Page:EB1922 - Volume 32.djvu/585

Rh

quantity of energy, ev, expressed in appropriate units. If such an electron bombards a neutral atom, it is found that no change takes place until v reaches a certain value, when there is a sudden radia- tion of energy corresponding to a particular line usually a strong flame line in the spectrum of the bombarded atom. According to Bohr's theory, the energy in this particular radiation is equal to hv, where v is the wave-number of the line, and h is Planck's constant. Experiments with several elements show that this critical value of v is determined by the relation ev = hv. This means that the energy of bombardment has been just sufficient to remove the electron in the atom from its normal position to the next orbit, and, on its return, the electron restores the energy in the form of mono- chromatic radiation of the appropriate frequency. If v is expressed in volts, the wave-number of the emitted line is numerically equal to VX8IO2. The value of v when emission first takes place is known as the " resonance " or " radiation " potential. Further, by increasing v, it is possible to remove an electron from the atom altogether. If the energy, ev, in this case is equated to hv, the resulting value of v is found to be equal to the largest term in the spectrum usually the term IS, the limit of the principal series. This term would correspond to the innermost orbit of the electron, and the result suggests that the energy which must be applied to remove an electron from the atom is just equal to the energy pos- sessed by the electron when revolving in its normal position. The potential required to remove an electron from the atom is known as the "ionizing potential." It has been determined experimentally for a number of elements and has been found to be in complete agreement with the orbital energy as calculated from the largest term in the spectrum. The Bohr theory thus presents a simple picture of the processes taking place in experiments of this type.

The Stark Effect. The resolution of spectral lines under the influence of intense magnetic fields usually known as the Zeeman effect has been extensively studied for a large number of elements. Somewhat similar effects, but much greater in magnitude, pro- duced by an electric field, have been brought to light by Stark, 1 and examined in considerable detail by a number of workers. Nicholson and Merton * have shown that the Stark effect may oper- ate to an appreciable extent in an ordinary vacuum-tube discharge, causing a broadening of the lines. Both the Zeeman and Stark effects have been treated on the basis of the Bohr theory, 3 with some success.

Spectra and the Periodic Table. Attention is being drawn more and more to the relation of the spectrum to the periodic table of the elements. While it cannot be said that the relation is known with any approach to completeness, a number of important facts have been noted which may ultimately prove of great service in the interpretation of the table. It has long been known that, when doublets or triplets occur in the spectra, the wave-number separa- tions of their components (which are constant in the sharp and diffuse series) are approximately proportional to the squares of the atomic weights of the elements producing the spectra so long as those elements belong to the same family group. The Zeeman effect also is generally the same for lines of corresponding series in the spectra of elements belonging to the same group But perhaps the most comprehensive connexion of spectra with the periodic table is established by the " displacement law " of Kossel and Sommerfeld. 4 It has been observed that the " complexity " of the lines of a series i.e. their character as singlets, doublets or triplets is constant throughout a group, but varies from one group to another. The displacement law states that, when an element is ionized, the enhanced series take on the same type of complexity as the arc series produced by the element to the left (i.e. in the preceding group) in the periodic table. It is assumed that electrons arrange themselves round the nucleus in rings, and that spectrum phenomena are produced by electrons in the outer ring. If the outer ring con- tains an odd number of electrons, the spectrum will consist of doub- lets, while, if the number is even, the spectrum will show triplets and singlets. In the periodic table each element contains one outer electron more than its neighbour in the preceding group, while a group consists of elements having the same number of electrons in the outer ring. It follows that the removal of an electron from an

1 Ann. d. Phys., vol. xliii., p. 965 (1914), etc.

J Phil. Trans. A. vol. ccxvi., p. 459 (1916).

'Bohr, Danish Acad. Sc. iv., I, part ii., pp. i-ioo (1918); H. A. Kramers, Memoires Acad. Sc., Copenhagen, 8th ser., hi., No. 3, pp. 287-384 (1919) ; Epstein, Ann. d. Phys., vol. 1, pp., 489-520, 815-840 (1916).

