Mr. Arnot recommends the following method for alkali testing: "The sample, which should be a fair average of the drum or cask from which it is drawn, should, in the case of caustic soda, be quickly crushed into small fragments, and returned to the stoppered bottle in which it was collected for testing. It need not be finely ground, and, indeed, should not be, as it very readily attracts moisture from the air. The contents of the drum are usually pretty uniform, and the crushing recommended will give the operator a sample quite fit to work upon. Samples of soda-ash and soda crystals will, of course, be fairly representative of the casks from which they are drawn. One hundred grains of the prepared sample must be weighed out upon a watch-glass or slip of glazed paper, and transferred to a porcelain basin, with at least half a pint of boiling water. The watch-glass is preferable for caustic soda, and the weighing in the case of that agent must be done expeditiously. While the sample is dissolving the burette will be charged with the standard acid. To the soda solution a few drops of solution of litmus, sufficient to colour it distinctly, will be added. The acid will then be run into the blue soda liquor; at first, within reasonable limits, this may be done rapidly, but towards the close of the operation the acid must be added cautiously, and the solution kept well stirred. In the case of caustic, when the blue has distinctly changed to red, the operation may be considered completed, and the measures may be read off the burette; and this is, without calculation, the result required. When the soda in the sample is a carbonate, the blue colour of the litmus will be changed to pink before all the soda is neutralised, owing to a portion of the liberated carbonic acid remaining in the solution; this must be eliminated by placing the basin over a Bunsen burner and boiling the solution. The blue colour will thus be restored, and more acid must be added, repeating the boiling from time to time, until the red colour becomes permanent. It is sometimes necessary to filter the soda solution before testing; this applies specially to recovered soda, and, although in a less degree, to soda-ash." When the soda solution is filtered, it will be necessary to thoroughly wash out the liquor absorbed by the filtering paper, the washings being added to the bulk of the liquor as before. The best plan is to allow the soda solution to stand for some time until all the sediment has deposited, and then to pour off as much of the liquor as possible, and then to wash the sediment into a very small filter, in which it will receive further washing, until no trace of alkali can be detected in the last wash water.
Estimation of Chlorine in Bleaching Powder.—It is desirable that the manager or foreman of a paper-mill should have at his command some ready means by which he may test the percentage of chlorine in samples of bleaching powder, or chloride of lime, delivered at the mill, not alone to enable him to determine the proportions to be used in making up his bleaching liquors, but also to ensure his employers against possible loss in case of inferior qualities being delivered at the mill. Bleaching powders being purchased according to percentage, it is absolutely necessary that the purchaser should have this determined to his own satisfaction before either using or paying for the material. Good chloride of lime should contain 35 per cent. of available chlorine, but the powder should not be accepted which contains less than 32 per cent. There are several methods of estimating the percentage of chlorine in bleaching powder, which is composed of hypochlorite of lime, chloride of calcium, and hydrate of lime, the latter substances being of no service in the bleaching process.
According to Fresenius, in freshly prepared and perfectly normal chloride of lime, the quantities of hypochlorite of lime and chloride of calcium present stand to each other in the proportion of their equivalents. When such chloride of lime is brought into contact with dilute sulphuric acid, the whole of the chlorine it contains is liberated in the elementary form. On keeping chloride of lime, however, the proportion between hypochlorite of lime and chloride of calcium gradually changes: the former decreases, the latter increases. Hence from this cause alone, to say nothing of original difference, the commercial article is not of uniform quality, and on treatment with acid gives sometimes more, and sometimes less, chlorine. As the value of bleaching powder depends entirely upon the amount of chlorine set free on treatment with acids, chemists have devised very simple methods of determining the available amount of chlorine in any given sample, these methods having received the name of chlorimetry. The method of Fresenius is generally considered both practicable and reliable.
Fresenius' Method of preparing the solution of bleaching powder to be tested is as follows:—Carefully weigh out 10 grains of the sample, and finely triturate it in a mortar with a little cold water, gradually adding more water; next allow the liquor to settle, then pour the liquid into a litre flask, and triturate the residue again with a little water, and rinse the contents of the mortar carefully into the flask, which should then be filled with water up to the graduated mark. Now shake the milky fluid and proceed to examine it while in the turbid state; and each time, before measuring off a fresh portion, the vessel must be again shaken to prevent the material from depositing. The results obtained with the solution in its turbid condition are considered more accurate and reliable than when the clear liquid alone is treated, even though the deposit be frequently washed. This may be proved, Fresenius says, by making two separate experiments, one with the decanted clear liquor, and another with the residuary turbid mixture. In an experiment made in his own laboratory the decanted clear fluid gives 22·6 of chlorine, the residuary mixture 25·0, and the uniformly mixed turbid solution 24·5. One cubic centimètre of the solution of chloride of lime so prepared corresponds to 0·01 gramme of chloride of lime.
