Page images
PDF
EPUB

The whole question has been recently reviewed by Herr Carl Barus, for the U. S. Geological Survey. (Bulletin No. 54, 1889).

Assuming the reliability of Pouillet's investigations, there appears little doubt that, so far as Carnelley's observations extend (900° C., about), they are probably the most reliable yet made; although it is probable—almost certain, indeed-that there is a small minus error increasing with the temperature of melting (see account of his method loc. cit.) It is, perhaps, a further proof of their reliability that their agreement inter se appears very close. If we assume the co-efficient of expansion of platinum to preserve a fairly uniform value over this range, we find, from observations on the meldometer, that his data closely lie along a common line. The agreement is still closer if we assume a slow, fairly uniform, rate of increase of the co-efficient of expansion (à priori probable); and this assumption enables us, also, to extend the line, through the careful results of Violle, for the highest temperatures. Thus, silver 954° C., gold 1045° C., and palladium 1500° C. fall upon such a slow curve as is seen in the plate. Violle's results which, as observed, are also calorimetric, and are admittedly of much value, are, in fact, singled out by Carnelley in his Melting and Boiling-point Tables, as specially reliable. By using these points of Violle's we may extend our knowledge of the melting points of minerals up to (and safely, somewhat beyond) the temperature of 1500° C. with the same degree of accuracy attained in the research of Violle. Pure palladium is supplied by Messrs. Johnson and Matthey. Pure precipitated gold 1045° C. should be used if this substance is experimented on. It is to be spread in fine dust upon the ribbon, and the moment at which the particles break down, seized as the melting point. I have found by interpolation upon the curve that precipitated black oxide of copper, CuO., melts at nearly the same temperature (at 1055° C.), and as this substance breaks down at once, I suggest it as a substitute, giving a convenient point upon the curve between silver and palladium. Its first break-down must be observed, as upon remelting phenomena occur which render the melting point difficult to distinguish. Copper (1054° C.) is difficult to deal with, owing to oxidation; on the meldometer it is hard to distinguish its melting point from that of gold.

A single numerical example of the method of finding a point upon the curve may be of service. Thus in the case of silver chloride, at starting we find the head of the screw reading 1098 divisions from the zero point, the temperature being 15° C.; when the salt is melting we find the head reads 1242 or the advance has been 144. Now each of these represents an advance of the both of an inch, the thread

being pitched roo", and the head carrying 100 divisions. measure when cold the ribbon by a divider, and find it reads 3·80′′. In taking this measurement we are not particular to scale the exact whole length; for although the ribbon is cut away at the ends in the manner before described (to favour uniform heating), it is seen to remain dark where quite close up to the forceps. It therefore conduces to accuracy to read the length about th of an inch less than its full length. The correction is evidently a very immaterial one in any case from the comparative nature of the observations throughout. The fraction l / has now the value 144/38000 = 0.00379 which is accordingly plotted; conveniently to the scale 0·0001 to the millimetre. This length is set off perpendicularly from the horizontal scale of temperature at the point 436° C. The melting point is indeed given by Carnelley as 451° C., but the interval of temperature corresponding to the observed expansion is this value less the initial temperature of 15° C. This is the most convenient course to pursue if we wish to carry the curve through 0° C., which should be done if we wish to preserve it for continued use. Fifteen degrees higher up on the curve we may write AgCl, as is done on Plate VI. In this manner the curve is constructed. An unknown temperature is evidently determined from an observed expansion by calculating the value of / corresponding to this expansion, and finding from the curve what temperature corresponds to such an ordinate. To this must be added a number of degrees corresponding to the initial temperature.

At the very highest temperatures the platinum ribbon reveals a certain amount of viscous stretching. It was very important to determine the amount of this, and its influence in setting a limit to the range of measurements. Observations were accordingly made in which the temperature was raised by different amounts, and the behaviour of the ribbon observed over intervals of five minutes. Where viscous extension took place the ribbon, of course, did not return to its initial length. The true expansion in such cases was deduced by deducting from the whole extension the observed amount by which it failed to return to the initial length. From the extensions the temperature was computed as described, and the following table obtained :

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][ocr errors][merged small][ocr errors][merged small]

A viscous extension of 0.0001 introduces an error of about 3° C. It was observed, too, that at the temperature of melting palladium (1500°) a small viscous extension was sometimes apparent, and also upon prolonged exposure to the temperature of melting quartz. But it is evident that the error introduced by this effect, even if neglected, would not at a temperature so high as 1500° C. maintained for five minutes be a source of serious error. I think the allowance for it which can readily be made by reading the new length 1, after the observation, may be considered to eliminate its effects. Thus producing the curve beyond Violle's result for the melting point of platinum temperatures so high as 1600° C. are determinable, and for their accuracy may with considerable assurance be referred to the degree of accuracy attained by Violle in his determination of the melting point of palladium. So high as 1700° the results obtained with this apparatus will still possess an approximation to this same degree of accuracy.

I submit, then, that contrasting this new method with the scale of Van Kobell, something has been gained. It must be remembered, also, that the apparatus is simple in nature and easily applied. It enables us to determine in a few minutes the melting point of a substance on the same basis of accuracy as that attained by the most careful observers. Only a very minute quantity of material is required, a condition often desirable in dealing with mineral bodies, and often again with bodies prepared by the chemist. Perhaps the peculiar function of the meldometer will be understood from the consideration that from any three or four of his numerous and laborious determinations, Carnelley might, by the use of such an apparatus, have at once arrived at the melting points of the many other substances he dealt with, by interpolation upon the curve of expansions in the same manner in which I propose that mineralogists should determine the melting points of mineral bodies. It is seen that in the apparatus described, the mechanical arrangements are such that there is no magnification of the expansion at the point at which the micrometer is applied. If it was desired to apply this form of platinum-thermometer to more minute measurements, as over a more limited range of temperature, it is of course easy to secure more delicate indications by lengthening the spring, so that we exaggerate at the micrometer the extension of the ribbon, or by lengthening the ribbon itself. Thus doubling the length of the spring with the present dimensions of the ribbon secures a reading closer than one degree centigrade. The applications of such an apparatus-necessitating only the most minute quantities of substances for the determination of their

melting points-might be many in chemical research. On the other hand, however, it must be observed that it appears scarcely applicable to the determination of boiling points, at any rate of volatile liquids, my limited experience in this direction indicating that the small quantity of the liquid is soon evaporated, and the indications of boiling difficult to catch.

For determining the curve of melting points, I have found the following substances convenient :

[merged small][ocr errors][merged small][merged small][merged small][merged small][ocr errors]

-

[merged small][merged small][ocr errors][ocr errors]
[merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][ocr errors][merged small][ocr errors][merged small]

For verification of a new ribbon it is sufficient to select silver chlorides, gold (or cupric oxide), and palladium.

The following list of melting points is further added for convenience of reference. It contains melting points principally determined by Carnelley and Violle, and some others signalized by Carnelley in his Melting and Boiling-point Tables as being specially reliable. Those tried and found suitable are marked with an asterisk. Such are free from phenomena of viscosity or decomposition, the substance having generally well-marked change of appearance upon melting.

[ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][ocr errors][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small]

Metals which oxidize at high temperature as copper, zinc, lead, tin, &c., are

difficult of exact observation.

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][merged small][ocr errors][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][merged small][ocr errors][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][merged small][ocr errors][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][ocr errors][ocr errors][merged small][merged small][merged small][merged small]
« PreviousContinue »