It is possible to discover the least traces of these metals in admixture together, by means of their property of imparting different colours to a flame. The presence of a salt of sodium in a flame is recognised by a brilliant yellow coloration, and a pure potassium salt colours a colourless flame violet. However, in the presence of a sodium salt the pale violet coloration given by a potassium salt is quite undistinguishable, and it is at first sight impossible in this case to discover the potassium salt in the presence of that of sodium. But by decomposing the light given by a flame coloured by these metals or a mixture of them, by means of a prism, they are both easily distinguishable, because the yellow light emitted by the sodium salt depends on a group of light rays having a definite index of refraction which corresponds with the yellow portion of the solar spectrum, having the index of refraction of the Fraunhofer line (strictly speaking, group of lines) D, whilst the salts of potassium give a light from which these rays are entirely absent, but which contain rays of a red and violet colour. Therefore, if a potassium salt occur in a flame, on decomposing the light (after passing it through a narrow slit) by means of a prism, there will be seen red and violet bands of light situated at a considerable distance from each other; whilst if a sodium salt be present a yellow line will also appear. If both metals simultaneously occur in a flame and emit light, the spectrum lines corresponding to the potassium and the sodium will appear simultaneously.
Fig. 72.—Spectroscope. The prism and table are covered with an opaque cover. The spectrum obtained from the flame coloured by a substance introduced on the wire is viewed through B. A light is placed before the scale D in order to illuminate the image of the scale reflected through B by the side of the prism.
For convenience in carrying on this kind of testing, spectroscopes (fig. [72]) are constructed,[23] consisting of a refracting prism and three tubes placed in the plane of the refracting angle of the prism. One of the tubes, C, has a vertical slit at the end, giving access to the light to be tested, which then passes into the tube (collimator), containing a lens which gives the rays a parallel direction. The rays of light having passed through the slit, and having become parallel, are refracted and dispersed in the prism, and the spectrum formed is observed through the eye-piece of the other telescope B. The third tube D contains a horizontal transparent scale (at the outer end) which is divided into equal divisions. The light from a source such as a gas burner or candle placed before this tube, passes through the scale, and is reflected on that face of the prism which stands before the telescope B, so that the image of the scale is seen through this telescope simultaneously with the spectrum given by the rays passing through the slit of the tube C. In this manner the image of the scale and the spectrum given by the source of light under investigation are seen simultaneously. If the sun's rays be directed through the slit of the tube C, then the observer looking through the eye-piece of B will see the solar spectrum, and (if the aperture of the slit be narrow and the apparatus correctly adjusted) the dark Fraunhofer lines in it.[24] Small-sized spectroscopes are usually so adjusted that (looking through B) the violet portion of the spectrum is seen to the right and the red portion to the left, and the Fraunhofer line D (in the bright yellow portion of the spectrum) is situated on the 50th division of the scale.[25] If the light emitted by an incandescent solid—for example, the Drummond light—be passed through the spectroscope, then all the colours of the solar spectrum are seen, but not the Fraunhofer lines. To observe the result given by a flame coloured by various salts a Bunsen gas burner (or the pale flame of hydrogen gas issuing from a platinum orifice) giving so pale a flame that its spectrum will be practically invisible is placed before the slit. If any compound of sodium be placed in the flame of the gas burner (for which purpose a platinum wire on whose end sodium chloride is fused is fixed to the stand), then the flame is coloured yellow, and on looking through the spectroscope the observer will see a bright yellow line falling upon the 50th division of the scale, which is seen together with the spectrum in the telescope. No yellow lines of other refractive index, nor any rays of any other colour, will be seen, and, therefore, the spectrum corresponding with sodium compounds consists of yellow rays of that index of refraction which belong to the Fraunhofer (black) line D of the solar spectrum. If a potassium salt be introduced into the flame instead of a sodium salt, then two bands will be seen which are much feebler than the bright sodium band—namely, one red line near the Fraunhofer line A and another violet line. Besides which, a pale, almost continuous, spectrum will be observed in the central portions of the scale. If a mixture of sodium and potassium salts be now introduced into the flame, three lines will be seen simultaneously—namely, the red and pale violet lines of potassium and the yellow line of sodium. In this manner it is possible, by the aid of the spectroscope, to determine the relation between the spectra of metals and known portions of the solar spectrum. The continuity of the latter is interrupted by dark lines (that is, by an absence of light of a definite index of refraction), termed the Fraunhofer lines of the solar spectrum. It has been shown by careful observations (by Fraunhofer, Brewster, Foucault, Ångstrom, Kirchhoff, Cornu, Lockyer, Dewar, and others) that there exists an exact agreement between the spectra of certain metals and certain of the Fraunhofer lines. Thus the bright yellow sodium line exactly corresponds with the dark Fraunhofer line D of the solar spectrum. A similar agreement is observed in the case of many other metals. This is not an approximate or chance correlation. In fact, if a spectroscope having a large number of refracting prisms and a high magnifying power be used, it is seen that the dark line D of the solar spectrum consists of an entire system of closely adjacent but definitely situated fine and wide (sharp, distinct) dark lines,[26] and an exactly similar group of bright lines is obtained when the yellow sodium line is examined through the same apparatus, so that each bright sodium line exactly corresponds with a dark line in the solar spectrum.