A S A L Y TICAL EDI T I O X
38
It will be noted that the agreement between the figures for the tungsten steel sample 1 using the old and new methods is satisfactory. The results for illium using the old method are low, indicating incomplete burning of nonferrous alloys of this type unless the induction furnace method is applied. The analysis of the blank suggests the determination of
Vol. 2, s o . 1
carbon in sheet steel without the use of drillings as sample and indicates the comparative accuracy of the two methods. Literature Cited (1) Marr,Iron SteelInd , a, 184 (1929).
Determination of Cobalt in Driers, Japans, and Alloys’ Oscar Heim 119-01 NEW
YORK
BLVD.,J A M A I C A , LOSC
T
HE method here described is generally useful and
time-saving in determining cobalt in the presence of a large number of other metals, as is often necessary with the inorganic constituents of varnishes and japans, especially those that have been extracted by organic solvents from the pigment. Sometimes appreciable quantities of metals are extracted in the form of their soaps which originally belonged t o the pigment. The results obtained by the nitroso-/?-naphthol method are as unsatisfactory as those obtained by Carnot’s method (precipitation by means of ammonium molybdate as ammonium cobaltic molybdate). T o isolate cobalt quantitatively, the writer has macle quantitative the well-known qualitative cobalt test devised by T’ogel ( 2 ) . The solution used contained cobalt, nickel, iron, aluminum, chromium, manganese, zinc, tin, lead, copper, titanium, and vanadium. Procedure
HYDROCHLORIC ACID SOLUTIOXOF hIETaLs-Treat the ash of the japan with hydrochloric acid, preferably by oxidizing 10 grams of the material, in a tall beaker or an Erlenmeyer, with about 40 cc. of sulfuric acid (concentrated) and 20 cc. of hydrogen peroxide (30 per cent). After the violent reaction has ceased drive off the excess of water by boiling until white sulfuric acid fumes begin to form. Cool somewhat, again add 20 cc. of hydrogen peroxide, and treat as before. When the dark liquid has finally become light, indicating the absence of organic matter, drive off the bulk of the sulfuric acid (hood). Cool, dilute, and add ammonium hydroxide and then hydrochloric acid, each in slight excess. If metals other than cobalt are to be determined also, make u p the solution in a volumetric flask. To the dilute hydrochloric acid solution a t 50” C. add pure zinc oxide in very small quantities, until only a trace of the oxide remains undissolved. If no iron is present, add a few drops of 10 per cent ferric chloride solution. (Even a small excess of zinc oxide will precipitate a little cobalt.) DETERMINATION OF COBALT-Filter the precipitate and wash it with warm water. (It may contain iron, aluminum, chromium, copper, vanadium, titanium, and some lead.) If properly treated with zinc oxide the precipitate should give a negative “Vogel” reaction for cobalt, which is sensitive t o 0.02 mg. The filtrate may contain cobalt, nickel, manganese, and some lead. Reduce the volume t o about 20 cc. and transfer it, with several portions of water, quantitatively to a separatory funnel so that the total volume is not more than 50 cc. Add and dissolve about 30 grams of ammonium thiocyanate. Shake out the solution with a mixture of ether and amyl alcohol, 9 : 1, until exhausted, which is indicated by the disappearance of the blue color of the ammonium cobaltous thiocyanate. (A Rothe estraccor serves 1
Received March 26, 1929.
IsLaND,
h-,Y .
best.) Shake the ether solution with 15 to 20 cc. of sulfuric acid (10 per cent) and wash several times with water. Evaporate the excess water from the aqueous solution, neutralize the remainder with ammonium hydroxide, and electrolyze it. Or make the solution while hot alkaline with sodium hydroxide, filter the precipitate, wash thoroughly with boiling water, ash, and weigh as cobalt oxide. It is possible also to determine the cobalt, as it is the only metal in solution, by means of its 3,5-dimethylpyrasole compound, according to the following method devised by the Siemens Laboratory in Berlin. Pour a 2 per cent solution of the specific cobalt reagent (cold) into the previously nearly neutralized (KaOH) cobalt solution. It should still be faintly acid. Then add about 5 cc. of 0.5 N sodium hydroxide, whereupon all the cobalt settles out as a beautiful purple precipitate, which is analogous t o nickel dimethylglyoxime. Filter this precipitate, wash with cold water, and dry in a Gooch crucible a t as low a temperature as possible. 3,5-Dimethylpyrazole, SH N H3C.C
C.CHa
CH which is a derivative of pyrazole, CK
ecH
CH reveaIed, when its reactions were studied by Fischer (f), a very characteristic affinity toward cobalt. It may be conveniently prepared (after Knorr and Rosengarten) by condensation of acetylacetone with hydrazine hydrate. It may be obtained from the Chemische Werke Schubert in Goerlitz, Germany. It is advisable to recrystallize it before using if it does not show a melting point of 107”. The purple cobalt compound CH C.CH3 C.CH3 CH \SCoN( \ I I / C.CHI N N C.CHs obtained, multiplied by 0.23483, gives the quantity of cobalt. By the foregoing method cobalt contents u p to about 20 per cent (based on the sum of the inorganic constituents of driers) have been satisfactorily determined. The maximum percentage of cobalt that can accurately be determined by i t has not been ascertained, but cobalt can be removed quantitatively from a solution containing a pure cobalt salt when properly treated in this manner. Literature Cited
I
(1) Fischer, Ti’iss. T’eroflegentlzch. Siemens-Konzern, Bd. IV, Heft I1 (1925). ( 2 ) Treadwell and Hall, “Analytical Chemistry,” Vol. I, p. 192, Wiley, 1927.