potosiite/franckeite

Can anyone please explain to a mere geologist in simple English what is the difference between Potosiite and Franckeite? I have good crystallized franckeite from Bolivia and relatively well-crystallized potosiite from Japan, and can not visually see any difference. I have had two different labs do a microprobe analysis which, given the possible substitution between divalent tin and lead, is indeterminate - the analyses could easily fit either species. Also, three people have done XRD work on the materials. Single crystal X-ray of the Japanese material gives true potosiite. Powder XRD of either material gives indeterminable results - supposedly because the material is soft and scaly, so it is difficult to grind them to a truly randomly oriented powder. I wonder at the wisdom of treating these as separate species when it is so difficult to differentiate them. Could they perhaps be only polytypes? (I must admit I have only understood half of the previous discussion here as to what constitutes a polytype.)

To further complicate the issue, I have two very obviously cylindrical cylindrites - from Oruro and Llallagua, not the cylindrite type locality of Poopo. It would never occur to me to waste time doing XRD on something so visually obvious as cylindrite, but I have fanatical collector friends in Boston and Prague who habitually x-ray every specimen that comes their way, and they both claim that BOTH the Oruro and LLallagua "cylindrites" are clearly franckeite! How much other franckeite out there is masquerading as cylindrite? And, again, what's the significant difference between the two?

I'll bet if I dig some more, I can find some more headaches in this group of minerals. A few years ago, at the Tucson show, I met a group of researchers from the University of Texas in El Paso who were studying the cylindrite group (with a grant from the U.S. Navy, of all people). One of these gentlemen cryptically mentioned that one member of the "cylindrite group" might need to be discredited as a species, but he wouldn't say which one.

Since I collect a couple times a year in tin/silver mines with members of this group, I have a special interest in them.

Greetings from Alfredo

Dear Alfredo,
at the moment I can only add to the confusion: several years ago I x-rayed both a "franckeite" and a "teallite" from Bolivia, and both turned out to be potosiite-H (at least the fit with the ICDD-PDF card for potosiite-H was perfect, but maybe the nomenclature has changed subsequently). The franckeite sample may also have had admixed stannite(?).

Cheers, Uwe

PS Hope to find out more information on this issue and post it here.

Dear Uwe,

Thanks for your interesting response. The "teallite" I can easily believe was something else. I have long been annoyed by all the franckeite (or potosiite?) being sold by many dealers at shows labelled "teallite". I am sure that 75% of "systematic collectors" who think they have a teallite, really do not have it. Actually teallite is not difficult to visually distinguish from franckeite - the morphology is different, and teallite does not have striations. And they never occur together in the same specimen. A teallite from Carguaycollo mine is probably really teallite; from anywhere else it's probably franckeite (or potosiite?).

I'll patiently wait for more news from you. Thanks!

Hello Alfredo

If the chemical formulas given in Dana 8th are to be believed, the difference is in the structural position of the Fe. In Franckeite it is in the A site, while in Potosiite it is in the C site.

Dana also gives different cell dimensions and angles:

Franckeite - a = 46.9, b = 5.82, c = 17.3, alpha = 90.0, beta = 94.68, gamma = 90.0, Z = 8, D(calc) = 5.88.

Potosiite - a = 5.915, b = 5.938, c = 17.239, alpha = 91.62, beta = 91.02, gamma = 90.84, Z = 1, D(calc) = 6.20

I don't see how franckeite can be considered Triclinic with two 90 degree angles in it - but, then, I'm not a crystallographer.

From what they give, it looks to me like the potosiite analysis is the more rigorous of the two. Perhaps an equally rigorous examination of franckeite would show it to actually be potosiite?

I'd add cylindrite to the question as well - it is very close in structure (and identical in it's chemical constituents) to the other two.

The main question is the structural position - and valence - of the Fe.

Just thought I'd help cloud the issue further... :~}

Regards

Al Plante

Dear Alfredo,
I have sent various references on potosiite, franckeite, incaite and cylindrite to your e-mail address.
Clearly, these minerals are nothing easy to identify.

