Just an additional simple thought: at which sequence do you measure Na? Other than first may produce severe lossing. Measure it first and the rest of the elements.
----------------------------------------------------------------
Ioannis Baziotis [log in to unmask]
Assistant Professor of Mineralogy and Petrology
Agricultural University of Athens
Laboratory of Mineralogy and Geology
Athens, 11855 Greece
(+30)210-529-4155
Twitter: https://twitter.com/EGU_GMPV
Personal site: http://ibaziotis.wix.com/ioannis-baziotis
----------------------------------------------------------------
----- Original Message -----
Από: Allaz Julien <[log in to unmask]>
Ημερομηνία: Wednesday, January 30, 2019 8:57 am
Θέμα: Re: severe alkali-loss during microprobe analyses of low alkali andesitic glasses
Προς: [log in to unmask]
> Hi Silvio,
>
> Yes, Na (or K) in hydrous glass can be tricky. Using the EDS and a low
> current is probably the safest way to get decent analysis (we do this
> regularly on our 30 mm2 Thermo Ultradry detector at 1.5 to 2 nA and 15
> keV), but if you work on the electron microprobe here are a couple
> suggestions (in addition to the one already suggested, notably the use
> of a cryostage which is your best chance):
>
> 1) Lower the electron beam even more - 2 nA and 5 um is our normal
> routine for very sensitive and small-sized material, especially
> hydrous alkali glass, and it is acceptable in most cases, even highly
> hydrous andesitic glass.
>
> 2) Lower the counting time on Na. I had in the past analysed even more
> challenging minerals (jarosite and alunite, which are alkali-rich
> hydrosulfate), and the most accurate analyses were obtained with a 3
> second counting time on Na at 2 to 3 nA with an almost focused beam
> (we couldn’t even afford a 5 um beam size due to strong zoning in
> these 10 um sized minerals). Beyond this, the analyses were
> inaccurate… Such results are very low precision, but you have here to
> choose between accuracy and precision, and once again, if you can get
> multiple analysis point, then the average will be more acceptable in
> term of precision.
>
> 3) If you are using a Cameca instrument, and if you have access to the
> PC0 crystal (2d = 45 Å), then go for it! You will have a much higher
> count rate, and can get very good results even with just 3 seconds
> counting time.
>
> 4) Use a Time Dependent Intensity correction. I believe JEOL and
> Cameca now offers this option (on the newest instrument at least, and
> often as an additional software package), otherwise it is available on
> the software "Probe for EPMA” from John Donovan.
>
> 5) If none of the above works, lower even more the beam current. I
> would strongly advise to monitor the count rate of Na using the Chart
> Recorder and determine at which point the Na intensity is decreasing.
>
> Note that I am not certain that lowering the acceleration voltage
> would help: doing so, you will strongly reduce the analytical volume,
> and the beam damage will be increase. The damage is a function not
> just of the counting time and the current, but also the impacted
> volume (electron density). If you increase the electron density, you
> will increase the heat effect and thus the damage…
>
> Best,
>
> Julien
>
>
> P.S. You can also address your question on one of the following
> probe-specific email list which regroup many microprobe users:
> - Cameca: [log in to unmask]<
> - JEOL: [log in to unmask]<
>
>
> ###########################
> Dr. Julien Allaz
> Head assistant for SEM/EPMA
> Inst. für Geochemie und Petrologie
> ETH Zürich
> NW E 84
> Clausiusstrasse 25
> 8092 Zürich
> Switzerland
>
> Tel: +41 44 632 37 20
> Fax: +41 44 632 16 36
> Email: [log in to unmask]<
> ###########################
>
>
> On 30 Jan 2019, at 13:16, Clemens, JD, Prof [[log in to unmask]< wrote:
>
> Hi Silvio,
>
> I am guessing that the granitic glasses you were referring to may not
> have been hydrous or perhaps they had much lower H2O contents than
> your andesitic sample. Na counting losses correlate positively with Na
> abundance and with H2O content of the glass.
>
> I assume you are using EDS. You can virtually eliminate the problem if
> you:
>
> lower the beam voltage (10 kV?)
> move the nose piece of the detector closer to the sample to get more
> counts and shorten the time for analysis
> most critically, use a cooling stage to lower the sample temperature
> to that of liquid N2.
>
> I have tested an albite-composition glass with 9 wt% H2O in it under
> these conditions and easily got stoichiometrically perfect Ab analyses
> out of it.
>
> So, you may have to invest in the freezing stage if you want to
> continue using a finely focused beam. There really is very little
> choice about that.
>
> I don’t know who might make these things, but people do have them.
> Manchester and Stellenbosch both have (or had) such facilities, for example.
>
>
> Prof. John D. Clemens
> Dept of Earth Sciences
> University of Stellenbosch
> +27 (0)21 808 3159
>
> On 30 Jan 2019, at 13:44, Silvio Ferrero
> <[log in to unmask]< wrote:
>
> Dear All,
>
> We are experiencing almost complete alkali loss during microprobe
> analyses of andesitic glasses with narrow beam (diameter around 1
> micron), at 9 nA and 15 kV. The glass under investigation is hydrous
> (around 5 wt% H2O), and based on other data we were expecting 1-2 wt%
> Na2O. Our analyses show instead a remarkably constant amount of
> 0.05-0.1 wt% Na2O over 20 analyses. Analyses on granitic standards
> during exactly the same session show a very limited loss for Na
> (around 20% relative), and in both cases the Na is only measured for 6
> seconds. Few analyses were conducted with diameter of 3-4 microns, and
> still the Na was in the same range of values. Larger beams cannot be
> used because we are analyzing small melt inclusions in garnet.
>
> I´m fully aware of the alkali loss problem in alkali-rich glasses, but
> I never experienced a almost complete (around 90% relative I would
> say) loss of Na in any previous case study. Anyone has experienced
> such a problem analyzing hydrous andesitic glasses with very narrow beams?
>
> Any insights, suggestion or comment is welcome!
>
> Silvio
>
> --
> Silvio Ferrero
> tel. 0049(0)3319775705
> Universität Potsdam
> Institut für Erd- und Umweltwissenschaften
> - Dept. of Earth and Environmental Sciences -
> Haus 27, Karl-Liebknecht-Straße 24-25
> 14476 Potsdam
>
> "In those early days people laughed at me. They quoted Saussure who had
> said that it was not a proper thing to examine mountains with microscopes,
> and ridiculed my action in every way. Most luckily I took no notice of
> them"
> (Henry Clifton Sorby)
>
> To many petrologists a volatile component is exactly like a Maxwell
> daemon; it does just what one may wish it to do.
> (The evolution of the igneous rocks, N.L. Bowen, 1928)
>
> ########################################################################
>
> To unsubscribe from the GEO-METAMORPHISM list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=GEO-METAMORPHISM&A=1
>
> [
>
> The integrity and confidentiality of this email are governed by these
> terms. Disclaimer<
> Die integriteit en vertroulikheid van hierdie e-pos word deur die
> volgende bepalings bereël. Vrywaringsklousule<
>
> ________________________________
>
> To unsubscribe from the GEO-METAMORPHISM list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=GEO-METAMORPHISM&A=1
>
>
> ########################################################################
>
> To unsubscribe from the GEO-METAMORPHISM list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=GEO-METAMORPHISM&A=1
########################################################################
To unsubscribe from the GEO-METAMORPHISM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=GEO-METAMORPHISM&A=1
|