Part 51

Sir D. Brewster first noticed in 1832 that certain coloured gases have the power of absorbing some of the sun’s rays, so that the spectrum, when the rays are made to pass through such a gas before falling on the prism, is crossed by a series of dark lines—altogether different from Fraunhofer’s lines, though these are also present. The gas in which this property was first noticed is that called “nitric peroxide”—a brownish-red gas, of which even a thin stratum produces a well-marked series of dark lines. The same property was soon discovered in the vapours of bromine, iodine, and a certain compound of chlorine and oxygen. Each substance furnishes a system of lines peculiar to itself: thus the vapour of bromine, although it has almost exactly the same colour as nitric peroxide, gives a totally different set of lines. These, therefore, do not depend on the mere colour of the gas or vapour, and this is conclusively proved by the fact of many coloured vapours producing no dark lines whatever: the vapour of tungsten chloride, for example, although in colour so exactly like bromine vapour that the two cannot be distinguished by the eye, yields no lines whatever.

In Fig. 218 is represented a lamp for burning coal-gas, which is constantly used by chemists as a source of heat. It is known as “Bunsen’s burner,” from its inventor the celebrated German chemist. It consists of a metal tube, 3 in. or 4 in. long, and ⅓ in. in diameter, at the bottom of which the gas is admitted by a small jet communicating with the elastic tube which brings the gas to the apparatus. A little below the level of the jet there are two lateral openings which admit air to the tube. The gas, therefore, becomes mixed with air within the tube, and this inflammable mixture streams from the top of the tube and readily ignites on the approach of a flame, the mixture burning with a pale bluish flame of a very high temperature. This little apparatus is not only the most useful pieces of chemical apparatus ever devised, but it furnishes highly instructive illustrations of several points in chemical and physical science; and to some of these we invite the reader’s attention, as they have an immediate bearing on our present subject. Coal-gas is a mixture of various compounds of the two elementary bodies, hydrogen and carbon; and when the gas burns, these substances are respectively uniting with the oxygen of the air, producing water and carbonic acid gas. Now, when coal-gas is burnt in the ordinary manner as a source of light, the supply of oxygen is too small to admit of the complete combustion of all its constituents; and as the oxygen more eagerly seizes upon the hydrogen than upon the carbon, a large proportion of the latter thus set free from its hydrogen compound is deposited in the flame in the solid form, and is there intensely heated. The presence of solid carbon in an ordinary gas flame is easily proved by holding in it a cold fragment of porcelain, or a piece of metal, which will become covered with soot. In the flame of the Bunsen burner there is no soot, because the increased supply of oxygen, afforded by previously mixing the gas with air, enables the whole of the constituents of the gas to be completely burnt; and this is of the greatest advantage to the chemist, who always desires to have the vessels he heats free from soot, in order that he may observe what is taking place within them. The flame of Bunsen’s lamp becomes that of an ordinary sooty gas flame, when the two orifices which admit the air at the bottom of the tube are closed up, and then, of course, the temperature cannot be so high as when the whole constituents of the gas are completely burnt, but the flame becomes highly luminous; whereas when the orifices are open it gives so little light, that in a dark room one cannot see a finger held 20 in. from the lamp. Plainly the cause of this difference is connected with the presence or absence of the heated particles of solid carbon. The non-luminous flame contains no solid particles; the bright part of the other flame is full of them. To these heated particles of solid carbon we are, then, indebted for the light which burning coal-gas supplies. And, since we are able by such artificial illumination to distinguish colours, the white-hot carbon must give off rays of all degrees of refrangibility, and we should expect to find in the spectrum produced by such a flame, the red, yellow, green, and other coloured rays. And such is indeed the spectrum which these incandescent carbon particles produce: it resembles the solar spectrum, but _there is an entire absence of dark lines_, so that the appearance is that represented in No. 1, Plate XVII., if we suppose the Fraunhofer lines removed. If the pale blue flame of the Bunsen’s burner be similarly examined, the spectrum, No. 14, Plate XVII., shows that only a few rays of certain refrangibilities are emitted, forming bright lines here and there, but of little intensity, while the whole of the other rays are absent. This shows that while the highly heated solid gives off all rays from red to violet without interruption, the still more highly heated gases give off only a few selected rays.

