we are trying to setup a measurement program at low voltages (10 kV/5 kV) to analyse small mineral grains. So we use L-lines for elements such as Ti, Mn, Cr or Fe. There is one problem with Mn: in an ol standard with Mn below the d.l. (for L-lines), Mn appears with almost 4 wt%. I cannot find a sensible interference to explain this. The Mg SKA4 and 5 lines are close, but imho not close enough. When I measure a mineral with higher Mn-concentration (20 or 40 wt%), I obtain perfect Mn results. The count rates for Mn are really low – maybe that is a problem?
Using the Standard application (included with the free CalcZAF software) I selected an olivine standard (USNM 2566), edited the Mn line from Ka to La, changed the crystal from LiF to TAP, and selected the Misc Options | Interferences menu dialog and calculated the possible interferences:
![](https://probesoftware.com/smf/gallery/395_12_12_23_8_52_25.png)
The reported composition for this standard is:
St 474 Olivine USNM 2566 (Fo 83) Springwater
TakeOff = 40.0 KiloVolt = 15.0 Density = 3.300 Type = olivine
Analysis (wet chemistry) by Gene Jarosewich
Oxide and Elemental Composition
Average Total Oxygen: 41.817 Average Total Weight%: 99.470
Average Calculated Oxygen: 41.820 Average Atomic Number: 12.535
Average Excess Oxygen: -.003 Average Atomic Weight: 21.723
ELEM: SiO2 FeO MgO MnO Cr2O3 O
XRAY: ka ka ka la ka ka
OXWT: 38.951 16.620 43.582 .300 .020 -.003
ELWT: 18.207 12.919 26.281 .232 .014 41.817
ATWT: 28.086 55.847 24.305 54.938 51.996 16.000
KFAC: .1310 .1099 .1758 .0007 .0001 .2366
ZCOR: 1.3901 1.1758 1.4952 3.1287 1.1083 1.7678
AT% : 14.157 5.052 23.615 .092 .006 57.078
24 O: 5.953 2.124 9.929 .039 .002 24.000
The results from the (nominal) interference calculations are:
On Peak Interferences for : St 474 Olivine USNM 2566 (Fo 83) Springwater
For Si ka TAP at 7.14182 angstroms, at an assumed concentration of 18.207 wt.%
Interference by Fe SKB IV at 6.98090 ( 75.9239) ( -1.7502) = .1%
Interference by Fe SKB`` IV at 6.99440 ( 76.0708) ( -1.6034) = .2%
Interference by Fe KB3 IV at 7.02780 ( 76.4340) ( -1.2401) = 1.2%
Interference by Fe KB1 IV at 7.02780 ( 76.4340) ( -1.2401) = 2.1%
Interference by Fe SKB` IV at 7.03850 ( 76.5504) ( -1.1237) = .2%
Interference by Fe SKBN IV at 7.13340 ( 77.5825) ( -.09161) = .2%
On Peak Position ------------- at 7.14182 ( 77.6741)
For Fe ka LiF at 1.93746 angstroms, at an assumed concentration of 12.919 wt.%
On Peak Position ------------- at 1.93746 ( 134.724)
For Mg ka TAP at 9.91164 angstroms, at an assumed concentration of 26.281 wt.%
Interference by Fe KA1 V at 9.68180 ( 105.299) ( -2.4997) = 1.3%
Interference by Fe KA2 V at 9.70130 ( 105.511) ( -2.2877) = .9%
On Peak Position ------------- at 9.91164 ( 107.799)
For Mn la TAP at 19.4925 angstroms, at an assumed concentration of 0.2323353 wt.%
Interference by Mg SKB` II at 19.3330 ( 210.265) ( -1.7347) = .8%
Interference by Mg SKBN II at 19.3450 ( 210.395) ( -1.6042) = 1.3%
Interference by Mg SKA9 II at 19.3950 ( 210.939) ( -1.0604) = 6.4%
Interference by Mg SKA6 II at 19.4560 ( 211.603) ( -.39694) = 19.0%
Interference by Mg SKA7 II at 19.4850 ( 211.918) ( -.08153) = 22.5%
On Peak Position ------------- at 19.4925 ( 212.000)
Interference by Mg SKA5 II at 19.5080 ( 212.168) ( .168610) = 21.9%
Interference by Mg SKA8 II at 19.5370 ( 212.484) ( .484024) = 17.4%
Interference by Mg SKA4 II at 19.6160 ( 213.343) ( 1.34322) = 24.1%
Interference by Mg SKA3 II at 19.6490 ( 213.702) ( 1.70213) = 7.1%
Interference by Mg SKA` II at 19.6950 ( 214.202) ( 2.20242) = .3%
For Cr ka LiF at 2.29113 angstroms, at an assumed concentration of 0.014 wt.%
On Peak Position ------------- at 2.29113 ( 159.317)
For O ka WSi60 at 23.9756 angstroms, at an assumed concentration of 41.817 wt.%
On Peak Position ------------- at 23.9756 ( 110.052)
Note the interference of Mn La by the Mg Ka (II), second order line. That's probably your problem. You'll want to apply the quantitative interference correction in Probe for EPMA using an interference standard that contains Mg, but no Mn, say MgO:
https://probesoftware.com/smf/index.php?topic=69.0Alternatively you could utilize the "low overvoltage" method for reducing the size of the analytical volume. For example analyzing the Mn Ka line just above the edge energy of the Mn K edge, which is 6.539 keV, so say using a beam energy of 7 or 8 keV.
This also has the benefit of not only reducing the interaction volume, but one still retains Ka quantitative accuracy for the matrix correction, not to mention avoiding peak shape and shift issues. See here for more tips on improving spatial resolution:
https://probesoftware.com/smf/index.php?topic=1354.0Note that you can quickly model interaction volume sizes using this dialog in CalcZAF:
https://probesoftware.com/smf/index.php?topic=86.msg309#msg309though for rigorous modeling of the interaction volume you'll want to use a good Monte Carlo software such as Casino or Penepma.