Ti in Quartz (without PFE)

[Probeman]

Hi John
I will attempt to get some long count time wavescans later this weekend on all 5 PETs and post what I find...

cheers

Hi Ben,
I would be very interested to see what you find.

Does anyone else have both normal and large area PET crystals?  Please try acquiring a long (300 seconds per point) wavescan on a very pure SiO2 material using all your spectrometers and let's see what we get.
john

[BenjaminWade]

Just to query/confirm, it doesn't necessarily have to be pure SiO2 does it? It just has to be "clean" of interferences in that region of the spectrum? So could be something more conductive and beam stable like a metal right? Or am I wrong?

[Probeman]

That's a reasonable question. My thoughts are that if it is a Bragg crystal artifact, whether it's secondary Bragg diffraction effect or some sort of crystal distortion from bending during manufacturing, I agree it shouldn't matter, so long as the scanned area is centered on the Ti Ka emission line.

Now if it's a some weird sample absorption edge artifact, such as the famous Au Ma in pyrite hole in the continuum, then the sample composition does matter.  But my feeling is that it is a Bragg crystal artifact, so the sample composition shouldn't matter. So a Si wafer should work as well as anything. In fact maybe better since the higher average Z of Si compared to SiO2 will produce a bit more continuum intensity.

[BenjaminWade]

Ok no problems, I take your point with the absorption edge. I have started running the long scan on my spectrosil. I will let you know Friday how it went.

Cheers

[BenjaminWade]

Hi John
Here are the results from both a focussed and defocussed (50um) beam at 300s/point and 200nA on my Spectrosil glass. Again I still can't see any of the holes so suspect I may have got lucky..? As I mentioned my Spectrosil glass isn't completely pure (few ppm Ti), but I find it unlikely that it contains exactly enough concentration to "fill the holes" as it were on all five spectrometers which would in reality likely have different magnitudes of the artifact.




Thoughts?

Cheers

[Probeman]

...but I find it unlikely that it contains exactly enough concentration to "fill the holes" as it were on all five spectrometers which would in reality likely have different magnitudes of the artifact.

Hi Ben,
Yes, I agree. Besides, while a small Ti peak might just fill the small hole right under the Ti peak position, it would not fill the larger hole just to the right.

So I would say either you got very lucky with your crystal manufacturing, or maybe Cameca got better at bending their crystals over time.  Your instrument being more recent vintage than mine (and Shui-Yuan's 8900).

I would still love to see similar scan data from JEOL (and other Cameca) instruments with PET crystals.  Can anyone else post some data scanning pure TiO2 with conditions similar to what Ben used?
john

[Probeman]

Looking for anyone else who has tried using SiO2 glass (e.g., SpectoSil) as a blank standard for measuring Ti in quartz. Please share with us your quantitative experiences.

[Karsten Goemann]

Yes, I’ve always been using Spectrosil silica glass for the blank correction in quartz. According to the datasheet it has <10 ppb of each of the trace elements, except Cl (<1 ppm).

I’m running Smithsonian/Harvard Hot Springs Quartz as secondary standard, which based on LA-ICPMS done here has ~1 ppm Ti (at least in the chip that we've mounted) and I can reproduce that well using the Spectrosil blank correction. Recently I’ve also started using Shandong Quartz from Andreas Audetat (57 ppm Ti) as another secondary standard and can reproduce that well, too. I can do it the other way round (Hot Springs Qz for blank corr) with the same result.

Yes, there is some beam damage, but I’m not using quite as severe settings compared to some of the other posts.
The most severe settings are probably 15 kV, 200 nA, 5 um with 120 seconds on peak and 2x60 seconds on background. For those I can see no significant difference between doing normal peak/off-peak and alternating on/off peak, so any damage doesn't seem to significantly affect peak or background intensities.

We’re typically using 2 large PETs for Ti Ka, 2 TAPs for Al Ka and large LiF for Fe Ka. Si Ka either by EDS or just specified 100% SiO2.

That gives us 9 ppm detection limit for Ti and 4 ppm 1-sigma precision, which is good enough for us as the quartz we analyse usually has 10s to 100s ppm of Ti. We also do CL on the quartz (panchromatic or colour imaging or spectral CL) and then target zones distinct in CL in those grains with the EPMA point analyses. Where we can we put 2-3 points next to each other in the same zone to see if we can reproduce the composition and get better statistics.

[Probeman]

Yes, I’ve always been using Spectrosil silica glass for the blank correction in quartz. According to the datasheet it has <10 ppb of each of the trace elements, except Cl (<1 ppm).

...

Yes, there is some beam damage, but I’m not using quite as severe settings compared to some of the other posts.
The most severe settings are probably 15 kV, 200 nA, 5 um with 120 seconds on peak and 2x60 seconds on background. For those I can see no significant difference between doing normal peak/off-peak and alternating on/off peak, so any damage doesn't seem to significantly affect peak or background intensities.

That is really good to have this confirmed.  Thank-you!

The good thing being that this SpectroSil glass is much more readily available than synthetic SiO2 single crystal and at <10 ppb we have an excellent blank for any number of elements.

It is interesting that the glass and quartz behave so similarly for other trace elements even though the Si and O Ka intensities do change over time (with enough beam current and time), but apparently not the continuum intensities for trace elements...