The simple fact is that observing a weak change in diode voltage corresponding to a change in electron energy is not enough to prove that Cerenkov radiation was indeed being observed. What I tried to accomplish with this project was, first and foremost, to show that the microwave diode chosen for this project would respond in the vacuum. This was determined by testing the diode's responsiveness to an external microwave generator. Even though the diode's voltage reacted quite strangely, it was still 100% responsive to the external generator. After this was shown, I was able to move on to generating Cerenkov radiation. Once the sensitivity of the data acquisition was improved to the point of being able to detect a signal as small as the Cerenkov radiation was predicted to be, observing this radiation was merely a matter of being able to reliably generate a 3000 V electron beam. This was taken care of by further insulating the control box. The signal generated in response to the energy of the electron beam was small, just as predicted, but not so small as to be hidden by the resolution of the experiment.

I’d like to see a deflection system (for the new design see the Apparatus page, for the old design see Appendix D) finished and installed. If it works as designed, the beam pulsing should be clearly separated from the noise. I have a real-time fast Fourier transform program in LabWindows that could be easily adapted for use I this project. By allowing for the live plotting of the FFT of the diode signal, one could tune the chopping frequency of the electron beam and see the radiation spike in the frequency spectrum move in step. Such a demonstration, in addition to similar trials without the dielectric present (running the gun at both voltages above and below the 3000 V threshold would allow one to investigate the effects of just the beam pulsing on the diode voltage) would provide definitive evidence for radiation. Coupling that with a better theoretical prediction for the behavior of the radiation under such conditions, and one could safely say that what was observed was Cerenkov radiation.

Probably the easiest way to improve the data precision further would be to pass the diode signal through an amplifier before it is recorded by the multimeter. The resolution of this meter was a primary impedance to the interpretation of the data. Of course, the two data points per second limitation imposed by the multimeter will have to be overcome if any sort of high frequency chopping or bunching is considered. An A/D (analog to digital) board can be attached to the computer to allow for faster data retrieval, but doing so introduces a considerable amount of noise. When looking for such a weak signal, noise is a serious threat. A new data acquisition system will probably have to be designed, one that has a high precision, fast sampling and most of all, low-noise voltage sensor. Of course, such a change is beyond both the scope and the budget of this project.

A more stable diode mount would also seem appropriate. Several designs have come across my mind, but none are very attractive. The diode’s inherent polarization sensitivity should be exploited. To do so, the diode must be mounted so that it can be easily rotated by 45°, 90°, and 135°. This, in addition to wavelength information (to be determined by some other method), allows one to define the four Stokes parameters for the radiation. One would also want to move rotate the diode through the plane of the radiation to verify its distinct axial distribution. By doing so one can fully characterize the radiation and gauge quite well whether what is being observed is Cerenkov radiation or something else. Personally, I am very interested in seeing what polarization the radiation has. By simple analysis, the radiation should be linearly polarized, but I have not seen any treatment of the theory behind this interaction that actually acknowledges that fact.

To increase the radiation intensity further, a second dielectric slab could be placed on the opposite side of the electron beam from the one currently in place. Theoretical work in this area [6], based on the work by Linhart [4], indicates that this method utilizes a larger portion of the electron beam than the single sheet method.

This project was initiated with the goal of making a Cerenkov laser, that is, a device that generates coherent radiation via the Cerenkov effect. Some models that have been explored involve placing the dielectric plate and the beam either in between two conducting plates [7] or putting the dielectric sitting on top of a conducting plate [8,9]. There is also much focus on curling this model into a cylindrically symmetric design [10, 11, 12], allowing for an even more uniform beam contact (remember, the distance between the beam and the dielectric is a primary factor in the properties of the radiation produced). What’s more, X-band Cerenkov masers using the cylindrical model have also been theorized [13].