Cheers, Uwe

Alfredo:

These look pretty beastly to discriminate with XRD, except for single crystal work. The enormous unit cell length of 46.9 Angstroms is going to require scanning really low angles. If they are all soft, lamellar and deformable, one also worries about actually inducing crystallographic transformations in the grinding process; thus further muddying the waters (e.g., maybe creating mixtures of polytypes). Preferential orientation per se can usually be prevented by the trick of spraying the mineral powder with plastic fixative, and then sieving out the little plastic spheres with crystal platelets stuck to them at random orientations (this works for Micas!).

If one had authentic material to work with, Moessbauer spectroscopy (ca. $350 a pop, unfortunately) would be a wonderful tool for examining this group of minerals. Not only could one look at the usual Iron-57 Moessbauer spectra, but tin and antimony are also readily accessible Moessbauer elements. Unfortunately, few research groups remain in the United States that can study all three Moessbauer elements (I'm not sure B.J. Evans still does all these since he left the University of Michigan). Maybe one of the German Universities can still do this kind of multiple element work. At any rate, Moessbauer spectroscopy (gamma ray absorption spectroscopy) is frequently capable of detecting different oxidation states , and different coordination numbers of certain elements with "Moessbauer - friendly" isotopes.

The mineral data base (webmineral.com) seems to have left off the iron in the Franckeite formula. Also it's not clear to me what all the oxidation states should be in these minerals. Ferrous iron perhaps, but what of lead, tin, and antimony. It could be really interesting.

Although Ag does not apear in the franckeite formula (too little of it), franckeite from 6 different mines in Bolivia always has between 0.5 and 1.0 wt% Ag. I don't know where the Ag fits in the structure. It might perhaps be essential to the structure? Or is it perhaps not in the franckeite at all, but rather as micro inclusions of incaite or some other Ag-bearing phase? I'm curious as to whether franckeite from California or other places also contains any silver.

Thanks, Alfredo for highlighting this really interesting group of minerals!

Getting back to your original question: What is the difference between franckeite and potosiite? I've done a little more checking of formulas, etc. The Athena site is especially helpful here, because the oxidation states are more completely listed.

Frankeite: [Pb, Sn+2]6 [Fe+2][Sn+4]2 [Sb+3]2 S14
Contains some stannous tin on the lead sites.

Potosiite: Pb6[Sn4+]2[Fe2+][Sb5+]2S16
Contains only stannic tin. Also note the presence of antimony in the +5 oxidation state. An interesting mix: two elements, Sn and Sb in their higher oxidation states, but iron still as ferrous!

Incaite is given as (Pb,Ag)4Sn4FeSb2S15 with some silver on the lead sites, but with no indication of how the charge compensation for silver is accomplished. Formulated with no silver, I think there has to be a mixture (ordered or disordered?) of Sn2+ and 3Sn4+ cations to achieve charge balance.

Cylindrite is given as Pb4[Fe2+][Sn4+]4[Sb3+]2S16

The IMA has used the group as an example of what is a mineral:
"Composite structures of members of the cylindrite group are formed of two kinds of layers, pseudo-hexagonal (H) and pseudo-tetragonal (Q). Cylindrite and franckeite have the same Q-H-Q-H sequence of stratification, but in franckeite the width of the Q layer is twice that of the Q layer of cylindrite. The two minerals are therefore regarded as separate species." See:
[www.minsocam.org]

Nowhere have I yet found any mention of whether the ferrous iron in these minerals is high-spin or low-spin. It makes a big difference in the magnetic properties. High-spin ferrous iron (as in pyrrhotite) has a large paramagnetic moment; low-spin ferrous iron (as in pyrite) yields essentially a non-magnetic (or diamagnetic) material.

It would be very, very difficult with elementary methods to detect the oxidation states of the different elements in this group of minerals. The Moessbauer spectroscopy of Fe, Sn and Sb should yield the best discrimination of oxidation states in the bulk phases.

