Poly Nets

Download as pps, pdf, or txt
Download as pps, pdf, or txt
You are on page 1of 23

Polyhedral Nets

A simple example of a polyhedral


network or polynet, constructed from
truncated octahedra and hexagonal
prisms.
The building process indicated produces a labyrinth. The labyrinth graph happens to be a
diamond net, with truncated octahedron centres at the nodes and the prism axes along the edges.
This particular example gives a representation of the zeolite framework FAU (faujasite). The
edges of the structure form a 4-connected net; in faujasite Si and Al atoms are centered at the
polyhedron vertices and the polyhedron edges correspond to the oxygen linkages. Note the
large voids and channels in the structure, the characteristic feature of zeolites.

An icosahedron with
octahedra on eight of its faces
can be extended to a polynet
with icosahedral nodes
centered on a bcc lattice. The
labyrinth graph is 8-connected.

In Al12Mo aluminium atoms are centered at the vertices of


this structure. Molybdenum atoms are at the centers of the
icosahedra.

Alpha manganese
Another intricate bcc structure is -Mn. The structural units are centered truncated
tetrahedra as in a Laves phase. They occur both as nodes at the bcc positions and also as
the links between the nodes. Recall that, in the regular space filling of tetrahedra and
truncated tetrahedra (equivalently: in a Laves phase) the hexagonal faces are shared by pairs
of polyhedra related by inversion. In -Mn this this does not occur. Instead, the contiguous
polyhedra are related by reflection in hexagonal or in triangular faces.

The pictures show two ways in which diamond type polynets can be built with Friaufs as
both nodes and links. In -Mn these are combined to give an 8-connected net ...

Basic polynet structure of -Mn. The


centers of the white units are at the vertices
and at the center of the cubic unit cell.

The net extended further.


The central atoms of all the polyhedra are 16-coordinated. The atoms at the centers of the nodal
units (white) are have 12 neighbours at the polyhedron vertices and 4 in the centers of
neighbouring polyhedra as in a Laves phase. The centers of link polyhedra (purple) are
coordinated to the 12 vertices and to 4 vertices of nearby polyhedra. Most of the polyhedron
vertices are 12-coordinated.
Sadoc, J. F. & Mosseri, R. Geometrical Frustration. Cambridge Univ. Press (1999).

How and why manganese atoms arrange themselves in such an intricate pattern is a
mystery. Another manganese phase, -Mn is just as intricate but quite different:

A stereo pair of images illustrating the -Mn structure. It has been identified
by Nyman, Carroll & Hyde as a packing of helical rods of tetrahedra.

Nyman, H., Carroll, C. E. & Hyde, B. G. Rectilinear rods of face-sharing tetrahedra and the structure of -Mn. Z. Kristallogr. 196 (1991) 39-46.

Polyhedral D-nets
The structure of very many crystalline materials can be described in terms of polyhedral nets.
Those with labyrinth graphs of diamond type are common.
Left: unit cell of a pair of complementary diamond type
nets (D-nets).
In polynets with D-nets as labyrinth graphs the two
complementary structures may or may not be equivalent.
Moreover, the nodes of a D-net may be of two varieties in
an alternating arrangement (as in the zinc blende ZnS).

Right: A portion of a polynet of D-type.


Inthis example the nodes and the links
connecting them are octahedra. This
structure occurs in the mineral
pyrochlore.

The structure of the silicate -crystobalite. A


D-net of vertex-connected tetrahedra. Silicon
atoms lie at tetrahedron centers, each
coordinated to four oxygen atoms at the
vertices.
The tetrahedra in the uninodal tiling of space
by tetrahedra and truncated tetrahedra form
this pattern.

The basic structure of pyrochlore can be


described as a pair of complementary
(interwoven) D-type polynets. One formed of
octahedra as both nodes and links, the other
formed of vertex-connected tetrahedra as in crystobalite. Nyman & Andersson have
identified and described several materials with
this basic structure.

A pyrochlore unit.
Four octahedra round
a central octahedron.

Nyman, H. & Andersson, S. the pyrochlore structure and its relatives. J. Solid State Chem. 26 (1978) 123-131.

A unit cell of the D-net of


octahedra in the pyrochlore
structure. Nodal octahedra
omitted for clarity (and because
in pyrochlore they lack a central
atom).

In the generalised pyrochlore geometry identified


in a variety of materials, a D-net of stella
quadrangulae replaces the D-net of tetrahedra.

The structure of W2Fe3C


Two interwoven D-nets. One of octahedra and one of vertex-connected stella quadrangulae.
(The edge length of the tetrahedra is 4/5 the edge length of the octahedra.)

Surprisingly, the generalised pyrochlore


structure just described is a D-net of -brass
clusters!
The pictures above show a gamma brass
cluster composed of four interpenetrating
icosahedra with minimal distortion, and one
with slightly more distortion in which the
hexagonal regions indicated in grey have an
exact inversion symmetry. These polyhedra
can pack together with the grey hexagons in
contact, to form a D-type polynet.
In what follows, the cluster of four
interpenetrating icosahedra will be referred
to simply as a -unit.

