F Ka ROI artifact

[Probeman]

The detailed wavescan at F peak position on rutile on PC0 also finished and is attached, but its confusing me. I don't see any interfering peak there at all like we do in the TAP scans, yet in my MAN fits for PC0 my rutile definitely sits above the fitted line, and those MAN fits were constructed using only peak counting times of 12 seconds, so I should definitely see it in the wavescan where it was 40 seconds per point? I am now thoroughly confused...

Hi Ben,
The TiO2 in the F Ka MAN plots on PC1/PC0 don't look much like an outlier to me. Plot them with their error bars and see if you agree. 

This makes some sense to me because, if the peak we are seeing is a higher order reflection, the multi-layer crystals will suppress their reflection. Whereas the TAP crystal reflects higher orders quite well.
john

[BenjaminWade]

Hmm I agree with your reasoning, but not sure if I would agree with the TiO2 being within error. Looking at the attached MAN plot with 3 sigma error bars, I think I would argue its definitely outside of error. Also if I now go back and have a look at all my other MAN fits, TiO2 is consistently systematically higher. Having said that I cannot explain why I can't see any peak in PC0 though....

I ran the wavescan on Ti metal (attached) and as you can see the peak is there but higher intensity. I guess we have to conclude that its probably a obscure Ti interference not in the database. I did end up running detailed wavescans over the Nb, Mo, Ta, W, Cd, P peak positions, but nothing there. I guess probably Nb would have been the only realistic option given that it can be many 1000's ppm in rutile.

The PC0 is till bugging me though. I might try some different MP acquisitions on TiO2 and see if I can measure it.

cheers

[Probeman]

Hmm I agree with your reasoning, but not sure if I would agree with the TiO2 being within error. Looking at the attached MAN plot with 3 sigma error bars, I think I would argue its definitely outside of error. Also if I now go back and have a look at all my other MAN fits, TiO2 is consistently systematically higher. Having said that I cannot explain why I can't see any peak in PC0 though....

I ran the wavescan on Ti metal (attached) and as you can see the peak is there but higher intensity. I guess we have to conclude that its probably a obscure Ti interference not in the database. I did end up running detailed wavescans over the Nb, Mo, Ta, W, Cd, P peak positions, but nothing there. I guess probably Nb would have been the only realistic option given that it can be many 1000's ppm in rutile.

The PC0 is till bugging me though. I might try some different MP acquisitions on TiO2 and see if I can measure it.

cheers

Hi Ben,
Awesome work-  I was just making a (hopeful) guess on the PC1/PC0 MAN curves.  I agree TiO2 still looks outside of the error range on PC0.  But since the PC (and LDE) crystals do suppress higher order relfections, the peak may still be there, just much lower in intensity than TAP.  And indeed TiO2 is a higher outlier on the TAP MAN plots...

Note that I ran my wavescans at 80 sec because one often acquires say 3 or 4 points on each MAN standard and if one is using a 20 sec count time that is pretty high precision (the MAN intensities are the average of all points in each MAN standard).  Did you run your wavescan on TiO2 (or Ti) using PC0/PC1 with that sort of precision?

OK, this is interesting.  If I look at the x-ray database table in CalcZAF and increase the order limit to 20th order reflections I *do* see a Ti Kb1 and Kb3 lines, but *below* F Ka (not above as we are observing).  But this table isn't corrected for refractive index of the Bragg crystal, as it is in the PFE Plot! window!

Alas, if we plot the F Ka in TiO2 on TAP data with the Ti KLM markers on:



So the mystery emission line does not reveal its true identity easily!  Who will solve this conundrum?   
john

[BenjaminWade]

Hi John
Yes you are right, I forgot the MAN point is the average of my 4 repeat analyses. I am only using 12 sec peak count times for those MAN fits though, and my wavescan was 40sec/pt so would have thought I may have still seen it. I guess I should repeat it with 80sec just to make doubly sure!

Interesting, yes I played around cranking it to XX as well but didn't see anything there. But it still has to be Ti doesn't it given we see it on the Ti metal scan?

[Probeman]

A Ti emission line seems possible.   But I looked in a Penepma 40 hour simulation and I see no Ti lines near 675 eV.  Can't be oxygen since the mystery peak got larger in Ti metal.

