I am trying to determine MACs for Cka in U using the approach outlined in the XMAC program of SAMx. Unfortunately in UC, at higher voltages there is a U peak (N iv O iv)on the PC2 crystal that nearly completely obscures the Cka peak. However, by measuring U on its usual crystal (PET or QTZ), I can know how much U is in the sample. By knowing the U concentration in the sample, the program should be able to determine the size of the U peak (in cps/na) that must be obscuring the C peak and subtract it out. The remaining counts should be those from CKa.
BUT......
1. If the U obscures the C peak, I cannot empirically determine an APF, therefore I must use the integral. Correct?
2. In order to use the xMAC program, I need to know the background corrected Cka in cps/na in U without the added counts from U. How do I find out the magnitude of the U overlap correction so that I know how much cps/na to remove from the C signal? Is it available in the software? Do you think this is the correct approach?
Thanks in advance for your thoughts!
Karen
Quote from: wrigke on April 11, 2019, 12:51:34 PM
I am trying to determine MACs for Cka in U using the approach outlined in the XMAC program of SAMx. Unfortunately in UC, at higher voltages there is a U peak (N iv O iv)on the PC2 crystal that nearly completely obscures the Cka peak. However, by measuring U on its usual crystal (PET or QTZ), I can know how much U is in the sample. By knowing the U concentration in the sample, the program should be able to determine the size of the U peak (in cps/na) that must be obscuring the C peak and subtract it out. The remaining counts should be those from CKa.
BUT......
1. If the U obscures the C peak, I cannot empirically determine an APF, therefore I must use the integral. Correct?
2. In order to use the xMAC program, I need to know the background corrected Cka in cps/na in U without the added counts from U. How do I find out the magnitude of the U overlap correction so that I know how much cps/na to remove from the C signal? Is it available in the software? Do you think this is the correct approach?
Thanks in advance for your thoughts!
Karen
Hi Karen,
Wow, this is a tough one I must say. I saw it yesterday when you posted it, but needed to think about it overnight. I think I have some solutions for you.
For those that don't know, the XMAC program is a small app distributed by SAMX based on code by Pouchou and Pichoir that allows the user to calculate MACs (mass absorption coefficients) from intensity measurements made over range of beam energies in a compound containing the emitting element and the absorbing element. It works best for low energy emission lines. For these low energy emission lines it generally produces MAC values that are quite accurate. For a list of some empirically determined MACs generated by this app, see the Empirical MACs menu in CalcZAF or Probe for EPMA. Here's an example:
https://smf.probesoftware.com/index.php?topic=667.msg4064#msg4064
Now as far as the interference of U on C Ka in determining the MAC, I think the solution is going to be that you'll have to measure both C Ka and U Ma quantitatively at each beam energy and specify the interference correction(s). How does this help, well it's not generally known but if you specify interference corrections in Probe for EPMA, the correction for spectral interference occurs during the matrix correction quantitatvely, so in the end you have not only concentrations corrected for interferences, but *also* you'll have k-ratios corrected for interference as well! See note below.
By the way, this is why acquiring thin film intensities in Probe for EPMA works so well for spectral interferences on thin film samples. The k-ratios exported into STRATAGem are *already* corrected for spectral interferences!
Now XMAC doesn't care what units the intensities are in, so you could just take the spectral interference corrected k-ratios from Probe for EPMA and utilize those intensities for the empirical determination of your MAC. It might end up being an interative process because you might want to enter the empirical MAC determined from XMAC into PFE to improve the accuracy of the calculation.
On the APF question for changes in the carbon peak shape, I think you should just utilize the integrated intensities option in Probe for EPMA so that you won't need to use APF values at all. See the option Use Integrated Intensities in the Elements/Cations dialog. Here's an example using integrating intensities on sulfur and also aggregating multiple spectrometers:
https://smf.probesoftware.com/index.php?topic=42.msg4932#msg4932
Note: I looked at the (default) NIST x-ray database and I don't see a uranium line near C Ka (between 40 and 50 angstroms) , though I do see a 2nd order O Ka line. I believe you but in what table are you seeing the U line near carbon Ka? Is this the modified xray.mdb that you and Philipp developed for PFE?
