AW: Questions regarding simulation of ABF-STEM

[Christoph Koch]

Hello,

 

Thank you for your question. I hope you are OK with me copying my anonymized reply to the qstem mailing list, since I have seen other people having similar questions.

 

The detector angles  in the simulation should correspond to exactly those angles your detector collects. To know what exactly those angles are you need to make sure that your microscope’s detector is properly calibrated. Some detectors also have different sensitivity for different scattering angles. This variation is not (yet) included in qstem. However, you could include those as well, by setting the save level to 2, and multiplying the generated diffraction patterns for each probe position by the detector sensitivity you have obtained by one of the various approaches to obtain this information (see work by the groups of J. LeBeau or A. Rosenauer, and also others).

 

However, I noticed that you are defining a HAAD-detector that collects scattering up to 200 mrad, but you set the resolution in your probe window such that the maximum scattering angle is only 130 mrad. This means that qstem will not compute any scattering into angles beyond 130 mrad. At small specimen thicknesses this may not be very critical, but at larger specimen thicknesses more electrons are scattered into high angles. This may cause part of the discrepancy between simulation and experiment and you are observing.

 

I suggest that you increase the number of pixels for your probe array while keeping the field of view (window size) constant. For Cs = 1 mm you will need to have a window size not much smaller than 20 A. You may verify whether this is the proper setting by doing a CBED calculation and look at the wave function and diffraction pattern for the thickness you expect your sample to have.

 

Using the simulation parameters you have used, I see strong aliasing artefacts. You can check those by looking at the probe generated using your parameters (click on “higher orders)”:

 

This probe looks so funny and has square symmetry, because of the large convergence angle you are using and the resulting undersampling of reciprocal space. This leads to aliasing in reciprocal space. You can check this by looking at the phase map that is shown to the right of the probe intensity:

 

If you define the proper convergence angle that leads to the minimum probe size (at 300 kV and Cs = 1 mm, alpha should be about 10 mrad), things will look much more reasonable:

 

If you really want to use 18 mrad convergence angle, you have to increase the number of pixels and increase the window size (decreases the reciprocal space resolution in the probe array). Here I have used 1024 pixels, and a field of view of 60 A:

The aliasing artefacts which make the probe square are now gone (I zoomed in to show the details better).

 

Now regarding your questions:

1.      Detector angles in ABF-STEM: These detector angles depend on the convergence angle of your probe. If I remember correctly, the outer angle in ABF-STEM should coincide with your convergence angle.  Also, 18 mrad seems a very high value for a non-corrected microscope (Cs = 1).

2.      The resolution is determined by how many pixels you use to sample the size of the probe array (window size). And the pixel size also determines the scattering angle (inversely proportional to the pixel size). It only makes sense to define detectors up to the maximum scattering angle, since beyond that there will be no electrons in th simulated diffraction patterns.

3.      I suggest to look into the literature for optimum defocus values for STEM experiments. For HAAD-STEM experiments, you probably chose a defocus that maximizes your contrast. You can also check by simulation what this defocus would be. In BF-STEM Scherzer defocus might give you the most interpretable image. But I always suggest to record a focal series.

 

With best regards,

Christoph.

 

Von:
Gesendet: Donnerstag, 20. Juli 2017 13:47
An: christo...@physik.hu-berlin.de
Betreff: Questions regarding simulation of ABF-STEM

 

Dear Prof. Koch,

 

My three main questions are :

• How do the detector collection angles in the software correspond to real life scenario?
• ABF-STEM normally has a collection angle range around 10-20 mrad, but I am getting contrast more similar to MAADF-STEM, odd contrast change with increasing thickness.
• How does resolution/scattering angle affect the collection angles?
• What is the ideal defocus value for STEM image simulation?
• Different for ABF & HAADF?
• Dependence on thickness?

 

Thank you very much in advance.

 

Best regards,

…