[Probeman]

Just to show how tiny this mystery peak is, here a scan of fluor-phlogopite and TiO2 on TAP:

[DavidAdams]

I figured I will post my MAN question under this thread. I've been doing some extensive MANing recently and I thought I would take a more systematic approach to see how each of my crystals and spectrometers respond. I haven't gotten around the the LDE crystals yet, however. The PET and LiF crystals I have perform beautifully and make a fairly straight line across the standards I have been using. My issue is with my TAP crystal.  They all are showing some very non-linear behaviour. I have attached a file to show what I mean. I am a bit stumped to explain what's happening. Any suggestions or thoughts would be really appreciated!

[John Donovan]

I figured I will post my MAN question under this thread. I've been doing some extensive MANing recently and I thought I would take a more systematic approach to see how each of my crystals and spectrometers respond. I haven't gotten around the the LDE crystals yet, however. The PET and LiF crystals I have perform beautifully and make a fairly straight line across the standards I have been using. My issue is with my TAP crystal.  They all are showing some very non-linear behaviour. I have attached a file to show what I mean. I am a bit stumped to explain what's happening. Any suggestions or thoughts would be really appreciated!

Hi Dave,
This is an entirely appropriate topic to ask these questions.  The emission lines on the TAP crystal are lower energy and hence more subject to absorption (and surface contamination).

If you go back a few pages you can find Julien Allaz discussing MAN using TAP crystals here:

http://probesoftware.com/smf/index.php?topic=4.msg5295#msg5295

And Ben Buse discussing F Ka on TAP versus LDE1 using MAN here:

http://probesoftware.com/smf/index.php?topic=4.msg6230#msg6230

So *without* a blank correction the accuracy of the MAN method in silicates and oxides is limited to about 100 to 200 PPM. The accuracy will be less in higher Z materials due to the larger background correction.

But remember, the MAN correction can never overestimate your background correction, it can only underestimate it, that is if you have a subtle interference (or previously unsuspected contamination) in your standards used for the MAN calibration).

By the way I just run some trace elements in quartz using MAN (20 keV, 200 nA and 450 seconds count time) and here are the calculated detection limits for 5 points on my quartz standard (1.4 PPM Ti, 6 PPM Fe, 15 PPM Al and 0.03 PPM K from ICP-MS):

Un    7 std 14 as unk
TakeOff = 40.0  KiloVolt = 20.0  Beam Current = 200.  Beam Size =   20
(Magnification (analytical) =  24000),        Beam Mode = Analog  Spot
(Magnification (default) =      600, Magnification (imaging) =    616)
Image Shift (X,Y):                                         .00,    .00
Number of Data Lines:   5             Number of 'Good' Data Lines:   5
First/Last Date-Time: 10/04/2017 01:02:50 AM to 10/04/2017 01:40:49 AM
WARNING- No MAN Background Count Data Calculated for ti ka
WARNING- Using Blank Trace Correction
WARNING- Using Alternating On and Off Peak Acquisition
WARNING- Using Aggregate Intensities for Duplicate Elements

Average Total Oxygen:       53.257     Average Total Weight%:  100.000
Average Calculated Oxygen:  53.257     Average Atomic Number:   10.805
Average Excess Oxygen:        .000     Average Atomic Weight:   20.029
Average ZAF Iteration:        1.00     Average Quant Iterate:     4.00

Oxygen Calculated by Cation Stoichiometry and Included in the Matrix Correction
Element Si is Calculated by Difference from 100%

Un    7 std 14 as unk, Results in Elemental Weight Percents
 
ELEM:       Ti      Fe      Al      Ti       K      Si       O
TYPE:     ANAL    ANAL    ANAL    ANAL    ANAL    DIFF    CALC
BGDS:      MAN     MAN     MAN     MAN     MAN
TIME:   450.00  450.00  450.00     .00  450.00     ---     ---
BEAM:   200.37  200.37  200.37     .00  200.37     ---     ---
AGGR:        2                                     ---     ---

ELEM:       Ti      Fe      Al      Ti       K      Si       O   SUM 
XRAY:     (ka)    (ka)    (ka)    (ka)    (ka)      ()      ()
    98  .00016  .00043  .00180  .00000 -.00015 46.7411 53.2567 100.000
    99 -.00005  .00080  .00198  .00000  .00004 46.7408 53.2564 100.000
   100  .00013  .00081  .00193  .00000  .00025 46.7406 53.2563 100.000
   101  .00021  .00078  .00129  .00000 -.00022 46.7413 53.2566100.0000
   102  .00021  .00018  .00112  .00000  .00006 46.7417 53.2567 100.000

