Estimation of Titanium Dioxide.-Fifty cc. (equal to five-tenths gram) of the main solution are carefully neutralized with ammonia, the ammonia being added until there is a slight permanent precipitate. This precipitate is dissolved in sulphuric acid, added carefully so that there shall be but a slight excess; the volume is then made up to 350 to 400 cc. and boiled for one hour. If much iron is present, shown by the color of the main solution, it is best to deoxidize with sulphur dioxide, led into the boiling solution in which the titanium dioxide is being precipitated, but if there is only a small amount of iron present sulphurous acid may be used. In either case, however, it is well to maintain the volume of the boiling solution at about the same point, by addition of fresh portions of water, and to be sure that it smells of sulphur dioxide during the boiling. By observing these precautions the titanium dioxide is completely precipitated practically free of iron. Filter hot, through double filters, and wash thoroughly with hot water, dry, ignite, and weigh as titanium dioxide.

Estimation of Ferric Oxide.-This is best determined in the filtrate from titanium dioxide, by reduction with zinc and titration with permanganate.

The weight of TiO, + Fe,O, subtracted from the weight of Fe,0, + A1,0, + TiO, gives the A1,O,, and this is the "available alumina."

Estimation of "Free Alumina."--Two grams bauxite, ten cc. 50° B. sulphuric acid, in a four-ounce Erlenmeyer flask, provided with a perforated stopper. Heat for one hour in a waterbath at a temperature of 95° to 100° C. with frequent shaking. Add 100 cc. hot water and keep in bath for ten minutes. Filter, wash thoroughly with hot water, and for the estimation of alumina proceed as before, with this difference, however, that the titanium dioxide does not have to be determined, since none of it goes into solution. In the solution we have merely the free alumina, and a small amount of ferric oxide, which can be estimated in the usual manner and its weight subtracted from the weight of the A1,0, + Fe,O,.

Combined Alumina.-The difference between the available alumina and the free alumina is the combined alumina.

We do not at this time wish to give an opinion as to the form in which this alumina exists. It may be as clay, as a lower hydrate, or a mixture of these and other alumina compounds. We have thus far disregarded the insoluble residue, whether from the acid carried to fumes, or from the treatment at 100° C. In case it is desired to determine the total alumina, the total titanic acid, the total ferric oxide, and the silica, the insoluble residue from the treatment to fumes may be folded up in the paper, and ignited. When the paper is consumed, fuse the residue with potassium bisulphate, and allow to stand in cold water containing at least five per cent. of strong sulphuric acid several hours, or until completely decomposed. Filter off the silica, which should be perfectly white, wash thoroughly with cold water, and ignite the silica. In the filtrate the alumina, titanium dioxide and ferric oxide are determined as in the filtrate containing the available alumina. Adding these several amounts to those found in the estimation of the available alumina gives, of course, the totals.

We can not advise the determination of silica in the insoluble residue by treatment with hydrofluoric acid and sulphuric acid, as we have found that some of the titanium dioxide is also volatilized, and the loss is SiO,+TiO,.

So far as we are aware Riley was the first to point out the volatilization of titanium dioxide as fluoride. It may be that some if not all of what is volatilized goes off as hydrogen titanofluoride, as Roscoe and Schorlemmer' state, that when titanium dioxide is dissolved in hydrofluoric acid, a sirupy liquid is obtained which is probably hydrogen titanofluoride, H,TiF, Crookes,' in speaking of the estimation of titanium dioxide in silicious residues, says that on treating them with hydrofluoric acid "a titanium fluoride is formed at the same time, which can not be heated without the larger portion going off." He quotes an experiment of Riley's in which 2.235 grams of titanium dioxide were dissolved in hydrofluoric acid, evaporated and ignited, and the residue weighed only 0.99 gram, a loss of 1.245 grams or 55.70 per cent. of the original weight.

By seven successive treatments of titanium dioxide with 1 Treatise on Chemistry, 2, [II], 262.

2 Select Methods in Chemical Analysis, Second Edition, p. 193.

hydrofluoric acid we found the following losses on 0.0700 gram;

While there is a gain in time by treating the insoluble residue with hydrofluoric acid and sulphuric acid and taking the loss as silica, and while the error introduced is, perhaps, of no consequence commercially, yet the fusion with potassium bisulphate is to be preferred. But unless the silica, and the total alumina, titanium dioxide and iron are required, it is not necessary to make a fusion, the data obtained by the two separate treatments with acid being, in most cases, entirely sufficient for the valuation of bauxite.

