On a search for radiation accompanying the scattering of comparatively slow electrons at the surface of incandescent solids

In a earlier paper published in these ‘Proceedings,’ I gave a report of some measurements of the energy losses suffered by electrons scattered from in-candescent solids. The substances tested were copper, silver, gold, platinum and the oxides of magnesium, calcium, strontium and barium. In every case the energy distribution curves for the electrons after scattering were found to exhibit characteristic maxima, different for different substances. The energy loss by the scattered electrons, indicated by the position of any particular maximum, remained the same as the initial energy of the bombarding electrons was varied within wide limits. This fact, together with other evidence, led to the conclusion that the corresponding scattering processes must involve the excitation of definite energy levels pertaining to and characteristic of the scattering substance. Particularly the three last-mentioned oxides were found to yield several rather definite and persistent maxima. The present conceptions as regards the mechanism of light emission would make one expect such emission to occur as a result of, or in connection with, the return of the disturbed scattering system to its normal state, the frequency of the radiation being related to the characteristic energy values concerned by the Bohr relation. As pointed out in the previous report, however, existing experimental data by different investigators fail to show any evidence of this radiation for bombarding voltages of the order 0-1000 volts, at any rate in the case of metals. In view of the importance of establishing the true nature of these energy levels, I have made a fresh attempt to look for radiation from solid substances bombared with electrons of velocities of the same order as was used in the scattering experiments referred to. For certain reasons I had to confine the investigation to the region of the visible and the ultra- violet above 2300 A., which is accessible by a quartz spectrograph. However, some of the characteristic losses previously observed were small enough to correspond to this region, and transitions in which a pair of the measured energy values were concerned should often yield frequencies in the range studied. The chief difference of the present investigation from earlier work by others, rests in the fact that the bombarded substances are maintained at incandescence during the whole time of exposure. This method was adopted since the electron scattering experiments had shown, that allowing the target to cool down often caused the characteristic maxima to diminish and gradually disappear, the distribution curve being replaced by a new one, the same for all kinds of targets. The details of this change were such as to suggest that an adsorbed film of some substance, present as a vapour of low pressure in the tube in spite of the high vacuum, was slowly formed on the surface of the target and that this film, once formed, was largely responsible for the scattering of electrons with small energy loss observed with cold targets. To enable heating of the substance under investigation this was therefore deposited on a strip of thin platinum foil, which could be heated by the passage of an electric current from a battery of accumulators. However, the dissipation of heat from the bombarded area of the long strip was so slow that with the heavy bombarding currents used the foil was often maintained at incandescence by the discharge alone.