Probe Software Users Forum
General EPMA => Discussion of General EPMA Issues => Topic started by: Probeman on April 28, 2014, 12:24:26 PM
-
I'd like to start a discussion of analyzing for boron using EPMA.
Of course, much of what we know from characterization of other light elements (O, N, C) also applies to boron, though usually without the contamination or native oxide issues! On the other hand, the absorption corrections for boron Ka are enormous, with absorption coefficients often as large as 10^5 cm2/ug.
In addition there can be large interferences from some other elements, particularly mid-z transition elements such as Mo.
To start things off I've attached a pdf of a paper just published on carbon doping of these magnesium borides and also a separate white paper on the details of the area peak factor (APF) shape corrections we utilized for previous work on the precursor magnesium borides, MgB2 and MgB4.
I'd be interested in seeing other efforts to quantify boron and what difficulties were encountered...
-
What is the best approach to analysing low amounts (1000s ppm to several weight percent) B in silicate glasses where there is an overlap with the 3rd order Ka1 oxygen peak?
-
What is the best approach to analysing low amounts (1000s ppm to several weight percent) B in silicate glasses where there is an overlap with the 3rd order Ka1 oxygen peak?
Hi Jeremy,
I know you don't have my software, but I assume you've tried the PeakSight interference correction and it didn't perform well enough?
I wonder if the PFE quantitative interference correction would do a better job?
http://probesoftware.com/smf/index.php?topic=69.0
Do you have a good test standard for this situation?
Or have you tried a blank correction, assuming you can obtain a suitable blank standard?
http://probesoftware.com/smf/index.php?topic=204.0
-
I would at least try measuring both oxygen and boron on a test sample and see how well the quant interference correction works. It often will surprise you... please try it and let us know what you find.
-
I grabbed some danburite (CaB2Si2O8) out of the museum a while ago--I will get back to this after I have submitted.
Edit by John: Yes, a nice crystal of that should be a good test. Is it very beam sensitive do you think?
You might want to check this post that has attached a pdf of a paper NIST published a few years ago on analyzing boron on a beam sensitive hydrated material:
http://probesoftware.com/smf/index.php?topic=61.0.msg229#msg229
The other thing that should be mentioned is that it is especially important, with these low energy interferences, to make sure the interference correction is matrix corrected because the unknown and the standard used for the matrix correction, by definition, have to be different compositions!
-
I have a copy of:
"Quantitative Electron probe Microanalysis of Boron in Binary Borides"
the details below may be of use for those trying to obtain this from libraries.
Bastin G.F. and Heijligers H.J.M.
University of Technology Eindhoven
Laboratory for Physical Chemistry
P.O. Box 513, 5600 MB Eindhoven
The Netherlands
ISBN 90-6819-006-7 CIP
SISO 542 UDC 541.1
March 1st 1986
This is a study (128 pages) with analytical data from 24 binary borides (does not include Mg Borides). Has a table of Mass Absorption Coefficients For B ka X-rays from a number of sources (Henke (74) and (82) but also includes a column labelled "present work".
Hope this is of help.
-
This is very interesting. The Henke MACs from 1985 (unpublished) are what I use in the LINEMU.DAT table, but I would be very interested in getting a copy of this. Can it be scanned as a pdf?
-
You can find the Bastin & Heijligers PDFs here:
http://alexandria.tue.nl/repository/books/304562.pdf
http://alexandria.tue.nl/repository/freearticles/620800.pdf
-
Hopefully I have attached a bmp of the MAC table from the first edition, which differs from the second.
John
-
Hopefully I have attached a bmp of the MAC table from the first edition, which differs from the second.
This is what I have in my Help file for boron, however, the empirical MACs table included with our CalcZAF distribution is attached below.
(https://probesoftware.com/smf/oldpics/i61.tinypic.com/2hekcuc.jpg)
-
Does anyone have an LDE2 on a JEOL, 140mm rowland circle, and a tourmaline standard? If so, could you please do a wavescan from 180 to 205mm at 10kV, 10nA, 50um step, 200ms dwell? I'm having trouble picking out the B peak in what a users says is a tourmaline (I can see it clearly in my BN std). My xtal hasn't been checked for reference quality by the service engineer in many years...I'm curious to see how my spectrum and intensity compares to yours. I'm only getting 30cps via LDE2 on BN.
