Chemical elements
    Physical Properties
    Chemical Properties
      Silicon Tetrahydride
      Silicon Tetrafluoride
      Hydrofluosilicic Acid
      Silicon Subfluoride
      Silicon Tetrachloride
      Silicon Tetrabromide
      Silicon Tetra-iodide
      Mixed Halides of Silicon
      Halogen Derivatives of Silico-ethane
      Halogen Derivatives of Silicopropane
      Halogen Derivatives of Silicobutane
      Halogen Derivatives of Silicopentane and Silicohexane
      Silicon Oxychlorides
      Silicon Dioxide
      Silicoformic Anhydride
      Silico-oxalic Acid
      Silicomes-oxalic Acid
      Silicon Disulphide
      Silicon Monosulphide
      Silicon Oxysulphide
      Silicon Thiochloride
      Silicon Thiobromide
      Silicon Chloroitydrosulphide
      Silicon Selenide
      Silicon Tetramide
      Silicon Di-imide
      Silicon Nitrimide
      Siliconitrogen Hydride
      Silicon Nitrides
      Crystalline Silicon Monocarbide
      Silicon Dicarbide
      Silicon Carboxide
      Borides of Silicon
    PDB 1fuq-4ehr


It has been seen that silica is the anhydride of a very feeble acid, silicic acid, which is supposed to exist in the ortho- and meta-forms, Si(OH)4 and SiO(OH)2 respectively; but that the acids actually obtainable approximate rather than actually conform to these simple types, because the chemical union between silica and water is of such a feeble and complex character as apparently to lack definiteness. The silicates are, however, more definite in composition than the acid itself, though only in the case of a few organic silicates are their molecular weights known. These, however, are of the utmost importance because they alone throw any certain light upon the constitution of silicates in general.

Thus the existence of ethyl orthosilicate, whose molecular formula is Si(OC2H5)4, to a certain extent justifies the formula

for the mineral olivine whose empirical composition is 2MgO.SiO2, whilst ethyl metasilicate, SiO(OC2H5)2, represents a type to which belongs a class of natural silicates of which wollastonite, CaSiO3, is an example.

The preparation by Kipping of the organo-silicic acids SiPh2(OH)2, HO.SiPh2.O.SiPh2.OH and HO.SiPh2.O.SiPh2.O.SiPh2.OH, as well as the cyclic anhydrides:

and (Ph = C6H5),

affords valuable information as to the probable structure of the complex mineral silicates, and suggests the way in which silicic acids may condense to produce silica, which may thus be a complex mixture of different polymers. No closed chain compound has yet been obtained, however, containing in its molecule more than four atoms of silicon and four of oxygen.

Having regard to the way in which natural silicates have crystallised from molten magmas, Mendeleeff has likened their formation, whether produced naturally, as for example in granites and lavas, or artificially, as in slags and glasses, to that of alloys, from which definite compounds of the constituent metals crystallise.

The earliest classification of the natural silicates, that of Berzelius, was purely empirical, and was upon a dualistic basis. Recognising that silicates are compounds of silica with other oxides, more or less basic, Berzelius employed the oxygen ratio as a means of classification. An example will make this clear:

R2SiO4 = 2RO:SiO2 oxygen ratio is 1:1

RSiO3 = RO:SiO2 oxygen ratio is 1:2

R2Si3O8 = 2RO:3SiO2 oxygen ratio is 1:3

This method is not satisfactory, since it separates silicates that are naturally allied. For example, albite and anorthite are isomorphous felspars, but they are separated in Berzelius's system, thus:

Albite = NaAlSi3O8 or Na2O,Al2O3:6SiO2 oxygen ratio = 1:3

Anorthite = CaAl2Si2O8 or CaO,Al2O3:2SiO2 oxygen ratio = 1:1

This raises the question whether or not alumina should be regarded as basic in such silicates - a question which will be dealt with later.

A more modern method of classification consists in formulating a number of hypothetical silicic acids and arranging the natural silicates in groups according to these types. In this way a large number of silicates can be brought together under a few types.