Apparently Moessbauer spectroscopy has already been done for three of the minerals:

Moessbauer Data Center
[www.unca.edu]
franckeite Fe,Sb,Sn
cylindrite Fe,Sb,Sn
potosiite Fe,Sb,Sn
Data for these has been collected in an expensive book. I have not tracked down the original publications.

A quick search on the web turned up the talk listed below.


ISIAME'96

International Symposium on the Industrial Applications of the Mössbauer Effect
Johannesburg, South Africa - November 4-8, 1996
Communications on Mineral and Mineral Processing


H. Mehner
Mössbauer Investigations on Minerals of the Frankeite-Cylindrite Group

Bundesanstalt für Materialforschung und -prüfung, Rudower Chaussee 5, D-12489 Berlin-Adlershof, Germany

sunsite.wits.ac.za/conferen/isiame/ isiame_backup/Third26.html

Thank you so much for such a fascinating thread.

A database search for Pb AND Sn AND Sb brings up a Cylindrite group containing Leviclaudite, Incaite, Potosiite, Franckeite, Cylindrite. These are obviously sheet structures and clearly related species. The compositions posted here allow for a straight forward probe distinguishability between them, apart from the visually distinguishable Cylindrite. The obvious tealite and other tin sulfides are ommited. A text book example of a simply defined chemical group!!!! Wrongo!!!!

MIneralogy has gone from distinguishing mineral species with a keen eye and clear mind to the technology of hardness kits, loupes and streak plates; through blow pipes and wet chemistry; to flame spectroscopy, XRD and electron probes and now Mossbauer spectroscopy!!!! I love the technology. I'm hungry for the information it produces. But are the differences this information produces sufficient to define new species????

What is so upsetting about this lot is the visually obvious cylindrite x-raying out as potosiite. Pseudomorphs? Unlikely. I think William is on the right track with sample prep. Sulfide glasses crystallize on grinding, yielding false powder patterns! With the excellent cleavage and extreme maleability, I wouldn't trust any powder data on the cylindrite group. The lines must be very diffuse; and the crucial low angle lines probably lost, causing Cylindrites to x-ray like Franckeites. Perhaps a single xl collected under liquid nitrogen might tell the tale.

William, I'm surprised at your suggestion of polytypism on grinding. Please correct me if I'm wrong. I thought polytypes were growth features resulting from stacking faults. Are you suggesting a recrystallization on grinding?.

I'm trying to understand homologous structures and struggling with the idea of mixed structures. I can see twinned crystals (especially polysynthetic twinning, sometimes due to stacking faults), stacking faults per se (polytypism), epitaxi ( not just the fortuitous fitting of two totally different structures, but a fitting due to a common structural element of both structures as in Binnental Pb sulfarsenides) and exsolution lamellae which share a common structural element with the host; all as examples of structural mixtures. Is this right? My first experience with homologous structures was in the semseyite group where PbS slabs in the structure get thicker by fixed amounts as one proceeds through fulloppite, heteromophite, plagionite, to semseyite. Are these structures as clean as the literature would have us believe? How do they grow? Why does the series stop at semseyite? The homologous series with the Pb Bi sulfides are mind boggling. I'm pleased to see that their validity rests on predicted members that have yet to be found. Little or no attention seems to be paid to the conduction electrons, except for the caveate that the predicted structures may require more or different cations to "materialize". Could the homologous series be regarded as structural mixtures and therefore interesting growth curiosities rather than new species?

Al, here's a model of a tricinic xl with all axial angles equal to 90 degrees. Take a normal brick with point group 2/m2/m2/m. Stack a pile of such bricks in a regular way to get an orthrhombic xl with 90 degree axial angles and point group 2/m2/m2/m. Now replace each brick in the xl with identical warty potatoes that don't even have an inversion centre. Voila! a triclinic potatoeite xl with 90 degree axial angles and point group 1.