Another view of the -unit, emphasising its


internal stella quadrangula.

The tetrahedral cluster of five units that


constitutes the node structure of the D-net..
The complementary D-net is the D-net of
octahedra. The picture on the left shows how a
pyrochlore unit and an icosahedron belonging
to a -unit fit together.

Four pyrochlore units attached to a -unit.

The generalised pyrochlore structure can be


described as a space filling of octahedra and
-units.
This structure occurs in a large number of
complex alloys of Ti2Ni type. In Ti2Ni the
nickel atoms are at the stella quadrangulae
vertices, the rest are titanium. There are 96
atoms per cubic unit cell.

Nyman and Anderson have analysed and described the amazingly intricate
structure of Mg3Cr2Al18. The following pictures illustrate its subtle geometry.

A D-net in which the nodes are


alternately pyrochlore units and
truncated tetrahedra.

Two interwoven (complementary)


nets of pyrochlore units and truncated
tetrahedra. Observe that they do not
touch.

Nyman, H. & Andersson, S. the pyrochlore structure and its relatives. J. Solid State Chem. 26 (1978) 123-131.

Left: the two interwoven


D-nets, distinguished by
color for greater clarity.
Right: the gaps between
close pairs of pyrochlore
units are nearly exact
regular icosahedra!

Left: one of the D-nets


and the related set of
icosahedra.
Right: The
configuration of
icosahedra with the Dnets removed for
clarity. A new way of
understanding the
structure emerges...

Recall the L-unit introduced by


Kreiner & Franzen: a tetrahedron
of vertex-sharing icosahedra.
Kreiner, G. & Franzen, H. F. J. Alloys and Compounds 221 (1995) 15-36.

The structure of Mg3Cr2Al18 is


a D-net of Kreiner-Franzen Lunits linked by their vertices
like the tetrahedra in a Laves
phase.

NaZn13
The stella quadrangula or tetrahedral staris a cluster of five tetrahedra made by
placing a tetrahedron on each face of a tetrahedron. Two icosahedra can be joined
by a stella quadrangula. Below, left: a network of icosahedra , centered on a
primitive cubic lattice, linked in this way.
The complement of this polynet is an array of snub cubes:

In NaZn13 the icosahedra are 13-atom Zn clusters; the Na atoms are at the centers
of the snub cubes. A large number of intermetallics have this kind of geometry.
Hassermann, U., Svensson, C. and Lidin, S. Tetrahedral stars as flexible basis clusters in sp-bonded intermetallic frameworks and the compound
BaLi7Al6 with the NaZn13 structure. J. Am. Chem. Soc. 120 (1998) 3867-3880.

Polynets with dodecahedra


The dihedral angle of a regular dodecahedron
116.6 close to 120 . Hence three
dodecahedra sharing an edge can be slightly
deformed to bring them into face contact.
The vertex angle in a regular pentagon is 108
close to the tetrahedral coordination angle 109.4 . A
tetrahedral cluster of four face-sharing almost
regular dodecahedra is possible:

The structure can be extended to produce a D-net


a portion of which is shown on the right.
Pearce, P. Structure in Nature is a Strategy for Design. MIT press (1978).

The complementary labyrinth of the D-net of dodecahedra is a D-net of hexakaidodecahedra. This 16-faced polyhedron has 12 pentagonal faces and 4 hexagonal faces.
It is one of the typical shapes of the Voronoi cells surrounding the atoms of a FrankKasper phase. The dodecahedra and 16-hedra in this space filling arrangement
represent the Voronoi regions surrounding the atoms in a Laves phase. The 4connected network of polyhedron vertices and edges is the zeolite framework MTN.

Voronoi cells of Frank-Kasper phases


From left to right:
Dodecahedron. 12 pentagons. Maximal symmetry icosahedral.
14-hedron. 12 pentagons and two hexagons. Maximal symmetry 12m2
15-hedron. 12 pentagons and 3 hexagons. Maximal symmetry
16-hedron. 12 pentagons and 4 hexagons. Maximal symmetry tetrahedral.

A space filling of dodecahedra and 14-hedra. Rods built of


14-hedra (grey) in three mutually perpendicular directions are
packed together. The voids are dodecahedra (yellow) centered
on a bcc lattice. A unit cell is outlined in green.
The edges form a 4-connected network. The hydrogen bonds in
chlorine hydrate form such a network. (The vertices correspond
to water molecules; chlorine atoms are at centers of the 14-hedra.
This polyhedral space-filling also represents the pattern made by
the Voronoi regions of beta-tungsten (-W).

A tiling of space with 12-, 14- and 15-hedra. Columns of 14-hedra run parallel to the hexagonal
symmetry axis. Layers of 12- and 15-hedra alternate.

Williams, R. The Geometrical Foundations of Natural Structure. Dover (1979).

You might also like