Author Topic: Chem Age / Chemical age mapping of monazite  (Read 1231 times)

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Chem Age / Chemical age mapping of monazite
« on: February 19, 2018, 09:53:46 pm »
Hi all
I am hoping the experts can weigh in on a question I have regarding chemical age mapping of monazite. More specifically, in the examples I have seen it appears that only Th/U/Pb/Y have been mapped, however wouldn't we have to map for the major elements (ie P/LREE/MREE) as well for matrix corrections?

I have done some tests from some old monazite spot data I have, and although it doesn't change the chemical age much it does change the absolute amounts.

Madel Monazite (~513 Ma) with all P/REE elements measured:
Th ~9.78 wt%
U ~0.32 wt%
Pb ~0.24 wt%
Chemical age average ~505 Ma

Madel Monazite (~513 Ma) WITHOUT all P/REE elements measured (ie Quant disabled):
Th ~8.85 wt%
U ~0.35 wt%
Pb ~0.22 wt%
Chemical age average ~507 Ma

Its the same for a number of other standards that I had run, doesn't change the age much (well within error) but quite drastically changes the absolute amounts.

So my questions is for anyone that has done chemical age mapping, do you use specified concentrations and put in some kind of average monazite composition, or composition from measured spot analysis?

Any input would be appreciated.

Cheers

Mike Jercinovic

  • Professor
  • ****
  • Posts: 58
    • UMass Geosciences Microprobe-SEM Facility
Re: Chem Age / Chemical age mapping of monazite
« Reply #1 on: February 20, 2018, 09:35:36 am »
Right, for matrix corrections to  work, you at least need all the major elements. You also need something to relate background to the Th/LREE concentration and subtractions for the Pb counts for interference with Ylg and ThMz, and adjusting UMb counts for ThMg.  We have not done age mapping in some time, but basically we used empirical Th/REE background (you can collect background maps too, although I think this just adds error to such minor concentrations for Pb and U) and interference corrections on the counts based on spot analyses, then applied Z A F factors also based on results from full major element analysis.  All kind of crude, but where age domain differences are large enough, can make for a nice map.  I would imagine with ProbeImage one could get the EDS spectrum to roughly quantify P and REEs to do this much more accurately.  That way you could actually do real interference corrections and the MAN background might do well enough.  Overall, you are better off using a single bkg value for a single compositional domain (where Th/LREE and Y do not vary much) rather than pixel by pixel values with large count uncertainty, so maybe a good way would be to assign some sort of count range of Th and Y to a single background value, and another range to another bkg value, and so on.

Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #2 on: February 20, 2018, 10:01:05 am »
Hi Ben,
I suspect I haven't looked at this Chem Age stuff in as much detail as Mike and Julien, but on the question of matrix corrections for monazite chem age dating, I believe I found that whether I acquired and analyzed all the REE elements and other matrix elements for a good total, or if I only acquired and analyzed U, Th, Pb (and Y for the interference on Pb) and simply specified Ce2O3 by difference (or was it CePO4 by difference?), I got almost exactly the same chemical ages.

So it appears that the matrix correction itself isn't a huge deal, but the background corrections are the biggest issue for accuracy. This is why Mike, Julien, Karsten and I went to so much trouble to implement the multi-point background method.  Slight changes in curvature of the background fit can significantly affect trace element accuracy. Julien Allaz has a post here that discusses some of these issues:

http://probesoftware.com/smf/index.php?topic=186.msg802#msg802

If you haven't read it, also see Mike's paper here on trace element analysis:

http://www.probesoftware.com/download/EMAS_proceedings_Trace_Analysis_MJJ_v4.pdf

So bottom line: a matrix correction is good to do (especially for absolute accuracy of the concentrations), but even just CeO3 by difference in the matrix correction is good enough for U, Th, Pb. But the interference correction and background corrections are the most important thing.
« Last Edit: February 20, 2018, 10:23:49 am by Probeman »
The only stupid question is the one not asked!

Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #3 on: February 20, 2018, 10:22:24 am »
Hi Ben,
I went back to a monazite mapping run I did some time ago just to see what I did and here is the quant setup:



Note that I added La as an analyzed elements, but I suspect just because I had an extra spectrometer free.  I also specified CePO4 by difference, but because this run was from before I had implemented the "formula by difference" method, I used this slightly complicated method in the Calculation Options dialog:



I hope this helps.
The only stupid question is the one not asked!

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #4 on: February 20, 2018, 04:21:14 pm »
Hi Mike and John
Many thanks for the replies, this info is exactly what I needed. It is ~1.7Ga monazite with domains that have been partially to wholly reset down to ~0.3Ga. I have U-Pb spot data on it, and can/have age mapped them via laser as well, but for the smaller ones would be good to do the equivalent via EPMA.

I think I will try a variety (background maps, MAN fit, single bkdg value from compositional domains) and see how it changes things accuracy and precision wise. I am quietly/desperately hopeful that the MAN might be good enough for the monazite in question, but we will see. I was thinking of running a restricted Z man fit ranging from ~35-45 using the REEPO4 standards. Not sure if that would be any better accuracy wise than the normal large Z range MAN fit. Will see.

John perhaps you also mapped for La for the La 2nd order intf on Pb Ma? What I really want to do is sum Pb spectrometers but this would preclude mapping for La and wouldn't be able to use in that for intf corrections.

Anyway I will try out some things and see how I go. Thanks again.

Cheers

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #5 on: February 22, 2018, 06:28:32 pm »
Hi John
Slightly embarassing...I can't find an option/checkbox anywhere to get it to calculate a grd file of chemical ages. I am looking in Calcimage -> Specify Quantitative Parameters -> Calculation Options and cant see anything? In normal PfE its under the calculation options window but is it different in Calcimage?

Cheers

EDIT: Ignore that...found it...
« Last Edit: February 22, 2018, 06:30:29 pm by BenjaminWade »

Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #6 on: February 22, 2018, 06:34:54 pm »
Hi John
Slightly embarassing...I can't find an option/checkbox anywhere to get it to calculate a grd file of chemical ages. I am looking in Calcimage -> Specify Quantitative Parameters -> Calculation Options and cant see anything? In normal PfE its under the calculation options window but is it different in Calcimage?

Cheers

EDIT: Ignore that...found it...

There definitely are a lot of menus in CalcImage,  but that's only because there are so many data types one can output!   :P 
The only stupid question is the one not asked!

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #7 on: February 22, 2018, 09:40:47 pm »
For those interested, here are some very rough preliminary results comparing traditional OFFpeak/MP background data with MAN fit for chemical age dating.
Here are the MAN fits for U/Pb/Th/Y/La using Jared's RPI phosphate standards:







Below is an excel spreadsheet of some example spot data processed using traditional offpeak and with MAN. Pretty close really for a first pass.
I am in the process of comparing the ChemAge maps using off peak acquired maps vs MAN fit maps, and as one might expect PPM maps are much less noisy using the MAN method, and seems to give a less "noisier" map.

Cheers
« Last Edit: February 22, 2018, 09:44:36 pm by BenjaminWade »

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #8 on: February 22, 2018, 10:10:16 pm »
And here is the comparison of the Calcimage maps using OFFPeak background maps, and those using MAN. Caveat is I had limited time to conduct this test...so my on peak count time was 1s, and my background maps were acquired at 0.5s each. Even so I think it illustrates what we see from the OFFPEAK vs MAN spot data.

Cheers

OFFPEAK




MAN



Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #9 on: February 23, 2018, 02:23:27 pm »
Hi Ben,
Very impressive. Especially for only 1 sec dwell time per pixel.  What was the mapping beam current?

