General EPMA > EPMA Standard Materials

An Open Letter to the Microanalysis Community

<< < (3/7) > >>

Probeman:

--- Quote from: jon_wade on November 22, 2021, 08:25:53 AM ---my comments re: sulfides and metal was really focused on trace elements in standards.  These are rarely homogenous, but there is a desperate need for such in the LA-ICPMS community (et al) where matrix matching is a bigger issue.  Of course, 'pure' standards' should be pure - stands to reason, which is why I favour a good metals block over some of our 'pure' synthetic stuff (MgO is a good example of a commercially available single crystal that we've found is often not as 'pure' as all that).  Sulfides are a particular problem as noted above, and I would be pleasantly surprised if you can make a significant amount that are both 'pure' and stoichiometrically identical. 

--- End quote ---

I can imagine that some synthetic materials might be inhomogeneous in trace elements especially at ICPMS sensitivity levels, but if we start with high purity materials we might mitigate much of that, especially for EPMA where anything below a few PPM is essentially undetectable.

As for LA-ICPMS standards I agree metals or oxides would work well as primary standards. Pretty much the same situation in EPMA, though some people still haven't thought this question through sufficiently and are still seeking some trace doped standard to test their trace accuracy. The best accuracy test for trace levels is a zero blank. See here for more details:

https://probesoftware.com/smf/index.php?topic=928.msg8498#msg8498

As for testing trace accuracy in LA-ICPMS, I would similarly ask why not a (roughly) matrix matched (high purity) blank as we are (or should be!) using in EPMA?

As for commercial MgO, today it is quite easy to find MgO with close to zero Ca.  This was not true in the past. Likewise, it used to be almost impossible to find Zr without a percent or so of Hf, but this material is now available in 99.999% pure form:

https://www.americanelements.com/zirconium-metal-7440-67-7

Question: I know nothing about growing synthetic sulfides, but in talking with John Hanchar he has indicated to me that it's the purity of the starting materials that matter (and cost) the most!


--- Quote from: jon_wade on November 22, 2021, 08:25:53 AM ---I  think a lot could be done to educate current probe users about the role of standard (ahem, SJIO), background selection and instrument operating conditions which would go someway to mitigating a lot of issues (honestly, theres still papers published where dead times in olivine are hit. worrying about your standard is a little moot when that happens!).  I honestly feel the cost/benefits of this mammoth effort isn't there without embracing other microanalytical communities. Perhaps it would be worth doing a thorough market survey of demand (and not just 'I'd like some!' but 'how much would you like some?').  It may also be instructive as my gut feeling is the EPMA community at a research funding level is not in such rude health. :(

--- End quote ---

I agree that for many laboratories the dead time calibrations are probably a major area of inaccuracy. Paul Carpenter has tried to point this out for decades. If anyone out there is still using the "default" dead time calibrations provided to them at the time their instrument was installed, I can promise them that they have large accuracy problems. These gas detectors age quite dramatically over time and one should be re-running these dead time calibrations every year to two at most for reasonable accuracy.  This issue has actually gotten worse over time as both Cameca and JEOL have migrated to larger and larger Bragg crystals with increasing geometric efficiency.

So unless you always run a 10 nA or less, here is a link to Paul Carpenter's dead time spreadsheet which is very nice for performing your own dead time constant calculations using any software:

https://probesoftware.com/smf/index.php?topic=1160.0

As for other microanalysis communities, we welcome all. I note that Zack Gainsforth at UC Space Sciences just wrote to Will Nachlas volunteering his TEM time, would you be willing to work with us on laser ablation standard characterization?

We need all the help we can get.

Probeman:

--- Quote from: NicholasRitchie on November 22, 2021, 12:21:57 PM ---
--- Quote ---2. In order to provide a long term and stable repository for these synthetic materials, we have been informed that the Smithsonian Institution would be pleased to provide this service for our global standards project, but they cannot have any association with commercial providers that would be making a profit from the sale of these materials.
--- End quote ---
It seems to me that for this to be a success (meaning we actually reach those people tempted to use standards as long as it isn't too hard), we need to involve commercial vendors.  This suggests that, as kind as the Smithsonian's offer is, we should probably look for someone else to handle the material who isn't averse to associating with commercial vendors.

--- End quote ---

Really?  I'm not opposed to having this global material hosted in several locations, but it's not just the Smithsonian, it's also NSF. I'm pretty sure NSF would not fund an effort that involves for profit activity.

Maybe once we've got a stable of high purity global standards that have been shared with the labs that actually care about accuracy, then we can have discussions with commercial providers.

Probeman:
While discussing WDS detector dead time calibration issues someone recently made the following point which I'd like share here:


--- Quote ---From my perspective it's exactly these sorts of instrumental calibration issues that has over time tended to corral people into finding matrix matched standards even as the matrix corrections themselves have become more accurate. As has been pointed out previously, if ones standard is exactly the same composition as ones unknown, *all* corrections are exactly 1.000!