So far as our observation goes, and we analyzed many different kinds of bauxite, titanium dioxide is always present, and in some bauxites may rise to four per cent. The clay immediately next to the bauxite may contain one per cent. of titanium dioxide, as we have found eight-tenths per cent. In other clays associated with bauxite we have found mere traces, and in the nearby soils it rarely exceeds five-tenths per cent. The concentration of the titanium dioxide in the bauxite is a very interesting phenomenon, and worthy of more extended investigation.

The titanium dioxide in bauxite is partly soluble in sulphuric acid brought to fumes. By using the Pittsburg method in bauxite acid analysis, or sulphuric acid alone to fumes, varying amounts of titanium dioxide are dissolved, as the following results show:

Total titanium dioxide. Soluble titanium dioxide.

Per cent. soluble.

82.50

93.33

62.50

The percentage of soluble titanium dioxide, calculated on the total amount present, may fall as low as fifty per cent., but for the

most part appears to be about seventy-five per cent. No titanium dioxide seems to be dissolved by the acid at 100° C. during one hour, but as the temperature rises the solubility increases, and when brought to fumes considerably more than half is dissolved. The variations in the amount dissolved may be due to variations of the form in which the titanium dioxide is present, but on this point we have no definite information. In summing up the chief points in this paper, we desire to call attention to the following:

1. The importance of discriminating in the analysis of bauxite between the trihydrate of alumina, easily soluble in sulphuric acid, and the other compounds of alumina, difficultly soluble.

2. A change in the basis of valuation, by which the trihydrate of alumina is to be given a higher value than the alumina present in some other form.

3. The adoption of a new nomenclature, calling the alumina soluble in 50° B. sulphuric acid at 100° C. during one hour "free alumina," and that soluble in sulphuric acid to fumes "available alumina," the difference between them to be known as" combined alumina."

4. The adoption of a standard method of analysis binding alike on the producer, the broker, and the consumer of bauxite for alum-making.

THE PHILLIPS TESTING LABORATORY,

BIRMINGHAM, ALA.

UPON THE SALTS OF HYDRONITRIC ACID.

By L. M. DENNIS AND C. H. BENEDICT, with Crystallographic Notes by A. C. GILL.

IN

Received January 20, 1898.

N the year 1890 Th. Curtius announced' his successful preparation of hydronitric acid (HN,), and briefly described the trinitrides of barium (BaN,), silver (AgN,), and mercury (Hg,N). He mentioned also the existence of trinitrides of copper, iron, sodium, and ammonium. Since that time no further work upon the inorganic compounds of this most interesting acid was published until last year when thallous trinitride (TIN,) and thallous thallic trinitride (TIN,.TIN,) were added to the list. Having a large amount of the acid on hand, we have entered upon the systematic study of the compounds of hydroni1 Ber. d. chem. Ges., 23, 3023.

2 Dennis and Doan: this Journal, 18, 970.

tric acid, and in this paper are given the results obtained with the elements in the principal group of Group I and the alkaline earths of Group II.

GROUP I.

Lithium Trinitride, LiN,.H,O.-This was made by neutralizing lithium hydroxide with hydronitric acid and allowing the solution to evaporate in the air. It separated in the form of colorless, glistening needles. It is very soluble in water and is hygroscopic. It is also soluble in alcohol. On being gently heated, the crystals lose their luster and water of crystallization is driven off. On further heating the salt breaks down, the decomposition being somewhat more violent than with the other compounds in this group. Upon standing, the lithium trinitride. loses some hydronitric acid.

The hydronitric acid was determined by dissolving the salt in water, precipitating with silver nitrate, and converting the silver trinitride into silver chloride.' To the filtrate and washings from the first precipitation hydrochloric acid was added to precipitate the excess of silver, the silver chloride was filtered off, and the filtrate and washings were evaporated to dryness with sulphuric acid. The lithium was then weighed as the anhydrous sulphate. Analysis gave as follows:

Sodium Trinitride, NaN,.-This salt is mentioned by Curtius but he gives no analysis. It was prepared by the same method as that used for lithium, and upon evaporation of the solution it separates in clear colorless crystals. A freshly prepared aqueous solution of the crystals is neutral in reaction, but, upon standing, it becomes alkaline. The salt does not decompose easily when heated, but on the contrary may be melted and kept in a state of fusion for some hours without breaking down.

The analysis, which was made as described under lithium, gave as follows:

1 Dennis: This Journal, 18, 950.