[G] Insufficiently soluble in water to form a normal solution.
EXAMINATION OF WATER.
Hardness (Hehner's process). (a) Temporary Hardness.—As has been stated ([p. 84]), this consists of lime and magnesia carbonates. As methyl orange is not affected by carbonic acid, bicarbonates of alkaline earths have an alkaline reaction, and may be estimated in solution by standard acid like the alkalies themselves. 100 c.c., or in soft waters 200 c.c., of the water is measured into a beaker, a drop or two of solution of methyl orange added, and decinormal hydrochloric or sulphuric acid run in from the burette with constant stirring till the colour just changes to pink. This is repeated, and the average taken. The two determinations should not at the most differ more than 1/10 c.c. Each c.c. represents 5 parts per 100,000 of CaCO3 or 2·8 parts of CaO; or corresponding quantities of magnesia (4·2 parts of MgCO3 or 2 parts MgO), when 100 c.c. of water are used.
(b) Permanent Hardness.—200 c.c. are measured into a beaker and boiled for 15 minutes with 40 c.c. decinormal sodic carbonate. The mixture is then allowed to cool and made up to 250 c.c.; or the flask and its contents may be weighed before boiling and made up again to the same weight. It is then filtered, and 60 c.c. representing 50 c.c. of the original water, is twice titrated with decinormal acid and the result added. If the water were pure, exactly 10 c.c. should be required to neutralise the 10 c.c. of sodic carbonate, but if there be permanent hardness a part of the sodic carbonate will be already neutralised with the acids of the lime and magnesia salts, which have been precipitated as carbonates together with the carbonates of these bases originally present in the water. The hardness will therefore be represented by the loss, i. e. the number of c.c. of acid used for 100 c.c. of the original water must be subtracted from 20 and the remainder calculated as before, or if calculated as sulphates, each c.c. represents 6·8 parts of CaSO4 or 6 parts of MgSO4 per 100,000. If, as is sometimes the case, more acid is required than is needed for the sodic carbonate used, the excess corresponds to sodic carbonate originally present in the water. In this case there can be no permanent hardness.
Chlorine in Water.—If silver nitrate be added to a solution of any chloride, the silver is precipitated as white curdy insoluble silver chloride. As indicator, a few drops of neutral potassic chromate are used. So long as any chloride is present the red silver chromate which forms is at once decomposed, and the silver converted into white chloride. But as soon as all the chloride is exhausted, the red chromate becomes permanent. To prepare a standard decinormal solution of silver, 17 grm. of pure recrystallised silver nitrate are dissolved in 1 litre of distilled water. To perform the estimation 50 c.c. of water are measured into a beaker, 2 or 3 drops of strong solution of pure yellow potassic chromate are added, and then silver nitrate from the burette till a permanent red is formed. This is repeated, and the results are added together, representing 100 c.c. of water. Each c.c. of silver nitrate used represents 3·55 parts of chlorine, or 5·85 parts of sodic chloride per 100,000. If more than 10 c.c. of silver solution are required to 50 c.c., it is advisable to use a smaller quantity of water. If the process be applied to other liquids than natural water, it must be borne in mind that the solution must not contain free acids or alkalies except carbonic acid. If this is not the case the liquid may be rendered faintly alkaline, with lime-water free from chlorides, and the excess of lime removed by passing carbonic acid through it; or it may be slightly acidified with sulphuric acid, and shaken with a little pure precipitated calcic or baric carbonate.
Detection of other Impurities.—Sulphuric acid (as sulphates) is seldom wholly absent, but its presence may be proved, by adding excess of barium chloride to the water slightly acidified with hydrochloric acid (2-3 c.c. of saturated solution of BaCl2 are sufficient for any ordinary water); if the mixture be allowed to stand overnight in a 100 c.c. cylinder beside a solution containing a known, and not very different quantity of decinormal sulphuric acid, the quantity present may be roughly compared by measuring the bulk of the precipitates.
Lime may be similarly detected and roughly measured by precipitation with excess of ammonic oxalate in presence of ammonium chloride, to hinder precipitation of magnesia. Lime-water, which may be used as a standard, contains about 128 parts of lime per 100,000.
Magnesia is detected by adding ammonium phosphate to the filtrate from the precipitated oxalate of lime. If the mixture be allowed to stand in a warm place for 24 hours all the magnesia will be precipitated as ammonio-magnesic phosphate.
Silica, &c.—100 c.c. of the water is acidified with a little HCl evaporated to dryness, moistened with HCl, and treated with a little hot water. The silica or silicic acid is left undissolved. The solution from which the silicic acid has been filtered off is evaporated to small bulk and ammonia added, when iron will be precipitated as brown ferric oxide, which is coloured black by tannin or tanning liquor. If copper be present it will give a blue solution with the ammonia. Iron may also be recognised by evaporating the water to small bulk with a trace of HCl, and adding a little sodium acetate, when if iron be present it will be coloured black by tannin, red by ammonium sulphocyanide, and blue by potassium ferrocyanide (prussiate of potash). Its quantity may be estimated (Thomson, Chem. Soc. Abstracts, May 1885) by measuring 100 c.c. of the water to be tested and 100 c.c. distilled water into two similar cylinders, adding to each 5 c.c. of dilute hydrochloric acid (1:5) and 15 c.c. of a solution of potassium sulphocyanide (40 grm. per litre), and then adding to the distilled water cylinder a very dilute standard solution of ferric salt, till its colour matches the other. If the iron contained in the water is in the ferrous condition, it must be oxidised with potassic permanganate before testing.
A suitable ferric standard solution may be made by dissolving 0·1 grm. of clean, bright, soft iron wire in a little hydrochloric acid in a long-necked flask, adding nitric acid so long as red fumes are produced, evaporating nearly to dryness, and making up to 1 litre (more accurately 996 c.c.). Each c.c. will then equal 0·0001 grm. Fe.