82 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 48, NO. 1, JANUARY 2001 Transient Ion Disturbances in Traveling Wave Tubes William Tighe, Dan M. Goebel, Fellow, IEEE, and Charles B. Thorington Abstract—It is well known that the presence of ions in the electron beam of a traveling wave tube (TWT) can lead to periodic variations in the output power, phase and the body (or helix) current. This has been referred to as ion noise or jitter. Recently, we have observed a different form of jitter, and while it is still observed as a small variation in the TWT output (typically 0.5 dB in power and 2 in phase), it is not periodic. We refer to this phenomenon as random jitter, since its random nature in time is a defining characteristic. Other characteristics include a relatively fast onset ( 1 ms) and slow ( 500 ms) recovery. It was found that random jitter was due to the spurious release of extremely small amounts of trapped gas inside the TWT. The source of the gas was identified and the problem was resolved. The observed level of fluctuations in power and phase had no effect on digital traffic and the small quantity of gas was found to have no measurable impact on cathode life. Index Terms—Communication systems, electron tubes, ions, noise. I. INTRODUCTION E FFECTS due to ions in vacuum electron devices (VEDs) have been observed from their early history [1]–[3]. In spite of the quality of the vacuum in the device, some level of background neutral gas will exist and with the interaction of the electron beam, a population of ions will be created. In some cases, these ions may be lost rapidly or be small enough in number that they do not significantly influence the performance of the device. In traveling wave tubes (TWTs), however, electrostatic traps allow a build-up of the ion population resulting in an observable interaction with the electron beam. For example, the trapping of ions in the central potential depression of the beam is known to lead to “ion focusing” of the electron beam [4], [16], [17]. In the mid-1950s, Cutler [5] described noiselike modulations of the electron beam that were explained by the presence of ions in the beam. Higher frequency phenomena were caused by ion plasma oscillations and identified through the presence of ion sidebands to the driving radio frequency (RF). Cutler’s mechanism for lower frequency oscillations (up to 100 kHz with torr) was a relaxation oscillation in a device pressure of which ions are trapped by the space charge of the beam. When a critical ion density is reached, the trapped ions are released. The trap is then refilled by further ionization and the process repeats. Shortly after Cutler, alternative mechanisms [6]–[8] were put forward to explain the low frequency oscillations. These involved the formation and decay of a virtual cathode or a Manuscript received July 26, 2000; revised September 7, 2000. This work was supported by HED management, including T. Fong, J. Porzuki, J. Dayton and M. Patterson. The review of this paper was arranged by Editor D. Goebel. The authors are with Boeing Electron Dynamic Devices, Inc. (formerly Hughes Electron Dynamics), Torrance, CA 90505 USA (e-mail: william.g.tighe@boeing.com). Publisher Item Identifier S 0018-9383(01)00307-0. bistable behavior with the beam shifting between two operational modes. Additionally, dielectric rod charging has been proposed as a potential mechanism. In this case, scattered electrons are assumed to collect on dielectric rods and the material charges up until a path to ground develops. In some cases, conductive coatings on the rods seemed to have a significant effect on the oscillations [9]. Over the course of time, the frequency range associated with jitter appears to be coming down as the base pressure of the device decreases. In very early days, frequencies up to 100 kHz were noted. In the 1970s, this was more typically limited to 1 kHz while frequencies are at present generally found to be less than 100 Hz. More recent investigations [4], [10]–[13], [16]–[18] tend, in large part, to support the original proposals of Cutler. While specific details may require further study, the primary mechanism driving these low frequency oscillations appears to be a relaxation oscillation caused by ion formation, trapping and de-trapping in the VED. At Hughes Electron Dynamics (HED) 1 , the low frequency oscillation is now termed periodic jitter. This is to distinguish it from a different, recently discovered form of jitter. This latter form is termed random or spiky jitter and is characterized by sudden, random noiselike bursts on the helix current and the output power and phase. Though similar in many respects to