So I think that what started out as a necessity for dealing with sub par matrix correction physics, has over the decades slowly become a crutch to avoid making sure our instruments are properly calibrated in other respects (dead time, effective takeoff, beam current linearity, etc.).

The good news is that by utilizing standards that are not exactly matrix matched, but instead accurately characterized for composition and purity, these instrumental calibration issues (and matrix correction physics) will become better understood and therefore more easily able to be improved.
--- End quote ---

sem-geologist:

--- Quote from: Probeman on November 22, 2021, 10:35:08 AM ---I know that high purity BaF2 and MgF2 are easily available. But with LaB6 is it high purity?  Please find out for us.

--- End quote ---
BaF2 and MgF2 are not so good EDS standards for the same reasons as La oxide-bearing phases (The Ba can emit Ma, especially if working with low voltages). Sincerely, I currently have no Ba standard which would allow to deal with this problem. One of candidates I would look for is Ba carbide (BaC2, https://www.americanelements.com/barium-carbide-50813-65-5), the same for Cs (Cs2C2) - I however have no idea how stable those would be. Currently to overcome these shortcomings I set DTSA-II to ignore M lines of Ba, but for any low voltage work that will bring huge impact on F and O EDS direct quantization.
As far of concerning LaB6 purity it needs to be pure as that is main material of that type of cathodes - any contaminants would cripple the stable emission. As from EPMA WDS perspective - I have wavescans and it is pure from that point of view.

Now When I say "wavescans" I mean very high current wavescans acquired using all available electron juice on our field emission SXFiveFE. Normally that is around 800nA and more (up to 1µA) which makes tops of peaks to blunt (unaccounted pile-up), but it exposes the backgrounds very clearly and any spectral artifacts or impurities at 10-50 ppm level is visible (depends from position and XTAL, (2048 channels 1 second per channel). The beam is defocused to 50um. The carbon coat is done with Leica coater with multi-pulse mode which makes a composite carbon layer which allows the coating to withstand 40+ minutes at these harsh conditions easily for most of standards. (Coating - this is again one of these underrated very crucial steps which if done wrongly will ruin analysis). Exception is minerals which breaks down (i.e. apatite) at this beam. I actually more believe my wavescans than some LA-ICP-MS where I have no control on fractionation, data reduction, laser stability. Lots of stuff there (at least what I had witnessed with my limited experience there) is quite a black box. And LA-ICP-MS can have all kind of nasty interference (mass interference). What I would be interested more for low concentration detection is µXRF. Don't understand me wrong, LA-ICP-MS is really robust method for comparative analysis (REE spider plot patterns, Isotope ratios, etc..), but I had never ever seen a reliable results for absolute values which would sum to 100 % (rather very far from it). Maybe there will be difference with new generation femto-second lasers, but previous generation with all that fractionation stuff is very unreliable as for absolute values, and I would take EPMA or µXRF (if done correctly) values without any doubt if would need to chose from values obtained by different methods. In that sense EPMA-WDS wavescan is quite a powerful and more reliable tool IMHO for trace detection down to tens of ppm especially when there is no interference hell (standards) and interpretation of such wavescan is simple.

Getting back on topic, as For EPMA-WDS REE I see no need of those REE borates, as REEPO4 scratch my all possible itches. (and there are also those REEP4O14, which I find less stable than REEPO4, and use it not at all). I don't know where we got these REEPO4 as they show absolutely no wavescan-detectable Pb (unfortunately, there is a very sad story behind that, why I don't know where it comes from). REE-borates would be handy for low voltage EDS, or/and substances where F, O, Na needs to be measured directly. 


--- Quote from: Probeman on November 22, 2021, 10:35:08 AM ---Curious: why would high purity synthetic pyrite be such a bad standard material? Seems quite stable when I've used it at 30 nA and 20 keV.  Then again I usually run my standards slightly defocused or turn on the TDI correction.

--- End quote ---
Let me put Your another quote below from your previous post:

--- Quote from: Probeman ---(our lab re-polishes and re-coats our standard mounts every one to two years)

--- End quote ---
If Your unknown pyrites was re-polished some weeks before EPMA session, then both your standard and Unknown pyrites are oxidized. You get away from it with 20kV generating X-rays from deeper of sample (for that reason alone I do all sulphides at 25kV). You will notice, however, that same calibrations does not give good result for chalkopyrite (even with corrected S position). That is as chalkopyrite oxidizes at different rate and extent. So basically, while your Unknown will be similarly oxidised as your standards you will get away with it, But try to do freshly polished Unknown samples and you should get some surprise.
 