I'm actually analysing C in U right at the moment and talked about this very interference problem at the AMAS symposium a couple of months ago. You can't completely separate the C and U lines but you can reduce the degree of interference by using the 2nd order C Ka line on either an LDE1/PC1 or a Pb-stearate. This puts the line at the top end of the spectrometer range, where the resolution is better. The stearate has the highest resolution of the light element crystals, and almost completely separates the two peaks but the intensity is really lousy (<200th that of the LDE2/PC2). Both still need overlap correction (really easy to set up in PfE, thanks John!). I'm using the LDE1/PC1 as a compromise between level of correction and useable count rate (~20th that of LDE2/PC2). Thanks to Ben Buse for suggesting this method. As John says, the k-ratio values that PfE outputs include the overlap correction.
I have a sort of linked question though: The interfering U line is the N6-O4 at 0.286keV, with a relative intensity of 0.01% (Bearden, 1967), but there should also be a N6-O5 line at 0.294keV which Bearden lists with a relative intensity of 1% (i.e. 100 times bigger than the N6-O4 line), but I see no evidence of this peak. My working theory is that the O5 shell in U isn't normally occupied so the line can't be fluoresced and the 1% intensity in Bearden is just a default value when the actual intensity hasn't been measured. However, my understanding of how shells are occupied is very limited and I might be completely misunderstanding the diagrams. Can anyone confirm or deny this hypothesis?
Mike
OK, after discussions with Xavier, it looks like I am mistaken about the shell occupancies: For U the 5f shell is partially filled and this corresponds to the O6 and O7 levels (I was wrongly equating the 5f with the O4 level) so the N6-O5 transition should be possible. However, unlike Bearden, the Penelope database lists the intensity of this transition as 1/10th that of the N6-O4 which may explain why I haven't seen it. That does make more sense since why would Bearden list it if it wasn't a possible transition. I guess the 1% relative intensity is just a default value if the actual value isn't known. Since I have some DU in the probe at the moment I'll try a slow scan to see if I can tease this peak out of the background.
Mike
Xavier Llovet responded to both Mike and I and said we could post his comments. Here they are:
QuoteHi John, Mike:
I've checked the electronic configuration of U and is 5f3 6d1 7s2 so in principle the O4 and O5 levels (5d) are filled. Those that are partially filled are the 5f levels, which would correspond to O6 and O7 levels. So if I'm correct, both two N6-O5 and N6-O4 lines should exist.
The problem could be in the relative intensity reported by Bearden. For instance, according to the PENELOPE database, the N6-O4 line (with energy 289 eV) is 10 times more intense that the N6-O5 line (with energy 298 eV), just the opposite than Bearden's data! You can find this information in the U.mat file:
15 20 0 7.40479E-05 2.89740E+02
15 21 0 4.72978E-06 2.98070E+02
(the transition "15 20 0" would correspond to the N6-O4 line and the transition "15 21 0" to the N6-O5 line)
So perhaps this could explain what you're seeing?
Regards,
Xavier
The (modified) NIST x-ray database included with Probe for EPMA does include some N emission lines, but not all apparently. I guess we need a KLMN x-ray database!
Pd MZ2 43.3622 .285930 1.00000 ES
Pd MZ1 43.3622 .285930 10.0000 ES