The following are these types, with a few examples of each:


I. Orthosilicic acid Type, H4SiO4 or Si(OH)4.

Olivine – Mg2SiO4.
Willemite – Zn2SiO4.
Anorthite – CaAl2(SiO4)2.
Garnet – Ca3Al2(SiO4)3.

II. Metasilicic acid Type, H2SiO3 or SiO(OH)2.

Wollastonite – CaSiO3.
Leucite – KAl(SiO3)2.
Enstatite – MgSiO3.
Beryl – Be3Al2(SiO3)6.


III. Orthodisilicic acid Type, H6Si2O7 or (HO)3Si-O-Si(OH)3.

Barysilite – Pb3Si2O7.

IV. Metadisilicic acid Type, H2Si2O5 or HOOSi-O-SiOOH.

Titanite – CaSiTiO5.
Petalite – LiAl(Si2O5)2.


V. Orthotrisilicic acid Type, H8Si3O10 or (HO)3Si-O-Si(OH)2-O-Si(OH)3.

Melilith – Ca4Si3O10.

VI. Trisilicic acid Type, H4Si3O8 or .

Orthoclase – KAlSi3O8.
Albite – NaAlSi3O8.

A large number of natural silicates can be included in the above classes, and so far the system seems satisfactory. There are, however, some important silicates that appear not to conform to any of these types, prominent among which are cyanite and andalusite, both of which possess the composition Al2SiO5, and kaolinite, or china clay, which is H4Al2Si2O9, to which serpentine, H4Mg3Si2O9, corresponds.

It is usual, however, to consider cyanite a basic metasilicate and andalusite a basic orthosilicate, thus:

Cyanite – (AlO)2Si03 or

Andalusite – Al(AlO)SiO4 or

whilst kaolinite and serpentine are hydrated forms of orthodisilicates, Al2Si2O7.2H2O and Mg3Si2O7.2H2O respectively. Al2SiO5 has been prepared artificially by Shepherd, Rankin, and Wright, but it is identical with the mineral sillimanite, into which andalusite and cyanite are slowly transformed when they are heated above 1300° C.

The micas are, moreover, an important class of natural silicates which appear not to come within any of the above categories. They are regarded by F. W. Clarke as derivatives, or isomorphous mixtures of derivatives, of the aluminium silicates Al4(SiO4)3 and Al4(Si3O8)3, and therefore as belonging to the orthosilicic or trisilicic type. Muscovite, for example, which occurs in granite, is H2KAl3(SiO4)3. Tschermak, however, supposes the micas to consist of isomorphous mixtures of three types of molecules, symbolised respectively by K, M, and S, thus:

K = H2KAl3(SiO4)3 (muscovite)

M = Mg6(SiO4)3 (olivine)

S = H4Si5O12.

It will be observed that S introduces an additional type of silicic-acid molecule.

These considerations suffice to show that the classification of the natural silicates is a matter of great complexity, and that to refer them to a few types of hypothetical silicic acids constructed a priori on theoretical grounds is not to account for them completely. Especially is it to be noted that this chemical classification is not the kind of classification which would be evolved by a mineralogist who took account of natural relationships. It has already been pointed out that classification by oxygen ratio separates albite and anorthite, which are naturally allied; chemical classification according to types of silicic acid also separates them. Clarke, by classifying natural silicates as derivatives of Al4(SiO4)3, gives the following formulae:

Lime garnet: Al2Ca3(SiO4)3
Anorthite: Al6Ca3(SiO4)6
Albite: Al3Na3(Si3O8)3;

but this arrangement brings together garnet and anorthite, which are dissimilar, and still separates anorthite and albite, which are similar in physical properties.

According to the system of Goldschmidt, silicates may be derived from polymerised silica molecules such as Si6O12 or S8O16 by the replacement of silicon atoms by equivalent metallic or other atoms. Thus these three silicates are formulated:

Lime garnet:Si3Ca2(CaAl2)O12
[Si3Si (Si2)O12]

and the system has the merit of classifying anorthite and albite together and separating them from garnet.

