FMF - Friends of Minerals Forum, discussion and message board

[Jesse Fisher]

I think we understand the geometry of spinel twinning (though nice model!). What is in question is just why spinel-twinned fluorite always seems to have a sutured composition plane rather than a straight one as seen in the twinning of other octahedral crystals such as spinels and diamonds.

Below are a couple more photos of spinel-twinned fluorites from Naica.

[Roger Warin]

Hi,
Pete could certainly better answer than me. But here's how I understand the phenomenon.
A first remark, Nature is more capricious than when we try to explain it. Also to simplify the facts, we take extreme situations.
When we say that there is a composition plane in a contact twin, it is assumed that individuals are far from each other, but they touch anyway.
If the distance between the centers of two crystals (or individuals) decreases, a partial interpenetration occurs. The edifice is deformed from the wooden model.
The distance between the centers may still decrease, resulting in a greater interpenetration.
Ultimately, the two centers overlap and we have the perfect interpenetration, another limit of this evolution.
At the atomic level, the twin plane is the same, the face of a {111} octahedron. It is always the same twin.
The contact twin is a mirror twin plane.
Interpenetrating individuals do have an irregular composition surface. For this reason, the twin is defined in this case by a twin direction axis, e.g. [111] for the spinel law.
[…] is used to define a vector, {…} for a crystallographic form, and (…) for a defined plane.
Roger.

[Ru Smith]

Another excellent Naica specimen, Jesse, once again highlighting the degree of interpenetration and absence of a single composition plane in these twins. Going back to the very first image in this thread, I've taken a moment to colour-code the two members of that Erongo twin.

[Ru Smith]

Back to some real spinel twins, these Peruvian alabandites show spinel twinning of cubeoctahedra, octahedra and some other forms I can't decipher.

[Ru Smith]

Another one with a clear composition plane.

[Ru Smith]

I now have one of the little Naica gems to look at closely and here is a quick batch of photos.

The first thing to note is that these are not spinel twins, but beautiful penetration twins (the usual [111] 60 degree rotation). I wonder where the spinel twin idea came from for these? The flattening?

They are complex crystals with numerous small faces. An untwinned crystal on the back has cube, dodecahedron and trapezoidal icositetrahedron. I think all of these forms are probably present in the twins, plus the etched octahedral faces. As you can see, there is selective etching of faces. In the untwinned crystal on the back the icositetrahedron faces are etched and the cube faces not.

looking inside, there are the curving wires of probable sulphosalt (4th photo) mentioned above by Peter Megaw and also a bubble (5th photo) that wanders up and down its tube as you tilt the crystal.

Quite a lot going on here. Good fun!

[Ru Smith]

By the way, each of the large apparent octahedral faces on the Naica specimen is composed of numerous minute cube corners; there are no reflective surfaces parallel to the octahedron. This is very similar in character to the famous "c-faces" of quartz from Diamantina or amethyst from Four Peaks, Arizona. The dodecahedron and cube faces in the Naica specimen are smooth and lustrous. The trapezoidal icositetrahedron faces are rough.

The attached gem pink fluorite from Pakistan is also not a spinel twin, but rather a penetration twin - also a challenge to identify all the faces on this one, though octahedron, dodecahedron and trapezoidal icositetrahedron faces become clear after turning it around for a while.

Do spinel twins of fluorite exist? I've yet to see one.

[Ru Smith]

Some explanation of the Naica twin.

[Ru Smith]

In the Naica image above O is octahedron, C is cube, D is dodecahedron, and I is icositetrahedron.

By way of contrast, here's a spinel law contact-twin of sphalerite.

[Ru Smith]

Checking Dana's Textbook of Mineralogy (pages 165-166) I see that he distinguishes contact-twins (spinel-twins) from penetration twins in isometric crystals, both with "the twinning axis an octahedral axis". He shows one example of each for Galena; 401 being the contact twin and 404 being the penetration twin (attached figure).

I found some nice examples of the penetration twin form on a specimen of galena from Naica, Mexico.

[Ru Smith]

So, I think I've found an actual spinel twin of fluorite. It's from Naica and involves cube-dodecahedral crystals. The twin plane is parallel to an octahedron face, but there are no octahedral faces expressed, so the twin looks very different from the flattened pseudohexagonal penetration twins. Nice!

[Ru Smith]

Another view. The trace of the twin plane is in the reentrant just beneath the quartz to the right and is shown again by the reentrant on the left. A cube face of the upper member of the twin is facing us, with dodecahedron faces above to the left and above to the right. The twin axis passes through the triple junction of these two dodecahedron faces and the third one just out of sight at the back.