periodic jitter, the main differences arise from the source of the ions. For the TWTs used in our studies, it is the interaction of the electron beam with the background gas that provides a continuous source of ions for periodic jitter. In random jitter the ions are produced from the beam ionization of random gas bursts from the cathode. In this paper, the results from a detailed investigation into random jitter will be presented. To provide some background into ion trapping and detrapping processes involved in both forms of jitter, a discussion of our studies of periodic jitter will be presented in Section II. This will be followed by sections on the characteristics of random jitter events, their effects on the TWT, and the cause and elimination of the phenomenon. II. CHARACTERISTICS OF PERIODIC JITTER There is a long history and experience with periodic jitter [4], [16], [17]. It has been observed in virtually all periodic permanent magnet (PPM) focused TWT designs and operating frequency bands, as well as in solenoidally focused TWTs [13] and klystrons [10]. Attempts to explain it as rod charging have not been generally successful. It has been determined that helix jitter is a benign issue that reduces with time, presumably as the vacuum in the tube improves with time. The general solution 1Hughes Electron Dynamics is now Boeing Electron Dynamic Devices, Inc. 0018–9383/01$10.00 © 2001 IEEE TIGHE et al.: TRANSIENT ION DISTURBANCES Fig. 1. Sketch of the potential distribution in a typical TWT showing the large space charge depression of the central potential and possible locations of axial, electrostatic ion traps. has been to operate the tube for extended periods of time until the level of jitter reaches an acceptable level. Jitter is very sensitive to focusing of the electron beam and at various times during processing the phenomenon can be reduced by careful focusing of the electron beam, though it may return at a later point in processing. Placement of magnetic shunts at various locations, apparently related to the scallops in the beam, have been observed to strongly affect the nature (frequency and amplitude) of the jitter. The presence of periodic jitter has also been found to be dependent on the environmental temperature and, while the tube pressure was not measured directly, it has been noted that jitter appears to occur only in a limited pressure region. The source term for periodic jitter is the rate at which ions are produced and made available to an ion trap. Fig. 1 shows a sketch of the potential distribution along the axis of a TWT. As is well known, the central potential provides a radial ion trap, that is; ions electrostatically confined to the central part of the electron beam are unable to move outward to the walls of the tube. Without any other variations in the potential, the ions would, however, be free to move axially toward the cathode and the collector. As indicated in Fig. 1, traps along the length of the TWT inhibit this motion. Because the motion is prevented along the length of the tube, these traps are called axial ion traps. Once a trap is filled, the ions can be released axially or, if the ion number density is sufficient to neutralize the central potential, radially to the wall. There can be numerous ion traps in TWTs and their number and magnitude can vary significantly from tube to tube. In many TWTs, a positive anode is placed near the entrance to the helix to create an electrostatic barrier and prevent ion bombardment of the cathode. Furthermore, in many designs the radial location of the ground potential wall moves significantly closer to the beam edge as it enters the helix. As illustrated in Fig. 1, this axial variation in the radial location of ground potential coupled with the positive anode produces a significant axial ion trap at this location. In designs utilizing a severed helix [4], [16], [17], a similar axial trap may exist in the sever region near the middle of the tube. In addition to these two large traps, a series of smaller, axial ion traps associated with the scalloping of the electron beam [4], [16], [17] can form along the helix region of the TWT. The 83 scallops are due to a focusing mismatch at the entrance to the helix combined with the confining magnetic field structure. As the beam radius changes due to scalloping there is a change in charge density and a variation in distance to ground, leading to the formation of axial ion traps. In a TWT, then, there is significant potential to trap ions along the length of the device. While scallop traps may be minimized with careful focusing technique, and other traps might be modified by tube design considerations, complete