Does Beam Size Matter?

[Probeman]

We all know that if the beam size is defocused too much, we will see a drop in intensity related to Bragg Defocusing...

However before that point is reached, I wonder if there is anything else we might detect due to sub-surface charge distribution as the beam size is increased. Obviously the TiO2 is the only non-conductor measured, but I did pure Ti and Ag coated Ti also. The following Ti Ka measurements where all performed at 15 keV...

First is a plot of uncoated pure Ti metal at a variety of beam sizes:



Next a plot of Ti coated with 120 nm of Ag (I do realize that I need to do another measurement of Ag La too!):



Finally a plot of carbon coated TiO2:



We would expect (hope) that beam size does not affect the analysis, and to a reasonable precision that seems to be the case.  More measurements would be needed to decide if any of these trends are significant and/or interesting...

[Probeman]

Unfortunately I was in a bit of a rush last night and only digitized a single point for each beam size, so I'm re-doing the measurements tonight with more points... but the results for the Ag (coating material) La x-ray line, seem to indicate that all is well by assuming that beam size does *not* matter for conductive bulk and thin films and substrates to around 0.2% precision levels:



This is consistent with what we know from electron-solid physics, but insulating materials may tell a different story. So I also want to try with some good insulators such as Al2O3 that are not beam sensitive.

[John Donovan]

Ok, I re-ran the 120 nm Ag on Ti with 5 points each at beam sizes ranging from 0 (focussed) to 25 um and here are the results for Ti ka and Ag La:



So it would appear that beam size does not affect intensity in conductive materials, although there may be a slight drop off in intensity at the largest beam sizes (20 to 25 um) for Ti ka due to Bragg defocussing. I hope to have results for Mg Ka in MgO and Al Ka in Al2O3 tomorrow...

[Probeman]

Finally here are the results for insulating materials measured over a range of beam sizes. I choose synthetic Al2O3, synthetic "ruby" (Al2O3 plus 0.4% Cr) and synthetic MgO. The idea being that pure Al2O3 and Al2O3 doped with 4000 PPM of Cr will have very different electrical conductivity (does anyone know the actual electrical conductivities?). These effects could possibly produce a difference in x-ray intensity due to differences in sub surface charge distribution. Let's test that idea...

Using a Cameca SX50, the O Ka was measured on a PC1 crystal, the Al Ka and Mg Ka on a TAP crystal and the keV was 15 and beam current 30 nA. All materials were carbon coated with 20 nm of carbon and each measurement was performed on a clean area.

The results for synthetic Al2O3 are here:



As you can see except for a small intensity drop for Al Ka and O ka at the largest beam sizes of 20 and 25 um, the intensities seem stable over the range of beam sizes. This is not unexpected due to Bragg defocusing effects.

So are we done? Well I thought so until I saw the next plot of synthetic "ruby" (Al2O3 with 0.4% Cr):



What the heck is going on with the O Ka intensities in "ruby"?

A couple of things to note: first the Al Ka intensity in "ruby" seems consistent, but the O Ka intensity starts at the smallest beam sizes with a significantly higher intensity than pure Al2O3, but by the time the beam is defocused to 20 or 25 um, the intensity is back to that we obtained from pure Al2O3.

I also note that the Cr Ln lines is close to the oxygen line, but why would that make the intensity higher at smaller beam sizes? I am running some wavescans at different beam energies now to look at the peak shapes.

Oh, and the MgO plot? That is attached below. It shows a similar effect for oxygen as the Al2O3 with Cr (but not for Mg Ka!), and yes, the Mg Ka II order in close to O Ka but not that close...  so what is going on?

[John Donovan]

So I ran wavescans at 8, 16 and 22 keV on Al2O3 and Al2O3 + 0.4% Cr, but I'll be darned if I can see any difference that would cause the intensity to change as beam size is increased on Al2O3 + Cr but not on Al2O3!

I also double checked and my beam center was moving only a few microns when changing the beam size from 0 to 25 um.

[Probeman]

Zack Gainsforth thinks this might be related to channeling effects and I think he may be right. Here is some more data which may indicate that grain orientation, not composition, is a contributing factor. Look at these graphs and then consider that the Al2O3 + 0.4% Cr is a single crystal, but the pure Al2O3 is a polycrystalline material:

Here is pure Al2O3 (on a different grain than above):



Here is Al2O3 plus 0.4% Cr (same single crystal):

[John Donovan]

I really don't want to spend much more time on this because I suspect the effects I'm seeing are partly due to beam alignment issues and also possibly carbon contamination issues as well.

But I ran some acquisitions on the Sx100 to compare to the Sx50 and the results are interesting if fairly ambiguous!

Let's start with the multiple size analysis (MSA) on Al2O3 on the SX50:



and here run on the Sx100, but with the acquisition order reversed so the large beam sizes get run first:



Now the Al2O3 _ 0.4% Cr on the SX50:



and the same on the Sx100 again in reverse acquisition order:



and now TiO2 on the Sx50:



and again, TiO2 on the Sx100, again in reverse acquisition order:



I think this is enough for now (I also have MSA acquisitions for SiO2, MnO, CoO and NiO), but if someone is still interested in pursuing this further just let me know.

[Les Moore]

A quirky thing I noted in the past is that C deposition on a refractory type material lays down a donut shaped deposit but on a steel lays down a gaussian shaped blob as you would expect. I believe that for refractory materials, the spot gets too hot to deposit C and it deposits on the region around the spot.  For Steels, the conductivity is sufficient for the spot to remain cool enough to deposit C.

I did maps of these and will post if I can find them.

This gets me to thinking are you in fact using a million dollar carbon coater?  The deposit would increase in the space to be mapped and thus provide an increasing attenuation as the map progresses.
1. Map for C (during and after run)
2. Use an oxyjet & cryo finger.