is essential for net positive energy exchange with the rf field. A negative space charge cloud near the photocathode may also provide the repelling voltage necessary for multiple oscillations in the self-oscillation mode. The self-oscillation mode cannot be fully understood in terms of the simplified analysis represented by Eq. (1). A fully self-consistent model including space charge is required. V. THE PROSPECT FOR HIGHER FREQUENCIES The present test model photoklystron with an 8 mm grid separation self-oscillates at about 5 MHz or lower, and in the biased mode, it has been operated at up to 240 MHz. Somewhat higher frequency oscillations are presumably possible by winding smaller inductors. To make the leap to ultra-high or microwave frequencies with the grid type device would require changes in the grid separation distance. From Eq. (1) we see that microwave frequency operation requires a substantial reduction in grid separation distance. One design with a grid separation of 0.5 mm was run on the computer and found to provide oscillations at 2.45 GHz, however, the parasitic capacitance of the grids at this distance is prohibitively high. Clearly the discreet elements must be replaced by a resonant cavity at these frequencies. This appears possible, however, additional research is necessary to determine if cavities with such narrow gaps are practical. VI. ADDITIONAL RESEARCH A problem found in some previous efforts to utilize photoelectric free electron devices for de solar energy conversion has been the low quantum yield (3). The problem is that thick photocathodes which provide a high photon interaction probability leave a long escape path for the photoelectron. Negative electron affinity photoemitters have been tested which have quantum yields approaching 50% over a substantial portion of the visible spectrum, however, they are carefully prepared crystal surfaces. To solve the problem of low quantum yield, a photoklystron design may be possible which allows more than one photoemission surface to contribute to the electron beam. It has been found that oscillation modes exist in which the reflector electrode voltage is the same as the photocathode voltage relative to the grids. In this case, the electric fields are fully symmetric about a plane halfway between the grids. The reflector electrode can now be a photoemitter and contribute an independent stream of photoelectrons generated by photons which pass through the front photoemitter. Moreover, if the reflector electrode photoemitter is backed by a mirror, still unused photons traverse the tube backwards and can further liberate photoelectrons. In this way, it may be possible to design a photoklystron with a very high effective quantum yield. An additional area for future research is the determination of an optimum photoemitter combining high quantum yield, stability and low cost. In order to maximize the effectiveness of the conversion device, photoemissive materials must be used which possess the lowest possible work function. A systematic search for stable and economical materials is presently underway. At this time, certain interstitial transition-metal compounds coated with alkali metals and their oxides are being tested for their photoemissive properties.
RkJQdWJsaXNoZXIy MTU5NjU0Mg==