It has long been known that some substances impart certain colours to flames, and such substances have been long employed to produce coloured effects in fireworks, &c. But coloured flames do not appear to have been examined by the prism until 1822, when Sir John Herschel described the spectra of strontium, copper, and of some other substances, remarking that “The colours thus communicated by the different bases to flame afford in many cases a ready and neat way of detecting extremely minute quantities of them.” A few years later, Fox Talbot described the method of obtaining a monochromatic flame, by using in a spirit-lamp diluted alcohol in which a little salt has been dissolved. The paper in which he describes this and other observations concludes thus: “If this opinion should be correct and applicable to the other definite rays, a glance at the prismatic spectrum of flame may show it to contain substances which it would otherwise require a laborious chemical analysis to detect.” Here we have the first hint of that spectrum analysis which has provided the chemist with a method of surpassing delicacy for the detection of metallic elements. The spectra of coloured flames were also subsequently examined and described by Professor W. A. Miller, but the most complete investigation into the subject was made by Professors Kirchhoff and Bunsen, who also contrived a convenient instrument, or _spectroscope_, for the examination and comparison of different spectra. The instrument has received many improvements and modifications, but the essential parts are one or more prisms; a slit, through which the light to be examined is allowed to enter; a tube, having at the other end a lens to render parallel the rays from the slit; a telescope, through which the spectrum is viewed; and usually some apparatus by which the positions of the different lines may be identified.

A very elegant instrument, made by Mr. John Browning, of the Strand, is represented in Fig. 219. It has a single prism, made of glass, of great power in dispersing the rays. The prism is supported on a little stage, placed in the middle of a horizontal circular brass table about 6 in. in diameter. On the left is seen a tube, about 15 in. long, at the outer extremity of which is the slit, formed of pieces of metal very accurately shaped. One of these pieces slides in a direction at right angles to the slit, and, by means of a spring and a fine screw, can be very nicely adjusted, so that an opening of any degree of fineness can be readily obtained. In front of the slit is a small glass prism, with its edges parallel to the slit, but only half its height. The bases of this prism are formed of two sides of a square and its diagonal, and, as shown in the figure, one side is parallel to the face of the slit, and the other to the axis of the tube. Rays of light coming from a source on the left of the slit (as seen in the figure) will, therefore, enter this little prism, and be totally reflected (see page 399) by the diagonal surface, down the axis of the tube through the lower half only of the slit. This is the only office of this prism, which has nothing to do with the dispersion of the rays: the use to which it is put will be seen presently. It is fixed in such a manner that, when required, it can be turned aside with the touch of a finger, and the _whole_ length of the slit exposed. A peculiarity in these instruments of Mr. Browning’s is the admirable arrangement for determining the position of any line in a spectrum. For this purpose, the eye-piece of the telescope is provided with a pair of cross-wires, and the telescope itself, which is about 18 in. in length, moves in a horizontal plane round the axis of the circular brass table, from which an arm projects, carrying a ring into which the telescope screws. This arm carries a _vernier_ along the limb of the circular table, which is very accurately divided into thirds of degrees, so that with the aid of the vernier the angular position of the telescope can be read off to a minute, that is, to 1/60th of a degree. The arm carrying the telescope is provided with a screw for clamping it in any desired position while the readings are taken. On placing in front of the slit the flame of a Bunsen’s burner, the spectrum produced by any substance in this flame will, when the instrument is in proper adjustment, be seen on looking through the telescope, and the cross-wires being also in view, the point of their intersection may be brought into coincidence with any line of the spectrum, and the telescope being clamped in this position, the angular reading thus taken determines the position of the line. Thus, for example, the angular positions in which the principal Fraunhofer’s lines are seen having been observed and recorded, the angular position of any line in another spectrum will at once determine its position among the Fraunhofer lines; or the spectrum may be mapped by laying down the angular readings of the lines by means of a scale of equal parts. And, again, in the little prism in front of the slit we have the means of bringing two spectra in view at once, one being from a light directly in front, and the other from a light at the side. The two spectra are seen one above the other, and the coincidence or difference of their lines may be directly observed. When the instrument is in use, the prism and the ends of the tube are covered with a black cloth, loosely thrown over them, by which all stray light is shut out. The author has had in use for several years one of these instruments, and he cannot forbear expressing his perfect satisfaction with its powers, which he finds amply sufficient for all ordinary chemical purposes, while the accuracy of the workmanship is really wonderful, considering the very moderate price of the instrument.

