Nasty Boundary Fluorescence Analytical Situations

[Anette von der Handt]

As always, the Edinburgh SIMS lab webpage on epoxies is useful here. They don't have absolute Cl concentrations but they identify the epoxy resins with elevated Cl contents.

https://www.ed.ac.uk/geosciences/about/facilities/all/ionprobe/instrument-capabilities-and-sample-requirements/specimen-requirements/epoxy-resins/results/compositions

It is not trivial in some cases. I definitely ran into this problem when doing analyses of apatite grain mounts where the apatites would get more Cl-rich towards the rims, although that was probably more a problem of mixed analyses than secondary fluorescence.

[Probeman]

Interesting.

Any ideas on how we could measure some epoxy Cl concentrations?

[Probeman]

The Athens 2022 EPMA just concluded and there were questions about how to correct data for modeled secondary fluorescence effects.

This post here provides one example:

https://probesoftware.com/smf/index.php?topic=58.msg5603#msg5603

It should be noted however, that ideally one should perform a matrix correction after the concentrations have been modified after this subtraction as the "apparent" concentration of the fluoresced element is of course not actually present in the matrix. However, so long as the subtracted concentration is a fraction of a percent the effect should be minimal, particularly on the measured element itself.

In fact the CalcZAF software actually includes this matrix iteration in the secondary fluorescence correction based on defined distances from the boundary. See here for details:

https://probesoftware.com/smf/index.php?topic=58.msg223#msg223

Unfortunately this SF subtraction/iteration GUI not yet implemented in the Probe for EPMA software, though soon perhaps.

But it should also be pointed out that the modeled secondary fluorescence concentrations from Fanal or Penepma are only accurate for EDS detectors (without any Bragg defocusing) and WDS detectors in the two directions parallel to the spectrometer orientation. This is described by Ben Buse in the paper linked to this post here:

https://probesoftware.com/smf/index.php?topic=1447.msg10718#msg10718

To avoid this WDS defocusing effect on the secondary fluorescence from a boundary, one should orient the sample so the boundary is parallel to the spectrometer orientation, hence one of the reasons we are trying to document this spectrometer orientation in JEOL and Cameca microprobes.

[Probeman]

Llovet et al. (2022), just published a new paper on correction of secondary boundary fluorescence effects near grain boundaries. 

If you ever perform trace/minor element analysis near a boundary, then you should read this new paper attached below (login to see attachments).

I also include one other earlier paper on this subject that is referenced in the above new paper.

[Yishen Zhang]

Hello all,

I went through the SF correction and have successfully modeled my own standards file (so many hours!)

However, from what I understand, to do SF correction for my own data would require a DAT file that can be easily done with Probe for EPMA? unfortunately our probe doesn't install this but could only export csv file, and I also always copy the raw files (.cor, .wt, .qnt, and .cnd) file. I was wondering in this case can I still do the SF correction with these files or I would have to install Probe for EPMA in our probe PC?

Thanks in advance,

Eason

[Probeman]

I went through the SF correction and have successfully modeled my own standards file (so many hours!)

So you have run your standards to create .PAR files for your standard compositions using Standard.exe?  Nice!

Note that for the purposes of simply modeling a boundary fluorescence to get an idea if there is a significant boundary fluorescence issue, you can just utilize the closest compositions you can find that are already modeled.  But it's fine to model your own standard compositions of course, though many/most/some will be pretty much identical to the compositions already provided.

But for the purposes of performing a boundary fluorescence correction you need to model the actual compositions of the boundary in your unknown samples.  You will need to model both the beam incident and the boundary materials from your probe analyses. 

Now it may be that your standard compositions are very similar to your unknown boundary materials, e.g., SiO2 adjacent to TiO2.  In that case you are all set to proceed with the boundary fluorescence correction process.

