Latest version changes for Probe for EPMA (and CalcImage) v. 13.8.3

[John Donovan]

Working again with Scott Boroughs, we've slightly modified the layout of the Standard Assignments and Elements/Cations main dialog to show more information. For example, when opening the Standard Assignments dialog from either the Acquire! or Analyze! windows, you are now presented with this dialog:



Note that the TDI assignments are now shown in the last (new) column.  But because this dialog is also used for the Elements/Cations dialog, we had to add additional information to these dialogs, which are slightly different depending on whether they are loaded from the Acquire! window or the Analyze! window. Here is the new Elements/Cations dialog when opened from the Acquire! window:



Note that the background acquisition type is now shown, and here is the Elements/Cations dialog when loaded from the Analyze! window:



This dialog now displays the spectrometer number and crystal name in the last column. Hopefully, others will find these changes useful.

[John Donovan]

Recently Scott Boroughs and I were chatting about a minor bug (or is it a feature?) in the Automate! window, where when a user has selected multiple position samples, using the multiple sample setups acquisition option, and is using the "Run Multiple Sample Setups One at a Time" option as described here:

https://probesoftware.com/smf/index.php?topic=40.msg10899#msg10899

Which is intended primarily for the acquisition of "constant k-ratios" for instrument calibration, as described here:

https://probesoftware.com/smf/index.php?topic=1466.msg11100#msg11100

Anyway, this "Run Multiple Sample Setups One at a Time" flag is for use when one wants to acquire all the selected position samples using the first assigned sample setup, then all the selected position samples using the next sample setup, etc., etc. This is great when you want to acquire each sample condition "one at a time" for all position samples.  Without the flag set, the automation will acquire each position sample, using all assigned samples setups, for example when acquiring multiple keVs for thin film analysis:

https://probesoftware.com/smf/index.php?topic=111.0

But, if they have assigned different numbers of sample setups to some of the selected position samples, the software will acquire the multiple sample setups as expected, but when it reaches a position sample where the last assigned sample setup was acquired, it will simply stop and say "Automation Completed".  Because of course, there are no more sample setups to run for that position sample.

Now maybe that is what the user intended, but on the other hand maybe you just made a mistake and accidentally did not assign the same number of sample setups to all the selected samples?

Anyway, now the software checks for the number of multiple sample setups assigned to the selected position samples and if any position samples have different numbers of multiple setups assigned it will prompt the user:



If you click No it will allow you to adjust your automation [parameters, but if you click Yes, it will proceed to automate the selected samples until it comes across a position sample where all the multiple sample setup have been acquired already.

Simple, right?

[John Donovan]

Sometimes it's hard to keep track of the many features in Probe for EPMA... more on that later.  Ok, as many of you know, the Analyze! window Report button in Probe for EPMA can be very useful for providing a summary of the analytical conditions for a sample acquisition along with appropriate published references and is documented here:

https://probesoftware.com/smf/index.php?topic=8.msg8398#msg8398

https://probesoftware.com/smf/index.php?topic=42.msg5982#msg5982

For example a colleague of ours recently requested that we also add a report format for wavescan samples just as we do already for standard or unknown samples as described above.  So we said, sure why not? 

A few hours of coding later we now can handle wavescan samples from the same Report button. So if a wavescan sample is selected in the Analyze! window, and you click the Report button, you will get a file output which when loaded into Excel looks like this:



So far so good, right?  Well it was only after we implemented this new code that we discovered that there was already a report format for wavscan samples in the Analyze! window!  This feature is a little hidden as it is accessed by using a right mouse click as seen here:



Oh well. So we decided to keep both wavescan report formats because they are slightly different and maybe one format is preferable to the other for some people.

[John Donovan]

We wanted to warn everyone that we found a serious bug in Probe for EPMA v. 13.4.2 and later when using the sample split feature button in the Analyze! window.

Basically if you split a sample into separate lines from the Analyze! window using v. 13.4.2 or later, the sample will be unreadable and if you close the file and re-open it, because Probe for EPMA tries to read the last sample when opening an MDB file, it will fail and the run won't open.

