During solution of the "melt" and evaporation (which may be carried on together), a clear solution will not be obtained, a flocculent silica will separate out, and towards the end of the evaporation the mass will get gelatinous. The drying of the jelly must be finished on the water-bath; first, because at this temperature the silica is rendered insoluble in hydrochloric acid, whilst the solubility of the alumina, iron, &c., is unaffected, which would not be the case at a much higher temperature; and second, because the gelatinous residue requires very cautious drying to prevent loss from spirting.
When dry, the substance is moistened, and heated with strong hydrochloric acid, and the sides of the dish are washed down with water. The silica is washed by decantation two or three times with hydrochloric acid and hot water, before being thrown on to the filter. The filtrate is again evaporated to dryness, taken up with a little hydrochloric acid and water and again filtered. The residue on the filter is silica. The two lots of silica are washed free from chlorides with hot water, dried on an air-bath, transferred to a platinum-crucible, ignited gently at first, at last strongly over the blast or in a muffle, cooled in a desiccator, and weighed.
The white powdery precipitate is silica (SiO2), and its weight, multiplied by 100, and divided by the weight of ore taken, gives the percentage of silica in the sample. Where the percentage of silicon is wanted, which is very rarely the case, it is got by multiplying this result by 0.4667. It is always necessary to examine the purity of the body weighed as silica. This is done by re-fusing the material weighed, and re-determining the silica in it; or, better, by mixing a weighed portion in a platinum-dish with a little strong sulphuric acid, covering with hydrofluoric acid, and evaporating. In the latter case, the silica will be converted into fluoride, which will be driven off, and the impurities will be left behind as sulphates of barium, phosphate and oxide of tin, titanium, &c. This must be weighed and deducted from the weight of the silica. In a complete examination of a silicate it should be treated with the precipitate containing alumina, ferric oxide, &c.
EXAMINATION OF SILICATES.
The student interested in the analysis of rocks and rock-forming minerals is advised to consult a valuable paper by Dr. W.F. Hillebrand in the Bulletin of the United States Geological Survey, No. 148, to which I am very largely indebted in the revision of the following pages.
Moisture.—Five grams of the powdered sample is dried between watch-glasses in the water-oven for two hours, or till its weight is constant; and the loss is reported as water lost at 100° C. The rest of the determinations are made on this dried mineral.
Combined Water, &c.—Weigh up 1 gram of the substance, and ignite over the blowpipe for some time in a platinum-crucible, cool in a desiccator, and weigh. Record the loss as "loss on ignition," not as "combined water."
Silica.—The ignition should have been performed in an oxidising atmosphere in a muffle or over a slanting blowpipe flame; this will ensure the oxidation of any pyrites or other sulphide present, which if unoxidised would injure the crucible in the next operation. The ignited residue is mixed with 6 or 7 grams of anhydrous sodium carbonate. This reagent should be the purest obtainable, but its purity should be checked, or rather its impurities should be determined by running a "check" or "blank" assay with 10 grams of it through the stages of the analysis; the impurities will be chiefly silica, alumina and lime, and altogether they ought not to exceed 1 milligram. The crucible with the mixture is heated at first gently over a Bunsen and afterwards more strongly in an oxidising atmosphere in a muffle or over the blowpipe. The fused mass is allowed to cool in the crucible, and is then dissolved out in a basin with water and a small excess of hydrochloric acid. After the removal and cleaning of the crucible, the liquor is evaporated almost to dryness. Dr. Hillebrand advises stopping short of complete dryness. The residue is taken up with a little hydrochloric acid and water and filtered and washed. The liquor, including the washings, is again evaporated and taken up with water and a little acid. Usually about 1 per cent. of silica will be thus recovered. It is to be filtered off and washed and added to the main silica. The filtrate is reserved. The silica, thoroughly washed, is dried and ignited at a high temperature for twenty or thirty minutes. It is then weighed in a platinum crucible. After weighing it is treated with hydrofluoric acid and a little sulphuric, carefully evaporated and ignited strongly. The residue, which in extreme cases may amount to 2 or 3 per cent. of the rock, is weighed and deducted from the weight of the impure silica. It is retained in the crucible.
Alumina, &c.—The filtrate from silica is treated by the basic acetate method. That is, it is first treated by a cautious addition of a solution of soda, almost to the point of producing a precipitate, in order to neutralise the excess of acid; 2 or 3 grams of sodium acetate are added, and the whole boiled for a minute or so. The precipitate is filtered off and washed only slightly. Save the filtrate. The precipitate is dissolved in hydrochloric, or, perhaps better, in nitric acid; and is reprecipitated by adding an excess of ammonia and boiling. The precipitate is filtered and washed with water containing 2 per cent. of ammonium nitrate. Both filtrates are evaporated separately to a small bulk, a drop or two of ammonia being added to the second towards the finish. They are next filtered into a 6 or 8-ounce flask through a small filter, the second filtrate coming after, and serving in a manner as wash water for the first[113]. The two washed alumina precipitates are dried and placed in the platinum crucible containing the residue from silica after treatment with hydrofluoric acid. They are then ignited in an oxidising atmosphere at a high temperature for about 10 minutes. The weight, including that of the residue from the silica, is noted as that of "alumina, &c."
The weighed oxides are next fused with bisulphate of potash for some hours. The bisulphate should have been first fused, apart, until the effervescence from the escape of steam has stopped. The melt is dissolved out with cold water and dilute sulphuric acid, and any insoluble residue is filtered off, washed, ignited and weighed. The filtrate is reserved for determinations of iron and titanium. The residue, after weighing, may be treated with hydrofluoric and sulphuric acids for any silica,[114] which would be determined by loss. It may be tested for barium sulphate by treatment with hot strong sulphuric acid; in which this salt dissolves, but is again insoluble (and so comes out as a white precipitate) on diluting with cold water; the acid also must be cold before adding the water. The filtrate containing the iron is reduced with sulphuretted hydrogen, boiled till free from that gas, filtered and titrated with a standard solution of permanganate of potassium. The iron found is calculated to ferric oxide by dividing by .7. The iron solution after titration serves for the determination of titanium oxide (TiO2). This is done colorimetrically, by adding peroxide of hydrogen free from hydrofluoric acid, and comparing the brown colour produced with that produced by the addition of a standard solution of titanium to an equal volume of water containing sulphuric acid.[115] The alumina is determined by difference. From the weight of the combined precipitate which has been recorded as "Alumina, &c.," deduct (1) the residue, insoluble, after fusion with bisulphate; (2) the ferric oxide; (3) the titanium oxide; and (4) the phosphoric oxide (P2O5), the amount of which is subsequently determined in a separate portion. This gives the alumina.