Using Probe for EPMA software in "demonstration mode" to teach EPMA

[Mike Matthews]

Another thought occurs to me: EDS detectors are usually at a fixed distance from the sample, but for WDS spectrometers, there is obviously a large decrease in geometric efficiency as the spectrometer is driven to high sin thetas and the distance to the sample increases.

That's a simple enough math problem, but there's a complication: the crystal is essentially perpendicular to the sample at high sin thetas, but tilts as it approaches the sample at lower sin thetas, therefore it presents a smaller area to the x-rays coming from the sample.  This is decrease in "aspect" geometric efficiency is inverse, compared to the increase in geometric efficiency from decreasing the distance to the sample at lower sin thetas.

I'm not sure how to calculate this change in geometric efficiency as a function of sin theta. Has anyone stumbled a cross a table or plot of this change in WDS geometric efficiency that I could use for my WDS simulation code?
john

I can provide some information on this: Detector efficiency is a combination of 3 factors: i) As John pointed out, the effective size of the crystal (solid angle) as it moves and tilts along the range, ii) crystal reflectivity, and iii) counter efficiency.

The solid angle can be calculated (see slide 33 in the attached pdf) but decreases fairly steadily from low to high angles. For the crystal reflectivity I've only found sketchy experimental data, but again generally decreases from high to low angles (slide 34). I only have half the curve for TAP, so if anyone has the complete curve for TAP I'd be interested to know what it does in the upper half of the range. Counter efficiency is itself a function of window absorption and gas absorption (slides 24 and 25). Slides 35 and 36 give the resulting intensities for Ka and La x-rays respectively, generated by extracting ideal peak intensity values for pure elements from Stephen Reed's Virtual WDS program so are based on experimental results from Cameca spectrometers. The La plot very nicely shows the effect of the Ar absorption edge.

[John Donovan]

Another thought occurs to me: EDS detectors are usually at a fixed distance from the sample, but for WDS spectrometers, there is obviously a large decrease in geometric efficiency as the spectrometer is driven to high sin thetas and the distance to the sample increases.

That's a simple enough math problem, but there's a complication: the crystal is essentially perpendicular to the sample at high sin thetas, but tilts as it approaches the sample at lower sin thetas, therefore it presents a smaller area to the x-rays coming from the sample.  This is decrease in "aspect" geometric efficiency is inverse, compared to the increase in geometric efficiency from decreasing the distance to the sample at lower sin thetas.

I'm not sure how to calculate this change in geometric efficiency as a function of sin theta. Has anyone stumbled a cross a table or plot of this change in WDS geometric efficiency that I could use for my WDS simulation code?
john

I can provide some information on this: Detector efficiency is a combination of 3 factors: i) As John pointed out, the effective size of the crystal (solid angle) as it moves and tilts along the range, ii) crystal reflectivity, and iii) counter efficiency.

The solid angle can be calculated (see slide 33 in the attached pdf) but decreases fairly steadily from low to high angles. For the crystal reflectivity I've only found sketchy experimental data, but again generally decreases from high to low angles (slide 34). I only have half the curve for TAP, so if anyone has the complete curve for TAP I'd be interested to know what it does in the upper half of the range. Counter efficiency is itself a function of window absorption and gas absorption (slides 24 and 25). Slides 35 and 36 give the resulting intensities for Ka and La x-rays respectively, generated by extracting ideal peak intensity values for pure elements from Stephen Reed's Virtual WDS program so are based on experimental results from Cameca spectrometers. The La plot very nicely shows the effect of the Ar absorption edge.

Hi Mike,
Right now I'm only interested in the geometric efficiency of the Bragg crystal (solid angle) as a function of spectrometer position, so your slide 33 is perfect for my purposes.  Thanks!
john

[Probeman]

I recently improved the EDS (demo) simulation mode to produce more realistic count rates when the element in question is not produced within the simulation time. This is primarily for elements that are measured but not present in the simulated composition, e.g., Ti in pure SiO2, or trace elements that do not produce an emission in the allotted simulation time.  See highlighted text below.
john

St   12 Set   6 MgO synthetic, Results in Elemental Weight Percents
 
ELEM:       Si      Fe      Mg      Ti      Si      Fe      Mg      Ti      Ca       O
TYPE:     ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    SPEC    SPEC
BGDS:      LIN     LIN     LIN     LIN     EDS     EDS     EDS     EDS
TIME:      ---     ---     ---     ---  120.00  120.00  120.00  120.00     ---     ---
BEAM:      ---     ---     ---     ---   30.00   30.00   30.00   30.00     ---     ---