The natural silicates may now be considered from another point of view. It will have been observed that a large number of them contain alumina, an amphoteric oxide possessing distinctly acidic properties which are. shown in the existence of aluminates, of which the spinels, e.g. MgAl2O4, are natural examples. Thus alumina approaches silica in chemical properties, with which indeed, in the crystalline form, it is isomorphous. Consequently it cannot be accurate to regard alumina as entirely a basic oxide in the aluminous silicates; and so several other points of view present themselves, which can be stated in the form of three distinct theories:

(i) The aluminous silicates are of the nature of double salts analogous to the alums, aluminium silicate and a silicate of a stronger base than alumina being united together by a kind of residual affinity whose strength depends upon the difference in basic power between alumina and the other metallic oxide, just as the union of sulphates in the alums depends upon such a difference.

(ii) The aluminous silicates are isomorphous mixtures of silicates and aluminates.

(iii) The aluminous silicates are salts of strongly basic radicles with complex silico-aluminic acids, which are analogous to the complex acids formed by molybdenum and tungsten - such as, for example, the silicomolybdate

It is doubtful, however, whether alumina forms complex acids analogous to those of the trioxides of the sixth group; rather will it take the place of silica itself in such a compound as the above. Nevertheless various attempts have been made to express the constitution of the aluminous silicates on the assumption that they are individual compounds and not isomorphous mixtures. Orthoclase felspar, perhaps the most important of these compounds, will serve as an example. Tschermak writes the formula for orthoclase thus:


and constitutionally:

whilst P. Groth, adopting the simplest empirical formula, represents the silicic acid to be entirely in the meta-form, thus:

Clarke, in accordance with his theory that aluminous silicates are derivatives of aluminium silicates, formulates felspar thus:

and Vernadsky, considering this silicate to be the potassium salt of the complex silico-aluminic acid H2O.Al2O3.6SiO2, develops its structural formula thus:

Clay, moreover, is believed by Loew to be an anhydrosilico-aluminic acid, thus:

Al2Si2O7.2H2O =

The subject is carried much further in the hexite-pentite theory of W. and D. Asch.

According to this theory silica and alumina form closed chains containing 6 or 5 atoms of silicon or aluminium alternated with oxygen atoms. The hydrated forms of silica and alumina hexite are:

to which there correspond the anhydrides:


The hydrohexites are represented more briefly thus:

or or

and or or

Similarly the silica and alumina pentites are, in their abbreviated form:


whilst the hexite and pentite symbols in their briefest form are:

and respectively.

The fundamental assumption of the theory of W. and D. Asch is that all the natural silicates contain ten- or twelve-membered rings of alternate silicon or aluminium and oxygen atoms, and by means of this theory a very large number of silicates can be formulated. This may be shown by a few examples (Ho = HO):

8H2O.6Al2O3.12SiO2 =

10H2O.6Al2O3.12SiO2 =

8H2O.6Al2O3.10SiO2 =

12H2O.3Al2O3.18SiO2 = =

Kaolin = 6H2O.6Al2O3.12SiO2.6H2O =

Albite = 3H2O.3Al2O3.18SiO2 =

Muscovite = 4H2O.2K2O.6Al2O3.12SiO2 =

Apart from theoretical views upon the constitution of the natural silicates, these minerals may be practically divided into two great classes: anhydrosilicates and hydro silicates or zeolites respectively. The former silicates contain no combined water, and are decomposed only by prolonged heating with concentrated sulphuric acid, or, better, by fusion with excess of alkali carbonate. Hydrosilicates intumesce when heated before the blowpipe, with loss of water, and are decomposed by concentrated hydrochloric acid with separation of gelatinous silica. Their hardness (3.5 to 5) is considerably less than that of the anhydrosilicates (4.5 to 7), and this, together with their greater reactivity, shows them to have a molecular texture less dense than that of the anhydrosilicates.

