Fig. 1. Schematic diagram of a photoklystron. reinforcing the oscillations in a tuned LC circuit (2). The time average of the electron energy imparted to the rf field is nonzero and positive because the electrons have been bunched. If reflex klystron theory could be scaled to very low electron energies, a few eV, and if a spread in energies comparable to that expected from a photoemitter could be accommodated, it seemed possible that the photoelectron kinetic energy could be used to drive oscillations. If we chose the appropriate frequency for oscillation it appeared that conventional phototube photoemitters could provide sufficient photocurrent to sustain oscillations. To demonstrate that the concept was sound we had custom manufactured a proof-of-concept test model. Fig. 1 is a schematic of this device. The photoemitter is a standard S-4, CsSb photocathode deposited on a glass window. The two grids are 0.8 cm apart and separated 1 cm from the photocathode and reflector electrode. The grids are coupled by an air core inductor. Fig. 2 is a photograph of the test photoklystron. II. TEST RESULTS The initial tests were conducted with a small accelerating bias voltage on the grids, positive relative to the photocathode. We found no trouble obtaining a variety of modes of oscillations in the frequency range from 8 to about 240 MHz. To our surprise, most of these modes (combinations of accelerating and reflection bias voltages) did not correspond to what would be expected from reflex klystron theory. For example, we found that the photoklystron would oscillate with the reflection voltage less than the accelerating voltage. Fig. 3 is a mode chart showing the unconventional modes. Throughout these tests a small tungsten microscope lamp, producing about 10 mW of light at the photocathode, was used. With no tuned antenna but with the inductor serving as a poor magnetic antenna, harmonics of the rf signal are detectable with a small transistor radio several meters away. Oscilloscope and rf voltmeter measurements indicate that the oscillations are strong and start spontaneously. A search coil pickup has been used to measure the output power under a 50 fl load. The measured electrical efficiency is about 1%. Using an rf voltmeter, the output voltages for the strongest modes are about 2.0 V rms. An overall efficiency (including light energy input) for this particular tube is not very meaningful since the photocathode has been damaged and its quantum efficiency is now less than 1%. However, assuming an
RkJQdWJsaXNoZXIy MTU5NjU0Mg==