Probe Software Users Forum

A couple of weeks ago, I had a user ask me if I had a vanadium silicate standard that wasn't a glass like the old standards from the University of Oregon. To my knowledge, there still isn't a commercially available vanadium silicate standard. In lieu of not having the equipment to try and grow my own material, I got to thinking about what we might be able to use for such a material. In the end, I decided to test out some cavansite, CaVSi4O11*4(H2O), a modulated layer phyllosilicate. There is quite a bit of this material available on ebay and through other sources that actually comes from the type locality in India.

I received two ~1 cm balls of the material for $20. I was actually quite amazed to find that it is extremely clean. At most, I found maybe three or four 20-40 µm grains of quartz in the material that I disaggregated for analysis (~1/2 a ball). Once I mounted up the material and got it in the probe, I was even more surprised at just how clean it is. Aside from the quality of my quick polish, there appears to be no real evidence for zoning or inclusions throughout the sample (see attached SE image). As well, a couple of element maps revealed no apparent zoning (see element map attachment).

So far, I have put down 74 analyses across 16 grains in the sample and everything has been really homogeneous. The results are summarized below:
   
AccV   15.0            
Curr. (nA)   20.0
Ox = 22               
         Avg      STDV   min        max     WebMineral
V2O5    20.138   0.702   18.588   22.123   20.15
MgO      0.242   0.116    0.061    0.445      -
SiO2    53.146   1.239   51.159   55.272   53.24
CaO     11.998   0.312   11.185   12.42    12.42
Total   85.696   1.343   83.221   87.946   85.81

I am pleasantly surprised with the results thus far. Element mapping before and after analysis reveals no obvious beam mobilization of the elements. The beam does cause a bit of damage to the specimen, however, the damage is not apparent in the post-analysis element maps. The lack of any zoning after the analysis would suggest to me that the specimen is just locally dewatering around the analysis spot and damaging the carbon coat (see attached beam effect.png). This should be expected as the mineral is a phyllosilicate.

Anyways, I am curious to see what everyone else thinks about the potential of cavansite as a V-silicate standard. If enough people are interested, then I'd be more than willing to send around some of the material for a "round-robin" test. There is plenty of it to go around.

Hi Matthew, I would be very interested in participating in such a round robin.
We've mostly been using synthetic Ca vanadate Ca3(VO4)2 for V calibration in oxide matrix, but having a V bearing silicate would be great for silicate analysis.

Best regards,
Karsten

Hi Matthew,
You might also to try adding the H2O to the matrix correction since an 85% total could be affected by "missing water". See here for more details:

https://smf.probesoftware.com/index.php?topic=92.msg8516#msg8516

john

Hi Matt,

very interesting. I have vanadinite (Pb5(VO4)3Cl) but a vanadium silicate would be very interesting. I would be interested in testing this material too.

Cheers,
Anette

Thanks for the quick replies! I took John's advice and recalculated the analyses with H2O by difference (WBD) and this is what the bulk stats look like, now:
         Avg           STDV        min      max    WebMin.   increase
V2O5   20.969   0.714   19.52      23.19     20.15      0.831
MgO      0.254   0.122      0.064     0.47          -          0.012
SiO2    55.537   1.088    53.756   57.422    53.24      2.391
CaO    12.464   0.306    11.639   12.877    12.42      0.466
H2O    10.61      1.043     8.874   12.545    15.96         -

So, we see an overall increase of 3.7% with each oxide increasing by ~3-5%. Compared to the Web Mineral stoichiometric composition, this still looks really good. I am going to do a little bit more work on this mount to see how the cavansite runs under different conditions.

My provider for the Cavansite has shipped ~10-15 grams of material to me that should arrive in about 2 weeks. If you are interested in having me send you some of the material to run, then you can send me an email and we will get everything setup.

It always amazes me how much an effect missing water can have on the matrix correction of the other elements...

For those wondering about hydrous glasses see this paper:

https://link.springer.com/article/10.1007/s00445-005-0003-z

Note that, if the formula unit of the compound is known, there is a really easy way to estimate the amount of H₂O present.  Just as it's possible to calculate the ratio of Fe₂O₃ to FeO for certain minerals in the normalization step within the iterations required to determine the matrix correction factors, it is also possible to calculate a mass of H₂O that is consistent with a formula unit.  For instance, when analyzing amphiboles in cases in which it's unlikely that any O^2- substitutes for OH^-, one can assume that nOH^- = 2 - nF^- - nCl^- (per 23 anhydrous oxygens) and apply this to the mass normalization for that iteration.  This approach, which I use myself, is incredibly useful when analyzing amphiboles, chlorite, biotite, muscovite, lawsonite (which also contains molecular H2O), and many other minerals.  I also find it to be a very intuitive approach.

Exactly.  All "unanalyzed" elements should be included in the matrix correction for best accuracy.

And as I'm sure many of you know, one can specify H2O or H2O (or any ratio for that matter!) by stoichiometry to any element, in Probe for EPMA from the Analyze! window Calculation Options dialog:

https://smf.probesoftware.com/index.php?topic=92.msg8485#msg8485

And one can even calculate hydrogen by stoichiometry to "excess oxygen":

https://smf.probesoftware.com/index.php?topic=61.msg229#msg229

and also in the CalcZAF app, again from the Calculation Options dialog.

Quote from: Probeman on October 08, 2020, 05:13:38 PM
And as I'm sure many of you know, one can specify H2O or H2O (or any ratio for that matter!) by stoichiometry to any element, in Probe for EPMA from the Analyze! window Calculation Options dialog:

I'm describing something different.  There is no need to specify a molar ratio of hydrogen to any other particular element.  This approach becomes inconvenient when significant F^- and/or Cl^- may substitute for OH^-.

Sorry, thought we were talking about vanadinate.

Yes, when halogens are present that is different.  See here for the oxygen-halogen matrix correction in Probe for EPMA:

https://smf.probesoftware.com/index.php?topic=1247.msg9305#msg9305

Basically it is especially significant when fluorine replaces stoichiometric oxygen.

Quote from: Probeman on October 08, 2020, 06:57:59 PM
Sorry, thought we were talking about vanadinate.

Yes, when halogens are present that is different.  See here for the oxygen-halogen matrix correction in Probe for EPMA:

https://smf.probesoftware.com/index.php?topic=1247.msg9305#msg9305

Basically it is especially significant when fluorine replaces stoichiometric oxygen.

Yes, I agree that this is an important correction that must be applied within the recursive algorithm used to determine the matrix correction factors; I apply it in my own programming.

But this is still not the point I was trying to make.  I'll start a new topic either tomorrow (if I get a chance) or over the weekend.

I do get the point that once one has a mineral formula, one has access to methods for calculating "unanalyzed" elements such as water in conjunction with halogens.  As you say, similarly to how excess oxygen from ferric iron is calculated once the number of cations and oxygens are specified e.g., Droop (1987):

https://smf.probesoftware.com/index.php?topic=92.msg8593#msg8593

I look forward to your topic.  Your insights are always interesting, especially to those like myself that are not geologists!