Ne KB1 III 43.3814 .857410 .500000 JD
Se SLB1`` V 43.4488 1.42680 .250000 JD
Dy M2-N1 V 43.4610 1.42640 1.25000 JD
Ni SLA3 III 43.4691 .855680 .500000 JD
La MB III 43.5016 .855040 45.0000 JD
Bi N5-N6 43.5801 .284500 1.00000 ES
Eu M4-O2 IV 43.5993 1.13750 .064000 JD
Eu MA1 IV 43.6146 1.13710 64.0000 JD
Ag MG II 43.6422 .568190 16.0000 JD
Eu MA2 IV 43.6492 1.13620 64.0000 JD
Tm MZ1 IV 43.6607 1.13590 3.84000 JD
Er M4-O2 V 43.6630 1.41980 .025000 JD
Tl N4-N6 43.6707 .283910 1.00000 ES
Tm MZ2 IV 43.6761 1.13550 .640000 JD
Se LB1 V 43.6876 1.41900 9.71200 JD
Ni LA2 III 43.7081 .851000 5.72500 JD
Ni LA1 III 43.7081 .851000 50.0000 JD
Mn Ln II 43.7338 .567000 5.14300 JD
Ga L2-N3 IV 43.7724 1.13300 .173000 JD
Ce M2-N4 IV 43.8266 1.13160 5.12000 JD
Ga SLB1`` IV 43.8421 1.13120 .640000 JD
Ne KA1 III 43.8467 .848310 50.0000 JD
Ne KA2 III 43.8467 .848310 25.0000 JD
Ga LG5 IV 43.8886 1.13000 .141000 JD
Tm M3-N1 V 43.9260 1.41130 .250000 JD
Pr MG IV 43.9938 1.12730 38.4000 JD
C KA1 44.0023 .281770 100.000 ES <-- carbon emission
C KA2 44.0023 .281770 50.0000 ES <-- carbon emission
Er MA1 V 44.0039 1.40880 25.0000 JD
Er MA2 V 44.0039 1.40880 25.0000 JD
Cs M2-N1 III 44.0049 .845260 7.00000 JD
Cs MZ1 II 44.0063 .563490 8.00000 JD
Ga LB1 IV 44.0837 1.12500 10.6910 JD
La M3-N1 III 44.1759 .841990 .650000 JD
Pd M2-N4 II 44.2016 .561000 .800000 JD
Gd MG V 44.2205 1.40190 6.52500 JD
Zr M3-N1 44.2789 .280010 1.00000 ES
As Ll IV 44.2805 1.12000 3.15500 JD
Sm MZ2 III 44.4259 .837250 .500000 JD
Ge LG3 V 44.4392 1.39500 .028000 JD
Mn Ll II 44.5991 .556000 10.6000 JD
Se SLA4 V 44.6119 1.38960 .250000 JD
La MA1 III 44.6414 .833210 50.0000 JD
La MA2 III 44.6414 .833210 50.0000 JD
Eu M3-N1 IV 44.6432 1.11090 .640000 JD
Se SLA5 V 44.6569 1.38820 .250000 JD
As LB4 V 44.6633 1.38800 .650000 JD
As LB3 V 44.6633 1.38800 1.19200 JD
Ho SMB2 V 44.6859 1.38730 .250000 JD
Cu Ln III 44.7068 .831990 1.37500 JD
Kr Ll V 44.7601 1.38500 1.12700 JD
Ga SLA4 IV 44.7601 1.10800 .640000 JD
Sm MZ1 III 44.7865 .830510 3.00000 JD
Se SLA3 V 44.7924 1.38400 .250000 JD
Zn LB4 IV 44.8005 1.10700 1.51000 JD
Zn LB3 IV 44.8005 1.10700 .250000 JD
Ru M4-O2 44.8021 .276740 .010000 ES
Ga SLA5 IV 44.8086 1.10680 .640000 JD
W MZ1 V 44.8118 1.38340 .336000 JD
Ho MB V 44.8280 1.38290 14.8610 JD
La M4-O2 III 44.8313 .829680 .050000 JD
Sm M2-N4 V 44.8832 1.38120 .300000 JD
Se LA2 V 44.9548 1.37900 2.85500 JD
Se LA1 V 44.9548 1.37900 25.0000 JD
W MZ2 V 44.9679 1.37860 1.12800 JD
Ga SLA3 IV 44.9956 1.10220 .640000 JD
Pb N5-N6 45.0021 .275510 1.00000 ES
Sm MB IV 45.0815 1.10010 56.3200 JD
I MG III 45.1064 .824620 10.0000 JD
Ga LA2 IV 45.1677 1.09800 7.30900 JD
Ga LA1 IV 45.1677 1.09800 64.0000 JD
Sb M2-M4 45.2023 .274290 .010000 ES
Hg N4-N6 45.2023 .274290 1.00000 ES
Na SKB^4 IV 45.2584 1.09580 .640000 JD
Tb M2-N1 V 45.3495 1.36700 1.25000 JD
Te M2-N4 III 45.3511 .820170 .500000 JD
Nd M2-N1 IV 45.3993 1.09240 2.56000 JD
Er MZ1 IV 45.4867 1.09030 3.84000 JD
Er MZ2 IV 45.5118 1.08970 .640000 JD
Ho M4-O2 V 45.5728 1.36030 .025000 JD
Er M3-N1 V 45.5996 1.35950 .250000 JD
Sm MA1 IV 45.7173 1.08480 64.0000 JD
Sm MA2 IV 45.7173 1.08480 64.0000 JD
Sm M4-O2 IV 45.7216 1.08470 .064000 JD
Cd M2-N1 II 45.8018 .541400 4.00000 JD
Ho MA2 V 45.8085 1.35330 25.0000 JD
Ho MA1 V 45.8085 1.35330 25.0000 JD
Cu Ll III 45.8639 .811000 2.07000 JD
In M3-N1 II 45.9503 .539650 .800000 JD
By the way, I also searched Philipp Poeml's modified replacement NIST x-ray database, which has additional emission lines for the actinide elements, and it doesn't have these N emission lines for U either.