I used 2.2 sec per pixel (and 0.6 sec per off-peak) and 100 nA for my monazite test here (I should probably merge these topics...):

http://probesoftware.com/smf/index.php?topic=100.msg366#msg366

But I didn't try using MAN!   In my MAN paper I looked at zircon (and quartz) comparing off-peak to MAN, but I never tried MAN with monazite simply because I'm not sure how to check accuracy using a "blank" standard.  What is your evaluation of the accuracy with MAN for monazites?

Also I didn't try MAN with monazite because it's such a high average atomic number at ~41, while zircon is only ~24.  On the other hand I know Karsten has been looking at REE rich minerals with very high average Zs using the MAN background method and he has some interesting data regarding that. Particularly the off-peak interference issues for REEs, because with the MAN method there are no off-peak measurements to interfere with!

Finally I'm glad you showed the background intensity maps for off-peak vs. MAN.  It's quite striking how much better the MAN method is compared to off-peak.  Of course that is most because the off-peak background statistics are limited by the continuum intensities, while with the MAN method statistics the background intensity is based on the concentrations of the major elements, which have much better statistics of course.  See fig 14  in my MAN paper here:

http://epmalab.uoregon.edu/publ/A%20new%20EPMA%20method%20for%20fast%20trace%20element%20analysis%20in%20simple%20matrices.pdf

The variation in the Hf really shows up in the MAN bgd intensities.

Of course the best thing about MAN is that is takes half the acquisition time compared to off-peak acquisition (and with better precision)!   :)
« Last Edit: February 23, 2018, 02:25:37 pm by Probeman »
The only stupid question is the one not asked!

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #10 on: February 24, 2018, 03:17:10 am »
Hi John
Yes I was a bit rushed for time so I only did 1sOn/0.5sOff...but cranked it to 200nA. Next time to do the Offpeak maps justice I will count for longer on the background maps (and peak) and see how it compares with the MAN.

Yes I have read your MAN paper many times. You are correct in that at the moment my comparison is a bit circular. My spot measurement OFFPEAK data was measuring La and using Ce for difference, which of course wouldnt be reality. And I am comparing my MAN data to it which is assuming the same Ce by difference, so the potential inaccuracy of the MAN wt% numbers due to z estimation is a bit negated by this. This is probably a topic for the standard material thread but given that single element phosphate xtals can be hydrothermally grown, is there anything stopping a mixed REE phosphate being grown Th/Pb/U free with an "average" monazite composition to use as a blank standard?

A test I guess would be to measure one of my monazite standards with a full element package (of which I have U/Th/Pb/REE quantified via other analytical techniques), then analyse the same monazite standard using MAN and measuring just U/Th/Pb/La and doing Ce by difference and see how close the numbers are then. If I was a smart man I would have done this first, but I'm not. I just jumped into and unknown as it was a bit of a play around with it.

Also the other thing worth mentioning is that the unknown monazite I have used is by no means a "low Pb" monazite, so would be interesting to see how well the MAN works for monazite with lower Pb. When I have some more free time on the instrument I will repeat the experiment on some monazite standards

Cheers


Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #11 on: February 25, 2018, 12:02:14 pm »
A test I guess would be to measure one of my monazite standards with a full element package (of which I have U/Th/Pb/REE quantified via other analytical techniques), then analyse the same monazite standard using MAN and measuring just U/Th/Pb/La and doing Ce by difference and see how close the numbers are then. If I was a smart man I would have done this first, but I'm not. I just jumped into and unknown as it was a bit of a play around with it.

Also the other thing worth mentioning is that the unknown monazite I have used is by no means a "low Pb" monazite, so would be interesting to see how well the MAN works for monazite with lower Pb. When I have some more free time on the instrument I will repeat the experiment on some monazite standards

Hi Ben,
I agree one should first test their analytical setup on known monazite age standards. 

But I'm a little skeptical that MAN will be accurate enough for monazite dating (though I would be pleased to be proven wrong!), unless one has a matrix matched blank, but I just don't think that will be easy to produce.  I once asked John Hanchar about it and he thought originally it might be possible but he later told me that due to zoning during crystal growth it didn't work out very well.

