SPECTRA AND THE PERIODIC TABLE General Relationships.—The spectra of a large number of elements have now been at least partially analysed, and it is possible to draw some general conclusions with regard to the rela tions between spectra and the periodic table. That some relations exist was indeed evident many years ago, when it was discovered that the spectra of elements in certain chemical families showed marked resemblances to one another. Thus the spectra of the alkali metals all show well-developed series of doublets; those of the alkaline earths show similar series of singlets and triplets, and so on. These now appear to be special cases of a general rule that elements of the same chemical family yield spectra with the same multiplicities. Furthermore, in progressing across the table from left to right, the multiplicities encountered are alter nately even and odd, as may be seen by the following example taken from the third row of the table :— The figures preceding the chemical symbols of the elements are the atomic numbers. Within a chemical family there is, in general, an increase in separation of component terms, roughly propor tional to the squares of the atomic numbers, and at the same time a movement of corresponding lines towards the region of greater wave-lengths. It will be noticed that the low multiplicities occur mainly among the elements at the left of the table, and it was among these elements that series were first detected. The existence of complicated multiplets was first discovered by Cata lan in the spectrum of manganese, which is derived from quartet, sextet and octet terms.
Enhanced Spectra.—It has already been mentioned that each element possesses a number of spectra, which are produced suc cessively by gradually increasing the exciting stimulus. They are denoted, in the case of a typical element, Z, by the symbols, ZI, ZII, ZIII, etc. The type of structure which has been described appertains to any of these spectra, but we meet with an important modification of the series formula in passing from one of them to the next, viz.—the value of the Rydberg constant, which is approximately R=109678 for ZI, becomes approximately 4R for ZII, 9R for ZIII, and in general, for the nth spectrum. Since the other constants in the expression for the terms have the same general order of magnitude for all the spectra, it follows that the term values increase rapidly from one spectrum to the next.
There are strong reasons for believing that the number of such spectra which an element yields is equal to its atomic number. Thus, no increase of stimulus can produce a successor to the red spectrum of hydrogen (HI), or more than two spectra of helium, three spectra of lithium, etc. Owing to practical difficulties, it is only for the lightest two elements—hydrogen (H), and helium (He)—that lines of all the possible spectra have been observed, but the results in these cases illustrate a very important rule which from theoretical considerations, is believed to be quite general, namely : if n be the atomic number of an element, then its nth spectrum is as simple as the first spectrum of hydrogen, consist ing only of terms of the form lel?' / , where R' is almost, but not quite, equal to R for hydrogen, and is very slightly different for each element. The lines of HeII were observed before this
rule was known—first in the spectra of heavenly bodies and afterwards in the laboratory by A. Fowler—and, from the sim plicity of the terms, were mistakenly attributed to hydrogen. It was. not until the foundations of the theory of spectra were laid by Niels Bohr in 1913 that the true origin of the lines was indicated. Their connection with helium was afterwards confirmed by experiment.
Displacement Law.—There are certain significant relation ships between the successive spectra of the same and neighbour ing elements which must now be considered. Passing along the sequence, ZI, ZII . . . , we meet with an alternation of multi plicities similar to that already noted in the spectra of elements forming a row of the periodic table. The resemblance is not acci dental, but is implied in what is known as the displacement law of A. Kossel and A. Sommerfeld, which in its generalized form states that the spectrum ZII will have the same multiplicity and general character as the first spectrum, VI, of the element pre ceding Z in the table; and further, that ZIII will resemble YII and XI—the first spectrum of the element preceding Y; and so on. The correspondence between these similar spectra is very close as regards multiplicities and the types of the most prominent terms, but it does not extend to the numerical values of the terms, which increase rapidly with the stimulus required to produce the spectrum.
As a particular example, consider the successive elements— sodium (Na), magnesium (Mg), aluminium (Al) and silicon (Si), details of which have been investigated by A. Fowler and F. Paschen. Here the spectra, NaI, MgII, AlIII and SiIV are strik ingly similar to one another, and may profitably be compared with the sequence of spectra of the alkali metals, NaI, KI, RbI, CsI. These two sequences have the spectrum NaI in common. They both consist of doublets, the separations between the com ponents of which increase rapidly with advancing position in the sequence. But whereas, in the spectra of the alkali metals, corresponding lines move towards the red with increasing atomic number, the drift in the other sequence is towards the violet. Moreover, both the increase in separation of component terms, and the changes of position in the spectrum, are more regular in the NaI, MgII . . . sequence, as may readily be seen from fig. 17, in which the doublets are shown on a scale of wave-number. It appears as though the atomic systems which are responsible for the spectra NaI, MgII, etc., are more regu larly related to one another than the atoms of elements of the same chemical family.