Water Analysis

On account of the considerable effect of certain impurities on the value of water for drinking and other purposes, the detection and estimation of these impurities is of the greatest importance.

The water is first examined visually for colour or turbidity, and by taste. Thus, for example, a green colour may be due to algae; brown, to peat or possibly to ferruginous material. Iron salts impart a distinct and bitter flavour to water, 1 part of iron per million of water being perceptible to the average person. In the case of common salt, 75 grains per gallon or approximately no parts per 100,000 are recognisable in this manner. Hard waters are supposed to have a more refreshing or palatable taste than soft. Less than 116 mg. of carbon dioxide per litre cannot be detected by the taste, whilst more than 246 mg. are distinctly perceptible at 15° to 17° C. Different persons show varying susceptibilities towards intermediate quantities. Carbon dioxide can be detected by the taste at lower concentrations in hard waters than in distilled water. In the last named, suspicion is aroused by 126 mg. per litre, but below 264 mg. of carbon dioxide, the carbonic acid is not distinctly tasted as such.

Any suspended solid, the presence of which should be regarded with suspicion, is removed by filtration. The quantity of dissolved solid is determined by evaporating a measured volume to dryness. From a practical view-point, the portion of the dissolved solids which affects the behaviour of a water towards lathering when treated with soap, is of especial interest, these constituents producing the so-called "hardness" of a water.

Useful information as to the suitability of water for various purposes may be rapidly obtained by means of a few qualitative tests. The presence of chlorides is revealed by addition of a few drops of concentrated silver nitrate solution acidified with nitric acid, when a white haze or turbidity results.

Lime gives a white turbidity on addition of saturated ammonium oxalate solution, and sulphates with barium chloride acidified with hydrochloric acid. A useful reagent for nitrites is metaphenylene diamine, 5 grams of which are dissolved in water, acidified with dilute sulphuric acid, and made up to one litre. It may be necessary to previously decolorise the solution with charcoal. If nitrites are present in the water to be tested, on addition of the diamine, a yellow colour is produced, either immediately or upon standing. Starch-iodide solution acidified with dilute sulphuric acid may also be used, the characteristic blue colour of the starch-iodine complex indicating nitrites, but this test is not altogether satisfactory.

Nitrates are readily detected by adding a few drops 0.1 per cent, brucine solution to the sample of water, and then pouring concentrated sulphuric acid to the bottom of the tube in as gentle a manner as possible; a pink and yellow zone forms at the junction of the acid and water if nitrates are present.

Nessler's solution gives a yellowish-brown coloration in the presence of ammonia. Traces of lead and copper give a dark colour with ammonium sulphide, acids being unable to destroy it. Discoloration due to copper may be removed by addition of potassium cyanide. The recognition of traces of lead by the above process, however, is sometimes impossible in the case of peaty waters, the brown colour of which entirely masks the reaction. In such cases a convenient method consists in adding permanganate to the water until it is distinctly pink. The water is then rendered alkaline with ammonia and kept for about forty-eight hours, when a precipitate will have formed, containing the whole of the lead. The supernatant liquid will be colourless unless too much permanganate has been added. The precipitate is collected on a filter, dissolved in hydrochloric acid, and, after dilution, tested with alkali sulphide in the usual manner. The composition of the precipitate has not been studied, but it appears probable that an oxide of lead is formed which is either carried down mechanically with the oxide of manganese or possibly as a compound.1 Lead is also detected by addition of a few crystals of potassium bichromate, an immediate yellow turbidity occurring in the presence of 0.1 grain of lead per gallon (0.14 parts per 100,000). On standing for half an hour a turbidity is detectable with 0.02 grain of lead.

The time-consuming operation of evaporating water in order to increase the concentration of lead which is necessary when the quantity is less than 0.1 mg. per litre, can be avoided by the procedure introduced by G. Frerichs.

When water is filtered through pure cotton-wool, any lead is completely retained by the latter. By filtering a litre or more of the water through a plug of cotton-wool, and subsequently extracting the lead from the plug by washing it with a little hot dilute acetic acid, it is possible without loss to obtain a solution in which the proportion of lead is many times as great as in the original water. The test for lead, whether qualitative or quantitative, can then be applied in the usual manner.

The amount of lead in drinking water frequently diminishes on standing, probably in consequence of precipitation as lead carbonate.

Iron gives a blue colour with a few drops of potassium ferro or ferricyanide solution acidified preferably with dilute sulphuric acid. Colorimetric tests for iron are often uncertain in the presence of copper, etc., but if suitable precautions are taken this difficulty can be overcome.

If the solution is rendered alkaline with ammonia, boiled and any precipitate removed by filtration, the presence of zinc may be demonstrated by addition of potassium ferrocyanide, when the liquid becomes turbid.

These reactions are summarised in the following table:

Qualitative tests for water

Reagent.	Result.	Conclusion.	Delicacy of Test. Parts detectable per 100,000.
Silver nitrate	White haze White turbidity	Chloride Chloride	1.5 6
Ammonium oxalate	White turbidity White ppte.	Lime Lime	9 20
Barium chloride	White turbidity	Sulphate	. . .
Metaphenylene diamine	Yellow colour	Nitrite	. . .
Brucine	Pink colour	Nitrate	0.7
Nessler's solution	Yellow - brown colour	Ammonia	. . .
Ammonium sulphide	Dark colour stable towards acid	Lead or copper	. . .
Crystal of potassium bichromate	Yellow turbidity	Lead	0.14 (immediately) 0.03 (on standing)
Potassium ferro or ferricyanide acidified	Blue colour	Iron	. . .