This procedure is explained here:

https://probesoftware.com/smf/index.php?topic=58.msg214#msg214

However, from what I understand, to do SF correction for my own data would require a DAT file that can be easily done with Probe for EPMA? unfortunately our probe doesn't install this but could only export csv file, and I also always copy the raw files (.cor, .wt, .qnt, and .cnd) file. I was wondering in this case can I still do the SF correction with these files or I would have to install Probe for EPMA in our probe PC?

It should be pointed out first of all, that if your boundary fluorescence effect is relatively small, e.g., a trace or minor level fluorescence, then one can simply subtract the fluorescence profile from your analysis measurements. There is no need for a full matrix correction of the composition after subtracting the modeled boundary fluorescence, since you will be subtracting such small concentrations from the matrix. You simply need to relate the analysis points to a linear distance scale in microns.  An example of this is found here:

https://probesoftware.com/smf/index.php?topic=58.msg5603#msg5603

However, if your boundary  fluorescence is a significant contribution to the matrix effects, say greater than a few weight percent, you might want to perform the fluorescence correction in CalcZAF with an iterative matrix correction.  This is described here:

https://probesoftware.com/smf/index.php?topic=58.msg223#msg223

I don't know if the files you describe above are JEOL or Cameca files, but in any case you will need to create your own CalcZAF input files containing the k-ratios (and stage coordinates) in order to perform this boundary fluorescence correction in CalcZAF.

Maybe someone on this forum has a application which can read these OEM files and output a CalcZAF input file format? In any case, here is a post which describes the CalcZAF input file format:

https://probesoftware.com/smf/index.php?topic=81.msg9330#msg9330

Good luck and please feel free to share your results with us.

[Probeman]

This relatively recent paper by Llovet et al.  is a very complete treatment of secondary fluorescence near mineral boundaries:

https://www.cambridge.org/core/journals/microscopy-and-microanalysis/article/abs/correction-of-secondary-fluorescence-across-phase-boundaries-in-electron-probe-microanalysis-of-mineral-inclusions/788EBDD9322BEE557AB4E3F39C28C09D

Xavier Llovet with Cesc Salvat developed the PENFLUOR and FANAL software to quickly model secondary fluorescence across boundaries and is implemented with an easy to use GUI in the CalcZAF/Standard open source package distributed by Probe Software and can be downloaded here:

https://probesoftware.com/download/CalcZAF.msi

And instructions for generating these models are found here:

https://probesoftware.com/smf/index.php?topic=58.msg214#msg214

It should be noted that although these boundary fluorescence effects are easily modeled in the Standard application, EPMA (or EDS) measurement data can also be corrected for these secondary fluorescence effects using numerical or graphical methods in the CalcZAF application as described here:

https://probesoftware.com/smf/index.php?topic=58.msg223#msg223

[Max Gavrilenko]

This new paper "Secondary fluorescence effect quantification of EPMA analyses of olivine grains embedded in basaltic glass" makes life easier for experimental petrologists who deal with trace element chemistry of synthetic olivine crystals:

https://www.sciencedirect.com/science/article/abs/pii/S0009254123000281

A new correction algorithm (a system of analytical expressions), integrated into a single MS Excel file (attached here), makes it possible to correct EMP analyses (Ca, Ti, Al) of olivines embedded in a basaltic glass matrix.

The only assumption is real olivine grains have to be equant (+/-), close to the hemispherical (1/2 buried sphere) geometry used in PENEPMA modeling.

[Ben Buse]

Fast fluorensence check using FANAL and the par files in compound which derive from the default standard.mdb materials, seem to be missing a spinel MgAl2O4, any plans to add one.

I will have to simulate my local standard   ,

Are Osc Strength & Osc Energy zero?

[Probeman]

Fast fluorensence check using FANAL and the par files in compound which derive from the default standard.mdb materials, seem to be missing a spinel MgAl2O4, any plans to add one.

I will have to simulate my local standard   ,

Are Osc Strength & Osc Energy zero?

Hi Ben,
Sure, there are lots of compounds missing.  The ones provided are just a sampling of what is possible. 

As for the Oscillator Strength and Oscillator Energy, always set those to zero.  I believe they get calculated automatically if they are zero. At least that's what I think Cesc Salvat told me once.