removal of the traps may be impossible. Computer simulations were undertaken at HED to better understand the trapping and de-trapping of ions in a PPM focused TWT. Here we provide a brief description of the model and its results, the details of which are reported elsewhere [14]. The simulation is based on a two-dimensional (2-D), time-dependent, electrostatic, particle-in-cell (PIC) code. Ions can be handled in two ways. In one case, populations of multiple ion species can be introduced at a user specified time step and beam location. Alternatively, ions can produced throughout the beam from the solution of the ionization equation. The main results from the modeling are 1) the observation of trapped ions in electrostatic wells due to both the scalloping of the beam and the axial variation in the radial location of the ground potential at the entrance to the helix, and 2) a very rapid, radial release of the trapped ions after a threshold level for the population density is exceeded. There is no attempt to model the time dependence of either the ion build-up time or the ion release time within the code. Nevertheless, for the conditions in the simulation, the build-up time due to ionization would be expected to be several orders of magnitude longer than the ion release time. The ion density threshold for detrapping was determined by the artificial injection of increasing numbers of ions until an unstable situation developed. The threshold density was determined to be approximately half of the space charge density of the electron beam. Neutralization of the trap and additional processes [14] such as ion–electron viscous forces and plasma oscillations then provide the ions with sufficient energy to move rapidly radially to the wall. The time required to produce the needed population of ions was calculated separately from the fine-time-step beam runs using the ionization rate equation. For the conditions expected in present day TWTs, the estimated time required to replenish the ion population would be in the range of 0.25 to 2.0 s [14] yielding oscillation frequencies from 0.5 to 4.0 Hz, in good agreement with that observed experimentally. From the modeling, one can qualitatively describe the behavior of the tube due to the presence of ions. Initially, as the ion population builds up in the trap, one would expect a slow change in the beam location and/or shape. This would be followed by a very fast return to the initial condition due to the sudden radial release of ions once a threshold density has been reached. The process would then repeat as a relaxation oscillation. The precise waveform that results from the relaxation oscillation will depend on many tube parameters. The depth of the potential well will affect the number of ions trapped and released, while the ionization rate will depend on the beam size and current and the density of background neutrals. The response of the 84 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 48, NO. 1, JANUARY 2001 Fig. 2. Strip chart of the helix current from a TWT exhibiting simple periodic jitter. Fig. 3. Strip chart of the helix current showing a somewhat more complex waveform associated with periodic jitter. Two distinct oscillation frequencies are observed suggesting simultaneous relaxation oscillations from two different ion traps. tube to the presence and release of these ions is difficult to predict, since it depends on the location of the traps, the number of traps, and the sensitivity of the tube design to the beam disturbance. Over the years, periodic jitter has been observed for a wide range of tube parameters. Here, we present two examples taken from a recent study of periodic jitter in C-band space TWTs. These tubes were typically operated with a beam current of about 50 mA at a beam voltage 5 kV. The tubes utilized a positive anode in the gun to form an axial ion trap at the entrance, and contained another ion trap in the sever region of the helix. The estimated depths of the potential traps were 20 to 30 eV at the helix entrance and 1 to 3 eV in the scallops. The actual trap size and shape depends on the tube construction, operation and focusing. There would be expected to be significant variation in these characteristics from tube-to-tube. In fact, during the processing of a single tube, one would expect changes in both the trap sizes and the ionization rates as the focusing and vacuum conditions varied. While the dependence of periodic jitter on beam current, beam voltage, anode voltage, collector voltages were examined, this data and its analysis will be the subject of a later paper. The results generally support the model presented here, but are not critical