A. K. Kal’e.

A New Conductometric Method for the Determination of Magnesium in the Presence of Alkalis

(Laboratory of Quantitative Analysis, TII).

Gravimetric methods for determining magnesium in the form of ammonium magnesium phosphate and weighing it as magnesium pyrophosphate, as is known, have always been rather lengthy processes; moreover, the accuracy of the magnesium determination itself, judging from the work of Majdel (Zeit. an. Chemie, 82, 1930, p. 425), is comparatively low and depends on the content of ammonium salt. Majdel took a definite quantity of magnesium salt (theoretically corresponding to 0.3195 g of magnesium pyrophosphate) and, working by Schmitz’s method, determined what quantities of magnesium pyrophosphate were obtained in practice at varying contents of ammonium salt; he succeeded in establishing precisely that, at a low content of ammonium chloride, the quantity of magnesium pyrophosphate obtained is 3% below the theoretical value, and that the accuracy of the determination increases with an increased quantity of ammonium chloride, not exceeding, however, 0.22% in the best case.

Volumetric methods for determining magnesium in this respect likewise offer few advantages (see, for example, Rauch, Zeit. an. Ch. 84; 1931, p. 336).

It would be far more convenient and simpler to use a method of conductometric titration. However, the old method of Duboux (Compr. rend. de l’Acad. de Sc 159,320, 1914), based on conductometric titration of a magnesium salt with phosphoric acid in the presence of ammonia and ammonium chloride, offers no advantages in accuracy compared with the gravimetric method; moreover, when this method is used, acid phosphate salts are introduced, which hinder the subsequent determination of alkalis.

The new method of Harned (Jour. Amer. Chem. Soc. 39,252, 1917, or Zeit. an. Ch. 75, 1928, p. 160) does indeed give excellent results, but it is applicable only in the absence of alkalis; this method is based on conductometric titration of magnesium sulfate by means of barium hydroxide hydrate.

I carried out a series of experiments, one of which I give below, with the aim of establishing the possibility of determining the quantity of magnesium in an aqueous solution in the presence of alkalis; however, in all cases smooth curves were obtained which did not give a sufficiently sharp break point that could indicate the end of the titration.

For the experiments a solution of barium hydroxide hydrate was taken, one cubic centimeter of which corresponded to 5.81 mg of anhydrous magnesium sulfate. For the experiment described below, 50 mg of magnesium sulfate and 19.5 mg of sodium sulfate were taken (calculated as anhydrous salts).

In this experiment, as in the subsequent ones, the form of vessel found practically convenient for titration was that whose construction is shown in Fig. 1.

Into this vessel (of about 100 cubic cm capacity) two platinum electrodes are sealed (platinized before the experiment); the outlet tube serves for stirring the solution placed in the vessel by means of air freed from carbon dioxide by prior passage through a concentrated solution of sodium hydroxide.

Every minute, from a burette correspondingly protected from carbon dioxide, one cubic cm of barium hydroxide solution was added, and every half minute, by Wheatstone’s bridge method, the minimum of sound was observed by means of an ordinary telephone. The resistance of the box was taken as constant, equal to 1000 ohms.

For the above-described method of conductometric titration in aqueous solution, the following data for the sound minimum were obtained:

Plotting these data on a diagram, we obtain the curve of Diagram 1. As is seen from the diagram, no sharp minimum can be discerned on the curve. Instead of the theoretically expected point of inflection for the given amount of magnesium sulfate (30 mg) at 8.6 cm³, we do not obtain it even for 9 cm³ of added barium hydroxide solution. Obviously, Harned’s method is not applicable for the determination (and separation) of magnesium from alkalis. The point of inflection is not obtained at the moment of complete precipitation of magnesium, of course, because the precipitation of the well-dissociated sodium (or potassium) sulfate by barium hydroxide continues.

Diagram 1

The picture of the graphs changed completely when I added to the sulfate solution an equal volume of ethyl alcohol (96%); in this case, as is seen from the following example, a curve with a sharp break was obtained at the moment of complete precipitation of magnesium hydroxide; the presence of the most varied amounts of alkali sulfates did not interfere with the accuracy of the determination.

Below are given the data and the graph of one of the titration experiments in a water-alcohol solution of 80 mg of magnesium sulfate and 19.5 mg of sodium sulfate (always calculated as anhydrous salts):

Diagram 2

On the basis of these data the graph in Diagram 2 was constructed.

As a result, we obtain an inflection point at the point corresponding to 13.8 cu. cm of Ba(OH)₂, which corresponds to 80.2 mg of magnesium sulfate, i.e., the error is +0.25%.

Subsequently, without presenting graphs, I give the numerical material and the results of three more conductometric titrations with potassium sulfate and in a mixture of potassium and sodium sulfates.

Without citing a number of other data, it may be said that the method described makes it possible to determine the amount of magnesium with an accuracy of up to ±1.0%. The accuracy of the determination can be considerably increased by taking a more dilute solution of barium hydroxide hydrate, and also when there is a higher content of magnesium salt in the sample (up to several tenths of a percent).

The method developed is of especially great importance for the analysis of drinking water, if one makes use of Kolthoff’s method (Chem. Centr. 1918, I, 654), which consists, in general outline, in evaporating the dry residue with sulfuric acid, igniting the residue to constant weight (using ammonium carbonate), and then treating it with 50% ethyl alcohol; in this process magnesium and alkali sulfates are obtained in the filtrate. The filtrate thus obtained, according to the method developed above by me, can then be used directly for determining the amount of magnesium. Parallel analyses of various natural waters, carried out by the usual method and by my method, gave discrepancies only in tenths of a milligram.

Conclusions:

1. A method has been developed and tested for the conductometric determination of magnesium in the presence of alkalis.

2. The inapplicability of Harned’s method for the conductometric determination of magnesium in aqueous solution in the presence of alkalis has been established.

3. The method developed accelerates the quantitative determination of magnesium tens of times over, as compared with the gravimetric determination methods used up to now, without being inferior to the latter in accuracy.

4. The most convenient and rapid route for determining the amount of magnesium in water is indicated.

Work is continuing on the application of the developed method to the determination of magnesium and alkalis in silicates.

Tomsk
1/VI—1935.