Polytypes are a subtle form of polymorphism especially characteristic of layered substances that potentially can be stacked in a variety of ways. The crystal structures are distinct; however, the free energy differences between polytypes can be almost infinitesimal because the arrangement of nearest neighbor atoms produces the biggest free energy effects; whereas, the next-nearest neighbors produce much smaller differences. We may know that only one of the polytypes is actually the stable phase under given conditions of temperature and pressure, but proving it thermodynamically can be very challenging because it may require extraordinarily precise calorimetry and thermal property measurements.

That said, it is not at all mysterious why grinding a sample can induce changes in polytypes. Grinding actually can deposit a considerable amount of energy in a small volume, and can convert one polymorph into another (let alone polytype!). In extreme cases, e.g., the long grinding of alpha quartz, one can produce total amorphization of the lattice. Glaciers do this sort of work with rock flour all the time.

A healthy skepticism about powder patterns for certain sensitive substances should be encouraged! The softer the materials, perhaps the more skeptical one should be.

Addendum on polytypism:
[www.iucr.org]
This is a fairly compact treatment of the nomenclature of polytypism.

There are problems with a strictly geometric approach: What if the layers differ slightly in composition? How big a composition difference should be "allowed"?

The Magneli phases (or crystal shear phases) taken from inorganic chemistry are similar to polytypes except that compositional differences exist between layers. Examples are found in the countably infinite number of oxides intermediate between Ti2O3 and TiO2. Similar series show up for the oxides of molybdenum and uranium.

William , thank you so much. Polytypes are not structural mixtures! Denying them species status must amount to saying that the differences in nearest neighbour environment between two polytypes are on the average too slight to define a species. This fits well with the idea that minor chemical impurites produce such slight differences on the average in nearest neighbour environments that separate species are unwarranted.

I think there is something deep here and want to consider it further. I have always been bothered my minerals with large unit cells. As the cryatals form how does it know after 20 angstroms growth that it is time to insert, say, the next Cu atom? Almost all the action, from hardness to colour centres, occurs in the nearest neighbour environment. Growth steps from spiral dislocations go along way to explaining large unit cells, as well as polytypes, but how do they get started? Some years ago, the first scanning tunneling microscope pictures of hematite and galena cleavages showed surface undulations with ~10 angstom periodities. Perhaps micro xls of hundreds of thousnds of atoms that close pack and have such oscillations, act as templates to nucleate the large unit cell minerals. But how would one detect even a few million nucleating atoms in the better part of Avogadro's number of mineral atoms? Have the new atomic microscopes shed anymore light on this?

The iucr site was really helpful. They list some polytypes and to my joy I found that argentopyrite is a polytype of sternbergite. Argentopyrite always froms pseudohexagonal trillings and is immediately recognizable. Unfortunately the trillings are almost always limonite pseudomorphs after argentopyrite. Most labeled argentopyrites are these pseudos. All sternbergites are orthorhombic tablets and frequently suffer the same pseudomorphosis. Check your specimens by poking the xl that you would sacrifice for an x-ray, with a needle. If it bends or deforms easily, you have the real thing. If it crumbles, you've got the pseudo. I have never seen a specimen containing both polytypes. Cubanite is the Cu analogue of sternbergite and occurs as fine acicular hairs, orthorhombic tablets (often twinned), and trillings; sometimes on the same specimen. Mark Mauthner found significant Ag in a cubanite needle from the Sylvana Mine, indicating a possible series between cubanite and sternbergite. The silver concentration increased towards one end. If there is a series, this gives another example of the "species" oscillation in the same xl.

I had never heard of "Magnelli phases" before and a Google search of that phrase brought up five pages of the most diverse listings imaginable; from the art world to molecular biology, mostly having nothing to do with William's topic. However, the very second listing has blown me away. It is a short paper on Sulfosalt Classification !!!! The paper is entitled "Sulfosalt mineralogy today" presented by N.N.Mosgova in June 2000 at an IMA COM short course in Finland. There are loads of references and the first explanation I've ever seen of homologous series as precipitations from non-stoichiometric phases. Mosgova also suggests that that an entire homologous series be given species status with its members being structural varieties. Thus fulloppite, heteromorphite, and semseyite would be varieties of Plagionite, first described in 1831. How do we get the IMA to reverse its pronouncement that homologues are separate species? When that happens potosiite and franckeite will be a variety of Cylindrite, first descibed in 1899. William's pertinant question on "allowed" compositional differences must be answered before the Ag and Cu bearing structures also become varieties. Alfredo's observation that franckeites from 6 different Bolivian mines all have .5-1 % Ag points to the answer. Incaite has only 2% Ag in the average!!! of 5 analyses.