If the water is first concentrated to one-fiftieth of its bulk, tests may be carried out for magnesia and phosphates. The former is precipitated as magnesium ammonium phosphate on standing for some twenty-four hours after addition of sodium phosphate solution to the water rendered alkaline with ammonium hydroxide in the presence of chloride. It is assumed that any lime has previously been removed with ammonium oxalate. Phosphates are precipitated as yellow phosphomolybdate on adding excess of ammonium molybdate solution to the water acidified with nitric acid, and warming.

The correct interpretation of water analyses is largely a matter of experience, and can only be arrived at after a thorough knowledge of local conditions has been attained. The presence of chlorides is usually to be regarded with suspicion as indicating possible .contamination with sewage. But perfectly good potable waters may contain chlorides due to proximity to the sea or salt deposits. Thus the water in the lower reaches of the Severn invariably contains sodium chloride resulting from the triassic salt springs of the neighbourhood round Droitwich. Again, chlorides may result from deep wells and mineral springs, or from waste effluents from factories. Ammonia and nitrites suggest recent contamination with animal refuse. They are gradually oxidised to nitrates. Whilst, therefore, nitrated water may be quite safe, owing to oxidation of all dangerous organisms, it should be regarded with suspicion until confirmatory evidence is available.

Lead in any appreciable quantity is a very dangerous constituent in potable waters, for the poisonous action of lead compounds is cumulative, so that repeated small doses may prove more serious even than one large dose. Chronic lead poisoning may result merely from drinking such water as has been conveyed in ordinary lead pipes. Waters containing 0.02 grain per gallon are dangerous.

Iron is apt to be troublesome when present in quantities of 1 part per 100,000 and upwards. The metal oxidises, and hydrated oxide (rust) precipitates out on standing; this may block the pipes conveying the water. This oxidation is assisted by certain lowly organisms known as iron bacteria. Iron salts are not toxic, but have a certain medicinal value and impart a bitter taste to the water. Copper salts are frequently employed to remove algae, 0.3 parts per 100,000 being about the minimum effective concentration of copper sulphate for this purpose. At such dilutions the salt is not prejudicial to the human organism.

The inhibiting action of copper salts upon the growth of bacilli can be detected even at such dilutions as one part of copper in ten million of water. With 0.5 parts of copper per 100,000 a marked effect is observed; zinc has a small effect, but lead and iron appear to be without effect at these concentrations.

Chlorides present in a water are estimated by titration with a dilute standard solution of silver nitrate.

Nitrates and nitrites in water are frequently estimated together, e.g. by reduction to ammonia, which can be determined in the manner described below; alternative processes are based on the reduction of these salts to nitric oxide which may be measured volumetrically, and on reduction of the nitrate to nitrite when the total nitrite may be estimated colorimetrically by the addition of sulphanilic acid and α-naphthylamine. For the estimation of nitrites and nitrates separately, organic colorimetric methods are usually applied.

The nitrites and nitrates in natural water generally owe their existence to the oxidation of ammonia or ammonium salts which have themselves been produced by the decomposition of nitrogenous organic matter. The " free " ammonia is expelled by distilling with the addition of potassium hydroxide solution, whilst the nitrogenous organic compounds are decomposed with formation of ammonia (distinguished as albuminoid ammonia) by boiling with an alkaline solution of potassium permanganate; in both cases the amount of ammonia obtained in the distillate is measured colorimetrically with Nessler's solution. As the common source of nitrogenous organic matter is animal refuse in a state of decomposition, the presence of these nitrogen compounds is suggestive of possible contamination in a water, the period which has elapsed since contamination being reflected roughly in the relation between the amounts of nitrates and nitrites, free ammonia, and albuminoid ammonia. The total quantity of oxidisable organic matter, not necessarily nitrogenous, may be determined by direct " combustion " of the residue obtained by evaporation; it is also measured by observing the extent of reduction suffered by a known volume of standard potassium permanganate solution when kept at a definite temperature (often 26.7° C.) for several hours with a measured volume of the water in the presence of sulphuric acid (Forchhammer's process); an empirical factor is necessary for the calculation of the percentage of organic matter.

Sulphates, silica, iron, calcium, magnesium, and the alkali metals are also sometimes estimated, as are also the dissolved gases, chiefly oxygen, carbon dioxide, and nitrogen.

None of the chemical tests, however, with the exception of such as reveal the presence of poisonous substances, e.g. lead, or copper compounds, can be regarded as final evidence of the suitability or otherwise of a water for drinking purposes, and for a definite decision on this point a bacteriological examination is necessary.

For the estimation of dissolved gases in water the former are generally first removed by boiling the water or by generating carbon dioxide in it; in the latter case the bubbles of this gas carry out any dissolved gas, from which the carbon dioxide is easily removed subsequently. Carbon dioxide itself may be conveniently estimated by titrating a measured volume of the water with sodium carbonate solution until phenolphthalein becomes coloured, the method depending on the neutrality of sodium hydrogen carbonate towards phenolphthalein.