Thanks,
Deon.
-
Hi Deon,
One problem is that B Ka is very heavily absorbed by Si:
john
"Table of MACs (mass absorption coefficients) from C:\ProgramData\Probe Software\Probe for EPMA\LINEMU.DAT"
"Emitting Element: B "
Absorber ka
H 1723.04
He 10694.59
Li 31412.30
Be 60146.45
B 3333.97
C 6337.93
N 11196.24
O 16431.25
F 23306.02
Ne 35353.24
Na 43812.10
Mg 59524.69
Al 64297.13
Si 83890.06
P 68960.19
S 74033.98
Cl 7535.20
Ar 9334.41
-
Hi Deon,
One problem is that B Ka is very heavily absorbed by Si:
john
Another problem is that a typical tourmaline only contains about 3.2 wt% B, while stoichiometric BN contains 43.56 wt% B. I'm about to collect a wavelength scan using LDE2 under the stated conditions on a tourmaline that I analyzed last Thursday. By stoichiometry it contains about 10.3 wt% B2O3.
-
Hi Deon,
One problem is that B Ka is very heavily absorbed by Si:
john
Another problem is that a typical tourmaline only contains about 3.2 wt% B, while stoichiometric BN contains 43.56 wt% B. I'm about to collect a wavelength scan using LDE2 under the stated conditions on a tourmaline that I analyzed last Thursday. By stoichiometry it contains about 10.3 wt% B2O3.
Here are two scans collected using LDE2 under the stated conditions on the same tourmaline grain. A further complication is that B Kα occurs at high Bragg angle on LDE2. L ≈ 190 mm corresponds to sine(θ) around 0.68.
(https://probesoftware.com/smf/gallery/381_20_08_18_8_49_03.png)
(https://probesoftware.com/smf/gallery/381_20_08_18_8_49_35.png)
-
Here are two scans collected using LDE2 under the stated conditions on the same tourmaline grain. A further complication is that B Kα occurs at high Bragg angle on LDE2. L ≈ 190 mm corresponds to sine(θ) around 0.68.
Thank you Brian, I appreciate it!
I attach my presumably-tourmaline raw and smoothed scans. The KLM database puts the peak at about 191 as you can see, while the BN (not attached) is peaking at about 193.X .
So what do you guys think....is that small bump at 191 is the B peak in tourmaline? And what is the big peak at about 195...is that a shifted B? O-III?
Deon.
-
Here are two scans collected using LDE2 under the stated conditions on the same tourmaline grain. A further complication is that B Kα occurs at high Bragg angle on LDE2. L ≈ 190 mm corresponds to sine(θ) around 0.68.
Thank you Brian, I appreciate it!
I attach my presumably-tourmaline raw and smoothed scans. The KLM database puts the peak at about 191 as you can see, while the BN (not attached) is peaking at about 193.X .
So what do you guys think....is that small bump at 191 is the B peak in tourmaline? And what is the big peak at about 195...is that a shifted B? O-III?
Deon.
Hi Deon,
What about optical properties? Is it uniaxial negative, length fast? What about the ED spectrum? Tourmaline can vary greatly in composition, but here is a spectrum collected on a typical schorl from the Fairbanks schist:
(https://probesoftware.com/smf/gallery/381_21_08_18_6_30_46.jpeg)
If you want to verify the presence of boron using LDE2, then you need to increase the dwell time, perhaps to 1 s, and bump the beam current way up, maybe to 200 nA or so -- tourmaline can handle it, though I'd defocus the beam to 10 microns.
Brian
-
If you want to verify the presence of boron using LDE2, then you need to increase the dwell time, perhaps to 1 s, and bump the beam current way up, maybe to 200 nA or so -- tourmaline can handle it, though I'd defocus the beam to 10 microns.
Brian
Thanks for the advice Brian. I'll try it.
