OD CHARACTER OF KERMESITE (Sb2S2O) STRUCTURE
S. Durovic
Institute of Inorganic Chemistry, Slovak
Academy of Sciences, SK-842 36 Bratislava, Slovakia
e-mail: uachduro@savba.sk
Keywords: kermesite, OD structures, polytypism, crystal structure
The crystal structure of kermesite Sb2S2O has been solved by Kupcik in 1967 [1] from the intensity data obtained photometrically from Weissenberg films. Although the structure is triclinic, Kupcik used for its description an F centred cell with monoclinic geometry (a = 20.97(3), b = 8.16(1), c = 20.38(3) A, b = 101.8(3)o) and with two additional lattice points at 1/4 and 3/4 along the a b diagonal, thus an 8-fold cell, space group F1_, R = 0.096. The structure has been refined twenty years later by Bonazzi, Menchetti and Sabelli [2] using single-crystal diffractometer data. They have chosen a primitive triclinic cell (a = 8.147(1), b = 10.709(1), c = 5.785(1) A, a = 102.78(2), b= 110.63(2), g= 101.00(1)o), space group P1_, R = 0.057, related to the Kupcik's cell by the matrix 104/1_00/1_ 2_ 2_. The refinement revealed some minor errors in the Kupcik's model, yielded considerably better structural parameters/interatomic distances, and, on the whole, confirmed its correctness. Bonazzi et al. were also first to find twinning in this mineral.
The unusual extinctions given by Kupcik: hkl present only for h+k=0 (mod,4) ^ k+l = 0 (mod 2) are due to the non-conventional choice of his cell. However, a closer look at his atomic coordinates revealed - in a good approximation - partial mirror planes at y =0 and 1/2. Accordingly, the structure can be considered as consisting of layers parallel (100), with symmetry A(1) 2/m 1, stacked so that any layer is shifted relative to the preceding one by +b/4. It is thus an OD structure [3] of equivalent layers with the OD groupoid family symbol
A (1) 2/m1
{ (1) (21/2 /a2) 1 }
belonging to the category Ia. From the NFZ relation (N=4, F=2 => Z=2) it follows that for any layer there are two translationally non-equivalent positions which lead to geometrically equivalent pairs of adjacent layers. These can be achieved by a shift by +b /4 or -b/4 of a layer relative to the preceding one. Since the layers are all translationally equivalent, a Hägg symbol can be used to describe the stacking sequence in any kermesite polytype, as a sequence of signs + and/or -.
The family structure is identical with a two-fold superposition structure
r^ (xyz) = (1/2) [r(xyz) + r (x,y+1/2,z)]
a^ = a/2, b^ = b/2, c^ = c/2 (a, b , c- Kupcik's basis vectors) and space group C(1)(2/m)1. The set of family diffractions (k= 0 (mod 2), etc. has thus a monoclinic symmetry.
There are two MDO polytypes in this family: MDO1 found by Kupcik [1], with Hägg symbol |+|, to which also a twin structure |-| as reported by Bonazzi et al. [2] exists. The MDO2 with a2=a/2, b2=b, c2=c with Hägg symbol |+-| and space group A(1)(2/a)1 has not been found as yet. The probable reason is the desymmetrization which may prefer the MDO1 arrangement.