ELEM:     Si-D    Fe-D    Mg-D    Ti-D      Si      Fe      Mg      Ti      Ca       O   SUM
XRAY:     (ka)    (ka)    (ka)    (ka)    (ka)    (ka)    (ka)    (ka)      ()      ()
    21     ---     ---     ---     ---    .001    .008  60.309    .004    .140  39.693 100.155
    22     ---     ---     ---     ---    .001    .009  60.181   -.001    .140  39.693 100.024
    23     ---     ---     ---     ---   -.001    .000  60.336    .000    .140  39.693 100.168

AVER:      ---     ---     ---     ---    .000    .006  60.275    .001    .140  39.693 100.115
SDEV:      ---     ---     ---     ---    .001    .005    .082    .003    .000    .000    .080
SERR:      ---     ---     ---     ---    .001    .003    .048    .002    .000    .000
%RSD:      ---     ---     ---     ---  410.99   87.44     .14  230.29     .00     .00

PUBL:     n.a.    n.a.    n.a.    n.a.    .005    .008  60.280    n.a.    .140  39.693 100.126
%VAR:      ---     ---     ---     ---  -93.38  -28.90  (-.01)     ---     .00     .00
DIFF:      ---     ---     ---     ---   -.005   -.002   (.00)     ---    .000    .000
STDS:      ---     ---     ---     ---      14     162      12      22     ---     ---

If you haven't tried the new EDS and WDS simulation in PFE you should-  just update PFE on any off-line computer in your lab or office or home, using the Help menu.  Also update the Penepma files from the same Help menu dialog by checking the Update Penepma Monte Carlo Files Only checkbox.

Then (if the InterfaceType=0 in the Probewin.ini file) simply click Yes when asked "Do you want to connect to the instrument?" and then from the Acquire! window, add some WDS elements and/or check the acquire EDS acquisition checkbox under Acquisition Options.  You will automatically acquire simulated WDS and/or EDS spectra.

[John Donovan]

Hi all,
With Xavier's help I figured out how to change the random seed for the EDS simulation using Penepma.  Turns out that not only does one need to specifiy a different negative number for the seed, but the "RSEED" line in the Penepma input file needs to be *before* the the "TIME" line!

Anyway, now replicate points more accurately reflect the actual counting statistics for the EDS simulation as seen here:

St  160 Set   2 NBS K-412 mineral glass, Results in Elemental Weight Percents
 
ELEM:       Si      Fe      Mg       O      Ca      Al      Mn
TYPE:     ANAL    ANAL    ANAL    ANAL    SPEC    SPEC    SPEC
BGDS:      EDS     EDS     EDS     EDS
TIME:   120.00  120.00  120.00  120.00     ---     ---     ---
BEAM:    30.01   30.01   30.01   30.01     ---     ---     ---

ELEM:       Si      Fe      Mg       O      Ca      Al      Mn   SUM 
    36  22.960   4.290  10.825  45.909  10.899   4.906    .077  99.866
    37  20.849   8.305  10.669  44.633  10.899   4.906    .077 100.338
    38  22.890   9.928  12.470  45.578  10.899   4.906    .077 106.748

AVER:   22.233   7.507  11.321  45.373  10.899   4.906    .077 102.317
SDEV:    1.199   2.902    .998    .662    .000    .000    .000   3.844
SERR:     .692   1.676    .576    .382    .000    .000    .000
%RSD:     5.39   38.66    8.81    1.46     .00     .00     .00

PUBL:   21.199   7.742  11.657  43.597  10.899   4.906    .077 100.077
%VAR:     4.88   -3.03   -2.88    4.07     .00     .00     .00
DIFF:    1.034   -.235   -.336   1.776    .000    .000    .000
STDS:       14     162      12      14     ---     ---     ---

STKF:    .4101   .0950   .4736   .2664     ---     ---     ---
STCT:   100.01    3.81  142.18   75.57     ---     ---     ---

UNKF:    .1708   .0634   .0755   .1822     ---     ---     ---
UNCT:    41.66    2.55   22.66   51.69     ---     ---     ---
UNBG:      .03     .00     .04     .12     ---     ---     ---

ZCOR:   1.3016  1.1854  1.4995  2.4908     ---     ---     ---
KRAW:    .4166   .6672   .1594   .6839     ---     ---     ---
PKBG:  1450.95  575.62  529.20  439.17     ---     ---     ---

Remember, to get EPMA like statistics, one needs to run the Penepma EDS simulation some 10 or 15 minutes or more!
john

PS to utilize this new random seed feature, you will need to update your Penepma files using the Help | Update Probe for EPMA menu and check the box that says: Update Penepma Monte Carlo Files Only.  And of course you'll need to update PFE again also.

[John Donovan]

I've improved the WDS simulation to now include detector absorption edges for Ar and Xe. 