This suggests another expanation for the large unit cells - a simple ordering on cooling!.

Rob: Watch your spelling. Magneli phase on Google yields a pretty good haul of examples. Magneli's original work was on the molybdenum oxide series, but the name has been generalized to other systems of oxides of multivalent metals.

The series between Ti2O3 and TiO2 (rutile) is a good example. The members are written Ti(n)O(2n-1) where n = 2,3,4,...
[really Ti3O5 doesn't seem to belong structurally]. The true members start with Ti4O7. Several higher ones have been characterized. Chemically they behave as if one is mixing Ti2O3 (corundum-type) layers with rutile layers with the corundum-type layers inserted along certain crystal shear planes. So here, there is the combination of shearing and oxygen loss. In polytypes, sensu stricto, there is just shearing of the stacking layers; this way and that, to produce the different stacking sequences.

Your point about communication between rather distant parts of the crystal lattice is one of the key questions about equilibrium between different polytypes. How does one layer "know" what's going on hundreds of angstroms away? Because of the small free energy differences between higher order polytypes, there is no doubt a strong tendency to freeze in disorder of various kinds, and the kinetics of annealing out the disorder may be excruciatingly slow.

William,Thank you again. I admire your patience. The right spelling gives much better google results, but I'm glad Mosgova made the same mistake. Thanks to your efforts I'm getting a grip on Magneli phases and polytypes.

Mosgova's recommendation:" A Homologous series, containing the same chemical elements, was reccomended to be considered as a mineral species and its members as structural varieties" would seem to make rutile and Ti2O3 with corundum structure the same species. Thank god the corundum analogue doesn't occur! I wonder why not? (Nature is subtle, but not malicious?). If it did, I suppose people would want to apply some 50/50 rule and create two species. In view of the layering in cylindrite and Neyite, it is no stretch to imagine some ternary Magneli phases and follow Nickel judging each case on its own merits!

Ramdohr in Ore minerals 2 laments the lack of correlation between synthetic and natural platinum group compounds, blaming it on long diffusion times. Mosgova suggests suggests that these homologous phases exsolve from a nonstoichiometric precursor. In general this makes sense for large unit cell phases. Relaxing the conduction electrons structure by thermally diffusing the "ions" into large unit cells seems energetically possible at 500K with diffusion times of a few million years. This kind of potential enenergy lowering can produce long range order. However, this would have to be reconciled with lab growth times and the morphology of the precursor xl agreeing with the symmetry of the ordered phase. The latter should present little problem since most of these are monoclinic or of low symmetry.

Are the layers in these phases always perpendicular to the axis of slowest growth?

I recently added a picture of potosiite, and was suprised to see it indicated as a variety of franckeite. I then stumbled on this interesting discussion, from almost 6 years ago !?! I am curiuos as to whether there have been any updates on the status of the cylindrite group

thanks, Jeff

Jeffrey,
Check the "Sulfosalt systematics: a review. Report of the sulfosalt sub-committee of the IMA Commission of Ore Mineralogy", European J. Min. 20, 7-46 (2008). I don't believe it's been posted to the IMA website yet.

Potosiite and incaite are considered Sn²⁺-poor and Sn²⁺-rich varieties of franckeite, respectively. All incaite compositions show Pb > Sn²⁺, and could be revalidated if Sn²⁺ > Pb can be found naturally.

The sulfosalt report was uploaded today on the CNMNC website.

Great timing, thanks!

Are all of the minerals listed as valid now accepted by the IMA - for example, plumasite?

thanks