to the understanding of random jitter which is the main topic of this paper. Fig. 2 shows a strip chart of the helix current from the C-Band TWT exhibiting periodic jitter. The sawtooth waveform in this example is fairly typical, although more complex shapes have been observed. In this simple example, the jitter has a frequency of approximately 1.5 Hz. The extreme regularity of the oscillation is perhaps its most dominant characteristic. It is possible to identify in this waveform a slow ion build-up portion of the cycle that is followed by a relatively rapid release. Following the release, the helix current returns to its original level and the slower build-up phase begins again, producing the sawtooth waveform. The response of the TWT to the release of ions is occasionally more dramatic, resulting in a significant, short-lived spike at this point in the waveform. In such cases, the build-up phase is often masked by the release event and the waveform form is characterized by a regular series of ticks. Consequently, this type of jitter is sometimes called ticking. The waveform of periodic jitter is sometimes more complex than that shown in Fig. 2. In some cases, two or more oscillation frequencies can exist at the same time. A series of “double spikes” has also been observed. Fig. 3 shows an example of a strip chart in which two frequency components can be identified. This sort of behavior suggests the presence of more than one ion trap. One interpretation of this waveform is that the high frequency, low amplitude oscillation arises from a shallower ion trap while the higher amplitude, lower frequency oscillation comes from the action of a deeper trap releasing more ions less frequently. In these cases, the oscillators may not be “coupled” but are likely “connected,” as ions would be expected to move axially from one trap to the next. Finally, it is important to note that in all TWTs exhibiting jitter at HED, the response of the tube has been such that the disturbances have been clearly associated with the presence of ions trapped in the beam. Extensive life testing has demonstrated that the phenomenon is benign with respect to the life of the device, and effectively results in only small changes to the “spurious noise” level in the tube output. III. CHARACTERISTICS OF RANDOM JITTER From time to time, over periods of many years, periodic jitter has been observed and treated in TWTs. Recently, TWTs at HED began displaying a new type of disturbance that appeared as random, spiky events observed in the helix current and the output power and phase. A strip chart capturing several of these events is presented in Fig. 4, where the power spikes were typically less than 0.1 dB in magnitude and the helix current spikes were about 10% of the total helix current. The events were infrequent, generally occurring about once per minute to as seldom as once per hour. The magnitude and waveshape of these events were similar in many ways to periodic jitter, strongly suggesting that they were caused by the sudden release of ions in the tube. However, the random nature and extremely low occurrence rate differentiated them from periodic jitter and the phenomenon was termed “spiky” or random jitter. A typical, single event is shown in Fig. 5. A rapid initial excursion on the order of 1 ms, followed TIGHE et al.: TRANSIENT ION DISTURBANCES Fig. 4. Strip chart showing the output power and helix current of a TWT exhibiting random jitter. Several spikes are captured here showing the random nature of the events. Fig. 5. Single captured random jitter event showing the response of the output power and phase. The events are characterized by an initial sudden excursion followed by a slow recovery to the steady-state condition. by a slow recovery of typically 0.1 to 1 s to the steady state condition, was found to be characteristic of this phenomenon. The isolated event appeared very similar to a single cycle of the sawtooth waveform found in periodic jitter. In order to fully understand the cause and effects of random jitter, a very detailed investigation was undertaken and the nature of the random jitter events was carefully examined. Within a single model or type of TWT, the characteristics (rates, rise and fall times) were observed to be consistent from tube to tube. For different TWT models, however, there was significant variation. Certain TWT designs appeared more sensitive to the beam perturbation events resulting in larger amplitude power and phase glitches. As with periodic jitter, some dependence on focusing and environmental conditions was also observed. The PIC code simulation runs [14], discussed in the previous section, showed a dramatic