I was brought an unknown mineral and asked to scan for Boron. So I first tried to see the B peak on a tourmaline (I assume the user who gave it to me correctly identified it optically; I don't have an EDS :'().
I'm trying to wrap my head around what's possible with LDE2. At 10nA there appears to be peaks at about 190-ish and 195 in various blanks, and the tourmaline (if those are in fact peaks and not noise). At 30nA some of the "peaks" begin to merge. I'll see what 200nA kicks out.
Have you tried to quantify B in tourmaline with LDE2?
Deon.
-
If you want to verify the presence of boron using LDE2, then you need to increase the dwell time, perhaps to 1 s, and bump the beam current way up, maybe to 200 nA or so -- tourmaline can handle it, though I'd defocus the beam to 10 microns.
Brian
Thanks for the advice Brian. I'll try it.
I was brought an unknown mineral and asked to scan for Boron. So I first tried to see the B peak on a tourmaline (I assume the user who gave it to me correctly identified it optically; I don't have an EDS :'().
I'm trying to wrap my head around what's possible with LDE2. At 10nA there appears to be peaks at about 190-ish and 195 in various blanks, and the tourmaline (if those are in fact peaks and not noise). At 30nA some of the "peaks" begin to merge. I'll see what 200nA kicks out.
Have you tried to quantify B in tourmaline with LDE2?
Deon.
Hi Deon,
I've never analyzed for boron in tourmaline because generally it's reasonable to assume wt% B2O3 that gives 3 B3+ per 29 anhydrous oxygens. Like John pointed out, the absorption correction when using BN as a boron standard will be outrageously large and will likely lead to inaccurate wt% B2O3 (not to mention problems with variable peak position, variable peak shape, and poor counting statistics). Many tourmalines fall at least roughly within the dravite-schorl solid solution. For the dravite end-member (NaMg3Al6(BO3)3Si6O18(OH)4), wt% B2O3 is 10.89, and wt% H2O is 3.76. For end-member schorl (NaFe3Al6(BO3)3Si6O18(OH)4), wt% B2O3 is 9.91, and wt% H2O is 3.42.
One approach is simply to analyze for Si, Al, Ti, Cr, Fe, Mn, Mg, Ca, Na, K, F, and maybe Cl (not likely to be present in measurable quantity) and adjust assumed wt% B2O3 and wt% H2O (taking halogens into account if present) and see if you can produce a reasonable-looking tourmaline formula unit. Keep in mind, though, that Al3+ is somewhat variable, as it can substitute for Si4+ in a coupled substitution that also replaces a divalent cation with VIAl3+. Significant Fe2O3 may be present as well. Also, Li+ can substitute on the octahedral sites, and this can complicate interpretation of the analysis -- it also requires adding octahedral Al3+ to maintain charge balance such that elbaite (Na(VIAl,Li)3Al6(BO3)3Si6O18(OH)4) may contain in excess of ~43 wt% Al2O3.
Maybe you already know about this, but a good on-line resource is the Handbook of Mineralogy (http://www.handbookofmineralogy.org/search.html?p=all), as it gives representative analyses of most of the listed minerals.
Brian
-
A couple of years ago I was experimenting with fully quantified EDS analyses of tourmaline measuring boron and oxygen, as the latest generation SDDs are nicely sensitive to lower energy than older SiLi detectors. For analysis of tourmaline with a beam energy of 6 keV, I could get really good signal to noise, and fitting was pretty good, the values were not far off from an idealized stoichiometric result. So, as with so many EPMA cases the issue is having an appropriate tourmaline standard to use. The interesting issue with this specific situation is that analyzing for boron in a silicate matrix like tourmaline, there doesn't seem to be a method capable of getting the Boron without some kind of method concern (use of borate fluxes for wet chemical analyses, neutron absorption issues with INAA, no measured standard for SIMS or LA-ICPMS). My question to the group is, do you know of a method to analyze tourmaline that doesn't have a major complication?
In the end, I feel that one is probably left with selecting a particular tourmaline, analyzing excessively and determining Boron by difference, considering it stoichiometric and using it for your additional tourmaline analyses. Not perfect, but at least one could generate a laboratory-internal consistent set of tourmaline analyses.