Standards Which Should Be Developed For EPMA

[John Donovan]

This topic is for the development of new standard materials.

At the recent Cameca user's meeting (which included a few JEOL users as well!), we discussed some ideas regarding standards. Anette Von der Handt suggested that we start to organize the community to create a priority list of new standard materials that should be developed. Suggestions for new materials should be made to this topic.

The idea would be to organize an effort to prioritize the development of a kilogram or so of material sufficient for distribution to every probe lab globally.

To start things off I can suggest two standard materials which I deem important and suitable for development. Specifically a Rb and a Cs standard.  In fact I had worked with a chemistry grad student a number of years ago at Berkeley to develop an excellent Rb standard, but we only produced a gram or so of the material which is RbTiOPO4.  But what a material!

The crystals were beautiful water clear and double terminated. RbTiOPO4 is also insoluble in water and completely robust under electron bombardment. It is also perfectly homogeneous and apparently homogeneous in Ti and P. Based on my own use of this material I think this could be the perfect standard for Rb.

Unfortunately there is no Cs analog, so more research and discussion is necessary.

[BenH]

Have you considered glass standards for these elements?

[Probeman]

Have you considered glass standards for these elements?

Hi Ben,
Yes, you be the "glassman", since that is your "bread and butter" as they say...

My only issue with glass standards (and don't get me wrong- I love some, e.g., the NIST mineral glasses), but although we can have purity constraints on synthetic glasses, we cannot get a handle on stochiometry.  Since I mentioned the NIST glass, I will point out that the NIST analyses are in error to the extent that some of the Fe is Fe2O3, not FeO. See here:

St  160 NBS K-412 mineral glass
TakeOff = 40.0  KiloVolt = 15.0  Density =  2.600

SRM 470, NIST
C.M. Taylor (Photometry?) FeO 2.77, Fe2O3 8.15
Total as FeO 10.10, Excess O 0.815
Na = 430 PPM (EPMA by JJD)
Oxide and Elemental Composition

Average Total Oxygen:       43.597     Average Total Weight%:  100.120
Average Calculated Oxygen:  42.797     Average Atomic Number:   12.694
Average Excess Oxygen:        .800     Average Atomic Weight:   21.981

ELEM:     SiO2     FeO     MgO     CaO   Al2O3     MnO       O    Na2O
XRAY:      ka      ka      ka      ka      ka      ka      ka      ka
OXWT:   45.352   9.960  19.331  15.250   9.270    .099    .800    .058
ELWT:   21.199   7.742  11.657  10.899   4.906    .077  43.597    .043
KFAC:    .1621   .0654   .0776   .1008   .0334   .0006   .1738   .0002
ZCOR:   1.3079  1.1840  1.5026  1.0818  1.4678  1.2046  2.5078  1.9914
AT% :   16.571   3.044  10.530   5.970   3.992    .031  59.822    .041
24 O:    6.648   1.221   4.224   2.395   1.601    .012  24.000    .016

St  162 NBS K-411 mineral glass
TakeOff = 40.0  KiloVolt = 15.0  Density =  2.600

SRM 470, NIST
C.M. Taylor (Photometry?) FeO 4.39, Fe2O3 11.23
Total as FeO 14.49, Excess O 1.12
Oxide and Elemental Composition

Average Total Oxygen:       43.558     Average Total Weight%:  100.183
Average Calculated Oxygen:  42.438     Average Atomic Number:   13.227
Average Excess Oxygen:       1.120     Average Atomic Weight:   22.412

ELEM:     SiO2     FeO     MgO     CaO   Al2O3     MnO       O
XRAY:      ka      ka      ka      ka      ka      ka      ka
OXWT:   54.301  14.420  14.671  15.471    .100    .099   1.120
ELWT:   25.382  11.209   8.847  11.057    .053    .077  43.558
KFAC:    .2018   .0950   .0568   .1027   .0004   .0006   .1735
ZCOR:   1.2578  1.1793  1.5587  1.0770  1.4589  1.2001  2.5106
AT% :   20.217   4.490   8.143   6.172    .044    .031  60.903
24 O:    7.967   1.769   3.209   2.432    .017    .012  24.000


So, once we make the glass, then our problems really begin.  What did we make?  This is not an inexpensive question to answer.