release of ions once a critical level of ion density is reached. The clear difference in Fig. 4 between the fast onset and slow recovery suggests that the random jitter event is initiated by a sudden release of the trapped ion population on millisecond time scales. This is followed by a slow build-up or recovery period with 0.1 to 1 second time scales that reestablishes the population of ions in the traps to their steady-state level. This response and the time scales involved were very suggestive of a mechanism involving the sudden release and ionization of gas in the TWT. This event would then trigger the 85 release of the body of trapped ions resulting in the initial, sharp response of the event. Slow ionization of the background gas then refills the traps to their steady-state level. During operation of the tubes in the small signal regime, the magnitude of the random jitter events ranged from 0.05 dB to 0.5 dB in the output power, while the phase could shift by several degrees. The shape of any individual event was very similar whether observed in output power, phase or helix current. Within a single device, the direction of the initial excursion would remain constant from event to event; however, it was found that with different devices the amplitude or current could either increase or decrease during the jitter event. Since the number density of ions released in a single event is equivalent to one-half of the beam density, an estimate of the beam density ( ) and the trap volume (0.1 ) as for the TWTs in our study provides an estimate of the number of ions produced per random jitter event. From the known rates ( 1 per min), worst-case estimates of the number of ions released over the life of the tube can be made and the impact on the cathode calculated. The calculations demonstrated that these events would not adversely affect cathode life or performance. Life-tests with cathodes operating at an accelerated temperature [15] with the presence of random jitter confirmed this analysis, showing no negative impact on cathode life or TWT performance. Life-tests with the cathode at nominal temperature were also started one to two years ago, and to date have not shown any negative effects due to random jitter. Questions concerning the effect of random jitter on communications performance of the TWT were addressed. The magnitude of these jitter events (from 0.1–0.5 dB in the output power and 1 to 2 degrees in phase) was typically at or below the discrimination threshold of the communication systems. In addition, the relatively slow nature of the events allowed them to be tracked and corrected by the communication system. It should be pointed out, however, that the ability to control any type of jitter or noise will become more important as the demands of these systems continue to increase. IV. RANDOM JITTER CAUSE Determination of the precise cause and the ultimate elimination of random jitter were of primary importance to this investigation. Various alternatives to a gas release mechanism were investigated. Detailed studies of every component of the TWT to identify possible gas sources were undertaken. Fig. 6 shows data that originally suggested to the authors that the cathode was the component responsible for the gas release events in the tube. It was discovered that the cathode temperature dramatically affected the number of random jitter events observed per minute. It was found that the rate dependence followed an Arrhenius behavior; that is, the logarithm of the random jitter rate varied linearly with the inverse of the true cathode temperature. This behavior is typical of gas emission and indicated that the cathode was indeed the source of the gas bursts responsible for random jitter. Further investigations of the cathode behavior produced a direct observation of these bursts of gas with a line-of-sight mass spectrometer. Furthermore, the species of the released gas was identified as argon, 86 IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 48, NO. 1, JANUARY 2001 Fig. 6. Dependence of the number of random jitter events observed per minute on the brightness temperature of the cathode, suggesting the heated component as the source of the gas burst involved in the jitter process. which had been embedded in the M-layer during improper deposition of the layer. Specific process changes were implemented that eliminated the Ar gas from the cathode and, subsequently, random jitter from the TWT. There have been no further incidences of random jitter since these process changes were put in place. V. REDUCING ALL FORMS OF JITTER From the present understanding of the jitter phenomena, it is clear that in order to eliminate