Help|Log In|Register|
Home PageMindat NewsThe Mindat ManualHistory of MindatCopyright StatusManagement TeamContact UsAdvertise on Mindat
Donate to MindatSponsor a PageSponsored PagesTop Available PagesMindat AdvertisersAdvertise on MindatThe Mindat Store
Minerals by PropertiesMinerals by ChemistryRandom MineralSearch by minIDLocalities Near MeSearch ArticlesSearch GlossaryMore Search Options
Search For:
Mineral Name:
Locality Name:
The Mindat ManualAdd a New PhotoRate PhotosLocality Edit ReportCoordinate Completion ReportAdd Glossary Item
StatisticsThe ElementsMember ListBooks & MagazinesMineral Shows & EventsThe Mindat DirectoryHow to Link to MindatDevice Settings
Photo SearchPhoto GalleriesNew Photos TodayNew Photos YesterdayMembers' Photo GalleriesPast Photo of the Day Gallery

What do you do when XRD & EDS indicate different minerals?

Posted by Fred E. Davis  
Fred E. Davis June 26, 2009 03:46AM
I've got a curious predicament. I collected an interesting specimen from a Connecticut pegmatite. The XRD pattern is an excellent match to betafite. The sample was heated because it was partially metamict; it showed a major peak around 30.2 degrees 2-theta prior to heating which developed into the major peak after annealing. I have XRD patterns from a friend in Basel, Switzerland, and my own data; they are virtually identical. Some other patterns I compared against that did not fit at all include tanteuxenite-(Y), tantal-aeschynite-(Y), uranmicrolite, polycrase-(Y), yttrobetafite-(Y) and pyrochlore among others; betafite is the best fit, and also chosen by an automated search/match.

Now consider an EDS analysis that returned the following Wt % oxides:
Y2O3 9.51%, Nb2O5 16.47%, ThO2 2.38%, UO2 3.91%, CaO 5.21%, TiO2 18.02%, Nd2O3 3.66%, Ce2O3 1.79%, La2O3 1.20%, Fe2O3 2.59%, Ta2O5 35.25%.

From this, I derived structural formulas, with the best fit based on six oxygen atoms:
(Y+REE0.447 Ca0.366 Fe0.142 U0.057 Ce0.043 Th0.036){1.091} (Ti0.889 Ta0.629 Nb0.488){2.007} O6

I will become rapidly clear that if it's betafite, there is a problem. For one, betafite has two atoms at the A site, while my best fit provides only one. The next problem is that my specimen is Ti>Ta>Nb, compared to Ti>Nb>Ta for betafite. In fact, I could find only one Ti>Ta>Nb mineral poking around Strunz & Dana (it's not even close *and* it was discredited in 1977).

The A site is dominated by Y, so let's consider for a moment betafite-(Y). Just in terms of elemental composition, my sample is fairly close to yttrobetafite-(Y) *except* for Ta>Nb, two A site atoms vs. one, and the XRD pattern for yttrobetafite-(Y) is a very poor fit.

Suggestions? Ideas?
Donald Vaughn June 26, 2009 05:35AM
I doubt any minerals in this group will ever be "ideal" formulas so relax give it a generic designation as a betafite subgroup in the pyrochlore group stick the formula after and be happy. but that is just the opinion of a rank amateur, someone with more knowledge and wisdom feel free to call me a moron I won't break.
jacques jedwab June 26, 2009 08:37AM
If you manage to get a polished or transparent section of your mineral, you will be surprised by the various microscopical phases present, reflecting chemical and structural "anomalies". Metamictization (partial) is frequent. (Find with google-scholar:: Van der Veen: "A study of pyrochlore" (1963) and Lumpkin & Ewing: "Geochemical alteration of pyrochlore group minerals: betafite subgroup" - American Mineralogist, 1996.
Pavel Kartashov June 26, 2009 08:49AM
Fred, you have tantaloan yttrobetafite-(Y) according to classification of Hogarth. In coming soon new classification of pyrochlore group this mineral will obtain much less attractive name. :( So you have some month to enjoy this yttrobetafite. :)
What about some difference in XRD patterns of your mineral and yttrobetafite from ASTM card, it may greatly depend on annealing conditions in both cases. During a heating quite different daughter phases able to form together with recrystallization of the main metamict phase. Their nature and quontity can be quite different in both cases. So you obtain different difractograms.
What about Ti>Ta>Nb succession in the B site. Relations of the second and the third elemens don't effect on position in systematic even in Hogath's nomenclature. In coming new nomenclature this relations will means nothing. Only the first/main element able to produce mineral specie name.
By the way most of hatchettolites from carbonatites are strong enriched with Ta and may to have Nb~Ta~Ti. But subgroup name (pyrochlore-microlite or betafite) depends from ratios 2Ti and Nb+Ta (according to Hogarth) or Ti and Nb+Ta (according to new nomenclature). So maximum what you able to obtain from this composition even in the network of Hogarth nomenclature is only adjective "tantaloan".
What about total population of A site. Compare both formulas of my betafite with one from textbooks - they also are quite different. :)

Kind regards,
Fred E. Davis June 26, 2009 01:16PM
Thank you all for your replies. I truly appreciate your input.

