tolerable accuracy. The toil of making a complete estimation may be lessened somewhat in the following manner, but the speed is not much increased, and it is easier for the final precipitate to contain accidental impurities. Dissolve the phospho-molybdate in ammonia as before, and pour the solution and washings into 50 c.c. of an acetic acid solution of lead acetate. On the assumption that all the phosphorus is precipitated as lead phosphate, the composition of the white precipitate would correspond to Pb(PO). 24PbMoO,, which contains 0.644 per cent. phosphorus. But although only a portion of the phosphorus is precipitated under the above conditions, the deficiency introduces no appreciable error; in fact, the separate authors regularly use the two forms of the process respectively, and obtain results in perfect agreement. The precipitate may be ignited with the same freedom as before.

CHROMIUM.

PRELIMINARY SUMMARY.

Separation of Iron and Chromium.-To separate chromium from iron as a salt of the sesqui-oxide, the iron must be in the ferrous condition. The chromium may then be precipitated by adding barium carbonate, magnesia, zinc oxide, or sodium phosphate. The separation with barium carbonate to be complete requires the filtrate to be re-treated, so that at best the operation is a troublesome one and very little practised in spite of its frequent inclusion in text books. Chromic oxide may also be precipitated from iron solutions with nitroso-ẞ-naphthol or phenyl-hydrazine. Chromic and ferric

oxides may be separated in the dry way by volatilizing the latter in a stream of hydrochloric acid gas at 200-300° C.; or the ignited oxides may be digested with hydrochloric acid and the insoluble chromic oxide filtered off. The treatment of the metallic borings with a neutral solution of cuprammonium chloride in a carbon dioxide atmosphere leaves all the chromium as a double carbide with iron; a further separation of the residual iron has therefore to be made (mainly by fusion) before the chromium can be estimated.

Those separations wherein the iron is eliminated after converting the chromium to chromic acid are very convenient. Hillebrand has

1 Lead acetate crystals, 8 grams, and B. P. acetic acid, 400 c.c., made up to a litre with water.

observed that the characteristic green colour formed on fusing a salt of manganese with sodium carbonate may fail to appear when manganese is certainly present. We have observed that the formation of chromic acid when mixtures of ferric and chromic oxides are fused with sodium carbonate may also fail occasionally, and that the complete separation of the two oxides in this way may fail frequently.

The separation of iron from chromic acid solutions is a misunderstood process. An excess of alkaline acetate, ammonia, carbonate or bi-carbonate of sodium invariably precipitates part of the chromic acid with the iron, unless the amount of the latter is small or the excess of reagent large. An early and widely-published mode of making the separation is to oxidize the chromic oxide solution with bromine or chlorine, and to precipitate the iron with acetate. This operation may leave only a mere fraction of the chromic acid in solution; it is one of the most noticeable instances, unfortunately too common in analytical chemistry, of a serious misconception passing muster for many years in the best text books, which might have been corrected by a simple experiment on a synthetic mixture.

All separations of iron from chromic acid by means of alkalis or alkaline salts are subject to error through the formation of basic ferric chromates. The tendency of this compound to form is so decided that alkaline chromates may (in some cases with advantage) be used instead of acetates, sulphates, phosphates, etc., for precipitating neutralized ferric solutions when manganese, nickel, cobalt, zinc, or copper has to be estimated in the filtrate. For the same reason as obtains in the separation of iron and molybdenum, it is always best to make the separation by pouring the mixture into the caustic alkali or alkaline carbonate used to precipitate the iron. The following results obtained by adding alkali of binormal strength to a mixture of ferric chloride (equal to one gram of iron) and 30 c.c. of potassium chromate (12.5 grams per litre), illustrate the efficiency of some common reagents. Sodium hydrate is obviously the best.

The Gravimetric Estimation of Chromium is generally made by adding ammonia, collecting the precipitated hydrate, and igniting it to chromic oxide. The precipitated hydrate is rather troublesome to wash, but it is improved in that respect if the precipitation is made in the presence of hydroxylamine. The precipitate also retains fixed alkalis, sulphuric and phosphoric acids, magnesia and other impurities, if they are in solution; and therefore the estimation in this form is limited almost entirely to pure solutions of chromic salts. By precipitating as phosphate, or basic phosphate (3Cr2O. 2P2O5), a purer precipitate is assured. Chromic oxide may also be precipitated as basic sulphite, and, if pure, ignited directly and weighed.

To make chromium the acid of the weighed compound it may be precipitated with a solution of a lead, barium, or mercurous salt. The use of either of the two first reagents is not interfered with by the presence of calcium, magnesium, zinc, nickel, cobalt, or manganese. Mercurous nitrate cannot, of course, be used satisfactorily if ammonium salts are present.

The Volumetric Estimation almost invariably requires the chromium to exist as chromic acid. The following reagents have been recommended for effecting the oxidation in the dry way:

Sodium carbonate alone or mixed with KNO, KCIO3, K2CO3 + KClO3, KClO3+ MgO, K„CO2+ MgO, K,CO, + borax glass, CaO, CaCO3, BaO, NaHO+CaO, NaHO + MgO.