it one must either eliminate the gas source (as was done with random jitter) or the ion traps. While the use of the best vacuum techniques can help to minimize the background gas level, complete removal of the gas in TWTs is not possible. In any case, reduction of the background gas would only be expected to reduce the frequency of this form of jitter. The magnitude of a jitter event depends on the depth and location of the ion traps and the sensitivity of the TWT design to the presence and release of ions. While design modifications resulting in the complete removal of ion traps may be challenging, if not simply impractical, this approach is needed in order to reduce the impact of jitter on TWT performance. Such steps have been taken at HED and found to be very effective in reducing the amplitude and phase change that results from both forms of jitter. In Fig. 7, a comparison of the jitter level, with and without the design modifications, is made. Initially [Fig. 7(a)], the TWT displayed significant levels of both periodic and random jitter. The TWT was then modified by removing the gun, implementing several design changes to minimize or eliminate the ion traps, reattaching the original gun (and cathode), and then reprocessing the TWT. The subsequent level of jitter, shown in Fig. 7(b), was dramatically reduced and is now lost in the measurement noise. VI. CONCLUSIONS Our present understanding of low frequency ion disturbances, called periodic jitter, in TWTs has been presented. The presence of ions, ion traps and detrapping mechanisms lead to a relaxation oscillation responsible for periodic jitter. The precise waveform Fig. 7. Reduction of random and periodic jitter using comparable vehicles. A) With ion traps and poor focusing—jitter is apparent. B) Design changes to remove ion traps and improve focusing—jitter is no longer detectable and is now in the noise. The cathode temperature in both cases was elevated resulting in a 40x increase in random jitter occurrence rate.  that results from the oscillation has been found to be very device dependent. We have observed a new type of ion disturbance termed “random jitter.” The cause and characteristics of random jitter have been identified following a very extensive investigation. The cause was found to be the random release of Ar gas from the cathode. Specific changes to the cathode production process corrected the problem and eliminated random jitter from the TWT. Techniques employed to reduce the size of ion traps also significantly reduced all forms of jitter. With fluctuations in power at the level of 0.1 to 0.5 dB and phase events at 1 to 2 degrees, random jitter had no significant effect on digital traffic or the operation of the communications systems. The extremely small quantity of gas released during the random jitter event did not result in any mission life issues related to the cathode or the traveling wave tube. At present, HED cathodes are being produced in a new, state of the art facility, and tests are in place to ensure quality. The present generation of cathodes is exhibiting very good work function and emission stability. TWTs using these cathodes do not exhibit random jitter; it has been eliminated. ACKNOWLEDGMENT The authors would like to acknowledge the hard work of the investigative team, including S. Bosma, D. Danner, W. Rubio, J. Vaszari, F. Churchill, B. Hannon and B. Campbell. They would also like to recognize R. Longo, D. Komm, A. S. Gilmour, and E. A. Adler, for their helpful discussions and contributions to the work presented here. TIGHE et al.: TRANSIENT ION DISTURBANCES REFERENCES [1] J. R. Pierce, “Possible fluctuations in electron streams due to ions,” J. Appl. Phys., vol. 19, pp. 231–236, 1948. [2] E. G. Linder and K. G. Hernqvist, “Space charge effects in electron beams and their reduction by positive ion trapping,” J. Appl. Phys., vol. 21, pp. 1088–1097, 1950. [3] K. G. Hernqvist, “Space charge and ion-trapping effects in tetrodes,” Proc. IRE, vol. 39, pp. 1541–1547, 1951. [4] A. S. Gilmour Jr., “Principles of traveling wave tubes,” in Ion Effects in Electron Beams. Boston, MA: Artech House, 1994, p. 625. [5] C. C. Cutler, “Spurious modulation of electron beams,” Proc. IRE, vol. 44, pp. 61–64, Jan. 1956. [6] A. D. Sutherland, “Relaxation instabilities in high-perveance electron beams,” IRE Trans. Electron Devices, vol. ED-8, pp. 268–273, Oct. 1960. [7] V. I. Volosok, “On a certain type of relaxation oscillations in electron beams,” in Proc. Acad. Sci. USSR, vol. 4, 1960, pp. 1019–1022. [8] T. G. Mihran, “Positive ion oscillations in long electron beams,” IRE Trans. Electron Devices, pp. 117–121, July 1956. [9] D. W. Maurer, E. E. Francois, and