[Ru Smith]

found this in Chang, Howie & Zussman (1996).

[crazy.stone]

beautiful especially color. pretty nice.

[Ru Smith]

Thanks crazy stone.

There are lots of interesting colours here, but this thread is mostly about shapes - stimulating spatial mental puzzles!

[Ru Smith]

Here is a very large one of the Erongo pseudohexagonal penetration twins.

[Ru Smith]

Here I've added some backlighting. The cube faces are naturally colour-coded in dark purple, making the twin easier to decipher.

[Ru Smith]

This is approximately what's going on with the interpenetrating twin elements. The blue-shaded element has purple cube-faces adjacent to the NW, NE and S ("N" assumed to be up) edges of the front shared hexagonal octahedron face. The unshaded element has purple cube faces adjacent to the SW, N and SE edges of the shared frontal octahedron face. The pattern is similar on the reverse side of this floater, but I'm too lazy to draw it.

[Josele]

This is a very interesting thread. Ru's twinned crystals collection is a truly treasure and explanations and drawings are very clear and instructive, thanks for your effort.

Concerning spinel and penetration twin in fluorite, seems that is not easy to establish a border line between them. Theoretically spinel twins have a contact plane interface but actually this surface is not plane and occupies a 3D space, sometimes very intrincated. Where is the difference with penetration twins? In the fact that penetration twins cross to the other side? In the degree of intrincation? Could be that they are the same thing? Could be that this is only a morphological classification easily applicable in ideal geometric forms but perhaps is not valid in some wayward species as fluorite?
After all, mystery of twins is how they manage to match both crystal lattices and this miracle occurs in both cases...

[Ru Smith]

Many thanks Josele.

Nice model of the penetrating cubes. I still haven’t seen the ideal example of this in nature, though a fluorite found by Arthur Scoble in the Blackdene Mine comes close.

I didn’t know quite what to expect from the so-called spinel twins in fluorite and was surprised by that first Erongo one.

It is true that all twins shown above by various people in this thread share one property: rotation of 60 or 180 degrees (whichever you prefer) around [111]. "The twinning axis is an octahedral axis" as Dana puts it. So, if we are “lumpers”, sure, these all belong in one big family.

Now there is then a clear class to be seen in the sample of specimens shown above in which the contact between twin elements forms a continuous trace around the twin, representing the intersection of a plane perpendicular to [111] with the outside of the twin. How much deviation from a plane there is inside the crystal is not accessible to us without some dissection or perhaps tomography. The view from the outside, however, suggests a very close approximation to a plane. The examples showing this above are: spinel, sphalerite, alabandite, diamond, Dana’s drawing of a galena spinel contact-twin, a wooden model, and the single example of a fluorite crystal showing this that I could find (the cube-dodecahedron from Naica). "This is called the spinel twin because it is so common for the mineral spinel" - and numerous similar statements in the various textbooks.

The other end-member would be your wooden model and the classic English fluorite interpenetrating cube twins, which entirely lack the twin boundary perpendicular to [111]. The pseudohexagonal fluorite twins also lack this boundary between twin elements and each element penetrates completely through the other. These twins look a bit different from the interpenetrating cubes since there’s an octahedral face bounding top and bottom of the twin, shared between the two twin elements on both sides (let’s say “top and bottom”) of each twin. This was a revelation to me on seeing the first Erongo specimen and the Naica and Chumar Bakhoor examples shown above confirmed the motif (Jesse by the way clicked to this very quickly). The delightful fluorite “berry” from Colorado showed one more variation on this theme, with octahedron faces missing and the combination of cube, dodecahedron and icositetrahedron making the crystal shape tend towards a sphere.

So far we have seen two discrete classes: one with a (near-) planar boundary between the twin elements (not one of my examples illustrated above shows a deviation from a plane that is visible to the eye - can't speak for Roger's diamond); twins in the other class all lack this planar boundary and show interpenetration between the two twin elements. Presence or absence of octahedral faces makes a large difference in the appearance of this second family of penetration twins.

The remaining distinctive motif (belonging to the larger family in which "the twinning axis is an octahedral axis") in the sample of specimens shown in the thread is Dana’s penetration twin of galena (his illustration 404). One element of these twins is rotated around [111], but does not penetrate out to the octahedral faces through which the twinning axis passes. I’m guessing that large areas of the boundary between elements in such twins would be roughly perpendicular to [111], so, in this sense, these could possibly be considered transitional.