Yes, annealing can cause problems, especially with excessive temperatures (introduces new phases). Annealing was first performed by my friend in Basel (plot attached, SlocumXRDSwc). The three scans show before annealing (blue), and at two stages of heating (600C red, then 800C black). The reference lines are betafite. Also attached (Betaf-13-0197) is one of my own patterns, with ICDD 13-0197 as the reference. I heated it at 750C for about 2 hours.
open | download - SlocumXRDSwc.jpg (319 KB)
open | download - Betaf-13-0197.jpg (66.5 KB)
Uwe Kolitsch June 26, 2009 03:33PM
On how many spot analyses is the EDS-derived formula based?
Was any chemical inhomogeneity or zonation observed?
Fred E. Davis June 26, 2009 04:19PM
I cannot answer the question - I sent a specimen to Excalibur Minerals for analysis (see attached). It's certain to be fewer points than I would hope, but that's what I got for $50. When the opportunity arises, I plan to run a serious EPMA, element mapping to look for zonation, etc.
open | download - SlocumEDSpg2a.JPG (118.7 KB)
Chris Stefano July 26, 2009 03:48PM
That's a single point EDS spectrum- you have no idea if there is any chemical heterogeneity at all.
Rob Woodside July 26, 2009 05:56PM
That's an impressive EDS result. Total is 100% and no standards mentioned. You are right to look for WDS or EPMA.
Chris Stefano July 26, 2009 07:15PM
It is most likely a standarless analysis- in which case the instrument will normalize to 100%- not impressive at all- just how it's done. You should never trust a quantitative analysis from an SEM- they are really only semi-quantitative at best. If you really need quantitative, you not only need to use an electron microprobe, but you need to do it on a polished section and standardize with well-known standards that are similar to the material being analyzed (no pure oxide standards to study silicates for example).
Fred E. Davis July 26, 2009 07:48PM
Excellent advice from one and all. Many thanks to all who responded!
Rob Woodside July 26, 2009 10:39PM
Thanks Chris, I was being sarcastic.
Pavel Kartashov July 27, 2009 12:27AM
But both analyses show to us yttrobetafite-(Y). ;) May be it is impure, zonal etc., but yttrobetafite present in this mixture.

Only presence in XRD of lines belonging to TiO2 polymorphs may be fatal for this identification. In such case it able turn out yttropyrochlore.

Also i have a question. Form this mineral any crystals typical for cubic niobates - octahedrons, cubooctahedrons or it was met as massive grains? Very often cubic niobates replace (during albitization process) more former rhombic precursors - euxenite-aeschynite, samarskite, fergusonite. In such case resulting pyrochlore mineral is enriched with Y+HREE and with exeption of fergusonite as precursor by Ti. By the way, very often during such pyrochlorisation U and Th form own uraninite-thorianite phase.

When you make EDS from unpolished grain, it is too simple to analyze later pyrochlore and rombic precursor together. In polished preparate these phases has quite different reflectance and are well visible in reflected light. Also they should be well distinguishable in reflected electrons on microprobe.

Edited 1 time(s). Last edit at 07/27/2009 12:44AM by Pavel Kartashov.
Fred E. Davis July 27, 2009 12:47PM
No individual crystals were found. It's essentially massive, but does have a distinct, almost linear structure (see photo, compressed for MinDat). "Almost linear" means that the lines are not perfectly straight, but sometimes curve slightly. The appearance in the hand is a material that was extruded (only an analogy to its appearance). The specimen in the photo is about 50 mm by 35 mm. Yellowish areas at the edges are a golden muscovite. The specimen is a very dark brown in massive areas, and a translucent yellowish-brown in thin pieces. Small pieces (1 mm to 5 mm) are available (in limited quantity, of course) to anyone who wants to take a "closer look". ;) (PM a mailing address)
open | download - DSCF4060a.JPG (831.7 KB)
Pavel Kartashov July 27, 2009 04:49PM
I am see typical features of initial euxenite-aeschynite mineral agregate - "linear structure". So your yttrobetafite is of secondary origin. I'd investigate a replacement of euxenite-(Y) by yttropyrochlore-(Y) and uraninite in specimen of "yttrotantalite" from Southern Agder, Norway.