With a synthetic single crystal of say Al2SiO5 or RbTiOPO4 there is no doubt what we made accuracy wise based on purity and an x-ray diffraction pattern.

[Jeremy Wykes]

Are the conditions and method of synthesis of the K-series CMASFe glasses available?

For Fe-bearing glasses the speciation of Fe will always be a problem. Accessing oxygen fugacities where Fe is 100% Fe²⁺ or 100% Fe³⁺ is difficult to intractable for large scale synthesis (for compositions that are somewhat like natural magmas anyway). For 100% Fe³⁺ oxygen pressures greater than 1 bar are necessary, and 100% Fe²⁺ cannot be produced by equilibration with pure CO gas (in addition to the problem of graphite precipitation).

At ANU we are in the process of small scale synthesis of some S-bearing CMAS glasses with controlled S-speciation (100% S^2- and 100% S⁶⁺ compositions) for redistribution as a standard. I will have more details when the synthesis is complete and I am ready to send them out.

[BenH]

Hi John.

I get your point about the advantages to stoichiometry.  In the case of elements in common rock forming materials, I think minerals are very preferable to glasses.  You pointed out the need for a Rb standard and proposed a reasonable mineral. If you are measuring something very different in composition, you rely heavily on ZAF.  In many cases this works quite nicely.  In many cases the matrix corrections result in small inaccuracy.  There is no perfect standard.  The trade off thus becomes using a standard that is more like the unknown versus a confidently constrained crystalline standard that is very different than your unknown.  Which results in more uncertainty?  That is an issue I commonly face measuring the materials I face.  Any input to this dilemma would be appreciated.

Thanks John.

[John Donovan]

I get your point about the advantages to stoichiometry.  In the case of elements in common rock forming materials, I think minerals are very preferable to glasses.  You pointed out the need for a Rb standard and proposed a reasonable mineral. If you are measuring something very different in composition, you rely heavily on ZAF.  In many cases this works quite nicely.  In many cases the matrix corrections result in small inaccuracy.  There is no perfect standard.  The trade off thus becomes using a standard that is more like the unknown versus a confidently constrained crystalline standard that is very different than your unknown.  Which results in more uncertainty?  That is an issue I commonly face measuring the materials I face.  Any input to this dilemma would be appreciated.

Hi Ben,
Yes, that is exactly the dilemma.

That is: is the accuracy of the std composition or the accuracy of the physics the dominating factor?  And I would agree with you that the answer sometimes is "it depends". Though for trace and minor elements, I would argue that precision is more often the culprit than std accuracy or physics...  but that's another story.

But if we assume that we can obtain stoichiometric oxides, silicates and other minerals, then these materials are more than adequate for many compositions we generally face, especially when armed with the current physics.  Which the accuracy of, I will contend, has improved immensely in the last decade based on my understanding of the literature. 

Now, there are still going to be chemical state issues for some materials, especially for low energy lines, and to be sure, there are also a few "black holes" in the periodic table (e.g., Si ka in hafnium) where an emission line is just above the critical excitation energy of an absorber edge in addition to large atomic number effects, but in general I think it makes sense to try and eliminate inaccuracy in many of our standards and more importantly, make these materials available in large quantities to all labs world wide, so we are all at least "on the same page" so to speak.

[Probeman]

I am discussing how we might synthesize some of these standard materials with Marc Schrier who now works/operates

http://www.calchemist.com/

growing crystals of all sorts.

He was the original student at UC Berkeley that synthesized the RbTiOPO4 material discussed above (that some of you have) for my UCB EPMA lab years ago when he was a chemistry student there and I was learning what an EPMA was.

