Totals both too low and too high for Fe-Ti oxides, rutile and spinel?

[GarnetSnowball]

Hi everyone,

I'm quite new to EPMA, having done some analyses as part of my Bsc(Hons) thesis project.

I analysed many silicates from some samples and received good totals.

On a separate run, but on the same instrument, I analysed oxides (namely Fe-Ti oxides like ilmenite and magnetite, but also rutile, spinel and corundum).
The totals are, for the most part, atrociously bad. The relevant file is attached.

Specifically, the totals on some analyses are ridiculously low, and others (in particular a nasty looking rutile with some ilmenite exsolution(?)) would be more or less correct except for >100% TiO2 and about 10% SiO2 and ~3% Al2O3 which make the totals far too high.

I also tried using the Droop's method option under calculation options but this didn't seem to improve things much at all.

I also tried analysing some different samples (a rutile) in another machine, and the situation was much the same.

The standards all seem OK.

There seems to be 2 situations:

1) totals too high - the TiO2 totals are way over 100%, and we have SiO2 and Al2O3 where I only expect TiO2 and Fe oxides. Could this be secondary fluorescence effects? I have read a little and not found a situation where the totals are increased to the numbers I see.

2) totals are too low (in some cases 50-60%), with just FeO and TiO2.  No idea?

If anyone is able to enlighten me as to the possible causes and/or ways to diagnose this problem, that would be appreciate. As I mentioned, I'm only new to EPMA so I want to treat this as a learning opportunity.

Also, if there is any other information I should provide, please let me know.

[Probeman]

This run is a "hot mess", you should start over and try again.  Good first effort though (my first runs were much worse!).   

To begin with all your unknowns are missing oxygen by stoichiometry.   This is easily remedied using the "Calculate with Stoichiometric Oxygen" option in the Calculations Options dialog in Analyze!.

But your standards are not OK.  First I would avoid Fe metal as a standard because there is a subtle issue with chemical peak shifting as described here:

https://probesoftware.com/smf/index.php?topic=1423.0

Also you have apparently analyzed some standards incorrectly as seen here:

St  811 Set   1 Spinel, Results in Elemental Weight Percents
 
ELEM:       Al      Ni      Mg      Si      Ti       V      Cr      Mn      Fe      Zn      Ca       O
TYPE:     ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    ANAL    SPEC
BGDS:      MAN     MAN     MAN     MAN     MAN     MAN     MAN     MAN     MAN     MAN     MAN
TIME:    40.00   40.00   40.00   40.00   40.00   40.00   40.00   40.00   40.00   40.00   40.00     ---
BEAM:    20.28   20.28   20.28   20.28   20.28   20.28   20.28   20.28   20.28   20.28   20.28     ---

ELEM:       Al      Ni      Mg      Si      Ti       V      Cr      Mn      Fe      Zn      Ca       O   SUM 
    55  47.111   -.008   6.211   -.006    .002   -.003   -.030    .968   -.006    .016    .000  44.990  99.245
    56  47.083    .000   6.216   -.001   -.022   -.001   -.024    .929    .006   -.030   -.003  44.990  99.143
    57  47.162    .007   6.221   -.011    .033   -.006   -.031    .948   -.018   -.011   -.007  44.990  99.276

AVER:   47.119    .000   6.216   -.006    .004   -.003   -.028    .948   -.006   -.009   -.003  44.990  99.221
SDEV:     .040    .008    .005    .005    .027    .003    .004    .019    .012    .023    .004    .000    .069
SERR:     .023    .004    .003    .003    .016    .002    .002    .011    .007    .013    .002    .000
%RSD:      .08-2979.23     .08  -82.44  653.42  -79.27  -12.66    2.05 -209.03 -271.07 -119.25     .00

PUBL:   37.930    n.a.  17.080    n.a.    n.a.    n.a.    n.a.    n.a.    n.a.    n.a.    n.a.  44.990 100.000
%VAR:    24.22     ---  -63.61     ---     ---     ---     ---     ---     ---     ---     ---     .00
DIFF:    9.189     --- -10.864     ---     ---     ---     ---     ---     ---     ---     ---    .000
STDS:      319     211     537     537     140     212     222     506     243     226     214     ---

In addition, your Ti standard has excessive drift:

Drift array standard intensities (cps/19.85nA) (background corrected) (TDI corrected):
ELMXRY:    al ka   ni ka   mg ka   si ka   ti ka    v ka   cr ka   mn ka   fe ka   zn ka   ca ka
MOTCRY:  2   TAP 5   LiF 2   TAP 2   TAP 4   LiF 3  LiFH 3  LiFH 3  LiFH 5   LiF 5   LiF 1  PETJ
STDASS:      319     211     537     537     140     212     222     506     243     226     214
STDVIR:        0       0       0       0       0       0       0       0       0       0       0
         29349.5  8067.8  5827.4  8877.2  1670.7 27630.8 12122.2 34156.3  5090.2  6881.3  9417.5
              -   7997.4      -       -   1102.3 27491.8      -       -       -       -       -

Remember, the "List Standard Intensities" button is your friend...   