AVER:   .00013  .00060  .00162  .00000  .00000  46.741  53.257 100.000
SDEV:   .00011  .00028  .00039  .00000  .00018    .000    .000  .00000
SERR:   .00005  .00013  .00018  .00000  .00008  .00020  .00007
%RSD:  79.9429 47.4139 24.1568   .0000 -8046.5  .00096  .00031
STDS:       22     395     374       0     374     ---     ---

STKF:    .5616   .6862   .0628   .0000   .1102     ---     ---
STCT:  71054.0 32507.0  3454.7      .0  3953.0     ---     ---

UNKF:    .0000   .0000   .0000   .0000   .0000     ---     ---
UNCT:       .1      .2      .7      .0      .0     ---     ---
UNBG:     89.4    34.1    34.6      .0    15.9     ---     ---

ZCOR:   1.1969  1.1746  1.3479   .0000  1.2098     ---     ---
KRAW:   .00000  .00001  .00019  .00000  .00000     ---     ---
PKBG:  1.00160 1.00712 1.01917  .00000  .99996     ---     ---
INT%:     ----    ----    ----    ----    ----     ---     ---
BLNK#:       7       7       7       0       7     ---     ---
BLNKL: .000140 .000600 .001500       0 .000000     ---     ---
BLNKV: -.00181 -.00314 -.00262       0 .001041     ---     ---

Detection limit at 99 % Confidence in Elemental Weight Percent (Single Line):

ELEM:       Ti      Fe      Al      Ti       K
    98  .00022  .00053  .00053  .00000  .00049
    99  .00033  .00053  .00053  .00000  .00049
   100  .00033  .00053  .00053  .00000  .00049
   101  .00033  .00053  .00053  .00000  .00049
   102  .00033  .00053  .00053  .00000  .00049

AVER:   .00030  .00053  .00053  .00000  .00049
SDEV:   .00005  .00000  .00000  .00000  .00000
SERR:   .00002  .00000  .00000  .00000  .00000

Detection Limit (t-test) in Elemental Weight Percent (Average of Sample):

ELEM:       Ti      Fe      Al      Ti       K
  60ci  .00005  .00015  .00022     ---  .00011
  80ci  .00009  .00024  .00035     ---  .00017
  90ci  .00012  .00034  .00049     ---  .00024
  95ci  .00016  .00044  .00064     ---  .00032
  99ci  .00026  .00073  .00106     ---  .00052

Note that Ti is acquired on two spectrometers using the aggregate method, but it bests my previous record of 2-3 PPM using off-peak measurements on all *five* spectrometers for Ti.  That is a 2.6 PPM t-test detection limit on Ti isn't too bad, using MAN on only two spectrometers!

What I think you are seeing in your TAP MAN plots (going up to Z = 90?), are very small spectral interferences (or contamination in some cases).  Remember, those high order Bragg reflections are ubiquitous on TAP crystals...

The fact that they show some beautiful "progressions" with Z, tells me these are spectra interferences.

Here's what I tell my students: because background is *by definition* the lowest thing we can measure, any points that plot above the trend are either interference or contamination (or a little of both). Therefore, you can only remove those standards that plot above the trend in the MAN calibration curve.

Because the MAN intensities are corrected for absorption, there should be no sample absorption edges, so the lowest points should represent the actual background.
john

[Probeman]

Hey, at least one can only have interferences on the on-peak intensities when using MAN background methods!

As opposed to the off-peak method which can have interferences both on the peak and also on the off-peak positions!

[John Donovan]

By the way, in the above run I only had Al ka on TAP, and while one could wish for more intermediate points around Z = 14 to 18, the fit is pretty smooth:



Unfortunately my TiO2 has about 100 PPM of Al so it cannot be used.
john

[John Donovan]

Unfortunately my TiO2 has about 100 PPM of Al so it cannot be used.
john

If you have an Al-free piece of V2O5, that's at about clocks in at Zbar ~ 16.4, or one of the synthetic Cl-apatites (~14.5).

