Other Spectra

In ordinary absorption the electron is ejected outwards by light of the proper frequency in the incident beam. A similar ejection can be brought about by electron bombardment, as in the experiments of Franck and Hertz and others. (See art., RESONANCE POTENTIAL.) In these and similar experiments the energy is usually imparted to the bombarding electron by an electric field, and the potential difference through which the elec tron has to fall to acquire sufficient energy to excite the atom from its normal state to the next possible state is known as the resonance potential of the atom (q.v.). Similarly, the poten tial difference necessary to enable the bombarding electron to ionise the atom is known as the ionisation potential of the atom. Clearly the ionisation potential is a measure of the energy of the atom in its normal state, and is, therefore, proportional to the largest term in the spectrum. It is given in volts by dividing the largest term, expressed in wave-number units, by 8,102.

Interpretation of Terms.

In applying the ideas of the Bohr theory to the facts concerning the spectra of the elements already described, the outstanding problem is to interpret the four num bers which specify a spectroscopic term, namely, nrki. The con that each of these numbers may be identified with a definite characteristic of an electron orbit is clearly incomplete, inasmuch as the movements of a point charge, such as an electron is conceived to be, can be completely defined by three quantities, whereas observation shows that four are to be accounted for. This difficulty has been partly overcome by assuming that an electron can rotate about an axis, thus giving it an additional "degree of freedom," but the more advanced developments of the theory seem to indicate that even this amplification is insufficient to interpret all the facts of observation, and the whole conception of electron orbits may have to be regarded as merely a symbolical one. For that reason the phrase "energy of an orbit" is often discarded in favour of the more general one, "energy level of the atom." The simple conception of spinning electrons in orbital motion, however, gives a fairly clear picture of what is happening when a spectrum is emitted.

In the earlier days of the theory, it was supposed that the spec trum was generated by the transitions of a single electron between different orbits, and that the k value of a term (i for S, 2 for P . . .) was always simply related to the corresponding orbit ; so that, for example, if the term were 3P the orbit was consid ered to be specified by 32. The principal quantum number n was

then regarded as an index of the major axis of the orbit (sup posed in the general case to be an ellipse), and k was considered to represent the minor axis, nk thus indicating the size and shape of the orbit associated with a particular term. The interpretation of r and j, however, remained somewhat vague and incomplete. An important indication of the need for further development of the theory was also subsequently given by the analysis of the more complex spectra, it being then found that the term repre senting the normal state of an atom was sometimes of a type which was very unlikely to represent the orbit of the most loosely bound electron in the unexcited atom.

In identifying the quantum numbers of the terms with the elements of a single electron orbit, it is implied that when an atom is excited, only a single electron is disturbed; but it is obvi ously conceivable that the exciting agency might disturb two or more electrons simultaneously, with the result that the energy level of the atom might be altered in a way inexpressible in terms of the movements of a single electron. Evidence that this actu ally occurs was first brought to light by Henry Norris Russell and F. A. Saunders in a study of the spectrum of calcium. They showed in effect that terms existed in the spectrum larger than the term corresponding to the ionisation potential of the element. This meant that the atom contained more than enough energy to ionise it if all the energy were given to one electron, and it naturally followed, since the atom was not ionised, that the energy was shared between two or more electrons.

A more general problem than that so far considered was thus opened up. A spectrum term could no longer be regarded as neces sarily characterized by the elements of a single orbit; it had to be conceived as measuring, in the most general case, the sum of the energies of all the electrons in the atom, each of which might be changed when the atom was excited to a new state. Fortunately it appears that, with the ordinary methods of exciting spectra at least, only a few of the electrons are disturbed, and remarkable progress has been made in correlating the distribu tion of electron orbits with the spectrum terms. It is now possible to state what terms correspond to any given configuration of electrons, and so to predict the possible terms in the spectrum of an element when the probable orbits of its electrons are known.