I have leucites and some pyroxenes immersed in a foiditic residual glass (red in the image). I melted everything in an oven, so in a closed system, and I got a tephri-phonolitic bulk (in blue). Both analyzes were performed with a microprobe. In green is the whole rock analysis (powder) with the XRF. Can someone explain this behavior?
Dear Sivia,
At first sight it looks to me that some alkalis loss occurred while making the XRF pills at high temperature.
yes. I know... I was hoping that with the xrf (pressing the powder into a pill, not melting) I had a minor loss of alkali rather than with the melt at high temperature on the oven, but it would seem the opposite. Any idea? and especially .. the decrease of silica?
Did you notice any major element amout that is higher than normal in your XRF results? This may explain the lower content of SiO2 (and alkalis too).
yes it is possible because your CPX are SiO2 rich and are SiO2-saturated, what is not the case of the parent liquid (tephritic one). The residual glass can be more and more undersaturated during the crystallization of cpx, givng way to a strongly undersaturated residal glass.
Bernard Platevoet
Univ. Paris-Sud
Thanks everyone for your answers.
Mattia, I don't notice any major element amout that is higher than normal, but rather I think the Na is lower (alkali loss).
Bernard, I think you hit the point. But this is possible also if I have about 30% of lecites and only the 2-3% of Cpx?
seeing your TAS diagram II aggree with mattia : probably loss of alkalies
Loss of alkalis either for the bulk by microprobe that for the total rock by xrf?
Dear Silvia,
The alkalis loss can occur with XRF while you are melting your pills at high temperature (alkalis are rather volatile at high temperature). It depends on the temperature you use for pill preparation and how the material within the crucibles is stirred (also mechanical stirring/mixing can produce of migration of alkalis, sometimes).
The alkalis loss can also occur during microprobe analysis. It depends on the analysis time, beam diameter, beam current and acceleration voltage you have used with the probe. To limit or avoid the alkalis loss, I would suggest to use the following analysis conditions with the probe when dealing with silicate glasses:
- analysis time = 20 s on the peak and 10 s on the background
- beam diameter = 10-20 microns
- beam current = 2-5 nA
- acceleration voltage = 15 kV
I have used such conditions by using a JEOL-JXA-8200 at ETH in Zürich with synthetic haplogranitic glasses and phonolites and tephriphonolites from Vesuvius. I did not encounter any problem.
I hope this helps you.
Cheers,
Mattia
Hi Silvia
XRF analysis of bulk powders can be fraught with issues including particle size effects particularly on the low Z elements such as Na. The 'matrix corrections' may not fully account for the observed values. Most labs use a fusion with various fluxes eg Li metaborate or mixtures of Li salts with possible addition of a heavy absorber (La based to reduce absorbtion from heavier elements eg Fe).
Nevertheless I believe that Mattia is on the right track. If you are using a finely focussed electron beam( of say 1-2 microns) on the glass then light element loss might occur. HIs EPMA settings for glass would suggest a defocussed beam and this might aid in determination of alkalies . I seem to remember a long time ago that one of the GURUS of EPMA at the time suggested that the volatiles (eg. Na ) migrate away from the beam.
Check your total analyses to see if alkali loss is compensated for by Si gain-if so it suggests that you have a volatile loss situation
Hi Silvia
(nice to talk about leucite problems again)
as pointed by all people in the discussion: melting processes and both xrf and mpa analyses are characterized by loss of alkali (above all Na in respect to K).
Highlighted that, you have to solve the problem and use datas!
Transition from red diamonds (foiditic residual glass) to blue ones (tephri-phonolitic bulk melts) is "normal", in the meaning that when you melt the foiditic glass you melt also leucite and pyroxenes, thus producing a relative increase of SiO2, Al2O3 and CaO in respect to a relative loss of alkali on the total major elements set.
Have you tried to produce a Mass-Balance in which the parent is the foiditic glass, the daughter is the tephriphonolitic bulk and the phases involved are leucite, clinopyroxene and Fe-spinel? (leucite is famous for its important amount of spinel inclusions).
If this mass-balance matches, you can evaluate the XRF analysis (the green one) as unusefull.
Ciao
F
Thanks Fede!
il problema è proprio con i bilanci di massa, non tornano..ma probabilmente meto troppo poco ossido di Fe, l'avrò sottostimato! Cheers!!!
Il problema potrebbe insidiarsi nella ripartizione tra ferro bivalente e trivalente. Cosa che influisce tantissimo sui bilanci di massa.
Inoltre dovreti far finta che nella matrice il P2O5 è sempre 0% perchè non conosci l'eventuale presenza di fosfati vari (le rocce vulcaniche della Provincia Romana sono famose per la grande varietà di silico-fosfati).
Inoltre dovresti provare un bilancio di massa a 4 o 5 fasi:
leucite con minimo sodio
leucite massimo sodio
cpx massimo sodio
cpx minimo sodio
spinello ferrifero (tipo magnetite- che è tipico delle inclusioni della leucite)
secondo me è solo questione di azzeccare il mood giusto.
Ciao!
F
P.S. sorry for not using english!
- Badges
- Science topic