The substances the spectra of which are most conveniently examined are the metals of the alkalies and alkaline earths. Small quantities of the salts of these metals, placed in a loop of fine platinum wire, impart characteristic colours to the flame of a Bunsen burner or to that of a spirit-lamp. For the examination of the spectra the former is to be preferred, as the lines come out much more vividly. Indeed, at temperatures higher than that of the Bunsen’s burner, such as in the flame of pure hydrogen, or in the voltaic arc, some substances give out additional lines. In Plate XVII., Nos. 2 to 9, is shown the appearance of the spectra produced by the Bunsen’s burner when salts of the metals are held in the flame in the manner already mentioned, and the spectra are examined with the instrument just described. One of the simplest of these spectra is that produced by sodium compounds, such as common salt. The smallest particle of this substance imparts an intense yellow colour to the flame, and the spectrum is found to take the form of a single bright yellow line—No. 3. It has been estimated that the presence of the (1/100000000)_th part of a grain_ of sodium can be detected by the production of this line. Indeed, the very delicacy of this sodium reaction renders it almost impossible to get rid of this line, for sodium is found to be present in almost everything,—a fact the earlier observers of spectra were not aware of, for they attributed this yellow line to water, which was the only substance they knew to be so generally diffused. If a platinum wire be heated in the flame of the Bunsen burner until all the sodium indications have disappeared, it suffices to remove the wire, and, without allowing it to come into contact with anything, to leave it exposed to the air for a few minutes, to cause it again to give the characteristic yellow colour when again plunged into the flame. This is due to the fact that the element is contained in all the floating particles which pervade the atmosphere. The spectroscope is not required to show the presence of the sodium on the platinum which has been exposed to the air, the colour imparted to the flame being plainly visible to the eye, and it needs only the Bunsen burner and 2 in. of platinum wire to prove the fact, and also to show that mere contact with the fingers is enough to highly charge the wire with sodium compounds. Any volatile compound of potassium gives the spectrum represented by No. 2, the principal lines being a red line and one in the extreme violet, the latter being somewhat difficult to observe. There is also a third rather ill-defined red line, and a portion of a faint continuous spectrum. Salts of strontium impart a bright red colour to the flame, and the spectrum they produce is shown by No. 6, in which are seen several bright red lines and a fainter blue one. Calcium, which also gives a reddish colour to flame, furnishes an entirely different set of lines (No. 5), and barium salt another, containing numerous lines, especially some very vivid green ones.

In all the cases we have named, and whenever bright-lined spectra are furnished by substances placed in the flame of a lamp, or in burning hydrogen gas, or in the intensely hot voltaic arc, there is evidence that the substances are converted into vapour or gas. We have already seen how hot solid carbon gives a continuous spectrum, while carbon in the state of gaseous combination gives most of the bright lines seen in the spectrum of coal-gas (No. 14). It is observed also that the more readily volatized are the salts, the more vivid are the bright lines they produce when heated in a flame. It must be understood that each element gives it own characteristic lines, that these are always in precisely the same position in the spectrum, that no substance produces a line in exactly the same position as another, however near two lines due to different substances may, in some cases, appear; and also, that however the salts of the different metals are mixed together, each produces its own lines, and each ingredient may be recognized. And this is done in an instant by an experienced observer—a mere glance at the superposed spectra of, perhaps, half a dozen metals, suffices to inform him which are present. There is also a peculiarity in this optical mode of recognizing the presence of bodies which gives the subject the highest interest, namely, the circumstance that the spectrum is produced and the bodies recognized, however far from the observer the luminous gas may be placed, the only condition required being that the rays reach the instrument.

Until Kirchhoff and Bunsen’s spectroscopic investigations, lithium was supposed to be a rare metal, occurring only in a few minerals. It happens that this substance yields a remarkable spectrum (No. 4), for it gives an extremely vivid line of a splendid red colour, accompanied by only one other, a feeble yellow line; and the reaction is of very great delicacy, for 1/6000000 of a grain can easily be detected, and an eye which has once seen the red line readily recognizes it again. A single drop of a mineral water containing lithium has been found to distinctly produce the red line, in cases where the quantity contained in a quart of the water would have escaped ordinary chemical analysis. The spectroscope has shown that lithium, so far from occurring in only four or five minerals, is a substance very widely diffused in nature. In the waters of the ocean, in mineral and river waters, in most plants, in wines, tea, coffee, milk, blood, and muscle, this metal has been found. Dr. Roscoe states that the ash of a cigar, when moistened with hydrochloric acid, and held in a platinum wire in the flame of the Bunsen’s burner, at once shows the principal lines of sodium, potassium, calcium, and lithium. Salts of lithium and of strontium both impart a rich crimson tint to flames, and it is hardly possible to detect any difference in these colours with the naked eye; but, as the reader may see on comparing spectra No. 4 and No. 6, the prism makes a wide distinction.