4 Verh. Deut. Phys. Gesell. (1919).

element will make the outer ring similar to that of the immediate forerunner of the element in the table, and so make the enhanced lines of the first element of the same type of complexity as the arc lines of the second. Removal of a second electron would restore the arc type of complexity, for the number of outer electrons would again become odd or even, as the case might be. A second ioniza- tion is difficult to bring about in most cases, but with silicon it is probable that one, two, and even three electrons have been removed, step by step, thus making possible four distinct spectra. These appear to show the alternation of complexity required by the dis- placement law. The table above gives the types of series pro- duced by the neutral and ionized elements of the various groups, so far as they are known at present.

The spectra of the higher groups are much more complex than those of the lower ones. Their series, if they possess any, are possibly of a different type from those with which we are familiar. The dis- placement law, however, suggests that, by repeated ionizations, series and therefore terms might be detected in such spectra, of the same kind as those of the groups of elements on the left. But since, with each successive ionization, the term constant, N, is multiplied in the ratio I 4:9:16 etc., the chief series lines might tend rapidly to approach the far ultra-violet and become difficult to observe.

Band Spectra. Several band spectra have been studied in further detail, but it does not appear that any very fundamental advance in our knowledge of the structure of these spectra has been made. The discovery of a band spectrum of helium, 6 however, is probably of considerable importance. It has been shown by Fowler 6 that, while the individual bands follow the ordinary laws of band spectra, the heads of some of them are arranged in accordance with the laws of line series. In this respect the helium bands appear to be quite unique. Unlike the lines of helium, the bands have not yet been traced in any celestial source.

The Solar Spectrum. A striking feature of continued work on the solar spectrum is the identification of a large number of faint lines with lines composing the bands of certain compounds, in addition to the band lines of carbon and cyanogen previously recognized by Rowland and Lockyer. The peculiarities of the region about the G group of Fraunhofer have been shown by Newall 7 to be due to the absorption of the well-known hydro-carbon band X43I5, and the group P has been found by Fowler and Gregory 8 to include the strong ultra-violet band of ammonia having its maximum near X33&O. In addition, the band of luminous water vapour beginning at X3O64 has been found by Fowler 9 to be present in the solar spectrum. A large number of previously unknown solar lines have thus been accounted for, and it is not improbable that the thousands of faint lines which remain unidentified may eventually be traced to other band spectra.

An interesting application of modern theories of spectra to solar problems has been made by M. N. Saha. 10 On the reasonable assump- tion that the composition of the sun is essentially the same as that of the earth, it remains to account for the absence of spectral indica- tions of many of the elements. Dr. Saha urges that the varying representation of different elements arises from the varying response of these elements to the solar stimulus, depending upon the struc- ture of their atoms, and the consequent difference in their ionizing potentials. Caesium, for example, has a low ionizing potential and is considered to be completely ionized in the sun, so that the familiar lines do not appear, while the chief lines of the ionized ele- ment are out of range. In contrast, sodium has a higher ionizing potential and is only partially ionized in the sun, so that the lines of the neutral atoms appear strongly. Other elements, such as neon and argon, have very high ionization potentials, and are not excited at all. Dr. Saha finds support for his views in calculations of the percentage ionizations of various elements at different temperatures and pressures, and it is possible that the peculiarities of the solar spectrum may be satisfactorily explained by these considerations.

Stellar Spectroscopy. Our detailed knowledge of the spectra of the stars has been greatly advanced by the use of the large tele- scopes which have been erected, and considerable progress has also

W. E. Curtis, Proc. Roy. Soc., Ixxxix., 146 (1913); E. Goldstein, Verh. Deut. Phys. Ges., xv., 10 (1913).

Proc. Roy. Soc., xci., 209 (1915). Monthly Notices R. A. S., Ixxvi., 640 (1916). Phil. Trans. A ccxviii., 351 (1918). Proc. Roy. Soc. A. xciv., 472 (1918). 10 Phil. Mag. xl., 809 (1920).