How we do sulphides? as for sulphur reference we use ZnS which does not oxidizes, and does not require to be often re-polished (we re-polish our standards as You do - once a year). We don't use neither pyrite neither chalkopyrite standards as the only good correct analysis (intensity) possible to obtain from them is at first days after re-polishing of standard block. Samples of Unknown needs to be polished a day (for best results) before session. Differently than other samples, sulphide (for analysis) bearing samples are not placed into heater for riddance of water vapour, as oxidation of pyrite/chalkopyrite and other sulphides increase with temperature. We dry samples with stream of nitrogen gas, and then keep it longer in coater in vacuum to get away the vapour residual (like 1-2 hours). After coating it should not lay days on shelf (or worse in the heater) as coating is not enough to prevent pyrite from oxidation. The best it should be analysed the same day. We use Fe,Co, Ni, Cu and other metals from normal oxide standards like Fe2O3, CoO, NiO... or metals (i.e. Ag, Au) and sulphur only from ZnS. Analysis in this way closes around 100%, and most importantly the atomic composition makes complete sens. If we would try to repeat analysis after a week, there would be huge discrepancies.


--- Quote from: Probeman ---
--- Quote from: sem-geologist on November 22, 2021, 03:33:06 AM ---From other side we need to do analysis to keep the labs running, we can't just stop everything and say we are waiting for a full set of excellent reference materials covering (near-)whole element table. 

--- End quote ---
And no one is saying we have to stop all current lab work immediately until this project is complete. Where did you see that mentioned in the open letter?   >:(

--- End quote ---
These are inner personal regrets sometimes. Especially when looking to some mine, one of the first analysis which were published. At least I find myself pretty often in situation when I am asked for results immediately! Even if there are some clear analytical biases or artifacts (which were not obvious previously), which would imply to redo some analysis at changed settings. Not everyone would say "Ok, take time and investigate and make sure these next analyses are correct, or redo those and make it more correct", rather "We already published this paper with these standards and these settings, we don't want to describe new methodology, we just will cite that old one, we don't ***** care... it should be completely the same. Period!", because our system is publish-or-perish, and incremental improvement of method does not fit well with that system.

So that is why I wish to have synth's covering all elements already yesterday. Of course I understand that it will take time.

--- Quote from: Probeman ---I am reminded of occasions when I have been contacted by some EPMAers attempting to utilize the MAN background correction (Donovan et al, 2017), and been told that they had no idea that their standards were so contaminated with minor and trace elements...   :(

--- End quote ---
That is why I don't use MAN, but single background and precise and universal slope - This way I can get down to 10 ppm, but without: any hassle of LA-ICP-MS of standards, making background correction curves, etc. Simple, elegant, reliable, works same independently from matrix.
 

--- Quote from: Probeman ---So the plan is that these materials will be provided for free to qualifying laboratories. A good first step in that qualification process is to join the FIGMAS:

https://figmas.org/about.php

--- End quote ---
Do I understand correct, I should join at first MicroAnalysis Society (MSA, as European joining MAS would make little sense), then I  should join FIGMAS?

As dead time was picked few times... I probably should get back and finish my MC simulation of pile-ups and come up at last with non-linear equation which actually works for whole range of possible count rates (0-1Mcps), and different generation of counting electronics.

Probeman:
If anyone is interested in our current Google spreadsheet of commercially available synthetic standard material candidates, as shown here:



Please use this link:

https://docs.google.com/spreadsheets/d/19AeXvxIaP6qvChbE7cxK05B_6rkZSN14T-7ZJ7nkm8M/edit*gid=0

If you would like to add to this spreadsheet any additional commercial (or academic or institutional) sources of potential synthetic standard candidates, please contact Nicholas Ritchie and he can add your email to the approved  "edit list".  Please read the instructions carefully so this spreadsheet remains well  organized.

Ultimately we are looking for high purity synthetic materials in 500 to 1000 gram quantities, enough for true global standards with extra material for future generations. This material does *not* need to be crystallographically oriented or polished. Also these do not need to be single crystals, they can be broken in pieces as long as the individual crystals are at least millimeter(s) in size.

We're looking for: Mg2SiO4, YAG, RbTiOPO4, KTiPO4, MnO, Fe3O4, NiO, ZnO, LaAlO3, MnPSe3, LiTaO3, ZrSiO4 (zircon), ZrO2 (zirconia), HfSiO4 (hafnon), HfO2 (hafnia), ThSiO4 (tetragonal thorite), ThSiO4 (monoclinic huttonite), Fe2SiO4 (fayalite), Mn2SiO4 (tephroite), CaMgSi2O6 (diopside), Al2SiO5 (sillimanite), NaAlSiO4 (nepheline), KAlSi3O8 (sanidine), KAlSi2O6 (leucite), KAlSi3O8 (orthoclase), NaAlSi3O8 (albite), CaAl2Si2O8 (anorthite), Fe3Al2Si3O12 (almandine), PbSiO3 (alamosite), CaAl2O4 (krotite), CaAl4O7 (grossite), CaAl12O19 (hibonite), CaSiO3 (wollastonite), MgSiO3 (enstatite), FeSiO3 (ferrosilite), sulphides and sulfosalts, etc., etc.

We would like pricing on 500 to 1000 gram amounts and also, for some initial testing, pricing on amounts in the range of 3 to 5 grams.

We need your help in tracking down suitable materials! 

Navigation

[0] Message Index

[#] Next page

[*] Previous page