You should consult with your lab manager and re-run all your standards and unknown materials.  You made a good start, but please try again.

By the way, for unknowns such as hematite, you will obtain totals around 88 wt% (once you've included stoichiometric oxygen), but you'll still need to include oxygen from ferric iron.  See here for more discussion:

https://probesoftware.com/smf/index.php?topic=92.msg8593#msg8593

And finally, the analyses I saw with high totals (~105%) look like rhyolite glass. Not sure why the totals are high, but probably bad standards.

[GarnetSnowball]

Thank you Probeman.

Yes, you're right. I've added stoichiometric oxygen now.

I've now changed the standard assignment from Fe metal to Hematite, using the Calculate Excess Oxygen Correction.. as you suggested.

The spinel was a dub but was not being used as a standard for anything so this shouldn't be an issue right?

Can the drift on the titanium standard be corrected for at all? I should have added another Ti standard in...

[Probeman]

The spinel was a dub but was not being used as a standard for anything so this shouldn't be an issue right?

As long as it's not utilized as a standard (primary, interference or MAN) for anything.

Can the drift on the titanium standard be corrected for at all? I should have added another Ti standard in...

The question really is: why did the Ti standard drift so much?

I didn't look at it that closely, but if you can be sure that one of those standardizations is correct, then you could disable the other standardization.  Or just ignore those samples with significant Ti.

Or re-run the Ti rich samples is probably your best bet.

[GarnetSnowball]

Probeman,

I don't quite understand what you're referring to when you say standardizations. Do you refer to the two numbers you highlighted in your first reply?

The 'rhyolite glass' is unusual. The rock is a garnet-sillimanite gneiss with rutile and ilmenite. The grains which gave these high totals look really horrible in BSE - lots of little inclusions/exsolution textures. The probe results give ~2-3% SiO2 and ~1-2% Al2O3, which is strange, and these elements show up in the EDS spectrum too. Could there be secondary fluorescence from the slide glass or inclusions that increase the totals?

At this point I'm just trying to assemble some results for a presentation in a couple of weeks - I still have some time left for additional analyses at the end of the year so will re-run then. I will definitely discuss these issues I've had during my presentation, it's been a good learning experience.

[Probeman]

I don't quite understand what you're referring to when you say standardizations. Do you refer to the two numbers you highlighted in your first reply?

A standardization is a standard sample that is assigned as a primary (or interference or MAN) standard for an element. Here the 1670.7 intensity is the first standardization (set 1) and the 1102.3 intensity is the second standardization (set 2). Of course there won't be a one to one mapping of the order and set number if a standard sample was acquired and then disabled or more standard sets were acquired for one element compared to another element.

Drift array standard intensities (cps/19.85nA) (background corrected) (TDI corrected):
ELMXRY:    al ka   ni ka   mg ka   si ka   ti ka    v ka   cr ka   mn ka   fe ka   zn ka   ca ka
MOTCRY:  2   TAP 5   LiF 2   TAP 2   TAP 4   LiF 3  LiFH 3  LiFH 3  LiFH 5   LiF 5   LiF 1  PETJ
STDASS:      319     211     537     537     140     212     222     506     243     226     214
STDVIR:        0       0       0       0       0       0       0       0       0       0       0
         29349.5  8067.8  5827.4  8877.2  1670.7 27630.8 12122.2 34156.3  5090.2  6881.3  9417.5
              -   7997.4      -       -   1102.3 27491.8      -       -       -       -       -

The 'rhyolite glass' is unusual. The rock is a garnet-sillimanite gneiss with rutile and ilmenite. The grains which gave these high totals look really horrible in BSE - lots of little inclusions/exsolution textures. The probe results give ~2-3% SiO2 and ~1-2% Al2O3, which is strange, and these elements show up in the EDS spectrum too. Could there be secondary fluorescence from the slide glass or inclusions that increase the totals?

Yes, it is common to have high totals when the interaction volume is not homogeneous. Which is why the so-called "defocus beam" technique does not yield quantitative results:

https://probesoftware.com/smf/index.php?topic=198.msg8908#msg8908

At this point I'm just trying to assemble some results for a presentation in a couple of weeks - I still have some time left for additional analyses at the end of the year so will re-run then. I will definitely discuss these issues I've had during my presentation, it's been a good learning experience.

That's the right attitude- good for you!