It did take some digging to work out what the overlapping peak was, none of the software tools showed any U lines in that region of the spectrum. We had to resort, as the late Douglas Adams put it, to twig-technology and dig out the hard copy x-ray tables.
Thank you all for your suggestions. I will let you know what I find out when I try these suggestions.
I do have a question about the overlapping line though. My Bearden 1967 clearly describes the U line at 286 ev as "N4O4", not N6O4. Is that a misprint in the article?
Karen
Hi Karen,
Sorry for the delay in replying. I got my info from the NIST x-ray database which does quote Bearden 1967 as one of its sources, but this gives the line as N6-O4. Looks like there's been a transcription error somewhere. Does anyone have the big Cameca book of x-ray tables to hand to see what that one says is the U-line at 43.4 A, 0.286 keV?
Mike
Quote from: Mike Matthews on April 24, 2019, 04:02:30 AM
Hi Karen,
Sorry for the delay in replying. I got my info from the NIST x-ray database which does quote Bearden 1967 as one of its sources, but this gives the line as N6-O4. Looks like there's been a transcription error somewhere. Does anyone have the big Cameca book of x-ray tables to hand to see what that one says is the U-line at 43.4 A, 0.286 keV?
Mike
I only have the BRGM "little book" which only goes down to the U M lines, but I do have a White and Johnson (2nd edition, May 1970) book which reports:
U N6-O4 0.286 KeV 43.29999 angstroms
OK, I also found my Bearden 1964 book which lists:
U N6-O4 0.286 KeV 43.6 angstroms (literally NVI OIV, but if my Roman numerals are correct that's N6-O4)
Has anyone got a reference or a weblink for the NIST x-ray database. I want to cite it but I can't remember where I picked my copy up from and a search isn't turning it up.
Is this what you want? https://www.nist.gov/pml/x-ray-transition-energies-database
Unfortunately not. I did see that one but it's only got K and L lines. The one I'm after was compiled by Chuck Fiori. Failing that, is there another citable comprehensive database? I've got Bearden but I'm hoping for something more recent and with more lines at the top end of the periodic table.
Hi Mike,
The table of emission lines that were first utilized by Probe Software for plotting KLM markers and other purposes were obtained from Dale Newbury and/or Nicholas Ritchie (I think!), back in 2000 or so (originally created by Chuck Fiori).
There are two versions. The first version was a text file called MASTER.LIN. It contains 4985 (1st order) emission lines, along with absorption edges. See attachments below. The file MASTER.TXT has a description of the data. Here is what it says:
This data base contains 4985 entries and includes all the measurable
X-ray lines, satellites and absorption edges from under 100 eV to
over 120 keV. Additionally, most of the X-ray lines and satellites
are assigned a relative intensity (relative to the alpha-1 line in
each family). The data base was assembled primarily from four
sources:
1.) B.L. Doyle, W.F. Chambers, T.M. Christensen, J.M. Hall and G.H.
Pepper "SINE THETA SETTINGS FOR X-RAY SPECTROMETERS", Atomic Data and
Nucleur Data Tables Vol. 24, No 5, 1979.
2.) E.W. White, G.V. Gibbs, G.G. Johnson Jr. and G.R. Zechman "X-RAY
WAVELENGTHS AND CRYSTAL INTERCHANGE SETTINGS FOR WAVELENGTH GEARED
CURVED CRYSTAL SPECTROMETERS" Report of the Pennsylvania State Univ.,
1964.
3.) J.A. Bearden "X-RAY WAVELENGTHS AND X-RAY ATOMIC ENERGY LEVELS"
Rev. Mod. Phys., Vol. 39, No. 78, 1967.
4.) J.A Bearden and A.F. Burr,"REEVALUATION OF X-RAY ATOMIC ENERGY
LEVELS", Rev. Mod. Phys., Vol. 31, No. 1, 1967.
Each X-ray line or edge series as a function of atomic number was fit
to a fourth degree polynomial. The fit was subtracted from the
appropriate data and the residuals plotted and examined. In this way
rogue entries could be identified and corrected. The resulting data
base is considered to be sufficiently accurate for any application
involving the Si(Li) X-ray detector and single crystal wavelength
spectrometers.
