How to Quant B4C Stoichiometry

[Michael Lance]

Hello. A customer of mine is interested in measuring the stoichiometry of B4C. It is a semiconductor so it will need to be coated. How might I approach this experiment? I presume I can’t coat with carbon. Should I try gold or nickel?

Has anyone attempted something similar?

Thanks in advance.

-Michael

[Probeman]

Hi Michael,

Lots to mention, but here's a start. First see this discussion of the analysis of boron, specifically in Mg borides, that I performed some years ago:

https://probesoftware.com/smf/index.php?topic=248.msg1186#msg1186

To deal with peak shift and shape changes for B and C emission lines, one can utilize so called area peak factors (APFs), which are necessary if the standard and unknown are not closely matrix matched and significant peak shifting is visible. They come in two flavors, first "specified" area peak factors (Bastin's method). One can also utilize "compound" APFs (from Donovan), which are synthesized automatically from weighted values of binary APF measurements from the APF table in Probe for EPMA. There's a discussion here about the differences between these two types of APFs:

https://probesoftware.com/smf/index.php?topic=536.0

This post is also worth reading and near the bottom I cite a paper published with a student at Ohio State on the aforementioned Mg borides:

https://probesoftware.com/smf/index.php?topic=1185.msg8154#msg8154

Integrated intensity measurements might also be a better option than APFs:

https://probesoftware.com/smf/index.php?topic=705.msg4326#msg4326

Integrated intensity measurements are much more time consuming than peak intensity acquisitions because the peak is scanned for each analysis, but this method automatically accounts for peak shift/shape changes. And another question is: are the published MACs accurate enough for determining stoichiometry?  Fortunately for you the MACs for B ka in C are *much* lower than the MACs for B ka in Mg that I had to deal with:

MAC value for B ka in C =    6337.93  (LINEMU   Henke (LBL, 1985) < 10KeV / CITZMU > 10KeV)
MAC value for B ka in C =    6456.00  (CITZMU   Heinrich (1966) and Henke and Ebisu (1974))
MAC value for B ka in C =        .00  (MCMASTER McMaster (LLL, 1969) (modified by Rivers))
MAC value for B ka in C =    5944.80  (MAC30    Heinrich (Fit to Goldstein tables, 1987))
MAC value for B ka in C =    6992.63  (MACJTA   Armstrong (FRAME equations, 1992))
MAC value for B ka in C =    5875.21  (FFAST    Chantler (NIST v 2.1, 2005))
MAC value for B ka in C =    5875.21  (USERMAC  User Defined MAC Table)

MAC value for B ka in Mg =   59524.69  (LINEMU   Henke (LBL, 1985) < 10KeV / CITZMU > 10KeV)
MAC value for B ka in Mg =   58170.00  (CITZMU   Heinrich (1966) and Henke and Ebisu (1974))
MAC value for B ka in Mg =        .00  (MCMASTER McMaster (LLL, 1969) (modified by Rivers))
MAC value for B ka in Mg =   54026.43  (MAC30    Heinrich (Fit to Goldstein tables, 1987))
MAC value for B ka in Mg =   89684.59  (MACJTA   Armstrong (FRAME equations, 1992))
MAC value for B ka in Mg =   51835.16  (FFAST    Chantler (NIST v 2.1, 2005))
MAC value for B ka in Mg =   51835.16  (USERMAC  User Defined MAC Table)

In any case I would utilize the empirically measured MACs from the literature (built into Probe for EPMA- see the Analytical menu):

https://probesoftware.com/smf/index.php?topic=347.msg1811#msg1811

Unfortunately, the MACs for C ka in B are pretty large:

MAC value for C Ka in B =   36816.89  (LINEMU   Henke (LBL, 1985) < 10KeV / CITZMU > 10KeV)
MAC value for C Ka in B =   37020.00  (CITZMU   Heinrich (1966) and Henke and Ebisu (1974))
MAC value for C Ka in B =        .00  (MCMASTER McMaster (LLL, 1969) (modified by Rivers))
MAC value for C Ka in B =   39834.31  (MAC30    Heinrich (Fit to Goldstein tables, 1987))
MAC value for C Ka in B =   38497.84  (MACJTA   Armstrong (FRAME equations, 1992))
MAC value for C Ka in B =   33723.39  (FFAST    Chantler (NIST v 2.1, 2005))
MAC value for C Ka in B =   33723.39  (USERMAC  User Defined MAC Table)

In any case stick with the empirically determined MACs and you should be OK.

As for standards, you ideally should find a single crystal boron carbide and use that as your primary standard, but note the orientation sensitivity of some borides:

https://probesoftware.com/smf/index.php?topic=667.msg4112#msg4112

Sintered borides can have almost any chemistry and lots of impurities. If you have to use a pure boron metal or borides other than boron carbide as a standard, you will have to utilize either APFs or integrated intensity acquisitions on your WDS spectrometers.

Don't despair, this is totally doable, but it is not going to be a walk in the park!  Please keep us informed on your efforts, it would be a useful exercise for many in the field that have not attempted boron analyses.

As for coating, just look at the MAC table in CalcZAF and see what element has relatively low MACs for B ka and C ka of the elements you have available to coat with.

john