Ordinarily one would conclude that the highly satisfactory | agreement of the determinations by so many different observers, using different methods, as summarised in Table IV., would settle this point. A careful study of the gas thermometers and methods used, however, shows that they are open to many criticisms, and many refinements are now possible that were not available or whose importance was not so clearly recognised at the time these determinations were carried out. Thus, in some of these determinations air was used as the gas instead of nitrogen. Glass or porcelain bulbs, so subject to changes in volume or to the emission or absorption of gases, were employed instead of metal bulbs. The coefficient of expansion of the bulb was not always determined with sufficient accuracy over the whole of the temperature range, and in some cases the coefficient used was found by an unwarranted extrapolation.

The use of a uniform scale by different experimenters is always most desirable. In view of the very general use of the value 444'7° as the boiling point of sulphur on the scale of the constant volume gas thermometer, and in the light of the evidence that is now available, no change in this point is possible at this time, It is well to bear in mind, however, that the standard gas scale is not yet definitely and certainly fixed to an accuracy of 1° at 450° C.

The following table of fixed points represents the temperature scale which appears to best satisfy the available observations to o 1° for the reproduction of temperatures in the interval 100° to 500° C. :

-

As these temperatures are on the constant volume scale, the reduction to the thermodynamic scale may be effected

approximately if desired by adding 0·2° at the sulphur point and a proportional amount at the lower temperatures, remembering that this correction is zero at 100° C. All of these substances may readily be obtained of sufficient purity to give a reproducibility of 0.05° C. from one sample to another.

I.

References.

"Proces Verbeaux du Committee International des Poids et Mesures, Séance due 15 Octobre, 1887." 2. Waidner and Burgess," Platinum Resistance Thermometry at High Temperatures," Bull. Bureau of Standards, 1910, vi., pp. 149-230.

3. J. M. Crafts, "Les Mesures Thermometriques et la Determination des Points de Fusion et d'Ebullition," Bull. Soc. Chim., 1883, xxxix., pp. 277–289.

4. Callendar and Griffiths, "On the Determination of the Boiling point of Sulphur, and on a Method of Standardising Platinum Resistance Thermometers by Reference to it," Phil. Trans., 1891, A, clxxxii., pp. 119-157.

5.

E. H. Griffiths (note to Ref. 4), Phil. Trans., 1891, A, clxxxii., 151.

6. Jaquerod and Wassmer, "Points d'Ebullition sous Diverses Pressions de la Naphthaline, du Biphenyle et de la Benzophenone, Determiné au Moyen du Thermomètre à Hydrogène," Journ. Chim. Phys., 1904, ii., pp. 52–78.

7. Travers and Gwyer, Proc. Roy. Soc., 1905, lxxiv., 528.

8. Holborn and Henning, "Uber das Platinthermometer und den Sättigungsdruck des Wasserdampfes zwischen 50° and 200°," Ann. Phys., 1908, xxvi., pp. 833-883.

9. Regnault, "Relation des Expériences," 1862, ii., 526.

10. Thätigheit der P. T. Reichsanstalt, Zeit. f. Instru mentenkunde, 1894, xiv., 304.

CHEMICAL NEWS,

Jan. 20, 1911

and moistened with hydrochloric acid. It is diluted with water, and the silica is filtered off, weighed, and this with that previously obtained is the total silica.

Iron, Alumina, and Manganese.-The solution from the silica is raised to boiling and the iron and aluminium are precipitated as hydroxides. Then 5 cc. of bromine water is added, and the boiling continued for five minutes. The precipitate is dried on filter-paper and ignited separately from it in a weighed platinum crucible, to which the ash of the filter-paper is afterwards added. The precipitate consists of Al2O3, Fe2O3, and Mn2O3, and is weighed as such. It is then fused with fifteen times its weight of potassium pyrosulphate over a low flame for three hours with the crucible covered. The crucible, contents, and cover are placed in a beaker, and dilute sulphuric acid (101) is added. By warming and continued shaking of liquid complete solution may be obtained. It is then drawn through a Jones reductor to change all the iron to ferrous, and titrated with N/10 potassium permanganate solution. The iron is calculated to Fe2O3, and the alumina determined by difference.

If manganese is present it is determined in a separate sample in a method given later, and is subtracted from the iron in the above. In white enamels containing only a trace of iron the manganese may be determined in the solution from the pyrosulphate fusion. A freshly prepared solution of potassium ferricyanide is added to oxidise the manganese, then the solution is made alkaline with sodium hydroxide solution, and the manganese-dioxide thus formed is filtered off. The solution is then made acid, and the ferrocyanide is titrated with N/10 potassium permanganate solution (1 cc. KMnO4 0.00435 grm. MnO).

=

Calcium Oxide.-The filtrate from the iron and alumina is raised to boiling, treated with boiling ammonium oxalate solution, and digested on water-bath until precipitate readily and quickly settles after being stirred. The calcium oxalate is now filtered off, and ignited wet in platinum to constant weight over a strong blast.

Magnesium Oxide. The solution is evaporated to dryness, and the residue ignited to remove ammonium salts. The residue is treated with a few drops of hydrochloric acid and taken up with boiling water and filtered from the carbonaceous residue. To the boiling solution is added drop by drop a solution of sodium ammonium phosphate, and is allowed to cool. Half as much concentrated ammonium hydroxide is added as there is solution, and it is allowed to stand over night. The precipitate is collected on a filter, washed with 3 per cent ammonia water, dried in oven, and ignited separate from the filter. The heat is applied gently at first, and finally with the highest heat of a good Bunsen burner. It is then weighed as Mg2P2O7: I grm. Mg2P207 0.3625 grms. MgO.

=

The alkalis are determined by the method of J. Lawrence Smith from a grm. sample finely powdered. This method is standard and need not be given here.

Separation and Determination of Antimony, Tin, Manganese, and Cobalt in Enamel. Decomposition.-Two grms. finely powdered sample are transferred to a platinum dish, and after moistening with a little water, pure hydrofluoric acid is added, and the whole is mixed with a platinum spatula. The dish is digested on steam-bath for five hours covered with platinum cover (a larger platinum dish may be used for cover if no other is at hand). After the decomposition is complete the solution is evaporated to dryness on steam-bath. The residue is moistened with enough dilute sulphuric acid (1 : 1) to make a thin paste, and evaporated as far as possible on a steambath, and then on a hot plate, all the time being covered to prevent spirting. As soon as fumes of sulphuric anhydride cease to be evolved the cover is strongly heated until fumes cease to be driven off, when it is removed. The contents are heated by bringing the dish to dull redness directly over a Bunsen burner. The sulphates

29

thus formed are moistened with strong hydrochloric acid, a little hot water is added, and the solution boiled with repeated additions of acid and water until completely in solution. In some enamels-especially those with high melting-points-the stannic oxide remains undissolved, and a fusion of the residue with sulphur and sodium carbonate may be necessary.

Treatment with H2S.-The solution containing at least 30 cc. double normal hydrochloric acid is transferred to a 500 cc. Erlenmeyer flask fitted with a double bored stopper. Through one of the holes a right-angled piece of glass tubing is introduced that just reaches to the lower edge of the stopper, while through another hole another rightangled glass tube is fixed so that it almost reaches the bottom of the flask.

A Kipp H2S generator is connected to the longer tube and H2S is passed through for half-an-hour, and the solution is let stand for another half-an-hour, after which the sulphides of antimony and tin are transferred to a filterpaper, and the solution is kept for the determination of manganese and cobalt.

Antimony and Tin.-The precipitated sulphides are dissolved in a solution of potassium polysulphide-if any lead or copper is present it will remain undissolved and may be determined separately-by pouring this successively through the filter into a 300 cc. Jena beaker, and finally washing with water containing a small amount of potassium polysulphide.

Antimony. The antimony and tin in this solution are separated by F. W. Clark's method as modified by F. Henz (Treadwell, vol. ii., p. 188), as follows:

To the solution in the Jena beaker 6 grms. caustic potash and 3 grms. tartaric acid are added. To this mixture twice as much 30 per cent hydrogen peroxide is added as is necessary to completely decolorise the solution, and the latter is now heated to boiling and kept there until the evolution of oxygen is over, in order to oxidise the thiosulphate formed. All of the excess of peroxide cannot be removed successfully at this point. The solution is cooled somewhat, the beaker covered with a watch-glass, and a hot solution of 15 grms. pure re-crystallised oxalic acid is cautiously added (5 grms. for o'i grm. of the mixed metals). This causes the evolution of considerable carbon dioxide. Now, in order to completely remove the excess of hydrogen peroxide the solution is boiled vigorously for ten minutes. The volume of the liquid should amount to from 80 to IOO CC. After this a rapid stream of hydrogen sulphide is conducted into the boiling solution, and for some time there will be no precipitation, but only a white turbidity formed. At the end of five or ten minutes the solution becomes orange coloured, and the antimony begins to precipitate, and from this point the time is taken. At the end of fifteen minutes the solution is diluted with hot water to a volume of 250 cc., at the end of another fifteen minutes the flame is removed, and ten minutes later the current of hydrogen sulphide is stopped. The precipitated antimony pentasulphide is filtered off through a Gooch crucible, which, before weighing and after drying, has been heated in a stream of carbon dioxide at 300° C. for at least one hour. The precipitate is washed twice by decantation with 1 per cent oxalic acid and twice with very dilute acetic acid before bringing it in the crucible. Both of these wash liquids should be boiling hot and saturated with hydrogen sulphide.

The crucible is heated in a current or carbon dioxide (free from air) to constant weight, and its contents weighed as Sb2S3.

The filtrate is evaporated to a volume of about 225 cc., transferred to a weighed unpolished platinum dish, and electrolysed at 60° to 80° C. with a current of o'2 to 0.3 ampère (corresponding to 2 to 3 volts). For very small amounts of tin, a current of not over o.2 ampère should be used. At the end of six hours 8 cc. of sulphuric acid (1 : 1 are added, and at the end of twenty-four hours the solution is transferred to another dish. The deposited tin ha a beautiful appearance, similar to silver.

Tin.

The solution containing the cobalt and manganese is boiled until free from H2S. The iron is oxidised back to the ferric state by the addition of bromine water and boiling until the excess of the latter is expelled. Ten cc. double normal ammonium chloride is added, and the iron and alumina are precipitated by the addition of ammonia, and are filtered off. (The iron alumina may be determined from this precipitate if desired).

The solution still containing the manganese and cobalt is transferred to an Erlenmeyer flask fitted for passing in H2S, as before described, and 3 cc. strong ammonia is added. H2S is passed through for some time, and after precipitation ceases 3 cc. more of ammonia are added, and the flask is filled to the neck (300 cc. flask), is corked, and set aside for twelve hours at least. The precipitate is collected and washed on a small filter with water containing ammonium chloride and sulphide.

Manganese. -The manganese is extracted from the precipitate on the filter by pouring through it strong H2S water acidified with one-fifth its volume hydrochloric acid (sp. gr. 1.11). This solution from the extraction is evaporated to dryness, ammonium salts are destroyed by evaporation with a few drops of sodium carbonate solution, hydrochloric acid and a drop of sulphurous acid are added to decompose excess of carbonate and to dissolve the precipitated manganese, and the latter is re-precipitated at boiling heat by sodium carbonate after evaporating off the hydrochloric acid. The manganese is weighed as Mn304 and calculated to MnO2, in which form it is probably present in the enamel.

The residue of cobalt sulphide left after extracting the manganese is burned in a porcelain crucible, dissolved in aqua regia, and evaporated with hydrochloric acid; the platinum-and copper, if any is present-are thrown down by heating and passing in hydrogen sulphide. The filtrate from the platinum and copper is made ammoniacal, and cobalt is thrown down by hydrogen sulphide. This is filtered off and washed with water containing ammonium sulphide. This is either ignited and weighed as oxide or more accurately determined by dissolving in an ammoniacal solution of ammonium sulphate, containing 10 grma. of ammonium sulphate and 40 cc. of concentrated ammonia for each 0.3 grm. of cobalt, and electrolysing in a weighed platinum dish at room temperature with a current of 0.5 to 15 ampère, and an electromotive force of 2.8 to 3.3 volts. The electrolysis is finished in three hours. The circuit is broken and the liquid poured off, and the platinum dish is washed with water, then with absolute alcohol (distilled one hour), and finally with ether, allowed to dry in oven at 95° for one minute, and then weighed. The metallic cobalt is calculated as CoO, in which form it is present in the enamel.

The Determination of Boric Anhydride in Enamel. The boron is determined in a separate sample of about 0.3 grm. This finely pulverised sample is fused with 3 grms. sodium carbonate for fifteen minutes, is taken up with 30 cc. dilute hydrochloric acid and a few drops of nitric acid. The melt is heated in a 250 cc. round bottomed flask almost to boiling, and dry precipitated calcium carbonate is added in moderate excess. The solution is boiled in the flask after it has been connected with a 6-inch worm reflux condenser. The precipitate is filtered on an 8 cm. Buchner funnel (see method of Wherry and Chapin, Journ. Am. Chem. Soc., xxx., 1688, for "Determination of Boron in Silicates "), and is washed several times with hot water, taking care that the total volume of the liquid does not exceed 100 cc.

The filtrate is returned to the flask, a pinch of calcium carbonate is added, and the solution is heated to boiling to emove the free carbon dioxide. This is best done by connecting the flask to a suction-pump, and the suction is

Lead. The enamel for cooking utensils should never contain lead. To determine whether a cooking utensil contains lead, E. Adam gives the following simple qualitative method:-A small piece of filter-paper moistened with hydrofluoric acid is placed upon the enamel and allowed to remain for some minutes; the paper, together with any pasty mass adhering to the enamel, is then washed off into a small platinum basin, diluted with water, and tested for lead by passing H2S through the solution. J. Grünwald (Oesterr. Chem. Zeit., viii., 46) gives another quick test for lead :-Wet small portion of surface with HNO3 (conc.), and heat until acid is evaporated. Add several drops of water and a few drops 10 per cent potassium iodide solution, and if even a trace of lead is present yellow lead-iodide will be produced.

Determination of Phosphoric Anhydride in Enamel. Enamels containing bone-ash to give opaqueness are analysed for P2O5 as follows:

To a grm. sample of very finely pulverised enamel in a platinum crucible I cc. of sulphuric acid is added and the crucible is filled half full (about 10 cc. are required) with hydrofluoric acid. The crucible is heated on the waterbath until most of the solution is evaporated, and then gently on a hot plate to remove all the fluorine as silicontetrafluoride and as hydrofluoric acid, but no sulphuric acid fumes should evolve, as P2O5 is volatile. The residue is dissolved in nitric acid and taken to dryness, moistened with nitric acid, diluted with water, filtered, and washed with a very little water.

Add aqueous ammonia to the solution from above until the precipitate of calcium phosphate first produced just fails to re-dissolve, and then dissolve this by adding a few drops of nitric acid. Warm the solution to about 70° C., and add 50 cc. ammonium molybdate solution (70 grms. MoO3 per litre). Allow the mixture to digest at 50° for twelve hours. Filter off precipitate, washing by decantation with a solution of ammonium nitrate made acid with nitric acid.

The precipitate on the filter is dissolved by pouring through it dilute ammonia solution (1 volume of 0.96 sp. gr. ammonia to 3 volumes of water).

The solution is received in the beaker containing the bulk of the precipitate, all of which is dissolved in the ammonia solution.

An excess of magnesium ammonium chloride ("magnesia mixture") solution is added very slowly and with constant stirring. Let solution stand over night. Decant clear solution through a filter and wash by decantation with ammonia water (1:3). Dissolve the precipitate by pouring a little hydrochloric acid (sp. gr. 1'12) through the filter, allowing the acid solution to run into the beaker containing most of the precipitate. When all the precipitate on the filter and in the beaker is dissolved, wash the filter-paper with a little hot water. To the solution add 2 cc. magnesia mixture and then strong ammonia, drop by drop, with constant stirring until distinctly ammoniacal. Stir several minutes, then add strong ammonia equal to onethird of the liquid, let stand two hours, and filter off the precipitate of magnesium ammonium phosphate. Wash with dilute ammonia water, dry the precipitate, ignite separately from the filter, first at low temperature, and gradually raise to full blast. Weigh precipitate as Mg2P207 and calculate as P2O5 in sample.

CHEMICAL NEWS,

Jan. 20, 1911
FLUXES AS APPLIED TO THE BRASS

FOUNDRY.*

By ERWIN S. SPERRY.

IN the early days of brass founding two things were guarded jealously: the mixtures and the fluxes. At that time it was not an uncommon occurrence to hear one say, when speaking of a new proprietor of a brass foundry, "He will get along all right. His father left him all his formulas for mixtures and fluxes." It seemed to be taken for granted that even capital itself was subsidiary to those heirlooms. Soon the chemist began to make serious inroads into the mixtures, and their secrecy faded away gradually but surely. The mystery of the fluxes was more difficult to eliminate, as, unlike the castings themselves, they did not go beyond the foundry. They could not be analysed without obtaining at least a small quantity, and this was difficult to do, as the brass founder carefully guarded them and the materials from which they were made. In the course of time, however, the secret fluxes

went the way of the brass mixtures, so that by the process of evolution "secrets" finally became general technical knowledge. Few brass founders are now found who claim to have anything original or remarkable in the flux line. When fluxes are mentioned, it is frequently asked, What is their advantage, and are they actually necessary? To answer this question, I will say that I believe the flux question is greatly overdone and likewise imperfectly understood. It was only a short time ago that I discovered a brass founder using lime in making an 88-10-2 mixture from new metals. He said "somebody" had told him that it was good for bronze. Although cheap substance that it is, he was wasting his lime, as it remained on the

metal like so much dirt, and had no more effect than sand. In another instance I found that a brass ingot maker used a couple of handfuls of a mixture of common salt and borax on the top of a pot of yellow brass after it had been skimmed and immediately before pouring. The mixture was thrown on and then at once skimmed off without waiting for the borax even to stop swelling. He, too, would have been equally as well off without the flux. Another, but diametrically opposite case, was a brass founder who attempted to melt his emery grindings in a crucible without a flux and obtained about a pint of metal and half a bushel of dross. In this case a flux was needed. It hardly will be advisable to go into a detailed enumeration of all the fluxes known to mankind and that can be used in brass melting, as it would be of little value. Many substances once used for this purpose have now become obsolete, and many others have proven to be valueless. It seems preferable, therefore, to give a description of those fluxes that the test of time has proved valuable and the manner in which they should be used.

Flux for Aluminium. -For years those who melted aluminium used no fluxes at all on it, not even charcoal, as it was soon found that this material did more harm than

good. On account of the lightness of aluminium, charcoal does not readily free itself, and is apt to become entangled in the metal and produce small black spots in the casting. It is within the past few years that a flux has become used. For aluminium the flux that is most extensively used, and which has proved to be so valuable, is chloride of zinc. It seems to react with the aluminium, forming chloride of aluminium and metallic zinc, which alloys with the aluminium. When this takes place the dross is changed to a fine granular condition, which is readily skimmed off. When aluminium is melted, the surface is covered with a rather thick mass, but the chloride of zinc will change it to a perfectly clear one closely resembling in appearance molten tin or lead. It is needless to say that such clean metal gives better castings.

The method of using chloride of zinc as a flux in melting

Paper read before the American Brass Founders' Association.From the Chemical Engineer, xii., No. 6.

31

aluminium is simple. Small pieces are thrown on the suiface after the melting has been completed. Enough has been added when the surface is clear. A very small amount usually suffices, and for 50 lbs. of aluminium a piece of the size of a walnut is generally enough. The metal is stirred immediately after the addition, and then skimmed. Those who have not used chloride of zinc should try it, as it is an excellent material.

Flux for Nickel-The flux used by makers of nickel anodes has proved to be a good one. It was first used by A. M. Hill, of New Haven, Conn., one of the first in the United States to make nickel anodes and the inventor of the well known "Hill-Barrel" for grinding brass furnace ashes. Mr. Hill has retired from the anode business, but for many years he used this flux with the best results. It is not only good but cheap. It is composed of the following :

Lime .. Fluor-spar.

The manner of making it is to take the lime and slake fluor-spar and allow it to become solid. It is then broken it as though mortar were to be made. Then stir in : up into small pieces for use.

While fluor-spar alone is a good flux, it becomes very fluid when melted, and rapidly attacks a crucible. I have seen a new plumbago crucible ruined in one heat when fluor-spar was used alone. It seemed to soak in and dissolve out the clay from the crucible mixture and leave nothing but the graphite. The crucible collapsed like an egg shell when grasped with the tongs. The use of the lime with the fluor-spar is to increase the melting-point so that it will not so readily attack the crucible. The proportions previously mentioned have been found satisfactory for nickel. Less lime will render it more fusible.

This flux has been found particularly serviceable in melting old anodes, as it dissolves any earthy matter that may be on them. It is used for both new and old material, however, and may be called the standard flux for nickel. The proportions used are about a pint or a good handful for a new nickel, and twice this quantity for old material. It must not be imagined because the fluor-spar is toned down with lime that the flux will not act on the crucible, for it certainly will. The crucibles last only five or six heats. In this connection it should be borne in mind that

all fluxes act on the crucible to a greater or less extent, otherwise they would not be of value as a flux.

Flux for Copper.-There probably have been more fluxes proposed or used for copper than for any other metal or its alloys. The fact that copper cannot be melted alone and obtain sound castings from it has brought about this fact. In the Practically every known chemical has been tried. selection of a flux for copper it should be known whether pure copper castings are to be made, or whether it is to be alloyed to make brass or bronze.

To make sound copper castings with a flux alone, and without the use of "physic" like silicon-copper, magnesium, or similar materials (which strictly speaking are found that yellow prussiate of potassium (potassium ferronot fluxes) is a difficult matter. For this purpose I have cyanide) is excellent. With it sound copper castings can be made, but I do not advise it, as far better results may be obtained by the usual deoxidising agents, such as silicon-copper, magnesium, phosphorus, &c.

Its

In melting copper for producing brass or bronze, the For this question is different from the preceding one. purpose there is nothing better than common salt. value lies in the fact that it possesses the property of reducing any oxide of copper which may form during the melting. It has been used for years in the brass industry, and the memory of the "oldest inhabitant" fails to indicate the date of its inception.

Common salt is so efficacious in reducing oxide of copper that, as far back as 1882, R. Monger in the CHEMICAL NEWS proposed it as a means for determining the quantity of oxide that copper contains. His method