We decided to maybe start with making a much larger amount of RbTiOPO4 and so he started strategizing and asked first if there was any trace Pt in the material since he used a Pt crucible...  other crucible materials might be possible also (and cheaper).

I analyzed the RbTiOPO4 material last night for Pt and at 15 keV I got:

0.029 wt% +/- 0.030

at 20 keV I got:

0.007 wt% +/- 0.013

So below detection limit.  This was using the Pt La line at 100 nA.

Turns out the Pt Ma line is interfered with by the Rb LG3 line!

I will keep you all informed as to progress.  In the meantime I would like to ask, if we try to "crowd source" this material so we can create enough material for every EPMA (and SEM?) lab in the world (~1 Kg?), how much money would each lab be able to "cough up" to accomplish this?

I have attached a poll to this topic... see the top of this page

On an slightly related topic please check this post also:

http://probesoftware.com/smf/index.php?topic=449.msg2479#msg2479

[AndrewLocock]

The orthorhombic form of CsTiOAsO4 might be a possible material to produce as a Cs standard.
According to Loiacono et al. (1993. Journal of Crystal Growth 131, 323-330) it can be grown by methods similar to KTiOPO4; they report 30 mm as the largest dimension.
In a couple of papers, Cheng et al. (1993. Journal of Crystal Growth 132, 280-288 and 289-296) suggest self-fluxing growth is best, and crystals up to 30-32 mm.

[Probeman]

The orthorhombic form of CsTiOAsO4 might be a possible material to produce as a Cs standard.
According to Loiacono et al. (1993. Journal of Crystal Growth 131, 323-330) it can be grown by methods similar to KTiOPO4; they report 30 mm as the largest dimension.
In a couple of papers, Cheng et al. (1993. Journal of Crystal Growth 132, 280-288 and 289-296) suggest self-fluxing growth is best, and crystals up to 30-32 mm.

Hi Andrew,
Apparently great minds think alike, because Marc Schrier wrote to me by email last week and also suggested the arsenate!  Here are some of his comments:

CsTiOAsO4, 39.6% Cs by weight
CsZrOPO4, 39.7% Cs by weight
for comparison sake, the RbTiOPO4 is 29.7% Rb by weight.

I assume you prefer a high weight %, but these are nearly the same.  Do these sound of interest?  Do you have any preference between the two?  Apparently they are both water insoluble.  The phosphate is probably an easier synthesis, but I suspect the two will be similar to the RbTiOPO4 preparation.

We'd probably prefer the phosphate just because arsenic is no fun to play with!
john

[Probeman]

Talking with Marc Schrier just now and he might have found a source for quantities of RbTiOPO4 for us, but I'm checking into pricing...

In the meantime I think we should start making a "wishlist" of ideal materials for various elements.  My thinking is that for some elements identifying materials that are relatively beam stable and *not* water soluble can be daunting.  But there are some "bad boys" of the periodic table in my experience:

The halogens: F, Cl, Br and I

The alkalies: especially Rb, Cs but also an oxygen free Na standard would be ideal to be utilized as an interference for Na on oxygen.

Also semi-metals such as In, Sn, P etc.

What standards are you interested in obtaining?

[Marc Schrier]

Hi Guys,
     Just before making contact with a RbTiOPO4 (RTP) optics supplier (that John's going to chat with further), I started a ~1 week long flux prep with the ~17 year old flux I used to grow the RTP crystals John has been working with.  I got a great yield (486.4 mg) (at this scale), but the crystals are clear and yellow/gold instead of clear and colorless (see attached picture if I got it to work).  I have a feeling the flux is now depleted in P2O5, and the excess Rb2O is dissolving the platinum crucible (there was a 3.3 mg loss).  There is also a chance the crucible had some trace contamination from the last thing I ran in it or that this TiO2 is not pure enough.  For now I think we should just wait and see how John's discussion goes, but if it does not go well (prohibitively expensive, etc.), here's what John and I were planning:  Unfortunately the current synthesis is in a ~10 cc platinum crucible, and scaled up a little would only give about 2 g RTP per week.  Unfortunately a giant platinum crucible would be prohibitively expensive, so that's a lot of runs to prepare John's goal of 1 kg!  -Maybe the community can convince John to bringing that amount down.  Our plan was to prepare some RTP from the platinum crucible as a baseline, and then try some alternative (and cheaper) crucibles like alumina, zirconia, silver, and quartz.  And then maybe even nickel and graphite in an inert atmosphere if needed.  If any work (and John does not see any contamination in the microprobe), then they could be scaled up for a much larger batch.