That said, there are monazite age standards, such as the Moacyr material from Brazil which has an age around 500 mya.  I think Mike Jercinovic might be able to provide you with some material.

Back to the MAN correction and the effect of the average Z for the MAN background for a measured matrix versus an assumed matrix of say CePO4, I note that the Moacyr monazite average Z is ~41.6, while CePO4 average Z is 39.18.  How much of a difference that makes in the MAN intensity calculation for different elements will depend on the slope of the MAN curve, and it would be very interesting to hear from you exactly what you find.
john
The only stupid question is the one not asked!

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #12 on: February 25, 2018, 08:58:55 pm »
Hi John
Yes it needs to be redone with some of my monazite standards. I have a number of them so all good on that front I think for the time being. I have probed/lasered/some SHRIMP'd/some TIMS'd them for majors/traces/age so have a set I can work with.

Looking at my standard database, my monazite standards range in average Z from 38.5-42.8, so a reasonable range there mostly governed by LREE/MREE mixtures. Using the equations for the polynomials from the current MAN fits, the increase in calculated background (cps/nA) going from average Z 38  to 43 is:

Pb - 0.506 to 0.535 = ~170ppm difference
Th - 0.338 to 0.371 = ~500ppm difference
U - 1.964 to 2.140 = ~630ppm difference
Y - 0.298 to 0.301 = ~20ppm difference
La - 8.172 to 8.719 = ~520ppm difference

What those differences equate to in age hard to say I guess as they are interrelated in the age equation, but I guess its probably not insubstantial. Regardless I might still have a go on the monazite standards and see what happens.

Cheers

Probeman

  • Emeritus
  • *****
  • Posts: 1690
  • Never sleeps...
    • John Donovan
Re: Chem Age / Chemical age mapping of monazite
« Reply #13 on: February 25, 2018, 09:36:00 pm »
Looking at my standard database, my monazite standards range in average Z from 38.5-42.8, so a reasonable range there mostly governed by LREE/MREE mixtures. Using the equations for the polynomials from the current MAN fits, the increase in calculated background (cps/nA) going from average Z 38  to 43 is:

Pb - 0.506 to 0.535 = ~170ppm difference
Th - 0.338 to 0.371 = ~500ppm difference
U - 1.964 to 2.140 = ~630ppm difference
Y - 0.298 to 0.301 = ~20ppm difference
La - 8.172 to 8.719 = ~520ppm difference

You should try the multi-point bgd next.  The advantage being that one can see if some off-peak backgrounds are inconsistent with the others.  In fact, why don't you take the off-peak data you already have and turn on the "shared bgd" feature?   Then show us some MPB plots.

http://probesoftware.com/smf/index.php?topic=9.msg6709#msg6709
« Last Edit: February 25, 2018, 10:57:50 pm by Probeman »
The only stupid question is the one not asked!

BenjaminWade

  • Professor
  • ****
  • Posts: 148
Re: Chem Age / Chemical age mapping of monazite
« Reply #14 on: February 26, 2018, 12:51:02 am »
Hi John
The difference in calculated background cps/nA I posted were just to illustrate the minimum/maximum calculated background values from my current MAN fit going from Z 38 to 43, which I have taken as a "typical" range in monazite composition.

I usually do MPB for U and Pb, and carefully selected two point offpeak for Th/+all REE in my current monazite package. This lets me turn on the shared for Th and all the REE, and pick and choose. For this ChemAge package though I just used my most consistent two point OFFPEAK backgrounds ascertained from past monazite work, as I knew I was going to do offpeak background maps and could just read them in in ProbeImage. I wasn't planning on doing any point analyses but I did in the end. The shared backgrounds method is defintely an awesome feature and I have used it a lot for weird and wonderful REE minerals in the past.

Unfortunately the shared wont work in this instance as I only measured one element per spectrometer! I will do a proper calibration measuring everything with MBP/Shared then compare it to MAN fit data. When the instrument is free next...

Cheers