Accuracy and Precision in Using MAN Background Corrections

[Rom]

Please, let me dream.
May be number of questions about accuracy and precision will decrease if we can adjust MAN BG curve to current equipment parameters.
What does it mean?
It means we should make 2 working steps.
Step 1. We need develop a perfect MAN BG curve with 5, 10, ...., n standards. It is common step for nowadays.
The shape of the curve is quite stable which were discussed a lot. We can even create the equation of the curve and place it to the database.

Good job! let's have a rest a day, a month, a year...

Step 2. We need to measure just few (1-3) MAN standards in the session with our unknowns and adjust the whole MAN curve (without changing the shape/equation) which we developed earlier to current equipment parameters (off course, the energy, current should be the same; beam size close to).
For extremally worrying researchers the operation could be repeated each several hours of their measurements with different standards.

If I am right, today we can only check the MAN BG curve for current session or add from current session some extra points to earlier collected points set. But we haven't option "adjust MAN curve without its shape changing".

I am very sorry if it is already realised (or doesn't make sense) but I don't know about it - could you shortly correct my view.
Thank you!

[John Donovan]

Please, let me dream.
May be number of questions about accuracy and precision will decrease if we can adjust MAN BG curve to current equipment parameters.
What does it mean?
It means we should make 2 working steps.
Step 1. We need develop a perfect MAN BG curve with 5, 10, ...., n standards. It is common step for nowadays.
The shape of the curve is quite stable which were discussed a lot. We can even create the equation of the curve and place it to the database.

Good job! let's have a rest a day, a month, a year...

Step 2. We need to measure just few (1-3) MAN standards in the session with our unknowns and adjust the whole MAN curve (without changing the shape/equation) which we developed earlier to current equipment parameters (off course, the energy, current should be the same; beam size close to).
For extremally worrying researchers the operation could be repeated each several hours of their measurements with different standards.

If I am right, today we can only check the MAN BG curve for current session or add from current session some extra points to earlier collected points set. But we haven't option "adjust MAN curve without its shape changing".

I am very sorry if it is already realised (or doesn't make sense) but I don't know about it - could you shortly correct my view.
Thank you!

Dream on!   

Seriously, your suggestion is unnecessary for several reasons.

1. The MAN standard intensities are automatically drift corrected in time.

This means that if your MAN background intensities did change over time, the software will automatically calculate this intensity drift for each data point based on its time of acquisition. Assuming of course that you have MAN standard intensities acquired at some time both before and after the samples.  This is why there is a "Acquire Standard Samples" and a "Acquire Standard Samples (again) option in the Automate! window.

However if your MAN background intensities are noticeably drifting, you have bigger problems, i.e., your primary standard intensities!  So you will need to re-acquire them!

2. If you want to re-use your MAN standard intensities in another probe run, simply use the Load File Setup option from the Acquire! window | New Sample/Setups button.

And when asked if you want to load standard intensities, just click "Yes". This will automatically load all primary, interference and MAN standard intensities for that sample.   Take a few mouse clicks and you are ready to go with new unknown samples. Many users say this is one of the best features in Probe for EPMA.

If you suspect for some reason that your standard intensities might have drifted, simply re-acquire those standard intensities.  In practice I find that at most only one or two primary standard intensities (for major elements) might possibly have drifted enough to be noticeable.

One can disable the Use Standard Drift Correction option from the Analytical | Analysis Options dialog, (in which case only the standards acquired before the sample in question are utilized for analysis).  But I always leave it on, except in very special circumstances, for example when analyzing samples acquired for the constant k-ratio method to be used for calibration of ones detector dead times.

https://smf.probesoftware.com/index.php?topic=1466.msg11196#msg11196
https://smf.probesoftware.com/index.php?topic=1466.msg11858#msg11858

But hey, we can all dream!   

[Probeman]

I recently had some discussions with one of our colleagues regarding using the trace element blank correction with the MAN background correction.  Because it is worth discussing the application of these two corrections with regards to accuracy versus precision:

https://smf.probesoftware.com/index.php?topic=307.msg11321#msg11321

The MAN background correction is useful because it decreases acquisition time while improving precision! For the statistical details see here:

https://pubs.geoscienceworld.org/msa/ammin/article-abstract/101/8/1839/264218/A-new-EPMA-method-for-fast-trace-element-analysis

At the same time it also improves accuracy because off-peak interferences are no longer a problem, because they are not measured!

So the advantage of the MAN background correction is that one can improve sensitivity and/or reduce acquisition time, especially for major and minor elements where the background correction is often small, e.g., large P/B.

However, for trace elements, especially in high Z matrices, the background correction becomes the most important correction:

https://smf.probesoftware.com/index.php?topic=1535.msg12121#msg12121

In these cases the accuracy of the MAN background correction becomes critically important.  Because the MAN background correction is a best fit to a number of standards, each not containing the element of interest, but because of spectral interferences, and/or undocumented trace contamination, the accuracy of the MAN background correction may be problematic at trace levels.

Enter the blank correction!  In the blank correction we obtain a material which has been certified as having a zero (or non-zero but low level) of the element in question. Basically the blank correction tells us how accurately we can measure zero.  We then apply this zero calibration to our unknown samples to improve the accuracy of the background correction, whether that be the off-peak or MAN background correction.

It is also important to keep in mind whether the background artifact which requires a blank correction, is due to the sample or the instrument. In the case of an instrumental artifact (a secondary Bragg diffraction (negative) peak under the Ti peak on certain PET crystals), we do not need a matrix matched zero blank material. But in the case of a sample induced background artifact, e.g., the "hole" in the continuum near the Au La peak in pyrites, we do require a matrix matched blank material, e.g., a synthetic high purity FeS2 material:

https://smf.probesoftware.com/index.php?topic=592.msg6414#msg6414

In the example of measuring Ti in natural quartz samples, we do not need a matrix matched blank material, since it is easy to obtain synthetic SiO2, we might want to obtain a synthetic quartz glass for the blank measurement, such as SpectroSil:

https://smf.probesoftware.com/index.php?topic=928.msg8498#msg8498

for use as our blank material, and measure it at the same beam conditions and counting time as our unknowns. Likewise, the same background method for each element (off-peak vs MAN vs MPB) utilized for the unknown, should also be utilized for the blank measurement material.  If we are correcting for background artifacts, the same background correction should be applied to both.

This SpectroSil blank material, which contains around 50 PPB Ti, so it is essentially a zero measurement for the microprobe! We also do not detect a difference between SiO2 glass and SiO2 crystal for this blank measurement for Ti, but beam damage is more problematic in SiO2 crystal, so beware at high beam currents with a focused beam.