So, if you have v. 13.4.2 or later, please do not try to split any samples from the Analyze! window until you update to the latest v. 13.6.0 Probe for EPMA using the help menu as usual!

Thanks and sorry for the trouble.

[John Donovan]

We wanted to warn everyone that we found a serious bug in Probe for EPMA v. 13.4.2 and later when using the sample split feature button in the Analyze! window.

Basically if you split a sample into separate lines from the Analyze! window using v. 13.4.2 or later, the sample will be unreadable and if you close the file and re-open it, because Probe for EPMA tries to read the last sample when opening an MDB file, it will fail and the run won't open.

So, if you have v. 13.4.2 or later, please do not try to split any samples from the Analyze! window until you update to the latest v. 13.6.0 Probe for EPMA using the help menu as usual!

Thanks and sorry for the trouble.

As stated previously if you have Probe for EPMA v. 13.4.2 or later you should immediately update to the latest Probe for EPMA using the Help menu as usual.

Using the Separate Selected Samples Into New Samples with Single Point button will make your MDB unreadable due to a bug which is now fixed in the latest version.

In the meantime we tested the new separate samples code and noticed that the TDI, CL and EDS intensity data are not always transferred over to new samples properly when using using the Combine (elements or lines) the Selected Samples into a New Sample, the Separate Selected Samples Into New Samples with Single Point or the Create New Samples from Selected Lines buttons in the Analyze! window as shown here:



The new code now properly transfers the TDI, CL and EDS intensities into the new samples, so please update as soon as possible.

[John Donovan]

Working with Mike Matthews we modified the STRATAGem and BadgerFilm k-ratio export GUI a bit:



The new GUI has better text explanations for the two geometry models (homogeneous vs. replicate) and also now includes the x-ray emission line in the element list. 

Note however, that even though you can select/deselect elements with more than one x-ray line, the export will contain all the emission lines for that element if they are present with two exceptions:

1. If the element channel is disabled for quant (or acquisition) it will be skipped. This can be done from the Analyze! Elements/cations dialog as the tool tip popup help shows in the above image.

2. If one has duplicate elements with the same x-ray line all element channels will be exported if Aggregate mode is off. If aggregate mode is on, the extra element channels will not be exported and will not be listed either.

See here for more details:

https://probesoftware.com/smf/index.php?topic=111.0
https://probesoftware.com/smf/index.php?topic=502.0

Finally, we just wanted to note that if one acquires ones thin film samples properly using the acquire using multiple sample setups option in the Automate! window (as described in the first link above), the samples will be listed with the same name, one for each Kev acquired as seen here:



In these cases the Output All button is really cool as it will automatically select which samples to output and bang them out to your hard drive!

[John Donovan]

The latest version of Probe for EPMA v. 13.7.0 now distributes a default set of USERMAC.DAT and USERMAC2.DAT MAC files.  These files contain MACs for actinide emitters and absorbers compiled by Philipp Poeml and Karen Wright The USERMAC.DAT file contains mass absorption coefficients for the Ka, Kb, la, Lb, Ma and Mb emission lines.  While the USERMAC2.DAT file contains MACs for the more exotic Ll, Ln, etc emission lines.

However, if you already have these files installed they will not be overwritten by the installer as these files are user editable files and marked as "do not overwrite" and "permanent".

Update as usual from the Help menu.

[Nico Vanhove]

Is it possible to make these MAC files available separately?
Where are the MAC values coming from?

Thanks in advance,

Nico

[John Donovan]

As stated above these values are compiled by Philipp Poeml (at ITU) and Karen Wright (at INL).  You would have to ask them where exactly specific values are from.