Many examples of anhydrosilicates have already been given; a few typical hydrosilicates are:

Natrolite – H4Na2Al2Si3O12 – rhombic
Chabazite – H8CaAl2Si4O16.2H2O – rhombohedral
Phillipsite – H2CaAl2Si5O15.4H2O – monoclinic
Stilbite – H4CaAl2Si6O18.3H2O – monoclinic.

The constitution of numerous natural silicates has been investigated by Tschermak in the following manner: The powdered silicate was decomposed by prolonged digestion with concentrated hydrochloric acid at a temperature approaching 60° C.; the liberated silicic acid was thoroughly washed, and then allowed to dry slowly in the air. The rate at which loss of weight proceeded was found to change at a certain point which was believed to indicate transition from aqueous evaporation to dehydration; and the composition of the silicic acid at this point was supposed to correspond to that of the mineral which was a salt of the acid. Thus, for example, whilst the hydrolysis of silicon tetrachloride yielded orthosilicic acid, H4SiO4, of density 1.576, the same acid was obtained from natrolite, Na2H4Al2Si3O12, and from dioptase, H2CuSiO4, which are consequently orthosilicates. Hemimorphite, H2ZnSiO5, also, in spite of its empirical formula, is shown to be an orthosilicate by this method; whilst anorthite, CaAl2Si2O8, yields metasilicic acid, H2SiO3 (density 1.813), and leucite, KAlSi2O6, H4Si2O6, or "leucite" acid, a polymer of metasilicic acid, of density 1.809. "Garnet" acid, H4Si3O8, has the density 1.910 and "albite" acid, H2Si3O7, the density 2.043. The different acids are distinguished by their behaviour towards methylene blue. The conclusions of Tschermak have been adversely criticised by Mugge.

The preparation of artificial silicates, corresponding to the natural minerals, by fusing together their constituents is a method of study which has been pursued by Vogt, and by Doelter and his pupils.

Vogt applies the principles of physical chemistry to the crystallisation of silicates from fused magmas, and especially the conclusions of Roozeboom concerning the solidifying-points of mixed crystals. Consequently he believes that the identity of the silicates is preserved in the fused state, and that their molecular weights may be determined from the depression of the freezing-points of the molten magmas. Doelter, on the other hand, who has determined the melting-points of a large number of silicates, regards them as highly dissociated in the liquid state, and dissents from some of Vogt's conclusions.

A number of "hydrothermal" silicates have been prepared by Baur and Becke by heating amorphous silica, alumina, and lime with sodium or potassium hydroxide to 350°-450° C. in a steel cylinder closed with a steel screw. The following are among the "minerals" so obtained: quartz, opal, orthoclase, albite, oligoclase, andalusite, muscovite.

Few definitely crystallised artificial silicates exist, and these are confined to the alkali metals.

Sodium Metasilicate

Sodium Metasilicate may be obtained by inoculating its strongly alkaline solution with the solid salt, and when recrystallised from warm 2-3 per cent, solution of sodium hydroxide yields rhombic crystals having the composition Na2SiO3,9H2O. These crystals melt at 48° C., and slowly lose six molecules of water when kept over sulphuric acid. The trihydrate Na2SiO3.3H2O thus obtained yields the hexa- hydrate Na2SiO3.6H2O when allowed to stand over 25 per cent, sodium hydroxide solution. High temperature is necessary for the complete dehydration of the salt.

The aqueous solution of this salt is strongly alkaline, and is believed to have been hydrolysed according to the reaction:

2Na2SiO3 + H2ONa2S2O5 + 2NaOH.

A dilute aqueous solution of sodium silicate may be titrated with decinormal acid as if it were sodium hydroxide, when methyl orange is used as indicator.

In commercial sodium silicate the ratio by weight of silica to soda may be as high as 3 to 1. When the proportion of silica is large the silicate dissolves only very imperfectly in much water.
© Copyright 2008-2012 by