Here is an actual scan on my Sx100 on a pure synthetic zircon crystal:



and here is a WDS simulation on the same synthetic zircon:



Not too bad actually.  Here is the experimental scan with the KLM markers for Si, O and Zr for reference:



We're not seeing the satellite lines as expected, since Penepma only does singly ionized atoms, but good enough for teaching purposes!
john

[John Donovan]

One more thing I should point out: the WDS simulation in PFE makes a new composition whenever a new standard is selected.  However, if the sample setup does not change, the program will utilize the same composition for unknowns and wavescans. 

But, only if the sample setup has *not* changed.  This means that not only must the keV be the same, but also the WDS elements and any specified (non-analyzed) elements. If the sample setup has changed, the program will synthesize a random composition based on the current WDS elements.

For example, in the above zircon standard, I added Zr as a specified element to the unknown (and wavescan) sample setups, because it is present in the composition, but is not one of the WDS elements being analyzed in the simulation.

That way, once the zircon standard has run, I simply create a new unknown (or wavescan), with the previously specified element Zr, and the program does not create a random unknown (or wavescan) composition for simulation, but instead utilizes the previous standard composition.

I hope that makes sense, please let me know if not.
john

[John Donovan]

Here's an EDS simulation of a NIST glass running the Penepma Monte Carlo with the random seed added:

St  160 Set   3 NBS K-412 mineral glass, Results in Elemental Weight Percents
 
ELEM:       Si      Fe      Mg       O      Ca      Al      Mn
TYPE:     ANAL    ANAL    ANAL    ANAL    SPEC    SPEC    SPEC
BGDS:      EDS     EDS     EDS     EDS
TIME:   600.00  600.00  600.00  600.00     ---     ---     ---
BEAM:    30.00   30.00   30.00   30.00     ---     ---     ---

ELEM:       Si      Fe      Mg       O      Ca      Al      Mn   SUM 
    53  21.126   8.607  11.594  44.923  10.899   4.906    .077 102.132
    54  20.967   7.960  11.088  43.319  10.899   4.906    .077  99.216
    55  21.034   6.699  12.027  43.500  10.899   4.906    .077  99.141

AVER:   21.042   7.755  11.570  43.914  10.899   4.906    .077 100.163
SDEV:     .080    .970    .470    .878    .000    .000    .000   1.706
SERR:     .046    .560    .271    .507    .000    .000    .000
%RSD:      .38   12.51    4.06    2.00     .00     .00     .00

PUBL:   21.199   7.742  11.657  43.597  10.899   4.906    .077 100.077
%VAR:     -.74     .17    -.75     .73     .00     .00     .00
DIFF:    -.157    .013   -.087    .317    .000    .000    .000
STDS:       14     162      12      14     ---     ---     ---

Close to typical EPMA precision and accuracy in 600 sec (10 min) per point of simulation. And it will improve with faster CPUs.

[John Donovan]

Just for fun last night I ran 1200 sec per point with the EDS (and WDS) simulation methods and got very pretty looking data:

St  162 Set   1 NBS K-411 mineral glass, Results in Elemental Weight Percents
 
ELEM:       Ti      Mn      Ca      Fe      Mg      Al      Si       O
TYPE:     ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    SPEC
BGDS:      LIN     LIN     LIN     LIN     EDS     EDS     EDS
TIME:    30.00   30.00   30.00   30.00 1200.00 1200.00 1200.00     ---
BEAM:    29.99   29.99   29.99   29.99   29.99   29.99   29.99     ---

ELEM:       Ti      Mn      Ca      Fe      Mg      Al      Si       O   SUM 
    21    .002    .079  10.964  11.394   8.981    .077  25.264  43.558 100.320
    22    .032    .071  11.109  11.285   8.374    .090  25.345  43.558  99.864
    23    .022    .080  10.953  11.395   8.688    .030  25.740  43.558 100.466
    24    .003    .083  11.119  11.415   9.032    .045  24.684  43.558  99.939
    25   -.018    .081  11.115  11.383   8.661    .038  25.107  43.558  99.925

AVER:     .008    .079  11.052  11.374   8.747    .056  25.228  43.558 100.103
SDEV:     .019    .005    .086    .051    .268    .026    .383    .000    .271
SERR:     .009    .002    .038    .023    .120    .012    .171    .000
%RSD:   238.76    5.99     .77     .45    3.06   46.69    1.52     .00

PUBL:     n.a.    .077  11.057  11.209   8.847    .053  25.382  43.558 100.183
%VAR:      ---    2.38  (-.04)    1.47   -1.13    5.61    -.61     .00
DIFF:      ---    .002   (.00)    .165   -.100    .003   -.154    .000
STDS:       22     160     162     263      12      13      14     ---

Even the trace Al by EDS (simulation) looks quite good, better than actual EDS!   

Remember, you will need to update PFE and the Penepma Monte Carlo files from the PFE Help menu to obtain these new simulation features.

[Probeman]

The only trouble we sometimes run into when students install CalcZAF/Probe for EPMA on their laptops is that some of the larger windows (in particular the Automate! window in PFE), are a little bigger than what the default screen resolution can display.

I don't know much about the newer operating systems such as Windows 10, but on my Win 10 laptop which has a fairly large screen, I can see the enough of the Automate! window so it's not a problem.  But some student laptop screens are smaller and the entire Automate! window doesn't display, so they can't reach the Run Selected Samples button at the bottom...

One might right click the top of the window, select Move and use the cursor keys to move the window around I guess, but does anyone have any better suggestions on what desktop or video card options should be modified to get a somewhat larger "virtual" desktop, so one can scroll the desktop a bit to see a little more on the sides and top/bottom of the visible desktop?
john

[Probeman]

The latest Penepma12 update contains pure element spectra for all the rare earths and many actinides, so I tried a WDS simulation of a monazite composition.





Can you tell which one is experimental and which one is simulation?
john

[John Donovan]

One "feature/problem" of the WDS simulations synthesized from Penepma pure element spectra in PFE is that because the peaks are convolved using a simple Gaussian method, the long Lorentzian tails that we normally see in our scans are not present. This means that spectral overlaps are smaller than expected and most overlaps simply do not appear in these synthetic spectra as shown here in an example of the Ba-Ti overlap in benitoite (Ba-Ti silicate):



These Lorentzian tails on the emission peaks are mostly from randomized recrystallization of micro domains in the Bragg analyzing crystals during the "polygonization" step in manufacturing where the Bragg crystals are thermally cycled to improve reflectivity...  see the attached figure below for details (I cannot remember where I got this figure so please remind me if you recognize it, so I can properly attribute it).

However, our old nemesis (Pb La - As Ka) is still a significant interference even with these simulated Gaussian convolutions:



john

[John Donovan]

If I run another simulation on benitoite, but this time using PET crystals for Ti and Ba, we get this:



Notice how the La2 line is now convolved with the PET crystal, while it was quite well separated with the LIF crystal plot in the previous post. Still not a significant overlap due to a lack of the above mentioned Lorentzian tails, but maybe good enough for teaching?
john

[Probeman]

As you know, Probe for EPMA, when run in "demo" mode (the keyword InterfaceType=0 in the probewin.ini file), can simulate both EDS and WDS spectra for performing spectrum simulations for teaching EPMA in the classroom.

The latest version of Probe for EPMA now has a complete set of pure element WDS spectrum simulations at 15 keV from the Penepma Monte Carlo software and is distributed in the Penepma12.zip update which can be downloaded using the Help menu in Probe for EPMA as described here:

http://probesoftware.com/smf/index.php?topic=366.msg1936#msg1936

These pure element spectra are utilized to synthesize WDS spectra for simulation of wavescans on compounds.  I am currently running simulations also at 5, 10, 20 and 25 keV and many of these are already in the Penepma12.zip distributions though they are not 100% complete.
john

[John Donovan]

The latest version of Probe for EPMA now has additional menus to easily switch between JEOL and Cameca simulation modes as seen here:



These menus are only available when the InterfaceType keyword in the Probewin.ini file is zero.  The JEOL simulation mode is the default interface option in Probe for EPMA when PFE is installed the first time.

The idea being that you can distribute PFE to all your students in class and if necessary switch to Cameca simulation mode (and back again to JEOL simulation mode  if desired), with a single menu click.
john

[Ben Buse]

As you know, Probe for EPMA, when run in "demo" mode (the keyword InterfaceType=0 in the probewin.ini file), can simulate both EDS and WDS spectra for performing spectrum simulations for teaching EPMA in the classroom.

The latest version of Probe for EPMA now has a complete set of pure element WDS spectrum simulations at 15 keV from the Penepma Monte Carlo software and is distributed in the Penepma12.zip update which can be downloaded using the Help menu in Probe for EPMA as described here:

http://probesoftware.com/smf/index.php?topic=366.msg1936#msg1936

These pure element spectra are utilized to synthesize WDS spectra for simulation of wavescans on compounds.  I am currently running simulations also at 5, 10, 20 and 25 keV and many of these are already in the Penepma12.zip distributions though they are not 100% complete.
john

Hi John,

I'm struggling to figure out how to get compound spectra - I have Si element but the wavescan is just for Si Metal. Standards added are plagioclase, hornblende etc

Thanks

Ben