settings. These differences get much worse in presence of fluorescent emission, but normally correct properly with flat field correction.