[Ben Buse]

Thanks John

[Probeman]

Once you've created a .MAT file, just look at any .MAT file using a text editor. These .MAT file are found in the Penepma12\pendbase folder, and you'll see something like this (in this case a NiP compound):

PENELOPE (v. 2012)  Material data file ...............
 Material: NiP                                                           
 Mass density = 8.12000000E+00 g/cm**3
 Number of elements in the molecule =  2
   atomic number = 28,  atoms/molecule = 1.00000000E+00
   atomic number = 15,  atoms/molecule = 9.99658540E-01
Mean excitation energy = 2.53469861E+02 eV
 Number of oscillators = 13 (E/P inelastic model)
   1  4.99897562E+00  0.00000000E+00  1.93895546E+01   0  30
   2  9.99931708E+00  1.13766238E+01  3.05590405E+01   0  30
   3  4.00000000E+00  7.30000000E+01  1.84760010E+02  28   7
   4  2.00000000E+00  7.50000000E+01  1.89528229E+02  28   6
   5  2.00000000E+00  1.17000000E+02  2.95420833E+02  28   5
   6  3.99863416E+00  1.34000000E+02  3.38447105E+02  15   4
   7  1.99931708E+00  1.35000000E+02  3.40821399E+02  15   3
   8  1.99931708E+00  1.91000000E+02  4.82095035E+02  15   2
   9  4.00000000E+00  8.60000000E+02  2.17026773E+03  28   4
  10  2.00000000E+00  8.77000000E+02  2.21314391E+03  28   3
  11  2.00000000E+00  1.01500000E+03  2.56138568E+03  28   2
  12  1.99931708E+00  2.14800000E+03  5.42051604E+03  15   1
  13  2.00000000E+00  8.33800000E+03  2.10410571E+04  28   1
 Number of shells = 13 (Compton IA model)
   1  4.99897562E+00  0.00000000E+00  6.85706871E-01   0  30
   2  9.99931708E+00  1.13766238E+01  4.58123112E-01   0  30
   3  4.00000000E+00  7.30000000E+01  2.02278000E-01  28   7
   4  2.00000000E+00  7.50000000E+01  1.99846000E-01  28   6
   5  2.00000000E+00  1.17000000E+02  3.20679000E-01  28   5
   6  3.99863416E+00  1.34000000E+02  1.34312000E-01  15   4
   7  1.99931708E+00  1.35000000E+02  1.33772000E-01  15   3
   8  1.99931708E+00  1.91000000E+02  2.49749000E-01  15   2
   9  4.00000000E+00  8.60000000E+02  5.93958000E-02  28   4
  10  2.00000000E+00  8.77000000E+02  5.86296000E-02  28   3
  11  2.00000000E+00  1.01500000E+03  1.15436000E-01  28   2
  12  1.99931708E+00  2.14800000E+03  5.88756000E-02  15   1
  13  2.00000000E+00  8.33800000E+03  3.06276000E-02  28   1

I'm not sure why one would edit these values from the defaults, but you could ask Xavier.

[John Donovan]

We found and fixed a small bug related to the oxygen-sulfur equivalence code implemented in v. 13.2.7 of Standard.exe and described here:

https://probesoftware.com/smf/index.php?topic=1247.msg11664#msg11664

which caused an error when calculating boundary fluorescence models for compounds containing sulfur. This error is now fixed.  If you have v. 13.2.7 of CalcZAF/Standard please use the Help | Update CalcZAF menu to update your CalcZAF/Standard applications.

[Probeman]

Note that if you have Probe for EPMA you can perform a boundary fluorescence correction directly on your unknown data from the Analyze! Window:

https://probesoftware.com/smf/index.php?topic=1545.0

Note that the SF boundary correction does not yet incorporate a Bragg defocus calibration, so at the moment it is a worse case scenario correction. Which depends on the orientation of the boundary relative to the Bragg crystal on the spectrometer being utilized.

But I believe a Bragg defocus calibration is currently being developed…