[26 bis] This conformity of the bright lines formed by sodium with the dark lines of the solar spectrum cannot be accidental. This conclusion is further confirmed by the fact that the bright lines of other metals correspond with dark lines of the solar spectrum. Thus, for example, a series of sparks passing between the iron electrodes of a Ruhmkorff coil gives 450 very distinct lines characterising this metal. All these 450 bright lines, constituting the whole spectrum corresponding with iron, are repeated, as Kirchhoff showed, in the solar spectrum as dark Fraunhofer lines which occur in exactly the same situations as the bright lines in the iron spectrum, just as the sodium lines correspond with the band D in the solar spectrum. Many observers have in this manner studied the solar spectrum and the spectra of different metals simultaneously, and discovered in the former lines which correspond not only with sodium and iron, but also with many other metals.[27] The spectra of such elements as hydrogen, oxygen, nitrogen, and other gases may be observed in the so-called Geissler's tubes—that is, in glass tubes containing rarefied gases, through which the discharge of a Ruhmkorff's coil is passed. Thus hydrogen gives a spectrum composed of three lines—a red line corresponding with the Fraunhofer line C, a green line corresponding with the line F, and a violet line corresponding with one of the lines between G and H. Of these rays the red is the brightest, and therefore the general colour of luminous hydrogen (with an electric discharge through a Geissler tube) is reddish.
Fig. 73.—Absorption spectrum (Lecoq de Boisbaudran) of salts of didymium in concentrated and dilute solutions.
The correlation of the Fraunhofer lines with the spectra of metals depends on the phenomenon of the so-called reversal of the spectrum. This phenomenon consists in this, that instead of the bright spectrum corresponding with a metal, under certain circumstances a similar dark spectrum in the form of Fraunhofer lines may be obtained, as will be explained directly. In order to clearly understand the phenomenon of reversed spectra, it must be known that when light passes through certain transparent substances these substances retain rays of a certain refrangibility. The colour of solutions is a proof of this. Light which has passed through a yellow solution of a uranium salt contains no violet rays, and after having passed through a red solution of a permanganate, does not contain many rays in the yellow, blue, and green portions of the spectrum. Solutions of copper salts absorb nearly all red rays. Sometimes colourless solutions also absorb rays of certain definite refractive indexes, and give absorption spectra. Thus solutions of salts of didymium absorb rays of a certain refrangibility, and therefore an impression of black lines is received,[28] as shown in fig. [73]. Many vapours (iodine) and gases (nitric peroxide) give similar spectra. Light which has passed through a deep layer of aqueous vapour, oxygen, or nitrogen also gives an absorption spectrum. For this reason the peculiar (winter) dark lines discovered by Brewster are observed in sunlight, especially in the evening and morning, when the sun's rays pass through the atmosphere (containing these substances) by a longer path than at mid-day. It is evident that the Fraunhofer lines may be ascribed to the absorption of certain rays of light in its passage from the luminous mass of the sun to the earth. The remarkable progress made in all spectroscopic research dates from the investigations made by Kirchhoff (1859) on the relation between absorption spectra and the spectra of luminous incandescent gases. It had already been observed long before (by Fraunhofer, Foucault, Ångstrom) that the bright spectrum of the sodium flame gives two bright lines which are in exactly the same position as two black lines known as D in the solar spectrum, which evidently belong to an absorption spectrum. When Kirchhoff caused diffused sunlight to fall upon the slit of a spectroscope, and placed a sodium flame before it, a perfect superposition was observed—the bright sodium lines completely covered the black lines D of the solar spectrum. When further the continuous spectrum of a Drummond light showed the black line D on placing a sodium flame between it and the slit of the spectroscope—that is, when the Fraunhofer line of the solar spectrum was artificially produced—then there was no doubt that its appearance in the solar spectrum was due to the light passing somewhere through incandescent vapours of sodium. Hence a new theory of reversed spectra[29] arose—that is, of the relation between the waves of light emitted and absorbed by a substance under given conditions of temperature; this is expressed by Kirchhoff's law, discovered by a careful analysis of the phenomena. This law may be formulated in an elementary way as follows: At a given temperature the relation between the intensity of the light emitted (of a definite wave-length) and the absorptive capacity with respect to the same colour (of the same wave-length) is a constant quantity.[30] As a black dull surface emits and also absorbs a considerable quantity of heat rays whilst a polished metallic surface both absorbs and emits but few, so a flame coloured by sodium emits a considerable quantity of yellow rays of a definite refrangibility, and has the property of absorbing a considerable quantity of the rays of the same refractive index. In general, the medium which emits definite rays also absorbs them.