Potassium carbonate, mixed with nitrate or caustic soda.

Caustic soda, mixed with lime or magnesia.

Caustic potash alone, or mixed with nitrate, or placed in an electric. circuit.

Potassium nitrate.

Sodium peroxide, alone or mixed with barium peroxide.

Calcium chloride mixed with lime.

Soda lime, and potassium chlorate.

The following have been suggested for oxidizing the chromic oxide in the wet way:-Nitric acid and potassium chlorate, permanganate, manganese dioxide, lead peroxide, bromine, chlorine, iodine, mercuric oxide, hydrogen peroxide, sodium peroxide, and hypochlorites.

The list presents as great a variety, for working in either acid or alkaline liquids, as the analyst is likely to need.

Having converted the chromium to chromic acid, and separated from iron or not as occasion demands, it may be estimated, and most commonly is, by ferrous sulphate and permanganate or bichromate. But

stannous chloride, potassium iodide and thiosulphate, thiosulphate alone using iodide and starch as spot indicator, or iodine and arsenious oxide in bicarbonate or strong acetate solutions, have also been suggested. Some less used and apparently less convenient means of estimating chromic acid are:

1. Addition of baryta water, removal of excess with carbon dioxide, and titration of the free alkali, which is equivalent to the precipitated chromic acid.

2. Direct precipitation with barium chloride, using haematoxylin or logwood as indicator.

3. Processes depending on the addition of a reagent (e.g. a lead salt) which precipitates chromic acid, the excess used being estimated. 4. Two gasometric processes depending on the liberation of nitrogen on adding acidified hydrazine sulphate, or of oxygen on adding hydrogen peroxide, the evolved gases being measured

at N.T.P.

The only noteworthy method for titrating solutions of chromic oxide is operated by running the solution into hot alkaline standard permanganate until a clear yellow colour is observed.

ESTIMATION OF CHROMIUM.

Chromium is more easily and quickly, and for everyday purposes more accurately, estimated by a volumetric than a gravimetric process. Oxidation with permanganate in an acid solution, and titration of the chromic acid with ferrous sulphate, has been shown by experience to be a suitable form of process for estimating chromium in steel.

Chromium steels (but not ferro-chromium alloys) are decomposed by boiling with dilute sulphuric acid, and all the chromium goes into solution except what is associated with the black particles seen floating in the liquid. The addition of permanganate up to a certain point oxidizes the iron only according to the equation

10FeSO4 + 8H2SO4 + 2KMnO4 = 5Fe2(SO4)3 + K2SO4 + 2MnSO4 + 8H2O. When this reaction is completed the further addition of permanganate oxidizes the chromic oxide to chromic acid. But before this oxidation is really completed the excess of permanganate and the manganous sulphate in solution are mutually decomposed (see page 28) with formation of peroxide of manganese, and it is this precipitate that completes the oxidation in the boiling solution.

The speed of oxidation depends on the temperature, the excess of permanganate or manganese dioxide and the acidity of the solution. The change can be wrought with only a few drops of permanganate, (25 grams per litre), in excess of that theoretically needed, by boiling long enough; but with an excess of several cubic centimetres permanganate the oxidation is completed by a few minutes boiling. Increased acidity favours the oxidation; but this advantage is limited by the fact that, in strong sulphuric acid solutions, the peroxide of manganese may be partly dissolved without being completely decomposed, and is thus able to augment the measured oxidizing effect of the chromic acid. Moreover, in still stronger acid solutions, the complete conversion to chromic acid is impossible.

of

When the acid permanganate process was introduced the power the precipitated dioxide of manganese to speedily and completely oxidize the chromium appears to have been overlooked, and it was customary to add permanganate to the boiling solution until its pink colour showed that free permanganic acid was present. This misconception led to a great deal of trouble, because no permanent pink colour is obtainable until all the manganous sulphate, which is formed by oxidizing the iron, is transformed to the dioxide, and this entails a big precipitate which invariably retains chromic acid.

The precipitated manganese dioxide was responsible, according to its amount, for a deficiency in each assay of from two to twenty per cent. In 1893 Mr. Galbraith proposed to free the precipitate from chromic acid by adding an excess of sodium hydrate and titrating the filtered alkaline chromate. A serious objection to this proposal is that unless all the manganous sulphate has been precipitated by permanganate, the added soda forms manganous hydrate, which, being a strong reducing agent, destroys a portion of the chromic acid, and may thus introduce a negative error many times greater than the one it was designed to avoid. On the other hand, if more permanganate than is required by the manganous sulphate is added, some additional device for destroying the excess is necessary in order to prevent permanganic being estimated as chromic acid. Galbraith's is perhaps the least happy modification of his original process, but it brings out a useful fact, viz. that chromic acid associated with manganese dioxide can be easily separated by treatment with caustic soda.

Stead's Modification.

Mr. Stead showed that the black chromiferous particles seen in the sulphuric acid solution of the steel are decomposed during the perman