By the way yellow powder on sides of the specimen and on muscovite is kasolite or uranophane.

My adwice is to make polished section from large (5 or more mm) fragment of this aggregate in direction perpendicular to "linear structure" and look on it under reflected light or in BSE on SEM.
Fred E. Davis July 27, 2009 07:52PM
Thanks, Pavel. Are you referring to the 1931 article by Andersen for the Norwegian yttrotantalite?

There is no kasolite or uranophane present, only golden muscovite. Compressing the image for MinDat may have blurred such details.

Two additional photos show close-ups of quasi-linear surface details attached. #4247 width ~ 12 mm. #4253 width ~25 mm.
open | download - DSCF4247c.jpg (904.4 KB)
open | download - DSCF4253b.jpg (497.6 KB)
Fred E. Davis July 29, 2009 08:29PM
Not to confuse the issue or anything ... 8-)

I've been studying an interesting article: T. Scott Ercit, "Identification and alteration trends of granitic-pegmatite-hosted (Y,REE,U,Th)-(Nb,Ta,Ti) oxide minerals: a statistical approach," CanMin Vol. 43, pp. 1291-1303. This seemed to be quite appropriate to my situation. Accepting the fact that I don't have great EDS data to work from, I decided to plug in my values into the author's three-group model just to see what fell out. The plot is attached. This suggests my unknown falls in the euxenite-aeschynite groups. To further refine that, the author gives another comparison, and that result suggests euxenite. Curiouser and curiouser. . .

As an easy reality check, I put in data for a known samarskite-(Y) which appears right where it should. I recently added pyrochlore, fergusonite-(Y), and two tanteuxenite-(Y) points from reference data sheets.

Edited 3 time(s). Last edit at 08/03/2009 05:02PM by Fred E. Davis.
open | download - SlocChem5a.jpg (391.8 KB)
Johan Kjellman July 29, 2009 11:29PM
I agree with Scott - your sample is, or rather was, an AB2O6 oxide (aeshynite or euxenite group). Your analysis is of course not complete and the sample is for sure at least partly altered but still give a quite decent formula with A1.07B1.99O6.
I also agree with a lot of Pavels comments, these minerals become "pyrochlorized", mainly due to Ca-exchange. BUT remember Ca can also be a primary component within the AB2O6 framework as a "fersmite" component (Ca(Nb,Ta)2O6).

The pyrochlore politics aren't even settled yet - so why would you like to pursue this further?

Pavel Kartashov July 30, 2009 12:08AM
OK! It is tanteuxenite-(Y). You had calculate the formula on euxenite-aeschynite O6 basis from the beginning and were right. Pyrochlore basis is O7 or O6-x(F,OH,F)1+x where x may to be 0-2 or more. So nobody calculate pyrochlore formulas by oxygen method.

Now try to recalculate it by cation method on basis Ti+Ta+Nb+Fe3+=2.

Aparently you have only slightly "pyrochlorised", completely metamict euxenite. After heating, pyrochlore phase was recrystallized, but euxenite one hadn't. So you obtain pyrochlore pattern from mixture of pyrochlore+glass+REE2O3+TiO2. Pyrochlore phase in this mixture faster of all has yttropyrochlore-(Y) composition (with low Ti content).

You need in further microprobe investigation of polished section of the mineral to find out compositional heterogeneity of it.

Look at - such "stibiomicrolite" also will give PXRD patterns of usual microlite with some additional/exessive lines... ;)
Fred E. Davis July 30, 2009 01:13AM
Johan - "... - so why would you like to pursue this further? "

Inquiring minds want to know! I've learned a great deal tilting at this windmill, and hope to learn more.

Pavel - Yes, I hope to get some SEM time in the near future. Polished cross section & element mapping should be interesting. A description of tanteuxenite-(Y) crystals mention striations along {100} and {110}; interesting.

I truly appreciate all of your thoughts and suggestions.
Sorry, only registered users may post in this forum.

Click here to login

Mineral and/or Locality is an outreach project of the Hudson Institute of Mineralogy, a 501(c)(3) not-for-profit organization.
Copyright © and the Hudson Institute of Mineralogy 1993-2017, except where stated. relies on the contributions of thousands of members and supporters.
Privacy Policy - Terms & Conditions - Contact Us Current server date and time: May 24, 2017 07:14:18
Go to top of page