The data base is comprised of three data files: MASTER.LIN,
MASTER.TRS, and MASTER.ENG. These three files are identical except
that they have been sorted in different ways. MASTER.LIN is sorted
such that all entries belonging to a particular transition such as
KA2 are grouped together in ascending atomic number. MASTER.TRS is
sorted in such a manner that all the lines and edges associated with
a particular element are grouped together. These groups are in
ascending atomic number. And finally, MASTER.ENG is sorted in
ascending wavelength in Angstrom units.
Each data file is organized in the following manner: The first column
contains the atomic number and the second contains the atomic symbol.
The third column is the transistion or edge. Where possible the
transition is given in Siegbahn notation. If the entry is an
absorption edge the forth column contains the letters ABS, otherwise
the column is blank. the fifth column contains the wavelength in
Angstrom units. The sixth column contains the relative transition
probability expressed as a percentage of the principal line within
each family ie K, L or M alpha 1. Absorption edge entries contain the
value zero for this column with the anticipation of a future
inclusion of jump ratios. Finally, the last column gives a code for
the source of the entry. If the column is blank the source is
reference 2. If the column contains the letter "C" the source is
reference 1. If the letters "BB" appear, the source is reference 4.
The letters "W,F" mean that reference 2 was used but the relative
transition probability has been adjusted by Fiori. Reference 3 was
used as a check since it is the source of many of the entries of
reference 1.
In column 3 the notation KA1,2 means the entry is the weighted sum of
the KA1 and KA2 in the ratio 2 to 1. For low atomic number the
entries are not self consistent since the data is from different
sources. If the column begins with the capital letter S then the
entry is a satelite line due to doubly ionized atoms. The relative
transition values for these entries are only valid for electron
excited specimens, and are, at best, estimates.
For more information call Chuck Fiori (301-496 2599) or write to me
at Rm 3W-13 Bldg. 13, National Institutes of Health, Bethesda, Md.
20205. In any event, if you find this data useful please keep in
contact since we plan to write our work up formally and will no doubt
update and improve what is in the present files.
Th following are Siegbahn to shell-transition notation conversions:
You will have to use your imagination to discover which arabic letters
we used to correspond to the Siegbahn Greek notation:
KA =KA1+KA2+KA3
KA1,2=(2*KA1+KA2)/3
KA1 =K-L3
KA2 =K-L2
KA3 =K-L1
KB =SUM(KBn)
KBX =Metal
KB1 =K-M3
KB1' =KB1+KB3+KB5
KB2 =(K-N3)+(K-N2)
KB2' =K-N3
KB2''=K-N2
KB3 =K-M2
KB4 =(K-N4)+(K-N5)
KB5 =(K-M4)+(K-M5)
KB5' =K-M5
KB5''=K-M4
Kd1 =K-O3
Kd2 =K-O2
LA =LA1+LA2
LA1 =L3-M5
LA2 =L3-M4
LB1 =L2-M4
LB10 =L1-M4
LB15 =L3-N4
LB17 =L2-M3
LB2 =L3-N5
LB3 =L1-M3
LB4 =L1-M2
LB5 =(L3-O4)+(L3-O5)
LB6 =L3-N1
LB7 =L3-O1
LB9 =L1-M5
LG1 =L2-N4
LG11 =L1-N5
LG2 =L1-N2
LG3 =L1-N3
LG4 =L1-O3
LG4' =L1-O2
LG6 =L2-O4
LG8 =L2-O1
Ll =L3-M1
Ln =L2-M1
Ls =L3-M3
Lt =L3-M2
Lu =(L3-N6)+(L3-N7)
Lv =L2-N6
MA1 =M5-N7
MA2 =M5-N6
MB =M4-N6
MG =M3-N5
MG2 =M3-N4
MZ1 =M5-N3
MZ2 =M4-N2
Md =M2-N4
Me =M3-O5
Eventually we discovered that a number of Mz and Mg lines were missing from Chuck's original table and so Nicholas Ritchie created a new table which is also attached below. This table has 5713 emission lines, but no absorption edges.
But I don't have any documentation of this newer table, though Nicholas can help I am sure.
Finally, I attach below the XRAY.ELM file which is a text file I generated back in 2006, with all these lines and edges, and also higher order reflections with "nominal" intensities.
Hope this helps.
Thanks John, that's very useful :). If Nicholas can give a weblink I can cite that.