Gay-Lussac's Method.—This method, which is known as the arsenious acid process, has been much adopted for the determination of chlorine in bleaching powders, and is conducted as follows:—
The Test-liquor.—This is prepared by dissolving 100 grains of pure arsenious acid in about 4 ounces of pure hydrochloric acid, and the solution is to be diluted with water until, on being poured into a graduated 10,000 grains measure-glass, it occupies the volume of 700 grains measure marked on the scale. Each 1,000 grains measure of this liquid now contains 14·29 grains of arsenious acid, corresponding to 10 grains of chlorine, or 1/10 grain of chlorine for every division or degree of the scale of the chlorimeter, for which purpose a Mohr's burette of the above capacity may be used, or a graduated tube of the form shown in Fig. 85 may be employed.
Testing the Sample.—100 grains of the chloride of lime to be tested are next dissolved in water, and poured into a tube graduated up to 2,000 grains measure. The whole must be well shaken in order to obtain a uniformly turbid solution, and half of it (1,000 grains measure) transferred to a graduated chlorimeter, which is, therefore, thus filled up to 0°, or the zero of the scale, and contains exactly 50 grains of the chloride of lime under examination, whilst each degree or division of the scale contains only ½ grain. 1,000 grains measure of the arsenious acid test-liquor are now poured into a glass beaker, and a few drops of a solution of sulphate of indigo added, in order to impart a faint, but distinct, blue colour to it; the glass is then to be shaken so as to give a circular movement to the liquid, and whilst it is whirling round the chloride of lime solution from the chlorimeter is gradually and cautiously added until the blue tinge given to the arsenious acid test-liquor is destroyed, care being taken to stir the mixture well with a glass rod during the whole process, and to stop as soon as the decoloration is complete. We will assume that in order to destroy the blue colour of 1,000 grains measure of the arsenious acid test-liquor 90 divisions or degrees of the chloride of lime solution have been employed. These 90 divisions, therefore, contained the 10 grains of chlorine required to destroy the colour of the test solution; and since each division represents ½ grain of chloride of lime, 45 grains of chloride of lime (10 grains of chlorine) were present in the 90 divisions so employed, from which the percentage strength may be ascertained:—
For 45 : 10 :: 100 : 22·22.
The chloride of lime examined, therefore, contained 22¼ per cent. (nearly) of chlorine. This method is extremely simple and trustworthy when properly employed, but to ensure accuracy certain precautions must be adopted. Instead of pouring the test liquor into the solution of the sample (as in alkalimetry), the solution of the sample must be poured into the test-liquor. If the contrary plan were adopted the hydrochloric acid of the test-liquor would liberate chlorine gas so fast that much would be lost, and the result rendered incorrect. By pouring, on the contrary, the chloride of lime solution into the arsenious acid solution the chlorine is disengaged in small portions at a time, and meets with an abundance of arsenious acid to react on. The mixture of chloride of lime should also be employed turbid.
Estimation of Alumina in Alum Cake, etc.—Mr. Rowland Williams, F.C.S., in a paper read before the Chemical Society in June, 1888, describes a method of estimating the alumina in alums, alum cakes, and sulphate of alumina, by which he obtained more accurate results than are obtained by the ordinary ammonia method of estimation. After pointing out several objections to the method of precipitating the alumina by ammonia, he proceeds:—"There is another method for the estimation of alumina which is not so well known as the above. This is by means of sodium thiosulphate. Having had a very extensive and successful experience of this process, I can recommend it with confidence. Considerable practice is, however, necessary in order to secure good results, as certain conditions must be carefully attended to, otherwise the precipitation will be incomplete. The estimation is made in a moderately dilute solution. In the case of alum cake and sulphate of alumina I dissolve 400 grains in water, filter, dilute to 10,000 grains. I use 1,000 grains of this solution (equal to 40 grains of the sample) for estimating the alumina. If any free acid is present it is neutralised by a few drops of carbonate of soda solution, and the whole diluted to about 8 ounces measure. A large quantity of crystallized thiosulphate of soda is then added, and the liquid boiled for at least half-an-hour, constantly replacing the water lost by evaporation. By the end of that time all the alumina will be precipitated in a finely-divided form, along with more or less free sulphur. The precipitate is then filtered off and washed well with boiling water. The filtration and washing take place very rapidly, and may generally be accomplished in about twenty minutes, this being a great saving of time in comparison with the long and tedious washing by decantation, which is necessary in the case of gelatinous alumina. Before filtration, it is advisable to add a drop or two of carbonate of soda solution, lest the liquid should have become slightly acid during boiling."