A. Zacharias, “A solution to the problem of spontaneous deflection of the beam in TWT’s with low helix voltage,” IEEE Trans. Electron Devices, vol. ED-13, pp. 40–44, Jan. 1965. [10] E. W. McCune, “Ion oscillations for pulsed klystron amplifiers,” in IEDM Tech. Dig., 1983, pp. 148–150. [11] Y. Y. Lau, D. Chernin, and W. Manheimer, “A note on noise in linear beam microwave tubes,” in APS Conf. Plasma Physics, 1998. [12] D. Thelan, “Experimental observations of ion noise,” in APS Conf. Plasma Physics, 1998. [13] W. M. Manheimer, “On the theory of ion noise in microwave tubes,” IEEE Trans. Plasma Sci., vol. 27, pp. 1146–1163, July 1999. [14] C. B. Thorington, “Computer simulation of ion trapping and detrapping in a PPM focused travelling wave tube,” IEEE Trans. Electron Devices, vol. 48, pp. 56–61, Jan. 2001. [15] R. T. Longo, E. A. Adler, and L. R. Falce, “Dispenser cathode life prediction model,” in IEDM Tech. Dig., 1984, pp. 318–321. [16] A. S. Gilmour Jr., “Rev. section 7.5.3 ‘Low frequency instabilities’,” unpublished, 1998. [17] K. G. Hernqvist, “Space charge and ion-trapping effects in tetrodes,” Proc. IRE, vol. 39, pp. 1541–1547, 1951. [18] Y. Y. Lau, D. Chernin, and W. Manheimer, “A note on noise in linear beam microwave tubes”, to be published. William Tighe received the B.Sc. degree from Carleton University, Ottawa, ON, Canada, and the M.Sc. degree in physics and Ph.D. degree in electrical engineering from the University of Alberta, Edmonton, AB, Canada, in 1980 and 1985, respectively. In 1985, he joined the Plasma Physics Laboratory, Princeton University, Princeton, NJ, where he worked on X-ray lasers and spectroscopic diagnostics of fusion devices. He joined Hughes Electron Dynamics (now Boeing Electron Dynamic Devices, Inc.), Torrance, CA, in 1996, where he studies the physics of thermionic emitters. 87 Dan M. Goebel (M’93–SM’96–F’99) received the B.S. degree in physics in 1977, the M.S.degree in electrical engineering in 1978, and the Ph.D. in electrical engineering in 1981, all from the University of California, Los Angeles (UCLA). He joined the research staff at UCLA in 1982, where he invented the PISCES plasma device, which provided the first experimental laboratory simulation of fusion confinement-device edge and divertor plasmas. In 1986, he co-founded PMT, Inc., a large manufacturer of plasma-processing equipment for the thin-film and semiconductor industries which is now called Trikon. In 1987, he led an international fusion research program on edge-plasma characterization and plasma-facing components in the TEXTOR tokamak at KFA-Jülich in Germany. In 1989, he invented the APS plasma source, which is used extensively by Leybold, Germany, in their optical coating and ion-assisted deposition systems. In 1988, he joined Hughes Research Laboratories where he led research and development programs on novel high-power plasma-filled microwave sources, plasma-cathode E-guns, and pulsed-power plasma switching devices. In 1997, he moved to Hughes Electron Dynamics (now Boeing Electron Dynamic Devices, Inc.), Torrance, CA, to work on novel microwave sources for communications applications and investigate the role of plasma in various types of microwave devices. He has won several outstanding technical publication and patent awards while at Hughes. He is the author or co-author of 70 technical journal papers, over 50 conference papers and reports, two book chapters, and holds 32 patents with five patents pending. Dr. Goebel received the IEEE Nuclear and Plasma Sciences Graduate Student Award in 1979, and was named the Outstanding Ph.D. Candidate in Engineering at UCLA in 1981. He is a life member of the American Physical Society and a member of Sigma Xi Research Society. He is an Editor for the IEEE TRANSACTIONS ON ELECTRON DEVICES, and has acted as Chairman and Program Chairman for several technical conferences and workshops. Charles B. Thorington joined Hughes Aircraft Company in 1981, working at various times for Hughes Electon Dynamics Division, Hughes Research Laboratory, and Hughes Radar Systems. While at Hughes Radar Systems, he rewrote Hughes’ code for processing ISAR data in real-time air-to-air imaging, and developed an advanced ray-tracing program in which the electric fields and the principal radii of curvature of astigmatic ray-tubes could vary across the cross sections of the ray-tubes. At HED, he has written both steady state and time-dependent electron gun and depressed collector design codes, and is currently working on a completely 3-D signal code for helix traveling wave tubes. He is currently with Boeing Electron Dynamic Devices (formerly Hughes Electron Dynamics), Torrance, CA.