The files are included in the free CalcZAF distribution (the USERMAC files are installed to the C:\ProgramData\Probe Software\Probe for EPMA folder):

https://probesoftware.com/Update.html

USERMAC.TXT    ' text file of Ka, Kb, La, Lb, Ma, Mb emitters
USERMAC2.TXT   ' text file of Ln, Lg, Lv, Ll, Mg, Mz emitters
USERMAC.DAT    ' binary file used by applications
USERMAC2.DAT   ' binary file used by applications

See here for more details:

https://probesoftware.com/smf/index.php?topic=625.msg3656#msg3656

[John Donovan]

Sometimes it's hard to keep track of the many features in Probe for EPMA... more on that later.  Ok, as many of you know, the Analyze! window Report button in Probe for EPMA can be very useful for providing a summary of the analytical conditions for a sample acquisition along with appropriate published references and is documented here:

https://probesoftware.com/smf/index.php?topic=8.msg8398#msg8398

https://probesoftware.com/smf/index.php?topic=42.msg5982#msg5982

For example a colleague of ours recently requested that we also add a report format for wavescan samples just as we do already for standard or unknown samples as described above.  So we said, sure why not? 

A few hours of coding later we now can handle wavescan samples from the same Report button. So if a wavescan sample is selected in the Analyze! window, and you click the Report button, you will get a file output which when loaded into Excel looks like this:



So far so good, right?  Well it was only after we implemented this new code that we discovered that there was already a report format for wavscan samples in the Analyze! window!  This feature is a little hidden as it is accessed by using a right mouse click as seen here:



Oh well. So we decided to keep both wavescan report formats because they are slightly different and maybe one format is preferable to the other for some people.

As you know, one can export all wavescans from a probe run using the Output | Save Wavescan Output menus.  However, these menus export all wavescan samples, so if you want to export only selected wavescans, you can use the Plot! window and change the Output Target from Graph Window to ASCII Text File.

But as requested recently by Scott Boroughs the latest version of Probe for EPMA can also output selected wavescans from the Analyze! window using the right mouse button click:

[John Donovan]

A very minor tweak to the MAN plot window:



Update from the Help menu as usual.

[AndrewLocock]

A very minor tweak to the MAN plot window:



Update from the Help menu as usual.

Thanks for doing this John!

The rationale behind this request is that the best fit line is effectively a least-squares quadratic regression of the form f(x) = c + bx +ax²,
for which the intercept is c, the slope is b, and the curvature is a.

The attached PDF gives short definitions of the various parameters such as the Coefficient of Determination.

These 3 extra statistical parameters - R², R²_adj - and the Mean Square Error of the Residuals, should help in deciding whether or not to keep a particular point when fitting the MAN background curve.
By tracking the changes in these parameters, as well as the Rel% Deviation, one can determine the effects of omitting a particular point.

Of course plotting the error bars on each point is also very helpful!

Thanks again, Andrew

[John Donovan]

We bumped the PFE version number today to 13.8.2:

https://probesoftware.com/smf/index.php?topic=40.0

The main changes were to the JEOL MEC EDS interface but some other general bug fixes worth updating your PFE using the Help menu when you get a chance.

[AndrewLocock]

A very minor tweak to the MAN plot window:



Update from the Help menu as usual.

Thanks for doing this John!

The rationale behind this request is that the best fit line is effectively a least-squares quadratic regression of the form f(x) = c + bx +ax²,
for which the intercept is c, the slope is b, and the curvature is a.

The attached PDF gives short definitions of the various parameters such as the Coefficient of Determination.

These 3 extra statistical parameters - R², R²_adj - and the Mean Square Error of the Residuals, should help in deciding whether or not to keep a particular point when fitting the MAN background curve.
By tracking the changes in these parameters, as well as the Rel% Deviation, one can determine the effects of omitting a particular point.

Of course plotting the error bars on each point is also very helpful!

Thanks again, Andrew

A couple of references that may prove useful with regard to linear and quadratic regressions:

Rawski, R.I., Sanecki, P.T., Kijowska, K.M., Skitat, P.M. and Saletnik, D.E., 2016. Regression analysis in analytical chemistry. Determination and validation of linear and quadratic regression dependencies. South African Journal of Chemistry, 69, pp.166-173.

Raposo, F. and Barcelo, D., 2021. Assessment of goodness-of-fit for the main analytical calibration models: Guidelines and case studies. Trends in Analytical Chemistry, 143, 116373, 12 p.