Since Si invariably occurs in tetrahedral coordination the fundamental unit of the silicate structure is the Si-O tetrahedra. The different types of silicate structure arise from the ways in which these tetrahedra are arranged: they may exist as seperate unlinked entitites, as linked finite arrays, as infinite 1-dimensional chains, as infinite 2-dimensional sheets or as infinite 3-dimensional frameworks. These possibilities give rise to the six primary structural types of silicates, each with a characteristic Si : O ratio (more strictly this should be the ratio of tetrahedral cations to oxygen, the reasons for which are discusssed later):
Orthosilicates (or neosilicates): independent Si-O tetrahedra, Si : 0 = 1 : 4, for example the olivine group.
Sorosilicates : two linked Si-O tetrahedra sharing one oxygen, Si : 0 = 2 : 7
Cyclosilicates : closed rings of linked Si-O tetrahedra sharing two oxygens, Si : 0 = 1 : 3, for example beryl.
Inosilicates : continuous chains of Si-O tetrahedra, sharing two oxygens (single chains, Si : 0 = 1 : 3, for example the pyroxene group) or alternately sharing two and three oxygens (double chains, Si : 0 = 4 : 11, for example the amphibole group)
Phyllosilicates: Continuous sheets of Si-O tetrahedra sharing three oxygens, Si : O = 2 : 5, for example the mica group.
Tektosilicates: Continuous framework of Si-O tetrahedra sharing all four oxygens, Si : O = 1 : 2, for example the feldspar group.
The requirement for charge balance or electronic neutrality in these different structural types is maintained by the dispersal of other cations in 6-fold (octahedral), 8-fold (cubic) or 12-fold (icosahedral or close packed) coordintaion between the individual tetrahedra or arrays of tetrahedra in the silicate structure. For example, in single chained inosilicates (Si : O = 1 : 3) there is a net excess of two negative charges per tetrahedra.
where n is the number of tetrahedra in the chain, and 2n- represents the charge excess of the chain forming elements. Theoretically the charge excess could be alleviated in a number of different ways, for example by adding one bivalent cation or two univalent cations per 3 oxygens. However, the location of these cations, which must reside in spaces (termed sites) between the individual chains of Si-O tetrahedra, must be such that they simultaneously satisfy the requirement for electronic neutrality of all oxygens in the structure. Fortunately, this constraint severely limits the range of compositions and structures found in the inosilicates (as indeed it does with all other silicates).
If we look at the detail of the single chain structure we find that in each tetrahdron there are two linking O atoms that are each bonded to two Si atoms and two peripheral O atoms each with bonds to only one Si atom. Since each Si atom shares its 4+ charge with the surrounding four oxygens of the tetrahedron, the requirement for the electronic neutrality for each of the two linking O atoms is completely satisfied. In contrast, each of the peripheral O atoms have a net excess of one negative charge. In order to satisfy this each of these oxygens can be bonded with 3 neighbouring bivalent cations in octahedral coordination (as shown above) or with four bivalent cations in 8-fold coordination. In the pyroxene group both possibilities occur, each placing profound constraints on the way in which the adjacent chains are located with respect to each other.
In double chain inosilicates (Si : 0 = 4 : 11) the chain forming elements give rise to the basic formula:
where the net excess of charges per 11 oxygens is equivalent to 6 negative charges. The sheet forming elements in phyllosilicates (Si : 0 = 2 : 5) give rise to the basic formula: