SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. On Platinum and the metals which accompany it.-H. ST. CLAIRE DEVILLE and DEBRAY have published a very interesting and valuable memoir on the platinum metals, considering the subject rather from a metallurgical than from a purely chemical point of view. For the details of the processes employed we must refer to the original memoir, which rarely admits of abridgment, and which is in the highest degree instructive. The authors employ exclusively the dry method of investigation and operate at temperatures much higher than any which have hitherto been obtained upon a working scale. By a new arrangement of the oxyhydrogen blowpipe most of the members of the platinum group may be fused-platinum even in larger quantities than was accomplished by Dr. Hare. By the same apparatus properly employed, the authors purify the metals and their alloys from more volatile elements with which they may be mixed. Osmium has a density of 21.3 to 214, and when dissolved in tin exhibits traces of a crystalline structure. It is bluish white, has no odor, and gives off vapors of osmic acid only above the temperature of melting zinc. At the temperature of melting ruthenium, osmium is sensibly volatilized but it does not fuse, and hence resembles arsenic in having its boiling point lower than its point of fusion. Two determinations of the density of the vapor of osmic acid gave 8.88 and 8.89, corresponding to 2 vols. Next to osmium ruthenium is most difficult of fusion, but may yet be obtained in small fused masses when its density is from 11 to 114. The authors give analyses of the protoxyd of ruthenium and of the crystallized deutoxyd isomorphous with stannic acid. They also describe a beautiful alloy of ruthenium and tin having the formula RuSn2 and crystallizing in cubes. Palladium fuses even more readily than platinum, and volatilizes at the temperature at which iridium melts. It also absorbs oxygen when fused like silver without becoming oxydized. Its density at 22.5° is 11.4. With tin it forms an alloy crystallizing in small brilliant plates having the formula Sn2Pdз. Rhodium fuses less easily than platinum; it has about the color of aluminum and when pure is malleable and ductile; its density is 12.1. It forms crystalline alloys with zinc and tin.

Platinum may (as first shown by Dr. Hare) be fused in large quantities before the oxy-hydrogen blowpipe. Deville and Debray give a detailed description of the apparatus by which this metal may be fused in quantities of not less than 12 to 15 kilograms at an expense of from 0.24 fr. to 0-40 fr. per kilogram. (The late Dr. Hare fused 28 ounces at one operation.) The fusion of platinum is best accomplished in crucibles of lime, which serve to refine the metal by absorbing the impurities. When fused and refined, platinum is as soft as copper; it is whiter than ordinary platinum and free from porosity; its density is 21.15. With tin platinum forms a beautiful crystallized alloy having the formula Pt2Sn3. Iridium requires an extremely high temperature for its fusion, but when fused has a pure white color, and is brittle under the hammer like a crystalline metal; its density is the same as that of platinum, namely 21.15.

SECOND SERIES, VOL. XXIX, No. 85.-JAN., 1860.

(Dr. Hare, who long ago succeeded in fusing iridium, found its density 21.83.) Iridium forms with tin a beautiful alloy crystallizing in cubes having the formula IrSn2.

The authors remark that the alloys of platinum with iridium and rhodium are much more valuable in the arts than pure platinum, many of them resisting the action of aqua-regia, and possessing a considerable degree of hardness and rigidity. The memoir contains in addition numerous elaborate analyses of different specimens of platinum ore and of osmiridium, as well as new processes for the treatment of platinum ores upon the larger scale, the preparation of pure platinum for industrial purposes, and of alloys of platinum with its associate metals possessing useful properties. For these we must refer to the original.-Ann. de Chimie et de Physique, lvi, 385, Aug. 1857.

[Note. A memoir read before the Am. Association for the Advancement of Science at its meeting in August 1859, and shortly to appear in the 12th volume of the Transactions of the Smithsonian Institution, contains entirely new processes for the separation of all the platinum metals in a state of absolute purity. These processes are in the wet way; they are very simple and easy of execution, and they not only apply to the separation but to the qualitative analysis of mixtures of the different metals of this group in almost any proportions. The methods in question involve the preparation and properties of a new and remarkable series of salts, and will I hope be found to remove completely the difficulties which have hitherto surrounded the subject.-w. G.]

2. Blowpipe experiments.-BUNSEN has contributed some very interesting additions to our knowledge of the use of the blowpipe in quantitative as well as qualitative analysis. The author employs the peculiar form of gas burner, first introduced by him, and now used in all laboratories, instead of the blowpipe. The lower part of the flame is surrounded by a conical sheet iron chimney, 30mm, in diameter above, and 55mm, below, so that the burner tube is in the axis of the chimney and 45mm, below the upper base of the cone. The cock is to be regulated so that the apex of the inner non-luminous cone of gas within the flame exactly reaches the level of the upper base of the chimney. In this manner we obtain a flame of perfectly constant dimensions which is immovable, sharply defined in all its parts, and which may always be obtained of uniform character. The outer cone of flame has a very faint sky-blue color, which is invisible even by feeble daylight. The inner cone of flame is less intensely blue than the outer. The object to be submitted to the action of the flame must never be larger than the half or the third of a grain of millet seed. It is to be introduced into the flame by means of a little loop on the end of a platinum wire which is attached to a holder by which it may be moved gently and steadily, so that the object may be introduced into any part of the flame. The loop is to be moistened with water, when a grain or a little of the powdered substance will readily adhere to it.

The author remarks that the temperature which the flame is capable of producing depends principally upon the constitution of the gas consumed. The temperatures corresponding to gas analyzed in the Heidelberg laboratory on four different occasions were 2369 C., 2352 C., 2391

C., 2386 C., or as a mean, 2350 C., so that the temperature of the flame where the quantity of air is exactly sufficient for the combustion of the gas, may be assumed in round numbers as 2300 C. It is easy to see, however, that the temperature will vary in different parts of the flame. The author gives a simple and elegant method of determining the point of maximum temperature by introducing a platinum wire into the flame, and determining at what point the light emitted by this is most intense. In this manner it is found that the zone of maximum temperature lies in the external cone of flame, a few millimetres above and below the apex of the internal non-luminous cone, which is on a level with the upper base of the chimney. The author employs this zone to investigate the action of a temperature of 2300 C. upon different substances, and terms it the melting space. The outer border of this melting space acts as an oxydizing flame, the inner as a reducing flame, the reduction being most powerful immediately above the apex of the innermost cone. The great constancy which the flame exhibits in all its parts allows us to observe and estimate the volatility of substances at the very high temperature of 2300 C. For this purpose a mass of matter having a measured diameter of 1 millimeter is introduced into the flame, and the time required for complete evaporation determined by means of a seconds pendulum or a metronome. The size of the mass introduced upon the platinum wire is easily regulated by adding new substance or by evaporation in the flame, and may be measured under the microscope. The author takes the volatility of carbonate of soda as unity, and gives a table of the comparative volatility of different substances in terms of this unit: thus the volatility of chlorid of potassium is 15.33; of chlorid of sodium 6.57; of phosphoric acid 23.00. Other substances are more or less completely decomposed at the temperature of 2300 C.

Besides its use in experiments on volatility the flame may also be employed for a series of other very valuable blowpipe reactions, among which the author cites the quantitative determination of soda in the presence of potash and lithia.

For the simple recognition of soda in its volatile salts, it is sufficient to introduce a small bead of the substance into the melting space and then to illuminate a crystal of bicromate of potash with the light produced. The salt appears perfectly colorless, transparent, and with a diamond lustre so long as the rays of the soda flanie fall upon it, and this even by ordinary lamplight or daylight. A still more delicate reaction is obtained by using paper, covered with iodid of mercury, a square centimeter of which may be attached to the chimney in front of the flame by a movable arm. If we introduce the smallest quantity of a soda compound into the melting space, the red paper appears white, with a faint tinge of tawny yellow. When the soda salt is in solution the loop of the fine platinum wire may be flattened under the hammer to a little ring. This ring introduced into the liquid will take up a drop which must be gently evaporated to dryness and then tested as before. In this manner Too of a milligram of common salt may be easily detected. Volatile potash compounds, as is well known, communicate a bluish violet tint which is completely concealed by small quantities of soda. In this case the potash may easily be detected by means of Cartmell's

reaction, that is, by looking at the flame through a deep blue cobalt glass, when a violet or ponceau-red color appears. This potash test is even more delicate than that for soda, Too parts of chlorid of potassium may be detected with perfect distinctness.

The detection of lithia in the presence of potash and soda is best effected by looking at the flame through a hollow prism filled with a solution of sulphate of indigo. The carmine-red color of the lithia flame disappears when a certain thickness of solution is reached, and if a mark be put upon the prism, all the layers of liquid above the mark will allow only the red potash-rays to pass through. Lime and soda have no influence on this reaction. When potash and lithia are both present the flame from the salt to be tested should be compared with that of pure potash. The flame containing lithia and potash appears through thin layers of liquid redder than the potash flame alone; through thicker layers the potash flame appears scarcely weakened. In this manner some thousandths of lithia may the discovered in potash salts.

To render these processes available for silicates Bunsen mixes the mineral with pure gypsum and heats in the melting space, when silicate of lime and volatile sulphate of potash are formed. By comparing the intensity of the color produced by a mineral to be tested with that of a series of silicates whose percentages of alkali are known, it is easy even with very small fragments to determine the relative quantities of potash, soda and lithia with tolerable approximation.

The author in the first place determines whether the mineral to be examined contains lithia or not by the method already explained. The minerals of the first group are those containing no lithia, and the author arranges a scale of minerals for comparison according to the content of soda in each. These are

These are to be ignited, pulverized and preserved with their numbers as blowpipe reagents. One of these and the mineral to be tested, with or without gypsum, are to be introduced into the flame together so that small equal ends of wire are ignited; the iodid of mercury paper then appears more or less bleached. Remove the test from the flame, if now the paper shows a reddish tint, the test contains more soda than the mineral used for comparison. If the paper however becomes paler the contrary is the case. In this manner it is easy to determine between what two minerals the test lies, so that the percentage of soda may be estimated within a few per cent. The substances to be compared must be as nearly as possible in equal quantities, the ignited lengths of wire be the same and the soda-flame of the same size and form. The eye must be accustomed to distinguish the different degrees of brightness of the same tint

from actual differences of tint. When the iodid paper is intensely bleached it may also be illuminated by a candle flame so that the sodaflame produces with this foreign light a white tint nearer to red.

The potash color-test is not as accurate quantitatively as that for soda. It is sufficient for all purposes to distinguish a slight, a strong, and a very strong potash reaction, using for comparison in succession the flame of oligoclase, orthoclase and leucite heated with gypsum in the same flame with the test. The indigo prism is to be used and the dimensions, color and duration of the red flames observed.

Lazulite gives a stronger soda reaction than nepheline because it contains sulphuric acid, and it is always necessary to determine beforehand whether the test contains sulphuric acid, chlorine or fluorine. This is best accomplished by the common blowpipe.

For further details we must refer to the original memoir, which must create an entirely new department in blowpipe analysis.-Ann. der Chemie und Pharm., cxi, 257.

[Note. It is easy to see that Bunsen's memoir contains the solution of many chemical and physical problems of great interest. Thus it is easy to produce at will a flame which shall have any required temperature, at least between certain limits. Since according to Bunsen's calculations (Gasometrische Analyse, p. 254) the flame of hydrogen burning freely in the air has a temperature of 3259° C., and that of olefiant gas has a temperature of 5413° C., we may as readily experiment at these temperatures as at 2300° C. Very much higher temperatures are of course produced when these gases burn with pure oxygen-in the case of hydrogen 8061° C.; in that of olefiant gas 9187° C. Hence by burning mixtures of these gases with oxygen and varying the proportion of nitrogen arbitrarily, we can make a scale of flames the temperatures of which shall range from less than 2000° C. up to at least 9000° C. and the melting point as well as degree of volatility of almost all metals and mineral substances may be thus assigned within quite narrow limits. It is easy to calculate the percentage of nitrogen or carbonic acid to be added in each case to the combustible gas in order to produce the required temperature. It appears probable that at high temperatures the radiating power of a body for heat is proportional to its radiating power for light. By directly comparing the intensity of the light radiated from platinum heated in a furnace with the intensity of the light radiated when the platinum is heated to a known temperature in a gas-flame, the temperature of the furnace might be approximately determined. Interesting results could also be obtained as to the exact temperature at which bodies become luminous and as to the relative quantities of light which different substances emit when heated to the same temperature, or the same substance at different temperatures. With respect to Bunsen's scale or series of minerals containing different but known proportions of soda I will suggest that perhaps a series of glasses could be made containing perfectly definite quantities of either soda or potash from 2 up to 30 per cent, so that each number would contain 2 per cent more alkali that the next lowest number. These would then serve as universal standards of comparison and give much precision to blowpipe observations.--W. G.]

W. G.