Matter for a very interesting chapter in the history of prismatic analysis has been furnished by the discovery of four new elements by means of the spectroscope. In 1860 Bunsen observed that the residue, after evaporation, of a certain mineral water, yielded spectra with bright lines which he had not seen before. He concluded that they were due to some unknown elements, and, in order to separate these, he evaporated many tons of the water, and was rewarded by the discovery of two alkaline metals, _cæsium_ and _rubidium_. The delicacy of the spectrum reaction may be inferred from the fact of a ton weight of the water containing only three grains of the salts of each of these substances. Rubidium gives a splendid spectrum, containing red, yellow, and green lines, and also two characteristic violet lines; while cæsium has orange, yellow, and green lines, and two very beautiful blue lines, by which it is easily recognized.

About the same time, Mr. W. Crookes discovered, in a mineral from the Hartz, another elementary body, the existence of which was first indicated to him by the characteristic spectrum it produces, namely, a single splendid green line (No. 8 spectrum). In 1864 two German chemists discovered, also in the Hartz, a fourth new element, which was detected by two well-defined lines in the more refrangible end of the spectrum—(see spectrum No. 9, in the plate). This metal was named Indium, in reference to the colour of its lines, and the names of the other three—cæsium, rubidium, and thallium, are also derived from the colours of their characteristic lines.

Although the reader may, from such representations of the spectra as those given in Plate XVII., form some idea of their appearance, he would find his knowledge of the subject much clearer if he had the opportunity of examining for himself the actual phenomena. We have already recommended the performance of certain easy experiments involving no outlay, but, in the matter of spectroscopes, carefully finished optical and mechanical work is absolutely necessary in the appliances. It fortunately happens that one eminent optician, at least, has made it his study to produce good spectroscopic apparatus at the lowest possible cost, and if the reader be interested in this subject, and desirous of trying experiments himself, he can, for a very moderate sum, be equipped with all the appliances for examining the phenomena we have described. He has only to procure, in the first place, a small direct-vision spectroscope, such as that represented of its actual size in Fig. 220, which is sold by Mr. Browning for twenty-two shillings; secondly, a Bunsen’s burner, a few feet of india-rubber tubing, two inches of platinum wire, and a few grains of the salts of lithium, strontium, thallium, &c. The whole expense will probably be covered by adding four shillings to the cost of the spectroscope, and the reader will then be in a position to see for himself the principal Fraunhofer lines, the spectra of the metals already referred to, and the absorption bands of the gases which have been mentioned, as well as the absorption bands in liquids which will be spoken of in the sequel.

The splitting up of a beam of light into its elements—which it is the office of the prism to produce—is accomplished by a single prism to a certain degree only. It separates the red from the green, for example; but the colours pass into each by insensible gradations through orange, yellow, and greenish yellow. If we allow the rays to fall upon a second prism after emerging from the first, the separation is carried further; the red, for instance, is spread out into different kinds of red, and so on with the rest. And the greater the number of prisms, the greater is the extension which is given to the spectrum. Now, just as by increasing the power of the telescope, new stars become visible, whose light was before too faint, and nebulæ, or stars which before seemed single, are resolved into clusters of individual stars—so, by increasing the power of the spectroscope by employing two, four, or more prisms, lines which appear single by the less powerful instruments are, in some instances, resolved into groups of lines, and new lines come into view, which before were too faint to show themselves. For example, if we view the Fraunhofer lines through a spectroscope like that in Fig. 220, but having two prisms instead of one, we shall see that the D line is not really a single line, but is formed of two lines close together. If we use greater dispersive power by employing a greater number of prisms, we shall observe with solar light that when these two D lines are sufficiently separated, several other lines make their appearance between them. In this way the number of dark lines in sunlight, which have been carefully mapped by Kirchhoff and others, amount to upwards of 2,000; and no doubt there are many more lines waiting a still more powerful instrument. Fig. 221 is copied from a large spectroscope made by Mr. Browning for Mr. Gassiot. It has nine or more highly dispersive glass prisms; the telescope and the tube bearing the slit have focal lengths of 18 in., the lenses having a diameter of 1½ in.; the telescope is provided with a slow motion for taking the angular position; and there is a third tube provided with a micrometer, by which the position of the lines can be measured to 1/10000th of an inch.