For now I'll dabble with growing a cesium standard, but if anyone has a material they need prepared and wants to chat about it, drop me a line.  I've got gear to run molten salt fluxes and hydrothermal syntheses that should be adequate for most crystals.

Edit by John: The guy at Coherent is out for a week and he's apparently the one I need to talk to. Yes, please send me a crystal to check for Pt.  FYI: 1 kg is an upper limit!

[Probeman]

Just spoke with a guy at Coherent and he thinks they have scraps of pure single crystal RbTiOPO4 that they can make available for us.  The only possible contaminant is K which is intrinsic to the RbCO3 starting material.

Will let you all know how this pans out...

[John Donovan]

Just spoke with a guy at Coherent and he thinks they have scraps of pure single crystal RbTiOPO4 that they can make available for us.  The only possible contaminant is K which is intrinsic to the RbCO3 starting material.

Will let you all know how this pans out...

It's in the mail.

As soon as I check it for purity and run an XRD I'll make it available for a nominal fee, to cover shipping and handling and a little more so we can start to build up a "kitty" of money for developing further synthetic materials in large quantities. I want to make this material available to to everyone in say, gram quantities, which should last most labs a lifetime or more, I'm thinking... what do you all think?

The poll above has $100 as the most popular "crowd source" amount, so maybe we'll start with a gram of RbTiOPO4 for $100 to start with and see if that is popular enough...

Marc Schrier is thinking that the CsZrOPO4 might be a good material to try and synthesize next?

[Brian Joy]

I'd certainly be interested in a good Cs standard.  Currently I use a natural pollucite from Hebron, Maine that I've characterized by a combination of microprobe analysis, stoichiometry, and solution ICP-MS (particularly to get Li2O).  My characterization of it is not really as satisfactory as I'd like it to be, especially since I've had to apply a number of assumptions, namely nAl + nSi =3, n large cations = nAl, and nCs + nH2O = 1.  I also have some Cs2SiF6 for which I paid an arm and a leg to get from SPI; however it turns out to be too water-soluble for water-based polishing (learned this the hard way) and is quite beam-sensitive.  The only other one I have is Corning Glass 95-IRW, which only contains 0.7 wt% Cs2O.

What have other people been using as a Cs standard?

[Marc Schrier]

In looking over what starting materials I have here, I don't have any good cesium starting materials.  Before I purchase some, I figured I'd bounce it off the group in case someone has some to share (or maybe access to a reuse facility).  To start, the materials do not need to be particularly pure, and only once everything is working well (prepare crystals (PXRD), it's water insoluble, polishes nicely, holds up to the beam, looks homogeneous...), then I can pick up purer starting materials.  To prepare the flux(es), any of the following would work as a cesium source: Cs, Cs2O, CsOH, CsOH•H2O, Cs2CO3, CsHCO3, CsOAc (CH3COOCs, cesium acetate), HCOOCs (cesium formate), Cs3PO4, Cs2HPO4, or CsH2PO4.  If you have any you can share, let me know, otherwise I'll try to place an order for some starting materials soon.  I'm not sure if CsZrOPO4 will work, but at least Cs2Zr(PO4)2 and CsZr2(PO4)3 have been prepared.  There are also some Cs/Ti/P/O's.  I probably ought to do some literature searches too and see what I can dig up!