Is that 100ppm real, or is that the Ti Ka III? Even in diff mode with a tight-ish window, I usually ended up excluding Ti-bearing standards from Al MAN curves.

Hi Owen,
The Al in TiO2 could certainly could be an interference from Ti Ka III, but the note in my standard database says that the 110 PPM Al is an "interference corrected" value...  but that measurement was performed over 25 years ago so it should probably be revisited!

My V2O3 is from Purdue University:

Harrison, et al. 1980, Mat. Res. Bull., v. 15, pp. 571-580

It's behind a paywall so I can't read it and see how much Al might be present.  I have no mention of Al in my standard database for this material.
john

[JohnF]

A question by John Fournelle was particularly interesting and I had some additional thoughts on it. His question was: you said that background is generally the lowest thing we can measure, does the MAN correction handle situations where there are "holes" in the continuum?

Actually I wasn't referring to a "hole in the background" explicitly, rather to a "low spot" in the monazite bkg wavescan you showed, where the modeled background was higher than that low spot....I figured there must be some reason for the low spot, which could perhaps be a hole or perhaps not, but something else which might have another physical explanation.

[Probeman]

Actually I wasn't referring to a "hole in the background" explicitly, rather to a "low spot" in the monazite bkg wavescan you showed, where the modeled background was higher than that low spot....I figured there must be some reason for the low spot, which could perhaps be a hole or perhaps not, but something else which might have another physical explanation.

OK, that is probably the Ar absorption edge from the detector (unlike the hole near the Au La from secondary Bragg diffraction). This paper by Jercinovic does discuss the issue of absorption edges creating "holes" in the continuum (see fig 23 for example):

http://www.geology.wisc.edu/~johnf/g777/AmMin/Jercinovic_2005.pdf

Of course this is the reason (traditional off-peaks or multi-point bgds), we generally don't want to interpolate off-peak measurements across an absorption edge.  Wow, absorption edges from the sample, absorption edges from the Bragg crystal and absorption edges also from the detector. It's a jungle out there!

The interesting thing to me was that if the "hole" in the background is due to secondary diffraction from the Bragg crystal (or the Ar absorption edge from the detector), it won't affect the MAN method, but holes in the background due to an absorption edge in the sample could. 

Though now that I think about it more, I am going to bet that the continuum absorption correction in the MAN background correction will take care of these sample absorption edges automatically!  I mean that's what an absorption correction is for, right?  But I wouldn't try the MAN bgd correction for monazite because it would be too difficult to obtain a suitable blank for checking accuracy.  But assuming one can find a suitable blank, the MAN method might deal with all three types of absorption edges automatically!

I can't wait to get back in the lab and try measuring Au La in pyrite using MAN.  I've never tried it, but my pure pyrite standard will make a great blank so it's worth a try...

[Probeman]

Last month Ben Wade posted this interesting observation of a mystery emission line near F Ka in TiO2:

http://probesoftware.com/smf/index.php?topic=4.msg6305#msg6305

Here is a plot of fluor-phlogopite and TiO2 normalized intensity.  I wanted to see if I could say whether the mystery line was a first order emission line or not:



Just eyeballing it I'd say it looks about as wide as the F Ka emission, so yes, I think it is a first order line, but next I'll try some really high precision wavescans.
john

[Probeman]

I have a question about the shape of the MAN background calibration curve.  Sometimes when one plots up a large range of average Z, the trends show a positive curvature where the continuum intensities decrease with increasing Z as seen here:

http://probesoftware.com/smf/index.php?topic=4.msg5136#msg5136

For light elements or other low energy emission lines this could be due to increasing absorption in higher Z matrices, but the MAN absorption correction should deal with that I would think.  So what else could it be?  Obviously there is more backscatter at high Z, so there are fewer electrons staying in the matrix and hence less continuum x-ray generation. Also higher Z elements tend to produce less intensity for a given emission line since there are so many other probabilities for absorption and fluorescence (and emission).

From that same zircon run above here is the MAN curve for Hf Ma (a relatively low energy emission line) with the continuum absorption correction applied:



and here is the same data without the continuum absorption correction applied:



But the difference in curvature is slight. We need a suite of low energy x-ray emission lines measured over a large range of average Z to see better what is going on with continuum emission.